po

INTEGRATED DESIGN
AND OPERATION OF
WATER TREATMENT
FACILITIES
SECOND EDITION

Susumu Kawamura

New M okawk waler Ircal mcnl plam ( 100 mgd). Tu lsa, Oklahoma.

JOHN WILEY & SONS, INC.

New York • Chichester • Weinheim • Brisbane • Singapore . Toronto
 CONTENTS

Preface xv

I. Management of a Design Project

1.1 Role of Ihe Professio nal Eng ineer
1.2 Selecl ion of Consullanls 2
1.3 Fee Structure for Serv ices 3
1.4 Projeci Conl ro l 5
1.5 Value Eng ineerin g 10
1.5. 1 Scope and Melhodology of Value Eng ineerin g II
1.5.2 Informati on Phase 12
Copyright © 2000 by John Wiley & Sons, Inc. All rights reserved.
1.5.3 Creali ve Phase 12
Published silllulianeolisly in Canada. 1.5.4 Judgmenl Phase 13
1.5.5 Developmenl Phase 13
No part of this publication may be reproduced, stored in a retrieval system or transmitted in any 1.5.6 Recommendation Phase 13
fortn or by any mea ns, electronic. mechanical, photocopying, record ing, scanning or otherwise,
except as permilled under Sect ions 107 or IDS of the 1976 United States Copyright Act, without 1.6 Privatization Delivery Process 13
either the prior written permission of the Publi sher. or aut horizat ion through payment of the
appropriate per-copy fcc to the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 2. Preliminary Studies IS
01923, (97S) 750-8400. fax (97S) 750-4774. Requests to the Publi sher for permission shou ld be
2. 1 Masler Planning 15
addressed to the Perm issions Department, John Wiley & Sons, Inc .. 605 Thi rd Avenue , New York,
NY 10 158-00 12. (2 12) 850-6011. fax (212) 850-6008. E·Mail : PERMR EQ @WILEYCOM _ 2.2 Feasibility SlUdy 16
2.2. 1 Planning Period 16
Th is publicat ion is designed to provide accurate and authoritati ve informat ion in regard to the 2.2.2 Water S upply Areas 16
subject matter covered . It is sold with the understanding that the publisher is not e ngaged in
2_2.3 FUlure Populalion 17
rendering professional services. If professional advice or other cxpert assistance is required , the
serviccs o f a competent professional person should be sough1. 2 .2.4 Maximum Daily Waler Demand 17
2_2.5 Evalual io n and SClcclion of Ihe Waler Sou rce 17
Library of COl/gress Calalogillg-ill-Publication Data: 2.2.6 Size of the Water Treatm ent Plant 17
Kawamura, Susumu . 18
2.2.7 Trealmenl Planl Si le
In tegrated design and operation of water treatme nt facilities I Susumu Kaw<ll1lura.-2 nd cd.
2_2 _8 Financ in g 18
p. Clll.
In cludes bibliographi cal reference and index. 2.3 Bench-Scale and Pilol-Planl Siudies 20
ISBN 0-47 1-35093 - 1 (cloth:alk. paper) 2.3.1 Bench-Scale Siudies 20
I. Water treatment plants- Design and construction . I. Tit le .
TH4538 .K382000
2.3_2 Pilol-Plant SlUdies 22
628. 1'62- dc2 1 99-089640 2.4 Preliminary Eng ineeri ng SlUdy 26
2.4.1 QualilY and Trealabi lily of Ihe Raw Waler 27
2.4.2 Objecli ves for Ihe Fini shed Waler Qualil y 28
10 9 8 Swface Water Treatment Rille 29
v
vi CONTE NTS
I CONTENTS vii

3.2.4 Flocculation Process 105

Disinfectants/Disill/eclioll By-products Rule 29
E"hanced Coagulatioll RequiremeJ/ts 30 Plllpose 105
Enhanced Swface Waler Treatmellt Rule 31 COllsideratiotls 105
Other Water Quality- Related Itellls 32 Type alld Selectioll Guide 109
Federal Drinkillg Wafer Regulatiolls 33 Discllssion of Alternatives 1/3
36
Design Criteria 1/ 7
2.4.3 Additional Goals and O bjec ti ves
38
Example Design Calclllations 126
2.4.4 Restrictions and Constraint s on Plant Des ign
38
Operation alld Maintenance 137
2.4.5 Alternative Treatment Processes
Preliminary Design Criteri a for Treatment Unit 3.2.5 Sedimentation (Clarification) Process 139
2.4.6
Processes 39 Purpose 139
2.4.7 Hydrauli c Grade Across the Plant 39 Considerations 140
2.4.8 Geotechnical Considerations 41 Type alld Selectioll Guide 166
2.4.9 Structural Design Conditions and Criteri a 41 Discussion of Altern atives 168
2.4. 10 Plant Waste Handli ng and Disposal 49 Design Criteria 171
2.4. 11 Instrumentation and Control Systems 50 Example Design Calculations 172
2.4. 12 Preliminary Cost Estimates 51 Operatioll alld Mailllenallce 184
2.4. 13 Recommended Treatment Processes 53 3.2.6 Dissolved Air Flotation Process 189
2.4. 14 Plant Layo ut and Architectural Desig n 54 Purpose 189
2.4. 15 Management Planning of the Design Phase 55 Considerations 189
2.4. 16 Environmental Impact Statcment 56 Type and Selection Guide 192
2.5 Watershed Management 56 Design Criteria 193
3.2.7 Granular Mcdium Filtration Process 194
3. Design of Basic Treatment Process Units 59 Purpose 194
3. 1 Common Basic Water Treatment Processes 59 Considerations 194
59
Type and Selection Guide 206
3. 1.1 Design Procedure
Basic Hydra ulics 225
3.2 Basic Approach 61
Design Criteria 229
3.2. 1 Introduction 64 Example Design Calculations 244
3.2.2 General Considerations 64 Operation and Mailllell all ce 260
Plant Layout 65 3.2.8 Specific Types of Filter 28 1
Pmcess Flow Diagram 66
SlolV Sa"d Filters 281
Prelimillmy Hydraulic Grade Line Acmss the Plallt 68
Pmprietary Fillers 286
Design of the Basic Pmcesses 69
Two·Stage Filtratioll 287
Modification of the Existing Plant 73
Miscellaneous Filters 290
3.2.3 Coagul ation and Chem ical Application Considerat ions 74
3.2.9 Disi nfection Process 292
Pll1pose 74
Purpose 292
Considerations 75
Considerations 292
Type and Selectioll Guide 84
84 Alternative Disinfectants 293
Discllssion of Alternatives
88 Chlori"e 293
Design Criteria
CIlIol'{lmines 295
Chemical Diffllsers 89
Chlorin e Dioxide 295
Enhanced Coagulation Process 89
Ozone 296
Example Design Calculations 90
Ultraviolet Light 297
Operation and Maintenance 101
viii CONTENTS CONTENTS ix

Design Criteria 30 1 COllsitJeratio/ls 384

Dechlorination 309 Sludge Characterisfics 386
Example Design Calculatioll s 309 Redltctioll of Sludge Productioll Rate and Volume 387
Operation amI Maillfenam:e 310 SllIdge Handling 388
313 Ultimate Disposal 39 1
:1.2. 10 Fi ltered Water Reservo ir (Cleal"Wcll )
313 Design Criteria 394
Purpose
314 Example Design Ca /culations 397
Considerations
316 Operatioll and Mail/tellallce 400
Type and Setectioll Glfide
318 4.4 Intake System
Des igll Criteria 404
318 PllIpose 404
3.2. 11 Instrumentati on and Process Control
318 Consideratiolls 404
Plfrpose Rive r /makes
Symbols and Manufacturers 319 405
319 Lake and Rese rvoir Intakes 405
COll sideratio/ls
325 Groul/dwater Intakes 406
Commoll Measurements
326 Design Criteria 406
Basic COli trois
327 Operation alld Ma;mellallce 412
Computer-Based Monitorillg and COlllro l
Design of the Illstrum entatioll mill COlll rol System 330 4.5 Gri t Chamber 4 13
Security System 334 PUl pose 413
335 COIIsideratiOlls 414
3.2. 12 Electrical System
335 Desigll Criteria 416
Purpose
335 Operatioll and Mail/tel/ance 416
COIlsideratiolls
336 4.6 Operations Building
Desig/l Criteria 4 17
PIt/pose 417
343 COllsideratiolls
4. Subord inate Plant Facili ties 417
343 Design Crite ria 422
4. 1 Chemi cal Feed System
343 4.7 Pumping System
Purpose 423
Water Trea tment Chem icals 343 PllIpose 423
COIlsideratiolls
345 COl/sideratiol/s 423
Overview of the Chemical Feed Systems 346 PUIllP Specifications 426
Required Capacity for the Chemical Feeder alld Storage 354 Special Considerations 428
Specific Design Issues
356 Operatioll al/d Mail/tel/alice 43 1
Desigll Criteria
358 4.8 Ozonation System 433
Risk Management Plalllling
36 1 PllI pose
362 433
Example Design Calculatiolls Considerations 434
Ope ratio/! and Maintenance 367 Design Criteria
372
447
4.2 Handling Facility for Waste Wash Water Exalllple Design Calculations 449
Purpose
372 Operation al/d Mailllen(fllce 453
COllsidemtiolls
373
377 5. Design of Plant Components
Design Criteria 455
Example Design Calclllatiowi 380 5. 1 Flow Measurement
382 455
Operatiol/al Issues PIlIpose
384 455
4.3 Sludge Hand ling and Disposal COl/sideratiol/s 455
384 Types of FloHl Meters
Purpose 455
x CONTENTS CONTENTS xl

Example Design Calculations 460 Direct Costs 504

463 Indirect Costs 504
5.2 Level Measurement
463 Estimate SummGlY 504
Purpose
464 6.4 Supplemental Studies 505
Float System
Pressure Elemellls
464 CeotecJlIlical/Ceological Study 507
Bubbler Tube System 464 Cold Weather Desigll 508
Ultrasonic Level Detector 465 Co n vsioJ/ Swdies 509
Valve Selection 466 Acoustic Studies 509
5.3
466 SlIrveying 509
Pllrpose Landscaping
Considerations 466 509
Types of Valves 468
469 7. Specific Water Treatment Processes SIO
Flow Cofltml Valve Considerations
Cavita tion Considerations 475 7.1 Lime- Soda Ash Softening 510
Example Design Calculations 479
7.1.1 General Discussion 510
Piping System 481 7. 1.2 Basic Chemical Reactions
5.4 511
P1lI1JOse
481 7.1.3 Softening Process Alternatives 513
Considerations
481 Altem ath1e I: Partial Lime Softening 513
Raw Water Lin e 482 Alternative 2: Excess Lime So/telling
482 513
Plallt Yanl Pipillg Alternative 3: Lime- Soda Ash Softellillg 514
Plant Utility and Wat er Distribution Lin es 486 Alternative 4: Caustic Soda Softellillg
486 514
Chemica l Feed Lin es Alternative 5: SO/Ienillg with Both Lime and
MiscellaneoIls Pipings 487 Caustic Soda
487 514
Pipin g Materials 7.1.4 Overall Softening Treatment Process 514
Standby Power Supply 488 7.1.5
5.5 Design Criteria 516
488 7. 1.6 Operational Considerations
Purpose 518
Considerations 489 Chemical Dosages 519
491 Sludge Halldlillg alld Disposal 520
6. Elements of Detailed Designs 7.1.7 Example Calculations 520
Plant Hydrau lics 49 1
6. 1 7.2 Iron and Manganese Removal 523
Purpose
491
7.2. 1 Alternative Met hods 524
Considerations
491
492 Oxidation 524
Example Desigll Calcllla tiolls
1011 Exchallge (Zeolite Process) 526
6.2 Specifications 497
Sequestering Process 527
Contellls 0/ Specifications 497 Lime So/tellillg
498 527
Types of COlltracts 7.2.2 Case History 527
Con tract Revisions 499
500 7.3 Taste and Odor Con trol 530
Stallda rd Specificatiolls
Master Specifications 500 7.3. 1 Major Taste- and Odor-Producing Substances 530
Genera t Provisions 500 7.3.2 Control Measures 531
50 1 Preventive Measures 531
Technical Specifications
Construction Specificatiolls In stifllte 502 Removal at the Treatm ent Plallt 534
Final Cost Estimates 502 7.3.3 Hyd rogen Sulfid e Removal 538
6.3
Development of Strategy and Data 503 Aeration 539
 CONTENTS xIII
xii CONTENTS

Chlorillatioll 540 pH COlltrol by Alkali Cllemicals 563

540 7.6.3 Corrosion Coupon Tesl 565
Ozollatioll
Potassium PemulIlgal/ate Oxidation 540 7.6.4 Corrosion (Degradalion) of Concrele 566
541 Pllysical Callses 566
7.3.4 Control in the Distribution System
Disinfeclion By-producl Conl rol: THMs, HA A" and vacs 543 Cllemical Attacks 566
7.4
543 Bacterial Action 567
7.4.1 THMs
543 7.6.5 Special Construction and Design Considerations 567
Ill stantaneolls THM
Terminal THM 543 7.7 Granular Acti vated Carbon Adsorption 569
THM Formatioll Polelllial 544 7.7. 1 Purpose 569
Maximum Total THM Potelltial 544 7.7.2 Characteristics of GAC 569
THM Control Measures 545 7.7.3 Design Considerations 571
7.4.2
546 7.7.4 GAC Rege neralion 575
7.4.3 HAA,
Operational Issues 547 Steam Regenera tion 576
7.4.4
Quarterly Samplillg [or Sill/ace Water SlIpplies 547 Th ermal Regeneratioll 576
Quarterly Sampling for Gro uI/dwater Supplies 547 Chemical Regeneratioll 577
548 7.7.5 Capital and Operational Costs 577
7.4.5 vacs
Trichloroethylelle 548 7.8 Ion Exchange and Membrane Separation Processes 579
Tetracllioroethyiell e 548 7.8.1 Ion Exc hange 580
Carbon Tetrachloride (TetrachlorometllOll e) 548 Characteristics 580
I, /, /-Tricilloroetilalle (Metllylchloro[orm ) 548 7.8.2 Membrane Separation Process 583
1,2-Dichloroetllalle 549
549 Microfiltratioll (MF) 585
Villyl Cilloride (Chloretllylell e) Ultrafiltratioll (UF) 585
Bellzelle (Bellzol) 549
Nallofiltratioll (NF) 588
549
7.4.6 voe Control Measures Reverse Osmosis (RO) 588
Managemell t COlltrol Measures 549 Electrodialysis (ED) 591
Treatment Control Measures 549 7.9 Inorganics (Toxics) Removal Process 595
Fluoridation and Fluoride Removal 553
7.5
554 8. Management of Procurement and Construction Phases 596
7.5. 1 Fluoridation
General Considerations 554
8. 1 Procuremenl Phase (Bid Phase) 596
Example Design Calculatioll 555
557 8. 1.1 Bid Adverlising 596
7.5.2 Fluoride Removal 8.1.2 Bid Opening 597
Chemical Precipitation 557 8. I.3 Bid Evalualion and Award of Conlracl 598
1011 Exchange Process 557
8.2 Construction Phase 598
Corrosion Control
559
7.6 8.2.1 Traditional Architect/Engineer Contract 598
7.6. 1 Corrosion of Metals 559 8.2.2 Design/Construction Manager Contract 598
General Corrosion of Meta ls 560 8.2.3 Professional Construction Manager Contract 598
Secondwy Corros ion Facto rs 561 8.2.4 Design- Build Contract 599
562 8.3 Construction Management 599
7.6.2 Corrosion Control for Metal s
Desigll Efforts 562 8.4 Responsibi lily of Residenl Engineers 600
Applicatioll of Corrosion Inhibitors 562 8.5 Prolesls and Claims 601
Cathodic Protectioll 562 8.6 Projecl Closeolll 601
Sacrificial Anodes 563 8.7 Major Closeollt Activ ities 602
xiv CONTENTS

8.8 Construction Phase Chec klist 603

9. Operations and Maintenance Manual and Operator Training 60S

9. 1 Opcrati ons and Maintcnancc Ma nual (O&M Manual ) 605 PREFACE

9.2 Operator Training Sefore Plant Startup 6 10

10. Plant Startup and Follow-Up Services 611

10. 1 Startup 6 11
10.2 Feed bac k 611 My book cntitled Integrated Design of Water Treatm ent Facilities was published in
10.3 Operator Training Serviccs 6 13 1991 . lt has ga incd a good reputati on alllong professio nal wa ter trea tme nt engineers,
10.4 Plant Opera ti on and Managcment Serviccs 6 14 plant operators, as well as many college st ud en ts bcca use it covers a broad spec trum
of water trcatm ent process dcsign (in detail ) and important operational iss ues. 1 have
Appendixes 616 tri ed to make thi s book uniquc by incorporatin g many practical idcas that are not
cove red in many dcsig n books. During th e past 49 years I have bee n in volved in all
I. Abb rcv iat ions 6 16
facc ts of watc r trea tm e nt iss lles-pil ot studi es, dcsig n work , co nstru c ti o n
2. Sasic Physical Tables and G reek Alphabet 62 1
managc ment , and fOll r years of act ual plant opera ti on- in va ri ous reg ions (U nited
3. Metri c (S I) Syste m and Se lec ted Useful Co nversion Factors 624
States, Asia, SOllth America, and Australi a) and wish to share Illy knowledge and
4. Tcmpcrature Conversions 630
cxpc ri cnce with the readc rs. As a tes tim ony to its va lue and usefulness, th c book has
5. Pcriodic Table of the Elements 631
been translated into Japancse and Korean , and a Spani sh ve rsion is curre ntl y in
6. Sasic Properties o f Air and Water 632
prepara tion .
7. Gas Law 635
Treatment process technology has advanced rapidl y during the past 10 ycars, due
8. Ozo ne Co ncentrati on Convcrsion 636
to var ious rccent trcatment iss ues and more stringen t drinking water qu ality
9. Tracer Test 637
standard s. For thi s reason I have dec ided 10 create a seco nd edition o f the ori gi nal
10. Recommcnded Jar Test Procedure 645
book. T he title has been modi lied to II/teg rated Design alld Operation of Wmer
I I. Core Sampling of th e Filter Sed 654
Tre{/flllent Facilities 10 rcllect th e many important plant operational iss ues that ca n
12. Venturi Flow Meter Selection 657
be ne fit plant operators. The boo k has o nly one author, thu s retaining a uniform style
13. Water Treatmcnt Chemicals and Preparation of for easy readability. Moreovcr, I have tri ed to indica tc clearl y how to sclec t, from
Certain Perccnt Solutions 659
many alternat ives, thc proper unit process and equipment in order to ma ximi ze
14. Hydrauli cs: Basic Data and Formulas 666
ovcrall process e fli cicncy, thereby minimi zing operati on a nd main te nancc e ffort s
15. Disc harge Flow Rate Correc ti on for Submerged Weirs 672
and costs.
16. Head- Disc hargc Curvcs of Telescoping Valves 673
This new ed ition cove rs several new subjects: va lue c nginee rin g, watershed
17. Nomograpilical De te rminatio n of pH and La nge lier Indcx 674
man age mc nt, di ssolved a ir n ota ti on process, liltered water rcscrvo ir (c lea rwe ll)
18. Tertiary Filtcr Performance 675
dcsign, elect ri cal sys tem design , ultrav iolct ray di sinfection, di chl ori nat ion, and four
19. Marston 's Soil Coefficicnt for Pipe Trench Condition s 676
new appe ndi ces of useful informatio n. There has also been a signili ca nt degree o f
20. Load Cocfli c ie nt for Pos iti ve Projecting Embankment Condition 677
Illodilica tion and addition for cert ain subjec ts: privatiza ti on de li very process
2 1. USA Standard Sieves 678
objec ti ves for lini shcd water quality, treatment process selcction, enha nced
22. Standard Elec tri c Motors and Their Characteris ti cs 679
coag ul ation , hi gh·speed microsa nd settling process, basic hydra uli cs o f filtrati on,
23. Geometri cal Formulas 680
tastc-a nd- odor co ntrol , instrume ntati on and cont ro l, di sinfec ti on process, ozona ti on,
24. A n Exa mple of a Coarse Deep Sed Filter Design 68 1
disinfec ti on by· product co ntrol , GAC process, and membrane liltrati on process.
The hi story of modern water treatm ent fo r the publi c wa ter supply began
Index 683
approximately durin g the mid· nin ctee nth century in Europe with con tro l of
pathoge nic bacteria by c hl orinatio n. Several decadcs late r, thc lI SC of che mi cal
floccu lation, sedim ent ati on, and the rapid sa nd liltration process tra in (prior to
chl orinati on) were established in the United Stat cs early in the twe ntie th century to

xv
xvi PRE FACE

make public waters sa fe from pathogenic bacteri a. No significant changes occurred

until the early 1970s; the Safe Drinking Water Act (1 974) ushered in a new era for
th e fie ld of water trea tment in the United States by controlling disinfec ti on by-
products (DBPs) created th rough the use of chl ori ne disin fec ti on as well as
ACKNOWLEDGMENTS
pathogens. Since the mid-1 970s new meth ods, such as ozonation, the enhanced
coag ulati on process to control precursors of DBPs, the C X t concept for good fin al
di sinfec ti on, and high rate filtration w ith du al med ia and coarse deep media beds
with sy nth eti c polymers, have been developed to meet the requirement s. H owever,
another new water trea tmen t issue emerged around 1990 due to outbreaks of
CI)lplosporic/osis in both the United States and other maj or developed co untries. Thi s book is dedicated to my parents, for th eir guidance, excepti onal caring . and
Since the Clyptoparvlflll oocyst cann ot be inac ti vated by regular chlorin ati on, there devotion to providi ng a ll th eir childre n an edu cati on to the hi ghes t level.
Illu st be a high degree of removal of suspended solids, includ ing the oocysts, pri or I would like to express my deepes t apprec iation to my daughter, Mika Kawamura,
to fin al chlorin ation or even ozonati on. M oreover, th e IiIter wash was te also requires for her devoted efforts in editing a nd organizing th e manu script. I am also grat efu l
a high degree of treatment prior to recycling. to Mrs. Debbie M ontesanti for typin g the manu script. Particul ar appreciation is
M embrane filtrati on, whi ch has been prac ticed in both the medical fi eld and given to many of my colleag ues, especiall y to Murli Tolaney, Chairma n of the Board
certa in food industri es, has now gained recogni tion as a new water trealment of M ontgomery Watson, I nc .• for hi s encou ragement. Fin ally, I wish to express my
process, due to the effecti ve removal of almost all suspended solids, including all sincere tribute to my wife, Rei ko, and my elder brother, D r. Shigeru Tanaka, for their
pathogens, such as C,yptosporidilllll oocy sts, without the treatmen t processes of spiritual support and understanding.
coag ulati on and sedimentation. M embrane nitration is also a highl y automated
system that saves operator time and attention and allows for easy di sposal of
residuals. Despite the few limitati ons of the membra ne process, it will undoubtedly
become an importan t as well as prac tical water treat men t process into the new
millennium .

SUSUMU K AWAMURA

Sail Cab riel, CA

JlIll e 2000

xvII
INTEGRATED DESIGN
AND OPERATION OF
WATER TREATMENT
FACILITIES
 CHAPTER 1

Management of a Design Project

1.1 ROLE OFTHE PROFESSIONAL ENGINEER

For projec ts that arc nol pri vati zed, the owne r mll st pre pare a budget to lin anec the
o perat io n and a time schedul e for compl eti o n. T he desig n ph ase o f most muni cipal
wa ler treatm ent pl ants lypi call y costs 5 to 10% o f the estim ated co nstructi o n cost.
Th e lime requi red to d es ig n and construc t a treatmen t pl ant is approx im ate ly ::\ years,
including a 9- to 12- l11onlh design peri od as practi ced in the United States. If the
ow ne r o f the water treatment pl ant wishes to restrict bot h the cos t and time req uired
to compl ete the proj ec t to these li g ures, he mus t hire wc ll -q ualili cd co ns ulti ng e n-
gineers or equi pment manufac turers and depend on their ex peri ence, j udgment, and
ingenuit y.
Mos t Euro pea n and As ia n countries generall y rely on eit her equipm ent manu -
facture rs o r co ntrac tors o fferin g packaged propri etary processes to des ig n, co nstruct,
and start opera ti on o f a water trea tmc nt pl a nt. T he owner or hired co ns ulta nt o ft en
establishes the pl ant site, th e hyd rauli c profil e across th e uni t processes, a nd a few
bas ic des ig n criteria, s uch as the hydraul ic loadin g rates for the bas ins and the til -
tel'S and the fini shed water qu ality s tand ard . Based 0 11 the criteria establis hed, the
successful bidder the n furni shes the major plant process fac ilities and all anc ill ary
fac ilities req uired to compl ete the pl ant.
In th e Euro pea n and As ian method, tend ers are in vit ed to bid o n the projec t a f-
ter onl y a few des ig n crit eri a have bee n estab lis hed. Th e tend er has th e frcedo m to
choose the process con li g urati o ns, plant layo ut , ty pe o f process unit s, num be r an cl
dim ens io ns o f the unit process, mode o f the control sys te ms, and building materi -
al s. Th e hyd rauli c load ings o f the trea tm ent processes are occasio nall y upg raded
from th ose dictated by the ow ners by submittin g the res ults o f pilo t studi es th at meet
the c riteri a establi shed for the linishcd water qu ality. The proposals arc th en evalu -
ated by the ow ner and judged according to the ir qu ality a nd adcqu acy and the csti-
mated desig n a nd co nstructi o n COS IS.
Few large publ ic wa ter purveyors in Europe or As ia arc capabl c o f performing
all phases o f a projec t, inc ludin g detailed des ig n, constru cti o n man age ment , and
pl a nt startup, lI s ing their own staff. In many cases, detail ed predes ig n studi es s uch
as evaluat io n o f altern ati ve processes and s ite se lec ti o n tend to be less thoro ug h tl1[ln
work perform cd by cons ulta nts in the Un ited States. In regard to Illunic ipal and puh-
li call y ow ncd water treatm ent pl ant s. s tand ard prac tice in the United States is ge n·

1
2 MANAGEMENT OF A DESIGN PROJECT
c rall y quite different fr0111 that in Europe, except in the fo ll ow i ~ g cases: sn~a.1I plants
processing less tha n 10 million ga llons per day (mgd) (0 .4 I\,- /s) of mUnlctpal wa-
ter, any size plant used in the industri al ficl~. Of w aler ~rc~tlnent pla~ts own ed .~Y
investors such as land developers. In the United States It IS Clistommy for p~ot cs­
'onal e ngineers to complete 80 to 85% o f the detailed e nginee ring design pTior to
Sl , . I . I . b Ik f the de-
bidding. In co ntrast, the European and Asmn approach comp c l es t 1C U 0
tailed des ign (70 to 80%) after bidding by the contrac tor.
In the United States. professional e ngineers arc reqUiTed t~ pr~pare d?c umc nts
that arc co mpl ete with a series of draw ings and a sct of specificati ons pnor to be-
ing in vited to bid on a project. The drawin gs al one may tota l 250 to 500 sheets o~
civil , architec tural, struclU ral, mechanical, a nd elec tn cal dCSl g ~ . as we ll as drawlO.g
of the instrum entation and land scaping. The printed s~cific~tlO.ns may ~e I to 3 Ill .
thi ck prior to bidding and must present the variou s deSign cntena es tabl.lshed by the
pro fessional e ngineers: the hydrauli c profil e across the plant, the van ous .p rocess
trains, and the site plan. This ~ocu~ent shoul~ als~ includ e. a detai led deSlgn t:~at
j
defines the exact number, configuratIOn, and size 01 each unit process,. yard p p g
sc he me, civil works, and necessary ancillary fac ilit.ies. Once. the bid has been
awarded, the e ngineer merely reviews the shop draw mgs submitted by the manu-
facturers and manages the general constructi on work. . . . fee stru ctu re of a projec t, as this is in the rea lm of the projec t engineer or his or her
In the United States a co nsultin g engineer is a profeSSIO nal who I.S ex penellced
in appl ying scientific prin ciples to engi neering proble ms. As a profess lO~al , the con-
sultin g engineer has a duty to both the public and the. cli~nt. Thus the eng.lIleer s h~lIld
be registered, that is, pass the required state examlJl atlOn s. The Al11 e r~ can ~oc l ety
of Civil Engineers (ASCE) publishes a guide on how to engage the engmeeTing ser-
vi ces of consulting engineers; refer to Ma nual 45 ( 198 1 and 1996).
1.2 SELECTION OF CONSULTANTS
In principle, a consultant does not competiti vely bid his or h~r e n g.il~eeri.ng services.
If co nsultants are selec ted on the basis of price, the owner n sks hm.ng Jncompet~ nt
and inadequ ate serv ices. It is therefore important for th~ ow ne r to Illre the best Pl ~­
fessional service that is available by pay ing an appropnate fee. The fee struc ture IS
di sc ussed in Sec ti on 1.3. . .
Ow ners of water treatment plants should follow th e normal procedure 01 select-
ing a professional consultant in the followin g order:
I. Iss ue a reqllest fo r qllalification (RFQ) for the project and review the qualili-
cati ons submitted by eac h firm .
2. Compil e a short list of three to fi ve firm s based on ex perience, kn owledge,
and abilit y to undertake the projec t.
3. Issue a reqllest fo r proposal (RFP) to each Ii Tin selected._Each comPdanYbslh~U:~
be asked to submit a deta iled presentati on of Its qu alificati ons an a 1 .1 Y
undertake the project. The proposal must include inlon~atj o n on th e Size of
the firm , number of staff me mbers, availability of qualified personnel to be
FEE STRUCTURE FOR SE RVICES 3
assigned to the project, and experi ence of these engi neers in similar li nes of
work .
4. Se lec t the most qualifi ed firm and a backup in case contrac t negotiati ons fail
with the first choice.
5. Notify the firm o f its selec ti on a nd begin negotiati ng the fee for the projec t
and a detailed scope of the professional servi ces that are to be re ndered.
Several important points should be evaluated during selecti on of the consult ant:
techn ical qu alifi cati ons of the firm ; personality and administrative skill s of key en-
gineers, such as the projec t manager a nd project e ngineer; existin g workl oad of the
firm (i.e., the ability o f the firm to absorb the additi onal workl oad in relati on to its
capac ity); fin anc ial stability o f the firm ; and experie nce, reputation, and pas t ac-
complishm ents of the firm in si mil ar lines of work .
1.3 FEE STRUCTURE FOR SERVICES
Although most engineers ass ig ned to a proj ec t are not involved in negoti ating the
superi ors, it is still import ant for all e ngineers to be famili ar with thi s subj ect. After
all , design eng ineers are responsible for bot h the fi scal and tec hn ical as pects of the
project. The fees charged by th e finn are based on acceptable compe nsati on cu rves
or ya rdsti cks and the prevailin g standards of the industry. For more deta il on thi s
subjec t, refer to the ASCE Manflal and Repo rt 0 11 Engineering Practice, Mallual
45 , and the Am erican Co nsultin g Eng ineers Counc il 's Compilation of Fees.
ll1ere are four basic method s by whic h fees are determin ed: (I ) salary cost X a
multi plier, plus other direc t costs; (2) cost plus a fixed fee; (3) fi xed lump sum; and
(4) a percentage of the co nstru c ti on cost. Most civil e ngineeri ng contracts are based
on the first two me thods.
I. Salat)' cost X a l1Iultipliel; pillS otller direct costs. This type of contrac t is
comm onl y used whe n the scope o f work cannot be de fined accurately. In thi s me thod,
reimbursement is calc ul ated by multiplying salary cost, payroll fac tor, and a multi-
plier. Salary cos t is defi ned as the direc t payroll plus frin ge benefit s g iven to the
projec t tea m mem bers; total benefi ts are appro ximately 130 to 140% o f actual salary
cost. The multiplier applied to the salary cost is a factor that co mpensates for over-
head and suppli es, a reaso nable margin for continge ncies, interest on invested cap-
ital, insuran ce, and pro fit. This multiplier varies with many factors, including geo-
graphi c area, and is negoti able. T he mult ipl ier is approx imately 2.3 to 2.5 times the
salary cos t but is subjec t to peri odi c reevaluation . For instance, if the average salary
rate o f a projec t tea m is $25 per hour, the pay roll fac tor is 130%, a nd the multiplier
IS 2.4, the payment prov ision in the contrac t should stipul ate that the consulta nt be
reimbursed $78 per hour for the designer's tim e:
($25 X 1.3) X 2.4 = $78 or $25 X 3. 12 = $78
 •

4 MANAGEMENT OF A DESIGN PROJECT

Other identifiable costs or direct nonsalary expe nses include travel, computer time,
telephone calls, subcontractors' charges. printing and binding, and a processing fee
for executing the work. These expenses must be reimbursed by the client at actual
cost plus a service charge of 10 to 15%.
This particular method is not limited to projects with an undefined scope o f work
but may also be applied in situations where the scope is fairly well defined. With
~ ~

thi s type of contract, consultants stand a good cha nce of not losing money.
.' ';
!J
il
.. ........
"'o=8~s:t:;;:s;o
..:~

2. Cost pillS a fixed fee. This type of contract is used in cases where the scope
of the work cannot readily be defined. Yet it is important to define the scope of th e
~ HU~U~
~:
work as completely as possible, for only the n will the owner agree to reimburse the
consultant for direct cost plus a fee. Reimbursable costs include the technical pay-
gi
!~
~ "~U ~ V
roll and actual expenditures for the project that are incurred directly. The fixed fee
-
~
contains profit. nonallowable costs, and allowable costs. The floflaf(owable costs in-
clude contingencies. interest on invested capital, and the avai labil ity of the co nsult- V
ing team . Allowable costs arc defined as the d irect labor costs, direct project costs, ./ ;;
and indircct costs incurred by the labor base. •g UJ
The fixed fee varies with the size of the projec t as well as th e complex ity and U
scope of the job. It is traditionall y calculated as a percentage of engineering cos ts '"
«
and varies from 10 to 25%. Due to the nature of this type of contract, it is impor-
tant to eliminate any increase in the e ngineering fee un less the scope of the engi-
~
= , = . ~
~
E

neering services is revi sed .

3. Fixed lump-sllm fee. This method is frequently used for investigations and
studics or desig n projects with a well-defined scope and complexity. The fixed fee 1=
~~~I ~ c
o
.~

'<t , - a- '.: g
is computed by estimating the work-hours required to execute the project and the
;] ::I! ~ ~ ~ l c
anticipated costs for renderin g the service. For water treatme nt plant design proj- J - ..... 0606'"'" II 8.
E
~
ec ts, any experienced professional engineer can esti lTlate the cost for services by an-
• 8
alyzing the demands of the project and the total number of required drawings.
The fixed fee may also be calculated as a percentage of the estimated construc-
l=i
~
~ t~ HUUU
~
U§§§§§§
- . . g~g v
-j
tion cost. However, this particular method is not ge nerall y recommended because
the figures te nd to be inaccurate, especially with res pec t to plant expa nsion projects. ~ -
•

-
.;
Unless otherwise specilied . the lixed lump-sum fee con tract has the client pay the
linn on a monthly basis during executio n of the work. It is therefore important that
the contract include a time limit for th e professional services and a provision for
I
~ ~
.;
adju sting the fee in case the projec t is delayed by factors th at are beyond the co n-
----
I~
trol of the engineer.
4. Percentage of construction cost. This method is a variatio n of the fa xed lu mp-
sum fee . It was very popular when the eco nomy was stable and the available tech~ ~
= , = 2 • • , ~
•
;;o
nology was not changing rapid ly. Yet thi s type of contract is still used for some de-
sign projects. This method of assessing fees relies heavily on the engineering
reputation of the consultin g finn and the customary percentage by the industry; re-
fer to the compensation curves published by ASCE (Figure 1.3-1).
This method is not recommended for plant ex pansion projects because the de-
sign work is generall y more detailed and th erefore more cosIly than for new plants.

5
 PROJ ECT CONTROL 7
6 MANAGEMENT OF A DESIGN PROJECT

TABLE 1.4·1 (Continued)

TABLE 1.4-1 Contract Brief and Bil ling Summary
CONTRACT BRIEF BILLING SUMMARY

Prollt Center D.te Opened E. Job UDDer Lim"

Job Humber
FMsS _ _ Firm _ _ fsllr'NI..:I BWgatedCOSI I _ _ _ _ _ __ P,Ohl _ ".
A.
Contracl lnlormatlo n
(BucIQ'l Worll.shnts al'l'IOl ttquiritd on ConlllC'l1 whh Itn 11'1101'1 110.000 in tin)
For contracts alruay In 01,/1 d.alabau. provide contnct number, and proce.d to SKtion 8 .
C
Convill' (,fI'\4IIln6er 01 Seelion A IOf n.w c:onUJC!5 whet! art nOlitl our daUlbU. F. Involclna Informallon
Contract Oncripllon Bil on 1I.' NIlnvolc;t i1 job no . _ __ _ _ _ __ _ _ Show jobl StPillllllly on II'IYOicl
ConU act Amount S
p,or~ e.nt., PrGpOul Number
E""lOy" N....·T'(I4f WNrn to bill: monthly In Slagn upon completion
Conlrl.C1 Manag.r N,"" G. Contract TvNt (select one
Clilnt Addrus
Zip G· \. Saliry Cost 0·2. Call. Fi:I:.d F.. G·3. Hourly Rat G·4 . Lump Sum
S1318

, ,
COy OVlrl'lud/marlt·I,IP ,.. (lvllllu<l!mOlrlt·up % tNtll oup~. _ _ c;IIl(i.,..., it
Middl. tnliilL- LutH."...
CII.nl Com.d: flnt Naml lindlu~llIm
Phon. Number
TI~'
Division
BIIII~Group Period "rTf.S
Depart"",n!
P'flr\I e benl Number
0110. labor -----¥. "4
R.II/lld by Client Number FII~ lAl)or
-----"
Note. to proce.d r.quir,d? Slandatd f"1 & Cond~ ions .uc:l'led1 FCH'mHo.
Contract Financial Elemenllnformatlon
Ov.ruu l alXlr
R.lllad Labor
-----"
---_%
".%" %
B. Oullicl. S.lVicAs
-----" "4 %

• FOf . 1.menIS Ih;J(!y In 0\,11 da1abu • . plovid.,lnanci.I , I."",M !lumber, .and prooeed \0 SediOn C.
OII'I.r Cllarg.s (ODC) -----y.
- ~C~': ':~ '~'~S<~'":;'~"7'~~%~====::
¥.
COrr9 I,I, I,m,indtr 01 Section e 101 MW . lfmel'llS which 'f' not In our d,ll1bas. _ _ PIOV. Ov.rl'lnd lal.
El,menlOUetiplion s",llty COlI _ _ " FludFM'--- _ Hillorcal 5\0.· coM
EflVIoy" Numot·
PfOje!;t MOInlQlf HI"" 0\1'1,1 (anlch nl, $th.)· c;ode
Conlrat1 Typa Ellmanl Amount: 1
PfOUl Clnlal
Clial'll Au1horiulion: H. Other Information
~\I'ce 01 Funds'
\II "ccoroanea with:
_ _ Our po'oposal counllnign,d by Clilnl dat.o
First Sou".. R'"ntion . ~:-::==== ~.
Inlllni Rat.
Mil,_,.
% An.r No. 01 Oayl _ _ _ _ _ __
R.,.·. _ _ JMM Sl~nd~rd.)/ ___ ,

SKondSou""
Our aQ,. . manl daltcl Bldl.up (only i/ f1quittd by clllnt) lloo, ODC
Third Source
lenar 01 authoriutlon dOilad
Clianl's "contract No. oatad I. Soecla l lnstructlons
Ctianl's P.O. No. d:altd

C. Job Information
Co"1lu1ar Job [)ascription
Job Oesc~lion J. Billing Contact Information
EfTlIloyn Numoar
Proje!;t Eoginn' Namt (complete only If different Irom address on reverse side)
BUllnass Elamanl Plojacl Control EI,""nl
$lila l~ SlrM! Addflu
Job Sita localion: C~y
_OA/OCJob
Raqunlld CCr.! mambe":
P,rcane ollan _ _ %
N_~
CCM sdlldu!ad: oall
Emp. No.' N"'"
location
._-
EI'I'9. No
ffr'Q .No.: _ _ _
"'"
County
$'111'
COlmlry
l ip

N,n. EfTlJ. No.: --- N.~

(atlad\ additional shltt it nassary)
Middlt Inlial
PhOh, NutrCer ( I
Lasl Na""

Sl./beon!raClo": K. Invoice Description

To _ _ _ _ __
(ll'IacflOiOditional ,1111'. nnury) ProflnlonaJ S.IVQS R'l'IdefiO dul1ngll'll PI/loci of: SIan 01[1

D. Aoorovals
Conlfat1 ManOlQII 0.11
PlojKt MOinaQI'
Dap!. Mgr. or BOM "'"
0 •• Group Managar
V.nlitd by Corp. CounStI
0101'
0.11
CA"2(3189)
V.fif..o by BiUing Oal.

C A.l2(31B9)
8 MANAGEMENT OF A DESIGN PROJECT PROJECT CONTROL 9

The exi sting plant must be evaluated ex tensively, and all new parts must be com- TAIlLE 1.4-2 Project Budget Worksheet
patible with existin g parts. This detailed analys is does not allow freedom in design
and the engineer must spend more time on the project. JobNumIM,III II.ltlll
It is very important to note that final agreement on the fcc (by both parties) should
Upper limit S
always be negoti ated after a detailed disc ussion of both the scope of the work and SC+ • Sal ary COSI PlIJ' '!Ii.
• Percent 01 ConstrIJction, ASCE CIJrve _ __
th e elements of the engineerin g costs. OIl'Ier

1.4 PROJECT CONTROL

Profess ional engineers manage a broad range of acti vities, including technical in-
vestigation and analysis, environmental and prelimi nary studies, planning, design,
constructi on, compilation of the operati on and maintenance manual, and plant
startup, as well as providing advice and general co nsultation . Regardless of the type
of service, each project must be completed on schedule, within budget, and with a
minimal number of errors and omissions-ideally, error free. A ri gorous control
measure is therefore vital for any project.
To execu te effective control over a project, a solid system of control must be es-
tablished. However, if the firm already has a standard proced ure, all projects should
follow this method . An exampl e of a design project control system is as follows:

I . Upon receipt of a signed contract, assign a job number to the project, in-
cluding the year if possible. This numbcr should be used on all work, in-
cluding calculati ons, drawings, and correspondence. All costs and charges
pertainin g to th e project should be identified by thi s number.
2. Prepare a contract brief, billing summary. and a budget worksheet for the
project. Refer to Tables 1.4- 1 and 1.4-2 for examples of the form s.
3. Create project files.
4. Pre pare a co ntrol sc hedule for the project, including the period of ac ti vity,
the budget for each discipline of the project team, meetin g dates for coordi-
nating each section of the project, a final check date, a date for obtaining the
final au thorized signature, and a date for printing and binding the proposal.
Although several methods are used in scheduling the workforce requirements
during the design phase. the simplest and most cOlllmon arc the bar chart, a
graphic representation of each discipline of the job plotted against time, the
cri ti cal path method (CPM), the program evaluati on and review technique
(PERT), and the integrated budget and sc hed ul e monit ori ng tec hnique
(IBSM ; Figure 1.4-1 ). 9. Determine the number of tec hnical advi sory meetings that wi ll be held dur-
5. Investigate all req uiremen ts established by local, state, and federal agencies. Ing the project. The first meeting should take place after approximately 5 to
6. Organize a project team . 10% of the projec t is completed .
7. Arrange for all necessary outside services, such as soil analysis and site sur- 10. Prepare a memorandum after each meeting.
vey. II . Encourage active input from the cl ient. Keep the client informed on the
8. Select members for the technical advi sory comm ittee and the value engi- progress of th e project by holding periodic meetings and exc ha nge ideas and
neerin g team. expenences through work shops.
 ..

10 MANAGEME NT OF A DESIGN PROJECT

VALUE ENGINEERING 11

lOOr--------------------------------------, 100 Finall y, it is esse nti al to sc hedul e the total workl oad of a firm as we ll as those
170 for indi vidual proj ecls. Wilh oul proper scheduling. a linn will operate inefli cie nll y
100 W and will have proble ms in effeclively controlling ongo ing projecls.
150
00
~ 14O

~130 70 1.5 VALUE ENGINEERING

)( 120
ellO
Value analys is was devel oped by staff members at General Elec tric in the late 1940s
~, 100
to improve product efficie ncy. Thi s process inte nti onall y seeks substitut e materi als
] w
and melhods 10 replace more costly component s. Until the early 1970s. Ihe eng i-
~ 80
nee rin g and co nstructi on industries had limited interest in the process o f va lue a naly-
'70
.~ sis or value engineering. Howeve r, value engineering (VE) beca me an import ant part
;; 00 Legend
, 50 30
o Projected ellpenditures of design and construclio n in March 1972 when the Public Buildin g Service of the
~ 40 ]s b. Project progress 20
Ge neral Service Adminislrati on (PBS/GSA) made VE a requi rement for mos t new
30 o Actual ellpenditur es architec tural, e ngineering, a nd const ructi on management contrac ts. Both Ihe Society
20 A Schedule status (2.3 months behind)
10 of Ameri ca n Value Engineers (S AV E) a nd Ihe de fe nse induslry have made greal con-
B Budget status ($1 2,000 over budget)
10 tri buti ons to the developme nt of va lue engineerin g.
O~I~J~F~M~A~~M~J~J~A~.S~O~N~D'I~J~F~M. .A~M~rJ.,J. .A-,S~1 o O ne early study ide nl ifi ed seven reasons/opportuniti es for Ihe appl icati on o f VE:
I 1983 . 1984

Figure 1.4-1 Integraled budget and sc hedule mon itoring (lBS M).
12. Review the coS! of the projeci al the end of eac h mon lh .
13. Prepare co nstru cti on specifications based on the sta ndard speci fi cati ons and
edit the m to make the j ob more specific.
14. Have the co mpl eted drawings and spec ifica ti ons chec ked by the projec t e n-
gineer and an indepe nde nt chec ker. Eac h sheet should be red and ye llow,
li ned.
15. Ed it the bid doc um ents and submit the prelim inary drawi ngs and specilica-
ti ons to the clie nt for review.
16. Arra nge a n estimate for the constructi on cos ts.
17. Schedule the producti on o f construc ti on docum ents.
18. Obtai n signatures from the co mpany ofiicer and the clie nt on the linal trac-
ings.
19. Presen t the fi nal drawi ngs and speci licatiolls to the clie nt and appropri ate
governmental age ncy (local, state, or federal) as required.
20. Arrange for advert ising of bids and bid openi ngs .
I. Ad va nces in lec hnology
2. Evaluation of speci fi cati ons for worth (va lue)
3. Introdu cti on of altern ati ve design co ncepts, skills, a nd inform ation
4. Recog niti on of a change in Ihe needs of Ihe c lient
5. Recog nition of a design with add ed value
6. Feedbac k from pilot a nd fi e ld tes t studi es, in additi on to ac tu al operati ons
7. Recog ni tion o f current design de fi ciencies
The va lue e ng ineering review tea m shou ld consis t of a min imum of one or two
licensed VE engi neers and several ex perts in a ll tec hni cal di scipli nes being a nalyzed.
It is recomme nded that the inclusion of pro fessors and scien tiSIs be limited only to
those who possess both theoreti cal and ac tual experience in design and construc-
tion. The VE review process is time consum ing; the du rati on of a small projec t is
at least I week. TIle foc us o f the process is to evaluate the func tional capability o f
Ihe fac ililY and second ary considerations such as aes theti cs and specilic preferences
re laling 10 design and operalions may be excl uded from Ihe VE review. A we ll -run
VE review generally res ults in recomm end ati ons prod ucing a cost savings of 15 to
30% per project.

It is imperative that the calc ul ati ons, draw ings, a nd specificati ons be chec ked
pri or to fin ali zing the bid doc um e nts. Failure to do so may result in erro rs that could
pote nti all y have d isastrous effects on plan t constru cti on a nd operatio n or eve n re-
sult in a lawsuit. Since many speci fi cati ons are similar among various projects, in
both outlin e and tec hni cal provisions, it is strong ly adv ised th at lirms sta ndardi ze
their specilications as a mea ns o f projec t co nt rol.
1.5.1 Scope and Methodology of Value Engineering
Va lue engineering is both a systemat ic a nd creati ve method of identify ing unneces-
sa ry cos t within a projec t. The goa l is to ac hieve sav ings in construc ti on cos ts, as
we ll as operation a nd ma intenance costs, without sacri lici ng the effec ti vc ness a nd
reliability of the proj ec t. The parame ters are tig ht: Maintain necessary funct ional
 p

12 MANAGEMENT OF A DESIGN PROJECT

PRIVATIZATION DELIVERY PROCESS 13

PRE STUDY PI lASE

tionale for the design, project co nstraints, design co nsidera tion s, raw water qua lit y
pnOJEC I rnESIUOY COST TO and treatme nt goa ls, treatability studi es, fac tors a ffec tin g decision makin g, and the
COOROINAl'lON PREPARATION VAI.IDATIOH WOIIK SIIOIJ
PI1AS[ rati onale for select ion of the processes, hydrau lic prolile, and layout of the treat-
ment plant.
After oral present atio n, the VE tea m members shou ld be more know ledgeable

I ORIENT ATiOH INfOlV1ATION

PIIASE
I CREATI VE
PHASE
JUOGI1EIIT
PHASE
11 It"""" I
Q(vClOf'M( NT
PIIASE PHASE
A TlO>!
abo ut the desig n and begin to criti cally analyze th e va ri o us cos ts and effec ti veness
of the proposed treatment process train, each process unit, equipme nt , and control
sys tems. After eval uatin g the treatment fac ilities, eac h tea m member prepares a fu nc-
tiona l analysis. It is importa nt for the tea m to review and assess the preli minary cost
I.IST ID EAS BRAIN
GUI[IlATEO
II STORMING
VALUE ENGINEERING STUDIES
es tim ates of the proj ect, as thi s figure serves as a benc hm ark for the vario us alter-
nati ves ge nerated during the revi ew works hop.

II USE T ASK FLOW DIAGRAM

I CtIECKl lS T
r OR IDEAS
1.5.3 Creative Phase

This seco nd step o f the VE review is a brainstormin g sess ion which lists areas that
-
POST V[ STUDY PROCEDURE
could poten ti all y be improved whil e incurring less cost. The purpose of thi s phase
VE STUDY FINAL is to ge nerate a number o f new ideas (alternatives); those worthy of furth er evalua-
REPORT CONSUL 1 AT ION ACCEPTANCE
tion arc identified in a later phase.

Figure 1.5·1 Typical task now diagra m of va lue enginee ring studies. 1.5.4 Judgment Phase

capacit y witho ut waste. Not all VE reviews res ult in cost reductions; however, ad-
ditional "value" may still be derived if the recommendations increase the capabili-
ti es of th e projec t. It is importa nt to note that VE is not a detai led rev iew of all proj-
ect fea tures. Rat her, VE evaluates selected high-cost items; areas in whi ch either
cos ts may be reduced w ithout sac rilicillg functiona l capabi lities or areas in whic h
the capabilities may be e nha nced w ithou t an in crease in cost.
The VE team only provides recommendations. The design eng ineer and the client
must co ndu ct a th orough analysis of the app li cabi lity, feas ibilit y, and viab ilit y of
eac h recolllmendat io n before deciding to adopt or rejec t them. The VE st udy is com-
posed o f two segments : workshop and prese ntation. The work plan is set in the lirst
seg ment (worksho p) and provides an organized structure for identifying potential
review areas for alterna ti ve desig n and deve lopment. The work plan shou ld contain
the fo ll ow in g fi ve phases: ( I) informati o n, (2) c reative, (3) judgme nt , (4) develop-
ment, and (5) recomme ndati on. Fig ure 1.5. 1 is a task fl ow diag ram of a VE study.
1.5.2 Information Phase
The projec t e nginee r must prepare handouts contai nin g all esse ntial project infor-
mation. These shou ld be distributed to each member of the VE review team I week
prior to the session. The review session beg ins with an oral overv iew prese nted by
the projec t e nginee r. It is esse nti al that the entire VE review team understand both
the bac kgrou nd and deci sions that influence the project desig n: site conditi o ns, ra-
The variolls design alternatives are now eva lu ated and refined. Each is ass ig ned an
associa ted cost, th en ranked in order of their potenti al cos ts savi ngs. Secondary co n-
siderat ions stic h as aesthetics, eflic ie ncy, reliabi lit y, and durability may be eva lu ated
and app lied to the ranked altern ati ves.
1.5.5 Development Phase
Design altern ati ves recei vi ng high ranking are further deve loped durin g th is phase.
l11e team should stud y the most promising a lternatives in detail and co nduc t a cos t
comparison in relat ion to th e original design . A life-cyc le cos t ana lysis sho uld be
co ndu cted in cases where the capi ta l cos t differe nti a l is not sig nificant , but th e op-
erat io n and maintenance cost differential is no tab le.
1.5.6 Recommendation Phase
The recomm endation phase is the last step of the worksho p seg ment. Afte r the VE
revi ew team has carefu ll y restudied and rereviewed all feasib le altern atives, a writ-
ten recommenda ti on is composed . An oral presentati on of th ese recomm end ati ons
is then de li vered to the projec t e ngineer and the ow ner/rep resen tatives. There sho uld
not be any debates or arg ulll ents during thi s presentation. As mentioned previollsly,
these reco mmendations are no t final deci sions, o nl y sugges tion s. It is th erefo re es-
se nti al that the VE tea m present its linding and recommendations in a n inte lli gent ,
cOll vin c in g, a nd cred ible manner. This can be ac hi eved only if the VE tea m has a
thorough understand in g of th e projec t goa ls, the function of all unit processes, in-
14 MANAGEMENT OF A DESIGN PROJECT
clud in g the pe rformance of all equipme nt and co ntro l systems, a nd the projec t con-
stra ints. Delivery of an e laborate VE report !lUlSt foll ow the oral presentation.
Preparation of thi s fin al report may last I to 2 weeks, depending o n the magnitude
of the report.
1.6 PRIVATIZATION DELIVERY PROCESS
Alth oug h th e private sec tor utilized items suc h as des ig n· build-operated (DBO).
design -build-maintain (DBM). public-private partners hips (PPPs), and lon g-term
co ntrac t o pe rati o ns, these were not practiced in the domestic water industry until
the ea rl y I990s. Some 35% of U .S. utiliti es now co ns ider using th ese approaches
at th e end o f the twentieth century. These project de li ve ry processes have di stinct
advan tages over the traditional processes of desig n, bid, build- hav ing RFQ, FRP,
eva luat io ns, negoti ations , biddin g, building, as we ll as protests and claims. TIle tra-
ditional process is both time co nsuming and costl y.
The privatized processes have three basic advantages over the traditi o nal
processes: lower cos t, lower ri s k, and an acce lerated sc hedule . The ex pected sav-
ings in these three categories are approximately \5- 30%. Yet, the disadvantages are
severe penalties if the project does no t meet co ntracted req uiremen ts, co ncern s re-
gardi ng quality control and the potential reduction in plant operatio nal flexibilit y
due to the low cost of the projec t.
The recent e mergence of the privati zed deli ve ry process in the domestic water
industry is the result of internal and external competiti o n. Factors such as regula-
tory req uire ments governing the upgrade o f existing plants or th e building of new
plants, the consequences of deferred maintenance, public resistance to rate inc reases,
and co mmunities embroiled in financial cris is, force many public utiliti es to con-
sider the o ptio ns prov ided by the privatization processes so that th ey ca n meet their
goa ls. This tre nd toward privati zati o n is li kely to become more popular o n a g lobal
sca le.
BIBLIOGRAPHY
Abdall ah, E. T.. Preparing Specifications for Design- /Jid-Build Projects, ASCE Manual.
January 2000.
"DBO and DBM Gain Popularity," Roundtable, J.AIVIVA. 9 1:4: 18(Ap ril t999).
Dysad, J. A., "Trends in Privatization," J.AIVIVA . 9 t : It:44 (November t999).
General Servi ces Administration. Vallie EIIgilleer;IIg, Public Building Service/GS A Handbook
P8OO0.0, U.S. Govern ment Printing Office, Washingto n. DC. 1972.
Haarmcycr, D .• " Privati zati on Inrrastructurc: O pti on for Municipal Wat er Syste ms," J.AWWA,
86:3:42( March 1994).
O'Brien. J. J.. Vallie AI/alysis ill Design al/(I COIISII'IIClioll. McGraw-Hill. New York. 1976.
"Publi c-Private Partnerships," J.AlVlVA . 88:4:26 (Aprit t 996).
CHAPTER 2
Preliminary Studies
2.1 MASTER PLANNING
The first s tep in the design of a regional water s upply system is mas ter pl annin g.
The fo ll ow in g areas mu st be covered: s tudy of the water treatm en t facilities. eval-
uati o n o f the system convey ing raw water, evaluati on of the fini shed water dis tri-
bution sys te m, a nd eva luatio n o f a few basic infras tructures of the reg ion (sewer
system and sewage treatment facilities).
T he number, size, and location of the local water distribution reservo irs depend
on the an ticipated pressure zo ne of th e service area. S in ce a large portion of potab le
water is converted into sewage an d various o ther wastewater, the conveyance and
treatment of wastewater rnu st also be evaluated. Ot her infrastructures that sho uld be
co nsidered are new access roads, power suppl y, and utility systems near the plant
si te. It is important to remember that a water treatment plant is a "factory" that han.
dies a large amount of chemicals, inc luding hazard o us chemicals s uch as ch lorine
and caustic soda, in addition to producing wastes. Thus the project engineer mu st
cons ider whether it is necessa ry to reclassify the ex isting zoning or modify existing
codes. The master deve lopment pl an s ho uld also s tudy the influence of a nd co n-
nec ti on to ex ist in g waterworks and sewage ha ndlin g facilities. In some cases, issues
suc h as water ri ght s and s ludge di sposa l may become difficult to solve when they
a ffec t ne ig hboring communities.
During the early s tages o f pla nning, the project engineer mu st evaluate potential
sources o f water and assess the qu alit y and safe yield o f each. Water may be ob-
ta ined from a river, lake, reservoir, or groundwater. In general , groundwater is pre-
ferred over surface water due to its superior and steady water qu ality. Althoug h these
charac teri sti cs make ground water eas ier and less costl y to treat, and result s in a
smaller amount of waste, the ava il able safe yield from gro undwater wells is ge ner-
all y limited. No nethe less, the master plan should evaluate gro undwater sources and
iss ues pertaining to groundwater contamin atio n, ground s ubs idence, and seawater
intrus ion due to excessive withd rawa l o f groundwater. Groundwater contamination
may result from human, animal, or industrial waste, agricultural activity, or geo-
logic co ndition s. Recently, the analysis o f rado n levels in groundwater has become
an important issue , due to its class ifi cat io n as a ca rcin ogen.
The mas ter pl an mu st also include the following items for s tudy : proposed wa-
tershed management , a brief eva lu ati on of site condi ti o ns, including topography, ge-
15
16 PRELIMINARY STUDIES
ologic conditions, history of flooding and earthquakes, site accessibility, land use
of the site, and preliminary archeological and environmental assessments. One of
the most difficult items in the master plan is the prediction of the region 's future
water demand, Although there are basic mathematical equations, variables unique
to the region (changes in socioeconomic conditions, sign ificant change in weather
pattern) and global variables may greatly fluctuate over a period of IOta IS years.
It is unfortunate that these variables are not easily predicted, as demonstrated by
many recent case histories throughout the world. Consequently, prediction of water
demand 50 years into the future is inaccurate and has little value,
In rece nt years, other unique concerns have found their way into the master plan,
The first is the anticipated enforcement of the more stringent Drinking Water Quality
Standards beyond the year 2000. A second is emerging water treatment technolo-
gies suc h as the membrane filtration process. In the past, master development plans
demonstrated the accommodation of future increases ill plant capacity merely as ad-
ditions to the original conventional treatm ent process module, This concept is no
longer valid, due to the emerge nce of new water treatm en t technolog ies and the need
for more sophisticated processes that may require a different (larger or smaller) over-
all footprint. For this reaso n, estimating the required acreage for the proposed wa-
ter treatment plant si te and the selection of future water treatm ent processes has be-
come more difficult. Yet it is safe to allocate an extra area of the plant site exclusively
for the addition of future water treatme nt processes on the master plan. The impor-
tance of the master development plan must not be overlooked. The master plan pro-
vides an early rough estimate of the overall capital expenditure, duration of the proj-
ect, and regional planning sc heme.
2.2 FEASIBILITY STUDY
The feasibility study is the startin g point for all water treatment design projects. It
is often conducted as part of the master development plan for the region or regions,
Since thi s subject is not the focal point of thi s book, it is discussed only briefly.
2.2.1 Planning Period
A spa n of 10 to 20 years is considered to be a reasonable period for planning a wa-
ter supply system, including the capacity of the water treatment plant. It is impor-
tant to note that some regions, such as southern California and the Bay area of north -
ern California, have already exceeded the water demand predicted in the early 1980s.
Yet some cities in the United States are well below the predicted levels, due to de-
clining industrial activities and severe weather co nditions. However, a span of 15
years is generally considered to be an acceptable planning period for immed iatc sys-
tem design and construction,
2.2.2 Water Supply Areas
When determining the areas to be served by the water supply system , e ngineers
should limit their evaluation to areas considered to receive water during the design
FEASIBILITY STUDY 17
p~riod. However, adj~cent areas and the potential amalgamation of neighboring dis-
tncts should also be IIlcluded in the overall design.
2.2.3 Future Population
T~ere are several methods to predict the growth or decline of the future population.
SlIlce most college textbooks cover thi s subject in great detail, it is not discussed
here.
2,2.4 Maximum Daily Water Demand
The lirst task in. the feasibility study is the estimation of the maximum daily water
dem~nd per capIta. For a new system, this estimation is generally derived from data
acqUIred from other communities sharing similar backgrounds, characteristics, and
developmental trends. In the case of plant expansion, the best estimation is provided
by ~he ~rends of ~he existing ~ater supply system. The maximum daily demand per
caplla, IS the baSIS f~r evaluatmg the water treatment capacity. Once this figure is
established, the max Imum daily demand may be calculated as the product of the
maximum daily per capita demand and the estimated popUlation to be served within
the planning peri.od. An average annual rate of use in gallons pcr capita per day
(gpcd) of 180 (wllh a range of ISO to 210) or 680 liters per capita per day (with a
range of 570 to 795) is common in the United States, and the maximum daily de-
mand is usually 150% of the average annual rate,
2.2.5 Evaluation and Selection of the Water Source
The source of the raw water may be a river, Jake, artificial reservoir, groundwater,
and in some cases, reclaimed sewage or seawater. The evaluation and selection of
the proper water source should be based on the following issues: (I) quantity of wa.
ter requ"e~, (2) quality of the raw water, (3) climatic conditions (i.e., icing), (4) po-
tential dtfflcultles III constructing the intake, (5) operator safety. (6) providing min-
Imal operatiOl~s a.nd maintenance costs for the treatment plant, (7) possibility of
future contamlllatJOn of the water so urce, and (8) ease of enlarging the intake if re-
quired at a future date.
Underground water, including water collected by an infiltration gallery, should
be conSIdered seno~sly as a water Source because it can potentially save a signifi -
cant amount of capllal as well as operations and maintenance costs for small com-
munities. The quality of underground water is generally su perior to that of surface
water, thus resulting in minimal waste (sludge) and no problems associated with ic-
ing during the winter months-an added benetit for cold-weather regions.
2.2.6 Size of the Water Treatment Plant
The capacity of a water treatment plant is determined by the maximulll daily de-
mand placed on the system . However, future water demand of the area being served
should also be taken into account when calculating the size of the plant and the site
 •

18 PRELIMINARY STUDIES
FEASIBILITY STUDY 19
area. Additional iss ues that an'ect plant s ize are the reliabi lity of the water supply
and the cost-effec ti ve ness of suppl yi ng wa ter from one large plant versus two or ccg'. has ,e~l1~rged. ~n the pu~lic w~ter-utiljty industry the term turnkey constructioll
eve n three medium -s ized plants at different loca ti o ns and different elevations. As a ~ ~:.s to hMv.rng a Sing le c nllt.y dcs lg n, co nstruc t, and de li ver a properl y fun cti o ning
rul e o f thumb, the s ite area required for a conven tional pl an t may be estimated by
.a ~ I . I IY t, at. IS r~ady for .ser~I CC. This type o f contract has the advantages of ma in.
th e formu la
t,unln~ cngllleellog contl/llllty 3I~d ~ualit y control. Privatization of the water treat -
ment IIldust ry has allowed org ani zation s from the public sector to e nter' t .
ket that has tradition ally been dominated by pub!"' f r p.' III ~ d l11ar·
r Ie Un( IIlg. IIvate co mpani es Illay
now mance, own , construct, and o perate the fac ilities Si nce these facilit 'les . '
as a business cnt . ' I ' d . " ' . ale run
where A is area in acres (1 acre = 00405 ha) and Q is the ultimate plant capacity in , , '.' , elpns~, t le II'. eS lg n, co nstruc ti on, and operation are genera II morc
cos t-clllclc nt than their publJc co unt erpart s How "ve' I . I . Y
milli ons o f ga lions per day [mgd; 1 mgd = 3.784 c ubi c meters per day (m'/d)]. ))(0 f' . h . ' .. C I, oca govelillnents have the
; -' I ~ n ~ PUle, as.lng these tYI~es o f f~1e iliti es at fair market value at the end of the
ease telm s. Despite the ta x re form bill I,assed by the U S H f' R . .
. Db ' " ollse 0 epleSe ntatlves
2.2.7 Treatment Plant Site 1;1 . ecem er 1986--which removed a large pan o f the ta x exemption b~llefits for
t1C
Illveshtor and .owners- th e ta x-cxemp t int erest of thi s in come, for the investor or
Eva luati on of the treatment plant si te is based primarily on the di stance from the in- Iessor, s ould stI li attract private invcs tors.
take, the layout of the treatment process unit s, the e nvironm ental impac t of the treat-
men t pl ant , and the method of water di stribution (grav it y versus pumping). Moreover,
the following items must be included in evaluation o f the treatment plant site: (I) geo-
BIBLIOGRAPHY
graphi ca l locati o n, (2) information obtained from the geo logical study, (3) avail-
ability o f e lectric power and utilities, (4) access ibility to major hi g hways, (5) hi s-
Alia. P. M. J., and Manzi I) "S)'dney Wal ' P bl ' I"
tory of Ilooding or the presence of ea rthquake fault s, (6) constru ction costs, (7) s ite (Apri l 1996). . ... < cr s 1I IC- n vate Partncrship," 1. AWWA. 88:4: 108

maintenance costs, (8) operator safe ty and the safety o f ne ig hborin g homes, (9) pro-
AW WA Joint Disc ussion "Pros d C . 'I'
vis io ns for future plant expansion, and ( 10) e nviro nmenta l impact study, includin g ,an OilS 01 urn -Key Construction " J AWWA 65' 12'766
( Dcce mber 1973). .. ".
an archeo logical investi ga tion.
AWWA Round Tabl e Discllssion s, "Gai ning Access to Capital Markets" J AWWA 87'2'21
(Febnrnry 1995). .. . . ..
2.2.8 Financing AWWA Round Table Discuss ions. "Privnti zation," 1. AWWA, 79:11:12 (November 1987)
Ayres. I). W. and Thorpe S "F . C . .
The preliminary site evaluation mus t prese nt variolls means by which to fund the 199 1). , . " Inanclllg aplla llmprove ment s"..J AWWA . 83'8'\0
" " (Augusl
project. Several methods are used to obtai n necessary fundin g fo r co nstructing wa-
ter treatm ent plants: reven ue bond s, general obli gati o n bonds, spec ial assessment BCn8\~18':4Ro' (B·, "Etl cct lve Reve nue Bond Markellll g. Put Your Bes t Foot Forward," J AIVWA
.. . A ugusl t 991).
bo nds, state and federal aid fund s, o peratin g reve nues, con tributi o ns by customers,
and private funds. The most common method is the issuan ce of revenue bonds, s ince
~;:ker. A.. ::Briti~h. Pri vat~z.ation:
Balancing Needs," 1. AWWA, 86:3:56 (March 1994).
water treatme nt plants are revenue -producing facilities. :y. F. A" Munrclpal Ull ir ty Bond Financing." J. AIVIVA. 62:8:468 (A uguSi t970)
Bnce, R. L.. and Unangst E R "L R F' " .
The lirs t s tep in obtainin g financ ing requires the eng ineer to provide an estimate 1. AWIVA. 8 1:5:48 (May' 1989)." ong- ange lIlallcml Planl1l1lg for Water Utilities,"
o f the plant construction costs. A team of co nsulta nts (e ngi neering. financial , and CaSi:1SllS C "P ' t·· · h M .
legal) Illust then study the proposal. It is imperative that the team include a finan- , ., nva IZlllg t e eX lca n Wntcr Industry." 1. AWWA 86' 3'69 (M h 1994
Doctor R D "P ' S . ' '. . arc ).
c ial s pecialist all d lega l expert so that the utility can dec ide if it ca n issue revenue , . " n val e ec tor FllulI1cing for Water Systems " J AIVWA 78'2"47 (F b
1986). ,. .. . e !lIary
bonds or if the community s ho uld be asked to iss ue gc neral obli gat ions bonds. A
Earl. T E and Greenste in S A "E 'C'
bond attorney is requircd to draw up the proceedings and to issue the approving "
i
62' 1'7"05 (N ' . .. "conOllll C ollsldel"<l ti ons in Facility Planning" J AIVWA
ovembe r 1970). , . ,
o pinion , without which the bonds ca nnot be marketed. Finally, the preliminary e n-
Gull et. B. M" and Bean D 0 "TI CI 'I M .
g ineerin g. finan c ial , and legal studies mu s t be submitted to the city cOlillcil and (A pril 1997). '.... le lalOU e odcll or Competiti ons," 1. AWWA. 89:4: 100
should di sc lose whether the projec t can be acco mpli shed at a reaso nabl e cost through Haarmey·c r D "p.' ', ' ,
acceptable linan cin g met hod s within the ex istin g statut ory limitatio ns. Th ese stud- '1'4 2 '(M" Ilvatlzing Inlrastructure: Opti ons for Municipal Water System s." 1. AWWA
86 " . arch 1994). .
ies ofte n save the gove rnin g board o f th e utility from ulln eccssary embarrassme nt
JelTrey. J., "Privatization in England and Wales," 1. AWWA. 86:3:64 (March 1994)
and keeps the project from becoming a political football .
In rece nt yca rs, privatizatiofl, whi ch in volves private fin anci ng and turnkey proj- Pon8'ge ~4'. R .
and ~Vehmeyer, M" "Competit ion: Strategies ror a Public Utilily " J'
.. 11 9 (Api'll 1997). . . AIVIVA.

20 PRELlMtNARY STUDtES
son J H and Willi ams P. C .. "Financing Strategies for Sma\1 Systems,"
Sagravcs. B. R., Peter ' . ., ' The Ilocculant optimi zati on test should be designed to determine the most ef-
1. AIVIVA, 80:8: 40 (August 1988). ..
"Privati zati on: An Emerging Manage ment and Financi ng Tre nd. Waler Ell g.
. ea cr, D..
S d
Mal/age., p. 44 (March 1984). .
Shankcr. A.. and Rod man. L.. "Pu blic- Private Partncrships." 1. AIVWA, 88:4: 102 (April 1996).
2.3 BENCH-SCALE AND PILOT-PLANT STUDIES
The prcdesign efforts must include bench-scale ~nd pilot-plant .studi~s whenever
possible. These studies allow the engineer to obtain va luable baSIC deSign, data anc~
either to optimi ze or confirm the proposed raw water and tilter waste-~vash .Ir~at
ment processes. The bench-scale stud y may ,be cond~c.tcd in a short period at tIlnc
and with a limited budget. However, it requi res a 1llllliITIUm of 6 months to 1 y~ar
for a pilol slUdy 10 yield meanin gful and reliable dala. Consequenll y, Ihese studt es
arc quite expensive.
2.3.1 Bench-Scale Studies
A proper bench-scale study addresses the following objec~ives:. oPlim iz~tion o f
chemical coagulants, the chemical application sequence, conll,rm atlOn of the p roper
mi xi ng conditi ons for fl occul ation, estim ati on of the hydraulic s~lrracc lo~d.mg for
the sedimentation process through the measurement of Iloc sellilng velOCIti es. the
total level of trihalomcthanc potenti ally produced by the water, and the ,control of
tastc- and odor-producing compounds through the usc of ox i~lant~ or ,ac ti vated car-
bon. The Phipps- Bird jar tester is the instrumenl empl oyed pnmanl y lor Ihese stud-
.les all(I a mmlinUIll
.' 0
f 200 work-hou rs. is generall y requtred to complele the bench-
sc;\c studies, whi ch includes the time required to write the report. . sonal sludge produ ction rate. T he fo urlh objec ti ve of bench-scale S1 udi es-fl oc
The coagulant optimi zation study involves selecting the most effecll ve l~pe of
coagulant and the appropriate dosage th at produces go~d settl ed water q~ahty, as
well as optimal conditions for the filtrati on pro~css. Engl11eers ~hould reah~e that ~
conventional water treatment process is compnsed of many Unit processes. coagu
lation, flocculati on, sedimentation, and fiit ration, and th at these subsy~tems must ~e
interrelated and compatible if the sequence is to produce good-quahty. water. It IS
recommended that the effec tiveness of coagulants, such as alu~l1 , ~ernc salts, and
cati onic polymers, be evaluated both individually an~ as a con1bl1l~110n of alum and.
polymer. The recommended jar tesling procedure tS prese nted to Example I 01
Secti on 3.2.7 a nd in Appendi x 10. The suggested coagul ant dosages for the fi rst raw
water j ar test run are as follows:
• All/III or ferric chloride dosages (mgIL): 3, 6, 12, 20. 40, 60
• Ca tionic polYlll er alone (mglL): 0.25. 0.5, I, 2. 4, 8
• Alliin lVitli cationic polYlller (mgIL): 6 to 0.25, 12 10 0.25, 20 to 0 .25. 6 to 0.5.
12 10 0.5 , and 20 10 0.5
BENCH-SCALE AND PILOT-PLANT STUDIES 21
See Table A I0- 1 of Appe ndi x 10 for sample jar test results.
fecti ve type of polymer or additives (such as bentonite clay), optimum chemical
dosage, and proper applicati on sequ ence for both coagul ant and c hemical floccula-
ti on aids. The mos t effec tive Ilocculant aids are usually anionic polymers or non-
ionic polymers because they can greatl y increase both fl oc size and settling veloc-
ity. However, the timin g of Iheir application is very imporl ant; they should be fed
when pinpoint fl oc masses have formed, which is approx imately 5 min after alum
fl occul alio n begins. Table 3.2.3- 1 illustrales Ihis issue clearly. The seq uence of chem-
ical applicati on is a critical consideration for other water treatment chemicals, such
as lime and alum, chlorine and polymer, and chlorine and powdered activated car-
bon (PAC).
The mixing conditions for the fl occul ation process may be optimi zed by a jar
test, and data obtained in this manner are useful for the plant design. The jar test
allows the engineer to determine the optimum level of the mixing energy input
(G value), the mixing tim e, and the energy input pathway (tapered mi xi ng require-
ment). Although 1110st ex isting conventional treatment plants use an average G value
of about 35 with a mi xing time o f 30 min, a wide range of numbers should be leSied
at optimum coagul ant and fl occulant dosages to maxi mize the conditions. Tapered
mi xing- the use of high-intensity mixing at the beginning of fl occulati on, fo llowed
by a gradu al reducti on toward the end- helps promote Ihe form alion of large noc
masses and should therefore be studi ed as well.
Four major parameters th at significantl y innuence the optimum mixing condi-
tions are water temperature, raw water quality, type of coagulant , and coagul ant
dosage. Thus it is recommended that bench-scale lests be performed during the win-
ter, summer, and either spring or autumn to better comprehend the need to tailor co-
agulant dosages to the changing raw water charac teri sti cs and to estimate the sea-
settling veloc ities- must be evaluated at optimal levels of coagulant and mixing
co nd ilions. The tes t should be conducted du ring bOlh cold- and hOI -weather monlhs,
for reasons stated previously.
It is relati vel y simple to produce a settling velocity distribution curve from a j ar
lest. The turbidil y of Ihe seuled waler is measured for sampl es draw n from 10 cm
below the surface of a 2-L beaker at 1, 2, 4, 8, and 16 min after the j ar tes t has been
initiated. The samples collec ted at these times represent the respec ti ve settlin g ve-
lociti es of 10, 5, 2.5, 1.25, and 0.625 cm/min. A seuling veloci ty of 4 c mlmin is
equivalen t to a surface loading of I gpm/ ft2. The measurement of noc settlin g ve-
locities is di scussed in delail by Camp ( 1946) and Hudson ( 198 1) (see the
Bibliogra phy at Ihe end o f this section). Engineers must note that bench-scale test
results may not be relevant if the turbidity is prod uced arti fic ially or if the water is
not fresh, more th an I to 2 days old, since the test results do not accurately portray
ac tu al water treatment conditions. Refer to Sec ti on 7.4 for a discussion of the total
potential THM form ation and Sec ti on 7.3 for tas te and odor contro l by ox idants or
ac tivaled carbon.

22 PRELIMINARY STUDIES
2.3.2 Pilot-Plant Studies
In recent years, engineers have bee n forced to evaluate nonconvc nti onal trealm,ent
processes due to new EPA drinking watcr reg~lalions, public ~oncc rn ov~r lh~ ~al et,~
of drinki ng waler, the ever-increasing costs ot plant constru cti on" and Ihl! elllcI~cn<.:c
of new water trealment technologies. Consequen tl y, the use of pilot-plant stu(lIes to,
evalu ate proposed treatment processes is increasingly becom ing a necessary part 0 1
the design process. . . ._ ..
Alt hough pilot studi es can poten ti all y result 111 slgllltlcant cost s~v ln gs III plan t
constru ction and operations, the cost of the pi lot studies themselves IS, rath~~ large:
A good pilot study may cost anywhere from $100,000 to $1,000,000. lhus Ii mean-
ingfu l resu lts arc to be obtained from a pi lot stud y, the most ap~ropn ate type and
best manufactured equ ipme nt must be selected. Moreover, the pilot plant must bc
operated by hi ghly qualified personnel for a period of at least 6 to 12 month s.
Yet, a good pilot study ca n save millions of dollars In pla~1t construct ion cost a~ld
result in a significa nt reduction in plant operation and mall1tenan~e cos~ s b~ PIO-
viding a simple, but effective high-rate unit process with a small loot-pnnt lor the
plant site; as supported by a few c a s e s . . . , . filter performance for the lilter bed conliguration selected. The physical vari ables
Pilot-plant studies may be necessa ry to ac hi eve some or all 01 the lollowlllg ob-
jectives:
. or variances
1, Obtain penl11ts . from regulatory age ncies for nonconve nti onal
processes.
2. Evalu ate the practicability of a new treatmcnt process.
3. Compare the effec tiveness of alternati ve processes.
4. Obtain a gu ide for process design criteria and operating costs.
5. Imp rove existing processes.
6, Investi gate the cause of problems,
7. Co nfirm the effectiveness of the proposed treatment process.
8. Discover unforeseen problems resu ltin g from the implementatio n of the pro-
posed treatment process.
Although pilot-plant studies are a valuable g uide in establish ing bot.h l~ro~cs~ de-
sig n c riteria and operati onal parameters of a pla nt , they also have thclr 11I111tat lons.
The major problems are (I) difficulty in testing the raw water on a year-round ba-
sis; (2) use of an improper type of clay when simulating abnor mal I,'aw water CO Il -
ditio ns (hi gh turbidity); (3) usin g raw water that has bee n stored lor over I to 2
days; (4) differences in operati onal conditions, including the tec hni cal and opera-
ti onal knowledge of the plant operators-that is, pilot plant versus ac tu al treatme n.t
plant; (5) problems enco unt ered in scalin g up; (6) 1~1i lure:o foresee long- term c l-
fects of the new process; and (7) arriving at conclUSIOns bmsed by personal expec-
tations. 1'1 I I"
Pilo t-plant stud ies may be classified into two basic cutegor.ies: .study 0 lye rau IC
charac teristics and study of un it process performance. T he IlI'st Includes the stud y
BENCH-SCALE AND PILOT-PLANT STUDIES 23
of flow c harac ter istics in c1earwells or sedime ntati on tanks, and the seco nd includes
pilot studi es on corrosio n and the processes of filtrat io n, flocculati on, dissolved ai r
floatation (OAF), GAC adsorptio n, and ownati o n.
Once a pilot study is planned, the engineer must determine the variables to be
stud ied and what their characteristi cs should be. Since noccul at ion, sedimentati o n,
DAF, and preozonation are filtration pretreatment processes, it is very important to
select a combinati on of these unit processes that yields both max imum efliciency
and optimum economic benefits for the overall treatme nt system . From a practical
standpoin t, a pilot stud y usually ca nn ot exa mine all the relevant factors d ue to time
and budgetary constrai nt s. T hu s it is necessary to reduce the num ber of variables
by selecting onl y a few very important items. However, engi neers must first estab-
lish the following issues during the initial stages of the study: ( I ) the purpose of the
study, (2) the duration and cost of the experimen t, (3) the avai lability of eq uipm ent
and techni cal staff, and (4) the importan t variables of the study.
A basic pilot plant has a coagul ati o n, fl occ ulati o n, clarification, filtrati o n, a nd li-
nal disinfection process. The study shou ld first opti mi ze the dosage of the coag u-
lant and other required chemicals, that is, determine the dosage that yields the best
of the pretreatment process, such as mi xing intensity, dete nti o n time, and hyd raulic
load ing, should then be evaluated using the optimum coagulant dosage. The opt i-
mum dosage is based o n resu lts obtained from jar test studies. However, the use of
ord in ary pilot plants is virtually useless in predicting settl ed water quality or in op-
timi zin g ho rizontal fl ow sediment at ion tanks. ll1is subj ect is discussed later.
The piloting filter, including the GAC adsorption process, yie lds meaningful data
and is rel at ively easy to monitor. Filter performance may be cffec ti vely evalu ated
through the use of small-diameter columns [2 to 4 in. (5 to 10 e m)]. However, if
filter was hing is co nduc ted for study purposes, large columns should be used to min-
imi ze the "sidewall" e ffect. A rul e of thumb is to use a column diameter th at is 1000
times the average size of the gran ular filter medium. For exa mpl e, if the average
size of the filter sand is 0.8 mOl, the size of the lilter column should be 80 cm
(30 in.) in diameter. In the case of filtrati o n studi es, the size of the filter column
should be 100 times the effective size of the filter medium.
The pilot filter st udy sho uld address the following issues: opti mi zat ion of the lil-
ter bed and filter rate, eva lu at io n of the filter washing cond itions, and cont ro l of ini-
tial and terminal turbi d ity breakt hroug h. In so me cases, declining rate vers us con-
stan t rate filtrati o n may be included as a study item. To ach ieve these objec ti ves, a
minimum of three filter columns is abso lutely necessary; five to six columns are
preferred, with one of the unit s as a control column , Media opt imi zation involves
the evalu ation of different types of medium materials, determination of thc size-to-
depth relationship, and co mparison of the effec ti veness of monomediulll ve rsus Illul-
timedia. Basic information on size, depth, a nd mate ri al of filter beds is prese nted in
Table 3.2.7- 1.
One of the most popular filter beds in th e last half of the twentieth century is a
dual -media filter bed with a depth of 30 in. (0.75 m). The bed is composed of a
layer of anth raci te coal that is 20 in. (0.5 m) thi ck ( I mm effec tive size) and lO in .
 >

24 PRELIM INARY STUDIES BENCH-SCALE AND PILOT-PLANT STUDIES 25

(0.25 m) of filter sand (0.5 mm e ffecti ve size). Thi s bed s hould be uscd as a co n-
tro l for comparin g the performance of ot her types of bed,
Engi neers are adv ised to use a co nstant-rate filtrati on process for the pilot filter
stud y. Filtration rates of 3, 6, and 9 gpm/ft2 (7.5, IS, and 22.5 m/h) should be eval-
uated sin ce a filtratio n rate of 6 gpm/ft2 ( 15 m/h) is generall y considered to be an ac-
ceptab le filtration rate for hig h-rate tiltrat ion using dual -media beds. Figure 2.3.2- t
is a sc hemati c di agra m of the recommended sta ndard pilot-plant co nfi g urati o n.
The quality of the filter e muent obtained fro m pilot filters is o ften sig nifi ca ntl y
below that produced by actua l o perational filte rs. Under identical conditio ns, the
turbidity of the pilot filter e muent may be tw ice as hi gh as the ac tu al Ii Iter: 0.2
nepherometri c turbidit y unit (nt u) for pilot-plant filters versus 0.1 ntu for actual fil-
te rs. Eng ineers should also be aware th at the initial turbidity breakthrough, at the
beg innin g of the filt rati o n cycle, is usuall y quite severe for pilot plants. This may
be attributed to the degree o f ripening or maturati on of the filter bed . Unlike ope r-
ational treatment plants, the filters of the pilot plant are alway s thorou ghl y washed.
Co nseq uently, the filt er beds are seldom all owed to beco me well ripened. The rea-
sons for the frequent washin gs are ( I ) minimi zin g th e condit io ns present in previ-
ous pilot filter tests, and (2) lack of operati o nal kn owledge by the pilot filter oper-
ators, who are usuall y e ngi neers or sc ientists. Scientists and eng ineers ge nerally
was h the filter beds until the waste wash water becomes clear, in co ntrast to the
practice of actua l plant operators, who all ow a turbidity of 5 to 10 ntu to remain in
the lilter wash was te.
As me nti o ned prev io usly, pilo t tests for hori zo nt al sedime nt ati on tanks have lim-
itations. If the pu rpose o f the pilo ting is to study th e hyd rauli c c haracteristi cs of the
tank , a hydrauli c-scale model th at appli es Frotlde 's similarity law should be adop ted.
CONSTANT HEAD FLOCCULATION ' SEDIMENTATION
TANK TANK TANK
However, if the o bject of the study is to evalu ate th e ab ilit y of the tank to re move
turbidit y or o ther co ns tituent s, it is difli cuit to scal e down the horizon tal sed imen-
tati o n tanks due to problems assoc iated wit h adjustin g the ph ys ical c harac te ri s ti cs
ort he /lac, eve n if th e hydra uli c /low characteristics have been properly establi shed.
Co nsequently, traditi onal design criteria , sll ch as the hydrau li c ove rll ow rat e and de-
te nti on time , are used to size the pilot unit s based o n a manageable (sma ll ) Il ow
rate.
Pilot sediment ati o n tanks whose des ig ns arc based o n actua l basi n parameters al -
ways have geometri call y distorted shapes, fo r exa mpl e, very deep and short.
Furtherm ore, unlike th eir prototypes these basins fail to produce effluent of accept-
able quality and are incapable of produc in g a good sludge deposition profile. The
way to produ ce accep tab le quality e fflue nt is to improve Il oc settl eabi lity th roug h
the app li cati o n of an appropriate type of polymer (as a Il occ ul an t a id ) in conjunc-
ti on w ith alum , in add iti o n to installing a hi g h-rate se ttl er module, such as a tube
se ttler in the pilot tank . The addition of th e hi gh-rate settl er mod ul e provides a n
overllow rate of appro ximately I to 2 gpm/ft' (2.5 to 5.0 m/h) for th e surface area
of the tank covered by the settl er modul e.
Based on data obtain ed from numero us pil ot-scale upll ow sludge blanke t c1ad -
tiers, it may be s tated that the overllow rate of the prototype pilot scale unit Ill ay
be lIsed , La an acce ptable deg ree, to model th e upllow sludge blanket claritiers and
the hi g h-rate settler mod ules. The relat ionship between the upward !low ve loc ity
and th e fl oc se ttlin g veloc ity is the major clarification mechan ism for bot h upn ow
reac tor-cla rifi ers Hnd the hi g h-rate settlers. This simple re lati o nship e nables the pi-
lot study to be poss ible, provided th at there is no di stin ct fl ow short-circuiti ng within
the unit.
Hydraulic-scale mode l studies arc not freq ue ntl y conduc ted eve n for hU'ge-scal e
projects beca use they are always tail ored for s pec ilic projec ts and the re fore ca nn o t
FILTERS
eas il y be used for other projects, thus making the m ve ry costl y. However, the hy-
drau li c-scale model study is th e only pos iti ve mean s o f studying the now charac-
ter istics of a sed ime ntatio n tank , including the selection of an e ffec ti ve now di ffuse r
and bartle wa lls. A fcw acade mi c a nd resea rch institutes co ndu ct hydraulic-scale
model studies pure ly for acade mi c interests or for large a nd costly projec ts.
Frol/de's similarity /all' s hould be appli ed to a ll scale-mode l studies :

~
. ' -J'~"
1'\N~"'AV8E ~[0VI:'("1[f1E
" lO>l'lIl'T
r Vl:N 1>t:lUGI< GAAYlTVn a N IS nor M ST JlOOA
a

..........
(I f' 8 I]

~"
Fr
Lg

where v is the characteri s ti c ve loc ity of the syste m, L the charac teri sti c linea r di -
, TV' )
"
~ mc nsion, and g the accelerati on o f g ravity. Fro ud e's law expresses the co nditi o n fo r
l'
@

••.~.";.,,:,~,
:J
IUI'IIIIO·" 'U( A
CD FLO>I'''' El£~
l:l 1.If'(I'11010"""'" Y.. similarity between th e forces of gravity a nd inertia. In o ther words , whe n the Froude
H """Vi
Cl P ~ E 5SU~ E • .ooc;..o.TI'IG "i 'O~OE~
T...~ (
OII WAST!
number is th e same for both the mod el an d th e prototype, the g ravity force to iner-
Al l ' ;PE s <zu I~ M!l.lIf.'nEAS
tial force ratios are the sa me; the paths of now are similar. Th e fund ame ntal rela-
.... l ",. .;G ~ l lIE PVC ' OWASa:
&CH£",,"-( .., ti onships for model s of struc ture , based o n Fr"oude's law, are prese nted in Tahle
Figure 2.3.2-1 Pilot-plant setup. 2.3 .2-f.
 PRELIMINARY ENGINEERING STUDY 27
26 PRELIMINARY STUDIES

TAULE 2.3.2-1 Relationships Between Hydraulic Scale Model and Prototype 10. Plant waste handling and disposa l
Example II . Instrument ati o n and con tro l sys tems
Rc lat ion(/ ( 1/25 Scale Model) 12. Preli min ary cost estim ates
Scale
1:0: I: 1/25 13. Recom mended treatm en t processes as well as instrumentati o n and control ,
Lenglh
1:0: 2 I: 1/625 and waste handling and disposa l
Area
Volume 1:o:J 1:1/15 .625 14. Plant layout and architect ural design
Time 1:0: 0 .:" I: 1/5
15 . Ma nage ment planning o f the design phase
Velocity 1:0 0 .:" I: 1/5
1:3. 125 16. Environmental impact statement (EIS) for the new pl ant
Di scharge 1:0:2.:"
Acceleralion 1:0 0 = 1:1
1:oJ 1:1/15.625 The pre liminary study must conclude with concrete recom mendati ons with re-
Force
1:04 I: 1/390.625 spec t to the proper treatment process train , hydrauli c loading values for each unit
Work
process, and all major eq uipment and appurtenan t fac ilities. At this s tage of the proj-
"0. is the size reductio n rat io for the hydraulic scale model. ect, proper procedure dictates that each of the recommendati o ns be discu ssed with
the appropriate regu lato ry agencies to so licit thei r opinions.
BIBLIOGRAPHY
2.4.1 Quality and Treatability of the Raw Water
AWWA. "Design of Pil ot- Pl ant Studies." AWWA Seminar Proceedings, No. 20 164, America n
Water Work s Assoc iat ion. Denver. CO. May 1982. Data on the surface water quality, take n over a s uffic ient period o f tim e (5 to 10
Ca mp , T. R., "Sedi menl ali on and the Design of SCliling Tanks," ASCE TraIlS., 3:985 (1946). years) , should be both reviewed and eva luated to assess the phys ical, chemical, mi -
Hickox, G. H.. " Hydraul ic Models," in Handbook of Applied Hydra ulics, 2nd cd .. C. V. Davic. crobiol ogical, and radiological characteris ti cs o f the raw water. A risk assess ment
cd.-in-chief. McGraw-Hili , New Yo rk . 1952. mu st also be made in regard to possible contam inati o n o f the water s upply by chem -
Hudson. H . E.. Jr.. Water Clariflcatioll Processes, Van Nostrand Reinhold . New York. 198 1. ica l sp ills or radioactive wastes. Moreover, the degree of present and future la nd de-
Kawamura, S.• " Hydrauli c Scale-Model Simul at ion in the Sed imc ntation Process." 1. AWWA , velopme nt in the watershed mu st be studied .
1':':7:417 (July 198 1) and Proceedi/lgs of the 3nl Amlllal COllference of the Japan \Vater If gro undwater is selec ted as the source of the process water, th e same consid-
Works Association. May 1952. erati o ns associated with su rface water apply. Groundwater as a raw water source ne-
cess itates additio nal studi es, such as the geolog ical conditions, water table, the draw-
dow n o f the water table as the res ult of pumping, problems associated with seawater
2.4 PRELIMINARY ENGINEERING STUDY intrus io n, and the po ten ti al leaching o f indus tri al wastes, domestic wastes, agricul -
tural chemicals, and fertili zers into the grou nd water.
Once the owner has dec id ed to proceed with the co nstruc tion phase of a project For the reaso ns just menti oned, dat a analys is is a very important aspect of eval-
based o n the feasibilit y study, a preliminary eng ineerin g study, inclu ding pl annin g, uatin g the quality of a water so urce. Although arithmetic averages are often pre-
mu st be initiated . This study should address the foll ow ing issues: se nted, a stati sti cal prese ntati o n is ge nerally more meanin g ful. The statistical pre-
se ntation Illay be constructed by pl o tting data on an ar ithmetic or probabi lity scale.
I. Quality a nd treatabilit y of the raw water TIle two most important aspec ts of data analy sis are that it is a means of describ -
2. O bjeclives for the fini shed water quality ing the central tendency o f the data and it allows engineers to determine the degree
3. Additio nal goa ls and objecti ves of variability in the dat a. Components in the raw water whose ma ximum concen-
tration levels are limited by the Drinking Water Quality Standards mu st be a naly zed
4. Restri ctions and constraints on plant design
and evalu ated carefull y.
5 . Alternative treatment processes The treatability of Ihe raw water may be evalu ated through th e use of bench- scale
6. Prelin1inary des ign criteria for trea tme nt unit processes, in cludin g chemi cal tests and a pil o t study. T hese issues are di sc ussed in Section 2.3. If there is an ex-
feed systems is tin g wa ter treatment pl ant in the vic inity of the proposed pl a nt site , the desig n e n-
7. Hydrauli c grade across the pl ant gineers should cons ult the operational data of the ex ist ing plant beca use it will pro-
8. Geotechnical co ns ideration s vide valuable informati o n on the treat abilit y of the raw water.
9. Structural des ig n conditio ns and criteria The 1986 Amendment to the Safe Drinking Water ACI (SDWA-PL-99-3 39) re-
28 PRELIMINARY STUDIES
quired the U.S. Environmental Protection Agency (EPA) to promulgate a National
Primary Drinking Water Regulation (NPDWR) by December 19, 1987. This amend-
ment specified the criteria by which filtration was ordered as a standard treatment
technique for all public water systems supplied by any source of surface water. The
most significant regulations which have the greatest impact on the treatment process
design are the Surface Water Treatment Rule (SWTR) and the Enhanced Surface
Treatment Rule (ESWTR).
Specific parameters of importance include the range of water temperature, cer-
tain microbiological organisms, total organic carbon (TOC), and either toxic or sus-
pected cancer-causing synthetic organic compounds (SOCs) and inorganic COIll-
pounds (lOCs). Microbiological characteristics of raw water thaI must be monitored
arc viruses, Giardia cysts, and C,yptosporidilllll oocysts.
2,4.2 Objectives for the Finished Water Quality
The objective of a public water supply waleI' system is to provide safe and aesthet-
ically appealing water to consumers without interruption and at a reasonable cost.
The required quality is measured against the standards set by state and federal gov-
ernments; both regulate drinking water quality. Since only the national regulations
are discussed in this section, engineers are strongly advised to contact local regula-
tory agencies for further restrictions. In 1974 the EPA established the National
Interim Primary Drinking Water Regulations (NIPDWR) and thereby set the maxi-
mum contaminant levels (MCLs) for a variety of inorganic and organic chemicals,
as well as physical, microbiological, and radioactive contaminants. These primary
standards have been effect since June 1977 and are designed to protect the public
health. Compliance with these standards is therefore mandatory. The National
Interim Secondary Drinking Water Regulations generally relate to the aesthetic qual-
ity of a water supply and thus are recommended goals.
After the enactment of the 1986 SDWA amendments, the EPA promulgatcd many
regulations. They include the following:
1. Phase I: criteria for eight volatile organic compounds (YOCs) and Iluoride
2. TOlal Coli/orlll Rille: MCLG = 0
3. SlIlface WaleI' Trealmenl Rille (SWTR)
4. Phase 2: 38 YOCs, synthetic organic chemicals (SOCs), inorganic chemicals
(laCs). and microbial contaminants
5. Lead alld Copper RI/le
6. Phase 5: 23 YOCs, sacs, and laCs
7. IllformatiolJ Collection Rille: monitoring for microbial pathogens and disin-
fection by-products (Certain large systems are required 10 conduct bench- or
pilot studies of GAC or membrane filtration to evaluate the formation of DBP
precursor and TOC removal.)
In 1996, Congress reauthorized the Safe Drinking Water Act. These amendments
made several modifications to the EPA's drinking water program; the EPA must eval-
PRELIMINARY ENGINEERING STUDY 29
uate five contaminants for possible regulation from a list of drinking water COIl-
taminant candidates. Sulfate, radon, and arsenic are the major items.
Surface Water Treatment Rule (SWTR) The SWTR applics to all public wa-
ter systems using surface water sources or groundwater sources "under the direct
influence of surface water." The SWTR establishes a Maximum Contaminant Level
Goal (MCLG) for the following microbiological organisms:
Giardia lamblia: 0
• Viruses: 0
Legiollella: 0
Since these microorganisms are neither easily monitored nor detected, the SWTR
specifies a treatment technique (disinfection and/or filtration) instead of MCLs.
Criteria for the removal or inactivation of microbiological organisms for public wa-
ler systems are established in the basic requirements of the SWTR.
• 3 log (99.9%) removal and/or inactivation of Giardia lamblia cysts
• 4 log (99.99%) removal and/or inactivation of viruses
In systems utilizing filtration, compliance to a specified treatment level is judged
by the type of filtration process and C . I values, where C is the residual concen-
tration of disinfectant (mglL) and I is the effective contact time 110 (min). The ef-
fective contact time is determined by tracer lest or by using the bafning factors pre-
sented in the SWTR guidance manual. The removal/inactivation of microbiological
contaminants, as recommended by the EPA, arc presented in Table 2.4.2-1 (June
1989) as a function of treatment process.
Under the SWTR, a public water system using filtration will achieve compliance
with the SWTR by meeting the following performance standards:
I. The combined filter effluent must be less than 0.5 ntu in 95% of monthly sam-
ples and must never exceed 5 nlu for conventional and direct filtration treat-
ment plants.
2. The system must maintain a 0.2 Illg/L disinfectant residual entering the dis-
tribution system and must never fall below 0.2 mglL for more than 4 h.
Heterotrophic plate counts (HPCs) can be measured in lieu of disinfectant
residual. For compliance purposes. a sample with a HPC of less than 500
colonies/mL is considered to have detectable residual.
Disinfectants/Disinfection By-products Rule (D/DBP Rule) The DIDBP
Rule (published in the Federal Regisler December 16, 1998) is shown in Table
2.4.2-2. The use of best available technology (BAT) was proposed to reduce or con-
trol all four contaminants. Bromate and chlorite could be controlled by a specific
mode of the ozonation method. However, enhanced coagulation or enhanced lime
softening would be required to reduce/control TfHM and HAA precursors.
30 PRELIM INARY STUD IES PRELIMINARY ENGINEERING STUDY 31

TABLE 2 .4.2~1 Assumed Log Removallhwctiv.:ttion by Disinfection TABLE 2.4.2-3 Required TOe Removal
Assumed Recommended Source Water Alkalinity (0/0) for:
Log Re moval Minimum Log In act ivation
Source Water 0-60 > 60- 120 > 120
by Filtration by Disinfection
TOe (lllglL)" IllgiL mg/L rng/L
Filtrati on Type Giardia Viruses Giard ia Viruses
2-4 35 25 15
Conven ti ona l 2.5 2.0 0.5 2.0 > 4-8 45 35 25
Direct 2.0 1.0 1.0 3.0 > 8 50 40 30
Slow sand 2.0 2.0 1.0 2.0
1.0 1.0 3.0 IINo acti on whcn TOe is less than 2.0 IllglL.
Diatomaceous ea rth 2.0

2. The TOC for treated water must be less than 2 mg/L.

The limits on di sinfectants as dictated by the D/DBP rule are as follows: 3. Raw water TOe must be less than 4 mg/L, alkalinity must be greater than 60
mg/L, and the di stribution sys tem ITHM/HAA leve ls mll st be less than
Chlorille: 4.0 mglL 0.04/0.0:1 mg/L.
Clllorctm ines: 4.0 mg/L as total chl orine 4. Distribution syste m TTHM/HAA levels must be less than 0.04/0.03 mg/L ,
Ch lorille dioxide: 0.8 mg/L with free chl orine as both primary and seco ndary disinfectant (not chlorinel
chloramincs)

IL sho uld be noted that the e nh anced coagu lation requi rement does not app ly to
Enhanced Coagulation Requirements Enhanced coagulati on is defined as
treat ment systems using direct lilt ration o r sys tems obtai nin g a waiver from the fil -
the add iti on of excess coag ul ant to improve the removal of di sinfection by-product
tration requirements of SWTR.
prec ursors through the use of co nvent ional water treatme nt processes. Specific treat-
ment systems musl be utilized to ac hi eve the percent reduction of TOe between ~he
Enhanced Surface Water Treatment Rule (ESWTR) The linal Interim
raw wa ler and treated water prior to final co ntinuous di si nfection. A system us mg
ESWTR was sc heduled to be published in December 1998 and the rule wi ll beco me
ozone and biologically active filters (BAFs) must also ac hi eve the percentage re-
effect ive in the year 200 1. T he rul e wi ll app ly to surface water and groundwater that
duction of TOe prior to the add itio n of a residu al di sinfectant (chlorinc) unless. the
are under the d irect influence of surface water sys tem s serving over 10,000 people.
avo idance criteri a are met. Table 2.4.2-3 (August 1997) illustrates how the requlJ"ed
The EPA is developing a long-term ESWTR th at wi ll be composed of two rules ,
TOe removal targets are linked to source water TOe and the alk alinit y level of the
Long Term (I) and Long Term (2). Curre ntl y, th e requirement s of Lo ng Term (I)
source water.
wi ll be applicabl e to system s serving less th an 10,000 people, with the poss ibility
T he avo idance criter ia for enhanced coagu latio n are as fo ll ows:
of future modifications. Long Term (2) requirements will app ly 10 sys tems serving
over 10,000 people and will incorporate the res ults of the Informat ion Coll ect ion
I. Raw wate r SUVA must be less than 2.0 L/mg per meter, where SUVA is equal
Ru le. Long Term s tandards are expec ted to be published in the year 2002 and be-
to (UV254ITOC) X 100. It shou ld be noted that preozonated raw wa ter may co me e ffect ive in th e year 2005.
be co nsidered as the raw water for subsequent conventional treatment process. The main features of th e ESWTR are as follows:

I. The minimum removal of Cryptosporidilll1l is 2 log.

TAHLE 2.4.2-2 Maximum Co ntaminant Levels of DBPs (mg/L)
2. The combi ned tiltered water turbidity Illust be less than 0.3 ntu in at least 95%
Stage 2 of th e measurements taken each month and mu st never exceed 1.0 ntu.
Con taminant Stage I (Proposed)
3. Systems that emp loy rapid gran ul ar liltration, slow sa nd fiitration, a nd di -
Tota l THM s (TTHMs) 0.080 0.040 atomaceous earth liltration wi ll be gran ted a 2 log reduction credit of
Five haloacetic acids (HAA~s)" 0.060 0.030 Cryptosporidilllll if they meet th e turbidity requirements prese nted above.
Bromate ion 0.010 4. A ll rapid gra nul ar liltratio n surface wate r system s serving ove r J 0,000 peo-
C hlorite ion 1.0
ple mu st monitor Ih e turbidity of eac h filter continuously a nd furnish an ex-
<l HAA ~ includcs trichl oroacet ic acid. dichloroacclic acid. Illonochloroacctic acid, dibromoacctic acid, and cept ions report (violat ions o f the turbid it y criteria) 10 the state on a monthly
llIonobrol110aCClic acid . basis.
 32 PRELIMINARY STUDIES
Other Water Quality-Related Items Design engineers must a lso pay partic-
ular attent ion to several water quality issues. These arc listed below.
Microbial Regrowth Biologically stable finished water is particu larl y impo rtant
for any treatment alternati ves utilizing ozonc. Biological stability requires a reduc-
tion in ei ther assi milable organic carbon (AOC) or biodegradable dissolved organic
ca rbo n (BDOq to raw water leve ls.
MX MX is a sy no nym for 3-chloro-4(diehloromethyl)-5-hydroxy-2[5 HJ-furano ne,
which is a by-product of chlorine disinfection. MX has been identified as a stron g
mutagen in the TA 100 strain of Salmonella typhil1l11riul1I. Additionally, MX studi es
have indicated th at it can induce chromosomal aberrati ons in hamsters, however, it
is not cu rre ntly being cons idered for regulation in ES WTR .
Methyl Tertiary Butyl Ether (MTBE) MTBE is a gasoli ne add iti ve used in the
United States. Leakage of underground storage tank s has resulted in MTB E contam-
ination of groundwaters in certain areas. Currently. there is no MCL enforced for
MTBE; however. studies havc suggcsted a maximum drinking watcr level ranging
from 40 to 200 11g/L. Some states have set more s trin ge nt limits, ranging from 50 to
100 IJ..glL. Three treatment processcs are currently bcing considered for thc trcatm cnt
of MTBE: air stripping. GAC adsorption, and advanced oxidati on. Some studies in-
dicate that conventional air stripping is vcry ineffi cient for MTBE removal. GAC ad-
sorption is an effective process but a costl y meth od of MTB E removal. Ad vanced
oxidatio n processes include the UVIH,O, and O,IH,O, ox idati o n processes. Although
ox idati on processes are equally effective and havc high capitalizati on and operation
cos ts, the 03/H,O, process is considered to be more cost-effecti ve.
Perchlorate Perchlorate (ClO.) is an electroc hem ical ox idan t (exp los ive) that is
in so lid state. It is used in solid rocket fuel (N H.C IO.), lireworks, and vari o us py-
rotech nics. Perchlorate is extremely stable when dissolved in water at concen tra-
ti ons as high as 1000 mglL. In very high doses, it has been used in medicine to treat
Grave's disease and there fore has great biological impac t; when injec ted, perchlo-
rate inhibits the abi lity of the thyroid gland to perform its normal functi on. Based
o n 1995 eva lu ations of perchlorate by the EPA and the States' Drinking Water
Program, a ma xi mum level of 0 .018 mglL has been set to protect publi c health . It
is certai nly possible that the action level could be lowered in the future. Currently,
the issue of perchlorate is limited primarily to groundwater contam ination of a few
regions in the United States. There are four possible treatment processes: anaerobic
biodegradation, reverse osmosis, GAC adsorpti on, and ion exchange.
Arsenic The current MCL for arsenic is 0.05 mg/L. However, it could be reduced
to 0.005 mg/L. The linal rule is to be re leased in Janu ary 200 I.
Filter Wash Wastewater Recycling Rule In Jul y 1998 the EPA held a public
meeting to discuss thi s rule. The following componen ts are considered to be po-
L
PRELIMINARY ENGINEEAING STUDY 33
tential issues for future regu lati ons: ( I ) requirement to recycle filler wash waste-
water (0 the plant head works, (2) requiremen t (0 practice fl ow equalization of the
recycle s trea m, (3) a ban o n the recycling of slud ge, and (4) a rcquirementto pro-
vide effec ti ve treatmen t prior to the recycling of filt er wash wastewater. Thi s regu -
lation is sc hedu led to be published in the summe r of 1999 and the final rule pub-
lished hy August 2000.
Algal Toxins Certai n blue·green algae produce hepatotoxin and neurotox ins.
Hepa totox ins are primarily pept idcs. Coml11onl y identified compounds are micro-
cystins re leased by MiclVcyslis aergil/Gsa and nodularin produced by Noduralia
damage. These toxins can cause death or liver cancer. Neurotoxins are alkaloids
slIch as the anatox ins released by Allabaellaflos aquae. Neurotoxins can cause rapid
death through paralys is of th e respiratory and cardiac muscles.
Number of Particles in Filtered Water (2 to 10 11m Range) Cryptosporidil/I/l
oocys ts (2 to 4 11m) a nd Gianlia cysts (4 to 10 11m), are not easily di sinfec ted by
conventional chlorination. Although olonation and membrane filtrati on are very ef-
fective agai nst these protozoans, many existin g plants have not adopted either of
these processes, for many reasons, 3mong them high capital cost, ti ght hydraulic
conditi ons, and hi gh plan t output. Recent pilot stud ies, however, indi cate that a reg-
ular conventional treatm en t process is capable of remov ing particles in the range 2
to 10 jJ.ITI range at a 3 to 4 log remova l rate provided that the operation of each
process is optimi zed. The current goal of filtered water turbidity of less th an 0.1 ntl!
is achieved primarily because of the high level of particle removal. However, ccr.
tain states are starting to adopt more stringent guidelines based on a number of par-
ticles per milliliter. The regulatory agency of New York has set a criteri on of less
than 50 particles/mL for particles in the range 2 to 10 f..I.l1l . M ajor consulting engi-
neers are setting a goal of 25 particles/mL for the same range of particles for th eir
pi lot filter stud ies.
Federal Drinking Water Regulations T he c urrent Trea tment Rules and
Drinking Water Quality Standards are subjec t to be mod ified upon the publication
of various studi es. For instance, the fCR studi es conducted by the EPA are antici-
pated to be published early in the twenty-first century; some of its conclusions may
alter a few of th e curre nt MCLs for certain contaminants. Under these circumstances,
engi neers should be aware of current federa l and state regu lati ons. An excellent sum-
mary of updat ed Rules and Drinking Water Quality Standards, wit h BAT informa-
ti on, is published on an annual basis by the Journal of tile American Water Works
Association (JAWWA).
The currently enforceable USEPA Primary Drinking Water Standards are shown
in Table 2.4.2-4 and the Secondary Standards in Table 2.4.2-5 . Table 2.4.2-4 also
prese nts Europea n Community (EU) Drinking Water Gu ide levels in addition to the
World Healt h Orga ni zati o n (WH O) g uide li nes for projec ts outside the United States.
TABLE 2.4.2-4 Primary Drinking Water Standards or U.S. EPA (1997) .md TABLE 2.4.2-4 (Col/til/lied)
International Drinking Water Guidelines (1989)
u.s.
u.s. Environmental
Env ironment al World Health Protection
World Health Protect ion Organization Agency
Organization Agency Parameter EU Guide levels Guidelines Standards
Parameter EU Guide levels Guidelines Standards
Radiollllclides
Microbials Rad ium-226 +
< 5 pCi/L
Bacteria < JOhnl at 37' C Radium-228
< 100/ml at 22°C Gross Alpha
< 15 pCi/L
Coliform Group Not detec ted in NO! detected in Not present Beta particle & ph oton
4 mRem/yr
100 ml 100 ml activity
ex posure
E. coli Not present
Giardia 99.9% removall Orgallics
in act ivat ion Phenol
Viruses < 0.0005 mglL *
99.99% Total Trihalomethanes
remova ll < 0.1 mglL
1.1, I-Trichloroethane 2 mg/L (P)
in acti vation < 0.2 mglL
Tetrac hloroethylene < 0.04 mg/L < 0.005 mglL
Trichloroethylcne < 0.07 mg/L
II/orgal/ics < 0.005 mglL
Dichloromethane < 0.02 mglL < 0.005 mglL
Lead < 0.05 mglL* < 0.01 rnglL < 0.01 5 & Benzene < 0.01 mglL < 0.005 mglL
treat ment Toluene < 0.7 mglL < I mglL
standard Ethylbcnzene < 0.3 mglL < 0.7 mglL
Fluoride < 1.5 mg/L* < 1.5 mg/L < 2.0 rng/L ** Xyl enes < 0.5 mglL < JO mg/L
8-12' C I. I-Dichloroethyl ene < 0.03 mg/L < 0.007 mglL
< 0.7 mglL* Carbon tetrachl oride < 0.002IllgiL < 0.005 IllgiL
25-30' C Aldrin and Dieldrin
Arsenic < 0.00003 mglL
< 0.05 mglL* < 0.0 I mglL (P) < 0.05 IllgiL Benzo(a)pyrene
Selenium < 0.0 1 mglL* < 0.01 mglL < 0.0007 IllgiL < 0.0002 IllgiL
< 0.05 mglL Chl oroform < 0.2 mglL See total
Mercury < 0.00 I mg/L* < 0.001 mg/L < 0.002 mg/L
Cyanide < 0.05 mglL* < 0.07 mglL < 0.2 mg/L trih alolllcthanes
2A-D < 0.03 mglL
Chromium (Cr6 +) < 0.05 mglL* < 0.05 mglL (P) < 0.1 mg/L *** < 0.07 mglL
DDT < 0.002 IllgiL
NH, - N < 0.05 mglL as < 1.5 rng/L as I.I -Dich lorocthene < 0.003 mglL
N H, NHJ Heptachlor epoxide and
< lmglL < 0.00003 mglL < 0.0002 mg/L
Heptachlor
< 25 mglL as < 50 mg/L as < 10 mglL & < 0.0004
NO, - NO,- mglL.
Total Nitrate + Nitrite < IO mg/L respectively
Hexachlorobenzene
Cadmium < 0.005 mglL * < 0.003 mg/L < 0.005 mglL < 0.00 I mglL < 0.001 mglL
Lindane < 0.002 mglL
Fluoride < 4.0 mglL < 0.0002 mg/L
Antimony Methoxychlor < 0.02 mglL
< 0.006 mglL < 0.04 mg/L
Asbestos Pentachlorophenol < 0.009 mglL (P)
< 7 MFL < 0.00 1 mglL
Barium 2,4,6-Trichlorophcnol < 0.2 mglL
< 2 mglL
Beryllium < 0.004 mglL 1,1 .2-Trichloroeth ane
< 0.005 mglL
Nickel < 0. 1 mglL 1.2-Dichloroethane
< 0.005 mglL
Thallium < 0.002IllgiL 1.2-Dichloropropane
< 0.005 IllgiL
Turbidi ty 2 < OAJU < 5 NTU < ~ 0.5 NTU 1,2,4-Trichlorobenzene
< 0.07 mglL
in 95% of cis-I,2-Diehloroethylene
< 0.07 mglL
sa mpl es per trans-I,2-Dichloroethylene
< 0. 1 mglL
month Monochlorobenzene
< 0.1 mglL
34 35
 PRELIMINARY ENGI NEERING STUDY 37
36 PRELIMINARY STUDIES

TABLE 2.4.2-5 EI)A Secondnry Drinking Waler Siandards

TAIILE 2.4.2·4 (COll/illlled)
U.S. Contaminant Ma ximum contaminant level
Environmental
A luminum 0.05 to 0.2 mglU'
World Heal th Protecti on
Chloride 250 IllgiL
Organization Agency
Co lor 15 col or unit s
Parameter EU Guide level s Guideli nes Standards
Copper I IllgiL
< 0.6 mglL Corrosiv ity Neither corrosi ve nor sca le fOflning
o- Dichlorobenzene
< 0.075 mglL Flu ori de 2.0 mglL
p-Dichlorobenzc ne
< 0. 1 mgIL Foaming age nt s 0.5 Illg/L
Styrene
< 0. 002 mglL Iron 0.3 IllgiL
Vinyl chloride
< 3 X 10- ' mglL Man ganese 0.05 Illg/L
V.7.S-TCDD (Dioxin)
< 0.05 mglL Odor Three th reshold odor numbers
2,4,5-TP (Silvex )
< 0.002 mglL pH 6.5 to 8.5
A lachl or
< 0.003 mglL Si l ve r 0. 1 IllgiL
A trazine
< 0.04 mglL Sulfate 250 mg/L
Carbofuran
< 0.002 mg/L Total di sso lved solid s 500 Illg/L
Chlordanc
< 0.2 mglL Z inc 5 mg/ L
Dalapon
< 0.5 mg/L
Di(2 -cthylhex yl)adipatc "Level se lected ro r alu minUIll depends on Ihe discn.:tion of the slate primac)' agency.
Di(2-cthylhcx yl)phthalatc
< 0.006 mglL
< 0.0002 mglL
Dibromochl oropropanc
(DBCP)
< 0.007 mglL
Dinoseb Design engi neers have trad iti onall y considered each unit process as an indi vid-
< 0.02 mglL
Di'luat ual item and have optimized eac h unit sepa rat ely on the premi se Ihat thi s wou ld re-
< 0. 1 mglL
Endothall sult in the most e flic ie nt overall plant design. However, suc h suboptimi zati on usu-
< 0 .002 mglL
Endrin all y does not result in oplim iza ti on of th e total plant design. Additionally, o~iec tives
< 0.00005 mglL
Ethylene dibromide
< 0.7 mglL such as optimi za ti on of initial costs, life-cycle cos ts, and co nstru ction tim e should
Gl yphosphatc
Hexac hlorocyc lopcnl ad iene
< 0.05 mglL be evaluated in thc preliminary cngineeri ng study.
< 0.2 mglL During the early stages of the projec t, e ngineers should establish basic desig n
Oxamy l (vydate)
< 0.5 mglL parameters fo r major fini shed wa ter qualit y and a few plant operatio nal issues.
Picloram
< 0.0005 mglL Suggested goals are:
Polychl orinated biphenyl
(PCB)
< 0.004 mglL
Simazine 1. Thc ran ge of plant !low rate should be no more than 4: [.
< 0.003 mglL
Toxaphene
< 0.001 mglL 2. The c larifier crllue nt turbidity must be less than 2 nlu .
Hexachlorobenzene
3. The frequency of filter washing should be limited to once a day per lilter al
(P) I)rovisional value the designcd filtration ratc.
"' Maximum Ad missible Concentrati on
"'. Secondary standard 4. Filtered water turbidity of less than 0.1 ntu.
"'* Tolal for all spec ies 5. Partic les in the liltcred water shou ld be less than 50 particles/mL (those larger
than 2 lun).
2.4.3 Additional Goals and Objectives 6. After bac kwash, di scard filtered wa te r wi th turbidit y above 0 .25 ntu .
Beyond th e lini shed water qualit y objecti ves, the ow ner 's go~l s for the water treat- 7. Aluminum in fini shed water should be less than 0.05 mg/L.
ment plant mu st be identifi ed and es tablished. The goals 01 the overall tr~.atme nt 8. TOe o f liltered water should be less th an 2.0 mg/L (or less than 2.0 SUVA).
plant should state what the plant is to accomplish and ho.w .th ese g~al s allect t~1e 9. TTHMs should be less th an 40 f.lg/L.
water di stributi on system s and the e nvironment. Once till S IS es~abh s hcd , th~ le-
10. HAA, should be less than JO f.lglL.
quired fun ction of eac h unit process (of the treatment proce~s tram) must be Iden-
tili ed and th e objec ti ves of eac h of these unit s should be de fin ed . II. Bromate should be less th an 10 f.lg/L.
38 PRELIM INARY STUDIES
2.4.4 Restrictions and Constraints on Plant Design
Project engineers arc rarely. if at all, given compl ete freedom in th eir designs. Thus
res tri cti ons and constraints must be clearl y defined at the beginning of the design
phase. Restrictions may be due to economi c, physical, chemical, temporal. climatic,
geo logical, soc iolog ical, legal, or aes thetic co nsiderati ons and Illay be imposed by
local, state, or federal age ncies.
Constraints are defined as restrictions placed on the acceptable values of design
va ri ables (properlies of Ihe syslem Ihal are co nl ro lled by Ihe designers). Examples
of constraints arc building codes, zoning laws, and OSHA regulations and standards.
For instance, a :B-in. (84 cm)-diameter steel pipe is not standard size, nor is a 70-
hp (53- kW) electric motor a stand ard item . These restrictions and co nstraints may
fi x Ihe componenl properlies or eSiablish a range in whi ch Ihey mllsl fall. Allhough
these restrictions are ge nerall y made for reasons of sta ndard iza ti on and safety, se-
rious design problems often ari se when designing treat ment plants for developing
countries, due to the limited number of available com pone nt s, materials, tec hn ol-
ogy, and qualified person nel, in add ition to the bureaucratic red tape associated with
Ihe impOrialion of goods.
2.4.5 Alternative Treatment Processes
The basis for select in g treatm ent process alternatives is established by the charac-
teristi cs of the raw water and the finished water quality goa ls. Conside rati on must
be given to future implementation of more strin ge nt EPA drinking water quality
standards and to possible changes and variabi lity in the raw water quality. Thus the
goals and object ives, as well as the restrictions and constraints defi ned in the pre-
ceding sec ti on, all bear upon the selection of alternative processes. Furthermore, the
avai lability of major eq uipmen t, postinstaJlation serv ices, a nd the capability of op-
erators and maintenance personnel , as well as the waste handlin g requirements a nd
the availabi lit y a nd cost of water treatment c hem icals, all great ly affec L the selec-
tion of the water treaLment process, especiall y in remo Le regions and developing
count ri es.
Fi nal selecti on of the most appropriate treatment process sc heme must be based
on reliability, co nstructabi lity, case of ope rati on, simple maintenance, and most im-
portant, cost. Value analysis may be practiced during the selec ti on process, with re-
spect to the whole system and the individual unit processes. A cost analysis of the
various alternati ves should be based on two criteria: prese nt worth and useful life.
Bench·scale sl ud ies and pilol sill dies. cond uclcd during e ilhcr Ihe fcasibilily slud y
or preliminary e ngineeri ng study phase, arc quite va luable in confir min g the cflcc-
tiveness o f the alternati ve processes selected if the e ngineer is unable to draw on
previous experiences in treating the raw water. Experience acqu ired from ex isting
water treat ment plants (treating the same source of water) and plant-scale sim ul a-
tion tests provide important guides in selecti ng both the treatment process sc heme
a nd the hydraulic load ing of each unit process.
There arc thrce basic water purification processes with six modified schemes: ( I)
micro membrane filt rat ion, (2) direct filtration, and (3) conve nti onal complete treat-
b
PRELIMINARY ENG INEER ING STUDY 39
ment. Modifi~d vers ions an: ultra membrane filtration, nano membrane filtration,
rever~e osmosIs .membrane hltration, in-line (co ntac t) filt ration, two-stage filtrat ion ,
an~ lime softelllng process. The first three arc modified microfiltration processes
uSlOg smaller pore membranes. In-line fill ralion is allained by deleling Ille n
I . . " occu-
atlO.n stage I.rom the d.lre~t-hne, filtrati on process. Two-stage filtrati on uses a coa rse
medl~ roug hing. filter III lie u 0 1 the reg ul ar flocculation process of direct filtrat ion.
The Illne softelll~g process is sim~lar. to a conven ti ona l complete trea tment process
except that .the. pnmary co~~ulan t IS hm e, recarbonation is achi eved through the usc
of carbon diOXide, a nd c lanl lCat.io n is conduc ted before filtration (standard scheme).
Table 2.4.5-1 presents. the applicable raw water qualiti es for Some practical water
~rea~ment processe~. Flgur~ 2.4.5- 1 illustrates six water treatme nt process trains. It
IS vita l that the project engllleer provide both space and exira hydrau lic head across
the pro.posed treatl~ent process train during the final design stage, as the rul es and
regulallons goverm ng waler qualily wi ll be modified in Ihe fulure.
2.4.6 Preliminary DeSign Criteria for Treatment Unit Processes
The pre~iminary size of all major treatment unit processes mu st be established in
p!~parall~n ~or the cost co mparisons ~nd site plans used in evaluati ng si te adequacy.
~ hiS prelllnma? re port should co nSider the coag ulati on, fl occulati on, sed imenta-
tIOn, a~~ filtr~tlOn pro~~sscs; the disinfection process, including type of disinfec-
ta~t, diSinfecting c?nd lt lon, a nd by-products; the filter washing, was te-wash han-
dling. sludge ha ndling, and chem ical feed and slorage facililies; Ihe clearwell; a nd
~he contr~1 bUllchng: T.he ozonati on a nd GAC adsorption processes shou ld a lso be
IIlcluded III the prehl11mary design report if they are integrated into the treatment
plant. Table 2.4.6-1 is an example of preliminary design criteria.
2.4.7 Hydraulic Grade Across the Plant
'.t i~ import3l~t to ~stabli sh the hydraul ic grade line across the plant early in the pre-
IlIll.mary engll1eer~ng .s~udy because both the proper select ion of the plant site e le-
vat ion and the suitability of the site (for executing balanced cu ts and fills to ac-
commodate all process. unit s ~equiring speci fi c water elcvat ions and depths of
struc tu res) depend on thiS conSiderat ion.
Mosl co nvenliona l waler Irealmenl planl s require 1610 17 fI (4.9 10 5.2 m) of
head loss across Ihe planl. ThiS means Ihal a difference of 1610 17 fI mllsl exisl be.
Iween Ihe waler level al Ihe head o f Ihe planl and Ihe hi gh waler level in Ihe clear.
well, w h~ch is the tail .end of the treatment plant process tra in . Modern treatm ent
plants ~Slflg preozonatlon, as well as poslolonalion and gra nul ar ac ti vated carbo n
acisorpllon processes, require a lmosl 25 fI (7.6 Ill) of available head across Ihe planl.
Under the~e c lrculllstances, If the plant site is flat, the following criteria must be
mct; the 11Igh water level in the clearwell must be se t at ground level because of the
grou nd waler lable; Ihe waler leVel a l Ihe head of Ihe process Irain mu sl be 25 fI
(7.6 m) above Ihe ground level; and Ihe majorily of Ihe unil processes in Ihe firsl
half of the process trai n must necessarily be in the air unless a pumping station is
 PREUMINARY ENGIN EER ING STUDY 41

included in the process train. It is therefore obvioLls Ihal a fl il! and level site is not
the bes t choice for this type of trealment plant. The ideal plant site wi ll have a 3 10
5% one· way slope and a gro und e levation lhal sati sfi es the necessary e leva tio ns.
Fig ure 6. I -2 is an example of the hydrauli c grade line across a plant .

2.4.8 Geotechnical Cons iderations

Ground considerations of the proposed plant site greatly influence the entire con-
struction cost, the durati on of construction, and plant maintenance. Items of major
imporlance are the availability of information necessary to design the foundati ons
of the treatment faci lities, ground characteristics that cou ld affect construction, and
so il charact eri stics that allow e ng ineers to provide measures against cut-a nd-fill
slopes as well as corrosion of pipes and concrete structures.
OONOO.., ~
°888 oo r-:r....:booo The fo ll owi ng items can be o nly obtained from a geotechnical investigation.
Y-NN
vVV
VVVVV VVV
I. Soil pressure: a llowable bearing pressures under foo tings, latera l earth pres-
sure against cantilever walls, lateral earth press ure aga inst fully restrained
walls, passive res istance force, friction coefficient between footing bases and
the soil. and anticipated deg ree of sett lement
~
~

~ 2. Data 0 1/ excavation and fill: cut slope for excavation. permanent cut-and-fi ll
~
slo pes, shoring c ut angle fo r pipe tren ches, type of equipment req uired for ex-
...e cavation , a nd compac tio n requirements under structures, around struc tures,
c~
a nd under roadways
e
;; J. Groundwater level: ex pected high groundwater level and seasona l fluctuati ons
~ 4. Site seismicity: peak ground acceleration force and the seismi c surcharge earth
;; load against the walls of buried struc tures ; must be used in the structural
.~ design
1:1
..."
~

.£ 2.4.9 Structural Design Conditions and Criteria

.£
;; The preliminary engineering study includes both general and specific stru ctura l de-
=
c;
or. - N 1""0 8 8 b
0 ci sign consideratio ns and c riteria for the proposed plant site. Genera ll y, all struct ures
8 -or.or.vvvv-
III 0 0
__ -
f"'l
~
.'!l Vv VV vvvv mu st be structura ll y sound, that is, capable of withsta ndi ng dead weig ht. live loads,
water pressure, earth press ure, vario us forces resultin g from earthquakes, vibrations
$ produced by movi ng mechanical equipment, loads arising from the snow pack, ice
~
=-:" pressure, wind pressure. and the anticipated loading and impact during construction.
] Since a maj ority of the structures are water bearing (tanks and basins). they must
~ be leakproof. Moreover, structures holdi ng filtered and fini shed water must no t be
:2 polluted by untreated water or wastewater. Finally. strllctures situated in areas with
=
<J}
hi g h underground water leve ls or where the underground water levels become hi gh
-'"~
, during the rainy seasons sho uld be designed to resist buoyancy, especially when the
tanks are empty. When all these requirements arc considered, the minimum thic k-
I>l
ness of the reinforced concrete walls and slabs of water-bearing stru ct ures sho uld
o-l be 8 in. (200 mm), regardless of the loading conditions.
i:i'" (le.tt COl/IiI/lies 01/ page 49)
40
 PRELIMINARY ENGINEERING STUDY 43
42 PRELIMINARY STUDIES

Sewer
Chemical Cleaning Waste

Micro Membrane Filtration Process

Sludge
Two-Stage Filtration Process

i!

Wash Waste
Holding Tank
Stud e
ThIckener

Direct Filtration Process with Preozonation

Direct Filtration Process

L_...,,==--+--!
Recycle
Clatifler
Conventional Treatment Process with Ozonation and GAC Filters
(Micro Filter as Alt.)
(b)
conventional Treatment Process
Figure 2.4.5·) (Col1 til1ued)
(a)
Figure 2.4.5-1 (a) Three basic water treat ment processes: (b) modifi ed processes.
TAULE 2.4.6-1 Exnmple of Design Criteria 'fAilLE 2.4.6-1 (Colltilllled)

FAIRFAX COUNTY WAT ER AUTHORIT Y

FREDERI C K P. GRIFFITH. JR . Mixing timc
WATER TR EATM ENT PLA NT @126 mgd 1.1 sec
@40 mgd
3.5 sec
Mi xi ng pump
Plallt capacity Number 2 ea
Design capacit y 126 mgd Type: Horizontal end suction
87.500 gpm Capac ity 1.200 gpm
195 cfs TDH 48 ft
11.697 cfm Nominal motor size 20 hp
Average dail y fl ow 75 mgd Flocculation basins
52.083 8pm Number
11 6 cfs 4 ea
Stagcsibasin 3
6.963 cfm ea
Water depth 15
40 mgd ft
Minimum fl ow Basin volumc (each) 105.300
27.778 gpm ClI ft
Total volumc 42 1.200
62 cfs Cll ft
Detenti on time
3.7 13 cfm @ 126 mgd 36 min
Influent flow metering @40mgd 11 3 min
Number 2 ea Slages I & 2
Type: Magneti c fl ow meter Ty pe: Hydra ulic flocculation
Si ze 54 in (barn ed hori zont al serpentine fl ow)
Range 20-85 mgd Bames/slage 24 ea
Encrgy input (G X t)
Flasl! mixing 1st stagc
Primary system @ 126 mgd
ea 4 1.1 50
Number @40 rngd
Type: Hydraulic pump di ffu sion 24.400
in 2nd stage
Si zc 72
scc - I @126 mgd 28.800
Mi xing energy (G) 730
@40 mgd 16.600
Mi xing zone lcngth 9.0 ft
Slage 3
Mi xing zone di ameter 6.0 ft
cu ft Type: Mechanica l fl occulation
Mi xing zone volume 254.3
(vari able speed. vertica l hydro foil impellers)
Mi xing time Number (per basin)
@ 126 mgd 1.3 sec 6 ea
Compartment size 20 x 19.5
@40mgd 4.1 scc ft x ft
Compartmcnts/basin 6
Mixing pump ea
ea Compartment volume 5.850
Number ell ft
Depth to equivalent di ametcr rat io 0. 77
Type: Hori zolllal cnd sucti on Energy inpul (G x I)
Capaci ty 2.300 gplll
@ 126 mgd va riable
TDH 70 ft
@40mgd variabl e
Nominal motor size 60 hp
Motor size
Secondary systcm hp
Number 2 ea Sellimelllatioll basins
Ty pe: Hydraulic pump diffu sion Type: rectangular
54 in Number 4
Si ze
690 sec
-, Size
ea
Mi xing energy (G) 295 X 120 ft X ft
Mi xing zone length 6.8 ft Watcr depth 16 ft
Mi xing zone di amcter 4.5 ft Basin volume (each) 566.400 Cli ft
107 .0 cu n Total volume
Mi xing zone volume 2.285.600 ell n
44 45
 TABLE 2.4.6· 1 (COl/iiI/lied) TABLE 2.4.6· 1 (Co lltinu ed)

Yl!.l.u£ !!D.iJ.s Value

Detention Time N umber !!D.iJ.s
3 ca
@126 Il1gd 194 min Air scour rale
3 se rm/sq ft
@40 mgd 6 10 min Capac ity
2.000 scfm
Horizonta l flow veloc ity Pressure
8.5 ps i
@126 mgd 1.5 rpm Molor size
125 hI'
@40 mgd 0.5 fpm Clearwell
Surface load ing Type: rectangular. buried cast· in·pJace cone
@126mgd 0.6 gpm/sq ft Nu mber of cells
@40 mgd 0.2 gplll/sq ft 2 ea
Ca pac ity (each)
Basin length to wi dth rati o 4.9: 1 6 Illg
TOlal capac ity
Weir loading 12 mg
Dimensions
Effl uent wei r lengt h (each) 100 ft 208 X 352 ft X ft
Water depth (max imum)
400 20 ft
Effl uen t weir lengt h (tota l) ft
Finished Water Pumping 51(1(;01/
@126 mgd 219 gp mlft
High service pumps
@49 Il1gd 69 gp mlft
Type: vertica l diffusio n vane (turbine)
Sludge wi thdrawa l
Num be r
Type: chain and fl ight . 3 ca
Capaci ty (each)
Number of longitudinal coll ec tors per bas1Il 12 en 21.600 (3 1) gpm (mgd)
Total capac ity
Number of cross coll ectors per bas in en 64.800 (93) gp m (mgd)
TDH
270 ft
Filters MOlor size
1.800 hI'
Type: deep bed grav it y Ma in service pumps
Number 14 ea Type; vertical diffu sion va ne (turbine)
Number of cellS/fiiter 2 co Number
3 ca
Ins ide dimcnti ons (eac h cell) 34 x 17 ft X ft Capac ity (eac h)
16.000 (23) gpm (mgd)
Medi a surface area (eac h filter) 1. 156 sq ft Total capacity
48.000 (69) gpm (mgd)
Medi a surface area (total) 16. 184 sq ft TDH
95 ft
Filtrati on rates (@ 126 mgd) Motor size
500 hp
Two fi lt ers off lin e 6.3 gplll/sq n Ozone System
All filters on line 5.4 gpm/sq ft General
Filter media Applied Total
Type: gra nu lar act ivated carbon/sand Plant Ozone
0" Gas now
Ozone Ozone 0" How per
Deplh 72110 in Flow Dosage Conccn. Production Flow Con lac lor No. of
P'"
Diffuser
Oxygen
Condi fion (mgd) (mgIL)
Required
Effecti ve size 1.4/0.54 mm (%) ObJday) (sefm) (sefm) Diffusers (sefm) «pd)
Uni fo rmity coc ffi ce nt < 1.4/< 1.4 Desig n 126 3.8 12 4.000 3()() 75 70 1.1 17
Speci fi c grav ity 1.35/2.65 Average 75 3.8 7 2.4()() 3()() 75 70 1.1 17
Mi nimum 40 2.5 2.5
Underdrai n type Dua l lateral 830 3()() 75 70 1.1 17
Maximulll 126 5.7 8 6.000
Filter backwash 630 160 140 1.1 37.5
Uhilllale 168 5.7 8 8.000
Maximu m rate 22 gpll1/sq ft 840 140 140 1.0 50
Normal rate 16- 18 gp m/sq n Oxyge n suppy
Durati on (total) 15 min Vapsa plant capacity
17 Ipd
Backwash pumps Min im ulll oxygen concentrati on 90 W1. %
Type of pump: vert ical ce ntri fuga l Back up lox storage volume 13.000 ga l
Number 3 ea Ozone ge nerati on
Capacil y (eac h) 12.500 gp m Num ber of ge nerators
3 ea
TDH 30 n Capaci ty (each) @12% wt. ozone concentrati on 2.000
Motor size hp Ib/day
125 Capacity (each) @8% wt. ozone concentration 3.000 Ib/day
Air scour blowers Applied ozone dosage
5.7 mglL
Ty pe: posit ive displae ment Number of future generators
ea
46

..
47

b
 TAIlLE 2.4.6·1 (CoIl/iII/led)
PRELIMINARY E NGINEERI NG STUDY 49

~ ~
Ozo ne contractors
Number of co nlaclors 4 ea
31.5 mgd
Capacity (eac h)
10 ea
Num ber of cell s
10 min
Detenti on lillle
22 rt
Water depth
Number of diffu sers in first cell 70 ea
Number of diffu se rs in second ce ll 70 eo
Typical gas n ow per diffuser I scfm
Number of future ozone contac tors 2 ea
Ozo ne destructi on
Number of ozone destru ct unit s 3 ea
300 scfm
Capacity (each)
3·5 in H2O
Vacuum o n conlaclors
Number of future ozone dCSllllct unit s ea
Cooling wate r
420 gplll
Flow per ozone generator
60 gpm
Flow per power suppl y unit
35 gpm
Flow 10 Vapsa system
85 OF
Ma ximum inl et tempe rature
Total coo ling water Il ow (des ign) 1.355 gp m
1.795 gpm
Total cooling wate r now (future)
Dosage (mglL)
Chem ical
Point of Application Minimum Average Maximum
Primary !lash mi x 20 30
Fe rric chl oride (Pri)
Primary !lash mi x 0.5 2 4
Cationi c polymer (A ll)
Prim ary Il ash mix 3 8
Lime (A It )
4 5
Sod ium hypochl orite (A ll) Primary fl ash mi x
Secondary flash mi x 20 30
Ferric chl oride (AIt)
Secondary Ilash mi x 0.5 2 4
Cationi c polymer (Pri )
Secondary nash mi x 3 8
Lime (Pri)
Flocculation (2 nd stage) 0. 1 0.25
An ionic Polymer (Pri)
Fl occulati on O rd stage) 0. 1 0.25
A ni on ic Pol ymer (AIt)
Ozone contactor cfllu cn t 0. 11 0.44
Sodium Bisulfite (Pri)
Filter innuent 0.0 1 0.D2
Noni oni c polymer (Pri )
4 5
Sodiulll hypoc hl orite (Pri ) Clearwe ll in!lucnt
Clearwcll inll uent 15 30
Sodium hydrox ide (Pri )
0.5
I-I ydrolluosili cic aci d (pri) Clcarwell inlluent
Clearwell inllucnt 1.25
Aqueous amm onia (Pri )
4 5
Sodium hypochlorite (A ll ) Clcarwell cflluent
I \.25
Aqueoll s ammoni a (A It) Clearwcll cfllu ent
Hi gh serv ice pu mp 0.5 2 4
Zinc orthophosph ate (Pri)
di scharges
Main service pump 0.5 2 4
Zi nc 0I1 hophosph atc (Pri)
di scharge
Washwater suppl y 0.1 0.25
Non ioni c polymer (A ll)
4 5
Sodium hypochl oritc (A ll ) Washwatcr supply
With the excepti on o f tropica l reg ions, the ambi e nt and wa ter te mperatures c hange
sig nifi ca ntl y with th e seasons. T hu s the plant des ig n mu st acco unt fo r thermal
stresses and th e expa nsion and contraction of th e stru ctures. T he loca l weather co n-
diti ons de te rmine the te mpera ture ra nge, and thi s in turn stri c tly di ctates th e spac-
ing of the ex pans ion j o ints for bo th co ncre te and metal structures.
The expansio n joints for reinforced co ncrete stru ctures are usuall y s paced every
33 to 65 ft ( 10 to 20 m ), depe ndin g o n fac tors suc h as if the struc ture is ex posed to
sunli ght, fill ed with water, or located underg ro und. These jo ints should provide a
space o f 0.5 to I in . ( 12 to 25 111m ) to a llow move me nt o f the structures. Each ex-
pans ion joint mus t have a wate r stop, joint liller materi al and sea lant , a nd s pec ia l
measures that all ow the re inforci ng concrete bars to move at the joint (e.g., a re in -
forc in g bar in a pi pe sleeve fill ed wi th grease). Although currently not cons idered,
eng ineers should be aware that inves ti gation o f rece nt earthqu ake damage has ide n-
tified a s tandard ex pans ion jo int desi g n us in g a p lasti c sheet across the jo int as a
water sto p barrier, as a weak po int in water- ho lding stru ctures during stron g earth -
quakes. Th e No rthridge, Ca lifo rni a ea rlhquake o f 1994 (mag nilud e 7 .3) and the
Kobe . Japan ea rthquake of 1995 (mag nitud e 7.9) damaged over 50% of the pl asti c
wate r sto p barriers used in tank s and tilters, there by res ulting in Il ooding of the til -
ler pipe gall e ri es.
With respec t to concrete and reinforced concrel e structures, the water/ce me nt ra -
lio is a critical factor in bot h the streng th o f the concrete, corrosion of co ncrete s ur-
face, and the degree of shrink age. This rati o is ge nerally limited to less than 0.5.
Thi s li gure a lso ma int ains the compress io n stre ng th o f the wa ll s and s labs at a min-
imum of 4000 psi (280 kg/cnl ). T he allowabl e s hrinkage rate o f wate r treatme nt
65 concrete tanks is usually limit ed to 0.04 to 0.05%. When des ig ning trea tme nt plants
for reg io ns ex periencing earthquakes, engi nee rs Illllst use the proper se is mic acce l-
20
erati on coefli cien ts (those recomme nded by the local building codes and geotec h-
10
nical cons ultants) in all structural des ig n co mputa ti o ns. Should a pl ant s ite have hi g h
65
levels of sulfate in the soil and/or underground water, type 2 or ty pe 5 ceme nt may
20 be considered. Genera l-p urpose type I ce ment is usually not capabl e o f withstand-
0.5 ing the corros ive con diti o ns.
0.5
2.2
2.4 .10 Plant Waste Handling and Disposal
0.05
10
Water treatment plants y ie ld two basic types o f wa ste: recove rable waste and nonre-
45
covera ble waste . The first type include filter was h wastes, wastewate r from the fil -
I
ter-to-waste process, the supernatant o f the s lud ge drying bed s, and pl ant overllow.
2.5
T he latter include s ludge from bot h the clarifie rs and the tilter wash-waste ho lding
10
2.5 tanks, sa nit ary and chemical wastes, and wastes produced by o ther unit processes,
such as the sludge press or ion exchange r. The acceptable modes of waste han dlin g
and process in g are determined by the Water Pollution Cont rol Act A me ndm ent of
1972 and the National Primary Drinking Water Reg ulali o n ( 1986 Ame ndme nl s). Thu s
the preliminary eng in eerin g s tudy mu st be conducted with these regul ati o ns in mind.
0.5 In the pas t, the rec overable wastes were co llec ted in a ho ldin g tan k, then recy-
10
c led to the head o f the plant witho ut furth er treatment , except in cases where nat-

48

.. >
50 PRE LIMINARY STUDIES
ural plain sedime ntati o n was empl oyed . However, the new E PA drinkin g water treat-
ment guidelin es now consider thi s prac tice to be un acceptab le. Sin ce the wastes usu-
ally contain high level s of hazardous material s, such as microorganisms, sacs,
DBPs, and heavy metal s, the wastes must now be tfeated (Il occ ul ati o n, sediment a-
ti o n, and di sinfec ti o n) prior to recycling. T he botto m line is that the qu ality of the
recycled water mu st be beHcr than or as good as the raw wate r e ntering the pla nt.
However, the Water Po llutio n Contro l Ac t al so all ows the treated recoverable wastes
to be discharged to a nearby water course as lo ng as a perm it is acqui red from the
regional adm inistrator.
Handling of no nrecoverable wastes is a more compli cated iss ue. Several alter-
natives ex ist. The simplest is di scharge into the sewer system. However, thi s method
is often im prac ti cal because of the ex pensive fines levi ed by was tewater departme nt s,
particul arl y when they are not located in the same ci ty as the treatm en t plant. The
fund ame nt al methods of processing and handlin g no nrecoverable wastes are COIl-
centrati o n of the was tes. treatm ent and di sposal of separated liquids. di sposal of con-
ce ntrated materia ls, a nd the recovery o f by- produ cts from the wastes. By- products
th atlll ay be recovered are aluminum sulfate and calciulll ox ide; the wastes also yie ld
inert co mpound s that can be used to produ ce bricks and aggregates for road con-
struc ti o n. The mos t co mmo n handlin g and process ing practi ces are gravity thi cken-
ing, phys ical and chemical separati o n, heat treatment , and other ty pes of chemi cal
eng inee ring processes.
Generall y. o nly a limited number of altern ati ves exist for the di sposal o f the fi -
nal residual wastes. Thus the e ng ineer mu st carefull y research the ava il ability of ap-
propriate di sposal si tes, requirements for di sposal , methods of tra nsporting the fin al
res idue, and the costs associated with fin al di sposal. The pre liminary en g ineerin g
study mu st evaluate all feasible altern atives for the treatm ent , ha ndlin g, and ultimate
di sposal of all waste produced by the water treatment plant. Furtherm ore , the sludy
mu st prese nt a rea listi c reco mmend ation based o n cost co nsiderati o ns and the ef-
fec tiveness of the proposed waste processin g and handlin g sc he me .
2.4.11 Instrumentation and Control Systems
The gene ral moti ves for installing a sophisticated ins trume ntati o n and co nt rol sys-
tem for a treatme nt plant and water supply system arc ( I ) the co ntinuou s produc-
ti o n and suppl y o f safe drinkin g water, (2 ) the auto matic exec uti o n of correc tive
measures and auto matic respo nse to potenti all y d isastrous si tuati o ns, (3) minim iz-
ing the pote nti a l for hum an e rror, (4) the capability to quic kl y so lve anal yti c prob·
lems, and (5) the ability to diag nose problems in rem otely localed equipm ent be-
fore a malfun ction occurs.
Because of cost cons traint s, managers of all utiliti es, includ ing those in the wa-
ter suppl y busin ess, are seekin g means to cut both the opera ti o n and mainten ance
costs; the redu cti o n of personnel is usually their target. Moreove r, managers of wa-
ter prod ucti on ano s uppl y facilities are troubled by the new strin gent water qua lity
standard s set by the E PA, includi ng the requi remen t for freq ue nt or poss ibl y co n-
tinuous mo nitorin g of water qu ality for certa in parameters.
PRELIMINARY ENGINEER ING STUDY 51
Modern wa ter treatment a nd supply co nt ro l sys tems are co mm only re ferred to as
Supervisory Cont ro l and Dala Acqui sition (SCADA) systems. Allhough the SC ADA
systems have powerful features, it is important to be certa in that these features are
Irul y bcne li c ial to the o perati on and mainte nance of the fac ilities. SC ADA sys lems
that are poorly des ig ned or those w ith no previo us applicatio ns, such as a new ly de-
sig ned syste m, can potenti ally cause numerou s problems and the ex penditure of un -
necessary amo unts of mo ney. E ng ineers sho uld not be swayed by the promoti o nal
sales pitch to acquire state-of- the-a rt equipm ent because these system s may become
obsolele within a peri od of 10 years, Ihus making replacemen t parts hard to find or
even un avail able.
An other consideration is the compl ex ity of the water treatment and s upply sys-
tems. In cont rast to most industri al fac ilities, the water sys tem is re lati vely simple
and the qu alificali ons to become a plant o perator arc less s tringent. In lig ht of thi s
fac t, Ihe rea l need for a high-l evel SCADA sys te m is o fle n questionable; the o nl y
appropriate situati o n would be if the facility is very large a nd has a n exorbitant
amo unt of data that must be processed qui ckly. Alth ough the new computer-
controll ed SC ADA lec hn o logy is exci tin g, project e ngi neers must avo id the potc n-
tial pitfalls.
Fo ur bas ic types of pla nt instrument ation and contro l syste ms are currently avail-
ab le: manu al, se miauto mati c, auto matic, and superv isory. T heir res pecti ve de lini-
tions are as fo ll ows: manua l init ia ti on of a fun cti o n by th e operator; manua l initia-
ti on of a n aut omatic fun c tio n; the use of sensors, limit sw itc hes, timers, analyti cal
instruments, controll er, and co nt rol logic dev ices such as relays or programm able
cont rollers; and remo te co nt rol over plant operati ons whil e on the plant si te or fro l11
locati ons away from the plant site.
The preliminary en g ineerin g study s hould evaluate all four level s of the instru -
mentati o n and control syste m based o n the local conditi o ns, size and compl exity of
the treatment and water suppl y systems, the manage ment phil osophy of the ow ner,
and the anticipated fina nc ia l co ns tra ints. The fin al recommendati on should be so me-
what based o n owner pre ference. However, fin al selec ti on must al so be based o n
sound eng ineerin g judg ment.
2.4.12 Preliminary Cost Estimates
The best system amo ng the vari o us water trea tme nt altern atives may be selec ted on
the bas is of cos t-effec tive co nstruction and the costs assoc iated with pla nt main te-
nance and operati on. The pre liminary cost estimate o f the treatment pl ant is al so
necessary to assess the capital th at mus t be ra ised to construc t the plant.
The construc tio n cost estim ate is compri sed of all cos ts assoc iated with the con-
structi o n contrac t, the overhead and profit of the cont rac tor, but excludin g the e n-
gineerin g a nd legal expe nses. Thi s es timate s hould also ass ume co mpetitive bidding.
A cO lllmo n me thod in formul atin g prel iminary cost es tim ates is to use the cost
es tim ati o n curves develo ped by the E PA (Estimatillg Water Treatment Costs,
EPN600/2-79- 162 b, Augus t 1979). O th er methods e mpl oy constructi o n cost esti -
mation da "., s uch as Fig ure 2.4. 12-1 , supplied by re putab le co nsulting eng ineering
 PRELIMINARY ENGINEER ING STUDY 53
52 PRELIMINARY STUDIES

ness of a proposed treatment system for a community is to be evaluated, the engi-

'" ne~r.~ ust also ~erfonn a value analysis and appraise the Irealmen t plant and rel ated
/' faclhlles (e.g., Intake, low ser.vice pumping facility, and hi g h service pumping sta-
• • lio ns). Furthermore, the locatio n of the water distribution reservoirs and the wale

,. ./
V distribution pipe system must be analyzed . r
The ope~a ti on an.d maintena nce cos ts o f a water treatmen t plant typically include
costs associated ~lIh (I) labor, supervision, and ad mini stration ; (2) chemicals;
"" r7 (3) power; (4) mainte na nce and repair; a nd (5) mi scell aneous supp li es and serv ices.
~ "
&0 Add it IOnal factors th at a ff~ct ac tual operati o n and maintenance costs are the policy
L.-
8 50 of the owner, the complexlly of the sys te m, and the local environment and weather
C
~ ~o These costs ris~ o n an a~nual basis due to co ntinuing innationary trends in labor:
i V po~er, and. eq Ulpme?t. Figure 2.4.12-2 is a cost estimation c urve for expenses as-
.S )0
• V sOC iated with operatIOns a nd mainten ance. T he c urves arc established o n data ob-
~ / ta~ned from a b~sic. conventional water treat men t plant processing raw water of
o •
V fairly good quality with no sig nificant bui ldup of scale in the pumping facilit ies. It
ENR Index · 6,500

"•
/
•
IS l111po~tan~ to remember that these cost estim ates are hi gh ly site specific and are
only g Uldelllles.

V
8 ~ , I 9 10

Capacity (MOD)
2.4.13 Recommended Treatment Processes
After the fi rst II steps of the preliminary engineering study have been completed
and thoro ug hl y evalu ated, the project engineer is ready to make a recommendatio n

Figure 2.4.12. 1 Construction cost ror conventi onal walcr treatment pl ant. ••
,
F • -
.!!
;8 •• -
./
firms. T he cost figure obtained from th e curves may be adjusted to a geographical o< "
area in the United States and to current standard s through the applicati on of a spe- ::
i 11
V
cial cost index . The two best recognized cos t indexes arc the Engineering News-
Record Construction Ind ex and the Handy- Whitman In dex of Water Ut ilit y
:§.
& ,. /'
Construction Costs. Figure 2.4. 12- 1 is based on the ENR Index of 6500. This index
was created in 193 1 and developed for 20 different c ities in the Un ited Stales and i
c
VV
two Canadia n cities, each representing a local geog raph ic area. This index is des- ~ I:
ignated as 100 and appears in the Engineering News-Record, a magazine published i
1!
two to four tim es a year by McGraw-Hili Book Compa ny. •& IS ./
It is important that engi nee rs real ize that preliminary estimates can not account ~ •.,
for all factors unique to the plant site. The preliminary estim ates shou ld therefore
be considered as a budget est imate with an expected acc uracy of approximately
& 0.1
./
+ 30% to - 15%. Cost estimates prepared for the purpose of compa rin g two or th ree ~ ./ ENR Index·6,500
C U
alternat ive systems must take into aCCOllnt the useful life, salvage va lue, and annual ~ V
/'
reven ue of the system, in addition to the initi al constructi on costs. This ca n be
achieved e ither by convert in g all costs to equivalen t, uniform annu al costs and in - .., ,
come or by convert ing all costs and revenues to presen t worth at time zero. Presellt • , • J • 11110 411'11100

PI.nt Capacity (MGD)

worth is defined as the yield att ai ned at a later date of money inves ted at time zero
X the required costs and revenues at a spec ified interest rate. If the cost-efrec li ve-
Figure 2.4.12-2 Plant operation and maintenance cost of conventi onal pJant.

, 7

54 PRELIMI NARY STUDIES
to the ow ner in regard to the most appropriate type of water treatment system . The
steps di scussed up to this point may be g rouped into four basic procedures : data
collecti o n and proble m formul ati on; sy nthesis and analysis of various subsystem s
find sys tem alternatives; valu e analysis; and appra isal of the alternatives chosen. The
recomme nded treatme nt syste m must be effecti ve, reliable, simple in both desig n
and operati on, cos t-effec tive, and a proven type of sys tem with minimal operati o n
and mainte nance costs.
2.4.14 Plant Layout and Architectural Design
The estahlishmcnt of the plant layout is one of the fin al steps o f the preliminary de-
sig n stud y. Once the altern ative treatment processes are narrowed dow n to one or
two, efforts to desig n the plant layout for the proposed site or sites should beg in.
The e ngineer must attempt to obtain input from qualified architects for aesthetic rea-
so ns, from superintend ent s o f local waterworks for their o perational preferences,
and from the ow ner for hi s o r her partic ul ar pre fere nces o n thi s iss ue. The usc of a
comput er-aided design and draftin g sys tem (CADD) is the mos t effec tive method
for conducting thi s task.
Although the layout is greatly innue nced by the ge neral site topography and the
ex tent o f future plant ex pansions, there are three basic plant layouts: cluster, satel-
lite or college campu s, and an intermedi ary form of the cluster and satellite plant
layouts. Figures of the different plant layout s are show n in Section 3.2.2. Regardless
of the layo ut pattern , the water should now thro ughout the en tire process tra in by
mea ns of g ravit y unl ess pumping betwee n the unit processes is absolutely neces-
sary.
Fro m an engineering standpoint , there are eight important characteristics that
must be considered durin g work on the layout. The tirst is the minimi zati o n of costs
associated with the c ivil wo rk s. However, the des ign engineer may not delete the
careful stud y of the site conditions, natu ral water course, c ut -and-fill requireme nts,
access roads to the site, the slope of these road s, and the drainage and protecti o n of
the proposed site against fl ooding.
The second item to be co nsidered is the ease of constru cti on. Thi s item addresses
iss ues such as providing a simple arran ge me nt o f water treatment unit s, providing
o nly the minimum number of req uired units, allowing easy access to and between
th e units for co nstructi o n equipment, avoiding a yard piping design that is too deep
or that cross over eac h other, and selec tio n of simple and uncomplicated individual
structures,
Automatic, equ al hydrauli c loading to each ta nk or filter is the third considera-
ti o n. Thi s goal ca n o nly be attained thro ugh proper layout of th e unit s. Over 50%
of all ex istin g water treatme nt plants have so me degree of imbalance in their hy-
draulic loadin gs from o ne treatme nt unit to another, primarily as the result of im-
proper desig n of th e layout.
The fourth item is the ce ntrali zation of con trol and operati on. Veteran plant op-
erators unanimously prefer thi s type of layou t because it allows for easier and mo re
effec ti ve plant supervision and control. Experi ence has show n that it is more e ffec-
PRELIM INARY ENG IN EERING STUDY 55
ti ve to arrange subsyste ms requiring close supervision, such as lilters, che mi cal feed-
ers, and the Il occ ulati o n process, in close proximity to th e co ntrol/ad mini strati o n
building rathe r than in a widely scattered pattern .
The lifth consideration is the physical separation o f the majority unit process
structures. Massive co ncrete stru ctures of different depths and weig ht mu st be phys-
ically separated to preve nt the formation o f serious cracks due to un eve n subside nce
and to fac ilitate easy access by constructi o n eq uipment for the installatio n a nd re-
pair of the plant equ ipme nt. Physical separation is very important for plant sites that
have soft grou nd or frequent earthqu ake ac tivity. The maj o rity of plant sites have
sig nifi ca nt difTcren ti al ground subsidence between existing a nd newly constru cted
structures; thi s fac t mu st be give n careful co nsiderati on.
The development of a mas ter plan for th e ultim ate plant layou t and yard piping
sc heme is the sixth co nsideration . E ngineers mu st allocate areas for future process
unit s in order to minimi ze any interference and inconve ni ence to the ex isting treat-
me nt sys tem and daily operational activities. Fro m a safety standpoint, two inde-
pendent mea ns of access and two alternative roads within the site sho uld be pro-
vided to th e plant. This sc he me is benefici al for both future plant ex pansion and the
transport ation o f che micals in large trailer truc ks during norm al pla nt operations.
The seventh significant consideration is the c limat ic co nditions that exist at th e
plant site. For exa mple, trea tme nt plants located in very cold reg io ns must ho use a ll
unit processes to protec t the m from freezin g. Co nseq uently, the layo ut mu st be a
compact sc heme that avo ids large separa ti o ns between the major process units to
minimi ze the size of the hOll sing structures. If the unit processes mu st be se parated,
underground passage tunnel s should be co nsidered ; it there fo re foll ows that the
process units mu st be arranged in a way that minimi zes tunn el construction costs.
The e ig hth and last, but not least, consideration is architectural dcsig n. The aes-
thetic fea tures o f the plan t layout and building a nd the landscaping o f the plant site
are superfi cial but very important iss ues. Most water treatment plants in the United
States arc public ly owned. They arc frequently vis ited by public groups, sc hool·
children , and engi neers from other muni c ipalities, states, or countries. Furthermore,
the plant sites are often near res idential areas or areas of natural preservation. Thus
the proposed treat men t plant may meet resistance from groups of local c iti zens or
ot her orga ni zat io ns. The pl ant mu st therefo re be functional a nd safe and must ul so
blend into the surrounding env ironmen t, as well as be aesthet ically pleasing wh en
viewed fro m th e s urrounding nei g hborh oods. It is very important to co nsider spend-
ing at least 2 to 3% o f the to tal des ig n cos t o n impro vin g the aesthetics of th e plant.
2.4.15 Management Planning of the Design Phase
For the desig n of a water treatment system to be exec uted e ffecti ve ly and success-
fully, a balanced tea m of hi g hl y qualified spec iali sts mu st be asse mbl ed. Pre liminary
selec ti o n o f expe rts ,md management planning o f the project mu st be accomplished
at th e end of the preliminary engi neering study phase. O ne tool widely used in proj -
ect management is the bar c hart. Howeve r, th e Prog ram Evaluation and Review
Techniqu e (PERT). th e Integrated Sys tem of Proj ec t Ma nage me nt (ISPM ), and th e
56 PRELIMINARY STUDIES
".. Fob Ma. Apr May
1999
Juno July Aug Sop
Figure 2.4.15-1 Example of project control schedule.
Critical Path Method (CPM) have also been lIsed in the ma nagement of complex
projects; these methods may also be used in project schedul ing., PERT emphasizes
the co ntrol phase of project manageme nt; CPM delines the durallon of a project and
thereby minimi zes the cost, and ISPM stresses the aspect of project cost con trol.
Figure 2.4. 15- 1 is an example of a project co nt rol schedule.
2.4.16 Environmental Impact Statement
In accorda nce with the National Environmenta l Protec tion Act of 1969, all water
treatm e nt and s upply projects arc req uired to lile an e nvironment al impact statement
(EIS) prior to impleme ntation . The EIS mu st include detailed studi es and an analy -
sis o f the env ironmen tal impact of the faci lity. A nonbuild altern ative IS al so re-
quired. Preparation of an EIS requires a team of many specialists (a biologist, a ~lY­
drologist, arc heologists, and economists) to both devel op and complete the studI es.
Projects that improve the existing environment or those with a negative impact statc-
ment, indicat ing no e nvironme nt al impact by the proposed project, can proceed to
the design stage rather quickl y. However, under certai n circ um stanccs, an EIS may
require a significant amou nt of tim e and money. The findings of the EIS may also
have a tremendous impact on the design, co nstructi on schedule, and total cost of
the project.
2.5 WATERSHED MANAGEMENT
The Surface Treatment Rule not only demand s the e nhanceme nt of drinking water
quality through improvements in water treat me nt processes and monitoring of treated
water quality, but also emphasizes the protec ti on of water so urces. Municipalities
WATERSHED MANAGEMENT 57
supplying unliltered water with di sinfec ti on need onl y demonstrate a fail-safe wa-
tershed manageme nt program in order to omi t filtration. The Delaware and Catskill
water supply systems of New York C ity are prime examples of this scenario. For
water purveyors that face the prospec t of adding new or upgrading existin g treat-
men t processes, a watershed protection program can play an importa nt role in main-
tai ning or improving raw water quality. Additionally, a robust watershed manage-
ment is a very cost-effective option, as it may circ umvent the need to cons truct
ex pensive advanced water treatme nt processes.
Although surface water is primarily considered to be the main source of water,
underground water, including th e infilt rati on gallery, should be evaluated as an al-
tern ative source. The geographic conditi ons affec tin g the underground water, a nd
possible contamination by huma n was te as well as livestock-fanning, agri cultural
was tes, and indust ry wastes, are important issues. Additional issues that must be
evalu ated are potential gro und subside nce and seawater intrusion into underground
aq uifers along the coastal areas. Both phe nomena are due to the overd ra fting of
gro und water.
It is unfortunate that most wa ter sources in the world have become polluted at
an acce lera ted rate due to popUlation explosion and the rapid development of vari-
ous industries without rega rd for adeq uate waste cont rol. In the United States and
Canada, a few lakes a nd reservoirs (sources) have been well protec ted. However,
urba n development and other land-usc ac ti vities now threaten these watersheds.
T he most effec tive way to ens ure the long- term protection of water Sources is
land ow nership by water purveyors and cooperat ive public jurisdictions. However,
the cost of purchasing valuable forests and highly productive agricultu ral land s sur-
rounding the watershed is ofte n too expensive for this to occur. This is especially
tru e for sma ller systems suffering fin ancia l proble ms du e to ri sing operational costs
and th e need to comply with new regulation s. Furthermore, the political pressures
to develop the land may be too strong to overcome. Conseq ue ntly, only a few wa-
ter purveyors own or co ntrol most or a ll of the land within their water supply wa-
tcrsheds. The most practical mea ns of meeting EPA and state requirements for hi gh-
quality drinking water is a combinati on o f wa ters hed control, reservo ir manageme nt ,
and optimi zati on of water treatment processes.
In ge neral , there are four basic measures in ac hieving wa ters hed control : general
cont rol measures, ag ricultural land con trol , forest management, a nd control of ur-
ban deve lopme nt. General contlVls are app li cable to most wa tersheds. These mea-
Sures co nt rol la nd use by restricting th e use of reservoirs and river and reservoir
buffers; in addit ion to land acquisition, trespass con trol, watershed inspection pro-
grams, plan review, public ed ucation and invol ve ment programs, and written agree-
ments with landowners.
Agricultural /and control is the second basic measure in watershed control. Soil
conservat ion practices encompass various practices, rangin g from soil conservation
of cu ltivated cropland , to conservation structures Stich as farm ponds a nd gully co n-
tro l, to animal waste control and management.
The third basic measure, forest management, ca n be difficult to manage since
logg ing activ ities are generally di spersed within the wa tershed. Nonetheless, the
58 PRELIMINARY STUDIES

managemen t plan should include buffer strips, skid trail s, and landin gs. Most at-
tenti on durin g the planning, constructi on, an d monitorin g stages should be directed
to the roads and skid trails since they cause the Illost signilicant ground erosion o r
all logging aCl ivilies.
CHAPTER 3
The last basic meas ure of wate rshed con trol is Ilrball developmen t. Strict en-
forcement of state public health codes and local ordinances in regard 10 locati on,
design. installati on. and maintenance of septic systems in urbani zed watersheds is Design of Basic Treatment
required. Local reg ul ati ons should establi sh zoning ordinances thaI restri ct the type,
location, and intensity of development . In add ition, polluti on control stru ctures sti ch Process Units
as det enti on bas ins, infiltration systems, and storm water diversions designed to di -
vert or reduce the fl ow of contaminants to the watershed should be provided.
Watershed protection plans are site specilic. The success of th e plans rely heav-
il y on the support from elected ofli cials, cooperat ing agencies, special interest
groups, and public awareness. 3.1 COMMON BASIC WATER TREATMENT PROCESSES

3.1.1 Design Procedure

BIBLIOGRAPHY
Efficient and successful design work can only be accompli shed throu gh strict ad-
herence 10 Ihe sland ard procedures of Ihe pro;eci as eSlabll'shed by Ih e orgalllzation.
AWWA , Water Treatmen t Plmll Desigll, American Water Work s Assoc iati on. New York . 1969.
The key aC l/vUles are as follows:
AWWA and ASCE, Water Treatm ellt Plant Design. 3rd cd .. McG raw- Hill. New York. 1997.
Carney. M., "European Drinking Water Siand ards," 1. AWWA. 81:6:48. (June 199 1).
Cook, G. D., and Carlson, R. E., Reselvoir Management for Water Qualit)' and THM Precursor
COlltm!, AWWA Research Foundation Report , America n Water Works Assoc iati on. De nver.
CO, 1989.
Dani el, F. 8., el aI. , "Toxicological Studi es o n MX , a Di sin fection By-product." 1. AWWA ,
86:3: 101 (March 1994).
EPA , "Nati onal Primary Drinking Water Regulati ons: Interim Enh anced Surface Water
Treatment : Fina l Rul e," Federal Reg islel; 40 CFR Parts 9. 141 and 9. 142 (Decembe r 16,
1998).
Fl ynn , K. c., and William s, T. . " Watershed Manage ment ," £E&T, pp. 36-40 (J uly 1994).
George, C. E ., "Effec ti ve Project Sc hedul ing," Water SelVage Works, p. 386 (December 1971).
Goll nlz, W. D" "Source Proleclion and Ihe Small Ulil ily," 1. AIVIVA, 80:6:52 (June 1988).
G rant. E. L.. Prin ciples of Engineering Ecollomy, 3rd cd .. Ro nald Press. New York. 1950.
Kawamura. S .. Integrated Desigll of Water Treatment Facilities, Wi ley. New York , 199 1.
Kawamura. S .• " Pil ot Stud ies o r Mi xi ng and Settling." AWWA Semillar Proceedillgs: Design
of Pilot Plallt Studies, American Water Works Assoc iation. Denver. CO, 1982. p. 25.
Merrill. F. S .• ed .• Standard Handbook/or Civil Engineers, 3rd ed .. McGraw-Hi ll. New York .
1983.
Montgomery. J. M .• Consulting Engineers. Wal er Treatmell t: Prillciples and Desigll. Wi ley.
New York, 1985.
Olson. R. V" "Crilical Palh Melhod of Work Schedulin g." 1. AIVWA . 6 1:9:447 (Scplelllber
1969).
Pontius, F. W., "New Hori zons in Federal Regulati on," 1. AWWA, 90:3:38 (Marc h 1998).
Robbins. R. W.• et aI. , "Erfecti ve Wate rshed Management ror Surrace Water Supp lies,"
1. AWWA, 81: 12:34 (December 199 1).
..
. • • J "
I. Assign job numbers.
2. ESlablish budge I and projeci sc hedules (see Table 1.4.2 and Figure 2.4 .15- 1).
3. Lisl all required drawin gs.
4. Orga ni ze a design team .
5. Allocale a budge I for each disc ipline of Ihe projeci learn.
6. Firml y eSlablish Ihe process and process and inslrumenlalion (P& I) diagra ms
dUring Ihe early slage of Ihe projecl.
7. Perform all necessary calculations and establish a reaso nably detailed basic
skeldl for each process unil. These should be circu laled 10 each of Ihe dis-
Ciplines (refer 10 Figure 3. 1.1 - 1).
8. Establish the details of all subordinate plant faciliti es, including the chemi-
cal feed system and thc waste and sludge handling system.
9. Disc uss an.d finali ze items 6 th roug h 8 with the head of each di sciplinc be-
fore rcicaslllg them to thc design teams.
10. Begin design work based on Ihe projeci schedule.
II. Es tablish ti ght coordinati on among the vari ous project disciplines.
12. Maintai n stri ct con trol of the overall project.
I ~. Edit .all speci~:ati~ns writtcn by other members of the team. Use the stan-
dardi zed specdlcatlons of th e finn whenever possible.
14. Carefull y review the design calculations, drawin gs, and spccificati ons with
the help of ex pcrts. Do not sign documcnts until each itcm has been checked
and corrected.

59
60 DES IGN OF BASIC TREATMEN T PROCESS UNITS
COMMON BASIC WATER TREATMENT PROCESSES 61

15. Es tim ate the construc ti o n cost and arrange to have Ihe constru ction doc u- TAIlLE 3.1.1-2 Normal Design Time per Drnwing
ment s produced .
Total Hours/Drawing Breakdow n of
Disc ipline (Design and Dra fling)
The projec t e ngineer will complete the des ig n projec t successfull y if the foll ow- the Hours
in g co nl rol methods are understood and fo ll owed. Project engineering 19 Specifications 6h
Cos t cstimate 3h
I . Have a c lea r und erstanding of th e projec t o bjec ti ves and the project goal. Engineering 10 h
Civil 85
2. Es tablish a good and balan ced des ig n team . Engineering 45 h
3. Have availabl e an adequate budget, reaso nabl e des ig n period, and sufilci ent Drafling 40 h
Architectural 80
perso nn el to complete thc projec t. Engineering 35 h
Drafting 45 h
4. Coo rdinate the various des ig n di sc iplines . Stru ctural 70 Engineerin g 35 h
5. Keep to a minimum c hanges to and dev iati ons fro m the originally es tablished Drafting 35 h
Mechani cal 75
desig n co ncept and sketc hes. Engineering 30 h
6. Foll ow th e project control procedu re stri ctl y. Secti o n 1.4 presents a detailed Dra fting 45 h
Elec lri cal and 60
exa mpl e of thi s ite m. Enginee ring 20 h
instrulllentat ion
Drafting 40 h
7. Both the project e ngineer and projec t ma nager mu st be kno wledgeable, ex- Landscaping 40 Enginee ring 15 h
peri e nced, willing to devote their time to the projec t, a nd bring a pos iti ve at- Drafting 25 h
titude to the des ign team .
Nol l'S: ( I) The ~ngineerjng li~le for all disciplines include s the hours required 10 check the design cal-
c ulallo~s and draWings and the li llie required to coordinate e fforls with other disc iplines o f
Tab les 3. 1.1 - 1 thro ug h 3.3. 1-3 and Figure 3. 1.1 - 1 may aid inexpe ri enced engi- the project.
neers to grasp so me of the bas ic des ig n requireme nt s and Illay ultimatel y help cre- (2) If ~ g~ de~ign emcien~y is to be ma intained. the wei ghted average hours per drawing. for
the englllcc nng and drafllllg phase. arc 80 h.
ate rea listi c projec t co ntrol sc he mes. Table 3. 1. 1- 1 prov ides an es timate of the tOlal
(3) The cost of one s heet of drawin g ranges fro m $1000 to $7000 ( 1989 ') 1 I'
number of drawings required for cert ain -s ized pl ant s usin g a conventional complete . ,. . . ". cos ( epcl1( lIIg o n the
natu re of the ~rawlllg. the reqUired deSign li me. the complexily of the drawing. and the nUIll-
ber o f co rrectIOn s [hat llluSt be Illade.
TABLE 3.1.1-1 Anticipated Number of Ormvings for n Water Trentment Plant Design
Percent age Number of Drawings if tre~ trn ent process. Tab le 3. 1.1 -2 lists th e normal des ig n tim e and work-ho urs rc-
Discipline of Total Total Is 200 Drawin gs qUlred t? complete a sheet o f dra_wings for each o f the va ri o us di sciplines. Table
3. 1. 1-3 IS a general breakdown of the co nstru c ti on costs. Fig ure 1 I I _I . . •
General 5 10 I fb ' . _. . IS .111 ex-
Civil 10 20 ami: e 0 aS lc de.s lg n sketches for a unit process; thi s prov ides a basis for deta iled
Architec tu ra l 8 16 deSIg n by eac h 01 the di sc iplines of the desig n lea m.
St ructural 30 60
Mechanical 20 40
Electrical and 25 50 3.2 BAS IC A PPROACH
instrulllcntati on
Land scaping 2 4 A .wa.tef Ircatn~e.nl plant sho~lld be des ig ned to produce a continu o us suppl y o f sa fe
N(}/l'.~ : ( I) The total number of draw in gs may vary dependi ng on the cond itions o f the pl ant site. local d~l.nkll1g walel legardl ess 01 the ra w wa ter charac teri sti cs and enviro nm ental co n-
weather conditions. the compl ex ity of the design. and the pl ant sizc. dit IO ns. Thus,. the ~pplicali on of space-age tec hn Ology in the process des ig n and sub-
(2) Small co nventional plants. those processing less than 10 mgd. may havc 70 10 120 lotal draw- system selec tIon IS no t alway s th e best approac h. The ultim ate plant des ign has a
ings depe nding on the degree 10 whi ch the propri etary treatment unit s arc specified. ~y~ tell~.~I.l (.1 s ubsy~ t e lll I.hat aJ..e proven to be simple, e ffective, reliab le, durable , and
C') Medium sized conve ntional plants ( 10 to 50 mgd) require anywhere from 120 10 200 draw-
in gs. but generally over 150.
Dos t~e llec t l :e. ~hl S baSIC phl l osoph~ is a~) pli c abl e to all c ircum stances and cases.
(4) La rge conventional plants (over 50 IIIgd) req uire 200 to 350 s heets of drawings . eSlg n englneels should study the 1011 0WIIl g bas ic rul es prior to des ign work :
(5) If an intake pumping station. hi gh-service pum pi ng stati on. or ozonation system is included
in the plant design. the total number of drawin gs will be in creased approximately by 7%. 7%. l. Eva lu ate local conditi o ns care full y.
and 9%. respectively.
2. C rea te a reaso nabl y conse rva ti ve des ig n th at is cost-effec ti ve to construct.
62 DESIG N OF BASIC TREATMENT PROCESS UN ITS BASIC APPROACH 63

TABLE 3.1.1-3 Brcukdowll or Wal er Treatment Plant Construction Costs ,.

(Approximate) .L
Civil work 7.0% T
(earthwork. grad ing, pavi ng. fenci ng)
Yard pipings 8.0%
Landscap ing and irri gation 1.0%
O perations building 10.0%
(chemi cal feed system included)
Flocculation and sed imentation basins 17 .0%
Filters 20.0%
Clcarwel l 8.0%
Pumping faciliti es 7.0%
Meter vaults (L. S.) 2.0%
Filter washwaste holdi ng and recycling 3.0% .,. o·
Sludge drying beds 2.0%
-=,,,,.,_.,
8--4-1 ".
Mi scell aneous items 0.3 % ,., III
r.;.,
Chemical storage faci lities 1.0% (I!O
" T
r-"ff
I ,-
.
2' I
~
12.0% ~
Electri cal and instrumentation works
Testing and di sinfecting works 0.2%
r
.
'.-' j W9 ~
•
C~lllflllullel _ _.- .
;
'2 '
~

Move on and move off

(contractors)
1.5%
-= Ii= IlL JllJITJ[] I II I Ill . t.!
No/n: ( I) The table docs not include the overhead and prolit of the contractor; Ihese are generall y 20% Figure 3.1.1-1 Dual-mcd ia bed alternativc (16 fill crs) cqualloading. constant-level control
of the total cost shown above. filters (su rface wash by watcr jcts).
(2) The fi gures above are based on a high-rate conventional process_

3. Create a simple, reliable, effec ti ve, and prove n syste m by appl yin g the best
know ledge and skill to the desig n.
4. Design a plant that is easy and safe to construc t as we ll as simple and safe
to operate_
5. Allow for max imum operati o nal fl ex ibility and minim al operati ons and mai n-
tenance costs.
6. Desig n a plant that is aesthetically pleas ing and does not have a negative im-
pact o n the enviro nment.
7. Acquire a plant site that adequ ately salislies the basic criteri a and is free of
all potential di sasters.
8. Do not be e mbarrassed to obtain help from qualified associates, consult ant s,
or spec ialists; an engineer should o nly perform services th at are in his or her
area of compete nce.
9. Be certain th at the project meets all pertine nt legal requirements and e ngi-
neeri ng sta ndards.
10. Respect the wishes of the ow ner.
A water treatme nt process is an assemblage of unit processes th at e ffec ti ve ly pro-
duces a n abundant supply of safe drinkin g water. The treatment system generally
co nsists of four or more interre lated and compatible compo nen ts. Si nce the unit
processes must be interrel ated , the operation o f each component affects the perfor-
mance of o thers, that is, the entire system. Most water treatme nt tex tbooks and de-
sig n g uide manuals present ly on the market tend to view each co mponent as an in-
dividual unit. ConsequelHly, readers may erroneously believe that the most efficient
overall plant performance is obtained by optimi zing the design of the individual
compo ne nt s. This type of suboptimi zati o n gene ra ll y does not achi eve optimum to-
tal plant des ign. Since th e pivo tal purificatio n process is filtration , all pretreatment
unit processes sho uld be designed and operated to max imi ze the e ffic iency of the
filt rat ion process and the subseque nt process of di sinfectio n.
As mentio ned previous ly, the desig n engineer sho uld o nl y work in hi s o r her area
of competence. The des ign team is usually composed of civil , structural , mec hani -
cal, electrical, and instrumentation e ng ineers and e ngineers spec ializing in comput -
ers, soil fou ndation, and hydraul ics, as we ll as an arc hitect. T he team may also in-
clude a chemical engineer if necessary. It is very important that bo th the project
engineer and projec t manager be famili ar with the outlines of work performed by
each of the disciplines since they are respo nsible for coord in ating the progress of
the vari ous sections o n a dail y basis (ideall y).
Desig n e ng ineers mu st reme mber to utili ze commercially available components
whenever feasible because thi s helps to keep down projec t costs. Components such
as gates, va lves, pumps, pipes, and heating and ventilation equ ipment, in addition
to certain equipment such as flow meters, c hemical feeders, mi xers, sludge collec-
ti on units, and instrume ntati o n and con trol units, shoul d be eval uated for substitu-
ti on by commerc ia lly avai lable unit s. T he key to creati ng a good water treatme nt
64 DESIGN OF BASIC TREATMENT PROCESS UNITS
plant desig n is to pay attention to important detail s and avoid s pendin g time 011
sig nifi ca nt items.
3.2.1 Introduction
The phrase commoll water treatment processes con notes a system Ih~1 .c~m~ines t.he
processes of coagulation, fl occul ati o n, sedimentation, filtrati o n, a ~d dl s ln~ ec t lO n \~ lIh
the necessary process control a nd instrum entation measures. Un ll.1 the 11lId-~":entl et.h
cent ury the desig n of water treatment plants was han dl~d exc l ~lslvel y by civil eng i-
neers, and emphasis was placed o n the c ivil , hydrauli c, lounda tl o n, and st ru ctu ra l as-
pects of des ign. It is therefore unde rstandable why these ea rl y treatmen t pl ants tend
10 lack chemical e ngineering, mi cro biolog ical, electrical , mechanical, and arc illtcc-
tural considerations. Conseq ue ntl y, the perfo rm ances of these plan ts do not always
meet the expec tations of the des ig n a nd are not always aestheticall y pleas ing. T he
prese nt tre nd is to have e nvironmental engi neers design the treat ment pro ~css~s and
the overall plant design because they are familiar w ith a broader ra nge 01 ~ ub.l .ec t s.
The e ffi c ient and s uccessful execu ti on of a modern wate r treatmcnt deSig n IS dc-
pendent on thc co ncerted e fforts of var iolls s pecialists: T he tea m Ill.embe rs. mu st.
therefore pool their knowledge and skill s so th at the pcrlorll1an ~e and In~eractlOI1 ~ I
eac h component are designed in a co mplementary mann er. -:""S
type 01 ~O l11mllnl­
catio n is essen ti al because the desig n of each compo ne nt allec ts the deSig n o f the
e ntire treatme nt system. . .
The cost-effectiveness of the desig n phase Illay be improved by estabhshlll g a
value engi neerin g team , in add iti on to the design team. The projec t members may.
also opt to include both construct ion experts and experienced plan t opera tors (of,
plants similar to th e plant being desig ned). Figure 3.2. 1- 1 prese nt s the me mbers 01
a typ ical desig n team.
3.2.2 General Considerations
When design ing basic treatment process units, the most important considerations arc
the fundamental design philosophy, the design procedures, and co ntrol over the d~­
sig n phase of the project; these issues are discussed in Section 3. 1. However, ce rl 31 n
r-----------
I
Own'f
h dl.n ic.1
Projtd m.MI" Id"i_,
_min"
Figure 3.2. 1-1 Typical water treatment system design team.
BAStC APPROACH 65
less items mu st be es tab lished during the early stage of any design project: plant layout ,
process fl ow diagrams, hydra uli c profile across the plant, process trains, and the is-
sue o f modifying and/or expa nding the existing plant. The project engi neer and man-
ager are responsible for establishing these items based o n so me basic rules and con-
siderations. A few of the more importa nt general desig n co ncepts are listed below.
I. Create a master plan of the treatment plant site, includin g future process unit s
and major yard piping.
2. Usc the mod ul e expansio n concept if the plant is to be expanded in stages. To
avo id any operational confusion, eac h mod ul e sho uld have th e sa me size and
e levat ion and ideally have identical processes.
.1. Provide adequate access roads to each unit and mod ule of the process. This will
facilitate simple plant maintenance and allow easy installati on of future mod ul es.
4. Special attenti on mu st be g ive n to the hydraulics of th e plant. For example,
provide equa l hydraulic load ing to eac h tan k and every process unit a nd main-
tain a uniform backwas h flow distribution to each filter.
5. Study the plant des ig n to ensure that the safety of both plant operatio n a nd
the env ironm ent have been addressed.
6. Resea rch the flood level of the pla nt site, ca re full y plan the access roads to
the plant s ite, and establish the vari o us utility hookups.
7. Establish an adequ ate area for the plant site by calculating the area required
by th e sludge handlin g fac ilities and the area required for the future in stall a-
ti o n of new facilities, sllch as an ozonati on system, GAC adsorptio n bed, a nd
re lated facilities.
Plant Layout The plant layou t is dictated primarily by the available area and
topography of the plant site. Thus, poor site selection severely restricts the choice
of alternati ve layouts. The des ign phase should o nl y refine the preliminary layou t
that was establi shed in the predesign stage. The design phase focuses on the ant ic-
ipated traffic fl ow a nd operator movement, on establishin g hydrau li c balance among
the process unit s and basi ns, o n the aesthetics of the plant, and on th e locat io n o f
the main con trol building; the main co ntrol building shou ld be situated as near as
possible to the pretreatment process, filters, and c hem ical feed system.
Two impo rtant considerati ons mllst be addressed whe n work in g o n the plan t lay-
out. The first is to provide a single chemical application point for a ll pretreatment
process unit s. The second is to provide a single chemi cal feed point to th e filter in-
flue nt for all the filters and use thi s scheme for the combin ed filter wa ter prior to
the clearwell. Although spec ial atte nti on must be given to the yard piping a nd lay-
out of the process units, the sing le-point application scheme is recommend ed be-
calise it si mplifi es the chemical feed system sig nifi ca ntly and provides uniform wa-
ter quality co ntrol. If the chemicals are divided to more th an two points, th e desig n
will require addi ti onal feeders or flow-splitting devices such as rotall1cters, and the
feed rate to eac h proccss unit mu st also be paced to the actu al process n ow rates.
66 DESIGN OF BAS IC TREATMENT PROCESS UN ITS
CLIUWUL

T here are two bas ic plant layo ut s, the cam pus type (Figure 3.2.2- 1) and the c lus-
ter type (Figure 3.2.2-2). In ge neral, the campus type layout has a few adva nt ages
over the cluste r layout: ( I) constructi on work is easier beca use eac h treatment process
st ru cture is read il y access ible; (2) there is no structural damage between di fferen t
unit processes of differing depth and weight duc to uneve n ground settl emcnt ;
(3) new, advan ced treatme nt process unit s are easil y insta lled betwee n ex isting flOC. TANKI
processes if requ ired; (4) civil wo rk may poten tiall y requi re less cut and fill if the
site is not !l at; (5) plant operators are provided with additiona l safe ty measures in
case o f chemi cal spill s and/or ti re; and (6) case hi stories show less structu ral dam-
age du e to ea rthq uakes.
T he ca mpus layout al so has di sadva nt ages: ( 1) it has a larger plant site require-
me nt; (2) it req uires ex tra ya rd pipings, thereby yie ld ing a hi gher loss of head across
the entire process tra in ; (3) plant operators must trave l a grea te r di stance between
eac h process uni t; and (4) co nstru ctio n cos ts may be slightl y grea ter. SCHEME I
Rega rdl ess of the pla nt layo ut sc heme, e ngineers are advi sed to provide two dif-
fe re nt access roads into th e pl ant for safety and tra ffi c considerati ons. T he ma in ac-
flOC. TAHK
cess road shou ld lead pla nt operators and visi tors to the cen tral co nt rol/adm ini stra-
ti on building. The seco nd access road is used primaril y by c he mi cal de li very trucks
lED .
and slud ge haulin g and constru cti on vehi c les. Separa ti on of the differe nt acti vities TANKS
is safc r and aes th eti call y pleasing to the publ ic.
In rece nt years the public has dc manded safe and good drinking water, thu s reg-
ulatory age nc ies arc se tting strin ge nt standards fo r public water suppl y. In li ght of FIL TEAS
these c hanges, treatme nt plan ts need a !lex ible treatm ent process tra in that provides
good re li able treatme nt. To mee t th ese requiremen ts, pl an ts usin g water fro m reser-
vo irs or lakes may co nsider the layo ut presented in Figure 3.2.2-3. This layout has
two independ ent tra ins (re li able treatm cnt) and operati onal nex ibilit y. The plant may
opera te in four di ffere nt treatm en t modes, depe ndi ng on raw water quality: (I) in-
line liltrati on with ozo nati on, (2) direct filtrati on with or withou t ozonati on. (3) reg-
SCHEME 2
ul ar co nve ntional trea tment. and (4) regul ar co nve nt ional process with ozo nati on. T he
additi onal operati onal nex ibilitics of thi s layo ut are ( I) differen t types of coag ul ant
and coag ulant dosages may be used for each process train, (2) eac h process train may
have differe nt hyd rauli c loadi ng, and (3) the two process tra ins may have differen t
modes of treatme nt. A suggested process unit layout is illustrated in Figure 3.2.2-4.

Process Flow Diagram T he process !l ow diagram is the fu ndament al basis for

the desig n work . T his is especiall y tru e for thc design of th e instrum ent ati on and
co nt ro l sys tems a nd the subsys te ms. A bas ic process n ow di agra m should be es-
tablished during the pre liminary design phase and relined at the beginning of the •
n-'-----r'I
tr-'---- ---., I
linal design phase. T he instrum en tation a nd co nt rol speciali st and the mec hani cal '--+t r -:- -- --H t
It-....,--- - ... ,
LL..! ____ JJ IL ____ JI
speciali st may the n begi n the ir des ign work .
T he process n ow di agram should include the foll owin g items: (I) all unit
processes in th e co rrec t sequ ence; (2) all major pi pe co nnec tions with the fl ow di -
rec ti ons; (3) all che mi cals that arc to be lIsed a nd the applicati on point(s) of each; SCHEME 3
(4) a ll major wa ter sa mplin g points th at are nccessary to ma intain qua lity control;
Figure 3.2.2.-1 T hree basic ca mpus pl am layouts.
(5) th e location and size of all major !low meters, va lves, an d co nnec ting pipes in
67
68 DESIGN OF BASIC TR EATMENT PROCESS UNITS BASIC APPROACH 69

c,eaLens

I
'Iow Spilt Willi 801
...w.k Bypl" Giitl

8 .. \" 8WI" Line

Clelr_n.
"• • ~ "• x• \CT.Tln\)

~u: i!l! ."•

<J)
0
"- ~ i!

Figure 3.2.2·3 Process train sc heme for max imum treatment nex ibility.
Chemical &
Control Building
in addition to the elevations of the wa lkway (the top of the structure), the water level,
the bottom elevation, the invert, and possibly the crown levels of all connecting pipes
and channels. Moreover, the fini shed ground level aga inst each structure of the process
SCHEME I unit must be shown to facilitate the work of the remaining di sciplines of the design
team, especially the civil, structural, and mec hanical engineers. It must be emphasized
that this diagram is a key component of the entire design project because it provides
the basic information for the design team. Any significant modifications to this di a-
gram during the midd le or latter stages of the design period will have a negative im-
pact on design efforts, especially with respect to the cost and deadline of the project.

FloclSed Tanks
Chemical
&
Fillers
II Crear
Design of the Basic Processes Thc detai led design work of each process
Conlrol welis
should begin as early in the design phase as possible. The basic design criteria of
6uildlng
eac h process are based on the hydraulic loading, which should have been established
durin g the preliminary desig n phase. Thus the size of each process unit and the type
of major equipment arc loosely established. However, at thi s stage all the details,
including the dimensions and shape of each process, must be firm ly set. Moreover,
all maj or equipment must be selected to ensure that they will fit comfortably into
each unit struc ture. The detailed design phase should therefore begin wi th the chief
SCHEME 2 engineer of the mechanical di sci pline. This is particularly important with respect to
Figure 3.2.2-2 Two basic cluster pl ant layouts. process structures that house numerous pieces of equipment.
The process structures must be reviewed carefully because certain types of unit s
the process train ; (6) the locati on of all major pumps, hlowers, sc reen s, and other require special foundation s, a space around the unit, temperature co ntrol, dust con-
such ite ms of the process train; and (7) the control points for the pressure, water trol, co mp lete isolation, a large watcr and power capacity, or an explosionproof room.
level, Ilow rate, and water quality of the process. All these items are essential for All equipment, particularly the mechani cal a nd e lectrical units, require servicing or
designing the process and instrumentation diagram (P&ID). Detai ls of the P&ID are may need to bc replaced. Consequent ly, adequate spacing must be provided between
di sc ussed in Section 3.2.1 1. Refer to Figure 3.2.2-5 for an example of a water treat- the units. There have been a few cases where the valves in the filter pipe ga llery
me nt plant process diagram . were so inaccessible, due to a complicated piping sc heme, that it was a major un -
dertaking to try to serv ice or replace them.
Preliminary Hydraulic Grade Line Across the Plant Once the process Ilow Design engineers should also design the unit process structures based on the largest
diagram is finali zed, the design engineer should produce a diagram of the hydraulic and heaviest units available on the market because the exact manufacturer of the in-
grade line across the treatment process train. This diagralllmust show all unit processes stalled equipment will not be known unti l after the plant construction bid is awarded to
70 DESIGN OF BASIC TR EATMENT PROCESS UN ITS

No ...... ~ r .......... N
,-:
J
I-----t----,,
,
,,, Ozonalion ,,
Future , .
,12
•
.
~
"

~ Si
~
.: - - - - -f - ---, ;; '5

.. ·•
~ ~
•m m
c c Q 0
,"! ,"! ,, I-- ----r -1------1 ,
,,
N N

,, ,, , " "
c
C C
,, ,, , c
,g ,2 ~ ~ E
g ;;;
,,
,, I , ,I ,,, m U m
c
•E "•§ ,,, F~t~re ,,
,
'","!
~
C Q. •"" '","!
~
C

, ,, ,
i3
• "• ,I
, ,, ,,I ,, <D
en en ,, c c

I , •
,·
,Q .Q r
,•,
;;c ;; - I • •0
E I i! ~
~
t I
I i3
•E 0
§ _, ~8 •
; .,
~.
t
i •
~
en• "•
en
%
• d• E
•00
Flocculalor
(typ" ------i
x.
rDitfuser Wall"

, ,.,
x ---1 - -
X X X X
Fillers
- -- --- ----- - •
~ ~
<

I
u
•• 3
•
e
"'"
c:, ,, :)'t~ C:' ,,, ,
1-;" - I ~~ '0

(~~)
~,
h . 0
I' e! ~
,) g~ 0
• ~: 0
I-'- ~ If, - -~ c
f~ n~
- - i
t...
I / ' ,~ <0
.0%
~
~
v
Flash Mixer
(typ,) Chemical POlymer r~ • &
u

Operations
PAC. NH 3
Addition Points U0=-1:
"
, ' .:
"
.
,
!
10
~ "0
AawWater
r 0·'l5~ \
' ~
. ~
v
0.

n
Clear.vells
,,
::.:;
;~
\
I
J
e
I
'"
!Ii
W
a
E
UJ
"x
,r-
~

=
r:
,
'1, r-
, ~,

; •e
., .....;N'"
~r
~
IPump House I It.
::;~
~

..:'"
Figure 3.2.2-4 Suggested treat ment process unit layout. The sc heme provides ( I) ope rationa l
flexibi lity (convent ional D, F, and in-linc titration); (2) sidc-by-side comparison tests (coagula-
,
.
u~
;!;:
ti on, mi xing. filtration rale, filt rati on rate, elc.); (3) compact arrangement and easy access 10 all
process units; (4) future ozonation and filt er additi on (GAC adsorption related) are considered. ...
~.
u~
- v~
~:i
l<
v
a contractor; contractors always select the most cost-effec ti ve unit that meets the design
specilications. There have been several embarrassing cases where the equipment was
delivered during construction but could not be installed in the stru cture because the door
size was too small or where the repl acement unit could not be installed due to the lack
5r
~~
of adequate spacing between the equipment. 1llesc types of problems should be dis- •
covered and corrected durin g the checking procedure of the latter stages of the design
phase. However, thi s requires the expenditure of more time and money. It is therefore
ex tremely important to sort out these issues at the beginning of the design phase.
The project manager and the project engi nee r shou ld also set lip a good sys tem
of design documentatio n, including a format for th e calculati on sheets and til ing
system, and di sseminate thi s program amon g the members of the desig n teu m. The 71

bz
 BASIC APPROAC H 73
72 DESIGN OF BASIC TREATMENT PROCESS UN ITS

design calculations and the record of eq uipment selecti on arc essen.ti al in ~ase of an tirst. evaluate the existing plant based on available doculllents and fi eld trips; second.
acc iden t and/or lawsuit. The project engineer should be capable 01 handling a\l the perf orm lield testing of the raw watcr quality and thc ac tual unit process ellicicncy. and
basic design calculations and selecti on of the principal equipment. Yet he or she study the local conditions; and third , evaluate and finali ze the plant modi fication schemc
should onl y handle the area of hi s or her competence. Qualilied experts shoul~ be bascd on the regulatory requirements of thc local and national agencies.
consulted for the other areas; the project engineer must clearly convey the obJec- The follow ing steps must be performed when evaluatin g an ex istin g plant:
ti ves and limitations of th e system and subsystems. Table 3.2.2- 1 contains some ba-
I . Collect, ana lyze, and evaluate all available doc um en ts concerning the design,
sic design criteri a used to size certain unit processes and pipelines. operati on, and hi stori cal raw fini shed water qu alit y of the plan t.
2. Interview the chief opcrator of the plant regarding operational ex peri ences and
Modification of the Existing Plant An ex isting plant may be modilied to ex-
obtain a list of both positi ve and ncgati ve comments and changes th at he or
pnnd plan t capacit y and/or be rehabi litated. The approach and considerations used
she would like to have implemented.
in this situat ion are quite different from those used in designing a new treatment
plant since the Illodilication Illust be based on the original plant design. Difficult 3. Obtain field meas urements of thc dimensions of all major process units and
pro blems may arise if the ori gi nal plant was designed by another firm many years mcasure the elevations of the wa ter surface level at selected poi nts in the
ago, when governm ental regulations were more lenient. . . process train , at the designed now rate of the plant, and at a different now
The lllodili cation of ex isting treatment plants is approached baSically 1Illhree phases: rat e. Do not rely on the as-built drawi ngs since the plant may have been mod-
ilied beyond what the as-built drawin gs show.
TABLE 3.2.2· 1 Basic Design Criteria 4. Evaluate all ex isting major equipment, conduct an in ve nt ory of the major
equipment . and evaluate their condition and utilit y. .
Guide Jor Design Flow Rare Jor Wmerworks 5. Condu ct an inventory of and evaluate the electri cal, in strumentati on, and CO Il -
trol systems.
Water treatmc nt proccsses including
intake facilit y 6. Inspect all major stru ctures for structural integrit y and evaluate their life ex-
Plant hydraulic capac ity 1.25- 1.5 X QUI••x.,.rn y pectan cy,
Equal or large r th an 0.15 X Q01ax-....;.y
Clearwell capacit y
or firefi ghting capacity (local code) Field testing is th e next phase in modifyi ng an existin g plant. The follow ing fi ve
Hi gh-service pump stat ion Q01••x-.. .'.y (Qmax.hour in some cases) steps must be perform cd:
Water di stributi on reservoir Q rn ••x-hour
I . Evaluate the treatabilit y of the raw watcr by conducting bench-scale tests. Be
(in the ci ty)
sure to evaluate the type of coagulant, opt imum coagulant dosage, optimum
Noles: Q:I\'c-...." y is an avcrage ann ual dail y now rate. mi xing conditions (G X t) . Iloc settlin g rate. and the liltcra bility of the set-
Q01"x..... :.y = 1.5 X Q",·c..... "y tled and Ilocculated water.
Qma., -huu r = 1.5 to 2.0 X Q11l:1.~ .rJa)· 2. Evaluate the effect iveness of the existing lilter was hin g system by obtaining
Qn.i..... . "y = 0.25 X Q,.vc ....... y core samples of th e lilter bed both before an d aft er filter washing. Al so ana-
lyze the act ual backwash rate, effecti ve sizes. uniformity coefli cients, and spe-
GHide for Pipe Size Selectioll Jar Waterworks
cific gravi ti es of the filter media. Furthermore, the mud deposition profile
Raw water main 6- 7 fl/s across the bed de pth should be assessed,
1.0- 1.5 ft/s fo r conventio nal rapid sand
Flocc ul ate water line 3. Evaluate the actu al now-th rough time of th e Ilocculati on and sedim entati on
filt er with alum fl oc
processes and thc clearwell by conducting tracer tests.
:to ft/s for direct nitrat ion process or
filt ers with reverse-graded media bed 4. Survey and investi ga te the geotec hnical aspect of th e locations whe re the ad-
2.5- 3.5 fl/s in general. 3.5-4.5 fl /s for ditional process units will be constructed.
Filte r intlucllt line
po lymer fed filt er influent 5. Verify the hyd rau lic bottlenecks and the performance of the unit proccsses at
Filter efllu ent line 5.0-6.0 fl /s hi gher th an designed now rates by performin g hi gh-now simulati on tes ts o n
Filte r wash-watcr main 8.0-9.0 fUs an isolated number of basins and IiIters.
Wash-waste main 6.0 fl/s
Di stribution main 6.0 [Us The linal phase in modifying an ex isting plant is evaluating the res ults obtained
Pump suctio n line 4.0-6.0 fl/s from the first two phases and establishin g a plan t modili cati on sc heme. Some of th e
Pump discharge line 7,0-9,0 ft/s morc important considerati ons are as fo ll ows:

74 DESIGN OF BASIC TREATMENT PROCESS UNITS
I. Evaluate and establish a balanced flow rate and process performance betwee n
the existing facilities and the new additions. . fore be compl eted in a fraction of a second. TI1C time rcquirement for other chemicals,
2. Try to use the ex isting tilter washin~ sys~cm by spe~difYing the· s~me-I·slzep' l~~~
ters at the same elevati on. This prac ti ce w ill also avOi any con f u SlOn or ,
operators.
:.. Check to see th at the design c riteria of all the unit ~roccssdcSI meet t:.l~:r~~~:t
requirements set by state, local, and fede ral agencies an t lose a e .-
seeable future.
4. Verify th at the owner' s requests are incorporated in the modification sC,hc me.
5. If any part of the old system is salvageable. try to integrate the old IIlstru -
mentation and control system into the new scheme.
6. Establish a basic plant modificati on schcmc so that the cxistin g plant opera-
tions arc minim all y di srupted by the constructi on.
Lastly, for individuals who wish to evaluate their ow n pl ant condi ti ons, the fol-
lowi ng steps are recommended .
I . Situation audit : addresses regulatory issues and analyzes Ihe plant's compli ance.
2. Plant capac ity audit: evaluates the hyd rauli c load.ing 0 f eac h luni: p~oce~s n:\~
analy ze the ac tu al hydra ulic profile across the e ntire p1ant. <~t Ile (eslgne .
rate and al so at higher plant fl ow rates to sim ul ate cOl1llillons prese nt dunng
emergency plant operati ons. . mi xi ng has littl e effect on settled water quality when operati ng under co nditi ons of
:.. Process fncility nudit : analyze the ph ysica l integrity of all process umts and
the ir conformance to all applicable regulations and codes.
4. Equipm cnt audit: assess the adeq uacy and reliabilit y 0df . ~II maj~ pr~cc:~
equipment and auxiliary eq uipment under operating ca n Illons an . e~~ ua.
the ir compliance to nil applicable codes, including safety and re liabilit y 111
e mergency situat ions.
BIBLIOGRAPHY
Kawamura. 5 .. et al.. " Improv ing Waler Sup~l: in TI.lai land:' 1. AWWA. 80:6~5~ (June 198~~.
Merritt . F. 5., cd., Standard Handbook/or Co,,1 £lIgllleers, 3rd cd .. McGraw HilI. Ncw Yo .
1983.
3.2.3 Coagulation and Chemical Application Considerations
Coagulation is defined as the destabilization of charge on coll~i~s aJ~d slls.pendcd
solids, including bacteria and viruses, by a coagul ant. Flash m.l xll1 g IS .an ~ntegral
part of coagul ati on. In thi s chapter we discu ss the role of che mi cal application and
n ash mixin g in e ffec tive water treatment.
Purpose The purpose of flash mi xi ng is to di sperse water-trc~ti~lg c~lemical~ quic.kly
and uniformly throughout the proccss water. Efl'ecti ve Ilas~ mlXI1l~ IS ~sl)ccla\l~ lin-
port ant when using metal coagulants such as alum and fem c chlonde, s1l1ce thclr hy-
BASIC APPROACH 75
drolysis occurs within a seco nd , and subseq uent adsorption to colloidal particles is al-
most immediate. In practical design , the dispersion of metal coagul ants should there-
such as polymers (polyelectrolytes), chlorine, alkali chemicals, ozone, and potassium
pcmlanga nate, is not as cri tical since they do not undergo hyd rolytic reactions. Thus,
from a practical point of view, the di spersion time of these chemicals may be com-
pleted within several seconds or less. Although flash mixing is an important unit process
in water treatment, most conven ti onal treatmen t plants are capable of producing qual -
ity watcr eve n with the use of a less than optimum flash mi xer prov ided that coagulant
is overfed in excess of 25 to 30%. It should be noted that Ilash mi xing of coagulant is
very different from the normal concept of liquid- liquid mi xing used in chemical in-
dustries. Under normal co nditi ons the rat io of liquid alum volume to raw water vol-
ume is I : 50,000 (in the water treatment field). Thus the instanlaneous dispersal of such
a small amount of alum into the raw water is a very unique situation.
A word of ca ution is necessary wi th regard to overemphasis of the coagulat ion
thcory and the ideal design. Many treatment plants with trad itional complete treat-
ment processes (which produces sweep fl oc), as well as rcactor-clarifiers. which of-
ten lack a flash mixer, demonstrate good sedi me nt ati on and filter performance. There
are also many case histories invo lving conve nti onal treatme nt process plants that
were poorly designed or co nstructed without fl ash mixing systems, th at yield ac-
ceptab le fl occ ul ati on, sedim entation, and filter performance. Moreover, Kawamura,
Amirtharaj ah and Mills, a nd C lark et al. (see the Bibli og raphy) have found that Ilash
the sweep fl oc and enhanced coag ul ati on mode.
Considerations The fo ll owing ilems should be evaluated carefully when de-
signing an effec tive Ilash mi xi ng system: ( I) type of coag ulant to be e mployed,
(2) number of c hemicals to be fed and the charac teri stics of each, (3) local condi -
ti ons, (4) raw water charac teri sti cs, (5) type of che mical diffusers, (6) available head-
loss for the n ash mi xer, (7) variations in th e pl ant now rate, (8) type of subsequent
process, (9) cost, and (10) other miscellaneo us items.
Type of Coagulant The selection of coag ul ant is ex tremely important in two re-
spects: to set design criteria for the flash mi xing syste m a nd for effective fl occula-
tion and clarification . The mos t commonly used coagulants are the metal salt co-
agul ants: aluminum sulfate, ferr ic chl oride, and ferric su lfate. Synthetic polymers
such as polydiallyl dimethyl amm onium (PDADMA) and natural cationic polymers
such as chit osa n (made from cru stacea n she lls) are also employed .
The difference betwee n metal salt coagulants and cati onic polymers is their hy-
drolytic reac ti on with water. Metal salts undergo hydrolysis whe n they are fed to
the process wa ler; yet a ll polymers do nol. T hi s hydrolytic reac ti on produces hy-
droxocomplexcs such as AI(H ,O)~+ , Fe( H,O)j+, AI(OH)2+ , and Fe(OH)2+ in a pH
ra nge found in the process trains of most water treat men t plants. The formation of
the hydro lytic products occurs in a ve ry short period of time: less than I s. These
products read il y adsorb onto colloid parti cles and cause destabilization o f their clec-
76 DESIGN OF BASIC TREATMENT PROCESS UN ITS BASIC APPROACH 77

trical charge. In turn , the hydrolyti c prod ucts are qui ckly polymeri zed th roug h hy- TABLE 3.2.3-1 Iml)Ofla nce of Time Delay Secluencing for Anionic Polymers as
drolytic reac tions. Floccuhtnt Aids
In summary, nash mixin g or instant aneo us mixin g is important because (I) hy-
droly sis a nd polymerizati on are very fast reactions, (2) a uniform supp ly of coagu- (aJ. Jar Test of Sill/ace Wat er for Turbidi,y Removal
lant and a uniform process wa te r pH are essen ti al for the production o f hydrolyti c Raw water: turbidity = 18 nil! . pH = 7.9. atkalinity = 84 mglL
produ cts, and (3) adsorption of thesc species to colloidal particles occ urs quickly.
When cationi c polymers arc used as coagul an ts, ins tant aneous mi xing is not c riti - Turbidity (ntl!) for
cal because hydro lyti c reac ti ons do not occur; the rate of co ll o id adsorption is much Time of Polymer Addition Alum Dosage (mg/L) of:
slower because of their large physical size. T hu s, from a practi cal point of view, the wit h Respect to Alum 0 to U 16
mixing time for polymcrs may be 2 to 5 s. No polymer addition 18 10.5 6 4
Ani oni c potymer (8 18 1) at 0.2 mglL
Number of Chemicals and Their Characteristics More than two or three chemi- 5 min before 7.5 4.5
cals are com monly fed into the nas h mi xer, whi ch is located at the head of the treat- Same time 6.5 3.5
ment plant. The most frequentl y used chemicals are alum, cation ic polym ers, potas- 5 min aftcr 2.5 1.2
sium pcrmanganate, chlorine, powdered act ivated carbon (PAC), am moni a, lime or 10 min after 1. 2 0.7
caustic soda, and ani onic and nonionic polymers. Design eng ineers shou ld evaluate the
characteristics of each of these chemi cals and class ify them into two or three categories (bJ. Jar Test of Imll Removal by Chlorine Oxidation for Well Wat er
based on the following cri teria: necessity of instantaneous mixing, potential chemi cal
Raw wate r; turbidity = 0.8 ntu. pH = 6.7. alkalinity = 198 mg/L
reactio ns that prevent the process chem icals from acting individuall y, and the chemi - hardness = 188 mg/L, Fe = 4.4 mg/L
cal addition sequence (to the raw water) that yields the most eO'ecti ve results. Chemica t dosages: Ci , = 5 mglL. NaOH = 15 mglL
Proper chemical appli cati on seq uence is very impo rtant because it produces re-
Timc of Polymer Addition Turbidity Iron
markable result s and also reduces total chemical cost. For exa mple, whe n an io ni c
with Respect to Chlorine (ntu) (mglL) pH
polymers are fed to the process water durin g pinpoint noc formation , the size and
we ig ht of the alum fl oc are o ft en improved dramatically. Genera ll y, a 5- to IO- min No polymer addition 5 2.5 8.0
lag tim e is esse nti al fo r optim um fl oc fo rmatio n. Table 3.2.3- 1 illustrates thi s point. Ani onic polymer (7763) at 0.2 mglL
Selec ti o n of a n appropriate chemical app lication seq ue nce is also vital in cases 30 s before 5 2.2 8.0
where the raw water does not have s ufli c ient alkalinit y. In thi s type of situat io n, al- 2.5 min artcr 3 1.6 8.0
5 min after 2 1.4 8.0
kali chemicals suc h as lime or causti c soda are fed to the water in conjunct io n w ith
alum . The mos t reasonable c he mi cal applicati o n sequence is first to lowe r the pH
of the raw water through the addit ion of alum , thereby all ow ing di va le nt and tri va-
lent aluminum hydroco mplexes to form ; these complexes effec ti vely red uce coll o id Raw Water Characteristics The design of the coag ulation process is determin ed
charge. The next step is to adju st th e pH of the water to th e ran ge of minimulll alu - by the raw water c harac te ri sti cs. The mos t commo n problem associa ted with the
minum solubility to allow alumi nu m hyd roxide ( l1 oc) formation . Fig ure 3.2.3- 1 process of coag ul at ion is c logg ing of the c hemic al diffu se r orifi ces a nd feed lines.
presents the so lubility eq uilibria of aluminum and iro n hydrox ide s, while Fig ure This is related directly to the hardness of the process wa ter (metal hydrox ides)
3.2 .3-2 illustrates the steps invo lved in a proper coag ul at io n and Il occ ulation process. and , in part, to di sso lved and s us pended soli ds. If the water hardness exceeds
30 mg/L, the metal salt and al ka li che mi cal feed lines and the diffu se r or ilices
Local Conditions The local conditi o ns that s hould be investi gated duri ng plan- will becomc clogged wit hin several months or less. The design o f the sys tem
ning of the process desig n are the avail ability of equipment and parts, th e avai l- s h ~uld th ere fore fac ilitate simple cleanin g o f the feed lin es, injec tion nozzles, a nd
ab ilit y of serv ices w ith in the area, and th e reli ability of the powcr supply. Other con- on ti ces. Although low wa te r te mpe ratures have an adverse e ffec t o n c he mical di s-
ditions that inlluence the choice and desi g n of the fl ash mixin g system are the persio n and reacti o n rates, unde r norma l co ndition s this influe nce is minimal.
frequency of sa ndsto rm s and the winter te mperatures. For insta nce, in remo te areas However~ raw water alkalinity and pH are critica l factors in both coag ul ati o n and
and developing countri es, the use of hyd raulic j umps and in-lin e sta tic mixers is the fl~ccu lat,on. Appropriate pH ra nges fo r met al salt coag ul a nt s are illustrated in
most appropriate cho ice. If the system is located aboveground, it should be pro- Figure 3.2.3- 1.
tec ted by a building struc ture . Howeve r, if the plant is located in a n area th at expe-
ri ences frequent sandstorms, severe w inter wea ther, is in a quiet resid e nti al neigh - Type of Chemical Diffuser T he most comm on ly used water treatment chem icals
borhood or an area wi th a hi g h rate of vandalis m, and in some cases, the potential may be ca tegorized into three basic groups: gases, liquids. and solids. M ost of these
for milit ary attack, the fl as h mi xing systelllmust be located in an underg round vault. chemicals are converted into soluti o n prior to being fed to the raw watcr, wi th the
 BASIC APPROACH 79

OAY "lUM S TO 20"" SOL' N (411'11- """XI

~(S04)3 16H:l0 ,\1215°4

pH pH
(0) (b)

0.70
I
""

0.60

lET" POTENTI"lllp) _ - 20 mV lE T" POTENTt"l - -1I mV lET" POTENTI"l _ O TO-IO IIIV

0.50 SUlIlE COLlOIO"L P"ATtClE OESUBLIZEO P"ATlClE P"RTlClES tNM ESHEO ~ FLOC

\ J Figure 3.2.3-2 Al um coagu lation and Ilocculatioll process.

0 ! excep ti on of ozo ne, carbon dioxide, lime, and powdered activated ca rbon (PAC).
Both ozo ne a nd carbon dioxide (w ith the except ion of liquened CO 2 ) are genera ll y
manufactu red on site; these gases usually contain on ly a small percentage of the es-

0.20
1\ V sential components and their solubilit y is very low. T hey are therefore fed to the
process water th rough deep-channel co ntact tanks and are released from the tank
bottom as fine bubbles to ensure eflic ien t gas transfer. Furth er discussion o f gas

0.1 0
\ J mixing is not covered in this c hapter since nash mixing usually does not ac hi eve
thi s func ti on e ffec ti ve ly. Lime and PAC are ge nerall y fed to raw wa ter in the form
of a slurry. A few small plants lllay feed lime to the finished water as satura ted cal-

I~ ....;:/
~ ciu m hydroxide water, provided that a lime saturator is used .
In summary, water treatme nt chemicals are fed to the nash mixing system in ei-
0
5.0 6.5 7.5 8 .0
ther of two forms, solution or slurry. For both pipe and cha nn el n ow application,
5.5 6.0 7.0
pH
the chemical diffuser (for solutions) is most often a pipe with Illultiple minces. If
Ellect 01 pH on solubility 01 alum lloc the chem icals are in sl urry form, two types of diffuser may be used: For pipe ap-
(constant alum dosage of 1.0 s.ll/gal(l7.1 mgIL). plication the ope n end of the diffuser pipe acts as the di ffuser, whil e chan nel Ilow
pH adjusted with acetic aCid and sodium hyd(OKide
(, ) appli cation uses a trough with notches or holes at the bottom.
Special attent ion must be give n 10 chem ical solutions that produce scale si nce
Figure 3.2.3-1 Solubilit y of common noc versus pH. (a,b) Equi librium compositi on of so-
luti ons in contact wi th freshly precipitated Al(OHh and Fe(OHh. Calcu lated using repre-
they ca n c log the diffuser orinces. Metal sa lt coagu lan ts, suc h as alum or ferric chlo-
sentative values for the equilibrium constants for solubilit y and hydrolysis equi li bria. Shaded ride, and alkali chem icals (caustic soda, soda ash, lime-saturated soluti ons, and am-
areas are app rox imate operating regions in wute r treatment practice; coagu lation in these sys- monia solutions) are scale-formin g c he mi cals. When these chemi cals are employed,
telll s occurs under conditions of aversaturation w ith respect to the metal hydrox ide. (From a provision for period ic c lea ning shou ld be inc luded in the design of the feed sys-
Stumm ancl O'Meli a. \968.) (c) Effect of pH on so lu bi lity of alum floc. [Constant alum dosage tem; an additio nal diffuser must be specified so th at plant operat ions are not inter-
of 1.0 g/ga l ( 17.1 mg/L). pH is adjusted with acet ic acid and sodium hydroxide.] rupted whe n one diffuser is being cleaned. Figure 3.2.3-3 depicts the most com-

78
 BASIC APPROACH 81
80 DESIGN OF BASIC TREATMENT PROCESS UNITS

HUD LOSS THRU ORIFICE - F[[T

HoI., "aWI .....IOfI .....
,_,.,""
",$,~--:.:t!~ r_.........._ .. to

-"~
... .....
~~.;::.~;:.,
,""" .... Ilo<'ft
: ,..........
__ .................
plu to._ .
NC_IIfY. ...
'_._>Il
N t _ .......
0<0
.,' . -= ,.,.."Of _ _
11IIUI. f IOoplO"""

. N ...... oI ..
~ ... ..., 6"M""""'"""", ~<I'II<"'''''

I Z Ho4o. _ Ii 0/>0000 (hoIH"

_ " ... .,: .. lObt..,. .. <OIlod) ~.
T, .. e: ........ iR_ _ )

_"",....-_...
UG .... . _

