United States
Environmental Protection
Agencv
EPA-600 / 8-87-039b
January 1989
&EPA Research and
Development
PREVENTION REFERENCE MANUAL:
CONTROL TECHNOLOGIES
VOLUME 2. POST-RELEASE MITIGATION
MEASURES FOR CONTROLLING
ACCIDENTAL RELEASES OF AIR TOXICS
Prepared for
Office of Air Quality Planning and Standards
Prepared by
Air and Energy Engineering Research
Laboratory
Research Triangle Park NC 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development. U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are;
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the SPECIAL REPORTS series. This series is
reserved tor reports which are intended to meet the technical information needs
of specifically targeted user groups. Reports in this series include Problem Orient-
ed Reports. Research Application Reports, and Executive Summary Documents.
Typical of these reports include state-of-the-art analyses, technology assess-
ments, reports on the results of major research and development efforts, design
manuals, and user manuals.
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents necessarily
reflect the views and policy of the Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service. Springfield. Virginia 22161.
-------
EPA-600 /8-87-039b
January 1989
PREVENTION REFERENCE MANUAL:
CONTROL TECHNOLOGIES
VOLUME 2: POST-RELEASE MITIGATION
MEASURES FOR CONTROLLING ACCIDENTAL
RELEASES OF AIR TOXICS
By:
0. S. Davis
G. B. DeWolf
K. A. Ferland
D. L. Harper
R. C. Keeney
J. 0. Quui
Radian Corporation
Austin, Texas 78720-1088
Contract No. 68-02-3994
Work Assignment 102
and
Contract No. 68-02-4286
Work Assignment 41
EPA Project Officer
T. Kelly Janes
Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park. North Carolina 27711
Prepared for:
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
-------
ABSTRACT
Reducing the probability of accidental toxic chemical releases reduces
the possibility of harm to human health and to the environment. When such a
release does occur, however, its consequences must be reduced. This can be
accomplished by means of a variety of "mitigation" measures that contain,
capture, destroy, divert, or disperse the released toxic chemical.
Mitigation measures begin with the initial siting and layout of a
facility to decrease the area that would be affected by a release. The extent
of the areas potentially affected, concentrations of toxic chemicals reaching
those areas, and the duration of exposure can be estimated by vapor or gas
dispersion modeling. Detection and warning systems give the first notice of a
release, and the actual extent and magnitude can be determined by
meteorological instruments. These systems, along with emergency planning and
training, are the first steps in the total mitigation process. Other measures
involve the use of specific mitigation techniques, such as leak plugging,
containment systems, and spray or foam systems.
Mitigation measures are described and discussed in terms of their
applicability, performance, state of development, and general application
costs. This manual is part of a multi-volume series of manuals that address
the prevention of accidental releases of toxic chemicals.
ii
-------
ACKNOVLEDG EMENTS
This manual was prepared under the overall guidance and direcCion of T.
Kelly Janes, Project Officer, with the active participation of Robert P.
Hangebrauck, William J. Rhodes, and Jane M. Bare, all of U.S. EPA. In
addition, other EPA personnel served as reviewers. Radian Corporation
principal contributors were Graham E. Harris (Program Manager), Glenn B.
DeVolf (Project Director), Daniel S. Davis, D. L. Harper, R. C. Keeney,
Jeffrey D. Quass, and Sharon L. Wevill. Contributions were also made by other
staff members. Secretarial support was provided by Roberta J. Brouwer and
Bonita Cross. Special thanks are due many other people, both in government
and industry, who served in the Technical Advisory Group and as peer
reviewers.
iii
-------
CONTENTS
Abstract ^7
Acknowledgements
Figures viii
Tables x
1. Introduction 1
1.1 Fundamental Concepts 2
1.2 Organization of the Manual 5
1.3 References 6
2. Emergency Planning and Training 7
2.1 Background 1
2.2 Emergency Planning 1
2.2.1 Description 1
2.2.2 Implementation 12
2.3 Training 18
2.3.1 Training Overview 18
2.3.2 Training Coordination with Outside Agencies 20
2.3.3 Emergency Exercise Program 21
2.4 References 26
3. Facility Siting and Layout 27
3.1 Background 27
3.2 Siting 28
3.2.1 Population Density Near Facility Site 29
3.2.2 Meteorology and Climate 31
3.2.3 Emergency Response 32
3.2.4 Topography and Drainage 33
3.2.5 Accessibility to Off-Site Emergency Response
Vehicles 33
3.2.6 Utility Service 33
3.3 Layout 34
3.3.1 Adequate Spacing Between Units 35
3.3.2 Grouping or Isolation of Facility Units 38
3.3.3 Transportation 39
3.3.4 General Layout 39
3.4 Costs 41
3.5 References 43
4. Detection and Warning Systems 45
4.1 Background 46
iv
-------
4.2 Description 47
4.2.1 Combustible/Flammable Gas Detectors 48
4.2.2 Toxic Vapor Detection Systems 51
4.2.3 Remote Monitoring 53
4.3 Design 55
4.4 Applicability and Performance 57
4.5 Reliability 57
4.6 Secondary Hazards 58
4.7 Costs 58
4.8 References 60
5. Vapor Dispersion Modeling 62
5.1 Background 62
5.2 Descriptions 64
5.2.1 Source Models 65
5.2.2 Dispersion Models 69
5.3 Applicability and Performance 74
5.3.1 Source Models 76
5.3.2 Dispersion Models 78
5.4 Reliability 81
5.5 Costs 82
5.6 References 84
6. Meteorological Instrumentation 86
6.1 Background 86
6.1.1 Historical Data 87
6.1.2 Real-Tim* Meteorological Data 89
6.2 Description 89
6.2.1 Wind Direction 90
6.2.2 Wind Speed 90
6.2.3 Ambient Temperature 91
6.2.4 Stability 91
6.2.5 Meteorological Towers 92
6.2.6 Data Acquisition Systems 92
6.3 Applicability and Performance 92
6.3.1 Instrument Selection 93
6.3.2 Siting of Meteorological Instrumentation 94
6.4 Reliability 97
6.5 Costs 98
6.6 References 99
7. Secondary Containment 100
7.1 Background 100
7.2 Description 102
7.2.1 Leak-Plugging Techniques 102
7.2.2 Physical Barriers and Containment Systems 108
7.3 Applicability and Performance 114
7.3.1 Applicability and Performance of Leak-Plugging
Devices 114
v
-------
7.3.2 Applicability and Performance of Physical
Barriers and Containment Systems 122
7.4 Reliability 128
7.4.1 Reliability of Leak-Plugging Techniques 128
7.4.2 Reliability of Physical Barriers and
Containment Systems 129
7.5 Secondary Hazards 129
7.5.1 Secondary Hazards of Leak-Plugging Techniques... 129
7.5.2 Secondary Hazards of Physical Barriers and
Containment Systems 130
7.6 Costs I30
7.7 References 133
8. Spray, Dilution, and Dispersion Systems 135
8.1 Background 136
8.1.1 Spray Systems 136
8.1.2 Steam Curtains 137
8.2 Description 137
8.2.1 Spray Systems 138
8.2.2 St*am. Curtains
8.3 Design 149
8.3.1 Spray Systems 1^9
8.3.2 Steam Curtains 152
8.4 Applicability and Performance 155
8.4.1 Spray Systems 155
8.4.2 Steam Curtains 161
8.5 Reliability 163
8.5.1 Spray Systems 163
8.5.2 Steam Curtains 164
8.6 Secondary Hazards 164
8.6.1 Spray Systems 164
8.6.2 Steam Curtains 165
8.7 Costs 165
8.7.1 Spray Systems 165
8.7.2 Steam Curtains 168
8.8 References 170
9. Foam Systems 173
9.1 Background 173
9.2 Description 176
9.2.1 Types of Foams 176
9.2.2 Foam Quality 181
9.2.3 Foam Application Systems 181
9.3 Design 182
9.3.1 Application Equipment and Methods for
Low-/Medium-Expansion Foams 183
9.3.2 Application Equipment and Methods for
High-Expansion Foams 184
vi
-------
9.4 Applicability and Performance 185
9.5 Reliability 196
9.6 Secondary Hazards 197
9.7 Costs 198
9.8 References 201
APPENDIX A - Metric (SI) Conversion Factors 202
vii
-------
FIGURES
1-1 The role of various accidental release control measures in
reducing the consequences of an accidental release 3
4-1 Schematic representations of typical remote monitoring
techniques for hazardous vapors 54
5-1 Potential release scenarios 67
7-1 Foam-injected leak plugging device 1-03
7-2 Examples of mechanical leak repairing devices offered by the
Chlorine Institute 105
7-3 Jacket that bolts around a leaking pipe, forming a
leak-tight seal 106
7-4 Mechanical sealing ring with Injected chemical sealant to
fit around and seal a leaking flange 107
7-5 Conceptual diagram of a typical low-wall dike 110
7-6 Conceptual diked storage facility with integrated
containment sump 113
7-7 Conceptual diagram of potential layouts for a remote
neutralization basin system 115
7-8 Conceptual diagram of a double-walled tank system 116
7-9 Envelope of practical applications for sealing leaks in
nonsubmerged containers 119
7-10 Vapor dispersion patterns for a release of chlorine from a
2-inch hole in a 10,000 gallon refrigerated storage tank 126
7-11 Vapor dispersion patterns for a release of chlorine from a
2-inch hole in a 10,000 gallon refrigerated storage tank
surrounded by a 25 ft diameter dike 127
8-1 Conceptual diagram of water spray barrier using downward
spraying nozzles jjg
viii
-------
8-2 Conceptual diagram of fixed-spray barrier surrounding
protected area 140
8-3 Conceptual diagram of fixed-spray barrier with activation
according to wind direction 142
8-4 Conceptual diagram of water sprays incorporated into
semi-enclosure 143
8-5 Typical mobile water spray system 144
8-6 Conceptual diagram of mobile water spray barrier and
associated "chimney" effect 146
8-7 Conceptual diagram of a steam curtain 148
8-8 Typical hollow-cone spray nozzle 153
8-9 Typical fan-tail spray nozzle 154
8-10 Conceptual diagram of steam jets emerging from steam
curtain 156
9-1 Benzene vapor concentration versus time 195
ix
-------
TABLES
2-1 Example Outline of a Community Emergency Response Plan 13
2-2 Exercise Types- -Advantages and Disadvantages 25
4-1 Comparison of Characteristics of Various Combustible/
Flammable Vapor Detection Techniques 50
4-2 Examples of Instruments Available for Various Toxic
Chaaicals 52
4-3 Examples of Estimated Costs 59
5-1 Examples of Source Models 68
5-2 Dispersion Models 71
5-3 Source and Dispersion Modal Packages 72
5-4 Commercial Modeling Packages 73
7-1 Hazardous Chaaicals Tasted and Small-Scale Sealing Test Results.. 118
7-2 Costs of Various Leak-Plugging Devices (1986 Dollars) 131
8-1 Results of Several Experimental Studies on the Effectiveness
of Water Sprays 158
8-2 Effectiveness of Steam Curtains 162
8-3 Typical Costs Associated with Fixed and Mobile Water Spray
Systems 166
8-4 Equipment Specifications for Mobile and Fixed Water Spray
Systems Used in Cost Estimates 167
8-5 Estimated Costs for a Typical Steam Curtain System 169
9-1 Foams for Control of Flamable/Toxic Vapors 187
9-2 Performance Sunary of Three Universal and Specialty Foams 194
9-3 Approximate Costs for Various Foams 199
9-4 Examples of Foam Application Equipment Costs 200
x
-------
SECTION 1
INTRODUCTION
Increasing concern about Che potentially disastrous consequences of major
accidental releases of toxic chemicals resulted from the Bhopal, India, methyl
isocyanate release on December 3, 1984, which killed approximately 2,000
people and injured thousands more. Concern about the safety of process
facilities that handle hazardous materials increased further after the
accident at the Chernobyl nuclear power plant in the Soviet Union in April of
1986.
While headlines of these incidents have created the current awareness of
toxic release problems, other, perhaps less dramatic, incidents have occurred
in the past. These accidents contributed to the development of the field of
loss prevention as a recognized specialty area within the general realm of
engineering science. Interest in reducing the probability and consequences of
accidental toxic chemical releases that might harm workers within a process
facility and people in the surrounding community prompted preparation of this
manual and a series of companion manuals.
This manual, which addresses the post-release mitigation measures for
toxic chemical releases, is designed to give the reader useful information
about the state of the art in this area. Post-release mitigation measures are
defined as any measures that may reduce the consequences of an accidental
toxic chemical release after it has occurred. Mitigation measures decrease
the quantity of a released chemical that can reach receptors, decrease the
area exposed to the chemical, and/or reduce the duration of exposure. The
emphasis is on post-release mitigation measures for facilities that handle
toxic chemicals so that the potential risk to surrounding communities can be
minimized.
1
-------
Companion volumes to this manual include:
Prevention Reference Manual:
Prevention Reference Manual:
Prevention Reference Manual:
User's Guide Overview for Controlling
Accidental Releases of Air Toxics (1)
Chemical Specific, Volume 8: Control
of Accidental Releases of Hydrogen
Fluoride (3)
Control Technologies, Volume 1:
Prevention and Protection Measures for
Controlling Accidental Releases of Air
Toxics (2)
The series also contains several chemical-specific manuals on controlling
accidental releases of specific chemicals. An example is the Prevention
Reference Manual: Chemical Specific, Volume 8: Control of Accidental
Releases of Hydrogen Fluoride (3).
1.1 FUNDAMENTAL CONCEPTS
The physical release of a chemical is the final event in a sequence of
events leading to the release. Such a sequence of events begins with the
initiating, or primary, event and propagates through enabling events. The
final event can be prevented by preventing the initiating event or the
enabling events. Success in preventing a release depends on correctly
identifying primary and enabling events and event chains, on knowing the
relative probability of the events, and on the skill and knowledge of the
individuals charged with conducting the analysis. If prevention and
protection meausures fail and a release occurs, mitigation measures must be
invoked.
Figure 1-1 illustrates the place of mitigation, among other measures, in
reducing the ultimate consequences of a potential release. As seen in the
figure, mitigation can be viewed as one of several barriers between a release
hazard and realization of its consequences. Before appropriate mitigation
2
-------
RELEASE
HAZAROS
HAZARO IDENTIFICATION
PRE • RELEASE PREVENTION
m
PRE • RELEASE PROTECTION
I
POST • RELEASE MITIGATION
COMMUNITY RESPONSE
ULTIMATE CONSEQUENCES
Figur* 1-1. Tha role of various accidental release control atuuru
in reducing the consequences of an accidental ralaaaa.
3
-------
measures can be determined, the potential consequences of a release must be
determined. The mitigation measures addressed in this manual include: 1)
emergency planning; 2) siting and layout; 3) dispersion modeling; 4) detection
and warning systems; 5) meteorological instrumentation; and 6) technical
mitigation measures that can effectively control a release.
Emergency planning and training ensure the rapid and correct response of
the people charged with applying other mitigation measures and define what
mitigation measures should be employed.
Plant siting and layout concerns the placement of hazardous facilities
relative to sensitive receptors in the surrounding community and within plant
boundaries. Important considerations in this area, besides the obvious
distance factor, include taking advantage of terrain features such as hills
that might act as natural barriers, and avoiding the funneling effects of
valleys.
Detection and warning systems give advance notice that a release is
incipient or has occurred and define the magnitude and location of the release
so that other mitigation measures can be taken. Meteorological
instrumentation serves a similar purpose by providing data needed to monitor
the physical location and movement of a release and the ambient
characteristics that affect the movement and dispersion of a vapor or gas
plume or cloud.
Vapor dispersion modeling is used to predict the extent, duration, and
concentration of the plume or cloud of released toxic vapor or gas. Numerous
dispersion models of varying levels of sophistication, accuracy, and
verification by actual field data are available. The results predicted bv
these models depend on a source term which describes the characteristics of
the initial release, and a dispersion term chat describes the characteristics
of the resultant cloud or plume. Characteristics of the release are primarily
process and process equipment related, while dispersion characteristics are
related to the properties of the chemical vapor or gas and to meteorological
4
-------
conditions. Such models can inform decisions about plant siting and layout,
the placement of detection and warning systems and meteorological
instrumentation, and the selection of technical mitigation measures.
Containment systems reduce the area that could be exposed to vapors from
a release and confine the liquid portion of a release until measures can be
taken to recontain or destroy the released chemical. While some containment
may be successful with gases, containment is probably more applicable to
spilled volatile liquids.
Spray systems disperse, dilute, and absorb a released airborne chemical.
Spray systems rely on fixed or mobile equipment that applies a spray of water,
other materials, or a condensing cloud of steam directly to the cloud or plume
of noxious chemical. Some spray systems for toxic gas releases are similar to
fire fighting systems.
Foam systems are primarily used to contain volatile liquid evaporation
from pools of spilled liquid. These systems are based on specialty chemical
materials that generate foams whose specific characteristics are tailored to
the chemical properties of the material to which they are applied. Foams ace
as a physical barrier to prevent or decrease evaporation from liquid surfaces.
1.2 ORGANIZATION OF THE MANUAL
The remainder of this manual is divided into 8 sections, beginning with
Section 2. Sections 2 through 6 deal with mitigation measures that enable a
plant to prepare for an accidental release and to respond appropriately if one
occurs. Sections 7 through 9 discuss the technologies used to control the
liquid and vapor potions of an accidental release.
Section 2 addresses emergency planning and training, which are essential
components of the mitigation effort. In Section 3, facility siting and layout
as related to mitigation are discussed. Detection and warning systems within
the facility and outside the facility boundary are discussed in Section 4.
5
-------
Seccion 5 covers the basic types of vapor dispersion models and discusses the
applicability of these models to the mitigation problem. Meteorological
instrumentation, as described in Section 6, allows real-time estimates of
where, when, and to what extent a release will affect the community.
Section 7 introduces the topic of secondary containment measures, or
those measures that can control the spread of a toxic release once it occurs.
The last three sections of the manual discuss three technologies that may be
used to reduce hazardous vapor concentrations, thereby minimizing their effect
on surrounding communities. These technologies are spray and steam systems,
covered in Section 8, and foam systems, covered in Section 9.
Appendices contain tables for converting between English units of measure
(used throughout the manual) and SI metric units. A glossary of terms also
appears in the appendix.
1.3 REFERENCES
1. Davis, D.S., G.B. DeWolf, and J.D. Quass. Prevention Reference Manual:
User's Guide Overview for Controlling Accidental Releases of Air Toxics.
EPA-600/8-87-028 (NTIS PB87-232112), U.S. Environmental Protection
Agency, AEERL, Research Triangle Park, North Carolina, 1987.
2. Davis, D.S., G.B. DeWolf, and J.D. Quass. Prevention Reference Manual:
Control Technologies, Volume 1. Prevention and Protection Technologies
for Controlling Accidental Releases of Air Toxics. EPA-600/8-87-039a
(NTIS PB87-229656), U.S. Environmental Protection Agency, AEERL, Research
Triangle Park, North Carolina, 1987.
3. Davis, D.S., G.B. DeWolf, and J.D. Quass. Prevention Reference Manual:
Chemical Specific, Volume 8. Control of Accidental Releases of Hydrogen
Fluoride. EPA-600/8-87-034h (NTIS PB87-234530), U.S. Environmental
Protection Agency, AEERL, Research Triangle Park, North Carolina, 1987.
6
-------
SECTION 2
EMERGENCY PLANNING AND TRAINING
Emergency planning and craining are an integral part of mitigation.
Reducing the consequences of an accidental release depends on a timely,
effective response by people at the scene who know the what, when, where, and
how of various countermeasures. This Includes the application of various
technical mitigation measures discussed elsewhere in this manual, as well as
calling for assistance, notifying authorities, and, if necessary, evacuating
people from the affected areas.
2.1 BACKGROUND
Numerous recent publications on emergency preparedness include sections
on emergency planning and training. The U.S. EPA has published a guidance
manual on the Chemical Emergency Preparedness Program (1) . The Chemical
Manufacturer's Association has sponsored the Community Awareness and Emergency
Response (CAER) program (2,3). Other publications also deal with the issue
(e.g., Reference 4).
2.2 EMERGENCY PLANNING
2.2.1 Description
The details of emergency planning and training will vary from facility to
facility and community to community. In general, however, any program should
address certain fundamental elements:
e Program initiation;
7
-------
• Hazard evaluation;
• The identificacion, evaluation, selection, and implementation of
countermeasures;
• Resource requirements and availability;
• Organization; and
• Mobilization and demobilization.
Emergency programs may be initiated by facility or corporate personnel or
by an outside authority. Once the program is started, an overall plan can be
developed to address each of the other fundamental elements listed above. The
plan should address community involvement as well as on-site emergency
planning and response activities. This is discussed further in Subsection
2.2.2.
Evaluating the potential hzards of a particular facility, which involves
both hazard identification and analysis, is an essential starting point in
actual plan development. Hazard identification begins with a list of desig-
nated chemicals, a determination of which ones are present at the facility,
the quantities, and their physical states. The identification of potential
release hazards is based on physical and toxicological data and on how the
chemicals are used at the facility. The potential hazards are then analyzed
to determine how release incidents might occur and what the consequences might
be. These topics are examined more fully in other manuals in the Prevention
Reference Manual series, as well as in the technical literature. Various
formal procedures can be followed to more effectively determine the relative
probability and consequences of various release scenarios.
8
-------
Once Che potential accidental release scenarios are known,
countermeasures can be proposed for incorporation into the emergency plan.
Since there may be more than one acceptable response to a given emergency
situation, the various response measures must be evaluated to select those
that appear most effective for the specific scenarios under consideration.
The final step is to implement the measure(s) when an emergency occurs.
Two other important subjects to consider when designing an emergency plan
are resource requirements and availability, and emergency organization.
Resource requirements include personnel, equipment, and money to fund both the
plan and its implementation. Personnel requirements and equipment needs
should be explicitly noted, while funds should be allocated in management
budgets but not explicitly addressed in the plan. An on-going financial
commitment will be required for periodic reviews and updates of the plan, as
well as for training programs. Availability addresses what resource additions
may be required that are not already present in a facility, company, or
community. The emergency organization must be clearly defined. The function
of the emergency organization is to establish clear lines of communication and
define responsibilities and authority so that confusion, delays, and
inappropriate actions can be avoided during an emergency.
Finally, a complete plan must address mobilization--how the plan will be
activated in an emergency--and demobilization--how the plan will be concluded
after an emergency.
The plan should be documented; the emergency response plan (ERF) docu-
ments the planning activities and provides a ready reference for use during
emergencies. The ERF must be a "living document," that is, frequently updated
to reflect changes in equipment, staff, and the facility. The ERF should
include the following major topics:
• General facility description, including plot plans and area maps;
9
-------
DescripCion of hazards associated with facilicy operations (detailed
assessments should be included or attached);
Description of the organization of personnel involved in emergency
response;
Functions, responsibility, and authority of key personnel;
Location and equipment of each emergency control center;
Alarm systems and notification;
Description of all communication systems and emergency backups;
Ranked list of notifications that must be made for each type of
accident;
Evacuation and/or sheltering plans for non-essential personnel;
Assembly points and procedures for accounting for all employees,
visitors, contractors, and others;
Locations of and equipment stored at each on-site first aid center;
Transportation facilities and likely locations of vehicles; includ-
ing specialty vehicles like trackmobiles, cranes, front-end loaders,
and others;
Security precautions;
Continuation of utilities and services;
Local emergency response agencies;
10
-------
Mutual aid organizations;
• Possibility of and procedures for community evacuation;
• Community shelter-in-place potential;
• Community emergency education;
• Public relations in an emergency;
• Procedures and equipment for responding to
fires and/or explosions,
toxic materials releases,
floods,
hurricanes (if applicable to location),
ice storms,
other natural phenomena (earthquakes, tornados, etc.),
civil disturbances;
e Restoration of normal operations;
e Training in emergency response procedures;
• Drilling in emergency response procedures;
• Emergency systems training and maintenance; and
• Emergency response plan updates.
The appropriate depth of planning- for each of these topics depends on the
nature/severity of the potential hazard, the size and complexity of the plan,
and the use of the facility relative to the use of outside agencies for fire
fighting, first aid, etc.
11
-------
One sample outline of a community emergency response plan published by
the U.S. EPA is shown in Table 2-1 (1). The details of this approach are
discussed in the reference. Comparison of this approach with the previous
discussion illustrates that, while certain topics are common to both
industrial and community plans, individual details may be addressed in
different ways.
2.2.2 Implementation
Implementation of an emergency response plan, including coordination
between the facility and the community, has been addressed by both the U.S.
EPA and the chemical industry (1,2,3).
One formal, collective chemical industry initiative is the Community
Awareness and Emergency Response program developed under the auspices of the
Chemical Manufacturers Association (3). Its purpose is Co prepare communities
to respond effectively to man-made or natural disasters. The goals of the
program are to improve community awareness and to integrate industrial
emergency response plans with those of the community. Some objectives of CAER
are to:
e Inform the public about hazardous chemicals;
• Review, renew, or establish emergency response plans;
e Integrate chemical facility emergency response plans with those of
the community to form an overall plan for handling all emergencies;
and
• Involve members of the local community in developing and
implementing overall emergency response planning.
12
-------
TABLE 2-1. EXAMPLE OUTLINE OF A COMMUNITY EMERGENCY RESPONSE PLAN
i. Emergency Response Notification Summary
ii. Record of Amendments
Hi. Letter of Promulgation
iv. Acknowledgment
v. Table of Contents
A. Abbreviations and Definitions
B. Purpose
C. Relationship to Other Plans
D. Assumptions/Planning Factors
E. Concept of Operations
1. Governing Principles
2. Organizational Roles
II. Emergency Response Operations
A. Notification of Release
B. Initiation of Action
C. Coordination of Decision-Making
D. Public Information/Community Relations
E. Personal Protection/Evacuation
F. Resource Management
G. Personnel Safety
H. Acutely Toxic Chemicals
I. Countermeasures
J. Response Action Checklist
K. Attachments
1. Emergency Assistance Telephone Roster
2. Siren Coverage
3. Emergency Broadcasting System Messages
4. Evacuation Routes
5. Traffic Control Points
6. Access Control Points
7. Evacuation Routes for Special Populations
III. Annendlces
A. Basic Support Documents
1. Legal Authority and Responsibility for Responding
2. Acutely Toxic Chemicals Information
3. Hazards Identification and Analysis
4. Response Organization Structure/Coordination
5. Laboratory, Consultant, and Other Technical Support
Resources
6. Computer Utilization
B. Post-Emergency Operations
1. Documentation of Accidental Releases
2. Investigative Follow-Up
13
-------
A CAER Handbook has been prepared that gives detailed information on
organizing and implementing a CAER Program (3). Major steps in a CAER program
are:
• Community status review and program organization;
• Facility status review;
• Implementation;
• Community involvement; and
• Emergency exercises.
Community Status Review--
This review involves a survey of the status of local emergency planning
in communities within a few miles around a facility. It seeks to determine
whether integrated community emergency response planning is a complex activity
(multifacility/hazard) or a local chemical facility function. Factors that
should be considered are:
• Are there significant natural disaster risks in the area and do
emergency plans exist for these risks?
• Are there other hazardous industrial activities, such as nuclear
power generation, hazardous waste disposal, or significant
transportation activities, and do emergency plans exist for these
hazards?
• Are there other chemical facilities in the area?
14
-------
• Does the facility's own emergency response plan clearly define Che
roles and responsibilities of community officials and off-sice
responders?
Based on chis information, a status reporc can be prepared ChaC describes
Che facility, the key elements of the facility emergency response capability,
the objecCives of CAER, and other pertinent issues. It should stress the need
for off-site emergency response planning to be integraced with che existing
facility emergency response capability, and that an integrated facility-
community plan can be used to improve emergency response to all hazards.
• As a facility initiative in getting the CAER program underway, this
document can be distributed and discussed with selected community
contacts. The CAER program is explained to other people who should
be involved in community emergency response planning as they are
identified. These people will probably include first responders
such as the fire chief, police chief, ambulance service workers,
mutual aid program participants, local officials, business leaders,
and facility employees, who may also be elected officials. Local
disaster preparedness officials may be helpful as initial contacts.
• These initial contacts will identify che appointed and elected
officials whose cooperation will be necessary to complete the
integrated community emergency planning process. Meetings with key
officials should be used to enlist their support in establishing a
coordinating group. In many locations, the equivalent of this group
or of a well-defined emergency response planning agency may already
be in place.
A more formal preliminary organizational meeting for the coordinating group is
a next step. Participants should include members of the groups listed above.
Other participants could be public health officials, hospital administrators,
Red Cross personnel, or local civic leaders. In most areas, chief elected
15
-------
officials have a general charter to protect public safety. Gaining official
recognition from the local community is an important prerequisite for the
coordinating group. This gives the group the necessary authority to begin the
integrated planning process described later in this discussion.
Facility Status Review--
A number of activities are required. A first step is to review the
facility emergency response plan to ensure that interactions with off-site
agencies and responders are clearly defined. Responsibilities and
notification procedures should be clearly defined.
• Employee awareness of emergency response planning must be provided
for by ensuring that basic emergency concepts are included in
initial training .and periodic safety training. The facility news-
letter or bulletin boards can announce that the facility has
completed a new or revised company plan and has started an inte-
grated emergency response planning process.
• A written community relations plan should be prepared. This can
begin with a list of local organizations, government agencies,
community groups (business, civic, public interest) and local media
who should be briefed on the CAER program. Past local media cover-
age of the facility and national industry issues should be reviewed.
Past and ongoing facility communications efforts should also be
reviewed and it should be determined whether the community and its
leaders are likely to be knowledgeable about the facility. Using
this information, assign priority to community awareness efforts.
Pay special attention to relations with local officials, off-site
responder groups, neighborhood associations and labor groups. It
may be appropriate to obtain expert assistance in community rela-
tions if this is a new activity for the facility.
16
-------
Assured Chat Che internal emergency planning activities are in order and that
the appropriate coordinating group has been established, the facility's first
contribution should be a presentation of a ten-step implementation process to
the coordinating group.
Implementation Steps--
Ten implementation steps are defined under the CAER program:
• Identify the emergency response participants and establish their
roles, resources and concerns;
• Evaluate the risks and hazards that may lead to emergency situations
in the community;
e Have participants review their own emergency plan for adequacy
relative to a coordinated response;
e Identify the required response tasks not covered by existing plans;
• Match these tasks to the resources available from the identified
participants;
• Make the changes necessary to improve existing plans, integrate them
into an overall community plan and achieve agreement;
• Commit the integrated community plan to writing and obtain approvals
from local governments;
• Educate participating groups about the integrated plan and ensure
that all emergency responders are trained;
e Establish procedures for periodic testing, review, and updating of
the plan; and
17
-------
• Educate Che general conmuniCy about Che inCegraCed plan.
2.3 TRAINING
The emergency response plan, which reflects the character of a specific
facility, determines the basic training needs, which will vary with different
groups within the facility according to their designated roles in an
emergency. The general objectives of training are to increase the awareness,
knowledge, and skills of management, and of operating, maintenance, and spe-
cialized emergency response personnel. Emergency exercise activities, sug-
gested by CAER, involve Integrated training with community emergency
personnel.
2.3.1 Training Overview
As defined by one author, training must address employee roles in three
stages of an emergency (4):
• Raising the alarm;
• Declaring that an emergency exists; and
• Implementing the emergency plan.
