Professional Documents
Culture Documents
P O W E R P L A N T
W A T E R
CHEMISTRY
A PRACTICAL GUIDE
P O W E R P L A N T
A PRACTICAL GUIDE
by
Brad Buecker
Copyright© 2006 by
PennWell Corporation
1421 South Sheridan Road
Tulsa, Oklahoma 74112-6600 USA
800.752.9764
+1.918.831.9421
sales@pennwell.com
www.pennwellbooks.com
www.pennwell.com
Marketing Manager: Julie Simmons
National Account Executive: Barbara McGee
Buecker, Brad.
Power plant water chemistry : a practical guide / Brad Buecker.
p. cm.
Includes bibliographical references and index.
ISBN 0-87814-619-9
ISBN 978-0-87814-619-2
1. Steam power plants. 2. Chemical engineering I. Title
TJ403.B82 1997
621.31’2132--dc21 97-35403
CIP
All rights reserved. No part of this book may be reproduced, stored in a retrieval system,
or transcribed in any form or by any means, electronic or mechanical, including photo-
copying and recording, without the prior written permission of the publisher.
4 5 6 7 8 12 11 10 09 08
Dedication
To Nancy and Alyssa, each of whose creativity easily exceeds my own,
and to my parents whose own lives have shown me the value
and rewards of hard work.
Table of
Contents
Figures and Tables
Acknowledgments xi
Preface xii
4 STEAM CHEMISTRY 83
Introduction 83
Primary Carryover Products 84
Copper 84
Sodium Hydroxide 85
Table of Contents ix
x Table of Contents
Acknowledgements xi
7 SAMPLING 205
Introduction 205
The Need for Sampling 206
Sample Point Selection 206
Makeup System Effluent 206
Condensate Storage Tank Effluent 211
Condensate Pump Discharge 211
Condensate Polisher Effluent 212
Deaerator Inlet 212
Deaerator Outlet/Boiler Feed Pump Suction 212
xii Preface
Bibliography 229
Index 233
Acknowledgments xiii
Acknowledgments
No book of any substance can be written without assistance from others. I
would especially like to thank Ray Post of BetzDearborn for his insightful com-
ments on cooling water chemistry. They immeasurably helped me with the cool-
ing water chapter. I would also like to thank Nissen Cahan and the people at
Purolite for allowing me to use data about their ion exchange resins. The data
was enormously useful. Similar thanks go to Dr. Barry Dooley and his staff at the
Electric Power Research Institute for allowing me to use illustrations from some
of their technical reports.
Others who provided very helpful data, illustrations, photos, or advice
include Terry Dwyer at the Marley Cooling Tower Company, Don Walter from
Osmonics, Tom Svoboda of the Sentry Equipment Corporation, Lee Machemer
from Jonas & Consultants, Phil Di Vietro from ASME, Jim King from DB Riley,
and several people from U.S. Filter.
I would also like to thank COADE Engineering Software of Houston, Texas
for allowing me to make use of their unit conversion program in preparing this
book. I have used this program many times over the last several years. I would
also like to thank John Meinders of the Kansas City, Kansas Board of Public
Utilities for allowing my colleague John Wofford and I to take photos at his plant.
A special thanks also to John Wofford for being able to answer virtually every
question I had about boilers while preparing this book.
Last, but not least, I would like to acknowledge my friends and former co-
workers at City Water, Light & Power in Springfield, Illinois, and in particular
Tom Bee, Ellis Loper, Ed Riordan, Dave Arnold, and the late Charles Hartman.
During my 12-year
career at the utility, I was given the opportunity to work on a wide variety
of projects related to analytical chemistry, water and wastewater treatment, flue
gas desulfurization, and plant engineering. Without this experience I could never
have even attempted to write this book.
xiv Preface
Preface
When PennWell Publishing asked me to write this book, I was unsure at first
how to arrange the contents. Two concepts emerged as I began to gather mater-
ial and put my thoughts in writing. One was to provide as much practical infor-
mation as possible regarding the core areas of steam generation chemistry. It is
my hope that utility and industrial personnel, and especially those who may be
somewhat new to this type of work, can open the book and find solid guidelines
and examples to follow. I included a number of case histories from my own expe-
riences and those of several colleagues to illustrate common, and a few not-so-
common, difficulties faced by steam generating personnel.
Additionally, I tried to include new information regarding steam generation
chemistry, makeup water techniques, and boiler water chemistry to give the read-
er an overview of trends within these areas. For instance, coordinated/congruent
phosphate boiler water treatment is losing ground to other programs; reverse
osmosis has grabbed a big share of the makeup water treatment market, and
steam chemistry guidelines keep tightening as researchers learn more about the
effects of steam contaminants on turbine components. Steam generation person-
nel need to stay abreast of these developments.
Much of the data in this book represents the latest ideas regarding steam
generation chemistry. However, this book makes no guarantees regarding unit
performance, and I encourage readers to use the data as a guideline and not as
an absolute for their system(s). Even sister units exhibit different operating char-
acteristics, so treatment and operating chemistry must, to some extent, be spe-
cific to each unit.
I eagerly invite readers of this book to send me suggestions, comments, and
even their own case histories for inclusion in a possible future edition. I firmly
believe that the best teacher is experience, and I would be very pleased to hear
from those of you who deal with steam generation chemistry on a daily basis.
Chapter 1
Introduction to
Steam Generation
Water Chemistry
Systems
● Makeup system
● Condenser cooling
peratures can reach 600˚F or higher and steam temperatures 1050˚F, highly puri-
fied feedwater, dosed with carefully controlled treatment chemicals, is required
if the boiler is to operate properly. Table 1–1 illustrates some of the effects that
contaminants have on boiler water systems.
The vastly different conditions between water systems, and the complexity
of a steam generating system, make the chemist’s job very lively. This book pro-
vides practical examples of water chemistry issues and problems for steam gen-
erating systems, and illustrates techniques and methods to control chemistry. It
also provides details on many of the latest trends, findings, and developments in
the areas of boiler water chemistry, steam sampling, and makeup water produc-
tion. Utility chemists and researchers have made many discoveries and improve-
ments to steam generation chemistry within the last decade. A number of these
have challenged traditional ideas. Some of the developments that industrial or
utility steam generating personnel should be aware of include:
● Boiler water treatment has undergone many changes. For years, coordi-
nated or congruent phosphate treatment was popular for many boilers.
These programs have been found to have some serious deficiencies and
are being replaced with alternative phosphate programs.
● Oxygenated treatment (OT), where oxygen is deliberately injected into the
boiler feedwater, is becoming very popular in once-through units in the
United States. OT, which was developed in Europe, has been shown to
greatly reduce iron transport from the feedwater system to the boiler.
● Ion exchange is no longer the only reliable method for producing high-puri-
ty water. Other techniques such as reverse osmosis (RO) and electrodialysis
are available for this process. Often, a combination of these techniques, such
as RO plus ion exchange, may be the most economical arrangement.
● Diverse opinions exist regarding chemical oxygen scavenging in boiler
feedwater systems. The reducing environment produced by oxygen scav-
engers is known to influence flow-accelerated corrosion (FAC), in which
the pipe wall gradually erodes. Several catastrophic failures, some of
which have caused fatalities, have occurred in recent years due to FAC.
Yet, the same reducing environment greatly lowers copper dissolution and
transport in those systems that have copper-alloy feedwater heaters.
Hydrazine, the most common and effective oxygen scavenger for many
years, is now listed as a hazardous chemical. However, alternative organ-
ic scavengers (and pH-controlling amines) can break down in boiler sys-
tems to produce organic acids and carbon dioxide, which in turn can
cause corrosion of afterboiler components including turbine blades.
● Combined-cycle or cogeneration systems with heat recovery steam gener-
ators (HRSGs) have become very popular. HRSGs, however, are often
designed with two or three steam generating circuits, all at different pres-
sures. Chemical treatment requirements for the various circuits are also
different and may be dependent not only upon the pressure of the circuits,
but also upon the configuration of the HRSG.
These are but some of the issues that I have addressed in this book. Research
still continues on these and many other items, and our knowledge of steam gen-
eration chemistry will only improve in the future.
Table 1-1
Common Steam Generating System Contaminants
Compound Effect on Plant Equipment and Operation
Oxygen Oxygen is often the principal corrodent in water systems. It causes pitting and failures
of pipes and heat exchangers. Oxygen corrosion in boiler systems generates particulates
that travel to the boiler where they precipitate and cause further problems.
Calcium Calcium can combine with a number of anions to form deposits and scales. In cool-
ing water systems the most common deposits include calcium carbonate, calcium
phosphate, and calcium sulfate. These scales retard heat transfer in condensers and
other heat exchangers, and may cause underdeposit corrosion. Calcium scale is even
more problematic in boilers, as the high temperatures greatly accelerate deposition and
corrosion mechanisms.
Magnesium Magnesium will react with carbonates and silicates to form compounds of low solubility.
Magnesium salts that leak into a boiler can react at high temperatures with water to
produce acid. The corrosiveness of acidic solutions is greatly increased at the high
temperatures found in boilers.
Silica Silica combines with a wide variety of elements to produce silicates, or it may form
deposits on its own. Silicates form tenacious deposits in cooling water systems, boiler
tubes, and on turbine blades. The scales are inert to most chemical cleaning solutions
with the exception of hydrofluoric acid. This is an extremely dangerous compound,
and makes prevention of silica deposition even more important.
Organics Organics are usually found in surface waters and are the result of decaying vegetation
or farm runoff. Organics break down in the boiler to form organic acids. The resultant
low pH can be quite deleterious. Organic acids and carbon dioxide produced by
decomposition can carry over to steam turbines and corrode the blades. Organics may
also be found in the condensate return at industrial and cogeneration facilities. These
organics are usually much shorter chained than surface water organics and may require
different treatments.
Suspended
solids Suspended solids, which are also generally found in surface waters, will foul makeup
treatment equipment including reverse osmosis units and ion exchangers. They will
also form deposits in cooling towers and cooling water heat exchangers, a process that
is exacerbated by the presence of microbiological organisms.
Microbes Microbiological fouling is principally troublesome exchanger tubes and cooling tower film
fill. The slime produced by microbiological organisms will trap silt and suspended solids,
further aggravating the situation. Microbes are a leading cause of under-deposit corrosion.
Chapter 2
Condensate/
Feedwater
Chemistry
Introduction
The preboiler system of a typical utility steam generating unit (Fig. 2–1)
includes a steam surface condenser, several closed tube-in-shell feedwater
heaters, a deaerating feedwater heater, and sometimes an economizer. For indus-
trial systems, feedwater heaters, with the exception of the deaerator, are often
omitted unless the steam drives a turbine. The preboiler circuit condenses the
turbine exhaust steam and prepares the condensate for return to the boiler. The
condensation process significantly improves the efficiency of a unit, as is outlined
in greater detail in Supplement 2–1 at the end of this chapter.
The potential for contaminant introduction to a steam generating plant is
greatest in the preboiler system, especially at the condenser or via condensate
return from an industrial process. This makes chemistry control and monitoring
of condensate and feedwater extremely important.
Figure 2-1
>
MAIN STEAM TO TURBINE
SUPERHEATER ECONOMIZER
TURBINE
BOILER
CONDENSER
VENT
>
>
>
STEAM >
DEAERATOR
Preboiler/Boiler/Afterboiler Schematic
Condensate/Feedwater System
Construction Materials
A variety of materials have been used for feedwater heater and condenser
tubes. The most common include 90-10 and 70-30 copper nickel, 304 stainless
steel, carbon steel, Admiralty metal (70% copper and 29% to 30% zinc depend-
ing on whether 1% arsenic or tin has been added for increased corrosion resis-
tance), and Monel (70% nickel, 30% copper). In previous years, copper alloys
were widely selected due to the excellent heat transfer properties of these mate-
rials. Recently, as copper corrosion and its effects on downstream components
have become better recognized, the use of copper alloys has declined. Stainless
steel is becoming favored for feedwater heater tubes; and stainless steel, titanium,
or even the duplex alloys are being used for new condensers.
The behavior of these materials in solution significantly influences conden-
sate/feedwater chemistry guidelines and treatment methods. Regardless of the
material, one of the toughest challenges for a plant chemist is control of dissolved
oxygen.
Condensate/Feedwater Chemistry 7
Oxygen that enters the condensate system will oxidize the protective layers
on iron and copper to films that are much less stable. In the case of mild steel,
oxygen converts magnetite to ferric oxide (Fe2O3), which is not protective.
Other complexing agents include chloride and sulfate, however, these are
usually at extremely low concentrations and do not participate in copper corro-
sion mechanisms. Oxygen/ammonia attack is much more common, primarily
because ammonia or organic amines are the preferential choice for feedwater pH
conditioning.
Carbon dioxide carries over with the steam and then redissolves in conden-
sate return lines. Inorganic chemists debate over whether CO2 actually forms an
acid or whether it exists as a discrete hydrated molecule, but the net effect of this
dissolution is shown in the following equation:
The absorption of carbon dioxide into the condensate drives the pH down-
ward. The acid solution directly attacks the pipe walls via the formation of fer-
rous bicarbonate:
Any oxygen present in the feedwater will oxidize the ferrous bicarbonate to
ferric oxide and liberate carbon dioxide to repeat the process.
Oxygen, carbon dioxide, and ammonia are the most common corrodents in
steam generating systems. Methods to remove or treat these compounds are out-
lined in the following sections.
Condensate/Feedwater Chemistry 9
Figure 2-2
Figure 2-3
Condensate Dissolved Oxygen Concentrations vs. Air Inleakage Copyright © 1986. Electric Power Research
Institute. EPRI CS-4629. Interim Consensus Guidelines on Fossil Plant Cycle Chemistry. Reprinted with permission.
Condensate/Feedwater Chemistry 11
Figure 2-4
Change in Dissolved Oxygen Concentration with Unit Load Copyright © 1986. Electric Power Research
Institute. EPRI CS-4629. Interim Consensus Guidelines on Fossil Plant Cycle Chemistry. Reprinted with permission.
shroud to pull gases into the compartment, upon which they are extracted and
vented to atmosphere. The shroud is situated around a block of tubes to con-
dense steam that may enter.
The Heat Exchange Institute (HEI) has established a design guideline of 7
ppb dissolved oxygen in condensate from a properly operating condenser.
However, excess air inleakage can overload the air removal system, which, as
Figure 2–3 shows, dramatically increases the dissolved oxygen concentration.
Several techniques are available to determine if excess air inleakage is a prob-
lem. Simple flow monitoring, wherein the extraction line from the air removal sec-
tion is equipped with an air flow rotameter, is one such method. A guideline estab-
lished by the HEI suggests that under normal conditions, air removal through the
extraction system should average about 1 standard cubic foot per minute (SCFM)
per 100 MW of capacity. Significantly higher values than this may indicate prob-
lems, although each unit should be evaluated on a case-by-case basis. It is more
important to establish baseline conditions, which can then be used for comparison
during times of suspected upsets.
Plant chemists can often detect excess inleakage by analyses of dissolved
oxygen (DO) levels in the condensate pump discharge. However, this data must
Figure 2-5
External View of a Deaerator. Photo taken with permission of the Kansas City, Kansas, Board of Public Utilities,
Quindaro Power Station.
Condensate/Feedwater Chemistry 13
Oxygen Scavengers
One of the first practical chemicals to be used was sodium sulfite (Na2SO3).
Sodium sulfite reacts with oxygen to produce sodium sulfate:
Sodium sulfite has a molecular weight almost four times higher than that of
oxygen and reacts in a 2 to 1 molar ratio, so theoretically 8 parts per million
(ppm) of Na2O3 are needed to remove 1 ppm of oxygen. However, sulfite resid-
uals are often maintained at 30 ppm or higher to provide adequate protection.
Figure 2-6a
Reaction Time of Hydrazine with Oxygen Copyright © 1986. Electric Power Research Institute. EPRI CS-
4629. Interim Consensus Guidelines on Fossil Plant Cycle Chemistry. Reprinted with permission.
The primary advantages of sodium sulfite are that it is a common and easi-
ly obtained chemical, is nontoxic, and can be used for water treatment where
steam is extracted for food processing or other FDA-regulated applications.
Sodium sulfite is primarily used in low-pressure industrial boilers (<600 psig)
because it adds too many dissolved solids to high-pressure boiler water. Also, in
boilers that operate above 900 psig, sodium sulfite will thermally decompose to
produce hydrogen sulfide (H2S) and sulfur dioxide (SO2), both of which are
quite corrosive.
For utility and industrial boilers that operate at pressures above 900 psig,
alternative chemicals are more suitable for oxygen scavenging. The workhorse for
many years has been hydrazine (N2H4), which reacts with oxygen as follows:
Condensate/Feedwater Chemistry 15
Figure 2-6b
Effect of pH on the Hydrazine-Oxygen Reaction for a Reaction Time of 0.13 Minutes at 300˚F in
Carbon-Steel Tubing Copyright © 1986. Electric Power Research Institute. EPRI CS-4629. Interim Consensus
Guidelines on Fossil Plant Cycle Chemistry. Reprinted with permission.
Condensate/Feedwater Chemistry 17
Due to its organic structure, hydroquinone and its oxidized products can
break down to a variety of organic acids and carbon dioxide in a boiler system.
Methyl ethyl ketoxime is an organic scavenger that reacts with oxygen to
produce methyl ethyl ketone (MEK).
Figure 2-10
External View of a High-Pressure Feedwater Heater. Photo taken with permission of the Kansas City,
Kansas, Board of Public Utilities, Quindaro Power Station.
day tank, is diluted with water, and is then pumped to the system.
The portable feed concept is gaining great acceptance. The water treatment
firm is responsible for delivering full tanks to the plant site and hauling away the
empty vessels. Plant personnel do not have to deal with drums and related haz-
ardous waste disposal requirements. The capacity of portable containers is typi-
cally within a range of 200 to 400 gallons. A method to minimize handling is to
stack one portable feed container on top of a primary feed container. The top
container is plumbed such that it drains into the bottom vessel. When the top
container empties, it is replaced with a full vessel. This arrangement provides two
distinct advantages. First, the bottom container can be permanently piped to the
metering pump or connected to a day tank. Second, the top container drains
completely without compromising the performance of the system. No residual
chemical remains in the vessel.
Condensate/Feedwater Chemistry 19
Flow-Accelerated Corrosion
A relatively new phenomenon, at least in terms of it being communicated to
steam generation personnel, is flow-accelerated corrosion (FAC). FAC has raised
additional questions about oxygen scavenging. FAC is produced by a combina-
tion of chemical and mechanical factors, and occurs in high turbulence regions
of the feedwater system with strongly reducing environments. These locations
include elbows in feedwater lines, economizer inlet headers, and high-pressure
heater drains. In simple terms, FAC develops when oxygen is completely
removed by the scavenger. This inhibits formation of the protective magnetite
layer on the pipe surface. In high turbulence zones, the flow gradually
erodes/corrodes the base metal until the wall becomes too thin to withstand the
fluid pressure. Several catastrophic failures have occurred within the past few
years and in some cases have caused fatalities. This is quite serious, and it is my
hope that this book will alert readers to this potential problem. The Electric
Power Research Institute (EPRI) is making a strong effort to inform their mem-
bers and others about FAC. Table 2-2
Several methods have
Relative Basicity
been proposed to combat Kb x 10
FAC in scavenger-treated
Compound 72˚F 298˚F 338˚F
feedwater systems. One Ammonia 20.6 6.9 4.6
method is to reduce the oxy- Cyclohexylamine 489 61 32
gen-scavenger feed, such that Diethylaminoethanol 68 11.3 9.2
a 1 to 2 ppb oxygen residual Morpholine 3.4 4.9 3.8
remains in the feedwater. This Basicities of Various Feedwater pH Conditioners.
helps preserve the magnetite Source: Betz Handbook of Industrial Water. Conditioning, Ninth
Edition. BetzDearborn, Inc., Horsham, PA.
layer without subjecting the
system to corrosive levels of oxygen. To inhibit FAC of high-pressure heater
drains, EPRI recommends that high-pressure feedwater heater (Fig. 2–10) vents
be closed. This, however, has caused problems at some utilities and must be eval-
uated on a case-by-case basis.
The balance between oxygen scavenging and prevention of FAC presents
one of the greatest difficulties in mixed-metallurgy feedwater systems, where
copper-alloy corrosion is greatly dependent on the oxidation reduction potential
(ORP) of the solution. This particular aspect of feedwater chemistry is discussed
below.
ular on the high-pressure turbine blades. Reports abound of electric utility units
whose output decreased over an operating period because of copper deposition
on the turbine blades. Copper corrosion and deposition also cause other diffi-
culties. For example, copper deposition in boiler tubes can greatly increase the
difficulty of a boiler chemical clean. The tube deposits may become layered with
copper and magnetite, which can require a multistaged chemical cleaning
process for complete removal. Copper inclusion in deposits can also increase the
potential for under-deposit corrosion.
As Equation 2.4 illustrated, copper corrosion is initiated when copper metal
or the protective cuprous oxide film is oxidized to cupric oxide. For this to hap-
pen, the oxidation potential (as measured in millivolts or volts) of the solution
must be greater than zero. Continual feed of the oxygen scavenger will maintain
a reducing environment and keep the ORP well below zero. However, a reducing
environment is the primary triggering mechanism for flow-accelerated corrosion.
Copper corrosion is exacerbated in cycling units. Several mechanisms may
be at work here. Mechanical stresses imposed during frequent load changes or
startups and shutdowns will fracture the cuprous and cupric oxide films on
heater tubes and cause corrosion and spalling. Oxygen corrosion of unprotected
feedwater heaters during an outage has caused severe corrosion and exfoliation
of heater tubes. In fact, this problem can be so serious that the heater steam-side
should be blanketed with steam or nitrogen during extended outages. More
details about off-line corrosion protection are presented in chapter 3.
The upshot of these difficulties with copper corrosion is that many experts
now recommend that feedwater heater tubes be fabricated from stainless steel, or
that existing copper-alloy heaters be replaced. Sometimes the latter recommen-
dation may not be economical, at least in the short term. Good water treatment
practices are a fallback for such situations.
Condensate/Feedwater Chemistry 21
laminoethanol and ammonia lies in between. Thus, the chemist has a number of
choices for finding a product that maintains condensate/feedwater pH within rec-
ommended guidelines.
Amines will decompose to produce ammonia in feedwater. Whether the
ammonia is generated by direct ammonia feed or amine decomposition, a gener-
al rule of thumb suggests that ammonia concentrations be limited to 0.5 ppm in
systems with copper-alloy tubed heaters. Even this level may be too high where
frequent inleakage of air is a problem.
In high-pressure utility boilers, where the steam is quite pure, decomposi-
tion of amines can potentially introduce unwanted organic acids and CO2 to the
turbine. For this reason, ammonia is widely recommended as the best pH-con-
ditioning chemical. The situation may be vastly different in an industrial boiler
where bicarbonate decomposition causes heavy carryover of CO2 to condensate
lines. Amines can be a good product for neutralization of the carbon dioxide. A
particular aspect of importance is the amine distribution ratio. This is the per-
centage of amine that carries over with the steam versus that which remains in
the boiler water. The distribution ratio varies with the product and with the pres-
sure of the boiler (Fig. 2–11). Selection of the neutralizing chemical based on dis-
tribution ratio can be very important. If protection of afterboiler condensate lines
is required, an amine with a high steam-to-liquid distribution ratio is best.
Where corrosion prevention in the boiler is more critical, a more appropriate
neutralizing amine is one whose distribution ratio allows most of it to remain in
the boiler water. Often, a blend of two amines is used to provide universal pro-
tection.
This section would not be complete without a brief discussion of filming
amines. These products are long chain amines (18 to 20 carbon atoms) that do
not neutralize acids but rather form a film on the pipe wall. The amines provide
a physical barrier between the material surface and the process liquid. The amine
group attaches to the metal surface while the nonpolar organic portion acts as the
film. Octadecylamine [CH3(CH2)16CH2NH2] is the most popular of the filming
amines. Filming amines are primarily used for protection of condensate return
lines at industrial facilities. They must be employed with caution, however.
Overfeed can introduce excess organics to the boiler water. Additionally, filming
amines do not work well in high velocity condensate return lines, because they
can be easily stripped from the metal surface.