~~~~~.~~~[::
Sool'M' ...,~
co-c.. ,,~"')
$1......... '·
lllU. """ 1.. 1" ,,, .1
mo.t,!"""'" ... ~con<>."
I ,=====::I~- -~,
..,... ~""I

' .... ( , (;< . ",..f. . """'.f,....h_

' .... 0. .. . " ......, .. .. . 11, .... "'. w
.,.._
Figure 3.2.3-3 Chemical diffusers.

ma nly used chemical diffusers, Figure 3.2.3.4 illustrates the basic design criteria for
multiorificc diffusers.

Available Headlass far the Flash Mixer Water treatment plants may occasion-
ally have excess pressure in the raw water line. This pressure must be reduced by
means of a special valve or elec tric power gene rator. When head loss can be afforded ,
a hydraulic mi xing system such as an in-line static mixer, weir, or hydraulic jump
should be considered . In any case, a headloss of approximately 2 to 3 feet (0.6 to
0.9 m) is generall y required to ensure good mixing. However, if a plant has ti ght
hydrau lic conditions that can not afford headloss, a special syste m such as pump dif-
fus ion should be employed . If instantaneous mixing is not absolute ly necessary, a
prope rly designed pipe diffu ser system is the correct choice.
Since mec hani cal mixers do not produce significant headloss, they IHay be con-
sidered as a feasible alternati ve to the nash mixer. However, many studi es indicate
_ , -"'"O ••f( _ " "''':''011 '01' ..00
that most mechanical nash mixing systems designs are ineffective du e to n ow short- ""'000 _ ,, ' 00 .. , , _ .... t , tooo", 001",""

circuiting and frequent maintenance and repairs. In recent years, several artic les have , ~ ,

.. 00" ... . . ....-

stressed the inability of mechani cal mixing systems to ach ieve instan taneou s mi x-
ing in tanks lac kin g the proper type of bafning.
~
.~~. -"".·o_..._-
" 10"

C'OI U
~

·IJO.
' .0 I ••
Variations in Plant Flow Rate Variations in water demand cause fluctu ati ons in
the plant now rate on a daily and seasonal basis. Since variat ions in plant !low rate
TABLE XIll
can eas il y exceed the I : 2 ratio. the mix ing e nergy of the hydraulic flash mi xin g sys- WAll-'Ce, TIERN.-.N DIVISION
P£NN'oW.l.T CORPORATION DESIGN OF PERFORATED
tem must also change in proportion to the plant flow rate. Recent studies in elTec tive
IElLEVlLlE , NE~ ~~~E~T:~ TYPE DIFFUSORS
!lash mixin g have show n that the va lues for the product of mi xin g energy (G value (#.1. NO. 5 .290
expressed in seco nds- I) and mixing time (1 in second s) should range from approxi- " ",,,1 ·11
mately 300 to 1600. Hydraulic mixing ca n be used if G X t fall s within thi s range. If Figure 3.2.3-4 Diffuser pipe design guide ,
the seaso nal plant !low fluctuation is too great to satisfy the proper G X 1 ran ge, the
design must provide two paralle l hydraulic fl ash mi xing systems: The design criteria
will be sati sfi ed if only one of the two systems is used during a low-flow season.
82 OESIGN OF BASIC TR EATM ENT PROCESS UNITS BASIC APPROACH 83

Due to variation in mixing time, plant flow variation becomes an importa nt dc- TABLE 3.2.3-2 Effect of Flash Mixing on Clarilication
s ign factor in other mixing sys tem s, s uch as pump diffusion and mcc hani ca l mix-
ers. T he mixing energy must therefore be adjusted to satisfy the proper G X , range.

tflrE~![
The pum p diffusion sys tem uses e ither a flow-regu lating valve or a pump with VFD
(variable frequency dri ve) con trol to regulate the pUlnp discharge flow, while me-
chanical fl ash mixers employ a variab le mi xin g speed co ntro l unit.

Type of Subsequent Process The desig n engi neer Illust carefully co ns ider the J.. 'H', ""1' "'10,. W""
MOdI'
lype of treatment processes that wi ll be subseq ue nt to th e Ilash mix ing system si nce
the coagu lant dosage and th e degree of pretreatme nt ca n be quite differe nt. As ide Value 01 the flash mixer was approxi.
mately 1,000 s-'; mixing propeller rota .
from sludge cond itioning and waste-wash-water recl amati on, four basic process tion was set to give upward now.
trains are used in modern water treatment: in-line filt ration , direc t filtrati o n, two-
stage filtrat ion, and complete co nventional treatment. Lime so fte nin g and hi g h- leve l
comple te treatme nt processes, such as ozonation and g ra nular ac ti vated ca rbon ad - ED<et 01 Fftuh Mlrln, by .• B",ch.Typt Ttn.
sorption , are considered to be modified complete co nve ntional treatm ents. Iro n and Fo=<

• , • 7 • • ••
.
TC'St No. Anti,.
ma nganese remova l (th roug h oxidati o n) may be ac hi eved throu gh e ith er direc t 1iJ-
• 2 I
trati on or the conventional complete process train.
The in -line, direct, and two-s tage filtrati o n processes have two di stinct cha rac-
QuaJity of raw "'ater
l\Irbldil1.....Jtv
pH •
'.7
2•
7Jl
I. , ..
"... ••• ••• "••• ... ".••• .., '" ... lOG

•• , •• " ,. " 20.. ,... ••

7.' U
AllcaIWIY-m,/I
teri sti cs; they have no sedime ntati o n process and requ ire very low leve ls of coagu- (II C.aCo,) 17 2J U 21 2. 2. 11
lant (2 to 6 Illg/L). Inadequate flash Illi xi ng of coagul a nt results in hi gher lilterc" Alum. dOUIe-m,/I
TIme ft4uirtd to (orm. •• 11 2J
19.'

water turbidity because the s ma ll amount of coag ul an t fails to d isperse unifor ml y.

In co ntrast, th e conven ti onal compl ete process always in clu dcs sed imc nt ation and
risible noc-mlll
Cue •
Cue 2
,.,
U
U
•.> ,..... ...... .., 2.' 1.' • '.7 1.1 I.l
U 2 32., 2.' 2.1 I
thereby all ows the forma tion of good settleable fl oc, provided th at an adeq uate 1\Irbldity of SUpcmaUnt
dosage of alum is appli ed (15 to 30 mg/L). Hence, if s uffic ie nt alulll is fed to th e
process water, the role o f flash mi xin g may not be obv ious. Ye t an e lTcctive nas h
-J",
Cue •
Cue 2
I 1.0
1 1.0 2
... , • ,, 2
7
I
• 3•••
.•
11
17 •
14 • ,.
11 I..
·CaM I iI with Club mid.a I Lad 1 is whllout nub mb.l.a ••
mi xing system requires 25 to 30% less al um . Thi s reduction has been confirmed by U.$C

several operati onal treatment plants after the insta ll ati on o f eftcc ti ve fl ash mi xers.
If the conve nti o nal compl ete proccss is uscd in treati ng raw wate r with low tur-
bidity, the benelits o f the flas h mi x in g syste lll Illay not be apparen t. However, if the EDtet 01 Flash Mixing T~sItJ on ,IIt
raw water turbidit y is hi gh, use of an e ffective flash mi xin g sys tem dramati ca ll y im- Acrual Piallr-Kabllloyama Plant,·
Kob~, Japan
proves the turbidity of the settled water (Tab le 3.2.3-2).
S lu dge condit io ni ng and was h waste clarification mu st be treated differentl y from fU10r Tnt' Tt'St 2 Tell)
ordin ary water treatment because of the hi gh co ncentrati o n of so lid s. The coagulan t Quality or r.a .... wattr
Turbidity- Jill
•••• 7.' ..,
used in alum slu dge co nditi o ning is a polymcr, no t alum . Poly me rs do not req uire
instant mixing. In fact , inte nse mixin g may advcrse ly di sintegrate th e ex is tin g fl oc
pH
AJIr:.allnlly-mrl l
(u Caeo,)
••
16.S ..,. ..
6.U 6.7S

structure within the s ludge. Unfavorable conditi o ns will no t ari se if the G va lue of
slu dge conditioning is 100 to 150 s - I with a mixin g time of 2 min or less.
Alum dos"e-mrll
Tutbldlly or
J'locculatinl W:Htr
-It/l
••
,.
i 11

Cost Cost-effec ti ve ness is alway s an important co nsideration ; it is part icul arl y cru -
Fluh mixer on
Flu.b miser ort
TurbiditY or Hulin.
", ,.•
3.• U
2.'

....,,
c ial when designing plants fo r develop in g countri es. With thi s in mind , the fl ash mi x- unk dDuent-/tu

in g system should be reliab le, e ffe cti ve, and easy to ope rate and maintain. The mec h-
Fluh miur OQ
Auh cniJ.cr 011
I..
I.. •I..••
anism of the flash mixing sys tem sho uld be simple so that service and repair ca n ·,)'.) • ......04u mlo c.apac 11.
..
eas il y be accomplished by the operatio nal staff. Experience has s hown that pla nt o p- S~//~ e. Repnnled from JO//rl/al Ameru.:em l\11ler \Vorh Anodmioll , Vol. 65. No.6 (June 1973). by per.
erators will not use the fl ash mix ing system if it is overly complicatcd, req uires fre- mISSIon. Copyright © 1973. Alllcircan Water Work Associalion .
quen t ma inte nance, or ge nerates too much no ise. Also, the opcrators will no t use the
84 DES IGN OF BASIC TRE ATMENT PROCESS UNITS
fl as h mi xer if they feel that it does not produ ce a di scernibl e di fference in the treat~
me nt process. For these reaso ns, the cost of the flash mi xin g system should be ap-
pro ximately 0.3% of the co nstructi on cost unl ess there are some justifiabl e reasons.
Miscellaneous Items Although some arti cles cmphas ize the ad verse effec t of back-
mi xing during the fl ash mi xin g process, thi s theory is somew hat debatable. Backmi xing
implies the ble nding of chemi cals to the mi xlme of raw water and recycled che mical-
lade n raw water. A typical exampl e of bac kmixing is a tank with a mechanical mi xer
o n ce nter ( for mi xin g alum), prov iding 1 to 2 min of dete ntion tim e.
Operato rs of many plants using mec hani cal mi xers as the fl as h mi xin g system
obse rve no measurable be nefits by operating the mi xer. In fac t, mechani cal prob-
le ms, s uc h as noise and energy waste, and hi gh maintenance cost are often cited as
reasons no t 10 run the mixer. Fl ash mi xin g by mea ns of co mpressed air has bee n
tri ed. However, thi s meth od has never become popular because there is some do ubt
th at it ca n achi eve instant aneous and uniform di spersion of coagulant and because
of the noise ge nerated by the co mpressor.
Type and Selection Guide In thi s secti o n we prcse nt comm onl y used alterna-
ti ves in the fl as h mi xing process and guid elines for making the proper selecti o ns.
Available Alternatives Flash mi xing may be ac hieved by means of a hydraulic ,
mechanical, or pump diffu sion syste m. Fig ure 3.2.3-5 shows the most common types
of fl ash mi xer: mec hanical mi xe r, hydraulic mi xer, in ~ lin e mec hanic al mi xer, in ~ lin e
stati c mi xer, diffu sion mi xin g by press ured water jets, and mi scellaneous-di ffu sion
by pipe grid .
Selection Criteria Selec ti o n of the n ash mi xin g syste m sho uld be based o n the
considerat io ns di sc ussed earli er in thi s sec ti o n.
Order of Preference When es tablishin g th e des ig n of a n ash mi xing process, the
order of prefe re nce is based o n effec ti ve ness, reli ability, minimal mainte nan ce, and
cos l. Thus the order of choi ce is as foll ows:
I. Diffusio n mi xin g by press ured water jets
2. In ~ lin e stati c mi xing
3. in-line mec hani cal mi xin g
4. Hydraulic mi xin g
5. Mec hani cal fl as h mi xin g
6. Diffusion by pipe grid
Discussion of Alternatives
Diffusion Mixing by Pressured Water Jets For var io us reaso ns, d iffusion mix -
in g by press ured water jets is the first choice among the alternati ve fl ash mi xing
BASIC APPROACH 85
processes. T he ad va ntage of thi s type o f fl as h mi xin g system is that there is esse n ~
ti all y no additi onal head loss by the mi xer, it is very effec ti ve, it has a co ntroll able
degree of mi xin g, its power co nsumpti o n is less th an half that of a mec hanical mix er
system, it is cos t ~e ITec ti ve, and all parts of the system arc oIT~ th e~s h e lf items.
The source of the pressured wa ter may be e ither the pla nt utility water line, the
di sc harge of th e hi gh service pumps, or pumped process water. The raw water should
not be used w ith dilute a nd d iffuse coag ul a nt unl ess the turbidity of the raw wate r
is continu ally low (less than 5 ntu ) and virtu all y free o f any suspended solids, in-
cludin g fi sh and freshwater clams. If the forego in g criteri a arc not met, serious noz-
zle clogg ing proble ms will occ ur unl ess a bas ket strain er is provided in the pump
sucti on line. Ideall y, se ttl ed or filt ered water shoul d be pumped. The press ure used
to diffu se the chemi cals mu st be a minimum of 10 psi (0.7 kg/cm 2).
One d isad vantage of thi s alternati ve is the po te ntial for coag ulant and debris pres~
ent in th e pumped water to clog th e nozzles. A seco nd di sadvantage is th e difli cult y
in applying it to a sys te m with ex tra large pipes or chann e ls, for exa mpl e, th ose hav-
in g a di ameter greater than 100 in . (2500 mm). In a n extra large syste m, o ther mean s
of di spersio n, such as multiple no zz les or a pi pe g rid with inj ec ti o n nozzles, sho uld
be empl oyed. To avo id c logg ing the injec ti o n nozzle with me tal hyd rox ide scale, it
is ex tremely important to e nsure that the metal salt coag ul ant s are not excessive ly
diluted prior to inj ec ti on. In gene ra l, alum so lutio ns should not be diluted beyo nd
1% (5% for ferri c c hl oride). T he diluti o ns may be adjusted if the diluti o n water is
very soft and if the hi gh level of diluti o n docs not shift the pH o f the coagulant into
the ran ge of tl oc form ati o n. These metal coagulant s are bes t fe d to the wat er j ettin g
fro m the nozzle as neat soluti on by means of a separate feed line.
In~Line Static Mixers In-line static mi xers are also kn ow n as mo ti onl ess stati c
mi xers. Half a doze n different types of in ~ lin e mixers are curre ntl y avail able, and if
selec ted and appli ed properl y, they arc quite effec tive in the coag ul ati on process.
The primary appli ca ti on of th ese unit s is in ind ustry, because of the ir e ffec ti ve ness.
T he ad vantages of these mi xers are the ir lac k of mov ing parts and th e fact th at th ey
do not require ex ternal energy to be input into the system. However, two di sad va n-
tages exist: ( I) the degree of mi xin g and the mi xin g tim e are a fun cti o n of fl ow rate,
and (2) the unit s are proprietary items a nd th e des ig n e nginee r mu st there fore rely
on the perform ance claims of the m a nut~lc turc r. Fro m a prac tical po int of view, a
mi xin g tim e of I to 3 s and a max imum head loss of 2 to 3 ft (0.6 to 0.9 Ill) ac ross
the unit are normall y quite acce ptable for waleI' treatme nt. T he treatm e nt pl ant d e~
sig n should have an intake sc reen upstrea m from the stati c mi xe r so that large de-
bris will not obstru ct the stati c mi xer. It is al so important to specify the mi xin g e l ~
ements as removable to faci litate th e necessa ry c leanin g o f debri s and scale.
Hydraulic Mixing Pars hall Ilumes, Venturi meters, and we irs arc categorized as
hyd rauli c mixing devi ces. When raw wa ter Il o ws into a plant throu gh an open c h a n ~
nel, the plant flow ca n be metered by either a Parshall or Palmer- Bowl es Ilume
sin ce th ey provide hydraulic j umps and large~scal e turbul e nce. Howeve r, if the raw
Water is pi ped to the plant, Ve nturi or orifice meters should be placed within the
 BAS'C APPROACH 87

influen t pipeline; these meters produce differential head and result in a certain de-
Tellon or Kynar
element
gree of mixing- more so with the orilice meter. Weirs are seldom used as fl ow-
retainer disk measuring devices in modern treatment plants; nevertheless, they also produce down-
stream turbulence.
These hydraulic fl ow-measuring devices may also be used for chemical fl ash mix-
ing; they have been used in thi s capacity, particularly in developing countries. However,
Removable element assembly the drawback to thi s procedure is that the degree of turbulence is a function of plant
made from solid Kynar, coated metal,
or special material-e .g., Carpenter 20 flow rate, and thus there is no positive control over the degrcc of mixing.

~ Orive (direct or
Mechanical Flash Mixing The mechanical flash mi xer is the most frequently used
i' I~~
'2~V-belt)
~
Auxiliary chemical type of flash mixing unit in the water treatment industry. It consists of a tank or chan-
feed lines ncl with onc or more mcchanical mixers. The common design parameters are as fol-
lows: G = 300 s- ' , a mixing time of 10 to 30 s, and a power requirement of 0.85 to
Rotameter panel
1.0 horsepower (hp) per million gallons per day. Recent studies have shown th at thi s
LJ ~cLI""_-Chemical type of rapid mixing is not preferred for use in a continuous fl ow process because of
solution

~
.jJ[-rl 'eed ';ne
'<_ Raw
(I) the lack of instantaneous mi xing characteristics, (2) the production of significant
flow short-circuiting, (3) a mi xing period that is too long for metal salt coagulants,
l water
-' and (4) back-mixing, which may adversely affect coag ulati on. Furthermore, there is
Flow contraction
some difficulty in analyzing the various forces acting on the mixer shaft and impeller;
Removable chemical
baffle feed place and hand thi s has resulted in shaft problems and gear drive failures in numerous installations,
leed
especially in large-scale plants. If marginal equipment is employed, it may create rel-
line atively high operation and maintenance costs and noise pollution. For the foregoing
reasons, mechanical !lash mixers rank low in system preference.

Diffusion by Pipe Grid This type of flash mi xer depe nds on the wake turbulen ce
created by the pipe grids; coagulant or other chemicals are added to the flow stream
through injec tion orifices in the grid. Kaufman et al. (1972) have suggested that a min-
imum density of one orifice per square inch is required to yield satisfactory coagula-
tion and fl occulation results. However, the majority of actual installations employing
this criterion experience problems with scale clogging the orifices after several months
to a year of plant operation. Therefore, under normal conditions, it is not recommended
Chemical that thi s parameter be included in plant design, ex.cept in pilot-scale installations.
diffuser

Design Criteria The selec ti on of an effective, yet trouble-free nash mixing sys-
tcm is not a simple task, espec ially when designing large-scale treatment plants. The
appropriate design criteria for the fl ash mixer process must be selected carefully.
Althoug h the velocity grar;lient (G value) has bee n used as a design guide, it is some-
what ambiguous in determining the process because the effectiveness of coagula-
tion is greatl y affec ted by the following criteria: shape of the reactor tank, degree
of fl ow short-circuiting, type of mi xing element, degree of e nergy input, and effec-
tive mi xing time. Appropriate design parameters should be based on (I) chemical
Figure 3.2,3-5 Vari oll s types of flash mi xers. diffusion rates, (2) the degree of turbul ence associated with the inertial forces (char-
acteri zed by the power number), and (3) the Reynold s number, which represe nts the
viscous forces and flow characteristi cs. Unfortunately, most practi cal design cri te-
86 ria are not based on these parameters.

88 DESIGN OF BASIC TREATMENT PROCESS UNITS
The widely accepted ge neral desig n criterion for nash mi xin g is the velocity gra-
die nt C:
C -
_(~)O.5
",V
where P is the power input, J..L the absolute viscosity of water, and V the vo lume of
the mi xin g zone. The most effecti ve velocity g rad ie nt G (in second s- I) and mi xin g
time t (i n second s), is
C X t = 500 - 1600
If pump diffusio n is used as the flash mi xing syste m, the fo llow in g desig n cri-
teria should be used:
1. C X t = 400 - 1600 ( 1000 average) .
2 . Mixing jet velocity of 20 to 25 fils (6 to 7.6 m/s) at the m ilice.
3. If neat alum solution is diluted prior to applicati o n to the process wa te r, the
pH of the dilut ed alum should be lower th an 3.0. If ferri c so lts are used as co-
agulant, th e pH should be less than 2.0. However, diluti o n of metallic coag-
ul ant is not recom mended unl ess absolute ly necessary.
For in-line stati c mi xers, the foll ow in g des ig n c riteri a sho uld be used :
1. C X I = 350 - 1700 (1000 average)
2. I = I to 5 s
The foregoing conditio ns may be sati sfied wit h a single in -line static mi xer if the
plant fl ow rate variati on is less th an 50%.
Chemical Diffusers Chemical diffu sers are a co mponent of all 11ash mi xin g sys-
te ms that most design e ngineers fa il to note. Thus, many che mi cal diffu se rs become
obstructed for vari o us reaso ns. Metal salt coagul ants, such as alum and ferric s ul -
fate, and alkaline chem icals, sll ch as lime, caustic soda , and ammo nia, wi ll clog
small orifices within a short period of time unl ess the process water is e xtremely
soft and has very low turbidity.
Selec tion of the proper chemical diffuser is a functi on of the coagulant and the
type of process e mpl oyed (refer to Figure 3.2.3-3 fo r the fo ll ow ing di scussio n). For
metal salt coag ul ant s and alk ali ne chemi cals, the correct type of chemi cal diffu ser
is ge nera ll y a type 0 diffuser, since it minimi zes the problem of clogging. Type E
may also be used for aq ua amm o nia in cases where the area of che mi cal app li ca-
ti on is not easil y accessi ble a nd if fumes are not all owed to escape into the env i-
ro nme nt thro ugh the installati o n of subme rged curta in walls both upstream and
downs tream. Types A and B diffu sers may be used if mixin g is achi eved solely by
BASIC APPROACH 89
diffusion and not by means of pos iti ve flash mixing. since the diffusers may easily
be di sconnccted. Types A a nd B diffusers may also be used with c hlorine, polymers,
potass ium permanganate, and fluoride solutio ns because these chemicals usuall y do
not produce scale at the o rifices. Type E diffusers may be used wi th lime or PAC
for channel n ow application upstream of the flash mi xer. Type F diffu sers are
used exclusively for gaseous chemical s such as ozone and ca rbo n dioxide. Fig ure
3.2.3-4 illustrates the proper design of mult ioriflce diffu sers.
Enhanced Coagulation Process E nh anced coagulation is loosely defined by
the EPA as the addition of excess coagulant to improve rem oval ofTOC. This process
has been in practice for many years as a means of removing colo r from water COIl-
taining a hi g her level of natural organic matter (NOM) such as humi c acid and tan-
nin . Removal of the hydrophilic NOM is attained by means of coagulatio n and noc-
cuiat ion using alum or ferric salls. It is well known that effective removal of color is
ac hieved through the use of an optimum dosage of inorganic coag ul ant to make the
zeta potential of the colored substance zero (the isoelectric point). The pH of iso-
electric poin t for color removal is in the range 5.5 to 6.5 using alum as a coagulant.
As mentioned previously, the Surface Water Treatment Rule calls for e nh anced
coagulation to reduce the total organic carbon (TOC) of raw water und er spec ific
conditions. The isoelectrie point varies depending on the nature of the colloid and
the characteristics of water. Thus the pH of the isoelectric poi nt is no t necessarily
5.5, as s uggested by the EPA . In ge neral, the usc of ferric c hloride is more effec-
tive than alum in remov ing TOC, as show n in Figure 3.2.3-6. The pH of the iso-
elec tric point is generall y 6.0 o r less for water treatme nt practice. The addition of
acid to create a pH of about 6,0 (for enhanced coag ulati o n process) ofte n improves
the e fficien cy of TOe remova l a nd saves in coagulant dosage, as illustrated in Fig ure
3.2.3-7.
Negative aspec ts of e nh anced coagulati on include the prod uc tion of ve ry corro-
sive wa ter fo r metals and concrete and the production of hi g h levels of di ssolved
aluminum in se ttled water whe n a lum is used as coagulant. Thus the pH of se ttl ed
water mu st be co rrected by a lka li substance (lime or caustic soda) to prevent severe
corros ion of submerged met als and concrete, and control hig h levels of dissolved
alum inum. Compared to the co nventi onal treatment mode, the production of two to
three times morc sludge is another drawback of enhanced coagul ati on.
Example Design Calculations
Example I Mechanical Flash Mixing System
Give"
Plant flow rate 1.85 m 3/s (42 mgd) max imum per day
C X I 1000
Mixing tim e Approximately I s at 1. 85 m' /s
Average wa ter tempera ture 20· C
 BASIC APPROACH 91
90 DES IGN OF BAS IC TREATMENT PROCESS UN ITS

Determin e SO llllioll (i) Provide a l1ash mi xing tank that is 10 ft (3 m) square and 13.5 ft
(4 . 1 m) deep at the head of the plant. An intermediate slab containing a 4.5- ft
(i) The size of the tank (1.37-m)-diameter hole at midheight should be provided to create an elTective nash
(ii) The horsepower of the mi xer and type of mix ing blades Illi xi ng zo ne. Attach a 4-ft ( 1.22- m) length of pipe (Figure 3.2.3-8) at the ope ning
to act as a mixi ng reactor tube. Flow veloc it y through the hole is
(iii) The application poi nt of aluill
(iv) The size of the inlet c hannel
1.85 -i- 1T X (J.37)' = 1.25 m/s
4

7.----------------------------------, and the mixi ng ti me is

1.22 m -i- 1.25 mls = 0.98 s

pH 7.8
3,---------------------------------
6 ~~--------------------------------~ [J Natural pH
~pH6.5
~pH 6 . 0
~ pH 5.5

5 7 .1

--'"
::;-
E
:J' 2

-
u
a,
E
u
0 ~
~
< \-- - - - - '"""' -\,-----~:-----------~
..
co
.,a:"
/(' "C
Cl
<:1,< .
·iii
0", .
~------~/~ ~------~~~---__1
_ _ _ _ _ _ _ _ _ _ _ pH 5.9

J~----------------~~----------~~~

20 30 40 60
o 20 37 <0 60
Ferric chloride dosage (mg/L)
Coagulant dose (mg/L)
Figure 3.2.3-7 Effect of pH during coagu lation process on TOe rcmova l.
Figure 3.2.3-6 Example of enh anced coagulat ion j ar test result.
92 DESIGN OF BASIC TREATMENT PROCESS UNITS
BASIC APPROACH 93

(ii) The power requirement for the mixer horsepower and mixing blades is P ==

~
C'Vflo. When C = lO00s - '.

~Slalor
V= 'IT X (1.37)' X 1.22 = 1.8 m'
4
baffle Assume that the minimum water temperature is 50C.
Enlarged Detail
flo = 0.0015 . Ns/m' al 5°C
4.5' Diameter hole
for flash mixer p = (I (00)' X 1.8 X (1.5 X 10- ') = 2700 joules (J)/s
10'
Inlet channel
Since I kW is 10(0)ls.

P = 2.7 kW or 3.6 hp

Assuming an overall efficiency of 80%,

13.3' 13.3' 6.7'
P = 3.6 -;- 0.8 = 4.5, say, 5 hp
Flocculation tank
40' 10' 40' which is the minimum required mixer horsepower.

A~ Utility tunnel Four 45 pilehed blade lurbines (PBTs) wilh an overall impeller diameter of ap-
proximately 1.5 ft (0.76 m) should be selccled for Ihe mixing blades. The blades are
~ (option)
located at the bottom of a 4.5-ft (I A-m)-diameter flash mixing pipe. TIlC inlerior of the
(a) mixing reactor tube must have four vertical stator baffies, each approximately 3 in. (75
mm) in width. The mixer shaft must rotate in such a manner that the water is lifted,
Alum_ minimizing the loss of head through the flash mixing system; shaft runout should be
limited to ~ in. (3 mm) and an underwater shaft bearing should nol be provided at the
end of the shali. TIle details of Ihe mixer design, including shaft rotation speed, should
be refelTed to a reputable mixer manufacturer (based on the foregoing criteria).
(iii) The alum feed line is located below the mixing blade as shown in Figure
3.23-8. To prevent clogging of the feed line, employ a solution with a concentra-
tion greater than 5%. An additional alum feed line may be included at the entrance
Primary reed line
of the raw water line into the tank; thi s acts as an alternative application point should
the main feed line become clogged or damaged. The end of the main chemical feed
Standby reed line line has a simple 45° cut (Type D, Figure 3.2.3-~) to minimize clogging. The end
of the auxiliary feed line is equipped with a diffuser pipe across the inlet pipe (Type
B, Figure ~.2.3-3).
(iv) Assume that there are two Ilocculation tanks and that Ilash mixing occurs
in the center of the two tanks. An influent channel is located between the flash mixer
lank and the flocculation tanks as shown in Figure ~.2.3-8. To assess the size of the
influent channel leading to the Ilocculation tanks, the rul e of thumb is to maintain
less than 0.3 in. (7.5 mm) of the velocity head exisling within Ihe influenl channel
1~ 7'
.~I----~--_+--~-- When the system undergoes maximum flow rate. This practice ensures that the wa-
(b) ter surface elevation remains fairly constant across the entire channel. It is impor-
tant to note that if one of the two flocculation tanks is isolated for maintenance or
Figure 3.2.3-8 Flash mixing tank and inlet channel to the Ilocculation tank.
repair, a flow of 1.85 m]/s may travel through one side of the channel.
94 DESIGN OF BASIC TREATMENT PROCE SS UN ITS BASIC APPROACH 95

Select a wid th of 7 ft (2. 1 m) fo r the midpoint of Ihe c han nel and 3 ft (0.9 m) The mixing tim e at the plant fl ow rate (I) is
for each end. The channels should not be designed with a wi dth of less th an 2.5 n
(0.75 111 ) since thi s causes difliculties in construction. The average water depth is I ~ 1. 18 m' + 1.5 m'/s ~ 0.79 s
assumed to be 7.5 ft (2.3 m).
(ii) T he req uired wa ler horsepower is fo und as fo ll ows:
LOCATI ON WITH IN T HE C HAN NEL p ~ G'2)J. V

Center of Item Entrance Midpo int Endpoilll )J. ~ 1.336 X 10- ' N . sim' at 10°C

Q 65 ft ' /s 43 ft'/s 2 1.7 ft' /s p ~ 750' X ( 1.336 X 10- ') X 1.18 ~ 887 Jls ~ 0.89 k W ~ 1.1 hI'
(1. 85 n,'/s) ( 1.23 n,'/s) (0.62 n,'/s)
(iii) To achieve sa ti sfac tory diffusion of the primary coagulant in a short period
A 52.5 ft' 37.5 ft' 22 .5 ft'
of tim e, the ru le of thumb is to des ign the pump fl ow rate to be 2 to 5% of the plant
u 1.24 rtls 1.1 5 ft /s 0.96 ft/s fl ow rate:
u' l2g 0.024 rt 0.02 ft 0.01 4 ft
(7.3 mm ) (6.1 mm ) (4.3 mm ) 1.5 m' ls X 0.03 ~ 0.045 m'ls or 71 0 gpm

The maximum velocity head of the channel is 0.29 in . (7 .3 ITIm). T herefore, the rul c- (iv) The veloc ity head of the Il ow issued fro m Ihe blendin g nozzle is used in
of-thumb crit eri on is met. When th e two Ilocculation tanks arc on- line, th e fl ow ve- computing the mix ing energy, The requ ired mi xing jet velocity is calculated by lIS-
loci ty at the cha nne l e ntrance is 0 .62 ft /s a nd the velocil y head is 0.07 in . ( 1.8 mm); ing the basic pum p formu la:
the in let condition to the flocculati on tank is good.
k W ~ 9.8 1 X Q X H
E xa m ple 2 Pump Diffusio u as t he Fla sh Mixer e
Given where kW i,s the hors;power of the pump (i,n thi s case it is water horsepower),
Q Ihe pump ing ra le (nr Is), H the total dynamiC head (m), and e the effi Cie ncy- in
Plant Il ow ralC 1.5 m' ls (34 mgd) max imum per day thi s case 100% since the theoreti cal water horsepower is used 10 determin e nozzle
G X I A pprox imate ly 1000 (G ~ 750 s - ' ) size, The mi xing jet velocity is based on the water horsepower required to produce
Pipe diameter 1000 mm (40 in. ) the given mix in g e ne rgy of 0 .89 kW and a pumping ca pac ity of 45 Lis. There fore,
Diffusion jet ve loc ity 6 to 7.5 mls (20 10 25 fils)
Minimum wa ler temperature 10°C 0. 89 ~ 9.8 1 X 0.045 X H
A lum dosage 8 to 50 mg/L
H ~ 2 m or 6.6 fI
Determ in e
Since the vel oc ity head H ~ u' /2g,
(i) T he mi xing time
u ~ 12 X (2 X 9.8 1)J O.5 ~ 6.26 ml s or 20 fi ls
(ii) Th e water horsepower requi red
(iii ) The pump capaci ty Multiple nozzles prov ide better e ffic iency of mix ing. However, du e to c logging prob-
(iv) The jet velocity and size of orifice required lems caused by debri s, use on ly one nozzle Ihal sprays a full cone pallcrn of 90°,
(v) The type of nozzle The diamete r of the o rifice is 3.75 in:
(v i) T he horsepower of the pump molor
Q )0.5 ( 0.045 )0.5
(v ii ) T he locati o n of Ihe alum feed li ne d ~ ( 0 .785 X u ~ 0 .785 X 6.26 ~ 0.095 m

SO iulioll (i) T he assumed le ngth (L ) o f the mix in g zo ne is

(v) Model 8 HF700, a full jet nozz le, manu faclu rcd by Spray ing Syste m
L ~ I .5D ~ 1.5 X I ~ 1.5 m Company, is selec ted . Th is nozzle has 92° ful l co ne spray ing, type J 16 sta inless
steel, construction, a 3,6-in , (92-l11m) nozzle diameter, a minimum free passage of
V ~ 0.785 X D' X L ~ 0.785 X I' X 1. 5 ~ 1.1 8 m3 1.9 111 . (48 mm) solids, 8-in . (200-mm ) Ila nge pipe co nnec li on, and a 700-g pm
96 DESIGN OF BASIC TREATMENT PROCESS UN ITS BASIC APPROACH 97

(0.044-m'/s) capacity at 7 psi (0.5 kg/c m' ) back pressure. Figure 3.2.3-9 illustrates Pump
this system and the nozzle.
We now check the energy input :

P(hp) = (0.97Cd au') -;- 550

where P is the wa ter horsepowe r, ell the coe flicient of discharge = 0.75, a the ori -
--
lice area ( fI'), and u the jet velocit y (fi /s).

P = [0.97 X 0.75 X 0.07 X (22.3)'] -;- 550 = 1.02 hp

Based on the G value, the required horsepower is computed to be 1.1 hI'. The en-
ergy input is a little less but nonetheless acceptable.
(vi) We now determine the pump motor horsepower:

TDH = required pressure at nozzle + ex it loss of jet wa ter + pipe fricti on loss fUll CONE
$nAT 'ATTEIN

NOTE: The stati c head for th e pump is ze ro due to the siphon effec t (see Figure
3.2.3-9), and the pipe fric ti on loss is ass um ed to be I fI, but the actu al value should
be calcu lated. Th e exit loss of spray jet may be added but not incl udcd here. FUll CONE
SPRAY PATTERN
TDH = (7 X 2.3 1) + [I X (22.3' -;- 64.4)] + I = 25 fI or 7.6 m

BHP = Q XH 700 X 25 = 5.2 hp = 3.88 kW

3960 X e 3960 X 0.85

Therefore , use a 5-hp (3. 7-kW) motor.

(vi i) We now determine the locati on of the alu m feed line.

FEED FLOW RATE OF ALUM SOLUT ION

Ma ximum = 50 mglL X 8.34 X 34 mgd = 14, 178 Ib/day

Average = 15 mg/L X 8.34 X (34 -;- 1.5) = 2835 Ib/day
Minimum = 8 mg/L X 8.34 X [34/( 1.5 X 4)] = 378 Ib/day
Assume that Qmal\ = 1.5 X Qa\'~ and Qmin = 0.25 x Qa\'c' Si nce commercial liquid
alum contains 5.4 Ib of dry alum per ga llon, the actual feed now rate is 2625, 525, Figure 3.2.3-9 Pump diffu sion nash mixi ng system and nozzle.
and 70 gal/day for the maximum , average, and minimulll now rates, respecti vely.

( I) (2) For most cases the strength of the alum solut ion should be greater than I % up-
Alum Feed PUlllpi ng Col. ( I) Alum stream of th e injection nozzle, and the pH of th e solution shou ld be less than ~ .O.
Rate Ratc -;- (2) Solution If these recom mendations are followed, th e nozzle and pipe may not become clogged
Dosage (gpm) (gpm) X 100 Strengt h (%) by alumi nulll hydrox ide scaling. Ex perience indica tes that if th ese conditions arc
not met, heavy cak ing by aluill hydroxides begins approxi mate ly 10 to 12 in. (0.25
Maximulll 1.82 700 0.26 0.13 to 0.3 ~) downstream of where the alum feed line term inates into the pump dis-
Average 0.36 700 0 .05 0.026 charge hne. Therefore, the end of the alum feed li ne should be located at the Ilange
M in i1lllll11 0.05 700 0 .007 0.0035 of the injecti on nozzle. Altern ati vely, th e feed line exit may be located just olltside
98 DESIGN OF BASIC TR EATMENT PROCESS UNITS
BASIC APPROACH 99

the nozzle. However, the dispersion of alum is not as clTcctivc as the first choice. Based on the manufac turer's information, a slatic mixer with two mixing elements
Note that sludge caking in the discharge line clcarly confirms that alum coagulation wou ld give the fo llowing results:
is completed within 0. 1 seconds after application.
Flow Rate Head Loss G at 10°C t
Exa m ple 3 Static Mixer as t he Flas h Mixer (mgd) (ft) (5 - 1) (5) G Xt
Given 5 0.17 346 3.2 1107
10 1.34 1340 1.6 2144
Plant flow rate 15 mgd (0.63 m'ls) ma ximum per day 15 3.0 2520 1.1 2770
G X t Approxima tely 1500
Pipe diameter 24 in . (600 mm) However, the desired G X t va lue should be below 2000 at the maximum plant
Maximum head loss 1.5 ft at 15 mgd flow rate and approx imately J 300 with an average flow rate. For th is reason, the
M inimum water temperature 10°C pressu re drop characteristic of the clement should be based on the desired G X t
values; the manufacturer can reduce the pressure drop across each cl ement by re-
Determin e moving a portion of the barne plates. The foll owing tabl e demonstrates how the de-
sired pressure drop is achieved.
(i) The type of in-li ne static mixer
(ii) The number of mixing elements Approx imate
Headloss
(ii i) The energy input (G X t)
Across
Flow Rate Mixer Gat 10°C t
SO/lilioll (i) Several types of in-line stalic mixers are available for water treatm ent
(mgd) (rt) (5 - ') (s) G Xt
application, but a simple unit with low-clogging characteri stics should be selec ted.
A provision for easy disasse mbly of the mi xing baflles (for cleani ng) must al so be 5 0.028 140 3.2 450
considered. Therefore, a motionless mixer, such as the one manufac tured by Koma x 10 0.46 806 1.6 1290
Systems, Inc., is preferred. 15 1.5 1782 1. 1 1960
(ii) Under normal flow conditions the practi cal mixing tim e is limited to ap-
proximately 2 to 3 s for pipes with d iameters larger than 5 rt ( 1.5 m). Thus, the to- Exa mple 4 C hemical Ap plication
tal length of the mixing unit can be approx imatcd by two timcs the fl ow velocity in
the pipeline. The flow veloci ty for a plant influent line generally ran ges from 6 to Given The foll owing water treatment chem icals are selected to be fed at the head
8 rtls (1.8 to 2.4 m/s). The total mixer length therefore ranges from 12 to 16 ft (3.6 of the plant: alum, cationic polymers, anionic polymers, nonionic polymers, sodium
to 4.9 Ill). A rule of thumb in estimating the length of one clement is to designate hydroxide, chlorine, ammonia, PAC, and potassium permanganate.
the leng th as 1.5 to 2.5 times the pipe diameter. Based on this c riteri on, the length
of one element is in th e range 3 to 5 ft (0.9 to 1.5 m). Thus two or three elements Determine
may be used.
(iii ) We now determine the energy input : (i) The chemical application seq uence
(ii) The number of pote ntia l app lication points
P )0.5 P = QIV"
(iii) The type of mixing sys tem
G = ( fLV and
~ol~ilion (i) Chemicals that potentially interact if fed together are alum and an-
where P is the power input (lb ' Ibis), Q the 110w rate ( rt'/s), IV the uni t weight of l?I1iC polymers, alum and sodium hydroxide, cat ionic and anionic polymers, chlo-
water (62.4 Ibm' ), and" the press ure drop (rt). If the total le ngth of the mi xing unit fine and PAC, ammonia with chlorine and alum, and chlorine with cerlain cationic
is assumed to be 8 ft (2.4 m), the vol ume of the unit is computed by the following polymers. Alum requires instantaneous dispersion . PAC and potassiulll perman-
formu la: Assume a 24-in . (0.6-m) pipeline diameter: ganate must be fed ahead of the coagulant to enhance their effectiveness.
(ii) Based on the various considerations discussed earlier the scheme shown in
[0.785 X (2)2] X 8 = 25. 12 rt' Figure 3.2.3- 10 is established. Ideally, each application poi~t should be separated
 BASIC APPROACH 101
100 DESIGN OF BASIC TREATMENT PROCESS UNITS

The second topic that must be stressed is the constant monitoring of the chemi-
cal feed a nd flash mixing systems. The most common water treatment problems
arise from the clogging of chemical feed lines and diffusers. Malfunctioning of me-
chanical mixers can easil y be detected by the plant operator. However, clogged sta-
_ o
'2 tic mixers, broken underwater chem ical feed lines, and clogged diffuser orifices are
~
N
e.g
, ~ not as easily detected unless the water is drained. Thus, scheduled plant shutdowns
O'<u must be exercised during periods of low water demand.
I
3-5 min 3-5 min The third consideration in proper operation and maintenance is adjustment of the
I- -I 14 -I mixing energy. Mechanical mixers and pump diffusion flash mixers must have their
) mixing energy altered with seasonal changes in raw water quality. Cold water is
Point no. t Point no. 2 ~ I Point no. 3
more viscous and theoretically requires a higher energy input to achieve the same
Diffusers to Instant mixing Hydraulic mixing mixing effect as obtained during summer months. Extra power is also required to
disperse chemicals by flash mixer or mechanical flash
disperse coagulant when processi ng raw water that is highly turbid or colored.
mixer
Treatment plants that are equipped with mechanical flash mixers sho ul d have the
Figure 3.2.3·10 Scheme for Example 4. following items inspected on a regular basis: the motor and speed reducer, mixer
shaft runout, vibrations, and noises. The lubrication oil should also be changed at
by a time lag of at least 5 min, but this is often diflicu lt to attain. The first appli - certain time in terva ls, or by seasons, based on the recommendations of the manu-
cation point should therefore be as far upstream from the plant as practical. The sec- facturer. It must be emphasized that coagulat ion, flocculation, and clarification
ond point is located at the head of the plant. It is also difficult to place the third ap- processes are pretreatments to filtration. Plant operators generally tend to produce
plication point so that it meets the time lag criterion; thu s, it shou ld be situated as large flocs and attempt to make the settled water as clear as possible by app lying
far as possible from the second point. Anionic and non ionic polymers are good aids rather large amounts of coagulant. Consequently, the effect of the flash mixer is of-
in floc formation if added when pinpoi nt flocs are formed by aluill. This process ten undetected and therefore turned off.
generally requ ires 5 min. For this reason , the appl icat ion points for anionic and non - The role of the flash mixer mayor may not be evident, depending on the type
ionic polymers are best located in the second stage of flocculation. of process. The conventional water treatment process uses a large amount of coag-
(iii) As illustrated in Figure 3.2.3- 10, there arc three application points: ulant that is more readily dispersed, forming large floes and clear settled water.
However, most modern high-rate filtration processes use 20 to 30% less alum than
Point J,' dispersion of chemical by d iffusers the conventional dosage; thus, vigorous flash mixing is essential for instant disper-
Poi1ll 2: instant mixing by static mixers or pump diffusion sion. Since direct filtration processes use such small amounts of alum (4 to 6 mg/L)
Poi1ll 3: mixing by mechanical or hydraulic mixer compared to the conventional processes, they produce much smaller floc and high
settled water turbidity. Most modern high-rate filters are capable of liltering sett led
Operation and Maintenance The most important operation and maintenance water with turbidity greater than I to 2 ntu, and in many cases 3 to 6 ntu is con-
aspect of the initial flash mixing process is the selection of the proper chemicals sidered to be optimum. The primary objective of flash mixing is to effectively co-
and the application of the correct quantities to an ever-changing raw water quality agu late colloidal matter present in raw water by applying the proper amount of co-
and plant flow rate. Jar tests have been used effectively in optimizing coagulant agu lant through good flash mixing prior to the flocculation process. This is
dosage for conventional processes. However, they do not appear to supply the plant particularly important for direct and in-line filtration processes.
operator with the proper coagulant dosage for direct filtration and in-line filtration
processes- hence the development of glass tube pilot Iilters. Although the glass tube Coagulation Control Test The jar test is the most widely practiced and proven
pilot Iilter is widely used, approximate ly 50% of plants equipped with this coagu- method of evaluating, as well as controlling, the coagulation, flocculation, and clar-
lant control system do not use it because the system requires a great deal of oper- ification processes. Although widely used, jar tests provide the operator with little
information as to the mode of filtration (direct or in-line filtration) or whether the
ator judgment and attention.
Zeta-potential measurements have also been tried , with some success. In recent tests are being performed improperly. When designed and performed correct ly, jar
years, streaming current detector (SCD) units have improved on the zeta potential tests yield very meaningful data that ca n ass ist engineers in designing new treat-
measurement and appear to be successful in coagu lant control. SCD units can be ment plants or in modifying existing plants. Plant operators also benefit from jar
used for any type of water treatment process lIsing coagulant. Yet it sti ll requires a tests since they will be able to optimize the efficiency of the treatment process. The
jar test procedure recommended is presented in Appendix 10.
significant amount of operator attention.
102 DESIGN OF BASIC TREATMENT PROCESS UN ITS
Jar tests have traditionally been perform ed under fixed test conditions to deter-
mine the relation between coagul ant dosage and settled water turbidity. However, it
is ve ry import ant to evaluate the IiltcrabiJity of the settled water by filter paper fil-
trati on (simulated IiIler) since watc r with the lowest turbidity is not necessarily the
best tilter intlucnt with respec t to lilter efficiency.
There arc livc basic objectives for the jar test: (1) optimization of coagulant,
(2) optimi zation of the chc;nical ap plicati on sequ ence, (3) optimi zati on of the mix-
ing energy and lime (G x I), (4) evaluation o f clarifier and filtration performance,
and (5) eva luation of corrosive c haracteri stics of settled water.
OPTIMIZATION OF COAG ULANT To optimi ze the coagulation process, operators must
choose the most appropriate type of coagulant (alum, ferric chloride, cationic poly-
mers) and evaluate the optimum dosage under standard mi xing conditions (to be de-
scribed later). Coagulant solution strengths should also be considered since they may
affcct the decis io n. For example, by using a combination of alum and cati onic poly-
mers, the amollnt of sludge prod uction is o ften reduced. However, cationic poly-
mers cost 6 to 10 times more th an alum, and the polymer dosage is therefore usu-
all y limited to 2 mg/L as a matter of economics.
OPTIM IZATI ON OF T HE CHEM ICA L APPLICATION SEQUENCE Optimizati on of the chem-
ical application sequence is a very important consideration, yet little attention has
been give n to thi s matter. For instance, the sequ ence of appli cati on is extremely im-
port an t when using alkali chemicals and anionic or nonionic polymers (i n conjunc-
ti on with alum) as Ilocculation aids. As desc ribed earlier, these polymers help pro-
duce excellent noc onl y if they are fed a few minutes after alum is added to the
process water.
Ol71'1MIZATI ON OF MIX ING ENERGY, MI XING PA1TERN, AN D M IXING T IM E The optimum
mi xing condition depends on the type of overall treatment process train , the type .of
filter bed of the plant, and the raw water charac teristics. In general , a tapered mI x-
ing-the use of hi gh-intensit y mi xing at the beginning, followed by a gradual reduc-
ti on toward the end- helps promote better tl occul ation und er any condition. Under
ac tu al plant cond itions, it is comlllon for the convent io nal treatment process to have
a velocity gradi ent path of 50 S- I at the beginning and 10 S- I at the end of 30 to 40
min of mi xing: G X t = 4 X 104 to 5 X 104 . For the direct liltration process, the ac-
cepted range for the velocity gradi ent is from 65 to 15 S- I; for 15 to 30 min of mix-
ing, G X t = 4 X 104 to 6 X 10"' These ran ges should be kept in mind when con-
ducting the jar test. Due to excessive shear force, Iloccul ati on times that are too long
(over 45 min ) generall y produce Iloc types that are less than optimum for settling.
EVALUAT ION OF CLAR IFtER AND FI LTRATI ON PERFOR MA NCE The design criteria for new
sedimentation tanks or the performance of ex isting clarilier units may be evaluated
by measuring the noe settling velocit y through use of the jar test. A settling veloc-
it y of 40 mm/min ( 1.5 in .lmin ), whieh is the max imum load ing rate of ordinary sed-
b
BASIC APPROACH 103
imentati on ta~'k s, for alul11 noe durin g the Slllllmer months. if the settling ve locity
of alum Il ?~ I S ll1~c h slower than the corres pondin g surface loading rate of the ex-
istin g c1anller UIlII , the noc characteristi cs may be improved by either adding an-
ionic polymer (0. 15 to 0.3 mg/L) to the alull1 fl occulation process or by substitut-
ing ternc chlonde tor alum ; these must be evaluated by a se ri es of jar tests. Refer
to Camp (1946), Hudso n ( 198 1), and Sing ley ( 1981 ) for di sc ussion s of fl oc sCllling
ve locity meas urement.
EVA L UATION O F CORROS I VE CHARACTERI STICS OF SETTL ED WATER II is important to
evaluate ~he corrosivity of the settled water. Regul ar inorga ni c coagu lan t will reduce
the pH 01 th e settled ~ater, thereby decreasing its alkalinity. Filter performance may
be ev~luated by passing. the supernatant throu gh What man No. I filt er paper. The
best filter pret reatment I S one that res ults in filtrate turbidity at or below the tur-
bidity goal (0. 1 to 0.2 ntu) within the shortest filtrati o n till1e.
As a rule of thumb, the following items should be recorded durin g the jar tes t:
I. Time at which pinpoi nt 11 0c appears
2. Si ze of the floc at the end of the fl occulation period
3. Turbidity, color (optional), pH, alkalinity, and temperature of the raw wa ter
and of selected settled water
4. Turbidity of tiltered water and the time required to tilter 50 mL of selected
settled water
Figure A I 0-2 ill Appendi x 10 prese nts a scale for noc size idcntitication. Table
A 10-1 is an example of jar test results.
BIBLIOGRAPHY
Amirtharajah. A.. and Mi ll s. K. M.. "Rapid Mixing Des ign for Mec hani sm of Alum
Coagulation," J. AWWA, 74:4:2 10 (April 1982).
ASCE, AWWA . and CSSE. \ValeI' Treallllenl Planl Design, American Water Work s
AsSocialion. New York . 1969. 1990. 1997.
AWWA Researc h Foundalion Report . Seleclion and Desigll oj Mixing Process for
Coagll/atiOI/, American WUlel' Works Association. Denve r. CO. 1994.
Ca mp. T. R.. "Scdimentation and the Design of Settling Tanks." ASCE TrailS., 1:895 (1946).
Chao, J. L. , and Stone. B. G .. " In itial Mixing by JCllnjecti on Blending." 1. AWWA, 71 : 10:570
(Oelober 1979).
Clark , M. M.. el a!.. Selec fion and Design of Mixing Process for Coagularion, AWWA
Research Fou ndat ion. American WaleI' Work s Associati on. Denve r. CO. 1994.
Culp, W. , and Culp. R. L. , Hal/dbook of Public WaleI' Syslellls, Van Noslrancl Reinh old. New
York, 1986.
Fair, G. M.. et al.. Elemel1ls of lii'tler Supply (llld Disposal, Wi ley. New York. 197 1.
104 DESIGN OF BASIC TREATMENT PROCESS UNITS
Guven. 0 .. and Benefield. L.. "The Design of In-Li ne Jet Injection Blenders." 1. AWlVA.
75:7:357 (Jul y 1983).
Hudson. H. E. Jr.. Waler Clarijicalion Process. Van Nostrand Rei nh old. New York. 1981.
Kaufman. W. J., et al.. Initial Mixing {/lui Coagll fcllioll Process, Repo rt 72-2. Uni versit y of
Cali fornia. Berkeley, CA. February 1972.
Kawamura , S., "Coagulation Considerati ons," 1. AlVWA. 65:6:4 17 (J une 1973).
Kawa mura. S .• "Considerati ons for Imp rovi ng Flocculation." 1. AWWA, 68:6:328 (June 1976).
Kawamura, S .• "Optimi zat ion of Basic Water Treatment Processes- Design and Operation:
Coagulati on and Flocculati on," 1. Wale,. SRT-Aqlla, 45:1:35-47 ( 1996).
Montgomery. J. M.. Consulting Engineers, WaleI' Trealmel/I: Principles al1d Desigll, Wiley.
New York. 1985.
Sa nks. R. L.. Wafer Trealmellf Plal1t Design for Practicing Engineers, Ann Arbor Science.
Ann Arbor. MI. 1979.
Si ngley. J. E.. "Coagulation Cont rol Us ing Jar Tests." AIVIVA Proceedillgs. No. 20 155,
Ame rican Water Works Association. Denver, CO. 1981.
Stumm , W.. and O·Melia. C. R.. "Stoichi ometry of Coagulation ," 1. AWWA. 60:5:514 (May
1968).
Weber. W. 1.. Physicochemical Pmcesses, Wiley-Interscicnce. New York. 1972.
3.2.4 Flocculation Process
Purpose Flocc ulati on is the ge ntl e mi xin g phase th at fo ll ows the rap id disper-
sion of coag ul ant by the !l ash mi xi ng unit. Its purpose is to accelerate the rate of
parti cle collisio ns, causin g th e agglomeratio n of e lec trol yticall y des tabili zed col-
lo idal particlcs into settl eab le and filterable sizes. . noec ul a" o n to produce good se ttleable floc with a filter innue nt turbidity of less
The term s coaglllation and floCCIl /(I tioll are sometimes used interchangeably In
tec hnic al literature. However, the aggregati o n of pa rti culate mate ri al is ac tu all y a
two-step process. The initial step involves the additi o n of coagulant , to reduce or
e liminate the interparticul ate fo rces respo nsible for the stability of the particulates.
This process is call ed coagulation. The subseque nt particulate collisions a nd en-
meshment of particulates into fl ocs occ ur as the resu lt of molec ular moti on and the
physical mi xing of the liquid. This isjloccllialioll.
Considerations The following factors mus t be taken into cons iderat ion when
designing a proper fl occ ul ation process: ( t ) raw water quality and fl occ ul at ion c har-
acteristics, (2) treatm ent process and finished water quality goa ls, (3) avai lable hy-
draulic head loss and plan t fl ow variations, (4) c hoice of Ilocc ulati on tank inlets,
(5) local condi ti o ns, (6) cost, (7) relation to existi ng treatmen t facilities , and (8) mis-
ce ll a neous items. These items are discussed in detail.
Raw Water Quality and Flocculation Characteristics These are the lirst con-
sideration s when designing a fl occulati on facility. T he desi gn eng ineer s ho uld ide-
all y have access to data on the year-round quality of the raw water, spa nning ~ ~e­
rioel of 5 years or eve n lo nger. 11le seven important water qu al it y aspects are turbidity,
BAStC APPROACH 105
IOta I orga nic compound s, pH, a lk alinity, color, algae cou nt s, and temperature. T he
natu re of the colloids, particularly the co ll o id al organic compound s, and the parti-
cle size distribution of the turbidity are c harac teri stics that are preferably known
since these factors grea tl y affec t the fl occ ulat io n characteri stics. The magn itude of
the raw water turbidity alo ne is not li kely to be a su rrogate to coag ul ant dosage re-
quirements (sec Figu res ~.2.4- 1 and ~.2.4-2).
If hi storic water quality data arc no t available, an imlllediate sampling should be
establis hed prior to or as c lose as possible to the beginning of the predesign study
period. Data o n the seve n bas ic water quality parameters should the n be collec ted for
a period of I year. However, if thi s sched ule ca nn ot be implemented, a n atte mpt
should be made to at least obtai n the water quality data in both dry and wet seasons.
FI~cul.ation charac teri stics may be evaluated by the jar test, a process com mo nly
pracliced In water treatmen t plants. The jar test must be conducted in a systemat ic
fash ion so that it will yie ld the necessary basic data for the desig n of the fl occul at io n
process, as discussed in Section ~.2'). The duration of the test should be as long as
can be afforded. Yet, as a minimum, repeatability of the results mu st be demon strat ed.
Treatment Process and Finished Water Quality Goals The overa ll trea tment
process and fini shed water goals are the second most important considera ti ons be-
cause both affect Il occ ulati o n. For insta nce, in a di rec t filtration process, the fl oc-
~u lati (~n tank s s hou ld not produce large se ttl eab le fl ocs since a sed imen tat io n process
IS not Involved. Instead, the flocculation basins sho uld produce an appropriate type
of noe so that in-depth lilt rati o n can be achi eved by a reverse-graded filter bed . In
contrast, . the co nve nti o nal co mplete and lime softe nin g processes require thorough
than :\ n~u , regardless of the raw water quality. This is particularly true for si ngle
sand rapid filters th at are characteri zed by surface filtrati o n.
The fini shed water qu ality goa ls also innuence the deg ree of fl occu lati o n.
Excessive aJ~ounts ?f the fo ll owing substa nces in the raw water may effectively be
removed by Improvlllg the Il occ ulati o n and sedim en tati on steps of the process train:
color as a prec ursor to DBPs, algae as a source of taste and odor asbestos fibers
(limitat io ns), and certain tox ic metals a nd compou nd s. '
Available Hydraulic Headloss and Plant Flow Variations Other co nsidera ti ons
In the design of the Ilocc ul ati o n process are th e allowable head loss across the floc-
cul ation process and the magnitude of plant fl ow variation. If the a ll owable head-
loss i~ limited, hyd ra uli c fl occ ulati o n methods are rul ed o ut. Thus, the design e ng i-
nee.r IS forced to use mechanical methods. S imilarl y, plant flow rate fluc tuation is
an Importan t co nsiderat io n. If the fluc tu ati on is re lative ly min or yea r-round that is
50o/t ..•. . ' ,
o v.a n atl on lrom the dally average tl ow ra te, hydraulic fl occ ul at io n is appl icable
and will perform e ffect ively.
Ch . .
olce of FlocculatIon Tank Inlets Almost all proprieta ry reactor-c1ariliers are
deSigned .h . I . . .
wit a slllg e pipe Inl et to the flocc ul atIOn zo ne. Yet most muni c ipa l wa-
 BASIC APPROACH 107
Kaolinite
100 /...- ....... - 1',.,.....-..0 zs Group A Croup 8
./' /

~,d'
,P'"c
%.,
Y.."
. /
/
~~~
ZO
\ \ .. ,
\
"---.tt
!\.
pH : \ &/
J!!{
.. .
\,".
-:.~~
"
,-0
.....
\ 'to
-;;
\ 00

.
\".\ 0;.',..
-;
ii ! \A\~
I 9 / o<eo" ' earth 9..
\
10 10

1 ...
0\<:\\0':,. _ - - -
J ,,-- ~\\\
~
\\-,;~ ~

~•
\ .
.\ .I ~
. .. --"l""V 10
., '\f' \ .~.
\ i! \~ \....,..... /
"'1=.' \ f-I •
•
0: 3\\~ ,~
\'\ I , •.,.,." / \
5
\b ~'.... .........
"-
.......,
,\
\
pH . 7.25
•oJijI"

__i·\~-~~":'~P~H~:~7.~2:5~~~
oL 10 20
".."
.,~ \

· ·~P~H~6l.~~~~~~~~~·~'~~~F=~'~~
30 40 50 60 70 80
I
6 pH : 6 .0

0l-------Z~0~------.~0-------6~0~----~8~0-------IOrO~----~1~0------,~40
..... ---. • ____
.
Atum dosage (mg/l ) Alum Douce (mg/l )

Turbidity Alkalinity
pH meq/L
NTU
0.4 Raw Water
Bentonite 75 8.3
75 9.0 0.4 Group A Group B
Diatomaceous
Kaolinite 75 8.7 0.4 Turbidity JOWTU 165 N'TU

Topsoil 75 6.7 0.4 Alkalinity 0.64 mcq/l 1.0 mcq/l

OJentangy RW. 65 8.02 3.26 pH 6.5 9.'
(a)
Figure 3.2.4-2 Effect of particle size on Il occulation. (From S. Kawamura . Thesis. Ohio
State University. 196 1.)
Raw wate r: Temperature = 3-270 C,
pH = 6.8-7 .2, Alkalinity = 22 - 35 mg/L
COD (KMnO. method) = 2.2- 22.8 mg/L ter treatment plants have a tradi ti onal rectangu lar fl occulatio n tank loca ted upstream
30
..'". ... of the sed imentati o n tanks or tlotation tanks (DAF system). In thi s case the mos t

.. ..........
.. ...
26 simple and cost-e ffec ti ve design is a si ng le inlet to each flocculation tank. This
'iiO
.s 22
...._............... ..... sc heme ca n be used if the flow pattern in lhe flocculation tank is a serpentine type
""~ 18 '" ............ ...
0 14 ..............
_..........
........ .. • (see Figure 3.2.4-8). However. many desig n engi neers provide more than two inlets
to eac h flocc ul ati on tank from the influe nt channel or pipe. Regardless of the num -
"E,
«
10
6
..• . ber of inle ts to eac h tank , th e most important design issue is eq ual hydra uli c load-
ing to each lank.
2 150 200 Eng ineers have two basic cho ices whe n designin g the tank inlet: a se ries of iden-
0 50 100
Raw water turbidity (T U)
tical individual weirs set at the sa me elevatio n or a series of s ubm crged or ifi ces each
fitted wit h an isolation va lve or gate. The weir inlet is attrac ti ve from a theoret ical
(b) viewpoint, but there are four basic problems associated with thi s c hoice: ( I) the in-
Figure 3.2.4.1 «(I) Natu re of turbid ity and allllll dosages on Il occu lation. [Adapted t:r~m let channe l must be largc eno ugh to ma intai n nearl y equa l wa ter surface elevatio n
S. Kawamura. J. AWWA, 68(6):328 (J une 1976).} (b) Relation between raw water turbidity for lhe ent ire length of the cha nnel; (2) heavy suspe nded solids accuill ul aic in the
and optimum alum dosage forYodo River. Japan . I Adaptcd from S. Kawamura , J. )1'11. \VWA, channel due to the large size of the chan ne l a nd the low fl ow veloci ty w ithin the
239: 17 (September 1954).]

106

>
108 DESIGN OF BASIC TREATMENT PROCESS UN ITS

chan nel; (3) eac h lank is oflen iso hlled by means of slap logs which are nol leak-
li ght and create ex tra work for plant operators unless dow n-opening slide gates are
!~
used (down -openi ng slides do not rcliab ly reset at exactly th e same elevation after
.~ :r:: ~ :r:: ~ :r:: ~ ~ ­
M 'e;:j d ~0~~;;~
thcy arc rai sed for isolat ion); and (4) sli ght changes in water depth ( in the channel) ""on 00 ~,
o--...;t-o- -
..". 00 ('I Q\
M -
resull in signi lica nt changes in !low rate over the weir whenever all the weirs are vi v) 0\ 0\
set at the same elevatio n; some plants set the weirs at diffe ren t elevations along the
inlet c hann e l, bu t thi s is not rccomm ended, as the set e levati on is valid for onl y a
certa in plant !low rate.
The second choi ce in tank inlct design, submerged orifices fitted wit h isolation
va lves/valve operators or gates, arc furnished ncar the bottom elevation of the in- :t:~
_ cl'
flu en t channel. Compared to the weir inlet design , thi s type has two drawbacks, that '" co:
of a sli ghtly hi gher capital cost and sli ghtl y more loss of head. However, thi s de- :z:g-
0
sign has four adva ntages over weir inlets: ( I) the flow rate is no t sig nifican tl y af- ,...:
fec ted by sli ght changes in wa ter level from one end of the innue nt c ha nnel to the
other, (2) the size of the inllue nt cha nnel may be much smaller and a Ilow velocity
of 2 10 3 fils (0.6 10 0.9 m/s) may be allowed, (3) Ihere is minimum deposilion of
suspendcd solid s in thc in!lucnt channcl due to higher !low ve locity and because
the !l ow ex its ncar th e bottom of the channel, and (4) the isolatio n valves or gates
provide virtuall y driptight clos ure.
Sensiti vity of !low rate va ri atio n due to reduct ion or drop in wate r elevation in
the ill!luent channel are prese nted in Table 3.2.4-1, a compariso n of the resulls for
both weir inlct and submerged orilice inlet. As illustrated in the table (we ir inlet),
a drop of I in . (2.54 cm) in wate r level at the last inlet to the basin correlates to a
now ra le over Ihe weir of 83.7% and becomes onl y 66.3% wilh a 2-in . (5. I-em)
on
drop. In conlraSl , an 18- in . (45-cm) orifice yie lds a now ra le of 96.3% wilh a I- in . ,...:
drop and 92.5% with a 2-i n. drop in water level. In summary, the most practical II

(best) inlct from the influent cha nn el to the flocculation tank is the orifice inlet. :t:
_r-
i
'" ('i
",,0-
",, -
Local Conditions Local condili ons should al so be analyzed whe n selecling an ap- 0'
propriate type o r fl occ ul ati on process. The fi ve major factors th at should be evalu-
ated arc site topography, climatic conditions, availabi lit y of services, capability of .~
th e operatio nal person nel, and levcl of thc local techn ology. .~ ""<i
M

"" II
Cost Cos t is always an important co nsiderati on. Both the initial cost (capi ta l) and ~ . . . . :r:: ..--, ~ _

the operati o n and maintenance expenses should be evaluated. .... ~~~ .... ~
~8~8::8
- 0 - - 0 - 1.0 -
Relation to Exis ting Treatment Facilities In the case of planl expansion, Ihe re- o 0-0-0 -
latio n of the new !locc ul ati on process to the existin g process is a realisti c problem .
The basic rul e is to make all the !locc ul ation units identical so that uniform perfor-
mance and operat ion and maintc nance procedures ca n be maintained.

Miscellaneous Items Miscellancous co nsiderati ons. such as the hydrauli c char-

acteristics of the Il occu lat ion tank and sc um and sludge removal, must be incorpo-
rated dur ing the des ign phase. T he preventio n of !low short -circui ti ng, the creation
of e lTecti vc eddies a nd turbule nce wit hin the tank, and minimi zati on of hi gh shear
109
110 DESIGN OF BASIC TREATMENT PROCESS UN ITS
BASIC APPROACH 111

forces c reated by the mi xi ng blades arc all im portant hyd rauli c problems. Im port ant
operat ional problems inc lude the case of sc um removal from th e water surface a nd
of sludge a nd sill fro m Ihe bOllo m of Ihe basi n.
-
C I '7\
So me chemi sts also e mphasize the effec t of io n co nce ntrati ons, found within the 'ill (
raw water, on Il occ ul atio n as the salt leve ls o f the water inllucllce the coagulant
dosage require me nt s. Since the drin king wa ter q uality s tand ards for the TDS of raw
wa ter already lim its the presence of excess ive amount s of io ns, thi s issue is not of ..v () ~
great co ncern for most des ign e ng ineers. In rece nt years, the process of ozo nati o n
Rou n d~the·en d barned channel type ( plan)
has become popu lar d ue 10 Ihe slringenl limilali o n o n DBPs (THMs and HAAS) in Over· and undcrbamed chunnel type (section)
d rin king water. Ozone is capable of both oxid izing subs tances a nd fl occ ul ating cer~
tain orga ni c an d in orga nic substances. T herefore, the fl occ ul atio n process is gener- Baffled chann el floccula ti on system
all y more effec tive if it is preccded by ozonat ion.

Type and Selection Guide In Ih is secli o n we descr ibe Ihe co mm on ly avail ab le

altern ati ves to the fl occ ul ati o n process , as wel1 as gu ide li nes for mak ing the proper A
selecti ons.

Available Alternatives Flocc ul alion mi xin g may be prov ided by e ither mec han- / - ..... II ".- ...... II
II c> ,, .
ical mi xers or bafll es. T hese uni ts red uce n ow s h o rt ~circ uiti ng a nd ind uce the col-
I \ " I
I
\ I
II :.
\, I"\ II i~~i'.; :/It::
lisio n and agg lo meratio n of particles. Fig ure 3 . 2.4 ~ 3 prese nt s the 1110st com l11 on typcs
of mec han ical mi xer and bafll e used in the li cld of water treatme nt. The typical cat- " " '" :[:
. . ......
ego ri es of mi xin g sys tems are li sted below.
A
I . Mechani cal mi xing Section A~ A
(a) Verti cal s haft with turbi ne or prope ll er bl ades Hori zontal shaft with paddle whee l flocc ulat ion system
(b) Paddl e Iy pe wi lh e ilher hori zo nl al or verli cal shaft s
(c) Pro pri elary unil s suc h as Walking Beam, Flocs il alor, and NU-Tre,"
2. Baflled cha nn el bas ins
(a) Hor izon lall y ba m ed c hanne ls
FlOCCULA TJON B.... SIN
(b) Venica ll y bafll ed c hannel INLET CHANNEL FLOCCULATOR MOTOR
AND CEARBOX CT YP.J
3. Reactor clarilier propri etary systems
4. Co ni aci Il occ ul al io n (gravel- packed liller)
5. D iffused a ir or water jet ag it ation

Selection Criteria Seleclio n of Ih e Ilocc ul alion process s ho ul d be based o n Ihe

fo ll owi ng criteri a:

BAFFLE
I . Ty pe of treatment process: for exa mpl e , conve nti o nal, d issolved air fl o tation, WALL
d irec t filt ra tio n, softe ning, o r sludge co ndi tio ni ng Mixing Blade
2. Raw water qu ali ty: for exa mpl e. turbid ity, color, and te mperat ure
Vertical shaft with hydrofo il mixi ng bl ade flocculatio n system
3. Flocc ul ati o n characteris tics in response to c ha nges in mix ing inte nsit y and
mix ing times Figure 3.2 .4~3 Three hasic Iy pes or flocc ul ation systems.
112 DESIGN OF BASIC TREATMENT PROCESS UN ITS
The foll ow ing crit eria s hould be used when selecting the type of mixin g:
I. Loca l cond itions s lich as hi gh wind, ice buildup, or increased viscosi ty o f the
gear o il during extremely cold peri ods
2. Availabl e headloss across the plant
3. Shape and depth of the basin
4. Cap it al a nd operatio n and maintenance (O&M) cos ts
Order of Preference Based on the des ig n of the most recent installations, the or-
der of preference for the selection of the flocculation process equ ipment is as fol-
lows:
I. Vertical-shaft Ilocculators (w ith hydrofoil mi x in g blades) in either direc t !low
th rough horizontal tanks wi th proper co mpartme ntalization in eac h Il occ ul a-
lion stage or a serpentine fl ow path , unless the total number of the units be-
comes excessive
2. Paddle nocculators in hori zontal Ilow tanks with proper compartmentali za-
ti on in eac h stage
3. A co mbination of baflled c hannel s (he li co idal ll ow) in a se rpentine !l ow path
wi th nocculators in the las t s tage o f the !locc ulation process
4. Ba rn ed c hannel Ilocculation process (around-the-e nd type) provided th at the
mixin g c rite ria of G x t exceeds :1 x 104 during the low-plant -!low period
Reactor-clarifi ers and co ntac t !l occ ulation systems arc in a category of th eir ow n;
these are usually proprietary ite ms o f equ ipm ent manufac turers. Both sys te ms may
be co ns id ered under certain co nditi o ns based o n plant s ize (usually s mall ), loca l
co ndit io ns, and the type o f process (i.e., reactor-clarilier for lime so fte nin g and co n-
tac t fl occ ul ati on for direct filtrati on with Illultimedia).
Discussion of Alternatives
Mechanical Mixing Systems Select io n of a proper Il occulati on unit depends on
the ove ra ll unit process th at is se lected. For in stan ce, if th e raw water quality is good
a nd direct filtration is feasible, the lilter bed w ill probabl y be a reverse-graded hi g h-
rate filte r bed s uch as a dual -media type. In thi s s itu atio n, the !loc s hould be s mall,
yet phys ica ll y stro ng to resis t the hi gh shearing forces w ith in the bed. Verti ca l-sha n
hi gh-energy flocc ul ators produce a !loc that meets these requireme nts. By co ntrast,
whe n the raw wa ter quality is poor (i. e ., polluted or has medium to hi g h turb id it y),
the suspended matter is re moved from the raw water th rough th e producti o n of good
!lac and e ffe ctive sediment at ion; thi s is especially tru e for co ld -water reg io ns. In
this case, paddle Ilocculators ge nera ll y produce large r and heav ier !loc beca use o f
the larger number of blades, larger to tal blade perimeter, a nd larger blade s urface
area. This type o f flocculator ro tates at slow speeds and is capable of c reatin g mo rc
BASIC APPROACH 113
locali zed eddies with lower shearing forces. However, a vertical-s haft !loccu la tor
with hydrofoil mixin g blades is capable of producing large and heavy /lac if an an-
ionic po lymer is used in conjunc tion with alum in a specific application sequence.
These examp les serve only to s tress the importance of the relation ship between the
selection o f process and the fl occulati on equ ipment.
When a mec hanica l mi xer is se lected as a fl occ ul ator, the mi xer s hould have the
following characteri s tics: (I) it must de liver the G va lues spec ifi ed for each stage
of flocculatIOn; (2) II must provide sufficient eddies and turbul ence so th at the re-
quired vel oc ity gradients are produced; (3) it mu s t provide low shear forces at the
edges of the mixing blades, especially during th e last two stages of fl occu lation '
and (4) it should have low ma intenance and operati o nal costs. Certain types of mix ~
ers, s llch as plate turbine and 45° pitched blades, produce a relatively un ifo rm flow
moti on or a rotati o nal Ill otion because of thei r rapidly rotating blades. Yet the lo-
cali zed shear forces at the tips of the blades arc hi g h. Therefore, theses types o f
mixers are no t sUlled for the ac tual flo cculati o n process.
Rega rdless of the type o f flocculator, tapered mi x ing across the Il occ ul at io n tank
is always an important co ns iderat io n. The last stage of th e flocculation process of-
ten requires a delicate and gentl e mixing condition to promote the grow th o f fl oc
size. For thi s reason, it is best to spec ify an infinitely variab le speed contro l unit
(frequency co ntrol) fo r the lin al-stage Ilocculators. Exa min ation of records from var-
~ous plants reveals that the first and second stages do not require freq uent cha nges
111 speed. Therefore, a two-speed mo tor control mi xer may be used.
Another aspect that s ho uld be considered is the reliability o f the Il occula ti on sys -
tem. Vertical-shaft flocculati on tank s con tain a large number o f mixer unit s. Fa ilure
of one or ~w.o of the. Il occulator drive units usua lly does no t s ignificantly a ffe ct the
overall efl,c,ency 01 the system. Conversely, th e hori zon tal -shaft paddle Il oec ula-
tors generally have only o ne drive unit for several paddles; failure of o ne drive will
tenn.inate the movement o f several flocculator paddles. It is therefore important to
speCify an extre mely reliabl e drive unit when usi ng horizontal-shaft !loccu lators.
In summ~ry, all other fac tors bein g equal , vert ica l-shaft floculators arc usuall y
the ~r.s~ chOice for the followin g reasons : (l) minim al maintenance, (2) operatio nal
fleXibi lity, (3) very little head loss across th e tank , (4) easy control of mixin g int en-
Sity, (5) e ffectiveness, and (6) minimal impact to the overa ll performance if one unit
malfunctions. The drive unit s shou ld have hydrofo il mixing blades, and a ll s haft
beanngs s hould be located above the water s urface o f the bas in. Indepe ndent o f the
i,·
type of mech anical mixer arc the requi rements for s ufficien t walkway s pace (4 ft
2 m)~ aroun~ each mi xing unit for the control panels, for power connec tio ns, and
or t~e II1s tallatlOn , removal, a nd reg ul ar mai nte nance of eq uipment. O ne drawback
of tillS sys tem is th at large plants may require many mixing units. If a plant requires
~ver ~O verti ~al- shaft fl occu lato rs, it would genera lly be co ns idered to be excessive
oth 111 quanlity and cost co mpared with other types of proven flocculati on system s.
Baffled Channel Basins The mai n fac tors in determining the app li cabi lity of the
baffled channel fl occ ul ati o n process are the local co nditi ons a nd the prese nce o f a
 BASIC APPROAC H 115
114 DESIGN OF BASIC TREATMENT PROCESS UNITS

moderate amo unt of allowable head loss across the tank. For example, many devel- 1.0
oping countries commonly have problems with proper maintenance of the equip-
me nt. Since the ba rn ed chann el fl occ ul ation process docs not have any mec hani cal .... I •
0
G =20
G =40
'
....
parts, it should be co nsidered seriollsly in thi s type of situatio n.
I G =60

-- .. •
When des ig ned properly, both the horizo ntall y bamcd (arollnd -the-end 110w) and G =80
the verticall y bartled (over- and undcrll ow) channels perform well and exhibit good
plug fl ow characteri sti cs. However, traditi onal barned channels have two di sadvan-
tages: There is a sig nili cant hcad loss of 2 to 3.3 ft (0.6 to I m) across the tank , and
::::l
I-
......
% "*' 1
• G =100
Raw Water Turbidity = 5.8 ru
mi xing intensity is a fun ctio n of plant fl ow rate. Helico id al or tangenti al fl ow baffled
channels create less tha n a 50% loss of head than tradi tional des igns without sacri- -
,->
.--
"C
_. a Alum Dose = 16mg/1
pH = 7.0
tic ing performance. Many e ng ineers who evaluate barned channel flocculati on sys-
tems gene rall y beli eve that the system is inadequate when they see a low G value -...
::I • ~I I 1 I •
across the system under low- n ow co nditi ons. However, th is impression is o ften in-
...
I-
0.5
correct because they have not calcu lated G X I . In general , good fl occ ulati on occurs
when G X t ranges from 3 X 104 to 2 X 10' (Figure 3.2 .4-4). Thus an average G of
the system may be as low as 12 S - I, but if mix ing time t is no less th an 45 min (2700
-:::
Q.I
<0
\.1 1 1 I 1
s), G X t will be greater than 3 X 10 4 ; therefore, the system should perform well.
This di sc ussio n is not meant to impl y that use of the ba rned chann el fl occ ul a-
-
"C

--
Q.I
"V!'-~ I I I.
~I.... ,.I • I /
tio n process is restricted to developin g co untri es. If the conditio ns of fl ow rate vari - Q.I
Cf.j
ation and avai labl e head loss can be met, thi s process s hould al so be considered for
use in deve loped countries. Some of the newer plants in bo th the U nited States and I -::
Japa n use ba rned channe l tl occul ati o n tanks.
I I I -j
Baffled Channels and Flocculators Combination System This co ncept com-
bines the best featu res of bo th systems. It offe rs a large reductio n in mechanical I I I
Il occ ulat or unit s so there is n savin gs in capital and maintenance cos ts, and in creased
reli abilit y. The Il occ ulat or units sho uld be located in the tinal s tage of the fl occula- o 6 12 18 24 30 35
ti o n process. This all ows floc to remain in sllspe nsio n and supplies additi o nal mi x-
in g energy during low process Ilow. Des ig n eng ineers should provide at least l.5 ft
(0 .45 m) of loss of head across the Il occ ul ati o n syste m at a Ilow rate des igned fo r Mixing Intensity (Gl x 104)
hcli coid al fl ow bam ed channe ls. Figure 3.2 .4-5 depicts thi s fl occ ulatio n system. Figure 3.2.4-4 EV<l ll1<1ti oll of Ilocc lilation co nditi ons b<lsed on G x I <IS a P<lf<l 111ctcr.
T here are curre ntly two large plants in th e United States usin g the co mbi nati on
ba rtl ed c hannel and 110cc ul ator system; the Ho us to n East plant [150 mgd (6.5 m'/s) ]
Some equipme nt manufac turers, s uc h as Ne ptune M icrotl oc and C ull igan, have pro-
began operatin g in 1993. These plant s have a settl ed water turbidity as low as 0.3
moted thi s sys te m in the ir pac kaged wate r treatme nt unit s. The ad va nt age of co n-
ntll. The New Mohawk plant [100 mgd (4.4 m' /s)] of Tulsa, Oklah oma does not
tac t fl occ ulati o n is its co mpact size and its lac k of mov in g parts. Hydrauli c load-
have problems w ith the wate r in the channe l freezing even tho ugh the plants are eX - .Ings 0 f 10 to 15 gpm/ ft 2 (2 5 to 37.5 m/ h) are poss ible and the sys te m requires an
posed to ambie nt te mperatures below 6°F (- 15°C) duri ng the winter. Plant opera-
empty bed co nt act tim e o f o nl y 3 to 5 min to produ ce a fl oc with th e proper char-
ti onal data from these two plant s demonstrat e th at process e ffic ie ncy does not change
ac.te~·i sti cs. This type o f tl occ ulatio n sys tem depe nds on plant !l ow rate and requires
durin g the win ter whe n plant n ow rate is 30% of design fl ow rate. However, oper-
~1J1 l1na l vari ati o ns in fl ow rat e and water te mpe rature for adcquat c mi xing inte n-
ators o fte n isolate a certain number of basins durin g peri ods of low water demand
sity. Co nt act fl occ ul ati o n also requires frequ ent bac kwashin g with air scour.
to maintain a flo w rate th at is 50% of des ig ned fl ow rate to prevent n oc deposition
with in the channels. Diffused Air or Water Jet Agitation O f all the processes, the least pre ferred are
dlf~used air a nd water jet fl occ ul at io n. The di ffu sed a ir system is c harac teri zed by
Contact Flocculation Cont ac t 110cclil ati o n (grave l-pac ked lilter) is another de-
a hi gh rate of e nergy consumpti o n and is also an ine ffi c ie nt Il occ ul ati on process.
sign th at requi res minim al maintenance and is similar to the in-lin e liltratio n process.
116 DESIGN OF BASIC TREATM ENT PROCESS UNITS
BASIC APPROACH 1 17
. t--i .. 1i00IOS

- - - x- ,... ,... The desig n of the Ilocc ulation tank is commonly based on o nly two criteria: de-
tention time and mixing energy level. The energy leve l is the G value or ve loc ity

~~ - - gradient and is defin ed by Camp et al. ( 1943) as

I X-~ -I- -
- -- sm ENT n ON _
N N x- G -_ ( - P )0.5
-
I
- - X-
I- -
-l-
",V

- - -- ~- -I- - where P is the power input to the nuid, V the volume of the tank, and f..L the ab-
I
- -- floc 1110- Lx- ,... p\dl y
-
0--
solu.e viscosi'y of the Iluid.
The flocculat ion process mu st be designed to provide maximulll interpartic le con-
- 1 - -- X-" )\. tact. These contac ts are a fUllction o f the ve loc ity g radi ents within the fluid; the ve-

~~
I-
I .
I - - X- I- -
I-
locity gradients are produced by hydraulic o r mechani cal mix ing (orthokinetic noc-
culation). The number of particle contacts is expressed as

- I
_ SEO( ENT nON _
BRS N N X- - I- -
--
- I - "
- x-· - - N = 1/]1/,
G
'6 (d . + d, )'

~ ~ I - - I-
Xo · - - - where N is the number of contacts betwee n III particles of diameter d l and 1/ 2 par-
-
- I - - ~-. , ___ ' 0 ticles of diameter (/2 in a unit of time. Therefore, the rate of nocc ulation increases
with the number and s ize of the parti cles a nd with the power input , but decreases
Figure 3.2.4-5 Fl occulation system which has a combinati on of bafl1cd channels and me- with the viscosity of the Ouid.
chanical flocculators.
The mean velocity g radient (G value) of certain types of Il occ ulati on systems
may be compu.ed as fo ll ows. For the bafned channel,
There fore, (his system should serve as an auxiliary or temporary meas ure. Use of a
water jet mixing sys tem is not recommended for Ilocc ulation because the high shear-
ing force on the jet path restricts the growth of floc size.

Design Criteria When desi gning a Iloccu lation process, selec.ion of .he mode of
where v is the kin emati c viscosity of fluid , 1 the mea n detention time, g the gravity
mix ing and determination of the physical relations and characteri stics of the Iloccula-
acce leration constant, and II the head loss across the tank.
lioll tanks and clarifiers (sedimentation tanks) are among the first decisions to be made.
For the mec hanical mi xers with paddles:
As described previollsly. either hydraulic mixing or mechanica l mixing Illay be cho-
sen. In evaluating the physical characteristics, a deci sion mu st be made as to whether
the fl occ ulation tank will be built as part of the clarilier or as an independent structure.
G = ( C"Av' )0.5
2vV
After establishing these two issues, the design may be relined; establish the number of
tanks; considcr the total number of tanks when the plant capacity reaches the ultimate
where Co is a drag coefficient that depend s on the paddle shape and now co ndi-
capacity; deline the number of mixing slages and the mixing cnergy level in each stage;
ti ons (use 1.8), A the cross-sec tion al area o f the paddles. v the relat ive ve locity of
and determine the tyfX! of bame that wi ll be used to minimi ze flow shall-circuit ing.
.he paddle with res pec t to .he Ilu id, v .he kinem at ic velocity of .he Iluid, and V the
One of th e Illost efficient and economical flocc ulation tank des ig ns incorporates
vo lume of the flocculatio n tank . It is estimated that \I is from 0.5 to 0.75 of the pe-
the basin into the c larifier unit. The majority of equipment manufacturers desig n ripheral ve loci.y of' .he paddle.
their unit s exc lusive ly in thi s manner. For custom des ign , the floccul ation basin may
Proper application of the parameters prese nted above is described in the foll ow-
be combined with the c larifier by placing it at the influent end o f a rec tang ular sed- ing section.
imentation tank. Thi s arrangement offers Ill any advantages, sllch as minima l land
requirements, minima l Ooc breakage hetween the two processes, simplicity o f de- General The genera l design criteria for a basic rectangular Oocc ulation tank are
sig n, a nd above all, simple and economica l constru ction . as fOll ows:
118 DESIGN OF BASIC TREATMENT PROC ESS UNITS
BASIC APPROACH 119

Energy input C X t = 3 X 104 to 2 X 10' 80

4
t is in seconds (5 X 10 average)
orC = IOt070s - (30s - 1 average)
1

Detention times 20 to 30 min at maximum dail y tl ow rate

70
Water depth 10 to 15 ft (3 to 4 .5 m)
Flocc ul ation stages Two to six stages (comm only three or four)

T he fo llow ing factors wi ll determine the numbe r of tl occ ulati o n stages to be in· g 60
eluded in the design. c
.g
~
.~\,
~

E \ "':!.,
I . Type of subsequent treatm ent unit (sedi me ntation, filt er, or slud ge dryin g bed) ~
c
~
50 \ "\
and the type of overall process (d irect liltration or comple te process) 't, , .'
[I
2. Quality and treatability of the raw wa ter: turbidit y levels, the nature o f the ~
c \ o '~.,~
turbidity, and water te mperature d urin g the cold mon ths w
3. Degree o f fl ow short -c ircuiting across the Il occ ulati on basin (with or with out
barnes) and the type of bames
40

"
",',

"
""',
'tl..
x Type c

Type b ........ Type,

4. Local conditio ns: regulat ory requirement s and the operat io n a nd maint enance
30
"'- ......
of the ll occ ulat ion process -- - ---DType d

Vertical-Shaft Flocculators In a mec hani cal mixer the diameter of the mix ing
°0~------~2~0-------4~0~----~6~0~----~BO
blade (D), in rel ati on to the tank diameter (eq ui va le nt diam eter T ), is important and
G (,-.1)
DIT s ho uld be greater than 0.35. Another ite m to wa tc h is the max imum Il ow ve·
loc ity induced by the mi xing blade: It sho uld be less than 8 ft/s (2.5 m/s) in the lirst
stage and less than 2 I'tis (0.6 I11/s) in the last stage of the fl occ ul ati on process if
good and settleab le floc fo rm at ion is to be ex pec ted. Mi xi ng eq uipme nt man ufac-
tu rers have establi shed the des ig n criteri a for mechanic al mix ers based on the ir rc-
search and experie nce. However, their mai n objec ti ves are for liquid- liquid, liq -
uid- gas, a nd solid- liq uid mi xing as a batc h o peration. T hus, manufacturers often
do not have a good understanding of floccu lator des ig n for a con tinuo us-flow mi x-
in g process and the product ion of frag ile ll ocs. When produced properly, th e fl ocs
should have the charac teri sti cs and appearancc of s nowllakes.
There are two bas ic philosophies in spec ify ing mec hanical tl occ ul ators. The first
is simpl y to s pecify the function and a few ph ys ical characteristics, such as G va lue,
tip speed (excluding hyd rofo il blades), and th e DIT ratio. T he o ther is to write de-
tailed specifications for the eq uipment, installati o n proced ures, and exac t power re-
quirements, including the dimensions and elevati on of the mixin g blades. It appears
c

,-~ ~I'
th at the first approac h is better fo r th e majority of e ngi neers sin ce the proper design
of vertical-sha ft Ilocc ulators is somewhat state-of- the-art. When using hydrofo il mi x-
ing blades, tip speed is irre leva nt since the blades lift or push dow n on the water,
depending on the setting. Figure 3.2.4·6 s hows the energy tran smittal e ffi c iency for
so me mi xing blades when used as ll occ ul ator unit s.
•• '"." "••',.".,.,
(b) small propeller, (c) large propeller.
(d) turbine . Size in mm. d

Horizontal-Shaft Flocculators The ad vantage of the ho ri zontal -s haft unit is that Figure 3.2.4-6 Mi xing erticicncy as a functio n of G va lue and type of im peller. (A Ber
o ne shaft can opera te a number of agitators and thereby reduce the numbe r of drive LcCnt raar. 1980.)
120 DESIGN OF BASIC TREATMENT PROCESS UNITS
BASIC APPROACH 121

units. However, its reliability is low because fa ilure of o ne sha n results in a 25 to 'fAilLE 3.2.4-2 Gcncntl Dcsign Critcria
30% loss in mi xin g capabi lit y. T he horizo ntal -s haft units can be equipped wit h e i-
ther paddl es or turbine blades. The basic des ig n criteria fo r hori zo ntal -shaft Il occll- Mec han ica l Floccul ators
lators are the sallle as Ih ose for Ihe verli cal type, exeepl for the Iype of paddle unit s. Bortle" Hori zontal Shaft Veni cal Shart
Channel wit h Padd les with Blades
I. T he to tal paddle area sho uld be 10 to 25% of the tank cross-sec ti o nal area. If G (s ' ) 50- 10 tapered 50- 10 tape red 70- 10 tapered
the values are ove r 30%, the padd les will produce a di stin ct rotating motion T (mi n) ) 0-45 )0--40 20--40
in the water, ca using a reduction in local edd ies and turbulence along the edge Floccu lati on stages 6- 10 )- 6 2-4
of th e paddles. (number of chan nels or
number of horizontal
2. Eac h arm should have a minilllulll of three paddles, so that dead space, es pe- shafts)
ciall y nca r the shaft , w ill be minimi zed. Mix energy cont rol Flow passage Variable mi xing Variable mi xin g
3. The peripheral speed of the padd les shou ld be be tween 0.5 and 3.3 ft/s (0. 15 variat ion speed speed
to I Ill/s). Maximum now veloci ty ) 6-9
4. The des ig ned G value is lowe r than that of verti cal-sha n Il occ ul ators; it sho uld or mixe r tip speed (fps)
Blade area/tank area (%) 5- 20 0.1-0.2
initia ll y start at about 55 s - I (hi g hest), then be reduced to abou t lO S- I in the
Blade: DIT 0.5- 0.75 0.2-0.4
las t stage of th e Ilocculatio n tank. H igher th an 60 S - 1 wi ll result in a high
Shan rpm 1- 5 8-25
shaft rotati o n speed; often , thi s leads to severe wear a nd lear of the shaft bear- Major app li cat ion Conventional Co nventional Direct liltration
ings within a short peri od o f time. The workmanship o r shaft install ati o n is co mpl ete compl ete and conventi onal
of utmost im portance for paddle Ilocculators. treatment trea tmen t com plete Ireallnent
5. For a direc t nitrati o n process, whic h does not have c larifi ers, the fl occulati on
ta nks may be designed with around-the-e nd channe ls. T hi s type of desig n min-
In general, each baflle is plnced across the tank perpe ndi c ular to the n ow path.
imi zes Il ow short-circui tin g ror ho ri zontal-shaft ll occ ulatOl's. If the Il occ ul a-
The top of the bafllcd wall is sli ghtly submerged [0.5 in . ( 1.3 Clll)] to all ow the sc um
ti o n tank is part of a rec tangu lar clarineI' tank , the last stage of the Il occula-
to fl ow over, and the bottom o f the wa ll sho uld have a space of 1.0 in. (2.5 cm) fo r
ti on tan k should be des ig ned to approxi mate the fl ow direction round in
easy drainage and slud ge re mova l. To preve nt !low short -c irc uitin g, engi nee rs shou ld
th e c larin er; thi s will prevent tan gential inll ow into th e clarifi er (see Figure
not desig n a large passageway; th e passage Illay be large e no ug h to acco mm odate
3.2.5-4).
a person if it is blocked off during normal operation or if a portion o f the diffu se r
6. Most dri ve unit fa ilures arc caused when the unit is started at the top rota- wa ll is ca pable of swi ngi ng open (by hin ges) o nl y during the clean ing opera ti o n. If'
ti o nal speed . Like all eq uipme nt , it should be started at a low s peed to avoid timber is lIsed to co nstru ct the barne, it w ill te nd to warp if the spnn is too long
very hi gh torqu e force and a hi gh power require ment. [over 10 ft (3 m)] o r if is too th in Iless than 2 in . (5 em) ne t] .
If the nocc ulati on tank is des ig ned as pan o f th e sedim en tation tank , the barned
Tab les 3.2.4-2 and 3.2.4-3 prese nt the general design c rite ria and compare the wa ll loca ted between the tan ks is called the diffu ser wa ll. In this case, th e ba rtl ed
major characteris ti cs among barned channe l, horizo ntal-s ha ft , and vertical-s haft wall shou ld produce a !low rate ve loci ty of no more th an 0.8 ft/s (0.25 m/s) to avoid
Il occ ulato rs. excessive breakage of the deve lo ped fl oc. Another advan tage of thi s type of bartled
wall is the production of add itional mixing e nergy. Barnes th at are based on the cri -
BAFFLED WALLS When mec hanical mi xe rs are selec ted for th e nocculati on process, teria above prov ide a G value of 5 to 25 s - I , dependin g 011 th e flow ve loc it y throu gh
(he usc of ba rn es, for compartme nt ali zati o n, improves Il occ ul ati o n. The shape and the bartles. Figure 3.2.4-7 illustra tes the resu lts of a scale- mode l hydrnuli c stud y 0 11
loca ti o n of the barn es are de termi ned by the desig n e ngineer because there are no the determination of an e ffec tive baflled wa ll system.
prac ti cal des ig n guides. Desig n e ngineers are strongly adv ised to compart men tali ze
eac h stage of the mec han ical mixer Il occulati o n process w ith a proper baffled wall. STATOR []i\FFLES The eflicie ncy o f ve rti cal-s ha ft mi xers can be improved through
Eac h barne sho uld have oritiees th at arc 4 to 6 in . (0.1 to 0.15111) in diamet er uni- the installati o n of stator bnrtles. T he purpose of sta tor barnes is to increase the e f-
forml y di stributed across its verti cal surface and a veloc it y of 1.2 to 1.8 ft/s (0.35 fec ti ve turbulence as well as to prevent rotational Iluid motion within th e tank . A
to 0.55 m/s) should be produced through the orifice hol es of eac h barne at maxi - proper sta tor bafll e desig n spec ifies th e ins tall ati on of two or four pairs of bafll es
mum fl ow rate; the hi gher veloc it y of 1.8 ft/s should be applied to the first barne on oppos ite sides of the compartme nt. The w id th of eac h indi vidua l stntor baflle
and the lower ve locity of 1.2 ft /s to the last baflle. should be ! to II! of the di stan ce across the indi vidua l mi xer compartme nts.
 BASIC APPROACH 1 23

Baffled Channel (Hydraulic) Flocculators The around-the-end baffled c hanne l

and the over-and-under baffled channel are the two basic types of baffled channel
flocculators. The around -Ihe-end baffled channel is considered to be more practical
since the latter has more problems with the acculllu lation of SC um and silt/grit be-
hind th e bafllc.
When barned channels are designed to provide uniform flow velocity through each
stage of channels (plug flow type), the headloss per 180 bend is approximately 3.2
to 3.5 times the velocity head in the adjacent channels. However, if the bamed chan-
nels are designed to produce the required flow velocity at only the narrowed paths.
such as sli ts or submerged porls, the head loss will be approximately 1.5 times the
velocity head of the restri cted area; this type of baffled channel is called helicoidal
or tangential flow because the stream of water enters tangentially into each chamber,
producing a helicoidal fl ow pattern toward the outlet downstream of each turn . To
facilitate easy construction and cleaning. the distance between the barnes should be
a minimum of 2.5 n (0.75 m). The minimum water depth should be 3.3 n ( I 111).
In general, a minimum of six channels should be provided and Ihe head loss across
the flocc ulati on tank must be approximately I to 2 ft (0.3 to 0.6 m) so that a n av-
erage G va lue of 30 to 40 S- I will be established. The process shou ld have a ta-
pered mixing paltern and a minimum residence lime of 20 min at maximum plant
now. The G value of the baffled channel Oocculation process may be es timated
quickly using the following rormu la (using English units):

~
~

i1
~
...c
~

where t is the residence tim e (min), Ir the headl oss (ft), and G the velocity grad ie nt
0
.~ (S - I). Tracer tests have indicated that there is little flow short -circuiting in baffled
:; channels. Case 8 in Figure 3.2.4-7 illustrates thi s condition.
v 0;
~
0
0
Ii;
v
"
.c
U
Example Design Calculations
.~
~
co '"
~

E Example I Mec hanical Mixer (Vertical-Shaft Flocculators)

~ "
.c
.... '" Given
'0
c
0 Plant fl ow rate
~
.;: 20 rngd (0.88 m'!s) da ily average
~ Detention time
20 min at maximum dai ly fl ow rate
'"
E
0
Number of tanks Two rec tangular tanks
u Stage of Oocculation Three stages
.., Energy input
60 to 20 S- I ror both stages I and 2, G = 30 to
..,~ Type or nocc ulator
lO s - I ror the last stage
Vertical shaft with either 45° pitched turbine
'"
....l blade
(PTB) or hydrofoil propeller blades; variab le-
""~ speed drive control
Minimum water tempera ture 50 0 r (10°C)
122

b
124 DESIGN OF BASIC TR EATMENT PROCESS UN ITS BASIC APPROACH 125

tHEORETICAL
tHEORETICAL OU[NTION 111
DE tE NTION (T)

P
'MIN . )0 HC. BAFFLE WALL
6 "'IN. 30 SEC.

, " , "
u
+.
CASE

0.e3~
I u
~
_,. +.
CASE

0 .85
5
-! 0 •0-
,•
~ 00000
•_ '0
,

0P
~
~
e e

:
u
•
~
25
CASE 2-

+-. O.H~ 10': ~15

u
••
CASE

i-. 0 .88
6

~ '~.'" •"o .0
0.0
z
o NOTE RAW WATER CONCENTRATION •
2 NOT[ : RAW WAT [R CONCENTRATION
IOrnt/lt
10mo/i t ~ "' ~ ' FCASE C;
~ •~ , ........
,,
CASE I
•!r: I~
I. '
~ I 'MIN n SEC. '"

8•
~

~ f ,f,lUNJ CASE 5 /'--~_-':"':'~ ....... _~

I .........
8
o • 10
,. .0
l0f--------;~------~-------.~~~~~----~
o , 10
" 20
"
TilliE OF fLOW TO OUTLET IMINUTESI
tiME OF" FLOW TO OUTLET (MINUTES)
(c)
(a)

('OR CASE 7)
THEORlTlCAL
OU[NTION (TI
SAFFLE WALL 2 6NIN. 30 I~(C .
(50Y. OPENING) w
• , , SAffLE WALL
THEORETICAL ~
~
r" ,
~,
f~t0P ,: \,
OEtENTION ITI ~,
CAS[ 3
6 MIN. 30 SEC,
+ .0 , \ CASE .,

, "
0 .82
,,
1
1
,
,
,, + II.I

u
~
o
,•_ '0 CASE 4 11'0:e;'pO" OPE"'"" '0
(CASE 81:
T,
;
~
,
~
CASE 8
+". 0 . 81
'0
& MIN.
,
, t . 0."
"
~u
~ \, NOT[ , RAW WATER CONCENTRATION
~ lom9 / 1t
•o 2.
CASE 7
NOT[ : RAW WAT[" CONCENTRATION
z 10mt/lt
o
~ MIN. 20SEt. I
• INUN)
~w " 'MIN. I'SEC. I tI
u .•. ..
~
IMEANI

'0
,. 20 ,.
o • 10
"
20
TiNE OF FLOW TO OUTLET (MINUTES)
TINE OF" FLOW TO OUTLET (MINUTES) (d)
(0 )
Figure 3.2.4-7 (Col/tillllcd )
Figure 3.2.4-7 (a,e) Fl occul ator arrange ments versus detenti on time; (b ,d ) t1 occul ators with
baffl es versus detenti on time.
 BAS IC APPROACH 1 27
126 DESIGN OF BAS IC TREATME NT PROCESS UN ITS

Batfle walt (typical)

Alternative 1
Determille
I I E
(i) The size of each fl occulation basin , including water depth I
I ® ® 1_
I
",
-
(ii) The lotal number of Ilocculators I I
(i ii ) The motor horsepower of each Ilocculator
I J E
I I <Xl

(iv) The details of the diffu ser walls I

I ® I -@
I N

(i) Assume that the maxi mum daily now rate is 150% of the daily aver·
SOlll lioll
Ir E-. 3 @8m = 24 m
-~:...:'!..::.::::...::..::.::::~--;,~ I Flocculator (typical)
age now rate. The vo lume required for each lank is

1.5 X 0.88 X 60 X 20 = 792 m' (27,970 ft ' )

2
Alternative 2

Make the average water depth 14 n (4 .25 Ill). The total tank area required is

792 .;- 4.25 = 186.4 m 2 (2005 ft2)

The size (shape) of eac h basin is as follows (see Figure 3.2.4-8):

• Alternative /: 8 m X 24 m ( W X L) or 26 ft X 78 ft
• Altern ative 2: II m X 16.5 m (W X L) or 36 ft X 54 ft
• Alternative 3: II m X 18.5 m (W X L) or 36 ft X 6 1 ft
:
1....:
3
_,-00_-.J~~
I
@ 5.5 m =:. 16.5 m Baffle wall

A ltern ative 1 is the 111 0 s1 economica l design, but it is not the most effective tank.
I E ., .

A lternati ve 2 is not as cost-effective to build as altern ative I . However, the larger

number of llocculators gives it a greater safety factor and more effective floccul a-
ti on. Therefore, adopt altern ati ve 2. Alternative:' is composed of n series of com- Alternative 3
plete mi xing reac tors that mi nimi ze the problem of fl ow short-circ uiting. Even
though thi s is the best type of fl occulation tank , thi s scheme requires a flow di stri-
/O\~
'<Y -1-,..
bution channel to be co nstructed , lead ing to the sedimen tati on tank . The now di- -- 4 e
recti on in thi s chan nel is not ideal because it is perpendicu lar to the fl ow direc ti on
in the clarilier. H owever, the use of a clarilier inlet diffu ser wall , which is described
0-'y0 t-
in Chapter 4, can produce a reasonably good in let fl ow co nditi on. I
(ii) T he total number of flocculators is 12 since alternati ve 2 has been selec ted 0""0 -y-y-e
because of its cost-e ffec ti veness.
(iii ) M otor horsepower of the fl occulato rs. For all Ilocculators in stages I and 2,
y-
0-.;(1:
P = G\I.V BaHle
(optional)
I. 3 @ 5.5 m = 16.5 m
) '.
II ~2 m

whe re Il. = 0.00 13 N . s/m2 (2.73 X 10- 5

Ib . s1ft2) (N, new ton = kg . m/s2)
Figu re 3.2.4-8 Alternati ve basins for Examp le \.
P = (60)2 X (1.3 X 10-') X [(5.5)2 X 4.25]
= 602 l /s pe r Ilocc ul ator
 BASIC APPROACH 129
128 DES IGN OF BAS IC TREATMENT PROCESS UNITS

Si nce I kW = 1000 l is, P = 0 .6 kW or 0.8 hp (I hp = 0.746 kW). Assu ming that

the overall efti ciency of a mechanical mixer is 75%, the mi nimum required motor
horsepower for each fl occulator in stages I and 2 is
0425 m BaHI ewalnO. l
. ~
0.6 -;- 0.75 = 0 .8 kW ( I hpj
E @ @ @ @ @ @ @ @ @ @ @ @ @ @
For the fl occulators in stage 3, the molor horsepower required can be computed in "
,.;
@ @ @ @ @ @ @ @ @ @ @ @ @ @
E I
the same way.
'" .,E
N
@ @ @ @ @ @ @ @ @ @ @ @ @ @
P = (30)' X (1.3 X 10- 3) X [(5 .5)' X 4.25 1 = 150 l is " '" 0 @ @ @ @ @ @ @ @ @ @ @ @ @ @

Therefore, P = 0. 15 kW -;- 0.75 = 0.2 kW C hpj, minimum for each n occul ator
!- @ @ @ @ @ @ @ @ @ @ @ @ @ @

l3 @0.8m - l0.4m
in stage 3. 0.3 m-j . I . 11m , I,r03m
(iv) There will be th ree barn es: call them No. I , No.2, and No.3 barnes, co unt-
ing from the inlet side. The n ow velocit y th rough eac h baflle will be set as show n
and the total orifice area required can be computed. 0.375 m
Baffle wall no. 2
0 .15 m diameter orifice
Barn ed Barned Bafll ed
Wall 2 Wall 3 @ @ @ @ @ @ @= @ @ @ @ @ @

~l
Wall I
@ @ @ @ @ @ @ @ @ @ @ @ @
0. 55 0.45 0.35
Maximum n ow veloc ity (m/s) E @ @ @ @ @ @ @ @ @ @ @ @ @
2
Total orifice area for both tanks (m ) 2.4 2.93 3.77
.
'"
N @
@
@
@
@
@
@
@
@
@
@
@
@
@
@
@
@
@
@
@
@
@
@
@
@
@ @
The size of a single orifice should be 0. 1 to 0. 15 III in diameter to prevent clog· "' - @ @ @ @ @ @ @ @ @ @ @ @ @ @
gi ng by algae or suspended matter. A reasonable di stance must also be kept between
the orifices so that the structural integri ty of the wall is maintained. Thus select 0.1 5
m as the diameter fo r each ori fice. The following table lists the total number of ori·
1,1. 13 @ 0.8 m ... 10.4 m
11 .0 m
~ I, r-0.375 m

lices per tank, the total orifice area, and the head loss per baffle for both the maxi·
mum and average dail y pl ant fl ow rates . 0.375 m
Baffle wall no. 3

Barned Bafll ed Baflled

Wall I Wall 2 Wall 3 ~I @ @ @ @ @ @ @ @ @ @ @ @ @ @ 181 @ @ @
,.; @@@@@@@@@@@@@@@@@@
140 168 2 16
Total number of orinces ~ ; @@@@@@@@@@@@@@@@@@
2 2.47 2.97 3.82
Total orifice area (m )
0.44 0.35 "ci 181@@@@@@@@@@@@@@@@@
Velocity at Qmax (111/s) 0 .53 @ @@@@@@@@@@@@@@@@@@
0.30 0.23
Veloc ity at Q,,, (m/s) 0.36
0.023 0.0 16 0.01 0 .L.: @@@@@@@@@@@@@@@@@@
Headloss at Qlll<lX (Ill)
0.007 0.004 _I. l7 @ 0.60m ~ 10.2m I k-, oAm
Headloss at Qave (m) 0.0 10

Figure 3.2.4-9 shows the details of three bafll ed walls in eac h ta nk. The rati o of
I
. < Il.0m ', l ·
Fi gure 3.2.4-9 Detail s of the three baffled walls of eac h tank.
the orifice area to the cross·sectional area of the tank is 2.5%, 3.2%, and 4. 1% for
walls No. I, No.2, and No. 3, respec ti vely. The head loss thro ugh the orifice is com-
puted fro m the formula Q = CA(2g h)o.s, where C = 0.8.
130 DESIGN OF BASIC TREATMENT PROCESS UNITS

Example 2 Horizontal Shaft with Paddle Flocculator

0"

Given

Plant !low ratc 1.85 m' /s (42 mgd) max imum pc r day ...,
Number of tanks T hree rectangular tank s
Stage of fl occulation Two stages due to site limitation s
Detention time 32 min at 1.85 m'/s
Shape of eac h tan k 5 m (W) X 24 m (L) X 5.0 m ( IV . D)
Type of fl occ ulator Hori zo ntal shaft with paddles wi th variab le-speed
co nt rol
E nergy input ( I) G = 50 S- I ma xim um
(2) Displacement factor (Bea n, 1953) = 30 N ~ •

maximum
Minimum water temperature lOoe (50°F)

Delermine

(i) The number of paddles per shaft

(i i) The details of the paddle
(iii) The maximum number of shaft revolutions per minute
(iv) The required horsepower of eac h drive unit Ir - - -- wO ...:_ _ _ _ ~

~~~ci~ci~ci~J-W o'-4 "Ol

Soilltioll (i) Considering the number of sha fl supports and lhe integri ty of the pad-
dle, a reasonable figure would be e ight paddles per shaft , or a total of 48 paddles w

for thi s Ilocculation process.

(ii) After a few trial s the shape and dimensions of the paddle show n in Figure
3.2.4- 10 are adopted .
I

NOTES:

I. Si nce there arc only two stages of flocculation. each stage is compart mentali zed with bar-
tl ed wall s simil ar to baftled wa ll s 1 and 2 of Example I (see Figure 3.2.4~9).
2. The number of paddles used is as many as is pract ica l because the drag force and eddies
produced by the paddles are proportio nal to the total length of the paddle edges.
3. The paddles are arranged diagonally to min imi ze !low short~circuiting; the travel speed of
th e paddle is minimal ncar the shaft.
4. Each paddle wheel con tains 3.97 m2 of the paddle area. Thi s is just under 25% of the
cross-sectional area of the tank .

(iii) Since eac h paddl e wheel has more than o ne blade, G is co mput ed by

AV' )0.5
G = Cf) (L 2vV 131
132 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 133

where A is the area of each paddle perpendicular to flow, v the relative velocity of each where
paddle to water, LJ the kin ematic veloc ity of the fluid , and V the volume of the tank.
D = 2, d
v = (0.75 X 2'TT rN) -;- 60 = 0.0785 rN d = a X s

where N is the rotational speed (rpm ) and r is the radiu s. Try 0.5 Ill /s as th e pe- S = 27rr X rpm
ripheral speed of eac h paddle wheel. Incide nlally, Ihe rpm of Ihe paddle 01 a ve loc-
ily of 0.5 m/s is (0.5 X 60) -;- (2'TT X 2. 1) = 2.27. The lirsl paddle's and

I
C = Y [I .8 X (0.24 X 2 X 48) X 0.372- J : [2 X (1.3 1 X 10 6) X 3600] d = d isplacemenl (pumping) capac ily of each paddle (fl 3/ min )

= 15 .04 S- I a = area of eac h paddl e (fI' )

s = Irave ling di Siance of Ihe paddle area A (fl/ min )

In like manner:
r = distance (ft) between the center of gravity of the paddle area (a) and the
The second paddle's G, = 12 .5 $ - 1 center of the shan
The Ihird padd le 's C3 = 10.0 S- I
F = !low rale (fl 3/min )
The fo unh paddle's C4 = 7.7 S- I

The liflh paddle's C, = 5.6 S- I

C 6 = 4.3 s- 1 Since a paddle speed of 2. :13 rpm sa ti slies the G vaillc, use thi s specd to check the
Two diagonal paddles
displaccment factor.
Thus, the G value for the total padd le wheels in the tank is calculated as the sum
d , = 0.24 X 2'TT . 2.33 X 2.05 = 7.20 m3/min
of the above:
;- 6
d, = 0.24 X 2'TT . 2.33 X 1.8 = 6.37
C = I Ci = 55.3 S- I = 55 S- I d, = 0.24 X 2'TT . 2.33 X 1.55 = 5.44
i- I
d 4 = 0.24 X 2'TT . 2.33 X 1.3 = 4.56
Make the max imum rotational speed of the shaft 2.:n rpm to mee t the required G
value of 55 S- I . d, = 0.24 X 2'TT . 2.33 X 1.05 = 3.69
d6 = 0.25 X 2'TT . 2.33 X 0 .55 = 2.0 I
C X t = 55.3 X (32 X 60) = 1.06 X 10'
d, = 0.25 X 2'TT . 2.33 X 0.55 = 2.0 1
The power requirement is d 8 = 3 X [(0.065 X 2.1) X (2'TT . 2.33 X 1.05)] = 6.29
i~8

p = C' fJ. V 2 X I
i= l
d; = 2 X 37.52 = 75.04 m' /min for one paddle
= 55.3' (1.3 1 X JO -3 ) X 3600
Since therc are 48 paddles in thi s Il occulation process,
= 14 .437 l is
= 14.4 kW D = 75.04 X 48 = 3602 m'/min
F = 1.85 X 60 = III n,.' / min
(iv) Assume that overall mechanical cfti ciency is 75%. Thu s P = 14.4 -7- 0.75
19.2 kW (25 hpj. Based on Bean's desig n crileria of Ihe di splacemenl faCiOI', Thus DlF = 3602 -;- III = 32 . 1 > 30. Ass umin g Ihal Ihe ove rall equipm e nl e rti-
cieney is 70%, we tind that th ere arc eight paddles mounted per shaft. A largc ini-
D "'- 30 tial torque is created when the illataI' is switched on. For thi s reason, a 22-k W (30-
F hp) motor is selec tcd for each of the six dri vc units.
134 DESIGN OF BASIC TREATMENT PROCESS UNITS BAS IC APPROACH 135

NOTE: Exa mple 2 is based on an actual plant design. The tl occulation system was
constructed as outlined above and has been producing excellent floc si nce 1957.

Example 3 Damed Channel Flocculation Process

~
4

,
Given Stage 1
(20 turns) e
Plant tl ow rate I m~/s (23 mgd) maximulll daily N

Detention time
Number of tanks
25 min at I n? ls
Two tanks Stage 2
-"
..;

Three stages 116turn5) e

Stage of flocculation N
oj
Energy input (at I m'!s) G = 70 s - I for stage
@
G = 34 s - I for stage 2
G = 20 S- I for stage 3 Stage 3
(12 turns)
'"
Type of channel Around -the-end type
Minimum water temperature 10' C or 50' F

Determine
~ ~
28.5 m
/ ·1
(i) The size of the tank and the number of channels c
(ii) The size of the channels and the arrangement of baffles
Figure 3.2.4·11 Baffled channel flocculati on system.
(iii) The headloss and the openings at each baffle

Stage 3 channels require that G be 20 s - I. Thus

SO/Ilfioll (i ) The total volume req uired is (1 X 60) X 25 = 1500 n? Since there
are two tanks, the vo lume of one tank is 750 m3 . Establi sh the average water depth h, = [20 2 X (U X 10 - 6) X 250] + (9.8 1 X 0.5) = 0.027 m
as 2 III and provide six channels (2.2 m wide) per tank. The length of each channel
should be The total head loss across the tank is ('

750 + [(2.2 X 6) X 2] = 28.4 m i= 3

I
i= l
hi = 0.433 m
Make the length 28.5 m.
(ii) Figure 3.2.4· 11 illustrates all the information necessary for the design: size
The opening dimen sion at each baftle o f the stage I channels may be computed
of the channels and arrangement of baffles. from the required head loss in each turn, This particu lar example contains a total of
(iii) Stage I tloeculation channels require that G be 70 s - I at 0.5 m'!s per tank.
20 turns, thereby producing a head loss of 0. 325 m (a head loss of 0.0163 m per turn ).
Therefore, the head loss across the stage I channels is The head loss per turn is computed as
2
hi = G vV = [702 X (U X 10 - 6) X 250] + (9.81 X 0.5) Kv 2
gQ
h =--
2g
= 0.325 m
where K = 1.5 because th e mean tlow ve locity of each channel is only 0.114 m!s
Stage 2 channels require that G be 35 s - I at 0.5 m'!s. Therefore, and the flow enters tangentially into the next chamber; the flow enters in such a
manner that a helicoidal tlow pallern is developed and a clear 180' plug now bend
h2 = [35 2 (1.3 X 10 - 6) X 250] + (9.81 X 0.5) = 0.081 m does not develop. The ve locity required at each slit in the stage I channels is
 \

136 DESIGN OF BASIC TREATM ENT PROCESS UN ITS BASIC APPROACH 137

v, = (2f~' r = ( 19.62 x
0.00 163
1.5
)0.5 thercfore, proper fl oc observati on requires a strong spotli ght to be aimcd toward the
lank fl oor. This lighl should be localed sli ghll y above Ihe walerlin e. If Ihese re-
qui remenlS are mel, Ihe Iloc can be visualized clearly due 10 Ihe Tyndall effec!.
= 0.462 m/s However, the prcfcrred setup is to havc an underwater lamp placed 12 to 18 in . be-
low the waler level.
The required width for each slit in the stage I channels is calculated to be Rccently, all major trcatmcnt plants in Japan havc installed undcrwater telcvision
cameras in the fl occulati on tanks to monitor floc size with hi gh magnificati on on a
IV, = JL = ~ X 2 = 0.54 m continuous basis. Computers automatically compare ac tu al fl oc sizc, as obscrved by
vH 0.462 the cameras, to a preprog rammcd standard size fl oc and warn pl ant operators when
floc size is less than adequate in addition to automatic correcti on of coagulati on
In stage 2 fl occulati on there is a total of 16 turns and th e total head loss across the feed rale. This Iypc of sYSIem provides reliable planl operalions a nd should be co n-
slage is 0.08 1 m. Thus sidercd for new treatm ent plants.
The majority of water treatm en t plants cxperi cnce problems with scum accumu-
v, = 0.258 m/s
lalion unl ess Ihe fl occ ulalion lank has been des igned 10 avo id Ihis. Allhough Ihe
and Ihe wi cllh of each slil is IV, = 0.97 m. Siage 3 Iloccul alion has 12 ba rn es and scum does not havc a significant cffect on thc watcr treatment proccss, it is clearly
muSl produce a lOla I headloss of 0.027 m; 0 .00223 m al eac h bafllc. There fore, ", vi siblc and is unsightly to mos t vi sitors. Another problem common to man y plants
= 0 .1 7 m/s and Ihe widlh of Ihe slil should be 1.47 m. The following lable Stllll- is Ihe growlh of algae on Ihe lank wa lls a nd ba rn es. Al gae growlh is un sighll y a nd
mari zes the charac teri sti cs of the baffled chan nel hydraulic fl occulati on process. can give the water an objec ti onable odor. Algae growth may be prevented through
Ihe usc of prechl orinali on: A conSlanl le vel o f 0.3 mglL or more o f res idual chlo-
Max imum Day, Ave rage Day, rine should be prcsent at all timcs. However, for many treatm ent plants in the Unitcd
I m'/s 0 .667 m' /s States the use of prcchlorinati on is limited due to restricti ons placed on THM for-
mati on. Thc growth of algae is most severe in the settling basin. One solution to
Siage Siage Slagc Siage Stage Siage Ihis problem is Ihe peri od ic applicali on of a hi gh dose of chl orine du ring Ihe ni ghl
I 2 3 I 2 3 for a short durati on (shock trcatm en t).
Slil size (m) 0 .54 0.97 1.47 0.54 0.97 1.47 In regard to maintenance, the baffl ed channcl process requires only an annllal dc-
Flow velocily al sli l 0.47 0.26 0. 17 0.31 0. 17 0. 11 watering of the tank (cvery 12 1110nlhs), during which any necessary repairs can be
(m/s) performed and sludge removal can be accomplished. Sludge removal is easier with
Headloss (m) pCI' slil 0.0 163 0.005 1 0.0022 0.0073 0.0023 0.00 1 the hori zontal turning (around-the-end) barn ed channels th an with the vcrtical tUfll-
Number of slilS/SIage 20 16 12 20 16 12 ~ng (over-and-under) chanllels. The mechanical mixer fl occulati on units are signif-
G-value/Slage (s - ') 70 35 20 38 19 II ~ea ntl y more maintenance intensive than the baffl ed channel type. These mi xers typ-
Head loss (m) across 0.43 0. 19 Ically requ ire the lubricati on oil of the speed-reducing units to be changed on a
process seasonal basis. repair of any oil leaks, and the repair and/or adjustment of the un-
derwater shaft bearin gs when necessary.
A survey of the opera ti onal staff for 30 plants revealed that most operators did
NOTE: The design presented in Example:.1 is based 0 11 the assum ption of 100% nol frequenll y change Ihe n occ ul alion speed. Allhe rn ajorily o f Ihe pla nlS surveyed,
effi ciency. The actu al design should allow for 80% effi ciency. Therefore, stages 2 t~e InI Xlllg speed was not even adj usted with changes in th e season. despite an ob-
and 3 should have 10 turn s per channel, not 8 and 6 as shown. The lotal number of V I O~IS change in wa ter temperature. The main reason given by the operators was that
lurn s for Ihe e nlire lank becomes 60, no l 48 as shown (Figure 3.2.4- 1 I). the efli ciency of th e process did not appear to improve even when attempts were
made to optimi ze the specd . Consequently, certain mechanical speed-reducing units
Operation and Maintenance The three basic operati onal procedures common to were often "frozen" at one speed and were no longer capable of changi ng speeds.
all fl occulation systems are checkin g fl oc size by way of visual observati on, removal The infinite range speed-reducing units used for the fl occulalors requirc the most
of scum from the water surface, and control of algae growth on the tank walls and baf- malntenancc. Con sequent ly, a significant number of plants have replaced them with
fl es. Ifmee hanicalmixing is prov ided, the mi xe r speed conlrol should also be inspecled. two-speed motor contro l units. However. somc recent infinite range speed reduccrs
Floc observati on during the night shift requires a specifi c type of lighting. The are more re liable as well as cost-effec ti ve. A variable- frequency control sys tem for
reg ular lamps located at the walkway are inadequate because the light is rellec tcd; the motor speed adjustm en t is an example of thi s type of unit.
138 DESIGN OF BASIC TREATMENT PROCESS UNITS
BIBLIOGRAPHY
Amirlharaj ah, A .. " Desig n of Flocc ulati on Syste ms," in Water Tre(llmClI1 Pla1/t Design for the
Practicing £" gillcCl ; It L. . Sacks. cd .. Ann Arbor Sc ie nce. A nn Arbor, MI. 1979, pp.
195- 229.
ASCE. Wale,- Tre(l/IIICIll Plallt Design. ASCE Manual of Engineering Pract ice 19. American
Society of Civil Engineers. New York. 1952.
Bea n. E. L.. "S tudy of Physical Factors Affecting Flocculation," Warer WaSlewater EII8·, p.
33 (Janu ary 1953).
Brown, G. G .. "Agitati on," Chapler 34 ill Ullil Opera/iolls, Wiley, New York, 1953.
Ca mp. T. R.o "Flocculation and Flocc ulati on Basins," TrailS. ASCE, 20: 1 ( 1955).
Ca mp. T. R. oct al. . "Veloc ity Gradients and Intern al Work in Fluid Motion," 1. Bostoll Soc.
Civil EI/g .. 30:2 19 ( 1943).
Cheng. R. c..
ct al.. "Enhanced Coagul ati on: A Preliminary Evalu ati on," 1. AWWA, 87:2:9 1
(February 1995).
O'ozes, G., et aI. , "Enhanced Coagu lation: Its Effec t on NOM Removal and Chemi cal Costs."
1. A IVIVA . 87: I :78 (January 1995).
Davis. C. v.. "Water Treatment," Chapter 38 in Handbook 0/ Applied Hydraulics, 3rd cd.,
M cGraw- l-lill , New York. 1973.
Dempsey. B. A .. et aI. , " Pol yal uminulll Chloride and A lum Coagul at ion of Clay-Ful vic Acid
Suspensions." 1. AIVIVA. 77:3:74 (March 1985).
Dempsey, B. A., et al .. " The Coagu lation of Humic Substances by Means of A luminum Salts,"
1. AWlVA. 76:4: 141 (April 1984).
Dentel. S. K .. et al.. "Usin g Streaming Curren t Detectors in Water Treatment," 1. AWWA,
8 1:3:85 (March 1989).
Edwards, M. , " Predi ctin g DOC Removal During Enh anced Coagul ati on," 1. AWWA, 89:5:78
(M ay 1997).
Fair. G. M .. and Geyer, I. C, Water SUPI)ly and Wastewater Disposal, Wil ey, New York. 1954.
Fai r. G. M .. et al.. Warer and Wa stewater EI/gil/eering, Vol. 2, Wiley, New York. 1968.
Haberer. K .. and Schmidt. S. N., "The Haberer Process: Combin in g Contact Floccul ati on,
Filtration and PAC Adsorption." 1. AlVlVA. 83:9:82 (Se ptember 199 1).
Hanson, A. T.. and Cleasby, J. L.. "Effect of Temperaturc on Turbulent Floccul ation: Fluid
Dy nami cs and Chemi stry," 1. AWWA . 82: II :56 (Novc mber 1990).
Hudson. I-I. E.. Jr. , Wat er Clarification: Processes, Practiwl Design, and Evalua tion. Van
Nostrand Reinhold . New York. 198 1. pp. 40-64. 54-74. 75- 100, 111 - 122.
Hundt. T. R .. and O' Melia, C R" "Aluminulll-Ful vic Acid Interaction: M echani cs and
Applications." 1. AIVIVA. 80:4: 176 (April 1988).
Jamcs. C. R .. et nl.. "Considering Sludge Producti on in Sclcct ion of Coagu lants," 1. AWWA,
74:3: 148 (March 1982).
Kawamura, S., "Appli cati on of Coll oid Tit rati on Tcc hni que to Flocculati on Co ntrol," 1. AWWA,
59:8: 10m (August 1967).
Kawamura, S.. "Coagulati on Cons iderati ons," 1. AWWA, 65:6:4 17 (June 1973).
Kawamura, S.. "Optimization of Basic Watcr Treat ment Proccsscs-Dcsign and Operation:
Coagul at ion and Flocculation," 1. Water SRI-Aqua, 45: I :35-49 ( 1996).
BAS IC APPROACH 139
Knw:1IllUfa, S. "Remova l of Color by A lum Coagulation," Pa rts I and 2. \Vater Sewage Worh,
p. 282 (August 1967). p . .124 (September 1967).
Kawmnu ra, S.. " The Fund ament als orA lum Floccu l:lI ioll as Applied to Water Purification."
thesis, O hi o State Uni ve rsity, Columbu s, 0 1-1 . 1961.
Lcciltraar. 1.. et al.. "Some Dimensionless Para meters of Impel ler Power ill
Coagul ation- Floccul at ion Process," Water Res. , 14: 140 ( 1980).
Montgomcry, J. M .. Consultin g Engineers, Wal eI' TreMmel/t: Prill ciples alld Design, Wiley,
New York. 1985. pp. 11 6-134.514-523,
Naj lll. I.. et al.. "Optimi zati on Ee wit h PAC Addition: A Casc Study." 1. AWWA, 90: 10:88
(October 1998).
Oldsiluc, 1. Y. "Flocc ul ator Impcll ers: A Co mpari son," e ltell/. Ellg. Pmg., p. 72 ( 1983).
Oldshue, 1. Y.. FllIid Mixin g Technology, McGraw-Hili. Ncw York. 1983.
Schul z, C. R., et al .. "Evalu at in g Buoyant Coarsc Media Flocculation," 1. AWIVA , 86:8:5 1
(A ugust 1994).
Tckippc, R. J.. and Ham. R. K .. "Velocity Gradicnt Paths in Coag ul ation." 1. AWWA, 63:7:439
(Jul y 197 1).
Thompson, J. D .. ct aI. , "Enh anced SoflCning: Factors Inllucnci ng DBP Precursor Rcmovn l."
1. AlVlVA. 89:6:94 (June 1997).
Von Essen. 1. A. , "Encrgy Saving with Exotic Vessel and Impell er Dcsign," PlVceedill gs of
the 9th Engineerillg FO/Illda /ioll Conference 0 1/ Mixil/g, New England Collegc.
Vrijcnhock. E. M .. et aI. , " Remov ing Particles and THM Precursors by Enhanced
Coagul at ion." 1. AIVIVA. 90:4: 139 (April 1998).
Walker. J. D .. " Hi gh Energy Fl occul ation U nit ." 1. AIVIVA . 60: II: 127 1 (November 1968).
Whitc, M. C.. et aI. , "Evalu ating Criteria for Enh anccd Coagulati on Compli ance ," 1. AWWA .
89:5:78 (May 1997).
3.2.5 Sedimentation (Clarification) Process
Purpose One of the 1110st C0 l111110n wa ter an d wastcwater treatm ent unit processes
is sedimentation, also known as clarification. Sed imentation is broadly deli ned as
the sepnra ti o n of a suspension in to a clarilicd llu id and a more concentrated sus-
pension. The sedimentati o n process i s designed to rem ove a majorit y o f the se t-
tleable solid s by g ravitati o nal se ttling, thereby ma x imi zing the d ow nst rea m unit
processes suc h as liltration . Fl oa tati o n also separates liqui ds from solid s within a
suspensio n, However, buoya ncy i s the key to thi s process.
The sed i men tatio n process is di v ided int o two c lassifications: grit c hamber ( pl ain
sedim entati on) and sedim ent ati on tank s (c lariliers). The c riteri a for these classili -
catio ns are thc size, quantit y, and specilic gravity o f th e suspen ded solid s to be sep-
arat ed. In thi s sec ti on w e d i sc uss the des i gn o f a sed imenta ti o n process hased O il
both theory and the experience of the author, ill addition to lield o pera ti onal data.
I! must bc emphasi zed that the efli c iency of the sed im entati on process i s g rea tl y in-
Ilucnced by the level and adequacy of raw water conditionin g upstream from the
scdim entation process . The key to effecti ve c larilication i s proper coag ul ation and
Ilocc ul ati o n of sus pended matter within the raw wa ter.
140 DES IGN OF BASIC TREATM ENT PROCESS UNITS BAS IC APPROACH 141

There are three main conliguralions for sedimentation tanks: hori zo ntal rectan- to 0
gular bas ins, uptlow sedimentation tanks, and upnow reactor clari fi ers with sludge 0
9 0
blanket. A rectangular tank with a horizo ntal fl ow is ge nerall y the favored tank con-
liguration due to its hydraulic stability (if properl y des igned) and its to lerance of
shoc k loadings. T his type of tank also has a predi c table performa nce and is capa- ...
:;:
8
0
0
0
0

ble of withstandin g a !low rate that is twice the ordinary design rate recolllmend ed ~
.~ 7
by the regulatory age ncies, without significant deteri oration in the quality of the set- ' §' 0
,
tled water. Furthermore, the horizo ntal rectangular basins arc simple 10 operate and ~
6 0 0
:;
adapt eas il y to the ins tallat ion of hi gh-ral c scttler modules.
•
"~;; 5
0 0
0
0 0
Considerations T here are man y important considerati ons that direc tly affect the ~
0
0
0
4 0 0
design of the sedimentation process: (I ) overall treatment process, (2) nature of the '0 0
1:: 0 0
suspe nded matter within the raw water, (3) sett ling velocity o f the suspe nded parti- 'Q 0
cles to be removed, (4) local climatic co nditi ons, (5) raw water characte ri stics, (6) :e, 3 0
0
0
00 0
0

geological characteristics of the plant site, (7) vari ations in the plant fl ow rate, (8) ....
2 0 0 0
occ urre nce of fl ow short-circ uitin g within the tank , (9) type and overall con fi gura- 0
0 0
tion of the sedimentation tank, (10) design of the tank inlet and outlet, ( I I) type 0 0 0
and selection of high-rate settling modules, ( 12) method of sludge removal, and ( 13) 0 0 0
cost and shape of the tank. T hese subj ects, along with other miscell aneo us items, 0
0 0 .2 0 .4 0 .6 . 0 .8 1.0 1.2 1.4 1.6 1.8 2.0
are di sc ussed in deta il.
Turbidit y of filtered water (silica scale)

Overa ll Treatment Process When designi ng an overall treatment process, the
source of the wate r mu st be e va luated care fu ll y. If the source is a ri ve r with
e pisodes of fl as h fl oods, the treatme nt process tra in should include a grit cham-
ber so that sand and silt will se n le out nca r the intake. A grit chamber is a plain
sedime nt ati o n process th at re moves disc re te particl es larger tha n 15 fun in di am-
e ter. If the tIlte r bed is designed as an ordin ary ra pid sand bcd, the degree o f clar-
ificati o n is import a nt because the turbidit y o f th e filtered water is directly pro-
portional to the settl ed water turbidit y (Figure 3.2.5- I). Thus a conservati vely
designed sedim e nt ati on process should be used to obtai n a se ttl ed water turbidity
of less than 2 ntu .
As di sc ussed in Secti on 3.2.4, n ocs th at are both sma ll in size and ph ys ically
strong, obtained by adding polymer as a filtrat ion aid, are needed for multimedi a or
coa rse, dee p filter beds. With these characteristics in mind , the criteri a used in the
design of the sedime nt ation basin mu st be d iffe rent from those used in the des ign
of class ic rapid sa nd filters, so that the overall process e llic iency is max imi zed .
Data obtained from many ex istin g plants de monstrate that dual-m edi a filters re-
main very e ni c ient des pite filter influent turbidities rang ing from 5 to 7 ntu, pro-
vided that solids in the raw water are pro perly co nditi oned by coag ulant (Figure
3.2.5-2). The direct liltration process also supports this charac teristic. The sedi-
mentati on process can there fore tolerate much higher hydraulic loading rates when
using mult imedia or coa rse, deep media filt er beds. In general, when du al-media fil -
ters are used, the hydraulic loading rate of the sedime nt ati on process ca n be twice
or even greater than the rate Llsed in rapid sa nd filter system.
Figure 3.2.5-1 Relat ion ship between turbi dity of settled and filtered water on co nventi onal
mpid sand filter. Notes: (I) Filter rate ranged from 4.2 to 6.7 mlh ( 1.7 to 2.7 gp mlft').
(2) A lum flocculation and sed imentation before filtrati on. No filter aid was appli ed. [Reprinted
from JOlIl'/la / American WaleI' Works AssocialiOIl, Vol. 67. No. 10 (October 197 5). by per-
mission. Copyright © 1975. A merican Water Works Associat ion.]
Nature of Suspended Ma tter Raw water contains two bas ic types of susJlended
matter: particl es th at are di scre te in nature (nontl occ ul able) and colloidal suspen-
sions. Examples of lIolljlocculable parti cles are sa nd and si lt. Colloidal suspe nsio ns
incl ude clay, microorga ni sms, and substances that cause color. Colloidal suspe n-
sions mu st be fl occ ulat ed with coagulant so th at their partic les wi ll be re moved dur-
ing sedim ent ati on and fi lt rati on. Di sc re te part icles may be removed by a cyclone
separator, pla in sedime nt ati on, or th roug h the use of a grit chamber. Howeve r, col-
loidal sLlspensions req uire a different type of settling bas in (clarifier), one that has
a much more co nservati ve hydrauli c loading rate. This is expla ined in the secti on
covering des ign cri teri a.
Settling Velocity of the Particles The sedimentation process is based on gravita-
tional settli ng of the part icles. For this reason, the design engineer should know the
settling velocities of each partic le that is to be removed in the antici pated water te m-
perature range. T he sett ling velocity of each partic le in sti ll water may be measured
under cont rolled co ndit ions within the laboratory. Be nch-scale fl occ ul ation tests (jar
tests) and the measureme nt of fl oc settling ve locities are there fore strongly recom-
142 DESIGN OF BASIC TR EATM ENT PROCESS UNITS t'!":
\Q Q'\ N J. -.b
;£g;~g~~Q'\.nMOOOO

~
8 0 :c
.~
S c.
.... E
~ 7 o
, u
•
;;; 00
0

u• 6
0
0

-" 0
5
R 0
0 0
"0
!: 4
0 00
'0 0 0 - M
:0

r--MNN-O\oN~~~8i~
0 0
:; 3 0
.... 80 0 .§.
o\...tNoO-.:itri.....:.....:OOOOOO
2 0
0

0
0 0.05 0.10
~
Turbidity alliltered water (NTU) E

Figure 3.2.5-2 Relationship between turbidity of settled and filtered water on typical dual·
media filter. Notes: (I) Filter rate ranged from 10 to 24 mIll (4. 1 to 9.4 gpm/ft 2).
(2) Polyelectrolyte (as filter aid) dosage ranged from 7 to 23 ppb. lRcprinted from JOllmai
American Water Works Association, Vol. 67, No. 10 (October 1975). by permission. Copyright E
© 1975. American Water Works Association.] E

mended. As previollsly mentioned, an appropriate safety t~\ctor should be applied

since floc settling velocities can vary widely depending on several parameters: type
of coagulant employed, mixing conditions during the Ilocculation process, and the
nature of the colloidal matter contained in the raw water. If these tests cannot be per-
formed, for whatever reason, the numbers listed in Table 3.2.5-\ may be used.
According to Hazen (1904), Ihe efticiency of an ideal horizonla l flow sedimen-
tation tank is a function of the settling velocity (va) of the particle to be removed, - N

the surface area of the lank (A), and the rate of flow (Q) through the basin. This re- ~~:2~~:2:2~:2:2:2~88
N0iNNNNNN0i0iNN""':""':
lation can be expressed as

Va = Q
A

QIA is commonly known as the surface loading or overflow rate and is expressed
in m/h or gpm/ft2 Once Ihe selliing rale of Ihe suspended maller (being removed)
is determined and the rate of flow through the tank is known , a physical basi n size
(area of the tank) can be calculated. A minimum safety factor of 1.5 should be in-
cluded in the calculation of the area.
143
144 DE SIGN OF BASIC TREATMENT PROCESS UN ITS BASIC APPROACH 145

According to Ha ze n, the e ffi ciency of the tank is indepe nde nt of its depth and
detention tim c. Howeve r, a shall ow de pth theoreti call y fa vors settlin g of the parti.
c les. Dete nti o n tim c also affects tank e ffi c iency sin ce fl occ ul ated particles become
large r and heavie r with time, due to the mi xin g effect in the tank , and there fore set·
ti c mll ch faster. Co nseque ntl y, the dete nt io n time of the bas in is sig nificant with re·
gard to tl occ ul a nt partic les.

Local Climatic Conditions Several local climat ic conditi o ns must be considered

whe n des ig ning a sedime ntati on process: dail y and seasonal temperat ure fluctu a·
ti on, water and air te mperature ran ge, deg ree of rainfall , and inte nsity and directi on
of the winds. If the dail y temperatures fluc tu ate widely, within a day or a week, the
bas in will 1110s t like ly ex perie nce tl ow short·circuiting, d ue to a de nsity c urrent.
Moreover, regio ns w ith severe w inter conditi o ns should ge nerall y have the tank ei·
ther cove red or housed in a heated building.
Water treatme nt pla nts located in reg ions of warm o r ho t wea ther wi ll have prob·
lems wi th alga l bloom and have a heavy gro wth o f vege tati o n in the watershed . If
th e reg io n is bo th wa rm and sunny. the treatment plant will always have problems
with al gal grow th o n th e basin wall s and launders. Thi s can be preve nted by e ither
covering th e bas in or pro viding sufti c ient chlorine res idual to the influent water.
However. chl or ine sho uld be added with ex treme cauti on beca use of the likelihood
that T HM formati o n w ill exceed the tolera ble limits.
Treatment plant s located in reg io ns of hi gh rainfa ll te nd to have large amo unts
of sus pe nded matter in the raw water. The des ig n sho uld there fore include a con·
tinuo us sludge removal syste m th at will re move the acc umul ated sludge before it
red uces the e ffec ti ve tank volume and be fo re it becomes septic. In cases where the
raw water cont ains a hi g h conce ntrati o n of g rit and silt , a presedime ntation bas in
must be in cluded in the des ig n.
In regio ns other than the tro pics, a major ity of th e local vegetati o n sheds leaves
into the water so urce, as well as the bas in, d urin g the winter, causing o rganic color
to be imparted to the water, creating low pH and clogg ing the lilter bed. Floc formed
from hi ghl y colo red water throu gh the usc of hi gh co ncentrati o ns of alum o r ferric
sa lts (enhanced coag ul ati on) is very frag ile and difficult to settl e. Therefore, both
submerged and s urface metal s and co ncrete surfaces will corrode unl ess spec ial de·
sig n alte rati o ns arc made.
Reg ions th at ex peri e nce strong w inds co ntinually have des ig n problems of the ir
ow n. Stron g winds can greatl y di sturb th e water Il ow within the cl ari fi ers and may
induce fl ow short-c ircuiting. depending o n the direc ti o n of the wind and orientati on
of the bas ins. Furthermo re, if trees are located near the settl ing bas ins. the w inds
Ill ay also blow leaves a nd branches into them .
Raw Water Characteristics One of th c mos t dramati c e ffec ts of raw water char-
acteri stics with regard to bo th hori zo nt al Il ow and upll ow c1 m'iliers is th e s udden
change in water tc mperature. A change of 1°F (O.SOC) ca n ca use a severe density
Il ow (F ig ure 3.2.5-3). If the inllue nt water is colder, it w ill trave l along the bottom
-c
r- S O::C:...:::~~ 3
?";,;--/-17!.:..oc
6 5 4 t
17.6 °C I
Side view
Modified tank under den si ty flow conditions
Figure 3.2.5·) Original and modified tank s under densit y now condi tions. Observations are
based on a hyd raul ic scal e model study. Notes: ( I ) Computed detenli on time is 49 min or
4 h 15 min for the ac tu al tank . (2) Observed Il ow.through tillle is 9 min for the ori gin al tank
model and 17 min for the modifi ed mode l. [R eprinted fro m )01/1'1/(11 American Wmer Works
Association, Vol. 73. No. 7 (J ul y 198 1). by pe rmi ssion. Copy ri ght © 198 1. A merica n Water
Works Assoc iation.]
and ri se up ward toward th e e nd of the lank . If th e wa tcr is wann er, fl ow short·
circuitin g will occ ur at the surface of the tank. This co nditi o n is particul arl y seri -
ous whe n a treatm ent plant has more than o ne source and these sources arc oft cn
sw itched. or if the ble nding rati o of two or three sources is frequ e ntly c hanged.
Turbidity s pikes also cause densit y c urrent s th at are ide nti cal to th ose c reated by
cold influe nt wa ter. Hi g hl y turbid wate r produces a large vo lume of fl oc aft er 110c·
cul ati on. Since th ese fl ocs ha ve a hi ghe r s pec ific grav it y tha n the water in the tank ,
the influent di ves to the bOllo m, creating a di stinct de nsit y fl ow. A raw wa te r tur-
bidity of over 30 ntu usuall y produces thi s type of densit y fl ow.
Att enti o n sho uld al so be g ive n to the aggress iveness of the raw water to sub·
merged parts (both concrete and metal ). The des ig n engineer sho uld either choosc
an appro priate type of cement and prov ide protec ti on by appl yin g a spec ial coating
or lise noncorrosive materi al sllc h as type 18·8 stainl ess steel. Subme rged steel part s
146 DESIGN OF BASIC TREATMENT PROCESS UNITS
must be properly painted and preferably have cathodic protection . Although type
18-8 stainless steel is significantly more expensive, its usc generall y will eliminate
the need for costly operation and mainlcrwllce procedures, slich as painting and ca-
thodic protection .
Soil Characteristics Consideration must also be g ive to the soil characteristics
of the plant site. Of special interest arc the bearing capacity of the native soil , the
presence of geological faults, the level of the undergrou nd watcr table, and the salt
content of the soil. Inadequate geotechnical studies have much too often resulted in
the subs idence of the clarifiers and the development of large cracks in the basin
walls and !loor slabs.
A high salt content within the soil quickly corrodes metals, including reinforc·
ing steel bars embedded in concrete. This problem may be minimized by providing
a concrete cover that is greater than 2 in. A soil resisti vi ty below 1500 J.d1/cm is
generally considered to be highly corrosive, and the use of steel shell tanks should
be avoided unless special measures for corrosio n protection are implemented. Under
these conditions, special cement (type II or V) and a low-shrinkage concrete mix-
ture must be specified for the concrete.
Variation in Plant Flow Rate Sudden changes in plant now rate, also known as
"ydraulic s"ock loads, may affect the now condition and disturb the settling effi-
ciency of the tank. This is especially true with sharp increases in flow rate during
short periods of time. Clarifiers that are most susceptible to hyd raulic shock load-
ing are the upfl ow clarifier and the sludge blanket reactor clarifier. The rectangular
horizon tal !low clarifiers generally tolerate hydraulic shock load ings much better
than the types mentioned previously.
Flow Short-Circuiting Three basic types of !low short-c ircuiting may occur within
the clarifiers. The first is caused by improper design . The second, although not se-
ri ous, is a chronic fl ow short-circuiting th at becomes evident when a sign ificant
amOllnt of floc is carried over to the filters. The third type is the creation of a den-
sity fl ow within an ordinary sedimenlation tank. This density flow is caused by a
difference in density between the influent water and the water wi lhin Ihe tank; the
influent water Illay be colder or wanner, contain a hi gher concentration of suspended
solids, or create a salinity pulse, that is. an influx of NaC) due to the intrusion of
seawater.
The first type of !low short- circuiting is typified by a tank design that has both
a poor inlet design and a short distance between the inlet and the outlet. Other fac-
tors thai exacerbate this problem are a large lank width in conjunction with a short
lank length and a deep water depth. A lank designed in thi s manner will always have
a high Reynolds number and very low !low stability.
The second type of flow short-circuiting is observed in many sedimentation tanks
whose plants are based on the prevailing tank design critcria set by the regulatory
agenc ies. T he ord inary plant visitor may not notice the fl ow short-circuiting, but
plant operators observe that (during the day) the influent watcr tends to dive down
BASIC APPROACH 147
at the inlet of the lank and rise at the effluent, carrying much floc with it. The op-
erators call this phenomenon I/ool/day IlImovel: In actuality, the most seriolls fl ow
short-circuitin g tends to occur mainly from midnight to dawn. In this case, the in-
fluent water short-circuits at the surface due to its slightl y higher water tempera-
turc. As long as the top of the tank is exposed to the weather, both the first and sec-
ond types of !low short-circuiting cannot be controlled effecti vely by the placement
of diffuser walls at both the inlet and outlet of the tank.
The third type of now short-circuiting is a severe case of the second. This den-
sity fl ow problem is commonly caused by switching from one source of water
supply to another, suddenly changing the mixing ratio of two and sometimes three
different so urces, abruptly shifting the reservoir intake elevation, and having a
hi gh raw water turbidity (over 50 ntu). Thi s particular type of now short-circu it-
ing may also be induced by strong winds blowing along the longitudinal axi s of
the tank or by a sali nity pulse. Nothing can stop this type of density now once it
has started.
This third type of flow short-c ircuiting can be minimi zed by installing interme-
diate diffuser walls perpendicular to the !low directi on, in the middle or at two- thirds
of the tank length . The minimization is achieved by increasing the detention time
and mixing effect in front of each intermediate diffuser wall (see Figure 3.2.5-3).
The mag nitude of the density current may be evaluated by Harleman's formllla:
v = (8g ~P I( I Its+ <X) )0.5
where
v = ve locity of the density !low (ft/s)
g = acceleration of gravity (32.2 ft/s')
~p = density difference between the two liquids (lbIft3)
p = density of the innuen t (lbIft3)
" = depth of the densit y current !low (ft)
s = slope of the channel bottom
I = Darcy- Weisbach friction factor
ex = correction factor for kinetic energy rang ing from 0 to I
(0.43 for turbulent !low)
A similar equation is recommended by the Tokyo Metropolitan Government
Report (1963) for computing the velocity of the density now:
v = 2g=--
~n ~,, <x )0.5
( p k
148 DESIGN OF BASIC TREATM ENT PROCESS UNITS
where
v = velocity of the density tlow (m/s)
~p = differcnce in density (glcm')
p = density of the intluent (approximately 1.0 glcm')
g = acceleration of gravity (9.81 m/s2)
~h = depth of the density tlow (m)
a = correction factor for kinetic energy (approximately 0.5)
k = dilution factor (normally 1.5 to 2)
According to the field measurements conducted by this author, the dirference in wa~
ter temperature (near the surface of the tank and the inlluent water) is usually 0.2
to 0.5°C during the day and the tlow velocity of the density flow is 2.6 to 6 ft/min
(0.8 to 1.8 m/min) despite a design flow velocity of 1.3 ft/min (0.4 m/min).
Type of Sedimentation Tank Upflow clarifiers and rcac tor clarifiers perform very
well if both the raw water characteristics and the hydraulic loading rates are con~
stant. The reactor clarilier is best suited for lime or lime with soda ash softening of
the water because of the seeding effect of the unit. Consequently, lime softening of
underground water is the perfect application for the reactor clarifier. However, these
two types of sedimentation tanks are very susceptible to hydraulic and solids shock
loading. Wind and uneven exposure to sun can also di sturb the elTectiveness of the
clarification process.
The upllow clarifier and the reactor clarifier are usually compact and predesigned
by the equipment manufacturer. The overall cost Illay therefore be cheaper than us~
ing custom~designed rectangular tanks. Furthermore, the sludge call easily be re~
moved from these tanks, due to the continuously Iluidized sludge blanket. On the
other hand, a signilicant amount of metal is used on the inside of these tanks and
the problems of corrosion and mainte nance mllst be addressed. It is necessary to
emphasize that upflow and reactor claritiers are not capable of functioning at rates
well over the designed rates because floc removal is based on th e balance between
particle settling velocity and vertical flow velocity.
Unlike the upllow and reactor clarifiers, the horizontal rectangular basin can tol~
erate both hydraulic and solid shock loadings. In fact, rectangular basins can han~
die a flow rate that is 50 to 100% over the design rate without signilicantly de-
grading the quality of the settled water. The flow condition of the basin, under higher
flow rates, is more stable because of an increase in the Froude number. Most large
municipal water treatment plants use hori zo ntal rectangular clarifiers primarily be~
cause of their flexible performance, predictable settling efficiency, and minimum
maintenance cost. A comparison between the upllow and rectangular clariiier is pre~
sented in Table 3.2.5-2.
BASIC APPROACH 149
overall Configuration The preferred configuration of the sedimentation basins is
a battery of rectangular tanks sharing comlllon sidewalls. The upstream portion of
each basin contains a flocculation process, and each rec tangular tank is connected
by a common influent and effluent channel; both channels span all the basins. This
configuration olTers several advantages: a reduction in construction cost, minimum
site requirement due to the lack of space between the basins, easy distribution of
the plant influent water to each basin, minimum floc breakage, minimum floc set~
tlement between the flocculation and sedimentation basins, simple relilling of the
basins with clean settled water from the effluent end via a valve, easy single~point
application of water treatment chemicals to the combined settled water (prior to the
filters), and minimum yard piping.
Another configuration that is often adopted is a circular array of reactor clarifier
basins arranged in a grid point fashion or that follow the topography of the plant
site. The only disadvantage with this type of configuration is that each tank requires
individual inlet and outlet yard pipings, and adequate spacing must therefore be pro ~
vided between the basins. The reactor clarifier basins have a flocculation zo ne in
the center and sedimentation zone in the outer portion; they are proprietary units
designed by the equipment manufacturers. The reactor clarifier basin can be either
circular or square in shape. The configuration of the basin is strongly influenced by
the shape, topography (slope), and geographic condition of the plant site. Thus the
design engineer may not have much of a choice in determining the basin configu~
ration.
Inlet and Outlet of the Basin Regardless of the shape and the type of clarifier,
hydraulic control of the basin influent is one of the most important design factors.
As expounded in many technical articles, any flow imbalance at the basin inlet will
lead to flow short~circuiting, jetting, turbulences, and an overall lack of hydraulic
stability within the settling zone. Although various baming methods have been tested
for distributing the water to the settling basin at the basin inlet, the most simple and
effective method to date is the perforated bame. The most effective perforated baf-
fle, for a specific basin shape, can only be determined through hydraulic scale model
studies. However, certain general rules may be used in designing the baffle walls
for most basins. The author conducted an extensive hydraulic-scale model study in
the late 1950s and determined the general criteria for the perforated bame. Based
on this study the major design requirements are as follows:
I. The ports should be uniformly distributed across the bame wall, which in turn
covers the entire cross section of the basin.
2. A maximum number of ports should be provided so that the length of the jets
is minimized and the dead zone between ports is reduced.
3. The head loss through the ports should be approximately 0. 12 to 0.35 in. (0.3
to 0.9 mm), to equalize the flow distribution across the entire cross section of
the tank inlet with minimum floc breakage.
'"
o

TABLE 3.2.5-2 Selection G uide for Some Basic Types of Clarifiers

Type of Clarifier Some Design Criteria Advantages and Di sadvantages Proper Appli cation
Rectangular basin Surface loadi ng: 0.34-1 gpmlft 2 I. More tolerance to shock loads Most municipal and industria l
(horizon tal fl ow) (0.83- 2.5 m/h) 2. Predictable performance under most water works
Water depth: 3- 5 m conditi ons Particul arly suited to largc r capacity
Detention time: 1.5-3 h 3. Easy operation and low mai ntenance plants
Width/length: > 1/5 costs
We ir load ing: <15 g pmlft 4. Easy adaptat ion to high-rate sett le r modul es
( I I mlm· h) a. Subject to density flow creation in the basi n
b. Requires carefu l design of th e inlet and oulet
structures
c. Usually requi res separate tlocculation
faci li ties
Upfl ow type Circular or square in shape I. Econo mi cal co mpact geometry Small to mid-sized muni cipal and
(radial -upflow Su rface loadi ng: 0.5-0.75 gpm/ft 2 2. Easy sludge removal indu stri al treatment pl ants
type) (1.3- 1. 9 mlh) 3. Hi gh cl arificati on e fticiency Best suited where the rate of fl ow
Water depth: 3-5 m a. Problems of tl ow Short-circuiting and raw water qu ality are constant
Settling time: 1-3 h b. Less tol erance to shock loads
Weir loading: 10 gpmlft c. A need for more careful operation
(7 mlm . h) d. Limi(3tio n on the practical size of the unit
e. May require separate flocculat ion facilities

Reactor clarifiers R occu/alion time: approx 20 min
Settling time: 1-2 h
Surface loadi ng: 0.8- 1.2 gpmlft2
(2-3 mlh)
Wei r loadi ng: 10-20 gprnlft
(7.3- 15 mlm . h)
Uptlow velocity: <50 mm/min
Sludge blanket Fl occu lat ion time: approx 20 min
clarifiers Sett li ng time: 1-2 h
Surface loading: 0.8- 1.2 gpmlft
2 3.
(2-3 m/h)
Weir load ing: 10-20 gpmlft
(7.3 to 15 mlm . h)
Upt10w veloc ity: < 10 mm/ min
Slurry circulation rate: up to 3-5
times the raw water intlow rale
I. Incorporates fl occ ulation and clarilication in Water softening ( \ .5-2 gpmlfl)
one unit A plant thaI treats a steady quality
2. Good flocculation and clarification effic iency of raw water
due to a seedi ng effect
3. Some ability to take shock loads
a. Requires greater operator skill
b. Less reliabi lity than conventional due to a
dependency on one mixing motor
c. Subject to upsets due to thermal effects
I. Good so fte ning and turbidity re moval Water so fte ning
2. Compact and econo mical design Floccul alionlsedi mentat ion treatment
Tolerates limited changes in raw water of raw water with a constan t quality
quality and now ratc and rate of now
a. Very sensitive to shoc k loads Plant treatin g a raw water with a low
b. Sensi tive to temperature change content of solids
c. Several days required to build up the
necessary sl udge bl anket
d. Plant operati on depends o n a sing le mixing
tlocculati on unit motor
c. Hi gher maintenance costs and a need for
greater operator skill

Notes: ( I) The reactor clarifiers and the sludge blankcttype clarifiers are often considered to be in the same category.
(2) Surface loading: m/ h = (mJ/m:!:· d) -:- 24: therefore. 60 m J/ m . d = I gpmlft 2.
2

(3) Wei r loadi ng: rn/h . ttl = ( m~/d . Ill) -:- 24: therefore, 175 m-~/d . m = 10 fpm/f1.