The training must ensure that personnel understand their roles in the general
areas of:
• Communications and control procedures;
• Individual responsibilities;
• Specific emergency operating and countermeasure procedures;
18
-------
• Coordination with outside services; and
• Public relations.
There are two distinct functional aspects of training: procedures and
equipment. Procedures include the safe shutdown of a facility,
communications, first aid and medical response, evacuation, traffic control,
rescue, fire fighting, and the implementation of other counter measures.
Equipment-oriented training deals with both protective and remedial
equipment, including: protective clothing and breathing apparatus, facility
shutdown and control, fire fighting, leak control, spill control, meteorologi-
cal monitoring, movement of equipment, emergency construction, and actuation
of hardware safety systems.
To he effective, a training program must be ongoing, periodically re-
viewed, updated and revised, and have full management support to ensure that
the necessary training resources are available. Management provides the
•nvironmenC for adequate training development and presentation through
guidelines and policies and also the enforcement and incentives to make the
program work. The initial training of personnel should give a thorough
grounding in fundamentals and be followed by subsequent enhanced training in
critical emergency areas. The development of an effective training program
depends on a thorough task analysis of the various personnel roles defined in
the emergency response plan.
The conduct of training should include on-the-job training,
seminar/discussion sessions, formal classroom training, and field exercises.
The support that vendors of equipment and technical services can provide
should not e overlooked. Such outside influences can inspire new interest in
otherwise routine in-house programs. Full advantage should also be taken of
newer training tools such as video tapes and computer simulation.
19
-------
In-depth training should be given to employees soon after they are hired
and should include a review of facility hazards, general emergency response
activities, the emergency response plan, and the specific duties of the new
employee. This level of training should be repeated each time an employee
moves to a new position in the facility with different hazards and different
emergency response duties.
Refresher training should be given regularly. The frequency of refresher
training should be based on the nature of the hazards and on the type of
emergency response duties assigned to the employee.
Drilling is an important part of training. While the objective of
training is to teach employees the appropriate response, the objective of
drilling is to improve response time. Drills should be held for each type of
hazard present at the facility. A combination of announced and surprise
drills should be used. Outside agencies should be included in some drills.
The employees should be given feedback on drill performance and supplemental
training when performance is seriously deficient.
Decisions about the frequency of drills should be based on the
nature/severity of the hazards, the size/complexity of the facility, and the
amount of reliance on outside emergency aid.
Drills are discussed further as part of emergency exercise programs in
Subsection 2.3.3.
2-3-2 Tralntm dwrdlmti™ »
-------
involved. Representatives of these agencies should regularly tour the facil-
ity and be trained in its hazard-related features. A spirit of open
communications and cooperation should be fostered.
2.3.3 Emergency Exercise Program
An emergency exercise program is an essential part of CAER. It completes
the cycle begun when the facility, along with community and industry represen-
tatives, formed a community coordinating group to manage emergencies. The
coordinating group integrated all emergency plans into one community emergency
response plan. Emergency responders have their tasks, and resources are
committed. An emergency response system is in place, but will it work?
Besides a real emergency, only the simulation of an emergency can answer
that question. Simulation and evaluation exercises, by testing all or part of
a system against the plan, complete the cycle the facility and the community
started. There are several types of exercises and each has its use, but no
single exercise can adequately test all of the elements of an emergency
management system. Properly implemented, an ongoing exercise program will:
• Help evaluate emergency plans and response capability;
e Provide the information necessary to improve plans and procedures;
• Train the participants;
• Improve coordination between on-site and off-site personnel;
e Ensure the continued involvement of key community organizations;
21
-------
• Provide a way to inform and involve the public and the media; and
• Serve as a visible demonstration of industry and community commit-
ment to protect the public.
Facility Management Role--
Facility management has a pivotal role. Exercises are complex events
that require the cooperation of many people and organizations. Without the
active support and encouragement of the chief executives of these organiza-
tions, there is little chance for success. The most important parts of the
task are committing the necessary time and resources of the organization and
gaining the cooperation of the chief officials in the community.
Components of the Program*-
Four types of exercises should be the major components of the program:
• Tabletop;
• Emergency operations simulation;
• Drill; and
• Field exercise.
They range from the simple to the complex and each has a definite function.
Most industrial managers are familiar with the concept of a drill or
field exercise. Such exercises typically involve actual response by personnel
and the use of protective equipment, emergency apparatus, and field
communications. A drill focuses on a single aspect of an emergency response
system, while a field exercise tests all or most of the response system.
22
-------
The tabletop and emergency operation simulation exercises are less
commonly used and are characterized by the simulation of all or most of the
actions to be performed in emergency response. They provide substantial
benefits within an overall exercise program because they are usually less
complex and require less- time and effort to plan and conduct. They are also
more flexible as learning experiences and can build confidence and capability
for more involved exercises.
• A tabletop is primarily a learning exercise that takes place in a
classroom or meeting room setting. Situations and problems generate
discussion of plans, procedures, policies, and resources. Tabletop
exercises are an excellent way to familiarize a group of
organizations and agencies with their assigned roles. This type of
exercise is also a good method for testing the logic and content of
the plan. It is sometimes referred to as a "What if" exercise.
• An emergency operation simulation (EOS) is a full-scale emergency
simulation that uses various forms of message traffic (telephones,
radios, message forms, etc.) to create a realistic, high-pressure,
emergency environment. EOS exercises stimulate decision-making and
interaction by emergency managers in response to simulated emergency
conditions. An EOS is conducted In an Emergency Operating Center or
In another suitable facility and involves command, control, and
decision making functions. All field response activitj.es are
simulated, with activity limited to the controlled environment of
the facility used. They are very useful for testing direction and
control functions and for evaluating how well the total response
system is coordinated.
23
-------
• The drill is a supervised activity that tests, develops or maintains
skills in a single emergency response function (i.e., communications
drills, fire drills, command post drills, medical emergency drills,
etc.). These frequently involve actual field response, activation
of emergency communications networks, and equipment and apparatus
that would be used in a real emergency. The effectiveness of a
drill is its focus on a single, or relatively limited, portion of
the overall response system in order to evaluate and improve it.
• A field exercise practices all or most of the basic functions of the
response system simultaneously and the ability of the different
organizations to work together. A full field exercise involves at
least a partial mobilization of each organization's resources and a
demonstration of their response actions. l£ also includes
activation of all emergency facilities, including emergency
operations centers, communications centers, command posts,
hospitals, media centers, first response units, etc. A field
exercise may exclude some specific functions previously tested, but
it should include all response functions that require significant
coordination with any others being tested.
Table 2-2 summarizes some of the advantages and disadvantages of each type of
exercise.
24
-------
TABLE 2-2. EXERCISE TYPES--ADVANTAGES AND DISADVANTAGES
Advantages
Disadvantages
1. Tabletoo Exercise
Modest time, cost, and resource
commitment. Effective method of
reviewing plans, procedures, and
policies. Good training for key
personnel in responsibilities
and procedures, helps build
coordination and consensus.
2. Emergency Prorations Simulation (EOS')
Moderate time, cost, and
resource commitment. More
realism than a tabletop
exercise. Tests integrated
response of entire emergency
management system. Good method
of testing command and control
before "going public."
Lacks realism. Does not provide
a true test of emergency system
operation.
Heavily dependent on written
scenario. Realism/effectiveness
can suffer when not well planned
and executed.
3. Drill
Moderate time, cost, and
resource commitment. Easiest
exercise to design. May be very
realistic. Provides good
hands-on training. Allows a
single component of the system
to be isolated and practiced in
depth.
4. Field Exarclae
Good realism. Good test of
integrated communications.
Provides means to evaluate
mobilization of resources and
first responder capability.
Good opportunity to increase
public awareness of program.
Does not test entire system.
Player safety is critical.
Major time, cost, and resource
commitment. Scenario development
is very critical. Player/public
safety is critical. Can create
problems when poorly planned.
25
-------
2.4 REFERENCES
1. U.S. EPA Chemical Emergency Preparedness Program, Interim Guidance,
Revision 1, 9223.0-1A, November 1985.
2. Cathcart, Christoper. U.S. Chemical Industry Emergency Response Initia-
tives, Avoiding and Managing Environmental Damage from Major Industrial
Accidents. In: Proceedings from Air Pollution Control Association
International Conference, Vancouver, British Columbia, Canada, November
1985.
3. Chemical Manufacturers Association, CAER Handbook, Washington, D.C.,
1985.
4. Lees, F.P. Loss Prevention in the Process Industries, Vol. 2,
Butterworth's, London, England, 1980.
26
-------
SECTION 3
FACILITY SITING AND LAYOUT
The consequences of an accidental release can be reduced by incorporating
mitigative features into the design of a facility's layout and siting. Siting
is the location of the process facility within a community or region, while
layout is the positioning of equipment within the process facility. Although
it is important to consider siting and layout when designing a new facility,
the guidelines discussed in Subsection 3.2 apply equally to facility
expansions. Siting could also be an important factor when deciding whether to
expand a particular facility.
3.1 BACKGROUND
Siting and layout design principles that can reduce the effects of an
accidental release include the following:
e Locating the facility within the community or region in a way that
reduces the number of people that could be affected; and/or
e Arranging equipment within the facility in a way that reduces the
effect that a potential release from one location within the
facility would have on the rest of the facility, or on the
surrounding community.
One study of the contribution of different hazard factors to accidental
releases found that poor facility siting played a role in 5.8X of the cases
and that poor facility layout was a factor in 3.9X of the cases (1). The
study did not distinguish between factors that could have prevented the
27
-------
incident and those that could have mitigated its effects. Although this
section focuses on the mitigation of accidental releases, the information
presented in Subsections 3.2 and 3.3 may also help prevent the occurrence of
an accidental release. Preventive measures for facility layout and siting are
discussed in the Control Technologies Manual of this Prevention Reference
Manual series (2).
3.2 SITING
Siting refers to the location of the process facility within a community
or region. Often, siting is carefully examined only for new facilities;
however, the expansion of an existing facility may require a reevaluation of
the site to assess the suitability of the expansion. This will be especially
true if the expansion involves chemicals or processes that pose more hazard
than presented by the original facility. One author discusses siting from a
process hazard perspective and cites a number of references (3).
A study of factors that have contributed to losses in several hundred
large fires and explosions identified the following inadequacies in facility
siting: poor utility service; poor emergency response and fire protection;
off-site traffic congestion hindering response by emergency vehicles; and poor
drainage (1). The effect of these inadequacies can be magnified by meteor-
ologic factors and population densities around the facility, two factors that
will determine the effects of an accidental release on off-site populations
(3,4).
Literature on chemical facility siting has generally been limited to
discussions of traditional siting criteria, i.e., distance from raw material
resources and markets and suitability of the transportation system and
utilities (5,6,7). However, the siting of facilities such as nuclear power
plants and hazardous waste disposal units has been more problematic historic-
ally, and specific siting criteria have been developed to address public and
28
-------
environmental health issues in these industries. Some of the issues
identified in Che relaCed literature could be relevant Co Che siting of a
chemical facilicy chac uses or produces toxic chemicals.
3.2.1 Population lWv«itv Near Facility Site
Clearly, the fewer people that live near an industrial facility, the
fewer chac would be affecced by an accidental release. One way to minimize
Che number of people affecced by a pocencial accidental release is Co maincain
a buffer zone immediately around or downwind of the facilicy. The buffer zone
for a nuclear power facilicy, for insCance, is a circle wich a 0.3 Co 1.3 mile
radius (6). A discance of five miles downwind was required during Che sicing
of a spill CesC facilicy for liquefied fuels, ammonia, and chlorine (4).
Seccing aside a controlled buffer zone, however, has noc been sCandard
praccice in industrial facilicy sicing. Although induscrial facilities may be
siced in what is initially a rural area, the facility iCself may become che
focus of later development; therefore, setting aside a buffer area is one way
for a hazardous industrial facility to protect itself against the problems of
future urbanization.
Another way Co reduce che population density around a facility is to
maximize che discance Co population centers. This issue has not been expli-
citly addressed in the siting of hazardous industrial facilities. The site of
che Bhopal Union Carbide facility, where a MIC release resulted in .over 2,000
deaths, was found In 1975 to be "not suitable for hazardous industries" be-
cause of the population density around the facility (8). Two hundred thousand
people lived within a 4-mile radius (9). The facility did, however, receive
approval to operate despite this evaluation. Contrast this with siting
requirements for commercial nuclear power reactors or hazardous waste disposal
facilities. The New Jersey hazardous waste facility siting act requires a
minimum distance of 2,000 feet between the facility and the closest residence
29
-------
or school. For some types of hazardous waste disposal facilities, a half mile
distance from residences or schools is specified (10).
The regulations governing commercial nuclear reactors take a different
approach to specifying distances from population centers. The regulations
specify two zones, calculations of which are based on maximum exposure from a
major accident and a population center distance. The population center
distance is the distance to the nearest densely populated center with more
than 25,000 residents. The population center distance must be at least one
and a third times the distance from the reactor to the outer boundary of the
second zone (11). Thus, the distance to the population center will be a
function of the design of the facility and the definition of a major accident.
Increased distance from population centers also allows more time to
respond to an emergency and evacuate or inform the local population. This
relationship is addressed indirectly by the nuclear plant siting regulations
discussed above. However, the literature does not examine the relationship
between time available for emergency response and facility distance from
population centers. Clearly, the distance could not be fixed arbitrarily, but
would be a function of factors such as emergency preparedness and population
mobility (i.e., access to public transportation, cars per household, road
network).
An example of a risk-based approach to facility siting can be found in
the Canvey study, an investigation of a heavily industrialized area in England
(12). The study examined existing and proposed facilities in the Canvey
Island/Thurrock area and calculated the individual and total risk posed by the
petrochemical and shipping facilities to area residents. The Commission
recommended that the proposed facilities be approved on the condition that
risk reduction measures be undertaken by the existing facilities. The Canvey
study illustrates how quantitative risk analysis can be used during the
facility siting process. This type of approach would be especially suited to
facility expansion decisions.
30
-------
The licecature on Che siting of hazardous industrial facilities does not
provide guidelines as to what constitutes a safe distance or an effective
buffer zone. Examples of buffer zones considered during the facility siting
process are given in the literature about other fields, but these examples may
not directly apply to chemical processing facilities.
3.2.2 MftgeorolQfY Climate
Meteorologic considerations involve identifying the prevailing wind
direction, wind speed, and atmospheric stability. Meteorologic data can be
used to rule out inappropriate facility sites or to locate units within the
facility boundaries. Hazardous units should be located to minimize the
effects of a release on off-site and on-site populations and should be
downwind (for the prevailing wind direction) from major population
concentrations when possible.
Wind patterns can be characterized in three ways. One is the large-scale
movement of major air pressure systems. An intermediate scale involves air
movements caused by regional topograhic features and differential heating. A
more local effect can be caused by the presence of ridges and valleys. Data
on all three types of patterns may be used to predict areas affected by
accidental releases (4). However, of most interest will be site-specific wind
characteristics. For example, ridges and valleys may channel surface winds
that would carry heavier-than-air releases. Ridges serve as barriers in this
instance; whereas valleys are possible conduits for the accidental release.
The reader is referred to Section 5 on dispersion modeling and Section 6 on
meteorological instrumentation for a discussion of the use of meteorological
data and the modeling of release impacts.
Temporal variability in wind direction may occur dally or seasonally.
Temporal variability may be an important characteristic when there is a basis
31
-------
for determining Che most likely time for a release; for example, if the
chemical is only stored or processed on site for a specific period.
Otherwise, it could be assumed that the most frequently occurring wind
direction is the one of greatest interest in modeling impacts. However, it
will often be necessary to investigate the effects of other than prevailing
wind conditions.
Examples of accidental release incidents in which the meteorological con-
ditions aggravated the release's impact can be found in the literature. At
Bhopal, for instance, the area was experiencing an atmospheric inversion at
the time of the release. An inversion increases the time needed to disperse a
chemical cloud to safe levels. A region that experiences inversions
frequently may not be an appropriate site for a hazardous industrial facility.
Other incidents in the literature illustrate how unstable or neutral
atmospheric conditions are beneficial, since the wind movement disperses
dangerous concentrations quickly; see case history A-24, Lees (3). Another
example of basing the selection of a site on good dispersion characteristics
(reliable wind direction and wind speed) can be found in. Miller, (4).
3.2.3 Emergency Response
A rapid response by facility or local emergency response personnel can
mitigate the effects of a release. If a facility is to be located in an area
that does not have an adequate emergency response system or if the addition of
the facility will strain the area's emergency response capabilities, the
facility owners might consider giving support to local emergency response
services. Mew Jersey now requires this of major hazardous waste facilities.
The New Jersey hazardous waste facilities siting act specifies that five
percent of a facility's gross receipts will be paid to the host municipality
for the cost of additional emergency response training and equipment (10).
Several chemical companies also report supporting the emergency response
capabilities of host communities (6).
32
-------
3.2.4 Topography and Drainage
A fairly level site is necessary to prevent flammable liquids from
flowing down a sloped terrain to other equipment or buildings. Also, the
facility site should not drain highway runoff, since flammable liquids from a
highway accident may spread a fire. Hills may be used as natural barricades,
separating operations with potentially explosive chemicals (13). The effect
of the terrain on wind patterns should be considered also, sihce local
features such as valleys and ridges will channel winds. Hills or ridges can
also block, divert, and help disperse an air release of a toxic chemical.
3.2.5 Acceaaibllji-" Off-Site Emergency Vehicles
If the facility is to be served by off-site emergency response personnel,
it must be easily accessible. Facility entrances should be free of traffic
congestion or blockage by trains (moving, stopped, or derailed); the facility
should have multiple entrances/exits, ideally onto different streets or roads
so that one accident cannot block all escape routes (13). It should also be
considered whether the facility 'site is accessible year round. In areas with
snow and ice, reliance on measures in addition to or instead of outside
emergency response services might be warranted.
3.2.6 nttltev Service
Water, gas, power, and any other utilities should be reliable. An
adequate supply of water for fire fighting is essential. Failure of several
utilities services simultaneously can aggravate a release incident (13).
Water and electricity are the most important utilities, since they are
essential to fire fighting, cooling systems, operation of pumps, lighting, and
instruments, etc. Back-up electrical power and looped water mains are typical
safeguards. A looped water main originates at a source, is routed through the
33
-------
facility and then returns to the source. This arrangement allows water to
flow from the source in either direction through the loop. Thus, a break in
the line at any point will not cut off the rest of the loop from the water
source.
3.3 LAYOUT
Layout concerns the placement and spacing of the components of a process
facility, including the individual equipment in each component. A properly
designed facility will minimize the consequences of an accidental release by
facilitating process operability and by segregating hazardous areas. Both the
Chemical Manufacturers Association (CMA) and the National Fire Protection
Association (NFPA) have issued standards and guidelines for facility layout
(14). Key features to be considered during the layout of a facility site are
its boundaries, the work boundaries, through railway lines, ignition sources,
control rooms, buildings that concentrate personnel, production units,
storage, loading and unloading, pipebridges, roadways and waste disposal
areas.
A study of factors that have contributed to losses in several hundred
large fires and explosions identified hazard factors and grouped these into
nine categories. Problems with facility layout and spacing were characterized
in the following way: congested process and storage areas; lack of isolation
of extra-hazardous operations, lack of proper emergency exit facilities;
sources of ignition too close to hazardous critical facility areas; and
inadequate hazard classification of facility areas (1). These problems can be
addrtssed by applying the general principles discussed below.
34
-------
3.3.1 Adequate Snacing Between Units
It is important to increase the distance between process units to reduce
the impact of an accidental release. An explosion in one process unit can
result in an accidental release in another unit because of damage from the
blast wave. The blast wave pressure front travels rapidly, but its intensity
decreases rapidly with distance. A vapor cloud associated with an accidental
release travels much more slowly than does a blast wave but usually affects a
larger area before it is diluted enough to present no danger. Explosions of
vapor clouds can be particularly destructive in congested plants.
A quantitative method may be used to specify the spacing between facility
units. First, a relative hazard ranking is developed of the individual
process units, including manufacturing units, unloading/loading operations for
tankers or drums, major pipebridge sections and major cross-country pipelines.
One common method of quantifying the potential fire and explosion hazard of a
given process unit is the Dow Index (15). The Dow Index has been modified to
address the potential hazards of toxic chemical releases. This Mond Index is
one of several methods developed for quantifying spacing requirements between
facility units (15,16). Other methods for quantifying spacing requirements
have also been developed (15,16).
Safe spacing requirements can be developed for a particular site, as des-
cribed above, or reference may be made to published sources, a number of which
recommend minimum safe separation distances for process and storage units.
Reference 3 lists several other sources that discuss safe separation
distances. Reference 3 notes that most of the published guidelines for
spacing requirements give little explanation of the basis for the distances
and that many sources incorporate recommendations from each other.
35
-------
The cvo principal faccors that have been traditionally used for determin-
ing safe separation distances are: 1) the heat produced by the burning liquid
and 2) the potential to ignite the accidentally released gas or vapor.
Separation of units from identified ignition sources, such as furnaces,
electrical switchgear, flarestacks or transport vehicles is one of the most
common recommendations encountered in the literature. A review of case
studies in Reference 3 shows that the presence of an ignition source near an
accidental release source can turn an initial release incident into a fire
chat can cause a greater loss of life and property. For instance, in the
Feyzin refinery case cited in Reference 3, a source 160 meters distant from
Che plant unit is suspected of igniting a cloud of propane. Mote that the
toxic hazard associated with an accidental release may be more severe than the
hazards associated with ics heat of combustion or flantmability. In such
cases, traditional methods of determining safe separation distances may not be
appropriate.
Some special considerations adapted from the literature for the safe
spacing of facility units are:
e In general, units with the most toxic materials should be suitably
distant from all facility buildings containing appreciable
concentrations of personnel, and also from activities outside the
works boundary. This particularly applies to off-site population
centers such as schools, hospitals, places of entertainment, etc.
• Units having the greatest potential for explosion should not be
located close to a facility boundary; they should be separated by
areas occupied by low-risk activities with low population densities
(up to 25 people per acre).
36
-------
• Independent elements should be spaced so that a fire or explosion in
one unit has a minimum effect on other units. When units cannot be
adequately spaced and segregated, the use of barricades, protective
construction, and fixed fire protection is recommended. For further
information see Rindner and Wachtell (18) and Lewis (17).
• Ignition sources, including trucks, railroads, boilers, personnel
smoking areas and direct-fired equipment should be located away from
and upwind of hazard areas processing or storing flammable liquids,
gases, or vapors. Some standards call for at least 10 feet of
separation between vapor hazard areas and ignition sources, although
this distance may be reduced by using vapor barriers or other
special protective systems (13). Yet in the Feysin refinery
example, the ignition source was thought to be 530 feet away, which
illustrates the importance of evaluating each situation individually
and identifying high-hazard areas.
• Facility work units should not be located at the facility boundary.
Distance to the facility boundary will vary, depending on the hazard
of the unit (17).
Finally, the initial design must allow sufficient space for future
facility expansions, and the expansions must also adhere to safe layout
principles.
There are two schools of thought on the safest spacing of interconnected
process units. One, which emphasizes the danger of individual process units
and specifies minimum safe distances between types of units, focuses on the
dangers of fire and explosion. Another school of thought emphasizes the
danger of releases from long pipe runs and considers specific maximum spacing
between the units to minimize such long runs. This approach applies more to
non-flammable, acutely-toxic materials. In many cases, both fire and toxic
release hazards are present and spacing design will be a compromise between
37-
-------
them. Both hazards can be addressed by planning for enough space (to satisfy
fire/explosion concerns) and by breaking up the resulting long pipe runs with
several emergency excess flow valves (to reduce the size of a release if the
pipe should rupture at any point).
3.3.2 Grouping or Isolation of Facility Units
How units are arranged within the facility is important, since this
partly determines whether the initial accidental release remains localized or
spreads throughout the facility processes. General guidelines for locating
units within a facility are noted below (13,17);
• Whenever possible, high hazard-units should be separated from each
other by units of mild, low, or medium hazard.
• Control rooms, amenity buildings, workshops, laboratories and offi-
ces should be adjacent to units of mild or low hazard, which act as
a barrier from higher-hazard units. Medium-hazard units are only
acceptable adjacent to populated buildings (a) if lower-hazard units
are not available to separate them, and (b) if the hazard level is
only just inside the "medium" band of overall hazard rating, as
assigned by the Mond Index.
• Storage units, such as tanks and container storage areas,, should be
adequately separated from operational areas and, as far as possible,
located away from road and rail traffic routes within the plant
site.
• Major pipebridges with a medium to high overall hazard rating
assigned by the Mond Index method should be protected from accidents
Chat could happen on tall process units and from vehicle collisions.
Incidents on tall process units can spread by the collapse of the
tall unit onto adjacent pipebridges.
38
-------
3.3.3 Transportation
Planning an adequate transportation system is an integral part of plane
layout. An adequace transportation system mitigates Che effects of an acci-
dental release in two ways: 1) emergency responsiveness depends on ready
access to the affected units to reduce the spread of the emergency in the
plant, and 2) a timely evacuation of potentially affected populations reduces
the number of people affected. Design principles of an adequate
transportation system are described below (13,17).
• Roads should allow entry to the facility from at least two points on
the site perimeter, preferably from opposite sides and onto
different roads/streets.
• All units in the facility with a moderate or high fire risk should
be accessible to for emergency vehicles from at least two
directions.
• Control rooms should have access to an easy exit, since control room
personnel may be the last plant personnel to leave the facility
site.
• Facilities with access from heavily traveled super-highways should
also have alternative emergency entrances.
3.3.4 General Layout
A few other recommendations appear below (13,17):
e Control rooms, amenity buildings, workshops, laboratories and
offices should be close to the site perimeter and away from
hazardous storage and process areas.
39
-------
Pipebridges should be laid out to minimize the transfer of accidents
from one unit to another. Key pipebridges should be assessed
individually for hazard potential, independently of adjacent units.
As far as possible, units should be laid out for a logical process
flow to minimize the pipebridge requirements.
Fipebridge routes should be chosen so that they are not likely to
contribute to the spread of an accident. Alternative routes should
be selected to retain as much process control as possible to be
retained in the event of an incident occurring.
A rectangular block layout is often used for chemical and petro-
chemical plants so that emergency response vehicles can gain access
to all units. Roads can also act as a fire break.
Vater systems should be adequate in all parts of the facility. For
a larger facility, the water system may consist of a number of
loops, rather than one large loop.
Sewers and drainage should be laid out to allow rapid and safe
removal of spilled chemicals and water used in fire fighting.
Critically important facility elements (power plants, computer and
control rooms, special process units) should be given """flrmim
protection, which might include extra space dividing them from other
units, proper location, or barricades.
Hazardous units in the facility should be accessible to fire
stations and other emergency facilities.
AO
-------
• Waste disposal and hazardous units processing, scoring, or disposing
of toxic gases and liquids should be located downwind of concentra-
tions of people on site and off site.
For a detailed discussion of facility layout, refer to Lees (3) and
Mecklenburgh (19).
3.4 COSTS
The cost of incorporating mitigation measures into facility siting and
layout cannot be estimated quantitatively since the costs will be specific to
the site. A general discussion of the types of costs to be encountered is
presented here.
Mitigation measures such as increased spacing between units increase
facility construction costs in two ways. Most directly, the facility might
require additional land to create adequate spacing between all units. Also
considered a mitigation coat is the additional expense of longer piping and
utility runs within the facility. Additional spacing between units will
result in increased pipe length and longer utility lines.
Directly measurable are the costs of including additional land as a
buffer for the facility, which is a function of land prices and of the size of
the area designated as a buffer.
It is difficult to estimate the c.ost of designing mitigation measures
into the facility siting without considering a particular site and facility.
Sites rated high according to mitigation criteria (i.e, not upwind of a
populated area, for example) would not necessarily cost more to develop than
sices selected according to traditional criteria. However, the initial site
cost is only a portion of the facility costs. If transportation costs
increase because of a greater distance to market, there is a direct cost to
the firm for locating the facility in the site selected for mitigation
41
-------
purposes. To evaluate the overall effects mitigation will have on costs, an
analysis of the facility's construction and operating costs would be needed.
Comparing the total costs at two sites, one selected according to mitigation
criteria, the other according to traditional siting criteria, is a way to
determine how the incorporation of specific mitigation measures affects
overall costs.
42
-------
3.5 REFERENCES
1. American Insurance Association: "Hazard Factors" in Hazard Survey o£ the
Chemical and Allied Industries. Engineering and Safety Service.
Technical Survey No. 3, 1979.
2. Davis, D.S., G.B. DeWolf, and J.D. Quass. Prevention Reference Manual:
Control Technologies. Vol. 1. Prevention and Protection Technologies for
Controlling Accidental Releases of Air Toxics. EPA-600/8-87-039a (NTIS
PB87-229656), U.S. Environmental Protection Agency, Research Triangle
Park, NC, August 1987.
3. Lees, Frank P. Loss Prevention in the Process Industries. Volumes 1 and
2, Butterworth's, London, England, 1983.
4. Miller, C.F., F.A. Leone, and W. Bryan. Liquified Gaseous Fuels Spill
Test Facility: Site Evaluation Study. Lawrence Livermore National
Laboratory, February 1985.
5. Urang, Sally. How Experts Pick a Chemical Plant Site. Chemical
Business, July 26, 1982. pp. 17-22.
6. McDavid, Catherine. "Prevention of Environmental and Public Health
Damage from Major Industrial Accidents through Innovative Siting
Approaches" in Avoiding and Managing Environmental Damage from Major
Industrial Accidents. Air Pollution Control Association Conference,
Vancouver, Canada, November 1985. pp. 237-251.
7. RPC, Inc. Siting Industrial Facilities on the Texas Coast. Texas
Coastal Management Program, Texas General Land Office, September 15,
1978.
8. Shrivastava, Paul. The Accident at Union Carbide Plant in Bhopal: A
Case Study in Avoiding and Managing Environmental Damage from Major
Industrial Accidents. Air Pollution Control Association Conference,
Vancouver, Canada, November 1985. pp. 90-95, 68.
9. Pesticide Plant Leak Wreaks Disaster in India. Nature, Vol. 312 (5995):
581, December 13, 1984.
10. New Jersey Statutes Annotated. 13:lE-49.
11. Nuclear Regulatory Commission. Reactor Site Criteria, 10 CFR 100, pp.
860.
12. Health and Safety Executive. Canvey: An Investigation of Potential
Hazards from Operations in the Canvey Island/Thurrock Area. London,
England, May 1978.
43
-------
13. American Insurance Association. Hazard Survey of Che Chemical and Allied
Industries. Engineering and Safety Service, Technical Survey No. 3,
1979.
14. National Fire Protection Association. Flammable and Combustible Liquids
Code, NFPA 30. Quincy, MA, 1984.
15. Battelle Columbus Division. Guidelines for Hazard Evaluation Procedures.
American Institute of Chemical Engineers, The Center for Chemical Plant
Safety, New York, NY, 1985.
16. Kletz, T.A., Plant Layout and Location: Methods for Taking Hazardous
Occurrences into Account. Loss Prevention, Volume 13, American Institute
of Chemical Engineers, 1980.
17. Lewis, D.J., The Mond Fire, Explosion and Toxicity Index Applied to Plant
Layout and Spacing. Loss Prevention, Volume 13, American Institute of
Chemical Engineers, 1980.
