Daniel A. Crowl/Joseph F.

Lowar

C'-

A m aam -

11-111-1-1

Process

I 5econd

Edition

Prentice Hl International Series al in the Physical and Chemlcal Engineering Sciences

I

Chemical Process Safety

ISBN 0-13-OZ817b-5

9 780130 1 8 1 7 6 3

PRENTICE HALL INTERNATIONAL SERIES IN THE PHYSICAL AND CHEMICAL ENGINEERING SCIENCES NEALR. AMUNDSON, SERIES EDITOR, University of Houston

ANDREAS ACRIVOS, Stanford University JOHN DAHLER, Universityof Minnesota H. SCOTT FOGLER, University of Michigan THOMAS HANRATTY, J. University of Illinois JOHN . PRAUSNITZ, M University of California L. E. SCRIVEN, University of Minnesota

BALZHISER, SAMUELS, ELIASSEN Chemical Engineering Thermodynamics AND BEQUETTE Process Control: Modeling, Design and Simulation BEQUETTE Process Dynamics BIEGLER, GROSSMAN, WESTERBERGSystematic Methods of Chemical Process AND Design BROSILOW JOSEPH Techniques of Model-Based Control AND CROWL AND LOUVAR Chemical Process Safety: Fundamentals with Applications, 2nd edition CONSTANTINIDES MOSTOUFI Numerical Methods for Chemical Engineers AND with MATLAB Applications CUTLIP AND SHACHAM Problem Solving in Chemical Engineering with Numerical Methods DENN Process Fluid Mechanics DOYLE Process Control Modules: A Software Laboratory for Control Design ELLIOT AND LIRA Introductory Chemical Engineering Thermodynamics FOGLER Elements of Chemical Reaction Engineering, 3rd edition Basic Principles and Calculations in Chemical Engineering, 6th edition HIMMELBLAU HINESAND MADDOX Mass Transfer KYLE Chemical and Process Thermodynamics, 3rd edition PRAUSNITZ, LICHTENTHALER, DE AZEVEDO Molecular Thermodynamics AND of Fluid-Phase Equilibria, 3rd edition PRENTICE Electrochemical Engineering Principles SHULER AND KARGI Bioprocess Engineering, 2nd edition STEPHANOPOULOS Chemical Process Control TESTER AND MODELL Thennodynainics and Its Applications, 3rd edition TURTON, BAILIE,WHITING, AND SHAEIWITZAnalysis, Synthesis and Design of Chemical Processes WILKES Fluid Mechanics for Chemical Engineering

Prentice Hall International Series in the Physical and Chemical Engineering Sciences

Chemical Process Safety
Fundamentals with Applications
Second Edition

Daniel A. Crow1
Michigan Technological University

Joseph F. Louvar
Wayne State University

Prentice Hall PTR Upper Saddle River, New Jersey 07458 www.phptr.com

Library of Congress Cataloging-in-Publication data Crowl, Daniel A. Chemical process safety : fundamentals with applications I Daniel A. Crowl, Joseph F. Louvar. - 2nd ed. p. cm. - (Prentice Hall international series in the physical and chemical engineering sciences) Includes bibliographical references and index. ISBN 0-13-018176-5 1. Chemical plants -Safety measures. I. Louvar, Joseph F. 11. Title. 111. Series.

EditoriallProduction Supervision: G & S Typesetters, Inc. Acquisitions Editor: Bernard Goodwin Marketing Manager: Dan DePasquale Cover Design Direction: Jerry Votta Art Director: Gail Cocker-Bogusz Manufacturing Manager: Alexis R. Heydt-Long Editorial Assistant: Michelle Vincenti Project Coordinator: Anne R. Garcia

O 2002 by Prentice Hall PTR Prentice Hall, Inc. Upper Saddle River, NJ 07458
Prentice Hall books are widely used by corporations and government agencies for training, marketing, and resale. The publisher offers discounts on this book when ordered in bulk quantities. For more information, contact Corporate Sales Department, phone: 800-382-3419; fax: 201-236-7141;email: corpsales@prenhall.com Or write: Prentice Hall PTR Corporate Sales Department One Lake Street Upper Saddle River, NJ 07458 All rights reserved. No part of this book may be reproduced, in any form or by any means, without permission in writing from the publisher. Printed in the United States of America 10987654321 ISBN: 0-13-018176-5

Pearson Education Ltd. Pearson Education Australia PTY, Ltd. Pearson Education Singapore, Pte. Ltd. Pearson Education North Asia Ltd. Pearson Education Canada, Ltd. Pearson Educaci6n de Mexico, S.A. de C.V. Pearson Education - Japan Pearson Education Malaysia, Pte. Ltd. New Jersey Pearson Education, Upper Saddle

Contents

Preface xiii Nomenclature xv

1

Introduction 1
Safety Programs 2 Engineering Ethics 4 Accident and Loss Statistics 4 Acceptable Risk 12 Public Perceptions 14 The Nature of the Accident Process Inherent Safety 20 Four Significant Disasters 23 Flixborough, England 23 Bhopal, India 25 Seveso, Italy 26 Pasadena, Texas 27 Suggested Reading 29 Problems 30

15

2

Toxicology 35
2-1 How Toxicants Enter Biological Organisms 36 Gastrointestinal Tract 37 Skin 37 Respiratory System 38 How Toxicants Are Eliminated from Biological Organisms 39 Effects of Toxicants on Biological Organisms 40 Toxicological Studies 41 Dose versus Response 42

2-2 2-3 2-4 2-5

vi

Contents

2-6 2-7 2-8

Models for Dose and Response Curves 48 Relative Toxicity 54 Threshold Limit Values 54 Suggested Reading 59 Problems 59

3

Industrial Hygiene 63
3-1 Government Regulations 64 Laws and Regulations 64 Creating a Law 64 Creating a Regulation 64 OSHA: Process Safety Management 68 EPA: Risk Management Plan 71 Industrial Hygiene: Identification 74 Material Safety Data Sheets 74 Industrial Hygiene: Evaluation 78 Evaluating Exposures to Volatile Toxicants by Monitoring 79 Evaluation of Worker Exposures to Dusts 83 Evaluating Worker Exposures to Noise 84 Estimating Worker Exposures to Toxic Vapors 85 Industrial Hygiene: Control 94 Respirators 96 Ventilation 97 Suggested Reading 103 Problems 104

3-2 3-3

3-4

Source Models 109
4-1 4-2 4-3 4-4 Introduction to Source Models 109 Flow of Liquid through a Hole 112 Flow of Liquid through a Hole in a Tank 116 Flow of Liquids through Pipes 121 2-K Method 124 4-5 Flow of Vapor through Holes 130 4-6 Flow of Gases through Pipes 136 Adiabatic Flows 136 Isothermal Flows 143 4-7 Flashing Liquids 151 4-8 Liquid Pool Evaporation or Boiling 157 4-9 Realistic and Worst-Case Releases 159 4-10 Conservative Analysis 159 Suggested Reading 161 Problems 162

Contents

vii

5

Toxic Release and Dispersion Models 171
5-1 Parameters Affecting Dispersion 172 5-2 Neutrally Buoyant Dispersion Models 176 Case 1:Steady-State Continuous Point Release with No Wind 180 Case 2: Puff with No Wind 181 Case 3: Non-Steady-State Continuous Point Release with No Wind 182 Case 4: Steady-State Continuous Point Source Release with Wind 183 Case 5: Puff with No Wind and Eddy Diffusivity Is a Function of Direction 183 Case 6: Steady-State Continuous Point Source Release with Wind and Eddy Diffusivity Is a Function of Direction 184 Case 7: Puff with Wind 184 Case 8: Puff with No Wind and with Source on Ground 185 Case 9: Steady-State Plume with Source on Ground 185 Case 10: Continuous Steady-State Source with Source at Height Hr above the Ground 186 Pasquill-Gifford Model 186 Case 11: Puff with Instantaneous Point Source at Ground Level, Coordinates Fixed at Release Point, Constant Wind Only in x Direction with Constant Velocity u 190 Case 12: Plume with Continuous Steady-State Source at Ground Level and Wind Moving in x Direction at Constant Velocity u 191 Case 13: Plume with Continuous Steady-State Source at Height Hr above Ground Level and Wind Moving in x Direction at Constant Velocity u 192 Case 14: Puff with Instantaneous Point Source at Height Hr above Ground Level and a Coordinate System on the Ground That Moves with the Puff 193 Case 15: Puff with Instantaneous Point Source at Height Hr above Ground Level and a Coordinate System Fixed on the Ground at the Release Point 194 Worst-Case Conditions 194 Limitations to Pasquill-Gifford Dispersion Modeling 194 5-3 Dense Gas Dispersion 195 5-4 Toxic Effect Criteria 199 5-5 Effect of Release Momentum and Buoyancy 212 5-6 Release Mitigation 213 Suggested Reading 214 Problems 215

6

Fires and Explosions 225
6-1 6-2 The Fire Triangle 225 Distinction between Fires and Explosions 227

-

-

-

-

-

viii

Contents

6-3 6-4

6-5 6-6 6-7 6-8 6-9 6-10 6-11 6-12 6-13

Definitions 227 Flammability Characteristics of Liquids and Vapors 229 Liquids 230 Gases and Vapors 233 Vapor Mixtures 233 Flammability Limit Dependence on Temperature 235 Flammability Limit Dependence on Pressure 236 Estimating Flammability Limits 236 Limiting Oxygen Concentration and Inerting 238 Flammability Diagram 240 Ignition Energy 248 Autoignition 249 Auto-Oxidation 249 Adiabatic Compression 249 Ignition Sources 251 Sprays and Mists 252 Explosions 252 Detonation and Deflagration 253 Confined Explosions 255 Blast Damage Resulting from Overpressure 265 TNT Equivalency 269 TNO Multi-Energy Method 271 Energy of Chemical Explosions 274 Energy of Mechanical Explosions 276 Missile Damage 279 Blast Damage to People 279 Vapor Cloud Explosions 281 Boiling-Liquid Expanding-Vapor Explosions 282 Suggested Reading 282 Problems 283

7

Designs to Prevent Fires and Explosions 291
7-1 Inerting 292 Vacuum Purging 292 Pressure Purging 295 Combined Pressure-Vacuum Purging 297 Vacuum and Pressure Purging with Impure Nitrogen 298 Advantages and Disadvantages of the Various Pressure and Vacuum Tnerting Procedures 299 Sweep-Through Purging 299 Siphon Purging 301 Using the Flammability Diagram To Avoid Flammable Atmospheres 301 Static Electricity 307 Fundamentals of Static Charge 307

7-2

--

Contents

ix

7-3

7-4

7-5

7-6 7-7

Charge Accumulation 308 Electrostatic Discharges 309 Energy from Electrostatic Discharges 311 Energy of Electrostatic Ignition Sources 312 Streaming Current 313 Electrostatic Voltage Drops 316 Energy of Charged Capacitors 316 Capacitance of a Body 321 Balance of Charges 324 Controlling Static Electricity 330 General Design Methods To Prevent Electrostatic Ignitions 333 Relaxation 332 Bonding and Grounding 332 Dip Pipes 333 Increasing Conductivity with Additives 336 Handling Solids without Flammable Vapors 337 Handling Solids with Flammable Vapors 337 Explosion-Proof Equipment and Instruments 337 Explosion-Proof Housings 339 Area and Material Classification 339 Design of an XP Area 340 Ventilation 340 Open-Air Plants 340 Plants Inside Buildings 341 Sprinkler Systems 343 Miscellaneous Designs for Preventing Fires and Explosions 347 Suggested Reading 347 Problems 348

8

Introduction to Reliefs 353
8-1 Relief Concepts 354 8-2 Definitions 356 8-3 Location of Reliefs 357 8-4 Relief Types 360 8-5 Relief Scenarios 364 8-6 Data for Sizing Reliefs 365 8-7 Relief Systems 368 Relief Installation Practices 368 Relief Design Considerations 368 Horizontal Knockout Drum 371 Flares 375 Scrubbers 376 Condensers 376 Suggested Reading 376 Problems 377

x

Contents

9

Relief Sizing 383
9-1 9-2 9-3 9-4 9-5 9-6 Conventional Spring-Operated Reliefs in Liquid Service 384 Conventional Spring-Operated Reliefs in Vapor or Gas Service 389 Rupture Disc Reliefs in Liquid Service 394 Rupture Disc Reliefs in Vapor or Gas Service 394 Two-Phase Flow during Runaway Reaction Relief 395 Simplified Nomograph Method 401 Deflagration Venting for Dust and Vapor Explosions 404 Vents for Low-Pressure Structures 406 Vents for High-Pressure Structures 408 Venting for Fires External to Process Vessels 411 Reliefs for Thermal Expansion of Process Fluids 415 Suggested Reading 418 Problems 420

9-7 9-8

10

Hazards Identification 429
10-1 10-2 10-3 10-4 10-5 Process Hazards Checklists 432 Hazards Surveys 432 Hazards and Operability Studies 448 Safety Reviews 454 Other Methods 459 Suggested Reading 460 Problems 460

11

Risk Assessment 471
11-1 Review of Probability Theory 472 Interactions between Process Units 474 Revealed and Unrevealed Failures 480 Probability of Coincidence 484 Redundancy 486 Common Mode Failures 486 11-2 Event Trees 486 11-3 Fault Trees 491 Determining the Minimal Cut Sets 494 Quantitative Calculations Using the Fault Tree 497 Advantages and Disadvantages of Fault Trees 498 Relationship between Fault Trees and Event Trees 498 11-4 QRA and LOPA 499 Quantitative Risk Analysis 499 Layer of Protection Analysis 500 Consequence 503 Frequency 503 Suggested Reading 507 Problems 508

Contents

xi

Accident Investigations 515
Learning from Accidents 515 Layered Investigations 516 Investigation Process 518 Investigation Summary 519 Aids for Diagnosis 521 Fires 522 Explosions 522 Sources of Ignition in Vessels 523 Pressure Effects 523 Medical Evidence 525 Miscellaneous Aids to Diagnosis 525 12-6 Aids for Recommendations 528 Control Plant Modifications 528 User-Friendly Designs 529 Block Valves 529 Double Block and Bleed 530 Preventive Maintenance 530 Analyzers 531 Suggested Reading 532 Problems 532 12-1 12-2 12-3 12-4 12-5

13

Case Histories 535
13-1 Static Electricity 536 Tank Car Loading Explosion 536 Explosion in a Centrifuge 536 Duct System Explosion 537 Conductor in a Solids Storage Bin 537 Pigment and Filter 536 Pipefitter's Helper 536 Lessons Learned 536 13-2 Chemical Reactivity 540 Bottle of Isopropyl Ether 540 Nitrobenzene Sulfonic Acid Decomposition 540 Organic Oxidation 541 Lessons Learned 541 13-3 System Designs 546 Ethylene Oxide Explosion 546 Ethylene Explosion 546 Butadiene Explosion 546 Light Hydrocarbon Explosion 547 Pump Vibration 547 Pump Failure 547 Ethylene Explosion (1) 548

xii

Contents

Ethylene Explosion (2) 548 Ethylene Oxide Explosion 548 Lessons Learned 549 13-4 Procedures 551 Leak Testing a Vessel 552 Man Working in Vessel 552 Vinyl Chloride Explosion 552 Dangerous Water Expansion 553 Phenol-Formaldehyde Runaway Reaction 553 Conditions and Secondary Reaction Cause Explosion 554 Fuel-Blending Tank Explosion 555 Lessons Learned 556 13-5 Conclusion 556 Suggested Reading 557 Problems 557

Appendix A: Unit Conversion Constants 561 Appendix B: Flammability Data for Selected Hydrocarbons 565 Appendix C: Detailed Equations for Flammability Diagrams 571
Equations Useful for Placing Vessels into and out of Service 576

Appendix D: Formal Safety Review Report for Example 10-4 581 Appendix E: Saturation Vapor Pressure Data 591

More material is presented than can be accommodated in a 3-credit course. a senior-level undergraduate course. The second edition has been rewritten to include new process safety technology and new references that have appeared since the first edition was published in 1990. source modeling. We continue to believe that a textbook on safety is possible only with both industrial and academic inputs. A complete set of problem solutions is available to instructors using the book in their curriculum. explosion venting. The industrial input ensures that the material is industrially relevant. fundamentals of electrostatics. It can be used by anyone interested in improving chemical process safety. The primary objective of this textbook is to encapsulate the important technical fundamentals of chemical process safety.Preface his second edition of Chemical Process Safety is designed to enhance the process of teaching and applying the fundamentals of chemical process safety. flammability characterization. and case histories. It is appropriate for an industrial reference. or a graduate course in chemical process safety. The T . providing instructors with the opportunity to emphasize their topics of interest. This application requires a significant quantity of fundamental knowledge and technology. These changes fulfill the requests of many professors who have used this textbook. This new edition also includes selected materials from the latest AICHE Center for Chemical Process Safety (CCPS) books and is now an excellent introduction to the CCPS library. It also includes our combined experiences of teaching process safety in both industry and academia during the past 10 years. This second edition also includes more problems (now 30 per chapter). Significant modifications were made to the following topics: dispersion modeling. including chemical and mechanical engineers and chemists. The emphasis on the fundamentals will help the student and practicing scientist to understand the concepts and apply them accordingly.

who have extensive industrial experience and are now consultants. many universities have developed courses or course content in chemical process safety. we continue to acknowledge our families.) and the other (D. Boicourt and J. Since the first edition was published. Some will remark that our presentation is not complete or that some details are missing. This new emphasis on process safety is the result of the positive influences from industry and the Accreditation Board for Engineering and Technology (ABET). understanding. We also continue to acknowledge and thank all the members of the Undergraduate Education Committee of the Center for Chemical Process Safety and the Safety and Loss Prevention Committee of the American Institute of Chemical Engineers. Howard and S. Based on faculty feedback. A. Although professors normally have little background in chemical process safety. C. Grossel. Professors have also found that industrial employees are enthusiastic and willing to give specific lectures on safety to enhance their courses. however. Crowl and Joseph E: Louvar . Wehman of the BASF Corporation. D. The purpose of this book. their enthusiasm and knowledge have been truly educational and a key inspiration to the development of this text. has a companion text titled Health and Environmental Risk Analysis that extends the topics relevant to risk analysis. Darby of Texas A&M University. who provided patience. is not to be complete but to provide a starting point for those who wish to learn about this important area. We are honored to be members of both committees. Although the authors are (now) both from universities. for example. The members of these committees are the experts in safety. this textbook is an excellent application of the fundamental topics that are taught in the first three years of the undergraduate education. we continue to believe that chemical process safety should be part of every undergraduate and graduate course in chemistry and chemical and mechanical engineering. We hope that this textbook helps prevent chemical plant and university accidents and contributes to a much safer future. Powers from Dow Chemical Company. Daniel A. Those who have been especially helpful include G. and encouragement throughout the writing of the first and second editions. they have found that the concepts in this text and the accompanying problems and solutions are easy to learn and teach. This text is an excellent reference for these courses. This textbook can also be used as a reference for a design course. F.xiv Preface academic input ensures that the material is presented on a fundamental basis to help professors and students understand the concepts. just as it is a part of all the industrial experiences. Welker of the University of Arkansas. and R. W. R. B. Willey of Northeastern University. This textbook is designed for a dedicated course in chemical process safety.) has accumulated significant industrial experience since the writing of the first edition. However. This book. Finally. one has over 30 years of relevant experience in industry (J. We thank many of our friends who continue to teach us the fundamentals of chemical process safety. Hendershot from Rohm and Haas. L. R.

Dtid E a ERPG velocity of sound (lengthltime) area (length2)or Helmholtz free energy (energy).Nomenclature Do D". or process component availability tank cross sectional area (length2) change in Helmoltz free energy (energylmole) mass concentration (masslvolume) or capacitance (Farads) discharge coefficients (unitless) or concentration at a specified time (mass/volume) concentration of dense gas (volume fraction) heat capacity at constant pressure (energylmass deg) heat capacity at constant volume (energylmass deg) concentration in parts per million by volume deflagration vent constant (pressure1'*) average or mean mass concentration (mass/volume) diameter (length) particle diameter (length) diameter of flare stack (length) diffusion coefficient (arealtime characteristic source dimension for continuous releases of dense gases (length) characteristic source dimension for instantaneous releases of dense gas (length) reference diffusion coefficient (arealtime) molecular diffusivity (area /time) total integrated dose due to a passing puff of vapor (mass timelvolume) activation energy (energylmole) emergency response planning guideline (see Table 5-6) .

y or z direction (areattime) overpressure correction for relief sizing (unitless) explosion constant for dusts (length pressureltime) viscosity correction for relief sizing (unitless) .4) reference sound intensity (decibels) streaming current (amps) in-service oxygen concentration (volume percent oxygen) number of inerting purge cycles (unitless) electrical work (energy) non-ideal mixing factor for ventilation (unitless) constants in probit a equations thermal conductivity of soil (energyllength time deg) mass transfer coefficient (lengthltime) backpressure correction for relief sizing (unitless) excess head loss for fluid flow (dimensionless) constants in excess head loss. Krn KG Kj KP Kst K" emergency exposure guidance levels (see section 5.4) Fanning friction factor (unitless) or frequency (lltime) failure density function mass fraction of vapor (unitless) frictional fluid flow loss term (energy mass) or force or environment factor fatal accident rate (fatalitiesIlO8 hours) forced expired volume (literslsec) forced vital capacity (liters) gravitational acceleration (lengthltime2) gravitational constant initial cloud buoyancy factor (lengthltime2) Gibbs free energy (energylmole) or mass flux (masslarea time) mass flux during relief (masstarea time) change in Gibbs free energy (energylmole) specific enthalpy (energylmass) fluid level above leak in tank (length) initial fluid level above leak in tank (length) leak height above ground level (length) enthalpy (energylrnole) or height (length) flare height (length) effective release height in plume model (length) change in enthalpy (energylrnole) heat of combustion (energylmass) release height correction given by Equation 5-64 enthalpy of vaporization (energylmass) sound intensity (decibels) pipe internal diameter (length) immediately dangerous to life and health (see section 5. AH" I ID IDLH 10 Is ISOC j J k kl. k2 ks K Kb Kf Ki. given by Equation 4-38 explosion constant for vapors (length pressureltime) eddy diffusivity in x.xvi Nomenclature EEGL f f(t) fv F FAR FEV FVC g gc go G GT A G h h~ h"L hs H H f H I AH AC H AH.

Nomenclature xvii KO K* L LEL LFL = LEL LOC m "no ~ T N T mv M Mo Ma MOC. Q. Qv r R R R d RHI rf R P reference mass transfer coefficient (lengthltime) constant eddy diffusivity (arealtime) length lower explosion limit (volume %) lower flammability limit (volume %) limiting oxygen concentration (volume percent oxygen) mass total mass contained in reactor vessel (mass) mass of TNT mass of vapor molecular weight (masslmole) reference molecular weight (masslmole) Mach number (unitless) See LOC mean time between coincidence (time) mean time between failure (time) number of moles out of service fuel concentration (volume percent fuel) partial pressure (forcelarea) number of dangerous process episodes scaled overpressure for explosions (unitless) total pressure or probability backpressure for relief sizing (psig) permissable exposure level (see section 5. defined by equation 6-25 (unitless) release duration for heavy gas releases (time) reaction hazard index defined by Equation 13-1 vessel filling rate (time-') ideal gas constant (pressure volume/mole deg) .4) probability of failure on demand gauge pressure (forcelarea) maximum pressure for relief sizing (psig) set pressure for relief sizing (psig) saturation vapor pressure heat (energylmass) or heat intensity (energylarea time) heat intensity of flare (energyltime area) heat flux from ground (energylarea time) specific energy release rate at set pressure during reactor relief (energylmass) heat (energy) or electrical charge (coulombs) mass discharge rate (massltime) instantaneous mass release (mass) ventilation rate (volumeltime) radius (length) electrical resistance (ohms) or reliability Sachs scaled distance. MSOC MTBC MTBF n OSFC P P d P s P Pb PEL PFD P P Pmax f's pa st 4 4 f q g 4 s Q Q".

t" tw At" T Td T i TLV Tm TWA TXD U - Ud U (4 U UEL UFL = UEL v Vf vg vfs v v c W w e w s X X f Y Y YG Z Reynolds number (unitless) entropy (energylmole deg) or stress (forcelarea) material strength (forcelarea) short term public exposure guideline (see section 5.4) velocity (lengthltime) dropout velocity of a particle (lengthltime) average velocity (lengthltime) mean or average velocity (lengthhime) internal energy (energylmole) or overall heat transfer coefficient (energylarea time) or process component unavailability upper explosion limit (volume %) upper flammability limit (volume %) specific volume (volumelmass) specific volume of liquid (volumelmass) specific volume of vapor (volumelmass) specific volume change with liquid vaporization (volumelmass) total volume or electrical potential (volts) container volume width (length) expansion work (energy) shaft work (energy) mole fraction or Cartesian coordinate (length) distance from flare at grade (length) mole fraction of vapor (unitless) or Cartesian coordinate (length) probit variable (unitless) gas expansion factor (unitless) height above datum (length) or Cartesian coordinate (length) or compressibility (unitless) scaled distance for explosions (lengthlma~sl'~) .4) time positive phase duration of a blast (time) emptying time time to form a puff of vapor vessel wall thickness (length) worker shift time venting time for reactor relief temperature (deg) material decomposition temperature (deg) time interval threshold limit value (ppm or mg/m3 by volume) maximum temperature during reactor relief (deg) saturation temperature at set pressure during reactor relief (deg) time weighted average (ppm or mg/m3 by volume) toxic dispersion method (see section 5.xviii Nomenclature Re S s m SPEGL t td te tl.

Nomenclature xix Greek Letters velocity correction factor (unitless) or thermal diffusivity (arealtime) thermal expansion coefficient (deg-') double layer thickness (length) pipe roughness (length) or emissivity (unitless) relative dielectric constant (unitless) permittivity constant for free space (charge2/forcelength2) explosion efficiency (unitless) nonideal filling factor (unitless) heat capacity ratio (unitless) conductivity (mholcm) function defined by Equation 9-6 frequency of dangerous episodes average frequency of dangerous episodes viscosity (mass/length/time) or mean value or failure rate (faultsltime) vapor viscosity (mass/length/time) overall discharge coefficient used in Equation 9-15 (unitless) density (mass/volume) liquid density (mass/volume) reference density for specific gravity (mass/volume) vapor density (mass/volume) standard deviation (unitless) dispersion coefficient (length) relaxation time inspection period for unrevealed failures operation period for a process component period required to repair a component period of unavailability for unrevealed failures zeta potential (volts) Subscripts ambient combustion formation or liquid vapor or gas higher pressure initiating event purges lower pressure maximum set pressure initial or reference Superscripts 0 I standard stochastic or random variable .

.

I . Safety andAccident Prevention in Chemical Operations. Since 1950.Introduction n 1987. H. Many industrialists even believe that the development and application of safety technology is actually a constraint on the growth of the chemical industry. More complex processes require more complex safety technology. 1. dispersion models representing the spread of toxic vapor through a plant after a release. p. (New York: Wiley. Fawcett and W. Solow. H. 2d ed. and 'H. As chemical process technology becomes more complex. Professor Solow concluded that the bulk of an economy's growth is the result of technological advances. It is reasonable to conclude that the growth of an industry is also dependent on technological advances. Robert M. Fawcett said. which is entering an era of more complex processes: higher pressure.significant technological advances have been made in chemical process safety. and exotic chemistry. Today. chemical engineers will need a more detailed and fundamental understanding of safety. 1982). H. This is especially true in the chemical industry. an economist at the Massachusetts Institute of Technology. Examples of the technology of safety include hydrodynamic models representing two-phase flow through a vessel relief." l This book sets out the fundamentals of chemical process safety. safety is equal in importance to production and has developed into a scientific discipline that includes many highly technical and complex theories and practices. received the Nobel Prize in economics for his work in determining the sources of economic growth. S. "To know is to survive and to ignore fundamentals is to court disaster. more reactive chemicals. Wood.

and the design of new engineering features to prevent loss. or economic loss in terms of both the incident likelihood and the magnitude of the loss or injury. but for convenience. "safety" has been replaced by "loss prevention. Risk: a measure of human injury. and risk are frequently-used terms in chemical process safety." This term includes hazard identification. or moving equipment. First. technical evaluation. or the environment. chemical plants are the safest of all manufacturing facilities. and toxic hazards. safety shoes. The word "safety" used to mean the older strategy of accident prevention through the use of hard hats. Recent advances in chemical plant safety emphasize the use of appropriate technological tools to provide information for making safety decisions with respect to plant design and operation. reactivity hazards. there are chemical hazards. Much more recently. These ingredients are System Attitude Fundamentals Experience Time You . Hazard: a chemical or physical condition that has the potential to cause damage to people. property. Second. The subject of this text is loss prevention. Their definitions are Safety or loss prevention: the prevention of accidents through the use of appropriate technologies to identify the hazards of a chemical plant and eliminate them before an accident occurs. there are the usual mechanical hazards that cause worker injuries from tripping. However. As will be shown later. Safety. Despite substantial safety programs by the chemical industry.- 2 Chapter I Introduction mathematical techniques to determine the various ways that processes can fail and the probability of failure. falling. Chemical plants contain a large variety of hazards. The main emphasis was on worker safety. as shown in Figure 1-2. and a variety of rules and regulations. 1-1 Safety Programs A successful safety program requires several ingredients. headlines of the type shown in Figure 1-1 continue to appear in the newspapers. These include fire and explosion hazards. the words "safety" and "loss prevention" will be used synonymously throughout. the potential always exists for an accident of catastrophic proportions. environmental damage. hazard.

It is especially recommended that employees (1)read and understand Fundamentals Attitude \ Experience Figure 1-2 The ingredients of a successful safety program.1-1 Safety Programs 3 Figure 1-1 Headlines are indicative of the public's concern over chemical safety. the participants must understand and use the fundamentals of chemical process safety in the design. the participants must have a positive attitude. Third. (2) to do what needs to be done. Fourth. . and (3) to record that the required tasks are done. and operation of their plants. construction. everyone must learn from the experience of history or be doomed to repeat it. the program needs a system (1)to record what needs to be done to have an outstanding safety program. Second. First. This includes the willingness to do some of the thankless work that is required for success.

and time to train or be trained. These statistics are valuable for determining whether a process is safe or whether a safety procedure is working effectively. 1-2 Engineering Ethics Most engineers are employed by private companies that provide wages and benefits for their services. Engineers have a responsibility to themselves. The three systems considered here are . Unfortunately. It is important to recognize the distinction between a good and an outstanding safety program. A good safety program eliminates the existing hazards as they are identified. Fifth. and the engineering profession. no single method is capable of measuring all required aspects. Safety must be given importance equal to production. family. whereas an outstanding safety program prevents the existence of a hazard in the first place. These statistics must be used carefully. Engineers are responsible for minimizing losses and providing a safe and secure environment for the company's employees. time to share experiences. everyone should recognize that safety takes time. safety audits. Many statistical methods are available to characterize accident and loss performance. shown in Table 1-1. Like most statistics they are only averages and do not reflect the potential for single episodes involving substantial losses. 1-3 Accident and Loss Statistics Accident and loss statistics are important measures of the effectiveness of safety programs.4 Chapter 1 Introduction case histories of past accidents and (2) ask people in their own and other organizations for their experience and advice. everyone (you) should take the responsibility to contribute to the safety program. Sixth. A good safety program identifies and eliminates existing safety hazards. and engineers must provide a service to the company by maintaining and improving these profits. All employees have the responsibility to be knowledgeable about safety and to practice safety. The commonly used management systems directed toward eliminating the existence of hazards include safety reviews. and proper application of technical knowledge. checklists. time to do the work. The most effective means of implementing a safety program is to make it everyone's responsibility in a chemical process plant. An outstanding safety program has management systems that prevent the existence of safety hazards. time to record results (for history). Part of this responsibility is described in the Engineering Ethics statement developed by the American Institute of Chemical Engineers (AICHE). A safety program must have the commitment from all levels within the organization. The older concept of identifying a few employees to be responsible for safety is inadequate by today's standards. fellow workers. community. hazard identification techniques. This includes time to study. The company earns profits for its shareholders.

5. using their knowledge and skill for the enhancement of human welfare. Engineers shall act in professional matters for each employer or client as faithful agents or trustees.1-3 Accident and Loss Statistics 5 Table 1-1 American Institute of Chemical Engineers Code of Professional Ethics Fundamental principles Engineers shall uphold and advance the integrity. All three methods report the number of accidents and/or fatalities for a fixed number of workers during a specified period. Engineers shall perform services only in areas of their competence. Fundamental canons 1. The OSHA incidence rate is calculated from the number of occupational injuries and illnesses and the total number of employee hours worked during the applicable period.000 Total hours worked by all employees during period covered. 6. and dignity of the engineering profession. their employers. 2. 3. striving to increase the competence and prestige of the engineering profession. 3. OSHA is responsible for ensuring that workers are provided with a safe working environment. and shall avoid conflicts of interest. A worker year is assumed to contain 2000 hours (50 work weekslyear X 40 hourslweek). and clients. and welfare of the public in the performance of their professional duties. Engineers shall continue their professional development throughout their careers and shall provide opportunities for the professional development of those engineers under their supervision. = . and dignity of the engineering profession by 1. Engineers shall act in such a manner as to uphold and enhance the honor. The OSHA incidence rate is based on cases per 100 worker years. Engineers shall hold paramount the safety. Engineers shall build their professional reputations on the merits of their services. or deaths per person per year. health. The OSHA incidence rate is therefore based on 200. The following equation is used: OSHA incidence rate (based on injuries and illness) Number of injuries and illnesses X 200. OSHA incidence rate. 7. Table 1-2 contains several OSHA definitions applicable to accident statistics. OSHA stands for the Occupational Safety andHealth Administration of the United States government. 4.000 hours of worker exposure to a hazard. Engineers shall issue public statements only in an objective and truthful manner. honor. being honest and impartial and serving with fidelity the public. and fatality rate. 2. integrity. fatal accident rate (FAR).

Injuries that result in the injured person not being able to perform their regular duties but being able to perform duties consistent with their normal work. Medical injuries include cuts requiring stitches. Injuries that require treatment that must be administered by a physician or under the standing orders of a physician. ingestion. caused by exposure to environmental factors associated with employment. absorption. Any injury such as a cut. ZOSHAregulations. other than one resulting from an occupational injury. Number of days (consecutive or not) after but not including the day of injury or illness during which the employee would have worked but could not do so. Number of occupational injuries and/or illnesses or lost workdays per 100 full-time employees. Cases of occupational injury or illness that do not involve fatalities or lost workdays but do result in (1) transfer to another job or termination of employment or (2) medical treatment other than first aid or (3) diagnosis of occupational illness or (4) loss of consciousness or (5) restriction of work or motion. ~ Term First aid Definition Any one-time treatment and any follow-up visits for the purpose of observation of minor scratches. 1999 ed. burns. (Chicago: National Safety Council. 29 CFR 1904. 151. It includes acute and chronic illnesses or diseases that may be caused by inhalation. during which the employee could not perform all or any part of his or her normal assignment during all or any part of the workday or shift because of the occupational injury or illness. The injured person is able to return to work and perform his or her regular duties. Incident rate Lost workdays Medical treatment Occupational injury Occupational illness Recordable cases Recordable fatality cases Recordable nonfatal cases without lost workdays Recordable lost workday cases due to restricted duty Recordable cases with days away from work Recordable medical cases 'Injury Facts. injury requiring prescription medication. Medical treatment does not include first aid treatment even though provided by a physician or registered professional personnel. Any abnormal condition or disorder. Treatment administered by a physician or by registered professional personnel under the standing orders of a physician. . or direct contact. and injury with loss of consciousness. sprain. Such one-time treatment and follow-up visits for the purpose of observation are considered first aid even though provided by a physician or registered professional personnel. 1999). splinters.12. Cases involving an occupational injury or occupational illness. including deaths. Injuries that result in death. p. Injuries that result in the injured person not being able to return to work on their next regular workday. second-degree burns (burns with blisters). broken bones. cuts. and so forth that do not ordinarily require medical care. or burn that results from a work accident or from a single instantaneous exposure in the work environment.Table 1-2 Glossary of Terms Used by OSHA and Industry to Represent Work-Related L o s s e ~ ~ . that is. regardless of the time between the injury and death or the length of the illness.

A FAR can be converted to a fatality rate (or vice versa) if the number of exposed hours is known. The definition of a lost workday is given in Table 1-2. . This provides a better representation of worker accidents than systems based on fatalities alone. The OSHA incidence rate provides information on all types of work-related injuries and illnesses. The resulting equation is Number of fatalities X 10' FAR = Total hours worked by all employees during period covered. including fatalities. Number of lost workdays X 200. The FAR reports the number of fatalities based on 1000 employees working their entire lifetime. fatality data cannot be extracted from the OSHA incidence rate without additional information. This approach is useful for performing calculations on the general population. For this case OSHA incidence rate (based on lost workdavs) .000 Total hours worked by all employees during period covered. This system is independent of the number of hours actually worked and reports only the number of fatalities expected per person per year. a plant might experience many small accidents with resulting injuries but no fatalities. This statistic is used here because there are some useful and interesting FAR data available in the open literature. The applicable equation is Number of fatalities per year Fatality rate = Total number of people in applicable population. = . where the number of exposed hours is poorly defined. TheFAR is usedmostly by the British chemicalindustry. On the other hand. The employees are assumed to work a total of 50 years. The OSHA incidence rate cannot be readily converted to a FAR or fatality rate because it contains both injury and fatality information.1-3 Accident and Loss Statistics 7 An incidence rate can also be based on lost workdays instead of injuries and illnesses. Thus the FAR is based on 10' working hours. Both the OSHA incidence rate and the FAR depend on the number of exposed hours. For instance. The last method considered is the fatality rate or deaths per person per year. An employee working a ten-hour shift is at greater total risk than one working an eight-hour shift.

A." Chemical Engineering (Apr. 1985 ed. One of these deaths will be due to direct chemical exposure. compute the deaths per person per year. A FAR of 1. Kletz. 1999 ed.27 7. However.49 1. 1986).3 8. Lees. 219. "Eliminating Potential Process Hazards.35 0. p.^ The FAR figures show that if 1000 workers begin employment in the chemical industry. If an employee works a standard 8-hr shift 300 days per year. 2 of the workers will die as a result of their employment throughout all of their working lifetimes.81 0. (Chicago: National Safety Council.53 5.3 5.28 0.8 X lo-'.p. Approximately half these deaths are due to ordinary industrial accidents (falling down stairs.54 2. 2d ed.2 0. (London: Butterworths.26 1. 1. Loss Prevention in the Process Industries (London: Butterworths.2 is reported in Table 1-3for the chemical industry.10 1. "rank P.68 19863 4. 4FrankP.06 2. Lees. 66.6 0.96 2. ZInjury Facts. Example 1-1 A process has a reported FAR of 2. 30.6 7. 20 of these same 2T. the other half to chemical exposure^.1985). 1999). .p.2 'Accident Facts.0 40 67 10 19904 1.0 3.88 4. 1996).p.22 3.28 3. 1985).35 6. being run over). Typical accident statistics for various industries are shown in Table 1-3. Loss Prevention in the Process Industries.07 1.0 1.7 1. 177.08 1.8 Chapter 1 Introduction Table 1-3 Accident Statistics for Selected Industries OSHA incident rate (cases involving days away from work and deaths) Industry Chemicals and allied products Motor vehicles Steel Paper Coal mining Food Construction Agricultural Meat products Trucking All manufacturing FAR (deaths) 1985l 0. Solution Deaths per person per year = (8 hrlday) x (300 dayslyr) x (2 deaths/108hr) = 4. (Chicago: National Safety Council.28 19982 0.89 0.

it appears that individuals are willing to take a substantially greater risk if it is voluntary. despite general perceptions otherwise. 2d ed. Of those that perish from disease. Lees. and (d) the total number of fatalities? "letz. Table 1-4 indicates that canoeing is much more dangerous than traveling by motorcycle.1-3 Accident and Loss Statistics 9 Table 1-4 Fatality Statistics for Common Nonindustrial Acti~itiesl. Based on these data.p. Lees.) Run over by vehicle 57 96 240 660 1000 4000 17 X lo-" 4 x lo-5 500 X 6 X lo-" 1 x lo-7 150 X lo-' 600 X lo-' 'Frank P.~ Table 1-4 lists the FARs for various common activities. (London: Buttenvorths. 1000 people will die as a result of nonindustrial accidents (mostly at home or on the road) and 370 will die from disease. 9/96.p. For example. ZFrank P. "Eliminating Potential Process Hazards. 40 will die as a direct result of ~moking. The total number of motorcycle fatalities is larger because more people travel by motorcycle than canoe.'' . (c) the fatality rate. Canoeing produces more fatalities per hour of activity than traveling by motorcycle. (b) the FAR.K.) Fire (U. Example 1-2 If twice as many people used motorcycles for the same average amount of time each. Loss Prevention in the Process Industries (London: Butterworths. This phenomenon is due to the number of exposed hours. 1996).~ FAR (deaths110' hours) Fatality rate (deaths per person per year) Activity Voluntary activity Staying at home Traveling by Car Bicycle Air Motorcycle Canoeing Rock climbing Smoking (20 cigaretteslday) Involuntary activity Struck by meteorite Struck by lightning (U. Loss Prevention in the Process Industries. 1986). The table is divided into voluntary and involuntary risks. what will happen to (a) the OSHA incidence rate. It is also evident that many common everyday activities are substantially more dangerous than working in a chemical plant. 178.K.

The total number of fatalities will double. part a. The OSHA incidence rate will remain the same. Recognizing that the chemical industry is safe. However. but the total number of hours exposed will double as well. the operator is the sole fatality. d. The second plant employs 10 operators. Both plants have a probability of explosion and complete devastation once every 1000 years.4 It is human nature to perceive the accident with the greater loss of life as the greater tragedy. and (d) the total number of fatalities? Solution a. or deaths per person per year. The fatality rate does not depend on exposed hours. tragedy. b. but there are a correspondingly larger number of exposed hours. "Eliminating Potential Process Hazards. Example 1-3 If all riders used their motorcycles twice as much. (b) the FAR. c. as. Rock climbing produces many more fatalities per exposed hour than traveling by car. for example. As a result. The fatality rate. The number of fatalities will double. Accident statistics can be somewhat misleading in this respect.10 Chapter 1 Introduction Solution a. the second accident kills more people. When this plant explodes all 10 operators succumb. will double. 4Kletz. In both cases the FAR and OSHA incidence rate are the same. The same reasoning applies as for Example 1-2. (c) the fatality rate. For example. The number of injuries and deaths will double. The FAR will remain unchanged for the same reason as in part a. Example 1-4 A friend states that more rock climbers are killed traveling by automobile than are killed rock climbing. The fatality rate will double. d. Twice as many fatalities will occur within this group. The FAR will remain unchanged for the same reason as in part a. India. b. The potential for large loss of life gives the perception that the chemical industry is unsafe. the statement might be correct but more data are required. The OSHA incidence rate will remain the same. Accident statistics do not include information on the total number of deaths from a single incident. The first plant employs a single operator. why is there so much concern about chemical plant safety? The concern has to do with the industry's potential for many deaths. in the Bhopal. c. When the plant explodes. In both cases the risk taken by an individual operator is the same. consider two separate chemical plants. Is this statement supported by the accident statistics? Solution The data from Table 1-4 show that traveling by car (FAR = 57) is safer than rock climbing (FAR = 4000). what will happen to (a) the OSHA incidence rate." . the rock climbers probably spend more time traveling by car than rock climbing.

p. initial expenditures are made ~ on safety. excessive expenditures for safety equipment to solve single safety problems may make the system unduly complex and consequently may cause new safety SLarge Property Damage Losses in the Hydrocarbon-Chemical Industries: A Thirty-Year Review (New York: J & H Marsh & McLennan Inc. Accidents of this type are much more common than fatalities. but the cost is decreased sales. and the average amount lost per incident have steadily increased. location. This is demonstrated in the accident pyramid shown in Figure 1-3.1-3 Accident and Loss Statistics N u m b e r o f Accidents Figure 1-3 The accident pyramid. This is part of risk management. an increase in chemical plant size. then the return increases more. and time. Indeed.A." . even higher safety expenditures result in uncompetitive product pricing: The company will go out of business. This results in increased return because of reduced loss expenditures. and an increase in the use of more complicated and dangerous chemicals. Each company needs to determine an appropriate level for safety expenditures. The increases are mostly due to an expansion in the number of chemical plants. If safety expenditures increase further. The exact numbers vary by industry. Finally. As shown by K l e t ~if. Loss data5 published for losses after 1966 and in 10-year increments indicate that the total number of losses. "No Damage" accidents are frequently called "near misses" and provide a good opportunity for companies to determine that a problem exists and to correct it before a more serious accident occurs. 2. "Eliminating Potential Process Hazards. It is frequently said that "the cause of an accident is visible the day before it occurs. has an optimal level of activity beyond which there are diminishing returns. the price of the product increases and sales diminish. Safety is good business and. These losses can be substantial. people are spared from injury (good humanity). the total dollar amount lost. safety reviews and careful evaluation of near misses will identify hazardous conditions that can be corrected before real accidents occur." Inspections. 1998).. 6T. If safety expenditures increase. The total loss figure has doubled every 10 years despite increased efforts by the chemical process industry to improve safety. plants are prevented from blowing up and experienced workers are spared. From a technical viewpoint. like most business situations.The numbers provided are only approximate. Property damage and loss of production must also be considered in loss prevention. but it may not be as much as before and not as much as achieved by spending money elsewhere. Kletz.

Engineers need to also consider other alternatives when designing safety improvements. if not all. When organizations focus on the root causes of worker injuries. assaults. company safety programs are directed toward preventing injuries to employees. 1999 ed. Although the emphasis of this book is the prevention of chemical-related accidents. mechanical and chemical exposures.p. the data in Figure 1-4 show that safety programs need to include training to prevent injuries resulting from transportation.200 (43. It is also important to recognize the causes of accidental deaths. 2. (Chicago: National Safety Council. At some point in the design stage someone needs to decide if the risks are "accept- . and fires and explosions. Every chemical process has a certain amount of risk associated with it. 1999).100 200 41.100 (5. 2Public accidents are any accidents other than motor-vehicle accidents that occur in the use of public facilities or premises (swimming.) and deaths resulting from natural disasters even if they happened in the home.100 5. This excessive expense could have a higher safety return if assigned to a different safety problem. the programs should include off-the-job safety.200 200 28.000 20. etc. especially training to prevent accidents with motor vehicles. 1-4 Acceptable Risk We cannot eliminate risk entirely. as shown in Table 1-5.4%) 28.0%) 20.2003 lZnjury Facts.000 (21.1%) 92.5%) 3.Because most. it is helpful to analyze the manner in which workplace fatalities occur (see Figure 1-4). 3The true total is lower than the sum of the subtotals because some accidents are in more than one category. falling.12 Chapter I Introduction Table 1-5 All Accidental Deaths1 Type of death Motor-vehicle Public nonwork Work Home Subtotal Work Nonmotor-vehicle Motor-vehicle Subtotal Home Nonmotor-vehicle Motor-vehicle Subtotal Public2 Subtotal Total accidental deaths 1998 Deaths 38. problems because of this complexity.000 2.900 2.400 (30. hunting.

Is it satisfactory to design a process with a risk comparable to the risk of sitting at home? For a single chemical process in a plant composed of several processes. The total number of workplace fatalities was 6026.630) Assaults and violent acts (n = 960) Contact with objects and equipment (n = 941) Falls (n = 702) Exposure due to harmful substances and environments (n = 572) Fires and explosions (n = 205) I I Highway Worker struck by vehicle highway Aircraft Other Homicide Suicide Struck by object Other Electrocutions Other I I I I I I I 5 10 15 20 25 Accidents (%) 30 35 40 45 Figure 1-4 The manner in which workplace fatalities occurred in 1998. 1999).' 7Modern site layouts require sufficient separation of plants within the site to minimize risks of multiple exposures. USDL 99-208 (Washington. this risk may be too high because the risks resulting from multiple exposures are additive. Aug. DC: US Department of Labor.1-4 Acceptable Risk 13 Transportation incidents (n = 2." That is. . able. Source: News. 4. are the risks greater than the normal day-to-day risks taken by individuals in their nonindustrial environment? Certainly it would require a substantial effort and considerable expense to design a process with a risk comparable to being struck by lightning (see Table 1-4).

more harm than good. The results show an almost even three-way split. A. is to eliminate synthetic fibers produced by chemicals and use natural fibers such as cotton. Kletz. Some naturalists suggest eliminating chemical plant hazards by "returning to nature. despite any statistics. more harm than good. 1-5 Public Perceptions The general public has great difficulty with the concept of acceptable risk. The results of a public opinion survey on the hazards of chemicals are shown in Figure 1-5. As suggested by Kletz? accident statistics demonstrate that this will result in a greater number of fatalities because the FAR for agriculture is higher. 8T.-29% --------More Harm Than Good -- 38% Same Amount of Good a n d H a r m Figure 1-5 Results from a public opinion survey asking the question "Would you say chemicals do more good than harm. Engineers must make every effort to minimize risks within the economic constraints of the process.This survey asked the participants if they would say chemicals do more good than harm. Chemical plant designers who specify the acceptable risk are assuming that these risks are satisfactory to the civilians living near the plant." One alternative. or about the same amount of each. with a small margin to those who considered the good and harm to be equal. or about the same amount of each?" Source: The Detroit News. "Eliminating Potential Process Hazards. The major objection is due to the involuntary nature of acceptable risk.14 Chapter 1 Introduction 28% More Good Than H a r m . for example." . No engineer should ever design a process that he or she knows will result in certain human loss or injury. Frequently these civilians are not aware that there is any risk at all.

fires are the most common. the order reverses. where a large cloud of volatile and flammable vapor is released and dispersed throughout the plant site followed by ignition and explosion of the cloud. Used by permission of Marsh Inc. 1-6 The Nature of the Accident Process Chemical plant accidents follow typical patterns. . The most damaging type of explosion is an unconfined vapor cloud explosion. As shown in Table 1-6. As illustrated. With respect to fatalities. Source: Large Property Damage Losses in the Hydrocarbon-Chemical Industries: A Thirty-Year Review (New York: Marsh Inc. gasoline. b. It is important to study these patterns in order to anticipate the types of accidents that will occur. paper. vapor cloud explosions ac- r Other 3% Figure 1-6 Types of loss for large hydrocarbonchemical plant accidents. Solution Penicillin.. concrete. An analysis of the largest chemical plant accidents (based on worldwide accidents and 1998 dollars) is provided in Figure 1-6. 2. plastic. 1998). Economic loss is consistently high for accidents involving explosions. synthetic rubber.1-6 The Nature of the Accident Process Table 1-6 Type of accident Fire Explosion Toxic release Three Types of Chemical Plant Accidents Probability of occurrence High Intermediate Low Potential for fatalities Low Intermediate High Potential for economic loss Intermediate High Low Example 1-5 List six different products produced by chemical engineers that are of significant benefit to mankind. followed by explosion and toxic release. with toxic release having the greatest potential for fatalities.

2. and cleanup liabilities can be significant. mechanical failures could all be due to human error as a result of improper maintenance or inspection. Pumps. valves are not opened or closed in the proper sequence or reactants are not charged to a reactor in the correct order. can be attributed to human error. Used by permission of Marsh Inc. Personnel injuries. The "other" category of Figure 1-6includes losses resulting from floods and windstorms. p. The second largest cause is operator error. Figure 1-7 presents the causes of losses for the largest chemical accidents. Process upsets caused by. valves. 1998). employee losses. Almost all accidents.. For example. legal compensation. and control equipment will fail if not properly maintained.16 Chapter I Introduction count for the largest percentage of these large losses. The Mechanical Operator error Unknown Process upsets Natural hazards Design Sabotage and arson Figure 1-7 Causes of losses in the largest hydrocarbon-chemicalplant accidents. By far the largest cause of loss in a chemical plant is due to mechanical failure. Failures of this type are usually due to a problem with maintenance. Source: Large Property Damage Losses in the Hydrocarbon-ChemicalIndustries: A Thirty-Year Review (New York: J & H Marsh & McLennan Inc. Toxic release typically results in little damage to capital equipment. power or cooling water failures account for 11% of the losses. for example. except those caused by natural hazards. For instance. . Human error is frequently used to describe a cause of losses.

The number and magnitude of the losses increase over each consecutive 10-year period for the past 30 years. An interesting result of this study is that the most complicated mechanical components (pumps and compressors) are minimally responsible for large losses. is likely the result of governmental regulations that were implemented in the United States during this time. includes human errors made on-site that lead directly to the loss. 1987)." used in Figure 1-7.1992. that is. Source: A Thirty-Year Review of One Hundred of the Largest Property Damage Losses in the Hydrocarbon-Chemical Industries (New York: Marsh Inc. The lower losses in the last 5-year period. Reprinted by permission. compared to the previous 5 years between 1987 and 1996. The loss distribution for the hydrocarbon and chemical industry over 5-year intervals is shown in Figure 1-9.. Piping system failure represents the bulk of the accidents. This increase corresponds to the trend of building larger and more complex plants. Figure 1-8 presents a survey of the type of hardware associated with large accidents. . followed by storage tanks and reactors." This rule became effective on May 26. on February 24. OSHA published its final rule "Process Safety Management of Highly Hazardous Chemicals.1-6 The Nature of the Accident Process 17 Piping systems Miscellaneous or unknown Storage tanks Reactor piping systems Process holding tanks Heat exchangers Valves Process towers Compressors Pumps Gauges 0 5 10 15 20 25 30 35 40 45 50 I Number of accidents Figure 1-8 Hardware associated with largest losses. term "operator error.

and termination (the event or events that stop the accident or diminish it in size). the valve stem shears off and flammable liquid begins to spew out. A cloud of flammable vapor rapidly forms and is ignited by a nearby truck. 1992. K . A worker walking across a high walkway in a process plant stumbles and falls toward the edge. The explosion and fire quickly spread to nearby equipment. propagation (the event or events that maintain or expand the accident).04) 27 Losses (1.5 .. and the plant is completely destroyed.It demonstrates an important point: Even the simplest accident can result in a major catastrophe.18 Chapter 1 Introduction .161. Most accidents follow a three-step sequence: initiation (the event that starts the accident). I n - P B 0.44) 17 Losses 18 Losses 16 Losses 9 Losses Figure 1-9 Loss distribution for onshore accidents for 5-year intervals over a 30-year period.5 - (0. Unfortunately. To prevent the fall.) Source: Large Property Damage Losses in the Hydrocarbon-Chemical Industries: A Thirty-Year Review (New York: J & H Marsh & McLennan Inc. 2. The accident was propagated by the shearing of the valve and the resulting explosion and growing fire. The following chemical plant accident illustrates these steps. The impact of these regulations occurred in subsequent years.0 - a a 1.L g a . This disaster occurred in 1969yand led to an economic loss of $4.48) " 1. Used by permission of Marsh Inc. In the example the worker tripped to initiate the accident. The resulting fire lasts for six days until all flammable materials in the plant are consumed. p. (1.0 . 3. 1986). 9 0 n e Hundred Largest Losses: A Thirty-Year Review of Property Lo.sses in the Hydrocarbon-Chemical Industries (Chicago: M & M Protection Consultants. Other countries are adopting similar regulations. Accidents follow a three-step process. he grabs a nearby valve stem.34) (1.39) 5 Losses (0. 2.000. . 1998). p.0 - r. (There were also 7 offshore accidents in this 30-year period. The event was terminated by consumption of all flammable materials.

Identify the initiation. A much more effective approach is to work on all three areas to ensure that accidents.35 MCFID) gas producing plant allowed the release of rich oil and gas at 850 psi and -40°F. . The 75' high X 10' diameter absorber tower eventually collapsed across the pipe rack and on two exchanger trains. Breaking pipelines added more fuel to the fire. Table 1-7 presents a few ways to accomplish this. waste heat recovery and super-heater train resulted in its near total destruction. accidents can be stopped by eliminating the initiating step. 10. Example 1-6 The following accident report has been filed lo: Failure of a threaded 1%" drain connection on a rich oil line at the base of an absorber tower in a large (1. Severe flame impingement on an 11.1-6 The Nature of the Accident Process 19 Table 1-7 Step Initiation Defeating the Accident Process Desired effect Diminish Procedure -- Grounding and bonding Inerting Explosion proof electrical Guardrails and guards Maintenance procedures Hot work permits Human factors design Process design Awareness of dangerous properties of chemicals Emergency material transfer Reduce inventories of flammable materials Equipment spacing and layout Nonflammable construction materials Installation of check and emergency shutoff valves Firefighting equipment and procedures Relief systems Sprinkler systems Installation of check and emergency shutoff valves Propagation Diminish Termination Increase Safety engineering involves eliminating the initiating step and replacing the propagation steps with termination events. propagation. The resulting vapor cloud probably ignited from the ignition system of enginedriven recompressors. do not propagate and will terminate as quickly as possible. p. l0One Hundred Largest Losses. once initiated.000-horsepower gas turbine-driven compressor. and termination steps for this accident. In theory. In practice this is not effective: It is unrealistic to expect elimination of all initiations.

redundancy. operated at less severe temperatures and pressures. the safety of a process relies on multiple layers of protection. safety shutdown systems. Inherent safety is a part of all layers of protection. Although a process or plant can be modified to increase inherent safety at any time in its life cycle. it is especially directed toward process design features.Inherently Safer Chemical Processes: A Life Cycle Approach (New York: American Institute of Chemical Engineers. A process that does not require complex safety interlocks and elaborate procedures is simpler.20 Chapter 1 introduction Failure of threaded 172"drain connection Release of rich oil and gas. Solution Initiation: Propagation: 1-7 Inherent Safety An inherently safe plant11J2relies on chemistry and physics to prevent accidents rather than on control systems. the study of case histories is an especially important step in the process of accident prevention. and they are free to consider basic process alternatives. interlocks. Subsequent layers include control systems. however. interlocks. ignition of vapor cloud by recompressors. formation of vapor cloud. Guidelines for Engineering Design for Process Safety (New York: American Institute of Chemical Engineers. 1993). Inherently safer plants are tolerant of errors and are often the most cost effective. The first layer of protection is the process design features. it is helpful to know the definitions of terms that are commonly used in the descriptions (see Table 1-8). . In general. An inherently safer plant is more tolerant of operator errors and abnormal conditions. protective systems. easier to operate. To understand these histories. The best approach to prevent accidents is to add process design features to prevent hazardous situations. At these early stages process engineers and chemists have the maximum degree of freedom in the plant and process specifications. and more reliable. and emergency response plans. and special operating procedures to prevent accidents. The major approach to inherently safer process designs is divided into the following categories: intensification substitution attenuation limitation of effects simplification/error tolerance llCCPS. has lower capital and operating costs. Smaller equipment. 1996). collapse of absorber tower across pipe rack Termination: Consumption of combustible materials in process As mentioned previously. alarms. 12CCPS. such as changes to the fundamental chemistry and technology. the potential for major improvements is the greatest at the earliest stages of process development.

the following four words have recently been recommended to describe inherent safety: minimize (intensification) substitute (substitution) moderate (attenuation and limitation of effects) simplify (simplification and error tolerance). A measure of human injury. or the environment. The process by which the results of a risk analysis are used to make decisions. . either through a relative ranking of risk reduction strategies or through comparison with risk targets. These five categories are the predominant ones used since the development of this concept. 1999). distillation columns. environmental damage. Some companies add or subtract categories to their program to fine-tune their understanding and application. or a conditional probability. Risk Risk analysis Risk assessment Scenario 'CCPS. When possible. a probability of occurrence during some time interval. A description of the events that result in an accident or incident. Minimizing entails reducing the hazards by using smaller quantities of hazardous substances in the reactors. The types of inherent safety techniques that are used in the chemical industry are illustrated in Table 1-9 and are described more fully in what follows. or economic loss in terms of both the incident likelihood and the magnitude of the loss or injury. Guidelines for Consequence Analysis of Chemical Releases (New York: American Institute of Chemical Engineers. or property damage. storage vessels. The description should contain information relevant to defining the root causes. property. The development of a quantitative estimate of risk based on an engineering evaluation and mathematical techniques for combining estimates of incident consequences and frequencies. A measure of the expected probability or frequency of occurrence of an event. death. A chemical or physical condition that has the potential for causing damage to people. hazardous materials should be produced and consumed in situ. This minimizes the storage and transportation of hazardous raw materials and intermediates. not all events propagate into incidents. not the result of the event.1-7 Inherent Safety 21 Table 1-8 Term Accident Hazard Incident Consequence Likelihood Definitions for Case Histories1 Definition The occurrence of a sequence of events that produce unintended injury. not all incidents propagate into accidents. A measure of the expected effects of the results of an incident. In an attempt to make these categories more understandable. This may be expressed as a frequency. The loss of containment of material or energy. "Accident" refers to the event. and pipelines.

mercury Use chemicals with higher flash points. boiling points. that is. When possible. This can be accomplished by using alternative chemistry that allows the use of less hazardous materials or less severe processing conditions. While minimization possibilities are being investigated. that is. toxic or flammable solvents should be replaced with less hazardous solvents (for example. Smaller tanks also reduce the hazards of a release.and explosion-resistant barricades Separate systems and controls into blocks that arc easy to comprehend and understand Label pipes for easy "walking the line" Label vessels and controls to enhance understanding Minimize (intensification) Substitute (substitution) Moderate (attenuation and limitation of effects) Simplify (simplification and error tolerance) Vapor released from spills can be minimized by designing dikes so that flammable and toxic materials will not accumulate around leaking tanks. flanged Use solvents that are less toxic Use mechanical gauges vs.Chapter 1 Introduction Table 1-9 TYpe Inherent Safety Techniques Typical techniques Change from large batch reactor to a smaller continuous reactor Reduce storage inventory of raw materials Improve control to reduce inventory of hazardous intermediate chemicals Reduce process hold-up Use mechanical pump seals vs. Another alternative to substitution is moderation. substitutions should also be considered as an alternative or companion concept. using a hazardous material un- . and other less hazardous properties Use water as a heat transfer fluid instead of hot oil Use vacuum to reduce boiling point Reduce process temperatures and pressures Refrigerate storage vessels Dissolve hazardous material in safe solvent Operate at conditions where reactor runaway is not possible Place control rooms away from operations Separate pump rooms from other rooms Acoustically insulate noisy lines and equipment Barricade control rooms and tanks Keep piping systems neat and visually easy to follow Design control panels that are easy to comprehend Design plants for easy and safe maintenance Pick equipment that requires less maintenance Pick equipment with low failure rates Add fire. packing Use welded pipe vs. water-based paints and adhesives and aqueous or dry flowable formulations for agricultural chemicals). safer materials should be used in place of hazardous ones.

such as remote controls. Often. The process uses cyclohexane. (2) transfer systems can be designed to minimize the potential for leaks. Flixborough. Simpler plants are friendlier than complex plants because they provide fewer opportunities for error and because they contain less equipment that can cause problems. Containment buildings are sometimes used to moderate the impact of a spill of an especially toxic material.1-8 Four Significant Disasters 23 der less hazardous conditions. For example. Less hazardous conditions or less hazardous forms of a material include (1) diluting to a lower vapor pressure to reduce the release concentration. special precautions are included to ensure worker protection. (3) handling larger particle size solids to minimize dust. and bellows. The Flixborough accident is perhaps the most documented chemical plant disaster. a basic raw material for the production of nylon. (2) refrigerating to lower the vapor pressure. This information is used to improve procedures to prevent similar accidents in the future. continuous monitoring. Italy. Seveso. and using proper support of lines to minimize stress and subsequent failures. using welded pipes for flammable and toxic chemicals and avoiding the use of threaded pipe. and restricted access. flexible connectors. England. England. and (4) processing under less severe temperature or pressure conditions. (3) process steps and units can be separated to prevent the domino effect. and (6) the status of the process can be made visible and clear at all times. and Pasadena. Chapter 13 presents case histories in considerably more detail. Although it was not reported to any great extent in the United States. the reason for complexity in a plant is the need to add equipment and automation to control the hazards. 1-8 Four Significant Disasters The study of case histories provides valuable information to chemical engineers involved with safety. Bhopal. As a result of the accident. All these accidents had a significant impact on public perceptions and the chemical engineering profession that added new emphasis and standards in the practice of safety. The design o an inherently safe and simple piping system includes minimizing the use of f sight glasses. England The accident at Flixborough. When containment is used. . The Flixborough Works of Nypro Limited was designed to produce 70.000 tons per year of caprolactam. using spiral wound gaskets and flexible graphitetype gaskets that are less prone to catastrophic failures. (4) fail-safe valves can be added. (1) piping systems can be designed to minimize leaks or failures. Simplification reduces the opportunities for errors and misoperation. The four most cited accidents (Flixborough. India. The British government insisted on an extensive investigation. occurred on a Saturday in June 1974. (5) equipment and controls can be placed in a logical order. it had a major impact on chemical engineering in the United Kingdom. Texas) are presented here. safety achieved a much higher priority in that country.

and 36 others were injured. Several months before the accident occurred.24 Chapter 1 Introduction which has properties similar to gasoline. The liquid reaction mass was gravity-fed through the series of reactors. Damage extended to 1821 nearby houses and 167 shops and factories. Under the process conditions in use at Flixborough (155°C and 7. The decision was made to remove the reactor for repairs. This accident could have been prevented by following proper safety procedures. In these reactors cyclohexane was oxidized to cyclohexanone and then to cyclohexanol using injected air in the presence of a catalyst. . an estimated 30 tons of cyclohexane volatilized and formed a large vapor cloud.9 atm). the connections to reactor 4 and reactor 6 were made using flexible bellows-type piping. Eighteen of these fatalities occurred in the main control room when the ceiling collapsed. Fifty-three civilians were reported injured. Each reactor normally contained about 20 tons of cyclohexane. including the administrative offices. The resulting fire in the plant burned for over 10 days. Upon rupture of the bypass. The loss of the reactor would reduce the yield but would enable continued production because unreacted cyclohexane is separated and recycled at a later stage. The process where the accident occurred consisted of six reactors in series. An additional decision was made to continue operating by connecting reactor 4 directly to reactor 6 in the series.It is hypothesized that the bypass pipe section ruptured because of inadequate support and overflexing of the pipe section as a result of internal reactor pressures. The resulting explosion leveled the entire plant facility. the Bellows Temporary Pipe Section Figure 1-10 A failure of a temporary pipe section replacing reactor 5 caused the Flixborough accident. First. as shown in Figure 1-10. Twenty-eight people died. The feed pipes connecting the reactors were 28 inches in diameter. Because only 20-inch pipe stock was available at the plant. the cyclohexane volatilizes immediately when depressurized to atmospheric conditions. The cloud was ignited by an unknown source an estimated 45 seconds after the release. Loss of life would have been substantially greater had the accident occurred on a weekday when the administrative offices were filled with employees. Inspection showed a vertical crack in its stainless steel structure. reactor 5 in the series was found to be leaking.

11.5 miles away when the plant was constructed. and flammable.400 gallons of benzene. with inadequate cooling the temperature will increase and the MIC will boil. on December 3. MIC reacts exothermically with water. MIC storage tanks are typically refrigerated to prevent this problem. accident. Bhopal.1°C.1-8 Four Significant Disasters 25 bypass line was installed without a safety review or adequate supervision by experienced engineering personnel. . killing over 2000 civilians and injuring an estimated 20. No plant workers were injured or killed. particularly intermediates. 26. Finally. toxic. Unfortunately. a shantytown eventually grew around the immediate area. The unit using the MIC was not operating because of a local labor dispute. The scrubber and flare system should have been fully operational to prevent the release. A chemical reaction heated the MIC to a temperature past its boiling point. If the accident was caused by a problem with the process. The maximum exposure concentration of MIC for workers over an 8-hour period is 0. Inventories of dangerous chemicals. Death at large concentrations of vapor is due to respiratory distress. the scrubber and flare systems were not operating. As a rule. the plant site contained excessively large inventories of dangerous compounds. any modifications should be of the same quality as the construction of the remainder of the plant. and 450 gallons of gasoline. This included 330. for a variety of reasons. The plant pesticides. The plant was partially owned by Union Carbide and partially owned locally. No plant equipment was damaged. The toxic cloud spread to the adjacent town.000 more. India. The bypass was sketched on the floor of the machine shop using chalk! Second. Although the reaction rate is slow. has received considerably more attention than the Flixborough accident. 66. should also have been minimized. MIC is an extremely dangerous compound. a well-executed safety review could have identified the problem. The nearest civilian inhabitants were 1. This is due to the more than 2000 civilian casualties that resulted. An estimated 25 tons of toxic MIC vapor was released. 1984. An intermediate compound in this process is methyl isocyanate (MIC). Somehow a storage tank containing a large amount of MIC became contaminated with water or some other substance.000 gallons of cyclohexane. volatile. Because the plant was the dominant source of employment in the area.and it has a vapor pressure of 348 mm Hg at 20°C. The MIC vapors traveled through a pressure relief system and into a scrubber and flare system installed to consume the MIC in the event of a release. Individuals exposed to concentrations of MIC vapors above 21 ppm experience severe irritation of the nose and throat. the bypass modification was substandard in design. The exact cause of the contamination of the MIC is not known. The Bhopal plant is in the state of Madhya Pradesh in central India.000 gallons of naphtha.000 gallons of toluene. MIC demonstrates a number of dangerous physical properties. India The Bhopal. ensuring that the vapors will stay close to the ground once released. Its boiling point at atmospheric conditions is 39.02 ppm (parts per million). The vapor is about twice as heavy as air. These inventories contributed to the fires after the initial blast. It is reactive.

Italy Seveso is a small town of approximately 17. Adapted from Chemical and Engineering News (Feb. a bactericide. a . 1985).26 Chapter 1 Introduction Methyl isocyanate route CH3NH2+ COCI. 1 1 . The reaction scheme used at Bhopal is shown at the top of Figure 1-11 and includes the dangerous intermediate MIC. Another solution is to redesign the process to reduce the inventory of hazardous MIC. During normal operation. The plant was owned by the Icmesa Chemical Company. p.000 inhabitants. 15 miles from Milan. with trichlorophenol produced as an intermediate. Methylamine Phosgene - CH3N = C = 0 Methyl isocyanate + 2HCI 0 II OH I 0 . Italy. with an inventory of MIC of less than 20 pounds. The lower reaction suggests an alternative reaction scheme using a less hazardous intermediate. Seveso. 30.CNHCH3 I CH3N=C=0 + a-Naphthol Carbaryl Nonmethyl isocyanate route a-Na~hthol chloroformate Figure 1-11 The upper reaction is the methyl isocyanate route used at Bhopal. An alternative reaction scheme is shown at the bottom of the figure and involves a less dangerous chloroformate intermediate. The product was hexachlorophene. One such design produces and consumes the MIC in a highly localized area of the process.

resulted in 23 fatalities. Because TCDD is also insoluble in water. and the contamination was confined to the plant. by following proper procedures. Animal studies have shown TCDD to be fatal in doses as small as lo-' times the body weight. by using proper hazard evaluation procedures. an acne-like disease that can persist for several years. 1989. Over 600 people were evacuated. The release formed a large gas cloud instantaneously because the system was under high pressure and temperature. isobutane. decontamination is difficult. The release occurred in the polyethylene product takeoff system.8-tetrachlorodibenzoparadioxin) produced in the reactor as is an undesirable side-product. Pasadena. Nonlethal doses of TCDD result in chloracne. TCDD is perhaps the most potent toxin known to humans. the initiation steps would not have occurred. civilian evacuation was not started until several days later. India. A subsequent heavy rain washed the TCDD into the soil.Usually the polyethylene particles (product) settle in the settling leg and are removed through the product takeoff valve. hexane. Because of poor communications with local authorities. Occasionally. the product plugs the settling leg. On July 10. By then. The damage resulting from the explosion made it impossible to reconstruct the actual accident scenario. The cloud was ignited about 2 minutes after the release by an unidentified ignition source. Second. The Seveso and Duphar accidents could have been avoided if proper containment systems had been used to contain the reactor releases. encased in concrete and dumped into the ocean. An estimated 2 kg of TCDD was released through a relief system in a white cloud over Seveso. on October 23. the hazards could have been identified and corrected before the accidents occurred. the condition it remains in today. 314 injuries. and the plug . the trichlorophenol reactor went out of control. Of the 50 men assigned to clean up the release. and capital losses of over $715 million. The most severely contaminated area immediately adjacent to the plant was fenced. The proper application of fundamental engineering safety principles would have prevented the two accidents. evidence showed that the standard operating procedures were not appropriately followed. in 1963 the plant was finally disassembled brick by brick.3.7. However. Texas. First. This explosion occurred in a high-density polyethylene plant after the accidental release of 85. 1976. Approximately 10 square miles were contaminated.1-8 Four Significant Disasters 27 small amount of TCDD (2. and hydrogen. TCDD is so toxic and persistent that for a smaller but similar release of TCDD in Duphar. resulting in a higher than normal operating temperature and increased production of TCDD. as illustrated in Figure 1-12. 4 eventually died from the exposure. over 250 cases of chloracne were reported.000 pounds of a flammable mixture containing ethylene. Texas A massive explosion in Pasadena. and an additional 2000 people were given blood tests. Less than 200 g of TCDD was released.

. specifically. the DEMCO valve was in the open position. 1990). The OSHA investigation13found that (1)no process hazard analysis had been performed in the polyethylene plant. The accident investigation evidence showed that this safe procedure was not followed. The normal . DC: US Department of Labor. Then the product takeoff valve is removed to give access to the plugged leg. and locking the valve in the closed position._ I.procedure includes closing the DEMCO valve. (2) the single-block (DEMCO) valve on the settling leg was not designed to fail to a safe closed position when the air failed. many serious safety deficiencies were ignored or overlooked. (3) rather than relying on a single-block valve. \ L DEMCO valve (ball valve) r n--~. the product takeoff valve was removed. a double130ccupational Safety and Health Administration. is removed by maintenance personnel.28 Chapter I Introduction A Air to open Lock-out device Air to close CfX-\ d/ -..and safe . This scenario was a serious violation of well-established and well-understood procedures and created the conditions that permitted the release and subsequent explosion. removing the air lines.-& Removed before accident (standard procedure) take. Figure 1-12 Polyethylene plant settling leg and product takeoff system. The Pasadena Accident: A Report to the President (Washington. and as a result. and the lockout device was removed.

1980). 1994). eds. Wood. and (5) no permanent combustible gas detection and alarm system was located in the region of the reactors. 1. General Case Histories Trevor A. and (3) crowded process equipment. app. Bethea. 16-18.Suggested Reading 29 block-and-bleed valving arrangement or a blind flange after the single-block valve should have been used. v. Howard H. Bethea." Chemical Engineering (May 16. 2d ed. and Other Zncidents (New York: American Institute of Chemical Engineers. Loss Prevention in the Process Industries. Sambeth.1983). 23. Loss Prevention in the Process Industries. 2.1985). 27. Frank P. 1985). Lees. p. Bhopal Chemical and Engineering News (Feb.1976). 3. 1982). 1998). Loss Prevention in the Process Industries. What Went Wrong? Case Histories of Process Plant Disasters (Houston: Gulf Publishing. Frank P. 1. v. Fawcett and William S. implementation. and enforcement of effective permit systems (for example. Pasadena. Texas (New York: American Institute of Chemical Engineers. The Bhopal Disaster (New York: American Institute of Chemical Engineers. ch. 1 (London: Butterworths. (London: Butterworth-Heinemann. (New York: Wiley. 44-47 Flixborough Robert M. Lees. . p.. 1996). Lees. 2. 11. v. Safety and Accident Prevention in Chemical Operations. app. 14. Lees. Ronald J. ch. ch. 1-5. Willey. "What Really Happened at Seveso. pp. Loss Prevention in the Process Industries. J. Other factors that contributed to the severity of this disaster were also cited: (1)proximity of high-occupancy structures (control rooms) to hazardous operation. 1996). Process Safety Management with Case Histories: Flixborough. line opening). (2) inadequate separation between buildings. Kletz. Suggested Reading General Aspects of Chemical Process Safety Robert M. 2d ed. Explosion and Fire at Pasadena. Seveso Chemical and Engineering News (Aug. (4) no provision was made for the development.

. How many hours must be traveled by car for each hour of rock climbing to make the risks of fatality by car equal to the risk of fatality by rock climbing? 1-7. Liquid products from 5 of 26 salt dome caverns fed the fire with an estimated 18. flashing back to the flange. When adequate circulation could not be maintained in a deisobutanizer heater circuit. which ignited. respectively.40% in process area 2. propagation. The ensuing fire exposed a fractionation tower and horizontal receiver drums. What is the overall FAR for the plant. The units have FARs of 0. it was decided to clean the strainer. These vapors were carried to a furnace 100 ft away.5. and 40% in process area 3.0.0. How many industrialrelated deaths are expected each year? 1-6. Should the worker be satisfied or alarmed? What is the FAR (assuming normal working hours) and the deaths per person per year? What should her chances be.14 Suggest ways to prevent and contain the accidents. b. destroying the fire pump engine. assuming worker exposure to all three units simultaneously? b.3. and 1. Three process units are in a plant. Assuming that a car travels at an average speed of 50 miles per hour. Workers had depressurized the pipe and removed all but three of the flange bolts when a pressure release blew a black material from the flange. a. what is the expected deaths per person per year? 1-2. which added more fuel. and termination steps for the following accident reports. what is his overall FAR? 1-3. assuming an average chemical plant? 1-5. A contractor accidentally cut into a 10-in propane line operating at 800 psi at a natural gas liquids terminal. 200 days per year. The large vapor cloud estimated to cover an area of 44 acres was ignited about 4-5 min later by an unknown source. rupturing pipelines. followed by butane vapors. where they ignited. one because intense radiated heat damaged its ignition wires and the other because the explosion broke a sight glass fuel gauge.000 gal of LPGs for almost 6 hr before being blocked in and the fires extinguished. how many miles must be driven before a fatality is expected? 1-4. These drums exploded. spilling diesel fuel.000-30. Both engine-driven fire pumps failed. a. If this employee works a 4-hr shift. Introduction Problems 1-1. An employee works in a plant with a FAR of 4. If a worker spends 20% of his time in process area 1. Identify the initiation. Assume now that the units are far enough apart that an accident in one would not affect the workers in another.- - 30 Chapter 1 . The explosions and heat caused loss of insulation from the 8-ft x 122-ft fractionator tower. An alkylation unit was being started up after shutdown because of an electrical outage. causing it to weaken and fall across two major pipeI40ne Hundred Largest Losses. A worker is told her chances of being killed by a particular process are 1in every 500 years. Consider Example 1-4. A plant employs 1500full-time workers in a process with aFAR of 5.

The fault was traced to a 10-in valve that had been prevented from closing the last %-inch by a fine powder of carbon and iron oxide. The union. After reviewing the answers of Problems 1-10 and 1-11. fearing that the plant might explode at any moment. e. achieved basically by isolating the fuel sources. b.000 passenger miles. 1-10. including some that could lead to catastrophic results. An explosion has occurred in your plant and an employee has been killed. a. 1-8. Compute the OSHA incidence rate based on injuries and lost workdays. preferring not to do any business there. What conclusions can you derive from Figure 1-9? 1-14. An investigation shows that the accident was the fault of the dead employee. "We've operated this plant for 40 years without an accident. d. You walk into a store and after a short while you decide to leave. Based on workplace fatalities (Figure 1-4) and assuming you are responsible for a safety program of an organization. The other employees elsewhere in the plant site. When the flange was opened. who manually charged the wrong ingredient to a reactor vessel. What did you observe to make you leave? What conclusions might you reach about the attitudes of the people who manage and operate this store? b. he says. You have just begun work at a chemical plant. Do the accident statistics support this claim? In 1984 the airline industry posted 4 deaths per 10. can inherent safety help? 1-13. What is the appropriate response from the following groups? a. The airline industry claims commercial airline transport has fewer deaths per mile than any other means of transportation.000. A university has 1200 full-time employees. After all. You walk into a chemical plant and after a short while you decide to leave. allowing liquid butane to be released.Problems 31 lines. Middle management. He is a few years away from retirement after working his way up from the very bottom. f. The president of the company. what would you emphasize in a safety program? 1-12. Also. You bring up these problems to your immediate supervisor. 1-16. but he decides to do nothing for fear that the plant manager will be upset. Extinguishment. What is the worst thing that could happen to you as a chemical engineer in industry? 1-15. What additional information is required to compute a FAR? a fatality rate? 1-9." What would you do in this situation? 1-17. The other employees who work in the process area affected. a number of unsafe practices are performed at the plant. what would you emphasize? 1-11. Upper management. took 2% hours. this powder blew out. Based on the causes of the largest losses (Figure 1-7). In a particular year this university had 38 reportable lost-time injuries with a resulting 274 lost workdays. breaking piping -which added more fuel to the fire. c. After several weeks on the job you determine that the plant manager runs the plant with an iron fist. What did you observe tomake youleave? What .

Clearly the pump unit should have been isolated. What will happen if the pump continues to operate? 1-24. One of the categories of inherent safety is simplificationlerror tolerance. What is this procedure? 1-21. The liquid level in a tank 10 m in height is determined by measuring the pressure at the bottom of the tank. which flashes into vapor and ruptures the tank. With the installation of the alarm.8. What is the ratio in volume between water vapor at 300 K and liquid water at 300 K at saturated conditions? b. Pumps can be shut-in by closing the valves on the inlet and outlet sides of the pump. Can you explain? 1-19. A particular pump contains 4 kg of water. If the pump is rated at 1HP. Unfortunately. When the original allocation of numbers ran out the new equipment was numbered JAlOOl upward. What happened? 1-20. an alarm was installed on the level gauge to alert the operator to a high-level condition.9. the tank contains residual water. If the usual liquid is replaced with a new liquid with a specific gravity of 0. the operator closes the valve to stop the filling. The cover flies off suddenly. what is the reading on the level gauge? Is it possible that the tank will overflow without the level gauge indicating the situation? 1-22. The fitter removes all eight bolts only to find the cover plate stuck on the housing. however. The operator first opens a valve on a supply line and carefully watches the level on a level indicator until the tank is filled (a long time later). A pipe fitter is instructed to repair the pump. If the tank is 10 m in diameter and 5 m high.32 Chapter 1 Introduction conclusions might you reach about the attitudes of the people who manage and operate this chemical plant? Comment on the similarities of parts a and b. Unfortunately. A cover plate on a pump housing is held in place by eight bolts. and toxic liquid sprays throughout the work area. and cleaned before repair. 1-18. will the tank be overfilled or underfilled? If the actual liquid level is 8 m. how many kilograms of water at 300 K are required to produce . What instrumentation could you add to the tank described in Problem 1-21 to eliminate problems? 1-23. he prepared pump Jl001 instead. causing an upset in the plant. the tank now overfills twice per year. Once a year the operator is distracted and the tank is overfilled. An operator was verbally told to prepare pump JAlOOl for repairs. A screwdriver is used to pry off the cover. Water will flash into vapor almost explosively if heated under certain conditions. Hot oil is accidentally pumped into a storage vessel. There is. The level gauge was calibrated to work with a liquid having a specific gravity of 0. what is the maximum temperature increase expected in the water in "Clhr? Assume a constant water heat capacity of 1kcal/kg/"C. a. drained. Once the filling is complete. On one unit the equipment was numbered JlOOl upward. A large storage tank is filled manually by an operator. To prevent this. This can lead to pump damage and/or a rapid increase in the temperature of the liquid shut inside the pump. Careful numbering of process equipment is important to avoid confusion. a better procedure for removing the cover plate.

Develop a procedure to handle this hazard. what is the LTIR? 1-26. An experienced maintenance person suggests that she can remove the slip plate without depressurizing the line. what is the resulting pressure (in inches of water gauge) if the person's weight is distributed across the entire roof? b. Since you are in charge. A particular tank is 30 ft in diameter. There is considerable pressure from the sales department to deliver product. At about 4 A. Consider liquid water at 25°C and 1 atm. If the plant had 2 lost-time incidents last year. followed by considerable damage to the column. Typically they are constructed to withstand no more than 8 in of water gauge pressure and 2. or lost-time injury rate. one valve was opened and the hot oil came in contact with the cold water. A column was used to strip low-volatile materials from a high-temperature heat transfer fluid. Problems 33 enough water vapor to pressurize the tank to 8 in of water gauge pressure. You are unable to identify the contents of the drum. An alternative travel plan is to fly on a commercial airline for 4%hr. 1-29. based on the FAR? 1-27. During normal operation. There are many ways to solve this problem. . a problem develops. A slip plate or blind has accidentally been left in one of the process lines. the burst pressure of the tank? 1-25.M. This is identical to the OSHA incidence rate based on incidents in which the employee is unable to continue their normal duties.. A plant site has 1200 full-time employees working 40 hrlweek and 50 weekslyr. If a 200-lb person stands in the middle of the tank roof. How did this happen? Note: A person can easily blow to a pressure of greater than 20 in of water gauge. A large storage tank was sucked in when the vent to the outside became plugged and the operator turned on the pump to empty the tank. She said that she routinely performed this operation years ago. what is the total weight (in pounds) of the water? c. How many times does the volume increase if the water is vaporized at 100°C and 1 atm? 1-28. Large storage tanks are designed to withstand low pressures and vacuums.5 in of water gauge vacuum. If the roof was flooded with 8 in of water (equivalent to the maximum pressure). what would you do? . The plant has been down for extensive maintenance and repair. Please describe just one approach. 1-30. The result was almost sudden vaporization of the water. a. During a maintenance procedure. A 50-gal drum with bulged ends is found in the storage yard of your plant. A car leaves New York City and travels the 2800-mi distance to Los Angeles at an average speed of 50 mph. You are in charge of bringing the plant up and on-line. water was trapped between two valves. What are the FARs for the two methods of transportation? Which travel method is safest. Another way of measuring accident performance is by the LTIR.

.

A toxicant can be a chemical or physical agent. chemical engineers must be knowledgeable about the way toxicants enter biological organisms. stated the problem: "All substances are poisons. including dusts. toxicology was defined as the science of poisons. there is none which is not a poison. The toxic hazard of B . Unfortunately. can become fatal if delivered to the biological organism in large enough doses. A good example of a physical agent is asbestos fiber. The toxicity of a chemical or physical agent is a property of the agent describing its effect on biological organisms. the way toxicants are eliminated from biological organisms. Paracelsus. The first three areas are related to toxicology. such as water. and methods to prevent or reduce the entry of toxicants into biological organisms. a topic considered in chapter 3.Toxicology ecause of the quantity and variety of chemicals used by the chemical process industries. the effects of toxicants on biological organisms. a known cause of lung damage and cancer. fibers. only harmless ways of using substances. toxicology is more adequately defined as the qualitative and quantitative study of the adverse effects of toxicants on biological organisms. the word poison could not be defined adequately. The right dose differentiates a poison and a remedy. A fundamental principle of toxicology is There are no harmless substances. The last area is essentially industrial hygiene. Many years ago. Toxic hazard is the likelihood of damage to biological organisms based on exposure resulting from transport and other physical factors of usage. and radiation. Today. noise." Harmless substances. an early investigator of toxicology during the 1500s.

for instance. drinking.Chapter 2 Toxicology a substance can be reduced by the application of appropriate industrial hygiene techniques. and dermal absorption generally result in the toxicant entering the bloodstream unaltered. For corrosive chemicals the damage to the organism can occur without absorption or transport through the bloodstream. Toxicants enter biological organisms by the following routes: ingestion: through the mouth into the stomach. injection: through cuts into the skin. A common misconception is that damage occurs in the organ where the toxicant is most concentrated. hoods. summarized in Table 2-1. The toxicity. and other personal protection equipment Proper protective clothing Proper protective clothing Dermal absorption . Figure 2-1 shows the expected blood-level concentration as a function of time and route of entry. and smoking Ventilation. The blood-level concentration is a function of a wide range of parameters. The damage is exerted at the target organ. Toxicants that enter by injection and dermal absorption are difficult to measure and quantify. the usual exposure is by vapor. 2-1 How Toxicants Enter Biological Organisms For higher-order organisms the path of the chemical agent through the body is well defined. so large variations in this behavior are expected. is stored in humans mostly in the bone structure. Lead. but the damage occurs in many organs. the inhalation and dermal routes are the most significant to industrial facilities. Of the four routes of entry. Inhalation is the easiest to quantify by the direct measurement of airborne concentrations. dermal absorption: through skin membrane. Toxicants entering through ingestion are frequently modified or excreted in bile. cannot be changed. After the toxicant enters the organism. it moves into the bloodstream and is eventually eliminated or it is transported to the target organ. respirators. however. inhalation. Injection. These control techniques are discussed in more detail in chapter 3 on industrial hygiene. All these entry routes are controlled by the application of proper industrial hygiene techniques. Injection usually results in the highest blood-level con- Table 2-1 Entry route Ingestion Inhalation Injection Entry Routes for Toxicants and Methods for Control Entry organ Mouth or stomach Mouth or nose Cuts in skin Skin Method for control Enforcement of rules on eating. inhalation: through the mouth or nose into the lungs. Some toxicants are absorbed rapidly through the skin. but small solid and liquid particles can also contribute.

The type of chemical. followed by inhalation. centration. and absorption. followed by inhalation. and excretion. ingestion. The peak concentration generally occurs earliest with injection. Injection includes both entry by absorption through cuts and mechanical injection with hypodermic needles. rate of movement through the GI tract. Airborne particles (either solid or liquid) can also lodge in the mucus of the upper respiratory tract and be swallowed. This layer consists of dead. distribution. its molecular weight. Absorption also occurs through .-and absorption. Gastrointestinal Tract The GI tract plays the most significant role in toxicants entering the body through ingestion. The gastrointestinal (GI) tract. and many other factors affect the rate of absorption.2-1 How Toxicants Enter Biological Organisms 37 Figure 2-1 Toxic blood level concentration a s a function o route o exposure. Mechanical injection can occur as a result of improper hypodermic needle storage in a laboratory drawer. molecule size and shape. Food or drink is the usual mechanism of exposure. The skin is composed of an outer layer called the stratum corneum. ingestion. and the respiratory system play significant roles in the various routes of entry. Wide variations f f f are expected as a result o rate and extent of absorption. dried cells that are resistant to permeation by toxicants. acidity. biotransformation. Skin The skin plays important roles in both the dermal absorption and injection routes of entry. The rate and selectivity of absorption by the GI tract are highly dependent on many conditions. susceptibility to attack by intestinal flora. the skin.

for example. In 1minute a normal person at rest uses an estimated 250 ml of oxygen and expels approximately 200 ml of carbon dioxide. The presence of water increases the skin hydration and results in increased permeability and absorption. Phenol. this skin demonstrates increased porosity. and hydroxides (ammonium hydroxide. Only a fraction of the total air within the lung is exchanged with each breath. These materials either react or dissolve in the mucus to form acids and bases. The upper respiratory tract is affected mostly by toxicants that are water soluble. Lower respiratory toxicants include monomers (such as acrylonitrile). resulting in higher toxicant absorption. halides (fluorine. however. an estimated 100 ml of blood is in the capillaries at any moment. oxides (nitrogen oxides. These demands increase significantly with physical exertion. reacts with the water on the alveoli wall to produce HCl and carbon monoxide. for example. including the bronchi and the alveoli. bromine). The mucus lining the upper respiratory tract assists in filtering. sinuses. The absorption properties of the skin vary as a function of location and the degree of hydration. An estimated 300 million alveoli are found in a normal lung. Phosgene gas. Most chemicals are not absorbed readily by the skin. The respiratory system is divided into two areas: the upper and the lower respiratory system. chlorine. Small capillaries found in the walls of the alveoli transport the blood.38 Chapter 2 Toxicology the hair follicles and sweat glands. Approximately 8 L of air are breathed per minute. sulfur oxides. The main function of the respiratory system is to exchange oxygen and carbon dioxide between the blood and the inhaled air. hydrogen bromide). The alveoli are small blind air sacs where the gas exchange with the blood occurs. larynx (the voice box). Upper respiratory toxicants include hydrogen halides (hydrogen chloride. do show remarkable skin permeability. Toxicants in the lower respiratory tract affect the alveoli by physically blocking the transfer of gases (as with insoluble dusts) or reacting with the wall of the alveoli to produce corrosive or toxic substances. The bronchial tubes carry fresh air from the trachea through a series of branching tubes to the alveoli. but this is normally negligible. and the trachea or windpipe. pharynx (section between the mouth and esophagus). The upper and lower respiratory tracts respond differently to the presence of toxicants. and potassium hydroxides). mouth. The lower respiratory system is composed of the lungs and its smaller structures. These alveoli contribute a total surface area of approximately 70 m2. requires only a small area of skin for the body to absorb an adequate amount to result in death. The upper respiratory system is composed of the nose. A few chemicals. Respiratory System The respiratory system plays a significant role in toxicants entering the body through inhalation. sodium dusts. However. Fresh air brought in through the nose is completely saturated with water and regulated to the proper temperature by the time it reaches the larynx. heating. and humidifying the air. and other miscellaneous . The upper respiratory tract is responsible for filtering. The skin on the palm of the hand is thicker than skin found elsewhere. sodium oxide).

Other routes of excretion are the skin (by means of sweat). inhalation. or other organs. whereas substances that enter through inhalation. damage can occur to the kidneys. or lungs. The liver is the dominant organ in the detoxification process. the stored chemical agents will be released into the bloodstream. in which the chemical agents are transformed by reaction into either harmless or less harmful substances. particularly those that are volatile. liver. liver. are excreted partially by this route. Compounds with lower molecular weights enter the bloodstream and are excreted by the kidneys. Dusts and other insoluble materials present a particular difficulty to the lungs. or dermal absorption generally arrive in the bloodstream unchanged. Toxicants that are ingested into the digestive tract are frequently excreted by the liver. chemical compounds with molecular weights greater than about 300 are excreted by the liver into bile. For massive exposures to chemical agents. and soot. hair. skin. lungs. and lungs. injection. and kidney. Particles greater than 5 km in diameter are usually filtered by the upper respiratory system. liver. Particles less than 1pm in diameter can reach the alveoli. Particles that enter the alveoli are removed slowly. kidneys. Storage can create a future problem if the organism's food supply is reduced and the fatty deposits are metabolized. the farther they penetrate into the respiratory system. Chloroform and alcohol. storage: in the fatty tissue. and dermal absorption. The kidneys are the dominant means of excretion in the human body. silica. methyl cyanide. detoxification: by changing the chemical into something less harmful by biotransformation. acrolein. The toxicants are extracted by the kidneys from the bloodstream and are excreted in the urine. significantly reducing the organism's ability to excrete the substance. The detoxification occurs by biotransformation. . 2-2 How Toxicants Are Eliminated from Biological Organisms Toxicants are eliminated or rendered inactive by the following routes: excretion: through the kidneys. These routes are usually minor compared to the excretion processes of the kidneys. liver. phosgene. The lungs are also a means for elimination of substances. injection. intestinal tract wall. They eliminate substances that enter the body by ingestion. for example. The final mechanism for elimination is storage. and nails. asbestos dust. blood. In general. This process involves depositing the chemical agent mostly in the fatty areas of the organism but also in the bones.2-2 How Toxicants Are Eliminated from Biological Organisms 39 substances such as hydrogen sulfide. The digestive tract tends to selectively detoxify certain agents. resulting in possible damage. For dusts the following simple rule usually applies: The smaller the dust particles. Particles with diameters betwcen 2 and 5 pm generally reach the bronchial system. Biotransformation reactions can also occur in the blood. and other organs.

such as asthma or bronchitis. The spirometer measures (1)the total volume exhaled. The problem is to determine whether exposures have occurred before substantial symptoms are present. and platelet count are performed easily and inexpensively. . called the forced vital capacity (FVC). and (4) the ratio of the observed FEV. pigmentation. The results from these tests must be compared to a medical baseline study. and hair and nail appearance are indicative of possible toxic exposures. Nervous system disorders are diagnosed by examining the patient's mental status. An increase in the RV is indicative of deterioration of the alveoli. This is accomplished through a variety of medical tests. measured in liters per second. An electroencephalogram (EEG) tests higher brain and nervous system functions. The patient exhales as hard and as fast as possible into the device. hemoglobin content. (3) forced expiratory flow in the middle range of the vital capacity (FEV 25-75%). marked changes are seen only after substantial exposure and damage. with units in liters. However. blood counts are frequently insensitive to toxic exposure. Many chemical companies perform baseline studies on new employees before employment. motor system reflexes.wic affects liver Nephrotoxic affects kidneys Neurotoxic affects nervous system Pulmonotoxic affects lungs 2-3 Effects of Toxicants on Biological Organisms Table 2-2 lists some of the effects or responses from toxic exposure./FVC%). cranial nerve function.). Blood counts are also used to determine toxic exposures. with units in liters per second. possibly because of emphysema.40 Chapter 2 Toxicology Table 2-2 Various Responses to Toxicants Effects that are irreversible Carcinogen causes cancer Mutagen causes chromosome damage Reproductive hazard causes damage to reproductive system Teratogen causes birth defects Effects that may or may not be reversible Dermatotoxic affects skin Hernotoxic affects blood Hepatoto. Reductions in FVC are due to reduction in the lung or chest volume. vascularity. Respiratory problems are diagnosed using a spirometer. (2) the forced expired volume measured at 1second (FEV. The air remaining in the lung after exhalation is called the residual volume (RV). and sensory systems. Changes in skin texture. performed before any exposure. to FVC X 100 (FEV. The RV measurement requires a specialized tracer test with helium. Reductions in expiration flow rate are indicative of bronchial disease. Measurements of the red and white blood cells. possibly as a result of fibrosis (an increase in the interstitial fibrous tissue in the lung).

For most toxicological studies animals are used. The period of the test depends on whether long. The toxicological study can be complicated by latency. the period of the test. Once the effects of a suspect agent have been quantified. The toxicant must be identified with respect to its chemical composition and its physical state. the dose range. The selection depends on the effects considered and other factors such as the cost and availability of the test organism. kidneys. For studies of genetic effects.or short-term effects are of interest. For gaseous airborne substances the dose is measured in either parts per million (ppm) or milligrams of agent per cubic meter of air (mg/m3).For airborne particulates the dose is measured in milligrams of agent per cubic meter of air (mg/m3)or millions of particles per cubic foot (mppcf). higher organisms are a necessity. For early kidney damage proteins or sugars are found in the urine. Chronic toxicity is the effect of multiple exposures occurring over a long period of time. Before undertaking a toxicological study. For example. 2-4 Toxicological Studies A major objective of a toxicological study is to quantify the effects of the suspect toxicant on a target organism. benzene can exist in either liquid or vapor form. The dose units depend on the method of delivery. single-cell organisms might be satisfactory. For substances delivered directly into the organism (by ingestion or injection). an exposure that results in a delayed response. Chronic toxicity studies are difficult to perform because of the time involved. Liver function is determined through a variety of chemical tests on the blood and urine. usually with the hope that the results can be extrapolated to humans. the following items must be identified: the toxicant. For studies determining the effects on specific organs such as the lungs. . or liver. appropriate procedures are established to ensure that the agent is handled properly. This enables researchers to apply the results obtained from small animals such as mice (fractions of a kilogram in body weight) to humans (about 70 kg for males and 60 kg for females). Each physical state preferentially enters the body by a different route and requires a different toxicological study. the target or test organism. The test organism can range from a simple single cell up through the higher animals.2-4 Toxicological Studies 41 Kidney function is determined through a variety of tests that measure the chemical content and quantity of urine. the effect or response to be monitored. Acute toxicity is the effect of a single exposure or a series of exposures close together in a short period of time. most toxicological studies are based on acute exposures. the dose is measured in milligrams of agent per kilogram of body weight.

and p is the mean. some individuals will barely notice any irritation (weak or low response). sex. Curves of the form shown in Figure 2-2 are frequently represented by a normal or Gaussian distribution. Given the same dose of vapors. . general health. whereas other individuals will be severely irritated (high response). given by the equation where f(x) is the probability (or fraction) of individuals experiencing a specific response. The fraction or percentage of individuals experiencing a specific response is plotted. consider the effects of an irritant vapor on human eyes. A plot of the type shown in Figure 2-2 is prepared with the data. Consider a toxicological test run on a large number of individuals. These differences are a result of age. diet.42 Chapter 2 Toxicology 2-5 Dose versus Response Biological organisms respond differently to the same dose of a toxicant. a is the standard deviation. Each individual is exposed to the same dose and the response is recorded. x is the response. weight. and other factors. L I Low response Average response High response Figure 2-2 A Gaussian or normal distribution representing the biological response to exposure to a toxicant. For example.

The quantity u 2is called the variance. the distribution curve becomes more pronounced around the mean value. A response interval of 2 standard deviations represents 95. The area under the curve of Figure 2-2 represents the percentage of individuals affected for a specified response interval. They are computed from the original data fixi) using the equations where n is the number of data points. Figure 2-3 Effect of the standard deviation on a normal distribution with a mean of 0. and the standard deviation determines the shape. In particular. Figure 2-3 shows the effect of the standard deviation on the shape. respectively. .2-5 Dose versus Response 43 The standard deviation and mean characterize the shape and the location of the normal distribution curve. The mean determines the location of the curve with respect to the x axis.5% of the total individuals (Figure 2-4b). as shown in Figure 2-4a. the response interval within 1 standard deviation of the mean represents 68% of the individuals.The distribution becomes more pronounced around the mean as the standard deviation decreases. The area under the entire curve represents 100% of the individuals. As the standard deviation decreases.

Example 2-1 Seventy-five people are tested for skin irritation because of a specific dose of a substance. The number of individuals exhibiting a specific response is given in the following table: . The responses are recorded on a scale from 0 to 10.44 Chapter 2 Toxicology p-20 ~1 p+2u (b) Figure 2-4 Percentage of individuals affected based on a response between one and two standard deviations of the mean. with 0 indicating no response and 10 indicating a high response.

2-5 Dose versus Response 45 Response Number of individuals affected a. c. The histogram is shown in Figure 2-5. Response Figure 2-5 Percentage of individuals affected based on response. The number of individuals affected is plotted versus the response. Plot a histogram of the number of individuals affected versus the response. b. An alternative method is to plot the percentage of individuals versus the response. . Solution a. Plot the normal distribution on the histogram of the original data. Determine the mean and the standard deviation.

4. Table 2-3 Theoretical Frequency and Number of People Affected for Each Response for Example 2-1 .51)~(13) c.46 Chapter 2 Toxicology b. in this case 75.51)'(5) + (2 . The corresponding values are shown in Table 2-3 and Figure 2-5. we find The distribution is converted to a function representing the number of individuals affected by multiplying by the total number of individuals. The normal distribution is computed using Equation 2-1.51)~(10)+ (3 .4. The mean is computed using Equation 2-2: The standard deviation is computed using Equation 2-3: u2 = [(I - 4. Substituting the mean and standard deviations.

Error bars are drawn at +a around the mean. and normal curves similar to Figure 2-3 are drawn. This form presents a much straighter function than the one shown in Figure 2-6. For comparison purposes the dose that results in 50% Logarithm of the dose Figure 2-7 Response versus log dose curve. The toxicological experiment is repeated for a number of different doses. . the response is plotted versus the logarithm of the dose. A complete dose-response curve is produced by plotting the cumulative mean response at each dose. The bars around the data points represent the standard deviation in response to a specific dose. A typical result is shown in Figure 2-6.2-5 Dose versus Response 47 Dose Figure 2-6 Dose-response curve. The standard deviation and mean response are determined from the data for each dose. If the response of interest is death or lethality. This form provides a much straighter line in the middle of the response curve than the simple response versus dose form of Figure 2-6. as shown in Figure 2-7. For convenience. the response versus log dose curve of Figure 2-7 is called a lethal dose curve.

pressure. Chronic toxicity data are usually considerably different. and sound.48 Chapter 2 Toxicology a . log f dose curves. ED. lethal dose. or antagonistically (both counteract each other). dose (lethal dose for 50% of the subjects). the toxicants might interact additively (the combined effect is the sum of the individual effects). response-dose curves are developed using acute toxicity data. TD. including exposure to heat. the response-log dose curve is called the effective dose (ED) curve. LC (lethal concentration) is used. 1 0 Q (I) 50 F X 0 .- 10 EDlo ED50 TD50 LD50 Logarithm of the dose Figure 2-8 The various types o response vs... synergistically (the combined effect is more than the individual effects). For gases. Finally. and so forth are also used. ED. impact. . If several chemicals are involved. the data are complicated by differences in group age. or T D curve. the response-log dose curve is called the toxic dose. Values for ED5.. The relationship between the various types of response-log dose curves is shown in Figure 2-8. an analytical equation is preferred. if the response to the agent is toxic (an undesirable response that is not lethal but is irreversible. such as liver or lung damage). Furthermore. potentiately (presence of one increases the effect of the other). For computational purposes the response versus dose curve is not convenient... LC (lethal concentration) data are used. sex. 2-6 Models for Dose and Response Curves Response versus dose curves can be drawn for a wide variety of exposures. are sometimes also reported. Other values such as LD. lethality of the subjects is frequently reported. LD. Most often. effective dose. and method of delivery. toxic dose. This is called the LD. radiation. If the response to the chemical or agent is minor and reversible (such as minor eye irritation). For gases. or LD.

379. 1971). . This relationship is plotted in Figure 2-9 and tabulated in Table 2-4. p.' For single exposures the probit (probit = probability unit) method is particularly suited. p. J. (Cambridge: Cambridge University Press. Probit Analysis (Cambridge: Cambridge University Press. 1985). Industrial Toxicology: Safety and Health Applications in the Workplace (New York: Van Nostrand Reinhold. Table 2-5 lists a variety of probit equations for a number of different types of exposures. 3d ed. Probit Analysis. (Source: D. 23. eds. The probit relationship of Equation 2-4 transforms the sigmoid shape of the normal response versus dose curve into a straight line when plotted using a linear probit scale.. p. Williams and James L. Many methods exist for representing the response-dose curve.2-6 Models for Dose and Response Curves Percentages Figure 2-9 The relationship between percentages and probits. Reprinted by permission. The probit variable Y is related to the probability P by2 Equation 2-4 provides a relationship between the probability P and the probit variable Y. as shown in Figure 2-10. 1971). 23. providing a straight-line equivalent to the response-dose curve. Finney. 2D. Standard curve-fittingtechniques are used to determine the best-fitting straight line. J. ! The causative factor represents the dose t The probit variable Y is computed from lPhillip L. Finney. Burson.

25. Prohlt Analysi~.(Cambridge: Cambridge University Press. 1971). Finney. log dose curve into a straight line when plotted on a linear probit scale. 24. Source: D.. Finney. p. J. Log Dose Figure 2-10 The probit transformation converts the sigmoidal response vs. p.50 Chapter 2 Toxicology Table 2-4 Transformation from Percentages to Probits1 . J.- - - - 'D. 1971). (Cambridge: Cambridge University Press. Probit Analysis.. Reprinted by permission. . Reprinted by permission. 3d ed.