'"
~
152 DES IGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 153

4. The head loss th ro ugh the port s should be less th an 0.4 in . ( I cm) to prevent

X X X ~ X}t
Ilac breakage. ~
5. The size of the ports should be uniform in di ameter, 3 to 8 in . (0 .075 to 0.2
m), to avo id clogging by algae and other debri s. -:--.. 1
-- - --- _ _ - - - - 1 T:--- Diffuse Wall
6. The ports should be spaced approx imately 10 to 20 in . (0.25 to 0.5 m) on ce n- ~ ~ (typical)
te r to provide struc tural strength for the diffu ser wa ll. II
7. Port cOllligurati oll should be arranged in such a manner th at th e parallel jets X X X X 1'1 L

wi ll direc t the Ilow towa rd the bas in outlet. - - - +--- --- -- -I' I
8. If the basins arc fed from a common channel or if there is an inlet feed chan-
nel zone between the fl occulati on and sedimentati on basins, the inlet diffuser
wa ll should be located approx imate ly 6.5 to 8 ft (2 to 2.5 m) downstrea m of
X X XI'L
'-./ 1
the inlet. T he inlet zone significantly increases fl oc size by allowi ng the proper
/'>.
ge nt ly mi xing effec t 10 occ ur (Figure 3.2.5-4). X /'>. X X~ -- 1

The water ex iting the sedimentati on basin should be uniformly collec ted across
1 - - - f- - -- - - - --~
-- 1
1
an area th at is perpendi cul ar to the proper !low directi on. Beginning ill the late 19505
X X X X~
~ -
1 L
a number of long launders (skimmin g trough) were installed in the sedimentati on
tanks in an att empt to achieve uniform collecti on of the settled water. This prac ti ce
is quite varied for the upllow clari fi ers bu t gen erally not for the rec tangular sett ling
f- - - f - - - - - - - ---~
-- 1
1
tanks. The installati on of long launders in hori zontal fl ow rec tangular tanks is vir-
X X X X I 1
tu all y ineffec ti ve when either a bott om-fl owi ng or surface-running density current
ex ists in the tank. f "I
~ Sedimentation
As discllssed previollsly, mos t sedimentati on basins ex hi bit a bott om-fl owing
density curre nt ; the bas in end wall defl ects thi s current, which then ri ses to the I- Flocculation - - - -- ....,1
Tank ~--
lntet
. Tank

surface, carryi ng fl oc, and fl ows into the end port ion of each launder. Thi s phe-
nomenon has been observed in both operati onal basins and in hydraulic-sc ale
model studi es. In terms of design, thi s phenomenon suggests th at tra nsverse laun-
ders placed away fro m the end wall of th e basin might be more effec ti ve th an fin-
ger (lo ng) la unders. A fe w opera ti onal plants have supported th is proposal. Yet
long launders have three maj or advan tages when used in conj uncti on with rec -
tangular bas ins: th e wa ter level of th e tank remains virtuall y constant regardless
of changes in th e plant fl ow rate; wave ac ti on created by winds is minimi zed; and
high-rate se ttl er modul es, sllch as tube settlers, are eas ily hung fro m the launder
sys tem.
V-notched weir plates are generally att ached to the launders. The weirs create a
free di scharge fl ow th at enters the launder sys tem, along wi th any fl oatin g matter.
Co nseque nt ly, there is a tendency to break the fragile alum fl oc. Unlike th e we irs,
submerged orilices do 110t allow Iloating scum to fl ow into the launder and th ey also
tolerate less acc urate installation. Therefore, in some cases, submerged ori fi ces are
drilled into the sidewalls of the launders. The fl oc break age characteri stics of sub-
merged orilices may be slightly better than those of weirs, but the fl ow velocity
th ro ugh the orifi ce must ra nge from 1.6 to 2.5 fils (0.5 to 0.75 m/s) in order to pro-
Channel
Note: The per/ora ted baffle walls shown in be tween th e f1occulators a re optiona l.
Whe n hydrofoil type mixing blades are used, the re is no need for those ba ffl es.
Figure 3.2.5·4 Inlet channel between Oocculat ion and sedimentati on tanks.
vide a co ntrolled head loss of approx imately 1.5 in . (38 mm ). Some Ooc breakage
is therefore unavoidable.
The previously held idea that the installation of perforated barnes at the tank out-
let is equi va lent to the install ati on of inlet baflles (to collect seu led water uniforml y)
still persists. On the surface thi s concept appears to be reasonable and sound.
However, observa tions of hydraulic-scale model studies and operati onal plants
demonstrate th at the ernue nt baflle is unex pected ly ineffec ti ve. The primary prob-
lem is that the emucnt baflle is only capable of prov iding an insignificant correc-
tion factor for a density curre nt. It is therefore common to see numerous flocs down-
stream of the eflluent bafll e, although the wa ter upstrea m may be very clear. For
thi s reason , the lise of the effluent baffle at the end part of the basin is not recom-
mended. In fac t, the efllu ent bafll e is now seldom empl oyed.
154 DESIGN OF BASIC TREATMENT PROCESS UN ITS

High-Rate Settling Modules The term high-rale sellier refers to any small in-
clined tubes or tilled parallcl plates which permit effect ive gravitational settling of
suspended particles within the modul es. The detention time in the settling modules
ranges from 5 1020 min , depending on the type of module and the surface load ing
rate. Surface loading is conveniently calc ul ated from the area of the basin that is
covered by the settlin g modul e and is ge nera ll y 2 to 3.5 gpm!ft2 (5 to 8.8 m/ h). The
scttling e fficie ncy of modules with such hi gh hydrauli c loading values is equi va lent
to a conventional sedimentation basin with a designed detention ti me of 3 to 4 h
and a surface loading rate of 0.5 gpmlft 2 ( 1.3 m/h ).
The concept of an ideal settlin g lank (a nd the advantages of shallow tank s) was
first advocated by Hazen ( 1904) at the beginning of the twenti et h century. Camp
( 1953) later crea ted the "double- try" settl ing tank whil e applying thi s co ncept.
Around 1955 the fi rs t high-rate settling modul e was commerc iali zed in Japan un·
der the name Uno Separator; the mod ule co ns isted of slo ping parall el phlles s paced
25 to 50 mm apart. Similar modules were later marke ted in Swede llunder the name
Lamella Separator and a tube sellier mod ul e was also marketed by Micro Floc, Inc.
(in 1969) and ot her firm s (United States) .
Short ly after release of the first hi gh-rate se ttl er, a mod ified parall e l-plate mod-
ule was released. This versio n had a se ri es of denector baflles allac hed to one side
of the plates, whic h were des igned to promote Il occu lati on with in th e modul e (Figure
3.2.5·5) to permit grea ter efliciency in c lar ilic ation . However, data from side-by·
side pilot studies s uggest that the effectiveness o f thi s mod ul e appea rs to be o nly
marginall y beller than modul es wit ho ut ba rn es.
Although the hi gh-rate settlers are desig ned to deli ver a certain level of perfor-
mance, th ey do not always fun cti on accordin g to the ir specifications, The primary
reasons for thi s aberrant behavior are ( I) poor Ilocc ulati o n, (2) un even now di stri·
bution to the inlet of the se ttler module, (3) heavy algal g row th in the mod ul es, and
(4) heavy scale (CaC0 3 ) depos its o n the mod ul e. The heavy scalin g c logs the fl ow
paths and, more important, eventu all y leads to corrupti o n of the sys te m.
Dependin g o n the type of parall e l-plate settling module , wa te r e nters either hor-
izontall y and parall el to the plates or upward alon g the 60· tilted pl ates. In the tube
settlers the directi on of fl ow is always upward at 60° to th e horizon tal plane, al-
tho ugh some earli er proto types have the ir plates at a 15° an gle, Under normal ci r-
c umstances, the settl ed noc slides down the inclined plates or tubes to the tank bot-
tOIll for subsequent mechanical slud ge remova l. T he design of th e hi g h-rate settler
modules mu st add ress the foll ow ing issues: settling ve loc ity and charac teri stics for
the suspe nded matte r, Ilow velocity w ithin the settl er mod ul e, s url~lce loading,
Reynolds number a nd Froude number fo r the n o~ within the mod ul e, selecti o n of
the appropriate slud ge collection unit (to be in sta ll ed beneat h the settler modu le),
spacing of the launde r when install ed above the settling modul e, and the support -
in g system,
Ex perience has show n that data from sma ll-sca le pil ot plants (area of 1 In or less)
cannot be used as meaningful des ig n c riteria because pilot studi es ofte n give very
optimistic loadin g rates; the studies are usuall y co nduc led unde r care full y co ntrolled
and ideal conditi o ns, The ac tual desig n c riteri a should be mo re co nservati ve becau se Figure 3.2.5-5 Several lypeS or high-rale scnl cr module.
155
156 DESIGN OF BASIC TREATMENT PROC ESS UNITS
of the potential for nonuni form !l ow cond itions at the inlet to the settler modules as
the res ult of density 110w, im proper tank des ig n, or unfavorable coagul ant dosage
or Il occul ati o n e lli cicncy. Dependin g o n the s urrace area o f the lank that is covered
by the settl er mod ul es, the maxi mum surface load ing value should be limited to 2
to 2.5 gpm/ft' (5 to 6.3 m/h) for co ld reg io ns and 3 to 3.5 gpmlft' (7.5 to 8.8 m/h)
for warm -Io-hot reg ions.
Selec ti on o f the proper materi al for the fabricat io n o f sellier mod ul es is very im-
portant. A modul e I~l brica l e d with paper-thin PVC pl ates o fte n deteri orat es within
10 to 15 yea rs due to wea therin g and wave ac tions in the basins. Some seltler mod-
ul es are made wit h thi c k plas ti c plates and have a mod ul e heig ht over 6 ft (2 m).
Alth oug h morc durable, the cost is ve ry hi gh and o fte n $250 to 300 for eac h 1 ft'
(0.3 m') o r tank s urrace area to cover the mod ul e (late 1990 cost o n the U .S . west
coast). T he preferred material is thin stainless steel pl ates (type 304) because o f the
structural stre ngth and noncorrosive nat ure, with a smooth surface to minimi ze the
adherence o f tl oc and algae grow th .
T he des ig n of the inlet (to th e settler mod ule) is very imporlant because it is es-
sen ti al to have a uniform Ilow into the enti re modul e. In the case of a rec tangular
tan k, the first q uarter o f the basin leng th is ge nerall y devo id o f seHler mod ul es. This
all ows for the insta ll ati on an d main tenance o f s ludge co ll ec ti on sys tems such as
chain-and-ni ght un its, and for heav y Il oe to settl e naturall y and also prov ides an in-
let buffer zo ne that improves the inl et now conditi o n to the settl ers. In tu be settler
mod ul es, an average Il ow ve loc ity o r app roxi mate ly 0 .5 to 0.65 ftl min (0. 15 to 0.2
Ill /min ) th at is normall y used in settlin g 1110s1 alum fl oc. Moreover, an approaching
ll olV ve loc ity of approx im ate ly 2 fl/ min (0.6 m/min ) sho uld be used in the tank up-
strea m o f the modul e install ati o n.
The detenti o n tim e o f the settler mod ul e vari es as a fun cti o n o f the ty pe o f mod-
ul e . Tube settl ers ge ne rall y have a detenti o n ti me o f 3.5 to 5 m in, whi le tilted pl ate
settl ers provide a lo nge r period of time; so m e unit s eve n req uire 15 to 20 min . The
Rey no lds and Frou de num bers are not co ns idered to be part of the desig n criteri a,
yet they act as good g uides in des igni ng the hi gh-rate settl ers. T he Reynolds num-
ber ro r the se ttl ers sho uld be kept be low 200 , pre rerably around 50, and the Froude
number sho uld be greater than 10- 5 .
W he n se lcc ting a con tinu oll s slud ge re moval systcm, th e num ber o f avail ab le al-
tern ati ves is ra the r limited. T he most app li ca ble types o r s ludge collec tor, ro r the
sedim e nt ati o n process, are th e c hain -a nd -fli g ht (fiberg lass-rein forced plas ti c) and
the cabl c-o pcrated un derwa ter bogics with sq uecgees. It sho uld be no ted th at these
types o f s ludge co ll ec tin g systems generall y req uire a min imum water depth (in the
n
tank ) o f 12 (3 .7 m) to acco mm odate th eir install ati on.
Use of the s kim mi ng tro ugh syste m is requi red because it uniforml y collec ts clar-
ifi ed wa te r from the area of the ta nk th at is covered by the seuie r modules. However,
the se ttl er modul es may be hun g from the launders, prov ided th at th e laund ers are
des ig ned to to lerate thi s s trai n. T he launders are usuall y spaced IO ta 13 ft (3 to 4
m) o n ce nter, an d hydraulic loadin g is us uall y less tha n 17 g p mlft ( 12.5 m 3/ m . h).
Fig ure 3.2.5-5 dep icts a few hig h-rate settl e rs.
BAS IC APPROAC H 157
Sludge Removal Method T he settlin g characteristics of a lum rerric noc arc typ-
ili ed by three stages: th e hindered settling, tra nsiti o n, and co mpac tin g stages (Fi g ure
3.2.5-6). Durin g the hinde red settlin g s tage th e n oc slides d own th e s moo th s url"cc
of the pl ates (or tubes), whi ch are in clin ed at a minimum of 60 0 from the horizo n-
tal pl ane, In the tra nsiti on and co mpac ti o n stages the n oc no lo nger slides d ow n as
eas il y, and the a ngle or repose (o f the s ludge) is close to 90' . In most cases o nl y a
few hours are necessary for the settl ed fl oc to reach the co mpactio n stage .
A des ig n tl aw th at is cO lllm o nl y made by in experi e nced eng ineers is to prov ide
a series o f sha ll ow-ang le sludge hoppers on the bott o m o f the basin. In thi s sc heme,
eac h hop per is li tted w ith a pipe co ntainin g a shut o ff va lve as the mea ns o f slud ge
rem ova l. Accordin g to thi s des ig n, the val ve is ope ned two to three times a day a nd
the slud ge is washed ou t with th e water. However, eve n when th e sid ewalls o f the
hoppers arc built 45 to 600 fro m the hor izo nta l, th is pa rt icul ar des ig n docs not work.
The reaso n for the des ig n I~'l ilure is that a wate r pass age, the same s ize as the pi pe
di ame ter, is created in the settl ed s lu dge whe n the va lve is o pe ncd and the major-
ity of the s ludge therefore rema ins settl ed and undi sturbed. To salvage thi s des ig n,
the slud ge Illust be kept in a Iluid co nditi o n at all tim es. Co nseque ntl y, the bott o m
of the sedime ntat io n La nk s hould be des ig ned as a n at surface with or with out a
sli ght slope toward the loca ti on of the s lu dge hoppe r. depe nding o n the me thod o f
mec hani cal slu dge re mova l.
In the Uni ted S lates the three co ntinu ous s ludge re moval sys tems that have bee n
proven and wide ly used in recent yea rs are ce nter-pi voted rotating rakes, the c ha in-
and-ni ght collec tor, and th e trave lin g brid ge. T he ce nter-pi voted rakes have a good
hi story o f serv ice and low main tenance cost. Yet there are four major drawbac ks to
thi s partic ular sys te m: ( I ) a large amount of metal is lIsed for the rakin g assem bly,
creating corros io n and assoc iated maint e nance pro bl e ms ; (2) the ce nt er-pi vot po r-
tion of the tan ks Ill llst be very dee p (over 20 rt ) wi th a conicall y shaped bo tt o m,
which signilica ntl y inc reases ta nk co nstruc ti o n cost; (3) even if the ex pandable arm s
are used to cover the corn ers o f the sq uare area, there is in adequate s lu dge remo va l
in the corners; and (4) the cap ital cos t for th e system ca n pote nti a ll y be hi g h if
the recta ngul ar ta nk has a le ng th/width ra ti o of 3 : I o r larger. Furthe rm ore, pl a nt
Compaction
Tim e
Figu re 3.2.5.6 Selliing characleristi cs of al ulll noc. A, Supe rnalant zone: B. discrete set-
tl ing ZOne; C. hindered sett ling zone; D, compression zone.
 158 DESIGN OF BASIC TREATMENT PROCESS UNITS
operators of treatment plants that have only th e fro nt half of the tank covered with
the rot atin g rakes are not sati sfied with its performance. The chain-and-fli ght col-
lec tor has a ma ximum le ngth of approx imately 230ft (70 m) per unit. The travel-
ing bridge collec tor can be installed to any le ngth of tank. It must be emphas ized
that a tank equipped with high-rate settler modules requires a compatibl e continu-
ous sludge removal unit, due to the problem of rapid sludge accumulation. Thi s high
rate of sludge accumulati on may completely obstruct modules such as tube settlers
unless the sludge is removed continuously. Figure 3.2.5-7 depicts the sludge removal
units that are available from reputable equipment manufacturers.
Cost When the geological and civil works are norm al for a plant site, an ordinary
rec tangular sedimentation basin will cost from $80 to $90 per square foot of the
bas in surface area (la te 1990 cos t on the U.S. west coast). As a rul e of thumb, the
cos t of a nocculati on/sedimentati on basin process is approximately 20% of th e to-
tal plant constructi on cost. Th e reac tor clarifiers are ra ther compac t units and man-
ufac turers cl aim th at they are cost-effecti ve, renderin g a 10 to 20% savings in cost.
Studi es of water treatment plant sites show that the ac tu al costs of these two lypes
(a)
(b)
(d)
Figure 3.2.5-7 «I ) Center-pi voted rotating rake collector; (b) chai n-an d- tlight collec tor;
(e) traveling bridge collec tor (suction type); Cd) cable-operated underwater bogie collector.
BAS IC APPROACH 159
of clarifi ers are generally the same. Yet the reactor clarifi ers are preengineered by
the manufacturers, up to 150 ft (4 5 m) in diameter, and the design cost of these units
is therefore virtually eliminated, excluding the foundations and pipings.
High-rate seulers ea n be effective if they are both properl y designed and installed .
There is a vast differe nce in cost be tween the tube seuler modul e system ($ 1OOm')
and some parallel-plate seul er syste ms ($3 50m '; late 1990 cost on the U.S. west
coast). Thu s, selec ti on should be based on a sound engineering evaluati on.
Selec ti on of the sludge removal units is similarl y affected by cost. If the le ngth
of the rectangul ar sedim entati on tank exceeds 230 ft (70 m), the traveling bridge is
more cos t effecti ve: two chain-and-night units (in series) are required 10 service the
sa me dista nce. A traveling bridge unit ean easil y span 100 ft (30 m), and the capi-
tal cost for the equ ipme nt is signifi ca ntl y reduced by hav ing one bridge span two
or three tanks. This type of arrangement can be achieved because the width of each
tank is usuall y less th an 40 ft ( 12 m). Co nversely, the chain-a nd-ni ght unit has a
maxi mum wid th of approx imately 20 ft (6 m). The tank width should therefore be
the aliquot of 16 to 20 ft (5 to 6 m) in orde r to economi ze on equipm ent cost.
Generally, the chain-and-night unit is more economical than the traveling bridge
onl y if the length of the tank is less than 230 ft (70 m); eve n if an addition al cross
collec tor is placed at one end of the tank, th ere is still a savings of approximately
30 to 50%.
Intermediate Diffuser Bafffe Waifs A rec tangul ar tank with one or two inter-
med iate diffuser walls was first introduced in Japan by the author in 1952, and a
number of functi onal large basins were designed soon after. The intermed iate di f-
fuser walls hclp to minimi ze the form ation of densi ty currents because the gentle
verti cal mi xi ng effec t upstream of the di ffu ser wall reduced the temperature differ-
ence between the inlluent wa ter and the water in the tank. This gentle mi xing also
significantl y pro motes the growth of noc size and weight. Consequen tl y. the noc
settl es very well , both in th e first compartm en t and upstream of each diffu ser wall.
Furthermore, stable fl ow patterns occur downstream of each diffuser wall and the
wind effect on the now in the lank is reduced. It must bc emphasized that overu se
of intermedi ate din'user wa lls created difficulties during installation of the continu-
ous sludge removal equipment. The aforementioned type of basin was installed in
the cities of Tokyo, Kobe, and Kanagawa (J apan) under bogie-type sludge collec-
tors. The collectors were placed between the diffu ser walls (ac ross the fl ow direc-
tion), in contrast to the longitudinal direc ti on of 1110s t ordinary tanks.
Okuno et al. ( 1980) (Tokyo Metropolita n Pu blic Work s) co mpared the opera-
tional efficiency of four Iypes of full -scale wastewater sedimentati on tanks; a cen-
tral feed and peripheral collec ting circular tank, a peripheral circular tank, a rec-
tangular tank with a tray (doubl e-decked ), and a rectangul ar ta nk with two
intermediate diffuser wall s. The results of th at study showed th at tank s with two in-
termediate diffuser walls were far superior and th at the double-decked tanks did not
perform sat isfac toril y (Figure 3.2.5-8), due pri marily to the lI nstable fl ow charac-
teri sti cs of th e tank.
 160 DESIGN OF BAS IC TREATMENT PROC ESS UNITS 8ASIC APPROAC H 161

50 Shape of the Tank When tirst designing a rec tangular tank, it appears that there
250
is much freedom in selecting the shape of the basi n. However, experience has shown
that basins that are both wide and deep tend to exhibit flow instability and a dis-
40
200 tinct density flow pattern. On the other hand, a basin that is narrow, shallow, and
long will have flow stability and a minimal amount of fl ow short-circuiting.
30 Rectangular tanks with a 1800 turn at the mid length and crescent-shaped tanks arc
150 both ineffec ti ve because of Ilow short-circuiting; these designs ca n be salvaged only
"
g
,- • g" if a very long detention time is provided. The actual basin design must also take the
20 ••• ," .... 00 following items into considerati on: the installation of sludge removal equipment,
100
~.

~o
the potential installation of high-rate settler modules, the presence of turbu le nce

~
caused by wind s, the scouring of bottom sediments, and the cost of the tank.
10 TIle following guidelines can help to determine the physical shape of the sedimen-
50
tati on tanks: (I) the water depth should be 10 to 16 ft (3 to 5 m); (2) the ratio between
the length (L) and the width (W) should be UW = 6: I , with a minimum of 4: I ;
o (3) the ratio between width (W) and the water depth (H) should be WIH = 3: I, with
60
Overflow rate
Overflow rate a maximum of 6 : I . From an economical as well as a practical viewpoint, the preferred
(a)
(b) height of the tank freeboard is approximately 2 ft (0.6 m). The freeboard acts as a wind
barrier and helps to prevent wind-induced waves from splashing on the walkway.
The now characteri sti cs of the regular sedimentation basin can be estimated by
the Reynolds (Re) and Froude numbers (Fr):

100 vR
Re = -;- < 20,000
,
Fr = -
v > 10- 5
80
gR
20 o
where
~ 60 15 o o v = di splacement flow veloci ty (m/s)
~ R = hydrau lic radius , or R = alP (m)
I
_ ' 40 • 0 •
•
Opo- a = area of flow passage (m2)

20
0~'O~~0 00
P = wetted perimeter (m)
0 0 v = kinematic viscosity (m2/s)
o
o
g = gravity constant (9.8 1 m/s')
o 40 50 60 20 30 50 60
Overflow rate
Overflow rate (m J/m 2 • d) It is rather surprising to discover that an ordinary basin has a Rey nolds number
(c)
(d) that is over 25,000 a nd a Froude number that is less than 10- 6 . Consequently, the
flow characteristi cs of an ordinary basi n are far inferior to that 1110st engineers would
Figure 3.2.5-8 Hydraulic charac teri sti cs of li ve types of clarifi er: 0 CFPC, ce ntral feed pe_
expect. The !low characteristics of an ordinary basin can be improved significantly
ripheral collect; • PFPC, peripheral feed peripheral collect; ... RDD-1. rec tangul ar double
dec k-I ; • RDD-2 , rectangular double deck -2 ; 0 RTMD , rectangular tank with multidiffuser
if longitudina l baffles are insta lled parallel to the sidewalls o f the basin . This arrange-
wa ll. (ll) Short circuit; (b) mean retention time: (c) dead space ratio: (d) Fair's 1/ (stabili za- ment divides the basin into a series of narrow channels, thereby decreasing the fl ow
ti on index). (A dapted from N. Okullo et aI., Analysis of Existill g Clarifiers alld Desigll passage area (a). The proper installation of longitudinal baffles in thi s manner will
Considerations /or Beller Pe,jormallce. Tokyo Metropolitan Gove rnment Report, 1980.) reduce the Rey nolds number to 50% of that of ordinary basins and will increase the
Froude number to the desired range of 10- '.
162 DES IGN OF BASIC TREAT MENT PROCESS UNITS BASIC APPROACH 163

The minimulll wid th between the lo ngitudina l barnes should be 10 ft (3 Ill), to TAUL E 3.2.5·3 Selection Guid e of Prol'r iet;.,ry Cla r iliers
all ow for the install ati on of slud ge re moval eq uipme nt , easy ma intenance, and eco~ Disadvan tages
Type o f Un it Ad va nt ages
nomic considerati o ns. The barnes may be co mposed of thin concrete walls or even
wooden plank s. O ne of th e least des ired shapes is a bas in th at has a 1800 turn at its Dctrillis tank Pote nti al low capiwl cost Fl ow short-c ircuiting. potential clogging
o f the grit re moval pipc. co rrosion o f
Illid leng th. A sig nilica nt number of o perati o nal basins have bee n desig ned in thi s
the rak ing system
mann er, primaril y for space-sav ing reaso ns. T hese tank s have either a horizo ntal or
Ce nte r feed and Low capi tal cost. easy Flow s hort-c ircuitin g. lim itations on the
ve rti ca l turn at the mid length and alway s ex hibit poor e rti ciency. T hi s inefficiency practical size of th e unit. less tolerant
peripheral sludge rClll ov8.1
is due to the prese nce of turbul ence near the area where the water mu st make the o f shock loadin g
co ll ection
turn a nd the ex iste nce of large dead spaces within the basin . Peripheral feed Good performance for raw Potential for !low short-ci rcuitin g.
and periphera l water wi th a hi gh performance rel ics on th e ma nufac tu r-
Upf/ow and Reactor Clarifiers T he upO ow and reae tor c larifiers are proprietary coll ect ion sus pension of so li ds er's des ign. limit ati ons on the practi ·
units that have the ir basic size and blue pri nts preestabli shed by the eq uipment man ~ ca l size of the uni t
ufact urers (based o n the fl ow rates). These unit s ca n be grouped into three categories: Reactor cl arifi er Compact and economica1. Corrosion o f unde rwa ter meta l pa rt s.
simple upll ow c larifiers, reac tor c larili ers, and sludge bl anke t reac tor c lariliers. The with high tolcrant of shock load ing. less reli able due to its depe ndency
upll ow and reac tor clarifiers a lTer several ad va nt ages over the recta ngul ar bas ins: (I ) reci rcul at ion good clarificati on on o ne dri ve uni t. requires ex tens ive
and Illcchan· due to a seedi ng effect yard pi ping to connec t each unit.
compac t des ig n and economical use, since no des ig n e ffort is required by the e n g i ~
ical sludgc pract ica l size lim ita ti o ns
neeI'; (2) sim pler sludge removal; (3) e ffic ient c1arilicati ol1 due to the seeding effect-
plow
thi s seedin g effec t is ac hi eved th roug h recircul ati on of the fl oc and adsorpti o n of the Very se nsiti ve to shoc k loadi ng. req uires
Reactor cl arili er Good lime soft ening and
fl oc by the slud ge bl anket, as lo ng as the Il ow ra te and raw water q uality arc fairl y turbidity removal. 2-4 days to buil d up the necessa ry
with slu dge
co nsta nt; (4) betler clarili ca ti o n e rtic ic ncy duc to improved fl occ ul ati on and ad so rp ~ blanket zone compac t and economical sludge blanket. req uircs g reater
tion effec t by the sludge blanket; and (5) eapabilit y or delaying deg radation of ef- and mec han- design. severa l manu fac- opera tor skills. corrosion of under-
Ilucnt wate r quality, in case of improper coag ul ant dosage or eve n a short interrup· ical slud ge turers to choose fro m water metal parts
tion of coag ul an t feed, d ue to the buffering effect of thc sl udge blanket zone. plow
However, th ere arc a lso seve ral disad va nt ages: (I ) compared to the ho ri zontal Reactor clarilicr Good pe rformance under Sl udge will choke the un it if heavy
tanks they req uire more stri ngent ope rati onal co nt ro l; (2) there is a rap id loss of e f~ wit h hyd rau lic the proper conditi ons. no so lids arc present in the raw water. no
li ciency durin g hydrauli c and solid overl oadin g and hydrauli c shock loading; (3) tl occu lati on mainte nance of mec han· operator cont ro l over the tl occ ul ation
and slud gc ica l equ ipmcnt process. process fun ct ions will
freque nt tempe rature Iluctu ati ons indu ce di stinct fl ow sho rt-c irc uitin g; (4) if the un it
remova l terminate if the wate r dOll si ng syste m
is not eq uipped with a positi ve rakin g mec hani sm, it is difficult to remove the sludge
fai ls. prod uced by onl y o ne man ufac-
du ring peri ods of high turbidity; and (5) more time is required to prod uce the sludge tu re r (patented)
blanket wheneve r th e sludge blanket clarilic r is drained . Moreo ver, a sig nifica nt
amoun t of mctal is used on the interior of th ese tanks and the problem of cOITosion
and mainte nance Illust be add ressed. whenever there is a hi ghe r qu anti ty of s uspe nd ed parti cles and a hi g her propo rti o n
T hese types o f c lari fi ers arc us ua ll y empl oyed in industri al and mu nic ipal appli- of large r suspe nded solids in th e raw watc r d ue to ra in . Fig urc 3.2.4~ I b illustra tes
ca ti o ns, such as lime softe nin g of gro und wa ter and clarifi ca tion of tu rb id water. It thi s pheno meno n; the nocc lil ati o n theory also proves thi s point. T he process ex-
mu st be e mphas ized that the fl ow rates and the raw wate r q ua lity in these situati ons ploits this phe nome non by feedin g microsa nd (20 to 200 /-un ) toget her w ith inor-
are nearly co nstant at a ll times. Ta ble 3.2.5-3 is a selec ti o n g uide for these un its. ga ni c coagul a nt a nd catio ni oc polymer, the reby all owin g the noc 10 att ac h to the
Re fer to Fig ure 3.2.5-9 fo r illustra tio ns of eac h type of unit. surface of the mic rosa nd parti cles. Since the settlin g ve loc ity o f thc mi crosa nd is 20
to 60 tim es hi g her than th at of alum n oc, the hyd ra uli c surface loadin g of th e clar-
High-Speed Microsand Settling Process Altho ug h the prine iple is no t new, in ifier may be 20 to 60 times hi g her th an that of regu lar c larili ers.
rece nt yea rs the process has a lso bee n call ed th e ballasted sa nd process. The prin- The hi g her-s peed microsa nd se ttling process incl udes both Il occ ul ati o n a nd c1a r~
c iple was applied to wa ter trea tm e nt in the I 970s by a French finn and marketed as ificati o n stages. T he Il occ ul ati o n stage has 9 to 10 minutes of mix ing timc with a
the Cyclolloc process. Th is was th en improved by Kruge r, Inc. a nd pate nted in the G value of 400 to 150 s - 1 and tapered mi xin g by VC l'li cal ~s ha ft Il oce ul ato rs. The
Unit ed States (Ac till o). US Filter has commerciali zed a similar process. clari fi catio n stage has 10 minutes of settling time w ith lame ll a plate settl er modules
Exa mi natio n of regu lar pl ant treatment reco rd s ind ica te th at bett er fl occ ul ati on with hyd raul ie surl"ec loading of 14 to 25 gpmlft2 (35 to 62 m/h). In ge neral, the
is ac hi eved, without increasing a lum dosage in proporti on to the level of turbidity, mi crosand feed ra te is 0.1 5 to 0.4% of the process in g fl ow rate and the amo un t of
 ,e!urnpip<!
'-'
,,"",,"

Slurry PooI lndiealed by Shaded Areas

(c)
Rec::ipfoea!ing ,ake mechanism Organic ,elurn
(9f ~ washif>g mtChanism) mechalliSln
Variable Speed
(a) Impelle r O,ive ---l-.!><~

Sludge ToWasle

(d)
Vacoom Vacuum Waler L.... e l eolleclioo Sludge Clar~ied
Raw Vacuum
Ex1 .. ~iDn Waler Chaooel
Cenle, Fe ed CIa,li...
Walef

''''''
Chambe< f>unlI Ve nt
,
float ConIroi Launders
P""
. ~ _I
ffislriMing
Ring Inlel Ring

~
(' ( OuIle!

St.xJge Oisdwge . --
S~CoIumn

Rim Ff ed CIa,~ier

(b) , i \ ~w Wale'
SIi · 9 OisIribulOr AawWalef
Olstributors
5Concentralor
'_ Oisllibulion
Channel
Figure 3.2.5-9 Vari oll s types of clarifier: (a) detritu s tank wi th grit washer (Dorr- Oli ver); Batf1es
"-"
(b) conce pt of ce nter feed and peri pheral feed clarifiers: (e) reactor-cla rifier with hig h reci r- (e)
culation (l nfilco): (d) slu dge blanket reactor-clarifier (Grave r Wate r Company): (e) reactor-
Figure 3.2.5-9 (Coll/illl/ed)
clarifier with hydrauli c tl occ ulatio n and sludge removal (Degrernont) .
165
164
 BASIC APPROACH 167
166 DESIGN OF BASIC TREATMENT PROCESS UNITS

sand in the sludge zone is 10 to 120/,o. F Igure 3.2.5-10 .Illustrates the principle of thl's Selection Criteria Selection of the sedimentation tanks should be based on the
process. following criteria:
The advantagesof the high-speed microsand settling process include: (I) rcquires
abveryt small
15 ' footpnnt;
. ' (2) has good performance'' (3)
. has a very qUI' c k proccss startup I. The type of suspended matter that is to be removed: for example, grit, silt
ad ou mIn, and (4) enJoys a potential savings in capital costs. However thc d' ~ (plain sedimentation), chemically tlocculated suspended matter, or biologi-
a v~ntag~s are: (I) heavy dependence on mechanical equipment and a sl~ort ;~ cal floc
cessmg time; (2) the entire process must be shut down when ther . P - 2. The overall process train and the role of sedimentation: for example, pre-
age lasting m th 10 . e IS a power out-
o ore an mmutes; (3) requires a higher coagulant dosa c tha treatment for rapid sand filters, pretreatment for multimedia high-rate filters,
conventIOnal processes with a high proportion of polymers' and (4) . ' g. n clarification of lime-softened floc prior to liltration, or removal of sand and
by patents. ' use IS restricted
silt upstream of the intake pumps
Th,e ~otent,ial applications for this process would be in situations where there is a 3. The topography and ground conditions of the plant site
~;r~ ln~lted site are~ for the clarification process, if the filter wash waste is to be cIa ~ 4. The potential for hydraulic shock loading and the degree of fluctuation in
I Ie pnor to recyclIng, and for the clarification of combined sewer overtlow (CSO;.
the quality of the influent water
Type and Selection Guide In this section we present a brief summar of th 5. The requirements for the quality of the settled water with respect to subse-
?PIC~ sedlm~ntatlOn tanks used in the water treatment process and a fe': gUide~ quent processes
mes lor makmg the proper selection. 6. The nature and amount of sludge that will be produced
7. The local climatic and geological conditions
Available Alternatives 8. The utilization of site dimensions and considerations with respect to future
plant expansion
Horizontal flow Long rectangular tanks 9. The capital required for the sedimentation tanks and the operation and main-
Horizontal flow Center-feed tanks (circular or square)
Horizontal flow tenance costs
Penpheral-feed circular tanks
Uptlow clarifiers 10. The time period that is allowed for the design and building of the treatment
Proprietary units
Reactor clarifiers Units: proprietary, recirculation, sludge blanket plant
High-rate settler modules Propr~etary items, tube settlers, parallel-plate modules
High-speed microsand Propnetary patented itcms, very high hydraulic
loading settling process, and a short startup Order of Preference In selecting the sedimentation process, the order of prefer-
ence is established by both the performance and the operation and maintenance costs
of each type. Based on past experience, the order of preference among the various
types of plain sedimentation tank is (l) a long rectangular tank, (2) a rectangular
Ipotymlr ). _____ ..... _n_. n tank containing high-rate settler modules, and (3) proprietary units. Similarly, the
order of preference among the clarifiers is (1) a long rectangular tank, (2) a rectan-
gular tank containing high-rate settler modules, and (3) a reactor clarifier or high-
speed microsand clarifier.
A rectangular tank with a horizontal tlow is generally favored because of its hy-
draulic stability (if properly designed) and its tolerance of shock loadings. This type
of tank also has a predictable performance and is capable of withstanding a tlow
,, rate that is twice the designed rate without significant deterioration in the quality of
the settled water. Additionally, this type of tank is easily operated, has low mainte-
nance, adapts easily to the installation of high-rate settler modules, and uses space
iii
Tube SltUar
efficiently, especially with the common-wall design.
with Scr-.per ~",,====d!' The use of reactor clarifiers should be considered and evaluated if (1) the plant
s,.,..-q/floc;J<t>I1aIIctI
flow rate and raw water quality are relatively constant; (2) the raw water quality is
Figure 3.2.5-10 Concept of high-speed micros and settling process. ""'" relatively constant, particularly with no sudden changes in water temperature, and

168 DESIGN OF BASIC TREATMENT PROCESS UNITS
the raw water turbidit y is always be tween 5 and 300 ntu ; (3) lI si ng lime so fte ning
or o thcr processes that benelit from the seedin g effec t; (4) the unit s are installed in
a hous ing; (5) the re are time constraints o n both the des ig n and construction, as well
as budgct constraints; and (6) th e ow ner prefers th eir use.
Three types of c lariliers are not recommended for removing alum floc because
o f the ir hydraulic in stability: hori zontal flow with ce nter feed, peripheral feed , and
th e s imple uptl ow cla rifi ers. The use o f rec tang ul ar tanks w ith trays (i ncludin g the
doubl e·dec k type) is ge nera lly no t recomme nded unl ess the available site area is
very limited s ince they do not meet the expected level of settling e rfi ciency, have
hi gh co nstru cti o n costs, and prese nt difliculties when in stalling mec hanical sludge
removal equipme nt o n eac h tray (unless the clearance is over 6 ft ).
Discussion of Alternatives
Horizontal Flow and Long Rectangular Basins As di sc usscd prcviously, in the
sec tion "Overall Configuration" und cr "Co nsiderati o ns," the most cos t·cffec tive and
s pace·sav in g co nli g urati o l1 s arc rectangular basins arranged lo ngitudinall y and s ide
by side. The important des ig n require ment is a unifo rm di s tributi o n of feed water at
the e ntrance of the basin. E illue nt control by men ns o f a series of lo ng s kimming
troug hs an d a ba ill e wa ll is not as e ffec ti ve as so me textbooks and design g uid es
published by regulatory age ncies s uggest. These s ubjec ts are desc ribed in the sec·
tion " Inl e t and Outlet of the Bas in" und er "Consi d erati o ns."
Two major pi eces of basin equipme nt comlllo nl y used are the sludge collection
syste m and hi g h-rate se ttler modul es (sec the ea rlier di scussio n). Pro per installation
of th e tube settl er modules requires a minimum tank de pth of 12 ft (3.6 m), de-
pending on th e type of slud ge co ll ec ti o n system under the modules. A typical in-
s tallati o n prolilc is illus trated in Figure 3.2.5· 11.
Sludge Collection Systems The man ufac ture rs o f rec tan gul ar and ci rcul ar basi ns
offer several diffe re nt types of co ntinu ous mec hani ca l s ludge co ll ectors. The main
types arc (I) chain -a nd -Ili ght , (2) trave lin g bridge with sludge scra ping squeegees
and a cross co ll ec tor at o ne end o f the tank , (3) trave ling bridge wi th s ludge suc-
Solid baHlewall
(entire front end) f----+~~~---...,/_.1
Flow_
---L. -'---
Sludge collector ~ --=- ...L_
Figure 3.2.5-1 I Req uired tank depth ror lube se ttler in stallation.
BASIC APPROACH 169
ti on headers and pump or s ipho n mec hani s m pulled by wires, (4) lloat·supportcd
sucti o n heade rs, (5) underwa te r bog ies with sq ueegee (pulled by wi res), and (6) cen·
ter.pivotcd rake.
Selecti on o f the proper slud gc collecti on system depend s on the local wea the r
conditi o ns , the nature and quantit y of sll spe nded so lids in the raw water, th e type
and amount o f coag ulant used, the shape o f the tank , the install ation (or lac k) of
hi glH·ate settl ers, aest he ti cs, and cos t. In regio ns of co ld weather the fo rm ati o n of
icc o n the surface o f the water prohibits the use o f movin g parts. T here fore, the trav-
eling bridge and Il oat·supported sludge headers shou ld 110t be ll sed in these regi ~ n s
unl ess the tank s are ei ther cove red or housed in a building. There are eve n a lew
case hi stori es where the travelin g brid ge has been derailed by ga le forc e wind s.
O ther important facto rs th at s ho uld be co ns idered are th e natu re of the s us pe nd ed
solid s and the quantity of solids ex pected to settle in the tank . For exa mpl e, ( I) wa·
ter co ntainin g large amounts o f freshwater clams will have shc ll s settle in the grit
chamber and settlin g tank s; (2) Sllc ti o n s lud ge coll ec tor systems require the orili ces
of their head ers to be c leaned periodically beca use the s ize o f th e milices is usually
less than 1 in . (20 mm ), and there fo re the orilices becomc frequ en tl y clogged; (3)
lime so fte ning o f hard wate r te nd s to produce large amounts of heavy sludge, whic h
may become troubl esome for sucti o n units; and (4) the chain-a nd -fli ght unit s re·
quire frequent maintenance whenever large amounts of abrasive grit are handl ed on
a co ntinuou s bas is (frequent maintenance may also be indic ative of improper in ·
staJlati on and/or use of inferior parts).
The shape o f the tank al so limits the type o f slud ge collection unit that may be
installed. A unit with a cen ter-pi voted rotating rake is best suited for circul ar or
square basins. S ince mos t sed im en tati o n tanks are lon g and rec tang ular in shape,
traveling brid ge or c hain·and · lli ght unit s are co mmonly cmpl oyed . The choice be-
twee n the two types is determ in ed by cos t, aest hetics, and the presence o f settler
modules in the ta nk . Most c hain -and-llig ht unit s ca n se rvi ce a tank le ngth o f ap·
proximately 200 ft (60 m ) if it is run a ll a co ntinu o us bas is to re move a heavy ac -
cumulation of slud ge. Howeve r, o ne manufacture r claims that it can provide a unit
that can cover 300 ft (90 111), but it mu st be operated in an inte rmitt ent 1~1s hi o n , sti ch
as in the re moval o f a li ght acc umulati o n o f s ludge. If th e basin le ngth is lo nge r
than these limits, two sets of co ll ec to rs are required. Basins cont aining hi g h-ra te se t·
tiers wi ll have space limitati o ns and the refore req uire e ither the c ha iIH.lI1d· lli g ht or
underwater bogies to be insta ll ed.
Unlike chain -a nd· lli g ht modul es, trave lin g bridge coll ec tors can serv ice any
length o f tank. Howeve r, it is morc cost-effec ti ve if th e bas in length exceeds 260 to
300 ft (80 to 90 Ill). Yet res idents livi ng near these types o f plant may complain that
traveling brid ge collectors arc un s ig htl y. The bridge o fte n tra vels o n rail s with a cog
drive mec hani sm but can be o n solid rubber tire wheels without rai ls .
A sucti o n sys tem with a travelin g bridge is ge nerally capa bl e of withdrawing
more co ncentrated slud ge than the sc raper. This type of unit also brings the slud ge
to a channel th a( is above th e wa ter level, thereby allowing the plant operator to vi·
sually inspec t the co ns iste ncy o f the s lud gc and smell a ny ex traordinary odo r.
Telescoping va lves ac hieve th e same e ffec t when used in co njunction w ith a s lud gc
170 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 171

scraper unit (Figure 3.2.5-12). If the sludge is sucked up by means o f pump or slope may range from I : 300 to I : 600. When a slud ge sucti on mechanism is adopted,
siphon, the discharge rate of the sludge sho uld be designed to be approx imately 10 the entire bottom of the tank may be level. Operators of plants using thi s type of
gpm per linear foot ( 120 Lim . min) of the tank width. system seldom express adverse comments regarding th~ design of the basin . The
In summary, a traveling bridge with suction headers is the best choice for large underllow rate (assoc iated wit h slud ge removal) of hOri zontal Ilow and long rec-
tanks that do not have high-rate se ttlers but have normal levels of alum or ferric tangular basins is usually 0.1 to 0.2% of the plant now rate . The concentration of
sludge deposits. In region s with moderate to warm weather, plant sites fur from res~ solids in the slud ge is 0.5 to 2%, depending on the type of mechanical unit e m-
idcntial areas arc prime candidates for a traveling bridge sucti on collector. Chain- ployed and operational adj ustme nts.
and-night units may be used under any conditions. Howeve r. their application to
very long and wide tank s is not cos t-effective because the maximum length of one Design Criteria T his secti on present s the basic design criteria for Ihree types of
chain-and- ni gh t unit is usually limited to 200 ft (60 m) and width of th e ni ghts is clarifier: grit chamber, rectangular sedimentati on tank, and sedimentation tank with
limited to 20 ft (6 m). Additionally, water containing large amounts of abrasive grit high-rate settlers.
would necessitate a high maintenance ratc. If cost contro l is a major issue, float-
supported sliction headers and underwater bogie units may be considered as alter- Grit Chamber (Removal of sand and silt )
nati ves.
The traveling speed of a continuous sludge removal unit is usually I ft/min (O.j Type Rec tangular tank with hori zontal n ow
m/min ). This speed ensures that sludge resuspensi oll does not occur. Units that can Minimum size of grit to be removed 0.1 mm
lift the sludge collect ion header onto a return path may travel at speeds that are 5 Minimum number of tanks Two (One tank with a bypass channel may
to 10 times higher than on the sludge collecti on path. For developing coun tries or be used if the anticipated amount of grit
small plants in isolated areas, the bottom of the sedimentation lank should have a is small.)
slope of at least I : joo toward the tank inlet to ensure easy manual sludge removal Water depth 10 to 16 ft (3 to 5 m)
through the use of wat er jets. If mechanical sludge scraper equipmen t is used , the Mean Ilow ve loc ity IOta 15 ft/min (3 to 4.5 m/min)
Detention tim e 6 to 15 min
Surface loadi ng 4 to 10 gpm/ft2 ( 10 to 25 m/h)
RACK AND PINION RACK AND PINION
SLUDGE WITHDRAWAL VALVE OFFSET STANO ARRANGEMENT Length/width rati o (UW) 4: I to 8: I
Minimum of I : 8
r~t:
Water depth/length rat io

~D
Rectangular Sedimentation Tank (Re moval of tlocculated suspended matter)
FLOOR STAND
DRAW·OFF VALVE
Minimum number of tanks Two
'( Water depth
Mean now velocity
10 to 15 ft (3 to 4.5 m)
I to 3.5 ft/min (0.3 to 1. 1 m/min)
Detention time 1.5 to 4 h
SLUDGE WELL Surface loading 0 .5 to 1.0 gpm/ft2 ( 1.25 to 2.5 m/h)
Length/wid th ratio (U W) Minimum of 4: I
SLUDGE PIPE Water depth/length rati o Minimum of I : 15
Launder weir load ing 12 to 18 gpmlft2 (9 to 13 m' lm . h) (could possibly
be 20 times hi gher)
SLUDGE HOPPER Sludge collector speed
For the collec ti on path I to 3 ftlmin (0 .3 to 0.9 In/min)
For the return path 5 to 10 ftlmin ( 1.5 to 3 m/min)
SIudge draw-oft yalyas consist 01 a III', G" or S' ~Iamerer seamless brass,
'lainless steel or pvC tube whiCh IS ral$8d and lowered by a rack and
pinIOn assembly. As Itte tube is lowered below the water \eyer the sludoe Sedimentation Tank with High-Rate Seltier (Tube se ttl er)
is discharged by hydrostatic ptessure.

Minimum number of tank s Two

Figure 3.2.5-12 Telescopic values for sludge withdrawa l. Water depth of the tank 12 to 15 ft (3.6 to 4.5 m)
172 DESIGN OF BASIC TREATM ENT PROCESS UNITS BAStC APPROACH 173

Surrace loading of the tank area I.S to 3 gpmlft' (3.8 to 7.S m/h) Example 2 Grit Chamber Design
covered by the settler
Given
Mean velocity in the tube settler module Maximum of O.S fllmin (O.IS Ill/min)
Detention time in the settler module Minimum of 4 min Plant now rate 100 mgd (4.4 n,'/s)
Launder weir loading S to 10 gplll1ft 2 (3.7 to 7.S 111'/111 . h) Minimum size of the sand to be removed 0.1 mm
Water temperature 10°C (SO°F)

NOTE: The design criteria or surl~1ce loading, as well as the inlet and outlet Determine
design, are predetermined by the manufacturer. These proprietary products generally
have surface loading of 3 to S gpmlft' (7.S to 12.S Ill/h) for the tank surl:,ce area (i) The number and shape of the tanks
covered by the settler modules. In some cases the space between the plates and the (ii) The size and water depth of each tank
length of the plates may be customized. (iii) The conligllration of the inlet valves and diffuser wall

Example Design Calculations Solution (i) Provide two rectangular tanks for the most cost-effective design.
(ii) The size of each tank is calculated as follows. The setlling velocity of 0.1
Example 1 Estimated Density Flow Velocity 10m of sand is Vo = I.S7 fllmin (8 mm/s) (refer to Table 3.2.S-I). Make the water
depth of each tank 12 fI (3.7 01) and the width of each tank 40 ft (12 10). The mean
Given flow velocity is

Water depth of the sedimentation lank 13.1 ft(4m) Q = (100 X I.SS) -;- 2 = 46S0 fI' /s
Water temperature 20°C (68°F)
Water temperature of the inlluent 19°C (66°F) v = Q = 46S0 -;- (12 X 40) = 9.7 ft/min
a
Determine The estimated velocity of the density flow in the sedimentation tank
L = K~ where K = I.S
vov
SO/Ill ionUsc the density flow formula recommended by the Japan Society of Civil
Engineering (1963): L = (I.S)(~) X 9.7 = 111.2 ft
I.S7

V = 2g ~ -
t1/p'
- )0.5 Say 110 ft (33.5 m) for the tank length. The detention time is
( p k
VlQ = (12 X 40 X 110) -;- 46S0
Assume that
= ft' -;- (ft'/min) = 11.3 min
" = O.S
and the surface loading is
k =2
tJ.,. = 2 III
QIA = 34,7S0 -;- (40 X 110)

= gpm -;- ft'

tJ.p = 0.99843 - 0.99823 = 0.0002 g/cm'
= 7.9 gpm/ft' < 10 gpm/ft2 OK
Therefore,
Check calculations via the Reynolds (Re) and Froude (Fr) numbers. For the
v = (2 X 9.81 X
0.0002
X
2 OS
2
X )0.5= 0.044 m/s Reynolds number:
0.99823
vR
Re = -;;-
= 2.66 m/min
 174 DESIGN OF BASIC TREATMENT PROCESS UNITS

where
~{
,
---
T .·
BASIC APPROACH 175

....

i
~
R=-"-12X
P 40 +
v = 1.41 X 10- 5 ft'ls
40
(2 X 12)

v = 9.7 -;- 60 = 0.162 ft/s

=7.5ft
tt-l
1'1'\

~J -
6--11-<-1·-~1I0 -----1"\
+ ....

Plan
Therefore,

0.162 X 7.5
Re = 1.41 X 10 5 = 86,200 OK for grit chamber ~Diffuser wall

For the Froude number:

~~. =! =

Fr = -
v' elevation
gR
Figure 3.2.5-13 Grit chamber.
0.162'
= ---'-'-==-- . T k Without a Tube Settler
32.2 X 7.5 Example 3 Rectangular SedimentatIOn an
1.09 X 10- 4 > 10- 5 Good
Given
(iii) For the inlet valves and diffuser walls, use two 36-in. rectangular butterfly 2 m 3/s (46 mgd) maximum daily rate
Plant !low rate
valves at each tank inlet to facilitate easy removal of grit in the inlet channel. Avoid Suspended solids to be removed Alum !lac
Llsing sluice gates because they tcnd to "freeze up" with infrequent operation. The 30 nun/min at I Doe (50"F)
Measured !loc settling velo~lty
velocity through the valve is
Expected settled water turbIdIty
2 ntu or less for rapid sand fillers
Minimum ambient temperatu,re -3°e (27"F)
3

v = :1 = (50 X 1.55) -;- (2 X 3') = 4.3 ftls

Potential future plant expanSiOn
Location of plant site
Up to 4 m /s
Far from residential areas

headloss through the valve = - I ( 38.75 )' Determine

2g 0.7 X 9
(i) The number of tanks
= 0.59 ft (0.18 m)
(ii) The size of each tank . .. II
(iii) . f h t
The configuratIon 0 t e an k inlet and dIll user wa
Provide a straight-line diffuser wall that is parallel to the end wall of the inlet with
a space of 6 ft (1.8 m). A total of 420 oritices [6 in. (0.15 m) in. diameter] are re- (iv) The type of sludge removal system
quired for the diffuser wall to allow a tlow velocity of 2 ftls (0.6 mls) through each (v) The configuration of the effluent structure
orilice at a tlow rate of 100 mgd. The oritices should be spaced 15 in. (0.4 m) apart
(center to center) in vertical rows and 21 in. (53 cm) (center to center) for the hor-
izontal direction. The ratio of the orifice area to the tank cross section is approxi- . . e tanks may be selected. Although three tanks pro-
Solution (i) Either two or thle b i'i'-line due to repairs), two tanks
mately 10%. The head loss through the orifice is approximately 1.5 in. (3.8 cm) at . 1 'bT (one tank may e o . '11.'d
vide more optIOnal f ext t Ity (1) future plant expanSIOn wi PIOVI e
100 mgd. Figure 3.2.5-13 illustrates the contiguration of the grit chamber. should be chosen for this example because
 BASIC APPROACH 177
176 DES IGN OF BASIC TR EATM ENT PROC ESS UNITS

the additi onal tanks, (2) the dail y ma ximum fl ow rate is 2 m3/s and it is safe to as- Tank le ngth/water depth ratio = 105 m to 4 m
sume that the daily winter fl ow rate will be approx imately 1.0 to 0.7 111 3/s (more = 26.3: 1 > 15:1 OK
suited to two tanks), and (3) the t wo ~ t a nk design is morc cost-e ffec tive.
(ii) Consideri ng that the standard maximum wi dth of the chain-a nd-fli ght sludge
collector is 6 m (20 ft ), a n aliquot o f 6 m should be evaluated for the width of the Mean flow velocity = Q = (2 X 60) -;- 2 X ( 18 X 4)
A
ta nk : 12, 18, a nd 24 m (40, 60, and 80 ft, res pec ti vely).
= (m3/min) -;- m2
For a trave ling bridge sludge collec tor, a tank width in the ran ge 12 to 30 m (40
to 100 ft) is considered to be cost-effective. The tank should be 18m (60 ft) wide = 0 .83 m/min < 1.1 mlmin OK
to maintain a good tan k le ngth/w idth ratio. Based on fl oc settlin g veloc ity mea-
surements (bench-scale studies conducted during the wint er season), the hydraulic vR
surface loadi ng rate of th e tank can be set at 1.8 m/h (0 .72 gp mlli') with a floc set- Rey nolds number: Re = --;;-
tlin g velocit y o f 30 mm/ min. Although one mi ght think that th is va lue is too con-
servati ve for the winter season (a fl ow th at is less than I mJ/s). the pl ant may lise
only one tank to treat th e plant fl ow, thereby allowing the lank loading rate to re- whcrc
main norm al. Conseque ntl y, 1.9 m/h (0.75 gpmlli' ) is used as the design criteri on v = 0 .83 m/ min = 1.39 X 10- 2 m/s
of the tank. This number includes a comfortable safety margin for all seasons.
The total tank surface area requi red is R = 4 X 18
_2...:.:"":':'- = 2.77 m
(2 X 4) + 18
(2 X 3600) -;- 1.9 = (m3/h) -;- (m/ h) = 3790 m'

Since eac h tank has a width of 18 m (60 ft ), the length required would be Therefore,
U 9 X 10- 2 X 2.77
3790 -;- (2 X 18) = (m' ) -;- (m) = 105 m (343 ft ) Re = 1.31 X 10 6

The water depth of eac h tank should be 4 m ( 13 ft) to faci litate the possible instal- = 29,620 No
lati on of tube settler modules in the future.
Chec k the va lues: Re is too hi gh and the flow in the tank is hi ghl y turbul ent. The preferred Rey nolds
number is less than 18,000.
(1 8 X 105 X 4) X 2
Detention timc at Qrnax
2 X 60 v2
Froude numbcr: Fr = -
gR

where
126 min (2 h 6 min)
v = 1.39 X 10 - ' m/s
Dctcntion tim c at Qave 126 min 1.5*
/I = 2.77 m
189 min (3 h 9 mi n)
g = 9.8 1 kg/s'

* Thc maximum daily flow ratc is usually 1.5 timcs thc avcragc
daily !low rate. Therefore,
(U9 X 10- 2)'
Fr = 9 .8 1 X 2.77
Ta nk le ngth/w idth rati o = 105 m to 18 m
= 5.8: I > 4:1 OK
= 7.1 X 10- 6 < 10- 5 No
 BASIC APPROACH 179
178 DESIGN OF BASIC TREATMENT PROCESS UNITS

The number is too small and the !low is therefore unstable. (iii) When the tank is fed by a common inlet channel, two basic types of sct-
How much will the !low charactcristics in the tank improve if longitudinal baf- tling tank inlet may be used: !low splitting weir and submerged ports with valves.
tles (partition walls) are installed in each tank? If the tlocculation tank is the head end of the sedimentation tank, a diffuser wall
may be used to separate the two tanks. Where the tanks are fed directly from a feed
Trial 1: One Longitudinal Wall in the Middle channel, the design of the channel should be large enough to avoid producing sig-
nificant headloss from one end of the channel to the other. This condition is partic-
R = --.!(;::4-,-X,-9,,-,)-,-X.:...::.2_ m' ularly important for the !low splitting weir.
=- = 2.12 m The weir inlet has problems not only with leakage every time a tank is isolated
(9 + 2 X 4) X 2 m
with stop logs but also with sludge accumulation in the feed channel. From a prac-
Since v is practically unchanged, tical aspect, the best decision is to place submerged ports at the bottom elevation of
the inlet channel by means of butter!ly valves. The inlet ports should provide at least
1.39 X 10-' X 2.12 4 in. (0.10 m) of controlled headloss to ensure even !low distribution to each tank.
Re =
1.31 X 10- 6 = 22,670 > 20,000
The design of the inlet channel is delineated in "Choice of Flocculation Tank Inlet",
Section 3.204.
This Reynolds number is better than before but still too high. Provide three submerged holes with butterfly valves as the tank inlet. Each hole
should be 0.6 m (24 in.) in diameter so that the tlow velocity through each orifice is
Fr = (1.39 X 10-')' = 9.2 X 10- 6 < 10-5
9.81 X 2.12 v = JL = ...,.----.,.:-::-:2":---::-~, = 0.708 mls (2.3 ftls)
La 6 X (0.785 X 0.6 )
The Froude number is still a little too low.

Trial 2: Two Longitudinal Walls 6 III Apart (on Celller) Forming Three Channels
__ I
- 2g
(.!L)'
ca -
__
I (
19.62
2"'" 6
0.7 X 0.283
)'

R 1.71 m v is practically unchanged = 0.144 m (5.7 in.) OK

Re 18,820 < 20,000 Tolerable
The diffuser wall should provide approximately 0.23 mls (0.75 ftls) of flow veloc-
Fr 1.2 X 10- 5 > 10- 5 OK
ity through each port at the maximum daily tlow rate.
Trial 3: Six Longitudinal Walls 3 III Apart (on Cellter) Forming Seven Chanllels . Q 2 m 3/s ,
ReqUired total port area = - = 023 I = 8.7 m
v .. m s
R = 1.1 m Flow velocity (v) is a little faster but
practically unchanged The total number of 0.12-m (diameter) ports that are required is
Re = 11,850 < 20,000 Good
Fr = 1.8 X 10- 5 Good 8.7 m'
----'''-'--~---,-:-
= 770
0.785 X (0.12 m)2
As demonstrated, a tank with a long and narrow channel is hydraulically supe-
rior. It is therefore understandable why a tube settler or a parallel-plate settler per- Provide 10 rows of ports in the vertical direction with a spacing of 004 m and 39
forms well. Thus, from a practical perspective with cost as a factor, the choice would rows in the horizontal direction spaced 0045 m apart (13 rows in each 6-m segment).
be to use two longitudinal baftles to create three channels that are 20 ft (6 m) in In doing this, the actual total number or ports is 390 for each tank, or a total of 780
width. The revised tank lengthlwidth ratio (L/W) is lOS m to 6 m or 17.5: 1. (total area of 8.82 m'). See Figure 3.2.5-14 for the design.
In summary, the features of each tank are as follows: Check the values:

Average water depth 4m(13ft) Flow velocity in each port at Qm" = 2 m3/s "'" 8.82 m' = 0.227 mls
Tank width 18 m (60 ft) with two longitudinal bames < 0.25 mls OK
Tank length lOS m (343 ft)
Detention time
Surface loading
2 h 6 min
1.9 mlh (0.75 gpm/fl' )
Headloss through diffuser wall = ;g (0.7; 8.82)'
= 0.0053 m (0.2 in.) = 0.0055 m OK
Mean !low velocity 0.83 mlmin (2.7 f1/min)
 BASIC APPROACH 181
180 DESIGN OF BASIC TREATMENT PROCESS UNITS

$IudS. channel 1200.45m _ 5.4"' Example 4 Rectangular Tank with Tube Sclllers
I' I r

::::::::::::: .......••....
(TYj)iul)
f. • fll? ....
.............
..............
· Given
i~::lI ··...........
........... .
. ... .
......•.•.... ::::::::::::: The same c riteria appl y as in Exampl e 3, with the additi o n of the fo ll ow ing :

•••.......•.• .....•••.......
';e· :::::::::::::
_~ E .....••..•.••
:::::::::::::
........•••. ·... ::::
•••• After the sedim ent ati on process
........... .. ....
Dual-media lilters
·2..,.
co'············
~
........•....
•••••••••••••
v
Or ............ . ....•.••••... ............
...•••...•...
••••••••••••• ••••••••••••• ••••
• ••• Minimum wa te r te mperature 5°C
6
~
01
.-
6" ' .- IBm(1.0.)
Minimum ambi ent temperature
Locati o n of plant si te
- 10°C ( 14°F)
Nea r residenti al areas; site area is small

Determin e
Figure 3.2.5-14 Inl et diffuse r wa ll (e levati on).
(i) The numbe r o f tanks
(ii ) The lank area that will be covered by the settler modul es
(iv) The trave li ng bridge with pump s ucti on is the 1110s t preferred ty pe of sludge
(iii ) T he size of each tank
removal syste m because ( 1) the tank leng th is o ver 100 In and there fore the travel-
(iv) The confi gurati o n of the tank inlet and diffu se r wall
in g bridge is the mos t cos t-effective; (2) o ne lInit can span two 18- 1ll (width) tanks,
and th is is very cos t-effec ti ve; (3) heavy ice for mati o n (w hic h pro hi bi ts move ment (v) Th e co nli g urat ion of the launde rs and slud ge coll ec ti o n sys tcm
of the bridges) is no t a n isslle since the lowest anti c ipated temperature is o nl y - 3°C;
(4) the tank bo tto m docs not have 10 be sloped w ith thi s system, thu s simplifyin g SOlltlioll (i) The num ber of ta nks is the sa me as in Exampl e 3- two tanks.
the des ig n and minimi zing co nstructi o n COS I; and (5) th e aesth eti cs of the bridge (ii ) The tank area th at is to be cove red by the tube scul er modul e may be cal-
(the industrial equipme nt look) is not an important iss ue becau se the plant is far cul ated. Th e fo ll owin g relati o n ca n be establi shcd fro m Fig ure 3.2.5- 15:
fro m reside ntial areas.
(v) Some regul ato ry agenc ies require an ex tensive launder syste m to be included
in the plant des ig n as part of the conli g uratio n of the e fllu ent stru cture. An accept-
abl e hyd rauli c we ir load ing rale is approx imately 10.8 1ll 3/ m . h (15 gpm/ f!) for th e
launders. Th e required we ir leng th (based o n the criteri on) is Sin ce tube sc ttl er modules have a = 60°,

(2 X 3600) -;- 10.8 = 666 m fo r two lanks So = Q IV

A 0.5/1 + 0.25 IV
A few e ng ineers also beli eve th at the install ati on of a baflle wall , one simil ar to
th e inlluent diffu se r wall at the e fllu cllt end of the tank , minimi zes the density fl ow.
However, ex perience in d ica tes that ne ither the lo ng launders no r the ba ffl e wa ll (lo-
cated at the e fflu ent e nd ) sig nilica ntl y improves the se ttling e fli c ie ncy of lo ng rec-
"'-w
c...
II', ,-,
tangul ar se ttlin g tank s. Co nseque ntl y, th e end wa lls of the tanks Ill.ay be built as an -,---- - -_ .,' I ' . ...,.
I B
ove rll ow of the e fllu ent wa ll.
I
It is also necessary to provide three 0.2-111 (di ameter) holes equipped with but-
ter ll y va lves in th e end wall near the bott om elevati on of the tank . T hese valves are
used in backlillin g the tank from the commo n e fflue nt channe l a fter it is clean ed
and bac kll as hing durin g the clea nin g process. Bac klillin g of the tank is es pec iall y h

L
im portant for fl occ ul ation-sedime ntati o n combinati on tanks th at have vertical- shaft
ll occ ul ators; the ll occ ulat ors will not fun cti on until the water leve l in the tank reaches
half the depth of the tank . If turbid wa ter is int rodu ced from the inlet side during
back fillin g of the tank , the unfl occ ul ated water wi ll now to the end of the ta nk un -
til the wa ter leve l reac hes half th e depth o f the ta nk . The untreated water will there-
fore Il ow di rec tl y il1l o th e lillers afl er Ihe tank is IiIl ed. Figure 3.2.5·15 Elemen t of tube settl er.
 BAS IC APPROACH 183
182 DESIGN OF BASIC TREATMENT PROCESS UNITS

where ~ is the s urface area of the settling tank that is covered by the settler mod. ulcs and the remaining 25% left as open space, 10 settle heavy fl oc, to improve the
ul e (m ) and Q is the now rate of the tank (m'/s). flow inlet co nditi on to the settler, and provide easy access to the mechanical sludge
. A standard tube seuler module has the foll ow ing dimensio ns: II ~ 0.55 m ( I n romo vat system. The total tank le ngth should therefore be
9 m.) and W ~ 0.05 m (2 in.). Therefore.
L ~ 38 m + 0.75 ~ 50.7 m (166 fl )
So ~ Q 0.05 ~ Q
A (0.5 X 0.55 ) + (0.25 X 0.05) A (0174)
. Say 50 m ( 164 fl) . The water depth in the tank sho uld be 4 m (13 fl) . as illustrated
in Figure 3.2. 5- 11 . With respect to the longi tudinal baflle walls. the. now charac-
The average fl oc se ttling velocity, as measured in the laboratory, is 0.0005 mls (re- teristics of the tank are no longer critical because the flow charactenstlcs III the tube
fer to Table 3.2.5-1). However, actual tank co nditi o ns arc not as well co ntrolled and settler are significant. Thus, only one lo ngitudinal baffle wall is required at the cen-
therefore a value of So = 0.00025 m/s is used, providing an approximate safety' fac- terline of the tank.
tor of 2.0: In summary, the features of each tank are as foll ows:

18 m (59 fl)
So ~ 0.00025 ~ 0.174 X I Tank w idth
A Tank len gth 50 m ( 164 fl)
Water depth 4 m (13 fl)
A ~ 696 m' or 7490 n' for each tank One lo ngitudinal baflle wall (UW) 5.5: I

S ince the width of the tank is ass umed to be 18 m. the length of the tank that is to
Tube seuler modules are installed over 75% of the tank area (38 m from the end wall ).
be covered by the tube seUl er is 696 m' + 18 m ~ 38.7 m. Make it 38 m. The ac-
(iv) The tank inlet and diffu ser wall are the same as in Example 3.
tua l tank area for the tube seuler is therefo re 18 m X 38 m ~ 684 m' per tank .
(v) Since an extensive skimming launder syste m is required with the tube set-
Chec k the values. The surface loadin g for the tank area that is covered by the
tler mod ul es, prov ide six laundcr troughs for each basin. Each launder mu st cover
scttl er mod ul e is
the entire length o f the scHier module; thus. the len gth of each launder should be
38 m. The now rate in each launder trough is I + 6 ~ 0.167 m /s . For a trou gh
3
Q I X (60)'
684 ~ 5.3 m/h (2. 1 gpm/ft' ) OK with a rec tang ul ar sec tion and a level invert,
A

Flow ve loc ity in the settler module IS Q ~ 1.38 bhb '

,
'Vo ~ -,---'Q,--- I X 60 where b is the width (m) and "0
is the water depth at the far upstream of the trou gh.
684 X 0.866 ~ 0. 1 mlmin (0.00 167 m/s) Selec t 0.5 m as the width for the launder trough:
A sin 0:

lib ' ~ 0.167 ~ 0.242

The hydra uli c radiu s of the tube seul er modul e is 1.38 X 0.5

R ~ !!.. ~ (0.05)' 110 ~ 0.39 m

p 4 X 0.05 ~ 0.0125 m
Make the interior height of the trough 0.5 m (0. 11 m freeboard). Each trough should
(1.67 X 10- 3 ) X ( 1.25 X 10- ' )
Re ~ 14 < 50 Ve ry good be supported by pedestal columns spaced 3.8 m (12.5 fl) o n center.
1.519 X 10 6
The chain-and- ni ght sludge collection system must be e mployed in thi s des ig n
because of th e use of tube settler modules, the formati o n of thick iee durin g severe
(1.67 X 10- 3 ) '
Fr ~ "9-.8-'-1'"'X"-'--
( 1.:....2:...5=X-ILO~
') ~ 2.3 X JO - S > I0 - s Good winter co nditi o ns, and for aesthetic reasons. Since onc tank is partitioned into two
channels (9 III wide), two sets of chain-and-Ilight unit s are installed in each chan-
~ incc. a standard-size modul e has tubes that arc 0.61 III (2 ft) long, the detention
nel. The end wall of each tank should have two 0.25-m (d iameter) holes with but-
tllne IS 0,6 1 In -:- 0.1 m/min or 6.1 min. terny valves for backfilling of the basin . Ideall y. the e ntire basin should be hou sed
(ii,i) Two identical ~anks that are 18 III wide will be used as discussed in Example to protect the system from ex tremel y cold temperatures. An illustration of the sed-
3. It IS co mlll on prac ll ce to have 75% of the tank area covered by the settl er mod- imentati on tank is provided in Figure 3.2.5- 16.
 184 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 185

~
~------50 m (I .D.) - -- -----:-11 For many plants it is common for the water in the middle of the tank to be clear,

~ m -~'t_--(-C-O-Ve-r-ed-;~h~U-b-e-'-et-lI-e-r)----l1
yet have a lot of fl oc ri se at the end of the tank. This type of pattern indicates the
2 '" 250 mm presence of a distinct density fl ow. Plant operators have only three possible c hoices
,-- B.V.
to counteract this situation. The lirst is to block off the las t 10 to 20 ft (3 to 6 m)

Tr~r;''''~fl~~~~~~~~~~~~~~~ii
~
~:
:
~-'-
~---~~~-~~===l=au=n=d=e=r=tr=ou=g=h=(=tY=P=i~='=)~~~~~~'
(Backfill valve) of V-notches or submerged ori lices, on the laund ers. This will work only if laun-
ders are used to collect both settled and clear waler from the midsec ti on of the tank.

~~ i~_j~~-~
~--i~~~~~~~~~~~~~~~~
111is simple modification often improves the situation significantly. The second
E ~ ... V channel
Effluent choice is to feed 0.2 to 0.5 mglL of anionic polymer to the second stage of the 110c-
: ~
- - " Longitudinal batfle wall eulation process. The polymer acts as a fl occulation aid th at helps to prod uce read·
~:~ "'_,-
@-rt=====================~~ H
1 iJy settleable fl oc. The last choice pertains to plants that have more than one source
of water: Minimi ze switching water sources or minimize changes in the blending
ratio of the two sources. By doing so, the fl ow shorl·circuiting through th e ba si n
(t he product of a density c urrent) ca n be reduced .
Influent channel It is very important that plant operators detect and evaluate any abnormal phe·
(a) nomena occurri ng within the tank. This includes the sudden appearance of fl oati ng
sludge (sludge bulking), scu m, midge fly larvae, and algae growt h on the tank wa lls
Chain-and-flight and barnes. Sludge bulking occurs when gas accumulates in the settl ed sludge as a
system
V-notch weir plate Tube settler module result of anaerobic decomposition by bacteria or supersaturati on of the raw water
(by air) so that it fl oats rather th an sinks. S ludge bulking is usually con trolled by
prechlorination: 2 to 3 mg/L of c hl orine. Midge fly larvae bloom occ urs during ce r-
-- tain seasons if the sludge is not removed frequen tl y. Algae will grow on the tank
walls, launders, and barnes during warm months, especially in nutrient·rich water.
Both the algae and larvae proble ms can be co nt rolled effectively by prechlorination
Sludge pipe
(I to 2 mglL). However, recent limitation on the use of THM restricts the use of
(6) prechlorination for many plant s. A feasible alternative is to provide occasional shock
Figure 3.2.5-16 Sed imentation lank wit h tube settl er mod ul es: (0) plan: (b) elevation. treatments with 2 to 3 mg/L of chl ori ne.
The accumulati on of heavy scum can be attributed to many causes; changes in
raw water quality, changes in coagul ant (from alum to ferric c hl oride), and super-
Operation and Maintenance The operaliona l slafr of water trCai ment plants saturation of raw waleI' by ai r or gas. Supersaturation is the most difficult problem
should be aware of the following items: to solve. The only practical solution is to prevent the entrainment of air in the raw
water.
I. Floc settling co ndit io ns in the tank Another phenomenon that should be noted on a regular basis is corrosion.
2. Abnormal phenomena Concrete surfaces and submerged metal parts will severely corrode with repeated
exposure to a combination of aggressive water, chlorine, and alum . In thi s situati on,
3. Eq ual hyd rau lic load ing to each tank
a proper coating must be applied to these surfaces or, in the case of the metal s, ca·
4. Sched ul e for sludge removal thodic protection should be introduced . A severe pitting type of corrosion may oc·
5. Spec ial design fea tu res of the tank cas ion ally be observed in submerged stain less steel parts (type 3(6). This occ urs
only under anoxic conditions and can be minimized by specifying the usc of type
. If the suspended solids are well 110cc ulated, a majority of the 110c should se ttle 304 stainless steel or by making the conditions aerobic.
I~ ~he first halfofthe tank. Consequently, visual observa ti on or the fl oc settling con- The design of the tank is based on eq ual hydraulic loadi ng to all tanks. Yet this
dl~l ~n (:fOm ~he walkway s) is important and the coagulant dosage or Il occulator may not occur due to erroneollSdesign or improper adj ustment of the weir and valve
mlxlIlg IIltensily should be adj usted accordingly ; inadeq uate floccul ati on will allow openings. Plant operators sho uld therefore check the fl ow rate th rough eac h tank by
small fl oc to leave the flocculation tanks. However, turbul ence in the basin increases visual observati on or by measuring the water height over the effluent weirs. An un·
fl oc size in the area from the middle to the end of the tank. As a result, there is a even fl ow rate to each lank can be roughly corrected by adjusting ei ther the fl ow·
heavy carryover of fl oc to the filters. Splitting weirs at the tank inlet or the inlet valve openings.
 186 DESIGN OF BASIC TREATMENT PROCESS UNITS
BASIC APPROACH 187

Certain factors can create undesirable flow patterns. For example. uneven level-
TABLE 3.2.5-4 Guide for Traveling Bridge Operation
ing of each launder trough or effluent end weir wall can cause a skewing flow to
be created. Uneven tank subsidence may tilt the tank and thereby create horizontal A BaU-length ((ront halO sweeping (47 metrea)
(al (Table sho'wa .woeping cycle and corrupondlng .weeping apeed)
flow short-circuiting. Improper design of the inlet channel and the tank inlet will
also cause a great imbalance in the flow rate among several tanks that are in paral-
lel. In certain plants a difference of as much as 50% has been observed, even though SweepIng Speed (m/mtn)
the elevations of the launder troughs at each end of the tank were virtually the same. 0.75 1.0 1, 5 ~. ~5 3.0
Perhaps one of the most important issues for the sedimentation tank is opti-
Total sweeping Time (1)
mization of the sludge withdrawal process. In general, tanks that lack high-rate set- (mlnutOl) 62 H H.3 ~1 15.7
tler modules will have 60 to SO% of the sludge deposited in the front half of the
tank, provided that the flocculation process is functioning well. Sludge sweeping
Volume or Sludge R.emoved (Zl
should therefore be performed more frequently for the front half of the tank; a trav- (1) X 4.B m3/m1n ~97,6 ~~5.6 150, ~ 10 t 71.4
eling bridge can be programmed to sweep the front half two or three times before
running the entire length (in order to sweep the last half). In contrast, other types
of sludge scraping mechanisms, such as a chain-and-flight unit, cannot perform par- 1% aoUd sludge
tial sweeps. (1310 m 3 /d)
Required cycle
The proper sludge sweep cycle and the duration may be determined from sev- or lIZ length.
eral trial runs, by noting the concentration of the solids and the texture and smeU sweeping 2!.!.
~.
of the sludge (for freshness). The most appropriate types of unit for this test are the
8.7
sludge suction sweep units and the telescoping valves (for sludge discharge) with
chain-and-flight collectors. These units allow the plant operator to determine the
7.0
proper sludge sweeping cycle by allowing easy observation and testing of sludge
consistency. Other methods of testing include the installation of magnetic or ultra- ~•.••••.•w ..~ Recommonded Ranll.
·:·u.·...·.·.
sonic sludge density and flow meters in the sludge discharge line or the installation
of a piece of transparent tubing in the sludge line to allow visual observation. (b) A F"1l1ength aweeplng rollowed by lIZ length aweeps
However, the second method (transparent pipe) is simply not practical because the 4 a 3/ in {Total or two brldgOl}

--
pipe becomes coated with sludge in a short period of time. Total pump capacity;: • m m d' d c ¥taln condItlona}
(Table .howl an approprlate .weeptna .pee • un or 0
Table 3.2.5-4 is an example of a plant operator's operational guide for a travel- Sludae Conentratlon ('V•• olida)
ing bridge titted with sludge suction pumps. In areas where the water table is high,
1'. I. S'1. Z~. Z. S.,..
the tank cannot be drained at will unless the tank is designed to resist uplift. For

t---- ----_.
this reason, some plant designs include an underdrainage system for each tank. Each Total volume or Ca.e I. 3 ~IB.
145.5 87.3 109,~

underdrain line has an observation point where the operator can visually inspect the dudge to be removed
(m 3 )
------
Case Z+
-----
327.5
-----
ZIB.3 163.8 131
flow of water to see if the water is running or standing. This type of system allows
the operators to check any tank for leaks and allows them to note unusually high Req'd pumping time Caao 1. 45 30.3 n.8 18.~
underground water tables. to remove the .ludgc
by thi 5 path (min. 1 Cue Z40 68.Z
-- --
--- - -- -- -- -- - -- - Z7.3
45.5 34.1
Finally, the operator should be familiar with certain design features of the tank.
For example, the diffuser wall should be designed to have a restricted water pas- Req'd sweeping Caee 1* Z.1 3. 1 4.1 x 4. aX
sage (facilitated by uniformly distributed ports) to produce uniform distribution. In speed (m/min) }----- ---- - - -- f-- -- --;:--
(or each case (approx ) Cue Z+ 1 •• Z.1 Z.8 3 ••
a few cases, the plant operators had a passageway cut in the diffuser wall [3 by 6
ft (0.9 by I.S m)l to allow easy access for maintenance crews. In another case, the Note: .Ca. .. 1:Five conucutlve l!Z.hngth .weepinia beroro the
bottom I to 2 Ii (0.3 to 0.6 m) of the baffle wall were removed completely in an at- fu1l1enath .weeping.
tempt to make the task of sludge removal and tank drainage more convenient. All +Case Z: Three conaecutlve 1/Z-length sweeping. before the
these modifications caused the diffuser wall to become nonfunctional. (ull length .weeptna
xBeyond the capabtuty of tuveUl.ng bridge
Since the majority of plant operators are not fully aware of the actual flow char-
acteristics of conventional sedimentation basins, Appendix 9 (tracer test) has been
 188 DESIGN OF BASIC TREATMENT PROCESS UNITS
added to this book. As shown in the figures in Appendix 9, the actual flow-through
time (I) IS only about 30% of the mean detention time (T); the performance of most
sedimentation tanks is represented by these graphs. Although plant operators can-
not improve this situation, they should be aware of this fact and should try to im-
prove the settling effiCiency of the tank by improving flocculation.
BIBLIOGRAPHY
ASCE, AWWA, and eSSE, Water Treatment Plant Desigll, American Water Works
Association, Denver. CO, 1969, 1990, and 1997.
Bernardo, L. D., et aI., "Use of Perforated Baftle at the Inlet of Rectangular Settling Basins"
J. AWWA, 72:9:528 (Seplember 1980).
Camp, T. R.. "Sedimemation and Design of Settling Tanks," ASCE Trails. 111:895-958
(1946). '
Camp. T. R., "Studies of Sedimentation Basin Design," Sewage Ind. Wastes, 25:1-12 (1953).
Culp, K. Y.. et a1.. ''Tube Clarification Process: Operating Experiences," ASCE P,ve., p. 1179
(January 1955).
Degremont. Water Treatment Handbook, 6th ed., Lavoisier Publishing, Paris. 1991.
Edzwald. J. K., et al.. Removal of Humic Substances and Algae by Dissolved Air Flotation,
EPA/600/2-89-032. U.S. Environmental Protection Agency, Washington. DC. 1989.
Hazen, A. M., "On Sedimentation," ASCE Trails., 63:45 (1904).
Hong, S. N., "Treatment of Combined Sewer Overflow with High Speed Microsand Settling"
WEFTEC '96, Dalias. TX. 1996. ' the lake/reservoir is nutrient rich; overturning of the stored water during the spring
Hong, S. N., et aI., "ACTIFLO Process for Drinking Water and Storm Water Treatment"
WEAUTECH '96, SI. George, UT, June 1996. ' manganese, and taste and odor problems. The regular conventional flocculation and
Hudson, H. E. Jr., Water Clarification Processes: Practical Design and Evaluation Van
Nostrand Reinhold, New York, 1981, pp. 123-138. ' slowly unless a very low surface loading rate is specified and costly polymers are
Huisman, L., Sedimentation and Floatation, DelFt University of Technology, Delft, The
Netherlands, 1973.
Ingersol, A. c., et aI., "Fundamental Concepts of Rectangular Settling Tanks," ASCE Proc.,
1179 (January 1955).
Kawamura, S., "Hydraulic Scale Model Simulation of the Sedimentation Process," J. AWWA,
73:7:372 (July 1981).
Kawamura, S., et aI., "Evaluation of Launders in Rectangular Sedimentation Tanks" J WPCF.
58: 12: 1124 (December 1986). ,. , higher (2 to 3%) than that of the sludge produced by sedimentation (0.5%). The DAF
LePonder, N:, and Binot, P., ''Treatment of Combined Sewer Overflow (CSO) with High
Speed Mlcrosand Settling," WEF Special Conference, Montreal, Ontario Canada July
1995. ' ,
Monk, R. D. G., et aI., "Designing Water Treatment FaCilities," 1 AWWA, 79:2:45 (February
1987).
Montgomery, 1. M., Consulting Engineers, Water Treatment: Principles and Design, Wiley,
New York, 1985, pp. 135-151,523-532.
Okuno, N., et aI., Analysis of Existing Clarifiers and Design Considerations for Better
Performance, Tokyo Metropolitan Government Report, 1980.
BASIC APPROACH 189
O'Melia, C. R, "Particles, Pretreatment, and Performance in Water Filtration," 1. Environ.
Ellg. Dill ASCE, 111:6 (December 1985).
Reed, G. D., and Mery, P. c., "Influence of Floc Size Distribution on Clarification," 1. AWWA,
78:8:75 (August 1986).
U.S. Patent 5,730,864. "Installation for Treating an Untreated Flow by Simple Sedimentation
After Baliasting with Fine Sand," March 24, 1998.
Walker, J. D., "Sedimentation," in Water Treatmem Plant Design for the Practicing Engineer,
R. L. Sanks, ed., Ann Arbor Science, Ann Arbor, MI, 1979, pp. 149-182.
Weber, W. 1.. Physiochemical Process for Water Quality Control, WileYMinterscience, New
York, 1972, pp. 111-138.
Yao, K. M., "Design of High Rate Settlers," ASCE Sail it. Ellg. Div., 99:621-637 (October
1973).
'
Yee, L. y, et aI., "Inlet Design for Rectangular Settling Tank by Physical Modeling," 1. WPCF,
57:12:1168 (December 1985).
Zabel, T., "The Advantages of Dissolved Air Flotation for Water Treatment:' J. AWWA, 77:5:46
(May 1985).
3.2,6 Dissolved Air Flotation Process
Purpose A significant number of water treatment plants obtain raw water from
lakes and reservoirs. In general, these sources contain low-density suspended solids
such as algae and organic compounds, and the water is often highly colored.
Although turbidity of lake/reservoir water is normally low, the hypolimnion zone of
and fall of each year often results in algae blooms in addition to dissolved iron,
sedimentation process is not very effective because the flocs are light and settle
fed to the process train.
Dissolved air flotation (DAF) is an effective alternative to sedimentation because
it uses minute air bubbles to float light flocs; the floated solids are skimmed off, leav-
ing clear water near the bottom of the tank. The time required for flocculation is shorter
than for the conventional settling process, and the hydraulic surface loading rate of
the flotation tank is generally 10 times or more that for conventional sedimentation
tanks. In addition, the solids concentration of the produced sludge is generally much
process has been widely used to thicken waste/activated sludge in biological waste-
water treatment as well as in the pulp/paper and mining industries in the United States.
Application of the DAF process to drinking water treatment began in the United States
in 1993, however, the process has been used successfully in European countries since
the 1960s. Figure 3.2.6-1 shows a general idea of the DAF system.
Considerations Most of the considerations described previously under regular
sedimentation are also applicable to the DAF process. However, a few specific is-
sues are important to this process. The issues include (I) raw water quality, (2) floc-
 190 DESIGN OF BASIC TREATMENT PROCESS UNITS
FJoC::ulatcr
Flocculation
Tanks
s:ueg. Ei'Uuent
Figure 3.2.6-1 Dissolved air flotation system. (Courtesy of PURAC Engineering, Inc.)
culation process design, (3) flotation tank inlet design, (4) recycling need and the
system design, (5) housing needs, air saturation system selection, (6) floating solid
(float) remov~l, (7) flotation tank effluent design, (8) limitations, and (9) cost. The
major lImItatIOns of OAF are the inability to treat highly turbid water and the ne-
cessity to house the flotation tanks.
Raw Water Quality The OAF process is most suited for treatment of algae-laden,
hIghly colored water which also has relatively low turbidity. In general, turbidity of
up to 50 ntu IS comfortable and the upper limit is as high as 100 ntu with a short
duration. The TOe level of raw water is important because the OAF process does
not operate effectively in the enhanced coagulation mode because many floes be-
come too heavy and will not float due to a high dosage of inorganic coagulant and
polymer. A TOe value of 8 to 10 mglL should be the maximum level for the OAF
process. However, removal of Cyrptosporidiwn oocysts and Giardia cysts can be
attamed, and accordmg to recent studies, a 2 to 4 log removal may be achieved.
Flocculation Process There are a few significant differences in the design of the
floccul~tlOn process for a DAF versus regular sedimentation process. Unlike sedi-
mentatIOn, the OAF process must produce small and light floes which attach to a
large numberof air bubbles that float to the surface. Three important criteria must
be met to achIeve this objective: (I) high-energy mixing, (2) shorter mixing period,
BASIC APPROACH 191
and (3) installation of an inclined baffle at the eflluent of the tank so that the bub-
ble-Iloc agglomerate are directed toward the surface of the flotation tank.
A mixing energy level of 30 to 120 S -I of G, which is double that of a conven-
tional Ilocculation process, should be used in the design of the OAF Ilocculation
process. The mixing time is generally IS to 20 min, and this short mixing time re-
quires a good baffle wall system to compartmentalize each stage of flocculation. In
general, only two-stage Ilocculation is provided. The effluent design is quite dif-
ferent from that of a conventional sedimentation process design. A critical design
parameter is the placement of an inclined baffle at the effluent of the tank which is
60 to 700 to the horizontal.
Flotation Tank The width of the flotation tank is limited by the Iloated floc
(sludge) removal equipment, and the length of the tank is dictated by the distance
that the bubble-floc agglomeration remains afloat. Therefore, in general, the tank
should not be longer than 40 ft (12 m). The maximum depth of the tank is normally
8.5 ft (2.6 m); however, a tank that is too shallow may carry air bubbles into the ef-
fluent due to influent momentum.
The general practice is a detention time of 10 to IS min with a surface loading
of 4 to 6 gpm/ft2 (10 to IS m/h). Recent pilot studies in North America have shown
that a surface loading of 6 to 8 gpm/ft2 (15 to 20 m/h) may be used when polymer
is added in conjunction with regular inorganic coagulant. Therefore, the footprint
of the flotation tanks is about 10% that for regular sedimentation basins.
The inlet to the flotation tank plays a critical role in the efficiency of the flota-
tion process. The inclined baftle wall inlet should introduce Ilocculated water uni-
formly along the width of the tank into the recycle dispersion zone near the bottom
of the tank. The top of this baffle should be approximately I ft (0.3 m) below the
water surface in the tank. Recycle flow is introduced continuously through a distri-
bution system of proprietary nozzles, orifices, or other means. When the recycle
flow pressure of 60 to 90 psi (415 to 620 kPa) is suddenly reduced to near atmos-
pheric pressure, the saturated air in the recycle flow is released in the form of mi-
crobubbles with an average size of 40 to 50 j.Lm. Water containing micro bubbles
has a milky appearance. These microbubbles are adsorbed on the Ilocs and these
bubble-Iloc agglomerates float to the surface of the water in the flotation tanks.
It is important to remember that the flotation tanks must be covered by a hous-
ing, as both rain and wind will break up the floated floes. Moreover, the flotation
tank must have a drain valve for removal of water as well as of accumulated sludge.
Air Saturation System Air saturation of water may be attained by various means:
injectors, a pressure vessel saturator, or a pressured vessel with packed materials.
Of the three, the packed saturator system is considered to be the most cost-effec-
tive. The recycling rate is generally 7 to 8% of the processing flow rate, which cor-
responds to approximately 8 to 109 of air per cubic meter of processing water. The
air/solids ratio is approximately 380 mL of air to gram solids for surface water; this
is about IS times higher than that for activated sludge thickening.
 BASIC APPROACH 193
192 DESIGN OF BASIC TREATMENT PROCESS UNITS

Floated Solids Removal Floating solids may be removed intermittently by tank must be coupled with a filter. Therefore, it is generally better to design the OAF
mechanical skimming devices or through flooding. A mechanical skimmer is ex~ process as an independent process uni.t. . .' .
pensive but creates sludge with a higher content of solids (I to 3%). The most com- A packed saturation tank system With a reciprocatlllg or a fu.ll-len~th cham-and-
m~)n types are chain and flight, oscillating, reciprocating skimmers, and rotating
flight float skimmer is often selected. However, the former system IS preferable. Pumps
skimmers. The use of flooding eliminates the need for mechanical equipment and required for sludge removal require special consid~ration, due to l.he large amounts of
has minimal effect on treated water quality. However, this method creates a very air contained in the sludge; a WEMCO or screw lin peller pump IS recommended.
low sludge solids concentration (0.2 to 0.5%) and produces a high percentage of
wastcwater (up to 2%). Design Criteria
It is important to note that raw water quality can dictate the selection of the ap-
propriate sludge removal system. Sludge produced from soft and highly colored wa- Flocculation Process
ter is easily breakable and must be skimmed off within 30 minutes of formation; a
full-length reciprocating skimmer should therefore be used frequently to skim off Minimum nllInber of tanks Two
the sludge at a speed of I to 2.5 ft/min (0.3 to 0.76 m/min). Sludge produced from Number of stages per tank Less than three (two normal)
turbid water or algae-laden water is stable, with sludge solids concentrations as high Water depth 12 to 15 ft (3.6 to 4.5 m)
Width of each tank Up to 30 ft (10 m): match to flotation tank width
as 2 to 3%. Therefore, the sludge may be removed on an intermittent basis.
Detention time 15 to 20 min
Effluent Design The clarified water in the tlotation tank must be collected from Energy input range (G) 30 to 120 S-1
Type of tlocculator Vertical-shaft tlocculator with hydrofoil blades
near the tank bottom under a curtain wall or by means of pipe laterals with orifices.
It is important to keep the water level in the flotation tank nearly constant to facil-
itate effective float skimming, and an eftluent water-level control weir must be pro- Flotation Tank
vided downstream of the claritied water collection system. Moreover, the effluent
collection level must be at least I ft (0.3 m) above the floor of the tlotation tank to Minimum number of tanks Two
minimize the chance of scouring any sludge settled on the bottom. Tank width Up to 30 ft (10 m)
Tank length In general, 40 ft (12 m) maximum
Cost The OAF system has a much smaller footprint than that of the conventional Water depth 8.5 ft (2.6 m) normal
Surface loading 4 to 8 gpm/ft2 (10 to 20 m/h)
flocculation and sedimentation process. This aspect makes the capital cost of DAF
less than that of the conventional flocculation and sedimentation process, however, Detention lime 10 to 15 min
the addition of mechanical equipment and housing may make the cost equivalent to
that of a conventional process. The operational costs of a DAF process is also higher Recycling System
than that for conventional sedimentation systems, due to the higher power use of its
Recycling rate 6 to 10% of clarified water
air saturators, recycling pump, and sludge skimmers; the DAF power requirement
10 g/m3
is approximately 3.5 kWh/mgd (2.5 kWh per 10,000 m3 /d), in contrast to 1.0 kWh! Maximum air loading 2
25 gpm/ft2 (62.5 n,'/m h)
mgd (0.7 kWh per 10,000 m 3/d) for regular tlocculation and sedimentation processes. Saturator loading rate
Operating pressure 60 to 90 psi (415 to 620 kPa)
Coagulant dosage is also somewhat higher-the process requires 13 to 15 mg/L of
alum-even when the raw water turbidity is less than 3 ntu. Overall, the cost of
DAF is similar to or higher than that of a conventional sedimentation process.
BIBLIOGRAPHY
Type and Selection Guide There is a special application of the OAF process,
together with the tiltration process, which results in the accumulation of tloc on the Braghetta, A.. et at., "OAF Pretreatment: Its Effect on MF," J. AWWA. 89:5:78 (May t997).
surface of the water above the filter bed, allowing for the tloc to be skimmed off Edzwald, J. K., et aI., "Flocculation and Air Requirements for Dissolved Air Flotation." 1.
and the clarified water to be filtered. This combined OAF and filtration process has AWWA, 84:3:92 (March t 992).
four basic drawbacks: (I) design and operation of the filtration system is compli- Johnson, B. A" et a1., "Pilot Plant Testing of DAF for Treating Boston's Low Turbidity Surfac.c
cated; (2) OAF loading is controlled by the filtration rate, so the OAF process can- Water Supply," IAWQ~IWSA-AWWA Joint Specialized ConferelJce, Orlando, FL, Apnl
not be optimized; (3) plant operators are not able to visually inspect the quality of 1994.
the clarified water (filter influent), due to tloating solids; and (4) each tlocculation Krofta, M., and Wang. L. K., "Flntation and Filtration." J. AAWA, 74:6:305 (June 1982).
 194 DESIGN OF BASIC TREATMENT PROCESS UNITS
Plummer, 1. D., et ai., "Removing Cryptosporidium by Dissolved Air Flotation," 1. AWWA
87:9:85 (September 1995).
Zabel, T.. "The Advantages of Dissolved Air Flotation for Water Treatment," 1. AWWA, 77:5:42
(May 1985).
3,2.7 Granular Medium Filtration Process
Purpose In a water treatment process train, the fundamental system that removes
particulate matter is filtration. The most common filtration process employs a gran-
ular medium of a certain size and depth. The pretreated water passes through the
filter bed, where a majority of the particulates are removed in the top portion, as
well as throughout the entire depth of the bed.
The primary objective of the design engineer should be to deliver a tilter design
that provides a steady production of high-quality water with minimal capital and
minimal operation and maintenance costs. When designing filters, most engineers
rely on either their own experience, the experience of others, or textbook design cri-
teria. However, engineers should expand certain design variables, such as the size
and thickness of the medium, tiltration rate. filter underdrains, filter washing Con-
ditions. and available headloss for filtration. These variables must be developed based
on the raw water quality. type of filter bed, and desired tiltered water quality.
It should be noted that there is a basic common denominator among the coagu-
lation, flocculation, and filtration processes. Under certain raw water conditions, ad-
equate treatment of the raw water can be carried out in the filter alone, and the need
for ordinary flocculation and sedimentation processes may be eliminated.
In this section we focus on the common granular media filters. Other types of
filters are discussed in Sections 3.2.8, 7.7, and 7.8.
Considerations Several factors must be taken into consideration when design-
ing a proper granular medium filtration process. They are (I) local conditions. (2)
design guidelines set by regulatory agencies such as the state department of health,
(3) site topography, (4) plant size, (5) raw water quality. (6) type of pretreatment
process, (7) new and proven types of filter, (8) provisions for future modification or
addition of filters, (9) type of filter wash system, (10) control of the filtration rate,
(II) type of filter bed, (12) chemical application points, and other miscellaneous
items. These subjects are discussed in detail.
Local Conditions Perhaps the single most important issue in filter design are the
local conditions. Factors such as the availability of qualified technicians and raw
materials, the local climate, and the soil and geotechnical characteristics of the plant
site all affect the design of the granular medium filtration process. For example, a
basic type of tilter bed should be chosen for a treatment plant located on a small
resort island or in a town of a developing country. Under these conditions, the best
type of filter is one that has a simple control system. It is tragic to see that many
design engineers fail to take this into consideration by designing treatment plants
with expensive filters and sophisticated control systems for regions lacking both
BASIC APPROACH 195
qualified plant operators and support pers?nnel. Not only are these plants produc-
' ing substandard quality water. they are rapidly detenoratlllg, due to the la.ck of qual-
ified maintenance personnel. The filter bed design should also be determllled by the
availability of filter materials at a reasonable cost.
Climate is the second issue that must be studied prior to filter design. In regions
of cold weather, the filters must be either covered or housed in a building to pre-
vent the water from freezing. Although the control cabinets are supposed to be her-
metically sealed, plants located in regions experiencing severe dust storms should
have a housing over the filter control system. The filter bed should also be covered
to prevent debris and sand from entering the filter, thereby preventing both opera-
tional and maintenance problems.
Finally, the soil and geotechnical characteristics of the plant site should be ana-
lyzed. Of primary interest are the level of the underground water table and the con-
dition of the soil because the filters are relatively deep and heavy structures. There
are a few documented cases where the filters became tilted or subsided as a direct
result of a high groundwater table and a poor foundation.
Large cities that are surrounded by a wealth of technology and qualified techni-
cians are prime candidates for treatment plants equipped with both a first-class fil-
ter system and a sophisticated backup system. Some of the filters used in the treat-
ment plants of these large metropolitan areas are capable of operating at a rat~ of 8
to 13 gpm/ft 2 (20 to 33 m/h), which is well above the average filtratIOn rate III the
United States of 5 to 6 gpm/ft2 or 13 to 15 m/h. However, both pilot studies and ac-
tual plant operational data show a significant increase of number of particles (2 to
10 11m) in filtered water at filtration rates above 8 to 10 gpm/ft2 (20 to 25 mlh).
Optimization of filtration has attained grave importance in recent years, due to the
need to remove Cryptosporidium oocysts and Giardia cysts. Most water purveyors
are now limiting their filtration rate to 6 gpm/ft2 (15 mlh), especially during cold
weather months.
Guidelines Set by the Regulatory Agencies Guidelines for designing treat-
ment plants can be found on both the state and county levels. They are especi~lIy
stringent with respect to filter design. A few of the requirements are worth ~otlllg:
maximum filtration rate, continuous monitoring of the filter effluent turbidity, re-
wash (filter-to-waste) requirements, and the types of filter that can be employed. For
example, the state of California discourages the use of pressure filters in public wa-
ter treatment plants because the condition of the filter bed cannot readily be ob-
served. Therefore, the bed could potentially be in poor condition without the oper-
ator's knowledge, California also discourages the use of declining-rate filtration
control systems because of the likelihood for high-turbidity breakthrough at the be-
ginning of the filter cycle because of a filtration rate that is twice or more the nom-
inal rate at the beginning of the process.
A few states also dictate the physical features of filters. such as media size, depth
of the filter bed, spacing and height of the wash trough. and double-walled con-
struction of the upper and lower gullet with a drainage pipe to assist operators in
detecting leakage of unfiltered water from the upper gullet. The Tefl States Standards
 196 DESIGN OF BASIC TREATMENT PROCESS UNITS
and the rules set by the state of Utah are examples of this type of design control
On the national level, the EPA is concerned primarily with the removal of harmfui
microbes, such as Giardia lamblia, CryPfosporidium, and Legionella, and viruses
from the water supply. This can be achieved through coagulation followed by fil-
tration and disinfection.
Site Topography The topography of the plant site must be evaluated carefully
prior to design since it determines the location, type, and arrangement of each unit
in the filtration process. For example, the head loss across a water treatment plant
typically ranges from 15 to 25 ft. Therefore, if the plant site is flat, the filter struc-
ture and dearwell must be'located deep within the ground unless a pump is placed
In the mIddle of the process train. If a gravity flow is to be maintained across the
plant, the only other alternative would be to elevate the pretreatment unit processes
above the ground.
Topography also dictates the type of filter backwash facility that may be used. If
a suitable hill exists on the plant site, the wash tank can be built on it. However, a
direct pump wash or self-backwash system should be seriously evaluated if the plant
site is flat. This minimizes problems with freezing and avoids the unsightly ap-
pearance of the elevated wash tank. Additionally, the energy effectiveness of the
self-backwash filters is gained. The ideal plant site has a constant slope of 2 to 3%,
which allows the filter wash tank, filters, and wash-waste holding tank to be situ-
ated easily and built economically since there is no need for excessive excavation.
The slope also helps to satisfy the water level required for each process. A steeply
sloped topography not only severely restricts the arrangemcnt of the unit processes
but greatly increases the probability of landslide.
Plant Size The size of the water treatment plant is instrumental in detcrmining
both the physical and performance requirements of filters. Plant size usually dic-
tates the total number of filters, the size of each filter, the degree of instrumental-
ization of the filter control system, and the method by which the liltration rate is
controlled. Some existing large plants, those processing over 200 mgd (9 m'/s), have
a total of more than 30 filters. However, the total number of filters may be limited
if either the size of each filter is increased [up to a practical limit of 2000 ft2 (100
2
m )] or if a high filtration rate such as 8 gpm/ft2 (20 m/h) is adopted, or if both
methods are implemented.
The capacity of the treatment plant also aids the design engineer in selecting the
most appropriate type of filter. For example, plants operating on a small scale [less
than 3 mgd (130 Lis)] can generally use proprietary lilters supplied by equipment
manufacturers, also known as package filters. These types of lilters are generally
cost-effective, easy to operate and maintain, and perform well for plants of this size
as long as the package filters are selected properly, that is, based on a rigorous en-
gineering evaluation.
Raw Water Quality and the Type of Pretreatment The raw water characteris-
tics should also be scrutinized when designing the filter bed since these character-
BASIC APPROACH 197
istics determine the type of pretreatment, which in turn determines the type of
filter bed. For instance, if the underground water contains high levels of iron or man-
ganese, the best method for removing these compounds is by oxidation with chlo-
rine or potassium permanganate, followed by pressure filters; the safe yield Irom a
well is generally less than 3 mgd (0.13 m 3/s). This method is the primary choice
for well-water application because it preserves the pressure of the pumped water
and allows for backwashing of the filters using the effluent from other pressure fil-
ters on the line; additional backwash pumps are therefore not required. Iron and
manganese removal may also be achieved through direct filtration with preoxida-
tion by chlorine, ozone, or permanganate. Green sand has been used as a filter
medium in this type of situation, although sand may be used as an alternative.
To optimize the performance of the filtration process, the filter bed design should
complement the pretreatment process. For instance, under conditions where direct
filtration or in-line filtration is applicable, the filter bed should be designed with a
larger storage capacity for suspended solids than that for conventional beds.
Therefore, a reverse-graded or coarse and deep bed should be selected. The pres-
ence or lack of a high concentration of suspended solids in the raw water requires
special consideration when selecting the type of filter bed. Filtration of algae-laden
water is very difficult even after flocculation and sedimentation, due to rapid clog-
ging of the filters by algae. Use of a coarse deep bed or a reverse-graded filter bed
should be evaluated under these circumstances. Conversely, a common monosand
bed (often referred to as a rapid salld bed) may be considered when the filter in-
fluent is very good, the result of a conservatively designed pretreatment process.
Since the bed depth of these types of filter is relatively shallow [less than 24 in. (0.6
m)], surface filtration, not in-depth filtration, plays a major role in the monosand
beds. Two-stage filtration should also be evaluated if the raw water quality is good
(i.e., has low turbidity), thereby allowing both the flocculation and sedimentation
processes to be eliminated.
Water temperature has a great impact on any filtration process. For example, dur-
ing the winter when water temperatures are below 40°F (4.SOC) the filter run length
is almost half that of the summer months (over 70°F or 21°C) at the same filtration
rate. Filtered water turbidity is also higher in the winter. Moreover, in the winter,
filter bed expansion is almost twice that of the summer under the same backwash
rate. Therefore, a significant amount of media loss may occur during the winter un-
less the wash rate is adjusted.
Due to the limited amount of water for domestic use in large urban regions, ter-
tiary treatment of sewage for water reuse has become an important issue. The filter
for biologically treated sewage behaves somewhat differently from that used in the
water supply field, although the physical characteristics are the same. The graphs
in Appendix 18 illustrate the typical performance of tertiary filters.
New and Proven Types of Filters Like any other type of equipment, new forms
of filtration systems, filter media, and filter control systems become available on the
market at any given time. Yet the design engineer must be extremely careful with
regard to using the new products. Although these new products are appealing, there

198 DESIGN OF BASIC TREATMENT PROCESS UNITS
is likelihood that problems will emerge after a year or two of continuous operation.
For example, one company marketed small aluminum chips as a filter medium.
According to the manufacturer, chlorine (added to the filter influent) slowly dis-
solved the aluminum and allowed coagulation and flocculation to occur within the
tilter bed, eliminating the need for coagulant. The concept was very appealing and
the tilters performed well during the first 3 to 6 months of operation. However, af-
ter 9 to 12 months the filter beds became a large chunk of fused aluminum chips.
Similar unforeseen problems have also occurred with new types of filter underdrain
systems.
Despite convincing laboratory test results, engineers should remember that real
lilters are subjected to many different conditions that cannot be duplicated in either
a laboratory or a piloting facility. Since most critical problems occur within a year
or two of operation, it is prudent to wait until these problems are isolated and re-
solved. Design engineers should not hesitate to research new products, but they are
strongly advised to select items that are proven to be reliable, with low maintenance
and ease of repair.
Allowing for Future Filter Modifications or Additions The design engineer
should anticipate an increasing demand on the plant output by including a provi-
sion for future modifications of the filters or addition of more filters. The process
should therefore be designed so that future construction will have minimal impact
on plant operations. In many cases, water treatment plants increase their operating
capacity by adding basins and filters in stages. If the original treatment unit processes
are not designed as a module, the new lilters must be added onto the original filter
bank. In situations such as this, the lilter influent channel, effluent channel, and
backwash system must be designed to accommodate the new flow rate, but based
on the original plant specifications.
Construction during plant expansion should have minimal impact on the exist-
ing facilities. For this reason, the module expansion scheme must have a carefully
conceived master plan for the yard pipings. The plan should maintain equal distri-
bution of water flow to each module and should not have any physical interference
between the major pipelines. The need to provide for future filter modifications has
become more of a reality due to the EPA's proposal to make future water treatment
regulations more stringent. It is also probable that the old monosand bed will be re-
placed with dual-media reverse-graded beds, such as anthracite and sand, to increase
the output of the filters.
Presently, the most important consideration is the possibility that granular acti-
vated carbon (GAC) beds will be required to adsorb objectionable organic com-
pounds-an EPA rule that is anticipated to be implemented in the near future. One
method would be to replace the existing filter medium with GAC medium. However,
the GAC bed requires an empty bed contact time (EBCT) of at least 10 min, thus
necessitating a bed depth of 6 to 8 ft as long as the filtration rate is at a level rate
of 4 to 6 gpmlft2 (10 to 15 m/h). Under no circumstances can such a depth be pro-
vided by ordinary tilters unless the filter cells are designed initially with the option
to convert to GAC adsorption fillers.
BASIC APPROACH 199
Deep bed filters also demand special design features, such as air and water back-
wash (also referred to as air-scourillg wash); the air-scouring backwash system re-
quires a specific type of filter underdrain system. Mo.reover, rece~t prototype stud-
ies indicate that deep bed filters should be designed with both an alf-scour wash and
surface wash system whenever polymers are used as part of the coagulant to pre-
vent mud ball formation.
Filter Washing System There are four basic schemes in filter washing: elevated
wash tank, direct pump, self-backwash (Greenleaf type), and continuous backwash
(Hardinge or Dyna Sand filters). The lirst two systems are the traditional and proven
lilter washing schemes. The self-backwash filter may be either a proprietary Item
or custom designed by the engineer. Continuous backwash tilters are proprietary
items. Filter selection is based on plant capacity, site topography, local conditions,
and other factors, such as energy efficiency, the wash-waste handling situation, and
owner preference. . ..
The elevated wash tank and direct pump wash systems have ample tlexlblltty
in adjusting the backwash rate. The elevated tank scheme is particularly suited
for a plant site with high ground on which to locate the tank. The advantages at
the elevated tank scheme are its effective use of energy, a smail-capaCity pump
that fills the tank within an hour or two, and emergency reserves of pressurized
water in case of power outage. However, the elevated tank system requires a large
tank to satisfy the two filter wash requirements: a large and long backwash pipe
and a minimum elevation of 33 n (10 m) above the filter wash troughs.
Conversely, the direct pump wash system eliminates the need for a large eleva~ed
wash tank, and control of the wash rate is relatively easy. However, both the Size
of the pump and the horsepower required become quite large for large filters. The
direct pump system is therefore best suited for small filters or air-~coUl'ing ~ash
systems since these must change the backwash rate two or three limes dUring a
washing cycle.
With respect to the washing system, the basic alternatives arc backwash ~lone, a
combination of surface wash and backwash, ordinary air-scour wash and Slll1ulta-
neous air and water wash. Backwash alone cannot adequately maintain clean filter
bed conditions unless coagulation and tlocculation are achieved with a minute
amount of alum (without polymers). Rapid sand tilters, ordinary dual-media beds,
and tertiary filters are all cleaned effectively by backwash supplemented by surface
wash. In contrast, coarse deep bed and tertiary filters are more suited to ordinary
air-scour wash or simultaneous air and water wash. The latter requires special buffer
plates around each wash trough (see Figure 3.2.7-6).
Some authors claim that the air-scouring wash (see Figure 3.2.7-7) is a far su-
perior system than surface wash. But this claim loses credibility after reviewing the
performance of several operational plants: the data show that a surface wash sys-
tem that is properly designed and operated can match or even excee~ the ~~rfo~'­
mance of an air-scouring system. For example, a review of four plants In Cali forma
revealed that tertiary tilters fitted with dual-arm surface agitators were devoid of
mud balls and had clean bed conditions despite 15 years of operation. In contrast,
200 DESIGN OF BASIC TREATMENT PROCESS UNITS

filters in both the United States and Europe that employed air-scour wash systems
occasionally had a significant mass of mud balls whenever polymers were used as
HWl
coagulant or filtration aid.
Design engineers should not be misled by the impressive boiling action at the rrller bed

surface of the filter bed during air-scouring wash because this violent boiling action lWl

actually occurs in only the top 6 to 8 in. (0.15 to 0.2 m) of the filter bed. Therefore, CIUrweU
mud balls that have descended below this boiling zone remain unbroken. It is worth-
while to mention that several water treatment plants in the United States, originally Constant Rate F.ltratlon
(Flow meter & lIow modulation val'ltl
designed with ordinary air-scouring filters, have added the surface wash system in
response to the problem of mud ball formation,

Filtration Rate Control There are basically two modes of filtration: constant-rate
filtration and declining-rate filtration, Despite claims to the contrary, both constant-
and declining-rate filters are capable of producing filtrates with less than 0,1 ntu HWl
turbidity. However, the colloidal particles must be conditioned properly in the pre-
treatment process, and the filter bed must be both designed and maintained prop-
lWl
erly.
Only four basic types of rate-control systems have supplied consistent and reli-
able performance: (I) constant-rate filtration with a flow meter and a flow modula- Constant ltvel Filtration .
tion valve; (2) constant-level filtration with equal-flow splitting inlet weirs, a wa- (Influent control, level sensor & modulating wive)

ter-level sensor, and a flow modulator valve; (3) declining-rate filtration with a
submerged inlet and a flow restriction in each emuent pipe that mayor may not be Variable
fitted with a weir to control the emuent level; (4) constant-rate filtration with equal-
flow splitting inlet weirs and a weir to control the common effluent level. These HWl
four basic types of systems are represented in Figure 3,2,7-1. _BaCkwash
Each of these control systems has both good and bad features. The last two sys-
tems are simple and require very low maintenance. However, only a plant with more lWl
than six filters can fully maximize these positive features. The first two systems have -=- Clearwell
been the standard flow control measures for the past 50 years and are still quite pop-
ular among design engineers. Therefore, a large percentage of operational filters are Declining Rate Filtration
equipped with these types of systems, despite their high capital cost of approxi- (No Influent control. no modulatlnc ",I~. an Ofl'lce plate)
mately $25,000 (average) per filter.
Of the four rate control systems, the first (historically) has the highest mainte-
2-3 Feet
nance cost and requires frequent operator attention. Another negative aspect is that
the filtration mode changes from the designed constant-rate filtration to declining- i
rate filtration when the control system fails, Many old constant-rate filters have been '" HWl

operating in the declining-rate mode because their /low control systems are beyond t
repair, This is particularly true for treatment plants in developing countries.
The declining-rate filtration system is very simple to design and build and gen- 'Q lWl
erally produces good-quality water. However, to meet the present high standards for "iT

filtered water quality, this mode of filtration requires vigilant plant operators be-
cause the water level in the filter cells tends to fluctuate widely and creates the po- RiSinl Level. SeIt-8ackwUh FillerS i I
!Influent control, no modulatlnl v.l1~, no backwash P p nc)
tential for partial drainage, overflow, and potential initial turbidity breakthrough at Ad d f m Water Treatment: Principles and
beginning of the filtration cycles, unless the design has been made very carefully. Figure 3.2.7-1 Basic filter control systems. ( . apte r? New York 1985)
It is precisely because of these problems that some regulatory agencies, such as in Design by 1. M. Montgomery. Consulting Engmeers, Wiley. "
201

202 DESIGN OF BASIC TREATMENT PROCESS UNITS
the ,state of California, discourage the use of declining-rate filtration in new filter
deSIgns.
One operational problem associated with declining-rate filters is that there are no
clear parameters for filter washing: the head loss for all the filters on-line is nearl
equal, there ,IS, no flow metering device to indicate the filtration rate of each filte~
and the turbidity of the filtered water does not increase with time. Thus operators
usually wash the filters based on the length of the filter run. It is an undeniable fact
t?at very few filters in industrialized countries arc designed as declining-rate filtra-
t~on system, beca~s~ there are no positive means of controlling the filtering condi-
tions. ,Yet t~e dechnlll~-rate syst~m still satisfies the needs of many filtration plants,
espe~lally In developmg countnes, because of its simplicity and low maintenance
reqUirements and the less stringent governmental requirements.
. The fourth type of system, constant rate with equal loading to each filter with a
nSIng water level, IS cap~ble of backwashing a lilter by using the emuent from the
rest of the filters on the hne. These are commonly referred to as self-backwash fil.
ters. A few equipment manufacturers supply these types of lilter with four filters
per module: The major drawbacks of the self-backwash filters are the limited flex-
1~lhty 10 adjustIng the backwash rate, lack of Ii Iter to waste capacity, and the deeper
hlter cells; they are at least 4 ft deeper than the normal lilter design. However, there
are posItive features to this type of system: the lack of a flow modulation valve; the
~lmple ope~~ng and closing of the inlet and waste-wash valves to initiate filter wash-
I~g; the ablhty to function without the system of backwash pumps, tanks, and wash
hnes; and the easy determination of headloss through the filter (by measuring the
water level In a filter cell). Stnce the water depth increases with the duration of the
filter run, the problem of air binding is minimized. The self-backwash Ii Iter is there-
fore particularly suited for situations where the raw water is saturated with air that
is ~ntr~in~d in the water. Moreover, this type should always be evaluated because
of Its limited use of mechanical equipment, including backwash pumps and valves.
Also, the filter-to-w~ste feature. can be incorporated fairly easily. Presently, there are
three large plants, wIth capacities exceeding 400 mgd (19.5 m/s), using this type of
filter: two 10 Cahfornla and another in Manila, Philippines.
Type of Filter Bed The selection of a tilter bed is first contingent on the avail-
ablhty of bed material. Other factors, such as the primary purpose of liltration the
type of filtratIOn process, anticipated filter run time, and the filtration rate at: the
system, also influence the decision.
The first step in selecting a filter bed is to study the local conditions. In the United
States, almost any type of common material is available at a reasonable cost: sand,
crushed anthracite coal, granular activated carbon (GAC), and even garnet sand. This
~s not the case for most developing countries and a few industrialized nations. For
Instance, a large part of Africa and Southeast Asia use sand and gravel as the pri-
mary filter matenals because other types are simply too expensive. Although only
three types .of granular media are generally used, lilter media can be composed of
other matenals, such as pumIce or synthetic material, provided that there is no mea-
surable leakage of unwanted substances from the media, no significant attrition loss,
BASIC APPROACH 203
and the cost is reasonable. For instance, porn-porn-like synthetic fiber balls called
Marima are used in a patented filter system manufactured in Japan.
The second step in selecting an appropriate filter medium is to delineate the pri-
mary purpose of the filtration process. Sand and the combination of sand and an-
thracite coal are the two basic and proven alternatives with respect to ordmary wa-
ter purification. Alternatively, the more expensive garnet grains may be used in
conjunction with sand or anthracite coal. If the main purpose is the filtration of
sewage, as in the tertiary filtration of sewage, monosand or mono-anthracIte coal IS
the most appropriate filter bed medium. A dual-media filter bed may also be con-
sidered because of its proven performance, but the disadvantage is that it tends to
have problems with sludge accumulation in the sand-<:oal interface. However, this
problem may be eliminated by employing air-scouring backwash or dual-arm sur-
face agitators to wash the sludge out of the interface; but the alf-scounng wash sys-
tem requires the lilter bed to be restratified with a high backwash rate because the
sand and coal are intermixed during the wash, so the advantages of smaller back-
wash pipe lines, valves, and waste-wash tank arc lost. Conversely, a monomedium
bed does not require restratification (i.e., may not require a wash rate step-up con-
trol system), and the advantages of the air-scouring wash can be fully utilized.
The filtration process also affects the selection of the filter bed because of the
special requirements of each type of process. The direct and in-line liltration
processes must have filter beds with a large floc holding capacity.. A rev~rse-graded
filter bed, such as a dual-media or coarse deep bed, satIsfies thiS reqUirement. In
two-stage filtration the filter bed of the lirst stage acts as a roughing filter and c~r­
ries out the flocculation process. Data obtained from pilot filter tests and actual 10-
stallations using the two-stage filtration process indicate that the first-stage filter bed
may be designed in the same fashion as an ordinary filter.
The fourth step in selection of the filter bed is based on the filtration rate. For
slow sand filtration the most appropriate choice is a fine-grained monosand bed
since the filtration rate for slow sand filters is less than 0.2 gpm/ft2 (0.5 m/h). Rapid
sand filtration, with filtration rates ranging from 2 to 3 gpm/ft2 (5 to 7.5 m/h), usu-
ally have medium-sized sand (0.5 mm E.S.). Filters displaying filtration rates of 5
to 10 gpmlft2 (12.5 to 25 m/h), commonly known as high-rate filters, always con-
sist of a reverse-graded filter bed or a deep, large-grain, monomedium bed. Details
of these filters are discussed elsewhere in this chapter.
Chemical Application Points Disinfectant, filter aids, adsorbent (for taste and odor
control) and alkali chemicals (for pH control) are the most common chemicals added
to the fi'lter influent to aid the process of filtration. Three important design consid-
erations must be addressed with regard to this subject: Minimize the number of chem-
ical application points, recognize that certain chemicals have the potential to "break
through" into the filtered water, and ensure that the chemicals are adequately diS-
tributed in the water. Finally, the design engineer should be cognizant that the loca-
tions of the application points determine the type of chemicals that may be applied.
The lirst item is to minimize the number of chemical application points. Hlgh-
rate filtration often employs a polymer as a filtration aid. Polymer and chlorine are
 BASIC APPROACH 205
204 DESIGN OF BASIC TREATMENT PROCESS UNITS

h ve wash trou hs since the filters are effectively cleaned, pr~­

fed to the filter influent to promote filter efficiency and keep the filter bed clean. If tial wash and do not a . gl h 12 ft (4 01) a monosand filter bed IS
the plant design provides only one water passage between the pretreatment process vided that the width of the fdters IS ess/It~; inlluent Th'is type of system has op-
units and the filter bank, these two chemicals may be fed to the water at this single used, and no polymer IS applied 10 the I t (s~e Figure 3 2 7-7). The choice
d · E since the early twentlet1 cen ury " . II
location via a mixing device. In many treatment plants there are several passages erate rn urope . . ontingent on tilter type (especia y
between the settled water channel and the tilters. This requires the chemicals to be b tween the two types of washmg systems .IS c
e h f h fi It bed) and deSigner preference.
distributed to these several application points and necessitates the use of additional the size and dept 0 tel er , '. should real ize that the various
chemical feeders and tlow-splitting equipment. When designing the filtration process'b~~~~n:~~ss Gravity filters generally have
The second important issue is the potential breakthrough of the added chemicals filters provide different ~mount~ of ~1~0~~2 5 to 3 m). 'Other types of fiiters, such as
into the filtered water; this phenomenon degrades the quality of the filtered water. an available headloss at about to k . d fi"lters do not provide more than 6 ft
Some treatment plants routinely add alkali chemicals such as lime or sodium hy- proprietary, valveless, or monovalve pa.c ~g\wash t~lters (Hardinge type) typically
droxide to the filter influent to raise the pH to 8 or above, following flocculation (1 8 01) of available headloss. Automatic ae.
. . flO ft (030 01) of avarlable headloss.
with alum and sedimentation. This practice always produces a high aluminum con- provide a maXImum 0 , , .: ' '1 in their filter washing design and the
centration in the finished water due to the pH and the basic characteristic of the alu- Filter underdrain systems dlfter pnmMan Yt . cipalities in the United States use
f fiI uired by each system os mum f
minum hydroxide (see Figure 3.2.3-1). Another common practice is to add pow- types 0 Iter req . . ' riet of underdrain systems, ranging rom
dered activated carbon (PAC) to the filter influent as a means for taste and odor large filters r"nd ~~~~~~I::~hp~a;i~d~I~~kS, ~recast concrete laterals, false bottoms
control. However, a significant amount of carbon breaks through the filter and ap- clay to rern orce I' the full potential of the treatment plant
pears in the finished water. Particle counters may be used to monitor the filter ef- with strainer, to porous plates, Toldreallze nderdrain system that is most com-
d ' n engineers shall se ec t an U · . d
tluent, but nepherometric turbidimeters are virtually incapable of detecting carbon processes: eS1g,. h sical characteristics of the filter medIa, an
leakage. patible With the Size ot the filter, the p y. I nted If air scouring is chosen as
The third issue, adequate mixing of the chemicals at the application point, can the type of filter washing system that IS lln P eme ted 'type of underdrain is a false
h' t the most common and accep ,
be resolved by including either an appropriate type of chemical diffuser pipe in con- the was mg sys em, . H wever there are several disadvantages 10
junction with a baffled channel or an in-line static mixer (for the pipe). bottom fitted with long leg strall~;rs~f t~e strai'ner heads may become clogged by
The last issue, with respect to the application of chemicals to the process water, using this type of system: the SIS h ' d the false bottom during construc-
is the selection of the feeding points. The filter influent, the filter effluent, and the scale, filter media or debns le,ft l~ t e Plp:s~~~n false-bottom floors became loose
backwash line are the three potential points. Chemicals that may be fed to the in- tion. There have also been a tew mst~~ce 11) avoid clogging the strainer heads, it
fluent are chlorine, polymers, and to some extent PAC. Chlorine, alkali chemicals, and lifted up after repeated filter was rngs' (.' s thoroughly after construction.
. . I t lean the underdrarns and pipe . d
corrosion inhibitors such as sodium hexametaphosphate, and tluoride compounds IS essentl3 a c , . f 5 ' (015 mm) or an orifice SIze (un er-
(for fluoridation) should ideally be applied after filtration, because of their function Mo~eover, a slit sizehof ~_ ~1l1(I;n~:~)oin ~di~~~et~~ should be selected whenever pos-
and effectiveness. The two chemicals that may be fed to the backwash water to dis- dram blocks) larger t an IX 10.
infect and condition the filter beds are chlorine and polymers. It should be noted sible. . -I d rd rain system are uniform flow
that the recent EPA standards may force many treatment plants with filters lacking The four essential factors in sele~t~ng a t~ te~ ,un :nd cost-effectiveness. The flow
a filter-ta-waste function to feed polymers to the backwash line to minimize the ini- distribution of the backwash, durabllItY'1 relt~albllItY'11 is kept uniform by (I) making
. . . f' h b k ash water rn t 1e I tel' ce
tial turbidity breakthrough of the filters. dlstnbutlOn 0 t e ac w II h to introduce a con-
the orifice or slits of the filter UI~derdrai~ syste~n cSi~nain ~~~~~essurized conduit up-
Miscellaneous Items Design engineers must also consider several additional is- trolled headloss, and (2) decreasmg the tho~ v~r~uli~ grade and energy lines of tlow
sues when designing a proper granular medium filtration process: the use of wash stream of the underdrall1 system so th~; t e.~ The false-bottom underdrain sys-
troughs; the amount of allowable headloss for filtration; and the type of filter un- entering the llnderdrain system are fau y lim O~I~. s minimal maintenance costs.
derdrain, type of filter, and waste-wash-water handling facility. In the United States, tern usually provides the best performance an I
n
~ determines the type of filter that
wash troughs are an essential part of the filter design, due to the predominant use The overall capacity of a water treatme~t p,a automatic backwash filters,
t
of a high backwash rate. Filter baekwashing, with or without surface jets as the aux- may be included in the filtration process. or ms ancel' ed t' mall plants those
fl' frequent Y us or s "
iliary scouring system, necessitates the use of wash troughs because prompt removal self-backwash tilters, and pressure I ~~I S ~~ t es of filter are proprietary items
of the dirty wash waste is essential in minimizing both the washing time and the processing less than 15 mgd (0.65 m s). Th eseot~'~r advantages such as low capital
y
formation of mud balls in the filter bed. Conversely, an ordinary air-scouring wash marketed by a number of manutacturers. (e t' pr'ssure filters) good filtered
system uses a low backwash rate, and the washing time is therefore twice as long cost, minimal headloss across t~e filters ~xcep In ~edillm to lar~e plants, those
water quality, and simple operatIon and mamtenance.
as for high-rate back washing. Most air-scouring systems in Europe utilize sequen-

206 DESIGN OF BASIC TREATMENT PROCESS UNITS
processing over 10 mgd, generally use gravity down flow tilters with dual-media fil-
ters constructed of reinforced-concrete structures.
Upflow filters, where the water passes upward through the filter bed, have also
been used on a limited basis. One obvious advantage of upllow filters is that they
can achieve coarse to fine filtration with a single medium such as sand, and thereby
eliminate the need for more costly multimedia beds. However, one serious disad-
vantage of an upllow filter is that the filter bed expands as it becomes clogged and
allows previously removed solids to escape into the eftluent. An extra deep bed (min-
imum of 6 ft) containing coarse sand (1.5 to 2 mm) or a system of restraining grids,
placed at the surface of the filter bed, may be used to overcome this problem if
normal-sized medium is used. This particular type of design is seldom considered
for municipal treatment plant applications because of the extreme care that must be
exercised during operation. Therefore, its use is restricted by regulatory agencies.
If a treatment plant undergoes expansion, the design engineer should try to match
both the size and elevation of each new filter to existing filters so that the old back-
wash facility may still be used. This not only reduces the cost of expansion but also
maintains continuity in operating procedures. For ordinary filters, the size of the
waste-wash-water handling facility is directly proportional to the plant capacity and
the size of each filter cell. The automatic backwash filters sequentially clean small
portions of the tilter bed and thereby constitute a continuous filtration process; this
type of filtration process produces a small amount of wash waste on a continuous
basis. For this reason, the waste can easily be handled by a very small facility. In
contrast, the large filters used at large plants require a sizable handling facility be-
cause these filters produce a significant amount of instantaneous waste flow during
a washing period. Regardless of the type of filter, 2 to 3% of the plant Ilow rate is
waste flow and the designcr must therefore evaluate alternative types of filters and
the size of each filter whenever the plant site does not have cnough areas to ac-
commodate a large waste handling facility.
Type and Selection Guide In this section we briefly summarize the common
filter units and provide guidelines for making the proper selection.
Available Alternatives There are three basic types of granular medium filters:
(1) slow sand tilters, (2) rapid sand filters, and (3) high-rate filters with either a
coarse deep monomedium or reverse-graded multimedia bed. Both the rapid sand
and high-rate filters require chemical coagulation and, in most cases, clarification
prior to filtration so as to optimize filter performance. Coagulation and clarification
pretreatment remove excessive amounts of suspended solids, color, and some min-
erai components. The pretreatment also removes a large portion of microorganisms
such as algae and Giardia.
A commonly encountered design Ilaw is that the type and degree of pretreatment
do not match the filter requirements. For instance, if the source is a large lake with
consistently good water quality, a slow sand filtration or a direct in-line filtration
system, as well as a two-stage fdtration system with high-rate filters, may be adopted.
However, if the quality of the raw water is generally good with occasional high tur-
BASIC APPROACH 207
bidity spikes, algal bloom, or taste and odor problems, a complete treatment process
must be considered; the treatment process train should include a OAF process or a
settling process. In this type of situation the sedimentation tanks may be either by-
passed or used as a large channel between the flocculation and filtration process,
but only during periods of good raw water quality, by using the direct tiltration mode
and by employing small amounts of coagulan!. From these examples it is evident
that design engineers must review both the raw water characteristics and the sea-
sonal variations in the water quality and should seriously consider running a bench-
scale or pilot study to evaluate the treatability of the raw water. Only then can a
proper pretreatment scheme be established.
The filters can be classified into four descriptive categories based on the filtra-
tion rate, the type of filter washing required, the type of filtration rate control, and
the type of filter bed selected. As discussed previously, slow sand, rapid sand, and
high-rate filters are the three commonly recognized classifications. Yet from the
viewpoint of the design engineer, a more detailed division is required.
The first group comprises gravity filters: slow sand, rapid sand, and high-rate fil-
ters containing a multimedia or a coarse, deep medium bed. The second division
consists of pressure filters, which encompass proprietary items such as the hori-
zontal single cell, horizontal multicell, vertical single cell, and vertical billow fil-
ters. The third group is that of automatic backwash filters: these are exclusively com-
posed of proprietary items. Finally, custom-designed filters include standard rapid
sand filters, multimedia high-rate filters, coarse deep bed high-rate filters, self-back-
wash gravity filters, constant-rate gravity filters, declining-rate gravity filters, GAC
adsorption filters, and ion exchange resin filters.
Selection Criteria Before designing a filtration system, engineers must establish
the following items: type of filter, size and number of filters, filtration rate and ter-
minal headloss, control of the fdter flow rate, characteristics of the filter bed, type
of filter wash system and wash trough, filter underdrains, auxiliary scouring, and
filter appurtenance systems. Specification of these design elements should be based
on the following conditions:
1. Local conditions, such as weather, the level of available technical service,
the capability of operators, and the topography of the plant site
2. Regulatory constraints
3. Plant capacity [Does the plant process more than or less than 10 mgd (0.5
m3/s)?]
4. The quality of the water source: turbidity, color, temperature, amount of dis-
solved air, algae count, taste and odor, and Iluctuations in the water quality
5. The final quality of the process water, that is, the Drinking Water Quality
Standards
6. The type of overall treatment process train: conventional complete treatment,
direct filtration, iron and manganese removal, organic compound removal,
or demineralization
 BASIC APPROACH 209

208 DESIGN OF BASIC TREATMENT PROCESS UNITS

. due to washing or repair. The design en-
ne or two of the filters are placed off-hne umber of filters and the proper Size for
7. The condition of the plant site and the effective use of space o. r must therefore select an adequate n .n·lng units will not be exceSSive
8. The amount of available hydraulic headloss across the filters gmee . t of the remal d
r sO that the filtration ra e . ded that no more than 33%, an
eaeh filte
I , . It IS recommen
9. The initial capital cost and the cost of operation and maintenance when at least one is out of serVice. f h draulic surcharge be allowed to flow to
10. Owner preference referably a maximum of 15 to 20%, .o . ~ff-line. This practice minimizes any tur-
P .. tilt whenever one UOi t IS
II. The potential for future plant expansion and modification the remaining I ers occur during shock loading.
bidity breakthrough that may .
TYPE OF FILTER Small to medium-sized plants processing less than 15 mgd (0.65 Contrar to popular belief, the filtratIOn
m3/s) may use proprietary filters; these should be evaluated along with custom- FtLTRATtON RATE AND TERMINAL HEADWSS ntl affec/the quality of the filtered water
high-rate filters does not slgmfic~ . Y I ' f the suspended colloids (In
designed filters. The proprietary units offer several advantages: lower capital cost, rate af , 25 m!h) ThIS IS true on y I II
easy operation and maintenance, a filtered water quality ranging from acceptable to for rates up to 10 gpmlft ( . a combination of coagulant and a sma
the raW water) are properly coagulated by ded solids are removed during pretreat-
good, and shorter plant design and construction periods due to the predesigned fea-
mount of polymer and if the excess susp~~ s is developed during higher fiItrati?n
ture of these units. The main disadvantage is that they generally have a limited life
span of 20 to 30 years. There is also a possibility that future use of certain propri- ~ent. However, the rate at whl~h th~h~~~It~:tion rates over 10 gpm/ft' are consl~~
etary units may be restricted because they may not be able to meet the more strin~ rates is quite dramatic, so muc so I edia filters Figures 3.2.7-2 and 3.2.7 .
ered to be impractical for normal dua -~f hi h-rate fllters. When processing phys-
gent guidelines of the regulatory agencies. Other disadvantages include their ques-
illustrate the performance characterlS~"cs alit; of the fliter effluent tends to degrade
tionable aesthetics and limited flexibility with respect to filter design and operation.
ically weak floc such as alum floc, t e qu
Automatic backwash filters, self-backwash filters with a module design, and steel-
shelled gravity or pressure filters are all proprietary types of filter. WT = 13·C
The most common types of filter that are included in the design of large treat- . . 1 TU Alkalinity = 120 mglL, pH = 8.1, ..
Conditions: Raw water: Turbld.ltb~ ~~l~~tion '(pilot filter with/dual media bed)
ment plants, those processing a minimum of 15 mgd, are the gravity-down flow and Treatment process. ~r~ gilL Cat Floc-T -= 0.35 mg!L
Coagulants: Alum = . m ,
dual-media (sand and coal) lilters. The dual-media lilter with a reinforced-concrete
cell structure is the standard design that is used by engineers because its life cycle
0.5
spans 50 years or more.

SIZE AND NUMBER OF EACH ALTER The size of each filter and the total number of fil-
ters are interrelated subjects. The maximum filter size is dictated by the difticulties en-
countered in providing unifonn flow distribution of the backwash water over the en-
tire filter bed, by the economically feasible size of the filter backwash tank and pumps,
st;.
and the cost of the waste-wash handling facilities. In general, the practical maximum
size of an individual filter bed is approximately 1500 ft' (150 m'). A minimum of four
'":2'"
,Q

filters should be provided for medium-sized to large plants, those processing over 15
mgd. In contrast, very small plants processing a maximum of 2 to 3 mgd (0.09 to 0.13
m3/s) may use as few as two filters if there are financial constraints. TI1e fewer the to-
E
-••
!l
'D
0.2
tal number of filters, the more cost-effective the plant is to construct. However, this
means that the larger plants require very large individual lilters. Refer to the design cri-
iu:
teria section of this chapter for the equation used to determine the required number of 0.1
lilters in relation to the plant capacity. If a self-backwash type of tilter is selected, the 12.5 mlh (5 gpmm')
minimum number of filters per module must be four, regardless of plant size, and prefer-
ably six. Otherwise, the system may not be capable of producing the required volume 12 14
7 8 9 10
of backwash water (to wash one filter) during periods of low plant flow. However, these 3 4 5 6
2
Hours of filter run
numbers may be disregarded if either an auxiliary backwash pump or a tank is pro-
F MWD Southern
vided to supplement the inadequate production of wash water. .d· versuS hours of filter run. ( rom
F ·
Igure 3.2.7. 2 F,·ltered water turbt tty
It is important to remember that if a plant is designed with a small number of
California, Internal Report, 1978.)
filters, the filtration rate in the remaining filters increases substantially whenever
 210 DESIGN OF BASIC TREATMENT PROCESS UNITS
Conditions: Raw water: Turbidity:= 1.9-2.1 lU, Alkalinity = 120 mg/L. pH = B.l. W.T .... 130C
Treatment process: Direct filtration (pilot filter with dual media bed)
Coagulants: Alum = 2.5 mg/l, CatFloc-T = 0.35 mgll
2.5
2.0
0.5C._ _ ~ maintained at less than 0.25 ntu. Regardless of the conditions, filter washing should
Hours 01 filter run length
Figure 3.2.7-3 Headloss versus filter run at four different filtration rates. (From MWD
SOllthern California, Internal Report, 1978.)
as the filtration rate rises above 4 gpm/ft 2 (10 m/h). This phenomenon is more fre-
quent during periods of cold weather and is more apparent for rapid sand filters.
Consequently, a designed tiltration rate of 6 gpm/ft' (IS m/h) is generally recom-
mended for ordinary high-rate tilters and 3 gpm/ft2 (7.5 m/h) for rapid sand filters.
A tiltration rate of 10 gpm/ft' (25 mlh) is reasonable for coarse deep bed high-rate
tilters if a small amount of polymer is applied as a filter aid. However, the design
engineer should discuss the filtration rate chosen with the local regulatory agency
prior to design.
Ordinary rapid sand tilter beds generally have a filtration rate of 2 gpmlft'
(5 m/h), which translates to a Reynolds number of approximately 1.0; this number
is well within the limits of a laminar type of Ilow through a granular bed. The typ-
ical high-rate tilter with a standard doal-media filter bed has a filtration rate of
6 gpm/ft' (IS m/h) and a Reynolds number of approximately 10, which is still within
the upper limits of a laminar now characteristic. Both coarse deep bed and first-
stage roughing tilters of a two-stage tiltration process have filtration rates of
15 gpm/J"t' (37.5 m/h). The Reynolds numbers for these two types of unit are 30 to
40; these tigures border on the laminar tlow regime (far end) and the turbulent type
of tlow (beginning). This transitional type of tlow may explain why high-rate filter
beds have remarkable flocculation characteristics, in addition to the ability to retain
a majority of the tlocculated suspended matter when a small amount of coagulant
or tlocculant (polymer) is added to the filter inlluen!. These characteristics are gen-
erally attributed to semilaminar types of now.
BASIC APPROACH 211
The available head loss for tiltration is generally designed to be 8 to 10ft (2.5 to
3 m) for gravity tilters but only I ft (0.3 m) for automatic backwash tilters; the lat-
ter are predesigned by equipment manufacturers. Pressure filters usually provide a
terminal headloss of over 30 ft (9.3 m), but their application is not practical for large
treatment plants because of the large number of units required.
The filtration rate and terminal head loss for a particular type of tilter and tilter
medium design should be selected after thoroughly analyzing the economic trade-
offs: the total required area of the lilter bed, the available hydraulic loss across the
tilter process, the anticipated terminal head loss prior to. turbidity breakthrough In
the tilter bed, and the length of the tilter run. All these tactors must fall WIthIn the
limits dictated by the quality of the emuent water. It is important to mention that
the eftluent turbidity of gravity filters often begins to increase when the net head-
loss is over 6 ft (1.8 m). Consequently, the lilters must be backwashed, despite an
available headloss of 8 ft (2.4 m) or more, if the turbidity of the tiltrate is to be
be initiated automatically or manually by any of the following: turbidity break-
through, tilter run time, or terminal headloss.
CONTROL OF THE FILTER FLOW RATE The two basic schemes for controlling tilter
flow rate are constant- and declining-rate filtration. In recent years, the most popu-
lar scheme is the constant-rate filtration system since it provides better operational
control over the filters, has proven performance, and is preferred overwhelmingly
by plant operators. Declining-rate filtration has some me.rits but lacks positive til-
tration control measures. Thus, owners of modern nitratIOn plants seldom choose
this type of system. .
A typical scheme for the constant-tlow type of control, developed In reccnt years,
is equal-loading self-backwash filtration (see Figure 3.2.7-1). ThIS deSIgn scheme
should be evaluated for any situation because of its simplicity, efticient use of en-
ergy, uncomplicated instrumentation, simple tilter control system, and simple oper-
ation and maintenance. This system also eliminates the need for backwash headers,
valves, pumps, and an elevated wash-water storage tank.
DEPTH SIZE, AND COMPOSITION OF THE FILTER BED Filter etliciency is determined by
such t~ctors as cel1ain physical characteristics of the tilter bcd, the conditions of the
bed, and the etTectiveness of coagulation pretreatment. The design engineer only has
control over the physical parameters of the tilter bed: media size and shape, bed poros-
ity, total surface area of the filter media grains, and bed depth/media gram size ratIo.
Once the size of the media and the depth of the bed are selected, both the bed poros-
ity and the total surface area of the media grains becomes tixed. Thus, selection of the
appropriate size and depth of media are of crucial imp0l1ance to the .desig~ eng~neer.
Three basic methods arc used to select the proper depth and size of medIa: on
the basis of existing data obtained from tilters treating the same or a similar type
of raw water, on the basis of data obtained from pilot-plant studies, and on the ac-
cepted ratio between the bed depth and media size. The tirst option is often used
because it is a foolproof method. However, it is both crude and unscientific. The
 212 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 213

1.2
second approach is tempting but requires a minimum of 6 months to obtain reliable 30
data based on the different raw water qualities exhibited during both warm and cold II
It S.G.I.~/
weather. Consequently, a significant sum of money (at least $Hx),OOO) must be bUd-
II
geted to complete the pilot study. Moreover, a pilot study may not always be feasi- II / 1.0
ble for various reasons: a limited schedule for completing the project, unavailabil- 25 1/ / /
1/ ./
-<~/
ity of fresh raw water because of the remoteness of the project site, and incomplete II
water sources due to factors such as the drilling of wells or the construction of a ,,"1 /v.
""'!It"': .!!
new reservoir (dam).
20 "q ~
./"'?/
/ ./
,/ 0.8 ~
The third method is based on the relationship between the established size of the N c:t, '"
VJ, ./ ./. ./ :I:
a:
media grain and the filter bed. This relationship has been studied since the mid-
twentieth century by two or three researchers using pilot filter columns in conjunc-
II<

~ ,/<f1 ~
0/·
,,;;/./ ...,y,/
./ 0-
tion with actual filter performances. This method is considered to be the proven
technique in selecting the appropriate depth and size of the media, and a detailed ~
fii
15 , ",~" /

'/'. ..'/
(,. ' /
,O'' '.' l. \'""'/ 0.6
1
.!!
discussion is presented in the section covering design criteria. The relationship be- ~ ;S' J><1. ./ ."'l l!
~u of ,i>~ ' / / :f:~ ~
~
tween the size and the depth of the filter media implies that the available surface
.,~. ;;- 0.4 ~u
..~ ..
~
III ~
area of the filter bed media grains is a significant factor in selecting the type of fil- 10 ./ ~

(,~ .. y/./
III
ter bed. Ordinary high-rate filters have a porosity ratio of 0.45, a sphericity of 0.8,
and a bed depth/effective media size ratio of 1000. Thus the total surface area of . /"'./
the filter media grains per unit area is approximately 3000 ft' (3000 m') in area per .~/ 0.2
~/
square foot (square meter of bed area). When a dual- or multimedia filter bed is
specified, the size of the medium grains and the specific gravity of each layer be-
5
-,;../
come extremely important with respect to filter backwashing and the potential for ~/'
media loss. If an inappropriate combination of media is specified, only a Iraction
of the filter bed will be cleaned adequately. The remaining part of the bed will be 00 0.5 1.0 1.5 Z.O 2.5
either compacted and dirty or washed away at a particular backwash rate. To ensure 60~ Weight grain size (mm) (effective size x uniformity coefficient)
that the two Ii Iter media particles with sizes d l and d, and densities PI and p, have
Figure 3.2.7-4 Appropriate filter backwash rate at water temperature of 200C (68°F).
the same settling velocity, the following relationship should exist:

!!.L = (£c..ll)O.667
d, PI - P
where P is the density of water.
Figure 3.2.7-4 will aid engineers in selecting the proper combination of common
filter materials at a particular backwash rate. Note that Figure 3.2.7-4 does not ap-
ply to a sequential system using air-scouring backwash. The ligure is applicable to
dual- or trimedia beds with air-scouring wash because the bed must be fluidized in
order for it to become restratified. Figure 3.2.7-5 compares the performances of five
different types of filter bed.
FILTER WASH SYSTEMS AND WASH TROUGHS There are three basic types of filter wash
systems: (I) fluidized-bed backwash with or without surface wash, (2) sequential
air-scouring wash, and (3) concurrent air-scouring wash. Most European filters use
sequential air-scouring wash; Americans use concurrent wash. The effectiveness of
both types of air-scouring system is about the same, but the application of sequen-
tial wash is generally limited to monosand beds that are less than 3.5 ft (I m) deep.
The sequential air-scouring wash begins with 3 cfm/ft' (0.9 m/min) of air scouring
without backwash for 4 to 5 min, followed by backwash (rinse wash) for 10 to 15
min at a rate of 8 to 12 gpm/ft' (20 to 30 m/h), depending on the media size. The
concurrent air-scouring system starts with 6 to 8 gpm/ft' (15 to 20 mth) of slow
backwash together with approximately 3 cfm/ft' (0.9 m/h) of compressed air for 4
to 6 min, followed by 18 to 22 gpm/ft' (45 to 55 mlh) backwash to purge air bub-
bles trapped in the filter bed and to also restratify the dual- or trimedia filter bed.
This type of system has been used extensively in the United States.
The first type, fluidized-bed backwash with or without surface wash, has been used
extensively in the United States, South America, and Japan. The backwash troughs
are arranged 6 to 8 ft apart (1.8 to 2.4 m) and 2.5 to 4 ft (0.8 to 1.2 m) above the fil-
ter bed. The second type is practiced primarily in Europe and typically does not pro-
vide wash troughs above the filter bed. Rather, a single overflow side wall is provided
for the discharge of the backwash waste into a channel, and the top of the sidewall is
usually about 20 in. (0.5 m) above the filter bed. The third type provides backwash
troughs similar to the first but often with one or two fewer troughs. The system of
continuous concurrent water and air washing (air scouring) has very limited appJica-
 214 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 215

fill., M~la Sp<'Cm~3I1oJlS

Effective Depth Uniform;ly

Name Mal~r;,1 Sil<' (in.) C"",ffidcnl UD
A Smull dual Amhr~cit .. '0 10 1.45 [()lb
modi, Sand 05

""'" "
B-Imumediate Anthracite 1.48 U W~j
dual media Sand 0.75 U
C-Large dual Anthracite 2.0
media
D-Mixed media
Sdnd
Amh" ..,...
'.0 '"
211 "
1.25
1016

Sand
Gam...
'0
OA2
0.25
'"• US
U ''''
) I.n
E-Monome- Anthracile
dium
'0
" "
1016

Nor" ub corre'p"nd. to the ratiu of 1h. deplh of Ihe m(dium (Ll to the dfe<:li,"c size or the
medium (D).

".ooo,-;:-;;;;--;;:;;;;--~::-;;:;- _______________--,
11.97°11.340 12.420 13,600
Unit filter 10.000 8.640
run volum,
2
(&iIltt *run) 5,000

(A)
0.';,------________________---,
BACKWASH MOOE OPERATION MOOE
Filter
efftul!1lt
t't~~~IY 0.2 - - - ___________________ _ 0.26 Figure 3.2.7-6 Special baffles for concurrent air and water-wash troughs. Schematic dia-
"" gram shows the operation of the Multi wash filter process. During backwash (left) water and
air are simultaneously passed through the media (A). The dirty backwash water is collected
by the backwash trough (8) after air is separated from the water stream by the exclusive
'.000',------------------0::- Multiwash baffles (C). After backwash the filter is placed back into normal service (right).
F,lttf
effluent (Courtesy of General Filter Company.)
part,tleSlmL 500
(t.S-60,.m)

9S1)!T11112
FeCIJdose~ \.5mgll.
Fille, a,d dose = 0.02 mgll.
Figure 3.2.7-5 Results of filtration rate and medium comparison tests. (Adapted from
Elekutria WTP pilot study report by 1.M. Montgomery, Consulting Engineers.)
tion. This method applies air and water during the entire wash period, the backwash
rate is high enough to fluidize the bed, and special baffled plates are provided on ei-
ther side of each trough to minimize the loss of media during air scouring (Figure
3.2.7-6). This air-scouring wash system is suited for washing tertiary filters of sewage
treatment plants that filter strong biological floc. If a fluidized-bed backwash system
IS selected, the spacing and height of the troughs must be determined carefully to en-
sure that dislodged floc is efficiently washed away and so that the process of back-
wash and surface wash does not flush Ii Iter media over the sides of the troughs.
According to the Ten States Standards (1968), the distance Irom the top of the
trough to the surface of a rapid sand filter bed should be equivalent to the rate of
backwash expressed in inches per minute. The rules also specify that the troughs
should be spaced in such a manner that the floc does not travel more than 3 ft (hor-
izontally) before reaching a trough. These criteria are sufficient when designing rapid
sand filter beds. If the troughs are designed according to these standards, there will
be excessive loss of the anthracite coal layers. Therefore, for dual-media filters, the
top of the troughs should generally be located 3.5 to 4 ft (1.1 to 1.2 m) above the
surface of the filter bed and should be spaced 6 to 10 ft (1.8 to 3 m) apart.
There are basically two types of wash troughs: those with a shallow and wide
cross section and those whose cross section is both narrow and deep with a V-shaped
or slightly V-shaped bottom. Wash troughs with a wide cross section tend to have
higher upflow velocities whenever the backwash flow exceeds the elevation of the
trough bottom; thus the suspended solids are quickly removed. Troughs that are both
narrow and deep may be constructed with thinner walls, due to a higher moment of
inertia and greater structural integrity. However, the bottom of the wash troughs should
not be flat (for both types) because froth and sludge tend to accumulate beneath the
trough. This accumulated mud then falls back onto the filter bed and forms mud balls.
Engineers must design wash troughs so that they will accommodate the maxi-
mum expected wash rate for both backwash and surface wash, in addition to a free-
fall into the main collection gullet at the lower end of 4 to 6 in. (10 to 25 cm). The

216 DESIGN OF BASIC TREATMENT PROCESS UNITS
bottom of the trough (interior) may be either horizontal or sloping, A horizontal bot-
tom is much simpler to construct and install, The width of the filter cell should be
less than 20 ft (6 m) so that olf-the-shelf types such as plastic troughs reinforced
with fiberglass may be used, The required height for a trough with level inverts and
a rectangular cross section may be computed by using the following formula:
Q )0,667
h = --
° ( 2.49B
where Q is the total rate of discharge per trough (ft 3/s), B the interior width of the
trough (ft), and ho the depth (ft) of the process water at the upstream portion of the
trough, A freeboard of at least 3 in, (7,5 cm) should be added to ho, If metric units
are used, the formula above becomes
_ ( Q )0.667
ho - - -
1.38B
where Q is in m3/s and 110 and B are in meters.
The distance between the lip of the trough and the top of the bed (conventional
Ii Iter bed) should be
(0,5L + D) < H o < (L + D)
Spacing of the troughs from center to center is generally determined by
1,5Ho < S < 2,5Ho
where L is the depth of the filter bed media, D the depth of the trough, H o the dis-
tance between the lip of the trough and the top of the bed, and S the spacing dis-
tance of the troughs from center to center.
Unlike the fluidized-bed backwash system, the European standard air-scouring
wash does not use a series of wash troughs but has a single V-shaped trough along
the sidewall opposite the wash-waste overtlow wall, which provides the surface
sweeping function (Figure 3,2,7-7), If the wash waste is to be removed effectively
without the use of wash troughs, there is a limitation to the width of each filter cell:
\3 to 15ft (4 to 4,5 m) is considered to be the maximum width, The design engi-
neer should also be aware that filter media migration has frequently been reported
in cases where the filters lack the regular wash troughs; in this case, media migra-
tion is caused by the momentum of the lilter intluent.
Since the 1960s, many old rapid sand filters have been modi lied to become high-
rate lilters by converting the sand bed to a dual-media bed, A common problem
caused by this change is excessive loss of anthracite coal medium under regular fil-
ter washing practices, A pilot study guided by the author found that a simple baf-
tle configuration on both sides of each wash trough can reduce media loss by as
much as 70% (Kawamura et aI., 1997),
BASIC APPROACH 217
Figure 3.2.7-7 Air-scouring wash filter. (From Degremont, Water Treatmellf Handbook,
Halsted Press, New York, 1979,)
FILTER UNDERDRAINS The selection of the appropriate type of filter underdr~in
nust be based on whether or not air-scouring wash is chosen as the filter washmg
~ystem. The underdrain system used in air-scouring wash is ~esigned specifically
for this purpose, It is therefore prudent to use one of the proprIetary types, one that
has a long history of success, instead of a custom-designed underdram; thiS will
ensure that the requirements of the system are met without failure, Figures 3,2,7-
8 and 3,2,7-9 depict a few commercially available types of underdrain: false bot-
toms with long leg strainers and fiberglass-reinforced plastic blocks with channels
(interior), , '
The strainer system is the predominant type of mr-scourmg wash filter under-
drain because of its long history of effectiveness and dependability, It is extremely
important to select strainers that are physically strong, Furthermore, they preferably
have a mechanism for adjusting the stem height, even if this feature reqUires the ex-
penditure of more money. Strainers that are physically weak or brittle tend to br~ak
during construction or filter backwashing, resulting in costly and length~ repm.rs.
Even when the bottom slab of the tilter underdrain is not level, due to ImpreCise
construction or uneven settlement of tilter structures, certain types of strainer allow
the height of the strainer legs to be adjusted by approximately; in, (20 mm), Another
important feature of the strainer type of underdrain is that it has a false floor on
which the strainer nozzles are mounted.
Engineers are advised to design the air-scouring un~erd~'ain as a m~nolithic re-
inforced concrete floor since this will reduce the potenlial tor underdraw upset. As
described previously, a signilicant number of tilters have had problems with large
amounts of air emanating from the structural joints during air scounng, and their
floors have consequently become damaged. This was especially true in cases where
precast concrete blocks were secured with anchor bolts and the area between the
blocks was cement grouted, The plenum should have a minimum height of 2 ft (0,6
m) to facilitate inspection of the nozzle legs and the condition of the false bottom,
This height also allows the plenum to be constructed easily, ,
Finally, the design engineer must give special consideration ~o the size .of the
strainer head slits, If these slits are less than 0, I in, (0,25 mm) WIde, they WIll eas-
ily become clogged by line media grains and other debris that may be left in the
backwash line or plenum during construction. The engineer is therefore adVised to
 BASIC APPROACH 219

·':>;1 :'"fi ::~~J~: :f~~ill~~'

Iyp. anchors

Stainless
steel
Insert
3000 PSI 7-sack concrete
G·day water cured

Precast Concrete Perlorated Lateral System

Manifold and PerfO(ated lateral System

'".
I In. ,lin, 1 1M
'V" rn. or
/Yl~ in. diam. holes r-it-lt 'n.~t-tt in-l~

............. , ..
··................
· .1............ .
..............
.~

·................
.............. .
Ptan of Filter Block

Cross Section of Block

CombtnahOn 01
block lengths to
surt filter size
...... 2 ,n II,er ~f. "'.-no 10 mt,h
~_2 '" "'ttl'" In _',\ In
".- ~2 In. Ilyfl :r.
In _'I. In
2 ,n 11~tf v. "I. __I: In

Filter effluent
.. ;': ..:~ and back wash -. '. .
.!",:< (-: supply channel ;:;:~~;:..:•• ":•.•,,..~-':;..:.• ,,:~.;::;"~•..,,.. ~""c:.:;..:."-.:.:.:c:.....:;.·~.:.::,::·W:·
.' ...:::' :;; .•.•. ;:'.•..
leopold System Filler medIum Figure 3.2.7-9 Filter underdrains for air scouring: (a) Leopold's universal underdrain block

16'~,~O~,~'=··~'~'I~
(courtesy ofEB. Leopold Company. BulletinASY·lOO); (b) PCl's K·Floor with nozzles (cour·
"I~,n IOJ/.,n
-+--
~ ~ tesy of Paterson Candy International).
Wheeler System ~t;
J"
select the largest possible width. The appropriate slit size is half the effective size
(3) 3 in. spheres
(2) 11/. m. sphere of the filter medium covering the strainer heads. Thus, if the effective size of the
(1) 1 Va in, spheres ~~:;":'rIi!:""--'';';:.ll:::''';! sand is I mm, the size of the slits should be 0.5 mm. Manufacturers often recom-
PorcelaIn mend covering the strainers with 6 in. (150 mm) of pea-sized gravel because this
Ihlmble
practice minimizes clogging of the strainers as well as deformation of the slits by
Fi~u~ 3.2.7-8 Filter underdrains with non-air scouring. (Adapted from Wi t 71· . the scrubbing action of the filler media. and distributes the backwash flow more ef-
Prl1lclples alld D . b JM M a er realment.
eSlgll y . . ontgomery, Consulting Engineers, Wiley, New York. 1985.)
ficiently, but only for certain types of strainer nozzle.
218
 220 DESIGN OF BASIC TREATMENT PROCESS UNITS
BASIC APPROACH 221

The plastic block type of underdrain of the air-scouring wash system has been
used successfully since 1980. This system requires a specifically graded gravel bed
with an approximate depth of 12 to 15 in. (30 to 38 em), to be located directly abov~
the underdrain because the size of the orifice is too large to retain the tilter media.
However. plant operators must strictly follow the manufacturer's instructions for
both the wash rates and the sequence of the air and water. Otherwise, the gravel bed
will be upset and the expensive process of reconstructing the filter gravel bed will
be required.
. Recently, a few manufacturers have introduced a porous plate in lieu of a specif-
Ically graded gravel bed. The porous plates attach directly to the top of the plastic
underdrain blocks. This type not only eliminates mounding and disturbance of the
gravel bed but also gains 12 to 18 in. (0.3 to 0.45 m) of extra Ii Iter bed depth due
to the elimination of the gravel bed. However, potential clogging of the porous plate
by scale, especially by calcium carbonate scale, is a concern.
A number of alternative types of underdrain systems are available for the regu-
lar backwash system: Precast concrete laterals, pipe laterals, and Leopold's tile block
system (Figure 3.2.7-8) are proven types that are used extensively. Unlike the air-
scou:Ing wash, these types of underdrain are not limited to the proprietary types,
provIded that they are properly designed. Moreover, any reputable contractor can
fabricate most of these underdrains on the job site. The design of the lilter under-
drain system is based exclusively on the hydraulics of the filter backwash. The de-
tails are discussed elsewhere.
AUXILIARY SCOURING OF THE FILTER WASH Auxiliary scouring is absolutely essential
during filter backwash whenever coagulants are used in the pretreatment process, es-
pecially if polymer is used as a filter aid. Backwash alone cannot maintain the Ii Iter
beds in a reasonably clean condition for longer than several months, regardless of
adjustments in the backwash rate and control of the bed expansion rate,
Regular surface wash systems are capable of adequately cleaning ordinary rapid
sand and dual-media filters. The advantages of this type of system are its simplic-
Ity, effectiveness, and easy operation and maintenance, Unlike air-scouring wash, it
does not require a delicate auxiliary scouring and backwash sequence, nor does it
require the backwash rates to be adjusted during a wash cycle. The air-scouring wash
The second and third types are designed specifically for dual- or multimedia lilter
beds. In the United States, only one manufacturer has had over 20 years of success-
ful operation for the three rotating arm systems. The mechanical rake system has
been tried, but its bulky mechanism, the limited freedom in design, and its frequent
maintenance requirements have made this system unpopular. Figure 3.2.7-10 illus-
trates the effectiveness of surface wash.
AIR-SCOURING FILTER WASH In recent years, some pub1ications and equipment man-
ufacturers have given readers the impression that the air-scouring tilter wash system
is the only effective means of maintaining clean filter beds for any filter type. This
is, however, not always true. The air-scouring wash vigorously agitates only the top
6 to lOin. (0.15 to 0.25 m) of the filter bed. This boiling phenomenon misleads plant
operators and visitors to believe that the violent agitation occurs throughout the fil-
ter bed. However, some mud balls that have fallen below the boiling zone remain un-
broken. When utilizing strainer nozzle underdrains it is important to know that com-
pressed air emanating from the bottom strainer nozzles rise directly above each nozzle
and does not spread sideways. Bed media, located in between the uprising streams,
is pulled down (mass balance law). This behavior is more distinct when a slow back-
wash is applied with compressed air (Figure 3.2.7-11). Both pilot filter studies and
core sampling of actual filters using air-scouring wash confirm this phenomenon. and
there is a slight increase in sludge deposition at the bottom of the bed.
In a worst-case scenario, oocysts of Cryptosporidium deposited on top of the fil-
ter bed are dragged downward, breaking through the filter bed. Therefore, it is worth-
while to consider the use of a surface wash system in conjunction with an air-scour-
ing system that has strainer nozzles for the underdrain. This scheme will break up
the top several inches (20 cm) of compacted media before air scouring when sur-
3000 4000
E
00
~:-~-- • ____--_x
~

:8 20
10
"
} x---
, ----
,'>
system should be employed for any deep bed filters with coarse media and may be !30
0
{ I .'
the preferred system for ordinary Ii Iter beds processing biologically pretreated -40 \ I \

~ ..... ..
sewage.
50 \ I ,\
The two basic types of surface wash systems are the lixed grid and rotating arms.
Both types have been used since the mid-twentieth century and have been proven
effectIVe when properly designed, fabricated, installed, and operated. The lixed grid
.c
~60
70
( , I
I

I
I
type of surface wash may be designed by the engineer and requires much less main- x- _ _ Before filter wash
tenance because of the lack of moving parts, A rotating arm system should neither
be designed by engineers nor fabricated in the machine shop because of its sensi-
0---0 After filter wash II

tive design and specifications.

The three types of rotating agitators are surface agitators, subsurface agitators, and
Prachin Suri WTP
(with surface wash)
I Bang Moon Nak WTP
(no surface wash)
dual-arm agitators. The tirst is used primarily in conjunction with rapid sand filters, Figure 3.2.7-10 Sludge profile in filter beds. Amount of sludge is shown as turbidity (ntu).
 222 DESIGN OF BASIC TREATMENT PROCESS UNITS
BASIC APPROACH 223

any air-scouring wash system. The required height of the overflow pipe is normally
6 Ii (1.8 m) above the top slab of tilter structure; 2 to 3 Ii above the hydraulic grade
line during backwash.
It is the author's opinion that any filter bed with a depth less than 30 in.
Boil
(0.75 m) does not require an air-scouring wash system because surface wash (fixed
Bed level nozzle type) can provide adequate cleaning. A surface wash system designed by
John Baylis and installed at South District Water Treatment Plant [500 mgd (22
m'/s)] in the 1950s has kept all of the filter bed very clean after almost 50 years of
operation. However, when the depth of tilter bed exceeds 3 ft (I m), installation of
an air-scouring wash system is a proper choice.
The important issues for selecting an air-scouring wash systems design are (l) a
long and proven history of satisfactory performance, (2) an adequate anchor sys-
tem, (3) manufactured by a company with good quality control, and (4) strictly su-
pervised construction of the underdrain system by an experienced construction su-
pervisor throughout the duration of the job. If a plenum type of underdrain system
is selected, it should have a monolithically cast reinforced concrete false floor. Avoid
the use of precast elements that must be bolted onto pedestals, as this will present
Ascending Air
With the problem of massive leakage of air around the blocks due to the failure of the
Mushroom cement grout after over 10 years of use.
Shape
APPURTENANT SYSTEMS The major filter appurtenant systems are the filter back-
wash tank, pumps, auxiliary scouring systems, wash flow control system, and wash-
waste handling systems. Since the last item is a subject in itself, it is discussed sep-
arately in Section 4.2. The Iilter backwash tank must have a large enough capacity
to hold a sufficient amount of water for two filter washes. The tank must also pro-
vide an adequate depth of water above the tank exit so that a vortex will not be
formed toward the exit pipe. Formation of a vortex leads to the entrainment of air
_--,,~Air
in the wash water, which in turn reduces the flow rate and hinders the functioning
of the flow rate controller. Therefore, it is always a good practice to provide an an-
tivortex assembly in the wash tank at the entrance of the wash main.
The lowest water level in the backwash tank should generally be about 40 ft
(12 m) above the top of the wash-water troughs located in the Iilter bank. Under nor-
mal conditions, the fixed grid type of surt~lce wash can obtain water from the back-
Figure 3.2.7-11 Fil!er bed movements under air-scouring wash. wash tank. In contrast, the rotating arm surface wash system requires a minimum of
80 psi (5.5 kg/cm' or 540 kPa) and therefore usually requires a booster pump system.
face The source of the filter backwash may be one of the following: an elevated wash
wash is applied first, and slow backwash can keep suspended solids above the
bcd at all tImes. tank, the pressurized main water distribution line, the water from the direct pump
of the clearwell, or the effluent from the rest of the IiIters on the line (as in the self-
. Ch~osing an air-scouring system complicates not only the design and construc-
backwash filters). The choice among these alternatives depends on factors such as
tion of the filters, but also filter operation: air blowers, air piping, and air and back-
site topography, cost considerations, and aesthetics. The elevated tank is the most
~aSh sequence .control. FIlter washing time is also affected, twice to three times as
common source of backwash. Depending on the topography of the site, the tank is
o~g ~ t~e surface wash system. Furthermore, there have been several catastrophic either on stilts or situated on a hill and may also be installed over the control build-
~n. e~ ral~ upsets due to the compressive and explosive nature of compressed air
ing. The tank may also be filled by a smali'capacity pump between backwashes .
. t 1St ere ore strongly recommended that a standpipe no smaller than 12 in (03 m)
An alternative to the elevated tank system is the direct pump backwash. In gen-
In diameter be placed in the backwash pipe main as a pressure relief me~su;; for
eral, this system is rather expensive to operate due to its high electric consumption.
 224 DESIGN OF BASIC TREATMENT PROCESS UNITS
In fact, some plants cannot operate the filter backwash pumps during periods of
peak power demand. Therefore, the direct pump wash system must be accompanied
by a standby power generator. It is also advisable to place a pressure relief stand-
pipe in the pump discharge line to avoid filter underdrain failure, which may result
from excessive pr~ssure. The power demand of the direct pump wash system may
be reduced by deslgnrng the filter to backwash only half a filter cell at a time. This
allows the plant to use a wash pump that is only half the usual size.
Air-scouring backwash filters are often designed with a pump backwash system
for two reasons: the backwash rate is half that of an ordinary fluidized backwash
system, and the wash rate is a two-step system. Two-step backwash may be achieved
by using one pump for the first wash and a second pump during the second step
rather t~an a," elaborate flow rate control system. Air-scouring wash is attained by
emploYing air blowers. The most common type of blower is the positive-displace_
ment blower (Roots type), despite its high level of operational noise.
The backwash flow control system is necessary for setting the proper washing
rate regardless of the type of auxiliary scouring employed during backwash. The
two most common systems are the mechanical flow rate controller and the master
!low control valve located in the backwash main. The mechanical !low rate control
system is a combination of a flow meter and a !low-modulating butterfly valve. The
second system is generally a plug or ball valve which is opened in a slow and con-
trolled manner to purge air out of the filter bed at the beginning of a wash and to
increase the wash rate slowly to a maximum (preset) rate; the preset rate is can.
trolled by a stopper located at the valve operator. Use of the master flow control
valve is recommended for most cases. However, the mechanical flow rate controller
~hould be employed if the water level in the tank varies over 10 ft (3 m) during a
hlter wash or If the backwash pipeline is both large in size and short in length.
Otherwise, an insufficiently controlled headloss and a large variation in the back-
wash rate (during the washing cycle) will result. It is important to note that the back-
wash valve should be opened slowly to avoid disturbing the gravel bed (filter) and
the underdrain system. If a self-backwash filter design is adopted, the filter wash-
ing system may be simplified dramatically since it does not require a pump system,
a backwash flow control system, a backwash line, or valves.
FlLTER-TO-WASTE The purpose of the filter-ta-waste system is to divert inferior-
quality filtered water to the waste line. Inferior filtered water may be generated un-
der both normal and abnormal conditions. Under normal conditions, turbidity break-
through IS often observed at operational treatment plants whenever there is
inadequate pretreatment andlor when the filter beds are in unripened condition.
Inferior filter water may be produced when the following abnormal plant condi-
tions exist: (I) immediately after the installation of new filter media due to the ex-
istence of an unripened bed (Le., the bed is not properly conditioned to effectively
capture suspended solids by means of a small amount of polymer fed as filtration
aid or the deposition of appropriately charged suspended solids and the correct quan-
tity of floc on the filter media), (2) after disturbance of the filter bed due to repair
or modification of the gravel bed or underdrain system, (3) accidental initiation of
BASIC APPROACH 225
the surface wash or air scouring during the filtration cycle, (4) liquefaction of the
filter beds resulting from strong earthquakes, and (5) administration of incorrect co-
agulant dosage. Although virtually all existing filters are equipped with a. drain pipe,
this pipe is too small to divert normal filtered water flow to the waste hne. .
In recent years regulatory agencies, including the U.S. EnVironmental ProtectIOn
Agency (EPA), are deeply concerned about the presence of cysts and oocysts of proto-
zoans, especially Ctyptosporidium. Consequently, the goal of filtered water IS now set
at 0.1 ntu. The Department of Health of some states, including California, set more strin-
ent guidelines by advocating the wasting of filtered water with turbidity greater than
~.25 ntu. The California Department of Health Services (DOH) also endorses slow start-
ing of tilters after filter washing in order to circumvent initial turbidity breakthrough.
It is important to note that EPA and DOH of a handful of states allow the addi-
tion of coagulant or polymer to the backwash water as an alternattve to filter-to-
waste. This practice conditions the filter bed prior to starting the filtration cycle.
Unfortunately, it is not a robust method and therefore not always successful, a fact
supported by both experienced pilot filter operators and engineers: A fe.w treatment
plants also report problems with floc formation and sludge depOSitIOn 10 the dear-
wells when using this alternative. Yet despite the drawbacks, both deSign englOeers
and plant managers often choose to add polymer as coagulant because it is a cheaper
method of meeting regulatory requirements.
It is possible to add polymer at the influent of each filter at the beginning of the
filtration cycle rather than the backwash water. This will help to effecUvely minI-
mize initial turbidity breakthrough with a simple setup. The disadvantage is the high
number of application points. If a filter requires an influent turbidity of less than I
ntu in order for it to produce good filtered water consistently, it is important to in-
vestigate whether the filter is suboptimally functioning due to insufficient coagula-
tion of suspended solids in the raw water andlor inadequately conditioned filter beds.
It is the author's opinion that each filter should be fitted with a filter-to-waste
piping system to manage initial turbidity breakthrough under both normal and ab-
normal plant operations. Although there are several methods to Implement filter-to-
waste, the system depicted in Figure 3.2.7-12 should be considered.
Basic Hydraulics One of the most important items of filter design as well as
operation is the hydraulics through the filter structure. Design engineers must know
the headloss across the filter bed, filter bed expansion rate during backwash, and a
common flow velocity for filter pipings and channels. This section will aid the en-
gineer to determine these items.
Initial Loss of Head Loss of head at the beginning of the filtration cycle is called
initial headloss: the resistance created by granular media to the passage of water.
The initial headloss through a filter bed at various filtration rates under different wa-
ter temperatures can be estimated by Darcy's equation:
M= LQ (Darcy)
kA
 BASIC APPROACH 227
226 DESIGN OF BASIC TREATMENT PROCESS UNITS

O.3-m air gap above the water sulfa Q = /low rate (em'/s)
(no cross connection) c~r 2
A = area of the bed (em )

I Influent Channel
'I k = permeability eoeftieient (em/s)

c = coefficient
i i , <-~ = water temperature (0C)

de = effective size of filter medium

~ To washwater holding tank

Filters *~articulate
mpling tap for
counter

kS Turbidimeter
NOTE: For most liltcr media sizes, e is 124 and varies only from 123 to 125 for
effective size (de) of 0.3 to 2.0 mm.

Flowmeter Example Compute the headloss for a 30-in. (76-em)-deep filter bed with 0.75 mm
j
(0.075 em) effective size medium under a 6 gpmlft2 (15 m/h) filtration rate at 68°F
7 'I
1 (20°C).

t~ tt , ::l , t1 p Filter!o waste pump(s)

Sollltion
2
Compute the headloss through a unit surface area (I em ) of the bed.

rrr , M

~
'-@ Yhli
(VFD controlled pump)
k = 124(0.7 + 0.03
= 0.906
X 20)0.075
2

ClealWell ) 2
= (15 = 0.417 em'/s
\ Filtration rate X 100) .,. 60

~ Thus

Illl = (76 X 0.417) .,. (0.906 X I)

Figure 3.2.7-12 Filter-to-waste schematic S . I f
wash-water holding tank(s)' (2) sl fiI . P~CIa eatures: (I) No additional capacity for = 34.5 em (1.13 ft)
7.5 -) 10 -) 12.5 m1h); (3) filter~:-~a~~: ~~~~mg can be achieved (3 -) 4 -) 5 gpmlft or
2

rate; (4) instrumentation/control can be sim lifie~.an be m~t~hed to. actual seasonal filtration As a quick reference, the approximate initial headloss across the common filter beds
termine the FIW time' (6) di p , (5) turbIdImeter III the recycle line can dc-
, scharge the FTW-to-waste line when its quality is questionable. at 60°F (15°C) is presented below.

Filtration Rate Headloss

and
ft m Type of Filter Bed
gpm/ft2 m/h
k = e(0.7 + 0.03/)de 2 (Hazen) 5 1.0 0.3 Regular rapid sand
2
1.45 0.45 Regular rapid sand
3 7.5
where 0.3 Standard dual media
4 10 1.0
1.45 0.45 Standard dual media
6 12.5
h = headloss (em) 1.9 0.58 Standard dual media
8 20
2.45 0.75 Standard dual media
L = depth of the Ii Iter bed (em) 10 25
 228 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 229

Filter Bed Expansion Rate and Loss of Head by Backwash The bed expan- ,- ' ' ' ecitied with a surface washing system, What is an es-
speclltc gravIty 01 1.65 IS sp .. b k ashing conditions? Assume that the
sion rate under optimum backwash rate can be estimated by the following equation, 'mated loss of head under an appropllate ac w( ,
where f is the porosity ratio of the bed: ~and bed expands 37% and a 25% expansion for the anthraCIte coal bed.
(0.1 )0.22 - f 0.6 - f Solution The headloss by the sand bed is
bed expansion rate =
I - (0.1 )0.22 0.4
h = (12 X 1.37) (2.65 - I) (I - 0.6)
In general. the porosity ratio of the sand bed is 0.43 to 0.45 and 0.48 to 0.50 for
anthracite coal. = 10.8 in. (27.4 cm)

Some treatment plants employ an alternative method to determine the appropri-

ate backwash rate: 50% lilter bed expansion rates during backwash. During the late and the head loss by the anthracite bed is
1950s Baylis and others emphasized that 50% sand expansion should be achieved
h = (24 X 1.25) (1.65 - I) (I - 0.6)
for rapid sand lilters. This requirement is justified when the sand bed is composed
of rather small (E.S. = 0.40 to 0.45 mm) grains, yet it is important to note that a = 7.8 in. (19.8 cm)
majority of tilters were not fitted with auxiliary scouring (surface wash) systems
during the late I 950s. However, even with 50% bed expansion during backwash, Thus the total headloss is approximately 18.6 in. (47 cm).
most of the rapid sand Ii Iter beds constructed in the I 950s and 1960s are extremely
dirty, Without proper surface wash or air scouring, the filter beds accumulate mud Flow Velocity The maximum flow vel,~cities (approximate) sustained by the fil-
balls after 2 to 3 years of operation. The targeted bed expansion rate for modern ter piping or channels and valves are as follows:
high-rate lilters during backwash with surface wash is approximately 37% for a typ-
ical sand bed with a porosity ratio of 0.45; the expansion rate for an anthracite coal Ordinary Filters Self-Backwash Filters
bed with a porosity ratio of 0.5 is about 25%. These figures have been verified by mls ft/s mls
ft/s
well-operated actual tiltration plants.
The depth of the expanded bed during backwash may be estimated using the Influent channellpipe 3 0.9 1.5 0.45
equation 5 1.5 5 1.5
Inlet valve
Forebay channel 0.5 0.15 0.5 0.15
Effluent valve 5 1.5 2 0.6
Emuent channel 5 1.5 2 0.6
Backwash main 8 2.4 3 0.9
where I is the bed depth at rest. I,. the depth of the expanded bed,Jthe porosity ra- Backwash valve 8 2.4 6 1.8
tio of the bed at rest, and f,. the porosity ratio of the expanded bed. When the sand Surface wash line 8 2.4 8 2.4
bed is expanded 37% during the backwash, f,. becomes approximately 0.6. Also, Wash-waste main 8 2.4 8 2.4
when an anthracite coal bed is expanded 25%, f,. of the bed is about 0.6, under ap- Wash-waste valve 6 2.4 6 1.8
propriate backwash rate for both cases (refer to Figure 3.2.7-4 for the appropriate Filter underdrain lateral 4.5 1.4 4.5 1.4
backwash rate).
Loss of head for the expanded bed can be estimated as
. I i
in the inlluent pipe or channel may be 4 Itls . (1.2 mI s) if
II NOTE. The flow ve oc t y . _ fll nt line or if the filtration system IS a
I = (Pm - 1)(1 -f,.) flow rate controllers are located In each hlter e . ue
. . If h fi It . nlet is a welf for equa
I load·,ng design (sec Figure
<
declining-rate system. . I e I er I h I hould be less than 1.5 ft/s (0.45 m/s).
3.2.7-1), the Ilow velocIty In the Inlluent c anne s
where h is the head loss and Pili is the specific gravity of the tilter medium,

~ ,
Example A dual-media filter bed comprised of 12 in. (0.3 m) of sand with an ef- Design Criteria . dcSlgn
This section covers the baSIC en'te ria for granular
. . medium
f· the
fective size of 0.5 mm, porosity of 0.45, and specific gravity of 2.65 and 24 in. . f h t e of filter the charactenstlcs a
filters: the number of filters, the sIze a eac yp th flt;s the type of lilter back-
(61 cm) of anthracite coal with an effective size of 1.0 mm, porosity of 0.5, and filter bed, the filtration rate, the head loss across e 1 e"
 230 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 231

washing system, the specifications for the gravel support bed, the basic h draul' Width of each cell 8 to 25 in. (0.2 to 0.64 m)
and a basIc filter design scheme. Y les, II to 48 in. (16 in. or 0.4 m standard)
Depth of filter bed
(0.25 to 1.2 m)
Number of Filters The minimum number of tilters for very small plants ro Traveling speed of the washing 21 in.lmin (0.53 m/min) standard and indexing
c~sslng less than 2 mgd (90 Lis) is two. If the capacity of the plant exceeds 2 ~ d- carriage available for deeper than 16 in. bed depth
t e mln.mum number of filters becomes four. Note that the required number of ~I'
ters IS calculated from the equation I-
PRESSURE FILTERS

N = 1.2Qo., Diameter of a filter cell 4 to 20 ft (10 ft or 3 m average) (1.2 to 6 m)

Length of a filter cell 8 to 50 ft (20 ft or 6 m average) (2.4 to 15 m)
where N is the total number of filters and Q is the maximum plant flow rate in d Depth of a filter bed 2 to 3 ft (0.6 to 0.9 m)
(I mgd = 3785 m'/d). mg

Size of Each Filter Th e d'.menSlons

. of four types of filter are discussed below. NOTE: Both horizontal and vertical cells are available. However, vertical cells are more
common fOf larger units because they are cleaned to a greater extent during filter waShing. The
ORDINARY GRAVITY FILTERS
multicell (four-cell) horizontal units have the following advantages: a self-backwash system.
Width of a filter cell 10 to 20 ft (16 ft or 4.9 m average) (3 to 6 m) space-saving features. and cost-effectiveness.
Lengthlwidth ratio 2:1 t04:1 (3:1 average)
Area of a tilter cell 250 to 1000 ft' (600 ft' or 56 m' average) Filter Bed: Types of Medium and Depth Silica sand and anthracite coal are the
(25 to 100 m') most commonly used types of filter medium. The American Water Works Association
Depth of the filter 15 to 25 ft (17 ft or 5.2 m average) (4.5 to 7.6 m) has set the quality standards for both these materials (refer to AWWA Standards
Bloo). Garnet, ilmenite, pumice, and synthetic materials may also be used as filter
media but to a lesser extent due to their limited availability and relative expense.
NOTES: General silica sand and anthracite coal are the most commonly used types of filter
medium. The depth of the filter bed is a function of medium size and generally fol-
I. :~; ~lters for medium-sized to large plants, those processing a minimum of 20 mgd (09 lows the relationship shown below:
S , are generally composed of two cells per filter and have a central gullet down .h·e
center.
1000 for ordinary monosand and dual-media beds
2. iohefi~at~~ de~th above the, fi.lte~ bed should be equal or deeper than the head loss provided
r I ra IOn In order to minimiZe air binding. Since 6 to 8 ft (1 8. 24 ). 1250 for regular trimedia (coal, sand, and garnet) beds
head los ~ fI . . 0 . m IS a common
s or I tratIon, a water depth of 8 ft is preferred. for most coarse deep monomedium beds
1250 to 1500
(d, is 1.2 to 1.4 mm)
SELF-BACKWASH FILTERS (GRAVITY FILTER)
1500 to 2000 for very coarse monomedium beds
Width of a filter cell 10 to 20 ft (16 ft average)
Lengthlwidth ratio (d, is 1.5 to 2.0 mm)
2: I to 4: I
Area of a filter cell 250 to 800 ft' (25 to 80 m')
Depth of the tilter where I is the depth of the filter bed (mm) and d" is the effective size of the filter
18 to 25 ft (22 ft average) (5.5 to 7.5 m)
medium (mm).