18. Ridner, R.M. and S. Wachtell. Establishment of Design Criteria for Safe
Processing of Hazardous Materials. Loss Prevention, Volume 7, Chemical
Engineering Progress, New York, NY, 1973.
19. Mecklenburgh, J.C. Plant Layout. Leonard Hill, London, England, 1973.
44
-------
SECTION 4
DETECTION AND WARNING SYSTEMS
Detection and warning systems are widely used in the chemical process
industries to alert plant operators and personnel of process upsets and
potentially hazardous situations that could lead to loss of production,
property, or in the extreme case, human life. Common to most chemical
processing facilities are detection and warning systems built into the process
control systems. Such systems monitor process operating conditions such as
the temperature, pressure, and flow rate, and trigger audible and visual
alarms when these process variables exceed design limits. Other detection
systems identify hazards after a release from a process or storage tank has
occurred. These systems may detect leakage of the chemical itself, flame,
heat, smoke, or gaseous products of combustion (1,2). Plant personnel also
play a role in detecting and warning of hazardous releases. However, in the
absence of a visible plume or fire, the reliability of plant personnel depends
greatly on the particular material being released, ambient conditions, and the
individual's physical senses. Reliance on plant personnel for detection and
warning can result in dangerous exposure of plant personnel to hazardous
materials. Visual observations, however, either directly by personnel in the
vicinity of the release or by personnel monitoring closed-circuit television
systems, can provide information on the direction and speed of a hazardous
release (3).
This section discusses systems used to detect and quantify concentrations
of leaks or releases of chemicals to the atmosphere after the release occurs
but before an actual fire and/or explosion. Because of the large number of
sampling and analytical methods available for collecting and analyzing
hazardous gases and vapors, the discussion is limited to direct reading
instruments or indicators, which, along with their sensing elements, can yield
45
-------
test results in the short time needed to quickly respond to a hazardous
release incident. Post release detection systems are important because the
more quickly an airborne release of a hazardous material is detected, the
greater the opportunity to minimize the effects on the community by
implementing on-site and off-sice emergency response procedures.
4.1 BACKGROUND
The purpose of a gas or vapor detection system is to provide a quantifi-
able indication of the concentration and location of a released chemical that
could lead to fire, explosion, or exposure of facility personnel or the
general public to hazardous concentrations of an airborne chemical. With
adequate warning from an automatic or manually operated warning system,
facility personnel and off-site emergency response personnel can take
appropriate action to reduce the effects of the chemical release. Specific
mitigation procedures taken in response to a leak or release are discussed in
other sections of the manual.
The general types of gas detection systems include (1,4,5):
• Multiple devices for sampling the air at various points
throughout the facility, using vacuum pumps and tubing which
then convey the sample to a common analyzer;
• Portable sampling/analyzing/monitoring devices that are
battery or manually operated and carried by plant personnel to
determine the concentration of a gas at a particular point of
interest in the facility;
• Devices where the sensing element and readout are contained in a
single unit mounted in the area of the hazard;
46
-------
• Multiple sensing elements wired to a single control panel located at
a point remote from the area being monitored; and
• Remote optical systems.
These systems often include an audible or visual alarm that activates
when it detects a release (2). Audible alarms include bells, whistles,
sirens, horns, or easily understood voice messages. In high-noise areas,
visual alarms in the form of lights are often used.
Once an alarm signals that a fire or release has occurred, the emergency
must be analyzed, and facility field personnel and/or the process operator in
the control room must take appropriate action to manually control the release.
In some cases, the detection system may activate mitigation devices such as
steam curtains or automatically released firefighting agents (1). A fully
automated computer system may be able to detect the release, determine whether
the release constitutes a hazard, display the location of the release, sound
warning alarms, provide instruction concerning mitigation of the release to
the operator, or automatically engage control devices, and telephone emergency
messages to the surrounding community public service agencies (1,2,3).
However, the danger and the damage that could occur if emergency response
procedures were falsely activated means that an automated system must be
highly reliable.
4.2 DESCRIPTION
Detection instruments use a variety of operating principles, including
(6):
e Electrical conductivity, potentiometry, calorimetry, and ionization;
e Radioactivity;
47
-------
Thermal conductivity and heat of combustion;
• Electromagnetic methods such as infrared photometry, ultraviolec
photometry, and other photometric methods;
• Chemi-electromagnetic such as colorimetry and chemiluminescent
methods;
• Magnetic methods such as paramagnetic analysis and mass
spectroscopy; and
• Gas chromatography.
In cases where releases of toxic chemicals occur at remote locations,
such as during transportation accidents, direct reading colorimetric
indicators are sometimes used for qualitative and quantitative measurements of
the released chemical. Methods include liquid reagents, chemically treated
papers, and glass indicating tubes containing solid chemicals that change
color and/or exhibit a length of stain that can be compared with a standard to
determine the concentration of the toxic gas (7). These types of manually
operated indicators are used in addition to other direct-reading instruments
based on the principles previously cited (8).
A detailed description of the numerous detection principles and instru-
ments available for detecting flammable and/or toxic gases is included in
References 6, 7, and 8. The following subsections are summary discussions of
the types of instruments typically used to detect flammable and hazardous
gases associated with an accidental release.
4.2.1 Combustible/Flammable Gas Detectors
Combustible/flammable gas detectors are commonly used to measure the
lower explosive limit (LEL) or lower flammable limit (LFL) of flammable gases
in air. (According to Reference 1 the terms are used interchangeably.) The
48
-------
below which the mixture cannot be ignited to yield a self-sustaining flame
(9). Examples of detectors are catalytic combustion sensors, solid-state
electrolytic sensors, and infrared analyzers (10). Table 4-1 presents the
characteristics of these and other combustible/flammable gas detection
systems.
Catalytic combustion sensors use a sensing element such as an elec-
trically heated platinum wire across which a sample of gas is drawn. The
flammable gas burns on the wire, which produces heat in direct proportion to
the concentration of the flammable gas in the air. The temperature of the
wire changes the electrical resistance. Electronic circuitry in the
instrument creates a readout indication as a percent of the LEL or percent
concentration. If the LEL is exceeded, alarm circuits and warning signals may
be activated to alert personnel of dangerous concentrations of gas in the
vicinity.
In a solid-state electrolytic sensor, the gas diffuses into the sensing
element (a semiconductor), resulting in a decrease in the electrical resis-
tance of the cell (10). The current flow resulting from this diffusion is
related to the concentration of the gas in the air, which is shown on a meter
as percent LEL.
Infrared analyzers draw air samples into a cell where an infrared light
shines perpendicular to the flow of the liquid. A detector on the side of the
cell opposite the light source measures the intensity of the light. The gas
of interest absorbs infrared radiation at a certain wavelength ranging from 2
to 15 micrometers (10,11). Thus, the concentration of the gas in air will be
proportional to the amount of infrared light absorbed. Two types of infrared
analyzers are available: the nondispersive type that operates at a specific
wavelength, and the dispersive type that can operate at several different
wavelengths. Infrared analyzers may draw air samples from several locations
or they may be set up to monitor only one location.
49
-------
TABLE 4-1. COMPARISON OF CHARACTERISTICS OF VARIOUS COMBUSTIBLE/FLAMMABLE VAPOR
DETECTION TECHNIQUES
TECHNIQUE
Characteriatic
Gaa
Chroaat ography
Hydrogen flaae
IonUation
Detector
Thermal
Conductivity
Interferoaetry
Infrared
Sea iconduc tor*
Catalytic
Coabuation
Senaitivity
Excellent
Beat
Poor to Good
Good
Vary good
Excellent
Very good
••liability
Good
Good
Good
Good
Good
Poor
Very good
Selectivity
Excellent
Vary Good**
Poor to Good
Poorc,d
Vary good^
Poor*1
Very good
Reaponae tiae
Poor
Good
Good
Good
Good
Excellent
Excellent
Stability
Good
Vary Good
Poor to Good
Good
Very good
Good
Good
Siaplicity
Very coaplex
Mediua
complexity
Slept*
Mediua
coaplexity
Mediua
coaplexity
Very eiaple
No
Saaple ay a tea
required
Tea
Tea
Tea
Tea
Tea
No
Very aiapli
Relative coat
Very high
Mediua
to high
Mediua
Mediua
Mediua
Low
Lou
Rangeability
ppa. to 100X
ppa. to
low X
Hide
Hide
Very Hide
Very wide
Below UEL*
Maintenance
High
Mediua
Mediua
Mediua
Mediua
Low
Low
Auxiliary gaa
auppliea
Yea
Tea
Soaetlaea
No
No
No
No
'Diffusion head, type Minor.
Sllll not ditKt tydrogen.
cHydrocarbona and hydrogen of oppoaite polarity aignale.
^Reaponde to product* of aaiaalon.
eUppar exploaive Halt.
Sourcet Reference 5.
-------
4.2.2 Toxic Vapor Detection Systems
Detecting toxic vapors requires much greater sensitivity than is required
for detecting flammable vapors. Detection systems for toxic vapors often rely
on detectors that respond to a specific chemical. Less often, a single device
can be used to monitor a variety of hazardous vapors. These devices measure
the concentration of the toxic gas and allow the operator or automated system
to compare the measured value with some set ceiling value. If this
concentration value is exceeded, an alarm system is activated.
The design of detection systems is based on a number of principles and
methods, including optical absorption, photoionization, flame ionization, mass
spectroscopy, gas chromatography, pH, measurement, and infrared
spectrophotometry (12). Detailed descriptions of operating principles and
available instruments appear in References 6, 7, and 8. Before committing to
any type of detection system, discussions should be held with equipment
vendors or manufacturers to determine the correct application of the detection
device to the vapors or gases to be monitored. For vapors and gases with both
a lower explosive limit (LEL) and an employee exposure limit, an instrument
capable of measuring the LEL may not be appropriate for measuring the
exposure limit concentration. The LEL may cover only a single decade of vapor
concentrations, for instance 1 to 10 percent, while the toxicity-based
exposure limit may include concentrations from fractions of a part per million
(ppm) to several thousand ppm (5). The wide range of concentrations may
require different types of detection systems.
For example, Table 4-2 lists the various types of instruments available
from one manufacturer for monitoring several toxic gases. A survey instrument
is typically a portable, battery-powered device used to locate leaks or to
measure ambient concentrations of toxic gases for a short time. Personal
sampling instruments are devices that are typically worn by personnel who may
be exposed to high concentrations of toxic gases. These devices are usually
equipped with alarms to warn the wearer when dangerous concentrations are
51
-------
TABLE 4-2. EXAMPLES OF INSTRUMENTS AVAILABLE FOR VARIOUS TOXIC CHEMICALS*
Substanca
Survay
Inittuaant
Parsonal Sampling
Inatruaant
Acaa Monitoring Instrumant
Anoonia
Autospot
PSM-8 Multipoint Monitor, PSM-8a
Multipoint Monitor, Sariaa 7000
Contlauoua Monitor. Sariaa 7100
Continuous Monitor
Chloclna
Autoapot
Moaitox/Chroaotox
PSM-8 Multipoint Monitor, PSM-Sa
Multipoint Monitor, Sariaa 7000
Continuous Monitor, Sarias 7100
Coatlnuoua Monitor, Statox Multipoint
Monitor
Bydraxlna
Bydrogaa
Cyanida
Bydrogaa
FluorIda
Bydrogaa
Sulfida
Autospot
Autospot
Monitox/Ctaronotoz
Moai tox/Chronotox
Autospot
Moaitox/Chroaotox
Phosgaai
Autospot
Mlalatura Coatlnuoua
Monitor, Monitox/
Chronotox
Sulfur
Toxie Vapors
(Saaaral)
Autospot
Vaeu-Saaplar
Aoeuhalar, Law Flow,
'•¦Flint Ptwp
Sarias 7000 Continuous Monitor, Sarias
7100 Continuous Monitor
PSM"8 Multipoint Monitor, PSM-8a
Multipoint Monitor, Sarias 7100
Coatiauous Moaitoe, Statox Multipoint
Monitor
PSM-•• Multipoint Monitor. Sarias 7100
Continuous Monitor
PSM-8 Multipoint Monitor. PSM-Sa
Multipoint Monitor, Sariaa 7000
Continuous Monitor, Sariaa 7100
Contlnuoua Monitor, Statox Multipoint
Monitor, Statox E Multipoint Monitor
PSM-8 Multipoint Monitor, FSM-6a
Multipoint Monitor, Sariaa 7000
Continuous Monitor, Sarias 7100
Coatiauous Monitor, Statox Multipoint
Monitor
PSM-8 Multipoint Monitor, PSM-8#
Multipoint Monitor, Sarias 7000
Coatlnuoua Monitor, Sarias 7100
Continuous Monitor
Rota:
Tlastruaaats listad ara aaaufacturad by MO A Seiantitle, Ine. Thair listing doaa not
eonstituta an andorsaaant.
52
-------
reached. Area monitoring instruments are designed for continuous monitoring
at fixfed locations. A single area monitor can sample several locations if a
switching valve and a sampling manifold are used.
4.2.3 Rflir*-* Mnnigoring
The previous detection methods are applied at a specific single location
or set of locations within a facility. However, a number of techniques have
been developed for remotely detecting hazardous vapors and gases. Remote
detection uses sensors not physically located at the point of release or in
the area directly covered by the plume or cloud.
Most of these systems use some form of light beam directed through the
plume or cloud. These detectors cover an area rather than a single location.
A variety of remote detection systems are available (4,8). Two of the most
widely used laser-based optical monitoring technique are the Differential
Absorption Light Detection and Ranging (LIDAR, also abbreviated as DIAL) and
the Raman LIDAR remote detection technique. Figure 4-1 is a schematic
representation of these systems.
The DIAL technique uses two lasers at different wavelengths. The first
wavelength coincides with band peak of the target gas and the other coincides
with a band where the gas does not absorb. The lasers are positioned so that
scattered or reflected light from natural or positioned targets is detected by
a receiver, which determines the concentration of the target gas by comparing
the wavelengths of the reflected light (since the concentration is
proportional to the difference between the return signals at the two specified
wavelengths).
The p--»" LIDAR technique differs from the DIAL technique In that only
one laser beam is used. The advantage of this technique is that the beam does
not have to be tuned to any specific wavelength. Operation of the Raman LIDAR
technique is based on the phenomenon that light scattered by gas molecules
53
-------
REFERENCE 1
DIAL remote detection system
Raman LIDAR remote detection system
Figure 4-1. Schematic representations of typical remote monitoring
techniques for hazardous vapors.
54
-------
produces a shift In wavelength. The wavelength shi£ts that occur are
independent of the incident wavelength and are characteristic of specific
compounds. Signals reflected off a background are collected by a receiving
telescope and are sent to a detector for concentration determination. The
g^nnati frequency of nitrogen in the air is used as a reference frequency, which
allows the system to discriminate between scattering from the target gas and
that from other atmospheric compounds.
References 4 and 8 discuss in more detail these and other remote
detection systems.
4.3 DESIGN
The design of an effective detection and warning system for hazardous
vapor releases must consider the following (5,13,14,15):
e Selection of the proper detection equipment (for the vapor of
interest and its particular hazard) which meets the appropriate code
or is approved for use within the facility environment;
e Location of sampling points and positioning of detectors;
• Installation of the detection system;
• Type of alarm system; and
• Maintenance, checking, and calibration.
The selection of equipment is based on the likely concentrations of the
specific chemicals to be detected. In some situations, combinations of
devices might be required. The selection process should include information
from equipment manufacturers and require certification that the equipment will
perforin properly in the given application.
55
-------
Equipment used should meet the National Electrical Code and have Factory
Mutual, Underwriters Laboratory, Canadian Standards, or Bureau of Mines
approval for use in Class 1, Division 1, Group A,B,C, and D classified areas
(14).
Detector and sampling points should be located either around the peri-
meter of a potential source or at the source. Sources may include pumps,
valves, flanges, compressor seals, rupture discs, vents, pressure relief
valves, storage containers, stacks, and ventilation and air conditioning
system intakes (5,14). The location of detectors or sampling points should
take into consideration the local wind conditions and structures that may
affect air flow in the immediate vicinity of the source. Multiple detectors
should be located no more than 30 feet apart (13). One reference suggests
that sampling points or detectors should be located 1-1/2 feet above ground
for heavier-than-air combustible compounds and 6 to 8 feet above ground for
lighter-than-air combustible compounds (13). For monitoring measurements of
work place exposure detection points should be selected at the height of
breathing levels (5).
Although most equipment is designed for simple installation, care must be
taken to follow the manufacturer's instructions concerning proper installation
and coordination with other equipment. When sampling lines are used, these
should be of a material that will neither contaminate nor degrade the sample
being routed to the analyzer. Detection sensors and sampling probes should be
protected from the weather, including extremes of temperatures (5).
Installation of equipment should be located so that maintenance can be
performed easily.
Alarm systems can vary greatly in complexity. A simple alarm system may
provide visual and/or audible warning signals once a single or multiple preset
concentration of vapor or gas is exceeded. Facility personnel can respond
appropriately. Complex alarm systems may initiate automatic shutdown of the
process or signal automatic mitigation efforts.
56
-------
Maintenance should be conducted periodically according to the
manufacturer's instructions. The system should be checked daily to ensure
that the detector or sampling lines are not plugged. Calibration and
adjustment of equipment should be performed periodically with vapor or gas
standards of known concentrations. Proper operation and maintenance of the
equipment requires proper training of facility personnel. This training may
be included in the contract for the equipment, where plant personnel can learn
the proper calibration and maintenance techniques from the supplier actually
performing the maintenance during the equipment's warranty period (13).
4.4 APPLICABILITY AND PERFORMANCE
A variety of equipment is available to monitor flammable and/or toxic
vapors from releases within plant environments and to use at remote chemical
spill locations. As indicated in Subsection 4.2, these systems operate
according to numerous detection principles. References 6 and 7 include
detailed descriptions of available equipment. In these references,
approximately 150 different direct-reading instruments and colorimetric indi-
cators are described and the operating principles, performance data, and
inferences are discussed. In addition, Reference 8 discusses several single-
component and multicomponent vapor monitoring devices for monitoring hazardous
releases from railroad accidents.
4.5 RELIABILITY
The reliability of detection and warning systems depends primarily on
proper system maintenance and on visual and functional tests. For catalytic
combustion detector elements, a 3 percent failure per month can be expected
(14). For the portion of the instrument where adjustments for the span, zero,
and alarm set points can be made, an 0.5 percent failure per month can be
expected (14).
57
-------
The reliability of detection systems may be affected by contaminants.
Sulfur compounds may form acids that permanently damage the sensor catalyst,
while halides can coat the sensor and cause a loss of sensitivity (14). For
other types of detection systems, there may be chemical compounds that inter-
fere with the accurate detection of the hazardous vapor being analyzed.
4.6 SECONDARY HAZARDS
Because of potential fire hazards, the use of detection and warning sys*
teas should follow all applicable electrical codes. Where necessary, systems
should be approved for use in classified hazardous areas.
Personnel maintaining and checking release alarm situations should wear
approved safety equipment to prevent exposure to hazardous concentrations of
vapors and physical injuries.
An indirect secondary hazard of detection and warning systems is that
they may engender complacency. These systems should not be substitutes for
proper accidental release prevention measures.
4.7 COSTS
The costs of detection equipment vary widely, depending on whether the
device is monitoring a single vapor or gas or several. Some typical estimated
costs of different types of detection systems are presented in Table 4-3
(A,8,16).
58
-------
TABLE 4-3. EXAMPLES OF ESTIMATED COSTS
Equipment Capital Cost ($)
Manually operated portable sampling pump for
colorimetric indicating tubes 135
Battery powered portable combustible gas
analyzer 300 - 7,000
Multicomponent monitoring devices (gas
chromatographs and infrared analyzers) 5,000 - 35,000
Remote area-wide detection systems (laser based,
infrared, and ultraviolet systems) 15,000 - 175,000
Mobile mass spectrometer 500,000
59
-------
4.8 REFERENCES
1. Lees, F.P. Loss Prevention in the Process Industries. Butterworth's
London, England, 1983.
2. Soden, J.E. Basics of Fire-Safety Design. Fire Protection Manual for
Hydrocarbon Processing Plants, Volume 1, 3rd Edition. Gulf Publishing
Company, Houston, TX, 1985.
3. Prugh, R.W. Post-Release Mitigation Design for Mitigation of Releases.
Presented at International Symposium on Preventing Major Chemical
Accidents, American Institute of Chemical Engineers, New York, NY,
February 1987.
4. Atallah, S. and E. Guzman. Remote Optical Sensing of Fire and Hazardous
Gases. Presented at International Symposium on Preventing Major Chemical
Accidents, American Institute of Chemical Engineers, Washington, D.C.,
February 1987.
5. Dailey, W.V. Area Monitoring for Flammable and Toxic Hazards. Loss Pre-
vention, Volume 10, American Institute of Chemical Engineers, New York,
NY, 1976.
6. Nader, J.S., et al. Direct Reading Instruments for Analyzing Airborne
Gases and Vapors. Air Sampling Instruments for Evaluation of Atmospheric
Contaminants, 6th Edition. American Conference of Governmental Indus-
trial Hygienists, Cincinnati, OH, 1983.
7. Saltzman, B.E. Direct Reading Colorimetric Indicators. Air Sampling
Instruments for Evaluation of Atmospheric Contaminants, 6th Edition.
American Conference of Governmental Industrial Hygienists, Cincinnati,
OH, 1983.
8. Hobbs, J.R. Monitoring Devices for Railroad Emergency Responses Teams.
Department of Transportation. Federal Railroad Administration Report
Number D0T/FRA/0RD-86/02. Cambridge, MA, February 1986.
9. National Fire Protection Association. Fire Protection Handbook, 15th
Edition. Quincy, MA, 1981.
10. Brown, L.E. and L.M. Romine. Flammable Liquid Gases. Hazardous Mater-
ials Spill Handbook. McGraw-Hill, New York, NY, 1982.
11. Rodgers, S.J. Commercially Available Monitors of Airborne Hazardous
Chemicals. Hazardous Materials Spill Handbook. McGraw-Hill, New York,
NY, 1982.
12. Rome, D. Personnel Safety Equipment, Hazardous Materials Spill Handbook.
McGraw-Hill, New York, NY, 1982.
60
-------
13. MDA Scientific, Inc. Toxic Substance Detection and Measurement
Instrumentation. Cataloque No. 970358, October, 1985.
14. Johanson, K.A. Design of a Gas Monitoring System. Loss Prevention,
Volume 10, American Institute of Chemical Engineers, New York, NY, 1976.
15. St. John K. Use Combustible Gas Analyzers. Fire Protection Manual for
Hydrocarbon Processing Plants, Volume 1, Third Edition. Gulf Publishing
Company, Houston, TX, 1985.
16. Schaeffer, J. Use Flammable Vapor Sensors? Fire Protection Manual for
Hydrocarbon Processing Plants, Volume 1, Third Edition. Gulf Publishing
Company, Houston, TX, 1985.
17. Telephone Conversation between R..C. Keeney of Radian Corporation and a
representative of Mine Safety Appliances Company, Houston, TX, May 1987.
61
-------
SECTION 5
VAPOR DISPERSION MODELING
Mathematical models that accurately simulate the movement of a hazardous
vapor cloud from an accidental release are valuable for predicting the effect
of various accidental release scenarios and, to a limited extent during an
actual release event, for determining whether community evacuation is
required. Models are available today that support both of these tasks.
However, the usefulness of a model during an actual release event is often
limited by a lack of information and by the time required to activate the
modeling process. Vapor dispersion models should not be used as the only
source of information for decision mating during an actual release event, but
they can help the emergency response expert choose the appropriate response
for the specific situation.
This section of the manual discusses vapor dispersion models for predict-
ing the impact of an accidental release. Types of models are discussed in
terms of their applicability, performance capabilities, and reliability.
5.1 BACKGROUND
Mathematical vapor dispersion modeling to predict the potential impact of
an accidental release is a relatively new field. Traditional vapor dispersion
modeling has focused on the atmospheric dispersion of pollutants from an
elevated stack. Such models deal with low concentrations of pollutants in
air continuously emitted over a long period of time. Models for predicting
the effects of an accidental release, however, must be able to handle
short-term releases at both high and low concentrations and at variable
release rates. They must be capable of modeling a release/dispersion of
heavier-than-air and lighter-than-air materials, and materials that have the
62
-------
sane density as air. Such models should simulate a variety of possible
release forms, such as a release from a boiling pool of liquid, or a release
from a hole in a pressurized vessel.
One purpose of a vapor dispersion model is to predict the area that might
be adversely affected by the vapors from a release. The adverse effects will
depend partly on the properties of the released chemical. For flammable
materials, the presence of vapor concentrations within the flammable limits of
the material are of concern because such concentrations can be ignited. For
toxic materials, the presence of vapor concentrations greater than the
compounds IDLH (immediately dangerous to life and health) will be of concern
(1).
One modeling method used to predict vapor cloud movement is experimental
physical modeling, such as in wind tunnels or water channels. In these
experiments, a scale model is used to simulate the release, using air and the
actual chemical of interest or liquid substitutes. Some success has been
achieved using these methods (2,3,4); however, the rest of this section
focuses on mathematical models because of their convenience and intense
development in recent years.
Mathematical vapor dispersion models may be used for two purposes: to
assess hazards and plan an emergency response, and to provide emergency
response guidance information during an actual accidental release. Modeling
may be used to predict the effects of various accidental release scenarios and
to estimate which scenarios present the greatest threat to plant personnel and
to the community. Modeling may be used to estimate what portions of the
community might be affected during various accidental release scenarios. This
information can be used to develop community emergency response plans. Models
may also be vised during an actual release event to help the emergency response
coordinator decisions about what action should be taken if a release
occurs.
63
-------
A distinction should be made at this point between individual dispersion
models and what could be called emergency response modeling packages. Many of
the models sold commercially would come under the category of emergency
response modeling packages. Such a package will typically contain a set of
source models, a dispersion model with a number of submodels to handle
non-idealities, a database with chemical property data for a variety of
potential release candidates, the ability to use real-time meteorological data
if an actual release occurs, and a visual display that shows the physical
motion of the cloud. Such packages are often customized for specific sites
with a map of the area shown under the released vapor cloud. Some packages
can be set up to communicate directly with detection and emergency response
notification systems.
Some mention of "real-time" response models should be also be made. For
a model to be "real time" it oust react or change its simulation as actual
events occur. In most accidental release situations, a model is activated
some time after the release has begun (from minutes to hours). Information
describing the actual event, necessary for model input, may also not be
available for some time after the release has begun (if at all). A true
"real-time" model would have to be monitoring sensors at possible event sites.
When it appears that a release may be imminent, the calculations are acti-
vated. The necessary data would be obtained by the model itself. Such a
system would be quite complex, and no such system is known to exist. The
existing "real-time" response models usually only use real-time meteorological
data in a predefined release scenario.
5.2 DESCRIPTIONS
Two key components of models must be available to fully describe an
accidental chemical release: a source model component and a vapor dispersion
model component.
64
-------
A source model is required to calculate the effect of the initial release
conditions on vapor dispersion. For example, a release from a refrigerated
storage vessel containing a liquefied gas will behave differently than a
release of the same material from a pressurized, non-refrigerated storage
vessel. The difference between these two releases is not a difference in the
material released, but in its physical state. Source models must be used to
calculate the effects of these differences. Examples of the types of
information generated from a source model which becomes information that a
dispersion model can use include the following: the rate of vapor generation,
the height of the release, the initial velocity and temperature of the vapor,
and the initial concentration of the released vapor.
Each type of release will require a different set of mathematical
equations to define the nature of the release. The physical state of the
released material and the way in which the loss of containment occurs will
both affect the calculations required to define the release. Examples of
possible physical states of released materials are:
• Release of a pressurized liquefied gas;
• Release of a chilled liquefied gas, stored at atmospheric pressure;
• Release of a pressurized gas;
e Release of a liquid stored at ambient or elevated temperature' and
pressure; and
• Release of a two-phase slurry.
Some examples of different ways a release may occur are:
65
-------
• Release from a rupcured pipe;
• Release from a hole in a vessel;
• Release from an acCivaCed pressure relief device;
• Sudden release of Che entire contents of a vessel;
• Opening a valve that results in a release from a non-blanked pipe;
and
• Any of the above releases within a diked area, or any of the above
released within an enclosure.
A wide variety of release scenarios can be modelled by combining the
physical state of the released material with the conditions under which the
release occurs. Figure 5-1 illustrates a few possible release situations.
Ideally, a source model will be able to take these combinations and ambient
weather conditions and determine whether a boiling pool of liquid is formed,
whether the release is entirely gaseous, whether an aerosol is formed or if
the release is some combination of each. This information is then used to
define the initial conditions of the vapor from the release.
Source models can be obtained either as part of a model package that
contains both source and dispersion models, or as separate packages. Table
5-1 gives the names of some source models, although this list is not exhaus-
tive. Examples of modeling packages that contain both source and dispersion
models are presented in Subsection 5.2.2 on dispersion models. Additional
information about these models will be discussed in Subsections 5.3 and 5.4.
66
-------
SMALL HOLE M THE VAPOR
SPACE OF A PRESSURIZED TANK
CATASTROPHIC FALURE OF
A PRESSURIZED TANK
TWO-PHASE RBJEASE
FROM A PRESSURIZED TANK
SPREADM8P00L0F
REFRIGeUTH)UQUD
SP&L OF REFRIGERATED
LIOUD NTO DIKED AREA
HIGH VELOCITY JET OF
REFRIGERATED UQUD
Figure 5-1. Potential release scenarios.
67
-------
TABLE 5-1. EXAMPLES
OF SOURCE MODELS
Authors
Sponsor
Application
Tang eC al (1980)
U.S. Coast Guard
Oleum/HjSO^ spills on water
Clewell (1983)
U.S. Air Force (ESL)
Parameters presented for
approximately 25 chemicals
Shen (1982)
New York State Dept.
of Env. Conservation
Volatilization from exposed water
surface
tfhittaker (1977)
Alberta Environment
H2S risk assessment modeling
Kahler et al
(1978)
U.S. Air Force
Specific to ^0^ (pool evaporation)
WhiCaere and
.Myriski (1982)
U.S. Army
Toxic liquid spill (pool
evaporation)
Ills and Springer
(1978)
U.S. Air Force
Dispersion modeling of propellant
spills (pool evaporation)
Vu and Schroy
(1979)
Monsanto Company
Pool evaporation, chemical spills
Reid and Wang
(1978)
American Gas
Association
LNG spills in different substrates
Shan and Brisiac
(1978)
U.K. Safety and
Reliability
directorate
LNG applications with suggestions
other liquids
Reid and Smith
(1978)
California Public
Utilities Commission
Estimates of LNG release- rates for
subsurface spills
Georgakis at al
(1979)
Distrigas Corp.
LNG and Gasoline
Source: Reference 5
68
-------
5.2.2 Dispersion Models
Of Che two extreme approaches for modeling vapor dispersion, accempcs to
solve Che problem exaccly by using differencial equations chat describe each
process affecting Che mocion and dispersion of Che vapor in air. The
equaCions are integrated wich time over three-dimensional space. This type of
model is called a three-dimensional numerical model.
The second excreme is one that makes simplifying assumptions for each
differencial tern used in the three-dimensional numerical model until the
equations can be combined and simplified Co form a single equation.