0~ C C1.1 -27.1 -35. p. Loss Prevention in the Process Industries (London: Butterworths. p.OT C C2.19 -16.29 -6.79 -19.'T -46.51 1.9 k 2 ire' Burn deaths from flash fire Burn deaths from pool burning Explosion' Deaths from lung hemorrhage Eardrum ruptures Deaths from impact Injuries from impact Injuries from flying fragments Structural damage Glass breakage Toxic release2 Ammonia deaths Carbon monoxide deaths Chlorine deaths Ethylene oxide deaths" Hydrogen chloride deaths Nitrogen dioxide deaths Phosgene deaths Propylene oxide deaths Sulfur dioxide deaths Toluene teI:'3/104 tZ4'"1o4 PO 2.92 1. 254.'T Z C1. 208.'T Z C1.8 -18.OT Z C2. Lees.'T : 2 C2.56 2.9 -37. "Quantitative Evaluation of Inhalation Toxicity Hazards.79 1.82 4.9 . 1999). = effective radiation intensity (W/m2) t = time duration of pool burning (s) I = radiation intensity from pool burning (W/mz) p" = peak overpressure (N/m2) J = impulse (N s/m2) C = concentration (ppm) T = time interval (min) 'Selected from Frank P.2-6 Models for Dose and Response Curves 51 Table 2-5 Probit Correlations for a Variety of Exposures (The causative variable is representative of the magnitude of the exposure.41 ! = effective time duration (s) . ZCCPS.85 3."T I C'.85 -13.1 -39.Guidelines for Consequence Analysis of Chemical Releases (New York: American Institute of Chemical Engineers. I.69 0.'T Z c2. .14.1 -23.42 -15.92 2.O 0.45 4.98 -8.56 Po J J J Po Po C C2.0 1.79 4.1995).67 -6.4 3. July 31. 1986).26 2. Purgh." in Proceedings of the29th Loss Prevention Symposium (American Institute of Chemical Engineers.27 -7.7 0. 3Richard W.) Probit parameters Type of injury or damage Causative variable 4 -14.'T Z C1.O 2.

A. Eisenberg. Solution The percentage from Table 2-4 is 27%. Vulnerability Model: A Simulutton System for Assessing Damage Resulting from Marine Spills.500 19.300 Confirm the probit correlation for this type of exposure. and other software programs.52 Chapter 2 Toxicology For spreadsheet computations a more useful expression for performing the conversion from probits to percentage is given by where erf is the error function.3 Eisenberg? reported the following data on the effect of explosion peak overpressures on eardrum rupture in humans: Percentage affected 1 10 50 90 Peak overpressure (N/m2) 16.39. Example 2-2 Determine the percentage of people who will die as a result of burns from pool burning if the probit variable Y is 4.500 84. The same percentage can be computed using Equation 2-6.300 43. The results are: Percentage Probit ?N. NTIS Report AD-A015-245 (Springfield. . VA: National Technical Information Service. Compare results from Table 2-4 and Equation 2-6. Solution The percentage is converted to a probit variable using Table 2-4. as follows: where the error function is a mathematical function found in spreadsheets. 1975). as shown in Table 2-5. Mathcad. Example 2.

N/mZ) Figure 2-11 Percentage affected versus the natural logarithm of the peak overpressure for Example 2-3. Figure 2-11 is a plot of the percentage affected versus the natural logarithm of the peak overpressure. ~ / r n ~ ) Figure 2-12 Probit versus the natural logarithm of the peak overpressure for Example 2-3. In (Overpressure. The sigmoid curve of Figure 2-11 is drawn after converting the probit correlation back to percentages. This demonstrates the classical sigmoid shape of the response versus log dose curve. 7 + C I a e. . 302010- 10 11 9 12 In (Overpressure. The straight line verifies the values reported in Table 2-5. 706050. 40- - +- U a. Figure 2-12 is a plot of the probit variable (with a linear probit scale) versus the natural logarithm of the peak overpressure.2-6 Models for Dose and Response Curves 53 90 - 80 - u I J a.

1984). If the response-dose curve for chemical A is to the right of the response-dose curve for chemical B.000 mg/kg. called threshold limit values (TLVs). Toxicants are compared for relative toxicity based on the LD. and TLV-C) with precise definition~ provided in Table 2-7.0 mg 1. 2-8 Threshold Limit Values The lowest value on the response versus dose curve is called the threshold dose. the situation shown in Figure 2-13 might occur.. p. then chemical A is more toxic. The complete data show that chemical B is more toxic at lower doses. The TLV was formerly called the maximum allowable concentration (MAC). OSHA has defined its own threshold dose. but for small doses the response is not detectable. Care must be taken when comparing two response-dose curves when partial data are available. The exposure occurs only during normal working hours.5-5 g 5-15 g >15 g Probable lethal dose for a 70-kg person A taste A teaspoonful An ounce A pint A quart More than a quart 'N. More TLV-TWA data are available than TWA-STEL or TLV-C data. it might appear that chemical A is always more toxic than chemical B. TLV-STEL. The TLV refers to airborne concentrations that correspond to conditions under which no adverse effects are normally expected during a worker's lifetime. ED.0 -50 mg 50 -500 mg 0. per kilogram of body weight <1. If the slopes of the curves differ substantially. If only a single data point is available in the upper part of the curves. Below this dose the body is able to detoxify and eliminate the agent without any detectable effects. In reaIity the response is only identically zero when the dose is zero. or TD curves. The American Conference of Governmental Industrial Hygienists (ACGIH) has established threshold doses. 1.0 mg/kg to 15. This table covers a range of doses from 1. . for a large number of chemical agents.Chapter 2 Toxicology Table 2-6 Hodge-Sterner Table for Degree of Toxicity1 Degree of toxicity Dangerously toxic Seriously toxic Highly toxic Moderately toxic Slightly toxic Extremely low toxicity Experimental LD. Irving Sax. There are three different types of TLVs (TLV-TWA. called a permissible exposure level (PEL). Dangerous Properties of Industrial Materials (New York: Van Nostrand Reinhold. eight hours per day and five days per week. 2-7 Relative Toxicity Table 2-6 shows the Hodge-Sterner table for the degree of toxicity.

Excursions above the limit are allowed if compensated by excursions below the limit. The maximum concentration to which workers can be exposed for a period of up to 15 minutes continuously without suffering (1) intolerable irritation. (3) narcosis of sufficient degree to increase accident proneness. even instantaneously. provided that no more than 4 excursions per day are permitted. . with at least 60 minutes between exposure periods. uninterrupted exposure. without adverse effects. Toxicant A is more toxic at high doses. and provided that the daily TLV-TWA is not exceeded.2-8 Threshold Limit Values 55 Logarithm of the Dose Figure 2-13 Two toxicants with differing relative toxicities at different doses. to which nearly all workers can be exposed. whereas toxicant B is more toxic at low doses. or (3) assessment of toxic hazard from continuous. or materially reduce worker efficiency. day after day. TLV-STEL TLV-C 'TLVs should not be used for (1) a relative index of toxicity. (2) chronic or irreversible tissue change. Ceiling limit. The concentration that should not be exceeded. ( 2 ) air pollution work. Table 2-7 TLV type TLV-TWA Definitions for Threshold Limit Values (TLVs)' Definition Time-weighted average for a normal 8-hour workday or 40-hour work week. Short-term exposure limit. impair self-rescue.

and M is the molecular weight in g/g-mol. mg/m3 is converted to ppm using the equation c. the PEL values are not as numerous and are not updated as frequently. mg/m3 (milligrams of vapor per cubic meter of air). or. Another quantity frequently reported is the amount immediately dangerous to life and health (IDLH). For vapors.4 T M 273 (-) (i) (mgim3) where T is the temperature in degrees Kelvin.56 Chapter 2 Toxicology PEL values follow the TLV-TWA of the ACGIH closely.. Exposures to this quantity and above should be avoided under any circumstances. = Concentration in ppm = - 22. mg/m3or mppcf (millions of particles per cubic foot of air).. TLV and PEL values for a variety of toxicants are provided in Table 2-8. for dusts. 25°C . TLVs are often somewhat more conservative.. 25°C PPm mg/m3. These toxicants have zero thresholds. For some toxicants (particularly carcinogens) exposures at any level are not permitted. Table 2-8 TLVs and PELSfor a Variety of Chemical Substances TLV-TWA2 OSHA PEL3 PPm Acetaldehyde Acetic acid Acetone Acrolein Acrylic acid (skin) Acrylonitrile4(skin) Ammonia Aniline (skin) Arsine ~enzene~ Biphenyl Bromine Butane Caprolactum (vapor) Carbon dioxide Carbon monoxide Carbon tetrachloride4 (skin) Chlorine mg lm3. TLVs are reported using ppm (parts per million by volume). P i s the absolute pressure in atm. However.

25°C 240 1050 200 200 1015 0.5 11 2.9 1 606 PPm 50 300 50 50 300 0.6 1.6 1.2 Dichloroethylene Diethylamine Diethyl ketone Dimethylamine Dioxane (skin) Ethyl acetate Ethylamine Ethyl benzene Ethyl bromide Ethyl chloride Ethylene dichloride Ethylene oxide4 Ethyl ether Ethyl mercaptan Fluorine Formaldehyde4 Formic acid Furfural (skin) Gasoline Heptane Hexachloroethane Hexane Hydrogen chloride Hydrogen cyanide (skin) Hydrogen fluoride Hydrogen peroxide Hydrogen sulfide Iodine Isobutyl alcohol Isopropyl alcohol Isopropyl ether Ketene Maleic anhydride Methyl acetate PPm 10 300 50 25 300 600 0.37 9.2 435 22 2600 40 2 1200 1.5 0.9 1 610 (continued) .5 0.25 200 mg/m3.1 50 400 250 0.1 100 400 500 0. 25°C 50 1030 200 100 1010 1720 0.l Dichloroethane 1.4 28 1 300 980 2100 0.75 5 5 500 1 500 5 10 3 1 20 0.7 176 7.3 5 2 300 400 1 50 TLV-C: 5 TLV-C: 10 TLV-C: 3 1 10 TLV-C: 0.4 14 1 150 983 1040 0.5 1 0.1 400 790 75 18 360 1400 18 435 890 2600 4 2 1200 25 0.4 8 900 1600 9.1 100 200 5 200 10 25 400 5 100 5 1000 10 1 400 0.1 9 20 2000 10 1800 7 11 2.1 0.1 405 790 15 705 18 90 1400 9.2-8 Threshold Limit Values 57 Table 2-8 (continued) TLV-TWA~ OSHA PEL3 Substance' Chloroform4 Cyclohexane Cyclohexanol (skin) Cyclohexanone (skin) Cyclohexene Cyclopentane Diborane 1.1 100 200 25 10 100 400 10 100 200 1000 1 1 400 10 0.25 200 mg/m3.2 1.3 2 0.

25°C 'Latest NIOSH Pocket Guide information is available at the NIOSH web site: https//www. 'NIOSH Pocket Guide to Chemical Hazards (Cincinnati: National Institute for Occupational Safety and Health. 1991-1994). . 'Human carcinogen. (Cincinnati: American Conference o f Governmental Industrial Hygienists. 2000). 2Documentation of the Threshold Limit Values and Biological Exposure Indices.cdc. 5th ed.58 Chapter 2 Toxicology Table 2-8 (continued) TLV-TWA2 OSHA PEL3 PPm Methyl acetylene Methyl alcohol Methylamine Methyl bromide (skin) Methyl chloride Methylene chloride4 Methyl ethyl ketone Methyl formate Methyl isocyanate (skin) Methyl mercaptan Naphthalene Nitric acid Nitric oxide Nitrobenzene (skin) Nitrogen dioxide Nitromethane Nonane Octane Oxalic acid Ozone Pentane Phenol (skin) Phosgene Phosphine Phosphoric acid Phthalic anhydride Pyridine Styrene Sulfur dioxide Toluene (skin) Trichloroethylene Triethylamine Turpentine Vinyl acetate Vinyl chlorides Xylene mg lm3.gov/niosh.25°C PPm mg/m3. 4Possible carcinogen.

The TLV assumes that workers are exposed only during a normal eight-hour workday. (1942).Problems 59 The ACGIH clearly points out that the TLVs should not be used as a relative index of toxicity (see Figure 2-8). 1991). p. Problems 2-1. Frank P. 14. 1984). J. Burson." Ann. (London: Butterworths. Frank P. Industrial Toxicology: Safety and Health Applications in the Workplace (New York: Van Nostrand Reinhold. Dangerous Properties of Industrial Materials.p. OERR 9200. 29: 69-81. 1985). 1. 207. Fawcett and William S. Suggested Reading Toxicology Howard H. 2d ed. Appl. 6th ed. Wood. Health Effects Assessment Summary Tables (HEASH). should not be used for air pollution work. Every effort must be made to reduce worker exposures to toxicants to below the PEL and lower if possible. Phillip L. 1971). Finney. T. 1986). and cannot be used to assess the impact of continuous exposures to toxicants.6-303 (Cincinnati: Center for Environmental Research Information. with Further Observations on the Rotenone Equivalent Method of Assessing the Toxicity of Derris Root. 1982). Probit Analysis D. Irving Sax. Biol.15. 2d ed. ch. VI. (Cincinnati: American Conference of Governmental Industrial Hygienists. Safety and Accident Prevention in Chemical Operations. 5J. (New York: Wiley. updated regularly). (New York: Van Nostrand Reinhold. Loss Prevention in the Process Industries. Williams and James L. sec. Loss Prevention in the Process Industries (London: Butterworths. eds. Finney4 reported the data of Martin5involving the toxicity of rotenone to the insect species Macrosiphoniella sanborni. and 25. 1996). Probit Analysis (Cambridge: Cambridge University Press. 1986). Threshold Limit Values Documentation of the Threshold Limit Values and Biological Exposure Indices. N. Finney. The rotenone was applied in a medium of 0.. 1971). J. Integrated Risk Information System (IRIS) (Cincinnati: Center for Environmental Research Information. The Toxicity of 1Elliptone and of Poisons Applied Jointly. 20.. 5th ed. Probit Analysis (Cambridge: Cambridge University Press. 2-2. Lees. eds.5% saponin. "The Problem of the Evaluation of Rotenone-Containing Plants. Lees. . Derive Equation 2-7. 4D. Martin.

and plot the probit versus the natural logarithm of the dose. plot the percentage of insects affected versus the natural logarithm of the dose.8 log r. Compare the probit and number of insects affected predicted by the straight-line fit to the actual data. determine a straight line that fits the data.000 N/m2.1. Assume a volatile species molecular weight of 100. the data were represented by the normal distribution function . Under most circumstances the air in a room cannot be assumed to be well mixed. A certain volatile substance evaporates from an open container into a room of volume 1000 ft3. 2-3. where P is the overpressure in psi and r is the distance from the blast in feet. how many ft3/min of fresh air must be supplied to ensure that the concentration of the volatile is maintained below its TLV of 100 ppm? The temperature is 77°F and the pressure is 1atm. The evaporation rate is determined to be 100 mglmin. From the given data. 2-5. The insects were examined and classified one day after spraying. part c. What fraction of structures will be damaged by exposure to this overpressure? What fraction of people exposed will die as a result of lung hemorrhage? What fraction will have eardrums ruptured? What conclusions about the effects of this blast can be drawn? 2-4. The peak overpressure expected as a result of the explosion of a tank in a plant facility is approximated by the equation log P = 4. The obtained data were: Dose of rotenone (mg/l) Number of insects Number affected a. Convert the data to a probit variable. b.2 . A blast produces a peak overpressure of 47. Be sure to state any additional assumptions. How would poor mixing affect the quantity of air required? 2-6. Estimate the number of fatalities expected as a result of lung hemorrhage and the number of eardrums ruptured as a result of this blast. The plant employs 500 people who work in an area from 10 to 500 ft from the potential blast site. If the result is linear. If the air in the room is assumed to be well mixed. In Example 2-1.60 Chapter 2 Toxicology containing 5% alcohol.

gov/niosh) to acquire the meaning and definition of IDLH concentration. carbon monoxide. 1000 pprn for 1hour. 2-11. Compare the result to the results from the probit equation (Table 2-5). Determine the concentration of ethylene oxide that will cause a 50% fatality rate if the exposure occurs for 30 min. carbon monoxide.Problems 61 Use this distribution function to determine the fraction of individuals demonstrating a response between the range of 2. and specific gravity. 2-16. The NIOSH web site states that deaths occur as a result of ammonia exposures between 5. A group of 100 people is exposed to phosgene in two consecutive periods as follows: (a) 10 pprn for 30 min and (b) 1 pprn for 300 min. Use the NIOSH web site to determine and compare the PEL. Assume 25°C and 1 atm. 2000 pprn for 2 hours. Determine the duration times.for ammonia. 2-18. If 500 workers in a plant are exposed to the following concentrations of ammonia for the given number of hours. c. Use the probit equation (Equation 2-5) to determine the expected fatalities for people exposed for 2 hours to each of the IDLH concentrations of ammonia. ethylene oxide.5. The following physical property data are for acetone: molecular weight. and hydrogen chloride. d. 2-20. and LC5(.1. chlorine. b. IDLH concentration. 300 pprn for 3 hours. 2-12.7899. 2-13. 0. 2-7. 2-10. that a group of 100 people can be exposed to 1500 pprn of carbon monoxide to result in (a) 0% fatalities and (b) 50% fatalities. how many deaths will be expected? a. 2-8. Use the NIOSH web site to determine the number of deaths that occurred in 1992 as a result of asbestos. and ethylene oxide. Use the NIOSH web site (www. 2-9. Use the NIOSH web site to determine and compare the PEL. and chlorine. ethanol. 2-14. 2-15. ethylene trichloride.cdc. Determine the expected number of fatalities. 150 pprn for 2 hours. benzene. in minutes. .000 and 10. Use the NIOSH web site to determine an escape time period for a person subjected to an IDLH concentration. 2-17. Use the NIOSH web site to determine and compare the PEL and the IDLH concentration of ethylene oxide and ethanol. IDLH concentration. How much acetone liquid (in milliliters) is required to produce a vapor concentration of 200 pprn in a room of dimension 3 X 4 X 10 m? The temperature is 25°C and the pressure is 1 atm. 58. and hydrogen chloride.000 pprn over a 30-min period. Use Equation 2-7 to convert the TLV in pprn to the TLV in mg/m3for benzene.5 to 7. 2-19. and TLV for ethylene oxide. fluorine.

50 pprn for 2 min. Estimate the exposure concentration in pprn that will result in fatalities for 80% of the exposed individuals if they are exposed to phosgene for 4 min. "2-30. 5. use a lookup function or an equivalent function. percentages. 100 pprn for 5 min. 2-25. 287288. Use Equation 2-6 to convert probits of 3.. Lotus. low) for the inhalation of toxic chemicals. to find the toxicity levels (high. determine the probit constants and the LC. 2-27. Use Louvar and Louvar. c.8 2. to find the toxicity levels (high. 287-288.0. Use Joseph F.7 5. and 6. 2-23.2 7. and compare with the values shown in Table 2-4.28 to percentage affected. Excel) to solve Problem 2-4.6 Number of insects 50 49 46 48 50 Number affected (deaths) 44 42 24 16 6 . Diane Louvar. Your spreadsheet output should have designated columns for the distance. NJ: Prentice Hall. Louvar and B. You should also have two spreadsheet cells that provide the total number of individuals with eardrum ruptures and lung hemorrhage deaths. Use a small enough distance increment so that the results are essentially independent of the increment size. probit values. 50 pprn for 20 min.1 3.2 2 . 2-26. Health and Environmental Risk Analysis. Estimate the exposure concentration in pprn that will result in fatalities for 80% of the exposed individuals if they are exposed to chlorine for 4 min. 2-28.: Dose of rotenone (mg/l) 10. 2-29. 200 pprn for 150 min. pp. b. c. medium.72. 1998). 200 pprn for 15 min. Break the distance from 10 ft to 500 ft into several intervals.62 Chapter 2 Toxicology '2-21. Use the results of Problem 2-21 to establish the recommended distance between the control room and the tank if the control room is designed to withstand overpressures of (a) 1psi and (b) 3 psi. Health and Environmental Risk Analysis: Fundamentals with Applications (Upper Saddle River. pressure. Use a spreadsheet program (such as QuattroPro. 4 . and the number of individuals affected for each increment. pp. For converting from probits to percentages. Using the following data. low) for the single dose of a chemical that causes 50% deaths. medium. Determine the potential deaths resulting from the following exposure to chlorine: a. b. 2-24. 100 pprn for 50 min. Determine the potential deaths resulting from the following exposure to chlorine: a.

developing procedures for the handling of hazardous materials. In this chapter we describe the relationship between laws and regulations as an introduction to industrial hygiene. Typical projects involving industrial hygiene are monitoring toxic airborne vapor concentrations. and other physical factors to ensure that workers are not exposed to harmful levels. In chemical plants and laboratories the industrial hygienist works closely with safety professionals as an integral part of a safety and loss prevention program. reducing toxic airborne vapors through the use of ventilation. Industrial hygiene is a science devoted to the identification. Control: application of appropriate technology to reduce workplace exposures to acceptable levels. heat. evaluation. radiation. the industrial hygienist makes recommendations relevant to control I . Evaluation: determination of the magnitude of the exposure. Industrial hygienists are also responsible for selecting and using instrumentation to monitor the workplace during the identification and control phases of industrial hygiene projects. Many of the results in this area are due to civic concern and ethics. After identifying and evaluating the hazards.Industrial Hygiene ndustry and society are continuing to focus on reducing personnel and environmental damage resulting from accidents. and monitoring and reducing noise. sometimes manifested in laws and regulations. and control of occupational conditions that cause sickness and injury. evaluation. selecting proper personal protective equipment to prevent worker exposure. The three phases in any industrial hygiene project are identification. and control: Identification: determination of the presence or possibility of workplace exposures.

To put these laws into effect. to create regulations and/or standards. Congress authorizes governmental organizations. including the EPA and OSHA. Step 3: The complete text of the law is published in the United States Code (USC). To make the laws work on a day-today level. quantities of the toxic chemicals that are legally emitted into the air.] The code is the official record of all federal laws. Congress is responsible for passing laws that govern the United States. If approved. the EPA has the responsibility (1) to help citizens comply with the law and (2) to enforce the regulation. safety professionals. how is it put into practice? Laws often do not include the details for compliance. Regulations set specific rules about what is legal and what is not legal. it becomes a law that is called an act. a regulation relevant to the Clean Air Act will specify levels of specific toxic chemicals that are safe. and what penalties are given if the legal limits are exceeded. Creating a Law A law is created with a three-step process: Step 1:A member of Congress proposes a bill. who has the option to either approve it or veto it. becomes a law. Congress authorizes certain government organizations. After the regulation is in effect. if approved. The industrial hygienist. Creating a Regulation After the law is official. For example. It has been clearly demonstrated that toxic chemicals can be handled safely when principles of industrial hygiene are appropriately applied. it is sent to the president. the USC requires the appropriate respirator protection. Step 2: If both houses of Congress approve the bill. For example. including the Environmental Protection Agency (EPA) and OSHA. . A bill is a document that. and plant operations personnel work together to ensure that the control measures are applied and maintained. 3-1 Government Regulations Laws and Regulations Laws and regulations are major tools for protecting people and the environment. but it does not specify the detailed types or limitations of respirators. to create and enforce regulations.64 Chapter 3 Industrial Hygiene techniques.

758-777. sufficient funding was committed to create and support the Occupational Safety and Health Administration (OSHA). and (4) training professionals for administering the programs of the act. This produced relatively inconsistent and ineffective results. To appreciate the significance of the OSHAct. In 1936 the federal government enacted the Walsh-Healy Act to establish federal safety and health standards for activities relating to federal contracts. Step 2: After a regulation is rewritten. develops. and developing and maintaining safety and health records. were the forerunners of our current occupational health and safety regulations. recognition. These comments are used to revise the regulation. (2) developing criteria for handling toxic materials. Some laws and regulations particularly relevant to the chemical indus2J. This 1936 act also initiated significant research related to the cause. The proposal is listed in the Federal Register (FR) so that the public can evaluate it and send comments to the organization. and OSHA uses these data to promulgate standards. and the National Institute for Occupational Safety and Health (NIOSH). Although some progress was made.3-1 Government Regulations 65 The process for creating a regulation andlor standard has two steps: Step 1:The authorized organization or agency decides when a regulation is needed. Examples of NIOSH responsibilities include (1) measuring health effects of exposure in the work environment. and it is simultaneously codified by publishing it in the Code of Federal Regulations (CFR). although not adequate by today's standards. these regulations were never sufficientlysupported to carry out a satisfactory program. The OSHAct of 1970 was developed to solve these problems and to give a nationally consistent program with the funding necessary to manage it effectively. In 1970 the US Congress enacted a health and safety law that continues to have a significant impact on the practices of industrial hygiene in the chemical industry: the Occupational Safety and Health Act of 1970 (OSHAct). (Chicago:National Safety Council. The concepts promulgated by the Walsh-Healy Act. it is posted in the Federal Register as a final rule. (3) establishing safe levels of exposure. The organization then researches. Before 1936 regulations concerning occupational health were poorly administered by state and local governmental agencies. During this era. and control of occupational disease. . staffs and funds were too small to carry out effective programs. As a result of the OSHAct. ' NIOSH develops data and information regarding hazards. it is helpful to review regulations and practices2 before 1970. pp. 1979). B. Fundamentals of Industrial Hygiene. and proposes a regulation. conducting investigations for compliance. which manages and administers the government's responsibilities specified in the OSHAct.ed. Olishifski. 2d ed. which conducts research and technical assistance programs for improving the protection and maintenance of workers' health.. This act defined clear procedures for establishing regulations. Between 1936and 1970a number of states enacted their own safety and health regulations.

Government regulation will continue to be a significant part of the practice of chemical process safety. HAZMAT (1989) Exposure to Hazardous Chemicals in Laboratories (1990) Annual Inventory of Hazardous Chemicals. . Highlights of OSHA enforcement rights are illustrated in Table 3-2. these standards will continue to create an environment for improving process designs and process conditions relevant to the safety and health of workers and the surrounding communities. OSHA is authorized. Hazardous Materials Technician. the distinction between laws (USC) and regulations (CFR) is global versus detail. interpretations. Since the OSHAct was signed into law. The OSHAct makes employers responsible for providing safe and healthy working conditions for their employees. and when violations of the safety and health standards are found.1450 40 CFR 370. SARA 311 (1991) Process Safety Management (1992) Risk Management Program (1996) Respirator Program (1998) 29 USC 651 42 USC 7401 33 USC 1251 42 USC 7401 15 USC 2601 42 USC 300f 40 CFR 280. OSHA officers finding conditions of imminent danger may request plant shutdowns.134 try are shown in Table 3-1. SARA 313 (1989) Training. The implications. Criminal penalties can be invoked.66 Chapter 3 Industrial Hygiene Table 3-1 Number A Few Laws (USC) and Regulations (CFR) Description Occupational Safety and Health Act (1970) Clean Air Act (1970) Clean Water Act (1977) Clean Air Act Amendments (1990) Toxic Substances Control Act I1 (1992) Safe Drinking Water Act Amendment (1996) Underground Storage Tank Leak Tests (1988) Annual Toxic Release Report.120 29 CFR 1910.119 40 CFR 68. OSHA's right of inspection includes safety and health records.20 40 CFR 370.20 29 CFR 1910. however. Especially within the chemical industry. A search warrant may be required to show probable cause. and applications of the OSHAct will continue to develop as standards are promulgated. substantial new legislation controlling the workplace and community environment has been enacted. Table 3-3 provides a summary of Table 3-2 Highlights of OSHA's Right of Enforcement Employers must admit OSHA compliance officers into their plant sites for safety inspections with no advance notice. they can issue citations and financial penalties.30 29 CFR 1910. As illustrated in this table.65 29 CFR 1910. to conduct inspections.

and Liability Act of 1980 (Superfund) Hazardous and Solid Waste Amendments Asbestos Hazard Emergency Response Act Superfund Amendments and Reauthorization Act Emergency Planning and Community Right-to-Know Act Toxic Substances Control Act (continued) . Research. and Cosmetic Act Federal Insecticide. and Rodenticide Act Dangerous Cargo Act Federal Water Pollution Control Act Flammable Fabrics Act Atomic Energy Act Fish and Wildlife Act of 1956 Federal Hazardous Substances Labeling Act Solid Waste Disposal Act ' Metal and Non-Metallic Mine Safety Act National Environmental Policy Act Federal Coal Mine Health and Safety Act Clean Air Act Poison Prevention Packaging Act of 1970 Water Quality Improvement Act of 1970 Federal Railroad Safety Act of 1970 Resource Recovery Act of 1970 Occupational Safety and Health Act Noise Control Act of 1972 Federal Environmental Pollution Control Act Hazardous Materials Transportation Act Consumer Product Safety Act Marine Protection. and Sanctuary Act of 1972 Clean Water Act Coastal Zone Management Act Endangered Species Act of 1973 Safe Drinking Water Act Transportation Safety Act of 1974 Energy Supply and Environmental Coordination Act Toxic Substances Control Act Resource Conservation and Recovery Act Federal Mine Safety and Health Act Surface Mine Control and Reclamation Act Uranium Mill Tailings Control Act Port and Tanker Safety Act Comprehensive Environmental Response.3-1 Government Regulations 67 Table 3-3 Federal Legislation Relevant to Chemical Process Safety1 Date Abbreviation RHA FDCA FIFRA DCA FWPCA FFA AEA FWA FHSA SWDA MNMSA NEPA I'MHSA "AA I'PPA WQI RSA RRA OSHA NCA FEPCA IlMTA CPSA MPRSA CWA CZMA ESA SDWA TSA ESECA TSCA RCRA FMSHA SMCRA UMTCA PTSA CERCLA HSWA AHERA SARA EPCRA TSCA Act River and Harbor Act Federal Food. Compensation. Fungicide. Drug.

The PSM standard has 14 major sections: employee participation. management of change. emergency planning and response. pressures. it sets general requirements for the management of hazardous chemicals. and Figure 3-1 shows how the amount of legislation has increased." This standard is performance oriented.html. it is important that chemical engineers be aware of the law to ensure that their facilities comply. Employee participation requires active employee participation in all the major elements of PSM. process safety information. However. Process safety management (PSM) was developed after the Bhopal accident (1985). A brief description of each section is given in what follows.1992. A description of this legislation is well beyond the scope and goals of this textbook. This . Water Quality Act Oil Pollution Act of 1990 Clean Air Act Amendments Pollution Prevention Act of 1990 Residential Lead-Based Paint Hazard Abatement Act Federal Facility Compliance Act National Environmental Policy Act Hazardous Materials Transportation Act Amendments Safe Drinking Water Act Amendment Food Quality Protection Act lInformation from the EPA and OSHA web sites: www. operating procedures. pre-startup safety review. OSHA published the final rule "Process Safety Management of Highly Hazardous Chemicals. flows. process hazard analysis. Employers must develop and document a plan of action to specify this participation.gov/epuhome and www.68 Chapter 3 Industrial Hygiene Table 3-3 (continued) Date Abbreviation Act WQA OPA CAAA PPA TSCA-TI - NEPA HMTAA SDWA FQPA . process chemistry. incident investigations. audits. hot work permits.osha-slc. This information includes block flow diagrams or process flow diagrams. such as temperatures. OSHA: Process Safety Management On February 24.epa. mechanical integrity.gov/oshsrrl-toclosha-std-toc relevant safety legislation. Consequences of process deviations are also required. to prevent similar accidents. and trade secrets. and process limitations. and compositions. contractors. It is recognized by industry and the government as an excellent regulation that will reduce the number and magnitude of accidents -if it is understood and practiced as intended. training. that is. Process safety information is compiled and made available to all employees to facilitate the understanding and identification of hazards.

management of change. neers. such as what-if scenarios. and for less complex processes a less rigorous process. industrial hygienists. and other appropriate and experienced specialists. checklists. Process hazard analysis ( P H A ) must be performed by a team of experts. and accident investigations. . including engiI . a hazards and operability (HAZOP) study for a complex process. chemists. process hazards analysis. . The PHA needs to include a method that fits the complexity of the process. operators. or fault trees. process safety information is needed before training.3-1 Government Regulations 69 X 0 1900 1910 X I I I I I I I I 1920 1930 1940 1950 Year 1960 1970 1980 1990 I 2000 Figure 3-1 Number of federal laws relevant to chemical process safety. failure mode and effects analysis.

An effective training program helps employees understand the hazards associated with the tasks they perform. (2) the safety. an inspection before the work.70 Chapter 3 Industrial Hygiene Employers must ensure that the recommendations from the PHA are acted on in a timely manner. the employees need to consider the contractors' safety performance in addition to their skills. A pre-startup safety review is a special safety review that is conducted after a modification to the process or operating conditions has been made and before the startup. the operators decide on the frequency of the refresher training. temporary operations. Even when selecting contractors. These instructions need to be clearly written and consistent with the process safety information. . exposure precautions. or using spark-producing equipment). (3) periodic inspections and testing based on vendor recommendations. including emergency shutdowns. initial startup. special or unique hazards. identification of fire watches. such as hot work. and alarms are mechanically sound and operational. normal operations. maintenance. These operating procedures are updated frequently. and confined space. Operators need to understand the hazards associated with every task. hazardous properties of the chemicals. (3) the appropriate training is completed. startups. identification of openings where sparks may drop. Maintenance and operations personnel receive initial training and refresher training. They need to cover. Contractors are trained to perform their tasks safely to the same extent as employees. and ( 5 ) a process to ensure that all equipment and spare parts are suitable. Every PSM process needs an updated PHA at least every five years after the initial analysis is completed. at a minimum. In this review a team of reviewers ensures that (1) the system is constructed in accordance with the design specifications. Refresher training is given every three years and more often if necessary. . with the frequency being set by the operating personnel. and normal operations. and emergency procedures are in place. Operating procedures that facilitate the safe operation of the plant must be documented. The mechanical integrity section of the PSM standard ensures that the equipment. operating. the equipment involved in the work. normal shutdown. The permit requires dates authorized for hot work. The PSM standard ascertains that a system is in place to prepare and issue hot workpermits before conducting hot work activities (welding.. safety and health considerations. emergency operations. quality control specifications for all chemicals. controls. the types and numbers of fire extinguishers. and (4) the recommendations from the PHA are implemented or resolved. engineering and administrative controls. relief systems. piping. The requirements include (1)written procedures to maintain functioning systems. lockout/tagout. Safe work practices also need to be documented. emergency shutdown. grinding. (2) training regarding preventive maintenance. (4) a process to correct deficiencies. startup after normal and emergency shutdowns. operating limits and consequences of deviations. authorization signatures. identification of flammable materials in the area. and safety control systems and functions. a system to maintain and document certification.

The regulation requires an investigation team composed of people. including operators. all the affected employees are trained. The RMP is required for plant sites that use more "ode of Federal Regulations. 20. The trade secrets section of the PSM standard ensures that all contractors are given all the information relevant to operating in the plant safely. 40 CFR 68. The RMP regulation is aimed at decreasing the number and magnitude of accidental releases of toxic and flammable substances. identification and closure of open vessels or ducts.3-1 Government Regulations 71 verification that the surrounding area is not explosive. Under the management of change section of the PSM standard employees are required to develop and implement documented procedures to manage changes in the process chemistry. The PSM standard mandates incident investigation. Under the audits section of the PSM standard employers are required to certify that they have evaluated their compliance with the standard at least every three years.1996). The intent of the PSM element for emergency planning and response is to require employers to respond effectively to the release of highly hazardous chemicals. the EPA published the Risk Management Plan (RMP) as a final rule. this element should be part of a program for even the smallest organizations that handle hazardous chemicals. Some personnel may need to sign secrecy agreements before they receive this information.? This regulationis also a response to the Bhopal accident. Although the RMP is similar to the PSM regulation i n many respects. The audit reports need to be retained as long as the process exists.1996. Employers must investigate all incidents that have or could have resulted in a major release or accident within 48 hours of the event. and operating procedures. The recommendations from the audit must be followed. who are knowledgeable about the system. and veritication that welded walls are not flammable. DC: US Government Printing Office. Before a change occurs (except for replacementin-kind). and a pre-startup review is conducted. Jun. EPA: Risk Management Plan On June 20. the employers are required to appropriately use the investigation recommendations. the RMP is designed to protect off-site people and the environment. Although the regulation requires this activity for companies with more than 10 employees. subpart B (Washington. it must be reviewed to ascertain that it will not affect the safety of the operation. whereas PSM is designed to protect on-site people. . It is recognized by industry and the government as an excellent regulation that will reduce the number and magnitude of accidents if it is understood and practiced as intended. After the change has been made. verification that combustible materials are removed or covered appropriately. After the investigation. process equipment.

documentation that is maintained on the site and submitted to federal. Alternative release cases for toxic substances cover scenarios with toxic concentrations beyond the fenceline. because risk is composed of both conse- . and only one flammable substance is analyzed for other more likely scenarios. The worst-case scenario is based on releasing the entire contents of a vessel or piping system in a 10-minute period under worst-case meteorological conditions (F stability and 1. The fourth part. emergency response program. is self-explanatory. The releases must include the worst-case scenario and the more likely but significant accident release scenarios. dikes) can be used in the calculation process. therefore the release rate for liquid spills corresponds to the evaporation rate.5 m/s wind speed). including the history of such releases at the facility. Therefore the results are not a true determination of risk. Those with no off-site potential damage are not required to be reported. state. and local authorities. The EPA requires the following consequence analyses: (1) A single worst-case release scenario is analyzed for all covered flammable materials on the site. these concentrations are the basis for determining the consequences resulting from toxicity. prevention program. The RMP requires only an analysis of the consequence and not the probability. and/or explosions. Dispersion model calculations are normally used to estimate downwind concentrations. For those not interested in using dispersion models. The first three parts of the regulation are described briefly in the following paragraphs. and (2) a single worst-case release scenario is analyzed for all toxic substances on the site. whereas PSM covers every covered process on the site. The RMP document is updated when the process or chemistry changes or when a governmental audit requests an update. This information is also shared with the local community. documentation.72 Chapter 3 Industrial Hygiene than a specified threshold quantity of regulated highly hazardous chemicals. The RMP has the following elements: hazard assessment. Passive mitigation measures (for example. fires. The release scenarios that have a potential to reach the public are of the greatest concern. and more likely releases are analyzed for each toxic substance covered by the rule. The RMP is a site responsibility (the site may have several processes). Alternative cases for flammable substances cover scenarios that may cause substantial damage off site and on site. A risk matrix can be used to characterize the worst-case and more likely scenarios. the standard includes lookup tables for all the listed substances to help a facility determine the impact distances for specific release scenarios. Hazard assessment is a consequence analysis for a range of potential hazardous chemical releases.

A more detailed description of the required consequence analyses can be found el~ewhere.~ The second requirement of the RMP is a prevention program. Training is for all employees on the topics relevant to emergency response. The first column in Table 3-4 lists each element of the PSM program. many of these elements are duplicated. the Resource Conservation and Recovery Act (RCRA). the EPA made a deliberate attempt to retain the same requirements wherever possible. As with similar OSHA regulations.3-1 Government Regulations 73 Table 3-4 Comparison of the PSM and RMP Prevention Programs RMP (EPA) Process safety information Hazard evaluation Standard operating procedures (No equivalence) Training (No equivalence) Pre-startup review Maintenance (No equivalence) Management of change Accident investigations Emergency response Safety audits (No equivalence) Risk assessment PSM program (OSHA) Process safety information Process hazards analysis Operating procedures Employee participation Training Contractors Pre-startup review Mechanical integrity Hot work permit Management of change Incident investigations Emergency planning and response Compliance audits Trade secrets (No equivalence) quence and probability. It also establishes procedures for notifying the local community and the appropriate emergency response agencies. The prevention program has 11 elements. Crowl. although differences exist because the EPA and OSHA have different responsibilities. The requirements include drills to test the plan and to evaluate its effectiveness. As shown in Table 3-4. Fortunately. . and the plan must be revised based on the findings of these drills. and the 4Daniel A. 1-5. pp." Process Safety Progress (Spring 1997). "Consequence Modeling for the EPA Risk Management Plan (RMP). The plan must be coordinated with local emergency response plans developed by Local Emergency Planning Committees (LEPCs) and local emergency response agencies. The emergency response program delineates the steps to be taken by the facility's employees in response to accidental releases of hazardous materials. and the second column shows the corresponding element of the prevention program (some elements have no equivalence). compared to the 14 elements of the PSM standard.

especially because the listed hazards can also act in combination. is so complex that this task requires the concerted efforts of industrial hygienists. The identification step requires a thorough study of the chemical process. laboratory personnel. and management. operating conditions. all potential hazards must be identified and controlled.74 Chapter 3 Industrial Hygiene Spill Prevention Control under the Clean Water Act. operating instructions. and information from operating personnel. The quality of this identification step is often a function of the number of resources used and the quality of the questions asked.in large plants there may be thousands. operators. Determining the potential for hazards to result in an accident (risk assessment) is frequently part of the identification step (see chapter 11). and attention to detail. information from chemical suppliers. To achieve this operating success. 3-2 Industrial Hygiene: Identification One of the major responsibilities of the industrial hygienist is to identify and solve potential health problems within plants. skill. To be safe under these conditions requires discipline. Many hazardous chemicals are handled safely on a daily basis within chemical plants. however. engineers. Resources for evaluating the hazards and developing control methods are allocated on a priority basis. As illustrated in Table 3-5. the potential hazards and methods of contact are identified and recorded. Material Safety Data Sheets One of the most important references used during an industrial hygiene study involving toxic chemicals is the material safety data sheet (MSDS). When toxic and/or flammable chemicals are handled. equipment vendor descriptions. safety reviews. The industrial hygienist helps the effectiveness of the overall program by working with these plant personnel. process designers. concern.This list of potential hazards and their risk is used during the evaluation and control phase of the project. the emergency response plan must be maintained at the facility and must include descriptions of all mitigating systems. The sources of information include process design descriptions. In this identification step it is often necessary to collate and integrate the available information to identify new potential problems resulting from the combined effects of multiple exposures. the potentially hazardous conditions may be numerous . giving the appropriate time and attention to the most significant hazards. Chemical process technology. The different resources may have different operating and technical emphases unique to pieces of equipment or specific chemicals. and operating procedures. the potential hazards are numerous. This list of potential hazards together with the required data for hazard identification (see Table 3-6) is commonly used during the identification step of industrial hygiene projects. A sample MSDS is shown in Fig- . For these reasons industrial hygiene (particularly identification) must be a part of the education process of chemists. During the identification step. and managers.

. The MSDS lists the physical properties of a substance that may be required to determine the potential hazards of the substance. and vapors Noise levels of equipment Types and degree of radiation ure 3-2. (2) a commercial source. Fundamentals of Industrial Hygiene. MSDSs are available from (1) the chemical manufacturer. pp. dusts. The industrial hygienist or safety professional must interpret the physical and toxicological properties to determine the hazards associated with a chemical. Table 3-6 Data Useful for Health ldentification p p -. or (3) a private library developed by the chemical plant. 24-26.3-2 Industrial Hygiene: Identification Table 3-5 ldentification of Potential Hazards1 Potential hazards Liquids Vapors Dusts Fumes Noise Radiation Temperature Mechanical Entry mode of toxicants Inhalation Body absorption (skin or eyes) Ingestion Injection Potential damage Lungs Ears Nervous system Kidneys Circulatory system Skin Eyes Liver Reproductive organs Other organs 'Olishifski.- - - - Threshold limit values (TLVs) Odor threshold for vapors Physical state Vapor pressure of liquids Sensitivity of chemical to temperature or impact Rates and heats of reaction Hazardous by-products Reactivity with other chemicals Explosive concentrations of chemicals. These properties are also used to develop a strategy for the proper control and handling of these chemicals.

swat~on nal (Non-Mandatory Form) Form Approved OM0 No 1218-0072 @ Note Blank spaces are not perm.F i r e a n d E x p l o s ~ o nHazard Data Flammable Limlls LEL UEL plash Point (MetOod Us&) Exttngu~snlng Wedla Specla1 Ffre Flgntlng P<aeducer Unusual Ftre and Ezolos on Hazards (Reproduce locallyi OSHA 174 S e ~ l1985 Figure 3-2 Material safety data sheet. Most companies use their own MSDS format.Chapter 3 Industrial Hygiene 5-210 MATERIAL SAFETY DATA SHEET 31 9203 Mater~al Safety Data Sheet May be used to comply with OSHA s Hazard Communlcat~onStandard 25 CFR 1910 1200 Slandard must be requsements consulted lor spec111c IDENTITY (As UsW on LaOel and bslJ Sectlon I Manulacturer s Name Address (Number Street Cl3. Sfare and ZIP Code) U.S av81Iable rhe space mvsr be maned ro indica e rnar Emergency TeIephone Number Telephone Numcer lor Intormat~on Dale Preparea S~gnalure Prepare.nea It any item e nor appncable or Pa inlormaron .S.1i I 1 Vapor Pressure lmm Hg I Vapor Densty (AIR = 1) Melting Polnl €vaporarcon Rate IBu3.Hazardous l n g r e d ~ e n t s / l d e n t l t ylnformatlon OSHA PEL ACGlH TLV Other Ltm8ls Recommended Hazardous Components (Specific Chemcal ldenrtry Common Narnels)) P o ioplronali Sectton Ill Botllng Poinl . (optional) of S e c t l o n II .P h y s ~ c a l / C h e m ~ cC h a r a c t e n s t ~ c s al Speclhc Gravlv I H D .l Acetale 1) - bluQ8itty tn Water Appearance and Wor Sect~on IV . . Department of Labor O c c u ~ a ~ ~ oSafety and Health Adrnn.

Control Measures Local Exhaust Specla1 Olher Eye P~MBnlcn Resp~r~dwy Pmtat~on (Spany Tlpe) Mechanical (Gemw) PrMBnlva Gloves Olhw PfUBntve Cblhlng or Equtpment WwWHyglenlc Pracltcer PW 2 Q usoro 916 .Health nszard Data I Skln'J Ingestdon? Route(s)ol Enlv Health Hazards (Acvm and C h m ) Carcnnogen!ctty NTP7 IARC Morographs? OSHA Regulaled? Sagns and Symptoms or Exposure -- Medeal Condanons Generally Aggravated b .- - 3-2 Industrial Hygiene: Identification 77 - 31 9204 REFERENCE FILE Scctlon V SlablLn\ .Relcthrlty Data U M W Stable C o n d W to Avad --- lncompattblllty (Malertals 10 Awe H~ardous DewmpOQ~t~On BypmduM or Hazardws PolymertzaUon May Occur Wnll C o n d l t ~ Avotd lo Na Occur Inhalaton? Section Vi . i .+l-m -5. Exposure Emergency and Ftrst Aid Procedures Section VII Precaut~ons Safe Handling and Use for Steps to Be Taken in Case Maler8al Is Released or Spvlled - Wane Dnsposal M a h M Plscautlons lo Be Taken m Handllng and Slorlng Ofher Precaullons - Section Vlll Ventllatlon .

Clear. Throat irritation occurs with concentrations of 35 ppm. Large doses can cause death. Most ethers are dangerously flammable and explosive. Lewis. No hazard. 5R.. Solution Sax5 provided the technical information required to solve this problem. . phenol. toluene. High concentrations produce a narcotic effect. Clear. Reacts violently with a number of substances. Corrosive action on all body tissues. and ether. Highly reactive with a wide variety of substances. Clear. Acute and chronic exposures occur with concentrations greater than 200 ppm. crystalline mass that is most frequently found in solution form. Absorbed readily through the skin. Control techniques are presented in more detail in section 3-4. the various types of existing control measures and their effectiveness are also studied. colorless liquid with a dangerous fire hazard and a moderate explosion hazard. benzene. but it is also absorbed through the skin. the likelihood of large and small leaks must be considered. I. During the evaluation study. colorless liquid with no fire or explosion hazard. It is a moderate irritant to the skin. The following table summarizes the results: Chemical Sodium chloride Toluene Description and potential hazard Common table salt. 2000). Hydrochloric acid Phenol Sodium hydroxide Benzene Ether 3-3 Industrial Hygiene: Evaluation The evaluation phase determines the extent and degree of employee exposure to toxicants and physical hazards in the workplace environment. Identify the potential hazards in this laboratory. A skin and eye irritant. ed. and mucous membranes and by ingestion and inhalation. Entry into the body is mostly by inhalation. Sax's Dangerous Properties of Industrial Materials. sodium hydroxide. eyes. During the evaluation phase. hydrochloric acid. Irritant to skin and eyes. 10th ed. A wide variety of organic compounds that are mostly narcotic in effect.Chapter 3 Industrial Hygiene Example 3-1 A survey of a laboratory is made and the following chemical species are identified: Sodium chloride. It is a possible carcinogen. (New York: Wiley. Emits toxic fumes when heated. Exposures to skin areas as small as 64 in2have resulted in death in less than 1 hr. Entry into the body is mostly by vapor inhalation. A white. colorless liquid with a slight fire hazard and moderate explosion hazard. It is a moderate fire hazard.

or fits of coughing. In these circumstances some provision for continuous evaluation is necessary. PELS. if a worker is exposed for 12 hours to a concentration of chemical equal to the TLV-TWA. arise from repeated exposures to low concentrations. such as TLVs. Thus. then the TLV-TWA has been exceeded. To establish the effectiveness of existing controls. Small leaks of these substances might not become obvious for months or even years. is the worker shift time in hours. Special attention must be directed toward preventing and controlling low concentrations of toxic gases. that is. independent of the length of time actually worked in the shift. it is necessary to compare actual exposure levels to acceptable occupational health standards. . Continuous monitoring is not the usual situation because most facilities do not have the necessary equipment available. Evaluating Exposures to Volatile Toxicants by Monitoring A direct method for determining worker exposures is by continuously monitoring the air concentrations of toxicants on-line in a work environment. C(t) is the concentration (in ppm or mg/m3) of the chemical in the air and t. These standards together with the actual concentrations are used to identify the potential hazards requiring better or more control measures. For continuous concentration data C(t) the TWA (time-weighted average) concentration is computed using the equation TWA = where 1 6 '"C ( t )dt.or IDLH concentrations. burning eyes. If problems are evident. may lead to immediate acute effects. samples are taken to determine the workers' exposure to conditions that may be harmful. temporary controls such as personal protective equipment can be used. There is rarely lasting damage to individuals if they are removed promptly from the contaminated area. because the computation is normalized to 8 hours. such as unconsciousness. The integral is always divided by 8 hours. through large leaks. mostly by small leaks. controls must be implemented immediately. There may be permanent and serious impairments from such exposures. however. Chronic effects. Longer term and permanent controls are subsequently developed. continuous or frequent and periodic sampling and analysis is important. After the exposure data are obtained.3-3 Industrial Hygiene: Evaluation 79 Sudden exposures to high concentrations. Many toxic chemical vapors are colorless and odorless (or the toxic concentration might be below the odor threshold). In this case ready access to a clean environment is important.

then the workers are overexposed. + + 8 hr All monitoring systems have drawbacks because (1) the workers move in and out of the exposed workplace and (2) the concentration of toxicants may vary at different locations in the work area. If we assume that the concentration Ci fixed (6r averaged) the over the period of time Ti. The mixture TLV-TWA can be computed from If the sum of the concentrations of the toxicants in the mixture exceeds this amount. For mixtures of toxicants with different effects (such as an acid vapor mixed with lead fume) the TLVs cannot be assumed to be additive.80 Chapter 3 Industrial Hygiene The more usual case is for intermittent samples to be obtained. Ci the concentration of chemical i with respect to the other toxicants. . 20 ppm of cyclohexanol (TLV-TWA of 50 ppm). The combined exposures from multiple toxicants with different TLV-TWAs is determined from the equation where n is the total number of toxicants. is the TLV-TWA for chemical species i.then the workers are overexposed. TWA concentration is computed by TWA = CITl C2T2 . Industrial hygienists play an important role in the selection and placement of workplace monitoring equipment and the interpretation of the data. If more than one chemical is present in the workplace. representing worker exis posures at fixed points in time. . one procedure is to assume that the effects of the toxicants are additive (unless other information to the contrary is available). What is the mixture TLV-TWA and has this level been exceeded? . Example 3-2 Air contains 5 ppm of diethylamine (TLV-TWA of 10 ppm). and 10 ppm of propylene oxide (TLV-TWA of 20 ppm). If the sum in Equation 3-3 exceeds 1. and is (TLV-TWA). + CnT.

the worker is overexposed. Because the TLV for toluene is 100 ppm. The total mixture concentration is 5 + 20 10 = 35 ppm. (TLV-TWA)... Additional control measures need to be developed. the TLV-TWA has been exceeded. An alternative approach is to use Equation 3-3: + Because this quantity is greater than 1. . Example 3-3 Determine the 8-hr TWA worker exposure if the worker is exposed to toluene vapors as follows: Duration of exposure (hr) Measured concentration (PP~) Solution Using Equation 3-2. = 5 5 + 20 + 10 20 10 = 25 ppm.3-3 Industrial Hygiene: Evaluation 81 Solution From Equation 3-4. The workers are overexposed under these circumstances. On a temporary and immediate basis all employees working in this environment need to wear the appropriate respirators. TWA = CITl + C2T2+ C3T3 8 llO(2) - + 330(2) + 90(4) 8 = 155 ppm.

p. = 46. = pto. 685.622 .82 Chapter 3 Industrial Hygiene Example 3-4 Determine the mixture TLV at 25OC and 1 atm pressure of a mixture derived from the following liquid: Component Heptane Toluene Mole percent 50 50 Species TLV (ppm) 400 50 Solution The solution requires the concentration of the heptane and toluene in the vapor phase.2 mm Hg.5)(46.PYt. Basic Principles and Calculations in Chemical Engineering. 1989). . 5th ed.2 mm Hg) = 14. NJ: Prentice-Hall. 0. Himmelblauh provided the following data at the specified temperature: PEpta. The total pressure of the toxicants is (23.uene= (0.. Using Raoult's law.7 ppm. Himmelblau.0.5)(28..1) = 37. = x. pheptane (0. From Dalton's law the mole fractions on a toxicant basis are The mixture TLV is computed using Equation 3-4: TLVmix = 1 = 109.2 + 14. Assuming that Raoult's and Dalton's laws apply to this system under these conditions. psat ol. the vapor composition is determined directly from the saturation vapor pressures of the pure components. Assuming that the composition of the liquid does not change as it evaporates (the quantity is large). the TLVs for the individual species in the mixture are "avid M.3 mm Hg.2mm Hg.4 mm Hg.378 400 50 + Because the vapor will always be the same concentration. the partial pressures in the vapor are determined: p..1 mm Hg. = 28..4 mm Hg) = 23. (Englewood Cliffs. the vapor composition is computed using standard vapor-liquid equilibrium calculations..

- - - 3-3 Industrial Hygiene: Evaluation TLV.2 ppm. c 2 ----.5 ppm. . industrial hygienists. fit this category. Instead of using ppm as a concentration unit. who are specialists in this technology. = (0. Dust evaluation calculations are performed in a manner identical to that used for volatile vapors.5 p m (see chapter 2). The main reason for sampling for atmospheric particulates is to estimate the concentrations that are inhaled and deposited in the lungs. Particles larger than 0.2-0.7 ppm) = 41. .%) (mppcf) 70 30 20 2.8 mppcf. more control measures will be needed. TLV.7 Solution From Equation 3-4: TLV of mixture = 1 c. of course.2 pm settle out too slowly and are mostly exhaled with the air. = (0. Special control measures will be required when the actual particle count (of the size range specified in the standards or by an industrial hygienist) exceeds 6. whereas those smaller than 0. dusts.5 pm are usually unable to penetrate the lungs. For mixtures of vapors the individual species' TLVs in the mixture are significantly reduced from the TLVs of the pure substance. = 6.7 ppm) = 68.8 mppcf.+ TLV. mg/m3or mppcf (millions of particles per cubic foot) is more convenient.622)(109. Toxicological theory teaches that dust particles that present the greatest hazard to the lungs are normally in the respirable particle size range of 0. Example 3-5 Determine the TLV for a uniform mixture of dusts containing the following particles: Type of dust Nonasbestiform talc Quartz Concentration TLV (wt. Sampling methods and the interpretation of data relevant to health hazards are relatively complex.378)(109. TLV. .. If the actual concentration exceeds these levels. should be consulted when confronted with this type of problem. Evaluation of Worker Exposures to Dusts Industrial hygiene studies include any contaminant that may cause health injuries.

Table 3-7 contains dBA levels for a variety of common activities.. 3d ed. ed. = - (t). An absolute sound scale (in dBA for absolute decibels) is defined by establishing an intensity reference. the hearing threshold is set at 0 dBA. then the difference in intensity levels in decibels is given by Noise intensity (dB) 10 log. 1988). For convenience... Noise evaluation calculations are performed identically to calculations for vapors. . the industrial hygienist should immediately make the appropriate noise measurements and develop recommendations. (Chicago: National Safety Council.Chapter 3 Industrial Hygiene Evaluating Worker Exposures to Noise Noise problems are common in chemical plants. Noise levels are measured in decibels.s oj'lndustrial Hygiene. 168. Table 3-7 Sound Intensity Levels for a Variety of Common Activities l Sound intensity level (dB) Source of noise Riveting (painful) Punch press Passing truck Factory Noisy office Conventional speech Private office Average residence Recording studio Whisper Threshold of good hearing Threshold of excellent youthful hearing - 'B. Plog. If one sound is at intensity I and another sound is at intensity I. this type of problem is also evaluated by industrial hygienists. Some permissible noise exposure levels for single sources are provided in Table 3-8. A decibel (dB) is a relative logarithmic scale used to compare the intensities of two sounds. A. p. If a noise problem is suspected. Fundamenta1. Thus a sound 10 times as intense as another has an intensity level 10 dB greater. except that dBA is used instead of ppm and hours of exposure is used instead of concentration.

3-3 Industrial Hygiene: Evaluation

85

Table 3-8

Permissible Noise Exposures
Maximum exposure (hr)

Sound level (dB4

B. A. Plog, ed., Fundamentals of Industrial Hygiene, 3d ed. (Chicago: National Safety Council, 1988), p. 176.

'

Example 3-6
Determine whether the following noise level is permissible with no additional control features:

Noise level (dBA)
85 95 110

Duration (hr)
3.6 3.0 0.5

Maximum allowed (hr)
no limit 4 0.5

Solution
From Equation 3-3:

2

3

c,

-

-

(TLV-TWA),

3.6 nolimit

3 0.5 + - + - = 1.75. 4 05

Because the sum exceeds 1.0, employees in this environment are immediately required to wear ear protection. On a longer-term basis, noise reduction control methods should be developed for the specific pieces of equipment with excessive noise levels.

Estimating Worker Exposures to Toxic Vapors
The best procedure to determine exposures to toxic vapors is to measure the vapor concentrations directly. For design purposes estimates of vapor concentrations are frequently required in enclosed spaces, above open containers, where drums are filled, and in the area of spills.

86

Chapter 3

Industrial Hygiene

C o n c e n t r a t i o n of V o l a t i l e , (Mass/Volurne)

C

Volatile

R a t e Out, Time)

kO vC

E v o l u t i o n R a t e of V o l a t i l e , (Mass/Time)

0

Figure 3-3

Mass balance for volatile vapor in an enclosure.

Consider the enclosed volume shown in Figure 3-3. This enclosure is ventilated by a constant volume airflow. Volatile vapors are evolved within the enclosure. An estimate of the concentration of volatile in the air is required. Let C be the concentration of volatile vapor in the enclosure (masslvolume), V be the volume of the enclosure (volume), Q , be the ventilation rate (volume/time), k be the nonideal mixing factor (unitless), and Q , be the evolution rate of volatile material (massltime). The nonideal mixing factor k accounts for conditions in the enclosure that are less than well mixed. It follows that Total mass of volatile in volume Accumulation of mass of volatile Mass rate of volatile material resulting from evolution Mass rate of volatile material out
=

VC,

= = =

d(VC) dt

dC V?

Q,, kQ,C.

3-3 Industrial Hygiene: Evaluation

87

Because accumulation equals mass in minus mass out, the dynamic mass balance on the volatile species is

At steady state the accumulation term is 0, and Equation 3-6 is solved for C:

Equation 3-7 is converted to the more convenient concentration units of ppm by direct application of thk ideal gas law. Let rn represent mass, p represent density, and the subscripts v and b denote the volatile and bulk gas species, respectively. Then:

where

R, is the ideal gas constant, T is the absolute ambient temperature, P is the absolute pressure, and M is the molecular weight of the volatile species.
The term m,/Vb is identical to the concentration of volatile computed using Equation 3-7. Substituting Equation 3-7 into Equation 3-8 yields

Equation 3-9 is used to determine the average concentration (in ppm) of any volatile species in an enclosure given a source term Q , and a ventilation rate Q,. It can be applied to the following types of exposures: a worker standing near a pool of volatile liquid, a worker standing near an opening to a storage tank, or a worker standing near an open container of volatile liquid. Equation 3-9 includes the following important assumptions: The calculated concentration is an average concentration in the enclosure. Localized conditions could result in significantly higher concentrations; workers directly above an open container might be exposed to higher concentrations. A steady-state condition is assumed; that is, the accumulation term in the mass balance is zero.

Chapter 3

Industrial Hygiene

The nonideal mixing factor varies from 0.1 to 0.5 for most practical situation^.^ For perfect mixing k = 1. Example 3-7
An open toluene container in an enclosure is weighed as a function of time, and it is determined that the average evaporation rate is 0.1 g/min. The ventilation rate is 100 ft"/min. The temperature is 80°F and the pressure is 1atm. Estimate the concentration of toluene vapor in the enclosure, and compare your answer to the TLV for toluene of 50 ppm.

Solution
Because the value of k is not known directly, it must be used as a parameter. From Equation 3-9

From the data provided Qm = 0.1 g/min = 2.20 X lb,/min, R, = 0.7302 ft3 atmllb-mol OR, T = 80°F = 54OoR, Q, = 100 ft"/min, M = 92 lbm/lb-mol, P = 1 atm. Substituting into the equation for kcppm:

=

9.43 ppm.

Because k varies from 0.1 to 0.5, the concentration is expected to vary from 18.9 ppm to 94.3 ppm. Actual vapor sampling is recommended to ensure that the TLV is not exceeded.

Estimating the Vaporization Rate of a Liquid Liquids with high saturation vapor pressures evaporate faster. As a result, the evaporation rate (massltime) is expected to be a function of the saturation vapor pressure. In reality, for vaporization into stagnant air, the vaporization rate is proportional to the difference be-

7R. Craig Matthiessen, "Estimating Chemical Exposure Levels in the Workplace," Chemical Engineering
Progress (April 1986), p. 30.

3-3 Industrial Hygiene: Evaluation

tween the saturation vapor pressure and the partial pressure of the vapor in the stagnant air; that is

where

PSat the saturation vapor pressure of the pure liquid at the temperature of the liquid and is p is the partial pressure of the vapor in the bulk stagnant gas above the liquid.

A more generalized expression for the vaporization rate is available8:

where
Q, is the evaporation rate (massltime),

M is the molecular weight of the volatile substance, K is a mass transfer coefficient (lengthltime) for an area A, R, is the ideal gas constant, and TLis the absolute temperature of the liquid.
For many situations, Pmt >>p, and Equation 3-11 is simplified to

Equation 3-12 is used to estimate the vaporization rate of volatile from an open vessel or from a spill of liquid. The vaporization rate or source term, determined by Equation 3-12, is used in Equation 3-9 to estimate the concentration (in ppm) of a volatile in an enclosure resulting from evaporation of a liquid:

%even R. Hanna and Peter J. Drivas, Guidelines for the Use of Vapor Cloud Dispersion Models (New York: American Institute of Chemical Engineers, 1987).

90

Chapter 3

Industrial Hygiene

For most situations T = TL,and Equation 3-13 is simplified to

The gas mass transfer coefficient is estimated using the relationshipy

where

a is a constant and D is the gas-phase diffusion coefficient.
Equation 3-15 is used to determine the ratio of the mass transfer coefficients between the species of interest K and a reference species KO:

The gas-phase diffusion coefficientsare estimated from the molecular weights M of the species lo:

Equation 3-17 is combined with Equation 3-16, giving

Water is most frequently used as a reference substance; it has a mass transfer coefficientll of 0.83 cmls.

Example 3-8
A large open tank with a 5-ft diameter contains toluene. Estimate the evaporation rate from this tank assuming a temperature of 77°F and a pressure of 1 atm. If the ventilation rate is 3000 ft3/min,
estimate the concentration of toluene in this workplace enclosure. "ouis J. Thibodeaux, Chemodynamics (New York: Wiley, 1979), p. 85. 'OGordon M. Barrow, Physical Chemistry, 2d ed. (New York: McGraw-Hill, 1966), p. 19. llMatthiessen, "Estimating Chemical Exposure," p. 33.

3-3 Industrial Hygiene: Evaluation

Solution
The molecular weight of toluene is 92. The mass transfer coefficient is estimated from Equation 3-18 using water as a reference:

The saturation vapor pressure is given in Example 3-4:

P: ",= 28.2 mm Hg :: ,,
The pool area is

=

0.0371 atm.

The evaporation rate is computed using Equation 3-12:

- (92 lb,/lb-mo1)(0.949 ft/min)(l9.6 ft2)(0.0371atm) (0.7302 ft" atmllb-mol0R)(537"R)

The concentration is estimated using Equation 3-14 with k as a parameter:

-

(0.949 ft/min)(l9.6 ft2)(0.0371 atm) (3000 ft"min)(l atm)

x lo6

=

230 ppm.

The concentration will range from 460 ppm to 2300 ppm, depending on the value of k. Because the TLV for toluene is 50 ppm, additional ventilation is recommended, or the amount of exposed surface area should be reduced. The amount of ventilation required to reduce the worst-case concentration (2300 ppm) to 50 ppm is
Q, = (3000 ft3/min)

ppm (2300ppm )

=

138,000 fr3/min.

This represents an impractical level of general ventilation. Potential solutions to this problem include containing the toluene in a closed vessel or using local ventilation at the vessel opening.

Chapter 3

Industrial Hygiene

Total S o u r c e = E v a p o r a t i o n + Displaced Air
\

Y

Liquid

Filling

Vapor

Evaporation

-

w_ L
-

-

Liquid

-- -

-4

-

Figure 3-4

-

D r u m o r Vessel

Evaporation and displacement from a filling vessel.

Estimating Worker Exposures during Vessel Filling Operations For vessels being filled with liquid, volatile emissions are generated from two sources, as shown in Figure 3-4. These sources are
evaporation of the liquid, represented by Equation 3-14, and displacement of the vapor in the vapor space by the liquid filling the vessel. The net generation of volatile is the sum of the two sources:

where (Q,), represents the source resulting from evaporation and (Q,), represents the source resulting from displacement. The source term (Q,), is computed using Equation 3-12. (Q,), is determined by assuming that the vapor is completely saturated with the volatile. An adjustment is introduced later for less than saturated conditions. Let

V , be the volume of the container (volume), r, be the constant filling rate of the vessel (time-'),

3-3 Industrial Hygiene: Evaluation

93

Pmt be the saturation vapor pressure of the volatile liquid, and TLbe the absolute temperature of the container and liquid.
It follows that rfVcis the volumetric rate of bulk vapor being displaced from the drum (volumeltime). Also, if pv is the density of the volatile vapor, rfVcpv the mass rate of volatile disis placed from the container (massltime). Using the ideal gas law,

and it follows that

Equation 3-21 can be modified for container vapors that are not saturated with the volatile. Let 4 represent this adjustment factor; then,

For splash filling (filling from the top of a container with the liquid splashing to the bottom), 4 = 1. For subsurface filling12(by a dip leg to the bottom of the tank), 4 = 0.5. The net source term resulting from filling is derived by combining Equations 3-12 and 3-22 with Equation 3-19:

This source term is substituted into Equation 3-9 to compute the vapor concentration (in ppm) in an enclosure resulting from a filling operation. The assumption that T = TLis also invoked. The result is

kQVP

(c$rfVc+ K A )

X

10'.

For many practical situations the evaporation term KA is much smaller than the displacement term and can be neglected.
I2Matthiessen, "Estimating Chemical Exposure," p. 33

94

Chapter 3

Industrial Hygiene

Example 3-9
Railroad cars are being splash-filled with toluene. The 10,000-galcars are being filled at the rate of one every 8 hr. The filling hole in the tank car is 4 in. in diameter. Estimate the concentration of toluene vapor as a result of this filling operation. The ventilation rate is estimated at 3000 ftqmin. The temperature is 77°F and the pressure is 1atm.

Solution
The concentration is estimated using Equation 3-24. From Example 3-8, K P"& 0.0371 atm. The area of the filling hole is
=

0.949 ft/min and

Thus

KA
The filling rate r,, is

=

(0.949 ft/min)(0.0872 ft2) = 0.0827 ft3/min.

For splash filling the nonideal filling factor 4 is 1.0. The displacement term in Equation 3-24 is 4 r f V , = (1.0)(0.00208min-')(10,000 gal) (7.4fgal)
=

2.78 ftvmin.

As expected, the evaporation term is small compared to the displacement term. The concentration is computed from Equation 3-24, using k as a parameter:
= -=
=

Psat+rfV,
QvP

(0.0371 atm)(2.78 ft"/min) (3000 ft3/min)(latm)

kc,,,

x

lo6

34.4 ppm.

The actual concentration could range from 69 ppm to 344 ppm, depending on the value of k . Sampling to ensure that the concentration is below 50 ppm is recommended. For subsurface filling, = 0.5, and the concentration range is reduced to 35-172 ppm.

+

3-4 Industrial Hygiene: Control
After potential health hazards a r e identified a n d evaluated, t h e appropriate control techniques must b e developed a n d installed. This requires the application of appropriate technology for reducing workplace exposures.

3-4 Industrial Hygiene: Control

95

Table 3-9

Chemical Plant Control Techniques
Typical techniques

Type and explanation Enclosures Enclose room or equipment and place under negative pressure.

Enclose hazardous operations such as sample points. Seal rooms, sewers, ventilation, and the like. Use analyzers and instruments to observe inside equipment. Shield high-temperature surfaces. Pneumatically convey dusty material. Use properly designed hoods. Use hoods for charging and discharging. Use ventilation at drumming station. Use local exhaust at sample points. Keep exhaust systems under negative pressure. Design locker rooms with good ventilation and special areas or enclosures for contaminated clothing. Design ventilation to isolate operations from rooms and offices. Design filter press rooms with directional ventilation. Clean vessels chemically vs. sandblasting. Use water sprays for cleaning. Clean areas frequently. Use water sprays to shield trenches or pump seals. Use dikes around tanks and pumps. Provide water and steam connections for area washing. Provide lines for flushing and cleaning. Provide well-designed sewer system with emergency containment. Use safety glasses and face shields. Use aprons, arm shields, and space suits. Wear appropriate respirators; airline respirators are required when oxygen concentration is less than 19.5%.

Local ventilation Contain and exhaust hazardous substances.

Dilution ventilation Design ventilation systems to control low-level toxics.

Wet methods Use wet methods to minimize contamination with dusts.

Good housekeeping Keep toxicants and dusts contained.

Personal protection As last line of defense.

The types of control techniques used in the chemical industry are illustrated in Table 3-9. Designing control methods is an important and creative task. During the design process, the designer must pay particular attention to ensure that the newly designed control technique provides the desired control and that the new control technique itself does not create another hazard, sometimes even more hazardous than the original problem. The two major control techniques are environmental controls and personal protection. Environmental control reduces exposure by reducing the concentration of toxicants in the workplace environment. This includes enclosure, local ventilation, dilution ventilation, wet methods,

96

Chapter 3

Industrial Hygiene

Table 3-10

Personal Protective Equipment, Not Including Respirators1
Description
-

Hard hat Safety glasses Chemical splash goggles, gas-tight Steel-toed safety shoes Wraparound face shield Vinyl apron Splash suit Umbilical cord suit Rubber oversleeves PVC-coa ted gloves PVC and nitrile knee boots Ear plugs must always be consulted.

Protects head from falling equipment and bumps Impact-resistant lenses Suitable for liquids and fumes Protects against dropped equipment Fiberglass, resistant to most chemicals Resists most chemicals Viton or butyl rubber for nonflammable exposures Used with external air supply Protects forearms Resists acids and bases Resists acids, oils, and greases Protects against high noise levels

Lab Safety Supply Catalog (Janesvelle,WI:Lab Safety Supply Inc.).Manufacturers' technical specifications

and good housekeeping, as discussed previously. Personal protection prevents or reduces exposure by providing a barrier between the worker and the workplace environment. This barrier is usually worn by the worker, hence the designation "personal." Typical types of personal protective equipment are listed in Table 3-10.

Respirators
Respirators are routinely found in chemical laboratories and plants. Respirators should be used only on a temporary basis, until regular control methods can be implemented; as emergency equipment, to ensure worker safety in the event of an accident; as a last resort, in the event that environmental control techniques are unable to provide satisfactory protection. Respirators always compromise worker ability. A worker with a respirator is unable to perform or respond as well as a worker without one. Various types of respirators are listed in Table 3-11. Respirators can be used improperly and/or can be damaged to the extent that they do not provide the needed protection. OSHA and NIOSH have developed standards for using respirators,l"ncluding fit testing (to ensure that the device does not leak excessively), periodic in-

13NIOSH Respirator Decision Logic, DHHS-NIOSH Publication 87-1-8 (Washington, DC: US Department of Health and Human Services, May 1987).

3-4 Industrial Hygiene: Control

97

Table 3-11
TYPe Mouth and nose dust mask

Respirators Useful to Chemical Industry
Example of commercial brand
MSA Dustfoe@ 88' MSA Comfo Classic cartridge

Limitations
0, > 19.5%; single use; PEL > 0.05 mg/m3 0, > 19.5%; GMA cartridge (black) for concentrations less than the IDLH concentration for organic vapors; GMC cartridge (orange) for concentrations less than the IDLH concentration for Cl,, HC1, and SO, 0, > 19.5%; type N canister for concentrations less than 100 times PEL and less than the IDLH concentration for acid gases, CO, ammonia, and organic vapors; escape concentrations of 2% for acid gases, CO, and organic vapors and 3% for ammonia; escape capacity less than 6 min
Good for toxic and noxious gases with concentrations below and above the IDLH concentration. Capacity between 30 and 60 min per specifications

Mouth and nose with chemical cartridge

Full face mask with chemical canister

MSA Industrial Canister, Gas Mask3

Self-contained breathing apparatus (SCBA)

MSA MMR XtremeB Air Mask4

'MSA Home Page 2000, Air-purifying Respirators, Conventionally Maintained, DustfoeB Respirator (Pittsburgh, PA: MSA International). 2MSA Home Page 2000, Air-purifying Respirators, Conventionally Maintained, Comfo Classic. 'MSA Home Page 2000, Air-purifying Respirators, Conventionally Maintained, Replacement Canisters for Gas Masks. 4MSA Home Page 2000, Supplied Air Respirators, Self-Contained Breathing Apparatus.

spections (to ensure that the equipment works properly), specified use applications (to ensure that the equipment is used for the correct job), training (to ensure that it is used properly), and record keeping (to ensure that the program is operating efficiently). All industrial users of respirators are legally bound to understand and fulfill these OSHA requirements.

Ventilation
For environmental control of airborne toxic material the most common method of choice is ventilation, for the following reasons: Ventilation can quickly remove dangerous concentrations of flammable and toxic materials. Ventilation can be highly localized, reducing the quantity of air moved and the equipment size. Ventilation equipment is readily available and can be easily installed. Ventilation equipment can be added to an existing facility.

Ventilation systems are composed of fans and ducts. The negative pressure system ensures that contaminants do not leak into workplace environments. with the fans located at the exhaust end of the system. These operating costs need to be considered when evaluating alternatives. 2nd Floor --z---- Leakage into Ducts 1st Floor -s-- H o d Intake Negative Pressure Ventilation Figure 3-5 The difference between a positive and a negative pressure ventilation system. . Exhaust 2 n d Floor Leakage O u t of Ducts Blower I st Floor / Hood Intake Positive Pressure Ventilation Exhaust A ------. pulling air out. The best system is a negative pressure system. This ensures that leaks in the system draw air in from the workplace rather than expel contaminated air from the ducts into the workplace. There are two types of ventilation techniques: local and dilution ventilation. and the cost to heat or cool the large quantities of fresh air can be large. Substantial electrical energy may be needed to drive the potentially large fans.98 Chapter 3 Industrial Hygiene The major disadvantage of ventilation is the operating cost. and (2) remove the contaminant before workers are exposed. Ventilation is based on two principles: (1)dilute the contaminant below the target concentration.1 psi) that moves the air. The fans produce a small pressure drop (less than 0. This is shown in Figure 3-5.

Likewise. The airflow profiles within the hood are highly dependent on the location of the window sash. The baffle at the rear of the hood ensures that contaminants are removed from the working surface and the rear lower corner. Dangerous Properties of lndustrial Materials. 74. Irving Sax. A push-pull hood uses a stream of air from a supply to push contaminants toward an exhaust system. A standard laboratory utility hood is shown in Figure 3-6. 4th ed. A receiving hood is an exterior hood that uses the discharge motion of the contaminant for collection. Inc. shown in Figure 3-7.3-4 Industrial Hygiene: Control Local Ventilation The most common example of local ventilation is the hood. There are several types of hoods: An enclosed hood completely contains the source of contaminant. the sash should never be fully opened because contaminants might escape. Fresh air is drawn through the window area of the hood and is removed out the top through a duct. The most common example of an enclosed hood is the laboratory hood. It is important to keep the sash open a few inches. p. minimally. 1975). Airflow patterns and control velocity are dependent on sash height. For this design bypass air is supplied through a grill at the top of the hood. An exterior hood continuously draws contaminants into an exhaust from some distance away. to ensure adequate fresh air. . Another type of laboratory hood is the bypass hood. This ensures the availability of fresh Figure 3-6 Standard utility laboratory hood. Reprinted by permisdon of John Wiley & Sons. A hood is a device that either completely encloses the source of contaminant and/or moves the air in such a fashion as to carry the contaminant to an exhaust device. (New York: Van Nostrand Reinhold. Source: N.

Irving Sax. The advantages of enclosed hoods are that they completely eliminate exposure to workers. provide a containment device in the event of fire or explosion. p. Q. = LWU.100 Chapter 3 Industrial Hygiene Figure 3-7 Standard bypass laboratory hood. The bypass air is controlled by the height of the sash. 4th ed. (New York: Van Nostrand Reinhold. bench-scale or pilot plant equipment. Source: N. Most hood calculations assume plug flow. is computed from For a rectangular duct of width Wand length L. The bypass air supply is reduced as the hood sash is opened. 1975). (3-26) . require minimal airflow. Reprinted by permission of John Wiley & Sons. air to sweep out contaminants in the hood. The disadvantages of hoods are that they limit workspace and can be used only for small. and provide a shield to the worker by means of a sliding door on the hood. Inc. Dangerous Properties of Industrial Materials. 75. the volume of air moved per unit time Q. is determined using the equation Q. For a duct of cross-sectional area A and average air velocity E (distanceltime).

The airflow velocity is a function of the sash height and the blower speed. L = Length W = Width i = Required Control Velocity Figure 3-8 Determining the total volumetric air flow rate for a boxtype hood. and the evolution rate of the contaminant. ~ o l u m e / T i r n e . 19th ed. Instruments are available for measuring the airflow velocity at specific points of the hood window opening. However. . the depth of the hood. Design equations are available for a wide variety of hood and duct shapes. Shallower hoods need higher control velocities to prevent contaminants from exiting the front. Testing is an OSHA requirement. The required control velocity depends on the toxicity of the material. The design strategy is to provide a fixed velocity of air at the opening of the hood. experience has shown that higher velocities can lead to the formation of a turbulent eddy from the bottom of the sash. For general operation a control velocity of between 80 and 120 feet per minute (fpm) is desired. 1986). (Cincinnati: American Conference of Governmental Industrial Hygienists. backflow of contaminated air is possible. This face or control velocity (referring to the face of the hood) ensures that contaminants do not exit from the hood.3-4 Industrial Hygiene: Control 101 Exhaust Q = Volumetric Flow Rote. Consider the simple box-type enclosed hood shown in Figure 3-8. Arrows are frequently used to indicate the proper sash height to ensure a specified face velocity.14 141ndustrial Ventilation: A Manual of Recommended Practice. For general operation a control velocity between 80 and 120 feet per minute (fpm) is suggested.

3-12. Unlike hood ventilation. Dangerous Properties. The ventilation quality is rated as average. Example 3-10 Xylene is used as a solvent in paint. Workers must remain a suitable distance from the source to ensure proper dilution of the contaminant. . The following restrictions should be considered before implementing dilution ventilation: The contaminant must not be highly toxic. where the airflow prevents worker exposure. The total dilution requirement is the sum of the individual dilution requirements..102 Chapter 3 . These methods will most likely expose workers to toxicants. and 3-14 are used to compute the ventilation rates required. The values reported here are the reciprocal of Sax's values. the TLV-TWA. The elephant trunk is simply a flexible vent duct that is positioned near a source of contaminant. 6th ed. dilution ventilation is necessary. For exposures to multiple sources the dilution air requirement is computed for each individual source. Determine the quantity of dilution ventilation air required to maintain the xylene concentration below 100 ppm. but in diluted amounts. Free-hanging canopies and plenums can be either fixed in position or attached to a flexible duct to enable movement. Dilution Ventilation If the contaminant cannot be placed in a hood and must be used in an open area or room. Scrubbing systems must not be required to treat the air before exhaust into the environment. Dilution ventilation always requires more airflow than local ventilation. It is most frequently used for loading and unloading toxic materials from drums and vessels. Table 3-12 lists values for k. The contaminant must be evolved at a uniform rate. operating expenses can be substantial. Equations 3-9. the nonideal mixing factor used with these equations. 29. p. Other types of local ventilation methods include "elephant trunks" and free-hanging canopies and plenums. A certain painting operation evaporates an estimated 3 gal of xylene in an 8-hr shift. Industrial Hygiene Table 3-12 Nonideal Mixing Factor k for Various Dilution Ventilation Conditions1 Mixing factor: Ventilation condition Poor 117 118 1/11 Vapor Dust concentration concentration (PP~) (mppcf) over 500 101-500 0-100 Average 114 115 118 Good 113 114 117 Excellent 112 50 20 5 113 116 'N. dilution ventilation always exposes the worker but in amounts diluted by fresh air. Irving Sax.

125. The specific gravity of the xylene is 0. 1984). 6th ed. Health. compute the air required if the operation is carried out in an enclosed hood with an opening of 50 ft2 and a face velocity of 100 ftlmin. J. 1984). A. and Environmental Protection (Boston: WCBIMcGraw-Hill. 1988). Engineering Design for the Control of Workplace Hazards (New York: McGraw-Hill. N. ed. B. (0.4 Ib. 2d ed. = 13. 1998). For a hood with an open area of 50 ft2. sec. ed. (New York: Van Nostrand Reinhold. Suggested Reading Industrial Hygiene Lewis J. C. 1-3 (New York: Macmillan. A. Fundamentals of Industrial Hygiene.1337 ft" (T) 62.300 ft3/min required dilution air. Irving Sax. Solution The evaporation rate of xylene is Q = m (g ) (&) 0. Cralley and Lester V. 1987). for average ventilation and a vapor concentration of 100 ppm.864) From Table 3-12. eds. and its molecular weight is 106. (Chicago: National Safety Council... 2 and 3. 3d ed. eds. we get The hood requires significantly less airflow than dilution ventilation and prevents worker exposure completely.864. Fundamentals of Industrial Hygiene.: = 118 = 0.. B. Safety.using Equation 3-25 and assuming a required control velocity of 100 fpm. (Chicago: National Safety Council. Richard A. Industrial Hygiene Aspects of Plant Operations.. . Olishifski. k With Equation 3-9. Plog. 1979). Wadden and Peter A. Scheff. The temperature is 77°F and the pressure is 1 atm. Dangerous Properties of Industrial Materials. v. Wentz. Cralley. we solve for Q.Suggested Reading 103 Also.

M. Determine (a) whether the following chemicals are covered under the PSM regulation (29 CFR 1910. 3-5. 3-10. propane. Wadden and Scheff. hydrogen selenide. hydrogen chloride. and ethanol. Review the RMP (40 CFR 68) to determine the conditions that need to be used for dispersion modeling for the worst-case scenarios. and describe the endpoint parameters for consequence analyses for the worst-case scenarios. propane. Engineering Design. 3-12. ethylene oxide. A one-hour lunch break is included between 12 noon and 1P. 1986).119) and (b) their threshold quantities: acrolein. 3-2. 3-9. A worker begins a work shift at 8 A. Determine (a) whether the following chemicals are covered under the PSM regulation and (b) their threshold quantities: ammonia (anhydrous). methyl ethyl ketone. and ethanol. 3-7. Review the details of the RMP (40 CFR 68). where it can be assumed that no exposure to the chemical occurs. phosgene. If they are listed. 3-8. Review the RMP (40 CFR 68) to determine the conditions that need to be used for dispersion modeling for the alternative case scenarios. and describe the endpoint parameters for consequence analyses for the alternative case scenarios. . and methanol. The TLV-TWA for a substance is 150 ppm. phosgene. (c) what are their threshold quantities? The chemicals are ammonia (anhydrous).M. and methanol. A process plant inventories the following chemicals: vinyl chloride. 19th ed. Determine whether the following chemicals (a) are covered under the RMP (40 CFR 68. hydrogen chloride. Determine whether the following chemicals (a) are covered under the RMP and (b) are listed as toxic or flammable. and completes the shift at 5 P. styrene. 3-3. What additional information might you request to perform an appropriate assessment of the risk associated with these chemicals? 3-13. ch. formaldehyde.130) and (b) are listed as toxic or flammable.M. If they are listed. ethylene oxide. and cyclohexane. Determine the hazards associated with these chemicals. 5. methane. In reviewing the results of Problems 3-1 to 3-4. 3-6. (c) what are their threshold quantities? The chemicals are acrolein. (Cincinnati: American Conference of Governmental Industrial Hygienists. describe why the threshold quantities are lower for the PSM-regulated chemicals than for the RMP-regulated chemicals. Problems 3-1.104 Chapter 3 Industrial Hygiene Ventilation Industrial Ventilation: A Manual of Recommended Practice. Describe several typical alternative case scenarios for an RMP study. formaldehyde.. methane. 3-4. 3-11. Review the details of the RMP (40 CFR 68). hydrogen selenide. and describe the three program categories that are used for consequence modeling. ethylene oxide. Review the details of the RMP (40 CFR 68).

The data in the following table were taken in a work area: Time 8:01 A.M. p.M.M.M. 12:08 P. what are the violations? Assume that the worker is at lunch between the hours of 12 noon to 1 P. and a TLV-C of 300 ppm. 1:17 P. Show that this equation is the same as Equation 3-9.M. 10:07 A. 3:09 P.Problems 105 The data were taken in the work area at the times indicated. Compute the mixture TLV. Has the worker exceeded the TLV specification? Time 8:10 A.M. (molecular weight)(TLV)(k) where CFM is the ft3/minof dilution air required. 2:03 P. 2:05 P. Sax. I. 5:00 P.M. A worker on an 8-hour shift is exposed to this toxic vapor.M.M. 1984). 11:20 A. Concentration (PP~) 110 130 143 162 142 157 159 165 153 130 3-14.87 X 108)(lb. Air contains 4 ppm of carbon tetrachloride and 25 ppm of 1. 3-16. 3:13 P. 1:06 P. Sax" provided the following working equation for determining the dilution air requirements resulting from evaporation of a solvent: (3.M. 1212 P. a TLV-STEL of 250 ppm. 28. 6th ed. A substance has a TLV-TWA of 200 ppm. 10:05 A. 9:17 A.M.M. and is not exposed to the chemical during that time. What assumptions are inherent in this equation? CFM = "N. 3-15.M.M. and determine whether this value has been exceeded. (New York: Van Nostrand Reinhold. 4:01 P. Is the exposure within compliance? If not.M.l-dichloroethane. Dangerous Properties of Industrial Materials.M.M.of liquid evaporatedlmin) . .M.M. 11:22 A.M. Concentration (PP~) 5:05 P.M. 9:05 A. 400 P.

If the lid of the drum is left open (lid diameter = 3 ft).8794 10 92. The ventilation quality within the vicinity of this operation is average. c. Estimate the concentration of toluene (in ppm) near the drum if the local ventilation rate is 1000 ft3/min.) Molecular weight Specific gravity TLV ( P P ~ ) 78. Determine the ventilation rate required to maintain the vapor concentration below the TLV. . 3-20. Compute the concentration (in ppm) of the saturated vapor with air above a solution of pure toluene.51 3096. The temperature is 85°F.36 -53. 3-22.The pressure is 1 atm. B.52 -52. Determine the ventilation rate required to maintain the vapor concentration below the TLV. Do the densities of these two gases differ enough to ensure a higher concentration on floors and other low spots? The temperature is 70°F and the pressure is 1 atm. determine the time required to evaporate all the toluene in the drum. The following data are wailable for these materials: Benzene (C6H6) Toluene (C. given by the following: Benzene Toluene 15. The temperature is 80°F.866 50 Saturation vapor pressures: where Pmt is the saturation vapor pressure in mm Hg. T is the temperature in K. A drum with a 2-in-diameter bung is used to contain the mixture. a.Chapter 3 Industrial Hygiene Problems 3-17 through 3-22 apply to toluene and benzene. Determine the evaporation rate per unit area for this mixture. Equations 3-12 and 3-14 represent the evaporation of a pure liquid.9008 16. and the pressure is 1 atm. Benzene and toluene form an ideal liquid mixture. 3-19.0137 2788. A certain plant operation evaporates 2 pintlhr of toluene and 1 pint18-hr shift of benzene. and the pressure is 1 atm. 3-18.67 3-17. and A. Compute the concentration (inppm) of the equilibriumvaporwith air above a solution of 50 mol % toluene and benzene.11 0. A mixture composed of 50 mol % benzene is used in a chemical plant.13 0. Modify these equations to represent the evaporation of a mixture of ideal miscible liquids. The temperature is 80°F.H. 3-21. A drum contains 42 gal of toluene. and C a r e the constants. The temperature is 80°F and the total pressure is 1 atm. Determine the mixture TLV b. Compute the density of pure air and the density of air contaminated with 100 ppm benzene.

6 psi at these conditions. Equations 3-12 and 3-14 can be applied to nonenclosed exposures by using an effective ventilation rate. 3-25. A storage tank of 1000 ft3 capacity must be cleaned before reuse. 3-26. how many gallons are recovered? The molecular weight of gasoline is about 94. Assume an average automobile tank size of 14 gal. and concentrations below 16% can cause distress. Proper procedures must be used to ensure that the oxygen concentration of the air within the tank is adequate. To reduce air pollution. Respiratory equipment without self-contained air supplies must never be used in atmospheres below 19.000gal of delivered gasoline. "Estimating Chemical Exposure. Assume a ventilation rate of 3000 ft3/min.7.Problems 107 3-23.5% oxygen. and its vapor pressure at 77'F is 4. This is accomplished by an elephant trunk ventilation system installed as part of the filler hose. Estimate the concentration (in ppm) of gasoline vapor as a result of this filling operation.5% and (b) 16%. The drums are being splashfilled at the rate of 30 drums per hour. The human body is sensitive to reductions in oxygen concentration. The passageway area is 7 ft2.6 mm Hg. Normal air contains about 21% oxygen by volume. 3-24.6 mm Hg under these conditions.16 A worker is standing near an open passageway of a tank containing 2-butoxyethanol (molecular weight = 118). Oxygen concentrations within tanks and enclosures can be reduced significantly by small amounts of inert elements! I6Matthieson. how many gallons of gasoline are recovered free for the station owner with each fill-up? For 10. The bung opening through which the drums are being filled has an area of 8 cm2.The vapor pressure of 2-butoxyethanol is 0. 3-27. gasoline filling stations are installing scavenger systems to remove the gasoline vapors ejected from the automobile tank during the filling operation. Compute the cubic feet of additional nitrogen at 77OF and 1 atm that will reduce the oxygen concentration within the tank to (a) 19. A 6-in-diameter elephant trunk is used to remove contaminants near the open bung of a drum during a filling operation." p. The molecular weight of gasoline is approximately 94. If the vapor in the tank is saturated with gasoline vapor at a vapor pressure of 4. 33.5% are dangerous. The air velocity required at the end of the elephant trunk is 100 ftlmin. and its liquid specific gravity is 0. Compute the volumetric flow rate of air required.Estimate the ambient vapor concentration if the ventilation rate is 3000 ft3/min. Estimate the concentration (in ppm) of the vapor near the passageway opening. The vapor pressure of the 2-butoxyethanol is 0. . 3-28.6 psi. Fifty-five-gallon drums are being filled with 2-butoxyethanol.The TLV for gasoline is 300 ppm. concentrations below 19. A gasoline tank in a standard automobile contains about 14 gal of gasoline and can be filled in about 3 min. The effective ventilation rate for outside exposures has been estimated at 3000 ft3/min.

4.3 lb per hour. The amount of TCE evaporated within the hood is 5. The TCE will be used in liquid form at room temperature. It is desired to operate the hood of Problem 3-29 so that the vapor concentration in the hood plenum is below the lower explosion limit of 12. The temperature is 70°F and the pressure is 1 atm. This hood will be used for an operation involving trichloroethylene (TCE) (TLVTWA: 50 ppm). and calculate the total air flow rate. Estimate the minimum control velocity required to achieve this objective. A laboratory hood has an opening 4 ft in length by 3 ft in height.5% by volume. The molecular weight of TCE is 131. . 3-30. The hood depth is 18 in. Determine an appropriate control velocity for this hood.- -- Chapter 3 Industrial Hygiene 3-29.

runaway reaction. For flammable releases fire and explosion models convert the source model information on the release into energy hazard potentials. a hole in a tank or pipe. which might include toxic properties and energy content. Environmental impacts could also be considered. the total quantity discharged (or total time of discharge). and explosive materials. More details are provided elsewhere. For a reasonably complex 'Guidelines for Consequence Analysis of Chemical Releases (New York: American Institute of Chemical Engineers. and sheltering or evacuation. but we do not do so here. which tend to reduce the magnitude of potential effects in real incidents. such as thermal radiation and explosion overpressures. 1999). foam systems. Source models are an important part of the consequence modeling procedure shown in Figure 4-1. A dispersion model is subsequently used to describe how the material is transported downwind and dispersed to some concentration levels.Source Models ost accidents in chemical plants result in spills of toxic. Effect models convert these incident-specific results into effects on people (injury or death) and structures. The material has hazardous properties. and the state of the discharge (that is.' Accidents begin with an incident. source models are selected to describe how materials are discharged from the process. M 4-1 Introduction to Source Models Source models are constructed from fundamental or empirical equations representing the physicochemical processes occurring during the release of materials. Additional refinement is provided by mitigation factors. which usually results in the loss of containment of material from the process. or a combination). 109 . flammable. vapor. The source model provides a description of the rate of discharge. Typical incidents might include the rupture or break of a pipeline. Once the incident is known. liquid. such as water sprays. or fire external to the vessel. solid.

Adapted from Guidelines for Consequence Analysis for Chemical Releases (New 'fork: American Institute of Chemical Engineers. of Individuals Affected Property Damage CHAPTER Selection of Fire and Explosion Model TNT Equivalency Multi-Energy Explosion Fireball Baker-Strehlow Others Results may Include: Blast Overpressure Radiant Heat Flux CHAPTER 2 C Mitigation Factors: Escape Emergency Response Shelter in Place Containment Dikes Other I 1 Consequence Model Figure 4-1 Consequence analysis procedure. . Dose Probit Model Others Results may Include: Toxic Response No. 1999).Chapter 4 Source Models Selection of a Release Incident Rupture or Break in Pipeline Hole in a Tank or Pipeline Runaway Reaction Fire External to Vessel Other Selection of Source Model to Describe Release Incident Results may Include: Total Quantity Released (or Release Duration) Release Rate Material Phase 1 CHAPTER 4 C Selection of Dispersion Model (if applicable) Neutrally Buoyant Heavier than Air Others Results may Include: Downwind Concentration Area Affected Duration CHAPTER 5 I C Selection of Effect Model Response vs.

If uncertainty exists. designed to prevent the overpressuring of tanks and process vessels. For gases or vapors stored in a tank. For liquids a leak below the liquid level in the tank results in a stream of escaping liquid. the parameters should be selected to maximize the release rate and quantity. Release mechanisms are classified into wide and limited aperture releases. and severed or ruptured pipes. This ensures that a design is on the safe side. releasing a substantial amount of material in a short time. Small liquid droplets or aerosols might also form from the flashing stream.4-1 Introduction to Source Models 111 Crack Valve (Body a n d Seals) - Severed o r Ruptured Pipe Pump (Body and Seals) Pipe Connection Figure 4-2 Various types of limited aperture releases. plant many source models are needed to describe the release. For these releases material is ejected from holes and cracks in tanks and pipes. If the liquid is stored under pressure above its atmospheric boiling point. the assumption of constant upstream pressure is frequently valid. Limited aperture releases are conceptualized in Figure 4-2. Relief systems. with the possibility of transport away from the leak by wind currents. For the limited aperture case material is released at a slow enough rate that upstream conditions are not immediately affected. and pumps. In the wide aperture case a large hole develops in the process unit. a leak results in a jet of gas or vapor. leaks in flanges. A leak in the vapor space above the liquid can result in either a vapor stream or a two-phase stream composed of vapor and liquid. Figure 4-3 shows how the physical state of the material affects the release mechanism. depending on the physical properties of the material. Some development and modification of the original models is normally required to fit the specific situation. valves. An excellent example is the overpressuring and explosion of a storage tank. are also potential sources of released material. Frequently the results are only estimates because the physical properties of the materials are not adequately characterized or because the physical processes themselves are not completely understood. . a leak below the liquid level will result in a stream of liquid flashing partially into vapor.

and liquid pool evaporation or boiling. flow of vapor through holes. -Jd!fkL Liquid o r Liquid Flashing into V a p o r Liquid Figure 4-3 Vapor and liquid are ejected from process units in either single. Other source models. There are several basic source models that are used repeatedly and will be developed in detail here. These source models are flow of liquid through a hole.Chapter 4 Source Models Gas Vapor or Two-Phase Vapor/Liquid \ .or two-phase states. specific to certain materials. flashing liquids. flow of liquid through a hole in a tank. 4-2 Flow of Liquid through a Hole A mechanical energy balance describes the various energy forms associated with flowing fluids: . flow of liquids through pipes. are introduced in subsequent chapters. flow of gases through pipes.

and Consider a process unit that develops a small hole. and a + 1. The change in elevation of the fluid during the discharge through the hole is also negligible. a = 1. g. p is the fluid density (mass/volume). and m is the mass flow rate (massltime). For this limited aperture release.. resulting in a reduced velocity. The A function represents the final minus the initial state. within the process unit. The frictional losses in the leak are approximated by a constant discharge coefficient C. and the velocity of the fluid within the process unit is assumed negligible. The shaft work is zero. z is the height above datum (length). F is the net frictional loss term (length forcelmass).0 for turbulent flow. W. For incompressible liquids the density is constant.5 for laminar flow. a is the unitless velocity profile correction factor with the following values: a = 0. so Az = 0. Frictional forces between the moving liquid and the wall of the leak convert some of the kinetic energy of the liquid into thermal energy...the average discharge velocity from the leak: i A new discharge coefficient C. u is the average instantaneous velocity of the fluid (lengthltime). is defined as .- - - - 4-2 Flow of Liquid through a Hole 113 where P is the pressure (forcelarea). is the shaft work (force length).0 for plug flow. so A P = P. assume a constant gauge pressure P. defined as The modifications are substituted into the mechanical energy balance (Equation 4-1) to determine i . The external pressure is atmospheric. as shown in Figure 4-4. The pressure of the liquid contained within the process unit is converted to kinetic energy as the fluid escapes through the leak. g is the acceleration due to gravity (lengthltime2).. is the gravitational constant (length masslforce time2).

Loss Prevention in the Process Industries.= 0 A = Leak Area p = Liquid Density Figure 4-4 Liquid escaping through a hole in a process unit. . with some frictional flow losses in the hole.0 - P = 1 atm u2= U Az = 0 W. (London: Butterworths. The discharge coefficient C. resulting from a hole of area A is given by The total mass of liquid spilled depends on the total time that the leak is active. 1996). p. 2d ed. 1517.114 Chapter 4 Source Models Liquid Pressurized within Process Unit External Surroundings P = Pg U1 . The resulting equation for the velocity of fluid exiting the leak is The mass flow rate Q . Lees. is a complicated function of the Reynolds number of the Auid escaping through the leak and the diameter of the hole. The following guidelines are suggested: 2Frank P. The energy of the liquid resulting from its pressure in the vessel is converted to kinetic energy.

The specific gravity of benzene is 0. More details on discharge coefficients for these types of liquid discharges are provided elsewhere. the discharge coefficient is approximately 0.. = 1090 gal.8794.. 10-16.and 2:30 P. A discharge coefficient of 0. The area of the hole is The density of the benzene is The leak mass flow rate is given by Equation 4-7. 7th ed.pp.- The total quantity of benzene spilled is (1. For these conditions the exit velocity of the fluid is independent of the size of the hole. Estimate the total amount of benzene spilled.- - - 4-2 Flow of Liquid through a Hole For sharp-edged orifices and for Reynolds numbers greater than 30.61. (New York: McGrawHill. At 2:30 P. For a well-rounded nozzle the discharge coefficient approaches 1.M.M. When the discharge coefficient is unknown or uncertain.0 to maximize the computed flows. The pressure is immediately restored to 100 psig.48 lbm/s)(90min)(60 slmin) = 7990 lb.000. use a value of 1.a total of 90 minutes. C.81. For short sections of pipe attached to a vessel (with a length-diameter ratio not less than 3). a 114-in-diameter leak is found in the pipeline and immediately repaired. approaches the value 0. Perry's Chemical Engineers Handbook. 1997).' Example 4-1 At 1P. Solution The drop in pressure observed at 1 P. Green. . Perry and Don W. 'Robert H.is indicative of a leak in the pipeline.61 is assumed for this orifice-type leak: e m = AC. The leak is assumed to be active between 1 P.M.M.M. the plant operator notices a drop in pressure in a pipeline transporting benzene.

A new discharge coefficient C. is zero.116 Chapter 4 Source Models Process Vessel . A hole develops at a height hLbelow the fluid level.. and the external gauge pressure is atmospheric. the average instantaneous discharge velocity from the leak: where hL is the liquid height above the leak. A dimensionless discharge coefficient C. or 0. The shaft work W. = O Liquid D e n s i t y A = L e a k Cross Sectional Area atm Figure 4-5 An orifice-type leak in a process vessel. is defined as The mechanical energy balance (Equation 4-1) is solved for E. 4-3 Flow of Liquid through a Hole in a Tank A storage tank is shown in Figure 4-5. Some energy is lost because of frictional fluid flow. is defined as . The flow of liquid through this hole is represented by the mechanical energy balance (Equation 4-1) and the incompressible assumption. and the velocity of the fluid in the tank is zero.. as shown in Equation 4-2. The gauge pressure on the tank is P. The energy resulting from the pressure of the fluid height above the leak is converted to kinetic energy as the fluid exits through the hole.

resulting from a hole of area A is given by As the tank empties. This would occur if the vessel was padded with an inert gas to prevent explosion or was vented to the atmosphere. the total mass of liquid in the tank above the leak is The rate of change of mass within the tank is where Q.: This equation is integrated to .4-3 Flow of Liquid through a Hole in a Tank 117 The resulting equation for the instantaneous velocity of fluid exiting the leak is The instantaneous mass flow rate Q. we can obtain a differential equation representing the change in the fluid height: Equation 4-15 is rearranged and integrated from an initial height hO. Assume that the gauge pressure P. on the surface of the liquid is constant. For a tank of constant cross-sectional area A. By substituting Equations 4-12 and 4-13 into Equation 4-14 and by assuming constant tank cross-section and liquid density.to any height h. is given by Equation 4-12. the liquid height decreases and the velocity and mass flow rate decrease..

The specific gravity of benzene at these conditions is 0. A 1-in puncture occurs in the tank 5 ft off the ground because of the careless driving of a forklift truck.. Estimate (a) the gallons of benzene spilled.118 Chapter 4 Source Models Solving for h. the liquid level height in the tank. Solution The density of the benzene is . = 0: If the vessel is at atmospheric pressure. The tank is padded with nitrogen to a constant regulated pressure of 1 atm gauge to prevent explosion. (b) the time required for the benzene to leak out. The time t.8794. P. yields Equation 4-18 is substituted into Equation 4-12 to obtain the mass discharge rate at any time t: The first term on the right-hand side of Equation 4-19 is the initial mass discharge rate at hL = hL. and (c) the maximum mass flow rate of benzene through the leak. for the vessel to empty to the level of the leak is found by solving Equation 4-18 for t after setting h. The liquid level within the tank is presently at 17 ft. = 0 and Equation 4-20 reduces to Example 4-2 A cylindrical tank 20 ft high and 8 ft in diameter is used to store benzene.

17 ft-lbm/lbf-s2)(2. Equation 4-19 is used to compute the mass flow rate: . I This appears to be more than adequate time to stop the leak or to invoke an emergency procedure to reduce the impact of the leak./ft2) 54.4 min.22ft2/s2)= 3386 s 1 - v'Zis2 = 56. This is the total benzene that will leak out.17 ftls2)(12ft) = (469 s2/ft)(7. However. The maximum discharge occurs at t = 0 at a liquid level of 17.2 ft2)(17ft - 5 ft)(7.4-3 Flow of Liquid through a Hole in a Tank 119 The area of the tank is The area of the leak is The gauge pressure is a. The length of time for the benzene to leak out is given by Equation 4-20: (2)(32.9 lbm/ft3 + (2)(32. c. b.0 ft. the maximum discharge occurs when the hole is first opened. The volume of benzene above the leak is V = A.48 gal/ft3) = 4506 gal.12x 10"lb.hE = (50.

Assume that the head space above the liquid is at atmospheric pressure. we get By rearranging and integrating. .. then combining Equations 4-12 and 4-14. we have By substituting into Equation 4-24. we obtain which results in the general equation for the draining time for any vessel: Equation 4-24 does not assume that the hole is at the bottom of the vessel. which is the same result as Equation 4-21. we obtain If the hole is at the bottom of the vessel. For a vessel with the shape of a vertical cylinder. then Equation 4-26 is integrated from h = 0 to h Equation 4-26 then provides the emptying time for the vessel: = h.120 Chapter 4 Source Models A general equation to represent the draining time for any vessel of any geometry is developed as follows.

The frictional flow losses between the fluid and the pipe wall result in a pressure drop across the pipe length.4-4 Flow of Liquids through Pipes 121 Figure 4-6 Liquid flowing through a pipe. fittings such as valves. For each frictional device a loss term of the following form is used: where K. and pipe entrances and exits. . The net result is The frictional loss term F i n Equation 4-28 represents the loss of mechanical energy resulting from friction and includes losses resulting from flow through lengths of pipe. Kinetic energy changes are frequently negligible. Flow of incompressible liquids through pipes is described by the mechanical energy balance (Equation 4-1) combined with the incompressible fluid assumption (Equation 4-2). Frictional forces between the liquid and the wall of the pipe convert kinetic energy into thermal energy. A pressure gradient across the pipe is the driving force for the movement of liquid. 4-4 Flow of Liquids through Pipes A pipe transporting liquid is shown in Figure 4-6. orifices. is the excess head loss due to the pipe or pipe fitting (dimensionless) and u is the fluid velocity (lengthltime). elbows. This results in a decrease in the liquid velocity and a decrease in the liquid pressure.

p. 22. The Fanning friction factor f is a function of the Reynolds number Re and the roughness of the pipe E . L is the flow path length (length). Engineering Flow and Heat Exchange (New York: Plenum Press. is . as a parameter. Table 4-1 provides values of E for various types of clean pipe. E (mm) For fluids flowing through pipes the excess head loss term Kf given by is where f is the Fanning friction factor (unitless). 1984).122 Chapter 4 Source Models Table 4-1 Roughness Factor E for Clean Pipes1 Pipe material Riveted steel Concrete Cast iron Galvanized iron Commercial steel Wrought iron Drawn tubing Glass Plastic 'Selected from Octave Levenspiel. &Id. useful for determining the Reynolds number from the friction factor f . For laminar flow the Fanning friction factor is given by f =- 16 Re' For turbulent flow the data shown in Figure 4-7 are represented by the Colebrook equation: An alternative form of Equation 4-32. Figure 4-7 is a plot of the Fanning friction factor versus Reynolds number with the pipe roughness. and d is the flow path diameter (length).