NOT~: Th~ minimum number of filters is four, but six to eight are preferred. Each filter
norma y consIsts of only one cell due to limitation in the backwash water supply. NOTES:

AUTOMATIC BACKWASH FILTERS (HARDINGE TYPE) 1. When the size of the medium exceeds 1.5 mm, the space between the filter grains be-
comes very large compared to those in regular filter beds. The void triples when the di-
Width of a filter Standard 16 ft (5 m) ameter of the grain is doubled. Therefore, the fide ratio should be used as an estimate only
Length of a filter Up to approximately 120 ft (37 m) when the medium is greater than 1.5 mm, and actual figures should be obtained through
Depth of a filter cell 7 to II ft (2.1 to 3.3 m) pilot study. Also, it is strongly recommended to place a thin [12 in. (0.3 m)) sand layer at
 BASIC APPROACH 233
232 DESIGN OF BASIC TREATMENT PROCESS UNITS

the bottom of the coarse deep mOl1omedium bed to act as a barrier to prevent small par~ BACKWASH RATE
tides such as oocysts of protozoans from breaking through. 15 to 18 gpm/f(2 (37.5 to 45 m/h)
Ordinary rapid sand bed
2. If the turbidity of the filtered water must be less than 0.1 ntll, without the use of pOly. 18 to 22 gpm/ft2 (45 to 55 m/h)
Ordinary dual-media bed (trimedia included)
mer as a filtration aid, it is recommended that the /Ide ratio be increased by 15%. Table 12 to 16 gpm/ft2 (30 to 40 m/h)
3.2.7-1 illustrates the general design criteria for common filter beds. Ordinary GAC bed
3. Design engineers should anticipate stringent drinking water standards for future plant op- SURFACE WASH RATE
eration and are therefore advised to include a provision for replacing the regular filter beds
with GAC medium. Due to the longer contact time required by the GAC bed, a minimum Fixed nozzle type
EBCT of 10 min, the depth of the regular filter bed must be made deeper to accommo- Flow rate 3 to 4 gpmlft2 (7.5 to 10m/h)
date the GAC medium. A depth of 5.5 ft (1.7 01) is needed to provide 10 min of contact Pressure 8 to 12 psi (55 to 83 kPa)
time at a loading rate of 4 gpm/ft' (10 m/h).
Rotating ann type
Single Arm
Filtration Rate As mentioned previously, regulatory agencies restrict the maxi- 0.5 to 0.7 gpm/ft2 (1.25 to 1.75 m/h)
Flow rate
mum filtration rate for a filter design. However, the agencies will consider a higher 70 to 100 psi (480 to 690 kPa)
Pressure
filtration rate if a well-managed long-term pilot study can prove that the higher rate
Dual arms
satisfies all the requirements. The proven and accepted filtration rates for common 1.3 to 1.5 gpm/f(2 (3.25 to 3.75 mlh)
Flow rate
filter beds are listed in Table 3.2.7- I. 80 to 100 psi (550 to 690 kPa)
Pressure
Engineers should realize that an acceptable filtration rate is a function of factors
such as the size of the medium, the degree of pretreatment, and whether polymer is
used as a filtration aid. Design engineers who must comply with very conservative
design criteria should include a provision for the possible increase in filtration rate.
NOTES:
Modern dual-media filter beds are capable of accommodating rates that are 6 to 8
1 The pressures shown are at the water jet discharge points. .
gpm/ft2 (15 to 20 m/h), whereas coarse deep beds may accommodate liltration rates
2 2: Place the nozzles 2 to 3.5 in. (5 to 9 cm) above the top of the bed [rotating arm type IS 2
of 8 to 13 gpmlft (20 to 30 mlh) with or without preozonation. Hence the size of
in. (5 em)].
the influent valve, effluent valves, pipes, and channels must be sclected properly
when implementing the higher flow rate.
AIR-SCOURING BACKWASH

Headloss Across the Filter

BACKWASH RATE

TOTAL HEADLOSS ACROSS EACH FILTER Air-scouring stage for ordinary beds 2 to 4 gpmlft' (5 to 10 mlh)
Air-scouring stage for coarse deep beds 8 to 10 gpm/ft2 (20 to 25 m/h)
Ordinary gravity filters 9 to 15 ft (12 ft or 3.7 m average) (2.7 to 4.5 m)
Rinse stage for ordinary beds 10 to 12 gpm/ft2 (25 to 30 m/h)
Pressure filters 50 to 100 ft (75 ft or 23 m average) (15 to 30) m)
Rinse stage for coarse deep beds 16 to 24 gpm/ft2 (40 to 60 m/h)
Automatic backwash lilters 2 to 3 ft (0.6 to 0.9 m)
AIR-SCOURING RATE
NET HEADLOSS AVAILABLE fOR FILTRATION

For ordinary filter beds 2.5 to 3 ftJ/min/ft2 (46 to 55 m3 /hr)

Ordinary gravity filters 6 to 12 ft (9 ft or 2.8 m average) (1.8 to 3.6 m)
For coarse deep beds 3 to 4 ft 3/min/ft2 (55 to 73 mJ/hr)
Pressure filters 25 to 50 ft (7.5 to 15 m)
Automatic backwash filters 0.5 to 1.0 ft (1.0 to 0.3 m)
Underdrain System Selection of the underdrain system varies with the type of
Filter Washing filter system that is chosen. Normal backwash and self-back,:,ash filters may use the
same type of underdrain system but with one important modIficatIOn-the. latter has
BACKWASH WITH SURFACE WASH The appropriate backwash rate should be determined
a larger orifice size. Air-scouring backwash filters are most compatible WIth plastic
by the specific gravity of the medium, the size of the media grains, and the water tem- blocks or a strainer type of underdrain system. Although m~ny other types of filter
underdrain systems are available, those listed here are conSidered to be the proven
perature. Figure 3.2.7-4 may be used in detennining the backwash rate. TIle effect of
water temperature on the backwash rate is significant, as shown in Table 3.2.7-2. and most effective types.
....""'"

TABLE 3.2.7-1 Types of Medium and Applications

Filter Medium Type of Filter Medium Design Criteria Advantages/Disadvantages
Fine sand Slow sand filter 0.05-0.17 gpmlft' Effective size: 0.25-0.35 1. Simple design and construction
(0.13-0.42 m/h) filtration rate U.c.: 2-3 2. Good effluent quality without pretreatment
Depth: 3.3-4 ft a. Requires a large filter bed area
(1.0-1.2 m) b. Applicable only for good quality raw water
S.G. a :=:= 2.63 c. Requires frequent scraping off of surface layer
(every 20-30 days)
Medium sand Rapid sand filters 2-3 gpm/ft 2 Effective size: 0.45-0.65 I. A proven and widely accepted tiltration process
(5-7.5 m1h) filtration rate U.c.: 1.4-1.7 2. A wide application range if pretreatment is provided
Depth: 2-2.5 ft a. Rather short filter runs due to surface filtration
(0.6--0.75 m) b. Always a need for coagulation pretreatment and an
S.O. "" 2.63 auxiliary washing system
Coarse sand High-rate filters 5-12 gpmlft 2 Effective size: 0.8-2.0 1. An effective high-rate tiltration process with very
(12.5-30 mlh) filtration rate U.C.: 1.4-1.7 long filter runs
Direct filtration Depth: 2.6--7 ft 2. A wide application range with polymer pretreatment
(0.8-2 m) a. Auxiliary wash system is limited to air-scour type
S.O. "" 2.63 b. Requires deep filter cells and a special underdrain
Multimedia High-rate filters 4-10 gpmlft' Sand 1. An effective high-rate filtration process with long
coal-sand (10-25 mIh) filtration rate Effective size: 0.45-0.65 filter runs
dual or coal- Direct or in-line filtration U.C.: 1.4-1.5 2. A proven and widely accepted filtration process
sand-gamet Depth: I ft (0.3 m) a. Either surface wash or air-scour wash is required as
trimedia Anthracite coal an auxiliary washing system and a polymer is required
Effective size: 0.9-1.4 as a filter aid

,~----------------

b. Proper selection of each medium is important

Removal of organic
Granular activated
contaminants 3-6 gpmlft
2 U.c.: 1.5-2.5
carbon (OAC)
(7.5-15 m1h) tiltration rate
Contact time: 10-15 min
Variety of types. including
Proprietary type
media green sand and synthetic media
u.c.: 1.4-1.5
c. Requires a high backwash rate for restratiflcation
Depth: 1.5 ft (0.45 m)
S.O. "" 1.5 to 1.6
Camel
Effective size: 0.25-0.3
U.C.: 1.2-1.5
Depth: 0.25 ft (0.0075 m)
S.O. "" 4.0 to 4.1
1. A proven and accepted process for specific removal
Effective size: 0.5-1.0
of organic contaminants (i.e., taste and odors, THMs,
and pesticides)
Depth: 6--12 ft
(1.8-3.6 m) 2. Can also operate effectively as a conventional tilter
a. Must be regenerated or replaced when adsorption
S.O. "" 1.35 to 1.37
capacity is depleted
b. High initial and maimenance costs
I. Design and efficiency guaranteed by the manufacturer
Depends on the purpose
a. Limited number of suppliers
b. Mostly patented items

"s.c .. specific gravity.

'"
'"""
 236 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 237

TABLE 3.2.7.2 Backwash Rate Adjustment for Various Water 'IIemperatures lateral length of 16 ft
Wash-Water (4.9 m)
Water 12 gpm/f(' 15 gpmln' 18 gpmln' 21 gpmlft'
Temperature Viscosity 2t06ft ~~to!in. 12-in. lateral spacing
(30 mlh) (37.5 m/h) (45 mlh) (52.5 m/h) Dual-parallel
CC) (cP)" (0.6 to 1.8 m) (4 to 6 mm) 48 or 18 orifices/ft2
at 22°C at 22°C at 22°C at 22°C lateral blocks
area; maximum
1.728 9.8 12.3 (tile or RFP
14.8 17.3 lateral length of 50 ft
2 1.671 9.9 12.4 blocks)
14.9 17.5 (15.3 m)
3 1.618 10.0 12.6 15.1 17.7 12-in. lateral spacing;
4 1.567 10.1 12.7 Pipe laterals 3t05ft !to~in.
15.2 17.9 orifices are spaced
5 1.519 10.2 (0.9 to 1.5 m) (6 to 10 mm)
12.8 15.4 18.1
6 1.472 10.3 3 to 4 in. apart and
13.0 15.6 18.3
7 1.428 10.5 13.1 45° down-angle from
15.7 18.4
8 1.386 10.6 13.2 the horizontal on both
15.9 18.6
9 1.346 10.7 13.4 sides of the lateral;
16.0 18.8
10 1.307 10.8 13.5 maximum lateral
16.2 19.0
II 1.271 10.9 13.6 length of 20 ft (6 m)
16.3 19.2
12 1.235 11.0 13.8 16.5 19.4 4t07ft 0.25 to 0.75 mm Plenum or lateral bottom
13 1.202 11.1 Strainer nozzle
13.9 16.6 19.5 strainers spaced 6 to
14 1.169 11.2 (It02m)
14.0 16.8 19.7 10 in. apart; a space
15 1.139 11.3 14.1 16.9 19.9 less than 10 in. for
16 I. 109 11.4 14.3 17.1 20.1 the lateral
17 1.081 11.5 14.4 17.2 20.2
18 1.053 11.6 14.5 17.4 20.4
19 1.027 11.7 14.6 Self-Backwash Filters The only difference in design criteria for these types of
17.5 20.6
20 1.002 11.8 14.8 17.7 20.8 underdrain, as compared to normal backwash filters, is the required headloss across
21 0.9779 11.9 14.9 17.8 20.9 each filter. This in turn affects the orifice size of the underdrain system. The re-
22 0.9548 12.0 15.0 18.0 21.1 quired headloss, at the designed backwash rate, should be 0.5 to I ft (0.15 to
23 0.9325 12.0 15.1 18.1 21.3 0.3 m) while maintaining no more than a ±5% of water flow uniformity across a
24 0.9111 12.1 15.2 18.3 21.4 filter bed area. In the case of precast concrete laterals, this requirement is satisfied
25 0.8904 12.2 15.4 18.4 21.6
26 0.8705 12.3 15.5 by using orifices that are; in. (12 mm) in diameter and spaced 3 in. or 7.5 cm (cen-
18.5 21.8
27 0.8513 12.4 15.6 ter to center) on each side of the laterals. All other types of underdrain must be de-
18.7 21.9
28 0.8327 12.5 15.7 signed specifically to meet the required headloss.
18.8 22.1
29 0.8148 12.6 15.8 19.0 22.2
30 0.7975 12.7 15.9 19.1 22.4 Air-Scouring Backwash Filters The backwash filters listed below are the proven
"cP. centipoise. and most effective types available.

NORMAL BACKWASH FILTERS: WITH OR WITHOUT SURFACE WASHING Headloss

During Orifice
Headloss Washing Size Remarks
Type
at Ordinary Orifice
25 and 40 in. : in. (6 mm) 12-in. lateral spacing;
Backwash Size Dual-lateral
Type under simultaneous 23 dispersion
Rates (Diameter) blocks
Remarks orifices/ft2 of bed
(plastic air and water
Precast concrete 3t05ft i~ to ~ in. area; maximum
12-in. lateral spacing blocks) backwash at
laterals (I to 1.5 m) (810 10 mm) jointed lateral length
3-in. orifice spacing 3 ft3/ m in per square
(teepees) is 50 ft (15.3 m);
on either side of the foot constant air
lateral; maximum rate and 15 to 20 gravel layers
 BASIC APPROACH 239
238 DESIGN OF BASIC TREATMENT PROCESS UNITS

No.6 sieve No. 12 sieve 3 in. (75 mm) Top layer

Headloss 5
During (3 mOl) (1.7 mOl) Total depth:
Orifice
Type Washing 16 to 18 in.
Size Remarks
(400 to 450 mOl)
gpm/ft2 backwash required on top of
rate the blocks
Strainer 15 to 24 in. 0.25 to 0.77 nll11 The preferred height of NOTE: A few departments of health in the midwestern region of the United States demand
under 20 (slit size) the plenum bottom is 2 to 3 in. of 0.8- to 2.0-mm pea gravel (torpedo sand) as the top layer of the gravel bed, as
gpm!ft2 3 ft; durable and shown above.
wash rate proven types of
For the dual-parallel lateral block type of underdrain (Leopold. Inc.). the fol-
strainer should be
selected; a gravel lowing values are the most appropriate:
bed is usually not Depth of
Layer Passing Retaining
required above the Layer Note
Number Screen Size Screen Size
strainers
,. , In.
. 3 in. Bottom layer
I , In.
4"
. ,4 .In.
2"
3 in.
2 i-In.
NOTES: ' In.
. No.6 sieve 3 in.
3 4

No.6 sieve No. 12 sieve 3 in. Top layer

4
I. The porous plate type of underdrain system has been used successfully in conjunction with Total depth:
the automatic backwash filter, a proprietary unit. However, this type of underdrain will 12 in. (300 mm)
eventually become clogged if moderate to hard water (over 150 mglL of CaeD3) is til-
tered continuously. Periodic cleaning of the underdrain with acid, 5% solution of Hel, is
For the dual-lateral block type of underdrain with air-scouring wash (Leopold.
therefore mandatory.
Inc.). the following values are recommended:
2. The lateral type of underdrain should provide air release holes or slits at the top, end, and
midpoint of each lateral. Moreover, each end of the lateral should be anchored to prevent
Layer Passing Retaining Depth of
the laterals from moving during air scouring.
Screen Size Screen Size Layer Note
3. The most important design consideration for the underdrain system is the hydraulics dur- Number
ing backwash. The flow velocity at the entrance to each lateral should not exceed 3.5 ftls ,. ,. 2 in. Bottom layer
(1 mls), and the velocity at the entrance of the manifold is preferably less than 5 ft/s (1.5
4 In.
,. ,.
2 (fl.
2 in.
2 4" (fl. s 10.
m/s), so that a nearly uniform flow distribution is maintained in the filter bed.
3 H .
' 10. No. 10 sieve 2 in.
4. A plenum beneath the filter underdrain system ensures better !low distribution. Yet the un- (2 mm)
derdrain slab should be monolithic in nature, to minimize the problems of the underdrain , 10.
. ,. 2 in.
system. 4
, In.
.4"
,.
IiIn.
2 in.
5 ,2 .10.
4 10.
,2 . 2 in. Top layer
Gravel Support Bed For most underdrain systems the following values are com- 6 4" 10.
Total depth:
manly used.
12 in. (300 mOl)
Layer Passing Retaining Depth of
Number Screen Size Screen Size Layer Note
NOTES:
I; in. (40 mOl) ; in. (20 mm) 4 to 6 in. Bottom layer
(100 to 150 mOl) 1. Some manufacturers and textbooks recommend using a gravel bed that is shallower than
2 1 in. (20 mm) ~ in.(l2 mm) 3 in. (75 mm) 16 in. for ordinary underdrains, including the Wheeler bottom. ~owever, numerous case
3 in. (12 mOl) 1in.(6 mm) 3 in. (75 mOl) histories have proved that shallow gravel beds tend to become disrupted. .
4 in. (6 mm) No.6 sieve 3 in. (75 mm) 2. Larger gravel. such as the bottom two layers. must be at least. 3 to 4 in. (7 ..5 to t a em) 10

(3 mOl) depth to compensate for any inaccuracies that may occur dunng construction.
 240 DESIGN OF BASIC TREATMENT PROCESS UNITS
3. Th~ size a~d depth of the first layer should be adjusted according to the type of under_
dram that IS selected. Precast teepees and pipe laterals [6 in. (15 cm) in diameter] m
have a depth of 6 in. for the tirst layer to ensure that the laterals are covered. ust
4. Most strainer nozzle types of underdrains can eliminate the need for a gravel bed if the
nozzl~s ~re spaced closely together. However, if the slit size is greater than i&- in. (0.25
mm), It IS r~commended th~t t~e nozzle domcs be covered by a layer of pea-sized gravel
~ot onl.y to Improve good dlstnbution of backwash flow but also to avoid erosion and de-
~ormatlOn of the slits by the sandblasting effects caused by fasHnoving filter media duro
Ing backwash.
5. If air in the. ~nderdrain system and main backwash line is not purged prior to backwash,
the compositIOn of the gravel bed may be disfllpted during back washing, by the explosive
mov,ement of large volumes of compresscd air. This is the main cause of gravel bed up-
set tor regular backwash filters.
6. A rev~rse-~rad~d grav~1 bed of the type illustrated for the dual-lateral block type of un-
derdram with aJr-scounng wash, developed in Chicago by John Baylis, is called hourgla
g~adatioll. It effectively prevents gravel bed upset by using controlled amounts of airflo~
with backwash .. However, if ~his type of gravel bed is used in conjunction with ordinary
backwash, the flOe gravel (nllddle) layer is likely to accumulate sludge.
7. The quality of the gravel must meet the standards set by the AWWA (refer to AWWA
St~n.d~rds. B 100). The use of limestone must be avoided, especially when the filter bed is
aCidifIed m preparation for cleaning.
Miscellaneous Items
I. The water depth above the filter bed must be a minimum of 6 ft (1.8 m) to
minimize potential air binding of the filter bed.
2. The lowest water level in the backwash tanks should be approximately 35 ft
(10.7 m) above the lip of the wash-water troughs. The capacity of the back-
wash tank must be large enough to hold water for two filter backwashings.
3. The elevation of the weir, which controls the lilter eflluent water level, should
be set ~t a height that produces the appropriate hydraulic gradient at the top
01 the lilter bed at the designed liltration rate of the filters.
4. The maximum water level in the wash-waste holding tank should be properly
set so that the backup of water in the forebay of the filter is prevented when
the most remote filter is undergoing backwashing. The end of each wash trough
must have free flow at all times to maintain even flow conditions through each
trough.
Copying the European design (Figure 3.2.7-7), wash troughs need not be used if
ai:-scour~ng wash is adopted. However, a coarse deep bed filter with air scouring
will require the use of wash troughs; the troughs may be spaced beyond what is al-
lowed by the formula Si!,ee it applies to regular backwash systems. The wash troughs
Improve dlstnbutIon of the filter Influent, and consequently, media migration does
not occur. This is possible because liltration rates above 8 or 10 gpm/ft' (20 or 25
m/h) Induce a high intlow rate, thus changing the hydraulic characteristics of the
intluent. The filter-to-waste system is important for lilter design. This system may
BASIC APPROACH 241
not be used all the time under normal operational conditions but is needed under
abnormal situations. Abnormal conditions include bed reconstruction, disturbance
of the Ii Iter bed (including gravel bed), turbidity breakthrough, and repairs to the
undcrdrain system.
Basic Filter Design Scheme In this section we provide a few general tips and
describe brietly some basic tilter designs.
GENERAL The possibility of contaminating the treated water, due to leakages, cross
connections, or operator error, is minimized by providing a clear physical separa-
tion between the unfiltered water line and the filtered water line. Engineers are there-
fore advised not to locate these pipelines or conduits in the same tilter pipe gallery.
When designing a filtration system, the engineer must consider the accessibility of
the various mechanical items. The pipe gallery should have a clear and wide pas-
sage so that operators can maneuver vehicles, which may be necess~ry to provide
adequate maintenance and repair work for large-scale plants. All maJor valves and
valve operators should be situated for ease of visual observation and service by the
plant operator. Additionally, an access hole should be provided in the Ii Iter under-
drain to facilitate inspection. The backwash main in the filter pipe gallery should be
located below the water level within the filter cells to avoid the accumulation of air
in the backwash main. Also, an air release valve should be furnished at the end of
the wash-water line, and air relief valves should be supplied for the filtered water
line and the main wash-waste lines. If air is allowed to accumulate within these pipe
lines, the !low may be reduced by an air blockage and the explosive escape of the
air could damage the filters.
The top of the forebay walls should be higher than the highest water level in the
filter cell. This design feature minimizes tilter media migration, which is initiated
by the inertia of the intluent tlow. However. most European air~scouring wash fil-
ters do not follow this criterion and the majority of these tilters therefore have prob-
lems of varying degrees with filter media migration. Engineers should try to reeval-,
uate the advantage of the flow rate controller in each filter effluent line because ot
the recent requirement of practicing tilter~to-waste. The tlow rate controller is al-
most an essential item when switching from lilter-to-waste to filtration mode in or-
der to avoid hydraulic shock loading (refer to Figure 3.7.2-12).
Problems associated with air binding are minimized when a minimum water depth
of 6 ft (2 m) is provided above the IiIter bed. This depth also induces a higher avail-
able headloss for filtration. The problem of air binding may also be prevented by
installing a control weir for the eflluent level of the filter; this scheme will provide
positive back pressure in the entire bed at all times. Additionally. this control well'
prevents the disturbance of the gravel bed by entrained air during backwash by con-
tinually !loading the lilter bed. A more detailed discussion of air binding is pre-
sented later in this chapter.
Based on the considerations described above, a typical tilter arrangement will be
similar to that illustrated in Figures 3.2.7-13 and 3.2.7-14.
 242 DESIGN OF BASIC TREATMENT PROCESS UNITS TOP 01 'OAE,",Y WALL

..."""
GALLI"V

,WASH
INFL'--NT
(TVPJ

/TROOOHS GRATING
IACKWAIH 000 000

fiLTERED fiLTERED
WATER MAIN ORAIN LINE WATER
CAL TERNA liVE CpNOUlT
TO CHANNEl) WASH WASTE

(,)

INFLUENT------~~ GALLERY
CHANNEL (TOP) GLASS BlOCK
/SKVl.I GH' / ' 0 ' OF FORE lAY WALL

~
INflUEN - QAUBlY
, '"' (... AN Of"TION
VALVE fDA
~LAN Iinow WALKWAY LEVU)
~- Ii """""NO
ORATINOS
CHANN!
FILTER
OPERATION CONSOLE BACKWAIH MAIN
t: 0 0 0
~
0 0 0
RATE'
Fll.TRATlOH

HANDRAILS
V
=
~~~IT~Y~P~.)~~~~~~~~~t__ .-
WASH
WASTE
j. lit;; _.:
FIL TEA INFLUENT CHANNEL
CHANNEL WASH TROUGH raTU.TO··] \ , ij \OWER O.....l£T
ITY~.) WASTI ....
o 0 0 0 0 EFFlUENT CHAN'£l.
(')
UNDERDRAINI
ITYP.) SKYLIGHT
BACKWASH
INFLueNT
WEIR
MAINI I~~~~~~ -",)1"1"1---'::'---1 •o
A CHANNEL) (TY',) 0 0
IICTION A - A
•• 1

WASH WASTE
CONOUIT EFFLUENT
(TVPJ CONOUIT FitTER TO WASTE
(0)

$KYUGHt
INFLUENT AIR PIPE
AJR PIPE

WASH
WASTE

.ICTION I - I

RL TER TO .A~=!oi..i
ORIFICE FOR
Figure 3.2.7-13 Basic filter design scheme. WASHWATER
WASTE
ALT .• ...-.i_ ALT.S

BASIC FILTER STRUCTURES When the total number of filters exceed six, they are
generaUy arranged on both sides of the filter gaUery, primarily for cost-saving rea-
sons. FIgure 3.2.7-14 depIcts five basic design schemes. Alternatives I and 2 per-
tam to the self-backwash filters and differ only in the design of the filtered water
coUection main. Alternative 3 is a conventional type of filter with an equal-Ioad-
mg, constant-level control system. Each filter in this scheme has a flow-modulat-
ing valve in its effluent line but lacks a flow metering device. Alternatives 4 and 5
(d)
Figure 3.2.7-14 (a) Alternative t: self·backwash filter with surface wash. A clear separa-
tion of filtered and unfiltered pipings; the structure is approximately 25 ft deep. (b) Alternative
2: self-backwash filter with surface wash. A central filtered water channel arrangement; the
structure is approximately 25 ft deep. (e) Alternative 3: conventional filter with surface wash.
Approxima(ely 17 f( for filter cell and 25 n height for pipe gallery. (tf) Allerna(ives 4 and 5:
air-scouring wash filters. Approximately 17 ft for filter cell and pipe gallery, except alterna-
live 4, which has 25 f( for height of gallery.
243
 244 DESIGN OF BASIC TREATMENT PROCESS UNITS
relate to air-scouring wash filters. Although the lilter-to-waste line is not shown'
alter~ative 5, it can easi.ly be provided at the intersection of the backwash and fill~
ter effluent valves. One Important consideration is elevation of the air supply head~
ers. The headers should be located at least 2 ft (0.6 m) above the highest wat
level. in the lilter cells, thereby reducing the chance of water back-siphoning in~~
the air hne.
See Appendix 24 for an example of filter design.
Example DeSign Calculations Three examples are discussed in this section
They concern selection of the most appropriate filtration process and filter desig~
and present the best method for determining the appropriate combination of lilte
media for multimedia beds. r
Example 1 Filtration Process
Select the most appropriate type of filtration process and overalJ process train based
on the SOurce of the water and the following local conditions.
C~se 1. A treatment plant in the Midwest has an ultimate capacity of 90 mgd
(4 m Is). The Source IS a large fIVer that exhibits high-turbidity spikes, up to 500
ntu, dUrIng the summer months. The source also has an average coliform Count of
780 MPN (most probable number), an average total organic carbon (TOC) value of
3.5 mgiL, a seasonal algae bloom, and associated taste and odor problems that are
prImarily a function of diatoms.
3
Case 2. A 150-mgd (6.5-m /s) plant located in the Northwest obtains water Irom
a large lake that consistently has good water quality and an average turbidity of
1:3 ntu. How~v~r, turno~er of the lake water occurs twice a year. During these pe-
nods the turbldl!Y level IS expected to rise as high as 10 ntu and the iron and man-
ganese levels may become as high as 1.2 and 0.2 mglL, respectively. Additionally,
the algae count can be expected to be as high as 1000 areal standard units
(ASU)/mL.
Case 3. A small town in a developing Asian country requires a 3-mgd (0. I 3_m 3/s)
plant. The raw water may be obtained from a local lake. The quality of this water is
good except dunng the monsoon season. Although accurate records are not available
the turbidity is expected to be approximately 50 ntu during the monsoon season. '
Case 4. A town proposes to treat 5 mgd (0.22 m 3/s) of water obtained from three
wells. The well water contains an average of 2.2 mglL iron and 0.4 mglL manganese
and has a hardness of 550 mglL. Other elements, including turbidity, are always un-
der the maximum contammant level (MCL) of the drinking water standards. The
plant IS located in Alaska.
Case 5. A large camp site, located in a national park in the Central Mountain re-
gion, requires a maximum of3 mgd (0.13 m 3/s) of drinking water during the tourist
~ea~on. The water source is a local impoundment fed by streams. The water qual-
Ity IS reasonably good but has occasional turbidity spikes of up to 50 ntu during
BASIC APPROACH 245
thunderstorms. The water contains moderate coliform and algae counts and may
contain Giardia. The plant is shut down from November to April.
Solution Case I. A coarse deep monomedium filter bed with air-scouring wash is
the most appropriate choice and a provision for future conversion of these filter beds
to GAC beds should be considered seriously. The bed should have 6 to 8 ft (2 to 2.4
m) of sand or anthracite coal monomedium, with an effective size of l.2 to l.5 mm,
depending on the depth of the bed. The monomedium should have a uniformity co-
efficient below 1.4 and a filtration rate of 6 to 8 gpm/ft2 (15 to 20 mlh); these fig-
ures are reasonable choices. A filter bed that is 6 ft (l.8 m) deep provides an empty
bed contact time (EBCT) of 15 min at a filtration rate of3 gpmlft2 (7.5 mlh). However,
if the rate is 6 gpm/ft2 (15 mlh), the depth of the bed must be 8 ft (2.4 m) to main-
tain 10 min. contact time. Placement of a 12-in. (O.3-m) sand barrier (0.7 mm ef-
fective size) at the bottom of the filter bed should also be considered.
The overall process train should be preozonation, coagulation, flocculation, sed-
imentation, filtration, and chloramination. The provision for converting the filter bed
to GAC adsorption should include the addition of filter cells to compensate for the
lower allowable filtration rate of the GAC filters. Since the GAC filter effluent con-
tains very high counts of bacteria, additional postozonation may be implemented
prior to postchloramination. A provision for feeding hydrogen peroxide during pre-
ozonation and the addition of a surface wash system may also be mcluded 10 the
design to enhance the efficiency of ozonation and to eliminate the problem of mud
ball formation within the deep filter bed.
Case 2. In this case the recommended system is a standard dual-media lilter bed
with surface wash as the auxiliary scouring during backwash. The self-backwash
type of filter is selected because of its simplicity, energy efficiency, and proven per-
formance. These characteristics minimize operation and maintenance efforts. The
treatment process train includes prechlorination, coagulation, and roughing filters
prior to the dual-media filters and chlorination process; this type of process train is
called two-stage filtration. A bypass channel should be provided around the rough-
ing filters because the plant will be operated as an in-line filtration process during
most of the year. However, the roughing filter must be employed during seasons
with high algae counts and periods of high turbidity.
A provision for future addition of the preozonation and chloramination processes,
as well as granular GAC capping of the filter beds, should be included to control
taste and odor problems. Direct filtration is not the appropriate choice for this situ-
ation since this process has no defense against very short filter runs due to algae
blooms. The roughing filter of the two-stage filtration performs as a flocculation
process and also removes 50 to 75% of suspended matter. The process of two-stage
filtration is discussed in detail in Section 3.2.8.
Case 3. Considering the local conditions and the raw water quality, slow sand
filtration with a prescdimentation reservoir should be selected as the treatment
scheme. This system is simple to operate and maintain and docs not require costly
maintenance and repairs. The presedimentation reservoir should provide a storage
 BASIC APPROACH 247
246 DESIGN OF BASIC TREATMENT PROCESS UNITS

period of I to 2 days because this will facilitate the removal of large suspended par- Determine
ticles during periods of turbidity spikes (in the raw water). An alternative choice is
a membrane filtration system. (i) The total number of tilters _.
Case 4. The appropriate choice for this situation is a standard dual-media bed or (ii) The type of filter bed, the filtration rate, and the hlter wash1l1g scheme
a medium-sized monosand bed, with surl~'lce wash scouring during filter wash. The (iii) The size of each filter
first choice for the overall treatment process is lime softening (lime flocculation and
(iv) The filter arrangement
c1arilication) by means of a sludge blanket type of reactor clarifier, followed by acid
pH control or carbon dioxide (for recarbonation), then clarifier, all prior to the fil- (v) The size and type of the major filter valves
ters. Lime softening reduces the water hardness to a range of 80 mg/L and effec- (vi) The lilter influent channel, pipings, and central gallery
tively removes iron and manganese at the levels dictated in the problem. (vii) The type of lilter media, gravel bed, and underdrain
The second choice for the overall treatment process begins with the oxidation of The backwash rate, the type of wash troughs, and auxiliary scour wash
(viii)
iron and manganese with either chlorine or potassium permanganate, or both. The (ix) The plan and sections of the filter _ . .
second step is a regular pretreatment to the dual-media bed liltration, followed by The elevation of the tilter effluent control weir and the configuratIOn of
(x)
sidestream softening via an ion exchange process. Finally. the softened and filtered the ernuent channel
water are blended to produce a water with medium hardness. Flocculation and sed- The device used to measure the backwash flow rate .
imentation are preceded by oxidation of the reduced metals.
(xi)
The filter piping scheme that allows for the lilter-to-waste teature and can
The pH of the treatment process water should be kept in the range 8 to 8.5 by (xii)
tinuous sampling of the lilter effluent
using ferric iron coagulant, cationic polymer, or both. Water softening may be
achieved through a cation exchange resin process or a membrane technique. with
proper pH adjustment and the addition of a scale inhibitor, due to the rather small . . The total number of filters in the f1l1al stage of this plant is N =
output of the treatment plant. All processes must be housed since the ambient tem-
SI02li('lftOOO")O}I~
12 Therefore six of these filters arc constructed in the IIlItJaI stagde.
peratures are extremely cold during the winter months. . . , . f I t I' a standard dual-me Ja
Case 5. If enough land is available, a slow sand filtration process should be con- (")
II
Due to the relatively good quality a t 1e raW wa e, .' .
. ' fI
. .
b d h s a history 01 reliable pertOlmance
sidered because of its simplicity, reliability, and low operation and maintenance filter bed iSprefetTed·tThlS tit~es::'fa~~e~a:h a~d does not employ noisy air blow-
costs. However, it is very unlikely that the area required by the slow sand filtration when used 111 conJunc IOn w . g the peace of the neighborhood. The
will be available in the mountainous location. Due to the size of the plant, the sea- ers for the air-scour process, thereby !"eservln . 6 m/ft' (15 m/h). Under the
sonal operation, and the limited maintenance services, a proven type of semiauto- designed liltration rate, with all the hlters ~n ~1I1e, 1\gp latory agencies. Note that
matic plant, including a membrane filtration process, is recommended. given conditions, this rate IS wl~e:~ a~~:ist~~O~I;~~ ~~~~ larger than the designed
The filter bed should be a simple monosand bed (rapid sand bed) because of the
low initial cost and simple operation. The primary choice for the process is mem-
~~~es~~ea~!~I~em:S!~~ ~l~~:e~~ial~y ahigher loading rate to each tilter during future
brane microliltration. The second choice is a package type of flocculation/sedi-
mentation pretreatment with a self-backwash type of pressure filter. The final choice
ope;~I~i~rn;"'ashing is achieved through normal backwa~h: a sel~-~:c~:~Sh;~:ers~I~:
si n with surface wash as the aUXiliary scounng ur1l1g a .
b:ckwash scheme is selected because it ~esolves importan.t ~~cto~s, ~u~~ea~~~~:~
is a roughing filter followed by proprietary automatic backwash filters. The last
two processes require skilled and knowledgeable operators. Selection of the treat-
ment process is strongly influenced by the requirements set by the local regulatory thetic and noise considerations that are dIctated by ~he n~lgt dar a~~ the slope of
agency. bility of the water freezing should an ele:ated ~atk de sce:cc:v~tion since the til-
the site. The average 2% slope rcduc~s the nee or ~~p efficient wash tank sys-
gy
Example 2 Filter Design tel's must be located at a lower elevation. The most enet -' d only I'j' the plant
k H . 't should be canst d ere
tern is the elevated wash tan.. ~wevel, I . d If th wash tank is an elevated
Give" A 50-mgd (2.2-m'/s) filtration plant with an ultimate capacity of 100 mgd site has a suitable hill on which It can be situate '. e '. .' n
tank on legs or a high cylindrical tank, this results In a negative Visual Impact 0
(4.4 m'/s) is located in a quiet and good residential area in the Northeast. The plant
site has an average 2% slope (one way). The treatment process train is a conven- the neighborhood. . 11583 ft' (1080 m') or (100
tional complete process and pumps lake water that has a reasonably good quality
throughout the year.
X ~~~) ~:: t~a~ ~ep;::~~~2d ~:~e:e~~~ ~~~:I a~ull~~e~~1 ~I~ter~ is -12, each fIlter bed area

.!.iIi.
 248 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 249

2
is 965 fe (90 m ). Each filter will have two filter cells and a central gullet; this con-
figuration ensures favorable hydraulic characteristics during backwash. The filter
area of each filter cell is 965 7 2 = 482.5 ft2 (45 m2). Therefore, make each filter
0; vi) To establish an approximately equal water level in each inlet channel, the
velocity at the beginning of the influent channel must not exceed 1.5 ftls (0.45
fl /s ) Since each filter is equipped with identically siz~d recta~gular weir tOlets at
cell 15 ft (4.5 m) wide and 32 ft (10 m) long, so that the area of each cell is 480 ~e s"ame elevation, this ensures that all filters on the hne receive almost the same
2
ft2 (45 m ). If the regulatory agency requires a filtration rate of 6 gpm/ft 2 when one flow rate (i.e., equal loading). Id b
filter is off line, each filter cell should have a bed area of 520 ft2 (48.5 m2). We The influent channel is located directly above the pipe gallery and shou e ap-
check the actual filtration rate by calculating: . t I 22 ft (6 7 m) in width to allow for a low flow velOCity. The gallery there-
prOXima e Y . I d th uip-
fore has ample space for inspection and maintenance of the va ~es an o. er eq
69,500 gpm 7 (480 X 2 X 12) ft2 = 6.03 gpm/ft2 = 15 mlh OK ment. For this type of filter, the total depth of the filter str~cture IS approximately 2~
ft 7.6 m), which allows sufficient water depth to be prOVided above the bed. As 11_
(
lustrate d I'n F'lgure .327-15
. . , the channel width, the height
" of the gallery, and
. thef th10
(iv) The filters are arranged as illustrated in alternative I of Figure 3.2.7-14. A
double row of filters is constructed along a central pipe gallery where the piping for . f the I'nlet and wash-waste pipelines are all dictated by the dimenSIOns 0 e
catton 0 . f h t lis
the filter influent and wash waste are located. The filtered water channels are situ- pipe fittings and the valves, in addition to the thickness 0 t e concre e wa .
ated opposite from the central gallery so that there is no possibility of contaminat-
ing the filtered water by cross connection. The filtered water may be collected into - <t
a channel instead of a large, expensive pipeline and fittings because of the high flow
rate of this system and to allow the valves to be accessible for easy maintenance
Slits for stop logs

f+--,<O-----22·-0·------..,·1
I
and repair. t
(v) The size and types of major filter valves are listed below. iP:' 'I," t' 00,'.4'

Inlluent Ilow rate = 6.03 gpm/ft2 X 960 ft2 = 5790 gpm = 12.9 ft 3/s
Backwash flow rate = 22 gpm/fe X 960 ft2 = 21,120 gpm = 47.1 ft 3/s •
o 9' -6' (Width of channet)
Surface wash flow rate = 3 gpm/ft 2 X 960 ft2 = 2880 gpm = 6.4 ft 3/s ~ 1
Water jets for agitation
""~
Based on these data, the following selections are made: •.
:.~~ ~.~:~~!

Velocity Through
the Valve
Valve Size and Type (fils)
Filter to waste
Filter influent 24-in. butterfly valve 4.1 OK outlet (it needed)
Filter effluent 36-in. butterfly valve 1.83 OK
Backwash (maximum) Filter eflluent value is also the 6.7 OK
backwash value for the
self-backwash filters
Surface wash 12-in. butterfly valve 8.2 OK
Filter-ta-waste 20-in. butterfly valve 5.9 OK
Wash-waste drain 36-in. butterfly valve 7.6 for back and surface
wash OK <t ot tilter gallery
Gutter Mechanical coupling (typical)

Wash-waste main"""'"
NOTE: The sluice gate and the gate valve arc bulky and expensive items that may require
frequent maintenance and repair. in contrast to the buuerlly valves specified by the AWWA Figure 3.2.7.15 Filter piping. influenl. and pipe gallery design (only the left half. from the
standards.
centerline, is shown),
 250 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 251

.Constructability, the ease with which the valves can be removed during major re- For this example the tilter underdrain system of choice is the precast concrete
pmrs, the wall pipes for leakage water stops, and the slope of the floor tow~rd the laterals with a triangular cross section (Figure 3.2.7-8). Dual lateral blocks or pipe
dra~nage gutt~rs are ~ll important design considerations. Unfortunately, many inex- laterals may be employed if the owner prefers their use. Regardless of the type of
pe~tenced deslg~ engineers disregard topics such as providing a minimum space re- underdrain, the headloss should not exceed 12 in. (0.3 m) at the designed backwash
qUIrement of 9 m. (22.5 em) between the wall and the adjacent pipe flanges, for rate. In addition, a uniform flow distribution must be maintained over the entire fil-
easy InstallatIOn of the nuts and bolts, for supplying mcchanical couplings to facil- ter bed; the difference should be less than 5%. The higher the headloss, the deeper
Itate minor adjustments in pipe alignment and easy removal of the valves, and fi- the filter structure. Precast concrete laterals that are 14.5 ft (4.4 m) long are placed
nally, for adequate drainage of water Irom the gallery floor. at 12-in. (O.3-m) intervals, and both sides of the sloped walls should have orifices
. The influent channel has a cross-sectional area that is 6 X 9.5 = 57 ft' (5.3 m'). that are t in. in diameter and 3 in. on center.
SInce the settled water IS dIVIded among the six lirst-stage tilters and the six linal- Each filter cell is 32 ft (9.6 m) in length and there are 64 laterals in each lilter.
sta~e ~lters, ~he lOO-mgd flow is split into two directions. The maximum flow ve- The lower gullets should be designed with a reasonably large cross section for three
lOCity In the Influent channel is therefore reasons: (I) to provide adequate space for construction workers to work, (2) to al-
low easy inspection and cleaning of the inside of each underdrain lateral, and (3) to
3
(50 X 1.55) ft /s .,. 57 ft' = 1.36 ftls Or 0.4 mls OK provide a low backwash flow velocity in the gullet to ensure uniform flow distrib-
ution to each lateral. As a rule of thumb, the backwash flow velocity in the gullet
(vii) The specitications for the tilter media area as follows: should be less than 2 ftls. Also, the velocity should be less than half of the inlet
flow velocity to each underdrain lateral. Thus, make the lower gullet size 5 ft in
Anthracite Coal Filter Sand width and 5 ft in height (for the wash-water connection side) and 3 ft in height at
(Top Layer) (Bottom Layer) Remarks the other end.
The maximum backwash flow rate for each lateral is
Effective size (mm) 1.0 :!: 0.05 0.55 :!: 0.02 For purchase
specifications 3
Uniformity :51.4 21,120 gpm .,. 64 = 330 gpm or 0.735 ft /s
:51.5 For purchase
coefticient specifications
Specilic gravity 1.65 :!: 0.05 2':2.63 The cross-sectional area of the lateral (a triangle cross-section) is
After 24 h of soaking
Depth (in.) 20 10 After scraping off the
(0.615 X 0.6) .,. 2 = 0.185 ft'
top layer of tine
particles
The flow velocity at the inlet of the lateral is
NOTES:
0.735 .,. 0.185 = 4.0 ftls < 2 X 1.9 ftls = 3.8 ftls OK
1. When each medium is first placed in the filters, an extra 1.5 to 2 in. should bc added to
~he depth to allow for removal of the fine particles (scraping) after backwashing, as spec- Make the size of the lower gullet 5 ft in width and 5 ft in height at the etTIuent chan-
Ified bel~w. The process of scraping is vital for proper functioning of the tiltcr bed. If the nel side (the entrance of the backwash water). The maximum flow velocity in the
fine particles are not removed from the surface of the bed, it will result in very short fil- lower gullet during backwash is
ter runs.
2. Following the ini~ial placement of the sand, the filter should be backwashed twice at a 47.1 ft 3/s + (5 ft X 5 ft) = 1.9 ftls
rate of 22 gpmlft (55 mlh). After the first backwash, the top ~ in. of medium should be
removed. Repeat the backwash process and scrape an additional ~- in. off the top. Thus the
The total number of orifices Gin. diametcr) per lateral is
final depth of the sand layer will be lOin. (25 em).
3. Place the coal layer on top of the sand layer and backwash the filter bed three times at 22
2 X 58 = 116 or 7424 oritices per tilter
gpm/ft2 (55 m/~~, scraping off the top layer after each wash. Remove the top I in. after
the first wash, i 111. after the second wash, and 1in. after the third wash.
4. The AWWA filter material speCifications (AWWA standards B 100) should be used for the
The flow rate from each orifice is
other necessary properties of the media. The gravel bed is composed of five layers of
graded gravel and the total depth is 18 in. (45 cm). See item 8(a) for the details. 0.735 .,. 116 = 0.0063 ft 3/s
 252 DESIGN OF BASIC TREATMENT PROCESS UNITS
BASIC APPROACH 253

The cross-sectional area of each orifice is

flow per trough. ill/sec

(0,5112), X 0,785 ~ 0,00136 ft' r-.oocooo

o"';"";<"i<"il"'irri"': ..; ..; uiui..o .,; ..: 0'> C'i_ .... _ .... _
20
The backwash flow velocity through each oritice is t9
t8
0,0063 + 0,00 136 ~ 4,63 ft/s 17
t6
The head loss through each orifice hole is 5 '<--
4
<>
'J.'"
~'

~ ~~ ~
2 3 ,>'
tJ.h K 2.4 4,63 0,80 ft < I ft OK 2
'\.
"'- ~
2g 64.4

It is imperative that each lateral be anchored to the noor even if the weight of the
lateral exceeds the normal uplift force that is generated during an average backwash.
,.
c
_ 11
-, t 0
9
8 ~ "
...
t:'
~... t-I~"

~
"
.~
~
~
<0'
'>
~

1- ,B I
~.'imum
water
level

8~ ~J1
The orifices should be located in the upper portion of the inner triangular section and 7
directed horizontally so that air will not be trapped in the lateral and so that the wa- 6 "I
ter jets are not directed toward the surface of the tilter bed, Both ends of each ori- approx~ ~~-
tiee should be reamed to provide rounded edges, minimizing the headloss. Computer
5
4
m.tely
2 ,n,
'("<1./ W + B.2
I
analysis of this type of underdrain reveals an approximate maldistribution of ±5%. 3
Typical ~ss secttOn
(viii) The designed backwash rate is the maximum wash rate and should only
be used during the summer months when the water temperature is high. Based on
8ggg8g8~8~~§~8~~~~8~~§~g2~
0 0 0 0 0
Figure 3,2,7-4 and the specitied tilter media, the maximum backwash rate is 22 ~~~~~~~_NNNNN~M~~M~~~~~~~~
gpmlft' (55 m/h), The maximum backwash now rate should be at least 30 mgd if flow per trough, gal/min
the maximum backwash rate is used, Due to the design of the self-backwash filter, Figure 3.2.7-16 Wash trough sizing diagram. (Courtesy of F.B, Leopold Company.)
27 mgd must be liltered to produce a backwash !low rate of 19 gpmlft', Whenever
the plant is operating at less than the minimum plant flow rate required to backwash
one tilter, the required ratc may be met simply by increasing the plant flow rate to Make it 4 ft.
the rate required to wash one tilter. The spacing (5) of the trough is
Each tilter cell is provided with four wash troughs composed of tiberglass-
reinforced plastic, The lightweight and durable troughs facilitate easy installation (1.5 X 4) < 5 < (2,5 X 4)
and simple adjustment of their level during construction and provide trouble-free
maintenance. The maximum flow rate for each trough is 6ft<5<IOft

(21,120 + 2,880) gpm + 8 ~ 3,000 gpm Use a spacing of 8 ft on center. . _. . :>

By using the trough sizing diagram (Figure 3,2,7-16), a trough with a width of
21 in, and an overall height of 24 in" with a half-circle-shaped bottom, is selected,
The elevation of the trough should be high enough so that media grains kicked
up by the surface wash jets are not easily washed out during the simultaneous sur-
lace wash and backwash period, The height (Ha) of the troughs above the filter bed
is (refer to the discussion of filter wash troughs)
(0,5 X 2,5) + 2 < Ha < (2,5 + 2)
3,25 ft < Ho < 4.5 ft
Auxiliary scouring will be provided by a surtace wash sy~te~. T~e tllSt ch~lce
is either a fixed nozzle 01' rotating arm type. The second chOice IS either a ~urta~e
wash or a subsurface wash unit. Despite its etfectiveness, the Baylis typ~ at auxil-
iary scouring (Figure 3.2.7-17), consisting of a series of ~ertical pi.pes .wlth nozzle
caps at each end, is not adopted because an intense jet at water will hit the gra~el
bed directly and upset the stratitication whenever the caps are lost; several case hIS-
tories exist for this problem. .' .
In this application a tixed nozzle lype of auxiliary scounng IS used Simply be-
cause of fewer maintenance requirements. Three-inch-diameter schedu~e 80 PVC or
stainless steel lateral pipes should be horizontally arra,nged at 4-ft, IOtervals and
3 in. above the surface of the bed. The orifices must be ~ 111. (6 mm) 111 diUmetel at
 254 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 255

The area of each orifice is

(0.25 .;- 12)' X 0.785 = 0.00034 ft'

while the flow velocity at each orifice is

0.0091 .;- 0.00034 = 26.8 ftls

To satisfy the design criteria, the jet velocity for surface scouring should be between
20 and 30 ftls (6 and 9 m/s).
The loss of hydraulic head through the orifice is
Elevation of ,urface-wash system piping in
plant filter No.2
l1h = 2A( 26.8') = 27 ft or 11.7 psi
6404

The source water may be either settled or filtered water. The latter is more common
Sections through
perforated and must be supplied by a pump; the use of filtered water also requires a special
plastic cop design in order to avoid the possible backflow of wash waste into the filtered wa-
Figure 3.2.7-17 Example of the Baylis surface washing system. ter tank. The total dynamic head of the pump is a function of the size and length of
the supply pipe and the elevation of the water source. The pump generally has a
approximately 8-in. intervals. The jets should be directed 25 to 35° below the hor- minimum total dynamic head of 65 ft (20 m).
izontal plane so that the coal and sand interface will be scoured. (ix) The plan, sections, and details of the designed surface wash system are shown
The subsurface wash systems have a slightly better performance when used in in Figure 3.2.7-18.
conjunction with dual-media beds. Yet experience over the last 20 years has shown (x) The elevation of the effluent level control weir is an important factor in set-
that the difference is marginal. In fact, the subsurface wash system potentially has ting the filter backwash rate control weir. The hydraulics of the filter backwash is
problems with the nozzles becoming clogged whenever the rubber caps are lost. The evaluated on a wash rate of 22 gpm/ft2 (48.2 ftls). The 36-in. butterfly valve (BV)
flow rate for each filter during surface wash is 2880 gpm or 604 ft 3/s (based on entrance loss, BY loss, and exit loss are given by
3 gpm/ft2). Since there are 8 laterals in each cell and 16 laterals in each filter, the
flow rate to each lateral is .
h, = (0.5 + 0.25 + 1.0) !L = 1.75 (6.7)2 = 1.22 ft
2g 2g
2880 .;- 16 = 180 gpm (0040 ft'/s)
Headloss in the lower gullet is
Therefore, the flow velocity in the 3-in. lateral is 8.2 ftis, which is a reasonable figure.
Each lateral in the surface wash system should have !-in. orifices drilled in a v = 47.1 .;- 5' = 1.9 ftls maximum (negligible loss)
staggered fashion at 8-in. intervals on either surface of its 14.5-ft length. The entire
width of the filter cell, 15 ft, may not be covered by the orifices due to the maneu- Entrance to the underdrain lateral is
v~ring space that is required to install fittings such as elbows and end caps, and pos-
SIble repair of the piping system. h, = 0.5!L = 0.5 (4.0)2 = 0.12 ft
The total number of orifices per lateral is 2g 2g

[(14.5 X 12) .;- 8] X 2 = 44 Headloss in the underdrain lateral is negligible. Headloss at 0.5-in. orifices of the
underdrain is given by
The flow rate through each orifice is

h, = 204 (4'263 )' = 0.80 ft

0040 .;- 44 = 0.0091 ft'ls . g
 256 DESIGN OF BASIC TREATMENT PROCESS UNITS
BASIC APPROACH 257

Handrails The height of water flows into the wash troughs is 2 in. (0.167 ft). Thus
J

f'~;~:,,~9H~WL~S~~~~~'~;~~~P/25'OO '" = 0.17 ft

;~ 12~~ce
Total headloss is
It main
.:iL,LWL fiF~~_~/;",~~__
(End filter Beginning of >' 0 wash
~ I
Top of bed ~
EI 4 33 " 0 3' PVC
cycle)

surface
filter cycle

8 ,;
I i=7
Li=
i=1
4.2 ft

Coak
Sa. d'-~~I~!!!i!laitelrial!i!iit:::!lijli'i!!ijl!ii!;:;w:a:S:;htrough
~ .. ' . ., ...... "'0
'" ( ) ::
' ••

Gra~el ;:~"'.::'::'.";.~;:,;.,,,.;,.:.;:,;':.'::>:,;,: ":. '. . .

';
I

undefdrain
Since a headloss of approximately 4.2 ft (1.28 m) will be provided during one
backwashing at a rate of 22 gpm/ft2 (55 m/h), the effluent level control weir should
be set so that the water level in the eflluent channel is 4.2 ft (1.28 m) higher than

, the top of the wash-water troughs. To adjust the overflow elevation to a 4-ft span,
two weirs 10 ft long (3 m), and two down-opening slide gates are installed in the
(a> wall of the filter eflluent channel. As illustrated in Figure 3.2.7-19, section A-A,
the top of the notched concrete wall should be set at an elevation of 10.25, that is,
03' PVC pipe I ft II in. above the top of the wash troughs, which are located in the filters. The

~
I,~II~'" 3' PVC
operator may manually raise the down-opening slide gates from an elevation of 10.25
to 13.25 whenever the backwash rate must be adjusted. Under these circumstances
21xl!Xl! the gate acts as a weir. The required hydraulic head for a plant flow rate of 50 mgd,
222
Stainless steel 4'1 'd'lamet er hole using two weirs that are 10 ft (3 m) in length, is

H1.S = (50 X 1.55) + 2 l.l6

3.33 X 10

~
........ ,. ~rout(4t)
Thus H is l.l ft.
&> A p: T The top edge of the slide gates should roughly be set at an elevation of 11.43
4!N
Underdrains Concrete underdrams 4
(say, 11.45) if the plant is to produce a wash rate of 22 gpmlft2 (55 m/h) when the
o 3' plant flow rate is 50 mgd. The effluent channel is 6 ft (1.8 m) wide with an aver-
suriace lateral age water depth of 12 ft (3.6 m). At a plant flow rate of 50 mgd, the maximum flow
(b> velocity in the channel is
Figure 3.~.7-18 Filter structure design: (a) cross section of the tilter; (b) details of the sur-
face washmg system. (50 X 1.55) ft'ls (6 X 12) ft = 1.08 ftls < 2 ftls

Headloss Ihrough the gravel bed is

NOTE: As described earlier, a new filter bank, identical to this filter bank, will be added
114 = 0.38 ft at a future date. The settled water will flow into the filters between the two filter banks. In the
future, this second filter bank will have an identical filter effluent level control system and the
obtained from a pilot study conducted at 16°C. Headloss through the tilter bed is filter effluent from the two banks will be combined. The flow rate will be measured using a
Venturi meter.
"s = 1.42 ft
(xi) The rate of backwash flow is measured by a venturi meter that is placed in
obtained from a pilot study conducted at 16°C. Headloss over the wash trough the 54-in. filter eflluent main. When one filter is undergoing backwash, the meter
(troughs as weirs) is will indicate a decrease in flow rate. This decreased flow rate is the actual wash rate
for the filter. The backwash rate may be adjusted by either raising or lowering the
h6 = 0.17 ft down-opening slide gates of the filter effluent channel.
258 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 259

(xii) Allernative 1 of Figure 3.2.7-14 should be adopted iflilter-to-waste is prac-

32'-0' 5'_0. 9 ' ticed at the beginning of each filler cycle or if there is continuous sampling of til-
'" 54' tiS • '[Hr- It tered water from each filter. The design in Figure 3.2,7-19 should be modi lied to
0 0
r-- provide a service gallery. The volume of the wastewater can be quite large; how-
I=F 4' 3@S'c24'
4' ever, the quality of this watcr is quite good. Therefore, it is better to recycle it back
,

-[
to the lilter intluent rather than directing it to the wash-waste holding tank and
;, thereby reducing the effective holding capacity of the tank.
.., 0
Access
hole
-
on
~ Example 3 Selecting the Proper Combination of Filter Media for Multimedia
Beds
Lf C\ Central gullet
0

on
0 [)10' 0 r
,
A
Given A dual-media filter bed is to be designed. However, the local regulatory

-[
agency has limited the effective size of the sand to 0.45 mm and the uniformity co-
. r-- I efficient to less than 1.7. The available tilter medium, anthracite coal, has a specitic
0
. ....... Ught and
gravity of 1.63, determined by the AWWA's standard test method .
- - 0
on
air in take
hoi
! • Determine The effective sizc and uniformity coefficient for the coal media

S'-O'
I
\ II , tw l Solution First, the design engineer should select the lower uniformity coetlicient
of sand; this figure may be obtained from the supplier. Once this is obtained, the
\ Down openmg slide gate (typlca~
optimum tilter performance can be anticipated and the sand and coal mixing zone
as effluent level control weir
can be reduced. Any supplier of filter media should be able to provide sand and coal
(a) with a uniformity coefficient of 1.4 without increasing the unit cost. However, a uni-
formity coefficient of 1.3 is now available with a slight increase in unit cost.
Top of filter EI 2500 Engineers should set the sand specifications at an effective size of 0.45 mm, a uni-
Building jail
f/ ~I formity coefticient of 1.5, and a specific gravity of 2.63. The matching coal size

EI. IS.\ Top 01 wall 12.33

EI, \0,25 -- HWL (EI. 21.0) n
~
-"-. I
0
may be determined by either computation or by using the nomograph:

I Top 01 lor.bay wall (EI. 16."'> I~I. 16.50 !!J.. = (~)0.667

~•
d2 PI - P
\ LL. .& LWL (EI. 13.5) =
/ 6TOPOIOUgh: EI. 8. 33 ~ Pipe gallery where do = 0.45 mm; d 2 is the size that must be determined; PI = 2.63, the den-
sity of particles of size d,; P2 = 1.63, the density of particles of size d,; and P =
0 0 tiiJ I 1.0, the density of the tluid.
0 I ~' Lower gullet ~
'
r
1=
/1.0;00
0.45 = ( 1.63 - : )0.667
d2 2,63 -
Effluent 'Channel

(b)
ts I

or d2 = 0.85 mm, the effective size of coal.

Alternatively, Figure 3.2.7-4 may be used. First, locate the size of 0.675 on the
Figure 3.2.7-19 Filter structure design: (n) plan at elevation 17.00; (b) section A-A.
horizontal scale. Then move vertically to a specific gravity of 2.63, which may be
established between 2.60 and 2.65 for sand. Move horizontally to the right to a spe-
cilic gravity of 1.63, which is assumed to be between 1.6 and 1.65 of anthracite
coal. Finally, move down vertically to the horizontal scale and tind 1.27, The ef-
fective size is therefore 1.27 ~ 1.5 = 0,85 mm.
260 DESIGN OF BASIC TREATMENT PROCESS UNITS
Operation and Maintenance In this section we discuss the various aspects of
(I) evaluating regular filter performance, (2) optimizing the pretreatment process
(3) performing jar test procedures, and (4) evaluating and adjusting the filter wash:
ing process. Other important operational parameters and common filter problems
are also covered.
Regular Filter Performance To optimize the filtration process, plant operators
should routinely evaluate filter performance. The three indicators used in evaluat-
ing plant performance are the filtered water turbidity, the length of the filter run
and the ratio of the volume of backwash water used to the volume of filtered wa:
ter. Under normal conditions, the filtered water turbidity should always be less than
0.1 ntu, with 0.1 ntu as the target turbidity level. The second item, filter run length,
is a function of the size of filter medium and the influent water quality, which in
turn is a function of the solids content, water temperature, filtration rate, and the
condition of the filter bed. A treatment plant that is both well designed and oper-
ated should have a filter run length longer than 24 h whenever the filtration rate is
less than 6 gpmlft2 (15 mlh). If a reverse-graded filter bed is selected, the filters
should be run for an average of 2 days at a rate of 4 to 6 gpm/ft2 (10 to 15 mth),
as long as the treatment process is not a direct or in-line type of filtration.
Shorter filter run lengths can be attributed to eight major factors: the accumula-
tion of fine media particles on the surface of the filter bed, an effective size of fil-
ter medium that is too small for the filtration rate, too much floc and suspended
matter (in the filter influent), a filter bed that is dirty and full of mud balls, and over-
abundance of filter-clogging algae in the water source, air binding, a higher than
optimum dosage of polymer (as either a flocculant aid or a filter aid), and recycling
substandard clarified wash waste to the filter influent. With respect to direct and in-
line filtration, extremely short filter runs can easily be induced by unusually high
levels of suspended matter in the raw water ancVor overfeeding of coagulant.
One of the best indicators of filter performance is the ratio of the water used for
filter washing to the amount of water that is filtered prior to filter washing. Under
normal conditions, this ratio is less than 3%; a ratio that is less than 2% is consid-
ered to be very good. If the ratio exceeds 5%, the filter performance is judged poor.
Operators should realize that filter performance is much better during months with
warm ambient temperatures, due to faster floc formation and quicker floc settling
in the settling tanks. A well-maintained filter may have a ratio of 2.5 to 3% during
the winter but only 1.5 to 2% in the summer.
Another indicator of filter performance is the volume of water that is filtered per
unit area of filter bed during a filter run. This is also referred to as the unit filter
rUII volume (UFRV), which is the product of the filtration rate (gpm/ft2) and the fil-
ter run length (minutes) and is expressed as gallons per square foot of filter bed
area. In general, a UFRV that is less than 5000 gal (200 m3 /m 2 ) is unacceptable
even with direct filtration because of the extremely short filter run length. A UFRV
that is over 10,000 gal (410 m3/m 2 ) is indicative of better performance, and over
15,000 gal (610 m 3 /m 2 ) is good filter performance for conventional complete treat-
ment processes.
BASIC APPROACH 261
Pretreatment Optimization Plant operators have two methods by which they can
optimize filter performance. The first involves the coagulant, while the second con-
cerns the mixing energy level and mixing time. In both cases, jar tests are used to
evaluate the various factors, especially for the conventional treatment process. To
optimize the coagulant, the operator must evaluate both the type of coagulant and
the dosage. Alum and polymerized aluminum chloride generally do not produce floc
that is physically strong enough to endure high-rate filters. Therefore, small amounts
(15 to 30 ILglL) of non ionic polymer are usually added to strengthen the floc and
to condition the filter beds. Ferric iron salts, such as ferric chloride or ferric sulfate,
should always be evaluated as a primary coagulant whenever the raw water has a
pH that is over 7.5 (high) and an alkalinity that is over 50 mglL. The smaller dose
of ferric coagulant produces floc that is both physically stronger and heavier than
those produced with alum. Moreover, ferric floc is virtually insoluble in water over
a wide range of pH: a range of 5 to II for ferric floc, compared to 5.5 to 7.2 for
alum floc.
Whenever the required dosage of metallic coagulant is low (less than 6 mglL),
the solubility of the floc becomes an important factor in the filtered water quality.
Many treatment plants using a low dosage of alum in treating water with a high pH
exhibit high concentrations of residual aluminum in the filtered water. Treatment
plants that practice direct filtration or in-line filtration should therefore give special
consideration to this issue. Plant operators should remember that a high dosage of
alum during periods of high turbidity also creates a high concentration of residual
alum because of suppressed pH (a pH that is less than 5.5). Consequently, proper
pH control by alkali chemicals is required during the flocculation process.
Finally, evaluation of cationic polymers, approved by the EPA for the purpose of
water treatment, is essential in selecting the type of coagulant to be employed. The
positive aspects of the cationic polymers are that they are not affected by the pH of
the regular process water; they produce a tougher floc; and the sludge is easier to
handle than those produced by alum. Since the effective dosage is small, rarely ex-
ceeding 2 mg/L, the volume of sludge that must be handled and disposed will be
reduced.
In many cases, a combination of alum and cationic polymer is used because of
better clarification and cost-effectiveness. Most cationic polymers cost approxi-
mately eight times more per unit weight of the product than alum. Consequently,
the polymer dosage is usually limited to 0.5 mglL and is supplemented by 2 to
4 mglL of alum. This combination provides both good performance and cost-
effectiveness when applied to a direct filtration process. As indicated earlier, final
selection of the appropriate coagulant is a function of the effectiveness of the chem-
ical(s) as a coagulant, the cost of treating the water, problems associated with sludge
disposal, and the ease with which the coagulant is handled by plant operators.
The second important factor in optimizing the coagulant is determination of the
optimum dosage. Generally, the optimum coagulant dosage for a water treatment
plant is the dosage that results in a filter influent turbidity that is less than I ntu, or
the dosage that produces the lowest supernatant turbidity after jar testing. Due to cost
considerations, plant operators usually select the minimum dosage required to pro-
262 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 263

duce the second lowest supernatant turbidity (based on jar test results). This practice
is justifiable for rapid sand filtration because a straight relationship exists between
the filtered water turbidity and the settled water turbidity. However, modern filter
beds, such as the reverse-graded beds and coarse deep beds, require a very different
way of thinking. The modern high-rate filters require polymer to be applied to the
filter influent with mixing to maximize their efficiency; the second flocculation (in-
cidental flocculation) actually occurs in the filter inlet channels, pipes, and Ii Iter bed
(primarily in the filter bed). The polymer may be used with or without metal salt co-
agulants, such as alum, and the dosage that produces a supernatant turbidity (after
jar testing) of less than 1 ntu is generally not the best for the overall lilter perfor-
mance. Data obtained from operational high-rate filters indicate that the optimum co-
agulant dosage, that which produces properly coagulated water (does not produce
large floc), is only one-half to one-third the optimum dosage required by rapid sand
filtration. The proper coagulant dosage for high-rate filters generally yields a super-
natant turbidity of 2 to 4 ntu by the jar test. Figure 3.2.7-20 summarizes the relation
between proper coagulant dosage and types of filter (treatment process).
Turbid Pinpoint
water
floc
The most difficult situation in selecting the optimum coagulant dosage is en-
countered with the in-line and direct filtration process. The dosage at which pin-
point floc is produced is already an overdose that will induce a short filter run.
Several equipment manufacturers and researchers have suggested that for these types
of filtration process, plant operators use pilot filter column tests to determine the
proper coagulant dosage. However, experience has shown that very few operators
of plants equipped with the pilot filter columns use them regularly enough to war-
rant the cost. Plant operators respond that the pilot columns are simply not practi-
cal enough.
Another option is to use a small glass tube pilot filter. This type of test filter is
composed of a half dozen l-in.-diameter glass tubes. The glass tubes are filled with
small-grained sand to a depth of 1 ft (0.3 m). The filterability of the supernatant,
obtained from jar tests, is determined by passing it through the pilot filter. Yet plant
operators rarely employ this apparatus on a regular basis because it is both unreli-
able and time consuming. The small glass pilot filter yields unreliable data because
the sand in the glass columns is not ripened, as in actual filter beds. An alternative
to the sand has been employed by the author and others. In this case, filter paper
(Whatman No.1) is used as the filter medium. This option is much quicker and the
results correlate reasonably well with actual filter bed performance. It must be em-
phasized that the ordinary jar test procedure is a good control tool for the plant

IT I 'l
Direct
filter J I
.'...
......
:~
'.'
process, but only for conventional processes. It does not work well for direct or in-
direct filtration, unless the flocculated water is first filtered through filter paper, such
as Whatman No.1 or a similar type of paper with a pore size of approximately 10
11m. Incidentally, the pore size of regular rapid sand filter beds is considered to be
20 approximately 15 11m.
~ Two-stage ~ The second method of optimizing filter performance involves the mixing energy
filtration
r-- Conventional level and the mixing time (G X f). Although these parameters are supplied to the
process ---1 operator by the design engineer, there is still some room for adjustment.
15
5
>- Recommended Jar Test Procedures Although some plant operators believe
~
~ that jar testing does not yield reliable results, it is a useful test for controlling co-
'C
agulant dosage. Factors such as mixing conditions, alum solution strength, method
:e
~
10
of chemical feeding, and the flash mixing system all affect the correlation between
,
m
the jar test results and actual plant performance.
U
.~

~
If the mixing conditions (G X f) of the jar tests are not similar to those of the
a:
5 actual process, the results may be quite different. Most modern treatment plants use
a 50% solution of liquid alum. However, a 1% solution of liquid alum is often used
in the jar tests, potentially distorting the correlation between the jar test and the ac-
tual plant process. Experience has shown that a 0.1 % a solution of alum is better
oo;-----~5------7.10~----~1~5----~2~0----~25 suited for the jar tests as long as it is freshly made (Kawamura, 1973). The feeding

FeCI, dosage (mgIL)

.. condition of the alum (i.e., continuous feed or intermittent feed), the actual type of
flash mixing system, and the presence of flow short-circuiting in the actual tank also
or alum dosage greatly affect the correlation between the jar test results and plant performance. Refer
Figure 3.2.7-20 Proper coagulant dosage range for the three treatment processes. to Appendix 10 for a recommended jar test procedure.
 BASIC APPROACH 265
264 DESIGN OF BASIC TREATMENT PROCESS UNITS

Evaluation and Adjustment of Filter Washing Procedure To optimize the fil- 700
tration process, the condition of the filter bed must be evaluated routinely. Operators
may maintain the proper bed conditions by adjusting the Ii Iter washing procedure. Clean filter 2
5 min backwash (lS.5 gpm/ft )
There are three basic methods by which to cvaluate the effectiveness of the filter 600 with 3 min surface wash (both
backwash and surface w~sh together
washing procedure: (I) visually inspecting the filter bed before and after filter wash- in the first 3 mIn)
ing, (2) measuring the turbidity of the backwash waste at I-min intervals after ini-
tiating backwash, and (3) core sampling of the Ii Iter bed both before and after fil- 500
ter washing.
A lilter bed that is properly maintained is clean (i.e., practically devoid of mud ...\
balls) but is also in ripened condition; the media grains of a ripened filter bed are
400 \
coated with a proper amount of floc or polymer. During visual inspection of the fil-
t \
tcr bed, plant operators must look for mud balls, cracks, and any other unusual con- I

V
\ Oirty filter
ditions. The bed is in poor condition, with mud balls and mud accumulation, if the t \ 5 min backwash only
I at 18 5 gpmlft'
operator linds lumps and cracks on the surface of the filter bed prior to filter wash- 300 I
ing and lumps even after washing. The most obvious indicators of a dirty filter bed I
I \
are the existence of worms and debris on its surface, If a tilter cell exhibits any I \
mounding, concaved areas, or obvious shifts in the media, there is either a scour- 200 t •\
I \
ing action by filter influent, problems with the surface washing system, or a maI- I
distribution of the backwash water and boiling action in the filter bed. The opera- \\
tor should immediately investigate the causes of these unfavorable phenomena.
I
I ,,
The second method, measuring the turbidity of the backwash waste, is used to
100
, I ,
.',
evaluate the effectiveness of the filter washing procedure in removing floc from the
Ii Iter bed. A valuable tool in assessing the washing etliciency and the optimum du-
J ---._---
-'" I 5
00 2 3 4
ration of the Ii Iter washing procedure is the turbidity protile of the backwash waste
plotted against wash time. A low profile with a low peak curve is indicative of in- Figure 3.2.7-21 Backwashing waste characteristics (con~entional sand I1lter~). (N~f~: The
effective washing. A high profile curve with a high peak curve is characteristic of water level in the filter cell was lowered approximately 6 \11. above the top at the tlltcr bed
effective washing. Figure 3.2.7-21 illustrates these two types of curves. prior to filter washing.)
Many plants wash their filters for an excessively long period of time, until the
operator can clearly see thc surface of the bed. This method of washing is actual " 000 to 400 .
where G ranges t lom.1
S -I Energy input due to the water jets of surface
.
detrimental to the filtration process because the filter often exhibits a distinct initial
wash is
turbidity breakthrough, for a period of 15 to 30 min or longer, until the bed regains
its ripened condition. Studies based on both pilot-scale and actual filters indicate G = pgv, tip
that this initial turbidity breakthrough may be minimized or even eliminated by ter- ",aL
minating Ii Iter washing when the turbidity of the wash waste ranges from 10 to 15
ntu; this usually occurs after 5 to 6 min of regular washing (not air-scouring wash- where G ranges from 1000 to 1300 S-I. . . '
ing). Adverse effects, such as a shorter filter run length and the accumulation of When backwash is applied to a tilter bcd, the headloss begms to mcrease with
mud within the tilter bcd, will not occur if the lilter wash rate is properly selected . e t'n the wash rate Yet when a specific backwash rate IS reached and the
an mcreas ' . . . d d
and a surface wash is implemented. wash rate continues to increase. the hydraulic headloss will beg.m t~ ecrease ue
The energy input generated by the backwash process, and by either surface wash to fluidization of the Ii Iter bed. This characteristic is illustrated In Figure 3.2.7-22.
or air scour, may be estimated by using the following equations. Energy input due Data obtained from both pilot and actual tilter studics show that the optimum back-
to the backwashing process is wash rate is approximately twice the wash rate at which maximum headloss IS at-
. . . F·
tamed For mstance, 111 19ure...
3 2 7-22 the proper wash rates tor sand II. and sand
I are 2'0 and 15 gpm/ft' (50 and 37.5 mlh), respectively. ~eyond these optimum ~'ates
G= there are no appreciable benelits, primarily because the hlter bed IS highly tluldlzed,
 BASIC APPROACH 267
60
the media grains are no longer in contact with each other and the scrubbing action
of the filter bed is therefore minimized. This same logic can be applied during the
Sand I (W.T. 22.5'C) rinse period after air-scouring wash. However, operational experiences indicate that
the backwash rate at which the maximum headloss is produced is the most effec-
20 tive concurrent wash water rate because the majority of the sludge, attached to the
50 media grains. is stripped during air scouring and transported into the water. The rec-
ommended filter washing procedures, with auxiliary scouring, are illustrated in
Figure 3.2.7-23. The mud deposition profile of Figure 3.2.7-10, both before and af-

40
Surface Wash
0

I
2
I I
Together
3
I ..
,
Surface Washing
(minutes)

.. -I

~~
15
g Backwash I
0
I
1 2
I
3
I I
4 5
I I
6
.,
"0
.D
Backwash
., Normal E (minutes)
£ backwash l:.
Notes: (1) Draw down water level to approx. 4 inches (10 em) above the bed
""
0
range
30 .Q" before surtace wash starts.
~ "0
u
~ (2) Longer concurrent wash will result in higher media loss.
"'"
J:
10
2" (a)
"0
.,'"
J:
Air Purge and 40
20 Re-stratification Washing
Legend
pause
E.S. (mm) UC. L (in.) (option) .. I
Sand I 0.52 1.6 24
5 Sand II 0.64 1.3 24
10

o 5 10 15
Washing Duration (minutes)

Notes: (1) Draw down water level to approx. 4 inches (10 cm) above the bed.

00 0 (2) The first slow backwashing must be stopped before the

10 20 30 40 50 water level gets up to six cinches below the top waste weir.
Backwash rate (in./min)
(3) Backwash in rinsing stage would require approx. 18 gpm/sf (45 m/h)
I to purge air in coarse media bed as well as re-stratification of dual
0 5 10 15 20 25 30 media bed.
Backwash rate (gpm/ft2) (b)
FigureJ.2.7·22 Relation between backwash rate and headloss. (Adapted from P. A. Norman, Figure 3.2.7-23 Recommended filter washing sequence: (a) recommended surface and
"Backwash Instability in Rapid Sand Filters," M.S. thesis. University of New South Wales backwash sequence; (b) normal concurrent air and water air-scouring wash for common-size
Australia, 1970.) , media bed.
266
 BASIC APPROACH 269
268 DESIGN OF BASIC TREATMENT PROCESS UNITS
,
ter filter washing, clearly demonstrates the effectiveness of the washing procedure.
-- ----
Based on the discussion above, plant operators should be able to evaluate their pres-
--B ,--: ~
,- ---- - --- ---- -- -- ,
ent washing practices and make the necessary adjustments. In most cases the filter ~
bed expansion rate during backwash ranges from 20 to 35%. However, bed expan-
sion is less important if a surface wash system is installed; when air-scouring wash "
j
l '\ , ,, PtloN f"tf w .. ~

,
is utilized, the importance of the bed expansion is almost diminished. -- 'I~
h
~
"'- ~, -
The third method used to evaluate the effectiveness of washing conditions is core
sampling, which allows plant operators to establish a mud prolile for the entire depth
" --B- --
;:,;. ':.'
•
,~
of the filter bed. Core samples should be obtained from selected Ii Iter beds at regular
intervals according to a rigorous schedule. This procedure is very important because
. --- - -,.-- -
,,
,
-
it allows the operator to check the actual depth of the medium, to obtain a medium-
]
'x'
r ... ,,o.k S<<>uI1r'oJ wa,h
"
L
sized distribution profile across the entire depth of the filter bed, to evaluate filter •
,/
washing efficiency, to check lilter bed conditions, and to evaluate the movement of /' ,
-
f.lter gravel. These data can only be obtained by taking core samples from multiple
locations within the filter cell. Moreover, it is crucial to obtain samples at various
depths from the same location. If one core sample is taken across the entire depth of • ,~ ,.~ 3,000 '.~ '.- 10,000

the filter bed, from one location, it may not be representative of the tilter bed since it
Amount of Floe I Sludge Retained (Turbidity in NTU) on 100 ml of Filter Media
does not show the accumulation of fine medium at the surface of the bed. The rec-
ommended sampling techniques and a simple quantitative analysis of the mud that ac- Figure 3.2.7-24 Floc retention profile for a filter bed with a large coal and sand mixed zone,
cumulates on the surface of the filter bed are described in Appendix II.
washing conditions must be reevaluated. Turbidity above 300 ntu is indicative of a
Mud Deposition Profile Both the effectiveness of the filter backwash process and mud ball problem. Figure 3.2.7-25 can be used to judge the cleanli~ess of the filter
the condition of the filter bed may be clearly evaluated from the mud protile graphs. bed. The mud deposition analysis procedure .s descnbed m AppendIX II.
For example. a filter without a surface washing system cannot adequately wash out When mud balls are found in the filter bed, measure the percent volume of mud
floc that is retained during a filter cycle, thus allowing the sludge to accumulate balls in the bed. If the percentage is less than 0.1 %, the filter bed is clean. If the
slowly. Conversely, a filter with a surface wash system is effectively cleaned through- mud balls are 0.1 to 0.5% by volume, the filter bed is in good conditlO~; 0.5 to 1.0%
out the entire depth of the bed, as indicated in Figure 3.2.7-10. Many case histories is indicative of a fairly clean bed; 1 to 5% indicates that the filter bed .s m bad con-
have shown that both regular rapid sand filters and standard dual-media lilters can
be kept clean with installation of the proper surface wash system and application of
the correct backwash rate. Improperly designed dual-media lilters with air-scouring Very cleanlbed
No! well ripened
wash systems have problems with mud accumulation in the coal and sand mixed
zone. For example, when the size of the two filter media is not hydraulically matched
\
C',," bw \
Pjlf1iai,ly
and the backwash rate after air scouring is inadequate to fluidize the bed (such as Ripened Reasonabl
clean
12 gpm/ft2 or 30 m/h), a large portion of the coal and sand mixed zone will develop Hipened

heavy mud accumulation. Figure 3.2.7-24 is an example of this problem at a lime Dlrl~ bed

softening plant in California. Well ripened

A filter bed that is properly conditioned is also in a ripened stage; each grain of ~ud
0'" ""
balls \
Well ripened
'ro.""" "~"'''"''''''Y
scouring in addition to badlwash
media is coated with a thin film of coagulant hydroxide or polymer. A bed that is too
clean usually exhibits a distinct turbidity breakthrough at the beginning of each fil- V9r~dlrt~~
ter cycle, lasting anywhere from 30 to 60 min. Based on the author's experience, a M'"Y~T"
turbidity ono to 60 ntu, determined after filter washing by the sludge retention analy- NEe~~~~I~$~~c
sis test method, is indicative of a clean and ripened bed if polymer is continually fed 3000 4000 5000

to the filter intluent. A filter bed that is slightly dirty (less than ideal but no need for 10 20
'" " 50 "'0 "" "'0
400 500 '000
""'"
floe deposition on 100 mI 01 IiIler media artar liller wash (NTU)
concern) will have a turbidity between 60 and 120 ntu. However, a lilter bed ex-
Figure 3.2.7-25 Filter bed condition range: cleanliness evaluation,
hibiting turbidity over 120 ntu is dirty, and both the Ii Iter washing system and filter

270 DESIGN OF BASIC TREATMENT PROCESS UNITS
dition. If the volume of mud balls is over 5%, the bed must be replaced with new
media.
Restoration of the Filter Bed Filter bed medium is commonly lost after repeated
filter washings. Rapid sand filters generally do not lose a significant amount of sand'
a normal loss of I to 2% of the total bed depth (original) occurs annually. Yet th~
loss of anthracite coal is much more significant; an annual loss of 5 to 7% is not
unusual for ordinary dual- or trimedia beds.
Excessive media loss is often caused for the following reasons: air binding in the
filter bed; an unnecessarily long and overlapping surface wash and backwash (over
3 min); an excessive backwash rate, especially during months of cold water tem-
peratures; a mismatch in the size of the coal and sand; a wash trough elevation that
is too low; improper sequential control of both the air-scouring and rinsing back-
wash processes; and leakage of the tilter media through the underdrain system. If
the media loss is excessive, the plant operator should first locate and correct the
causc. The original depth of thc filter bed should be restored by adding the appro-
priate media to the surface of the tilter bed and backwashing the filters a minimum
of two runs; this will fluidize the bed and ensure a proper gradation profile.
Many engineers and plant operators ask what the tolerable amount of media loss
should be. Losing up to 20% of the original depth, or a loss of 6 in. (15 cm) in
depth, whichever is attained first, is generally a tolerable loss if total bed depth is
approximately 30 in. (76 cm). Filters with surface wash systems begin losing the
advantages of surface washing when the distance between the top of the bed and
the elevation of the jet nozzlcs exceeds 6 in. Filter performance also begins to de-
crease when media loss is greater than 20% of the original depth.
The purchase of replacement sand for rapid sand filters is quite simple; specify
the same size medium grain as the original but with a lower uniformity coefficient
(1.5), to avoid the short tilter run lengths characteristic of tine medium grains.
However, dual- and multimedia beds must have the replacement media properly
matched in physical characteristics so that all the layer will fluidize to the same de-
gree during backwash. The specific gravity and size of both anthracite coal and gar-
net grains are very important considerations. Based on past experience, the specific
gravity of sand is always near 2.63, but this is not the case for commercially avail-
able anthracite coal and garnet grains. Over the past 30 years the specitic gravity of
anthracite coal, when used as filtering material, has ranged from 1.50 to 1.72.
Therefore, the size of the available anthracite coal must be adjusted to match that
of the sand. Figure 3.2.7-4 will help determine the complementary media sizes. First,
locate the representative sand size (0.55 X 1.5 = 0.83 mm). Then move vertically
up to the specific gravity of the sand (2.63). Next, move horizontally to the right,
to a specific gravity of 1.65 (assuming that this is the available type of anthracite),
and vertically down to 1.53, the represcntative particle size of the coal. Assuming
that the uniformity coefticient is specitied as 1.5, we tind that the effective size of
the coal is
1.53 1.5 = 1.02 mm, say, I mm
BASIC APPROACH 271
The following formula may also be used to compute the matched size of two dif-
ferent types of tilter bed medium:
0.55 = ( 1.65 - : )0.667
d2 2.63 -
or d, = 1.015: say, 1.0 mm.
If other kinds of filter material, such as garnet or GAe, are to be used in con-
·unction with sand, thc appropriate media sizes may be found in the same manner.
·~s was demonstrated in Example I, it is imperative that the media specitications
not allow for a wide range of media sizes and specific gravities.
Gravel Bed Upset in the Filter Major disturbances in the gravel bed arc usually
caused by errors in the design phasc, construction phase, or during operation. Plant
operators have no control over the first two. Improper design is th~ _res~ons!bililY
of the design engineer. The obvious design tlaws are improper speclticatlOn for the
gradation and thickness of each gravel layer. Based on case hist~ries, disturbances
in the gravel bed and leakage of tilter media through the supporting gravel bed are
generally caused by improper construction of the gravel bed, including the omis-
sion of one or even two layers. Gravel bed upsets are also caused by Improper op-
eration of the backwash process.
The most common operational error is to quickly initiate full-scale backwashing
for a filter containing no water. This type of error usually occurs when the filter is be-
ing filled with water for the first time by someone who docs not have adequate kn~)~l­
edge of the tilter, by the contractor, or by an inexperienced operator. The compost.tlOn
of the gravel bed may also be disturbed by backtilling a tilter that has been drained
at full backwashing rate without slowly operating the backwash valve. In either case,
the combination of the quickly escaping compressed air (in the tiltcr underdrain sys-
tem), the short-circuiting of the water, and the escape of air can cause major distur-
bances in the arrangement of the gravel bed. It may evcn damage or break the un-
dcrdrain system. Once the gravel bed is disturbed, it is irreversible until reconstructed.
It is therefore essential to backtil1 the filters slowly until the water level covers the
2
surface of the tilter bed. The initial backfilling rate should not exceed 5 gallft (12.5
mlh). The backwash valves must also be opened very slowly to avoid lifting the tiI-
ter media bed off the gravel layer. Otherwise, an open space will be formed between
the media bed and the gravel bed. If this should occur, the filter bed will break at sev-
eral points, causing a boiling action, and the upper portion of ~he sup~ol:ting g~a:vel
bed will be disturbed (Figure 3.2.7-26). If the plant operator notices a dIStinct boilmg
action during tilter back washing, the filter must be draincd and t~e condition of ~oth
the gravel bed and the inside of the underdrain system must be mspected. The hiler
bed should be dug up by using a 4 ft X 4 ft (1.2 m X 1.2 m) bottomless box as a
shoring frame. Failure of the gravel bed is also clearly evidenced when tilter media
accumulation is found both inside and downstrcam of the underdrain system.
In the mid-twentieth century, John Baylis proposed the use of a reverse-graded
gravel bed. Although this type of bed is etlective in preventing the movement of
 272 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 273

Expanded sand level Water surface

Underdrain

Pressure (It. water gauge)

(a) Pressure versuS depth in a gravity filter at various times

floor

water level

35- -
Present operating level

Hydraulic gradient: no flow

g 5.8'
Figure 3.2.7 26 Filter gravel bed upset by operational error. (From AWWA, Water Quality 3.8'
.~
a

and Treatment, 3rd ed., McGraw·HiIl. New York, 1971.)

~
iii 30- =,.~_ _L-_ _ _- I
gravel, it has problems with mud accumulation in the fine gravel layers of the gravel
bed (middle layer) for ordinary backwashing filters.
"--,,,I~!:~~:~!~ic gradient
Air Binding Air binding is a phenomenon in which large amounts of air bubbles 'ilter clogged
accumulate in the tilter bed. Air dissolves in the water near or even over the satu-
ration point. The pressure in the upper portion of the filter bed may be reduced to
Zone of negative head
below atmospheric pressure due to the creation of negative pressure during the lat- causing air-binding
ter stages of filtration. When this occurs, the air in the water no longer remains dis-
solved but accumulates within the filter bed, rapidly increasing the headloss through- (b) ·Causes of air-binding

out the bed. Figure 3.2.7-27 illustrates the development of negative pressure within Figure 3.2.7-27 Air binding in the filter bed. {(b) From H. E. Hudson, Water Clarification
the filter bed. If a high water level is maintained in the filter, air binding will be de- Process, Van Nostrand Reinhold, New York, 1981.]
layed. Therefore, a shallow filter with a water depth that is 4 to 5 ft (1.2 to 1.5 m)
above the filter bed and a filter that lacks an etlluent water control weir tend to pro-
 274 DESIGN OF BASIC TREATMENT PROCESS UNITS
mote air binding. This type of filter is shown in Figure 3.2.7-1 and represents a con-
stant-rate tiltration scheme.
When air binding, occurs in th~ filter bed, gaseous air violently escapes when the
filter effluent valve IS closed. ThIs phenomenon causes media loss. Therefore, the
water In _the filter should be lowered to a level below the top of the wash troughs,
and the hlter waste valve should not be opened until the majority of the air has es-
caped from the filter bed. Moreover, Ii Iter washing should be delayed until most of
t~e 31r has escaped. The degree of air binding may be reduced or even eliminated
If filter washIng IS frequently initiated whenever the head loss reaches 4 to 5 ft (1.2
to 1.5 m). This practIce prevents the creation of negative pressure in any part of the
filt~r bed. In many cases, air binding tends to occur during certain seasons such as
sprrng: when there is a high degree of dissolved air. Air binding may also b~ a func-
tIOn of the degree ?f aeration due to hydraulic power plant activity upstream of the
plant, sUctIOn of aIr ,from air vacuum protection valves in the raw water main, or
the degree of cascadIng at the intake.
Air binding will not occur if the lilters are designed properly. For example,
the dechnrng-rate filter~ and the rising level self-backwash filters shown in Figure
3.2.7-1 can ehmInate alf bIndIng due to a high water depth (above the bed) andlor
because the effluent water level control weir of the clearwell is situated high enough
so that negative pressure is not created during a normal filtering cycle.
Filter Performance Surveillance The four basic tools used in monitoring filter
performance are the tur?ldlmet~r. pH meter, particle sizing and counting equipment,
and filter headloss prohle monitor. Each type is discussed in detail. Turbidimeters
are used not only to monitor the filter effluent turbidity of individual lilters but also
to obtaIn samples from the middle (depth) of the filter bed and the lilter wash waste.
MOnItonng the turbidity of the water at the coal-sand interface, located approxi-
mately halfway down the filter bed, is recommended because the turbidity of this
water can give the operator advance warning of a turbidity breakthrough across the
filter bed. The filter turbidity of the wash waste should be measured because it is a
good parameter for adjusting the duration of filter washing.
. ~ephelometric measurement of turbidity is required because the standard tur-
b~dlty measurement in the United States is expressed in ntu. The nephelometric tech-
nique measures the degree by which the incident light beam is scattered by the sus-
pended particles; .the hght beam is at a right angle to the particles. Several companies
manufacture turb,d,meters, but the most popular is produced by the Hach Company
(United States). Extreme care must be taken not to introduce any air bubbles to the
turbldll~eter because this will cause inaccurate readings. Each individual tilter ef-
Iluent lIne should be monitored by a reliable turbidimeter that is also capable of
measurIng very low levels of turbidity; the new EPA drinking water standards de-
~and a turbIdity of less than 0.3 ntu 95% of the time. If this primary standard is
VIOlated, a harsh penalty will be enforced by the EPA.
The pH meter is indirectly related to filter performance because it detects the po-
tential for corrosIOn and gives early warning of other problems in filter performance.
If the pH of the filtered water is above 8, a significant amount of dissolved alu-
BASIC APPROACH 275
minum may break through the filter bed. The same situation will occur if the pH is
below 5. A low pH also gives rise to corrosion problems, and special attention should
be directed to the problem of lead corrosion. A high filtered water pH will also re-
duce the effectiveness of chlorine disinfection. It is standard practice to measure the
pH of the combined filtered water for the reasons given above.
A unique tool in evaluating filter performance, in conjunction with the tur-
bidimeter, is particle sizing and counting equipment. Such an analyzer is especially
valuable in evaluating the filtered water characteristics that cannot be detected by
the turbidimeter. For example, the turbidity readings may be the same for the fil-
tered water of plant A and plant B. Yet the particle size and the number of particles
in each of the waters may be quite different. Since the design of the turbidimeter is
based on nephelometry-that is, it measures the degree to which the incident light
is scattered by the suspended particles-the size, shape, and color of the suspended
particles greatly affect the reading. The turbidimeter therefore is incapable of mea-
suring the true content of dark suspended particles such as PAC. However, a good
correlation generally exists between the turbidity and the particle count (Figure
3.2.7-29). Based on the HiaclRoyco Model 8QOO, MET ONE Model 210, HACH
1900, or equivalent units, particle counts of less than 50 mL -I are generally con-
sidered excellent. Particle counts (larger than 2 I'm) of 50 to 150 mL -I are good,
but counts over 200 are unfavorable.
The newest particle sizing and counting equipment sold by Met One, for exam-
ple, is able to measure in the range 0.5 to 15 I'm. This is important because most
protozoa (Giardia and Cryp/Osporidillm) are in this size range, but most filters with
granular beds are considered to be somewhat ineffective in removing the particu-
lates in this range. These units are rather expensive, over $20,000 per unit; they are
actually best suited for study and research and may be too expensive to be used in
each filter as a surveillance tool for daily filter performance. However, it is advan-
tageous to have one unit in the plant process control laboratory or on one selected
filter.
The filter headloss profile monitor supplies the plant operator with a headloss
profile across the filter bed. The monitor should be arranged so that there are sev-
eral pressure taps at various depths of the filter bed. These pressure taps should be
connected to glass tubes, creating a piezometer board. If high-tech equipment is in-
cluded in the plant design, the headloss profile may be shown on a cathode ray tube
(CRT) in the control room, thus allowing for continuous observation and giving early
warning of air-binding conditions. Although this unit is a good performance moni-
toring tool, it is not necessary for all the filters to be supplied with one. One select
filter should be designed to receive this unit if the plant owner expresses interest in
its use.
Common Filter Behavior and Control Measures In contrast to all the other
processes in an ordinary water treatment process train, the filtration process is gen-
erally a batch process. The filtering cycle usually lasts from I to 3 days, followed
by filter washing. The washing cycle lasts a total of 15 to 30 min, including the
drawdown of water level prior to washing and refilling with water. After the wash-
 BASIC APPROACH 277
276 DESIGN OF BASIC TREATMENT PROCESS UNITS

ing cycle is completed, the lilter cycle is repeated. The special performance behav- 0.'
ior of the !ilter may be attributed to the batch operation process. As illustrated in
FILTER A
0.'
Figures 3.2.7-28 and 3.2.7-29, the filtered water often exhibits a sharp turbidity
breakthrough at the beginning of the lilter cycle. It is usually short, 10 to 20 min.
After this period the turbidity of the filtered water drops below 0.3 ntu and eventu-
ally settles in the vicinity of 0.15 ntu; this level is both the practical and the upper
"~
0.3

0.2
.J~- -0

turbidity limit delined by most regulatory agencies. However, the turbidity of the
liltered water gradually increases with time and exceeds 0.3 ntu after 24 h. The tur-
bidity breakthrough at both the beginning and late stages of a liltering cycle is char-
0.1
0 10
" ,0
nmtl (min.)
. so .
acteristic of most constant-rate tilters employing alum coagulation. Design engi-
neers and plant operators must be cognizant of this particular filter behavior and
0.5
should analyze the methods used to control the turbidity breakthrough.
The initial turbidity breakthrough may be controlled by any of four methods: 0.'
wasting the filtered water, slowly opening the Ii Iter emuent valve, adjusting the fil- FILTER B
. 0.3

~
ter washing practice, or adding polymer to the !ilter backwash water. The first two
methods are related to the design of the lilter system. The last two are operational ":z: 0.2

manipulations, except proper design of the polymer feed system to the backwash 0.1 .., • •
water. The lirst method isolates the liltered water from the clearwell for a period of
0.0
10 to 20 min, until the turbidity of the water falls below 0.25 ntu. The second slowly
opens the lilter effluent valve, taking 10 to 20 min to arrive from a fully closed valve
0 10
T1m. (min.)
15
"
position to the appropriate open position, a position that permits the designed fil-
tration rate to occur. The third method, adjusting the filter washing practice, tenni-
nates the filter wash cycle when the turbidity of the wash waste is approximately
10 to IS nlU, thus maintaining the ripened condition of the filter bed. The last method I.'
applies approximately 0.3 mglL of non ionic polymer during the last 3 min of back- 0.'
FILTER C

r~
wash. This last method is more of a trial-and-error type of technique and is not al- 0.'
ways successful because the polymers are very selective with respect to the char- j:

:~
acteristics of the suspended matter. In some cases, the addition of 0.2 to 0.3 mglL :z: 0.'

of cationic polymer to the Ii Iter intluent for 30 min, from the beginning of the fil- 0.2
-0
tration cycle, can control the breakthrough. 0.0
Turbidity breakthrough in the late stages of the filter cycle may be controlled by oa
strengthening the Iloc and increasing the adsorption capability of the filter bed. This
0 10 '0
TIm. (min.' "
may be achieved by feeding cationic polymer as a coagulant, with or without alum,
or by adding minute amounts of nonionic polymer to the filter inlluent Cas a filtra-
tion aid); alum must be the sole coagulant. Plant operators must be careful to limit 1.0
the dosage of the nonionic polymer (filtration aid) to 0.015 to 0.025 mglL, since
0.'
overfeeding will result in extremely short lilter run lengths. According to Figure FILTER D
3.2.7-30, the late-stage turbidity breakthrough of the previous filter is eliminated
completely when the coagulant is switched from alum to cationic polymer. .." :z:
0••

0.'

0.2
Operational Procedures Following Plant Shutdown If the treatment plant is
shut down for more than 3 days during any season Cfor any reason) or if the plant 0.0
tlow rate is only 10 to 20% of the designed rate for a period of I month or more, 0 10 '0
nm~(min.)
,0 40

the majority of the filters may be drained or all the filter effluent valves may be
closed, allowing the water to stand over the filter bed. However, if the plant is shut Figure 3.2.7-28 Turbidity breakthrough at beginning of tilter run.
 278 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 279

I I I 1

1000 1000 I- - 10
10
Notes: (I) Filtration rille .. B gpm/tt l ,
(2) Coagulant: cationic; polymer - 0.2 mg/L;
.Ium ,. O.

'[ E,
0
"',
~
Sl
~
, Particle counts
s
C.
"' .~
cl-' (I) Filtration rale _ 8 Ullnltt l , I:!.
'"- '0

~ 100
(2) Coagulants; Alum. 4 mgll;
no polymer,
-'
~ 100 r- I.O -
:is
....
~

~
1:: ~
l!. ~ 50 0.5
50

- Turbidity

10 '---4~--:-8----:11:----:'="6--,!20"""""--,l240 .1
2 1
Time since start (h) -
Continuous monitoring 01 filter eHluent
Figure 3.2.7-29 Filter performance evaluation by turbidity and particle counts with con-
tinuous monitoring of filter effluent (Adapted from MWD Southern California, (nlernal
Report. 1978.)
down for only 2 or 3 weeks, the filters should not be drained because it will take a
considerable period of time to regain the proper (mature) filter bed conditions if the
bed has been allowed to dry up. When a plant is treating an extremely low flow rate
for long periods of time, a few operators choose to use only one or two filters and
the rest are left on standby; these standby filters contain standing water. However,
it is better to use all the filters, allowing water to move through all the filters, since
this will prevent an explosive growth of microorganisms in the Ii Iter bed and will
prevent anaerobic conditions from occurring in the tilter bed. Anaerobic conditions
often lead to objectionable taste and odor problems as well as the reduction of ox-
idized iron and manganese in the filter bed. In any case, if the filters are left stand-
ing for over 3 days without being drained (during warm months), they should be
backwashed prior to reintroduction into the filtration cycle. If the lilters cannot be
backwashcd, due to an insufficient amount of water, it is essential to drain the stand-
ing water completely and to backwash the filters for approximately 3 min or more
after backfilling with backwash water. Since a rate of 2 gpm/ft 2 (5 m/h) is equiva-
lent to 3.2 in.lmin (8 cm/min), it would take approximately 30 min to drain a filter
completely, assuming that the water depth in the filter is 8 ft (2.4 m) and the aver-
age rate of the drain is 16 ftlhr (5 m/h).
1
10 L--41---!S'---'1:1:::--,--,11'6::---:;!::0--::24 0 .1
2
Time since start (h) -
Continuous monitoring of filter effluent
Figure 3.2.7-30 No turbidity breakthrough except at beginning. There is continuous mon-
itoring of filter effluent. (Adapted from MWD Southern California, Internal Report, 1978.)
If the filter underdrain system or the gravel beds are reconstructed or disturbed
by workers, the filter-to-waste line should also be employed. For example, if a por-
tion of the filter underdrain is repaired, a filtered water turbidity of almost 30 ntu or
more may result from the disturbance of the gravel bed, which is usually quit~ filthy
after several years of service. The design of the filter system should therefore Include
a filter-to-waste line, even though it may not be used for normal daily operations.
Liquefaction of Filter Bed Due to Strong Earthquakes An interesting and im-
portant behavior of granular medium filters is the potential for high-turbidity break-
through during strong earthquakes (magnitudes above 5) due to scrubbing and liq-
uefaction phenomena. Plant operators should be aware of this fact, and whenever
these phenomena are observed, they should (I) stop plant operations, if possible,
or (2) route the filter effluent through the filter-to-waste line. The importance of
these emergency procedures is illustrated by the case of two modern water treat-
ment plants in southern California, located near the epicenter of a 5.5 magnitude
earthquake, which recorded an effluent turbidity as high as 5.0 ntu ImmedIately
after the quake.
Before a system is put back into operation, each filter should be lightly back-
washed for about 3 min.

280 DESIGN OF BASIC TREATMENT PROCESS UNITS
BIBLIOGRAPHY
Ahmad, R., et aI., "Effect of 8ackwashing on Biological Filters," J. AWWA, 90: 12:62
(December 1998).
AWWA, Water Quality and Treatment, 3rd ed., McGraw· Hill, New York, 1971.
Barrett, E., et al., "Comparing Plant Scale Dual and Mixed Media Filters, J. AWWA, 84:6:76
(June 1992).
Baylis,!. R., "Nalure and Effect of Filler Backwashing," J. AWWA, 51: II: 1433 (November 1959).
Brigano, F. A., et aI., "Dual-Stage Fillration Proves Cost Effective," J. AWWA, 86:5:57 (May
1994).
Clark, S. L., et aI., "Contact Filtration: Particle Size and Ripening," 1. AWWA, 84:12:61
(December 1992).
Cleasby, J. L., "Declining-Rate Filtration;' J. AWWA, 73:9:484 (September 1981).
Cleasby, J. L., et aI., "Effectiveness of Backwashing of Wastewater Filters," 1. Environ. Eng.
Div. ASCE, 104:4:79 (August 1978).
Degremont, Water Treatment Handbook 5th ed,. Halsted Press, New York, 1979.
ForgeJ, D., et aI., "Removing Giardia and CryptosporidilII/l by Slow Sand Filtration,"
J. AWWA, 85:11:77 (November 1993).
Fujita, K., "Hydraulics of Filtration Theory," J. Jpn. WWA, 445:2 (August 1973).
Ginn, T. M., Jr., et aI., "Effects of Particle Detachment in Granular Media Filtration," 1. AWWA,
84:2:66 (February 1992).
Hancock, C. M., et aI., "Assessing Plant Performance Using Microscopic Particulate
Analysis," J. AWWA, 88:12:24 (December 1996).
Hargesheimer, E. E., et aI., "Tracking Filter Performance with Particle Counting," 1. AWWA,
90: 12:32 (December 1998).
Hozalski, R. M., et al., 'TOC Removal in Biological Filtraters," J. AWWA, 87: 12:40 (December
1995).
Hudson, H. E., Water Clarification Processes, Van Nostrand Reinhold, New York, 1981.
Ives, K. J., The Scientific Basis of Filtration, Noordhoff~Leyden, Groningen The Netherlands
1975. ' ,
Kawamura. S., "A Study on Filter Washing Efficiency," Proceedings of the 6th Annual Japan
WWA COllference, May 1955.
Kawamura. S., "A Pilot Study on Filter Bed and Underdrain System," Pmceedillgs of the
14th Annual Japan WWA Conference, May 1963.
Kawamura, S., "Design and Operation of High-Rate Filters," J. AWWA, 91: 12:77 (1999).
Kawamura, S., "Design and Operation of High-Rate Filters," J. AWWA, 67: 10:535 (October 1975).
Kawamura, S., "Optimization of Basic Water Treatment Processes-Design and Operation:
Sedimentation and Filtration," J. Water SRT.Aqua, 45:3:130-142 (1996).
Kawamura, S., "Two~Stage Filtration," J. AWWA, 77: 12:42 (December 1985).
Kawamura, S., et ai., "Modifying a Backwash Trough to Cut Media Loss During Filter
Washing," J. AWWA, 89:12:60 (December 1997).
Lang, J. S., et aI., "Investigating Filter Performance as a Function of the Ratio Filter Size to
Media Size," J. AWWA, 85:10:123 (October 1993).
Maran, D. c., et al., "Particle Behavior in Deep Bed Fihration," Parts 1 and 2, 1. AWWA.
85: 12:69 and 85: 12:82 (December 1993).
BASIC APPROACH 281
Mayhugh, 1. R., et a1., "Filter Media Rehabilitation at a Lime-Softening Plant," 1. AWWA,
88:8:64 (August 1996).
Monk, R. D. G., "Design Options for Water Filtration," 1. AWWA, 79:9:93 (September 1987).
Monk, R. D. G., "Improved Methods of Designing Filter Boxes," 1. AWWA, 76:8:54 (August
1984).
Nieminski, E. c., and Ongerth, J. E., "Removing Giardia and CI)'I"OSporidiulli by
Conventional Treatment and Direct Filtration," J. AWWA, 87:9:96 (September 1995).
Niquette, P., et aI., "Shutdown of BAC Filters: Effects on Water Quality," J. AWWA, 90: 12:32
(December 1998).
Norman, P. A., "Backwash Instability in Rapid Sand Filters," M.S. thesis. University of New
South Wales, Australia, 1970.
Pontius, F.. "Regulating Filter Backwash Water;' 1. AWWA, 89:8:14 (August 1997).
Suthaker, S .• et aI., "Optimization of Filter Ripening Sequence," 1. Water SRI·Aqua,
47:3: I 07-118 (1998).
Tobiason. 1. E .• et al.. "Pilot Study of the Effect of Ozone and Peroxone on In~Line Direct
Filtration;' J. AWWA, 84:12:72 (December 1992).
Trussell, R. R.. and Chang, M., "Review of Flow through Porous Media as Applied to Head
Loss in Water Filters," J. Ellv. EIlgrg .. ASCE, 125: II :998 (Nov. 1999).
Valade, M. 1., et aI., "Particle Removal by Floatation and Filtration: Pretreatment Effects,"
J. AWWA, 88: 12:36 (December 1996).
Visscher, J. 1.. "Slow Sand Filtration: Design, Operation and Maintenance," 1. AWWA, 82:6:67
(June 1990).
Wang, J. Z., et at., "Biofiltration Performance," Parts I and 2, J. AWWA, 87:12:55 and 64
(December 1995).
Weber, W. J., Physicochemical Pmcesses, Wiley-Interscience, New York, 1972.
3,2.8 Specific Types of Filters
This section discusses the slow sand filters, proprietary filters (in general), the two-
stage filtration process, and a miscellaneous type of proprietary filter-the automatic
backwash filter.
Slow Sand Filters Slow sand filters have been used in water treatment since the
early nineteenth century. They have been proven to be effective under various condi-
tions as long as the filters are properly designed and applied to appropriate situations.
The system is simple, cost-effective, reliable, and easy to operate. Thus, their use should
be considered seriously for remote areas or small to midsized water supply systems in
developing countries. The filtration rate of slow sand filters is SO to 100 times slower
than that of ordinary rapid sand and high-rate filters. Consequently, a considerably larger
area (filter bed) is necessary to produce the same amount of water.
A unique feature of slow sand filters, as described in much of the literature, is
the presence of a thin layer of medium on the surface of the tilter bed, known as
the schmutzdecke, loosely translated as "dirty maL" This special layer contains a
large variety of microorganisms and enables these filters to remove bacteria by a
2
faetor of 10' to 104 and E. coli by a factor of 10 to 10'. The schmutzdecke also
removes organic matter and reduces the turbidity of the raw water.
 282 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 283

Slow sand lilters olTer several advantages: (I) they are simple to build and operate Filter bed depth 3 to 5 ft (I to 1.5 m)
and arc cost-elfective, (2) the Iiltration system is reliable and produces good-quality Support gravel depth I to 1.5 t\ (four layers) 0.3 to 0.45 m
dflnkmg water, (3) the system docs not require highly trained operators and has mini- Water depth above the bed 3.3 to 5 ft (I to 1.5 111)
mal power requirements, and (4) the tilters can tolerate hydraulic and solid shock load- Height of the freeboard 2':8 in. or 0.2 m
ings provided that they are not excessive. Yet the disadvantages tcnd to restrict their Height of the filter box 8.2 to 13 ft, average of 10.5 ft; (2.5 to 4 111,
use: (I) they require a large amount of land, (2) the filters are easily clogged by ex- average of 3.2 111)
cessive amounts of algae, and (3) intermittent operation of the filters may degrade the Filter sand
quality of the tilter eflluent by promoting anaerobic conditions within the Ii Iter bed. Effective size 0.15 to 0.35 mm
The slow sand tilter system is a batch process and the tilters must therefore be Uniformity coefticient Less than 3, preferably near 2
cleaned periodically. Since the majority of the suspended solids (in the raw water) Filter gravel
are removed in the top 0.8 to 1.2 in. (2 to 3 cm) of the bed, the tilter must be cleaned Top layer 0.4 to 0.6 mm with a depth of 10 cm (4 in.)
by scraping a thin layer of sand off its surface. Baekwashing is not employed since Second layer 1.5 to 2.0 mm with a depth of 10 em
this will restratify the bed hydraulically. The lilter bed may require scraping every Third layer 5 to 8 mm with a depth of 10 cm
2 to 6 months, depending on the turbidity and the concentration of algae in the raw Bottom layer 15 to 25 mm with a depth of 10 em
water. The tilter cleaning operation is usually performed on a manual basis and COm- Underdrains The slow sand filtration system generally has a
pleted in I to 2 days. Since manual cleaning of the tilters does not require highly main manifold and many laterals; the later-
tramed personnel, this particular feature is best suited for regions with an abundance als may be porous drain tiles or glazed pipes
of inexpensive labor. Industrialized nations should use mechanized systems, such as laid with open joints, but PVC pipes are
tractors with special skimming buckets or traveling bridges fitted with scraper mech- more often used; 4 m (13 ft) maximum lat-
anisms, to make the cleaning operations more economical. eral spacing
After the bed is cleaned, a period of up to I week is required to ripen the filter Filter eftluent-Ievel control weir The etIluent weir should be at least 0.2 m
bed (build the schmutzdecke) so that an acceptable quality of water is produced. above the top of the initial height of the fd-
Treatment plants that usc slow sand filters and process raw water with seasonally ter bed
high turbidity spikes and algal blooms must use pretreatment with chemical Iloc-
culation and settling, in conjunction with prechlorination. Figure 3.2.8-1 is a longitudinal section of a slow sand lilter with a lilter cflluent level
control weir and major valves. The tilter effluent level control weir may be a down-
Acceptable Raw Water Quality for Slow Sand Filtration The acceptable raw opening slide gate, f~r more flexible operation, or a telescoping tube lype of valve.
water quality for slow sand filters is generally ba<;ed on the same criteria as those
used for the direct filtration process of high-rate filters. The criteria are listed in
Table 2.4.5-1. Slow sand filters can tolerate raw water with turbidities greater than Scum removal and
overflow va Ive
15 ntu as long as the turbidity spikes are less than 50 ntu and last no more than 2 Air vent
to 3 days. However, if the spikes exceed these criteria or if frequent high algal counts
are to be expected, an appropriate pretreatment process train must be implemented. O.2m
0.05 m
General Design Criteria The general design criteria for the slow sand tilter sys- l...
tern arc listed below. R.w rir::::='-i

Required filter bed area

Number of filters
0.5 to 1.6 ft2 per capita per day (0.5 to 0.15
m' per capita per day)
/I = 0.25Qo.s, where Ii is the number of lilters
w.ter Dra;n 1r;.,J:~:-'~ "~ ~> ;i~ :~ i;~'\~ \~·.~..> .~ '~:.............
.......
•• -.
-

nderdraln
'-w" ' .• _':0
§':~ '-~'; ~';·~= ~~~§-?(~~~~~~~~~-jn2....0
, ... ;
-
cle.rwell

and is greater than or equal to 2, and Q is

Area of each Ii Iter bed

the plant Ilow rate (m 3/h) Drain l +Drain
Less than 32,300 ft' with an average of 1075
Flow rate
to 2150 ft' or less than 3000 m' with an control valve Backfilling valve
(also weir bypass)
average of 100 to 200 m'
Filtration rate 0.04 to 0.08 gpm/ft' (0.1 to 0.2 m/h) Figure 3.2.8-1 Section view of slow sand II Iter.

-_",-I
 284 DESIGN OF BASIC TREATMENT PROCESS UNITS
Operation and Maintenance of Slow Sand Filters In this section we discuss
the (I) startup, (2) filtration, (3) filter cleaning, and (4) resanding processes.
STARTUP When construction of the filter is complete, the sand in the bed is both
clean and dry. In this condition, the filter is incapable of water purification because
it must first undergo a maturation process to build up the schmutzdecke. The mat-
uration process has two distinct phases: backfilling and gradual initiation of filtra-
tion.
The first step, slow backfilling, drives the air out of the sand bed. When the raw
water reaches a level of approximately 3 ft or (I m) above the surface of the bed
backfilling is terminated and the filter inlet valve is opened. introducing raw wate;
from the inlet instead of the outlet. The backfilling process is extremely important
because it prevents the filter bed from being scoured or disturbed by the cascading
effect of the raw water onto the filter bed. When the water reaches normal opera-
tIOnal levels, the second step is initiated.
The second step in the maturation process is gradual initiation of filtration. This
procedure is achieved by partially opening the filter effluent valve; the suggested
filtration rate is IOta 30% of the designed filtration rate. During this step the fil-
tered water must be continuously removed via a drain valve until the quality of the
water meets the drinking water quality standards. This step usually takes several
weeks in regions of hot weather and considerably longer in colder regions, due to
the suboptimal levels of bacterial nutrients and low temperature.
The maturation or ripening period may be shortened by seeding the new filter
sand with schmutzdecke from another filter. Filter bed maturation may be detected
as a slight, yet distinct, increase in the headloss across the filter bed. Once the bed
becomes ripe, the effluent valve should be opened to produce the designed flow rate,
and the quality of the eflluent water should be rechecked. When the eflluent water
meets the drinking water quality standards, the drain valve is closed and all the fil-
tered water is directed to the water supply system. The quality of the eflluent wa-
ter must always be checked whenever the filtration process is interrupted for more
than I or 2 weeks.
FILTRATION PERIOD The main task of the operator is to control the flow rate through
each filter by first adjusting the opening of the filter effluent valve and then by mak-
Ing the necessary adjustment at the filter influent valve. At the beginning of each
filtration cycle, the filter effluent valve should be at a partially closed position since
there is very little head loss across the filter bed. The headloss increases with the ac-
cumulation of suspended matter on the surface of the bed-with the creation of the
schmutzdecke. It is therefore necessary to provide some type of flow indicator near
the filter eflluent valve, which allows the operator to gauge the Ilow rate and ad-
just the valve opening so that the proper Ilow rate can be met. The simplest method
is to install a gauge stall' upstream of the filter eflluent-Ievel control weir. It may
be a gauge glass or a Iloat in a f1oatwell, but it must be located outside the box
housing the control weir since the operator must be able to check the reading eas-
ily while making the flow adjustments. Alternatively, a flow meter may be located
BASIC APPROACH 285
immediately upstream of the filter eflluent valve. This is an ideal configuration, and
all filters should implement this scheme to provide more accurate Ilow control, with
a few exceptions. The exceptions, due to cost considerations, are small plants and
developing countries. The filter inlluent valve should also be adjusted periodically
to keep the influent and eflluent flow rates in balance. If the operator fails to do so,
an exorbitant amount of water will Ilow through the overflow pipe or the water
above the filter bed will be very shallow.
Plant operators must maintain the filters in good working order by routinely mon-
itoring the filter system. For instance, algal blooms in the raw water should be con-
trolled at the source by adding 0.15 mg/L of copper sulfate, or through other means,
before the filters become clogged. Operators should routinely clean the filters and
their surroundings, in addition to removing any floating scum, leaves, and debris.
The operator should also routinely analyze samples of both the raw water and the
filtered water at prescheduled intervals.
FILTER CLEANING As mentioned previously, slow sand filters may require cleaning
every 2 to 6 months, depending on the algal concentration and turbidity of the raw
water. The cleaning process is initiated at the end of the last filter run and IS com-
posed of three phases: draining the filter bed, removing the sand, and reripening the
bed. The end of a filter run is characterized by a rapid increase in headloss across
the bed that is simultaneous with a reduction in the filter eflluent flow rate, despite
a fully opened valve. There will also be a slight degradation in the quality of the
filter eflluent water. The filter inlluent valve is closed at the end of a filter run and
the effluent valves are opened to drain the water level; the water level above the bed
will drop to the height of the outlet weir overnight. The filter effluent valve should
then be closed and both drain valves, just downstream of the filter inlluent valve
and upstream of the filter effluent valve, should be opened to speed the drainage
process. The water level must be 6 to 8 in. (15 to 20 cm) below the surface of the
bed before the cleaning operation can be initiated. As soon as the top of the bed IS
sufficiently dry, the second step in the cleaning operation may be started. The scrap-
ing process should not be delayed because the guano of scavenging birds contami-
nates the filter bed to a greater depth.
The scraping operation removes the top 0.8 to 1.2 in. (2 to 3 cm) of the bed by
means of manual labor or mechanical equipment. In either case, the schmutzdecke
and the surface sand adhering to it should be quickly stripped and carefully stacked
into ridges. If the bed is cleaned manually, the waste material should be removed
with square blade shovels and the debris transported by wheelbarrow or conveyor
belt. If the filter beds are large and the cost of personnel is high, the scrapIng may
be performed by a tractor fitted with a specifically designed scraping blade. Both
mechanical equipment and wheelbarrows must be run on protective planks so that
the top portion of the filter bed is not damaged and the microorganisms (in the bed)
are not killed. Upon completion of the scraping operation, the surface of the filter
bed should be leveled.
If the schmutzdecke is composed largely of filamentous algae, the cleaning op-
eration will be easy. However, if the predominant species are nonfilamentous or-
 286 DESIGN OF BASIC TREATMENT PROCESS UNITS
ganisms, such as diatoms, the cleaning process will be more difficult. Workers must
take precautions to avoid being exposed to harmful microorganisms contained in the
schmutzdecke. Similarly, the workers must be careful not to contaminate the tilter
bed with bodily excretions or spit.
The third step in the cleaning process is reripening of the lilter bed. This process
is the same as described in the startup process, although the maturation period is
greatly accelerated.
RESANDING After approximately 20 to 30 scrapings, the depth of the tilter bed is
reduced to 20 to 28 in. (0.5 to 0.7 m). This is the minimum dcpth that should re-
main above the supporting gravel bed. When the bed is reduced to this level, the fil-
ter should be drained and new sand must be added to restore the original medium
depth. The resanding operation does not simply add new sand on top of the origi-
nal bed. First, 12 to 20 in. (0.3 to 0.5 m) of the old bed is moved to one side of the
filter. Next, new sand is added on top of the old sand; the depth of this old sand
layer should be approximately 8 in. (0.2 m). Finally, the new sand layer is covered
by the old sand that was initially set aside. This practice retains most of the valu-
able microorganisms, thus shortening the reripening period.
In regions where sand is difficult to obtain or is expensive, it is common to re-
cycle the sand that was removed during the cleaning process. This contaminated
layer is washed by water jets or by a hydraulic educator system and sorted for fu-
ture resanding operations. The scrapings must be washed immediately after removal
or it will become putrid and produce an objectionable odor; it is quite difficult to
remove the coating of debris if the scrapings are allowed to dry. Sludge from the
scrapings contains numerous bacteria and organic substances, thus is very odorif-
erous and attracts many flies. It is therefore prudent to schedule the recycling and
resanding operations during the winter season.
Proprietary Filters Many water treatment equipment manufacturers supply var-
ious types of filter, mainly for industrial purposes, but a few market filters have been
especially designed for water treatment plants. In the United States, the proprietary
lilters are used primarily in small plants (those processing less than 15 mgd).
However, this is not necessarily true for some overseas countries.
The two basic types of proprietary unit used in water treatment are the gravity
filter and pressure filter. Gravity filters may be composed of a variety of materials,
but pressure filters are cylindrical pressure vessels that are fabricated exclusively
from steel plates. The use of pressure filters may be considered if the high water
pressure of the plant inlet is to be preserved, this avoids having to repump to ob-
tain the water pressure necessary to distribute the filtered water. Since pressure fil-
ters use an equivalent of approximately IS psi (1 kg/cm') of line pressure, provided
that the inlet pressure is over 35 psi (2.5 kg/cm') of water head, the Ii Iter effluent
line still has enough high pressure to backwash the filters without the use of wash
pumps. Additionally, there is enough line pressure remaining in the system to trans-
mit filtered water to a remote distribution reservoir. If the gravity tilters are used
under these same conditions, the filter effluent line virtually has no line pressure
BASIC APPROACH 287
left. Engineers should therefore evaluate the possible use of pressure filters when
designing small to midsized plants under the aforementioned conditions, despite the
fact that state regulatory agencies generally discourage the use of pressure filters
because of the difficulty in inspecting filter bed conditions.
In the water supply business, one of the best applications for pressure filters is
the filtration of pumped, deep-well water containing iron and manganese. This type
of filter will remove these compounds through oxidation by chlorine or potassium
permanganate, and the well pumps usually supply a high level of extra pressure to
the pump discharge line. The proprietary filters that are generally used in public wa-
ter treatment plants are open-top gravity filters, since these allow easy observation
and maintenance of the filter bed. Yet each manufacturer markets a somewhat unique
filter: self-backwash, continuous automatic washing, air-scouring wash, bumping
bed, coarse media bed for oil removal, ultra-high-rate, and so on. The size and de-
tails of the proprietary filter units are generally preengineered by the manufacturer
based on a wide range of flow rates; these units are not cataloged based on filter
size and appurtenances, but on flow rates. Since the manufacturer has predesigned
the blueprints, the completion time of the project is much shorter compared to us-
ing custom-designed filters. Additionally, many of these filters are self-contained
units that require rather simple power and pipe connections to make them functional.
Another advantage of these units is that the manufacturers always give the customers
a limited guarantee on both the performance of the units and the equipment.
However, the customer usually cannot request significant design modifications with-
out forfeiting the guarantee. This can be a potential problem when the job site is in
a remote area; repair and replacement of certain parts may be difficult if they are
patented items. Moreover, many proprietary filters are fabricated with metal plates,
and corrosion protection is therefore a serious problem. Although proprietary filters
may be custom designed, this negates the advantage of using these units: the sav-
ings in time and money.
The decision to use proprietary filters versus custom-designed units is the re-
sponsibility of either the engineer or the owner. If the engineer is authorized to make
the decision, he or she must conduct a thorough evaluation of the units that are on
the market. Special attention should be given to the number of years the unit has
been in operation, feedback from plant operators using these types of unit, quality
control, follow-up servicing, cost of each type of unit, and the financial stability of
the manufacturer. The engineer may choose to conduct prequalification of certain
manufacturers prior to bidding in order to avoid purchasing an inferior unit, although
the unit may meet the specifications. Figure 3.2.8-2 depicts an example of a pro-
prietary pressure filter.
TwO-Stage Filtration The two-stage filtration process has historically been ap-
plied to the slow sand filtration process to improve the overall efficiency of filtra-
tion. Not until the late 1970s did this process gain attention as a substitute for the
high-rate filters used in both direct filtration and in-line filtration: these two types
of filter are susceptible to turbidity spikes and algal blooms in the raw water. The
two-stage slow filtration process has also been used successfully in Europe in the
 288 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 289

OPERATING
WATER lEVEL

I. Ra", "'QI~rl~111 ' 111>...

2. R"w-~'''I'' /~/el Qnd
"'''Sh'M'''I'' iJIIlkl
.htifl.
3. N""I.z.
4. Air/nl" Vb',,,,
.5. Sludgt IX""<I/on
"111'1.
6. AIM""'" / ...1pipe.
7. Ai, rNni.
8. Rfnu "11/"1 /0' fil. Fll TER
Ined M'ol., b'bdt' efflUENT
p/~J)·."m.
9. WOlh'k'l1ler inl~1