A large number of models can be formed with characteristics of both these
extremes. By making a limited set of simplifying assumptions, models can be
created whose solutions cover a specific range of accidental release cases.
In this manual, such models are called similarity models.
Similarity models
Similarity models avoid the complexity of specifying the details of how
the physical properties of the cloud vary from point to point and, instead,
are written in terms of the bulk or integrated properties of the cloud.
Instead of calculating how the velocity, concentration, and other
characteristics change along a stream line, they determine how these bulk
properties change as the cloud moves downwind. These models generally assume
that the material behaves as a mass leas quantity that in no way influences the
turbulent atmosphere in which it disperses. One o£ the best known examples of
this type of model is the gaussian plume model.
When the cloud formation time is very small compared with the time of
cloud travel, the cloud may be approximated by the formation of an
instantaneous cloud or "puff", as it is typically referred to. If the release
is continuous, some similarity models will approximate the release by modeling
a series of instantaneous "puffs." The concentration at any point may be
determined by summing the contributions at the point from each of the puffs.
69
-------
By making che assumption of continuous source strength and a constant wind
speed and direction, this summing can lead to a single expression describing
the concentration anywhere in the plume. This expression is known as the
continuous Gaussian plume model. By using puffs, noncontinuous time-dependent
source strengths and meteorology can be simulated (5,6).
In practice, similarity models often solve differential equations with
simplifying assumptions. For example, a similarity model may account for the
change in the volume of a puff caused by entrainment by using differential
equations of motion and entrainment, but make the simplifying assumption that
the concentration in the puff is Gaussian in nature.
A wide number of similarity models are available both publicly and from
commercial sources. The differences, or lack of difference, between these
models was summarized by one author as follows (7);
The current generation of public domain and commercially available source
or dispersion models is not notably innovative in formulation. This is a
result of the current state of the science of dispersion meteorology.
Differences among models do not represent innovation, but are related to
the selection and combination of different theoretical expressions for
efficient programming or the use of different frameworks for data
handling and presentation.
Table 5-2 presents examples of similarity dispersion models. Table 5-3
presents examples of modeling packages that contain source and similarity
dispersion models. Table 5-4 shows examples of commercial modeling packages
that use similarity dispersion models. These lists are not exhaustive.
Additional information is presented in Subsections 5.3 and 5.4.
Three-dimensional numerical models
In t^ree"dlJB*nsional models, the distributions of velocity and concentra-
tion are determined by the solutions of the conservation equations of fluid
flow and the prescribed boundary conditions. These models take a much more
70
-------
TABLE 5-2. DISPERSION MODELS
Model Name
Sponsoring Organization
British Gas
(Chlorine Institute)
DENZ
DEGADIS
INPUFF
(IEPA)
HEGADIS-II
OB/DG
EIDSVIK
SLUMPING
SLAB
SRI PUFF
WINDS
Cremer and Warner
Chlorine Institute (CI)
U.K. Safety & Reliability Directorate (SRD)
Dense Gas Dispersion Model U.S. Coast Guard
EPA Integrated Puff Model
Illinois EPA
Heavy Gas Dispersion from Area Sources Koninklijke/
Shell Lab
Ocean Breeze/Dry Gulch (U.S.A.F.)
Norwegian Institute for Air Research
Royal Netherland Meteorological Institute
Lawrence Livermore National (LLNL)
SRI International
Savannah River Lab (SRL)
Source: Reference 5
71
-------
TABLE 5-3. SOURCE AND DISPERSION MODEL PACKAGES
Model Name
Sponsoring Organization
AGA (LNG)
AGA
DDESB
ICARIS
HACS
FUMING ACID
(Standard Oil Co.)
TOXCOP
SPILLS
American Gas Association (AGA)
Germeles & Drake
Department of Defense Explosives Safety Board (DDESB)
Industrial Chemical Accident Response Information
System of the Association of American Railroads
Hazard Assessment Computer System U.S. Coast Guard
Proctor & Gamble
Standard Oil Co. (Indiana)
Toxic Corridor Prediction U.S. Army Atmospheric
Sciences Lab (ASL)
Shall Development Co.
Source: Reference 5
72
-------
TABLE 5-4. COMMERCIAL MODELING PACKAGES
Model Name Sponsoring Organization
Dispersion Models
DAISY Dispersion Analysis Information System,
Dow Chemical
TRPUF Trinity Consultants
Source and Dispersion Model Packages
Energy Impact Associates
Complex Hazardous Air Release Model
Radian Corporation
Concord Scientific Corporation
Dames & Moore
Environmental Systems Corporation (ESC)
Environmental Research and Technology (ERT)
ETA, Incorporated
Pickard, Lowe, and Garrick, Inc.
IMPELL Corporation
Safer Emergency Systems
Source: Reference 5
73
CEES/ERADAS
CHARM
COBRA
CARE
AIRTOX/HASTE
ETA
MIDAS
MESOCHEM
SAFER
-------
rigorous and theoretical approach to the diffusion problem than do similarity
models. A primary use of these models is to predict the dispersion of a vapor
cloud around obstructions and irregular terrain.
Several three-dimensional numerical models have been developed for use on
toxic releases. All are experimental at this point. Additional information
may be found elsewhere (6,8).
5.3 APPLICABILITY AND PERFORMANCE
The models presented in Tables 5-1 thru 5-4 are not all representative of
the state of the art; some have been superseded by further developments.
Also, some of these models were developed for very narrow applications and
will be of limited use for general accidental release modeling. Some models
were developed for a purpose that did not require the accuracy necessary for
modeling accidental releases of toxic materials. These models incorporate
simplifications that are not appropriate for accidental release modeling. The
potential users must determine their requirements and select a model
accordingly. One source has summarized the important questions in model
selection as follows (7);
• What is the time frame in which results are required?
• Who is the intended user?
e What is the source configuration and what are likely accident
scenarios?
• What are the chemical and physical characteristics of the material
released?
• Does local terrain influence dispersive patterns and what is the
local dispersion climate?
74
-------
The present generation of models is most valuable as an emergency
response planning tool. Models should not be used as the sole decision-making
tool during an accidental release, and with the current limitations of
"real-time" modeling they may be of little use during an actual event.
Instead, model results should be used to help the emergency response expert
during an accidental release situation.
Ideally, before a release occurs, single page reports containing response
information for likely releases would be created by using a model. These
single pages could then be used by nontechnical persons in an actual event
until more information was available for a refined estimate of impact.
Only those models that use an empirical or simplified approach to model
vapor dispersion (similarity models) are useful in the emergency response
model because of the calculation time involved. These simplifications
restrict the accuracy of the models. However, these same simplifications
allow the models to fairly rapidly produce a result accurate enough for the
purposes mentioned above. At this time, models that take a more rigorous
approach (three-dimensional models) cannot produce rapid results. Also, these
models require large quantities of data to achieve accurate results and such
data may not be available.
Several potentially desirable modeling features are either not available,
or are not fully developed at this time. Some of these features are discussed
below:
e The effects of post-release mitigation measures such as foams
applied to the spill surface, or water sprays;
• Adequate complex terrain models;
e Modeling of atmospheric turbulence;
75
-------
• Models for certain types of sources, such as high-pressure jets, or
two phase releases;
• The ability to model multi-component releases;
• Accounting for chemical reactions, such as the effect of atmospheric
moisture on water reactive releases; and
• The effect of fire on the release and the dispersion of the result-
ing fire decomposition products.
5.3.1 Swrcs
Host source models used in conjunction with air models are designed to
calculate the evaporation rate from pools of spilled liquid. These models
calculate an evaporation rate based on mass and heat transfer equations. All
evaporation source models apply some simplifying assumptions. Some models
assume that the pool forms instantaneously and is of a fixed radius defined by
the model user. Other models calculate the size of the pool as a function of
time. Most models have restrictions on the range of materials that can be
modeled. Additional simplifications made by most vaporization models include
(5);
• Models are for single component systems;
• Ice formation in the soil is not considered;
• Percolation into the soil is specified (as opposed to calculated),
or not considered; and
• Runoff from the spill area is specified or not considered.
76
-------
For models Chat calculate Che size of a spill pool, calculation of the
race of spilled maCerial inCo Che pool is based on user-defined spill
paramecers. Most ofcen Chese parameters are Che size and locaCion of a hole
in a vessel. The models will cypically use an orifice flow equation Co
calculate Che flow into the pool. Source models Chat handle vapor releases
will also use some type of orifice flow equation to calculate the rate of
vapor release.
The modeling of more complicated sources has not been incorporated into
vapor dispersion modeling. Examples of complicated sources would be the
release of two-phased flow or the release from a high-pressure jet. There are
two reasons why these types of flow have not been Incorporated into
atmospheric dispersion models. First, the behavior of such releases is not
well understood and the mathematics of the sources have not been developed.
Second, the complexity of these sources would make it difficult to get mean-
ingful results when used with similarity models. A complex source would
probably need to be used with the more complex three-dimensional models, and
three-dimensional models themselves are not yet practical for modeling acci-
dental releases.
Another issue of concern when considering source model reliability is the
selection of an appropriate source model for a given situation. Each poten-
tial type of release will require a unique combination of calculation proce-
dures to model the result of the release. The issue of source modeil selection
is of particular concern when a release is being modeled as it occurs. In
this event, the modeling package used must not only have a source model for
the type of release that has actually occurred, but must be able to select
that model and set ell of the model parameters to be consistent with what has
actually occurred. This shows one of the limitations of using dispersion
models for modeling "real-time" events.
77
-------
5.3.2 Dispersion Models
Similarity Models
As was discussed above, similarity models use a simplified approach for
solving Che problem of Che dispersion of gases from an accidenCal release.
Even with these simplifications, similarity models offer the best available
option for modeling such releases. The simplifying assumptions allow the
models to operate rapidly enough on a computer to be used to predict the
movement of a release even after the release has begun. All of the emergency
response modeling packages sold commercially incorporate similarity models.
As will be discussed below, modifications to these models have allowed some of
them to handle some of the nonideal conditions encountered in real release
situations.
A state-of-the-art similarity model, interfaced with a state-of-the-art
source modeling package will produce accurate enough results to be used for
the following purposes;
e By modeling a series of credible worst case releases, plant design-
ers can determine whether the siting and layout of a potential plant
is appropriate.
e Plant designers can also use these modeled events to evaluate what
prevention, protection, or mitigation technologies might be appro-
priate at various locations within the plant.
e A series of credible worst-case releases can be modeled for an
existing plant. Plant management and engineers can use the results
to evaluate the same features mentioned above for a new plant. A
heavier emphasis would be placed on prevention, protection, and
mitigation than on siting and layout. The effect of some of these
control measures could be tested using the model. For example, many
models allow the effect of diking to be tested.
78
-------
• An emergency response system thac incorporates a modeling package
could be sec up Co help Che emergency response coordinaCor decide
vhac actions would be appropriate in the event of an accidental
release.
As has been discussed, similarity models are not accurate enough Co serve
as the sole basis for decision making during an actual accidental release
evenc.
The assumptions and simplifications incorporated into similarity models
impose basic restrictions on prediction accuracy. Individual models will
incorporate varying degrees of complexity and varying degrees of accuracy.
However, certain simplifying assumptions will be common among most similarity
One source has summarized these assumptions as follows (5);
Meteorological conditions are assumed to be constant during the
simulation period, which is typically one hour. The effects of any
systematic change or trend in wind speed, wind direction, or stabil-
ity conditions during an hour would not be described by the model.
Vinds and turbulence are assumed to be the same at all locations
throughout the boundary layer. The effects of wind speed, or wind
direction shear, or of changes in turbulence with height are not
considered within the Gaussian formulation.
The basic model averaging time is assumed to be long, compared with
the time scale of turbulent atmospheric motion and wich the
transport time from source to receptor.
Pollutant mass is conserved within the Gaussian formulation.
Processes that add or remove mass, such as deposition, decay, or
chemical transformations, are assumed to be of secondary importance.
models.
e
e
79
-------
Many models have incorporated modifications in an actempc to improve che
performance for non-ideal situations. Some incorporate mathematical expres-
sions to account for the movement of gases that are heavier than air (many
toxic materials fall into this category) (5,9,10). Methods have been
developed to account for simple wind shear and for occasional changes in wind
speed and direction (5). Some work has been done to allow for the deposition
of materials from the vapor cloud, or for the time-dependent decay of the
material. A few similarity models have incorporated methods for simulating
the effect of certain types of complex terrain. All of these modifications
are made using simplified approaches to complex phenomena. However, most of
these phenomena have not been modeled using more rigorous approaches, and for
the present time, the simplified approaches offer the only practical solution.
Naturally, the accuracy of predictions made by a similarity model will
depend somewhat on the accuracy of the source model used. As a result,
predictions made using an adequate source model can be misleading.
Additionally, the way the source model interacts with the dispersion model may
affect the overall accuracy.
Three-Dimensional Models
Three-dimensional models can, in principle, simulate the dispersion of
vapors as a function of space and time without the simplifications incorporat-
ed into the similarity models. These models should be able to account for the
effects of terrain and wake turbulence. Although they appear to be
theoretically correct, several factors limit their usefulness at this time.
First, the model theories and assumptions have been developed from an
inadequate experimental data base. A variety of methods have been developed
for handling the effects of turbulence on diffusion, yet the accuracy of the
assumptions in each case are difficult to evaluate because of the present lack
of experimental data (6).
80
-------
Second, these models require large quantities of data, large computer
storage and computational capacity, and long calculation times, features that
tc difficult to run multiple comparative runs for planning purposes, and
that make their use during actual release events impractical.
Finally, most of these models have yet to develop computational routines
that give highly accurate results. For example, most of these models compute
diffusion by dividing the physical vapor cloud and time into many pieces or
elements and performing momentum, mass, and energy balances around each
element. An error will be introduced into the calculation that is inversely
proportional to the number of elements that the cloud is divided into and the
size of the time increments chosen. An infinite number of elements will
result in no error, while dividing the cloud into one element will result in a
very large error. Also, there are rounding errors produced by carrying a
finite number of significant figures in each calculation. These errors
accumulate and can become quite significant. The size of the accumulated
error will be inversely proportional to the amount of computer computation
time required. This compounds the problem mentioned in the previous
paragraph.
Because of their potential accuracy and ability to model complex situa-
tions, three-dimensional models may eventually replace similarity models as
the preferred method for modeling accidental releases; however, the areas of
dispersion theory, numerical methods, and computer technology need
improvements before this will occur.
5.4 RELIABILITY
One way to measure model reliability is to compare actual test data with
model prediction. A limited number of such tests have been performed, and
comparisons of the data with several models have been published. A major
comparison of models was done for the Chemical Manufacturers Association (5).
81
-------
The list of models mentioned in the report is quite large, but actual data
comparisons were only performed for nine of the models. The material released
was Freon.
The U.S. Air Force has compared other models with field data. The
results are shown in References 11 and 12. Comparisons were for nitrogen
tetroxide spills and an inert tracer. In 1986, Amoco Chemical Company
performed a series of full-scale hydrogen fluoride releases at the Department
of Energy's spill test center in Nevada. Liquid hydrogen fluoride, stored
under pressure and at ambient temperature, was released to determine if such a
release could be modeled accurately by existing source and dispersion models.
Test results were presented in late 1987 (13).
There are plans to continue test releases of toxic substances to gather
data for model comparison/development. For the most part, these tests are
carried out at the Department of Energy's Nevada Test Site.
5.5 COSTS
Models in the public domain are available for a copying fee that may
range up to several hundred dollars. The main expense associated with most of
the public domain models is the time the user needs to become familiar with
the model and adapt it to specific requirements. Most of the models in the
public domain have limited applicability; therefore, several models may be
required to meet all the needs of the user.
Most of the commercial modeling packages are assembled so that compara-
tively little time is needed to become familiar with the model. Also, many
companies that market the commercial modeling packages will customize the
model for the specific needs of the user. Several man-months may be required
for such an effort. Such features are reflected in the higher cost of
commercial modeling packages.
82
-------
Commercial modeling packages may be divided into two cost groups. The
first is composed of modeling packages that sell for less than $50,000
(includes hardware and software). At the bottom end of the cost range is one
company that has packaged the EPA puff model (with no heavy gas dispersion
capability) for use on a personal computer. At the top end of this cost range
are packages that include multiple source models, data bases with the physical
and chemical properties of many different chemicals, and a meteorological
tower that sends data to the system to be used during an actual accidental
release. These higher-priced packages incorporate a degree of flexibility so
that the user can adapt the package for specific needs or use the model for a
number of different release situations.
The second group of commercial modeling packages are those that sell for
over $100,000 (including hardware and software). These systems have the
capabilities of the higher-priced models from the first group, while adding
extensive customization for the specific location where they will be applied.
The packages can also include a number of features that aid the emergency
response effort but that are not part of the modeling system. Such features
may include recorded phone messages that are automatically sent to local fire
and police in the event of an emergency.
83
-------
5.6 REFERENCES
1. Mackison, F.W. , Scricoff, R.S. (eds.). NIOSH/OSHA Pocket Guide to
Chemical Hazards. U.S. Government Printing Office, DHHS (NIOSH) Publica-
tion No. 78-210, Washington, D.C., 1980.
2. Wighus, R. Simulation of Propane Dispersion from Spills in a Petrochemi-
cal Plant Using a Water Tunnel Technique. Presented at Fourth
International Symposium on Loss Prevention and Safety Promotion in the
Process Industries. The Institution of Chemical Engineers, Symposium
Series No. SO, London, England, 1983.
3. Cheah, S.C., J.W., Cleaver, and A. Millward. The Physical Modeling of
Heavy Gas Plumes in a Water Channel. Presented at Fourth International
Symposium on Loss Prevention and Safety Promotion in the Process
Industries. The Institution of Chemical Engineers, Symposium Series No.
80, London, England, 1983.
4. Bardley, C.I., and R.J. Carpenter. The Simulation of Dense Vapour Cloud
Dispersion Using Wind Tunnels and Water Flumes. Presented at Fourth
International Symposium on Loss Prevention and Safety Promotion in the
Process Industries. The Institution of Chemical Engineers, Symposium
Series No. 80, London, England, 1983.
5. McNaughton, D.J., A.A. Marshall, P.M. Bodner, and G.G. Worley.
Evaluation and Assessment of Models for Emergency Response Planning,
Prepared for Chemical Manufacturers Association by TRC Environmental
Consultants, Inc., 1986.
6. Havens, J.A. Mathematical Models for Atmospheric Dispersion of Hazardous
Chemical Gas Releases: An Overview. Presented at International
Symposium on Preventing Major Chemical Accidents. American Institute of
Chemical Engineers, New York, NY, 1987.
7. McNaughton, D.J., G.G. Worley, and P.M. Bodner. Evaluating Emergency
Response Models for the Chemical Industry. Chemical Engineering
Progress, January 1987.
8. Havens, J.A. Evaluation of 3-D Hydrodynamic Computer Models for Predic-
tion of LNG Vapor Dispersion in the Atmosphere. Gas Research Institute.
Contract No. 5083-252-0788, Annual Report, March 1984 - February 1985.
9. Van Olden, A.P. On the Spreading of a Heavy Gas Released Near the
Ground. Presented at First International Loss Symposium, The Hague,
Netherlands, 1974.
10. Van Olden, A.P. A New Bulk Model for Dense Gas Dispersion:
Two-Dimensional Spread in Still Air. I.U.T.A.M. Symposium on Atmospheric
Dispersion of Heavy Gases and Small Particles. Delft University of
Technology, The Netherlands, 1983.
84
-------
11. Carney, T.A., and M.M. Lukes. A Comparative Study and Evaluation of Four
Atmospheric Dispersion Models with Present or Potential Utility in Air
Force Operations. Air Force Office of Scientific Research (AFOSR), 1987.
12. McRae, T.G., et al. Eagle Series Data Report: 1983 Nitrogen Tetroxide
Spills. Lawrence Livermore National Laboratories. Report Number
UCID-20063, June 1984.
13. Results presented at International Conference on Vapor Cloud Modeling.
American Institute of Chemical Engineers (Center for Chemical Process
Safety) and U.S. Environmental Protection Agency. Cambridge, MA,
November 2-4, 1987.
85
-------
SECTION 6
METEOROLOGICAL INSTRUMENTATION
Meteorological data can serve several roles In Che micigatlon of
accidental releases. First, historical data can be used to run vapor/gas
dispersion studies for planning emergency response procedures. Second,
real-time meteorological data are essential for choosing correct mitigation
and emergency response actions at the time of an actual release.
This section presents guidelines for gathering and using meteorological
data. General specifications for instrumentation are recommended, and basic
considerations for the siting of meteorological systems are discussed.
The information presented in this section is general in nature; meteoro-
logical data needs and the design of meteorological systems for specific
applications must be assessed case by case.
6.1 BACKGROUND
When applied to facility design and emergency response planning,
meteorological data are useful for selecting mitigation measures for
accidental releases, tfhen an actual release occurs, these data can facilitate
emergency response by real-time estimation of toxic cloud transport and of the
potential effects on sensitive receptors. Meteorological data can also be
used to analyze past events and to predict the consequences of various
accidental release scenarios of possible future events. The following
subsections discuss the utility of historical and real-time meteorological
data to a facility or the surrounding community for accidental release
applications.
86
-------
6.1.1 H
-------
careful consideration of meteorological data at this point can reduce Che
emergency response planning effort if potential toxic release points are
located. In the case of an existing facility, meteorological data can still
be used to help determine the potential effects on sensitive receptors, which
makes possible an assessment of the Cype and degree of emergency response
planning needed. The methodology for using meteorological data for emergency
response planning is similar to that described in the previous subsection for
facility design.
Historical Analysis of Accidental Releases--
After an actual release, meteorological data from the release period can
be combined with other data and observations of the actual release to better
understand what has occurred. Meteorological data used for this purpose can
provide comparative information on observed effects versus predicted effects.
This information may be useful for mitigating future hazardous releases. For
example, the placement of potential toxic release points in future facility
designs or emergency response planning might be influenced by the analysis of
an actual release.
A simplified approach to using meteorological data for historical
analysis involves using wind direction and wind speed data for arriving at a
crude estimate of the transport path of the released cloud. A more in-depth
analysis would use hazardous release dispersion models for predicting the path
of transport and the specific effects on selected receptors along the
transport path.
On-Site versus Off-Site Historical Data--
Historical data used for the purposes described in the previous subsec-
tions should be obtained from on-site instrumentation if the meteorological
instruments are sited properly (see Subsection 6.3.4, Siting of Meteorological
Instrumentation). In many cases, however, historical data gathered on site
may not be available or may be insufficient. In such cases, off-site data may
88
-------
have to suffice. If off-site data are used, the adequacy of the data must be
carefully assessed. Off-site meteorological data collected near the plant
would be more representative of conditions at the plant than off-site data
collected far from the plant. Data obtained from a distant source or from an
area with significantly different topography may not be useful.
6.1.2 Real-Time oelcal Data
Real-time meteorological data at the time a release is occurring can be
used by the emergency response team in appropriate models to predict the
concentration levels and areas affected by a cloud or plume, thus enabling the
response team to ir'^* more effective decisions about mitigation action.
The ability of the emergency response team to use real time data and a
dispersion model depends on the conditions of the emergency (e.g., available
reaction time, character of the release). In some cases, the usefulness of
the model may be limited to a subsequent analysis of the event, as described
in the previous section.
6.2 DESCRIPTION
This section describes the meteorological equipment that may be used for
accidental release applications. Stability measurement techniques (inferred
from measurement of other variables) and meteorological towers and data
acquisition systems are also included. Much of the discussion of
meteorological instrumentation is extracted from guidelines previously
published by the U.S. EPA (2).
The intended use for the data determines which meteorological parameters
should be monitored. When no dispersion model will be used, wind direction
and wind speed data may be sufficient. When a dispersion model is used,
additional variables such as temperature and humidity will probably be
necessary. The variables needed depend on the model used (see Section 5).
89
-------
Relative humidity and barometric pressure data are used in some hazardous
release dispersion models; however, the sensitivity of the models to
variations in these two parameters is minimal, and climatological averages can
often be used. Thus, it is generally not necessary to obtain actual
measurements of these parameters.
6.2.1 Winfl Direction
Wind direction must be known in order to determine which direction the
vapors from an accidental release will travel. Wind direction is measured
with a wind vane connected to the necessary electronic circuitry to provide a
remote readout and to record, if necessary, wind direction. Though rarely
used for remote sensing, an elevated wind sock should also be used as a visual
indicator of wind speed and direction. An elevated wind sock allows plant
personnel to get an instant visual indication of wind direction from many
locations at the facility, and it can also be seen off site, depending on its
placement.
6.2.2 wmd Speed
Wind speed determines how quickly the plume or cloud from an accidental
release will reach off-site receptors. A number of wind speed measurement
systems that operate on different principles are available. The most commonly
used sensors are the rotational cup and propeller anemometers. Cup
anemometers operate on the principle that net torque (lift greater than drag)
causes a rate of rotation roughly proportional to wind speed. The three
conical cup design performs best. With propeller anemometers, the propeller
turns at a rate almost directly proportional to the wind component parallel to
its axis. Electronic circuitry converts the rotary motion of the cups or
propeller to an electrical signal for remote readout equipment.
90
-------
6.2.3 Ambient Temperature
Ambient temperature data are used In dispersion modeling to calculate
cloud buoyancy and the heat available for evaporating volatile liquids. The
most appropriate temperature measurement devices for meteorological
applications are linear thermistors and resistance temperature detectors.
Thermistors are electronic semiconductors made from metallic oxides. The
resistance of a thermistor varies inversely with absolute temperature. The
system configuration is typically designed with Che thermistor connected to a
bridge circuit that provides an output voltage that varies directly with the
temperature of the thermistor.
Resistance temperature detectors (RTDs) operate on the principle that the
electrical resistance of a pure metal increases with temperature. An RTD
operates connected to a bridge circuit that provides an output voltage that
varies directly with the temperature of the RTD.
6.2.4 Stability
When accidental release dispersion modeling is conducted, data for
atmospheric stability are essential for estimating the spread of the plume.
Stability data are derived indirectly from measurements of other parameters.
There are several methods for computing atmospheric stability. For example,
one method calculates stability by evaluating the cloud cover, ceiling height
and wind speed (Turner's 1964 method). Another method uses the standard
deviation of the vertical wind direction; one uses the standard deviation of
the horizontal wind direction; another uses vertical temperature differences.
These methods are discussed extensively in the literature and in U.S. EPA
guidelines (3).
91
-------
The method used for calculating stability will depend on the method used
in the specific vapor dispersion model. Some models allow a choice of
methods.
6.2.5 Meteorological Tow»r«
Several types of meteorological towers are commercially available for
mounting wind and temperature instrumentation. Free-standing, guyed towers,
as well as towers that can be attached to fixed structures, are available from
a number of vendors. In some cases, meteorological instruments can be mounted
on an existing elevated structure. Lightning protection kits, available from
tower vendors, are recommended for areas where cloud-to-ground lightning is
common.
6.2.6 Data Acquisition
The most practical type of data acquisition system for the meteorological
applications discussed in this manual is a digital system that uses magnetic
media such as diskettes, cassettes, bubble memory, and semiconductor memory.
Such systems allow immediate retrieval of data during actual emergency situa-
tions and provide the convenience of data retrieval and processing at any
time. Many types of digital data acquisition systems with a wide range of
memory capability are commercially available. System needs will have to be
determined case by case.
6.3 APPLICABILITY AND PERFORMANCE
The utility of on-site meteorological data for hazardous release applica-
tions depends on the careful design of the meteorological data acquisition
systems. Important considerations, discussed in the following sections, are
variable selection, instrumentation selection, and instrumentation siting.
92
-------
6.3.1 Instrument Selection
A number of wind vane manufacturers offer wind sensors with a rated
accuracy of +3 degrees, which meets EPA recommendations for prevention of
significant deterioration (PSD) monitoring in standard air pollution appli-
cations (A), and which should be adequate for hazardous release applications.
As with all meteorological sensors, the accuracy is actually the accuracy of
the entire measurement system, including the readout device. Additional
information on important wind vane performance criteria may be found in
References 5, 6, and 7.
A wind socle is often used as a visual indication of wind direction.
Although wind socks are rarely used for remotely sensing wind direction,
are extremely valuable to facility personnel as quick visual indicators,
should be used wherever an accidental releAse could present a hazard to
facility personnel or to the surrounding community.
Many manufacturers offer wind speed sensors with a rated accuracy of
±0.82 feet per second or better, which meets EPA guidelines. As in the case
of wind direction, the accuracy guidelines apply to the entire measurement
system. Additional information on wind speed performance criteria may be
found in Reference 8.
A wind sock may be used as a visual guide for estimating wind speed.
Host manufacturers offer guidelines as to how the deflection of the wind sock
correlates with wind speed. As mentioned in the previous discussion, wind
socks should be used wherever an accidental release could present a hazard to
the facility personnel or the surrounding community. Even when more sophisti-
cated instruments are available, the wind sock should be used as a backup
visual indicator.
they
They
93
-------
Temperature sensor accuracies of 0.9*F or better are easily achievable
(9). A number of manufacturers offer sensors with much better accuracy,
although an accuracy of 0.9*F should be adequate for most accidental release
applications (an exception is the case in which delta temperature measurements
are needed for stability calculations; in this case, an accuracy of 0.18'F or
better would be appropriate). It is important to note that temperature sensor
accuracy is often limited by poor sensor exposure, improper coupling, and
signal interference, rather than by the accuracy of the sensor itself. A
serious problem in temperature measurements is radiation error, which can
amount to several degrees at midday. Aspirated temperature radiation shields,
readily available from manufacturers, can minimize radiation error.
6.3.2 Siting of Meteorological Instrunmntaclon
The objective of proper meteorological instrumentation siting is to place
the instruments at locations where measured data represent the atmospheric
conditions in the area of interest. Proper instrumentation siting is of
special importance for accidental release applications, since the data may be
used for making critical decisions that affect human lives and property.
This section discusses instrumentation siting criteria, including the
number and location of measurement sites, distances from obstructions,
measurement heights and topographical considerations. The discussion focuses
on the variables of wind direction, wind speed, and ambient temperature.
Siting considerations for wind direction and speed receive the greatest
attention, since data for these parameters are most sensitive to siting
factors.
It is often difficult to locate a site that meets all the desired siting
criteria. In any instance, site accessibility and security should never be
overriding considerations; the integrity of the data may be significantly
compromised if factors such as obstructions and topography are given too low a
priority.
94
-------
Siting for Wind Direction and Speed--
For accidental release applications, it is of utmost importance that
reliable wind data be measured, since wind data (including stability calcu-
lated from wind data) determine the path and spread of the toxic cloud. The
reliability of the wind data depends on Che appropriateness of wind sensor
siting.
Usually, the more wind data sites available, the more reliable the
modeled estimates of accidental release transport. Some vapor dispersion
models take advantage of multiple sites by weighting wind values according to
the locations of the wind sites relative to the position of the plume at a
given time.
The number and location of wind data measurement sites should be based on
consideration of probable paths of cloud transport, the demographics of the
facility vicinity, and expected variations in wind speed and direction caused
by obstructions and non-uniform topography.
Using historical wind frequency distributions, the meteorological network
designer should determine the relative probability that populations or
properties in various directions from the facility will be located in the path
of a hazardous cloud release so that the possible number of wind data sites
and their locations can be determined. Wind data sites should be located as
close as possible to the facility property boundary line, since data from a
property line location should provide the best estimate of cloud transport
into adjacent areas once the cloud has cleared the obstructions created by
facility structures.