20. Source: Octave Levenspiel. Engineering Flow and Heat Exchange (New York: Plenum Press. as shown by the nearly constant friction factors at high Reynolds number in Figure 4-7. f is independent of the Reynolds number.4-4 Flow of Liquids through Pipes 123 Figure 4-7 Plot of Fanning friction factor fversus Reynolds number. p. Reprinted by permission. 1984). For fully developed turbulent flow in rough pipes. For this case Equation 4-33 is simplified to .

Hooper. 1988)..0452 log A ) (3. Hooper. The 2-K method defines the excess head loss in terms of two constants. This equation is -1 . E = 0 and Equation 4-32 reduces to .000 the following Blasius approximation to Equation 4-35 is useful: A single equation has been proposed by Chen4 to provide the friction factor f over the entire range of Reynolds numbers shown in Figure 4-7. 6 ~B. K. pp. and outlets. (Aug. Chemical Engineering. 96-100. 24.= 4 log- 1 ~e fl 1. pp. The problem with this method is that the specified length is coupled to the friction factor.7065 Re where 2-K Method For pipe fittings. Industrial Engineering and Chemistry Fundamentals (1979). Chen. B.255 fl For smooth pipe with a Reynolds number less than 100. 89-92. is the excess head loss (dimensionless). Chemical Engineering. (Nov. 5W. and K. . the Reynolds number and the pipe internal diameter: where K. .1981). 7. 18: 296.. are constants (dimensionless). and other flow obstructions the traditional method has been to use an equivalent pipe length Le.-410g - fl ~ i d 5. valves. inlets. An improved approach is to use the 2-K which uses the actual flow path length in Equation 4-30 -equivalent lengths are not used and provides a more detailed approach for pipe fittings. in Equation 4-30. H.Chapter 4 Source Models For smooth pipes. 4N.

25 0.. P = 0.00 2.15 0. p. threaded Standard. Hooper.60 0.00 2.30 0.20 1. 97. Re is the Reynolds number (dimensionless). P = 1. flangedlwelded Stub-in branch Threaded Flangedlwelded Stub-in branch Full line size.20 0.40 0. all types KI 800 800 800 1000 800 800 800 800 500 500 500 500 1000 1000 1000 Km 0.25 0.35 0.30 Tees Used as elbows Standard.9 Reduced trim.00 0. threaded Standard (rlD = I).50 Run-through Valves Gate.0 Reduced trim. 1weld (45") Mitered..5") Standard (rlD = I). threaded Long radius. and ID. Chemical Engineering.25 0. flangedlwelded Long radius (rlD = IS).5") 5 welds (18") Standard (rlD = I).25 10. threaded Standard (rlD = I).35 0.15 0.1981). all types Mitered (rlD = 1.4-4 Flow of Liquids through Pipes 125 Table 4-2 Fittings Elbows 90" 2-K Constants for Loss Coefficients in Fittings and Valves1 Description of fitting Standard (rlD = I). 24. ball or plug Globe Diaphragm Butterfly Check lWilliam B.27 0.8 Standard Angle or Y-type Dam type Lift Swing Tilting disk 1500 1000 1000 800 2000 1500 1000 4. P = 0.5) Mitered. Table 4-2 contains a list of K values for use in Equation 4-38 for various types of fittings and ..5): 1 weld (90") 2 welds (45") 3 welds (30") 4 welds (22. (Aug. is the internal diameter of the flow path (inches)..15 0.50 0. flangedlwelded Long radius (rlD = IS). 2 welds (22. all types Long radius (rlD = 1.0 1.

For Reynolds numbers greater than 10. diameter. including entrances. exits. pipe lengths. (Nov. and from Equation 4-40. 96-100 8W. pp. Determine the initial fluid velocity at point 1. 2.. K. etc. work input or output to the fluid resulting from pumps. Re > 10.0 for the exit. K. B. and type of pipe. Given: the length. provided by Equations 4-30. = 1. For high Reynolds numbers (that is. Specify the initial point (point 1) and the final point (point 2). so no additional kinetic energy terms in the mechanical energy balance must be considered. For a simple hole in a tank with no pipe connections or fittings the friction is caused only by the entrance and exit effects of the hole. Re < 50) the first term dominates and Kf = K.5 for the entrance and Kf = 1. 89-92.000) the first term in Equation 4-39 is negligible and Kf = K. This must be done carefully because the individual terms in Equation 4-28 are highly dependent on this specification. The solution procedure to determine the mass flow rate of discharged material from a piping system is as follows: 1. turbines. Chemical Engineering.0 for a Bordatype entrance.1988). number and type of fittings in the pipe.IRe. and 4-39. Thus 2 K. = 1.50 for a normal entrance and K.1981). including density and viscosity. properties of the fluid. .5.63.126 Chapter 4 Source Models For pipe entrances and exits Equation 4-38 is modified to account for the change in kinetic energy: For pipe entrances. 7.8 The 2-K method also represents liquid discharge through holes. The K factors for the entrance and exit effects account for the changes in kinetic energy through these piping changes. B. 24. 3. Hooper. For low Reynolds numbers (that is. From the 2-K method an expression for the discharge coefficient for liquid discharge through a hole can be determined. Hooper..0. Determine the pressures and elevations at points 1and 2. Chemical Engineering. For pipe exits. = 0 and K. Equations are also available for orifices7 and for changes in pipe sizes. = 0. The result is where C Kf is the sum of all excess head loss terms. 7W. which nearly matches the suggested value of 0. = 160 and K. = 1. (Aug. = 0. pressures and elevation changes across the piping system.000. and fittings.61. pp. Kf = 0. C.4-38.

If the liquid level in the tank is 5. leaving the velocity at point 2 as a variable. Compute values for all the terms in Equation 4-28. Determine the friction factor for the pipe using Equations 4-31 through 4-37. 5. Determine the excess head loss terms for the pipe (using Equation 4-30). Assuming negligible kinetic energy changks. 6. 8. compute the flow rate of material escaping from the pipe. If the sum of all the terms in Equation 4-28 is zero. If fully developed turbulent flow is expected. no pressure changes. and compute the net frictional loss term using Equation 4-29. The entire pipe assembly is mostly horizontal. and no shaft work.- -I Draining geometry for example 4-3. Substitute the known terms into Equation 4-28. 7. for the fittings (using Equation 4-38). and the pipe is accidently severed 33 m from the tank. Solution The draining operation is shown in Figure 4-8. If fully developed turbulent flow is expected. then this is not required. If not. the mechanical energy balance (Equation 4-28) applied between points 1and 2 reduces to Gate Valve ID C o m m e r c i a l Steel Pipe Figure 4-8 33 m . Use the velocity at point 2. Sum the head loss terms. . and substitute into the equation. Solve for the velocity directly. the solution is direct. Guess a value for the velocity at point 2.- 4-4 Flow of Liquids through Pipes 4. go back to step 4 and repeat the calculation. Example 4-3 Water contaminated with small amounts of hazardous waste is gravity-drained out of a large storage tank through a straight commercial steel pipe. The pipe is 100m long with a gate valve near the tank. then the computation is completed. Determine the mass flow rate using the equation m = pGA. and for any entrance and exit effects (using Equation 4-39).8 m above the pipe outlet. 100mm ID (internal diameter).

128 Chapter 4 Source Models For water The K factors for the entrance and exit effects are determined using Equation 4-39. Thus 2 Kf = 1320f + 1. The gravitational term in the mechanical energy equation is given by . For the pipe exit. and the K factor for the pipe length is given by Equation 4-30. The K factor for the gate valve is found in Table 4-2. and it follows that = (660f + 0. Summing the K factors gives For Re > 10. For the gate valve.60. For the pipe entrance.000 the first term in the equation is small.80)Z2. For the pipe length.

the solution is found by trial and error. The table also shows that the friction factor f changes little with the Reynolds number. Equation 4-34 produces a friction factor value of 0.66 mls. from Table 4-1. . The trial and error solution is shown in the following table: Guessed ii (m 1s) Re f F Calculated ii (m 1s) Thus the velocity of the liquid discharging from the pipe is 3.0041.. For commercial steel pipe. Thus we can approximate it using Equation 4-34 for fully developed turbulent flow in rough pipes. the mechanical energy balance (Equation 2-28) reduces to Solving for the exit velocity and substituting for the height change gives The Reynolds number is given by Re=-= dip P (0. Then By substituting and solving. we obtain This result is close to the more exact trial and error solution.0 x 10-"glm s = 1.4-4 Flow of Liquids through Pipes Because there is no pressure change and no pump or shaft work. E = 0.0 x lo5.1 m)(G)(1000 kg/m3) 1.0046 mrn and Because the friction factor f and the frictional loss term Fare functions of the Reynolds number and velocity.

pressure. For free expansion releases most of the pressure energy is converted to kinetic energy.130 Chapter 4 Source Models The-cross-sectionalarea of the pipe is The mass flow rate is given by This represents a significant flow rate. For throttling releases the gas issues through a small crack with large frictional losses. In addition to the material released by the flow. Assuming a 15-min emergency response period to stop the release. a total of 26. For flowing gases and vapors these assumptions are valid only for small pressure changes (P11P2< 2) and low velocities (<0. they are not considered here. The mechanical energy balance (Equation 4-1) describes the flow of compressible gases and vapors. the assumption of isentropic behavior is usually valid. The density. This could include a reduction in the emergency response period. Source models for throttling releases require detailed information on the physical structure of the leak. replacement of the pipe by one with a smaller diameter. the liquid contained within the pipe between the valve and the rupture will also spill. A free expansion leak is shown in Figure 4-9. Free expansion release source models require only the diameter of the leak. or modification of the piping system to include additional control valves to stop the flow. 4-5 Flow of Vapor through Holes For flowing liquids the kinetic energy changes are frequently negligible and the physical properties (particularly the density) are constant. Gas and vapor discharges are classified into throttling and free expansion releases. and temperature change as the gas or vapor exits through the leak.3 times the speed of sound in gas). little of the energy inherent to the gas pressure is converted to kinetic energy. Assuming negligible potential energy changes and no shaft work results in a reduced form of the mechanical energy balance describing compressible flow through holes: . Energy contained within the gas or vapor as a result of its pressure is converted into kinetic energy as the gas or vapor escapes and expands through the hole.000kg of hazardous waste will be spilled. An alternative system must be designed to limit the release.

(4-44) py where y is the ratio of the heat capacities. The integration is carried to any arbitrary final point (denoted without a subscript). y = C..IC.is defined in a similiar fashion to the coefficient defined in section 4-2: Equation 4-42 is combined with Equation 4-41 and integrated between any two convenient points. Substituting Equation 4-44 into Equation 4-43.4-5 Flow of Vapor through Holes 131 Gas Pressurized Within Process Unit External Surroundings At Throat: Figure 4-9 A free expansion gas leak. The gas properties (P. An initial point (denoted by subscript "0") is selected where the velocity is zero and the pressure is Po. The result is For any ideal gas undergoing an isentropic expansion. T) and velocity change during the expansion. pvy = -= P constant.. and inte- . The gas expands isentropically through the hole. A discharge coefficient C. identical to that in Equation 4-5. defining a new discharge coefficient C.

the following statements are valid: (1) The velocity of the fluid at the throat of the leak is the velocity of sound at the prevailing conditions. and Tois the temperature of the source. The result is solved for the pressure ratio resulting in the maximum flow: The choked pressure Pchoked the maximum downstream pressure resulting in maximum flow is through the hole or pipe... and setting the derivative equal to zero. is the ideal gas constant. For many safety studies the maximum flow rate of vapor through the hole is required. Thus: . Using the continuity equation and the ideal gas law for isentropic expansions in the form results in an expression for the mass flow rate: Equation 4-48 describes the mass flow rate at any point during the isentropic expansion.. and (2) the velocity and mass flow rate cannot be increased further by reducing the downstream pressure. R. or sonicjlow and is illustrated in Figure 4-10. they are independent of the downstream conditions. For downstream pressures less than PC.. An interesting feature of Equation 4-49 is that for ideal gases the choked pressure is a function only of the heat capacity ratio y. critical.132 Chapter 4 Source Models grating results in an equation representing the velocity of the fluid at any point during the isentropic expansion: The second form incorporates the ideal gas law for the initial density p. This type of flow is called choked.. This is determined by differentiating Equation 4-48 with respect to PIP.

67 ~1.710.528P0 0. the flow will be choked and maximized through the leak. The mass flow rate is independent of the downstream pressure.487P0 0.7 psia). and R.1 psig.4-5 Flow of Vapor through Holes 133 Gas Pressurized Within Unit External Surroundi Process < 'choked At Throat: 'choked - U r Sonic Velocity Figure 4-10 Choked flow of gas through a hole. is the ideal gas constant. The gas velocity is sonic at the throat. Conditions leading to choked flow are common in the process industries. The maximum flow is determined by substituting Equation 4-49 into Equation 4-48: where M is the molecular weight of the escapilig vapor or gas. To is the temperature of the source. .542P0 For an air leak to atmospheric conditions (PChoked = 14. if the upstream pressure is greater than 14.40 -1.32 Pchoked 0.528 = 27. Gas Monotonic Diatomic and air Triatomic Y -1. or 13.8 psia.

Chemical Engineers Handbook. is uncertain. Gas Acetylene Air Ammonia Argon Butane Carbon dioxide Carbon monoxide Chlorine Ethane Ethylene Helium Hydrogen chloride Hydrogen Hydrogen sulfide Methane Methyl chloride Natural gas Nitric oxide Nitrogen Nitrous oxide Oxygen Propane Propene (propylene) Sulfur dioxide -- - - - - - lCrane Co. of 0. 7th ed. For sharp-edged orifices with Reynolds numbers greater than 30.0 is recommended.. Chilton. 10-16. a constant discharge coefficient C.IC.61 is indicated.000 (and not choked).^ For these flows and for situations where C. 1986). Values for the heat capacity ratio y for a variety of gases are provided in Table 4-3. .. Fzttzngs.. However. Flow of Flutds Through Valves. 1997). pp.134 Chapter 4 Source Models Table 4-3 Heat Capacity Ratios y for Selected Gases1 Chemical formula or symbol Approximate molecular weight (M) Heat capacity ratio y = C. 9Robert H. Technical Paper 410 (New York: Crane Co. and Pipes. Perry and Cecil H. for choked flows the discharge coefficient increases as the downstream pressure decrease^. (New York: McGraw-Hill. a conservative value of 1.

using Equation 4-50.7) psia = 113.. Po = 200 + 14.1 psia will result in choked flow through the leak.. Then.1 psia An external pressure less than 113.0. Because the external pressure is atmospheric in this case. choked flow is expected and Equation 4-50 applies. is assumed to be 1.1-in hole forms in a tank containing nitrogen at 200 psig and 80°F. for nitrogen y = 1.7 = 214. T o = 80 + 460 = 540°R.. Then from Equation 4-49 Thus PCh. Solution From Table 4-3. Also. . Determine the mass flow rate through this leak.7 psia.4-5 Flow of Vapor through Holes 135 Example 4-4 A 0. = 0. The area of the hole is The discharge coefficient C.41.527(200 + 14.

it expands because of a decrease in pressure. Unfortunately. For this particular case the outlet velocity is less than the sonic velocity.136 Chapter 4 Source Models 4-6 Flow of Gases through Pipes Vapor flow through pipes is modeled using two special cases: adiabatic and isothermal behavior. For this case it is more conveniently written in the form . an underwater pipeline is an excellent example. The flow is driven by a pressure gradient across the pipe. The velocity of sound is determined using the thermodynamic relationship which for an ideal gas is equivalent to which demonstrates that for ideal gases the sonic velocity is a function of temperature only. This expansion leads to an increase in velocity and an increase in the kinetic energy of the gas. frictional forces are present between the gas and the pipe wall. For air at 20°C the velocity of sound is 344 m/s (1129 ftls). either an increase or a decrease in the gas temperature is possible. These frictional forces increase the temperature of the gas. For both the isothermal and adiabatic cases it is convenient to define a Mach (Ma) number as the ratio of the gas velocity to the velocity of sound in the gas at the prevailing conditions: where a is the velocity of sound. a decrease in temperature occurs. The mechanical energy balance (Equation 4-1) also applies to adiabatic flows. The adiabatic case corresponds to rapid vapor flow through an insulated pipe. The kinetic energy is extracted from the thermal energy of the gas. The isothermal case corresponds to flow through an uninsulated pipe maintained at a constant temperature. the real case is difficult to model and no generalized and useful equations are available. However. Adiabatic Flows An adiabatic pipe containing a flowing vapor is shown in Figure 4-11. As the gas flows through the pipe. Real vapor flows behave somewhere between the adiabatic and isothermal cases. Depending on the magnitude of the kinetic and frictional energy terms.

> Sonic Velocity Ma. Ma I . > 'choked Figure 4-1 1 Adiabatic nonchoked flow of gas through a pipe.4-6 Flow of Gases through Pipes 137 '2 > 'choked ---------- U2> U. age 6WS m gc . Equations 4-29 and 4-30 can be combined and then differentiated to result in Because no mechanical linkages are present. An important part of the frictional loss term is the assumption of a constant Fanning friction factor f across the length of the pipe. This assumption is valid only at high Reynolds numbers. depending on the magnitude of the frictional losses. Ma. < 1 For Surroundings. Assuming a straight pipe without any valves or fittings. For this open steady flow process the total energy balance is given by iidii g dh+-+-dz=6q--. A total energy balance is useful for describing the temperature changes within the flowing gas. The gas temperature might increase or decrease. The following assumptions are valid for this case: is valid for gases. 1 -I P = [ U2 < Ma.

p. Engineering Flow and Heat Exchange (New York: Plenum Press. 43. 1986). where Y.1 Maf./E/I \ = Ma2P2 YgcM RgT2' where G is the mass flux with units of mass/(area time) and (4-60) kinetic energy compressibility friction Equation 4-60 relates the Mach numbers to the frictional losses in the pipe. glg. integrated (between the initial point denoted by subscript "0" and any arbitrary final point).. The equations are combined. The following assumptions are invoked: dh = C. after considerable effort. PI Ma2 = G pii = MalPl . using Equations 4-56 through 4-58: "'Octave Levenspiel. Equations 4-59 and 4-60 are converted to a more convenient and useful form by replacing the Mach numbers with temperatures and pressures.138 Chapter 4 Source Models where h is the enthalpy of the gas and q is the heat. These assumptions are applied to Equations 4-55 and 4-54. Y 2 Ma2 Y . . and manipulated to yield. The various energy contributions are identified.1° T2 -- TI -- 3. The compressibility term accounts for the change in velocity resulting from the expansion of the gas. 6Ws = 0 because no mechanical linkages are present. d T for an ideal gas. 6q = 0 because the pipe is adiabatic. dz = 0 is valid for gases. = 1 + 2 P.

independent of the downstream pressure. This assumption can be checked later but is normally valid. P < PChoked * Q=O /\ TI --------------- 7/7\ ' " 1 II Ii \\ * - - 3 Choked 12 --------------. For choked flow Equations 4-56 through 4-60 are simplified by setting Ma2 = 1.. Compute the total mass flux G from Equation 4-62. The gas exiting the to pipe makes an abrupt change from Pchoked the ambient pressure. The pressure at the end of the pipe will remain at PC. the gas flow is called choked. When the sonic velocity is reached. Compute &Id. This case is shown in Figure 4-12. Determine the Fanning friction factor f from Equation 4-34. 3. The gas reaches the sonic velocity at the end of the pipe. To compute the mass flux G. and downstream pressure (P2)are known.\ \ t u2 = Sonic Velocity Ma2 = 1 Ma1 Sonic Velocity Reached at Exit of Pipe / f Figure 4-12 Adiabatic choked flow of gas through a pipe. The results are Tchoked -2Y1 - TI y + 1' y+l' Pchoked -Ma. the procedure is as follows: 1. upstream temperature ( T I )and pressure (P.. 4. This assumes fully developed turbulent flow at high Reynolds numbers. the gas velocity at the end of the pipe remains constant at the sonic velocity. inside diameter ( d ) . The maximum velocity is reached at the end of the pipe. even if this pressure is greater than the ambient pressure. For most problems the pipe length (L). 2.0. For long pipes or for large pressure differences across the pipe the velocity of the gas can approach the sonic velocity.- 4-6 Flow of Gases through Pipes For Surroundings.). If the upstream pressure is increased or if the downstream pressure is decreased.. If the downstream pressure is decreased below the choked pressure Pchoked. flow through the pipe remains choked and the constant. Determine pipe roughness E from Table 4-1. PI Pchoked -PI +1 . Determine T2from Equation 4-61.

For most problems involving choked adiabatic flows the pipe length ( L ) . Flow of Fluids Through Valves. is checked using EquaThis tion 4-64. llCrane Co. from 4. p. are the procedure is as follows: 1.inside diameter (d). and Pipes. is a gas expansion factor (unitless). Crane Co. and fittings (unitless). This assumption can be checked later but is normally valid. is the gravitational constant (force/mass-acceleration).. is the upstream gas density (mass/volume). is the downstream gas pressure (forcelarea). Fittings. g. For ideal gas flow the mass flow for both sonic and nonsonic conditions is represented by the Darcy formula: l1 where G is the mass flux (masslarea-time). m is the mass flow rate of gas (massltime). 2. . are the excess head loss terms. Determine the mass flux Gchoked Equation 4-66. A is the area of the discharge (length2). Determine the Fanning friction factor f using Equation 4-34. 1986).Chapter 4 Source Models Choked flow occurs if the downstream pressure is less than Pchoked. from Equations 4-63 through 4-67 for adiabatic pipe flow can be modified to use the 2-K method discussed previously by substituting C Kf for 4fLld. Determine Ma. This assumes fully developed turbulent flow at high Reynolds numbers. 3. pipe lengths.. The procedure can be simplified by defining a gas expansion factor Y. and C K. Technical Report 410 (New York. Y. including pipe entrances and exits.. To compute the mass flux G. and upstream pressure (PI) and temperature ( T I ) known. from Equation 4-67. Determine Pchoked Equation 5-64 to confirm operation at choked conditions. PI is the upstream gas pressure (forcelarea). P.

4-6 Flow of Gases through Pipes 141 The excess head loss terms Z Kfare found using the 2-K method presented earlier in section 4-4. within 1%. First.For most accidental discharges of gases the flow is fully developed turbulent flow. varying by less than 1% from the value at y = 1. . in Equation 4-68depends only on the heat capacity ratio of the gas y and the frictional elements in the flow path C Kf. total number and type of fittings. This computation can be done once and for all with the reAs the sults shown in Figures 4-13 and 4-14. Figure 4-14shows the expansion factor for y = 1. the upstream Mach number Ma. is determined using Equation 4-67. If less than the ratio from Equation 4-64.)IP. then the flow is not sonic and the actual pressure drop ratio is used. This is found from Equation 4-64. shown in Figure 4-13.) + D. The functional results of Figures 4-13and 4-14can be fitted using an equation of the form In Y. the expansion factor Y . are specified. Given: y based on the type of gas. pressure ratio (P. pipe entrances and exits..67. C. equation for the gas expansion An factor for choked flow is obtained by equating Equation 4-68to Equation 4-66 and solving for Y. where A. The Reynolds number can be calculated at the completion of the calculation to check this assumption. is calculated from Equation 4-69.. This means that for pipes the friction factor is independent of the Reynolds number and that for fittings K. The solution is obtained by trial and error. The calculation to determine the expansion factor can be completed once y and the frictional loss terms C K. The expansion factor Y. 2. ranges indicated. has little dependence on y. Sum the individual excess head loss terms to get C. is a weak function of the heat capacity ratio y. The next step in the procedure is to determine the sonic pressure ratio. B. This can be easily done using a spreadsheet. pipe length. The results are shown in Table 4-4and are valid for the K. The result is where Ma. diameter. Assume fully developed turbulent flow to determine the friction factor for the pipe and the excess head loss terms for the fittings and pipe entrances and exits. The gas expansion factor Y.2 to y = 1. by guessing values of the upstream Mach number and determining whether the guessed value meets the equation objectives.4. and the solution is direct. and type. Finally. The procedure to determine the gas expansion factor is as follows. The procedure to determine the adiabatic mass flow rate through a pipe or hole is as follows: 1. then the flow is sonic or choked and the pressure drop predicted by Equation 4-64is used to continue the calculation. and D are constants. upstream gas density.P.If the actual ratio is greater than the ratio from Equation 4-64. with the value of Y . total pressure drop. K. . Kf Z must be substituted for 4fLld to include the effects of pipes and fittings. = A(ln Kf)3+ B(ln K~)'+ C(ln K. is the upstream Mach number.4 over the range from y = 1. = K.

From AICHE/CCPS.2 All points at or above function are sonic flow conditions. Guidelines for Consequence Analysis of Chemical Releases (New York: American Institute of Chemical Engineers. Guidelines for Consequence Analysis of Chemical Releases (New York: American Institute of Chemical Engineers.4. 1999).142 Chapter 4 Source Models 0. . I I I I 0 0 10 20 30 Excess head loss. 1999). Kf 40 I 50 Figure 4-13 Sonic pressure drop for adiabatic pipe flow for various heat capacity ratios. From AICHE/CCPS. Kf 15 20 Figure 4-14 The expansion factor Y. for adiabatic pipe flow for y = 1. 0 1 0 I I I I 5 10 Excess head loss.

1141 0.0009 0.7248 -0. This method is applicable to gas discharges through piping systems and holes. and the Sonic Pressure Drop Ratio (P.1-300 0. A pressure gradient across T = Constanl For Surroundings. or using the equations provided in Table 4-4.P.P. lsothermal Flows Isothermal flow of gas in a pipe with friction is shown in Figure 4-15. Use the sonic choking pressure determined in step 3 in this expression.0308 -0. interpolating a value from the table. Calculate the mass flow rate using Equation 4-68. All areas above the curves in Figure 4-13 represent sonic flow. .6455 -0. by using Figure 4-13 directly.4 Sonic pressure drop ratio y = 1.4-6 Flow of Gases through Pipes 143 Table 4-4 Correlations1for the Expansion Factor Y.213 D 0. P 1 P2 > P choked Ma2 <1 Figure 4-15 Isothermal nonchoked flow of gas through a pipe.0011 0.5304 -0. Sonic pressure drop ratio y = 1. Calculate (PI .)lP..0287 C 0. Check this value against Figure 4-13 to determine whether the flow is sonic.0006 0.2 Sonic pressure drop ratio y = 1. 0.1-100 0.261 0. 5.0013 B -0.238 0. 4. Either read the value off of the figure.1-300 'The correlations are within 1%of the actual value in the specified range. ZTheequation used to fit the expansion factor and the sonic pressure drop ratio is of the form 3..67 A 0.5633 Range of K. Determine the sonic choking pressure P.0302 -0. Determine the expansion factor from Figure 4-14.0185 -0.1-100 0. as a Function of the Excess Head Loss K.2 Function value y Expansion factor Y .)IPI from the specified pressure drop. For this case the gas velocity is assumed to be well below the sonic velocity of the gas. or use the equation provided in Table 4-4. . interpolate it from the table.

-PI Maz' Ma. and because no mechanical linkages are present. we obtain12 T2 = TI. and. 12Levenspiel. Y~CM friction where G is the mass flux with units of mass/(area time). Isothermal flow is represented by the mechanical energy balance in the form shown in Equation 4-54. P z Ma1 --PI Ma2' P 2 Ma1 . The pressure at the end of the pipe is equal to the pressure of the surroundings. 46.144 Chapter 4 Source Models the pipe provides the driving force for the gas transport./g. . and kinetic energy compressibility The various energy terms in Equation 4-74 have been identified. PI G = pii - = . The temperature is constant across the entire pipe length. As the gas expands through the pressure gradient. p. Engineering Flow. the velocity must increase to maintain the same mass flow rate. assuming constant f. by combining Equations 4-29 and 4-30 and differentiating. A total energy balance is not required because the temperature is constant. The following assumptions are valid for this case: is valid for gases. By applying the assumptions to Equation 4-54 and manipulating them considerably.

46.*. - Uchoked -- 1 Ma1*' = Gchoked pU = plUl = MalPl d z - pchoked E. Levenspiell3 showed that the maximum velocity possible during the isothermal flow of gas in a pipe is not the sonic velocity. .inside diameter ( d ) . The result is A typical problem is to determine the mass flux G given the pipe length ( L ) . as shown in Figure 4-16. This assumption can be checked later but is usually valid. 2. Engineering Flow. In terms of the Mach number the maximum velocity is This result is shown by starting with the mechanical energy balance and rearranging it into the following form: The quantity -(dPldL) +CCJwhen Ma -+ 1 / G .and upstream and downstream pressures (PI and PI). as in the adiabatic case.Ma. This form is achieved by using Equations 4-70 through 4-72.4-6 Flow of Gases through Pipes 145 A more convenient form of Equation 4-74 is in terms of pressure instead of Mach numbers. p. Thus for choked flow in an isothermal pipe. the following equations apply: Tchoked = T 1 ? Pchoked -- . Determine the Fanning friction factor f using Equation 4-34. Compute the mass flux G from Equation 4-75. p 1 -- Pchoked P1 El - Malfi. 13Levenspiel. This assumes fully developed turbulent flow at high Reynolds numbers.The procedure is as follows: 1.

the actual flow rate is less than the adiabatic prediction and greater than the isothermal prediction. This assumes fully developed turbulent flow at high Reynolds numbers.. . P < PChoked T p1 TI U I = Constant / 7/ 1I I\ \\ ---------------t --------------- P2 = b P~hoked T2 = TI u2 = aly Ma2 = l l y Ma1 /" Velocity of a17 Reached at Exit of Pipe ' Figure 4-16 Isothermal choked flow of gas through a pipe. Engineering Flow. The mass flux G is determined using the following procedure: 1. and is 1 For most typical problems the pipe length (L). For both cases the velocity of the gas increases because of the expansion of the gas as the pressure decreases. This assumption can be checked later but is usually valid.146 Chapter 4 Source Models For Surroundings. Levenspiel l 4 showed that the adiabatic model always predicts a flow larger than the actual flow. 2. The maximum velocity is reached at the end of the pipe. The Crane Co. upstream pressure ( P . ) . Determine Ma. Example 4-5 shows that for pipe flow problems the difference between the adiabatic and the isothermal results is generally small. Determine the mass flux G from Equation 4-82. 45. from Equation 4-83. Flow of Fluids. l5Crane Co. For adiabatic flows the temperature of the gas may increase or decrease. p. and temperature (T) are known. 3. Determine the Fanning friction factor using Equation 4-34. where Gchokedthe mass flux with units of mass/(area time). For gas releases through pipes the issue of whether the release occurs adiabatically or isothermally is important. provided that the source pressure and temperature are the same. depending on the relative magnitude of the frictional and kinetic energy terms. For real pipe flows from a source at a fixed pressure and temperature. inside diameter (d). For choked flows the adiabatic choking pressure is less than the isothermal choking pressure.15 reported that "when compressible fluids discharge from the end of a reason14~evenspiel.

Equa: tion 4-50 provides the maximum mass flow rate. For nitrogen.. for a diatomic gas. In the event of a failure of the nitrogen regulator.. P.4 psia 1. Choked flow can be expected because the system is venting to atmospheric conditions.4-6 Flow of Gases through Pipes 147 ably short pipe of uniform cross-sectional area into an area of larger cross section. Which mass flow rate should be used? Solution a. (b) an adiabatic pipe. y = 1. Decide which result most closely corresponds to the real situation. the vessel will be exposed to the full 200psig pressure from the nitrogen source. Example 4-5 The vapor space above liquid ethylene oxide (EO) in storage tanks must be purged of oxygen and then padded with 81-psig nitrogen to prevent explosion.7 psia = 3. As a result..528)(214. This will exceed the pressure rating of the storage vessel. and (c) an isothermal pipe. t h e adiabatic flow model is t h e model of choice for compressible gas discharges through pipes.. The choked pressure from Equation 4-49 is. it must be equipped with a relief device to vent this nitrogen./ft2. Finally. the flow is usually considered t o b e adiabatic. The nitrogen in a particular facility is supplied from a 200-psig source. Determine the mass flow rate assuming (a) an orifice with a throat diameter equal to the pipe diameter. under choked sonic flow conditions isothermal conditions are difficult t o achieve practically because of t h e rapid speed of t h e gas flow. To prevent rupture of the storage vessel. = = (0.7 = 214." C r a n e supported this statement with experimental d a t a on pipes having lengths of 130 a n d 220 pipe diameters discharging air t o t h e atmosphere.63 X lo4lbf/ftZ.. It is regulated to 81-psig and supplied to the storage vessel through 33 ft of commercial steel pipe with an internal diameter of 1. Determine the required minimum mass flow rate of nitrogen through the relief device to prevent the pressure from rising within the tank in the event of a regulator failure. The area of the pipe is The absolute pressure of the nitrogen source is Po = 200 + 14. The maximum flow rate through the orifice occurs under choked conditions.4 and .7 psia) = 113.09 X lo4 lb.049 in.

4 + 1)Ma2 1 (5 - - 1 +1. The results are tabulated as follows: Guessed Ma Value of left-hand side of equation .4 mmlin) = 26.049 in)(l ft/l2 in) This equation is solved by trial and error for the value of Ma.148 Chapter 4 Source Models The molecular weight of nitrogen is 28 lbm/lb-mol. Thus b. The upstream Mach number is determined from Equation 4-67: with Y . from Table 4-1.4+1 --ln[2 2+(1. y = 1. = 1.4-1)Ma2 (1. Assume adiabatic choked flow conditions.046 mm.00564)(33 ft ) (1.4 ) [ (4)(0. assume a unit discharge coefficient C. given by Equation 4-56. E = 0.0. For commercial steel pipe.049 in) (25. The diameter of the pipe in millimeters is (1.Without any additional information.4. Substituting the numbers provided gives 1.6 mm. Thus From Equation 4-34 For nitrogen.

037 1b/ft3. The excess head loss resulting from the pipe length is given by Equation 4-30. The first consideration is whether the flow is sonic. The mass flux is computed using Equation 4-66: The simplified procedure with a direct solution can also be used.4 and K. The gas density under the upstream conditions is PI = P1M RgT (214. = 8.4-6 Flow of Gases through Pipes 149 This last guessed Mach number gives a result close to zero.56 p.PI p 2 - 0. The sonic pressure ratio is given in Figure 4-13 (or the equations in Table 4-4).770 * P2 = 49.4 psia.4 psia = 7.4 psia.69.4 psia to ensure choked flow.7 psia)(28 lbllb-mol) = 1. From Figure 4-14 (or Table 4-4) the gas expansion factor Y . (10. = 0. The friction factor f has already been determined: For this solution only the pipe friction will be considered and the exit effects will be ignored.7 psia) = = 49. Then from Equation 4-56 and from Equations 4-63 and 4-64 PChoked (0. It follows that the flow is sonic because the downstream pressure is less than 49. For y = 1.11 X 10"lbf/ft2 The pipe outlet pressure must be less than 49. -.731 psia ft3/lb-molo~)(5400~) .230)(214.

245 0.0 psia = = The mass flow rate is computed using Equation 4-82: = 8. Substituting the numbers provided. from Equation 4-79. . c.526 -0. For the isothermal case the upstream Mach number is given by Equation 4-83. f i = ( 2 1 4 . ? l b ..25 0.097 0. we obtain (aThe solution is found by trial and error: 1 Guessed Ma 0.005 t Final result - The choked pressure is. we obtain This result is essentially identical to the previous result.93 x 10' lbf/ft2. / i n 2 ) ( 0 .150 Chapter 4 Source Models By substituting this value into Equation 4-68 and using the choking pressure determined for P.362 0.24 0. Pchaked P . ~ a . 2 4 4 ) ~ 62. although with a lot less effort.244 Value of left-hand side of equation 0.

The results show that this approach results in a large result for this case. is easier to apply.4 62. the mass of liquid vaporized m . T . then the excess energy contained in the superheated liquid is given by This energy vaporizes the liquid.16 1. Flashing occurs so rapidly that the process is assumed to be adiabatic. the heat capacity of the liquid (energylmass deg). If the tank. however. is the heat of vaporization of the liquid. requiring only the pipe diameter and the upstream supply pressure and temperature. If AH.81 1. and T. ensuring a conservative safety design. C. sometimes explosively. For most real situations the heat transfer characteristics cannot be easily determined.4-7 Flashing Liquids 151 The results are summarized in the following table: Pchoked Qm Case Orifice Adiabatic pipe Isothermal pipe (psi4 113. The configurational details of the piping are not required.76 A standard procedure for these types of problems is to represent the discharge through the pipe as an orifice.0 (Ibm/s) 4. the temperature of the liquid before depressurization. The orifice method always produces a larger value than the adiabatic pipe method. it will always produce the larger number for a conservative safety design. Finally. is given by . pipe. as in the adiabatic and isothermal pipe methods. Thus the adiabatic pipe method is the method of choice. If m is the mass of original liquid. with a substantial difference between the orifice and the adiabaticlisothermal cases. The excess energy contained in the superheated liquid vaporizes the liquid and lowers the temperature to the new boiling point. the liquid will partially flash into vapor. 4-7 Flashing Liquids Liquids stored under pressure above their normal boiling point temperature present substantial problems because of flashing. note that the adiabatic and isothermal pipe methods produce results that are reasonably close.. or other containment device develops a leak.4 49. The orifice calculation. A choking design based on an orifice calculation might not be choked in reality because of high downstream pressures. Also note that the computed choked pressures differ for each case. the depressurized boiling point of the liquid.

.6 Btu/lb. Solution a.Solving for the fraction of the liquid vaporized.): where and are the mean heat capacity and the mean latent heat of vaporization. we obtain Example 4-6 One lb. The initial state is saturated liquid water at To = 350°F.m.152 Chapter 4 Source Models The fraction of the liquid vaporized is Equation 4-86 assumes constant physical properties over the temperature range To to Tb. The change in liquid mass m resulting from a change in temperature T is given by Equation 4-87 is integrated between the initial temperature To (with liquid mass m) and the final boiling point temperature Tb (with liquid mass m . (b) Equation 4-86. and (c) Equation 4-90. fv = m. At this temperature and under saturated conditions . Compute the fraction of material vaporized using (a) the steam tables. H = 321. or 212°F. The final temperature is the boiling point at 1atm. over the temperature range Toto Tb. A more general expression without this assumption is derived as follows.6 psia. respectively. The vessel ruptures and the pressure is reduced to 1 atm.lm. of saturated liquid water is contained in a vessel at 350°F. From the steam tables P = 134.

OF.0.4-7 Flashing Liquids 153 H.(T.. Hinitial the fraction of vapor (or quality) = and is computed from that is.e x p [ .8557 Both expressions work about as well compared to the actual value from the steam table. are - C. .01 Btullb.3 Btullb. "F)(350 . AH.04 Btullb.T. = Because the process occurs adiabatically. .)I=] = = = 1 . = 1. .. = 970. HIiquid 180. = 920.(1.04 Btullb.3 Btullb. "F)(350 .... 14. = 1150. b. Hfina.7 Btullb.07 Btu llb. OF.T. Substituting into Equation 4-90 gives f.4 Btullb.. 1 .01 Btullb. f" = i c.. AH.59% of the mass of the original liquid is vaporized.exp[-(1.)] 1 .7 Btullb. A H .) . For liquid water at 212'F C. = 1.G ( ~ .212)"F/(920. From Equation 4-86 C.212)"F 970. The mean properties for liquid water between To and T.

Procedures are available to solve this problem. Introduction to Chemical Engineering Thermodynamics. the flash calculation is complicated considerably. The equations describing incompressible fluid flow through holes apply (see section 4-2).968 X 106Pa. and its density is 603 kg/m3. C.p. Then 16J. Assume a discharge coefficient of 0. (New York: McGraw-Hill." Plantloperations Progress (July 1985).17If the fluid path length of the release is short (through a hole in a thin-walled container). Equilibrium flashing conditions can be assumed. 4th ed. nonequilibrium conditions exist. Van Ness. If the fluid path length through the release is greater than 10 cm (through a pipe or thickwalled container). and the liquid does not have time to flash within the hole. is the discharge coefficient (unitless). The result will be valid only for liquids stored at a pressure higher than the saturation vapor pressure. 314.M. 1987). Smith and H. With this assumption the mass flow rate is given by where A is the area of the release. A good approximation is to assume a choked pressure equal to the saturation vapor pressure of the flashing liquid. is the density of the liquid (mass/volume).61. A pipe of diameter 0. equilibrium flashing conditions are achieved and the flow is choked. P is the pressure within the tank. "Hans K. p. is Example 4-7 Liquid ammonia is stored in a tank at 24'C and a pressure of 1. C. p. the fluid flashes external to the hole.16 Flashing liquids escaping through holes and pipes require special consideration because two-phase flow conditions may be present. . allowing the flashing ammonia to escape.4 X lo6 Pa. because the more volatile components flash preferentially.Determine the mass flow rate through the leak. Fauske. Several special cases need consideration. "Flashing Flows or: Some Practical Guidelines for Emergency Releases. The saturation vapor pressure of liquid ammonia at this temperature is 0. Solution Equation 4-91 applies for the case of equilibrium flashing conditions. and PSat the saturation vapor eressure of the flashing liquid at ambient temperature.154 Chapter 4 Source Models For flashing liquids composed of many miscible substances.0945 m breaks off a short distance from the vessel (the tank). 133.

we obtain Equation 4-95 contains a maximum. Assume that kinetic energy is dominant and that potential energy effects are negligible. from a mechanical energy balance (Equation 4-l). and Equation 4-91 is no longer valid. A much more detailed approach is required. and realizing that the specific volume (with units of volume/mass) v = llp. Under choked flow conditions. Consider a fluid that is initially quiescent and is accelerated through the leak. at which choked flow occurs. Then.4-7 Flashing Liquids 155 For liquids stored at their saturation vapor pressure. we can write A mass velocity G with units of mass/(area time) is defined by Combining Equation 4-93 with Equation 4-92 and assuming that the mass velocity is constant results in Solving for the mass velocity G and assuming that point 2 can be defined at any point along the flow path. Differentiating Equation 4-95 and setting the result equal to zero gives . P = P". dGldP = 0.

. "Unconfined Vapor Cloud Explosions." in Eleventh Loss Prevention Symposium (New York: American Institute of Chemical Engineers. = -A H . dT' and from the Clausius-Clapyron equation. C* d f . These aerosol droplets are readily entrained by the wind and transported away from the release site. TC. 1977). and f is the mass fraction of vapor.. : dP AH.156 Chapter 4 Source Models Solving Equation 4-97 for G. The assumption that the quantity of droplets formed is equal to the amount of material flashed is frequently made.-- dv v. we obtain The two-phase specific volume is given by where vfgis the difference in specific volume between vapor and liquid. v. at saturation dP dT - AH.. Kletz. The mass flow rate is determined by combining Equation 4-103 with Equation 4-98: Note that the temperature Tin Equation 4-104 is the absolute temperature from the ClausiusClapyron equation and is not associated with the heat capacity. is the liquid specific volume. . Differentiating Equation 4-99 with respect to pressure gives But. Small droplets of liquid also form in a jet of flashing vapor.. from Equation 4-86.lx IsTrevor A. Tvk' Substituting Equations 4-102 and z-101 into Equation 4-100 yields .

AH.15 x lo6 Pa. A is the area of exposure. PSat 1. is R. is the mass vaporization rate (massltime). vfg = 0.18 x 10"Jlkg K)(298 K)(1 N mlJ) 4-8 Liquid Pool Evaporation or Boiling The case for evaporation of a volatile from a pool of liquid has already been considered in chapter 3. Pat the saturation vapor pressure of the liquid. M is the molecular weight of the pure material.4-8 Liquid Pool Evaporation or Boiling 157 Example 4-8 Propylene is stored at 25°C in a tank at its saturation pressure. The area of the leak is Using Equation 4-104. = C.042 m"kg. is the ideal gas constant.18 X 10" Jlkg K. Estimate the mass flow rate through the hole under these conditions for propylene: = 3. The total mass flow rate from the evaporating pool is given by Equation 3-12: where Q. . Solution Equation 4-104 applies to this case. A 1-cm-diameter hole develops in the tank. K is the mass transfer coefficient (lengthltime). = 2.34 X 10"Jlkg. and TLis the temperature of the liquid. we obtain x (2.

The higher hydrocarbons (C. a. q. The rate of boiling is determined by assuming that all the heat is used to boil the liquid. which can significantly alter the heat transfer behavior. is the mass boiling rate (massltime). solar heat fluxes and convective heat transfer from the atmosphere become important. At later times.158 Chapter 4 Source Models For liquids boiling from a pool the boiling rate is limited by the heat transfer from the surroundings to the liquid in the pool. A is the area of the pool (area). T. Heat is transferred (1) from the ground by conduction. and (3) by radiation from the sun andtor adjacent sources such as a fire. determined by Equation 4-105 (energylarea-time). This model also neglects possible water freezing effects in the ground. is the heat of vaporization of the liquid in the pool (energylmass).19 19Guidelinesfor Consequence Analysis of Chemical Releases (1999). More details on boiling pools is provided elsewhere. is the heat flux from the ground (energylarea-time). The heat transfer from the ground is modeled with a simple onedimensional heat conduction equation. This is especially true for a spill of liquid with a normal boiling point below ambient temperature or ground temperature. T is the temperature of the liquid pool (degree). is the temperature of the soil (degree). is the thermal conductivity of the soil (energytlength-time-degree). Equation 4-105 is not considered conservative. and AH. is the heat transfer for the pool from the ground. and t is the time after spill (time). This approach seems to work adequately for liquefied natural gas (LNG) and perhaps for ethane and ethylene. (2) from the air by conduction and convection. Q. is the thermal diffusivity of the soil (areattime). The initial stage of boiling is usually controlled by the heat transfer from the ground. For a spill onto an insulated dike floor these fluxes may be the only energy contributions. k. . given by where q. and above) require a more detailed heat transfer mechanism. Thus where .

then the realistic cases would be selected. To illustrate conservative modeling. The realistic releases represent the incident outcomes with a high probability of occurring. By doing so. it is more realistic to assume that a high probability exists that the release will occur from the disconnection of the largest pipe connected to the tank.4-10 Conservative Analysis 159 4-9 Realistic and Worst-Case Releases Table 4-5 lists a number of realistic and worst-case releases. then the worst-case releases must be used. however. defining the limits of the design envelope. These uncertainties arise because of (1)an incomplete understanding of the geometry of the release (that is. (2) the pressure within and outside the tank. have uncertainties. For example. including (1)the hole area. with the intent of estimating the toxicological impact. The discharge rate depends on a number of parameters. and (4) unknown or poorly understood mixture behavior. An equation of state (perhaps nonideal) is required to determine the tank pressure given the total mass. (3) the physical properties of the gas. If an internal company study is being completed to determine the actual consequences of plant releases. (2) unknown or poorly characterized physical properties. 4-10 Conservative Analysis All models. with the discharge rate decreasing as a function of time as the pressure within the tank decreases. Uncertainties that arise during the consequence modeling procedure are treated by assigning conservative values to some of these unknowns. requiring a mass discharge model cross-coupled to a material balance on the contents of the tank. rather than assuming that an entire storage vessel fails catastrophically. the hole size). dispersion modeling based on a groundlevel release will maximize the consequence for the surrounding community but will not maximize the consequence for plant workers at the top of a process structure. For any particular modeling study several receptors might be present that require different decisions for conservative design. consider a problem requiring an estimate of the gas discharge rate from a hole in a storage tank. (3) a poor understanding of the chemical or release process. to name a few. to name a few. and (4) the temperature of the gas. should be made to achieve a result consistent with the demands of the problem. This ensures that the resulting engineering design to mitigate or remove the hazard is overdesigned. The worst-case releases are those that assume almost catastrophic failure of the process. The complete dynamic solution to this problem is difficult. Thus. This discharge rate is used to estimate the downwind concentrations of the gas. resulting in near instantaneous release of the entire process inventory or release over a short period of time. However. The reality of the situation is that the maximum discharge rate of gas occurs when the leak first occurs. The selection of the release case depends on the requirements of the consequence study. a conservative estimate of the consequence is obtained. . including consequence models. if a study is being completed to meet the requirements of the EPA Risk Management Plan. Every effort.

or at process temperature. RMP Offsite Consequence Analysis Guidance (Washington. as appropriate. or the incidents can be developed from the outcome of a review or derived from hazard analysis studies. assume the entire quantity is released within 10 min. assume a full bore rupture. unless meteorological data indicate otherwise.160 Chapter 4 Source Models Table 4-5 Guidelines for Selection of Process Incidents Guideline Rupture of the largest diameter process pipe as follows: For diameters smaller than 2 in. Incident characteristic Realistic release incidents Process pipes Hoses Pressure relief devices relieving directly to the atmosphere Vessels Other Assume full bore rupture.DC: US Environmental Protection Agency. assuming a fixed temperature and pressure within the tank equal to the initial temperature and pressure. For diameters 2-4 in.5 m/s wind speed. Assume urban or rural topography. Consider liquids to be released at the highest daily maximum temperature. whichever is highest. Refer to pressure relief calculation. based on data for the previous 3 years. For diameters greater than 4 in. 1994). The actual discharge rate at later times will always be less. assume rupture area equal to 20% of the pipe cross-sectional area. A modeling effort of this detail is not necessarily required to estimate the consequence. . Assume release of the largest quantity of substance handled on site in a single process vessel at any time. Complicated temperature effects are also possible. To estimate the release rate. 1. Assume that the release occurs at ground level. 2USEPA. 1996). A much simpler procedure is to calculate the mass discharge rate at the instant the leak occurs. Use the pipe criteria. Assume a rupture based on the largest diameter process pipe attached to the vessel. Assume the highest daily maximum temperature and average humidity. In this fashion a conservative result is ensured. Incidents can be established based on the plant's experience. Assume F stability. Assume that gases liquefied by refrigeration at atmospheric pressure are released at their boiling points. assume rupture equal to that of a 2-inchdiameter pipe. All material released is assumed to be airborne. and the downwind concentrations will always be less. Worst-case incidents2 Quantity Wind speed 1 stability Ambient temperature I humidity Height of release Topography Temperature of release substance 'Dowk Chemical Exposure Index Guide (New York: American Institute of Chemical Engineers. Use calculated total release rate at set pressure.

24-32. in particular. Levenspiel. with each decision producing a maximum result. 2. This procedure is continued until all the model parameters are specified. 15/10. leading to a potential overdesign of the mitigation procedures or safety systems. This occurs. Loss Prevention. McCabe. because pipe disconnections are a frequent source of tank leaks. Smith. Lees. 1999). Flashing Liquids Steven R. 1984). (London: Butterworths. 2001). Julian C. 2d ed. Unit Operations of Chemical Engineering. 1996). if several decisions are made during the analysis. 1996). Flow of Liquid through Pipes Octave Levenspiel. Engineering Flow. Flow of Vapor through Holes Lees. 3. (New York: McGraw-Hill. 15/22. Warren L. tempered with a good dose of reality and common sense. Lees. 5. p. Engineering Flow. Guidelines for Consequence Analysis of Chemical Releases (New York: American Institute of Chemical Engineers. Again. pp. ch. Suggested Reading Consequence Modeling AICHE/CCPS. 6th ed. 2d ed. (New York: American Institute of Chemical Engineers. Hanna and Peter J. Drivas. Flow of Vapor through Pipes Levenspiel. pp. Engineering Flow and Heat Exchange (New York: Plenum Press. 1516. Flow of Liquid through Holes Frank P.Suggested Reading 161 For the hole area a possible decision is to consider the area of the largest pipe connected to the tank. 2000). Loss Prevention. For this reason. ch. consequence analysis should be approached with intelligence. Loss Prevention in the Process Industries. AICHEICCPS. Guidelines for Chemical Process Quantitative Risk Analysis (New York: American Institute of Chemical Engineers. 48-51. this maximizes the consequence and ensures a conservative result. ch. . p. this procedure can result in a consequence that is many times larger than the actual. p. Unfortunately. and Peter Harriott. Guidelines for Use of Vapor Dispersion Models.

If a 10-mm hole forms 3 m above the ground. the upstream pressure in the pipeline is 50 psig and the downstream pressure is 40 psig. The pressure in the pipeline at the point of the leak is 100 psig. and other hardware affect this result? 4-6. and g is the acceleration due to gravity. The specific gravity of benzene is 0. . At the time. The liquid in the tank has a density of 490 kg/m3. pp. valves.1 in.1 bar gauge to prevent a flammable atmosphere within the tank. 4-5. Problems 4-1. At a particular time the liquid level is 5 m high within the tank. h is the initial height above the ground. 4-4.866. The local ventilation rate is 2000 ft3/min and is deemed average.20-in hole develops in a pipeline containing toluene. For incompressible flow in a horizontal pipe of constant diameter and without fittings or valves show that the pressure is a linear function of pipe length. The tank is pressurized with nitrogen to 0. The specific gravity of toluene is 0. Guidelines. A 0. Hint: For a freely falling body the time to reach the ground is given by where t is the time. Estimate the distance from the tank the stream of liquid will hit the ground. What other assumptions are required for this result? Is this result valid for nonhorizontal pipes? How will the presence of fittings. a. A storage tank is 10 m high. 31. what is the initial mass discharge rate of liquid (in kgls)? b. Assume choked flow and a constant tank gas temperature of To. Estimate the diameter of a hole in the tank leading to a local hydrogen sulfide concentration equal to the TLV. The ambient pressure is 1 atm. Determine the leakage rate. Hydrogen sulfide gas is stored in a tank at 100 psig and 80°F. Determine whether this stream will be contained by a 1-m-high dike located 1 m from the tank wall. The diameter of the leak is estimated to be 0.8794. Estimate the mass flow rate of benzene through the leak. A 100-ft-long horizontal pipeline transporting benzene develops a leak 43 ft from the high-pressure end.39. The TLV-TWA for hydrogen sulfide gas is 10 ppm. 4-3. A tank contains pressurized gas. 4-2. Develop an equation describing the gas pressure as a function of time if the tank develops a leak.162 Chapter 4 Source Models Liquid Pool Evaporation and Boiling Hanna and Drivas.

(b) the maximum spill rate. The morning inspection of the tank farm finds a leak in the turpentine tank. and the density is 1600 kg/m3.000 L/hr. Determine (a) the total amount of turpentine spilled. If the pressure at one point in the pipe is 103 psig and a small leak develops 22 ft downstream. The pipe is plastic PVC pipe with an internal diameter of 50 mm. The density of turpentine at these conditions is 55 lb/ft3.004 kg/m s. For water at these conditions the viscosity is 1. An investigation finds that the leak was 0. 4-9. The pipe is commercial steel pipe with an internal diameter of 50 mm.000 liters - Per Hour Leo4 O c c u r s H e r e Figure 4-17 Process configuration for Problem 4-1 1 .5% hydrochloric acid solution is pumped from one storage tank to another. 4-11.1 m above the pipe outlet. Vent 1 m 1- 1- "I-. This point is 27 m in equivalent pipe length from the first tank. The tank diameter is 15 ft. Water is pumped through a 1-in schedule 40 pipe (internal diameter = 1.0 centipoise and the density is 62. Because of an accident./ft3. 4-10.0ft after the leak was repaired. compute the fluid pressure at the leak. A 31.3 ft before the leak occurred and 13. compute the pressure assuming that the valve is wide open.049 in) at 4-8. The viscosity of the solution is 1. 400 gallhr.8 X lop3kg/m s. The flow rate through the pipe is 10. The liquid in the pipe is crude oil with a density of 928 kg/m3and a viscosity of 0. and (c) the total time the leak was active. At a certain time the liquid level in the first tank is 4.4 lb.1 m below the pipe outlet of the first tank. The tank is vented to the atmosphere. I 10. Compute the flow rate (in kgls) from the leak. Records show that the turpentine level in the tank was 17. The power input to the pump is 2 kW and is 50% efficient.. at a point 2.1 in in diameter and 7 ft above the tank bottom. nd . If a globe valve is added to the pipe section of Problem 4-7.. The leak is repaired.Problems 163 4-7. the pipe is severed between the pump and the second tank. Compute the pressure in the pipe at the location shown on Figure 4-17. The pipe section is horizontal and without fittings or valves.

25-in (internal diameter) drawncopper tubing. estimate the flow of nitrogen from the tubing. allowing the tank contents to leak out. Unfortunately.164 Chapter 4 Source Models 1 5 0 mm ID P i p e Figure 4-18 Tank draining pro- cess for Problem 4-12. A tank with a drain pipe is shown in Figure 4-18. The heating coils consist of 20 ft of the pipe wound in a coil within the reactor. The coils are 0. If the cylinder falls and the valve is sheared off. If the heating coil pipe shears accidently. the reactor vessel will be exposed to the full 125-psigpressure of the steam. Because of a failure of the heat control. a. The tank contains crude oil. saturated. If the pipe shears off at the tank bottom. A home hot water heater contains 40 gal of water. A laboratory apparatus uses nitrogen at 250 psig. heat is continuously applied to the water in the tank.004 kglm s. Steam is supplied to the heating coils of a reactor vessel at 125 psig. the relief valve is clogged and the pressure rises past the maxi- . leaving a 50-mm hole. 4-13. estimate the initial mass flow rate of nitrogen from the tank. estimate the initial mass flow rate. the reactor must be equipped with a relief system to discharge the steam in the event of a coil shear. The steam is supplied from a main header through similar pipe with an equivalent length of 53 ft. Compute the maximum mass flow rate of steam from the sheared coils using two approaches: a. A cylinder in the laboratory contains nitrogen at 2200 psia.5 in. The crude oil has a density of 928 kg/m3and a viscosity of 0. Assuming adiabatic flow through the pipe. Assume a hole diameter of 0.546 in). If the tubing separates from the apparatus. The nitrogen is supplied from a cylinder. 4-15. exceeding the vessel's pressure rating. estimate the initial mass flow rate of material out of the drain pipe. through a regulator. 4-12. and the oil level is 7 m at the time. increasing the temperature and pressure. to the apparatus through 15 ft of 0. What is the force created by the jet of nitrogen? 4-14. 4-16. Assuming the leak in the coil is represented by an orifice. The nitrogen in the tank is at 75°F. If the drain pipe shears 2 meters below the tank.5-in schedule 80 pipe (internal diameter = 0. b. b. As a result. and there is concern that the drain pipe might shear off below the tank.

A tank 100 ft in diameter and 20 ft tall is filled with crude oil to within 2 ft of the top of the tank. Propane b.048 0.33 x 1. Use an average temperature and an effective ventilation rate of 3000 ft"/min. . Estimate the quantity of water flashed. Estimate the rate of emission of toluene (in kgls). estimate the maximum amount of crude oil (in gallons) leaked. PSat 36. allowing crude oil to drain out. and THis the final absolute temperature.09 Heat capacity (J/ kg K) 2. these vents also allow volatile materials to escape.36 X 103 c. If the temperature changes from 4°C to 30°C over a period = of 12 hr. One accident scenario is that a 6-in-diameter line connected to the bottom of the tank might break loose from the tank. If a 30-min emergency response time is required to stop the leak. and the specific gravity of crude oil is 0. Pmt is the saturation vapor pressure of the liquid.Problems 165 mum pressure of the vessel.. is the ideal gas constant. M is the molecular weight of the volatile.is the vapor volume of the tank. At 250 psig the tank ruptures. Methyl chloride d.75 x 3.4 mm Hg).9. The tank is vented to the atmosphere. V.56 X lo5 10' 10' 10' v.077 0.95 x 1X 0. 4-17. It is currently half full of toluene (M = 92. Ammonia Pressure (Pa) 0. A storage tank is 15 m in diameter and 10 m tall. R. b. (m3/kg) 0. If a worker is standing near the vent. An expression that can be used to estimate the volatile emission rate in a storage tank resulting from a single change in temperature is given by where rn is the total mass of volatile released.23 x 4.56 X 0.39 X 10' lo6 10" lo6 Heat of vaporization (J/kg) 3. a. Sulfur dioxide 10" lo" lo3 4-18.127 0. Large storage tanks need a breather vent (technically called a conservation vent) to allow air to move into and out of the tank as a result of temperature and pressure changes and a change in the tank liquid level.49 x 1.. c. estimate the concentration (in ppm) of toluene in the air.5 X 1. resulting in potential worker exposures.17 X 3. Is the worker overexposed? 4-19. Unfortunately. Calculate the mass flux (kg/m2s) for the following tank leaks given that the storage pressure is equal to the vapor pressure at 25°C: Toxic material a. Derive the equation form. TLis the initial low absolute temperature.

Integrate the equation to determine an expression for the liquid level height as a function of time. Estimate the maximum mass flow rate (in kgls) of the gas if the initial pressure in the tank is 800 kPa gauge. What is the final liquid level (in m)? c. how long will it take to empty the tank? c. Estimate the minimum pump size (in HP) required to pump the entire tank empty in 1hr. in which the material is transported away from the accident site before it becomes involved. C. A. What is the maximum spill rate of liquid (in kg/s)? Assume that the tank is vented. or the same? d.. One accident mitigation procedure is called emergency material transfer. b. b. and the ambient pressure is 1atm. A 3-cm (internal diameter) pipe has broken off of a 1-ton pig (or tank) of nitrogen.166 Chapter 4 Source Models 4-20. If a 3-cm (internal diameter) feed pipe at the bottom of the tank breaks off. is given by the following differential equation: where hLis the liquid level height above the leak.004 kg/m s. or the same? . At the current time the liquid height is within 1m of the top of the tank. What is the driving force that pushes the water out of the hole in the tank? c. 4-23. the liquid level h. a. does the liquid level change faster. A is the cross-sectional area of the leak. What conclusions can be drawn about the viability of this approach? The density of the crude oil is 928 kg/m3with a viscosity of 0. The temperature is 25"C. does the liquid level change faster. is the constant discharge coefficient (= 0. how much liquid (in m3) is spilled if an emergency response procedure requires 30 min to stop the flow? b. and g is the acceleration due to gravity. slower. Assume an initial liquid level above the leak of h. a. If a hole develops in the tank. We plan on mitigating a crude oil tank fire scenario by pumping the tank empty in 1hr total time. 4-22. The transfer will be accomplished by pumping the crude oil through a 200-mm (internal diameter) commercial steel pipe to another tank 40 m in diameter and 10 m high. If the cross-sectional area of the tank is increased. Assume a pump efficiency of 80%. A storage tank is vented to the atmosphere. If a 100-HPpump (80% efficient) is available.61). is the cross-sectional area of the tank. and the crude oil is typically at a level of 9 m. The pipeline represents 50 m of equivalent pipe. a. If the liquid level is increased. A storage tank contains water contaminated with a small quantity of a soluble hazardous waste material. The crude oil storage tank is 30 m in diameter. slower. 4-21. The tank is 3 m in diameter and 6 m high.

If the liquid level in the tank at the end of the spill was 8./s. b. Use a mechanical energy balance to show that the pump work required to pump a liquid through a pipe from one tank to another is given by where W . What new safety or operational problems might arise as a result of installing the orifice? 4-26.Problems 167 e. What would be the significance of the leak if the liquid were flammable? toxic? 4-24. estimate the length of time for the leak. b. A leak occurred in the tank because a 0. m is the mass flow rate. c. and A is the cross-sectional area of the pipe. By the time the liquid flow was stopped.1-m-diameter pipe located 1m above the bottom of the tank was accidentally disconnected. In Example 4-5 the maximum flow through the nitrogen line was determined in order to size the relief device. What was the original liquid level in the tank? 4-27. If a 1-in puncture occurs 2 ft off the bottom of the tank. is the work input to the pump.79 m./s. Calculate the orifice diameter required to reduce the flow from the nitrogen line to 0. Show that for any pump the maximum liquid discharge velocity is given by . Suppose that the reactor of Example 4-5 is equipped with a relief device capable of relieving nitrogen from the reactor vessel at the rate of 0. A 10-m-diameter round tank sits on the ground within a 20-m-square diked area. p is the density of the liquid. a. This is not enough to prevent overpressuring of the reactor in the event of a regulator failure. The liquid level in the tank is initially at 7 ft. a. One inherently safer design method is to install an orifice in the nitrogen supply line to limit the flow of nitrogen. a. how long will it take for the water to drain down to the leak? What is the total amount of liquid (in gallons) discharged? f. The tank is vented to the atmosphere.5 lb. This means that the process is designed in such a fashion as to prevent hazards from resulting in an accident.5 m above the tank bottom. L is the length of the pipe. The tank contains a hazardous material dissolved in mostly water. Estimate the total amount of liquid spilled (in m3 and in kg). f is the Fanning friction factor.5 Ib. An important concept in process safety is inherent safety. Be sure to list clearly your assumptions! 4-25. d is the diameter of the pipe. the liquid level in the diked area had reached a height of 0. A cylindrical tank 10 ft high and 20 ft in diameter is used to store water.

is the pump shaft work. 3 cm internal diameter. what is the maximum temperature increase expected in the water in "Clhr? Assume a constant heat capacity for the water of 1kcal/kg°C. If the pump is rated at 1HP. W. Make sure you list your assumptions in your solution. Comment on the magnitude of the result and the necessity for large pipe sizes in a chemical plant. Make sure that you clearly state and justify any assumptions! Remember. d. No other pumping or filling operations occur during this time. 5 cm internal diameter. 50 cm internal diameter.93. 25 cm internal diameter. assuming a gas velocity of 3 m/s. 4-31. If the pressure in the well reservoir is a constant 500 psig. The tank is 10 m high and 10 m in diameter. c. What is the maximum velocity of the liquid from this pump? What is the maximum discharge rate (in kgls)? Consider an oil well drilled to a depth of 1000 ft and connected to the surface with 4-in (internal diameter) commercial steel pipe. What is the total quantity (in kg) leaked? . b. 50 cm internal diameter. and a molecular weight of 44: a. and it contains a liquid with a specific gravity of 0.- - - 168 Chapter 4 Source Models 4-28. 4-30.9. 3 cm internal diameter. what is the expected flow rate of liquid oil (in barrels per day) at the surface if the surface pipe is open to the atmosphere? Assume 1000 ft of equivalent pipe and no gas flow with the oil. 25 cm internal diameter. A 1-kW pump discharges water through a 50-mm (internal diameter) pump outlet.4 centipoise. The strip chart in Figure 4-19 displays the history of a leak in a storage tank. At what height is the leak? c. When did the leak start. The specific gravity of the oil is 0. d. b. b. Calculate the number of kilograms per year of ideal gas that can be transported through the following pipe sizes. a pressure of 689 kPa gauge. and about how long did it last? b. assuming a constant liquid velocity of 1mls: a. where u is the maximum liquid discharge velocity. and its viscosity is 0. 4-29. 5 cm internal diameter. Comment on the magnitude of the result and the necessity for large pipe sizes in a chemical plant. a. What will happen if the pump continues to operate? Calculate the number of liters per year of liquid that can be transported through the following pipe sizes. an oil barrel is 42 gal. A pump contains 4 kg of water. p is the density of the liquid. Pumps can be blocked in by closing valves on the inlet and outlet sides of the pump. c. This can lead to a rapid increase in the temperature of the liquid blocked inside the pump. and A is the pump outlet discharge area. a temperature of 25"C. 4-32.

) is contained within a diked area with dimensions of 10 m x 10 m. assuming an effective ventilation rate for the outdoors of 3000 ft3/min. Estimate the leak hole diameter (in cm). A storage vessel containing carbon tetrachloride (CCI. e. A is the area of the hole. and the ambient pressure is 1atm. and g is the acceleration due to gravity. the time for the liquid to drain from a hole in the bottom of the sphere is given by where D is the inside diameter of the sphere. Show that for a spherical storage vessel containing liquid at an initial height h. estimate the CC14 vapor concentration near the vessel (in ppm). is the discharge coefficient. .Problems 169 1- 0 I I I I I I I I I I I I 1 I I I 0 200 400 600 800 1000 Time after midnight (min) 1200 1400 1600 Figure 4-19 Strip chart data for Problem 4-32. The temperature of the liquid is 35"C. 4-34.4. If one of the accident scenarios for this vessel results in a leak with a discharge rate of 1 kgls. The storage tank is in a horizontal bullet configuration with legs to raise the vessel well above the dike floor. a. C. Estimate the maximum discharge rate of the fluid (in kgts). What spill rate (in kgls) from the storage vessel is required to completely fill the floor of the dike with liquid? b. The atomic weight of chlorine is 35. 4-33. d.

16 X lo-' m2/s Thermal conductivity of soil: 0. You will need the following data: Thermal diffusivity of soil: 4. Estimate the vaporization rate resulting from heating from the ground at 10 s after the instantaneous spill of 1500 m3 of liquefied natural gas (LNG) into a rectangular concrete dike of dimensions 7 m by 10 m.170 Chapter 4 Source Models 4-35.92 WI m K Temperature of liquid pool: 109 K Temperature of soil: 293 K Heat of vaporization of pool: 498 kJ1kg at 109 K .

Toxic release models are routinely used to estimate the effects of a release on the plant and community environments. and rupture of a train or truck transportation tank following an accident. estimating the downwind concentrations of the toxic material using a dispersion model (once the downwind concentrations are known. rupture of a tank containing toxic material stored above its atmospheric boiling point. An excellent safety program strives to identify problems before they occur. A few examples are explosive rupture of a process vessel as a result of excessive pressure caused by a runaway reaction. These steps are D 1. identifying the release incident (what process situations can lead to a release? This was described in sections 4-9 and 4-10). developing a source model to describe how materials are released and the rate of release (this was detailed in chapter 4). This requires a toxic release model. process equipment can release toxic materials quickly and in significant enough quantities to spread in dangerous clouds throughout a plant site and the local community. rupture of a pipeline containing toxic materials at high pressure. several criteria are available to estimate the impact or effect. . Serious accidents (such as Bhopal) emphasize the importance of planning for emergencies and of designing plants to minimize the occurrence and consequences of a toxic release. Chemical engineers must understand all aspects of toxic release to prevent the existence of release situations and to reduce the impact of a release if one occurs. Toxic release and dispersion models are an important part of the consequence modeling procedure shown in Figure 4-1. 2. The toxic release model represents the first three steps in the consequence modeling procedure.Toxic Release and Dispersion Models uring an accident. and 3. as discussed in section 5-4).

Atmospheric stability is classified according to three stability classes: unstable. ground conditions (buildings. During the day. The air temperature difference does not influence atmospheric mechanical turbulence. At night the temperature decrease is less. This most often occurs at night because the ground cools rapidly as a result of thermal radiation. the plume in Figure 5-1 becomes longer and narrower. as shown in Figure 5-2. for example. the substance is carried downwind faster but is diluted faster by a larger quantity of air. For neutral stability the air above the ground warms and the wind speed increases. the air temperature decreases rapidly with height. as shown in Figure 5-1.172 Chapter 5 Toxic Release and Dispersion Models Various options are available. The maximum concentration of toxic material occurs at the release point (which may not be at ground level). After a release the airborne toxic material is carried away by the wind in a characteristic plume. water. (2) developing engineering modifications to eliminate the source of the release. A wide variety of parameters affect atmospheric dispersion of toxic materials: wind speed. neutral. therefore . encouraging vertical motions. resulting in less vertical motion. the temperature increases with height. (1) developing an emergency response plan with the surrounding community. trees). This results in unstable stability because air of lower density is below air of greater density. Temperature profiles for day and night situations are shown in Figure 5-3. and stable. atmospheric stability. or a puff. These options are discussed in more detail in section 5-6 on release mitigation. because of turbulent mixing and dispersion of the toxic substance with air. as might be observed in the early morning hours. and (5) adding area monitors to detect incipient leaks and providing block valves and engineering controls to eliminate hazardous levels of spills and leaks. height of the release above ground level. As the wind speed increases. (3) enclosing the potential release and adding appropriate vent scrubbers or other vapor removal equipment. For stable atmospheric conditions the sun cannot heat the ground as fast as the ground cools. This influence of buoyancy enhances atmospheric mechanical turbulence. resulting in minimal vertical motion. During an inversion. For unstable atmospheric conditions the sun heats the ground faster than the heat can be removed so that the air temperature near the ground is higher than the air temperature at higher elevations. or insolation. Concentrations downwind are less. Sometimes an inversion occurs. (4) reducing inventories of hazardous materials to reduce the quantity released. reducing the effect of solar energy input. momentum and buoyancy of the initial material released. Atmospheric stability relates to vertical mixing of the air. based on the predictions of the toxic release model. 5-1 Parameters Affecting Dispersion Dispersion models describe the airborne transport of toxic materials away from the accident site and into the plant and community.

.

.

p. OC Figure 5-3 Air temperature as a function of altitude for day and night conditions. and Welfare. The temperature gradient affects the vertical air motion. and Welfare. 1970). 1970). 1. Bruce Turner. Figure 5-4 shows the change in wind speed versus height for a variety of surface conditions. The influence of buoyancy suppresses mechanical turbulence. Bruce Turner. This condition is stable because the air of higher density is below air of lower density. Workbook of Atmospheric Dispersion Estimates (Cincinnati: US Department of Health. Trees and buildings increase mixing. whereas lakes and open areas decrease it. Workbook of Atmospheric Dispersion Estimates. Adapted from D. Adapted from D. p. the temperature near the ground is lower than the air temperature at higher elevations. .5-1 Parameters Affecting Dispersion T e m p e r a t u r e . Education. Ground conditions affect the mechanical mixing at the surface and the wind profile with height. Education. (Cincinnati: US Department of Health. ' l W i n d G r a d i e n t s U r b a n S u b u r b s F l a t Figure 5-4 Effect of ground conditions on vertical wind gradient. 2.

The distinction between the two .176 Chapter 5 Toxic Release and Dispersion Models Continuous release source Wind direction I I I As release height increases. Figure 5-5 Increased release height decreases the ground concentration. resulting in a much higher effective release height. this distance increases. as the gas travels downwind and is mixed with fresh air. Thus these models apply to gases at low concentrations. The puff model describes the temporal concentration of material from a single release of a fixed amount of material. The plume model describes the steady-state concentration of material released from a continuous source. ground-level concentrations are reduced because the plume must disperse a greater distance vertically. As the release height increases. The momentum of a high-velocity jet will carry the gas higher than the point of release. The release height significantlyaffects ground-level concentrations. Figure 5-6 demonstrates these effects. then the released gas will initially be negatively buoyant and will slump toward the ground. For all gases. a point will eventually be reached where the gas has been diluted adequately to be considered neutrally buoyant. the released gas will initially be positively buoyant and will lift upward. 5-2 Neutrally Buoyant Dispersion Models Neutrally buoyant dispersion models are used to estimate the concentrations downwind of a release in which the gas is mixed with fresh air to the point that the resulting mixture is neutrally buoyant. The buoyancy and momentum of the material released change the effective height of the release. Two types of neutrally buoyant vapor cloud dispersion models are commonly used: the plume and the puff models. typically in the parts per million range. If the gas has a density less than air.The increased distance leads to greater dispersion and a lower concentration at ground level. If the gas has a density greater than air. The temperature and molecular weight of the released gas determine the gas density relative to that of air (with a molecular weight of 28. At this point the dispersion is dominated by ambient turbulence.97). This is shown in Figure 5-5.

into an infinite expanse of air (a ground surface will be added later).y. the puff model must be used.5-2 Neutrally Buoyant Dispersion Models 177 Initial acceleration and dilution -.Dominance of / . the concentration C of material resulting from this release is given by the advection equation where uj is the velocity of the air and the subscript j represents the summation over all coordinate directions x. the plume model is recommended because it is easier to use. A large vapor cloud is formed that moves away from the rupture point. The coordinate system is fixed at the source. The puff model can be used to describe a plume. \ \ \ \ \ \ Dominance of internal buoyancy .. if steady-state plume information is all that is required. Q. For the plume model a typical example is the continuous release of gases from a smokestack. . and 2 . Guidelines for Use of Vapor Cloud Dispersion Models (New York: American Institute of Chemical Engineers. . . / / . we would find that the material disperses much faster than . 1987). the effect on a plume of a change in wind direction)._ _ . For studies involving dynamic plumes (for instance. Hanna and Peter J. models is shown graphically in Figures 5-1 and 5-2. The dispersion models discussed in this chapter represent only ambient turbulence. Consider the instantaneous release of a fixed mass of material. a plume is simply the release of continuous puffs. . A steady-state plume is formed downwind from the smokestack. Assuming no reaction or molecular diffusion. . For the puff model a typical example is the sudden release of a fixed amount of material because of the rupture of a storage vessel. However. 6. p._ Wind . Adapted from Steven R. Drivas.. \ ambient turbulence \ \ Release source Transition from dominance of internal buoyancy to dominance of ambient turbulence Figure 5-6 The initial acceleration and buoyancy of the released material affects the plume character. If the velocity uj in Equation 5-1 is set equal to the average wind velocity and the equation is solved.

where (C)is the mean concentration and C' is the stochastic fluctuation.C1) not necessarily zero and remains in the equation. (5-3) (C') = 0.) = 0. The usual approach is to define an eddy diffusivity K. including the effects resulting from turbulence.)C1) (u. it follows that (u.178 Chapter 5 Toxic Release and Dispersion Models predicted. so C = (C) + C'. This is due to turbulence in the velocity field. (with units of areahime) such that . Equation 5-1 would predict the correct concentration.)C1 and u. an approximation is used. Substituting Equations 5-2 and 5-3 into Equation 5-1 and averaging the result over time yields The terms ju.(C)are zero when averaged (((u. no models are currently available to adequately describe turbulence. is An additional equation is required to describe the turbulent flux.)(C1r O). Because the fluctuations in both C and uj are around the average or mean values.) is the average velocity and u. Unfortunately. is the stochastic fluctuation resulting from turbulence. Let the velocity be represented by an average (or mean) and stochastic quantity where (u. If we could specify the wind velocity exactly with time and position. It follows that the concentration C will also fluctuate as a result of the velocity field. As a result. but the turbulent = = flux term (u.

then and Equation 5-7 becomes Equation 5-9 together with appropriate boundary and initial conditions forms the fundamental basis for dispersion modeling. . Note the coordinate system: xis downwind direction. The coordinate system used for the dispersion models is shown in Figures 5-7 and 5-8.- 5-2 Neutrally Buoyant Dispersion Models Substituting Equation 5-6 into Equation 5-5 yields If the atmosphere is assumed to be incompressible. They axis is the distance off the centerline.. and the z axis is the elevation I Continuous Release o f Occurs Material / a t R a t e Q. and z is vertical direction. The x axis is the centerline directly downwind from the release point and is rotated for different wind directions. y is off-wind direction. Here Wind Direction \ with Wind S p e e d .- . This equation will be solved for a variety of cases. u Figure 5-7 Steady-state continuous point source release with wind.

0) are level with the release point.180 Chapter 5 Toxic Release and Dispersion Models Wind Direction Wind S p e e d . and the coordinates (x. z ) = (0.0) are along the centerline. u I n s t a n t a n e o u s Release o f Material Here a t t Q. Transforming Equa- . no wind ((u. For this case Equation 5-9 reduces to the form Equation 5-10 is more tractable by defining a radius as r2 = x2 + y 2 tion 5-10 in terms of r yields + z2. the puff moves with the wind. y. Case 1: Steady-State Continuous Point Release with No Wind The applicable conditions are constant mass release rate (Q. . or x axis. steady state (d(C)ldt = 0).y.0. The point (x. above the release point. 0.) = 0). The coordinates (x. and constant eddy diffusivity (K. = K* in all directions). After the initial instantaneous release. Occurs 3 : L t Initial Puff at t = 0 P u f f Moves D o w n w i n d a t u Velocity = 0 z Puff at t = t 1 X Concentrations Are E q u a l o n B o t h " R o u g h " Surfaces Figure 5-8 Puff with wind.0) is at the release point. = constant).

Equation 5-15 is transformed to rectangular coordinates to yield Case 2: Puff with No Wind The applicable conditions are puff release. Equation 5-9 reduces for this case to . that is. and constant eddy diffusivity (K. = K* in all directions).) = O). = m: (5-13) Equation 5-12 is separated and integrated between any point r and r Solving Equation 5-14 for (C)yields It is easy to verify by substitution that Equation 5-15 is also a solution to Equation 5-11 and thus a solution to this case.5-2 Neutrally Buoyant Dispersion Models 181 For a continuous steady-state release the concentration flux at any point r from the origin must equal the release rate Qm(with units of massltime). instantaneous release of a fixed mass of material Q: (with units of mass). no wind ((u. (C)-0. This is represented mathematically by the following flux boundary condition: The remaining boundary condition is As r -+m.

The solution is found by integrating the instantaneous solution (Equation 5-19 or 5-20) with respect to time. 256.182 Chapter 5 Toxic Release and Dispersion Models The initial condition required to solve Equation 5-17 is y. Jaeger.Y . . z. 261. p. C. 'H. 1959). Carslaw and J. For this case Equation 5-9 reduces to Equation 5-17 with the initial condition expressed by Equation 5-18 and the boundary condition expressed by Equation 5-13. (C)(x. no wind ((ui) O ) . and = constant eddy diffusivity (K. Conduction of Heat. 2Carslaw and Jaeger. = K* in all directions). S. p. The solution to Equation 5-17 in spherical coordinates1 is and in rectangular coordinates it is Case 3: Non-Steady-State Continuous Point Release with No Wind The applicable conditions are constant mass release rate (Q. Equations 5-21 and 5-22 reduce to the corresponding steady-state solutions (Equations 5-15 and 5-16). Conduction of Heat in Solids (London: Oxford University Press. 2 . t ) = 0 at t = 0. = constant). The result in spherical coordinates2 is and in rectangular coordinates it is Qm (C)(x. t ) = 4 r r K * d x 2 + y2 + z2 erfc ( ' x2 + Y 2 + z2 22/K*t (5-22) As t -+ oo.

= constant). K.) = u = constant). The applicable conditions are continuous release ( Q .)= O). = constant). Conduction of Heat. . = K* in all directions). p. and Case 5: Puff with No Wind and Eddy Diffusivity Is a Function of Direction This case is the same as case 2 but with eddy diffusivity a function of direction.Equation 5-24 is simplified to Along the centerline of this plume.) = (u. then by using . y = z = 0. that is. The applicable conditions are 1 puff release (Q. T a r s l a w and Jaeger. no wind ((u. The solution for the average concentration at any point3 is If a slender plume is assumed (the plume is long and slender and is not far removed from the x axis). For this case Equation 5-9 reduces to Equation 5-23 is solved together with boundary conditions expressed by Equations 5-12 and 5-13.. = 1 + aI2. 267.5-2 Neutrally Buoyant Dispersion Models 183 Case 4: Steady-State Continuous Point Source Release with Wind This case is shown in Figure 5-7. and constant eddy diffusivity (K. and K.. wind blowing in x direction only ((u.). and each coordinate direction has a different but constant eddy diffusivity (K.

. p. wind blowing in x direction only ((u. and K. = constant). Loss Prevention in the Process Industries. k "ees.) = u = constant). K. K. and each coordinate direction has a different but constant eddy diffusivity (K.- -- Chapter 5 Toxic Release and Dispersion Models Equation 5-9 reduces to the following equation for this case: The solution is4 Case 6: Steady-State Continuous Point Source Release with Wind and Eddy Diffusivity Is a Function of Direction This case is the same as case 4 but with eddy diffusivity a function of direction. 4 ~ r a n P.). 1996).)= (u. Equation 5-9 reduces to The solution is Along the centerline of this plume. Loss Prevention. 2d ed. Figure 5-8 shows the geometry. . steady-state (d(C)lat= 0). = constant).. 151106.. wind blowing in x direction only ((uj) (u. and K.. The applicable conditions are puff release (Q. Lees.).p. and the average concentration is given by Case 7: Puff with Wind This case is the same as case 5 but with wind. The applicable conditions are continuous release (Q. 151107.) = u = constant). y = z = 0.. (London:Butterworths. = each coordinate direction has a different but constant eddy diffusivity (K. and slender plume approximation (Equation 5-25).

The solution is 2 times Equation 5-29: Qi (C)(x7 Y' z' ') = 4 ( ~ t 1 x2 y2 -+ -+ ) exp[-%~ ( K . The solution is simply Equation 5-29. As a result. If the puff moves with the wind along the x axis. The solution to case 5 represents a puff fixed around the release point. transformed into this new coordinate system: Case 8: Puff with No Wind and with Source on Ground This case is the same as case 5 but with the source on the ground. and u is the wind velocity. As a result. the concentration is twice the concentration in case 5. The solution is 2 times Equation 5-31: z Wind Continuous Release of Material Direction with Wind Speed.ut. u Figure 5-9 Steady-state plume with source at ground level.5-2 Neutrally Buoyant Dispersion Models 185 The solution to this problem is found by a simple transformation of coordinates. that moves with the wind velocity. the solution to this case is found by replacing the existing coordinate x by a new coordinate system. The ground represents an impervious boundary. The variable t is the time since the release of the puff. The concentration is twice the concentration of a plume without the ground. .(5-34) Case 9: Steady-State Plume with Source on Ground This case is the same as case 6 but with the release source on the ground. the concentration is twice the concentration in case 6. x . The ground represents an impervious boundary. " ~ K Kz i)]. as shown in Figure 5-9.

This is due to a change in vertical temperature profiles from day to night. for a continuous source are given in Figures 5-10 and . shown in Table 5-1. and prevailing weather conditions. . G. 286.186 Chapter 5 Toxic Release and Dispersion Models Case 10: Continuous Steady-State Source with Source at Height H. = 0. whereas at night the reverse is true. the standard deviations of the concentration in the downwind. The atmospheric conditions are classified according to six different stability classes.. p. respectively. Values for a are not provided 6Lees. represent . a . and vertical (x. . Micrometeorology (New York:McGraw-Hill. The dispersion coefficients are a function of atmospheric conditions and the distance downwind from the release. increased wind speed results in greater atmospheric stability. 151107. it is not convenient experimentally and does not provide a useful framework for correlation. The stability classes depend on wind speed and quantity of sunlight. Values for the dispersion coefficients are much easier to obtain experimentally than eddy diffusivities. Although the eddy diffusivity approach is useful theoretically. Equation 5-36 reduces to Equation 5-35 for a source on the ground. wind velocity. Pasquill-Gifford Model Cases 1 through 10 all depend on the specification of a value for the eddy diffusivity K. 1953). with the corresponding correlations given in Table 5-2. time. During the day. The solution is" fi)(x. crosswind. . Y. z ) directions. z ) = e m eXP(--) uy" 4pXw If H. and a.Loss Prevention. above the Ground For this case the ground acts as an impervious boundary at a distance H from the source. . p. y . Kj changes with position. The dispersion coefficients a and a. Sutton7 solved this difficulty by proposing the following definition for a dispersion coeficient: with similiar expressions given for a and uZ The dispersion coefficients a. 5-11. 70. Sutton. In general.

km 10 Figure 5-1 0 Dispersion coefficients for . neutrally stable E. A." Nuclear Safety (1976). Slight insolation to similar conditions in midwinter. I o4 A 1o3 Ei C 102 D E F 10' . 2(4): 47.1 1 Distance downwind.Giffordplume model for rural releases. for overcast conditions during day or night and for any sky conditions during the hour before or after sunset or sunrise. respectively. slightly stable D.- 0. 4Night refers to the period 1 hour before sunset and 1 hour after dawn. 2F.A. 3Strong insolation corresponds to a sunny midday in midsummer in England. slightly stable E moderately stable IF. 17(1): 68. moderately unstable C. "These values are filled in to complete the table.1 1 Distance downwind.2 Nighttime conditions4 Surface wind speed (m 1s) <2 2-3 3-4 4-6 >6 Daytime insolation3 Strong Moderate A-B B B-C C-D D6 Slight B C C D6 D6 Thin overcast 5318 or >4/8 low cloud cloudiness A A-B B C C F5 E D D6 D F F E Dh D6 Stability classes: A. regardless of wind speed. "Use of Routine Meteorological Observations for Estimating Atmospheric Dispersion. Gifford. extremely unstable B. "Turbulent Diffusion-Typing Schemes: A Review.5-2 Neutrally Buoyant Dispersion Models 187 Table 5-1 Atmospheric Stability Classes for Use with the Pasquill-Gifford Dispersion Modell. . Gifford." Nuclear Safety (1961). 6Theneutral category D should be used. km 10 0.

Atmospheric Turbulence.1 1 10 Distance downwind." Atmospheric Environment (1994). 'R. Report ATDL-106 (Washington. km Figure 5-11 Distance downwind. 2G.E Griffiths. Dispersion coefficients for Pasquill- Table 5-2 Recommended Equations for Pasquill-Gifford Dispersion Coefficients for Plume Dispersionl.188 Chapter 5 Toxic Release and Dispersion Models A-B A-B C D E-F E n E-F 1 0. Diffusion Estimation for Small Emissions. . 1974). km -Giffordplume model for urban releases. and Diffusion Laboratory. DC: Air Resources. "Errors in the Use of the Briggs Parameterization for Atmospheric Dispersion Coefficients. Environmental Research Laboratories. Briggs. 28(17):2861-2865.2 (the downwind distance x has units of meters) Pasquill-Gifford stability class (m) ffz(m) Rural conditions A B C D E F Urban conditions A-B D D E -F - A-F are defined in Table 5-1. A.

F. Environmental ResearchLaboratories. Griffiths. km 10 0. . 28(17): 2861-2865. DC: Air Resources.5-2 Neutrally Buoyant Dispersion Models 189 0.'~ 0. Report ATDL-106 (Washington.1 1 Distance downwind.1 1 Distance downwind.~" 0. (m) or a. ~ (the downwind distance x has units of meters) Pasquill-Gifford stability class A B C D E F a." 0. andDiffusionLaboratory. .53x07" 0 ." Atmospheric Environment (1994).'~ 0 . 1974). = a. Atmospheric Turbulence.18~~.'~ 0. 'G. km 10 Figure 5-12 Dispersion coefficients for Pasquill-Giffordpuff model. "Errorsin the Use of the Briggs ParameterizationforAtmospheric Dispersion Coefficients. 'R. Briggs. Diffusion Estimation for Small Emissions. The puff dispersion coefficients are based on limited data (shown in Table 5-2) and should not be considered precise. 1 0 h5 ~ ~ 0.~' A-F are defined in Table 5-1.05~'.34x0. The dispersion coefficients a and a. 1 0 ~ ~ ~ ~ ~ 0. because it is reasonable to assume that a. A.'~ 0. (m) U Z(m) 0. for a puff . Table 5-3 Recommended Equations for PasquillGifford Dispersion Coefficients for Puff D i s p e r s i ~ n ~ .04~O. ~ " 0.60~O.06x~.1~ x O .14~O. release are given in Figure 5-12 and the equations are provided in Table 5-3.

. 1962). received by an individual standing at fixed coordinates (x.O. The center of the cloud is found at coordinates (ut. The concentration at the center of this moving cloud is given by The total integrated dose D.Pasquill.y. Q. The solution has a form similar to Equation 5-33: The ground-level concentration is given at z = 0: The ground-level concentration along the x axis is given at y = z = 0: (C)(x. Coordinates Fixed at Release Point. 1 x-ut v2%-3'2ffxuy~. These equations along with the correlations for the dispersion coefficients are known as the Pasquill-Gifford model.Chapter 5 Toxic Release and Dispersion Models The equations for cases 1through 10 were rederived by Pasquills using expressions of the form of Equation 5-37.2 1. (o. Atmospheric Diffusion (London: Van Nostrand.O). .) 4 . t ) = O. z ) is the time integral of the concentration: 8F. Case 11: Puff with Instantaneous Point Source at Ground Level. Constant Wind Only in x Direction with Constant Velocity u This case is identical to case 7..0.

The procedure is repeated for each value of t required. Set (C)(x. and t.5-2 Neutrally Buoyant Dispersion Models The total integrated dose at ground level is found by integrating Equation 5-39 according to Equation 5-42. t ) along the x axis using Equation 5-40. For a specified concentration (C)* the isopleths at ground level are determined by dividing the equation for the centerline concentration (Equation 5-40) by the equation for the general ground-level concentration (Equation 5-39). The line connecting points of equal concentration around the cloud boundary is called an isopleth. Specify (C)*. 3. 2. The solution has a form similar to Equation 5-35: . t ) = (C)" in Equation 5-45.0 .y. 0 . The result is The total integrated dose along the x axis on the ground is Frequently the cloud boundary defined by a fixed concentration is required.u. Determine the concentrations (C)(x. Define the boundary of the cloud along the x axis. and determine the values of y at each centerline point determined in step 2. Case 12: Plume with Continuous Steady-State Source at Ground Level and Wind Moving in x Direction at Constant Velocity u This case is identical to case 9. 0 . This equation is solved directly for y: The procedure is 1.

For continuous ground-level releases the maximum concentration occurs at the release point. Case 13: Plume with Continuous Steady-State Source at Height H...192 Chapter 5 Toxic Release and Dispersion Models The ground-level concentration is given at z = 0: The concentration along the centerline of the plume directly downwind is given at y = z = 0: The isopleths are found using a procedure identical to the isopleth procedure used for case 11. above Ground Level and Wind Moving in x Direction at Constant Velocity u This case is identical to case 10. The solution has a form similar to Equation 5-36: - The ground-level concentration is found by setting z = 0: The ground-level centerline concentrations are found by setting y = z = 0: The maximum ground-level concentration along the x axis (C). is found using .

because their values change as the puff moves downwind from the release point. At ground level. The result is . followed by using Equation 5-52 to determine the maximum concentration. Case 14: Puff with Instantaneous Point Source at Height Hrabove Ground Level and a Coordinate System on the Ground That Moves with the Puff For this case the center of the puff is found at x = ut. The average concentration is given by The time dependence is achieved through the dispersion coefficients. and the concentration is computed using The concentration along the ground at the centerline is given at y = z = 0: The total integrated dose at ground level is found by applying Equation 5-42 to Equation 5-55. If wind is absent ( u = O). Equation 5-54 does not predict the correct result.5-2 Neutrally Buoyant Dispersion Models 193 The distance downwind at which the maximum ground-level concentration occurs is found from The procedure for finding the maximum concentration and the downwind distance is to use Equation 5-53 to determine the distance. z = 0.

The assumptions used in the calculation must be clearly stated.5 m/s. If the release occurs above ground level. we can see that the smallest dispersion coefficients occur with F stability. . "PA. the wind speed cannot be zero. reaches a maximum. Worst-Case Conditions For a plume the highest concentration is always found at the release point. the highest concentration is estimated. DC: US Environmental Protection Agency. By inspecting Figures 5-10 through 5-12.Chapter 5 Toxic Release and Dispersion Models Case 15: Puff with Instantaneous Point Source at Height Hrabove Ground Level and a Coordinate System Fixed on the Ground at the Release Point For this case the result is obtained using a transformation of coordinates similar to the transformation used for case 7. For a puff isopleth the isopleth is close to circular as it moves downwind. If weather conditions are not known or are not specified. R M P Offsite Consequence Analysis Guidance (Washington. Limitations to Pasquill-Gifford Dispersion Modeling Pasquill-Gifford or Gaussian dispersion applies only to neutrally buoyant dispersion of gases in which the turbulent mixing is the dominant feature of the dispersion. then certain assumptions can be made to result in a worst-case result. The result is 'C)(x''' '' ') = system (Equations 5-54 through 5-56)] where t is the time since the release of the puff. The EPA9suggests that F stability can exist with wind speeds as low as 1. The weather conditions in the Pasquill-Gifford dispersion equations are included by means of the dispersion coefficients and the wind speed. then the highest concentration on the ground is found at a point downwind from the release. It is typically valid only for a distance of 0. For a release above ground level the puff center will move parallel to the ground and the maximum concentration on the ground will occur directly below the puff center. Some risk analysts use a wind speed of 2 m/s. The diameter of the isopleth increases initially as the puff travels downwind. Thus the maximum concentration is estimated by selecting the weather conditions and wind speed that result in the smallest values of the dispersion coefficients and the wind speed. Clearly. that is. so a finite value must be selected.1-10 km from the release point. 1996). For a puff the maximum concentration is always found at the puff center. it is readily evident that the dispersion coefficients and wind speed are in the denominator. By examining the Pasguill-Gifford dispersion equations for estimating the concentrations. and then decreases in diameter.

5-3 Dense Gas Dispersion A dense gas is defined as any gas whose density is greater than the density of the ambient air through which it is being dispersed. is the density of ambient air (mass/volume). . Also required is the wind speed at a height of 10 m. Actual instantaneous concentrations may vary by as much as a factor of 2 from the concentrations computed using Gaussian models. The model requires a specification of the initial cloud volume. a cloud having similar vertical and horizontal dimensions (near the source) may form. g is the acceleration due to gravity (length/time2). E.5-3 Dense Gas Dispersion 195 The concentrations predicted by the Gaussian models are time averages. The first step is to determine whether the dense gas model is applicable. is the initial density of released material (mass/volume). 1°R. After sufficient dilution occurs. p. The initial cloud buoyancy is defined as where g. Atmospheric stability was found to have little effect on the results and is not a part of the model. Workbook on the Dispersion of Dense Gases (Sheffield. Thus it is possible for instantaneous local concentrations to exceed the average values predicted . The release is assumed to occur at ambient temperature and without aerosol or liquid droplet formation. normal atmospheric turbulence predominates over gravitational forces and typical Gaussian dispersion characteristics are exhibited. Considerable initial dilution occurs because of the gravity-driven intrusion of the cloud into the ambient air. Subsequently the cloud height increases because of further entrainment of air across both the vertical and the horizontal interfaces. increasing its diameter and reducing its height. and p. the initial plume volume flux. The models presented here assume a 10-minute time average. McQuaid. The Britter and McQuaid10model was developed by performing a dimensional analysis and correlating existing data on dense cloud dispersion. and the initial gas density. Most of the data came from dispersion tests in remote rural areas on mostly flat terrain. The dense cloud slumps toward the ground under the influence of gravity. Following a typical puff release. The model is best suited for instantaneous or continuous ground-level releases of dense gases. the distance downwind. and the ambient gas density. This result can be due to a gas with a molecular weight greater than that of air or a gas with a low temperature resulting from autorefrigeration during release or other processes. the duration of release. 1988). United Kingdom: Health and Safety Executive. is the initial buoyancy factor (length/time2).this might be important for emergency response. Thus the results are not applicable to areas where terrain effects are significant. Britter and J.

. is the initial plume volume flux for dense gas dispersion (volume/time). Tables 5-4 and 5-5 provide equations for the correlations in these figures. The criteria for a sufficiently dense cloud to require a dense cloud representation are. for instantaneous releases. for continuous releases. For instantaneous releases the characteristic source dimension is defined as where Di is the characteristic source dimension for instantaneous releases of dense gases (length) and V. is the characteristic source dimension for continuous releases of dense gases (length).is the initial volume of released dense gas material (length3). The criteria for determining whether the release is continuous or instantaneous is calculated using the following group: where R. is the release duration (time) and x is the downwind distance in dimensional space (length).196 Chapter 5 Toxic Release and Dispersion Models A characteristic source dimension. and u is the wind speed at 10 m elevation (lengthltime). q. . dependent on the type of release. For continuous releases where D. If these criteria are satisfied. can also be defined. and. then Figures 5-13 and 5-14 are used to estimate the downwind concentrations.

5-3 Dense Gas Dispersion 197 Figure 5-13 Britter-McQuaid dimensional correlation for dispersion of dense gas plumes. - - 10' - - I 0. .1 I Passive limit I data region Figure 5-14 Britter-McQuaid dimensional correlation for dispersion of dense gas puffs.

concentration is used. If the value lies in-between.6. or most pessimistic. then the release is considered instantaneous. then the nonisothermal effects are assumed negligible. then the calculation providing the maximum. If the difference between the two calculations is small. For the first calculation a correction term is applied to the initial concentration (see Example 5-3). For gases lighter than air (such as methane or liquefied natural gas) the second calculation might be meaningless. For the second calculation heat addition is assumed at the source to bring the source material to ambient temperature. which provides a limit to the effect of heat transfer. If the group value is less than or equal to 0. then the concentrations are calculated using both continuous and instantaneous models and the maximum concentration result is selected. For nonisothermal releases the Britter-McQuaid model recommends two slightly different calculations.5. If the two calculations are within a factor of 2.Chapter 5 Toxic Release and Dispersion Models Table 5-4 Equations Used to Approximate the Curves in the Britter-McQuaidCorrelations Provided in Figure 5-13 for Plumes Concentration ratio (CmICo) Valid range for P = log [-I x (qo/u)'" If this group has a value greater than or equal to 2. If the . then the dense gas release is considered continuous.

and wind speed profiles. ground roughness. then the maximum. The model is also unable to account for the effects of parameters such as release height. These models are also capable of including the effects resulting from transient changes in toxic concentrations. the model is based only on data from flat rural terrain and is applicable only to these types of releases. but further investigation using more detailed methods (such as a computer code) may be worthwhile. are overly conservative and are designed for worker exposures. based on a correlation developed from experimental data. concentration is selected. Unfor- . or most pessimistic. However. discussed in chapter 2.5-4 Toxic Effect Criteria 199 Table 5-5 Equations Used to Approximate the Curves in the Britter-McQuaid Correlations Provided in Figure 5-14 for Puffs Concentration ratio (CmICo) Valid range for difference is very large (greater than a factor of 2). the question arises: What concentration is considered dangerous? Concentrations based on TLV-TWA values. The Britter-McQuaid model is a dimensional analysis technique. One approach is to use the probit models developed in chapter 2. not short-term exposures under emergency conditions. 5-4 Toxic Effect Criteria Once the dispersion calculations are completed.

The following paragraphs define these criteria and describe some of their features. To date. and the data show wide variations from the correlations. Three concentration ranges are provided as a consequence of exposure to a specific substance: 1. One simplified approach is to specify a toxic concentration criterion above which it is assumed that individuals exposed to this value will be in danger.and short-term human exposures. ERPGs are prepared by an industry task force and are published by the AIHA. ERPG-3 is the maximum airborne concentration below which it is believed nearly all individuals could be exposed for up to 1hr without experiencing or developing life-threatening health effects (similar to EEGLs). Because of the comprehensive effort to develop acute toxicity values. NIOSH publishes IDLH concentrations to be used as acute toxicity measures for common industrial gases. ERPG-1 is the maximum airborne concentration below which it is believed nearly all individuals could be exposed for up to 1 hr without experiencing effects other than mild transient adverse health effects or perceiving a clearly defined objectionable odor. These criteria and methods are based on a combination of results from animal experiments. 2. and toxic endpoints promulgated by the EPA as part of the RMP. ERPGs are becoming an acceptable industry1 government norm. ERPG-2 is the maximum airborne concentration below which it is believed nearly all individuals could be exposed for up to 1 hr without experiencing or developing irreversible or other serious health effects or symptoms that could impair their abilities to take protective action. emergency exposure guidance levels (EEGLs) and short-term public emergency guidance levels (SPEGLs) issued by the National Academy of Sciences/National Research Council.200 Chapter 5 Toxic Release and Dispersion Models tunately. This approach has led to many criteria promulgated by several government agencies and private associations. observations of long. and expanded by an AIHA peer review task force. 3. and expert judgment. An IDLH exposure condition is defined as a condition "that poses a threat of . IDLH levels established by NIOSH. PELS promulgated by OSHA. ERPG data are shown in Table 5-6. updated. toxicity dispersion (TXDS) methods used by the New Jersey Department of Environmental Protection. Some of these criteria and methods include emergency response planning guidelines (ERPGs) for air contaminants issued by the American Industrial Hygiene Association (AIHA). TLVs established by the ACGIH. 47 ERPGs have been developed and are being reviewed. published correlations are available for only a few chemicals. including short-term exposure limits (TLV-STELs) and ceiling concentrations (TLV-Cs).

000 1000 4000 125 (continued) .5-4 Toxic Effect Criteria Table 5-6 Emergency Response Planning Guidelines (ERPGs)l (all values are in ppm unless otherwise noted) Chemical Acetaldehyde Acrolein Acrylic acid Acrylonitrile Ally1 chloride Ammonia Benzene Benzyl chloride Bromine 1.3-Butadiene n-Butyl acrylate n-Butyl isocyanate Carbon disulfide Carbon tetrachloride Chlorine Chlorine trifluoride Chloroacetyl chloride Chloropicrin Chlorosulfonic acid Chlorotrifluoroethylene Crotonaldehyde Diborane Diketene Dimethylamine Dimethylchlorosilane Dimethyl disulfide Epichlorohydrin Ethylene oxide Formaldehyde Hexachlorobutadiene Hexafluoroacetone Hexafluoropropylene Hydrogen chloride Hydrogen cyanide Hydrogen fluoride Hydrogen sulfide Isobutyronitrile 2-Isocyanatoethyl methacrylate Lithium hydride Methanol Methyl chloride Methylene chloride Methyl iodide ERPG-1 ERPG-2 ERPGQ 1000 3 750 75 300 1000 1000 25 5 5000 250 1 500 750 20 10 10 3 30 mg/m3 300 50 3 50 500 25 250 100 500 25 30 50 500 100 25 50 100 200 1 500 pglm3 5.