FILTRATION MODE
10. Pu·fillt,rd ",,,It, /0
WI/'I. jilltt.
/bkt,
II. FlI/tled ",,,/e, brad"
~~~C:;=:a~""',
pl~JYJ"m.
n. Flluud k'ble, 01111"
I'a/'I. WASH CYCLE
WATER lEVEL

Permutit Type E Vertical Filter

(Permutit CO)
- OMalb,d"rI/'
~'1,t;~;~rp;~~:~~_ '/1'1'''1 ,h,
Ir"',lrnfD.g,emQnl/

Figure 3.2.8-2 Proprietary pressure filters. (From bulletins of Permutit Company and
Innlco-Degremont International.)

treatment of poor-quality raw water, as well as raw water with seasonal degrada- BACKWASH CYCLE
tions in quality. Figure 3.2.8-3 Commercialized two-stage filtration system. (Courtesy of Neptune
The two-stage filtration process for high-rate filtration is composed of two dis- Microtloc. Inc.)
tinct filtration steps: first-stage filters and final-stage filters. These stages consist of
a bank of coarse bed roughing filters and a bank of ordinary high-rate filters, re-
spectively. The ability of the coarse media bed to act as a flocculation system has tration process in relation to direct tiltration and a conventional complete treatment
been recognized and demonstrated in a number of studies since the early 1970s. The process. Figure 3.2.8-3 is an example of the two-stage liltration system.
function of the roughing tilters is flocculation: to produce floc with the proper char-
acteristics for the subsequent high-rate filters, in addition to removing approximately General DeSign Criteria
50 to 80% of suspended matter. The tilters therefore depend on coagulant dosage
and the nature of the raw water. ROUGHING FILTER (FIRST-STAGE FILTERS)
The major advantage of the two-stage process are the substantial savings in co-
agulant-only 4 to 6 mg/L of alum is required-and the small amount of sludge pro- Filtration rate 12 to 15 gpm/ft2 (30 to 37.5 m/h)
duction. In contrast to the direct and in-line filtration processes, the two-stage process Filter medium Pea gravel, crushed anthracite, or plastic chips
is also capable of withstanding turbidity spikes of short duration and algal blooms. Size of medium, d 3 to 6 mm, effective size
A few equipment manufacturers have recognized these advantages and have com- Uniformity coetlicient $1.5
mercialized this system with proprietary designs and patents. The one disadvantage Bcd depth, I 500 < lid > 250
of these tilters is that wash waste is produced by two sets of filters. Refer to Table Filter washing Air-scouring backwash
2.4.5-1 for a general idea of the applicable raw water quality for the two-stage process. Backwash rate 15 to 20 gpm/ft2 (37.5 to 50 m/h)
Figure 3.2.7-20 contains the proper range of coagulant dosage for the two-stage til- Air-scouring rate 3 to 5 ft'lmin per square foot (55 to 90 m/h)
 BASIC APPROACH 291
290 DESIGN OF BASIC TREATMENT PROCESS UNITS

Maximum available headloss 3.3 ft (I m) The advantages of this filter are that it does not need a backwashing facility, in
for filtration the regular sense, and the wash waste is very easy to handle due to the low tlow
Underdrains Air-scouring type rate: only 200 gpm (0.23 m 3/h) for a period of 30 to 90 min. One drawback with
Wash-water troughs Use of the troughs is preferred respect to potable water production is the need for a housing over the filter bank.
Direction of flow Downflow is normal; however, upflow and The housing is necessary because of the potential for contamination, since the fil-
horizontal tlow have been applied tered water flows in an open-top channel.
The maintenance requirements of the ABF are very simple: periodic inspection
and lubrication of the drive mechanism and pumps, and occasional touch-up paint-
FINAL FILTER (SECOND-STAGE FILTERS)
ing. Yet if the process water has a medium to high hardness, the porous plate un-
The design criteria for the final filters are exactly the same as those of the high-rate derdrains will become clogged by calcium carbonate depOSits over a perrod of 15
filters (refer to Section 3.2.7). to 20 years, depending on the conditions. Experience has shown that an acidifica-
tion treatment-application of 5 to 10% hydrochloric acid-restores the capacity of
Miscellaneous Filters Diatomite, bitlow, and automatic backwash filters such the filters. Figure 3.2.8-4 presents a schematic view of an ABF. Discussion of the
as the Hardinge filter, ABF (manufactured by Infilco-Degremont and other compa- other types of miscellaneous filters will not be attempted since the application of
nies), Dyna Sand (by Parkson Corp. and other firms), and diatomaceous earth pre- these filters to the public water supply is very limited due to their serious disad-
coat filters are the other types of filter that are used with limited applications in the vantages.
processing of the public drinking water supply. Of these units, the ABF and di- One of the newly emerging filtration techniques for potable water treatment is
atomaceous earth precoat filter are the only types that have considerable acceptance the membrane technique: microfiltration, ultrafiltration and nanofiltration. The pore
for industrial applications, as well as potable water plants for small public water size of these membranes is small enough to separate solids from the raw water and
supplies, when the use of package plants is appropriate. capable of filtering out bacteria, viruses, Giardia, and CYlptosporidum. It is impor-
When the patent on the ABF design expired in the 1960s, a total of four man- tant to note that humic substances and THMs can be controlled by nanofiltratlon.
ufacturers began marketing very similar units. This competition helped to gener- These new water treatment systems have been evaluated by pilot-scale plants dur-
ate a much wider acceptance of this type of filter, especially for use in tertiary ing the late 1980s, and the results are promising. Membrane filtration is discussed
wash-water filters. Yet certain states discourage the use of ABF for potable water further in Section 7.8.2
because of its declining-rate mode of filtration and the lack of a provision for fil-
ter-ta-waste mechanism. However, a recent unit is equipped with filter-ta-waste
[I. !:(fluenl d",'hatH" hnc \1 Wa'h"'~lcrdl~(htrK. PiJ~
A 1"lluem tine
coupled with a turbidimeter and has the option of adding polymer to each cell af- [J. InOuen! pom 1. U8<1......·a'h'·alv. :-:. WMhwalcrrrou,h
J. l"l.co:v.·uhl'umJ'~.,.mbl) O. Wa.h ... \cnli.,."harge
C. Innuent ,hanne!
ter backwashing. tl Comp4rcmcnle.Jfl!l~rl)e.J K. W.,h .... IH hoW P M.chanism dm·, mOlor
U n~c\wa~h suppon ret.in,n; sprlng\
E. S'~lIonalil.du~dtN",n I .. w••II'I.. altrp"mp~"~01!11y
The original ABF design has a depth of only II in. (0.28 m) of 0.6 mm (effec- 1'. fJflucnlandblrk .. uhrorl> R. rtd'u~("OfIlrol.p.rinp

o. [,!fluenl eh.nnel s. C"Qnlrolinn,umenlG1l0n

T. Travc!inlb.,·kwnh mechMnism
tive size) sand. However, a depth of 16 to 24 in. (0.4 to 0.6 m) of dual media is now
available. The ABF filter bed is supported by a porous plate underdrain and is di-
vided into numerous filter cells that are approximately 9 in. wide. Each cell is washed
individually by a traveling bridge equipped with a backwash pump and a wash-waste
pump, thus allowing the rest of the filter cells to remain in service. If the filter bed
depth exceeds 16 in., the traveling bridge stops at each filter cell (indexing) to com-
plete filter backwash. The standard width of one automatic backwash filter is 16 ft
(5 m), and the recommended maximum length of each filter is 150 ft (45 m).
The basic principle on which the automatic backwash filters operate is a surface
filtration that facilitates shallow penetration of tloc into the filter bed. The filter
therefore operates at a low headloss of 6 to 12 in. (0.15 to 0.3 m) with a filter rate
2
of 2 to 3 gpm/ft (5 to 7.5 m/h). Filter backwashing is initiated by either a headloss
or a timer. A typical wash cycle is every 2 to 6 h and the surface of the bed never
becomes compacted because of this frequent wash cycle. Consequently, surface wash
is not required and air binding does not occur. Figure 3.2.8-4 Automatic backwash filter. (Courtesy of Infilco-Degremont.)
 292 DESIGN OF BASIC TREATMENT PROCESS UNITS
BIBLIOGRAPHY
Bersillan,1. L., et al.. "Ultrafiltration Applied to Drinking Water Treatment: Case of Small
System," AWWA Annual Conference, Los Angeles, June 1989.
Burns, D. E. et aI., "Particulate Removal on Coated Filter Media," 1. AWWA, 63:2:121
(February 1970).
Cleasby, 1. L., "Slow Sand and Direct In-Line Filtration of a Surface Water," 1. AWWA,
76: 12:44 (December 1986).
Degremont, Water Treatment Handbook, 5th ed., Halstead Press, New York, 1979.
Dillingham, 1. L.. et aI., "Optimum Design and Operation of Diatomite Filtration Plants"
1. AWWA, 58:6:657 (June 1966). ' viruses and protozoan cysts. Consequently, the EPA has proposed that the standard
Graham, N. 1. D., ed .• Slow Sand Filtration, Ellis Horwood, New York, 1988.
Huisman, L., et aI., Slow Sand Filtration, World Health Organization, Geneva, 1974.
Jacangelo. 1. G., et aI., "Assessing Hollow-Fiber Ultrafiltration For P3Iticulate Removal"
1. AWWA, 81:11:68 (November 1989). ' heterotrophic plate count (HPC) for water in distribution systems should be less than
Kawamura, S., "1\vo-Slage Filtralion," 1. AWWA, 77: 12:42 (December 1985).
Sanks, L. R., ed., Water Treatment Plant Design for the Practicing Engineel; Ann Arbor
Science, Ann Arbor, MI, 1978.
Schulz, C. R., and Okun, D. A., Sw/ace Water Treatmem for Communities ill Developing
Countries, Wiley, New York, 1984.
Vaillant, C. J., "Simple Water Treatment Processes," lWSA SOUli1ern Asia Regional
Conference, Bombay, October t 981.
3_2,9 Disinfection Process
Purpose The purpose of the disinfection process is to kill the growing form of
pathogenic microorganisms, not necessarily the resistant spore forms, through the
use of chemicals, ozone, or ultraviolet (UV) irradiation. Palhogens are defined as
any type of microorganism capable of producing disease. The process of disinfec-
tion should not be confused with sterilization, which refers to the total destruction
or total removal of alI microorganisms from the treated medium, including the spore
forms. Strictly speaking, heat treatment, and gamma or x-ray irradiation, should be
included in the disinfection process. However, these processes are generally prac-
ticed by various industries and are not used in treating Ihe public water supply, pri-
manly because they do not have the residual disinfecting power required to combat
postcontamination within the distribution systems. Thus, these types of disinfecting
processes, except UV irradiation, are not discussed in this book.
ConSiderations Major considerations in selecting a disinfection process are (I)
the presence of surrogate organisms in the drinking water supply, (2) the feasibility
of using alternative disinfectanls, (3) the disinfeclant residual""ontact time rela-
tionship (C X f), (4) the formation of disinfectant by-products and their magnitude,
(5) the quality of the process water, (6) safety problems associated with the disin-
fectants, and (7) the cost of each disinfection alternative. Each of these considera-
tions is discussed in turn.
BASIC APPROACH 293
surrogate Organisms Coliform organisms, particularly Escherichia coli
(E. coli), are continuously present in the human intestine in large numbers. Billions
of these organisms are excreted by an average human being or animal each day.
Thus their presence in the process water is indicative of fecal pollution. The col-
iform group of bacteria is defined as all aerobic and facultative anaerobic, gram-
negative, non-spore-forming bacilli that produce acid and gas through the fermen-
tation of lactose.
The most common method of assessing the safety of the drinking water supply
is the coliform test. However, microbiologists generally agree that coliforms may
be more rapidly inactivated at much lower disinfectant dosages than some enteric
plate count be instituted as a routine test to screen for bacteria regrowth. The pres-
ence of a large population of organisms, as indicated by the plate count (heterotrophic
bacteria), would certainly be indicative of an ineffective disinfection process. The
500 colonies/mL. The EPA is also emphasizing the inactivation of protozoan cysts
such as Elltamoeba his/oly/ica and Giardia lamblia, as well as Cryptosporidium,
Legionella, and enteric viruses.
Alternative Disinfectants Chlorine, chloramines, chlorine dioxide, and ozone are
practical alternative disinfectants. The selection process is based on the effective-
ness of the alternatives as a disinfectant, formation of harmful by-products, resid-
ual disinfecting capabilities, and cost-effectiveness.
CHLORINE Chlorination of potable water has been practiced in the United States
since 1903; therefore, much knowledge is available concerning the application tech-
nology and its effectiveness. Despite growing concerns regarding THM, a suspected
carcinogen that forms during the chlorination process, chlorination will undoubt-
edly continue to be the primary water disinfectant for years to come, especially in
developing countries. The basis for this reasoning is that all potable water should
provide disinfectant residual in the distribution system, and chlorine can easily sat-
isfy this requirement. Additionally, chlorine has been proven to be effective as a dis-
infectant at reasonable cost for almost a century.
Chlorine is generally applied to the process water as an aqueous chlorine solu-
tion: Chlorine gas is hydrolyzed to form hypochlorous acid (HOCl) and hypochlo-
rite ion (OCI-) in ammonia-free water. Hypochlorous acid in water is commonly
known as free residual chlorine. The relative concentration of these two species
is a function of the pH of the water and, to some degree, the temperature (Figure
3.2.9-1). The source of the chlorine is usually liquefied chlorine gas. However, re-
gions that prohibit the use of chorine (for safely reasons) and small plants com-
monly use calcium hypochlorite, sodium hypochlorite, or chlorinated lime as sources
of chlorine. Free residual chlorine reacts rapidly and indiscriminately with many
substances in the water, including microorganisms. Thus, effective disinfection and
minimum THM and haloacetic acid (HAA) formation may be achieved if chlorine
is added to clarified water with low Toe such as filtered water.
 BASIC APPROACH 295
294 DESIGN OF BASIC TREATMENT PROCESS UNITS

o CHLORAMINES Chloramination of drinking water has been used successfully in

tOO
many water utilities and has proven to be an etfective disinfecting process. However,
90 ~ 10 chloramines are less effective than chlorination as bactericides and viricides.

~
Moreover, neither the enteric viruses nor protozoan cysts are effectively inactivated
in short periods of time (less than 30 min) at low dosages (I to 2 mglL). Yet chlo-
eo 20
rami nation does have the advantage of producing very small amounts of THM-
levels that arc insignificant to human health. During the late 1970s. the EPA pro-
70
posed that chloramination not be used as the primary disinfectant, due to its
40 suboptimal performance as a bactericidal and viricidal agent, but this proposal waS
60
.... later modified, leaving implementation to the discretion of each state. Thus, the chlo-
u

.
~50
rami nation process may be used as a prcdisinfectant and/or a postdisinfectant prior
to distribution.
One of the most sensible uses of chloramination is in combination with other dis-
0 6o
infectants, particularly in the case of postdisinfection, since the chloramine residu-
2O'C als last much longer than free chlorine and do not form signit1cant amounts ofTHM.
~ o Chloramination is normally practiced at a ratio of approximately I part of ammo-
O"C nia to 4 parts of chlorine (using a mglL base) to ensure monochloramine formation.
20 8o
Thus. if a postchlorinated water has 1.0 mglL of free residual chlorine. 0.25 mglL

0
\\ 90
of ammonia is to be added. However, if 0.3 mg/L of ammonia is added prior to

~ I::..
postchlorination, a minimum of 1.2 mg/L of chlorine must be added.

o 1
• 6 7
pH
8 9 10 CHLORINE DIOXIDE Chlorine dioxide (Cl02) has been used as a primary disinfec-
tant by several large systems in Europe. However, it has had limited application as
Figure 3.2.9-1 Effect of pH on the distribution of hypochlorous acid and hypochlorite ion a water disinfectant in the United States; it has been used primarily as a means of
in water. taste and odor control, oxidation of iron and manganese, and control ofTTHMs and
HAAs. The basic reaction for the production of chlorine dioxide is

2NaCIO, + CI, --> 2CIO, + 2NaCI

Many factors affect the disinfection efficiency: oxidizable substances in the ~hl<ll'il1c dioxi<.k
sodium ~hlmilc ~hlorioc
process water, particulate concentration, pH, temperature, contact time, and the level
of the residual chlorine. The ideal disinfecting condition has a low pH (6 to 7). a Therefore. 1.34 Ib of sodium chlorite reacts with 0.5 Ib of chlorine to produce
relallvely hIgh water temperature (20 to 25"C). a longer contact time (over 30 min). 1.0 lb of chlorine dioxide. However, excess chlorine oxidizes chlorine dioxide to
and higher levels of residual chlorine (over 0.5 mglL). According to Morris (1975).
hypochlorous acid is 100 times more effective than hypochlorite ion as a germicide chlorate ion:
tor enteric bactena and amoebic cysts. Yet hypochlorous acid is less effective in
HOCI + CIO, --> CIO, + HCI
combating viruses, as well as amoebic cysts and bacterial spores, compared with
other types of disinfectant. such as ozone or chlorine dioxide. Although hypochlo- Chlorine dioxide in water produces chloratc ion and chlorite ion with a slow reac-
rite ion is much less effective than hypochlorous acid compared with monochlo~
ramine, its germicidal power is greater by a factor of 2. tion:
Chlorination is a proven. reliable. and mature technology. However. THM for- 2CIO, + H 2 0 --> CIO, + CIO, + 2H+
mati?n and ~ub1ic safety are major concerns for both the engineer and the general chlorale ~hlorilc
pubhc. StudIes have shown that THM formation is reduced substantially when the
application point of chlorine is changed Irom the head of the plant to postfiltration. Chlorine dioxide has a high oxidation power. It is an effective bactericidal agent and
The potential leakage of chlorine gas has been minimized with improvements in the its killing power of microorganisms is equal to or better than hypochlorous acid in wa-
chlorine feeders and chlorine storage. Improvements in both technology and oper- ter. For example, compared to free chlorine, chlorine dioxide is more effective as a bac-
atmg practices will provide a more reliable and safe disinfecting process.
296 DESIGN OF BASIC TREATMENT PROCESS UNITS
tericide and viricide at a higher pH range (8.5 to 9) than at a pH of 7. One definite ad-
vantage of chlorine dioxide is that it requires the pH of the filtered water to be above
8, thereby minimizing lead corrosion and the corrosion of other metals, in accordance
with the recent EPA guideline. Also in this category is the disinfection of softened Wa-
ter by the lime softening process. Another advantage of chlorine dioxide is its ability
to maintain a residual in the distribution system for an extended period of time.
The disadvantages of chlorine dioxide are its higher operational costs ($2.2 per
pound for sodium chlorite) relative to chlorine ($0.13 per pound for chlorine)-late
1990s cost. Additionally, excess chlorine added to the water will form THMs and
undesirable by-products whenever chlorine is used in excessive amounts to ensure
complete reaction with sodium chlorite. This practice was commonly employed in
the United States to produce chlorine dioxide.
Since the EPA has set the criteria for chlorine dioxide residual to be less than 0.8
mg/L and a MCL for chlorite at 1.0 mgIL, whenever chlorine dioxide is used as a dis-
infectant or in taste and odor control, the dosage of chlorine dioxide and optimization
of the generator performance are extremely important. However, CI02 and CIa, can
be controlled by ferrous iron additon. Chlorine dioxide has potentially explosive char-
acteristics with a slight change in environment. Therefore, it should not be stored in
pressurized containers but should be produced on site. Preengineered and packaged
chlorine dioxide-generating equipment is offered by a few equipment manufacturers.
OZONE Among the commonly available oxidants used in water treatment, ozone is
the most powerful oxidant and disinfectant. Ozone has been used as a disinfectant
worldwide in over 1100 plants, including very small installations, primarily in
Europe. In practice. ozone is often used in conjunction with a secondary disinfec-
tant such as chlorine or chloramine, since it is not capable of providing a residual
disinfectant to combat the proliferation of microorganisms within the distribution
system. Although ozone is used as the sole disinfectant in a few European countries
and other parts of the world, the use of ozone alone is not allowed as the final dis-
infection before distribution in the United States, precisely because of the lack of
protection against postcontamination.
The effectiveness of ozone disinfection is a function of the pH and temperature
of the process water and the method of ozone application. The required residual-
contact time relation (C X t) of ozone disinfection is approximately 1.5% of free
chlorine residual. Yet the type of mixing, the degree of mixing, and the rate of mass
transfer greatly influence the effectiveness of ozonation because ozone has a low
solubility in water.
The advantages of using ozone as a disinfectant are its strong disinfecting power,
potentially better flocculation and filter performance, a signiiicant reduction in the for-
mation of THM precursors prior to final chlorination, a limited number of by-prod-
ucts (aldehydes are a by-product that may have adverse effects on human health), and
the low dosages (I to 2 mgIL) required to complete disinfection. However, ozone also
has the following disadvantages: the process does not provide residual protection af-
ter ozonation due to the rapid decomposition of ozone; the ozone must be produced
on-site because it cannot be stored; and the ozonation system has a high capital cost
($3000 per pound of ozone production per day for most medium-sized plants) and
BASIC APPROACH 297
frequent high levels of bacteria regrowth in the distribution line if adequate residual
chlorine is not used. If cost is not a factor, the effectiveness of ozone disinfection may
be enhanced by the use of a catalyst such as UV light or hydrogen peroxide.
ULTRAVIOLET LIGHT UV light is in the region of the electromagnetic spectrum be-
tween visible light, which has wavelengths greater than 400 nm, and x-rays (less
than 100 nm). UV has three spectral bands: (I) 100 to 280 nm, which is short-wave
UV; (2) 280 to 315 nm, middle-wave UV; and (3) 315 to 400 nm, long-wave Uv.
Disinfection by UV light is performed by short-wave UV, which is capable of rup-
turing the deoxyribonucleic acid (DNA) of cells. This disruption of DNA prevents
cell reproduction, thereby inactivating pathogenic microorganisms.
Regular low-pressure UV lamp systems should use 40- to 60-UV lamps per mil-
lion gallons per day (3790 m 3/d). The power requirement for each lamp ranges from
65 to 80 W. The intensity of the UV light emitted by a lamp is expressed as milli-
watts per square centimeter (mW/cm 2 ). UV dose is calculated by intensity times the
exposure time in seconds and is expressed as mW . stcm 2 . The dose required to dis-
infect water varies with the quality of the water, but generally ranges from 24 to 45
mW . s/cm 2 The lamps operate at temperatures ranging from 40 to 50°C and pro-
duce nearly monochromatic radiation at 254 nm.
Medium-pressure UV lamp systems have been employed in Europe since 1980
and is gaining popularity in the United States because this system requires fewer
lamps than the low-pressure UV lamp system. These medium-pressure lamps oper-
ate in a temperature range of 600 to 800°C and produce a high-intensity broad-
spectrum light. Suppliers offer a choice of three transformer power level settings in
the range 60 to 100%, thereby allowing plant operators to operate the lamps at a
lower setting (save energy) during periods of low plant flow.
The pulsed UV system is an emerging UV disinfection system that delivers high-
intensity microsecond bursts of broadband emission with a power flux of 10 mil-
lion mW/cm 2 • Pulsed UV flash lamps contain xenon, which is intermittently ener-
gized to form plasma. The temperature in the lamp rises to approximately 10,000
K, which can be compared to sunlight. At this temperature the plasma emits an in-
tense, broadband spectrum (UV, visible, and infrared wavelengths). However, the
outer surface of the lamp is reduced to near-ambient temperature by an internal cool-
ing system, to minimize problems associated with scaling. A recent study conducted
by the University of South Florida and the Electric Power Research Institute demon-
strated the effectiveness of the pulsed UV disinfection system:
Log Inactivation
Total Dose (m W . s/cm 2 ) Bacteria Virus Cryptosporidium
30 6 2 4.3
50 >9 3 4.8
100 >9 5.6 5.6
Some manufacturers of low-pressure UV systems are Atlantic UV Corp. and
Calgon Carbon Corp. Low-pressure, high-intensity lamps are manufactured by
 298 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 299

UltraGuard in Canada and WEDECO in England. Trojan Technologies of Canada TABLE 3.2.9-1 Recommended Minimum Log RemovaUInactivation by Disinrection
and Aquionics of Erlanger, Kentucky produce medium-pressure UV systems. Pulsed and Filtration Schemes
UV systems are supplied by Pure-Pulse of San Diego, California and Inovatech of Assumed Recommended
EI Cajon, California. Log Removal Minimum Log Inactivation
by Filtration by Disinfection
Disinfectant Residual-Contact Time Relationship Two factors are of primary Giardia Viruses
Filtration Type Giardia Viruses
importance in disinfection: the concentration of the disinfectant residual and the
contact time. The relationship may be expressed as conventional 2.5 2.0 0.5 2.0
Direct 2.0 1.0 1.0 3.0
Slow sand 2.0 2.0 1.0 2.0
eXt = constant
Diatomaceous earth 2.0 1.0 1.0 3.0
where C is the disinfectant residual (mg/L) and I is the contact time (min). The im- Source: Adapted from 54 FR 27511 (June 29. 1989).
portant issue is that a low concentration of disinfectant, with a long contact time,
accomplishes the same goal as using a high residual concentration with a short con- Quality of the Water Being Processed If the water being process~d contains
tact time. However, other factors, such as the pH and temperature of the process turbidity, color, ferrous iron, manganese, and organic substances, all dl~tnfectants
water, also affect the efficiency of the disinfection process. described in this section will react with them. Consequently: the ~ffectlveness of
According to the Surface Water Treatment Rule (final guidance manual issued disinfection becomes very low and a significant amount of disInfection by-produc.ts
by the EPA in 1989) all water treatment systems must comply with the C X I re- will probably be formed. Turbidity and the presence of other particulate matter In
quirements. Yet compliance is a function of the type of treatment used. The EPA
recommends selecting the disinfection by inactivation process based on the type of
filtration system that is used by the filtration plant. Table 3.2.9-1 presents re- TABLE 3.2.9-2 ext Values to Achieve Giardia lamblia Inactivation
movallinactivation as a function of the type of filtration. Table 3.2.9-2 illustrates I Temperature (C)
log (90%) and 0.5 log (68%) inactivation of GiGldia at various temperatures and 0.5 5 10 15 20 25
pH
pH levels. Table 3.2.9-3 presents 2 to 4 log virus inactivation.
Let us assume that a treatment plant has a conventional process, the winter water (a) J log (90%) Inactivation
temperature is 0.5°C (33°F), and the pH of the process water is 8.0. Based on Table
Free chlorinea 6 49 35 26 18 13 9
3.2.9-1, 2.5 log for Giardia and 2.0 log for viruses are credited for the treatment 19 12
7 70 50 37 35
processes; therefore, inactivation (disinfection) will require 0.5 log for Giardia and 54 36 27 18
8 101 72
2.0 log for viruses. From Table 3.2.9-2b the required minimum C X I (mg/L . min) 146 104 78 52 39 26
9
for Giardia by free chlorine disinfection is 51 and 0.48 for ozone. According to Table Ozone 6-9 0.97 0.63 0.48 0.32 0.24 0.16
3.2.9-3, the C X I requirement for 2 log inactivation of viruses by free chlorine is 6 Chlorine dioxide 6-9 21 8.7 7.7 6.3 5 3.7
and 0.9 by ozonation. However, if chloramines are used for the final disinfection, the Chloramines 6-9 1270 735 615 500 370 250
C X I must be 690 for Giardia and 1243 for viruses. The maximum level of resid- (preformed)
ual allowed by the EPA is 4 mg/L for both free chlorine and chloramines.
(b) 0.5 log (68%) lllactivalioll
Disinfection By-Products All disinfectants react with any element that is oxi- Free chlorinea 6 25 18 13 9 7 4
dizable in water. Thus the water supply industry is now faced with the problem of 7 35 25 19 13 9 6
controlling disinfection by-products. THMs and other halogenated organics are pro- 8 51 36 27 18 14 9
duced by chlorination. Aldehydes are formed by ozonation; other compounds, such 9 73 52 39 26 20 13
as haloacids, haloacetonitriles, haloketones, and bromate, may also be formed, de- Ozone 6-9 0.48 0.32 0.23 0.16 0.12 0.08
pending on the type of elements present in the process water. Although the magni- Chlorine dioxide 6-9 10 7 5 3.3 3 1.7
Chloramines 6-9 690 363 337 250 181 130
tude of the ozonation by-products is generally below the level of health concerns,
it may be only a matter of time before science discovers which ozonation by- (preformed)
products significantly affect human health. The EPA has currently set the maximum Suurce: Adapted from EPA's Guidance Mal/llal (March 31. 1989). . .
allowable level ofTHM as 0.08 mg/L. However, this level will probably be reduced ae X t values will vary depending on the free chlorine concentratio~. Values mdlcated are for 1.0 mglL
to 0.04 mg/L in the near future. free chlorine. (For other free chlorine concentrations, see EPS's GlIldtlllCe Malluol.)
 300 DESIGN OF BASIC TREATMENT PROCESS UNITS
BASIC APPROACH 301

TABLE 3.2.9.3 C . t Values to Achieve Virus Inactivation

dioxide, and ozone disinfecting systems, as established by the EPA and slightly mod-
Log ified, are listed in Table 3.2.9-4. The cost does not include the construction of sup-
Inactivation Temperature (OC) porting systems, nor does it include the construction of the contact tank. As Table
at pH 6-9 0.5 5 10 15 20 25 3.2.9-4 clearly indicates, construction of the ozonation system is by far the most ex-
Free chlorine 2.0 6 4
pensive of the four alternatives. Refer to Table 3.2.9-5 for the total cost of disin-
3 2
3.0 9 6 4 3 3
fection, including the operation and maintenance of the feed system for each of the
1
4.0 12 8 6 4 3 four alternatives. These costs were developcd by the EPA.
2
Ozone 2.0 0.9 0.6 0.5 0.3 0.25 0.15
3.0 1.4 0.9 0.8 0.5 0.4 0.25 Design Criteria The dcsign criteria for the chlorination, chloramination, and chlo-
4.0 1.8 1.2 1.0 0.6 0.5 0.3 rine dioxide systems are discussed below. The fourth alternative disinfection system,
Chlorine dioxide 2.0 8.4 5.6 4.2 2.8 2.1 ozonation, is discussed in Section 4.8. Therefore, no attempt is made to describe it here.
1.4
3.0 25.6 17.1 12.8 8.6 6.4 4.3
4.0 50.3 33.5 25.1 16.8 12.6 8.4 Chlorination System (Note: I kg = 2.2 lb.)
Chloramines 2.0 1243 837 643 428 321 214
(preformed) 3.0 2063 1423 1067 712 534 356 Dosage I to 5 mg/L (2.5 mglL average).
4.0 2883 1988 1491 994 746 497
Number of chlorine feeders Minimum of two, one standby is required.
Source: Adapted from EPA's Guidal1ce Manuul (March 31, 1989). Standard-sized feeders are 10, 20, 30, 50, 75,
100, 150, 250, 500, 1000, 1500, 2000, 3000,
4000, 5000, 6000, 8000, and 10,000 Ib/day
the process water arc very significant with respect to etTective disinfection because and in the 20: I feed range.
the microorganisms living within these materials may not be killed. Thus, water eo- Evaporators These are normally required if the daily dosage
tering the disinfection process must have both high quality and low turbidity. The exceeds 1200 Ib/day. A safe yield of chlorine
water should therefore be tiltered prior to tinal disinfection so that the disinfecting gas from one I-ton cylinder is 400 Ib/day
agents can work effectively without producing significant amounts of by-products. without an evaporator under normal tempera-
ture.
Safety of Operation All disinfectants are harmful to human health and animal life. Gas cylinder weighing scale Minimum of one, as required by regulatory
The design and handling of the disinfectant system should therefore adhere strictly agencies.
to the local, state, and federal (OSHA) safety rules and codes, in addition to those Chlorine expansion tank Required for liquid chlorine line.
listed in the chemical handling manuals provided by the chemical industries. If a Chlorine gas strainer Required.
chlorination system is located in a residential area, a neutralization system (for chlo- Chlorine gas pressure Required.
rine gas) should seriously be considered, to combat chlorine gas leakage. The deci- regulation valve
sion to integrate a neutralization system into the overall plant design should be Chlorine gas pressure gauges Required.
weighed for the sake of public safety, even though the governmental codes do not Chlorine eductors Required.
specify their usc. Certain municipalities also prohibit the transport of chlorine gas to
specific areas of the city because of safety considerations. Under these circumstances,
the choice of disinfectant is limited to hypochlorites of calcium or sodium. However,
TABLE 3.2.9-4 Cost of Constructing Disinfection Systems ($)
hypochlorites may be generated electrolytically from brine or salt solutions using
specially designed cells and an ordinary power source. The locations of the ozone Plant Size Chlorine Chloramine Ozone Chlorine Dioxide
manufacturing facilities and the ozone contact basins arc also dictated by safety con- (mgd) (2 mg/L) (2 mg/L) (1 mg/L) (1 mg/L)
siderations. These are generally isolated from the operations building because of the 1 15.000 20,000 55,000 25.000
potential for lire and health hazards. No matter the type of disinfectant, plant oper- 10 32,000 40,000 150,000 50,000
ators should plan and frequently practice safety maneuver drills. 100 130.000 145.000 800,000 170,000
150 150,000 170,000 1,200.000 200,000
Cost Cost is a prime consideration in selecting the type of disinfection process. Source: Data from EPA Publication 60012-79-1 62b. August 1979. The costs are based on ENR Index
The approximate construction costs for the chlorination, chloramination, chlorine 2850.
 BASIC APPROACH 303

302 DESIGN OF BASIC TREATMENT PROCESS UNITS

"C (60"P) This feature is vital if a
ture0 f 15 ' .
TABLE 3.2.9·5 Costs of Disinfection (¢/IOOO gal) s stem does not have a chlonne evaporato~,
E : wash, shower, and gas masks ?f the c~mster
Plant Chlorine
Ozone (I mg/L) Safety equipment ytype must be provided in complIance WIth
Size Chlorine Chloramines Dioxide
(mgd) (2 mg/L) (2 mg/L) Air Feed Oxygen Feed (I mg/L) OSHA regulations.
3.8 7 6 7.8 8.5 Chlorine neutralization system
Optional (Ceilcote Co., Inc., of Ohio, Ametec
Inc of Ontario, Canada, and EST Corp., 0 r
10 1.7 1.9 2.2 2.3 2.2
Qu~kertown, Pennsylvania manufacture the
100 0.48 0.9 1.4 1.36 1.7
150 0.45 0.8 1.2 1.15 1.5 scrubber). See Appendix 13. . r
All cylinders must be anchored by cham~ fo
SOlln'e: Data taken frolll EPA Publication 600/2-81-56. 1981. The costs do not include the contact tanks. Chlorine cylinders safety reasons. This measure must be Imple-
nes
mented in plants located in earthquakezo .
.' . st be proVIded at
Chlorine applIcatIOn pomts mu. d
Water supply to the eductor Dual supply of pressurized water or a standby Application points the bead of the plant, the filter mfiu~nt, an
booster pump is required. Additionally, a de- the filtered water, even though chlonne may
vice to prevent backfiow should be provided. not be applied on a continuous baSIS.
Chlorine solution Totometers Optional-as needed.
Chlorine gas leak detectors Required. . f hlorine feed system. d
Chlorine residual analyzers Required. Figure 3.2.9-2 is a schematIC 0 a c d f the reed system is generally dissolve
m
The measured chlorine gas rdease ro t nsported to the application point
Chlori De containers One-ton cylinders are preferred whenever the dr ullc eductor, th en ra h d'
f
in water by means 0 a Y ,ah 000 fL) However, in cases where t e IS-
plant uses over 200 Ib/day of chlorine. A
as chlorine solution (approXImately 3 . mg 1'50 ft (45 m) the eductor is placed
I-ton cylinder weighs 1.9 IOns, including the d feed pomt IS over • h h
weight of the cylinder. A 20-ton stationary tance between the f eed er a n . the chlorine gas is pulled t roug
, U d r these CIrcumstances, (.
tank or a 17-lOn tank trailer or a 90-lOn tank at the injection poml. n e , ' s s u r e for reasons of safety m case
the pVC reed pipe under a slIghtly negatIve P~~ematic in regions experiencing cold
car system is used for large plants.
the pipe is cracked). This scheme becomes pro
Housing Chlorine storage and feeder rooms must not be
located underground. The cylinders and feed-
ers should be located in separate rooms. The
inventory should include enough chlorine
cylinders for a minimum of 15 days of opera- R010METER

tion. "'"
Ventilation Each room must have an air ventilation system LIQUID CHLOfUNE SYSTEM
CHLORINE GAS SYSTEM
that provides one complete air change per
1OCHLORINE
minute. The air outlets should be located near a, VACUUM FEEDER l'a,
GAUGE EXPANSION If
the fioor and exhausted in a safe direction. "'''
The air inlet should be through gravity louvers
J_,,!2:~=':'::::::ll..l--::;~
0QWl:I SLO 10 FEEDER
CHLORINE
RUPTURE DISC
located near the ceiling. I'

Inspection window The door and wall of both the chlorine storage ,o,uTOMATIC SWlTCHOVER
ANDVACUUM REGULATOR
and feeder rooms should have clear glass win- UNIT(TYP)

dows to permit viewing without having to en-

f_iw~ BACK Flem
ter the rooms. PREVENTER

Access door All access doors must be outside the building. If

access is required from within the main build- UTILITY WATER

ing, the door must have an airtight seal. LOAD CELL

Heating The chlorination building must have a heating Figure 3.2.9~2 Chlorine feed system.
system that will maintain an ambient tempera-
 304 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 305

winter weather because the cold temperature around the feed pipe causes the warm
~a~ to condense, forming a very strong chlorine solution. Consequently, the gas flow
IS Interrupted. Moreover, the potential for a severe chlorine leak is created due to
the dissolving of the PVC joint solvent. For these reasons, design engineers must
'I'IUSUI'IE'VACUUM ""MOV.... la
specify the feed pipes to be well insulated, and the feed line must be encased by an IIIEUE.v...... vE

additional pipe.
YENT A.Oya
IULDtlQ,
M"RATION
,,,......
~""'"
FOR DlFFUIlI'I

YMOI'IlIIfl\II'IN UNf CO\.UMN

Chloramination System In the chloramination system, ammonia is added prior

to or af~er chlOrInatIOn. The source of ammonia may be anhydrous ammonia, aqua
ammOnIa, or ammonium salts such as ammonium sulfate. However, the use of am.
[
moni~m salts is limited to small installations because of their hygroscopic nature
and high cost; the cost does not outweigh the fact that ammonium salts are safe to }
handle. Generally, aqua ammonia is safer to handle than anhydrous ammonia and
."
CONNECTION
LO,t..OCElL VlN1'UfII n,.
is also generally more cost-effective, due to its wide availability. In regions where ''''"
IoIE"1II

the ambient temperature exceeds 90°F (32°C), it is recommended that 25% or even
20% ammonia solution be stored in a tank located outside to avoid problems asso.
ciated with ammonia vaporization.
The main design criteria for an aqua ammonia feed system are listed below.
Dosage 0.2 to 1.5 mglL (25 to 30% of chlorine residual).
Number of ammonia feeders Minimum of two; one standby is required.
Type of feeder Solution feeder (metering pump) for 25 or 30%
strength aqua ammonia.
Storage tank(s) A minimum of 15 days' storage: a minimum
shipment of 4000 gal (15 m3 ) if in bulk ship-
ment. The tank(s) must be composed of steel or
fiberglass. The tanks must be painted white and
have a sun shield to prevent the contents from
heating up. The tank will be slightly pressured
(25 psi or 1.75 kg/cm') to prevent the escape of
ammonia vapor.
Safety features Eyewash, shower, and a dike around the storage
tank for containment purposes.
Feeding ammonia Ammonia will react with the hardness of water
and form scale. Thus the ammonia solution is
fed from above the water level, if possible, to
avoid clogging the feed pipe.
Figure 3.2.9-3 illustrates the aqua ammonia feed system.
Chlorine Dioxide System Chlorine dioxide is prepared by reacting sodium chlo-
rite with chlorine gas in a glass column packed with ceramic Rashig rings. A self-
contained chlorine dioxide feed system is commercially available. Bottled chlorine
dioxide gas is not commercially available, due to its explosive nature. Theoretically,
a sodium chlorite solution containing 1.34 kg of sodium chlorite will react with
0.5 kg of chlorine to produce I kg of chlorine dioxide. If the reaction is to be com-
Figure 3.2.9-3 Aqua ammonia feed system.
pleted in a short period of time, say, less than I min, the pH of the mixture should
range from 2 to 4. Acid feeding is therefore usually required.
Dosage 1.5 to 3 mglL, but depends on the chlorine
dioxide demand of water (EPA limits the max-
imum residual to 0.8 mglL and 1.0 mglL for
chlorite).
Number of feeders Minimum of two; one standby is required.
Sodium chlorite solution tank A fiberglass tank that is large enough to produce
a I-day supply of 17% solution at maximum
dosage.
Number of chlorite tanks Minimum of two.
Chlorite feeders A solution metering pump with a minimum of
two feeders; one standby is required.
Safety features Eyewash, shower, and gas masks of the canister
type as dictated by OSHA regulations.
Additionally, some caution must be exercised
when handling sodium chlorite because it be-
comes combustible on contact with organic
matter.
Refer to Figure 3.2.9-4 for a schematic representation of the chlorine dioxide feed
system.
Hypochorination Hypochlorination has been practiced in Europe since the mid-
nineteenth century. Yet hypochlorination only began in the United States starting in
the late 1920s, when a dry bleaching chemical called calcium hypochlorite (con-
taining about 70% available chlorine) became obtainable under the name of HTH.
 306 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 307

~.o. SOLU"O~ DI!<:".",,£

L'~[ TD ",OCU5 Since additional caustic soda is added as a stabilizer, the commercial product has a
very high pH (greater than 11.2). The common strength of sodium hypochlorite used
in hypochlorination is either 12.5% or 15% solution. Solution strength is usually ex-
pressed as percent by volume (trade percent). However, the weight percent of avail-
able chlorine is important for the disinfection process. Weight percent is detined as
the trade percent divided by the specific gravity of the solution. For instance, a 15%
solution (trade) has a specific gravity of l.l68 at 20°C (68°F); thus the weight per-
cent of this solution is 12.84% and it contains approximately 1.25 Ib of available
chlorine per gallon. The most popular strength of sodium hypochlorite for disinfec-
tion purposes is 12.5 % since it contains approximately I Ib of available chlorine per
gallon and the stability of the solution is better than a 15% solution.
A major problem with the sodium hypochlorite solution is its significant loss of
available chlorine within I week unless the solution strength is less than 5% and the
temperature is below 10°C (50°F). The deterioration rate increases with higher so-
lution strength and higher temperatures. The 12.5% solution will lose a fourth (25%)
of its available chlorine content within 20 days during summer months. The stabil-
ity of the solution is also greatly affected by light, pH, and heavy metal cations such
as iron, copper, and nickel. Therefore, the dilution water should not have more than
0.3 mg/L of iron and no more than 1 mg/L of copper. The freezing temperature of
wun ~(SSO~(
R[GUlAT,OtG VAlVr
sodium hypochlorite are -12°C or 10°F for a 10% (trade concentration) solution,
-l7°C or 2°F for a 12.5% solution, and -22°C or -8°F for a 15% solution.
WUER SVP ....y
The use of granular (dry) calcium hypochlorite for hypochlorination is limited
primarily to developing countries, where the supply of sodium hypochlorite is scarce
Figure 3.2.9-4 Schematic illustration of chlorine dioxide feed system produced by Wallace
& Tiernan Company. or, in the United States, to the disinfection of newly constructed til tel'S, clearwells,
and distribution lines. The cost of 100 Ib of calcium hypochlorite (70% solution)
ranges from $110 to $130. Under normal conditions the solution will lose only 3 to
Liquid bleach (sodium hypochlorite) also became available about the same time and 5% available chlorine per year. The decomposition releases oxygen and produces
is widely used today (Clorox) for home laundry and general disinfection purposes. high heat (177°C or 350°F). It should be noted that hypochlorination can make the
Although sodium hypochlorite is approximately three times more expensive than pH of the water higher than its original pH when the buffer capacity of the water is
chlorine gas and deteriorates rapidly in a warm to hot environment, many water treat- low (low alkalinity), thereby affecting the optimum pH of alum flocculation under
ment and wastewater treatment plants have converted from chlorination to hypochlo- prechlorination as well as the C X f value of final disinfection.
rination in recent years. The two main reasons are (I) safety concerns with regard to
chlorine by regulatory agencies, backed by the public; and (2) increasing difficulty Dosage I to 5 mg/L (3 mg/L) average.
in the use of chlorine due to the risk management rule promulgated by the EPA. Number of feeders Minimum of two; one standby is required.
Type of feeders Diaphragm pump, peri static pump, metering pump with
Sodium hypochlorite: NaOCI + H2 0 --> HOCI + NaOH
100: 1 motor speed control, magnetic flow meter with
Calcium hypochlorite: Ca(OCI), + H2 0 --> 2HOCI + Ca(OH), needle valve flow control with pump recycling (padded or
pressurized storage tank is not recommended, based on
However, HOCI will convert to hypochlorite ion in water with a high pH (refer to experience).
Figure 3.2.9-1). Type of valves Diaphragm valve. If a ball valve is used, placement of a
Almost all water purveyors utilize sodium hypochlorite because it is easier to vent between the valves is necessary to release the pres-
handle and poses fewer problems when spilled. Its cost ranges from $0.27 to $0.35 sure caused by oxygen accumulation.
per pound of available chlorine, depending on the conditions. Sodium hypochlorite Pipe lines Schedule 80 PVC is most commonly used. Provide air
is produced by a reaction between caustic soda and chlorine gas: (oxygen) vents at strategic points to release gas pressure.
Hydraulic eductors identical to the chlorination system to
2NaOH + Ch --> NaOCI + NaCI + H2 0 dilute the solution. Transporting to the application points
308 DESIGN OF BASIC TREATMENT PROCESS UNITS
can cause heavy scale formation in the pipeline if Water
hardness exceeds 100 mg/L; softened water is required.
Diffuser In general, provide a perforated pipe across the pipe or
channel (see Figure 3.2.3-3, type A or B). However, high
hardness of the treating water will form scale at each ori-
fice hole. Provide a special means of cleaning scale or
two diffusers so that one diffuser may be removed for
cleaning when it is clogged.
Storage tank Minimum of two tanks for maintenance. Capacity of tank:
Alternative A: Enclosed room with air conditioning:
lOOC or 50°F maximum. 6000 gal (23 m') per tank for
tank truck delivery or 11,000 gal (42 m') per tank for
large plants. Note: The tank truck has either a 4000-gal
(lS-m') or SOOO-gal (l9-m') tank.
Alternative B: Tanks are exposed to ambient
temperatures. One tank capacity should be limited to
10 days' usage during summer months.
NOTES:
I. Both sodium and calcium hypochlorite produce oxygen under certain conditions; there-
fore, the piping system becomes pressurized, giving rise to potential leakage and break-
age problems.
2. Conversion of regular chlorination systems to hypochlorination systems requires certain
considerations and evaluation of the existing system, including storage, the type of valves,
type of metering pump, size of the feed line. vents, and the quality of the dilution water.
Disinfectant Contact Tank The design of the disinfectant contact tank is contin-
gent on the type of primary disinfectant selected. The eXt values provided by the
EPA clearly indicate that the size of the contact tank drastically varies among chlo-
ramines, ozone, and chlorine. Nonetheless, the fundamental design rule is to mini-
mize flow short-circuiting within the contact tank.
A recent EPA draft guideline manual suggests that the actual contact time of a con-
tact tank be verified by running tracer tests. When these tests are performed, they es-
tablish that most existing c1earwells, which serve as contact tanks, have a severe flow
short-circuiting problem and that the actual flow-through time is often only 10% of the
calculated detention time during conditions with lowest water depth and highest water
demand. The contact tank must therefore be designed to have as near a plug flow con-
dition as possible. One of the best tank designs is identical to a baffled channel type
of flocculation tank (refer to Section 3.2.4). Refer to Appendix 9 for a tracer test.
Number of tanks An absolute minimum of one; two tanks are preferred.
Type of flow path A serpentine flow path with or without baffles.
Water depth 12 to 20 ft (3.5 to 6 m).
BASIC APPROACH 309
End of the tank An overflow wall must be provided to maintain a constant
water depth.
Detention time This depends on the type of disinfectant and the eXt
requirement.
Dechlorination Recent regulations do not allow water with residual chlorine to be
discharged into public water courses, to protect fish inhabiting these waters.
Dechlorination is required when residual chlorine is too high; MCL is 4 mg/L for both
free residual and chloramine residuals. The most practical method of dechlorination
is by sulfur dioxide (SO,) or sulfite compounds such as sodium bisulfite (NaHSO,)
and sodium metasulfite (Na,S,O,). The sulfur dioxide dechlorination process is shown
below. The first reaction is the formation of sulfurous acid in the sulfonator injector:
SO, + H,O --> H,SO,
Sulfurous acid then reacts with various chlorine residual species as illustrated below.
For the free chlorine residual: HOCI + H,SO, --> HCI + H, S0 4
For monochloramine:
As shown above, acids are produced, thereby lowering the pH of water, and 2.8
mg/L of alkalinity is lost for each part of chlorine removed.
The stoichiometric relationship requires 0.9 part of SO, to remove 1 part of chlo-
rine residual, but a I : I ratio is used for the actual declorination practice. The re-
action between sodium bisulfate and chlorine residual is as follows:
NaHSO, + CI, + H,O --> NaHS04 + 2HCI
Each part of the chlorine residual reacts with 1.46 parts of sodium bisulfate.
Activated carbon can adsorb chlorine so that dechlorination can be achieved. However,
this is not the most practical method, due to capital cost and other limitations.
Example Design Calculations
Example 1
Given A new SO-mgd or 2.2-m'/s plant (daily average flow rate) located in Denver,
Colorado, requires a new chlorine disinfection system. Assume that the minimum,
average, and maximum dosages are 1.5, 2.5, and 5 mg/L, respectively.
Determine
(i) The considerations for the system design
(ii) The number and capacity of the chlorine feeders and the appurtenant units
(iii) The size and number of chlorine containers (storage)
 310 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 311

Solution (i) There are various design considerations for the new chlorine disin_
fection system: the potential interruption of the chlorine supply due to severe win-
tcr weather conditions; the extreme toxicity of chlorine; the need to adhere strictly
to the safety codes and EPA regulations; specifying a large curvature for the access
road, in addition to providing adequate maneuvering space in front of the chlorine
cylinders; providing adequate height (14 ft) below the monorail to facilitate the Un-
loading of chlorine cylinders from delivery trucks; and providing a heating system
in the chlorine storage and feeder rooms.
(ii) The number and capacity of the feeders are as follows:

maximum plant flow rate = 1.5 X 50 = 75 mgd

maximum dosage = (5 X 8.34) X 75 = 3128 Ib/day (1422 kg/d)
average dosage = (2.5 X 8.34) X 50 = 1043 Ib/day (474 kg/d)
minimum dosage = (1.5 X 8.34) X (0.25 X 50) = 156 Ib/day (71 kg/d)
= = 20:

1
ratio of maximum to minimum dosages 3128: 156 I

Chlalne I' ~
detector \, I / Ir---="flOiO. IlItOin.

~ I crfi"i"
NOTE: Commercially available chlorine feeders have a 20 : I feed range for manual ~l~n~r{h 5~f:J
angle slops
control and a 10 : I feed range for automatic control. <y"odm

Based on these figures, select three identical 1500-lb/day (700-kg/day) feeders.
This choice allows for maximum and minimum dosages, and the third feeder serves
as an automatic control standby unit. Additionally, provide two identicaI4000-lb/day
(l800-kg/day) chlorine evaporators. The maximum chlorine dosage cannot be han-
dled without the use of evaporators; a safe yield of chlorine gas from a I-ton cylin-
der is 400 Ib/day (180 kg/d) without heating.
(iii) The optimum storage of chlorine is based on the maximum dosage of the
disinfectant at the average plant flow rate. Since the delivery of chlorine may be in-
terrupted during winter months due to snowstorms, a 2-week stockpile must be pro-
vided under the conditions described above; the plant flow rate during the winter
months is less than half the average annual rate. Thus actual storage (under normal
dosage) will be approximately 60 days.
[(5 X 8.34) X 50J 15 = 15.6 tons, say, 16 tons
The storage area should also accommodate the empty I-ton cylinders. Since each
truck carries a minimum of eight cylinders, provide eight additional spaces for the
empty cylinders. Consequently, a total of 24 storage racks (for the I-ton cylinders)
should be provided in the storage room. Figure 3.2.9-5 is an illustration of both a
typical chlorine storage room and a feeder room.
Operation and Maintenance The operation of the disinfection process is con-
tingent on the type of disinfectant that is employed, the plant size, regional charac-
teristics, and the raw water characteristics. Since chlorine is the primary disinfectant
7f10in·~3f14in.
,-,=_ _ _ 1
Typical Cross Section
ChtOfine Storage Room
Figure 3.2.9-5 Typical cross sections of chlorine storage room (bottom) and chlorinator
room (top).
in water treatment, this discussion is centered on the process of chlorination. If
prechlorination can be practiced without producing significant levels of THM and
HAA, a residual chlorine concentration of approximately 0.3 mg/L should be main-
tained in the settled water at all times. This practice will prevent the growth of algae
on the basin walls and will prevent the growth of microorganisms in the filter bed.
The level of the residual chlorine in the c1earwell should be at least 0.5 mg/L so
that the chlorine residual at the end of the distribution system will be 0.2 to 0.3
mg/L at all times. Plant operators must record the level of residual chlorine in the
cIearwell and distribution system at regular intervals, since a log of these figures
must be submitted to the local regulatory agency on a monthly basis. The chlorine
feeders and appurtenant systems must be operated and maintained in strict accor-
dance with the operation manual furnished by the equipment manufacturers and the
manual provided by the Chlorine Institute; this includes the handling of chlorine
cylinders. If the chlorine feed system should fail, the unchlorinated water must not
be allowed to enter the distribution system because the Safe Drinking Water
Regulations demand a minimum residual chlorine of 0.2 mg/L to be present at the
end of the distribution system .

.,~.,

312 DESIGN OF BASIC TREATMENT PROCESS UNITS
Maintenance of the chlorination system requires a thorough understanding of the
literature provided by the manufacturer, in addition to the operation and mainte-
nance manual prepared by the design engineers. The daily, weekly, and monthly op-
eration and maintenance procedures are generally described in these manuals. It is
imperative that operators not attempt to perform any unfamiliar maintenance work
or work that is beyond their qualifications. Many serious accidents have occurred
because someone has ignored this very basic rule.
In the case of chlorine leakage, corrective actions must only be taken by trained
operators wearing the appropriate safety equipment. It is essential that all plant op-
erators be trained to repair chlorine leaks. Operators should always work in pairs
when searching for and repairing chlorine leaks. All other personnel must be evac-
uated until the conditions are safe. If the chlorine leak is large, the fire department
and police must be notified immediately~ all inhabitants in the adjacent area must
be warned and evacuated.
Alkali chemicals such as caustic soda and lime may be used to neutralize very
small chlorine gas leaks. For example, stoichiometrically, 1.13 Ib of caustic soda is
required to react I Ib of chlorine. Therefore, a I-ton chlorine leak would theoreti-
cally require 2260 Ib of caustic soda. However, the scrubber manufacturers recom·
mend 3600 gallons of 7% caustic soda. In the case of hydrated lime, 125 Ib is re-
quired to neutralize a I OO-Ib cylinder of chlorine, and 2500 Ib is required to neutralize
I ton of chlorine under ideal mixing conditions.
It is extremely important that treatment plants establish a formal safety program.
All personnel should perform "hands-on" training periodically using the safety
equipment. The American Water Works Association provides training materials, in·
c1uding videos. The Chemical Transportation Emergency Center (CHEMTREC) pro-
vides immediate advice for those at the scene of an emergency and quickly alerts
experts for more detailed assistance and appropriate follow-up whenever the situa-
tion becomes uncontrollable. CHEMTREC has a toll-free number (1-800-424-9300)
that all plant operators should keep in a readily accessible place. An example of a
chlorine gas scrubbing system is shown in Figure A 13-2.
As described in Section 4.1, if any plant uses, handles, or stores more than 2500
Ib (1135 kg) of chlorine and 5000 Ib (2270 kg) of sulfur dioxide at the plant site, the
plant operators must be familiar with and comply by the Risk Management Planning
(RMP) rule established by the EPA; this rule has been in effect since June 1999.
BIBLIOGRAPHY
Aieta, E. M., et al., "A Review of Chlorine Dioxide in Drinking Water Treatment," 1. AWWA,
78:6:62 (June 1986).
AWWA Committee Report, "Disinfection," J. AWWA, 74:7:376 (July 1982).
Barnes, D., "Alternatives to Chlorination for Water Disinfection," Water J. (Australia), p. 12
(September 1983).
Chiu, K. P., et aJ., "Effect of UV System Moditications on Disinfection Performance,"
J. Ellviroll. Ellg. Div. ASCE, 125:5:459--469 (May 1999).
BASIC APPROACH 313
Clancey, 1. L., et aI., "UV Light Inactivation of CryptosporidiufI/ Oocysts," 1. AWWA, 90:9:92
(September 1998).
Dyksen, 1. E., et aI., "Advanced UV Light for Inactivating O),ptosporidillll1 Oocysts;'
J. AWWA, 90:9:103 (September 1998).
EPA, Municipal Wastewater Disinfection: Design Manual, EPA/62SII·86-021, U.S.
Environmental Protection Agency, Washington, DC, October 1986.
Glaze, W, H.. et al., "The Chemistry of Water Treatment Processes Involving Ozone, Hydrogen
Peroxide, and Ultraviolet Radiation," 1. Ozone Sci. Eng., 9:335 (1987).
Hoff, J. C, The Relationship of Turbidity to Disinfection of Potable Wate,;" EPA/370/9-78-
DOc, U.S. Environmental Protection Agency, Washington, DC, 1977.
Hunter, G. L., et al.. "Medium-Pressure Ultraviolet Lamps and Other Systems Are Considered
for Wastewater Application," J. Water Engineering and Technology, pp. 41-44 (June 1998).
Longley, K. E., et aI., "The Role of Mixing in Wastewater Disinfection." Proceedings of the
ASC Meeling, Anaheim, CA, 1978.
Marshall, T., "Deadly Pulses," 1. Water Engineerillg and Technology, pp. 36-41 (April 1999).
Metcalf & Eddy, Wastewater Engineering: Treatment, Disposal, Reuse, 2nd cd., McGraw-
Hill, New York, t979.
Morris, 1. C, Aspects of the Quantitative Assesslllent of Germicidal Efficiency Disinfection:
Water and Wastewater, Ann Arbor Science, Ann Arbor, MI, 1975.
Parrotta, M. C, and Bekdash, "UV Disinfection of Small Ground Water Supplies," 1. AWWA,
90:2:71 (February 1998).
Shull, K. E, "Experience with Chloramines as Primary Disinfectants," J. AWWA, 73:2:101
(February 1981).
White, G. C, Handbook of Chlorination, Van Nostrand Reinhold, New York, 1972.
3,2,10 Filtered Water Reservoir (Clearwell)
Purpose In the United States the finished water reservoir in the water treatment
plant is customarily called the clearwell, There are four basic reasons for requir-
ing c1earwells: (I) to meet water peak demands, (2) to provide a sufllcient volume
of water for plant operations including filter washing, (3) to ensure adequate chlo-
rine contact time (C x I), and (4) to store enough water for lirelighting. Water
treatment processes are designed based on the maximum daily water demand rate,
not the maximum hour rate; therefore, an adequate amount of water must be stored
in the clearwell to meet the peak demand (maximum hour rate). Usually, the max-
imum hour water demand is 1.5 to 2.0 times the maximum day demand, and its
duration and demand vary due to the local conditions, Treatment plants must use
a certain amount of water to service the operational needs, and filter washing re-
quires the largest use of water; filter wash can be several flIters in succession. All
water treatment plants must meet the eXt requirement for the disinfection process
by the Surface Water Treatment Rule established by the EPA. In general, a c1ear-
well should provide the required chlorine contact time and have proper baftling in-
side the c1earwel1 to minimize flow short-circuiting and dead water zones. The stor-
age of sufficient water for firefighting is a very important factor for small
communities that do not have adequate local water distribution tanks, Adequate
 314 DESIGN OF BASIC TREATMENT PROCESS UNITS
water storage in the clearw~1I improves filtered water quality and filter operation
smce the filters do not reqUIre frequent changes in filtration rates to meet the peak
demands.
Considerations Clearwell storage should be sized in conjunction with the dis-
tribution system storage in the city. The mass diagram or Ripple method for the de-
t~rmination of storage requirements for an impounding reservoir may be applied to
size a clearwell. One of the important considerations is the location of the clear.
well. The clearwell should be located above the possible flood level and the bottom
slab should be placed above the groundwater table. At least 50% of the water depth
should be above grade, if possible. Sewers, sludge lagoons, a wash-waste holding
tank, or a Similar source of possible contamination should be kept away from the
clearwell [a minimum of 50 ft (15 m)l to avoid any contamination of finished wa-
ter. Recent EPA regulation does not allow the use of an open-top reservoir to store
finished water, the clearwell must be covered, and the top of the clearwell should
be at least 2 ft (0.6 m) above the finished grade of the plant or any possible flood
level. The clearwell cover should be a suitable watertight roof that prevents the en-
try of birds, animals, insects, and excessive dust.
The c1earwell also requires drains, overflow protection. and access with hand rail-
ings. The vents should prevent the entrance of surface water and rainwater (use of
an inverted-V pipe is recommended), and the opening level should be about at least
2 ft (0.6 m) above the top of roof; the opening should be covered with a fine SS-
316 insect screen (No. 18 or I-mm openings).
Pipe connections to the clearwell should have a reliable flexible coupling or an
artICulated pipe arrangement to avoid breakage of the pipe connection to the side-
wall of the clearwell. Many pipe connections to the reservoir have broken in earth-
quake-prone regions such as California and Japan, due to inadequate flexibility in
the pipe connection to the .sidewall of the clearwell. The top of the clearwell may
be utilized as a public service area, for example, a tennis court, baseball court, and
even a parking lot. However, safety against vandalism and potential pollution of fin-
Ished water IS a concern. The local laws and regulations of the site area should be
examined during the planning stage.
Treatment plants that incorporate ozonation as a part of the treatment process do
not need to meet the regular eXt requirement for the required final disinfection
by chlorination because of the highly effective disinfection by ozone upstream of
the final chlorine disinfection process. Therefore, a few treatment plants provide a
smaller clearwell capac.ity than the regular chlorination practice. However, this prac-
tIce IS not a robust deslgn~ ozonation systems can fail due to natural disasters such
as earthquakes, tornadoes, hurricanes, and even sabotage. For instance, a 600-mgd
(26-m'/s) water treatment plant in Los Angeles could not operate its ozonation sys-
tem after the 1995 Northridge earthquake (magnitude 7.3) for about 3 weeks due to
damage of the oxygen production facilities and the power supply system to the ozone
generators. Regardless of whether or not an ozonation system is included in the
treatment process, the size of the clearwell should be large enough to perform chlo-
rine disinfection in case the ozonation system fails; this design provides good reJi-
BASIC APPROACH 315
ability. If the clearwell is an aboveground tank fabricated from steel plate, the de-
sign should follow the current AWWA Standards. Vse of steel tanks requires that
they have a protective coating both inside and outSide, and engIneers should also
consider to provide cathodic protective system to prevent corroSIOn of steel plates.
The painting system should be in accordance with NSF Standard 61 and be accepted
by the local state authority. .
All new clearwells must be disinfected in accordance With current AWWA
Standard C652. However, the heavily chlorinated water should be carefully disposed
of since it can kill the fish if disposed of into the water course without dechlorina-
tion. Also, the water may contain a high level of DBPs. One of the purposes of a
clearwell is to fulfill the final disinfection process with the necessary eXt re-
quirement. However, traditional clearwell design has often disregarded the flow char-
acteristics in the tank so that severe flow short-clrcUltmg as well as stagnant zones
are found. Knowing this condition, SWTR allows only 10% of hydraulic detention
time as the reaction time of t unless a tracer test to come up with actual flow-through
time of existing clearwells. SWTR also shows the baffling classifications table as a
general guideline of actual disinfectant time efficiency factor in the clearwell, which
is expressed as TldT, as shown below, where TIO is the actual detentIOn time (flo,",:-
through time) and T is the hydraulic detention time, which is the volume of tank di-
vided by the flow rate.
BAFFLING CLASSIFICATIONS
Baffling Condition TldT Baffling Description
Unbaffled (mixed flow) 0.1 None, agitated basin, very low length/width
ratio, high inlet and outlet flow velocities
Poor 0.3 Single or mUltiple unbaffled inlets and outlets,
no intrabasin baffles
Average 0.5 Baffled inlet or outlet with some intrabasin
baffles
Superior 0.7 Perforated inlet baffle, serpentine or perforated
intrabasin baffles, outlet weir, or perforated
launders
Perfect (plug flow) 1.0 Very high length/width ratio (pipeline flow),
perforated inlet, outlet, and intrabasin baf-
fles)
The rule of thumb for designing a good clearwell (to provide a good flow char-
acteristic for disinfection) is to furnish 2 to 4 longitudinal baffles dividing the tank
into 3 to 5 channels, assuming that the shape of the clearwell is either square or rec-
tangular. This arrangement gives each channel a length/width ratio of 5 : I to 15: I,
so that the total ratio of the tank is between 25 and 50, with a minimum water depth
of 10 ft (3 m). The chlorine contact tank portion of the clearwell should provide a
minimum of 30 min of net contact time with I mglL free chlorine residual for
 316 DESIGN OF BASIC TREATMENT PROCESS UNITS
BASIC APPROACH 317

Giardia inactivation, and the water depth is maintained by the overflow wall at the
end of the contact tank. However, If chloramination is the disinfection process
net contact time of 210 min should be provided at O.5'C (33'F) with 3 mg/L 'o~
residual for Giardia inactivation. Therefore, to determine the COrrect detention tim
the water depth must be monitored constantly to calculate the volume of Water e,
thb '(
e aSIn ' clearwell volume) necessary to meet the required C X t. Effluent
entlfe In
overflow cannot be used because the c1earwell is a flow balancing basin; thus the
wate~ level must be fluctuated. When modifying existing c1earwells to meet C X t
requlfcments, engineers should provide as many baffles as practical and also Can.
sider installing a perforated diffuser wall at the beginning of each channel wherever
the flow turns 180'. The minimum number of c1earwells is two, thereby satisfying
reliability and maintenance needs.
. Additional design considerations for the clearwell are the amalgamation (com-
binatIOn) of c1earwell structure and the pumping station, which provides high ser-
vice pumps and utility pumps, including filter wash-water supply pumps. This
scheme not only reduces construction costs but also makes a compact site layout.
The preferable design is to locate the vertical shaft turbine pumps on top of a con-
crete slab of the end portion of the clearwell, with or without housing. Since the
turb~ne pumps must be pulled out vertically for service and repair, engineers should
furnish a removable hatch on the roof of the housing directly above the pump to
~revent creating a buildi~g with a very high ceiling structure. In most cases, plants
In ~he wa~m weathe,r regions do not provide any housing over the pumps. Even in
regions with cold WInter weather, the housing is deleted if the ambient temperature
I~ not as. severe ~nd snow and ice are not great problems. This type of pumping sta-
tIOn desl~n. provides easy access to the pumps and avoids the need for a dry well,
pump pnmmg system, and floor drainage pumps; thus the construction and main-
tenance costs are generally minimized. However, a dry pump room utilizing the end-
wall of the clearwell as a part of the room with the instal!ation of horizontal cen-
trifugal pumps is also a good proven design. Centrifugal pumps usual!y last longer
However, in most cases, transfer pumps are required to transfer filtered water into
the tanks. The second type is suitable in cases where the groundwater level is rea-
sonably low and a rather large volume of water must be stored. In general, filtered
water flows into the c1earwell by gravity. Unless designed carefully, there is a po-
tential danger that the finished water could be contaminated by floodwater or con-
taminated groundwater, and there is the risk of damage to the structure or floating
due to high groundwater levels during rainy seasons. The third basic type, c1earwell
underneath the filter module, is an old design concept that has two serious draw-
backs: (I) it is very difticult or impossible to calculate reliable disinfection time due
to filter effluent flows into the clearwell at many locations, including adjacent to the
effluent side of the clearwell, and (2) c1earwel! capacity is often inadequate for dis-
infection for Giardia and especially Cryptosporidiuln oocysts .
A few c1earwel! designs furnish only one pipe connection to the clearwell; there-
fore, the water level in the tank is floating and is dependent on the amount of in-
flow and outflow. Consequently, the flow characteristics in the tank is the worst for
disinfection. Under certain conditions, such as the peak-hour period, chlorinated wa-
ter does not flow into the tank and no disinfection time is provided by the clear-
well. This type of pipe connection is not recoin mended. A c1earwell should have
both inlet and outlet connections at strategic locations, and the tank should have
a proper baffle wal! system in order to make the TldT ratio at least 0.70. Figure
3.2.10-1 shows an example of a ground-level clearwell with a reinforced concrete
structure design.
An Orifice Flow Pusage
With 3 Chemic 810muser
Pipufor;
~
I ~ 'Chlorine Itrim ming)
'Caustic Soda (prim.!V)
.~~
.2 .~
De:.
~•
'0
0
"
II:
t / Submarged Weir
(Typical!
' / ,..--- Energv Ois~pation Pipe
.Ammonia Ipri marvl

than turbine pumps. However, this type of design is generally more costly than the
deSign scheme descrIbed above and there is the potential problem of flooding in the From t t /'
(Lower Side Open)
I
--" ~fj
~\o.
t~
dry room. r-
t ~;t 't
Filters
t "- ~~a '-... ~

Type and Selection GUide There are three basic types of clearwel!: (I) ground
level up to approximately 30 ft (9 m) in height and a steel or reinforced concrete
~•
0_
"'i<'
u • .0
o~
0 0
. t
.-.'E
I
---. Overflow Pipe
("'r-- J
-g
~~
.~
tank-these are mostly cylindrical in shape; (2) ground level, a rectangular or square •" u0
deep baSIn [25 ft (7.6 m)] with a reinforced Concrete structure having vertical side-
u
Iii!!!
•
--" --"
~alls or a trapezoidal cross section with sidewalls of rather thin concrete and pe-
rIpheral walls composed of reinforced concrete that are 8 ft (2.4 m) high to support Notes: 1.1 Submerged weir controls walur level in the filter beds.
the roofing system Or an anchored floating membrane cover; and (3) a rectangular, 2.J Advantages of this scheme are:
• Smeller CxT due to free chlorina residual in the clearwell.
rather shallow [10 ft (3 m)] reinforced-concrete tank, located directly underneath • Smaller exT because of the lower pH in the clearwelt.
the filter structure. • Single chemical feed location in both inlet and outlet 01 clearwell.
. The first type of clearwell is usually built in cases where the groundwater level 3.1 Provide drain valve at 8 proper location of each tank.
ISnear the ground surface. A cylindrical tank is chosen because of less capital cost. Figure 3.2.10~1 Suggested clearweJl design and chemical feed locations.
 318 DESIGN OF BASIC TREATMENT PROCESS UNITS
BASIC APPROACH 319

Design Criteria
equipment damage because the status of each system is monitored continuously, and
Number of clearwells should the situation go wrong, the plant is automatically shut down.
Minimum of two, or one large clearwell with a divider Computerization of the treatment plant frees plant operators from monitoring the
wall to make it two cells
Capacity various gauges or charts and from updating the operational logs continuously. The
Approximately 15% of the plant design flow rate per day system required by water treatment plants is generally simple in nature, unlike
Baffling
Mlmmum TldT = 0.7 wIth proper baffling the complex systems used in the oil, chemical, and rocket industries. The safety and
security of plant personnel and the plant itself may be improved by installing a
NOTES: closed-circuit television monitoring system. The installation of modern instrumen-
tation in a treatment plant permits operators to concentrate on the overall control of
the treatment plant and on effective plant operation.
I. ~apac!ty can b: estimated as the sum of the volume for 2 h of plant operational needs
m~ludmg two tllter washes, and the volume of water necessary to meet the e x '
qUirement by free chlorine residual and the unusable bottom portion of water [usuau/ t~~ Symbols and Manufacturers The symbols used in designing the instrumen-
bottom 4 to 5 ft (1.2 to 1.5 m) water depth]. ' e tation system are based on those provided by the Instrument Society of America
2. Miscellaneous required items are described in the "Considerations" section. (IS A), the Institute of Electrical and Electronic Engineers (IEEE), and the National
Electrical Manufacturers Association (NEMA). Refer to Figure 3.2.11-1 for an ex-
ample. The field of instrumentation is a rapidly evolving technology in which new
BIBLIOGRAPHY components and equipment are increasingly being geared toward the field of water
treatment. Thus, it is very rare that a single manufacturer will he able to furnish the
Bishop
. , M' M
" et 31 ., "I [proving
n . th e D"IstmctlOn
. Detention Time of a Water Plant" J AWWA bulk of the necessary equipment. Yet it is very important to place the overall re-
85:3:68 (March 1993). ,. , sponsibility of installation and system startup with a single manufacturer. Firms that
Boulos, P., et aI., "Hydraulic Mixing and Free Chlorine Residual in Reservoirs" J AWWA have been dominant providers of service in the United States are Fisher and Poter
88:7:48 (July 1996). ,. , Company, Bristol Babcock, Inc., Westinghouse Company, and Robert Shaw Controls
Kennedy, M. S" et al ., "A ssessmg
. th e Effects of. Storage Tank Design on Water Quality" Company. Major manufacturers of instrumentation control equipment such as
1. AWWA. 85:7:78 (July 1993). ' Foxboro, Leeds, and Northrop mayor may not be involved in water treatment plant
Sykes, R. G., and Coate, A. R., "PCBs in Sealants in Water Distribution Reservoirs" 1 AWWA projects, due to the level of competition and the required bidding procedures for ser-
87: I 2: 10 (December 1995). ,. , vices of this nature. However, this is beginning to change.

3.2.11 Instrumentation and Process Control
Purpose Due to increased public awareness on the subject of safe drinking wa-
ter, regulatory agencIes have made the drinking water standards more stringent.
Co~sequently, water treatment plants must have more sophisticated processes and
eqUIpment, and plant operators are forced to operate and maintain systems that are
vastly morecomplicated than in the past. As the costs of energy chemicals and per-
sonnel c.ontln~e to increase, managers of water purveyors mu;t be conce;ned with
the quality of the wat.er supply as well as improving the efficiency of production
a~d s~lpply. One sol.utlOn to these Issues is to provide the treatment plants and dis-
tfibutlOn systems ~Jth modern. s.upervisory control and data acquisition systems.
. The ~dv~ntage m computeflz.ln~ trcat~nent plants is that plant operators are sup-
plied .wlth tImely and accurate mformallon on which to base their decisions thus
allowmg them to achieve etlective process control. Process variables such ~s the
raw water quality, plant tlow rate, chemical feed rate, and finished water quality
~na~ be mOl1Jt~red by o.n-.llne I~slrument~,. which relay the information to the oper-
tOl on a contmuous baSIS. ThIS type of Instrumentation may also help to reduce
Considerations The major considerations in designing the instrumentation and
control system are (I) the local conditions, (2) the benefits of the system, (3) the
selection of the system, (4) the needs of the system, (5) the function of the instru-
mentation and control system, (6) the availability of the system, and (7) the under-
standing and commitment of the operators and management to the system. Each is-
sue is discussed in detail.
Local Conditions The instrumentation system is a highly sophisticated item. For
this reason, the following issues must be evaluated carefully: availability of the
equipment, availability of qualified operators, follow-up services, weather condi-
tions, and the living standard of the region, because this dictates the drinking water
quality standards. Due to the high cost of systems acquisition, as well as the cost
of repair and replacement parts, the system should be operated only by highly skilled
and knowledgeable operation and maintenance personnel. Moreover, the environ-
ment of the control room must be strictly controlled to optimize the performance of
the computers (microchips); the room must be maintained at a specified tempera-
ture and humidity and have a dust-free environment. Extensive instrumentation and
 INSTRUMCNTATION SYMBOLS '" «O""ON}
DESIGNUION
SUCCEEDING LEttER

4
INsrRUlrlrNT OR OTNE" COMPONENT TO 8E
QUIPltUN r NUIrI8ER
1rI0UNTEO IN THE FIELD, SEE PUNS. FIRST lETTER
"
/'i4V60HI
INSTRUMENT OR OTNeR COIr#PONENr TO 8E
1rI000NrED INSIDE MAIN CONTROL 10ltRO 0"
OrHER PlANEL ItS NAffKEo.
IJ/ INDICATES QUANTITY OF SIGNALS IrHRCf;} T· 8V·20 '"
"
INSTffUItlENT Off OTHER COMPONENT TO IE
/tI()UNTEO ON /tiCS FRONT PltNEL 0" OTNER - - - EUCTRIC SIGNAt
EO/JIPItlENT .o.88REVI.o.TlON INsrRUItlENT •
20 PSI
PANEL AS ItIA"KCo (8l/TfERfV VALVE) LOOP 1lU1rI8ER

SINGLE INSTRUMENT OR OTHCR

COMpONCNT HAVING MU"IF'tE FIINCTIONS -t--t--t- HYDRAULIC liNE
MECHANICAL EOUIPMENT CAUOUT INSTRUMENT TAG NUMBER
INSTRUIrIENr OR OTHCR COMPONENT
ACCCSSED THROIJOH COMPUTER CffT
- PROCESS PIPING

COMPUTe" INTERFACE
~ SONIC SIGNAL

RELATED OCVlce PERFORMS liNEARIZING

OR SQUARe ROOT FUNCTION
- - - - - - CAplU.MRY rUBlNG IFlLlED SYSTEM}

REureo OCVICe PeRFORMS MUtTiJILTING

SUMItIAnNQ OR DIVISION FUNCTION.
(ii\
QCNE!tAUZCD FOR COMpUX INTeRl.OCK
/.D(jIC·sec eUCT!t1C SCHeMATICS DR
\!!)
speCIFICATIONS. AIR SUI'PLY /20P5[G -EKCEPT AS NorED}

INTERLOCK IF ANY ONE OR MORE 1

INI'UTS EXISr.
PRIMARY HccrRICAL SUPPlY (120V 50HI/
1
1
INTERLOCK IF AU INPUTS E;r(lST
• INolCItTEO COMPONENTS ARE NOT SPECITlEo
ItS PARr OF THE INSTRUMENT PACKltGE I I
STArvS OR FlU)T LIGHT, I I I
R ·REO G-GREEN A-AIrI8ER --@- C118tE (MVLTlCONOUeTOR OR CDl.XIALI
FURNISHEO WITH EQUIPlltENT I ~T MC8\ I1 t ·fLiifJ
ev CONTIN)I.UO VARIASLE PO PtltsE OURATION
1----1. I~

l-l4t~r~I:;<U'!)~-·~=~
E VOLTAGE PF PUUE F!tEouCNcr
1 CURRENT
51' SEt POINT
P PNEUMATIC

'~~~~@-~~L'll
PRIMARY CLEMCNT SYMBOLS

~
~
~ CLECTROMAGNETIC
PROPeLLER

SONIC
e>f.aOtNSlTr MtrER

6-;"
y SONIC LEVEL SENSOR

~
I
I
k
tJ:D ~
~ '0
4Z

~
1I
1._
~----.. ~ l---i :.".
_\.!;./

B
I
TEMPERItTURE HEMENT WITH WHL
\!U
~ SAMPLE LOOP OR..!'..~ {)!~~

~
VORTEX
LEVEl rRANSMITT£R FlO.o.T TYPE -----"-5.0. STANOARO 55.11

~
~
PITOT TUSE (ANNUlAR)
PRESSURE TRANSMITTER IfICHCMICAI. SEAL

~V£NTURI $~ "Wt'SCAtE WCiGHT TRANSMITTER

(HYDRAULIC TYPE}

_V-",_..,
>' FLUME
A
~""EIR

~ ROT.o.METER

Figure 3.2.11-1 Instrumentation symbols. Figure 3,2,11·1 (Colitinued)

320 321
 322 DESIGN OF BASIC TREATMENT PROCESS UNITS
contr~l systems are no~ recommended for developing countries or small plants lo~
cated In fcmote areas Slllce they generally cannot meet the criteria listed above.
Benefits of the System Although all plant operators would like to have a mod_
ern Instrumentation and control system, in each case the benefits of the system m
b . h d . ust
e wClg e agamst the ~ost factor. For instance, a simple water treatment process
such as slow sand filtratIOn cannot justify the installation of high-level instrumen_
tatIon. Yet an ozonation system, reverse osmosis process, pumping plant filt
process, and water distribution system all greatly benefit from a highly sophis;icat:~
instrumentatIOn and control system. Manual OM-6, Process InSfrtlme11lation and
COlltrol Systems, by the Water Pollution Control Federation (1978), lists a number
of benefits derived from instrumentation; refer to Table 3.2.11-1.
Selection of the System Although there are various ways to improve plant
. d h' P
~ratlOns an to ae leve the benefits listed in Table 3.2.11-1, modification of an ex-
Isling, outdated system is usually very difficult, due primarily to outdated sensors
~ransmitters. and control signalers. In fact, complete replacement of an old systeJ~
IS. oftenmore cost-,ef:ecti.ve than updating certain parts. However, the owner may
wish to keep the eXisting Instrumentation system, including the 7-day recording cir-
cular chart and the relay logic control system, as a backup to the new computerized
system and/or a training tool for plant operators. In this type of situation, the de-
sIgn engineer should respect the wishes of the Owner.
When an engineer is designing a new plant, he or she should evaluate the most
appropriate degree of instrumentation and control for that particular plant. Although
th~ state-of-the-art system may appeal to both the engineer and the owner, these ex-
otiC systems may become obsolete after several years due to rapid technological ad-
TABLE 3.2.11-1 Benefits of the Instrumentation System
Purpose
Benefits
Process
Improved process results
Efficient use of energy
Efficient use of chemicals
Automation of some process adjustments
Greater ability to control complex processes
Process changes detected in a timely manner
Personnel Timely and accurate process information
Safer operation
Increased security
Efticient use of labor
Allows for an overview of plant operations
Decrease in manual paperwork
More complete records
Equipment
Increase in running time
Status is known at all times
Automatic shutdown to prevent major damage
BASIC APPROACH 323
vances, and replacement parts may become unavailable. The basic rule in selecting
a new instrumentation and control system is to anticipate the future needs of the
treatment plant, including plant expansion, the potential addition of new treatment
processes, and owner preference or corporate policy. If the new system must be com-
patible with the systems of neighboring districts, engineers should also consider the
type of system employed by adjacent water districts and the system used by the city;
for example, determine if the city has a main computer that controls the overall diS-
tribution of water.
It is important to remember that most water treatment unit processes are slow re-
acting in nature and have high system lag times. In fact, fast response times create
unstable or undesirable operational conditions, such as "hunting" actions. Thus, de-
sign engineers should keep this fundamental nature of water treatment processes in
mind when selecting the instrumentation and process control system. Also, the units
0 _ and system to be selected should be "user friendly"; otherwise, the design system
will never be used effectively.
Needs of the System Design engineers must collaborate with owners to iden-
tify the needs of the system. For instance, treatment plants consisting of a simple
treatment process require instrumentation only for the plant flow rate, monitoring
of raw and filtered water turbidity, and monitoring of the residual chlorine content
of the finished water. Conversely, plants consisting of rather complex processes,
such as ozonation or a large number of filters and pumps, require an extensive in-
strumentation and control system to simplify plant operation and to make the plant
efficient and cost-effective.
The needs of a system may be classified into three categories: essential for plant
operation, useful for plant operations, and luxury item. The essential items are plant
flow rate control, chemical feed rate control, pump rate control, ozonation system
control, filter control, analysis and recording of water quality, control and record-
ing of residual chlorine, and detection of chlorine gas leakage (with an alarm). The
useful items are those that reduce the mental and physical activity of the operators
or those that can save operational costs. They include computerized data logging,
programmable control of the filter backwashing process, closed-loop automatic con-
trol of residual chlorine and the pH of the water being processed, and devices that
check the coagulant dosages such as streaming current detectors and automatic plant
controllers with advanced data-handling capability (the latter requires extensive cus-
tom-designed software). All these items are not vital to normal plant operation.
Finally, luxury items are defined as items that make plant operation easier without
providing compensatory savings in cost. The design engineer should discuss the
needs of the system with both plant operators and the owner prior to making any
decisions.
Task of the Instrumentation and Control System After the preceding items
have been considered, the type and purpose of the system must be defined. At a very
basic level the system can be used to serve operational, maintenance, safety, and
documentation purposes. The operational goals of the instrumentation and con-
324 DESIGN OF BASIC TREATMENT PROCESS UNITS
trol system should be to provide the best control quickly and without human errors.
The system can inform the operator of all current activities, where they are hap-
pening, and the duration. The system is also capable of making corrective or pre-
ventive actions. It can detect and warn the operator of changes in temperature, ex-
cessive vibrations, and abnormal levels of power consumption by individual
operating equipment and thereby prevent serious damage, costly repairs, and down-
time from occurring; these all relate to maintenance work. With respect to the safety
aspect of plant operation, a surveillance system is useful in meeting the security re-
quirements and the safety of the plant operator, who must operate the equipment
alone. Fatal accidents may also be minimized by installing systems that can detect
and warn the operator of chlorine gas, ozone, and oxygen leaks. Finally, documen-
tation is very important for both the operator and plant manager in identifying the
control modes and running time, in analyzing the history of repair work, and in tak-
ing inventory of available spare parts.
System Availability Three basic types of systems are currently available: digital,
analog, and control instruments. Digital systems have primarily two positions:
on-off, open-dosed, high-low, or alarm-normal. This type of signal is used to in-
dicate a change in status or to act as an alarm. The signals may be activated by a
position, limit, float, or pressure switch. Analog data systems cover a range of val-
ues from liquid or gas Ilow rate, liquid levels, liquid or gas pressure, to the turbid-
ity or pH of the water. The analog data may be sensed and transmitted in their ana-
log form or may be converted into digital form. The control instruments range from
a simple on-off switch to a programmable controller with a computer backup.
Many factors are involved in selecting the most appropriate type of system for
each treatment plant. The characteristics and the need of the treatment process train
may determine both the type of control and the complexity of the control system.
Factors such as the frequency of information, technical limitation, and local condi-
tions may also limit the type of system designed. Other important considerations
are the capital cost of the system and the associated operation and maintenance costs.
Finally, the case and frequency of calibration, the difficulty in maintenance, and the
availability of timely servicing must also be evaluated. A conventional monitoring
and control instrumentation system uses 4- to 20-mA analog signals (standard) in
transmitting measurement and control signals. Voltage or air pressure signals are
rarely used today.
Understanding and Commitment For any instrumentation and control system
to operate at its fullest potential, operators must have a complete understanding of
the system. The facilities must therefore have well-organized training courses that
not only familiarize operators with the system but also teach them the steps to be
taken when the system fails: After all, the system is not infallible. Thus backup con-
trol and monitoring methods must be an integral part of the plant design, and op-
erators must be familiar with these methods.
System maintenance must be conducted by either the operators or maintenance
personnel. However, if the system is highly complex, it is in the best interest of the
BASIC APPROACH 325
owner to sign a service contract with the equipment supplier or the service agency.
Since the operation and maintenance of a sophisticated instrumentation and control
system requires an adequate budget just to keep it in service, management must
clearly understand the benefits of a properly functioning system. If management is
committed to the system, it must also be committed to providing adequate resources
for the system: financial resources and the personnel required to design, operate,
and maintain it. Without both a clear understanding of the benet;ts and a thorough
commitment to the system by the operators and management, the instrumentation
and control system will not function at its full potential, and many future problems
may be anticipated.
Common Measurements The level, pressure, Ilow rate, turbidity, pH, and chlo-
rine residual of the process water are major process variables that are measured on
a continuous basis in all water treatment plants. Each variable is discussed briefly.
Level measurements are commonly made by float-operated transmitters, pressure
sensors, capacitance probes, sonic or ultrasonic units, or a bubble tube system. The
pressure within a system may be measured as absolute pressure, where vacuum' is
zero pressure; as gauge pressure, where ambient atmospheric pressure is designated
as zero; or as a differential pressure. The sensors may be U-tube manometers, bour-
don tube gauges, bellow gauges, pneumatic or electric force-balance pressure trans-
mitters, or strain gauge transmitters.
Flow measurement may be conducted by weirs and !lumes, such as the Parshall
flume, which is used to measure open-channel flow. However, the flow of liquid or
gas in pipelines should be measured by differential flow meters, acoustic (ultrasonic)
flow meters, magnetic flow meters, vortex shedding now meters, or propeller or tur-
bine flow meters. Yet Venturi tlow meters are the most commonly used type of flow
meter due to their high accuracy (0.5 to 2%), low headloss, and wide availability.
Engineers should note that selection of the proper size or capacity of a flow meter
is vital to obtaining accurate flow measurement. Most flow meters have a turndown
(maximum to minimum flow rate) of 5: 1 to 10: I. A common mistake of many in-
experienced design engineers is to design the meter size based on future maximum
flow rates; this practice does not allow for the measurement of low flows during the
plant startup period and other special occasions.
In the United States, turbidity is measured by nephelometric turbidimeters. The
principle of the turbidimeters is to detect and measure the scattered light beam or
the portion of a light beam that has not been absorbed by suspended particles after
an incident light has been passed through the process water. Thus the size, shape,
refractive indexes, absorption capacity, and viewing angles of the suspended malleI'
greatly inlluence the turbidity reading. The nephelometric turbidimeter may be used
in situations where the turbidity of the water is approximately 30 ntu. It will not be
very accurate in situations where high turbidities are expected (over 30 ntu) since
the meter will go blind at these levels and give false readings.
Turbidity has a fairly linear direct response: zero signal at zero turbidity. True
color does not register as turbidity but can cause negative errors. The surface scat-
ter type of turbidimeter may be used for water with high levels of turbidity but must
 326 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 327

be installed in a vibration-free environment. This type of meter is not appropriate

for measuring turbidities lower than I ntu because the response is nonlinear. Since
any type of turbidimeter will give a false high reading whenever the water sample
[Operator
(a' ·1 Valve

contains air bubbles, an air bubble trap should be installed in the sample line up-
stream of the meter.
[operator
·1 Motor ,I Valve

(bl
pH meters are generally composed of three electrodes: a pH sensing electrode,
a reference electrode, and an electrode that compensates for the temperature. These [operator , 1
ContrOl
p!nel
(oJ
,I Motor ,I Valve
electrodes are usually mounted in a chamber and installed in the sample line or sub-
merged in a tank or channel. The most recent models have a signal preamplifier Set point
Control
within the electrode assembly to minimize the electrical noise of the high- panel
impedance electrical circuits.
(d'
Finally, free and combined chlorine residuals may be measured by residual chlo-
rine analyzers. These units are commercially available from several manufacturers.
The principle used in measurement is amperometric titration through oxidation-
reduction titration procedures.

Basic Controls When selecting the type and control mode for a specific applica-
tion, the engineer must be familiar with the purpose and the hydraulic characteristics
of the particular treatment process. There are three basic considerations that engineers
should remember when selecting a control system: the nature of the required control,
the degree of disturbance that the process can tolerate, and the cost. There are four
types of control methods: manual, semiautomatic, automatic, and supervisory.
Semiautomatic control requires manual initiation of the automatic function. A typical
example would be manual startup of the automatic filter backwash sequence. The au-
tomatic control method uses sensors, limit switches, timers, analytical instruments,
controllers, and logic control devices (i.e., programmable controllers) to control the
process or equipment automatically. Supervisory control refers basically to a system
that is remotely monitored and adjusted; either in-plant or remote from the plant.
The two basic schemes for the control loops are the open loop and the closed
loop. Open-loop control does not possess a measurement device within the circuit,
and the treatment process is not contained in the loop (Figure 3.2.11-2, scheme cJ.
A prime example of this type of control is the !low-paced chlorine feed system be-
cause the chlorine feed rate is proportional to the flow rate of the processing water.
Chemical feeders commonly use open-loop control because of the ease and effec-
tiveness of the system. In contrast, the closed-loop control circuit includes a device
that measures the effects of the process. This information is then relayed to the con-
troller, which compares the result with a set point and performs the necessary ad-
justments. For example, a closed-loop chlorination system will be produced if a sig-
nal is relayed from the residual chlorine analyzer to the controller. Schemes d, e,
and f in Figure 3.2.11-2 are examples of closed-loop circuits. As mentioned previ-
ously, most water treatment processes usually have a long lag time and are there-
fore not well suited for closed-loop control systems.
Several types of control modes may be used, depending on the purpose and hy-
draulic characteristics of the process. The three basic modes are proportional control
and its variations, cascade control, and ratio control. Proportional control is defined
computer
(f)
Figure 3.2.11-2 Control systems: (a) manual control; (b) manual control made easier;
(c) remote manual control; (d) automatic feedback control; (e) automatic feedforward con-
trol; (J) automatic feed forward-feedback computer control. (Adapted from WPCF Manual
OM-6, 1978.)
as a system in which the controller continuously adjusts a controlling device, such
as a valve, to balance the process input with the process demand. An example of pro-
portional control is control of the filtration rate of filters with respect to the water
level in the clearwell. There are also several variations of proportional control: pro-
portional control with reset control or proportional control with derivative control.
Cascade control is widely used in the water treatment field, especially in pumping
control, where the output of one controller adjusts the set point of another controller.
In contrast, the ratio control scheme maintains a constant ratio between two variables;
the primary variable is normally uncontrolled. The ammonia feed rate of the chlo-
ramination process is controlled in proportion to the quantity of the residual chlorine.
Although engineers should select control systems that are simple and cost-
effective, one type of control system should never be used for chemical feeding:
on--off control. The impulse-driven control scheme is an on--off control that has been
employed widely over many years, but it tends to feed chemicals only part of the
time, and therefore a portion of the water remains untreated. This mode of chemi-
cal feeding is detrimental to coagulants such as alum or ferric chloride and to other
chemicals, such as chlorine and chloramines.
Computer-Based Monitoring and Control Computer-based monitoring and
control equipment uses digital signals. A large number of signals may be transmit-
 328 DESIGN OF BASIC TREATMENT PROCESS UNITS
ted by installing a remote terminal unit (RTU) at each major unit process. The RTU
receives analog or digital signals from nearby sensors, then sequences transmission
of these signals so that a single coaxial cable then transmits these measurement and
control signals to a central processing unit (CPU) located in the control center, where
they are displayed on a color cathode ray tube (CRT) and logged on a printer or
stored on a disk.
Control of treatment processes is accomplished by coding commands into the
video terminal keyboard. The microprocessor-based CPU communicates this infor-
mation to all the elements of the system: the RTUs and peripherals such as bulk
storage devices and printers. The CPU is capable of handling real-time commands'
that is, the computer system is capable of handling all multiuser requests, syste~
alarms, process control commands, and much more, almost on an instantaneous ba-
sis. The system operator is able to view the system data at the computer terminal.
The use of programmable controllers (PC) and programmable logic controllers
(PLC), which facilitate local process control, is another aspect of the computer-based
control approach. For example, the tilter backwash cycle may be controlled by a
PC, and the operator will be informed of the progress status of the backwash cycle.
Dlstnbuted control systems (DCSs) and supervisory control and data acquisition
(SCADA) systems have gained popularity in the lield of treatment (plant), and wa-
ter distribution systems and many water purveyors are now adopting their use,
Conventional Versus Computer-Based Systems Conventional monitoring and
control instrumentation is a proven and reliable method of operating a water treat~
ment plant. The computer-based monitoring and control system is also a proven
technology; however, its use has become widespread only since the late I 970s. Both
methods have advantages and disadvantages. These are listed below.
Compared to a computer-based system, a conventional monitoring and control
system has the following advantages: (I) the technology is familiar; thus operation
personnel do not have to be retrained, (2) troubleshooting is generally easier for op-
erators because the system is familiar, (3) much of the equipment is standardized
within the water treatment industry, and (4) the overall capital costs are generally
lower than for the computer-based system. However, it also has the following dis-
advantages: (I) the maintenance requirements are higher, due to the use of mechan-
ical relays and timers, (2) separate hard-wired connections arc required for each mea-
surement or control signal, (3) possible problems with electrical noise may exist if
the analog signal wire comes in close contact with a source of high voltage, and
(4) the incorporation of new instrumentation is more dillicult and costly during plant
expansion unless the initial plant design and construction are properly planned.
A computer-based system has the following advantages compared to a conven-
tional monitoring and controls system: (I) the system is more reliable and requires
less maintenance due to the use of solid-state components which are housed in sealed
and noise-proof enclosures; (2) the system is more tlexible due to the ability to mod-
ify the software to changing needs and the ability to incorporate new signals easily
when required; (3) the system is easily expanded to accommodate future process
units, in addition to accommodating water distribution data telemetcred from loca-
BASIC APPROACH 329
tions remote from the water treatment plant; (4) there is a reduction in the number
of hard-wired connections, thus there may be a cost savings in electrical construc-
tion due to the fact that the coaxial cables are capable of carrying much more in-
formation; (5) analog (conventional) or digital signals may be incorporated through
the use of RTUs; (6) the presentation of operational data is improved through the
use of a video terminal capable of color graphics; (7) the operating staff will po-
tentially respond more effectively to impending emergency situations because they
will be reacting to trends rather than critical alarms; (8) reports are generated auto-
matically at the end of each shift, the end of each day, and the end of each month,
as required by regulatory agencies; (9) the staff requirements may potentially be re-
duced due to the decrease in paperwork and instrumentation maintenance and the
ease with which the automatic functions are incorporated; and (10) plant manage-
ment and planning are improved because the presentation and integration of the plant
process data are vastly improved. However, computer-based systems also have the
following disadvantages: (I) should the computer system malfunction, the operator
could lose both monitoring and control capabilities throughout the plant;
(2) the overall capital costs may be somewhat higher than those for a conventional
control system; (3) the operators must be retrained to use the computer system;
(4) key plant functions must usually have conventional controls for backup opera-
tion; and (5) the computer systems are not standardized in the water treatment in-
dustry. Thus, expansion of hardware may be difficult and expensive, due to the need
to match components. If the new equipment is limited to those offered by the in-
stalling manufacturer, the engineer may not be able to obtain competitive bids.
Levels of Computer-Based Monitoring and Control The size of the computer
system required by the plant is not a function of the plant size. Rather, it is deter-
mined by the number of analog and digital signals that are to be fed into the com-
puter: the number of signals increases in logarithmic fashion with an increase in the
number of tasks. The computer-related tasks are data logging, report generation,
alarm, plant graphic displays, analog variable displays, and manual and automatic
plant control. There are six computer systems that are different combinations of
these tasks:
I. Report generation
2. Data acquisition and logging/report generation/alarm indication
3. Data acquisition and logging/report generation/alarm indication/plant graphic
display
4. Advanced display and data handling: data acquisition and logging/report gen-
eration/alarm indication/plant graphic displays/analog variable displays
5. Manual plant control and advanced data handling: data acquisition and log-
ging/report generation/alarm indication/plant graphic displays/analog variable
displays/manual plant control
6. Automatic plant control: data acquisition and logging/report generation/alarm in-
dication/plant graphic displays/analog variable displays/automatic plant control.
 BASIC APPROACH 331
330 DESIGN OF BASIC TREATMENT PROCESS UNITS

The tifth and sixth computer systems arc commonly known as DCSs or SCADA
systems. The first system is a very basic computer system that only generates rc-
ports. It is not hard-wired to the plant and the data must therefore be inputted man- -------------------------.
ually. Systems 2 through 6 are built around a mainframe or a powerful personal OPE~A1>ON$
IIU"-O",,O I
,,
computer and may be expanded to handle many additional tasks. ... oo~o HR""N"'l
... 100 P~I"TER
I
•
The hardware costs of the last five configurations are somewhat dependent on the 0"'''' HBUl.TIO... .
::~l~:~e~ .'::. o,. :
plant size and layout. Selection of the memory size is determined by the number of ,,
analog and digital signals that must be processed and stored by the system. The mode ,,
of control and transmission (analog or digital) also influences the size and cost of the ,
,,
computer system. The use of PCs, as opposed to standard relay logic control, increases ,,
the cost (PCs are normally considered as part of the system) but reduces the cost of
the electrical system because there are fewer relays, circuits, and wiring. If the plant
,,
size is large and the unit processes are spread alit, the use of RTUs is cost-effective
since single coaxial cables may be used to transmit both the analog and digital signals.
Other features that should be considered are the standard "vendor" report formats ver-
sus custom-editable report formats, vendor O&M software versus custom-editable
O&M software, and the inclusion of a separate engineering or laboratory terminal.
For budgetary purposes, the estimated costs (1989 costs) for the six alternative '"
,e' p~OOM"'''''''BU CONTROllER
systems for a 2-m 3/s (50~mgd) ordinary conventional process water treatment plant RIU' REMOU TERMINAL VNIT

at an ENR index of 5000 are as follows: .. IV _ ..... IN TfA"'''''''l UNIT

CPU _ CENTRAL PROCt$SI~Q UN,T ._------------------------
Figure 3.2.11 ~3 Schematic diagram of monitoring and control system (principal functions).
System I: $5000
System 2: $100,000
System 3: $125,000 3, Instrumentation and input-output summaries (IIOSs)
System 4: $150,000 4. Instrumentation specitication sheets (ISSs)
System 5: $600,000 5. Logic diagrams
System 6: $800,000-$1,000,000 6. Panc1layout drawings
7. Loop interconnection drawings (LIDs)
If a conventional monitoring and control system is adopted, the estimated cost for 8. Instrument installation details (liDs)
systems I through 4 would be approximately $220,000. Figure 3.2.11-3 is an ex-
ample of system 5. The P&ID is a functional schematic presentation of the treatment process and
the required instruments and controls. Its purpose is to illustrate the f~~ction of the
Design of the Instrumentation and Control (I&C) System The design of treatment plant without referring to the actual hardware; the peD speClhes the mon-
the instrumentation and control system must be conducted by I&e engineers. itoring and control loops,
However, the project engineer should maintain an open line of communication with The purpose of [he IIOS is to provide continuity between [he P&ID and IIOS.
the I&C engineers and perform a thorough check of the final design. Since the sys- These documents specify both the quantity and characteristics of the analog and dig-
tem must be operated and maintained by plant operators, it is beneficial to elicit in~ ital instrumentation, Similarly, the ISS provides detailed specification for each task
put from both the operators and the owner during the design phase. and panel instrument. Logic diagrams for the control panels and programmable logic
Design work usually produces the following documents in the sequence listed controls (PLCs) should be obtained by the instrumentation contractor. LIDs are used
below. The process flow sheets (PFS) prepared by the project engineer are the ba- primarily in projects where the instrumentation and control s~stem mus,t be~ II1te~­
sis for the P&ID diagrams. faced with existing control equipmcnt. A typical P&ID for ordlllary gravIty hlter IS
presented in Figure 3.2.11-4.
I. Process and instrumentation diagram (P&ID) During the late twentieth century, most water treatment plant control systems use
2. Process control diagram (peD) one of the three basic control systems: PLC, DeS, and a hybrid Des system. The
 332 DESIGN OF BASIC TREATMENT PROCESS UNITS
BASIC APPROACH 333

grated system. PLCs are excellent for performing sequential logic control and are
capable of providing reasonably good regulatory or closed-loop control. However,
they often do not provide sophisticated control or the precision required for more
demanding functions such as energy or process optimization. Ordinarily, PLC sys-
tems require a lower initial cost than that of the other two systems.
DCS (Distributed Control Systems) generally consist of process control modules
(peM), operator workstation/personal computers, a communication network, and
OU! software. The DCS package is a standard product offered by one DCS manu-
facturer; thus there is the advantage of a single source of responsibility for both the
initial purchase and ongoing long-term support services. The most reliable data ac-
quisition and control system currently available is DCS configured with redundant
PCMs and a redundant data highway. Features such as advanced control, compre-
:, :, hensive diagnostics, global database, fault tolerance, backup power supplies, and ad-
vanced alarming are standard features of a DCS.
DCS is a standard system and the manufacturer usually has extensive training
and support services with program upgrades to keep the system relatively current.
,,, However, since the DCS system does not have a large sales volume such as the

,, ,, ,
,
Intellution GUI for PLCs, the system may not be upgraded as frequently and there-
fore may lack the latest features. In general, the initial cost of DCS is higher, due
to the manufacturer's higher overhead and much higher R&D costs required to de-
velop a more sophisticated standard system.
A hybrid distributed system is a DCS that utilizes PLCs for the field controller.
This hybrid system offers the advantages and flexibility of the DCS and data high-
way as well as single-source responsibility. This type of system is usually supplied
when there is a preference for using existing PLCs or because of the owner's fa-
,, miliarity with programming and maintenance of PLCs. The hybrid system offers ex-
, .sa ,1"-- -- cellent sequential logic control and fair regulatory control or closed-loop control.
·~I~ ~ However, troubleshooting is not as easy as with DeS because the comprehensive
diagnostic function of the DCS will not extend to the PLC printed circuit card; the
"" I~ A
diagnostics will sound an alarm whenever a PLC fails. The cost of the hybrid sys-
G§ tem is usually between those of DCS and PLC systems.
The servers, workstation, and PLCs in the plant communicate with each other over
a shared local area network (LAN), a real-time network and an information network.
There are three basic LANs: Ethernet, token ring, and Fiber Distributed Data Interface
To F1aslllhber (FDDI). However, the Ethernet is usually the network of choice, due to its through-
Figure 3.2.11·4 Example of P&ID drawing. put and reliability, ease of maintenance, and feasibility. It is important to note that
the process information network (PIN) and facility information network (FIN) should
be kept separate because the facility information/data traffic is much more dynamic
PLC-based system consists of programmable logic controllers (PLCs) networked
and subject to large peaks, which can lead to a response slowdown. Also, FIN traf-
with personal computers through a communication network (data highway). The
fic may have high-bandwidth applications such as video and multimedia. Figure
computer workstatIOns are the operator's window into the system. These systems
3.2.11-5 is an illustration of the control and automation block diagram.
usually have graphical user interface (GUI) software packages (lntellution. Factory
Generally, the design engineer should follow the basic design philosophy of KISS-
Link) which run on the personal computer platforms. It is vital that a competent "keep it simple, stupid." There have been several case histories where the plant oper-
person purchase and assemble the PLCs, personal computer, communication hard-
ator was incapable of handling the situation when there was a sudden and complete
ware, GUI software, communication software, and system auxiliaries into an inte-
failure of the SCADA system during the night or weekend shift. Moreover, plant op-
 334 DESIGN OF BASIC TREATMENT PROCESS UNITS
BASIC APPROACH 335
PJ.AHt COHfRIJI. CeNtcR

AWWARF and Japan WWA, Instrumellfafiofl alld Computer llilegratioll of Water Utility
Operations, AWWARF/AWWA Report, American Water Works Association, Denver, CO,
1993.
1--;';;;;'.;-" ~oo·",-·-
Babcock, R. H., "Instrumentation and Control," special issue, Water and Wastewater
Engineering, 1968.
Computer Advances Research Committee, "Computers in the Water Industry, 1 AWWA,
81 :2:74 (February 1989).
Considine, D M., Process Illstrumelifs and Controls Handbook, 3rd ed., McGraw-Hili, New
York, 1985.
Dobbs, D. M., et aI., "Quick Payback Comes After PLC Installed," Water Ellg. Manage,
p. 25 (October 1988).
Eckman, D. P.. Industriallnstfllmentatioll, Wiley. New York, 1975.
Heywood, C. H., "SCADA Systems Today and Tomorrow," Water EI/g. Mal/age, p. 18 (July
1986).
Jentgen, L. A., and Wehmeyer, M. G.. "SCADA Trends and Integration Prospectives,"
1. AWWA, 86:7:44 (July 1993).
Liptak, B. G.. ed., Process Measllremelll Illstl'llment Engineering Handbook, rev. ed .. Chilton
Radnor, PA, 1982.
Stephen, J., Nix, and Lollins, A. G.. "Expert System in Water Treatment Plant Operation,"
I 1. AWWA, 83:2:43 (February 1991).
i WPCF, Process Ilisfrumematioll and Control Systems, Manual of Practice OM-6, Water
Pollution Control Federation, 1978.
I
I 3.2,12 Electrical System
Figure 3.2.11-5
I
Basic design block diagram of control and automation system. Purpose The modern water treatment plant cannot be operated without electri-
cal power; almost all equipment and buildings and all control and instrumentation
erators should not be allowed to become overl d . systems require electricity. Therefore, the plant electrical system must be reliable
terns because this will make th 'II Y epe~dent on automatic control sys- and safe. There are two basic standards and codes in the United States, the National
ern I prepared to act In an emergency situation.
Fire Protection Association (NFPA) Standard 70 and the National Electrical Code
Security System To assure the r b T . (NEC). Additionally, there are many local codes and regulations that govern the
lies, all plants must be prov'd d rhe Ja , Ityand safely of water treatment facili-
I e Wit a secunty syst A .. safety of the electrical system.
security system should includ' em. t a ITIlnUTIUm, the plant In general, most water treatment plants have a higher electrical load component
. . e a penmeter fence and electric l'
CirCUIt television (CCTY) w,'th f II' f .' en 'y gates, c1osed- and a smaller electrical load block. Higher load blocks consist of intake pumps, high
a u view 0 the mtak f T d .
and intrusion sensors on all bU'ld' e aCI Hy an mam entry gate, service pumps, and ozone generators. Flocculators, sludge collectors in sedimenta-
I mg gates.
tion tanks, various equipment and valves in the filter bank, chemical feed systems,
mixers, small pumps, lighting and control systems for processes, and electricity for
BIBLIOGRAPHY
bUildings are in the smaller load block. The electrical system must be reliable and
safe, as well as efficient, clean, user friendly, and expandable.
Amy, 1. P, and Cesarian A L "Corn A" .
85:4:129 (ApriI1993)'. ' " puter ppiJcatlOns TO the Water Industry," 1. AWWA,
Considerations Engineers must evaluate both the reliability of the power sup-
Andreas. M., Schulte, and Maim A P "Inte rati H .
for System Planning and Co~tr~I.:: J A~WA ng . y~rauhc Modeling and SCADA Systems
ply and of the power distribution at the plant when designing the electrical system
Arndt D J "M k' . . , 85.7.62 (July 1993). of a water treatment plant.
, '" a lUg the Right Moves in SCADA S I . "
(October 1985). e eetlOn, WaleI' Eng. Manage., p. 25
AWWA, Automau'on and Instrumel1fatioll Manual M 2 . Reliability of Power Supply Power distribution systems from power stations are
Denver, CO, 1982. , - ,Amen can Water Works Association, generally very reliable. The utility transmission system (voltage higher than 34.S
kY) is also usually reliable. Historically, however, the power supply from utility dis-
 336 DESIGN OF BASIC TREATMENT PROCESS UNITS BAStC APPROACH 337

tribution systems (15 to 25 kV) is less reliable, therefore, the engineer should first SIN C L [ liN £; 0 III G R .I. ~ SCHE~ATIC DIAGRAM

evaluate the historical data of the local power supply system to the plant. Frequency "l-" =:~'~~!f~~::.:t<S".1 ort_
m
M

~::s~.[~\.::e::"'.,~:::!. I """"II _ ~ ;:~~I:'" ,"01(" .. 1M 0",,",".'.

of outages, a range of outage length, and the average length of outages for last past
5 years must be investigated. @. ~w: ~!!c:.~:~~.:;::!~~~
u... ·u . no .-.,:
100
~II ~.
~'n,
.. ••
OOI'VI(

l!DI ""."......,0<
.

The regulatory agency's guidelines for water treatment plant systems requires ei~ 0... ~"""u""" I'" ((J 'fL\%nlL~~lI.' '''"''''''U'
....
U "'In.

Lgj ".t.c"'''n ::--0- .I"INI·rO·IUI ..x"..: ,CHr

ther two sources of power supply from separate substations or a single supply source ''''1II.l UIlt>o ,~"n ,,!H'
t
~ .><or
and an emergency power generator system. Design engineers should first contact the .", -""'T oro! L'" 10"' •

local power company and investigate the standard service. The engineer should then _W.'" (l.ono LN' ""'H
<j> no .. 1IP( UAoJ LI.n ""<;II.
inquire about extra charges for two-line service from separate substations to the pro- (} <LOs..: ...... oe LliI'

.1_
nou
.... um. 100reIL -+- C"D'1"" 01 <OtGA:l"'I·IIO' COIHtC110

.1)1 "["(II.' 'o\.[ ..... UI '0"

110"1 , _
"00;\1.'\ -t- I_e,.,.. 01 tOtGA:l001.''''''''''
posed plant site. If reliable service from the power company is not available or it is
•
.. _·.u.....
.....0£0 (AI( C"'CU'
O'Ol.'lStllO.to.
hI:
........ "'rM; ...... ....
"";;'
.OCI.UI
c,o~
lOt 'M;INIt U",IL
011 IIIJrC ,""!(.III
••
"' ..... 1•

" _ , ,..... IIC<"

too expensive, installation of an on~site emergency power generator should be con- IIw-A) n~~~:r~~·:ar~lrn T
wOCl.Ul ... ,"I:ISIM lO""'"
OO1._ .....IoOIQ'OIIWOt "",-"II-.n"",oOO;III
o 11_1.... 11 [I".OIIL(I'
""'I ""too.
sidered. In this situation, the power company may provide the treatment plant with
...
y Ti"'tlUlUOI: .ctU"IO

rate incentives for operating this generator during the power company's peak power
-<>1)-0-
C\OIIle 011 .. _
1I1f1t:IUI.I!t
~CIIIUU
"'"'''lOOt
ntIC.., t'l
~ 1)""0111 lIIt"tI\, Nt 'S 00110 ~OII",OO;II"""""-t
• UIIU1 ,,"LV ,_COo:>or

...
to • \ItIIO'u IUU~O;'\U GO,.un
demand as a means of shedding power load. However, on~site emergency power gen-
''''LOOOII_1I0.'0:(
4O-IH:> I0Il_ GfI _ 001."1)[ ""'LO "t' 'lO.' ... "' ....... 'I .. U< ...
nDI ... 0II1l0o _lAlI ,no 011 JltIOl'£ I0.Il110 I+:"":!
..!. ~ uc..,_ ",,'m .. n•• ,."tM ....... ".. "c. . <",UCI U K>l.OU.I'I.C.II11 011 \eI'
erators have a high capital cost and require regular exercise and maintenance efforts. ®
001._ 011 "GO IOCIIU$[
110'11110 ...... 11. ;;'-; _"'ll'
OI'[~
_HUG.G'II
M>eII"'"
Moreover, the design of the fuel storage must comply with all local code requirements, ~ :::"ir::\,:::,f'~DnK~":f:.~'. ~ _.l\'nol€\l,...,...."...
wqlOl! OPt~"lO .... 1
la-' ~W"~H (Of<UC' OC'illOPIIt'"
_lO,,", OPOI
(}r1!..1QO·'OIl}l'''\ .. ....,,_".
,,·s """ID t_T CCIITOCT

_. • -... .., ,....

~ 10110 _ '"-"'£II 01 IIO/IC ..... UKO N$HII.IUOII OU... • ..... 'LU OO'UI ""I"O<&'l
Reliability of Power Distribution at the Plant There are several design methods t ~., ,.""'_..
,~ nOlI ~.
,.,
to ensure reliability of the power supply to individual equipment in a treatment plant.
.
CI
~ .01'" _ _ " 01 C"S AI ... 110
~
I.
..' '·:=.":
"-.VI...
..!',;.~~~.:.;f::L.lac.L.
"",.O ... ,
~.

~
<V ~ mc~":::::.::::l"'-ac. ~. a.OU .oK
Higher reliability generally requires higher capital expenditure. A typical water treat- 1,","~1101! stUtTOII
'I .... 01 '.0'0
'OI"IL
"'
ment plant has four load blocks: two large loads and two small loads. If an ozona- 1'"1£1-'0111_ "'_ IIU"",·10·<"""
_~1,",. COI< .. " IOI.OII"UICI....... Uo\
""·"":Ue.I
tion process is included in the process train, it should be designated to the large load
blocks. Large loads are typically supplied at a primary voltage of 4160 V, with 480
V of secondary supply through transformers for the smaller loads. 6
-F!-..... u .. cnOll. ....... , IIOlto

The power distribution within a treatment plant is always shown in a single-line e ...........u.
(one-line) diagram using standardized symbols. Figure 3.2.12-1 depicts the standard e '0\.'11[11'

"l."i.t.
__ .... Of"E.N IOC ...... a.~."

symbols. Figure 3.2.12-2 is a single-line block diagram of a plant power system. " "X'i.o. 1OO"'lLT (l.01t0 01'. 'oe DIU . . . ~

Examples of low-voltage (secondary loads) power supply line diagrams are shown "
tp>
.000'IoO!'LO ''''01
SURe.( .... 1.0)<'_ ....u.'" ~.o.
in Figure 3.2.12-3. There are three types of power distribution design schemes: ra-
dial electric, primary selective, and secondary selective systems. A radial system is Figure 3.2.12-1 Major electrical symbols.
the simplest and least expensive power distribution system. However, failure at any
point in the system interrupts all power downstream of that point. The primary se- tribution and the required voltage, and (5) selection of the loc~tion of the main trans-
lective system is capable of isolating a fault and switching to a different path to pro- former within the plant site-this is a site-specific issue and IS not addressed below.
vide uninterrupted power; this system has redundancy for the upstream part of the
system, where failure causes the greatest disruption of power. The secondary selec- Power Supply Conditions The IIrst task is to investigate the availability of the
tive system adds redundancy to the downstream portion of the distribution system, proper type of power within a reasonable cost. The capacity and locatIOn nearby ?t
providing an alternative path for power to secondary loads. Another type of reliable .
substatIons, . ot
locatIOn ' th
e mam' powel. I·me 0 I' nearest power pole relative to the
system is the double-ended or main-tie-main arrangement. This system starts with plant service main, frequency and duration of power ou~age over!he pas.~ seve~al
two sources of power and two main breakers, each of which supply half the load, years, and power charge structure of the local power supplIer (especJally dUl tng pedk
thus half of the system is not affected by the failure of the other system. hours) should be studied during the early stage 01 the prolect. Power supply condt:
tions greatly affect both construction and plant operational costs, and .lI1t~~ence the
Design Criteria When designing a new water treatment plant, the development selection of equipment drive units and the size of the standby power unit (It needed).
and evaluation of power are divided into five subissues: (I) investigation of power
supply conditions at the plant site, (2) estimation of power load or power require- Load The load or power requirements are dictated by plant size, type and num-
ment, (3) evaluation of standby power requirement, (4) determination of power dis- ber of unit processes, whelher the treated water is distributed by gravity or by hIgh
 338 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 339

.
TABLE 3.2.12-1 Estimated Power ReqUirement 0 r Nor th Bay Regional Water
Treatment Plant (30 to 60 mgd Ultimate)
Total Connected
Load
Number
of Units hp kW
Location/Item
Treated water pump station
I 15
Lighting panel
Exhaust fan I 1.5
Backwash pumps 2 400
Treated water pumps 9 2700 (5400 ult)
Filters
10
Lighting panel
Exhaust fan I 1.5
Sump pumps 4 12
Actuated val ves 16 36
Operations building
2 40
Lighting panels
30
HVAC panel
20
Instrument panel
~·A "'__ ..
HN·I.!W PAC I 10
40
U
.
~
.
~
-"". ...",~
Power panels 2
Chlorine monorail 2
Figure 3.2.12·2 SingJe·line power system block diagram. ATS. automatic transfer switch;
50
TS. transfer switch; 1M, inphase monitor~ C&S, control and sensing system; A, ammeter; AS, Chlorine scrubber I
ammeter switch; VS, voltmeter switch; SIG. ground fault sensor. Roll-up doors 3 3
Spare chemicals mixer I I
Activated carbon mixers 2 50
service pumps, and the hydraulic grade line across the treatment plant. The electri- 2 6
Dust collectors
cal load is determined by the horsepower of the individual equipment components. 20 16
Metering pumps
Other miscellaneous electrical powered items, such as instrumentation and control, Exhaust fans 4 6
valve operators, chemical feeders, and lighting, are based on the actual number and Chemical transfer pumps 2 4
scale of each unit. Table 3.2.12-1 presents an example of estimated power require- Actuated valves 7 10
Miscellaneous
,,. .•...•.. •.. ..... -...... _- - - - - - .-. - - •.. -.... ,. - -
, _
Wash-water recovery pumps 3 40
Sludge decant recovery pumps 2 4
Sludge pumps 2 1.5
Bar screens 2 2
Fuel oil pump I I
Security gates 2 2
Sample pumps 6 2
Boiler pump I I
Primary A/C unit 1.5
Condenser 0.5

! Floc/sed basins
1,/ Flash mix pump 10
1II Flocculators 36 54
Chain-and-tlight collectors 3 6
Sludge cross collectors 3 3
Figure 3.2.12-3 480-V single-line diagrams. Actuated val ves 6 4
 340 DESIGN OF BASIC TREATMENT PROCESS UNITS BASIC APPROACH 341

TABLE 3.2.12-1 (ColI/illlled) TABLE 3.2.12·2 Approximate Relation Between Plant Size and Power Requirement
(kVA)
Total Connected
Number Load Conventional Treatment Process
Location/Item of Units Plant Size Without Ozonation With Ozonation
hp kW
Generator/blower building 5 to \0 mgd 2100 2400
Lighting panel (0.22 to 0.44 m'/,)
Blowers
30
2 150 \0 to 25 mgd 3300 3600
Air compressors 2 20 (0.44 to 1.1 m'/,)
Exhaust fan 1.5 25 to 50 mgd 3400 3800
Ozonation building (1.1 to 2.2 m'/,)
Lighting panel 50 to 100 mgd 5600 6500
20
Power panel (2.2 to 4.4 m'/s)
30
Ozone system (750 Ib/day)
400 kVA _ V3 x volts x ampers
Raw water supply Note: - 1000
NBA pumps 3 175
Totals Voltage: The larger the load, the highter the voltage requirements, so that large pumps and ozone gen-
3787 (6487 ult.) 635 erators need high-voltage service. According to Honeycutt et aI., the approximate voltage requirements
for certain sizes of single motors, including starter and conductors, are as follows:
For motors up to 500 hp: 480 V
For motors 500 to 5000 hp: 5 kV (2400 or 4160 V)
ments for a30-mgd (U2-m'/s) plant. This plant requires 3800-kVA process mo- For motors larger than 5000 hp: 15 kV (12,470, 13,200 or 13,800 V)
tors. ozonatlOn. and 500-kVA lighting and miscellaneous loads. Table 3.2.12-2 il- For an entire plant, the approximate lotal horsepower values for different voltages are:
Up to 2500 hp: 480 V
lustrates approximate relation between plant size and estimated power requirements.
From 2500 to 10,000 hp: 5 kV
Above 10,000 hp: 15 kV
Standby P0w.er As mentioned previously. water treatment plants should be op-
erated In a reliable manner; therefore, standby power is required in case of power
outage unless the power to the plant is supplied by two different substations. If the filters are washed by direct pump wash, the wash pumps will require standby
Although this is a site-specific issue. a general list of essential equipment that should power if potential power outage is expected to last more than one day.
be connected to the standby power system is listed below. Economic evaluation of the various peak shaving alternatives depends on the rate
structure of the utility. There are essentially four cases of peak shaving operation:
• Instrumentation and control system (I) load shedding can be accomplished by reducing load using either a computer or
• Major chemical feeders (coagulants and disinfectant) microprocessor-based controller; (2) utility peak demand reduction involves the
• Electrically operated major valves transfer of a portion of the load from the utility source to the plan!"s engine gener-
Flash mixer ator sets during times when the demand for electricity is at its highest; (3) when
paralleling with the utility, the generators are directly connected to the normal dis-
Lighting, heating, ventilating, and air-conditioning systems in the operations
building tribution system-peaking power is provided directly into the system and the amount
of power furnished is determined by the engine governor controls; (4) a prime power
Safety light for all areas with process units and the security system cogeneration system designates the on-site generation system as the prime source
50% of mechanical flocculators of power and only certain dedicated loads are connected to the on-site generators:
25 to 50% of ozone generators The utility may be used as a standby source through the use of an automatic trans-
25 to 50% of intake and high-service pumps fer switch; however, this alternative is more appropriate for wastewater treatment
Minimum of one utility pump but all sump pumps plants since methane gas produced by the sludge digestion process can drive the on-
site generators.
One pump for filling the backwash tank
All water sampling pumps and water quality analyzers Voltage-Drop Issue The ~urrent required to start a motor with full voltage is as
• Refrigerators, incubators, and other essential analyzers in the laboratory high as six times the full-load running current; therefore, a protection measure should
 342 DESIGN OF BASIC TREATMENT PROCESS UNITS

~e provided. However, a 15% drop is generally considered to be tolerable The th

.aslc reduced-voltage starters for motor starting control measures are' (I) ree
tl ansformers , (2) wye-<I eta
I
starters, and (3) solid-state starters Autotrans' auto-
can be set at 80% 65% d ,. .ormers
• 0, 0, an 50% of full voltage for starting (the 65% se!tin i
com~'0nlY used). Wye-<lelta starters arc able to control starting current to appr; : CHAPTER 4
male. y two ~lInes the running current, as compared to six times. Solid-state s XI-
provIde tlexlble adjustment of the starting condition; therefore, there is no p t~~ers Subordinate Plant Facilities
wIth tranSitIOn current. ro em

Miscellaneous Issues Correction of the power factor improves the

sump!" f . h power Con-
. IOn 0 motOi. s; t erefore, electrical engineers should provide power factor c _
pacltors. The , maxImum power factor' . is lOb, ut some eXlstlllg
. . plants have a rna a. 4.1 CHEMICAL FEED SYSTEM
mum, ~ower factor of 0.9 or less. Selection of the motor is another important i XI-
Ind~c;lOn motors are generally more cost-effective than synchronous motors an~sue. Purpose Most water treatment plants use a variety of chemical engineering
use. or. smaller to midsized pumps. Synchronous motors are more favorab1e ;~e processes to produce large quantities of safe drinking water within a short period of
use III hIgh-horsepower and s10w-speed applications· however tl
about 0 8 Th " "
,!
1e power .actor IS time. Therefore, the design of the chemical feed system, including handling, stor-
• '. > us corre~tlOn ot the power factor is essential. Installation of v01ts am- ing, feeding, and mixing, requires special attention. A properly designed chemical
pere reactIve (VARs) IS a common solution for this problem. feed system can improve treatment efficiency and thereby reduce operation and
maintenance costs. Three basic forms of chemical are used in water treatment: dry,
liquid, and gas. It therefore follows that the basic types of chemical feeders are dry,
BIBLIOGRAPHY liquid, and gas feeders. Liquid volumetric feeders are preferred because of their
compact size, accuracy, and convenience in handling. Generally, chemical feeders
Honeycutt, D. E., and Jones, J., "Electrical Systems" Chapter 21 in Wi: t 'T' must have a feeding capacity that is broad enough to accommodate the variations
Design 3rd ed AWWA d ASCE' .' a er lrealment Plaflt
, ., an . McGraw-Hili, New York, 1997. in plant flow rate and chemical dosages,
Square D Company, Motor COlltml Fundamentals, Square D, Oklahoma 1974.

Water Treatment Chemicals Water treatment chemical should be of a high-

grade food additive class or those approved by the EPA for the treatment of drink-
ing water. The quality standards for the major water treatment chemicals are estab-
lished by the American Water Works Association (AWWA). The chemicals may be
classified according to their application in water treatment. The chemical names,
commercially available form, and approximate unit cost (bulk purchase: 1998) are
listed below. Some important characteristics of the chemicals may be found in Table
4.1-1 and Appendix 13.

COAGULANTS AND FLOCCULANTS

Aluminum sulfate (alum) Dry or 49% liquid $250 per ton in dry form
Ferric chloride 35 to 45% liquid $250 to $350 per ton in dry form
Ferric sulfate Dry form $350 per ton
Cationic polymers Liquid form $1500 per ton
Anionic polymers Primarily dry form $5000 per ton
Nonionic polymers Primarily dry form $5000 per ton
Lime (for softening) Dry form $70 per ton
Soda ash Dry form $400 per ton
343
 344 SUBORDINATE PLANT FACILITIES
CHEMICAL FEED SYSTEM 345

DISINFECTANTS
considerations Prior to designing the chemical feed system, the engineer should
Chlorine Liquid form be familiar with the raw water quality, the desired finished water quality, and the
$250 per ton
Chloramines (CI, + NH,) treatability of the water. These factors determine the types of chemical that can be
$300 per ton for NH, alone
Chlorine dioxide - employed and their dosages. The design engineer may then address the following
$4500 per ton for dry
(CI, + NaCIO,) issues: the characteristics of each chemical, local conditions, mode of delivery, avail-
NaCIO, alone
Ozone (on-site gas ability of the chemicals, cost, local and federal safety codes, and procedures for han-
$2000 per ton for cost of
generation) dling the storage.
generation
Liquid oxygen (LOX) Chemical characteristics that should be analyzed are the stability (storage life),
$60 to $80 per ton
Sodium hypochlorite Liquid form corrosivity, crystallization temperature (freezing), safety, potential hazard to health,
$500 to $700 per ton
CI, equivalent hygroscopicity, and flammable or explosive nature. Similar to the chemical charac-
teristics, local conditions affect the cost and the form in which the chemicals are
PH CONTROL transported and stored. The local conditions may also affect the reliability of the
Lime [Ca(OHhl Dry form chemical supply. Additional design considerations include the ambient temperature,
$100 per ton
Caustic soda 50% liquid humidity, amount of snowfall, potential for flooding, distance from major freeways
$600 per ton as NaOH
Soda ash Dry form and rail lines, and the proximity of the plant site to residential and commercial
$400 per ton
Sulfuric acid 98% liquid areas.
$200 per ton
The method of chemical transportation is an important issue because it deter-
TASTE AND ODOR CONTROL mines the design of the receiving facility. The treatment chemicals may be deliv-
Powdered activated ered in the form of bags via truck or in bulk form by semitrailer, rail, or barge. If
Dry form $1300 per ton
carbon (PAC) the chemicals are delivered by trucks, the access road must be wide, with an ade-
Granular activated quate turning radius. The engineer must also establish the proper grade and the type
Dry form $2100 per ton
carbon (GAC) of pavement (access road), in addition to researching the weight limits of the pub-
KMnO, lic road and bridges.
Dry form $1800 per ton or $700 per cubic yard
Ozone (on-site gas The availability of the water treatment chemical is a very important considera-
$2000 per ton tion for plants located in remote areas and developing countries. For example, poly-
generation)
mers and liquid chemicals, such as liquid alum and liquid caustic soda, are seldom
CORROSION CONTROL
available in these regions. Engineers are therefore advised to research the local avail-
ability of particular chemicals prior to designing the system.
Sodium metaphosphate Dry form $1800 per ton An issue related to chemical availability is cost. Depending on the availability,
Zinc orthophosphate Liquid form $1500 per ton the difference in cost between two chemicals may be quite significant. For exam-
Caustic soda 50% liquid $600 per ton as NaOH ple, caustic soda is generally 10 times more expensive than lime. Chemical cost is
also affected by the manner in which it is purchased; bulk versus small amounts.
FLUORIDATION
Engineers are encouraged to study market trends in order to predict possible short-
Hydroflurosilicic acid 23% liquid ages in chemical supply, which increases their cost substantially.
$850 per ton as F
Sodium fluoride Dry With regard to the handling and storage of water treatment chemicals, the design
$400 per ton
Sodium silicofluoride Dry engineer must be thoroughly familiar with the local, state, and federal safety regu-
$1300 per ton as F
lations. In compliance with OSHA regulation, plants should have emergency show-
SOFTENING ers, eyewashes, exhaust fans, and special clothing and gear for the safety of the
Lime (CaO) Dry workers. Additionally, chemical storage tanks must be located within a bermed
$60 per ton area-to contain chemical spills-and acid holding tanks must be isolated from al-
Caustic soda 50% liquid $600 per ton kali chemical storage tanks by a separation berm. OSHA also demands that oxidants
Carbon dioxide Liquid $70 per ton such as chlorine and potassium permanganate be stored in a separate room from or-
Sodium silicate Viscous liquid, S.G. = 1.39 $150 per ton ganic chemicals such as polymers and activated carbon.
ALGAE CONTROL The handling and storage of treatment chemicals must be based on the nature of
Copper sulfate the chemicals. Hygroscopic chemicals such as ferric chloride, ferric sulfate, potas-
Dry $1200 per ton sium permanganate, soda ash, calcium hypochlorite, and alum are generally shipped

I
 346 SUBORDINATE PLANT FACILITIES
and stored in dry form and therefore require special packaging and storage to pro~
tect them from moisture. The powdered form of all substances is potentially haz-
ardous. If the dust concentration is allowed to reach a certain level, one spark will
cause an explosIOn. Ozone, oxygen, calcium oxide (quick lime), sodium chlorite
and act~vated carbon are all considered to be fire hazards; some are also explosiv~
and tOXIC. They must therefore be handled and stored in strict accordance with the
safety codes and rules provIded by the manufacturer.
There .are two other miscellaneous but important design concerns: clogging of
the feed lIne and the. mode of chemical feeding. Clogging of the chemical feed pipe
IS a common, yet seflOus problem. Alkali chemicals that include hydroxide and metal
coagulants often produce h~avy scale in the feed Ii nes and diffusers. Many treat_
ment plants dIlute the chemIcals downstream of the feeder to obtain better diffusion
of t?e chemical within the process water, to minimize the feed lag time, and to avoid
havmg the slurry settle in the feed line. However, extreme caution must be exer~
cised when diluting metal coagulants such as alum and ferric chloride, due to their
pH and solubility relationship (see Figure 3.2.3-1). In the case of alum and ferric
chloride, solutions up to I and 5% (using ordinary filtered water), respectively, may
be the maximum safe dilution. Alkaline chemicals such as lime, caustic soda, and
aqua ammonia must not be diluted by carrying water unless the total hardness of
the carrying water is less than 20 to 30 mglL; otherwise, calcium carbonate scales
will form rapidly, due to the softening reaction.
Regardless of the metered solution concentration of the chemicals the diffuser
pipes can still potentially become clogged as the result of instant me;al hydroxide
formatIOn ~r softening reaction th~t occurs when these chemicals hit the processing
water, provIded that the hardness IS moderate or high. For this reason, they should
be deSIgned for easy cleaning, and if the diffusers are not easily accessible, an al-
ternatIve dIffuser should be provided as a backup. The ideal design is to locate the
feeders as close as possible to the application points and to apply these chemicals
from above the water surface.
It is worth restating the importance of continuous chemical feed versus the on-off
mode. The puls~tion pump is not the primary choice as a feeder unless a pulsation
damper can be Installed to smooth out the discharge flow: noncontinuous chemical
feedi.ng will overdose one particular part of the process water while leaving the rest
?evOld ?f chemicals. Certain chemicals, such as alum, undergo instantaneous and
Irreverslbl~ reactions. From the view of the plant operators, chemical feed pipes
should be mst~lIed at every possible location. The convenience of changing chem-
Ical feed locatIOns will offset the initial cost of installation. If there is a possibility
of future pla.nt expansion, the design should include the space and capacity required
by the addItIOnal feeders, feed lines, and appurtenances.
Overview of the Chemical Feed Systems Three types of chemical feed sys-
tems are used in water treatment: (I) liquid chemical, (2) dry chemical, and (3) gas
feed. Of the three types, the liquid chemical feed system is preferred because it is
clean and compact, is not labor intensive to operate, has easy automatic controls,
and does not have problems with chemical dust or fumes.
CHEMICAL FEED SYSTEM 347
Liquid Chemical Feed System The simplest, proven, and most cost-effective
liquid chemical feed system has a pump capacity that is at least 10% larger than the
maximum feed rate (alternative I, Figure 4.1-1). This system contains a rotameter,
flow control valves, shutoff valve(s), and a recycling pipeline; the recycling line can
return the excess flow either to the tank or to the pump suction line. This type of
system is not suggested for use with chemicals that coat the interior of the rotame-
ter (glass tube). Chemicals classified under this category are caustic soda, fernc
chloride, potassium permanganate, and certain types of polymer. It is recommended
that rotameters not be installed in locations exposed to sunlight, since algae will
grow within the glass tube.
One major drawback of this system is that there is no simple way to pace the
chemical feed rate to changes in the plant flow rate. However, this problem can be
avoided by using a variation of alternative I-installing a rotodip type of feeder (al-
ternative 2 of Figure 4.1-1). The rotodip feeder receives a 4- to 20-mA signal from
the plant flow meter and is therefore capable of pacing itself to the plant flow rate.
The feeder has a set of revolving cups and contains a unique feature, a wide turn~
down ratio. A ratio of 1500: I may be achieved by changing the revolving speed,
the submergence of the cups, and the speed reduction gears. It requires minimal
maintenance and can easily last more than 20 years under normal operating condi-
tions.
The majority of modern water treatment plants use metering pumps. Most such
pumps have alSO: I feed range, including adjustment of the pump stroke length
(15: I) and motor speed variation (10: I). The advantages of a metering pump are
its simple pacing of the chemical feed rate to the plant flow rate, a very high feed-
ing accuracy (I %), a discharge pressure up to ISO psi (1035 kPa) or more, and com-
pact size. Refer to Figure 4.1-2 for an illustration of this second type of feed
system.
A third type of chemical feed system combines a magnetic flow meter with a
needle valve (for flow control) and a pressurized storage tank (Figure 4.1-3). The
pressure level within the air padded tank depends on the back pressure from the pro-
cessing waterline. For example. if a chemical is fed into an open channel, an air
padding pressure of 10 psi (70 kPa) will be sufficient for the storage tank. This sys-
tem is simple and has proved to be effective, but it has two major disadvantages: a
limited turn-down ratio with an overall practical limit of 15: I, and a potential for
clogging of the flow control valve and meters. A variation of this system uses a re-
cycling pump rather than the pressurized storage tank, in much the same manner as
the first feed system containing the rotameter. Use of a recycling pump is cost-
effective since there is no need to specify a thick shell wall tank, due to the lack of
internal pressure.
Since exposure to some water treatment chemicals is hazardous to human health,
the plant design must conform to the local safety codes and OSHA regulations. All
storage and handling facilities should therefore have eyewashes, showers, clear
warning signs, berms for containments, and adequate lighting and ventilation. All
storage tanks must have tags labeling their contents, an access hole, fill line, drain,
overflow, discharge valve, vent, and content indicators, such as a load cell or a dif~
 348 SUBORDINATE PLANT FACILITIES CHEMICAL FEED SYSTEM 349

l'Ions:
Alternative 1 I. JANMS. ~a.q'ZOkT.o.l OR ~ER"CAt
1 now FACING· yAl\JA8lE SPEED MOtOll
~. MAON(TIC flOW IoIEfEII rOA
IoIONIYOflINQ{OPTION"'~)
oos ... o~ SlnlNO· PUMP 'TROKE 5. PROORUStvE CAVITY tyPE I't.I"'P
). vOlU,,"E 01 fNClOSURE ... AO\.IND ...... y BE USED IF R... NOE OF fEEO
STORAGE T4>II( SlW.l BE 5lIGNTl Y CONOITIONS CAN IE ...n U,..,rvW"'HR
GRf",l[R THAI< lIQUID VOLUI<IE OF '~N~ IHUStllNG 011 CMlUTIONj
To Application
Point
VENT
Recycle
Fill Line
DIFfUSER
Rotameter (typical)
(1:10 Feed Range for Each:
Storage
Tank
ORAl'"! Y
Q '·20m'

VE'"!TlIRI I"VPiE
lIT"""Y F~OWME1ER
W ... TER

Drain Figure 4.1-2 Liquid chemical feed system (metcling pump system).

terials used in constructing the metering unit, valves, piping, and storage tank. Other
design considerations include the ease of operation, reliability, safety, cost, and the
Alternative 2 method by which the chemicals are fed: It is essential to use continuous feeding and
not intermittent feeding of chemicals.
Fill Line
Dry Chemical Feed System The dry chemical feed system is a well-established
Rotodip Feeder and reliable method of adding dry or solid chemicals. It has been used successfully
O.F.
(1:1000 Feed Range) in industrial and water treatment applications for many years. This type of system
may easily be automated and problems associated with dust control can be mini-
Storage mized with the installation of an appropriate type of control system. Yet the dry
Tank
HOTES:

1 'ANK. HORIZONT"'L OR VERTICAL

2. VOLUME OF EHCLOSURE SI1,t,ll BE
SIGNAL

r To Application SLIGHTlY GREATER TH,t,H

TA"'NSMlnER
LIQUID VOLUME OF "'NK
Point •.!!"",m"-'_ _~Ol\?
_.!; VENTURI TYPE
CONTROLLER ,. FLOW METER
Drain

Figure 4.1-1 Two basic liquid chemical feed systems.

terential pressure cell, or sonic-level indicators. Optionally, the tanks may contain
a mixing device. The use of glass-sided gauges is not recommended, due to the po-
LIOUIO
tential for breakage and because they become fogged by the chemicals. Use of a CNEMICAL STOR,t,OE

bubbler tube system should be avoided, as it may introduce oxygen, carbon diox- j
ide, or bacteria into the stored solutions and therefore degrade certain chemicals,
lO,t,O CELL
such as caustic soda and polymers.
For any of these systems, the most important design consideration is the selec- Figure 4_1-3 Flow pacing solution feed system (closed-loop control with magnetic flow
tion of materials that are most compatible with the chemicals to be handled-ma- meter).
 CHEMICAL FEED SYSTEM 351
350 SUBORDINATE PLANT FACILITIES

chemical feed system is more labor intensive (operation and maintenance) than the
liquid chemical feed system.
The dry feed system is usually not preferred for most plant operations. However
certain chemicals, such as lime, potassium permanganate, PAC, and most anioni~
and nonionic polymers, are available only in dry form. Moreover, a few plants can
only utilize chemicals that are in dry or solid form, due to problems associated with
the liquid form: freezing, high delivery cost, degradation of the liquid form of the
chemicals when stored for long periods of time, or a limited selection (polymers)
compared to the dry form.
The design considerations for the dry chemical feed system focus primarily on
the storage silo and dissolving tank. The major concerns of the engineer are explo-
SIon-proof handling of dusty materials and the health and safety of plant operators.
An appropriate type of dust collector unit and an exhaust fan must be installed in
the feed rooms. Engineers should realize that all bag-type dust collectors require the
dust to be completely dry. If this cannot be accomplished, a water scrubber type of
collector should be specified.
For any type of solid material, especially the powdered form, the angle of the
hopper bottom is an important issue. The bottom of the hopper should have an an-
gIe no greater than 300 from the vertical and should be fitted with ordinary vibra-
tors. However, if a bin activator, manufactured by Vibra Screw, Inc., is installed in
the bottom, the hopper angle may be more shallow, due to the effective control in
bridging and flashing the dry material. Since the surface of the dry material stored
in the silo is generally not level, it is very difficult to select a reliable inventory in-
dICator. Currently, the best and most reliable method of keeping inventory is the in-
stallation of load cells to weigh both the silo and its contents.
Selection of the dissolving or slurry tank should be based on the characteristics of
h chemical chosen. The necessary design considerations include the solubility of the
~:emical at the lowest water temperature, the time required to dissolve the chemical at
the maximum feed rate, the appropriate ratio of water to chemical that creates .a solu-
. or slurry with the proper concentration, the amount of heat generated dunng hy-
- tion the possibility of flow short-circuiting within the tank, and msta
. II'
atlon 0 f over-
dfa • . I'
flow and drainage pipe connections that are adequate in size and In the proper ~atlOn.
The feed line should provide c1eanout fittings, in addition to a connectIOn for
flushing water, so that any solids or grit will be flushed out before. they clog the
feed pipe. The flow velocity in the feed line should be approxI~ately 3 ftls
(0.9 m/s). If the slurry form of the chemical is to be fed to the processIng water, the
pipeline should avoid sharp turns and steep nses and falls becau~e these features
romote caking and increase the headloss. In other words, the deSign should mml-
~ize erosion and the sedimentation of slurry within the pipeline.
Figure 4.1-4 is a schematic illustration of the general dry chemical feed system.
Figure 4.1-5 is a lime feed system that is installed directly above the applicatIOn
OUST COLLECTOR
/ NOTE:
1. HVORA.UUC INJECTOR
BA.G LOADER _IF BA.CK PRESSURE EXISTS
OR _PROVIDES BEnEA MIXING Of
PNEUMATIC LOA.DING CHEMICAl. W/PROCESS WATER.
SYSTEM
SIGNAL
;r,;~-I.-- BIN ACTIVATOR TAANSMlnER

In the case of hygroscopic chemicals, the storage silo should be made as her- 4-20 ...
metic as possible; this will minimize caking caused by atmospheric moisture. The
silo should also have a discharge bin gate to isolate the feeder for service. This gate
should be followed by a flexible connection, which isolates the vibrations of the
feeder and facilitates the simple connection of the feeder to the bin outlet. If a pneu- ~

matic system is used to load the dry chemicals into a bulk storage silo, the fill con-
nection must be of a suitable size and type so that it is compatible with the deliv-
iIi!
ery trucks of the anticipated chemical suppliers. Other essential items that should
be incorporated into the design are the use of elbows with a long radius in the fill /
/ " HYDRAULIC
INJECTOR
CD i-
line and a dust collector, which should be installed in an easily accessible location.
Recently, a special type of elbow (Vortice Ell by Hammer Tek Corp.) has become
available and has demonstrated very good performance.
Engineers are also advised to investigate commercially available types of pack- y s
DRAIN
aged bulk storage which are combined with feed and mixing systems. One exam-
ple is the ChemTower by suppl iers such as Smith and Loveless, Advance Industries,
Inc., and others. For better chemical dispersion a dissolver or slurry tank is gener-
ally required to create a chemical solution or slurry prior to application to the pro-
UTILITY _ _J.--------1I01--{~M!!MQ_-'
cessing flow. Thus the dry chemical feed system (storage bins or bag storage area, WATER
(OPTIONALI
dust collectors, feeder, dissolvers, duct or pipe works, and grit or waste bins) occu-
pies a considerable amount of floor space and height. Figure 4.1.4 Dry chemical feed system with dissolving tank.
 352 SUBORDINATE PLANT FACILiTIES CHEMICAL FEED SYSTEM 353

VENT
~ -01,1"
. COlUClOII

DUST COLLECTOR

LARGE AAOIUS , _ _ _ _M_.H_.

ELBOW (TYP)
t~'--I'-...11

LIME
SILO

-
LOAD CELL (OPTtONAl)

PNEUMATIC
CONYEYING PIPE
(:(IH'f(CIOII uTl.ITY W41(II

CIol"""'oQH
(0l1W1<
Itt
iii -"
r'---':"'~_ BIN ACTIVATOR
... ,eM
iii'!"
O•

" i
•
.... DIA.
FILL CONNECTION TO
i
BULK DELIVERY TRUCK
9tN GATE
"c / FLEXIBLE COUPLING

GRAVIMETRIC DRY eHEM. FEEDER

r---L--L-~'::'

Figure 4.1-6 Nonionic/anionic polyelectrolyte feed system.

The basic types of dry chemical feeder are gravimetric and volumetric feeders. The
GRIT
gravimetric feeder is preferred because of its wide turn-down ratio (100: I), high ac-
DRUM curacy (I %), easy automation, and ability to feed information directly to a computer.
Moreover, plant operators do not have to refer to a calibration chart to set the chemi-
: .. : i.·.. :::.::.~'.:..•...• ·.·• ::: . . ':.:":'~:,:,': cal dosage. However, these units are significantly more expensive than the volumetric
dry feeders. Typical gravimetric dry feeders are the belt type and loss in weight type.
LIME SLURRY Volumetric feeders include the oscillating hopper, vibrating feeders, grooved disk
FEEDING TROUGH --rCHANNEL
DRAIN Y feeders, screw (helix) feeders, and belt feeders. The first two types should not be
considered for use in the water treatment field because of their low feeding accu-
PROCESS WATER _____
racy (10%). The grooved disk feeders are highly accurate but have a limited ca-

.•.: •. ' .•.......• -'...•...••. ..' ......•...

"
'
", '. ..','.•...
",'.'
Figure 4.1-5 Lime (quick lime) feed system.
point. Figure 4.1-6 is a dry polymer feed system that may also be used as a liquid
polymer feed system. ThIS last system is proven and performs well with minimal
operation and maintenance effort. However, the skid-mounted, self-contained sys-
tem generally costs about $70,000. There are a few types of inexpensive dry poly-
mer feed systems that utthze hIgh-speed blenders without an aging tank. However,
these systems are not effective, due to the high shear forces generated by the blender.
pacity. Their application is therefore limited to small plants. The screw type and belt
, . ,:.:':.
'
feeders are reasonably accurate (± I to 3%). The feeding range of the screw feeder
is 20: I and 10: I for the belt feeder.
The last two types are most commonly used in water treatment because of their
lower cost compared to the gravimetric feeders. Yet the volumetric feeder also has
a disadvantage: The feed rate setting must be calibrated frequently, due to variations
in bulk density from batch to batch. This difference is the result of purchasing the
same chemical from different suppliers, since the purchase of the chemicals is based
on bids. The variations may also be attributed to the particle size, density, shape,
and moisture content of each batch. If different types of chemicals are to be fed by
means of a volumetric feeder, each chemical must have its own calibration chart, so
that its dosage can be set for the feeder.
 354 SUBORDINATE PLANT FACILITIES

00 0.
Gas Feed System With the exception of ozone and carbon dioxide, the gaseous fonn r...:ci
of a chemical is usually fed to the process water as a solution, provided that the gas is
reasonably soluble in water. Although chlorine, anhydrous ammonia, and sulfur dioxide
are shipped and stored in steel cylinders as liquefied gas, they are commonly made into
a solution by means of a hydraulic eductor and fed to the process water as a solution.
However, the liquetied gas is generally evaporated to fonn a gas, then injected into the
process water by means of a vacuum pressure or direct gas feed system.
Gas feed systems are generally simple in mechanism and easy to operate. The
solution feed vacuum type of feeder is manufactured as a proprietary type of feeder lrlt=:'t=:'
and is commonly used in chlorination, ammoniation, and dechlorination, and in feed.
~ 00 ' " ("'". f"";
I I I I I
'-"'-"'-"-
ing chlorine dioxide. This type of gas feeder shuts off the gas automatically upon a
~
0.
loss in vacuum, loss in operating water pressure, or when the solution stops flow- I
ing from the discharge line. Direct gas feeders are seldom used but may be utilized
in situations where pressure water and/or electricity is not available. It is therefore
used primarily in very small facilities such as campsites or small plants located in
remote areas or in developing countries.
Both carbon dioxide and ozone, particularly ozone, have extremely low solubil- ~~~
~ ~ ~

ity in water (refer to Table 4.1-1). These gases must be fed to the process water via
r", 0\
'" '" '"
" 00
r", 'r".
00 00 tr. '"
-.0
a rotameter, where they are released as fine gas bubbles from the bottom of the con- .."f . d N -.ci N
tact channels. The contact channels have a water depth ranging from 12 to 18 ft
(3.5 to 5.5 m). The ozone contact channel must have a water depth of over 18 ft
(5.5 m), due to the very poor solubility of ozone in water. The chlorine gas feed
system is a typical gas feed system and is illustrated in Figure 3.2.9-2. The ozone
feed system is discussed in detail in Section 4.8.

Required Capacity for the Chemical Feeder and Storage The feeder t-
0\
t-
N (",",
000\0\0
should be able to cover both the maximum and minimum chemical feed require- d ci d '
ments. If the turn-down is not large enough to meet the maximum and minimum
requirements, two sets of feeders may be needed to cover the entire chemical feed
range adequately. The feed range may be computed by the following method (Ib/hr):

maximum feed rate = [(maximum dosage X 8.34) X Qrnax-day]

average feed rate = [(average dosage X 8.34) X Q,"e-d,yl
minimum feed rate = [(minimum dosage X 8.34) X Qrnin-day]
Iblhr = [(mg/L X 8.34) X mgd] 0- 24

NOTES:

I. Minimum plant tlow occurs during winter nights or during the period of plant startup.
2. Many chemical feeders have a limited feed range of 10: I to 15: I.
3. In the United States, the capacity of chemical feeders is commonly expressed as poundS
per hoUl" or gallons per hour.

355
 356 SUBORDINATE PLANT FACILITIES
, The quantity of water treatment chemicals that should be stored at the plant site
IS Influenced by local weather conditions, raw water quality, and the local SUPply
situatIOn. Treatment plants are generally considered to be adequately stocked if they
have a 15-day supply; thiS figure IS based on the maximum dosage and the average
dally flow rate. Remote areas or regions with frequent tlooding or heavy snowfall
should have a much larger supply.
storage capacity (Ib) = [(maximum dosagc X 8.34) X Qm-d,yl X 15 days
= [(mg/L X 8.34) X mgdl X 15 days
If the maximum dosage and the maximum daily flow rate are expected to coincide
for more than I week, the following equation may be used to calculate the storage
reqUirement:
storage capacity = [(maximum dosage X 8.34) X Qm"'d,yl X 15 days
Specific Design Issues Due to the different nature of the water treatment chem-
icals, each type of feed system has a few specific design issues that must be addressed
by the engineer. This section discusses special considerations associated with the (I)
polymer, (2) PAC, (3) potassium permanganate, and (4) lime feed systems.
Polymer Feed System Cationic polymers are generally shipped as a solution in
drums or by bulk delivery; bulk deliveries are approximately 3000 gal, but trucks
with 1000-gal tanks are also available in some areas. Nonionic and anionic poly-
mers may be shipped in either liquid or dry form. The dry form is preferred by some
plants because it occupies less storage area and can be stored for a longer period of
time than the maximum effective storage of I year for the liquid form.
The system should be designed to feed both liquid and dry polymers as shown
in Figure 4.1-6, especially in cold weather regions and remote areas. This design
gives the plant operators maximum flexibility in feeding the polymer. All polymers
are very viscous and slippery if spilled on wet concrete floors. Therefore, a concrete
curb should be provided around the polymer feed system, in addition to hose bibs
and floor drains to contain and wash away the spilled material.
Mechanical mixers with relatively low speeds should be used in preparing poly-
mer solutions. The possibility of breaking the polymer chains is decreased if the
mixers are operated at 400 rpm or less. Dilute polymer solutions (I % or less) are
subject to rapid degradation. Thus, the capacity of the working solution (day tank)
should be sized to require a new batch every I to 2 days.
POWdered Activated Carbon (PAC) Feed System There are two basic PAC
feed systems.
• Allernative 1: dry feeder with bag-loading hopper, extension hopper, dust col-
lector, and either a dissolving tank or a vortex mixer tank
• Alternative 2: PAC slurry storage tank with a metering pump
CHEMICAL FEED SYSTEM 357
If PAC feeding is infrequent and the feed rate is less than 150 Iblh, alternative I
h Id be considered because the capital investment is much less. However, If the
sou '2'h
tern is used frequently and the feed rate exceeds 150 Iblh, alternative IS t e
sys . d Wit. h i ' 2'. so I'd
proper choice. Two major proble~s. have b~en aSSOCiate a ter?atI~~ I M
ification of the PAC slurry when It IS not mixed frequently and the Inablhty of hy-
draulic eductors to convey the PAC slurry. The mixer for the PAC slurry tanks should
be designed to prepare 1.5 Ib/gal of slurry (lIb/gal IS standard). . . .
PAC adsorbs organic compounds, including airborne gases. ThiS characterIStic
may reduce its effectiveness, and engineers must incorporate special st~rage con-
siderations into the plant design. A water scrubber type of dust collectIOn system
hould be installed for most PACs and a bag dust collection system should be used
:xclusively for dry PAC. Carrying water is added to the PAC feed line so that a flow
velocity of 5 ftls (1.5 m/s) is maintained. The feed system should also have cleanouts
and large radius elbows. . . .
PAC particles are very abrasive and wet PAC IS extremely corrosive. ~e deSign
of the feed system should take these characteristics into acc~unt. A rotodlP feed~r
'th a hydraulic eductor system has a very successful operational hiStOry, and thiS
~stem should be considered whenever applicable. A 50% moistured PAC is used ex-
clusively in all major water treatment plants 10 Japan With practically no dust prob-
lems. The product has the consistency of natural table salt and IS easy to handle.
Potassium Permanganate Feed System There are two basic potassium per-
manganate feed systems.
• Alternative 1: volumetric dry feed with a bag-loading hopper, extension hop-
per, dust collector, and dissolving tank (no metering pump)
• Alternative 2: a batch system consisting of two large dissolving tanks and a
metering pump
Like other chemicals, the optimum dosage of potassium pennanganate usually
covers a wide band and a deviation of :!:5% from the desired dose does not gener-
ally affect the result. For this reason, alternative. I is preferred. Also, the floor area
required by alternative I is usually half that reqUired by alternative 2 because potas-
sium permanganate has a very low solubility, 3 g per 100 mL at 10°C (50°F) and
2.8 g per 100 mL at 5°C (41°F). Alternative 2 usually requires large dlssolvmg tanks.
Lime Feed System The most effective form of adding calcium hydroxide to
process water is quick lime (CaO). Quick lime with a purity of 85% and hydrated
lime [Ca(OHhl with a purity of 90% are easily obtainable. Compared to hydrated
lime both the storage and dosage requirements of quick lime are lower because
I m~1 (56 g) of CaO produces I mol (74 g) of Ca(OHh after hydration with ~ater.
Thus quick lime should be considered if lime is to be used frequently at a mInimum
dosage of I ton/day. .
Quick lime may be shipped as pebbles or in granular form; the approximate bulk
density is 60 Ib/ft'. In contrast, slaked lime has a bulk density of only 30 Ib/ft3 but
 358 SUBORDINATE PLANT FACILITIES CHEMICAL FEED SYSTEM 359

is available only as a light powder. The quick lime storage silo may therefore b
much smaller, and "flushing" and "bridging" problems associated with silo-store~
slaked hme may be aVOIded. Yet quick lime must be slaked prior to use and grit
must be removed from the slaker on a routine basis.
The lime feed system should consist of a storage silo with mechanical pneumatic
loadIng, dust .collector,
.
bin activator such as those produced by Vi bra Screw, b'In
gate, content Indl~ator (load cells), volumetric or gravimetric dry feeder, slaker (for
qUIck hOle), or dIssolVIng tank wIlh mechanical mixer (for slaked lime). For opti-
mal results, the feeder should be located directly above the application point. This
arrangement will prevent "caking" within the feed pipes.
In recent years, 35 to 40% solids lime slurry has become available for water treat_
ment fo~ certain regions (especially in Texas). The Houston East Plant [350 mgd
(15.4 m-/s)] was the first large plant to convert successfully from the old quick lime
teed system to the hOle slurry system, in 1995. Since then, several large water treat-
ment plants in both the Houston and Dallas areas have converted to the lime slurry
system; these plants are satisfied with the slurry system, despite a few initial de-
bugging problems.
The slurry system has several advantages: (I) no problems relating to lime dust.
(2) much smaller storage and feeding system (overall) compared to dry lime sys:
tem;. (3) practIcally no problems with lime scale formation within the storage and
feedIng systems, provided that dilution of neat slurry is not made; (4) virtually no
grIt; and (5) much less operator attention is required by this system. The disadvan-
tages are that there are a limited number of suppliers, the cost of the solids lime
slurry is approximately 12 to 15% higher than for dry lime, and there is increased
traffic by tank trailers to fill the storage tanks.
The solids lime slurry is generally stored in at least two to three tanks which
have two levels of continuous slow mixing (about 35 rpm) through the use of large-
diameter mixing blades (50% of the storage tank diameter). The power requirement
is approximately I hp per 1000 gal of slurry.
necessary. However, local manual control must be
provided.
Accuracy :t 1.0% is preferred, based on the maximum feed rate.
Storage
Type Noncorrosive material. Alternatively, a steel tank may be
used if it has a protective Jining. The engineer must also
incorporate earthquake, freezing, safety, and aesthetic
considerations into the design.
Capacity Normally 15 days of storage, based on the maximum
dosage and average daily flow rate.
Control A content indicator should be provided for each tank and
possibly another indicator for the main control hoard.
Feed line Noncorrosive materials such as PVC or type 316 SS are
commonly used. If pulsation pumps are used, the feed
line should be sized to accommodate the peak pulsation
flow rate-this rate is approximately three times higher
than the rated pump capacity-regardless of whether a
pulsation damper is used or not. Flushing water
connections and c1eanouts are also necessary items.
Diffuser The type of diffuser depends on the type of chemical
used. A perforated-pipe diffuser should be avoided if
the chemical has scale-forming characteristics and types
C, D, and E of Figure 3.2.3-3 should be considered.
Non-scale-forming chemicals will be well dispersed
by the perforated-pipe diffuser (refer to types A and B
of Figure 3.2.3-3). The diffuser design is based on
Figure 3.2.3-4.
Dry Chemical Feed System

Design Criteria Depending on the form of the chemical, the design parameters
Feeder
for the feeder, storage, feed line, and diffuser may be different. This section discusses
Type Belt feeders for large feed capacities; screw or helix
these issues for (I) liquid chemical, (2) dry chemical, and (3) gas feed systems. The
feeders for small capacities; grooved disk feeders for
design criteria for the treatment plant access road and loading dock are also presented.
small capacities; loss in weight type or vibration feeders
in special cases,
Liquid Chemical Feed System
Feed range Each type of feeder has an inherent feed range and
Feeder accuracy. The following table summarizes the data.
Type Metering pumps, rotodip, magnetic tlow meter or rotameter.
Feed range Capacity Feed Range Accuracy
100: I. This figure includes a range of 10: I for the dosage
setting and a range of 10: I for plant flow pacing. If the
COMMON FEEDERS
feeder does not have a 100: I feed range, in most cases,
more than one feeder must be provided to cover the Volumetric (belt) Minimum, 0.05 ft'lmin 10: I normal :t I % of set rate
range. Maximum, 60 ft'/min 10: I optional
Control Ideally, the plant will have automatic tlow pacing and Gravimetric (belt) Minimum, 0.05 Iblmin 100: I normal :t I % of set rate
remote dosage setting control, but this is not absolutely Maximum, 2,000 Ih/min
 360 SUBORDINATE PLANT FACILITIES CHEMICAL FEED SYSTEM 361

Screw (helix type) Minimum, 0.02 ft 31h 20: I normal ±I% by volume Design Criteria for the Access Road and Loading Dock The dimensions of
volumetric fceder Maximum, 40 ft 31h ±3% by weight typical delivery trucks arc presented below.
Grooved disk type Minimum, 0.17 ft 31h 100: I normal ± I% by volume
volumetric feeder Maximum, 17 ft 31h Straight Conventional
±3% by weight
Van Body Semitrailer
MISCELLANEOUS
Length (ft) 15 to 20 17 to 35 55
Loss in weight type Minimum, 20 Iblh Width (ft) 7 8 8
10: I normal ±1% of set fate
(gravimetric) Maximum, 500 Iblh Height (ftfin.) 7 13/6 13/6
Floor height (ftlin) 2 to 2/8 3 to 4 4 to 4/4
Vibration feeder Minimum, 0.7 ft 31h 40: I normal ± 10% by weight
(volumetric) Maximum 40 ft 31h Track (ftlin.) 5 to 5/4 5/10 6/6
Rear axle (ftfin.) 1/3 to 12 4 to 12
Control The same as liquid feeders.
Accuracy The same as liquid feeders. The allowable maximum gross weight (lb) may be calculated from 1025 (D +
Storage This is basically the same as the liquid feed system with 24) - 30D2, where D is the distance (in feet) between the tirst and last axle of the
respect to type, capacity, and control. Design vehicle or combination of vehicles (AASHO). Generally, the maximum cargo weight
~o~si~erations are covered in the preceding section.
is 18 tons. The required turning radius for a 55-ft semitrailer and truck combination
Feed line This IS lundamentally the same as for the liquid feed system, is illustrated in Figure 4.1-7.
except that a removable top channel with gravity flow is
often used for lime slurry. Risk Management Planning Water and wastewater treatment plants must com-
Diffuser The design of the dilTuser is the same as that for the liquid feed ply with the EPA Risk Management Program (RMP) rule (40 CFR 68) by June 21,
system. Type E of Figure 3.2.3-3 illustrates a trough type of 1999 if they use, handle, or store more than 2500 Ib (1135 kg) of chlorine andlor
diffuser. store 5000 Ib (2270 kg) of sulfur dioxide at their facility. Propane has an RMP
threshold of 10,000 Ib (4500 kg). This rule does not apply to small facilities.
Gas Feed System The RMP rule is catagorized under the Clean Air Act Amendments of 1990,
Section 112, but is executed differently throughout the United States. The EPA has
Feeder established three different program levels for facilities affected by RMP. Level I fa-
Type cilities present minimal risk to surrounding communities and therefore have reduced
Rotameter system, venturi tlow meter system.
Feed range compliance requirements. Level 2 and 3 facilities present increasing risk, and their
20: I preferred, with a minimum of 10: I.
Control requirements for compliance arc more complex. Facilities covered by OSHA gen-
Aut(~matic. flow pacing is required; a closed-loop control
erally must comply with the provisions of level 3.
With reSidual analyzer and plant flow meter is preferred.
Accuracy Levels 2 and 3 have a number of similiar RMP requirements: A RMP plan must
±2% based on the maximum feed rate.
Storage be written, available to all personnel at the facility, and updated whenever there are
Type changes in the usc of threshold chemicals. The plant must document the manner in
A high-pressure steel cylinder is standardly used in
which the chemicals are used at the facility and include the following five items:
storin~ most liquefied gas. For large installations, a larger-
capacity tank mounted on a trailer or a tank railcar
I. A detailed hazard assessment that evaluates the possible hazards of using the
(stationary storage tanks) may be used. Ozone must be
chemical at the plant and how these hazards can be reduced
generated on site.
Capacity This is basically the same as the liquid feed system. 2. Worst-case and most-likely-case scenarios with reasonably good estimates of
Control The same as the liquid feed system. the popUlation and sensitive receptors (schools, hospitals, nursing homes, and
Feed line Fundamentally the same as the liquid feed system. offices) within the affected area
However, in the case of ozone, type 316 SS is the only 3. A 5-year history of accidents which reports any outside releases and the extent
material that may be used. 4. A written emergency response plan that is coordinated with the facility's lo-
Diffuser Similar to the liquid feed system, but type F of Figure cal emergency planning committees if plant personnel will respond to acci-
3.2.3-3 must be used for carbon dioxide and ozone. dental releases
 362 SUBOROINATE PLANT FACILITIES CHEMICAL FEEO SYSTEM 363

L = See table for maximum length~ Determine

I Average = 55'-0· x S'-O·wide

~
I
Road tractor
or city tractor
(i) The number and capacity of each feeder

+&r--T"~~~~
(ii) The number and capacity of the storage tank(s)
(iii) The major design considerations

Solution The maximum, average. and minimum plant flow rates should be deter-
f .. .Bulk tank and tractor mined first.

.
Tire size approximately 41 " ::t diameter x 10·:t wide

'f
Qmo> = 1.5 X 20 = 30 mgd
N
Q"e = 20 mgd
...
'v.
--- -- ......... ~ goo Qm;n = 0.25 X 20 = 5 mgd

--0 --',_----f'----
,\ Path of
Commercially available liquid alum is used for operator convenience. Since liquid
...: ' \ left front alum contains 5.6 Ib of dry alum per gallon,
.,,-:; wheel

\ r-\ Path of
overhang
maximum dosage rate = 50 X 8.34 X 30 = 12,510 Ib/day or 93 gallh
average dosage rate = 15 X 8.34 X 20 = 2502 Ib/day or 18.6 gallh
Path of
50'-0" Minimum right rear minimum dosage rate = 8 X 8.34 X 5 = 334 Ib/day or 2.5 gallh
p~actical turning side
radius to centerpoint r turn-down ratio = 93: 2.5 gallh or 37 : I
of left front wheel I
, I
, I (i) The system should have a hydraulic diaphragm type of metering pump with
45'-0· Semitrailer
SO'-O' x S'-O' unit \ !: • I
a 100-gallh capacity. a 100: I turn-down unit, and a second metering pump with a
capacity of 50 gallh. A third metering pump (50 gal/h) is specified as a spare since
II
"-'-
I 'lil'
\ I
alum is one of the most important chemicals in the treatment process. Thus, one
--.l ~s'-O' loo-gal/h and two 50-gallh metering pumps are specified.
5~';-O· Semitrailer and tractor combination (ii) The storage capacity is calculated to be [(50 X 8.34 X 20) X IS] + 5.6 =
MInimum practical turning radius of SO' -0·
22,340 gal. Thus one 23,000-gal or two 12,000-gal fiberglass storage tanks should
Figure 4.1-7 Turning radius required for a 55 ft semitrailer and tractor.
w

be provided. Freezing is not a likely problem in this location.

(iii) The major design considerations are listed below.

5. ~ prevention, program that covers safety information, hazard review, opera- (a) The alum dosage should always be based on dry alum even when liquid alum
tIons and mamtenance, and management practices, training, incident investi- is used.
gation, and compliance audits (b) Alum is a very corrosive chemical. PVC, FRP, type 316 SS, or Hasteloy C
SS are therefore the most appropriate types of material.
Example Design Calculations
(c) Alum should be fed as a continuous flow.
Example 1 (d) An effective flash mixing system is essential at the application point.
(e) To avoid clogging in both the feed lines and the diffuser. it is recommended
Given A treatment plant with a daily average flow rate of 20 mgd is located in that neat alum be fed to the process water.
North Carolina and requires an alum feed and storage syslem. The maximum _ (f) The alum fill line should be a type 316 SS pipe instead of PVC because liq-
d.. ,0
erage. an mmlmum alum dosages are 50, 15, and 8 mglL, respectively. uid alum may be delivered at temperatures over I30°F (54°C). The exits of
 364 SUBORDINATE PLANT FACILITIES
the fill line should be above the highest liquid level in the tank because there
is a tendency for the liquid alum to shoot out from the air vent at the last
moment during the unloading operation; this is a function of the massive air
supply from the delivery truck blower.
(g) The feed system should be designed in accordance with the safety codes and
rules, including OSHA.
(h) The design should allow for ferric chloride or ferric sulfate (alternative co-
agulant) to be used in the alum feed system. Since ferric salts are more cor.
rosive than alum and may contain some grit, engineers must take these is.
sues into account when designing the feed system. Examples are the lise of
a hopper tank bottom to remove grit, Hasteloy C stainless steel as the metal,
and Kynar for the plastic parts. Ferric chloride solution freezes at O°C (320F)
if the solution strength is over 45%, but a 40% solution will not freeze un-
til the temperature reaches 10°F (-12°C), approximately the same freezing
point as that of commercial liquid alum (49% solution).
The concept of using an alternative coagulant is important because any treatment
plant that achieves alum tlocculation above a pH of 7.5 will often have a high con-
centration of residual aluminum, equal to or above 0.2 mgIL, in the process water.
The general public and the EPA are concerned about high levels of aluminum in
potable water due to its possible connection to Alzheimer's disease. It should also
be noted that ferric coagulant is much more effective than alum for flocculation and
clarification in cold months.
Example 2
Given A treatment plant with a maximum daily !low rate of 50 mgd (2.2 m'/s) is
located in California and requires a caustic soda feed system for pH control. The
maximum, average, and minimum dosages are 10, 5, and 2 mg/L, respectively.
Dete,mine
(i) The number and capacity of each feeder
(ii) The number and capacity of the storage tanks
(iii) The major design considerations
Solution The maximum, average, and minimum flow rates are as follows:
Qm" = 50 mgd
Q"e = 50 -;- 1.5 = 33 mgd
Qm;" = 0.25 X 33 = 8.3 mgd, say, 8 mgd
Commercially available 50% liquid caustic soda is purchased and stored as a 25%
solution to avoid crystallization during the winter. The 25% solution contains
2.7 Ib of dry caustic soda per gallon.
CHEMICAL FEED SYSTEM 365
maximum dosage rate 10 X 8.34 X 50 = 4170 Ib/day or 64 gallh
average dosage rate =5 X 8.34 X 33 = 1376 Ib/day or 21 gallh
minimum dosage rate =2 X 8.34 X 8 = 133 Ih/day or 2 gallh
turn-down ratio = 64: 2 or 32 : I
(i) The capacity of the pump selection is the same as for Example I, except that
a spare pump is not specified because this chemical is not absolutely essential in
water treatment.
(ii) The storage capacity is calculated to be [(10 X 8.34 X 34) X 15] -;- 2.7 =
15,753 gal of 25% solution. Provide two identical 10,OOD-gal tanks because a dilu-
tion step is required even during mild California winters.
(iii) The major considerations for a caustic soda feed system are as follows.
(a) The dosage is based on the dry weight and expressed in mg/L.
(b) A 50% solution of caustic soda will freeze at 55°F (13°C); however, a 25%
solution will freeze at O°F ( - 18°C); refer to Figure 4.1-8.
(c) Liquid caustic soda is not corrosive to steel.
(d) Caustic soda is toxic and dangerous to handle.
(e) Scale formation will occur when caustic soda is diluted with water unless
the water is very soft. Also, heat is generated by dilution.
(f) Caustic soda generally costs eight times more than lime, but the cost of re-
quired dosage for pH control is approximately the same.
(g) Plant operators usually prefer to use liquid caustic soda rather than lime since
it is easier to handle.
(h) The design of the feed system should be in accordance with current safety
rules and regulations.
Example 3
Given A treatment plant with a daily average tlow rate of 20 mgd (0.88 m'/s) is
located in Michigan. The plant requires a polymer feed system to function as a tloc-
eulation aid, as well as a filtration aid. Pilot tests confirm that an anionic or non-
ionic type of polymer feed system is needed. The maximum, average, and minimum
dosages are 0.5, 0.2, and 0.05 mg/L, respectively.
Determine
(i) The number and capacity of each feeder
(ii) The amount of required storage
(iii) The major design considerations
Solution Commercially available anionic or non ionic polymers are primarily of
the dry form; however, a limited number of these polymers are available in liquid
form. The system design should allow for both dry and liquid forms to be used, even
366 SUBORDINATE PLANT FACILITIES CHEMICAL FEED SYSTEM 367

maximum feed rate = 0.5 X 8.34 X 30 = 125 Ib/day

= 125 -c 0.042 = 2976 gal/day or 124 gallh
1 Ice pluS a to 19~ NaOH in solution average feed rate = 0.2 X 8.34 X 20 = 33 Ib/day or 33 gal/h
2 Ice plus NaOH.7H2<)
3 NaOH plus 19 to 22" NaOH in solution
4 NaOH.5HzQ plus 22 to 25% NaOH in solution minimum feed rate = 0.05 X 8.34 X 5 = 2 Ib/day or 2 gallh
5 NaOH·SHl<) plus NaOH.7H20
6 NaOH.4H~ plus 25 to 32" NaOH in solution turn-down ratio = 124: 2 or 62: I
7 NaOH.4HzQ plus NaOH.SH20
8 --+--+-.j
NaOH.3.5H20 plus 32 to 46% NaOH in solution
9 NaOH.3.5H 20 plus NaOH.4H20
10 NaOH·3.5HzO plus NaOH.2HzO (i) A hydraulic diaphragm type of metering pump with a 150-gal/h capacity
11 NaOH·2HzO plus 46 to 51 % NaOH in solution
12 NaOH.HzO plus 51 to 74% NaOH in solution -7""q:r:~::= 100: I turn-down, and another unit of 50-gallh capacity should be installed.
13 NaOH.Hl<) plus NaOH-2H20
14 NaOH plus 74 to 100% NaOH in solution
(ii) The amount of storage is calculated to be [(0.5 X 8.34) X 20J X 15 = 1251
15 NaOH.H20 plus NaOH Ib of dry polymer. An ordinary bag of polymer weighs 30 lb. Therefore, 1251 .,. 30
= 42 bags, let us say 40 bags. Twenty bags should be stacked on a pallet for easy
hauling and to protect the bags from the wet floor. Additionally, floor space for four
55-gal drums should be provided for the liquid form of the polymer.
(iii) The design considerations for the polymer are as follows:

15 (a) The dosage is based on the dry weight of the chemical. In the case of liquid
polymer, the dosage should be determined by running jar tests, assuming that
the stock solution is a 100% solution, for operator convenience.
(b) It is difficult to produce a polymer solution that forms "fISh eyes" when mak-
ing solution from dry polymers using ordinary dissolving tanks.
(c) Polymers that are EPA approved for the treatment of drinking water should
be selected and the maximum dosages should conform with the criteria set
by the EPA.
% NaOH
Cd) Polymers act selectively; care must be taken to choose a polymer that per-
forms best under the given conditions.
Figure 4.1-8 Graph of freezing points and solubility for a caustic soda solution. Complete
(e) The application sequence plays an extremely important role in the effective-
solution prevails at points on or above the freezing point line. Below this line, some of the
caustic socia exists as a solid, anhydrous or hydrated, depending on the temperature and com- ness of the polymer (see Table 3.2.3-1.)
position. Numbered areas and table inset indicate the nature of the mixtures. (Courtesy of (I') Never overfeed polymers; their optimum dosages are located in a very nar-
Stauffer Chemicals, New York, Callslic Soda Bulletin, 1965.) row range. A minute amount of overfeeding will cause the turbidity colloids
to restabilize, thus resulting in a sharp increase in turbidity.
(g) Since diluted polymer solutions deteriorate quickly, (during hot weather)
though the severe winter temperatures are not favorable for liquid polymer because
every batch should be used within 2 days.
of the potential for freezing.
(h) Some polymers are corrosive in nature.
Qrn" = 1.5 X 20 = 30 mgd
Operation and Maintenance The major issues in operation and maintenance
Q,", = 20 mgd
are (I) safety programs and training, (2) good housekeeping, and (3) good record
Qrn;" = 0.25 X 20 = 5 mgd keeping. Each of these issues is discussed in detail.

Most dry polymers are very viscous, at concentrations over I%. Therefore, a 0.5% Safety Programs and Training The chemicals used in water treatment processes
concentration is usually recommended; the 0.5% solution should contain 0.042 Ib (Table 4.1-2) are for the most part harmful to human health. Thus the water pro-
of dry chemical in I gallon of solution. duction manager and water supply department must establish a formal safety pro-
 TABLE 4.1-2 Major Water Treatment Chemicals

'"enco Available Solubility Major

Bulk Weight or
Dry Chemical in
Suitable
Name Materials Strengths (giIOO mL) Suppliers Solution Remarks
Aluminum PVc. rubber. Dry (17% A1 2 0 3 ) 78 at 10°C AlIied Chemicals. 60 Ib/CF (dry). Mostly liquid alum.
sulfate (alum) FRP. 316 SS or 49% liquid Staffer Chemical 5.5 Ib of dry S.G. ~ 1.33 (LA).
alum (LA) alum/gal of corrosive
49% solution
Ferric chloride PVC. Hypalon 35-45% solution Infinite Dow Chemical. 4.6 Ib dry EC.lgal S.G ~ 1.46 (42%
Hastelloy C. Pennwalt Corporation of 42% solution solution), very
FRP DuPont corrosive
Ferric sulfate PVC. rubber, Dry (68-78%) or 20.5 at 10°C Stauffer Chemical. 70 Ib/CF (dry). 6.3 Ib Hygroscopic, S.G. =
FRP, 316 SS 50% liquid Tennessee Chemical dry ES. in I gal of J.5 (50% solution).
50% solution corrosive
Cationic FRP. PVC, Mostly liquid Infinite Dow Chemical, Nalco 8.5-9.5 Ib/gal Eye and skin irritant.
polymer 316 SS Chemical, American may be corrosive
Cyanamid
Anionic polymer FRP, PVC. Mostly dry Intinite Dow Chemical, Nalco 45-551b/CF Eye and skin irritant
316 SS Chemical. American
Cyanamid
Nonionic FRP. PVc. Mostly dry Infinite Dow Chemical, Nalco 42-481b/CF Eye and skin irritant
polymer 316 SS Chemical. American
Cyanamid
Chlorine gas Steel. copper 99.8% purity 0.98 at 10°C Dow Chemical. Jones d ~ 3.214 (air ~ I) Poisonous and very
for gas: Chemical, McKeeson corrosive when wet
Hastelloy C. Chemical
PVC, Vitor
for solution
Sodium Rubber. PVC, 12-15% Infinite Jones Chemical, Hill 10.2 Ib/gal (15%) Highly alkaline short
hypochlorite vinyl. Brothers company life, I gal (12.5%)
Hastelloy C Occidental ~ lIb ofCI,

Dry form with 22% at 10°C Delta Chemical. 70 Ib/ft3 (granules) Hygroscopic
Calcium Rubber. PVc.
hypochlorite Hastelloy C 70% CI, Jones Chemical.
Penn walt Corporation
Infinite Essex Chemical. 10.5 Ib/gal Corrosive. toxic.
Fluosilicic acid Rubber. PVC. 20-30%
(16-24% F) Thompson Chemical (30% solution) attacks glass
Hypalon.
Carp. 20
95-98% 4% at 10°C, Allied Chemical, 90-105 Ib/CF Poisonous
Sodium fluoride Rubber. 316
SS, PVC, (43-44% F) 4.1% at Ashland Chemical. (crystal)
Hypalon 25°C J.T. Baker Chemical
98.5% (60% F) 0.46 at 10°C, Ashland Chemical, 85-1051b/CF Poisonous
Sodium Rubber, 316
silicofluoride SS, PVC, 0.76 at Delta Chemical,
Hypalon 25°C Olin Corporation
97-99% 3.3 at 10°C. Ashland Chemical. 86-100 Ib/CF Hygroscopic.
Potassium Steel, 316 SS
5.0 at 20°C Carns Chemical oxidant, toxic
permanganate FRP, PVC
Infinite Ashland Chemical, 64-100 Ib/CF, glass Hygroscopic,
Sodium 316 SS, rubber. 67% PzO,
FMC Corporation, or flake sequestering agent
hexameta- PVC, RFP (minmum)
phosphate
Monsanto Company
8% zinc in sulfate Infinite Technical Products 10.5 Ib/gal, S.G. ~ Acidic solution, avoid
Zinc RFP. PVC,
type, 25% zinc Company. Monsanto 1.27, skin/eye contact
orthophosphate 316 SS
in chloride type Industrial Chemicals, z.. ~ 0.83 Ib/gal
Conray Chemical
25.2 at 15°C Ashland Chemical, 75-90 Ib/CF Poisonous. algaecide
Copper sulfate Rubber, PVC. 99%
RFP, 316 SS Jones Chemical.
Westco Chemical
0.35 at 10°C. Union Carbide, Arco d ~ 1.53 (air ~ 1) Liquefied CO2 in
Carbon dioxide Steel, 316 SS 99.5% liquid
CO z,10-14% 0.15 at Chemical cylinder, recarbon-
PVC
in combustion 25°C ation use

'"en
'"
Co>
..... TABLE 4.1-2 (Continued)
0
Bulk Weight or
Suitable Available Solubility Major Dry Chemical in
Name Materials Strengths (gll00 mL) Suppliers Solution Remarks

Sulfur dioxide PVC, ceramic, 100% 16.2 at 10°C. Delta Chemical, Jones d - 2.26 (air = 1) Liquefied S02' in
Hypalon, 11.3 at Chemical. Stauffer cylinder. dechlorin-
Carp. 20 20°C Chemical ation use
Air (reference) 14 mg/L at d = 1.29 giL (dry)
10°C,
10 mgIL at
25°C
Ozone gas Ceramic. Teflon 2-8% depending 6.5 mgIL at On-site generation only 2.14 gIL at O°c, Toxic. fire hazard.
316 SS. glass on generator 10°C. d = 1.66 (air = 1) strongest oxidant
3.5 mg/L at
2YC
Hydrogen Aluminum. 35, 50. and 70% Infinite FMC. DuPont, Inferox S.G. = 1.29 (70% Potentially a fire!
peroxide type 5254; solutions solution) explosion hazard.
polyethylene; eye and skin
porcelain; irritant
316SS (fair)
Ammonia gas Steel. 316 SS. 99% NH, 40 at 10°C Delta Chemical, 0.77 giL. d = 0.56 Pungent, toxic.
neoprene Monsanto Industrial (air = I) explosive
Chemicals. Olin
Corporation
Aqua ammonia Iron. 316 SS. 30% NH, (26 Bc) Intinite Allied Chemical. Hill S.G. = 0.897 Unstable. strongly
PVC Brothers. Jones alkaline
Chemical
Sodium chlorite Glass. PVC. Dry (81%) 35 at 10°C Olin Corporation, 25% solution weigh Hygroscopic,
FRP. or 25% Vulcan Chemical 10.25 Iblgal combustible with
Hastelloy C solution Drew Chemical organic matter

,~

22 at lOoC AC Industry, J.T. Baker. 601b/ft' Emits heat upon

Quick lime Iron. concrete, 82-95%
Kraft Chemical contact with water
(CaO) PVC
22 at 10°C AC Industry. Asher- 301b/ft' Dusty. irritant
Hydrated lime Iron. concrete. 82-95%
PVC Moore Company,
[Ca(OHhl
Rugger Chemical
51 at 10°C Stauffer Chemical, 60 Ib/ft' or 2.7 Ib Toxic, dangerous
Sodium hydroxide Steel, iron Dry or 50%
Allied Chemical. NaOH in 25% to handle
(caustic soda) liquid
FMC Corporation solution
12.5 at 10°C Hacros Chemical. 40-601b/CF Hygroscopic
Sodium carbonate Steel. rubber. 99% pure
PVC Ashland Chemical,
(soda ash)
Stauffer Chemical
Infinite Allied Chemical, 15.2 lbigal (66 Be). Very corrosive.
Sulfuric acid Steel. PVC for 60 Be; 93%
Ashland Chemical. 14.2 Ib/gal (60 Be) hygroscopic
92% acid~ H 2 S04 ; 60 Be;
Carp. C. 78% H2SO4 ; Monsanto Chemical
Viton for 50 Be; 62%
<92% H2SO4
22 at O°C Stauffer Chemicals. 2.26 = d (air = 1.0); Acid gas; vapor
Sulfur dioxide Steel for dry 100% S02
16 at 10°C Virginia Clum S.G. = 1.363 pressure: 15 psi at
gas; wet gas;
11 at 2ifC (H 20) = 1.0 40°F. 70 psi at
glass, Carp.
100°F
20, PVC. 8 at 30°C
ceramic.Viton
Calgon Corporation. 13-46 Ib/fr' (20 avg.). Dusty. can be
Activated carbon 316 SS, FRP. Power (pAC) 200
mesh size or American Novelty 25--26 Ib/ft 3 explosive
rubber.
Hastelloy C. granules Company. Trans-
epoxy paint, (GAL); E.S.; Pacific Carbon
Saran 0.6-0.9 mm; Corporation. Calgon
UC; 1 1'>-24

Co>
.....
~
 372 SUBORDINATE PLANT FACILITIES
gram and conduct periodic hands~on training. It is essential that every treatment
plant have a custom-made operation and maintenance manual for each type of chem-
ical feed system. For example, potassium permanganate (KMn04) and ferric chlo_
ride (FeCI,) must not be mixed together since toxic chlorine gas (CI,) is formed.
Furthermore, the necessary type and number of special clothing, emergency show-
ers, and eyewashes should be checked routinely.
Good Housekeeping Good housekeeping methods are an important issue since
the chemicals are essentially hazardous in addition to being toxic. Any chemical
spill should be cleaned immediately and the accumulation of gas and dust must be
avoided through the frequent use of exhaust fans and a vacuum cleaner. Polymer
solution spilled on the floor will result in extremely slippery conditions; all spills
should therefore be washed out immediately by means of the floor drains. Operators
should note that wet concrete floors arc also very slippery. Appropriate hazard signs
should be posted in highly visible locations at all times. Finally, regular mainte-
nance of all mechanical equipment must be performed on time, and the chemical
feed lines and diffusers should be checked periodically for clogging.
Record Keeping Record keeping is a vital task because it is required by local,
state, and federal regulatory agencies. Hourly, daily, and monthly records must be
kept with respect to the type and dosage of each chemical used. The chemical in-
ventory must also be known at all times in order to maintain an adequate supply of
Irea.tmen! chemicals on site. Moreover, the operating performance of each piece of
equipment must be logged on a daily basis and all pertinent information should be
convcyed to the next shift operator through verbal communication and in writing
(on the daily opcrational data sheet). Computcr~assisted instrumentation and con-
trol can greatly assist in record keeping.
BIBLIOGRAPHY
AWWA, Water Fluoridation: Principles and Practices, Manual M-14, American Water Works
Association, Denver, CO, 1977.
AWWA. Water Quality and Treatment: A Handbook of Public Water Supplies, 3rd ed.,
McGraW-Hili, New York, 1971.
BIF, Chemicals Used in Treatment of Water and Wastewatel; engineering data, 8IF,
Providence, RI, August 1982.
Williams, R. 8., ancl Cu1p. G. L., eds., Handbook of Public Water Systems, Van Nostrand
Reinhold, New York. 1988.
4.2 HANDLING FACILITY FOR WASTE WASH WATER
Purpose Conventional water treatment processes ordinarily produce 2 to 3% of
thc trcated water as tilter waste wash water. Prior to 1965 it was common practice
to discharge this waste back to the water source or other public water course.
HANDLING FACILITY FOR WASTE WASH WATER 373
However, due to the passage of Public Law 92-500 and the Water Pollution Control
Act Amendments of 1972, this practice is now prohibited. Thus, both the waste wash
watcr and sludge from the clarifiers are categorized as industrial wastes, and a per-
mit must be obtained from the local regulatory agency to discharge them; the agency
is governed by the National Pollutant Discharge Elimination System (NPDES).
Since the passage of Public Law 92-500, many treatment plants have been recy-
cling the wash waste with or without clarification, and thi~ practice h~s. not had ad-
verse etfects on the efticiency of the process or the quality of the hl1ls~ed water.
However, a recent EPA proposal seeks to restrict the rules governmg surf~ce wat~r
treatment by requiring a high degree of treatment-including tloccul.atlOn, sedI-
mentation, and disinfection-before recycling. Furthermore, the recyclmg must be
directed to the head of the plant and not the filter intluen!. . . ..
This section describes how to design a waste-wash-water handling facility so that
it mects the recent requiremcnts set by the regulatory agcncies: The desig~ of a
waste recycling system, where the waste is routed to the head at the plan~ Wilhout
a purification process, is also discusscd because this practice is still applicable to
most watcr treatment plants outside the United States.
Considerations The quality and quantity of the wash waste, the desired quality
of the recyclcd or discharged (treated waste), and the type and capacity ot the tr~at­
ment process are the major design considerations for the waste-wa~h-water handl~ng
facility. Alternatives such as lagooning, tlocculation and sedimenta~lon, and recyclmg
without a purification process arc also important issues and are discussed as well.
Quality and Quantity Many design engineers have chosen to mix together the
waste wash water and the sludge from the basins. Yet thiS practice IS undeSirable
becausc the solids concentrated in the sludge are allowed to disperse in ~he waste-
water. 0.5 to 1.5% of thc sludge is composed of concentrated solids: IllIcroorgan-
isms, organic compounds, and heavy metals that have be~n extracted ~nd settled
from the raw water. In contrast, waste wash water has a solid concentration of 0.01
to 0.04%. If the two arc mixed, the turbidity of the water increascs and more chem-
icals are required to treat the sludge and waste mixture. This can ~otenti~lly lead to
the recycling of objectionable compounds and the impo~tation ot un?esuable taste
and odor (of the sludge) to the water. Thus the sludge trom the baSinS should not
be mixed with the tilter wash waste (which is recycled) but should be diScharged to
the sludge handling system.
The quality of the wash waste depends on the raw watcr quality and pretreatment
unit processes. For a conventional watcr purification plant using alum as the coag-
ulant, the typical propertics of the waste are
Turbidity: 150 to 250 ntu
BOD: 2 to 10 mg/L
COD: 30 to 150 mg/L
• pH: 6.7 to 7.5
• TDS: 0.01 to 0.05%
 --
374 SUBORDINATE PLANT FACILITIES HANDLING FACILITY FOR WASTE WASH WATER 375

The quantity or production rate of the wash waste is also a function of the raw wa~
ter quality and the type of pretreatment processes. in addition to the efficiency of
200
the processes. The percentage of waste to the amount of water treated is usually
about 2% during the warm months and 3% during the cold. Direct filtration and in-
line filtration plants generally produce higher percentages of waste. and figures
100
greater than 10% have been reported during episodes of algal bloom and turbidity
spikes in the water source. Engineers are strongly advised to recycle the waste wash
water rather than discharging it after the clarification process because the cost of 50
water is often rather significant.

Quality Goal of the Recycled or Discharged Treated Waste The quality goal ...51'; 30
of the recycled treated waste should be equal or better than the average raw water ~
20
quality. However. if the waste wash water is to be discharged to a public water
~
course. the quality of the treated waste should basically be equal to or better than
the quality of the receiving body. Nonetheless. the NPDES permit may include other
-
...
~

10
special requirements.

Type and Capacity of the Treatment Process The type of the waste treatment
process is fundamentally the same as that of a conventional tlocculation and sedi-
mentation process with disinfection. Proprietary reactor-claritiers, which have a large
internal flow recirculation, have good process efficiency and are tolerant of shock
loadings. Other advantages include the ability to thicken the sludge. easy sludge re-
moval, and a small site requirement. Regular horizontal-flow rectangular tanks are
also effective waste handling facilities. although they may require a larger site area
than the proprietary types. due to their lower allowable hydraulic loading rates. The
high-speed microsand settling process (ACTIFLO) should also be considered.
No matter the type of unit. the best process performance is always achieved if
the units are operated on a continuous basis and not in an on~and-off fashion. This
ideal condition may be attained by providing a large waste holding tank and by
pumping the waste into the treatment unit at a constant tlow rate. The capacity of
the waste treatment facility should be based on both the maximum waste produc-
tion rate and the tolerable recycling rate of the treatment plant. The practical max-
imum recycling rate is 20% of the plant flow rate. yet the design should be based
on a recycling rate of 10% or less. so that the hydraulic loading of each process unit
is not increased significantly. The sludge produced by the waste treatment facility.
from the sedimentation basins. should be discharged to a sludge handling facility
for further treatment.
Alternatives Under normal conditions the suspended matter in the raw water is
already well coagulated and flocculated prior to filtration. Furthermore. additional
effective flocculation usually occurs in the filter bed during filtration. Consequently.
the backwashing wastewater contains well-tlocculated particles that settle very well.
Figure 4.2-1 presents the batch settling characteristics of wastewater without addi-
tional flocculation with coagulant. The figure shows that the supernatant reaches
nearly 10 ntu. from an original turbidity of 210 ntu. after 30 min of settling time.
5
3
2
10L-~20--40J--60~------1~20~----~180~------~
Settling time (min)
Figure 4.2-1 Settling characteristic of waste wash water.
A settling time of I to 2 h yields a supernatant of nearly 5 ntu. Knowing the quick-
settli ng characteristics of the waste wash water, two basic treatment alternatives may
be used: (I) lagooning with disinfection and (2) ordinary tlocculation. sedimenta-
tion. and disinfection. Moreover. the wash waste may be recycled to the head of a
conventional complete process without settling, as long as the waste is fresh; this
practice is still a feasible alternative under certain conditions.
LAGOONING Lagooning is an effective method of clarification if the plant site has
adequate space and if the filter backwash intervals are reasonably spaced. The la-
goon should be large enough to hold the wash waste of approximately 10 back-
washes, or a series of three or more smaller lagoons, each holding three or four fil-
ter wash wastes. may be used. All lagoons should be elongated in shape to provide
the maximum distance between the inlet and outlet. This configuration minimizes
the effect of the suspended sludge. which is caused by the stirring action of the in-
coming backwash waste. The inlet of each lagoon should have an energy dissipa-
tor. The outlet should be designed to decant as well as drain the lagoon. and should
 376 SUBORDINATE PLANT FACILITIES
also act as an overnow facility. The sidewalls of the lagoon must be protected from
erosion, rainfall, and wave action. Each lagoon should also have an adequate access
road, which allows vehicles to remove accumulated sludge. Finally, a disinfection
facility, together with a recycling pump, should be provided at the eftluen! of the
lagoon. Furthermore, engineers should realize that the lagoons tend to breed mos-
quitos and other aquatic organisms, in addition to taste and odor problems charac~
teristic of certain regions.
FLOCCULATION AND SEDIMENTATION PROCESSES A conventional flocculation and sed-
imentation pretreatment unit, with a disinfection facility, is required whenever the
site does not have adequate space for lagoons or when settleability of the waste is
unacceptable, unless conventional flocculation and sedimentation are implemented.
As discussed previously, either the upflow type of reactor-clarifier or a conventional
horizontal-flow rectangular flocculation and sedimentation basin may be used. In
most cases, the preferred chemical is either 15 mgIL of alum Or 2 mg/L of cationic
polymer. The process will have 20 min of flocculation, followed by a surface load-
ing of 0.8 to 2.5 gpm!ft' (2 to 6 m!h) and a settling time of 0.5 to 2 h, depending on
the type of unit provided. A disinfection facility should be included in the design.
The ~onventional pretreatment process should have a backwash waste hOlding
tank. ThiS tank must be large enough to accommodate the water of two to three back-
washes, depending on the anticipated frequency of filter washing. Floc removal should
be achieved during the subsequent process (clarification), not in the holding tank.
The holding tank may be square or circular rather than elongated. These shapes will
prevent the floc from settling since the inflow creates a stirring effect. Experience
has shown that the ideal holding tank is cylindrical in shape and has an inlet pipe
that produces a tangential flow direction, thus creating a swirling or vortex flow pat-
tern within the tank (consider the Coriolis ellect for the direction of the swirling tlow
pattern). To be able to drain the tank completely and collect tilter medium, which is
very abrasive to the pump impeller, a sump must be provided at the center of the
noor of the holding tank. Furthermore, an overtlow facility should bc included in the
design. Several actual installations of this type of holding tank have performed well.
Tanks that have the transfer pump located in the center of the circular tank have
been observed to be virtually free of sludge accumulation. The capacity of the trans-
fer pump used in the flocculation and sedimentation units is dictated by the antici-
pated maximum filter washing frequency; the waste tlowing into the holding tank
must be pumped out to accommodate the next filter washing. Similarly, the rated
capacity of the treatment process is determined by the maximum pumping rate of
the transfer pump.
Ideally, the waste handling facility is located at a higher elevation than that of
the main treatment plant and the treated water is recycled back to the head of the
main plant by gravity. If this scheme cannot be adopted, recycling pumps must be
mstalled. Yet the recycling pumps should be used only as a last resort, because the
double-pumping scheme tends to have "jogging" problems from the transfer pump
to the recycling pump-repeated starting and stopping of the pump at frequent in-
tervals.
HANDLING FACILITY FOR WASTE WASH WATER 377
RECYCLING OF THE WASTE WITHOUT A PURIFICATION PROCESS This alternative has been
applied to many existing plants and has proved to be effective provided that the re-
cycling rate and other factors are correctly chosen. This is especially true for con-
ventional treatment processes that must produce large and easily settleable tloc. There
is an abundance of clear evidence which conlirms that recirculation of performed noc
remarkably improves both nocculation and sedimentation; the majority of companies
that design and manufacture reactor-clarifier units are exploiting floc recirculation.
The key design guides are provided by bench-scale tests. The test results obtained
from a study conducted by the author are presented in Figure 4.2-2. The wastewater
used in the experiment was not the supernatant but the well-stirred turbid wash wa-
ter with a turbidity of 210 ntu (see Figure 4.2-1). The test results demonstrate that
both nocculation and sedimentation were improved dramatically as the amount of
wash waste added approached 30% of the total water volume. However, in actual
treatment plants the practical blending ratio of unsettled waste wash water is up to
10%, due to the effect of hydraulic shock loading and the limited detention time of
the units. Fresh settled tloc (sludge) also improves the settling efficiency if the blend
volume ratio is no more than 2%. However, it should be remembered that the dan-
ger associated with sludge recycling lies in the potential for recycling objectionable
taste, odor, and other undesirable substances. Moreover, sludge that is over 2 to 3
days old often loses its effectiveness completely (Figure 4.2-3).
In summary, the waste wash water should be stored in a holding tank in the man-
ner described previously and should be recycled, without clarification, at a blend-
ing rate that is 5 to 10% of the plant innow rate. It is always a good practice to dis-
infect the recycling waste.
Design Criteria The design criteria for proprietary reactor-clarifiers and regular
horizontal now rectangular tanks are not discussed fully in this section because they
were covered in Sections 3.2.3 through 3.2.5. However, the criteria for the alterna-
tive system of lagooning are presented, and the flocculation and sedimentation
processes and disinfection facility are described brieny.
Lagooning
Number of lagoons A minimum of two to satisfy the cleaning requirement.
Size of each lagoon Large enough to hold the wastewater of a minimum of 10
filter backwashes.
Shape of the lagoon Free form or rectangular. The length/width ratio should
be 4: I or greater.
Depth of the lagoon A minimum of 10 ft (3 m).
Inlet Isolation valve and an energy dissipator.
Outlet Should function in decanting, draining, and overnow
protection.
Miscellaneous The sidewall should be protected from erosion and
checked for seepage periodically. If the area is open to
the public, provide a safety fenCing around the lagoons.
378 SUBORDINATE PLANT FACILITIES HANDLING FACILITY FOR WASTE WASH WATER 379

5 Condition Turbidity Color pH Alkalinity

.Note: -Fresh Wash Waste and
No Settling Allowed Raw water @ 100% raw water 6 7.5 6.R 16.5
(reservoir)
S
l-
4
@ Fresh wash waste 230 220 6.65 27
e © 12 h old waste
~ 3 Filter wash waste @ 1 day old waste
:!l ® 2 day old waste
.c
~

® 4 day old waste

F
'iO 2
:::I

.
'0
'iii
a:
Nole: Experimental conditions
waste to raw water.
@ through ® were evaluated using 10% addition or "aged" wash

0 8
4 6 8 10 12

~
Alum Dosage mg/L 7

~\
Raw Water 1. (Turb. 11, pH 7.0, Alk. 25)

\""\
,,'.\•..".: .
.... \ .
\

~
5 \'.

- o '\ '~.:~
'.-,"
~ '"'''''''::'::'-
::;) \.
..
- -
I-
Z 4 C , " ' .... ~,'.

~ "'" ---.-...-= .... -

.~
:!l
.c 3 I B ........ ,.,......~-
F
~

'iO O~~~L-~~--~~~~~~~
o 2 4 6 8 10 12 14 16 18 20
--
:::I
.,..
:!l 2 Alum dosage (mglL)

a: Figure 4.2·3 Settled water turbidity as a function of wash-waste admixture age.

Flocculation and Sedimentation Processes Despite the high turbidity of the

Note: 8 mg/l Alum Dosage
waste wash water, the waste is easily tlocculated and clarified. The design criteria
O+-----.-----r----.-----r----~--~ of each unit process are fundamentally the same as presented in Sections 3.2.3
o 1 2 3
through 3.2.5. If the site area for the waste reclamation facility is very limited, a
Percent (Volume) of Sludge Densadag reactor·c1arifier should be evaluated seriously. This unit, manufactured by
Settled Floc (sludge) of the Waste Infileo Dregremont, Inc., provides a surfacc loading rate as high as 5 to 6 gpm/ft'
(After 24 hour Settling)
(12.5 to 15 m/h) for the c1arificr and is capable of producing an alum sludge com·
Figure 4.2·2 Improvement of flocculation by wash-waste mixing. posed of 2 to 3% solids by means of a rather high polymcr dosage. The most rc·
cent supcrhigh rate clarification unit, ACTIFLO by KrUger, can produce settlcd wa·
ter turbidities of I ntu or less with about 0.5 to 2 mg/L of cationic polymer at a
 HANDLING FACILITY FOR WASTE WASH WATER 361
360 SUBORDINATE PLANT FACILITIES

TRANSFER PUMPS
surface loading rale of 20 10 25 gpmlfl2 (50 10 62.5 m/h). Consequenlly, Ihe fool-
print of the unit is very small. Type Vertical turbine pump
Number Three idenlical pumps (one pump as a standby)
Disinfection Facility Chlorinalion is Ihe prime candidale for Ihe disinfeclion of Capacily 1000 gpm (3.8 m' /min) each
recyclable waler. Since Ihe recycling rale is generally only 2 10 3% of Ihe planl flow
rale, Ihe THMs produced by chlorinalion should nol be a problem unless Ihe raw TREATMENT I'ROCESSES
water characteristics pose a threat. Refer 10 Section 3.2.9 for the design criteria for
Tolal capacily 2000 gpm (7.6 m'/mi n)
Ihe di sinfeclion facilily.
Flash mixing 10 in. (25.4 cm) in-line stalic mixer
Two idenlical barned channel lypeS: round-Ihe-end Iype
Example Design Calculations Flocculation tank s
Size of each lank 20 fl (W) X 20 ft (L) X 10 ft (SWD)
Exa mple 1 or 6. 1 m X 6.1 m X 3.05 m
Number of channels Four channels wilh a 100ai of 16 barnes for each lank
Givell The filters in Example 2 of Seclion 3.2.7 arc pari of a 50-mgd filtralion Volume of each 4000 ft' (113m') or 30,000 gal
plant. The treatment plant requires a waste wash water trealment facili ty and are· An average of 30 s- I- tapered mixing
G
cycling facilily.
SEDIM ENTATION TANKS
Determine
Number Two idenlical lanks wil h lube sellier modules
Size of each 20 ft (W) X 35 ft (L) X 10 ft (SWD)
(i) The Iype of wasle reclamalion facilily
or 6.1 m X 10.7 m X 3 m
(ii) The Irealmenl process
Tube sellier 500 ft2 (46.5 m 2) of each lank is covered by Ihe
(iii) The basic design crileria for each unil process settlers
Surface loading 2 gpm/ft 2 (5 m/h) for Ihe area covered by Ihe
Solulioll (i) A conventional pretreatment process should be selec ted si nce the treat-
sellicrs
ment plant is located in a good residential area and the cost of the land docs not
Launders Two launders per lank, each 25 ft (7.6 m) long and
permit the plant to purchase large extra parcels of properly. Furthermore, the resi- a loading rale of 10 gpmllinear fool (7.5 m'lm • h)
dents will probably nol lolerale having a sludge lagoon in Iheir neighborhood, due
Sludge colleclor One chain-and-flighl Iype for each lank
to aesthetic and health considerations and the potential mosquito problem.
(ii) The treatment process should have a waste holding tank located ncar the filter TarAL VOLUM E OF TH E WASTE WASH WATER
bank, waste transfer pumps on the holding tank, and conventional pretreatment units at
an elevation that is several fcct higher than the water level at the head of the main plant. (22 X 960 X 7) + (3 X 960 X 3) ; 156,480 gal/wash
The clarified water should be chlorinated and transported to the head of the plant via
gravity flow. The head of the plant should have a flow meter in the discharge pipe line. Since the water source is a large lake, the potential for algal ?Ioom exists. Therefore,
(iii) The six basic design criteria for the treatment process units are presented be· the worst scenario would be that each filter is washed three tllnes a day. The ave~agc
low. wash cycle is assumed 10 be 24 h for Ihe filters. The waste holding lank should be SIzed
to hold three backwashes because two to three filters may become c1~gged at the ~ame
DATA ON HIE FILTERS
time, requiring the filters to be washed consecutively. nlC total reqUired volume IS
2
Total number of filters Six, each wilh a filter bed area of 960 fl2 (89 m )
Backwash condition 22 gpm/ft2 (55 m/h) for 7 min (collservalive) 156,480 X 3 ; 469,440 gal or 62,800 ft3 (1780 m')
Surface wash condition 3 gpmlfl2 (7.5 m/h) for 3 min
The lank should be cylindrical in shape wilh a maximum waler deplh of 12.5 ft
WASTE WASH WATER HOLDING TANK
(3.8 m). The diameler of Ihe lank is
Number One lank wilh a 62,800-ft' (1780-m') or 470,OOO-gal capacily
Size 80 fl (24.4 m) in diameler wilh a waler deplh of 12.5 ft (3.8 m) 10 ; (62,800 -i- 12.5)0.5; 80 ft (24.4 m)
D 0.785
hold Ihree filter washings
 382 SUBORDINATE PLANT FACILITIES HANDLING FACILITY FOR WASTE WASH WATER 383

Under the worst conditions, a total of 18 filter washes will be performed per day;
thus 24 h -;- 18 = 1.33 h or an average of I h 20 min between Ii Iter washings, In
+ Inlet
Transfer pump (typical)
other words, the total volume of wash waste produced by one lilter washing must
be pumped out within 1.33 h. The required capacity of the transfer pump is

156,480 -;- 80 = 1956 gpm (7.4 m 3 /min)

Make this 2000 gpm or 2.88 mgd (7.6 m 3/min).

Under normal conditions, each filter is washed once a day. Since the raw water is
obtained from a large lake, it is assumed that the treatment mode of the plant (for Sump
the majority of the year) is a direct filtration simulation process. In this case, the
Section A-A
average time between the filter washing is 24 -;- 6 = 4 h. The required recycling
pump capacity is therefore la) Ib)

['_~...:T.::ube:: ,el'.:;tI.:.ce'':''_"1
156,480 -;- 240 = 652 gpm (2.47 m'/min) i' 25'

However, the pumping capacity should be roughly 50% larger (1000 gpm) since
(-r)f -)
(') ~i- (-
-'--
~
I
.. %,<,< <"
'. :-:~};~;.;'~>,
'/- ~;:

.»,;:
T
~
plant operators generally prefer to wash as many lilters as possible during the day -I- - - ) I
shift; the graveyard shift has only the absolute minimum number of operators, and From
holding-
c::" 1-) I;: '--} <; T .' -.. ~ ~;~-:.('..;>///~::~
". ,·~,;·~~1~;::'~~-~-;~)~/. -
head
To
of the

\LH\-.-I) J..I
in many cases it is only one or two persons. tank plant
I
Provide three identical transfer pumps, each with a capacity of 1000 gpm
3
(1.44 mgd or 3.8 m /min), at the center of the circular tank. One of these pumps
should act as a standby. The total design capacity of the treatment facility should
SI;de
gate
~~i-
Ily pical) -
\
=..L....
.
~
r) I -
['
;, ;>:({;
.... ,.....,.
'/;'

~
i<l
l
3
be 2000 gpm (7.6 m /min). Therefore, provide two parallel units, each consisting
of a 1000-gpm process train.
Iflocculation tank I Sedlmentat..,o.::n...:l.:.an"k'-_~"l
20'-~'-f1~'-----35'-
launder
(typical)
(c)
Since the recycling rate is rather constant, that is, only one pump is operated re-

..~
gardless of the water level in the holding tank, a bamed channel tlocculation process
od I V-Notch plate on launder

-~'-
is the most cost-effective method for producing good tloc; it is also cost-elfective
to build and operate. An alternative to the baflled channels is the vertical shaft tloc-
culator; refer to Section 3.2.4 for further information.
Each sedimentation tank should be designed with both a tube settler module and
chain-and-tlight collector units. The tube settler modules are used for their space-
saving feature and their effectiveness. The blending ratio of the recycled wash waste
J I~I~II j:>:J~~ Id)
should always remain below 10% of the anticipated normal operational conditions. Figure 4.2-4 Waste-wash~water treatment process: (a) plan of holding tank~ (b) section
The blending ratio is generally 5 to 8%; thus there is little concern for hydraulic A-A; (e) plan; (d) longitudinal section.
shock loading to the main process units. Figure 4.2-4 illustrates the general features
of the waste holding tank and the waste treatment processes.
of media during the washing process. The operator should then modify the filter
Operational Issues An issue that is important but often neglected is the mon- washing conditions or adjust the height of the wash troughs. AccumulatIOn of filter
itoring of filter media accumulation is the wastewater holding tank. Plant operators media should be detected at an early stage because the media are very abraSive and
should drain the tank every 6 months and estimate the volume of accumulated sand will erode the impeller of the transfer pumps. .
and/or anthracite coal at the bottom of the tank. If the operator finds an excessive With regard to the treatment of recycled water, the coagulant d?sage of the malll
amount of media, it is clearly evident that the filters are losing a significant amount process would will not need to be adjusted if the blending ratIO IS less than 10%,

384 SUBORDINATE PLANT FACILITIES
regardless of the turbidity of the recycled water. The turbid form of wash waste con-
tains a great deal of floc, and this generally improves the efficiency of the floccu-
lation and sedimentation processes due to the seeding effect (Figure 4.2-2).
Moreover, the dosages of other water treatment chemicals usually do not have to be
readjusted as long as their blending ratios remain under 10%. Operation and main-
tenance of the pumps and the flocculation and sedimentation units do not require
special operator considerations. Refer to Sections 3.2.4, 3.2.5, and 4.7 for further
information on the operation and maintenance aspects.
BIBLIOGRAPHY
Kawamura, S., "Considerations on Improving Flocculation," J. AWWA, 68:6:328 (June 1976).
Kawamura, S., "Effect of Filter Wash Waste Recycling on Both Alum Flocculation and Cost
of Plant Operation," l. lpn. WWA. 228:28 (September 1958).
4.3 SLUDGE HANDLING AND DISPOSAL
Purpose The sludge produced by water treatment plants is categorized as chem-
ical sludge and consists primarily of inert materials. It is a by-product of the water
treatment process, and until the 1980s, treatment plants returned it to the water source
downstream of the intake. Currently, federal and state legislation classify most types
of sludge as pollutants and prohibit their disposal into the public water supply.
The majority of water treatment processes employ aluminum sulfate as a coag-
ulant, and sludge is therefore commonly referred to as alum sludge. Alum sludge is
diflicult to dewater, and dried sludge can revert to slurry when wet and disturbed.
due to its thixotropic nature. In this section we present some design guides for cre-
ating the most cost-effective and feasible sludge treatment (conditioning, thicken-
ing. and dewatering) and disposal methods.
Considerations The major considerations facing every design engineer are
(I) the Water Pollution Control Act, (2) the estimation of the sludge production rate,
(3) characteristics of the sludge, (4) methods of minimizing the sludge production
rate, (5) dewatering methods, and (6) the ultimate disposal of the sludge. In this sec-
tion we address the first five issues and present several sludge handling methods.
Water Pollution Control Act Public Law 92-500, thc Water Pollution Control Act
Amendment of 1972, categorizes sludge from water treatment plants as an indus-
trial waste requiring compliance with the provisions stated in this act. Under the
NPDES (National Pollutant Discharge Elimination System) provisions, a permit
must be obtained to discharge materials such as sludge and waste wash water. Wide
variations are found in the limitations of the permit among the various states. The
parameters include the pH, BOD, turbidity, temperature, flow of discharge, and the
total concentration of the following substances: suspended solids, settleable solids.
iron, manganese, aluminum, chlorides, sulfate, residual chlorine, floating solids, and
SLUDGE HANDLING AND DISPOSAL 385
foam. According to this amendment, no utility is allowed to discharge waste after
1985, unless it can show good cause for not recycling the treated wastewater. Thus
engineers must grapple with the problem of the ultimate disposal of sludge.
Estimation of the Sludge Production Rate It is rather dimcult to estimate the
sludge production rate since there are no concrete methods for predicting tur~i~ity
spikes and the chemical dosages required to treat them .. However, several empIrIcal
formulas may be employed. It is recommended that engIneers select parameters that
are similar to the operational conditions and use the basic formulas, such as those
shown below. [Note: In the formulas, 1.3 is the ratio between total suspended solids
(mg/L) and turbidity (ntu). This ratio ranges from 1.0 to 2.0.]
Dry alum sludge production rate (Ib per million gallons of water treated)
= [alum dosage (mg/L) X (0.26 X 8.34)]
+ [raw water turbidity (ntu) X U X 8.34]
Dry felTic sulfate sludge production rate (Ib per million gallons of water treated)
= [ferric sulfate dosage (mg/L) X (0.54 X 8.34)]
+ [raw water turbidity (ntu) X U X 8.34]
Dry ferric chloride sludge production rate (Ib per million gallons of water trcated)
= [ferric chloride dosage (mg/L) X (0.66 X 8.34)]
+ [raw water turbidity (ntu) X 1.3 X 8.34]
NOTES:
1. The equations above are based on the following chemical equations:
AI,( S0 4)' . 14H,O + 3Ca(HCO,h .-, 3CaS04 + 2AI(OH), + 6C0 2 + 14H,O
Fe2(SO,), + 3Ca (HCO,h .-, 3CaSO, + 2Fe (OH), + 6CO,
2FeCI, + 3Ca (HCO,), .-, 3CaCI, + 2Fe (OH), + 6CO,
2. According to the EPA Guidance Manual, Enhanced Coagulation alld Preci.pitate Softe/l~flg
(PB95-I00392), three waters of hydration are chemically bound to aiumlllum hydrOXide
in the residuals when the enhanced coagulation process is used. This manual shows that
the dry alum sludge production rate for each million gallons of water treated by the en-
hanced coagulation process can be established by the equalion
8.34 X [alum dosage (mglL) + TOC (mglL) + (1.5 X turbidity (ntu) + A]
where A represents additional chemical used (mg/L).
3. The sludge production rate is highly dependent on raw water quality, which is controlled
by weather conditions each ycar. Therefore, marc elaborate formulas for sludge produc-
tion rate do not mean that they arc better.
4. Annual sludge production rate should be computed using average tigures of turbidity, alum
dosage, and plant flow rate except for an expected high turbidity month.
 386 SUBORDINATE PLANT FACILITIES
SLUDGE HANDLING AND DISPOSAL 387

. In Ihe case ofsludge produced by lime softening, Ihe rale of sludge Produclio
IS largely a funcllon of Ihe hardness of Ihe raw waler and Ihe hardness level f hn 20-25 Semisolid Pasty 10 semisolid
softened waler. The basic formula for delermining Ihe lime sludge prod r 0 I e 25-30 Soft solid Soft solid to pasty
(pounds per million gallons of waler Irealed) is uc Ion rate 30-50 Crumbly cake Pasty to crumbly cake

Reduction of Sludge Production Rate and Volume To minimize capital

8.34 X (2.OCa + 2.6Mg + SS)
where Ca is Ihe calcium hardness removed as CaCO (mglL) M Ih . rate may be cut by 30 to 80% from the amount produced by a conventional com-
hardness removed as CaCO, (mglL), and SS the s~'spended' sol7ds ~n ~:~n~~~m
(mglL). In general, 2 to 3 mglL of solids will be produced for each mglL of har~~
~::::removCd. Thus average hme softening production rate may be estimated as fol-
dry lime sludge production rate (lb per million gallon of waler trealed)
~ 2.5 X 8.34 X tOlal hardness removed (mglL as CaCO,)
;Iudge CharacteristiCs The characteristics of waler treatment sludge differ
rom.locallon to 10callOn because they are largely dependent on Ihe raw water char
~cterIStlcs, the type and amount of coagulant, and other types of chemical aids use~
In trealIng the raw water.. In general, dried alum sludge is composed of 30% alu-
mInum and 50% Inert solIds such as sillica and calcium. A typical sample of alum
sludge contaInS Ihe componenls lisled in Table 4 3-1 Th I . I
rIme-so flene d sludge are presented in Table 4.3-2. .. . e yplca componenls of
The characteristics of lime sludge vary as a funclion of magnesium h droxide
content. The drYIng characterislics of lime sludge are similar in nature 10~hose of
~lum sludge. The concentrallon of magnesium hydroxide may range from' . 'f
Icant 10 30% by weight. Lime sludge containing low levels of magnesium I~S~~~:
~de may be dewatered 10 50 to 60% solids through Ihe use of drying beds Ye; Ihese
Igures WIll be reduced to 20 and 25% solids if the sludge has higher conc'entrations
of magnesIUm hydrOXIde. The characteristics of sludge are imporlanl to know be-
cause Ihey greatly Influence the methods of handling and disposaL For exam Ie,
semldned .alum a~d fernc sludges are Ihixotropic; the consislency of Ihe slu~ e
changes wIth phYSIcal force. ThIS characterislic makes alum and iron sludge p!_
IIcularly alum sludge, very difficult to handle. ' sludge may be minimized through effective dewatering techniques and by-product
Thedconsistency of drying sludge differs primarily due to Ihe amount of metal
hydroXI es and Ihe applied poly
'. mer (coagu I ) dosage. The general expression of
ant
s Iudge consistency IS as follows:
Solids Conlenl Sludge Consislency
(% of Solids) Alum Sludge
Lime Sludge
0-10 Liquid
10-15 Liquid
Viscous liquid
15-20 Viscous liquid
Pasty Pasty
expenditures and the operational costs of sludge handling and disposal, the first step
to be considered is reduction in the sludge production rate. The sludge production
plete Irealment process using alum (coagulant) if either of two melhods is imple-
menled: simultaneous use of polymer and alum (as coagulant) or adopting a direct
fillration type process to replace the conventional complete Irealment process. These
methods should be invesligated prior to design work.
The actual data from many operational water Irealment plants demonslrate that
if 0.1 to 1.5 mglL of cationic polymer is applied in conjunction with alum, the alum
dosage can be reduced by 30 to 50%. The combined application of 0.2 to 0.3 mglL
of anionic or nonionic polymer and alum often improves flocculation dramatically,
to the extent that alum dosage may be reduced by 15 to 30%. In the case of lime
softening. the use of caustic soda may decrease the production of calcium carbon-
ate sludge by 35 to 50%. However, the cost of caustic soda often outweighs the ben-
efits associates with a decrease in sludge production.
The generation of sludge is reduced most dramatically when the conventional
complete treatment process is replaced by two-stage, direct, or in-line filtration.
These processes may be used to treat raw water obtained from large lakes or reser-
voirs and require only 4 to 6 mglL of alum or the combined application of 2 to
4 mgIL of alum and 0.5 to I mglL of cationic polymer. In contrast, 12 to 16 mgIL of
alum is required by the conventional complete process for the same raw water. Thus
it is apparent that the sludge production rate may be reduced to levels that are only
10 to 20% of the conventional process. If the treatment plant is forced to adopt a
conventional process due to occasional turbidity spikes or algal bloom, the plant de-
sign could possibly have a bypass channel across the basins so that plant operators
have the ability to treat the water in direct filtration or in-line filtration mode for
most of the year. The second step in minimizing the costs of the sludge handling
and disposal processes is to decrease the volume of the sludge. The volume of the
recovery.
It is important for engineers to select an appropriate type of clarifier and sludge
handling facility. For example, the composition of alum sludge in rectangular sed-
imentation tanks is generally 0.3 to 1.5% solids. However, the composition is only
TABLE 4.3-1 Alum Sludge Characteristics
BOD COD Total Solids AI,O, Si02 and Inert Organics
(mgIL)
(mgIL) pH (%) (%) (%) (%)
30-300 30-5000 &-8 0.1-4 15-40 35-70 15-25
 388 SUBORDINATE PLANT FACILITIES
TABLE 4.3-2 Lime Sludge Characteristics
CaCO, Ca(OH), Mg(OH),
(%) Si0 2 and fne;
(%) (%)
(%)
85 93 oI 0.5-8 2-5
0.2 to 0.5% if a sludge blanket type of clarifier is used. Furthermore, basins lack_
m~ a contl~UOUS mechanical sludge removal unit usually have the sludge accumu_
latIng to heights of 10 to 12 ft (3 to 3.5 m) before the basins are cleaned. Thus the
sludge IS compressed to the extent that it is composed of 4 to 6% solids where
I d f d' . , as
s u ge .f?ll se l~entatlOn tanks containing mechanical sludge removal units has a
compositIOn rangIng from 0.5 to 1.5% solids, depending on the type of sludge re-
moval unit and the scraping cycle.
Under n~rmal condHions, approximately 0.1 to 0.5% of the water treated at a Wa-
ter punficatlOn plant Will be wasted by the sludge withdrawal operation unless re-
cycled. Therelation~hip between the suspended solid content (mg/L) and the mea-
sured turbidity (ntu) IS a function ~fthe type, size, shape, and color of the suspended
solids. Thus a defimte relatIOnship cannot be derived. The ratio of the total sus-
pended solids to the turbidity usually ranges from I to 2, with 1.3 being an average
figure. The solids content of hme softened sludge is much greater than for alum
sludge; It ranges from I to 10%, depending on the type of c1arilier unit. Sludge re-
moved from c1anfiers may be 3 to 4% solids, and 0.3 to 1.5% of the lime-softened
water is used in the sludge withdrawal cycles.
Sludge Handling A certain level of handling and treatment is required prior to
ultlmat.e disposal of the sludge. The d~gree of handling depends on the requirements
govermng ultimate disposal. Mechamcal dewatering systems may be used in situa-
tions where drying beds cannot be constructed, due to the lack of land or lack of cap-
Ital, frequ~nt raInfall, and/or a high required solids content. These systems include
gravl~Y thIckenmg, vacuum titters, tilter presses, centrifuges, and freeze-and-thaw
techniques. Yet in regions where Ihe wealher is mild to hot and evaporation exceeds
raInfall, the Simplest and most cost-effective methods are sludge lagoon drying beds
and the shallow-depth sludge drying beds. Properly designed and operated, these beds
are capable of drYIng both alum and lime sludge containing 45 to 50% solids.
The drying time for sludge drying beds may be reduced drastically if the beds
are aerated continuously at a rate of I to 2 ft'/min per square foot (18.3 to 36.5 m/h)
for the first 3 weeks of the drying cycle, until cracks appear on the surface of the
sludge. One Japanese equipment manufacturer has demonstrated that the drying time
for alum sludge may be shortened from 4 months (using ordinary weather drying)
to 6 weeks through the implementation of sludge aeration. Rather than using com-
pressed air, a vacu~m may be applied to the sludge beds. However, this practice
causes the underdraInage system to clog frequently. Similarly, sand drying beds may
cut the drymg lime of both alum and lime sludge significantly; however, in many
cases the beds must be reconstructed every 2 to 3 years, due to clogging.
SLUDGE HANDLING AND DISPOSAL 389
In locations where severely cold temperatures are predominant for one-third of the
year, freezing is a very effective method of sludge dewatering. Freez.ing can produce
alum sludge that is 25% solids, and the sludge IS no longer thixotropIc m nature. The
application rate of the sludge is approximately 2.5 Ib/ft2 (12 kg/m 2), and the wet sludge
is usually kept below a depth of 12 in. (0.3 m). A large-scale mechanical freeze-and-
thaw facility was built in Chiba [Kashiwai W.T.P., 150 mgd (6.6 m'/s)] and Sendai,
Japan, to treat alum sludge. However, widespread use of this facility has been limited
by the high initial cost and high operational cost ($100 per ton).
Gravity Thickening Gravity thickening may be attained through the use of la-
goons or mechanical thickening tanks, provided that a polymer is added to the liq-
uid sludge. Alum sludge treated with 2 to 3 mglL of cationic polymer can produce
a sludge containing 3 to 5% solids through lagooning. Lime conditioning of alum
sludge yields a sludge composition of 6 to 9% solids when processed by a thick-
ener tank at a feed rate of 0.07 to 0.14 gpm/ft2 (0.18 to 0.35 mlh) and approximately
5 to 10 Ib of solids per square foot of thickener surface per day (25 to 50 kg/m2 .
d). Lime sludge may be thickened to a greater degree than alum through the use of
thickeners; a loading rate of 30 to 60 Ib of solids per square foot of thickener sur-
face area per day (147 to 294 kg/m2 . d) may yield lime sludge that is 15 to 35%
solids. For alum or ferric hydroxide sludge, the standard solid loading rate for sludge
thickener is 10 Ib of solids per square foot per day (49 kg/m2 . d ) or 0.12 to 0.15
gpm/ft2 (0.3 to 0.38 mlh) hydraulic loading.
Centrifugation Among the several types of commercially available units, the most ap-
plicable units are the scroll discharge, solid bowl, and basket bowl. Basket centrifuges
produce alum sludge that is 10 to 11% solids without the use of polymers. However,
the sludge composition can be improved to 15% solids if the sludge is preconditioned
with polymer [4 to 5 Ib per ton of dry solids (1.8 to 2.3 kg/ton)]. Optimum precondi-
tioning of alum sludge calls for the application of I to 2 Ib of polymer per ton (0.45 to
0.9 kg/ton) of solids and results in a 25 to 30% cake. Bowl centrifugation allows lime-
softened sludge to be more easily dewatered than alum sludge. This technique produces
a cake that is 30 to 70% solids with or without polymer preconditioning. Yet the disad-
vantage is that it may require a high level of maintenance and has a high operating cost.
Drum Dehydrator (Dehydrum) The drum dehydrator is a proprietary unit that
has been used successfully in Japan. It transforms the sludge into pellets that are 20
to 35% solids, provided that polymer is the sludge conditioner. The shape and mo-
tion of this unit resemble those of a concrete mixer truck. Mechanically, the system
is simple and therefore has the lowest operation and maintenance costs of all me-
chanical dewatering systems.
Filter Press (Pressure Filtration) The filter press may be used to dewater both
alum and lime sludge. However, it is used to dewater alum sludge that has been pre-
conditioned with polymer or lime or precoated with diatomaceous earth. The solid
content of the cake is generally 25 to 50%, with a pressing cycle of 8 h. The life
 SLUDGE HANDLING AND DISPOSAL 391
390 SUBORDINATE PLANT FACILITIES

cycle of the Ii Iter cloth is normally 1.5 years. The filter press is a batch process that TABLE 4.3-4 Example or Sludge Dewatering Cost Analysis: Alum Sludge
Filter Sand
requires a considerable outlay of capital and has a high operational cost. Yet if the Belt Filter
Centrifuge Press Beds
sludge is required to be over 40% dry alum solids, this process is the most reliable Press
and quickest method available.
Capital costs (1990-2010) $ 905,000
$ 905.000 $ 905,000 $ 905,000
1. Thickeners 291,000
Belt Press The belt press is a relatively small unit requiring minimal attention and 291,000 291,000 291,000
2. Storage tank 133,000
maintenance once the system is optimized. It has been used to process coagulant 133,000 133.000 133,000
3. Sludge pumps 3,874,000
sludges and has yielded alum cakes that are 20 to 25% solids as long as polymer is 1,421,000 2,167,000 6,035,000
4. Sludge dewatering 64,000
added to the feed sludge. The Skinner Filtration Plant operated by the MWD of 64,000 64.000 64.000
5. Chemical feed 150,000
150,000 150,000 150,000
Southern California has been using two sets of belt presses (Winkle press: Andritz 6. Site work
$3,710,000 $ 7.578,000 $ 5,417,000
Model SMX-8-2m belt) with a processing capacity of 600 to 800 Iblh (270 to 360 Subtotal $2,964,000
kg/h) of alum sludge (dry solids base) for the past sixteen years. This plant precon- 758.000 542,000
10% Engineering 296,000 371,000
ditions the alum sludge, which is approximately 3% solids, with 2.5 to 3.2 Ib (1.1 to 928,000 1,894,000 1,354,000
25% Contingency 741,000
1.5 kg) of cationic polymer (Telacol) for every ton of dry solids. The product (cake) $ 7,313,000
$4.001,000 $5.009,000 $10.230,000
is 22 to 25% solids. The system requires only one operator for a period of about 2 Total capital
to 3 h/day. The belt cloth costs $2000 per unit and lasts approximately 2000 operat- 962,000 4,556,000 3,499,000
Year 2000 expansion 1,278,000
ing hours before any major rcpair work is required. Other treatment p1ants on the 306,000 1,452.000 1,115,000
Present worth 407,000
East Coast of the United States report the belt press as producing cakes that are 15 $5,315.000 $11,682,000 $ 8,428,000
Capital cost present worth $4,408.000
to 20% solids, requiring 10 to 15 Ib of polymer per ton (4.5 to 6.8 kg/ton) of solids.

Vacuum Filters Vacuum filters are used primarily in processing lime sludge. The Operation and maintenance
typical yield is a cake that is 40 to 50% solids. For low-magnesium lime sludge, the costs 195,000
145,900 219.000 211.500
loading rate is 40 to 90 Ibltt' per hour (195 to 440 kgl m' . h), but it is only 10 to Annual cost, 1990-2000 240,000
162,500 274.000 229.500
20 Iblft' per hour (49 to 98 kgl m' . h) if the magnesium level is high. Vacuum fil- Annual cost. 2000-2010 2,154,000
1,551,000 2,420.000 2.768.000
ters are generally ineffective in processing alum sludge. O&M worth
Table 4.3-3 summarizes the typical performance of several dewatering processes. Total capital and O&M
$7.735,000 $14.450,000 $10,582,000
Tables 4.3-4 and 4.3-5 show cost comparisons among various sludge dewatering present worth $5,959,000
systems for alum sludge and lime sludge.
Note: Average sludge production rate, 17 tons/day; maximum sludge production rate; 32 tons/day.
Filtrate Disposal The liquid waste from all mechanical dewatering processes pose
a problem because of their chemical characteristics: They contain a high concen-
tration of polymer, alkali chemicals, aluminum, and organic compounds and a sig- Health Services in certain U.S. states discourage this practice because of uncertainty
nilicant level of heavy metals. Recycling the liquid waste back to the head of the about the effect on human health of trace metal and other unfavorable com.poun.ds.
plant, alone or combined with the filter wash waste, is not recommended because Possible methods of disposal include discharge into the sanitar~ sewer, punficatlOn
it will adversely affect the flocculation process and will increase significantly the treatment, and discharge to sludge drying beds if they are on Stte.
level of unfavorable compounds in the processing water. Moreover, Departments of . I' d' I
Ultimate Disposal Four practical options are aVailable for the u lima:e ISposa
of sludge: (I) direct discharge to the sewer system, (2) hauling to a landitll, (3) on-
TARLE 4.3-3 Solids Content (%) of Sludge by the Dewatering Processes
site disposal, and (4) recovery of by-products.
Gravity Drying Filter Belt
Thickening Beds Centrifugation Press Press Freezing Discharge to the Sewer System The easiest solution for the treatment plant is
Alum 2-5 30-50 10-25 25-50 20-25 25-30 direct discharge of the sludge to a sanitary sewer. However, t~e waste .sludge flow,
sludge must be equalized prior to discharge because ordinary sewer ~Ines are Incapable ot
Lime 8-12 40-60 30-70 40-50 handling the flow rate (directly from the basins) and the solid content can be too
sludge thick. The equalized sludge may be pumped from the holding tank to the sewer line
392 SUBORDINATE PLANT FACILITIES SLUDGE HANDLING AND DISPOSAL 393

TABLE 4.3·5 Example of Sludge Dewatering Cost Analysis: Lime Sludge ment plants are under the same administrative jurisdiction, a connection fee will be
imposed for the direct discharge of sludge to the sewer system. Following is an ex-
Belt Filter Vacuum Sand
ample of the fee structure used in southern California in 1986:
Press Centrifuge Filter Beds
:'. 1',
Capital costs (1990-2010) fee = $375 X [0.0034 (gpd) + 0.115 (ppd COD) + 0.3595 (ppd SS)]
1. Thickeners $ 540,000 $ 540,000 $ 540,000 $ 540,000
2. Storage tank 274,000 274,000 274,000 274,000 where gpd is the sludge flow (gal/day), ppd COD the pounds per day of COD, and
3. Sludge pumps 143,000 143,000 143,000 143,000
ppd SS the pounds per day of suspended solids. I'
4. Sludge dewatering 1,421,000 2,283,000 1,008,000 801,000 I
In addition to the initial connection fee, the sewer authority will assess an an- , :1
5. Chemical feed 75,000 75,000 75,000
6. Site work 150,000 150,000 150,000 150,000
nual surcharge based on the established unit costs. An example is as follows:
Subtotal $2,603.000 $3,465.000 $2,115,000 $1,983,000 $200
cost per year per million gallons discharge
10% Engineering 260.000 346,000 212,000 198,000 cost per year per 1000 Ib of COD 12
25% Contingency 651,000 866,000 529,000 496,000 cost per year per 1000 Ib of SS 33
Total capital $3,514,000 $4,677,000 $2,856,000 $2,677,000 . ',1
In contrast to the difficulties encountered in the United States (except in Chicago
Year 2000 expansion 726,000 786,000 395,000 742,000 and a few other cities), the discharge of sludge to sanitary sewers is a popular method , I
'·'1:

Present worth 232,000 251.000 127,000 237,000 of disposal in Great Britain and the Greater Paris area in France. It has been claimed
I '

Capital cost present worth $3,746,000 $4,928,000 $2,983,000 $ 2,914,000 that this practice has very little effect on the wastewater treatment processes or on
subsequent sludge digestion.
Operation and maintenance
costs
Hauling Sludge to Landfills Selection of a sanitary landfill that is near a city is
Annual cost, 1990-2000 145.900 202,600 302,000 106,000 becoming increasingly difficult. In recent years, residents have become more con-
Annual cost, 2000-2010 154,700 243,600 369,000 119,000 cerned about environmental issues and issues that may affect land value. To further
O&M present worth 1,529,000 2,203,000 3,315,000 1,114,000 complicate matters, many existing sanitary landfills are reluctant to accept liquid and
semiliquid sludge in order to comply with the federal landfill guideline PL 91-512
Total capital and O&M
(Solid Waste Disposal Act) and because of their operational inconvenience. The Solid
present worth $5,275,000 $7,131,000 $6,298,000 $4.028,000
Waste Disposal Act recommends that landfill sites accept water treatment sludge only
Note: Average sludge production rate, 50 tons/day; maximum sludge production rate; 61 tons/day.
if it does not contain free moisture. Free moisture is defined as liquid that will drain
freely, by gravity, from solid material. Currently, most landfill sites that have the bot-
tom and sidewalls lined with impervious material only accept sludge with a mini-
at a constant rate that does not flood the sewer system. In several cases, the dis- mum solids content of 20 to 25%; some sites that do not have the lining demand a
charge of waste into the sewer lines was restricted to night, due to the limited ca· solids content of over 40%, due to the potential contamination of ground waters. It is
pacity of the existing sewer system. The actual effect of this practice is usually not important to note that alum sludge containing less than 20% solids generally cannot
noticeable unless sludge from the water treatment plant exceeds the total suspended easily be handled by sludge handling equipment or hauled from the plant in trucks. ,
,

solids received by the wastewater treatment plant (in the form of sewerage) by more Landfills generally charge a disposal fee of $5 to $10 for each ton of dried sludge. I

than a few percent. On certain occasions, the alum sludge may even be beneficial If the sludge is located in a remote area, the hauling charge becomes significantly
to wastewater treatment processes; it may increase the removal of phosphorous or higher. Design engineers should also investigate aiternative landfill sites in case the ,'I,

may aid in the removal of suspended solids due to the aluminum hydroxide. Water
treatment plants using a high proportion of organic polymer as coagulant will yield
sludge that tends to be more biodegradable. The negative effect of alum sludge on
wastewater treatment plants is the reduction in methane gas (caused by excessive
amounts of inert materials such as aluminum hydroxide and clay materials) being
brought to the sludge digesters; increased levels of heavy metals in the sludge de·
ters the biological treatment processes. Unless both the water and wastewater treat-
nearest site becomes unavailable. If sludge hauling trucks must pass through re-
spectable residential areas or busy commercial zones, the option of using landfills
may be eliminated by public hearings. Environmental impact studies may also pre-
vent implementation of this alternative.
On-Site Disposal On·site disposal may be considered as an alternative method
of sludge disposal only if the treatment plant has an adequate site area, has an old

'll
, '1I '1
I~
394 SUBORDINATE PLANT FACILITIES SLUDGE HANDLING AND DISPOSAL 395

quarry on the premises, produces very small amounts of sludge, or has a cornbina~ 5. Sludge containing 4 to 5% solids at the beginning of a drying cycle will shrink
tion of these conditions. If a plant is a candidate for on-site disposal, the following to approximately 15% of the original volume after drying; the dried sludge
Issues must be evaluated carefully prior to design work: the potential for contami_ will thus be 45% solids.
nating the linished water and/or groundwater, and the aesthetics and safety of the 6. The initial depth of the sludge, prior to drying, has a drastic effect on the length
facility. The disadvantages of on-site disposal may be illustrated by the following of the drying period. Thus the initial depth should not exceed 2 to 3 ft (0.6 to
example: ApphcatIon of alum sludge to the ground surface tends to clog the soil 0.9 m) if the drying cycle is established to be 3 to 4 months.
th~s preventing the passag~ of air to the roots of vegetation. However, lime SIUdg~ 7. The supernatant must be decanted continuously during the drying cycle to fa-
offers some advantages: Lime sludge has been shown to improve the condition of cilitate the formation of cracks on the surface of the drying sludge; these cracks
clay soil and neutralize acidic soil. will speed up the drying time by exposing more surface area for evaporation.
S. Figure 4.3-1 plots the unit volume weight (bulk density) of the alum sludge at
Recovery of By-products Acidic alum recovery was tirst tried on a large-scale various stages of the drying cycle. This chart should be used as a reference
basis in Japan: 15 plants were built between 1965 and 1972. Alum recovery can when designing a drying bed.
generally be attained at a rate of 80% or more, through sulfuric acid extraction at a
pH of 2.5. The recovered alum is nearly as effective as commercially available alum.
However, the quality of the alum often does not meet water treatment chemical stan- Lagoon Sludge Drying Bed
dards because other substances are also extracted during the acidification process:
iron, manganese, heavy metals, and a wide variety of organic compounds. Conse- Sizing guide The required net bed area for alum sludge is
quently, the alum recovered is used in conjunction with commercially available alum. estimated to be S Ib/ft2 (40 kg/m2) for wet regions
Due to the possible harmful effects of recycling and the accumulation of heavy met- and 16 Ib/ft2 (SO kg/m 2 ) for dry regions; these
als, this practice is now discouraged. The recovered by-products of alum sludge have ligures are based on dry sludge weight.
also been used in the production of low-grade bricks (in Brazil), as road stabilizer, Minimum number of Three, but four are preferred.
and as aggregates for low-grade road material. In Japan, alum sludge is used as a lagoons
raw material in the cement manufacturing industry, as a soil conditioner for certain Size of each lagoon Each lagoon should be capable of holding the
types of vegetation, and in high-grade bricks. amount of sludge produced (average) by the
In many locations, lime recovery is carried out by burning the lime sludge in treatment plant over a period of 3 to 4 months.
kilns. However, very few plants now practice this alternative because of recent con- Depth of the liquid A maximum depth of 6 ft (I.S m) during the lilling
cerns associated with air pollution and because of economic disadvantages. Lime sludge stage; however, a depth of 4 ft (1.2 m) is preferred.
sludge has been used successfully in the midwestern United States to neutralize Bottom slope A downward slope of 0.5 to I % toward the outlet.
acidic soil. Lime sludge may also be used to increase the porosity of tight soils, thus Shape of each lagoon Rectangular or free form with a minimum
allowing them to be more workable and productive for agricultural purposes. length/width ratio of 4 : I.
Underdrainage system This is an optional item. The system is composed of
Design Criteria The design criteria for (I) the properties of alum sludge, (2) la- perforated-pipe laterals with gravel packing and
goon drying beds, (3) sand drying beds, and (4) mechanical dewatering systems are I ft (0.3 m) of sand above the gravel bed. Note
now presented. that the underdrainage system tends to become
clogged and is often ripped away by the sludge
Properties of Alum Sludge removal equipment.
Lining The entire surface, including the bottom of the
I. The solids content of wet alum sludge obtained from ordinary rectangular sed- drying bed, should be covered by 2 to 3 in.
imentation tanks containing mechanical sludge collection systems is 0.5 to (5 to 7.5 cm) of gunite, asphalt cement, or soil
1.5%. cement to facilitate easy sludge removal (by
2. Alum sludge from sludge blanket clariners has a solids content of 0.1 to 0.5%. means of mechanical equipment), to prevent
3. Sludge lagoons are used primarily in gravity thickening. The solids content contamination of the groundwater, to provide
of sludge processed in this manner, with 3 to 4 months of continuous sludge protection from erosion, and for weed control.
filling and decantation operations, may be 3 to 4%. Distance between inlet A minimum of 100 ft (30 m) should be provided to
and outlet prevent flow short-circuiting.
4. A regular sludge thickener with stirring rakes can yield alum sludge that is 4
Inlet pipe Isolation valve and a simple energy dissipator.
to 6% solids with the addition of very little polymer.
 Bulk density of alum sludge SLUDGE HANDLING AND DISPOSAL 397
(empirical data)
1.0 1.1 1.2 1.3 1.4 Outlet Multilevel [9- to 12-in. or (23- to 30-cm) interval]
100
decantation facility, overflow protection, and
90 10 complete drainage of the lagoon.
80 20 Vehicle access A ramp should be provided for each bed to facilitate
~ ~
the removal of dried sludge.
~., 70 30 ~
., Access road A good access road, both to and around each lagoon,
"-,;
~
60 40 "-,;
~
should be included in the design.
'0 '0
Flow velocity in the pipe The tlow velocity of liquid sludge should be higher
~

C 50 50 ~
than 2.5 ft/s (0.75 m/s).
~ c
0 l'!
8 40 60 0
0
u Sand Drying Bed The design criteria for this type of drying bed are fundamen-
~ ~
~
30 70 tally the same as those listed for lagoon drying beds, with the exception of the fol-
1;;
'0
"~
lowing items:
:;;
20 80 '"
10 90 Depth of the liquid A maximum of 18 in. (45 cm), but 12 in. (30 cm) is
sludge preferred.
0 100 Perforated laterals with 6-in. (l5-cm) diameters and 10
60 70 75 80 85 90 95 100 (lbIff3) Underdrain laterals
Bulk density of alum sludge ft (3 m) on center.
(empirical data) Three layers with a total depth of 14 in. (36 cm):
Gravel bed
(a) 3-in. layer of ~-- to >in. gravel; 3-in. layer of ~- to
~-in. gravel; and 8-in. layer of }- to 1.5-in. gravel.
Sand bed Sand ranging from No. 10 to No. 30 mesh at a depth of
12 in.

NOTE:

Lime sludge is generally easier than alum sludge to dewater by the sand drying bed;
this method yields a sludge that is 25% solids and is therefore spadeable [weighs
120 Iblft' (1925 kg/m')].

Mechanical Dewatering Systems The belt press, tilter press, vacuum lilter, cen-
trifugation. and freeze-and-thaw processes are the mechanical dewatering systems that
arc applicable to water treatment. The first two units are mos1 frequently employed.
Since these systems are proprietary items, the project engineer need only determine
the functional specifications: the solids content of the dewatered sludge, floor space
requirements, and limitations on power and chemical use. A pilot study must be con-
ducted prior to writing the specifications: to confirm the effectiveness of the system
and to evaluate the type and required chemical dosage for preconditioning the sludge.

Example Design Calculations

Figure 4.3·1 Unit volume weight (bulk density): (a) alum sludge; (b) lagoon sludge; (e) Example 1
drymg alum sludge.
Givell A water treatment plant with a daily average tlow of 20 mgd (0.88 m'/s) is
396 located in southern California and requires a lagoon type of sludge drying bed. The
average turbidity of the raw water is 12 ntu and the average alum dosage is 10 mg/L.
398 SUBORDINATE PLANT FACILITIES SLUDGE HANDLING AND DISPOSAL 399

Determine In summary, provide four lagoons that have a net area of approximately 72,000
ft2 [6700 m' (1.7 acres)] and a depth of 6 ft (1.8 m). Other considerations, such as
(i) The required number of lagoons vehicle access, the building of dikes around the lagoons (sidewall slope of 2.5: I),
(ii) The depth of each lagoon and a vehicle access road that is lOft (3 m) wide at the top of the dikes, require the
land to be almost twice the net area required for the lagoons [3.5 acres (104 hall.
(iii) The size of each lagoon
Example 2
Solution By using the formula presented in the "Considerations" section, the sludge
production rate may be calculated: Given A IO-mgd (OA4-m 3/s) lime softening plant is located in the midwestern
United States. The plant requires that lime sludge drying beds be constructed. The
[(10 X 0.26 X 8.34) + (12 X 1.3 X 8.34)] X 20 average total hardness of the river water is 360 mg/L, with insignificant amounts of
= 3035 Ib (1380 kg) of dry sludge per day magnesium. The finished water hardness is, on average, 80 mg/L.

(i) Provide four lagoons with identical capacities. Each lagoon should be sized Determine
to hold sludge with an average of 4% solids for a duration of 3 months.
(i) The number of sludge drying beds
(ii) The maximum depth of the liquid sludge, during the filling stage, should be
set at an average of 6 ft (1.8 m). When the drying cycle begins, continuous de- (ii) The filling depth of the sludge
cantation should be initiated so that the depth of the sludge is reduced to 2 to 3 ft (iii) The size of each bed
(0.6 to 0.9 m) prior to the actual drying process.
In addition, provide a sketch of the sludge bed section
. (iii) The required net area of one lagoon may be estimated from the sizing guide
listed In the "Design Criteria" section. Since southern California is sunny for most Solution There are several general considerations that must be determined prior to
of the year, with a warm to hot climate, use a dry solid weight of 15 Ib/ft' (73 kg/m') making these four determinations. The following assumptions are made for this ex-
of reqUIred area to determine the required lagoon area. Based on a 3-month supply
ample.
of sludge, the average amount of dry solids is
I. For every I mg/L of hardness removed, 2.5 mg/L of dry solids (sludge) is pro-
(3 X 30) X 3035 = 273,150 Ib (124,160 kg)
duced.
2. Lime sludge containing insignificant amounts of magnesium hydroxide is eas-
The required net area for the lagoon is therefore
ily filtered through the sand bed and dries faster than alum sludge.
273,150 Ib -0- 15 Ib/ft' = 18,210 fe (1700 m') 3. The drying beds will yield sludge that is 50% dry solids (80 Ib/ft3 or 1280
kg/m'). Sludge that is 25% solids contains 72 Ib of dry solid per cubic foot
of sludge (1156 kg/ m3); this sludge is spadeable and is also easily handled by
Thus, make the lagoon rectangular in shape, 60 ft (18.3 m) wide and 300 ft (91.5
m) long, although a free shape, such as a kidney bean shape, is aesthetically beUer. mechanical sludge removal equipment.
Check the calculations. Based on Figure 4.3-1, the volume of the 4% sludge is 4. Assume that the average daily flow is 67% of the plant capacity of 10 mgd.
.
approxImate Iy 63.8 Ib/fr.
1
Therefore, the total volume of the sludge may be calcu- 5. The hardness removed is averaged to be 360 - 80 = 280 mg/L.
lated to be
The average sludge production rate (dry sludge) is 39,115 Ib/day (17.780 kg/d):
v = (273,150 -0- 0.04) -0- 63.8 = 107,034 ft3 (3030 m3)
[(2.5 X 280) X 8.34] X 6.7 = 39,115 Ib/day
Since the depth of the liquid sludge at the end of the filling cycle is 6 ft (1.8 m),
the area required for 136,480 ft 3 of 4% solids sludge is Now assume that the process unit yields a liquid lime sludge that is 5% solids with
a unit weight of 65 Ib/ft3 (1044 kg/m3). The volume of the 5% solids sludge pro-
107,034 -0- 6 = 17,840 ft' (1660 m') duced each day may be calculated as
3
which is essentially the same as 18,000 ft2 (1675 m') v = (39,115 -0- 0.05) -0- 65 = 12,035 ft 3/day (341 m /d)
400 SUBORDINATE PLANT FACILITIES SLUDGE HANDLING AND DISPOSAL 401

The volume of the 25% solids sludge is 39,115 7 72 = 543 ft'/day (15.4 m3 /d) f-_____ 8@40ft(10)=3201t.PIUSwallthickness-----.....J,11

-I
1
(i) A total of eight sludge beds is needed. Each bed should be able to hold the
25% solids sludge produced over an average of 15 days, so that a total of 4 months
storage is provided.
(ii) The maximum depth of accumulated sludge during the tilling cycle is 12 in.
I
l"""r«l~ <><-r"" i ..,... r"" ,... r« ! "'-~ r« ..... r"" I,.,. . r""

(0.3 m). I g
(iii) The size of each bed must be calculated as follows. The estimated volume
of the semi dried sludge (25% solids) for 15 days of storage is I =
I, I
v = 543 X 15 = 8145 ft3 (230 m 3) - I I
The depth of the bed is limited to I ft (0.3 m). If the width of each bed is desig-
nated to be 40 ft (12.2 m), the required length is
-Li..,... f-o< I 0<- f-... I ..... t- ..,...
>-1+ ..... f-I . . . f-o<
I
(8145 7 I) 7 40 = 203 ft, say, 200 ft (61 m) I I I I I
Therefore, the drying bed is 40 ft (W) X 200 ft (L) or 12.2 m (W) X 61 m (L) per
~
, +.....---6 in. Sludge line
sand bed. Figure 4.3-2 is a sketch of the designed sludge bed.
(a)

NOTES:
6 in. x 4 In. Reducer and
4 in. valve
1. The sludge beds tend to become clogged and the beds are easily disturbed by the heavy 6 in. Sludge pipe
equipment used during sludge removal. in. Elbow
2. The sludge lagoon drying bed is the most popular and practical method of sludge han-
dling.
3. The design criteria for the sludge lagoons are basically the same for lime sludge and alum
sludge. i·!
:'
4. A conventional lime sludge lagoon system has a minimum of three lagoons, each having
a depth of 10 ft (3m) and the capacity to hold nearly a I-year production of wet sludge. 6 in. PVC
The land requirement is approximately I acre per 100 mglL hardness removed (4050 m2 Underdrain pipe
per 100 mg/L) for each million gallons treated (3785 m\ A decantation facility is es-
sential for proper sludge drying. ~1.~ _______ 20·-_ _ _ _ _ _--1
1-________ 40 It (Width of the b e d ) - - - - - -
Operation and Maintenance Plant operators must pursue three major issues: (b)
(I) reduction of the sludge production rate, (2) withdrawal of highly concentrated
Figure 4.3-2 Lime sludge drying bed: (a) plan; (b) partial cross section.
sludge from the process, and (3) optimization of the sludge dewatering operation.

Reduction of the Sludge Production Rate To decrease the sludge production
rate, operators must select the appropriate coagulant, optimize the chemical dosage,
and possibly optimize the flocculation process. Since 26% of the added alum pro-
duces aluminum hydroxide (i.e., becomes part of the sludge), the primary goal of
the operator should be to reduce the alum dosage without degrading the tinished
water quality. The operator should first check to see if the polymers are being em-
ployed effectively. Since the cost of most cationic polymers is approximately six
times higher than alum, there will be no difference in the total chemical cost if I
part of polymer is used to replace 6 parts of alum. There are many case histories
which prove that a cationic polymer dosage of 0.5 mglL, together with alum, re-
duces the alum dosage by 3 to 5 mglL when compared to flocculation by alum alone.
The use of anionic and non ionic polymers as flocculation aids can also reduce the
402 SUBORDINATE PLANT FACILITIES
required alum dosage, as illustrated in Table A 10-1. In the case of direct or two_
stage filtration, the use of cationic polymers virtually replaces alum as the coagu_
lant. Since there is a minimal amount of aluminum hydroxide in the sludge, there
is a reduction in sludge production and the sludge is more easily handled.
The adoption of a direct Jiltration process during periods of good raw water qual-
ity is an alternative method of drastically reducing sludge production. Many water
trcatment plants that draw their water supply from large reservoirs or lakes should be
able to implement this mode of treatment, but only at certain times of the year. Rather
than feeding 15 mg/L of alum to raw water containing 3 ntu of turbidity to form set-
tleable floc, the direct filtration process requires an alum dosage of only 6 mg/L, or
a combination of alum (3 mg/L) and cationic polymer (0.5 mgIL) may be used to co-
agulate the turbidity. The flocculation and filtration processes occur subsequently in
the Ii Iter bed. To reiterate, there are many actual case histories of successful treatment
of water by this mode; technical articles on this subject may be found elsewhere.
The third item that the operator must be aware of is the effect of feeding lime
with alum at the head of thc plant. Many old tcxtbooks often illustrate this practice
without regard to the alkalinity and pH of the raw water, and a number of treatment
plants consider this practice as a given. Unless the alkalinity of the raw water is be-
low 15 mg/L and the required alum dosage exceeds 15 mg/L, there is no need to
add extra alkalinity. Flocculation in a pH range of 5 to 7 is actually ideal, except
for the potential for corrosion; however, the pH may be adjusted in the later stages.
If lime is not added with alum to the head of the plant, both residual alum in the
treated water and sludge production will be reduced. All cases described in this sec-
tion require bench-scale tests to be performed to verify and obtain the optimum
chemical dosage.
Withdrawal of Thick Sludge If the sludge delivered to the dcwatering system is
thick, the process is more cost-effective and more effective in concentrating the
solids, thus allowing the ultimate disposal operation to be more efficient. For ex-
ample, the operator may adjust the sludge collection cycle of the clarilier by ob-
serving the consistency of the liquid sludge, provided that the facility allows the op-
erator to do so. Excessive sludge withdrawal will result in an extremely low solids
content. Thus operators should adjust plant operations so that the sludge will have
a solids content of at least 0.5%; alum sludge obtained from the clarifIer preferably
has a solids content of 1%. A word of caution: Do not allow the sludge to remain
in the clarifier for a long period of time in an attempt to increase sludge concen-
tration, because the sludge may become septic and may impart an objectionable taste
and odor to the treated water.
Although it is common practice to discharge the sludge from the clarifiers to the
tilter waste-wash holding tank, it is highly discouraged for three reasons: (I) the
sludge becomes diluted, (2) unfavorable substances such as organic compounds, mi-
croorganisms, heavy metals, and taste- and odor-causing compounds may poten-
tially be recycled back to the treated water, and (3) the EPA's upcoming rule will
prohibit this practice. Sludge Irom the clarifiers should be directed to either the
sludge beds or the sewer system.
SLUDGE HANDLING AND DISPOSAL 403
Optimization of Sludge Thickening and Disposal Operations There are many
areas in which the plant operator can implement change; however, these depend on
the type of thickening system. Plant operators have the ability to optimize the sludge
loading rate to the thickening process, the thickness of the feed sludge, the belt
speed, and the pressure level to the Ii Iter press. However, the efficiency of the sludge
thickening process may be greatly improved by adding an appropriate type and
proper dosage of polymer to the liquid sludge, regardless of the .type of sludge be-
ing processed. The EPA's Filter Wash Waste Recyclmg Rule_ (ettectIve 111 the year
2003) prohibits discharge of sludge trom the clarIfiers to the hlter wash waste If the
water is recycled.
In regions where lime is readily available (at a reasonable cost) and where the
cost of personnel is also reasonable, the application of an additional 15% (or more)
of lime to alum sludge raises the pH of the sludge to 12. Experience has shown that
this practice dramatically improves the dewatering characteristics of the alum sludge.
BIBLIOGRAPHY
AWWA, Water Quality alld Trearlllellf, 3rd cd., McGraw~Hill, New York, 1971.
AWWA Committee Report, "Research Needs for Alum Sludge Discharge," 1. AWWA, 79:6:99
(J une t 987).
Bishop, M. M. et a!.. "Mechanical Dewatering of Alum Solids and Acidified Solids: An
EvaluaIion," 1. AWWA. 83:9:50 (SepIember 1991).
Bishop, M. M., et al., "Testing of Alum Recovery for Solids Reduction and Reuse," 1. AWWA,
79:6:76 (June t987).
Clark, D. K., and Anderson, C. A., "Shared Residual Processing Will Senetit Adjacent
Agencics,".I. AWWA. 89:6:48 (June 1997).
Cornwell, D. A., and Lee, R. G., "Water Stream Recycling: Effect on Water Quality," 1. AWWA,
86: II :50 (November 1994).
Cornwell, D. A., et al., "An Overview of Liquid Ion Exchange with Emphasis on Alum
Recovery," 1. AWWA, 71: t 2:74 (December 1979).
Culp. R. L" et aI., "Is Alum Sludge Advantageous in Wastewater Treatment?" Water Waste
El1g .• p. t6 (July t979).
Dulin, B. E., and Knocke, W. R., "The Impact of Incorporated Organic Matter on Dewatering
Characteristics of Aluminum Hydroxide Sludges," 1. AWWA, 81 :5:74 (May 1989).
Garabeck, R. 1., et a\., "Silvicultural Application of Alum Sludge," 1. AWWA, 79:6:84 (June
1987).
Geertsema, W. S., et al., "Long-Term Effects of Sludge Application to Land," 1. AWWA,
86: II :64 (November 1994).
Japan WWA, "Design Criteria on Handling Facilities for Water Treatment Wastes," 1. 11'11.
WWA. 489: 103 (June 1975).
Johnson, G., Buchanan, G. G., and Newkirk, D. D., "Optimizing Belt Press Dewatering at
the Skinner Filtration Plant;' 1. AWWA, 84: t I :47 (November 1992).
Knocke, W. R., "Effects of Coagulation on Sludge Thickening and Dewatering," 1. AWWA,
79:6:89 (June 1987).
404 SUBORDINATE PLANT FACILITIES
Koorse, S., ''The Role of Residuals Disposal Laws in Treatment Plant Design," 1. AWWA
85: 10:57 (October 1993).
Lucas, J. B., et a!., "Effects of Alum Sludge on Plant Growth," J. AWWA, 86: II :75 (November
1994).
Martel, J. c., "A Pilot Scale Study of Alum Sludge Dewatering in Freezing Bed," J. AWWA,
83:12:51 (December 1991).
Novak, 1. R.. and Bandak, N.. ''The Effect of Shear on Dewatering of Water Treatment
Residuals," J. AWWA, 86:11:84 (November 1994).
Novak, 1. T., et aL, "Use of Polymers for Chemical Sludge Dewatering on Sand Beds," J.
AWWA, 69:2:106 (February 1977).
Peters, G. H" Baumann, E. R, and Larson, M. A., "Effect of Various Parameters on the
Thickening or Softening of Plant Sludges," J. AWWA, 81:3:74 (March 1989).
Thompson, P. L., and Paulson. W. L., "Dewaterability of Alum and Ferric Coagulation
Sludges," J. AWWA, 90:4:164 (April 1998).
"Water Treatment Plant Sludges: An Update on the State of the Art," Parts I and 2, 1. AWWA,
70:9:498 and 70: 10:548 (September and October 1978).
Wilheim, J. T" et a!., "Freeze Treatment of Alum Sludge," J. AWWA, 68:6:312 (June 1976).
4.4 INTAKE SYSTEM
Purpose A water treatment plant ceases to function when the intake system fails
to supply water. If the plant is relying on a single intake system, malfunction of this
unit results in interruption of the water supply to the community. Therefore, the in-
take must be located in an easily accessible location and designed and built to sup-
ply a specified quantity of the best available quality of water; the source should be
reliable, that is, no interruption under any conditions. The intake system is often
combined with the grit chamber and raw water pumping station. However, in this
section we discuss only the intake system.
The intake structure is constructed at the water source for the sole purpose of ex-
tracting water for water treatment and water supply. The capacity of the intake should
be the maximum daily demand anticipated for the next 50 years; the permits re-
quired to build the intake and the cost of construction are both dimcult and costly
tasks. If an intake is twice the size required to satisfy present needs, it would cost
an additional 20 to 30%. However, a well-designed intake system, for middle-sized
to large plants, may cost 15 to 20% of the total cost of constructing the entire wa-
ter treatment plant.
Considerations The essential factors for the intake system arc reliability, safety,
and minimal operation and maintenance costs. Thus the design engineer must con-
duct extensive studies and use ingenuity in selecting the site and designing the in-
take works. Numerous studies must be performed prior to selection of the intake
site: water rights, the quality of the water source, climatic conditions, fluctuations
in the flow rate and water level, regulations set by departments of fish and game,
water navigation, geographical and geological information, and economic issues.
INTAKE SYSTEM 405
The intake may be located on a river, lake, or reservoir or may be designed to ex-
' tract groundwater. The basis for site selection varies according to the water source.
River Intakes Site selection for the intake system should be based on (I) obtain-
ing the best quality water through the implementation of procedures that avoid pol-
luting the water source; (2) forecasting possible changes in the flow and course of
the river; (3) minimizing the effects of icing, flooding, floating debris, river navi-
gation, and surges in flow; (4) providing easy access to the intake for maintenance
and repair; (5) providing adequate maneuvering space for vehicles; (6) allowing for
future additions to the facility; (7) maintaining a safe quantity of water flow during
periods of drought; (8) minimizing the effects of the facility on aquatic life; and
(9) obtaining the best or at least good, geological conditions.
If the intake site is located at a bend in the river, the outside bank of the curve
should be selected since it offers better conditions. The inside bank will probably be
troublesome because of the shallow water depth and the presence of sand bars and
accumulated debris. Most river intake structures are constructed as shore intakes: an
intake system and raw water pumping station combined. Yet several alternatives are
available; submerged intake, intake tower, suspended intake, siphon intake, floating
intake, and movable intake. The last two types are worth evaluating if the water level
of the river bed tends to fluctuate widely, if there is a frequent flooding problem, and
if the intake capacity is rather small-less than 5 mgd (0.2 m'/s).
The design of the intake system should include multilevel intake ports and bar
screens or racks. Additionally, engineers must provide protection against ice, float-
ing debris, boats, and barges, and the design must meet the requirements established
by the department of fish and game. The infiltration galleries or pipes are a unique
type of intake system. This system is a gravel-packed gallery containing many open-
ings (perforated pipes). It may be constructed under the river bed parallel to the
shore or in a gravel-packed well located on the river bed. The advantages are that
it avoids problems associated with ice formation, and water quality is substantially
better than that of the river water itself. The disadvantages are the potential reduc-
tion in yield due to clogging by the sand and gravel layer, and its limited capacity.
Lake and Reservoir Intakes Tower intakes and submerged intakes are two typ-
ical types of lake and reservoir intakes. Submerged intakes are generally less ex-
pensive than intake towers to build. These units also have no aesthetic impact and
do not interfere with river navigation. Submerged inlets that are properly designed
and located will have minimal problems with ice as long as the water depth exceeds
20 to 30 ft (6 to 9 m). The characteristics of tower intakes are described in the
"Design Criteria" section below.
An alternative to a tower and submerged intakes is a shore shaft. The shore shaft
is the part of the intake system that connects to the submerged inlet structure by
means of a pipe or conduit. It functions as a screen chamber as well as a pump suc-
tion well. Thus the shore shaft should be located at a depth that allows for the safe
withdrawal of the required maximum flow rate when the lake or reservoir is at its
lowest level.
 INTAKE SYSTEM 407
406 SUBORDINATE PLANT FACILITIES

Site selection for the lake and reservoir intake system should be based on the
considerations listed for river intakes and the following items: (I) the stratification -- ./ Removable hatch

~ Access bridge
characteristics of the water due to seasonal changes in water temperature; (2) the
water quality of minor streams, algal growth characteristics, and the growth cycle
-
of other microorganisms; (3) shore conditions, wind direction, and velocity; (4) the
conditions of the catchment area, including potential sources of pollution; (5) con-
ditions conducive to siltation in the reservoir; (6) the ultimate purpose of the reser_
Bar rack __ •
voir-multipurpose, such as hydroelectric power, irrigation, water supply, recre_ (IYPlcal) ~
ation-and the impact of joint users on the inlet location; (7) the level of sanctioned
recreational activities and the degree to which body contact sports will be allowed;
Pile
and (8) the flood level. In addition to these considerations, the design engineer must Verlicalturbine pump
also contend with issues such as a cooperative design arrangement with the joint tal
users (if the intake is for multipurpose use), the possible incorporation of an artifi-
cial water mixing device to destratify the water in the intake design, and special ice-
breaking measures (air agitation) at the intake location for cold weather regions,
Refer to Figures 4.4-1 and 4.4-2 for various types of intake structures. Traveling screen ___ ,_L..r-;---.

Groundwater Intakes The term groundwater includes water in artesian wells, HWl~
deep infiltration galleries or pipes, shallow wells, and deep wells. In all cases, the .~-
quality and quantity of the water should be investigated by a test well or other meth-
ods. The first step of the design project is to lind a reliable and suitable aquifier to
function as the source of groundwater, The initial study must also investigate (I) Bar rack
"- ",
-- Pump

the relationship between the level of the nearby river and the groundwater level of
the site during the dry season; (2) the quality of the groundwater during the dry sea-
son if the wells are close to the seashore; (3) the quality, yield, and geological con- tb)
dition of wells and infiltration galleries in the immediate vicinity; (4) the tempera-
ture and quality of the water and if artesian wells and springs are targeted as the Vertical turbine pump Removable skylight hatch
potential sources of water, the yield; and (5) the effect of industrial wastes and the
Vacuum pump
possible seepage of wastewater (nitrates) into the aquifer. Site selection for ground-
water intakes must also consider the potential intrusion of seawater, the effects of
having the groundwater level being drawn down by a cluster of wells, and the po- HWl
tential migration of pollutants from remote dump sites harboring industrial toxic
wastes,
After a well site has been selected, a test well should be drilled and pumping
tests must be performed to obtain a safe yield from the test well and to check the
quality of the well water. A safe yield is assumed to be approximately 60 to 70%
of the maximum yield obtained during the pumping test. Since the quality of the
groundwater is generally very good, all conventional treatment processes except Ie)
chlorination may be eliminated. Thus the design of the water supply well must be Various types of intake structures: (a) intake tower; (b) shore intake; (c) siphon
Figure 4.4-1
in strict accordance with all requirements dictated by the regulatory agencies,
well intake,

Design Criteria In this section we present the design criteria for the intake tower,
shore intake, intake crib, intake pipe, and infiltration gallery.
408 SUBORDINATE PLANT FACILITIES INTAKE SYSTEM 409

Intake Tower

Location As close to the shore as possible but at a minimum water

Vertical turbine pump
depth of 10 ft (3 m), with the exception of small intakes.
Shape and size The standard cross section of the tower should be circular
or oval. In the case of a river intake, the major axis of the
oval must be parallel to the flow direction. The top of the
tower should be at least 5 ft (1.5 m) above the highest
HWl.&, water level and the access bridge to the tower must also
have the same clearance. The interior diameter of the
lWl tower must be large enough to be able to install and
., service the intake gates and pumps (if pumps are
installed).
Structure The materials used to construct the tower should be strong
and durable, such as reinforced concrete, and should be
built on a solid foundation so that it may resist the force
of !looding.
Intake ports Gated intake ports must be provided at various depths; the
lowest of the ports should be approximately 2 ft (0.6 m)
above the bottom. The vertical interval of the ports should
be 10- to 15-ft (3 to 4.5 m). The !low velocity through the
Gross area of the ports, at the same elevation, should
generally not exceed I ft/s (0.3 m/s). In regions where the
hazards of icing are to be expected, a velocity of 0.5 ft/s
(0.15 m/s) or less is desirable. Sluice gates or butterfly
valves must be specified for the intake ports on the
(b)
interior or exterior of the tower. The gates are used for
isolation, but butterfly valves are preferred since the sluice
gates have a tendency to freeze up if they are not used on
Intake screen a regular basis.
Antivortex plate
Bar screen Bar screens (racks) should be provided at each port exterior
to the valve. The bars should be composed of ~- to >in.
Intake pipe
(13- to 19-mm)-diameter steel spaced 2 to 3 in. (5 to
7.5 cm) on center. Under normal conditions, the now
velocity through the net opening area of the bar screens

- must not exceed 2 ft/s (0.6 m/s). In certain cases the

velocity is limited to under 0.35 ft/s (0. I m/s) by
fish and game authorities to protect small fish from
being sucked in.
Fine screen If screens are not installed downstream of the tower, it will
(c)
be necessary to install fine screens to remove small
Figure 4.4-2 Variolls types of intake: (a) suspended intake; (b) floating intake; (e) intake tloating objects and to protect fish. In most cases, the
crib.
clear openings of these screens should be approximately
I~ to ~ in. (5 to 9.5 mm), and the maximum tlow
velocity through the net openings in the screens should be
410 SUBORDINATE PLANT FACILITIES
no more Ihan 2 ft/s (0,6 m/s) al the maximum designed
n ow rate. The usc of an automatic hydraulic clean ing
device is hi ghl y recom mend ed. Moreover, in regions with
a very cold climate, the intake tower and sc reen s should be
protected from frazil and anchor ice formation. The two
melhods thai are ofte n used arc compressed air and steam
injection.
Shore Intake
Location Minimum waler deplh of 6 ft (1,8 m),
Type Typical shore inlake-siphon well, suspended, or !loaling
Iype----<lepending on Ihe silualion,
Structure This is delermined by Ihe type of inlake bUI is basically Ihe
same as for the intake tower.
In lake hay The cross sec tion of the intake openi ng should be capable
of accommodating a maximum Ilow velocity of 1.5 ft/s
(0.45 m/s), The !low velocily shou ld be reduced 10 below
I ft/s (0.3 m/s) if an unusual amounl of debris or icing is
10 be expeeled,
Bar sc reen These should be provided at the inlake bay and sci
approxi mately 60° from the horizo ntal. The bar should be
t-
composed of to !- in . steel spaced 2 to 3 in. (5 to
7.5 em) on center. Under norm al co nditions, the fl ow
velocily Ihrough Ihe screens should nol exceed 2 ft/s
(0.6 m/s), The bar scree n has a trapezoidal cross sec tion ,
which prevents the acc umulati on o f solids, with
reciprocating action, that is, a climbing screen inclined
80° to the horizontal.
Fine scree n Fine screens are often installed just downstream of the bar
scree n or may be installed in the grit c hamber. The clear
openin gs in the scree n must be I~ to ~ in. (5 to 9.5 mm),
and under norm al co nditions, the maxi mum fl ow
ve locity through the net opening shou ld occ ur at minimum
submergence. As with the tower intake, in regions with
cold weather, the fine scree ns must be protected from
ic ing ha zard s through the usc of co mpressed air or steam
injectio n.
Intake Crib
Location Deeper than 10 ft (3 m) from the surface and in a location
where it will not be buried by sed imen t, washed away, or
hampered by problems associated with ice formation.
Structure In areas where Ihe waler deplh exceeds 10 ft (3 m), Ihe lOp
INTAKE SYSTEM 411
of Ihe inlake should be 3 fI (I m) above Ihe river bOllom,
However, if Ihe waler deplh is less Ihan 10 ft, Ihe crib
musl be buried anywhere from I 10 3 fI (0.3 10 I m)
below Ihe ri ver bed, Cribs are usually polygonal and built
with a firm frame of lumber or reinforced concrete.
All sides of Ihe crib should be prolecled by riprap or
a concrete slab. Crib ports must be sized to provide a
maximum velocily of 0,25 100.5 ftls (0,08 100,15 m/s),
The c rib should surround a bellmoulhed pipe Ihal is
connected to the intake conduit.
Intake Pipe or Conduit
Size To minimi ze the accu mulation of sedim ents, the pipe
or conduit must be capable of providing a minimum
now vel ocily of 3 10 4 fils (0,9 10 1,2 m/s),
Protection If the pipe or co nduit crosses a ri ver or lake to co nnect with
Ihe shore shafl, it should have 3 10 4 fI (0,9 10 1,2 m) of
cover over the top of the crown. Additionally. crushed
rocks should be placed above the trench; a rule of thumb
is 10 use I yd' of crus hed roc k per linear fool of pipe
or conduil (2.5 n,'/m),
Slope To avoid creating an air blockage in the line, the pipe or
conduit must be laid on a co ntinuousl y rising or falling
grade.
Permit A permit must be obtained from the appropriate authorities
prior to the design phase if the intale pipe or co nduit is
designed to cross the bottom and bank of a river, lake, or
reservOIr.
Infiltration Gallery
Direction At right angle to the ri ver or parallel with the river flow,
depending on Ihe undern ow pallern, Ihe an lici paled
difficulties in co nstruct ion, and the cost of constructin g
Ihe gallery,
Deplh The com mon deplh is generally 15 ft (4,5 m) below the
boltom of Ihe river or lake, Yel Ihe aC lual deplh should
be delermined Ihrough hydrogeological slUdies,
Orifice size The diameter of each perforation is generally: to t in.
(101020 mm), The number of orifices is 2 or 3 per sq uare
fOOl (20 10 30 per square meier) of the colleclor surface,
Lenglh The lenglh of Ihe gallery is based on Ihe assumplion Ihat
the now ve locity through eac h orifice, at ma ximum
inlake n ow ra le, musl be 0, I ft/s (0,03 m/s),
 412 SUBORDINATE PLANT FACILITIES
Structure The gallery must be constructed with perforated reinforced_
concrete pipe or conduit. A stainless steel well strainer
may be used for small-scale installation.
Slope The gallery can be laid on level ground, but a slight slope of
500: I will minimize the problem of air blockage.
Velocity The flow velocity must not exceed 3 ft/s (0.9 m/s) at the
outlet of the gallery.
Backfilling The collector is packed with cobblestone, gravel, and sand
similar to that used in tilter bed gradation. The thickness
of each layer must be a minimum of 20 in. (0.5 m).
The trench of the gallery collector is backtilled with the
specified sand, or native sand may be laid on top of the
gravel and sand packing.
Junction well This feature must be provided at the junction points of the
gallery and at the end of the gallery line. The diameter of
the junction well must be a minimum of 3 ft (I m) to
facilitate the inspection of the gallery.
Shallow Wells and Deep Wells The design and construction of the wells are ad-
dressed in many books. Therefore, no attempt is made to discuss this subject (see
the Bibliography). The regulatory agencies and the Ten States Standards provide de-
tailed instructions on the design of the wells.
Operation and Maintenance Under normal conditions, a properly designed
and constructed intake system requires very little in the way of operation and main-
tenance efforts. The usual operational tasks include regular maintenance of mc-
chanical equipment-traveling screen, intake pumps, and appurtenances-and the
removal of large floatables collected by the bar racks. If the intake is provided with
multilevel ports, the valves of each port must be operated on a regular basis. There
have been too many cases where the sluice gates attached to various levels of in-
take ports became inoperable because they were not operated periodically for main-
tenance and severely corroded valve operating stems and the gates.
Floods are a potential problem for the intake system because logs and other large
objects may jam the intake bar racks. In regions of cold weather, ice floes driven
by the wind may potentially block the ports of the shore intake. Operators must keep
careful watch for these problems and appropriate actions must be taken before any
damage occurs to the intake or before there is a drastic reduction in the capacity.
After each tlood or earthquake it is vital to send a diver down to examine the con-
dition of the underwater intake structures to determine if the ports have become
blocked by sand migration or if the foundation of the intake has been eroded or
damaged.
When the intake is located in a large reservoir the plant operator and laboratory
personnel should periodically check the quality of water obtained from various
depths of the lake. If the quality is better at a particular level, the intake level should
be adjusted accordingly. This practice reduces the chemical cost of water treatment
GRIT CHAMBER 413
and saves a considerable amount of extra labor by avoiding conditions associated
with algal bloom and taste and odor problems.
In some cases, the raw water must be conveyed to the treatment plant over a long
distance by means of a pipeline or conduit. One of the problems associated with
this practice is the growth of slimy organisms or freshwater muscles in the interior
of the pipeline. The growth of these organisms may dramatically reduce the carry-
iog capacity of the raw water aqueduct. A common solution is to chlorinate the wa-
ter at the intake site, leaving approximately 0.5 mglL of free residual for 15 to 30
min. Chlorination may be continuous or a shock treatment, depending on the situ-
ation; operators must give special consideration to the potential for THM formation.
Regardless of the potential for THM formation, it is a good practice for engineers
to incorporate a chlorine feed system, including a chlorine diffuser, during the de-
sign phase. If no such facility is provided, plant operators should consider adding a
chlorination system before the situation becomes uncontrollable. In regions such
as the Great Lakes area, zebra mussel infestation has caused great problems.
Potassium permanganate is a good control chemical (dosage rate as low as 0.3 to
0.5 mglL) since there is not THM formation by this chemical.
BIBLIOGRAPHY
Anderson, K. E., Water Well Handbook, 4th ed., Missouri Water Well and Pump Contractors
Association, Belle, MO, 1971.
AWWA, Groundwater, AWWA Manual M2J. American Water Works Association, Denver,
CO, 1973.
Davis, C. v., cd., Handbook of Applied Hydraulics, McGraw-Hili, New York, 1969.
DHV Consulting Engineers, Shallow Wells, DHV, Anersfoot, The Netherlands, 1979.
Fair, G. M., et aI., Water alld Wastewater Eflgineerillg, Wiley, New York, 1966.
Gibson, U. P., Water Well Manllal, Premier Press, Berkeley, CA, 1971.
Japan WWA, Design Criteria for Watenvork Facilities, Japan Water Works Association,
Tokyo, 1969.
National Water Well Association, Design and Construction of Water Wells, Van Nostrand
Reinhold, New York, 1988.
National Water Well Association, Committee on Water Well Standards, Water Well
Specifications, Premier Press, Berkeley, CA, 1981.
Richardson, W. H" "Intake Construction for Large Lakes and Rivers," 1. AWWA, 61:8:365
(August 1969).
Urquhart, L. c., ed., Civil Engineering Handbook, McGraw-Hili, New York, 1959.
4,5 GRIT CHAMBER
Purpose Grit is defined as a combination of silt, sand, gravel, shells, and other
abrasive materials. The grit chamber is a plain sedimentation tank that removes grit
by simple gravity sedimentation. The purpose of this chamber is to protect moving
414 SUBORDINATE PLANT FACILITIES GR IT CHAMBER 415

mechanical equipm en t (s uch as pum ps and mixers) from the abrasive substa nces
and to preve nt the acc ulllulation of g rit in the raw water line and the pret reatment
processes, incl ud in g the ozone cont act ta nks and basins. The grit c hamber should
1.8
be sit uated upstream of the raw water pUlllpS and as close to the intake struct ure as
poss ible. T hi s arrangemen t prevents problems assoc iated with siltation and si mpli_
li es grit ha nd lin g. As described in Secti o n 4.4, the g rit c hamber and inlet structure 1.6
are o ften constructed at th e intake site, and sc ree ns are often install ed at the inlet of
the g rit c ham ber. A hydrauli c scale model co nducted by the author c learl y de mon-
strated that whe n the sc reen is installed properl y, it func ti ons as a n inlet diffuser g 1.4
wa ll. .2
Eng inee rs sho uld no te th at th e grit chambe rs fo r wate r treatm ent processes are i!
c 1.2
distin ctl y differe nt from those used in sewage trea tme nt in regard to o ne function. .2

the se ttlin g of undesirable compou nds. Sewage plants are des ig ned primarily to re-
i!
'E
~
move the large inert co mpound s o f suspended matter. Thus, most of the organ ic de- u
C 1.0
o
composa ble solids are permitted to flow throug h the tank . The tank design there- u
~

fore docs nol need to provide good fl ow characteri stics for se ttlin g and th e tan k is .~
;;
o rt en ae rated to produce a spiral flow pattern. For thi s reaso n it is obv ious that en- 0 .8
&
ginee rs should not usc the blue prints o f a sewagc treatme nt grit chamber whe n de-
signin g a wate r g rit c hamber. 0 .6
•
Considerations The fou r basic considerat ions in des ig ning a gr it c hamber are •
( I ) the locat io n of the tank , (2) th e number of req uired tanks, (3) the shape of each 0.4 •
tank , a nd (4) the size of the grit to be rcmoved. The tirst two issues are di sc ussed • •
o nl y briefly. The mos t des irable loca ti on of the grit chambe r is at the intake si te •
0.2
••
since thi s will ma xim ize the purpose o f the tank . A sin gle div ided tank or two sep- •
arate tanks are ge nera ll y used to meet th e minimulll operati onal requ iremen ts. This
arra ngeme nt all ows one tan k o r s ide to be drain ed for clean in g or repair wh ile al-
----------------------
Backgroun d level of t racer ......'Ooo-D-OO

low ing the pl ant to re main opera tio nal. In sit uat io ns whe re sand carryover is not a
major problcm , a sin gle lank with a bypass chann e l or pipe is co nsidered to be sat-
isfacto ry.
G rit c hambers arc ava ilab le in two basic shapes: recta ngul ar and square. The rec-
tangu lar tank is identica l to th e rectangu lar hori zo ntal now sed iment at io n tank. The
square tank , also kn ow n as a detritus tOllk, is commonl y used in the treatment of
sewage (sec Figure 3.2.5-9). T he square ta nk is sati sfac tory fo r the purpose of waste-
water treatmc nt but does not perform we ll as a wa ter treatment g rit chamber. The
main problem with the det ritu s lan k is a function of its ph ysica l characterist ics. The
sho rt tank le ngth and the uneve n inflow to the ta nk ca use severe fl ow short·
ci rcuit ing. Figu re 4.5- 1 illustrates the res ult of a trace r test using a sq uare tank. The
graph c learly demon stra tes the poor flo w character isti cs in the ta nk . To prov ide good
n ow c haracteri sti cs for g rit sCllling, the cha mbe r should be rec tang ul ar with a con·
tracted inlet (F ig ure 4.5-2).
If there is a sig nili ca nt amo unt o f grit carryover, the grit may be removed by
c hai n-and- Ili g ht o r trave ling bridge sc rapers since these will re move the g rit con-
tinuou sly. If these conditions do not ex ist or arc no t expec ted to occ ur, a g utter
shoul d be provided at th e ce nte r of the tank bottom (long itudina ll y) to faci litate man-
° OL----0~.5----1---~1.~5--~2~--~2~.5----~3--~
Relative time ratio tiT
Figure 4.5-1 Gri l basin Iracc r cur ves for a squi re lank .
ual clea ning of g ri t (by hosi ng) and remova l by a grit pump or a hyd rauli c eductor
syste m. T he lowe r porti o n of the tank may be lo ngitudinall y slo ped 100 : I toward
the s ump and transve rsely sloped I : 50, or Slee per, toward the ce nter gutt er. II
A sc ree n may be co nstruc ted at the downstream po rti o n of the inl et diffuse r wa ll
at an angle of 70 0 from the hori zo nt al. Since the sc ree n ac ts as a diffu se r wa ll and
the direc ti on o f fl ow Uu st past the sc ree n) is perpe ndicular to the sc ree n a ngle, an
angle of less than 70 0 wo uld direc t the Il o w to the tank holto m and scour the sed-
iment (Fig ure 4.5-2). It is also importa nt to provide isolati o n va lves at both ends of
the tank and to install a we ir type of effluc nt so Ih at a minimum wate r depth ca n
be ma inta ined in the ta nk. The d iffuscr wa ll should havc a to tal orilicc area o f ap-
proximate ly 15% of the c ross-sec ti o nal arc;] of the tank . T hese c riteria are based o n
the hydra uli c scale model tes t of a rectangu lar grit chamber co nductcd by the au-
thor to o ptimi ze the diffu ser wa ll des ig n and sc ree n angles (see the Bibli ography).
 416 SUBORDINATE PLANT FACILITIES
OPERATIONS BUILDING 417

""~>------38.0 m-------1 detritus tank, require periodic adjustment to maintain uniform inlet flow conditions.

~
Yet it is virtually impossible to create a good inflow condition based solely on the
adjustment of the diffuser vane. Plant operators must observe the conditions of the
grit chamber on a daily basis. However, more attention should be given to the in-
Screen
walkway
take and the grit chamber after heavy rainfall, and tasks such as trash removal, ad-
justments in the grit removing cycle, and the quality of the incoming raw water must
" Diffuser wall
(curved)
be monitored before the water is permitted to flow to the treatment processes. If the
Plan
Screen plant design does not include an automatic screen cleaning system, the plant oper-
ator may be quite busy removing debris from the screen. Plant operators should also
consider installing a turbidimeter in the grit tank for early detection of turbidity
spikes if the design has failed to provide one.
Regardless of the grit chamber conditions, the interior of the tanks should be in-
'Saffle wall spected on an annual basis during periods of low water demand. All required main-
longitudinal Section tenance should also be performed at this time. It is always good practice to collect
grit samples from the tank because sieve analysis of the samples provides operators
Figure 4.5-2 Grit chamber.
with valuable information on tank performance. An example design calculation of
a grit chamber design is shown in Example 2 of Section 3.2.5.

Generally, the grit to be removed is sand or silt with diameters exceeding 0.1 mm
or I.arger than No. 100 mesh (0.15 mm). Grit any smaller than this does not pose a BIBLIOGRAPHY
m~lor threat to the pumps and pipelines. Figure 4.5-2 is a grit chamber designed to
treat 35 mgd (1.5 m'/s) of flow and has been performing well since 1955; the Degremont, Water Treatment Handbook, 5th ed .. Halstead Press. New York. 1979.
design is based on the hydraulic scale model study.
Kawamura. S.. "Hydraulic Scale Model Simulation of the Sedimentation Process," 1. AWWA,
73:7:372 (July 1981).
Design Criteria
Metcalf & Eddy. Wastewater Ellgilleering: Treatment, Disposal, Reuse, 2nd ed .. McGraw-
Hill, New York, 1979.
Number of tanks Two tanks (one with a bypass line may be used as an Montgomery. J. M.. Consulting Engineers. Water Treatme1lf: Principles alld Design, Wiley.
alternative) New York, 1985.
Water depth 10 to 13 ft (3 to 4 m) with a grit remover; 11.5 to 16 ft
(3.5 to 5 m) without a grit remover
Length/width ratio Minimum of 4: I 4,6 OPERATIONS BUILDING
Length/water depth Minimum of 6: I
ratio Purpose The operations building of many of the older, small to midsized plants
Mean flow velocity 10 to 15 ftlmin (3 to 4.5 m/min) were designed with little attention to architectural and human considerations. More
Detention time 6 to 15 min emphasis should be placed on designing a functional and pleasing environment since
Surface loading 4 to 10 gpm/ft' (10 to 25 m/h) this building is considered to be the human-oriented structure of the entire water
treatment plant and often dictates the theme of all the surrounding structures.

NOTE: Considerations The three basic issues that must be addressed are the (I) func-
tional, (2) architectural, and (3) structural considerations. The issues range from
If a gravity flow system is planned, the highest water level in the tank must be lower than building configuration and exterior finish to the design loads.
the lowest water level at the intake.
Functional Issues The operations building should have an administrative zone,
Operation and Maintenance A properly designed grit chamber does not re- operational management zone, product quality control zone (laboratory), and a me-
quire much operator attention. The inlet diffuser vanes of square tanks, such as the chanical/workshop zone. Additionally, the chemical feed system is often included
 418 SUBORDINATE PLANT FACILITIES OPERATIONS BUILDING 419

in the operations building. The administrative zone is composed of a lobby with Environmentallssues This category includes the topography of the site, weather
'h .. an conditions, the type and the quality of the surrounding neighborhood, and the view
area lor t e receptIOnIst or secretary, a storage room for records and office suppJi
toilet facilities for visitors, a conference room, and offices for the water Producti~~ of the process units from the control room. Most of these data will have been col-
manager, chief operator, and supervisor. lected during the plant site evaluation phase.
. The ~one of operational management minimally contains the control room with
Its momtors and data loggers, a process control laboratory, and a lunchroom, The SPACE REQUIREMENTS The size, type, and number of rooms are strongly influenced
laboratory should at least have a jar tester, turbidimeter, pH meter, and a water sam. by owner preference and real needs. The limits on the occupancy rate affects the
piing sink. A large meeting room suitable for training operators and accommodatin layout of the office, process, and maintenance zones. Two areas within the admin-
a. large group (30 people) of visitors is preferably included in the design. By pro~ istrative zone-the toilet/locker room and the laboratory-require larger than aver-
vldmg an, accordion type doo~, part of this meeting room may be sectioned off for age space. Handicap toilet facilities and separate locker rooms for both men and
use by shift operat~rs and ~amtena~ce personnel who are stationed at the facility. women are mandated by law, and the need for a larger laboratory space is a result
. A lunchroom wIth vend~ng area IS an essential part of modern treatment plants, of the Safe Drinking Water Amendment.
for the morale and well-bemg of the plant personnel. Locker room facilities, with Based on past experience, mid-sized plants [30 to 60 mgd (1.3 to 2.6 m3/s ) ca-
showers, must also be provided for both men and women operators. Modern su- pacity 1 should have the following major rooms and functional sizes:
pervisory and control systems produce a large quantity of records, thus an adequate
ro?m must be pro:id~d for the storage of operational records and computer sup- Lobby/reception area 18ftX20ft 360 ft'
phes. The Safe Dnnkmg Water Act was recently amended to increase the number Control room 15 ft X 20ft 300 ft'
of water sampling and analysis items from the present 49 to 183 items by the year Conference room 15 ft X 20 ft 300 ft 2
2000. Based on this knowledge, all major treatment plants should have a rather large Offices
laboratory zone unless the water samples are sent to an outside laboratory. Superintendent 15ftXI5ft 225 ft 2
The laboratory should have four distinct areas based on their specialty: general Plant supervisor 15 ft X 15 ft 225 ft2
chemistry lab, instrumentation lab, bacteriological lab, and an oflice for the water Supervisor 12ftxI5ft 180 ft'
quality control superintendent. One of the laboratories must have a sampling sink Chemist 12f1X15ft 180 ft2
with sampling taps from the head of the plant. settled water, filtered water, finished Maintenance 12ftXI5ft 180 ft2
water, and other necessary phases of the processing water. A minimum total lab Lunchroom 12ftXI2ft 144 ft'
space (floor area) of 1500 to 2000 ft' (140 to 190 m') is required to process the Laboratory 25 ft X 50 fI 1250 ft2
drastic increase in future water analysis. (minimum)
The fourth zone in the operations building is the mechanical equipment room ToiletlIocker room 16 ft X 32 ft 512 ft2
and the machine (maintenance) shop. The function of this zone is to house the me- HVAC equipment room 15ftxI6ft 240 ft2
chanical equipment-the heating, air conditioning, and ventilation-necessary for Electrical equipment room 12 ft X 16 ft 192 ft2
the entire operations building and the workshop for the process equipment. Engineers Machine shop 25ftX30ft 750 ft2
must design one room in this zone with a separate air-conditioning system. This Warehouse 24 ft X 30 ft 720 ft'
room does not need to be large, generally 200 fe (19 01'), but it must have a dust- Chemical feed room 40ft X 62ft 2480 ft2
free environment so that it may be used as an electronics and computer workshop. Chlorine feed room 15 ft X 20ft 300 ft2
The chemical storage and feeding system may be designed as part of the opera- Chlorine storage room 15 ft X 44ft 660 ftz
tions building. It is usually located at the rear of the administrative building, behind 9018 ft 2 (840 m2)
the administrative and operational management zones for reasons of safety and traf-
fic control. The advantages of having a separate building for the chemical feed sys-
tems are improved and convenient access for delivery trucks, shorter chemical feed- NOTE: I m ~ 3.28 ft, I m2 ~ 10.76 ft2.
lines and suction pipelines for the liquid chemical metering pumps, better
housekeeping, better climate control, and safety-in case of a severe chemical leak. The other rooms should be sized according to the type of equipment, usage, and
the requirements set by the manufacturer.
Architectural Issues There are eight m'\ior architectural issues in the design of
the operations facility. Some of these are common sense. However, many are a func- CONFIGURATION OF THE ROOMS Factors such as the location of the visitor parking lot,
tion of owner preference and state and federal laws. location of plant facilities, and the view of the treatment plant often influence the con-
420 SUBORDINATE PLANT FACILITIES OPERATIONS BUILDING 421

tigurations of the rooms in the operations building. In many cases, the plant superin_ Toilet/locker rooms Mosaic tile Ceramic tile Painted gypsum
tendent and control room operators demand that their rooms have a view of the main board
tre~tment. pr?cess units or the plant entrance. Thus the following is a list of several Laboratory Sheet vinyl Painted gypsum Acoustic tile
design cntena that may be used to arrange the space within the operations building. board
Electric room Concrete Painted gypsum Painted gypsum
I. The control room should be adjacent to the lobby. This will allow operators board board
to monitor and control visitor entry. Process room Concrete Scaled masonry Acoustic tile or
painted gypsum
2. Ideally, the control room should be located in the interior of the building to
board
aVOid exposing computer equipment to the effects of direct sunlight and fluc-
tuations in temperature. Moreover, the computer terminals are easier to view
if they are shielded Ii'om window glare. BUILDING CODE The building code divides the operations building into an occu-
pancy zone and a factory zone. The codes also dictate certain design features based
3. Access to the conference room, offices, and so on, should be convenient for
on the occupancy rate and safety considerations; these considerations include:
employees.
4. Visitor access t.o the .'oilet f~c.i1ities must be simple. A separate toilet facility • Number of exits
should be provided for all VISitOrS, and this room should be kept clean at all
times. Fire-rated corridor walls
• Fire-rated doors and windows
S. Separate rooms should be provided for meetings and lunch breaks.
• Area separations
6. The lunchroom should have the same appliances and storage space as those
found in a small residential kitchen. • Fire-rated ceilings
7. If the treatment plant is located in a residential area, the operations building Flame spread classifications
should be a one-story structure, to minimize the visual impact to the neigh- • Smoke detectors
borhood. However, the height of the building is a function of the situation and
owner preference. The factory zone must be separated from the occupancy zone by a fire-rated wall
as required by the Uniform Building Code (UBC). The heating, ventilation, and air
EXTERIOR FINISH The exterior of the building should be attractive, durable, and re- conditioning room (HVAC) requires a tire-rated wall containing sound-insulating
quire minimal or no maintenance. The walls should be constructed of a material that material. Additional issues are the amount of insulation in the walls and ceiling.
may be used for both the interior and exterior of the building; concrete or masonry Refer to the UBC for the specific requirements related to each of these considera-
meets these requirements. Masonry is relatively inexpensive and is more attractive tions.
than concrctc. The roof should be pitched to minimize the potential for leaking and
also to create an architectural feature. Many options exist for roofing material: clay BUILDING ARRANGEMENT It is in the best interest of all parties if the building
tile, shingle, or metal. Finally, the windows and window wall units should be spec- arrangement issue is developed by a qualii1ed architect who is familiar with the
ified to be insulated glass. Engineers may even wish to stipulate that all rooms re- functions of a water treatment plant or an architect who is guided by a project en-
quiring climate control have tinted windows. gineer. Figure 4.6-1 is a sample building arrangement.

Structural Issues The structural system of the operations building is affected di-
INTERIOR FINISH The materials selected for the interior of the building should be
rectly by the occupancy groups in the building. Each occupancy group has a max-
both durable and attractive. However, owner preference often strongly influences
the selection. The following is a typical room finish schedule.
imum allowable floor area; refer to the UBC for the maximum allowable floor area
for the tive different types of construction. Note that the washdown and durable ma-
terials used in the construction of water treatment plants often give the operations
Floor Wall Ceiling
facility type II or III building characteristics.
Lobby and corridors Quarry tile Wallpaper Acoustic tile The goal of the stlUctural design for hydraulic structures is to produce strong
Control room Sheet vinyl or tile Wallpaper Acoustic tile units that will remain serviceable and leak-free for the life of the plant. This design
Ortices Carpeting or Wallpaper Acoustic tile should follow the recommendations of the UBC and ACI-350, COliC rete Sallitary
sheet vinyl Engineering Sfmctures. Regardless of the specific application, the design and de-
 422 SUBORDINATE PLANT FACILITIES
PUMPING SYSTEM 423

Design Loads

Office, laboratory, and control room 100 Iblft2 (490 kgf/m2)

Areas with process units and mechanical 150 lb/ft' (730 kgf/m')
equipment
Areas with heavy equipment Equipment weight
plus 200 Ib/ft2 (980 kgf/m2)
Electrical control room 250 Ib/ft2 (1220 kgf/m2)
Traffic loading HS-20 (AASHTO)
Roof loads (URe requirement) 20 Ib/ft2 or 97 kgf/m2 (basic)
Wind loads (URe requirement) 100 mph (161 kmlh)
Seismic loads Recommendation of a
geotechnical expert

4.7 PUMPING SYSTEM

Figure 4.6-] Building tloor plan.
tails of the structure must minimize cracking due to shrinkage and limit direct crack-
ing at construction joints and expansion--{!ontraction joints. Since geotechnical con-
ditions of the plant site also affect the structural design of the plant facilities, rec-
ommendations should be obtained from a geotechnical consultant in regard to the
subgrade and groundwater level of the property. Advice should also be obtained on
shoring, earth pressures, sheet piles, and so on.
Design Criteria In this section we present very general design criteria for the
materials and the design loads used in constructing an operations building. Specific
recommendations cannot be presented because adjustments must be made accord-
ing to the local and specific conditions of each plant site.
Construction Materials
Purpose Although pumping is not a water treatment process per se, it is a ma-
jor ancillary item. The water treatment design engineer is responsible for selecting
and specifying the appropriate type of pump, for setting the system requirements,
writing the pump specifications, and developing all other necessary information.
Since the pumping system is usually not the design engineer's area of expertise, he
or she is strongly advised to refer to the literature published by pump manufactur-
ers and to consult with mechanical engineers to obtain more detailed information
and data. This section presents only basic guidelines for pump selection, designing
a pumping facility, preparing pump specifications, and other special considerations.
Considerations Several items must be addressed when designing a pump sys-
tem. These items, in addition to those used in selecting pump drive units and de-
termining the number of pumps and standby generators, are covered briefly in this
section.
Type of Fluid The primary concerns that must be addressed during pump selec-
tion are both the chemical and physical characteristics of the fluid to be pumped.
The main characteristics that strongly influence pump selection and system design
are viscosity, temperature, density, corrosivity, chemical stability, lubricating prop-
eny, and the type and amount of suspended material (in water).

Concrete 4000 psi (280 kgf/cm') compressive strength System-Head Curve The design engineer should establish the system-head
Reinforcing steel ASTM A615, grade 60 curve by analyzing the pumping system; the preliminary design of the system should
Structural steel ASTMA36 include a pump layout, piping scheme, and instrumentation diagram. A clear and
Masonry ASTM e90, grade N-I, lightweight block, 1500 psi (105 accurate curve of the pumping equipment characteristics, in the expected range of
kgUcm') combined compressive strength operation, may be obtained from the manufacturer. Engineers should remember that
Aluminum Alloy 6061-T6 hydraulic head loss is a function of the flow rate and the physical characteristics of
Stainless steel Type 304 the piping system. Therefore, a clear piping scheme will reduce the chance of ad-
 424 SUBORDINATE PLANT FACILITIES
PUM PING SYSTE M 425

cii.l ion.al headl~ss ,and ultimately minimi ze the opermi onal cos t o f pumpin g. The de.
tails lor, es tab lishing the system- head curve may be fou nd in most engineering and
hydrau lic textbooks. Efficiency
_-/_ T"\
Potential System Modifications In the majorit y of cases, the pumping system / ..1-
?cslg.ncd dUring the initial deSign phase is later expanded or modified. lffuture mod. ----
Power I
ificatJon of Ih~ pumping station is required by the mas ter plan , the pumping system
Capacity Capacity Capacity
should b~ deslgne~ to accommodate Ihem. This will allow the ori ginal desig n 10 still Centrifugal Vertical turbine Axial-fkJw
be effecll ve and will nOI hinder exi sting plant operat ions during con stru cti on.
ttl I I I 1 I I I I I I I I I 1
Operational Modes Important considerations for desig nin g a pumping system ~ §~~§~ ~ ~ ~ ' ~ ~ ~ ~ ~ ~~ §
~re th e dcg~ce of ~ ow or head Iluctu ali oll and the mode of operation , such as Can-
11I1~~US or .1~terl11lttenl. Th~ o~~ra li o nal modes dic tate the number of pumps and Vatues of specific speed
-
'"
th ell capaciti es and deterlllllle II booster pumps are requ ired in a fl ow path .

~equired Margins. Engine~rs must spec ify a pumping unitlhat has adequate mar-
gills so I~a.t the eq ulpl11~ nt ~ 1I1 be ab le to withstand infrequenl , short-li ved, abnor-
mal ~on(htlOns: fluctuatIOn III press ure, dips in electri c vo ltage and frequ ency, hy- Radial-vane Francis-screw MiXed- flow
A
Axial-flow
Axis of
rotation

d~auhc surges, or the loss of cooling wa ter. With respect 10 small electri cal Figure 4.7-1 'TYpe c harac teri sti cs and specifi c speed sca le for vari olls impeller designs.
di sturban ces, the margins may be added to the pump head and ca pac ity. Yet the (Adapted, with permi ssion, fro m R. P. Walker, Pump Selectioll, Bu ttcrworth-Hei nc mann .
pumps should onl y be spec ilied to be 15 to 20% over the des ign points because Stoneham. MA. 1972.)
overS ized pumps may develop mechanical and hydrau lic problems.

Pump Selection Differen t. types of app lications require spec ilic types of pump.
~llJll~ selectIOn IS based on fac tors slIch as th e flu id charac terislics, turn -dow n ra-
tIO, disc harge press ure and system requ irements, ava ilab ility of space, layo ut, life,
energy and pump costs, code requirements, and the materials used in construct ion
. R~ci ~rocating pumps, such as the plunger or dia phragm Iype, should be selected
t?r hqllld chemical l11~t ering and injecti on applicati ons. These types of pumps de-
li ve r a constant capacity of liquid over a wide range of system- head vari ations.
However, the capac ity of these pumps is ra ther small and the pumps are charac ter-
ized by a pulsatin g fl ow.
. ~ositive-displacement pumps, also known as rotary pumps, may be L1sed for both
liqUid and gas pumping. Although this type of pUIllP does not ex hibit a pulsating
I~p~ of fl ow and may also be used as a slurry or metering pum p, its application is
limited 10 a low to medium pressu re range.
Centrifugal pumps arc lIsed for a wide va ri ety of hydra uli c head and over a wide
ra nge of capac ~ty l:equi.rel11enls. However, in water treatm ent applications they are
generall y used III situati ons req uiring low to medium capaci ty with medium to hi gh
press ure. These pumps require priming unl ess installed below the wate r surface in
a dry we ll .
Ax.ial now pUI11~s. are best suited for app licati ons req uirin g low hyd rauli c head
and hi gh-fl ow conditions. The mi xed now type is used in situati ons with interme-
diate conditions, between th e axia l now and centrifugal pump, Refer to Figure
4.7- 1 for the pump characteristics.
Vertical tu rbine pumps require much less space than the centrifugal type of pump
and are also self-priming. Yel this lype of pump generally requires more head room.
Thus, if the pumps are installed in a housing, a rem ovable skyli ght or hatch must
be provided directly above each pump to faci litate repairs. Nonetheless, vertical
pumps such as turbine pumps, axia l now pumps, and mi xed now pumps are often
installed without a housing. Rather than using a pump well constru cti on, verti cal
turbine pumps may use a "canned" sucti on Iype of design in the suction pipeline,
to help minimi ze the construction costs of a pumping stati on.
For furth er information of the various aspects of pumps, refer to the Hydra ulic
Ilistitllte Stalldards, 14th edition and the more updated vers ions published in 1997
and 1998 (sec the Bibliography).
Drive Selection The ava ilable energy source, the cost of operati on, and the various
local conditions all affect selection of the pump drive unit. Pumps may be driven by
a variety of means--electric motor, internal combusti on engine, or steam- and may
have either constant- or vari able-speed dri ves. The internal combustion engine can be
the drive unit due to local conditions and to avoid increasing the energy demand dur-
ing peak hours of electrical consumption. An electric motor may be selected as part
of the dual dri ve, to ac t as a redundancy drive. Yet the majority of pumps are driven
by squirrel-cage inducti on motors, due to their versatility and availability.
Other miscell aneous types include sy nchronous motors, used in app licati ons re-
quiring a hi gh horsepower rating and where a power factor correction is required;
 PUMPING SYSTEM 427
426 SUBORDINATE PLANT FACIUTIES

and standards of the materials to be used. Industrial standards such as

wound-rotor motors, which are used occasionally in applications requiring variable- ~N~~d:l~d' AST~ are good sources for basic requirements relating to metallurgy,
speed drives; and reduced-voltage starting and low-in-rush current motors. The last
two types may be used as high-power motors in situations where an initial transient dimensions, tolerance, and tlange detmls... . e roblems with se-
reduction in-line voltage would adversely affect other equipment. It is important to note that a numb.er 01 deSign cngm~ers ia~he~naJ'ority of cases,
. . t t e of metal for pump constructIOn. n .
lectlng the aPtropn~ e;;;h excellent results. However, if the plant is pumping anaer-
Number of Pumps and Standby Generators Reliability considerations often stamless stee IS use , . ut the bottom portion of deep lakes,
tend to favor the use of multiple-pump schemes, A three-pump system employing obic sludge or deep well water or pumpmg 0 b'c condi-
identical electric motor-driven pumps, each having a capacity that is 50% of the type 316 stainless steel will become severely pitted under t~~~c o~n;~~~ :hould bc
. Therefore other types of stamless steel, such as typ, e , bl f'
maximum demand, is commonly selected. This type of system requires only that 11ons. , f ern IS the pro em 0
the power supply run two of the pumps at anyone time and maintain the third pump sp ecified with manufacturer
,
assistance. Another areba 0 colnc A ' rule of' thUll1b
I t' h type ru toget ler, s a '
as a standby. If frequent power outages are anticipated, an engine-driven pump or abrasion when stamless stee sot e samc d'ff f 50 Bhn on the
the hardness of the two adjacent metals should have a I erence 0
standby diesel generator should be incorporated into the pumping system design, If
a standby generator is selected, the system will require voltage starting and the en- Brinell s~al~ of hardness, > II divided into two maior parts, the technical
gine generator should be capable of supplying a starting current that is at least three The bldchng documents are usua Y " T
specifications and gencral infor.mat~on (~ommercialterm). The techlllcal speci Ica-
times the normal full-load current. However, a smaller emergency generator will be
sufficient if sequence starting is used, since each pump may be started in a stepwise tions should include the followmg Items.
fashion.
The size of the pumping system is also an important factor in determining the I. Scope of work , ' h bolts
number of pumps, Smaller systems may opt to utilize two to three pumps of the , I d d' the document· for example, foundations, anc or .,
2. Work not mc u e 111 •
same capacity in order to make the pumping system more flexible. This scheme will labor . I' h r 'I tern
enable the system to meet the water demands, will facilitate easier maintenance, and Ratings and service conditions such as the characteristics 0 t e IqU~l, n c-
will allow interchangeability of parts. However, larger installations experiencing 3. erature of the liquid, flow rate, total dynamic head, preferred spee ra g ,
high peak demands would benefit tram a two-bank system. In this scheme, one bank ~fflciency at the design point, load conditions, NPSH, overpressure, runout,
of pumps (constant-speed drives) supplies the base tlow to satisfy the normal de- and operating conditions .
mand. The second bank supplies the extra tlow necessary to satisfy the peak de- Design and construction brief detailing the codes, sta~dard~, materml I~SCd~
mands. The pumps of the second bank may be driven by internal combustion en- 4. t e of casin ,stage arrangement, balancing, nozzle onentatlon, type 0 s~p
gines, such as diesel engines, so that the expensive peak power costs (including the ::rt, vents a~d drains, type of bearings, shaft seals, baseplates, connecting
potential penalty charge) may be avoided. pipes, and instruments
5 The lubricating oil system d I
Pump Specifications The two basic types of pump specifications are those con- ' Itaa>c power supply an regu a-
cerned with pump construction and those related to performance. The first specifies de- 6. Details of the drive unit, such as motor vO ' °d , t I'llal boxes and elec-
. nel re uirements wiring stan d ar s, erm, ',
d ;he type ai, fuel, number of cylinders, cooling syste~"
tails such as the type of design, the design method, and pump construction, The per- cal
lion, 10d pa
formance specifications dictate the performance standards for the pump and contain a tncal eVlces, an . r d haust mut1ler tor
speed governing, starting requII'ements, coup mgs, an ex .
few details related to pump construction, sllch as the pump construction material.
Since the owner is primarily responsible for performance of the pump, it is more the internal combustion drives
than likely that the plant owner, not the manufacturer, will be held legally respon- 7. Cleaning and painting
sible if pump malfunctions develop, provided that the specifications describe the 8. Performance testing
pump design and construction. Thus, most water treatment plant projects use the 9. Drawings and supporting data
pump performance specifications when constructing the pump system. The project
10. Special tools
engineer can always seek help from pump manufacturers when developing the c1ar~
11. Criteria for evaluation
ifying specification. However, the engineer should remember that each supplier is
partial to its own product. It is always a good practice to include a series of alter~ , I d h> and address of the owner, thc
natives, in addition to the basic specifications, because it allows pump suppliers to The general information should InC u e t e name . (d d'ltl'on of the plant
d . f ation on the locatIOn an con
present their best offers and provides the owner with the advantage of obtaining applicable codes and standar s, m orm . . t bid dates manufactur-
site, and a schedule that explains the drawmg reqUlremen s, " (
bids from several manul~'lcturers. The pump specifications should refer to the qual-
 428 SUBORDINATE PLANT FACILITIES
ing time, and equipment delivery. The clause concerning the plant site should spec-
ify, among other things, the elevation of the plant site, the range of ambient temper_
atures, and humidity. It is recommended that the general information also clarify the
key people associated with the project: for example, owner, engineer, and seller.
Performance testing is generally not an important consideration for small pumps:
those with a dischargesize that is less than 6 in. (15 cm). Yet it is vital for larger pumps
because a loss of efficiency of I to 2% from the specified value results in a significant
increase in the annual power cost. It is very important that the owner and supplier agree
on the method of testing and whether testing of the pump should be witnessed.
Special Considerations In addition to the considerations mentioned previously,
the design engineer should address five other issues: (I) the pump starting condi-
tions, (2) the use of high-speed versus low-speed pumps, (3) the method used in
controlling hydraulic surcharge, (4) the design of the pump suction well, and (5) se-
lection of the sludge pump.
Pump Starting Conditions The discharge valve of centrifugal pumps must be
closed at a particular time during startup to prevent hydraulic surges. Under these
circumstances, the pumping unit will not require additional horsepower. As long as
an adequate net positive suction head is provided, this practice also removes the
concern with regard to cavitation. This is also true for mixed flow pumps, up to an
approximate specific speed of 5000.
Propeller pumps, including most types of mixed flow pump, require very high
horsepower at startup. Thus the pump discharge valve should be sufficiently opened
after the drive unit has operated at full speed for a few seconds. If the discharge
valve is allowed to remain in a closed position, the water in the bowls will become
heated, thereby allowing steam to displace the water in the column. For this reason,
valve control during pump startup must be evaluated carefully.
High-Speed Versus Low-Speed Pumps The pumping requirement for a par-
ticular tlow rate at a specific total dynamic head may be met by two different meth-
ods. The first is to use the combination of a small impeller, capable of high speeds
such as 3600 rpm, with a high-speed motor. The second alternative is to use a slower-
speed drive (800 to 1200 rpm) with a larger impeller. Although pump selection is
intluenced by owner preference and plant conditions, the high-speed pumps are
cheaper to manufacture and the owner therefore has the option to select the lowest
bidder. Despite their lower cost, high-speed pumps are generally maintenance in-
tensive, due to problems associated with wear, erosion, misalignment, and vibra-
tion. High-speed units also tend to produce a high-pitched noise. It is the responsi-
bility of the engineer to evaluate the various offers carefully, including the economic
life span of the pump and the associated maintenance costs.
Hydraulic Surge Control Two types of hydraulic surge have been associated with
the pumping system. The first occurs during pump startup and in small plants; this
type of surge may be eliminated by installing a surge control valve between the
PUMPING SYSTEM 429
pump and the check valve (Figure 4.7-2). The second type, the water hammer, is
produced if the pump is suddenly stopped by the operator or if all the pumps sud-
denly fail due to a power outage. If the facility is expected to have a significant
problem with water hammer due to power outages, a surge tank should be installed
to alleviate this problem. Figure 4.7-2b illustrates surge tank installation for a hy-
draulic power generation system. However, the basic idea of the surge tank is the
same for the pumping system.
A special technique used in controlling hydraulic surges is surge analysis.
Although the technique requires both knowledge and experience to be used effec-
tively, a rule of thumb is to close the surge control valve (during pump startup) and
to open the valve (during pump topping) after the shock wave has reflected. The re-
quired time may be calculated from the formula 2Ua, where a is the velocity of the
shock wave (4000 to 5000 ftls) and L is the length of the pipeline in feet. Computer
analysis of the surge is often implemented using software developed by the
University of Kentucky.
Design of the Pump Suction Well A critical part of the pumping system design, for
the engineer, is the pump well, because the pump impeller requires a particular type of
flow condition to exist. The engineer must verify that flow conditions are uniform and
ascertain that the design of the pump suction well will provide the required type of flow.
The pump well may be an open pit or large pipe. Detailed design criteria for optimum
pump suction configurations may be found in the Hydraulic Institute Standards (15th
edition, 1997). Figures 4.7-3 and 4.7-4 illustrate some of these design guides.
Sludge Pumps Three types of pump are best suited for sludge pumping: the pos-
itive-displacement type, such as the plunger and progressive cavity pumps, the cen-
trifugal type, and the torque flow type. Regardless of the type of pump chosen, the
sludgc piping should be no smaller than 6 in. (15 cm) in diameter and the velocity
should be 5 to 6 ftls (1.5 to 1.8 m/s), to prevent clogging problems.
The plunger type of pump has a satisfactory history of operation. These pumps
are self-priming and the pulsating action tends to concentrate the sludge in the hop-
pers (located prior to the pumps) and resuspends solids in the pipelines during low
flow velocity. The main disadvantages of the plunger pumps are their bulky size and
cost. For these reasons, this type of pump is rarely used today.
The use of progressive cavity pumps, such as the Moyno pump, has been par-
ticularly successful in pumping concentrated sludge. The Moyno pumps are pro-
gressive cavity pumps and are self-priming at suction lifts up to approximately 20
ft (6 m). However, the pump is usually installed under tlooded suction conditions
to prevent the rubber stator from burning out.
Specially designed centrifugal pumps, such as screw feed, bladeless, and torque
flow, have also been used successfully. The torque flow type of pump, such as the
Wemco pump, has fully recessed impellers and is very effective in pumping sludge
but has very low pump efficiency. Engineers should refer to Wastewater Engineering
(2nd edition) by Metcalf & Eddy, Inc., for more detailed information on sludge
pumps.
 430 SUBORDINATE PLANT FACILITIES
PUMPING SYSTEM 431

RECOMM[NOf.D NOT RECOMMENDED

•
Surge control valve
• - o
o Vr -2tp.,Up

Discharge o __ 5_I...,r020
oO· If ,. _ LESS THA1
8.
Header
o A

• - AOO WALL THICK·

NESS Toli Dl51.
- ~~~r~3&s~cJ.'1
AT REAR Of WALL
_ APPX.D/l