Oxygenated Treatment
With all of the previous discussion about the deleterious effect of dissolved
oxygen, it may come as a surprise to some readers that a feedwater treatment
technology has emerged in which oxygen is deliberately injected into the con-
Figure 2-11
Condensate/Feedwater Chemistry 23
Figure 2-12
Condensate Condenser
High Pressure Low Pressure Polisher
Heaters Deaerator Heaters
BFP
To Boiler V
Oxygen Oxygen
Figure 2-13
FeOOH
Fe3O4 Fe3O4
DEOXYGENATED OXYGENATED
TREATMENT TREATMENT
Oxide Layer Formed in an Oxygenated Treatment Program.
a 50 to 150 ppb residual. Ammonia is added to raise the pH within a range of 8.0
to 8.5. Typically, 20 to 70 ppb of ammonia will produce this pH.
The chemistry of oxygenated treatment is very interesting and explains why
the program has become popular. In conventional AVT units, even with very
good chemistry programs, some iron oxides are still generated in the condensate
or feedwater system. The iron oxides are transported to the boiler, where they
precipitate on the tube walls. With controlled oxygen injection however, the base
layer of magnetite becomes overlayed and interspersed with an even tighter film
of ferric oxide hydrate (FeOOH). (See Fig. 2–13). This compact layer is more sta-
ble than magnetite and releases very little dissolved iron or suspended iron-oxide
particles to the fluid. Some utilities that switched from AVT to OT have reported
that dissolved feedwater iron concentrations, which were often 10 ppb or high-
er on oxygen-scavenger programs, dropped to as low as 1 to 2 ppb once the OT
program was fully established.
The key to an OT program is control. Makeup water must be highly puri-
fied because deposits would cause differential oxygen cells to form, which would
in turn cause under-deposit corrosion and pitting. Very pure feedwater in once-
through units is usually a given, as these steam generating systems are always
equipped with condensate polishers. OT cannot be used in systems that contain
copper-alloy feedwater heater tubes, as the corrosion would be too great.
One benefit of OT is that the feedwater pH is a unit or unit and a half lower
than that in AVT programs. This is due to the stability of the FeOOH film. The
lower pH can greatly increase condensate polisher run lengths.
Table 2-3
Feedwater (7)
Dissolved oxygen ppm (mg/1) O
- measured before chemical
oxygen scavenger addition (B) <0.007 <0.007 <0.007
Total iron ppm (mg/l) Fe ≤0.1 ≤0.05 ≤0.03
Total copper ppm (mg/l) Cu ≤0.05 ≤0.025 ≤0.02
Total hardness ppm (mg/l)* ≤0.3 ≤0.3 ≤0.2
pH @ 25˚C 8.3-10.0 8.3-10.0 8.3-10.0
Chemicals for preboiler system
protection NS NS NS
Nonbolatile TOC ppm (mg/l) C (6) <1 <1 <0.5
Oily matter ppm (mg/l) <1 <1 <0.5
Boiler Water
Silica ppm (mg/l) SiO2 ≤150 ≤90 ≤40
Total alkalinity ppm (mg/l)* <350(3) <300(3) <250(3)
Free OH alkalinity ppm (mg/l)* (2) NS NS NS
Specific conductance (12)
mhos/cm (s/cm) 25˚C
without neutralization 5400-1100(5) 4600-900(5) 3800-800(5)
*as CaO3
NS = not specified
ND = not detectable
VAM = Use only volatile alkaline materials upstream of attemperation water source (10)
Reproduced from Consensus on Operating Practices for the Control of Feedwater and Boiler Water
Chemistry in Modern Industrial Boilers with permission from the American Society of Mechanical Engineers.
Ali Sadeghi Digital Library
Condensate/Feedwater Chemistry 25
≤30 ≤20 ≤8 ≤2 ≤1
<200(3) <150(3) <100(3) NS(4) NS(4)
NS NS NS ND(4) ND(4)
Condensate/Feedwater Chemistry 27
Table 2-4
Feedwater (7)
Dissolved oxygen ppm (mg/1) O2
- measured before chemical
oxygen scavenger addition (1)(2) <0.007 <0.007
Total iron ppm (mg/l) Fe ≤0.1 ≤0.05
Total copper ppm (mg/l) Cu ≤0.05 ≤0.025
Total hardness ppm (mg/l)* ≤0.3 ≤0.3
pH @ 25˚C 8.3-10.5 8.3-10.5
Boiler Water
Silica ppm (mg/l) SiO2 ≤150 ≤90
Total alkalinity ppm (mg/l)* <1000(5) <850(5)
Free OH alkalinity ppm (mg/l)* (2) NS NS
Specific conductance (12)
mhos/cm (s/cm) 25˚C
without neutralization <7000(5) <5500(5)
*as CaO3
NS = not specified
Reproduced from Consensus on Operating Practices for the Control of Feedwater and Boiler Water
Chemistry in Modern Industrial Boilers with permission from the American Society of Mechanical Engineers.
Condensate/Feedwater Chemistry 29
Table 2-5
Feedwater (3)
Dissolved oxygen ppm (mg/1) O2 - measured
before chemical oxygen scavenger addition (1)(2) <0.007
Total iron ppm (mg/l) Fe <0.1
Total copper ppm (mg/l) Cu <0.05
Total hardness ppm (mg/l)* <1.0
pH @ 25˚C 8.3-10.5
Boiler Water
Silica ppm (mg/l) SiO2 ≤150
Total alkalinity ppm (mg/l)* <700(5)
Free OH alkalinity ppm (mg/l)* (4) NS
Specific conductance (12) mhos/cm (s/cm)
25˚C without neutralization <7000(5)
*as CaO3
NS = not specified
Reproduced from Consensus on Operating Practices for the Control of Feedwater and Boiler Water
Chemistry in Modern Industrial Boilers with permission from the American Society of Mechanical Engineers.
Table 2-6
Condensate/Feedwater Chemistry 31
Figure 2-14
<
>
>
Many of these items are standard (check valves, isolation valves, and relief
valves) to protect the equipment and personnel from the process fluid, and to
allow the maintenance staff to work on the equipment safely. Other items, while
still not standard, are becoming more common. One example is the increasing
use of diaphragm metering pumps in place of piston pumps. Diaphragm pumps
are reportedly more accurate and can be regulated more precisely than piston
pumps. At least one major metering pump manufacturer is phasing out piston
pumps in favor of the diaphragm type.
Another increasingly popular item is the pulsation dampener. This device
converts the pulsed bursts from the metering pump to a more steady-state flow.
Pulsation dampeners are becoming more common for other chemical feed appli-
cations including those to cooling tower basins.
Chemical feed rates are usually very low, and the time for a dose of reagent
to leave the pump discharge and arrive at the injection point may be quite long.
For instance, consider the following actual example. The hydrazine/pH-condi-
tioner feed pump on a 2400 psig unit delivers the solution at 12.3 gallons per
hour (GPH). Flow is through 3/8 in. OD (0.245 in. ID) tubing. The linear flow
rate is 1.4 ft/sec. For a line length of 100 feet, it takes over a minute for a dose
to leave the pump discharge and reach the injection point. If the line length had
been 500 feet, the time would increase to almost six minutes. Long lag times can
seriously affect the responsiveness of the chemical feed system to signals for
increased or decreased dosages. The closer the feed system is to the injection
point, the more responsive it will be.
Regarding oxygenated feed systems, various arrangements are available for
delivery of oxygen to condensate. Most popular is a simple system in which oxy-
gen cylinders, like those used by welders, are manifolded together. The discharge
is routed through a flow regulator and on to the injection point. Valves are
arranged such that when one cylinder empties, another can be put into service
immediately. This type of system can be easily placed near the injection location.
Some plant personnel prefer liquid oxygen as the source of supply, with a flow
regulated system. At least one U.S. utility simply uses the dissolved oxygen in the
makeup water as the supply. Results to date appear to be favorable, although it
seems that this arrangement allows less control than the others mentioned above.
Condensate/Feedwater Chemistry 33
couple of hours. The plant chemists will begin to check for a condenser leak if
the concentration remains above 1 ppb for a more extended period of time. Case
History 2–4 outlines an interesting situation involving an intermittent condenser
tube leak.)
Condenser tube leaks may occur for several reasons. Waterside fouling, dis-
cussed in chapter 6, can cause pitting and tube failure. Steamside corrosion, an
example of which has been outlined in Case History 2–3, is a frequent problem.
Condenser tubes are usually rolled into the tube sheet, and leaks may occur at
these locations. Whatever the problem, even small leaks are cause for concern.
Supplement 2–3 provides a simple BASIC program for calculating the quantity of
a leak based on an analysis of sodium. It can give plant chemists an idea of the
seriousness of the leak and the effect it may have on boiler water chemistry. The
program can be used check leakage volumes based on other ions such as chloride.
Sometimes, contaminants may be introduced to the condensate from unlike-
ly locations. Off-line contamination is well documented. In two cases involving
sandblasting of a condenser and deaerator storage tank, respectively, plant per-
sonnel forgot to tell the contractor that an inert material such as ground walnut
shells should be used as the blasting media. Instead, the contractor, who would
not be expected to know boiler water chemistry, used a form of silica-based mate-
rial for blasting. Upon unit startup, silica gradually leached from sandy residue
remaining in the vessels and affected water chemistry. Contaminants can also be
introduced from other unsuspected locations, and Case History 2–5 illustrates a
unique example of one such occurrence.
Condensate/Feedwater Chemistry 35
lar is either a two- or three-vessel system. In one common design, the resin is
sluiced to a backwash vessel for cleaning and hydraulic classification of the
cation and anion resin. The anion resin is then transferred to a separate vessel,
upon which each resin is regenerated individually. The regenerated resins are
then pumped to a third vessel where they are mixed and stored for recharge to
the condensate polisher.
A problem formerly encountered with this arrangement was cross-contami-
nation of each resin by off-size or broken beads of the other. Thus, cation resin
that ended up in the anion regeneration system would come out in the sodium
form, and anion resin accidentally introduced to the cation regeneration system
would be converted to the sulfate form. These ions would then leak into the con-
densate once the resin was returned to service.
Newer technologies have been developed that minimize the cross-contami-
nation of resins. The use of uniformly sized cation resins and uniformly sized
anion resins allows better separation during the backwash cycle. Some polishing
systems are designed to accommodate resins containing a small quantity of inert
beads. The inert resin has a density between that of the cation and anion resins,
and keeps them physically separated. Other polishers, such as the Conesep sys-
tem by Graver, extract the cation resin rather than the anion resin. Very good
resin separation has been achieved with these types of systems.
Deep-bed units can be operated in the ammonia form. In this arrangement,
the polisher releases ammonium ions during the exchange process. The ammo-
nia produced by the unit helps condition the condensate and feedwater, and per-
forms the same function as ammonia directly injected to the condensate.
Operation of a polisher in the ammonium cycle is more complicated, however.
Regeneration of the cation resin becomes a two-stage process, wherein the resin
is first regenerated with acid and then with ammonia. During operation of an
ammonia-hydroxide cycle polisher, leakage of other cations, particularly sodium,
may be greater than in a hydrogen-hydroxide cycle unit. Control of condensate
pH is also more difficult. For these reasons, ammonia-hydroxide cycle polishers
are not strongly recommended for high-pressure boilers.
An aspect of importance for hydrogen-hydroxide deep-bed polisher opera-
tion is ammonia concentration of the influent. The higher the ammonia concen-
tration, the more quickly the cation resin exhausts. The ammonia concentration
is of course related to the dosage used for pH control of the feedwater. Polisher
run lengths can become particularly short as the condensate pH rises towards the
9.5 level. Once-through plant personnel who switched feedwater treatments
from AVT to OT have noticed a severalfold increase in polisher run lengths. This
is due to the change in pH from a range of 9.0–9.6 to 8.0–8.5.
Condensate/Feedwater Chemistry 37
Condensate/Feedwater Chemistry 39
Conclusion
Control of chemistry in the condensate/feedwater system is extremely criti-
cal for proper boiler operation and safety. This chapter is intended to give
chemists readily available information regarding monitoring and control of feed-
water chemistry. (More details on sampling and monitoring are provided in chap-
ter 7, Sampling.) The case histories outlined in this and other chapters will hope-
fully illustrate that creative thinking is important when solving feedwater, boiler
water, and steam chemistry problems.
Supplement 2-1
Why Condenser Performance Is Important
I have been asked on a number of occasions, “Why isn't the turbine exhaust
steam returned ‘as is’ to the boiler? Wouldn't that save a lot of money by elimi-
nating the condenser and other equipment?” The answer is thermodynamically
related.
Consider what happens in a condenser. Ideally, when steam leaves the tur-
bine it has used all of its available heat for work and is at a saturated condition.
The condenser removes the steam's latent heat and converts it to condensate,
which produces a very strong vacuum. As the following example shows, this is
quite important thermodynamically.
The steam tables show that the enthalpy of the turbine inlet steam is 1477
BTU/lbm and the entropy is 1.6325 BTU/lbm °R. Because the turbine is adiabat-
ic and reversible, the entropy of the exhaust steam is the same as the entropy of
the throttle steam since no heat is transferred during the reversible expansion.
Accordingly, the enthalpy of the exhaust steam can be calculated for various con-
ditions. In the hypothetical turbine, if the steam were taken from the turbine
exhaust at atmospheric pressure, its enthalpy would be almost 1070 BTU/lbm.
Thus, 407 BTU/lbm of heat would be available for work. Because the enthalpy of
the condensate at these conditions is 180 BTU/lbm, the turbine efficiency would
equal ([1477 - 180] - 1080) / 1477 - 100 = 31%. If, however, all incoming steam
exhausts into a vacuum, the situation changes noticeably. With relatively cool cir-
culating water, absolute condenser pressures can reach as low as 0.5 inHg. The
enthalpy of steam exhausting into these conditions is near 846 BTU/lbm. Thus,
the amount of heat available for work increases to 631 BTU/lbm, giving a turbine
efficiency of 44%. This is an enormous improvement over the previous scenario.
(Note: Too much cooling can cause a loss of efficiency. This phenomenon, known
as condensate subcooling, will not be discussed in this book. Condensate sub-
cooling does not cause nearly the efficiency losses as do other factors such as tube
fouling or excess air inleakage.)
Obviously, the hypothetical example just presented is exaggerated since no
utility turbine is reversible or is designed to exhaust to atmospheric conditions.
Also, in most circumstances some steam is extracted from the turbine for feed-
water heating, so an increase in condensate temperature would decrease the
quantity of steam needed for feedwater heating. Thus, higher condensate tem-
peratures would alter the heat transfer in the feedwater heaters. However, the
quantity of steam extracted for feedwater heating is much smaller than that
which passes through the condenser, so the hypothetical example listed above
still clearly illustrates the thermodynamic importance of the condensing process.
Supplement 2-2
BASIC Program for Monitoring
Condenser Performance
Poor condenser performance can subtract 1% or more from unit efficiency.
This can add up to a considerable loss of money in fuel costs alone. The prima-
ry causes of poor performance are fouling of the tubes on the waterside or excess
air inleakage into the steamside. Each can drastically reduce condenser perfor-
Condensate/Feedwater Chemistry 43
mance, as is described in more detail in this book. Utility personnel employ var-
ious methods to track condenser conditions, including monitoring of condenser
backpressure and waterside inlet and outlet differential pressures. However, the
attached BASIC program provides a very effective method for tracking condenser
efficiency.
Program Output
The program calculates an ideal heat transfer (Ui), which is then compared
to the actual heat transfer (Ua). This value is known as the cleanliness factor.
When condenser tubes are placed in service they quickly develop an oxide coat-
ing. This coating, which actually protects the metal substrate, retards heat trans-
fer. Thus, a condenser free from mineral or microbiological deposits will still only
achieve about 85% of the ideal heat transfer, so a cleanliness factor of 85% indi-
cates totally clean tubes. (Factors above 80% are quite good and indicate that the
condenser is achieving good heat transfer.) It is very important that accurate
temperature readings are available because slight variations in temperature can
significantly affect results. However, even if absolute values are not totally accu-
rate, the program is still good for indicating changes in condenser performance.
CONDPERF.BAS
10 CLS:LOCATE 5,1
20 PRINT “CONDENSER CLEANLINESS FACTOR PROGRAM”
30 PRINT
40 PRINT “ENTER THE CIRCULATING WATER INLET TEMPERATURE (¯F)”
50 INPUT TIN
60 PRINT
70 PRINT “ENTER THE DENSITY OF THE INLET COOLING WATER (LB/FT^2).”
80 PRINT “A VALUE OF 62.3 IS SUITABLE FOR ALL CONDITIONS.”
90 INPUT RHO
100 PRINT
110 PRINT “ENTER THE CIRCULATING WATER OUTLET TEMPERATURE (¯F)”
120 INPUT TOUT
130 PRINT
140 PRINT “ENTER THE CONDENSER STEAM TEMPERATURE”
150 INPUT TSAT
160 PRINT
170 PRINT “ENTER THE CIRCULATING WATER FLOW RATE (GPM)”
180 INPUT FLOW
190 PRINT
200 PRINT “ENTER THE CIRCULATING WATER CORRECTION FACTOR AS”
210 PRINT “OUTLINED IN SUPPLEMENT 2-2.”
220 INPUT CWCF
230 PRINT
240 PRINT “ENTER THE CONDENSER TUBE CORRECTION FACTOR AS”
250 PRINT “OUTLINED IN SUPPLEMENT 2-2.”
260 INPUT CTCF
270 PRINT
280 PRINT “ENTER THE NUMBER OF CONDENSER TUBES”
290 INPUT NT
300 PRINT
310 PRINT “ENTER THE NUMBER OF CONDENSER TUBE PASSES”
320 INPUT NP
330 PRINT
340 PRINT “ENTER THE INSIDE TUBE DIAMETER (IN.)”
350 INPUT ID
360 PRINT
370 PRINT “ENTER THE OUTSIDE TUBE DIAMETER (IN.)”
380 INPUT OD
390 PRINT
400 PRINT “ENTER THE TUBE LENGTH (FEET)”
410 INPUT L
420 PRINT
430 PRINT “IN THE CALCULATIONS WHICH FOLLOW, THE IDEAL HEAT TRANSFER”
440 PRINT “IS DETERMINED THROUGH THE USE OF AN EMPIRICAL CONSTANT”
450 PRINT “DETERMINED BY THE HEAT EXCHANGE INSTITUTE. THIS CONSTANT”
460 PRINT “KNOWN AS C, VARIES DEPENDING ON THE OUTSIDE TUBE DIAMETER.”
470 PRINT “THE VALUES OF C ARE AS FOLLOWS:”
480 PRINT “5/8 & 3/4 INCHES, C=267”
490 PRINT “7/8 & 1 INCHES, C=263”
500 PRINT “1-1/8 & 1-1/4 INCHES, C=259”
510 PRINT “1-3/8 & 1-1/2 INCHES, C=255”
520 PRINT “1-5/8 & 1-3/4 INCHES, C=251”
530 PRINT “1-7/8 & 2 INCHES, C=247”
540 PRINT
550 PRINT “ENTER A VALUE FOR C”
560 INPUT C
570 PRINT
Condensate/Feedwater Chemistry 45
Supplement 2-3
BASIC Program for Calculating the Rate of
Cooling Water Leakage into a Condenser
When a leak develops in a condenser the flow rate of the leak can be deter-
mined by comparing the concentration of a particular contaminant in the cool-
ing water to that in the condensate. Sodium is a good parameter for this mea-
surement. The following BASIC program will calculate the flow rate of the leak
based on the difference in sodium concentrations. Plant personnel can then
determine the potential effects of the leak on unit chemistry, and decide whether
the size of the leak calls for immediate action.
CONDLEAK.BAS
10 CLS:LOCATE 5,1
20 PRINT “CONDENSER INLEAKAGE CALCULATION PROGRAM”
30 PRINT
40 REM QFW = FEEDWATER FLOW IN POUNDS PER HOUR
50 REM QFWGPM = FEEDWATER FLOW IN GPM
60 REM NAFW = SODIUM CONCENTRATION IN THE FEEDWATER (PPM)
70 REM NACW = SODIUM CONCENTRATION IN THE COOLING WATER (PPM)
80 REM QCWL = COOLING WATER LEAK INTO CONDENSER (GPM)
90 INPUT “ENTER THE FEEDWATER FLOW (POUNDS PER HOUR)”;QFW
100 PRINT
110 INPUT “ENTER THE SODIUM CONCENTRATION IN THE FEEDWATER (PPM)”;NAFW
120 PRINT
130 INPUT “ENTER THE SODIUM CONCENTRATION IN THE COOLING WATER (PPM)”;NACW
140 PRINT
150 QFWGPM=QFW/(8.34*60)
160 QCWL=(QFWGPM*NAFW)/(NACW-NAFW)
170 PRINT “THE RATE OF COOLING WATER FLOW INTO THE CONDENSER = “;:PRINT
USING “###.##”;QCWL;:PRINT “ GPM”
Condensate/Feedwater Chemistry 47
any gases to vent to the atmosphere. When the trap stuck open, the strong con-
denser vacuum pulled outside air in through the vent. Maintenance personnel
quickly replaced the trap and the condenser performance problems disappeared.
Condensate/Feedwater Chemistry 49
Table 2-7
Condenser Tube Correction Factors
Tube Wall Gauge - BWG
Tube Material 24 22 20 18 16 14 12
Admiralty Metal 1.06 1.04 1.02 1.00 0.96 0.92 0.87
Arsenical Copper 1.06 1.04 1.02 1.00 0.96 0.92 0.87
Copper Iron 194 1.06 1.04 1.02 1.00 0.96 0.92 0.87
Aluminum Brass 1.03 1.02 1.00 0.97 0.94 0.90 0.84
Aluminum Bronze 1.03 1.02 1.00 0.97 0.94 0.90 0.84
90–10 Cu–Ni 0.99 0.97 0.94 0.90 0.85 0.80 0.74
70–30 Cu–Ni 0.93 0.90 0.87 0.82 0.77 0.71 0.64
Cold Rolled Low
Carbon Steel 1.00 0.98 0.95 0.91 0.86 0.80 0.74
Stainless Steels
Type 304/316 0.83 0.79 0.75 0.69 0.63 0.56 0.49
Titanium 0.85 0.81 0.77 0.71 – – –
Condenser Tube Correction Factors. Source: Heat Exchange Institute. Adapted from “Computer program pre-
dicts condenser cleanliness factors,” Power Engineering, June 1992.
Table 2-8
Cooling Water Correction Factors
Cooling Water Correction Factors. Source: Heat Exchange Institute. Adapted from “Computer program pre-
dicts condenser cleanliness factors,” Power Engineering, June 1992.
Ali Sadeghi Digital Library
02.qxd 8/8/06 1:51 PM Page 50
Chapter 3
Boiler Water
Chemistry
Introduction
Deposition of contaminants and corrosion are common in boilers and are
the cause of many forced outages. Material damage and lost power generation or
product output due to corrosion cost electric utilities and industries billions of
dollars per year. This chapter outlines the principal mechanisms of boiler chem-
istry upsets and discusses treatment methods to minimize them. Preceding this
discussion is a brief look at some of the common boiler types used in the fossil-
fired steam generating industry.
STEAM
Boilers
DRUM
Many industrial plants
generate steam only for process
use or to drive small turbines.
For these applications, precon-
structed, package boilers are
often sufficient. Figures 3-1
through 3-3 illustrate the sim-
plified circuitry of three of the
most common package boiler
designs, the “A,” “D,” and “O”
types. Natural gas or oil is the
MUD MUD principal fuel for these boilers.
DRUM DRUM Steaming rates typically range
General Circuitry of an A Type Boiler from 40,000 to 200,000
pounds per hour, with
100,000 pounds per hour being very common. As the figures indicate, each of
the boilers contains one or more mud drums. The mud drum(s) serves as a col-
lection site for precipitates formed by chemical treatment programs.
The greatest advantage of package boilers is that they can be added in stages
as plant capacity increases. Package boilers are relatively small in size, so site and
construction requirements are far more simple than for field-erected units. The
boilers can be equipped with superheaters to produce steam of the quality and
temperature needed to drive turbines or provide adequate heat for high temper-
ature industrial processes. Package boilers are typically rather short in height,
where their length may be modified depending on steaming requirements.
cally just before or in the gas passage from the boiler, although some horizontal-
ly arranged superheater and reheater tubes may be located further downstream
in the gas passage. Larger units are often equipped with an economizer that is
placed in the gas passage near the low-temperature superheaters and reheaters.
Figure 3-4
Utility Boiler. Photo taken with permission of the Kansas City, Kansas, Board of Public Utilities,
Quindaro Power Station
Figure 3-5
course of the day will gradually change from black to gray to nonexistent as the
particles are either removed by the boiler blowdown or precipitate on the tube
walls.