Also note that IDLH levels have not been peer-reviewed and that no substantive documentation for the values exists."~] IDLH values also take into consideration acute toxic reactions. Publication 94-116 (Washington. and sulfuric acid) Sulfur dioxide Tetrafluoroethylene Titanium tetrachloride Toluene Trimethylamine Uranium hexafluoride Vinyl acetate - 'AIHA. that could prevent escape. For flammable vapors the IDLH concentration is defined as one-tenth of the lower flammability limit (LFL) concentration. IDLH data are currently available for 380 materials. The IDLH level is considered a maximum concentration above which only a highly reliable breathing apparatus providing maximum worker protection is permitted. Exposure to concentrations at the EEGL may produce transient irritation or central nervous sys" N I O S H . or aerosol that is judged acceptable and that allows exposed individuals to perform specific tasks during emergency conditions lasting from 1to 24 hr. If IDLH values are exceeded. . all unprotected workers must leave the area immediately. Since the 1940s. vapor. disabled. Because IDLH values were developed to protect healthy worker populations. such as severe eye irritation. they must be adjusted for sensitive populations.202 Chapter 5 Toxic Release and Dispersion Models Table 5-6 Chemical (continued) ERPG-1 ERPG-2 ERPG-3 Methyl isocyanate Methyl mercaptan Methyltrichlorosilane Monomethylamine PerAuoroisobutylene Phenol Phosgene Phosphorus pentoxide Propylene oxide Styrene Sulfonic acid (oleum. NIOSH Pocket Guide to Chemical Hazards. Emergency Response Planning Guidelines and Workplace ~ivironmental Exposure Levels American Industrial Hygiene Association. such as older. the National Research Council's Committee on Toxicology has submitted EEGLs for 44 chemicals of special concern to the Department of Defense. DC: US Department of Health and Human Services. (Fairfax. sulfur trioxide. 1994). An EEGL is defined as a concentration of a gas. or ill populations. VA: exposure to airborne contaminants when that exposure is likely to cause death or immediate or delayed permanent adverse health effects or prevent escape from such an en~ironment. 1996).

and (3) the methods by which EEGLs and SPEGLs were developed are well documented in National Research Council publications. Certain (ACGIH) criteria may be appropriate for use as benchmarks. The LOC is considered "the maximum concentration of an extremely hazardous substance in air that will not cause serious irreversible health effects in the general population when exposed to the substance for relatively short duration. (2) the median lethal concentration (LC50) value from animal test data multiplied by 0. SPEGLs. OSHA-cited "acceptable ceiling concentrations. These were developed specifically to apply to general populations and to account for sensitive populations and scientific uncertainty in toxicologic data. EEGL and SPEGL values are shown in Table 5-7. the most directly relevant toxicologic criteria currently available. For incidents involving substances for which no I2EPA. defined as acceptable concentrations for exposures of members of the general public. The ACGIH threshold limit values . These criteria are discussed in chapter 2. are ERPGs. These criteria can be used for toxic gas dispersion but typically produce a conservative result because they are designed for worker exposures.5-4 Toxic Effect Criteria 203 tem effects but should not produce effects that are lasting or that would impair performance of a task. as proposed by the New Jersey Department of Environmental Protection. (1) the ERPG-2 or (2) the level of concern (LOC) promulgated by the Emergency Planning and Community Right-to-Know Act." "excursion limits. the National Research Council has developed SPEGLs.1. particularly for developing emergency response plans. as required by the New Jersey Toxic Catastrophe Prevention Act (TCPA)." Toxic endpoints are provided for 74 chemicals under the RMP rule and are shown in Table 5-8. In general. such effects include irritation and narcosis. These levels are similar to the ACGIH criteria for TLV-TWAs because they are also based on 8-hr time-weighted average exposures." or "action levels" may be appropriate for use as benchmarks. and EEGLs. in order of preference. 1996).are designed to protect workers from acute effects resulting from exposure to chemicals. . The EPA has promulgated a set of toxic endpoints to be used for air dispersion modeling for toxic gas releases as part of the EPA RMP. An acute toxic concentration (ATC) is defined as the concentration of a gas or vapor of a toxic substance that will result in acute health effects in the affected population and 1fatality out of 20 or less (5% or more) during a 1-hr exposure. The New Jersey Department of Environmental Protection uses the TXDS method of consequence analysis to estimate potentially catastrophic quantities of toxic substances. The advantages of using EEGLs and SPEGLs rather than IDLH values are (1) a SPEGL considers effects on sensitive populations. SPEGLs are generally set at 10-50% of the EEGL and are calculated to take account of the effects of exposure on sensitive heterogeneous populations. DC: US Environmental Protection Agency. The PELS are promulgated by OSHA and have force of law. ATC values.12 The toxic endpoint is.TLV-STELs and TLV-Cs . or (3) the IDLH value. RMP Offsite Consequence Analysis Guidance (Washington. (2) EEGLs and SPEGLs are developed for several different exposure durations. In addition to EEGLs. are estimated for 103 "extraordinarily hazardous substances" and are based on the lowest value of one of the following: (1) the lowest reported lethal concentration (LCLO) value for animal test data.

1-hr EEGL 24-hr EEGL Source NRC I NRC I N R C IV N R C VII NRC I N R C VI N R C 111 NRC I N R C IV N R C I1 N R C I1 NRC I N R C I1 N R C I1 N R C I1 NRC V N R C I1 N R C IV N R C VI NRC I NRC V N R C VII N R C IV N R C I1 N R C VII N R C VIII NRC I NRC I N R C IV NRC V N R C IV N R C IV NRC I N R C I1 N R C I1 N R C I1 NRC I N R C VII N R C VIII NRC T I N R C I1 N R C TI NRC T I 400 15 mglm" 100 pg lm" 200 0.000 1 0.l -Dimethylhydrazine Ethanolamine Ethylene glycol Ethylene oxide Fluorine Hydrazine Hydrogen chloride Hydrogen sulfide Isopropyl alcohol Lithium bromide Lithium chromate Mercury (vapor) Methane Methanol Methylhydrazine Nitrogen dioxide Nitrous oxide Ozone Phosgene Sodium hydroxide Sulfur dioxide Sulfuric acid Toluene Trichloroethylene Trichlorofluoromethane Trichlorotrifluoroethane Vinylidene chloride Xylene ISPEGL value.204 Chapter 5 Toxic Release and Dispersion Models Table 5-7 Emergency Exposure Guidance Levels (EEGLs) from the National Research Council (NRC) (all values are in ppm unless otherwise noted) - Compound Acetone Acrolein AIuminum oxide Ammonia Arsine Benzene Bromotrifluoromethane Carbon disulfide Carbon monoxide Chlorine Chlorine trifluoride Chloroform Dichlorodifluoromethane Dichlorofluoromethane Dichlorotetrafluoroethane l.24 1' 10.2 2 mg/m3 10 1 mg/m3 200 200 ppm 1500 1500 .

020 0.l-Dimethylhydrazine Epichlorohydrin Ethylenediamine Ethyleneimine Furan Hydrazine Iron.0019 0.14 0.10 0.0011 0.5-4 Toxic Effect Criteria 205 Table 5-8 Toxic Endpoints Specified by the EPA Risk Management Plan1 Toxic endpoint (mg/L) 0.026 0.022 0.028 0.0037 0.0087 0.26 - 0.0040 0.0012 0.0078 0.0070 0.012 0.010 0.011 0.0030 0.029 'EPA.0070 0. 1996) . DC: US Environmental Protection Agency.0092 0.0065 0.028 0.01 Chemical name Gases Ammonia (anhydrous) Arsine Boron trichloride Boron trifluoride Chlorine Chlorine dioxide Cyanogen chloride Diborane Ethylene oxide Fluorine Formaldehyde (anhydrous) Hydrocyanic acid Hydrogen chloride (anhydrous) Hydrogen fluoride (anhydrous) Hydrogen selenide Hydrogen sulfide Methyl chloride Methyl mercaptan Nitric oxide Phosgene Phosphine Sulfur dioxide (anhydrous) Sulfur tetrafluoride Liquids Acrolein Acrylonitrile Acrylyl chloride Allyl alcohol Allylamine Arsenuous trichloride Boron trifluoride compound with methyl ether (1 :1) Bromine Carbon disulfide Chloroform Chloromethyl ether Chloromethyl methyl ether Crotonaldehyde Chemical name Liquids (continued) Cyclohexylamine Dimethyldichlorosilane 1.023 0.076 0. pentacarbonylIsobutyronitrile Isopropyl chloroformate Methacrylonitrile Methyl chloroformate Methyl hydrazine Methyl isocyanate Methyl thiocyanate Methyltrichlorosilane Nickel carbonyl Nitric acid (100%) Peracetic acid Perchloromethylmercaptan Phosphorus oxychloride Phosphorus trichloride Piperidine Propionitrile Propyl chloroformate Propyleneimine Propylene oxide Sulfur trioxide Tetramethyllead Tetranitromethane Titanium tetrachloride Toluene 2.00044 0.076 0.030 0.030 0.0012 0.011 0.085 0.042 0.12 0.049 0.018 0.49 0.00081 0.82 0.00067 0.0039 0.031 0.16 0.0018 0.018 0.6-diisocyanate Toluene diisocyanate (unspecified) Trimethylchlorosilane Vinyl acetate monomer - Toxic endpoint (mg/L) 0.016 0.026 0.010 0.0070 0.00090 0.4-diisocyanate Toluene 2.0076 0.0045 0.0094 0.49 0. RMP Offsite Consequence Analysis Guidance (Washington.00025 0.0019 0.012 0.0028 0.0032 0.16 0.00066 0.0011 0.59 0.0027 0.0035 0.14 0.090 0.0040 0.050 0.036 0.010 0.

Davis. J. Davis. "Alternative Guideline Limits for Chemicals without Environmental Response Planning Guidelines. K. D.5 ppm. Hansen. DeVore. the IDLH level for chlorine dioxide is 5 ppm. A. Craig. J. This hierarchy is shown in Table 5-9. the EPA suggests that the identification of an effect zone should be based on exposure levels of one-tenth the IDLH level. Hansen. J. persons located outside the zone that is based on these limits can be assumed to be unaffected by the release. Effect zones resulting from the release of this gas are defined as any zone in which the concentration of chlorine dioxide is estimated to exceed 0. SPEGLs or EEGLs are available. 56. S. R. Petrocchi. and T.Chapter 5 Toxic Release and Dispersion Models Table 5-9 Primary guideline ERPG-3 Recommended Hierarchy of Alternative Concentration Guidelines1 Hierarchy of alternative guidelines EEGL (30-minute) IDLH Source AIHA NRC NIOSH AIHA NRC EPAIFEMAIDOT OSHA ACGIH ACGIH AIHA OSHA ACGIH ACGIH ERPG-2 EEGL (60 minute) LOC PEL-C TLV-C 5 X TLV-TWA ERPG-3 PEL-STEL TLV-STEL 3 X TLV-TWA AIHA: American Industrial Hygiene Association NIOSH: National Institute for Occupational Safety and Health NRC: National Research Council Committee on Toxicology EPA: Environmental Protection Agency FEMA: Federal Emergency Management Agency DOT: US Department of Transportation OSHA: US Occupational Safety and Health Administration ACGIH: American Conference of Governmental Industrial Hygienists ID. Of course. DeVore. J. IDLH levels provide alternative criteria. J. D. a more realistic approach is to use a constant-dose assumption for releases less than 30 min using the IDLH level." AIHA Journal (1995). Petrocchi. A. . 13D. For example. and T. J. Powell. because IDLH levels were not developed to account for sensitive populations and because they were based on a maximum 30-min exposure period. "Alternative Guideline Limits for Chemicals without Environmental Response Planning Guidelines. However.I3 provided a hierarchy of alternative concentration guidelines in the event that ERPG data are not available. 56. In general. J. K. the approach is conservative and gives unrealistic results." AIHA Journal (1995). S. Craig. R. Powell. J. such as sensory irritation or odor perception. The use of TLV-STELs and ceiling limits may be most appropriate if the objective is to identify effect zones in which the primary concerns include more transient effects. Craig et al.

we obtain b. Using Table 5-2: Substituting into Equation 5-51. The mean concentration 50 m crosswind is found by using Equation 5-50 and by setting y = 50. The ground concentration directly downwind is given by Equation 5-51: From Table 5-1 the stability class is D. The dispersion coefficients are obtained from either Figure 5-11 or Table 5-2. Good judgement should prevail. c. The wind speed is 6 m/s. The location and value of the maximum mean concentration on ground level directly downwind. The mean concentration of SO2 on the ground 500 m downwind. Determine a.5-4 Toxic Effect Criteria 207 These methods may result in some inconsistencies because the different methods are based on different concepts. The mean concentration on the ground 500 m downwind and 50 m crosswind. This is a continuous release. The location of the maximum concentration is found from Equation 5-53: . Solution a. b. The stack is located in a rural area. Example 5-1 On an overcast day a stack with an effective height of 60 m is releasing sulfur dioxide at the rate of 80 gls. The results from part a are applied directly: c.

Assume in all cases that the chlorine cloud released is neutrally buoyant (which might not be a valid assumption).208 Chapter 5 Toxic Release and Dispersion Models From Figure 5-10 for D stability. For a distance of 500 m and a wind speed of 2 m/s. b. The maximum concentration of chlorine in the residential area. the time required for the center of the cloud to reach the residential area is t = x 500m .0 ppm. at a point 5 km directly downwind on the ground. u 2 m/s This leaves very little time for emergency warning. The maximum concentration occurs at the center of the cloud directly downwind from the release. A residential area is 500 m away from the chlorine source. This requires dispersion coefficients of minimum value. A source model study indicates that for a particular accident scenario 1. Determine a. From Figure 5-12 the lowest value of either dispersion coefficient occurs with F stability conditions. Assume the conditions of part b.. Compare this with an ERPG-1 for chlorine of 1. The concentration is given by Equation 5-41: The stability conditions are selected to maximize (C)in Equation 5-41. d. The maximum concentration is determined using Equation 5-52: Example 5-2 Chlorine is used in a particular chemical process. This is for nighttime conditions with thin to light .= ---. b. Determine the distance the cloud must travel to disperse the cloud to a maximum concentration below the ERPG-I.0 kg of chlorine will be released instantaneously. a has this value at about 1200 m downwind. a = 88 m. What stability conditions and wind speed produces the maximum concentration? c. Determine the size of the cloud.250 s = 4. The release will occur at ground level. based on the ERPG-1. Assume a wind speed of 2 mls. Use the conditions of part b. . From Figure .2 min. The time required for the center of the cloud to reach the residential area. Solution a. 5-10 or Table 5-2.

From Table 2-7 the ERPG-1 of 1. 0. Thus for F stability Solving for x by trial and error results in x = 8. 1- The time required for the center of the plume to arrive is At a downwind distance of x = 5 km = 5000 m and assuming F stability conditions.0 X kg/m3. Thus From Equation 5-41 This is converted to ppm using Equation 2-6. the concentration in ppm is 798 ppm.a. we obtain The distance downwind is solved using the equations provided in Table 5-3. we calculate .0 km downwind. Assuming a pressure of 1 atm and a temperature of 298 K.0 mg/m3or 3. This is much higher than the ERPG-1 of 1. This occurs at a distance of 500m.0 ppm. t ) = Q: 1 x-ut f i ~ ~ ~ ~ aexp[-?(_) . The concentration at the center of the cloud is given by Equation 5-41. c. The maximum concentration in the puff also occurs at the closest point to the release in the residential area.0 ppm is 3. Substituting the known values.a. d. The downwind centerline concentration is given by Equation 5-40: ( C ) ( x .0.5-4 Toxic Effect Criteria 209 overcast and a wind speed less than 3 m/s. Any individuals within the immediate residential area and any personnel within the plant will be excessively exposed if they are outside and downwind from the source.

2m [ 11' The cloud is 87. . LNG vapor density at boiling point of -162°C: 1. The following data are available: Spill rate of liquid: 0. based on the ERPG-1 concentration.76 kg/m3. Example 5-314 Compute the distance downwind from the following liquefied natural gas (LNG) release to obtain a concentration equal to the lower flammability limit (LFL) of 5% vapor concentration by volume.): 174 s. United Kingdom: Health and Safety Executive. An appropriate emergency procedure would be to alert residents to stay indoors with the windows closed and ventilation off until the cloud passes. Workbook on the Dispersion of Dense Gases (Sheffield. Spill duration (R. 1988).210 Chapter 5 Toxic Release and Dispersion Models Substituting the numbers provided gives where x has units of meters.9 mls. The quantity (x . we obtain 1 x .2 ( 39. Britter and J.326 = exp -.6 kg/m3.5000) represents the width of the plume. LNG density: 425. Solving for this quantity.22 kg/m3. McQuaid. At 2 m/s it will take approximately to pass. An effort by the plant to reduce the quantity of chlorine released is also indicated. Solution The volumetric discharge rate is given by The ambient air density is computed from the ideal gas law and gives a result of 1.E. Wind speed at 10 m above ground (u): 10.8 m wide at this point. Thus from Equation 5-59: 14R. Assume ambient conditions of 298 K and 1atm.23 m3/s.5000 0.

= 0. Determine whether the release is considered continuous or instantaneous. From Figure 5-13.. and it follows that C. both in absolute temperature. and the quantity must be greater than 2. . Apply Figure 5-13 to determine the downwind distance. S t e p 4. Adjust the concentration for a nonisothermal release.019. Substituting the required numbers gives and it follows that for a continuous release The final distance must be less than this. For this case Equations 5-60 and 5-62 apply. Determine whether a dense cloud model applies. If the original concentration is C*. For our required concentration of 0. The initial concentration of gas C. Thus Co = 1.0.5 for a continuous release. then the effective concentration is given by where Tais the ambient temperature and Tois the source temperature.IC.019. the equation for Cgives an effective concentration of 0. For this case expression 5-64 applies.5-4 Toxic Effect Criteria 21 1 S t e p 1. S t e p 2.. Substituting the appropriate numbers gives and it is clear that the dense cloud model applies. The Britter-MacQuaid model provides an adjustment to the concentration to account for nonisothermal release of the vapor. Compute the dimensionless groups for Figure 5-13: and S t e p 5. is essentially pure LNG. S t e p 3.05.

and it is also buoyant. if the material is released above ground level.. The initial momentum and buoyancy change the effective height of release. Education. 31. Similarly. A release that occurs at ground level but in an upward spouting jet of vaporizing liquid has a greater effective height than a release without a jet. based on its upward velocity within the stack pipe. .26 m)(126) = 285 m. 1970). Bruce Turner. then the material will initially fall toward the ground until it disperses enough to reduce the cloud density. For smokestack releases Turner suggested using the empirical Holland formula to compute the additional height resulting from the buoyancy and momentum of the release: where AH. Both effects are demonstrated by the traditional smokestack release shown in Figure 5-15. u is the wind speed (in mls). 5-5 Effect of Release Momentum and Buoyancy Figure 5-6 indicates that the release characteristics of a puff or plume depend on the initial release momentum and buoyancy. This compares to an experimentally determined distance of 200 m. is the correction to the release height H. The material released from the smokestack contains momentum. p. and Welfare. Workbook of Atmospheric Dispersion Estimates (Cincinnati: U S Department of Health. I5D. d is the inside stack diameter (in m). because its temperature is higher than the ambient temperature. a release of vapor at a temperature higher than the ambient air temperature will rise because of buoyancy effects. increasing the effective height of the release. This demonstrates that dense gas dispersion estimates can easily be off by a factor of 2. - usis the stack gas exit velocity (in mls). P is the atmospheric pressure (in mb). Thus the material continues to rise after its release from the stack. The upward rise is slowed and eventually stopped as the released material cools and the momentum is dissipated.Chapter 5 Toxic Release and Dispersion Models and it follows that x = (2. - For heavier-than-air vapors. is the stack gas temperature (in K). and Tais the air temperature (in K). T.

Prugh and Robert W. reducing the time available for an emergency response procedure. Guidelines for Vapor Release Mitigation (New York: American Institute of Chemical Engineers. p. this impact can occur minutes after the initial release. 1988). f 5-6 Release Mitigation The purpose of the toxic release model is to provide a tool for performing release mitigation. Both the consequence and the probability must be included to assess risk. 2. Johnson. an effect model is used to estimate the impact of the release. which is a measure of the consequence. Table 5-10 contains a number of measures to mitigate a release. The example problems presented in this chapter demonstrate that a small release can result in significant downwind impact. o f h o t gases Figure 5-15 Smokestack plume demonstrating initial buoyant rise o hot gases. This is coupled to a dispersion model and subsequent models for fires or explosions." l 6 The release mitigation procedure is part of the consequence modeling procedure shown in Figure 4-1. and management should be the first issues considered in any release mitigation procedure. if the probability is low. Risk is composed of both consequence and probability. It is possible that a particular release incident might have high consequences. Finally. . Release mitigation is defined as "lessening the risk of a release incident by acting on the source (at the point of release) either (1) in a preventive way by reducing the likelihood of an event that could generate a hazardous vapor cloud or (2) in a protective way by reducing the magnitude of the release andlor the exposure of local persons or property. In addition. the effort might not be required. a source model is used to determine either the release rate or the total quantity released. Inherent safety. Clearly. 16Richard W. However.5-6 Release Mitigation 213 Gases mix C O O cool as t h e y N e u t r a l Buoyancy and d i l u t e with l a i r . leading to extensive plant mitigation efforts to reduce the consequence. After selection of a release incident. engineering design. it is better to prevent the release in the first place. Thus an estimate of the consequences of a release provides only half the total risk assessment.

and drills Major area Inherent safety Engineering design Management Early vapor detection and warning Countermeasures Emergency response 'Richard W. 1988). training. 1996).Chapter 5 Toxic Release and Dispersion Models Table 5-10 Release Mitigation Approaches1 Examples Inventory reduction: Less chemicals inventoried or less in process vessels Chemical substitution: Substitute a less hazardous chemical for one more hazardous Process attentuation: Use lower temperatures and pressures Plant physical integrity: Use better seals or materials of construction Process integrity: Ensure proper operating conditions and material purity Process design features for emergency control: Emergency relief systems Spill containment: Dikes and spill vessels Operating policies and procedures Training for vapor release prevention and control Audits and inspections Equipment testing Maintenance program Management of modifications and changes to prevent new hazards Security Detection by sensors Detection by personnel Water sprays Water curtains Steam curtains Air curtains Deliberate ignition of explosive cloud Dilution Foams On-site communications Emergency shutdown equipment and procedures Site evacuation Safe havens Personal protective equipment Medical treatment On-site emergency plans. . 2d ed. 1999). procedures. Prugh and Robert W. (New York: American Institute of Chemical Engineers. Johnson. Guidelines for Vapor Cloud Dispersion Models. Suggested Reading Vapor Cloud Modeling Guidelines for Consequence Analysis of Chemical Releases (New York: American Institute of Chemical Engineers. Guidelines for Vapor Release Mitigation (New York: American Institute of Chemical Engineers.

1999). A backyard barbeque grill contains a 20-lb tank of propane. 5-2. The propane-air mixture then arrives at the burner assembly. John H. Contaminated toluene is fed to a water wash system shown in Figure 5-16. Seinfeld. Release Mitigation Richard W. The toluene is pumped from a 50-gal drum into a countercurrent centrifugal extractor. . ch. Workbook of Atmospheric Dispersion Estimates (Cincinnati: US Department of Health. Lees. Loss Prevention in the Process Industries. Frank P. The propane leaves the tank through a valve and regulator and is fed through a 112-in rubber hose to a dual valve assembly. where it is burned. 2d ed. and 14.13. Education. ch. Problems 5-1. Johnson. Guidelines for Vapor Release Mitigation (New York: American Institute of Chemical Engineers. (London: Butterworths. Prugh and Robert W. Bruce Turner. The extractor W a t e r f r o m Supply a t 40 psig r Z Centrifugal Extractor C l e a n Toluene (50 Gallon Drum) Dirty W a t e r (50 G a l l o n Drum) Dirty Toluene (50 Gallon Drum) \ Figure 5-16 Toluene water wash process. 1988). 1986). D.Problems 215 International Conference and Workshop on Modeling the Consequences of Accidental Releases of Hazardous Materials (New York: American Institute of Chemical Engineers. 15 and 18. 1996). 12. 1970). Describe the possible propane release incidents for this equipment. After the valves the propane flows through a dual set of ejectors where it is mixed with air. and Welfare. Atmospheric Chemistry and Physics ofAir Pollution (New York: Wiley.

The washed toluene and contaminated water flow into 50-gal drums. What is the average concentration of oxides of nitrogen from this source directly downwind at a distance of 3 km on an overcast night with a wind speed of 7 m/s? Assume that this dump is a point ground-level source. It is a sunny day. 5-9. The location of the 5 mg/m3 isopleth. Determine. Estimate the amount of acrolein that must be instantaneously released at ground level to produce a concentration at the boundary of the residential area equal to the ERPG-1.- - - - - - 216 Chapter 5 Toxic Release and Dispersion Models 5-3.1 ppm) is located 1500 m from a residential area. The water enters the center of the extractor and flows to the periphery. determine the location of the 5 mg/m3 ground-level isopleth. 5-5. sunny day with a 5 mph wind. Reconsider Problem 5-10. Determine a number of release incidents for this equipment.0 X lops g/m3.urn diameter) at the rate of 750 lblhr from a 30-m stack. One facility uses a 500-lb bottle. 5-10. Diborane is used in silicon chip manufacture. 5-7. b. c. 5-8. Which way should you run with respect to the wind to minimize your exposure? An air sampling station is located at an azimuth of 203" from a cement plant at a distance of 1500 m. What is the concentration of particulates at the air sampling station when the wind is from 30" at 3 mls on a clear day in the late fall at 4:00 P. On a sunny day with a 2 mls wind the concentration of sulfur dioxide 200 m directly downwind is measured at 5.1 km perpendicular to the plume 2 km downwind. .? A storage tank containing acrolein (ERPG-1 = 0. 5-4.6 mg/m3. It is a clear. separates the water from the toluene by centrifugal force acting on the difference in densities. The cement plant releases fine particulates (less than 15 . Assume that the release is at ground level. 5-11. Consider again Problem 5-7. If the entire bottle is released continuously during a 20-min period. Assume now that the bottle ruptures and that the entire contents of diborane are released instantaneously. and the wind speed is 18 kmlhr. The contaminated toluene enters the extractor at the periphery and flows to the center.M. at 15 min after the release. A burning dump emits an estimated 3 g/s of oxides of nitrogen. A trash incinerator has an effective stack height of 100 m. but assume a continuous release at ground level. The concentration at the center of the cloud. What is the release rate required to produce an average concentration at the boundary to the residential area equal to the ERPG-I? The concentration of vinyl chloride 2 km downwind from a continuous release 25 m high is 1. Determine the average concentration 0. Also estimate the maximum sulfur dioxide concentration expected on the ground and its location downwind from the stack. The location of the vapor cloud. You have been suddenly enveloped by a plume of toxic material from a nearby chemical plant. Estimate the mass release rate (in gls) of sulfur dioxide from this stack. a. 5-6.

4 mph on a clear night. determine the distance downwind that must be evacuated. estimate the distance downwind that must be evacuated. d. Assuming all of the chemical is released during a 10-min period. nursing homes. You are required to develop an emergency plan for the community. e. Use a distance of 0. b. evacuation of the population must occur for areas where the vapor concentration exceeds the ERPG-1. g. 5-15. The total dosage received by an individual standing on the downwind axis at the 15min downwind location.000 Ib of solid acrylamide stored in a large bin. The liquid is near its boiling point. e. unless a smaller distance is indicated by part b.A chemical plant has 10. They report the following chemicals and amounts: 100 lb of hydrogen chloride and 100 gal of sulfuric acid. a. c. 5-12. estimate the area downwind that must be evacuated. estimate the area that must be evacuated. shopping centers. Determine evacuation routes for the areas surrounding the plant. Identify any high-risk intersections where accidents might occur. How far and long the cloud will need to travel to reduce the maximum concentration to 5 mg/m3. If evacuation must occur in areas where the particle concentration exceeds 110 mg/m3.You have been appointed emergency coordinator for the community of Smallville. The boiling rate of MIC has been estimated to be 20 lblmin.An 800-lb tank of chlorine is stored at a water treatment plant. A study of the release scenarios indicates that the entire tank contents could be released as vapor in a period of 10 min. About 20% by weight of the solid has a particle size less than 10 pm.A reactor in a pesticide plant contains 1000 lb of a liquid mixture of 50% by weight liquid methyl isocyanate (MIC). f. Identify any special concerns (schools. ABC Chemical Company is shown on the map. Determine which chemical presents the greater hazard to the community.Problems 217 d. If the wind speed is 3. A study of various release scenarios indicates that a rupture of the reactor will spill the liquid into a boiling pool on the ground. Identify locations that might be affected by a release incident at the plant or might contribute to the incident because of its proximity to the plant. Evacuation of the population must occur in areas where the vapor concentration exceeds ERPG-1. and the like) that appear in the transportation route vulnerable zone. 5-13.5 mi on either side of the route. Determine the vulnerable zone along the transportation routes identified in part d. . Without any additional information. For chlorine gas. shown in Figure 5-17. A scenario study indicates that all the fine particles could be airborne in a period of 10 min. 5-14. Determine transportation routes that will be used to transport hazardous materials into or out of the facility.

Chapter 5 Toxic Release and Dispersion Models r Church MI 5 2 Mega Shopping Center Muddy River -- A Home ---. . Determine how these people are to be moved and where they might be evacuated to. An average house. Identify the means required to warn the area. exchanges air with the surroundings equal to three times the volume of the house per hour (although wide variations are expected). Determine alternative traffic routes around the potential hazard. j.Derive Equation 5-43.H o s p i t a l MI 52 Little Railway -- Senior --citizens 7 City Hall -- ----- Picnic Area - Apple Road South Road Chemical Srnallville School 0 u in Miles 1 2 P r e v a i l i n g Wind Scale Figure 5-17 Map of Srnallville. Assume wellmixed behavior for the air.One response to a short-term release is to warn people to stay in their homes or offices with the windows closed and the ventilation off. k. N. a. Estimate the potential number of people evacuated during an emergency. with the windows closed.. h. Derive an equation for the concentration of chemical vapor within the house based on a parameter. 1 What other concerns might be important during a chemical emergency? . 5-17. and describe the content of an example warning message that could be used in an emergency at the facility. i. an initial zero concentration of vapor within the house. and a constant external concentration during the exposure period. equal to the number of volume exchanges per hour. 5-16. Determine the resources required to support the needs of parts g and h.

5-19.016 mg/L.9. What is the maximum hole diameter (in mm) in this tank that will result in a downwind concentration equal to the ERPG-1 at a downwind distance of 300 m? Assume 1 atm. 5-22. 2S°C. The emergency coordinator has decided that the appropriate emergency response to the immediate release of a toxic material is to alert people to stay in their homes. The toxic endpoint is specified in the RMP as 0. The dispersion calculations must be completed assuming F stability and 1. a. estimate the distance the plume must travel to reduce the concentration to 7.Problems 219 b. The coordinator has also indicated that homes 4000 m downwind must not be exposed to concentrations exceeding 0. What does this imply about the size of chemical inventories for chemical plants? 5-21. wind speed. Estimate the maximum instantaneous release of material (in kg) allowed for these specifications. The temperature is 90°F. If the wind is blowing at 2 mph and the plant is 1mi upwind from the community. each facility must determine the downwind distance to a toxic endpoint. Be sure to clearly state any assumptions about weather conditions.000 Pa). If the supply line ruptures. These results must be reported to the EPA and to the surrounding community. The EPA Risk Management Plan (RMP) defines a worst-case scenario as the catastrophic release of the entire process inventory in a 10-min period (assumed to be a continuous release). A tank of chlorine contains 1000 kg of chlorine at 50 bar gauge (1 bar = 100. what is the maximum time available to the plant personnel to stop or reduce the release to ensure that the concentrations within the homes do not exceed the 10 ppm value? 5-18. .10 mg/m3 of this material for any longer than 2 min.5 m/s wind speed. 5-20. A supply line (internal diameter = 0. It is an overcast day with a 9 mph wind. and that all the liquid chlorine vaporizes. c. A plant has a 100-lbtank of anhydrous hydrogen fluoride (molecular weight = 20). until the cloud has passed. c. Many plants have much larger vessels on site. One hundred pounds of HF is a small quantity. Comment on the viability of using a continuous release model for a 10-min release period. a molecular weight of chlorine of 70. Assume an overcast day with a 15 mph wind and a temperature of 80°F. etc. A vapor cloud with a maximum concentration of 20 ppm is moving through a community.493 in) containing chlorine gas is piped from a regulated supply at 50 psig. Estimate the evaporation rate and the distance affected downwind. The pool is approximately rectangular with dimensions of 20 ft by 30 ft.3 mglm3. Determine the distance downwind (in miles) to the toxic endpoint for an EPA worst-case release. Comment on how a larger quantity would affect the downwind distance and how this might affect the public's perception of your facility. Determine the time before the vapor concentration within an average house reaches 10 ppm. A tank has ruptured and a pool of benzene has formed. The release is near ground level. b. As part of the RMP rule. Define the plume boundary using the TLV-TWA of 10 ppm. with doors and windows closed.

Determine the maximum width of the plume (in ft) and the distance downwind (in mi) where it occurs. Use the Britter-McQuaid dense gas dispersion model to determine the distance to the 1% concentration for a release of chlorine gas. In this case the flow of gas has been calculated to be 1. and the flow through the relief is equivalent to the flow obtained through a 3-cm-diameter hole in the tank. The boiling point for the chlorine is -34"C. Determine the evaporation rate from the dike (in lbls). It is a cloudy night with a 5 mph wind. 5-27. the release occurs at ground level. c. Assume ambient conditions of 298 K and 1atm.76 kgls. at each point downwind. 5-28. 5-25. and isopleth locations (m). downwind centerline concentrations (ppm). It is predicted that if a particular pipeline were to rupture. 5-24. It is estimated that ammonia will release at the rate of 10 lbls if this pipeline ruptures. the leak occurred in a diked area and the liquid benzene is contained within the square 50 ft X 30 ft dike. and it is a clear night with a wind speed of 5. wind speed (mls).220 Chapter 5 Toxic Release and Dispersion Models 5-23. Assume that evacuation must occur in any location that exceeds the OSHA PEL. release height (m). pressure (atm). A cloud of material has formed downwind of the release. and isopleth concentration (ppm). and the density of the liquid at the boiling point is 1470 kg/m3. You have decided that anyone exposed to more than 100 ppm of ammonia must be evacuated until repairs are made. both y and z dispersion coefficients (m). Fortunately. Use a spreadsheet program to determine the location of a ground isopleth for a plume. The wind speed at 10 rn height is 10 mls. It is decided that anyone exposed to potential concentrations exceeding 500 ppm must be evacuated. In this case the relief device has a 3-cm diameter. Determine the distance downwind (in mi) that must be evacuated. Emergency plans are being formulated so that rapid action can be taken in the event of an equipment failure. Determine the distance downwind that must be evacuated (in km). . What evacuation distance downwind will you recommend? 5-26. For this release the hydrogen sulfide in the tank is at a pressure of 1MPa absolute and 2S°C. The spreadsheet output should include. You are developing emergency evacuation plans for the local community downwind of your plant. The spreadsheet should have specific cell inputs for release rate (gls). temperature (K). spatial increment (m). All areas downwind with a concentration exceeding 4 times the PEL must be evacuated. molecular weight of the released material. Assume that the release occurs over a duration of 500 s with a volumetric release rate of 1 m3/s. a. b. One scenario identified is the rupture of an ammonia pipeline. What recommendation will you make as to the evacuation distance downwind? Assume that the wind speed is 6 mph and that the sun is shining brightly.5 mls. A tank containing hydrogen sulfide gas (molecular weight 34) has been overpressured and the relief device has been opened. it would release ammonia at a rate of 100 Ibis. The temperature is 80°F and the ambient pressure is 1 atm. A pipeline carrying benzene has developed a large leak.

Use the following two cases for computations.5 kg Release height: 0 m Molecular weight of gas: 30 Ambient temperature: 298 K Ambient pressure: 1 atm Isopleth concentration: 1 ppm Atmospheric stability: F Run the spreadsheet for a number of different times. Answer the following questions: a. ambient temperature (K). wind speed (mls). The spreadsheet should contain specific cells for user input of the following quantities: time after release (s). The spreadsheet output should also include a graph of the isopleth location. The spreadsheet output should include. both y and z dispersion coefficients. molecular weight of released gas. Your submitted work should include a brief description of your method of solution. release height (m). downwind centerline concentration. and isopleth concentration (ppm). Use the spreadsheet for the following case: Release mass: 0. but release height = 10 m. and the maximum width of the plume. Compare the plume width. ambient pressure (atm).Problems 221 The spreadsheet should also have cells providing the downwind distance. 5-29. . and plot the maximum puff width as a function of distance downwind from the release. total release (kg). area. Develop a spreadsheet to determine the isopleths for a puff at a specified time after the release of material. Otherwise. outputs from the spreadsheet. at each point downwind. At what distance downwind does the puff reach its maximum width? b. For your spreadsheet construction. Comment on the difference between the two cases. and plots of the isopleth locations. we suggest that you set up the cells to move with the puff center. Estimate the total area swept out by the puff from initial release to dissipation. you will need a large number of cells. and isopleth distance off-center (+I-). all based on the isopleth values. and downwind distance for each case. the total area of the plume. downwind location. and assume worst-case stability conditions: Case a: Release rate: 200 g/s Release height: 0 m Molecular weight: 100 Temperature: 298 K Pressure: 1atm Isopleth concentration: 10 ppm Case b: Same as above. At what distance and time does the puff dissipate? c.

Your submitted work must include a single output of the spreadsheet for a total release of 5 kg. Use the total release quantity as a parameter. including the puff concentration profile and the percent fatalities. a plot of maximum puff width as a function of downwind distance. A fixed mass of toxic gas has been released almost instantaneously from a process unit. Supplemental information: Molecular weight of gas: 30 Temperature: 298 K Pressure: 1atm Release height: 0 Wind speed: 2 m/s Use a probit equation for fatalities of the form where Y is the probit variable. This is the concentration recorded at a fixed location as the cloud passes. and a complete discussion of your method and your results.00 instead of 2. and rerun your spreadsheet for a total release amount of 5 kg.222 Chapter 5 Toxic Release and Dispersion Models Your submitted work should include a description of your method of solution.. C is the concentration in ppm. a complete spreadsheet output at 2000 s after release. The release occurs at night with calm and clear conditions. You have been asked to determine the percentage of fatalities expected 2000 m downwind from the release. total quantity released. The concentration increases linearly to a maximum concentration C. Determine the percentage of fatalities at the 2000-m downwind location as a result of the passing puff. How sensitive are the results to this exponent? Hint: Assume that the puff shape and concentration profile remain essentially fixed as the puff passes. Change the concentration exponent value to 2. and T is the time interval (min). and then decreases linearly to zero. A particular release of chlorine gas has resulted in the concentration profile given in Figure 5-18 for the cloud moving downwind along the ground. a plot of the concentration profile for the 5-kg case vs. and the calculation of the total swept area.00. Record the results at enough points to provide an accurate plot of the percentage of fatalities vs. Vary the total release quantity to result in a range of fatalities from 0 to 100%.75 in the probit equation. Prepare a spreadsheet to calculate the concentration profile around the center of the puff 2000 m downwind from the release. a single output of the spreadsheet for a 5-kg release with a probit exponent of 2. quantity released. a plot of the percentage of fatalities vs. 5-30. the distance in meters from the center of the puff. . 5-31.

duration of exposure. 50. Set up your spreadsheet to include input parameters of cloud duration and maximum concentration. with this particular shape. Use your spreadsheet to draw a plot of the percentage of fatalities vs. 60. What conclusions can be drawn about the results? . as shown.Problems 223 Duration cmax Conc. Time Figure 5-18 Concentration profile for a chlorine gas release. and 100 ppm. passing a fixed location. The width of the cloud is represented by the duration or time to pass.70. Draw a curve on the plot for each of the maximum concentrations of 40.80. Develop a spreadsheet to compute the percentage of fatalities expected as a result of this cloud.

.

These elements are illustrated by the fire triangle. To prevent accidents resulting from fires and explosions.Organic solvents are the most common source of fires and explosions in the chemical industry. and procedures to reduce fire and explosion hazards. .Fires and Explosions hemicals present a substantial hazard in the form of fires and explosions. and an ignition source. We discuss procedures to reduce fire and explosion hazards in chapter 7. IJames C. The combustion of one gallon of toluene can destroy an ordinary chemistry laboratory in minutes. persons present may be killed. Large Property Damage Losses in the Hydrocarbon-Chemical1ndustry:A Thirty Year Review (New York: J.' Additional losses in life and business interruptions are also substantial. engineers must be familiar with the fire and explosion properties of materials. and toxic releases. Marsh & McLennan. ed.. H. In this chapter we cover the first two topics. with yearly property losses in the United States estimated at almost $300 million (1997 dollars). 1997). an oxidizer. shown in Figure 6-1. in that order (see chapter 1). The potential consequences of fires and explosions in pilot plants and plant environments are even greater. Chemical and hydrocarbon plant losses resulting from fires and explosions are substantial. c 6-1 The Fire Triangle The essential elements for combustion are fuel. the nature of the fire and explosion process. The three most common chemical plant accidents are fires. explosions. emphasizing definitions and calculation methods for estimating the magnitude and consequences of fires and explosions. Coco.

Two common examples of the three components of the fire triangle are wood. is the rapid exothermic oxidation of an ignited fuel. metal particles Gases: acetylene. and (3) the ignition source is not energetic enough to initiate the fire. perchloric acid Solids: metal peroxides. and a match. This means a fire will not occur if (1)fuel is not present or is not present in sufficient quantities. However. and gasoline. or vapor form. burning will occur. Various fuels. propane. \ Sides Ignition Source Fire. other. fibers.226 Chapter 6 Fires and Explosions A \ Fire: When A l l Are Connected / N o Fire: + Figure 6-1 The fire triangle. fluorine. ether. and ignition sources common in the chemical industry are Fuels Liquids: gasoline. and a spark. flames. The combustion always occurs in the vapor phase. wood dust. acetone. but vapor and liquid fuels are generally easier to ignite. (2) an oxidizer is not present or is not present in sufficient quantities. less obvious combinations of chemicals can lead to fires and explosions. or burning. The fuel can be in solid. heat . liquids are volatized and solids are decomposed into vapor before combustion. oxidizer. oxidizers. and an ignition source are present at the necessary levels. carbon monoxide. air. When fuel. hydrogen Oxidizers Gases: oxygen. chlorine Liquids: hydrogen peroxide. ammonium nitrite Ignition sources Sparks. pentane Solids: plastics. liquid. static electricity. air. nitric acid.

Practical experience has shown that this is not robust enough . At the flash point the vapor will burn but only briefly. A good example of how the energy release rate affects the consequences of an accident is a standard automobile tire. Autoignition temperature (AIT):A fixed temperature above which adequate energy is available in the environment to provide an ignition source. whereas explosions release energy rapidly. Part of the energy released is used to sustain the reaction. inadequate vapor is produced to maintain combustion. Fires can also result from explosions. current practice is to prevent fires and explosions by continuing to eliminate ignition sources while focusing efforts strongly on preventing flammable mixtures. The flash point generally increases with increasing pressure. and explosions can result from fires. The two most commonly used methods are open cup and closed cup. If the tire ruptures suddenly and all the energy within the compressed tire releases rapidly. the tire is harmlessly deflated.6-3 Definitions 227 In the past the sole method for controlling fires and explosions was elimination of or reduction in ignition sources. Flash point (FP):The flash point of a liquid is the lowest temperature at which it gives off enough vapor to form an ignitable mixture with air. the result is a dangerous explosion. depending on the physical configuration of the experimental equipment. 6-3 Definitions Some of the commonly used definitions related to fires and explosions are given in what follows. As a result. the fire point temperature is higher than the flash point. The compressed air within the tire contains energy. Firepoint: The fire point is the lowest temperature at which a vapor above a liquid will continue to burn once ignited. If the energy is released slowly through the nozzle. The open-cup flash point is a few degrees higher than the closed-cup flash point. Ignition: Ignition of a flammable mixture may be caused by a flammable mixture coming in contact with a source of ignition with sufficient energy or the gas reaching a temperature high enough to cause the gas to autoignite. 6-2 Distinction between Fires and Explosions The major distinction between fires and explosions is the rate of energy release. . Fires release energy slowly. Each method produces a somewhat different value. These definitions are discussed in greater detail in later sections. typically on the order of microseconds.the ignition energies for most flammable materials are too low and ignition sources too plentiful. There are several different experimental methods used to determine flash points. Combustion orfire: Combustion or fire is a chemical reaction in which a substance combines with an oxidant and releases energy.

Unconfined explosions are rarer than confined explosions because the explosive material is frequently diluted below the LFL by wind dispersion. These explosions are destructive because large quantities of gas and large areas are frequently involved. Explosion: An explosion is a rapid expansion of gases resulting in a rapidly moving pressure or shock wave. the mixture is too lean for combustion. The gas is dispersed and mixed with air until it comes in contact with an ignition source. The expansion can be mechanical (by means of a sudden rupture of a pressurized vessel). . Boiling-liquid expanding-vapor explosion (BLEVE):A BLEVE occurs if a vessel that contains a liquid at a temperature above its atmospheric pressure boiling point ruptures. the vapor pressure of the liquid within the tank increases and the tank's structural integrity is reduced because of the heating. the hot liquid volatilizes explosively. Mechanical explosion: An explosion resulting from the sudden failure of a vessel containing high-pressure nonreactive gas. If the tgnk ruptures. possibly followed by combustion or explosion of the vaporized cloud if it is combustible. The mixture is also not combustible when the composition is too rich. Unconfined explosion: Unconfined explosions occur in the open. A mixture is flammable only when the composition is between the LFL and the UFL. that is. Explosion damage is caused by the pressure or shock wave. Detonation: An explosion in which the reaction front moves at a speed greater than the speed of sound in the unreacted medium.228 Chapter 6 Fires and Explosions Flammability limits: Vapor-air mixtures will ignite and burn only over a well-specified range of compositions. Lower explosion limit (LEL) and upper explosion limit (UEL) are used interchangeably with LFL and UFL. Confined explosion: An explosion occurring within a vessel or a building. This type of explosion is usually the result of a flammable gas spill. Dejlagration: An explosion in which the reaction front moves at a speed less than the speed of sound in the unreacted medium. These are most common and usually result in injury to the building inhabitants and extensive damage. Many solid materials (including common metals such as iron and aluminum) become flammable when reduced to a fine powder. Dust explosion: This explosion results from the rapid combustion of fine solid particles. or it can be the result of a rapid chemical reaction. This type of explosion occurs when an external fire heats the contents of a tank of volatile material. As the tank contents heat. The mixture will not burn when the composition is lower than the lower flammable limit (LFL). when it is above the upper flammable limit (UFL). Commonly used units are volume percent fuel (percentage of fuel plus air). The subsequent BLEVE is the explosive vaporization of a large fraction of the vessel contents.

Typically. 6-4 Flammability Characteristics of Liquids and Vapors Flammability characteristics of some important organic chemicals (liquids and gases) are provided in appendix B. although experimental data do not always agree with this. A shock wave in open air is followed by a strong wind.. Figure 6-2 is a plot of concentration versus temperature and shows how several of these definitions are related. Not flammable Mists 1 / Flammable ( 1 Lower flammability limit Not flammable I I 1 Flashpoint temperature 1 Autoignition temperature (AIT) Temperature 1 Figure 6-2 Relationships between various flammability properties. . The pressure increase in the shock wave is so rapid that the process is mostly adiabatic. The flash point and flammability limits are not fundamental properties but are defined only by the specific experimental apparatus and procedure used. Shock wave: An abrupt pressure wave moving through a gas. The LFL theoretically intersects the saturation vapor pressure curve at the flash point. The exponential curve in Figure 6-2 represents the saturation vapor pressure curve for the liquid material. the combined shock wave and wind is called a blast wave. The autoignition temperature is actually the lowest temperature of an autoignition region.- - 6-4 Flammability Characteristics of Liquids and Vapors Saturation vapor pressure curve \. the UFL increases and the LFL decreases with temperature. Overpressure: The pressure on an object as a result of an impacting shock wave. The behavior of the autoignition region and the flammability limits at higher temperatures are not known.

. The test flame applicator is moved back and forth horizontally over the liquid sample. most modern flash point methods employ a closed-cup procedure. Eventually a temperature is reached at which the liquid is volatile enough to produce a flammable vapor. To prevent this. The problem with open-cup flash point procedures is that air movements over the open cup may change the vapor concentrations and increase the experimentally determined flash point. is required to produce a continuous flame.230 Chapter 6 Fires and Explosions Liquids The flash point temperature is one of the major quantities used to characterize the fire and explosion hazard of liquids. Flash point temperatures are determined using an open-cup apparatus. During heating. The liquid to be tested is placed in the open cup. A small flame is established on the end of a movable wand. shown in Figure 6-3. called the fire point temperature. For Figure 6-3 Cleveland open-cup flash point determination. The liquid temperature is measured with a thermometer while a Bunsen burner is used to heat the liquid. The temperature at which this first occurs is called the flash point temperature. a higher temperature. the wand is slowly moved back and forth over the open liquid pool. Note that at the fla'sh point temperature only a momentary flame occurs. and a momentary flashing flame occurs.

0 334.6 600.6-4 Flammability Characteristics of Liquids and Vapors 231 Table 6-1 Constants Used in Equation 6-1 for Predicting the Flash Point Chemical group Hydrocarbons Alcohols Amines Acids Ethers Sulfur Esters Ketones Halogens Aldehydes Phosphorus-containing Nitrogen-containing Petroleum fractions a 225. Satyanarayana and P. b." Journal of Hazardous Materials (1992). this apparatus a small. 32: 81-85.9 b 537. is the boiling point temperature of the material (K).5 416.1 230. Closed-cup methods typically result in lower flash points. Journal of Hazardous Materials (1992).4 c 2217 1780 1900 2970 2879 2297 2217 1908 2154 1970 1666 1645 1807 lK. is the flash point temperature (K). The shutter is then opened and the liquid is exposed to the flame. a. "Improved Equation to Estimate Flash Points of Organic Compounds.1 264. .1 700.7 237.0 414.9 238.9 449. They were able to fit the flash point for over 1200 compounds with an error of less than 1%using the equation where T.5 262.5 201.0 577.8 222. and T. Satyanarayana and Rao2 showed that the flash point temperatures for pure materials correlate well with the boiling point of the liquid.7 185.1 432.0 293. G.0 260. manually opened shutter is provided at the top of the cup. Rao.0 416. Rao.8 260. Satyanarayana and P. The liquid is placed in a preheated cup and allowed to sit for a fixed time period. Flash points can be estimated for multicomponent mixtures if only one component is flammable and if the flash point of the flammable component is known. and c are constants provided in Table 6-1 (K). In this case the flash point temperature is estimated by determining the temperature at which the vapor pressure of the flammable component in the mixture is equal to the pure component vapor pressure at its 2K. 32: 81-85.2 296.4 323. G.2 275.6 390.

the flash point of the solution is 20S°C.73 Mole fraction 0.g°F.37 0. Assuming a basis of 100 lb of solution. Temperature ( *c) Figure 6-4 Saturation vapor pressure for methanol . Experimentally determined flash points are recommended for multicomponent mixtures with more than o n e flammable component. or 68.232 Chapter 6 Fires and Explosions flash point.63 1.00 25 75 - Raoult's law is used to compute the vapor pressure (Pmt) pure methanol. we can construct the following: Pounds Water Methanol - Molecular weight 18 32 - Moles 1. and its vapor pressure at this temperature is 62 mm Hg. Example 6-1 Methanol has a flash point of 54"F. shown in Figure 6-4. Using a graph of the vapor pressure versus temperature.4 mm Hg.34 3. What is the flash point of a solution containing 75% methanol and 25% water by weight? Solution The mole fractions of each component are needed to apply Raoult's law. based on the partial presof sure required to flash: p = xpsat PSt = PIX = 6210.39 2.63 = 98.

Crowl. Mines (1891). 3H. This test is repeated with different concentrations to establish the range of flammability for the specific gas. A. Figure 6-5 shows the results for methane. Gases and Vapors Flammability limits for vapors are determined experimentally in a specially designed closed vessel apparatus (see Figure 6-14 on page 255). V. These mixture limits are computed using the Le Chatelier equation:3 where LFLi is the lower flammable limit for component i (in volume %) of component i in fuel and air. Process Safety Progress (1998). 8. The maximum explosion pressure is measured. Vapor-air mixtures of known concentration are added and then ignited. DC: US Bureau of Mines. Vapor Mixtures Frequently LFLs and UFLs for mixtures are needed. The flammability limits are defined at 1 psig maximum pressure. 1985). . M. ser. Le Chatelier. and n is the number of combustible species. yi is the mole fraction of component i on a combustible basis. Mashuga and D." Ann. Kuchta. Data from C. Investigation of Fire and Explosion Accidents in the Chemical. and J.6-4 Flammability Characteristics of Liquids and Vapors 233 LFL \ & . 17(3): 176-1 83. 19: 388-395. US Bureau of Mines Report 680 (Washington. and Fuel-Related Industries: A Manual. "Estimation of Firedamp by Flammability Limits. Mining. *a UFL Mashuga and Crowl Kuchta I l l ' l ' o l l l l r r l l l l r ' 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Methane volume % in air Figure 6-5 Maximum pressure for methane combustion in a 20-L sphere.

A. Proper usage of Le Chatelier's rule requires flammability limit data at the same temperature and pressure.1 4C.2 5.15 UFLi (vol. These assumptions were found to be reasonably valid at the LFL and less so at the UFL.0 0.61 0. The combustion kinetics of the pure species is independent and unchanged by the presence of other combustible species. 2.5 3.5% ethylene by volume? Solution The mole fractions on a fuel-only basis are calculated in the following table. The LFL and UFL data are obtained from appendix B. The number of moles of gas is constant." Process Safety Progress. is the upper flammable limit for component i (in volume %) of component i in fuel and air. %) 7. The derivation shows that the following assumptions are inherent in this equation: The product heat capacities are constant. Le Chatelier's equation is empirically derived and is not universally applicable.0% methane. 19(2): 112-117.24 0.8 2. (2000). Also. flammability data reported in the literature may be from disparate sources.5 15 32.8% hexane.3 3. where UFL. The adiabatic temperature rise at the flammability limit is the same for all species. %) 0. Mashuga and Crow14derived Le Chatelier's equation using thermodynamics.0 1. Example 6-2 What are the LFL and UFL of a gas mixture composed of 0. . Mashuga and D. V.7 Mole fraction on combustible LFLi basis (vol. Crowl. and 0.Chapter 6 Fires and Explosions Similarly. Volume % Hexane Methane Ethylene Total combustibles Air 0. Combining data from these disparate sources may cause unsatisfactory results. which may not be obvious to the user. with wide variability in the data.3 96. "Derivation of Le Chatelier's Mixing Rule for Flammable Limits.

G. it is flammable. the flammability range increases with temperature. 1959). "Flame Temperatures of Limit Mixtures. p. Equation 6-3 is used to determine the UFL of the mixture: = 12. Scott. 5M. 484." in Seventh Symposium on Combustion (London: Butterworths. Lambiris. Because the mixture contains 3. Flammability Limit Dependence on Temperature In general. and G. S. is the net heat of combustion (kcallmole) and T is the temperature ("C).75% by volume total combustibles.3% total combustibles. Zabetakis. .363 = 2.9% by volume total combustibles. S.6-4 Flammability Characteristics of Liquids and Vapors 235 Equation 6-2 is used to determine the LFL of the mixture: = 110.5 The following empirically derived equations are available for vapors: where AH.

. = UFL + 20. Engr. Chem.0% by volume at 0. 22(1): 1-26. (1938). "Fire and Explosion Hazards at Temperature and Pressure Extremes. (1965).6(log P + I). Flammability limits are easily measured. 2.. Jones7 found that for many hydrocarbon vapors the LFL and the UFL are a function of the stoichiometric concentration (C. . Proc.0 MPa gauge. experimental determination is always recommended. Symp. = UFL + 20. Symp.) of fuel: I I LFL = 0. Extreme Cond. The UFL is determined using Equation 6-6: UFL. Chem. 7G. % fuel in air. The UFL increases significantly as the pressure is increased.. where flames do not propagate.." AICHE Inst. where P i s the pressure (megapascals absolute) and UFL is the upper flammable limit (volume % of fuel plus air at 1 atm).55CS. what is the UFL at 6. I 1 where C.2 + 0. UFL = 3. Rev. ser.50Cs." Chem.6(log P + 1 ) = 48 vol. The stoichiometric concentration for most organic compounds is determined using the general combustion reaction 6M. 99-104.301 MPa. broadening the flammability range. pp. Jones. An empirical expression for the UFL for vapors as a function of pressure is a ~ a i l a b l e : ~ UFL. is volume % fuel in fuel plus air. Example 6-3 If the UFLfor a substance is 11. W.236 Chapter 6 Fires and Explosions Flammability Limit Dependence on Pressure Pressure has little effect on the LFL except at very low pressures ( 4 0 mm Hg absolute). Zabetakis. Estimating Flammability Limits For some situations it may be necessary to estimate the flammability limits without experimental data. G. Engr.101 = 6..2 MPa gauge? Solution The absolute pressure is P = 6. "Inflammation Limits and Their Practical Application in Hazardous Industrial Operations.

" Fire and Materials (1994). 18: 333-336. Suzuki and K.~ bustion of the fuel. nitrogen. + #T.6-4 Flammability Characteristics of Liquids and Vapors 237 It follows from the stoichiometry that where z has units of moles O.76m + 1 . 1 9 ~ 2 . Suzuki. oxygen. 9T.80. 3 8 ~+ 1 - Another m e t h ~ dcorrelates the flammability limits as a function of the heat of com~. (6-12 ) (6-13) = 6. . 23.0538AHz + 1. A good fit was obtained for 30 organic materials containing carbon. hydrogen. 18: 393-397.42 AH.30AHC + 0 ./mole fuel. "Empirical Relationship Between Lower Flammability Limits and Standard Enthalpies of Combustion of Organic Compounds. and sulfur. Koide. 5 6 7 A ~ .5.50(100) 4. The resulting correlations are LFL UFL =- -3. "Correlation between Upper Flammability Limits and Thermochemical Properties of Organic Compounds. Additional stoichiometric and unit changes are required to determine C. as a function of z : cst = moles fuel moles fuel + moles air x 100 - 100 moles air moles fuel 100 1 - + (A) ( moles fuel O2 ) Substituting z and applying Equations 6-7 and 6-8 yields LFL = UFL = 3.. + 0.569AHC + 0." Fire and Materials (1994).

% actual.76(6) UFL = + 1.19(14) + 11 = 7. % versus 7. 3. % fuel in air). If the heat of combustion is provided in kcallmol.5 vol. and compare the calculated limits to the actual values determined experimentally.57 vol.238 Chapter 6 Fires and Explosions where LFL and UFL are the lower and upper flammable limits (vol.19(14) + 11 1.76(6) = + 1. The prediction capability of Equations 6-6 through 6-13 is only modest at best.55(100)/[4. Example 6-4 Estimate the LFL and the UFL for hexane.19 vol. and y are found by balancing this chemical reaction using the definitions in Equation 6-9: The LFL and the UFL are determined by using Equations 6-10 and 6-11: LFL = 0. is the heat of combustion for the fuel (in lo3 kJ/mol). Solution The stoichiometry is and z . and AH. This is an especially useful result. For hydrogen the predictions are poor.184.9-23%. x .2 vol. oxygen is the key ingredient and there is a minimum oxygen concentration required to propagate a flame. However. Thus these methods should be used only for a quick initial estimate and should not replace actual experimental data.5(100)/[4. 6-5 Limiting Oxygen Concentration and lnerting The LFL is based on fuel in air. % actual. Equation 6-13 is applicable only over the UFL range of 4. m. For methane and the higher hydrocarbons the results are improved. it can be converted to kJ/mol by multiplying by 4. % versus 1. because explosions and fires can be prevented by reducing the oxygen concentration regardless of the . respectively.

concentration of the fuel.5 11. the LOC is estimated using the stoichiometry of the combustion rzaction and the LFL. .5 11. the maximum safe oxygen concentration (MSOC). Venting of Deflagrations (Quincy.6-5 Limiting Oxygen Concentration and lnerting 239 Table 6-2 Limiting Oxygen Concentrations (LOCs) (volume percent oxygen concentration above which combustion can occur)l Gas or vapor Methane Ethane Propane n-Butane Isobutane n-Pentane Isopentane n-Hexane n-Heptane Ethylene Propylene 1-Butene Isobutylene Butadiene 3-Methyl-1-butene Benzene Toluene Styrene Ethylbenzene Vinyltoluene Diethylbenzene Cyclopropane Gasoline (731100) (1001130) (1151145) 10 11. Below the limiting oxygen concentration (LOC) the reaction cannot generate enough energy to heat the entire mixture of gases (including the inert gases) to the extent required for the self-propagation of the flame.5 N2/Air C02/Air Gas or vapor Kerosene JP-1 fuel JP-3 fuel JP-4 fuel Natural gas n-Butyl chloride Methylene chloride Ethylene dichloride Methyl chloroform Trichloroethylene Acetone t-butanol Carbon disulfide Carbon monoxide Ethanol 2-Ethyl butanol Ethyl ether Hydrogen Hydrogen sulfide Isobutyl formate Methanol Methyl acetate N2/Air C02/Air 'Data from NFPA 68. 1994).5 12 10. The LOC has also been called the minimum oxygen concentration (MOC). MA: National Fire Protection Association. This procedure works for many hydrocarbons. The LOC depends on the inert gas species. and other names. The LOC has units of percentage of moles of oxygen in total moles. This concept is the basis for a common procedure called inerting (see chapter 7). Table 6-2 contains LOC values for a number of materials. If experimental data are not available.

The tick marks on the scales show the direction in which the scale moves across the figure. The stoichiometric line represents all stoichiometric combinations of fuel plus oxygen. a mixture of this composition is not flammable. and inert material (in volume or mole %) are plotted on the three axes. Each apex of the triangle represents either 100% fuel. Concentrations of fuel. The UFL and the LFL are shown as the intersection of the flammability zone boundary with the air line. 6-6 Flammability Diagram A general way to represent the flammability of a gas or vapor is by the triangle diagram shown in Figure 6-6. oxygen. 20% oxygen. which is the composition of pure air. From the stoichiometry moles fuel (total = moles) moles 0. is not recommended because any condition that condenses water would move the oxygen concentration back into the flammable region. and 20% nitrogen. or even water vapor until the oxygen concentration is below 12. The air line in Figure 6-6 represents all possible combinations of fuel plus air. oxygen.4%. The coml~ustion reaction can be written in the form Fuel + z 0 2 +combustion products. (6-14) . moles 0. The addition of water.9% by volume. The air line intersects the nitrogen axis at 79% nitrogen (and 21 % oxygen). carbon dioxide. we obtain The combustion of butane is preventable by adding nitrogen. Because point A lies outside the flammable zone. Solution The stoichiometry for this reaction is The LFL for butane (from appendix B) is 1. (moles fuel) = LFL( moles fuel ) By substitution. however.240 Chapter 6 Fires and Explosions Example 6-5 Estimate the LOC for butane (C4HI0). The zone enclosed by the dashed line represents all mixtures that are flammable. or nitrogen. Thus point A represents a mixture composed of 60% methane.

Source: C. V. and the mole or volume percent of oxygen is given by expression 6-15. Thus the flammability limits and the LOC also change with these parameters. A. Mashuga and D. any gas mixture containing oxygen below the LOC is not flammable. Clearly. "Application of the Flammability Diagram for Evaluation of Fire and Explosion Hazards of Flammable Vapors. pressure. Crowl. The shape and size of the flammability zone on a flammability diagram change with a number of parameters. including fuel type. The intersection of the stoichiometric line with the oxygen axis (in volume % oxygen) is given by The stoichiometric line is drawn from this point to the pure nitrogen apex. and inert species. The LOC is also shown in Figure 6-6. where z is the stoichiometric coefficient for oxygen. Thus the moles present are fuel (1 mole) plus oxygen (2 moles). . 17(3): 176-1 83." Process Safety Progress (1998). The total moles are thus 1 + z. temperature. Expression 6-15 is derived by realizing that on the oxygen axis no nitrogen is present.6-6 Flammability Diagram 241 0 20 40 60 100 NITROGEN Figure 6-6 Flammability diagram for methane at an initial temperature and pressure of 25°C and 1 atm.

For systems having composition points that fall on a straight line passing through an apex corresponding to one pure component. the mixture is flammable as the gas composition moves through the flammability zone. at infinite dilution the mixture composition is at point S. then a fire or explosion would result. consider a storage vessel containing pure methane whose inside walls must be inspected as part of its periodic maintenance procedure. If two gas mixtures R and S are combined. At this point the vessel contains 100% methane. This is discussed in more detail in chapter 7. For example. The LOC can be estimated by reading the oxygen concentration at the intersection of the stoichiometric line and a horizontal line drawn through the LFL (see appendix C). the resulting mixture composition lies on a line connecting the points R and S on the flammability diagram. pure air. An inerting procedure can be used to avoid the flammability zone for both cases. The procedure is reversed for placing the vessel back into service. 3. For this operation the methane must be removed from the vessel and replaced by air for the inspection workers to breathe. 4. the mixture composition follows along the straight line between points R and S on the flammability diagram. Note that at some point in this operation the gas composition passes through the flammability zone. Diagrams with experimental data . Again. with the vessel containing air. the other two components are present in a fixed ratio along the entire line length. This is equivalent to the equation LOC = z(LFL). The first step in the procedure is to depressurize the vessel to atmospheric pressure. The location of the final mixture on the straight line depends on the relative moles in the mixtures combined: If mixture S has more moles. If the vessel is closed and methane is pumped in. then the gas composition inside the vessel will follow the air line and finish at point A. This is identical to the lever rule used for phase diagrams. The determination of a complete flammability diagram requires several hundred tests using a specific testing apparatus (see Figure 6-14 on page 255). In this case the procedure begins at point B in Figure 6-7. If an ignition source of sufficient strength were present. represented by point A in Figure 6-7. If the vessel is opened and air is allowed to enter. the mixture composition moves closer and closer to point S. These results show that: 1.242 Chapter 6 Fires and Explosions Appendix C derives several equations that are useful for working with flammability diagrams. As the dilution continues. the final mixture point will lie closer to point S. These results are useful for tracking the gas composition during a process operation to determine whether a flammable mixture exists during the procedure. the composition of gas within the vessel will follow the air line in Figure 6-7 until the vessel gas composition eventually reaches point B. If a mixture R is continuously diluted with mixture S. Eventually. 2.

the flammability zone is much larger for ethylene than for methane.6-6 Flammability Diagram 243 A = Pure methane 100 NITROGEN Figure 6-7 Pure air: 79% N2 21% O2 = B The gas concentration during an operation to remove a vessel from service. This was done to obtain a more complete understanding of the pressure versus time behavior of the combustion over a wide range of mixtures. A number of important features are shown in Figures 6-8 and 6-9. the lower boundary of the flammability zone is mostly horizontal and can be approximated by the LFL. For most systems detailed experimental data of the type shown in Figure 6-8 or 6-9 are unavailable. Second. the combustion produces copious amounts of soot in the upper fuel-rich parts of the flammability zone. the UFL for ethylene is correspondingly higher. Several methods have been developed to approximate the flammability zone: . Data in the center region of the flammability zone are not available because the maximum pressure exceeds the pressure rating of the vessel or because unstable combustion or a transition to detonation is observed there. Finally. in accordance with ASTM E918. For these data a mixture is considered flammable if the pressure increase after ignition is greater than 7% of the original ambient pressure. This information is important for mitigation of the explosion. First. respectively. for methane and ethylene are shown in Figures 6-8 and 6-9. Note that many more data points are shown than required to define the flammability limits.

Mashuga.69 psia Initial temperature: 25°C Reactor volume: 20 liters lgnitor type: 1 cm 40 AWG SnCu / 500VA lgnitor energy: 10 J Ignitor location: Center Figure 6-8 Experimental flammability diagram for methane. 1999). Use expression 6-15 to locate the stoichiometric point on the oxygen axis.Chapter 6 Fires and Explosions 0 20 40 60 NITROGEN 80 100 Experimental conditions Initial pressure: 14. Method 1 (Figure 6-10): Given the flammability limits in air. The flammability zone derived from this approach is only an approximation of the actual zone. Ph. Draw a point at this intersection. the LOC. Draw flammability limits in pure oxygen as points on the oxygen scale. and draw the stoichiometric line from this point to the 100% nitrogen apex. 3. 4. This . 5. Note that the lines defining the zone limits in Figures 6-8 and 6-9 are not exactly straight. Draw flammability limits in air as points on the air line. (Source: C. Connect all the points shown. and flammability limits in pure oxygen. the procedure is as follows: 1.D. Locate the LOC on the oxygen axis. and draw a line parallel to the fuel axis until it intersects with the stoichiometric line. V. dissertation (Michigan Technological University. 2.

3. .6-6 Flammability Diagram 245 NITROGEN Experimental conditions lgnitor type: 1 cm 40 AWG SnCu 1500VA Initial pressure: 14. V.data that are not readily available. and usually (but not always) provides a conservative LOC. the procedure is as follows: Use steps 1 and 3 from method 1. although it extends all the way to the oxygen axis and typically expands in size. This is only an estimate. Method 2 (Figure 6-11): Given the flammability limits in air and the LOC.Estimate the LOC using Equation 6-16.In this case only the points at the nose of the flammability zone can be connected. Mashuga.D. Flammability limits in pure oxygen for a number of common hydrocarbons are provided in Table 6-3. dissertation (Michigan Technological University. the procedure is as follows: Use steps 1. Method 3 (Figure 6-12): Given the flammability limits in air. The flammability zone from the air line to the oxygen axis cannot be detailed without additional data. The lower boundary can also be approximated by the LFL. Ph. (Source: C. and 4 from method 1. 1999).69 psia Initial temperature: 25°C Ignitor energy: 10 J Reactor volume: 20 liters Ignitor location: Center Figure 6-9 Experimental flammability diagram for ethylene. method also requires flammability limits in pure oxygen .

. Lewis and G. and Explosions of Gases (New York: Harcourt Brace Jovanovich. 2The limits are insensitive to pHd. 1987). von Elbe. Table 6-3 Flammability Limits in Pure Oxygen Limits of flammability in pure oxygen Compound Hydrogen Deuterium Carbon monoxide2 Ammonia Methane Ethane Ethylene Propylene Cyclopropane Diethyl ether Divinyl ether Formula Lower Upper 'Data from B.above a few mm Hg. Combustion. Flames.246 Chapter 6 Fires and Explosions Flammability limits 100 0 NITROGEN Figure 6-10 Method 1 for the approximation of the flammability zone.

. Only the area to the right of the air line can be determined.6-6 Flammability Diagram 247 NITROGEN LFL 100 Figure 6-1 1 Method 2 for the approximation of the flammability zone. Only the area to the right of the air line can be determined. NITROGEN LFL 100 Figure 6-12 Method 3 for the approximation of the flammability zone.

Many hydrocarbons have MIEs of about 0. 3d ed. a static discharge of 22 mJ is initiated by walking across a rug. The MIE depends on the specific chemical or mixture. Combustion. in general. the concentration. pressure. and an increase in the nitrogen concentration increases the MIE. as a result of fluid flow. contributing to plant explosions (see chapter 7).25 mJ. and temperature. Electrostatic discharges. (New York: Academic Press. Experimental data indicate that the MIE decreases with an increase in pressure. and an ordinary spark plug has a discharge energy of 25 mJ. Glassman. For example. the MIE of dusts is. A few MIEs are given in Table 6-4. 1996). also have energy levels exceeding the MIEs of flammable materials and can provide an ignition source. This is low compared with sources of ignition.3-Butadiene n-Butane Cyclohexane Cyclopropane Ethane Ethene Ethylacetate Ethylene oxide n-Heptane Hexane Hydrogen Methane Methanol Methyl acetylene Methyl ethyl ketone n-Pentane 2-Pentane Propane - 'Data from I. . at energy levels somewhat higher than combustible gases.248 Chapter 6 Fires and Explosions Table 6-4 Chemical Minimum Ignition Energy for Selected Gases1 Minimum ignition energy (mJ) Acetylene Benzene 1. All flammable materials (including dusts) have MIEs. 6-7 Ignition Energy The minimum ignition energy (MIE) is the minimum energy input required to initiate combustion.

pressure of the system. presence of catalytic material. insulation on a steam pipe saturated with certain polymers. Some examples of auto-oxidation with a potential for spontaneous combustion include oils on a rag in a warm storage area. . Composition affects the AIT. Liquids with high volatility are less susceptible to autoignition because they self-cool as a result of evaporation. sometimes leading to autoignition if the energy is not removed from the system. Safeguards must be included in the process design to prevent undesirable fires that can result from adiabatic compression. an increase in pressure decreases AITs. Liquids with relatively low volatility are particularly susceptible to this problem. allowing auto-oxidation and eventual autoignition). A compressor is particularly susceptible to autoignition if it has a fouled after-cooler. and flow conditions. The autoignition temperature is a function of the concentration of vapor. rich or lean mixtures have higher AITs. These examples illustrate why special precautions must be taken to prevent fires that can result from auto-oxidation and autoignition. This is the cause of preignition knock in engines that are running too hot and too lean.6-10 Adiabatic Compression 249 6-8 Autoignition The autoignition temperature (AIT) of a vapor. This strong dependence on conditions illustrates the importance of exercising caution when using AIT data. 6-9 Auto-Oxidation Auto-oxidation is the process of slow oxidation with accompanying evolution of heat. AIT data are provided in appendix B. is the temperature at which the vapor ignites spontaneously from the energy of the environment. It is also the reason some overheated engines continue to run after the ignition is turned off. gasoline and air in an automobile cylinder will ignite if the vapors are compressed to an adiabatic temperature that exceeds the autoignition temperature. and increases in oxygen concentration decrease AITs. subsequent compression resulted in autoignition. referred to as spontaneous combustion. volume of vapor. It is essential to experimentally determine AITs at conditions as close as possible to process conditions. Larger system volumes decrease AITs. 6-10 Adiabatic Compression An additional means of ignition is adiabatic compression. and filter aid saturated with certain polymers (cases have been recorded where 10-year-old filter aid residues were ignited when the land-filled material was bulldozed. Many fires are initiated as a result of auto-oxidation. sometimes called the spontaneous ignition temperature (SIT). For example. Several large accidents have been caused by flammable vapors being sucked into the intake of air compressors.

3)(14.4. and y = C. is the final absolute pressure.7 psia to 500 psia if the initial temperature is 100°F? The AIT of hexane is 487°C (appendix B). Pi is the initial absolute pressure. resulting in an explosion. Example 6-7 The lubricating oil in piston-type compressors is always found in minute amounts in the cylinder bore. . we obtain = 17. The actual compression ratio or pressure should be kept well below this. This ratio represents an output pressure of only (17. The potential consequences of adiabatic temperature increases within a chemical plant are illustrated in the following two examples. is the final absolute temperature. A particular lubricating oil has an AIT of 400°C. Assume an initial air temperature of 25°C and 1atm.3. and y for air is 1. Example 6-6 What is the final temperature after compressing air over liquid hexane from 14.250 Chapter 6 Fires and Explosions The adiabatic temperature increase for an ideal gas is computed from the thermodynamic adiabatic compression equation: where T. compressor operations must always be maintained well below the AIT of the oil to prevent explosion. P. Compute the compression ratio required to raise the temperature of air to the AIT of this oil. Solution From Equation 6-17 we have This temperature exceeds the AIT for hexane.IC". Solving for the compression ratio. Solution Equation 6-17 applies. Ti is the initial absolute temperature.7 psia) = 254 psia.

for example. lamps. etc. The main reason for rendering a flammable liquid inert. etc. This is especially important today. The goal is to eliminate or minimize ignition sources because the probability of a fire 1°Accident Prevention Manual for Industrial Operations (Chicago: National Safety Council.) Burner flames (improper use of torches.) Cutting and welding (sparks. etc.) Molten substances (hot spills) Chemical action (processes not in control) Static sparks (release of accumulated energy) Lightning (where lightning rods are not used) Miscellaneous 'Accident Prevention Manual for Industrial Operations (Chicago: National Safety Council. and the need for periodic preventive maintenance programs when working with flammable gases and compressors.000 fires by the Factory Mutual Engineering Corporation and are summarized in Table 6-5. 6-11 Ignition Sources lo As illustrated by the fire triangle. In these cases a thorough safety analysis is required to eliminate all possible ignition sources in each of the units where flammable gases are present.) Combustion sparks (sparks and embers) Spontaneous ignition (rubbish. arcs. The elimination of the ignition sources with the greatest probability of occurrence (see Table 6-5) should be given the greatest attention. . Various ignition sources were tabulated for over 25.6-11 Ignition Sources 251 Table 6-5 Ignition Sources of Major Fires1 Electrical (wiring of motors) Smoking Friction (bearings or broken parts) Overheated materials (abnormally high temperatures) Hot surfaces (heat from boilers. heat. careful monitoring of conditions. etc. is to prevent a fire or explosion by ignition from an unidentified source. 1974) These examples illustrate the importance of careful design. because high-pressure process conditions are becoming more common in modern chemical plants. Combinations of sources must also be investigated. etc. consequently it is impossible to identify and eliminate them all. crushers. Some special situations might occur in a process facility where it is impossible to avoid flammable mixtures.) Exposure (fires jumping into new areas) Incendiarism (fires maliciously set) Mechanical sparks (grinders. Although all sources of ignition are not likely to be identified. engineers must still continue to identify and eliminate them. 1974). The sources of ignition are numerous. fires and explosions can be prevented by eliminating ignition sources.

Mists and sprays also affect flammability limits. In this situation. explosion behavior is still not completely understood.01 mm. For chemical plants much of the damage from explosions is due to this pressure wave. In experiments with larger drop diameters the LFL was less than one-tenth of the normal LFL. Many approaches to the problem have been undertaken. The effort required increases significantly as the size of the plant increases. This is true even at low temperatures where the liquid is nonvolatile and no vapor is present. the LFL is virtually the same as the substance in vapor form. and acoustic energy. The damage from an explosion is caused by the dissipating energy. Borgoyne. When sprays have drop diameters between 0. including formation of a pressure wave. An explosion results from the rapid release of energy.H. Thus. 6-12 Sprays and Mists l1 Static electricity is generated when mists or sprays pass through orifices. thermal radiation.2 mm the LFL decreases as the drop diameter increases. a fire or explosion will occur. Mists of this type are formed by condensation. I1Frank P. A charge may accumulate and discharge in a spark. projectiles. flame propagation is impossible. Despite these efforts. including theoretical. which results in damage to the surroundings. This energy is then dissipated by a variety of mechanisms. Practicing engineers. If flammable vapors are present. The energy release must be sudden enough to cause a local accumulation of energy at the site of the explosion." Chemical Process Hazards (1965). The pressure wave contains energy. For mechanically formed mists with drop diameters between 0.12 For suspensions with drop diameters less than 0. If the explosion occurs in a gas. the energy causes the gas to expand rapidly. 2: 1. 2d ed. . 6-13 Explosions Explosion behavior depends on a large number of parameters.01 mm and 0. Loss Prevention in the Process Industries. the presence of small drops andlor disturbances that shatter the larger drops may create a hazardous condition. however. should use extrapolated results cautiously and provide a suitable margin of safety in all designs. we must understand the dynamics of the pressure wave. This is important when inerting in the presence of mists. "The Flammability of Mists and Sprays. A summary of the more important parameters is shown in Table 6-6. (Boston: Butterworths. in order to understand explosion impacts. Lees.5 mm. potential ignition sources may be in the thousands. and empirical studies.6 mm and 1.252 Chapter 6 Fires and Explosions or explosion increases rapidly as the number of ignition sources increases. semiempirical. 1996). 12J. Explosion behavior is difficult to characterize. therefore. forcing back the surrounding gas and initiating a pressure wave that moves rapidly outward from the blast source.

but it can also occur from the sudden rupture of a pressure vessel. . This compression occurs rapidly. A shock wave is expected from highly explosive materials. and duration Geometry of surroundings: confined or unconfined Amount of combustible material Turbulence of combustible material Time before ignition Rate at which combustible material is released A pressure wave propagating in air is called a blast wave because the pressure wave is followed by a strong wind. Figure 6-13 shows the physical differences between a detonation and a deflagration for a combustion reaction that occurs in the gas phase in the open.6-13 Explosions 253 Table 6-6 Parameters Significantly Affecting the Behavior of Explosions Ambient temperature Ambient pressure Composition of explosive material Physical properties of explosive material Nature of ignition source: type. The difference depends on whether the reaction front propagates above or below the speed of sound in the unreacted gases. A shock front is found a short distance in front of the reaction front. This is classified as a detonation. In some combustion reactions the reaction front is propagated by a strong pressure wave. Fundamentally. These processes are relatively slow. energy. resulting in a reaction front and leading shock wave that propagates into the unreacted mixture at or above the sonic velocity. such as TNT. For ideal gases the speed of sound or sonic velocity is a function of temperature only and has a value of 344 m/s (1129 ft/s) at 20°C. For a deflagration the energy from the reaction is transferred to the unreacted mixture by heat conduction and molecular diffusion. resulting in an abrupt pressure change or shock in front of the reaction front. The maximum pressure over ambient pressure is called the peak overpressure. the sonic velocity is the speed at which information is transmitted through a gas. For a detonation the reaction front moves at a speed greater than the speed of sound. A shock wave or shock front results if the pressure front has an abrupt pressure change. The reaction front provides the energy for the shock front and continues to drive it at sonic or greater speeds. causing the reaction front to propagate at a speed less than the sonic velocity. which compresses the unreacted mixture in front of the reaction front above its autoignition temperature. Detonation and Deflagration The damage effects from an explosion depend highly on whether the explosion results from a detonation or a deflagration.

Both fronts move at the same speed. The pressure fronts produced by detonations and deflagrations are markedly different. Distance . the reaction front moves at a speed greater than the speed of sound. These pressure fronts move away from the reaction front at the speed of sound and accumulate together in a main pressure front.--> DEFLAGRATION Reaction front In a deflagration. The pressure front resulting from a deflagration is characteristically wide (many milliseconds in duration). and total duration of typically less than 1ms. One way to conceptualize the resulting pressure front is to consider the reaction front as producing a series of individual pressure fronts. The explosion is initiated to the far left. The main pressure front will continue to grow in size as additional energy and pressure fronts are produced by the reaction front. Distance . For a deflagration the reaction front propagates at a speed less than the speed of sound. driving the shock front immediately preceding it. while the pressure front moves away from the reaction front at the speed of sound. with an abrupt pressure rise. A detonation produces a shock front.-> Figure 6-13 Comparison of detonation and deflagration gas dynamics.254 Chapter 6 Fires and Explosions DETONATION ------- h Peak overpressure Reaction front Shock front lArnbient pressure In a detonation. a maximum pressure of greater than 10 atm. The pressure front moves at the speed of sound in the unreacted gas and moves away from the reaction front. flat (without an abrupt . the reaction front moves at a speed less than the speed of sound.

A deflagration can also evolve into a detonation. the rate of pressure rise after the flammable mixture is ignited. Empirical studies have shown that the nature of the explosion is a function of several experimentally determined characteristics. and with a maximum pressure much lower than the maximum pressure for a detonation (typically 1 or 2 atm). Different behavior occurs if the fronts propagate in a closed vessel. and the maximum pressure after ignition. The gas dynamic behavior for complex geometries is beyond the scope of this text. These characteristics depend on the explosive material used and include flammability or explosive limits. resulting in an increase in the adiabatic pressure rise. These characteristics are determined using two similar laboratory devices. . Gas Manifold @ "8' I Figure 6-14 Magnetic Stirrer I Test apparatus for acquiring vapor explosion data. a pipeline. The behaviors of the reaction and pressure fronts differ from those shown in Figure 6-13 depending on the local geometry constraining the fronts. The pressure builds and results in a full detonation. or through a congested process unit. The transition is particularly common in pipes but unlikely in vessels or open spaces. The two most common confined explosion scenarios involve explosive vapors and explosive dusts. Confined Explosions A confined explosion occurs in a confined space. In a piping system energy from a deflagration can feed forward to the pressure wave.6-13 Explosions 255 shock front). such as a vessel or a building. This is called a deflagration to detonation transition (DDT). shown in Figures 6-14 and 6-17.

6 bar - . Typically.- dP dt 7'6 bar 0.-) ( I . The test procedure includes (1)evacuating the vessel. the maximum pressure and pressure rates occur somewhere within the range of flammability (but not necessarily at the same concentration). = 8. (4) igniting the gas by a spark.256 Chapter 6 Fires and Explosions - - - . By using this relatively simple set of experiments. - . the reaction is terminated at the wall. in this example the flammability limits are between 2% and 8%.80 . The rate of pressure rise is indicative of the flame front propagation rate and thus of the magnitude of the explosion. The pressure rate and maximum pressure for each run are plotted versus concentration. (2) adjusting the temperature.P . the explosive characteristics can be completely established. and the maximum rate of pressure rise is 360 barls. A typical pressure versus time plot is shown in Figure 6-15.120 - 3 . Explosion Apparatus for Dusts The experimental apparatus used to characterize the explosive nature of dusts is shown in Figure 6-17. and ( 5 ) measuring the pressure as a function of time.60 F -. The pressure within the vessel is measured by a transducer located on the external wall.the maximum pressure is 7. After ignition the pressure wave moves outward within the vessel until it collides with the wall. The device is similar to the vapor explosion apparatus.4 bar. as shown in Figure 6-16.. with the exception of a . Explosion Apparatus for Vapors The apparatus used to determine the explosive nature of vapors is shown in Figure 6-14. (3) metering in the gases to obtain the proper mixture.5 bar . The maximum pressure and maximum rate of pressure rise are determined.024 s : V) = 316 bar/s - 0 0 I I I I I I I I 20 40 60 80 100 120 140 160 0 180 Time (ms) Figure 6-15 Typical pressure versus time data obtained from gas explosion apparatus shown in Figure 6-14. The pressure rate or slope is computed at the inflection point of the pressure curve. as shown in Figure 6-15. The experiment is repeated at different concentrations.100 4 AP = 7. Experiments of this type usuallyresult in a deflagration with a few atmospheres of pressure rise.

) Explosion Pressure (bar-gauge) .Rate of Pressure Increase (bar/?.

Explosion Characteristics The explosion characteristics determined using the vapor and dust explosion apparatus are used in the following way: 1.and lowspeed pressure transducers.987 bar) within the sphere at ignition time. The maximum rate of pressure increase indicates the robustness of an explosion.258 Chapter 6 Fires and Explosions D i s t r i b u t i' o n Dust Sample Air I . The maximum pressure and the maximum rate of pressure increase are determined. The data are collected and analyzed in the same fashion as for the vapor explosion apparatus. carbon dioxide. or Halon) to stop the combustion process. 2. The maximum rate is also used to design a vent for relieving a vessel during an explosion before the pressure ruptures the vessel or to establish the time interval for adding an explosion suppressant (water. The air used to drive the dust into the sphere is carefully metered to ensure a pressure of 1 atm (0. Thus the explosive behavior of different materials can be compared on a relative basis. The limits of flammability or explosivity are used to determine the safe concentrations for operation or the quantity of inert material required to control the concentration within safe regions. The computer measures the pressure as a function of time using high. as well as the flammability limits. . discharged. I I lnput/Output Computer Devices I Figure 6-17 Test apparatus for acquiring dust explosion data. A typical pressure versus time plot from the dust explosion apparatus is shown in Figure 6-18.

the deflagration indexes KG and Kst increase..V1'3 = constant = KG. A plot of the logarithm of the maximum pressure slope versus the logarithm of the vessel volume frequently produces a straight line of slope -113. where KG and Ks..0 10. This relationship is called the cubic law: (dPldt)m.... The cubic law states that ---.Hydrogen 1 0... respectively. 5..0 Volume of vessel (m3) Figure 6-19 Typical explosion data exhibiting the cubic law.5 1O ..6-13 Explosions - - ------------- Peak pressure . As the robustness of an explosion increases. are the deflagration indexes for gas and dust... .. as shown in Figure 6-19.1 atm 0 0 50 150 100 Time (ms) 200 250 Figure 6-18 Pressure data from dust explosion device...

Equations 6-18 and 6-19 are used to estimate the consequences of an explosion in a confined space. Table 6-7 shows that good agreement is found between different investigations for the maximum pressure but that only limited agreement is found for the KG values. Senecal and P.2 76 NFPA 68. and KG and Kst data for vapors and dusts are shown in Tables 6-7 and 6-8.9 7. Explosions-Schutz: Grundlagen und Anwendung (New York: Springer-Verlag.1 7.9 NFPA 68 (1997) 109 10 92 105 115 106 78 94 659 45 64 5 104 96 8.2 171 638 67 46 94 7. the pressure front takes longer to propagate through a larger vessel.9 7.05 5.5 7.0 6.4 6.4 7.4 8..8 7.5 55 75 104 100 94 7.8 7.65 7." in 31st Loss Prevention Symposium (New York: American Institute of Chemical Engineers.7 7. Beaulieu. depending on the value of the deflagration index.260 Chapter 6 Fires and Explosions Table 6-7 Maximum Pressures and Deflagration Indexes for a Number of Gases and Vapors1 Maximum pressure Pmax g) (bar Senecal and Bartknecht Beaulieu (1993) (1998) Deflagration index KG(bar-m Is) Senecal and Bartknecht Beaulieu (1993) (1998) Chemical Acetylene Ammonia Butane Carbon disulfide Diethyl ether Ethane Ethyl alcohol Ethyl benzene Ethylene Hydrogen Hydrogen sulfide Isobutane Methane Methyl alcohol Methylene chloride Pentane Propane Toluene lData selected from: NFPA 68 (1997) 10.8 7. 1997). "KG:Data and Analysis. Dusts are further classified into four classes.6 5.4 6.1 7.0 92 7. MA: National Fire Protection Association. P. respectively. such as a building or a vessel. A.4 106 96 550 78 8.8 7.6 6. 1997) W.4 8. A. Bartknecht. as follows: .0 6..0 6.8 7. It is postulated that the KG values are sensitive to experimental configuration and conditions.0 7. Venting ofDeflagrations (Quincy. 1993). These St classes are shown in Table 6-8. J.

1 7.5 - 97 178 - - <1 - (continued) . Ks.9 10.6-13 Explosions 261 Table 6-8 St Classes for Dusts and Combustion Data for Dust Clouds - Deflagration index.4 - 112 90 57 - 8.0 5.8 9.5 6. wood.2 9. resins.0 8.2 8.3 - - 250 100 7 - 5.4 8.. (~m) (g/m3) (bar g) 44 51 33 80 <10 80 200 400 80 165 - Kst (bar-rn Is) 24 66 18 18 27 - Minimum ignition energy (mJ) 100 60 - 7.1 7.7 8.. food Dextrose Fructose Fructose Wheat grain dust Milk powder Rice flour Wheat flour Milk sugar Coal. peat Cotton Cellulose Wood dust Wood dust Paper dust Feed.4 7. auxiliary materials Adipinic acid Naphthalene Salicylic acid Median Minimum particle explosive dust size concentration P. coal products Activated carbon Bituminous coal Plastics.(bar rn 1s) 0 1-200 200 -300 >300 St class St-0 St-1 St-2 St-3 Dust Cotton.8 75 44 55 12 194 67 20 101 38 110 23 156 75 >lo0 540 14 - 250 - - 30 200 30 - - <30 <10 95 - 60 15 30 8.3 8.3 8. rubber Polyacrylamide Polyester Polyethylene Polyethylene Polypropylene Polypropylene Polystyrene (copolymer) Polystyrene (hard foam) Polyurethane Intermediate products.4 - 60 125 - - 180 >4000 - 50 10 18 <lo 10 <lo 72 280 25 162 155 760 3 60 60 60 60 60 - 9.4 7.3 6.7 4.

2 17. Dust explosions demonstrate unique behavior.1 5.5 7. It is therefore necessary to run the experiments as close as possible to the actual conditions under consideration.Chapter 6 Fires and Explosions Table 6-8 (continued) Median Minimum particle explosive dust size concentration P. 1997). These explosions occur if finely divided particles of solid material are dispersed in air and ignited. Equation 6-20 allows the experimental results from the dust and vapor explosion apparatus to be applied to determining the explosive behavior of materials in buildings and process vessels. and (4) the energy of the ignition source. and the maximum pressure and the maximum pressure rate are linearly dependent on the initial pressure.0 11.2 6. are not physical properties of the material because they are dependent on (1)the composition of the mixture. a point is reached where the deflagration turns into a detonation. K. As the initial pressure is increased. Eckoff. The constants KG and Ks. This is discussed in more detail in chapter 9. The subscript "experimental" applies to data determined in the laboratory using either the vapor or dust explosion apparatus.8 11. The dust particles can be either an unwanted by-product or the product itself. alloys Aluminum powder Aluminum powder Bronze powder Iron (from dry filter) Magnesium Magnesium Silicon Zinc (dust from collector) Other inorganic products Graphite (99.2 9. This is shown in Figure 6-20. .0 10. (3) the shape of the reaction vessel.. The spikes in the curves indicate a detonation. Dust Explosions in the Process Industries (Oxford: Butterworth-Heinemann. as shown in Figure 6-21.8 8.2 11. (~m) (g/m3) (bar g) Minimum ignition energy (mJ) Kst Dust Other technical.. Experimental studies indicate that the maximum explosion pressure is usually not affected by changes in volume. chemical products Organic dyestuff (blue) Organic dyestuff (red) Organic dyestuff (red) Metals.5% C) Sulfur Toner (bar-mIs) <lo <lo 52 <10 22 18 12 28 240 <lo <10 7 20 <lo - 50 60 60 30 750 500 30 500 125 250 <30 30 60 9.9 6.5 4. (2) the mixing within the vessel.7 5.9 73 249 237 515 110 31 50 508 12 126 125 71 151 196 - - - - 54 - - 4 'Data selected from R. The subscript "in vessel" is for the reactor or building.

6-13 Explosions 263 Initial Pressure (bar) Figure 6-20 Effect of initial pressure on maximum explosion pressure and rate. grain storage. . Data from W. Bartknecht. Explosions (New York: Springer-Verlag. 171266. p. and coal mining industries. Loss Prevention in the Process Industries. Accidents involving dust explosions can be quite substantial. 1981). Explosions involving dusts are most common in the flour milling. a series of grain silo explosions in Westwego near New Orleans in 1977 killed 35 people.13 'Tees.

. In this fashion the explosion leapfrogs its way through a plant. An initial dust explosion can cause secondary explosions. ~ ' I " " 3 0 V cn n w 1000. . possibly resulting in a secondary explosion. Data from W. . 1981). I ~ - - - I .cn w - 2 3 cn V) 2 a C 500- 0 w + r Y 0 Propane Content in Air (vol %) Figure 6-21 Explosion data for propane showing peaks indicative of the onset of detonation. Many times the secondary explosions are more damaging than the primary explosion. Explosions (New York: Springer-Verlag. stirring up additional dust.264 Chapter 6 Fires and Explosions Propane Content (vol %) 1500 n = ~ - - - - I - - . Bartknecht. The primary explosion sends a shock wave through the plant.

so its value is important for estimating damage. t. the shock front has arrived and a peak overpressure is observed. The time period t. Explosions (New York: Springer-Verlag. the dust loading must be reasonably uniform. After attaining the maximum underpressure t. the reaction front terminates. For most dustslqhe lower explosion limit is between 20 glm3 and 60 g/m%nd the upper explosion limit is between 2 kg/m%nd 6 kg/m3. is called the arrival time. For dust particles the particles are of varying size and many orders of magnitude larger than molecules. 15W. To be explosive. For dusts deflagrations appear to be much more common than detonations.14 The pressure waves from dust deflagrations. Loss Prevention in the Process Industries. the pressure will approach ambient pressure at t4. After the combustible material is consumed. Blast Damage Resulting from Overpressure The explosion of a dust or gas (either as a deflagration or a detonation) results in a reaction front moving outward from the ignition source preceded by a shock wave or pressure front. The decreasing pressure continues to drop below ambient pressure to a maximum underpressure at time t.Bartknecht. The pressure quickly decreases to ambient pressure at time t. are powerful enough to destroy structures and kill or injure people.. This time. The underpressure for large explosions and nuclear explosions. can be quite large. At t. Figure 6-22 shows the variation in pressure with time for a typical shock wave at a fixed location some distance from the explosion site. but the pressure wave continues its outward movement. The explosion occurs at time to. however. p. the particle loading must be between certain limits.There exists a small but finite time t. For most of the underpressure period from t. before the shock front travels from its explosive origin to the affected location. At this time the blast wind and the direct destruction have terminated. 1981). It is the blast wave that causes most of the damage. resulting in considerable damage.. however. 27. a dust mixture must have the following characteristics: the particles must be below a certain minimum size. A blast wave is composed of the pressure wave and subsequent wind.. is called the shock duration. I4Lees.6-13 Explosions 265 Dust explosions are even more difficult to characterize than gaseous explosions. The shock duration is the period of greatest destruction to free-standing structures. Gravity also affects dust particle behavior. but because the maximum underpressure is only a few psi for typical explosions. . For a gas the molecules are small and of well-defined size.. p. 171265. the damage is much less than that of the overpressure period. to t. immediately followed by a strong transient wind. to t4 the blast wind reverses direction and flows toward the explosive origin. There is some damage associated with the underpressure period. but the wind continues in the same direction for a short time.

As illustrated. Blast damage is based on the determination of the peak side-on overpressure resulting from the pressure wave impacting on a structure. Damage estimates based on overpressures are given in Table 6-9. At a fixed location. significant damage is expected for even small overpressures. In general.tl = shock duration t3 = maximum underpressure to ti t2 t3 Time t4 Figure 6-22 Blast wave pressure at a fixed location. Good estimates of blast damage. the side-on overpressure increases abruptly to its maximum value (peak side-on overpressure) and then drops off as the blast wave passes. however. An important consideration is how the pressure is measured as the blast wave passes. If the pressure transducer is at right angles to the blast wave. In general.266 Chapter 6 Fires and Explosions ti = arrival time t2 . overpressure implies the side-on overpressure and frequently the peak side-on overpressure. Many references report overpressure data without clearly stating how the overpressure is measured. The reflected overpressure includes the side-on overpressure and the stagnation pressure. then the measured pressure is the reflected overpressure. are obtained using just the peak side-on overpressure. If the pressure transducer is placed facing toward the oncoming shock wave. shown in Figure 6-22. The reflected overpressure is a maximum when the blast wave arrives normal to the wall or object of concern and decreases as the angle changes from normal. The stagnation pressure is due to deceleration of the moving gas as it impacts the pressure transducer. the overpressure measured is called the side-on overpressure (sometimes called the free-field overpressure). the damage is also a function of the rate of pressure increase and the duration of the blast wave. The reflected pressure for low side-on overpressures is about twice the side-on overpressure and can reach as high as 8 or more times the side-on overpressure for strong shocks. .

projectile limit. panels blow in Steel frame of clad building slightly distorted Partial collapse of walls and roofs of houses Concrete or cinder block walls.7 1.6 34.7-27.5-48.8 17. followed by buckling.7 20. "Diagnostic Features of Explosion Damage. 1972).95 of no serious damage below this value). .000 lb) survive Limit of crater lip 0. Clancey.28 0.5 3 3-4 4 5 5 -7 7 7.3 2. steel frame buildings distort and pull away from foundations Frameless.0 1-2 2. not reinforced.0 13. tall hydraulic presses (40.8 9 10 9.3 2 2-3 2. wood panels (standard housing).2 48.8-20. 10-15 Hz) Occasional breaking of large glass windows already under strain Loud noise (143 dB).1 0. made uninhabitable Corrugated asbestos shatters.-- - 6-13 Explosions Table 6-9 Damage Estimates for Common Structures Based on Overpressure (these values are approximation^)^ Pressure psig 0. shatter Lower limit of serious structural damage 50% destruction of brickwork of houses Heavy machines (3000 lb) in industrial buildings suffer little damage." paper presented at the Sixth International Meeting of Forensic Sciences (Edinburgh. rupture of oil storage tanks Cladding of light industrial buildings ruptures Wooden utility poles snap. 10% window glass broken Limited minor structural damage Large and small windows usually shatter.5-1. not reinforced.8 13.1 62.07 Damage Annoying noise (137 dB if of low frequency.14 0.6 27.03 0.9 6. fastenings fail.4-6.76 3. self-framing steel panel buildings demolished.2-55. fail by shearing or flexure Loaded train boxcars completely demolished Probable total destruction of buildings. sonic boom. 8-12 in thick.02 0.000 lb) in buildings slightly damaged Nearly complete destruction of houses Loaded train wagons overturned Brick panels.0 0.2 48.9-13.03 2. glass failure Breakage of small windows under strain Typical pressure for glass breakage "Safe distance" (probability 0.9 4. heavy machine tools (7000 lb) moved and badly damaged.0 68.5 34. J. fastenings fail.21 0.15 0. some damage to house ceilings.3 kPa 0.04 0.7 15.4 0.2 20.8 1. occasional damage to window frames Minor damage to house structures Partial demolition of houses.69 1.8 6.9 300 2068 IV. very heavy machine tools (12. corrugated steel or aluminum panels.

The scaled overpressure p. Baker. J. Experiments with explosives have demonstrated1* that the overpressure can be estimated using an equivalent mass of TNT. Glasstone. multiply by 0. is given by 16W.. Source: G. 1962). denoted r. and the distance from the ground-zero . Figure 6-23 provides a correlation for the scaled overpressure p. DC: U S Atomic Energy Commission. Graham. The empirically derived scaling law is The equivalent energy of TNT is 1120 callg.3967. Explosions in Air (Austin: University of Texas Press. Kinney and K. denoted m. with units of m/kg"3. The Effects of Nuclear Weapons (Washington. 1973).E. point of the explosion.1 1 10 100 Scaled distance.01 0. To convert ftllb"' to mlkgl''.. 1985). versus scaled distance z.268 Chapter 6 Fires and Explosions 0. Explosive Shocks in Air (Berlin: Springer-Verlag. z ( m ~ k ~ ' ' ~ ) . S. . F. Figure 6-23 Correlation between scaled distance and explosion peak side-on overpressure for a TNT explosion occurring on a flat surface.

(2) convert the energy to an equivalent amount of TNT. and pais the ambient pressure. and (4) use Table 6-9 to estimate the damage. Example 6-8 One kilogram of TNT is exploded.0 kg)"" = 30 m kg-"" From Figure 6-23 the scaled overpressure is 0. Most explosions occurring in chemical plants are considered to originate on the ground.6-13 Explosions 269 where p. The data in Figure 6-23 are also represented by the empirical equation The procedure for estimating the overpressure at any distance r resulting from the explosion of a mass of material is as follows: (1) Compute the energy of the explosion using established thermodynamic procedures. Solution The value of the scaling parameter is determined using Equation 6-21: - 30 m (1.81 psi). is the peak side-on overpressure.5. then the resulting side-on overpressure is estimated at (0. the resulting overpressures from Figure 6-23 are multiplied by 0. if the ambient pressure is 1 atm. well above the ground. is the scaled overpressure (unitless). The data in Figure 6-23 are valid only for TNT explosions occurring on a flat surface. The approach is based on the assumption that an exploding fuel . p. Thus.3kPa) = 5. (3) use the scaling law and the correlations of Figure 6-23 to estimate the overpressure.055)(101.6 kPa (0. For explosions occurring in the open air. Compute the overpressure at a distance of 30 m from the explosion.055. From Table 6-9 this overpressure will cause minor damage to house structures. TNT Equivalency TNT equivalency is a simple method for equating a known energy of a combustible fuel to an equivalent mass of TNT.

The procedure to estimate the damage associated with an explosion using the TNT equivalency method is as follows: 1. including incomplete mixing with air of the combustible material and incomplete conversion of the thermal energy to mechanical energy. the overpressure curve for TNT tends to overpredict the overpressure near the VCE and to underpredict at distances away from the VCE. with most flammable cloud estimates varying between 1% and lo%. as reported by a number of sources. respectively. The advantage to the TNT equivalency method is that it is easy to apply because the calculations are simple. 4. Others have reported 5%. Use the scaling law given by Equation 6-21 and Figure 6-23 (or Equation 6-23) to estimate the peak side-on overpressure. Vapor cloud explosions (VCEs) are explosions that occur because of the release of flammable vapor over a large volume and are most commonly deflagrations. and ErnT is the energy of explosion of TNT. m is the mass of hydrocarbon (mass). and 15% for flammable clouds of propane. ~ is the equivalent mass of TNT (mass).- - - - 270 Chapter 6 Fires and Explosions mass behaves like exploding TNT on an equivalent energy basis. Use Table 6-9 to estimate the damage for common structures and process equipment. 2. Determine the total quantity of flammable material involved in the explosion. As a result. The equivalent mass of TNT is estimated using the following equation: where r n . such as ammonium nitrate. . A typical value for the energy of explosion of TNT is 1120 callg = 4686 kJ1kg = 2016 Btullb. The procedure can be applied in reverse to estimate the quantity of material involved based on damage estimates. and acetylene. diethyl ether. and calculate the equivalent mass of TNT using Equation 6-24. The TNT equivalency method also uses an overpressure curve that applies to point source detonations of TNT. Estimate the explosion efficiency. AHc is the energy of explosion of the flammable gas (energylmass). 7 is the empirical explosion efficiency (unitless). lo%. The heat of combustion for the flammable gas can be used in place of the energy of explosion for the combustible gas. In addition. 3. the method is unable to consider the effects of flame speed acceleration resulting from confinement. The explosion efficiency is used to adjust the estimate for a number of factors. Explosion efficiencies can also be defined for solid materials. The explosion efficiency is one of the major problems in the equivalency method. The explosion efficiency is empirical.

(c) estimating the volumes of fuel-air mixture present in the individual areas identified as blast sources (this estimate can be based on the overall dimensions of the areas and jets. Identify potential sources of strong blast within the area covered by the flammable cloud. In general. tunnels. sewage systems. spaces between extended parallel planes (for example. The remaining fuel-air mixture in the flammable cloud is assumed to produce a blast of minor strength. The basis for this model is that the energy of explosion depends highly on the level of congestion and depends less on the fuel in the cloud. Assign a number representative of the blast strength for each individual blast. spaces within tubelike structures (for example. (b) assuming that the full quantities of fuel-air mixture present within the partially confined/obstructed areas and jets. Some companies have defined procedures for this. stacks of crates or pallets. contribute to the blasts. 5. because of the limitations of dispersion modeling in congested areas. and open buildings. 3. assigns a relative degree of confinement. those beneath closely parked cars in parking lots. 4. and (d) calculating the combustion energy E (J) for each blast by multiplying the individual volumes of the mixture by 3. Perform a dispersion model to determine the extent of the cloud. Potential sources of strong blast include congested areas and buildings. corridors. culverts). and an intensely turbulent fuel-air mixture in a jet resulting from release at high pressure. multistory parking garages). this is done by assuming that equipment and buildings are not present. The procedure for using the multi-energy model for a VCE is as follows:17 1. Flash Fires. "Guidelines for Evaluating the Characteristics of Vapor Cloud Explosions. identified as blast sources in the cloud. Conduct a field inspection to identify the congested areas. however. note that the flammable mixture may not fill an entire blast source volume and that the volume of equipment should be considered where it represents an appreciable proportion of the whole volume). . bridges.6-13 Explosions 271 TNO Multi-Energy Method The TNO method identifies the confined volumes in a process. and pipe racks. Semi-empirical curves are used to determine the overpressure. and BLEVEs (New York: American Institute of Chemical Engineers. and then determines the contribution to the overpressure from this confined volume (TNO is the Netherlands Organization for Applied Scientific Research). Normally. 1994). many risk analysts use their own judgment. for instance. such as process equipment in chemical plants or refineries.5 x lo6~ I r n ~ (this value is typical for the heat of combustion of an average stoichiometric hydrocarbon-air mixture). heavy vapors tend to move downhill. Estimate the energy of equivalent fuel-air charges by (a) considering each blast source separately. 2.

The blast peak side-on overpressure and positive-phase duration are calculated from the Sachs-scaled overpressure and the Sachs-scaled positive-phase duration. and Po is the ambient pressure (Pa). is the side-on blast overpressure (Pa). because of the momentum of a fuel release). R is the distance from the charge (m).is the Sachs-scaled side-on blast overpressure (dimensionless). pa is the ambient pressure (Pa). the Sachs-scaled blast side-on overpressure and positivephase duration at some distance R from a blast source is read from the blast charts in Figure 6-24 after calculation of the Sachs-scaled distance: where - R is the Sachs-scaled distance from the charge (dimensionless). assume a minimum strength of 1. For extended and quiescent parts. a source strength of 7 seems to more accurately represent actual experience. E is the charge combustion energy (J).5 bar. for side-on overpressures below about 0. However. . no differences appear for source strengths ranging from 7 to 10. which are in lowintensity turbulent motion (for instance. assume a strength of 3. The blast resulting from the remaining unconfined and unobstructed parts of a cloud can be modeled by assuming a low initial strength. AP. Once the energy quantities E and the initial blast strengths of the individual equivalent fuel-air charges are estimated.is assumed. The overpressure is given by and the positive phase duration is given by where P. For more nonquiescent parts.272 Chapter 6 Fires and Explosions A safe and most conservative estimate of the strength of the sources of a strong blast can be made if a maximum strength of 10 -representative of a detonation . Furthermore. 6.