, MAXaclS'
0\
PREFERRED 0-10' c
(a)

S
Not Ittc4ll'llll'llftdtd Ulllm" W. 50 OR MORE, OR
'If'UUlftilti
V, _ O.2Ip,OR LtSS "NO
y • SAME AS CHARrT() lEFT
, I If
to I
" • 10 1111
.., o S • IS GREATtA THAN 40

Tank and riser levels at Quarter Cycle Net area of

tank = F Initial Quiescent
Two cases Yl > ciz and Yl < cvl lever in surge
I I tank and riser

'¥ _ ZIp'
I [
h, ht
Ttle Dimension 0 is ,.nerally the diameter siln. Reter to the pump manufacturer for spec
Conduit discharge .. QI; Port area = a
.. ...
" :. "
... Conduit area = A OiS(;harge from tank
of Ih, suction btll mlnured .t Ih. inlet. This
dimension m.y vlry dependlnl upon pump de·
fie dimensions.

Figure 4.7·3 Multiple pump pits. (Adapted from Hydraulic Institute Standards, 14th ed.,
through ports = Q,
1983.)
I
(b) Qw
Figure 4.7-2 Surge control valv d Operation and Maintenance As described earlier, (he starting and stopping of
R. P. Walker, Pump Selection A e an su~ge tank. (~) Surge control valve. (Adapted from the pump requires special consideration. These procedures are dependent on the
tank installation. (Adapted fr~mn~:;::~!CI~I'~e ~~~lShers, A~n Arbor, MI, 1973.) (b) Surge pump type and (he associated accessories. Pumps with low hydraulic head, mixed
Hill, New York, 1952.) OJ Pl'le Hydraulics, C. V. Davis, ed., McGraw-
flow, and of the propeller type should be started with the discharge valve wide open
against a check valve to prevent backflow. However, pumps with medium and high
 432 SUBORDI NATE PLANT FACILITIES OZONATION SYSTEM 433

lI[lOCAT( I'UMI'S AT lACK WALL ual provided by the manufac turer and lhe plant operat ion and maintenance manual.

o AS INDtCAl[O IY OAStilO LlNU '

I:~ cfb"Al The major areas of maintenance concern inspection and replaceme nt. Three types
of inspection mu st be practiced : hourl y and daily, sc mi a nnual , a nd annual. Hourly

o cl'ILOCAf(O
and daily inspections require the operator to note the no rmal sounds of the pump
and th e bearing temperature, and check for leak s in the stufling box. Semiannual
procedures involve inspection of the pumps, sllifting box , packing, and ali gnment
o ....... ADO P[RfOR.lnO BAHlE
"PPROJ,IMAT[LY AS INDICA TEO.
• and the refilling of th e lubricat ion oil and g rease. Annual inspection req uires that
all parts o f the pump asse mbly he checked and c leaned thoroughly, and the con-
necting pipes be reali g ned with the pump. Finally, th e pumping system should be
• overhauled co mpletely whenever necessary, usually every 2 to 4 years,
Record keeping is also a very important issue in the operation and maintenance
of a pumping system. Eac h pump should have its own co mplete record on til e. This
MIN , W l.
record should note the maintenance cos ts, the Ilumber and type of re pairs, and the

.. .. ~O:;;;::"_7t-"::::ICtNAl tolal operati ng ho urs. Experience has shown that photog raphs of the internal paris of
the pump, taken du rin g sem iannual and annual inspection , prove to be very va lu able .
o
o- BIBLIOGRAPHY

o COIIII[Cf[o Alli s·C halme rs. Reference Dala, A lli s·Chalmers Pump. Inc .. Cinc innati . OH. 1970 .
Colt Industries. Hydmulic flal/{/book, Fairbank s Morse Pump. KS. 1975.
Cone AdcJed to c o
Reduce Poss ibility Davis. C. V. • Handbook of Applied Hydraulics. 1 rd cd .. McGraw-Hill. New York. 1969.
01 Submetged Vor[u
Format ion Hydrau li c In stitulc. American Natiollal Standard for Cell/nf llgalNertical PIIII//IS. Hydrau li c
lnslitutc . Parsippany. NJ. 1997.
Hydrau li c lnstilute. American Stal/dard f or PI/IIII' IlI wke Design. Hydrauli c ln stitul c.
Parsipprmy. NJ. 1998.

o
o
o
R(MO.... [O

c)
" ""!lUNG" SHOW~ AS
DARK BlU[ SHAO(O PAR'S

'COEO
Hydraulic In stitute, Hydl'Cllllic Ills/illfle SWI/t/art/s, 14t h cd .. Hydraulic In stitut e, Cleve land ,
OH. 1983.
K:uass ik , I. J., ct aI., PUIIlP Halldbook, McG raw- Hill. New York . 1970.
Langtca u, R. R , "C larilicat ion Con trol in Pumps:' ) , WPCF, 18:4:585 (A pril 1966).
Metca lf & Eddy. Wastewater En gilleerillg: Tremmell/. Dis/1o.m /, Rellse', 2nd cd .. McGraw-
Hill. New York. 1979.
o Shaw. G . V.. and Loo mi s. A. W.. Call/elVlI Hydraulic Data, Compressed Ai r Magazine Co ..
Ingcrsoll -Rnnd . Phillipsbu rg, NJ. 1970.

Figure 4.7-4. Correction of exist ing sumps (part one). (Adapted from Hydraulic In stitute
Stalldards, 14th cd .. 1983.) 4.8 OZONATION SYSTEM

hydra uli c head sho uld be started against a c losed di scharge va lve to reduce the start-
ing load on the motor. If the discharge line is already press uri zed by another pump,
a check va lve will be equivalent to a closed va lve. II shou ld be remembered that
whenever a pump is started against a closed di sc harge valve th e recirculation by-
pass line mu st be ope ned to prevent overheating.
T he pump stoppin g procedure is al so dependent on pUIllP type and accessories.
The plant operator is therefore advi sed to foll ow stric tl y both the operational man-
Purpose The ozonation process recen tly ga ined popularity as a treatme nt alter-
native in the United States after the regulatory agencics placed restrictions on di s-
infec tion by-products, especially the acceptable levels o r trihalomethane. Ozonation
is nn established and proven disinfection alternative as well as a preoxidant for the
control o r THM and HAAs precursors. O zonalion al so has the add ed benelits of
( I) oxidnt io n and volatili zation of organics; (2) co ntrol of al gae and assoc iated taste-
and odor-producing compounds; (3) des tabili zati o n (mi croll occ ul ali on) of ce rtain
types of turbidity; (4) re moval of colo r·causing co mpo und s; (5) oxidation of iron
434 SUBORDINATE PLANT FACILITIES
and manganese; (6) very short disinfection times for all pathogens, including
Cryptosporidillm; (7) partial oxidation of organics for subsequent removal by mi-
croorganisms; and (8) methyl tertiary butyl ether (MTBE) removal. In this section
we discuss the design considerations for the ozonation system and present some de~
sign issues with respect to ozone generation and the ozone contact tanks.
Considerations Ozonation is considered to be a relatively new treatment process
in the United States. However, the process has a history of almost 100 years ofprac-
tical experience. Thus basic knowledge should first be obtained from the extensive
source of published articles and operational experiences prior to conducting design
works. The second step, provided that there is both the time and budget, is to per-
form a pilot study; this will furnish the engineer with valuable design guides.
The design issues are (I) selection of a feed gas system; (2) preparation of the
feed gas system; (3) selection of the ozone generator; (4) design of the ozone con-
tact basin; (5) destruction of off-gas ozone; (6) control of DBPs by ozonation, par-
ticularly bromate; (7) control of assimilable organic carbon (AOC) levels due to
ozonation in order to minimize after growth of microbes in the water distribution
system; and (8) safety. The issue of safety will also be discussed briefly since ozone
is both a toxic and a flammable gas. The two most practical articles (published) on
this subject so far are the Design Manual Oil Municipal Wastewater Disinfection
(EPN62511-821021) written by the EPA and "Ozanation of Drinking Water" by
Domenic Grasso [Ozone Sci. Eng., 9: 109 (1987)]. There are also many valuable
technical papers on the ozonation system as listed in the Bibliography.
Feed Gas Selection Ozone many be generated from air, oxygen-enriched air, or
oxygen. The concentration of ozone produced by air is 1.5 to 2.5% by weight. The
ozone concentration is increased to 3 to 5% if high-purity oxygen is processed by
the same low-frequency ozone generators. However, 8 to 12% ozone can be gener-
ated with modern medium-frequency generators containing ceramic dielectrodes.
Feed Gas Treatment Both air and pure oxygen must be treated prior to being
fed to the ozone generator; this maximizes the ozone production rate and minimizes
maintenance work on the ozonator. The pretreatment process includes the removal
of dust, moisture, oil, and in some instances, nitrogen gas: Dust reduces the effi-
ciency of ozone production; oil is capable of fouling the dielectric; and nitrogen has
the ability to damage the ozone generator because of the production of nitric acid
within the unit.
If the gas contains excessive moisture, the life span of the dielectric of the ozone
generator becomes reduced and the power required to produce a specified level of
ozone is increased. Thus the removal of moisture is the most important of the gas
pretreatments. This task is achieved by means of a dryer, which uses a combination
of pressure, temperature, and desiccant.
The traditional ozonation system for application to water treatment is generally
composed of a precompressor with a 5-f.Lm paper filter, main compressor, after-
cooler [less than 95"F or (35'C)], oil coalescer (if required), refrigerant dryer (op-
OZONATION SYSTEM 435
tional), heat-reactivated desiccant dryer with activated alumina and molecular sieves
or silica gel, l-fLm filter, hygrometer, gas flow meter, and pressure-regulating valve.
The efficiency of ozone generation is a function of the feed gas treatment. The
capability of the feed gas treatment may be maximized by reducing the moisture
content of the reed gas to below 0.00596 Ib of water per 1000 ft 3 of air (0.14 g/m3),
with a dewpoint of -40'F (-40'C). This dewpoint can only be attained if 99.71 %
of the moisture is removed. Yet gas containing this moisture content is still consid-
ered to be in poor condition for the ozone generator. A dewpoint of - 58'F (- 50'C)
is judged to be marginal for the generator. The desired feed gas humidity is achieved
at a dewpoint of -76'F (-60'C); at this temperature the feed gas contains only
0.00051 Ib of water per 1000 ft3 of air (0.01 g/m3) with 99.98% of the moisture re-
moved.
The feed gas compressing process generally removes approximately 60% of the
moisture content of the air; an additional 20% is removed by the refrigerant dryer
and the desiccant dryer removes the remaining moisture so that a total of 99.98%
is eliminated. Excessive moisture content in the feed gas (air) not only reduces ozone
production and increases generator maintenance, but also damages the internal com-
ponents of the generator. Since the pretreatment of the gas is vital to the efficient
production of ozone, sufficient monitoring procedures and control devices should
be implemented in the process train. Serious consideration should be given to du-
plicating critical elements such as the desiccant dryer. Table 4.8-1 lists the moisture
content of air at various ambient temperatures.
The most popular air feed gas systems used in ozone generation systems are clas-
sified by their operating pressure. The system generally used is the low-pressure
system. It operates at a pressure of 10 to 40 psig (70 to 275 kPa). High-pressure
systems operate at pressures of 70 to 100 psig (480 to 690 kPa) and decrease the
gas pressure prior to the ozone generator. These high-pressure systems are typically
employed in small to medium-sized applications.
As discussed, air feeding to the ozone generator is a complex, costly, and pre-
cisely controlled pretreatment process that is maintenance intensive. However, a
great majority of the feed gas treatment process may be eliminated if high-purity
oxygen such as liquefied oxygen (LOX) is used to generate ozone. LOX is presently
available at a very reasonable price ($3.00 per cubic foot or $0.04 per pound). On-
site generation of oxygen may also be achieved by means of cryogenic, pressure-
swing adsorption (PSA), or vacuum swing adsorption (VSA) processes. However,
these on-site oxygen generation processes require complex mechanical and control
units. These units also create a somewhat negative visual impact and emit a certain
degree of noise. Thus design engineers should seriously evaluate the use of LOX
and the oxygen feed ozone generation system for a plant of any size, especially for
small to midsized treatment plants. This particular type of process offers several ad-
vantages for both the design engineer and the plant owner: The design engineer ben-
efits from the simple overall design scheme, especially instrumentation and control,
and a required floor space that is approximately half that of the air feed system. The
owner benefits from the potentially significant savings in both capital and operation
and maintenance costs. Consequently, most new ozonation facilities designed after
436 SUBORDINATE PLANT FACILITIES OZONATION SYSTEM 437

LOX DELI VERY Water

TABLE 4.8·1 Moisture Content of Air
f'I ---------------------,
influent

Moisture I
P1I:ESSURE I
Moisture Weight Content') I CONTROL I

I
Air Temperature I VALVe I FLOW
Weight Volume" (lb H20 Per METER

°C of (Ib H20llb air) (lblft') 1000 ft') :I~==~

r r-~~--~,~F~LO~"-,r-(:]r-£o~--~11
I I CONTROL AIR
I VAPORIZER ~ VALVE FILTER
-80 -112 0.0000003168 0.07526 0.00002384
-75 -103 0.0000007713 0.07526 : I

-70 -94 0.000001640 0.07526

0.00005805 L_~~ !A~I(.! _______________ J Ozone treated
water
0.0001234 OWNED AND OPERATED BY orHERS
-65 -85 0.000003342 0.07526 0.0002515 (optIonal) , ________ -:±-, eHluent

-60 -76 0.000006743 0.07526 0.0005074

-55 -67 0.00001311 0.07526 0.0009866
-50 -58 0.00002464 0.07526 0.001854
-45 -49 0.00004455 0.07526 0.003352 SUPPLEMENTAL
NITROGEN SYSTEM
-40 -40 0.00007925 0.07526 0.005964 Water
influent
-35 -31 0.0001381 0.07526 0.010393 VACUUM SWING ADSORPTION (VSAI
-30 -22 0.0002344 0.07526 0.017640
-25
-20
-13
-4
0.0003903
0.0006731
0.07526
0.07526
0.029373
0.050665
r-------------------------,
I
flfR' C J--lJI---{}-'
-15 5 0.001020 0.07526 0.07676 I
-10 14 0.001606 0.07526 I
0.1209 I
-5 23 0.002485 0.07526 0.1870 I
Ozone treated
I
0 32 0.003788 0.07526 0.2851 I water
I effluent
5 41 0.005421 0.07526 0.4080 I
I
10 50 0.007658 0.07526 0.5763 I
I
15 59 0.01069 0.07526 0.8045 I RoOM
I CONTROl,.
20 68 0.01475 0.07526 1.1100 1 VAPORizeR VALVE
25 77 0.02016 0.07526 1.5171 L_______~~~~:~~!~!~______ _
30 86 0.02731 0.07526 2.0552 OHNED RNo OPERA TEO BY OTHERS
35 95 0.03673 0.07526 2.7642 (optional)

40 104 0.04911 0.07526 3.6959

SOllrce: Adapted from EPA's Design M(/Illlal, EPAl6251l-86/02, October 1986.
AIR FEED SYSTEM
"Air weight/volume is corrected to 68°F and I atm pressure. Air weight/volume = 1205 glm 3 .
b g/m3 = IbllOOO ft3 x 16.012.
Influent

1990 employ LOX as the feed gas. Figure 4.8-1 is a schematic illustration of the
Lox feed system, VSA feed system, and classic air feed system for ozone genera-
Elfluenl
tion. gas stream
Residual
Ozone Generator Ozone is generated by passing a high-voltage alternating cur- After-
0,
c,;estroyer
cooler Oesiccant
rent (6 to 20 kV) across a dielectric discharge gap through which oxygen or oxy- drying system Ozone treated
elUuent
gen-bearing gas is injected. There are currently three basic types of ozone genera-
tor for use in water treatment applications: (I) low-frequency units with variable Figure 4.8-1 Examples of three different feed gas systems.
voltage; (2) medium-frequency units with constant voltage, variable voltage, or fre-
quency control; and (3) high-frequency units. These units are tube generators, in
contrast to the plate generators used exclusively in small ozone generators.
 438 SUBORDINATE PLANT FACILITIES
OZONATION SYSTEM 439

Low-frequency units have a frequency of 50 to 60 Hz, single-phase power, and Much credence should not be given to the sales pitch of the manufacturers. The
a peak voltage of 10 to 20 kV. The performance of these units is proven. Medium_ most practical ozone generators are either low- or medium-frequency units, primar-
frequency units operate at a frequency of 200 to 1000 Hz, are three phase, and have ily because of their years of operational experience, reliability, and the large number
a peak voltage of 8 to 14 kV; their performance is proven effective, thus they are of suppliers from which to choose. Units requiring lower voltages, in~luding many
wIdely accepted III the mdustry today. Another ozone generator is a high-frequency medium-frequency units, generate less heat, thus minimizing the cooling water re-
unit operating at 600 to 2000 Hz, with three phases, and a peak voltage of 8 to quirements and possibly increasing the life span of the dielectric. Medium-frequency
10kY. However, they are manufactured by a very limited number of companies.
units have a better power to production turn-down ratio than the low-frequency gen-
Figure 4.8-2 provides a general concept of ozone-generating systems.
erators: Low-frequency units may achieve a production of 20% at 20% power if the
airflow rate is reduced, whereas medium-frequency units may achieve a 10% pro-
2 duction rate at 10% power. Yet medium-frequency units contain more electronic com-
1. Atr inlet.
2, Ozonized air outlet
ponents than do low-fi'equency units; failure of anyone of these components will
3. Cooling water inlet. cause the power supply to malfunction and terminate ozone production.
4. Coohng water outlet. Selection between low- and medium-frequency ozone generators should there-
5. Dielectric tube.
6. Silent discharge zone. fore be based on the following considerations: (\) reliability and maintenance, (2)
7. Tube support. energy cost differential, (3) turn-down ratio, (4) cooling water temperature, (5) ben-
8 8. High-voltage terminal.
9. Port-hole. etits attained by using oxygen-enriched feed gas, and (6) owner preference. Table
10. Metal cooting. 4.8-2 briefly summarizes the characteristics of the two types of ozone generator.
11. Contact brush.
In the overall production of ozone, electrical power is a critical element because the
7 10 6 5 greater the power consumption, the greater the heat generation. Gas flow also affects
t
: Ii? the generation of heat; a gas flow that is too low will cause the generator to heat up,
whereas an excessively high gas flow rate will decrease ozone production. Since 90 to
95% of the supplied power is converted to heat, the cooling system is a very important
issue with respect to the generator. F0I1unately, manufacturers are highly competitive in
(0 I

TABLE 4.8-2 Brief Comparison of Two Types of Ozone Generator

OZone concentration (% by weight) Low-Frcquency Unit Medium-Frequcncy Unit

Heat o 1 2 3 4
30r---~__~__~__~~ Electronics Lower Higher
~ complexity
i 25 Power requirement
to produce I Ib
Specific energy of 4.5-5.5 kWh;
a total of 12 kWh with air
Specific energy of 4-5 kWh;
a total of 5-6 kWh with O 2
.§
- Electrode Q. 20 of ozone feed but 8-10 kWh with O2
f:==~==~:=:::!:j
Discharge
-- Dielectric E, Ozone concentration 1-1.5 with air, 2-3 with oxygen 2-3 with air, 5-6 with glass
~ 15 (weight %) (can be 5-6 wilh O 2 ) diclcctrodes and O 2 , 8-12
--+- gap ~ 0 3
8 with ceramic dielectrodcs
:=<"\'=9
~Electrode f
c
10 and O 2
~ Capital cost Lower for smaller-capacity units. Higher for smaller-capacity
~ 5
~
<n 0 L--L_....L__.l---L_-L---..l 0 <n
0.
2
i higher for larger-capacity units
($370/Ib of ozone for
units, lower for largc-
capacily units ($280Ilb of
Heat o 10 20 30 40 50 60 2000-lb/day unit) ozone for 2000-1b/day unit)
Operation and Higher. $O.8-$1.2I1b of ozone Lower, $O.8-$I.O/lb of ozone
( bl OZone concentration (g/m 3 )
maintenance cost
(, I Reliability Excellent Good to very good
Figure 4.8-2 General concepts on ozone gcneration: (a) horizontal tube, voltage-controlled, Effect of generator Moderate Sensitive
watcr-~ooled ozonator; (b) corona cell arrangement for ozone production; (c) energy con- Cooling
sumptIOn vcrsus ozone production for an ozone gencrator. temperature

:i,A:
 440 SUBORDINATE PLANT FACILITIES OZONATION SYSTEM 441

their cooling designs. The cooling water temperature should generally he 68°F (200C) Complete Ozone System
or less and approximately 0.75 to I gal (3 to 4 L) of cooling water is required for each (includes generation, building. and cantactor)
gram of ozone generated (0.23 to 0.3 gpm of water for each pound/day of ozone). Refer
10.000 ~--'-~~"""'-rM-.--.,--.,.-r""rTTT1
to Figure 4.8-3 for the various co?ditions that affect the rate of ozone production. Figure
4.8-4 IS an example of the bid pnces of both low- and medium-frequency ozone gener-
ators In the Vnlted States In recent years. Figure 4.8-5 is a breakdown of the construc-
tIOn costs. However, If all anCillary Items such as a new power substation are included
the to.tal cost can he more than two times higher than that indicated in the figure. '
After the type of ozone generator unit is selected, the design engineer must es- • •
tabhsh the following key design issues:

• Size of the ozone generator

• Number of generators
• Cost of energy
Availability and cost of oxygen
• Low frequency, air
•
1.2 1.2 .. Medium frequency. air
• Medium freQueney, 02
.i 1.0 loooL-__L-~LLLU~~-L~-L~~
100 1000 10,000
5" 0 .S
:l Ozone Generation CapacIty (Ibfd)
.~ 0.6
Figure 4.8-4 Construction cost of ozonation system.
j 0.4

0.2
0.6 0.8 1.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 Type of feed gas treatment system
Relative corona power Relative teed gas flow rate
• Reliability of each component
(a) (b)
Operation and maintenance costs
105
Ozone contactor design
;; 100
C> • Destruction of the off-gas
;;
0 95 90 • Vse of catalysts such as VV and hydrogen peroxide
"g 90
~
2 80
:l 0 The major ozone generators currently are Emery-Trailigas (ET), Ozonia (formerly
.~ 85 "
.~ 70 IDI and ABB) Schmidding-Werke, PCI, Mitsubishi, Sumitomo, and Fuji.
:;;
II! 80 ~
0: 80
Ozone Contact Tank As in any other water treatment process unit, the effective
75 50
5 15 25 35 45 55 -80 -50 -40 -30 -20 -10 mixing of ozone and process water is critical in maximizing the performance of the
Cooling water inlet Feed gas dewpoint (Oe) ozonation system. Due to the extremely low solubility of ozone in water, a special
temperature (OC) Ozone contact tank must be designed. Table 4.8-3 lists the solubility of ozone in wa-
(d)
(e) ter. The contact tank is an effective means of transferring the ozone from the gas
Figure 4.8-3 Various conditions that affect ozone production rate: (a) ozone yield versus bUbble to the bulk of the process water: The net ozone content in the gas bubbles
corona power for a constant feed gas flow rate; (b) ozone yield versus gas flow rate for con- is usually only 5 to 10%. This effective transfer of ozone to water is a critical step
s~ant corona power; (c) effect of cooling water on ozone yield; (d) effect of moisture on ozone in the ozonation process.
Yield. (Adapted, wit~ p~rmission, from R. P. Rice and A. Netzer, Handbook of Ozone Several types of contactor have been developed: (I) diffused bubbles (concur-
Technology and AppItCa/IOIIS, Vol. I, Butterworth-Heinemann. Stoneham, MA, 1982.) rent and countercurrent), (2) positive pressure injection (V-tube), (3) negative pres-
442 SUBORDINATE PLANT FACILITIES OZONATION SYSTEM 443

Ozone System Components TABLE4.S·3 Solubility of Ozone in Water

10.000 ,.---,--,--,-,-,-rnn---,--r..-rr-rr, Ozone

Water Henry's Ozone
Concentration"
Temperature Constant Solubility
eC) (atmlmol) mglL ppm vol. (mglL)

0 1,940 12.07 6,044 8.31

5 2,180 12.07 6,044 7.39
to 2,480 12.07 6,044 6.50
§ Generation only 15 2,880 12.07 6,044 5.60
co~ 20 3,760 12.07 6,044 4.29
:e €
.--
,~
1000 25 4,570 12.07 6,044 3.53
30 5,980 12.07 6,044 2.70
9,069 12.47
.3 0
5
1,940
2,180
18.11
18.11 9,069 11.09
10 2,480 18.11 9,069 9.75
15 2,880 18.11 9,069 8.40
20 3,760 18.11 9,069 6.43
25 4,570 18.11 9,069 5.29
30 5,980 18.11 9,069 4.04
l00~--~~~~LW~--~~-L~~~
100 1000 10.000 0 1,940 24.14 12,088 16.62
5 2,180 24.14 12,088 14.79
Ozone Generation Capacity (IbId)
10 2,480 24.14 12,088 13.00
Figure 4.8-5 Breakdown of ozonation system construction cost. 15 2,880 24.14 12,088 11.19
20 3,760 24.14 12.088 8.57
25 4,570 24.14 12,088 7.05
sure (Venturi tube), (4) turbine mixer tank, (5) static mixer, and (6) packed tower. 30 5,980 24.14 12,088 5.39
The countercurrent bubble contactor is considered to be the most efficient and cost- 0 1,940 36.21 18,132 24.92
effective of the alternatives for use in metropolitan water treatment plants and were 5 2,180 36.21 18,132 22.18
designed in the manner illustrated in Figure 4.8-6 until the early 1990s. However, 10 2,480 36.21 18,132 19.50
an extensive ($13 million) study on the advanced ozone oxidation process conducted 15 2,880 36.21 18,132 16.79
by the Metropolitan Water District of Southern California (MWD) using a 5.5-mgd 20 3,760 36.21 18,132 12.86
36.21 18,132 10.58
(0.24-m 3/s) pilot facility, as well as special studies using three computer models, in- 25 4,570
30 5,980 36.21 18,132 8.09
cluding the Finite Element Analysis (FEA) hydrodynamic model, revealed that a se-
vere flow short-circuiting problem existed for the traditional ozone contact tank de- Source: Adapted from EPA's DesiXI! Manllal. EPAl625/1-86/02, October 1986.
sign. The effective contact tank design created in the 1995 study by the MWD, "The concentration of the ozone gas is determined at a standard temperature of 68°P (20°e) and a stan-
illustrated in Figure 4.8-7, was called the baseline contact geometry by the MWD; dard pressure of 1 atm (101 kPa).
a TlOIT of 0.75 can be achieved with this design. Due to the rapid reaction kinetics
of ozone, a hydraulic detention time (HDT) of 5 to 10 min is considered to be prac- above the water surface; all pipe (Iype 316L SS) connections should be welded or
tical for regular oxidation and disinfection purposes. However, inactivation of flanged inside the tank and a check valve should be installed in the gas feed line
Cryptosporidium would require 15 to 20 min of HDT even though the EPA will not prior to the contactor to prevent the backflow of water into the gas line; the water
release the required C X t value or the MCL until mid-2002. depth in the tank should be 18 to 20 ft (5.5 to 6 m) 10 provide a minimum ozone
The ozone contact tanks have several common design features: The tanks must transfer efficiency of 95%. Finally, if the ozone concentration in the feed gas is 10%
be enclosed completely; they are composed of concrete and are located primarily or above, a different system, such as the sidestream pump injection method or a hy-
outside; and each tank has several cells and is capable of handling 50% of the max- draulic eductor and in-line static mixer, may be considered.
imum daily flow of the plant. Additionally, each cell should have a drain to allow
periodic dewatering and cleaning; stainless steel hatches should be provided to fa- Off-Gas Destruction The ozone contact tank typically has a 90 to 95% transfer
cilitate inspection and maintenance; a slight negative pressure [2 in. (5 cm) of wa- of Ozone to the water; the remaining 5 to 10% resulls as off-gas. The ozone content
ter] should be maintained in the tank to ensure that there is no leakage of ozone of the off-gas must be reduced to levels below the OSHA and local Air Quality
 444 SUBORDINATE PLANT FACILITIES
OZONATION SYSTEM 445

r Fi". Bubbl.
'nnu."r Diffuser. ITyp.) B-,
'I Sit • s'!.l

t
~~(Typ.)

II
ltSlideglre;1
Trlln A

Train a
:~g
0000
0000
gg
DCII j @]

Sit
@] ~ @I
12.ft
~ I!l
r
~J
14 It

Ilr:tVP .1Jl lin - -

In
Plan: Otone Contact a•• in
Jl • TRACeR ADOmOM POINT
... TRACER cou.£CTlON POINT

l!J
0 •• neG., Baseline Contactor Geometry
--' ~f;;:~::':~' Off·Gas "':7: Efflu
1;01.: Pressure
Differenlil'
l...
He
~

511 01 1
IP
~
i'
Stlge 2
1m )\
St.ge 3
r-
:=
....e-
Fine
Bubble
Diffuse,s N
I
I
!
lL,"-V =>L~ =iI~\V
lb <= =
Section A.-A.

Train B Train A

To Outfall
Streamline Plot for Metropolitan Water District OOP Ozone Contactor

Section 8.8

Figure 4.8-6Schematic of a three-stage bubble diffuser ozone contact basin. (Adapted from
0.150 r------;=::;;:;:;o;;;;;r::==]l
0*2.75 mgo:I

EPA Design Manual. EPAl625/1·86/02. October 1986.) 0,12.5

no p i additicn
_FEAModei p~

o A<:WaI Tesl DatI

0.100
Management District (AQMD) standards prior to venting into the atmosphere.
According to3
OSHA, the maximum allowable ambient concentration of Ozone is I
~ 0,07$
0.0002 g/m (0.1 ppm by volume) for an 8·h working day. The off.gas normally
3 €
contains levels greater than I g/m (500 ppm by volume). Three basic methods of w
0.050
destruction have been used: thermal destruction, thermal destruction with a catalyst,
and catalytic destruction.
0.'" o
Thermal destruclion is used almost exclusively for the air feed gas system; the o
oxygen feed gas syslem supplies such a high content of oxygen in the off.gas that 0." J.,_r.t.0_-:-~~""':;x-"",;:;,_-;-_-:."
00.511.522.
usc of thermal destruction is considered to be a fire hazard. Temperatures ranging Normalized Tirre, 9
jrom 570 to 660°F (300 to 350°C) for a period of 3 s are required to obtain 99%
ozone destruction by thermal destruction. Due to the high temperature and the large Tracer Test Result
power consumption of the unit, a heat recovery unit is usually provided. A demis-
, . mmended six·stage bubble diffuser ozone contact tank,
leI', a stainless steel wire mesh, should also be provided upstream of the ozone de- Figure 4.8-7 Concept .01 a Iceo D' 't f Southern California, Report 1105.)
(Adapted from Metropolitan Water Istnc 0
structor to minimize the accumulation of foam on the heating elements. The heat
loss of the thermal destructor is approximately 30%.
 446 SUBORDINATE PLANT FACILITIES
OZONATION SYSTEM 447

The system of thennal destruction with catalyst consumes less energy than ther- (OJ ppm by volume) for more than 10 minutes." Local Air Quality Management
mal destruction alone. Thus this system has become more popular in recent years. Districts may have more stringent criteria that are part of the restfictlo~S on smog.
The catalyst may be metal (platinum or palladium), metal oxides (aluminum oxide A few other safety considerations include the installation of self-contaIned breath-
or manganese oxide), or hydroxides and peroxides. However, the hydroxides and ing apparatuses in the event of a severe ozone leak, the installati~n of eyewashes,
peroxides are used only for pilot-scale studies. The advantage of the metal catalyst and the availability of comprehensive safety and operation and maIntenance manu-
system is its lower operating cost. The metal oxide catalyst system can operate at als, which may be consulted during an emergency situation. Ideally, the ozone gen-
85 to 120°F (30 to 50°C); this is much lower than the thennal unit. Yet the disad- eration building and ozone contact tanks are isolated from the operations bUIldIng
vantage is its sensitivity to chlorine, chlorine derivatives, sulfides, and nitrogen; for obvious safety reasons.
these chemicals may react with and destroy the catalyst. Activated carbon adsorp-
tIOn has been employed, but the powdery carbon produced by the reactivation process Design Criteria The design criteria for the ozonation system are numerous. They
is potentially explosive. Thus this method has limited application. range from the ozone generator, to the contact tank, to the destructl~~ UOIL The re-
. . In certain cases the off-gas is recycled back to the feed gas after conditioning or is quirements for these and many more issues are presented below. AddIllonally, an ex-
IOJected IOta the filter waste backwash water, which is discharged to a holding lagoon. ample of a detailed design is presented. (Note: I psi = 0.0703 kg/cm2 = 6.895 kPa.)
It should be noted that these alternatives are used very rarely because of the need to
maintain an additional process and the lack of high reliability for the off-gas treatment. Ozone dosage 1.5 to 3 mg/L (normal) depending on the purpose
or 0.3 to 1.0 mg/L of ozone per mg/L of TOC
By-products Control The major products of ozonation arc brominated com- Number of ozone generators Minimum of two, preferably three; one always
pounds, including bromate, assimilable organic carbon (AOC), and aldehydes. The acts as a standby
levels of these compounds depend on the quality of raw water and level of ozone Ozone generator
dosage. A large increase in the level of AOe after ozonation is a concern because Minimum production 10 to 20% of rated capacity
this condition promotes the regrowth of bacteria in the water distribution system Maximum production 75% of rated capacity if a continuous run
when there is a regular level of chlorine residual in the finished water; AOC level Cooling water temperature Less than 75'F (24°C) at the inlet
increases two to five times that prior to ozonation have been reported. Vessel construction Pressure vessel [15 psig (104 kPa)] constructed
To control the major ozonation by-products (except brominated compounds), with type 304L or 316L stainless steel with
granular activated carbon (GAC) is placed in filter cells to make them biological Hypalon or Teflon gaskets
filters or bacterial active filters (BAFs). Microorganisms that flourish in the upper Type of generator Low frequency with variable voltage, medium
portion of the GAC bed remove AOC, aldehydes, as well as TOC and DBP pre- frequency with frequency control, or another
cursors, thereby reducing these compounds to the levels regulated. Regular dual- type, depending on the selection
media filters and sand lilters also act as biological filters after prefilter ozonation. Compressor pressure 10 to 40 psig if heat-reactive desiccant; 80 to 100
However, the regular lilter beds lose the capability of bioliltration for several hours psig if pressure swing desiccant
when backwashed with chlorinated water, whereas the GAC bed is virtually unaf- Number Minimum of two; one acts as standby
fected. Thus it is recommended that engineers consider the use of GAC beds after Type Liquid ring, < 1000 Ib/day (455 kg/d) ozone
ozonation. The GAC filter bed is discussed further in Section 7.7. production; centrifugal, > 1000 Ib/day
(455 kg/d) ozone production; piston type,
Safety The issue of safety must be addressed because ozone is both a toxic gas oil-free, or oil lubricated with oil removal
~nd a fire hazard. Moreover, if the system uses oxygen as the feed gas, the situa- filters, <100 Ib/day-small plant
tIOn becomes more dangerous. Although the ozonation system may be less haz- Filters
ardous than the chlorination system because it can be shut down if an ozone leak Before compressor Regular filter with silencer
develops, it may also be more dangerous because the system must use high-voltage Before dryer 3t05ILm
electrical power to generate the ozone.
Before generator 0.3 I'm
The American Industrial Hygiene Association (AIHA), OSHA, and other assO- Pressure drop 0.5 psi when clean, 2 psi for coalescing filter
ciations recommend the folloWing pennissible levels of ozone in the air: "Workers Moisture removal
will not be exposed
3
to ozone concentrations in excess of a time weighted average Refrigerant dryer 41°F (5°C) dewpoint (80% removal from air)
of 0.2 mg/m (0.1 ppm by volume) for eight hours or more per workday, and that Desiccant dryer -76'F (-60'C) dewpoint (99.98% removal from
no worker be exposed to a ceiling concentration of ozone in excess of 0.6 mg/m 3 air)
 448 SUBORDINATE PLANT FACILITiES
OZONATION SYSTEM 449

NOTES: Ambient air contain s approx imately 2.7 g of water per kil og rmll of air. A Ozone destruction unit
dewpoilll of - 58°F ( - SOOC) is margin al for the des iccant dryer. Air with a dcw point of Type Heat catalyst unit (m ost popular)
- 76°F ( - 60°C) contains 6.7 mg of water per kilogram of air. Tempcrature 80 to 100°F (27 to 38°C)
Catalyst Metal (platinum) or metal oxides: these catalys ts
Desiccant regeneration cycle (heat reacti vation type) are proprietary items
Minimum cycle lime 8h Si ze of ozonation system 2 to 4 ft' /Ib of 0 , production (0.4 to 0.8 m' /kg)
Design cycle tim e 16 h building (estimated)
Oxyge n suppl y alternati ve: +98% purit y, 4400-gal ( 16.5-m') tank truck, Overall power requirement 10 to 12 kWhilb of 0 3 (air feed) (2 2 to 26
Liquid oxygen (LOX) I3,OOO-gal (50_m3) sta ndard -tank -size storage kWh/ kg); 8 to 10 kWhllb of 0 , (0 , feed)
tank (17.5 to 22 kWh/ kg)
Pressure swing adsorpti on 15 to 30 tons/d gaseous oxygen (GOX)
(PSA) production , 90 10 92 % purity, .10% turndown Figure 4.8- 1 is a schemati c presentation of a typical ozonation system.
Vacuum swing adsorption 30 to 90 ton s/day GOX production capac ity,
(VSA) 90 to 94% purit y, 30% turndown Example Design Calculations
Cryogenic di stillation
100 tons/day GOX as well as LOX producti on,
+98% purity, 50% turndown , high capital COS1 Example 1
Ozone conlact tanks
Number of tanks Minimum of two Given The ozone concentration of the gas flowing from the ozone generator to the
Tran sfer efficien cy Minimum of 95% if possible; may range tank is customarily expressed as "pcrcentage of ozone by weight" rather than the
from 90 to 95% wei ght of ozone in gram s per cubic meter of gas. Assume that the ozone concen-
Detenti on time (HOT) 5 to 20 min (depending on appli ca ti on) trati on is 20 g/m]
Stage of contact 2 to 10 (depending on appli cation)
Water depth 18 to 20 ft (5.5 to 6. 1 111) Determin e The percentage weight of ozone for the air feed ozone generation sys-
Submerge nce of diffuser 16 to 18 ft (4.9 to 5.5 m) tem and the oxygen feed ozone generati on system. Assume that the gases are all at
Freeboard 4 to 6 ft ( 1.2 to 1.8 m) to allow deposition of 20°C (68°F) and und er I atm of pressure (STP conditions).
[oa111
Ozone diffuser
Sollltion
Materi al Type 304L stainl ess steel, glass, ceramic, or
Tenon Gram molecular weight of air = 29 glmol
Bubble size 2 to 5 mm
Gram molecular weight of oxygen = 32 g/mol
Gas Il ow 0.5 to 4 ft '/min ( 14 to 11 3 Llmin), depending on
the diffuser: Gram molecular weight of ozone = 48 g/mol
rod diffuse r (2.5 in . X 24 in. L), 4 ft ' / min
max imum (113 Llmin); di sk diffuser (7 in. (i) For the air feed ozone generati on system the density of air is
diameter), 1.25 n3/1llin ma ximulll (35 L/min);
di sk diffu ser (9 in . di ameter), 1.8 ft ' /min MP 29
--= - - - = 1.2056 gIL
maximulll (51 L/min) RT 24.053
Headloss Maximum o f 0.5 psi
Permeabilit y 8 to 20 ft' / min per square foot (225 to 565 1205 .6 glm' (sec Appendi x 7)
Llmin' )
Porosity 35 to 45 % and the density of ozone is
Residual ozone (ozonated wat er) monitor
Type UV ozone monitor (continuous monitoring) MP = ~ = 1.9956 giL
RT 24.053
such as PC I Model HC or potassium iodide
monitoring 1995.6 glm'
 450 SUBORDINATE PLANT FACILITIES OZONATION SYSTEM 451

The volume of 20 g of ozone is _ gRT _ 25,6 X 24,053

Volume of 25,6 mg/L of ozone - MP - 48 12,83 L
gRT = 20 X 24,053
10,02 L 1,28% of the gas
MP 48
1% of I m' The weight of the gas is calculated to be

Since the volume of 20 g of ozone is I % of m' of the gas, the weight of the gas is 1205,6 + [(1995,6 - 1205,6) X 0,0128] = 1215,7 g/m3

1205,6 + [(1995,6 - 1205,6) X 0,01] = 1213,5 g/m3 Thus the percent weight of the ozone is

Thus the percent weight is

(25,6 -;- 1215,7) X 100 = 2,1%

(20 -;- 1213,5) X 100 = 1.648%, say, 1.65%

Example 3
(ii) For the oxygen feed ozone generation system the density of oxygen is
Given An air feed ozone generator is operating at an airflow rate of 300 standard
cubic feet per minute (scfm; 8,5 m 3/min); the standard conditions are 20'C (68'F)
MP 32
- - = - - - = 13304 gIL and I atm (14.7 psi) of pressurc. The ozone conccntration is measured to h; 1.65%
RT 24,053 weight by the ozone monitor, The plant water !low rate is 20 mgd (0,88 m'/s),
1330.4 g/m'
Determine
Knowing that the volume of 20 g of ozone is I % of the gas, the weight of the gas
IS (i) The ozone production rate Ib/day
(ii) The ozone dosage in mglL
1330.4 + [(1995,6 - 1330.4) X 0,01] = 1337,05 glm'
SO/ilfion (i) The conversion factor for giL to Ibm' is 0,0753, The ozone prodoe-
The percent weight is tion rate is calculated to be

(20 -;- 1337,05) X 100 = 1.496%, say, 1.5%

1.65 Ib 0, X 0,0753 X (300 X 1440)= 537 Ib/day (244 kg/d)
100 Ib air
Example 2 (ii) The ozone dosage is

Given The ozone concentration of the gas from an air feed ozone generator is mea- 537 -;- (8,34 X 20) = 3,22 mg/L
sured by the potassium iodide wet-chemistry method as 25.6 mglL.
Example 4 Design the ozonation system of a 20-mgd (average) water treatment
Determine The percent weight of ozone at STP conditions. plant using the design criteria presented in the "Design Criteria" section.

So/utioll Solution Step 1.' Ozonatioll Production Rates

25,6 mglL = 25,6 g/m3 Maximum ozone production rate = 3 X 8,34 X 30 = 751 Ib/day (340 kg/d)
Weight of air = 1205,6 g/m' from Example I Average ozone production rate = 3 X 8,34 X 20 = 500 Ib/day (227 kg/d)
Weight of ozone = 1995,6 g/m' from Example I Minimum ozone production rate = 1,5 X 8,34 X 5 = 63 Ib/day (29 kg/d)
 452 SUBORDINATE PLANT FACILITIES OZONATION SYSTEM 453

Step 2: Type, Size, alld NumberojOzone Gellerators A medium-frequency ozone
generator is selected primarily because of its J0: I turn-down ratio, lower power con-
sumption, and its ability to enrich the feed air should the plant need to double the
ozone output at a future date. Generally, any ozone generator should be operated for
extended periods of time at approximately 75% of the rated capacity. The rated max-
imum daily production of ozone should be 751 + 0.75 or 1000 Ib/day. Three gen-
erators, each capable of 335 Ib/day (150 kg/d), should therefore be provided. This
type of generator can accommodate a turn-down ratio of 10: I so that the minimum
production rate is approximately 35 Ib/day (16 kg/d). Thus a rate of 63 Ib/day
(28 kg/d) is easily achieved. Since two of the generators can satisfy the required
500 Ib/day (227 kg/d), under normal conditions, one unit should be left on standby.
Step 3: Size and Number of Air Compressors Since the maximum ozone con-
centration is 2.5% by weight for most air feed medium-frequency generators, the
air flow at maximum condition is
1000 Ib/day + (0.0753 X 1440 X ~o~) = 369 scfm or 370 scfm (10.5 m3/min)
Add 60 scfm to this number for purge airflow to the dryers. Thus the maximum air-
now may be calculated as
370 + 60 = 430 scfm (12 m3/min)
The minimum ozone concentration is approximately I % by weight when this type
of generator is employed. Thus the minimum airflow rate is
75lb/day + (0.0753 X 1440 X I~O) =69 scfm or 70 scfm (2 m3/min)
If 2.5% by weight ozone is produced on a continuous basis, the minimum airflow
rate would be
75 Ib/day + (0.0753 X 1440 X ~o~) = 28 scfm (0.8 m3/min)
This value is the smaller of the two. Add 22 scfm (0.6 m3/min.), for purge airflow
to the dryer. Therefore, the minimum airflow is
28 + 22 = 50 scfm (1.4 m3/min)
Based on these calculations, the compressors may be either medium [35 to 50 psig
(243 to 345 kPa)] or low [10 to 30 psig (69 to 207 kPa)], depending on the type of
drying unit selected. Provide:
Step 4: Size and Number oj Desiccant Dryers The design basis is as follows:
I. Operating pressure of 20 to 25 psig for low-pressure systems and 50 psig for
medium-pressure systems
2. Adjustable site elevation for the pressure
3. Inlet air temperatures: 77"F (25°C) with after-cooler and 40°F (4.4°C) with
refrigerated dryer
4. A required maximum dewpoint temperature for the air flowing to the ozone
generator of -76°F (-60°C)
5. Designed desiccant dryer cycle time of 16 h; minimum drying time of 8 h
Operation and Maintenance The ozonation system is both very sensitive and
expensive compared to other types of water treatment equipment. Plant operators
must be very conscientious in maintaining optimum operational conditions, for rea-
sons of efficiency and safety. The operational conditions of the ozonation system
must be adjusted by the plant operator whenever the plant flow rate changes. This
will provide the system with optimal operational conditions, thus facilitating maxi-
mum process efficiency and minimal operation and maintenance costs. Inducted in
the optimization process is the determination of the optimum ozone dosage; a 0.5-
mg/L reduction in dosage translates into a significant savings in power consumption.
The regular performance of several other tasks also directly affects the efficiency of
the ozonation system: checking the temperature of the cooling water, checking the
dewpoint temperature at the desiccant dryer unit, checking and calibrating the ozone
concentration monitors, checking the off-gas concentration, and adjusting the ozone
dosage. The plant operation and maintenance manuals, as well as manuals furnished
by the equipment supplier, should describe these very important items in detail.
A common problem for many large metropolitan areas is air pollution due to
smog. In these situations, the local AQMD sets the maximum allowable ozone con-
centration in the air. Thus the efficiency of the off-gas destruction unit should be
maintained at a level high enough to meet the stringent limits. Moreover, the ex-
hausting air from the ozone contact tanks and the building housing the ozone gen-
erator must be monitored closely for excessive emission of ozonc. With regard to
the safety aspects of the ozone system, strict adherence to the safety codes and reg-
ulations released by OSHA, the American National Standards Institute/American
Society for Testing and Materials (ANSI!ASTM), the American Conference of
Governmental Industries (ACGI) and the American Industrial Hygiene Association
(AIHA) is absolutely necessary.
BIBLIOGRAPHY

• One 50-scfm compressor (1.4 m3/min) Chang, S. D., and Singer, P. c., "The Impact of Ozonation on Particle Stability and Removal
of TOe and THM Precursors," 1. AWWA, 83:3:71 (March 1991).
• Two 100-scfm compressors (2.8 m3/min)
Cipparone, L. A.. et al .. "Ozonation and BDDC Removal: Effect on Water Quality," J. AWWA,
• Two 150-scfm compressors (4.2 m3/m in) 89:2:84 (February 1997).
 454 SUBOR DINAT E PLANT FACILITIES

Dowbiggin, W. 8. , an~ Singer. P. c., "Effect of Natural Organic Matter and Ca lciulll 0
Ozone- Induced Particle Destabi li zat ion," 1. AWiVA , 8 1:6:77 (June 1989). n
Edwards, M., et al., "Effects of Ozonalion on Coag ulation of NOM Using Polymer Ala
and Polymer/Melal Sa il Mi xlUres." 1. AlVlVA, 86:1: 105 (January 1994). ne
Fe rguson, ~ . V:., et a l.. "Applying Ozone for Orga ni c Control and Di sinfec ti on: Ut ilit
CHAPTER 5
!'erspccllve.· 1. AIVWA, 83:5:32 (May 199 1). Y
Ferguson, D. W., c.t ~J.. "~o mparing Pcroxone and Ozone for the Control of Taste and Odor Design of Plant Components
Compounds, Dlsmfectl on By-products and Mi croorga ni sms" J AWWA 82'4' 181 (A .
1990). .. ,. . pnl

Jaca ngelo,,!.G .. et aI. , "Ozonation: Assessing Its Role in the Formation and Cont rol
D.B.!'s. 1. AWWA, 8 1:8:94 (AuguS! 1989). of
Kat zcnel s~,n , E., et aI. , " Inact ivation Kinetics of Viru ses and Bacteria in Water by Use of 5.1 FLOW MEASUREMENT
Ozone. 1. AlVIVA, 66: 12:725 (Dece mber 1974).
Krasner, .S'. W., et aI., "Formati on and Control of Brolll~He Durin g Ozollntion of Water Purpose Flow measure ment is thc most importa nt variablc meas ured in a treat-
Conlm lllng Bromide." 1. AWIVA, 85:1:73 (January 1993).
me nt plant. The three principa l appli cati ons for fl ow measuremcnt a re establishing
LePage, W. L., "A Treatment Plant Operator Assesses Ozonati on " J AWWA 77'8'4 1 (A the fl ow rate of the process and determining the liquid chemical and gas flow rates
1985). .. ,. . uguS!
of various subordin ate processes. Differential pressure prod ucers, direct disc ha rge
Li <ang , S
., 1 < lono fMTBEb y a zonea ndPeroxoneProcesscs,"J.AWWA 9 1'6' 10
elal ., "Ox'da"
measurement , pos iti ve vo lume di splacement meas ure ment , and fl ow ve locity- arca
(June 199 9). ' " 4
mcasurement are the four basic typcs of measuremcnt commonly used in watcr treat-
Ma lorcy, S. W. , et a I. , "Ozone-GAC Following Conven tional U.S. Drinking Water Treatmelll " ment pl a nts .
1. AWWA, 77:8:66 (AuguS! 1985). .
Metropolitan Water District of Southern Ca lifornia , Oxidation Rellvjit PlVgmm : Desigll
Report, Report 11 05, Novembe r 1995, Considerations Several questi o ns must be considered to se lec t the most appro-
Monk. R. D. F.. el al. . "Prepurchasing Ozo ne Equ ipmenl." 1. AlVlVA, 77:8:49 (AuguS! 1985). priate type of flow measure me nt equipment.
Rakn~s.s, K., L., "Ozone Disinfection," Chapt er 6 in EPA Mal/llal: Mllnicipal Wastewater
DlSlllfectlOn, EPA/625/1-86-021, U,S. Environmcnt .. 1 Prolection Agency. Washington DC I . Is liquid or gas n ow being measured ?
Oelober 1986. ' ,
2. Is the flow occurrin g in a pipe or in an open c hanne l?
Ri ce. R. G., "Uses of Ozone in Drinking Water Trc3tll1CIlt " J AWlVA 73' 1'44 (J' . 198 1)
R' . R G ' , . , . .. <lIlualY . 3. Whal is Ihe mag nilude of Ihe n ow rate?
ICC, . ",a nd Netzer. A.• Hal/dbook of Ozolle Techflology al/d Applicariolls, Vo l. I, Ann
Arbor SCience, Ann Arbor, MI, 1982. 4 . What is the range o f flow variatio n?
Rice. R. G., and Netze r. A., Handbook of Ozolle Tecl/ll ologV alld Applicariolls Vol. 2, 5. Is the liquid being measured clean, or does it contain suspe ndcd so lids or air
BUllerworih. Boston, 1984. . , bubbles?
Sc hylz, C. It, and Hawkins, R, A.. "Eva luating a High Effic iency Ozone Injec tion Comaclor " 6 . What is the accuracy requireme nt?
1. AWWA, 87:8:85 (Augusl 1995). '
7. Whal is the a llowab le head loss by the now meter?
Weinberg H S I I "F .
, . "e a " ormatIOn a nd Rcmova l of Aldehydes in Pl ant s That Use Ozone"
1. AlVlVA, 85:5:72 (May 1993). . 8 . Is the flow corros ive?
9. What types o f flow meters arc a vailable to the regi o n?
10. Whal Iypes o f poslin slalla lion service is avai lable 10 th e area?

Types of Flow Meters Flow me lers commonly used in Ihe field of waler Irea l-
mcnt arc the venturi meter, orilice me ter, propeller meter, mag netic fl ow meter, ul -
trasoni c fl ow meter, vortcx mete r, rotamctc r (variable-area me ter), flumes, a nd weirs.
Liquid chemical flow is a lmost cxclus ively measurcd by pos itive di splacement
pumps rathcr Ihan n ow me ters, with the cxception o f the rotameter wh ic h may be
used in certai n situations.
455
456 DESIGN OF PLANT COMPONENTS FLOW MEASUREMENT 457

Differential Pressure-Producing Meters Differential pressure-producing me- ferential; these values depend on the ratio of the orifice diameter to the pipe diam-
ters currently on the market are the venturi, Dall tube, Hershel venturi, universal eter. This type of flow measurement equipment is suitable for measuring both gas
venturi, and venturi inserts (Figure 5.1-1). The accuracy of these flow tubes is ± 1% and liquid !low.
for a flow range of 10: 1. However, since the square-root extractor is used before A third type of dilYerential producer is the !low tube (now nozzle). The throat
the 4- to 20-mA signal is transmitted, the range of accurate flow measurement is re- ratio determines the head loss of the !low tube; headloss ranges from 10 to 20% of
stricted to no more than 5: I. The head loss across a venturi flow meter is relatively the differentia1. However, its application is primarily in sewage works (Figure
small, ranging from 3 to 10% of the differential, depending on the ratio of the throat 5.1-1). To assure uniform flow through the meter, a straight pipe with a length equiv-
diameter to the inlet diameter (beta ratio). These !low tubes should be installed at alent to at least 10 pipe diameters should ideally be provided upstream of all dilc
an elevation that will sustain the meter continuously at full pipe flow and ideally ferential flow meters. Yet the meter manufacturer may only allow for a straight pipe
will not allow negative pressure to form at the throat during the maximum flow rate. that is five pipe diameters in length in conjunction with a flow straightener. Figure
Orifice plates are another type of differential pressure producer. Both concentric 5.1-2 presents the headloss characteristics of several venturi flow meters and their
and eccentric orifice plates are available. Their accuracy is ±2% for a flow range piping requirements.
of 4: I. The oritice plate yields a high headloss ranging from 40 to 90% of the dif-
Propeller and Turbine Meters The Sparling, Rockwell, Badger, and Neptune
meters have been on the market for decades. These types of metcr have a magnetic
coupling that senses the revolutions of the spinning rotor and sends oul electronic
pulses. The accuracy is I to 2% for a !low range of 10: I, provided that there is a
OUT~fT
ONE uniform and straight pipe of at least five pipe diameters and straightening vanes
INLET
SHOULDER
~rjji!~--: T UBr ahead of the meter. The minimum now velocity should generally be 0.7 ft/s
.w~'" LINER (0.2 m/s) and the maximum !low velocity must not exceed 6 ft/s (1.8 m/s). The pro-
peller and turbine meters should be employed in situations where the water is rela-
INLET tively clean and noncorrosive and for liquids that are not too viscous since they arc
CONE subject to wear, pm1icularly in cases where the water contains suspended solids such
INLET THROAT as silt. Thus the meter must be installed in a readily accessible location to facilitate
FI.ANGE PIEZOMETER
VENT easy maintenance. Main-line meters are available in sizes up to 24 in.
(0.6 m) in diameter; the saddle type, up to 72 in. (1.8 m); and the intake meters, up
to 78 in. (2 m) in size.
DALL TUBE
Magnetic Flow Meters The magnetic now meters operate by having a conduc-
tor, such as water flowing in a magnetic field, create a voltage that is proportional
-JI.-IH"- TAPERED to the velocity of relative motion. Fixed magnetic coils wrapped around the pipe
METERING TUBE
produce a magnetic tield that is perpendicular to the !low of liquid. The induced
voltage is detected by two electrodes embedded in the pipe. The accuracy of this
type of meter is :±: I % for a !low velocity range of 3 to 30 ft/s (0.9 to 9.1 I11/s) and
within :±:2% if the range is I to 3 ft/s (0.3 to 0.9 m/s). Yet this type of direct dis-
charge meter is relatively expensive and may require periodic calibration to correct
drifting of the signal.
FLOW The magnetic flow meter may be used for measuring most any lype of liquid, in-
IN
cluding the measurement of sludge !low. Sizes up to 48 in. (1.05 m) are readily
available, and larger sizes Illay be custom manufactured. Since there is no Ilow re-
striction inside the mcter, there is essentially no headloss within the unit. However,
ROTAMETER the size of the meter should be one size smaller than the size of the line, so that the
FLOW NOZZLE flow velocity requirements can be met. The two fundamental requirements are that
a uniform and straight pipe of at least five pipe diameters be placed ahead of the
Figure 5.1-1 Common flow meters. meter and that the unit should receive a full pipe now.
 FLOW MEASUREMENT 459

Ultrasonic Flow Meters The flow measurement of the ultrasonic flow meter is based
on the time required for an ultrasonic pulse to diagonal1y traverse a pipe or a channel
against the liquid flow. Application of this type of direct discharge meter should be to
clean water so that it measures the transit time or time of flight and not the reflection
•• %
~

of the pulse off particles or air bubbles (Doppler effect). The flow velocity is propor-
.." !" 0

tional to the difference in the time of the two pulses. The accuracy is ± I % for a flow
~"....f
~

01.1. ... r I
"
.0 «
~
velocity ranging from I to 25 ftls (0.3 to 7.6 mls), but the meter reading is greatly af-

if ••• •"~
I
fected by a change in the fluid composition. Thus this type of meter may only be em-
ployed for water that is free of particles and air bubbles as much as possible. In the
YENTURI ! I I
I
, case of large pipes Of a channel, more than onc pair of sensors is required.
0

Vortex Meters When a liquid flow hits a bluff object, eddies and vortices are
• ..l"i
i • formed; this is known as vortex shedding. The frequency at which the vortices are
generated is proportional to the velocity of the liquid flow. The resulting pressure
changes may be detected by probes mounted on the body of the bluff. This infor-
0.60 0.80
mation is then relayed to an indicating recorder by a 4- to 20-mA signal. The man-
ufacturers claim that this type of meter has an accuracy of ± 1% for a flow range
'00 of 12: I. The headloss is approximately two times the velocity head through the me-
ter. The cost of the vortex meter is moderate. However, this type of direct discharge
meter does not function well with viscous, dirty, abrasive, or corrosive fluids.
'0

Rotameters (Variable-Area Flow Meters) Rotameters consist of a tempered

'0 glass tube containing a freely moving float. This glass tube is vertically mounted
with the wide side up (Figure 5.1-1). Rotameters may be used in measuring both
70 gas and liquid flow rates. The flow rate of the particular gas or liquid is indicated
when the float reaches a position of equilibrium between its weight and the upward
~ force of the medium. The flow range for this type of meter is approximately 10: I,
.0 •
~ with an accuracy of ±2% of full scale. Yet the accuracy is highly dependent on the
Z
w
,. «
w skill of the person reading the meter.
•
'0 Q
• A rotameter may be applied in situations with very low flow rates, 0.1 to 140
OII"IC(

•0 gal/h (0.4 to 530 Llh) for water and I to 520 standard cubic feet per hour (0.028 to
14.7 m3/h) for air, and the accuracy of the meter is not affected by the piping con·
'0 .. ,
~
~
figuration immediately upstream of the rotameter. Basically, these meters are con-
0
u. ~ trolled manually and read or measured visually.

T,,,,,,,VHT lilT , YNTUItINO (

'" Flumes (Positive Volume Displacement Type) Flumes are used for open-
LO"G 'Oil"" ZO channel flow. The Parshall flume and Palmer-Bowles are the types used predomi-
yUll III T
nantly in wastewater treatment. These meters are used to measure small to very large
flow rates [0.02 to 3000 mgd (0.9 Lis to 130 m 3/s)] and are applicable to both clean
,. .. TUI '" and dirty water. The accuracy of the flumes is ±5% over a range of 20: I, but the
... , It .. II NT III' range may be expanded to 35: I with some sacrifice in accuracy. It is important that
o engineers provide a sufficient fall downstream of the flume to avoid the effects of
0.1 02 0.1 0.4 O.~ 06 OJ' 0.8 09
IUllO - THROAT 01.1.. I INLET 01.1.. (p RATIO J
the backwater, because the level measurement is conducted at a specific point in the
flume by either an ultrasonic level detector or a float located directly above the point.
Figure 5.1-2 Piping requirement and headloss for venturi flow meters. (Adapted from BIF
The flow rate may also be determined in a stilling well by a bubbler or float.
Catalog, 1968.)
458
 FLOW MEASUREMENT 461
460 DESIGN OF PLANT COMPONENTS

Weirs Several types of weirs-for example, the V·notch, rectangular, or trape. GPM MGD CFS
»»~-------------,
zoidal-are currently available for usc in the field of water treatment. The head 4000
»»
above the crest of the weir is usually measured by a float, bubbler system, acoustic 2000
meter, hook gauge mounted above the weir, or stalf board mounted on the sidewall
1000
of the channel. The practical tlow range of the weirs is approximately 30: I and the BOO
600
accuracy is ±3%. Rectangular and trapezoidal weirs have a head-tlow power rela. 400
tionship of approximately 1.5, while the V·notch weirs have a power relation of2.5 300
~n: to Poon! TO
(see Appendix 14 for the formulas). Therefore, any inaccuracy in the measurement 200 meUure meUure

'~~
dePth. H, dePTh.",
of the hydraulic head will greatly affect the accuracy of the tlow measurement, es.
pecially in the case of the V·notch weirs. Yet the V·notch weir plates may be used
100
eo
60 ;: I ~ X~_
TnTet •','

Ol-A~
I '

;;.~: I
_ W' \"\ r
for a much wider tlow range: an extreme case would be 3000: I. 40
30 ,I' i . 2llA
Although the weirs are easy to manufacture and are cost-effective, the main dis- 20 ~,
advantages are their high headloss and the tendency for settleable solids and large -M--8--F--C_
debris to collect on the upstream side of the weir. Figure 5.1·3 illustrates tlow curves
10
8 ,.~"' .'~
,
.'
_. ,: .~
,
for Parshall numes and a 90' V·notch weir. The equipment manufacturers can pro· 6 .
·N: ... K·
4 ~- .•», ~
vide the engineer with detailed information on the various types of flow meter cur-
,
3 ,
STOoe
rently available, and all engineers are advised to review the brochures prior to se-
lection. Virtually all types of meters, with the exception of the rotameter, require an 1.0
1·,4
,.)
0.8
adequate length of uniform and straight conditions upstream of their location. Most 06
of the now measurement meters are capable of sending a standard 4· to 20·mA 0.4
0.3
signal.
D.'
CFS
0.1
Example Design Calculations 0.08
0.06
0.04
Example 1 0.03 20
, ! 'ddt! " ''/ rid,!, 0.02
2 3 4 6 810 152030406080100 Flow
Given A water treatment plant has a daily maximum, average, and minimum de- Head·inches
I II.!.!" 111.1".1 0.01
sign tlow rate of 22, 15, and 4 mgd, respectively. However, this plant is designed .1 .2 .3,4 6.81.0 2 34 6810
to double its capacity in approximately 15 years. ,.)
Head-teel
"
Determine The appropriate size and elevation of the venturi tube, which is con-
nected to a plant inlet tank containing a water depth of 10 ft.

,
NOTES:
go, lf~noTtiI We,r
r." of!!ow In GPM
Conversion factors from English units to the metric system are as follows: 1 mgd o , JI)
0.044 m'/s: I ft ~ 0.305 m: I in. ~ 2.54 cm. For Tower CUrlfe (dolled line)

2. The selection guides for the venturi meter are as follows: o 500 1000 1~ 2000 2500 3000
FOf upper Curve (sCbd line)
(al
(a) Select a meter size that will allow for an inlet flow velocity of 6 to 10 ft/sec at the aver-
age daily flow rate. Figure S.1M3 (a) Parshall flume; (b) flow curves for Parshall flume; (e) V-notched weir;
(b) Select a proper beta ratio for the Venturi tube so that a differential pressure of approxi- (d) flow curves for 90° V-notched weir.
mately 80 in. will be produced at the average daily flow.
462 DESIGN OF PLANT COMPONENTS LEVE L MEASUREMENT 463

(c) Check the magnitude of the differential at the dail y maximum and daily minimum de. Hydraulic grade line
sign nows. The minimum differential shou ld be at least 3 in. and the negative pressure
should be no more than 2 ft at the th roa t of the venturi (al the differential value) during
th e maxim uill now rate.

SO /lIlioli The pl ant influent pipe shou ld be s ized for a future flow rate of 30 mgd
(daily average ll ow rate). Thus the lin e s ize s hou ld be 36 in . in diameter (v :::
6.6 ft/sec ). Yet the l10w veloc it y is only 3.3 ft/sec (I m/s) at 15 mgd; thi s is the av-
erage daily fl ow rate o f the initial s tage. The Venturi tube se lec ti on table (Appendix 20 In. Type 0 VentUri meter ~
20m
12) indicates that none of the 36· in. tubes will sat is fy the criteria; a type A throat
36 In
"- 36 in.
(beta ratio A) o f 36 in . is charac terized by a 120- in . differential at 16 mgd, but the --+-

differential is 240 in . (20 ft) at 22 mgd ; that is, the centerline o f the Venturi tube
mu st be install ed ap proxima tely 20 ft lower th an the water level in the inl et tank to
avo id producing a s ignificant amount of negative pressure at the throat o f the Venturi
I,I' t7 fL (to)--1

I Approxima tely 30 ft~

tube. The differential at the minimum des ign fl ow rate o f 4 mgd is
Figure 5.1-4 An example o f Vent uri n ow lUbe Installati on.
120 in . .;- ( 16/4)2 = 7.5 in . > 3 in . OK
BIBLIOGRAPHY
If the s ite condition allows thi s type o f co nfi g uration for the plant inlet pipe,
with out rai s in g the cost s ig nifica ntl y, a Venturi tube with a Type A throat o f 36 in, Andrew. W. G .• and Willi ams. H. B .. Applied Illstl'lIl11ell((l{ioll ill the Pl'Ocess Illdustries, 2nd
should be se lec ted for the initial plant. When the plant capacity is do ubled (at a fu- cd .. Vols. I and 2, Gulf Publishing. Houston, T X, 1980.
ture date), the Venturi may be repl aced with a Venturi tube with a Type C throat, Babcock. R. H .. " Instrumcntation and Control in Water Suppl y and Wnstcwa!cJ' Di sposal."
whi ch wo uld produce a differential o f 80 in. at a daily flow rate o f 32 mgd, 160 in, WaleI' Wastewater Ellg., ( 1968).
( 13.:1 ft) at 45 mgd (maximum daily 110w), and 5 in. at 8 mgd (minimum daily flow Department of Ci vil Engi nee ring. " Instrumentation, Control. and Automat ion for Wat er
rate). Supply and Wastewater Treatmcnt Systems," Proceedillgs of the 9111 Sal/itfll )' EI/gilleerillg
Alternati vely, the pipe s ize may be reduced from 36 in . to 20 in . with a mini- COllferellce, Uni ve rsi ty of Illinois. 1967.
mUIll pipe leng th of 30 ft, if the in!lu cllt line cann o t be in stalled g reater than 20 ft Knud sen, D. I. , cd., Process IlIstl'lflllell/(ltiofl al/d COI/lI'OI System, M anual of Practi ce OM -6,
below the water leve l of the tank due to the s ite con diti o n or cost. Under these cir- Water Pollution Control Federation, 1982.
c um stances, a 20- in . ve nturi tube (v = 10.5 ftlsec at 15 mgd) should be selected:
The Ve nturi tube se lectio n table indicates that the type D (20 in .) Venturi tube will
give a differe nti al of 80 in . at a flow rate of 16 mgd. The 110w rate at 15 mgd is 5.2 LEVEL MEASUREMENT

80 in . .;- ( 16/ 15i = 70.3 in . (6 ft)
The differential at the daily ma ximum !lawaI' 22 mgd is
80 in . X (22/16)2 = 151 in . (12.6 ft)
and th e differential at 4 mgd is 5 in . A minimum o f 17 ft o r uniform straight pipe
a nd a straightener s ho uld be prov id ed upstrea m of the Venturi tube (Figure 5.1·4),
When the plant capac ity is do ubled, the 20- in. line (co ntainin g the VenLuri tube)
is replaced by a 36· in. lin e and a 36· in . Type C Venturi tube. This alte rnative re-
quires that the centerline e levati o n be approximately 14 ft below the water level in
the inlet tank; thi s is easi ly achi eved in most cases.
Purpose Level meas ure ment is an esse nti a l item in plant operati o ns bccause the
plant operator mu st be aware o f the water level o f all the process unit s, the leve l.s
of all che mi cal storage tank s and sil os, and the pressure o f water or compressed all'
lines- th at is, the wate r leve l in the di s tribution mains and the utilit y lines. There
are several mean s o f meas urin g liquid leve ls; via a 110at, pressure e le me nt s, bubbler
syste m , or an ult raso ni c system . Meth ods used in meas urin g dry. powdery materi -
als include ultrasonic system s, pho tocell sys te ms, rotary paddle sw itc hes, diaphrag m
unit s, and wire s train gauge syste ms. However, it is o ft en vcry diflic ult to meas ure
the level of' powd ery materials held in storagc bins acc uratel y, due to the concavity
of the surface or irregular surface palLe rn . The 1110st acc urate and re li ab le me thod
of mo nito rin g any dry or liquid material in storage bins or tank s is to use load ce ll s
to mcas ure the to tal we ig ht (tank plus co ntent).
464 DESIGN OF PLANT COMPONENTS
In the remainder of this section we describe some features of the (I) float sys-
tem, (2) pressure elements, (3) bubbler tube system, and (4) ultrasonic level detec_
tor, Other methods of level measurement are not commonly employed in water treat-
ment and are therefore not discussed.
Float System In the past, the most common method of level measurement of
open tanks or channels was the float-operated transmitter. This is a simple and rea-
sonably accurate system. Nevertheless, the system is very unsightly, and the instal_
lation process is very time consuming and expensive. due to the need for a stilling
well and a collection of wires, wheels, and tackles. The float system also requires
periodic maintenance to assure friction-free motion of the float and cable assembly.
Pressure Elements Level measurement via pressure elements is very common
in the field of water treatment. This method determines liquid levels by measuring
the pressure produced by the head or by measuring the level above the pressure mea-
suring tap. A pressure transducer connected to the pressure elements measures the
water pressure at the base of the tank and reads the liquid level directly if properly
calibrated for the liquid.
NOTE: Conversion factors from the English unit system to metric and SI systems are as
follows: I psi - 0.0703 kg/em' ~ 6.895 kPa.
The bourdon tube, bellow element, diaphragm element, and manometer are ex-
amples of pressure element measurers. The bourdon tube has both helical and spi-
ral units and is suited for high-pressure (water depth included) measurements. The
pressure range of the spiral units is commonly 0 to 15 and 0 to 200 psi (0 to I and
o to 14 kg/cm2). The helical unit is typically used for measuring much higher pres-
sures: 0 to 200, 0 to 400, and 0 to 6000 psi (0 to 14, 0 to 28, and 0 to 420 kg/cm2)
or higher.
Bellow units measure pressures in the intermediate range. The units are COID-
monly manufactured to measure water levels (depth) of 0 to 100 in. of water (0 to
2.5 m) and a pressure range of 0 to 5 and 0 to 30 psi (0 to 0.35 and 0 to 2 kg/cm2).
The diaphragm type of pressure element is particularly suited for obtaining mea-
surements within a small range in the low-pressure zone. The available units can
cover a range of 0 to 10 and 0 to 20 inches (0 to 0.25 and 0 to 0.5 m), and 0 to 40
and 0 to 140 inches of water (0 to I and 0 to 3.6 m).
The use of manometers is limited to pilot studies or temporary use; when the
prime meter is out of service. Pressure elements often require the following condi-
tions to yield accurate readings: zero suppression, zero elevation, or differential pres-
sure for pressurized tanks (D/P cell).
Bubbler Tube System The bubbler tube system is used to measure the back
pressure of the hydrostatic head. This is equivalent to the hydrostatic head of the
liquid at the bottom of the bubbler tube. This type of level measurement is currently
LEVEL MEASUREMENT 465
the most widely used method applied to tanks. The bubbler tube system has a tube
placed inside a tank which runs from the top of the tank and opens approximately
3 in. (7.5 cm) from the bottom. During operation, compressed air is supplied to the
submerged bubbler tube via a regulator (pressure regulator, differential pressure reg-
ulator) or a purge rotameter. The bubble purge rates are very low and as long as the
bubbles escape periodically, the system is functioning properly. The purge rate is
generally 0.5 ft3 of free air per hour (14 L/h).
The source of the compressed air may be a small air compressor, a compressed
air tank, or a nitrogen gas cylinder. If electrical power cannot be provided, for what-
ever reason, the use of a compressed air cylinder is the most economical method; a
standard scuba diving compressed air tank can supply I month of air and is easily
refilled. The end of the bubbler tube must be set at a point of zero elevation or zero
level measurement. For example, if the end of the bubbler tube is located 12 ft
(3.6 m) below the surface of the water, the bubbler tube would indicate a water level
of 12 ft (3.6 m). If the water level drops by 5 ft (1.5 m), the measured depth would
be 7 ft (2.1 m). This system detects the water depth (pressure) by means of an ap-
propriately calibrated gauge or manometer.
The bubbler system has the advantages of (1) simple design, (2) easy accessi-
bility and little concern for the corrosion of the pressure sensing device, and (3) the
ability to be installed at the bottom of the tank where accessibility is blocked or im-
possible. Figure 5.2-1 shows a bubbler system used in open-tank level measure-
ment.
Ultrasonic Level Detector The ultrasonic system is used to monitor either the
water level in a tank or dry material stored in a storage bin open to the atmosphere.
The height of the water/material is measured by means of an acoustic pulse; the ul-
trasonic transmitter and receiver units are located above the maximum level of the
Clean\OutPIU~--------
'g I t ,n. Copper tubong
a L __ .l. __ Alarm
(typical)
1000 ft maximum
'::::::[)I<::J:::s
\-
Clean air
-orgas
Differential/'
pressure
'"
Pressure regulator
t
regulator
Figure 5.2-1 Bubbler tube system of tank level measurement.
 466 DESIGN OF PLANT COMPONENTS
object. The transmitter generates a pulse down through and perpendicular to the liq-
uid level or surface of the dry material. The time elapsed between pulse generation
and the detection of the reflected pulse energy is a function of the speed of sound
through air. Since the speed of sound in air is dependent on temperature (0.18% per
degree CelsIUS), use of a temperature correction factor is necessary for accurate mea~
surement. Prior to selecting the level measuring device, the design engineer should
be familiar with the following considerations: the required pressure range, required
accuracy, equipment reliability, environmental conditions of the plant site, and op-
erations and maintenance concerns.
BIBLIOGRAPHY
Andrew, W. G., and Williams, H. G., Applied Instrumentation ill the Process Industries, 2nd
ed., Vols. I and 2, Gulf Publishing, Houston, TX, 1980.
Babcock, R. H., "Instrumentation and Control in Water Supply and Wastewater Disposal,"
Water Wastewater Eng., (1968),
Department of Civil Engineering, "Instrumentation, Control, and Automation for Water
Supply and Wastewater Treatment Systems," Proceedings of the 91h SaniIGl)' Engineering
Conference, University of Illinois, 1967.
Knudsen, D. I., ed., Process Instrumentatiofl and COlllrol System, Manual of Practice OM-G,
Water Pollution Control Federation, 1982.
5.3 VALVE SELECTION
Purpose The flow of liquid or gas in pipes is regulated by valves, whercas gates
are used to regulate the flow of water from reservoirs, tanks, or channels. Valves are
used much more frequently because they are capable of performing a multitude of
important functions. Their primary functions include shutoff, throttle, prevention of
backflow, or a combination of these functions.
To be able to select an appropriate type of val ve, all design engineers must have
a basic knowledge of the various types of commonly used valves, the characteris-
tics of the fluid that is to be regulated, and the hydraulic characteristics of the pip-
ing system in which the valve is to be installed. Valves are manufactured out of a
wide variety of materials to meet a wide range of requirements, and the engineer is
responsible for specifying the appropriate valve type as well as valve materials.
Engineers should also be aware that all large [over 12 in. (0.3 m)] valves, includ-
ing gates, require adjustment after installation~ without the necessary adjustments,
most valves will leak to some degree. Design engineers should therefore include the
final valve specifications.
Considerations Items that must be considered during valve selection are (I) the
type of fluid or gas that is to be regulated, (2) the temperature of the system, (3) the
flow range, (4) the pressure of the system, (5) valve function, (6) valve location,
VALVE SELECTION 467
(7) the type of valve operator, and (8) the reliability and cost of the valve. Each of
these items is discussed.
Type of Fluid or Gas The characteristics of the liquid or gas are important when
selecting the appropriate type of valve. If the liquid or gas is corrosive in nature,
noncorrosive materials must be specified for all internal parts of the valve; type
18-8 stainless steel is typically selected. Valves used in ozone gas lines must be com-
posed of type 316 stainless steel, and Teflon seats, not rubber, must be used. If a
chemical slurry is used in the treatment process, the valve must not have internal
recesses which may become filled with slurry solids. Finally, if abrasive matter is
present in the liquid, the fluid passages must be composed of materials that are re-
sistant to this type of erosion.
Temperature Although temperature is not a major issue when selecting the ma-
jor valves of a water treatment process, it becomes a very important consideration
when valves are used in conjunction with auxiliary equipment such as heating boil-
ers and certain types of chemical feed systems-those that handle exothermic chem-
icals such as caustic soda and sulfuric acid. In these types of situation, valves ordi-
narily used in the water treatment process should not be employed because, due to
thermal distortion, they do not perform well at operating temperatures above 150"F
(65"C) unless special metal parts are specified.
Flow Range Flow range is an important issue when selecting throttling valves.
Control of the flow range is extremely important because most throttling valves have
a limited range in which they yield satisfactory performance. Unlike the throttling
valve, flow range is not an important consideration for multijet valves. These valves
are designed and manufactured to service a wide range of pressures and flow con-
trol without cavitation.
For the purpose of simple shutoff, the flow range is usually not a design con-
sideration. However, most valves are unsuitable for service when water velocity ex-
ceeds 35 ftls (10.5 m/s), based on the valve port area, and the engineer must there-
fore specify special instructions for valve construction.
Pressure The maximum differential pressure across the valve and the normal and
the extreme line pressures should be known prior to valve selection. Extremely high
pressures often result from hydraulic surges in the piping system, and selection of
the valve operator should therefore be based on the maximum line pressure.
Valve Function Each type of valve has an optimal function. The valve function
may be isolation of a line, drainage of a tank, prevention of backflow, reduction in
pressure, or flow modulation. Thus the design engineer should specify the most ap-
propriate type of valve for each application.
Valve Location Valves may be located in a variety of locations: in a valve vault,
a pipe gallery, in the wall at the entrance of a tank, at the exit of a pipeline, buried
 468 DESIGN OF PLANT COMPONENTS
in the ground, or submerged in the water at all times. There are also differences in
valve construction for buried and regular types of valve. Furthermore, valves installed
in the tank wall require only one flange rather than a double flange. Engineers must
select the proper type of valve and specify the conditions for valve installation.
Valve Operator The design engineer must specify whether a valve is to be oper-
ated manually or by power. If a manually operated valve is selected, the type of op-
erator (i.e., a wheel or a square nut with key) and the orientation of both the oper-
ator and stem supports must be specified. Certain types of valves, such as flow
modulating, automatic sequencing, and remotely operated, necessitate the Use of
powered operators because manual power is insutlicient.
Powered operators may be energized by means of electricity, compressed air, wa-
ter, or oil. However, the tirst two options are most commonly used in waterworks.
Strictly from the standpoint of reliability and maintenance, the use of air actuators
is preferred during a power failure or during repair work because of its simplicity.
However, an electrical motor offers better control over the modulating speed and is
easier to incorporate into the instrumentation and control system.
Reliability and Cost The degree of valve reliability may play an important role
in the control system, yet may be independent of the application. If a high degree
of reliability is essential to the treatment process, the cost of the valve is unimpor-
tant. However, the relative costs of the various sizes and types of valve should gen-
erally be compared for each application. A good cost comparison study will include
the cost of the valve itself as well as the projected maintenance costs and the cost
of replacing equipment when necessary.
Types of Valves There are tive basic types of valves: (I) slide, (2) rotary,
(3) globe, (4) swing, and (5) multijet. The slide valve has a sliding disk that travels
pcrpendicular to the flow direction. An example is a gate valve. Rotary valves operate
by having a plug or disk move in a rotary fashion. The butterfly, ball, plug, and cone
valves are classified as rotary type valves. Globe valves are used extensively in or-
dinary home plumbing tixtures. They have disks or plugs that move parallel to the
flow direction. Globe valves do not cavitate under most household usage conditions.
The swing valve is comprised of a swing check valve, which prevents reverse
flow. The swing check valve is fundamentally a combination of a rotary and a globe
valve. The multijet valve, also known as a sleeve valve, is composed of two con-
centric pipes. Both the inner and outer pipes are covered with a multitude of small
orifices. The inner pipe slides back and forth like a telescoping pipe, thus aligning
or staggering the holes. An alternative type has perforation only in the outer sleeve.
A third type of multijet valve has multiple orifices on the slide gate disk. The mul-
tijet valves became available in the early 1970s. They are used exclusively to re-
duce high pressure and to control flow rate without causing cavitation (under most
conditions). Table 5.3-1 summarizes valve characteristics and the main valve ap-
plications commonly used in water treatment.
VALVE SELECTION 469
Flow Control Valve Considerations The design engineer must carefully se-
lect the most appropriate type of valve for each application; the use of an improper
type of valve may result in cavitation, corrosion, or erosion and thereby degrade the
structural integrity of the valve itself. Valve selection may be conducted in two steps:
selection of the proper type of valve, followed by sizing.
Valve Selection To select the most appropriate type of valve for a given applica-
tion, the design engineer must evaluate the pressure drop characteristics and the work-
ing range of the valves. Figure 5.3-1 presents examples of valve rangeability. Figure
5.3-2 shows the !low characteristics of certain types of valves when throttled. Engineers
should not confuse rangeability with turn-down. Rangeabilify is the ratio between the
maximum controllable flow and the minimum controllable flow rate. Tum-down is a
ratio of the normal maximum flow rate versus the minimum controllable flow rate.
If a valve is to be used in a situation requiring watcr pressure control, valves
characterized by equal percentage !low, such as the ball and butterlly valves, should
be selected for ordinary cases-where there is a normal pressure drop of at least
15% but less than 30%. If a higher pressure drop, such as 50%, is expected, a valve
with linear characteristics (plug or l11ultijetlsleeve valve) should be specitied. In sit-
uations requiring a single valve to handle a pressure drop of over 50%, the sleeve
valves perform effectively in most applications without cavitation. For applications
requiring the control of liquid levels, a valve with linear characteristics, such as the
plug valve, is the most appropriate type in most cases.
Equal-percentage valves are most appropriate under the following conditions:
(I) the process is fast acting; (2) the situation requires high rangeability; (3) the dy-
namics of the system are not well known; and (4) in the case of heat exchangers,
an increase in the production rate requires a correlative and greater increase in the
volume of the heating or cooling medium.
Linear valves should be used (I) for slow-acting processes and (2) when there is
a system pressure drop of over 40% across the valve. Quick-opening valves should
be used (I) for on--off control and (2) when maximum valve capacity must be ob-
tained immediately. The quick-opening valve exhibits linear characteristics for ap-
proximately one-fourth of its travel from shutoff. Beyond this point, the quick-
opening valve is useful only in an on-off function.
Valve Sizing The sizing of control valves involves the application of basic sizing for-
mulas, evaluation of the process fluid characteristics, and use of basic data for the valves.
Some design engineers fail to take the size of the control valves into consideration and
end up with the wrong-size valve. Nevel1heless, system control by these valves is
achieved to an acceptable degree most of the time; this is possible because of the high
rangeability of the control valves and the tlexibility of the controller mode and because
the available pressure drops are often greater than the design points. However, optimum
system control may be achieved only if the control valve is capable of handling both
the minimum flow and a tlow that is over 100% (120%) of the maximum flow rate.
Another factor that should be considered when sizing the control valves is the
capacity factor (Cv). Cv is detined as the number of gallons of water [at room tem-
~

Cl
TABLE 5.3-1 Characteristics and Applications of Common Valves
Rated
Standard Pressure
Type Size Application (psi) Remarks
Gate valves 2-48 in. • OnJoff control but does not have 150 This class includes wedge, double disk
ANSIIAWWA (5-120 em) a driptight shut-off (10.5 kg/em') gate valves. and sluice gates
C 509-87 for • Not suitable for flow rate control Debris deposits in the seats will cause
gate valves by throttling valve closure problems
AWWA C 501-87 Full size openings allow pipe cleaning pigs
for sluice gate to pass through
• The valves often become inoperable when
left in an open or closed position for
extended periods of time
Double disk type should only be installed
vertically
Slow opening and closing speeds minimize
water hammer
• Requires a large space for installation
Butterfly valves 4-90 in. • On/off control with drip-tight 75 and 150 Easy to operate and at a low cost
ANSI/AWWA (10-230 em) shut-off (5.3-10.5 kg/em') • Minimal maintenance requirements
C 504-87 "A" Limited degree of now control • Requires a small space for installation
valves have a rated by throttling • Quick dosing causes water hammer
maximum line flow problem
velocity of 8 fps~ Applicable to both air and liquid
"8" valves, up to Water should not contain any large debris
16 fps Does not allow the pipe cleaning pig to
pass through

• Full size openings allow pipe cleaning pigs

Up to 300 to pass through
Driptight shut-off available • Very little head loss when fully opened.
i-48 in. Limited degree of flow control by
Ball valves (1.3-120 em) (21 kg/em') • The size of the valves is often necked down
ANSI/AWWA throttling to {o of the size of the connecting line
C 507-85 Services clean water as well as slurry
Large size valves are heavy
Fun size openings allow the pipe cleaning
Up to 300 pigs to pass through
On/off control with non-driptight
6--24 in. available • Operation of the valve begins to lift the
Cone valves closure (21 kg/em') plug off the seat before rotation of the
(15-60 em) Limited degree of flow control
by throttling cone
• The valve operator is rather complicated
and expensive
Should not be installed in an upside-down
position
• The port consists of a tapered plug or a
full-area circular port
• Plug valves are suitable for service of dirty
120-170, depend- water and slurry as well as clean water
• On/off control with driptight ing on the size
-~-36in. • Lubricated and eccentric plug valves are
Plug valves dosure (8.4-12 kg/cm')
(1.3-90 cm) Often used as a flow control available
• Lubricated type comes with cylindrical or
valve tapered plugs with a diamond shaped port
Mild throttling is acceptable
for throttling application
Lubricated types are easily operated, pro-
vided that they are frequently used~ how-
ever, the plug type tends to "freeze" if is
not operated for an extended period of time

~~
 :!::i TABLE 5.3-1 (Continued)
'" Rated
Standard Pressure
Type Size Application (psi) Remarks

Lubricated types require periodic lubrica-

tion with grease
Nonlubricated types are easier to operate
and require less maintenance
Eccentric plug valves have rectangular
ports whose area equals 80% of the pipe
area
Eccentric valves may be installed in any
position; however, the shaft must be kept
horizontal and the plug must remain in the
upper part of the valve body when the
valve is opened
Globe valves 1-16 in. On/off with a high pressure drop ISO but up to 300 Disk globe, angle valve. Y-valve, needle
(2.5-40 em) and driptight closure available valve, and altitude valves are in this class
Accurate throttling and tlow (10.5-21 kg/em') Application to clean water only
control The disk design is used for manual control
and the needle valves are used for accurate
and automatic control
Most all household valves are disk globe
valves
Check valves 2-60 in. • Prevents reversal of tlow ISO Lift, swing, and ball checks are the basic
ANSI/AWWA (5-150 em) (10.5 kg/em') types of check valve
506-78 and Internal spring-loaded types are called
ANSI/AWWA "silent" checks
508-82 Lift checks are often used with globe valves
for both Jjquid and gas. Angle Hft check

I
!"

valves are used for vertical line; horizontal

lift types are used for horizontal lines
Swing checks are often used with gate
valves and may be installed in either a
horizontal or vertical position
Ball checks are used for air vent, vacuum
valves. and chemical metering pumps-on
both the suction and discharge sides
Headloss across the check valves is signifi-
cant
Driptight shut-off for corrosive 150 but up to 300 The valves are used in lime slurry and PAC
Diaphragm val ves ~-24 in.
(1.3-{)() em) chemicals, slurry, and sewage available slurry lines
(10.5-21 kg/em') Intermittent operation may cause leakage
Poor flow control characteristics
Has a low cost but a high capacity
Flow control with a high pressure ISO Both in-line and submerged discharge types
Multijet valves 8--72 in.
(2()"180 em) drop across the valve without (10.5 kg/em') are available
cavitation Has a history of good performance; emits
low noise and has minimal maintenance
requirements
Limited number of valve manufacturers
These valves are expensive

Notes: (1) ANSUAWWA C 540-87 standard for power-activated devices (operators) for valves and sluice gates.
(2) ANSUAWWA C 550-81 standard for protective interior coating for valves and hydrants.

~
474 DESIGN OF PLANT COMPONENTS VALVE SELECTION 475

100 Ideal inherent valve characteristics and

-other valve configurations
90

80

70

60
~
•
~
50

~ 40
"-

30

20
o Stem position-
Stem position
Inherent Valve Characteristics Percentage of Maximum Flow
10 Versus Stem Travel for
Typicat Valve Types
0
0 10 20 30 40 50 60 70 80 90 100 Figure 5.3-2 Flow characteristics for common valve. (From W. G. Andrew, Applied
the Process Industries, Gulf Puhlishing, Houston, TX, 1980.)

L
Travel(%) flltrll/nellta/ioll ill

Rang.ablilty = 95/10 = 9.5 -----I
1 - - - - - Rangeability = 98/2.5 = 39.2-----1
Figure 5.3·1 Rangeabilities for common valves. (From W. G. Andrew, Applied
Ills/mlllel/fation illthe Process Industries, Gulf Publishing, Houston, TX, 1980.)
where Cv(~ is the capacity factor through a I-in. (2.5-cm) valve for a given valve
angle with a pressure drop of I psig (0.07 kg/cm') and D is the pipe diameter in
inches. Table 5.3-2 compares the capacity factor (gpm) for several types of valves.
Most major valve manufacturers will provide the Cv values of their products.
The control valve is generally selected to open only 25 to 60% at the designed
tlow rate. The following general guides should be used in valve selection:

I. To select the appropriate size valve, begin by comparing the value of Cv at

perature, 60"F (15.5°C)] that will pass through a given flow restriction per minute both maximum and minimum flow conditions.
at a pressure drop of I psi (0.07 kg/cm'). The formula used to compute Cv is es- 2. Next, select a valve size that is capable of handling the maximum Cv when
tablished by the Fluid Controls Institute, Inc. (FCI), and the formula for noncom- the valve is 85 to 90% open, and the minimum Cv when the valve is opened
pressive fluids such as water is only 10 to 15%.

G
Cv = Q ( (!,.p )0' (volume based)
Cavitation Considerations Cavitation is defined as the formation and subse-
quent collapse of vapor-tilled cavities resulting from the dynamic action in a liquid.
The cavities may be bubbles, vapor-tilled pockets, or a combination of the two.
where Cv is the capacity factor (gpm), Q the flow rate of the liquid (gpm), G the Dissolved gases are often liberated shortly before vaporization begins.
specitic gravity of the liquid (1.0 for water), and (!"P the differential pressure drop Cavitation causes noise and vibration, and if it is severe, both the disk and body
across the valve (psig). of the valve will become pitted. Under these unfavorable conditions, cavitation may
become so severe that the butterfly valve (flow modulator) must be replaced every
Cv"" = Cv few months. The design engineer must make sure that the valve selected for flow
'I' D' control does not cavitate under critical conditions, that is, when the valve is at a
 VALVE SELECTION 477
0
N £~ "
.....' 0
,,",
..... '"
M
N "''''
N M
hi gh throttlin g position. The cavitation index (a) is calculated from the following
equation:
..... 0 00 MO
0 00 N
;:b$
'" M
M "
0 '" Hd - H I'
- N
a =
AH + v'l'g
where IT is the cavitati on index, Hd the static head at the valve exit (absolute feet).
~8
""''''.....
"'0
00 M '" .....
'" M
- ..... Hv the vapor pressure of the liquid, AH the press ure drop across the valve (ft), and
v the velocity th rough the pipe, which is directly co nnected to the valve (ft/s).
g
~~8~~~~o
" '"
N
Vi M"""M~II'),,,,"-oS
;.,
.Q
NOTES:
R I . The barometric pressu res. expressed as feet of water column. at given ground elevations
b'
are as follows:
">
~ 10.332 m at sea level
~
~
33.9 ft at sea level
0 32.8 ft at 1000 ft elevation
9.997 m at 305 m
·c 9.631 m at 610 m
~ 3 1.6 ft at 2000 ft elevation
~
30.5 ft at 3000 ft elevation 9.296 m at 9 14 m
.E 8.96 1 mat 1220 III
c
0
29.4 n at 4000 n elevation
~ 8.625 m at 1524 m
·c 28.3 ft at 5000 ft elevati on
"E
Co

0 2. The absolutc water vapor pressures. expressed as feet of water column. are as follow s:
U
~
0
ti 40°F (4.4°C) 0.28 ft (0.085 Ill)
If 50°F (10°C) 0.4 1 ft (0.125 Ill)
> 0 .59 ft (0. 180 m)
U 60°F (15.6°C)
~ 70°F (2 1. 10c) 0 .83 ft (0.253 Ill)
~ 1.1 8 ft (0.360 m)
80°F (26.7°C)
"
'g 90°F (32.2°C) 1.6 1 ft (0.49 1 m)
=> 100°F (32.8°C) 2.2 1 ft (0.674 m)
"
~
~ The criteria for valve cavitati on are as follows:
0
c
In general, cavitation wi ll not occur and no particular noi se
·c5: cr > 2.5- 3
will devclop.
"c.
E 1l1cre is a development of noise; "hammer blow s" and a
0
2.5 < cr > 1.5
u steady roar persist.
Under these circumstances. the pipe and valve are apt to
.€ cr < 1.5
vibrate and cavitati on develops. Whcn (1 rcaches 0.5.
"
c.
cavit ati on will be full y developcd.
U"
~
on &
Figure 5.3-3 illustrates a multijet (sleeve) valve. This type of valve is designed
and used as a throttling valve in situations with a high pressure drop in the system.
"-l ~ TIle multijet valve does not develop cavitation until the cavitation index approaches
...:l
=
1!:
">
~
0.2. This fi gure al so compares the cavitation index of various types of lTIuitijet valves.

476
 Incipient Cavitation Index VS Valv e 0 penlng VALVE SELECTION 479
• .0
200
o•• '/ 00 Example Design Calculations
.,./
~ 1.0 50
II.• 0
/' 40

20 I Example 1
£
.•
, 0
c/
~
!7
10 .0 \ Given A 36-in. (0.9-m) pipeline connects a water reservoir to a water treatment
!•
•
o. 4 - f-... § V~ / .0
\ plant. The difference in static water level is 30 ft (9.1 m) A butterfly valve is se-
o. , .J. i?" /" I.--~
" "- lected to control the planned flow rate from 31.5 to 5 mgd (1.38 to 0.22 m'/s. The
,; 0
...... pipeline elevation, with the venturi flow tube and the flow control valve, is 15 ft
/
0.2
..,.- I. 0 (4.6 m) below the water surface at the head of the plant. Assume that the headloss
/' ~

o. I in the line is 14 ft (4.3 m) at the maximum flow rate and 0.5 ft (0.15 m) at the min·
imum flow rate. The plant elevation is 1000 ft (305 m) above sea level and the am-
o i i
0 10 20 '0 40 50 60 70 80 90 100 0
bient temperature is 80°F (2rC).
- v~rve Opening 1%1
10 20 30 40 ~ 60 70 ao 90 100
Mark. - - - - - Valv. O~n,nq 1%1

111 In·I..,n. Mull,pl. J.I SI ..... e

C'V'If!'O~ ,nO •• (0) • He!. 1-11> - Ii.
Determine The proper-size flow control valve, and evaluate the cavitation situa-
lSI OIObl l,nu. V~I... .
(9t V.'hul M~H;'il' SIU"" V,I .....
1101 Lub".:Iltd PllIg Val".
'"
wht,. Md· DOwn", • .,n"t'"'''' (IHII tion
(III e"IWlly Valvl 111>· e.,omeu,c 1)<'"'''' (1'.11
I!~) C.,"V,lvl H• • V,PO' ""U.. " II,.U
~H • H..d (I,OP 'C'O" •• h,. lIud Solution Step 1. Select a 36-in. (0.9-m) butterfly valve, which is the same size as
Lou·F~tlor Tabl.
Nail: r~. 10n1KIO' 0'''11 ... orh Ih, n n'Ib
the pipeline. The capacity factor Cv at the maximum flow rate of 31.5 mgd
II It of POIlI. SO j)ln., con ... !! "',beta" d"'ijn,nij p,,,.nl!.
(1.4 m'/s) is

OilChar91 Chamber Dilch.rge Tank

Cv = Q(~pr
SYltem Symm
= (31.5 X 695) [ I
(45 - 14 - 15) + 2.31
]0.5 = 8318

Cv ,
Cva = - , = 8318 + 36 = 6.4
D

Based on the graph in Figure 5.3-4, the disk angle of the valve is approximately
35°. For the purpose of flow control, the recommended disk angle of the butterfly
valve is generally 15 to 60°, and 35° is therefore an appropriate valve.
The Cv at the minimum flow rate of 5 mgd (0.22 m'/s) is

Cv = (5 X 695) [
I ]0.5 = 972
(30 - 0.5) + 2.31
Cv ,
Cva = D' = 972 + 36 = 0.75

In this case the disk angle is approximately 9°, which is a very throttled position.
This condition is unacceptable since it is less than the recommended angle of ISO.
Step 2. Since the rule of thumb is to size the flow control valve one size smaller
Figure 5.3·3 Mit" (I than the size of the line, begin by selecting a 24·in. (0.6-m) valve. The placement
brochure.) U lJet s eeve) valve and cavitation index of values. (From Bailey Valve
of a pipe size reducer and an increaser on the ends of the 24-in. valve will produce
478
 480 DESIGN OF PLANT COMPONENTS
PIPING SYSTEM 481

The water vapor pressure at 80"F is 1.18 ft (0.36 m); the tlow velocity through the
45 valve inlet is 2.46 ft/s at 5 mgd; and the pressure drop across the valve is calculated
to be 45 - 14 = 31 ft (9.45 m):
40
46.8 - 1.18
(J= 1.47
35 31 + [(2.46)' 7 64.4]

30 This value is definitely in the cavitation range. The important decision-making fac-
~
~ 25
tors are how often and how long this 5 mgd (0.22 m3/s) tlow requirement occurs.
If this low-tlow condition is infrequent and short in duration, the 24-in. (0.6-m) but-
tertly valve may be used with the knowledge that the cavitation condition exists and
l 20 that the valve disk will cavitate every time. However, a plug valve or even a multi-
jet valve should be considered if this low-tlow condition occurs frequently and if
15 the duration of the 5-mgd (0.22_ m3/s) tlow is expected to last throughout most of
the winter season.
10

5 BIBLIOGRAPHY

%~~~~~~~~~~~~
60 70 80 90
Andrew, W. G., and Williams, H. B., Applied Instrumentarion in the Process Industries, 2nd
ed., Vols. I and 2, Gulf Publishing, Houston, TX, 1980.
Val ....e disc position (in degrees) )0 Benjes, H. H., et aI., "How to Select Valves for Water and Wastewater," Water Wastewater
Figure 5.3·4 U't . f ' Ellg., 5:62 (May 1968).
m capacity actor (CvID ) versus approximate disk angle of butterfly valve.
Daneker, J. R., "Sizing Buttertly Valves," Parts I and 2. Water Wastes Eng. Ind., 0·10 (July
1970) and E-12 (September 1970).
extra head loss. Although the head loss Ii'om the . d .. . . .
wdl yield a head loss of approximatel I ft Ie ucel IS negligible, the illcreaser Edwards, J. A., "Valves, Pipes, and Fittings," Pollut. Eng., p. 22 (December 1974).
valve is therefore 45 - 14 _ 16 _
. -
f d
6(0.3 111), The plessure drop across the
t or .5 pSi (0.45 kg/cm'):
Hutchinson, J. W.,ISA Handbook of Control Valves, 2nd ed., Instrument Society of America,
Research Triangle Park, NC, 1976.