Once the number and general locations of meteorological sites have been
determined, the network designer must analyze obstructions and topography to
find specific sites that will produce data representative of areas potentially
affected by a hazardous release.
95
-------
It is standard meteorological practice to locate wind direction and wind
speed sensors where the distance between the instruments and any obstruction
is at least ten times the height of the obstruction (2). Examples of common
obstructions are industrial facility structures, buildings, and large trees.
Sites that are too close to obstructions will normally record wind speeds that
are too low, wind directions that are unrepresentative of the area, and wind
direction fluctuations that are too great (as indicated by the standard
deviation of the wind direction).
While the adverse effects of obstructions on wind instrumentation siting
can be determined and avoided rather easily in many cases, it is a significant
challenge, at least in complex terrain, to assess the effects of topography on
wind flow and to locate wind sensors where measured data are truly representa-
tive of the areas of interest. The meteorological network designer should
carefully review topographical maps and determine the potential of the local
topography (hills, creek beds, valleys, canyons) to cause channeling and
upslope and downslope flows. The network designer must then locate the wind
instrumentation sites so that the measured data are representative of the
areas that may be affected by a hazardous release. When there are no signifi-
cant obstructions to the flow of wind and the topography is level, wind
measurements at any one site in the facility area should provide representa-
tive wind data.
The standard height for surface wind measurements in scientific
applications is 33 feet (2). An appropriate wind measurement height for
accidental release applications, however, will depend on the expected height
of potential releases. For example, in the case of a non-buoyant toxic cloud
that may flow near the ground, a wind measurement height in the range of 7 to
17 feet may be more appropriate than one of 33 feet. If a toxic cloud could
be released at an elevated point, wind measurements at that height would be
appropriate, especially during night-time periods when wind directions can
vary significantly with height. When a facility contains multiple potential
96
-------
toxic release points at various heights, selection of an optimal wind
measurement height may be difficult. In such a case, a measurement height
that reflects the average of the release heights may be used.
Siting for Ambient Temperature--
The calculations made by hazardous release dispersion models of the
dispersion of the released cloud are based on the ambient temperature at the
time of release, and they assume that the temperature remains constant along
the path of transport. It is visually sufficient to measure the ambient
temperature at one location in the site vicinity.
The standard ambient temperature measurement height is 3 to 7 feet (2).
For accidental release applications, however, the measurement height should be
adjusted to correspond with the height of potential toxic releases. Such a
height adjustment is most important during nighttime periods when variations
in temperature with'height can be significant. In case there are multiple
potential toxic release points located at various heights, it is most
practical to select a single height that best represents the heights of the
release points.
Ambient temperature sensors should be sited so that they are not unduly
influenced by heat sources such as exhausts. Locations directly over concrete
or asphalt should also be avoided (2).
6.4 RELIABILITY
Meteorological instrumentation can be highly reliable if proper
preventive maintenance procedures, as prescribed by the manufacturers, are
followed. However, because of the importance of meteorological measurements
for accidental release applications, spare parts and sansors should always be
readily available.
97
-------
For a meteorological system to produce valid and useful data, a well-
designed program of quality assurance/quality control must be implemented and
maintained. A stringent quality assurance/quality control program is particu-
larly important for accidental release applications, since decisions affecting
human health and the environment will frequently be based on meteorological
data.
Examples of quality control measures include calibration checks, preven-
tive maintenance, and spot data checks by personnel directly involved in
operating the instrumentation. Quality assurance activities include equipment
audits and detailed screening of the data sets by parties who are independent
of routine system operations. A comprehensive set of quality
assurance/quality control procedures has been developed by the U.S. EPA (2).
These procedures are appropriate for meteorological systems deployed for
accidental release applications.
6.5 COSTS
The cost of equipment and accessories for a station measuring wind speed,
wind direction, and temperature typically ranges from $10,000 to $20,000. A
number of vendors offer a variety of equipment models and configurations.
Components that should be considered when projecting costs for
meteorological systems include the sensors; translator modules, racks, and
power supply; junction boxes; cabling; wind sensor mounting crossarms; surge
and lightning protection; data acquisition system; telemetry devices;
calibration equipment; spare parts; and equipment shelter.
98
-------
6.6 REFERENCES
1. Boykin, Raymond F.f and Mardyros Kazarians. Quantitative Risk Assessment
for Chemical Operations. Presented at International Symposium on
Preventing Major Chemical Accidents. Pickard, Lowe, and Garrick, Inc.,
Washington, D.C., 1987.
2. U.S. Environmental Protection Agency. Quality Assurance Handbook for Air
Pollution Measurement Systems, Volume IV: Meteorological Measurements.
EPA-600/4-82-060, February 1983.
3. U.S. Environmental Protection Agency. Guideline on Air Quality Models
(Revised). EPA-450/2-78-027R (NTIS PB86-245248), July 1986.
4. U.S. Environmental Protection Agency. Ambient Monitoring Guidelines for
Prevention of Significant Deterioration (PSD). EPA-450/4-80-012 (NTIS
PB81-153231), November 1980.
5. Finkelstein, Peter L. Measuring the Dynamic Performance of Wind Vanes.
Journal of Applied Meteorology, 20, pp. 588-594, 1981.
6. MacCready, P.B. Dynamic Response Characteristics of Meteorological
Sensors. Bulletin of American Meteorological Society, 46 (9), pp.
533-538, 1965.
7. Wierinza, J. Evaluation and Design of Wind Vanes. Journal of Applied
Meteorology, 6(G), pp. 1114-1122, 1967.
8. Lockhart, T.J. Bivanes and Direct Turbulence Sensors. MRI 70 Pa-928,
EPA Institute for Air Pollution Training, 1970.
9. U.S. Environmental Protection Agency. Strimaitis, David, G. Doffngole,
and A. Bass. On-Site Meteorological Instrumentation Requirements to
Characterize Diffusion from Point Sources. EPA-600/9-81-020 (NTIS
PB81-247223), C-l-C-5, 1981.
99
-------
SECTION 7
SECONDARY CONTAINMENT
In Che went of Che accidental release of a volatile, toxic and/or flamm-
able liquid, che spread of the liquid must be concrolled Co reduce the rate of
evaporation and the resulting generation of a vapor cloud. Physical barriers
provided by Judicious siting of storage and process facilities can contain and
thus reduce the effects of a hazardous release. However, such barriers often
do not exist or do not provide sufficient containment. In such cases, other
methods of secondary containment are needed. Stopping or reducing the flow of
a released chemical at the source will also help contain an accidental
release. Sometimes the flow of material can be stopped by closing a valve of
stopping a pump. At other times stopping the flow of released material
requires the Application of a leak plugging procedure.
This section of the manual discusses several methods for plugging leaks
and stopping the flow of material upstream of a leak. A similar discussion
and evaluation of several secondary or partial containment systems is also
included. A brief discussion of costs is presented.
7.1 BACKGROUND
Post-release mitigation measures attempt to reduce the effects of an
accidental release. The most effective method for reducing the effect of a
release is to stop the release before large quantities of material have
escaped. Using strategically placed, remotely operated emergency isolation
valves is the soit effective way of stopping the flow of material from an
accidental release; however, in some cases, it is not possible to use valves
to isolate a leaking portion of a process. An exu.pl. would be a release from
a hole in a large vessel. Such a release would continue even if che tank were
100
-------
isolated, as long as the liquid level was above the leak point or the internal
pressure was sufficient to force vapors out. In such cases, leak plugging
techniques night stop the flow of material.
Leak plugging specifically applied to accidental releases is not a common
practice, although the Chlorine Institute has designed a number of leak-
plugging devices for chlorine cylinders, one ton containers, and tank cars and
tank trucks used to transport chlorine (1,2,3). In the chemical process
industries, the desire to perform repairs without shutting down a process has
led to the development of several plugging methods. Although these methods
would have limited applicability to accidental release mitigation, the
principles may be useful. The use of various chemical sealing materials for
plugging chemical leak* has been investigated experimentally; none is in wide-
spread commercial use. Such methods could be potentially applied to a number
of accidental release situations. The following section discusses some of the
leak plugging methods that may apply to mitigating releases.
In addition to efforts to stop the flow of the release at the source, the
spread of any released liquid must be controlled. Limiting the spread of
released liquids will limit the ground surface area affected by the release
and will allow emergency response personnel to work close to the release
source. In addition, containing the release will limit the surface area of
the liquid, which will liait the rate of vapor generation, which can
substantially reduce downwind vapor concentrations of volatile liquids.
Finally, containment can be used to divert released liquids into a temporary
containment vessel, thus substantially reducing the impact of the release.
Secondary containment systems in the form of curbing, trenches, basins
and dikes have been successfully used for many years In the chemical Industry.
Many improvements have been made over the years. The following section
discusses some of the secondary containment systems available at this time.
101
-------
7.2 DESCRIPTION
7.2.1 Leak-Plugging Techniques
Leak-plugging techniques may be divided into three groups: chemical
patches and plugs, physical patches and plugs, and methods for stopping the
flow upstream of a leak. Examples of each of these categories are discussed
below.
Chemical Patches and Plugs--
A chemical patch or plug is made by filling a hole or crack with a
material that solidifies in place after application. No chemical patching
system has been developed for widespread commercial use. The concept has been
experimentally tested and requires further developmental work. Three
categories of materials have been tested for this application (4). Urethane
foams conform to an irregularly shaped hole and expand in place, resulting in
a tight seal. Nonexpanding materials react in place to achieve crosslinked
structures which are relatively resistant to chemical attack. Examples of
crosslinked materials include: various epoxy systems, quick-curing silicone
rubber, and catalyzed polysulfide rubbers. The third category of materials
that have been tested are thermosetting nonexpanding plastics (5,6,7).
Several different methods for applying chemical plugs and patches have
been tested. In one method, the patching material is forced into the leak and
held in place until it sets and forms a plug (4). A second technique has been
tested for improving the ability of nonexpanding materials to plug a hole. In
this method, the chemical patching material is used in conjunction with a
supporting or backing material such as nylon cloth (4). Another application
technique uses an elastomer "balloon" or cloth sack attached to the end of a
lance. The lance is inserted into the leak and the balloon or sack is filled
with a plugging chemical, thereby plugging the leak (4,8,9). This type of
device is illustrated in Figure 7-1. A variation of this last type of device
is a cloth bag that is wrapped loosely around a leaking valve or fitting with
102
-------
SPONGE FIRMLY
• ATTACHED TO TUBE
AT BOTH ENOS
ELASTOMER COATING
OR CLOTH BAG
TUBE ENO IS CLOSED
TUBING
OPENINGS IN TUBE
TO ALLOW FOAM
TO ENTER SPONGE
\.f
¦ FOAM INLET
OPEN - CELL
POLYURETHANE
SPONGE
Figure 7-1. Foam-injected leak plugging device.
Source: Adapted from Reference 10.
103
-------
a leak-plugging chemical injected into Che annular space between the cloth and
the leak. The cloth holds the chemical in place around the leaking fitting
until it can harden in place.
Physical Patches and Plugs--
A physical patch or plug is any prefabricated device vised to fit into or
around a leak; a multitude of such devices are available. The Chlorine
Institute has assembled three emergency kits used to temporarily stop leaks
from valves and fittings on cylinders, one-ton containers, and tank cars and
tank trucks (1,2,3). Figure 7-2 illustrates some of these patching and
plugging devices. The kits have patching devices for small punctures of
cylinders and one-ton containers. This type of leak repair equipment is
available from numerous other sources.
Several physical patching devices have been developed for leaking pipes,
flanges or other fittings, so that pipe repairs can be made without shutting
down the process. These devices, illustrated in Figure 7-3, are usually used
to repair leaks of nontoxic materials such as steam, but they may be useful
for preventing the release of toxic materials (6). This type of leak-plugging
equipment is usually bolted or clamped around the leaking pipe or fitting,
forming a leak-tight case. It is usually sized for a specific type of fitting
or pipe. These devices are often used in conjunction with a chemical sealing
material. For example, a metal jacket is bolted around a leaking pipe or
fitting and a chemical sealing compound is injected into the annular space to
more fully seal the leak (11). This type of arrangement is illustrated in
Figure 7-4.
Methods for Stopping Flow Upstream of a Leak--
Analogous to leak-plugging methods are techniques developed for stopping
Che flow of material upstream from a leak. The most practical method for
stopping flow would be shutting an emergency isolation valve. However, such
valves are not always available. A few methods have been developed for
stopping the flow of material in a pipe so that repairs can be made without
104
-------
TWHTewe tmw acww weasta we
MTO. HOOO OOMM ANO *OMW * OM TWHT
eOMR MM TM LtMONO ANOU VM.VS.
OMnrwrnuwrna
HOOO OONtKIS TM«
laak topalr Davie* for '—"-g
Cmr m CUorlaa Taak Car
<«jl« Till*
HaMTWM 1MM KMN MM TU ~«w
'Mu»m Txa cwfBuwaw wnu. ano raw *
CMflHM 1WKT M*k MOUNO A M«U UMt
Martiaalcal Laak lapair Oa*lca far Oaa Ton Chlorine » -—
With SaaU Slda HaU Laak waar
Figure 7-2. Examples of mechanical leak repairing devices offered
by the Chlorine Institute.
Source: Adapted from References 2 and 3.
105
-------
gasket between plates sealing ring
Figure 7-3. Jacket chat bolts around a leaking pipe,
forming a leak-tight seal.
106
-------
-4. Mechanical scaling ring wich injected chemical
sealant to fit around and seal a leaking flange.
Adapted froa Reference 12.
107
-------
shutting down a process. These techniques are probably not useful for most
accidental release situations. However, they are discussed here to illustrate
what Che state of the art is at this time.
Pipe freezing, one method for stopping the flow in a pipe, has been
developed primarily as a maintenance tool for making repairs in a pipe without
shutting down an entire process unit (13). A temporary jacket is placed
around the pipe where the freeze is to be made. The cooling liquid, such as
liquid carbon dioxide or liquid nitrogen, is injected into the jacket and
replenished until a solid, frozen plug is formed in the pipe. The plug is
maintained by monitoring the temperature in the jacket and replenishing the
coolant as required.
One company has developed a device for high-pressure steam applications
capable of shearing a pipe and sealing the shear in place, thereby stopping
the flow of material through the pipe (12). A similar method, called
"stoppling," has been used for stopping flow in a line. In this technique, a
reinforcing saddle is first welded or bolted around the line. A hole is
drilled into the side of the line and a plug is inserted to stop flow in the
line. The entire operation is kept tinder sufficient pressure to prevent any
liquid in the pipe from escaping.
7.2.2 Physical Barriers and Containment Systems
Secondary containment systems typically consist of the following:
curbing, trenches, excavated basins, natural basins, and earth, steel, or
concrete dikes. The type of containment system best suited for a particular
facility depends on the consequences associated with an accidental release
from that location. The inventory of toxic material and its proximity to
other portions of the plant and to the community are primary considerations in
the selection. The secondary containment system should be able to contain
spills with a minimum of damage to the facility and its surroundings and with
minimum potential for escalation of the event.
108
-------
In general, secondary containment systems:
• Prevent the spread of hazardous liquid from the immediate spill area
to nearby areas;
• Limit the size of the pool of spilled liquid;
• Reduce the rate and quantity of vapor generated;
• Prevent hazardous vapor concentrations from spreading to areas
outside the facility boundary; and
• Prevent the material from reaching a source of ignition, if it is
flaomable.
The most common type of secondary containment for storage and process
vessels is a low-wall dike (14). A dike, which is a physical barrier around
the perimeter of process equipment or storage vessels, is designed to confine
the spread of liquid spills. A dike can be little more than a curb or a high
wall rising to the top of a storage tank. It can be constructed of earth,
steel, or concrete. Figure 7-5 shows a typical use of a dike.
The area contained by the dike is usually sufficient to contain the
largest tank or process unit served. Generally, a limited number of process
units or storage tanks are enclosed within one diked area; as the number of
units increases, so also does the risk of an accidental release.
For most applications, dike heights range from three feet to twelve feet,
depending on the area available to achieve the required volumetric capacity
(15). For a specific capacity, a low-wall dike requires the greatest amount
of land area but is usually the least expensive to construct. On the other
hand, higher wall dikes may cost more but they may also result in some vapor
holdup and contribute to a reduction in the size of the vapor cloud (5).
109
-------
Figure 7-5. Conceptual diagram of a typical low-wall dike.
110
-------
Dike walls are designed Co be liquid CighC and Co viChsCand Che
hydrosCaCic pressure and temperature of Che spill. Dikes are Cypically
locaCed a sufficienC distance from the tank to allow access to the vessel and
to equipment near the vessel. Piping is routed over dike walls; penetrations
through walls are avoided if possible. Vapor fences may be situated on the
top of the dikes to give additional vapor retention capacity. In addition, if
there is more Chan one Cank or process vessel in the diked area, the vessels
are often elevated above the maximum level attainable in the impoundment if
more than one tank or vessel loses its entire contents. Common diking of
incompatible materials should be avoided.
Provision should be made for draining rainwater from diked areas. This
will involve the use of sumps and separate drainage pumps, since direct
drainage to stormwater sewers would allow any spilled material to follow the
same route. Alternatively, a sloped rain hood may be used over the diked area
that could also serve to direct the rising vapors to a single release point.
In areas where it is critical to minimize vapor generation, surface
insulation may be used in the diked area to further reduce heat transfer from
the floor and walls or the dike to the spilled liquid. In addition, earthwork
dikes may require a layer of clay, asphalt, plastic film, or similar material
to prevent the contained material from seeping into the ground (4).
The ground enclosed within a dike should be graded to cause the spilled
liquid to accumulate at one side or in one corner. This will help minimize
the area to which the liquid is exposed and from which it may gain heat.
High-wall dikes may be an appropriate secondary containment choice for
selected systems (8). High-wall dikes may be constructed of low-temperature
steel, reinforced concrete, or prestressed concrete. A weather shield may be
provided between the tank and wall, with the annular space remaining open to
the atmosphere. The available area surrounding a storage tank will dictate
the minimum height of the wall. High-wall dikes may be designed with a .
volumetric capacity greater than that of the tank or vessel to provide vapor
111
-------
containment. Increasing the height of the wall also raises the elevation of
any released vapor.
As with low-wall dikes, piping should be routed over the wall if possi-
ble. The closeness of the wall to the tank may necessitate placement of the
pump outside the wall, in which case the outlet (pump suction) line will have
to pass through the wall. In such a situation, a low dike encompassing the
pipe penetration and pump may be provided, or a low dike may be placed around
the entire wall.
An alternative, to simply accumulating the material in one corner of the
diked area is to provide a system such as the one depicted in Figure 7-6,
which consists of a concrete or steel basin located at the lower edge of the
dike. A one-way valve allows the toxic material to flow into but not out of
the basin. A submerged pump is used to remove the material from the basin,
and it is either sent to storage or to an appropriate treatment and disposal
system. In addition, the vapor generated from the spilled liquid could be
sent to a scrubber or flare. Such basins may be required to conform to
regulations for underground storage of toxic chemicals.
A remote basin is another type of secondary containment. It is espe-
cially well suited to storage systems where more than one tank or process
vessel is served and a relatively large site is available. The flow of a
liquid spill is directed to the basin by dikes and channels under the storage
tanks. These channels are designed to minimize exposure of the liquid to
other tanks and surrounding facilities. Often the channels that lead to the
remote basin, as well as to the basin itself, are covered to reduce the rate
of evaporation of the spilled liquid. For nonflammable materials, the
impounding basin is usually located near the tank to minimize the amount of
toxic material that evaporates as it travels to the basin. For flammable
materials, two conflicting spacing requirements must be considered. First,
the impounding basin must be close to the tank to minimize evaporation.
Second, the impounding basin must be far enough from the tank that a fire in
the basin would not expose the tank to excessive heat.
112
-------
Figure 7-6. Conceptual diked storage facility with integrated containment sun,p
-------
A neutralization basin is an alternative to an impounding basin. This
system is essentially the same as the impounding basin, except that the
spilled material is mixed with a neutralizing material to form a mixture that
can then be disposed of safely (9). For acidic materials, an appropriate
material might be lime, which could be stored in the basin as a premixed
slurry or as solid limestone. However, a high heat of reaction may cause an
increased rate of evaporation with some materials. For example, Figure 7-7
illustrates two layouts for a neutralization basin.
A final type of secondary containment is one structurally integrated with
the primary system to form a vapor-tight enclosure around the primary
container (15). Many arrangements are possible. A double-walled tank, as
illustrated in Figure 7-8, is an example of such an enclosure. A low-wall
dike placed around the entire systems may be used as a backup, in the event of
wall or foundation failure.
7.3 APPLICABILITY AND PERFORMANCE
7.3.1 Applicability and Performance of Leak-Plugging Devices
The leak plugging and patching techniques listed in this section cannot
be used for catastrophic failures of equipment. They perform best on small
leaks and most will work better on circular holes than on cracks. Four other
limitations common to most of these techniques are: first, personnel must be
able to get close enough to the leak to apply the patch or plug; second, it is
difficult to prevent the person applying the leak from being sprayed with the
leaking material; third, all of these techniques have limited capabilities:
no appropriate leak-plugging technique may be available at the time of a
specific release situation; finally, all of these techniques require time to
set up and implement, and a large amount of hazardous material may escape
before the leak can be stopped. It could be very hazardous to attempt to plug
the flow of a flammable material. Plugging such a leak would probably require
personnel to enter a cloud of material within the flammable range.
114
-------
CONCRETE DIKES ANO MO PARTIALLY LIMESTONE FILLED REMOTE
FILLED WITH LIMESTONE UNOER NEUTRALIZATION BASIN
STORAGE TANK (COULO EE COVERED)
LIMESTONE NEUTRALIZATION BASIN
CONCRETE DIKES AND MO
UNOCR STORAGE TANK
UMC. WATER SLURRY RIUCO
REMOTE NEUTRALIZATION BASIN
(COULO BE COVEREO)
DRAIN UNI
UME WATER NEUTRALIZATION SYSTEM
Figure 7-7. Conceptual diagram of potential layouts for a remote
neutralization basin system.
Source: Adapted from Reference 9.
115
-------
SUSPENDED DECK
WITH INSULATION "
DOUBLE WALL
STEEL TANK
INSULATION-v.
REINFORCED
CONCRETE WALL
OUTSIDE V
EARTHEN FILL iyl
(MAY 8E OMITTED) II
/VX<€v <=* J
p J " 1—i
TANK
FOUNDATION"
LOAD BEARING
INSULATION
Figure 7-8. Conceptual diagram of & double-vallfid tank system.
Source: Adapted from .Reference 15.
116
-------
Where hazardous materials are involved, a patch or plug should never be
regarded as a permanent form of repair. The purpose of leak plugging is to
stop the leak just long enough to shut down the process, drain the storage
tank or effect whatever changes are necessary to stop the flow of material up-
stream of the release.
Chemical Patches and Plugs--
A limited amount of experimental testing has been done to measure the
performance of chemical patches and plugs. One report prepared by the
Environmental Protection Agency summarized the results of leak-plugging tests
on a variety of expanding foams and nonexpanding sealants (5). In these
tests, the plugging chemical was pushed into the leak and held in place until
it hardened. Table 7-1 shows some of the test results. The EPA report also
calculated a practical applicability range where chemical plugs might be used;
Figure 7-9 shows this range.
Chemical patches and plugs have several advantages over alternate
plugging techniques: the plug can conform to irregularly shaped holes and
cracks; the plug can be applied to holes in irregularly shaped surfaces; and
no elaborate application devices are required. One disadvantage of chemical
plugs is that the chemical plugging material must be resistant to attack by
the chemical that is escaping; a facility may need to have several different
plugging materials on hand. In addition, a chemical plug will be subject to
failure by chemical attack or physical or thermal shock.
Several studies have tested the performance of applicators that use a
foam-filled elastomer or fabric bag for leak plugging (6,7). Using such a
device increases the physical strength of the plug. Also, since a narrow
lance is first inserted into the hole, less spray of leaking material may
occur than with other methods, thus reducing the risk of exposure to
personnel. The disadvantages of using this type of applicator are the same as
those associated with the use of chemical patches without an applicator. The
117
-------
TABLE 7-1. HAZARDOUS CHEMICALS TESTED AND SMALL-SCALE SEALING TEST RESULTS
Results of Sealing Tests
Hazardous Material
Starfoam® Urethane
Foam
Sea-Goin$
Epoxy Putty
Other Successful Sealants
Phenol
Methyl Alcohol
Insecticides, Rodenticides:
DDT 95* Solu./Water
Dieldrin
Acrylonitrile
Chlorosulfonic Acid
Benzene
Phosphorus Pentasulfide
Styrene
Acetone Cyanohydrin
Nonyl Phenol
Isoprene
Xylene
Nitrophenol
Sealed (A)+*
Sealed (A,U)
Sealed (A)
Failed (A)
Sealed (A,U)
Sealed (A)
Sealed (A,W)
Sealed (A)
Sealed (U)
Failed (A)
Sealed (A,U)
Sealed (ff)
Sealed (A)
Failed (A)
Failed (A)
Sealed (A,W)
Failed (A)
Sealed (A)
Sealed (A)
Sealed (A,W)
Sealed (A)
Sealed (A)
Sealed (W)
Butyl Rubber (A)
Epoxy Putty (A)
MSA Urethane (A)
Polysulfide Rubber (A)
MSA Urethane (A,W)
^Sealed leaks of aqueous solutions (50 to 90 percent) for 18 hours, then failed.
"A" and "W" indicate tests performed in air.and submerged in water, respectively.
Source: Reference 5
-------
Figure 7-9. Envelope of practical applications for sealing
leaks in nonsubmerged containers.
Source: Reference 5,
119
-------
applicator. The plug is still subject to failure by chemical attack or
thermal or physical shock.
Physical Patches and Plugs--
The primary advantage of physical patches and plugs is their ability to
form a very tight seal nearly as strong as the original container. They can
be constructed of material more resistant to chemical attack than a chemical
plug.
The primary disadvantage of physical leak-repairing devices is that most
are designed for a very specific use. For example, a leak from a flange can
be stopped by using a specially designed encasement constructed to form a
tight seal around the entire flange; this is a very effective method for
stopping leaks. However, such an encasement will work only for flanges and
often a different encasing device will be required for each size of flange. A
facility would need a large inventory of physical plugging devices to be able
to handle any accidental release that might occur; this is an impractical
solution for most situations.
Physical patches and plugs are most useful where only a few types of
leaks are possible, or where only a few types of leaks present the greatest
potential hazard. For example, the Chlorine Institute has assembled leak
repair kits for several transportable chlorine storage containers. The kits
contain leak-repairing devices for all the types of leaks most likely to occur
in each of the containers (5 to 9 different types of leaks).
Leak-plugging methods where a. physical jacket or shell is filled with a
chemical plugging material have some advantages over other types of devices.
The outer shell provides high physical strength for the plug. Also, the use
of the chemical sealing material, means that the shell does not have to form a
tight seal around the leaking equipment; therefore, one type of shell could be
used for plugging leaks from several line sizes or for several types of
fittings.
120
-------
One disadvantage of this method is that the chemical plugging material
must be chemically resistant to the leaking material and must be thermally
resistant to the temperature of the leaking material. These types of devices
are used for plugging leaks from pipes or other small diameter fittings by
enclosing the entire line. They are not used for filling a leak, and thus are
not practical for leaks from large diameter vessels.
Methods for Stopping Flow Upstream of a Leak--
The main advantage of any method that stops the flow of material upstream
of a leak is that it reduces the potential of personnel exposure. However,
all of these methods have severe limitations that restrict their usefulness in
the event of an accidental release.
Some technical literature addressing the use of pipe freezing for
plugging lines has examined the utility of pipe freezing for maintenance
purposes (15). Many organic materials cannot form a frozen plug with the
strength sufficient to stop flow. Many materials will require extremely low
temperatures to create a solid plug. Even if a material can form a physically
strong plug at an achievable temperature, the method would probably require
too much time to implement in the event of an accidental release.
Stoppling and line shearing and sealing techniques have the advantage of
forming a mechanically tight seal. These are proven methods that have a low
potential for exposing personnel to the leaking material. The disadvantage of
the methods is that they require elaborate and expensive equipment. Usually,
a company that specializes in this work is contracted to plug a line. Such an
arrangement would not be practical in the event of an accidental release
because of the time required for the plugging equipment to be brought to the
site.
121
-------
7.3.2 Applicability and Performance of Physical Barriers and Containment
Systems
Secondary containment systems can minimize the spread of a liquid spill
and/or the rate of toxic vapor generated from an accidental release. However,
the type of containment system best suited for a particular storage or process
unit will depend on the material being contained and on the risk associated
with an accidental release from that location. This subsection discusses the
factors that must be examined when selecting a secondary containment system.
In general, it is desirable to use some form of secondary containment
around storage facilities to prevent the unconfined spread of spilled liquid.
It is also beneficial to slope the surfaces away from storage tanks to a catch
basin so that liquid does not accumulate in the dike but is still confined to
an area near the dike. This is especially important if the material is toxic
and flammable, since an ignited pool of liquid beneath a vessel could lead to
a BLEVE (boiling liquid expanding vapor explosion).
In process areas, curbing around accumulators, reactors, or other vessels
containing large amounts of liquid serves to prevent the unconfined spread of
spilled liquid. For piping systems, the question of whether or not to provide
secondary containment depends partly on the pressure and temperature of the
material in the pipes and on the location of the piping system. Curbing or
trenches may be used along long pipe runs if a break could result in a large
accumulation of liquid. However, in such situations the curb or dike should
be designed to drain liquid away from the pipe. For a long pipe carrying
flammable materials this will reduce the potential for small leaks to result
in the accumulation of flammable materials around the pipe. Transfer areas
represent high-hazard areas because of the failure of temporary connections
used during transfer. Spills may occur frequently and a method of secondary
containment may be warranted.
122
-------
Each situation must be specifically examined. The secondary containment
system should at least be able to contain spills with a minimum of damage to
the facility and its surroundings and with a minimum potential for escalating
the event.
The applicability of diking to spills of volatile liquids is readily
apparent. By containing the liquid, the dike reduces the surface area
available for evaporation, at the same time allowing a liquid to be cooled by
evaporation so that the vapor release is diminished. In this way, diking can
reduce the rate at which a toxic material is released to the air. The
material can be allowed to evaporate at a manageable rate, collected into
alternate containers, or neutralized in place. A dike with a vapor fence will
give extra protection from the wind and will be even more effective at
reducing the rate of evaporation.
A dike will not reduce the impact of a gas or vapor release, A. dike also
creates the potential for toxic material and trapped water to mix, and,
depending on the material, this nay accelerate the rate of evaporation or the
formation of highly corrosive solutions. If materials that would react
violently with the released toxic material are stored within the same diked
area, the dike will increase the potential for mixing the materials in the
event of a simultaneous leak. A dike also limits access to the tank during a
spill.
High-wall diking, on the other hand, can be used to minimize vapor
generation rates and to protect the tank or vessel from external hazards such
as missiles from nearby explosions. Maximum vapor generation rates will
generally be lower for high-wall dikes than for low-wall dikes or remote
impoundments because of the reduced surface contact area. Insulation on the
wall and floor within the annular space can further reduce these rates.