(I) 1 0. and BLEVEs (New York: American Institute of Chemical Engineers. 1994).> . used by permission. . Source: Guidelines for Evaluating the Characteristics of Vapor Cloud Explosions. Flash Fires.5 g - Ro Combustion energy-scaled distance (R) Po = atmospheric pressure co = atmospheric sound speed E = amount of combustion energy Ro = charge radius Combustion energy-scaled distance (R) Figure 6-24 Sachs-scaled overpressure and Sachs-scaled positive-phase duration for the TNO multi-energy blast model.C a . .

Another popular method to estimate overpressures is the Baker-Strehlow method. A.ls The TNO multi-energy and Baker-Strehlow methods are essentially equivalent.274 Chapter 6 Fires and Explosions t. they may be initiated almost simultaneously. The stoichiometric equation is The initial number of moles on the left-hand side is 24. (2) the flame expansion characteristics of the process unit (which relates to confinement and spatial configuration). A complete description of the procedure is provided by Baker et al. the determination of a minimum distance between potential blast sources so that their individual blasts can be considered separately) is a factor in present research. which is selected based on three factors: (1)the reactivity of the released material. it is not clear how the results from each blast strength should be combined. "Recent Developments in the BakerStrehlow VCE Analysis Methodology. and the number of moles on the righthand side is 25." Process Safety Progress (1998). This expansion can be caused by two mechanisms: (1) thermal heating of the reaction products and (2) the change in the total number of moles by reaction. and (3) the obstacle density within the process unit. based on the degree of confinement. Further definition of this important issue (for instance. although the TNO method tends to predict a higher pressure in the near field and the Baker-Strehlow method tends to predict a higher pressure in the far field.A. is the Sachs-scaled positive-phase duration (dimensionless).8. C. Little guidance is provided for partial confinement geometries. In this case only a small pressure increase is expected as a result of the change in the number of moles. and c. consider the combustion of propane in air. Both methods require more information and detailed calculations than the TNT equivalency method. t. For example. Baker. Doolittle. is the ambient speed of sound (mls).8. Fitzgerald. J. and M. This method is based on a flame speed. If separate blast sources are located close to one another. Tang. and almost all the blast energy must be due to thermal energy release. The major problem with the application of the TNO multi-energy method is that the user must decide on the selection of a severity factor. E is the charge combustion energy (J). is the positive-phase duration (s). . 17(4): 297. M. The most conservative approach to this issue is to assume a maximum initial blast strength of 10 and to sum the combustion energy from each source in question. A set of semi-empirical curves is used to determine the overpressure. 18Q. G. and the respective blasts should be added. Furthermore. For most hydrocarbon combustion explosions in air the change in the number of moles is small. Energy of Chemical Explosions The blast wave resulting from a chemical explosion is generated by the rapid expansion of gases at the explosion site.

The released energy is equal to the work required to expand the gases.25. The energy of explosion for hexane is found in appendix B.016 kg)(818. Substituting into Equation 6-24.(0. Availability (Exergy) Analysis (Lexington. Crowl also showed how the energy of explosion could be determined for materials exploding at different gas compositions and nonambient temperatures and pressures. Crowl. Equation 6-24 applies. 19D.02)(1000 kg)(l moV0. Solution a. However.V." Journal of Loss Prevention in the Process Industries (1992). Crowl19reasoned that this expansion work is a form of mechanical energy. M A : Mulliken House. S(2): 109-118. A. . Sussman. Appendix B contains energy of explosion values based on thermodynamic availability. For most practical purposes the two properties can be used interchangeably. Determine (a) the equivalent amount of TNT and (b) the side-on peak overpressure at a distance of 50 m from the blast. Sussman20showed that the thermodynamic availability for a reacting system can be computed using the standard Gibbs energy of formation. the scaled overpressure is 0. . Example 6-9 One thousand kilograms of methane escapes from a storage vessel. Crowl then concluded that the energy of explosion for a material exploding at room temperature and pressure is equal to the standard Gibbs energy of formation.6-13 Explosions 275 The energy released during a reaction explosion is computed using standard thermodynamics. these adjustments are normally small. we obtain qmAH. 1981). mixes with air. 20M. and explodes. Thus the overpressure is This overpressure will demolish steel panel buildings. The thermodynamic availability is a state function used to determine the maximum mechanical energy extractable from a material as it moves into equilibrium with its surroundings. as shown in appendix B.7 kJ/mol) mTNT = = 218 kg TNT 4686 kJ/kg ETNT b. Equation 6-21 is used to determine the scaled distance: From Figure 6-23 (or Equation 6-23). Assume an explosion efficiency of 2%. "Calculating the Energy of Explosion Using Thermodynamic Availability. For many materials the heat of combustion and energy of explosion differ by less than lo%.

and thermodynamic availability." Physics of Fluids (1959). isentropic expansion. a blast strength of 7 is chosen. Determine the overpressure from this vapor cloud explosion at a distance of 100 m from the blast using the TNO multi-energy method.confined below the tank. To apply the TNO multi-energy method.5 ~ J l m ~ . Energy of Mechanical Explosions For mechanical explosions a reaction does not occur and the energy is obtained from the energy content of the contained substance. is the ambient pressure (forcelarea). If this energy is released rapidly. 21H. an explosion may result. Four methods are used to estimate the energy of explosion for a pressurized gas: Brode's equation. Brode.5 MJ/m3) = 7329 MJ. The result is The curve labeled 7 in Figure 6-24 is used to determine the scaled overpressure value of about 0.13)(101.2 kPa = 1. Examples of this type of explosion are the sudden failure of a tire full of compressed air and the sudden catastrophic rupture of a compressed gas tank. P.13. the resulting total energy is (2094 m"(3. "Blast Waves from a Spherical Charge. pa This is adequate to shatter concrete or cinder block walls. and by multiplying by the confined volume.L. isothermal expansion.3 kPa) = 13. . The tank is supported 1m off the ground by concrete piles. It determines the energy required to raise the pressure of the gas at constant volume from atmospheric pressure to the final gas pressure in the vessel. The Sachs-scaled energy is determined using Equation 6-25. The concentration of vapor in the cloud is assumed to be at stoichiometric concentrations.- - - 276 Chapter 6 Fires and Explosions Example 6-10 Consider the explosion of a propane-air vapor cloud confined beneath a storage tank.9 psi. representing the volume underneath the tank. = (0. 2: 17. Solution The heat of combustion of a stoichiometric hydrocarbon-air mixture is approximately 3. Assume a cloud volume of 2094 m3. Brode's method2I is perhaps the simplest approach. The resulting side-on overpressure is determined from Equation 6-26: po = AE. The resulting expression is where E is the energy of explosion (energy).

is the burst pressure of the vessel (forcelarea). Because P. The isentropic expansion method assumes that the gas expands isentropically from its initial to final state. with the addition of a correction term. Crowl. Thermodynamic availability represents the maximum mechanical energy extractable from a material as it comes into equilibrium with the environment.6-13 Explosions 277 P. "Calculating the Energy of Explosion. is the ideal gas constant and T. This is represented by the following equation: where R. This correction term accounts for the energy lost as a result of the second law of thermodynamics. The following equation represents this case: The isothermal expansion case assumes that the gas expands isothermally. >PI. is the ambient temperature (degrees)." . indicating that the energy is released to the surroundings during the vessel rupture. 22D. the energy calculated from Equation 6-28 is positive. V is the volume of expanding gas in the vessel (volume). The resulting overpressure from an explosion is a form of mechanical energy. The final method uses thermodynamic availability to estimate the energy of explosion. and y is the heat capacity ratio for the gas (unitless).A. Thus thermodynamic availability predicts a maximum upper bound to the mechanical energy available to produce an overpressure. 1 ~ batch thermodynamic availability resulted in the following An analysis by C r o ~ using ~ expression to predict the maximum explosion energy of a gas contained within a vessel: Note that Equation 6-31 is nearly the same as Equation 6-30 for an isothermal expansion.

It is thought that the Brode equation more closely predicts the potential explosion energy close to the explosion source. and that the isentropic expansion method predicts better the effects at a greater distance. . This is thermodynamically inconsistent because the final temperature is ambient. used by permission. For estimation purposes it is not uncommon to subtract 50% of the total potential energy to calculate the blast pressure effects from vessel burst. Crowl and C. The question arises as to which method to use. it is unclear where this transition occurs. Source: D. the expansion of an ideal gas from 200 psia to 14. A. The calculation assumes an inert gas initially at 298 K with y = 1. in press). according to the second law of thermodynamics. Figure 6-25 presents the energy of explosion using all four methods as a function of initial gas pressure in the vessel. or far field. Mashuga.7 psia results in a final temperature of 254"R. or near field. The thermodynamic availability method accounts for this loss through the correction term in Equation 6-31. some of the energy must be expelled as waste heat. However. The isentropic expansion results in a gas at a very low temperature. The isentropic method produces a low value for the energy of explosion. All four methods continue to be used to estimate the energy of explosion for compressed gases. UnderstandingExplosions in the Process Industries (New York: American Institute of Chemical Engineers. The gas expands into ambient air at 1 atm pressure. computed using four different methods. or -205°F. a portion of the potential explosion energy of vessel burst is converted into kinetic energy of the vessel pieces and other inefficiencies (such as strain energy in the form of heat in the vessel fragments).4. The isothermal expansion method predicts a large value for the energy of explosion because it assumes that all the energy of compression is available to perform work. Also.Chapter 6 Fires and Explosions Isothermal 0 0 100 200 300 Gas pressure (psia) 400 0 500 Figure 6-25 The energy of explosion for a compressed inert gas. In reality. V.

This debris strikes storage tanks.) Missile Damage An explosion occurring in a confined vessel or structure can rupture the vessel or structure. Unconfined explosions also create missiles by blast wave impact and subsequent translation of structures. 1972." paper presented at the Sixth International Meeting of Forensic Sciences (Edinburgh. can cause appreciable injury to people and damage to structures and process equipment. Blast Damage to People People can be injured by explosions from direct blast effects (including overpressure and thermal radiation) or indirect blast effects (mostly missile damage). 2 " J. This debris. process equipment. "Diagnostic Features of Explosion Damage. Clancey. resulting in the projection of debris over a wide area. resulting in secondary fires or explosions. This relationship is useful during accident investigations for calculating the energy level required to project fragments an observed distance. "Diagnostic Features of Explosive Damage. 1972)." paper presented at Sixth International Meeting of Forensic Science (Edinburgh. Clancey. Data from V. and pipelines. or missiles. discussed in section 2-6. as illustrated in Figure 6-26. Blast damage effects are estimated using probit analysis. A localized explosion in one part of the plant projects debris throughout the plant. . Missiles are frequently a means by which an accident propagates throughout a plant facility.6-13 Explosions 279 Distance (ft) Figure 6-26 Maximum horizontal range of blast fragments. Clancey2"eveloped an empirical relationship between the mass of explosive and the maximum horizontal range of the fragments.

21)(14. Injury to personnel is determined using probit equations from Table 2-5.280 Chapter 6 Fires and Explosions Example 6-11 A reactor contains the equivalent of 10. If it explodes. estimate the injury to people and the damage to structures 500 ft away. this explosion could project blast fragments a maximum distance of 6000 ft.21 and the overpressure is (0. . Solution The overpressure is determined using Equation 6-21 and Figure 6-23. Thus Substituting this value into the probit equations yields Table 2-4 converts the probit to percentages. Based on Figure 6-26.000 Ib of TNT. Table 6-9 indicates that steel-panel buildings will be demolished at this location. resulting in probable injuries and damage as a result of blast fragments. The scaled distance is From Figure 6-23 the scaled overpressure is 0.7 psia) = 3.1 psig. The result shows that there are no deaths and that less than 10% of the exposed people suffer eardrum ruptures. The probit equation for deaths resulting from lung hemorrhage is and the probit equation for eardrum rupture is where P is the overpressure in N/m2. This assumes complete conversion of explosion energy.

.000. ence on Vapor Cloud Modeling (New York: American Institute of Chemical Engineers.Loss Prevention in the Process Industries. Accidents occur under uncontrolled circumstances. 2. (3) the explosion efficiency is usually small (approximately 2% of the combustion energy is converted into a blast wave). is a classic example of a VCE. 26Lees. ignition of the resulting vapor cloud.27 From a safety standpoint the best approach is to prevent the release of material. efficiency of explosion. "Evaluation of Unconfined Vapor Cloud Explosion Hazards.000 and 140 fatalities (an average of almost 13 per year). 713. Qualitative studiesz6have shown that (1) the ignition probability increases as the size of the vapor cloud increases. "Evaluation of Unconfined Vapor Cloud Explosion Hazards. 714. 24Richard Prugh. Lees. 1987). sudden release of a large quantity of flammable vapor (typically this occurs when a vessel. 3. Some of the parameters that affect VCE behaviorzs are quantity of material released. A sudden failure of a 20-inch cyclohexane line between reactors led to vaporization of an estimated 30 tons of cyclohexane. despite any safety systems installed to prevent ignition. 27Pr~gh. volatile superheated liquid. p. and (4) turbulent mixing of vapor and air and ignition of the cloud at a point remote from the release increases the impact of the explosion. A summary of 29 V C E Sover the period 1974-1986 shows property losses for each event ~~ of between $5. dispersion of the vapor throughout the plant site while mixing with air.000. The accident at Flixborough. Data collected from real events are mostly unreliable and difficult to compare. existence of a threshold quantity of material. VCEs are difficult to characterize. (2) vapor cloud fires are more common than explosions. probability of explosion rather than fire. These explosions occur in a sequence of steps: 1. Any process containing quantities of liquefied gases. Loss Prevention in the Process Industries." International ConferW. 171155.000 and $100.p. 171156. A large cloud of combustible material is dangerous and almost impossible to control. The entire plant site was leveled and 28 people were killed. England. The vapor cloud dispersed throughout the plant site and was ignited by an unknown source 45 seconds after the release. VCEs have increased in number because of an increase in inventories of flammable materials in process plants and because of operations at more severe conditions. probability of ignition of the cloud. distance traveled by the cloud before ignition.6-13 Explosions 281 Vapor Cloud Explosions The most dangerous and destructive explosions in the chemical process industries are vapor cloud explosions (VCEs). 2"rank P. primarily because of the large number of parameters needed to describe an event. time delay before ignition of cloud. fraction of material vaporized." p. p. or high-pressure gases is considered a good candidate for a VCE. ruptures). and location of ignition source with respect to release. containing a superheated and pressurized liquid.

The fire heats the walls of the tank. The vapors might also be hazardous to personnel by means of skin burns or toxic effects. If the BLEVE is not caused by a fire. pronounced blel-vee) is a special type of accident that can release large quantities of materials. The tank walls below liquid level are cooled by the liquid. propelling vessel parts for great distances. The fraction vaporized is estimated using the methods discussed in section 4-7. W.1980). The tank ruptures. 1980). 2. using process conditions that minimize flashing if a vessel or pipeline is ruptured. the amount depends on the physical and thermodynamic conditions of the vessel contents. Boiling-Liquid Expanding-Vapor E x p l o s i ~ n s ~ ~ A boiling-liquid expanding-vapor explosion (BLEVE. Explosions (Berlin: Springer-Verlag. if they are toxic.Loss Prevention in the Process Industries. Suggested Reading W. resulting in a VCE. The steps are as follows: 1.p. 99. 1989). . Dust Explosions: Course.282 Chapter 6 Fires and Explosions Methods that are used to prevent VCEs include keeping low inventories of volatile. For either situation the energy released by the BLEVE process itself can result in considerable damage. A BLEVE occurs when a tank containing a liquid held above its atmospheric pressure boiling point ruptures. p. BLEVEs are caused by the sudden failure of the container as a result of any cause. Industrial Explosion Prevention and Protection (New York: McGraw-Hill. the liquid may ignite as the tank ruptures. When a BLEVE occurs in a vessel. If the flames reach the tank walls or roof where there is only vapor and no liquid to remove the heat. Industrial Explosion Prevention and Protection (New York: McGraw-Hill. resulting in the explosive vaporization of a large fraction of the tank contents. Bodurtha. a vapor cloud might form. the tank metal temperature rises until the tank loses it structural strength. 28Lees. a large area might be subjected to toxic materials. increasing the liquid temperature and the pressure in the tank. If the liquid is flammable and a fire is the cause of the BLEVE. explosively vaporizing its contents. a VCE might result. 4. Bodurtha. and installing automated block valves to shut systems down while the spill is in the incipient stage of development. Bartknecht. 5. 3. The most common type of BLEVE is caused by fire. using analyzers to detect leaks at low concentrations. flammable materials. If the materials are flammable. the boiling and burning liquid behaves as a rocket fuel. Bartknecht. 171178. Prevention.1980). only a fraction of the liquid vaporizes. Frank T. A fire develops adjacent to a tank containing a liquid. Protection (Berlin: Springer-Verlag. Often.

Problems

283

D. A. Crow1 and C. V. Mashuga, Understanding Explosions in the Process Industries (New York: American Institute of Chemical Engineers, in press). Rolf K. Eckhoff, Dust Explosions in the Process Industries, 2d ed. (London: Butterworth-Heinemann, 1997). Guidelines for Evaluating Process Plant Buildings for External Explosions and Fires (New York: American Institute of Chemical Engineers, 1996). Guidelines for Evaluating the Characteristicsof Vapor Cloud Explosions, Flash Fires, and BLEVEs (New York: American Institute of Chemical Engineers, 1994). Gilbert F. Kinney and Kenneth J. Graham, Explosive Shocks in Air, 2d ed. (Berlin: Springer-Verlag, 1985). Frank P. Lees, Loss Prevention in the Process Industries, 2d ed. (London: Butterworth-Heinemann, 1996), ch. 16 and 17. Bernard Lewis and Guenther von Elbe, Combustion, Flames, and Explosions of Gases, 3d ed. (New York: Academic Press, 1987). Daniel R. Stull, Fundamentals of Fire and Explosion, AICHE Monograph Series, no. 10, v. 73 (New York: American Institute of Chemical Engineers, 1977).

Problems
6-1. Estimate the flash point of a solution of 50 mol % water and 50 mol % methanol. 6-2. Estimate the flash point of a solution of 50 mol % water and 50 mol % ethanol. 6-3. Estimate the LFL and the UFL of the following mixtures:
All in volume %

Hexane Methane Ethylene Acetone Ethyl ether Total combustibles Air

0.5 1. O 0.5 0.0 0.0 2.0 98.0

0.0 0.0 0.5 1.O 0.5 2.0 98.0

1.0 1O . 1. O 0.0 0.0 3.0 97.0

0.0 0.0 1.0 1.0 1.0 3.0 97.0

6-4. Estimate the LFL and the UFL of Problem 6-3a at 50°C, 75"C, and 100°C. 6-5. Estimate the UFL of Problem 6-3a at 1atm, 5 atm, 10 atm, and 20 atm of pressure. 6-6. Estimate the LFL and the UFL using the stoichiometric concentrations for methane, propylene, ethyl ether, and acetone. Compare these estimates to actual values. 6-7. Estimate the LOC of propane, hydrogen, and methane. 6-8. Determine the LOC of a mixture of 2% hexane, 3% propane, and 2% methane by volume. 6-9. Determine the minimum compression ratio required to raise the temperature of air over hexane to its AIT. Assume an initial temperature of 100°C. 6-10. What will be the LFL of hexane in the presence of hexane mists with drops larger than 0.01 mm'?

284

Chapter 6

Fires and Explosions

6-11. Why do staged hydrogen compressors need interstage coolers? 6-12. Why do hot engines sometimes continue to run after the ignition is turned off? 6-13. A set of experiments is run on a flammable gas in a spherical vessel. The following data are obtained for two different vessel volumes. Estimate the value of KG for this combustible gas:

V = I m3
Time (s)

V = 20m3
Time (s)

P (bar)

P (bar)

-

-

-

-

-

-

6-14. Determine the energy of explosion for 1lb of gaseous n-butane. What is the TNT equivalent? 6-15. A gas cylinder contains 50 lb of propane. The cylinder accidentally falls over and ruptures, vaporizing the entire contents of the cylinder. The cloud is ignited and an explosion occurs. Determine the overpressure from this explosion 100 ft away. What type of damage is expected? 6-16. A VCE with methane destroyed a house structure 100 ft away from the ignition source. Estimate the amount of methane released. 6-17. A large cloud of propane is released and eventually ignited, producing a VCE. Estimate the quantity of propane released if the blast shattered windows 3 mi from the source of the ignition. 6-18. At 77°F gasoline has a vapor pressure of 4.6 psia. Why can gasoline be stored in vented storage vessels without the presence of flammable vapors above the liquid in the vessel? Comment on the EPA's effort to reduce gasoline volatility in order to reduce fugitive emissions. What will happen as the volatility is reduced?

Problems

285

6-19. An automobile assembly line includes an operation that involves filling the gas tanks with gasoline. Estimate the ventilation rate required to reduce the vapors from this operation to below the LFL for gasoline. Assume that each tank has a volume of 14 gal and that a tank can be filled in 3 min. Assume splash filling and that only one tank is filled at a time. The molecular weight of gasoline is about 94, and its vapor pressure at 77°F is 4.6 psia. Also, calculate the ventilation air required to reduce the concentration of the gasoline vapors to below the TLV-TWA. Which problem is more difficult? 6-20. A butane tank is located 500 ft from a residential area. Estimate the minimum instantaneous release of butane required to produce a vapor concentration in the residential area equal to the LFL for butane. What continuous release rate is required? Assume that the release occurs at ground level. Will the minimum amount increase, decrease, or stay the same if the release occurs above ground level? 6-21. Benzene is stored in an inside storage area, 15 ft long and 15 ft wide with an 8-ft ceiling. This storage area has a ventilation system that changes the air in the room completely six times per hour. The storage area is also equipped with a flammable vapor detector that sounds an alarm when the flammable vapor concentration reaches 25% of the LFL for benzene. What is the minimum benzene spill rate, in lblhr, that will set off the flammable vapor alarm in the room? Assume a pressure of 1 atm and a temperature of 80°F. Also assume average ventilation conditions. 6-22. A standard laboratory cylinder is about 5 ft high with an internal vessel diameter of about 6 in. Determine the total energy of explosion for this cylinder if it contains nitrogen compressed to 2500 psig. Assume initial and ambient conditions of 298 K and 1 atm. 6-23. Many chemical operators believe that the inerted vapors above a flammable liquid are not flammable when they mix with air. This is frequently not the case: If the inerted vapors escape from the vessel and mix with air or if the vessel is purged with air after emptying, the resulting mixture might be flammable. A storage vessel contains liquid benzene at 100°F. The vessel vapor space is inerted with pure nitrogen to a total pressure of 112-in of water gauge. Assume that the vapor space is saturated with benzene vapor. a. Determine the volume percent concentration of benzene in the vapor. b. Use a flammability diagram to show whether this mixture will become flammable or not when mixed with air. (Hint: 1 atm = 34.4 ft of water.) 6-24. An informal industry rule is to design occupied control rooms to withstand a 1-ton blast of TNT at 100 ft. a. What overpressure does this correspond to? b. What quantity (in pounds) of propane (C,H,) does this correspond to, based on an equivalent amount of energy? c. How far away (in ft) from this 1- ton blast must a residential home be in order to receive no more than minor damage to house structures?

-

-

-

-

286

Chapter 6

Fires and Explosions

6-25. According to fire code, propane storage tanks cannot be closer than 10 ft from a house. This requirement is designed to prevent flammable vapors from entering the house, not to protect the house from a potential explosion. What quantity of propane (in lb,) can be stored under these conditions that will cause no more than minor damage to a house in the event of an explosion? Be sure to list any assumptions. 6-26. Fires and explosions are substantial hazards in many chemical plants. a. Describe with examples the three ingredients of any fire. b. Create a checklist with at least six items to identify fire hazards in any workplace. c. List six common fire preventionlprotection features for chemical plants, and describe when they would be appropriate. 6-27. The following liquids are stored in a storage vessel at 1 atm and 25°C. The vessels are vented with air. Determine whether the equilibrium vapor above the liquid will be flammable. The liquids are: a. Acetone b. Benzene c. Cyclohexane d. Ethyl alcohol e. Heptane f. Hexane g. Pentane h. Toluene 6-28. A natural gas wellhead is located 400 m from an instrument control room. The control room is a potential ignition hazard in the event of a leak of natural gas (essentially pure methane). Studies have shown that a suitable safety margin is imposed if the downwind gas concentration is determined using one-half the LFL. For methane this represents a concentration of 2.5 vol. %. a. What is the minimum release rate of methane (in kgls) that will result in a concentration at the control room equal to half the LFL? Be sure to state your assumptions clearly. Assume a temperature of 298 K and an ambient pressure of 1 atm. b. If the methane pressure in the wellhead is at 10 atm pressure, what hole size (in cm) will produce the release rate of part a? c. If the largest pipe Size in the wellhead is 4 cm (internal diameter), comment on the likelihood of an ignition hazard from the control room. 6-29. In the TWA Flight 800 tragedy the accident is blamed on explosion of fuel vapors in the central fuel tank. The volume of the central fuel tank is 18,000 gal. a. If, at the time of the explosion, the fuel concentration in the tank is 1% by volume and the pressure inside the tank is 12.9 psia, determine the equivalent energy of explosion for the vapor (in pounds of TNT). Assume a temperature of 80°F. Be sure to state carefully any assumptions.

Problems

287

b. Estimate the overpressure at a distance of 50 ft from the fuel tank explosion from the

explosion of the vapors in part a. c. Estimate the limiting oxygen concentration (LOC) for the jet fuel, given that the LFL is 0.6% by volume and the stoichiometric coefficient for oxygen in the combustion equation is 18.9. For jet fuel the energy of explosion is 18,590 Btullb and the molecular weight is 160. 6-30. You have decided to purchase a 500-gal tank of liquid propane (C,H,) to heat your house during the winter. You are concerned about tank rupture and the possibility of a vapor cloud explosion of all the propane. How far away (in ft) must the tank be from the house to ensure that your house will receive only minor damage from an explosion? The specific gravity of liquid propane is 0.500, and the energy of explosion for propane is 503.9 kcallg-mol. 6-31. A liquid mixture containing 0.50 mole fraction benzene-toluene is contained in a storage vessel at 25°C and 1atm. The vessel is vented to the atmosphere. a. Is the vapor in the vessel flammable? b. What are your resulting concerns about fire and explosion hazards with this storage vessel? Hint: Benzene-toluene can be assumed to be an ideal liquid-vapor system. 6-32. A tank containing liquid butane (C,H,,) is located 500 ft from an electrical substation. One of the scenarios we are considering is the breaking of a 1-in schedule 40 pipe (internal diameter = 1.049 in) with discharge of the liquid butane. We are concerned that this leak will cause flammable vapor concentrations at the substation. Assume that all the liquid flashes to vapor. a. Estimate the discharge rate (in lb,/s) of butane from the 1-in broken pipe. b. Estimate the butane vapor concentrations at the substation. Is this likely to be a flammable hazard? The temperature is 80°F and the ambient pressure is 1 atm. Make sure you clearly state any assumptions. The vapor pressure of liquid butane at 80°F is 40 psia, and the specific gravity of liquid butane at 80°F is 0.571. 6-33 Acetone is used as a solvent in a laboratory. There is some concern about the fire hazards associated with the acetone. One solution is to dilute the pure acetone with water, thus providing an increased flash point. What mole fraction of water in a water-acetone mixture is required to increase the flash point of the mixture to 100°F? Acetone is completely soluble in water. 6-34 You have been assigned the task of assisting in relocating the new control room for your process. The new control room will be designed to withstand an explosive overpressure of 2 psig. Your attention is focused on a propane storage tank located 100 m from the proposed site for the new control room. What is the maximum quantity of propane (in kg) that can be stored in this tank without exceeding the overpressure rating of the control room? Make sure you state any assumptions used in your calculation.

--

288

Chapter 6

Fires and Explosions

6-35. Methyl alcohol liquid is stored in a vessel. Its vapor is inerted with nitrogen to a total pressure of 2 in of water gauge. Will the inerted vapor be flammable if it escapes the vessel? Assume a temperature of 25°C. 6-36. Draw a flammability diagram for n-butane. The experimentally reported LOC for n-butane is 12%. What must the oxygen concentration be reduced to before pumping in butane? What butane concentration must the vapor be reduced to before pumping air into the vessel before taking it out of service? 6-37. For flammable gases the minimum ignition energy is typically 0.1 mJ. The mass of a penny is typically 2.6 g. How far must this penny be dropped to contain the kinetic energy equal to 0.1 mJ? 6-38. During a particular accident, an estimated 39,000 kg of flammable material was released and ignited, resulting in an explosion and fireball and the subsequent fatalities and equipment damage. The publication Guidelines for Evaluating the Characteristics of Vapor Cloud Explosions, Flash Fires, and BLEVES (New York: American Institute of Chemical Engineers, 1994) provides a number of equations useful for estimating the effects of such an explosion. The heat radiation intensity (in kw/m2) from a ball of burning vapor is given by the empirical equation

The effective time duration (in seconds) of the burn is given by

The height of the center of the fireball (in meters) is assumed to be constant during the burn and is given by

Finally, the maximum diameter of the fireball (in meters) is given by

For these equations I, is the effective radiation intensity (kW/m2),m fis the mass of fuel (kg), L is the distance from the center of the fireball to the receptor (m), and t, is the effective time duration of the burn (s). Use a spreadsheet program (such as Quattro Pro or Excel) to estimate the total number of fatalities resulting from the burning fireball. Use the probit equations provided in the text. Assume that 400 people are distributed evenly at a distance of 75 m to 1000 m from the fireball center. Divide the distance interval into a number of small increments. Use a small enough distance increment so that the results are essentially independent of the increment size.

Problems

289

Your spreadsheet output should have designated columns for the distance from the fireball center, radiation intensity, probit values, and percentage and number of fatalities. You should also have a single designated spreadsheet cell providing the total number of fatalities. One way to simplify the procedure is to specify a threshold radiative flux. It is assumed that 100% fatalities will occur to anyone exposed to anything above this value. Anyone exposed to a lesser value will be unharmed. Estimate an approximate threshold radiative flux value that will result in the same number of fatalities as the detailed probit calculation. 6-39. The air in a 55-gal drum must be flushed and inerted with nitrogen before the drum is filled with a flammable liquid. This is accomplished by placing a nitrogen lance through the hole of the drum reaching to the bottom. A constant flow rate of nitrogen is used to achieve the inerting. a. Show that the concentration of oxygen in the drum is represented by

where C is the concentration of oxygen in the drum (mass/volume), k is the nonideal mixing factor (0.1 < k < I), and Q, is the volumetric flow of nitrogen (volume/time). b. Show that the time required to reach a target concentration Cf from an initial concentration C, is given by

c. Estimate the time it will take to inert a drum to 1% oxygen using 75 L/min of nitrogen. Use k as a parameter. 6-40. A container in a process using a flammable vapor has dimensions of 100 m by 100 m by 10 m high. Use the TNO multi-energy model to estimate the overpressure 100 m from the process resulting from the release and ignition of the flammable vapor. Assume that 20% of the process volume is moderately congested and that the remaining 80% is lightly congested. Be sure to state any additional assumptions.

Designs to Prevent Fires and Explosions

twofold strategy is used to limit the potential damage from fires and explosions: prevent the initiation of the fire or explosion and minimize the damage after a fire or explosion has occurred. This strategy is presented in this chapter. The specific topics include inerting, use of the flammability diagram introduced in chapter 6, static electricity, controlling static electricity, ventilation, explosion-proof equipment and instruments, sprinkler systems, and miscellaneous design features for preventing fires and explosions. For any fire or combustion explosion to occur, three conditions must be met (as shown in the fire triangle of Figure 6-1). First, a combustible or explosive material must be present. Second, oxygen must be present to support the combustion reaction. Finally, a source of ignition must be available to initiate the reaction. If any of the three conditions of the fire triangle is eliminated, the triangle is broken and it is impossible for a fire or combustion explosion to result. This is the basis for the first six design methods listed above. Damage resulting from fires and explosions is minimized by stopping fires or explosions as quickly as possible and also by designing the process equipment (and control centers) to withstand their effects. Miscellaneous design features are additional safety design methods that are usually included in the early design phases of new projects and are often the basis for safety improvements in existing plants.

A

- -

-

-

--

292

Chapter 7

Designs to Prevent Fires and Explosions

7-1 lnerting
Inerting is the process of adding an inert gas to a combustible mixture to reduce the concentration of oxygen below the limiting oxygen concentration (LOC). The inert gas is usually nitrogen or carbon dioxide, although steam is sometimes used. For many gases the LOC is approximately lo%, and for many dusts it is approximately 8%. Inerting begins with an initial purge of the vessel with inert gas to bring the oxygen concentration down to safe concentrations. A commonly used control point is 4% below the LOC, that is, 6 % oxygen if the LOC is 10%. After the empty vessel has been inerted, the flammable material is charged. An inerting system is required to maintain an inert atmosphere in the vapor space above the liquid. Ideally this system should include an automatic inert gas addition feature to control the oxygen concentration below the LOC. This control system should have an analyzer to continuously monitor the oxygen concentration in relationship to the LOC and a controlled inert gas feed system to add inert gas when the oxygen concentration approaches the LOC. More frequently, however, the inerting system consists only of a regulator designed to maintain a fixed positive inert pressure in the vapor space; this ensures that inert gas is always flowing out of the vessel rather than air flowing in. The analyzer system, however, results in a significant savings in inert gas usage without sacrificing safety. Consider an inerting system designed to maintain the oxygen concentration below 10%. As oxygen leaks into the vessel and the concentration rises to 8%, a signal from the oxygen sensor opens the inert gas feed valve. Once again the oxygen level is adjusted to 6%. This closedloop control system, with high (8%) and low (6%) inerting set points, maintains the oxygen concentration at safe levels with a reasonable margin of safety. NFPA recommendations are described at the end of this section. There are several purging methods used to initially reduce the oxygen concentration to the low set point: vacuum purging, pressure purging, combined pressure-vacuum purging, vacuum and pressure purging with impure nitrogen, sweep-through purging, and siphon purging.

Vacuum Purging
Vacuum purging is the most common inerting procedure for vessels. This procedure is not used for large storage vessels because they are usually not designed for vacuums and usually can withstand a pressure of only a few inches of water. Reactors, however, are often designed for full vacuum, that is -760 mm Hg gauge or 0.0 mm Hg absolute. Consequently, vacuum purging is a common procedure for reactors. The steps in a vacuum purging process include (1) drawing a vacuum*on the vessel until the desired vacuum is reached, (2) relieving the vacuum with an inert gas, such as nitrogen or carbon dioxide to atmospheric pressure, and (3) repeating steps 1 and 2 until the desired oxidant concentration is reached.

7-1 lnerting

293

End of first vacuum purge

End of second vacuum purge

concentration is constant Time

Number of moles of oxygen is constant

Figure 7-1

Vacuum purge cycles.

The initial oxidant concentration under vacuum (yo)is the same as the initial concentraand initial low pressure or vaction, and the number of moles at the initial high pressure ( p H ) uum (PL)are computed using an equation of state. The process for vacuum purging is clarified using the stepwise procedure shown in Figure 7-1. A vessel of known size is vacuum-purged from an initial oxygen concentration yo to a final target oxygen concentration yi. The vessel is initially at pressure P, and is vacuum-purged using a vacuum at pressure PL. The objective of the following calculation is to determine the number of cycles required to achieve the desired oxygen concentration. Assuming ideal gas behavior, the total moles a t each pressure are

where nH and nL are the total moles in the atmospheric and vacuum states, respectively. The number of moles of oxidant for the low pressure PL and high pressure P, are computed using Dalton's law:

where 1L and 1H are the first atmospheric and first vacuum states, respectively.

294

Chapter 7

Designs to Prevent Fires and Explosions

When the vacuum is relieved with pure nitrogen, the moles of oxidant are the same as in the vacuum state and the moles of nitrogen increase. The new (lower) oxidant concentration is

where y, is the oxygen concentration after the first purge with nitrogen. Substituting Equation 7-3 into Equation 7-5 gives

If the vacuum and inert relief process is repeated, the concentration after the second purge is

This process is repeated as often as required to decrease the oxidant concentration to a desired level. The concentration after j purge cycles, vacuum and relief, is given by the following general equation:

This equation assumes that the pressure limits P, and PL are identical for each cycle. The total moles of nitrogen added for each cycle is constant. For j cycles the total nitrogen is given by

Example 7-1
Use a vacuum purging technique to reduce the oxygen concentration within a 1000-gal vessel to 1ppm. Determine the number of purges required and the total nitrogen used. The temperature is 75"F,and the vessel is originally charged with air under ambient conditions. A vacuum pump is used that reaches 20 mm Hg absolute, and the vacuum is subsequently relieved with pure nitrogen until the pressure returns to 1 atm absolute.

Solution
The concentration of oxygen at the initial and final states is yo = 0.21 lb-mol 02/totalmol,
y,
=

1ppm

=

1x

1b-molO,/total mol.

7-1 lnerting

295

The required number of cycles is computed using Equation 7-6:

I

=

ln(10-610.21) = 3.37. ln(20 mm Hg/760 mm Hg)

Number of purges = j = 3.37. Four purge cycles are required to reduce the oxygen concentration to 1 ppm. The total nitrogen used is determined from Equation 7-7. The low pressure P, is

PI.=

20 (760mm Hg mm Hg

(14.7 psia) = 0.387 psia,

= =

4(14.7

-

0.387) psia
=

(1000 gal)(l ft3/7.48gal) (10.73 psia ft3/lb-mo1°~)(75 460)"R +

1.33 lb-mol

37.2 lb of nitrogen.

Pressure Purging
Vessels can be pressure-purged by adding inert gas under pressure. After this added gas is diffused throughout the vessel, it is vented to the atmosphere, usually down to atmospheric pressure. More than one pressure cycle may be necessary to reduce the oxidant content to the desired concentration. The cycles used to reduce the oxygen concentration to a target level are shown in Figure 7-2. In this case the vessel is initially at PL and is pressurized using a source of pure nitrogen at pH. The objective is to determine the number of pressure purge cycles required to reach the desired concentration. Because the vessel is pressurized with pure nitrogen, the number of moles of oxygen remains constant during pressurization, whereas the mole fraction decreases. During depressurization, the composition of the gas within the vessel remains constant, but the total number of moles is reduced. Thus the oxygen mole fraction remains unchanged. The relationship used for this purging process is identical to Equation 7-6, where nL is now the total moles at atmospheric pressure (low pressure) and n, is the total moles under pressure (high pressure). In this case, however, the initial concentration of oxidant in the vessel ( y o ) is computed after the vessel is pressurized (the first pressurized state). The number of moles for this pressurized state is n, and the number of moles for the atmospheric case is n ~ .

Pressure purging. we obtain yo = (0.296 Chapter 7 Designs to Prevent Fires and Explosions End of first pressure purge End o second f pressure purge Oxygen concentration is constant Time Figure 7-2 Number of moles of oxygen is constant Pressure purge cycles. Compare the quantities of nitrogen required for the two purging processes. however. . Also. Therefore the best purging process is selected based on cost and performance. The pressurization process is much more rapid compared to the relatively slow process of developing a vacuum.21) 14.0326. Substituting the numbers provided. Example 7-2 Use a pressure purging technique to reduce the oxygen concentration in the same vessel discussed in Example 7-1. The composition at the high-pressure condition is determined using the following equation: where Po is the starting pressure (here atmospheric). O n e practical advantage of pressure purging versus vacuum purging is the potential for cycle time reductions. Also.7) psia 1 = 0. The initial mole fraction of oxygen yo is now the concentration of oxygen at the end of the first pressurization cycle. Determine the number of purges required to reduce the oxygen concentration to 1ppm using pure nitrogen at a pressure of 80 psig and at a temperature of 75°F. the capacity of vacuum systems decreases significantly as the absolute vacuum is decreased. determine the total nitrogen required.7 psia [(80 + 14. Solution Equation 7-6 is used to determine the number of cycles required. uses more inert gas.

In this case the beginning of the cycle is defined as the end of the initial pressurization. The purging cycles for an evacuate-first purge are shown in Figure 7-4.0326) ln[14. Furthermore. the number of cycles j is the number of cycles after the initial evacuation. The computational procedure depends on whether the vessel is first evacuated or pressurized. the number of cycles j is the number of cycles after the initial pressurization.7-1 lnerting 297 The final oxygen concentration (y.21. . The j = 1n(10-~/0. This result illustrates the need for a cost performance comparison to determine whether the time saved in pressure purging justifies the added cost for nitrogen. The quantity of nitrogen used for this inerting operation is determined using Equation 7-7: An.) is specified to be 1ppm or number of cycles required is computed using Equation 7-6: lb-mol oxygenltotal lb-mol..2 lb of nitrogen for vacuum purging.7) psia 133.7) psia] ' The number of purge cycles is j = 5. Combined Pressure-Vacuum Purging In some cases both pressure and vacuum are available and are used simultaneously to purge a vessel. The purging cycles for a pressure-first purge are shown in Figure 7-3. If the initial oxygen mole fraction is 0.7 . = j(PH .T ) ~ = = 6(94.73 psia ft3/lb-mol0~)(53S0~) 11.14.6.7 ft" (10. However. the oxygen mole fraction at the end of this initial pressurization is given by At this point the remaining cycles are identical to pressure purging and Equation 7-6 applies. Pressure purging requires 6 purges and 311 Ib of nitrogen compared to 4 purges and 37. compared to four for the vacuum purge process. the remaining cycles are identical to the vacuum purge operation and Equation 7-6 is directly applicable. Thus six pressure purges are required.P ~ R.7 psiaI(80 + 14. In this case the beginning of the cycle is defined as the end of the initial evacuation.1 Ib-mol = 311 Ib of nitrogen. However. The oxygen mole fraction at this point is the same as the initial mole fraction.

Many of the nitrogen separation processes available today do not provide pure nitrogen. Time Figure 7-4 Vacuum-pressure purging with initial evacuation. they typically provide nitrogen in the 98% + range. Vacuum and Pressure Purging with Impure Nitrogen The equations developed for vacuum and pressure purging apply to the case of pure nitrogen only.Chapter 7 Designs to Prevent Fires and Explosions Time Figure 7-3 Vacuum-pressure purging with initial evacuation. .

7-1 lnerting 299 Assume that the nitrogen contains oxygen with a constant mole fraction of yo. especially if the initial cycle is a vacuum cycle. . When combining vacuum and pressure purging.. Sweep-Through Purging The sweep-through purging process adds purge gas into a vessel at one opening and withdraws the mixed gas from the vessel to the atmosphere (or scrubber) from another opening. Advantages and Disadvantages of the Various Pressure and Vacuum lnerting Procedures Pressure purging is faster because the pressure differentials are greater. Thus the mole fraction of oxygen at the end of this cycle is This result is generalized into the following recursive equation (Equation 7-11) and a generalized equation (Equation 7-12) for the oxygen concentration at the end of the jth pressure cycle: Equation 7-12 is used in place of Equation 7-6 for both pressure and vacuum purging. This purging process is commonly used when the vessel or equipment is not rated for pressure or vacuum. the purge gas is added and withdrawn at atmospheric pressure. less nitrogen is used compared to pressure purging. Vacuum purging uses less inert gas because the oxygen concentration is reduced primarily by vacuum. For a pressure purging procedure the total moles of oxygen present at the end of the first pressurization is given by the moles initially present plus the moles included with the nitrogen.. however. This amount is The total moles in the vessel at the end of the first pressurization are given in Equation 7-1. it uses more inert gas than vacuum purging.

t is determined using Equation 7-15: . Example 7-3 A storage vessel contains 100% air by volume and must be inerted with nitrogen until the oxygen concentration is below 1. constant temperature. C. Under these conditions the mass or volumetric flow rate for the exit stream is equal to the inlet stream.is the inlet oxidant concentration (mass or volumetric units). Equation 7-13 is rearranged and integrated: The volumetric quantity of inert gas required to reduce the oxidant concentration from C.01% oxygen? Solution The volume of nitrogen required Q... and the flow rate of oxidant exiting is CQ.Q. and it is determined using Equation 7-14: For many systems Co = 0. and t is time.. is the volumetric flow rate. The mass or volumetric flow rate of oxidant into the vessel is C.300 Chapter 7 Designs to Prevent Fires and Explosions Purging results are defined by assuming perfect mixing within the vessel. is Q. assuming the nitrogen contains 0. to C. C is the concentration of oxidant within the vessel (mass or volumetric units). The vessel volume is 1000 ft? How much nitrogen must be added. Q.25% by volume. The material balance around the vessel is where V is the vessel volume.t. and constant pressure.

01% oxygen). When using the siphon purging process. as shown in Figure 7-5. One approach is to continue the nitrogen flow until the vessel contains pure nitrogen. This could be expensive when purging large storage vessels. One might suggest an even more optimized procedure. A more robust design is to prevent the existence of flammable mixtures as the primary control. This involves first pumping air into the vessel until a point is reached on the air stoichiometric line above the UFL. the sweep-through process requires large quantities of nitrogen. this requires a large amount of nitrogen and is costly. Siphon purging is used to minimize this type of purging expense. and the rate of purging is equivalent to the volumetric rate of liquid discharge. The quantity of pure nitrogen required to reduce the oxygen concentration to 1. If air is used to purge the vessel.7-1 lnerting 301 This is the quantity of contaminated nitrogen added (containing 0. This is followed . and the gas composition follows along the line SR in Figure 7-5. ignition sources are too plentiful to use as the primary prevention mechanism. However. The vessel is initially at point A and contains pure fuel. As previously stated. The volume of purge gas is equal to the volume of the vessel. it may be desirable to first fill the vessel with liquid and then use the sweep-through purge process to remove oxygen from the residual head space. If nitrogen is first pumped into the vessel. the oxygen concentration is decreased to low concentrations with only a small added expense for the additional sweep-through purging.25% is Siphon Purging As illustrated in Example 7-3. the elimination of ignition sources alone is not enough to prevent fires and explosions. which crosses the flammability zone. By using this method. Then air can be introduced. the gas composition follows along line AS. The siphon purging process starts by filling the vessel with liquid . The procedure for taking a vessel out of service is illustrated in Figure 7-5. In this case the flammability zone is avoided and a safe vessel preparation procedure is ensured. followed by the elimination of ignition sources as a secondary control. The flammability diagram is important for determining whether a flammable mixture exists and for providing target concentrations for inerting and purging procedures. The purge gas is subsequently added to the vapor space of the vessel as the liquid is drained from the vessel.water or any liquid compatible with the product. Using the Flammability Diagram To Avoid Flammable Atmospheres The flammability diagram introduced in chapter 6 is an important tool to prevent the existence of flammable mixtures. A more efficient procedure is to inert with nitrogen until point S is reached. the composition follows line AR. The objective is to avoid the flammable region.

however.A : A Chapter 7 Designs to Prevent Fires and Explosions Nitrogen pumped in 20 Flammability zone 100 / 20 v \ 40 v N 60 0 Nitrogen. the composition at point S is determined graphically or with OSFC = . The flammability diagram reflects only the average gas composition within the vessel. is that the air forms a flammable mixture at the entry point as the pure air mixes with the fuel-rich gas mixture in the vessel. This approach avoids the nose of the flammability zone and minimizes the consumption of nitrogen. When using the nitrogen purge process. The problem with this approach. The approach is shown in Figure 7-6. volume percent Figure 7-5 qRO 0 100 Pure air A procedure for avoiding the flammability zone for taking a vessel out of service. Because the gas compositions at points R and M are known. Point S is approximated by a line starting at the pure air point R and connecting through a point M at the intersection of the LFL with the stoichiometric combustion line. by pumping nitrogen into the vessel followed by air. one must determine the location of point S in Figure 7-5. Using nitrogen first avoids this problem.

the fuel concentration at point S in Figure 7-6. Another approach is to estimate the fuel concentration at point S by extending the line from point R though the intersection of the minimum oxygen concentration (M) and the stoichiometric combustion line.7-1 lnerting 303 Figure 7-6 Estimating a target fuel concentration at point S for taking a vessel out of service. and z is the stoichiometric oxygen coefficient from the combustion reaction given by Equation 6-9. where OSFC is the out-of-service fuel concentration. that is. Point M is the intersection of the LFL line with the stoichiometric line. The derivation of Equation 7-16 is provided in appendix C. The analytical result is OSFC = . LFL is the volume percent of fuel in air at the lower flammability limit.

This is compared to an experimentally determined OSFC of 14. For ethylene." Process Safety Progress (1998). The vessel begins with air. an OSFC value of 14% is determined. following line SR until point R is reached. For all other species in Table 7-1.5% (Table 7-1). Equation 7-17 is derived in appendix C. 1. Figure 7-7 shows the procedure for placing a vessel into service. Equations 7-16 and 7-17 are approximations of the fuel concentration at point S. where LOC is the limiting oxygen concentration (also called the minimum oxygen concentration) in volume percent of oxygen. less than the experimentally determined OSFC value. then the ISOC is estimated. Then fuel is pumped in.7% fuel. for methane the LFL is 5. 17(3): 176. Crowl.3% and z is 2. A.51145) JP-4 Hydrogen Carbon monoxide 'C. and hydrogen Equation 7-17 predicts a higher OSFC than the experimentally determined value. Thus Equation 7-16 predicts an OSFC of 10. shown as point A. Mashuga and D. Nitrogen is pumped into the vessel until point S is reached. One approach is to use the intersection of the LFL with the stoichiometric combustion line. % Fuel) - Chemical Methane Ethane Propane Butane n-Pentane n-Hexane Natural gas Ethylene Propylene 2-Methylpropene 1-Butene 3-Methyl butene 1. that is. For instance.3-Butadiene Acetylene Benzene Chemical - ISOC (vol. with a small margin of safety. The in-service oxygen concentration (ISOC) represents the maximum oxygen concentration at point S in Figure 7-7 that just avoids the flammability zone.3-butadiene. % oxygen) OSFC (vol. they are usually conservative. Fortunately. The problem is to determine the oxygen (or nitrogen) concentration at point S. Equation 7-16 estimates an OSFC that is less than the experimental value. "Application of the Flammability Diagram for Evaluation of Fire and Explosion Hazards of Flammable Vapors. A line is drawn . If a detailed flammability diagram is lacking. % Fuel) ISOC (vol. V. By using an experimental LOC of 12%.304 Chapter 7 Designs to Prevent Fires and Explosions Table 7-1 Experimental In-Service Oxygen Concentrations (ISOCs) and Out-Of-Service Fuel Concentrations (OSFCs) l OSFC (vol. This is closer to the experimental value but still conservative. % oxygen) Cyclopropane Methyl alcohol Ethyl alcohol Dimethyl ether Diethyl ether Methyl formate Isobutyl formate Methyl acetate Acetone Methyl ethyl ketone Carbon disulfide Gasoline (11.

from the top apex of the triangle (R) through this intersection to the nitrogen axis. This is shown in Figure 7-8. and LFL is the fuel concentration at the lower flammability limit. z is the stoichiometric coefficient for oxygen given in Equation 6-9. The composition at S is determined graphically or with -z-j ISOC = - (0 1) where ISOC is the in-service oxygen concentration in volume % oxygen. volume percent ~ i i r o ~ pumped in en Figure 7-7 A procedure for avoiding the flammability zone for bringing a vessel into service.7-1 lnerting 305 Nitrogen. Equation 7-18 is derived in appendix C. in volume percent of fuel in air. .

Equation 7-19 predicts a lower . with the exception of methyl formate. Point M is the intersection of the LFL line with the stoichiometric combustion line. A comparison of the estimates using Equations 7-18 and 7-19 with the experimental values in Table 7-1 shows that Equation 7-18 predicts a lower oxygen value than the experimental values for all species.306 Chapter 7 Designs to Prevent Fires and Explosions Nitrogen. volume percent Figure 7-8 Estimating a target nitrogen concentration at point S for placing a vessel into service.'fi ISOC = z-- where LOC is the minimum oxygen concentration in volume percent oxygen. The analytical result is . The nitrogen concentration at point S is equal to 100 . An expression to estimate ISOC using the intersection of the minimum oxygen concentration and the stoichiometric line is also found using a similar procedure.ISOC.

3-butadiene. electrons are not as mobile and are trapped on one of the surfaces. The best design methods for preventing this type of ignition source are developed by understanding the fundamentals relevant to static charge and by using these fundamentals to design specific features within a plant to prevent the accumulation of static charge or to recognize situations where the buildup of static electricity is inevitable and unavoidable. 1. Despite considerable efforts. one or both of the materials are insulators or poor conductors. NFPA 69 recommends a target oxygen concentration for storage vessels of no more than 2% below the measured LOC. (Quincy. Household examples that result in a buildup of a static charge are walking across a rug. If. Standard on Explosion Prevention Systems. The calculated values are deliberately not shown in Table 7-1. 3-methyl-1-butane. When different materials touch each other. If both the materials are good conductors.isobutyl formate. more of the electrons remain on one surface than on the other. For instance. 7-2 Static Electricity A common ignition source within chemical plants is sparks resulting from static charge buildup and sudden discharge. lNFPA 69. and combing hair. and the magnitude of the charge is much greater. The clinging fabrics and sometimes audible sparks are the result of the buildup of static charge. . If the oxygen concentration is not continuously monitored. Direct and reliable experimental data under conditions as close as possible to process conditions is always recommended. placing different materials in a tumble dryer.er. removing a sweater. one material becomes positively charged and the other negatively charged. and acetone. Fundamentals of Static Charge Static charge buildup is a result of physically separating a poor conductor from a good conductor or another poor conductor. Upon separation. 1997 ed. then the equipment must not operate at more than 60% of the LOC. If the LOC is less than 5%. if the oxygen concentration is continually monitored. 1997). the target oxygen concentration is no more than 60% of the LOC. Static electricity is perhaps the most elusive of ignition sources.7-2 Static Electricity 307 oxygen concentration than the experimental value for all species in Table 7-1 with the exception of butane. serious explosions and fires caused by static ignition continue to plague the chemical process industry. For unavoidable static buildup design features are added to continuously and reliably inert the atmosphere around the regions where static sparks are likely. MA: National Fire Protection Association. or 40% of the LOC if the LOC is below 5 %. Other methods are available to estimate the target gas concentration for placing a vessel into or out of service. the electrons move across the interface from one surface to the oth. the charge buildup as a result of separation is small because the electrons are able to scurry between the surfaces. however.

and interface. the static buildup created by walking across a carpet averages about 20 mJ and exceeds several thousand volts. For industrial operations where flammable vapors may be present. particle. An electrostatic discharge occurs when two materials at different potentials or polarities come close enough together to generate a charge transfer. currents and potentials in flow systems. In an explosive environment this sudden transfer of charges may be energetic enough to be an ignition source.A. capacitance. and Applications (Bristol: Adam Higler. are brought into contact. 3. 1987). with one being an insulator. When a metal object is touched. and (3) how to estimate the resulting energy discharged in relation to the minimum ignition energy (MIE) of the explosive environment. pneumatically conveying solids. relaxation times. shoulders. Cross. Subsequent grounding produces large and energetic sparks. Static charges of this magnitude are easy to generate. Charging by transport: When charged liquid droplets or solid particles settle on an isolated object. (2) how charges discharge by means of charge transfer. These relationships may include field strengths produced by static charges. thus accumulating an equal quantity of positive charges on the opposite side of the body. gas-liquid. any charge accumulation exceeding 0. mixing immiscible liquids. Induction charging: This phenomenon is applicable only to materials that are electrically conductive. some of the charges remain separated. and leaking steam that contacts an ungrounded conductor. giving the two materials opposite but equal charges. . Problems. a pipe wall. Electrostuti~~r: Principles. 2J. Charge Accumulation There are four charge accumulation processes2 that are relevant to dangerous electrostatic discharges in a chemical plant: 1. Double-layer charging: Charge separation occurs on a microscopic scale in a liquid at any interface (solid-liquid. 2. for example. for example. If the two objects are then separated. and arms) migrate toward the positive charge of the vessel. creating a spark. one must understand (1) how charges accumulate on objects. Basic electrostatic relationships are used to understand and investigate the sample situations. Electrons in the person's body (head. may approach an overhead vessel that is positively charged (previously filled with positively charged solids). a charge separation occurs at the interface. This leaves the lower part of the body positively charged by induction. As the liquid flows.1 mJ is considered dangerous. The static charges in these examples accumulate to develop large voltages. The transferred charge is a function of the object's capacitance and of the conductivities of the droplet. A person with insulated shoes. and many more. To prevent these ignitions. or liquid-liquid). the object is charged.308 Chapter 7 Designs to Prevent Fires and Explosions Common industrial examples are pumping a nonconductive liquid through a pipe. there is a transfer of the electrons. 4. electrostatic potential. Contact and frictional charging: When two materials. it carries a charge and it leaves a charge of opposite sign on the other surface.

lining 1 Tramp metal II A I L C Solids Metal shell and metal nozzle Figure 7-9 Common electrostatic discharges. These discharges 3T. the electrons move to exit at a single point of the charged object. (2) propagating brush. Electrostatic Discharges A charged object can be discharged to a ground or to an oppositely charged object when the field intensity exceeds 3 MVIm (breakdown voltage of air) or when the surface reaches a maximum charge density of 2. Conical pile . and (6) corona discharges.7 X lop5C/m2by six methods: (1) spark. Jones and J. ( 5 ) lightning-like. 1991). Apropagating brush discharge (Figures 7-9 and 7-10) is a discharge from a grounded conductor when it approaches a charged insulator that is backed by a conductor. (4) brush. Because both objects are conductive. (3) conical pile (sometimes known as Maurer discharge). A spark discharge (Figure 7-9) is a discharge between two metallic objects. King. This is therefore an energetic spark that can ignite a flammable dust or gas. B. L.7-2 Static Electricity Grounded thermo well. discharge Nonconductive . . and they enter the second object at a single point. grounds all metal \ Solids Lightning-like Propagating brush discharge . Powder Handling and Electrostatics (Chelsea. M I : Lewis Publishers.

" Journal of Electrostatic's (1993). are energetic.Grounded v conductor Figure 7-10 Propagating brush discharge. 30: 123-134. Maurer. 3: 189-195. "Discharges Due to Electrostatic Charging of Particles in Large Storage Silos. "Ignition Tests with Discharges from Bulked Polymeric Granules in Silos (Cone Discharge).t--z-. ." German Chemical Engineering (1979). and they can ignite flammable gases and dusts.5 The necessary conditions for this discharge are (1) a pow4B. Glor and B. Side view Grounded conductor Layer of dense charge (before . 5M. Data show that propagating brush discharges are not possible if the breakdown voltage of the insulator is 4 kV or less4 A conicalpile discharge (Figure 7-9) is a form of a brush-type discharge that occurs at the conical surface of a pile of powder.p~ Chapter 7 Designs to Prevent Fires and Explosions Top view of non-conductor brush discharge ------_------------------. Maurer.

7-2 Static Electricity 311 Corona discharge Conductor / (diameter < 5 mm) Figure 7-11 Corona discharge. Maurer. propagating 6P. Luttgens. This discharge is less intense compared with the point-to-point spark discharge. These are relatively intense discharges with energies up to several hundred millijoules." Journal of Electrostatics (1977). therefore they can ignite flammable gases and dusts. It is known from experiments that lightning-like discharges do not occur in vessels with volumes less than 60 m3 or in silos with diameters less than 3 m. (3) a powder with a high charge to mass ratio (for example. Energy from Electrostatic Discharges The energy generated in electrostatic discharges compared with the minimum ignition energies of gases and vapors and dusts is illustrated in Figure 7-12. The electrode conductor has a sharp point. To ignite dusts.6 There is currently no physical evidence that lightning-like discharges have resulted in industrial deflagrations. B. the results illustrate that flammable gases and vapors can be ignited by spark. conical pile. . The discharge from such an electrode has sufficient energy to ignite only the most sensitive gases (for example. (2) a powder with coarse particles (>1mm in diameter). brush discharges can ignite flammable gases. In general. This discharge radiates from the conductor in a brush-like configuration. charged by pneumatic transport). G. der with a high resistivity (>lo1"ohm m). W. Wider. Lightning-like discharges (Figure 7-9) are discharges from a cloud in the air over the powder. However. A brush discharge (Figure 7-9) is a discharge between a relatively sharp-pointed conductor (radius of 0. 3: 303-310. "An Experimental Contribution to the Question of the Existence of Lightning-Like Discharges in Dust Clouds. brush. the coarse particles need a fraction of fines to give an explosive atmosphere. Boschung. and (4) filling rates above about 0.1-100 mm) and either another conductor or a charged insulated surface.5 kgls. and A. and propagating brush discharges and that flammable dusts can be ignited only by sparks. Hilgner. and it is unlikely to ignite dusts. hydrogen). A corona discharge (Figure 7-11) is similar to a brush discharge.

Electrostatic Hazards in Powder Handling (New York: Wiley. the capacitance of the object (C in farads). or voltage (V in volts) of the object. 1988). Glor and B. brush." Journal of Electrostatics (1993). These three variables are related by the expression C = QIV. Maurer. The regions enclosed by the dotted lines in Figure 7-12 indicate regions of uncertainty. Energy of Electrostatic Ignition Sources A spark is generated between two conductors when the distance between the conductors is small compared to the diameter of the conductors and when the electric field intensity between the conductors is approximately 3 MVlm. and conical pile discharges. and the potential. A brush discharge is generated if the distance between the conductors is large compared to the radius of curvature of the conductor. Glor.312 Chapter 7 Designs to Prevent Fires and Explosions Energies to ignite Energies generated Figure 7-12 Minimum ignition energies compared to electrostatic discharge energies. Adapted from M. The actual energy (expressed in joules) associated with the discharge process is given by . The energy of a spark discharge is a function of the accumulated charge (Q in coulombs) on the object. "Ignition Tests with Discharges from Bulked Polymeric Granules in Silos (Cone Discharge). 30: 123-134. and M.

an electrostatic charge develops on the streaming material. 1999). and r is the liquid relaxation time (seconds). this accumulated charge is usually the result of either contact or friction charging for flowing solids and double-layer charging for flowing liquids. Britton. capacitance and voltage are not defined in nonconductive systems. d is the pipe diameter (m). fluid velocity. A precise experimental determination of the MIE is often required under the specific conditions of the system. u is the velocity (mts). A criterion that is commonly used to estimate the potential hazard of a discharge is to compare the MIE of the fuel-air mixture to the equivalent energy of the discharge. . and fluid properties is given by where Is is the streaming current (amps). The relaxation time is the time required for a charge to dissipate by leakage. G. but it is used qualitatively for the other discharges. The static discharge energy is a function of the accumulated charge. The relation between a liquid streaming current and the pipe diameter. When a liquid or solid flows through a pipe (metal or glass). In each case the charge (electrons) is transported with the material. MIEs are shown in Table 6-4 for a number of flammable gases and in Table 6-8 for dusts. Therefore Equation 7-20 is valid only for capacitive sparks. L is the pipe length (m). pipe length. Avoiding Static Ignition Hazards in Chemical Operations (New York: American Institute of Chemical Engineers. It is determined using 7L.7-2 Static Electricity Equation 7-20 assumes a capacitance-type discharge (that is. In an industrial setting. Streaming Current A streaming current Is is the flow of electricity produced by transfering electrons from one surface to another by a flowing fluid or solid. a spark). This current is analogous to a current in an electrical circuit. The measure of this flow of electrons is a streaming current and is expressed in coulombs per second or amps. however.

1 ~ 1.33 x 1 0 . 3d ed. 1950).9-2. is the specific conductivity (mholcm).85 x lo-'' y.0 3.0 lJ. where T E. 2Resistance = l/conductivity = l/(mho/cm) = ohm cm. that is. Because solid geometries are almost always ill-defined.0-7.0 6. coulomb2 s = 8.5 x 10-1' 10-10to 10-l3 1.5 4.5 2.3 3. H. (New York: McGraw-Hill.8 1.2 X 10-l8 0.314 Chapter 7 Designs to Prevent Fires and Explosions Table 7-2 Material Liquids Benzene Toluene Xylene Heptane Hexane Methanol Ethanol Isopropanol Water Properties for Electrostatic Calculations1 Specific conductivity2 (mholcm) Dielectric constant Other materials and air Air Cellulose Pyrex Paraffin Rubber Slate Teflon Wood 1. is the relative dielectric constant (unitless). Specific conductivities and relative dielectric constants are listed in Table 7-2. The buildup results from the separation of solid particle surfaces.0 x 1. O 3.85 x 10-l~and N m2 ohm cm ' 8. p.9-7. is the permittivity constant. Perry. electrostatic calculations for solids are handled empirically. .0 x lo-s 0. Chemical Engineers' Handbook.0 x lo-14 10-l6 to 0. is the relaxation time (seconds). Charges also accumulate when solids are transported. 1734.

G. as shown by where I. and kgls is the solids flow rate.65 X 1 0 ~ ~ o u l o m b / c m 2 Voltage to produce spark between needle points 112 in apart Voltage to produce spark between plates 0. "Electrostatics.000V 350 V 2. Table 7-4 Accepted Electrostatic Values for Calculations1 14. 163. Mancini.1 V . 1988) 7(1): 24." Chemical Engineering (March 13. The streaming current that is generated while transporting solids is a function of the solids processing method (see Table 7-3) and the flow rate.01-0." Plant/Operations Progress (Jan.01 mrn apart Maximum charge density before corona discharge Minimum ignition energies (mJ) Vapors in air Mists in air Dusts in air Approximate capacitances C (micro-microfarads) Humans Automobiles Tank truck (2000 gal) Tank ( 1 2 4 diameter with insulation) Capacitance between two 2-in flanges (118-in gap) Contact zeta potentials IF.000 20 0. "The Use (and Misuse) of Bonding for Control of Static Ignition A. Hazards. Eichel. p. Some generally accepted guidelines for electrostatic calculations are shown in Table 7-4.1967). is coulombs1second or amps. 100 to 400 500 1000 100. coulombslkg is given in Table 7-3.7-2 Static Electricity 315 Table 7-3 Process Charge Buildup for Various Operations1 Charge (coulomb/ kg) l0-"0 Sieving Pouring Grinding Micronizing Sliding down an incline Pneumatic transport of solids lo-" lo-7 to to lo-7 to lo-7 to to lo-7 lR.

and substitutions give . and the area A of the conductor (in cm2): This relationship shows that as the area of the conductor increases. the integration gives Equation 7-20.316 Chapter 7 Designs to Prevent Fires and Explosions Glass Lined Pipe I Metal Pipe Wher esistance Glass Vessel Figure 7-13 Electrical charge accumulation in a feed line resulting from fluid flow. Because V = QIC. (in mholcm). Fluid flows through the feed line and drops into the tank. The streaming current builds up a charge and voltage in the feed line to the vessel and in the vessel itself. the length of the conductor L (in cm). where V is the potential difference and the charge is Q. The voltage from the electrical ground in the metal line to the end of the glass pipe is calculated using The resistance R (in ohms) is computed using the conductivity of the fluid y. Energy of Charged Capacitors The amount of work required to increase the charge on a capacitor from Q to Q + dQ is dl = V d Q . Electrostatic Voltage Drops Figure 7-13 illustrates a tank with a feed line. the resistance increases. and if the conductor length increases. the resistance decreases.

150 gpm The data are: Hose length: 20 ft Hose diameter: 2 in .. 1gpm b. small drums 50-100-gal containers Person Automobile Tank truck lR. as shown in Figure 7-14. beer can." Plant/Operations Progress (Jan. Equation 7-28 assumes that the system starts with no accumulation of charge. where I. Q in coulombs. Assuming a constant streaming current. Example 7-4 Determine the voltage developed between a charging nozzle and a grounded tank. is in amps and t is in seconds. Charges can accumulate as a result of a streaming current dQldt = I. V in volts. only one constant source of charge Is. 7(1): 24. tools Buckets. compute the energy stored in the nozzle and the energy accumulated in the liquid. Mancini. A. Capacitances of various materials used in the chemical industry are given in Table 7-5. and J in joules. Also. The units used in Equations 7-26 and 7-27 are usually C in farads.7-2 Static Electricity Table 7-5 Object Capacitance of Various Objects1 Capacitance (farad) 5x 20 x 100 X 200 X 500 X 1000 x 10-l2 10-l2 10-l2 10-l2 10-l2 lo-'* Small scoop. 1988). "The Use (and Misuse) of Bonding for Control of Static Ignition Hazards. Explain the potential hazards in this process for a flow rate of a. and no current or charge loss terms (see the section "Balance of Charges" for a more complex system).

The average velocity in the pipe is .00 X lo9ohm. we obtain = 3.14)(1 in)2(2. Because the hose and nozzle are not grounded.7 Density: 0. The streaming current is a function of the velocity and the pipe diameter.318 Chapter 7 Designs to Prevent Fires and Explosions I Non-Conducting Hose Steel Nozzle u S t o r a g e Vessel 300 G a l l o n S t e e l G r o u n d e d Vessel / Ground Figure 7-14 System for Example 7-4.54cmlin)' = 20.88 g/cm3 Solution a.The resistance is computed using Equation 7-25 for the conducting fluid with a resistance length equivalent to the hose length (from the ground near the pump to the nozzle) and a resistance area equivalent to the cross-sectional area of the conducting fluid: L = (20 ft)(12 in/ft)(2.3 cm2. Using Equation 7-25. the voltage generated at the nozzle tip is V = ZR.54 cmlin) = 610 cm. Liquid conductivity: mholcm Dielectric constant cr:25. A = rrr2 = (3.

28 )]'[I 0) = - exp(. assuming that the nozzle is grounded.6 x 10-~)(2. Method 2: Compute the energy accumulated in the capacitor formed by the tank of liquid. C = 20 x 10-l2 farads = 20 X lo-'' coulomb/volt.1 mJ). The voltage is therefore V = I R = (2.48 X 10-'I amp. The accumulated charge is computed using Equation 7-28: with the time equal to the filling time of the vessel: .49 X 10-l4 Joules. - YC lo-' mholcm The streaming current is now determined using Equation 7-21: 10 X amp ( m / ~ ) ' ( m ) ~ [(3.48 X lo-'' amp)(3. Method 1: Compute the energy accumulated in the capacitor formed between the flanges at the nozzle.27 x (9. This is significantly lower than the energy required to ignite a flammable gas (0. - The capacitance between the two 1-in flanges is given in Table 7-4.074 volt. 0. such as a grounded thermocouple.074)' I = 5.= 2 CV2 [20 X lO~'(O. The energy is determined using Equation 7-26: J = --.*that is. A spark between the flanges may be an ignition source.0 = X lo9ohm) .102 x 2. therefore there is no hazard at the nozzle.0. The voltage drop down the 20-ft line is the same as the voltage drop from the hose flange to the nozzle flange.7-2 Static Electricity The relaxation time is estimated using Equation 7-22: 7=-= . A brush discharge can jump from this liquid to a metal component.1 = ] 2 X lo-')( - 12 X 3.58X 10-')(I 2.

The accumulated energy is again computed using Equation 7-26: J=-= cv2 2 (20 X 10-12)(2.0 X lo9 ohm: The streaming current is 10 X ][(4. b. that is. This is greater than the energy required to ignite a ffammable gas (0. 150 gpm versus 1 gpm for case a: u = (0.102 p)(lgprn) 150 gpm = 4.000 s) 4. therefore C = 100 X 10-l2 farads = 100 X coulomb/volt.66. This case is identical to case a except that the flow rate is higher. Method 1: Compute the energy accumulated in the capacitor formed between the flanges at the nozzle: V = ZR = (8. and the accumulated energy is determined using Equation 7-20: J=-= e2 (4. shown in Table 7-4.99 mJ.46 X 2C co~lomb)~ = 9.08 x 10-~(1.0) 8.9 X 2(100 X 10-12 farads) Joule = 0.. .08 X amp)(3 X lo9ohm) = 2.46 x coulomb. m The resistance is the same as for case a.1)]'[1 (m/~)~(m)~ = amp - exp ( - 15.3 X 2.1 mJ).320 Chapter 7 Designs to Prevent Fires and Explosions Substitution into Equation 7-28 gives Q = Ist = (2. = The capacitance of the liquid is estimated to be one-tenth of the capacitance of a 2000-gal vessel.1 mJ).42 X lo6)' 2 = 117 Joules. 3.42 X volts.08 x amp.27 X lo-' = 8. In this situation the vessel should be purged with nitrogen to keep the concentration of the flammable vapor below the LFL.46)(6. This exceeds the energy required to ignite a flammable gas (0.48 x lo-'' amp)(18.

Several examples are shown in Figure 7-15. This problem illustrates the importance of inerting.7 r is the sphere radius. For example. In the chemical industry the properties of the developed capacitor are estimated by assuming parallel flat plate or spherical geometries. the capacitance of a tank or a person is estimated by assuming spherical geometries.7-2 Static Electricity 321 Method 2: Compute the energy accumulated in the capacitor formed by the tank of fluid: t = 150gpm min 300 -(-)gal 60 s = 160s.t = (8. the capacitance for a spherical body is where is the relative dielectric constant (unitless). The capacitance C of a body is QIV. because C = QIV.08 X 10-4)(160) = 0. E. and the capacitance of a person's shoe sole or of a noncorrosive tank lining is estimated assuming parallel flat plates.13 coulombs. It is relatively easy to build up energies that exceed 0. Q = I.1 mJ.1 mJ. This energy exceeds 0. is the permittivity (8. For a sphere with radius r the voltage developed when a charge Q is accumulated is derived with elementary physics: Therefore.85 X 10-l2 coulomb2/N m2 = 2. and C is the capacitance. For two parallel plates . EO X 10-l2 coulomb/volt ft). Capacitance of a Body The buildup of a charge on one surface relative to another surface produces a capacitor.

..l.ec . t J Shoe Sole Ground Glass Lining ( ~ i e l e c t r i c ) Metal Vessel Person (Not rounded) Figure 7-15 Different types of industrial capacitors.322 Chapter 7 Designs to Prevent Fires and Explosions Flat Plates Condenser Shoe + Tank Gro Spheres Air (~ielectric) ++ Sphere (Not Grounded) Glass o r P l a s t i c T a n k (Air is Dielectric) + - 9er .

Solution This person's capacitance is estimated assuming that the person's shape is spherical and that the "sphere" is surrounded by air (E.7 = X coulomb) (6. ft 10-l2 . Given Shoe sole area (ft2) = 2 shoes (0. Example 7-5 Estimate the capacitance of a person (6 ft.2 in Dielectric constant of shoe soles = 3.is 1. which for flat parallel plates is (36)(2. we have 2.1.0 for air).8 n2) volt ft ) = 4.54 x 10-lo farads.7-2 Static Electricity Therefore the capacitance between parallel plates is where A is the area of the surface and L is the thickness of the dielectric. Using Equation 7-30 for a sphere. Assume that the person's shoe soles separate the person from the floor.7 X lo-"--- coulomb (0. Example 7-6 Estimate the capacitance of a person standing on a conductive floor.05 X 10-lo- coulomb volt ' The calculated capacitance is close to the value listed for a person in Table 7-5. .volt ft ) 1. 2 in tall) standing on a dry wooden floor. the shoe sole is the dielectric of the capacitor. that is.5 Solution Use Equation 7-32.4 ft2each) Shoe sole thickness = 0.

and charge loss resulting from relaxation. using Equation 7-20.454 kg) lb = 1. as a result of a person (insulated from the floor) charging 30 lb of a dry powder. An example is illustrated in Figure 7-16.8 Joules.. The result is where is the streaming current entering the tank through a specific inlet line i from a set of n lines. Balance of Charges Some systems are more complex than those previously discussed. Solution This operation is a sliding-contact type operation. ..36 X LO-' coulomh. Assume that the person's capacitance is 300 X lo-'' farad. a vessel may have several inlet lines and several outlet lines. The accumulated energy. if a grounded object of any kind is placed close to the solids. From Table 7-3 this operation gives a charge of coulomb/kg.. using a scoop.36 X 2(300 X J = .. An equal and opposite charge is also accumulated in the powder in the insulated drum. the charge carried away by flows going out. farad) These results illustrate that the energy exceeds the requirement for generating a spark capable of igniting a flammable gas. This spark would be discharged if the person approached a ground with a hand or with the scoop. into a 20-gal insulated drum. The charge balance is developed by considering the currents streaming in. Therefore the charged powder is another ignition source.= coulomb^)^ 2C lo-'' = 30.. and accumulated energy. (Is). is the current leaving through one specific outlet line j from a set of m lines. Therefore the charge buildup is c0u10mbs)(30 lb) (0.324 Chapter 7 Designs to Prevent Fires and Explosions Example 7-7 Estimate the charge buildup. is Q2 (1. an energetic spark could be generated. for example. For this type of system a charge balance is required to establish the charge and accumulated energy as a function of time. For example.

and relaxation times are constant. - - \ Figure 7-16 Any Vessel Vessel with multiple inlets and outlets. is a function of the charge accumulated in the tank and the rate of discharge F from the specific outlet nozzle j: where V . (ls)i. streaming currents.7-2 Static Electricity 325 ( ' ( ' ~ IN ~ $-1 2 - - - ( ' S ) 1 OUT ( ' S 1 2 OUT ~ IN ~ - . Q/Tis the charge loss resulting from relaxation.ou. Equation 7-35 is a linear differential equation that can be solved using standard techniques. is the container or tank volume and Q is the total charge in the tank. and T is the relaxation time. Substituting Equation 7-34 into Equation 7-33 gives If the flows. The result is .

An example of a sequential operation is (1)charging benzene to a vessel at a specific rate through a specific line of known size. rates. Equation 7-33 reduces to . Solution Because there is only one inlet line and no outlet lines.5 X amp.000 gal) is being filled with toluene. In this case Q is computed for each step with the specified C (I.)&.4.000 gal) and where F = 100 gpm. and (4) discharging the batch through a different line at a specified rate. Liquid conductivity = 10-l4 mho cm-l. and C (FnIVc)for that particular step. is the initial charge in the tank at t = 0. is the result from the previous step. and materials of construction are known. Is = 1. Example 7-8 A large vessel (50. the potential hazard of each step of the operation can be estimated. and the initial Q. Compute Q and J during the filling operation when the vessel is half full (25. (3) holding the batch for a specified time. (2) charging methanol and toluene through two different lines at different rates.326 Chapter 7 Designs to Prevent Fires and Explosions where Q. If the line sizes. Equation 7-36 is also used when the filling and discharge rates are sequential. These equations plus the equations described previously (equations for Is and J ) are used to compute Q and J as a function of time. and Dielectric constant = 2. Therefore the hazards of relatively complex systems can be evaluated.

e-'IT) = 1.14 X X 10-12- coulomb volt ft )(9. The charge buildup as a function of time is Q(t) = lS7(1 .85 7 & Ell . Therefore Q(15.27 ft) = 10-In farads.4)(8. The relaxation time is computed using Equation 7-22: (2.l ) X lo-"- = 21. Using Equation 7-30 and assuming a dielectric of 1 for air. r = 4(3. ~ . therefore the operating conditions for this vessel are extremely hazardous.2 s)(l - e-1i21. .000 gal.2).2 s.48 gal ft-3 ) = 9. the elapsed time is 15.19 X coulombs.000 gal 1 47~ 7.10 mJ.27 ft. When the vessel contains 25.7 3.14)(1.7-2 Static Electricity 327 Therefore Because the vessel is initially empty. The capacitance of this vessel is estimated by assuming a spherical geometry surrounded by air: = ( 25.000 gal of toluene) is computed using Equation 7-20: The minimum condition for an ignition is 0.0) 2. we obtain C = 4 n e .000 s. The energy stored in this vessel (25.OOO s) = 3. Q. = 0.5 X lo-'- coulomb S )(21. = = YC ohm cm (10-l4 mho ~ m .

o. The residence time of this vessel is Residence time = ( .328 Chapter 7 Designs to Prevent Fires and Explosions Overflow - P Toluene - 1 0 = 2x10 -'Coulomb in I n i t i a l (Charge Liquid a t T i m e = 0) Figure 7-17 Charge buildup with complex vessel system.~ m ) l ~ ~ - (s) = 3. = 1. Q and J when the vessel fluid just reaches the overflow line (start with an empty vessel). .5 X amp (high value due to filter in line) Liquid conductivity = 10-14 mho/cm Dielectric constant = 2.00 s.4) (sss x lo-" .ohm cm 10-14 -mh0 cm Er 80 7=-= = Yc 21. Q and J under equilibrium conditions ( t = 0 ) 0. The time required to reduce the accumulated charge to half the equilibrium charge if the flows are stopped after equilibrium conditions are reached. Example 7-9 Figure 7-17 shows an in-line trap for removing water from a process stream.2 s. The relaxation time is determined using Equation 7-22: (2. Given: Volume of vessel = 5 gal Flow rate = 100 gpm toluene Streaming current I. Compute: a. c. The charge removed with the discharge under equilibrium conditions. b.4 Initial vessel charge = 2 X lo-' coulomb Solution a. d..

The vessel capacitance is calculated by assuming a spherical geometry with the surrounding air serving as the dielectric.5 X = coulomb S (21.84 X X coulomb 10-l2 .~. Equation 7-36 for this case reduces to (I(t = 00)= - - % ' = 3.55 mJ) greatly exceeds the quantity required for ignition of flammable materials." ~ ' .24 ~ ) ] e . the radius of this sphere is = 0. ' 3. Equations 7-35 and 7-36 reduce to Q(t) = I S r + (Q.55 mJ.e.7-2 Static Electricity 329 During the filling operation.volt ft )(0. with Q(t) in coulombs and tin seconds. b. 21.84 X lo-" farads) The accumulated energy (9.542 ft) = lo-'' farads. This system is operating under hazardous conditions. This vessel will gradually level off to steady-state equilibrium conditions when the operating time significantly exceeds the relaxation time.2 3 .2 s) - + [ 2 x lo-' = coulomb X 1.5 X amp(21. therefore the exponential term of Equation 7-36 is 0.2. Because 5 gal = 0.IS7)e-'/' 1. The energy accumulated in this vessel is estimated using Equation 7-20: J=-= e2 2C (5.r = 4 ~ ( 1 . before the liquid level reaches the discharge line. At 3 seconds This is the charge buildup just before reaching the overflow line. 0 )2.18 x lo-' .7 1. 2(1.98 10-~e-"~'. .93 x co~lomb)~ = 9.94 X 1U-' coulomb.542 ft.668 ft3. The capacitance is estimated using Equation 7-30: C = 4~re.

are zero. and (I.22 mJ.). c.86 X lo-' .. The energy is determined by using Equation 7-20: Q~ 2C J=-= (3.94 x 2(1. In practice.). Hopefully this emphasis on the fundamentals will make the subject less elusive and destructive.31 X coulomb S For this example the charge loss resulting from flow out of a system is greater than the loss resulting from relaxation.5.. resulting sparks. = 0. d. from the problem definition. however. design engineers recognize .94 x Charge loss via relaxation coulomb. Although there is an additional loss of charge with the overflowing liquid. The examples and fundamentals developed in these sections were designed to emphasize the importance of this subject. Sparks resulting from static charge and discharge continue to cause major fires and explosions within the chemical industry. (I. 7-3 Controlling Static Electricity Charge buildup.84 X lo-" farads. and =-= Q coulomb 1. S F Charge loss via the overflow = -Q v c = 1. After the inlet flow is stopped. and Equation 7-36 reduces to For QIQ. Q(t = co) = 3. Therefore it only takes about 15 s to reduce the accumulated charge to one-half its original charge.330 Chapter 7 Designs to Prevent Fires and Explosions From part a the capacitance is C = 1.84 coulomb^)^ X = lo-" farads) 4. Under equilibrium conditions Equation 7-35 is set to zero: and from part b.. and the ignition of flammable materials is an inevitable event if control methods are not appropriately used. the system is still operating under hazardous conditions.

General Design Methods To Prevent Electrostatic Ignitions The design objective is to prevent the buildup of charges on a product (liquid or powder) as well as on surrounding objects (equipment or personnel). Isolated objects are notorious for building up large potentials and energetic sparks when they are approached by another conductor at a lower potential. 2. Nonconductive coatings with a . to eliminate the accumulation of a high-density charge on the metallic interface and a countercharge on the nonconductor surface. It is always used when working with flammable liquids that are 5°C (or less) below the flash point (closed cup). When dangerous discharges cannot be eliminated. This method is generally used when handling powders. then prevent the possibility of an ignition by maintaining oxidant levels below the combustible levels (inerting) or by maintaining fuel levels below the LFL or above the UFL.) of 4 kV.The most effective way of preventing ignitions from conical pile discharges is inerting. Prevent charges from accumulating to dangerous levels by reducing the rate of charge generation and increasing the rate of charge relaxation. Three methods are used to achieve this objective: 1.7-3 Controlling Static Electricity 331 this problem and install special features to prevent (1) sparks by eliminating the buildup and accumulation of static charge and (2) ignition by inerting the surroundings. Grounding and bonding are used especially to prevent the existence of isolated metal parts or objects. These discharges are also prevented by keeping the metallic backings grounded. deflagration venting and explosion suppression). The special design features for preventing electrostatic ignitions are described in the following paragraphs. by decreasing the charge rate below 0. Sparks are prevented by grounding and bonding. For every charged object there exists its oppositely charged counterpart. Prevent charges from accumulating to dangerous levels by designing the system to include charge reduction by means of low-energy discharges. or by using containers with a volume less than 1m3. Methods for preventing charge buildup are described in the following paragraphs. This procedure prevents two metallic objects (close to each other) from having different potentials. This method is generally used when handling liquids. Brush discharges are prevented by keeping the nonconductive surfaces thin enough or conductive enough to have a breakdown voltage (U. 3. Conical pile discharges are prevented by increasing the conductivity (additives). Measures to mitigate the consequences of an explosion are also options for consideration (for example.5 kgls. Propagating brush discharges are prevented by keeping the nonconductive surfaces or coatings thin enough or conductive enough to have a breakdown voltage below 4 kV. Inerting (section 7-1) is the most effective and reliable method for preventing ignition.

provides a safety factor of 4. Bonding and Grounding The voltage difference between two conductive materials is reduced to zero by bonding the two materials. When comparing sets of bonded materials. the sets may have different voltages. Examples of grounding and bonding are illustrated in Figures 7-18 and 7-19. that is. by grounding. The residence time in this relaxation section of pipe should be about twice the relaxation time determined from Equation 7-22. however. "Electrostatics. and fixed charges cannot be transferred in a brush discharge. In actual practice: it was found that a hold time equal to or greater than one-half the calculated relaxation time is sufficient to eliminate charge buildup. therefore. Brush discharges from nonconductive liquids are prevented by increasing the conductivity using conductive additives.G. Lightning-like discharges are prevented by keeping the vessel volume to less than 60 m" or the vessel diameter to less than 3 m. less than 4 kV.5 m2/s for road tanker filling and 0. This fixes the charges accumulated on the nonconductor. If this condition is not met.332 Chapter 7 Designs to Prevent Fires and Explosions thickness greater than 2 mm. DC: American Petroleum Institute." Chemical Engineering (Mar. This hold provides time for charge reduction by relaxation. 8F. Relaxation When pumping fluids into a vessel through a pipe on top of the vessel. of less than 0. 153. are capable of brush discharges even with a U . then the system needs to be inerted. and Stray Currents (Washington. . eliminating the potential for static sparks. To prevent brush discharges. Protection Against Ignitions Arising Out of Static. This also eliminates the charge buildup between various parts of a system. the separation process produces a streaming current Is. 1991). which is the basis for charge buildup.1967). The voltage difference between sets is reduced to zero by bonding each set to ground. from Equation 7-21. that is. It is possible to substantially reduce this electrostatic hazard by adding an enlarged section of pipe just before entering the tank. 13. The "twice the relaxation time" rule. Lightning. Bonding and grounding reduces the voltage of an entire system to ground level or zero voltage. The American Petroleum Institute9 recommends a ud. a thickness of less than 2 mm is necessary.8 m2/sfor rail car filling. The most effective way of preventing ignitions from brush discharges is inerting. by bonding one end of a conducting wire to one of the materials and bonding the other end to the second material. Eichel. p. 9API RP 2003.

Another technique is to use an angle iron instead of a pipe and to let the liquid flow down the angle iron (see Figure 7-21). . These methods are also used when filling drums. In this case the fill line should extend to the bottom of the vessel (see Figure 7-21). This technique. reduces the electrical charge that accumulates when liquid is allowed to free fall. G. however. to help eliminate the charge generation (and accumulation) resulting from separation during the filling operation. Also. Eichel. When using dip pipes. is not effective when handling liquids with low conductivity. Dip Pipes An extended line. A commonly used method is to place a hole in the dip pipe near the top of the vessel. 1967). care must be taken to prevent siphoning back when the inlet flow is stopped. Adapted from F. 153. "Electrostatics. sometimes called a dip leg or dip pipe. (continued) Glass and plastic lined vessels are grounded using tantalum inserts or metal probes.- - - 7-3 Controlling Static Electricity Charging f r o m D r u m Figure 7-18 Bonding and grounding procedures for tanks and vessels. the inlet velocities should be low enough to minimize the charge generated by streaming current Is. p. 13. as illustrated in Figure 7-20." Chemical Engineering (Mar. however.

/ + Conductive Hose i r e s Bonded to Metal Pipe 2 Drum Type " A " Clamp Type " B " Clamp Figure 7-18 (continued) .334 Chapter 7 Designs to Prevent Fires and Explosions Grounded Water Line Conductive Hose Metal Pipe Type " A " Clamp 1 L T "B" Clamp ~ ~ ~ Grounded W a t e r Line Bond --- Type ' Clamp -K NF: Vacuum Metal T--I.

7-3 Controlling Static Electricity

Grounded Water

Llne

amp or

Strap

Fill

Line

Type

Cla

Wire

Bonded Tank

to

Metal

Pipe

Car o r Truck

Unloading Tank

Trucks

or C a r s

Figure 7-18

(continued)

C l a m p o r S t r a p o n Metal Pipe J-- C o n d u c t i v e H o s e

&
# 4 Wire

Metal Pipe to Metal Pipe Metal

'Glass or Plastic Valve

Solder Wire t o I

O n e Brass Bolt, N u t a n d Washer ~ e F l a n a e r

-

Glass

Pipe Gasket

Solder Wire t o

Brass Bolt

Non-Conducting

Figure 7-19 Bonding procedures for valves, pipes, and flanges. Adapted from F. G. Eichel, "Electrostatics," Chemical Engineering (Mar. 13, 1967), p. 153.

336

Chapter 7

Designs to Prevent Fires and Explosions

Probe

a V e s s ie l s _ s n_

Tantalum Insert

-

-

Ground

Ground

Figure 7-20

Grounding glass-lined vessels.

Hole f o r Siphon

z . -

'

Break Dip Leg
-

-

- - - ---

-

Liquid

Figure 7-21

Dip legs to prevent free fall and accumulation of static charge.

Increasing Conductivity with Additives
The conductivity of nonconducting organic materials can sometimes be increased using additives called antistatic additives. Examples of antistatic additives include water and polar solvents, such as alcohols. Water is effective only when it is soluble in the offending liquid, because an insoluble phase gives an additional source of separation and charge buildup.

7-4 Explosion-Proof Equipment and Instruments

337

Permissible:

Vessel and Surrounding Area Must be Free from Flammable Vapors, Gases, and Liquids

Figure 7-22 Handling solids with no flammable vapors present. Adapted from Expert Commission for Safety in the Swiss Chemical Industry, "Static Electricity: Rules for Plant Safety," Plant/Operations Progress (January 1988), p. 19.

Handling Solids without Flammable Vapors
Charging solids with a nongrounded and conductive chute can result in a buildup of a charge on the chute. This charge can accumulate and finally produce a spark that may ignite a dispersed and flammable dust. Solids are transferred safely by bonding and grounding all conductive parts and/or by using nonconductive parts (drum and chute). See Figure 7-22.

Handling Solids with Flammable Vapors
A safe design for this operation includes closed handling of the solids and liquids in an inert atmosphere (see Figure 7-23). For solvent-free solids the use of nonconductive containers is permitted. For solids containing flammable solvents, only conductive and grounded containers are recommended.1°

7-4 Explosion-Proof Equipment and lnstruments
All electrical devices are inherent ignition sources. Special design features are required to prevent the ignition of flammable vapors and dusts. The fire and explosion hazard is directly proportional to the number and type of electrically powered devices in a process area.
l0Expert Commission for Safety in Swiss Chemical Industry, "Static Electricity: Rules for Plant Safety," Plant/Operations Progress (January 1988),7(1): 1.

338

Chapter 7

Designs to Prevent Fires and Explosions

Bonds a n d Ground

Valve

Valve Nitrogen

I
Nitrogen

Charge Container Closed a n d Purged with Nitrogen

0 . 0 0 0 . 00.0

before Opening Valve

Shutoff

Valve

..* .

B o n d All G a s k e t e d F l a n g e s

(e.g.: B r a s s B o l t s )

do

\

- - -

Ground

Figure 7-23 Handling solids with flammable vapors present. Source: Expert Commission for Safety in the Swiss Chemical Industry, "Static Electricity: Rules for Plant Safety," Plant/Operations Progress (January 1988), p. 19. Reprinted by permission of the American Institute of Chemical Engineers, New York.

7-4 Explosion-Proof Equipment and Instruments

339

Most safety practices for electrical installations are based on the National Electric Code (NEC)." Although states, municipalities, and insurance companies may have their own installation requirements, they are usually based on the NEC. Process areas are divided into two major types of environments: XP and non-XP. XP, for explosion proof, means flammable materials (particularly vapors) might be present at certain times. Non-XP means that flammable materials are not present, even under abnormal conditions. For non-XP designated areas open flames, heated elements, and other sources of ignition may be present.

Explosion-Proof Housings
In an XP area the electrical equipment and some instrumentation must have special explosion-proof housings. The housings are not designed to prevent flammable vapors and gases from entering but are designed to withstand an internal explosion and prevent the combustion from spreading beyond the inside of the enclosure. A motor starter, for example, is enclosed in a heavy cast walled box with the strength needed to withstand explosive pressures. The explosion-proof design includes the use of conduit with special sealed connections around all junction boxes.

Area and Material Classification
The design of electrical equipment and instrumentation is based on the nature of the process hazards or specific process classifications. The classification method is defined in the National Electrical Code; it is a function of the nature and degree of the process hazards within a particular area. The rating method includes Classes I, 11, and 111, Groups A-G, and Divisions 1 or 2. The classes are related to the nature of the flammable material: Class I: Locations where flammable gases or vapors are present. Class 11: Same for combustible dusts. Class 111: Hazard locations where combustible fibers or dusts are present but not likely to be in suspension. The groups designate the presence of specific chemical types. Chemicals that are grouped have equivalent hazards: Group A: acetylene Group B: hydrogen, ethylene Group C: carbon monoxide, hydrogen sulfide

''NFPA 70, The National Electrical Code Handbook (Quincy, MA: National Fire Protection Association,
1996).

Chapter 7

Designs to Prevent Fires and Explosions

Group D: butane, ethane, ethyl alcohol Group E: aluminum dust Group F: carbon black Group G: flour Division designations are categorized in relationship to the probability of the material being within the flammable or explosive regions: Division 1:Probability of ignition is high; that is, flammable concentrations are normally present. Division 2: Hazardous only under abnormal conditions. Flammable materials are normally contained in closed containers or systems.

Design of an XP Area
When designing an XP area, all pieces of electrical equipment and instrumentation are specified for the class, group, and division, as discussed previously. All pieces of equipment and instrumentation within an area must be appropriately specified and installed. The overall classification is only as good as the piece of equipment in an area with the lowest classification.

7-5 Ventilation
Proper ventilation is another method used to prevent fires and explosions. The purpose of ventilation is to dilute the explosive vapors with air to prevent explosion and to confine the hazardous flammable mixtures.

Open-Air Plants
Open-air plants are recommended because the average wind velocities are high enough to safely dilute volatile chemical leaks that may exist within a plant. Although safety precautions are always practiced to minimize leaks, accidental releases from pump seals and other potential release points.

Example 7-10
A plant handling substantial quantities of flammable toluene is located 1000 ft from a residential area. There is some concern that a sizable leak of flammable vapors will form a flammable cloud with subsequent ignition in the residential area. Determine the minimum mass flow rate of toluene leakage required to produce a vapor cloud in the residential area with a concentration equal to the LFL. Assume a 5 mph wind and D atmospheric stability.

7-5 Ventilation

341

Solution
Assume a continuous leak at ground level. The plume concentration directly downwind along the cloud centerline is given by Equation 5-48:

Solving for Q,, the mass flow rate from the leak, we obtain

The LFL for toluene is 1.4% in air (from appendix B). Converting the units, we obtain

0,014

m3 toluene

1 g-mol toluene 22.4 X lo-"' toluene

92 g toluene

=

57.5 g/m3.

The wind speed is 5 mph = 2.23 m/s. The distance downwind is 1000 ft u, = 22 m and u, = 12 m. Substituting, we obtain Qm = (49.3 g/m3)(3.14)(22m)(12 m)(2.23 mls)
=

=

304 m. From Figure 5-10,

9.11

X

lo4 gls

Any leak with a flow rate greater than 201 lb/s is capable of producing a flammable cloud in the residential area. Of course, the toxic effects of this cloud must also be considered. The LEL of 57.5 gim3 is much above the TLV of 0.188 g/m3.

Plants Inside Buildings
Frequently, processes cannot be constructed outside. In this case local and dilution ventilation systems are required. These ventilation systems were discussed in detail in chapter 3, section 3-4. Local ventilation is the most effective method for controlling flammable gas releases. Dilution ventilation, however, is also used because the potential points of release are usually numerous and it may be mechanically or economically impossible to cover every potential release point with only local ventilation. There are empirically determined design criteria for designing ventilation systems for flammable materials inside storage and process areas. These design criteria are given in Table 7-6. The effectiveness of a ventilation system is determined using material balance equations, described in chapter 3 in the section "Estimating Worker Exposures to Toxic Vapors," and as illustrated in the following example.

342

Chapter 7

Designs to Prevent Fires and Explosions

Table 7-6
Type of area

Ventilation Data for Handling Flammable Materials1
Rate
1ft?min/ft2 of floor area

Conditions
(a) System interlocked to sound an alarm when ventilation fails (b) Locate inlet and exhausts to provide air movement across entire area (c) Recirculation is permitted but stopped when air concentrations exceed 25% of LFL (a) to (c) as for inside storage areas (d) Design ventilation system to keep concentrations at a 5-ft radius from all sources to below 25% of LFL

Ventilation for inside storage areas

Ventilation for inside process areas

1 ft"min/ft2 of floor area or more; see (d)

Class I: Flash-point (closed cup) below 37.8"C (100°F) 1 Class 1 :Flash-point from 373°C to 60°C (100°F to 140°F) Class 111: Flash-point above 60°C (140°F) 'Data taken from NFPA 30, Flammables and Combustible Liquids Code, 1996 ed. (Quincy, MA: National Fire Protection Association, 1996).

Example 7-1 1
Determine the concentration of toluene over a diked area (100 ft2) that contains toluene as a result of a spill. Assume that the process area (2500 ft2) is designed to handle Class I flammable materials and that the liquid and air temperature is 65°F. The vapor pressure of toluene at 65'F is 20 mm Hg. The LEL is 1.4% by volume.

Solution
The source models for spills are described in chapter 3, Equations 3-14 and 3-18. The concentration of volatiles in a ventilated area resulting from the evaporation from a pool is given by Equation 3-14:

where

K is the mass transfer coefficient, determined using Equation 3-18, A is the area of the pool, PSat the saturation vapor pressure of the liquid, is k is the nonideal mixing factor, Q, is the volumetric ventilation rate, and P is the pressure.
The ventilation capacity for this process area is based on the design criterion of 1 ft"min/ft2 (Table 7-6); therefore
1 ft'

ft3 (2500 ft2) = 2500 7. min

7-6 Sprinkler Systems

Also,

M

=

92,

PS"' 20 mm Hg, =
A
=

100 ft2.

The mass transfer coefficient is computed using Equation 3-18 with M,,, and KOfor water, that is, 18 and 0.83 cmls, respectively:

The nonideal mixing factor k ranges between 0.1 and 0.5. Because no information is given about the ventilation, k is used as a parameter. Substituting into Equation 3-14, we obtain

-

(0.948 ft/min)(100 ft2)(20/760) atm (2500 ft3/min)(l atm)

X

lo6
=

998 ppm.

The concentration range is estimated to be

CpP,=1996ppm=0.1996%byvolume, C,,,
=

fork=0.5, fork = 0.1.

9980 ppm = 0.998% by volume,

These concentrations are considerably below the LFL of 1.4% by volume, which illustrates that the specified ventilation rate for Class I liquids is satisfactory for handling relatively large spills of flammable materials. The concentrations do, however, exceed the TLV for this substance.

7-6 Sprinkler Systems
Sprinkler systems are an effective way to contain fires. The system consists of an array of sprinkler heads connected to a water supply. The heads are mounted in a high location (usually near ceilings) and disperse a fine spray of water over an area when activated. The heads are activated by a variety of methods. A common approach activates the heads individually by the melting of a fusible link holding a plug in the head assembly. Once activated, the sprinklers cannot be turned off unless the main water supply is stopped. This approach is called a wet pipe system. These systems are used for storage areas, laboratories, control rooms, and small pilot areas. Another approach activates the entire sprinkler array from a common control point. The control point is connected to an array of heat andlor smoke detectors that start the sprinklers when an abnormal condition is detected. If a fire is detected, the entire sprinkler array within an area

344

Chapter 7

Designs to Prevent Fires and Explosions

Table 7-7

Fire Protection for Chemical Plants1

Sprinkler system types Antifreeze sprinkler system: a wet pipe system that contains an antifreeze solution and that is connected to a water supply. Deluge sprinkler system: open sprinklers and an empty line that is connected to a water supply line through a valve that is opened upon detection of heat or a flammable material. Dry pipe sprinkler system: a system filled with nitrogen or air under pressure. When the sprinkler is opened by heat, the system is depressurized, allowing water to flow into the system and out the open sprinklers. Wet pipe sprinkler system: a system containing water that discharges through the opened sprinklers via heat. Design densities (see NFPA documents for details) Source of fire: not less than 0.50 gpm/ft2 of floor area. Pumps and related equipment: 0.50 gpm/ft2 of projected area. Vessels: 0.25 gpm/ft2of exposed surface, including top and bottom. Vertical distance of nozzle should not exceed 12 ft. Horizontal structural steel: 0.10 gpm/ft2of surface area. This may not be necessary if the steel is insulated or designed to withstand the worse-case scenario. Vertical structural steel: 0.25 gpm/ft2of surface area. This may not be necessary if the steel is insulated or designed to withstand the worse-case scenario. Metal pipe, tubing, and conduit: Not less than 0.15 gpm/ft2of surface areaand directed toward the undersides. Cable trays: Not less than 0.3 gpm/ft2 of projected plane area (horizontal or vertical). Combined systems: The NFPA standards specify acceptable methods for combining the above requirements. Nominal discharge rates for 0.5-in orifice spray nozzles are: gPm: psi:
18 10

25 20

34 35

50 75

58 100

lData taken from NFPA 13, Installation of Sprinkler Systems (Quincy, MA: National Fire Protection Associates, 1999); and NFPA 15, Standards for Water Spray Fixed Systems for Fire Protection (Quincy, MA: National Fire Protection Association, 1996).

is activated, possibly in areas not even affected by the fire. This approach is called a deluge system. This system is used for plant process areas and larger pilot plants. Sprinkler systems can cause considerable water damage when activated, depending on the contents of the building or process structure. Statistically, the amount of water damage is never as great as the damage from fires in areas that should have had sprinklers. Sprinkler systems require maintenance to ensure that they remain in service and have an adequate and uninterrupted water supply. There are various fire classes that require different sprinkler designs. The detailed descriptions of these classes and sprinkler specifications are given inNFPA 13.12 average chemiAn cal plant is classified as an ordinary hazard (Group 3) area. Various sprinkler specifications for this type of area are given in Table 7-7.
I2NFPA 13, Iizstallation o f Sprinkler Systems (Quincy, MA: National Fire Protection Association, 1999).

7-6 Sprinkler Systems

345

Sometimes vessels need special water protection to keep the vessel walls cool during fires. High surface temperatures can result in metal failure at pressures far below the vessel's maximum allowable working pressure (MAWP) with potentially disastrous consequences. In hydrocarbon spill fires unprotected vessels (no insulation or water spray) can fail within minutes. A water spray protection system around vessels is recommended to prevent this type of failure. These water spray protection systems, commonly called deluge systems, are designed to keep the vessel cool, flush away potentially hazardous spills, and help to knock down gas clouds.13 Deluge systems can also provide enough time to transfer material out of a storage tank into another (safe) area. Vessel deluge systems are usually designed as open-head systems, which are activated when a fire is detected and/or a flammable gas mixture is detected. The deluge system is usually opened when the flammable gas concentration is a fraction of the LFL (approximately 25 %) or when a fire is detected through heat. Table 7-7 provides descriptions and design specifications for these systems. Monitors are fixed water hydrants with an attached discharge gun. They are also installed in process areas and storage tank areas. Fire hydrants and monitors are spaced 150-250 ft apart around process units, located so that all areas of the plant can be covered by 2 streams. The monitor is usually located 50 ft from the equipment being protected.14Fire monitors discharge water at a rate of 500-2000 gpm.

Example 7-12
Determine the sprinkler requirements for a chemical process area within a building with an area of 100 ft by 30 ft that handles reactive solvents. Determine the number of sprinkler spray nozzles and pump specifications. Assume 0.5-in orifice sprinklers with 35 psig at each nozzle, giving 34 gpm each, a 10-psig frictional loss within the system, and a 15-ft elevation of the sprinkler system above the pump.

Solution
Data for designing this system are found in Table 7-7. Total water requirement
= =

(0.50 gpm/ft2)(100ft)(30 ft) 1500 gpm, (1500 gpm) = 44.L, (34 gpm/nozzle)

Number of sprinkler nozzles =

which is rounded to the next even number for layout convenience, or 44. The pressure required at the pump is the sum of the minimum pressure at the nozzle (specified as 35 psi), the pressure loss resulting from friction (10 psi), and the pressure resulting from the

13D.C. Kirby and J. L. De Roo, "Water Spray Protection for a Chemical Processing Unit: One Company's View," Plant/Operations Progress (October 1984), 13(4). I40rville M. Slye, "Loss Control Association," Paper presented at AICHE Symposium, New Orleans, Louisiana, March 6-10,1988.