Cv = (31.5 X 695) (_1_)05 = 8587

6.5 5.4 PIPING SYSTEM
Cva = 14,9
Purpose There are six basic piping systems in a water treatment plant: (I) raw wa-
Th~~isk angle of the valve is 53", which is within the accepted range ter and !inished waste distribution mains; (2) plant yard piping that connects the unit
e pressure drop across the valve fo' th " . .
same as in the preceding trial. I e mmmlUITI flow rate IS practically the processes; (3) plant utility, including Ihe lire hydrant lines; (4) chemical lines; (5) sewer
lines; and (6) miscellaneous pipings, such as drainage and irrigation lines. The cost of
the total plant piping system generally runs 5 to 7% of the total plant construction cost
CV<j> = 972 7 24' = 1.7 depending on the layout, site conditions, and winter temperatures. Careful attention
The valve is open 16". This valve is on the bard' '. . must be paid to the design of the piping system because if the pipes in the main processes
range. Now check the cavitatio . d h erline, but stili wIthin the proper fail, the entire plant will be shut down for an indeterminate length of time.
full throttle position. The hydra nr'nhexd(a) w en the 24-in, (0.6-m) valve is in the
. . UIC ea att h evalveexItlsl5-1 = 14f (4 Considerations The design considerations for the piping system are the func-
dnd the barometnc pressure at an elevation of 1000 ft is 32,8 ft; thus t.3 m)
tion of the specifics of the system. However, all piping systems share a few corn·
Hd = 14 ft + 32.8 ft = 46.8 ft (14.3 111) man issues: The pipe strength must be able to resist internal pressure, handling, and
earth and traffic loads; the pipe characteristics must enable the pipe to withstand
482 DESIGN OF PLANT COMPONENTS
PIPING SYSTEM 483

corrosion and abrasion and expansion and contraction of the pipeline (if the line is
exposed to atmospheric conditions); engineers must select the appropriate pipe sup-
port, bedding, and backfill conditions; the design must accollnt for the potential for
pipe failure at the connection point to the basins due to subsidence of a massive
structure; and the composition of the pipe must not give rise to any adverse effects
on the health of consumers. SingJeType

Raw Water Line Depending on the site specifics, the raw water line may be ei-
ther a gravity flow or a pressure line. Factors that can adversely affect the func-
tioning of the water line are the growth of clams and slime in the interior of pipes;
..... - _---
-- ...
accumulation of sediment within the pipeline; air blockage at high points of the (b)
pipeline; presence of a vacuum condition within the pipeline since this can lead to
Double Type
collapse of the pipeline; creation of a thrust force by the water flow, especially dur- ,
ing flow surges; and the expansion and contraction of exposed pipeline resulting
from changes in ambient temperatures.
Many case histories can be found to illustrate the problem concerning the growth
of aquatic organisms in the raw water lines and how the carrying capacity of such
pipelines can be drastically reduced. If there is a possibility that these conditions
could arise, a temporary or permanent potassium permanganate or chlorine feed sys·
tern may be incorporated into the piping system design, depending on the antici-
pated severity of the problem and DBP formation. ..., .. '
..
"~ ~;::=:.=;.";

Doubte type
••
DnlAr ~apUnr
Since the majority of pipelines seldom have a constant pipeline grade from the
intake to the plant, the air release and air vacuum valves should be installed at the
high points of the pipeline and a blow-off assembly should be provided at the low
points. If they must pass through areas with poor ground conditions, the pipelines
must be extensively supported, for example, by the use of pile supports.
Alternatively, the design may allow an adequate amount of flexibility to occur at
Double t:rpe
the pipe joints.
Whenever a pipeline must cross a geological fault, the design engineer is faced
with additional problems. The recommendations of geotechnical experts must be in-
cluded in the design of the pipeline. Yet there is one basic design rule: The design
should allow the pipeline to undergo a sufficient degree of movement; therefore, Note: .
flexible joints should be installed on either side of the fault and an articulated con- B.C. (bolt circle) and 0.0. (outside dillllCter) shall be sizc4
toensureO.' In. minimum durance around COURlin,.
figuration (Z-shaped pattern) should be used. Figure 5.4-1 presents two types of
flexible joints that can endure a certain degree of pipe displacement. The Dresser (c)
coupling provides for a deflection of approximately 6"; Flex-Tend by EBBA Iron (a)
and TAl-Flex by Taisei-kiko can tolerate a deflection of 15" and an expansion of
Figure 5.4-1 Flexible pipe joints and applications: (a) Tai-Flex joints (courtesy of Taisi
4 in. (10 cm).
Kika); (b) Flex-Tend joint (courtesy of EBBA Iron): (c) Dresser coupling.
Plant Yard Piping The plant yard piping is a low-pressure system. The main de-
sign concerns are the flow velocity in the pipeline, allowable headloss, connections too slow, sludge will settle in the pipelines. The suggested flow velocity for each
to the structures, physical strength of the pipes (in the case of deep installations), connecting pipeline (among treatment unit processes) is listed in Table 3.2.2-1.
corrosion of the pipes, and cost of the system. Flow velocity is an important con- Another important design issue is the pipe connection to the structures and tanks.
sideration because of its effect on the fragile alum floc contained in the process wa- Many pipe breakdowns occur within 6 ft (1.8 m) of the structure wall, due to differ-
ter. If the flow is too fast, the alum floc will be destroyed; if the flow velocity is ential settlement between the buried pipeline and the structure. The solution is either
 PIPING SYSTEM 485
484 DESIGN OF PLANT COMPONENTS

to allow a rotating action to occur at the point where the pipe meets the wall through
the use of a bell and spigot type of joint or to allow the pipeline to flex a couple of de-
Wall
grees by installing one or two pipe couplings in the pipeline adjacent to the wall. -l,:"!..C=.:-Ta p mechanical ioint
Nonetheless, sleeve couplings have a limited degree of allowable deflection; if the an- +,..,..--."..+ locate 2" min.
Paint all exposed st'l
bell lor stud bolt
(typical)
of bell ring w/epoxy • • I, away Irom
gle of deflection is greater than 3 to 4°, water may begin to leak from the pipe. Figure per specs. ,.~. wall reinlorcing st's.
Cast iron pipe
5.4-2 presents several examples of these designs. The Flex-Tend joint (Figure 5.4-1) is .', +-14
pipe manufacturer's
capable of tolerating three types of movement-rotation, deflection, and stretching or std. bell ring with ~.. ,'U!""""'F.!:~-'!Cast iron
contracting-in one assembly, thus eliminating the need for a joint harness assembly. wall flange, see -:4.;liie~ Reinlorced pipe
tabl' below. ...•... concrete Gasket, typical
The yard piping may have to be laid at a depth that is much greater than the stan- pipe
1· Chamfer ::::'5:";:' -J,f""'''''~~t2
c,'
Mechanical ioint cast iron thimble
dard; the standard depth is dictated by vehicle wheel loading. Additionally, the yard Field mortar
pipings are buried in twos and threes in a wide-open trench. In both these cases, the t- Min. Nole:
For JVli~bl, IIn8tlls c".~k
with m,nu'ie!ur,r (dow.
thickness of the pipe is determined by the earth load and not by the internal water Note:
8r,UI sleeve IfIIt.n.11y Americ.n !;fIt iron pipe.
belort pl"in, =el, or tqulll
pressure; thus the engineer must be careful when specifying the pipe thickness. Tap mechanical jOint
Furthermore, the design engineer should specify the proper bedding conditions and Bell rins and wall flange ben for stud bolt
plate thickness (tYPical)',,--.-.:!lI~Ib'l"· Plain end
details regarding backfill to avoid pipeline failure.
Pipe size Steel plate
There are cases when large trench excavations are required, due to a maze of
1,
piping or because the pipes must be laid alongside large buried structures-to which 30" & under
• \-:'...\.."llIch.1Piil Cast iron
they must connect. These types of yard piping conditions should be considered as
wide trench or embankment conditions. Under these circumstances, the pipes may
31" to 50" .1.,
"! .
Follower gland is""
outside wall
-"-""'=- thimble
Gasket
61" & over
be protected from high external loading through the use of three basic methods: •
Reinforced concrete pressure pipe
Mechanical ioint and plain end
Cast iron thimble
(I) reinforced-concrete encasement, (2) reinforced-concrete cradles, and (3) heavy- terminating in wall
walled pipe. Due to the generally congested nature of the yard pipings, the most
common choice is the installation of heavy-walled pipes that are specifically de- 3" Swing chk.
valve, typo
signed for embankment conditions. 3" Ecc. plug
-- .-- - I valve typo
Based on Marston's formula, the external loading resulting from backfill on the
pipe under trench and embankment conditions is as follows. For the trench condition: -r'~"i~~=::':===:(f""'1tr,4 ,. rrli· . ~t!=- Mech. cplg., typo

HWL I ,,~ EI. 1628.50

el. 1628.00 I
16"~ 12"~
// Open end. typo ·1-·
where Wd is the fill load (pounds per linear foot of pipe), Cd the load coetticient (a Inv. 1626.50 • to
Waste washwaler lagoon
function of HlBd and K~; refer to Appendix 19), W the unit weight of till material basin pump
station 24'Sl
(lb/ft 3 ), Bd the width of the trench at the level of the top of the pipe (ft). For the I

embankment conditions:
lWl el. 1609.00 ~
IllJ! J-
I 1 ................ 'nv. 1611.20

where W, is the load per unit length of pipe, C,. the load coetticient (a function of
HlB, and T,dP; refer to Appendix 20), W the unit weight of fill material, B" the ex-
ternal diameter of the pipe eft).
Pipe corrosion may occur internally and/or externally. The use of cement-lined
cast iron and steel pipes provides adequate protection against internal corrosion since
the process water of most water treatment plants is generally not highly corrosive.
However, certain soil characteristics are very corrosive to the pipes. For this reason,
the geotechnical study of the site must include soil resistivity measurements. If the
soil is "hot," the engineer should obtain expert recommendations. In situations where
~=~~
Figure 5.4M2 Typical methods of pipe connections to a structure.
corrosion poses a serious threat, the use of cathodic protection (as a backup to the
protective coatings) should be considered seriously for steel and cast iron pipes.
Since the internal pipe pressure is relatively insignificant for yard pipings, rein-
forced concrete pipes and steel cylinder concrete pipes should also be evaluated
along with the cast iron or cement-lined and cement-coated steel pipes. This rec-
ommendation considers the problem of corrosion, the need for flexibility at the pipe
joints, the pressure imparted by the heavy earth, and the cost of the system.
 486 DESIGN OF PLANT COMPONENTS
Plant Utility and Water Distribution Lines These pipelines are high-pressure lines
that must normally withstand pressures of 75 psi (5.3 kg/cm') or more. The economic
velocity of the distribution main is approximately 5 ft/s (1.5 m/s). However, a flow
velocity of about 7 fils (2.1 m/s) is quite common for utility lines carrying potable
water, as well as utility water (nonpotable) and lines servicing tire hydrants.
There are four specific issues that greatly affect the design of the piping system:
(I) extreme caution must be exercised when installing pipelines near existing sewer
lines, (2) cross connections with the chlorine injector actuation line and other util-
ity lines (not always potable water) must be avoided, (3) a joint harness must be
specitied when using a Dresser type pipe coupling so that the pipe will not slip out
from the coupling, and (4) pipe material that will not degrade the quality of the
potable water must be selected.
In most states, the Department of Health Services requires the water main and
the sewer line to be physically separated. The basic rules set by regulatory agencies
are as follows. The horizontal separation between the water line and the drain or
sewer line should be 10 ft (3 m) if the water main must cross a sewer line or storm
drain. The water line should be laid at an elevation that allows the bottom of the
water main to be 18 in. (0.45 m) above the top of the drain or sewer.
Regulatory agencies also dictate the type of cross connections that may be used
between potable and nonpotable water lines; this includes steam condensate, the cool-
ing water from the engine jacket or other heat exchanger, and pipe connections to
chemical tanks or injectors. If a direct connection is absolutely necessary, an approved
type of backtlow prevention valve must be placed in the pipeline or an air gap of at
least 6 in. (0,15 m) must be provided in all connections to chemical storage tanks.
The installation of surge tanks is often required at the beginning of the distribution
main whenever high-service pumps are used to pump potable water into a lengthy dis-
tribution main. The surge tanks function to protect the pipeline and pumps.
Dresser couplings are commonly used in the high-service pump discharge lines.
Whenever sleeve couplings such as the Dresser coupling are used in high-pressure
lines, the joint harness should be capable of resisting the unbalanced thrusts at the
elbows or to tolerate the line pull caused by, among others, hydraulic thrusts. A me-
chanical type of coupling such as the Victaulic coupling does not require that a har-
ness be installed; however, it does not allow the pipes to undergo much movement.
The recent EPA Drinking Water Quality Standards sct stringent rules with re-
spect to the allowable maximum concentration level of many items. Certain types
of pipe can release undesirable components, such as pipe material or coating mate-
rial, into the potable water. Of major concern are asbestos cement pipes and pipes
coated with bituminous material. Although lead pipes were used primarily in ser-
vicing residential homes in the past, engineers should be aware that small lead pipes
pose a threat to the health and safety of the public. All new pipe designs should
avoid specifying the aforementioned types of pipe when establishing the potable wa-
ter supply line.
Chemical Feed Lines The main concerns surrounding chemical feed lines are
corrosion, abrasion, formation of scale, and heat resistance. When plastic pipes are
exposed to the sun, they can potentially deteriorate as a result of ultraviolet rays and
PIPING SYSTEM 487
weathering. Moreover, these pipes should have a proper type of pipe suppon to min-
imize pipe sagging between the supports, especially when plastic pipes are used.
Selection of an appropriate type of pipe material depends on the type and solution
strength of each chemical that will tlow through the pipe; refer to Table 4. I -I and
Appendix 13. Any chemical that is fed in slurry form requires a minimum tlow ve-
locity of 5 ft/s (1.5 m/s), and if the chemical feed line is exposed to the atmosphere,
the pipes must be insulated from freezing and direct sunlight; certain chemical solu-
tions tend to crystallize when the ambient temperature is well above ::nop (0° C). When
employing scale-forming chemical solutions, the pipe design must have a minimum
number of sharp turns and should provide cleanouts and a water tlushing line.
If acid, alkali, oxidizing, and organic compound chemical lines are installed in
a pipe chase, double-walled chemical feed pipes must be used, as dictated by local
OSHA regulations or other codes. This practice will minimize the potential for con-
tamination of soil and fire or explosion in the event that a pipe ruptures. Similarly,
from a safety standpoint, chlorine should be fed to the application point as chlorine
gas (under negative pressure) rather than as a strong solution under pressure, unless
the feed line is short and double-walled pipe is used.
Miscellaneous Pipings Included in the category of miscellaneous pipings are
sewer, drain, storm drain, and irrigation pipelines. If these pipelines are designed to
enter any building, the uniform plumbing codes must be adhered to strictly. Should
these pipelines be laid outside, the following materials may be used, depending on
the conditions of the site: cast iron, salt-glazed vitritied clay, concrete, or PVC. With
respect to sewer lines, the standard sewer line design should be followed when es-
tablishing details such as the size, tlow velocity, minimum slopes, and access man-
hole requirement (refer to the Ten States Standards for Sewage Works).
Piping Materials The most commonly used materials for water supply arc cast
iron, steel, concrete, and PVC, In the past, asbestos cement pipes were employed
extensively; however, in June 1989 health concerns over the detrimental effects of
the asbestos fibers caused the EPA to ban their use.
Cast iron pipes have a certain resistance to corrosion, The standard sizes range
from 2 to 24 in. (5 to 60 cm) in diameter. Most larger cast iron pipes are cement
lined. The Hazen-William C value may be as high as 145 for new pipes larger than
8 in. (0.2 m) in diameter. The disadvantages of this type of pipe are its slightly brit-
tle nature, its heaviness, and its rather high cost, particularly for large pipes. Yet
ductile cast iron pipes may be manufactured as large as 78 in. (2 m) in diameter due
to their ductility. However, the length of large pipes must be limited to reduce the
weight of the pipes, and this is achieved by using a large number of pipe joints.
Mechanical joints are generally used in these types of circumstances rather than the
bell and spigot and the tlange joints normally associated with cast iron pipes
(ANSI/AWWA CliO to C151-86 for cast iron and ductile iron pipes and litting).
Steel pipes are commonly used when large and relatively small pipelines as well as
nonstandard large fittings are required. In most cases, both the interior and exterior of
these pipes are lined with cement for protection against corrosion. Both longitudinal and
spiral welded steel pipes are available for use. The joints betwecn the pipes are usually

488 DESIGN OF PLANT COMPONENTS
constructed by welding or through the use of a bell and spigot with rubber gasket, sealed
flanges, sleeve couplings, or mechanical couplings. The advantages of steel pipe inclUde
thinner walls, lighter weight, and in general, lower cost when large-diameter pipes are
specified. Yet this type of pipe also has several disadvantages: its inability to carry high
external loads, its tendency to collapse when subjected to negative pressures, and the
additional expense of providing corrosion protection measures such as cathodic protec-
tion (ANSIJAWWA C200 to C213-85, steel water pipe standards).
Concrete pipes are available in various forms: precast, reinforced concrete, pre-
stressed with steel bars, and steel cylinder pipes. These pipes have low maintenance
costs, are resistant to corrosion under normal conditions, and are capable of with-
standing external loads. However, the concrete pipes also have some undesirable
features, such as a short pipe length due to the weight, free lime being leached from
the concrete, tendency to crack under pressure and cause leaking, permeability of
the concrete, and corrosion in the presence of strong acids (ASTM C 14, C76, C361
and AWWA C300 to C303, specifications for concrete to modi tied concrete pipes).
Plastic pipes are used primarily as chemical solution, drain, sewer, and irrigation
lines. The standard specifications for plastic and glass fiber-reinforced pipes may be
found in ANSIJAWWA C900 to C950-81, and the thickness design may be found in
the ANSIJAWWA standards piping handbook or the reference data andlor catalogs
provided by the pipe manufacturer. The recommended pipe trench dimensions and
details concerning pipe support, bedding, compact till around the pipes, and backfill
may also be found in the aforementioned publications and various other publications.
BIBLIOGRAPHY
ACPA, Concrete Pipe Design Mallual, American Concrete Pipe Association, Vienna, VA,
1987.
AWWA, PVC Pipe: Design and Installation, Manual M-23, American Water Works
Association, Denver, CO, 1980.
AWWA, Steel Pipe: A Guide for Design and Installation, Manual M-II, American Water
Works Association, Denver, CO, 1985.
Bealey, M., and Lemons, J. D., eds., Concrete Pipe alld the Soil-Structllre System, STP 630,
American Society for Testing and Materials, Philadelphia, 1977.
IAPMO, Uniform Plumbing Code, 18th ed., International Association of Plumbing and
Mechanical Officials, Los Angeles, 1988.
King, R. c., and Crocker, S., Piping Handbook, 5th ed., McGraw-Hill. New York, 1973.
5.5 STANDBY POWER SUPPLY
Purpose All water treatment plants should be operated in a reliable and cost-
effective manner. Power outages will undoubtedly occur for a number of reasons,
including damage to the power generation system or power supply system as the
result of accidcnts such as earthquakes, fires, floods, tornadoes, hurricanes, light-
ning, or even sabotage. Certain equipment, those of the monitoring and control sys-
STANDBY POWER SUPPLY 489
terns, require a continuous power supply. This may be generated by a standby emer-
gency power supply unit. .
The standby unit also offers a monetary benefit for water treatment plants uSlOg
high-service pumping stations with large electric motors. Variations in water demand
affect pumping requirements, which in turn affect the power demand. Since the rate
structure or amount the utility charges for power is based in part on anticipated peak
demand, a standby power unit may be used to reduce the base rate by peak shaving.
Considerations
Power Outage 11 is important to investigate the reliability of the local power sup-
ply system over a period of at least 5 to I 0 years. Issues that should be studied prior
to the design work are the frequency of the power outages and the duratIOn of each
episode. The type and scope of the standby power supply system are dependent on
several factors: (I) the frequency of power outages, (2) the duration of outages,
(3) the amount of essential equipment, (4) the total power load required by the es-
sential equipment, (5) the types of fuel available for the power generator, (6) fuel
cost, and (7) the availability of alternative power supplies from other substations in
the vicinity of the plant.
Water treatment plants using raw water pumps will have a different-sized standby
power supply system than that of plants receiving raw water from the source via
gravity. The standby power system is installed in substantially greater numbers for
the latter type of treatment plant. This is because the raw water keeps flowing into
the plant regardless of whether the power is on or off. Thus the Ireatment process
units must be capable of processing the water based on standby power. Although the
plant flow rate and power consumption can be reduced, the plant flow may not be
allowed to be reduced if the water source is shared by more than two or three users,
even in the event of a power failure. Such a reduction would result in a significant
increase in pressure or a massive overflow at the transmission line. In most cases,
the daily flow cannot be changed without giving advanced 24-hr notice to the agency
controlling the source. For any given situation, the minimum capacity of the standby
unit should be capable of providing a power supply that will operate 25% of the de-
signcd total capacity of the plant. Yet specitic local conditions as well as the char-
acteristics of the water supply system can dictate the minimum supply quantity.
The following is a list of equipment whose functions must essentially be main-
tained during a power outage:
I. All equipment associated with the instrumentation system and the major units
of the control system.
2. The computer system; all data stored in the RAM of the system will be lost
during the time between the loss of power and the startup of the standby
generator; thus it is crucial to have an uninterrupted power supply.
3. The emergency lighting of all site structures.
4. Electrically operated valves.
5. Major chemical feeders such as coagulant and disinfectant feeders.
 490 DESIGN OF PLANT COMPONENTS
6. Chlorinator booster pumps, leak detectors, and emergency fans.
7. Flash mixer.
8. Twenty-five to 50% of the mechanical flocculators.
9. Sampling pumps.
10. Sump pumps.
11. A designated number of raw water and treated water pumps; less than 50%
of the units (for each type) must generally be functioning.
12. Certain laboratory equipment: refrigerators, incubators, compressors, vac-
uum pumps, room lights, and other essential analyzers.
13. The hot-water circulating pumps of the heating system.
Depending on the duration of the power outage, other major process units, such
as the filt~r backw~sh equipment, mayor may not require the services of a standby
power umt. If a rehable power supply system can be obtained from other electrical
substations in the vicinity, the treatment plant may opt to use a dual power supply
system rather than a standby generator.
Peak Power Shaving The economic evaluation of the various peak shaving al-
t~rnatlves IS Influenced by the rate structures of the utility companies. The four ba-
SIC treatment plant operational modes used to achieve peak shaving arc:
I. Load shedding. Significant load shedding may be accomplished by reducing
the load through the implementation of a computer or microprocessor-based
controller.
2. Reduction of the peak utility demand. This method involves the transfer of
part of the power load from the utility to the engine generator supply during
perrods of high electricity demand (peak).
3. Paralleling with utility. In this scenario, the power generator is connected di~
rectly to the normal distribution system. Energy required during peak power
consumption is supplied directly to the system and the amount of power fur-
nished is determined by the engine governor control.
4. Prime power cogeneration. A prime power system is a system in which the
on-site generator is the prime source of power. Only certain dedicated loads
may be connected to the on-site generator, and the utility may be used as a
standby source through the use of an automatic transfer switch.
Natural gas or diesel engine generators are generally used to supply the standby
power ~ecessary to maintain the essential equipment during power outages; under
these Circumstances, the second peak shaving method is normally considered. The
startup and shutdown functions of the engine generator should be fully automatic.
However, a time delay of approximately 60 s should be provided before the gener-
ator r~aches the. pro~er speed, to differentiate extremely short-term power outages
resultmg from hghtnmg and other causes from long-term power failures.
CHAPTER 6
Elements of Detailed Designs
6.1 PLANT HYDRAULICS
Purpose One of the major design efforts of a water treatment plant is the in-plant
hydraulics. Since each water treatment unit process is functionally dependent on hy-
draulic factors, a hydraulic analysis must be conducted during the design phase be-
cause it plays a crucial role in assuring overall plant function and the efficiency of
each treatment process.
Considerations Obtaining an even Ilow distribution to parallel process units is
an extremely important hydraulic consideration. It is rather amazing to find that a
number of existing treatment plants have problems with uneven /low distribution:
among sedimentation basins, between the treatment process modules, and even
among the filters. This has resulted in the presence of significant variations in flow
distribution. Differences as high as 47% have been observed among process units.
A sample computation for sizing the inlet channel may be found in Example I of
Section 3.2.3 and "Choice of Flocculation Tank Inlets" in Section 3.2.4.
Another common hydraulic problem is the occurrence of severe flow short-
circuiting in the tanks of process units such as flocculation and sedimentation as well as
the clearwel1 and disinfection contact tank. Flow short-circuiting in sedimentation tanks
has been studied extensively. These studies show that only 20 to 30% of the mean de-
tention time (tank volume divided by the flow rate) may be considered to be the actual
flow-through time. Figure 3.2.5-3 and Appendix 9 present some examples of flow short-
circuiting resulting from density /low and improper design of the inlet diffuser wall.
Many college engineering courses teach that selection of the pipings connecting two
separate tanks should be based on the most economical size that is capable of carry-
ing a particular flow rate from point A to point B using computer analysis. However,
practical and operational water treatment experience indicates that the proper size is
determined by two factors: the required /low velocity, as dictated by the process, and
the extreme operational conditions of the plant. For instance, all pipes or channels de-
signed to transport floc to the sedimentation tank must be sized neither to break the
fragile /loc nor allow floc settling. If the pipe is used to carry sludge, a minimum /low
velocity of 2 ftls (0.6 m/s) must be maintained to prevent the sludge from settling.
Another hydraulic consideration is the maximum and minimum plant flow rate.
Should one side of a module be taken off line, the remaining side is required to
491
492 ELEMENTS OF DETAILED DESIGNS PLANT HYDRAULICS 493

carry nearly twice the designed Ilow rate and an extremely high flow will occur.
Thus all pipes and channels should generally be sized at 150% of the designed flow
rate (maximum daily tlow rate). If this scheme is followed, the output of the treat-
ment plant may easily be increased by 50% (i.e., filtration rates of 4 to 6 gpm/ft2)
at some future date. The raw water main and the main distribution lines are gener-
ally sized to handle future plant flow rate, in anticipation of plant capacity expan-
sion after construction of the original plant.
The required available headloss across an ordinary conventional water treatment
plant averages approximately 16.5 ft (5 m). However, an advanced treatment plant with
prcozonation, postozonalion, and GAC adsorption processes requires an available head
of approximately 25 ft (7.5 m): available head is defined as the difference between the
water surface elevation in the tank at the head of the plant where the plant influent
flows in, and the high water level in the clearwell at the designed plant flow rate.
The tlow hydraulics through granular media beds, including tilter beds, have been
studied cxtensively and may be found in various textbooks and tcchnical journals.
However, it is important to note that most inexperienced design engineers tend to spend
many hours calculating the headloss through a filter bed at various filtration rates and
water temperatures, determining the results up to two decimal points. Yet this entire
process is quite unnecessary; the engineer actually only needs to calculate a conserv-
ative approximate figure. which may be assumed from similar existing filter beds or
pilot study data. Furthermore, design engineers should realize that the headloss across
the tilter is a function of many factors that come into play after the media is in place:
changes in media gradation aftcr numerous backwashings. changes in media size due
to atlrition loss and the loss of the fine grains, formation of mud balls within the bed,
and changes in the depth of the filter bed resulting from media loss.
Generally, the tiltration cycle requires a lilter design with over 8 ft (2.4 m) of
available headloss across the bed. Consequently, a difference of ±3 in. (75 mm) in
the initial headloss does not have a significant impact on the tiltration process from
an operational standpoint. The same analogy may be applied to computations on
backwash hydraulics and tilter bed expansion. The calculation of head loss through
the tilter underdrain orifices is another example where practical experience saves
the expenditure of a lot of unnecessary time and effort. The engineer should not
spend time evaluating the orifice coefficient when performing hydraulic calculations
because most filter underdrains are covered with a gravel bed. Consequently, over
30% of the orifices do not have a clean orifice area. Thus the most conservative co-
eflicient (0.65) should be used for the calculations.
The hydraulics of water troughs, gullets, channels, conduits, oritices, weirs, pipes,
20-in. bUllerfly valve are located just upstream of the head of the plant. Each tloc-
culation tank has two 14-in. (0.36-m) bullertly valves (as the inlet), and each of the
two bartle walls (one per tank) produces a O.I-ft (0.3-cl11) headloss acroSS the wall.
Each sedimentation tank is equipped with two launders that arc 35 ft (10.7 m) long
with a grand total of 280 V-notches (90°) to decant thc sellied water. Each lilter has
an inlet weir that is 4 ft (1.2 m) in width and a 14-in. (0.36-m) inlet bultertly valve
as shown in Figure 6.1-1. The required water elevation in the tilters is elevation 100.
Determine The hydraulic grade line between the tilter and the inlet to the 20-in.
3
(0.5-m) Venturi tlow meter at a flow rate of 10 mgd (0.44 m /s)
Solution Step 1. Determine the water level control points in the process train. In this
case, the filter inlet weir and the V-notched weirs in the sedimentation tanks are the level
control points. Both weirs should be set at an elevation that will allow a free-fall con-
dition (not submerged) to be maintained, resulting in a uniform flow rate at each weir.
Step 2. Compute the headloss or required head over the weirs between the level
control points. Use a computation sheet (tabulation form as shown here) so that the
values may easily be checked. Refer to Appendix 14 for the proper K factors and
the basic hydraulic formulas that arc applicable to this problem.
(a) Between the Filters and Settled Water Channel Refer to Table 6.1-1 a for the
calculations.
(b) Between the Settled Water Condllit and the Head of the Plalll The proper
size of each launder in the sedimentation tank can be determined quickly by using
Figure 3.2.7-16. Since each launder carries 2.5 mgd (1740 gpm), the width should
be 1.5 ft (0.46 m) and the height should be specitied as 1.67 ft (0.5 m). Set the bot-
tom elevation of all launders at an elevation of 104.75. Each V-notch on the laun-
ders should have a free discharge flow. Thus the elevation of the V-notch should be
set to give 106.50 water elevation in the tank. Since each tank has a total of 280
of a 3-inch height V-notched weirs, the tlow rate of each 90° V-notched weir is
3
(10 X 695) 7 560 = 12.4 gpm or 0.0277 fl /s
Since 0.0277 ft 3/s = 2.54H2.5, H = 0.164 ft or 1.97 in.
Inlet Weir
EI. 108.54 EI. 107.08 €I. 10\.00
EI. 106.50 EI. 100
90" V-notch Weir (given)

/
EI. 10\.44
and titting are not discussed in this book because they may be found in a number EL 104.41
I
of other textbooks.
I
'lr-fr--+O
,- 1-.35~
Example Design Calculations
\ Flocculation tank Sedimentation tank
~ 20 in.
Example 1

Given A 10-mgd (0.44-m 3/s) water treatment plant has two tlocculation/sedimen-
/
o 20 in. Venturi
~ 20 in. B.V. L=2oott
Filter -
tation tanks in parallel and four tilters. A 20-in. (0.5-m) venturi flow meter and a Figure 6.1·1 Process train and hydraulic grade line.
 ...
~

TABLE 6.1. 1 Calculations for Example 1

(a) Between the Filters and the Seuled Water Conduit

Item Friction Loss

Q Diameter v 'v2l2g K Water
Water level in L IJ.h L.O.H. Elevation
the fil ter cell
14 in. BV 100
2.5 mgd 14 in. 3.62
Inlet hole 2.5 mgd 0.2 0.35
14 in. 3.62 0.2 0.07 ft
1.6 100.D7
Inlet weir* (C = 0.8) 0.32 f' 100.39
Filter inlet 2.5 mgd = 3.97 cfs = 3.33 X 4 H I.'. where H = 0.45 ft
channel 0.45 ft 10 1.44
20 in. pipe ex it Neg ligible 10 1.44
10 mgd 20 in.
Pipeline 7.09 0. 78
10 mgd 20 in.
1.0
0.78 ft 102.22
200 ft 0.9 ftl lOO ft
Pipe inlet 1.8 ft 104.02
10 mgd 20 in. (C = 120)
Water level in [he 7.09 0.78 0.5
settled water 0.39 ft 104.4 1
conduit Negligible 104.4 1

*Sel the crest of the weir at 10 1.00 to ensure a free discharge condition .

- ~

(b) Between the Seuled Warer Conduit and the Head oj rile Plant

Water
hem Q Diameter v v 212g K L O.H. El evati on
Water level in the settl ed 104.41
water conduit
Water level over the V· notched 0. 164 f' 106.50
weirs· on the launders
Sedimentation lank Negligible 106.50
Diffuser wall (sedimentatio n 0.1 f' (give n) 106.60
lank inl et)
Baffle wall (fl occul ation tank) 0. 1 f' (given) 106.70
Inle' BY 2.5 mgd 14 in. 3.62 0.2 0.3 0.06 ft 106.76
Inlet hole 2.5 mgd 14 in. 3.62 0.2 1.6 0.32 f' 107.08
Inlet channel Negli gible 107.08
*The crest of the V-notch is set at an elevation of 106.34.

(c) In let of the Venturi Meter to the Inlet Challne/to the Flocculation Tanks
Friction Loss
Water
2
Item Q Diameter v v 12g K L M L.O.H. Elevati on
Water level in 107.08
the inlet cell
20 in. Pipe exit 10 mgd 20 in. 7. 1 fps 0.78 f' 1.0 0.78 ft 107.86
20 in. BY IO mgd 20 in. 7. 1 fps 0.78 f' 0.25 0.20 f' 108.06
(full open)
20 in. Venturi, 10 mgd 20 in. 0.34 [, 108.40
Pipeline 10 mgd 20 in. 15 ft 0.9 ftIlOO f' 0. 14 108.54
...
co
(C = 120)

'" tThe Venturi differential is 80 in. (6.67 ft) at 10 mgd and the head loss across the Venturi is approximately 5% of the differential.
 496 ELEMENTS OF DETAILED DESIGNS
SPECIFICATIONS 497

(c) Inlet of the Venturi Meter to the Inlet Chaunel to Floccillation Tanks. See Miller, D. S., Internal Flow Systems, Vol. 5 in the BHRA Auid Engineering Series, BHRA
Table 6.1-lc. Fluid Engineering, London, 1978.
Refer to earlier chapters for examples of hydraulic calculations on baffled walls Montgomery, 1. M., Consulting Engineers, Water Treatment: Principles and Design, Wiley,
in basins, density flow and flow stability (Froude number) of sedimentation tanks, New York, 1985.
and the basic hydraulics of filter structures. An example of the hydraulic grade line Shaw, O. v., and Loomis, A. W., Cameron Hydraulic Data, Compressed Air Magazine Co.,
across a plant is illustrated in Figure 6.1-2. Ingersoll-Rand, Phillipsburg, NJ, 1965.
Yao, K. M., "Hydraulic Control for Flow Distribution," Proc. Sanitary Eng. Div. ASCE,
2:275(April 1972).
BIBLIOGRAPHY

Benefield, L. D., et a!., Treatment Plallt HydrauUcs for ElIl'irol1mellfaf Engineers, Prentice 6.2 SPECIFICATIONS
Hall, Englewood Cliffs, NJ, 1984.
Chao, 1. L. and Trussell, R. R, "Hydraulic Design of Flow Distribution Channels," 1. Environ.
t Clear and precise specifications containing drawings are a mandatory part of any
EII8· Div. ASCE, 106:2:321 (February 1980). modern project. Vague or erroneous specifications result in inferior-quality prod-
Davis, C. v.. Handbook of Applied Hydraulics, 3rd ed., McGraw-Hili. New York, 1970. ucts, in addition to a large number of change orders and delays in the project.
Hudson, J. E., Water Clarification Processes: Practical Desigll alld Evaluatioll, Van Nostrand Moreover, they are a potential source of lawsuits. General guidance on the various
Reinhold, New York, 198 I. types of engineering specifications are available from several agencies, including
Huisman, L.. Sedimentation alld Floatation, Delft University of Technology, Delft, The the Construction Specification Institute, American Consulting Engineers Council,
Netherlands, 1973. American Society of Civil Engineers, National Society of Professional Engineers,
Hydraulic Institute, Engineering Data Book, Hydraulic Institute, Cleveland, OH, 1979. American Institute of Architects, and the General Services Administration.
Karassick, I. J., et aI., Pump Handbook, McGraw-Hili, New York, 1976. Project engineers should write the specifications in a manner that is both clear
King, H. W., Krutzsch, W. c., Fraser, W. H., and Messina, 1. p, Hal/dbook of Hydraulics, and convenient for the contractors since the specifications are written for the con-
McGraw-HilI, New York, 1976. tractors who must furnish all the materials and equipment, build the structures, per-
form the civil works, provide and assemble the components, and finish the electri-
cal work. For instance, a system such as ozonation or chlorination contains many
/ "., .. """""", subsystems; thus the system specifications should include all the components and
""'-"'" subsystems under one heading. This frees the contractor from having to search
..",j . . ".." ..... _ ....... through several other divisions and sections to find the specifications for each sub-
....... -,.".... ,""',."
system. If the specifications are not organized in this manner, the contractor will
waste precious time when formulating a bidding price for the project and may even
overlook some components and subsystems.
Finally, project engineers must be aware that the specifications of certain pro-
.""4il&r~."_'T_~ ::J:::;-::::~)""'"
.. ..,- , jects must be tailored to meet the particular needs of the system. Although the mas-
-.---------rr.~".~~
.," <oo"",-~,~,~. '. . ~, ........ ,t""'."
,~ ,..,~
..... ,"--_--- -.--
-""""_.j=
~"

.--_ .. ___ - '095

ter and standard specifications certainly save time, minor adjustments are always re-
quired in most treatment plant constructions .
Mail! process hydraulic profii9 ,;125 MGO
-""-
Contents of Specifications The specifications are generally comprised of two
major parts: the administrative requirements and the detailed requirements of the
work to be performed. Some major items that must be considered by the engineer
are the size of the project, the duration of construction, and the nature of the pro-
ject-whether it is a privately or publicly financed project. The outline of the spec-
ifications should contain the following items:

• Bid advertisement (invitation to bid)

Figure 6.1a2 Example of hydraulic profile. • Information to bidders

•
498 ELEMENTS OF DETAILED DESIGNS
Form of the proposal
Form of the contract agreement
Bond forms
Contract revisions
Standard specifications
Master specifications
Special provisions/conditions
General provisions
Technical specifications
Bid Advertisement It is standard practice for public agencies to publicly adver-
tise all public works projects. Sealed bids arc received through advertising in vari-
OllS news media for a required period of time. After the bids arc opened, publicly
read aloud, tabulated, and evaluated, the lowest bidder is selected.
Information to Bidders It is customary to issue a set of drawings and specifica-
tions to all prospective bidders for a reasonable charge.
Form of the Proposal In most cases, the proposals are accompanied by a guar-
antee in the form of a certified check or surety bond. This practice ensures that the
successful bidder will cnter into a contract; the guarantee is returned when the bid
is awarded. However, if the lowest bidder fails to execute the contract, the certified
check is forfeited as liquidated damages and the obligations of the surely (under the
bond) are enforced as compensation to the owner for the cost of awarding the con-
tract to the next bidder or for the additional cost of re-advertising. If the project is
financed by private owners, it is not necessary to obtain competitive bids, and ne-
gotiations are often conducted with selected contractors.
Types of Contracts For both public and privately linanced works, the two ba-
sic types of contract are the unit process and lump sum. Negotiated and special con-
tracts are also possible.
Unit Price Contract The unit price contract should be used when it is impossi-
ble to define the exact limits of the various work items encompassed in the contract
and/or bid documents. For payment purposes, the work is broken down into major
segments with respect to lhe type of work and trade involved. Each segment is des-
ignated as a payment item (estimated quantities) and is listed in the proposal. Bidders
are requested to submit a written bid for each unit. For example, a specific number
of cubic yards of concrete will be bid at a certain unit price per cubic yard. The to-
tal bid is the sum (in dollars) of all items listed in the proposal and is the basis by
which all received bids are compared to determine the lowest bidder. Payment to
the contractor is determined by the actual quantity measured for each item incor-
porated into the work at the contract unit price.
SPECIFICATIONS 499
Lump-Sum Contract The lump-sum contract is selected when it is possible to
accurately delineate, through drawings, the limits of the work encompassed in the
contract. With this type of contract, the bidder makes quantity takeoffs and presents
them as the basis for the bid. However, it is essential that the bidder be provided
with drawings and specifications that are extremely detailed with respect to all fea-
tures and requirements of the work. Payment is made in accordance with the lump-
sum bid to cover all work and services required by the drawings and specifications.
Negotiated Contract Private work contracts often take the form of negotiated
contracts. Several different payment methods are available: lump sum or unit price,
fixed fee plus reimbursable costs (with a ceiling price), fixed fee plus reimbursable
costs (no ceiling), reimbursable costs plus a percentage of the cost, and a construc-
tion-management contract. Furthermore, incentives such as a bonus for economic
construction and/or completion ahead of schedule may be added. Conversely, the
contractor is penalized for inefficiency and/or failure to meet the completion date.
Special Contract The use of a special contract may be applicable in specific cir-
cumstances such as the procurement and installation of highly specialized equip-
ment. In this case, the owner simply invites proposals from a select group of pre-
qualified contractors without advertising the project. However, contracting
procedures are generally followed for contracts procured through public bidding.
Contract Revisions There are four basic types of contract revision: (I) addenda,
(2) stipulations, (3) change orders, and (4) supplementary agreements.
Addenda Revisions to contract documents, made before bids are received from
contractors, are classitied as addenda. The revisions are usually changes to the con-
tract drawings and specifications resulting from errors, omissions, or the need for
clarilication. Bidders should be notified of any changes in the bid opening dates by
means of an addendum. All addenda should be delivered sufficiently in advance of
the bid opening date to all persons who have picked up the bid documents so that
they may adjust their proposals. In turn, the bidders must acknowledge receipt of
all addenda; otherwise, their bids will not be accepted.
Stipulations Stipulations are issued to the bidder whcn the contract is executed.
They contain modifications of the contract terms as proposed by the owner.
Change Orders This is a written order to the contractor authorizing a change in
work from the original drawings and specifications. It usually increases or decreases
the quantity of work to be performed. Thus a payment adjustment should be made
accordingly. The change order must be approved by the owner and signed by the
engineer. Good design work will have a total adjusted payment of only 2 to 3% of
the total contract price resulting from change orders. If the total cost of the changes
exceeds a specified percentage (such as 20%) of the total contract price, a supple-
mentary agreement, acceptable to both the owner and the contractor, should be ex-
ecuted before proceeding with the affected work.
 500 ELEMENTS OF DETAILED DESIGNS
Supplementary Agreements This written agreement encompasses all modifying
work that is beyond the scope and terms of the contract or work changes that are
within the scope of the contract yet exceed a set percentage of the original price. The
supplementary agreement must be signed by both the owner and the contractor.
Standard Specifications The standard specifications are often preestablished
by public agencies to ensure uniformity in administrative procedures and in the qual-
ity of the facilities. These criteria list the specific requirements for all materials and
work quality. The specifications are updated or amended periodically and published.
Although in some instances the standard specifications may be used without any
changes, they generally require minor modifications and additions. The assembled
moditications and additions are referred to as supplementary specifications, special
previsions, or special conditions.
Master Specifications Many consulting engineering firms provide master spec-
ifications that are used for projects owned by either private or local agencies. These
specifications are written in-house and pertain to general items. When the master
specifications are applied to a project, the engineer should delete requirements that
are irrelevant to the particular project. This practice helps to reduce the time re-
quired to write the contract specifications, as well as serving as a checklist, thereby
minimizing errors and omissions.
General Provisions The general provisions define the rights and responsibili-
ties of both the owner and the contractor to the construction contract, surety, and
requirements, as well as delineating the legal relationships and the authority of the
engineer. The general provisions are usually comprised of the following sections:
Definitions and abbreviations
Bidding requirements
Contract and subcontract procedures
Scope of the work
• Control of the work
Legal and public relations
Damage claims
Workers' compensation insurance
Comprehensive general liability of the contractor
• Laws, ordinances, and regulations
Responsibility for work
• Explosives
Sanitary provisions
Public safety and convenience
• Property damage
SPECIFICATIONS 501
Public utilities
Prosecution and progress
Time of completion
• Suspension of work
• Unavoidable delays
• Annulment and default of contract
• Liquidated damages
Time extensions
Measurement and payment
Payment
Termination of contractor responsibility
Guaranty against defective work
These sections are often referred to as boilerplate.
Technical Specifications The common types of technical specifications per-
tain to (I) materials and work quality, (2) material procurement, and (3) perfor-
mance. Each of these is discussed.
Materials and Work Quality Specifications The material and work quality spec-
ifications are included in most construction projects. They cover general and specific
conditions related to the performance of the work, material requirements, construc-
tion details, quantity measurements, and the criteria used for determining payment.
Materials Procurement Specifications These are usually written for special sit-
uations such as accelerated construction schedules, for projects requiring the use of
materials that have a long lead time, and for projects involving several general con-
struction contracts. This type of technical specification ensures that the materials
are available to the job site on time. Furthermore, it ensures uniformity in the ma-
terials. The technical specifications should clearly delineate all elements concern-
ing materials, work quality, and the mode and time of delivery. Examples are filter
media (i.e., anthracite and filter sand), filter underdrains, special valves, and certain
precast concrete items.
Performance Specifications The performance specifications are divided into two
categories. One type specifies the water treatment unit process or an entire water
treatment plant. These generally include a basic site plan, the maximum required
hydraulic loading rate for each unit process for the product water, a guaranteed wa-
ter quality limit (required) from each unit process, the time of completion, and other
miscellaneous basic requirements. This type of performance specification is com-
monly employed in most European and Asian countries.
The second type of performance specification concerns the prepurchase of equip-
ment by the owner prior to signing a construction contract; this is often practiced in
 502 ELEMENTS OF DETAILED DESIGNS
the United States. Certain types of equipment must be prepurchased due to their long
lead time or because the owner IS determined to obtain a particular type of equipment
that ~ay not be purchased through competitive bidding procedures. The performance
specificatIOns should therefore deline the quality, function, and all rcquirements pcr-
tammg to the water treatment equipment, in addition to the required tests, atlidavits,
and other supporting evidence of compliance. It is vital that the test conditions and
procedures be clearly defined to prevent conflict and/or future lawsuits.
Construction Specification Institute (CSI) The purpose of the CSI is to or-
g~nize and improve construction specifications. It was established by architects, en~
glneers, and certain manufacturers. The CSI recommends that companies use the
CSI format for all water treatment projects. Their format consists of 16 divisions
with several sections per division pertaining to specific construction activities. Th~
CSI format is widely issued in the construction of buildings and public works and
for product specifications released by manufacturcrs. The CSI also publishes a
M~l1ual of Prac~ice, which contains a great deal of information on writing specifi-
catlo?s. In.structl~n. documents and project manuals such as Specification Language,
Spec'fi~atlOn Wn~m~ and P,:oduction, Pe,formance Specifications, Methods of
Specifymg, and Blddmg ReqUIrements are valuable reference items for all project
engineers.
BIBLIOGRAPHY
Abdallah, E. T., Preparing Specifications for Design-Bid-Build Project, ASCE Manual Jan
2000. ' ble regulations with respect to taxes, import duties, port and freight charges, currency
Goldbl~om, 1., and White, J. J., "Specitications," in Section 3 of Standard Handbook/or Civil
Englfleers, 3rd ed., F. S. Merritt, ed., McGraw-Hili, New York, 1983.
White, 1. 1., "Specifications," in Section 3 of Standard Handbook/or Civil Engineers, Hicks,
T. G. editor. McGraW-Hili, New York, 1972.
6.3 FINAL COST ESTIMATES
~o~t estimat~ons for a project is generally required in three stages: feasibility, pre-
hmlllary deSign, and the design completion phase. Each of these phases would be
expected to have a respective level of accuracy of ±45, ±30, and ±7%. Cost esti-
~ates conduct~d during the early stages of a project are often used to prepare fund-
~ng for the project budget through the use of cost estimation curves and by adjust-
mg the cost to a geographical area as described in Section 2.4. In this section we
discuss the final cost estimate for bid evaluation.
!he ~nal cost estimate is not merely a simple summary of the anticipated costs of
vanous Items based on quantity takeoffs, extensive cost data, and statistics. Many
other factors are an integral part of the linal cost estimate: the complexity of the pro-
Je~t, site and climatic conditions, job site locations, judgment of the cost estimator
WIth respect to economic trends, dynamism of the construction industry, and the ex-
FINAL COST ESTIMATES 503
tent to which the project attracts contractors. Professional cost estimators generally
require 2 to 5 work-hours per sheet of construction drawings to complete an estimate.
If the project engineer is required to make the final construction estimate, he or
she should be aware of three characteristic weaknesses of most engineers: (I) the
tendency to devote too much time to the design aspects and not enough to cost es-
timation, (2) the tendency to focus on direct costs and to neglect the necessary in-
direct costs incurred by the contractor, and (3) the omission or insufficient allotment
of funds to cover specific business costs-risks, escalation, and legal requirements-
incurred by the contractor.
Development of Strategy and Data The first step in preparing a construction
cost estimate is to develop the strategy and data used in the estimating process. One
method involves dividing this task into four basic components: (I) site conditions,
(2) classification of the work and quantity takeoffs, (3) construction program, and
(4) unit cost development.
Site Conditions The topographical, geological, cultural, climatic, and socioeco-
nomic conditions of the site all affecl project cost directly. Site conditions such as
the topography, nature of the soil, physical characteristics of the soil, underground
water table, construction costs, the source of the imported materials, and the avail-
able disposal sites (for the excess soil) relate directly to the total design and con-
struction costs. Furthermore, items such as site access, working conditions, pro-
ductivity of the local labor, and cost of materials and labor (in the vicinity of the
site area) should not be overlooked.
If the project is in a foreign country, the cost estimator should review all applica-
conversion rates, local inflation rate, and limitations on money transferred between
countries. Other factors, such as specific laws or regulations governing construction
by foreign contractors and the employment of local labor, must also be evaluated.
Classification of Work and Quantify Takeoffs Once the design plans and spec-
ifications are completed, the engineer should divide the project into a series of spe-
cific bid items. Each of these items should then be broken down into the most logi-
cal and practical individual categories of work so that payments may be proportioned
as the work progresses. Since cach bid item generally includes a number of separatc
and distinct construction operations or classes of work, a separate cost estimate must
be prepared for each item. The quantity takeoffs are divided further to deline the
scope or size of each construction operation based on the site condition and the en-
gineer's judgment. The quantities (numerical figures) are then tabulated using the ap-
propriate units: feet, yards, cubic yards, or days. Be aware, however, that certain pro-
jects benefit from having the items expressed as a lump sum.
Construction Program Having executed the classification of work and quantity
takeotTs, the cost estimator must now develop a realistic construction program. Items
that must be tabulated are the required type and level of laborers and professionals,
required type and number of construction equipment, and the frequency of use and
 504 ELEMENTS OF DETAILED DESIGNS
duration of need (anticipated) for such equipment and the crews to operate them. The
cost estimator should establish a preliminary but fairly accurate overall construction
progress schedule, including sequential and concurrent construction operations. The
schedule should provide a reasonable time for mobilization and preparation at the
beginning of the project, in addition to demobilization and completion of the work.
Unit Cost Development In anticipating the direct costs of a project, the cost es-
timator must develop the unit cost for (I) direct labor, (2) equipment operations,
(3) consumable materials, and (4) nonconsumable materials. The unit cost for di-
rect labor is determined by establishing a schedule and dividing the project into var-
ious c,lassifications, such as crafts, skills, and labor. The availability, quality, and
rates tor the various trade classifications, in addition to the hourly rates of payment
for regular work hours, must be designated for each classification. Various taxes and
lringe benelits should be taken into account when determining the hourly rates.
Direct equipment operating costs include the cost of fuel, oil, tires, all equipment
parts, and labor-operators, maintenance, and repair personnel.
Consumable materials are defined as items that are used during the construction
operation; they are not part of the permanent installation. This category includes
items ~uch as small tools and accessories. Depending on the type of construction,
the unIt cost for consumable materials may be expressed as cost per hour of use or
as a percentage of the direct labor cost. Nonconsumable goods are defined as ma-
terials and equipment that are installed on a permanent basis. A schedule of these
items should be prepared early in the estimating work because a considerable amount
of time will be spent on obtaining quotes from all suppliers and manufacturers, in-
cluding transportation or freight charges.
Direct Costs Direct costs encompass direct labor, the operation of construction
equipment. consumable and nonconsumable materials. and subcontracted services.
If the project contains several construction operations, these costs must be estimated
separately for each operation and the costs then totaled to yield the cost for each
bid item. Later, the direct costs of each bid item are summarized.
Indirect Costs The principal indirect cost items are administrative and overhead
labor; transportation and equipment overhead; general office expenses; staff ex-
penses; expenses incurred in establishing an on-site project oftice, in drawing up the
construction plan and shop drawings~ expenses associated with the operation and
maintenance of the facilities; payroll taxes and burden; equipment mobilization and
demobilization; equipment ownership; insurance and bond; taxes; and licenses
and fees. Items such as depreciation, interest on investment. major repair and stor-
age costs, overhead, insurance, and taxes are categorized as equipment ownership
costs. If equipment is rented. rental charges are substituted for ownership costs.
Estimate Summary The estimate summary should be comprised of direct and
indirect costs, makeup, and cost escalation adjustment. Makellp is deli ned as the al-
lowance granted to the contractor for profit and contingencies. The time schedule,
SUPPLEMENTAL STUDIES 505
working conditions, and uncertain ground and climatic conditions, in addition to the
economic situation, all affect the amount of the makeup; the amount is generally
expressed as a percentage of the total cost and is approximately 20% for man~ wa-
ter treatment construction projects. The percentage will be higher for small projects,
and vice versa.
The final cost estimate described above is a general approach that is applicable
to almost any type of construction project. including the construction of water treat-
ment plants. Table 6.3-1 is an example of a cost estimate summary for ,a 50-mgd
(2.2-m'/s) simple direct liltration plant. Table 6.3-2 is another example 01 a cost es-
timate summary for a 40-mgd (I.75-m 3/s) conventional process plant wllh ozona-
tion and lilters with OAC beds. It should be noted, however, that both cost estimates
do not include the contingency costs (about 20% of total cost), bids evaluation, or
construction management services (eMS) by engineers (3 to 6% of total cost, de-
pending on the contract). Actual cost estimates for the construction of water t,reat-
ment plants should include these two items. Table 6.3-3 lists the ENR Index 01 ma-
jor U.S. cities (1999). .
Whether it is regarded as an art. science, or equal parts of good data, good Judg-
ment, and good fortune, cost estimators rely increasingly on computers to help pre-
pare the tinal cost estimate; since computers are more accurate and efficient, ~hcy "'e-
duce project costs. Approximately 170 different types of computerized cost e~tll~atlllg
software arc currently available on the market. The popular packages are dlstnbuted
by Contractors Management (CMS) of San Diego, California; Management Computer
Controls (MCC) of Memphis, Tennessee; Timberline Software Corporation and
Bidtek, Inc., of Wilsonville, Oregon; and 0-2, Inc., of Boise, Idaho.
BIBLIOGRAPHY
Dickson, R. D., in Water Treallllelll Plall! Design: Estimating Water Supply System Costs,
R. L. Sank, cd., Ann Arbor Science, Ann Arbor. MI, 1978.
Gunman, D. L., et aL, Compwer Cost Models for Potable Water TreatllJent Plants, EPA/60012-
78-181, U.S. Environmental Protection Agency, Washington, DC, September 1978.
Hopper, T. W., "Estimating Construction Costs," 1. COl/slllt. Eng., p. 94 (August 1968).
Montgomery, 1. M" Consulting Engineers, Water Treatment-Principles and Design, Wiley,
New York, 1985.
6.4 SUPPLEMENTAL STUDIES
Designing a water treatment plant requires the combined talents of experts from var-
ious disciplines. In many cases, the entire plant design can~?t be. hand~ed by one
design tirm since some arcas arc simply beyond the capability .ot the fIrm or be-
cause it is more cost-effective to subcontract thcm. These areas Include (l) the ge-
otechnical/geological study, (2) cold weather design, (3) corrosion study, (4) acoustic
study, (5) surveying, and (6) landscaping.
 506 ELEMENTS OF DETAILED DESIGNS
SUPPLEMENTAL STUDIES 507

TABLE 6,3.1 Detailed Cost Estimate of a 50·mgd Direct Filtration Plant TABLE 6.3-2 Detailed Cost Estimate of a 40-mgd Conventional Treatment Process
Utah Valley Water Fillrations Plant Including Ozonations and GAC Filter Beds
Detailed Construction Cost Estimate Summary. 1977 North Bay Regional Water Treatment Plant, California
Adjusted Detailed Construction Cost Estimate Summary, 1988
Description Estimate Estimate!l ($)% Description Estimate %
Earthwork, grading, paving $ 381,699 $ 343,529 4,8 Site works $ 3,270,000 11.3
Yard piping 1.225,249 1,102,724 15.3 Yard piping 4,500,000 15.6
Landscaping and irrigation 68,779 68,779 1.0 Flocculation/sedimentation basins 2,990,000 10.4
Operations building 844,730 824,938 11.5 Filters 3,180,000 11.0
Pumping station 722,961 702,206 9,7 Blowers and building 510,000 1.8
Meter vault I 297,612 286,484 4.0 Ozone generators with building 2,060,000 7.1
Meter vault 2 83,497 77,865 1.0 Preozonation tanks 910,000 3.2
Valve vault I 61.927 57,665 0.8 Postozonation tanks 840,000 2.9
Valve vault 2 71,639 66,597 1.0 High-service pump station 1.580,000 5.5
Valve vault 3 219,241 203.525 2.8 Clearwell 1,000,000 3.5
Valve vault 4 47,871 45,312 0.6 Operations building 3,350,000 11.6
Filters 1.735,277 1,628.577 22.5 Wash-waste reclamation building 410,000 1.4
Filter waste wash water 328,034 321,342 4.5 Wash-waste recycling facility 470,000 1.6
reclamation basin
Chemical feed system 1,040,000 3.6
Filter waste w8shwater basin 169,746 157,571 2.2 Chemical feed pipe trenches 380,000 1.3
Flocculation basins 374,815 349,080 4.8 Intake structures 460,000 1.6
Chemical storage and activated 226,435 218,832 3.0 Miscellaneous structures 1,140,000 4.0
carbon tanks
Miscellaneous 780,000 2.7
Treated water reservoir 465,491 426,906 6.0
Retaining wall 76,328
Total $28,870,000 100
68,695 1.0
L. P. gas tanks 14,330 12,897 0.2 Notes: (I) The cost is rounded up below $1000.
Testing and disinfecting 20,000 (2) No contractor's overhead and profit are included.
20,000 0.3
Fend ng and gates 25,675 (3) No contingency and CMS costs by the engineers are included.
23,107 0,3
Overflow structure to Salt Lake aqueduct 43,848 39,463 0.5
Sludge supernatant pumping station 16,620 15,412 0.2 Geotechnical/Geological Study The geotechnical/geological study and rec·
Move on and off 150,000 150,000 2.0 ommendations may be obtained from reputable local geotechnical consultants for a
Subtotal (I) 7,671,824 7,211,506 100 reasonable fee. A local firm or a large firm with a local office is generally preferred
Contractor's overhead and profit 1,534,703 1,442,301
(20% of subtotal) simply because it is assumed to be most familiar with local conditions and the his·
Subtotal tory of events. These firms usually provide such services as test borings of several
9,206,527 8,653,807
Electrical, instrumentation, telemetry or more strategic plant site locations and laboratory analysis of test samples, and
1,341,189 1,341,189
(approx. 17.5% of subtotal) furnish a report on their findings. Additionally, the firm should provide recommen·
Total $10,547,716 dations on the plant design, especially with respect to the structural design, the de·
$9,994,996
sign of the foundation, and the construction aspects of the civil engineering work.
:IAdjusted ~~wnward by 10% to account for lower Salt Lake City area costs for earthwork, paving, fenc-
Subcontracting firms are expected to supply the following information: in situ mois·
mg, yard plpmg, concrete, reinforcement steel, miscellaneolls metalwork, structural steel, waterstops and
certain other construction materials, and associated labor. ture, density, a laboratory maximum density and optimum moisture curve, direct
Nora: (I) No cost for the construction management services (CMS) by the engineer is included. shear tests, moisture sensitivity test. consolidation tests, resistivity, pH, and an analy-
(2) No contingency cost is included. sis of the soluble sulfate and chloride content of the soil.
The design recommendation should also address (I) seismic design parameters;
(2) structural design parameters such as the type and size of foundations, the vari·
ous anticipated earth pressures for subterranean walls, and the design of slabs on
various grades; (3) design data on earthwork, such as cut, fill, and compaction reo
 508 ELEMENTS OF DETAILED DESIGNS SUPPLEMENTAL STUDIES 509

TABLE 6.3·3 Example of ENR Construction Cost Index

capacity of the tank should be large enough to accommodale a required volume of
water or sludge below the ice caps. Another important design consideration con-
cerns pipelines. All pipes transporting liquid must be buried below Ihe local frosl
101111011111 IlIIT ,
1f11>101 JIIlI'll UII • JIIl~~~IIJI\~ : lu11~OI ~II:I. : IntlDlIlI!
JIIlI 'II HIlI.
MlT[RlIII
I~I '00 Hm. depth so Ihat the piping system functions on a continuous basis and to avoid pipe
1111 IIDU III! ~OU 11111 . IIDII IU! , ~DU IlIR IIIIEI III! ruplures which result from freezing of the water.
ATLANTA 3822.76 +4.5 2789.81 +5.7 6583.16 +4.1 3849.85 +6.3 2143.16 ... 5.2
BALTIMORE The design and construction of buildings and treatment plants in cold weather
BIRMIHOKAM
4498.88 +0.2 3061.75 .0.3 8442.11
••• 4671.47
••• 2078.85 +0.8
regions are unique tasks requiring very special knowledge. Thus projects located in
BOSTON
4245.00
6994.40
+1.3
+4.'
2649.57
3832.58
+3.1
+4.8
7052.03
15131.58
•••
+3.8
3658.86
6819.82
+2.0
+4.0
2033.S7 +4.4
arctic regions should employ project engineers who are licensed as cold weather de-
2007.10 +8.6
GHIGAGO 7192.96 +5.0 3915.04 +5.0 151G8.42 +4.6 6553.15 +4.1 sign specialists (in the engineering field). The engineer, as well as the archilect, me·
2304.77 ... 6.6
ClNGIHNATI 5872.81 +4.7 3229.27 +3.3 11789.47 +5.4 4839.04 +3.9 2248.46 +2.5 chanical engineer, and olher project staff members, should be either registered as
GlEVELANO 6481,51 +2.1 3583.05 .. 1.2 13350.00 +2.9 5782.58 +2.6 2239.53 -0.9 cold weather design specialists or hired/selected only if qualified to design and con-
DALLAS 3945.14 ".\ 2868.00 ...S B742.11 +1.8 3385.59 +2.5 2230.S7 -3.\
DENVER struct structures for cold weather regions.
DETROIT
4429.43
7042,05
-1.0
+3,5
2823.00
3956.43
-1,3
+4,3 i
8255,26
14742,11
•••
+3,7
4033,03
6&30,63
..0,5
+5,1
2084,57 - 3.4
2322.65 +2,8
KANSAS CITY Corrosion Studies Corrosion is a common concern in the field of water treat-
lOS ANGELES
MINNEAPOUS
6016,54
6823,97
6846,48
... 1.11
... 2,0
+2,9
3385,37
3589,02
3571.58
...
+3,6

+2,3
\
12052,63 +2,0
14457,89
14531.58
+3,2
+4.0
5135,14
5952,55
5938.94
+5,3
"'\,8
2317,0\ ... 1.2
2145,11 -2.9 ment. Corrosion of pipes due to the carrying water is rarely a serious problem be-
NEW ORLEANS
+4.7 2136.22 ~ 1.5 cause all steel and cast iron pipes, with the exception of pipes less than 4 in. (10
NEW YORK
3926.00
9301,94
-2.6
+7.0
2800.60
5093,38
-3•• 6842.11
+5.5 : 20597.37
•••
+5.9
~ 3358.56
9534.53
•••
+2.7
2138.71 -7.4
2378,93 +12.8
cm) in diameter, are now cement lined and coated. Moreover, in the United States,
PHILADElPHIA 7461.87 +2.1 4083,58
the pH of the filtered water is now commonly adjusted prior to dislribution to Ihe
+0.9 i 15736.04 +3.5 6839.64 '13,2 2399,79 - 2,8
PITTSBUROII 6087.39 +2.2 3514.25 +1,5 12497.37 +3.4 5733.33 +3.8 21511.70 -2.0
consumer, thereby significantly reducing internal corrosion in the main distribution
ST.lOUIS 6801.36 .3~ 3488.24 +5.5 ; 14552.83 3.\ 5841.44 lines. Nevertheless, certain sites, primarily marshy areas, can have very corrosive
+6.1 2050.511 +4.4
SAN FRANCISCO 6818.39 +0.8 3900.211 +1.5 14410.53 +1.8 6739.94 +4.0 soils. Under these circumstances, the issue of corrosion must be given special con-
2165.14 -3.1
SEAmE 6932.16 +2.8 3450.31 +0.4 j 14810.53 +4.0 5670.B7 +2.3 2103,48 - 2.4 sideration during design of the yard pipings and concrete structures. There are also
MONTREAL
TORONTO
6359.59
805t.54
+4.8
+1.3
3700.10
4099.07
+8.6 \ 12378.95
+1." : 16913.16 +2.0
••• 5387.39
8521.92
•••
+3.2
2670.31 +21.1
2620.23 - 1.3
a few cases where iron and steel pipes (buried) corroded severely due to stray cur-
rents emanating from an electric railroad track or as the result of galvanic action
Source: Engineering New.t-Record. July 12, 1999. along the pipelines. A corrosion expert should be involved in the project if pipe cor·
rosion is anticipated to be a major problem.

quire~e~ts, and th~ type of equipment required for excavation and compaction; (4)
c?rro.slvlty o.f the SOIl, complemented by recommendations on the major underground
pipelIne deSign and the required type of cement for concrete structures set in the
ground; (5) a prediction as 10 the groundwaler level during planl conslruclion and
after construction is completed; and (6) magnitude of seltling (for struelures) after
completion of construction.
Cold Weather Design If Ihe water Ireatmenl plan I is located in a region with
extre.mely .cold wInt~r weather conditions, the plant design requires many special
consideratIOns. For Instance, all major unit processes should either be housed or
buried in the ground (sedimenlalion tanks and the c1earwell) 10 prevent freezing of
the ~rocess wat~r. Certain items in the building and construction design should be
speCIfically outlIned, especially Ihe rooling system and Ihc vapor barriers (Ihese
should be deSIgned 10 reduce condenSalion) otherwise, the doors will freeze in place.
Heallraces should be conducled for all oUldoor equipmenl and chemical feed pipes;
these Ite~s should also be Insulated. Drain pipes, especially filter waste wash and
sludge pIpes, should be prolected from Ireezing and be capable of draining lanks or
lagoons thai may be Irozen solid. The exit of Ihese pipes must be located allhe bot-
tom of the tank, beneath the anticipated thickness of surface ice. Furthermore, the
Acoustic Studies If a water treatment plant is located in a quiet residential area,
the noise generated by certain water treatment equipment may be a significant is-
sue. Noise emanating from air compressors, blowers, high-service pumps, and en-
gine·driven power generators is loud and disturbing to both plant operators and
neighboring residents. OSHA recommends that Ihe maximum noise level be main-
tained below 80 decibels (dB) for the health and safety of operators, and this crite-
rion is generally much more stringent with respect to neighboring communities. The
design engineer should consult an acoustic control specialist for recommendations
on how to solve the noise problem properly in a cost·effective manner.
Surveying Although most engineering firms are capable of surveying, an out-
side firm of surveying specialists should be used if it is the more cost-effective al·
ternative. Whenever the job sile is remote or does not have easy access, or if it is
difficult to conduct surveying using regular surveying techniques, an aerial survey-
ing specialist should be chosen.
Landscaping Proper landscaping not only makes a plant site attractive but re-
duces maintenance and irrigation costs. Experience has shown that reputable local
landscaping professionals can do a good job at a reasonable cost.
 LlME·SODA ASH SOFTENING 511

90
Notes
CHAPTER 7 80 I. Raw water TOe = 4.8 mgIL
2. Ferric sulfate dosage
Specific Water Treatment Processes A=2.0mgIL
B =4.5 mgIL
70 e = 6.0mg/L
D=7.5 mgIL
E= 9.5 mgIL
60
7.1 LIME-SODA ASH SOFTENING

7.1.1 General Discussion 50 A

The main purpose of waler softening by means of lime-soda ash and lime softening is
to reduce the levels of calcium and magnesium in the process water to a total hardness 40
of approximately 80 mg/L <as CaCO,) as weli as reducing the magnesium hardness to
approximately 40 mg/L <as CaCO,). Over 1000 municipal water softening plants cur-
rently operate in the United States. the majority located in the Midwest and Florida. 30
Prior to the mid-twentieth century. the primary problem of the domestic con-
sumer was the high consumption of soap by hard water. However, due to the wide-
spread use of synthetic detergents, this is no longer a major issue today. Aside from 20
the more immediate consumer concerns, the additional benefits of lime softening
arc quite substantial, including the removal of heavy metals, metallic elements, and
organic compounds, as well as effectively killing bacteria, viruses, and algae. Lime
8 9 10
softening also improves water quality with respect to pipe corrosion, boiler feed,
and cooling waters. The spontaneous removal of elements such as iron, manganese,
lead, mercury, and chromium is particularly attractive to purveyors of potable wa-
pH
ter, while the reduction of silica and total dissolved solids is beneficial to industries. Figure 7.1.1-1 Enhanced softening process for TOe removal. [Data from S. R. Qasim. et
Since 1990, special attention has been given to the excess lime softening process al.. J. Environ. Eng. Oil'. ASCE, 1t8(3) (May-June t992).)
because of its effectiveness in controlling THMs and HAAs by removing natural or-
ganic matter (NOM) prior to chlorination. The softening process is termed enhanced strated an inverse correlation between the incidence of cardiovascular disease and
softelling by the EPA, and TOC removal can be as high as 80% at pH 10.5 or higher the hardness level of drinking water. Yet contlicting studies have also been reported.
with 10 to 15 mg/L offerric sulfate (Figure 7.1.1-1). Lime softening, as well as other processes such as ion exchange, electrodialysis,
Water hardness is defined as the amount of divalent metallic cations in the wa- nanofiltration, reverse osmosis, distillation, and freezing may be employed to soften
ter and is expressed in mg/L as CaCO,. The major divalent metallic cations that hard water. However, only the process of lime softening is discussed in this chap~
contribute to water hardness are calcium and magnesium. A few major European tel'. Yet engineers should be aware that there are a few technical articles that demon-
countries are adopting their own scales of water hardness, and this confusing situ- strate that the use of magnesium carbonate, in conjunction with lime softening, im-
ation is compounded by the fact that many industries in the United States still use proves the calcium carbonate sludge characteristics.
the old expression of hardness: grains per gallon. Table 7.1.1-1 presents the rela-
tionship between the expressions of hardness. 7.1.2 Basic Chemical Reactions
Engineers should note that hard water is not currently known to adversely af-
fecting human health in any significant manner. Quite a few studies performed dur- In this section we present the basic chemical reactions of (I) lime and lime-soda ash
ing the 19708 in the United States. Canada, United Kingdom. and Japan have demon- softening, (2) caustic soda softening, and (3) water stabilization of softened water.

510
 512 SPEC IFIC WATER TREATME NT PROCESSES
LlME ·SOOA ASH SOFTE NING 5 1 3

TABLE 7.1.1-1 Rela tionships Among Hardness Units Water Stabilization of Softened Water
United States mglL as CaCO J • which is 0.02 meq/L
I grain of CaCO,lgal ~ 17. I mglL as CaCO,
2Ca(OHh + 3CO, --> CaCOJ + Ca(HCO,h + H20 pH ~ 8.8 ( 15)
Germany Equiva lent 10 10 mglL o f lime expressed as CaO
I German degree = 17.86 mg/L as CaCO l CaCO, + CO2 + H20 --> Ca(HCO,h pH :5 8.3 ( 16)
Fra nce I French degree = 10 mglL as eaco l .

The level of waler hardness is generally cl assified as r~lI ows: 2CaCO, + H 2SO, --> Ca(HCO,h + CaSO, pH :5 8.3 ( 17)
0-75 mg/L So ft water
2CaCO, + 2HC I --> Ca(HCO, h + CaCI2 pH :5 8.3 (18)
75- 150 mglL Moderately hard wn ler
J50-300 mglL Hard \Vale T
Over 300 mglL Very hard wate r 7.1.3 Softening Process Alternati ves
There are fi ve alte rn ati ves to the softening process. These depe nd on the type o f
hardness, degree of softe ni ng, operat io nal conve nie nce, degree to whi ch the pro-
Lime and Lime-Soda Ash Softening du ctio n of lime sludge must be reduced, and desired savings in c he mical cost. The
alternati ves include ( I) part ial lime softening, (2) excess lime softe ning,
(3) li me-soda ash softe ning, (4) caustic soda softe ning, and (5) softe nin g with both
CaO + H20 --> Ca(OHh ( I) lime and ca usti c soda.
CO2 + Ca(OHh --> CaCO, + H 20 (2)
Ca(HCO,h + Ca(OHh --> 2CaCO, + 2H20 (3)
Alternative 1: Partial Lime Softening This process removes onl y calcium
pH '" 9.5
carbonate. Lime (alone) is fed to the process water to rai se the pH to approx imate ly
Mg( HCO,h + Ca(OH)2 --> CaC03 + MgCO, + 2H20 pH '" 9.5 (4) 9.5 [reacti on (3)]. The softened water is then stabili zed by recarbonati on [reacti ons
MgCO, + Ca(OHh --> Mg(OHh + CaCO,' pH '" I I (5) ( 15) and ( 16)] or th rough the add ition of ac id [reactions ( 17) and ( 18)] to reduce
the pH to appro ximate ly 8.8 or below, thus minimi zing the occurre nce of heav y scal-
' plus Ca(OH), due 10 a n excess lime dosage ing wit hin the pi pes and va lves.
MgSO, + Ca(OHh --> Mg(O Hh + CaSO, (6)
CaSO, + Na,CO, --> CaCO, + Na,SO, Alternative 2: Excess Lime Softening This alternati ve process is used in re-
(7)
moving the carbonate form of calcium and magnesium . An excess amount of lime is
CaCl, + Na,CO, --> CaCO, + 2NaC I (8) applied to Ihe process water at a minimum pH of 10.6, preferably II 10 11.3 [reaclions
(3) through (5)] since magnesium hydrox ide will not adequ ately form unless the pH
Caustic Soda Softening of the water is raised above II . Followi ng the formation of magnesium hydrox ide, the
process water should be recarbonated to reduce the pH to 8.7 to 8.8. 1l1is step will con-
co, + 2NaOH --> Na2COJ + H20 vert the hydrox ide form to carbonate and bicarbonate because the water has a calcium
(9)
hydrox ide conce ntration of 40 10 50 mglL at pH II and Ihe water is un stable.
Ca(HCO,h + 2NaOH --> CaCO, + 2Na,CO, + 2H, O ( 10) At pH 8.8, approxi mately 10% of the alkalin ity is carbonate and 90% is in the
Mg( HCO,h + 4NaOH --> Mg(O Hh + 2Na2CO, + 2H, O bicarbonate for m. Calcium bicarbonate is di ssolved in the water, yet calci um car-
( II )
bonate prec ipitates out and must be re moved by means of scdim ent ation or filt ra-
MgSO, + 2NaOH --> Mg(O Hh + Na2S0, (12) ti on to mini mize its effect of harde ning the softened water. The sedi men tati on process
CaSO, + 2NaO H --> Ca(OH h + Na,SO, generally follows the recarbonatio n process because the filtration process receives
( 13)
a rat her excessive amount of calciu m ca rbonate precipitate. When clear watcr such
CaCl, + 2NaO H --> Ca(OHh + 2NaC I ( 14) as underground water is softened by mea ns of th is process, a portion of the raw wa-
ter is ble nded to the softened water (called split treatm ent). The carbon di oxide and
bicarbonates (in the raw water) react with the calcium hydrox ide in the softe ned wa-
NOTE: Reacti ons (13) and ( 14) need to add CO2 10 form CaCO., prec ipitate. te r 10 yie ld the calcium carbonate prec ipitate.
 514 SPECIFIC WATER TREATMENT PROCESSES
LIME-SODA ASH SOFTENING 515

Alternative 3: Lime-Soda Ash Softening This alternative should be used to Design engineers should note that recycling lime sludge to th~ head of the soft-
remove both carbonate and noncarbonate hardness. Noncarbonate hardness is typi- ening process train always improves the efficiency of the softenmg process; re~y­
cally produced by calcium sulfate and calcium chloride; soda ash must be used to cling gives a seeding effect in the formation of calcium carbonate floc. The solids
remove these compounds [reactions (6) through (8)J. The pH of the softening wa- contact and sludge recirculation types of reactor clarifier are frequently use? m lime
ter should be a minimum of 10.6; therefore, rccarbonation or acidification is re- softening plants precisely because of this feature and are proven to be quite effec-
quired to stabilize the water prior to filtration.
tive in this capacity (Figure 7.1.4-1). If the carbon d,ox,de content of the raw wa-
ter exceeds 10 mg/L, an aeration process such as a coke tray aerator should be con-
Alternative 4: Caustic Soda Softening Caustic soda may be used as a sub-
stitute for lime in alternatives 1 through 3. The advantages of caustic soda are its
····0-'· pH IO.S.lime only
decreased sludge production, the elimination of problems associated with chemical
dust, and the option of utilizing simpler storage and feed systems. Another advan- - -.- - pH 10.5. 10 mgll reme sulrale
tage is the production of Na2C03 as shown in reactions (9) through (II). Therefore, - •••- pH 10, 10 mgll reme sulf.,e
this process removes noncarbonate hardness without feeding soda ash. However, --0-- pH 10.5. 10 mgll reme sulf.te. sludge recycle
caustic soda also has several major disadvantages; the cost of caustic soda is four ····0.. • pH 10, 10 mgll feme sulfate, sludge recycle
to six times higher than that of lime; it is a potential safety hazard to operators 1000
should there be a massive leak; and freezing problems occur for 50% solutions at
55°F (13°C).

Alternative 5: Softening with Both Lime and Caustic Soda This last al- V
.-
.. . .' 1'-'

.. .' .•.•.
~ '
ternative is a variation of alternative 4; it is used to reduce the overall chemical cost I--""
and to reduce the capital cost of the lime feed system. The chemical reaction in-
100
~ V '
volved in this alternative is as follows:
.
2Ca(HC03)2 + Ca(OH), + 2NaOH -) 3CaCO, + Na,C0
3
+ 4H 20
.cr'
.. '
.... ..
..
. ....
"
7.1.4 Overall Softening Treatment Process -;::;; I"
.... i.-
All five water softening alternatives basically have the same overall water treatment
process as the conventional complete water process: flash mixing, flocculation, sed-
10
.
imentation, recarbonation, second sedimentation, filtration, and chlorination, A me-
...Sl --::;;
chanical mixer should not be used as the flash mixer in the softening process be-
, I/' i.o"
..... ...... ..... 0
cause the mixing blades will be coated by calcium carbonate, resulting in a large
ball of CaC03. The appropriate type of flash mixing is achieved by hydraulic mix- ,.'j .... ....
l....- .....
f- ..... ~

ing through the use of baffled channel or water fall from weirs. If the softened wa-
ter is stabilized through acid rather than carbon dioxide, the recarbonation and sec-
I -. L-

ond sedimentation processes are deleted, However, the addition of acid increases
either the sulfate or chloride concentration and slightly increases water hardness due
to the dissolving of CaCO, by acids, whereas the use of carbon dioxide does not.
The normal lime softening process can manage occasional turbidity spikes of up
to 500 to 1000 ntu because the large dosage of lime (for softening) also acts as a .1
coagulant. Nevertheless, turbidity removal and tilter performance will both improve 0.0 0.2 0.4 0.6 0.8 1.0 1.4
if metal coagulants such as sodium aluminate or ferric sulfate are fed in conjunc-
tion with lime. If the turbidity frequently exceeds 1000 ntu, engineers should con- Loading Rate (gpm/sq ft)
sider installing an additional clarification process (alum flocculation and sedimen- Raw river water. NTU = 40.
tation) upstream of the regular lime softening process train. Figure 7.1.4-1 Effect of pH, ferric and sludge recycle on settling velocities.
 516 SPECIFIC WATER TREATMENT PROCESSES LIME-SODA ASH SOFTENING 517

sidered for use in gas stripping. This scheme reduces the dosage of lime or caustic Filtration (Dual Media Bed)
soda required ror the removal of carbon dioxide [reactions (2) and (9)J, as well as
reducmg the volume of sludge production. Filtration rate 4 to 6 gpm/ft' (10 to 15 m/h)
Anthracite coal
Depth 20 in. (0.5 m)
7,1,5 Design Criteria
E.S. 1.0 to 1.1 mm
The design criter!a for (I) !lash mixing, (2) flocculation, (3) sedimentation, (4) re- V.c. <1.4
carbonatIOn, (5) hltratlOn, (6) chlorination, and (7) sludge handling and disposal (for S.G. 1.67 to 1.7
lime-soda ash softening) are presented below. Sand
Depth lOin. (0.25 m)
Flash Mixing E.S. 0.55 to 0.6 mm
V.c. < 1.5
Mixing energy (G) 300 to 500 S-I S.G. >2.63
Mixing time Backwash rate 20 gpm/ft' average (50 m/h)
10 to 30 s

Flocculation
NOTES:
Mixing energy (G) 5t050s- 1
Mixing time 30 to 40 min I. A surface washing system must be installed because it plays an essential role in main-
Flow velocity through ports 0.5 to 1.2 rtls (0.15 to 0.36 m/s) taining the top portion of the tilter bed free of mud balls and mud pans; these form as the
result of the cementing action of the calcium carbonate precipitates.
2. The design criteria for the surface wash system arc identical to those used in designing
regular tilters (refer to Section 3.2.7).
NOTE: Adequate mixing time is velY important because of the slow reaction rate.
3. Based on the bed expansion, use of filter backwash (alone) is not an effective method.

Sedimentation
Chlorination Although the eXt rule issued by the EPA does not acknowledge
Surface loading 0.75 to I gpmlti', rectangular tank (1.9 to 2.5 m/h); I to its role, the softening process is always characterized by a good disinfecting capa-
1.75 gpmlft', sludge blanket type (2.5 to 4.4 m/h) bility; the level of THM and HAA formation due to subsequent chlorination is usu-
Detention time Minimum of 2 h, rectangular tank; minimum of I h, ally also quite low because of effective removal of TOe by the softening process.
sludge blanket type The required chlorine dosage is defmed as the minimum amount required to main-
Weir loading 15 gpm/ft, rectangular tank (II m3/m . h); 20 gpm/ft,
tain a chlorine level of 0.5 mg/L throughout the distribution system to tight post-
sludge blanket type (15 m'lm . h) contamination. This restriction is imposed because under high-pH conditions, resid-
Total water loss < 3% due to sludge withdrawal ual chlorine is in the hypochlorite form. The disinfecting power of this ion is mLlch
less effective than hypochlorous acid. Under a pH range 01'8.6 to 8.7, approximately
90% of the chlorine residual is in hypochlorite rorm. Thus the chlorine disinfecting
Recarbonation
power increases with a decrease in pH after recarbonation or acid neutralization.
Carbon dioxide diffusion tank 3 min minimum contact time
Recarbonation tank 20 min minimum detention time Sludge Handling and Disposal
Water depth Minimum of 12 ft (3.7 m)
pH of water after recarbonation 8.7 to 8.8 Sludge production rate 2 to 3.5 Ib of solids for each pound of hardness
removed (refer to Section 4.3)
Sludge dewatering Sludge lagoon type of drying bed or a
NOTE: The use of liquid carbon dioxide has become more commonly practiced for mechanical dewatering system (refer to
recarbonation. Section 4.3)
 LIME-SODA ASH SOFTENING 519
518 SPECIFIC WATER TR EATMENT PROCESSES

TABLE 7.1.6.t Molecular Weights of Major Items

TAIILE 7.1.5·1 pH versus Hydroxide Alkalinity Mg
NaO H CO, Ca
Ca(OH ), Na, CO, CaCO,
pH 9.7 10 10.2 10.4 10.5 10.6 10.8 11.0 CaO
40 44 40 24.3
74 106 100
OH(mglL 2- 3 4- 5 8-9 12- 13 14- 16 17- 20 26- 30 41 - 50 56
as CaCO,
successful water softe ni ng. One of the most importa nt operational parameters is
proper c hemical dosage. The molecular weights of the major items involved in the
NOTE: Lime recovery by lime sludge recalcifi cation has be .
However. due 10 recent laws against air p II t' < en pract iced to some deg ree, sofJening process are listed in Tablc 7. 1.6-1.
thi s process. 0 II lon, most urban areas severe ly limit the use of
Chemical Dosages When sofJening water con taini ng carbonate hardness, the
following general formu la will determine the lime dosage (the chemical feed delaiis
are di scussed in Section 4.1):

lime (CaO in mgfL) = (A + B + C) X OlE

where
7.1.6 Operational Considerations
A = CO, as CO, (mg/L) X 56144
between the clements of hard~~~se:sn~ th:s~~~~n~nstOlchlO~letric chemical reac.tion
The lime- soda ash sofJc ning p' . . b I ..

operational staff and a basic chemical lab I g chhcm lcals. A properly trar ned
B = HCO, alkalin ity as CaC0 3 (mg/L) X 56/100
( < ora ory arc t ercfore essent ial in ensuring
C = Mg as Mg (mg/L) X 56124.3

10 D == excess lime req ui red to raise pH to 11 to remove magnesium ; it is

generall y 10 to 20% of A + B + C: thu s 0 = 1.1 to 1.2
E = purity of the quick lime (0.88 to 0 .95); therefore, E = 0.88 to 0.95
9.5
For slaked lime [Ca(OH)'l:
9.0
slaked lime [Ca(OH), l = (N + B' + C') X OlE
!I: 8.5
0.
where
8.0 A ' = CO, as CO2 (mglL) X 74/44
B' = HCO, alkalinity as CaC03 X 74/100
7.5
0 20 40 60 80 100 C' = Mg as Mg (mglL) X 74/24.3
Carbonate (CO~-) as percentage of total alkalinity o = excess lime: 0 = 1.1 - 1.2
I tit I I E = purity of the lime: generall y, E = 0 .93 - 0 .95
100 80 60 40 20 o
Bicarbonate (HCOj') as percentage of total alkalinity The soda ash dosage for noncarbonate hardness is determined as foll ows:
I I I I
o 10 20 30 40 50 soda ash (Na,C03 in mg/L) = (NH - L) X 106/100
. P Alkalinity as percentage of M alkalinity
where NH is the noncarbonate hardness as CaC0 3 (mglL) and L is the noncarbon~
Figure 7.1.5-1 Relation ships belween
from Th e Na/co Water Handbook F N ~ H. d .
an carbonate and bicarbonate ions. (Adapted ate hardness remaining in the softened water (mg/L).
• . . e11lmer. cd .. McGraw- HilI. New York . 1979.)
 L1ME·SODA ASH SOFTENING 521
520 SPECIFIC WATER TREATMENT PROCESSES

Since commercially available soda ash is nearly 100% pure, a purity correction dioxide. Assume that the softened water delivered to. the consumer has a pH of 8.7
factor is usually not necessary. an d a tota I h ard ness 0 f 85 mgiL , of which 35
. mgiL IS carbonate hardness.
If caustic soda is employed instead of lime, in the same manner as described
above, refer to reactions (9) through (14) and use a molecular weight of 40 (for Determine The required chemical dosages, using commercially available chemi-
3
NaOH) rather than 56 (for CaO). cals, based on a plant flow rate of 15 mgd (0.7 m /s)
The carbon dioxide dosage for recarbonation is
Solution
COz(mg/L) ~ {[HA + (CA X R)J X 441100J X F
Dosage of quick lime (mg/L) ~ (A +B+ C) X DIE
where HA is the hydroxide alkalinity as CaC03 (mg/L), CA the carbonate alkalin-
ity as CaC0 3 (mg/L), R the ratio used to convert CO, to HC0 3 (generally ranges where
from 0 to 0.2), and F the excess CO, dosage factor: usually F ~ 1.2. The pH of the
A ~ 15 X 56/44 ~ 19 mg/L
recarbonated water must be approximately 8,8 in order to facilitate the formation of
HeO] and caeo3 : these compounds are vital in removing excess calcium hardness B ~ (250 - 35) X 56/100 ~ 120 mg/L
by means of settling and filtration.
C ~ 30 X 56/24.3 ~ 69 mg/L
The treatment plant operator should remember the basic relationships of hy-
droxide, carbonate, and bicarbonate alkalinity: D ~ 1.15 (15% excess lime)
E ~ 0.88 (88% purity is assumed)
I. When the phenolphthalein alkalinity (PA) is zero, the water does not contain
hydroxide or carbonate alkalinity. Methyl orange alkalinity (MA) is the bi-
carbonate form. Thus

2. If PA < 0.5MA, there is no hydroxide alkalinity. The carbonate alkalinity is CaO (mg/L) ~ (19 + 120 + 69) X (1.15/0.88) ~ 272 mg/L
2PA and the bicarbonate alkalinity is MA - 2PA.
dosage of soda ash (mg/L) ~ (NR - L) X 106/100
3. When PA ~ 0.5MA, both the hydroxide and bicarbonate alkalinities are zero,
and MA is all in carbonate form.
where NH ~ 75 and L ~ 35 mgiL. Thus
4. When PA > 0.5MA, the bicarbonate alkalinity is zero. The hydroxide alka-
linity is 2PA - MA and the carbonate alkalinity is 2(MA - PA). say, 43 mg/L
Na,C03 ~ 40 X (10611 00) ~ 42.2 mg/L,
5. When PA = MA, both the carbonate and bicarbonate alkalinities are zero and
the hydroxide alkalinity is equal to either PA or MA. It is assumed that the water qualities of the softened water and recarbonated water
are the same as those listed in Table 7.1.7-1. .... .
It is also worth remembering that the pH for phenolphthalein endpoint titration Recarbonation is achieved by means of liquid carbon dIoxide, SInce It IS nea.rI~
is 8.3.
100% pure, is easily handled, is easily fed, and has low maimenance and c~plla
costs when compared to flue gas; flue gas is attained by burn 109 011 or natura gas
Sludge Handling and Disposal This issue also has great bearing on the op- and only yields approximately 12% CO,. . . e
erations of the softening plant. In Section 4.3 we present a detailed discussion of The carbon dioxide dosage is determined from Table 7.1.7-1. Smce hydroxld
this topic.
alkalinity of 43 mg/L and a carbonate alkalinity of 2 mg/L must be converted to bi-
carbonate alkalinity,
7.1.7 Example Calculations
CO, (mg/L) ~ {[RA + (CA X R)J X (44/100)J X F
Example 1
where HA ~ 43 mg/L and CA X R ~ 2 mg/L or 32 X 0.063 ~ 2. Thus
Given The raw water contains the following: 250 mg/L of bicarbonate hardness,
30 mg/L of magnesium, 75 mg/L of noncarbonate hardness, and 15 mg/L of carbon CO, ~ [(43 + 2) X 0.44J X 1.2 ~ 24 mg/L
 522 SPECIFIC WATER TREATMENT PROCESSES IRON AND MANGANESE REMOVAL 523

TABLE 7.1.7-1 Changes in Water Characteristics During the Softening Process Nalional Research Council, Drinking Water alld Health: Water Hardness and Health, Vol. 3,
(Assumed) National Academy of Sciences, Washington, DC, 1980, p. 20.
Characteristic Raw Water Softened Water Recarbonated Water Reh, C. W., in Water Treatment Pla"t Desigll: Lime-Soda Softelling Pmcesses, R. L. Sanks,
cd., Ann Arbor Science, Ann Arbor, MI, 1982.
CO 2 15 0 0
pH Riehl, M. L., Hoover's Water Supply alld Treatmell/, 10th cd., National Lime Association,
7.5 II 8.7
Mg Washington. DC. 1970.
30 3 3
Total hardness 325 110 Schroeder, 1. A., "Municipal Drinking Water and Cardiovascular Death Rates," 1. Alii. Med.
85
Noncarbonate hardness 75 35 Res.• 195:2: 125 (February 1966).
35
Total alkalinity (MA) 250 75 50 Singhal, A. K., "Conventional Lime-Soda Ash Softening vs. Split Treatment," J. AWWA,
P alkalinity (PA) 0 59 15 69:3:158 (March 1977).
HCO J alkalinity 250 0 20 Snoeyink, V. L., and Jenkins, D., Water Chemis/I}', Wiley, New York, 1980.
COJ alkalinity 0 32 30 Thompson, C. G., et ai, "Magnesium Carbonate: A Recycled Coagulant," Parts I and 2,
OH alkalinity 0 43 0 1. AWWA. 64: I: II (January 1972) and 64:2:93 (February 1972).
Notes: (I) MA represents methyl orange alkalinity. Van del' Veen, c., and Graveland, A., "Central Softening by Crystallization in a Fluidizcd-
(2) PA is phenolphthalein alkalinity. Bed Process," 1. AWWA. 80:6:51 (June 1988).
(3) Magnesium cannot be completely removed by this process. Therefore, a residual of 3 mg/L Winton, E. F., and McCabe, L. J., "Studies Relating to Water Mineralization and Health,"
is assumed to be present in the softened water.
1. AWWA, 62:1:26 (January 1970).
Wood. F. 0 .. "Selecting a Sollening Process." 1. AWWA. 64: 12:820 (December 1972).
Wyness. D. K.• "The Helical Flow Reactor Claritier," 1. AWWA. 71: 10:580 (October 1979).
SUMMARY OF CHEMICAL DOSAGES

Lime (CaO) dosage = 272 X 8.34 X 15 = 34,027 Ib/day

7.2 IRON AND MANGANESE REMOVAL
Soda ash dosage = 43 X 8.34 X 15 = 5380 Ib/day
Iron and manganese are minerals commonly found in soil, where they exist in the
Carbon dioxide dosage = 24 X 8.34 X 15 = 3002 Ib/day
form of oxides: namely, insoluble ferric oxide and manganese oxide (very insol-
uble), However, when water contains carbon dioxide, or in acidic water, ferric
iron is reduced to the ferrous form under anaerobic conditions; this ion is soluble
BIBLIOGRAPHY
in water. Under the same conditions, the dioxide form of manganese is reduced
from a valence of 4 to a valence of 2 and is also soluble in water. Consequently,
AWWA/ASCE, Water Treatment Plant Design, American Water Works Association, New a significant amount of iron and manganese is often found in deep well water and
York. 1969.
the hypolimnion water of stratified lakes and reservoirs. Iron and manganese may
Cox, C. R., Operation and Confrol of Water Treatment Processes, World Health Organization, occasionally be found in groundwater containing hydrogen sulfide. Organically
Geneva, 1964.
bound iron and manganese may also be found in groundwater containing humic
Craun, G. E, and McCabe, L. J., "Problems Associated with Metals in Drinking Water," acid.
1. AWWA. 67: II :593 (November 1975).
The iron and manganese concentrations of surface water seldom exceed I mg/L.
Graveland, A., et aI., "Developments in Water Softening by Means of Pellet Reactors," The iron content of groundwater may reach levels up to 10 mg/L under conditions
1. AWWA, 75:12:619 (December 1983).
of low alkalinity (less than 50 mg/L) and up to 2 mg/L of manganese. The presence
Hess, J. S., "Lime and Caustic Soda Softening at Fremont, Ohio," 1. AWWA, 60:8:980 (August of iron and manganese in potable water is not known to cause health problems.
1968).
These minerals are associated primarily with aesthetic factors, such as unpleasant
Judkins, J. E, and Wynn, R. H., "Crystal-Speed Conditioning of Lime Softening Sludge," taste and odor, and domestic problems, such as staining of laundry and lixtures, and
1. AWWA. 64:5:306 (May 1972).
can potentially be a serious problem for many industries-for example, staining of
Kemmer, P. N., ed., Nalco Chemical Co., The Nalco Water Handbook, McGraw-Hili, New the product andlor formation of mineral deposits. Another potential problem is the
York. 1979.
growth of iron bacteria within the distribution main. The Secondary Drinking Water
National Research Council, Drinking Water alld Health: Water Hardness and Health, National Standards (EPA) recommends MCL levels of 0.3 mg/L iron and 0.05 mg/L man-
Acadcmy of Sciences, Washington, DC, 1977, p. 349. ganese.
 524 SPECIFIC WATER TREATMENT PROCESSES
IRON AND MANGAN ESE REMOVAL 525

7.2.1 Alternative Methods
Several alternatives exist for the removal of iron and manganese in the field of wa-
ter treatment. The basic methods are (I) oxidation followed by clarification and fiI.
tration. (2) ion exchange, (3) stabilization by means of a sequestering agent, and
(4) lime softening. Each of these schemes is discussed. A case history is also pre-
sented at the end of the section.
Oxidation Oxidation may be achieved through (I) aeration, (2) chlorination,
(3) chlorine dioxide, (4) potassium permanganate, or (5) ownation.
Aeration Aeration can be used to fulfill two purposes: transfer of oxygen to wa-
ter and the removal of volatile gases.
4Fe (HC0 3 ), + 0 , + 2H, O --> 4Fe (OH), + 8CO,
2MnS04 + 2Ca(OH), + 0, --> 2MnO, + 2CaS04 + 2H,O
The air dilTusion type of aeration uses air diffusers to supply air to the process
water at a depth of 12 to 15 ft (3 to 4.5 m). The airlwater volume ratio is 0.75: 1.0.
This type of unit has an average oxygen transfer efliciency of only 5 to 10% and is
therefore not considered to be effective.
The coke tray type of conlact aerator is a more effective method of oxidation. This
aerator consists of a series of trays that are each 12 to 18 in. deep (0.3 to 0.45 m). The
bottoms of the trays are perforated and coke, crushed stone, limestone, or plastic medium
is placed in each tray as the contact material; the loading rate is 15 to 20 gpmlft' (37
to 50 mlh). Theoretically, I mg or oxygen oxidizes 7 mg or divalent iron and 3.4 mg
of dIvalent manganese. Organically bound iron cannot be oxidized through aeration.
The rate of oxidation is a runction or the pH or the process water: the higher the
pH, the better the result. To complete the process or oxidation within 15 min, the
pH of the water should be higher than 7.5, prererably 8. Manganese oxidizes very
slowly and is not errectively oxidized at pH values below 9.5. In ract, it would take
over I h to oxidize manganese at pH 9.5.
Alum nocculation and sedimentation arc generally required after aeration. These
processes are then followed by filtration, especially in cases where the iron con-
centration of the raw water exceeds 5 mg/L.
The rates of the oxidation reactions are pH dependent. For iron to be oxidized
within 15 to 30 min, the pH or the process water should be approximately 8.0 to
83, preferably 8.5. Under these same conditions, manganese will be oxidized within
2 to 3 h. When the raw water contains ammonia, chloramines are formed upon ad-
dition or chlorine. Consequently, the oxidation rate is reduced. If THMs and HAA,
formation is discouraged, chlorination cannot be employed as an oxidant.
Generall y, the standard process train for iron and manganese removal is alum
flocculation, followed by clarification and filtration; the practice of prechlorinalion
leaves a minimum level of 0.5 mg/L of available free chlorine. This process train
quickly coats the filter medium grains with a layer of manganese dioxide and fer-
ric hydroxides or ferric oxides. Once the filter grains have been coated, the filter
bed is capable of effectively removing manganese and iron by means of continuous
prechlorination.
Chlorine Dioxide Chlorine dioxide is a strong oxidant that efTectively oxidizes
organically complexed iron and manganese. The pH level necessary for oxidation
is a very important factor in the reaction rate and should be a minimum of 7.0.
Fe(HCO,), + NaHC03 + CIO, --> Fe(OH)3 + NaCIO, + 3CO,
Mn(HC0 3 ) , + 2NaHC0 3 + 2CIO, --> MnO, + 2NaCIO, + 4CO, + 2H, O
Theoretically, I mglL of chlorine dioxide will oxidize 0.83 mglL of iron, and the
same amount or chlorine dioxide wi ll oxidize 0.41 mglL of manganese.
Engineers must remember that many regulatory agencies are limiting the use of
chlorine dioxide due to the potential toxicity of the chlorite and chlorate ions formed
during the production of chlorine dioxide. Another important issue is the operational
cost. The chlorine dioxide process is approximately three times as expensive as chlo-
rination.
Potassium Permanganate Potassium permanganate is a strong oxidant. The re-
action time is therefore fast in a wide range of pH. The oxidation time varies from
as short as 5 min ror both iron and manganese, provided that the pH is over 7.0.
3Fe(HC0 3h+ KMn04 + 7H, O --> 3Fe(OH), + MnO, + KHC0 3 + 5H,CO,

Chlorination Chlorine is often used in the oxidation of divalent iron and man- 3Mn(HC0 3h + 2KMn04 + 2H,O --> 5MnO, + 2KHC0 3 + 4H,C0 3
ganese because, unlike aeration, it has a faster oxidation rate and is capable of ox-
idi zing organically bound iron . The reactions above indicate that I mg/L of potassium permanganate oxidizes
1.06 mglL of iron and 0.52 mgIL or manganese. Ir both iron and manganese are pre-
2Fe(HC03 ), + Ca(HC03 ), + CI, --> 2Fe(OH), + CaCl, + 6CO, sent in the raw water, the oxidation process may be more cost-effective if chlorine
is added prior to feeding potassium permanganate; the chlorine oxid izes iron rather
Mn(HC03 ), + Ca(HC0 3), + CI, --> MnO, + CaCJ, + 4CO, + 2H, O easily, leaving the more expensive potassium permanganate to oxidize the manganese.
In the field of water treatment, it is well recognized that the addition of potas-
The equations above show that I mg/L of ch lorine oxidizes 1.58 mg/L of iron and sium permanganate to the head of a conventional complete treatment process helps
that 0.78 mglL or manganese is oxidized by I mglL or chlorine. to improve flocculation and sedimentation. This phenomenon is due to the seeding
526 SPECIFIC WATER TREATMENT PROCESSES
effect of the oxidized iron and manganese particles. Under these conditions, the me-
dia grains in the upper portion of the filter bed soon become coated with manganese
oxide and ferric hydroxide in the same manner as the prechlorination process.
Ozonation The use of ozone for the sole purpose of oxidizing ferrous iron and
manganese is very unlikely due to the availability of other cost-effective methods.
Yet ozonation has been practiced successfully in Europe for the removal of both iron
and manganese by means of a conventional complete treatment process coupled with
preozonation. Experience has shown that excessive ozone oxidation of manganese
forms pennanganate and the water in the clearwell becomes pink in color.
2Fe(HC0 3)2 + 0 3 + 2H 20 --> 2Fe(OH), + 0, + 4CO, + H 2 0
Mn(HC0 3)2 + 0 3 + 2H 20 --> MnO, + O 2 + 2CO, + 3H 20
Theoretically, I mg/L of ozone will oxidize 2.3 mg/L of iron and 1.15 mg/L of man-
ganese.
Ion Exchange When the raw water contains less than 0.5 mg/L of iron and man-
ganese, either a cation or hydrogen cation type of ion exchange unit may be em-
ployed to remove these elements; these units are also capable of removing hardness.
Yet if the processing water contains any amount of dissolved oxygen, there is al-
ways the danger of fouling the ion exchange material and clogging the exchanger
bed. Thus it is vital that the process be run in the absence of oxygen to prevent the
elements from oxidizing. The ion exchange process is generally limited to the pro-
cessing of industrial water.
Manganese Zeolite Filtration Green sand impregnated with manganese (glau-
conite) is commonly known as zeolite. Each grain of the sand is coated with iron
and manganese oxides. This type of nIter medium is a form of ion exchange com-
monly used in industries. In this process, potassium permanganate is fed to the fil-
ter influent on a continuous basis, oxidizing soluble iron and manganese, and con-
tinuously regenerating the Ii Iter medium. It is very important that a proper dosage
of potassium permanganate be applied to the process water, since an excessive dosage
yields an effluent that is pink in color, while an inadequate dosage results in the
leakage of manganese into the filter effluent.
In the case of groundwater treatment, zeolite is placed in pressure filler cells
rather than gravity filters, to preserve the discharge pressure of the well pump. This
scheme eliminates the problem of repumping and allows one Ii Iter to be backwashed
by using the pressured filtered water from the remaining filters. The filtration rate
of this process usually ranges from 3 to 4 gpm/ft2 (7.5 to 10m/h). Manganese ze-
olite filtration is generally limited to water containing a maximum level of I mg/L
of iron or manganese. However, it has been reported that zeolite has the capacity to
remove 0.09 Ib/ft3 (1.5 kg/m3) of either iron or manganese and to regenerate 0.18
Ib/ft3 (2.9 kg/m 3) of filter medium by potassium permanganate.
IRON AND MANGANESE REMOVAL 527
Sequestering Process The purpose of the sequestering process is to hold both
iron and manganese in solution by means of sequestering chemicals. This process
is generally applicable to water containing less than 2 mgIL of iron and manganese;
both elements must be present in their bicarbonate form. The sequestering agents
include compounds such as sodium silicate, trisodium phosphate, hexametaphos-
phate, and zinc orthophosphate.
Proper dosage of the sequestering chemicals should be calculated on the basis of
the manufacturer's suggested dosage. Approximately 2 mgIL of hexametaphosphate
may be used for I mg/L of iron. However, polyphosphate dosages are limited to less
than 5 mg/L due to its tendency to stimulate biological growth in the distribution
system. In this case, an adequate amount of residual chlorine must be present in the
process water so that the growth of the microorganisms can be controlled.
Engineers should note that the sequestering process is effective only on cold wa-
ter: the sequestering agents lose their dispersing properties if the water is either
highly heated or boiled. For this reason, this process is seldom employed in treat-
ing medium-sized to large domestic water supply systems. Yet small water systems
are often financially incapable of implementing oxidation and filtration or lime soft-
ening processes, and the much simpler and less costly (capital) sequestering process
may therefore be the appropriate choice for treating their well water.
Lime Softening As described in Section 7.1, iron and manganese are effectively
removed by the softening process provided that the pH of the process water is above
9.5. Based on the pH and solubility relationships, 83% of the iron precipitates at a
pH of 8.4, but 92 and 100% of the iron precipitates at the respective pH values of
8.8 and 9.6. Manganese is more difficult to remove: No manganese will precipitate
out at a pH of 8.8; however, a pH of 9.4 and 9.8 will remove 98 and 100% of the
manganese, respectively.
7.2.2 Case History
A water treatment plant located in the northern part of the United States treats shal-
low groundwater containing approximately 5 to 5.5 mg/L of ferrous ion and 0.6
mg/L of manganese. The water has a total hardness of 200 mg/L and a pH of ap-
proximately 7.1. A sludge blanket type of reactor clarifier and standard dual-media
filters are used to treat this water. The basic treatment process is partial softening
with lime. This scheme automatically reduces the iron level of the filtered water to
nearly 0.1 mg/L and also reduces the manganese level to below 0.05 mg/L.
The plant conducted an experimental study to evaluate the feasibility of reduc-
ing the required lime dosage as well as the feasibility of reducing the sludge pro-
duction rate based on the oxidation of ferrous iron through application of an ap-
propriate amount of chlorine. The results were very interesting and proved to be
educational for those people designing iron removal systems under similar condi-
tions. Table 7.2.2-1 presents some of the results. These results indicate that the chlo-
rinated water must have a pH of at least 8.2 if the treatment process is expected to
produce filtered water containing less than 0.2 mg/L iron within a reasonable
 528 SPECIFIC WATER TREATMENT PROCESSES IRON AND MANGANESE REMOVAL 529

TAIlLE 7.2.2-1 Main Results of the Iron Removal Test the table demonstrate two major issues. The first is an obvious improvement in both
Li me dosage- 45-55 55-65
the iron hydroxide characteristics and iron removal (via sedimentation) through the
65-75 75-85 85-95
range (mg/L) addition of anionic polymer in concentrations as low as 0.2 mglL. The second is~
pH of the settled 7.8-7.9 7.9-8.2 8.3-8.6 8.7-8.9 8.9-9.1 sue is the time lag between the addition of oxidant and polymer. With respect to the
water polymer, a time delay of 5 min signilicantly improves the removal of both turbid-
Turbidity of the 0.7-2.2 0.1-1.2 0.05-0.2 0.05-0.12 0.05-0.15 ity and iron by means of sedimentation.
ti Itered water
Iron (mg/L)
Settled water 3.9-4.1 3.1-3.9 2.6-3.1 2.1-2.5 2.0-2.5
Filtered water 0.5-0.8 0.1-0.5 0.01-0.2 0.01-0.05 0.01-0.05
BIBLIOGRAPHY
Filter run length (h) 10-16 12-22 26-30 24-34 24-30
Adams, R. 8., "Manganese Removal by Oxidalion with Potassium Permanganate," 1. AWWA,
NOles: (I) Approximately .5 mg/L of chlorine were fed with lime for all cases.
52:2:219 (February 1960).
(2) The hardness of the filtered water increased in proportion to the lime dosages for all cases
where the pH was less than 8.5. However, a 5-10 mg/L reduction in hardness only occurred AWWA, Water Treatment Plant Design, American Water Works Association, New York, 1969.
in cases where the pH exceeded 8.5. AWWA Committee Report, "Research Needs for the Treatment of Iron and Mn," 1. AWWA,
(3) The water temperature ranged from 6 to ?"'c. 79:9:119 (September 1987).
(4) The filtration rate was approximately 2 gpm/ft 2 ,
Boatby,1. R., "Optimum Mn Removal and Wash water Recovery at Filtration Plant in Brazil,"
J. AWWA, 80:12:71 (December 1984).
period of time: a tilter run length of 24 h. The increase in filtered water hardness at Conneley, E. 1., "Removal of Iron and Manganese," 1. AWWA, 50:5:697 (May 1958).
pH values below 8.5, after the addition of lime, is the obvious result of the conver- Culp, R. L., Wesner, M. R., Culp, G. L., Handbook of Pllblic Water Systems, Van Nostrand
sion of most of the lime (calcium hydroxide) to calcium bicarbonate~ the calcium Reinhold, New York, 1986.
bicarbonate is completely soluble in water at lower pH conditions. Jenkins, S. R., et a!., "Effective Manganese Removal Using Lime as an Additive." 1. AWWA,
Bench-scale studies were also conducted at the plant to evaluate the effective- 76:9:82 (September 1984).
ness of anionic polymer for improving the settleability of iron hydroxide prior to Knocke, W. R., et a!., "Examining the Reactions Between Soluble Iron, DOC, and Alternative
tiltration. Some of the test results are listed in Table 7.2.2-2. The results listed in Oxidants During Conventional Treatment." 1. AWWA, 86:1:117 (January 1994).
Knocke, W. R., et a!., "Soluble Mn Removal on Oxide~Coated Filter Media," 1. AWWA,
80: 12:65 (December 1988).
TABLE 7.2.2-2 Jar Test Results for the Polymer Evaluation Tests
Knocke, W. R., et aI., "Using Alternative Oxidants to Remove Dissolved Manganese from
Beaker CI, KMnO, Settled Water Water Laden with Organics," J. AWWA, 79:3:75 (March 1987).
Polymer
Number Dosage Dosage Dosage Floc Size Turbidity Iron Montgomery, 1. M., Consulting Engineers, Water Treatment: Principles alld Design, Wiley,
New York, 1985.
I a 5 a Small 5.8 4.5 Robinson, R. B. and Ronk, S. K., "The Treatability of Mn by Sodium Silicate and Chlorine,"
2 a 5 0.2 Medium to large 3.5 2.3
J. AWWA, 79: II :64 (November 1987).
3 2.5 2.5 0.2 Large 2.7 2.1
a Robinson, R. 8., et aI., "Iron and Manganese Sequestration Facilities Using Sodium Silicate,"
4 5 a Small 7.0 4.7
5 5 a 0.2 Medium to large 3 1.6
J. AWWA, 84:2:77 (February 1992).
6 5 0 0.2 Large 2 1.4 Sanks, R. L., Water Treatment Ptant Design, Ann Arbor Science, Ann Arbor, MI, 1978.
Victoreen, H. T.. "Controlling Corrosion by Controlling Bacteria Growth," 1. AWWA, 76:3:87
NOle5: (I) Iron conlent of Ihe raw waler was 5 mglL.
(2) The jar tesl was conducted at G X 1 = 3 X 104 , wilh a lolal mixing time of 10 min and a (March 1984).
settling time of 20 min. Vitaraghavan, T., et a!., "Removing Mn from Water at Fredericton, N.8., Canada," 1. AWWA,
(3) Polymer (Na1co 7763) was added approximately I min after the oxidant was applied to the 79:8:43 (August 1987).
beakers, with the exception of beaker 5 and 6. Welch, W.A., "Potassium Permanganale in WaterTreatmenl," J. AWWA, 55:6:735 (June 1963).
(4) The polymer was added 2.5 min after chlorine application for beaker 5 and 5 min after the
addition of chlorine for beaker 6.
Willey, 8. F., et a!., "Iron Removal with Potassium Permanganate," 1. AWWA, 55:6:729 (June
(5) The pH was adjusted to 8.2 by NaOH prior to the addition of oxidant. 1963).
(6) The temperature of the raw water was 5.6 C.
Q
Wong, 1. M., "Chlorination Filtration for Fe and Mn Removal." 1. AWWA, 76:1:76 (January
(7) The chemical dosages are all in mg/L. The unit of iron in water is also mgIL. 1984).
 TASTE AND ODOR CONTROL 531
530 SPEC IFIC WATER TREATMENT PROCESSES

7.3 TASTE AND ODOR CONTROL TABLE 7.3.1-1 Tnsle- .:HId Odor-Producing Algae and Characteristics of the Taste
and Odor
In the water s uppl y indus try. it is no t un us ual for water pur veyors to receive c us- Odor When Algae Are: Tongue
Alga l Algal
!Orner compl aints regard ing the objec tionable taste and odor of the po table wa ler. Abundant Taste Sensation
Genus Group Moderate
T hese com pl ai nt s are more frequ en t when the source is surface water rather th an
groundwater, because surface water is more li ke ly to be affec ted by microorga ni sms Alla/Jaella Blue-green Grassy. nasturt ium. Septic
and orga ni c substances as well as manu fac tu red wastes. Some co nsumers also con- musty
Blue-green Grassy Se pt ic Sweet
si der th e odor o f the res id ual chlorine in tap water to be offensive. AI/acystis
Blue-gree n Grassy, nasturtium, Septic Sweet Dry
The most freq uen t causes o f taste and odor in the water suppl y arc algae an d A"h(ll/izomelloll
musty
Actillomycetes (du ring cerlai n seasons). The second is decay ing vege tatio n. Taste Geranium. spicy Fishy
Astel'ioll elfa Diatom
and odor proble ms arc occasio nall y the result of hydrogen s ulfi de , agr icultural Fi shy Septic Bitter
Cera tium Flagcll ale
run o ff, ind ustrial chemical sp ills, ill egally d isc harged ind ustrial chemicals, and 1'01- Flage llate Violet Fishy Slick
Dil/ob'JolI
IUlion by sewage. The c urre nl ( 1999) d rinki ng waler qualily slandards sel lhe Ih resh· Oscillatoria Blue-green Grassy Mu sty.
o ld odor number (TON) 10 less Ih an 1. II is ge nerally d iftic ul! 10 co nl ro l lasle and spicy
odor problems o nce Ihey have developed. T he pl anl desig n should Ihere fo re pro- ScelJedesmus Green Grassy
vide the operator with a certain amount o f n cx ibility, that is, several avail ab le o p- Spirogyra Gree n Grassy
Flagellate Cucumber. Fishy Bitter Dry.
tio ns. It is ve ry im porta nt to a ll ow the use of a lternat ive c hcm ica ls, as we ll as s up- Sylllll'a
muskmelon. metall ic
plyi ng a ltern at ive feed poi nt s, so th at the locatio n and type of chemi ca l can bc
spicy sli ck
changed in accordance wi th c hanges in raw water qu ality. Moreover, th e int ake de-
Tabellal';a Diatom Gerani um Fishy
s ign sho uld have provis io ns that a ll ow the intake to wi thd raw water fro m selec ted Grassy
Ulothl'ix Green
depths in the lake or reservoi r; th is is an essen ti a l design fac to r. Fishy
Volvox Flagell ale Fishy

SO/lrce: Adapted rrom Palmer ( 1962).

7.3.1 Major Taste- and Odor-Producing Substances
T he detec ti on and int ens ity of tas te and odor ,lre very subjective issues: Some peo-
ple are very sensiti ve, whereas o th ers have a ve ry hi gh level o f to le ra nce. O ne meth od
Ih al has bee n developed in rece nl years is fl avor pro file analysis (FPA). FPA is de·
s ig ned to red uce or eliminate the subjec ti ve nature o f this pro ble m by lI s ing a pa ne l
of people who are trai ned to analy ze odors associa ted w ith processed water.
Mosl odor problems arc Ihe resul! o f algae or Actill omycetes. Table 7.1. 1· 1 li sls Ihe
major odor-produci ng compounds and the responsible species of al gae. A mo ng the
Iyp ical odor·produci ng compound s, geosmi n and 2·melhylisobornco l (M IS) often im·
parI objeclionable odor at very low concentrali o ns (1 0 nglL or 0.0 1 fLgIL). Therefore,
Ihe acce plable leve l o f Ihese compou nds is considered to be approx im alely 8 ngIL.
The tas te of the potable water ca n genera lly be auributed to the presence of small
amounts o f metals, such as iron and copper, and phe no lic compoun ds. For example,
0.05 100. 1 mgIL of di ssolved iron, 2 10 5 mglL o f di valenl coppcr, and 4 10 9 mgIL
of d issolved zinc are considered 10 be las le Ihresholds. In Ihe case o f pheno l, lhe Ihresh-
old level is I mgIL; il is 0.002 gIL for chlorophe nol (ch lorophenol is formed by Ihe
chlorin atio n of pheno l- bea ring water). Ge nerall y, tastes and odors are objec ti onable
to the consumer. However, under norm al condit io ns, they are not a threat to the health
of the consumer, prov ided that water treatm ent plants are operated properly.
Re fer 10 Table 7.1. 1·2 fo r a lisl of las Ie· and odor-prod uc ing algae and Ihe ir char-
ac teristic features. Ta ble 7.3. 1-3 presen ts the troublesome algae a nd the recom-
me nded copper su lfale and chlorine d osages Ihal sho uld be used in Iheir conlrol.
7.3.2 Control Measures
T hree basic taste and odor control measures are used in the treatment o r po table wa-
ler: ( I) preve nli o n al Ihe so urce, (2) removal at the Irealme nl pl ant, and (1) contro l
wilhin Ihe d istribuli o n sys lem (re fer 10 Seclio n 7.3.4). T he fund ame nlals of eac h o f
these meas ures are d isc ussed.
Preventive Measures Melhods used in prevenl ion al the sou rce include
(I) reservo ir mixi ng, (2) aq ualic planl co ntrol, and (1) waler pollul io n con lrol.
Reservoir Mixing Most lakes and reservoirs stratify during warm seasons, fre-
que ntl y causing Ihe bOll om layer (hy polimni on) to be oxyge n defic ie nt. Under such
anaerobic co nd it ions, both hydrogen suIfide and the degradati o n of orgalllc sub-
stances result in the emi ss io n o f objec ti o nable odors. Experience gained during the
last half of the twe nti eth ce ntu ry has show n th at artifi cial mi xi ng of lakes by mea ns
of mec hanical mi xers. sub mers ibl e pu mps , or air- li ft pum ps sig nifi ca ntl y im proves
water q uality, thereby reduc in g the burden o n water treatment plants in co nt rolling
tas te and odor problems. as well as iron and manganese. T he desig n engineer sho uld
th erefore evalu ate the feas ibility of installing a reservoir mi xin g device as a pre-
venti ve measure fo r tas te and od or.
 TASTE AND ODOR CONTROL 5 33
532 SPECIFIC WATER TREATMENT PROCESSES
. d Dosnges 0 f Control Chemicals
TABLE 7.3.1-2 Odor- Produci ng Algae a nd Chemica l Struclures of Odors TAIlLE 7.3.1-3 Troublesome Aigne Il nd Req UI re
Copper
Compound Structure Anocialc:d Organisms Sulfate Chlorine
Trouble (mg/L) (mgIL)
MC lhylisooomeol A('li,w,"yr:~ttS Organi sm
(MIDI CHC@CH.
CH. OsciJIal()fia C'lffuin'/JJ
OH OSI'I'IIOfl)ri(l I('/lIIi.(
Algae 0.5- 1.0
Odor: aromatic 0. 1- 0.5
Diatoms ASle,.iollella, Sytlel/ra,
CH ,I to fi shy
Tabellaria
Gc:osmin AC'linom:~a l n Turbidi(y 0. 1-0.3
Fragillaria, Navicula

d5
Sympioro mllS ('(J/lM
Turbidity 0.2 2.0
Oscilialorio /trwis Melosira
01611010';0 simp/ids· Elldorilla, a p(lIIdorilla" Odor: fi shy 2- 10
Grass-green 0.25 0.3- 1.0
simo Volvo..\JI Odor: fishy
CH. AnahatnlJ ,f('ht'fcmf'- Turbidity. sc um 0.1-0.5
I;rvi Chara, Cladophora
Turbidity. scum 0.1-0.3 1.0-1.5
Coelastrlllll, Spimgyra
Odor: moldy. 0. 1-0.5 0.5- 1.0
Blue-green A,wbaella, Aphal/iZOlllellOll
~
C,!.Oy0';('-
CH .I grassy, vile
Mucidone 0.5- 1.0
Odor: grassy, vi le 0.1-0.3
CH. Clatlllv cyslis,
CoeiosphaeriwlI 1.1
Isobulyl mercaptan Turbidily 0.2-0.5
CH·· .... CHCH ._SH Oscillatoria
CH,- . Odor: aromatic 0.2-0.5
M io(/('.\'s ,;" flO.Hlflll{/t Golden or CryptomOllai' 0.3- 1.0
/If-Bulyl mercaptan CH.l ICH:h-SH Odor: aromatic to 0.2
Osl'i/fUluriu dlll/,rhra yell ow- DiIlObl)'ol/
fi shy
Isopropyl mercaptan Mil"rtlqsris jlus-aquut brown 0.2- 0.5
Mallomolws Odor: aromatic
Taste: cucu mber 0.1-0.3 0.3- 1.0
Dim ethyl disulfide MiCfIJ('." .flis /I,H'fllIl/llt' Syllllrtl 0.3- 1.0
Osdllf/ff/rill "/llIlyh"" Umgieuopsis Odor: fishy. Taste: 0.1-0.2
Dimethyl sulfide Osdll/lIllfiu (' /lII/y bt'1I oily
"""hflt'tIIl Odor: fi shy. vile 0.2-0.3 0.3- 1.0
MC lh )" mc: rc .. ptan Mi,"""',"s,;,, jim'lllIl/lI l' Dinoflagell ates Cem/ill'"
D.H'ilt/ltvria c'lwh'hr'U Odor: fi shy 0.2-0.5
GlellodilliwII
Odor: fishy 0.5- 2.0
Peridilli///Il
Smm:e: Adapted from 1.M . Montgomery. Consulli llg Eng ineers. IVtlfer Trellllllelll: Prill ciples olld Odor: decayed. 5.0
Filamentous Beggiatoa (sulfur)
D eJigl/. Wi ley. New York. 1985. Pi pe growths
bacteri a 0.3-0.5 0.5
Crellotllrix (iron) Odor: decayed.
Pipe growths
Aquatic Plant Control The ca n(rol of marginal aqua(ic plan(s is a ve ry effeclive 1.0-3.0
Cyclops c
method in preservi ng th e (good) water quality of a reservoir. However, this task is Cmstacea 2.0 1.0- 3.0
Daphnia c
not easily ac hi eved when the lake or reservoir is large. One method of con tro ll ing c 15- 50
Miscell aneous CllirollOlllllS (bloodworm)
marg inal plants, inc lud ing algae, is to drasti ca ll y chan ge the water level several times c 0.3
CraspedaclIsla (jell yfi sh)
a year. When Ihe waler level is lowe red by 4 (0 5 feel (1.2 (0 1.5 m), Ihe raals of
th e margina l plants become exposed to sun and dry up. Source: Adapted frolll J.M. Montogolllcry. Consulting Engineers.
Thc most commo n mcthod of co ntro ll ing algae growth is the use of copper su l- II Thcsc organ isms are classified also as Oagcll atc protozoa.
fate. The required dosage va ri es with thc type o f algae a nd ranges from 0.05 to 0.8 "Classification unccr1ain. .
"These organisms are indi vidually visible and cause conSUlller complmnl s.
mglL. Se nsiti ve species suc h as bluc-g ree n algae, which frequen tl y calise intense
taste and odor, may be treated with lower dosages. Howcver, gree n algae, which oc-
casio nally causes taste a nd odor problems, req uire much higher dosages. Yet dosages . d' I I include PAC- Ia creale a "black-
O ther methods that have bee n tnc In t le pas _ .
hi gher than 0.8 mg/L are not recommended since they are let ha l to 1110St types of out" condition- and the application of chlorine compound s. Carbon trcatmen~ I~
fi sh. Potassium permanganale is an altern ati ve to copper su lfate. Its crfect ive dosage usuall y limi ted 10 small reservoirs and is a te mporary measure be~ause. Ih~ car 0
ra nges from 0.4 10 4 mIL. Nanelheless, fac(ors suc h as low salubi li(y (5 gl l OO mL settles ou t rather qu ic kly unless a certain degree of agitatio n is provl ~ed In t le rese~­
a( 20"C), hi gher cost ($2000 per (011), and Ihe deve lopmenl of a purple color gen- voir. The effec tive dosage of PAC ranges from 2 to 10 mglL. Chlonne compollfl( s
era ll y discourage e ngi nee rs and plant operators fl'0111 using potassium perma nga nate.
 534 SPECIFIC WATER TREATMENT PROCESSES
such as calcium hypochlorite and sodium hypochlorite may be used to control al-
gae with 0.2 to I mglL of residual Iree chlorine. Yet their effectiveness is quickly
lost with exposure to sunlight and waterborne organic materials.
Groundwater Source Management Groundwater is frequently contaminated by
pollutants such as gasoline, industrial solvents, and a wide range of volatile organ-
ics. The prevention of groundwater contamination has just begun in recent years,
but this issue will most likely continue being a source of engineering concern for
several more years, since today's problems are generally the result of practices and
politics that have been ongoing for the past 10 to 50 years.
Reservoir Management In regions that experience the four seasons, most lakes
and reservoirs undergo thermal stratification. Under these conditions, a stagnation
zone (hypolimnion) is formed and the water below 25 ft (7.5 m) is nearly stagnant
dunng both the warm and cold months. Since the stagnation zone has little to no
dissolved oxygen, the sulfates and nitrates serve as a source of oxygen for bio-
chemical oxidation by anaerobic bacteria and the sulfate ion is reduced to sulfide ion
to form hydrogen suUide. The level of hydrogen sulfide formed in this manner may
be as high as 5 to 6 mglL. However, the well water in some areas may have hydro-
gen sulfide levels as high as 20 to 30 mgIL. Conversely, the water of impoundments
overturns twice a year-spring overturning and autumn overturning-due to the
change in water temperature. This phenomenon brings the hydrogen sulfide, as well
as other odor-producing compounds and unfavorable elements, to the surface.
. Positive and effective control of hydrogen sulfide formation may be attained by
mixing the entire body of lake water, thus preventing the formation of a stagnation
zone within the reservoir. This artificial destratification control method has proved
to be effective for most reservoirs. Mixing may be achieved through the use of me-
chani~al n~ixers, pumps, or diffused air pipes. However, the mixing is not always
effectIVe since the shape, length, and depth of the lake, as well as the design and
location of the mixer, strongly affect effective mixing.
Removal at the Treatment Plant If the preventive measures are ineffective or
impractical, the problem of taste and odor must be controlled at the treatment plant.
The three basic measures are (I) aeration, (2) oxidation, and (3) adsorption.
Aeration Aeration is a practical solution in controlling taste and odor problems
caused by volatile compounds such as hydrogen sulfide. It is generally not the best
method in controlling geosmin and MIB, the most frequent taste- and odor-
producing compounds. Examples of the aeration process include diffused mechan-
ical nozzle spraying, multiple tray cascading, and packed tower type.
Oxidation In most cases, oxidation is the most practical and effective method of
controlling taste and odor problems. Oxidation can be potassium permanganate ox-
idation, ozonation, ozone and hydrogen peroxide oxidation, chlorine dioxide, or
chlorination. The most frequently used oxidizing chemical is potassium perman-
TASTE AND ODOR CONTROL 535
ganate. The dosage ranges from 0.5 to 5 mg/L: however, taste and odor are com-
monly reduced to an acceptable level of 3 TON through the use of 0.3 to 0.6 mgIL.
Ozonation is also an effective method of oxidation. This process actually changes
the characteristics of the odor and tlavor, in addition to reducing the level of the
odor-producing compounds. The effective dosage of ozone ranges from I to 5 mg/L,
with 2 mglL being the average crfective dosage. With this method, the key to suc-
cessful taste and odor control is to provide alternative chemical feed points and an
alternative chemical feed system in case it becomes necessary to use an alternative
oxidant. Several pilot studies conducted in recent years indicate that the concurrent
use of ozone and hydrogen peroxide mayor may not be superior to the use of ozone
alone. Thus bench test or pilot studies should be conducted to evaluate the effec-
tiveness of this method.
Chlorine dioxide has been used effectively in the past. However, its use is now
discouraged by both the EPA and the departments of health services in certain states
due to the potential harmful effects of chlorite on human health. Similarly, chlori-
nation is also an effective method of taste and odor control, but its use must be care-
fully evaluated due to the possible formation of DBPs and chlorophenol (chlorophe-
nol is produced when phenol is oxidized in water). Generally, if any taste and odor
problems are anticipated, the water treatment plant must have a potassium perJna~­
ganate and a chlorine feed system. The design of these chemical feed systems IS
covered in Section 4.1. It is also essential to conduct bench-scale studies to help se-
lect the optimum type and dosage of oxidant prior to plant design.
Figure 7.3.2-1 illustrates the relation between MIB concentration and ozone
dosage with a contact time of 2.5,5, and 10 min. Figure 7.3.2-2 presents the changes
in flavor at various ozone dosages and contact times. The required ozone dosage to
remove MIB is specific to individual water. However, a general idea of the dosage
and removal rate can be obtained from Figure 7.3.2-3.
Adsorption The two basic and effective adsorption processes in removing taste and
odor arc the addition of PAC to the process water and the use of GAC adsorption beds
(to tilter the water). The use of PAC adsorption is most appropriate in situations where
moderate taste and odor problems are infrequent: midsized to small plants. The two
basic types of PAC storage and feed systems are dry carbon storage and dry feeding.
The rule of thumb dictates that if the hourly feed rate of PAC is less than 150 lb, the
PAC dry feeder should be used in combination with a solution mixing tank or a vor-
tex mixer. If the hourly feed rate exceeds 150 lb and there is a need for li·equent PAC
addition, the use of a slulTY feed system should be considered. The PAC dosage varies
from I to 50 mgIL. However, based on past experience, 25 mgIL should be conSid-
ered to be the maximum allowable dosage, regardless of the type of feeding (dry or
slurry). Refer to Section 4.1 for the design consideration of the PAC feed system.
Whenever a moderate to severe chronic taste and odor problem eXists, the use
of GAC adsorption beds should be considered. GAC adsorption is similar to ordi-
nary tilters; however, the empty bed contact lime (EBCT) is a very important ~e­
sign issue. The EBCT values typically range from 3 to iO minutes (mostly 3 mm)
for the purpose of taste and odor removal; the filtration rate of the GAC adsorptIOn
 536 SPECIFIC WATER TREATMENT PROCESSES
TASTE AND ODOR CONTROL 537

~~------~-------r------~ 100r------r------r------r------r---~
,, _ 2.5 min °.J_C.JJ,....:.O_J_ C
°
,, Contact
5 min
80 .................... . ..· ...... ·.. ·..·*·.. 1···· ...... ·.... ·.. •..

.... '
••• 0-.. C
Contact
'0.
... ... _~_ 10 min C
..... '~ Contact
60 ...................(j
·····0 C+
40 ............................................. ·..if .. ·.. ·.... ·........................'1' .....................
Detection limit
------------------ -------
00o---------~I----------~2--------~3
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ------.. --i' -..---.... -.--
OL-____- L____
+*
____ ____ ____
Ozone dose (mg/L) ~L- ~ ~~ ~

Figure 7.3.2-1 Effect of ozone contact time on MIB removal. 0 2 3 4 5

Ozone Dosage ( mgIL)
2
beds ranges from 3 to 6 gpmlft (7.5 to 15 mlh). The standard size of the medium
Figure 7.3.2-3 MIB removal by ozone for several surFace waters.
IS approxImately thesame as.that of rapid sand filter beds, and the depth of the layer
I~ ~~pr?xl~ately tWIce the sIze. Operational experience gained from European fa-
~Ihtl~s IndIcates that the GAC beds need to be regenerated every 4 to 5 years. Further Figure 7.3.2-4 illustrates the effect of the BBCT on the removal of several dif-
eta I s on the GAC adsorptIOn process are given in Section 7.7. ferent types of odor based on a pilot study. It is very important to run a bench-scale
or pilot test so that the appropriate type and size of GAC may be selected and to
c::::J 0, Dose r::::::J 0, Dose find the optimum BBCT for dominant taste and odor compounds. Water treatment
Omg/L EZiZ3 0, Dos. _ O,Dose
plants that require frequent taste and odor removal by activated carbon should com-
0.5 mg/L 1.0 mg/L 2.0 mg/L
pare the cost and ease of operation and maintenance, as well as the level of main-
tenance work, of PAC versus GAC. Figure 7.3.2-5 presents an example cost com-
I Note: Chloramines were used In all tests.1I 1.5 mg/l parison between PAC and GAC for taste and odor control. The figure is created
I based on the following assumptions: an average plant flow rate of 114 mgd (5 m3 /s),
I a filtration (adsorption) rate of 5 gpmlft2 (12.5 m/h), cost of PAC at $1000 per ton
I
I and $1700 per ton for GAC. The capital costs for a PAC storage and feeding sys-
1 tem and GAC facilities with a regeneration cost of GAC are not included; thus the
I comparison is based only on the cost of material. This figure shows the initial place-
I Chlorine Chlorine
I ment cost of GAC for 3, 4, and 5 minutes of BBCT and an averaged annual cost of
I replacement (every 3 years). The cost of PAC is based on the duration of PAC feed-
ing from 0 to 365 days (12 months) at a feed rate of 5, 10, 15, and 20 mg/L. A gen-
Musty/Chlorine
eral idea for the cost of PAC and GAC under selected conditions can be obtained;
2 for instance, the graph shows the annual cost of GAC when replaced every 3 years
(this is considered normal) for 3 min of BBCT is approximately $270,000. The same
Chlorine
cost for PAC can be found where the four diagonal lines of PAC dosages, indicated
as 5, 10, 15, and 20 mg/L, intersect the vertical line of 3 min of BBCT. Thus if the
PAC dosage is 5 mg/L, a total of 3 months per year costs the same as GAC with 3
Raw o 2.5 5.0 min of BBCT. Also, PAC dosages of 10 mg/L for 45 days, 15 mg/L for I month,
Ozone contact time (min) and about 20 days at 20 mg/L yield the same cost as GAC with 3 min. If the days
of PAC feeding exceed these numbers of days, the cost of PAC is higher. Since PAC
Figure 7.3.2-2 Changes of flavor and odor intensity by ozonation.
ends up in the sludge and is disposed, there is no recovery cost. However, GAC can
 538 SPEC IFIC WATER TREATMENT PROCESSES TASTE AND ODOR CONTROL 539

c:::J Musty 1",,,',,,,1 Mustyl Conditions: a . 5m3/s (114 mgd)

Chlorine I'ZZZJ Sweet _ None Filtration rate .. 12.5 nVh or 300 mid (5 gpm'sf)
PAC . $l .tOMon

5 "G:AC~'~$1~~~~~o:n____-.______r-____~;=:::GA:c~~~oo~m:.n:I:Coo~'~~-,
12 ,
I
I
41- I 10

!)
I - oS•
c
'Vi I 0
E 6
c I .E
3
.5'" I - '"
.E
~

§
I 6
I "11:
,•
I - ,c
0
4
I E0
1 I "
I - 2
I

aw non-
I
o 1.02.0 3.0
I 0
0 4)(105 1.2)(106
4.0 5.0 6.0 7.0 8.0
GAC
filter -"E"E--_ _ __ Annual Cost of PAC or GAC In US Dollars
GAC eaCT (min) - - -_ _~-
Figure 7.3.2-5 Choice of PAC or GAC for taste and odor control.
Fig ure 7.3.2-4 Odor removal by GAC adsorpti on process.

be reactivated, and approximate ly 85% of the orig ina l GAC 1'0' . I I I I lide and almost 100% at pH 6. However, at pH levels of 8 and above, the reduced
noted th t h PAC' . . .. ' ., le usee. t SlOU d he
a t. e. . feed lllg and storage lac ri lly is bu lky and di rty and PAC s l .. su lfu r ex ists in the water as HS - and SO~ - ions, and Ihe amount of free H 2S is very
teneIs to so lrd ll y In the holdin t k ~ d I' . , lill Y small and ins igni fi ca nt.
will s rd 'f 'f I _ I . ,g an ,ee mc, and even the s ludge in the c larifier
ad 0 I I Yll . t lC ~ osa~e IS 11Igh. The effective PAC feeding rate for lllosllaSle and T he mas I common method of removi ng hydrogen sultide from water is oxida·
o r co ntro SlI uatlOn s IS 10 10 20 mg/L when MIB · d . . ti on. Adsorption by activated carbo n is also an elTective method but rather expe n·
the r bl ' " .In gCOSlll 1ll arc the cause of
7.3 ~_o em , Since the rel~lOva l rate of these c?lllpounds by PAC is abollt 50% (Tab le sive. The oxidation proccss used in Ihe fie ld o f water treatment inc ludes (l) aera-
. 1). Tab le 7.3.2- f p,ese nts the res ul ts 0 1 bench-sca le tes ts conducted to eV'rlu tio n, (2) chl ori nation, (3) ozo nation , and (4) potass ium perma nganate oxidation.
ale the el fccl lvcness of taste and odor remova l by P'AC alld 1'0 I . I ' -
d 'k ' ~ < ur Ol ler Iy plca oxi-
ant s, Spl IIl g 100 nglL eac h of MIB a nd geosmin to Jilt ered \Va ter. Aeration T here are three basic types of aeration: spraying, d iffused air, and mul·
tiple tray. T he aeratio n process is bOlh practical and sati sfactory if the level of total
7.3.3 Hydrogen Sulfide Removal sul fi des (expressed as H, S) is less than 3 to 4 mg/L. It is essential to adj ust the pH
of the wate r to below 7. preferably 6, prior to aeratio n. Under the proper pH con-
The rem ova l o f hydrogen sulfide i d ' d . d itio ns, aeratio n may li berate CO 2 and other volatile o rganic substances. The ef·
ad d ' S ISCUSSC separatel y lrom thc rcmova l o f othcr

u n~:t~~n~~II:gol~~:~~poun~s be~~use of.its u l~iqu~ c~laracteristics. In nat ura l water, fec li ve air/water fl ow rate ratio should be in the range 80 to 120: 1, and the hy -
SO H _ Ions, tlC s u ur species eX ists In live stable form s: HSO - S02- drau li c loadi ng for a tray aera tor- the type IllOSt commonl y used in well wa ler
a l;o e:~s't~ndtIHS . Otiller s pec ies s uch as thiosu lfate, po lysullid e, and pol;tilion~l~ treatme nt- is 7to 15 gp m/ft' ( 17.5 to 37.5 m/h). For diffused air aerat ion, the power
. I III le nal ura water Howevcr Ih . I requirement for the blower is approximately I kW per each mi lli o n ga ll ons of wa·
fo il . . I' ,. . ' , ey me nOl l lermodynamica lly stable. The
b I oW ing I~ allons li p CX lsts among the su lfur specics. In wa ler wilh a pH of 8 or ter treated ( I kW per 3,785 m'/d).
e OW, a nOlln al pH ra nge for nalura l waler H, S and HS - arc til d ' t'
of s If A H 7 ' ' " C Ol11l11anl arms
u ur. I P , approxllnalely 80% of the sulfur is in the form o f hydroge n suI . H, S + ;0, -; H, O + SO
 540 SPECIFIC WATER TREATMENT PROCESSES
TASTE AND ODOR CONTROL 541

TABLE 7.3.2-1 An Example of Geosmin and MIll Removal Rate by PAC, KMn0 ,
Ch, H 2 0 2 , and O,l 4 by ordinary filtration processes and the odor problem will remain. The use of sodium
pyrosulfite or sulfur oxide has bccn recommendcd by Monscvitz in treating this col-
Removal Rate (%) loidal form of sulfur.
Chemical Feed Rate (mg/L) Geosmin MIB
Powdered Activated 10 7.3.4 Control in the Distribution System
40 62
Carbon (PAC)
Powdered Activated Taste and odor problems may also occur in the distribution system, primarily as the
25 52 65
Carbon result of corrosion of pipe material and/or the growth of iron bacteria such as
Potassi um Permanganate Crellothrix and Lepfofhrix within the water main. Additionally, in cases where the
0.8 42 28
(KMn0 4 ) sulfatc content of the water is high and the water is allowed to stand in dcad-cnd
Chlorine (CI,) 2 mains, taste and odor problems may be compounded by sulfate-reducing bacteria.
45 33
Hydrogen Peroxide I These problems can be rectified only if design engineers minimize the use of dead-
50 72
(H,O,) end mains and provide blow-off and cleanout assemblies at strategic points in the
Ozone (OJ) 2.5 distribution system. Furthermore, maintenance crews must keep the water lines clean
94 77
Ozone and Hydrogen Peroxide 2.5 and 0.5 by periodical flushing of dcposits and microorganisms that accumulate within them.
97 95
Note: 100 ng/L each of Gcosmin and MIB were spiked to filtered water for these tests. In cases where the interior of the pipe is badly corroded, the main must be cleancd
by the "pig" and relined with new cement.
Watcr treatment plants can also minimize taste and odor problems by maintain-
Chlorination Chlorination is often used to oxidize hydrogen su!tide in well wa- ing an adequate level of residual chlorine to combat bacteria in thc distribution sys-
tel. The reactIOn IS
tcm and taking anticorrosion measures--either adjusting the pH of the water based
on the saturation index or through the addition of a corrosion inhibitor to the fin-
4CI, + H,S + 4H,O --> H,S04 + 8HCI
ished water.
Therefore, 8.3parts of chlorine are required to oxidize I part of hydrogen sulfide
Factors that affect this reaction include pH, temperature, and reaction time. The ox:
IdatlOn rate causes the pH to increase sharply to 6.5 to 7.0. BIBLIOGRAPHY

Oz~n~t~on The ~)zonation process is seldom used to oxidize hydrogen sulfide.

Ansleme, C. I. H., et aI., "Effects of Ozonation on Taste and Odor," 1. AWWA, 80: 10:45
(October 1988).
The elllctCncy of this process is affected strongly by the level of pH' the pH of th
water should be maintained at 7 or less. The reaction is ' e AWWA, Wafer Treatmellf Plant Design. American Water Works Association, New York, 1969.
AWWA Committee Report, "Research on Taste and Odors," 1. AWWA, 62: I :59 (January 1970).
H,S + 40 3 --> H,S04 + 40, AWWA Committee Report, "Artificial Destratilication in Reservoirs," 1. AWWA, 63:9:597
(September 1971).
The ratio of ozone to hydrogen sulfide is 5.7: I. Bernhardt, H. and Clasen, 1., "Flocculation of Micro-organisms." 1. Wafer SRT-Aq/la,
t

40:2:76-87 (1991).
PotassiumPermanganate Oxidation Potassium permanganate is a common Biederman, W. 1., et aI., "Destratification Using Air." J. AWWA, 63:7:462 (July 1971).
OXidant that IS oftcn used in the water treatment process. Craig, S. L., et aI., "Optimizing the Removal of Gcosmin and MIB by Powdered Activated
Carbon," J. AWWA, 80:3:73 (March 1988).
4KMn04 + 3H,S --> SO + 2K,S04 + 3MnO + MnO, + 3H,O Daughtery, J. D., and Morris. R. L., "Studies on the Removal of Aclinomycetes Musty Tastes
and Odors in Water Supply," 1. AWWA, 59:10:1320 (October 1967).
Th.is. reaction iIlustratcs that 6.2 parts of potassium permanganate are required to
Glasc, W. H., et aI., "Evaluating Oxidants for Removal of Model Taste and Odor Compounds,"
oXldtze I part of hydrogen sulfide. Once again, pH control plays an important role J. AWWA, 82:5:79 (May 1990).
In maintaining the pH in the range 6.5 to 7.
Khiari, D., et aI., "Sensory GC Analysis of Decaying Vegetation and Septic Odors," 1. AWWA,
Elemental_ sulfur exists in water in a virtually insoluble form and thus may be
89:4: 150 (April 1997).
rem?ved by. hltratlO~. Howev~r, sulfur may also exist in water in colloidal form, as
Knoppert, P. L., et aI., "Destratitication Experiment at Rotterdam," 1. AWWA, 62:7:448 (July
a mtlky whIte emulSton. In thtS form, the sulfur may not be removed satisfactorily 1970).
 542 SPECIFIC WATER TREATMENT PROCESSES
Krasner, S, W" et aI., ''Taste and Odors: The Flavor Prolile Method," J. AWWA, 77:3:34 (March
1985).
Lalezary. S., et aI., "Air Stripping of Taste and Odor Compounds from Water," 1. AWWA,
76:3:83 (March 1984).
Lalczary, S., et aI., "Evaluating Activated Carbons for Removing Low Concentrations of Taste
and Odor Producing Organics," J. AWWA, 78:11:76 (November 1986).
Lalezary, S., et al.. "Oxidation of Five Earthy-Musty Taste and Odor Compounds," J. AWWA.
78:3:2 (March 1986).
Laverty, G. L., et al.. "Quality Improvements by Reservoir Aerations," J. AWWA, 62:11:711
(November 1970).
Meadows, M. D., "Reservoir Management," J. AWWA, 79:2:26 (February 1987).
Means. E. G .• "An Early Warning System for Taste and Odor Control," 1. AWWA, 78:3:77
(March 1986).
Medsker, L. L., et a!.. "Odorous Compounds in Natural Water," Environ. Sci. Techno!.. 2:6:461
(June 1968).
Monscvitz, J. T., and Ainsworth. L. D.• "Treatment for Hydrogen Polysulfide," J. AWWA,
66:9:537 (September 1974).
Montgomery, 1. M., Consulting Engineers. Water Treatment: Principles alld Design, Wiley,
New York, 1985.
Mouchet. P.. and Bonnelye. v..
"Solving Algae Problems: French Expertise and World-wide,"
J. Water SRT-Aqua, 3: 125-141 (t 998).
Nelson, M. K., "Sulfide Odor Control," J. WPCF, p. 1285 (October 1963).
Nielson. 8. 1., "Reaeration Dynamics of Reservoir Destratilication," J. AWWA, 66:10:617
(October 1974).
Palmer, C: M,. "Algae in Water Supplies," in Water Supply, U.S. Department of Health,
EducatIOn, and Welfare, Washington, DC, 1962.
Powell, S. T., "Removal of Hydrogen Sulfide from Well Water," J. AWWA, 40: 12: 1277
(December 1948).
Raman, R. K., "Controlling Algae in Water Supply Impoundments," J. AWWA 77:8:41 (August
1985). '
Rashash, D. M. C., et al" "FPA of Selected Odorous Compounds," J. AWWA, 89:4: I 31 (April
1997).
Schiller, 8., "Vacuum Degasification of Water for Taste and Odor Control," 1. AWWA, 47:2: 124
(February 1955).
Silvey, J. K., "Growth and Odor: Production StUdies," J. AWWA, 64: I :35 (January 1972).
Steichen, 1. M., et aI., "The Effect of Lake Destratification on Water Quality," 1. AWWA,
71 :4:219 (April 1979).
Sunet, I. H., "An Evaluation of Activated Carbon for Drinking Water Treatment: A National
Academy of Sciences Report," J. AWWA, 72:1:41 (January 1980).
Symons, 1. M., et aI., "Impoundment Destratification of Raw Water Quality Control Using
Either Mechanical or Diffused Air Pumping," J. AWWA, 59:10:1268 (October 1967).
Symons, 1. M., et aI., "Mixing of Water Supply Reservoirs for Quality Control," J. AWWA,
62:5:322 (May 1970).
Tcerick, 1. R .• et aI., "Artificial Destratification in Reservoirs of the California State Water
Project," J. AWWA, 61 :9:436 (September 1969).
DISINFECTION BY-PRODUCT CONTROL: THMS. HAA" AND VOCS 543
Vajdic, A. H., "Gamma-Ray Treatment of Taste and Odors," J. AWWA, 63:7:459 (July 1971).
Walker, G. S.. et aI., "Chlorine Dioxide for Taste and Odor Control," 1. AWWA, 78:3:84 (March
1986).
7,4 DISINFECTION BY·PRODUCT CONTROL:
THMS, HAAs, AND VOCS
One of the dilemmas facing the water industry today is how to minimize the po-
tential DBP risks while maintaining protection from microbial contamination using
adequate disinfectants. If the level of applied disinfectant increases, the microbial
risk decreases but the DBP risk increases. Trihalomethanes (THMs) and haloacetic
acids (HAAs) are major by-products of chlorine disinfection. Synthetic volatile or-
ganic compounds (VOCs) are industrial wastes frequently found in many ground-
waters. These substances are categorized by the EPA as a threat to publIc health
since most are suspected of being carcinogenic compounds. Since THMs, HAAs,
and VOCs are organic contaminants that predominantly cause problems in the field
of water supply for both treatment practices and cost, among others, the control
methods are discussed briefly.
7,4,1 THMs
THMs are formed by chlorination of natural organic matter (NOM), often described
as an organic precursor, by free residual chlorine. Organic precursors are generally
humic substances and fluvic acids. Chloramines are also capable of producing
THMs but at such low levels that this process is usually considered to be a non-
THM-forming process.
In 1979, the EPA established a maximum contaminant level (MCL) of O. I mg/L
for the total trihalomethane (TIHM) concentration, but Stage I of the DIDBP Rule
has lowered it to 0.08 mg/L. TIHM is defined as the sum oftrihalomethane (CHCI,),
tribromomethane (CHBr,), bromodichloromethane (CHBrCl,), and dibromo-
chloromethane (CHBr,CI).
To establish THM control methods and to comply with THM regulation, the char-
acteristics of THM formation under different conditions must be understood. There
are four basic definitions for THM:
1. Instantaneous THM. Instantaneous THM is the THM concentration at the mo-
ment of sampling. Compliance with the MCL of THM is determined by this
value.
2. Terminal THM. THM concentration at the farthest end of the distribution main
is know as the terminal THM. If a bench-scale test is used to evaluate termi-
nal THM, the test should simulate the actual chlorine contact time in the treat-
ment plant, plus the anticipated detention time in the distribution system.
Furthermore, the temperature and pH of the sample water should be identical
to the actual water treatment and supply conditions.
 544 SPECIFIC WATER TREATMENT PROCESSES
3. THM [ormation potelltial. The THM formation potential is determined by sub-
tracting the mstantaneous THM concentration from the terminal THM.
4. Maximum fotal THM potential. This value represents the maximum THM Con-
centration level under the most favorable conditions for THM formation. The
test condition as defined by the EPA requires either a period of 7 days at a
water temperature of 77"F (25"C) with a minimum of 0.2 mg/L of free chlo-
rine residual, or an initial chlorine concentration of 5 mglL with a pH of 9 to
9.5 for a period of 7 days.
The rate of THM formation and the terminal THM concentration depend on six
major factors: (I) the amount of organic precursors, (2) level of free residual chlo-
nne, (3) water temperature, (4) water pH, (5) bromide concentration in the water
and (6) chlorine contact time. ' design of a new water treatment plant and during the modification of an existing
Precursors If organic precursors are not present in the water, there will be no THM
formation. T?us the concentration and type of precursors have a direct impact on
THM formatIOn. The purpose of the maximum total THM potential test is to mea-
sure the quantity of organic precursors available to react with free residual chlorine.
Level of Free Chlorine Residual Since THM is a by-product of chlorination
the amount of chlorine residual directly influences the level of THM formation:
Therefore, the initial mixing conditions and the design of the contact tank also af-
fect the rate of THM formation.
Water Temperature Water temperature directly influences the rate of THM for-
mation: a hig~er temperature induces a faster rate and therefore results in a higher
rate of formatIOn. In many cases, THM levels during the summer months, when the
water temperature is 77"F (25"C), is nearly twice as great as THM levels during the
WInter season [water temperature is generally 39"F (4"C)]. Thus no special THM
contr~1 meas~res are n~cessary during the winter months. Although water temper-
ature IS a major factor In the rate of THM formation, the type and concentration of
organic precursors also greatly affect this phenomenon.
pH of the Water The rate of THM formation and terminal THM are greatly af-
fected by the pH of the water. The higher the pH, the faster the reaction rate and
the higher the THM levels. This phenomenon is believed to be due to pH-induced
changes in the functional groups of the precursor molecules.
Bromide Concentration The rate of reaction between bromide and the precursor mol-
ecules IS faster than that of chlorine and the precursors: the level of THM increases
wit?in a shorter period of time. If bromide is present in the process water, the concen-
tratIOn of tenninal THM is higher than the chlorine compounds. The mechanism in-
volved in this reaction is that the bromide ion is oxidized by free residual chlorine to
form bromine, which is capable of quickly reacting with precursors to form THMs. In
DISINFECTION BY-PRODUCT CONTROL: THMS. HAA,. AND VOCS 545
coastal regions the bromide ions are often supplied to surface water, as weB as ground-
water, from seawater. Yet bromide ions may also be present in water obtained from many
other sources, such as in regions that were below sea level during prehistoric times.
Chlorine Contact Time The THM level is a function of chlorine contact time:
The longer the contact time, the higher the level of THM concentration. This phe-
nomenon is particularly distinct during the first several hours but generally tapers
off after 2 to 3 days.
7.4.2 THM Control Measures
THM control measures should be implemented in two basic types of situations: the
plant (due to high THM levels). The THM control strategies are the same for both
cases and the basic steps are as follows:
Step I. Look for alternative local water supplies such as groundwater. If alter-
native suppJies do not exist or are not economically feasible, the second
step should be evaluated.
Step 2. This step contains several options, which are modifications of the con-
ventional treatment technique.
(a) Improve flocculation and sedimentation to increase the removal of
organic precursors (enhanced coagulation).
(b) Change the points of chlorination to shorten the contact time.
(c) Substitute potassium permanganate for prechlorination and use small
amounts of chlorine dioxide as preoxidant.
(d) Use chloramines as an alternative disinfectant. This option should
include the addition of ammonia 10 to 20 min after chlorine appli-
cation and observation of the THM formation level.
(e) Add PAC to remove precursors and THM (seasonal use).
If any of these options appears feasible after bench-scale test evaluation, a pilot-
scale test should be conducted (or even testing in an actual plant) to confirm the ef-
fectiveness of the option and to establish the anticipated costs. Should any of the
second-step options fail, a more drastic and expensive treatment technique should
then be evaluated (step 3).
Step 3. The alternatives in this step of the THM control strategies are as follows:
(a) Use preozonation to oxidize and remove organic precursors.
(b) Strip THMs through the use of an aeration tower.
(c) Provide a GAC adsorption bed to remove THMs.
(d) Utilize a membrane process, either nanofiltration (NF) or reverse
osmosis (RO), to remove THMs.
546 SPECIFIC WATER TREATMENT PROCESSES DISINFECTION BY·PRODUCT CONTROL: THMS. HAAs. AND vocs 547