One disadvantage of high-wall impoundments is that the high walls around
a tank or vessel may hinder routine external observation. Furthermore, the
closer the wall is to the tank or vessel, the mora difficult it becomes to
reach the tank or vessel for inspection and maintenance.
123
-------
A remote basin, on the other hand, allows for removal of the spilled
liquid from the immediate tank or process area, allows access to the tank
during the spill, and reduces the probability that the spilled liquid will
damage the tank, vessel, piping, electrical equipment, pumps, or other
equipment. In addition, a covered impoundment will reduce the rate of
evaporation from the spill by protecting it from wind or from heating by
sunlight.
A limitation of a remote basin is that the potential still exists for
water or other incompatible materials to be trapped in the impoundment and mix
with the spilled material. Additionally, remote basins do not reduce the
effects of a gaseous toxic release.
The effectiveness of a secondary containment system will depend on the
specific material being contained. For some liquefied gases, such as
chlorine, a spill from a pressurized storage or process vessel results in a
significant amount of flashing and vapor generation and little liquid
accumulation (14). In such cases, conventional low-wall diking may not be
effective. All flashing gases "self-refrigerate" to some extent, and there
will likely be some liquid accumulation. Low-wall diking can be constructed
around a liquified gas tank in an area too small to contain the whole tank
volume, since some will flash off, and a reduction in vapor evolution should
be achieved. On the other hand, during releases of pressurized liquefied
gases with higher boiling points or during releases of refrigerated liquefied
gases, liquid will accumulate near the release point. In these cases, diking
will be effective. Each situation must be examined in detail to determine the
most appropriate applicable secondary containment system.
A limited amount of data is available concerning the performance of the
various types of secondary containment systems. However, dispersion modeling
can be used to illustrate the effects of diking and vapor holdup.
124
-------
For example, a leak from a vertical 10,000-gallon refrigerated storage
tank containing chlorine was modeled using the Complex Hazardous Release Model
(CHARM®) of Radian Corporation. The chlorine was stored at -40°F and one
atmosphere of pressure with a surrounding ambient temperature of 78°F. Vapor
dispersion patterns for the following scenarios were investigated:
• Spill of chlorine directly onto soil from a 2-inch hole and allowed
to spread unconfined; and
• Spill of chlorine directly onto soil from a 2-inch hole, confined in
a 25-foot diameter.
The results of the modeling are presented in Figures 7-10 and 7-11. The
results are based on a westerly wind of 10 mph and atmospheric conditions not
conducive to mixing and dispersing gas or vapor clouds. Similar results were
obtained by Eidsrick (16) for a ruptured pipe containing refrigerated propane.
Figures 7-10 and 7-11 show that confining the spill with a dike
dramatically reduces the size of the vapor cloud by limiting the spread of the
spill and reducing the surface area available for evaporation.
This example shows the potential effectiveness of a secondary containment
system. In general, the effectiveness of any system depends on the properties
of the particular chemical being contained and on the properties of, the system
itself. The secondary containment system should be able to contain spills
with as little damage to the facility and its surroundings as possible and
with a minimum potential for escalation of the event.
125
-------
gure -10. Vapor dispersion patterns for a release of chlorine
from a 2-inch hole in a 10,000 gallon refrigerated
storage tank using CHARM® model.
126
-------
o.a
alias
l
mils
1.0
miles
2
nll<
LBMENOS CHLMZNK
> as.0
» 100. wm
> t.oooc+oa pw
a - aouptcs
7-U Vapor dispersion patterns for a release of chlorine
8 " froa a 2-inch hole in a 10,000 gallon refrigerated
storage tank surrounded by a 25-ft diameter dike
using Charm® model.
127
-------
7.4 RELIABILITY
7.4.1 Reliability of Leak-Plugging Techniques
A qualitative evaluation of the reliability of a leak-plugging device can
be based on the ability of the method to securely stop the flow of material
from a leak. Reliability could also be assessed in terms of how quickly a
leak can be stopped in the event of an accidental release.
Chemical patches and plugs can be applied to a leak relatively quickly.
In most cases, some equipment will be required to generate the plugging
compound. Since this equipment may sit idle for very long periods of time,
routine inspection and maintenance of the equipment will be required to
maintain a quick response time.
Chemical patches and plugs are mechanically and chemically weak when
compared with other leak repair methods. Vibration, shock, or high pressure
can all result in the failure of a chemical plug. These plugs are also
subject to failure when exposed to high temperatures or to chemically
incompatible materials.
Mechanical patches and plugs can be applied to a leak relatively quickly.
They form seals that can sometimes have the structural strength of the
original process equipment; however, mechanical vibration, shock or improper
application can cause a mechanical leak plugging device to fail.
Methods for stopping the flow of material upstream of a leak are
characterized by long preparation and implementation times. Once in place,
however, these methods are reliable ways of stopping flow.
128
-------
7.4.2 Rp-lfability nf Physical Barriers and Containment Systems
The reliability of a secondary containment system is directly related to
how it is designed and constructed. In particular, the reliability will be
affected by the following:
• Ability to withstand the hydrostatic pressure and temperature of a
spill;
• Ability to withstand the dynamic stresses caused by fire or
explosion, in the case of flammable materials; and
• Ability to withstand damage caused by earthquakes, floods, high
winds, frost, or similar events.
Systems must be inspected and maintained regularly. Undetected corrosion
and cracks in dike walls, for example, could lead to the failure of a
secondary containment system during an accidental release.
7.5 SECONDARY HAZARDS
7.5.1 secondary ftagards of Leak-Plugging Techniques
All leak*plugging techniques can potentially expose plant personnel to
the leaking material. Often, material flowing from a leak will spray out
around the plugging device as it is being applied. Also, where flammable
vapors are involved, these personnel will probably have to enter a flammable
gas cloud when applying the plug.
Where hazardous materials are involved, a temporary leak repair should
never be treated as a permanent fix. Once the leak has been plugged, actions
should be taken to shut down the portion of the process involved with the
release so that permanent repairs can be made.
129
-------
7.5.2 Secondary Hazards of Physical Barriers and Containment Svsterns
Mixing incompatible materials is one potential secondary hazard associ-
ated with most types of secondary containment. This is a particular problem
where water-reactive materials are stored within the containment area.
Rainwater can collect in diked areas or in containment basins. Spilled
materials could mix with the water in the event of an accidental release.
Though not so likely, two incompatible chemicals could mix if they were both
stored within the same secondary containment area, and if both were released
simultaneously.
A spill from a vessel surrounded by a dike can fill up the dike to a
level that prevents access to the vessel, which could prevent repair or
emergency response crews from stopping the release at its source. Also, a
fire in this situation would expose the vessel to extreme heat, which could
result in an explosion. For both of these reasons, dikes should be designed
so that material accumulates away from vessels within the dike.
A covered collection basin can trap flammable vapors when a liquid is
involved. Fire could spread to such a basin and result in an explosion.
7.6 COSTS
Table 7-2 presents some order-of-magnitude costs for a few leak-plugging
devices. These costs are for equipment only and do not include installation.
Although the variety of types and sizes of secondary containment systems
makes it difficult to estimate costs, costs for dikes built around storage
facilities can be estimated easily. The cost of a rectangular dike can be
expressed as (17):
COST - (aD2 - *r2)F + 2(1 + a)DHW
130
-------
TABLE 7-2. COSTS OF VARIOUS LEAK PLUGGING DEVICES (1986 Dollars)
Cost
Equipment Item ($) Reference
Small transportable polyurethane 5,000 - 8,000 18
foam system
Chlorine Institute Emergency Repair 800 1
Kit for Chlorine Cylinders
Chlorine Institute Emergency Repair 900 2
Kit for Chlorine Ton Containers
Chlorine Institute Emergency Repair 1,000 3
Kit for Chlorine Tank Cars and
Tank Trucks
Stainless steel band clamp, for leaking 100
4" pipe, working pressure less than
150 psi
Carbon steel bolted sleeve, for leaking 500
4" pipe, working pressure less than
1000 psi
Clamp for sealing pinhole leaks on a 70
4" pipe, working pressure less than
1000 psi
12
18
18
231
-------
where:
a — ratio of dike length to dike width
D - dike width
r - tank radius
H — dike height
F - cost of dike floor per unit area
V - cost of dike wall per unit area.
In general, the cost of secondary containment systems will be small
compared with the costs associated with the vessels that they protect.
Besides the initial investment, additional costs are associated with the
necessary cleanup equipment needed such as pumps, for example, when a release
occurs within the containment systems.
132
-------
7.7 REFERENCES
1. The Chlorine Institute, Inc. Chlorine Institute Emergency Kit "A" for
100-lb and 150-lb. Chlorine Cylinders. Washington, D.C., 1986.
2. The Chlorine Institute, Inc. Chlorine Institute Emergency Kit "B" for
Chlorine Ton Containers. Washington, D.C., 1981.
3. The Chlorine Institute, Inc. Chlorine Institute Emergency Kit "C" for
Chlorine Tank Cars and Tank Trucks. Washington, D.C., 1981.
4. Feind, K. Reducing Vapor Loss in Ammonia Tank Spills. CEP Technical
Manual, Ammonia Plant Safety and Related Facilities. Volume 17.
American Institute of Chemical Engineers, New York, NY, 1975.
5. Mitchell, R.C. , Hamermesh, C.L., Lecce, J.V. Feasibility of Plastic Foam
Plugs for Sealing Leaking Chemical Containers. EPA-R2-73-251 (NTIS
PB 222627), U.S. Environmental Protection Agency, May 1973.
6. Vrolyk, J.J., Mitchell, R.C., Melvold, R.W. Prototype System for Plug-
ging Leaks in Ruptured Containers. EPA*600/2*76*300 (NTIS PB 267245),
U.S. Environmental Protection Agency, 1976.
7. Cook, R.L. , Melvold, R.W. Development of a Foam Plugging Device for
Hazardous Chemical Discharges. In: Proceedings of the 1980 National
Conference on Control of Hazardous Material Spills, Vanderbilt
University, Nashville, TN, 1980.
8. Hendriks, N.A. Safety Wall Systems for Ammonia Storage Protection. CEP
Technical Man""1 Ammonia Plant Safety and Related Facilities. Volume
21. American Institute of Chemical Engineers, New York, NY, 1979.
9. Benson, R. Hydrogen Fluoride Exposure - Prevention in the Operation of
HF Alkylation Plants. Industrial Medicine 13(1): 113-117, 1944.
10. Zajic, J.E., Himmelman, W.A. Highly Hazardous Materials Spills and
Emergency Planning. Marcel Dekker, Inc., New York, NY, 1978.
11. Bond J. Leak Sealing Using Compound Injection. Loss Prevention Bulle-
tin ' Issue #069. 1986. (No publisher or other identification provided
in the publication.)
12. The Pipe Line Development Company, PLIDCO. Piping Repair and Maintenance
Products. Cleveland, OH, 1987.
13. Schelling, Carter. Reducing Repair Downtime with Pipe Freezing. Plant
Engineering, July 24, 1986.
133
-------
14.
Lees, F.P. Loss Prevention in the Process Industries - Hazard
Identification, Assessment, and Control. Volumes 1 & 2. Butterworth's
London, England, 1983.
15. Aarts, J.J. and D.M. Morrison. Refrigerated Storage Tank Retainment
Walls. CEP Technical Manual, Ammonia Plant Safety and Related
Facilities. Volume 23. American Institute of Chemical Engineers, New
York, NY, 1983.
16. Eidsvik, K.J. A Model for Heavy Gas Dispersion in the Atmosphere.
Atmospheric Environment, Volume 14, 1980.
17. Johnson, D.W., and J.R. Welker. Diked-In Storage Areas, Revisited.
Chemical Engineering, July 31, 1978.
18. Batch Air Equipment, San Antonio, TX: vendor.
134
-------
SECTION 8
SPRAY, DILUTION, AND DISPERSION SYSTEMS
Water spray systems are routinely used in the chemical process industries
for a variety of fire protection purposes. However, they can also be used to
reduce the effects of toxic and/or flammable gas or vapor releases. Theoreti-
cal studies and experimental research have shown that such sprays can be
effective in aiding the dispersion and dilution of gas or vapor clouds result-
ing from an accidental release (1,2,3,4,5,6,7,8). However, many of the re-
sults are based on specific systems operating under a specific set of condi-
tions. Details concerning the overall effectiveness of these systems are
limited in scope. Additional research in this area is needed to refine the
present data.
Steam curtains are another technique for reducing the effects of toxic
and/or flammable gas/vapor releases. Steam curtains act similarly to water
sprays in that the primary dispersing mechanism is the dilution of the gas or
vapor with air. However, steam curtains provide enhanced buoyancy to the
toxic and/or flammable cloud by heating the gas or vapor passing through the
steam curtain.
This section of the manual presents a detailed discussion of the use of
water spray systems and steam curtains for the mitigation of toxic and/or
flammable gas or vapor releases. Both mobile and fixed installations are
discussed in terms of their applicability, performance capabilities,
reliability, secondary hazards, and costs. In addition, reactive spray
systems are briefly discussed.
135
-------
8.1 BACKGROUND
8.1.1 Spray Systems
A spray is defined as a dispersion of liquid droplets in a gas. Sprays
can achieve a variety of objectives, including, but not limited to, the forced
dispersion and dilution of a gas or vapor, absorption of water-soluble materi-
als, confinement of released vapor or gas in a particular area, or diversion
of the vapor or gas away from a particular area.
In mitigating toxic and/or flammable gas or vapor releases, the primary
purpose of sprays is to dilute the gas or vapor with air. This is brought
about by the entrainment action of the sprays. The momentum and energy of the
spray causes air to be pulled into the cloud and creates momentum and
turbulence in the gas or vapor, thus improving mixing and enhancing dilution.
The performance of a spray depends in part on the drop size distribution.
Fine droplets lose their momentum quickly and move large volumes of air
relatively slowly. Coarse droplets entrain less air but induce greater
velocities.
Dilution is also achieved to a lesser degree by absorption of the gas in
the liquid drops. Gases that are highly soluble in water (e.g., ammonia) can
be partially absorbed by a water spray. However, the mass transfer charac-
teristics of the system may limit the efficiency of absorption; it is diffi-
cult to create a sufficiently high liquid-to-gas ratio using sprays. A "fog"
consisting of fine water droplets often improves the mass transfer efficiency.
Adding a chemical to the spray solution that reacts with the vapor can improve
the absorption efficiency. An example would be using an alkaline spray
solution to absorb an acidic vapor release.
Finally, .pray-Indued »ralng of c#ld „.por cW> ^ ^ froa u
-------
individual water drops transfer heat to the cloud, thereby increasing both the
buoyancy of the gas and its dispersal.
In addition to dilution, liquid sprays are also useful for containment
and diversion. In some situations, the movement of vapor clouds, especially
if visible because of aerosol formation or condensation of moisture, can be
controlled by hand-held nozzles and fire monitors. A released gas cloud may
also be contained in a particular area by a series of spray nozzles around the
perimeter.
8.1.2 fftfffllB Chains
As with water sprays, the primary purpose of a steam curtain used to
mitigate toxic and/or flammable gas/vapor releases is to dilute the gas or
vapor with air. The energy of the steam causes air to be pulled into the
cloud and creates turbulence in the gas or vapor, thus improving mixing and
enhancing dilution.
Steam curtains will also heat the gas cloud and enhance its buoyancy.
Heating the cloud will decrease its density and help the cloud to rise from
the ground, which can result in decreased ground level concentrations of the
toxic vapors downwind of the release.
The following subsection describes spray systems and steam curtain
-«.«¦ in the mitigation of accidental vapor releases,
systems that may assist in cne 5 r
8.2 DESCRIPTION
-Bd attention has been focused on the use of
In recent years, increases
, curtains for the control and dispersion of toxic
aqueous sprays and steam cun»*»
and/or v.por r.l«— R""tch *cudl" h,v* r,v"l,d th" 'uch
«. .ffctlv.; h.«v.r, th. fchnoiogy U .till In eh. ..rly .t.g„ of
development.
137
-------
8.2.1 Sorav Systems
A spray system may consist of a single nozzle, such as a fire monitor, or
a series of specially designed nozzles that discharge in a predetermined spray
pattern. The type, design, and method of application of spray systems vary,
depending on the specific situation. Two systems are most common: fixed
water sprays and mobile water sprays. These systems may be used individually
or in combination. A reactive water solution is sometimes used with either a
fixed or mobile spray system.
Fixed Water Spray Systems--
Fixed water spray installations are similar to fire protection fixed pipe
deluge systems. They may be located in the open air or in enclosures where
there is a high probability of an accidental release of a toxic and/or
flammable gas/vapor or where there is a need to protect a certain area (i.e.,
a control room). These spray systems are used where there is a need for quick
application of water spray.
Fixed installations may consist of a series of spray nozzles elevated off
the ground with the spray directed downwards or located at ground level with
the spray directed upwards. Figure 8-1 is a conceptual drawing of a typical
system using downward spraying nozzles. Two facilities in England have
installed such fixed water spray barriers (9); however, they are limited to
the control of small scale leaks.
Figure 8-2 shows a system where the spray nozzles completely surround a
desired area. The advantage of a completely-surrounding system is that it
also provides a degree of containment. However, total spray barrier systems
may use excessive quantities of water since upwind sprays will have little
effect on the gas cloud. Spray barriers that totally surround a system often
are designed so that only downwind sections are activated in the event of a
138
-------
Figure 8—1« Conceptual diagram of water spray barrier using downward spraying
nozzles.
Source: Adapted from Reference 8.
-------
Figure 8-2. Conceptual diagram of fixed-spray barrier
surrounding protected area.
140
-------
release, as shown in Figure 8-3. Activation of fixed systems can be manual or
automatic (by using gas detectors and switches).
Fixed water sprays incorporated into semi-enclosed structures have also
been proposed as a possible mitigation technique, although this is still an
experimental concept (10). Figure 8-4 shows such a system. These systems
offer several potential advantages over other systems. First, it can provide
a degree of direct containment by partially enclosing the released vapors.
Second, release of the material occurs at a known location, rate, and con-
centration.
Thus, if sources with a high-release probability can be identified and
the exposed area and release direction can be defined, a properly designed
fixed water spray system can be effective in reducing the hazards of an
accidental toxic release.
Mobile Water Spray Systems--
A mobile water spray system, illustrated in Figure -8-5, consists of
hand-held firefighting nozzles and/or fire monitors. It can be used as an
alternative to a fixed water spray system or when a release has moved beyond
the reach of a fixed installation.
Mobile water sprays in the form of a "fog" have been used for absorbing
and diluting of water-soluble materials. A "fog" is defined as a water spray
having a mass median drop diameter of 0.03 in. or less (11). The fine spray
or "fog" allows for mora effective mass transfer. A technique for controlling
ammonia releases has been developed from several hundred outdoor ammonia
workshops for firefighters (11). As shown in Figure 8-5, fire hoses and fog
nozzles are used to create a "capture" zone downwind of a release by position-
ing the nozzles in designated locations. This technique was developed for a
specific type of release (i.e., an ammonia tank rupture) and thus may not
apply in all situations.
141
-------
A. CALM AND UGHT WINDS
FROM ANY DIRECTION
Activated Spray
Unacttvatad Spray
Figure 8-3. Conceptual diagram of fix«d-«pray barriar vitb activation
according to wind direction.
142
-------
LINE SPRAY
HEAD
, , conceptual diagram of water sprays Incorporated
Figure 8-0. seal-enclosure.
Source: Adapted from Reference 10.
143
-------
Figure 8-5. Typical mobile water spray system.
Source: Adapted from Reference 11.
144
-------
Mobile systems have also been used as effective barriers. In particular,
Beresford (12) has developed a system whereby coarse water spray discharging
from flat fan sprays and wide-angled spray monitors in an upward direction
completely surround a flammable and/or toxic gas leak in the form of a "chim-
ney." This technique is shown in Figure 8-6. Large amounts of air are
induced at ground level, and dilution of the gas is achieved as the sprays
push the gas out the top of the "chimney" allowing it to disperse safely.
A mobile water spray system must be capable of rapid deployment to the
source of a release. In many releases, the worst may be over by the time such
a system is operational. However, the cost of mobile systems are usually less
than fixed installations since much of the same equipment can also be used for
fire protection. In fixed installations, additional piping and nozzles may
have to be added to existing fire protection systems. Thus, for very large
installations, a mobile system may be the most cost-effective.
Reactive Spray Systems--
An alternative to a water spray system is the use of a mild aqueous
alkaline spray system (e.g., reactive spray system). Often the use of water
in absorbing and diluting a toxic gas or vapor is limited by the mass transfer
characteristics of the system. In addition, dilution of some materials with
water sprays results in the formation of highly corrosive mists. A mild
alkaline spray would both enhance absorption and act as a neutralizing agent
in the case of acid formation.
Limited data are currently available on the use of reactive spray
systems. Several major users of phosgene have experimented with ammonia-
injected water spray systems at their facilities (13). These systems consist
of an ammonia cylinder connected to a water feed line of a water spray system
via an injector system. When the spray system is activated, a valve opens
allowing ammonia, under pressure, to enter the water line and create the
alkaline solution.
145
-------
Figure 8-6. Conceptual diagram of mobile water spray
barrier and associated "chimney" effect.
146
-------
Other alkaline solutions could also be used. Several examples are sodium
carbonate, calcium hydroxide derived from slaked lime, or a weak solution of
sodium hydroxide. However, adequate supplies of these solutions would have to
be stored or a quick means of production would be required in an emergency
situation.
Additionally, the hazards associated with spraying an alkaline solution
into the air must be considered. Unless the spray is applied so that the
alkaline material remains in a confined area (e.g., a diked storage facility),
reactive sprays present an additional health hazard to plant personnel. Also,
such sprays may cause the corrosion of machinery and equipment.
In conclusion, reactive spray systems may be more efficient than water
sprays in applicable situations. However, more research is needed to define
potential applications, performance capabilities, and system design.
8.2.2 Steam Curtains
Steam curtains are fixed-pipe systems used to contain and disperse
releases of toxic and/or flammable gas or vapor. Caimey and Cude (14) and
Simpson (15,16) have described such systems in detail. A steam curtain
consists of a horizontal steam pipe with a row of small holes in the top and
mounted near the top of a wall. Figure 8-7 illustrates a typical system.
When operated, a wall of steam approximately 15 to 20 feet high is produced.
The steam pipe is designed so that all the individual jets combine to form a
continuous curtain of steam that entrains sufficient air to dilute the gas or
vapor concentration to below its toxic and/or flammable limit. Additionally,
the steam pipe is usually divided into sections that are individually supplied
with steam from a distribution main, allowing plant operators or an automated
activation system to select which sections of the steam curtain will be
activated in the event of a release.
147
-------
SOURCE OF
RELEASE
GO®'
ot *
iw
L**«
fc-1 *
-------
The steam curtain can be activated automatically or manually. However,
in practice, steam curtains are typically controlled manually by remotely
operated valves (17,18).
8.3 DESIGN
8.3.1 Sprav Systems
In recent years, much research and a number of experimental studies have
contributed to the development of several experimental design methods for
water spray systems (6,7,12,19). However, the optimum set of parameters for a
particular set of circumstances (i.e., nozzle type, size, location, number,
orientation) are not always clear, and, indeed, each situation requires a
unique set of parameters. This section of the manual focuses primarily on the
circumstances considered in designing water spray systems for mitigating toxic
and/or flammable gases and vapors. A brief discussion of several of the
proposed design methods mentioned above are also included.
Detailed procedures for the design of fixed water spray barriers have
been proposed. All of these methods use different approaches. The procedures
have been developed from limited experimental data and may not apply in all
situations, and since many factors, such as large obstructions (building,
process units, other structures), and shifting wind direction will modify the
dispersion characteristics of a gas cloud, an appropriate application and
design must be evaluated case by case. Therefore, the following discussion
merely indicates the types of procedures currently available for designing
water spray systems.
McQuaid (19) developed a method using an experimental correlation whereby
a water spray barrier can be sized by equating the total air entrained by a
barrier (a summation of the quantity of air entrained by each nozzle) ^o the
149
-------
amount of air required to dilute the gas or vapor cloud below a certain level.
The method is based on the assumption that adequate mixing requires an induced
air velocity of approximately 20 feet/second. The air flow rate and the water
flow rate at the nozzles are required. This method can give the type, number,
and position and direction of nozzles required for a given installation. This
method was derived from work carried out on the entrainment properties of
water sprays used in coal mines for auxiliary ventilation. Predicted results
were compared with test data on large-scale sprays as reported in the
literature (19).
In contrast, Moodie (7) developed a design technique based on the assump-
tion that a water spray barrier behaves as a jet in a crossflow. From this, a
characteristic length can be defined in terms of the square root of the
momentum flow rate of the barrier and a characteristic wind speed. For a
specific wind speed and specified degree of dilution, the momentum flow rate
can be determined from an experimental correlation. Using this information
and a procedure developed by Moodie (7) for nozzle selection, the barrier can
be sized in terms of its width, nozzle type, and water consumption. This
method was developed from extensive test data generated using water sprays to
mitigate carbon dioxide releases; therefore, it is based on substantial field
data.
Apart from these proposed design methods, the only other technique that
has been developed to aid in the design of fixed systems is the use of comput-
er models to predict the performance of various spray systems (20,21,22,23).
In these models, water sprays are treated as sources of momentum imparted to
the air at the same location as the actual water spray. The quantity of
momentum imparted is related to the flux of the air entrained by the water
spray. It is assumed that the air mixes instantaneously with the gas/air
mixture in the plume at the spray location, resulting in a sudden change in
geometry and composition of the plume. The accuracy of such models depend on
the accuracy of the dispersion model used. At best, they indicate how a spray
barrier can influence the downwind development of a gas plume. However, they
150
-------
can help the planner decide how Che dispersing effects of a water spray system
can best be achieved.
Whether the system is mobile or fixed, designed as a barrier or as a fog,
the following are important design characteristics:
• Water pressure,
• Water flow rate,
• Nozzle type,
• Nozzle spacing, and
• Nozzle orientation.
Research has shown that the efficiency of water spray systems increases
as the water pressure increases (8,24). One source suggests that a minimum
nozzle water pressure of approximately 145 psi be used for fixed systems (8),
which is consistent with experience in gas-cleaning systems where scrubbing
efficiency increases with energy input.
Research has also shown that the most efficient spray systems are de-
signed so that the adjacent sprays just impinge on each other (6,24). This
helps prevent the passage of the cloud between the sprays. Significant over-
lap is inefficient. Thus, the spacing will depend on the nozzle type, size,
and spray angle selected.
Water spray systems perform as deluge type systems with all nozzles open.
Based on the number of nozzles, nozzle spacing, and flow to each nozzle, up to
several thousand gallons per minute of water may be required. The water
supply system must be properly sized to permit operation at the designed
pressure and flow rate for spray system, a municipal water system may not be
able to deliver the necessary flowrate. a second water supply may be required
for the spray system. Many facilities will already have this type of water
system in place for fire protection. It may be appropriate to use this system
to supply the spray system as well. A backup water system will often be
151
-------
composed of a large water storage tank that is automatically kept full. The
spray system designer must decide which accidental release events are most
likely to occur and whether the waste supply contains a sufficient quantity of
water to mitigate these releases. In addition, drainage must be adequate to
prevent flood damage.
Finally, the nozzle type and size must be chosen to provide a density of
water spray at a velocity and droplet size that maximizes the dilution and/or
absorption of the gas cloud. There are many basic types of nozzles; however,
based on recent test data, the most useful are those producing hollow-cone and
fan-tail type sprays, as illustrated in Figures 8-8 and 8-9 (6,8,24). Hoodie
(7) has developed a scheme where the choice of nozzle, its size, spray angle,
and separation distance can be determined in a systematic fashion, based on
the specific momentum flow rate for a spray.
Additional consideration must be given to the piping system when fixed
systems are designed. Two general types are used: wet systems and dry sys-
tems. In a wet system, the pipes are kept full of water tinder pressure. In
dry systems, the pipes are empty until a master valve is opened when the
system is activated. If there is a danger of freezing, a dry system is re-
quired to ensure that the system is not rendered ineffective by freezing.
Consideration should also be given to other hazards, such as potentially
explosive processes nearby, when designing such systems.
8.3.2 Steam
Only limited information is currently available on the effective design
of steam curtains. Seifert (25) has developed a mathematical model that can
be used as a preliminary basis for designing such systems, and experimental
evidence verifies this model (25).
The model is based on the assumption that the toxic and/or flammable gas
cloud moves toward the curtain on one side while the steam jets entrain
152
-------
SWIRL
CHAMBER
LIQUID SPRAY
SHEATH
LIQUID
FLOW
SWIRL SLOT
ELEMENT
Ftgure 8-8. Typical hollow>cone spray nozzle.
Source: Adapted from Reference 26.
153
-------
Figure 8-9. Typical fan-tail spray nozzle.
Source: Adapted from Reference 5.
154
-------
surrounding air from the other side, as illustrated in Figure 8-10. When the
released cloud reaches the curtain, the steam jets mix the gas or vapor with
air and heat the vapors which provides buoyancy, causing the mixture to rise,
dilute, and disperse.
The design method requires knowledge of the mass flow rate and density of
the released gas or vapor. With these data, a subsequent system design can be
determined for a given reduction in concentration.
8.4 APPLICABILITY AND PERFORMANCE
8.4.1 Sprav Systems
Water sprays can be used to help mitigate accidental toxic and/or
flammable gas or vapor releases. However, such systems are often limited in
their scope and may not be capable of dealing with catastrophic releases.
Although much theoretical and experimental work has been done on the use of
these spray systems, the use of water spray systems is not a proven mitigation
technology for all types of toxic and/or flammable gas or vapor releases.
Hazard analysis can be used to determine where a spray system might be
appropriate. Such an analysis can identify the locations within a chemical
facility with the highest potential for an accidental release of a hazardous
vapor. The analysis can also indicate the potential Impact that each release
would have on the surrounding community. Sometimes there will be only a few
locations within a facility with a high accidental release risk. In these
situations, a fixed spray system at each potential release site may be
appropriate. However, most facilities will have a number of possible release
sites.
155
-------
v
CO'
AO' it
^ assv
o*
tv$*
.6®$
,c®
if
o*
*.e
£**'
kc®
So^v
ktc6-
V^>b
•c®
-------
Since mitigative spray system technology is still in a developmental
stage, there are no proven rules for designing an appropriate spray system
when there are multiple release sites.
If a fixed system is installed, the potential exists that a release will
occur in an area not covered by the system. Even if a release occurs in a
location with a fixed spray system, the individual spray nozzles may not be
directed directly at the vapor plume. The sprays from a mobile system can be
directed to optimize dilution of the plume. However, a mobile system cannot
be activated as quickly as a fixed system. A significant quantity of vapors
can be released before the mobile system is in place.
For some facilities it may be appropriate to install fixed spray systems
in the highest risk locations and have a mobile spray system on hand as a
backup.
For flammable vapors, a spray system can help dilute the vapors below the
flammable limit. However, the vapors can still ignite if they are present in
flammable vapor phase concentrations. Mobile spray systems should only be
used in situations where the personnel operating the sprays can be located
away from the flammable gas cloud (this is sometimes impossible to do). A
fixed spray system is safer to use for flammable vapors.
The results of the theoretical and experimental studies in recent years
show that water sprays can be used effectively under certain circumstances.