346

Chapter 7

Designs to Prevent Fires and Explosions

Table 7-8
Feature

Miscellaneous Designs for Preventing Fires and Explosions1
Explanation
The best way to prevent fires and explosions is to stop the release of flammable materials. Preventive maintenance programs are designed to upgrade system before failures occur. Insulate vessels, pipes, and structures to minimize damage resulting from fires. Add deluge systems and design to withstand some damage from fires and explosions; e.g., use multiple deluge systems with separate shutoffs. Design control rooms to withstand explosions. Provide supply for maximum demand. Consider many deluge systems running simultaneously. Diesel-engine pumps are recommended. Place shutoffs well away from process areas. Install hydrants, monitors, and deluge systems. Add good drainage. Separate (space) plants on a site, and separate units within plants. Provide access from two sides. Design steam, water, electricity, and air supplies to be available during emergencies. Place substations away from process areas. Locate personnel areas away from hazardous process and storage areas. Group units in rows. Design for safe operation and maintenance. Create islands of risk by concentrating hazardous process units in one area. Space units so hot work can be performed on one group while another is operating. Install isolation valves for safe shutdowns. Install in safe and accessible locations at edge of unit or group. Process equipment should be separated from flares and railroads. Place gas compressors downwind and separated from fired heaters. Locate flammable storage vessels at periphery of unit. Dike vessels to contain and carry away spills. Automated block valves should be placed to stop andlor control flows during emergencies. Ability to transfer hazardous materials from one area to another should be considered. Add appropriate on-line analyzers to (1) monitor the status of the process, (2) detect problems at the incipient stage, and (3) take appropriate action to minimize effects of problems while still in initial phase of development. All controls need to be designed to fail safely. Add safeguards for automated and safe shutdowns during emergencies.

Maintenance programs

Fireproofing

Control rooms Water supplies Control valves for deluge Manual fire protection Separate units Utilities Personnel areas Group units

Isolation valves Railroads and flares Compressors Dikes Block valves

On-line analyzers

Fail-safe designs

'John A. Davenport, "Prevent Vapor Cloud Explosions," Hydrocarbon Processing (March 1977), pp. 205-214; and Orville M. Slye, "Loss Prevention Fundamentals for Process Industry," Paper presented at AICHE Loss Prevention Symposium, New Orleans, LA, March 6-10,1988.

Suggested Reading

347

pipe elevation over the pump (15 ft water or 6.5 psi). Therefore the total pressure is 51.5 psi, which is rounded up to 52 psi. The pump power is now determined:

HP Horsepower = (25,029 ft-lbf/s)

Therefore this sprinkler requires a pump with a capacity of 1500 gpm and a 45.5-HP motor, assuming an efficiency of 100%. Actually, fire pumps are usually designed with discharge pressures of 100-125 psig so that the hose and monitor nozzle streams will have an effective reach. In addition, the size of the monitor is governed by requirements in the fire codes.15

7-7 Miscellaneous Designs for Preventing Fires and Explosions
The successful prevention of fires and explosions in chemical plants requires a combination of many design techniques, including those mentioned previously and many more. A complete description of these techniques is far beyond the scope of this text. A partial list, shown in Table 7-8, is given to illustrate that safety technology is relatively complex (the appropriate application requires significant knowledge and experience) and to serve as a checklist for engineers to help them include the critical features for preventing fires and explosions.

Suggested Reading
R. Beach, "Preventing Static Electricity Fires," Chemical Engineering (Dec. 21,1964), pp. 73-78; (Jan. 4, 1965), pp. 63-73; and (Feb. 2,1965), pp. 85-88. L. G. Britton, Avoiding Static Ignition Hazards in Chemical Operations (New York: American Institute of Chemical Engineers, 1999). H. Deichelmann, The Electrostatic Charge of Class-Lined Vessels and Piping, Pfaudler PWAG Report 326e. J.S. Dorsey, "Static Sparks: How to Exorcise the 'Go Devils,"' Chemical Engineering (Sept. 13,1976), pp. 203-205. Fire Protection Handbook, 14th ed. (Boston: National Fire Protection Association, 1976), ch. 5. S.K. Gallym, "Elements of Static Electricity," Gas (March 1949), pp. 12-46. M. Glor, Electrostatic Hazards in Powder Handling (New York: Wiley, 1988). H. Haase, Electrostatic Hazards (New York: Verlag Chemie-Weinheim, 1977). T. M. Kirby, "Overcoming Static Electricity Problems in Lined Vessels," Chemical Engineering (Dec. 27, 1971), p. 90.
l"ational Fire Codes, v. 1 (Quincy, MA: National Fire Protection Association, 1986).

A. Problems 7-1. Pratt. A 150-ft3tank containing air is to be inerted to 1% oxygen concentration. Reduce the oxygen concentration to 1%using pure nitrogen as the inert gas. 1. For the pressurization technique. Loeb. "Loss Prevention. W. Flammable liquid is being pumped out of a drum into a bucket using a hand pump. Because the tank's maximum allowable working pressure is 150 psia. pp. MacKeown and V. Bulletin 256 (Washington. 11. 7-2. 99-102. Wouk. J. 1959). Understanding and Controlling Static Electricity (Goethestrasse. For the pressurization technique. H." Industrial Engineering Chemistry (June 1942). Boicourt. What bonding and grounding procedures must be followed to transfer a drum of flammable solvent into a storage tank? 7-3. Circular C-438 (Washington. pp. Recommended Practice on Static Electricity. 1989). Static Electricity. pp. 7. Describe an appropriate grounding and bonding procedure. Using the sweep-through purging method.1959). DC: National Bureau of Standards. What volume of nitrogen is required? Assume nitrogen with no oxygen and a temperature of 77°F. "Electrical Charges Produced by Flowing Gasoline. F. 1958). b. 7-4. 659-664. Electrostatic Ignitions of Fires and Explosions (Marietta. "Tame Static Electricity. Determine the volume of nitrogen required for each technique. 1985)." Chemical Engineering Progress (November 1994). 7-5. DC: US Department of Labor. D. Describe a system and a procedure for transferring ethylene oxide from a tank car through a pumping system to a storage tank. Klinkenberg and J. Germany: Expert Verlag. 101-106. Develop a list of steps needed to convert a common kitchen into an XP area. A. 573-576. it is possible to use either the sweep-through or a pressurization technique. Van der Mine. determine the number of cycles required if the pressure purge includes increasing the pressure to 140 psia with nitrogen and then venting to 0 psig. 1963). a. Static Electricity. 29. 1997). The temperature is 80°F. S. L. G.B.B. and G. pp. 75 -81. Use a vacuum purging technique to purge oxygen from a 150-ft3tank containing air." Chemical Engineering Progress (1977). 7-7. Saletan. inert a 100-galvessel containing 100% air until the oxygen concentration is 1%. Electrostatics in the Petroleum Industry (New York: Elsevier.348 Chapter 7 Designs to Prevent Fires and Explosions T. 7-6. Include both inerting and purging as well as bonding and grounding procedures. pp. What Went Wrong? (Houston: Gulf Publishing. B. and (Jun.S. 1942).L. Pure nitrogen is available for the job. "Static Electricity Hazards. F. Louvar. Silsbee. Maurer. Kletz. Ethylene oxide is a flammable liquid having a normal boiling temperature below room temperature." Chemical Engineering (Jun. "The Basic Mechanisms of Static Electrification. with the tank being returned to atmospheric pressure at the end of each cycle. Glor. v. GA: Burgoyne. T. NFPA 77-1972(Boston: National Fire Protection Association.1945). Luttgens and M. 1972).I. The tempera- . multiple pressurization cycles might be necessary." Science (Dec.

Repeat Problem 7-7 using a combined vacuum and pressure purge. and the energy stored in the nozzle. b. b. Explain the potential hazard for cases a and b from the following table: Case a Case b Hose length (ft) Hose diameter (in) Flow rate (gpm) Liquid conductivity (mholcm) Dielectric constant Density (g/cm') 20 2 25 20 2 25 10-l8 2. 7-16. to determine the hose diameter required to eliminate the potential hazard resulting from static buildup. 7-12. Determine the number of vacuum purges required to reduce a vessel's oxygen concentration from 21 % to 1% if the nitrogen contains: a. What is the most effective way to reduce the hazard of this situation? 7-17. Assume that the vacuum purge goes from atmospheric pressure to 20 mm Hg absolute.Assume that the room is purged with air at a rate of 6 room volumes per hour. 7-11. Design an inerting system for a pressure vessel to maintain the inert atmosphere at 40 psig. Be sure to account for filling and emptying of the vessel. . Use the system described in Problem 7-13. Vacuum charging of material from a 55-gal drum. 7-13. regulators. Estimate the charge buildup and accumulated energy as a result of pneumatically conveying a dry powder through a Teflon duct. Draining the tank contents.000 ft3. Repeat Example 7-2 with a 40. pipes. How long will it take to complete this purge process? 7-10. Provide precise details on the location of valves. Indicate the precise location of valves. 0 ppm of oxygen. Determine the number of purge cycles required and the total moles of nitrogen used. Include the following design features: a. Assume that the vessel height is equal to the diameter.8 0. Review Problem 7-13. The powder is collected in an insulated vessel.4 19 0. Design a generalized pressure vessel storage tank for a flammable material. 7-15. Use the system $$$ described in Figure 7-14 to determine the voltage developed between the $$$ charging nozzle and the grounded tank. part b. 7-9.8 7-14. regulators. 9000 ppm of oxygen. and process lines. c. Vacuum and pressure purging. Use a vacuum of 20 mm Hg absolute and a pressure of 200 psig. etc. Use the sweep-through purging technique to reduce the concentration of toluene from an initial 20% to 1% in a room with a volume of 25. 7-8. Assume that your vacuum system goes down to 20 mm Hg absolute.Problems 349 ture is 80°F.000-gal storage vessel. part b.

Assume an average wind speed of 5 mph and stability class D. Some large storage vessels have a floating head.0. Determine the fire water requirements (gpm. describe what safety features should be added to this process area. number of sprinkler heads. See chapter 4 for the appropriate source model and chapter 3 for material balance models. Discuss ways to improve the safety of this situation. and (c) a full vessel with an overflow line. 7-24.5-in orifice and that the nozzle pressure is 75 psig. Repeat Problem 7-27 assuming that the nozzle pressure is 100 psig and that the rate is 58 gpm. Assume that the sprinkler nozzles have a 0. What are the reasons for this design? 7-21.5-in nozzles with a nozzle pressure of 35 psig. compute the accumulated charge and energy after 5 hr and after 20 hr. and Divisions 1and 2 motors? 7-23. The vessel must be inerted with pure nitrogen before storage of the ace- . number of sprinklers. 7-31.5-in orifice) are at 75 psig. (b) a full vessel. 7-18. if the inlet flow is stopped. Determine the water requirement (gprn) and number of nozzles for a deluge system required to protect a 10. 7-29. Toluene leaks as large as 200 gprn have been recorded. Repeat the calculation for (a) a half full vessel. and the rate is 50 gpm. and assume that the vessel contains a reactive solvent. Determine the sprinkler requirements for a chemical process area 150 ft by 150 ft. 7-27. The temperature is 80°F.Assume that the sprinkler nozzles have a 0. 7-30. Using the results of Problem 7-25.000-galvessel being filled with a fluid at a rate of 200 gprn and having a streaming current of 2 X lop6amp.350 Chapter 7 Designs to Prevent Fires and Explosions Repeat the calculation for a transport rate of 50 Iblmin and 100Iblmin for transport times of 1hr and 5 hr. determine the concentration of propane in the area as a function of time if at t = 0 a 314-in propane line breaks (the propane main header is at 100 psig). part c. a flat cover that floats on the liquid surface. Assume that the friction loss from the last sprinkler head to the pump is 50 psi and that the nozzles (0. As the liquid volume increases and decreases. 7-19. Use 0. 7-28. What electrical classification would be specified for an area that has Classes I and 11. 7-25. and pump horsepower) to protect an inside process area of 200 ft2. Acetone (C3H60)is to be stored in a cylindrical process vessel with a diameter of 5 ft and a height of 8 ft. 7-20. For Problem 7-18. 7-22. Compute the accumulated charge and energy for a 100. the nozzle pressure is 75 psig. Determine the number of sprinkler heads and the pump specifications for this system (HP and gprn). Determine the recommended ventilation rate for an inside process area (30. Determine the fire water requirements (gpm. Determine the recommended distance between a process area with toluene and an area with an open flame. Discuss the consequences of these results. the floating head rises and falls within the cylindrical shell of the vessel.000 ft3) that will handle Class I liquids and gases. Groups A and E. 7-26.000-galstorage tank that has a diameter of 15 ft. and pump horsepower) to protect an inside process area of 2000ft2. Make the calculation for a fluid having a conductivity of lo-'' mholcm and a dielectric constant of 2.5-in orifice. For the process area described in Problem 7-24.

Problems 351 tone. a. A vacuum is available at 30 mm Hg absolute pressure. c. Determine the number of cycles required for your selected procedure. d. Determine the total amount of nitrogen used. is the best procedure. The ambient temperature is 80°F. Decide whether a pressure or vacuum purge. b. Determine the target oxygen concentration for the inerting procedure. or a combination of both. . The final pressure in the tank after the inerting procedure is atmospheric. A limited supply of pure nitrogen is available at 80 psig and 80°F.

.

Introduction to Reliefs espite many safety precautions within chemical plants. including physical properties of the ejected material. D . If pressures rise too high. The motivation is to determine the material mass flow rate through the relief and the physical state of the material (liquid. Inherent safety. causing major releases of toxic or flammable chemicals. Second. an error in any step of this procedure can result in catastrophic failures. the appropriate relief device type must be selected. The relief system is composed of the relief device and the associated downstream process equipment to safely handle the material ejected. The type depends mostly on the nature of the material relieved and the relief characteristics required. A major effort is always directed toward controlling the process within safe operating regions. The defense against this type of accident is to prevent the accident in the first place. data are collected on the relief process. the worst-case scenario is selected and the final relief design is achieved. Every step in this method is critical to the development of a safe design. Next. scenarios are developed that describe the various ways in which a relief can occur. The method used for the safe installation of pressure relief devices is illustrated in Figure 8-1. The third line of defense against excessive pressures is to install relief systems to relieve liquids or gases before excessive pressures are developed. Third. Finally. is the first line of defense. Dangerous high-pressure excursions must be prevented or minimized. and the relief is sized. described in chapter 1. This can result in rupturing of process equipment. Definitive guidelines are available. The second line of defense is better process control. they may exceed the maximum strength of pipelines and vessels. The first step in the procedure is to specify where relief devices must be installed. vapor. or two phases). equipment failures or operator errors can cause increases in process pressures beyond safe levels.

'Marx Isaacs." Chemical Engineering (Feb.354 Chapter 8 Introduction to Reliefs I I L o c a t e Re1 i e f s I I u Choose e l i e f Types 0 Develop Re l i e f 5d Scenar ios Acquire Size Reliefs Size Rel iefs Sin l e Phase Two P h a s e s ( I Choose Worst Case Scenarios 1 I "Data may be simple physical properties or sophisticated calorimeter data to characterize two phase flow (including gassy systems and/or high viscosity laminar flow systems). 22. 8-1 Relief Concepts Pressure relief systems are required for the following reasons: to protect personnel from the dangers of overpressurizing equipment. In this chapter we introduce relief fundamentals and the steps in the relief design procedure. to prevent damage to equipment. pp. 113-124. "Pressure Relief Systems. .1971). Relief sizing methods are covered in chapter 9. D e s i gn Rel i e f System Figure 8-1 Relief method. to minimize chemical losses during pressure upsets.

(B) relieving froth (two-phase flow). After the reactants are consumed. Typical pressure versus time curves for runaway reactions are illustrated in Figure 8-2. Curve C assumes that the reactor is capable of withstanding the full pressure of the runaway reaction. leading to an increase in heat production. . with temperature and pressure increases of several hundred degrees per minute and several hundred psi per minute.8-1 Relief Concepts 355 i \ reactor (C) presrre Relief opening a AII vapbr (A) Time Figure 8-2 Pressure versus time for runaway reactions: (A) relieving vapor. the heat generation stops and the reactor cools. This self-accelerating mechanism results in a runaway reaction. As the temperature rises. respectively. For the curves in Figure 8-2 the cooling is lost at t = 0. the pressure subsequently drops. the pressure and temperature continue to rise until the reactants are completely consumed. or other scenario. to prevent damage to adjoining property. failure of a valve. If cooling is lost because of a loss of cooling water supply. as shown by curve C (Figure 8-2). and (C) closed reaction vessel. then the reactor temperature will rise. to reduce insurance premiums. Reaction runaways for large commercial reactors can occur in minutes. The pressure within the reactor increases because of increased vapor pressure of the liquid components and/or gaseous decomposition products resulting from the high temperature. If the reactor has no relief system. the reaction rate increases. and to comply with governmental regulations. Assume that an exothermic reaction is occurring within a reactor.

The pressure will increase inside the reactor until the relief device activates at the pressure indicated. This is illustrated by curve A (Figure 8-2) for vapor relief only and by curve B for a two-phase froth (vapor and liquid). 1997). This is sometimes called the design pressure. Guide for Pressure-Relieving and Depressuring Systems. The pressure drops immediately when the relief device opens because only a small amount of vapor discharge is required to decrease the pressure. The pressure drops until the relief valve closes. As the operating temperature increases. DC: American Petroleum Institute. (Washington. Definitions that are commonly used within the chemical industry to describe reliefs are given in the following paragraphs. the MAWP decreases because of metal embrittlement at lower temperatures. It is expressed as a percentage of the MAWP. this pressure difference is called the blowdown. Operating pressure: The gauge pressure during normal service. the pressure continues to rise as the relief valve opens. Likewise. Accumulation: The pressure increase over the MAWP of a vessel during the relief process. Maximum allowable working pressure (MAWP): The maximum gauge pressure~ermissible at the top of a vessel for a designated temperature. as the operating temperature decreases. 2API RP 521. the pressure response depends on the relief device characteristics and the properties of the fluid discharged through the relief. although vessel deformation may occur at as low as twice the MAWP. . usually 10% below the MAWP. the nature of the relieved material must be known in order to design a proper relief. Because the relief character of two-phase vapor-liquid material is markedly different from vapor relief. 4th ed. 1-3. Set pressure: The pressure at which the relief device begins to activate. The incremental pressure increase over the initial relief pressure is called overpressure. When froth is discharged (curve B in Figure 8-2).pp. the MAWP decreases because the vessel metal loses its strength at higher temperatures. Overpressure is equivalent to the accumulation when the set pressure is at the MAWP. Curve A is for vapor or gas discharged through the relief valve. It is expressed as a percentage of the set pressure. Overpressure: The pressure increase in the vessel over the set pressure during the relieving process.356 Chapter 8 Introductionto Reliefs If the reactor has a relief device. Vessel failure typically occurs at 4 or 5 times the MAWP.

. 187-194. .. flare. . . pp. The engineer must anticipate the potential problems that may result in increased pressures. including the pipe to the relief. Maximum allowable accumulated pressure: The sum of the MAWP and the allowable accumulation.. . Figure 8-3 Backpressure: The pressure at the outlet of the relief device during the relief process resulting from pressure in the discharge system. 8-3 Location of Reliefs3 The procedure for specifying the location of reliefs requires the review of every unit operation in the process and of every process operating step... that is. . "Pressure-Relief Valves for Process Plants. . 28. The type of questions asked in this review process are What happens with loss of cooling. ... \ \ Relief begins to open Description of overpressure and accumulation. . . . or agitation? What happens if the process is contaminated or has a mischarge of a catalyst or monomer? 3Robert Kern... . knockout drum. . . . scrubber... . . . or other types of equipment that assist in the safe relief process.. It is expressed as a percentage of the set pressure. . .. . The relationship between these terms is illustrated in Figures 8-3 and 8-4. .." Chemical Engineering (Feb. . . the relief device.8-3 Location of Reliefs 357 Overpressure (% of set pressure) Accumulation (% of set MAWP) ----f--+-L-\----. discharge pipelines. at points where upset conditions create pressures that may exceed the MAWP. 1977). . . Pressure relief devices are installed at every point identified as potentially hazardous... . . Blowdown: The pressure difference between the relief set pressure and the relief reseating pressure. Relief system: The network of components around a relief device. . heating. . . . .. MAWP Set pressure 4 .

W 9 Overpressure (Typical) 100 aximum A l l o w a b l e S e t P r e s s u r e or Single Valve (Average) S t a r t t o Open - U s u a l Maximum N o r m a l Operating Pressure .-.. DC: American Petroleum Institute.358 Chapter 8 Introduction to Reliefs Pressure Vessel Requirements Vessel Pressure Typical Characteristics o f Safety Rel i e f Valves Maximum R e l i e v i n g Pressure for F i r e Maximum A l l o u a b l e Accumulated Pressur Maximum A l l o w a b l e Accumulated Pressure for Multiple-valve Instal lation (Other Than F i r e Exposure) ' - . Adapted from API RP 521.-. .. What happens if the operator makes an error? What is the consequence of closing valves (block valves) on vessels or in lines that are filled with liquids and exposed to heat or refrigeration? What happens if a line fails..-..u L Y m D -n - - Max imum Re l i e v i n g Pressure for Process Multiple Valves Single Valves Overpressure (Maximum) Maximum A l l o u a b l e Working Pressure or Design Pressure (Hydrotest a t 150) - ._ c D o 95 11 Blowdown ( T y p i c a l ) I - - L eseat Pressure ( T y p i c a l ) for Single Valve -a 90 r Standard Leak Test Pressure Figure 8-4 Guidelines for relief pressures. 4th ed. for example.. 1997).4 -r .. (Washington. 30.* .A D rn 2 116 . Guide for PressureRelieving and Depressuring Systems. p.x 0 . a failure of a high-pressure gas line into a lowpressure vessel? What happens if the unit operation is engulfed in a fire? .

Reactor (R-1): A relief is installed on this reactor because. Drum (D-1): Again. Example 8-1 Specify the location of reliefs in the simple polymerization reactor system illustrated in Figure 8-5. and how are relief systems designed to handle the discharge as a result of runaway reactions? Some guidelines for locating reliefs are summarized in Table 8-1. Reliefs are installed in jackets to prevent excessive steam pressures due to operator error or regulator failure. This hazard is eliminated by adding PSV-3. Add PSV-5 to this coil. V-6. and turbines need reliefs on the discharge side.(3) adding monomer for a period of 3 hr. for example). Solution The review method for specifying the location of reliefs follows. The reason for the two relief devices PSV-1A and PSV-1B is described in the next section. and damaged if they are dead-headed without a pressure-relieving device (PSV-2). V-5. 113-124. Storage vessels need pressure and vacuum reliefs to protect against pumping in or out of a blocked-in vessel or against the generation of a vacuum by condensation. d. Reactor coil: This reactor coil can be pressure-ruptured when water is blocked in (V-4. 'Marx Isaacs. and V-7 are closed) and the coil is heated with steam or even the sun. Positive displacement pumps. pp. 22. including reactors. "Pressure-Relief Systems. towers. Positive displacement pump (P-1): Positive displacement pumps are overloaded. Refer to Figures 8-5 and 8-6 and Table 8-1 for relief locations. Vessel steam jackets are often rated for low-pressure steam. What conditions cause runaway reactions. storage tanks. all process vessels need relief valves. b. This relief is labeled PSV-1 for pressure safety valve 1. Blocked-in sections of cool liquid-filled lines that are exposed to heat (such as the sun) or refrigeration need reliefs. c. e." Chemical Engineering (Feb. The major steps in this polymerization process include (1) pumping 100 Ib of initiator into reactor R-1. Because the reaction is exothermic. PSV-4. a. every process vessel needs a relief. and (4) stripping the residual monomer by means of a vacuum using valve V-15. in general. Heat exchanger (E-1): Heat exchanger tubes can rupture from excessive pressures when water is blocked in (V-10 and V-11 are closed) and the exchanger is heated (by steam.8-3 Location of Reliefs 359 Table 8-1 - Guidelines for Specifying Relief Positions1 All vessels need reliefs. cooling during monomer addition with cooling water is necessary. This completes the specification of the relief locations for this relatively simple process. overheated.1971). . (2) heating to the reaction temperature of 24OoF. This type of relief discharge is usually recycled back to the feed vessel. and drums. compressors.

In engineering . liquids and gases.1 Spec ifi c a t ions Nitrogen 1 inch 0.-Il Monomer In ool ing W a t e r In Cooling Water Out (100 p s i g ) P. depending on the specific application. Example 8-1 illustrates the engineering rationale for installing relief valves at various locations within a chemical plant. solids. the type of relief is chosen. condenser. such as for liquids. they may be vented to the atmosphere or vented to containment systems (scrubber.360 Chapter 8 Introduction to Reliefs C o o l ing W a t e r O u t (10 g p m ) $. 8-4 Relief Types Specific types of relief devices are chosen for specific applications. After the relief locations are specified. and the like). incinerator. flare.5 i n c h Figure 8-5 Polymerization reactor without safety reliefs. and corrosive materials. gases.

process conditions. as illustrated in Figure 8-7. For the balanced-bellows design the bellows on the backside of the valve seat ensures that the pressure on that side of the seat is always atmospheric. To avoid the possibility of an unauthorized person changing this setting. For a conventional spring-operated relief. therefore. the flow through the balanced-bellows relief . There are two general categories of relief devices (spring-operated and rupture discs) and two major types of spring-operated valves (conventional and balanced-bellows). the valve opens based on the pressure drop across the valve seat. the set pressure is proportional to the pressure drop across the seat. that is. terms the type of relief device is specified on the basis of the details of the relief system. the set pressure will increase and the valve may not open at the correct pressure. the adjustable screw is covered with a threaded cap.8-4 Relief Types 361 Cooling Water Out (10 gpm) Monomer ( 5 gpm) I n o o l i n g Water I n Steam (125 psig) C o o l i n g Water Cool ing Hate ( 2 0 0 gpm) Out Steam Trap P. In addition. Thus.1 Figure 8-6 Polymerization reactor with safety reliefs. if the backpressure downstream of the valve increases. Thus the balanced-bellows valve will always open at the desired set pressure. The relief set pressure is usually specified at 10% above the normal operating pressure. is reduced as the backpressure increases. The flow through the relief. On spring-operated valves the adjustable spring tension offsets the inlet pressure. However. and physical properties of the relieved fluid. the flow through the conventional relief is proportional to the difference in pressure across the seat.

they remain open. Flexing could lead to premature failure at pressures below the set pressure. rupture discs are . It may also allow air to enter the process. Another problem with rupture disc systems is that once they open. They are used alone. in series. including exotic corrosion-resistant materials. or in parallel to spring-loaded relief devices. They can be made from a variety of materials. They usually consist of a calibrated sheet of metal designed to rupture at a well-specified pressure. Convent i ona l Spr i ng Operated Valve Rupture Disc and Assernbl y Bet low Balanced Bellows Figure 8-7 Major types of relief devices. Therefore the flow is reduced as the backpressure increases. Rupture discs are specially designed to rupture at a specified relief set pressure.Chapter 8 Introduction to Reliefs S P ~ Seat. For this reason some rupture disc systems are designed to operate at pressures well below the set pressure. In some accidents discs were ruptured without the process operator being aware of the situation. vacuum service may cause rupture disc failure if the relief system is not specifically designed for this service. leading to a possible fire and/or explosion. is proportional to the difference in pressure between the inlet and the outlet of the valve. This may lead to the complete discharge of process material. An important problem with rupture discs is the flexing of the metal as process pressures change. In addition. To prevent this problem.

The relief valve (liquid only) begins to open at the set pressure. 3. Rupture discs are available in much larger sizes than spring-operated relief valves. with commercial sizes available up to several feet in diameter. When rupture discs are used before a spring-loaded relief. This valve reaches full capacity when the pressure reaches 25% overpressure. The safety valve is for gas service. This telltale gauge is an indicator that shows when the disc ruptures. In either case the telltale gauge indicates that the disc needs to be replaced. This is accomplished by using a discharge nozzle that directs high-velocity material toward the valve seat. (4) to protect the relatively complex parts of a springloaded device from reactive monomers that could cause plugging. Rupture discs typically cost less than equivalently sized spring-operated relief valves. 2. PSV-la is a rupture disc to protect PSV-lb from the reactive monomers (plugging from polymerization). There are three subcategory types of spring-loaded pressure reliefs: 1. In all these examples the problems are eliminated if the rupture disc and system are specified and designed appropriately for the specific operating conditions of the process.8-4 Relief Types 363 available with embedded wires that are cut when the disc ruptures. Safety valves pop open when the pressure exceeds the set pressure. pieces of the disc may become dislodged. Example 8-2 Specify the types of relief devices needed for the polymerization reactor in Example 8-1 (see Figure 8-6). the valve reseats at approximately 4% below the set pressure. (2) to give absolute isolation when handling extremely toxic chemicals (spring-loaded reliefs may weep). After blowdown of the excess pressure. a pressure gauge is installed between the two devices. The valve closes as the pressure returns to the set pressure. the valve has a 4% blowdown. . Safety relief valves function as relief valves for liquids and as safety valves for gases. and (5) to relieve slurries that may plug spring-loaded devices. creating potential downstream plugging problems. when rupture discs rupture. (3) to give absolute isolation when handling flammable gases. Rupture discs are frequently installed in series to a spring-loaded relief (1)to protect an expensive spring-loaded device from a corrosive environment. The failure can be the result of a pressure excursion or of a pinhole caused by corrosion. this can activate an alarm in the control room to alert the operator. Solution Each relief is reviewed in relation to the relief system and the properties of the relieved fluids: a. Also. The safety relief valve is used for liquid and gas service. The relief valve is primarily for liquid service. Recent advances in rupture disc design have minimized this problem.

A conventional relief device is satisfactory in this service.- 364 Chapter 8 Introduction to Reliefs b. The scenarios developed for the reactor system described in Figure 8-6 are summarized in Table 8-2. This relief provides protection for the following scenario: The liquid is blocked in by closing all valves. The same pump relief is in a line with a nitrogen regulator. the relief scenarios are developed. c. A pump is dead-headed. After specifying the location and type of all relief devices. the relief is sized to handle steam injected into the exchanger under uncontrolled conditions. PSV-2 is a relief valve because this relief is in a liquid service line. the relief is sized to handle the nitrogen if the regulator fails. PSV-5 is a relief valve for liquid service only. a balanced bellows is specified. the pump relief is sized to handle the full pump capacity at its rated pressure. a steam regulator failure. for example. For each specific relief all possible scenarios are identified and cataloged. PSV-lb is a safety relief valve because a runaway reaction will give two-phase flow. d. The relief vent area is subsequently computed for each event (scenario) and the worst-case scenario is the event requiring the largest relief vent area. PSV-4 is a safety relief valve because liquid or vapor service is possible. PSV-3 is a relief valve because it is for liquid only. the heat of reaction increases the temperature of the surrounding reactor fluid. In Table 8-2 only three reliefs have mul- . and pressures are increased inside the coil because of thermal expansion. The worst cases are a subset of the overall developed scenarios for each relief. The worst-case scenarios are identified later by means of the computed maximum relief area for each scenario and relief (see chapter 9). e. 8-5 Relief Scenarios A relief scenario is a description of one specific relief event. 2. This is a list of scenarios for one specific relief. Usually each relief has more than one relief event. Examples of relief events are: 1. Because this vent will go to a scrubber with possibly large backpressures. A conventional valve is satisfactory. both liquid and vapor. 3. and the worst-case scenario is the scenario or event that requires the largest relief vent area. The same pump is connected to a heat exchanger with live steam. f. This step of the relief method is extremely important: The identification of the actual worst-case scenario frequently has a more significant effect on the relief size than the accuracy of relief sizing calculations.

V-1 is accidentally closed. (a) Nitrogen regulator fails. 8-6 Data for Sizing Reliefs Physical property data and sometimes reaction rate characteristics are required for making relief sizing calculations. (b) Note: The other R-1 scenarios will be relieved via PSV-1. the entire contents of the bottle may be "relieved. and heat of reaction causes thermal expansion. . therefore they are the worst-case scenarios." This result has also been verified for runaway reactions in the chemical industry. In the relief design." Chemical Engineering Progress (August 1985).8-6 Data for Sizing Reliefs 365 Table 8-2 Relief Scenarios for Example 8-2 (See Figure 8-6) Scenarios (a) Vessel full of liquid and pump P-1 is accidentally actuated. Data estimated using engineering assumptions are almost always acceptable when designing unit operations because the only result is poorer yields or poorer quality. These data are acquired with the apparatus already described in section 6-13. As a result of this complexity. 33-36. Water blocked inside coil. however. especially when conditions change rapidly. system needs relief for 100 gpm at 50 psig. giving critical flow through 0.5-in line. Fisher. The other three reliefs have only single scenarios. (d) Loss of cooling during reaction (runaway). A runaway reaction is another scenario that requires special data. Confined water line is heated with 125-psig steam. giving critical flow through 1-in line. When designing reliefs for gas or dust explosions. special deflagration data for the scenario conditions are required. special methods have 4Harold G. Relief identifications PSV-la and PSV-lb PSV-2 PSV-3 PSV-4 PSV-5 tiple scenarios that require the comparative calculations to establish the worst cases. It is known that runaway reactions nearly always result in two-phase flow reliefs4 The two phases discharge through the relief system similar to a champagne and carbon dioxide mixture exiting a freshly opened bottle. (c) Nitrogen regulator fails. these types of assumptions are not acceptable because an error may result in catastrophic and hazardous failures. "DIERS Research Program on Emergency Relief Systems. If the champagne is heated before opening. as in a runaway reaction scenario.pp. (b) Cooling coil is broken and water enters at 200 gpm and 50 psig. Two-phase flow calculations are relatively complex.

" Thermochimica Acta (1980). In the second heating mode the sample is heated at a fixed temperature rate and the calorimeter watches for a higher rate that identifies the initiation of the exothermic reaction. data acquisition hardware. Muller. In the first mode the sample is heated to a fixed incremental temperature. Forrest. I. 37: 1-30. 6H. A series of controlled 5D. R. Townsend and J. Each calorimeter has a different sample size. A. J. S. These include the accelerating rate calorimeter (ARC). E. maximum pressure rate. A. the reactive system screening tool (RSST). D. and B. G. The sample to be tested is heated by means of one of two modes.366 Chapter 8 Introduction to Reliefs Nitrogen Adiabatic Heater / \Magnetic 'Stirrer Figure 8-8 Vent sizing package (VSP) for acquiring runaway reaction data. the automatic pressure-tracking adiabatic calorimeter (APTAC). 1992). Grossel. If no reaction is detected. been developed for acquiring the relevant data5 and for making the relief vent sizing calculations6 (see chapter 9). Huff. Tilley. . The data obtained from the calorimeters include maximum self-heat rate. and the vent sizing package (VSP). The VSP (Figure 8-8) is essentially an adiabatic calorimeter. All of these calorimeters work essentially the same way. and temperature and pressure as a function of time. reaction onset temperature. container design. J. Shaw. J. Tou. C. Several commercial calorimeters are available to characterize runaway reactions. Noronha. A small amount of the material to be tested (30-80 mg) is loaded into a thin-walled reactor vessel. "Thermal Hazard Evaluation by an Accelerating Rate Calorimeter. Fisher. A. Emergency Relief System Design Using DZERS Technology (New York: American Institute of Chemical Engineers. and then the calorimeter maintains this temperature and waits a fixed time to determine whether an exothermic reaction occurs. H. and data sensitivity. S. S. then the temperature is increased by another increment. Some calorimeters use a mix of the two modes.

heaters increases the sample temperature to the runaway conditions.8-6 Data for Sizing Reliefs 367 Time ( minutes ) Figure 8-9 Runaway reaction temperature data acquired using the vent sizing package (VSP). this prevents the thin-walled sample container from rupturing. at the set Figure 8-10 Runaway reaction pressure data acquired using the vent sizing package. the VSP device tracks the pressure inside the can and maintains a similar pressure in the main containment vessel. The data acquired with this calorimeter are shown in Figures 8-9 and 8-10. Results of particular importance for relief sizing calculations include the temperature rate (dTldt). . During the runaway reaction.

8-7 Relief Systems After the relief type has been chosen and the relief size computed. and the cooperative efforts of the plant. it is essential to understand this technology. hopefully they will never need to operate. Pressure-relieving systems are unique compared with other systems within a chemical plant. Specific references have already been cited. This is particularly important because local government standards may vary. and the National Board of Fire Underwriters.During field construction. time. and insurance requirements. The engineer must take the responsibility for adhering to standard practices. Because the calorimeter starts with known weights and known compositions. This evolution requires creativity. Relief Design Considerations A designer of relief systems must be familiar with governmental codes. and process engineers. however. industrial standards. It is recommended that relief designers carefully consider all codes and. the American Petroleum Institute. hard work. including deciding how to install the relief in the system and how to dispose of the exiting liquids and vapors.- - - - - - - - - - 368 Chapter 8 Introduction to Reliefs pressure and the temperature increase AT corresponding to the overpressure 4P. a poor installation can result in completely unsatisfactory relief performance. they must do so flawlessly. Some installation guidelines are illustrated in Figure 8-11. The objective of this section is to give students and design engineers the details necessary for understanding relief systems. and tested. To develop the necessary optimum and reliable relief systems. the engineer takes the responsibility for completing the design of the relief system. where feasible. Relief Installation Practices Regardless of how carefully the relief is sized. Other systems. in this case the relief system development must be optimally designed and demonstrated within a research environment before the plant start-up. This same effort and creativity is essential when developing relief systems. usually evolve to their optimum performance and reliability. sometimes expediency or construction convenience leads to modifications and deviations from acceptable practice. Codes of particular interest are published by the American Society of Mechanical Engineers. design. . select the one that is most suited to the particular installation. such as extraction and distillation systems. but when they do. specified. the heat of reaction can also be determined from the Tversus t data (assuming that the heat capacities of the monomers and products are known). especially when installing relief systems. practical knowledge.

Disc Protects Against Corrosion. pp. M a t e r i a l s where Highly Toxic S p r i n g L o a d e d V a l v e M a y Weep. Adapted from Eric Jennett. I s Good f o r T h i s Type Design Polymerization Reactors Where P e r i o d i c C l e a n i n g I s N e c e s s a r y . Figure 8-11 Relief installation practices. T w o Rupture Discs with Special Valve W h i c h K e e p s One V a l v e A l w a y s D i r e c t l y Connected t o Vessel. The Rupture Toxicity or The S p r i n g Loaded R e l i e f C l o s e s and M i n i m i z e s Losses. . "Components of PressureRelieving Systems. for Back-up L a r g e r Re l i e f s . N o r m a l R e l i e f M a y Go T h r o u g h S p r i n g and R u p t u r e D i s c I s Loaded Device.8-7 Relief Systems 369 System Recommendations A Vesse l Rupture Disc O r for i n Corrosive Service. 1963)." Chemical Engineering (Aug. i n Extremely The 1 s t May P e r i o d i c a l l y Need t o be R e p l a c e d . u 9 23 T w o Rupture Discs Corrosive Service. R u p t u r e Disc and S p r i n g Loaded R e l i e f . 151-1 58. @ Two R e l i e f s i n S e r i e s . 19.

L Pipe 8 O r i f i c e Area o f a S i n g l e Safety Relief i n Vapor Service.370 Chapter 8 Introduction t o Reliefs Ti T Jg System A. Recommendations P r e s s u r e D r o p N o t M o r e T h a n 3% o f Set Pressure L o n g R a d i u s E l bow B ' I f D i s t a n c e i s G r e a t e r T h a n 10 Weight and R e a c t i o n F o r c e s Feet. Device Causing Turbulence Regulator or Valve: 25 i n Same P l a n e : 2 0 2 E l l s o r Bends N o t 2 E l I s o r B e n d s i n Same P I a n e : 1 E l l o r Bend: P u l s a t i o n Damper: 15 18 16 Figure 8-11 (continued) . -A 1 t 4. Vessel S h o u l d be S u p p o r t e d Below t h e Long Radius Elbow. A. Turbulence Causing Device D i m e n s i o n ( B ) shown below: M i n i m u m Number o f S t r a i g h t Pipe Diameters B. M u l t i p l e Valves with Staggered S e t t i n g s May b e R e q u i r e d . A. C. Should N o t E x c e e d 2% o f t h e C r o s s Sectional Area o f t h e Protected Line. Process L i n e s Should Not be Connected t o S a f e t y Valve I n l e t Piping. B.

Sizing. or a combination of them. and Installation of Pressure-Relieving Devices in Refineries. one is illustrated in Figure 8-12. Blowdown Drum Figure 8-12 Relief containment system with blowdown drum. . incinerator. this horizontal knockout drum system serves as a vapor-liquid separator as well as 'API 520. For this reason reliefs are now rarely vented to the atmosphere. normal stress analysis is the recommended method. Selection. DC: American Petroleum Institute. here the liquid is collected and the vapor is discharged to another treatment unit. As illustrated in Figure 8-12. The API RP 5207has some guidelines. flare. Another important consideration is the reaction forces generated when the relieved materials flow through the relief system at high speed. It is also important to recognize that company philosophy and the regulatory authorities have a significant influence on the design of the final disposal system. Total containment systems are commonly used. This type of system is called a total containment system. however. (Washington. Horizontal Knockout Drum Knockout drums are sometimes called catch tanks or blowdown drums. primarily from the standpoint of pollution. and they are becoming an industrial standard.8-7 Relief Systems 371 To Condenser. scrubber. In most cases a relief is first discharged to a knockout system to separate the liquid from the vapor. it may include a condenser. pt. 2. This subsequent vapor treatment unit depends on the hazards of the vapor. The blowdown drum separates the vapor from the liquid. 1994). 4th ed. Installation.

pp. (Washington. When space within a plant is limited. . The dropout velocity of a particle in a stream is 8S. as shown in Figure 8-13. 63-67." Plant/Operations Progress (July 1986). S. The two-phase mixture usually enters at one end. and the vapor leaves at the opposite end. 1997). "Design and Sizing of Knockout DrumsICatchtanks for Reactor Emergency Relief Systems. Grossel. a holdup vessel for the disengaged liquid. The design method for sizing this type of system was published by Grosse18 and in API 521. a tangential knockout drum is used. 4th ed. 9API RP 521.372 Chapter 8 Introduction to Reliefs Vent o r or t o Atmosphere F l a r e Stack Scrubber Long Knock-Out Radius Rupture L o c a t ed D i s c as s C lose t0 Vesse Poss l a i b le Figure 8-13 Tangential inlet knockout drum with separate liquid catch tank. DC: American Petroleum Institute. Guide for Pressure-Relievingatzd Depressurizing Systems. with a vapor exit in the center to minimize vapor velocities. Inlets may be provided at each end.9The method is based on the maximum allowable velocity for minimizing liquid entrainment.

and Cis the drag coefficient given by Figure 8-14. DC: American Petroleum Institute. is the vapor density. g is the acceleration due to gravity. where u. Guide for Pressure-Relieving and Depressurizing Systems. The abscissa of Figure 8-14 is where pv is the vapor viscosity in centipoise and C(Re)2 is unitless. p. Data from API RP 521. d. 1982). . is the dropout velocity.8-7 Relief Systems 373 10 . 2d ed. is the liquid density. 100 1 000 lb' 10" 1 b' C ( ~ e ) ~ Figure 8-14 Drag coefficient correlation. (Washington. p. is the particle diameter.

2 lbmlft3)(9. is subsequently computed using the velocity and the volumetric flow rate of the vapor.84 x ft.01 centipoise. pv = 0. = 300 pm = 9. The dropout velocity is determined by using Equation 8-1: The required vapor space area.where Vapor rate = 170 lb/hr. .01 centipoise)' - 0. The graph abscissa is computed using Equation 8-2: X (0.20 lb/ft3. The entire vessel design is determined as a function of this vapor area plus the liquid hold volume. the drag coefficient is first determined. using Figure 8-14.2) Ibm/ft3 Using Figure 8-14. pv = 0.374 Chapter 8 Introduction to Reliefs - -- - Example 8-3 Determine the maximum vapor velocity in a horizontal knockout drum to dropout liquid particles with particle diameters of 300 pm. perpendicular to the vapor path.3. we find that C = 1. Solution To determine the dropout velocity.84X ft)"30 (0. and the general geometric configuration of the vessel. and d. p~ = 30 Iblft3.

. 0 4 8 m . and a heating value of (in 20.8-7 Relief Systems 375 Flares l o Flares are sometimes used after knockout drums. the emissivity E . The diameter of the flare must be suitable to maintain a stable flame and to prevent a blowout (when vapor velocities are greater than 20% of the sonic velocity).000 Btullb. and the distance R from the flame: Assuming a flame height of 120df. The height of a flare is fixed on the basis of the heat generated and the resulting potential damage to equipment and humans. Tan.000 Iblhr. l0Soen H. The flare diameter is 4 ft. The usual design criterion is that the heat intensity at the base of the stack is not to exceed 1500 Btulhrlft2." Hydrocarbon Processing (January 1967). the flare load is 970. an emissivity E = 0 . (in lblhr): Example 8-4 Determine the stack height required to give a heat intensity of 1500Btu/hr/ft2at a distance of 410 ft from the base of the flare. The effects of thermal radiation are shown in the following table: Heat intensity (Btu/hr/ft2) 2000 5300 3000-4000 350 Effect Blisters in 20 s Blisters in 5 s Vegetation and wood are ignited Solar radiation Using the fundamentals of radiation. we know that the heat intensity q at a specific point is a function of the heat generated by the flame Q. "Flare System Design Simplified. The objective of a flare is to burn the combustible or toxic gas to produce combustion products that are neither toxic nor combustible. . Equation (8-3) can be algebraically modified to give the flare height Hi ft) as a function of the flare stack diameter df (in ft) and the desired heat intensity qf (in ~ t u / h r / f t ' ) at a distance Xf from the base of the flare (in ft) for a burning fuel with a molecular weight M and a vapor rate Q. and the molecular weight of the vapor is 44.

must first go to a knockout system. Scrubbers The fluid from reliefs. Mass Transfer Operations.Q. This alternative is particularly attractive if the vapors have a relatively high boiling point and if the recovered condensate is valuable. Louvar. Details of scrubber designs are covered by Treybal. 12D. or venturi-type systems. Emergency Relief System Design (New York: American Institute of Chemical Engineers. Liquids are subsequently collected and the vapors may or may not be vented. a flare (described previously) or a scrubber system may be required. "Emergency Relief System (ERS) Design: An Integrated Approach Using DIERS Methodology. ed. F. If the vapors are toxic. 1950). (New York: McGraw-Hill. sometimes two-phase flow. The design of condenser systems is covered by Kern. G. Boicourt. pp. 2000). If the vapors are nontoxic and nonflammable. S. Kern. where the liquids and vapors are separated. Scrubber systems can be packed columns. S. they may be vented unless some regulation prohibits this type of discharge. R. 1958). Richard Greene. Anderson." in Safe and EfJicient Plant Operations and Maintenance.ll Condensers A simple condenser is another possible alternative for treating exiting vapors." Process Safety Progress (April 1995). p. Grossel and J. "Pressure Relieving Devices. 207.376 Chapter 8 Introduction to Reliefs Solution The flare height is computed using Equation 8-4. 3d ed. Design for Overpressure and Underpressure Protection (New York: American Institute of Chemical Engineers. Treybal. . Darby. 1997). plate columns. Process Heat Transfer (New York: McGraw-Hill. This alternative should always be evaluated because it is simple and usually less expensive and because it minimizes the volume of material that may need additional post-treatment. "R. 1980). The units are consistent with those required: = 226 ft. (New York: McGraw-Hill.12 Suggested Reading General Articles on Relief Valves and Systems Floyd E. W. 93-106.E.

" Thermochimica Acta (1986). p.Problems 377 Marx Isaacs. "Pressure-Relief Valves for Process Plants. 8-5. "Simplified Vent Sizing Equations for Emergency Relief Requirements in Reactors and Storage Vessels. Review Figure 8-15 and Problem 8-9 to determine what types of relief devices should be used at each location. Review Figure 8-16 and Problems 8-10 and 8-12. 104: 13-29." AICHE Journal (Oct. 8-9. "Thermal Runaway Reactions in a Low Thermal Inertia Apparatus. 1977). "Pressure-Relief Systems. how would this affect the size (qualitatively) of this relief system? 8-8. 8-4. 113. 8-10. 22. 8-13. K. qualitatively describe the algorithm you would use to compute the relief size for this system. Fisher. In the future it is anticipated that insurance rates will be set as a function of the safety of a plant. determine the relief scenarios for each relief device. Using the results of Problems 8-10 and 8-12. Review Figure 8-16 and Problem 8-10 to determine what types of relief devices should be used at each location. 8-3. Give four examples of situations requiring a combination of spring-operated reliefs in series with rupture discs. and make recommendations for total containment systems. Fauske. J." Chemical Engineering (Feb. If a scrubber is installed after PSV-lb and it has a pressure drop of 30 psig. C. 8-15. Referring to Problem 8-7. Review Figure 8-16. . and make recommendations for total containment systems. 8-16. If the fluid being handled is extremely volatile and flammable. 1622-1634. Robert Kern. Leung. determine the relief scenarios for each relief device. 1986). Review Figure 8-15. Illustrate the kinds of plant statistics that you would cite to reduce your insurance costs. 8-14. Problems 8-1. Using the results of Problems 8-9 and 8-11. 8-11. Develop an example to illustrate the potential magnitude of the vacuum. what control features (safeguards) would you add to this reactor system? 8-7. what design modifications would you make to this relief system? 8-6. and determine the locations for relief devices. 8-12.1971). C. Review Figure 8-15 and Problems 8-9 and 8-11. p. pp. One defense against runaway reactions is better process control. 28. J. Leung. Using the system illustrated in Figure 8-6. PSV-2 of Figure 8-6 is a relief to protect the positive displacement pump P-1. H. and determine the locations for relief devices. Can gate valves be placed between a vessel relief and its vessel? 8-2. and H. G. 187. Describe the process of creating a vacuum in a storage vessel as a result of condensation." Chemical Engineering (Feb.

.

.

0 0. See Grossel and Louvar (2000). 8-21. . for the following three cases: Case a - Case b 70. Selection. 8-24. it is important to periodically check the pressure gauge to be sure that there is no pinhole leak in the rupture disc.05 50.000 50 5 Vapor flow (lblhr) Molecular weight Heat intensity (~tu/hr/ft') Distance from base (ft) Stack diameter (ft) 60.000 5 2 8-20. Describe the advantages and disadvantages of a buckling pin relief valve.000 80 4. and Installation (1994).02 400 0.01 300 0. Describe the advantages and disadvantages of a pilot-operated safety valve. When designing the inlet piping to a relief valve. 8-22.03 64. Sizing. See Grossel and Louvar (2000).000 60 3. See Grossel and Louvar (2000).5 0.0 0.04 64. Determine the vapor velocity inside a horizontal knockout drum for the following three systems: System System System b c a pv (lb/ft3) p~ (lb/ft3) Vapor viscosity (centipoise) Particle diameter (pm) 0.000 30 2. 8-23. Describe several methods to satisfy this requirement. what pressure losses are recommended? See API 520. assuming various maximum heat intensities at ground level at the specified distances from the flare.000 10 3 Case c 80. Develop sketches of reactor vent systems for the following four cases: Case a Reactor relief is vapor only Reactor relief is two-phase flow Reactor contents are corrosive Reactor contents are plugging type Relieved vapors are toxic Relieved vapors are high boilers Vapors are low boilers Case b x x x Case c x Case d x x x x x x x x x 8-18. When using a rupture disc followed by a spring-operated relief. Determine the height of a flare.Chapter 8 Introduction to Reliefs 8-17.01 350 8-19. Describe the advantages and disadvantages of a rupture disc followed by a springoperated relief valve.

and Installation (1994). . Describe a tempered runaway reaction. 8-26. What are these two stresses? See API 520. 8-31. Describe a runaway that is "gassy. 8-28. See Grossel and Louvar (2000). Describe a runaway that is a "hybrid. Selection." See Grossel and Louvar (2000). Sometimes isolation valves are needed between the vessel and the relief. Describe a runaway reaction scenario that is the result of a sleeper reaction. and Installation (1994). The outlet piping of a relief system is normally supported to resist two mechanical stresses.Problems 381 Figure 8-17 Relief valve configurations: What's wrong? 8-25. Identify the problems with the relief valve configurations shown in Figure 8-17. 8-29. What management system is recommended for isolation valves? See API 520. Sizing. 8-30. Selection." See Grossel and Louvar (2000). 8-27. Sizing. See Grossel and Louvar (2000).

.

second. or two-phase) and the type of relief device (spring or rupture disc). The relief vent area calculation depends on the type of flow (liquid. These overpressures frequently exceed the set pressure by 25% or more. vapor. The optimal vent area for a particular relief depends on the specific application. The overpressure specification is part of the relief design. Spring relief devices require 25-30% of maximum flow capacity to maintain the valve seat in the open position." caused by rapid opening and closing of the valve disc. R . This can lead to destruction of the relief device and a dangerous situation. depending on the requirements of the equipment protected and the type of material relieved. purchase. For this reason reliefs must be designed with the proper vent area. In chapter 8 we showed that for liquids and two-phase relief. the relieving process begins at the relief set pressure with the pressure normally continuing to rise past the set pressure (see curve B in Figure 8-2). manufacturers' technical specifications are used for selection. using an appropriate source model to determine the rate of material release through the relief device (see chapter 4) and. and installation. A relief device with an area that is too large for the required flow may chatter. neither too small nor too large. relief devices are specified for overpressures from 10% to 25%. Lower flows result in "chattering. This is one example that illustrates this typical result. Experimental data at the actual relief conditions are recommended for sizing relief vents for runaway reaction scenarios. The relief sizing calculation procedure involves. As always.Relief Sizing elief sizing calculations are performed to determine the vent area of the relief device. As shown in Figure 9-1. first. the relief vent area is reduced substantially as the overpressure increases. Normally. using an appropriate equation based on fundamental hydrodynamic principles to determine the relief device vent area. A relief device designed to maintain the pressure at the set pressure could require an excessively large vent area.

reliefs for dust and vapor explosions. two-phase flow during runaway reactor relief." AlCHE Journal (1986). 9-1 Conventional Spring-Operated Reliefs in Liquid Service Flow through spring-type reliefs is approximated as flow through an orifice. 32(10). C. The vent area is decreased appreciably a s t h e overpressure increases. rupture discs in liquid or vapor-gas service. reliefs for fires external to process vessels. Data from J. An equation representing this flow is derived from the mechanical energy balance (Equation 4-1). The result is similiar to Equation 4-6. and reliefs for thermal expansion of process fluids. In this chapter we present methods for calculating the relief device vent areas for the following configurations: conventional spring-operated reliefs in liquid or vapor-gas service. except that the pressure is represented by a pressure difference across the spring relief: . "Simplified Vent Sizing Equations for Emergency Relief Requirements in Reactors and Storage Vessels. Leung.384 Chapter 9 Relief Sizing Percent Overpressure Figure 9-1 Required vent area a s a function of overpressure for two-phase flow.

A.. . AP is the pressure drop across the relief. Finally. the relief spring is compressed. DC: American Petroleum Institute. (Washington. Selection. Many process fluids have high viscosities. Q. Substituting Equation 9-1 and solving for the vent area A of the relief. In reality. Equation 9-3 does not consider the viscosity of the fluid. A true orifice has a fixed area. The result is API RP 520. Equation 9-3 has been modified by the American Petroleum Institute to include corrections for the above situations. and AP is the pressure drop across the spring relief (lbf/in2). Recommended Practice for the Sizing. and Installation of Pressure Relieving Systems in Refineries. Equation 9-3 does not consider the special case of a balancedbellows-type relief. C. The relief vent area must increase as the fluid viscosity increases. 1993). As the pressure increases. The volumetric flow Q. Also. and p is the liquid density. flow through a spring-type relief is different from flow through an orifice. (p/pref)is the specific gravity of the liquid (unitless). or . 6th ed.is the discharge coefficient. C. we obtain A working equation with fixed units is derived from Equation 9-2 by (1)replacing the density p with the specific gravity (pip. is the volumetric flow through the relief (gpm).) and (2) making the appropriate substitutions for the unit conversions. is the discharge coefficient (unitless).9-1 Conventional Spring-Operated Reliefs in Liquid Service 385 where u is the liquid velocity through the spring relief. The result is where A is the computed relief area (in2). of liquid is the product of the velocity times the area. increasing the discharge area and increasing the flow..

section 4-2.and P. 16(2). a conservative value of 0. is the discharge coefficient. Equation 9-4 appears to assume a maximum pressure equal to 1. corrects for the additional frictional losses rQulting from flow of high-viscosity material through the valve. This correction is given in Figure 9-3. If this value is uncertain. For most reliefs the Reynolds number is greater than 5000 and the correction is near 1. is P. is the gauge backpressure (lbf/in2). "Viscosity Correction Factor for Safety Relief Valves. K. . the procedure is iterative. Darby and Molavi2 developed an equation to represent the viscosity correction factor shown in Figure 9-2. is the viscosity correction (unitless). The viscosity correction K. includes the effect of discharge pressures greater than the set pressure. The overpressure correction K. The required relief vent area becomes larger as the viscosity of the liquid increases (lower Reynolds numbers). is the backpressure correction (unitless). Because the Reynolds number is required to determine the viscosity correction and because the vent area is required to calculate the Reynolds number. Molavi. As the specified overpressure becomes smaller. This equation applies only to Reynolds numbers greater than 100: where K. The overpressure correction K.. Note that the A P term in Equation 9-3 has been replaced by a term involving the difference between the set pressure and the backpressure. is the gauge set pressure (lbf/in2). This correction is given in Figure 9-2. is a function of the overpressure specified for the design. Discharge at other maximum pressures is accounted for in the overpressure correction term K.25 times the set pressure. K. C.Darby and K." Process Safety Progress (1997). is the volumetric flow through the relief (gpm). Q. Specific guidelines for the selection of an appropriate value are given in chapter 4.61 is used to maximize the relief vent area. is the discharge coefficient (unitless). 2R. is the viscosity correction factor (unitless) and Re is the Reynolds number (unitless). (p/pref) the specific gravity of the liquid (unitless). is the overpressure correction (unitless). K.This assumption is frequently used as an initial estimate to begin the calculations.386 Chapter 9 Relief Sizing where A is the computed relief area (in2). C.

DC. DC. 6th ed.9-1 Conventional Spring-Operated Reliefs in Liquid Service 387 R = REYNOLDS NUMBER Figure 9-2 Viscosity correction factor K.3 0. sion of the American Petroleum Institute. ( 993). 9 0. 35.RecommendedPractice for the Sizing. Selection. Washington. 37. O lo Figure 9-3 Overpressure correction K. and lnstallation of Pressure-Relieving Systems in Refineries. Selection. (1993).Used by permisp.Recommended Practice for the Sizing. 6th ed. Source: API RP 520. and Installationof PressureRelieving Systems in Refineries. for spring-operated reliefs in liquid service.Used by permission of the American Petroleum 1 p. V) V) $ 0. for conventional reliefs in liquid service. Institute. Washington. Source: API RP 520.4 - e 8 0 ' l5 I I I I I I I I ' 45 I I 20 Overpressure (%) 25 30 35 40 50 .5 0.2 01 .

or 20 psig. Designs incorporating less than 10% overpressure are not recommended. . 'ItCK PRESSURE. Source: API RP 520. is used only for balanced-bellows-type spring reliefs and is given in Figure 9-4. The overpressure correction factor curve shown in Figure 9-3 shows that up to and including 25% overpressure. Solution a. Washington. Valves operating at low overpressures tend to chatter. Because a deadheaded pump can be easily damaged. The backpressure is specified as 20 psig and the overpressure is 10% of the set pressure. assuming a backpressure of 20 psig and (a) a 10% overpressure and (b) a 25% overpressure. the change in the orifice discharge coefficient. the relief device capacity is affected by the changing discharge area as the valve lifts. resulting in a larger relief area. for 25% overpressure on balanced-bellows reliefs in liquid service. The set pressure is 200 psig. and Installation of Pressure Relieving Systems in Refineries. Above 25% the valve capacity is affected only by the change in overpressure because the valve discharge area is constant and behaves as a true orifice. 6th ed. PSIG Figure 9-4 Backpressure correction K. Selection. The backpressure correction K. Recommended Practice for the Sizing. the correction value decreases.388 Chapter 9 Relief Sizing %GAGE BACK PRESSURE. (1993). so overpressures less than 10% should be avoided. DC. and the change in overpressure.PSIg1 ~ ) x SET PRESSURE. 35. Used by permission of the American Petroleum Institute. compute the area required to relieve the pump. Example 9-1 A positive displacement pump pumps water at 200 gpm at a pressure of 200 psig. This correction compensates for the absence of backpressure on the back of the relief vent disc. p.

. However. at 200 gpm the Reynolds number is assumed to be greater than 5000. Thus for an ideal gas Equation 4-50 is valid: . the downstream pressure must be checked to ensure that it is less than the choked pressure computed using Equation 4-49. K. The backpressure correction is not required because this is not a balanced-bellows spring relief. K.0. = 1. = 200 gpm. The overpressure correction K. Thus Kb = 1.0 (Figure 9-3).9-2 Conventional Spring-Operated Reliefs in Vapor or Gas Service 389 The discharge coefficient C. = 0. As expected. However. These numbers are substituted directly into Equation 9-4: . Manufacturers do not provide relief devices to the nearest 0.0 gpm (0. K.0. For an overpressure of 25%. for a conservative estimate a value of 0.61)(1.01 in.6.851 in. is given in Figure 9-3. 9-2 Conventional Spring-Operated Reliefs in Vapor or Gas Service For most vapor discharges through spring reliefs the flow is critical.61 is used. = 1. from Figure 9-3.0) b. The next largest available size is normally selected.6)(1.0)(0. The quantity of material relieved is the total flow of water. so Q. Because the percentage of overpressure is lo%. the relief vent area decreases as the overpressure increases. However.. For all relief devices the manufacturers' technical specifications must be checked before selection and installation. The Reynolds number through the relief device is not known. from Figure 9-2. is not specified. and _ Jv = 0. Thus the viscosity correction is.[in2(psi)'"] 200 gpm 38. Thus a selection must be made depending on relief device sizes available commercially.

. z is the compressibility factor (unitless). R. A is the area of the discharge. Qmis the discharge flow.). K.: A=- . is the discharge coefficient. and M is the average molecular weight of the discharge material. is the backpressure correction (unitless). y is the heat capacity ratio for the gas.. Qm / -- ygc[ TIM 2 )(~+l)'(~-l). P is the maximum absolute discharge pressure. g.. . The result is where A is the area of the relief vent. Equation 9-6 is simplified by defining a function X : Then the required relief vent area for an ideal gas is computed using a simplified form of Equation 9-6: For nonideal gases and real vents Equation 9-8 is modified by (1)including the compressibility factor z to represent a nonideal gas and (2) including a backpressure correction Kb.390 Chapter 9 Relief Sizing where (Q.. Cois the effective discharge coefficient. C. usually 0. T is the absolute temperature. cop. is the ideal gas constant. and T is the absolute temperature of the discharge. Equation 4-50 is solved for the area of the relief vent given a specified mass flow rate Q. P is the absolute upstream pressure. is the gravitational constant. is the discharge mass flow.975 (unitless). M is the molecular weight of the gas.

is the maximum gauge pressure in psig..33PSfor piping. Equation 9-9 must have the following fixed units: Q.. Used by permission of the American Petroleum Institute. 6th ed. For vapor reliefs the following guidelines are recommended3: P.. PSIA % A B S O L U T E BACK PRESSURE= SET PRESSURE + PSIA X 100 = r X 100 Figure 9-5 Backpressure correction K. The constant x is represented by Equation 9-7.2Psfor vessels exposed to fire. P = P. It is given for the fixed-unit case by. (9-11) where P.. 1. . 33. 1998).7.. p.9-2 Conventional Spring-Operated Reliefs in Vapor or Gas Service 391 . DC. Recommended Practice for the Sizing. is the backpressure correction and depends on the type of relief used. and Installation of Pressure-RelievingSystems in Refineries.. = = 1.and M in lb. Values are given in Figure 9-5 for conventional spring reliefs and in Figure 9-6 for balanced-bellows reliefs. Tin OR. Selection. K. It is conveniently calculated using the following fixed-unit expression: If Equation 9-10 is used. P. " Y BACK PRESSURE. P in psia./hr. (1993). The area computed is in in2.. The pressure used in Equation 9-9 is the maximum absolute relieving pressure./lb-mol. Source: API RP 520. 3 A S M E Boiler and Pressure Vessel Code (New York: American Society of Mechanical Engineers. + 14. Washington. for conventional spring-type reliefs in vapor or gas service.. in lb.

29. for balanced-bellows reliefs in vapor or gas service. p. Source: API RP 520. Recommended Practice for the Sizing. A relief vent must be installed to vent the nitrogen as fast as it is supplied through the 6-in line. If the regulator fails. 6th ed. Because no other information on the piping system is provided. Washington. Assume a relief backpressure of 20 psig. Selection. and the discharge coefficient C. DC. Used by permission of the American Petroleum Institute.392 Chapter 9 Relief Sizing Figure 9-6 Backpressure correction K. The source of the nitrogen is at 70°F and 150 psig. The API Pressure Vessel Code4 provides working equations that are equivalent to Equation 4-48. . (1993). Equation 4-50 describes this: 4API RP 520. The downstream pressure P i s now required. and Installation of Pressure-Relieving Systems in Refineries. Solution The nitrogen source is at 150 psig. Determine the diameter of a balanced-bellows spring-type vapor relief required to protect the reactor from this incident. must be estimated. The reactor relief is set at 50 psig. Example 9-2 A nitrogen regulator fails and allows nitrogen to enter a reactor through a 6-in-diameter line. the nitrogen will flood the reactor. Recommended Practice. For vapor flows that are not choked by sonic flow the area is determined using Equation 4-48. the flow from the pipe is initially assumed to be represented by critical flow through an orifice. increasing the pressure to a point where the vessel will fail.

1)(50 psig) = 55.975. Substituting into Equation 4-50. X 1W = (z) x 100 = 40%. The constant x is computed from Equation 9-10: . = (1. The pressure P is the maximum absolute pressure. This is a 10% overpressure.528)(150 + 14. and the flow from the 6-in line will be critical. The area of the relief vent is computed using Equations 9-9 and 9-10 with a backpressure correction K. The backpressure is 20 psig. y = 1. the choked pressure across the pipe must be determined to ensure critical flow. Thus the pressure in the reactor is less than the choked pressure.7) = 87.0 psia.psig ) set pressure.7 = 164.. The maximum relief design pressure within the reactor during the relief venting is. P.- - 9-2 Conventional Spring-Operated Reliefs in Vapor or Gas Service First. The required quantities for Equation 4-50 are P = 150 + 14. Thus P = 69. we obtain = 3.7 psia. The gas compressibility factor z is approximately 1at these pressures. From Figure 9-6.40 for diatomic gases. from Equation 9-12. Thus psig (backpressure. however. T = 70°F + 460 = 530°R..7 psia.7 psia. = l ./ hr.0 psig = 69. determined from Figure 9-6.86 for an overpressure of 10%.82 x 10' lb. The effective discharge coefficient is assumed to be 0.528P = = (0. l P . For diatomic gases the choked pressure is given as (see chapter 4) Pchoked 0. Kb = 0.

The result is Equation 9-13 assumes a discharge coefficient C. and the flow through the entire pipe system must be considered.125 in). This would likely be 1@ in (10. a procedure similar to the procedure used for liquid reliefs through rupture discs is required. scrubbers. Manufacturers provide relief devices only at convenient sizes. The calculation is performed identically to regular pipe flow (see chapter 4). or flares).Chapter 9 Relief Sizing The required vent area is computed using Equation 9-9: - 3. The calculation to determine the rupture disc area is iterative for this case. For rupture discs discharging into a relief system (which might include knockout drums.14) 10./hr (0.86)(69. Equations 9-2 and 9-3 apply to rupture discs discharging directly to the atmosphere. 22. p. 113. The next largest diameter closest to the one required is selected. 9-3 Rupture Disc Reliefs in Liquid Service For liquid reliefs through rupture discs without significantlengths of downstream piping the flow is represented by Equation 9-2 or by Equation 9-3 for flow through a sharp-edged orifice. x .0.975)(356)(0." Chemical Engineering (Feb. of 1. The procedure is iterative. Isaacs5 recommended assuming that the rupture disc is equivalent to 50 pipe diameters in the calculation. S ~ a rIsaacs. 9-4 Rupture Disc Reliefs in Vapor or Gas Service Flow of vapor through rupture discs is described using an orifice equation similar to Equation 9-9 but without the additional correction factors. No corrections are suggested.82 x 10' Ib.1971).1 in. the rupture disc is considered a flow restriction. If appreciable backpressures exist from downstream relief systems. "Pressure Relief Systems.7psia) The required vent diameter is = (3.

(3.Pb = 220 psig .0.0 gpm] 0.9-5 Two-Phase Flow during Runaway Reaction Relief 395 Example 9-3 Determine the diameter of a rupture disc required to relieve the pump of Example 9-1. two-phase flow should be expected during the relief process.32 in. Example 9-4 Compute the rupture disc vent diameter required to relieve the process of Example 9-2.14) This compares to a spring relief vent area of 1.20 psig = 200 psig. The area is therefore The rupture disc diameter is This compares to a spring relief diameter of 9. Substituting into Equation 9-3.gpm --. The solution is identical to Example 9-2.= [38. Solution The pressure drop across the rupture disc is AP = P. Solution The solution is provided by Equation 9-9. .61 0.881 in..J is 1.610 J z = The relief vent diameter is 0.61 is assumed. 9-5 Two-Phase Flow during Runaway Reaction Relief When a runaway reaction occurs within a reactor vessel. we obtain = in2(psi)"2 200 .10 in. with the exception of a deletion of the correction factor Kb. The specific gravity of the water (plp. A conservative discharge coefficient of 0. part a. The vent sizing package (VSP) laboratory apparatus described in chapter 8 provides the much needed temperature and pressure rise data for relief area sizing. .

is the mass flow through the relief. The first step in the relief sizing calculation for two-phase vents is to determine the mass flux through the relief. v. representing choked two-phase flow through a hole: where. is the heat of vaporization of the fluid. C. The runaway reactor is treated as entirely adiabatic. It is called "tempered" because the reactor contains a volatile liquid that vaporizes or flashes during the relieving process. AH. is the change of specific volume of the flashing liquid. for this case. The heat losses through the reactor walls are assumed negligible. called a tempered reactor.396 Chapter 9 Relief Sizing 2-Phase Flow Re l i e f Device Energy Loss by Vapor iz a t i o n o f Liquid Energy Generated by Reaction 1 I Energy Accumulated by Heating o f L i q u i d Beyond S e t Temperature Figure 9-7 A tempered reaction system showing the important energy terms. The energy terms include (1) energy accumulation resulting from the sensible heat of the reactor fluid as a result of its increased temperature due to overpressure and (2) the energy removal resulting from the vaporization of liquid in the reactor and subsequent discharge through the relief vent. Figure 9-7 shows the most common type of reactor system. and T. This vaporization removes energy by means of the heat of vaporization and tempers the rate of temperature rise resulting from the exothermic reaction. A is the area of the hole. is the heat capacity of the fluid. Q. . is the absolute saturation temperature of the fluid at the set pressure.. This is computed using Equation 4-104.

$ creases. correcting for two-phase flashing flow through pipes. 33(3): 524-527.6Leung7 showed that Equation 9-14 must be multiplied by a factor 0.9 to bring the value in line with the classic homogeneous equilibrium model (HEM). 7J. Leung and M." AICHE Journal (1986). K. 4(3). Fauske. 32(10): 1622. Grolmes. . A. "Flashing Flows or: Some Practical Guidelines for Emergency Releases. = 1. C." AICHE Journal (1987). Equation 9-14 is the so-called equilibrium rate model (ERM) for low-quality choked flow. For two-phase flow through pipes an overall dimensionless discharge coefficient $ is applied. The mass flux GTis given by Equation 9-14 applies to two-phase relief through a hole. "Simplified Vent Sizing Equations for Emergency Relief Requirements in Reactors and Storage Vessels.9-5 Two-Phase Flow during Runaway Reaction Relief Figure 9-8 Correction factor rC." Plant Operations Progress (July 1985). The result should be generally applicable to homogeneous venting of a reactor (low quality. not restricted to just liquid inlet condition): Values for $ are provided in Figure 9-8. C. For a pipe of length 0. As the pipe length in- 6H. Data from J. the value of $ decreases. "The Discharge of Two-Phase Flashing Flow in a Horizontal Duct. Leung.

" . V is the volume of the vessel. 8J. . q is the exothermic heat release rate per unit mass. C. "Simplified Vent Sizing. Leung. and C is the liquid heat capacity at constant volume.398 Chapter 9 Relief Sizing A somewhat more convenient expression is derived by rearranging Equation 4-102 to yield and substituting into Equation 9-15. we obtain The exact derivative is approximated by a finite-difference derivative to yield where A P is the overpressure and AT is temperature rise corresponding to the overpressure. The required vent area is computed by solving a particular form of the dynamic energy balance. is the total mass contained within the reactor vessel before relief. Details are provided e l s e ~ h e r eThe result is .~ An alternative form is derived by applying Equation 4-102: For Equations 9-19 and 9-20 the following additional variables are defined: m.

The heat input q resulting from an exothermic reaction is determined using fundamental kinetic information or from the DIERS VSP (see chapter 8). This applies to most reaction systems. The maximum allowable working pressure (MAWP) of the vessel is 5 bar. The best procedure is to convert all energy units to their mechanical equivalents before solving for the relief area. 3. the equation is applied. Given the following data. "Simplified Vent Sizing. 2. Example 9-5 Leung9 reported on the data of Huff1O involving a 3500-gal reactor with styrene monomer undergoing adiabatic polymerization after being heated inadvertently to 70°C. The equations assume that 1. "Emergency Venting Requirements. corresponds to the temperature rise at the maximum turnaround pressure.5"C = 492." 1°J. the second term corresponds to the heat absorbed as a result of increasing the temperature of the liquid because of the overpressure.7"CImin = 0.4 bar absolute: Data Volume (V): 3500 gal = 13. = 29. Huff.5 bar and a maximum pressure of 5.): 9500 kg Set temperature (T. AHv. Both derivatives are determined experimentally using the VSP. l(4): 211 .4"C = 482. the system is a tempered reactor system.): 219. E.662 Kls 'Leung. 4. = 39. Units are a particular problem when using the two-phase equations. uniform froth or homogeneous vessel venting occurs. where the derivative. The first term in the denominator corresponds to the net heat removed by the liquid and vapor leaving the system. physical properties Cv.493 Kls (dTldt). corresponds to the heating rate at the set pressure and the derivative." Plant/Operations Progress (1982).6"Clmin = 0. and vfg are constant. For data obtained using the VSP.5 K Data from VSP Maximum temperature (T. and 5.16 m3 Reaction mass (m. the mass flux GTvaries little during the relief. particularly when English engineering units are used. determine the relief vent diameter required. denoted by the subscript s. denoted by the subscript m. the reaction energy per unit mass q is treated as a constant.7 K (dTldt).): 209.9-5 Two-Phase Flow during Runaway Reaction Relief 399 For both Equations 9-19 and 9-20 the relief area is based on the total heat added to the system (numerator) and the heat removed or absorbed (denominator). Assume a set pressure of 4.

$ = 1. The change in temperature AT is Tm.4-bar peak Solution The heating rate q is determined using Equation 9-21: Assuming that C.0: = (0.0.9)(1.5-bar set 5.5 = 10..7 .600 J/kg)[l (N m)/J] X (0.001388) m3/kg [ I (kg m/s2)/N] (2470 Jlkg KX482.5 K)[1 (N m)/Jl The relief vent area is determined from Equation 9-19.400 Chapter 9 Relief Sizing Physical Property Data 4.T. we have The mass flux is given by Equation 9-15.482. Assuming LID = 0. = C.2 K: The required relief diameter is .08553 . = 492.0) (310.

= 1.32 and C." Plant/Operations Progress (1984).600 Jlkg) Equation 9-6 provides the required relief area. Simplified Nomograph Method Fauske l1 developed a simplified chart-driven approach to the two-phase reactor relief problem.9-5 Two-Phase Flow during Runaway Reaction Relief 401 Suppose that all vapor relief was assumed. Assume that y = 1." Plant/Operations Progress (1984). Thus. 3(3). "A Quick Approach to Reactor Vent Sizing. At the set temperature the heat release rate q is The vapor mass flow through the relief is then - (1218 Jlkg s)(9500 kg) (310. the result would be physically incorrect and the reactor would be severely tested during this runaway event. . He suggested the following equation for determining the relief area: "Hans K. a significantly smaller diameter than for two-phase flow. if the relief were sized assuming all vapor relief.0. Then (8314 Pa m3/kg-mol K)(482. The molecular weight of styrene is 104.176 m. The size of the required vapor phase rupture disc is determined by assuming that all the heat energy is absorbed by the vaporization of the liquid.5 K)[1 (N/m2)/Pa) X d r (1. 3(4). Fauske.32) [ I (kgm/s2)/~](104 kglkg-mol) This requires a relief device with a diameter of 0. and "Generalized Vent Sizing Nomogram for Runaway Chemical Reactions.

. J. C. and Atvis the venting time. assuming a typical liquid heat capacity of 2510 J/kg K for most organic materials. is the heat capacity. T is the temperature. Equation 9-22 was developed by Boylel"y defining the required vent area as the size that would empty the reactor before the pressure could rise above some allowable overpressure for a given vessel. and q. V is the reactor volume. and assuming a saturated water relationship. is the energy release rate per unit mass at the set pressure of the relief system. 63(8): 61. "Sizing Relief Area for Polymerization Reactors. GTis the mass flux through the relief.402 Chapter 9 Relief Sizing where A is the relief vent area. Boyle Jr.9-14. By considering the case of 20% absolute overpressure. 6(2). and 9-23 yields Equation 9-24 provides a conservative estimate of the required vent area. 135. we can obtain the following equation 13: A = (m2/1000kg) = Ps(baI-1 "w. Combining Equations 9-22." Chemical Engineering Progress (August 1967). p is the density of the reactants. . Fauske. K. The mass flux GTis given by Equation 9-15 or 9-18. and the venting time is given approximately by where AT is the temperature increase corresponding to the overpressure AP. C." Plant/Operations Progress (April 1987). Leung and H. "Runaway System Characterization and Vent Sizing Based on DIERS Methodology.

Used by permission of the American Institute of Chemical Engineers. Leung. Fisher and J." Plant/Operations Progress (1984). The Fauske nomograph is useful for performing quick estimates and for checking the results of the more rigorous computation. personal communication.9-5 Two-Phase Flow during Runaway Reaction Relief 403 I 1 0 SELF HEAT RATE. A simple nomograph of the results can be plotted and is shown in Figure 9-9. K. + 14H. . requires a suitable correction. January 1989. the set pressure. Source: H. Recent studies14suggest that the nomograph data of Figure 9-9 apply for a discharge coefficient of = 0. C.5. 100 OC/min Figure 9-9 Nomograph for sizing two-phase reactor reliefs. representing a discharge (LID) of 400. and the mass of reactants. "Generalized Vent Sizing Nomogram for Runaway Chemical Reactions. The required vent area is determined simply from the heating rate. Use of the nomograph at other discharge pipe lengths and different rC. as shown in the following example. 3(4). G. Fauske.

03 X 0. = 1. propagating the accident and resulting in increased injuries. m2/1000kg)(9500 kg) Figure 9-9 is applicable for 9 = 0. The result can be adjusted for other overpressures by multiplying the area by a ratio of 20/(new absolute percentage of overpressure). o r processing unit can eject high-velocity structural debris over a considerable area. however. Deflagration venting in buildings and process vessels is usually achieved by using blowout panels. storage bin. they d o represent the most accepted method available today. during the milling process to make flour from wheat.9869bar/atm)(l4.Thus the From Figure 9-9 the vent area required per 1000 kg of reactant is about 1. Solution The heating rate at the set temperature is specified as 29. For IC. The set pressure is 4.6"C/min.5 bar absolute.7 psialatm) = 65. A n uncontrolled dust explosion in a warehouse. substantial quantities of flammable dust are produced. For example. This result compares to a more rigorously computed area of 0.098 m2.5.0 the area is adjusted linearly: This assumes a 20% absolute overpressure. The blowout panel is designed to have less strength than the walls . 9-6 Deflagration Venting for Dust and Vapor Explosions Loss prevention means preventing the existence of hazards. The energy of the explosion is directed away from plant personnel and equipment. Furthermore.- - - - 404 Chapter 9 Relief Sizing Example 9-6 Estimate the relief vent area using the Fauske nomograph approach for the reaction system of Example 9-5. for some situations hazards are unavoidable. Two-phase flow through reliefs is much more complex than the introduction provided here. as shown in Figure 9-10. Deflagration venting reduces the impact of dust and vapor cloud explosions by controlling the release of the explosion energy. m2.084 m2. the technology is still undergoing substantial development. However. so P.3 psia.03 X total relief area is A = = (1. The equations presented here are not universally applicable.5 bar)(0. = (4.

of the structure. it is not unusual for the walls (and perhaps roof) of the entire structure to be constructed of blowout panels. Furthermore. Thus. depending on a number of design factors.9-6 DeflagrationVenting for Dust and Vapor Explosions 405 n J D u s t o r Vapor Explosion Occurs Structure Attachment o f Vent P r e v e n t s P r o j e c t i o n o f Vent D e b r i s D e f l a g r a t i o n V e n t Opens. the blowout panels are preferentially detached and the explosive energy is vented.'" l5 N F P A 68. 1998). These include the explosive behavior of the dust or vapor. The actual construction details of blowout panels is beyond the scope of the text. and the volume of the structure. Damage to the remaining structure and equipment is minimized. Construction details are available in manufacturers' literature. For particularly explosive dusts or vapors. . Therefore a mechanism must be provided to retain the panel during the deflagration process. thermal insulation of panels is also required. Guidefor Venting of Depagrations (Quincy. during an explosion. Design standards are available. the maximum overpressure allowable in the structure. R e d u c i n g O v e r p r e s s u r e and R e s u l t i n g Damage Vent Figure 9-10 Deflagration vents for structures and process vessels. A detached blowout panel moving at high velocity can cause considerable damage. M A : National Fire Protection Association. Blowout panels are designed to provide the proper relief area.

.. and P i s the maximum internal pressure that can be withstood by the weakest member of the enclosure. (New York: American Institute of Chemical Engineers.1 bar gauge).5 psig (0. is the inside surface area of the enclosure.*.Chapter 9 Relief Sizing Deflagration design is segregated into two categories: low-pressure and high-pressure structures.1 bar gauge). "Recent Developments in Deflagration Venting Design. including many important combustion features: where A is the required vent area. 16Richard R. is the smallest dimension of the rectangular building structure to be vented. L." Plant/Operations Progress (April 1987). 6(2). is L. "Performance of Low Pressure Explosion Vents. Woodward. is the second smallest dimension of the enclosure to be vented.a constant that depends on the nature of the combustible material. Low-pressure structures include structures with sheet metal sides and other lowstrength building materials. Swift and Epstein17 presented a more detailed equation. 1987). Schwab. Vents for Low-Pressure Structures For low-pressure structures that are capable of withstanding pressures of not more than 1. ed.101. A. and so forth that are capable of withstanding pressures greater than 1." in Proceedings of the International Symposium on Preventing Major Chemical Accidents. p.. High-pressure structures include steel process vessels.1 bar gauge). John L. original design techniques were based on the Runes (pronounced Rooness) equation16: where A is the required vent area. C. 171an Swift and Mike Epstein. concrete buildings. 3. These structures are capable of withstanding not more than 1.5 psi (0.5 psig (0.

1998). Equation 9-28 is remarkably similar to the Runes equation (Equation 9-26).51 INFPA.- - 9-6 Deflagration Venting for Dust and Vapor Explosions Table 9-1 Combustible Characteristic Constant for the Swift-Epstein Equation1 Combustible material Anhydrous ammonia Methane Aliphatic gases (excluding methane) or gases with a fundamental burning velocity less than 1.20 0. M A : National Fire Protection Association. Many of the variables in Equation 9-27 can be estimated or assumed. is the laminar burning velocity. Three of the walls are shared with an adjoining structure. G is the mass flux. P. The fourth and larger wall of the room is on the outer surface of the structure. p. These variables are regrouped to result in the following form: where P i s the maximum internal overpressure that can be withstood by the weakest structural element.3times that of propane.10 0. Example 9-7 A room is used for dispensing flammable liquids.17 0. The room is 9 m long by 6 m wide by 6 m in height.. Values for the constant C. is the maximum unvented explosion pressure.37 Gent 0..Venting of Deflagrations (Quincy. is the density of the unburned gas. Pois the initial pressure. S . Estimate the vent area required for this operation.14 Cvent (m) 0.12 0. C..30 0.26 0.. The liquids are expected to have fundamental burning velocities less than 1. P.is the discharge coefficient. are given in Table 9-1.05 0..13 0. The three inside walls are capable of withstanding a pressure of 0. and y is the heat capacity ratio.45 0. .05 bar. is the final peak pressure during the vent.3 times that of propane St-1 dusts St-2 dusts St-3 dusts (m) 0. h is the turbulent augmentation factor.

I8NFpA68. The total surface area of the room (including floor and ceiling) is m.18The length-to-diameter ratio LID of the enclosure determines the equation(s) used for calculating the necessary vent area. where A is the cross-sectional area normal to the longitudinal axis of the space. For noncircular enclosures the value used for the diameter is the equivalent diameter given by D = 2 . Tables of typical values were also provided. Vents for High-Pressure Structures High-pressure structures are capable of withstanding pressures of more than 1. (dPldt).. determined experimentally. .5 psig (0. We discussed the experimental procedure used to determine the deflagration indexes for gases and dusts in chapter 6. This would reduce the vent area required. Guidefor Venting of Deflagrations (Quincy. One option is to strengthen the three inner walls to withstand a higher pressure. max where KG is the deflagration index for gases and vapors. The vent design is based on the definition of a deflagration index for gases or dusts: KG or Kst = (s) V113.. is the maximum pressure increase. Kst is the deflagration index for dusts. 1998). The required vent area is This is larger than the area of the outer wall. Extensive testing with dusts and vapors has resulted in a detailed set of empirical equations for the relief vent area (published as NFPA 68). The venting constant for this flammable vapor is provided in Table 9-1 and has a value of 0.45 Equation 9-28 is used to estimate the required vent area.408 Chapter 9 Relief Sizing Solution The vent must be installed on the larger outer wall to vent the combustion away from the adjoining structure. and V is the volume of the vessel.1 bar gauge). MA: National Fire Protection Association.

Equation 9-30 has the following restrictions on its use: KG is less than 550 bar-mls. P. and P. is less than 2 bars.. KG is the deflagration index for the combusting gas (bar-mls). NFPA 68 must be consulted. is less than 0.. and 250 m3with an LID for all test vessels of approximately 1. ignition energy of 10 J. KG is the vapor deflagration index (bar-mls). The experimental conditions under which Equation 9-30 was developed are as follows: vessel volumes of 2. initial pressure atmospheric. For LID values greater than 5.. ranging from 0... and LID is the length to diameter ratio of the enclosure (unitless).5 bar. .1 bar to 0. is the vent area (m2). is the vent release pressure (bar)..4.- - - - 9-6 Deflagration Venting for Dust and Vapor Explosions For combusting vapors discharging through a low inertial vent and an enclosure LID of less than 2.5 bar. P. P.. and V is less than 1000 m3. the following equation from NFPA 68 applies: where A...10. and P. is the maximum pressure during venting (bar)... For enclosures with LID ranging from 2 to 5 the area calculated using Equation 9-30 is adjusted using where AA is the adjustment to the vent area of Equation 9-30 (m2)... quiescent gas mixture at time of ignition with no turbulence inducers..25.

) and that the KG of the vapor is 100 bar-mls...5 bars and higher Equation 9-32 should be used alone. is the maximum pressure during venting (bar)..... the additional vent area is large. P.1 and 1bar. Example 9-8 Consider again the flammable liquid dispensing room of Example 9-7. 0. values of 1. is the maximum pressure reached during deflagration of an optimum mixture of combustible dust and air in a closed vessel (bar).. The longitudinal axis runs the 9 m length of the room. For P.. the following equation from NFPA 68 applies: where A.. the vent area of Equation 9-32 is increased by.1 and 2 bars...4 bar (Pred). The cross-sectional area normal to this axis is (6 m)(6 m) = 36 m2. Estimate the vent area required to protect this enclosure. D (9-33) For The adjusted vent area (Equation 9-33) is sensitive to Pred. low values of PI. Pmax must be between 5 and 10 bars. is the vent release pressure (bar). Solution The LID ratio must first be determined for this enclosure. must be between 5 and 12 bars.(-4.2 bar (P..410 Chapter 9 Relief Sizing For venting of combustible dusts through a low inertial vent and an enclosure LID of less than 2... For LID values equal to or greater than 2 but less than 6 and for Predless than 1.305 log Pred+ 0. between 300 and 800 bar-m/s.758)log -. must be between 0. P. In this case the walls have Assume that the vent will operate at been reinforced to withstand a pressure of 0. The following limitations apply to Equation 9-32: for Kst between 10 and 300 bar-mls.. For long pipes and ducts where LID is greater than 6. L AA = A. Kst is the dust deflagration index (bar-mls). PI. must be between 0. P. Thus D = . NFPA 68 must be consulted. P.. for Ks..000 m3..5 bars (22 psi). P. is the vent area (m2). the enclosure volume must be between 0. and V is the volume of the enclosure (m3).1 and 10..

0567)(0. Substituting into Equation 9-30.4)-~. Two-phase flow during these reliefs is possible but not likely. = [(0.175(0. Two-phase flow during fire relief can therefore be prevented by providing a suitable vapor space above the liquid within the vessel. For most fires only a fraction of the external vessel i exposed to fire.~~~(0.1)](324)~'~ More than adequate area exists on the outer wall of the enclosure to accommodate this vent. we obtain A. Then LID = 9 m16. 9-7 Venting for Fires External to Process Vessels Fires external to process vessels can result in heating and boiling of process liquids. Thus liquid boiling will occur only next to the wall. s . Venting is required to f prevent vessel rupture. KG or Ks.0.0. and the resulting two-phase foam or froth at the liquid surface will not have a substantial thickness. Both vent sizing methods for gases and dusts require values for the deflagration indexes. For heating caused by external fire the heating occurs only at the surface of the vessel.- - 9-7 Venting for Fires External to Process Vessels 21/(36 m2)/. Venting is required to prevent explosion of these vessels.77 m = 1. For runaway reactor reliefs the energy is generated by reaction throughout the entire reactor liquid contents.77 m. as shown in Figure 9-11. We discussed the experimental procedure to determine these values and also provided tables of typical values for gases and dusts in chapter 6. The volume of the enclosure is (9 m)(6 m)(6 m) = 324 m3.2 .127 log 100 .3.rr = 6. and it follows that Equation 9-30 applies without further correction.~~~ + 0. Pressure i n Vessel B u i l d s Due t o I n c r e a s e d r Pressure and/or rnposition Products Figure 9-11 Heating o a process vessel a s a result of an external fire.4)-~.

is the liquid mass in the vessel. T. for Q = 21. 1 9 9 . "Sizing Relief Valves for Fire Emergencies. is the maximum temperature in the vessel. MA: National Fire Protection Association.. In this case the correct solution is the minimum mass flux GT.. T. Cv is the heat capacity at constant volume. This assumes a constant heat input rate Q: where T. Other standards are also a ~ a i l a b l eCro~ier. 1998. is the set temperature corresponding to the set pressure. DC: US Department of Labor. and Equation 9-34 reduces to Various relationships have been recommended for computing the heat added to a vessel that is engulfed in fire. m. Venting Atmospheric and Low-Pressure Storage Tanks (Nonrefrigerated and Refrigerated).1985).000~'. The solution to Equation 9-34 for GTAis done by an iterative or trial-and-error technique. Equation 9-34 is likely to produce multiple roots. 5th ed. V is the volume of the vessel. Flammable and Combustible Liquids (Washington. is the volume difference between the vapor and liquid phases.412 Chapter 9 Relief Sizing Two-phase fire relief equations are available for conservative design. A. Crozier.106. 1996). recom. Flammable and Combustible Liquids Code (Quincy.. GTis the mass flux through the relief. Leung19presented an equation for the maximum temperature based on an energy balance around the heated vessel. = T." Chemical Engineering (Oct. A is the area of the relief.'~ for A > 2800." ZOOSHA1910.106 criterion20 is mandatory. v.OOOA for 20 < A < 200.~~ after mended the following equations for determining the total heat input Q: Q Q = = 20. 2000).. 28. (Washington. (9-36) I9Leung. ~ ~ ' 1000 < A < 2800. DC: American Petroleum Institute.2~ analysis of the various standards. For regulated materials the OSHA 1910. 4 0 0 ~ ~ . "Simplified Vent Sizing Equations. is the heat of vaporization of the liquid. . 3 0 0 ~ ' . 21API Standard 2000. 22R. and NFPA 30. and AH. ~ ~for 200 < A < 1000. Q is the constant heat input rate. ~ Q = 9 3 6 .For the special case of no overpressure.

Vessels.9-7 Venting for Fires External to Process Vessels 413 Table 9-2 Environment factors F for Equations 9-37 and 9-38 Insulation thickness (in) Environment factor F where A is the area absorbing heat (in ft2) for the following geometries: for spheres. A number of values for various insulation thicknesses are shown in Table 9-2. 23APIR P 520. 1993). then the heat flux is estimated using If adequate fire fighting and drainage do not exist. 55% of total exposed area.Y. for horizontal tanks. If prompt fire fighting is available and if the flammable material is drained away from the vessel. for vertical tanks. The surface area A is the area of the vessel wetted by its internal liquid with a height less provided considerably more detail on how to deterthan 25 ft above the flame source. The mass flux GTis determined using Equation 9-15 or 9-18. (Washington. 100% of total exposed area for first 30 ft. and Installation of Pressure-Relieving Devices in Refineries. DC: American Petroleum Institute. . WongZ4 mine this and provided a number of equations for various vessel geometries. "Fires. 6th ed. Sizing. then the following equation is used: where Q is the total heat input through the surface of the vessel (Btuthr). The environment factor F is used to account for vessel protection from insulation. and Q is the total heat input to the vessel (in Btulhr). F is an environment factor (unitless). API 52OZ3 suggests a slightly different approach to estimate the heat flux to process equipment as a result of a fire. 75% of total exposed area. Wong. and A is the total wetted surface of the vessel (ft2). 24W. Selection." Chemical Engineering (May 2000). and the Pressure Relief Valve.

+ 25Leung.74 X 105J/kg.41 AH. based on the saturation pressure.55)(1121 ft2) = 616 ft2. Then Q = 2 1 .414 Chapter 9 Relief Sizing Example 9-9 Leung2"eported on the computation of the required relief area for a spherical propane vessel exposed to fire. we obtain X lo6 Btulhr. The vessel has a volume of 100 m3 and contains 50. 3.82 6. 0 0 0 ~ A ~ " (21. The surface area of the sphere is The area exposed to heat is given by the geometry factors provided with Equation 9-36: A = (0. then F = 1. Solution The problem is solved by assuming no overpressure during the relief.700 kg of propane.5 bars absolute is required.65 = ~ = which is close to the value estimated using Equation 9-36. A set pressure of 4.76m. The molecular weight of propane is 44.0.5 K. The total heat input is found using Equation 9-36: If we use Equation 9-37 and assume that the vessel is full of liquid.14) ] = 5." .0. At these conditions the following physical property data are reported: Cp = CV= 2.000)(1. then the entire vessel surface area is exposed to the fire. The relief vent area calculated is larger than the actual area required for a real relief device with overpressure.0)(1121ft2)0. = X 103 J/kg K. If we also assume that no insulation is present. This corresponds to a set temperature. of 271. From Equation 9-15 and assuming = 1. The diameter of the sphere is d = (?)"'= [ (6)(100 m") (3."Simplified Vent Sizing Equations.

damage to heat exchange systems can result in (1) contamination of product or intermediate substances. leading to damage to the cooling coils. (3) substantial plant outages. is defined as . two-phase flow discharges for fire scenarios are possible but not likely. p. and determine the vapor mass flow rate through the relief by dividing the heat input by the heat of vaporization of the liquid.308 m = 12.1 in. Failure in heat exchange equipment is also difficult to identify. and (4) large repair expenses. To size the relief for fire and a single-vapor phase. the water will expand when heated by the reactor contents. use the heat input determined from Equations 9-36 to 9-38. What initial fill fraction should be specified in the tank to avoid two-phase flow during a fire exposure incident? No tested correlations are presently available to compute the height of a foam layer above the boiling liquid. If the coils are filled with water and are accidentally blocked in. The expansion will damage pipes and vessels if the pipe or vessel is filled completely with fluid and the liquid is blocked in. This assumes that all the heat input from the fire is used to vaporize the liquid. A thermal expansion coefficient for liquids. An alternative way to look at the problem might be to ask the question.9-8 Reliefs for Thermal Expansion of Process Fluids 415 The required vent area is determined from Equation 9-35: The required diameter is = 0. and repairs are time consuming. For fire reliefs with single-phase vapor flow the equations provided in sections 9-2 and 9-4 are used to determine the size of the relief. 9-8 Reliefs for Thermal Expansion of Process Fluids Liquids contained within process vessels and piping will normally expand when heated. As mentioned previously. A typical situation is thermal expansion of water in cooling coils in a reactor. Although this may appear to be a minor problem. shown in Figure 9-12. (2) subsequent corrosion problems. The relief area is then determined using Equations 9-3 to 9-12. Relief vents are installed in these systems to prevent damage resulting from liquid expansion.

). through the relief resulting from thermal expansion is By applying the definition of the thermal expansion coefficient. we obtain For a pipe or process vessel heated externally by a hot fluid. Table 9-3 lists thermal expansion coefficients for a number of substances. where V is the volume of the fluid and T is the temperature. Coefficients for water are readily determined from the steam tables.- dT dt = UA(T - T. after which the thermal expansion coefficient increases with temperature. The thermal expansion coefficient decreases with increasing temperature up to about 4"C.416 Chapter 9 Relief Sizing Hot Proce L i q u i d I n Cool ing Water Cool ing Wa ter B o t h Wate Va I v e s C l o s e d Out In- A -Coo I i n g Co i l s F i l l e d with Liquid Heat Transfer from Hot Vessel F l u i d t o Cool ing L i q u i d Increases Cool ing L i q u i d Temperature Leading t o Thermal Expansion. The volumetric expansion rate Q. Water behaves in an unusual fashion. . an energy balance on the fluid is given by mC. Figure 9-12 Damage to cooling coils as a result of external heating of blocked-in cooling fluid. given by Equation 9-39.

).. (Englewood Cliffs. Elements of Physics. UA is an overall heat transfer coefficient. NJ: Prentice Hall. is the ambient temperature. sufficient for sizing a relief device. changing the value of T However. 4th ed. and.9-8 Reliefs for Thermal Expansion of Process Fluids 417 Table 9-3 Liquid Thermal Expansion Coefficients for a Variety of Liquids1 Density at 20°C (kg/m3) 791 792 877 1.261 13. invoking the definition of the liquid density p. it is apparent that Equation 9-45 provides the maximum thermal expansion rate. methyl Benzene Carbon tetrachloride Ether. Substituting into Equation 9-41.T.546 873 Thermal expansion coefficient ("C-') 112 x 120 x 124 X 124 x 166 X 51 X 18. and T.). 1965). .= -(T dT dt UA mCp . Shortley and D. Equation 9-45 describes the fluid expansion only at the beginning of heat transfer. The heat transfer will increase the temperature of the liquid. = -UA(T - T. The volumetric expansion rate Q. ethyl Glycerin Mercury Turpentine lo-5 lo-' lo-' lo-s lo-' lo-s lo-s 'G.2 x 97 x Alcohol. It follows that . 302. Williams. p. where T is the temperature of the fluid.). Q. Cpis the heat capacity of the liquid. = -UA(T PV - ~ C P T. when the fluid is initially exposed to the external temperature T. ethyl Alcohol. we obtain Q.595 714 1. is subsequently used in an appropriate equation to determine the relief vent size.

Given a heat transfer coefficient of 50 Btulhr-ft2-OF. However.estimate the volumetric expansion rate of the water in the cooling coils in gpm. A value between 32°F and some appropriate higher temperature will suffice.) The expansion coefficient is computed using Equation 9-39: The volumetric expansion rate is given by Equation 9-45: = 102 ft3/hr = 12. "Pressure Venting of Dust Explosions in Large Vessels." Plant/Operations Progress (October 1986).000 ft2. 5(4): 196. The relief device vent area must be designed to accommodate this volumetric flow. . This is estimated using liquid volumetric data from the steam tables over a short range of temperatures around 32°F. Solution The expansion coefficient P for water at 32°F should be used. Under the most severe conditions the coils can contain water at 32OF and can be exposed to superheated steam at 40OoF.418 Chapter 9 Relief Sizing Example 9-10 The cooling coil in a reactor has a surface area of 10. Suggested Reading Deflagration Vents W. From the steam tables: Temperature (OF) Specific volume (ft3/lb. the steam tables do not provide liquid water specific volume data below 32°F. Bartknecht.7 gpm.

K. Shaw. 1993)." Chemical Engineering Progress (February 1999). API RP 521. S. 32(10): 1622. Fauske. 3d ed.Suggested Reading 41 9 Frank T. K. "Generalized Vent Sizing Nomogram for Runaway Chemical Reactions. First and J. S. Guide for Pressure-Relieving and Depressurizing Systems. Venting Atmospheric and Low-Pressure Storage Tanks (Nonrefrigerated and Refrigerated). 32(10): 1743. and R. Leung. E. 14(2). C. Leung. 5th ed. and Installation of Pressure Relieving Systems in Refineries. D. "Properly Sized Vents for Nonreactive and Reactive Chemicals. DC: American Petroleum Institute. Huff. K. H. Bodurtha. H. K. Muller. Fisher. R. E. "Performance of Low Pressure Explosion Vents. Guide for Venting of Deflagrations (Quincy. Huff. Fauske and J." Chemical Engineering Progress (August 1985). (Washington. E. 33. . Forrest. Tilley. J. Industrial Explosion Prevention and Protection (New York: McGraw-Hill." AICHE Journal (October 1986). 3(4). "New Experimental Technique for Characterizing Runaway Chemical Reactions. API Standard 2000. DC: American Petroleum Institute. Noronha. Grossel. "Emergency Relief System (ERS) Design: An Integrated Approach Using DIERS Methodology. C. Fauske. J. J. A. "Emergency Relief System (ERS) Design. G. 1980). "DIERS Research Program on Emergency Relief Systems. 1990)." Chemical Engineering Progress (February 2000). Relief Codes API RP 520. ASME Boiler and Pressure Vessel Code (New York: American Society of Mechanical Engineers. Emergency Relief System Design Using DIERS Technology (New York: American Institute of Chemical Engineers. H. K. Fauske. Ian Swift and Mike Epstein. H. Fauske. 1998). Fauske." Chemical Engineering Progress (August 1985). 1998). "Flashing Flows or Some Practical Guidelines for Emergency Releases. J. A. "A Generalized Correlation for One-Component Homogeneous Equilibrium Flashing Choked Flow." Process Safety Progress (April 1995). Harold G." Chemical Engineering Progress (August 1985). Selection. "Determine Two-Phase Flows During Release." AICHE Journal (October 1986). "Design Charts for Two-Phase Flashing Flow in Emergency Pressure Relief Systems. C. (Washington. p. 1992). K. Recommended Practice for the Sizing." Plant/Operations Progress (July 1985)." Plant/Operations Progress (April 1987). H. Two-Phase Flow G. W. Fisher. (Washington. A. H. J. Boicourt. Leung. "Simplified Vent Sizing Equations for Emergency Relief Requirements in Reactors and Storage Vessels. K. DC: American Petroleum Institute." Plant/Operations Progress (October 1984). H. MA: National Fire Protection Association. 6th ed. NFPA 68. S. H. 1998)." Paper presented at 1988 AICHE Spring National Meeting. 6(2).

420 Chapter 9 Relief Sizing J. O 1.I 9-3. and the temperature is 100°F. "Two-Phase Flow Venting from Reactor Vessels. Assume a specific gravity of 1. Fisher. Problems 9-1. Determine the diameter of a spring-type vapor relief for the following conditions. 50 Set OverBackpressure pressure pressure (psig) (%I ("/. 100 b. 1000 b. 33(3): 524. Grolmes. 100 Pressure drop (psi) 100 100 50 50 ." AICHE Journal (March 1987).2 9-2. 1000 d.3. C. Determine the required diameter for rupture discs for the following conditions. Compressibility. 2(2): 78. "A Generalized Correlation for Flashing Choked Flow of Initially Subcooled Liquid. Grolmes.3 1. Estimate the diameter of spring-type liquid reliefs for the following conditions: Pump capacity at AP (gpm) a.I 50 100 50 20 20 10 10 30 40 Valve tY Pe Conventional Balanced-bellows Balanced-bellows (~I~ret) 1. J. C. Leung and M. 34(4): 688." Journal of Loss Prevention (April 1989). Leung and M. Liquid flow (gpm) a. Molecular z weight Mass flow (Iblhr) Overpressure (%I Backpressure ("/. A. Assume for each case that y = 1. C.2 for all cases. "The Discharge of Two-Phase Flashing Flows in a Horizontal Duct. 200 c. Leung and H." AICHE Journal (April 1988). A. J. 100 c. G. the set pressure is 100 psia.

ft3 Set pressure. "Fls Liquid specific volume. psia Set temperature.4 1. Assume a spherical vessel in each case.1 130 10.52 86 5000 15. "F (dTldt). ftylb Heat capacity.0 120 520 2.66 0.4 1. 100 d.000 200 100 500 0.6 0.40 72 5000 15.66 0. Btullb v. 200 c. Gas flow (Iblhr) a..5 120 520 0. Ib Set pressure. Molecular weight Volume.000 100 220 100 220 130 1.000 100 220 150 275 150 1.6 0. psia Maximum temperature. "Fls Maximum pressure. Btullb "F 72 5000 30.40 86 5000 15. Assume in all cases that LID = 0. Reaction mass.000 100 220 130 240 150 1.6 0. O F Maximum pressure. 200 Pressure (psi@ 100 100 50 50 9-5.6 0.Problems 9-4.1 130 10. Assume that nitrogen is the vent gas and that the temperature is 100°F.6 0.000 100 220 100 220 130 1.000 500 100 500 2.000 500 100 500 0.0.1 130 9-6. psia Set temperature. Determine the required diameter for rupture discs in vapor service for the following conditions. Btullb "F Heat of vaporization. Btullb 10.1 130 10.. O F (dTldt).000 500 100 500 2.02 1. ft3/lb Vapor specific volume. Determine the relief vent areas for the following two-phase fire scenarios. psia Maximum temperature.4 1. How is the overpressure included in the design of two-phase reliefs? 9-7.2 1..02 1.4 1. 100 b.02 1. ft"/lb C..0 140 550 2. O F Heat of vaporization.52 ..4 0. Determine the proper relief diameter for the following two-phase flow conditions. Ib Volume. ft" Initial mass.5 120 520 0.