7.4.4 Operational Issues

SOOr-------,-------~------~------~
1 I I
The plant operator must run the treatment plant in a manner that produces a processed
400 i- /O-.%.A... .f>-.o.. I - water that is in compliance with the MCL of the THM S1andard. Refer to Figure
_ Cf"">-<>-'</' \ .......... ..0.& 'o.C"'~....,j 7.4.4·1 for the appropriate steps. The lirst of these steps is the quarterly sampling
requirement. The quarters end on March 31, June 30, September 30, and December
~ 300 - 0 Raw water - 31. The THM levels of these samples must be submitted to the regulatory agency
~ , Settled
::& • Filtered (GAe) 30 days aftcr receipt of the analysis. The sampling requirements and the conditions
~ 200 I- ,
~
,\ ,I(. . __ x_x ,," ,"....,~
Jr""X' ". "
a Postozanated •
I I"
... • ..,t
,
- for reduced sampling frequency are different for surface-water and groundwater sup-
......./.... ..~
plies.
\1" ~ :c x ... """
100 '- ,:A:.!-

~~~~~~~"~~==:~'.'~~~"':":":~'=':"~:~~":'::'~:'=~~--~
Quarterly Sampling for Surface Water Supplies Each treatment plant is re·
quired to obtain and analyze a minimum of four samples per quarter, The samples
o~ 1000 2000 3000 4000
must be collected from the distribution system within 24 h of each other, and at least
(Oays)(O) (5) (10) (15) 25% must be collected from a location that retlects the maximum resident time
(20)
within the system. The remaining samples arc collected from representative loca-
Sed volumes fed
tions in the distribution system based on the person served, the water source, and
Figure 7.4.2·\ Total potential trihalomethanc (TPTHM) reduction through pilot plant the treatment methods used.
processes. Compliance with the MCL of THM and HAA standards is based on a running
average of the four quarterly averages. If the instantaneous TIHM concentration
never exceeds 0.08 and 0.06 mg/L for HAAs after one year of compliance moni·
(e) Change to a water supply source containing a minirnallevel of pre-
toring, a request to decrease the monitoring ti'equency can be made to the primary
cursors.
agency. If the agency agrees to a reduced sampling frequency of one sample per
quarter, this sample should be collected at a point (in the distribution system) re-
Use of ozone and the GAC adsorption bed have been reported to be an effective
tlecting the maximum resident Lime.
m~ans (~f controlling .THM levels since the mid-1970s; technical articles concerning
Ihls subJect may be found elsewhere. Figure 7.4.2·1 is an example of the effective
Quarterly Sampling for Groundwater Supplies The quality of groundwa·
eonlrol of THMs using a combination of GAC adsorption bed and postozonation. In
tel' is generally more consistent and has a lower concentration of precursors than
thIs case,. th~ THMs were removed by GAe, and postozonation was necessary to
surface water. For these reasons, the monitoring requirement for THMs and HAA
combat sIgnificant leakage of microorganisms tram the GAC bed; the GAC bed is
are less stringent than for surface water supplies. Distribution systems that use
literally a nest of microorganisms, due to the abundant supply of nutrients.
groundwater as their sole source of water may have a monitoring frequency as low
as one maximum total THM and HAA potential test per year per aquifier. This sam-
7.4.3 HAA5 ple must be collected from a location in the distribution system with the longest res-
idence time. However, the primary agency will agree to this reduced frequency only
Haloacetic acids (HAAs) consist of nine compounds. However, the MCL for HAA
jf the utility can substantiate that the maximum total THM and HAA5 potentials are
IS appllca~le o~ly t~ the first five compounds, so it is expressed as HAAs: mono-
chloroacetlc aCId, dlchloroacetic acid, trichloroacetic acid, monobromoacetic acid
dibromoacetic acid, and chloral hydrate. '
. Under Stage I of the DlDBP Rule for HAAs, the MCL is 0.06 mglL. The MCLs
lor both HAAs and THMs are regulated on a running annual average based on quar-
terly samples. The best available lechnology to meet Stage I MCL includes enhanced
coagulation (enhanced softening) or granular activated carbon adsorption with an
empty bed contact time of 10 min. Like THMs, HAA, formation is higher with ele·
vated temperature and TOe concentration in water. The GAC process removes HAA
rather effectively in conlrast to THMs, which have a low atlinity for GAC. Howeve:'
the control strategy for HAA, is basically the same as for THM control. Figure 7.4.4~1 Steps required to comply with MeL of the THM regulation.

548 SPECIFIC WATER TREATMENT PROCESSES
less than 0.08 and 0.06 mg/L during periods of maximum total THM and HAA for-
mation (high temperatures and pH) andlor during periods of highest total organic
compound (TOC) concentrations.
7.4.5 VOCs
It has been estimated that approximately 20% of all groundwater supplies in the
United States are contaminated by synthetic volatile organic compounds (VOCs).
This constitutes a real threat to public health since approximately 80% of public
water supplies depend on groundwater and 95% of rural areas use groundwater to
satisfy domestic water needs. Volatile organic chemicals are widely employed in in-
dustrial, agricultural, commercial, and domestic activities. Yet it is believed that the
majority of VOC contamination (of groundwater) results from improper disposal,
leaks, or spills of industrial wastes. In some instances, VOCs may enter the potable
water as a by-product of chlorination: Certain joint compounds used in reservoir
membrane liners and certain types of plastic pipe or plastic pipe liner react with
chlorine to yield carbon tetrachloride.
Once the VOCs enter an aquifier, they can be transported over great distances
due to their low affinity for soil, the lack of opportunity to evaporate, and the min-
imal chance for biodegradation. Recent federal and state studies reveal that the VOC
most frequently found in the highest concentrations is trichloroethylene (TCE); tetra-
chloroethylene (PCE) ranks second. The most common VOCs and their sources are
discussed brietly.
Trichloroethylene (TCE) TCE is a nonllammable liquid with an odor resem-
bling that of chloroform. Its primary application is as a degreasing solvent in the
metal industry. Some household products, such as spot removers, rug cleaners, dry
cleaning liquid, and air fresheners, also contain TCE.
Tetrachloroethylene (PCE) Tetrachloroethylene is also known as perchloreth-
ylene and is therefore commonly abbreviated as PCE to distinguish it from
trichloroethylene (TCE). PCE is a colorless, nonflammable liquid with an etherlike
odor. It is used primarily as a commercial dry cleaning solvent. Other usages vary
from rug and upholstery cleaner; stain, lipstick, and rust remover; degrcascr (mctal
industry); textile scouring; to vegetable fumigant.
Carbon Tetrachloride (Tetrachloromethane) This VOC is a colorless, non-
tlammable, heavy liquid used as a fire extinguisher; in the cleaning of cloth; as a
solvent for oil Of fat; and as a fumigant. Its application in consumer goods was
banned in 1970 and ceased being used as an aerosol propellant in 1978.
1,1,1-Trichloroethane (Methychloroform) This compound is a nontlamma-
ble liquid that is virtually insoluble in water. It is used primarily for cold type metal
cleaning. Other applications include the cleaning of plastic molds and drains, leather
tanning, degreasing of septic tanks, and as a shoe polish.
DISINFECTION BY·PRODUCT CONTROL: THMS. HAA" AND VOCS 549
1,2-Dichloroethane This chemical is an intermediary in the manufacture . of
vinyl chloride monomers and chlorinated solvents such as TCE and PCE. It tS used
as a solvent in cleaning textiles, PVC processing equipment, and so on. It is also
used in the manufacturing of paint, adhesives, and other similar substances.
Vinyl Chloride (Chlorethylene) This compound is a colorless, tlammable gas
that is soluble in alcohol. ether, carbon tetrachloride, and benzene; it is also slightly
soluble in water. Vinyl chloride is used primarily in the manufacturing of plastic and
PVC resins. Since plastic and PVC products are widely consumed in the modern
world, vinyl chloride is used extensively by many industries.
Benzene (Benzol) Benzene is a colorless and highly flammable liquid, pos-
sessing a characteristic odor, which is fairly soluble in water. Derived originally
from coal, benzene has been used in the manufacture of a wide variety of goods:
medicinal chemicals, dyes, artificial leather, linoleum, oil cloth, varnishes, lacquers,
and numerous other organic compounds and goods.
7.4.6 VOC Control Measures
The VOC control measures can be classified into two basic schemes: (I) manage-
ment and (2) treatment. Each of these is discussed.
Management Control Measures The management alternatives are (I) finding
a new source, (2) blending contaminated water with water of a good quality,
(3) containment of pollutants, and (4) elimination of the pollution source. The first
alternative is generally not feasible since it requires a significant expenditure of cap-
ital or is simply not practical. The blending alternative may be attractive to some
water purveyors: those who own multiple wells or own both groundwater and sur-
face water supplies. Yet this second alternative is limited by the lack of system tlex-
ibility, insufficient dilution capability, and lack of quality water sources. The third
and fourth alternatives are theoretically good but are very difficult to implement be-
cause there is no easy way of locating the source of contamination, predicting the
migration of pollutants, or stopping the migration.
Treatment Control Measures The treatment control strategies are (I) strip-
ping by aeration, (2) removal by adsorption, and (3) a combination of aeration and
adsorption.
Air-Stripping Method The process of aeration transfers volatile substances from wa-
ter to air or from air to water. The former process is defined as air stripping. Air strip-
ping has been used to strip carbon dioxide, hydrogen sultide, and certain taste- and odor-
producing compounds from water. It has also been used successfully m the removal of
VOCs. However, air stripping has certain limitations: The process is temperature de-
pendent and is therefore not applicable to treatment plants located in cold climates; its
exhaust pollutes the air; and a high air/water ratio is required to remove VOCs effec-
 550 SPECIFIC WATER TREATMENT PROCESSES
DISINFECTION BY·PRODUCT CONTROL: THM S. HAAs. AND VOCS 551

li vely. Olh,er minor draw~ac~s include the po te ntial degradation of water quality by air- mova !. The basic formul a for des ig ning a packed tower was deve loped by chemi -
bor~e particulates: ~he oX ld~tl on of the reduced form of metallic inorganics, thereby pro- ca l e ng ineers prior 10 the mid-twenticth ce ntury. Howeve r, modilicati o ns of this ba-
~ olll1g fl,oc ~Ol1n atlOn; the IIlcreased potential for cOITos ion-relatcd problems due to the sic des ig n (for special water treatment appl ications) have bcen published by several
IIlcrcase III dl~solved oxygen; and increased biological growth within the unit. people in the water treatm ent industry a fter 1980; these articles are lis ted in the
Three basIc types ~f air- stripping methods have been tested an d applied to actual Bibliography at Ihe end o f Ihis seclion .
Ireallnenl planls: (I) dIffu sed air, (2) spray, and (3) packed lower. According 10 Hess
el al. (1983) (Flgure7.4.6- 1), if over 90% o f Ihe VOCs mu SI be removed, Ihe packed Adsorption Methods The three basic mel hods o f adsorplion used in Ihe removal
tower type o ~ al~ Slnpplllg IS the only method capable of meeting this requirement. of VOCs arc PAC, synlhelic res in adsorplion, and GAC. PAC adsorplion is not very
When des igning air-strippin g units, live basic design co nsiderati ons (minimum) promi sing in the removal of VOCs, due to the nature of the medium: Co mpounds
mu st be addressed:
wilh high molec ular weighls lend 10 adsorb more readily 10 PAC, bUI VOCs arc
low-molecular-weig hl compounds Ih al do no l readily adsorb 10 PAC unless very
I. C harac teri s ti cs o f the VOCs to be removed high d osages are used. Synlhelic res ins such as Amerson XE-340, manufac lured by
2. Tempera ture o f the water and air Rohm & Hass, were des ig ned to remove no npol ar organic compounds wi th low
3. A ir/water ra tio molecu lar weights. This alternative has had some promi si ng res ults.
4. Contact tim e Currenlly, Ihe mosl effeclive adsorplion process for VOCs is Ihe GAC bed.
Resull s obtained from pilot studies indi ca le Ihat Ihis Iype o f bed e ffeclively removes
5. Required s urface area for mass trans fe r
PCE and oiller VOCs. T he prime fac lors Ihal affect Ihe efficiency of ad so rpli o n arc
so lubility and affinity. Thus a cruc ial co nsideration for design e ng ineers is the po-
Generally, an a ir/wa ler rali o [(fI'/min)/( n '/min ») rangi ng Ii-o m 20: I 10 25: I is
Icnli al desorplion o f VOCs, THM , and HAA from Ihe GAC bed. The desorplion
required to .o btaln over 90% removal of both TCE and PCE. Based o n the result s
process can result from co mpetitive adsorption and from reversal of the adsorption
of severa l pilot-sca le s tudi es, thi s air/water ratio IllUSt be at least 20: J for THM re-
process due to a decreased innuent co ncentratio n. Since the adsorption rate is a func-
ti on of water temperature, higher temperatures improve the adsorption rate; pH docs
not affcc i Ihe adsorplion rate of VOCs.
Spray lowers Diffu sed
aeration
90 Combination of Aeration and Adsorption This alternali ve is an e lTec live mel hod
Crossflowl o f VOC remo val. Aeration redu ces the load in g of co nlaminanls 10 Ihe GAC bed,
tower I thereby decreas ing the frequency of rege neration for the bed.
99 I
I
I Packed
"g
>
99.9
I
I
tower
BIBLIOGRAPHY
~ I
Aieta, E. M., ct a!.. "Ad vanced Oxidation Processes for Treating Groundwater Contaminated
~
~ 99.99 with Te E and PCE: Pil ot Scale Evaluation," J. AIVIVA, 80:5:64 (May 1988).
...
~

Arora, H.. el al.. "DBP Occurrence Survey." J. AIVIVA, 89:6:60 (June 1997) .
Not feasible AWWA. "Organic C he mi cal Contaminants in Ground Water: Transport and Removal." AWWA
99.999
Sell/illar Proceedillgs, No. 20156. American WaleI' Works Associati on. Denver. CO. June
198 1.
99.9999 AWWA Researc h Foundation. Advanced Oxidalioll Process for COIlfrol of Off-Gas Emissiolls
CHCI, peE from vac Slril'pillg, AWWA Researc h Fund/AWWA. J. AIVIVA, 78:2:62 (February 1986).
AWWA Resea rch Foundation. Occurrence alld Removal of VOC frolll Ddl/king Watel;
Cooperati ve Researc h Report. American Wate r Work s Association. Denver. CO. 1981.
0 .1 1.0 10 1000 10,000
Bil ell o. L. J. . and Singley, J. E., "Removal of Triha lomethanes by Packed-Column and
Henry'S constan t (ATM)
Diffu sed Aerati on," 1. AIVIVA, 78:2:62 (February 1986).
Figure 7.4.6-1 Selecti on ofa feasib le aeration process forVOCs. (Adapted fro m A. F.l-less Black. B. D., et al.. "Reducing Cance r Risks by Improv ing Organic Carbo n Re mova l."
el a l. , AWWA Research FOllndation Report. 1983.) J. AIVIVA, 88:6:40 (June 1996).
 552 SPECIFIC WATER TREATME NT PROCESSES
Carl son. M .. ct al.. "Controlling DBPs with MOl1ochl ora minc." J. AWWA. 90:2:95 (Fcbru ary
1998).
Crittendcn. 1. c., ct al.. "Dcsig n Co nsidcrm ions for GAC Trea tmcnt of Organi c C hcmi ca ls,"
J. AIVI VA, 79: I :74 (January 1987).
Crittende n. 1. C .. et al.. " Usi ng GAe to Rcmove VOCs from Ai r Stripper O ff-Gas." J. AWWA ,
80:5:73 (May 1988).
Crozcs. G .. el aI., "Enhanced Coag ul at ion: li s Effeci on NOM Removal and C hemical Costs,"
1. AIVIVA, 87: I :78 (January 1995).
Ferguson. D. W.. et a l. . "Co mparing Peroxone and Ozone for Co ntrol of Taste and Odor
Compounds. Di sinfec ti on By-prod ucts and Microorgani sms:' 1. AWWA, 82:4: 18 1 (April
1990).
Glaze. W. H.. and Kang. 1.. "Advanced Oxidation Proccss fo r Trcating Ground wate r
Comaminated with TCE and PCE: Laboratory St udi es," J. AIVIVA, 80:5: 57 (May 1988).
Glaze. W. H., ct a l.. " Dcterrn ining Hea lth Ri sks Associatcd with Di sinfcctants and Di sinfcction
By,products: Resea rch Needs," J. AIVIVA, 85:3:53 (March 1993).
Glaze. W. H.. cl aI., "Evaluation of Ozonali o n By-prod ucts fro m Two Californi a Surface
Waters ," J. AIVIVA , 8 1:8:66 (August 1989).
Gordon. G .. el al.. "Mi nimi zi ng C hl orite Ion and Chl orate Ion in Drinking Wate r Treatment
wi th C hl ori ne Dioxide," J. AIVIVA , 82:4: 160 (April 1990).
I-Ianel. D. W.. " Des ig n and Economic Eva luati on of a Treatment or vocs from a Cont am inated
Groundwater." J. AIVIVA. 78:9:87 (September 1986).
I-land, D. \V., et al.. " Desi gn and Eva lu ati on of an Air-S tripping Towe r for Re moving VOCs
from Grou nd wa ter." 1. AWWA. 78:9:87 (Septe mber 1986).
Hand . D. W.. el al.. "Des lru ction of DBP Precursors with Catalyt ic Oxidati on." 1. AWWA.
87:6:84 (June 1995).
Hand . D. W.. et al.. "S implilied Model for Design of Fixed-Bcd Adso rpti on Systems,"
J. EIII'il'OlI. Ellg., 11 0:4:440 (Apri l 1984).
Hess. A. F. et al.. COllt/vl Strategy- Aeration Tretlfmcll t Techniqlle: OCCllrrefJce and Removal
oj VO Cs JtV/1l Drillking Watel; AWWA Resea rch Foundati on Report . America n Water
Works Assoc iat ion. Denver. CO, 1981.
Jacan ge lo, 1. G .. et al.. "Ozonati on: Assess ing Its Rol e in the Formati on and Contro l of
D isinfection By' products," J. AIVIVA . 8 1:8:94 (August 1989).
Jacange lo. J. G.. el aI. , "Selected Processes for Re movin g NOM: An Overvi ew," 1. AWWA ,
87: I :64 (January 1995).
Jacange lo. 1. G .. et al.. "U P With Pretreat ment for Rem ov ing DBP Precursors." 1. AWWA,
87:3:100 (M arc h 1995).
Ka vana ug h. M. c., and Tru ssell. R. R.. " Desig n of Aerat ion Towe rs to Strip Volatil e
Contaminants fro m Drinking Water," J. AWWA, 72: 12:684 (Dece mber 1980).
Krasnc r. S. W.. et al.. "The Occurrence of Disinfecti on By-produ cts in U.S. Drink ing Water,"
J. AIVIVA , 8 1:8:4 1 (A ugust 1989).
Lykins. B. W.. Jr.. et al .. "GAC fo r Controlling THM 's." J. AIVIVA , 80:5:85 (May 1988).
Mean s. E. G., et al.. "D/DB P Rcg ulati on: Iss ue and Ramificati ons," J AWWA, 85:2:68
(Fcbmary 1993).
Miller. R. 1. . et al.. Disinfec tio n By- product Formati on and Control by Ozonation and
Biolreatmcnt ." J. AWWA, 84: II :51 (November J992).
FLUORIDATION AND flUORIDE REMOVAL 553
Naj m. I. N.. et al.. "Effcc t of Bro mide and NO M on By·product Formati o n." 1. AWWA,
87: I : I 06 (January 1995).
Najm, I. N.. et al.. "Evaluating Surrogates for Di si nfectio n By- products." J. AlVWA, 86:6:98
(June 1994).
Pc rry, R. H .. and Gree n. D. W.• Chemical Ellgilleerillg f/mulbook. 6th ed .. McG raw- Hili. Ncw
York . 1986.
Pontiu s, F. W .. " DI DIlP Rule to Set T ight Stand ards." J. AIVWA, 85: 11:22 (Nove mber 1991).
POlll'm oghaddas. H., et al.. "Effect of Bromi de Io n on Format ion of I-I AA .~ During
C hl orinati on." J. AIVIVA , 85: I :82 (Janu ary 1993).
Ram, N. M., et al .. Significance and Treatment of VO C ill Warer SlIpplies. Lewis Pu blishers.
Chelsea. M I. 1990.
Siddiqui s. M. S., c t nl., " Facto rsAI'fccting DBI' Format ion Durin g Ozone-Bromide React ion."
1. AIVIVA. 85: 1:63 (January 1993).
Singe r, P. c.,ct al .• "Correlati ons Bctween Trihaloll1ct hancs and To tal Organic Halides Formed
Duri ng Water Treatment ." 1. AWWA, 8 1:8:61 (August 1989).
Singley. J. E.. Trace Orgallics Remoll(lf by Air Stripping. AWWA Researc h Foundati o n Report.
Ameri can Water Work s Assoc iati on, Denver. CO. 1980.
S noeyink, v.. "Adsorpt ion as a Treatment Process for Organic Contaminant Rcmova l from
Groundwate r:' PlVceedillgs of the AWlVA Semill(ll; American Water Work s Association,
Denver. CO. 1981.
Sy mons. J. M., Treatment Techniques for COllltVllillg T/·IM 's ill Drinkillg Ware/; EPA/6oo/2·
8 1- 156. U.S. Env ironmenta l Protecti on Agency. Washington, DC. Se ptembe r 198 J.
Symons. J. M .. et al.. "A n Advanced Ox idm ion Process for DBP Conlro!." J. AWWA. 87: I I :66
(Nove mber 1995).
Symons. J. M .. et al.. " Precursor Cont rol in Wate r Containing Bro midc," 1. AW\VA . 86:6:48
(June 1994).
Tan. L.• et al.. ';Co mparing Ozonation and Membrane Se pnrat ion for Color Rcmova l and
D isinfect i on By, Proclu ct." .I. AIVWA , 83:5: 74 (May 199 1).
Thompson, 1. D.. ct al.. "En hanced Softening: FaclOrs Inlluenci ng BBP Prec ursor Remova l:'
J. AIVWA , 89:6:94 (Jun e 1997).
Um phries. M. D .. ct al. . "A Study of Trih alomcthane Remova l by Pac ked·Towe r Aerati on."
J. AWIVA, 75:8:4 14 (August 1983).
Vrijenhoe k. E. M .. ct al. . "Removing Parti cles and TI-IM Precu rsors hy Enhanccd
Coagul ation." J. AIVIVA . 90:4:139 (Apri l 1998).
Wein berg, H. S., at aI. , " Formali o n and Removal of Aldehydes in Pl ants That Usc Ozone,"
J. AIVIVA, 85:5:72 (May 1993).
Westerhoff. P.. et al.. "NOM's Rol e in Bromide and Bro mine Formati on Durin g Ozonation ."
J. AIVIVA , 90:2:82 (Febru ary 1998).
7.5 FLUORIDATION AND FLUORIDE REMOVAL
During the earl y twentieth ce ntury, studi es co nfirmed th at drinking wate r contain -
in g Iluoride io n co nce ntrati o ns in excess of 1.5 mglL caused dental Iluorosis .< mot-
tied e name l of hum a n teeth) and th at the severity of mo ttlin g inc reased Wi th an
 FLUORIDATION AND FLUORIDE REMOVAL 555
554 SPECIFIC WATER TREATMENT PROCESSES

increase in the fl uoride leve l. Other stu dies found that dental caries were Icss preva- TABLE 7.5.1-1 Fluoride COI11I)ound s a nd Feed System
lent in people with mottl ed enamel and concludcd that it was desi rable for the pub- Item Na2S iF6 NaF
li c water suppl y to havc a fluoride ion concentrati on o f approximatc ly I mg/L to Li quid
Co mmerciall y avail able form Dry Dry
promotc optimal dental health. 98-99 90--98 23- 24
COl11lllcri cal purity (%)
The flu oride conce ntrati on of potable water should be mainta ined at an optimal t88.05 42.0 t44.08
Molecut ar weight (M W)
level based on the ligure set by the local health departme nt; Iluoridati on of the pub- 1.220 at t 7.5°C
Specific gravi ty
lic water suppl y is determined by pu blic vote in man y states. Consequentl y, design (10. 17 Ib/gal)
e ng ineers mu st design e ither a fluoride feed sys tc m to ma intain an optim al level o f 60.7 45.25 79.2
Percentage o f Ouo ride in MW
flu oride or a process that rcmoves flu oride whcn the water suppl y co ntains exces- 0.762 4.05 In fi ni te
Solubility (g/ IOO mL at 25°C)
sive levels of Iluoride. Freezing point o f the soluti on aO
e aoe - 22°C
tOO tb bags or 125-400 tb 55 gal drums or in
Shippi ng cont ainers
d rums or in bul k form bulk fo rm
0.25 0.45 0.3
7.5.1 Fluoridation Approximate cost ($nb)
Dry chemical storage Liquid storage
Feed syste m co mponents
General Considerations As me nti oned prev iously, the reco mme nded level of • Water softener if total tanks such as
hardness of tap water steel tank with
fluoridc in potable water is I mglL. The Californi a Departmc nt of Health Services
exceeds 75 mglL as Neoprene linin g
suggcsts th at thi s fi gure be a n annu al average at water temperatures near 15°C; the • A day tank wi th
CaCO,
recomme nded flu oride conce ntrat ion fall s to 0.8 mg/L durin g hot mon ths (s umm er) a scale to moni tor
• Two soluti on tanks for a
a nd rises to 1.2 mglL during cold seasons. T he reason for adj usting the levels is that daily usage
batch system
the average water intake by co nsumers vari es wi th the seasons- more wa te r is co n- • A scale to monitor dail y A reli able
sumed during warm seasons th an the winter month s. dosage meteri ng pump
Fluoride is not found in the free state; it always exi ts in combination with other • A reli able metering pump
clements. All Iluori de compounds di ssociate to yield Iluoride ion whe n added to the • A du st coll ector for d ry
process water. T he 1110s t freque ntl y used flu oride compounds arc sodium flu oride feeder or solution
(NaF), sodium sil icolluoride (Na,S iF6 ), and hyd ro lluosili cic ac id (H, Si F6 ). These preparat ion area
th ree co mpound s are chosen because o f the ir availab ility, rcaso nab le cost, a nd safcty.
Thc solubility of the lirst two compounds is very low and the max imum solution
Example Design Calculations
co nce ntrati on is 4 % for sod ium flu oride and onl y 0 .76% for sod ium s ilicofluoride
at 25°C. Hydro flu osilic ic acid is availab le in liquid form a nd its solubilit y is there-
forc inlinitc. Howevcr, the co ncentrati on of commerciall y available soluti ons arc Example I
generall y onl y 23%. Refer to Table 7.5. 1- 1 for a summ ary of the chemical c harac- Given A 20- mgd (0.88-111 3/s) water treatme nt plant in O kl ahoma req~lires a new
tcristics and th c major componc nt s of the feed system. flu oridati on systcm . T hc water supply consistentl y has a natural flu on ~e c.once n-
Both sod ium flu oride and sod ium silicoflu orid e arc available in powc red for m or tration of 0.3 to 0 .4 mglL. Hydrolluosilicic acid will be used for lluortd atlOn be-
line c rystal. Thus problems assoc iated with du st must bc addressed durin g prepara-
ca use of its easy handli ng and safety (to the operator).
ti on of the solution. Plant operators mu st wea r the propcr type of protec tive gear
and use an ex haust fa n wh en handling storage bags because the dust is harmful to
Determille
hum an health . Furtherm ore, if the hardn ess of the di ssolvin g water exceeds 75 Illg/L
(as CaCO,), it must be softe ned prior to pre paring the Iluoridc solution or the Ilu- (i) The required type and ca pacity of the feeder
oridc ion will react with calcium and m ag ncs iul11 ~ the res ultin g precipitate is capa-
(i i) The capac ity of th e day tank
ble of cloggi ng the feeders and th e piping syste m.
Hydro lluosilicic ac id is much easicr to handl e th an the other two compounds but (iii) The capac it y of the storage tank
is a very corrosive che mical (strong ac id). Morcover, if the process wate r has low (iv) The applicati on point
alkalinity, th e appl icati on of a vcry sma ll amount of thi s che mi cal, 10 produce a
dosage of I mglL, ca n signilican tly depress the pH of th e water. The design of these
Solution (i) Since the commerciall y available produc ts is a 23% solution and the
perce nt age of Iluorine in the hydro lluosilicic ac id is 79.2%, based on the molec u-
chcmical feed systems should follow the guide lincs presented in Secti on 4. 1.
 556 SPECIFIC WATER TREATMENT PROCESSES
lar weight derived from the chemical formula, the fluorine content (%) of the prod-
uct IS
79.2 X 0.23 = 18.2%
The dosages are
maximum dosage as fluoride ion
= (I - 0.3) X 8.34 X 20 = 116.8 Ib/day (53 kg/d)
minimum dosage as fluoride ion
= (I - 0.4) X 8.34 X (20 X 0.25) = 25 Ib/day (0.97 kg/d)
Knowing that the specific gravity of the solutions is 1.220 at 17.5"C, the weight
of the solution is
8.34 X 1.22 = 10.17Ib/gal
Thus the actual chemical dosages in gallons per hour are
maximum dosage of 23% solution = 116.8 Ib/day + (10.17 X 0.182)
= 63 gal/day or 2.6 gal/day (9.84 Llh)
minimum dosage of 23% solution = 25 Ib/day + (10.17 X 0.182)
= 13.5 gal/day or 0.56 gal/day (2.1 Llh)
The required capacity for the metering pump is therefore 3 gal/h (11.4 Llh). Since
the turn-down ratio of any metering pump is to: I, the minimum dosage is 0.3 gal/h
(1.1 Llh).
The type of pump that should be specified is a diaphragm-plunger combination
pump, because it is best for handling very corrosive solutions. The pump should
have Hastelloy C stainless steel parts, a polyethylene or neoprene diaphragm and
PVC pipings. '
Two identical pumps are provided, with one acting as a standby. The metering
pumps are preferably paced to the plant tlow rate.
(ii) The maximum amount of chemical used per day is 63 gal (240 L). Therefore
provide a day tank with a minimum capacity of 80 gal (300 L). This tank should
be set on a weighing scale to facilitate the monitoring of daily use.
(iii) Bulk shipments of hydrofluDsilicic acid are most cost-effective and are safer:
Operators do not have to handle the 55-gal (210-L) drums frequently. Since the ca-
pacIty of a tank (trailer) is 4000 gal (15 m'), a storage tank with a 5000-gal
(l9-m') capacity should be installed. This storage tank should be fabricated from
steel. for protection from tornadoes and potential vandalism, and lined with neo-
prene rubber to prote?t the steel ,from corrosion. Based on the average dosage, the
4000-gal supply proVIdes approximately 100 days' worth of storage. The tank is in-
FLUORIDATION AND FLUORIDE REMOVAL 557
stalled outside for reasons of safety. A heating system is not provided since the am-
bient temperatures in this area does not fall below -8"F (-22"C), which is the
freezing point of this chemical.
(iv) The application point for hydrofluosilicic acid is at the inlet to the c1earwell.
A diffuser with a static mixer is provided for good mixing. This location is chosen
because application of tluoride, as well as lime (for pH control), at the head of the
plant results in the loss of some of the added fluoride due to alum flocculation. Refer
to Section 7.5.2.
7.5.2 Fluoride Removal
The two alternative methods for removing excess tluoride from the water supply are
(I) chemical precipitation and (2) the process of ion exchange.
Chemical Precipitation The fluoride content of the process water can be re-
duced to an acceptable level by means of either conventional alum tlocculation and
sedimentation, provided that a large alum dosage (200 to 300 mg/L) is used, or a
lime softening process. Results obtained from bench-scale tests indicate that fluo-
ride concentrations up to 3.5 mg/L (raw water) may be reduced to I mg/L by em-
ploying alum dosages up to 350 mg/L. Lime softening also effectively reduces tlu-
oride to a safe level by producing calcium fluoride precipitates.
Ion Exchange Process There are presently only three types of ion exchange
material that may have practical applications in defluoridation: (I) tricalcium phos-
phate, including bone char or bone meal; (2) activated alumina; and (3) ion exchange
resins. The effectiveness of these materials has been proven by actual plant-scale
tests and/or pilot-scale tests.
Bone Char and Bone Meal These two materials are very similar in that they
are both comprised of processed bone material. Their only difference is the slight
variation in the manufacturing process and products. An example of detluoridation
by ion exchange using bone char is as follows:
Bone char, expressed as Ca,(P0 4 )" can be regenerated through the application of
I % caustic soda.
Activated Alumina (AI2 0 3 ) Fluoride removal by means of activated alumina has
been demonstrated by an actual plant (Gila Bend, Arizona) and several pilot-scale
studies. In each of these cases, granular activated alumina (filter sand size)-
approximately 5 ft (1.5 m) deep-was used to reduce tluoride from an initial level
of 3 mg/L to less than 0.5 mg/L using a filtration rate of 5 to 7 gpm/ft2 (12.5 to 17.5
m/h). These studies also demonstrated that it was necessary to adjust the pH of the
raw water to 5.5 prior to filtration to maximize the efficiency of fluoride removal.
558 SPEC IFIC WATER TREATMENT PROCESSES CORROSION CONTROL 559

Rege nerati on or the bed was ac hi eved throug h the additi o n of I % causti c soda so- 7.6 CORROSION CONTROL
luti on.
Businesses in the industry of wate r treatment and suppl y are hampered by the seri olls
Ion Exchange Resin Process Fluoride can be rem oved fro m process water problem of corrosion (iron pipes and metals) and degradation of concrete, Corrosion
throu gh th e usc of ordin ary cation exchange and ani on exchange res in co lumns protecti on is therefore a major design considerati on for des ign engineers. Painting,
arran ged in seri es. The lirst column exchanges its polari zed hydroge n with sodium coating, proper maintenance of the cathodic protecti on system (for metal structures),
to form the equi valent acid : and the continuous fcedin g of corrosion inhibitors to the distribution system are the
major tas ks of the plant operation personnel. Corrosion is far from being a simple phe·
nomenon. It is a combination of many factors: elec trochemical , physical , chemical ,
biological, and metallurgical . Since cOlTosion and corrosion control are subjects which
The hydrogen ll110ride is sllbseqllellil y removed by the second column by the fol - on their own could lill a book, only a hrief disc ussion is prese nted here.
lowing reacti on:

2R- + H, F, -> 2RHF 7.6.1 Corrosion of Metals

The electrode potential is th e potential difference betwee n a metal and a solution of
Des ign engineers should note that one problem docs exisl for all three types of ion its ions. It is ex pressed by the Ne m st eqll atioll :
exc ha nge: the di s posal of the co ncent ra ted Iluoridc. These processes cannot be im-
plemented unl ess proper disposal meth ods are L1sed or an appropriate site is found . RT P RT C
As a linal note, it Illust be stressed that both bench· and pilot·scale tests must be E = - In - = - In -
/IF P' /IF C'
condu cted prior to the desig n of fluorid e rcmoval process. These procedures are of
ex treme importance bccause defluoridati on is affected significantl y by the presence
where
of other ions in the raw wa ter fo r th ese ions compete with the flu orid e removal re·
ac ti on. 'I = va lency of the metal ions

T = absolute tempcrature
BIBLIOGRAPHY
R = gas constant
AWWA. \ValeI' Fluoridation: Principles and PraClice, Manual M4. American Water Works P = pressure at whic h the metal dissolves
Associati on. Denver. CO. 1977.
P' = osmoti c pressure of the solution
Hannon, J. A., and Kaliclunan. S. G.. "Delluoridation of Drinking Water in Southern
Californi a." 1. AIVIVA. 57:2:245 (February 1965). C = a constant
Kreft . P. H.. "Removal of Fluoride by Ac tivcd Alumina," M.S. thesis, Loyola Uni versity, Los
Angeles. 1986. C' = acti vit y of the metal ions in the solution
Maier, F. J. , "Delluoridation of Municipal Water Supplies." J. AWWA , 45:8:879 (August 1953). F = Faraday number
Raboski. 1. B.. and Miller. J. P. . "Fluoride Removal by Lime Precipilation and Alum and
Polyelectrolyte Coagul ati on." PlVceedillgs of the 29th Pllrdlle Industrial Waste Conference, The electrode potential is standard for a standard solution of its ions and is ex pressed
Purdue Un iversi ty. Lafayette. IN. 1974. p. 669.
as Eo. thus simplifying the NCl'llst equ ati on to give the ge neral equati on dclining
Savinelli. E. A.. "Dell uoridat ion of Water with Activa ted Alumina." .I. AWWA, 50: I:13 (Janu ary
1958).
the electrode potential of a metal at 25°C:
Sawyer. C. N.. Chem isl l), fo r SallifOry Ell gincers, McGraw· Hili. New York, 1960.
Smith. H. R.. and Smith . L. C.. "Bone Cont:1ct Removes Fluoride." Water Works £ lIg., 90:5:600 E = Eo + 0.058 100
0
C'
/I
(May t937).
Williams. R. B. , and Cull'. G. L.. Hal/dbook of Pllblic Walel'Syslellls, Van Nostrand Reinhold,
New York. 1986. In the case of iron, this equation becomes
Wu. Y. C.. and Nitya. A.. "Water Dc nu oridati on and Acti vated Alumina." 1. £ lIvilv n. E"g.
Oi l'. ASCE. 105:2:359 ( 1979). E = - 0.44 patient 0.029 log Fc""
 CORROSION CONTROL 561
560 SPECIFIC WATER TREATMENT PROCESSES

Table 7.6.1-1 presents the oxidation reactions and potential for the major types site. These factors create a localized concentration of a specific element such as
of common metal at 25'C. Metals with positive electrode potentials are classified oxygen, sulfate, or chloride, which is significantly different from other portions of
as noble metals (cathodic); those with negative electrode potential are base metals the water. For instance, the oxygen concentration beneath the deposits of debris is
(anodic). The corrosion of iron is caused primarily by electrochemical cell action. less than the surrounding water; thus this portion of the metal becomes anodic and
The presence or absence of oxygen alters the rate of corrosion. Other factors that acts as the starting point for corrosion. This type of corrosion mechanism is called
influence the corrosion of iron are the total mineral concentration, the presence of an oxygen differential cell. Coexisting with most oxygen cells are the correspond-
dissolved gases, the temperature of the water, the surface condition of the iron, the ing sulfate and chloride concentration cells. These coexist because chloride and sul-
flow velocity across its surface (pipe flow), and biological action. fate ions penetrate the deposit and become concentrated. The corrosion rate is in-
fluenced by the surface area ratio of cathodic and anodic metals: The larger the
General Corrosion of Metals General corrosion of metals is induced by gal- cathode area, the higher the corrosion rate, and vice versa. For example, a large an-
vanic action. A good example is the case where two dissimilar metals are connected ode will produce a general corrosion condition, whereas a small anode yields a se-
and exposed to water: One metal becomes cathodic and the other anodic, thus set- vere pitting type of corrosion.
ting up a galvanic cell. For instance, when a steel and a copper pipe are connected As mentioned previously, the surface condition of the metal greatly influences
(which is often the case in most household plumbing), the steel pipe becomes the the conditions for corrosion. When a metal surface has deposits of debris or when
anode. Since metal loss occurs at the anode, the steel pipe corrodes. As shown in alum floc or lime particles create nonaerated zones (after precipitation) beneath the
Table 7.6.1-1, there is a O.96-V difference in electrode potential between steel and deposits, corrosion begins in these anodic regions. Furthermore, if bacteria are grow-
copper, and the corrosion is therefore severe. The corrosion rate of metals that arc ing under the deposits, they create an oxygen-deficient (reducing) condition, which
close to one another in the table is less than that of metals that are widely separated. intensifies the corrosion condition. Corrosion of stainless steel often occurs under
Just as in the case where dissimilar metals are exposed to water of a uniform this type of condition. Anaerobic microbes also playa role in corrosion. The pres-
concentration, a galvanic current is also produced when a single metal is exposed ence of sulfate-reducing bacteria within the deposit creates a very aggressive con-
to water of varying concentrations (ionic strengths). The portion of the metal lo- dition, which further accelerates metal loss. Some anaerobic bacteria, such as iron
cated in the more concentrated solution becomes the anode and therefore corrodes. bacteria (i.e., LeptllOthr;x), actually consume ferrous iron and thereby increase the
This process is known as concentration cell corrosion. rate of corrosion.
As long as a treatment plant is characterized by poor design and/or has deposits
of debris on the metal surface, concentration cell corrosion may occur at any plant Secondary Corrosion Factors Secondary factors that influence corrosion are
(I) the concentration of dissolved solids in the water, (2) the level of dissolved gases
TABLE 7.6.1-1 Oxidation Potentials Against Normal Hydrogen Electrode at 25°C (in water), (3) water temperature, (4) stress corrosion, (5) fatigue, and (6) impinge-
ment attacks. Each of these is described briefly.
Metal Electrode Reactions Equilibrium Potential (V)
Magnesium Mg ~ Mg'+ + 2e -2.34 Dissolved Solids The effect of the dissolved solids content on water corrosivity
Beryllium Be = Be 2 + + 2e- -1.70 is a complex issue. Both the species and the concentration of ions are important fac-
Aluminum AI = AIJ+ + 3e- -1.67 tors. Some species, such as carbonate and bicarbonate, reduce corrosion. whereas
Manganese Mn = Mn 2 + + 2e- -1.05 chloride, sulfate, bromide, and nitrate ions markedly accelerate corrosion. In regard
Zinc Zn = Zn2 + + 2e- -0.762 to concentration, less than I mglL of copper ion can increase the rate of corrosion
Chrome Cr = CrJ+ + 3e- -0.71
significantly: Copper ions are deposited on the metal and act as an anode, causing
Iron Fe = Fe 2 + + 2e- -0.440
a severe pitting type of corrosion to occur. Household aluminum utensils have been
Nickel Ni = NiH + 2e- -0.250
Lead Pb = Pb2 + + 2e- -0.126 known to corrode as the result of copper ions.
Hydrogen H, =2H++2e- -0.000 by convention
Copper Cu = Cu2+ + 2c- +0.345 Dissolved Gases Dissolved gases such as free carbon dioxide, oxygen, hydro-
Copper Cu =Cu++e- +0.522 gen sulfide, and ammonia strongly influence the corrosion of metals, especial1y iron
Silver Ag ~Ag-+e- +0.800 and copper.
Platinum Pt = Pt2+ + 2e- + 1.2 approx.
Gold Au = Au J + + 3e- + 1.42 Water Temperature Higher water temperatures accelerate the rate of corrosion
Gold Au =Au++e- + 1.68 by increasing the rate of cathodic reaction. The chemical reaction rate generally dou-
Source: Adapted from Degremonl, Water Trelltmellt Handbook, 5th ed .. Halsted Press. New York, 1979. bles for every 15°F (8°C) increase in temperature.
 562 SPEC IFIC WATER TREATMENT PROCESSES COR ROStON CONTROL 563

Stress Corrosion Metals that are under te nsile or bending stress, du e to ex ternal
loads or internal stress locked into the metal durin g fabri cation , will exhib it stress
IN,HIL.A n""" IHSVlATINC IL[EYIE ON STUD
wro, _1 Nor ALl
''''SVLATINQ WAY THROVCN. r.ACH
corrosion whe n s ubjected to a corrosive e nvironment. UlAT'N' WAI .. 5'U

WAhl, • •
Fatigue Corrosion Fa ti gue corrosio n occ urs when a me tal und ergoes repea ted
st ress in a corrosive environment.

Impingement Attacks CavitHlio n, erosion, abrasion by wat erborn e g rit and other
solids, and eros io n of the protec tive coatin g on metal s urfaces are examples o f illl.
pingement attacks.

7.6.2 Corrosion Control for Metals

Five bas ic measures ex ist for corrosion control : ( I) th e selection of constru ction ma.
teri als, coat in gs/linings, and ins ulati o n; (2) appl icat io n of corros io n inhibitors ;
(3) installation of cathodic protecti on sys te ms; (4) install ation of sac rificial anodes;
and (5) pH con trol by alka li c hemi cal s.

Design Efforts When corrosion problems are ex pected for a treatm ent plant, the
des ign e ngineer may co nsider using corrosio n·resistant materials such as stainless
steel. concrete, plastic, or copper as a su bstitute fo r carbo n steel. If these materials
ca nno t be implemented, special coatings and/or linings sho uld be applied to the
metal to sepa rate it frol11 the co rros ive water or environm ent. Paints, plastic sheet -
ing, rubber, and cement are often used in thi s capac ity. Galvanic corrosio n res ult.
in g from the connec ti o n of di ss imilar metal s mu st be avoided by insulati o n.
No nconducti ve material s such as plasti c or sy ntheti c rubber (Figure 7.6.2- 1) should
be inserted between the metals.
Application of Corrosion Inhibitors Despite the best desig n effo rt s, most wa-
ter treatme nt sys tems in clude a sys tem that applies chemical corrosio n inhibitors to
the process water. The inhibitors control corros io n by sto pping the anod ic and/or
cathod ic reacti o n. Chromate, orthophosphate. nitrate, and silicate are anodic in-
hibitors; calc ium carbonate, polyphosphate, and zi nc are typi cal cathodic inhibitors.
There are some chemicals that act as both anodic and ca thodi c inhibito rs: mercap.
toben zothiazole, ben zotriazo le, and tolytriazole. These c hemi cals form an orga ni c
film of amines of the surface o f the metal , there by inhibiting corrosio n (s imilar to
painting). The corrosion inhibito rs most commonl y lIsed in the lield o f water treat-
ment are pol yphos phate (cathodi c inhibitor requirin g a dosage of approximat ely
2 mglL), zin c orthophosphate (an odic inhi bitor requirin g a dosage of I Ill g/L or less).
and sodium silicate (anodic inhibitor req uiring a dosage of 8 to 10 mg!L as silica).
Cathodic Protection Cathodic protection is used to lower the e lec trod e poten·
tial o f metals and is used to protect buried pipes or equipme nt ex posed to corrosive
water, includin g seawater. Cathod ic protec ti on is ge nerally effec tive in app lyin g an
e lectrical potcntial of 0.85 to I V.
Figure 7.6.2-1 Suggested insulation flange assemb ly. (Ad apted from A. W. Peabody. COlltrol
of Pipelille Corrosioll, NACE. Houston. TX. 1967.)
Sacrificial Anodes Sacrificial anodes reduce galvanic cell corrosion for steel
or iron structures. They are composed of zinc or magnes ium and become anodi c un-
der corrosive conditions, thereby supplying electrons to cathodic surfaces.
pH Control by Alkali Chemicals The IllOSt comlllon mcthod of corrosion con-
trol for municipal water distribution systems is pH control of water. Th is method
prov ides a slight ly positive (+0.2) Lange li cr Index or a Stability Index (Ryznar
Index) of 6 or below. The chemical generally fed to the process water is lime since
it provides calcium and alkalin ity, in addition to rai sing the pH. However, lime par-
ticles tend to settle out in the di stribution main, creating a condition co ndu c ive to
concentratio n cell corrosion. Thus causti c soda is often e mployed as a substitute for
lime, due to its ease of handling and because it e lim inates the precipitati o n of lime
particles.
Langelier Index (L1) = pH - pH,
where pH is the pH of the sa mple water and pHs is the saturatio n pH .
Stability Index (S I) = 2pH, - pH
 COR ROStON CONTROL 565
564 SPEC IFIC WATER TREATME NT PROCESSES

Fig ure 7.6.2-2 and Appendix 17 are g raphic al determinati o ns of pH and LI. SI ca n TABLE 7.6.2-1 Co rrosion Characteristics Versus LI and SI
be determined by simple calc ul ati on aftc r LI is obta ined. Tab le 7.6.2- 1 illustrates LI 51
the relation be tween corrosio n characteris ti cs and LI and SI values. + 0.5 to + 1.0 4-5
As a mea ns of corrosion control , so mc treatment plants feed lime ahead o f fil - Med iulll to heavy scaling
+0.2 to + 0.3 5- 6
Slight scale format ion
trati o n (p H ~ 8) in the belie f that thi s sche me wi ll e liminate the after-prec ipitati on o 6-7
Equili brium 7- 7.5
of undi ssolved lime in the di stributi o n sys tem. In act ualit y, thi s prac ti ce increases - 0.2 to - 0.3
Sli ght ly corrosive
the linished wa te r co ncentra ti on of a luminum ; the alu m fl oc in the filter influe nt is - 0.5 to - 1.0 7.5- 8.5
Medium 10 heavy corros ion
di ssol ved by the lime du e to the solubility c haracteris ti cs of the noc (refer to Figure
3.2.3- 1), and the level of residual aluminum is often 0.3 mg/L or more.
Plant operators should rea li ze thaL an adju stme nt in pH based o n LI a nd SI docs 7.6.3 Corrosion Coupon Test
not mean th at corrosio n does not occur. Indeed, in many cases, iro n struct ures be- A Ilum ber of tools have bee ll developed to measure corrosio n by in?irc~ t l11 ea n ~. One.
ca me corroded, alth oug h the water had a slightly positi ve Lange li er Index. This sit- method is the corrosion coupon test: Fig ure 7.6.3- 1 prese nts a typical lIl.stali alion 01
uation o ften occ urs w ith son water. The best way to mo nito r me tal corrosio n and the corros io n coupon tcst asse mbly. The major steps of thi s test arc as lollows:
to o ptimi ze corrosion inhibitor dosage or pH co ntrol by alkali che mi cals is to con-
duc t a continu ous corrosio n co upo n test. I. The pretest weigh ts of the metal specimens arc recorded prior to plac in g the
metals into the system for a peri od of 30 to 100 days.
2. Following remova l, the specimens arc cleaned, rewe ig hed, and evaluated Th~
j
loss of weight and type of corrosion (ge ne ral , pitting, etc.) arc observc( all(
recorded.

500 400 300 200 100

Ca ha rdness. mgfl asCaC03
Example :
o 6 7
pH value at saturation (pHs)
9

Gl0~t ~o ! ~t or
-
.b:ti ~12" min ~
90 lt .,Ol,t \ l¢r;ij
Raw water calcium =< 240 mg/L
Alkalinity
pH
Temperature
= 190 mgfl
"" 6.8
:0; 70 °F
aO l.i...J
d@j
qIP,prPluq-1
2 P,pln9-' - -b l ockll on
3 • T· conntcllOn - ( Alternate
From chart: pHs = 7.3 installation
~ . fl ow rolt 01 8 Qpm or
Langelier index ,.. 6.8 - 7.3 = -0.5
VtIO("Y 3 1 1/~ t (
Stability index = (2 x 7.3) - 6.8 = 7.8
Figure 7.6.2-2 Gra ph ical dete rmi nati on of p H .~ and Lange li er Index. (Adapted rrom Th e Figure 7.6.3-1 Coupon test asse mbl y. (Adapted from Til e Nafco Water Handbook. F. N.
Na/eo Waler Halldbook. F. N. Kemmer. cd .. McGraw- HilI. New York . 1979.) Kemmcr. cd .. McGraw- HilI. New York. 1979.)
 566 SPEC IFIC WATER TREATMENT PROCESSES CORROSION CONTROL 567

3. The corrosion rate is expressed as metal loss in mils per year, a mil being to th e process water. The hydrochloric acid produced by ch lorine and ammoni a is
0.000 1 in. One millyr is equivale nt to 0.025 mm/y r. capable o f degradin g concrete, provided that the water is sort and has very little al-
kalinity, alt hough its dosage is so small that it will not quickly attack concrete ex-
The corrosion coupon lest is occasionall y run for visual observation only. The cept at th e ammonia feed points.
results of the coupon test can vary widely depending on the condition of the coupon: Certain regions have water supplies containing a high level of sulfates. Since sul-
whether the metal was sa ndblasted versus polished, the locatio n oflhe coupon within fates arc capable of reacting with aluminates-a n ingredient of cement-and cause
the test system, the length of exposure (duration of the test), the type of metallurgy, the origi nal volume to swell by a fac tor of 2 to 2.5, they may be considered to be
and the qua lity of the water. the primary cause of concrete deterioration in water treatment structures. A number
of exist ing plants, especially those located in regi ons with a hi gh sulfate content in
7.6.4 Corrosion (Degradation) of Concrete the water, exhibit severe degradation of their concrete surfaces, especially in areas
continuously exposed to cascading water and nowing water: Concrete troughs in
Today, almost all concrete struc tures are composed of steel bar- reinforced concrete. sedimentation basins and filters, and the sidewalls of channels and basins. Engineers
Theoretically, the steel bars do nol corrode, due to the positive potential in the con- are therefo re strongly advised to consider a den se and high-s trength concrete with
crete, that is, until the coveri ng concrete [approximately 2 in . (5 c m) thick] degrades. type 5 cement when designing plants for these regio ns (su lfate co ntent over 100
The electrode potential of the steel bars (i n concrete) is approximate ly + 0. I V and mglL). Table 7.6.4- I lists the available types of ceme nt and their major ingredients.
is referred to as the normal hydrogen electrode for concrete al a pH of approxi- Concrete can also be deteriorated by strong alka li chemicals such as caustic soda,
mately 11 .6. Yel , in reality, corrosion of the reinforcing stee l bars is an important potassium hydroxide. and soda ash because the a luminum oxides in the cement (sec
concern for engineers si nce degradation of concrete or the formation of cracks in Table 7.6.4- I) tend to become soluble when in con tact with water with a high pH:
the concrete can be attributed to three causes: ( I) physical causes, (2) chemical at- The strong alkali c he micals react with the sodium molec ule in the molec ular struc-
tac k, and (3) bacterial act ion. ture. Thus engineers are advised either to selec t cement with a low a luminum con-
te nt (type 2 or type 4) or to coat the concrete surface with an appropriate type of
Physical Causes The degradation of concrete wi ll be accelerated if the con- paint. If ge neral-purpose cement is spec ified, a PVC membrane should be used, pro-
crete is excessively permeable. This characteri stic result s from using the improper- vided th at the liquid pH exceeds 12.
size gravel; the wrong mixing proportion of cement, sand , and grave l; and poor con-
crete casti ng techniques. If the concrete is poorly constructed- that is, has cracks Bacterial Action The third cause of concrete degradation is bacterial action. A
and cavities, and insufficient thickness of concrete over the steel hars, or badly con- typical example is crown corrosion of concrete sewer pipes. In this situation, the
structed expansion joints that leak water- these defect s will contrihute to its degra- co ncrete is deteriorated by sulfuric acid formed by anaerobic bacteria, which con-
dation. The last of th e physical causes is erosion res ulting from heavy ve hicle traf- sume waterborne sulfate. This phenomenon rarely occurs in water treatment plants.
tic, scouring by sa nd or grit, or a high velocity of water now [over 15 fllscc or (4.6
m/s)].
7.6.5 Special Construction and Design Considerations
Chemical Attacks Carbon dioxide, strong ac ids, ammonia, sulfate, and strong The co nstructi on supervising engineer mu st give special consideration to the con-
a lkali chemicals all attack concrete. Of the five, carbon d ioxide has the least effect tact between steel reinforcing bars and iron pipes passing through concrete walls.
becau se the high lime residue of the cement simply converts the carbon dioxide into Absolutely no direct contact should ex ist between these two elements. since con-
calcium carbonate. However. if a concrete surface is ex posed co ntinuously to water
containing over 15 I11g1L of CO 2 • its surface wi ll begin to degrade. In or ncar
TAIILE 7.6.4·( Major Ingredienls of Pori land Cement
large c ities, concrete corrosion due to exhaust gas from industries and ca rs has be-
come a serious problem. Type I Type 3 Type 5
Strong acids such as sulfuric, hydrochloric, nitric, and phosphoric, as well as or- Composition General Type 2 Hi gh Early Type 4 Sulfate
ganic acids, arc highly destructive to concrete. The only effective method of pro- (%) Purpose Modilied Strength Low Heat Resisting
tecting concrete from the action of these acids is to line or coal it with a suitable 3CaOSiO, 53 47 58 26 38
type of material, such as plastic membrane, rubber, or paint. 2CaOSiO, 24 32 t6 54 43
Allllllonia degrades co ncrete by replacin g calcium (in co ncrete) with amlllonia. 3CaAI, O, 8 3 8 2 4
This loss of calciulll from cement causes the concrete to corrode rapidly. In recent 4CaOAI, OJFe, OJ 8 t2 8 12 8
years, many water treatment plants began feeding amlllonia (in addition to chlorine) Total 93 94 90 94 93
 568 SPECIFIC WATER TREATMENT PROCESSES
tac t wi.1I prod~ce ga lvanic ce ll corrosion in the iron pipe. The sa me type o f prob-
lem will oce lli whene~er stee l members arc embedded in the rei nforced co ncrete
~ Iru ct ure. If the end 01 a s lee l member, whi ch is ei ther exposed (0 watcr or buried
III the g ro und , Il1U.Sl be welded 10 re inforc ing s teel bars (in co ncrete), it mus t be well
coaled or cathodically protected. fr~m ga lva nic ce ll corros iOIl . Des ig n e ng ineers
:r
sho uld be ery careful when des igning these types of co nt ac ts and arc adv ised to
make nOlallons on the drawings.
Addi~i o ~ a l. des ig n consi deration s in clude the lise of insu latio n fl anges in cases
wl~ere d lss lITIJlar metal pipes mu st be connec ted and in situati ons where (he pipelin e
CX I~S a concre te wa ll 10 enter the ground or a Waler ta nk (see Figure 7.6.2-1 ).
FUithermore, stray current electro lys is type o f corrosio n may occ ur in are"s nea r d i-
r~ct current (dc) tran s it sys tem s. It is also very important to spec ify a ~cns~ and
11Igh-stre ngt h conc re te [compress ive s lre ngt h of over 5000 psi (350 kg/c m')] re-
ducmg the .w~ter/cement. rati o to less than 0.45. Recen t case hi stori es have show n a
c1e~r supenont~ o f th~ hig h-stre ng th re inforced concre te structures again st both cor-
r~slOn an.d eroSIOn, It IS a l s~ known that overuse of form re lease age nts have an o b-
VIOUS. effec!. o n . the rate ~f co ncret e corros ion, For further informati on, re fe r to
COl/ llO/ of I lpe/Ille Co n vSlO1/ (Pea bod y, 1967), re leased by Ihe National As soc ia tion
o f PIpe CorrosIO n (AAPC) Eng ineers.
BIBLIOGRAPHY
Degremonl , Wat er Trealmellllialldbook. 5th cd .. Hal sted Press, Ncw York, 1979.
Edwards, M., el aI., "Effect of Selecled Anions on Coppcr Corrosion," J. AWIVA 86' 12'71
(December 1994). ' . "
~dwards, M.. el al. . "The Pilling Corrosion of Copper," 1. AWWA. 86;7;74 (Jul y 1994).
aJr, O. M. , et aI. , Water Supply alld Wa stewater Disposal, Wil ey, New York, 1954.
Go~, R. ~. , ct al., Copillg lVitlt Deterioratioll of COl/crete Sill/aces ill Adelaide's Earl)' Warer
. ' Report by IIle Englllecnng
FlltratlOlI Pial/Is . . .' and Waler Supply Depa rt ment Ade laide
Soulh Au stralia. ' "
Hatc h, G. B., and Ri ce, 0., " In llucnce of Water Composit ion on th e Corros ion of Steel "
1. AWWA. 5 1;6;7 19 (June 1959). ' Drinking Water Act. These amendments set the maximum contaminant levels (MCLs)
Holm . T. R.. and Schock R " POIen t'fa I EI'I'ects 0 I' Polyph osphate WaterTreatmcm Products
" .' M,.,
. on Lead Solub ility III Plumbing Syslcms." 1. AIVIVA. 8J;7;76 (J uly 199 1).
JOint Task ~roup on .Calcium Carbonate Saluration, "S uggested Mcthod s for Ca lculming and
Interpreting Calc IUm Carbonate Saturation Index." 1. AWWA, 82:7:7 1 (Jul y 1990).
Kemmer, F. N. , cd., Na lco Chemica l Co .. The Nolco Water Halldbook McGraw- Hili N.
York, 1979. ' , . cw
Korshin. G. v., et aI., "lnl1u ence of NOM on Copper Corros ion," 1. AlVlVA. 88:7:16 (Ju ly
1996).
Larson, T. E., "Chemical Control of Corrosion." 1. AIVIVA, 58:1:354 (Marc h 1966).
Montgomery, J. M., Consulting Engi neers, Water Treatment: Prillcif,le and De'll'gll
New York. 1985. . . Wiley.
Peabody. A. W., COllfrol of Pipelill e CorrosiOIl, Nat ional Assoc iat ion of Corrosive Engi nec rs
Houston. TX. 1967. •
GRANULAR ACTIVATED CARBON ADSORPTION 569
Pisigan, R. A.. Jr., and Singley. 1. E., "Effects of Waler Quality Parameters on Corrosion of
Ga lvanized Sleel." 1. AIVIVA. 77; II ;76 (November 1985).
Pisigan. R. A.• Jr., and Singley. J. B. , "Influence of Buffer Capac ity. Ch lorine Resid ual, and
Flow Ratc on Corrosion of Mild Stcel and Copper," 1. AIVIVA. 79;2;62 (February 1987).
Reiber. S, H., et a!. , "Corros ion Monitoring and Control in the Pacific Northwest," 1. AlVlVA ,
79;2;7 1 (Febru ary 1987).
ROlh berg. M. R" et al.. "Compulerized Corrosion Co nlral," 1. AIVIVA. 8S;7;26 (Ju ly 199J).
Singley. 1. E.. "Electrochemical Nature of Lead Cont am inati on," 1. AWWA. 86:7:9 1 (July
1994).
Stone. A,. et al.. "The Effects of Short Term Changes in Water Quali ty on Copper and Zi nc
Corrosion Rates." 1. AIVWA. 79;2;7S (February 1987).
Strum, W.. "Corrosion Studies," Pllblic Works, p.78 ( 1957).
Strum, W., "Investigat ion on th e Corrosive Behavior of Water," 1. Sallil. Ellg. Div. ASCE,
86;27 (1960).
7.7 GRANULAR ACTIVATED CARBON ADSORPTION
7.7.1 Purpose
Granu lar activated carbon (GAC) has recen tly been used as a s ubstitute for gra nul ar
filter medium or as an add itional process in the conventio nal treatm ent process, fo r
the re moval of organic compounds, including disinfecti on by-products, those produc-
ing taste and odor, pesticides, and o ther synthetic orga nic compounds. GAC has been
proven effecti ve in removi ng these various organic compounds and even organ ic mer-
cury. The GAC filter bed is capable of functioni ng in the same manner as ordinary
filters (removal of suspended matter) but also removes organic compounds. However,
the Europeans use GAC adsorptio n beds for the removal o f taste and odor, other trace
organic compounds, and to tal organic carbon, by installin g GAC co ntactors down-
strea m o f ordinary filters (i n most cases) preceded by a preozonation process.
In recent years, the GAC adsorption process has s udde nly ga ined a g reat deal o f
atten ti o n in the United States as the result of the 1986 Ame ndme nts to the Safe
for over 50 organic compou nd s, of which 26 are syn thetic orga ni c compounds. The
water treatme nt industry in the United States has recently begun to lake large steps
in improv ing the aesthetic quality of drinking wate r, by reducing co lo r, objectionable
tastes and octo rs, in add iti o n to making it safer. The G AC adsorpti on process is cer-
tainly capable of meetin g these object ives, Yet the operation and maintenance costs
arc s ig nilica ntly hi g her, due primar il y to the need to regenerate the G AC.
7.7.2 Characteristics of GAC
GAC may be manufac tured from a variety of mat eri al: wood , nut s hell s, coal, peat ,
or pe tro leum residues. However, GAC used in water treatm ent is usua ll y produced
from bituminous or li g nite coal th roug h s low furn ace heating under anaerob ic co n-
diti o ns, T he absence of oxygen e nsures that the coa l does not burn but is trans-
 570 SPECIFIC WATER TREATMENT PROCESSES GRANULAR ACTIVATED CARBON ADSORPTION 571

form ed into a porous carbon mat erial. This product is then activated by exposure to the li ght specific grav ity of the mate rial. The GAC heds must be hackwashed peri-
a mixture o f steam and air at a temperature of I 500°F (815 °C). The activation process odically to re move suspended matter and mi croorga ni s ms, eve n thou g h the mass
oxidizes the surface of the carbon pores, thereby allowing the surl~1ce to atfract and tran sfer zo ne is destroyed completely by backwashing. Figure 7.7.2·lh depict s th e
hold organic compounds. head loss across a clean GAC bed at various filtration rates.
According to the Ca lgoll Blllletill (Calgon, 1976), the physical properties of GAC
are as follows:
7.7.3 Design Considerations
Total surface area of GAC 890 to 900 m'/g Design engineers are advised to con duct the following seq uence of activities when·
Bulk density of wetted GAC 30 Ibm 3 (480 kg/m3) eve r GAC con taclors are being co ns ide red as part of a treatment process train:
Bulk density of dry GAC 28 Ibm 3 (450 kg/m 3)
Particle density 1.4 to 1.5 g/cm)
I. Survey intens ive ly all available literature on the subject.
Effec tive size 0.8 to 0.9 mm for Filtrasorb 100, 0.55 to 0.65
2. Collect all available water analysis data on the subject water. Identify the pos·
mill for Filtrasorb 200
Uniformity coefficient sible options for removing the organic compo und s.
1.9 for Filtrasorb 100, 1.7 for Filtrasorb 200
Iodine number 850 to 875 ]. Procure a fresh sa mple of water and condu ct a co mpl ete waleI' analysis ce n·
Abrasion number 70 to 86 tered on the organic co mpound s.
Moisture as packed 0.5 to 2% 4. Reevaluate the preliminary options and establish both the carbon usage rate
Cos t of drying GAC $ 1.0 to $l.2l1b ($2.2 to $2.6/kg) or $760 to and the pilot tes ting co ndition s.
$905/yd 3 ($1000 to $ 1 I 85/m3) 5. Conduct a GAC pilot study.
6. Establi sh the design parameters of the process after analyzing the pilot tes t
Figure 7.7.2·la illustrates the hydraulic characteristics of a GAC bed. The opti- data.
mum backwash rate is approximately 30 to 40% of the bed ex pansion rate, due to

fllfAASORB 100 A~D FILTRASORB;roo

The major design considerations for GAC co ntactors are:
~ I-=~~~r=r::....-,--.
rFor otho< ", po. d oiormin. rO<i ulrO'd OOWNFLOW PIlf.SSURf DROPTHROIJGH A D... e~ WA SH~O
SEGAEGATE08EO OF FILTRASORB 100 AND FILTAASORB 200
now • • follow o
I. The designed now rate and available head loss for thi s process
10 [''') - [,,") [""), 1I--I-tI--+---1
fl'
II....
Ft'
O"ph
0$
2. The naWre and amount of all organic co mpound s in the water supply
3. The MCLs of the regu lat ed organic co mpounds
4. The overall treatm e nt process train and eac h unit process
5. The type of GAC
6. The size of GAC
7. The filtration rate (s uperiicial velocity)
8. The empty beel contact time (EBCT)
9. The type of contact unit s
,. 10. The loca tion of the GAC con tactor in the process train
I 1. The regeneration freq uency and th e method of regeneration
" I---h't'-+ .- 1- - - -'11111--1---1
Designed Flow Rate The designed !l ow rate should allow for the present re-
" "
Sup" " " . ' V.'ocily (()PINF1') "
5 110 20 30 quired now rat e as well as future expansion of the facility (if needed).
S"p .. f l. ~ 1 Yoloclty (OPINFl,)

(a) (b) Nature and Amount of All Organic Compounds Both the physical and c hem-
Figure 7.7.2-1 Hydraulic characteri sti cs of typi ca l GAC medium beds: (n) backwash rate ica l properties of the organic compounds st rong ly affect the type and leve l of pre·
versus GAC bed expa nsion rate: (b) head loss ve rsus backwash rates. (Ada pted from Cnlgoll treatment prior to th e GAC adsorption bed. Alum coagulation and flocc ulation with
BII/lelill 20·60. 1976.) clarifier are effective sc hemes for hi g IHll0I ecu IUl'~ weight organic compounds; th e
 572 SPECIFIC WATER TREATMENT PROCESSES
majorit y of highly volatil e, low-molecul ar-weight compounds may possibly be re-
moved by air stripping upstream of the G Ae contact beds.
MCL of Regulated Organic Compounds Refer 10 Seclio n 2.4 for" delailed di s-
cussion.
Overall Treatment Process Train and Each Unit Process Refe r 10 Seclion 2.4,
especiall y Figu re 2.4.5-1 b.
Type of GAC The pore size of Ihe GAC is an impOrlan l fac lor in Ihe removal of
organ ic substances. For example, GAe with large pore volumes may be clTcctivc
in remov ing waterborne compounds of high molecular weight. M oreover, activated
carbon adsorbs inorganic ions in a preferential order:
J
H + > AI + > Ca 2 + > Li + > Na + for cal ions
NO ) > CI - for anions
Engineers are therefore advised to consull the literatu re provided by manufacturers
prior to selec ting the GAe medium.
GAC BED AS FILTRATION/ADSORPTION PROCESS A recent design of tilt ers used after
preozonation utlizes GAC media instead of traditi onal sand and anthracite coal to
rem ove carci nogenic ozonation by-products and red uce the high level of AOC in
the ozollated wa ter to the allowable level. Furthermore. the replacement of ex isting
regular filler media wi lh GAC and GAC capping over exisling filler beds has bee n
evaluated and implemented in quite a few existin g plants. not because of pre-
olonation, but to control tas te and odor or re move SOCs found in agri cultural, ur-
ban, and industrial wastes/pollution.
The re moval of ozonal ion by- prod ucls by a GAC bed largely depends on Ihe mi -
croorgan ism growth within the bed. A tilter bed composed of biological ac tivated
carbon fillers is called a bacterial active filter (BAF) or biofilter. Regu lar sand and
anthracite filter beds can also become BAF with preozonation to some degree, but
the concentration of biomass is much less than for a GAC bed. M oreover, the bio-
mass in regular filter beds is easi ly washed out during backwash wilh residual chlo-
rin e, in contrast to a GAC bed, which is capable of holding mos t of Ihe biomass. It
is important to reali ze Ihal Illost TOC removal is achieved in Ihe top 6to 10 in. (15
10 20 cm) of Ihe bed (aboul 1.5 min of EBCT). Therefore, il is WOrlhwhile 10 con-
sider cappin g regular filler beds wil h GAC. If Ihe liller is equipped wil h a surface
wash system, surface washing should be applied for less than 2 to 3 min to retai n
the biomass.
T~ere are three important design issues thaI must be addressed when using GAC
cappll1g and GAC with a layer (barrier) of sand or garnet at the bottom of the fil -
ter bed. The lirst is to selec t the proper size and specific gravity of GAC so that
bOlh Ihe GAC bed and sa nd or anlh rac ile layers are Iluidized 10 Ihe same degree
GRANULAR ACTIVATED CARBON ADSORPTION 573
during backwash; Ihal is, bOlh GAC and olher Iypes o f media in a filler cell musl
be hydraulically matched during the filter washing period . Otherwi se, GAC is ei-
ther washed out or becomes dirty and clogged with suspended matter and biomass.
The second design issue is 10 provide a proper Ud ralio across Ihe filler bed.
Since the filter bed docs not act as an adsorption bed but as a filtration and ad-
sorpli on bcd, Ihe lurbidily o f Ihe filler eflluenl should be less Ihan 0.1 ntu as a goal.
The Ihird issue is Ihe size and gradalion of Ihe GAC grains. For Ihe adsorplion
process, GAC of small size and wil h a hi gh uniformily coefficient is preferable. The
liltration process, however, req uires a larger size with a sl11all uniformity coeffi-
cien t; thi s is essen tial for the longer filter run and to achieve good effiuent quality.
Modern filler beds, including GAC, are composed of med ia wilh an effeclive size
of 1.2 to 1.5 mill, with a uniformity coefficient of less than 1.4, and have a bed
deplh of 6 10 7 fI ( 1.8 10 2.1 m) wilh or wilhoul a sand barrie r al Ihe bonom.
S ince a GAC bed is loaded wilh biomass such as bacleria, prolozoans, rol ifers
and eve n nemalodes-whi c h may polen Ii all y break Ih rough Ihe GAC bed-il is
slrongly reco mme nded Ihal a Ihin [10 in. (25 cm)] sand barrier (0.55 to 0.60 mm
effecli ve size) be placed al Ihe bonom of Ihe bed. II is imporlalll 10 recogn ize Ihat
coal-based GAC is preferred 10 wood-based GAC for filler media and TOC removal.
The Ihree filler bed design issues are discussed Ihoroughly in Seclion 3.2.7.
There are Ihree major operalional iss ues for Ihe GAC filler: ( I) 10 conlrol Ihe
biomass, (2) 10 avoid feeding cenain chemicals prior 10 Ihe filler, a nd (3) 10 prevenl
anaerobic condilions wilhin Ihe GAC bed. Ozone Irealmenl of a hi gh-NOM (TOC)
wa ter produces a high concentrati on of biomass during a warm -water period. Despite
reg ular air-scouring wash, initial loss of head increases with time, thereby unrea-
so nably shon enin g Ihe fille r run le nglh . A praclical solulion 10 Ihis problem is 10
add c hl orine 10 Ihe bac kwas h waler (0.3 100.5 mg/L residual) unlil Ihe inilial head-
loss drops 10 a normalleve!. A few planls always backwash GAC fillers wilh chl o-
raminated water to control biomass without sacrificing the adsorption capac ity of
GAC 10 any signilica nt degree.
CaUl ion mu sl be exercised whe n selecling Ihe type of c hemical Ihal is 10 be fed
upslream of Ihe GAC filler bed. Polymer feeding as a filler aid is common praclice
for regular hi gh- rale fille rs, bUI Ihis praclice pOlenlially decreases Ihe adsorpllon ca-
pacity of GAC and also creates a possible release of monomerous acrylamide. All
anionic and nonionic types of polymers are made from polymerized acrylamide.
Microorganisms in the biomass can potentially break the polymer chains and pro-
duce monomerous products; monomerous acrylam ide is a kn own carcinogen.
Potassium peflnanganate is often fed to con trol iron and manganese as well as tas te
and odor. If Ihis is pracliced for a long peri od of lime, Ihe GAC grains become
coated with manganese oxide, thu s the adsorption capac ity is reduced or even lost.
The Ihird operalional iss ue is unique to Ihe GAC bed. A GAC bed that is lefl as
a slandby (w ilhoul draining Ihe waler) for a long peri od will have an anaerob ic con-
di lion wilhin Ihe bed. When Ihi s GAC !iller is pul back on line afler a long slandby
period, there will be a bad laste and odor problem, and soluble iron and manganese
wi ll be released from Ihe bed. T he lasle and odor problem is Ihe resull of decaying
organic matter and hydrogen sulfide. The reduced form of iron and manganese re-
 5 7 4 SPEC IFIC WATER TREATMENT PROCESSES GRANULAR ACTIVATED CARBON ADSORPTION 5 7 5

leased from a GAC lilter will produce a hi gh color when c hl orin e is fed to the wa-
ter during the li nal di sinfec ti o n stage. Therefore, if a GAC bed is not used for more
th an 2 to 3 weeks (es pecially durin g summ er months), it is rcco mm ened th at th e

.
. 00
watcr be drained and th e bed be kept in an aerobic co nditi on by appl ying air scour.
~ ::c~ -tIL ~
ing at certain intervals.
3
00
)
! 1/'- /
U 7 ....... t . ....
Size of GAC The size of the GAC stro ngly inll ue nces both th e depth of th e bed z nC T _
a nd the head loss across the bed give n a spec ific flow rate. Coarse GAC has definite
0

..
u
00
nCT
V
adva ntages for applicati on to dee p-bed gravity cont actors under high li lt ration rates.
However, smaller GAC is charac terized by a hi gher rate of adsorptio n, give n
o.
:I 50 : I ""'"
"...... : I , ./
.."..
an equ al level of acti vity with a shall owe r bcd, since it has a greater ac ti ve surface 40
area. Calgo n Carbon Corp., one of the major GAC produ cers, provides va ri ous size
Z
/j ,,; V
ra nges of GAC, as show n below.
..
o.
o.
20
V ~
f-"""
F- IOO F-200 F-300 F-400 F-820 F-8 16 F-6 14 r-
Effec ti ve size (mm ) 0 .85 0.63 0.95 0.64 1.0 1.4 1. 8 o 20 40 00 00 '00 12 0 140 11 0 180 200
UC 1.9 1.7 2.3 1.6 1. 5 1.4 1.4 Tlwe TO BREA KTHROUG H (DA YS)
Iodine num ber 850 850 900 1000 850 850 850 Figure 7.7.3-1 EBCT versus THM breakthrough again st lime. (Reprinted fro m AWWA
Semillar Pm ceedillgs, Grallllla r AClivated Carboll '" slallalio" s: COllceplioll 10 Operatioll, by
permi ssion. Copyri ght © 1987. American Water Works Assoc iati on.)
NOTE: Both effective size and uniformity coe ffi cient (UC) are approximate numbers.
F· IOO through F·400 are products used primari ly for adsorpti on purposes. The
filt ration/adsorption lISC GAC are F· 820. F-8 16. and F-6 14. Type of Contact Unit The confi gurati on of the GAC bed may be upO ow, dow n-
now in parallel tra ins, or dow nO ow in se ri es trains. The upOow co nfi gurati on is not
Filtration Rate Ge nerall y, th e most commo n li ltrati on ra te is 2 to 5 gp mtrt' (5 to suitable for application in a potable water treatm ent process unless the reg ul ar fil -
12.5 m/h), although hi gher ra tes are also used. ter is located dow nstream of the upll ow GAC bed . Otherwise, the GAC grain s will
be present in the fini shed water due to the escape of line GAC parti cles d uring the
EBCT A practi cal EBCT is o ne that ran ges from 7.5 to 10 min . The shorter EBCT latter stages of the lilt rati on cycle: the res ult of bed ex pansion due to the develop-
Illay be used whe n re movi ng sy nthet ic organic co mpound s (SaCs) since these are me nt of head loss across the bed.
easil y adsorbed, but the longer contac t ti me mllst be used in removi ng the Illore so l-
ub le and less adsorbable orga ni c compounds. Location of the GAC Contactor in the Process Train The GAC contac t bed
Figure 7.7.3- 1 illustra tes the EBCT versus the T HM brea kthrough at EBCT val- can be placed upstream of the Ilocc ul ati on process, before the ordin ary filters, or
lies of 5,7.5 , and 15 min. The graph shows th at the EBCT shou ld be a minimum dow nstrea m o f the regu lar filters, depe nding on the natu re of the substance to be
of 15 min, give n a reasonable rege nerati on cyc le (a 6-l11onth to 2-yea r cyc le). It is removed and the result of the pilot study. The GAC co nlac t bed is often used as
very interesti ng to sec that increasing the EBCT fro m 7.5 to 15 min , increases the both a filtrati on a nd adsorption bed . Figure 7.7.3-2 is a cross sec ti on of a GAC ad-
carbon life signi fica ntl y. Simila r results have bee n obtained fro m a nu mber of pilot sorpti on bed des igned for the ci ty of C inci nnati .
tests, incl udin g the test prese nted in an EPA publ icati on e nt itled Treatmellf
Techniqlles for Controlling THM's in Drinking Water (1 98 1).
7.7.4 GAC Regeneration
Most o f the design pa rame te rs ( I th rough 8) may be evaluated by a GAC pilot
study, prov ided th at th e project has both the tim e an d the budge t to do so. The pi- GAC is a very expensive product ($2000 per ton); thus the exhausted medium can not
lot study should test a va ri ety o f pretreatm en t options, such as ozona tio n, aera ti on, be di scarded like spent PAC. Fortun ately, GAC is fairly easy to regenerate. The need
alum coag ulati on, sediment ati on, and fi lt rati on. Since the ac tu al raw water usua ll y to regenerate the GAC ranges from 6 months to 5 years, dependi ng on the amount ,
co nt ains a uniqu e combinati on of suspe nded solids, orga ni c co mpound s, and inor- type of orga nic contaminants and level of biological acti vities in the GAC bed. The
ga ni c compounds ind igeno us to the region, the onl y way to prope rly estab lish most two basic alternati ves for GAC regenerati on are ofT-site regenerat ion and on- site re-
of the des ign crite ria is to co nduc t a pi lot-scale study on the pla nt site. generation. TIle on-site alternative is generally not cost-e lTective unless the carbon ex-
 576 SPECIFIC WATER TREATMENT PROCESSES GRANULAR ACTIVATED CARBON ADSORPTION 577

is generally lost during this entire process. The types. ~f furnace commonly ~sed in
.•. '. thermal regeneration of GAC are multiple hearth, Ihlldlzed bed. and rotary kIln. All
of these are very expensive. ' fh
Both spent and regenerated GAC is transferred in slurry form ?~ me,ans a y-.
d 'aulic eductors. The slurry is transported through stainless steel plpll1g for most of
I distance; stainless steel screw conveyers are used to I'e.ed th e spent GAC to the
the
furnace. Figure 7.7.4-1 is an example of the off-gas handhng system.

Chemical Regeneration This method uses a solvent to extract adsorbed compounds

at a temperature of approximately 21 O"F (I <XJ"C) by means ot pH. The process was de-
veloped by the French firm Dcgremont and Its apphcatlon IS very limIted at plesent.

7.7.5 Capital and Operational Costs

The capital cost of a GAC adsorption system can be est.imated ~ased on the total
volume of GAC medium required for the system. The caplt~1 cost mcludes the GAC
medium, reactor cells, pipings, pumping, flow control, elcctncal, and concrete works.
The estimated capital cost of the system is shown below.

Volume of GAC (ft') 2,500 5,000 10,000 25,000 50,000 100,000

Cost ($ million) 0.48 0.84 lAO 3.90 7.80 13.0

The costs above do not include storage of the GAC medium or ~he spent GAC re-
activation system because they are site specific. The total cost IS generally 2.5 to
Figure 7.7.3-2 Cross section of a GAC adsorption bed. (Reprinted from AWWA Seminar
Proceedings, Gmnlllar Activated Carbon IflStaffatiofls: Conception to Operation, by per-
mission. Copyright © 1987, American Water Works Association.)
DRVER
OFF-GAS
haustion rate is very large (over 5 tons/day), Moreover, the additional air pollution from CYCLONE
the furna<.:e otT-gas can pose a great problem in large metropolitan areas, which already
have poor-quality air. The off-site regeneration area can be any site outside urban ar-
TO SCRUBBERS
eas where air pollution is not a major issue. In most cases, the GAC manufacturer is
AND STACt<.
capable of regenerating the spent GAC at its facility on a contractual basis.
The three methods lIsed in GAC regeneration are (I) regeneration by steam,
(2) thermal regeneration, and (3) chemical regeneration. Since only an outline of
these methods is presented, refer to the Carboll AdsOIption Handbook edited by
RECUPERATOR
Cheremisinoff and Ellerbusch (1978) andlor the manut;,cturer for further details.
REGENERATION
Steam Regeneration This method is limited to carbon that has been saturated FURNACE
with a limited type and level of volatile compounds. For further details, refer to the
Carbon Adsorption Handbook.
AFTERBURNER

Thermal Regeneration This is a pyrolytic process that burns off adsorbed com-
pounds. The dewatered GAC is heated to approximately 1500"F (815"C) under con- · 774.1 Off-gas control system. (Adapted from AWWA Seminar Proceedil.lgs,
FIgllre •• . ' ,. . ' C fight
trolled conditions: steam is used to prevent the carbon from igniting; The GAC is Grallular Activated Carbon Installations: ConcepllOll 10 OperatIOn, by PCI mISSIOI1. opy ...
then quenched with water It)!" reuse. Approximately 10 to 15% of the original GAC © 1987, American Water Works Association.)
 578 SPEC IFIC WATER TREATME NT PROCESSES
ION EXCHANGE AND MEMBRANE SEPARATION PROCESSES 579

3.5 limes I~ighcr th an the ~ost s hown abo ve, depending 011 the size of the fac ility
McCarty. P. L.. et al.. "Operati onal Ex pcricnces wi th GAC Adsorbers at Water Factory 21."
and ot~cr lac lors, S,uch as If the sto rage and the reacti vat io n fac ilit ies are inc luded. J. AIVIVA, 7 1: 11 :74 1 (Nove mber 1979).
OpcrallOna,' costs melu,de ,labor. fucl , power, e lec tri cal and co nt rol, virg in GAe Oxc nford. J. L.. and Jykins B.W.. Jr.. "Confercnce SUlllmary: Pract ical Use of GAC,"
ll1akCll~, Inl sccllaneous md lrect s, and maintenance costs. The operati o na l cos t is usu- J. AIVIVA, 83: 1:58 (J anu ary 199 1).
a ll y CSllIll a te d based o n the GAe ex haustio n rate.
Quinn. J. E.. and Snocyink . V. L.. "Removal of Total Organic Halogen by GAC Adsorbers,"
1. AIVIVA, 72:8: 483 (A ugust 1980).
GAe ex hausti on rate 1000 2500 5000 10,000
(Ib/day)
25,000 50,000 100,000 Robe rts. P. v.. and Summers. R. S.. "Performance ofGAC for Total Organi c Carbon Removal."
J. AlVIVA, 74: 2:1 U (Fcbru ary 1982).
Cost ($lIb)
Suffell, !. H.. Activated Carboll Adsorptioll o/Orgallics/rom rite Aqlleolls Phase, VoIs. I and
O n-site 2.1 1.1 0.9 0.72 0.65 0.6 1 0.58 2. Ann Arbor Science. Ann Arbor, MI. 1980.
rege nerati o n
Suffelt, J. 1.. "An Evaluat ion of Activated Carbon for Drinking Water Treatment: A Nati onal
Off-s ite 2.7 2.65 2.6 2.6 2.6 2.6 2.6 Academy of Sciences Repon:' J. AIVIVA, 72: I :4 1 (January 1980).
rege nerati o n Summers. R. 5., et al., " Bench Scale Eva luati on ofGAC for NOM Control:' J. AIVWA, 87: 8:69
Th rowaway 4.5 4.45 4.43 ( 1995).
4.4 4.4 4.38 4 .35
Symons. J. J.. "A History of the Attempted Federal Regul ation Requ iring GAC Adsorpti on
BIBLIOGRAPHY of Water Treat ment." J. AIVIVA, 76:8:34 (A ugust 1984).
Waer. M. A.. et aI. , "Carbon Regenerati on: Depcndcnce on Timc and Temperat ure." 1. AWWA,
84:3:82 (March 1992).
Adams, J. Q.. and Clark, R. M.. "Cost Estimates for GAC Treatment Systems" j AWIVA Wang. J. Z., et aI. , "B iotiltrati on Performancc: Relati onship to Biomass," 1. AWWA, 87: 12:55
8 1: I:34 (January 1989). . . ,
(December 1995).
AWWA. AlVlV~ Seminar P,vceellillgs- GAC Installations: Conception 10 Operatioll annual Weber W. J.. Jr., and JoccJJah. A. M.. "Removing Humic Substances by Chcmical Treatment
conference III Kansas City, Ameri ca n Water Works Associat ion. Denver. CO. 1987. and Adsorpti on:' J. AIVIVA , 77: I: 132 (Janu ary 1985).
Bab lon. G:.P' . "Developing a Sand-GAC Filter to Achieve High Rate Biological Filtration in Wiesner. M. R. et al.. "Opti mizing Ihc Place ment of GAC Filtration Units," 1. AWWA,
France, J. AIVIVA, 80: 12:47 (December 1988).
79: 12:39 (December 1987).
~algOn Co rp .. Calg oll,,811I1etill,. 20-60, Calgon, Pittsburgh. PA. 1976.
arl so n. M. A.. et a!. . Companng Two GAC for Adsorpt ion and Bios tabi lization" I AWWA
86:3:9 1 (Marc h 1994). . .. ,
7.8 ION EXCHANGE AND MEMBR ANE SEPARATION PROCESSES
Chere~nisinoff. P. N.. and Ellerbusch. E. cd.. Carboll Ads0I1J1ioll Hall dbook Ann Arbor
SCience, Ann Arbor. MI. 1978. .
Crittenden, 1. c.. ct al.. "Design Considerati ons for GAC Treatment of Organi c Chemicals" The ion exc hange and membrane separati on processes are not ordinarily used in mu-
J. AIVIVA, 77: I: 132 (Ja nuary 1987). . ni cipal water treatment prac ti ces. However, membrane filtrati on process has gained
Culp, R. L. . and Clark. R. H.. "Granular Activated Carbon Installations." 1. AWWA 75'S'198 great attent ion because it has the potential to achieve the hi ghest removal rate of NOM
(August 1983). ' ... and DBP prec ursors and complete removal of Giardia cysts and C'yptosporidilll1l
EPA. T~atmellt Teclllliqll~s/or Controlling THM's ill Drinking Wate,; EPA/600I2-S I-156 U.S. oocysts. T hese processes are the products of highly speciali zed chemical eng ineering.
Envlronmcnta l Protection Age ncy, Washington. DC. September 19S I. Consequentl y, water treatment engineers not involved in the detailed design of these
Grease. S. L .. ct al.. "GAC Filter Ad sorbers," 1. AIVIVA, 79: 12:64 (Dcce mber 1987). processes- the design, fabrication, and operational aspects-arc advised to seck the
Hand. D. W.. et al.. "Designing Fixed- Bcd Adsorbers to Remove Mixtures of Organics" assistance of speciali zed equ ipment manufacturers and resin suppliers.
J. AIVIVA , 8 1: I :67 (J anuary 1989). ' . Prov ided th at the projec t has both the budget a nd the tim e, water trea tme nt e n-
Hass ler. 1. W., PlIrification with AClil'ared Carbo" , Chemical Pu blishing. Ncw York. 1974. gineers should run a pilot test o n one o f the processes to establish the design and
Hooper. S. M., el al.. "Improv ing GAC Performance by Optimized Coag ulation" j AWWA operati o nal pa ra me ters (for a spec ifi c w ater supply). Althoug h e ng ineers should un -
88:8: 107 (August 1996). .. ,
derstand the bas ic co ncept and import ant characteristi cs of these processes in order
Hoza lski. R. M.. el al.. "TOC Re mova l in Biologica l Fillers" j AIVIVA 87' 12'40 (D b to write performance specificati ons and possibly P&JD di agrams, it is very impor·
1995). .. '" eccm er
tant that engineers obtain assistance and advi ce from eq uipment manufac turers since
Hyde. R. A.. et al. , "GAC as Sand Replace ment in Rapid Grav it y Filtcrs." 1. AWWA 79' 12. :n they possess a great deal more knowledge and ex perience.
(Dece mber 1987). ' . ...
In this secti on we d isc uss so me fund amental features of ion exc hange processes
Lykins, B. W., et al.. "Granu lar Act ivated Ca rbon for Cont roll ing THMs" 1. AWWA 80'5'85
(May 1988). . , .. a nd li ve me m brane processes: ( I) mi c rofiltrati o n, (2 ) ult rafi ltralio n, (3) n. no filt ra-
ti on, (4) reverse os mosis, a nd (5) e lec trodialys is.
 580 SPECIFIC WATER TREATMENT PROCESSES ION EXCHANGE AND MEMBRANE SEPARATION PROCESSES 581

7.8.1 Ion Exchange A typical anion exchange resin, with a chloride cycle, used to remove nitrate
The process of ion exchange removes undesirable ions from raw water by exchang- ions is
ing them with a number of desirable ions; the desirable ions are stored in a solid ma-
lerial (ion exchanger), Since the ion exchanger has a limited capacily for exchange, NO, + CIR ..... N03 R + CI -
it must be washed with a regeneraling solution to regain its function, This cycle gen-
erally entails backwashing, regeneralion, rinsing, and service, The most primitive ion Ion exchange resins are characterized by selectivity: the .pre.I'er~ntial removal of
exchangers are natural soil and sodium aluminosilicates, The first synlhetic ion ex- certain ions. Seleclivity primarily varies with the type of reSJIl , IOniC st~ength o~ t,he
changer was composed of processed natural greensand (zeolite), However, insoluble solution , relative amounts of different ions, and w~ter temperature. ThIS sele~tlvlty
synthetic plaslic (resin) is used almost exclusively today, Most cOlllmercial resins arc is a very important characteristic in the regeneratIOn ~roc~ss. For exal~ple, III the
sodium cycle the exchanger has a preference for calcIUm 1011 over sodIUm at con-
composed of synthctic plastic materials sllch as copolymers of styrene and divinyl
benzene and are expected to survive 5 10 10 years of conlinuous use, .
centrallons 0 f 1000 mg/L or below. However
" at 100000 ,mg/L. the exchanger prefers

The resins are classified as either cation exchangers- that is, resins with nega- sodium ions over calcium; thus it can be regenerat~d ~Jth bnne. '.
tively charged sites that take up posilively charge ions- or anion exchangers. The The order of ion selectivily (in ordinary water) IS listed below. For callons.
resin has a porous and permeable structure; Ihus Ihe whole resin particle participates
Fe 3 + > A1 3 + > Pb2 + > Ba2 + > Sr2 + > Cd 2 + > Zn 2 + > Cu 2 +
in the exchange process. Resins are typically spherical in shape and range from 20
to 40 mesh (0.8 to 0.4 mm), Cation resins generally have an exchange capacity of > Fe 2 + > Mn 2 + > Ca 2 + > Mg2+ > K + > NH: > H+ > Li +
approximalely 2 milliequivalents per dry gram (mEq/g). The typical strong base an-
ion resins have a 1.3 mEq/g exchange capacity. For anions:
In the tield of water treatment, the ion exchange process is uscd for watcr soft-
ening, the removal of fluoride or nilrate, and the production of demineralized wa- Crol - > sol - > soj - > HPol - > CNS - > CNO- > NO,
ter for laboratory usc. The resinous ion exchange process is the only process, with
the exception of evaporation, that has the potential to remove minerals from water
> NO, > BR - > CI - > CN- > HCO, > HSIO, > OH - > F
completely, thus yielding ultrahigh·purity water.
The ion exchange units can be used to deionize. water simply by allowing the wa,-
ter to flow continllously through a cationic resin column (hydrogen cycle) and (1
Characteristics Whether ion exchange is performed in lixed beds, continuous
strong base anionic resin column (hydroxide cycle).
loops, pulsed beds, or by countercurrent eXlraction, it is still a batch process with
respect to exhaustion and regeneration of the medium. Resins have a discrete ex-
Ca (HC03 h
change capacity and must be regeneraled when exhausted. The cost of the chcmi-
Mg(HC03 h
cals per unit volume of water treated is almost directly proportional to the amount CO 2
CaS04
of minerals removed. H,O
MgS04
The reaction for the removal of waler hardncss is
CaC!,
+ HR ..... H,S04 ..... ROH ..... H,O
MgCI 2
J, HCI J,
ca'+} cal Ca } SiO, . H, O S04
Mg'+ + Na2R ..... Mg R + 2Na + NaCI
SiO, . H,O Mg R R{ CI
Na HSiO,
where R is the exchange resin. A more comprehensive expression of ion exchange,
The imponant issues that must be addressed when designing an ion ~xchange
with a hydrogen cycle (H-form exchange), used in treating a typical water cOlltaill-
process are reduction of chemical COSIS, reduction of waste products, and ~mprove­
ing a variety of iOlls is as follows:
ment of the processed water. However, it is vital that the wa,ter be ,?retleated, by
. of a proper process such as alum flocculation and sedJlllentallon (possIbly
Ca 2 + means
Ca with liltration) prior to ion exchange (reverse osmosis or I'Ime, so ftenlng
' ) t0 reduce
.
Mg2+ Mg the ionic loading to the exchanger system and to remove p~rllculates such as Iron,
Fe 2 + + H, R ..... Fe manganese, and organic substances which may foul the resins. ' _ _ '
2Na + Na2 Tables 7.8.1-1 and 7.8.1·2 can help in selecting the most appropnate type of IOn
2NH,j (NH4 ) 2 exchange process and alternative. The major U.S. suppliers of resll1 are Dow
582 SPECIFIC WAT ER TR EATM ENT PROC ESSES tON EXCHANGE AND MEMBRANE SEPARATION PROCESSES 563

TAIH. . E 7.8.1-1 Basic Types of Demineralizcr and Resins Uscd u TABLE 7.8.1-2 Ion Exchange DiminefHlizalion Unit Seleclion Clwrl
Aill ount of Impurit y Resins
RemOval
to be Rcmovcd, mglL Required Unit s to be Used
Fl ow rat e
gpm FM A co, C WIJ SIJ C DG A Mil

I. Any An y None NOllc x x ( I) x x

(2) x
B, seR B, WBR NOne
x x x x
2. Any Any Any None x ( I)
(2) X X
Any Any X X ( I) X X
3. 0-20 Any
(2) X

0-50 Any X X ( I) X X
4. 20-50 A ny
o S(R 0'
1. By SCR -BrSBR- (2) X
0...,(A 2 . 8,INCR· - 8, S8R ,n M 8 - - 5. Over 100 0- 100 Ove r An y X X ( I) X X X
seR ,n Me (2) X X X
100
Ove r An y X X X ( I) X X X
6. Ove r 100 Ove r
(2) X X X X
"Ih,ed oeeJ 200 100
S(I//I"/;(': Adapted from Th e NlI lt"o Wal e,. H(/I/dvllod. r. N. Kel1llller, cd., McG raw-Hili , New York , 1979 .
8, seA _8yS8R_
Nole.~.· (I) C, cation; A, anion ; DG. degasilier: MH. mi xed bed : WB. weak base : SB. Slrong base: FMA .
free mineral ac id ity (SO.j + C I + NO.1)'
(2) Numbers in parentheses: ( I) Illullibcd plant. (2) mi xed be<' for I f.LS e tnllenl.
(3) Intermediate ranges (50- 100 gr ill now, 50- 100 rnglL CO 2 , 100-200 mglL rM A) R'q uire ca n.>
ful evaluation for best balance in capilal and operatin g costs.
B, weFt 9, SBR

C hcm ical, Ro hm & Haas, a nd Zeo lite. Man y waleI' treatm ent equipm e nt manufac-
turers are al so capable o f producing ion exc hangc process unit s.
Tn'fll! OO!eJ

Br INCR·
By weR By SBR ,n MB
7.8.2 Membrane Separation Process
seR ,nMe
In the li cit! of water Ircalment , in-depth liltralion with g ranular medium is a regular
filtrati o n system capable o f remov in g particles that are approximately 0. 1 J.llll
Two beeJ or above in s ize, provided that chemical coagulation pre treatment is performcd .
• itl\ DC
1. 8y SeR B, saR BrDC'S8Ft In res ponse to the EPA's rece nt emphas is on safe drinking water, eng ineers have be-
Q)SSA or , 8r weR • a, S8R '" M8 Br DC· saR gun to reevaluate all th e best avai lable technol ogy (BAT), includin g the membrane
0·' seR in Me In M8
separation process. Membrane liit rati on, sometimes called a prcssure-dri ven mem-
brane process, has di stinct advantages over the conventionalircatllleni process: (I ) re-
mo val of sllspe nded solids, with no coagulunt , up to about 200 nlll turbidity ; (2) reli -
I By seR e,WBR 8,OC-S8R able productio n of good filtered water; (3) very hi gh " log removal" o f Cianlia - and
0S8A or
ClyptosporidiulII -s ized particles; (4 ) much less space (footprint) required than for the
CD .. e Z . eyWCR' By W8A B,DC'SBA
seR on Me ,n Me con ve ntional treatment process ~ (5 ) easy integration into the autom atic control sys-
tem; (6) minimum labor requirelt1Cnt so that it can maintain unatte nded operati o n l11 os1
of th e timc; (7) chemical-free backwash water lilat can often be di scharged to local
SOli/no: Adapled from TIle Nafm \Vtuer f/tllldb{lok. E N. Kelllmer, cd ., McGraw-Hili, New York, 1979 . water bodies; and (8) lon g-term compliance with drinkin g water regulati ons.
"SCR, strong cation resin : WCR, weak cation resin ; SBR , stron g base resin ; WBR , weak base resin ; DG. The membran e se paratio n process llses se mipermeable me mbran es to separate
{k gas ilier: Mil, mi xed bed .
impurities from water, Scmipcrm eabl e me mbranes are se lec tively permcab le to wa-
ter and ccrtain solutes and play an ex tremel y important rol e in nature: Nutri ent s and
584 SPECIFIC WATER TREATMENT PROCESSES
waste produels arc transferred across the semipermeable membrane of cells.
Similarly, the membrane separation process used in the licld of public water sup-
ply uses a driving force to pass the water through the membrane, leav ing the im-
purities behind as a concentrate. The degree of purilicalion is dictated by the type
of membrane, the type and level of driving force , and the characteristics of the wa-
ter (Table 7.8.2-1). However, the process does have several associated problems: (I)
membrane fouling (fouling by bacteria, chlorine residual , and cationic polymer for
certain types of membranes), (2) rcquires treatment of chemicall y washed waste be-
fore disposal. and (3) need for pretreatment of poor-quality raw waler, which is al-
most identical to that in the conventional treatment process, including ozonation and
activated carbon adsorption for taste and odor control.
Despite the problems, the number of membrane filtration plants has grown dramat-
ically each year and membrane systems arc replacing package lihration plants in some
small water systems. Low-pressure membrane IiItration provides a physical barrier to
turbidity and pathogens; thus regulatory agencies in all U.S. states are now allowing
the installation of this system provided that a pilot study is conducted. TIle economi-
cally applicable size of membrane filtration plants is no longer small and a few 23 mgd
(I m'/s) plants wi ll be on line at the end of the twentieth cent ury in the United States.
TIle development of the submerged microfiltration process (Zenon's ZeeWced and
CMF-S by US Filter) now makes it possible to design plant capacities of 50 to 100 mgd
(2.2 to 4.4 m3/s). The cost of the membrane filtration system has decreased steadily since
1990, and the capital cost of the microfiltration system without appurtenances is ap-
proximately $0.25 per ga llon per day ($65/m' . d) at the end of the twentieth century;
in comparison, the cost of regular high-rate filters are approxirmuely 50% of this figure.
There are several major suppliers of membrane liltration systems: US Filter
(Memcor), Inlilco Degrcmont (Aquasource), Asahi Chemical Ind., Co., Ltd. (Pa ll
Corp), Mitsubishi Rayon Co., Ltd., Leopold/PWT (FYNE System), Hydra-
nautics/NillO Denko Corp., OSl11on ics Corp., and Zenon Environmental Inc ..
Dupont's Pennasep is a well-known RO system. With regard to desalinatio n sys-
TABLE 7.8.2-1 RO Membrane Characteristi cs (Based on 90% Rejection)
Type Membrane
Cellulose
Triacetate Polyamide
Hollow Hollow
Fibers Fibers
Flux at 400 psi. gpd/sq ft 1.5 1.0
Back pressure, psi 75 SO
pH range 4-7.5 4-11
Maximum temperature. OF 86 85
Ch. maximum mglL 1.0 0. 1
Bio-resistance Good Excellent
Backtlushing Ineffecti ve Ineffective
Smm:e: Adapted from TI, e Nako \Vtlfer Halldbook. EN . Kellllllcr. cd .. McGraw- Hill. New York. 1979.
ION EXC HANGE AND MEMBRANE SEPARATION PROCESSES 585
terns, fonics Inc. , Dupont De Nemours Inc ., Culligan International Co., Osmonics
In c. , and US Filter are major suppliers.
One special characteristic of membrane filtration is the great influence of water
temperature on the production capacity (nux). Wintertime [39' F (4' C)] nux can be
about half of summer [n' F (22' C)] condition under the same applied pressure. The
department of health services in each state sets the allowable flux rate for membrane
filters; for instance, the state of California limits the nux to 0.7 gpm/m' ( 159 Llh .
m' ), whereas the state of Wisconsi n has a limit of 0.5 gpm/m' (115 Llh . m'). The
most important aspect of the membrane filters is the integrity of the system; thus the
system should have a reliable method of checking integrity. Table 7.8.2-2 brie ny
compares the characteristics and applications of these processes, and Figure 7.8.2-1
indicates the size ranges of suspended matter in raw water and selected processes.
Microfiltration (MF) Mi crofilter is the most popular type of membrane filter and
the US Filter (Memtec) system has dominated the water treatment field in the United
States during recent years. The purpose of microfiltration is to remove rather large
particles, those over approximately 0.5 fJ.m in diameter. One applica ti on of MF is
in the field of sanitary engineering: Millipore filters are used to isolate coliform bac-
teria from water to collect suspended solids from water samples. Another applica-
tion is in the removal of airborne particulates for air-feed ozone generators.
The MF membranes may be manufactured from a variety of material, including
cellulose acetate. The pore size of these membranes is generall y larger than O. I fJ.m
and a pore size of 0.2 ,..LIn is frequent ly employed for water treatment. Thus removal
of all colloidal matter (generally, 0.01 to I fJ.m in size) and dissolved materials can-
not be expected. Yet the membrane is capable of swelling and changing its charac-
teristics, thereby al lowing it to retain certai n types of colloidal and solu ble substances.
Whenever the MF membrane becomes clogged, the filtration rate drops below a
practical level, thereby necessitating back washing and chemical cleaning. A self-
cleaning cartridge filter with 500-fJ.m openings is ge nerally installed upstream of
membrane filters. Consequently, the water should be pretreated and have most of
its suspended solids removed prior to the membrane filtration process. Backwashing
of the membranes is normall y done by air-scouring wash every 20 min for a minute
or so. Moreover, chemical cleaning of the membrane should be performed every 2
to 3 months, depending on the raw water quality, by caustic soda or acid with de-
Acetate tergent when the pressure exceeds 15 psi (I kg/COl').
Spiral As discussed previously, the new larger capacity of MF modules and the more
Wound reasonable cost of the system enables engineers to design 100-mgd (4.4-m'/s) MF
15- 18 systems beyond the year 2000. Figure 7.8.2-2 is an artist's concept of a 30-mgd
0 ( 1.3-m'/s) MF system with six cells of a submerged type MF filter system .
4-6.5
85 Ultrafiltration (UF) Ult rafiltration uses membrancs with pore sizes that arc sig-
1.0 nificantly smaller than 0.1 J.lm and generall y require a driving pressure ranging from
Fair 10 to 90 psi (69 to 620 kPa or 0.7 to 6 kg/cm '). For water treatment application, a
Effective " loose" membrane is used so that 10 to 40 psi (0.7 to 2.8 kg/cm') pressure is re-
quired cven though thci r pore size is 0.0 1 to 0.02 fJ.m. UF is capable of removing
 '"ena>
TABLE 7.8.2-2 Comparison Between Membrance Separation and Ion Exchange Processes
Molecular
Pore Size Weight
Suitable or Cutoff Driving Removal
Process Water Resin Size (da ltons) Force Objects Notes
Microfiltration 500-~m 0.1-0.2 ~m . 300.000 10--20 psig Particu lates and Batch process, removal
(M F) self-cleaning 0.2 ~ m is (0.7-1.4 kg/cm') mi crob ial of particles over 0.5 j..Lm
cartridge filter common
Ultrafiltration 200 to 500- ~m 0.003-0.0 I ~m. 50.000 10-40 psig Molecular size Batch process. liquid- solid
(UF) self-cleaning 0.0 1 /-Lm is (0.7-7.8 kg/cm ') compounds. separation
cartridge fi lter common particulates and
microbial
Nanofiltration Regular filter 0.00 1-0.005 ~m 200-400 75- 150 psig NOM. including Batch process. DBPs
(NF) effl uent or MF (5.3-10.6 kg/cm') color. virus. control and soften i ng
filtrate Ca. Mg
Reverse osmosis Filtered water. < I nm >200 psig Ionized salt ions Continuous process. 90-95%
(RO) 100-36.000 (> 14 kg/cm') and collo idal inorganic salts and 95-99%
mglL salts matter organic matter
Electrodialysis Fi ltered water. < I nm Dc electric current. Ioni zed salt ion s Continuous process. complete
(ED) 500-8000 0.27-0.36 kWllb. removal of salts is not
mglL salts salt s (0.6-0.8 possi bl e
kW/kg)
Ion exchange Settled or filtered < I nm <7 psig Ionized ions. Batch process. complete
(IX) wate r. 50-1.000 «0.5 kg/cm') cationic and/or removal of sallS is poss ible
mglL salts anionic

I
I

3 3 g
::r -l"F
" ::r 3
n " ~ " " -
C3 30" 0_ .
~""
z..., .."
o
-.
g~
" ~
< .... ~
~m
~

~;:<
7
~
.~ ~

N
!"
Q
n :::J C-'
0 ::::I
"T1C"
3 ;
~
""
0-
~
"
(";l

... ~
(";l ....:.
~~

a~ ~zi:!~oiS!; i;;
~
n
...
.." ~
~ 8 ~ 3
~ 0.. (!j ~
~ "
: g
-
0"
g~ Oc
.., ........ N
-"' ,
~::l
~o
In ~O~rn~
>c:
;;:I
c;.... <
m
~
0
2
"~
...,
:: "
o· 0 "0 "~
(; (!j
"n ~ - -On
00",
2 ~

"8" " n "

_.
~
:;B 5'en
o - a "~
~ 0 ~
00 "
°~ a- ;; N' S
i;-> ~::r-
¥ '< Q. " ; -~
~::: > ;;

'"'
""0
o
(!j
:'l en
- ;-- "'5"' ~
§.
in ~
(!j
0"
~ (":l
n;:0::'" "'"
~
C
<
- .
~
~
-g
~

~. 3 ~
~
0
Z
~
•
~
"'~ '"
~~- . . ..:.••
B
~ - r~
>

Q
~~
(":l (";) z
~ N -
:: o· 0 c::: en :=. ~
~ ::::I til ~

_ . '-
~ 0 0
~
r p- i
§ is
g
""S
o
_ . ::::1
o
.? ~
< ::..
n - ~ "< ~
~ '<
~ n
Q.
•
en
0::
~
::::I
Q.
o ::: -,
? :::: ....
():; c
o - ~ ~
"il
~ ," ~• e;..
<
~
e
;:
z
m
1j
'"~. ,.~ "~
!)i
,,?S5-3° n ::r ,
••>
~ 0
.~ en n _ . .g 00'
"~
I
>-
o
3
c_
=.
tll
til n""O
;:0::'" 0 =. ::r-
<§. ::::.
c % Z
(!j Vl
0"'
- 3 - r.
Co
~ .- ~ i "
0
'm>-"
'a. ::::I
00
_. nl
til ..,
~O-
~ 0
(";l
"
oCS- 0 © § € ~
>
Q .,. ~0 z
s:." s;.
::::I ~ r.
!!.S :;~~
~ ~ _.
iL- a Ii<
~
.'
"I
o
;;:
~
~ n 0 -
~
'< 0" u.~ ""°_ . "~< . , " I. m
;;:
~
;; n 0.. 0 0 (JO ;..
»-~
p g CD

f-- ~ f-- f--

> :n
~~1=~
>
~~g ~- >-
? " "
(JO - .
,
~
~
z
no
~
~
-,..., ::r
n -- t;l
..
5 ' en ::::I ~
3
~
(";l
•
J:;<
o
(JO

0
.., :0:- n
fr" ,.~ ~-
:E t:: 5-
r- ' -
~

~
;:
n
"~ •
~s
~:
m
U>
m
~Oci ~
nCo.. ::::I ~ n ;:
o. 1Ii~
;l
~ ~- TF-
'<
~ @ til
~"g 3~ < "
~"
:,)
;; 5' Q
(")

~
<
~ > • :n
'<
2,
C
en ::::I 0 -
n o..dB
-. (!j -
..,

3
"
n'
n
..,
~ "
~ ~ -g
:::. ~
~ ~ " 5
...> ,
f-;-
n
is
z
;n
::::I
.., - Q ~
0.. ""0 ::l
o 3
:!! ° *~5.
»
S ,- •;; 0
."
~
:!l
.Q . ::l
" " "
a "< _"0
~.
£. 5 '
(I'Q i • ()

~
Q.:::tIlS-
(!j n til - .
'<
"0
0
" g ~ 3 • "g m
~
•~
0.. .., t: ::::I
2- "o "0-
~ ~.
=.
!::
~
> m
_ . ~

- 0"""'" ~ % ~

en Z:::$ 0 > U>

::::I
~
::::I
~ ~ o 0 a § %

-S
";" "<
;: ~ 0
r- VI
c::l
.,,~
? Z
~
0
" '"a>....
 ION EXCHANGE AND ME MBRANE SEPARATION PROCESSES 589
588 SPEC IFIC WATER TREATME NT PROCESSES

TAIlLE 7.8.2-3 T)'ll iclil Pc r forma ncc of Mcmbranc Fi lt ration Processes (Qlmlity of
riti es. T hi s increase in the densi ty and viscos it y of the solutio n at the surface of th e
me mbrane ca uses a rcducti o n in th e Ilu x ( Il ow) relte: The water producti o n rate is PCfmhttes)
Microfi lter Ult rn filter Na no liltcr
approx imately 50 ga l/day per sq uare foo t (2 Ill J/day . 111 2) of me mbrane surface area
(Me mcm) (Aquasourcc) (Os moni es)
and th e e nergy req uirement ranges fro m 0.4 to 2 Wh/ga l (0.1 to 0. 5 kWh/ m3) of the Parameter Raw Water
prod uct wate r, de pendin g o n the dead-e nd and c ross-fl ow nitratio n. 0.Q5 0.06 0.0 1
Turbid ity (nlll) U
UF mc mbranes may be fl at with a niter prcss type of s uppo rt , n at and s pi rall y 6 6 6
Te lllpenH un: (OC) 6
wo und , tu bul ar with a supporting cylinde r (i nside), o r a tubul ar (holl ow) tiber w ith 40 57 9
Partic les 4500
a n ex tern al di amCler of 0.5 to 1.5 mill th at does no t have a supporting member (the (> 3 JJ.mll OO mL)
2.2 2.2 t.2
raw watc r Il ows inside the tubes). T his las t conli g urati o n requires pre lilte ring throu gh TOe (mg/L) 2.6
0.07 1 0.073 0.0 17
Illembra ne filters with a pore size of 10 to 20 J.1.111 to preve nt th e membran es from UV 2:".I
0.087
0.093 0.011 0.0 12 0.0 11
beco ming qui ckly clogged. The UF process has been used in the med ical fi eld for Iron (mg/L)
0.032 0.033 0.036
Manganese (mg/L) 0.064
tilt.! se parati o n of blood plas ma and also has a number of indu stria l applicatio ns: 570 64
tlPe (cfu lmL) 54
co nce nt rati on of macromolec ul ar suspensions (i.e., polyv in yl alco hol), recovery of
cc rt ain substan ces fro m was tes, and co nce ntrati o n of e mulsions. Tti Ms 0.2
0.3 0.2 0.2
Chl orine resid ual
(mglL)
Nanofiltration (NFJ The po re size o f thi s me mbra ne is between th ose of UF a nd 66 61 60 14
C hl oro form (~g/L)
RO. Nanoliltrati o n me mbra nes fo r pota blc water appl icati o ns have a pore size of U 1.3 2.5
Dibromoc hl oromcl hallc U
0 .00 1 to 0.002 IJ.m and a molec ular we ig ht c utoff (MWCO) of approx imate ly 200 (JJ.glL)
da lt o ns. Na noliltratio n is capab le of re movi ng DBP prec ursors as well as calcium II II 5.5
Bro mod ichl oro mcth ane II
and mag nes ium io ns, and water softe nin g ca n be ac hi eved more cost-effecti vely than (JJ.gIL)
79 74 72 22
wi th the lime- soda as h softe nin g process. Perm eate recovery is 80 to 90%, a nd 50 Towl trihalo lllethanes
to 70% rc moval o f lotal disso lved solids ca n be ac hieved. Sin ce TOe removal is (JJ.g/L)
e lTcc ti ve ly ac hicved by the NF process, cnhanced coagul atio n is not necessary to HAA, 0.2
0.3 0.2 0.2
co nt ro l O BPs. Also, mi cro bial removal is 100%. There fore, NF is capable of re- Chlo rine residual (mgl L)
3 3 3
placi ng ozonat io n, GAC adsorptio n, and enhanced coagulation . However, the NF Bro tlloc hl oroaceli c ac id
mc mbra ne has a biofo ulin g problc m; thus e ffec ti vc pretreatme nt such as a co nvc n- (lJ.glL)
2 2 2 0
Brol1l odich loroacelic
ti o nal treatm en t process and mi cro n cartridge lilters or eve n micro nltrati o n pri or to
acid (JJ.gIL)
NF, ca n be very hc lpful. Table 7.8.2-3 shows typical perform ance of the membrane 0 0 0
Ch lorod ibro ll1oncet ic 0
lilt ra ti on proccss. ,cid (JJ.gI L)
25 25 4
Dichl oroaccti c ac id 25
Reverse Osmosis (ROJ When pure water !lows th roug h a semipermeable mem- (JJ.glL)
21 18 18 2
brane (se parat in g the two waters) into wa ter with a hi gh salt content , the water level Tric hl oroace tic ac id
in the salted water co mpartment rises unti l the os motic press ures o n e ither side of (JJ.glL)
7
46 43 43
the mc mbra ne arc eq ua li zcd : This is os mos is. In reverse os mosis, a n osmo tic pres- Total tl AA,
sure g reatc r th an that o f the salt ed water is applied to the co mpa rtme nt co nta ining
the salted water. This the n forces pu re water to now bac k ac ross the membra ne. T he
surc and the size of thc s ubstances that ca n be re moved . Na noliltratio n and ultra-
pressure rcqu ired depe nd s o n the diffe rence in salt co nce ntratio n but is often greater
nitrati o n are essent iall y incapable of rc mov in g solut es o f low to intermcd iate 1.11 0 1-
than }OO psig (20 70 kPa 01' 2 1 kg/c m'). The rule of thumb is to appl y approx imate ly
ec ular weight. Reverse os mos is ca n re move io ni c spccies with low molec ul ar we ig ht ,
10 psig (70 kPa) fo r eac h 1000 mg/L diffe rence in T DS.
The RO process may be used in cases where the wa ter has a salinity (TDS) of as we ll as all typcs of coll oids, bac tcria, and viruses. Thc ratio bc t wee ~ th e ~Jll ount
of demine rali zed wate r and the amo unt of raw water fed to the syste m IS delmcd as
less th an 500 mg/L to prod uce a 25-mglL T DS prod uct wa te r using a small a mo unt
of energy (co mpared to the EO process). RO is also capable o f de mi neralizin g sea- thc conversio n fac tor and is ex pressed as a perce ntage.
Two types of RO membranes arc currently bein g marketcd: cellll l os~ acctnte and aro-
water (}6,OOO mg/L TDS) to wa ter cont a ini ng approx imately }OO mg/L T DS (Table
mat ic polyami de. Membrancs composed of cellulose acela~e yield a hi gh now rate per
7.8.2-4). The RO proccss is occasionall y called hypelfi ltration si nce a hi gh press ure
unit of surface arca. T hey arc available in tube shape, as spirall y wound !lat sheets, and
is rcqu ircd to d rivc the sys te m. The major d iffe re nccs be twecn RO a nd Il anofilt ra-
ti o n, whic h also appl ies a fai rl y hi gh press urc, are the mag nitude o rthc applied pres- as holl ow libel'S. Conversely, polyamidc membranes have a low ratc of flow and arc
 ION EXCHANGE AND MEMBRANE SEPARATION PROCESSES 591
590 SPECIFIC WATER TREATMENT PROCESSES

TABLE 7.8.2·4 1Yllic~, 1 Performa nce of RO Process ineralization are lower for RO than for ED and io n exchange; it is also un iquely
characterized by a constant cost production factor.
(a) Dem inera lization (Softcning) by RO Process
Electrodialysis (ED) The se mipermeable dialysis membrane is impermeable to
Con stitll cJ1( (mglL) water but permeable to either cations or anions under different driv ing forces, The
Raw Finishcd
drivi ng force may be a concentration gradient (simple dialysis), pressure (piezo-
dialysis), or an electrical potential (electrodialysis). The most commonly used mem-
Hardness. as CaCOJ 380 20 branes are composed of hydrated cellophane, and selectivity of a give n membrane
Alka linity. as CaCOJ 2 15 16
is largely a function of pore size. Although dialysis is rarely used in the field of wa-
Tota l e lectrolytc. as CaCOJ 445 29
ter treatment and water reclamation , a few industries, such as the texti le industry,
Silica. as Si0 2 25 3
pH 7.2 6.0
use this process to reclaim caustic soda from merceri zing baths,
CO, 25 25 TIle clectrical potential is the driv ing force in the ED process. Membranes having
anion and calion exchange properties are stacked alternately in a press with a narrow
water passage between them , The channels are made as narrow as possible to minimize
electrical consumption. When direct CUITent is applied to the electrodes (located on ei-
ther side of the stacked membrane), the anions migrate to the anode, and vice versa.
Since the calion exchange membrane only permits cations to pass through, and only an-
Feedwatcr Purified Wate r
(mg/L) (mglL)
ions for the anion exchange membrane, alternating concentration and dilut ions are cre-
ated in the various (alternating) compm1me11ts of the membrane stack (Figure 7.8.2-3).
Ca2+ 329 1.5 0.5
Mg2+
Na+
K+
1,03 1
9,419
355
3
77
3
0.3
0.8
0.8
t
.t""od.
"n,ln, ",Id
Sr2+
S0,, -
13
2,200
0.2
0.05
concentrated
water
I demineralized
po1-
HCO,
CI -
9
68
0
26 38
~ 1+ r I I
~ + l - l +l-~
water

F-
15,825
0
120
0
38
I
I I
I
i I
I
:I I
'
NO,
Si02
J.4
0
0
0
I
I
I eU i ei-J: ~
Total salinity 32,680 185 0.6 I
I ! i : i
Conductivity (,...S/cm)
pH
37.250
6.3
390
6.3
anode
+ CI,
I
I
I
I
i i'
, : I
I
:
.
I
cathode

Source: (a) Adapted from The Nalco H'tller Ham/book, EN , Kemmer. cd., McG raw -Hili, New York,
1979. (b), Adapted fro m Dcgrc mon l, \Vtlll'r Trelllmeill Hwulbook, 5t h ed ., Hal sted Press, New York, 1979. Li@ lW@
. I
iU®
. I
i.
I
I
! : !I : !I
"r
marketed as hollow libel' to provide the maximum surface area per unit volume. TIle _,node: I I cathod.

t I +I t
nnse I
tit
~dvantage . of t l~c polyamide membrane is its excellent resistance to chemical degrada-
tion and bl ~loglca l attacks, thereby providing it wit h a longer service life (3 to 5 years) w.'"
to be concentrated t I I water
than that 01 cellulose acetate membranes (2 to :\ years); refcr to Table 7.8.2-1, However,
the p~l~a.mide. m~teria l in RO membranes becomes in'eversibly damaged when exposed
• 10 be demineralind

to oXldlzmg blocldes and most cationic biocides used in water treatment. electrode
rlnslng fluid
Since RO is a cOI1~i n uous process, the membranes must be replaced every 3 to
5 years. Howevcr, unlike ED, there are a number of membrane and RO process unit
manufacturers to choose from . Moreover, the operational costs assoc iated with dem- Figure 7.8.2-3 Concept of electrod ialysis.
 ION EXCHANGE AND MEMBRANE SEPARAT ION PROCESSES 593
592 SPEC IFIC WATE R TREATMENT PROCESSES

Figure 7.8.2-4 is an example of the de mi neralizati on c~st s assoc~ated with ~o n

Unlike ion exchange, ED is a continuolls process where thc amount of Sellt rcmoved
exc hange, elec trodi alysis, and the process of reverse os mOS IS, excludlllg the capital
from the watcr is proporti onal to the applicd e lectrical cun·en!. However, application
amorti zati on cost.
of an excessive amount of current results in electrolysis, formin g hydrogen ancl oxy-
gen gas, and thereby red uces overall e rticiency. ll1C erticiency of ED may also be re-
duced by concentrati on polari zation and by the formation of scale and deposits on the BIBLIOGRAPHY
membrane surface; these phenomc na increase the electrical cu rre nt requ iremcn t. For
exa mple, calcium carbonate and magnesiulll hyd rox ide scales will develop at the <lIl - Adham, S. S.• el al.. '·C haraclerislics and COSIS of MF and UF Planls," J. AIVIVA, 88:5:22
ia n exchange membrane, duc to the hi gh pH res ult i ng from the hydroxide ions. Although (May (996).
improveme nts in the stack design. as well as in the me mbnmcs, have helpcd to reduce Adham. S. S., Jacangclo, J. G., and Laine, J. M., " Low Pressure Membranes: Assessi ng
both polari zation an d scale formati on, they have not eliminatcd these problems. Inlegrily," J. AIVIVA, 87:3:62 (March 1995).
If the raw water cont ains noni onized molec ul es sllch as orga ni c co mpound s a nd Applegate, L. E., "Post Treatment or Reverse Os mosis Prod uct Waters," 1. AWWA, 78: 5 :59
colloids, they re main part iall y de mi nera li zed in thc wate r. It is therefo re necessnry (May (986).
to pre treat the raw wa ter to remove turbidit y, orga nic co mpound s. and iron, prior to Bergman. R. A.. "COSI of Membrane Softening in Florida." J. AIVIVA. 88:5:32 (May 1996).
bei ng applied to the ED process; otherwise, they will fo ul thc me mbranes. T he powe r Bersillon, J. L., et al.. "Ultrafiltration Applied to Drinking Water Treat ment: Case or a Small
co nsumpti on o f the ED process is approx imate ly 0.27 to 0.35 kWh per pound o r Systcm," Proceedil/gs of the AWWA Am I/wI COl/ventiol/, Los Ange les, 1989.
salt removed (0.6 100.8 kWh/ kg). Chellam, S., et al.. "Effect or Pretreatment on Surface Water Nanofiltration." 1. AlVlVA.
The ED process has three major limitati ons: ( I) it is incapable o f prod uc ing hi ghly 89: 10:77 (October (997).
de minerali zed water; (2) the cost of the process increases rap idl y with th e salinity of Chell am, S., et al.. "Estimating Cost fo r Int egrated Membrane Systems." 1. AWWA , 90: 11 :81
the feed wate r, with the maxi mum salinity limit for economical treatmen t being 8000 (November 1998).
mgIL; and (3) the num ber of equ ipment manufac turers is ve ry limited : Ion ix, Inc., Cicek, N., et al.. "Rccycling Munici pal Wastcwater Using a Me mbra ne Biorcactor." 1. AlVWA.
Aqua-C hem Inc. , and Dow C he mical USA are the on ly major suppliers in the United 90: I I : I05 (November 1998).
States. Figure 7.8.2-3 prese nts the basic principle of the elec trodi alys is sys tem. Clifford. D.. and Liu. X.. "Ion Exc hange for Nilrate Removal," J. AIVIVA. 85:4: 135 (April
1993).
Degrcmont, Water rreatmem flalU/book. 5th ed., Halstcd Prcss. New York. 1979.
3.0 Escobar, I. c., and Randall , A. A.. "Influencc or NF on Distribution System Biostability:'
: / f J. A\VIVA , 91:6:76 (June (99 1).
2.7 : #Two- bed weak base
Two:-bed strong ba se :. , ion exchange Fai r. G. M.. ct aI., Water SlIpply alld Wastewater Disposal, Wiley. New York. 1954.
2.4
Ion exchange --H I
Fu . P.. ct al.. "Sclccting Membrancs for Rcmovi ng NOM and DBP Prccursors," 1. AWlVA,
:. f
;;; 86: 12:55 (December 1994).
" 2. 1 !
: .
I I
I
Gere. A. R.. "Mi crofiltralion Operating COSIS." J. AIVIVA, 89: 10:40 (Oclober (997).
~ 1.8 ,: I I Haas, R. D., "Microfilt ration Bcats Conventi onal Treatmcnt," 1. AWlVA, 89: 10:10 (October
.
~
0
u
1. 5
: . I
,: / I
Mi xed-bed ... ' I
Electrod ialYSi j

J
1997).
acange al "Assessing Hollow-Fiber Ultrafi ltrati on for Palt icul atc Removal."
10 , J. G ., el ,.,
;;;
c
.9 1.2
ion exchange.:' .III
Three-stage RO
/ J. AIVIVA, 8 1:11:68 (November 1989).

~
....
.' , I
/ , Jacangclo, J. G., ct a!.. "Low Pressurc Membranc Filt ration ror I~c mov ing Giardia and
0 .9 / Mi crobial 1ndicalors," J. AIVIVA. 83:9:97 (Seplember (99 1).
~
0-
0 ... /
..../ . , / ' Two-stage RO / "Mechanism or CI)'IJtosIJOridilllll, Giardia, and MS2 Vi rus Removal
0.6 7 J acangc I0, J. G" el al ..

0 .3
.....:/•' ;"•'" > One- stage RO by MF and UF." J. AIVIVA. 87:9: 107 (Seplember 1995) .
Jacangelo, J. G.,ct aJ.. ''Thc Membranc Treatment ," Civil Eflg., pp. 40-45 (Septcmber 1 998~:

0
~... .- Jurca n.d V. L., el '., , ,ation Pret rcatmcnt for Ultrafi lt ration of a Surracc Water.
al "Coagul
10 50 100 500 1000 5000 10,000 J. AIVIVA. 82: 12:76 (December 1990).
Total di ssolved solids - TOS (mg/L) Karimi. A. A.. cl al.. " Microliltral ion Goes Holl ywood," J. AlVlVA. 91:6:90 (Ju ne 1999).
Kemmcr, F. N.. cd., Nalco Chcmical Co .. r ite Nalco Wa ter Ha ndbook. McG raw- HilI. New
Figure 7.8.2-4 Rough comparison or operational costs between RO and io n cxcha nge
processes. York. 1979.
 INOAGANICS REMOVAL PROCESSES 595
x '0
c
~
, '" " .~

""
~
Lauch, R. P.. and GUIer. G. A .. "\ on Exchan ge for RClllovn \ of Nitrate from WeJl Water,"
"c
~ U
"
.~ · 0
O~
'"
::;-
OJ,
E
'"
.Q

""'"~
~

C
-"
.Q
'0
C
" .~0 J. AWIVA. 78:5:52 (May 1986).
u" tE rf C " ~

c'"
>-
"~ c Marlin. H. J .• and Mi lic I', G. R.. "A Zero Di scharge Stea m Electric power Generating Stat ion ,"
'"'"
0 ~
v-. ~ 0 .2 " E
'"
'3 '0
c
.c " ".Q '0 L "E
'0
c U
::2 '0 1. AWWA . 75:5:52 (May 1986).
'"'"u '0
8 " v
0
.~
c "0> c" '" 0
U "
ct s ~ "
;;
E
E
~

'0
c
Metcalf. P. R.. and Schroeder. E. D.. "Dc nlritication Using a Mcmbrnnl..:- lmll1obi lizcd
u
'E
~
~
0 E
~ " .Q

r-:, "a
E a:
~

.e!
'0
c "
.~
u
""
~

.Q "
.Q
'0
'3
0
.c .2
"
C
Biolilm." 1. AWIVA. 87:3:77 (March 1995).
Montgomery. J. M.. Consulting En gineers. Walei' Trefl{IIIl'II': Principles alld Desigll, Wiley.
u<1:'" ." g '80
~
gfl ~1) "8 ~
;;
'0
c "> c <- C o 0 ';:: c '3 New York. 1985.
'" u 0
u
'"
u
.c
u Q " ~ OIJ -3 '" 2 en Mulford. L. A.. cl al.. "N F Performance at Full and Pilot Scale," 1. A\vWA. 9 1:6:64 (June
E ~v E +
:I:
.s "c0 .c .c ~ ell
C
V
>
g '" 0
u
~

:;;:
~

:;;: ~" '0 '0 '" ug li: 199 1).

u co CO
'" Point ius. E . "Regu latory Compliance Using Mcmbnlllc Processes," .I. AWWA . 88:5: 12 (May
1996)
Pointills. F.. cl al.. "Regulati on Gove rning Ml.!mbranc Concentrate Disposal." J. AWWA ,
88:5:44 (May 1996).
Rei nh ard. M. R .. ct al .. " Remov in g Trace Organics by RO Using Cellul ose Acetate and
I I I I Polya mide Membranes." .I. AIVIVA. 78:4: 163 (Apri l 1986).
Sinsigalli. P. D.. and McN utt . J. L . "Industrial Usc of Reverse Osmosis." 1. AWWA , 78:5:47
'80 (May 1986).
Snocy ink. Y. L.. ct a!.. "Strong-Ac id Ion Exchange for Remov ing Barium. Radiulll and
~
v~ '"
.e! .~ Hardness." 1. AWIVA. 79:8:66 (Au gust 1987).
~ u
~
U
c
e '"cE ~ Snyder. D. W.. el al.. "Weak-Acid Ion Ex change for Remov ing Bmiulll. Rad ium. and
c'" 0..
v
Hardness." J. AIVIVA. 78:9:98 (September 1996).
'ss ~
.~
'80
C
v
"
'0
~
~ Wiesner. M. R., et al.. "Cost Estimates for Membrane Filt ration and Conventi onal Treatment,"

u
0
c E
;;
~
f0-
!l
'"
.~
u
e"
c.
'"
.e!
X
J. AIVIVA, 86: 12:33 (Dece mber 1994).
.~ c
e v
'" .;;; c

-
~
0
.5
.~
u
:~
UJ
"
.Q

~ X
~ .2
;;
u 7.9 INORGANICS REMOVAL PROCESSES

'0 " In 1985 the EPA released its recommendati o ns on the Ina ximum co ntainme nt level
:0
'" C."
>
0 .~
c
"1
~
M "1
~

"1 (MeL) for 10 inorgan ic contalninanls: arsenic, barium, cadmium , chromium, Ilu o-
E
u c. o c ride, lead, me rcury, nitrate, selenium . and sil ver. This sectio n brielly cove rs the ap-
0: <1: .~ .g plicable and teSled processes lIsed to remove these co ntaminants to levels below the
u '3 c:
'"

-
-5
.£
~
u
~
~)
uo E
"
-:.0
'" v
.~ U'l
MCL.
In 1978, the EPA published a manual that prese nted the recomm ended treatme nt
techniques for meetin g the inte rim primary drinkin g water standards; these arc li sted
~
u E '0 in Tab le 7,9- 1. Si nce the EPA did no t present effective methods for rem ov in g as-
V C
2
0..
.c '"
U bes tos nbers, it has been added to the table .
..,>
u
u
~
BIBLIOGRAPHY
'" +
' ,-
<1:
+
<"
U
+
~
v ~

~
C '0
c E c
'0 I
u. '" :I:'" ,z
:I: '"c
'0
V
.Q

'"
EPA. M W/If(I / of Dri/l king \-ValeI' Trcallllellf Techlliqlles for Meeting fhe Interim Primary
"E Drinkillg Water Regulatioll , EPA-600J8-77 -00S. U.S. EPA, Cincinnati. OH . 1978.
'"'"
...l
C

g
c
C e,-
'"
.::! +
~ <1:
E
~ +
E ;!; 'E
•~ '" f'
b
.- +'"
~

E e.
o "
.1j.Q

'§ ~
u
'0
'"0
~,
c
U
.-

g'J) ~
_V E
~ +

c " ">
~'"
v
.
.- " <1:'"
~

~v

~
0
~ CO -gU loU
u::~
~ '" .= " .- .Q
~

U <1:
'" U U OZ
'"" Vi <1:

594

CHAPTER 8
Management of Procurement and
Construction Phases
The des ig n phase of a project is completed o nce the linal pl ans and spec ifi cati o ns
of the projec t are prese nted to the ow ner and the requi red reg ulatory age nc ies. As
pari o f the design phase, the design eng in eer sho uld compil e a com pIe Ie sc i of pro-
ject fil es, including all des ig n calc ul ati o ns and correspondence, s in ce these doc LI -
ments may prove to be esse nti al for the project in the latter stages and parti cularl y
in the case of futu re lawsuits.
Upon co mplcling the design phase. the e ng in eer mu st begin ma nag in g the pro-
c urement and construction phases. The sequ ence of events is (I) bid advcrlisillg, (2)
bid opening. (3) bid eva lu at ion and award of contract, and (4) the co nstru ctio n phasc.
There arc generall y three principal parties involved in the cons tructi on phase: the
ow ner, the des igner, and the bidder.
8.1 PROCUREMENT PHASE (BID PHASE)
As s tated previously, the proc ure ment phase is composed of bid adve rti s ing, bid
opening, and bid eva luati o n and award of contract.
8.1.1 Bid Advertising
During bid advertising th e project engi neer and the reside nt e ng in eer (inspcc tor)
s ho uld review the co ntrac t document s together to ensure that the co ns truc tabi lity of
the structures and important fi eld consideration have bee n incorpora ted int o the plant
des ig n and spec ifi ca ti ons. Since the period between bid ad verti s ing and bid open-
ing is generally a minimum o f 6 to 8 weeks (u nless the project is ve ry s l11 all)- to
allow the prospecti ve co ntrac to rs to perform the necessary activities- the e ng ineers
have tim e to review th e documents carefull y. If any omi ssio ns or errors are found
during thi s period, the project engineer sho uld issue an addendum to the plans or
spec ificati ons to all bidders a minimum of 10 days befo re bid opening. The project
ma nager sho uld also look ahead and mee t with the resident engi nee r to develop a n
agenda and a list o f key s ubjects that mu st be di sc ussed with the co ntrac to r at the
precollstrllction mec ting, which is he ld with the sllccessful bidder.
596
PROCUREMENT PHASE (810 PHASE) 597
During the bidding phase, the projec t and resident eng inee rs should perform the
items listed below.
I. Arra nge for bid ad verti s ing and se lect a bid ope ning date. Log the issuance
of bid documents and release the project data to constructio n trade papers.
2. Keep an accurate log of all sets o f co ntrac t documents iss ued : for example,
name o f the fi rm, address of eac h bidder, name o f the person acquiring the
docum ent s for the firm , date of pickup, and deposit fee rece ived .
:1. If addenda are issued to the bidding documents, a copy of the addend a should
be ma iled to each ho lder of a set of documen ts by cert ified mail w ith a return
receipt requested.
4. If the demand for the document s exceeds the suppl y prepared, exact repro-
ducti o ns o f th e ori g inal documents mu st be supplied to the requesting firm s.
For obvious reasons, it is vita l that all bidders receive identica l sets of bidding
documents and that the same inform ati on be provided to all bidders whenever they
req uest clarification by phon e. e- mail , or mail. To protect the owner against finan -
cial loss, whi ch may occ ur because th e bidder later refu ses or is unable to sig n th e
contrac t a fter submittin g a bid, the requirement for a bid bond (security) should be
in c luded in the bidding docume nts. The bid bond should be 5 to 10% of the co n-
trac tor' s bid price and may be in the form of a certifi ed or cashier's check and, oc-
cas ionall y, a negoti able secu rity. The bid securities are returned to the bidders after
all bids are ope ned and the co ntract is awarded.
All public works con tracts req uire the contractor selec ted to s ubmit performance
a nd payme nt (labor and material s) bonds o nce the contract is awarded. The perfor-
mance bo nd ac ts as a surety to cover cos ts assoc iated with completion o f th e con-
tract should the co ntrac tor fai l to execute the co ntractu al requi rement s. Under the
payment bond, if Ihe co ntractor becomes insolvent , the surety guara ntees payme nt
of all legi tim ate bills (labor or the acqu is iti on o f materia l) resulting from perfor-
mance of the con tract.
8.1.2 Bid Opening
T he bid-opening procedure generall y proceeds as follows:
I. Receipt o f sca led bids at the designated time and pl ace
2. Conlirmation that all bids are responsive
3. Acceptan ce and logg in g o f each bidder's name and the bid amount of all re-
spo nsive bids
4. SUlllmary of a ll line- item prices for all unit-price bids
After all the bids have been o pened, the non responsive or informal bids arc re-
jected. The project man age r the n identifies the apparent low bidder, w hi ch a fter care-
ful bid rev iew and evalu atio n is s ubj ec t to confirmation.
 598 MANAG EMENT OF PROCUREMENT AND CONSTRUCT ION PHASES
8.1.3 Bid Evaluation and Award of Contract
Following the bid ope nin g session, all bids are tabulated and eac h bidder is checked
for lin ancial responsibility, lice nsing, integrity, and reputation. T he remni ning tasks
for (he project and reside nt e ngineers are:
I. Analyze the bids and compare thcm with the enginee rs' es timate.
2. Make recolllme nd ati ons with regard to awardin g the co nt rac t.
3. Pre pare noti licati on of award o f cont rac t and procure bid bond s fro m the three
lowest bidders.
4. Remove copies of payment and perform ance bonds and certili ca te of insur-
ance from the file prepared on the selec ted co ntractor and have the contractor
submit the pay me nt and perform ance bond with a n ac tu al insura nce policy.
5. Have the own er sign th e contrac t and return all other bid bonds.
8.2 CONSTRUCTION PHASE
There are several types of co ntrac t for a co nstruct ion project. Fisk ( 1982) desc ri bes
the following Iypes as the mos t coml11 on Iy used : ( I) traditi onal architec t/cngineer
(AlE) contract. (2) design/constructi on manage r (D/eM) cont ract, (3) professional
constru ction manager (PCM) co ntrac t, a nd (4) des ign- build (turnkey) co ntrac t.
8.2.1 Traditional ArchitecUEngineer Contract
The trad itional A/E contract limits the e ngineering servi ces provided durin g con-
struc ti on to occasional visilS to the construct ion site and a few other functio ns, such
as review a nd approval of shop drawin gs, interpretati on of pl ans a nd spec ifica ti ons
durin g cons tru cti on, evaluati on of claims by co ntrac tors, and review of payment to
the co ntrac tor.
8.2.2 Design/Construction Manager Contract
The DieM contract is simil ar to the A/E co nt ract exce pt th at the projec t manage r
of the AlE firm is responsible to the ow ner durin g bo th design and constru cti on.
The projec t manager must sa ti sfy all projec t needs, incl uding sc hed ulin g, cos t con-
trol, quality cont rol, prepurchasi ng, and coordinati on of all work. The res ponsibil -
ity of the firm does not end until th e ow ner's linal accepta nce o f the projec t.
8.2.3 Professional Construction Manager Contract
Under the PCM contract, the profess ional construc ti on manageme nt finn on ly acts
as a representative o f the ow ner durin g the ent ire durati on of the projec t but is not
in volved in the design or construc ti on of the treatmen t plant. Under thi s type of co n-
tract, the const ruction manager operates as an orga nization, /l ot as an individual ,
CONSTRUCTION MANAGEM ENT 599
a nd is responsible for the total project time, cost and quality con trol, and supervi-
sio n; the constru ct ion manager also has control over the fun ction s of the AlE firm
and the con tractor.
8.2.4 Design-Build Contract
The design- build con trac t is commonly called turnkey cOIIstl'llctioll . The owner con-
trac ts wi th a single firm that plans, designs, and constructs the en tire project wit h
its own in-house capabilities. In additi on, the turnkey contrac tor often offers the
ow ner financial incentives. The major advantages o f turnkey cons tru cti on are (I)
the elimination of claims by the contractor against the owner for errors or omissio ns
in the plans and specifications, and (2) the ab ility to design co nstructi on since the
dcsig n of eac h indepcndcnt phasc of the project is completcd without havi ng to wa it
for completio n of the overall project design, which is often referred to as fa st track.
The di sadvantage of turnkey co ntracts is that there is no compet iti ve bidding process
for plant construc ti on. Thus turnkey co nstructi on cont rac ts are illegal in the United
States for all publicly fu nded construct ion projects, although it is commonly used
to build water treatme nt plants in the rest of the world and is applicable to any pro-
ject in the Un ited States providcd that the project uses onl y private funds. In recent
years, the design- build-{)perate (DBO) cont ract has become popular world wide, in-
cluding the United States, due to the increasing number of pri vatized projects.
If a design- build co ntract is chose n, the firm should be selec ted on the basis of
individual experti se, previous designing experie nce in the partic ul ar lype of projcct,
financial status, and the reputati on of the finn . Althou gh it appears that thi s concept
is beneficial and saves a great deal of cost to the ow ner, case histories have proved
that thi s is not necessarily true. Without competitive bidding procedures, the COI1-
structi on costs tend to be hi ghe r, and the quality of the design and specifications is
not necessarily superior.
8.3 CONSTRUCTION MANAGEMENT
Co nst ruction involves the combined ski ll s of e ngineeri ng, science, a nd organi za ti on,
in addition to the usc of educated guesses and calcu lated risks. NUlllcrous complex
deta ils must be carefully coordinated among various parties: the ow ner, engi neers,
arc hitec ts, ge neral and specialty contractors, manufac turers, materi al suppli ers,
equipm ent distributors, governme ntal agencies, utilities, labor, and so on.
Construction contractors are private enterprises that e ngage in various types of co n-
structi on projec ts, including wa ter suppl y. The prime objec ti ves of the co ntrac tor
are ( I) to complete the const ru cti on projec t to the sat isfacti on of the owner, wi thi n
the set budget and on sc hedul e, a nd (2) to make a pro fit. Since a co ntrac tor's goal
is to complete the project as quic kly and as econom icall y as possible, there is al-
ways the possibility of poor·quality work on the projcct.
Engineers, on the other hand, tcnd to belicve that if their design is COITCct and the
plans and spccilicati ons are prepared carefully, lield construction will proceed without
600 MANAGEMENT OF PROCUREMENT AND CONSTR UCT ION PHASES
problems. However, Illany case hi stories have proved thi s idea to be far from true and
e ngineers therefore need to e nsure that all work and constructio n practices are equal
tn, or better than, the standards dictated in the co nstruction co ntract documen ts. For
these reaso ns it is not surpri sing th at owners arc reluctant to place the responsibility
for overseeing the quality of work on the project solely in the hands of the contractor.
Thc co nstru cti o n manage ment program is a vital tool in co mpl eting the con-
stru cti o n phase of a project in a sa ti sfactory mann er, for it meets the requirement s
of bo th the ow ner and the e ngineers. As a matter of fact, many federal co nstruction
co ntracts req uire the resident in spector, who is employed by the co ntractor, to meet
th e co ntrac tor quality control requirements. Ideally, construction management is han-
dled by the linn that designed and prepared the plans and speciticati ons, since it is
intimatel y familiar with the plant site and design de tail s and there fore capabl e of
clTectively administering the const ruction phase, providing quality control for the
ow ner, sec urin g the public safet y, and maintainin g the professional reputation of
both the design firm and the co ntractor.
Co nstru ction administration and construction quality con trol should be charac-
ter ized by co ntinu o us o n-site inspec tion by o ne or more ex perie nced and competent
engi nee rs who are tec hni cally qualili ed . These inspectors are often know as res i-
dent enginee rs. The res ident engineer is respo nsible for ensuri ng that all the desig n
details shown on the plans and spec ilicatioll s, and on the shop drawings approved
by th e des ig n e ngineer, are constructed in strict accordance with the req uirements.
The res ide nt engi nee r mu st also ens ure that th e work qualit y and construction prac-
ti ces are equ al to or better than the standard s di ctated in the construction documents.
O ne important res tric ti o n is that both the resident project represe ntative and th e
inspector are no t authorized to chan ge the plans and specification s, nor are they al -
lowed to imple ment th eir ow n interpretati o ns of the document s. If errors or prob-
le ms associated with construction practice ar ise during construction, they should im-
med iately be brought to the attention o f the project manager for furth er clarification.
Whenever there are o bviolls errors or if th e changes vastly improve the situat io n or
save tim e and material without sac rifi ci ng qualit y, these chan ges should be imple-
me nted by havi ng the ow ner sig n a formal change order (consent). The c han ge or-
der ge nerall y invo lves an adjustment in the construction fee.
The co nstru ction ope rati o n is a dynamic process req uirin g ma ximum effi ciency.
speed, and economy for the co ntractor. T here fore, the res ident e ngineer(s) should
make qui c k, yet appropriate dec isio ns o n all min or tec hni cal problems so as not to
delay the opera ti on of th e con tract or; reside nt enginee rs have th e authority to ap-
prove material s and work. As in other fi elds, co mputeri zed projec t management pro-
gra ms are available for tasks such as sc hedulin g, budgeting, and the plotting of job
progress. Use of thi s type of software is recommended because the tim e required to
perform these tasks is accelerated to a short period of time.
8.4 RESPONSIBILITY OF RESIDENT ENGINEERS
The function of the res ident eng ineer is di sc ussed in Secti o n 8.3. However, the ma-
j or responsibilities are as foll ows:
PROJECT CLOSEOUT 601
I. Be thoro ughl y famili ar with th e plans and spec ificn ti ons.
2. When the work does no t mee t th e req uireme nt s, notify the contractor of th e
construction site immedi ately and record the infrac tion and date of notili cn-
ti o n in the daily log. If the co ntractor chooses to ig nore th e noti ce, nclvise
the projec t manage r of the inac ti o n immed iate ly.
3. Write n daily report that inclmh.:s the daily activities, the items inspec ted,
any instructi o ns given. any agreements macle with the co nstru cti o n tea m, th e
weather conditions, and ot her pertinen t items. The daily report and di ary/ log
boo ks have legal importance in the eve nt o f contract di sputes.
4. Perform all on-site testin g spec ified in the contract docllments. The tes ts
shou ld be performed in an expeditious and ca reful mann er. All tes t I~'l ilures
Illust be reported to the co ntractor without de lay.
5. Perform all inspec tion s in a tim ely manner.
6. Discover unacceptab le work in its early stages and report it to the co ntrac·
tor as soo n as possible.
7. All problems beyond the ca pabilit y of the res ident engineer should be re-
ported to the project manage r of the eng inee rin g tlnn.
8. Do not be forced into makin g hasty dec isions; ca refull y evaluate the situa-
tion and the possible co nsequ ence of the dec isio n.
9. Be prepared to stand behind any decision made in regard to the co ntractor 's
work.
10. Be concerned w ith th e safe ty of the public and all persons associated w ith
the co nstru cti o n projec t and attcmpt to be objec ti ve in all situation s (control
cmoti o ns).
8.5 PROTESTS AND CLAIMS
Protests are de lined as disputes arising from th e iss uance of a co ntrac t change or·
der by the e nginee r against the objec ti o n of th e contractor. When the e ngineering
firm receives a letter of protest from the co ntractor, it should be exa min ed ca re full y
before acknowledging its rece ipt to assure that the basic requirements of the spec-
ilicati on are in cluded in the leiter. The project manager and the owner shou ld th en
rev iew the merits of the protest and iss lie a letter ad visin g the co ntractor of hi s ri g ht s
under th e contract to lile a legal c laim. One of the most important ite ms that is con-
sidered before presenting any legal c laim is documentation. When negoti ati o ns be-
twee n the ow ner and th e co ntrac tor fail , the data documented by the resident cngi-
nee r(s) and the co ntractor become va luable in the litiga ti o n or arbitration procedures.
8.6 PROJECT CLOSEOUT
The main ac tivities during project closeout are final inspecti on, acceptance of the plant
for the ow ner, and final payment to the contractor. However, many important issues rc-

602 MANAGEMENT OF PROCURE MENT AND CONSTRUCT ION PHASES
mai n: review of (he record (as bililt) drawings, transfer of the gua rantee, obtaining Ihe
operation and maintenance instructions for the equipment, return of bonds, composing
the plant orx:ration and maintenance manual, and compiling the punch list items. The
punch list is a detailed checkoff list that enumerates all items still requiring completion
or correc ti on before the work will be accepted and a ccrt ilicate of completion is issued.
In some cases, th e projec t is accepted as be ing "substanti all y co mpl ete" where
onl y min or items re main to he compl eted or the "bene licial use" o f the faci lity is
acce pted by the owner. Be nefi cial use (occupancy) is de li ned as the ow ner' s use of
the fac ility prior to its full completion. The date of benelicial use is the beginning
of the I-year guarantee peri od and total responsibility for proper maintenance and
operati on of the facility is transferred to the owner. Eve n if the certificate of sub-
sta nti al completion is filed with the punch list, the owne r will not release the re-
ta ined fund s until all remaining de liciencies arc co rrec ted and a wa iver of liens is
co mpl eted. Final pay ment signifi es the re lease of all retainers held during the pro-
ject; thi s ge ne rall y amount s to 5 to 10% of th e total projec t cost.
Liquidated Damages If the contractor fail s to complete the projec t by the date
speci fied, he is linancially liable to the ow ne r for a preagreed slim for each day be-
yond the orig inal fini sh date. This amollnt of money represents the opportunity cost
o f the own er resulting from suc h delays, and in the case o f water treatm ent plants,
thi s is usuall y the loss of income from the sale of potable water to c ustomers.
Cleanup The con tractor is responsible for clea ning up the co nstructi on si te at the
e nd of the job. The work wi ll not be accepted until all te mporary utilities, co n-
structi on access roads, sec urit y fen ces (for materi al storage), fi eld offi ces, stoc k-
piles, surplu s materi als, and scraps are re movcd fro m the pl ant site. Site cleanup is
the obligati on of the co ntrac tor and must be conduc ted at the co ntrac tor's expense.
T he co ntractor should not be permitted to bury any waste or rub bish on the pl ant
site unl ess the ow ner iss ues written approva l.
8.7 MAJOR CLOSEOUT ACTIVITIES
The major closeo ut ac ti vities are as follows:
I. Perform closeout inspec ti ons as prese nted in the punch list.
2. Prepare clear and concise docum ent ation of the closeout inspect ion of each
item on the punch list.
3. Implement pa rti al redu ction of the fi eld offi ce inspect ion staff,
4. Co ndu ct a fin al inspec ti on with qu alified e ngineers: a fin al punch list mllst
be developed fo r any out standin g defi ciencies requirin g correcti o n.
S. Pre pare th e record drawing with the co ntrac tor.
6. O btain th e follow ing items from the co ntrac tor be fore acce pting the project :
(a) G uara ntees
(b) Certifi ca ti on of inspec tio n
CONSTRUCTION PHASE CHECKLIST 603
(c) O perating manuals and instructi ons for all equipme nt
(d) Keying sc hedul es
(e) Spare parts that are specilied to be received
(I) Record draw ings
(g) Bond s (guarantee bonds, mainte nance bonds, etc.)
(h) Certifi cate of inspec ti on and compli ance issued by local agencies
(i) Waivers of liens
(j) Co nse nt of surety for final pay ment
7. Receive the requ est for Iinal progress payme nt from the co ntrac tor.
8. Check that all remaining work requirin g compl eti on is fixed properly.
9. Obtain signatures from the AlE project manager, the co ntrac tor, and the own er
on the certilicate of compl eti on. File thi s certifi cate with the oflice of the
Co unt y Recorder.
10. Notify the ow ner that the projec t is complete and read y for operation.
II . Arrange for final pay ment and release of the reta iner by the owner.
12. C lea n up and close the field om ce.
8.8 CONSTRUCTION PHASE CHECKLIST
The projec t manager and res ident engi neer are responsible for the following major
tas ks during the constructi on phase:
I. Approve the cost breakd own doc ume nts submitted by the low bidder.
2. Identify and list the contrac tor's key personnel and the phone numbers at
whic h they may be cont acted,
3. Fi nali ze the qu ality cont rol and inspectio n plans.
4. Hold preconstru ction meetin gs wi th the ow ner and with the cont rac tor and
owner.
S. Arra nge for surveys, test labs (i.e., determin e co nc rete strength), and mill
and fac tory inspec ti ons.
6. Arrange for the co nstructi on doc umentation tools (photos, di ary, fie ld log).
7. Issue a notice to proceed .
8. Review the mobili zati on and co nstructi on sc hedule as submitted by the co n-
trac tor.
9. Prov ide the necessary co nstruction manage ment servi ce for on-site qu ality
assurance of materi als and work; the scope of thi s servi ce depends on the
type of constru ction contrac t.
10. Prepare dail y construction reports and a di aryllie ld log. Transmit these to the
projec t manager.
I I. Prepare and di stribute monthl y reports.
12. Rev iew and process shop drawi ngs with the transmitt als.
604 MANAGEMENT OF PROCUREMENT ANO CONSTRUCTION PHASES
13. Receive, log, and tran smit all s ubmitta ls from the co ntrac tor.
14. Review the record drawings and update th em o n a monthly basis.
15. Confirm the amount o f eac h payment.
16. Process a ll prog ress payments.
17. Process the relocat ion of retainer fees.
18 . Estimate the proposed cha nges and the ex tra work req uired by these cha nges.
19. Prepare, log, a nd process a ll change o rd ers.
20. Docu ment all extra work and work order protests.
2 1. Process and document time delays and requests for time exte nsions.
22. Assist in negoti ati o ns related to di sputed work.
23. Administer the ow ner' s acceptance of " beneficial usc."
24. Compile a punc h li st and arrange for fin al inspec tion and job acce ptance.
25. Prepare a no tice for projec t completion , fin al payme nt , and fin al billing.
Advise the ow ner th at a I-year g uara ntee bond is required from the con-
tractor prior to occ upation.
26. Finali ze the record (as built) d raw in gs.
27. Prepare a construction con tract sUlllm ary repo rt and liIe thi s in the con-
stru ct ion man age ment project docum ent file.
28. Inspec t the project prior to the end of the wa rranty; prepare a re port on the
findings and make it part of the projec t man agement til e.
29. Evaluate the perfo rm ance of th e co nstruction managc mc nt tea m.
30. File all pertinent documents for permanen t record.
BIBLIOGRAPHY
Fisk, E. R.. COllsl m elioll Projeci Admillislratiol/, 2nd cd .. Wiley. Ncw York. 1982.
Montgomery, J. M., Consulting Enginccrs. Projeci COl1lrol Mallllal , J.M. Mon tgo mery
Consulting Engi necrs, Pasadena. CA. 1976.
O'Brien, J. 1., CPM il/ COIlSlrIIc/ioll Mal/agell/elll, ~ rd cd .. McGraw- HilI. New York. 198 ~ .
PMI, PlVject Managemelll Seminar/Symposillm ; IIII/Ova/ioll, Project Management Institut e.
Drexel Hill . PA. 1984.
PM I. Research Goals with PlVjec/ Managemelll, Projcct Management Institutc, Drcxel Hill.
PA. 1987.
CHAPTER 9
Operations and Maintenance Manual
and Operator Training
Since water treatment plants generally require a period of 2 years to be constructed, ad-
equate time ex ists for the preparation of the operatio n and maintenance manual. TIlis
manual must be provided to the plant personnel at least I to 2 months before plant
startup. Traini ng sessions for the plant operators must begin during the final month of
plant construction, when a substanti al portion of the plant is completed. Thus the owner
shou ld ei ther hire or assign the key positions of the operational stalT by thi s date. All
new members of the plant operating stalT--especially the superintendent. chief opera-
tor, and maintenance supervisor-arc advised to visi t the construction site frequently,
beginning in the earl y stages of construction, so that they arc familiar wi th the location,
elevatio n. and method of installation of all pi pes. weirs. valves. and other items that are
not visible aner plant construction is completed. Most of these structures are either buried
in the ground or embedded in the concrete walls and slab once the plant is finished .
Today, all municipal and public works official s, as we ll as directors of private
water purveyors, are aware of the new and stringe nt drinking water quality stan -
dards set by the EPA and the states, and the requirements imposed on the water
treatme nt process. In th e mid st of all these complex interactio ns of needs and con-
cerns are the public oflicials and th e directors of the private sector, who have iden·
tified the needs of the ir communities for improved and more sophisticated wate r
treatment and co ntrol. Furthermore, the public is demanding safe drinking water.
Thus milli o ns of dollars have been funded for the construct io n of tec hnolog icall y
sophi sticated , complex, and ex pensive water treatm ent plants and the modification
of existing plants to meet the new req uirements. Yet use of hig h-quality eq uipme nt
and comput er-based instrum entati on and control does not guara ntee the co ntinu ous
productio n of safe water; th e plant mu st be operated by qu alified and well-traincd
personnel. Thus the need for a well·organi zed and practical plant operati ons and
mainte nance manual , as well as sys tematic and hi ghl y skilled o n·site training, is ob-
vio us. Unfortunate di sas ters suc h as Three Mile Island clearly demon strate thi s need .
9.1 OPERATIONS AND MAINTENANCE MANUAL
Too many o perati o ns and mai ntenance (O& M) manuals are re legated to occupyi ng
she lf space on the bookshelves of water treatm en t plants. They arc seld o m used as
605
 606 OPERATIONS AND MA INTE NANCE MANUAL AND OPERATOR TRAIN IN G
OPERATIONS AND MAINTENANCE MANUAL 607

a refe re nce beca use they arc ei ther abstracts of textbooks or "boilerplate" that has pocket gu ide th at summari zes the key subj ec ts. If thi s co ncep t is adopted , the O&M
little practical va lu e. A lth oug h th e EPA iss ued g uide lines for co mpos ing an O&M manllal may be o rga ni zed as show n below.
manual in 1974, the key to creatin g a lIseful ma nual is to be prac ti cal and spec ific.
An O&M manual should have practical fea tures such as discussions o f the basic Main O&M Manual
principles used in the plant; talk-through prese nt ati o ns of the process layo ut s; de-
tailed des<.:riptions of eac h pro<.:ess co mponent ; step-by-s tep operatin g instruct io ns I. Purpose of the treatm ent process
for eac h process and process co mpone nt ; di sc uss ions of th e variables th at arise dur-
2. Process descriptions
in g process control and process status evaluatio n; descriptions a nd ca talog cuts of
all major eq uipm ent; tabulated troubleshooting g uid es li stin g th e probable causes
and solution s for problems common ly e nco unte red in th e field; and e mergency re- TABLE 9.1-1 Operations Checklist (Example)
sponse procedures.
O,e:,.ATIONI CHECKLIST
T he fo rm al o f the O& M manu al can be any style; however, th e manual should
at leas t inc lud e th e following subj ec ts ~.,c· _____________

I. Technical literature (eq uipm ent)

"-"""
2. Process opera ting Illanual """"",,,,,lion EQvlpm'l'Il No. SlaNt
3. Preventive operati o n proced ures
4. Standard operatin g pro<.:edures
5. Maintenance serv icing sc hedul es
6. Eme rgency respo nse proced ures
7. Po li <.:y formulation procedures

T he process operatin g manu al listed above (po int 2) is the heart of an O&M man-
ual. This sec ti o n shou ld inc lude the following ite ms:

I. Process description
2. Design crit eria I

3. Operat ing procedures

4. Tro ubl es hootin g gu ides
5. Emerge ncy response progra m
6. Servicing and mainte nance req uire ments I

Photographs of th e co nt ro l sys te ms and the major equipment of eac h process train

and a narrative instructio n- which provides informati o n on how to start up , shut down,
and troubles hoot respec ti ve pieces of eq uipm ent- have been found to be very elTec-
ti ve in ed uca ting plant perso nn el. Use of suc h documents (videocassettes) provides
even new<.:omers to the plant with a visual direc ti o n o n how to operate and mai ntain
(l ilY syste m. The use of sc hemati c illustration s is always helpful in creatin g a better
understa nding of the elec tri cal, electronic contro l, and ex tensive valving schemes.
One of th e mos t popular and effective O&M m anuals is th at prepared by televi-
sio n and automobile manufacturers. Their conc ise manu als arc alm os t always on ly
a few pages lo ng; a su mm ary of the key items support ed by a detailed description
of eac h fun cti o n. The plan t O&M manual ca n also be arranged in a similar mann er.
Thi s type o f form at a ll ows the plant ope rat or to obta in key informatio n from a slllall
608 OPERATIONS AND MAINTENANCE MANUAL AND OPERATOR TRA INING OPERATIONS AND MAINTENANCE MANUAL 609

~. Component descriptions Operator Pocket Guide

4. Detai led operating procedures
5. Performance evaluation and process control I. Flow sc hematics
6. Troubleshooting guide 2. Concise operating procedures
7. Maintenance and servicing schedule 3. Troubleshooting guides
8. Emerge ncy response program 4. Sample problems, including emergency response

TABLE 9.1·3 Problem IdentilicCllion Form

ThllLE 9.1-2 Equipment Data Form

IQUIP .. I[HT UA TA ~"u

PROBLEM STATtMENT
P\altl ,..,. _ _ Equ~ De.a\ptk'll'l
co- C.I.·
EQ\I!ptMnt No.
c-....... " $ut>;oct
~don of ProCIlMI:tAftad'l.hlehi' 0I1tdd111Ot\&! ,hHU" nHdtd.)
SuP9ll., Add/ ...

e-_
UW'lul'K1v'"
.._ NO

....., No.
0
1~'Don 0.1.
""'. ••
ttfCOl'l'VNll6e<I s,p.,. p."" •, Suoo..ttd SofuISon
~ •
•,
.-
S9Md
TDH

Typo .-
ComtMn~
SubmM eM copy .Kh to:

Ptam SuPt"nltl'lC!l'" ProjKt CoordI"1Jaf' Sup. SUI. Engin",

DRMMOTOR

. "'.......... Su9Ph, Add'tu ,, AESOUJllON OF PAOBt..E.M (Oe<:\fion JCorndye .tdonI)

......
" - NO.

...........
No
00"
0
I
••

'-
"""-*" Coot I

....... DelAC Typo

v_
S9Md
""-"
"""'-
.... au.
""-
Cydoo
'.F.
.........
.., RN

eo-...
II
•
•"
••,
04STRl BlfTlC)N
Sup. Saft. EI9"-
Pro;.ct CootdInlW
PIIn( &.rPtMllnoJlrrt (JMWOCf)

P\a/It Su~landMtI
610 OPERATIONS AND MAINTENANCE MANUAL AND OPERATOR TRAIN ING
Maintenance Pocket Guide
1. Mai nte nance proced ures
2. Servicing sc hedules
3. Schematics o f major processes and control loops
The O&M manu al should additi onall y co nta in thc recomlllc ndcd types of form s
used for records, logs, and reports. Today, data logg in g by means of co mputers is
almost the standa rd method of producing the plant opera ti on log and report.
Therefore. an additional operatio n log shee t is not nccessary. Howevcr, it is rec-
omlllended th at a few addit ional fo rm s be di stributed to the operators so that they
can famil ia ri ze themselves with eac h piece of equipmcnt, its location, and its ph ys-
ical conditi on (e.g., the need for painting). Ove rreliance on a com puteri zed sys tem
wi ll make plant operators ill prepared for any emerge nc ies resulting from 1~1 illirc of
the computer. Tables 9.1 - 1 through 9. 1-3 arc exam ples of thi s type of for m.
9.2 OPERATOR TRAINING BEFORE PLANT STARTUP
There arc two key requirements for successful plant startup: ( I) good operat ion and
proper maintenance and (2) fam iliarity with a ll plant faci lities and a working kn ow l-
edge o f all water treatment processes. The goal of the operator training sessio ns
(prior to startup) is to educa te the plant opera tors in th ese two essenti al items. The
training sess ions should last a minimum of three working days with the O&M man -
ual as the tex tbook. The project engineer and the heads o f the mec han ical, e lect ri -
cal, instrumentation, and control disciplines are ass igned as instru ctors for eac h area
of experti se. Represen tati ves of major equipm en t m anu l~1c turcrs should also be in-
vited to participate in the trai ni ng sessions. Full use of audi ovisual aids such as
videotapes, color slides, and a chalkboard is hi gh ly recolllll1ended,
BIBLIOGRAPHY
EPA , COl1sideratiolls for Preparatioll of Operation alld Mail1lel/{/llce Manuals. EPA/43019-
74· 00 1, U.S. Environment al Protecti on Agency. Washingto n. DC. 1974,
CHAPTER 10
Plant Startup and Follow-Up Services
10.1 STARTUP
Evcn if the plant is co nstructed exactly according to the plans and specificatio ns,
plant operators must always debug a few items durin g plant startup. Moreover, cer-
tain processes, such as ordinary granular medium filters and sludge blanket reac tor-
cl ari fi ers, requ ire a period of maturation in order to function properly, Plant s~artll p
is a hec tic time for plant operators. Many equ ipment and process co ntrol 1l.ll1tS re-
quire adjustment. Therefore, se rious problems a nd complications are most hkel ~ 10
arise during thi s period, The pla nt ow ner is st rongly advi sed to co ntract the services
of a qualified professional consulting linn so that problems often associa~ed wi th
pla nt and process start up can be avoided, thereby quickly plac ing the plant Into ser-
vice with fewer complications; the firm may also troubl es hoot and optllTI1 Ze process
efficiency, The most appropriate choice wou ld be to hire the engineeri ng firm ,th ~t
conducted the design, as we ll as co nstructi on manage ment, simply bec~ u se It ,IS
the most familiar wit h the plant , prov ided that it is qu alified and experi enced III
doing so.
Optimi zati on of coagul ant dosage, adj ustment of mixing conditions for t~~cc u­
lati on as well as ozo nation and di sinfect ion, optimizati on of filter wash cond itIOn s,
ensuring the proper pump operation sequence, and careful ana,lys is and co nt ro l. of
the fini shed wate r quality prior to delivery to the customer are Ite ms that are eflec-
tively executed by the co nsulting firm . Plant startup should be initiated at 20 to 25%
of plant capacity, continuin g until the majority of debu gg ing is compl ete and the
plant is steadil y prod ucing good- qu ality water, and before deli vering the water to
the public. It is important that the plant man ager create a solid plan for startup and
has all ava ilable plant operational sta ff participate, It is very beneficJaI I ~ w,ell-
experi enced plant operators from other plants and consulti ng engineers are Invlled
to partic ipate durin g thi s phase.
10.2 FEEDBACK
Feedback from the ow ner (client) wi th respect to the quality of the design, the pe r-
forma nce o f the e ngineers, and the performance and e ffi ciency of the fac ility (wa-
ter treatment plant) is abso lutely necessary for maintaining quality engineering se r-
611
612 PLANT STARTUP AND FOLLOW-U P SERVICES OPERATOR TRAINI NG SERVICES 613

vi ces. Most ow ners me re lu ctant to vo ice their di ssati s l~1c ti on or their constructi ve and " How can we improve our se rvice?" ( feed bac k). It is a cold fact that the per~
opinions unl ess outrageous mi stakes have been mad e. Instead,. the unh appy c l.ic nt formance of th e engineers and thc lirm will ne ither improve nor prog ress unl ess
ncver awards another job 10 the e ng ineering finn and th e e nginee rs have no Idca pro per feedhack is o btained from the c li ent.
what ca used the di ssatisl~lction. Thus, from th e standpoint o f bOlh busincss and en ~ Ge nerally, once all th e debuggi ng has been performed and thc water treat men t
gineering, il is il1"lperali ve lh at e ngineers ask the cl ie nt "How havc wc been doing?" plant ha s oce n o peratin g in a sati sfac tory manner, the engineers who were assigned
to the project beg in work o n the next project. Seldom do engi neers have a cha nce
to vi ew the operati o nal plant unl ess they are asked by th e c li ent to re mcd y s i g nin ~
TABLE JO.2~1 Project Feedback Report Form (Ex ..,mlllc)
ca nt problcms. Howeve r, these pro blems m-e ortell min or in nature: problcms with
PROJECT FEEDBACK REPORT mec hani cal equipm e nt , problems assoc iated with th e e lec tric al or co ntro l syste ms,
FROM: ____________________
or problems wi th the c ivil work s o r architectura l items. Eng inee rs frolll these vari-
Oli S di sc ipl ines arc capable of handling these types of problem; thus the projec t en-
TO: ______________________
Dhision or Humin Resources Ind Technotogy ginee r and project manager are gencrally not in volved or even supplied wit h the dc~
tai ls of the proble m. This is espec iall y tru e for busy consultin g e ngineers.
PROJECT CIIEC~ APPROPIUAn SPACE Consequentl y, so me e nginee rs con tinu e makin g the sallle mi stakes ove r and ove r.
JOB NO. exlle,.1 Dnl,,, ProNtrn
For thi s reaso n, n project e nginee r shou ld visit eac h plan t th at he or she has de~
CLIENT Spccln.u,lo" Proble,,,
sig ned at least once a yea r for the first 5 yea rs. This allows the engineer (po tc n-
OBSERVER SUndlfd De,,1!
DATE INlnA TED Spc(ill Oculi tiall y) to obtain feedback from the plant operators. Furthermore, th e engineer shou ld
SPEC. SECT. cl PAGE Spec !,1 u.,) Problem co mpose a me mora ndum describing thc lindings during the plant visitation, in ~
DRAWINO SnEET NO. O.lIn eluding the ow ne r's criti cis ms and comments (Tnble 1O.2~ I). The mem o should be
liled in the project liIe and should a lso he disse minat ed among the key eng ineers
of the project team ; thi s a ll ows eve ry key e ng ineer to benefit from the lindings.
CASE HISTORY OR PROBLEM: (Auiel. skelches or .ddilionat in(olmltion)
When an error of a serious nature is di scove red, somc project enginee rs lind it
vc ry difli cult to di sc lose thi s informati o n to their boss and ot her e ng ineers and the
error te nd s to be cove red up. Thi s is not in the best interest of the engi neer or the
linn. A ll grave e rrors should be reported immediately to both the chi e f e nginee r and
the legal department o rthe firm . Numerous cases have need less ly e nded lip in court ,
cos tin g thc e ng ineering linn 10 or eve n 100 tim es more than if the problem had
quickly been remed ied ; Ihe laller case usuall y cosis less Ih an $50,000. Any delay in
correc ti ve ac ti on not o nl y aggravates and angers the client but ca n potentially lead
SUGGESTI:D IMrROVEMENT: to a laws uit.

10.3 OPERATOR TRAINING SERVICES

Both the EPA and public ofli c ial s recog ni ze thc import ance of opera tor tminin g for
was tewater treatme nt pl ant s in improving plant e lTcct ive ness and , more important,
in cultin g opera ti on and ma inte nance cos ts. T he EPA requires th at a ll publicly fund ed
ACTION TAKEN IN nils CASE: pl ant s pro vide operator train ing se rvices. Due to th e requirements imposed by both
the states and the EPA o n wa ter treatme nt pla nt s, many ad ministrators have beg un
to fcc I the same need for o perators of public fac ilities, inc lud ing water treatm ent
plant s. Thi s co ncep t is not a new idea: O pe rato r tra ining has bee n lIsed fo r a nUI1l ~
bel' of years to he lp plnnt o pera tors obtain th eir o pera tor ce rtili cat ion (req uired by
all slalcs).
T he ope rator tra ining program has been offe red CIS a s tate~spon sored short sc hool
prog ram to he lp operato rs pass the ce rtili cati o n exa min ati o n, as a corres pond ence
614 PLANT START UP AN D FOLLOW-UP SERVICES
course, as in-plant prac ti cal techni cal assistance, and as in-pla nt training. Of these
programs, the in-plant training is undoubtedly the mos t comprehe nsive and effec-
ti vc, since it is a combination of hands-on a nd classroom instructi on condu cted at
th e treatme nt plant. This type o f program provides operators with a sound under-
stan ding o f the wate r trcatm ent process and plant manage me nt, as well as useful in-
form ation with las ting value. Several we ll-quali!ied co nsulting !inns and profes-
sional orga ni zatio ns offer this type of trainin g service.
The operator training program generall y teaches pl ant operators how to analyze
plan t processes and prov ides instru ction in mathemati cs, chemistry, and biology.
Additional topics include housekeeping, safety equipm ent inspec ti on tec hniques,
and cmerge ncy procedu res. In ad diti on to the operator program, several other types
of specific instru cti on are avail able in categori es such as management , maintenance,
and la boratory procedures. Ma nageme nt training is inte nded for the pl ant manage-
ment sta ff a nd foc uses on hum an reso urces devel opment, quality cont ro l, qu ality as-
surance, process a nalys is, and writing skills. Maintenance training prov ides in-
struc ti on in equ ipment performance standards, elec trical and control system wiring
di agrams, the sc hedulin g of preventi ve and correc ti ve maintena nce, mainte nance
manageme nt prog ramming, mec hani cal and electrical troubl es hooting, and emer-
gency services.
The laboratory traini ng program is designed to prov ide detailed and fun cti onal
tra ining for plant operators in standard testin g for ro utine opera ti ons and on-site spe-
cific process co nt rol, including tasks such as jar tes tin g, core sampli ng of filt er bed,
and filter medium a nal ys is. Laboratory training should not be limited to pla nt op-
era tors; the laboratory staff should be prov ided with on-site instru ctio n as well as
suppl emental tra ining at a state-certifi ed laborat ory in Gianlia and virus monitor-
ing; the use of gas c hro matography/m ass spec trometer equipme nt for the analysis
of organic polluta nts; adso rpti on and emission spec troscopy, inducti ve ly coupled
plas ma, and other tec hniques used in heavy metal anal ys is; and tec hniqu es used in
conductin g as bestos fiber coun ts wi th the transmission elec tro n mi croscope. Since
the tracer test for evaluati ng the tan k detent ion tim e has becomc a n important iss ue
with respcc tto disinfec tion criteria, it is essential th at plan t personnel be trai ned to
perform the tes t and bc allowed to practi ce th e tec hniqu e.
10.4 PLANT OPERATION AND MANAGEMENT SERVICES
Before an y type o f training program can be imple me nted, th e ow ner must hire the
water treatm ent personnel. The e ntire plant operati on a nd manage ment is ofte n con-
ducted by a private pro fessional organi zati on. The primary reaso ns for this arra nge-
ment are cost- effec tive plant operati on, bett er process control , and the steady pro-
du ctio n of safe drinking water by a we ll-ed ucated and well-trained (licensed)
professional staff. The cost-effecti veness an d effecti ve performance provided by
qu alified private professional fi rms are both well kn ow n a nd proven in ma ny sec-
tors o f society. All major water treatmcnt plants in Paris are operated a nd managed
by a private professional firm (those in London have also implemented this option),
PLANT OPERATION AND MANAGEME NT SERVICES 615
and th is concept has also gained popularity in the Un ited States. Several consult ing
fi rms and pri vate professional orga ni zati ons offe r thi s type of service.
One o f the newest app roac hes to effec tive plant management is the comprehen-
sive mainte nance program using a computeri zed mai ntenance manageme nt system.
One type, the Computcr Assisted Manage ment Syste m for Envi ronme nt al Operati ons
(CA MEO), developed by Envirotech Operating Services, generates data on issues
such as eq uipme nt age and number of breakd owns. The master report al so provides
info rmati on on equipme nt cost and re pair. By using this type of system, an orga-
ni zed mainte nance prog ram, including preventi ve maintena ncc work and prioriti zed
work orders, ca n easil y be established. The effec ti ve use of thermographic analys is,
vi brati on detec tion, and analys is systems such as the Snapshot syste m can ide ntify
problems in major equipme nt before the proble ms become se rious, thus avoiding
costl y and major repairs. These types of ass istance make overall management of a
water treat ment pl ant more efficien t and eco nomical and are provided by a few pro-
fessional service firms.
Another spec ial type of se rvice that may become esse nti al for water purveyors
is th at of a compl ete laboratory which perform s tasks such as sample collecti on,
analysis, data evaluation, and report pre parati on. As a res ult of the new EPA regu-
lati on, based on the Safe Drinkin g Wate r Act, very speci fi c require ments are im-
posed on all public water systems in regard to sampling: The anal ys is of over 100
and over 180 regul ated cont aminants was required by the earl y 1990s and the year
2000. res pecti vely, and the report must follow a speci lic format. It is obvious that
most sma ll to midsized public water suppl y agenc ies are presently incapable of han-
dlin g the ana lysis of all contaminan ts, due to th e need for special sophisti cated an-
alyzers and well-qu ali fied scie ntists/chemists. In the author's opinion. one of the
best solutions is either to have the state use the services of certified laboratori es or
to have eac h water purveyor enter a servi ce contract with a certifi ed pri vate labo-
ratory th at will handle these requireme nts. Several certified laboratori es prese ntl y
operate in major metropolitan regio ns and are capable of meeting these new needs
for a reasonable fee.
 ABBREVIATIONS 617

DICM design/construction managemcn t

DCS distributed con trol systcm
DHS Department of Health Services
APPENDIX 1 DNA deoxyribonuclcic acid
DO di ssolved oxyge n
EBCT cmpty bed contact tim c
Abbreviations ED elcc trodialysis
EIS cnvironmental impact study or statement
ENR Ellgin eering News- Record
ENR Index Ellgineering News- Reco rd Construction Index
EPA U.S. Environmental Protection Agency
AAPC American Assoc iation of Pipe Corrosion ES Effective size (Hazen)
AASHTO American Association of th e State Highway Transport ation ESWTR Enhanced Surface Wat er Treatmcnt Rule
Ollicials OF Fahrenheit degree
ACCI American Conference of Governm ental Industri es FCI Fluid Controls Institute
AlE architec t/e ngineer FDDI liber-di stributed data interface
AIHA American Industrial Hygiene Associmion FEA finite clement analysi s
ANSI American National Standards Institute FIN facility informati on network
AOC assimilable organic carbon FPA fl avor profile analysis
AQMD Air Quality Management District Fr Froudc numbcr
ASCE American Society of Civi l Engineers ft foot or feet
ASTM Ameri can Society for Testing and Material s ft' /h cu bic foot per hour
ASU areal standard units ft'/min cubic foot pcr minute
AWWA American Water Works AssociaiiOIl ft'/s cubic foot per second
BAF bacteri al or biologically ac ti ve lilter g gram
BAT best available technology g acceleration of gravity
BDOC biodegradable diss olved organic carbon G velocity gradient or mi xing intensity (S - I)
BOM biodegradable organic maHer G X t product of G (S - I) and time t (s)
°C Celsius or Centigrade, degree CAC granular activated carbon
CADD computer-aided design and drafling COX gaseous oxygen
CCTV closed-circuit telev ision gph ga llons per hour
CMS construction management service gpm gallons per minute
COD chemical oxygen demand gpm/ft gallons per minute per linear fOOl
CPM criti cal path method gpm/ft 2 or gp m/s f gallons per minute per square foot
CPU central processing unit CUI graphical user interface
CRT cathode ray tube G value velocity gradient
CS I Constructi on Specificati on Institute HAA haloacctic acids
CSO combincd sewer overn ow HOT hydraulic detention tim e
CSSE Conference of State San itary Engineers HPC heterotrophic platc count
eX t product of concentrat ion (mg/L) and tim e (min) HVAC heatin g, ventilation, and air-conditioning
Cv capac it y factor for valves ICR Informalion Collec tion Rule
OAF di ssolved ai r fl otation 10 inside diameter
DBO design, build, and operate IEEE Institute of Electri cal and Electronic Engi neers
DBP disinfectan t by-product in. inch
dc direc t current IOC inorganic chemical
616
616 ABBREVIATIONS ABBREVIATIONS 619

ISA In strument Society of America NEC National Electril:al Code

ISPM integrated system of project management NEMA National E lec tri ca l Manul~1c ture rs Association
Jtu Jackson turbidity lIllit NF nanotiltration
K kelvin (abso lut e temperature) NFPA National Fire Protec tion Association
kg/e m' kilogram per square centimeter ng nanogram
kPa kilopascal ng/L nanograms per liter
kW kilowall NIPDWR National Int erim Primary Drinking Wat er Reg ulation s
9
kWh kilowallhour nm nanometer (10 - )
LAN loca l area network NOM natural organic matter
Ib pound NPDES National Pollutant Discharge E limination System
LCR Lead and Co pper Ru Ie NPDWR National Primary Drinking Water Regulations
LI Langelier Index ntu nepherometri c turbidity unit
LID loop illterconncc tion diagram f1 ohm
LOH loss of head O&M operation and maintenance
LOX liquid oxygen OSHA Occupational Safety and Health Administration
I-'g/L microgram s per liler PA phenolphthalein alkalinity
I-' W em microhm s per ce ntim eter PAC powdered activated carbon
6
fJ-111 mi crometer (10 - ) P&ID process and instrum entation diagram
I-'g/L mi crogram s per lil er PC programmable co ntroll er
m mei er PCD process co ntrol diagram
m3/s cubic meters per second PCE tetrachloroethyl ene (perchloroethylene)
MA methyl orange alkalinity pc/L pi coc uries per liter
max maximulll PCM process co ntrol modul e
MCL maximum contaminant level PCM professional co nstru ction man age ment
MCLG max imum co ntaminant leve l goal PD process diagram
mEq/g milliequivalent per gram PERT program eva luation and review technique
MF micro n I(ration PIN process information network
mg milligram PLC programmabl e logic controll er
mg/L milli gram s per liter POADMA pol ydialJ yl dimethyl ammoniulll
mgd million gallons per da y ppd pound s per da y
m/h meter per hour ppm part s per million
MII3 2· methy I isoborneol PSA press ure-sw in g adsorption
min minutes psi pound per square inch
Min minimum PTB pitched turbine blade
mL millilit er PVC poly vinyl chloride
MUd mega liters per day RCML reco mm ended maximulll co ntaminant leve l
mill millimeter Re Rey nolds number
MPa Ill ega pasca l RFP requ es t ror proposal
MPN most probabl e number RFQ request for qualifi ca ti on
MTBE methyl tertiary butyl ether RMP ri sk management plannin g
MWCO mol ec ular weight cutoff RO reverse osmosIs
mW· s/c m2 milliwalts per second per square centimeter (ultraviolet light) RTU remot e terminal unit
MX 3-chl oro-4( d ieh loromethy 1)-5-hydrox y- 2[5H]- furanone s seco nd s
N newton ( kg"m/s' ) SCADA superv isory co ntrol and data acquisition
NAS National Academy or Sciences SDWA Safe Drinking Water Ac t
620 ABBR EVIATIONS

Sl Stabi lity Index

Sl unil international system of units
SMCLs Secondary Maximum Contam inant Levels
SOCs sy nthetic organic chemicals
APPENDIX 2
SUVA spec ilic UV absorbance
SWTR Surface Water TreallllCnl Rule
TCE trichl oroethy lene Basic Physical Tables and
TCR tota l col i form rule
TDH total dynamic head Greek Alphabet
TDS total di ssolved solids
THMs trihalomethancs
TOC total organic carbon
TON threshold odor number
TTHMs Systems of Basic UnIts
total trihalomethancs
UBC Uniform Bu ildin g Code Systems
UF ultraliltrat ion Metric Internat ional
English
UFRV un it fi lter rlln volume Designation Dimensions (FPS) (MKS) (SI)
UV"4 ullravi olct radia tion of 254 11m
VE value engi neering Lengt h (L) foo t (ft) meter (m) meter (m)
VFD Mass (M) pound (Ib) kilogram (kg) kilogram (kg)
variable-frequency drive
Time (I) second (sec) second (s) second (s)
VOCs volatile organic compounds
Electric current (A) ampere (A) ampe re (A) ampere (A)
VSA vacuum swi ng ad sorption Temperature (7) degree degree deg ree
W watt Fahrenheit (OF) Celsius (0C) Kelvin (OK)
WHO World Health Organi zation Luminous in- (/) candela (cd) candela (cd) candela (cd)
tensity

Decimal Multiples and Fractions of SI Units

Factor Prefix Symbol Factor Prefix Symbol
10' deka D (da) 10 - , deci d
10' hecto h 10 - 1 cen ti c
lO' kilo k 10 - ) milli m
U)6 mega M 10-' micro ~
Ul' giga G 10 - ' nano n
1012 tera T 10 - 12 pico P
10 15 femta F 10- " femto f
10 18 alta A 10 - 18 alto a

62 1
622 BASIC PHYSICAL TABLES AND GREEK ALPHABET BASIC PHYSICAL TABLES AND GREEK ALPHABET 623

Systems of Derived Units, Geometry, Mass Alphabet Table

Designation Dimensions FPS MKS SI Greek Greek English Russian English
Letter Name Equivalent Letter Equivalent
Area (L)' ft' m' m'
Static moment of area (L)' ft' m' Ill) Aa Alpha (a) Aa (a)
Moment of inertia of area (L)' ft' m' m" 66 (b)
B~ Beta (b)
Product of inertia of area (L)' ft' m" m" Bn (v)
Polar moment of inertia of area r~ Gamma (g) rr (g)
(L)" ft' m' m'
Volume (L)' ft ' m' m' t.8 Delta (d) A~ (d)
Static momen t of volume (L)' It' m' m' Ee (ye)
E Epsilon (e)
Z,
E
Momen t of inertia of volume (L) ' ft' m' 10' lKlK (zh)
Product of inertia of volume (L)' ft' Zeta (z) 33 (z)
m' m'
Polar moment of inertia of volume (U ft ' 10 ' m' Ii'Yl Eta (a) H" (i, e)
Mass" =M = Wig (M) Ib kg kg aa Theta (th) l1ii (e)7
Static moment of mass (M)(L) Ib-ft kg-m kg-m KK (k)
Moment of inertia of mass I, Iota (e) iln (I)
(M)(L)' Ib-ft' kg-m' kg-m'
Product of inertia of mass (M)(L)' Ib-ft' kg_m 2 kg-m ' KK Kappa (k) MM (m)
Polar moment of inertia of mass (M)(L)' Ib-ft' kg-m' kg-m ' HI{ (n)
A), Lambda (I)
Curvature of a curve (L) " 11ft 11m 11m 00 (6,0)
Torsion of a curve (L) " 11ft 11m 11m M .,. Mu (m) nn (p)
Plane angleb (R) rad rad rad Nv Nu (n) Pp (r)
Solid angleb (S) Sf sr Sf Ce (5)
=~ Xi (ks)
TT (t)
"In the English system (FPS) and in the metric systcm (MKS), the mass M is a derived unit, 00 Omicron (0) Yy (00)
d~fincd as the weight W divided by the acce leration due to gravity g. Ib := pound mass , kg =
ki logram mass. n" Pi (p) ct><j> (I)
lYfhe unit of plane angle called radian (rad) and the unit of solid angle called steradian (sr) arc Xx (kh)
Pp Rho (r)
supplemental units. 1\1\ (ts)
1 (T < Sigma (s) 6 'I. (eh)
T7 Tau (t) lliw (sh)
lli w (sheh)
Yv Upsilon (ii , 00)
b" 8
ct><!> Phi (f) b I bl (0)
Xx Chi (H) bb 9
33 (e)
"'>I< Psi (ps) 1010 (ti)
Ow Omega (0) 1I. (ya)
 METR IC (S I) SYSTEM AND SE LECTED USEFUL CONVERSION FACTORS 625
APPENDIX 3
APPENDI X 3 (Co l/ tinu ed)
Metric (5 1) System and Selected SPECIAL li TO .... .... to", fOOl )7 . )1"1", ..1.. pe ' 101.. )1.)01 /1 0 .0161
0.H96
I .}U Jo .. I.. pe' , .. III I.H8 1/1

Useful Conversion Factors COMPOUNI>

UNITS
IITtI ....
"1 U pe,

1 P'Jun~ P'"
po .. "" of "' . ..
>q~ ... ruol pOI hoy , sa
).1 /u .. k, pt , "I"'" me l.,

I ~ S .• ' .. m. .... ~"b'" "Mimcl~' PC' ", <u lld

1158 J /o~ l

liS ·· II,m '.

0.1 16 7

O.OOB9
.~o",."" ,.bi< 'cc'pc . d., 01 O. I U~ ~;lOJ .. m PC' wI>", .... ... PC' .. cond O. I IS~ k,/Ill'. S.)9
0.0600
I ~ ..I>ic 10<1' pc. mi~"I' PC' 16.67 ",ie.uJ i, ... PC' Ii ••• pt . ..,·.. ftd 16.67 ~I/II
. h"" .."d . .. lIk r•• 1 <H 1661 , I W' 1"" .. ~nd · 16.67, 10·l ,., 60.000
I ".1'1< 10<>1 pe, "'"un~ r" " ... 69.'" Iilor pcr ",oond P'" ho,·,., e 6~ . 91 1/. h O.olH9
U! 6.'9~ mi •• o""''''' pcr .. ,·und, 6. 99 ' ~", /,
9. lS ml/tl
rUn9
0. 1010
1 .. 11,," PC' . " . 9. l$ millilil'" po, 1L<.Uf<
... 9. H 'Ulimc'e"" 9.)S "III " 0.10111
IU. l",i<. uhm . p. .... .. nd p'" nltl., 14).1.1/. '" (I .OO 6~6
1 .,'kon po' d., PO ' lin." I ......
Q ' O. IUl ><I.'" ",illi.... ,." pe' .."ond 0.1431 mm l {.
O.~lI S ml/ml ,
6.'6
1.12
U , ~ll S m,ll, lil e, r-<' ,qu~f< mol. , p'" ""ond

...
I _.lIun 1"" do, 1"" ",utr fw'
,,' U.UIS ... " 'N ....IO,, 1"" ...,'und 1I.• 1 IJ ~ml . 1.11
I pound "' . .. pc . , ~ •• 1.1 n kilo"u m pe' "'<1 . ,. 1. I!Jk,Jh II.K91
",0.11 n ,,~m per >qu' ...... " ., O.II U"", ·
16."1 ~ .. "" pe l Ii •• , 16.01", 0.0614
1 _"~ .... " pc. <"b,~ 1\><>4
I P'J,,~d m." pe ' 10"." r..... l I.U"ki\u~"m l""",,·I<J
I.USk,Im U.671
PHYSICAL CUSTOMARY I ""lin d PO ' ~"''''p<>w.'.h,,u, U. I6911 mil1i¥"", pe' ;",.,1. 0. I690 ... ,JJ $, 9 17
QUANTITY ( " 'rhl&h~ 01 "Imp.eri.ol")
. , .•. • 16.67 p•• I. in I.'IUI.IIOII. p.:. "",,"d •• 'loI ulhpl , 't.,n.;., k"".. n in m.;1",· ,,",II 1>, Ihn numb..." "" ' 1I' ;lhl1 C<I ..i •• knl.
I fool
11""
0.:1041 III
0.91Um
1.211 'L• .• 6 .9'11 ",i",,,m.,,,, 4<plll e~.ft"" ... , ... <""" "M,," ipl, q", nli'y known ,n 1If1, i". .. fti" by ,hi, n ..",bt. '0 1'" _ttic O'I"i.oI<"' .
1.094 · · , .•.• 9. )) «".imo'," depth
1 m~~ 1.609 kill 0 .611S
1"1"'" Ineb
lloqu",c fool

Y""
,.".
I Iqu". 0.1)6 ... •
4 ,041 11\ 1
0.4047 h
liquor. mil . U90km 1
15' ,0 II
I ""bi. in<h
I pin! (liquid) ,. 16. J9
4Jl .l
L qUOIt ~6 .• <ubk ~nlim'I'"
,,! 0,9464 lilt ,
1 ,.llon
I <'Ubi< ' 001
1.")
I)
""n
OlU <ubI< .... ,. ,
I cubic fI.d
I >...·f ool

'm",ul<
I "CHI.
60 o«ond.
J.6UIo»oCOfl'"
"
1 "'Y 8&.4\1""'<01><1, I),OIlS?
or O. OI~ ""'pl«ond I I .S7
' .... k 0.6048 m.p~<ond I.6B4
UUU .... ' ......nd .
1.591
JU 6

,. ..n.
HJ.6,.".,
0. 0.41)6 kiloi"m
"",rn
0.907 .....
'" 0.to1 ... " 1< ..,,,

H .U ft,,,,,,.,P" ",U." ....1.'

6.195 kilofou'lo,,, Pc' loCI"'" ........
Ill. ) .,"Ion, ptr 'qu". mo'"
1.919 kilo •• "'on, pt. 1<1 ..... molOl
O. IOll mom

POWER 74S.1 W 0.001}4 r

0.7451 \W I .H I
1.000 W ' .000
0.19l 1/. 1.41

I.' "O,'ff F"h.r.IM!il

"" fo r uch C.loi".
dOl''' , pl., n -r

Source': J. J. TUIll[l , f/w ulbook (/fPII)'.I"iwf ClIlt.:lillllioIiJ. M cG ra w- HilI. New York. 1976.

624
626 METR IC (SI) SYSTEM AND SELECTED USEFUL CO NVERS ION FACTORS METRIC (SI) SYSTE M AND SELECTED USEFUL CONVERSION FACTORS 627

Quantities and SI Units QUANTITIES AND SI UNITS (Continued)

Expressions Expressions
in Terms of in Terms of
Unit Type Quantity Unit Symbol Other Units Unit Type Quantity Unit Symbol Other Units
Base units Length meter m Some other Viscosity- dynamic pascal second Pa·s
Mass ki logram kg derived
Time second s units
Electric current ampere A (conl'd)
Thermodynamic tem- kelvin K Thermal conductivity watt per meter W/m ·K
perature kelvin
Amount of substance mole mol Thermal flux density watt per square W/m'
Luminous intensity candela cd or irradiancc meter
Supple- Plane angle radian rad Thermal capacity or joule per ke lvin 11K
mentary Solid angle steradian sr entropy
units Permeability henry per meter Him
Derived units Frequency hertz S- 1 Permittivity farad per meter F/m
Hz
having Force newton N kg ' m/s' Magnetic field ampere per Nm
special Pressure, Stress pascal Pa Nlm' strength meter
names Luminance candela per
Energy, Work , joule 1 square meter
N' m
Quantity of heat Molar entropy joule per mole J/mol ·K
Power watt W l is kelvin
Electric charge coulomb C A·s
Electric potential
Potential difference
Electromotive volt V W/A
force
Electric resistance ohm n VIA
Electric conductance siemens S AIY
Electric capacitance farad F CIY
Magnetic flux weber Wb V's
Inductance henry H Wb/A
Magnetic flux density tesJa T Wb/m'
Luminous flux lumen 1m cd ·sr
Illuminance lux Ix hn/m2
Some other Area square meter m'
derived Volume cubic meter m'
units Velocity - angular radian per second radls
Velocity-linear meter per second mls
Acceleration- angu· radian per second rad/52
lar squared
Acceleration- linear meter per second m/s2
squared
Density (mass per ki logram per kglm'
unit volume) cubic meter
Moment of force and newton meter N' m
torque
 METRIC (SI) SYSTEM AND SELECTED USEFUL CONVERSION FACTORS 629
626 METRIC (SI) SYSTEM AND SELECTED USEFUL CONVER SION FACTOR S

METRIC (SI) CONVERSION AND USEFUL CONVERSION FACTORS The following lable shows Ihe number of pounds of chemical 10 add 10 I gallon
of water to make % solutions
Lenglh: I mil =2.54 X 10 - 3
em % solution Iblgal % solution Iblgal % solution Iblgal
I in. =2.54 cm 20.0 2.20
0.1 0.008 3.0 0.258
I ft =30.48 em 0.440 25.0 2.76
0.2 0.017 5.0
Iyd =0.914m 0.5 0.042 6.0 0.553 30.0 3.56
I mile =1.61 km 1.0 0.084 10.0 0.927 40.0 5.52
Area: I ft' =0.093 nl 2.0 0.170 15.0 1.473 50.0 8.33
I yd' =0.836 m'
I mile 2 =2.59 km '
I ha =10,000 m' = 2.47 acres
Volume: I ft' =28.32 L = 7.48 gal
I gal =3.785 L
I yd' = 0.7646 m'
I acrc- ft = 0.326 million gallon s
= 1235 m'
I m' = 35.31 ft ' = 264 gal
Weight: I dalton = 1.65 X 10 - 24 grain
I Ib = 453.6 g
I Ion (shari) = 2000 Ib = 907.18 kg
Temperature: (32°F) X 5/9 = °C
Feeding rate: I mg/L = 8.34 Ib/million gallon s
Flow rate: I fl ' ls = 448.8 gpm
= 0.6463 mgd
n J/min = 28.32 Umin
mgd = 1.547 ft ' /s
= 3785 n,'/d
= 3.785 MUd
1 n?ls = 22.826 mgd
I gpm!ft' = 2.5 m/h = 60 mid
Pressure: I psi = 0.0703 kglJcm'
(Ibtfin' ) (kg/cm' )
I psig = 14.7 Ib/in' = 2.31 ft H, O
I kPa = 0.145 psi
(N/m ' )
I MPa 145 psi
I bar 1.0197 kglicm'
14.5 psi
Power: I hp = 550 ft -Ib/s
= 42.44 BllI/min
I hp = 0.7457 kW
Gravitational acceleration = 32.2 fils' = 9.807 mis'
en Equ ivalent Temperature Readings ror Fah renheit and Celsius Scales
of = ~ °C + 32°. oC = ~ (OF - 320)
'" n;t
I
o
Pabr~,,·
bd t
Cent i- Fa h r<:o -
De;t
Ceol;. J F"Ilr~o·
,,,
Cent;. l'alu.,o- Ctllti. I Fah' ca. e e n!;. Fall1"rn _ C enti. Fa !'!,..,,,. Ce nt i_ Fa bre ll. C eDl i.
3
'" -,,'21.
«Tade ('"ad e beit !trad e bei t heil , n d., lici t
,~
,,'
" ad"
,,' I:. a d e b.,it cn.c!e
,,'
b ~t Cnd ..
---- --- '" -,.'"
<! CII:

---4~~ .~
- 273 - 29. 4 '" m.
d"
---
'" d"
'". '""., "0
17 . $ 50. 13.3
I lS.gr. .' 35. .5... 7 172. 4

".IS. ".
- C3e. - 2bO. -2'. 78 211.
- 20. 2 18. -1. 8 67. ll.g 3.5. 6 13~ . 6 57. m:.
".
-
-us. -250. -20. - 2~ . 9 15.
=~:2
57. 2
,,,.
'35.
78.3 212. 100. CD ~
--<00.
- 382.
-2~ O .
- 23a,
- 13.
- 'S . t
-ZS .l
-28.
lY . ~
20 . -6.7 ".
". 14 . 4
". I ".
36. 1
57. 2
57.8
114.
174 . 2

"'.
78. 9
".
".,
213.
213 .'
100. 6
18t, ~ "tl
".
36 . • 13&. 4 58· 214.
". ".
Q) "tl
- 351 .
-34 e,
- 720.
- 210.
-18.
- 17.
- 2i.8
- 27. 2
2T .
21 . 2 -,.
- 6. 1 ]~. ~ ~8. & 137. 58. 3 17&. sa. 215. 101.1
m
- 32S.
- 310.
- 200.
- I~O .
- Hi &
- 1&.
-27.
- 26 7 ".". -,.
-oS .e
".
".8
62.
16 .
10. 1
16.7
100.
'100.4
37.2
37 .'
38.
13'.
138.2
\39.
58 .9
".
5".4
Tn.
177. S
178.
80. 6
81 .
11 1.1
21.s. 6
216.
217.
102.
102. 2
,,,. s:::: z
- 292.
-274 .
- ISO.
- 170.
- IS.
-14 .S
' - 20: 1
- 26.
14.
24 .S -,.
-4.4 62. 41
6l.
17. 101.
'02.
3S .3 140. 60.
w .•
179.
,,,.
81.7 2 17. 4
102. '
~
0
- 250.
-"".
-m.
-100.
-150.
- H.
-Jl.
I
- 2.'>.0
- 25. ".
".
- 3.9
-,.
-3.3 ".
6~. 4
1i .2
l7. S
18. ,,,. ".
102. 2
,0<.
)S • •

3'U
HJ.
IH .S
H2. ".
~ 1.1
179. 6
181.
82.
82. 2
82.'
218.
m.
219. 2
103. 3
103 . ~ CD .,.X
-I~O - 12. -2~ . 4 20.0
". 65.
".
18. 1
,, <a.
'''. .1.7
,,,.
18U
". ",.
",.
104.
- 202.
- IS4.
-130
-1 20.
-11.2
-11.
I - 2U
-24 .
28.
- t.s
-,.
_2.2 tiil. ~
18.V
IS.
105.
105.S
40. 6
<T.
143. 6
H4. ".
62.2 '83.
OJ.'
,,,.
HM . 4
105. (")
- leo.
-us.
- 1I0.
- l Oll.
- 10.
- 9. 4 1 -23 .3
- 21 .
28. 4
29 . - 1. 7 I G) .
&8 .
19 . 4
20. Ii 10& . 4l.1 145. 62. 8
". ,,,.
ISJ. 2
83.9
". ",.
m.!
222. I OS. 6
0
-". I w., ". 107. 41. 7 145. 4 84. 4 223.
-13g.
- 130.
-
-
95.
go. - , I -no
- 8. -T.!. 2
30. - 1. 1
- I.
ZU. S
21 .
107 6 42. H& . 63 .'
,,,.
'83.
"".. "'.
85.
106.1
106.7
:::l
".
(;~.S 108. 42. 2 141 .
-121.
-" ,. OJ. '
". ,ss.
224.6 107.

...,. I
- 7. 6 -0.6
-112 . - 80.
-- 7. I - 21.7 32.
10.
71.
:!I.I.
21.7
10' . 42.8
".
141.2
148.
186.8
",. 107. 2
<:
"."'. ,,,.
10'l .' 64.4 187. 86. 1
- 100.
- P4 . -- "70 . -
6.
5.8
-21. 1
-21. ,
33.
n .s
+ 0 .11
1-
71 .0
n.
22.
22.2
liD.
111.
43 .3
43 .9
14'.
,so. 86. 7 226. '
107. 8
108.
CD
- 8J .
- 76 . -,. 1-"'·' I
-
".
-- ".
M.
- 'I. - 20.
34 .
lS. U
7J .
1J .'
2'1.S
23.
111.2
112 .
«.
H .'
150. 8
". na.
IlIS .6 87.
87 .2
217.
228.
108.3
,,,.
108.' ~
- 67 .
- 5S.
, I ".
-311.
". "
". -
-
1.
2. ~
-I~ 4
-19 .
;15 .11
15.
n.'
n .•
113. <S.
lSI.
152.
lob. !
~6 . 7 190. 4
87.8

....
88.
228. 2
"'. 109. 4
-.
(J)
- 4g.
- 40.
~ff: 4
-

I ". ,.
45. -
-
2.
\.
- 18 .g
-18. 3 37 . ~
' .8 , 5.2
16. ". 114.
11' .8
45 .•
".'''. 1 '5<.
152. 6
'53.
67.
61. 2
191 .
T92.
88. 3 210.
231 •
11 0.
IIO . ~
0
" I r "'". I 111 .
115. 67.S 192 .2 89. 2.11 . 8
- 39. =

.. ,. 38. lIb
. ,,,.
1 - 0. 4 -1 8. 3 .3 25. 11 6. 46. 1 154.4
o. 68. 193. 232 .
-
-
38. 2
3~ .
- 39.
- 18. 9
1
+ 1.
- 11 .8
- 17 .2 lY.2
116 .6
". ,,,. ,SS.
."."., 68. ' ,0<.
"'.
89.4
90. 111 . 7 :::l
...
i!l.S

-". I -". I ,.
-17 .
,. -
'- ,
llI .] 1.4 - 17.
- HU
. <D . u
.
79. 118 .
!L II m.
26.7
41 .:
H .S
<S.
156.2 195. S
~0. 6
!I .
233. 6
234.
112.
112.3 (J)
,.••• II!U '57. '9 .4 11&. ~I.I '35. lI2. '

,.,...
- 3G. - :i7 .8 - l(i . 1 42 . SO. 6 21. 4S .3 '58. 70. m.
=~h =i~: 2 1
~. 2 -16. 1~ . 8
,, I SI . 27.'l 120. 4 .~. 9 i O. 6 197 .6
' 1. 1
n.
235. 4
23G.
113.
lI3 .3
- 3<. I --:l(;.l
,'-' j, - 1$ . 6
- IS.
4J .
0<. ' .7
62.
S~.•

I "'. 120. 2
121. ".
'"
15'.
159 .8
UO.
7'-
71.1 ",.
1'8. 92.2
9'l. B
2J7.
23 7. 2
113 . '
11-4,
- lJ.
- J 2.8 I - 3S. ' .8
,.
-1 4 .1
-14 .
H .G
<S.
7.
7. ~
SJ.
84. lj'
28.'
28.'
so. 161. 71.7
n.
19'1. 4 93 .
"'. 114 . 4
- 32.
- 3T. I - lS.
- J!>.6
I
7. - 13 .9
-13 .3
<G.
46. 4 ,. '-'
m.
S( .2
85. ".
".,
!21.
123.8
.~O . '
51.
I 'L l.
,,,.
1&2. 72. 2
200.
21)1.
' 3. 3 ..
93. 9
".
2U.
2(0.
115.
It.s . ,
-30.
-
-
29 . 2 , -
34. 4 '
34. ,.
8.' -11.
-12 .8 ".<S . ,. I 8. '
8. '
".
56.
81.
3O. 125.
51.1
51. 7
,,,.
163.'
,,,.
72 .S
n.
201. 2
202. 94 .4
240.8
24t.
11&.
116.1

I ss. ff: I
d 12:;. 6 13 .3 203. lS. 242. 116. 7
- 29 . j -- 3U TO. -12. '2 48 . 2 H .8
- 28.
-- 27.
"., I -- ".
33. 3 10. 4
II .
-1 2.
-11 .7 ".
,'"'-. ,'- ".
'J. 4
TO.
".
3L. 1
al.7
I
1Z&.
127.
52 .2
52. 8
,,,.
165. 2
13 .'
". '0<.
208 .8
,os.
,,s. 6
".
2'~ . 6 ·
,,,.
243.
117.
117.2
3U
". -1 1. 1 W.G
53.' 8'J .6 32:
127. 4 74. 4 ~6 . 1 117.'

".".
128 . 1&7. 15. 20&. 96. 7 2(, . 4 118.
= ~to I = H:
2 I ~ .~ -11. 5 r. 8
".". I '''. ".53.' 32.2 ' 68. 75.6 206. 6
". N5. IIS.3

=!l .• 1-=;::;1
13. -10. 11 11. 1 32.S
- 26. '"~ ".
IS.
-10 .
-
-- ,.
9. 4
".
.\o:l .G
11.1
".
9l.4
92.
33.
33.3
129. 2
130.
131.
M.' m.
I/,IL' 76.
76.1
207.
208.
""'.,
't7. 2
91 .8
246.
U6 . 2
liS .'
m.

I I: ".". I " II I
- 21. - 30.
' 5. S
16. '.9 ".
ss.
".,
1~. 2
12. 8 "., 93. 2
33 .' 132.
1st •• 66.
170.
170. 6
!71.
76.1
TI.
77.'
",.
210.
OS.
'1. 3
91.'
247.
,<S.
2(' .
'1'.1 .4
'20.
120.'
- 22 .
, - l D. 17. - S.3
I 13. 3404 '53. 66. 1 Tn. 77. 8 210 . 2
". 24'.' '21 .

_" 1_"I " 1.. 1" 1.. 1" 1 I " 1" 1_"1_"I_M I_k1_"1_'I~
I
"'C
1

,
+,
H -,
I.1.00791
AlGDicH,...bu _ • +
in +
_ o.ic!aI ;"' St.o~
z
He
i ~ ·~2
" _.
CD
~
Sr.bol- II:£Y TO CHART
0
ILi
+ 11,
Be
+2
Atom;c Weicht _ II U 9
I II-I t- - 1:1<"'-' C006c~" lioG hB l.C..--+i!!,NI',.--.;I· -·t-4
-'1N.
10 t~ o
+. 'IF
0
_.
a.
(")
~
"tl
+>
-, "tl
,., 1'.0067 - 21 ' 5 .~ ,,- m
.~. I'·"~
12.01115
'~~j"l j2<UP

~ z
1: · 5 -3 2-4 :- 7 2-0 I K-L

r:'~ + 1:-; +2 . .
T.. ...""'. [1.",,,10
T ....... lioD
~l ...... '" I13 +3,'" +2115 0
Si ::IP
Na
:~ ~;~t~2
P.+.
Mg
C"",p '
AI

~~n 2~:!, -:.~~ l!:! I K_lrW -

C"
CD
X
U1

--
20 +221 +l'U +2in +224 +2126 +2 28 H :n +2 %8 +2 H ·1-231 32 +233 +3 :U +t 3.5 +13e 0 - --
I ~ Ca :~ Se ~ Br !~Kr
+1 131 +J

-'-~2 ~':I"'1
+1 Sc Ti !!V :::! Cr t: Mn! !Fe +leo +lNi +3 eu HZn Ga Ge H As
.1 02 to.OII 4.I.S6 4; 110 I +~
1:./1 ' 42 ) 1 996 S. tlliO
+7 ~ ~ 7 Ir-:.1W2
•• , 58" .)46 6537 972 :-1.l9 ~U2 1 8 lA. M 79 .11CH !D."
0
&-&-1 •·... 2 _,..1 0-.2 +11-2 ' ' :1.-'1--''3+2 -ll-I t -2 ' 1-15-2 ..... 6-2 · 1--11-' .... ' S-l -"I H "'-II-.s -.. I..... -1·1'-7 • ..1.... 1-t.-Iot.S
, +\38 +,Iil+i fo +tfl +3~43 +6 " +l:u--+l U +2 4' +1 '8 +2 41 +3s;--+2~ .5t + 453 +I~
Rb Sr Y Zr Nb H Mo Tc t; Ru Rh PdHAg Cd In SnHSb :!:~ Te ~ I +~Xe
-, :::r
8$.47 ro.U a~ t:tos 1I~t4
11\ :2 97J 1~O
I 07 U!21iO$ 106 6 107.168 112tO 11 1 82 111.69 1~I .a 127.110 I:K.~' 131.30 CD
-1 .. 1-1 - 1..1-2 ~~ -1S- 12' 1 !~ 1~ -I.- I ~ I ~~ - Ia-I I- I ~ · II- I N ~~ ~~~~r. M-N-O
+156 +251" H 12 +4 13 H ,. H 1S H 16 Tl11 H 1S +2i"t+lst + 1 ~8~ + 283
ISS
:Cs Ba La HI Ta W
+1," +285
Re :::;Os +'Ir HPt
81 0
+ 6 Au +lHg +t T1 +l Pb HBi H pO +'At Rn m
132.!l0~ 117.34 IlS." 178. 49 180 961 18l8.S ISIt~ 1'0102 1922 IUOII IHRl!l 2lO~ 20437 207.19 208 .9110 (210) 2(9 ) (2"22)
~~'IH·2 -32-11).2~~~~~~:: ~~~~~_ ~~ -31-11-7 ~I~ CD
IS1
'Fr
+ 1 88 +2 S' ··
Ra Ae H 3

-
m~ll- I ~~~'2 ~~~~'2 -<>- p.. CD
:::l
,-
' UtoI~&tIid..
SIJ +3 SI
CeHPr
140.12 140.1107
+~IO
Nd
+3 II +~6~ H
144 .2'
I63 +214 +lJ6$
Sm +3 Eu +3 Gd Tb
( U ~) 1~. lS J51 .Gtl \3;.25 151.92.
HI" Dy
I6UO
+ 3 ..,
Ho
+1 68 +31'9 +3!'" + 211 +
Er
1&4.$30
Tm Yb +l Lu
157.:6 168.1134 173.04 174.97
(J)
-1 ~~2 -"lO-~' -n-I-' -Z;g-2 -25-~2 ·32·~2
I~
-:2+2 -23-1-' -3:-1-' -2HI-Z -2.J-1-2 -n-1-2 -30-1·2 · 31 -' -2
-- --- - - - - - - - - --- - --------
90 H l l + '2 +3 93 +3 9. +3 9S +' " +3'7 +398 +2" 100
-----
101 IOZ
- -103
!Tb Pa +tU .t;Np !; Pu t; Am!; Cm Bk +tel Es Fm Md Lw
" -'tl inidoo

"," ~" .,~-, -, ,"." ~,.-, I·" ~" .,- -

:
232.031 (231) %3' .03
+
(231)
+ +1 +
: 44 ) (:4J1 (%t7 ) 1~47 (UI I (2511 ( ~51) (l.'.6) (25<1 )
L--
..,.,.
en :-:~"'''''' i. !)&Ital" ' - .,~ .... 6WffI!>oro of ......1 ......,. ""'lOP< or u.., . I<.,.nl .
'"
~
 BASIC PROPERTIES OF AIR AND WATER 6 3 3

MASS DEN SITY OF WATER

5°C 10°C 15°C 20°C 25°C 30°C

O°C
1000.0 999.7 999. 1 998.2 997. 1 995.7
APPENDIX 6 p (k g/m' ) 999.9

KI NEMATIC V ISCOS IT Y OF WATER ( v): v == ~p

Basic Properties of Air and Water
5°C 10°C 15°C 20°C 25°C 30°C
O°C
COMPONENTS OF ATM OSPHERIC AIR (APPROX IM AT E)
(centistokes)
1.519 1.3 1 1.1 46 1.0 11 0.898 0.803
Content (%) Co ntent (ppm) cm'ls (X 10-') 1.792
1.0 11 0.898 0.803
ConstilUcnt by Volume by Volume m' ls ( X 10- 6) 1.792 1.519 1.3 1 1.1 46
1.633 1.408 1.232 1.087 0.965 0.864
ft' /s (X 10- ') 1.926
N, 78.08
0, 20.95
CO,
NOTE: To convert cenlistokcs to ft /s. multiply by 1.075 X lO ~ J.
0.33 2
Ar 18. 18
He 5.24 ABSOLUT E V I SCOS ITY OF WATER (f.1)
CH 4 2.0
H, 0.5 O°C 5°C 10°C 15°C 20°C 25°C 30°C
NO, 0.5
(centipoises)
1.792 1.519 1.310 1.145 1.009 0.895 0.800
MASS DENS IT Y AND V ISCOC ITY OF AIR glem . s ( X 10- ' )
1.51 9 1.310 1.1 45 1.009 0.895 0.800
kglm . s (X 10- ' ) 1.792
3.74 3.17 2.73 2.39 2.11 1.87 1.67
Temperature Density Ki nematic lb · sift' (X 10- ')
0.913 0.8 16
roC) (kg/m') Viscosity (m'ls ) newton ' slm' (X 10- ') 1.83 1.55 1.336 1.17 1.009

20 1.392 1. 152 X 10- 5

- 10 1.340 1.245 X 10 - 5 NOTES : To convert ccntipoiscs to lb · s1f( 2, multiply by 2.088 X 1O- ~. Newton =
o 1.293 1.319 X 10- 5
kg · Ill/5 2 = IO- ~ dyne.
10 1.247 1.412 X 10- 5
20 1.201 1.486 X 10 - '
30 1.1 60 1.607 X 10 - 5
40 1. 129 1.691 X 10 - 5

NOTE: For conversion to n2/s. use 1 mZ/s :: 10.764 n2/s; for conve rsion 10 Ibm J • use
I kg/Ill' ~ 0.06243 Iblft'

Atmospheric Pressure
I atm = 10. 1325 NicOl' = 1.03 32 kgf/c m' = 14.69 Ibfli n. (psi)
I atm = 760 mOl mercurylcm' @ 0° C = 406.8 in . watcr/i n.'
Notes:
Bar (b) = 10 Nlcm' = 1.0 197 kgf/cm'
Torr is the pressure of I mm of mercury per square centimeter at latitude 45 , sea
level. I torr = 1.333 X 10- ' Nlcm' = 1.3596 X 10- 3 kgf/cm'
632
634 BASIC PROPERT IES OF AIR AND WATER

TAIlLE A6·j Density and Viscosity of Water

Density Absolute Kinemati c
Temperature p, 'Y (glem'), viscosi ty. j.L viscosit y. v
ee) also & (eP)" (cS t)" APPENDIX 7
0 0.99987 1.792 1 1.7923

2
I 0.99993
0.99997
1.7320
1.6740
1.732 1
1.6741
Gas Law
3 0.99999 1.6 193 1.6 193
4 1.00000 1.5676 1.5676
5 0.99999 1.5 188 1.5188
6 0.99997 1.4726 1.4726
7 0.99993 1.4288 1.4288 Generali zed gas law:
8 0.99988 1.3872 1.3874
9 0.99981 1.3476 1.3479 PV = IIRT
10 0.99973 1.3097 1.3 10 1
II 0.99963 1.2735 1.2740 where
12 0.99952 1.2390 1. 2396
IJ 0.99940 1.206 1 1.2068 I' = atmospheric pressure (I' = I as standard pressure)
14 0.99927 1.1 748 1756 V = volume of a gas in liter
15 0.99913 1.1447 1.1 457 " = number of moles of gas
16 0.99897 1.11 56 1.11 68 R = molar gas constant
17 0.99880 1.0876 1.0888 T = absolute temperature (K) (O°C = 273.15K)
18 0.99862 1.0603 1.06 18
19 0.99843 1.0340 1.0356 Under standard temperature (20°C) and pressure (I atm), R and II arc given as follows:
20 0.99823 1.0087 1.0105
21 0.99802 0.9843 0.9863 R = PV = I atm X 22.4 L = 0.0821 K L I = 0.0821 UK· mol
22 0.99780 0.9608 0.9629 liT I mol X 273.15 K ' mo
23 0.99757 0.9380 0.9403
24 0.99733 0.9 16 1 0.9186
II = ..IL
25 0.99707 0.8949 0.8975 M
26 0.99681 0.8746 0.8774
27 0.99654 0.855 1 0.8581 where g is the mass of the gas (g) and M is the molecular weight of the gas, Under
28 0. 99626 0.8363 0.8394 standard temperature and pressure (STP), the density, volume, and molecular weight
29 0.99597 0.8181 0.8214 of a gas are given as follows:
30 0.99568 0.8004 0.8039
31 0.995 37 0.7834 0.7870 Density (gIL) of gas:
32 0.99505 0.7670 0.7708
33 0.99473 0.7511 0.7551 d = 1!.. = MP ord = ~
34 0.99440 0.7357 0.7398 V RT 24 ,053
35 0.99406 0.7208 0.7251
36 0.99371 0.7064
Volume (L) of gas:
0.7 109
37 0.99336 0.6925 0.6971
38 0.99299 0.6791 0.6839
39 0.99262 0.6661 0.6711
Molec ular weight (g/mol) of gas:
" I centipo ise = I 0- 2 (gram Imss)/(c m)(s).
1) 1 centi stoke = 10 - 2 cm2/s .
M = gRT or M = g X 24.053
PV V
635
 APPENDIX B APPENDIX 9

Ozone Concentration Conversion Tracer Test

TRACER TEST STUDY

Tracer tests a f C commo nl y performed by injecting a slug dose of dye, rad ioactive sub-
stance, or a salt solut io n into the inlet o f a basin or tank , the n measuring the co nce n-
trati on of the tracer in the crlluent at va ri ous time intervals. The diluent monitoring
should be carried on unt il substant ia ll y all the tracer has passed th rough the bas in . T he
type of tracer that can be used for water treatment and water suppl y systems is lim-
ited to lithium chl oride and Iluoride because of public health safe ty co ncern s.
C hl or id es such as sodiulll chl oride , whi ch is read il y avail able as we ll as cost-
effec ti ve , is not an e ffec ti ve tracer because of its hi g h spec ilic grav ity as a so luti o n,
and it prod uces a densit y !l ow in the tank or bas in . O fte n, lithium chl oride is used
as a tracer since it is safe to hum an health at tracer dosage, sa fe to handl e , and easy
to put in to so luti o n d ue to its hi gh so lubilit y (63 g pe rl OO miL in near freezin g wa-
ter, ex othermi c reacti o n). Howeve r, lithium chl oride is rather ex pe nsive ($3 per
po und ). Li thium chl oride is ava il abl e from Li thco, a subsidiary of FM C Corp. ,
Gasto ni a, North Carolina (pho ne : 70 4-868-5350). Commo nl y ava il able Ilu oride
~ ~ ~ ~~
... .......... N
~ ~ M;V;
,..jNNNN
compounds are sodium tluoride. sod ium silicollu oride, and hyd ro!luos iJic ic acid .
The lirst two arc ava il able in dry form, but the latter is ava il able o nl y as a liquid .
T he dry co mpo und s have very low solubilit y in water (4 and 0.76% , res pec ti ve ly,
al 25°C) and dust fro m lhese compound s ca n be hazardous to hea lth. Hyd ro nu osilicic
acid is avail able as a liquid (23 % so luti o n) and has inlinite so lubilit y, but is ex-
c
• ~g~g~ ~:~n;:~~ ~ ~ ~~p~
tremely ac idic and there fore creates a sa fe ha ndling proble m. Alth ough the cost o f
.E ~t oco d c:i 0"":"":"':"":
__ _ N N
these !luoride chemical s is about 10 to 30% of that for li thiulll chl or ide , the cost
.!!' 0
•
;t beneli! is o ften o n-set by the safety concern s a nd tro ubl eso me preparati o n of the
E tracer soluti o n. O ne impo rt ant charac teri sti c of an e l1'ective tracer is that it cann ot
..•
~
~ ~~:st;J ~~ ~ ~:2 Ml7lo.t'I_ ....
be re moved or co nsumed by the trea tme nt process. Fluo ride ca n be removed by a
hi g h dosage of alum fl occ ul ati on and 10 a signi li ca nt deg ree by the lime so ftenin g
~l ci d c:id ci 0 "':":"":"':
CO "' I7I _ N
.... N N Mro)
process. Lithium , however, is muc h more stabl e th an Ilu orid e.

TRACER TEST METHODS

T he two basic methods o f tracer additi o n arc the step-dose method and th e slu g-
dose method. In the step-dose me th od the tracer is dosed at a co nstant rate until the
637
636