One particular difficulty with iron oxide deposition is that the particles pre-
cipitate more heavily on the hot side of the tubes. This is the worst location, as
higher temperatures increase the potential for corrosion. Under-deposit corro-
sion mechanisms will proceed more rapidly on the hot side of the tubes.
Condenser Inleakage
A condenser tube leak allows many unwanted contaminants to enter the
steam generating system. These contaminants, upon reaching the boiler, react to
Figure 3-6
SATURATED
STEAM
>
STEAM
DRUM
DOWNCOMERS WATERWALL
TUBES
>
FLOW > FLOW
HEAT >
>
>
LOWER
HEADER
form scale, acid, or other deleterious compounds. The following equations illus-
trate some of the principal reactions.
Equations 3.1 and 3.2 are typical scale-forming reactions. Many hardness
compounds exhibit retrograde solubility as temperatures rise (Fig. 3-8) and will
precipitate on boiler internals. The effect of such deposition on heat transfer is
outlined in Table 3-1, which shows thermal conductivities for several boiler met-
als and scales. Even a relatively thin deposit layer will significantly reduce heat
transfer, and a boiler must be fired harder to achieve the same level of steam pro-
Figure 3-7
duction. This in turn can lead to overheating of the boiler tubes, which will
shorten tube life. In cases of severe deposition, heat transfer is so restricted that
overheating may occur very rapidly resulting in catastrophic tube failure.
Deposits are also the precursor to another phenomenon known as underde-
posit corrosion, which may occur via several different mechanisms. Boiler water
is normally maintained in an alkaline range (pH of 9 to 11 depending on boiler
pressure and type) to prevent acidic attack (Fig. 3-9). This pH range is moder-
ately basic, and small amounts of caustic alkalinity (OH-) may be present in the
bulk boiler water. In one form of caustic corrosion, water that enters a porous
deposit through some openings will boil off through other channels, leaving
sodium hydroxide behind. The concentrated NaOH then attacks the boiler metal
and protective magnetite film via the following reactions:
The localized attack may cause tube damage within a relatively short period
of time.
Deposits can also cause the concentration of other species including acid
chlorides, which are generated by reactions similar to that outlined in Equation
3.3. Not only does the acid corrode the boiler metal, but the reaction generates
hydrogen, which can lead to hydrogen damage of the tubes. In this mechanism,
hydrogen gas molecules, which are very small, penetrate into the metal wall
where they then react with carbon atoms in the steel to generate methane (CH4):
Figure 3-8
Figure 3-9
Phosphate ions (PO4-3) can exist in the mono, di, and trihydrogen state in
aqueous solutions, and thus can give up or accept hydrogen ions. This buffering
capability makes phosphate effective in preventing wide pH swings in boiler
water due to contaminant ingress.
Phosphate’s second major function is to absorb contaminants that enter the
boiler. Phosphate reacts directly with calcium to produce calcium hydrox-
yapetite:
10Ca+2 + 6PO4-3 + 2OH- → 3Ca3(PO4)2 • Ca(OH)2↓ (3.8)
Figure 3-10
Guidelines for Coordinated Phosphate Control Source: Betz Industrial Water Treatment
Seminar, BetzDearborn, Inc., Horsham, PA.
Hydroxyapetite and serpentine are much more benign and easy to remove via
blowdown than the hard scale or corrosive products which would otherwise form.
Figure 3-11
Guidelines for Coordinated Phosphate Control Adjusted for Ammonia Level. Chart
Provided by BetzDearborn, Inc., Horsham, PA.
This effect must be taken into account when monitoring the phosphate
treatment program. Several methods are possible. Figure 3-11 illustrates control
curves for coordinated phosphate treatments at various ammonia levels. By mon-
itoring the ammonia concentration in the boiler water, the chemist can prepare
phosphate dosages based on the curves. Where neutralizing amines are used, the
calculation becomes more complicated since various amines exhibit different
basicities. Holland Technologies of Jamison, Pennsylvania, has developed a com-
puter program that accounts for the effects of ammonia and the common amines
that are used in feedwater systems. Finally, measurement of both sodium and
phosphate allows direct calculation of the sodium-to-phosphate ratio. If done
on-line, the results can be incorporated into data acquisition systems, such as
that already mentioned in Case History 3-3.
Within the last decade or so, more and more difficulties with congruent
chemistry in high-pressure boilers have become apparent. This is due to phos-
phate hideout and the incongruent precipitation of sodium phosphate.
Phosphate Hideout
Figure 3-12 illustrates the reverse solubility of sodium phosphate. The effect
becomes very pronounced at the boiler water temperatures common to units that
Figure 3-12
Solubility of Trisodium Phosphate versus Temperature Copyright © 1986. Electric Power Research Institute.
EPRI CS-4629. Interim Consensus Guidelines on Fossil Plant Cycle Chemistry. Reprinted with permission.
operate above 2000 psi and/or are subjected to frequent load changes. Reports
by some utilities formerly using congruent treatment indicate that phosphate
concentrations would often drop well below 1 ppm when hideout was severe.
This depleted the boiler water of the chemical designed to control chemistry and
protect against contaminant introduction. Furthermore, researchers found that
the phosphate precipitated incongruently, with deposit sodium-to-phosphate
molar ratios of 2 to 1 or lower.
The effects of hideout are most graphically illustrated in a cycling unit.
Consider a boiler at low load with congruent treatment. As load is raised and
heat fluxes increase, sodium phosphate begins to precipitate onto the tube sur-
faces. Due to incongruent precipitation, the Na/PO4 ratio of the phosphate
remaining in solution rises. This increases the pH to perhaps several tenths of a
unit above congruent guidelines. The chemist may then add a di-/trisodium
blend to raise the phosphate concentration and lower the pH, but this phosphate
also hides out. When boiler load is reduced, the situation reverses and the pre-
cipitated phosphate redissolves. In this case, however, the sodium phosphates,
which have a low Na/PO4 ratio, drive the pH downward. Severe hideout and the
reverse dissolution process have been known to force boiler water pH below 7 in
cycling units. Since EPRI recommends unit shutdown if the pH drops below 8.0,
Phosphate Treatment
Phosphate treatment almost appears to be a combination of EPT and con-
gruent or coordinated treatment. Phosphate residuals are maintained between
2.5 to near 10 ppm, with a sodium-to-phosphate ratio of 2.8 to 1 or greater. As
with EPT, caustic alkalinity is held to 1 ppm maximum. Hideout in PT-condi-
tioned systems has been reported. Although such deposits are more alkaline in
nature than those in a congruent program, hideout can still create difficulties for
the chemist. Possibly for this reason, EPT appears to be the more preferred
replacement for congruent or coordinated programs.
Figure 3-13
HOOC-CH2 CH2-COOH
N-CH2-CH2-N
HOOC-CH2 CH2-COOH
Structure of the ETDA Molecule.
are 40 and 24.3. This indicates that 7.2 and 11.9 ppm of EDTA would theoreti-
cally react with 1 ppm of calcium and magnesium, respectively.
The use of chelants must be carefully considered because they can cause prob-
lems in boiler systems. For example, chelants are attacked by dissolved oxygen and
should not be added to the feedwater until after the deaerator. More importantly,
improper control and overfeed of chelants can cause dissolution of the protective
magnetite layer on the boiler tubes. Serious corrosion has been known to occur.
Polymers of the acrylate (Fig. 3-14), sulfonated styrene (Fig. 3-15), and
other families are also used for boiler water scale control. Most often the poly-
mers are employed as part of a combined program with phosphates or chelants.
The polymers modify the crystalline structure of precipitates, making them less
sticky and more easily removable by blowdown. These chemicals do not act on
a stoichiometric basis with the compounds being treated, and typical dosages
may range from 1 to 10 ppm.
One area where sequesterants such as polymethacrylate (Fig. 3-16) or sul-
fonated styrene appear to be particularly effective is iron control, especially after
unit startup. The sequesterants will bind free iron and allow it to be removed by
the boiler blowdown rather than precipitate on the tube walls.
Chelants and polymers have been
Figure 3-14 effectively used for boiler water chemistry
CH2 CH2 ~ control. However, proper selection of the
~ CH CH
correct chemical or combination of chem-
COOH COOH icals may not be easy. Monitoring of the
Structure of Polyacrylic Acid. chemical residuals is more difficult than in
phosphate programs. Many people in the
Figure 3-15
electric utility industry appear to favor the
H H H more “simple” phosphate or volatile treat-
~ C C C ~ ment programs. For low-pressure indus-
trial treatment, however, a chelant/poly-
H mer program may perform very well.
Certainly polymers should be considered
SO H SO H for iron control during unit startups.
3 3
Structure of Sulfonated Polystyrene.
All-Volatile Treatment
All-volatile treatment (AVT) was principally developed for once-through
boilers, since these units cannot tolerate dissolved solids. Because once-through
boilers have no drum, boiler water chemistry is a function of feedwater chem-
istry. AVT chemistry guidelines for a once-through unit call for a pH range of 9.3
to 9.6 with less than 2 ppm dissolved solids. Ammonia levels may range from 1
to 2 ppm. Condensate polishers are an absolute requirement for once-through
units because a condenser leak or demineralizer upset would introduce uncon-
trollable contamination to the boiler.
AVT is also used in some very high-pressure, drum-type units where the
pressure approaches critical. As boiler pressures increase to this value, the den-
sity of steam and water approach each other. Thus, it becomes much more diffi-
cult to separate water from steam in the boiler drum internals, so boiler water
must be quite pure to prevent carryover.
AVT minimizes mechanical carryover, but it does not protect drum boilers
from contaminant introduction due to a condenser leak or other problems.
Condensate polishers are the most effective buffer against chemistry upsets. For
units without polishers, the chemical feed system should contain an emergency
phosphate unit so that phosphate can be immediately injected into the drum in
the event of an upset. This must be considered as only a temporary measure,
because the sudden increase in dissolved solids due to contaminant inleakage
and phosphate injection will greatly increase the potential for carryover and
deposition.
Several problems with AVT have become evident. First, the volatile treat-
ment chemicals will carry over into the steam. Unfortunately, they also tend to
carry chloride and sulfate with them, which then deposit on the low-pressure
turbine blades. Chlorides and sulfates are prime contributors to stress-corrosion
cracking and corrosion fatigue. Second, the volatiles are less capable of neutral-
izing acidic turbine deposits than phosphate, which, if it carries over in small
quantities, may be beneficial due to its alkaline nature. Third, where amines and
organic oxygen scavengers are used, the breakdown products may include car-
bon dioxide and organic acids. These are potential corrodents of turbine blades
Oxygenated Treatment
OT, which was described in chapter 2, is basically a feedwater treatment. Its
effectiveness in boilers stems from the fact that OT greatly reduces iron transport.
This is quite important, as iron oxide usually makes up the bulk of boiler tube
deposits. European utilities that have been using OT for a long time have found
that chemical cleanings have been significantly reduced. This provides a great
economic incentive, when one considers that the recommended chemical clean-
ing frequency of a supercritical unit is every 18 months.
Case History 3-4 provides some specific details on the startup and initial
observations of an OT program at a supercritical steam generating unit. An
important point to remember is that the oxide layer formed in an OT program is
orange in color. This can come as a surprise when plant personnel open up a
feedwater system after initiating oxygenated treatment.
OT is currently being tested in one drum boiler in the United States. The
program has been in progress for over three years and results so far seem to be
promising. Iron transport has definitely been lowered. Look for this treatment to
be tested in other drum boilers.
The concept of oxygenated treatment can be difficult to accept, especially con-
sidering that for years all of the chemistry manuals emphasized removal of oxygen
from boiler feedwater. However, positive reports about the treatment are proliferat-
ing. OT has even been recommended for circuits in heat recovery steam generators.
Figure 3-17
Sampling
Sampling of boiler water is extremely important because the harsh environ-
ment in the boiler greatly magnifies the effects of corrosion and deposition mech-
anisms. The most important constituents to be monitored include pH, conduc-
tivity, silica, and phosphate. On-line monitoring is greatly preferred, with grab
sampling as a backup. For drum-type units, the sample is either taken from spe-
cial ports located in the drum, or more often from the boiler blowdown or down-
comer. Further details are outlined in chapter 7.
Conclusion
A number of effective programs are available for boiler water treatment.
Plant chemists should carefully monitor boiler water conditions to determine the
effectiveness of the program in use. Single upsets have been known to cause tube
failures and forced outages that have cost utilities millions of dollars for material
replacement and purchased power. In industry, boiler failures can curtail or shut
down production units.
Off-line corrosion can also be very destructive. Proper layup procedures are
important to maintain the integrity of the steam generating unit during outages.
Supplement 3-1
BASIC Program for Calculating
Sodium-to-Phosphate Ratios of Boiler Water
Sodium-to-phosphate ratio monitoring is very important for boilers that are
treated with coordinated or congruent phosphate. Chemists must maintain the
ratios within relatively narrow guidelines to properly control the boiler water
chemistry. The BASIC program on page 77 provides a simple and efficient method
for determining these ratios. The only inputs needed are pH and phosphate con-
centrations (ppm), however, ammonia may significantly affect the calculations.
The calculations first appeared in the November 1986 issue of Power
Engineering magazine, and the computer program then appeared in the May
1992 issue of this magazine. Since the publication of these two articles, utility
personnel and researchers have become much more aware of the effect of ammo-
nia on boiler water pH. At the 1996 International Water Conference, George
Verib of Ohio Edison presented an excellent paper on boiler water treatment,
part of which discussed ammonia and its relation to sodium/phosphate ratios.
Ammonia can have a very significant effect on the calculations, especially in
higher pressure units where the sodium-phosphate concentrations are low. Ohio
75
Ali Sadeghi Digital Library
03.qxd 8/8/06 1:52 PM Page 76
PHOS.BAS
10 CLS:LOCATE 5,1
20 PRINT “THIS PROGRAM WILL CALCULATE THE SODIUM/PHOSPHATE RATIO IN BOILER
WATER.”
30 REM - THE CALCULATIONS FOR THIS PROGRAM MAY BE FOUND ON PAGES 31-32
40 REM - OF THE NOVEMBER 1986 ISSUE OF POWER ENGINEERING
50 PRINT
60 INPUT “ENTER THE DRUM WATER pH” ; PH
70 PRINT
80 INPUT “ENTER THE PPM OF PHOSPHATE IN THE DRUM WATER” ; PO
90 PRINT
100 REM - THE MOLAR PHOSPHATE CONCENTRATION = PO/94971
110 MCPO=PO/94971!
120 REM - H = THE HYDROGEN ION CONCENTRATION
130 H=1-^(-1*PH)
140 POH=PH-14
150 REM - OH = THE HYDROXYL ION CONCENTRATION
160 OH=10^(PH-14)
170 REM - A = THE K1 CONSTANT FOR THE PHOSPHATE SERIES
180 A=10^-2.1
190 REM - B = K1 * K2
200 B=A*10^7.2
210 REM - C = K1 * K2 * K3
220 C=B*10^-12.3
230 REM - THE CALCULATIONS IN LINES 260-280 DETERMINE THE “APPLICABLE PHOS-
PHATE
240 REM - SPECIES DISTRIBUTION USING IONIZATION FRACTION METHOD FOR TRIPRO-
TIC
250 REM - ACID.”
260 E=H*H*H : F=H*H*A : G=H*B : D=C=E=F=G
270 AA=H*H : KA=A/D : KB=B/D : KC=C/D
280 DIPO=(AA*KA)*MCPO : MONO=(H*KB)*MCPO : ZERO=KC*MCPO
290 REM - THE FOLLOWING CALCULATIONS DETERMINE THE MOLAR CONCENTRATIONS OF
300 REM - THE VARIOUS PHOSPHATE SPECIES AND SODIUM
310 MSP=H+DIPO : TSP=OH+ZERO : DSP=MCPO-MSP-TSP
320 NA=MSP+(2*DSP)=(3*TSP)
330 R=NA/MCPO
340 PRINT “THE SODIUM TO PHOSPHATE RATIO IS ” ;:PRNT USING “##.##”; R
in Supplement 3–1. The monitoring system also is equipped with a PLC that
controls valve operation and collects data from the instruments for distribution
to remote computer terminals. PLC logic includes common arithmetic functions,
so I converted the calculations in the BASIC program to ladder logic and sent the
output to PLC registers. The values were then transmitted to the remote display
screens along with other water chemistry data.
This data distribution concept could be used for other boiler water treat-
ment programs. For any future applications involving congruent chemistry, the
effects of ammonia or amines would make the calculations more complicated. A
potential solution is to monitor both sodium and phosphate on-line and set up
the PLC logic to directly calculate sodium-to-phosphate ratios.
• The condensate piping was overlayed with the protective FeOOH film in
about one day, but it took almost a week for the feedwater piping to be
converted
• The cation conductivity of the condensate polishers remained below the
utility guideline of 0.15 MS
• Dissolved iron concentrations dropped from a range of 9.0–9.7 ppb to
around 3 ppb
• The formerly black color of the pipe surfaces (magnetite) became brown-
ish orange due to the formation of FeOOH
So far, the plant staff has been very pleased with the program.
Figure 3-18
MUD DRUM
STEAM INLET
HEADER
STEAM
OUTLET
Figure 3-19
Chapter 4
Steam
Chemistry
Introduction
The previous chapters have outlined the importance of feedwater and boil-
er water chemistry. Although control of chemistry in these systems is critical for
protection of the systems themselves, feedwater and boiler water chemistry have
a direct effect on steam purity. Deposition of solids and corrosion in superheaters
and turbines have been common problems. Even in industrial plants that have
no turbines, proper control of steam chemistry is still very important. Some of
the most intensive research in the steam generating industry is being directed
towards steam chemistry.
Steam produced in the boiling process is never completely pure, and even
at its best contains trace amounts of solids. The process by which solids are trans-
ferred to steam is known as carryover. Carryover is influenced by the following:
• Virtually all solids are at least slightly soluble in steam. Solids become
more soluble as boiler pressure increases, and some, particularly silica,
carry over extensively as a vapor.
• Even with the best steam separating devices, moisture droplets still enter
the steam.
• Sulfate and chloride will enter steam as ammoniated salts.
• Improper drum level control, poor drum design, or excessive solids
buildups in drum water will increase carryover.
Contaminants also enter steam via attemperator systems. Even if the drum
is operating properly and minimizing carryover, dissolved solids still have this
direct path to the steam system. Attemperator contamination is greatly exacer-
bated during upset conditions such as a condenser leak.
Another mechanism for steam contamination is exfoliation of superheater
and reheater tube walls, which introduces iron oxide particles to the steam.
Copper
Copper tends to precipitate in the high-pressure stage of the turbine, where
it forms copper oxide or metallic copper deposits. Over time, the deposits will
affect turbine performance, and many cases of unit derating due to copper depo-
Steam Chemistry 85
Figure 4-1
Steam Solids Concentration versus Steam Pressure. Copyright © 1986. Electric Power Research Institute.
EPRI CS-4629. Interim Consensus Guidelines on Fossil Plant Cycle Chemistry. Reprinted with permission.
sition have been recorded. This phenomenon has been under close scrutiny in
recent years, as chemists strive to understand the mechanisms and magnitude of
the problem.
Sodium Hydroxide
Sodium hydroxide is soluble in water and steam, and exists as a dissolved
compound in both phases. Sodium hydroxide is a primary culprit in stress cor-
rosion cracking (SCC), and, as Figure 4-3 shows, may cause turbine blade cor-
rosion in the area between the inlet of the low-pressure turbine and the begin-
ning of the saturation zone. Turbine blades are naturally under stress due to oper-
ating conditions, and thus are susceptible to SCC.
Figure 4-2
Solubility Characteristics of Various Carryover Products in the Steam Cycle. Copyright © 1987.
Electric Power Research Institute. EPRI CS-5275. Water, Steam, and Turbine Deposit Chemistries in Phosphate-
Treated Drum Boilers. Reprinted with permission.
Iron Oxides
Iron oxides transported by high-velocity steam cause solid particle erosion
(SPE) of the turbine blades. Turbine screens remove some of these particles, but
small particles may pass through the screens and erode the edges of the high-
pressure turbine blades. Exfoliation of superheater piping generates many of
these iron oxide particles.
Steam Chemistry 87
Figure 4-3
Concentration Zones of Contaminants in a Turbine. Copyright © 1986. Electric Power Research Institute.
EPRI CS-4629. Interim Consensus Guidelines on Fossil Plant Cycle Chemistry. Reprinted with permission.
Silica
Silica will coat turbine blades and gradually cause a decrease in turbine per-
formance. Although it is sometimes possible to water wash silica from the blades,
some of the deposits may be very tenacious. Chemical cleaning is required to
remove this adherent scale.
Sodium Phosphates
Sodium phosphates will carry over and deposit on intermediate- and low-
pressure turbine blades. If the deposits are not too heavy this can actually pre-
sent an advantage, as the phosphates help neutralize acids formed by other car-
ryover products. However, excessive phosphate carryover is detrimental. Many
cases of phosphate carryover have been reported, with some causing reheater
tube failures due to sodium phosphate accumulation in reheater U-bends.
Organics
Breakdown of organics in the boiler or steam lines introduces organic acids
and carbon dioxide to the turbine. These acids can cause corrosion in the low-
pressure blades. At chemical plants, excess organics in condensate return have
been known to either carry over to steam directly or influence boiler water chem-
Mechanical Carryover
During the steam generation process in drum boilers, water droplets become
entrained in the steam. If the droplets are allowed to remain, they will introduce
dissolved solids to the superheater, reheater, and turbine. Boiler drums, there-
fore, contain steam separating devices to remove the bulk of the entrained mois-
ture. Figure 4-4 shows a simplified diagram of a common steam separating
scheme. The separation equipment relies upon the difference in density between
steam and water. In the design shown, steam first passes through cyclonic sepa-
rators, which throw water droplets to the cylinder walls where they drain back
into the drum. The steam then passes through chevron-like scrubbers that
remove additional water. Other steam drum internals may represent those shown
in Figure 4-5.
Even these devices
Figure 4-4
will not prevent steam
STEAM contamination if boiler
OUTLET
water chemistry or operat-
ing conditions are not
> < properly controlled. The
most common factors that
< SCRUBBERS >
influence mechanical car-
>
<
Steam Chemistry 89
Figure 4-5
Inside of a Steam Drum. Photo taken with permission of the Kansas City, Kansas, Board of Public Utilities,
Quindaro Power Station.
the demand for steam changes more quickly than the unit can supply it. Priming
can also be caused by poor drum level control, wherein high drum levels over-
load the steam separating equipment. Foaming results from excessive concentra-
tions of solids in the boiler water. As the foam bubbles burst, they introduce
excessive quantities of water to the steam separating devices, which may not be
able to totally remove the added moisture. Foaming may be more prevalent at
industrial facilities, where makeup treatment systems are less sophisticated and
condensate polishing is minimal or nonexistent. Case History 4-2 outlines
superheater deposition problems at a chemical plant, that were believed to have
been caused by introduction of excessive organics to the steam boilers.
Inadequate drum design will influence all of the phenomena mentioned
above. One of the most important factors in drum design is that sufficient space
be provided between the operating water level and the steam separation devices.
If the drum is too small, slugs of water or excess water vapor will enter the steam
separating devices. Operation of the boiler at higher than rated capacity can also
overload the steam separating internals.
Proper operation of the steam separating devices is critical for production of
high-purity steam. Maintenance personnel should inspect the internals during
every planned maintenance outage where the steam drum is opened and repair
those components which may be damaged or out of position.
Vaporous Carryover
Figure 4-1 illustrated the increase of solids solubility with increasing steam
pressure. As unit pressure increases, boiler water chemistry must be more tight-
ly controlled to restrict vaporous carryover. Ultimate limits are reached in once-
through units, where concentrations of solids must be below steam saturation
values so that none of the solids precipitate in the waterwall tubes.
Silica is the most well-known vaporous carryover product. It is also one of
the most troublesome for several reasons. First, silica can form tenacious deposits
on turbine blades, that often can only be totally removed by chemical cleaning.
Second, the vaporous carryover potential greatly increases with increasing boiler
pressure. For example, in a 900 psig boiler without reheat, the recommended
maximum boiler water silica concentration is 7 ppm. In a 2400 psig boiler with
reheat, the recommended maximum is less than 0.2 ppm! Third, silica can easi-
ly enter a steam generating unit. It is the most weakly held ion on makeup sys-
tem anion resin and will break through first. More importantly, silica can exist in
colloidal form in water. The colloids will pass almost untouched through a dem-
ineralizer, but will break down to reactive silica from the heat of the boiler water.
(See Case History 4-3.)