422 Chapter 9 Relief Sizing 9-8. Determine the relief size required. is completely filled with water and blocked in. 4 ft in diameter and 10 ft long.4 1000 3 16. 9-13.6 c 1000 16. Assume 20% backpressure. Estimate the thermal expansion rate of the water if the water is at 50°F and .4 d 1000 16.6 1000 3 16. "F Heat transfer coefficient. part a. This time use alcohol as a liquid medium with a thermal expansion coefficient of 1. psia Gas Dusts - Methane a - Hydrogen b - Methane c - Hydrogen d - Volume.000 550 70 75 10. psia Maximum pressure.000 550 50 75 10.6 b 1000 16. 9-11.58 kcallkg "C.7 17. Home hot water heaters contain relief devices to provide protection in the event that the heater controls fail and the water is heated to a high temperature.12 x 10-3/0C.6 1000 1 16. Btulhr ft2O F 10. A cylindrical tank. What size line would you select if the nitrogen is available from a 500-psig source? The temperature is 80°F.7 29. and its density is 791 kg/m3.000 800 32 75 10.7 29.000 Btulhr. Also compute the relief vent size assuming all vapor relief. psia 1000 1 16.4 Volume. and no backpressure. Assume a set pressure of 500 psig. Hint: Two-phase flow is expected.7 20. compute the area required for relief. Consider Problem 9-9.7 17. A process vessel is equipped with a 2-in rupture disc set at 100 psig and designed for 10% overpressure.4 9-9. 9-12. Assume that the tubes can withstand a pressure of 1000 psig and that the normal operating pressure is 200 psig. Determine the relief size for spray dryers operating under the following conditions: Vapors a 1000 16. an overpressure of 20%. ft2 Maximum temperature. Assume no overpressure. Determine the size of relief required to protect the following cooling coils against thermal expansion.The heat capacity of the alcohol is 0.7 29. ft" Set pressure. A nitrogen line must be added to the vessel to provide the capability of purging and/or pressure discharging liquids.000 550 70 125 9-10. A typical water heater contains 40 gal of water and has a heat input of 42.7 20. If the heater is equipped with a 150-psia spring relief device. Blocked-in area. Water is used for each case.7 29. "F Minimum temperature. psia Maximum pressure. ft3 Dust class Set pressure.

1 psig. system to keep the beverage pressurized.- - Problems 423 the steel shell of the tank is suddenly heated to 100°F by the sun. bottle. Determine the most likely scenario contributing to overpressure of the beverage container. The liquefied CO. 9-15. system includes a small bottle of high-pressure liquefied CO. The ambient temperature is 25°C and the ambient pressure is 1atm. a 5-lb bottle of liquefied CO. The tank MAWP is 200 psig. A beverage dispensing system consists of a bottle of beverage. Assume a heat transfer coefficient of 50 Btulhr ft2 "F and that only the top half of the tank is heated. and a regulator to regulate the gas pressure delivery to the beverage. your calculations also show that the flow of nitrogen through the l-in supply pipe will be much greater than this. The relief device will be installed in the C 0 2line where it enters the beverage container. The cone is 5 ft long. If the deflagration index for the vitamin powder is 80 bar mls.2 bar gauge and that the maximum pressure is 0. A spray dryer is used to dry vitamins in powder form. The regulator is connected directly to the CO. what is the required relief area? Assume no overpressure. Must two-phase flow be considered? c. Determine the orifice diameter (in cm) required to achieve this flow. Assume a temperature of 80°F. Assume that the vent opens at 0. The dryer consists of a cylindrical section 12. and a 0. A 10-ft wide by 10-ft long by 10-ft high shed is used to store tanks of methane. You are currently reviewing the scenario involving the failure of a nitrogen regulator that provides inert padding to the vapor space of the reactor. thus limiting the flow to the maximum of 0.0 ft high and 5 ft in diameter.75-gal beverage bottle. A typical beverage system contains a 7. and a regulator set at 9 psig.. determine the area required for a deflagration vent.5 kg/s. The CO.. Also assume a spring-type relief. If the tank is exposed to fire. assuming a set pressure of 20 psig and an overpressure of 10%. saturation vapor pressure is 800 psia. What deflagration vent area is required? Assume a maximum internal overpressure of 0. . 9-14. Your calculations show that the maximum discharge rate of nitrogen through the existing relief system of the vessel is 0. and a CO. You have been assigned the task of reviewing the relief scenarios for a specific chemical reactor in your plant. Assume a nitrogen source supply pressure of 15 bar absolute.5 kgls. Thus under the current configuration a failure of the nitrogen regulator will result in an overpressuring of the reactor. Attached to the bottom of the cylindrical section is a cone section for collecting the dried powder. A pressure relief system must be designed to protect the beverage bottle from overpressure. Determine the vent area required. One way to solve the problem is to install an orifice plate in the nitrogen line. 9-17.5 bar gauge. a dispensing hose with valve.5-in (internal diameter) plastic hose connects the regulator to the beverage bottle. a. However. 9-16. b.

A laboratory test has shown that the reaction will not result in a two-phase relief.4 Btullb. What rupture disc diameter is required to relieve the vessel properly? Assume a discharge pressure of 10 psig. contains water. 10 ft long and 3 ft in diameter. Calculate the mass flux (kg/m2s) of two-phase material through the same leak under the same conditions of part a. Assume a 10% backpressure and a 10% overpressure in the relief system. a. calorimeter tests indicate that the maximum self-heat rate is 40°C/min. Assume that the discharge length is greater than 10 cm. b. a. Determine the maximum vaporization rate during a runaway reaction (in kgls). the heat of vaporization is 92. A relief system must be designed to protect the vessel in the event of fire exposure. A 500-gpm pump is used to provide water to a reactor vessel. If the pump continues to operate. its boiling point is 673"R. and that its molecular weight is 44. 9-20. Assume a MAWP of 7 bar gauge. Assume that the material is stored at 25OC. the reactor might be overfilled and overpressurized. What relief size is required to protect the vessel from fire exposure? Assume the following: vapor relief only.000 kg of reacting liquid material. Determine the diameter of a conventional spring-operated relief required to protect the vessel from a regulator failure. 9-22. 10% backpressure. The reactor system in a pilot plant contains stock tanks that are 24 in in diameter and 36 in high. 9-23. The molecular weight of the liquid is 162. A batch chemical reactor contains 10.5 x lo5 Pa abs). that it is discharged to 1 atm pressure. Assume a temperature of 200°C at the relief conditions. A relief device must be properly sized for a potential runaway reaction. The physical properties of the material are also reported: Heat capacity of liquid: 2. b. The vessel contains a flammable polymer material. Additional physical property data are: . Calculate the mass flux (kg/m2s)of gaseous material through a leak assuming that the material is stored at its vapor pressure within the vessel (9.30.2. Determine the relief diameter (in inches) required to protect the vessel. Determine the relief diameter (in m) required to vent the runaway reaction.424 Chapter 9 Relief Sizing 9-18. and the heat capacity ratio of the vapor is 1. A 1-in (internal diameter) pipe is used to supply water to a low-pressure tank with an MAWP of 20 psig.5 kJ/kg K Heat of vaporization: 300 kJ/kg Molecular weight: 100 Vapor acts as an ideal triatomic gas. 9-21. MAWP of 200 psig. 9-19. Furthermore. The water is supplied through a regulator from a source with a maximum pressure of 100 psig. Thus a vapor relief system must be designed. conventional spring-operated relief. and a conventional spring-operated relief. A horizontal vessel. Be sure to state clearly any assumptions and your justifications for them. Please state clearly any additional assumptions required for your calculation. The MAWP of the vessel is 100 psig.

f. It is essentially an adiabatic calorimeter. In general. If the reaction is first order in concentration. During an exothermic reaction. Compare the results of parts d and e. t is the time (s). is the ideal gas constant (energylmol deg).70 X lo3Jlkg K c. Assume that the energy is due to the sensible heat increase of the two-phase discharge stream and that the temperature of the discharge is 10 K higher. Calculate the energy discharge rate for the discharge of part b.: 0. e. and A is the pre-exponential factor (time-'). with the test sample heated at a constant temperature rate until an exothermic reaction is encountered. where T is the absolute temperature (K). Assume that the energy content of the vapor is due to the heat of vaporization of the liquid to gas. the increase in temperature will also be first order. Comment on the difference in flux rates between parts a and b. the energy of reaction heats the sample and increases its temperature. The RSST (reactive system screening tool) is a laboratory device used to characterize the reactive nature of liquid materials. is the maximum temperature (K).. Show that for a first-order system. 9-24. a. Assume that a sample is being heated in the RSST. if the heat of reaction is approximately constant. Calculate the energy discharge rate for the discharge of part a. R. relief systems designed for two-phase flow must be larger than those for all-vapor flow. Is this consistent with the results of parts a and b? Why? d.Problems 425 Heat of vaporization: 3. E is the activation energy (energy/mol). T. How many times larger must the area of the twophase discharge be in order to remove energy at the same rate as the single-phase relief? Comment on the implications for relief systems on reactor vessels..23 X lo3 J/kg K Heat capacity of vapor: 1.048 m3/kg Heat capacity of liquid: 2. a plot of versus -10001T will produce a straight line of slope (El1000 R.) and an intercept of In A. It is also assumed that the reaction rate follows a typical Arrhenius function as follows: .33 X lo5 Jlkg vf. then.

426 Chapter 9 Relief Sizing b. Is the reaction first order? If so. and its viscosity is 0. The MAWP is 250 psig. The following table provides a set of data collected from an RSST run. determine the activation energy and the pre-exponential factor for the reaction.928. The specific gravity of the oil is 0. Temperature Time (min) c'c) Pressure (psi4 9-25.004 kglm s. . and a maximum backpressure of 50 psig is expected. A relief device must be designed for a vessel to relieve 1800 gpm of crude oil liquid in the event that a discharge line is blocked.

a conventional spring-operated relief. The fundamental burning velocities of the liquid vapors are less than 1.672 ft2. Determine the relief diameter required. 9-26. and the set pressure is 75 psig. What type of relief should be used: a rupture disc. A relief device must be installed on a vessel to protect against an operational upset. . The room is used for dispensing flammable liquids.09. The dryer is a vertical cylinder 2 m in diameter and 3 m high.5 psi. 30 ft wide. b. Determine the vent area required if the maximum pressure that the room can withstand is 1. b.500 lblhr of hydrocarbon vapor. a. A vent must be designed for a room that is 20 ft long. and a heat capacity ratio of 1. 9-30. 9-29. Estimate the vent area required if the dryer is capable of withstanding a pressure of 0.69 psi.Problems 427 a. How does this area compare to the wall area available for the vent? b. The relief must discharge 53. Assume a vent release pressure of 0. Determine the vent area required if the maximum pressure that the room can withstand is 0.05 bar gauge. a compressibility of 0. a. What size nitrogen line would you select if the nitrogen is available from a 500-psig source? The ambient temperature is 80°F. A process vessel is equipped with a 2-in rupture disc set at 100 psig and designed for 10% overpressure. One wall of the room is located against the wall of another structure and is thus not available for a vent. The hydrocarbon vapor has a molecular weight of 65. Estimate the vent area required if the dryer is a low-pressure structure capable of withstanding 0.04 psi. Determine the diameter of the relief. Assume an overpressure of 10% and a backpressure of 0 psig. A spray dryer is used to dry a dust with a Kst value of 230 bar mls. Specify an appropriate set pressure and overpressure for this relief. in inches. The relief temperature is 167"F. and 20 ft high. or a balanced-bellows relief? c.84. Determine the vent area required for this structure if the total internal surface area of the structure (including floor and roof) is 24. 9-27. A nitrogen line must be added to the vessel to provide the capability of purging andlor pressure discharging liquids.15 bar gauge.2 bar gauge. A pole barn with thin metal walls must be fitted with a vent to safely vent a hydrocarbon deflagration from the combustion of a hydrocarbon similar to propane. The maximum pressure that this building can withstand is estimated at 0.3times that of propane. Is adequate wall area now available for the vent? 9-28.

.

the probability of those events. Unfortunately. What can go wrong and how? 3. damage to the environment. This is followed by a concurrent . The various scenarios by which an accident can occur are then determined.Risk assessment includes a determination of the events that can produce an accident. The consequences could include human injury or loss of life. a hazard is not always identified until an accident occurs. For each process in a chemical plant the following questions must be asked: H 1.C H A P T E R 1 0 Hazards Identification azards are everywhere. Risk assessment is sometimes called hazard analysis. and the consequences. The terminology used varies considerably. After a description of the process is available. or loss of production and capital equipment. considered in detail in chapter 11. Hazard identification and risk assessment are sometimes combined into a general category called hazard evaluation. What are the chances? 4. whereas a procedure that determines probability and consequences is called quantitative risk analysis (QRA). the hazards are identified. Figure 10-1 illustrates the normal procedure for using hazards identification and risk assessment. Question 2 is frequently called scenario identification. What are the hazards? 2. What are the consequences? The first question represents hazard identification. A risk assessment procedure that determines probabilities is frequently called probabilistic risk assessment (PRA). The last three questions are associated with risk assessment. It is essential to identify the hazards and reduce the risk well in advance of an accident.

then the system must be modified and the procedure is restarted.430 Chapter 10 Hazards Identification System description G Hazard identification I Scenario identification 1 J. If the risk is acceptable. Risk determination I hazard acceptance Modify: Process or plant Process operation Emergency response Other I Yes Build and/or operate Figure 10-1 Hazards identification and risk assessment procedure. . 1985). 1-9. This information is assembled into a final risk assessment. Accident consequence I J. Adapted from Guidelines for Hazards Evaluation Procedures (New York: American Institute of Chemical Engineers. Accident probability J. then the study is complete and the process is operated. If the risk is unacceptable. pp. study of both the probability and the consequences of an accident.

One outcome is that hazards of low probability and minimal consequences are identified and addressed with the result that the process is "gold-plated. However. Process hazards checklists: This is a list of items and possible problems in the process that must be checked. then risk assessment procedures cannot be fully applied." This means that potentially unnecessary and expensive safety equipment and procedures are implemented. Each organization using these procedures must have suitable criteria.l Only a few of the more popular approaches are considered here. If failure rate data on the applicable equipment are not available. An important part of the hazard identification procedure shown in Figure 10-1 is the risk acceptance step. and what should be done about them? For most facilities the probability of these hazards is small: No steps are required for prevention. 1992). much like an income tax form. 4. Various events are suggested for a specific piece of equipment with the participants determining whether and how the event could occur and whether the event creates any form of risk. that provide penalties for hazards and credits for safety equipment and procedures. This enables modifications to be easily incorporated into the final design. If the study is performed with the initial design. Guidelines for Hazard Evaluation Procedures. 3. Hazards and operability (HAZOP) studies: This approach allows the mind to go free in a controlled environment. Safety review: An effective but less formal type of HAZOP study. The selection of the best method requires experience. 2d ed. What are the chances of their occurrence. it should be done as soon as possible. Hazard identification can be performed independent of risk assessment. the best result is obtained if they are done together. Most plant sites (and even subunits within a plant) modify the procedure to fit their particular situation. flying aircraft and tornadoes are hazards to a chemical plant. Most companies use these methods or adaptations to suit their particular operation. Hazards surveys: This can be as simple as an inventory of hazardous materials. The Dow indexes are a formal rating system. (New York: American Institute of Chemical Engineers. . For instance.Chapter 10 Hazards Identification 431 The procedure described by Figure 10-1is frequently abbreviated based on circumstances. hazards with reasonable probability but minimal consequences are sometimes also neglected. Hazards identification and risk assessment studies can be performed at any stage during the initial dcsign or ongoing operation of a process. or it can be as detailed as the Dow indexes. No single approach is necessarily best suited for any particular application. The hazard identification methods described in this chapter include the following: 1. 2. Likewise. Many methods are available for performing hazard identification and risk assessment. The results are highly dependent on the experience and synergism of the group reviewing the process.

A checklist can be used during the design of a process to identify design hazards. This checklist might contain the following items: Check oil in engine. Check gasoline level in tank. A checklist intended for use during the initial design of the process will be considerably different from a checklist used for a process change. Check headlights and taillights. Note that three checkoff columns are provided. 1994). the effort expended in developing and using checklists can yield significant results. plant layout. Check fluid levels in brake system. Some companies have checklists for specific pieces of equipment. such as a heat exchanger or a distillation column. or it can be used before process operation. and so forth. Extensive notes on individual areas are kept separate from the checklist.432 Chapter 10 Hazards Identification 10-1 Process Hazards Checklists A process hazards checklist is simply a list of possible problems and areas to be checked. Checklists should be applied only during the preliminary stages of hazard identification and should not be used as a replacement for a more complete hazard identification procedure. electrical systems. Check fluid level in radiator. (New York: American Institute of Chemical Engineers. 7th ed. The second column is used for those items that do not apply to the particular process. A typical process design safety checklist is shown in Figure 10-2. Check air filter. as illustrated in the vacation example. Checklists are most effective in identifying hazards arising from process design. . But. Check air pressure in tires. The first column is used to indicate those areas that have been thoroughly investigated. 10-2 Hazards Surveys A hazards survey can be as simple as an inventory of hazardous materials in a facility or as com)~ plicated as a rigorous procedure such as the Dow Fire and Explosion Index ( F ~ L E Iand the 2Dow's Fire and Explosion Index Hazard ClassiJicationCode. The last column is used to mark those areas requiring further investigation. The list reminds the reviewer or operator of the potential problem areas. storage of chemicals. involving hundreds or even thousands of items. The design of the checklist depends on the intent. A classic example is an automobile checklist that one might review before driving away on a vacation. Check fluid level in windshield washer tank. Check exhaust system for leaks. Checklists for chemical processes can be detailed.

Hazardous overhead restrictions? 6. Adapted from Henry E. Special fume or dust hoods required? 3. Unstable materials properly stored? 4. Aisleways provided? 3. Chemistry of processes completely understood and reviewed? 8.Adequate platforms for safe maintenance operations? 11.10-2 Hazards Surveys 433 Further study required 1 Does not apply L Completed 1 General layout 1.Clearance for overhead power lines? Buildings 1. Webb. Access for emergency vehicles? 9. stairways and escapeways? 2. Consequences of exposure to. Hazardous underground obstructions? 5. Process laboratory checked for runaway explosive conditions? 5. Enough headroom? 8. ed. "What To Do When Disaster Strikes. . Provisions for rapid disposal of reactants in an emergency? 0 9. Failure of mechanical equipment possible cause of hazards? n o Figure 10-2 A typical process safety checklist. Emergency accesses and exits? 7. Adequate ladders. Head obstructions marked? 4. A list of this type is frequently used before a more complete analysis. Need for fireproofed structural steel? 0 ' o 0 D 0 o 0 D 0 Process 1. Fire doors required? 3. Provisions for protection from explosions? 6.Hoists and elevators properly designed and safeguarded? 12. Areas properly drained? 2. Safe storage space for raw materials and finished products? 10. Safety glass specified where necessary? 7. Need for ladder or stairway to roof? 6.adjacent operations considered? 2." in Safe and Efficient Plant Operation and Maintenance. Fire walls. Richard Greene. 1980). (New York: McGrawHill. Ventilation adequate? 5. dikes and special guardrails needed? 4. Hazardous reactions possible due to mistakes or contamination? 7.

fumes.Process diagrams correct and up-to-date? Piping 1.City water lines not connected to process pipes? 17.Emergency valves readily accessible? 12. Vent lines directed safely? 6. All overflow lines directed to safe areas? 5.Steam condensate piping safely designed? 14.Hazards possible from simultaneous loss of two or more utilities? 16.Public liability risks possible from sprays.Personnel protective insulation provided? 19.434 Chapter 10 Hazards Identification Further study required Does not apply L Completed . Corrosion allowance considered? D 1 D 0 13 !J 0 0 o D D D D 0 0 0 0 0 0 0 0 D 0 0 D Figure 10-2 (continued) . reviewed? 18.Relief valve piping designed to prevent plugging? 15.Hazards possible from gradual or sudden blockages in piping or equipment? 11.Drains to relieve pressure on suction and discharge of all process pumps? 16. Washing-down hoses needed? 8. mists or noise? 12.Long and large vent lines supported? 13.Safety factors altered by design revisions? 17. Check valves provided as needed? 9. Sprinkler systems required? 3.Possible deterioration of exterior of piping by chemicals? 11.Hazards involved in sewering material? 14.Provisions made for disposal of toxic materials? 13. Piping specifications followed? 7. Provisions for thermal expansion? 4.Material safety data sheets available for all chemical species? 15.Flammable fluids feeding production units shut off from a safe distance in case of fire or other emergency? 18. Designs correct for maximum operating pressure? 2.1 10.Consequences of reasonably worst incident. Safety showers and eye baths required? 2. Protection and identification of fragile pipe considered? 10.Hot steam lines insulated? Equipment 1. or combination of incidents.

Schedule for checking protective devices? 6. Relief valves protected from plugging by rupture disks? 6. Relief valves or rupture disks required? 2.Sufficient lighting for both outside and inside operations? 1l. Guards for belts. Telltale pressure gauges installed between rupture disks and relief valve? Instrument and Electrical 1.All equipment grounded? o 1 o o 0 0 o o o o 0 o o Figure 10-2 (continued) . configuration?) 4. Dikes for any storage tanks? 7. sheaves and gears? 0 5. Construction materials compatible with process chemicals? 9. direction.10-2 Hazards Surveys 435 Further study required Does not apply 1 Completed I 3.Lights provided for all sight glasses. Materials of construction corrosion resistant? 3. Guard rails for storage tanks? 8. Process safety affected by response lag? 7.Pipelines independently supported to relieve pumps and other equipment. Special isolation for hazardous equipment? 4.Emergency standby equipment needed? Venting 1. showers and eyebaths? 12. Flame arrestors required on vent lines? 5. Reclaimed and replacement equipment checked structurally and for process pressures? 10. Labels for all start-stop switches? 8. Equipment designed to permit lockout protection? 9. Dual indication of process variables necessary? 3. Tubing runs protected? 5. Vents properly designed? (Size. as necessary? 11. Safeguards provided for process control when an instrument must be taken out of service? 6.Breakers adequated for circuit protection? 13.Automatic lubrication of critical machinery? 12. Electrical failures cause unsafe conditions? 10. All controls fail safe? 2. pulleys. All equipment properly labelled? 4.

flammability. Any products hazardous from a toxic or fire standpoint? 4.All necessary communications equipment provided? 18. Colorimetric indicator tubes required? 5. Flammable vapor detection apparatus required? 6. etc? 6.436 Chapter 10 Hazards Identification Further study required Does not apply 1 Completed 1 14. Special respiratory equipment required? 3. stability.Emergency disconnect switches properly marked? 19. Any materials and products require special o handling equipment? 2. Any raw materials and products affected by extreme weather conditions? 3. Does warehouse have operating instructions covering each product regarded as critical? o Figure 10-2 (continued) .Emergency escape lighting required during power failure? 17. Proper containers being used? 5. Fire extinguishers required? 2.Special explosion proof electrical fixtures required? Safety Equipment 1. Consequences of bad spills considered? 7. Special instructions needed for containers or for storage and warehousing by distributors? 8.Emergency standby power on lighting equipment required? 16. Special emergency procedures and alarms required? 1 0 0 0 0 o o 0 0 Raw Materials 1. Containers properly labelled for toxicity. Diking material required? 4. Fire extinguishing materials compatible with process materials? 7.Special interlocks needed for safe operation? 15.

handling. the more flammable and/or explosive the material. The first column is the penalty column. shown in Figure 10-3. The next step is to determine the material factor (MF) for use in the form shown in Figure 10-3. The procedure begins with a material factor that is a function only of the type of chemical or chemicals used. for determining the relative magnitude of flammable hazards in a chemical plant. The F&EI also contains a mechanism for estimating the dollar loss in the event of an accident. A procedure is provided in the complete index for computing the material factor for other compounds not listed in Table 10-1 or provided in the Dow reference. This factor is adjusted for general and special process hazards. The final column is used for computation. These adjustments or penalties are based on conditions such as storage above the flash or boiling point. These are formal systematized approaches using a rating form. A process unit is a single pump. The usual approach is to select only the units that experience shows to have the highest likelihood of a hazard. This allows for a reduction or increase in the penalty based on extenuating circumstances not completely covered by the form.10-2 Hazards Surveys 437 Dow-Chemical Exposure Index (CEI)3. the higher the value of the MF. The main idea of this procedure is to provide a purely systematic approach.which are two popular forms of hazards survey. A process safety checklist or hazards survey is frequently used to select the most hazardous units for further analysis. 'Dow's Chemical Exposure Index Guide. It is not practical to apply the fire and explosion index to all these units. The final rating number provides a relative ranking of the hazard. The main forms used for the computations are shown in Figures 10-3 and 10-4. The highest value of the MF under the complete range of operating conditions is suggested. the complete penalty value from the first column is used. If mixtures of materials are used. and fired heaters. .Table 10-1lists MFs for a number of important compounds. consists of three columns of numbers.or exothermic reactions. The first step in the procedure is to conceptually divide the process into separate process units. In the event of uncertainty here. A large process results in hundreds of individual units. The Dow F&EI is designed for rating the relative hazards with the storage. This list also includes data on heat of combustion and flash and boiling point temperatures. 1994). endo. similar to an income tax form. The additional data are also used in the computation of the Dow F&EI. after the fire and explosion index has been determined. Credits for various safety systems and procedures are used for estimating the consequences of the hazard. or a storage tank. mostly independent of judgmental factors. The resulting MF value for the process is written in the space provided at the top of the form in Figure 10-3. Penalties for various unsafe situations are placed in this column. the MF is determined from the properties of the mixture. a reactor. (New York: American Institute of Chemical Engineers. The second column contains the penalty actually used. The form. In general. and processing of explosive and flammable materials. 4th ed.

30 - 0.50 Special Process Hazards Factor (F2) Process Unit Hazards Factor (F1 x F2) .20 to 0..nl) 1 L o n Pmv*ntlon) BY: ( .) YI R W W D BY: (Tmchdwy Cmtmr) E E E I REVIEWED BY: ( ........... Process upsetor Purge Failure 3. Pressure (See Figure 2) Operating Pressure psig or kPa gauge Retief Selling psig or kPa gauge F. Corrosion and Erosion I............... Reproduced by permission of the American Institute of Chemical Engineers..00 0.... Low Temperature G...START UP .......... Leakage Joints and Packing J.. Liquids or Gases in Process (See Figure 3) 2.. Combustible Solids in Storage. Source: Dow's Fire and Explosion Index Hazard Classification Guide.. Dust Explosion (See Table 3) E....15 0................nd. I 5by MATERIALSIN PROQSS UNlT STATE OF OPERAllON BASK: MATERIAL(S)FOR MATERIAL FACTOR I I - - DESION .... Figure 10-3 Form used in the Dow Fire and Explosion Index... .25 to 2.......438 Chapter 10 Hazards Identification FIRE & EXPLOSION INDEX AREA I O N R CUTY DIVISION I LOCATION I ME I Sm YANVACTURINO UNIT APPROVW P O E S UNrr R CS BUIWNO I PREPARED BY: REVIEWD I BY: (s. (1994)...10 to 0. ....SHUTDOWN MATERIALF C O (See Table 1 w Appendices A w B) Note requiremernswhen tmit temperature over 140 O F (60 O C J ATR - 2.... ..... 7th ed.... ........ Hot Oil Heat Exchange System (See Table 5) L...... Dust in Process (See Figure 5) H... Always in Flammable Range D..1510 1..75 0.... Liquids or Gases in Storage (See Figure 4) 3.. Rotating Equipment 0.80 0..50 0... = FQ..... .......... The figures and tables referenced in the form are provided in the index booklet. Quanliiy of FlammableNnstable Material: Quantity ....... ...NDRMAL DPERATIDN .10to 1.... Use of Fired Equipment (See Figure 6) K.........or kg ib Hc = B T U I b or kcallkg 1...win(................30 0..

....... Refer to Fire 8 Explosion Indew Hazard Classification Guide for details........................................(Actual MPPD) [6 x 71 ............ Value of Area of Exposure...(MPDO) ... .... (2) For no credit factor enter 1.............. Reproduced by permission of the American Institute of Chemical Engineers.... ... fl or m ft2 or m2 Area of Exposure.. ..............96 to 0..... 7th ed...........................................00......... Business Interruption ...(81) .97 0...... Source: Dowk Fire and Explosion lndex Hazard Classification Guide........96 Credit Factor Used(q c... Figure 10-4 Form used for consequences analysis.. Dump/Blowdown Credit Factor Range 0.(Base MPPD) [4 x 51 .......... Radius of Exposure ...98 Feature a................. Process Control Credit Factor (C1) c1Vaiue(3) 2.. Maximum Probable Days Outage ............................( 3.............. (3) $MM Product of all factors used............................................ Fire 8 Explosion Index (F8EI). $MM 5..... 9. Fire ProtectionCredit Factor (C3) n c3Value(3) 1 Loss Control Credit Factor = C1 X C2 X C3(3)= 7) on line below) (Enter 7 .. I I I Actual Maximum Probable Property Damage .....96 to 0...91 to 0........... (Figure 9) days $MM 10... 1994.......................................... Remote Control Valves b.................. 1 6.. Base Maximum Probable Property Damage ............. (See Front) ... $MM 7...........98 0............... PROCESS UNIT RISK ANALYSIS SUMMARY 1....... (Figure 8) ........ Loss Control Credit Factor...... ...... (See Above) 8.... Figure 7) 2..... ............... .. 4... .10-2 Hazards Surveys 439 LOSS CONTROL CREDIT FACTORS 1. ... ................................ Interlock ~2 Credit Factor Use*L ~alue(3) 3................... Drainage d................. Material Isolation Credit Factor (C2) 1 7 Feature Credit Factor Range 0................................... Damage Factor .... ............ .............. ..........

.3 0.7 17..0 18.440 Chapter 10 Hazards Identification Table 10-1 - Selected Data for the Dow Fire and Explosion Index1 .6 21.162 88 40 .8 18. 3.4 8. The next step is to determine the general process hazards. exothermic reactions that might self-heat.3 20.8 51.4 17.400 -423 -258 147 680 258 97 - -54 293 232 7 279 'Selected from Dow's Fire and Explosion Index Hazard Classification Guide.1 4.155 304-574 100.7 18.5 19.3 19.8 20.7 18.5 8.6 17.108 77 -313 -29 179 322 315 -128 .7 15.- Compound Acetone Acetylene Benzene Bromine 1. 2.0 18. material handling and transfer.4 20.0 19.2 19.- .3-Butadiene Butane Calcium carbide Carbon monoxide Chlorine Cyclohexane Cyclohexanol Diesel fuel Ethane Ethylene Fuel oil #1 Fuel oil #6 Gasoline Hydrogen Methane Methanol Mineral oil Nitroglycerine Octane Pentane Petroleum (crude) Propylene Styrene Toluene Vinyl chloride Xylene Material factor 16 29 16 Heat of combustion (Btullb x 12.6 Flash point (OF) -4 Gas 12 - Boiling point (OF) 133 -118 176 - 1 24 21 24 21 1 16 10 10 21 24 10 10 16 21 21 16 4 40 16 21 16 21 24 16 24 16 -105 Gas - 24 31 - .4 21.3 0. Penalties are applied for the following factors: 1.0 7. enclosed process units preventing dispersion of escaped vapors.0 17. 7th ed. including pumping and connection of transfer lines. 4.7 17.7 9. (New York: American Institute of Chemical Engineers. Gas Gas -4 154 100-130 Gas Gas 100-162 100-270 -45 Gas Gas 52 380 56 <-40 20-90 . endothermic reactions that could react because of an external heat source such as a fire. 1994.7 20.

dust explosion risks. large rotating equipment. leakage around joints and packings. 7.) are multiplied together to produce a unit hazard factor (F. The Dow F&EI is computed by multiplying the unit hazard factor by the MF. 2. which could impede fire fighting. 6. 10. limited access for emergency equipment. 3. use of fired heaters. less than atmospheric pressure operation with a risk of outside air entering. low-temperature operation with potential embrittlement of carbon steel vessels.). including pumps and compressors. 5. and 6. corrosion and erosion of process unit structures. This includes the maximum probable property damage (MPPD) and the maximum probable days outage (MPDO). quantity of flammable material. providing a ready ignition source. Table 10-2 Determining the Degree of Hazard from the Dow Fire and Explosion lndex Dow Fire and Explosion Index 1-60 61-96 97-127 128-158 159 and above Degree of hazard Light Moderate Intermediate Heavy Severe . and 12. higher than atmospheric pressure. 9. 8. 11. Penalties for special process hazards are determined next: 1. Table 10-2 provides the degree of hazard based on the index value. The general process hazard factor (F1) and special process hazard factor (F. 4. The computations are completed in the Risk Analysis Summary table at the bottom of the form. The consequences analysis is completed using the worksheet form shown in Figure 10-4.10-2 Hazards Surveys 441 5. hot oil heat exchange systems where the hot oil is above its ignition temperature. Detailed instructions and correlations for determining the general and special process hazards are provided in the complete Dow F&EI. operation in or near the flammable limits. The Dow F&EI can be used to determine the consequences of an accident. toxic materials. poor drainage of flammable materials away from the process unit.

442

Chapter 10

Hazards Identification

The damage radius is first estimated using a correlation published in the complete Dow index. This correlation is based on the previously determined F&EI. The dollar value of the equipment within this radius is determined. Next. a damage factor (based on a correlation provided) is applied to the fraction of the equipment actually damaged by the explosion or fire. Finally, a credit factor is applied based on safety systems. The final number, in dollars, is the MPPD value. This number is used to estimate the MPDO using a correlation. Details on the procedure are available in the complete Dow reference. The Dow indexes are useful for determining equipment spacing requirements. The F&EI uses an empirical correlation based entirely on the F&EI value to estimate the radius of exposure. It is assumed that any equipment located outside this distance would not be damaged by a fire or explosion. The CEI estimates the hazard distance for chemical exposure based on the emergency response planning guideline (ERPG) values for the particular material released.

Example 10-1
Your plant is considering the installation of a new railcar tank unloading facility. The facility will unload nominal 25,000-gal tank cars containing either pure butadiene or cyclohexane. The unloading system will be equipped with an emergency shutdown system with remotely operated block valves. The unloading operation will be done by computer control. The railcars are inerted with nitrogen to a pressure of 40 psig, and the railcar relief system has a set pressure of 75 psig. The unloading operating instructions are written and have been reviewed by the corporate technical staff. A reactive chemicals review has already been completed on the proposed facility. Combustible gas detectors will be located at the unloading station. A deluge system will be installed at the unloading site with an excellent water supply. A diking system will surround three sides of the facility, with any spills directed to a covered impounding area. Determine the Dow F&EI for this operation, and determine the minimum spacing from adjacent units.

Solution
The Dow Index contains most of the data required to complete the evaluation. The data for the chemical species used in this facility are:

Species
Butadiene Cyclohexane

NFPA Material factor health rating
24 16 2 1

Heat of combustion (Btullb)
19.2 x lo3 18.7 X lo3

Flash point
(OF)

-105 -4

Because the butadiene has the highest MF, it is the material we need to evaluate using the Dow F&EI. The completed F&EI form is shown in Figure 10-5. Each nonzero item on the form is discussed in what follows.

10-2 Hazar ds S u r v e y s

FIRE & EXPLOSION INDEX
-1

C N Y OU TR

OMSION

LOCATION

DATE

North America
SITE

North Central
MANUFACTURING UNIT

Arkansas
P O E S UNIT R CS

03/04/94
l eUliDlNC

No Loss
PEAE B: RPRD Y

Dow Polymer Alvin Doe

I APPROVED BY: (Superintendent)

Rail Car Unloading A-103
REVIEWED BY: (Safety L Loss Prevention)

John Smith
RMEWW

BY:(Y.nagement)

REVIEWEDBY:(TechnologyCenter)

Robert Big
MATEfUALS IN P OC S UNl R ES T

Bill Wright
BASIC MATER!AL(S)FOR MATERIAL FACTOR

Butadiene, Cyclohexane
S A E O OPERATION TT F

- DESIGN

- ST~VITw

-X-

NORMAL OPEFATION

- SHUTDOWN

I

Butadiene
I

MATERIAL FACTOR(See Table 1 or Appendices A or B) Note requirements

I General Process Hazards .

Penalty Fact o r Ranoe

Penalty Factor Usedri~

~~

- -

D.

Dust Explosion (See Table 3) Pressure (See Figure 2) Operating Pressure Relief Setting 40 75 psigpsig-

E.

I I

0.25 to 2.00

I

.28

F. Low Temperature G. Quantity of Flammable/UnstableMaterial:
1.

0.20 to 0.30 Quantity -130K- Ib-wkg -19.2K-BTUIlb or kcallkg

.76

Hc = Liquids or Gases in Process (See Figure 3) 2 Liquids or Gases in Storage (See Figure 4) 3. Combustible Solids in Storage, Dust in Process (See Figure 5) H. Corrosion and Erosion I. Leakage Joints and Packing J. Use of Fired Equipment (See Figure 6) K Hot Oil Heat Exchange System (See Table 5) L Rotating Equipment

-

-

0.10 to 0.75 O.1Oto 1.50 0.15to 1.15 0.50

. I .1

-

-

Special Process Hazards F a cto r (Fz) Process Unit Hazards Factor (FI

............................................................................................
=
F3

2.94 4.41 106.00

x Fp)

.............................................................................
..............................................................................

F i r e a n d Explosion Index (F3 x MF
FORU G D O RNlDi-04

=

FBEI)

(I) For no penalty use 0.00.

Figure 10-5 The Dow Fire and Explosion Index applied to the railcar unloading facility of Example 10-1.

-

-

-

444

Chapter 10

Hazards Identification

1.A. Exothermic chemical reactions: The reactive chemical review has determined that an exothermic chemical reaction here is not possible. The penalty is zero. 1.B. Endothermic chemical reactions: This penalty applies only to reactors, so the penalty is zero. 1.C. Material handling and transfer: The index documentation states: "Any loading and unloading operation involving Class I flammables or LPG-type materials where transfer lines are connected and disconnected receives a penalty of 0.50." 1.D. Enclosed or indoor process units: The unit is outdoors, so the penalty is zero. 1.E. Access: The unit will have emergency access from all sides, so the penalty is zero. 1.F. Drainage and spill control: No penalty is applied because the dike and impounding system is present. 2.A. Toxic materials: The index suggests using a penalty value of 0.20 X NFPA Health Rating. Because the rating is 2, the penalty value is 0.4. 2.B. Subatmospheric pressure: The operation is pressurized, so no penalty is applied here. 2.C. Operation in or near flammable range 1. Tank farms storage flammable liquids: The tanks are inerted with a closed vapor recovery system, so the penalty here is zero. 2. Process upset or purge failure: The unit relies on inert purging to keep it out of the flammable range, so a penalty of 0.30 is applied. 3. Always in flammable range: The process is not in the flammable range during normal operation, so the penalty is zero. 2.D. Dust explosion: No dusts are involved, so the penalty is zero. 2.E. Pressure: The Dow index provides a detailed procedure for determining this penalty. The operating pressure penalty is determined from Figure 2 in the Dow index booklet using the operating pressure. In this case the operating pressure of 40 psig results in a penalty of 0.24. Second, a penalty is determined at the relief set pressure (75 psig), again using Figure 2 in the Dow index booklet. This value is 0.27. The operating pressure penalty is then divided by the set pressure penalty to get a final pressure penalty adjustment. In this case the adjustment is 0.2410.27 = 0.8889. This is multiplied by the operating pressure penalty to obtain 0.24(0.8889) = 0.2133. Finally, this is multiplied by a correction factor of 1.3 because this is a liquefied flammable gas. The final penalty is 0.2133(1.3) = 0.28. 2.F. Low temperature: Low-temperature operating is not expected, so the penalty is zero. 2.6. Quantity of flammablelunstable material 1. Liquids or gases in process: This is not part of the process, so the penalty is zero. 2. Liquids or gases in storage: The total energy contained within the storage inventory is estimated in order to determine the penalty. This requires the specific gravity of butadiene, which can be found on the MSDS sheet or other reference. This value is 0.6263. Thus the total energy is (25,000 ga1)(8.345 lb/ga1)(0.6263) = 130,662 lb, (1.30
X

10" lb)(19.2

X

lo3 Btullb)

=

2.51

X

lo9 Btu.

From Figure 4, curve B, in the Dow index booklet, the penalty is 0.76. 3. Combustible solids in storage, dust in process: No solids are present here, so the penalty is zero.

10-2 Hazards Surveys

445

2.H. Corrosion and erosion: Corrosion and erosion is expected to be less than 0.5 millyr. Thus the penalty is 0.10. 2.1. Leakage -joints and packing: The pump and gland seals are expected to have some small but minor leakage. Thus the penalty here is 0.10. 2.5. Use of fired equipment: No fired equipment is present, so the penalty is zero. 2.K. Hot oil heat exchange system: Not present, so the penalty is zero. 2.L. Rotating equipment: No large rotating equipment is present, so the penalty is zero.
These penalties and factors are summarized in Figure 10-5. The resulting calculation shows an F&EI value of 106, which means that this unloading station is an intermediate hazard. Figure 7 in the Dow index booklet provides the radius of exposure based on the F&EI value. For this case the radius is 90 ft. Thus the unloading station must be located a minimum of 90 ft from any other equipment or processes.

The Dow CEI is a simple method of rating the relative acute health hazard potential for people in neighboring plants or communities arising from possible chemical release incidents. To use the CEI, the following items are required: an accurate plot plan of the plant and the surrounding area, a simplified process flow sheet showing the containment vessels, major piping, and chemical inventories, physical and chemical properties of the materials investigated, ERPG values, from Table 5-7, the CEI guide, and the CEI form shown in Figure 10-6.
A flowchart of the CEI procedure is shown in Figure 10-7. The procedure begins with a definition of possible release incidents. This includes releases from pipes, hoses, pressure relief devices relieving directly to the atmosphere, vessels, and tank overflows and spills. The CEI guide has detailed guidelines for these incidents, as shown in Table 4-5. The incidents are used with a number of simplified source models provided in the Dow guide4 to estimate the release rate of material. The ERPGs are then used with a simplified dispersion model to determine the CEI value and downwind hazard distances resulting from the release. Hazards surveys are suitable for identifying hazards associated with equipment design, layout, material storage, and so forth. They are not suitable for identifying hazards resulting from improper operation or upset conditions. On the other hand, this approach is fairly rigorous, requires little experience, is easy to apply, and provides a quick result.

4Dow's Chemical Exposure Index Guide.

CHEMICAL EXPOSURE INDEX SUMMARY
Plant Chemical Largest Single Containment Pressure Of Containment
1. Scenario Being Evaluated

Location Total Quantity In Plant

Temperature Of Containment

2. Airborne Release Rate from Scenario

kg/sec lblmin

3. Chemical Exposure Index
Concentration PPM
Hazard Distance meters feet

mghn3

ERPG- l/EEF'G- 1 ERPG-2/EEF'G-2 ERPG-3/EEPG-3

---meters
feet

5. Distances to:
Public (generally considered property line) Other in-company facility Non-company plant or business 6. The CEI and the Hazard Distance establish the level of review needed. 7. If further review is required, complete Containment and Mitigation Checklist (Chemical Exposure Index Guide, 2nd Edition -Appendix 2, page 26) and prepare Review Package.
8. List any sights, odors or sounds that might come from your facility and cause public concern or inquiries (e.g., smoke,large relief valves, odors below hazardous levels such as mercaptans or amines, etc.)

Prepared by: Reviewed by:
Date

Plant Superintendentor Manager Site Review Representative Additional Management Review (if required)

Figure 10-6 Form used for the Dow Chemical Exposure Index. Source: Dow's Chemical Exposure Index Guide (New York: American lnstitute of Chemical Engineers, 1994). Reproduced by permission of the American lnstitute of Chemical Engineers.

10-2 Hazards Surveys

447

i

Define possible chemical incidents

Determine the Airborne Quantity (AQ) for each scenario

Select the scenario with the largest Airborne Quantity (AQ)

Obtain ERPG-2 or ERPG3 values

I

Calculate the CEI

I

Calculate the hazard distance

Complete CEI summary sheets

Figure 10-7 Procedure for calculating the Chemical Exposure lndex (CEI). Source: Dow's Chemical Exposure lndex Guide (New York: American Institute of Chemical Engineers, 1994).

448

Chapter 10

Hazards Identification

10-3 Hazards and Operability Studies
The HAZOP study is a formal procedure to identify hazards in a chemical process fa~ility.~ The procedure is effective in identifying hazards and is well accepted by the chemical industry. The basic idea is to let the mind go free in a controlled fashion in order to consider all the possible ways that process and operational failures can occur. Before the HAZOP study is started, detailed information on the process must be available. This includes up-to-date process flow diagrams (PFDs), process and instrumentation diagrams (P&IDs), detailed equipment specifications, materials of construction, and mass and energy balances. The full HAZOP study requires a committee composed of a cross-section of experienced plant, laboratory, technical, and safety professionals. One individual must be a trained HAZOP leader and serves as the committee chair. This person leads the discussion and must be experienced with the HAZOP procedure and the chemical process under review. One individual must also be assigned the task of recording the results, although a number of vendors provide software to perform this function on a personal computer. The committee meets on a regular basis for a few hours each time. The meeting duration must be short enough to ensure continuing interest and input from all committee members. A large process might take several months of biweekly meetings to complete the HAZOP study. Obviously, a complete HAZOP study requires a large investment in time and effort, but the value of the result is well worth the effort. The HAZOP procedure uses the following steps to complete an analysis:

'

1. Begin with a detailed flow sheet. Break the flow sheet into a number of process units. Thus the reactor area might be one unit, and the storage tank another. Select a unit for study. 2. Choose a study node (vessel, line, operating instruction). 3. Describe the design intent of the study node. For example, vessel V-1 is designed to store the benzene feedstock and provide it on demand to the reactor. 4. Pick a process parameter: flow, level, temperature, pressure, concentration, pH, viscosity, state (solid, liquid, or gas), agitation, volume, reaction, sample, component, start, stop, stability, power, inert. 5. Apply a guide word to the process parameter to suggest possible deviations. A list of guide words is shown in Table 10-3. Some of the guide word process parameter combinations are meaningless, as shown in Tables 10-4 and 10-5 for process lines and vessels. 6. If the deviation is applicable, determine possible causes and note any protective systems. 7. Evaluate the consequences of the deviation (if any). 8. Recommend action (what? by whom? by when?) 9. Record all information.
SGuidelinesfor Hazard Evaluation Procedures, 2d ed. (New York: American Institute of Chemical Engineers, 1992).

-

-

-

-

10-3 Hazards and Operability Studies

Table 10-3
Guide words
NO, NOT, NONE

Guide Words Used for the HAZOP Procedure
Meaning
The complete negation of the intention Quantitative increase Quantitative decrease Qualitative increase

Comments
No part of the design intention is achieved, but nothing else happens. Applies to quantities such as flow rate and temperature and to activities such as heating and reaction. Applies to quantities such as flow rate and temperature and to activities such as heating and reaction. All the design and operating intentions are achieved along with some additional activity, such as contamination of process streams. Only some of the design intentions are achieved, some are not. Most applicable to activities such as flow or chemical reaction. Also applicable to substances, for example, poison instead of antidote. No part of the original intention is achieved - the original intention is replaced by something else. Applies to process steps or actions. Applies to process steps or actions. Applies to process locations, or locations in operating procedures.

MORE, HIGHER,

GREATER LESS, LOWER

AS WELL AS

PART O F

Qualitative decrease The logical opposite of

REVERSE

OTHER THAN

Complete substitution Too early or in the wrong order Too late or in the wrong order In additional locations

SOONER THAN

LATER THAN

WHERE ELSE

Table 10-4 Valid Guide Word and Process Parameter Combinations for Process Lines (x's represent valid combinations)
Process parameters
Flow Temperature Pressure Concentration pH Viscosity State

NO, More, not, higher, Less, none greater lower x x x x x x x x
X

As well as x x x

Part Other Sooner, Later, Where of Reverse than faster slower else x x x

x
x x x
X

x
x x x
X

x

x
X

x

x

x

x x

x

x

x

x

450

Chapter 10

Hazards Identification

Table 10-5 Valid Guide Word and Process Parameter Combinations for Process Vessels (x's represent valid combinations)
Process Parameters NO, More, not, higher, Less, none greater lower As well as Part Other Sooner, Later, Where of Reverse than faster slower else

Level Temperature Pressure Concentration pH Viscosity Agitation Volume Reaction State Sample

10. Repeat steps 5 through 9 until all applicable guide words have been applied to the chosen process parameter. 11. Repeat steps 4 through 10 until all applicable process parameters have been considered for the given study node. 12. Repeat steps 2 through 11 until all study nodes have been considered for the given section and proceed to the next section on the flow sheet.
The guide words AS WELL AS, PART OF, and OTHER THAN can sometimes be conceptually difficult to apply. As WELL AS means that something else happens in addition to the intended design intention. This could be boiling of a liquid, transfer of some additional component, or the transfer of some fluid somewhere else than expected. PARTOF means that one of the components is missing or the stream is being preferentially pumped to only part of the process. OTHER THAN applies to situations in which a material is substituted for the expected material, is transferred somewhere else, or the material solidifies and cannot be transported. The guide words SOONER THAN, LATER THAN, and WHERE ELSE are applicable to batch processing. An important part of the HAZOP procedure is the organization required to record and use the results. There are many methods to accomplish this and most companies custon~ize their approach to fit their particular way of doing things. Table 10-6 presents one type of basic HAZOP form. The first column, denoted "Item," is used to provide a unique identifier for each case considered. The numbering system used is a number-letter combination. Thus the designation "1 A" would designate the first study node and the first guide word. The second column lists the study node considered. The third column lists the process parameter, and the fourth column lists the deviations or guide words. The next three columns are the most important results of the analysis. The first column lists the possible

Table 1 0 - 6

HAZOP Form for Recording Data

Hazards and Operability Review

Chapter 10

Hazards Identification

Reactor
Monomer feed Cooling coils

\

Thermocouple

Figure 10-8

An exothermic reaction controlled by cooling water

causes. These causes a r e determined by t h e committee and a r e based on t h e specific deviation-guide word combination. T h e next column lists t h e possible consequences of t h e deviation. T h e last column lists t h e action required t o prevent t h e hazard from resulting in an accident. Notice that t h e items listed in these three columns a r e numbered consecutively. T h e last several columns a r e used t o track t h e work responsibility and completion of t h e work.

Example 10-2
Consider the reactor system shown in Figure 10-8.The reaction is exothermic, so a cooling system is provided to remove the excess energy of reaction. In the event that the cooling function is lost, the temperature of the reactor would increase. This would lead to an increase in reaction rate, leading to additional energy release. The result would be a runaway reaction with pressures exceeding the bursting pressure of the reactor vessel. The temperature within the reactor is measured and is used to control the cooling water flow rate by a valve. Perform a HAZOP study on this unit to improve the safety of the process. Use as study nodes the cooling coil (process parameters: flow and temperature) and the stirrer (process parameter: agitation).

Solution
The guide words are applied to the study node of the cooling coils and the stirrer with the designated process parameters. The HAZOP results are shown in Table 10-7, which is only a small part of the complete analysis.

Table 10-7

HAZOP S t u d y A p p l i e d t o t h e E x o t h e r m i c R e a c t o r of E x a m p l e 10-2.

Hazards and Operability Review
Project name: Example 10-2 Process: Reactor of Example 10-2 Section: Reactor shown in Example 10-2
Item Study node
1A Cooling coils

Date: 1/1/93

Page

1

of

2

Completed: No action: Reply date:

I
1
DAC 1/93 DAC 1/93 DAC 2/93 DAC 2/93 DAC 2193 DAC 1193 JFL 1/93

1

Reference drawing: Figure 10-8 Possible causes Possible consequences
1. Loss of cooling, possible runaway ,, 2.

Process parameters
Flow

Deviations (guide
No

Action required

to:

1. Control valve fails closed 2. Plugged cooling coils

3. Cooling water service failure 4. Controller fails and closes valve

3.
4.
5.

R o

1B

5.Air pressure fails, closing valve
High 1. Control valve fails open 2. Controller fails and opens valve 1. Partially plugged cooling line 2. Partial water source failure 3. Control valve fails to respond

1C

Low

1. Reactor cools, reactant conc. builds, possible runaway on heating 2. 1. Diminished cooling, possible runaway 2. 3. 1. Not possible here 1. Loss of cooling, possible runaway 2. 1. None 1. Temperature rises, possible runaway 1. None-controller handles 1. Cooling system capacity limited, temp, increases 1. No mixing, possible accumulation of unreacted materials 2. Monomer feed continues, possible accumulation of unreacted materials 1. None

1. Select valve to fail open 2. Install filter with maintenance procedure Install cooling water flow meter and low flow alarm Install high temperature alarm to alert operator 3. Check and monitor reliability of water service 4. Place controller on critical instrumentation list 5. See l A . l 1. Instruct operators and update procedures
2. See 1A.4 1. See 1A.2

1D 1E
1F

As well as, part of, reverse

1G 1H 1I 1J 1K 1L 2A Stirrer Temp. Agitation

1. Contamination of water supply 1. Covered under 1C 1. Failure of water source resulting in backflow 2. Backflow due to high backpressure Other than, 1. Not considered possible sooner than, 1. Cooling normally started early later than 1. Operator error Where else Low High No 1. Not considered possible 1. Low water supply temperature 1. High water supply temperature 1. Stirrer motor malfunction 2. Power failure

2. See 1A.2 3. Place valve on critical instrumentation list 1. None 1. See 1A.2 2. Install check valve

JFL 1193
X X

JFL 2193
X X

1. Interlock between cooling flow JW and reactor feed 1. None 1. Install high flow alarm andlor JW cooling water high temp. alarm 1. Interlock with feed line JW 2. Monomer feed valve must fail closed on power loss JW

1/93
X X

1193 1193 2/93
X

26

More

1. Stirrer motor controller fails,

resulting in high motor speed

454

Chapter 10

Hazards Identification

The potential process modifications resulting from this study (Example 10-2) are the following: install a high-temperature alarm to alert the operator in the event of cooling function loss, install a high-temperature shutdown system (this system would automatically shut down the process in the event of a high reactor temperature; the shutdown temperature would be higher than the alarm temperature to provide the operator with the opportunity to restore cooling before the reactor is shutdown), install a check valve in the cooling line to prevent reverse flow (a check valve could be installed both before and after the reactor to prevent the reactor contents from flowingupstream and to prevent the backflow in the event of a leak in the coils), periodically inspect the cooling coil to ensure its integrity, study the cooling water source to consider possible contamination and interruption of supply, install a cooling water flow meter and low-flow alarm (which will provide an immediate indication of cooling loss). In the event that the cooling water system fails (regardless of the source of the failure), the high-temperature alarm and emergency shutdown system prevents a runaway reaction. The review committee performing the HAZOP study decided that the installation of a backup controller and control valve was not essential. The high-temperature alarm and shutdown system prevents a runaway reaction in this event. Similarly, a loss of coolant water source or a plugged cooling line would be detected by either the alarm or the emergency shutdown system. The review committee suggested that all coolant water failures be properly reported and that if a particular cause occurred repeatedly, then additional process modifications were warranted. Example 10-2 demonstrates that the number of suggested process changes is great, although only a single process intention is considered. The advantage to this approach is that it provides a more complete identification of the hazards, including information on how hazards can develop as a result of operating procedures and operational upsets in the process. Companies that perform detailed HAZOPs studies find that their processes operate better and have less down time, that their product quality is improved, that less waste is produced, and that their employees are more confident in the safety of the process. The disadvantages are that the HAZOP approach is tedious to apply, requires considerable staff time, and can potentially identify hazards independent of the risk.

10-4 Safety Reviews
Another method that is commonly used to identify safety problems in laboratory and process areas and to develop solutions is the safety review. There are two types of safety reviews: the informal and the formal.

Phosgene is a colorless vapor with a boiling point of 46.8"F. well above the TLV. Example 10-3 Consider the laboratory reactor system shown in Figure 10-9. It includes the individual responsible for the process and one or two others not directly associated with the process but experienced with proper safety procedures. r U def lu-x Scrubber (NaOH) REACTOR Figure 10-9 Original design of phosgene reactor before informal safety review.) with aniline to produce isocyanate and HCl. Aniline is a liquid with a boiling point of 364°F. and its odor threshold is 0. The reaction is shown in Figure 10-10. This system is designed to react phosgene (COCI. Its TLV is 2 ppm.10-4 Safety Reviews VENT f COCI. The TLV for phosgene is 0. The isocyanate is used for the production of foams and plastics. The improvements must be implemented before the process is operated. Significant improvements should be summarized in a memo for others to reference in the future.1 ppm. .5-1 ppm. The informal safety review procedure usually involves just two or three people. The reviewers simply meet in an informal fashion to examine the process equipment and operating procedures and to offer suggestions on how the safety of the process might be improved. It is absorbed through the skin. Thus it is normally stored as a liquid in a container under pressure above its normal boiling point temperature. The idea is to provide a lively dialogue where ideas can be exchanged and safety improvements can be developed. Figure 10-10 Reaction stoichiometry for phosgene reactor. The informal safety review is used for small changes to existing processes and for small bench-scale or laboratory processes.

trap catches liquid phosgene. flow indicator provides visual indication of flow.OH REACTOR Figure reactor Figure 10-11 Final design of phosgene reactor after informal safety review. A caustic scrubber is used to remove the phosgene and HCl vapors from the exit vent stream. and processes that need an updated review. before the process is started. bubblers are used instead of scrubbers because they are more effective. The complete process is contained in a hood. the caustic would absorb the phosgene). Solution The safety review was completed by two individuals. 5. substantial changes in existing processes. In addition. and operating areas (this paper is normally white but turns brown when exposed to 0. 7. In the process shown in Figure 10-9 the phosgene is fed from the container through a valve into a fritted glass bubbler in the reactor. The changes and additions to the process are as follows: 1. Conduct an informal safety review on this process. ammonium hydroxide bubbler is more effective for absorbing phosgene. The formal safety review is used for new processes. 4. and (3) post an up-to-date process sketch near the process.456 Chapter 10 Hazards Identification Vacuum NaoH C NH. vacuum is added to reduce boiling temperature. the reviewers recommended the following: (1) Hang phosgene indicator paper around the hood. The reflux condenser condenses aniline vapors and returns them to the reactor. 2. pail of caustic is added (the phosgene cylinder would be dumped into this pail in the event of a cylinder or valve leak. 3. The final process design is shown in Figure 10-11. relief system is added with an outlet to a scrubber to prevent hazards resulting from a plugged fritted glass bubbler.1 ppm of phosgene). room. The formal safety review is a three-step pro- . 6. daily. (2) use a safety checklist.

This checklist is used before every startup. 111. and implementing the recommendations. Describes what to do for each case so that the system fails safely. D. and disposed. having a committee review the report and inspect the process. Provides the physical data and design information associated with the equipment. B. The formal safety review report includes the following sections: I. such as loss of steam. and air pressure. pressures. electricity. Equipment specifications: Identifies the equipment by manufacturer name and model number. 2. Normal operating procedures: Describes how the process is operated. C. This is done after the formal safety review is complete. Major release procedures: Describes what to do in the event of a major spill of toxic or flammable material. and loss of electricity. reactor runaway. Cleanup procedures: Describes how to cleatt the process after use. D. Safety checklist: Provides the complete safety checklist for the operator to complete before operation of the process. 1V. VI. and relevant physical property data for the materials used. Introduction A. Sketches of the equipment are provided. Process overview or summary: Provides a brief description of the process with an emphasis on the major hazards in the operation.10-4 Safety Reviews 457 cedure. Equipment setup A. Waste disposal procedure: Describes how toxic or hazardous materials are collected. Equipment description: Describes the configuration of the equipment. This consists in preparing a detailed formal safety review report. B. Reactions and stoichiometry: Provides the chemical reaction equations and stoichiometry. B. Raw materials and products: Refers to specific hazards and handling problems associated 1 with the raw materials and products. handled. Emergency shutdown: Describes the procedure used to shut down the equipment if an emergency should occur. 3. This includes: 1. C. or inert padding. Overview or summary: Provides a brief summary of the results of the formal safety review. air pressure. Material safety data sheets: Provided for each hazardous material used. water. V. 1 . . Procedures A. water. Safety procedures: Provides a description of the unique concerns associated with the equipment and materials and specific procedures used to minimize the risk. Engineering data: Provides operating temperatures. Fail-safe procedures: Examines the consequences of utility failures. This includes major leaks. Discusses procedures to minimize these hazards.

. Figure 10-13shows the modified process after the formal safety review has been completed. and ventilation. The extraction is conducted at 190°F. 3. provide dip legs in all drums to prevent the free fall of solvent resulting in the generation and accumulation of static charge. add a charge drum with grounding. because of a difference in densities. provide a vacuum connection to the dirty toluene storage for charging. This process is used to clean water-soluble impurities from contaminated toluene. The separation is achieved with a Podbielniak centrifuge. 8. The significant changes or additions added as a result of the review are as follows: 1 add grounding and bonding to all collection and storage drums and process vessels. 5. add a relief valve to the dirty toluene storage tank. The contaminated toluene is fed from a storage tank into the Pod. add inerting and purging to all drums. bonding. Perform a formal safety review on this process. The light phase (contaminated toluene) is fed to the periphery of the centrifuge and travels to the center. Example 10-4 A toluene water wash process is shown in Figure 10-12. or Pod. inerting. 7.458 I I TOLUENE Chapter 10 Hazards Identification DIRTY TOLUENE STORAGE Figure 10-12 Toluene water wash process before formal safety review. 6. and . Solution The complete safety review report is provided in appendix D. The heavy liquid out (contaminated water) is sent to waste treatment and the light liquid out (clean toluene) is collected in a 55-gal drum. The heavy phase (water) is fed to the center and travels countercurrent to the toluene to the periphery of the centrifuge. add heat exchangers to all outlet streams to cool the exit solvents below their flash point (this must include temperature gauges to ensure proper operation). add elephant trunks at all drums to provide ventilation. Both phases are mixed within the centrifuge and separated countercurrently. 4. 2.

9.10-5 Other Methods 459 an°F . provide a waste water collection drum to collect all waste water that might contain substantial amounts of toluene from upset conditions. The formal safety review can be used almost immediately. changing the emergency procedure for spills to include (a) activating the spill alarm.9 ~Vacuu - 77°F T POD DIRTY TOLUENE STORAGE 7 WATER "-? F DIRTY TOLUENE CHARGE Figure 10-13 Toluene water wash process after formal safety review. and is known to provide good results. the question might be. (b) increasing the ventilation to high speed. and (c) throwing the sewer isolation switch to prevent solvent from entering the main sewer lines. Additional changes were made in the operating and emergency procedure. the committee participants must be experienced in identifying safety problems. is relatively easy to apply. a more formal HAZOP study may be more effective in identifying the hazards. "What if" analysis: This less formal method of identifying hazards applies the words "what if" to a number of areas of investigation. They included 1. However. For less experienced committees. . For instance. 10-5 Other Methods Other methods that are available for identifying hazards are the following: 1. checking the room air periodically with colorimetric tubes to determine whether any toluene vapors are present and 2. What if the flow stops? The analysis team then decides what the potential consequences might be and how to solve any problems.

. H A Z O P and H A Z A N . ch. but in a different stoichiometric ratio than the hydrolysis reaction. 3. 2d ed. Water is also circulated by a centrifugal pump from the temperature bath through coils in the reactor vessel. The quenching is achieved by adding an excess of aniline to the sample. The quench reaction is The quenching reaction also forms acetic acid. The water is mixed with the acetic anhydride just before it enters the reactor. Kletz. Problems 10-1. 1994). Human error analysis: This method is used to identify the parts and the procedures of a process that have a higher than normal probability of human error. Because the titration procedure requires time (relative to the hydrolysis reaction time). 7th ed.460 Chapter 10 Hazards Identification 2. A temperature controller in the water bath maintains the temperature to within 1°Fof the desired temperature. 1992). Lees. 2d ed. In this reaction acetic anhydride [(CH. 3d ed. (London: Butterworths.COOH). Control panel layout is an excellent application for human error analysis because a control panel can be designed in such a fashion that human error is inevitable. effects. and criticality analysis (FMECA): This method tabulates a list of equipment in the process along with all the possible failure modes for each item. This maintains the reactor temperature at a fixed value. Failure mode. Thus it is possible to determine the acetic anhydride concentration at the time the sample was taken. England: Institution of Chemical Engineers. The concentration of acetic anhydride at any time in the CSTR is determined by titration with sodium hydroxide. (New York: American Institute of Chemical Engineers. (Warwickshire. it is necessary to quench the hydrolysis reaction as soon as the sample is taken.CO)20] reacts with water to produce acetic acid (CH. The effect of a particular failure is considered with respect to the process. Loss Prevention in the Process Industries. Suggested Reading Dow's Fire and Explosion Index Hazard Classification Guide. 1992). (New York: American Institute of Chemical Engineers. 1996). The hydrolysis of acetic anhydride is being studied in a laboratory-scale continuously stirred tank reactor (CSTR). Guidelines for Hazard Evaluation Procedures. Trevor A. Frank P. 8. Water and acetic anhydride are gravity-fed from reservoirs and through a set of rotameters. The initial experimental design is shown in Figure 10-14.

Perform a HAZOP study on the intention "hot solvent from heat exchanger. A heat exchanger is used to heat flammable.. Samples are withdrawn from the point shown and titrated manually in a hood.Problems 461 Acet ic Water 7 F Anhydr ide 7 F - -. ~3 Sample V5 I------- W--c P o i n t C i r c u l a t i n g Water Thermocouple Bath Figure 10-14 Acetic anhydride reactor system. Suggest modifications to improve the safety.. a. and a controller valve manipulates the amount of steam to the heat exchanger to achieve the desired set point temperature. The temperature of the outlet stream is measured by a thermocouple. 10-2. Perform an informal safety review on the experiment. Perform a HAZOP study on the laboratory process of Problem 1. Develop a safety checklist for use before operation of this experiment. volatile solvents.A + . I I . Identify the study nodes of the process. as shown in Figure 10-15. . Consider the intention "reactant flow to reactor" for your analysis. What safety equipment must be available? c." Recommend possible modifications to improve the safety of the process.- h c z . v1 I v2 -----i --- a II R l [ R2 - .. b. b. What specific recommendations can you make to improve the safety of this experiment? 10-3.-. a.

The gas flow is controlled by an electric solenoid valve connected to a thermostat. The hot combustion gases pass through a heat exchanger to heat fresh air for space heating.462 Chapter 10 Hazards Identification le a m /Control Valve i l e Solvent Condensate v G h Heat Exchanger Figure 10-15 Volatile solvent heating system. A gas-fired furnace is shown in Figure 10-16. 10-4. A high-temperature switch shuts off all gas in the event of high temperature in the fresh air plenum. The gas is ignited by a pilot light flame. . Chimney Heat Exchanger Cold Air I Burner I rmostat: n Valve Heat Close Valve Figure 10-16 if T > THish Furnace control system.

World War I1 submarines used torpedo tubes with outer and inner doors. A problem can also arise because of failure of the pilot light. resulting in flooding of the torpedo room and possible sinking of the sub. The presence of pressurized water was a direct indication that the outer door was open. the valve was opened momentarily to check for the presence of pressurized water in the tube. ice. Five process pumps are lined up in a row and numbered as shown in Figure 10-17. Identify as many failure modes as possible. What is the most likely failure mode? d. 10-5. a small pipe and valve were attached to the top of the torpedo tube in the torpedo room. Suggest at least two ways to prevent this problem. 10-7. The torpedo was loaded into the tube from the torpedo room using the inner door. Because no direct visible check was possible. The inner door was then closed. Use the HAZOP guide words to identify additional possibilities. What type of valve (normally open or normally closed) is recommended for the gas supply? c. and beverage (composed of syrup and water) in a sequential order. 10-6. Can you identify the hazard? A similar layout led to a serious accident by a maintenance P r o c e s s Pumps Figure 10-17 Pump layout. b. and chimney.Problems 463 a. . heat exchanger. and the torpedo launched. leading to the inner door being opened when the outer door was open. Determine a failure mode for this system. leading to excessive heating of the plenum and possible fire. leading to combustible gases in the furnace. Beverage dispensers are notorious for either taking one's money or not delivering the proper beverage. One problem was ensuring that the outer door was closed before the inner door was opened. Determine the various ways in which this system can fail. Before opening the inner door. Consider a beverage dispenser that delivers a paper cup. the outer door opened. The machine also makes change.

Because the flammable vapors were water soluble. 10-8. Stupid! This also applies to hazards. This configuration resulted in an explosion and fire because of loss of inert material.2 psi). Can you explain why? . Complicated designs are almost always more hazardous than simple ones.5 in of water gauge vacuum (0. Can you explain why? 10-11. The level controller and pump ensure that the sump level is maintained below a maximum height. both with independent level controllers. Figure 10-19 shows a storage tank blanketed with nitrogen. He connected a hose to the vent at the top of the tank and ran the hose down to the ground.464 Chapter 10 Hazards Identification Figure 10-18 Sump level control system. Can you explain what happened and how? 10-10. worker who was sprayed by hot solvent when he disconnected a pump line on the wrong pump. Can you suggest a much simpler system? 10-9. Figure 10-20 shows two tanks in series.1 psi) and about 6 in of water gauge pressure (0. During a subsequent operation that involved emptying the tank. The roof was equipped with a vent pipe with a flame arrestor. volatile liquid. Figure 10-18 shows a sump designed to collect process fluids. Storage tanks typically are not capable of withstanding much pressure or vacuum. an accident occurred. A welding operation was to occur on the roof of a storage vessel. he stuck the end of the hose in a drum full of water. The foreman recognized a possible hazard from flammable vapor escaping from the vent pipe and igniting on the sparks from the welding operation. An accident like this might be attributed to human error but is really a hazard resulting from poor layout. This configuration will result in the lower tank inevitably overflowing. The tank contained a flammable. A good acronym in chemical plant design is KISS -Keep It Simple. Standard storage tanks are designed for a maximum of 2.

. .. .. . . . .. . .. . .. . . . ... . This memo will be given to each committee member before the formal safety review committee meeting..Problems 465 S t o r a g e Tank . "Fail-safe" is a concept used to specify the position of process instrumentation in the event of power.. .. . .. . . For instance. .. . . Prepare a formal safety review memo for the gas-fired furnace described in Problem 104 and shown in Figure 10-16. air pressure... . .. . . . . . . . . 10-15. . H y d r o c a r ban:. . ... .. . .. .. .. . . . . identify the study nodes of the reactor process. .. . . .. Describe an informal safety review process for using a cylinder of phosgene to charge gaseous phosgene to a reactor. 10-13. . . . . . .:-:. .. 10-12. . . . . .... . .. . .. .. .. as shown... Nitrogen L 40 p s i Figure 10-19 Nitrogen padding system for a storage tank. .. . .. .. .. . In Figure 10-8. 10-16. Review up to the reactor only. . .. The intention of the checklist is to ensure the system is safe before operation. . . . .. . . . . . Develop a safety checklist for the system described in Example 10-3 and shown in Figure 10-11. .. . i q u i d :::::. . . . . 10-14. . . . . . . . or other utility failures. ..:. .. . the valve supplying cooling water to a chemical reactor would fail in the open position ("fail open") in the 1 Figure 10-20 Level tanks in series. . ... . . . .. . ... . . .:-:. ...

. and La La Pharmaceuticals has decided to produce this in their existing pilot plant operations in Lala Land. This would provide maximum cooling to the reactor and prevent dangerous high temperatures in the vessel. A flammable solvent is heated by steam in a heat exchanger. in their chemical laboratories. La La Pharmaceuticals has recently discovered a new drug. The reaction is exothermic. The valve controls the flow of steam to the exchanger.466 Chapter 10 Hazards Identification event of a power failure. Specify either fail open or fail close. A valve controls the pressure in a steam header. This led to an accidental release of toxic and flammable material. A valve cannot be closed until a furnace is shut down. For the next stage of clinical studies over 50 kg of Lalone is required." Determine at least 15 ways in which the operator might fail to perform the instructions correctly. Two valves cannot both be closed at the same time. Feed to a reactor cannot be started until the reactor vessel stirring motor is activated. f. Thermocouples in chemical plants are usually found in sheaths. c. Interlocks can be mechanical or electronic. In many cases they can be as simple as a lock and key. g. 10-17. Lalone is expected to be a blockbuster drug. Lalone. although they looked like the sheathed type. Specify the proper fail-safe positions for the valves in the following equipment. a. One chemical plant had some thermocouples that did not have sheaths. b. b. d. In one such tank the material stored in the tank was changed and an overflow resulted. c. Why? 10-21. As the safety director for the pilot plant operations. A valve controls the flow of natural gas to a utility furnace in a power station. Specify the simplest mechanical interlock capable of achieving the following functions: a. e. A valve must be closed before a pump is started. Can you explain why? 10-20. you are in charge of ensuring the safety of all operations. h. d. raking in billions of dollars each year. A remotely operated valve is used to fill a tank from a supply line. The reaction is endothermic. Interlocks are used to ensure that operations in a chemical plant are performed in the proper sequence. Liquid levels in storage tanks are frequently determined by measuring the pressure at the bottom of the tank. A valve controls the flow rate of reactant to a reactor vessel. 10-19. A valve controls the flow rate of reactant to a reactor vessel. These sheaths protect the thermocouple and also allow the thermocouple to be removed and replaced without shutting down the process. 10-18. A valve controls combustion air to a furnace. A remotely operated valve is connected to a drain on a storage tank. A process operator is given the following instructions: "Charge 10 lb of catalyst into batch reactor A at 3 hr into the cycle.

followed by drying to produce Lalone (all reactions are carried out in the liquid phase and require acetone as a solvent). Thus. The coffee product is collected in the coffee pot. b. Based on your safety knowledge and experience. the scale printed on the gauge read 0 to 1200 psig. A coffee maker has a reservoir where a quantity of clean water is poured. Perform a HAZOP study on a common coffee maker. and (6) the Engineering Division has already started writing the operating procedures for the eventual process. resulting in an accident.three sets of reactions to produce intermediates. c. The valve did not close when it should have. white powder. This gauge was accidentally overpressured. corresponding to the pneumatic signal sent from the plant. Pneumatic process equipment operates in the range of 3-15 psig. (4) so far. and the plant exploded. The scale actually indicated 0-100% of a temperature range of 0-200°C. a signal of 3 psig might represent 0 psig in the process and 15 psig might represent 1200 psig in the process. What was the set point temperature the operator actually set? 10-23. fresh-brewed coffee in the coffee pot. corresponding to the actual process pressures. An operator was told to control the temperature of a reactor at 60°C. identify the major hazards in this process that you would be concerned about.) In a devolatilizer. What additional information will you need to conduct the hazard analysis study? 10-22.Problems 467 During a meeting with the chemist who synthesized Lalone in the laboratory. A light in the control room of a chemical plant indicated whether a valve was closed or not. (5) management wants the pilot plant operations to be started as soon as possible. and identify the study nodes. A pneumatic pressure gauge was designed to operate in the range of 3-15 psig. However. A small heater percolates the water up to the top of the coffee maker. b. In reality it indicated only the status of the signal being sent to the valve. Liquid was discharged and injured the operator. and most physical and chemical properties are not known. (2) Lalone is synthesized by a batch process through a series of four major steps . Describe how you would structure a hazard study for the Lalone manufacturing process. a. where it drips down through the coffee grounds and filter assembly. (This problem requires student access to the Dow Fire and Explosion Index manual. for example. (3) the chemical reactions are not fully understood. 10-26. As the safety director of the pilot plant: a. a solvent (60% cyclohexane and 40% pentane) is removed from a . Use as a design objective hot. What happened? 10-24. you have learned the following: (1)Lalone is a fine. Draw a sketch of the coffee machine. This caused a runaway reaction that overpressured the vessel. Why? How would you prevent this problem? 10-25. Lalone has been manufactured only in the laboratory and in small quantities (less than 50 g). He set the set point of the temperature controller at 60.

50 per pound and an annual plant production rate of 35 million lb. The storage area has the required drainage to direct a spill away from the tank. cyclohexane. The transfer operations in and out of storage are monitored by computer control with emergency shutdown capability.000gal. The plant is also under computer control with emergency shutdown based on redundant inputs. estimate the business interruption loss. and a water deluge system.4 X 103B ~ U J I ~ . The process has several interlocks to prevent polymerization. The process is run above the flash point of the solvent at 300 mm Hg. written. estimate the maximum probable property damage. A diesel-driven fire pump is capable of handling the emergency demand for 4 hr. The storage system has remotely operated emergency block valves on all transfer lines into and out of the tank. and isopropene. The process area has a 1% sloping concrete surface with a remote impounding area capable of handling all of a spill and 30 min of fire water. The butadiene storage area is equipped with a chilled glycol cooling system that can be operated from an emergency generator if necessary. Assuming a product value of $1. isopentane. Backflow protection has been installed and is tested to prevent backflow into the transfer line and storage. The vapor space in the vessel is inerted. Loss control features include combustible gas detectors installed around the containment area and transfer system. A reactive chemicals review was completed recently. a. Diesel-powered fire water pumps can provide a maximum fire water demand for 4 hr. Determine the Dow F&EI value for this process to estimate the relative degree of hazard.) Consider a butadiene storage vessel in a tank farm area containing butadiene. The process unit has many loss control features. . (This problem requires student access to the Dow Fire and Explosion Index manual.468 Chapter 10 Hazards Identification polymer and sent to the solvent recycle section of the plant for treatment and recovery. The devolatilizer is located in an open structure with good access for fire fighting. with the relief valve set at 50 psig. The butadiene storage is diked separately from the other materials. The maximum butadiene storage capacity is 100. A water deluge system has been installed around the storage tank and transfer pump. The process area has combustible gas detectors. The process has complete. Cable trays are protected with deluge. 10-27. c. and up-to-date operating instructions. Assuming an equipment value within the radius of exposure of $1 million. Assume a potential spill of 8000 lb of flammable material with a heat of combustion of 19.6263. b. The vessel has a relief device set at 50 psig. and portable dry chemical extinguishers are in the process area. Vacuum is always broken with nitrogen. The plant has a diesel emergency power generator with an emergency cooling system. Operating instructions are current. The normal pressure of the butadiene storage vessel is 15 psig. styrene. and the system has been through a recent reactive chemicals review. fireproofing. The specific gravity of the butadiene is 0.

Frequently. A flammable liquid is to be stored in a large storage vessel. by the time the alarm sounds. it is too late. with the weakest part being the seam along the bottom of the tank. Assuming an equipment value within the radius of exposure of $1million. Your manufacturing plant has purchased a number of robots to facilitate production. Determine the Dow F&EI value for this process to estimate the relative degree of hazard. Two vessels are available.Problems 469 a. estimate the maximum probable property damage. Assuming a product value of $2. the temperature can rise quickly. Exothermic chemical reactions are frequently dangerous because of the potential for a runaway reaction. Loss of cooling can be detected by measuring the temperature within the reactor and sounding an alarm. b. What are the main hazards associated with robots? What are some effective safeguards against these hazards? . During a runaway reaction. resulting in a higher reaction rate and higher energy generation. 10-29. Draw the instrumentation diagram. Cooling coils are provided in batch reactors to remove the energy of reaction. 10-28. One vessel is called a weak seam roof tank. Design a better instrumentation and alarm configuration to detect loss of cooling more directly. the reactor temperature rises. with the weakest part of the vessel being the welded seam between the roof and the vertical wall of the tank. In the event of a cooling water failure. The result is a runaway reaction. Which tank is the best choice for storing this material? 10-30. c. estimate the business interruption loss. The other vessel is a domed roof tank.00 per pound for this plant and an annual production rate of 10 million Ib. resulting in dangerous pressures within the reactor and a possible explosion.

.

these numbers are obtained by some rather simple calculations involving published correlations. Incident identification describes how an accident occurs. show how the failure probabilities of individual hardware components contribute to the failure of a process. This is a form of incident identification. This is a form of consequences analysis. We focus on determining the frequency of accident scenarios. However. and describe the relationship between quantitative risk analysis (QRA) and LOPA. The Dow F&EI includes a calculation of the maximum probable property damage (MPPD) and the maximum probable days outage (MPDO). and the results R . damage to the environment or capital equipment. No probabilities or numbers are used with the typical HAZOP study. In this chapter we will review probability mathematics. including the mathematics of equipment failure. and days outage. This includes loss of life. although the experience of the review committee is used to decide on an appropriate course of action. describe two probabilistic methods (event trees and fault trees). The hazards identification procedures presented in chapter 10 include some aspects of risk assessment. Hazard and operability (HAZOP) studies provide information on how a particular accident occurs. The last two sections show how the frequencies are used in QRA and LOPA studies. It should be emphasized that the teachings of this chapter are all easy to use and to apply. describe the concepts of layer of protection analysis (LOPA). Consequence analysis describes the expected damage. It frequently includes an analysis of the probabilities. LOPA is a simplified QRA.C H A P T E R 1 1 Risk Assessment isk assessment includes incident identification and consequence analysis.

and it is given by The failure density function is defined as the derivative of the failure probability: The area under the complete failure density function is 1. E. Rijnsdorp. MI: Ann Arbor Science. This is called the average failure rate and is represented by p with units of faultsltime. p. The speed at which this occurs depends on the value of the failure rate p. f1 f1 f (t) dt = p/ I. As t +oo. This simple exponential distribution is the one that is used most commonly because it requires only a single parameter. The probability that the component will not fail during the time interval (0. The overall probability of a failure in a process depends highly on the nature of this interaction. Equation 11-1assumes a constant failure rate p.> e-@ dt = e-"' - ~ " 1 . the faster the reliability decreases. Process Dynamics. p. The failure density function is used to determine the probability P of at least one failure in the time period to to tl: P(r. With adequate data it can be shown that. on average. The complement of the reliability is called the failure probability (or sometimes the unreliability). Data are collected on the failure rate of a particular hardware component. ) = 1. 11-1 Review of Probability Theory Equipment failures or faults in a process occur as a result of a complex interaction of the individual components. 1982).the reliability goes to 0. the component fails after a certain period of time. Roffel and J. and Protection (Ann Arbor. t) is given by a Poisson distribution1: where R is the reliability. Control. The higher the failure rate. In this section we define the various types of interactions and describe how to perform failure probability computations. . P. Other and more complex distributions are available. 381. (11-4) lB.472 Chapter 11 Risk Assessment are often the basis for significantly improving the design and operation of chemical and petrochemical plants. +l .

- 11-1 Review of Probability Theory Failure Rate Failure Density Failure Probability Reliability Figure 11-1 Typical plots of (a) the failure rate p. (b) the failure density f(t).1 -- Infant Mortality 4 Old Age D 0 P Time Figure 11-2 A typical bathtub failure rate curve for process hardware. Between these two periods (denoted by the lines in Figure 11-2). f . shown in Figure 11-2. . The time interval between two failures of the component is called the mean time between failures (MTBF) and is given by the first moment of the failure density function: E ( t ) = MTBF = 1 Typical plots of the functions p. (c) the failure probability P(t). and (d) the reliability R(t). The failure rate is approximately constant over the midlife of the component. Failure Rate. and R are shown in Figure 11-1. The failure rate is highest when the component is new (infant mortality) and when it is old (old age). The integral represents the fraction of the total area under the failure density function between time to and tl. P. p (faults/time) 0 Period of Approximately Constant 1. the failure rate is reasonably constant and Equations 11-1 through 11-5 are valid. Equations 11-1 through 11-5 are valid only for a constant failure rate p. Many components exhibit a typical bathtub failure rate.

The overall process failure probability is computed from the individual component probabilities. The total reliability for parallel units is given by where Ri the reliability of an individual process component. is Process components also interact in series. The logical OR function represents this case. Process components interact in two different fashions. Equation 11-9 is expanded to The cross-product term P ( A ) P ( B )compensates for counting the overlapping cases twice. is the failure probability of each component. This rule is easily memorized because for parallel components the probabilities are multiplied. This parallel structure is represented by the logical AND function. This means that the failure probabilities for the individual components must be multiplied: where n is the total number of components and P. This means that a failure of any single component in the series of components will result in failure of the process. In this case .474 Chapter 11 Risk Assessment Interactions between Process Units Accidents in chemical plants are usually the result of a complicated interaction of a number of process components. For series components the overall process reliability is found by multiplying the reliabilities for the individual components: The overall failure probability is computed from For a system composed of two components A and B. In some cases a process failure requires the simultaneous failure of a number of components in parallel. Consider the example of tossing a single die and determining the probability that the number of points is even or divisible by 3.

Actual values would Table 11-1 Failure Rate Data for Various Selected Process Components1 Instrument Controller Control valve Flow measurement (fluids) Flow measurement (solids) Flow switch Gas-liquid chromatograph Hand valve Indicator lamp Level measurement (liquids) Level measurement (solids) Oxygen analyzer pH meter Pressure measurement Pressure relief valve Pressure switch Solenoid valve Stepper motor Strip chart recorder Thermocouple temperature measurement Thermometer temperature measurement Valve positioner Faultslyear 0.29 0.70 6.11-1 Review of Probability Theory 475 P(even or divisible by 3 ) = P(even) + P(divisib1e by 3) - P(even and divisible by 3 ) .022 0. If the failure probabilities are small (a common situation). .22 0.027 0. the term P(A)P(B) is negligible.044 0. 1986).42 0.86 5.6 0. Lees.12 30. These are average values determined at a typical chemical process facility. 313.65 5. Loss Prevention in the Process Industries (London: Butterworths.88 1.14 3. For this special case Equation 11-9 reduces to Failure rate data for a number of typical process components are provided in Table 11-1.13 0.44 'Selected from Frank P. and Equation 11-10 reduces to This result is generalized for any number of components.14 0.044 1.52 0.75 1. p.41 0. The last term subtracts the cases in which both conditions are satisfied.60 1.

hard. Figure 11-3 Computations for various types of component linkages. . = 1 . the environment.L n R ) / t Tpi n R = 1 .( n 1 .(1 .2 i = 1 P = 1 - 1 i : 1 Pi) R =TR~ i . the unreliability. and the MTBF for this system. 1 n !Pi S e r i e s L i n k o f Components: The f a i l u r e o f e i t h e r component a d d s t o the t o t a l system f a i l u r e . and not intermittent and that the failure of one device does not stress adjacent devices to the point that the failure probability is increased. and the control valve manipulates the flow of coolant. the reliability. P P = PIP2 P = i . the design. Determine the overall failure rate. the controller decides on an appropriate control strategy. and other factors. The assumptions in this analysis are that the failures are independent. there Note t h a t i s n o c o n v e n i e n t way t o c o m b i n e the failure rate. 1 = :I 3 . A summary of computations for parallel and series process components is shown in Figure 11-3. Assume a 1-yr period of operation. . materials of construction.476 Chapter 11 Risk Assessment Failure Probability Reliability F a i l u r e Rate p p1 z P = P 1 . depend on the manufacturer. The flow is measured by a differential pressure (DP) device. = .(1 RIJq-~ F:=P Rz - Pl)(l - P2) R = R1R2 P P = = PI+ P. Example 11-1 The water flow to a chemical reactor cooling coil is controlled by the system shown in Figure 11-4.) P = ( .T(I i = l Ri) P a r a l l e l L i n k o f Components: The f a i l u r e o f t h e s y s t e m r e q u i r e s t h e f a i l u r e o f b o t h components.

901yr.10) = 2. on average.55 0.10 = 0. The reliability and failure probability are computed for each component using Equations 11-1 and 11-2. The overall failure rate is computed using the definition of the reliability (Equation 11-1): 0. Solution These process components are related in series.11-1 Review of Probability Theory 477 Controller Control Flow Figure 11-4 Flow control system.30 failureslyr. The components of the control system are linked in series.75 0.10 = e-'' p = -ln(0.43 yr. . once every 0.41 Reliability R = e-Pt 0. The results are shown in the following table.R = 1 . the entire system fails.60 0. The failure rates are from Table 11-1.25 0.The result is The failure probability is computed from P = 1.0. Thus.24 Failure probability P=1-R 0. The MTBF is computed using Equation 11-5: MTBF =-= 1 P 0.29 1.76 The overall reliability for components in series is computed using Equation 11-8. This system is expected to fail. Component Control valve Controller DP cell Failure rate p (faultslyr) 0.43 yr.45 0. if any one of the components fails.

87 0. the failure probability.. .14 0. For additional safety an automatic high-pressure reactor shutdown system is installed.42 Reliability R = e-" 0.. 4..478 Chapter 11 Risk Assessment Pressure k/ A larm Switch I . The alarm and feed shutdown systems are linked in parallel.66 Failure probability P=1-R 0. The automatic system stops the flow of reactant in the event of dangerous pressures.34 .. Pressure switch 1 Alarm indicator Pressure switch 2 Solenoid valve Failure rate p (faultslyr) 0._ . Assume a I-yr period of operation.I .04 0. and the MTBF for a high-pressure condition.u Figure 11-5 A chemical reactor with an alarm and an inlet feed solenoid.044 0.13 0. It consists of a pressure switch within the reactor connected to an alarm light indicator.. This reactor contains a high-pressure alarm to alert the operator in the event of dangerous reactor pressures. .--. Example 11-2 A diagram of the safety systems in a certain chemical reactor is shown in Figure 11-5. 3. Also. the reliability.- u2L.-.The reliability and failure probabilities of each component are computed using Equations 11-1 and 11-2: Component 1. Solution Failure rate data are available from Table 11-1..13 0. This system is activated at a pressure somewhat higher than the alarm system and consists of a pressure switch connected to a solenoid valve in the reactor feed line..96 0. develop an expression for the overall failure probability based on the component failure probabilities. Compute the overall failure rate._ .87 0.--- - I Valve .14 0. 2...

1.426.56 yr. For the alarm system .930) = 0. MTBF = -= P 5.80 yr.574 = 0.11-1 Review of Probability Theory 479 A dangerous high-pressure reactor situation occurs only when both the alarm system and the shutdown system fail. with both systems in parallel the MTBF is significantly improved and a combined failure is expected every 13.0.R -1n R 1 = = 1 . p = -In R = -1n(0.96) = 0. These two components are in parallel.5 yr.555 faultslyr. 1 MTBF = . However. For the alarm system the components are in series: R P = = n 2 i=l R. P The two systems are combined using Equation 11-6: p = -In R = -ln(0. a failure is expected once every 1.073 faultslyr. p = -1n(0. for a reactor with a highpressure shutdown system alone. The overall failure probability is given by where P ( A ) is the failure probability of the alarm system and P ( S ) is the failure probability of the emergency shutdown system.835) = 0.835 = 0.80 yr. = (0.180 faultslyr.835.R = 1 . An alternative procedure is to invoke Equation 11-9 directly.7 yr.574) = 0.165. P For the alarm system alone a failure is expected once every 5.0.= 13. MTBF = - 1 = 1.87)(0.7 yr. For the shutdown system the components are also in series: P = 1. Similarly.

of the problem until the spare is needed. as shown. then P ( A ) = PI P ( S ) = P. This is called an unrevealed failure. Emergency alarms and shutdown systems are used only when a dangerous situation occurs. + P. and + P2. and P. After a failure occurs. Without regular and reliable equipment testing. alarm and emergency systems can fail without notice. . The time that the component is operational is called the period of operation and is denoted by T. Failures that are immediately obvious are called revealed failures. It is possible for the equipment to fail without the operator being aware of the situation. is required to repair the component.). However. The difference between this answer and the answer obtained previously is 14. a period of time..3%. If the products PIP. Revealed and Unrevealed Failures Example 11-2 assumes that all failures in either the alarm or the shutdown system are immediately obvious to the operator and are fixed in a negligible amount of time.480 Chapter 11 Risk Assessment For the shutdown system The overall failure probability is then Substituting the numbers provided in the example..P4 are assumed to be small. Figure 11-6 shows the nomenclature for revealed failures. we obtain This is the same answer as before. the spare tire in the trunk might also be flat without the driver being aware. The component probabilities are not small enough in this example to assume that the cross-products are negligible. The MTBF is the sum of the period of operation and the downtime. A flat tire on a car is immediately obvious to the driver. called the period of inactivity or downtime (T.

11 Operotionol 1 m w I ---q _O I I I I T - MTBF Tr . A failure requires a period of time for repair. the time before failure or period of operation is given by where rq is the period of operation between a particular set of failures. .. Similarly. The MTBF is the sum of the period of operation and the repair period: MTBF =-=T .11-1 Review of Probability Theory 481 Component Repoired V] Component Fails Component Repoired 3 I. I -t I I I I I Time Figure 11-6 Component cycles for revealed failures. . For revealed failures the period of inactivity or downtime for a particular component is computed by averaging the inactive period for a number of failures: where n is the number of times the failure or inactivity occurred and T . 1 I" + 7. is the period for repair for a particular failure.

we obtain . similarly. represents the period that the process is in operation. 7" = Combining with Equation 11-19. it follows that the availability is given by 7. It is obvious that The quantity 7. By definition. This situation is shown in Figure 11-7. then The average period of unavailability is computed from the failure probability: P ( t ) dt. is the average period of unavailability during the inspection interval and if ri is the inspection interval. the unavailability is By combining Equations 11-16 and 11-17 with the result of Equation 11-14. The unavailability U is the probability that the component or process is found not functioning. + 7. If 7. and the total time. The availability A is simply the probability that the component or process is found functioning. represents and.482 Chapter 11 Risk Assessment It is convenient to define an availability and unavailability. we can write the equations for the availability and unavailability for revealed failures: For unrevealed failures the failure becomes obvious only after regular inspection.

The failure probability P(t) is given by Equation 11-2. . A decrease in the inspection interval is shown to increase the availability of an unrevealed failure. The result is u=1 An expression for the availability is - -(I P7i 1 - p T 8 ) . 3 Operational I 1 until inspection I I I I I i 5 w C a. K I I I 0 a Failed I I I I 0 I I I I iQ----I -+ + I I T~ i 7. (11-23) If the term PT. This is substituted into Equation 11-23 and integrated. 1 Time Figure 11-7 Component cycles for unrevealed failures. < 1. It demonstrates that. This is a useful and convenient result. for unrevealed failures the process or component is unavailable during a period equal to half the inspection interval. (11-22) A = 1 -(I P7i .11-1 Review of Probability Theory Component Repoired Component Foils Failure not Noticed i n f. (11-24) and Equation 11-21 is integrated to give. for unrevealed failures.egT1). on average. then the failure probability is approximated by < P(t) - Pt.

Thus The shutdown system is available 97. The frequency of this episode is given by For an emergency system with unavailability U.2% of the time.55 faultslyr. Probability of Coincidence All process components demonstrate unavailability as a result of a failure. For the shutdown system p = 0. This requires a coincidence of events. Assume that a dangerous process episode occursp.U episodes.18 faultslyr. Assume that a maintenance inspection occurs once every month and that the repair time is negligible. The inspection period is 1/12 = 0. Solution Both systems demonstrate unrevealed failures. This is usually a valid assumption because on-line process equipment is generally repaired within hours. whereas the inspection intervals are usually monthly. This is every p. For the alarm system the failure rate is p = 0. . Example 11-3 Compute the availability and the unavailability for both the alarm and the shutdown systems of Example 11-2. The unavailability is computed using Equation 11-25: The alarm system is available 99. For alarms and emergency systems it is unlikely that these systems will be unavailable when a dangerous process episode occurs.7% of the time. times in a time interval Ti. a dangerous situation will occur only when the process episode occurs and the emergency system is unavailable.484 Chapter 11 Risk Assessment Equations 11-19 through 11-25 assume a ndgligible repair time.083 yr. The danger results only when a process upset occurs and the emergency system is unavailable.

Compute the MTBC for a high-pressure excursion and a failure in the emergency shutdown device. If the inspection interval T~ is halved. Substituting into Equation 11-27 yields The mean time between coincidences (MTBC) is the reciprocal of the average frequency of dangerous coincidences: Example 11-4 For the reactor of Example 11-3 a high-pressure incident is expected once every 14 months.50 yr.11-1 Review of Probability Theory 485 The average frequency of dangerous episodes Ad is the number of dangerous coincidences divided by the time period: For small failure rates U = $p7iandp.010. Solution The frequency of process episodes is given by Equation 11-26: A = 1 episode/[(l4 months)(l yr112 months)] = 0. = 0. Ad 0.. Assume that a maintenance inspection occurs every month. = ATi. A.023.8571yr. The unavailability is computed from Equation 11-25: The average frequency of dangerous coincidences is given by Equation 11-27: The MTBC is (from Equation 11-29) 1 1 MTBC = . and the resulting MTBC is 100 yr. . This is a significant improvement and shows why a proper and timely maintenance program is important.= ---.020 It is expected that a simultaneous high-pressure incident and failure of the emergency shutdown device will occur once every 50 yr. then U = 0.

2d ed. 1995). 1992). p. and actuators that ensure that the system operates safely and reliably. An example of a redundant temperature control loop is an additional temperature probe. Common Mode Failures Occasionally an incident occurs that results in a common mode failure. and 4. This is achieved with redundant controls. This approach is inductive. Crowl. The event tree approach includes the effects of an event initiation followed by the impact of the safety systems. For example. construct the event tree. The utility is connected to these systems via OR gates. The degree of redundancy depends on the hazards of the process and on the potential for economic losses. 2. eds. (New York: American Institute of Chemical Engineers. This is used effectively to determine the probability of a certain sequence of events and to decide what improvements are required. consider several flow control loops similar to Figure 11-4. The method provides information on how a failure can occur and the probability of occurrence. A common mode failure is the loss of electrical power or a loss of instrument air. A utility failure of this type can cause all the control loops to fail at the same time. including two or more measurements. processing paths. This is a single event that affects a number of pieces of hardware simultaneously. the procedure is used to assign numerical values to the various events. . Guidelines for Hazard Evaluation Procedures. When working with control systems. various safety systems come into play to prevent the accident from propagating. 11-2 Event Trees Event trees begin with an initiating event and work toward a final result. If appropriate data are available. describe the resulting accident event sequences. Grossel and D. The typical steps in an event tree analysis are3 1. A. identify the safety functions designed to deal with the initiating event. S. 264. and actuator (for example. identify an initiating event of interest. cooling water control valve). When an accident occurs in a plant. Handbook of Highly Toxic Materials Handling and Management (New York: Marcel Dekker. controller.486 Chapter I 1 Risk Assessment Systems are designed to function normally even when a single instrument or control function fails. one needs to deliberately design the systems to minimize common cause failures. 3. An example of a redundant temperature measurement is an additional temperature probe. These safety systems either fail or succeed. 2S. This increases the failure rate substantially.

the operator will already be aware of the high-temperature condition. At this point the safety function can either succeed or fail. where the operator notices the high temperature. the horizontal line is continued through the safety function without branching. Horizontal lines are drawn from these two states to the next safety function. . a successful operation is drawn by a straight line upward and a failure is drawn downward. The second safety function is the operator noticing the high reactor temperature during normal inspection. and the circles with the crosses represent unsafe conditions. the upper branch continues through the second function. These are written across the top of the sheet. The sequence description and consequences are indicated on the extreme right-hand side of the event tree.11-2 Event Trees Reactor Feed . except that a high-temperature alarm has been installed to warn the operator of a high temperature within the reactor. By convention. The final safety function is invoked by the operator performing an emergency shutdown of the reactor. The third safety function is the operator reestablishing the coolant flow by correcting the problem in time. 1 g Q G I ing 1 C o i Is Nat i COO e r n Out Cool i n Water I I Alarm ' Therrnocoup High Temperature A larm Figure 11-8 Reactor with high-temperature alarm and temperature controller. For this example. If the high-temperature alarm operates properly. These safety functions are written across the page in the order in which they logically occur.. If a safety function does not apply. Four safety functions are identified. The initiating event is written first in the center of the page on the left. The event tree is written from left to right. The event tree for a loss-of-coolant initiating event is shown in Figure 11-9.This system is identical to the system shown in Figure 10-6. Consider the chemical reactor system shown in Figure 11-8. The open circles indicate safe conditions. A line is drawn from the initiating event to the first safety function. The first safety function is the high-temperature alarm.

001688 0.7't25 I Con 0.01 0.0001875 + + 0.H i a h Temp Operator Notices H i g h Temp Operator Re-starts Cool ins Operator S h u t s Donn Reactor Saf c ty Func t ion: Alarm R l e r t s Operator R Ident if ier: B C D E A 0.0025 Shu Run Shutdown Runaway : 0.0005625 = 0.0000625 = 0. .02975 + + 0. t i a t ing Event: Run Con Loss o f Cool ing 1 Occurrcnce/y~.0075 Shu 0 . 0 0 1 8 7 5 RBOE Run Con 0. 0.2227 = 0. 0. Figure 11-9 Event tree for a loss-of-coolantaccident for the reactor of Figure 11-8.02500 occurrtnccS/yr.02't75 fl 1 I n .2250 o c c u r r c n c e ~ / y r .

The initiating event is always included as the first letter in the notation. Both of these cases represent a failure rate of 1time out of 4. the lettering sequence ADE represents initiating event A followed by failure of safety functions D and E.495 Failure o f 0. Also assume that the operator will notice the high reactor temperature 3 out of 4 times and that 3 out of 4 times the operator will be successful at reestablishing the coolant flow. The frequency associated with the lower branch is computed by multiplying the failure rate of the safety function times the frequency of the incoming branch. The letters indicate the sequence of failures of the safety systems.025 failure per year (sum of failures ADE. Figure 11-10 The computational sequence across a safety function in an event tree. The computational sequence performed at each junction is shown in Figure 11-10. The occurrence frequency for the initiating event is written below the line originating from the initiating event. This is a failure rate of 0. For this example assume that a loss-of-cooling event occurs once a year.5 = 0.01 failureldemand. This event tree analysis shows that a dangerous runaway reaction will occur on average 0.11-2 Event Trees 489 Safety Function 0. The frequency associated with the upper branch is computed by subtracting the failure rate of the safety function from 1 (giving the success rate of the safety function) and then multiplying by the frequency of the incoming branch. For the example here. Again. it is estimated that the operator successfully shuts down the system 9 out of 10 times. represents a successful safety function and the lower branch represents a failure. or once every 40 years. For this example the net frequency is estimated at 0.01 )*0. The failure rates for the safety functions are written below the column headings. (1 -0. A possible solution is the inclusion of a high-temperature reactor shutdown system. by convention. Let us also assume that the hardware safety functions fail 1%of the time they are placed in demand.01 Failures/Dernand Initiating Event 5 ~ u c c e s so f safety Function Occurrences/yr.01 *0. The event tree can be used quantitatively if data are available on the failure rates of the safety functions and the occurrence rate of the initiation event.5 = 0.025 time per year. An event tree for a different initiating event in this study would use a different letter.005 Safety Function Occurrences/yr. The lettering notation in the sequence description column is useful for identifying the particular event.10 failureldemand. Finally.25 failureldemand. or 0. ABDE. This is considered too high for this installation. The net frequency associated with the event tree shown in Figure 11-9 is the sum of the frequencies of the unsafe states (the states with the circles and x's). This is a failure rate of 0. and ABCDE). the upper branch. This control .

001856 RBDE 1 Occurrencc/yr.00001688 0.001875 ABCD t r n t i n u t Oprrmtisa 8. 2Lt50 S h u t Oonn 1 0. This includes a high-temperature shutdown system .001875 0. Operator ~ l ~ r t ~ s. t i t r s Safety Func t ion: Hilh Temp I d e n t i f i e r : B C 0.800625 0.01 0.000001875 Shu t Down Rvnowoy RBC 0.1 0.00750 0.002228 S h u t Down R 1 ABD S h u t Oonn I n i t i a t i n g Event: LOSS o f Cooling 0. I 0.H i g h Ttmp O p t r a t Re-stgrts Cee Shutdonn S h u t s Oena Optrater H i g h Temp Optrator RIor.8006187 S h u t Oonn S h u t Oonn Runaney Figure 11-11 Event tree for the reactor of Figure 11-8.25 8-81 8.8025 0.25 ling Rear t e r 0 E F -r ROE Fo i lure~/Demond: R0 0.

The runaway reaction is now estimated to occur 0.00025 time per year. For this example. If quantitative data are available. the tire failure is caused by either a defective tire or a worn tire. The tire failure is an intermediate event because it results from either a defective tire or a worn tire. a flat tire on an automobile is caused by two possible events. If a probabilistic computation is attempted. resulting in a huge event tree. The event tree for the modified process is shown in Figure 11-11. If an engineer is concerned about a particular consequence. mostly as a result of the successful experiences demonstrated by the nuclear industry. The additional safety function provides a backup in the event that the high-temperature alarm fails or the operator fails to notice the high temperature. The approach starts with a well-defined accident. This approach is becoming more popular in the chemical process industries. data must be available for every safety function in the event tree. driving over the road debris is a basic event because no further definition is possible. and intermediate events are events that can.the flat tire is caused by either debris on the road or tire failure.11-3 Fault Trees 491 system would automatically shut down the reactor in the event that the reactor temperature exceeds a fixed value. A fault tree for anything but the simplest of plants can be large. The emergency shutdown temperature would be higher than the alarm value to provide an opportunity for the operator to restore the coolant flow. The basic events are events that cannot be defined further. The flat tire is identified as the top event. Fault trees are a deductive method for identifying ways in which hazards can lead to accidents. It means that either of the input events will cause the output state to occur. or top event. The fishlike symbol represents the OR logic function. with a variety of computer programs commercially available to draw fault trees based on an interactive session. In one case the flat is due to driving over debris on the road. such as a nail. This is a substantial improvement obtained by the addition of a simple redundant shutdown system. The other possible cause is tire failure. This is perhaps the major disadvantage of event trees. and works backward toward the various scenarios that can cause the accident. shown in Figure 11-12. The two contributing causes are either basic or intermediate events. . or once every 400 years. Fortunately. an estimate can be made of the failure frequency. involving thousands of process events. An event tree begins with a specified failure and terminates with a number of resulting consequences. For instance. 11-3 Fault Trees Fault trees originated in the aerospace industry and have been used extensively by the nuclear power industry to qualify and quantify the hazards and risks associated with nuclear power plants. The flat tire example is pictured using a fault tree logic diagram. there is no certainty that the consequence of interest will actually result from the selected failure. The difficulty is that for most real processes the method can be extremely detailed. Similarly. The circles denote basic events and the rectangles denote intermediate events. this approach lends itself to computerization. The event tree is useful for providing scenarios of possible failure modes. This is used most successfully to modify the design to improve the safety. As shown in Figure 11-12.

Define the unallowed events. For reasonably complex chemical processes a number of additional logic functions are needed to construct a fault tree. whereas an event such as "leak in valve7'is too specific. Define precisely the top event. lightning. 4. Events such as "high reactor temperature" or "liquid level too high" are precise and appropriate. The INHIBIT function is useful for events that lead to a failure only part of the time. These are events that are unlikely or are not under consideration at the present. Define the existing event. What components are to be considered in the fault tree? . Define the physical bounds of the process. What conditions are sure to be present when the top event occurs? 3. For instance. and hurricanes. a number of preliminary steps must be taken: 1. This could include wiring failures. Events in a fault tree are not restricted to hardware failures. driving over debris in the road does not always lead to a flat tire. The AND logic function is important for describing processes that interact in parallel. and environmental factors. A detailed list is given in Figure 11-13. human.Chapter 11 Risk Assessment +0 Tire Failure Debr is r? Flat Tire Top E v e n t Figure 11-12 A fault tree describing the various events contributing to a flat tire. 2. They can also include software. The INHIBIT gate could be used in the fault tree of Figure 11-12 to represent this situation. tornadoes. Before the actual fault tree is drawn. Events such as "explosion of reactor" or "fire in process" are too vague. This means that the output state of the AND logic function is active only when both of the input states are active.

occurrence o f a l l 6 OR G a t e : The resulting output event o f any requires the individual occurrence event. event that needs no A fault further definition. 0 INTERMEDIATE E v e n t : An e v e n t that results from t h e o f other i n t e r a c t i o n o f a number events. Will the analysis consider just a valve. input OO 0 ~ ~ h i b Condition iI N H I B I T E v e n t : ~ The the output input event occurs w i l l occur if and t h e inhibit event BASIC E v e n t : occure. EXT ERNAL E v e n t : An e v e n t condition that i a a boundary fault tree. draw the top event at the top of the page. 5. a sheet Figure 11-13 The logic transfer components used in a fault tree. or will it be necessary to consider the valve components? The next step in the procedure is to draw the fault tree. Label it as the top event to avoid confusion later when the fault tree has spread out to several sheets of paper. 0 0 UNDEVELOPED E v e n t : An event that due t o cannot be developed suitable further lack o f informat ion. t o the TRANSFER S y m b o l s : Used t o t r a n e f e r i n t o and out o f the f a u lt o f tree paper. What valves are open or closed? What are the liquid levels? Is this a normal operation state? 6. Define the level of resolution. Define the equipment confiGration. . First.11-3 Fault Trees 493 AND G a t e : The r e s u l t i n g requires the output event simultaneous input events.