However, limited and often contradictory performance characteristics, optimal
system design, and reliability data have been reported. Table 8-1 lists some
of these studies, along with major results and performance characteristics.
Several investigations have attempted to quantify the performance of
water spray systems. The most comprehensive method is that developed by Rees
(6) and Moodie (7). It consists of ranking the spray system performance in
157
-------
TABLE 8-1. RESULTS OF SEVERAL EXPERIMENTAL STUDIES ON THE EFFECTIVENESS OF WATER SPRAYS
Description of Study
Summary of Results
Reference
Clouds of ethylene and vinyl chloride
vapor directed from ground level to-
wards 20-foot spray barrier, with
sprays directed downwards
If velocity of entrained air exceeds that
of vapor cloud, spray acts as barrier.
Vapor clouds with velocities of 3 and 9 mph
were stopped by sprays with 100 psig nozzle
pressure. Only lower velocity cloud stopped
with 40 psig pressure. Where cloud passed
through barrier, there was still an ap-
preciable dilution effect. Water spray
did not prevent ignition of flammable
cloud and flame speed actually increased.
1
(1976)
Clouds of propane vapor directed to-
wards spray barrier, with downward
directed sprey nozzles at 1, 5, 9, 15,
and 20 feet above the ground, 55-feet
long with nozzles 5 feet apart.
A properly designed and installed water
spray system can dilute a flammable mix-
ture belcw the LEL. However, the system
does not act as a flame arrester. Igni-
tion of vapor cloud depends on the spacing
of nozzles and water pressure.
2
(1976)
Effectiveness of water sprays in
suppressing combustion in mists of
heat transfer media, specifically
Dowthern A.
Combustion suppressed by a scrubbing mecha-
nism that reduced concentration of droplets
in the mist.
3
(1976)
A water curtain generated by mobile
monitors was used on a chlorine gas
cloud.
The curtain successfully protected a small
area against the gas cloud.
4
(1975)
A series of wide-angle, flat spray
nozzles was used to disperse LNG
vapors.
Two mechanisms are responsible for dilu-
tion: heating the vapor and increased tur-
bulence. Such sprays are effective for
small, confined area LNG spills. Large
spills may require systems too large to be
cost effective.
5
(1977)
(Continued)
-------
TABLE 8-1. (Continued)
Description of Study
Performance Characteritics
Reference
A series of eight experiments ranging
up to full scale were conducted to
evaluate the performance and effec-
tiveness of water sprays. Propane was
used as gas source.
A series of full-scale tests using
hollow-cone spray nozzles inclined
downwards were conducted using carbon
dioxide.
Upward water sprays are significantly more 6 (1981)
effective than downward. The best type
of nozzle is a conical* narrow angle, high
velocity.
Uater spray barriers can be effective in 7 (1985)
dispersing a heavy gas cloud. Nozzles
pointing in a vertical direction or angled
into the gas cloud at a 45° angle are most
effective. Effectiveness can be optimized
in terms of the water pressure.
Two water spray barriers* one using
downward-directed nozzles, the other
using upward-directed nozzles were
used to develop performance data for
full-scale barriers using carbon
dioxide.
High momentum discharges from flat-fan
nozzles with spray angles of 90° or less
are most effective. At a wind speed of 5
mph. a vertical system can produce a ratio
of gas concentration with and without
sprays of approximately 4. At a wind
speed of 20 mph, this ratio decreased to
about 1.5.
8 (1981)
-------
terms of a concentration reduction (CR) ratio (Hoodie (7)) or a forced dis-
persion (FD) factor (Rees (6)); both are a ratio of gas concentrations with
and without a water spray observed at a point downstream.
Moodie (7) demonstrated that, in general, high momentum discharges from
flat-fan spray nozzles with spray angles of 90° or less are the most effec-
tive. Experimental results using carbon dioxide indicated that at a wind
speed of 5 mph, a properly designed vertical water spray system directed
upwards produced a CR ratio of approximately 4. At a wind speed of 20 mph,
the effectiveness was reduced to 1.5.
Prugh (27) developed a systematic method in which the maximum theoretical'
effectiveness of a water spray barrier can be calculated. Material balances
are used to determine the mole-fraction of vapor in the air exiting the
barrier, which is then compared with the average crosswind concentrations of
the vapor in the air entering the barrier, resulting in a reduction factor for
the system. However, this method is limited to water-soluble vapors.
In general, water sprays can be an effective means of absorbing water-
soluble gas or vapors (e.g., ammonia). However, for many released chemicals
the efficiency of absorption may be low because of mass transfer limitations.
As a result, a second water spray barrier may also be required further
downstream.
Likewise, water spray barriers appear to be limited in their effective-
ness. It has been reported that these barriers are only marginally effective
in reducing the hazards of gases and vapors not soluble in water
(3,19,21,28,29,30). Others have gone as far as regarding them as being
limited to water-soluble, low-density, and/or non-flammable vapors (7).
160
-------
8.4.2 Steam Curtains
Steam curtains were initially designed to dilute and contain heavy
flammable vapors and are incorporated into several installations to prevent
highly flammable materials from reaching sources of ignition in the event of
an accidental release (17,18,31). In principle, steam curtains can also be
helpful in mitigating heaver-than-air releases of toxic and/or flammable
vapors. At the time this manual was prepared, no commercial steam curtain
installations were known to exist for mitigating toxic releases.
In general, the characteristics of water spray systems described in
Section 4 are similar to those of steam curtains. However, steam curtains
require large quantities of steam, typically 0.1S ton/hr per foot of curtain.
Therefore, they must often be limited to small-scale uses. Also, where steam
is not available or where the supply of steam is not reliable, water spray
systems may be more applicable.
The effectiveness of steam curtains has been investigated experimentally
by Rees (32), Seifert (26), and Rulkens (33). These investigations found that
steam curtains can reduce the concentration of a vapor cloud from an
accidental release. Table 8-2 shows the effectiveness of steam curtains for
diluting a flammable gas to levels below its lower flammability limit (9).
The primary advantage of steam curtains over water sprays is that steam
curtains enhance the buoyancy behavior of the cloud by heating the vapors
reaching the curtain.
However, steam curtains are reported to be less effective at mixing than
water sprays at the same supply pressure (32). More steam is required for
comparable mixing. This can be more important for toxic than flammable vapors
or gases since the former are hazardous at lower concentrations. The buoyancy
effect may still be beneficial, however. For example, for materials like
ammonia, whose density is near that of air, heating from the steam curtain may
161
-------
TABLE 8-2. EFFECTIVENESS OF STEAM CURTAINS®
Source Strength13 (lb/s)
44
132
264
Wind Velocity
(ft/s)
3.3
6.6
13.2
3.3
6.6
13.2
3.3
6.6
13.2
Distance from
Curtain to
Source (ft)
Steam Flow
(ton/hr)
165
25
+
-
0
+
-
-
-
-
-
50
+
+
0
+
-
-
+
-
-
100
+
+
0
+
+
-
+
+
-
330
25
+
+
0
+
-
0
-
-
-
50
+
+
0
+
+
0
+
-
-
100
+
+
0
+
+
0
+
+
-
500
25
+
0
0
+
-
0
+
-
0
50
+
0
0
+
+
0
+
+
0
100
+
0
0
+
+
0
+
+
0
Note: + effective
- not effective
0 sufficient dilution
by wind alone
£ Based on propylene release diluted below 2 percent.
Amount of propylene released at the source.
Source: Adapted from Reference 9.
162
-------
be sufficient to cause the plume to become buoyant and airborne, which would
result in a significant reduction in downwind ground level concentrations.
8.5 RELIABILITY
8.5.1 Spray System?
Reliability data for water spray release mitigation systems do not appear
to be readily available. In general, the reliability of water spray systems
should depend on the design of the system and on the reliability of the
individual components that make up the system. However, reliability is more
*
than the degree to which a piece of hardware is free of basic defects that may
prevent it from operating properly. It also includes human factors,
interactions, and controls that determine whether a systems is effective or
not..
System unreliability can often result from events other than simple
mechanical failure, including:
• Failure to properly define the hazard;
• Inadequate quality control;
• Poor maintenance;
• Incorrect operations of the system; and
• Changes made to the surrounding systems that were not accounted for
in the original spray system design.
Systems must be regularly inspected and maintained. Of particular
importance are the spray nozzles, which must be inspected and cleaned as
necessary, or provided with some type of nozzle protection to ensure that
their small water passages remain clear.
163
-------
8.5.2 Steam Curtains
A steam curtain system can be no more reliable than the steam source.
Steam pressure must be available. Since the system is dormant much of the
time, the mechanism that activates it could become inoperable if neglected.
Reliability depends on regular inspection, testing, and maintenance.
8.6 SECONDARY HAZARDS
8.6.1 Sorav Systems
Water spray systems pose some potential hazards if they are used in
certain circumstances. This subsection briefly discusses some of the
potential hazards.
For some materials (e.g., hydrogen chloride, hydrogen fluoride, and
chlorine), applying water directly at the leakage point will result in an
acceleration of the corrosive effect of the chemical. This can result in an
enlarged release hole and an increase in the rate of release. Similarly,
spraying water on a spill of liquified gas such as chlorine would add heat to
the subcooled pool of liquified gas, resulting in increased vaporization.
For some chemicals, water sprays can lead to the formation of highly
toxic, acidic, or alkaline liquid droplets that may be hazardous or
potentially more hazardous than the original release.
Finally, where mobile systems are used, personnel can be injured if
appropriate protective clothing and breathing apparatus are not worn.
164
-------
8.6.2 Steam Curtains
Steam curtains, particularly those with jets containing wet steam, can
create static electricity. The generation of static electricity in steam
curtains has been investigated by several individuals (26,33). Static
electricity could ignite a flammable vapor or gas. Static discharges from
steam jets can be prevented by grounding all pipework and other metal in
contact with the steam.
Additional secondary hazards are similar to those presented for spray
systems.
8.7 COSTS
8.7.1 Sprav Systems
The cost of a water spray system depends on the type of system used
(i.e., fixed or mobile) and will be similar to costs for fire protection
systems and equipment. Table 8-3 lists the costs of typical components for
both a fixed and a mobile system. Design bases for these costs are presented
in Table 8-4.
For fixed systems, the major portion of the actual system cost is that
associated with the piping and water supply system. If an existing supply of
water for fire protection systems can also be vised for mitigation purposes,
the only major additional cost is that of the extra pipework and nozzles
needed for the water spray system. These systems may, however, be limited in
their application size. Thus, for very large installations, mobile systems
may be more cost-effective.
165
-------
TABLE 8-3. TYPICAL COSTS
SPRAY SYSTEMS
ASSOCIATED WITH FIXED AND
MOBILE WATER
Item
Capital Cost Range
(1987 $)
Annual Cost Range
(1987 $/yr)
Mobile Hater Spray System:
Fire Hose
4-5 per linear
foot
0.35 - 0.90 per
linear foot
Fog Nozzle
250 - 300
20 - 25
Fire Hydrant
400 - 500
35 - 45
Fire Monitor
750 - 1.000
65 - 85
Fixed Water Spray System:
Deluge type, water spray
system
300 - 350 per
nozzle
25 - 30 per
nozzle
Spray nozzle
60 - 80
5-7
Deluge valve
1.400 - 1.500
120 - 130
Fire pump
25.000 - 30.000
2.100 - 2.600
Source: See Table 8-4 for references.
166
-------
TABLE 8-4. EQUIPMENT SPECIFICATIONS FOR MOBILE AND FIXED WATER
SPRAY SYSTEMS USED IN 00ST ESTIMATES
Item Specification Reference
Mobile Systems:
Fire Hose
Fog Nozzle
Fire Hydrant
Fire Monitor
High strength, 500 lb. test, 2-1/2 in. 36
diameter.
Adjustable nozzle, 1-in. booster inlet 36
Standard 4-in. hydrant with two 2-1/2 in. 36,37
hose connections.
Standard fire monitor with adjustable 36,37
nozzle, 500 gpm capacity.
Fixed Systems:
Water Spray
System
Spray Nozzle
Deluge Valve
Fire Pump
Deluge type, fixed-pipe system capable
of delivering 3,000 gpm of water to
source. Includes piping and supports.
Fan-tail or hollow-cone type, carbon
steel construction.
2-inch deluge valve, including trim,
pressure-operated relief, emergency
release, and pressure gauge.
5-inch centrifugal pump, 99 hp, 100
psi, 1,000 gpm, diesel operated.
38,39
36
36
36
167
-------
8.7.2 Steam Curtain
The costs of typical components associated with steam curtain systems are
shown in Table 8-5. Since there are many individual variations of components,
the system chosen is for illustrative purposes only. These costs provide an
order-of-magnitude basis for estimating the total cost of a steam curtain
system.
168
-------
TABLE 8-5. ESTIMATED COSTS FOR A TYPICAL STEAM CURTAIN SYSTEM®
Capital Cost Annual Cost
(1987 $) (1987 $/yr) References
Piping
7,000
1,200
30
Concrete wall
2,500
220
39
Globe valve
1,000
90
40
Remote shutoff valve
3,000
450
40
Steam supply*'
$4.50/hr
$68/hr
40
a Basis: 6-inch Schedule 40 carbon steel piping, 100 feet long with 3/16-inch
holes, 1 foot apart
250 psig steam supply
Concrete walls, 4 feet high, 100 feet long, 6 inches thick.
b Based on a steam requirement of 0.15 ton/hr per foot of curtain and a steam
cost of $15 per 1000 pounds.
169
-------
8.8 REFERENCES
1. Eggleston, L.A., W.R. Herrera, and M.D. Pish. Water Spray to Reduce
Vapor Cloud Spray. Loss Prevention, Volume 10. American Institute of
Chemical Engineers, New York, NY, 1976.
2. Watts, J.W. Effects of Water Spray on Unconfined Flammable Gas. Loss
Prevention, Volume 10. American Institute of Chemical Engineers, New
York, NY, 1976.
3. Vincent, G.C., et al. Hydrocarbon Mist Explosions - Part II, Prevention
by Water Fog. Loss Prevention, Volume 10, American Institute of Chemical
Engineers, New York, NY, 1976.
4. Experiments With Chlorine. Ministry of Social Affairs, Voorburg, Nether-
lands, 1975.
5. Martinsen, W.E., and S.P. Muhlenkanp. Disperse LNG Vapors With Water.
Hydrocarbon Processing, July 1977.
6. Moore, P.A.C., and W.D. Rees. Forced Dispersion of Gases by Water and
Steam. In: I. Chem. E. Symposium Proceeding. The Containment and
Dispersion of Gases by Water Sprays. Manchester, England, 1981.
7. Moodie, K. The Use of Water Spray Barriers to Disperse Spills of Heavy
Gases. Plant/Operations Progress, October 1985.
8. Moodie, K. Experimental Assessment of Full-Scale Water Spray Barriers
for Dispersing Dense Gases. In: I. Chem. E. Symposium Proceedings. The
Containment and Dispersion of Gases by Water Sprays. Manchester,
England, 1981.
9. Health and Safety Executive (U.K.). Canvey: An Investigation of
Potential Hazards. London, England, 1978.
10. Smith, J.M., and M. van Doom. Water Sprays in Confined Applications:
Mixing and Release from Enclosed Spaces. In: I. Chem. E. Symposium
Proceedings. The Containment and Dispersion of Gases by Water Sprays.
Manchester, England, 1981.
11. Greiner, M.L. Emergency Response Procedures for Anhydrous Ammonia Vapor
Release. Ammonia Plant Safety and Related Facilities. Volume 24. CEP
Technical Manual, AIChE, 1984.
12. Beresford, T.C. The Use of Water Spray Monitors and Fan Sprays for
Dispersing Gas Leakages. In: I. Chem. E. Symposium Proceedings. The
Containment and Dispersion of Gases by Water Sprays. Manchester,
England, 1981.
170
-------
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Personal Communication with industry representative. Name withheld by
request.
Cairney, E.M. , and A.L. Cude. The Safe Dispersal of Large Clouds of
Flammable Heavy Vapours. Institute of Chemical Engineers Symposium,
Series No. 34, London, England, 1971.
Simpson. H.G. The ICI-Vapour Barrier - A Means of Containing and
Dispersing Leakages of Flammable Vapour. Power and Works Engineering,
May 8, 1974.
Simpson, H.G., and A.L. Cude. U.S. Patent 3,882,943. May 13, 1975.
Lees, F.P. Loss Prevention in the Chemical Process Industries, Volumes 1
and 2. Buttersworth's, London, England, 1983.
Barker, G.F., T.A. Kletz, and H.A. Knight. Olefin Plant Safety During
the Last 15 Years. Loss Prevention, Volume II. American Institute of
Chemical Engineers, 1977.
McQuaid, J. The Design of Water-Spray Barriers for Chemical Plants.
Second International Loss Prevention Symposium. Heidelberg, West
Germany, September 1977.
Bucklin, S.M. Aerodynamic Behavior of Liquid Spray Design Method. VKI
Report No. 171, 1980.
Deaves, D.M. Experimental and Computational Assessment of Full-Scale
Water Spray Barriers for Dispersing Dense Gases. Fourth International
Loss Prevention Symposium. Harrogate, England, September 1983.
McQuaid, J. and R.D. Fitzpatrick. The Uses and Limitations of Water-
Spray Barriers. In: I. Chem. E. Symposium Proceedings. The Containment
and Dispersion of Gases by Water Sprays. Manchester, England, 1981.
Zalosh, R.G. Dispersal of LNG Vapor Clouds with Water Spray Curtains.
In: I. Chem. E. Symposium Proceedings. The Containment and Dispersion of
Gases by Water Sprays. Manchester, England, 1981.
Harris, N.C. The Design of Effective Water Sprays - What We Need to
Know. I. Chem. E. Symposium Proceedings. The Containment and Dispersion
of Gases by Water Sprays. Manchester, England, 1981.
Seifert, H.B. Maurer, and H. Giesbrecht. Steam Curtains - Effectiveness
and Electrostatic Hazards. Fourth International Symposium on Loss
Prevention. Harrogate, England, 1983.
Kirk, R.E., and D.F. Othmer. Encyclopedia of Chemical Technology, 3rd
Edition. John Wiley & Sons, Inc., 1980.
171
-------
27. Prugh, R.W. Mitigation of Vapor Cloud Hazards. Part II. Limiting the
Quantity Released and Countermeasures for Releases. Plant/Operations
Progress, July 1986.
28. Lees, F.P. Loss Prevention in the Process Industries - Hazard Identi-
fication, Assessment and Control. Volumes 1 and 2, Butterworth s,
London, England, 1983.
29. Emblem, K. , and O.K. Madsen. Full-Scale Test of a Water Curtain in
Operation. 5th Loss Prevention Symposium. Cannes, France, 1986.
30. Harris, N.C. The Control of Vapor Emissions from Liquified Gas Spil-
lages. 3rd Loss Prevention Symposium. Basle, Switzerland, 1980.
31. Bockman, T., G.H. Ingebrigtsen, and T. Hakstad. Safety Design of the
Ethylene Plant. Loss Prevention, Volume 14. American Institute of
Chemical Engineers, 1981.
32. Rees, W.D. , and P.A.C. Moore. Forced Dispersion of Gases by Water and
Steam. In: I. Chem. E. Symposium Proceedings. The Containment and
Dispersion of Gases by Water Sprays. Manchester, England, 1981.
33. Rulkens, P.F.M. , et al. The Application of Gas Curtains for Diluting
Flammable Gas Clouds to Prevent Their Ignition. Fourth International
Symposium on Loss Prevention. Harrogate, England, 1983.
34. R.S. Means Company, Inc. Building Construction Cost Data 1986, 44th
Edition, Kingston, MA.
35. American Valve and Hydrant. Arlington, TX: vendor.
36. Spraying Systems Company. Wheaton, IL: vendor.
37. Richardson Engineering Services, Inc. The Richardson Rapid Construction
Cost Estimating System. Volumes 1-4, San Marcos, CA, 1986.
38. Yamartino, J. Installed Cost of Corrosion-Resistant Piping--1978.
Chemical Engineering, November 20, 1978.
39. R.S. Means Company, Inc. Building Construction Cost Data 1986, 44th
Edition, Kingston, MA, 1986.
40. Peters, M.S. and K.D. Timmerhaus. Plant Design and Economics for
Chemical Engineers. McGraw-Hill Book Company, New York, NY, 1980.
172
-------
SECTION 9
FOAM SYSTEMS
Foams which are used in the chemical process industries to control and
extinguish'certain types of hydrocarbon fires involving spilled liquids, are
used when the fire may not be effectively controlled by water spray applica-
tion. Although originally developed for firefighting purposes, some of the
same properties that make foams effective for controlling fires also make them
useful for controlling the release of vapors from volatile chemical spills.
The application of a foam blanket to a liquid spill may prevent the release of
_ ,,«„or from reaching an ignition source in concentrations
a flammable gas or vap«t
, ~ -volosion or fire. In the case of a nonflammable
that could result in an exp*v
.. f „ nay help prevent personnel or general public exposure
toxic liquid spiii.
of a hazardous gas or vapor being emitted from the
to dangerous concentrations
surface of the liquid.
This section of the manual discusses the use of firefighting and
vaoor hazards from spilled volatile flammable
specialty foams to control r
. i- tVi« different types of foams and basic design
and/or toxic chemicals. Tiie an.
j as well as their applicability, performance
considerations are descrio . j
ij.kf-Htrv Secondary hazards that may limit the use of
capabilities, and reliability.
hazards from certain hazardous liquid spills are
foams for controlling P
also described.
9.1 BACKGROUND
Foijjj t„ firrflghtln* "f * "" °f
j - liauid hydrocarbon fire, control and extinguish the
which when appli*® co *
' . „ _f foam that make it effective for fighting fires are
flames. The propertie* oi
(1,2):
173
-------
• Ability to blanket the spilled liquid surface with a material
that has a lower density than the liquid, thereby extinguish-
ing an existing fire by cutting off the source of combustion
air;
• Suppression of potentially flammable vapors from being emitted
to the atmosphere and mixing with combustion-supporting air;
• Prevention of nearby flames from heating the spilled liquid
covered by the foams; and
• Cooling of the spilled liquid with water draining from the
foam and surrounding surfaces to help prevent reignition.
These properties are also helpful in mitigating the release of flammable and/
or toxic vapors before actual ignition or exposure of personnel.
Two varieties of foams, chemical and mechanical, are used for firefight-
ing. Chemical foam is produced by a chemical reaction, such as the reaction
that occurs when an aqueous solution of sodium bicarbonate is mixed with
aluminum sulfate and sulfuric acid (3,4). With the addition of other chemi-
cals acting as foam stabilizers, the reaction produces a foam consisting of
carbon dioxide-filled bubbles (1). Originally introduced to control coal
fires in the 19th century (2), chemical foam has been largely replaced by
mechanical foam, which consists of air-filled bubbles (2,3). This "air" foam
is produced by mechanical aeration of a mixture of foam concentrate and water
using a foam-making device. Mechanical foams developed for fire control in-
clude regular protein foam, fluoroprotein foam, surfactant foam, aqueous
film-forming foam (AFFF), and alcohol-type foam (ATF) (5). A discussion of
each of these conventional firefighting foams and other special foams
developed specifically to control vapors is included in Subsection 9.2. The
various types of equipment used to generate and apply these foams to spills
are also described in Subsection 9.2.
174
-------
An important quality of firefighting foam is the expansion ratio, which
is the ratio of the volume of foam produced to the volume of solution fed to
the foam-making device. Foams are classified as low-, medium-, and high-
expansion foams (6). Low-expansion foams have expansion ratios of from 2:1 to
15:1. Medium-expansion foams have expansion ratios from 15:1 to 100:1, while
high-expansion foams are those with expansion ratios of from 100:1 to 1000:1.
While these particular ranges of expansion ratios have been presented in the
literature for firefighting foams (6), the ranges are rather subjective and
can be different for the low-, medium-, and high-expansion foams used for
vapor control. These different range are discussed in Subsection 9.2.
Most firefighting foams, including regular protein, fluoroprotein, aque-
ous film-forming foam (AFFF), and alcohol-type foam (ATF), are available in 3
percent and 6 percent concentrations. This percent value indicates the amount
of foam concentrate that should be mixed with water to form the foam solution,
which is then aerated to create the actual foam. For example, to produce 100
gallons of foam solution, 3 gallons of 3 percent foam concentrate would be
mixed with 97 gallons of water (1). If the particular foam had an expansion
ratio of 10:1, 1000 gallons of foam could be produced for application to the
fire or spill.
Other important qualities of firefighting foams are fluidity and drainage
rate (4). The foam should flow easily around obstructions in the spill area
and quickly cover the liquid surface without breaking up the flame smothering
"blanket." A good firefighting foam possesses a shear stress value of 150 to
250 dynes/cm2, as measured on a torsional vane viscometer (4). The drainage
rate of a foam is a critical property for firefighting, as well as for vapor
control, as discussed in Subsection 9.2. The drainage rate is referred to as
the "25 percent" or "quarter drainage time." The "quarter drainage time" is
the time, in minutes, it takes for a foam to loose 25 percent of the liquid
used to make the foam. For a good firefighting foam, the drainage rate is 2
to 5 minutes (4).
175
-------
9.2 DESCRIPTION
The ability of foams to mitigate vapor released from volatile liquid
chemical spills has grown out of their ability to extinguish fires and prevent
reignition by isolating the spilled material and volatile vapors from ignition
sources. The foam blanket insulates the liquid from ignition sources and
radiant heat sources that cause vaporization. The foam acts as a physical
barrier, because of its limited permeability, and suppresses vapor loss to the
atmosphere. The foam may also absorb vapors being emitted from the spill
surface (3,7). For refrigerated liquid spills, the foam can help warm the
emitted vapor, which passes through the applied foam layer, thus allowing the
vapors to rise and disperse (8). Foams can also provide water for diluting
certain water-reactive chemical spills (3). On the other hand, the
accelerated boil-off of regrigerated liquids or violent reactions with
water-reactive chemical spills can occur. These secondary hazards are
discussed in more detail in Subsection 9.6.
The effectiveness of a foam depends on the type of chemical spilled, the
type of foam iselected to control the spill, the general qualities of foam that
help control the release of vapors, and foam generation equipment.
9-2.1 Types of Pflnma
Six types of foams are used to control vapors from chemical spills:
regular protein foams, fluoroprotein foams, surfactant foams, aqueous film-
forming foams (AFFF), alcohol-type (ATF) or polar solvent foams, and special
foams (9). The first five were developed for firefighting, while the special
foams have been developed specifically to control vapor releases from spills.
Regular Protein Foams--
Protein foams, derived from hydrolyzed protein, are the oldest of the
mechanical firefighting foams. The use of protein foam in controlling vapor
release grew out of its use as an extinguishing agent for aromatic hydrocarbon
fuel fires and its effectiveness in preventing the emission of ignitable
176
-------
vapors. Some of the important characteristics of regular protein foam are
listed below (2,3,6,10):
• Demonstrated spill control capability for water-immiscible,
non-polar hydrocarbon fuels such as petroleum products, gasoline
fuel oil, jet fuel, etc.;
• Designed for low expansion foam-making equipment;
• Desirable physical characteristics (good adhesion, stability, water
retention, elasticity, and mechanical strength);
• Good heat resistance, resistance to burnback (burning back of the
foam blanket if the blanket is broken) and resistance to reignition;
• Non-toxic and biodegradable after dilution;
• Among standard types of foam in use by fire service departments;
• Fresh or sea water can be used to make foam solution;
• Limited shelf life for foam concentrates;
• Poorer flowing ability and slower fire knockdown ability than other
foam types; and
• Can be used in temperature range of 20'F to 120°F.
Fluoroprotein Foams--
Fluoroprotein foams are similar to regular protein foams, but also con-
tain fluorinated surfactants, which allow the foam to shed fuel from its
surface if the foam becomes coated with hydrocarbon material (3,10). This
fuel-shedding characteristic is important when fighting hydrocarbon tank fires
177
-------
or in cases where the foam is injected below the surface of the fuel fire
(10). With the following exceptions, the characteristics of fluoroprotein
foams are similar to those of regular protein foam. These exceptions are
noted below (2,10);
• Better suppression of hydrocarbon fuel vapors than regular protein
and AFFF foam;
• Better resistance to burnback than protein foam;
• Better compatability than regular protein foam with dry chemical
agents; and
• Faster fire knockdown ability than regular protein foam, but slower
than AFFF foam..
Surfactant Foams--
Surfactant foams are produced for application as low-, medium-, and high-
expansion foams from detergent foam concentrates. These concentrates consist
of synthetic hydrocarbon surfactants (syndets) (10). Some important charac-
teristics of surfactant foams are shown below (3,10):
• Applicable to water-immiscible, non-polar hydrocarbons;
• Resistance to burnback and fire knockdown ability is not as good as
that other low-expansion foams;
• Capability of expansion ratios of up to 1000:1 for total flooding
capability;
• Readily available surfactant foams are mostly anionic;
• Non-toxic and biodegradable;
178
-------
• Unlimited shelf life if stored properly;
• Produce less odor when burned than do other foam types;
• Large variety of concentrates on market, with wide range of physical
properties, which can make proper selection difficult; and
• Available in concentrates to yield from 1 to 6 percent foam
solutions;
Aqueous Film Forming Foams (AFFF)
Concentrations for AFFF foams are composed of fluorinated surfactants and
for fast knockdown of water-immiscible, non-polar
hydrocarbon surfactants ror r
* i «„« Additional characteristics of AFFF foam are listed
hydrocarbon fuel fires.
below (2,3,6,10):
• Ability to be used in low- to high-expansion foam-making equipment;
• Ability to spread a water solution film across spill quickly to
ti guish fire, exclude combustion air, and stop fuel from
vaporizing or toxic vapors from being emitted;
lost more quickly after application than with
• Water of composing
other foams;
» Low viscosity;
, -~«,-»ntrates to yield 3 and 6 percent foam solutions;
. Available in concent
can be used to make foam solution;
. Fresh or water c
* ii-ri«eradable after dilution; and
. Hon- toxic andbioMgr
• Long shelf
179
-------
Alcohol-Type Foam (ATF)--
Alcohol-type foam was developed to use on fires involving water-miscible,
polar materials, where ordinary air foams exhibit rapid breakdown and loss of
effectiveness. Examples of polar materials for which ATF foams have been used
are alcohols, paint thinners, methyl ethyl ketone, acetone, isopropyl ether,
acrylonitrile, ethyl and butyl acetate, and amines and anhydrides (10). ATF
foams consist of a regular protein, fluoroprotein, surfactant, or AFFF
concentrate with additives of a metal stearate or a polar material resistant
polymer (2,3,6,10). ATF foams using polymers resistant to polar materials are
applicable to both non-polar hydrocarbon and polar solvents. When these types
of "universal" foams are applied to a spill, a gel is formed on the surface.
This gel forms an insoluble, low-permeability polymeric layer between the foam
and the spilled chemical that protects the film and foam from degradation, and
mitigates the release of vapor from the spill. ATF foams are generally
available in concentrations to yield 3 and 6 percent foam solutions. These
foams are normally used at temperatures of from 35°F to 120°F.
Special Foams- -
Special foams are those developed to suppress hazardous vapors from
liquid spills that typical firefighting foams cannot control effectively.