Vaporous carryover of copper oxides is another area of concern. This pre-
dominantly occurs at pressures over 2300 psig and can be quite troublesome in
high-pressure units. Even a few pounds of copper on turbine blades may be
enough to decrease the turbine capacity by several megawatts. The corrosion
control procedures outlined in chapter 2 are very important for prevention of
copper transport from the feedwater system to the boiler to the turbine. This
includes off-line corrosion prevention. A number of cases of severe copper cor-
rosion of feedwater heater tubes during unit outages have been reported. In some
instances, a subsequent chemical cleaning of the boiler generated several thou-
sand pounds of copper in the cleaning solution.
Table 4-1
Steam Chemistry 91
Superheater Exfoliation
The stresses experienced by steam generating equipment during shutdown,
start-up, and cycling duty increase the mechanical degradation of the equipment.
This occurs in the superheater and reheater, where the stress and harsh environ-
ment cause exfoliation of the magnetite layer on the tube surfaces. These parti-
cles may then pass to the turbine where they can cause erosion of the turbine
blades. This effect, known as solid particle erosion, is most commonly found in
the high-pressure end of the turbine where linear velocities are highest. Case
History 4-4 illustrates some of the products that may be found in superheater
and reheater tubes due to exfoliation and mechanical carryover.
Sometimes, corrosion of afterboiler components may occur due to excess
oxygen in the steam. A prime location for this corrosion is the crossover line
between the high-pressure and low-pressure ends of the turbine, where any
slight amount of moisture helps generate the corrosion reaction. Some utilities
inject an oxygen scavenger into the crossover line to prevent oxygen attack of
carbon steel components in this area and in the low-pressure turbine. This brings
up an interesting point regarding oxygenated treatment. Research to date indi-
cates the absence of oxygen corrosion in the turbine. This is probably due to the
carefully controlled conditions of the program.
Sodium 3 5
Cation Conductivity (/s/cm) <0.15 <0.3
Chloride (ppb) 3 3
Sulfate (ppb) 3 3
Silica (ppb) 10 10
TOC (ppb) 100 100
Recommended Steam Turbine Purity Limits for Utility Boilers. Source: Electric Power Research Institute.
Steam Chemistry 93
Table 4-3
Recommended Boiler Water Concentrations to Meet
Steam Purity Guidelines
Boiler Pressure Sodium Silica Chloride Sulfate
(psig) (ppm) (ppm) (ppm) (ppm)
900 3.3 3.6 0.33 0.33
1100 3.0 1.9 0.24 0.24
1300 2.7 1.3 0.17 0.17
1500 2.5 0.81 0.13 0.13
1700 2.2 0.57 0.086 0.086
1900 1.9 0.39 0.065 0.065
2100 1.6 0.27 0.048 0.048
2300 1.2 0.16 0.037 0.037
2500 0.71 0.14 0.028 0.028
2700 0.44 0.085 0.020 0.020
2900 0.27 0.051 0.014 0.014
Author’s Interpretation of Graphs in the EPRI Interim Consensus Guidelines for Fossil Plant Cycle
Chemistry. Adapted from Interim Consensus Guidelines on Fossil Plant Cycle Chemistry. Electric Power Research
Institute. EPRI CS-4629. Copyright © 1986.
inspection. The edges of the turbine blades exhibit an abraded pattern. This
problem can become more acute in cycling units because frequent startups, shut-
downs, or load changes impart additional stress to the tubes.
Superheater/reheater exfoliation is mostly a mechanical phenomenon, and a
chemical cleaning of the tubes may be necessary to remove the fractured mag-
netite layers. Superheater chemical cleaning is not simple, and requires tempo-
rary piping, well thought out procedures, and the services of a reliable chemical
cleaning firm. However, superheater/reheater cleaning may still be cost effective
if SPE is a severe problem. For boilers with drainable superheaters, it is possible
to remove some of the exfoliated material by water washing.
Steam Chemistry 95
using softened water for makeup can generate prodigious amounts of carbon
dioxide. When the CO2 dissolves in the condensate, the corrosivity increases.
Treatment with neutralizing or filming amines is often required to protect down-
stream condensate piping.
Conclusion
Control of steam chemistry is vitally important for protection of turbines
and other afterboiler components. Intensive research is being conducted into the
effects of various compounds in turbines, and recommended steam purity limits
continue to become lower. Utility chemists should be aware that tightening
steam chemistry guidelines will require improved boiler water chemistry control
and boiler water/steam monitoring.
The boilers generate superheated steam for use in the production of phenol
and phenol-derivative compounds. Approximately 80% of the condensate is
returned to the boilers. Neutralizing amines are injected into the condensate for
corrosion control. Steam is only used for process functions and does not drive
any turbines.
Sporadic, but abnormally frequent, failures have been occurring in the
superheater tubes. When a tube fails, the company removes and replaces the
entire superheater, a rather expensive procedure. Tubes taken from the removed
sections contained internal deposits whose thickness ranged from 1/8 inch to per-
haps 1/4 inch. An analysis of these deposits showed that they were primarily com-
posed of iron oxide, silica, sodium, and carbon. Each of the boiler drums con-
tains steam separating internals, and the company even added an external set of
separating devices to one of the most problematic boilers. The problems, how-
ever, still persist.
An examination of water chemistry records revealed the probable source of
the difficulties—organics. For boilers of this pressure, the recommended maxi-
mum TOC level in the boiler water is 0.5 ppm. Prior water chemistry records by
an outside vendor indicated that TOC levels in the condensate return sometimes
reached the 20 ppm level. Plant personnel admitted that other tests had shown
that TOC levels in the condensate return had risen above 200 ppm at times. This
data immediately suggested foaming as a probability, and indeed foam was evi-
dent in the boiler water sample line discharge.
• Upgrade the sampling system so that accurate steam samples are available.
The current sampling taps have been placed in locations that do not pro-
vide representative samples.
• Upgrade analytical procedures so that plant personnel or an outside vendor
can determine all of the major constituents in the condensate return, boiler
feedwater, boiler water, and steam. For example, of the five major organic
chemicals produced at the facility, the operators can only analyze for phenol.
The procedure involves a colorometric determination, and the operators
Steam Chemistry 99
report that the sample often develops a different color than that outlined in
the analysis procedure. This indicates contamination by other compounds.
• If the upgraded sample data confirms contamination by the condensate
return, install a condensate polisher. The type and capabilities of the pol-
isher would of course be designed around the constituency of the conta-
minants in the condensate return.
Chapter 5
High-Purity
Makeup Water
Treatment
Introduction
As the preceding chapters have indicated, the water supplied to a steam gen-
erating unit must be pure in order to prevent serious corrosion or scaling. Ion
exchange had been the backbone of the high-purity water treatment industry, but
with the improvement of membrane technologies, a variety of methods now
exists to produce pure water. In most cases, some form of pretreatment is need-
ed ahead of the makeup system to protect it from fouling, scaling, or microbio-
logical contamination.
Pretreatment
Water has been called the closest thing to a universal solvent because it will
dissolve, at least to some extent, most compounds. Raw waters, surface waters in
FEED >
>
• Microbiocide feed
• Clarification/softening
• Media filtration
• Activated carbon filtration
Microbiocide Feed
Chlorine has been the principal microbiological control agent for many
years, and has proven to be the most economical biocide. However, use of chlo-
rine gas has been curtailed due to safety and environmental concerns. Popular
replacement feed chemicals now include sodium hypochlorite (liquid chlorine),
bromine, chlorine dioxide, and in some cases ozone. (UV light is also a possibil-
ity. Supplement 5-1 outlines a UV light arrangement at a manufacturing facility).
The mechanisms by which chlorine and other oxidizing agents disinfect water
are described in chapter 6. The oxidizing biocide should be added in great
enough strength to maintain a slight residual (0.1 to 0.5 ppm) throughout the
pretreatment process. This helps prevent the growth of microbes in equipment
downstream of the injection point. Chlorine and other oxidants will attack ion
exchange resins and some types of reverse osmosis membranes, making it nec-
essary to remove the oxidant ahead of these devices. Activated carbon is an excel-
lent chlorine scavenger, but if the system contains no carbon filters, some other
method of oxidant removal is required. Most commonly, a dehalogenating or
reducing agent (sodium sulfite or sodium bisulfite) is injected ahead of the dem-
ineralizer or reverse osmosis unit to protect the equipment.
Figure 5-2
Mixer Motor
Chemical Feed
Raw Water
<
>
Clarified Water
Top of Sludge
Blanket
>
<
Sludge
in the range of 0.5 to 1.5 GPM/ft2. Retention times may range from 1 to 4 hours.
The effluent quality of clarified water may vary depending on influent quality, but
with proper operation a clarifier can reduce calcium to 35 to 40 ppm (as CaCO3).
Turbidities of less than 10 nephelometric turbidity units (NTU) should be
expected.
While inground circular clarifiers are often specified for large applications,
clarification/softening of water at flow rates of up to 1000 GPM or even more can
be accomplished in individual or parallel package clarifiers. These units are usu-
ally of rectangular design. The water is introduced to the rapid mix zone, flows
to a flocculation compartment, and then enters the main body of the clarifier. To
improve settling, the main clarifier compartment is equipped with a series of
inclined plates or tubes. The water is forced to flow up these tubes to the outlet.
The interaction of particles flowing upwards with heavier particles settling down-
wards enhances the contact between newly formed floc and the heavier solids.
Residence times in package clarifiers may be much shorter than in circular clarifiers.
An attractive feature of package clarifiers is cost. Package clarifiers are much
less costly than in-ground units, where installation costs may be significant.
When only pretreatment of the makeup influent is required, package units can
be quite attractive. Systems of less than 600 GPM or so are usually shipped as
one unit. Even large systems can be shipped in just a few pieces and be field
erected.
Filtration
With the proper selection of coagulants and flocculants, a clarifier will
remove most suspended solids. However, enough solids remain to potentially
clog ion exchange resins or RO membranes. These particles are removed by pass-
ing the effluent through a filter, usually containing several media. A common
multimedia filter arrangement contains anthracite, sand, and garnet in graded
layers (Supplement 5-2). Coarse material is placed at the top to remove large par-
ticles, with fine media below to remove small particles. These filters may operate
at up to 5 GPM/ft2 and produce a water with turbidities of less than 1 NTU.
Proper filter operation is critical to the performance of high-purity makeup sys-
tems, especially reverse osmosis units.
A common guideline recommends backwashing of the filter when the dif-
ferential pressure increases by 10 psi over normal. Backwash flow rates of 14 to
16 GPM/ft2 are typical, although the manufacturer’s guidelines must be followed
during this very important step. A poor backwash can generate mudballs in the
filter, which will then impair subsequent operation. Water density changes with
temperature, so backwash flow rates may have to be adjusted at various times
during the year.
Calcium - 15 ppm
Magnesium - 5 ppm
Silica - 1 to 2 ppm
nologies that have begun to emerge and show promise include electrodialysis
reversal (EDR) and electrodeionization (EDI).
Ion Exchange
The ion exchange process was first truly developed in the early 1900s by a
German scientist, Gans, who used synthetic zeolites (sodium aluminum silicates)
to exchange calcium and magnesium ions for sodium. This discovery led to the
development of water softeners. Then, in the 1930s and 1940s, plastic synthetic
ion exchange resins became available, which led to the industry that we know
today.
The ion exchange process is dependent upon the efficiency with which
water can flow through the exchange media and contact exchange sites. To
obtain maximum efficiency, the media must have several properties. These
include:
resin and give the resin its porous structure. Resins of different porosity and
strength are useful in different environments. For example, macroporous resins
are more durable in heavy-duty applications. The gel-type resins typically con-
tain more exchange sites. Both types are heavily used in the water treatment
industry.
Macroreticular resins with large pores and no exchange sites have been
developed for removal of organic complexes from raw water. The resin functions
by allowing the organics to penetrate the bead, where they adsorb onto the poly-
mer chains. The resin is nonregenerable and is discarded when it reaches exhaus-
tion. These resins show promise for removing relatively small chain organics
from condensate, which are otherwise difficult to treat.
Exchange Groups
Once the beads have been synthesized, exchange groups are added. The
principal exchange groups are:
The exchange groups give the resins their character and define their perfor-
mance in service.
Because the resin behaves as a strong acid, it will split salts and separate
cations from their corresponding anions. A water containing the ions calcium,
magnesium, sodium, chloride, sulfate, bicarbonate, and silica will exit the vessel
as a dilute solution of hydrochloric, sulfuric, carbonic, and silicic acids. The
order of affinity the resin has for ions is Ca+2 > Mg+2 > Na+.
HCl Cl-
H2SO4 + CH2N(CH3)3+OH- ⇔ R+ SO4= + H2O (5.6)
H2CO3 CO3=
H2SiO3 SiO3-
Degasifiers
Alkalinity that passes through a cation bed is converted to carbonic acid
(H2CO3), which is essentially just hydrated carbon dioxide (CO2 • xH2O). The
load on the anion exchanger can be reduced if the carbon dioxide is removed
upstream. Forced-draft or vacuum degasification will accomplish this. In the for-
mer mechanism, water is cascaded down a series of trays while air is blown
upward. This process will reduce carbon dioxide concentrations to approxi-
mately 10 ppm although it does saturate the water with oxygen. A vacuum
degasifier pulls gases from the tower as the water flows down the trays. Even
lower CO2-removal efficiencies may be obtained, but a vacuum degasifier is more
expensive. The cost of a degasifier must be weighed against the value obtained
by decreasing the load on the anion exchanger. Caustic prices generally tend to
be high, so a degasification system may be warranted for even moderately alka-
line waters.
•Introducing the regenerant at the bottom of the vessel and using a block-
ing flow of water at the top to keep the bed in place
• Introducing service water at the bottom of the vessel and injecting regen-
erant from the top
Figure 5-10
Strong Acid Cation Exchanger Specifications
Polymer Structure – Polystyrene cross-linked with divinylbenzene
Total Capacity Hydrogen Form – 39.2 kgr/ft3
Swelling (NaH) – 5%
Standard Operating Conditions
Operation Rate Solution Minutes Amount
Service 1–5 GPM/ft3 Influent water
Backwash 3–5 GPM/ft3 Influent water 5–20 10–25 gals/ft3
Regeneration 0.2–0.8 GPM/ft3 0.5–5% H2SO4 30 4–10 lbs/ft3
Slow rinse 0.2–0.8 GPM/ft3 Decationized 60 20 gals/ft3
Fast rinse 1–5 GPM/ft3 Decationized 60 30 gals/ft3
Backwash expansion 50–75%
*Data courtesy of The Purolite Company, Bala Cynwyd, Pennsylvania
er’s literature must be used to determine resin capacity for the system. As Figure
5-10 shows, the capacity of the resin as shipped is 39.2 kgr/ft3. However, due to
the equilibrium that develops between hydrogen ions and cations in an exchange
bed, once the resin has been used its recoverable operating capacity is well below
the preoperational value. Resin manufacturers supply graphs and charts to cal-
culate resin performance, and these are used in the following steps.
The first step is to calculate the normal operating capacity of the resin. The
factors which affect this value are regenerant dosage, the percentage of sodium in
the feed, and the ratio of calcium and magnesium in the feedwater. All can be
determined from Figure 5-11. Although regenerant concentrations are often
talked about in percent, it is the actual mass of acid introduced to the bed that
determines regeneration capacity. If more acid is supplied, more cations will be
removed. The units used for this calculation are pounds of acid per cubic foot of
resin. Although any dosage may be used, the optimum range is typically within
5 to 6 lbs/ft3. For this example, 5 lbs/ft3 is the regenerant dosage.
The header at the top of the graph in Figure 5-11 shows that the chart
applies to waters with 0 to 50% sodium. The ratio of sodium to other cations in
the water is important, because sodium elutes from the bed more easily than cal-
cium and magnesium. This effect is double-edged. As the sodium concentration
of the influent increases at the expense of hardness ions, the available capacity of
the resin increases because sodium is more easily regenerated. However, sodium
leakage during the service run will increase, because sodium’s weaker affinity for
the resin allows some of it to continually escape. Countercurrent regeneration
can mitigate this effect, as will be seen.
The graph also shows lines for magnesium and calcium. Regeneration effi-
ciency is dependent on this ratio, because magnesium elutes more easily than cal-
SAC Resin, Base Operating Capacity: 0–50% SAC Resin, Base Operating Capacity: 85% Sodium
Sodium in Influent. Courtesy of the Purolite Company, in Influent. Courtesy of the Purolite Company, Bala
Bala Cynwyd, PA. Cynwyd, PA.
cium. The lines indicate resin capacities for various magnesium/calcium ratios.
The base operating capacity can now be predicted. Influent sodium makes
up 50% of the cations, the ratio of calcium to magnesium is 1 to 1, and the regen-
erant dosage is 5 lbs/ft3. The point on the y-axis that corresponds to these con-
ditions is 14.2 kgr/ft3. Had the ratio of magnesium to calcium been higher, the
base operating capacity would have been greater. Similarly, Figure 5-12 reveals
that if the sodium ratio been higher, with all other factors the same, the predict-
ed capacity would have increased. (In this case, if sodium made up 85% of the
cations the base operating capacity rises to 15.7 kgr/ft3.)
Correction now must be made for the alkalinity of the influent water. As
water passes through a cation bed, it acquires hydrogen ions as the cations are
exchanged. If the cations are associated with weak bases such as chloride and
sulfate, then the hydrogen ions behave as a strong acid and tend to regenerate the
lower portion of the bed. If, however, the cations are associated with alkalinity,
the hydrogen ions combine appreciably with this stronger base and lose most
regenerant capabilities. This has an effect on capacity calculations, and is taken
SAC Resin Sodium Leakage, Co-Current SAC Resin Sodium Leakage, Co-Current
Regeneration, 5 lbs/ft2. Courtesy of the Regeneration, 6 lbs/ft2. Courtesy of the
Purolite Company, Bala Cynwyd, PA. Purolite Company, Bala Cynwyd, PA.
ent increases, this effect diminishes because the hydrogen ions are more attracted
to the stronger base. Sodium leakage, therefore, becomes a function of the influ-
ent sodium and alkalinity concentrations. Figure 5-15 can be used to calculate the
projected sodium leakage from the cation bed. With sodium comprising 50% of
the cations and alkalinity 40% of the anions, and a regenerant dosage of 5 lbs/ft3,
the projected sodium leakage during the service run is approximately 2.2 ppm.
Figure 5-16 shows that with a regenerant dosage of 6 lbs/ft3, the sodium leakage
could be reduced to around 1.7 ppm.
These calculations were all based on co-current regeneration. Figures 5-17
and 5-18 show the base capacity and projected sodium leakage from the same
resin with countercurrent regeneration. Base operating capacity is increased from
14.2 kgr/ft3 to 15.2 kgr/ft3. Using the previously established correction factors for
alkalinity and temperature, the operating cation exchanger capacity becomes
Figure 5-17
15.8 kgr/ft3. This value will be used in subsequent cation exchanger vessel siz-
ing. More important is the effect that countercurrent regeneration has on sodium
leakage. The countercurrent process reduces sodium leakage during the service
run from 2.2 ppm to less than 40 ppb. This can have a significant impact on
anion exchanger operation.
The reduction in sodium leakage due to countercurrent regeneration is eas-
ily explained. The sodium band forms in the lower portion of the resin. During
co-current regeneration, sodium not removed by the regenerant remains at the
bottom of the vessel. Therefore, some sodium ions that detach from exchange
sites have only a short distance to travel before reaching the exchanger effluent.
In countercurrent systems, the reverse flow of the regenerant moves sodium ions
further up in the bed. The bottom of the bed becomes more fully regenerated.
Upon resumption of the service cycle, sodium leakage is low.
Figure 5-18
Figure 5-19
Strong Base Anion Exchanger Specifications
Polymer Structure – Polystyrene cross-linked with divinylbenzene
Total Capacity Chloride Form – 28.3 kgr/ft3
Swelling (ClOH) – 20%
Standard Operating Conditions – Countercurrent Regeneration
Operation Rate Solution Minutes Amount
Service 1–5 GPM/ft3 Decationized water
Backwash 2–3 GPM/ft3 Decationized water 5–20 10–25 gals/ft3
Regeneration 0.25–0.5 GPM/ft3 3–6% NaOH 30–60 3–10 lbs/ft3
Slow rinse 0.25–0.5 GPM/ft3 Decationized water 30 (approx) 15–30 gals/ft3
Fast rinse 1–5 GPM/ft3 Decationized water 20 (approx) 25-45 gals/ft3
Backwash
expansion 50–75%
Figure 5-20
45˚ C (120˚ F)
SBA Resin Base Operating Capacity: 120˚F Regenerant. Courtesy of the Purolite Company, Bala
Cynwyd, PA.
The actual silica leakage can now be determined from Figures 5-22, 5-23,
and 5-24. Silica makes up 2.5% of the total anions. At a caustic regenerant level
of 5 lbs/ft3, the base silica leakage is projected to be 5 ppb or less. (Since the
graph is not explicit at this low level, 5 ppb will be assumed.) The actual silica
leakage is then determined by the equation:
where CF1 is a correction factor for sodium leakage from the cation bed and
CF2 is a temperature correction factor.
At this point it starts to become apparent how sodium leakage from a cation
exchanger can affect silica leakage. The previous cation exchanger calculations
showed that sodium leakage from a co-currently regenerated bed would be
approximately 2.2 ppm, while from a countercurrently regenerated bed would be
under 40 ppb. Figure 5-23 outlines correction factors for silica leakage based on
influent sodium. At 2.2 ppm sodium, the correction factor is 0.75, but at 40 ppb
of sodium leakage the correction factor rises to nearly 0.96.
The temperature correction factor as outlined in Figure 5-24 gives a value of
0.9 for 120˚F regenerant solution. The final projected silica leakage is:
SAC Resin Sodium Leakage Correction Factor. SAC Resin Regenerant Temperature Correction
Courtesy of the Purolite Company, Bala Cynwyd, PA. Factor. Courtesy of the Purolite Company, Bala
Cynwyd, PA.
Ali Sadeghi Digital Library
05.qxd 8/8/06 1:56 PM Page 125
Figure 5-25
Weak Acid Cation Exchanger Specifications
Polymer Structure - Acrylic divinylbenzene
Total Capacity Hydrogen Form - 91.6 kgr/ft3
Swelling (H Na) - 90%
(H Ca) - 20%
Standard Operating Conditions
Operation Rate Solution Minutes Amount
Service 1–5 GPM/ft3 Influent water
Backwash 3–5 GPM/ft3 Influent water 5–20 15–60 gals/ft3
Regeneration 1–2.5 GPM/ft3 0.5–1% H2SO4 30–45 100–120% of theory
Slow rinse 1–2.5 GPM/ft 3 Decationized 15 15–30 gals/ft3
Fast rinse 1.25–5 GPM/ft Decationized
3 22–45 gals/ft 3
WAC Resin Base Operating Capacity. Courtesy of WAC Resin Temperature Correction Factor.
the Purolite Company, Bala Cynwyd, PA. Courtesy of the Purolite Company, Bala Cynwyd, PA.
much closer to the design capacity than is the case with SAC resins, and reflects
the efficiency with which WAC resins may be regenerated.
Weak base anion resins remove weak bases as their acid conjugates (sulfu-
ric acid, hydrochloric acid). The resins do not remove alkalinity. For this exam-
ple then, a WBA exchanger would remove the 120 ppm of chloride and sulfate.
Figure 5-29
Resin VolumeWeak
and Base Anion Exchanger
Vessel Diameter Specifications
Calculations
Polymer Structure - Polystyrene cross-linked with divinylbenzene
For Capacity
Total a demineralizer to3 perform properly, certain hydraulic requirements
- 28.3 kgr/ft
must be met.Operating
Standard The ion Conditions
exchange process is not only governed by chemistry but
alsoOperation
by kinetics and maximum utilization
Rate Solution of the ion exchange
Minutessites. Figures
Amount5-10,
5-19, 5-25,
Service and 5-29 illustrated
8–40 BV/h the manufacturer’
Decationized water s recommended flow rates for
the Backwash
exchange resins. 5–7Table
m/h 5-3 is a supplement
Decationized waterto this data,
5–20 and provides
1.5–4 BVgener-
al hydraulic guidelines.
(39–77˚F)These guidelines are integral to the following calcula-
tions.
Regeneration 4 BV/h 2–4% NaOH 30 2–6 lbs/ft3
From the
Slow rinse calculations
4 BV/h performed above, the
Decationized water capacity
20 of the SAC
1–5 resin
BV was
determined
Fast rinse to be 1615.8
BV/hkgr/ft3. Decationized
The influentwater
water quality
15 contains4200 BV ppm of
cations
BV =as CaCO3,
bed volumesproduction is 100 GPM, and the service run time is 16 hours.