Continue developing the fault tree until all branches have been terminated by basic. 1. Example 11-5 Consider again the alarm indicator and emergency shutdown system of Example 11-2. If the new events cannot be related to the top event by a single logic function. they must be connected to the top event by an AND gate. Determining the Minimal Cut Sets Once the fault tree has been fully drawn. 4. Remember. Solution The first step is to define the problem. or external events. undeveloped. basic. undeveloped. Two events must occur for overpressuring: failure of the alarm indicator and failure of the emergency shutdown system. a number of computations can be performed. reactor feed flowing. Physical bounds: The equipment shown in Figure 11-5. The emergency shutdown system can fail by a failure of either pressure switch 2 or the solenoid valve. determine the major events that contribute to the top event. Level of resolution: Equipment as shown in Figure 11-5. Write these down as intermediate. 3. The first computation determines the minimal cut sets (or min cut sets). basic. the purpose of the fault tree is to determine the individual event steps that must occur to produce the top event. Equipment configuration: Solenoid valve open.494 Chapter 11 Risk Assessment Second. These events must occur together so they must be connected by an AND function. they must be connected by an OR gate. If these events are related in parallel (all events must occur in order for the top event to occur). the new events are probably improperly specified. electrical failures. hurricanes. The alarm indicator can fail by a failure of either pressure switch 1or the alarm indicator light. All intermediate events must be expanded. Now consider any one of the new intermediate events. Top event: Damage to reactor as a result of overpressuring. Then decide which logic function represents the interaction of these newest events. Unallowed events: Failure of mixer. tornadoes. These must be connected by OR functions. The top event is written at the top of the fault tree and is indicated as the top event (see Figure 11-14). What events must occur to contribute to this single event? Write these down as either intermediate. 2. If these events are related in series (any event can occur in order for the top event to occur). Existing event: High process pressure. The complete fault tree is shown in Figure 11-14. 6. These must also be connected by an OR function. or external events on the sheet. undeveloped. electrical storms. Draw a fault tree for this system. or external events on the tree. wiring failures. The minimal cut sets are . 5.

the minimal cut sets are ordered with respect to failure probability. In general. 15. For instance.11-3 Fault Trees 495 Overpressur ing o f Reactor T~~ E~~~~ P = 0. The mi imal cut sets are useful for determining the various ways in which a top event could occur. The minimal cut sets are determined using a procedure developed by Fussell and V e ~ e l y .1648 R = 0. Vesely. "% 4J. the various sets of events that could lead to the top event.8352 P 0. The different unique sets of events leading to the t event are the minimal cut sets. E. the top event could occur through a variety of different combinations of events. "A New Methodology for Obtaining Cut Sets for Fault Trees. a set involving human interaction is more likely to fail than one involving hardware alone. Fussell and W. Based on these simple rules. a set involving just two events is more likely than a set involving three." Transactions of the American Nuclear Society (1972). . B. The higher probability sets are examined carefully to determine whether additional safety systems are required.0702 I F a i l u r e o f Alarm Indicator I F a i l u r e o f Emergency Shutdown P = 0.4258 w w w w Figure 11-14 Fault tree for Example 11-5. Some of the mimimal cut sets have a higher probability than others. Similarly. ~ The procedure is best described using an example.

Logic gate A in Figure 11-14 has two inputs: one from gate B and the other from gate C. First. Because gate B is an OR gate. . This is shown in Figure 11-14. gate A is replaced by gates B and C: Gate B has inputs from event 1 and event 2. Be sure to copy the 1 from the other column of the first row: Finally. replace gate C in the first row by its inputs. This generates a fourth row: The cut sets are then This means that the top event occurs as a result of any one of these sets of basic events. and then create a second row below the first. replace gate B by one of the inputs. Copy into this new row all the entries in the remaining column of the first row: Note that the C in the second column of the first row is copied to the new row. Because gate A is an AND gate. replace gate C in the second row by its inputs.496 Chapter 11 Risk Assessment Example 11-6 Determine the minimal cut sets for the fault tree of Example 11-5. The first logic gate below the top event is written: AND gates increase the number of events in the cut sets. whereas OR gates lead to more sets. replace C by basic event 3 and then create a third row with the other event. Next. gate B is replaced by adding an additional row below the present row. Solution The first step in the procedure is to label all the gates using letters and to label all the basic events using numbers. Because gate C i s also an O R gate.

This procedure approaches the exact result only if the probabilities of all the events are small.3 or 1. This approach assumes that the probability crossproduct terms shown in Equation 11-10 are negligible. On other occasions the sets might include supersets.0169 P(2 AND 3 ) = (0. 2. Then the necessary computations are performed across the various logic gates. The failure probabilities for the basic events were obtained from Example 11-2. The supersets are eliminated to produce the minimal cut sets. Sometimes a set might be of the following form: 1.13) = 0.2 This is reduced to simply 1. The minimal cut sets represent the various failure modes. This is accomplished in two ways.0799 . For Example 11-6 events 1 . consider The second and third sets are supersets of the first basic set because events 1 and 2 are in common. and undeveloped events are written on the fault tree. Quantitative Calculations Using the Fault Tree The fault tree can be used to perform quantitative calculations to determine the probability of the top event. With the first approach the computations are performed using the fault tree diagram itself. The results of this procedure are shown in Figure 11-14. 3 or 2.0442 P(2 AND 4 ) = (0.13)(0. Remember that probabilities are multiplied across an AND gate and that reliabilities are multiplied across an OR gate.04)(0.13) = 0.13)(0.0052 P(l AND 4 ) = (0.2. For this case 4 P ( l AND 3 ) = (0.0136 Total 0.34) = 0.4 or 2.04)(0. The other procedure is to use t e minimal cut sets.1 1 -3 Fault Trees 497 The procedure does not always deliver the minimal cut sets. INHIBIT gates are considered a special case of an AND gate.34) = 0. external. The failure probabilities of all the basic. The symbol P represents the probability and R represents the reliability.4 could cause the top event. In general. For this example there are no supersets. To estimate the overall failure probability. this result provides a number that is larger than the actual probability. The computations are continued in this fashion until the top event is reached. For instance. the probabilities from the cut sets are added together.

measured in years. These probabilities are not usually known or are not known accurately. and calculating failure probabilities. This is opposed to the event tree approach. Furthermore. This top event is selected by the user to be specific to the failure of interest. the entire fault tree procedure enables the application of computers. Relationship between Fault Trees and Event Trees Event trees begin with an initiating event and work toward the top event (induction). Advantages and Disadvantages of Fault Trees The main disadvantage of using fault trees is that for any reasonably complicated process the fault tree will be enormous. This is an approximate result. Reference libraries containing failure probabilities for various types of process equipment can also be included. For Example 11-6 all the cut set probabilities were added. The different trees generally predict different failure probabilities. Fault trees are also used to determine the minimal cut sets.0702 obtained using the actual fault tree. then failure probabilities are needed for all the events in the fault tree. A major advantage of the fault tree approach is that it begins with a top event. This. This inexact nature of fault trees is a considerable problem. where the events resulting from a single failure might not be the events of specific interest to the user. Finally. The minimal cut sets provide enormous insight into the various ways for top events to occur. Fault trees of this size require a considerable amount of time. increases the reliability of the system significantly. A leaking valve is a good example of a partial failure. the approach assumes that a failure of one component does not stress the other components. More complete fault trees are usually developed by more experienced engineers. because the cross-product terms were neglected. the developer of a fault tree can never be certain that all the failure modes have been considered. Fault trees developed by different individuals are usually different in structure. determining the minimal cut sets. . to complete. Fault trees involving thousands of gates and intermediate events are not unusual. Also. If the fault tree is used to compute a failure probability for the top event. as shown by Equation 11-10. For small probabilities the cross-product terms are negligible and the addition will approach the true result. Some companies adopt a control strategy to have all their minimal cut sets be a product of four or more independent failures. Fault trees begin with a top event and work backward toward the initiating events (deduction). Software is available for graphically constructing fault trees. The cut sets are related to each other by the OR function." that aparticular item of hardware does not fail partially. resulting in a change in the component failure probabilities. Fault trees also assume that failures are "hard.- 498 Chapter 11 Risk Assessment This compares to the exact result of 0. of course.

from its initiating causes all the way to its final outcome. General description The initiating events are the causes of the incident. 2000). Both are used together to produce a complete picture of an incident. Guidelines for Chemical Process Quantitative Risk Analysis. Maximum benefits result when QRAs are used at the beginning "CPS. fault trees. Risk is usually described graphically. . however. In both methods the frequency of the release is determined using a combination of event trees. and the consequences of the exposure. (New York: Center for Chemical Process Safety. All companies decide their levels of acceptable risk and unacceptable risk. Other methods are sometimes used. 11-4 QRA and LOPA Risk is the product of the probability of a release. The two methods are related in that the top events for fault trees are the initiating events for the event trees. as shown in Figure 11-15. The actual risk of a process or plant is usually determined using quantitative risk analysis (QRA) or a layer of protection analysis (LOPA). 2d ed. or an appropriate adaptation. Quantitative Risk Analysis5 Q R A is a method that identifies where operations. Q R A and LOPA are the methods that are most commonly used. engineering.11-4 QRA and LOPA 499 Not acceptable Frequency Figure 11-15 of risk. or management systems can be modified to reduce risk. th&obability of exposure. AICHE. and the top events are the final outcomes. Probabilities and frequencies are attached to these diagrams. The complexity of a QRA depends on the objectives of the study and the available information.

maximum frequency of a specific category of a consequence. 2001) (in press). (New York: Center for Chemical Process Safety. The major steps of a QRA study include 1. D. estimating the risk by combining the impacts and frequencies. This method includes simplified methods to characterize the consequences and estimate the frequencies. then the application of LOPA methods may be more appropriate. and property. frequency of fires. QRAs are used to evaluate potential risks when qualitative methods cannot provide an adequate understanding of the risks. environment. Thus the results of a LOPA should always be more conservative than those from a QRA. for example. and human intervention. estimating the potential incident frequencies using event trees and fault trees. The criteria may include frequency of fatalities. Layer of Protection Analysis: Simplified Process Risk Assessment. 6CCPS. estimating the incident impacts on people. and 5. QRA is a relatively complex procedure that requires expertise and a substantial commitment of resources and time. The protection layers may include inherently safer concepts. 3. such as dikes or blast walls. Crowl. Layer of Protection Analysis6 LOPA is a semi-quantitative tool for analyzing and assessing risk. 2. safety instrumented functions. . AICHE. Various layers of protection are added to a process. ed. to lower the frequency of the undesired consequences. the basic process control system. and the effectiveness of the protection layers is also approximated. The results of both methods need to be used cautiously. passive devices. However. A. 4. then a full QRA may be justified. evaluating the incident consequences (the typical tools for this step include dispersion modeling and fire and explosion modeling). the results of QRA and LOPA studies are especially satisfactory when comparing alternatives. the frequencies are estimated. In some instances this complexity may not be warranted. and required number of independent layers of protection for a specific consequence category. In LOPA the consequences and effects are approximated by categories. The combined effects of the protection layers and the consequences are then compared against some risk tolerance criteria. The approximate values and categories are selected to provide conservative results. In general. and recording the risk using a graph similar to Figure 11-15. QRA is especially effective for evaluating alternative risk reduction strategies. If the LOPA results are unsatisfactory or if there is any uncertainty in the results. This concept of layers of protection is illustrated in Figure 11-16.- - - - - - - 500 Chapter 11 Risk Assessment of a project (conceptual review and design phases) and are maintained throughout the facility's life cycle. active devices. Individual companies use different criteria to establish the boundary between acceptable and unacceptable risk. defining the potential event sequences and potential incidents. The QRA method is designed to provide managers with a tool to help them evaluate the overall risk of a process. such as relief valves.

sufficient layers must be added to the process to reduce the risk to an acceptable level. depending on the process complexity and potential severity of an accident. A scenario may require one or many layers of protection. identifying the initiating event for the scenario and estimating the initiating event frequency (a simple method is described later). 2. The major steps of a LOPA study include 1. Note that for a given scenario only one layer must work successfully for the consequence to be prevented. . identifying a single consequence (a simple method to determine consequence categories is described later). identifying an accident scenario and cause associated with the consequence (the scenario consists of a single cause-consequence pair). Figure 11-16 does not include all possible layers of protection. The primary purpose of LOPA is to determine whether there are sufficient layers of protection against a specific accident scenario. Because no layer is perfectly effective. 3. As illustrated in Figure 11-16.11-4 QRA and LOPA 501 Figure 11-16 Layers of protection to lower the frequency of a specific accident scenario. many types of protective layers are possible. however.

000$100.000. . >10.000-lb release Category 5 Category 5 Category 5 Category 5 Category 3 Release characteristic Extremely toxic above BP1 Extremely toxic below BP or highly toxic above BP Highly toxic below BP or flammable above BP Flammable below BP Combustible liquid Spared or nonessential equipment Plant outage <1 month Category 3 Category 2 $10. Vessel rupture.000.000 gal.000 Category 2 Category 2 Category 2 $0 -$10. 100-300 psig >300 psig Category 4 Category 4 >$10.000$1.000.000-lb release Category 5 Category 5 Category 5 Category 4 Category 2 10.000 gal.000 Category 4 Vessel rupture.000 Category 3 Plant outage 1-3 months Plant outage >3 months Category 4 Category 4 $1.000-lb release >100.Table 11-2 Consequence size - Semi-Quantitative Consequences Categorization 1-10-lb release Category 3 Category 2 Category 2 Category 1 Category 1 Category 2 Category 1 Category 2 Category 1 Category 2 Category 3 Category 4 Category 3 Category 2 Category 4 Category 3 Category 5 Category 4 Category 5 Category 5 10-100-lb release 100-1.000$10.000 Category 5 Category 5 Category 5 Consequence characteristic Mechanical damage to large main product plant Mechanical damage to small by-product plant Consequence cost (US dollars) Categories 'BP. 3. atmospheric boiling point.000-10.00010.000100.000.000 Category 1 Category 4 Category 3 $100.000-lb release 1.

and the consequences are characterized with a category. The frequencies are also adjusted (reduced) to include the benefits of preventive maintenance. This procedure is repeated for other consequences and scenarios. In a LOPA study the consequences are estimated using one of the following methods: (1) semi-quantitative approach without the direct reference to human harm. This can occur through a variety of incidents. Adjust the failure frequency to include the probabilities of failure on demand (PFDs) for each independent layer of protection. It also identifies problems. When using the semi-quantitative method. 2. evaluating the risk for acceptability (if unacceptable. In a QRA study the consequences of these releases are quantified using dispersion modeling and a detailed analysis to determine the downwind consequences as a result of fires.and 7. for example. 5. Consequence The most common scenario of interest for LOPA in the chemical process industry is loss of containment of hazardous material. One of the less rigorous methods includes the following steps: 1. as shown in Table 11-2. several methods can be used to determine the frequency. for example. identifying the protection layers available for this particular consequence and estimating the probability of failure on demand for each protection layer.1 1 -4 QRA and LOPA 503 4. plotting the consequence versus the consequence frequency to estimate the risk (the risk is usually shown in a figure similar to Figure 11-15). a reactor failure frequency is divided by 12 if the reactor is used only 1 month during the entire year. such as a QRA. See footnote 6 for the detailed methods. additional layers of protection are required). the quantity of the release is estimated using source models. This is an easy method to use compared with QRA. (2) qualitative estimates with human harm. Frequency When conducting a LOPA study. or toxicity. then its failure frequency is divided by 4. combining the initiating event frequency with the probabilities of failure on demand for the independent protection layers to estimate a mitigated consequence frequency for this initiating event. . a control system is given preventive maintenance 4 times each year. A number of variations on this procedure are used. a ruptured pipeline. a gasket failure. such as a leak from a vessel. it clearly identifies problems that may need additional work. 6. Although the method is easy to use. 3. Adjust this frequency to include the demand. Determine the failure frequency of the initiating event. and (3) quantitative estimates with human harm. or release from a relief valve. which may be deemphasized because the consequences are insignificant. If. explosions.

respectively. The IPL is auditable.504 Chapter 11 Risk Assessment Table 11-3 Typical Frequency Values Assigned to lnitiating Events' Example of a Frequency range value chosen by a from literature company for use in LOPA (per yr) (per yr) Initiating event Pressure vessel residual failure Piping residual failure. 2. Some PFDs recommended by CCPS (see footnote 6) for screening are given in Tables 11-4 and 11-5. There are three rules for classifying a specific system or action of an IPL: 1. not fatigued) to 10-4/lift 10-3 to 10-2 to 1 to 10-2 lo-' to 10-2 1 to 10-2 1 to 10-2 1 to lo-' lo-' to 10-2 to lo-" 10-QO opportunity 10-1 to 10-31 opportunity 1 X 10-4(/lift) 1 x lo-" 1x 1 x lo-' 1 x 10-I 1 x 10-I 1 x lo-' 1 x 10-I 1 x lo-' 1 x 10-2 1 x lo-" (/opportunity) 1 x 10-2 (Iopportunity) IIndividual companies choose their own values.full breach Piping leak (10% section). 100 rn. 3. unstressed. consistent with the degree of conservatism or the company's risk tolerance criteria. well trained. The failure frequencies for the common initiating events of an accident scenario are shown in Table 11-3. . and documentation. etc. The IPL functions independently of the initiating event and the components of all other IPLs that are used for the same scenario. vehicle. 100 m Atmospheric tank failure Gasket /packing blowout Turbineldiesel engine overspeed with casing breach Third-party intervention (external impact by backhoe. The common practice is to use a PFD of lo-' unless experience shows it to be higher or lower. The PFD for each independent protection layer (IPL) varies from 10-' to for a weak IPL and a strong IPL. The IPL is effective in preventing the consequence when it functions as designed. the PFD of the IPL must be capable of validation including review.) Crane load drop Lightning strike Safety valve opens spuriously Cooling water failure Pump seal failure Unloading/loading hose failure BPCS instrument loop failure Regulator failure Small external fire (aggregate causes) Large external fire (aggregate causes) L O T 0 (lock-out tag-out) procedure failure (overall failure of a multiple element process) Operator failure (to execute routine procedure. that is. testing. Failure rates can also be greatly affected by preventive maintenance routines.

Crowl. and maintained. installed. If properly implemented. and maintenance procedures) Reduces the frequency of large consequences (widespread spill) of a tank overfill. D. f! is the initiating event frequency for the initiating event i. ed. Preventsoverpressure Reduces rate of heat input and provides additional time for depressurizing. rupture. can eliminate scenarios or significantly reduce the consequences associated with a scenario If properly designed. When there are multiple scenarios with the same consequence. Simplified Process Risk Assessment: Layer of Protection Analysis. A. etc. as described previously. inspections. and PFDij is the probability of failure of the jth IPL that protects against the specific consequence and the specific initiating event i. (New York: American Institute of Chemical Engineers. each scenario is evaluated individually using Equation 11-30. spill. The PFD is usually lo-'. can eliminate the potential for flashback through a piping system or into a vessel or tank Passive lPLs Dike PFDs from industry' 1 X lo-z to 1 x lo-' 1X to 1 x lo-3 1X 1x 1X 1x 1X 1x PFDs from CCPS1 1X Underground drainage system Open vent (no valve) Fireproofing Blast wall or bunker Inherently safer design Flame or detonation arrestors 1 X lo-z to lo-3 lo-' to lo-3 lo-' to 1X 1 X lo-' 1X lo-3 1 X lo-z 1 x lo-' to 1 x lo-6 1 x 10-I to 1 x lo-3 1x 'CCPS. rupture. buildings. etc. The frequency of a consequence of a specific scenario endpoint is computed using where ff is the mitigated consequence frequency for a specific consequence C for an initiating event i. Reduces the frequency of large consequences of an explosion by confining blast and by protecting equipment. spill. etc Reduces the frequency of large consequences (widespread spill) of a tank overfill.1 1 -4 QRA and LOPA 505 Table 1 1-4 PFDs for Passive lPLs Comments (assuming an adequate design basis. etc. fire fighting. The frequency of the consequence is subsequently determined using . 2001) (in press).

Simplified Process Risk Assessment: Layer of Protection Analysis. Example 11-7 Determine the consequence frequency for a cooling water failure if the system is designed with two IPLs. 2001). Solution The frequency of a cooling water failure is taken from Table 11-3. 2001) for life-cycle requirements and additional disc~ssion. 1998) and IEC 6151 1 (IEC." Simple well-documented action with clear and reliable indications that the action is required.The human response PFD is lo-' and the PFD for the BPCS is 10-I. inspections. training.506 Chapter 11 Risk Assessment Table 11-5 PFDs for Active lPLs and Human Actions Comments [assuming an adequate design basis. (New York: American Institute of Chemical Engineers. See IEC 1X (1998. Parts 1-7. 1 to 1 X 10-' 1 x lo-' 1 X 10-' to 1X 1 X lo-2 'CCPS. and testing procedures from (human action)] industry Active IPL or human action Relief valve PFDs from CCPS1 Rupture disc Basic process control system (BPCS) Safety instrumented functions (interlocks) Human action with 10 min response time Human action with 40 min response time 1 x 10-I to 1 x lo-2 Prevents system from exceeding specified over1 x lo-5 pressure. Parts 1-3. IEC 61511.' See IEC 61508 (IEC. The PFDs are estimated from Tables 11-4 and 11-5. that is. 'IEC (1998). where f: is the frequency of the Cth consequence for the ith initiating event and I is the total number of initiating events for the same consequence. 'IEC (2001). Crowl. The consequence frequency is found using Equation 11-30: . f : = lo-'. and maintenance procedures PFDs (active IPLs) and adequate documentation. 2001) (in press). Simple well-documented action with clear and reliable indications that the action is required. Functional Safety Instrumented Systems for the Process Industry Sector. Functional Safety of ElectricaNElectronic/Programmable Electronic Safety-related Systems.~. The IPLs are human interaction with 10-min response time and a basic process control system (BPCS). 1 X 10-' 1 X lo-' to Can be credited as an IPL if not associated with the initiating event being considered. (Draft in Process). D. 1 x lo-' to 1x Prevents system from exceeding specified overpressure. Effectiveness of this device is sensitive to service and experience. IEC 61508. Geneva: International Electrotechnical Commission. Geneva: International Electrotechnical Commission. A.2. ed. Effectiveness of this device can be 1 x lo-' sensitive to service and experience.

"A New Methodology for Obtaining Cut Sets for Fault Trees. Rijnsdorp. Health and Eizvironmental Risk Analysis: Fundamentals with Applications (Upper Saddle River.. including the sensor. Louvar. final control element. E. MI: Ann Arbor Science. 1982). (London: Butterworths. a relayed shutdown system). and using a deenergized trip system (i. Many companies find that they have a limited number of SIL3 systems because of the high cost normally associated with this architecture. J. SIL2 (PFD = to These SIFs are typically fully redundant. 3. 2d ed. D. testing the system components at specific intervals to reduce the probability of failures on demand by detecting hidden failures. ch. and Protection (Ann Arbor. and final control element. . 15. a single final control element. and requires periodic proof testing. and the system requires careful design and frequent validation tests to achieve the low PFD figures. Suggested Reading CCPS. Roffel and J.the failure frequency is determined easily by using LOPA methods.e. Loss Prevention in the Process Industries. SIL3 (PFD = loe3to lop4):SIL3 systems are typically fully redundant. B. P. Control. E. Process Dynamics. 19. logic solver. 1992). Guidelines for Hazard Evaluation Procedures. logic solver." Transactions ofthe American Nuclear Society (1972). Louvar and B. using multiple sensors with voting systems and redundant final control elements. and requires periodic proof testing. B. 1996). a single logic solver. F. NJ: Prentice Hall PTR. (New York: American Institute of Chemical Engineers. Lees. The concept of PFD is also used when designing emergency shutdown systems called safety instrumented functions (SIFs). 2d ed. SILl (PFD = lo-' to lop2):These SIFs are normally implemented with a single sensor. 2.Suggested Reading 507 As illustrated in Example 11-7. 1998). Guidelines for Consequence Analysis of Chemical Releases (New York: American Institute of Chemical Engineers. A SIF achieves low PFD figures by using redundant sensors and final redundant control elements. There are three safety integrity levels (SILs) that are generally accepted in the chemical process industry for emergency shutdown systems: 1. including the sensor. Fussell and W. F. J. Vesely. 1999).

b. estimate the number of overflows expected per year. Problems 11-1. Determine an expression for the probability of the top event in terms of the component failure probabilities. Overfilling of storage tanks is a common problem in the process industries. a. c. The storage tank system shown in Figure 11-18is used to store process feedstock. Given the fault tree gates shown in Figure 11-17 and the following set of failure probabilities: Component Failure probability a. 11-2. To prevent overfilling. Use the following data: . Determine the minimal cut sets. Given that the level indicator fails 4 timeslyr.508 Chapter 11 Risk Assessment Figure 11-17 Fault tree gates. Use both the expression of part a and the fault tree itself. the storage tank is equipped with a high-level alarm and a high-level shutdown system. Compute a value for the failure probability of the top event. Develop an event tree for this system using the "failure of level indicator" as the initiating event. The high-level shutdown system is connected to a solenoid valve that stops the flow of input stock.

The protection device fails once every 25 yr. A starter is connected to a motor that is connected to a pump.01 0. The starter fails once in 50 yr and requires 2 hr to repair. and the MTBC. Derive Equation 11-22. what is the new probability of the top event? 11-5. Calculate the unavailability. The problem of Example 11-5 is somewhat unrealistic in that it is highly likely that the operator will notice the high pressure even if the alarm and shutdown systems are not functioning. Develop a fault tree for the top event of "storage tank overflows. System High-level alarm Operator stops flow High-level switch system Failuresldemand 0. . Determine the minimal cut sets. Compute the MTBC for a high-level episode and a failure in the shutdown system.01 b." Use the data in Table 11-1 to estimate the failure probability of the top event and the expected number of occurrences per year. What are the most likely failure modes? Should the design be improved? 11-3.1 0. A reactor experiences trouble once every 16 months. Determine the overall failure frequency. the probability that the system will fail in the coming 2 yr. and the unavailability for this system. Inspection takes place once every month. The pump fails once per 10 yr and requires 4 hr to repair. assuming that a high-level episode occurs once every 6 months. 11-6. Compute the availability of the level indicator system and flow shutdown system for Problem 11-2. the frequency of dangerous coincidences. Show that for a process protected by two independent protection systems the frequency of dangerous coincidences is given by 11-7. the reliability. The motor fails once in 20 yr and requires 36 hr to repair. Draw a fault tree using an INHIBIT gate to include this situation. Determine the minimal cut sets. 11-8. If the operator fails to notice the high pressure in 1 out of 4 occasions. 11-4.Problems 509 Solenoid Flow meter Chart Valve Figure 11-18 Level control system with alarm. Assume a 1-month maintenance schedule.

4 P = 0.1 P = 0.E Top event 5 6 P = 0. . I Explosion I Chemical deflagration p = 0.1 faultlyr I I Flammable vapor 0.2 faultlyr flame p = 10 faultlyr p = Figure 11-20 Determine the MTBF of the top event.2 P = 0.3 Figure 11-19 Determine the failure characteristics of the top event.

and the MTBF for the top event of the system shown in Figure 11-21.Problems 511 11-9. -7 Overflow A Figure 11-21 Determine the failure characteristics of the top event. 11-11. reliability. p. Also list the minimal cut sets. and probability of failure of the top event of the system shown in Figure 11-19. Also show the minimal cut sets. 11-10. . Determine the MTBF of the top event (explosion) of the system shown in Figure 11-20. Determine P. R. failure rate. Compute the MTBF.

Determine the failure characteristics and the minimal cut sets for the system shown in Figure 11-22. someone turns off valve 7. 11-14. Determine the consequence frequency for a regular failure if the system is designed with three IPLs. 11-13. what will be the frequency if this regulator is given preventive maintenance once per month? . If a regulator has a consequence frequency of lop1failurelyr. 11-12. 11-15. draw the fault tree and determine the failure characteristics of the top event (vessel overflows). Using the system shown in Figure 11-23. draw the fault tree and determine the failure characteristics of the top event (vessel pressure exceeds MAWP). that is.respectively. 11-17. In this problem you have human intervention.and lop4. SIL2.512 Chapter 11 Risk Assessment Y Reactor explosion (T) Figure 11-22 Determine the failure characteristics of a reactor explosion. and SIL3 ratings with PFDs of lop3. when the alarm sounds. 11-16. Using the system shown in Figure 11-24. Determine the expected failure rates and MTBFs for control systems with SIL1.

Problems 513 Valve 8 opens at high pressure Valve 4 closes at high level Top event = vessel pressure exceeds MAWP Figure 11-23 A control system to prevent the pressure from exceeding the MAWP. Overflow I I I I 1 2 5 Worker closes valve 7 when level goes up Figure 11-24 Control system to prevent vessel overflow. .

If you have a plant with poor safety performance. c. Determine the MTBF for SIL1-3 systems. what steps would you take to improve the performance? 11-29. Using the rules for IPLs. b. Assume that a company decides to characterize an acceptable risk as a Category 1 failure every 2 years and a Category 5 failure every 1000 years. (Note: This is not the same as Problem 11-15.514 Chapter 11 Risk Assessment 11-18. Are the following scenarios acceptable or not: a. and respectively. What steps should be taken to decrease the MTBF of a critical control loop? . how many IPLs are needed to reduce this frequency to 11-22. lo-'. Potential facility damage of $1. describe what you would do to establish which parts of the plant need special attention. 11-20. 11-24. If your lock-outltag-out procedure has a failure frequency of what measures could be taken to reduce this frequency? 11-25. If you have a complex and dangerous plant. Release of 1000 pounds of phosgene. describe the process for determining the frequency of this specific event. b. Release of 1000 pounds of isopropanol at 75OF. What would you do to move the unacceptable scenarios into the acceptable region? b. What consequence categories do the following scenarios have? a.) 11-27. 11-26. Category 4 every 100 years. 11-28. per opportunity. Determine the consequence frequency of a cooling water system if it is used only 2 months every year and if it is given preventive maintenance each month of operation. Category 2 every 50 years. list four protective layers that are clearly IPLs. if they have PFDs of lo-'. Was the analysis of Problem 11-19 acceptable? 11-21.000. Using the results of Problem 11-19. a. If a specific consequence has two initiating events that give the same consequence.000. 11-19. 11-23. What redundant components should be added to a critical measurement and controller? 11-30. If a plant has a consequence frequency of lo-'.

T 12-1 Learning from Accidents Every member of an investigation team learns about problems that precipitate accidents. An important principle in safety states that the causes of accidents are visible the day before the accident. are commonly shared within the chemical industry. These causes are visible to professionals who "see" deficiencies. including learning from accidents. the investigation process. This vision (knowledge or awareness) is developed by the study and development of accident and nearmiss investigations. and aids for recommendations. investigation summary. Many professionals believe that this sharing of information about accidents has been a major contributor to the steady improvement in safety performance. including detailed descriptions and recommendations. layered investigations. This new knowledge helps every team member avoid similar situations in the future. Accident investigations. In this chapter we cover the more important techniques. . close calls) provides opportunities to learn how to prevent similar events in the future. In recent years important techniques have been developed for improving the effectiveness of investigations. aids for diagnosis.C H A P T E R 1 2 Accident Investigations he investigation of accidents and near misses (that is. If the investigation is appropriately reported. many others will also benefit.

(2) accumulating relevant facts. (3) analyzing the facts and developing potential causes of the accident. however. Investigation results are used to change hazardous practices and procedures and to develop management systems to use this new knowledge on a long-term and continuous basis. this older method developed recommendations that were relatively superficial. similar to a brainstorming session to develop new applications for a common house brick. 'T. . often leads to new conclusions and recommendations. A deeper analysis of the facts. It is much easier to correct minor problems before serious accidents occur than to correct them after they are manifested in major losses. Accident investigations are designed to enhance learning. Second layer: recommendations to avoid the hazards. New and interesting applications will continue to surface." Hydrocarbon Processing (November 1979). Kletz. Kletz. and one or two technical recommendations resulted. Unfortunately. The number of relevant facts accumulated in an accident investigation is usually limited. This recommended deeper analysis identifies underlying causes of the accident that are analyzed to develop a multilayered solution to the problem layered recommendations.516 Chapter 12 Accident Investigations This concept is also important for reporting minor accidents or near misses. (4) studying the system and operating methods relevant to the potential causes of the accident. 373-382. ( 5 ) developing the most likely causes. Good investigations help organizations use every accident as an opportunity to learn how to prevent future accidents. pp.' It is a technique that significantly improves the commonly used older methods. faultfinding creates an environment that is not conducive to learning. Minor accidents and near misses are excellent opportunities to obtain "free chances" to prevent larger accidents from occurring in the future. "Layered Accident Investigations. for example. The newer and better method includes a deeper analysis of the facts and additional levels or layers of recommendations. Further investigation usually cannot uncover additional facts. Third layer: recommendations to improve the management system. 12-2 Layered Investigations The important concept of layered investigations is emphasized by T. Their evidence supported their conclusions. and (7) using an investigation style that is fact-finding and not faultfinding. most accidents are investigated in this style. The fundamental steps in an investigation include (1) developing a detailed description of the accident. Learning from Accidents in Industry (Boston: Butterworths. (6) developing recommendations to eliminate recurrence of this type of accident. Kletz. According to Kletz. Kletz emphasized an extra effort to generate three levels of recommendations for preventing and mitigating accidents: First layer: immediate technical recommendations. This deeper analysis is. 1988). Older investigation methods identified only the relatively obvious causes of an accident. and T.

Solution The facts uncovered by an investigation team are 1. 7. Add more lifeguards. such as paint pool depths at the edges of the pool and add more lifeguards. Reduce the number of swimmers. almost all recommendations can be implemented without difficulty. 3. In this particular example. Approximately 100 children. ~ Example 12-1 Illustrate the layered investigation process to develop underlying causes of a municipal pool accident. ranging between 5 and 16 years old. Assign the supervisor to make formal (documented) audits on a regular basis. An older child unknowingly pushed a 5-year-old into the deep water. Paint pool depths at the pool edges. Second layer recommendations: Avoiding the hazard 1. 5. 3. 3.12-2 Layered Investigations 517 To fully utilize this layered technique. 2. Facts about the accident that support conclusions at all three levels are accumulated. the pool was overly crowded. Layered recommendations are the result of uncovering the underlying causes of the accident: First layer recommendations: Immediate technical recommendations 1. 4. 5. These technical improvements and new management systems will prevent future drown- . Third layer recommendations: Improving the management system 1. Add a roving lifeguard to monitor and control pool behavior. Zone pool to keep smaller children at shallow end of pool. Give all new swimmers (especially young children) a pool orientation. the younger child did not know how to swim. the older child was engaged in horseplay. and the pool did not offer swimming lessons. 2. An old-style accident investigation report would include only one or two recommendations. 6. Train lifeguards to alert supervision of observed potential problems. the pool did not have deep and shallow markings. The pool was relatively crowded. and the 5-year-old slipped under the water without being noticed by others. 2. Prohibit horseplay. including the lifeguard. the pool did not have an orientation program. the investigation process is conducted with an open mind. the lifeguard had many blind fields of vision. were in and around a pool (3 ft to 9 ft deep). Add swimming lessons for all age groups. 2. A drowning accident occurred during an open swim period. 4.

. One approach that can be used for most accidents is described here and shown in Table 12-2. 1988).)? How can leaks (etc. D. ~ 2Howard H. Safety and Accident Prevention in Chemical Operations (New York: Wiley. decomposed.p.518 Chapter 12 Accident Investigations Table 12-1 Questions for Layered Accident Investigations1 (1) What equipment failed? How can failures be prevented or made less likely? How can failures or approaching failure be detected? How can failures be controlled or consequences minimized? What does the equipment do? What other equipment can be used instead? What could we do instead? (2) What material leaked (exploded. This example also illustrates the value of having an open mind during the investigation. eds. 659-680. Learning from Accidents in Industry (Boston: Butterworths. A set of questions designed to help accident investigators find less obvious ways to prevent accidents is shown in Table 12-1. 153. Fawcett and William S.. 1982). pp. it is an adaptation of a process recommended by A. synergistic. 12-3 Investigation Process Different investigators use different approaches to accident investigations. Wood.) be prevented? How can leaks or approaching leak be detected? What does material do? Can volumes of material be reduced? What materials can be substituted? What could we do instead? (3) Which people could have performed better? What could they have done better? How can we help them to perform better? (4) What is the purpose of the operation involved in the accident? Why do we do this? What could we do instead? How else could we do it? Who else could do it? When else could we do it? - 'Trevor Kletz. C r a ~ e n . etc. The team approach of questioning and answering is especially important because the supportive. which is a requirement for uncovering underlying causes. and feedback approach by team members gives results that are always greater than the sum of the parts. ings and also prevent other types of accidents in this pool environment.

(1) Investigation team (2) Brief survey (3) Set objectives and delegate responsibilities (4) Preincident facts (5) Accident facts (6) Research and analyses (7) Discussion. (5) discussion. Items 5-7 appropriately contain the opinions of the authors (investigation team). In general. and hypothesize the sequence of events. Experience and affiliation are proportional to the magnitude of the accident. Flow sheets. photographs. These details. and technical. Keep report clear. and (7) layered recommendations. Record extent of damage. Make detailed examinations with photos. the format should be flexible and designed specifically to best explain the accident. and recommendations (8) Report The accident investigation report is the major result of the investigation. Resist development of potential causes to maintain momentum and objectivity while collecting data. inspections. conclusions. accurate. which is the major purpose of the investigation. (2) process description (equipment and chemistry). and interviews. Study (2) to (6) to develop conclusions and layered recommendations. the accident investigation report is a learning tool. 12-4 lnvestigation Summary The previously described accident investigation report is a logical and necessary result of an investigation. Initiate research-type experiments. are too focused for an average inquirer. This technical style allows readers to develop their own independent conclusions and recommendations. concise. (4) investigation results. . Develop accident investigation report. define the objectives and subobjectives of the investigation (the investigation team does this). Establish origin of accident and facts relevant to layered causes and recommendations. Make overview survey (maximum of 1 hr) to understand the type and value of the information needed to derive causes of the accident. It includes comprehensive details that are of particular interest to specialists. (6) conclusions. (3) incident description. Gather and organize preincident facts.12-4 Investigation Summary 519 Table 12-2 Step Accident lnvestigation Process Purpose Choose team as quickly as possible. The format may include the following sections: (1) introduction. procedures. Delegate responsibilities to team members with suggested completion times. As a result of these criteria. Do not smother key results. Based on (1) and (2). and analyze facts to clarify perplexing evidence. The accident investigation report is written using the principles of technical documentation. Items 1-4 are objective and should not include the authors' opinions. however. and data (interviews or recorded data) are used.

Kletz3 used a report format that summarizes the events and recommendations in a diagram. that is. Kletz. for example.520 Chapter 12 Accident Investigations Accident Summary Accident title: Major damage: Date: Location: Recommendations for ~reventiodmitiaation lst layer: Immediate technical recommendation 2nd layer: Recommendations for avoiding hazard 3rd layer: Recommendations for improving the management systems Events Major accident Events which precipitate accident Recommendations * * * * * Recommendations which will break the chain of events (layers 1. and on a long-term basis. These concepts are described in Example 12-2. Examples may be (1) a quar3T. This type of summary is shown in Figure 12-1. plant design procedures * Figure 12-1 Accident report summary. to break the link in the chain of events that led to the accident. . p. inspect. The illustrated format is similar to the one used by Kletz. 22. Management systems are designed to continuously. Learning from Accidents in Industry. It emphasizes underlying causes and layered recommendations. either prevent the accident or eliminate the hazardous conditions.2 & 3 Preaccident conditions Sequence of operations or steps to derive accident conditions * Recommendations to train. or change methods (layers 2 & 3) * * * * * Management influences (Decisions and practices) * * * * * Physical and management events to give preaccident conditions * * * * * * * Recommendations (and events) may address historical events. The third-layer recommendations shown in Figure 12-1emphasize the importance of management systems for preventing accidents.

Give periodic pool orientation and prohibit running and horseplay. . Safety and Accident Prevention. For this reason some experienced personnel are always assigned to investigation teams.12-5 Aids for Diagnosis 521 terly audit program t o ensure that recommendations are understood and used. I n most cases the evidence clearly supports one o r more causes. Add roving lifeguard. 668. eds. Park Z. Limit number of swimmers to improve visibility and control. Sometimes. Initiate periodic audit program to monitor adherence to rules and regulations. 5-year-oldfatality. Paint pool depths at edges of pool. Initiate rule to keep small children away from the deep region of the pool.. the evidence needs added 4Fawcettand Wood. and often they also make significant contributions through an open and probing discussion. (2) a semiannual orientation program to review and study accident reports. Initiate swimming lessons for all levels. o r (3) a checklist that is initiated by management and checked by operations on a daily basis. Recommendations for preventionlmitigation ~~~~~~ 1st layer: Immediate technical recommendation 2nd layer: Recommendations for avoiding hazard 3rd layer: Recommendations for improving the management systems Events Recommendations Drowning (5-yr-old) Child knocked in pool Crowded pool Horseplay Pool depth not marked No swimming lessons Add more lifeguards. 12-5 Aids for Diagnosis The data collected during an investigation are studied and analyzed to establish the causes of the accident and to develop recommendations to prevent a recurrence. p. Layered events and recommendations are developed primarily by experienced personnel. Example 12-2 Use the investigation described in Example 12-1to develop an investigation summary Solution Accident title: Major damage: Date: Location: Drowning in municipal pool. xx /xx /xx Detroit municipal pool. however. Inexperienced team members learn from the experienced personnel. Train lifeguards to watch for safety problems.

Lees. Loss Prevention in the Process Industries (Boston: Butterworths. For example. is approximately 1000°C. This phase of the investigation may require special techniques or aids to diagnosis to relate the evidence to specific causes. steel weakens at approximately 575°C. All aluminum products will therefore melt during fires. Therefore. and solvents. The pressure increases during deflagrations are approximately5 . "rank P. Aluminum and alloys of aluminum have very low melting points (660°C). Sometimes pits at the ends of conductors indicate high temperatures and vaporization of copper while arcing. If copper beads are found around electrical equipment. it may not be the primary source of the fire.522 Chapter 12 Accident Investigations analysis to uncover explanations. if the time of extinguishment is known and if the depth of char at various locations is known. 567. Fires The identification of the primary source of ignition is one of the major objectives of investigations. 1983). it may indicate that electrical arcs created temperatures greater than those observed in fires. plastic. and most woods burn at a rate of 1. as shown in the following discussion. such as wood. p. copper wire usually survives fires. . However. and deeper analysis in this regard should be avoided. The fire temperature for various materials. P2 P 2 P1 PI (12-1) . In this regard observations around the charred remains are helpful. The integrity of steelwork is not very useful evidence. therefore steelwork may be completely distorted. This evidence together with steelwork distortions is not useful.= 8 for hydrocarbon-air mixtures. Although this type of evidence indicates a source of ignition. the depth of wood charring is proportional to the duration of burning. then the region of the origin can be approximated. Further searching in this region may reveal possible causes of the fire.16 for hydrocarbon-oxygen mixtures. Explosions The classification of the explosion as either a deflagration or a detonation and the magnitude of the explosion may be useful for developing causes and recommendations during accident investigations.5 inlhr.Because pure copper melts at 1080°C. Iron and steel have high melting points (1300 -1 500°C) compared to fire temperatures. Deflagrations Breaks in pipes or vessels resulting from deflagrations or simple overpressurizations are usually tears with lengths no longer than a few pipe diameters.

With pressure piling. X 16 X 16 for Equations 12-1 and 12-2. In pipe systems explosions can initiate as deflagrations and the Aame front may accelerate to detonation speeds. They usually occur at pipe elbows or other pipe constrictions. therefore p. 6Lees. it is important to know the pressures required to create the damage and ultimately to determine the magnitude and source of energy. the source of ignition is at the point with the thinnest walls. The downstream pressure may be 8 to 16 times greater than the original upstream pressure. therefore. such as valves. x 8 x 8 andp. The damage is manifested as tears (detonations give shearing failures). drums. p. Therefore. may increase by as much as another factor of 20. and vessels great distances. respectively. therefore the failure occurs at the vessel's weakest point. the downstream p . the source of ignition is usually coincident with the point of maximum vessel thinning resulting from expansion. Blast pressures can shatter an elbow into many small fragments. Rapidly expanding gases give large tears and can propel missiles. Loss Prevention. Detonation failures in pipe networks are always downstream from the ignition source. p2 = p. 569. A detonation in light-gauge ductwork can tear the duct along seams and can also produce a large amount of structural distortion in the torn ducts. . which are destructive. = p. Detonation failures usually occur in pipelines or vessels with large length-to-diameter ratios. Sources of Ignition in Vessels When a vessel ruptures because of a deflagration. if the vessel parts are reconstructed. Hydraulic and Pneumatic Failures Hydraulic high-pressure failures also give relatively small tears compared to pneumatic failures. In a single vessel detonations increase pressures significantlyh: When a pipe network is involved. During deflagrations in vessels. Pressure Effects When pipe or vessel ruptures are investigated.12-5 Aids for Diagnosis 523 For pipe networks the pressure will increase in front of the flame front as the flame travels through the network. increases because of pressure piling. Detonations As described in chapter 6. detonations have a rapidly moving flame andlor pressure front. and the point of ignition has no relationship to the ultimate point of failure. This concept is called pressure piling. the pressure is uniform throughout the vessel.

S ." ChemicalEngineering (Nov.They are also used to determine the pressures required to produce failures by using tensile strengths for S M . p. 191.the following equation applies: For spherical vessels with pressures not exceeding 0. and r is the inside radius of the vessel.1968). Pan. is the strength of the material. "Designing Pressure Vessels. C. and the mechanical properties of the vessel walL7 For cylindrical vessels with the pressure p not exceeding 0. is the wall thickness of the vessel. 'Samuel Strelzoff and L.385SM. . the vessel diameter.Strength of material data are provided in Table 12-3. 4. t.665SM the equation is These formulas are also used to determine the pressure required to produce elastic deformations by using yield strengths for SM. For cylindrical vessels and pressures exceeding 0.665SM the equation is For spherical vessels and pressures exceeding 0.385 times the mechanical strength of the material SM where p is the internal gauge pressure.524 Chapter 12 Accident Investigations The pressure necessary to produce a specific stress in a vessel depends on the thickness of the vessel.

Perry and Cecil H. (4) signs of suffocation. Miscellaneous Aids to Diagnosis Other aids for identifying underlying causes of accidents are found throughout this text. eds. The ultimate source of this explosion energy is found by developing various reaction or mechanical hypotheses and comparing the reaction energy to the explosion energy until the most likely hypothesis is identified. etc. During an accident investigation. (3) magnitude of leaks depending on the source. ( 5 ) principles of grounding . (5) signs of heat exposure or heat exhaustion. and (6) signs of eye irritation. 1973). Chemical Engineers' Handbook (New York: McGraw-Hill. Chilton. and they must also make supporting calculations to evaluate various hypotheses. (3) type of poisoning (carbon monoxide. for example. (2) explosion limits. A brief review of safety fundamentals before the investigation is helpful. the explosive energy is computed. attention is placed on developing conditions to prevent the source of failure. After the energy and ignition sources are identified. The maximum internal pressure for pipes is calculated using Equations 12-4 and 12-5. This includes. pp. 6-96 and 6-97. toluene. Medical Evidence Medical examinations of the accident victims result in evidence that may be useful for identifying the source of the accident or for identifying some circumstances that may help to uncover underlying causes. (4) dispersion of vapors outside or inside plants. using Equation 6-29. (1) toxicity of chemicals or combinations of chemicals.).. the investigation team must watch for visible evidence. The types of medical data that help accident investigations include (1) type and level of toxic or abusive substances in the blood.12-5 Aids for Diagnosis 525 Table 12-3 Strength of Materials1 Tensile strength (psi) Yield point (psi) Material Borosilicate glass Carbon Duriron Hastelloy C Nickel Stainless 304 Stainless 316 Stainless 420 'Robert H. After the maximum internal pressure is computed. High-pressure failures are as likely to occur in a pipe or pipe system as they are in vessels. (2) location and magnitude of injuries.

This knowledge and information will be useful during an investigation. and a pressure of 1170 psi is required to rupture this vessel.t.75 in)(80.385SM. Therefore Equation 12-4 is applicable. What pressure ruptured the pipe.Because SM (from Table 12-3) is 85. a wall thickness of 0. Example 12-5 During an accident investigation.75 in.5 in) . and has a wall thickness of 0.6653.000psi) (5 ft)(12 inlft) + 0. Equation 12-4 or 12-5 is used by trial and error until the correct equation is identified. (6) principles of static electricity.237 in) - = 9348 psi. (85.000 psi.665(80. Equation 12-6 is applicable if the pressure is less than 0. and an inside diameter of 4. By substituting into Equation 12-4 for cylindrical vessels.(2.526 Chapter 12 Accident Investigations and bonding.385SM = 32.000 psi. and has a wall thickness of 0. = 0. Solution This problem is similar to Example 12-3. and r = 3 ft = 36 in and t. Deflagration tests in a small spherical vessel indicate a deflagration pressure of 500 psi.2(0.5 in.237 in) .026 in. . or 0.6(0. Equation 12-4 for cylinders is used to compute the pressure necessary to rupture this pipe: = y SM~" (85. It is hypothesized that a hydrogen and oxygen deflagration or a detonation was the cause of the accident.237 in. it is found that the source of the accident was an explosion that ruptured a 4-in-diameter stainless 316 schedule 40 pipe. From Table 12-3 the tensile strength SMfor stainless 316 is 85. The pressure criteria is met for this equation. we obtain = S. (7) design concepts for handling flammable materials.000 psi)(0. Solution Because the pressure is unknown.000 psi)(0. The pressure required to rupture this vessel is 1990 psi.000) = 53. and was it a deflagration or a detonation that caused this rupture? Solution A 4-in schedule 40 pipe has an outside diameter of 4. we obtain = y 2tvS~ - 2(0.013 in) + 0. 0.5 in) - = 1170 psi. has a radius of 3 ft.6(0. has a radius of 5 ft. Example 12-3 Determine the pressure required to rupture a cylindrical vessel if the vessel is stainless 316. Equation 12-4 is applicable for pressures below 0.5 in. and (8) methods for performing accident investigations.(36 in) + 0.75 in) = 1990 psi.700 psi. Using Equation 12-6 for spherical vessels.5 in. Example 12-4 Determine the pressure required to rupture a spherical vessel if the vessel is stainless 304.200 psi.

we can estimate the deflagration pressure in the pipe using Equation 12-2: p.700 psi. The pressure required to rupture this pipe. therefore. = 500 X 16 = 8000 psi. A 1000-ft3tank containing compressed air at 100 atm is suspected.4. Using the deflagration test data. Example 12-6 An explosion rips through a chemical plant. = 500116 = 31. representing the energy contained in a compressed gas.3 X 20 X 20 = 12. Substituting the known quantities. . which gave a p.27 X 10' call(ll20 callg TNT) = 1.500 psi. = 31. This pipe rupture was therefore due to a detonation. y = 1. A detonation with pressure piling is now computed using Equation 12-3: p. To estimate pressures resulting from a detonation and pressure piling. we obtain The equivalent amount of TNT is m . Is the mechanical explosion of this compressed air tank consistent with the damage reported. A small vessel could be used as a test. For air. The next step in the investigation would include searching for a chemical reaction that would give a detonation. and assuming pressure piling. Site damage indicates that the windows in a structure 100 yards away are shattered. or is the explosion the result of some other process? Solution From Equation 6-29.3 psi. using Equation 12-2: p. of 500 psi. is 9348 psi.13 X 10" g TNT 249 lb of TNT. we estimate the original deflagration test pressure p.12-5 Aids for Diagnosis 527 Equation 12-4 is applicable because the pressure is less than 0. = = 1.385SM= 32.

528 Chapter 12 Accident Investigations From Equation 6-21 the scaling factor is Substituting.9 mlkg"' From Figure 6-23 the overpressure is estimated at 1. changes to the system or new procedures may add complexities that result in a more hazardous situation rather than in an improvement. Evaluate each investment (money and time) to ensure that there is a true safety improvement proportional to the investment. safety reviews. including training. Some aids to recommendations are covered in the following sections.3psia. They are made to prevent a recurrence of the specific accident. Develop recommendations to improve the management system to prevent the existence of safety hazards. Develop layered recommendations with an appropriate emphasis on recommendations to eliminate underlying causes of accidents. In addition. and in these situations all efforts are directed toward a quick restart. From the data provided in Table 6-9 the estimated damage is consistent with the observed damage.66 ftllb"' = 18. 12-6 Aids for Recommendations Recommendations are the most important result of an accident investigation. = (249 lb)"" 300 ft = 47. plans. 4. One good accident investigation can prevent hundreds of accidents. There are four overriding principles that are used to influence accident investigation recommendations: 1. checklists. but they are also made to prevent similar accidents within the company and within the industry. quality. and organization as used in production programs. . 3. Develop recommendations to improve the management and staff support of safety with the same enthusiasm. inspections. and audits. The ultimate result of accident investigations is the elimination of the underlying causes of entire families of accidents. All the fundamentals described in this text are commonly used to develop recommendations. Make safety investments on a basis of cost and performance. Many accidents are the result of plant modifications. 2.we obtain z. attention. they are sometimes the result of mechanical problems that shut the plant down. Control Plant Modifications Modifications to plants are often not given the same attention and concern as a new plant design. If the designer is not careful.

4.to tolerate departures from the norm without creating hazardous conditions. p. . such as area monitors for detecting solvent leaks. and design specialists while the modification project is in the design phase. This allows (and encourages) safety changes to be made with minimum effort. / Block Valves Block valves are installed throughout plants to return a process to a safe condition under unusual circumstances. 5. User-Friendly Designs8 New plants or modifications to existing plants must be designed to be friendly . operators. 148. the leak is detected and the block valves on the reactor and sewer are immediately closed. Learning from Accidents in Industry. 8Kletz. and vent analyzers to detect high levels of contaminants. 3. the process shown in Figure 12-2 detects a hose leak by comparing flow rates at both ends of the hose. keeping temperatures below the flash point and below the boiling point at atmospheric conditions. plausible. Authorization: All modifications must be authorized by several levels of management. If the hose breaks. Examples of friendly designs include using nontoxic and nonflammable solvents when possible. keeping inventories low. Training: Engineers and operators need sufficient training to understand and appreciate the modified operation. 2. and designing for safe shutdowns during emergency situations (expect the unexpected). Design: The modification designs must be mechanically constructed with the same quality of equipment and pipes as the original design. reactor analyzers for detecting runaway reactions (a block valve can be opened to add a reaction inhibitor or to turn on a deluge system). Once the system is constructed. These five requirements are all part of the OSHA q'ocess Safety Management regulation discussed in chapter 3.12-6 Aids for Recommendations 529 Recommendations are especially designed to prevent this kind of problem: 1. economically sound. sewer analyzers for detecting high concentrations of contaminants. Safety reviews: A safety review (HAZOP study or equivalent study for hazardous operations) must be conducted by engineers. changes are difficult and costly to make. The designers must appreciate that for every problem there are many interesting. Block valves are often controlled on the basis of analyzer results. Original designs should be studied so that the consequences of any change are understood. and wrong solutions. For example. Audit: Every plant modification needs periodic audits to be sure that the modifications are made and maintained as designed.

for example.anywhere a positive break in a line is desired. . When the feed pump is shut off. shown in Figure 12-3. A little neglect can cause serious property damage and may be the genesis of serious accidents.is added to every feed line to a reactor. Double Block and Bleed A special double block and bleed system. even if the block valves leak. the block valves are open and the bleed line is closed. poorly maintained brake systems have inevitable consequences. especially those owning automobiles or homes. If the flow at both ends of the hose is not identical. During normal operating conditions. This prevents an unexpected chemical reaction in the storage tank.530 Chapter 12 Accident Investigations Hose Reactor Solvent Storage Sewer Curbs a n d Sump for Spill Control Figure 12-2 A block valve arrangement used to prevent leakage from the connecting hose. This system prevents the reactor contents from siphoning back into the monomer storage vessel. These systems are also commonly used for reactive intermediates and analyzer systems . the block valves are closed and the bleed line is open. Preventive Maintenance Most engineers are aware of the importance of preventive maintenance programs. the block valves are closed. Figure 12-3is a relatively simple example of a particularly important application of double block and bleed systems.

instruments. but bad results are apparent when plants are not appropriately maintained. Good results may not be immediately apparent. Analyzers Chemical analysis of reactor contents and of the surrounding environment is an important way to understand the status of a plant and to identify problems at the incipient state of development. . and deluge systems. it is found that protective equipment failed because it was neglected. managed. and fully supported by management. Many times. Every maintenance job requires a feedback mechanism based on the inspection of parts while conducting the maintenance. In plants one major cause of accidents is the failure of emergency protection equipment such as cooling water pumps. when evaluating underlying causes of accidents. Good maintenance programs include scheduled maintenance and a system to keep an inventory of critical maintenance parts. Preventive maintenance programs must be organized.12-6 Aids for Recommendations Double Block Pipe L Controller Figure 12-3 A double block and bleed arrangement used to prevent reactant from entering reactor vessel. adding more protective equipment or procedures might increase the likelihood of accidents. there was no preventive maintenance program. In this case new procedures or new equipment is not needed. The only improvement needed is upgrading the existing maintenance program. When problems are identified at an early stage. action can be taken to return the system to safe operating regions with no adverse consequences. The maintenance schedule is subsequently changed if more frequent maintenance is required.

Kletz. Use the Seveso. 12-4. 1988). Use the Seveso. 373382.532 Chapter 12 Accident Investigations In recent years new and better analyzers have been developed." Hydrocarbon Processing (November 1979). A Thirty Year Review of One Hundred ofthe Largest Property Damage Losses in the Hydrocarbon-Chemical Industry. 1996). 12-3. Use the Bhopal. Guidelines for Investigating Chemical Process Accidents (New York: American Institute of Chemical Engineers. 1996). Trevor A. Design engineers should always search for new opportunities to use process analyzers to improve operations and safety within plants.. Use the Flixborough Works accident and the investigation developed in Problem 12-1 to develop an investigation summary similar to Example 12-2. Problems 12-1. Kletz. Chemical Process Safety: Learning from Case Histories. quality. 2d ed. they will become the key control elements in chemical plants. 12-6. and yield improvements. Roy E. Suggested Reading CCPS. Include layered recommendations. (New York: American Institute of Chemical Engineers. 1992). Chemical Process Safety: Learningfvom Case Histories (Boston: Butterworth-Heinemann. "Layered Accident Investigations. Include layered recommendations to cover the accident causes and underlying causes. 1999). 12-2. Inherently Safer Chemical Processes (New York: American Institute of Chemical Engineers. . 1985). 1992). accident described in chapter 1 to develop an investigation similar to Example 12-1. Sanders. (Chicago: Marsh and McLennan Protection Consultants. Plant Guidelines for Technical Management of Chemical Process Safety (New York: American Institute of Chemical Engineers. Include layered recommendations. CCPS. India. Use the Flixborough Works accident described in chapter 1to develop an investigation similar to Example 12-1. CCPS. Use the Bhopal accident and the investigation developed in Problem 12-3to develop an investigation summary similar to Example 12-2. Trevor A. 1994). Italy. 12-5. 11th ed. What Went Wrong? Case Histories of Process Plant Disasters (Houston: Gulf Publishing. 18th ed. Trevor A. pp. 1998). accident and the investigation developed in Problem 12-5 to develop an investigation summary similar to Example 12-2. Kletz. particularly in regard to safety. (Chicago: J&H Marsh & McLennan Inc. 2d ed. Learning from Accidents in Industry. Large Property Damage Losses in the Hydrocarbon-Chemical Industries: A Thirty-Year Review. CCPS. accident described in chapter 1 to develop an investigation similar to Example 12-1. Italy. (Boston: Butterworth-Heinemann. As the reliability and applicability of analyzers are improved.

The system has two vessels that are connected with a 4-in vapor line. assuming that stoichiometric quantities of hydrogen and air explode. determine the wall thickness required to eliminate future failures. Using the results of Problem 12-13. Determine the vessel wall thickness required to contain an explosion of 2 lb of TNT.Problems 533 12-7. 12-21. An accident occurs that ruptures a high-pressure spherical vessel. is made of stainless 304.5 ft in diameter.25 in thick. Compute the theoretical maximum pressure obtained after igniting a stoichiometric quantity of methane and air in a spherical vessel that is 1. Describe a preventive maintenance program that is designed to prevent automobile accidents. Create your own brief accident scenario for this problem. The normal operating pressure is 2 atm. Using the data and results of Example 12-6. . 12-13. Determine the required thickness of a reactor with cylindrical walls that must be designed to safely contain a deflagration (hydrocarbon plus air). Develop some hypotheses regarding the causes of this accident. Describe why the second vessel requires a greater wall thickness. Glass is shattered 500 ft away. Describe the concept of using block valves to prevent detonation accidents in a system handling flammable gases.determine the vessel wall thickness required to contain an explosion in another vessel that is physically connected to the first vessel with a 1-in pipe. 12-20. 12-17. Assume an initial pressure of 1 atm. 12-12. 12-14. Develop an investigation similar to Example 12-1 and an investigation summary for an automobile accident that occurred as a result of a brake failure. 12-15. Determine the pressure required for a pipe to swell and the pressure required for a pipe failure. Compute the theoretical maximum pressure obtained after igniting a stoichiometric quantity of methane and oxygen in a spherical vessel that is 1. The vessel has a diameter of 4 ft and is constructed with stainless steel 304. 12-9. Using the results of Problem 12-11. The pipe is 3-in stainless 316 schedule 40 pipeline for transporting a gas mixture that is sometimes within the explosive composition range. 12-10. 12-19. Assume that the initial pressure is 1atm. Estimate the quantity of hydrogen gas initially released. 12-8.5 ft in diameter. Describe what features are added to water heaters to eliminate accidents.5 ft in diameter. A cloud of hydrogen gas is released and subsequently explodes. Assume that the vessel's cylindrical wall height is equal to the vessel's diameter. 12-18. 12-11.5 ft in diameter and is constructed with stainless steel 316. The vessel is 1.determine the required vessel wall thickness to contain this explosion if the vessel is made of stainless 316. Describe why accident investigation recommendations must include recommendations to improve the management system. Determine the pressure required to cause this failure. In the 1930s there were many accidents in homes because of the explosion of hot water heaters. and the walls are 0. 12-16. The spherical vessel is 1.

gov/ceppo/pubs/lodirecc. 12-25. epa. 238).htm. htm. What are the tangible benefits of a good communications system? Compare your answer to CCPS's (1992. See CCPS. See http:// www. p. As stated in section 12-4. . What facts should a near-miss accident report include? Compare your answer to CCPS's (1992. p. included previous industrial accidents with sodium hydrosulfite and aluminum. the three layers of recommendations for accident investigations include management systems to prevent similar accidents or to eliminate the hazardous conditions.epa. See http://www. 240). Compare your answer to CCPS's (1992. An accident investigation at the Tosco Refinery Company emphasized the importance of a management system.gov/. 12-23.chernsafety. 12-26. and develop three layers of recommendations. 12-28. An EPA-OSHA accident investigation at Napp Technologies Inc. Develop a definition for a major incident. p. The accident investigation at Lodi. Describe the accident.gov/ceppo/pubs/lodiintr. Evaluate the board's recommendations. 239). 238. Plant Guidelines for Technical Management of Chemical Process Safety (1992).chernsafety . developed the root causes and recommendations to address the root causes. New Jersey. The US Chemical Safety and Hazard Investigation Board investigated an accident at the Morton Specialty Chemical Company in 1998. 12-27. See http:15/ www. and break them down into three layers of recommendations. New Jersey.gov/. p. 12-30. This management system includes the delegation of responsibilities and followup. 12-29. in Lodi. Define nearmiss accidents. A management system for accident investigations includes good communications. Near-miss (close-call) accident investigation reports are also important. 12-24. What are the benefits of followup? Compare your answer to the benefits described in the CCPS (1992) reference on p. Describe the accident. 236.534 Chapter 12 Accident Investigations 12-22. See http://www. Summarize the findings of these accidents and develop a few management system recommendations for these industries. and compare it to CCPS's definition. and develop layered recommendations for this specific accident.

and methods required to prevent similar events. Anyone familiar with the specific equipment or procedures will appreciate the lessons learned. anyone working in the chemical industry can learn from case histories and avoid hazardous situations or ignore history and be involved in potentially life-threatening accidents. . In this chapter we cover case histories as reported in the literature. These are the people with the hands-on experience. References are provided for more thorough studies. Each of these sections includes descriptions of several accidents and a summary of the lessons learned. Santayana. The objective of this chapter is to illustrate.C H A P T E R 1 3 Case Histories ase histories are written descriptions of accidents. chemical reactivity. the ones who know and appreciate the accident and accident prevention methods. This is especially true for safety. the importance of applying the fundamentals of chemical process safety. The cause of a specific accident frequently places it in more than one category. To paraphrase G. consequences. These case histories are categorized into four sections: static electricity. and procedures. These accidents actually occurred. one learns from history or is doomed to repeat it. system design. through actual case histories. The following statements place the case histories into perspective: c 1. including the causes. The study of case histories is important in the area of safety. They are descriptions written by plant managers and operating personnel.

The accident investigation indicated that the explosion was caused by a static spark that jumped from the steel nozzle to the tank car.536 Chapter 13 Case Histories 2. DC: Manufacturing Chemists' Association. The lid was lifted and a flame was released between the centrifuge and the lid. 1 (Washington. The use of a nonmetallic hose probably also contributed. engineers are missing some of the important aspects of this subject. January 1966). 231. There is usually inadequate time to manually return a situation back into control after a significant deviation from the norm is observed. p. 2 (Washington. . running to a point 3 S from the centrifuge where there was a rubber sleeve connector. The fill line from the reactor to the centrifuge was Teflon-lined steel. Explosion in a Centrifuge2 A slurry containing a solvent mixture of 90% methylcyclohexane and 10% toluene was being fed into a basket centrifuge. the contents of the tank exploded. 2Case Histories of Accidents in the Chemical Industry. Those who believe that they can successfully control accident deviations manually are doomed to repeat history. The following series of case histories is given to illustrate the complexity of this topic and to give some important design requirements for preventing future accidents involving static electricity. The short line from the t sleeve to the centrifuge was steel. he sustained a fractured skull and multiple body burns and died from these injuries. v. p. The operator on top of the tank was thrown to the ground. DC: Manufacturing Chemists' Association. Case Histories of Accidents in the Chemical Industry. The nozzle was not bonded to the tank car to prevent static accumulation. 106. One operator was on the ground. Many of these accidents are repeats of previously recorded accidents. and the other was on top of the car with the nozzle end of a loading hose. Accidents occur rapidly and unexpectedly. 13-1 Static Electricity A large proportion of the reported fires and explosions are the result of a flammable mixture being ignited by a spark caused by static electricity. The centrifuge was lined. v. The foreman's hand was burned. Tank Car Loading Explosion Two plant operators were filling a tank car with vinyl acetate. A foreman was about to look into the centrifuge when it exploded. July 1962). A few seconds after the loading operation started.

Maurer. proper grounding and bonding techniques must be used to prevent static buildup. Later (and successful) processing was conducted in a grounded stainless steel centrifuge that was inerted with nitrogen. Duct System Explosion Two duct systems in the same vicinity contained dust transport lines. The open system emitted some methanol vapors. and after modifications are made. A piece of tramp metal entered the hopper with the solids. W. 3 (Washington. 4J. a magnetic trap or a screen. and G. which was grounded. v. As it rolled down the solids. The lined centrifuge was the source of ignition as a result of static accumulation and discharge. and a building blowout panel also ruptured. April 1970). An additional safeguard would be the addition of an inerting gas. pp. The other system was being charged through a funnel with a dry organic intermediate. At some point in the operation the tramp metal approached the metal wall of the hopper. The spark was energetic compared to the minimum ignition energy of the dust." Chemical Engineering Progress (November 1994). . Louvar. This explosion could have been prevented with a tramp metal collector. F. Several explosion vents were ruptured.13-1 Static Electricity 537 The accident investigation indicated that a flammable atmosphere had developed because of an air leak. Fortunately. Open lines should be blanked-off when the discharge of flammable vapors is possible. One system was recently repaired and left open. This accident points out the need for carefully reviewing systems before. B. and hoppers. The ignition source created violent explosions in both systems. The duct system that was being charged exploded violently. DC: Manufacturing Chemists' Association. Also. 75-81. The transportation of the intermediate dust through the glass and plastic line generated a static charge and spark. the dust exploded and the storage hopper ruptured. Conductor in a Solids Storage Bin4 A dry organic powder was collected in a hopper. and the explosion initiated other fires. p. "Tame Static Electricity. 95. for example. A spark jumped from the tramp metal to the grounded wall. dryers. v a s e Histories of Accidents in the Chemical Industry. during. it accumulated a charge by the charging method called separation. no one was seriously injured. The charge line consisted of a new glass pipe and a 6-ft section of plastic pipe. The accident investigation indicated that methanol vapors entered the second charging system. Boicourt. Because the storage hopper's atmosphere was air (plus the dust).

" r h ~ o u v aet al. Pipefitter's Helper6 Apipefitter's helper was transporting tools to the boss. Possibly. Using conductive shoes and changing the location of the toolbox solved this problem. and all metal parts were bonded. In reviewing approximately 30 additional case histories regarding static electricity. during an emergency situation. This example may have been a disaster if the pipefitter was repairing a flammable gas leak. (6) large voltages are generated when crumpling and shaking an empty polyethylene bag. It is also possible that the grounding rod wire was closer to the isolated drum than previously. ( 3 ) leather arch supporters make shoes ineffective against static." r . It is hypothesized that one of two scenarios could have created the ignition. Charge loss was prevented because the helper was wearing insulated shoes. some important lessons were identified: (1) A built-in ground line is rendered nonconductive by the use of a nonconductive pipe dope. The problem was the steam. A number of recommendations are also developed: (1) Operators must be cautioned against drawing pipes or tubing through their rubber gloves. The pigment drum was grounded by means of a grounding rod. resulting in static buildup.. (4) free-fall filling generates static charge and discharge. one day there was a fire.538 Chapter 13 Case Histories Pigment and Filter5 A low-flash-point solvent containing pigment was pumped through a bag filter into an open drum. in this case a spark could have jumped between the drum and the grounding wire. it became charged as it exited a manifold. Then the charge was transferred to the helper and to the tools when the helper passed through the steam cloud. ( 2 )clothing that generates static electricity must be prohibited.. "Tame Static Electricity. The helper walked through a cloud of steam before handing the tool to his boss. giving the conditions for a brush discharge between the filter and the grounding rod. ( 5 ) the use of nonmetallic hoses is a source of static buildup. Subsequent operations were incident-free. Upon each transfer. Lessons Learned Case histories involving static electricity emphasize the importance of understanding and using the fundamentals described in chapter 7. The boss was grounded because he was kneeling on a damp grounded grating. "Tame Static Electricity. the grounding rod was placed closer to the filter than previously. (3) recirculation lines must be extended 5 ~ o u v aet al. the boss received a rather large shock. ( 2 ) a potential is generated between two vessels that are not bonded. and (7) a weak grounding clamp may not penetrate the paint on a drum adequately to provide a good electrical contact. for example. This system was modified to include an inerting system and a dip pipe charging line. Although the operation ran successfully for some time.

Develop tank car loading procedures. Initiate a periodic (every six months) audit to ensure that all standards and procedures are effectively utilized. 4. Bond hose to tank car. and 5. contents at the top of vessel were flammable.13-1 Static Electricity 539 into the liquid to prevent static buildup. Solution The facts uncovered by the investigation are 1. (5) bonding. 2. a policy to give all new loading and unloading applications a safety review. humidification. develop the underlying causes of the tank car loading explosion discussed earlier in this section. grounding. Provide dip pipe design for charging tank cars. Add guardrails to charging platforms to prevent accidental falls from the top of the tank car to the ground. or combinations are recommended when static electricity is a fire hazard. (13) stainless steel centrifuges must be used when handling flammable materials. Develop and give operators special training so the hazards are understood for every loading and unloading operation. and bags should b e physically bonded and grounded. Initiate. a spark probably jumped between the charging nozzle and the tank car. (9) ground connections must be verified with a resistance tester. 2. 4. (6) a small water spray will rapidly drain electrical charges during chopping operations. (4) shoes with conductive soles are required when handling flammable materials. Third-layer recommendations: improving the management system 1. the explosion knocked the man off the tank car (the fatal injury was probably the fractured skull sustained in the fall). (10) Spring-loaded grounding or bonding clips should be replaced with screw type C-clamps. (12) sodium hydride must be handled in static-proof bags. the charging line was a nonconductive hose. . Second-layerrecommendations: avoiding the hazard 1. Use a conductive metal hose for transferring flammable fluids. 5. scoops. (11) conductive grease should be used in bearing seals that need to conduct static charges. 3. and (14) flanges in piping and duct systems must be bonded. no inspection or safety review procedure was in place to identify problems of this kind. 2. 3. ionization. and ground tank car and hose. Provide a means to nitrogen-pad the tank car during the filling operation. Initiate an immediate inspection of all loading and unloading operations. First-layer recommendations: immediate technical recommendations 1. as a standard practice. (8) drums. (7) inert gas blankets must b e used when handling flammable materials. Include engineers and operators in this review. 2. Layered recommendations are the result of uncovering the underlying causes of the accident. 3. Example 13-1 Using the layered accident investigation process discussed in chapter 12.

The following case histories illustrate the importance of understanding the complete chemistry of a reaction system. The victim remained conscious and. He unsuccessfully tried to open the bottle over a sink. and through the roof of an adjoining building. These materials are unstable. v. or the wrong temperature). they form peroxides. Bottle of lsopropyl Ether7 A chemist needed isopropyl ether. unexpected. coherently described how the accident happened. p. An accident investigation identified the cause of the accident to be the rapid decomposition of peroxides. He found a pint glass bottle. Case Histories. p. Just as the cap broke loose. Ethers should not be kept over 6 months. the bottle exploded. Light. The peroxides react further to form additional hazardous by-products. 3. especially isopropyl ether.540 Chapter 13 Case Histories 13-2 Chemical Reactivity Although accidents attributable to chemical reactivity are less frequent compared to fires and explosions. such as triacetone peroxide. Only small quantities should be purchased. 2. which was known to decompose at 200°C. and heat accelerate the formation of peroxides. Nitrobenzene Sulfonic Acid DecompositionX A 300-gal reactor experienced a violent reaction. The cap appeared to be stuck tightly. decomposition reactions. destructive. wrong concentrations. pressed it to his stomach and twisted the cap with his other hand. Ethers should be stored in metal containers. 111 . 6. All work with ethers should be done behind safety shields. Containers should be labeled and dated upon receipt. When working with chemicals. It is hypothesized that some of the peroxides crystallized in the threads of the cap and exploded when the cap was turned. practically disemboweling the man and tearing off several fingers. resulting in the tank being driven through the floor. The man was taken to a hospital and died within 2 hr of the accident of massive internal hemorrhage. the potential for unwanted. The reactor was designed to contain 60 gal of sulfuric acid and nitrobenzene sulfonic acid. 8Case Histories. out the wall of the building. As ethers age. and reactions resulting from the accidental and wrong combination of chemicals or reaction conditions (wrong type. the consequences are dramatic. and often injurious to personnel. Inhibitors should be used whenever possible. and hazardous reactions must always be recognized. which formed in the ether while the bottle sat in storage. v. air. including potential side reactions. and opened containers should be discarded after 3 months. in fact. so he grasped the bottle in one hand.

The rate of heating was slower than normal. He turned on the agitator. The two operators were drenched and both died from these injuries. Lessons Learned Case histories regarding reactive chemicals teach the importance of understanding the reactive properties of chemicals before working with them. Without agitation. Approximately 1 hour later. At this point the material in the kettle erupted through the manhole. Although previous tests indicated decomposition at 200°C.p. cooling is no longer efficient. This is achieved by adding redundant and remote temperature sensors and by adding electronic interlocks to prevent the agitator from being turned off while the reaction is still exothermic. Specific functional groups that contribute to the explosive properties of a chemical through rapid combustion or detonation are illustrated in Table 13-1. so heat-up occurs. 121. . the operator was ready to take a temperature reading through the manhole. This type of problem is currently preventable through better operator training and installation of electronic safeguards to prevent operators from making this mistake. the hotter chemicals mix and react violently. experimental testing is necessary. Steam was applied to the reactor jacket to heat the sulfuric acid and an organic material to a temperature of 70°C. The best source of data is the open literature. One operator went to find a thermometer. When the agitator is subsequently activated. If data are not available. Cuse Histories. The underlying cause of this accident was the lack of precise reaction decomposition data. The two operators turned the agitator off and also shut off the steam. Without agitation. Data of special interest include decomposition temperatures.13-2 Chemical Reactivity 541 The investigation indicated that the vessel contents were held for 11hr. impact shock sensitivity. 3. Organic Oxidation Chemical operators were preparing for an organic oxidation. engineers can design safeguards to absolutely prevent accidental heat-up. and flash point. With good data. The accident investigation stated that the agitator should never be turned off for this type of reaction. rate of reaction or activation energy. A steam leak into the jacket brought the temperature to about 150°C. segregation of chemicals also occurs. subsequent tests showed exothermic decomposition above 145°C. Functional Groups A preliminary indication of the potential hazards can be estimated by knowing something about the chemical structure. v.

Steere. PA: Division of Chemical Education. The RHI is related to the maximum adiabatic temperature reached by the products of a decomposition reaction. Jackson et al." in Safety in the Chernicul Industry. L. Norman V. 3. Steere. R. pp.Chapter 13 Case Histories Table 13-1 Azide Diazo Diazonium Nitro Nitroso Nitrite Nitrate Fulminate Peroxide Peracid Hydroperoxide Ozonide Reactive Functional Groups1 N3 -N=N-N2+X- -NO2 -NO -O N 0 -ON02 -O N C -0-o-C 0 3 H -0-0-H 0 3 -N-Cl Amine oxide Hypohalites Chlorates Acetylides of heavy metals X =NO - I ox C103 -C=CM 'Conrad Schuerch. 114 -1 17. v. Jackson et al. R. It is defined as RHI = 'OH. 3. Some examples of peroxidizable compounds are given in Table 13-3." in Safety in the Chemical Industry. Stullll developed a rating system to establish the relative potential hazards of specific chemicals.. pp. the rating is called the reaction hazard index (RHI). When peroxide concentrations increase to 20 ppm or greater. the solution is hazardous.1° Reaction Hazard Index D. "D. . PA: Division of Chemical Education. Stull. 106 -1 10. pp. "Control of Peroxidizable Compounds. L. 1974)." in Safety in the Chemical Laboratory. "Safe Practice in the Chemistry Laboratory: A Safety Manual. American Chemical Society. American Chemical Society. "Linking Thermodynamic and Kinetics to Predict Real Chemical Hazards. 22-25. 1974). (Easton. Norman V. v. ed. Peroxides Peroxides and peroxidizable compounds are dangerous sources of explosions. Structures of peroxidizable compounds are shown in Table 13-2. ed. (Easton. Methods for detecting and controlling peroxides are outlined by H.

pp. Jackson. particularly those with tertiary hydrogen: I I I >C=C-C=C< >C=C-CGCH >C- I H -C=O 5. and I. Smeltz. Vinyl monomers. Rondestvedt. amides. Dienes. 2R.13-2 Chemical Reactivity 543 Table 13-2 Peroxidizable corn pound^^^^ Organic material and structure H -C-0- 1." in Safety in [he Chemical Industry. "Review of Safety Guidelines for Peroxidizable Organic Chemicals. Ketones having an alpha-hydrogen: 11 0 I H Inorganic materials 1. 3. S. v. acetals: 2. L. Norman V. Ureas. Ethers. Organometallics 'H. Alkali metals. Paraffins and alkyl-aromatic hydrocarbons. Olefins with allylic hydrogen. Viele. 114-117. J. Kelly. tetrahydronaphthalene: I > C=C < 3. W. pp. Steere. C. (Easton. American Chemical Society. Aldehydes: I H 0 6. vinyl esters: H 11 -C-N-C I I / > C=C-C-C< 8. PA: Division of Chemical Education. chloro. particularly potassium 2. vinyl acetylenes: and 4. lactones: 7. acrylates. ed. 28-36. K. C. including vinyl halides. 1974)." Chemical Health and Safety (September-October 1996). . "Control of Peroxidizable Compounds. Alkali metal alkoxides and amides 3. B. E. methacrylates. McCormack. terpenes.and fluoroolefins.

PA: Division of Chemical Education. where T. ed. is the Arrhenius activation energy (kcallmol). (Easton.544 Chapter 13 Case Histories Table 13-3 Examples of Peroxidizable Compounds Peroxidizable hazard on storage Isopropyl ether Divinyl acetylene Vinylidene chloride Potassium metal Sodium amide Peroxidizable hazard on concentration Diethyl ether Tetrahydrofuran Dioxane Acetal Methyl i-butyl ketone Ethylene glycol dimethyl ether (glyme) Vinyl ethers Dicyctapentadiene Diacetylene Methyl acetylene Cumene Tetrahydronaphthalene Cyclohexane Methylcyclopentane Hazardous when exposed to oxygen due to peroxide formation and subsequent peroxide initiation of polymerization Styrene Butadiene Tetrafluoroethylene Chlorotrifluoroethylene Vinyl acetylene Vinyl acetate Vinyl chloride Vinyl pyridine Chloroprene H.. "Control of Peroxidizable Compounds." in Safety in the Chemrcal Industry. Norman V. 114-117. 1974). Some RHI data for various chemicals are provided in Table 13-4. is the decomposition temperature (K) and E. . 3. Jackson et al. v. Steere. L. pp. The RHI relationship (Equation 13-1) has a low value (1 to 3) for relatively low reactivities and higher values (5 to 8) for high reactivities. American Chemical Society.

as indicated in Table 13-3. Norman V. is 63. Stull. 1974). The units are compatible with Equation 13-1.0 46.82 2. v. T.25 3. (Easton. where. This RHI indicates a chemical with low reactivity.4 28.70 2.5 57.7 40. However.12 7.5 79. Example 13-2 Compute the RHI for isopropyl ether.5 78 63.5 46.81 7.5) which is the same as the value given in Table 13-4. ed.72 4.19 2. "Linking Thermodynamics and Kinetics to Predict Real Chemical Hazards. and compare the result to that shown in Table 13-4. 3.3 37. If we assume .98 6.3 Number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Formula CHC1.52 2. R.33 6.38 2.13-2 Chemical Reactivity 545 Table 13-4 Reaction Hazard Index Data1 Decomposition Activation temperature energy (K) (kcallmol) 683 597 859 634 866 712 1005 991 880 993 1338 1062 2317 2213 2898 2895 968 47 89. Steere. isopropyl ether is a peroxidizable compound. PA: Division of Chemical Education.3-butadiene vinyl ethyl ether styrene hydrazine ethylene oxide vinylacetylene cellulose nitrate acetylene nitroglycerine diethyl peroxide RHI 3. we obtain RHI = (10)(712) (712) + (30)(63.0 C8H8 N2H4 C2H40 C4H4 C12H~6N40~~ C2H2 C3H5N309 C4H~002 ID.2 60. Solution The RHI is computed using Equation 13-1: RHI = 10T..94 3. + 30Ea' T. is 712°K and E.33 4.4 44." in Safety in the Chemical Industry. American Chemical Society. C2H6 Compound chloroform ethane toluene acetic acid propylene isopropyl ether ethylene 1.64 C7H8 C2H402 C3H6 C~HI~O C2H4 C4H6 C4H. 106-110.05 7.4 19. Explain why the RHI is relatively low. from Table 13-4.5 40.5 85 67.26 1.05 4. pp.5 kcallmol. Substituting.

resulting in an accumulation of flammable vapors. dikes.64). Davenport. The following case histories emphasize the importance of these special safety design features. giving a partially confined vapor cloud explosion. and the like). Ethylene Oxide Explosion l 2 A process storage tank contained 6500 gal of ethylene oxide. It also appears that mitigation techniques were not part of the system (deluge systems. A vapor cloud was formed and almost immediately exploded. 13-3 System Designs When new plants are constructed or when modifications are needed in existing plants. and (3) lack of automated vapor detection analyzers and alarms.A.12-0. property loss was $6. giving an explosion equivalent to 0. The probable causes of this accident include (1)use of nonwelded pipe."A Survey of Vapor Cloud Incidents. '%avenport. It was accidentally contaminated with ammonia. Butadiene Explosionl4 A valve on the bottom of a reactor accidentally opened because of an air failure. The events happened so rapidly that personnel could not take appropriate cover. It created an explosive force equivalent to 18 tons of TNT. (2) installation of pipe in trenches. "A Survey of Vapor Cloud Incidents." I4Davenport. This accident took place in a courtyard. Two people were killed and 17 were injured. One person was killed and nine were injured. A cloud was formed and ignited.5 million. The tank ruptured and dispersed ethylene oxide into the air. detailed process designs are required. This example illustrates the importance of understanding the chemistry of the entire system. About 200 gal of butadiene spilled 12J. These designs must include special safety features to protect the system and operating personnel.546 Chapter 13 Case Histories an RHI equivalent to diethyl peroxide (RHI = 4." . "A Survey of Vapor Cloud Incidents.5 million. 54-63. as evidenced by the damage. the hazards of handling isopropyl ether are high even with peroxide concentrations as low as 20 ppm. property losses exceeded $16. Ethylene Explosion l3 Failure of a 3/8-in compression fitting on a 1000-2500-psi ethylene line in a pipe trench resulted in a spill of 200-500 lb of ethylene.30 ton of TNT. This accident was attributed to the lack of design protection to prevent the backup of ammonia into this storage tank. pp." Chemical Engineering Progress (September 1977). The spill generated a vapor cloud that was ignited 50 ft from the source.

Because the pump was equipped only with manually operated suction-side valves. A pump roller bearing failure in a crude oil refinery initiated the fracture of the motor shaft and the pump bearing bracket. Plant damage totaled over $15 million. (2) improved deluge system design. Secondary pipe and flange failures contributed fuel to the fire. One Hundred Largest Losses. p. This accident had two fatalities and nine injuries." IhWilliarnG. 1986). 7. which autoignited. I5Davenport. . block valves. p. (3) backup or more secure electrical supply. (Chicago: Marsh & McLennan Protection Consultants. and (4) installation of detection analyzers and block valves. The maximum overpressure was estimated from the damage to be 3.5-1 psi were estimated. One Hundred Largest Losses:A Thirty-Year Review ofproperty Damage Losses in the Hydrocarbon Chemical Industries. Overpressures of 0. gas analyzers.6 million. (3) lack of a block installed as a mitigating device. Pump Failure17 Numerous accidents are unfortunate duplicates of previous accidents. Light Hydrocarbon Explosion l5 A pipe failed and resulted in a spill of 16. and (4) failure to eliminate ignition sources in this operating region.- 13-3 System Designs 547 before ignition. "A Survey of Vapor Cloud Incidents. Significant damage resulted from the subsequent fires. Damage to the plant exceeded $23 million. giving an explosion yield of approximately 1%of the total energy source. A vapor cloud developed and ignited. The explosion knocked out the deluge systems and electrical supplies to the fire pumps. An equivalent of 1ton of TNT was estimated. adding fuel to the original fire. This accident could have been prevented by a good inspection and maintenance program. Probable causes of this accident include (1)installation of a fail-open valve instead of a fail-closed valve. The total damage was estimated to be $15. (2) lack of vapor detectors. as the following shows.800 lb of light hydrocarbons. "Garrison. Three people were killed and two were injured. 7. and deluge systems. The pump casing then broke. Pump Vibration I h Vibration from a bad pump bearing caused a pump seal to fail in a cumene section of a phenol acetone unit. 9th ed. The released flammable liquids and vapors ignited.5 psi at 120 ft. releasing hot oil.. the valves could not be reached during the fire. The magnitude of this accident could have been reduced with (1) improved pipe design. Garrison. Potential design improvements include vibration detectors. An explosion ruptured other process pipes.

blocking-in the flow of ethylene. This accident could have been mitigated if the gas detection analyzers alarmed at lower concentrations. Twelve buildings were destroyed. Ignition may have been by static electricity. Gaseous compositions are controlled carefully to keep the concentrations outside the explosion limits. a methane diluent. 20W. A good inspection and maintenance program would have prevented the accident. Ethylene ExplosionlY Ethylene was accidentally released from a 118-in stainless steel instrument tubing line leading to a gauge from a main line on a compressor system. H. 38-42. which was designed to wash the ethylene from the atmosphere. and recycled carbon dioxide to a continuous reactor. I9Garrison. The ethylene detonated in the engine.000 lb of gas had already escaped. Ethylene Explosion ls A drain fitting in a high-pressure (40 kpsi) compressor line broke." Instrument Technology (October 1972). "Instrument-Connected Losses in the CPI. Doyle. Ethylene Oxide E x p l o ~ i o n ~ ~ ) Ethylene oxide is produced by adding ethylene. One Hundred Largest Losses. in the layout design it should have been noticed that the compressor needed special consideration to eliminate this ignition source. The damage was estimated at over $15 million. p. 8. it failed to sound an alarm because of a faulty relay in the control room. This accident emphasizes the importance of adding gas detectors that measure flammable gases at low concentrations so that alarms and block valves can be actuated before large quantities of gas are released. However. but not before 45011. The ethylene cloud drifted and entered the intake system of an engine that was driving one of the compressors. This accident caused $21. Automatic equipment promptly detected the hazardous vapor and operated the automatic high-density water-spray system. The leak was too large for the spray system to handle.One Hundred Largest Losses. The tubing failed as a result of transverse fatigue caused by vibration from the reciprocating compressor. 3. . Automatic fail-safe valves functioned properly. Also. allowing ethylene to escape. IXGarrison. The explosions were felt 6 miles away. oxygen. and this explosion ignited the rest of the vapors.548 Chapter 13 Case Histories Automated block valves would have minimized damage in this fire. p. The unmanned compressor building was equipped with a combustible gas detection system. pp. and fire and explosion damage occurred throughout the polyethylene plant.8 million in damage.

A flange leak will douse the cables with liquid. r 22Walter B. all valve vent lines are now covered with bug screens to prevent blockage. Do not install pipes underground. "Process Safety Technology and the Responsibilities of Industry. Do not install liquid-filled flanges above electrical cables. The bleed line was opened and was noted on the control panel. Check all purchased instruments and equipment for integrity and functionality. In this particular case the accident could also have been prevented with appropriate inspection and maintenance procedures. 2 1 T r e v ~Kletz. Learning from Accidents in Industry (Boston: Butterworths. p. The gases in the ethylene oxide reactor moved into the explosive region while being above the autoignition temperature. It is now an industrial standard to use positive identification of the valve position on all important safety valves -limit switches that are tripped when the valve is open or shut. A violent explosion occurred. The air. Lessons Learned The case histories related to system design emphasize that (1) accidents occur rapidly. Trevor KletzZ1and Walter B. 25-33. The following recommendations also include design features from our own experiences: Use the appropriate materials of construction. Do not secure pipes too rigidly. requiring only minor process changes. welds) meets the required specifications. In addition. Pipes must be free to expand so that they will not damage other parts of the system. Howard. 1988).13-3 System Designs 549 One plant experienced an emergency situation. Although the vent valve was open. usually with inadequate time to manually return the system to control once the accident scenario is in progress. Be sure that the quality of construction (for example. the air could not escape. . pp. Howard22 have emphasized the special design features for safer plants. especially when using old systems for new applications. resulting in several injuries and significant plant damage. as indicated on the control panel. however. The oxygen control valve was normally closed by bleeding air out of the valve bonnet diaphragm (air to open)." Chemical Engineering Progress (September 1988). did not bleed off through the bonnet vent because a mud dauber wasp constructed mud cells over the vent hole. and (3) the time and effort required to develop a safe system design is justified: An engineer is hired for a fraction of the cost of most accidents. (2) the system designs required for preventing accidents or mitigating the consequences of accidents are frequently subtle. 143. The emergency procedures specified: Close the oxygen feed valve.

there is no opportunity to adjust or improve its operation. Use electronic or mechanical level gauges. Do not install traps in lines where water can collect and develop a corrosion problem. . Design systems for easy operation and easy maintenance. Install bellows carefully and according to manufacturers' specifications. Be sure that all tracing is covered. Also. for emergency shutdowns. Provide hand-operated or automatic block valves. Be sure that all lines that can catch water can be appropriately drained. not glass sight glasses. and replace them with properly designed welded plugs. Remove all temporary supports after construction is completed. When welding reinforcement pads to pipes or vessels. Example 13-3 Analyze the first ethylene explosion example (318-in fitting failure) to determine the percentage of fuel that actually exploded compared to the quantity of ethylene released in a vapor cloud. install manual valves within easy reach of the operators. inspect frequently and replace when necessary before they fail. Bellows should be used cautiously. Install bug screens on vent lines. nipples. Check to ensure that all equipment is assembled correctly. and plugs.550 Chapter 13 Case Histories Provide adequate supports for equipment and pipes. for example. Do not allow spring supports to be completely compressed. and design pipe networks for easy maintenance or with easy access to equipment requiring maintenance. Do not use screwed joints and fittings when handling hazardous chemicals. Critical safety instruments must have backups. or equivalent valves. Safety technology must work right the first time. or depressions where water can accumulate. If required. Add interlocks to decrease pressure before the doors can be opened. add visible pressure gauges at the doors. Remove all temporary startup or checkout branches. Do not install relief tailpipes too close to the ground where ice blockage may make them inoperable. trenches. ensure that trapped air can escape through a vent during heating. Make static and dynamic analyses of pipe systems to avoid excessive stresses or excessive vibrations. Add fail-safe block valves with a positive indication of the valve position (limit switches). Usually. Do not install pipes in pits. Design doors and lids so that they cannot be opened under pressure. Make structural analyses of relief systems to avoid structural damage during emergency reliefs. Do not let pipes touch the ground.

Therefore the fraction of energy manifested in the explosion is 0. In the study of these case histories.69 ton of TNT m . The higher energy conversion is a result of the explosion occurring in a partially confined area. The management systems commonly used in industry include safety reviews.13-4 Procedures 551 Solution The total energy contained in the vapor cloud is estimated by assuming the heat of combustion (appendix B). vm AHc ~ T N = T ETNT where Therefore (1)(227. The combustion reaction is Therefore the theoretical energy is AHc = 1411. The causes of all accidents can ultimately be attributed to a lack of management systems. = Based on the accident investigation. the explosive energy was equivalent to 0. is developed by implementing management systems to prevent the existence of safety problems in the first place. one must recognize that the existence of procedures is not enough.000 g)(12. 13-4 Procedures An organization can develop a good safety program if it has personnel who can identify and eliminate safety problems.69 = 11. An even better safety program. This 11. and maintenance procedures.3 ton TNT. Case histories that especially demonstrate this problem are illustrated in this section.046 callg) ~ I N = T 1. operating procedures. The tons of TNT based on this heat of combustion are calculated using Equation 6-24.2 kJ/mol = 12046 callg.2% is considerably higher than the 2% normally observed (see section 6-13) for unconfined vapor cloud explosions. . however.312.017 X 1OYcal/ton 2.2%.

The foreman and three employees set to work 23 Case Histories. with the key in the mechanics pocket. The main switch was left energized. 2 5 C a ~Histories. prevents this type of accident. Fortunately. As one mechanic was completing his work inside the mixer. p. A short time later. another operator on an adjoining floor pushed. one of the other start-stop buttons. The accident investigation found that the leak test should have been conducted with a hydraulic procedure and not air and that the vessel should have been protected with a relief device. They were instructed to use 5 psi of air pressure and a soap solution to identify the leaks. The gauge was incorrectly chosen for vacuum service and not pressure because the vacuum identifier was small. the fitters should have taken more time to check out the gauge to ensure that it was correct for this application. Vinyl Chloride E x p l o s i ~ n ~ ~ Two vinyl chloride polymerization reactors were being operated by the same team of operators. In addition. v. Man Working in Vessel24 Two maintenance workers were replacing part of a ribbon in a large ribbon mixer. p. A padlocked switch at the starter box disconnect. the mixer was stopped with one of three start-stop buttons. this is the "try" part of the lock-tag-and-try procedure. and the two fitters escaped injury. as the fitters were carrying out the tests.and used effectively. there was no fragmentation of the metal. v. by mistake. e . killing the mechanic between the ribbon flight and the shell of the vessel. 2. The mixer started. 225. 186. 113. Lock-tag-and-try procedures were developed to prevent accidents of this kind. 2.552 Chapter 13 Case Histories There must also be a system of checks in place to ensure that the procedures are actually used . Leak Testing a Vessel23 A 2-ft-diameter float was fabricated using stainless steel and welded seam construction. v. and reactor 4 was nearly full of monomer and in the polymerization phase. A busy instrument worker gave them a gauge. the mechanic should also verify the dead circuit by testing the push-button at all switches. p. the float ruptured violently. They clamped a 100-psi air hose to a nipple on the tank. 24CaseHistories. After the switch gear lockout. Reactor 3 was in the cool down and dump phase of the process. 2. Pipefitters were given the job of checking the welds for leaks.

The runaway reaction was triggered when. Before this particular start-up. In this case the pressure luckily did not exceed the vessel rupture pressure. for example. Report 550-F99-004 (August 1999). or adequate safeguards should be added to the system to prevent the specific hazard scenario. p. *'EPA.epa . a double block valve arrangement was installed in the bottom discharge line. that the second valve created a dead space between the two block valves and that water was trapped between them. The gaseous vinyl chloride monomer just in the process of polymerization burst out of the vessel. Phenol-Formaldehyde Runaway Reaction27 A plant had a runaway reaction with a phenol-formaldehyde polymerization reaction. 20. A bleed line and possibly a nitrogen blow-out line would have prevented the accumulation of this water. could have been prevented if the plant had used a safety review procedure during the design phase of this plant modification. presumably ignited by a spark from an electric motor or by static electricity generated by the escaping gas. . Dangerous Water E x p a n ~ i o n ~ ~ A hot oil distillation system was being prepared for operation. If these scenarios are possible. 1984).gov/ceppo/. A valve at the bottom of the tower was opened to initiate the transfer of heavy hot oil to a process pump.13-4 Procedures 553 to discharge the contents of reactor 3. When the bottom valve was opened. and shortly afterward exploded violently. Modern plants use interlocks or sequence controllers and other special safeguards to prevent this type of error. Problems similar to this are usually identified in safety reviews. and all the trays dropped within the tower. This accident resulted in four fatalities and ten injuries in and around the plant. The steam created excessive pressures at the bottom of the tower. It was not realized. booklet 1 (Chicago: Amoco Oil Company. The accident could have been prevented with better operating procedures and better training to make the operators appreciate the consequences of mistakes. HOW to Prevent Runaway Reactions. the tower was destroyed by this accident. Flashing steam surged upward through the tower. all the raw materials and 26Hazards of Water. relief valves should also be installed to mitigate these events. Available at www. however. This accident. The result was one fatality and seven injuries and environmental damage. contrary to standard operating procedures. filled the room. Although no injuries were sustained. The temperature was gradually raised to 500°F. the pocket of water came in contact with the hot oil. Consequences of contaminating hot and high boiling liquids with low boilers can be estimated using thermodynamics. but in error they opened vessel 4 instead.

The root cause of this accident was poor operating procedures and poor process info\r mation. . for example. This company's initial research for the process identified and described two exothermic chemical reactions: (1) The desired exothermic reaction is initiated at an onset temperature of 38"C.554 Chapter 13 Case Histories catalyst were charged to the reactor at once. followed by the addition of heat. (3) installing the appropriate controls and safety features to safely manage these reactive chemicals. Available at http://www. The other root causes were (1) the poor understanding of the chemistry. and it proceeds rapidly at 75°C.gov/ reports/2000/rnorton/inde.htm. and (3) the use of cooling water to control the reaction rate was not initiated soon enough. The primary reason for this accident was the lack of administrative controls to ensure that the standard operating procedures were used appropriately and that the operators were trained. but they did not include information on the undesirable decomposition reaction.u. and (3) no safeguard controls to prevent runaway reactions. (4) revising the operating procedures and training for handling these reactive chemicals to include descriptions of the possible consequences of deviations from normal 2XCSB. ortho-nitrochlorobenzene (0-NCB) and 2-ethylhexylamine (2-EHA). causing an explosion that blew the hatch off the reactor and allowed the release of the contents from the vessel. did not cover the safety consequences of deviations from the normal operating conditions.chemsufety. The investigation team found that the reaction accelerated beyond the heat-removal capacity of the reactor. Information on their MSDS was also misleading (mentioning a lower reactivity and a much lower boiling point than the actual values). (2) evaluating relief requirements using the appropriate technology published by the Design Institute for Emergency Relief Systems (DIERS) (see details in chapters 8 and 9). The operating plant was not aware of the decomposition reaction. A runaway reaction caused an explosion and flash fires that injured nine workers. The plant's operating and process information described the desired exothermic reaction. The recommendations from the investigation included (1) revalidating the safety data for all reactive chemicals. The resulting high temperature led to a secondary runaway decomposition reaction. such as the possibility of a runaway reaction and the specific steps to be taken to avoid or recover from such deviations. Chemical Manufacturing Incident. (2) an inadequate risk analysis. The operating procedure. The runaway reaction was the result of the following factors: (1) The reaction was started at a temperature higher than normal. Conditions and Secondary Reaction Cause Expl0sion2~ A plant manufactured a dye by mixing and reacting two chemicals. (2) an undesirable decomposition (the dye) reaction has an onset temperature of 195°C. This EPA case history also summarized seven similar accidents with phenol-formaldehyde reactions during a 10-year period (1988-1997). Report 1998-06-I-NY. (2) the steam used to initiate the reaction was left on for too long.

On the day of the accident. Available at www. two workers poured four drums of liquid waste into the blending vessel . perchlorates. and other flammable and combustible wastes. (3) properly training employees in the processes they work on using the standard operating procedures for the processes and job tasks. and nitrites. Before this accident the operating procedures included (1) adding about 500 gal of solvent before starting the agitator. Report 550-F00-001. halogenated hydrocarbons. (5) implementing a program to investigate and document safety incidents. (4) It was known that the addition of oxidizers could be hazardous if the oxidizers were added without a large quantity of liquid fuel in the blend tanks. Then they added solids into the top of the tank: about 2 lb each of chlorates. and (6) revising the MSDSs and distributing them to anyone needing this information.gov/ ceppoL . The potential consequences of mixing such incompatible materials are violent reactions. Prevention of Reactive Chemical Explosions. In the EPA's report of the investigation it was stated that strong oxidizers are generally considered incompatible with many organic substances because of the potential for dangerous reactions. The explosion and resulting fire caused extensive damage to the facility. (2) evaluating the chemical and process hazards before starting a process or procedure that has been changed or modified. One worker was killed and two others were injured. This facility had two 1000-galblend tanks to blend waste solvents. who died. and (6) ensuring that all employees understand the hazards of the chemical process. (5) ensuring that chemical and process hazards are understood and addressed. (3) Oxidizers were added only after the vessel was three-quarters full of solvent and the agitator was running. perchlorates. and a small quantity of oxidizers. The flammable vapor exploded. "EPA. nitrites. The EPA's recommendations for the prevention of this type of accident included (1) establishing standard operating procedures that are essential for safe operation.13-4 Procedures 555 operating conditions and the steps taken to correct the resulting problems. (4) ensuring that the chemicals and reaction mechanisms associated with the substances mixed or blended are well understood and documented. other organic compounds and solvents. liquid suddenly erupted out of the vessel manway. including emergency response action. fires. and explosions. according to an unwritten procedure. engulfing one employee.epa. including perchlorates. Chlorates. and injuring two others. cleaners. Thirty to 60 seconds after the oxidizers were added and while a fifth drum of solvent was being dumped into the top of the reactor. (2) No inert gas blanketing was used to lower the vapor concentration to below the LFL.about half the amount needed to reach the agitator. Fuel-Blending Tank Explosi0n2~ An accident occurred in a fuel-blending facility that provided a way to reuse flammable and hazardous wastes. and chlorates. and other strong oxidizers are potentially incompatible with alcohols.

When reviewing case histories it is apparent that safety procedures and standard operating procedures are needed to help the chemical industry (1)eliminate injury to personnel.14 ft'. the worst-case scenario assumes that all the water vapor collects beneath the bottom tray. 182-188. Example 13-4 Using the dangerous water expansion example. 13-5 Conclusion This chapter on case histories is brief and does not include all the lessons relevant to accidents. emergency. compute the approximate pressures that were developed in the bottom of this column.- 556 Chapter 13 Case Histories Lessons Learned Procedures are sometimes incorrectly perceived as bureaucratic regulations that impede progress. Kletz. . Assuming a tray spacing of 1 ft. case histories and safety literature are of no value unless they are studied. What Went Wrong? Case Histories of Process Plant Disasters (Houston: Gulf Publishing. Never use gas to open plugged lines. Use procedures effectively (lock-tag-and-try. Train operators and maintenance personnel to understand the consequences of deviations from the norm. (2) minimize incapacitating damage to facilities. 1985). additional lessons are identified3? Use a permit procedure for opening vessels that are normally under pressure. understood. Assume a column diameter of 2 ft. The references provide excellent information for more studies.A. hot work. we obtain = 1522 psia if all the water vaporized. the volume under the first tray is 3. Using an equation of state. Communicate operating changes to other operations that may be affected by the change. pp. vessel entry. Solution The areas of the column trays are 3.14 ft2. Use safety review procedures during the design phases of projects. and used appropriately. 30T. There is significant information in the open literature. However. In the review of case histories relevant to procedures. Make periodic and precise audits of procedures and equipment. a water slug of 1gal. and (3) maintain steady progress. and a column pressure of 10 psia. including new installations or modifications to existing systems. and the like). If the tray vapor paths are small openings.

Trevor A. v. Problems 13-1. July 1962). 3 (April 1970). Repeat Problem 13-1 for the vinyl chloride explosion described in section 13-4. 2d ed. Lees. Kletz. Managing Change in Chemical Plants: Learning from Case Histories (London: Butterworth-Heinemann." Chemical Engineering Progress (July 1989). (London: Butterworths. Therefore the maximum pressure is 680 psi if some water remains as liquid water. 8-26. Frank P. . 1993). E. This example explains why all the column trays were torn away from the supports and also illustrates