£
Examples of these are Hazmat NF® Number 1 and Hazmat NP® Number 2 produced by
National Foam, and Type V foam (VEE foam®) by MSA Research (11). Hazmat NF®
Number 1 is designated to use on spills involving alkaline materials, while
Hazmat NF® Number 2 is for acid material spills. MSAR Type V foam is for
controlling hazardous vapors from water-reactive chemicals.
aThese products are mentioned as examples only; such mention does not
constitute an endorsement.
180
-------
9.2.2 Foam Quality
Two important qualities of a foam will effect its ability to control
vapors from hazardous volatile chemical spills: the expansion ratio and the
quarter drainage time. The quarter drainage time is a measure of the
stability of the foam. More stable foams can prevent vapor from being emitted
from the spill for longer periods of time than can less stable foams.
As indicated in Subsection 9.1, expansion ratio classifications are
¦ . rhere is no definite agreement on what constitutes a
somewhat subjective and tnere x e,
w*.rh.evnansion ratio foam. Ratios can range from less than
low-, medium-, or mgn* r
20 1 t er 150-1 High-expansion foams have long quarter drainage times,
nne hour, but can be blown away by the wind, while
sometimes greater than one no«",
c .-d low-expansion foams have quarter drainage times of
medium-expansion foams ana j.
i.. is and 30 minutes, and between 3 and 12 minutes,
between approximately is
respectively.
Medium-expansion foams may be the most effective for controlling vapors
. ii mot-arl&ls. Th«y can be applied more easily than
from spilled volatile materials.
will not be blown away as easily by wind. They also
high-expansion foams an
- j^.inaee times than do low-expansion foams, and thus
exhibit longer quarter drainag
require less frequent reapplication.
9.2.3 Frfiffl Anoiii ntira Sysfrgaa
* - a stored quantity of foam concentrate, a water
Foam systems consist or ®
3 concentrate and a nozzle system to introduce air into
supply to mix with . eXpanded foam for application to the spill,
the foam solution to produce w
the foam soi ,efflifixed, or portable (6).
Foam systems may be •
181
-------
the foam solution to produce the expanded foam for application to the spill.
Foam systems may be fixed, semifixed, or portable (6).
Regardless of the type of foam system selected, a proportioning system or
proportioner is used to mix the foam concentrate with the make-up water to
form the foam solution. The expanded foam is generated by mechanically mixing
air with the foam solution within a nozzle. The specific types of equipment
used to make the foam depends on the type of foam and the expansion ratio.
Low- and medium-expansion foams can be generated using various foam types and
nozzles. Examples of portable air aspirating foam-making equipment typically
used to respond to hazardous chemical spills are described in References 3 and
10. Generally, foam application systems are sold as part of a total package
from the foam vendors, and design of a custom system is not necessary.
9.3 DESIGN
The design of a foam system depends primarily on the type of hazard
involved. In the case of a volatile chemical spill, an appropriate foam con-
centrate should be selected and on hand to control the release of hazardous
vapors from that particular material. For handling emergency releases in
chemical processing plants, the user needs to review the types of hazardous
chemicals and the potential areas for spill releases in order to select the
proper speciality foam and equipment. In the case of a transportation
accident and subsequent chemical release, fire service departments may have a
limited selection of foams to choose from, which may or may not be appropriate
in controlling vapor releases from the spill. A table showing the qualitative
performance abilities of various foams on several chemicals is shown in
Subsection 9.3.
If the type of hazard and foam to control this hazard have been defined,
the other major design criteria are:
• The water supply; and
• The magnitude of the hazard.
182
-------
Without an adequate supply of water at the required rate and pressure,
the foam application system may be inoperable or unable to produce the
necessary quantity of foam to control the release (6).
The potential size of the chemical spill is also an important design
criterion. If the size of a spill, for instance within a diked area, can be
determined before a release, the amount of water at adequate pressure and flow
rate can be planned for, along with the amount of foam concentrate that would
fully contain the release.
For contained and uncontained chemical spills, the size of the hazard
influences the selection of application equipment and methods, and suitable
foams of the appropriate quality for controlling vapors from the spill. These
are discussed below.
9.3.1
flrrUp*f:tpTl AT"* Methods for Low-/Medium-Rxnansion Foams
foam nozzles for applying low- and medium-expansion foams
Air-aspirating xoaw
should b. selected that che d"ln*S* °f U"uld £r°m "*« tom *nd
These include turret-mounted or hand-held nozzles
allow maximum expansion. These in
* _ <*e a stream or spray; however, use of fog-foam nozzles
that disperse the foam as a
.u- axoansion ability of the foam. Different nozzles
can effectively reduce Che expans /
e duality; therefore, it is important to choose the
may produce foam of vary & ...
t. Amaired auarter drainage time and expansion ratio with
nozzle that yields the
., j discussions with foam and equipment manufacturers or
a given foam. Detaiieo
A evaluate equipment and the actual quality of the foam
tests may be required
as applied.
. -hould be applied in such a way that the foam is not
Low-expansion foams snou*
-v,. arsill. The foam stream or spray should be directed
sprayed directly into tto» ®P^
* J .in -nd allowed to flow into the spill area or directed
lust in front of the spm ¦»
just m it behind the spill (7,9). This allows the
onto a wall or other surw. j
,fllP. with minimal disturbance of the spilled chemical,
foam to blanket the surface
183
-------
The selection of application equipment and the foam application rate can
be determined from the desired foam depth, the area of the spill, the time
required to achieve complete blanketing of the spill, and the expansion ratio
of the foam, as given by the formula shown below:
n 7,48 P&
R TE
Where:
R — Discharge capacity of foam solution for nozzles (U.S. gal/min)
D - Depth of foam (ft)
2
A - Spill area (ft )
T - Time required to completely blanket spill (min)
3
7.48 - conversion factor (U.S. gal/ft )
E - Expansion ratio of foam
For a 5000-square-foot spill to be covered by a 0.5 foot thick foam blan-
ket within five minutes, about 190 gallons per minute of foam solution with an
expansion ratio of 20:1 would be needed. This may require the use of several
smaller nozzles (i.e., 3-60 gallons per minute nozzles) or fewer larger flow
rate nozzles. For 60 minutes of operation using a foam concentrate that
yields a 6 percent foam solution, approximately 690 gallons (190 gal/min x 60
min x 0.06) of foam concentrate would be required to control the spill.
9.3.2 Application Equipment and for High-Expansion Foams
High-expansion foams are applied to a chemical spill by means of a foam
chute. They are used indoors to totally flood an area to extinguish fires.
Use of these foams outdoors to control vapor releases from spills is limited
to situations where the wind speed is less than 10 miles per hour, unless
precautions are taken to restrain displacement of the foam by the wind.
184
-------
The application rate for the generated foam should exceed 0.5 cubic feet per
minute per square foot of spill area. Foams should be applied to flammable
liquids to a thickness of at least 18 inches.
9.4 APPLICABILITY AND PERFORMANCE
Certain types of aqueous foams have been shown to be effective in con-
trolling the release of flammable and/or toxic vapors from various chemical
spills. A major criterion in selecting a foam is the compatibility of the
foam with the spilled chemical. Foams have limited effectiveness in con-
trolling vapor releases from flowing spills. To control vapor releases, the
spill must first be physically contained.
Currently, it does not appear that foams are used to control airborne
releases of hazardous chemicals; however, foams have been used for -gas
scrubbing (8). Some work has been done where a contaminated air stream has
been sent into a foam generation machine, where the air is incorporated into
the foam. The contaminant is then temporarily trapped in the foam, where its
rate of release is slowed, or where it is absorbed by the water in the foam.
Such a system could be used wherever the vapors from an accidental release
could be contained and routed into the foam-generation machine. In such
situations, however, an alternate treatment system such as a scrubber or flare
could also be used.
The use of aqueous foams for controlling vapors from water-Immiscible
hydrocarbon fuels is well established because of their use in extinguishing
fires. The use of typical commercial fire-fighting foams is limited to
non-polar materials with dielectric constants of less than three; otherwise,
rapid degradation of the foam will occur (3). For control of water-miscible,
polar material spills, alcohol-type foams (ATF) have been developed that
protect the foam blanket by creating a film barrier between the spilled
material and the foam. Universal or multipurpose foam agents have also been
produced that are effective against both polar and non-polar solvents, and
185
-------
chemically neutral material spills (7). Specialty foams have been developed
to control hazardous vapor releases from spills of alkaline-, acid-, and
water-reactive chemicals.
Another major point in foam selection is the drainage of liquid from the
foam, or the quarter drainage time. The longer the quarter drainage time, or
the longer the foam retains its water, the better its fluidity, spreading
capability, collapse resistance, and heat resistance (in the case of fire)
(11).
Because of the variety of fire-fighting and specialty foams available and
the numerous chemicals in commercial use, it is important to evaluate which
foams are applicable for controlling vapor from spills of particular chemi-
cals. Table 9-1 shows an overview of the types of foam which are useful for
controlling vapors from spills of specific chemicals. Shown in this table are
the chemical group, the particular chemical, the recommended type of foam,
satisfactory types of foam, and ineffective or dangerous types of foams. New
foams for vapor control will probably be developed for these and other chemi-
cals. Therefore, foam manufacturers should be consulted before selecting foam
concentrates, particularly for chemicals not listed in Table 9-1. Table 9-2
shows typical application information about some universal and specialty foams
available from some manufacturers.
Depending on the type of foam used and the material spilled, foams can
reduce vapor concentrations, measured just above the foam layer, by over 90
percent (8). However, this reduction is temporary and eventually the vapor
concentration above the applied foam layer will reach its uncontrolled
concentration as the foam breaks down over time. The length of time that a
foam can effectively mitigate vapors without reapplication can range from 5 to
120 minutes (8). An illustration of the mitigation of vapor released from
liquid benzene is illustrated in Figure 9-1 (3).
186
-------
TABLE 9-1. FOAMS FOR CONTROL OF FLAMMABLE/TOXIC VAPORS
Group
Chemical
Recomaended Foaa
Types (Minimum
Thickness/
Reapplication
Rate)(in/min)a
Satisfactory Foam
Types (Minimum
Thickness/
Reapplicat ion
Rate)(in/min)a
Inef fective/Dangerous
Foam Types (Reason)
ALCOHOLS
Butanol,
aethanol, and
propanol
Alcohol (ND)b
None
AFFF (collapse)
Fluoroprotein (collapse)
Protein (collapse)
Surfactant H&L (collapse)
Octanol
Alcohol (ND)
AFFF
Fluoroprotein
Protein
Surfactant H&L
ALDEHYDES
AMD
KETONES
Acetone,
¦ethyl ethyl
and aethyl
butyl ketones
Alcohol (10/60)
None
AFFF (collapse)
Fluoroprotein (collapse)
Protein (collapse)
Surfactant H&L (collapse)
AMINES
Ethylamines, Alcohol (ND) ,
ethylene diamine, Hazmat NF #1 (ND)
hydrazine, and MSA Type V (ND)
¦ethylanines
None
AFFF (collapse)
Fluoroprotein (early
breakthrough)
Protein (collapse)
Surfactant H&L (collapse)
(Continued)
-------
TABLE 9-1. (Continued)
Group
Chemical
Recommended Foam
Types (Minimum
Thickness/
Reappliestion
Rate)(in/min)a
Satisfactory Foam
Types (Minimum
Thickness/
Reapplication
Rate)(in/min)a
Ineffective/Dangerous
Foam Types (Reason)
ETHERS
Ethyl ether
Alcohol (5/25-120)
None
AFFF (collapse)
Fluoroprotein (collapse)
Protein (collapse)
Surfactant (collapse)
ESTERS
n-Butyl acetate
AFFF (5/120) None
Alcohol (5/120)
Fluoroprotein (5/120)
Protein (5/120)
Surfactant L (5/120)
Surfactant H (no
information)
Methyl acrylate
Alcohol (ND)
None
AFFF (collapse)
Fluoroprotein (collapse)
Protein (collapse)
Surfactant H&L (untested)
Vinyl acetate
(monomer)
Alcohol (ND)
Fluoroprotein (ND)
AFFF (ND)
Protein
Surfactant H&L
HYDRO-
CARBONS
(ALI-
PHATIC)
Ethane and
ethylene
Surfactant H (ND)
Alcohol (ND) AFFF (accelerates boil-
Fluoroprotein (ND) off)
Protein (ND)
Surfactant L (ND)
(Continued)
-------
TABLE 9-1. (Continued)
oo
to
Group
Chemical
Recommended Foam Satisfactory Foam
Types (Minimum Types (Minimum
Thickness/ Thickness/
Reapplication Reapplication
Rate)(in/min)a Rate)(in/min)a
Ineffective/Dangerous
Foam Types (Reason)
HYDRO-
CARBONS
(ALI-
PHATIC)
(con* t)
Heptane
Alcohol (ND)
Fluoroprotein (ND)
Protein (ND)
Surfactant H&L (ND)
AFFF (ND)
None
Hexane and
octane
Alcohol
AFFF
Fluoroprotein
Protein
Surfactant
None
HYDRO-
CARBONS
(AROMATIC)
Benzene
Alcohol (2.5/120)
Fluoroprotein
(5/60)
Protein (5/60)
Surfactant H
(25/50)
AFFF (early break-
through)
Surfactant L (early
breakthrough)
Ethylbenzene
AFFF (5/120)
Alcohol (5/120)
Fluoroprotein
(5/120)
Protein (5/120)
Surfactant H
(50/30)
None
(Continued)
-------
TABLE 9-1. (Continued)
Group
Chemical
Recommended Foam
Types (Minimum
Thickness/
Reapplication
Rate)(in/min)8
Satisfactory Foam
Types (Minimum
Thickness/
Reapplication
Rate)(in/min)8
Ineffective/Dangerous
Foam Types (Reason)
o
HYDRO-
CARBONS
(AROMATICS)
(Con't)
HYDRO-
CARBONS
(ALI-
CYCLIC)
HYDRO-
CARBONS
(INDUS-
TRIAL
Toluene
Cyclohezane
Gasoline
and
kerosene
Alcohol (ND)
Fluoroprotein (ND)
Protein (ND)
Alcohol (5/>120)
Fluoroprotein (5/>120)
Protein (5/>120)
Alcohol (ND)
Fluoroprotein (ND)
Protein (ND)
Surfactants (ND)
AFFF (5/40)
Surfactant H (50/30)
Surfactant L (5/30)
AFFF (5/60)
Surfactant H (ND)
Surfactant L (5/30)
AFFF (ND)
None
None
None
Naphtha
Alcohol (ND)
Fluoroprotein (ND)
Protein (ND)
AFFF (ND)
Surfactants (ND)
None
LIQUEFIED
ORGANIC
GASES
Ethylene
oxide
None
Alcohol (ND)
AFFF (collapse)
Protein (collapse)
Fluoroprotein
Surfactant H
(untested)
(Continued)
-------
TABLE 9-1. (Continued)
Group
Recommended Foam
Types (Minimum
Thickness/
Reapplication
Chemical Rate)(in/min)a
Satisfactory Foam
Types (Minimum
Thickness/
Reapplication
Rate)(in/min)a
Ineffective/Dangerous
Foam Types (Reason)
LIQUIFIED
ORGANIC
GASES
(con* t)
Liquefied
natural gas
Surfactant H (ND)
None
AFFF (early breakthrough)
Alcohol (early break-
through)
Fluoroprotein (early
breakthrough)
Protein (early break-
through)
Surfactant L (early
breakthrough)
INORGANICS
Carbon disulfide
Hazmat NF #2 (ND)
MSA Type V (ND)
None
AFFF (untested)
Alcohol (untested)
Fluoroprotein (untested)
Protein (early break-
through)
Surfactant H&L (untested)
Hydrochloric
acid and
hydrogen
chloride
(anhydrous)
Hazmat NF #2 (ND)
MSA type (V (ND)
None
AFFF (untested)
Alcohol (untested)
Fluoroprotein (untested)
Protein (untested)
Surfactant H&L (untested)
(Continued)
-------
TABLE 9-1. (Continued)
Group
Chemical
Recommended Foam
Types (Minimum
Thickness/
Reapplication
Rate)(in/min)a
Satisfactory Foam
Types (Minimum
Thickness/
Reapplication
Rate)(in/min)a
Ineffective/Dangerous
Foam Types (Reason)
INORGANICS
(con't)
Nitric acid
Hazmat NF #2 (ND)
MSA type V (ND)
None
AFFF (untested)
Alcohol (untested)
Fluoroprotein (untested)
Surfactant H&L (untested)
Silicon
tetrachloride
Surfactant H (ND)
None
AFFF (violent reaction)
Alcohol (violent reaction)
Fluoroprotein (violent
reaction)
Protein (violent reaction)
Surfactant L (violent
reaction)
Sulfur trioxide
Hazmat NF #2 (ND)
Surfactant H (ND)
MSA Type V (ND)
None
AFFF (violent reaction)
Alcohol (violent reaction)
Fluoroprotein (violent
reaction)
Protein (violent reaction)
Surfactant L (violent
'reaction)
Titanium
tetrachloride
Hazmat NF #2 (ND)
MSA Type V (ND)
None
AFFF (untested)
Alcohol (untested)
Fluoroprotein (untested)
Protein (untested)
Surfactant H&L (untested)
(Continued)
-------
TABLE 9—1. (Continued)
Group
Chemical
Recommended Foam
Types (Minimum
Thickness/
Reapplication
Rate)(in/min)a
Satisfactory Foam
Types (Minimum
Thickness/
Reapplication
Rate)(in/min)a
Inef f ective/Dangerous
Foam Types (Reason)
INORGANICS
(con'O
Ammonia
Surfactant H
(30/10)
AFFF (8/5) None
Alcohol (8/5)
Surfactant (8/5)
Fluoroprotein (8/5)
Protein (8/5)
INORGANIC
Bromine and
chlorine
Hazmat NF #2 (10))
Surfactant H (ND)
MSA Type V (ND)
None
Other foams are ineffec-
tive for use on chlorine,
causing accelerated boil-
off of heavier-than-air
chlorine gas.
0
Reapplication times are generally based on the amount of tine before a 1-percent vapor
concentration by volume is established above the spill surface.
k No data are available on foam thickness or reapplication rate.
£
H - high—expansion surfactant foam; L = law-expansion or medium-expansion Burfactant foam.
^ Ha*mat NF #1. Hazmat NF #2, and MSA Type V are specified by name because they were designed for
controlling hazardoua vapors. This is not an endorsement.
-------
TABLE 9-2. PERFORMANCE SUMMARY OF THREE UNIVERSAL AND SPECIALTY FOAMS
Foam: National Universal Foam
General Applicability
Multi-purpose foam for extinguishing hydrocarbon, alcohol, and polar
solvent fires; also for vapor control for hydrocarbons, alcohols, polar
solvents, and chemically neutral materials.
Chemicals for Which this Foam is Ineffective:
Chlorosulfonic Acid
Fluorosulfonic Acid
Phosgene
Phosphorus Oxychloride
Phosphorus Trichloride
Sulfuryl Chloride
Foam: Hazmat NF* Foam Number 1
General Applicability
Foam for vapor control for alkaline materials spills.
Chemicals for Which this Foam is Ineffective:
(see Table 9-1)
Foam; Hazmat NF* Foam Number 2
General Applicability
Foam for vapor control for acid material spills.
Chemicals for Which this Foam is Ineffective:
(see Table 9-1)
Foam: MSA VEEFOAM1"
General Applicability
Multi-purpose foam for extinguishing fires and control of vapors;
provides effective vapor control for water-immiscible organic liquids and
acid and alkaline inorganic materials; can be used in low- and high-
expansion foam generating equipment*
Chemicals for Which this Foam is Ineffective:
Polar compounds (with dielectric constant greater than 15)
Some organic acids
Some inorganic acis
Uncontained spills of liquified gases
194
-------
2.2 t
AFFF - Medium and low expansion
Foam height 2 in. - low expansion
Foam height 18 in. - medium expansion
3% National Foam
6% Light Water
3% low
exp.
TFT.
3% med. exp.
adjustable dectector
height 7 in. when
LEL reached
6% med. exp.
adjustable detector
height 6.5 in. when
LEL reached
70 80 90
Time (minutes)
Figure 9-1. Benzene vapor concentration versus tine.
Source: Adapted from Reference 3.
195
-------
9.5 RELIABILITY
Information concerning the reliability of foam systems for use in mitiga-
ting the release of flammable and/or toxic vapors from chemical spills does
not appear in the literature; however, in general, the reliability of any
system depends on the reliability of the system's components, human factors,
interactions, and controls, as discussed in the section on spray systems.
The reliability of foam systems in emergency situations, either for fire
or spill control, depends on regular inspection, maintenance, and testing of
foam concentrates and application equipment.
To evaluate the reliability of foam concentrates on hand for emergency
responses the following should be considered (10):
• Storage in accordance with the manufacturer's instruc-
tions, including shelf life;
• Protection from exposure to extremes of heat and cold;
• Adequate precautions to avoid contamination with other
materials; and
• Inspection of concentrate for formation of any precipi-
tates that may affect the foam's performance or render the
foam concentrate useless in an emergency.
For foam-generating and applicating equipment, adequate procedures should
be included for periodic inspection, maintenance, and testing of various com-
ponents of the system without actually producing the foam on a full-scale
emergency basis. Valving, piping, and any electrical controls can be checked
for proper operation and to ensure that corrosion has not rendered the system
inoperative. If special test equipment and qualified personnel are available,
196
-------
some important equipment performance characteristics can be checked against
the results of tests conducted when the equipment was first installed
accepted. These include (10):
• The foam equipment discharge pattern;
• The percent concentration of the foam concentrate in the foam
solution;
• The expansion ratio of the foam;
• The quarter drainage rate of the foam; and
• The film-forming characteristic of the concentrate.
9.6 SECONDARY HAZARDS
The primary purpose of using foam for controlling vapors from hazardous
chemical spills is to minimize hazards to plant personnel, emergency spill
response teams, and the general public. However, if the selected foam is
incompatible with the spilled material, more vapors may actually be released
than controlled. For example, certain types of foams may lose their liquid
quickly and form openings in the foam layer, causing rapid generation of
flammable vapors from some chemicals (i.e., liquified natural gas spills).
Selection of a slow-draining foam that maintains an effective blanket over the
spill but that allows the gas to become buoyant enough for safe dispersal over
a period of time can eliminate this major secondary hazard. Also, if the
wrong foam is used to respond to a spill of an inorganic water-reactive
chemical, more vapor could be released than if no foam had been used in the
first place.
197
-------
Foams that cause increased vaporization should not be used in totally
enclosed areas by personnel without adequate safety precautions, including
self-contained breathing apparatus and protective clothing. Alternatively, a
foam should be selected that is specifically formulated to reduce this
secondary vaporization hazard.
Protective clothing should be worn where mobile equipment is used.
Appropriate breathing apparatus should also be used by response personnel.
Personnel should be properly trained in the use of all safety equipment to
minimize exposure to spilled liquids and hazardous vapors.
9.7 COSTS
Costs for foam and foam application systems vary widely and depend on the
type of foam selected and the type and complexity of equipment used. Approxi-
mate costs, for illustrative purposes only, for various foams to industrial
customers are presented in Table 9-3 (12).
Note that the prices in Table 9-3 are approximations; the actual prices
may differ, depending on the volume of foam concentrate purchased by a
customer and on the particular distribution system.
Costs for application equipment also vary, ranging from small manually
operated portable systems to large automatically operated fixed systems and
self-contained mobile foam trucks. Examples of foam application equipment
costs are shown in Table 9-4 (12,13).
198
-------
TABLE 9-3. APPROXIMATE COSTS FOR VARIOUS FOAMS
Type of Foam
Cost ($/Gallon)
Regular protein foam (3 percent
formulation)
$8/gallon
Surfactant foam (1.5 percent
formulation)
$ll/gallon
Aqueous film forming foam
(3 and 6 percent formulations)
$9 - $12/gallon (depending
on formulation)
Alcohol type foam (3 and 6 percent
formulations)
$14 - $15/gallon (depending
on formulation)
Specialty foams (6 percent
formulations)
$15/gallon
199
-------
TABLE 9-4. EXAMPLES OF FOAM APPLICATION EQUIPMENT COSTS
Equipment
Cost 1987 ($)
Smallest manually operated portable
systems
Larger manually operated portable
system with nozzle and eductor
Modification of city fire department
pumper for 1 to 2 foam discharge lines
Skid-mounted foam system for semi-
fixed or permanent installation
Trailer-mounted foam system
Self-contained mobile foam trucks
with storage for foam concentrate
and water
Fully automatic high-expansion foam
generators for total flooding
Small portable high-expansion foam
generators
$ 125
$ 600
$1,200 to $1,500
$15,000 to $100,000, depending
on whether full automatic
control is required
$6,000 to $30,000, depending
on size
$120,000 to $400,000,
depending on additional safety
and spill control equipment
needed
$8,000 to $25,000
$1,000 to $1,500
200
-------
9.8 REFERENCES
1. Vervalin, C.H. Role of Foam in Fighting Flares F'
for Hydrocarbon Processing Plants, Volume 1 Third r-rManuaL
Publishing Company, Houston, TX, 1985. d Edition- Gulf
2. Basically Speaking, Foam Systems. The Sentinel, Second Quarter, 1986
3. Gross, S.S., and R.H. Hiltz. Evaluation of Foams for
Pollution from Hazardous Spills. EPA-600/2-82-029 (NTIS PB89ooi^s
U.S. Environmental Protection Agency, March 1982 ^B8Z-227117),
4. Lees, F.P. Loss Prevention in the Process Industrie n «...
London, England, 1983. industries, Butterworth's,
5. Chandnani, M.F. Design Fundamentals. Fire proterH
Hydrocarbon Processing Plants, Volume 1 3rd on Manual for
Conpany, Houston, TX. 1985. ' 3rd EdI"°". Gulf Publishing
6. Gillespie, P.J. and L.R. DiMaio. How Foam Can Protect hpt di
Protection Manual for Hydrocarbon Processin* rr t Plants. Fire
Publishing Company, Houston, TX, 1981. 8 "«ts, Volume 2. Gulf
7. National Foam Systems, Inc. Controlling Hazard™,* w.
literature, Section XIV, 1986. g rdOUS advertising
8. Evans, M.L., and H.A. Carroll. Handbook for Usin* .. „
Vapors from Hazardous Spills. EPA-600/8-86/019 (NTIS^RR7°iaSS?1
Environmental Protection Agency, July 1986. "»°7-145660), U. S.
9. Hiltz, R.H. The • Potential of Aqueous Foams to M-fi-i
from Released Volatile Chemicals. International <; ®ate t*ie VaPor Hazard
Major Chemical Accidents. American Institute. ymposium on Preventing
Washington, D.C., February 1987. C8 °f Chemical Engineers,
10. McKinnon, G.P., (ed.), The National Fire Protection ,
Fire Prevention Handbook, 15th Edition. Quincy MA 1981 Nationa'''
11. MSA Research Corporation. Advertising Literature, Data sheet 18-01-01.
12. Personal communication with E.C. Norman, National e ^
Llonvllle, PA, April 1987. "OMl F"« System, Inc.,
13. Personal communication with J. Headrick, Global Fire
Ft. Worth, TX, April 1987. " EqUlp"enC C™Pany,
201
-------
1023117
lb -3
a M
APPENDIX A
TABLE A-1. METRIC (SI) CONVERSION FACTORS
Quantity
To Convert From
To
Multiply By
Length:
Area:
Volume:
Mass (weight):
Pressure:
Temperature:
Caloric Value;
Enthalpy:
Specific-Heat
Capacity:
Density:
Concentration:
Flowrate:
Velocity:
Viscosity:
in
,f5
in,
ft3
in,
ft
gal
lb
short ton (ton)
short ton (ton)
atm
mm Hg
psia
psig
°F
°C
Btu/lb
Btu/lbmol
kcal/gmol
Btu/lb-8F
lb/ft3
lb/gal
oz/gal
quarts/gal
gal/min
gal/day
ft /min
ft/min
ft/sec
centipoise (CP)
cm
2
cm2
m3
cm3
m^
m
kg
Mg
metric ton (t)
kPa
kPa
kPa
kPa*
°C*
K*
kJ/kg
kJ/kgmol
kJ/kgmol
kJ/kg-8C
kg/m3
kg/® ^
k§/m3
c^ /m
m^/min
m3/day
m /min
m/min
m/sec
kg/m-s
2.54
0. 3048
6.4516
0.0929
16. 39
0.0283
0.0038
0.4536
0.9072
0.9072
101.3
0.133
6.895
(psig+14,696)x(6.895)
(5/9)x(°F-32)
"C+273.15
2.326
2.326
4.184
4.1868
16.02
¦119.8
25,000
0.0038
0.0038
0.0283
0.3048
0.3048
0.001
^Calculate as indicated
202
DATE DUE
-------
TECHNICAL REPORT DATA
(Please read [nuructions on the reverse before completing)
flTREPORT NO.
EPA-600/8-87-039b
3. RECIPIENT'S ACCESSION NO.
|4. TITLE ANO SUBTITLE i ^ .
Prevention Reference Manual: Control Technologies,
Volume 2: Post-release Mitigation Measures for
Controlling Accidental Releases of Air Toxics
9. REPORT DATE
January 1989
6. PERFORMING ORGANIZATION COOE
D. S. Davis, G. B. DeWolf, K. A. Ferland,
D. L. Harper, R. C. Keeney, and J. D. Quass
». PERFORMING ORGANIZATION REPORT NO.
PERFORMING ORGANIZATION NAME ANO AOORESS
Radian Corporation
Austin, Texas 78720-1088
10. PROGRAM 6LEMEN+ NO.
i!.C6KiTftAgT/6ftAKB, 9lS
[IS. SUPPLEMENTARY
541-2852.
116. A
ACT The report covers post-release mitigation -r
releases of air toxics. Reducing the possibility of amiri ^r.es Contr°l accidental
ses reduces the possibility of harm to human health hT toxic chemical relea-
such a release does occur, however, its conseauenc ^ ° ^ env*ronrnent- When
be accomplished by a variety of mitigation measures6^"?18* be reduced* This Can
troy, divert, or disperse the released chemical Mit- V- Can contain' capture, des- |
initial siting and layout of a facility to decrease th 1®atlon measures begin with the J
release. The extent of the area potentially affected th&t would. be ^ected by a
chemicals reaching those areas, and the duration of 6 concentrations of toxic
vapor or gas dispersion modeling. The extent and m exp°s"re can be estimated by
can be determined using meteorological instrument a^rh 311 actua* re*eaSe
emergency planning and training, are the first stenS'* .fSe ®ystenis, along with
Other measures involve the use of mitigation techni ^ mitigation process.
tainment systems, and spray or foam systems. Th qU6S SU°h &S *eak pluSSini' con"
these methods are discussed. * e ®eneral application costs of
Il7.
|a. DESCRIPTORS
pollution Maintenance
Chemical Plants Education
^mission Management
Toxicity
Chemical Compounds
Accidents
| Design
119. DISTRIBUTION STATEMENT
Release to Public
EPA Perm 2320-1 (>-73)
KEY WORDS AND OOCUMENT ANALYSIS
^b.IOENTIF16RS/OPEN ENOEO TERMS
Pollution Control
Stationary Sources
Air Toxics
Chemical Releases
Training
1ft. SECURITY CLASS (This Reporti
Unclassified
its ¦ -
203
Unclassified8
COSATi Field/Group
13B I5e"
07a 151
14G 15A
06T
07B.07C
13 L
21. NO. I
213
22. PRICE
------- |