OneBackwash
grain per gallon is
Expansion equivalent to 17.1 ppm of ions as CaCO3. Thus, the
35–50%
quantity of ions
*Data courtesy to Purolite
of The be removed
Company,is:
Bala Cynwyd, Pennsylvania.
Figure 5-30
lbs./cu.ft.
g/l
WBA Resin Base Operating Capacity. Courtesy of the Purolite Company, Bala Cynwyd, PA.
200 ppm ÷ 17.1 ppm per gr/gal x 100 GPM x 960 minutes = 1,122,807
grains (5.10)
3
The volume of SAC resin (capacity 15.8 kgr/ft ) needed to treat this quanti-
ty of water is 71 ft3. Most resin suppliers recommend that a safety factor of 0.85
to 0.9 be incorporated into resin volume design. Using a 0.9 factor, the final
design volume becomes:
A 2 GPM/ft3 volumetric flow rate as suggested in Table 5-3 gives a bed depth
of 48 inches with a tank diameter of 5 feet. This is within good design proce-
dures.
Calculations for the SBA resin are similar. The capacity for the resin was
determined to be 13.8 kgr/ft3. (Anion resins usually have less capacity than
Table 5-3
cation resins). Including silica, the water contains 210 ppm of anions as CaCO3.
The quantity of anion resin required is 94 ft3 including the 0.9 design factor. At
this point, a choice is possible. A 48-inch bed depth could again be used, which
would increase the diameter of the vessel to 5 feet. Or, the diameter of vessel
could be set at 5 feet to correspond with that of the cation vessel, and the bed
depth could be increased to 58 inches. Both would be satisfactory.
Similar calculations can be performed for WAC and WBA resins. Once the
resin capacity is determined, the volume can be calculated based upon the quan-
tity of ions to be removed.
Backwash Requirements
When resin exhausts it is backwashed before regeneration to remove resin
fines and particles that were filtered by the bed. Backwash water is introduced at
the bottom of the bed to lift and mix the resin. For backwashing to be effective,
the resin must be thoroughly agitated and expanded. This requires extra vessel
height, known as freeboard. Figure 5-10 shows that the manufacturer recom-
mends a bed expansion of 50 to 75% for the SAC resin. The bed depth has
already been calculated at 48 inches, so a 75% volume addition increases the
straight side-height of the vessel to 84 inches.
Figure 5-19 also recommends 50 to 75% bed expansion for the SBA resin.
Since the bed height in this vessel was calculated to be 58 inches, a 75% bed
expansion increases the straight-side height to 101 inches.
The efficiency of bed expansion is very dependent on the temperature of the
water. Colder water, being more dense, expands the bed more easily. Water flow
rates must be carefully regulated to prevent resin from being discharged through
the service water influent distributor. Resin manufacturers will supply charts and
equations for calculation of the backwash flow rate based on temperature.
Backwash times typically range from 5 to 20 minutes, with 15 minutes being
an average. This allows 1 to 3 bed volumes to pass through the vessel.
Regeneration Requirements
Many aspects of regeneration have already been discussed. Regeneration is
carried out at a much slower flow rate (0.2 to 0.8 GPM/ft3) than either the ser-
vice or backwash processes in order to give the regenerant time to be effective.
Regeneration of cation exchangers with sulfuric acid is often performed in a step-
wise process to prevent formation of calcium sulfate deposits within the resin.
The regenerant is initiated at a 2% concentration and is raised to 4% and perhaps
even 6% later on in the process. SBA regenerant strength is usually 4% through-
out the process. Regeneration proceeds more quickly in a cation bed than an
anion bed. Cation regeneration may be complete in only half an hour, whereas
anion regeneration may take an hour or more.
Regenerant concentrations for weak acid and weak base exchangers are sig-
nificantly lower due to each resin’s high efficiency of regeneration. In systems
where WAC and WBA exchangers are used in combination with SAC and SBA
exchangers, the waste regenerant from the strong exchangers may be used to
regenerate the weak resin exchangers. Recommended regenerant dosages for
SAC and SBA resins are 4 to 10 lbs/ft3 of acid and caustic, respectively, but for
weak acid and weak base resins, they may be as low as 2 lbs/ft3.
All resins undergo swelling as they convert from one ionic form to another.
SAC and SBA resins both swell during regeneration, but the change is usually
minor (5 to 8% for SAC resins and 10 to 20% for SBA resins). WAC and WBA
resins swell during the service run, and these changes are much more dramatic,
with swelling up to 100% being possible. In fact, vessel design for weak acid and
weak base resins may be dictated more by swelling than for backwash volume.
Rinsing
Upon completion of regeneration, the exchange resins are rinsed to remove
the regenerant. The first step is a slow rinse at the same flow rate as the regener-
ant. This allows the remaining regenerant to remove ions as it is being rinsed
from the bed. The slow rinse may take anywhere from 30 to 60 minutes. At the
end of the slow rinse, the water flow rate is increased, often to the service rate,
for the fast rinse. The fast rinse duration may also be 30 to 60 minutes. The fast
rinse puts several bed volumes of water through the resin.
Mixed-Bed Polishing
Mixed-bed (MB) exchangers contain both SAC and SBA resins intimately
intermixed. The primary purpose of a MB exchanger is to polish the already puri-
fied water. MB quality water is needed for high-pressure boiler makeup and at
some manufacturing facilities, particularly in the electronics and semiconductor
industries.
Because a MB exchanger has such a light ionic load, flow rates can be
increased. A typical cross-sectional flow rate is 20 GPM/ft2. The mixed bed is
sized to provide very long term operation with infrequent regenerations. This can
be done without excessive quantities of resin because the ion loading is so low.
Consider a SAC/SBA demineralizer with an effluent that contains 1 ppm of
cations as CaCO3. If the cation exchange resin from the earlier example is used
in the MB exchanger, 50 cubic feet could theoretically process 13,500,000 gal-
lons of water. At 100 GPM, the MB cation resin would not exhaust for three
Figure 5-31
CONDUCTIVITY (micromho)
SODIUM (ppm)
>
HARDNESS (ppm)
0 <
SERVICE CYCLE >
months. (In actuality, the higher flow rate can reduce the operating capacity
somewhat. Even so, very long run times are still normal.)
Although regeneration of mixed-beds is performed infrequently, it is a very
important process and must be handled with care, since two resins are involved.
Regeneration is usually performed in the exchange vessel. Anion resin is lighter
than cation resin, and when the MB resin is backwashed, the process causes the
two resins to settle into distinct layers. By design, a collector system is installed
at the cation/anion resin interface. During regeneration, acid is introduced below
the cation resin and flows upwards. Caustic is introduced above the anion resin
and flows downwards. The waste regenerant from each is collected at the inter-
face. It is very important that the resins settle properly and that the division
between them occurs at the central collector. If cation resin remains in the anion
zone, it will absorb sodium from the caustic regenerant. Conversely, anion resin
will pick up sulfate from the acid regenerant. A technique for providing good
separation is to design the system for and include a small amount (10% or so) of
inert resin that has a density between that of the cation and anion resin. The inert
resin will settle at the interface and provide separation between the active resins.
The mixed-bed regeneration process is similar to that for cation and anion
resins with regard to acid and caustic concentrations. Once the resin has been
rinsed, it is remixed with air. The air remix process is very important to re-estab-
lish the heterogeneity of the cation and anion resins. Typical guidelines suggest
an air flow rate of 7 to 10 SCFM for 15 minutes.
Figure 5-33
Distributors
The flow rate of water or regenerant across the resin should be uniform to
prevent channeling or areas of high and low flow. Distributor design is most
important. Distributors are usually supplied in a lateral type arrangement so that
the spray uniformly covers the entire surface area of the resin. The spray nozzles
and piping should be designed to provide the same quantity of water from each
opening.
Vessels
Ion exchanger vessels are rubber lined to prevent corrosion. The rubber lining
must be installed properly and spark tested to eliminate any holidays in the liner. If
liquid gets behind the liner, where it can attack a localized area of the vessel shell,
shell failure may result. Vessels must also be fabricated in accordance with pres-
sure-vessel design codes. This becomes particularly important for on-line conden-
sate polishers, which may operate at several hundred pounds of pressure.
Valves
The valving arrangement on a demineralizer can be fairly complicated.
Valves automatically open and close during changes between service, regenera-
tion, and rinse. Positive shutoff valves such as block valves are recommended so
that process fluids do not infiltrate unwanted areas.
Materials
Materials for piping, valves, and tanks must be corrosion resistant. PVC pip-
ing has been used on some demineralizers, although it can break if it is acciden-
tally hit with a piece of equipment. Stainless steel (316) is much more durable
and can handle most chemical solutions. (Stainless steel is not recommended if
hydrochloric acid is used as the cation regenerant.) Alloy 20 provides superior
resistance to dilute sulfuric acid solutions, although it is very expensive. Both
lined and unlined tanks have been used for concentrated acid and caustic stor-
age. A liner will help prevent corrosion.
Packed-Bed Demineralizers
The latest generation of demineralizers include systems based on packed-
bed technology. In these demineralizers, the vessel is almost completely filled
with resin, with perhaps only 10% freeboard. Two processes are available, both
based on countercurrent operation. In one, the process water flows from top to
bottom of the vessel and the regenerant from bottom to top. Unlike convention-
Figure 5-34
Short-Bed Ion Exchange Unit. Photo courtesy of Eco-Tec, Inc., Pickering, Ontario, Canada.
The obvious advantage of this system is its small size. For facilities with lim-
ited space, or where a demineralizer must be retrofitted into a confined area,
these systems can be very advantageous. Water quality and regeneration efficien-
cy are also reported to be very good. One drawback to these systems is that with
zero freeboard, the resins cannot be backwashed. The influent must be well fil-
tered to remove suspended solids.
Reverse Osmosis
Reverse osmosis technology has developed substantially. Systems have
become more reliable, and membrane performance has improved to a level where
RO systems can routinely remove 99% or greater of the influent dissolved solids.
Reverse osmosis has become particu-
Figure 5-35
FLOW
larly popular as a retrofit ahead of an
existing demineralizer, or as part of a
combined system, i.e., RO plus mixed
bed, for new installations. The economics
are particularly favorable when reverse
osmosis is used to pretreat high TDS
waters. The RO greatly reduces ion load-
ing on the demineralizer, which lowers
Conventional Filtration. the associated regeneration frequency and
chemical costs. Although reverse osmosis is considered to be a filtration tech-
nology, the process is more complex than conventional filtration. In an ordinary
depth or weave filter, water flows perpendicularly to the filter (Fig. 5-35).
Particles are removed throughout the depth of the filter. When the differential
pressure between the inlet and outlet becomes too large, the filter is replaced. In
an RO system, water flows parallel to the membrane surface. Applied pressure at
the influent forces a portion of the water through the membrane as it passes from
one end to the other. Solids are swept along with the water that does not pass
through the membrane. This water, which becomes increasingly concentrated,
flows to the end of the pressure vessel and is discharged (Fig. 5-36). This is
known as crossflow filtration.
Figure 5-36
FEED REJECT
PERMEATE
Cross-Flow Filtration.
RO Membrane Design
Two types of RO membrane design are most common today: hollow-fiber and
spiral-wound. In hollow-fiber configurations, a membrane element consists of
many fibers bundled together. Each fiber individually treats a portion of the feed-
water. The purified water (permeate) and concentrated water (reject) are collected
at the end of the vessel. Hollow-fiber systems have not proven popular for brack-
ish water systems and have not been extensively used for most of the industries
addressed by this book. By far, spiral-wound membrane configurations are more
common. The remainder of this discussion is based on spiral-wound systems.
Spiral-wound membranes are manufactured in flat sheets, which are wound
around a central core to produce a membrane element (Fig. 5-37). Several elements
are placed in series and are sealed in a pressure vessel (Fig. 5-38). Feedwater is
applied to the end of each element, and it flows along the spacers to the opposite
end. Permeate passes to the central core of the element, while the reject is collect-
ed and discharged at the element end cap.
RO Membrane Material
Two types of material are common for RO membrane configuration: cellu-
lose acetate/triacetate (CA) and polyamide. The latter material is typically layered
with other membranes for support. These layered polyamide membranes are
known as thin-film composites (TFC).
CA membranes were the first to be developed. The material provides advan-
tages and disadvantages, as compared to its polyamide TFC counterpart. The
important advantages of CA material include:
• The membrane is chlorine tolerant up to about 1.0 ppm and works well
in systems where microbiological fouling must be controlled.
Figure 5-38
FEEDWATER PRESSURE VESSEL PERMEATE
INLET REJECT
MEMBRANE MEMBRANE MEMBRANE >
> ELEMENT ELEMENT ELEMENT >
Figure 5-37
Disadvantages include:
• Feed pressure is higher than for a TFC membrane. Typical feed pressures
of CA systems for the applications described in this book range from 200
to 400 psig.
Disadvantages include:
RO Pretreatment
Reverse osmosis units are very fine filtration systems, which concentrate dis-
solved solids. This makes them subject to scaling, fouling, or chemical attack.
The effects of oxidants and pH have been mentioned above. Of equal concern is
fouling. Suspended solids are not compatible with RO membranes, although
reverse osmosis can be an excellent method for removing contaminants such as
colloidal silica, which passes through a demineralizer. The makeup water pre-
treatment methods mentioned at the beginning of this chapter will often produce
water suitable enough for feed to a RO. A rule of thumb says that waters with a
turbidity of less than 1 NTU will be suitable. This is only a general guideline.
Another measurement for determining the fouling potential of water is known as
the Silt Density Index (SDI). SDI is a measure of the effect suspended solids have
on water flowing through a filter. A SDI value of less than 5 usually indicates that
fouling will not be a problem. Supplement 5-5 outlines the SDI procedure and
calculations.
RO Design
A reverse osmosis unit has been called nothing more than a high-pressure
pump, some pressure vessels, and pipe. In truth, the operation is more compli-
cated than this. Spiral-wound membrane elements can come in different sizes.
The most popular size is 8 inches in diameter by 40 inches in length. These are
loaded in series into a pressure vessel (refer again to Fig. 5-38), with four, five,
or six elements per vessel being most common. Each element can pass a certain
amount of water, and this volume is usually measured in gallons per day (GPD).
Common values for 8 in. x 40 in. elements range from 4000 to over 13,000 GPD.
The rate at which water passes through the membrane is known as the flux and
is measured in gallons per square foot per day (GFD). Common flux rates for
water sources have been reported as follows:
For normal surface and groundwaters, each pressure vessel will produce
about 50% purified water (permeate) and 50% concentrated water (reject or con-
centrate). This does not seem very efficient. However, the concentrate is often
still pure enough to be treated again at another 50/50 split to produce 75%
capacity. Sometimes even the second concentrate can be treated to give an over-
all RO output of 87.5%.
With this background information, we will look at a simple reverse osmosis
system designed to produce 300 GPM water with 75% recovery of the influent.
Conditions are as follows:
• Surface water
• SDI < 5
• Five membranes per pressure vessel
• Membrane capacity - 10,000 GPD
12 ppm TDS. If this RO were retrofitted ahead of the hypothetical SAC/SBA dem-
ineralizer calculated earlier, the resin run lengths could be greatly extended (a
general rule of thumb is 20 times). Retrofitting is in fact what many plant man-
agers and engineers are now doing. The driving factor is economics. It is not
impossible for regenerant chemical costs to exceed $100,000 per year for even a
moderately sized (200 GPM) demineralizer. Operating costs for the RO pump
(power), perhaps a small amount of antiscalant feed, periodic chemical cleaning
of the RO membranes, and (reduced) regenerant chemical feed to the deminer-
alizer may be less than a fourth of the normal chemical regenerant cost. Payback
time for the RO might be as short as two years. One factor not included is the
cost for membrane replacement. Membranes typically last from three to seven
years, although longer lives have been recorded. Membrane replacement costs
may be a third of the original price of the RO.
A technique that is gaining acceptance for new high-purity water installa-
tions is reverse osmosis followed by mixed-bed polishing. This arrangement per-
forms very similarly to an SAC/SBA/MB demineralizer.
Various articles and technical papers have been presented over the last few
years that attempt to give a rule of thumb guideline for the water quality at which
RO is favored over SAC/SBA, when installed ahead of a mixed-bed. Although the
topic is debatable, it is evident that the economics have improved for RO.
Around 1990, it was suggested that an influent TDS of approximately 350 ppm
or higher made RO economical. Reports later in the decade indicated that this
level had dropped to as low as 150 ppm. The economics must be examined on
a case-by-case basis.
A technique for producing relatively high-purity water without the use of
any demineralization is double-pass RO. (Pass and stage should not be con-
fused.) In this system, the permeate from the first stage is sent to another set of
membranes for further purification. This process is capable of reducing solids
concentrations to less than 1 ppm. The concentrate from the second pass is recir-
culated to the influent, so the overall efficiency of the reverse osmosis system is
not diminished by adding this second pass. Two-pass RO permeate is potential-
ly suitable for feed to low- and medium-pressure boilers.
RO Components
Various components besides the membranes make up a reverse osmosis sys-
tem. They include the pump, pressure vessels, piping, and instruments (Fig. 5-
39). The following sections outline some of the most important aspects of this
equipment.
Pumps. Centrifugal, multistage pumps have proven best for RO applica-
tions. They can easily generate the pressures needed, although pressures have
Figure 5-39
RO Alarms
A number of circumstances might occur that could cause damage to a
reverse osmosis system, particularly the membranes. Therefore, RO systems are
typically equipped with a number of alarms or automatic shutdown devices to
protect the equipment. Some of these have been hinted at in the previous sec-
tion. A discussion of the most important ones follows.
High permeate pressure. Membranes are designed to function with flow in
one direction. Backpressure applied to the membranes from the permeate dis-
charge could tear the leaves of the membranes from their support. The system
can be designed to shut down if the permeate pressure reaches a certain limit.
The line can also be equipped with a relief valve for additional protection.
High membrane pressure. Backpressure may be exerted on the membrane
if the reject valve should somehow be closed. This condition must be corrected
quickly. An alternate alarm that would serve this and another purpose is low
reject flow, which would obviously occur if the valve was shut. Low reject flow
is serious because a decrease in flow rate reduces the ability of the concentrate to
sweep solids from the membrane surface. Scaling may result.
High/low influent pH or high ORP. These alarms indicate a problem with
the chemical feed systems that were installed to protect the membranes.
High permeate conductivity. A sudden increase in conductivity indicates
performance problems with one or more of the membranes. The increased salt con-
centration in the permeate could have serious effects on downstream equipment.
Low/high pressure RO feed pump pressure. Low pressure indicates a
problem with the pump. High pressure indicates a possible obstruction or flow
problem in the RO system.
RO Cleaning
Periodically, membrane performance may decline enough to require a chem-
ical cleaning. Because the fouling is on the concentrate side of the membrane, the
chemical cleaning solution is injected through the reject line. The chemical is
mixed up in an external tank and is circulated with a low-pressure pump over
the membranes. Usually, the cleaning system piping network is set up so that
only a portion of the pressure vessels are cleaned at one time. The recommend-
ed flow rate for an 8-inch diameter pressure vessel is 40 GPM.
Common cleaning chemicals include citric acid for calcium carbonate
removal, a carbonate/EDTA chelant for calcium sulfate, and alkaline phos-
phate/EDTA solutions for organically fouled resins. The strength of solution is
usually between 1 and 2%. The solutions work better when warm, so the mix-
ing tank is equipped with a heater to raise the temperature of the cleaning agent
(105˚F is recommended).
For clean up of microbiological foulants, recommended treatments include
a 10 ppm sodium hypochlorite solution for CA membranes and a 400 ppm per-
acetic acid solution for TFC membranes.
Figure 5-40
FEED
>
#1 #2 #3 #4 #5 #6
CATHODE ANODE
AN. AN. AN. AN. AN. AN.
The cost of an EDR system is roughly 30 to 50% higher than that of a reverse
osmosis unit. EDR is not extensively used for boiler makeup at this time, being
more popular in the potable water industry. EDR has also shown promise as an
ion concentrator for zero discharge systems.
Electrodeionization
The electrodialysis process has been carried a step further with the develop-
Figure 5-41
FEED
>
#1 #2 #3 #4 #5 #6
CATIONIC N ON C CATIONIC
C O C ANIONIC C O C N ON C CATIONIC
ANIONIC C O C
MEMBRANE MEMBRANE MEMBRANE
MEMBRANE MEMBRANE MEMBRANE MEMBRANE
MEMBRANE MEMBRANE MEMBRANE
• The addition of ion exchange resin greatly improves the process with
regard to silica removal.
• The electric field generates H+ and OH-, which regenerate the resin on-
line. This eliminates acid and caustic regeneration.
An EDI system can produce water of a mixed-bed quality. The units are pri-
marily used as polishers, and a system composed of a reverse osmosis unit fol-
lowed by an EDI unit might be ideal for generating high-purity water with no
chemical regeneration. EDI systems must be protected against the intrusion of
suspended solids because the resin cannot be backwashed as is possible in con-
ventional ion exchange units.
Conclusion
Production of makeup water suitable for feed to steam generating units
requires many different processes. A number of options are available to produce
high-purity water, and some treatment method can almost always be developed.
As always, proper design and operation are essential for good operation.
Supplement 5-1
UV-Light Disinfection
Figure 5-42
Makeup To RO Reject To
Treatment System Reclaim System
V
RO
Media V Storage Heat UV (3 Units) UV Storage
V
V V V V
Filter Tank Exchanger 254 nm Two-Stage 254 nm Tanks
V
254 nm Polisher Demineralizer Tank Filter 254 nm
V
V
Purified Water RO Water
To Plant To Plant
* Supplemental Feed
Supplement 5-2
Multimedia Filtration
Supplement 5-3
Activated Carbon
Supplement 5-4
Sulfuric Acid and Caustic Specifications
Listed below are maximum contaminant guidelines for sulfuric acid and
caustic purchased for demineralizer regeneration.
Sulfuric Acid
Iron - 20 ppm
Supplement 5-5
Silt Density Index
The silt density index (SDI) is a measure of the fouling potential of the water.
A simple test is available to calculate SDIs. In the test, water is filtered through a
0.45 micron membrane at a pressure of 30 psig. Measurement is taken of the
time for 500 ml of water to pass through the filter. The stream is allowed to con-
tinue flowing through the filter for a set period of time (5, 10, or 15 minutes),
after which the filtration rate is again measured. The SDI is calculated by the fol-
lowing equation:
ti = 34 seconds
tf = 66 seconds
T = 15 minutes
SDI = 100 x (1-34/66)/15 = 3.2
Listed below is a simple BASIC program that will automatically calculate the
SDI when the user inputs values for ti, tf, and T. This program can very useful if
the plant chemist must routinely monitor silt density indices.
SDI.BAS
10 CLS:LOCATE 5,1
20 PRINT “SILT DENSITY INDEX CALCULATION”
30 PRINT
40 PRINT “ENTER THE TIME IN SECONDS TO COLLECT 500 ML OF SAMPLE.”
50 INPUT TI
60 PRINT
70 PRINT “ENTER THE TIME N MINUTES BETWEEN READINGS. 15 MINUTES IS STAN-
DARD.”
80 INPUT T
90 PRINT
100 PRINT “ENTER THE TIME IN SECONDS TO COLLECT 500 ML DURING THE FINAL
SAMPLING.”
110 INPUT TF
120 PRINT
130 SDI=100*(1-(T1/TF))/T
140 PRINT “THE SDI AT “;:PRINT USING “##”;T;:PRINT “ MINUTES = “;:PRINT
USING “##.#”;SDI
Chapter 6
Cooling Water
Chemistry
Introduction
Cooling of fluids is an essential process at power-generation and industrial
plants. Control of cooling water chemistry is very critical in preventing corrosion
and fouling, and ensuring equipment reliability.
The most important uses for cooling water include:
The volume of cooling water required for these processes, especially steam
condensation, is often quite large. (Sufficient cooling water flow to steam con-
densers allows the condensed steam to reach its lowest temperature, which is
important with regard to the thermodynamic concepts outlined in Supplement
2-1.) High cooling water flow rates usually prohibit the mechanical purifying
techniques (demineralization, RO) that are used for smaller purposes such as
boiler makeup production. Yet, some form of treatment is required to protect
equipment. The following sections discuss types of cooling systems, water qual-
ity variations between different sources of supply, and treatment methods.
Changing environmental regulations and improvements in cooling water treat-
ment chemicals have added a complexity to treatment selection.
Cooling Systems
The type of cooling system most suitable for a process depends upon
process operation, flow requirements, availability and quality of water, and envi-
ronmental requirements regarding discharge. Most cooling systems belong to one
of the following categories:
• Once-through
• Open recirculating (evaporative cooling towers)
• Closed recirculating
Dry cooling systems for steam condensation are becoming more popular,
especially at facilities where makeup water is scarce or where liquid discharges
are restricted or banned. A detailed discussion of dry cooling systems is outside
the scope of this book.
Once-Through Systems
Once-through systems are most often used to cool turbine exhaust in a con-
denser and provide service water to the plant. Flow rates range from tens of thou-
sands to hundreds of thousands of gallons per minute. The source of supply for
once-through systems is usually a lake, river, or ocean due to the large quantities
of water needed.
The greatest advantage of once-through cooling is that the water does not
become concentrated as it does in a cooling tower, and thus has a much lower
scale-forming or corrosion potential. (Sometimes, as Case History 6-1 illustrates,
unusual events can upset the somewhat steady-state conditions of a once-
through source.) Another advantage is that the open body of water which serves
as the supply will usually be cooler than water from a corresponding cooling
tower. This improves the thermodynamic efficiency of the condenser.
The primary disadvantage of once-through cooling is that the water is
returned to the source at a higher temperature. Increasingly stringent regulations
on thermal discharge and the effects on aquatic organisms are restricting new
construction of once-through systems, and it may take a variance from environ-
Cooling Towers
Figure 6-1
EVAPORATION & DRIFT
Figure 6-2
Cooling Tower. Photo courtesy of the Marley Cooling Tower Company, Overland Park, KS.
Figure 6-3
>
AIR
OUT
CIRCULATING
WATER IN CIRCULATING
WATER IN
FILL
<
AIR AIR
> <
IN IN
>
TO
BASIN CONDENSER
Dual-Entry, Crossflow Cooling Tower Arrangement.
The heart of an open system is the cooling tower (Fig. 6-2), which cools
water by cascading it through air. Modern cooling towers can basically be divid-
ed into two categories, mechanical draft (air forced through the tower) and nat-
ural draft (air flows naturally through the tower). Mechanical draft towers can in
turn be divided into two other categories, crossflow and counterflow (single or
dual entry), in which the air is either pushed (forced-draft) or pulled (induced-
draft) through the tower. Figures 6-3 and 6-4 illustrate the basic outline of dual-
entry crossflow and counterflow towers, respectively. The principle of operation
is relatively straightforward. Warm cooling water is sprayed into flowing air,
which absorbs heat. The warmed air is ejected from the tower, while the cooled
water falls to a basin where it is recirculated. The falling water is broken up by
fill material in the tower, which enhances air/liquid contact and heat transfer.
Both the crossflow and counterflow design have advantages and disadvan-
tages. Air pressure drop and water pump head requirements may be greater in a
counterflow tower, but cooling efficiency is normally a bit higher as well.
Maintenance on crossflow towers is often easier, but these towers, being shorter
in height, are more susceptible to recirculation of discharged air to the air intake.
Fan location also has an effect on tower performance, particularly on power
usage and durability of the fan materials. In an induced-draft tower, where a fan
is located at the air outlet of each cell, the fan blades, shafts, and other compo-
nents are constantly subjected to very humid air flows containing trace amounts
of suspended solids. The tower structure must be strong to handle the fan vibra-
tion, and the elevated location of the fans can hamper maintenance. Forced-draft
fans, located at the air inlets, are not subjected to such a humid environment, but
they cannot produce the same air exit velocity. Lower exit velocities increase the
possibility of air recirculation to the cooling tower inlet. Both forced and induced
draft towers are equipped with louvers on the air inlets to regulate air flow. This
Figure 6-4
>
AIR
OUT
CIRCULATING WATER IN
< FILL
>
TO
BASIN CONDENSER
Figure 6-5
>
AIR
OUT
MIST ELIMINATORS
<
RETURN WATER
< DISTRIBUTION SYSTEM
FILL
<
breaks up the water flow to enhance air/water contact, and it comes in many
shapes and patterns. Wooden splash bars were once the primary material, but
these have mostly been supplanted by more sophisticated designs and materials,
as is illustrated in Figures 6-6 and 6-7. These include vertically aligned sheets of
corrugated plastic known as film fill, which is used in counterflow towers. Film
fill, while being very efficient, can be problematic if microbiological fouling or
scaling is not carefully controlled, as the close spacing of film fill layers makes the
material susceptible to plugging. In fact, the increasing use of film fill has been a
driving factor for improvements in treatment programs.
Figure 6-6
Plastic Splash Bars. Photo courtesy of the Marley Cooling Tower Company, Overland Park, KS.
cooling tower:
where
ƒ (average) = 0.75
ƒ (in summer) = 0.85
ƒ (in winter) = 0.65
1000 is the approximate latent heat of vaporization
The ƒ factor is the ratio of the amount of heat exchanged by evaporation ver-
sus sensible heat transfer, and is a function of wet bulb temperature and relative
humidity.
The number of times that water is concentrated in the cooling tower is
known as the cycles of concentration (C). C is usually determined by a compar-
ison of some very soluble ion, such as chloride or magnesium, in the makeup to
Figure 6-7
Film Fill. Photo courtesy of the Marley Cooling Tower Company, Overland Park, KS.
the recirculating water. Specific conductivity is also used for this determination,
although it is not completely linear with increasing dissolved solids.
Conductivity is also influenced by treatment chemicals added to the water.
The maximum cycles of concentration depend on the effectiveness of corro-
sion and scale inhibitor programs, and on the water quality of the makeup to the
tower. Soft water or softening of hard water may allow for increased cycles of
concentration due to the reduced scaling potential, although softening may
increase the corrosion potential. Cycles must be kept low when the makeup sup-
ply is scale forming or highly ionic. Seawater is a prime example of the latter.
Cycles of concentration in a seawater-cooled tower may be limited to less than
two due to the high ionic concentration of the recirculating water, and its poten-
tial for corrosion.
Regardless of the quality of the makeup water or effectiveness of a corrosion
or scale inhibitor program, cycles of concentration have a limit in any system.
Some water must be continually removed to prevent excessive buildups of dis-
solved solids. This is known as blowdown (BD). An additional, very small blow-
down occurs due to entrainment of water droplets in the air exiting the tower.
This is known as drift (D). Drift typically ranges from about 0.3 to 0.05% of the
recirculation rate, depending on the type and efficiency of the cooling tower.
Some of the more modern towers may have drift values below 0.01%.
Evaporation, blowdown, drift, and makeup comprise the water balance
around a cooling tower (Fig. 6-8), and a set of simple equations has been derived
to calculate this balance. Evaporation has already been given. The others are:
Figure 6-8
EVAPORATION
&
DRIFT
<
CIRCULATING LOOP
MAKEUP TO CONDENSER AND BACK
> >
>
BLOWDOWN
BD + D = E / (C – 1) (6.3)
MU = E + BD + D (6.4)
Figure 6-9
Relative Cooling Tower Size versus Approach Temperature. Reprinted with permission from Cooling
Tower Fundamentals by the Marley Cooling Tower Company, Overland Park, KS.
listed earlier, 75% of the heat exchange was due to evaporation and 25% to sen-
sible transfer. A 100% humid air could provide the same cooling if the wet-bulb
temperature were the same. For this to happen, the dry-bulb temperature would
have to be much lower. All heat exchange would be by sensible transfer. This is,
of course, an absurdity when designing cooling towers, but it does illustrate the
factors that must be taken into account.
Cooling towers can never reach 100% efficiency no matter where they are
located. The difference between the temperature of the cooled water and the wet-
bulb temperature is known as the approach. According to the Marley Cooling
Tower Company, the closest approach that cooling tower manufacturers are usu-
ally willing to guarantee is 5˚F. Figure 6-9 shows the tower size factor versus the
approach temperature. As can be seen, the typical design approach temperature
is 15˚F. An approach temperature of half that amount would double the size of
the cooling tower, and the minimum guaranteed approach of 5˚F would require
a tower over 2º times larger. The curve is asymptotic, which essentially precludes
closer approach temperatures than 5˚F due to economic and size constraints of
the tower. Figures 6-10 and 6-11 show an example of daily and seasonal varia-
tions of wet-bulb temperatures. These may play a very important part in the
design of the cooling tower. For example, seasonal operation of a cooling tower
might require an evaluation of several wet-bulb temperatures. Additional factors,
such as the proximity of the tower to sources that might affect the entering wet-
bulb temperature, i.e., other cooling towers, also may require adjustment of the
entering wet-bulb temperature. A reputable cooling tower firm will take all of
these factors into account when designing the tower. The performance of the
tower can change over time due to degradation if it is not properly maintained.
Figure 6-10
Illustration of Daily Wet-Bulb Temperature Variations. Reprinted with permission from Cooling Tower
Fundamentals by the Marley Cooling Tower Company, Overland Park, KS.
Figure 6-11
MONTH
Common Seasonal Wet-Bulb Temperature Variations. Reprinted with permission from Cooling Tower
Fundamentals by the Marley Cooling Tower Company, Overland Park, KS.
• Corrosion
• Scaling
• Fouling
• Microbiological fouling
These phenomena and their effects on cooling systems are often interrelat-
ed. For example, microbiological deposits restrict heat transfer similar to scales
and also generate under-deposit corrosion cells. Slime produced by microorgan-
isms will trap silt and increase fouling. Corrosion products may dislodge and
form deposits in heat exchangers. The source of water and the type of system
(once-through, recirculating, and closed) have a great impact on these phenom-
ena. Environmental considerations are playing a major role in control methods.
Corrosion
Corrosion control requires careful evaluation because of the many and var-
ied corrosion mechanisms that can occur. Cooling water systems may contain
several different metals, including carbon steel water lines, copper-alloy heat
exchanger tubes, galvanized structural supports, bolts, nuts, etc. These may all
suffer corrosion, sometimes from chemicals added for scale or foulant control.
Even concrete and wood components are subject to attack and degradation.
Corrosion of metals is an electrochemical reaction in which an electron
transfer takes place between the corroding material and the corrosive medium.
The driving force for corrosion reactions is the potential between the electron
acceptor (the corrosive medium) and the electron donor (the metal). Every metal
exhibits a different tendency to release or accept electrons in a corrosion cell.
This tendency is measured against the standard hydrogen ion half cell, in which
a one molar solution of hydrogen ions is assigned a half cell potential of zero.
• Iron releases electrons, which travel through the metal to another site
where they reduce oxygen and water to hydroxide ions. The site of reduc-
tion is known as the cathode.
• Oxidized iron atoms leave the metal substrate and enter the solution. This
site is known as the anode.
Figure 6-12
CATHODE ANODE
+2 +2
Fe Fe
+2
O2 + H2 O > O H- Fe +2
v
Fe
<
e-
PIPE WALL
Steel Corrosion Cell.
• The iron and hydroxide ions combine to form a precipitate. Under the con-
ditions shown, the precipitate eventually converts to rust (Fe2O3 • xH2O).
Also, copper is a noble metal and does not undergo the simple acid attack shown
in Equations 6.5 and 6.6. Copper alloys can corrode, however, in the presence
of compounds that complex or bond with copper. Ammonia, of course, is the
primary copper complexor, however its concentration in natural waters is usual-
ly very small. Ammonia can become a problem where treated sewage water is
used for cooling water makeup.
Sulfide is another problematic corrodent, and copper alloys should not be
used to handle waters that contain sulfides. Case History 6-2 discusses corrosion
of copper-alloy condenser tubes by sulfides from an unexpected source.
Reports suggest that the most common water-side corrosion problems are
due to dezinkification or denickelification of Admiralty metal and copper-nickel
alloys, respectively. These corrosion mechanisms are influenced by excessive
chlorination, low pH, and high chlorides. Some copper alloys, particularly
Admiralty, are soft and may suffer erosion from the flowing cooling water. This
condition typically occurs at the inlet end of the tubes.
Oxygenated acid solutions are also detrimental. Corrosion of copper-alloy
condenser tubes by these types of solutions has occurred in cooling systems
treated with sulfuric acid when the acid feed system malfunctioned and lowered
the pH of the oxygen-containing, cooling water.
Nonmetallic Corrosion
Other cooling water system materials can degrade as well, albeit by different
mechanisms than those shown above. Wood may be subject to degradation (rot)
by various types of fungi, which attack either the cellulose or lignin structure of
the material. Concrete degradation may be caused by too many or too few ions
in the water. Both chloride and sulfate can attack concrete, although special types
of concrete are available to resist these ions. Sulfate attack is prevalent in many
cooling towers operated at high cycles of concentration using sulfuric acid to
control pH. Chlorides at high concentration will attack the reinforcing bars and
cable in concrete.
Conversely, concrete may corrode if the water is too soft because calcium
ions will leave the concrete to establish equilibrium with the water. Supplemental
calcium additions to the water may be needed to minimize this reaction.
Protective coatings that serve as moisture barriers can also be beneficial.
Corrosion Inhibitors
Corrosion inhibitors work by protecting the material surface. Corrosion is
influenced by a wide variety of conditions, and it is often difficult to precisely
predict when and where it will occur. A corrosion inhibitor program should be
designed to provide total protection.
Inhibitors effectively depolarize (reduce or stop the electrical flow of) the
corrosion reaction. Individual corrosion inhibitors provide anodic or cathodic
protection, or both when blended. In general for cooling applications, cathodic
inhibitors precipitate at the locally high pH cathodic site to form a barrier that
limits the rate of oxygen reduction. Anodic inhibitors generally promote the for-
mation of a stable metal oxide at the anodic surface. This limits metal dissolu-
tion. The most common mild steel corrosion inhibitors include:
Anodic Cathodic
Molybdate Zinc
Orthophosphate Polyphosphate
Nitrite Phosphonate
Silicate
such as tolyltriazole (TTA, Fig. 6-13) or butyl-benzotriazole (Fig. 6-14) are most
common. The azoles function by bonding to the cuprous oxide layer through
nitrogen atoms on the molecule. The azole not only stabilizes the cuprous oxide
from further oxidation, but the planar azole molecules form a barrier on the
metal surface. Azole dosages of 1 to 2 ppm are common. The azoles, which dis-
solve much more readily in basic solutions, have become popular in blended for-
mulations of alkaline scale inhibitors.
Scale
Scale and deposit control are very important in cooling systems because of
the effects that deposits have on heat transfer, corrosion of metals, and fouling of
cooling tower film fill. The following sections outline methods for monitoring
and treatment of scale in cooling water systems.
Calcium Carbonate
The potential for scale formation is much Table 6-3
higher in open, recirculating cooling systems,
Most Common Cooling
where the solids concentration may be 5, 10, or
Water Scales
perhaps 20 times greater than that in the make-
up. The common cooling water scales are listed Calcium Phosphate
in Table 6-3. Most of these compounds are cal- Calcium Carbonate
cium based, with the two most common being Silica
calcium phosphate and calcium carbonate. Calcium Sulfate
Calcium carbonate scale forms when calci- Calcium Fluoride
um and alkalinity exceed saturation values and
begin to drop out as calcium carbonate (CaCO3). A number of programs have
been developed to calculate scaling potentials. In the 1930s, Langelier developed
the first set of calculations for this purpose, and the value derived from the cal-
culations is known as the Langelier Saturation Index (LSI). Calcium carbonate
precipitation is dependent not only on the calcium and alkalinity in solution but
also on the pH. Langelier’s equations take this into account:
where
pHs is the Langelier Saturation pH
pK2 is the second dissociation constant for carbonic acid
pKs is the solubility product constant for calcium carbonate
An LSI less than zero indicates that the water tends to dissolve scale, and
thus may be corrosive. An LSI greater than zero indicates that the water may be
scale forming.
Since Langelier introduced his equations, the ease of calculating the LSI has
been improved. For example, Equation 6.13 may now be written as:
Values for A, B, C, and D can be found in Table 6-4. Also, nomographs have
been developed for calculating the LSI. One of these is shown in Figure 6-15.
The user determines the values for pCa, pAlk, and C, and the sum of these val-
ues is then subtracted from the actual pH to give the LSI. The LSI can also be
approximated from the following equation:
Although the Langelier Saturation Index has proven to be useful, it does not
always provide the accuracy needed due to inherent errors in the calculations. In
particular, the calculations do not account for ionic interactions in the water.
To improve the accuracy of the calculation, Ryznar developed the Stability
Index:
The scale of the RSI is different than that of the LSI. Waters with an RSI of
less than 6 are scale-forming, while those above 7.5 exhibit corrosive tendencies.
The LSI and RSI are often used in combination, although the RSI is based on the
same calculations as the LSI and provides no additional improvement with
regard to ionic interactions. Supplement 6-2 outlines a BASIC program that cal-
culates LSI and RSI using Equation 6.15. Like the original calculations by
Table 6-4
Data for Rapid Calculation of Saturation & Stability Indexes (Based on Langelier for-
mulas, Larson-Buswell residue, temperature adjustments & arranged by Eskel Nordell)
A B C D
Total Solids Temperature Calcium M.O.
(ppm) (F.) Hardness Alkalinity
A B (ppm or CaCO3) C (ppm or CaCO3) D
50–300 0.1 32–34 2.6 10–11 0.6 10–11 1.0
400–100 0.2 36–42 2.5 12–13 0.7 12–13 1.1
44–48 2.4 14–17 0.8 14–17 1.2
50–56 2.3 18–22 0.9 18–22 1.3
58–62 2.2 23–27 1.0 23–27 1.4
64–70 2.1 28–34 1.1 28–34 1.5
72–80 2.0 35–43 1.2 35–43 1.6
82–88 1.9 44–55 1.3 44–55 1.7
90–98 1.8 56–69 1.4 56–69 1.8
100–110 1.7 70–87 1.5 70–87 1.9
112–122 1.6 88–110 1.6 88–110 2.0
124–132 1.5 111–138 1.7 111–138 2.1
134–146 1.4 139–174 1.8 139–174 2.2
148–160 1.3 175–220 1.9 175–220 2.3
162–178 1.2 230–270 2.0 230–270 2.4
280–340 2.1 280–340 2.5
Saturation Index = 350–430 2.2 350–430 2.6
pH (actual) – (9.3 + A + B) + (C + D) 440–550 2.3 440–550 2.7
Stability Index = 560–690 2.4 560–690 2.8
2 [(9.3 + A + B) – (C + D)] – pH (actual) 700–870 2.5 700–870 2.9
880–1,000 2.6 880–1,000 3.0
Values for A, B, C, and D for LSI Calculations. Adapted from Cooling Tower Fundamentals by the Marley
Cooling Tower Company, Overland Park, KS.
Figure 6-15
LSI Nomograph. Reprinted with permission from the Betz Handbook of Industrial Water Conditioning, Ninth
Edition, BetzDearborn, Inc., Horsham, PA.
Other Scales
Many other scales are possible in cooling water systems. The following are
reported guidelines for maximum levels of calcium and other anions besides
bicarbonate that can be tolerated without treatment.
Calcium phosphate. Calcium is limited by the following equation when
phosphate concentrations are 10 ppm or higher.
Magnesium silicate
The reader should note that these are only general recommended guidelines,
and as with all other guidelines in this book, each system must be evaluated indi-
vidually.
One obvious item in this list is magnesium silicate. The allowable concen-
tration greatly decreases as pH increases. Many utilities and industries are oper-
ating cooling towers at higher cycles of concentration to minimize discharge and
conserve water. However, this increases the potential for scale formation.
Scale Control
A number of scale control techniques and chemical programs are available.
These range from traditional programs such as acid feed, to treatment with com-
plex organic polymers. The range of chemicals gives the operator flexibility in
program selection.
Acid Feed
For many years, the primary method for controlling calcium carbonate scale
was sulfuric acid addition. Sulfuric acid reduces bicarbonate alkalinity as follows:
Table 6-6
Alkalinity Relationship
to “P” and “M” Readings
Alkaline treatments have become the most popular of all cooling water pro-
grams. In an alkaline program, the pH of the cooling water is allowed to equili-
brate at a natural level, or may even be pushed into a basic pH range by addition
of alkaline compounds such as caustic soda or soda ash. In some instances, how-
ever, acid feed is still required to prevent the recirculating water from becoming
too alkaline.
Alkaline treatment programs are designed to
Figure 6-16
either keep calcium in solution or to modify the
H
O O O crystalline structure of calcium precipitates so that
HO O H they form a sludge-like product, which can be
P C P blown down. Solubilizing compounds include
HO OH
phosphonates and polymers such as polyacrylate.
CH 3
Common phosphonates include hydroxyethyli-
HEDP.
dene diphosphoic acid (HEDP, Fig. 6-16) and
aminoethylenephosphonic acid (AMP). A newer
Figure 6-17 product, phosphono-butane-tricarboxylate (PBTC)
~ CH 2 CH CH 2 CH ~ is also proving to be effective. The phosphate por-
tion of each of these compounds adsorbs onto the
COOH COOH
incipient calcium carbonate crystal nuclei, distort-
Polyacrylate.
ing its shape and slowing crystal growth.
Phosphonates may produce some negative side effects. They are corrosive to
zinc and mildly corrosive to copper and aluminum. Phosphonates are also
degraded by oxidizing biocides, particularly at biocide concentrations above 1
ppm. Not only will this cause the phosphonates to lose effectiveness, but the
breakdown products include orthophosphate, which can combine with calcium
to form calcium phosphate scale. Despite these factors, phosphonates are still the
most popular scale control agents. PBTC in particular has made an impact. It is
reported to be less corrosive than other phosphates to copper, and more resistant
to degradation from oxidizing biocides or UV light.
Polyacrylates (Fig. 6-17) also keep calcium in suspension. The polymer binds
calcium through the partial negative charges on the oxygen atoms contained in the
molecules. A co-polymer of polyacrylate and acrylamido-2-
methylpropane-sulfonic acid has proven to be a very effective dispersant.
Crystal modifiers include polymaleates and co-polymers of these with sul-
fonated polystyrenes. These polymers allow calcium deposits to form, but alter
the structure so that the deposits do not precipitate as a hard scale. The deposits
are removed in the blowdown or with a sidestream filter. The crystal modifiers
are more expensive than other treatments and are not used nearly as often.
Table 6-7 illustrates guidelines for scale inhibitor residuals. With all of these
options, it is often difficult to decide on a treatment program. However, some
trends have emerged. Blended products of phosphonates and polymers seem
LSIs as high as 2.8 to 3.0 and PSIs as low as 3.5 may be obtained with the
correct chemical program. It has also been claimed that these new polymer for-
mulations allow higher silica levels of up to 300 ppm. Some water treatment per-
sonnel view this cautiously and still abide by the old silica limit of 150 ppm.
The selection of the best cooling water treatment is dependent upon water
quality, process conditions, and system metallurgy. The choice is sometimes not
easy. Any program should be monitored closely to determine its performance.
Fouling
Fouling is the deposition of suspended solids or buildup of microbiological
organisms within heat exchangers and cooling tower fill. Foulants can be intro-
duced to a cooling water system from a variety of sources. A makeup supply from
a lake or river may contain silt and debris that is stirred up during seasonal
changes or heavy rainfall. Surface water sources also contain many microorgan-
isms that will cause microbiological buildups even in once-through cooling sys-
tems. A cooling tower is an ideal source for foulant introduction because the
tower is an efficient air scrubber. Warm water, aeration, nutrients, and sunlight
transform a cooling tower into an efficient bio-reactor. The following sections
discuss fouling control methods.
Microbiological Fouling
Microbiological fouling may be the worst problem in cooling water systems.
Microbes can:
• Generate acids and corrodents that attack the underlying tube metal
• Secrete a protective gelatinous layer that coats the tube surface. The secre-
tions, along with silt that becomes trapped within, reduce heat transfer
and cause under-deposit corrosion.
HOCl is the active ingredient of this mixture, and it attacks organisms very
quickly. It has been calculated that the kill rate of a 1.0 ppm chlorine solution is
OCl- is much less potent than HOCl. Some scientists theorize that OCl-,
being a charged ion, has much more difficulty penetrating the cell wall and
attacking cell components. The dissociation of HOCl begins at a pH of about 5.2,
reaches 50% at a pH of 7.5, and is fully complete at a pH of 9.4. With the grow-
ing trend towards alkaline cooling water treatment programs, other biocides are
proving to be more effective.
Application of chlorine. Chlorine gas is very toxic and must be handled
with great care. A common method of introducing it into water is through an
eductor system, in which a sidestream of flowing water is used to pull the chlo-
rine gas from the chlorine cylinder into the stream. Chlorine is commonly
shipped in one-ton cylinders, which can be loaded in multiple units into a rack
containing a manifold system for quick transfer between full and empty contain-
ers. Chlorine systems are now required to have ambient air monitoring and alarm
systems that provide safety warnings in the event of a chlorine leak. The maxi-
mum allowable ambient air chlorine limit as established by OSHA is 1.0 mg/l.
Chlorine reacts irreversibly with a number of constituents in water, most
notably ammonia and organics. These reactions, which reduce the amount of free
chlorine available as a biocide, are known as the chlorine demand. The higher
the chlorine demand, the more chlorine that must be added to achieve the same
level of killing effectiveness. Ideally, a 0.1 to 0.5 ppm free chlorine residual is
most effective in controlling microorganisms, while minimizing degradation of
cooling water treatment chemicals or cooling tower materials.
Environmental aspects of chlorine usage are becoming more complicated.
The “technology-based limits” established by EPA for power plants essentially
limited the chlorine residual in the cooling system discharge to 0.2 ppm for a
maximum of two hours per day. These limits are gradually giving way to “ambi-
ent water quality standards” to be determined in the stream at the boundary of a
calculated mixing zone. The new limit is 0.019 mg/l in fresh water for intermit-
tent applications. Individual states may set more stringent limits, and often apply
the limits at “end-of-pipe” instead of allowing for dilution in the mixing zone.
These concentrations are much too low to be effective. A method to main-
tain higher free chlorine residuals is to dehalogenate the discharge with sodium
sulfite (Na2SO3), sodium bisulfite (NaHSO3), or sulfur dioxide gas. These react
with chlorine in a one-to-one stoichiometric fashion:
Bromine and Chlorine Dissociation Chart. Reprinted with permission from the Betz Handbook of Industrial
Water Conditioning, Ninth Edition, BetzDearborn, Inc., Horsham, PA.
• Powerful oxidizer
• Not pH sensitive like HOCl
• Does not form halogenated organics
• Does not react with ammonia
Of these production methods, the first two have been the most popular for
large cooling systems. However, the latter method offers an advantage because no
chlorine or hypochlorite is needed. Similar to gaseous chlorine or bromine treat-
ment, an eductor system is employed to mix the sodium chlorite with the acti-
popular being used today include isothiozolone (Fig. 6-23), quaternary amines
(Fig. 6-24), and bromonitropropanediol (BNPD, Fig. 6-25). These biocides work
by either reacting with the cell wall or interfering with the organism’s metabolic
processes. A cell structure is shown in Figure 6-26. Quaternary amines penetrate
into cell walls and disrupt transport of products through the cell wall. BNPD and
isothiozolone react with groups inside the cell or interfere with protein making
processes.
Microorganisms can develop a resistance to the nonoxidizing biocides, so
careful planning may be needed for them to be effective. One method is to use
them as a supplement to oxidizing biocides. A periodic batch dosage can be used
to shock the microbes and kill those which may have survived the oxidizing
chemical. Or, two nonoxidizers could be used on a more continuous basis, but
be alternated periodically (perhaps even daily) so that the microorganisms do not
build up any tolerance.
Just like oxidizing biocides, nonoxidizers can have an effect on the environ-
ment if they leave the plant in the cooling discharge. Various treatment methods
are possible to neutralize many of these chemicals. For instance, some of the
non-oxidizers, isothiozolone and BNPD in particular, are decomposed with sul-
fite. Bentonite clay, if added to the discharge, provides absorption sites for com-
pounds such as the quaternary amines.
Macrofouling
Fouling caused by organisms whose individual members are visible to the
naked eye is known as macrofouling. Macrofouling may be caused by a number
of different creatures, but the two most troublesome fresh-water species by far
have been Asiatic clams and even more importantly zebra mussels.
Macrofouling by clams occurs when the creature dies and the shell is
washed into sensitive flow areas such as a condenser. Asiatic clams proved trou-
blesome because shell sizes typically ranged from about 1/2 inch to 1 inch in size.
Since condenser and heat exchanger tubes often range from 1/2 inch to 1 inch in
Figure 6-26
diameter, the clam shells almost perfectly fit. A growth in Asiatic clams within a
cooling system could play havoc with operation.
This, however, seems almost minor compared to some of the horrors that
plant personnel have experienced in dealing with zebra mussels.
Zebra Mussels
These creatures were inadvertently introduced to the Great Lakes in the
mid-1980s. They are so called because their shell is alternately patterned with
light and dark stripes. Zebra mussels are freshwater bivalves that may grow to
about 5 cm in length.
Zebra mussels are extraordinarily troublesome, primarily because of their
colonization patterns. Zebra mussels are microscopic when first spawned, being
about 40 microns in length. They also have no shell at first. Within just a few
weeks after hatching, however, the mussels have become mature enough to look
for a place to settle. When the mussel has found a location, it extends fibers,
known as byssal threads, to attach permanently to the surface. Zebra mussels are
• The microscopic larvae (veligers) can be carried a long way by water cur-
rents.
• The veligers are much too small to be stopped by traveling screens at
water intakes.
• They like locations of flowing water, such as cooling water intakes, where
nutrient supplies are plentiful. However, if water flow rates are higher
than 3 or 4 feet per second they may not settle.
• They will attach to many different hard surfaces, including each other, to
form dense, thick masses.
This latter aspect is what has made zebra mussels particularly annoying.
Densities of up to 80,000 creatures per square foot have been found at locations
in the Great Lakes. The mussels can clog trash racks, traveling screens, auxiliary
cooling water lines, and fire lines. Masses that break loose in the main cooling
water line will plug condenser tubes.
Various control methods have been investigated for control of these creatures.
Some of them include electrification of trash racks, application of fouling or slick
coatings to intake pipes, and even use of acoustics. None have yet been adopted on
an industrial level. More effective have been chemical treatment and thermal shock.
Chemical treatment. If zebra mussels infest a cooling water system, it is
usually because the veligers have entered and then matured within the system.
Plant personnel can combat this by diligently operating the biocide system, even
though feed may not be continuous. Infant zebra mussels have no shell, and it
takes several days for one to develop. Regular oxidant feed to the cooling water
will kill some of the veligers before the shell fully develops.
To kill mature mussels takes a more vigorous effort. Adult mussels can sense
oxidizing biocides and will simply close up if they detect a residual oxidant. They
can stay closed for at least several days, and can certainly outwit an intermittent
oxidant feed. One method is to initiate continuous oxidant feed, and dehalo-
genate the discharge. Eventually, the mussel will have to open, at which time the
oxidant can kill it. Such a treatment can only be used if the plant’s environmen-
tal permit allows for it. However, the EPA has granted variances for this type of
application to facilities threatened by zebra mussels.
An alternative is feed of a nonoxidizing biocide. Mussels cannot detect these
chemicals and will filter the water unwittingly. Nonoxidizers damage zebra mus-
sel cells just as they do microbiological cells. Quaternary amines have proven to
be effective. Again, however, chemical feed must be approved by the environ-
mental protection governing board that oversees the plant.
Thermal shock. Zebra mussels are not very tolerant to really warm water
Conclusion
Cooling water treatment is often a complex issue, and is not being made any
easier by environmental regulations. Certainly chlorine as a biocide and metals
(zinc and chromate) for corrosion control have lost popularity. Alkaline cooling
water treatment programs appear to be the trend. Plant personnel must careful-
ly consider any treatment program and take into account such factors as econo-
my, safety, efficiency, and environmental regulations.
Supplement 6-1
Cooling Tower BASIC Program
The cooling tower BASIC program asks for four variables, recirculating
water flow rate, temperature difference between the warm return water and
cooled water in the tower basin, the evaporation correction factor, and cycles of
concentration. The program will display values for evaporation, blowdown, drift,
and makeup.
CTALCS.BAS
10 CLS:LOCATE 5,1
20 PRINT “COOLING TOWER CALCULATION PROGRAM:
30 PRINT “ENTER THE RECIRCULATING FLOW RATE IN GPM:
50 INPUT R
60 PRINT
70 PRINT “ENTER THE TEMPERATURE RANGE BETWEEN THE WARM RETURN WATER”
80 PRINT “AND THE COOLED WATER IN THE TOWER BASIN (DEGREES F):
09 INPUT DELTAT
100 PRINT
110 PRINT “ENTER THE CORRECTION FACTOR (f) FOR EVAPORATION”
Supplement 6-2
LSI and RSI BASIC Program
The program is based on the values for A, B, C, and D found in Table 6–4.
The user must input this data. The program calculates reasonably accurate val-
ues for both the LSI and RSI in waters that are not too highly concentrated. (TDS
below about 800 ppm.) More sophisticated programs, which take into account
ionic interactions, must be used to calculate the scaling potential of more highly
concentrated waters.
LSI.BAS
10 CLS:LOCATE 5,1
20 PRINT “THIS PROGRAM WILL CALCULATE THE LANGELIER SATURATION INDEX”
30 PRINT “AND AYZNAR INDEX FOR WATER.”
40 PRINT
50 PRINT “ENTER THE MEASURED pH”
60 INPUT PHA
70 PRINT
80 PRINT “ENTER THE TOTAL DISSOLVED SOLIDS AS PPM”
90 INPUT TDS
100 PRINT
110 PRINT “ENTER THE WATER TEMPERATURE AS DEGREES F”
120 INPUT T
130 PRINT
140 PRINT “ENTER THE CALCIUM HARDNESS AS PPM CaCO3”
150 INPUT CAH
160 PRINT
170 PRINT “ENTER THE ALKALINITY AS PPM CaCO3”
180 INPUT ALK
190 GOSUB
200 PRINT
210 PRINT
220 PRINT “*********************************************”
230 PRINT “THE LSI = “;:PRINT USING “##.##”;LSI
240 PRINT “THE AYZNAR INDEX = “;:PRINT USING “##.##”;RSI
250 PRINT “*********************************************”
260 PRINT
270 INPUT “ANOTHER ANALYSIS (Y/N)”;ANS$
280 IF ANS$=”Y” GOTO 10
290 IF ANS$=”y” GOTO 10
300 CLS:LOCATE 1,1:SYSTEM
310 IF TDS>=SO AND TDS<=3OO THEN A=.1
320 IF TDS>300 AND TDS<400 THEN A=(.001*TDS)-.2
330 IF TDS>=400 AND TDS<=1000 THEN A=.2
340 IF T >=32 AND T<=34 THEN B=2.6
350 IF T=35 THEN B=2.5S
360 IF T>=36 AND T<=42 THEN B=2.5
370 IF T=43 THEN B=2.45
380 IF T>= 44 AND T<=48 THEN B=Z.4196
390 IF T=49 THEN B=2.35
400 IF T>=SO AND T<=56 THEN B=2.3
410 IF T=57 THEN B=2.25
420 IF T>=58 AND T<=62 THEN B=2.2
430 IF T=63 THEN B=2. 15
440 IF T>=64 AND T<=70 THEN B=2.1
450 IF T=71 THEN B=2.05
460 IF T>=72 AND T<=80 THEN B=2
470 IF T=81 THEN B=1.95
480 IF T>=82 AND T<=88 THEN B=1.9
490 IF T=89 THEN B=1.85
500 IF T>=90 AND T<=98 THEN B=1.8
510 IF T=99 THEN B=1.75
520 IF T >= 100 AND T<=110 THEN B=1.7
530 IF T= 111 THEN B=1.65
540 IF T>=112 AND T<=12Z THEN B=1.6
5S0 IF T= 123 THEN B=1.55
560 IF T>= 124 AND T<=132 THEN B=1.5
570 IF T=133 THEN B=1.45
580 IF T>=134 AND T<=146 THEN B=1.4
590 IF T=147 THEN B=1.35
600 IF T>=1 48 AND T<=160 THEN B=1.3
610 IF T=161 THEN B=1.25
620 IF T>161 THEN B=1.2
630 IF CAH>=10 AND CAH<=11 THEN C=.6
640 IF CAH>=12 AND CAH<=13 THEN C=.7
650 IF CAH>=14 AND CAH<=17 THEN C=.8
660 IF CAH>=18 AND CAH<=22 THEN C=.9
670 IF CAH>=23 AND CAH<=27 THEN C=1
680 IF CAH>=28 AND CAH<=34 THEN C=1.1
690 IF CAH>=35 AND CAH<=43 THEN C=1.2
700 IF CAH>=44 AND CAH<=55 THEN C=1.3
710 IF CAH>=56 AND CAH<=69 THEN C=1.4
Supplement 6-3
Acid Feed Calculations
Alkalinity like all other ions will cycle up in a cooling tower unless it is con-
trolled. Sulfuric acid converts bicarbonates and carbonates to carbon dioxide,
which escapes from the tower. The amount of acid needed for conversion is
dependent upon the alkalinity in the makeup water and the desired concentra-
tion in the tower. Water treatment experts recommend an alkalinity of 20 to 40
ppm in the recirculating water to provide safe operation. Consider an example
where the makeup alkalinity is 100 ppm (as CaCO3), the cycles of concentration
are 6, and the desired alkalinity in the recirculating water is 30 ppm. The alka-
linity in the makeup would have to be lowered from 100 to 5 ppm.
The accompanying BASIC program will calculate the amount of acid need-
ed for alkalinity reduction. The program asks for makeup flow rate, alkalinity in
the makeup water, desired alkalinity in the recirculating water, and cycles of con-
centration. The calculations are made more simple by the fact that alkalinity is
routinely measured and reported as ppm CaCO3. (The molecular weight of
CaCO3 is 100, this is why it is so often used as a standard for water analysis cal-
culations.) Sulfuric acid has a molecular weight of 98 and an equivalency of 2,
so for most practical calculations 1 ppm of sulfuric acid can be considered to
react with 1 ppm of alkalinity as CaCO3. The BASIC program assumes this ratio
and does not correct for the slight difference in molecular weights.
ADICDALC.BAS
10 CLS:LOCATE 5,1
20 PRINT “ACID FEED CALCULATION PROGRAM.”
30 PRINT
40 PRINT “ENTER THE MAKEUP FLOW RATE IN GALLONS PER MINUTE”
50 INPUT GPM
60 GPD=GPM*1440
70 PRINT
80 PRINT “ENTER THE ALKALINITY OF THE MAKEUP WATER”
90 INPUT MUALK
100 PRINT
110 PRINT “ENTER THE DESIRED ALKALINITY IN THE RECIRCULATING WATER”
120 RWALK
130
140 PRINT “ENTER THE CYCLES OF CONCENTRATION”
150 INPUT C
160 MUALKD=RWALK/C
170 MINUSALK=MUALK-MUALKD
180 PRINT
190 LBACID = (GPD*MINUSALK)/(1000*120)
200 PRINT “THE AMOUNT OF SULFURIC ACID REQUIRED PER DAY = “;:PRINT USING
“#####”
;LBACID;:PRINT “ POUNDS”
Supplement 6-4
Sizing a Sidestream Filter
A simple set of equations is available to size sidestream filters. The calcula-
tions are based on percent solids reduction and the cooling tower blowdown
rate. As an example, consider a recirculating water that contains 20 ppm of sus-
pended solids. The desirable concentration is 5 ppm. Blowdown from the tower
is 100 GPM.
First calculate the percent solids reduction, where:
Supplement 6-5
Specifications for Sodium Hypochlorite
Sodium hypochlorite will decompose into oxygen, sodium chloride, and
sodium chlorate. The decomposition rate is affected by temperature and by the
action of metals, most notably iron and copper. A specification should contain
the following:
that afflicted most of the Midwest caused the lake level to drop dramatically.
Plant chemists calculated that the concentration of dissolved solids in the lake
water increased to four times the normal amount. However, they and other plant
personnel did not give thought to the possibility of scale formation, as none had
occurred before. During the summer, performance in the 10-year-old condenser
declined slowly but noticeably. When the unit came off line for an autumn out-
age, an inspection team found that the waterside of the tubes was completely
covered with a layer of scale, less than one millimeter in thickness. The scale was
primarily composed of calcium carbonate. The scale was the direct result of the
drought and the concentration of solids in the lake. The plant staff hired a con-
tractor to mechanically scrape the tubes. The problem has not reoccurred since.
An interesting aspect about this situation is that the scale developed in this one
condenser, which is tubed with copper-nickel alloy. No scale formed in the other
condensers, which are tubed with Admiralty metal. Because outlet cooling water
temperatures in the other condensers were at least as high as those in the scaled
condenser, the scaling may have also been influenced by metallurgy. No investi-
gation was conducted into this possibility, however.
system would not start-up because an electrical component had failed. Plant per-
sonnel had no way of knowing in advance about the failure, because the system
was operating properly when it was shut down the previous autumn. The
replacement part could not be procured for over a month, during which time the
plant results staff watched condenser performance steadily decline. The problem,
which was visually confirmed later, was caused by microbiological fouling of the
condenser tubes. The condenser efficiency dropped so low that plant personnel
chlorine-shocked each condenser once during the summer. The process, which
was performed while the condensers were on-line, was interesting. Unit load was
reduced so that one-half of the condenser could be taken out of service. The out-
let valve was closed, and then 25 to 50 pounds of calcium hypochlorite were
blend-filled into the condenser waterbox with diffusion into the condenser tubes.
The chemical was allowed to remain in the condenser for at least four hours, after
which that half of the condenser was valved back into service and the other half
was treated.
The process restored only about 50% of the condenser efficiency. Part of the
problem may have been caused by poor distribution of chemical in the lower
pass of the two-pass condensers. However, part of the problem was also caused
by the slime layer generated by the bacteria. A lot of it did not detach from the
tubes, even though they had been soaked with a concentrated chlorine solution.
That autumn, the plant maintenance crew mechanically scraped the tubes in
each condenser to remove the residual slime and entrained silt deposits.
This case clearly illustrates that proactive treatment of microorganisms is
better than reactive treatment. Once formed, microbiological deposits are diffi-
cult to remove.
Chapter 7
Sampling
Introduction
The most well-intentioned chemical treatment program may be virtually
worthless without representative sampling. However, sample collection and
transport to on-line instruments requires great care. Some of the key factors that
go into representative sampling include:
• Extracting the sample from the bulk solution so that it represents actual
conditions in the process fluid
• Conditioning the sample to prevent deposition of dissolved constituents
in the sample line
• Maintaining linear velocity of the sample within a suitable range to pre-
vent deposition or entrainment of suspended solids
• Further conditioning the sample to pressures and temperatures that allow
for accurate analyses by on-line instruments
Sampling 207
Table 7-1
Steam Sample Points and Recommended Analysis
for Utility Drum Boilers Treated with Phosphate
Table 7-2
Steam Sample Points and Recommended Analysis
for Utility Drum Boilers on All-Volatile Treatment
Sampling 209
Table 7-3
Steam Sample Points and Recommended Analysis
for Once-Through Utility Boilers
On All-Volatile-Treatment
Table 7-4
Steam Sample Points and Recommended Analysis
for Cogeneration Drum Boilers Treated with Phosphate,
which Drive Turbines
Sampling 211
ment outputs wired directly to the demineralizer and/or other control system.
Thus, the demineralizer operator can examine system chemistry data as he or she
operates the equipment.
Less stringent makeup treatment equipment and techniques, and therefore
less sophisticated instrumentation, are often satisfactory for lower-pressure units.
For example, sodium softening is frequently used to produce makeup for boilers
operating at less than 600 psig. At these lower pressures, carryover is much less
severe, so a higher amount of dissolved solids can be tolerated in the boiler water.
Sodium softeners are designed to remove hardness so the most important analy-
ses are those for calcium and magnesium. On-line monitors are available for
these measurements.
Water quality for intermediate- and high-pressure units is achievable not
only with demineralizers, but also with membrane processes, sometimes in
series, or with a mixed-bed polisher. Reverse osmosis (RO) is typically the back-
bone of a membrane treatment system. Because an RO unit uses pressure to
electromechanically filter dissolved solids, on-line measurements usually consist
of pressure, flow, temperature, and conductivity. Fouling, mechanical failure of
equipment, or degradation of membranes will cause significant changes in these
parameters. RO monitors and gauges are typically mounted on the equipment.
Signals may be sent to a local control panel or plant control system.
Monitoring of RO influent pH is frequently also necessary. For example, cel-
lulose acetate membranes must operate within a relatively narrow pH range (4 to
6) to prevent membrane degradation. Thin-film-composite (TFC) membranes
will operate over a much broader pH range, but are affected by even small quan-
tities of chlorine. Where TFC membranes are used to treat chlorinated water, a
dechlorination system is placed ahead of the RO to protect the membranes.
Chlorine residual or oxidation reduction potential (ORP) monitoring of the
dechlorination system effluent provides a safeguard against membrane attack.
ductivity, and dissolved oxygen. These will provide a quick indication of a con-
denser tube leak, demineralizer overrun, or an air in-leakage problem.
Given the effects that organics have on boiler water and steam chemistry,
boiler water experts now recommend that TOC be analyzed on a weekly grab
sample basis.
Deaerator Inlet
The two principal on-line analyses recommended for the deaerator inlet are
oxygen scavenger and dissolved oxygen. These analyses provide data on the per-
formance of the scavenger in the condensate system. Even if the oxygen scav-
enger is injected in the deaerator or deaerator outlet, on-line scavenger monitor-
ing of the deaerator inlet is still useful in determining scavenger carryover
through the steam system.
Sampling 213
Boiler Water
This is another very critical sample, for it is in the boiler that chemistry
upsets typically cause the most damage. Continuous analyses should include pH,
specific conductivity, silica, and phosphate for phosphate-treated units. As chap-
ter 3 discussed, a chemistry upset can rapidly deplete boiler water treatment
chemicals. When this happens, the boiler water pH will decrease (or sometimes
increase) to values outside of recommended ranges. A chemistry upset severe
enough to significantly alter boiler water pH has the potential to cause cata-
strophic tube corrosion. Any monitoring system should be equipped with alarm
capabilities for drum water pH.
For phosphate-treated boilers, phosphate measurements are obviously quite
critical. Phosphate provides the primary protection of the boiler water, and if an
upset depletes the phosphate concentration, severe corrosion or deposition may
result. Phosphate monitoring is also critical for controlling sodium-to-phosphate
ratios in coordinated/congruent treatment programs.
Silica is an extremely important parameter due to its vaporous carryover
qualities and tendency to form tenacious deposits on turbine blades. Silica is usu-
ally the limiting factor when a boiler is started up from outage (Case History 7–1).
Ammonia analyses of phosphate-treated boiler water allow the plant chemist
to calculate correct sodium-to-phosphate ratios. Ammonia analyses are not as
critical for AVT drum units because control is based on ammonia concentrations
in the feedwater.
Chloride and sulfate analyses of boiler water can help the plant chemist
monitor and prevent carryover. Given the problems these two ions can cause in
turbines, and the relative ease with which they carry over, the sampling recom-
mendations seem well justified.
Saturated Steam
A principal reason for sampling saturated steam is to verify that the steam pro-
duced by the boiler meets the turbine manufacturer’s steam quality guidelines. If
main or reheat steam samples are unavailable, then saturated steam sampling
assumes much greater importance. If saturated steam is not the only steam sample,
then saturated steam sampling can be restricted to silica and sodium, as these are
most effective for monitoring carryover. Where saturated steam is the only steam
sample, chloride, sulfate, TOC, and conductivity analyses should also be included.
Data from saturated steam and boiler water samples can be used to determine the
percentage of mechanical carryover from the drum water to the steam.
Main/Reheat Steam
Main or reheat steam sample data is often even more informative than satu-
rated steam data because any contaminants introduced by the attemperator sys-
tem will also be measured. For units without a reheat section, main steam is the
only choice, but for units with reheat, the reheat sample is preferred. As with sat-
urated steam, main/reheat samples are used to verify that steam quality is with-
in the turbine guidelines. Recommended continuous analyses include sodium,
silica, conductivity, and chloride, with sulfate and TOC being monitored on a
grab sample basis.
Cogeneration/Combined-Cycle/Industrial
Plant Sampling
At cogeneration, combined-cycle, or industrial steam generating facilities,
sample point selection may be slightly different. For example, heat recovery
steam generators often consist of two or three steam generating circuits and
drums. Steam and drum water samples from each circuit should be obtained if a
comprehensive analysis of the boiler water chemistry is desired.
Sampling of condensate return at industrial plants is very important because
many contaminants may be introduced by process heat exchangers or other
equipment. If, as is very common, the condensate return is treated in a polisher,
the polisher should be sampled similarly to those at electric utilities. However,
more emphasis on TOC analyses is needed when organics or oils have the poten-
tial to enter the condensate. As an example, at one petrochemical plant I had the
chance to visit, all of the condensate return lines to the boilers are equipped with
continuous TOC analyzers. If an analyzer detects organic concentrations above
plant guidelines, the analyzer triggers an automatic dump valve.
These paragraphs have covered the sample points and analyses for various
steam generating systems. The following sections discuss sample collection and
conditioning, topics that are extremely important with regard to chemistry mon-
itoring.
Sampling 215
these aspects is carefully considered in the sample system design, the reliability
of the samples will be seriously compromised.
Figure 7-1
Figure 7-2
Figure 7-3
Each nozzle is designed with considerations of vortex shedding, resonance, vibration, erosion,
and strength of the attachment to the pipe.
Materials: Ordering Information:
Carbon Steel (C1018 or A105, as specified) 1. Pressure, temperature, and mass flow
Stainless Steel (304 or 316, as specified) rate of the sampled fluid
Low Alloy Chrome Moly Steel 2. Pipe ID, wall thickness, and material.
(F11 or F22, as specified) 3. Desired sample flow.
Other materials available upon request 4. Attachment to the pipe: weld, thread
plus seal weld, flange, etc.
5. Nozzle material.
6. Thickness of thermal insulation.
Pressure-Temperature Rating
lbs. per sq. inch
Temperature - ˚F
Material 70˚ 200˚ 400˚ 600˚ 800˚ 1000˚ 1200˚
EPRI Isokinetic Steam Sampling Nozzle Supplied by Jonas & Consultants. Reprinted courtesy of Jonas &
Consultants, Wilmington, DE.
Sampling 217
Features of the nozzle include a beveled entry port, insertion to a depth that sam-
ples the bulk fluid but minimizes the moment of the flowing water on the noz-
zle, and a reinforcing boss to provide strength to the nozzle.
Liquid sampling is usually straightforward, while saturated steam, on the
other hand, is more difficult to sample. These difficulties are caused by the fact
that the fluid is at saturation temperature and may contain fractional water vapor,
or may develop more moisture as the steam cools in the sample line. Thus, the
sample is not totally homogeneous. Multiport sampling is needed to collect rep-
resentative saturated steam samples. At some utilities, the boiler drums have
been designed with multiple sample ports installed in the drum shell. Where this
has not been done, the generic multiport nozzle outlined in Figure 7-2 may be
installed in the saturated steam line. The multiple ports are designed to balance
out variabilities in the constituency of the fluid.
Some debate still exists over sample extraction techniques for superheated
steam. The sample itself is much more homogeneous than saturated steam
Figure 7-4
STEAM FLOW
>
DOUBLE ISOLATION
VALVES
NOZZLE
TO SAMPLE PANEL
>
SAMPLE COOLER
because superheating converts the fluid to a single phase, and passage through
the superheater or reheater tubes thoroughly mixes the fluid. The debate centers
around isokinetic sampling, whether it is needed and whether it is totally possi-
ble. Figure 7-3 illustrates the generic outline of several isokinetic nozzles
designed by EPRI. The nozzle opening is oriented in the path of the steam flow.
Jonas and Consultants of Wilmington, Delaware offer these isokinetic nozzles for
installation similar to that shown in Figure 7-4.
A few experts are not convinced that even these nozzles offer true isokinetic
sampling. Additionally, concern exists that the nozzle could potentially fracture
and be carried downstream to the turbine. Even so, the trend towards isokinetic
sampling is growing stronger as utility and industrial personnel realize the impor-
tance of representative sampling. However, any user must be aware of the poten-
tial hazards and should install the nozzles according to approved guidelines.
Sample Nozzle
Installation
Except for low-pressure lines such as condensate storage tank discharge,
condensate pump discharge, and deaerator inlet, the procedures for installing
sample taps must conform to appropriate boiler codes for welding and work on
high-pressure lines. The principal code that applies to this work is ANSI B31.1,
which discusses welding procedures for high-pressure pipe. It is absolutely vital
that personnel installing high-pressure taps follow this guideline, both to ensure
the integrity of the sample tap and to prevent problems such as nozzle breakage.
One important requirement of the code is that the pipe be heat treated so that
the welding does not compromise the integrity of the pipe material. These pro-
cedures make sample tap installation into high-pressure lines rather expensive.
Sampling 219
recommendations for both the 3:00 and 1:30 positions. Either should ensure a
good sample with minimal introduction of solids that may have settled along the
bottom of the pipe.
found to be best for maintaining the integrity of the samples, and for either pre-
venting deposition of products on the sample tube walls or scouring of previ-
ously deposited products. With this criteria as a guideline, the volumetric flow
rate and line size become dependent on flow requirements of the instruments.
Sample line sizes of 1/4 in. OD and 3/8 in. OD are most common. Volumetric flow
rates in these lines to maintain a 6 foot per second linear rate are 1200 cc/min
and 3300 cc/min, respectively. The Sentry Equipment Corporation recommends
the following line sizes for steam samples.
In some cases, the 6 ft/sec linear flow rate taxes the capability of the sample
conditioning equipment. Sampling researchers have found that flow rates can
often be lowered to 3 ft/sec without compromising the integrity of the sample.
When laying out sample line routes, the designer should attempt to mini-
mize the number of elbows and fittings. Horizontal lines should be installed on
level grade or set at a constant slope. Low spots in the lines create problems.
Sampling 221
Figure 7-5
PI
TI
FI
FI FI FI
TO GRAB TO ON-LINE
DRAIN SAMPLE INSTRUMENTS
to less than 100 psi. Further fine-tuning of the pressure is then accomplished
with the pressure regulator. The high-temperature solenoid is a safety device that
will cut off sample flow if a preset temperature is exceeded. It protects down-
stream equipment against high-temperatures caused by loss of cooling water to
the roughing coolers.
The low-pressure blowdown line is used to send samples to waste during
unit startup. At unit startup, power plant fluids, especially boiler water, contain
many particulates that could foul the downstream equipment. Sometimes the
blowdown line is installed ahead of the high-pressure regulator so that these par-
ticulates do not foul the high-pressure reducer and pressure regulator. However,
safety factors involved with blowing down high-pressure samples have generat-
ed greater interest in the low-pressure blowdown design.
The sampling system also includes a secondary cooler. The most accurate
readings are obtained if the sample is cooled to 77˚F, ±1 /2˚F. Some instrument
manufacturers claim that their temperature compensation devices eliminate the
need for secondary cooling, but the arguments are not yet convincing. Unless a
manager’s budget does not allow for secondary coolers, the coolers should be
installed to help ensure representative readings.
Figure 7-6
Complete Sample Conditioning System. Photo courtesy of the Sentry Equipment Corp., Oconomowoc, WI.
A chilled water supply is usually required for the secondary coolers. The
water can either be passed through the coolers and discharged or can be circu-
lated through an isothermal bath and be reused over and over. Recirculation
saves on water costs but may increase the size of the chiller.
The rotameters/flow controllers are used to set sample flow rates at recom-
mended values. The combination of a main sample rotameter and individual
flow rotameters provides flexibility in adjusting both the total sample flow rate
and the flow rate to each on-line instrument.
The backpressure regulator helps maintain a constant flow to the system,
and minimizes upsets.
Figures 7-6 and 7-7 illustrate how the equipment in a sample panel looks
when it is completely assembled.
Data Acquisition
Figure 7-8 shows the actual layout of the on-line water chemistry system at
the Dallman Power Station at my former utility. On-line instruments include
hydrazine, pH, specific conductivity, cation conductivity, sodium, silica, phos-
phate, and dissolved oxygen. Each analyzer sends continuous signals to a PLC
located in the analysis room. The PLC then communicates to various points in
the plant. It sends control signals to the hydrazine pumps of three generating
units. The pump stroking rate is controlled from signals sent by the deaerator
inlet hydrazine analyzers. The PLC also provides an audible alarm if condensate
pump discharge sodium levels exceed 5 ppb, or if drum pH drifts 0.4 units above
or below a predetermined range.
Sampling 223
Figure 7-7
Sample Conditioning Piping. Photo courtesy of the Sentry Equipment Corp., Oconomowoc, WI.
Finally, the PLC is the data collection source for continuous data display
screens located in the plant control room, main laboratory, and engineering
offices. Two of the screens are illustrated in Figures 7-9 and 7-10. The first is the
primary screen in the control room. Should any of the values exceed the ranges
shown, a red warning light will flash in the “Unit” boxes to the right. Not all of
the data collected by the analyzers is included on the screen, because the system
designers, including myself, did not want to overwhelm the operators with num-
bers. We wanted them to only be concerned with the most important readings.
Figure 7-10 is a simplified flow diagram of one of the units showing real-time
water chemistry data. This screen is particularly useful for personnel who wish
to look at any point in the system and immediately analyze the chemistry. Both
screens are color coded so that the plant staff can differentiate between units and
between process lines on the flow diagram.
Conclusion
Proper sampling of steam generation fluid samples is extremely important in
guaranteeing the overall reliability of the plant. Without representative sampling,
upsets could occur that would have potentially catastrophic results. This chapter
hopefully serves as a guideline for the equipment and procedures that go into the
design of a truly useful sampling system.
Ali Sadeghi Digital Library
07.qxd 8/8/06 1:59 PM Page 224
Figure 7-8
Laboratory Annex
Oxygen To Instrument
pH & Cond. Analyzers Analyzers Solenoid Valves
To Dallman
Sample Line Inlets PLC Control Room
(Data Display)
Si & Po4 (Alarm System)
Display Panel
Conductivity
pH, Oxygen &
Sink
Hydrazine Sodium Analyzers
Analyzers Analyzers To Main Laboratory
Communication (Data Display)
Routes = (Data Logging)
To Hydrazine
Pump Stroke
Actuators
Computer
Workstation
On-Line Sample System Arrangement at City Water, Light & Power, Dallman Generating Station
Springfield, IL.
Figure 7-9
Dallman Unit Chemistry
Unit 31
Unit 33 Unit32
Control Control
Unit 33 Range Unit32 Range Unit 31
Sampling 225
Figure 7-10
Sat. Steam
Na= 0.74 ppb Main Steam
Si= .15
Hot Reheat
Drum Water
pH = 9.39 Turbine
Si = 40 ppb
Phos = 0.0S
Spec.
Conc. = 15.1 Cold Reheat
E.C.P.D
Na = 0.0.S.
Cat.
Cond.= 0.210
Feedwater Heaters
Bibliography
Books
Applebaum, S.B., Demineralization by Ion Exchange, New York: Academic Press, 1968.
Byrne, W., Reverse Osmosis: A Practical Guide for Industrial Users, Littleton, Colorado:
Tall Oaks Publishing, Inc., 1995.
Cantafio, A.R., ed., Drew Principles of Industrial Water Treatment, eleventh edition,
Boonton, New Jersey: Drew Industrial Division Ashland Chemical Co., 1994.
Cohen, P., ed., The ASME Handbook on Water Technology for Thermal Power Systems,
New York: The American Society of Mechanical Engineers, 1989.
Dillon, C.P., Corrosion Control in the Chemical Process Industries, second edition,
Houston. Published for the Materials Technology Institute by NACE International, 1994.
Hensley, H.C., ed., Cooling Tower Fundamentals, second edition, Overland Park, Kansas:
The Marley Cooling Tower Company, 1985.
Herro, H.M., and R.D. Port, The Nalco Guide to Boiler Failure Analysis, New York:
McGraw-Hill.
Meller, F.H., ed., Electrodialysis (ED) & Electrodialysis Reversal Technology, Watertown,
Massachusetts: Ionics, Incorporated, 1984.
Owens, D.L., Practical Principles of Ion Exchange Water Treatment, Littleton, Colorado:
Tall Oaks Publishing, Inc., 1995.
230 Bibliography
Technical Reports
“Interim Consensus Guidelines on Fossil Plant Cycle Chemistry,” Palo Alto, California:
Electric Power Research Institute, June 1986. CS-4629.
Seminars
Borck, W., and G. Bartley, “Equilibrium Phosphate Treatment Conversion at TVA Gallatin
Fossil Plant.” In Proceedings of the 1996 International Water Conference, Pittsburgh,
1996.
Dooley, B., “The Cutting Edge of Cycle Chemistry for Fossil Plants.” In Proceedings of the
1996 International Water Conference, Pittsburgh, 1996.
Harfst, W., “Chlorine or Bromine: Which is Right for Your System?” Power, September
1993, pp. 70–72.
Jonas, O., L. DuPree, T. Gilchrist, G. Lawrence, and M. Stevens, “Copper Deposition and
MW Loss: Problem and Its Solutions.” In Proceedings of the 1996 International Water
Conference, Pittsburgh, 1996.
Jonas, O., and B. Dooley, “Steam Chemistry and Its Effects on Turbine Deposits and
Corrosion.” In Proceedings of the 1996 International Water Conference, Pittsburgh, 1996.
Bibliography 231
Port, R.D., and H.M. Herro, “An Overview of Water-Formed and Steam-Formed
Deposits,” Industrial Water Treatment, May/June 1992, pp. 33–39.
Robinson, J.O., and A.W. Fynsk, “A Practical Guide to Avoiding Steam Purity Problems in
the Industrial Plan,” Proceedings of the 1992 International Water Conference, Pittsburgh,
1992.
Smock, B., ed., “Water Treatment: Boiler Feedwater and Condensate.” Series of reprints
from Power Engineering magazine.
Stockdill D., and G. Reck, “Full Scale Evaluation of Hydrazine Alternatives in a High
Pressure Boiler,” Paper No. 204 presented at Corrosion 94. Published by the National
Association of Corrosion Engineers, Houston.
Stodola, J., “Ten Years of Equilibrium Phosphate Treatment.” In Proceedings of the 1996
International Water Conference, Pittsburgh, 1996.
Strauss, S., ed., S. Keen, and P. Puckorius, “Boiler Water Treatment for Low- and
Moderate-Pressure Plants,” Power Special Report, June 1987.
Thaler, J.O., and P.K. Sinha, “Control Parameter Model Speeds Water Treatment
Calculations,” Power Engineering, November 1987, pp. 31–32.
Wiltsey, D.G., and C.B. Batton, “Hydrazine Alternative Optimizes Corrosion Control: A
Twelve-Year History of Results in Utility Cycles.” Presented at Power-Gen Europe, Paris,
May 25–27, 1993.
Index
Acid feed,
for alkalinity reduction in cooling water .….….….….…178-179, 198-199
Acid phosphate corrosion of boiler tubes .….….….….….….….….….….… 66
Acrylates
boiler water treatment .….….….….….….….….….….….….….….….…68
cooling water treatment .….….….….….….….….….….….….….….…180
Activated carbon
guidelines for selection .….….….….….….….….….….….….….….…151
makeup water pretreatment .….….….….….….….….….…102-103, 106
Admiralty condenser tubes .….….….….….….….….….….…6, 47, 169, 201
Aeration .….….….….….….….….….….….….….….….….….….….….…106
Air inleakage to condensers and condensate/feedwater systems .…10, 46-47
Air removal from condensers .….….….….….….….….….….….….…10-13
Algae .….….….….….….….….….….….….….….….….….….….….…184
Alkaline cooling water treatment programs .….….….….….….….…179-181
Alkalinity,
bicarbonate .….….….….….….….….….….….….….….….….…178-179
carbonate .….….….….….….….….….….….….….….….….….…178-179
control in cooling towers .….….….….….….….….….….….….…178-179
effect on strong acid cation exchanger performance .….….….….…117-118
effect on weak acid cation exchanger performance .….….….….….…124
hydroxide .….….….….….….….….….….….….….….….….….…178-179
“M” .….….….….….….….….….….….….….….….….….….….…178-179
“P” .….….….….….….….….….….….….….….….….….….….… 178-179
All-volatile treatment (AVT) .….….….….….….….….….….…23, 34, 69-70
American Society of Mechanical Engineers boiler and feedwater chemistry guide-
lines .….….….….….….….….….….….….….….….….….….….…24-30
Amines
distribution ratio .….….….….….….….….….….….….….….….….…21
effect on sodium-to-phosphate ratios in boiler water .….….….….….…64
filming .….….….….….….….….….….….….….….….….….….….….…21
function of in steam-generating systems .….….….….….….….….…7, 20-21
neutralizing .….….….….….….….….….….….….….….….….….…20-21
used in boiler layup .….….….….….….….….….….….….….….….….…73
Aminoethylenephosphonic acid (AMP) .….….….….….….….….….….…180
Ammonia,
as a condensate/feedwater treatment .….….….….….….….….….….…7, 20
as a copper-alloy tubed heat exchanger corrodent .…7, 12, 15, 21, 47, 169
effects on condensate polishing .….….….….….….….….….….….….…36
effect on sodium-to-phosphate ratios in boiler water .….….….…63-64, 76
in oxygenated treatment programs .….….….….….….….….….….…22-23
reaction with chlorine in cooling water systems .….….….….….….….…185
used in boiler layup .….….….….….….….….….….….….….….….….…73
Anion exchange resin (see Ion exchange)
Anodic corrosion protection in cooling water systems .….….….….…170-171
Antiscalants for RO systems .….….….….….….….….….….….….….….…140
Attemperation
effects on steam purity .….….….….….….….….….….….….…84, 91, 93
detection of leaks .….….….….….….….….….….….….….….….….…228
Azoles
Copper corrosion inhibitors .….….….….….….….….….….…172-173, 181
Backwash
ion exchange resins .….….….….….….….….….….….….….….…128-129
media in filtration units .….….….….….….….….….….….….….….…105
Bacteria .….….….….….….….….….….….….….….….….….….….….…184
Barium sulfate scale on RO membranes .….….….….….….….….….….…140
Blowdown,
Boiler .….….….….….….….….….….….….….….….….….….….…56, 62
cooling tower .….….….….….….….….….….….….….….….….…162-163
Boilers
circulating fluidized bed (CFB) .….….….….….….….….….….….….…53
drum .….….….….….….….….….….….….….….….….….….….…51-54
field-erected .….….….….….….….….….….….….….….….….….…52-53
heat recovery steam generators (HRSG) .….….….….….….….….….…53-54
once-through .….….….….….….….….….….….….….….….….….….…54
package .….….….….….….….….….….….….….….….….….….….…52-54
Boiler chemical cleaning .….….….….….….….….….….….….….…20, 55, 78
Boiler drum internals .….….….….….….….….….….….….….….….…88-89
Boiler layup .….….….….….….….….….….….….….….….….…72-73, 80-81
Boiler tubes
corrosion .….….….….….….….….….….….….….….….….…57-60, 80-81
deposits.….….….….….….….….….….….….….….….….….….….…57-60
Boiler Water Chemistry .….….….….….….….….….….….….….….…51-82
ASME guidelines .….….….….….….….….….….….….….….….….…24-30
effects of condenser tube leaks .….….….….….….….….….….….…56-58
guidelines to control steam chemistry .….….….….….….….….….….…92
water chemistry upsets and effects on steam chemistry .….….….….…97-99
Boiler water treatment programs .….….….….….….….….….….….…2, 60-73
all-volatile-treatment (AVT) .….….….….….….….….….….….….…69-70
early programs .….….….….….….….….….….….….….….….….…61-62
caustic .….….….….….….….….….….….….….….….….….….….….…69
chelants and polymers .….….….….….….….….….….….….….….…67-68
coordinated/congruent phosphate .….….….….….….….….….….…62-64
equilibrium phosphate (EPT) .….….….….….….….….….….….….…66-67
oxygenated treatment (OT) .….….….….….….….….….….…21-23, 70, 79
phosphate treatment (PT) .….….….….….….….….….….….….….….…67
treatment for heat recovery steam generators .….….….….….….….…70-72
Bromine
cooling water microbiocide .….….….….….….….….….….….….…187-188
pretreatment for makeup water production .….….….….….….….….…102
Calcium
removal in ion exchange units .….….….….….….….….….…108-109, 112
Calcium carbonate scale
in boilers .….….….….….….….….….….….….….….….….….….….….…3
in cooling water systems .….….….….….….….….….….….….…3, 173-176
on RO membranes .….….….….….….….….….….….….….….….….…140
Calcium fluoride scale
in cooling water systems .….….….….….….….….….….….….….….…173
on RO membranes .….….….….….….….….….….….….….….….….…140
Calcium hypochlorite .….….….….….….….….….….….….….….….….…186
Calcium phosphate scale
in cooling water systems .….….….….….….….….….….….…3, 173, 177
Calcium sulfate scale
in cooling water systems .….….….….….….….….….….…3, 173, 177, 179
on RO membranes .….….….….….….….….….….….….….….….….…140
Carbohydrazide
metal passivator .….….….….….….….….….….….….….….….….….…16
oxygen scavenger .….….….….….….….….….….….….….….….….….…16
Carbon dioxide
corrosive effects of .….….….….….….….….….….….….….….…2, 3, 8, 17
produced by decomposition of water treatment chemicals 16-17, 18, 21, 70
Carbonic acid corrosion .….….….….….….….….….….….….….….….….…8
sample flow
flow rate recommendations .….….….….….….….….….….….…219-220
line size .….….….….….….….….….….….….….….….….….…219-220
sample nozzles
design of liquid sample nozzles .….….….….….….….….….….…215-216
design of steam sample nozzles .….….….….….….….….….….…216-218
installation recommendations .….….….….….….….….….….….….…218
sample point location to prevent flow interferences.….….….….….….…218
Sample parameters
ammonia .….….….….….….….….….….….….….….….….…33, 207-211
cation conductivity .….….….….….….….….….….….….….…33, 207-211
chloride .….….….….….….….….….….….….….….….….….….…207-211
conductivity .….….….….….….….….….….….….….….….….….….…33
copper .….….….….….….….….….….….….….….….….….….…207-211
degassed cation conductivity .….….….….….….….….….….….…207-211
dissolved oxygen .….….….….….….….….….….….….….….…33, 207-211
iron .….….….….….….….….….….….….….….….….….….….…207-211
oxygen scavengers .….….….….….….….….….….….….….…33, 207-211
pH .….….….….….….….….….….….….….….….….….….…33, 207-211
phosphate .….….….….….….….….….….….….….….….….…78, 207-211
silica .….….….….….….….….….….….….….….….….….…207-211, 227
sodium .….….….….….….….….….….….….….….….…33, 48, 207-211
specific conductivity .….….….….….….….….….….….….….….…207-211
sulfate .….….….….….….….….….….….….….….….….….….…207-211
total organic carbon (TOC) .….….….….….….….….….….….….…33, 38
Sample point selection
boiler water .….….….….….….….….….….….….….….…72, 207-211, 213
condensate pump discharge (CPD) .….….….….….…33, 207-211, 211-212
condensate polisher outlet (CPO) .….….….….….….….…33, 207-211, 212
condensate storage tank effluent.….….….….….….….….….….….….…211
deaerator inlet (DAI) .….….….….….….….….….….….…33, 207-211, 212
deaerator outlet (DAO) .….….….….….….….….….….…33, 207-211, 212
economizer inlet/feedwater .….….….….….….….…33, 207-211, 212-213
guidelines for various types of steam generating systems .….….…207-211
makeup system effluent .….….….….….….….….….….…131-132, 206-211
main steam .….….….….….….….….….….….….….….…94, 207-211, 214
reheat steam .….….….….….….….….….….….….…94, 100, 207-211, 214
saturated steam .….….….….….….….….….….….…94, 207-211, 213-214
Saturated steam .….….….….….….….….….….….….….….….….….….…94
Scale
control in boilers .….….….….….….….….….….….….….….….….…60-73
Titanium.….….….….….….….….….….….….….….….….….….….….….…6
Total Organic Carbon (TOC)
effects on boiler water and steam chemistry .….….….….….….….…98-99
in condensate return .….….….….….….….….….….….….….….….….…38
steam purity guidelines .….….….….….….….….….….….….….….….…92
Trisodium phosphate for boiler water treatment .….….….…61-63, 65, 66, 78
Turbidity .….….….….….….….….….….….….….….….….….….….….…105
Ultraviolet light for makeup treatment .….….….….….….….…102, 149-150
Underdeposit corrosion
in boilers .….….….….….….….….….….….….….…55, 56, 58-59, 67, 77
in cooling water systems .….….….….….….….….….….….….….….…168
Unit conversion program (COADE) .….….….….….….….….….….….…229
Vaporous carryover .….….….….….….….….….….….….….….….….….…90
Wet-bulb temperature .….….….….….….….….….….….….….….…163-165
Zebra mussels .….….….….….….….….….….….….….….….…186, 190-193
Zinc .….….….….….….….….….….….….….….….….….….….….…171-172