United States
Environmental Protection
Agency
Solid Waste
and Emergency Response
(OS-120)
July 1989
OSWER-89-008.1
Series 8, No. 1
4>EPA Why Accidents Occur
Insights From The
Accidental Release
Information Program
Chemical Accident
Prevention Bulletin
Printed on Recycled Paper
-------
-------
CHEMICAL ACCIDENT PREVENTION
AND
THE EMERGENCY PLANNING AND
COMMUNITY RIGHT-TO-KNOW ACT OF
1986 (TITLE III)
In the wake of the chemical tragedy in Bhopal, India, the U.S.
Congress enacted the Emergency Planning and Community
Right-to-Know Act of 1986, also known as SARA Title III.
Besides establishing state and local entities to plan for chemi-
cal emergencies, Title III requires facilities that handle hazard-
ous chemicals to make information on the presence of those
chemicals available to the public. Most of the initial steps in
the Title III process were directed toward preparedness
planning for possible chemical accidents. But the provision of
information on chemical hazards and the dialogue with the
public that such information creates can lead to a sharper
recognition of hazards, actions to reduce those hazards, and,
therefore, the prevention of chemical accidents.
EPA's report to Congress on chemical accident prevention,
Review of Emergency Systems (June 1988), found that preven-
tion of chemical accidents requires a holistic approach inte-
grating technologies, procedures, and management practices.
Sustaining a comprehensive approach depends on manage-
ment's awareness of risks and commitment to safety. While
the primary responsibility for accident prevention rests with
industry, the circle of those involved with chemical safety is
broad. It includes chemical manufacturers and distributors as
well as the multitude of other businesses, such as furniture
manufacturers and local dry cleaners, that routinely use
hazardous chemicals. It covers every level of government and
professional societies, trade associations, labor unions, the
research community, environmentalists, and the general
public.
Everyone has a role in encouraging industry to meet its re-
sponsibilities. Open communication, an understanding of the
hazards, and a willingness to work together can bring us
toward the ultimate prevention goal: no accidental releases.
-------
-------
WHY ACCIDENTS OCCUR:
INSIGHTS FROM THE ACCIDENTAL
RELEASE INFORMATION PROGRAM
INTRODUCTION
In this bulletin, EPA focuses on the causes of accidents, using
information gathered through a pilot program -- the Accidental
Release Information Program (ARIP). EPA established ARIP to
promote safety initiatives by industry and to develop a na-
tional database on the causes of chemical accidents, but more
important, to identify methods used to prevent recurrences.
The purpose of this bulletin is to present insights drawn from
the pilot program to local emergency planning committees
(LEPCs) to help in dialogues with local facilities.
PREVENTION AND LEPCS
For you and for facilities in your community, the importance of
understanding ways to prevent accidents is clear: preventing
accidents is more effective in reducing risks and far less ex-
pensive than responding to accidents. The primary responsi-
bility for preventing accidents must lie with industry. LEPCs
and the public can, however, play an important role. With the
information you have gained through the Title III planning
process, you know which facilities in your community handle
hazardous substances that may pose a threat to the commu-
nity should an accident occur. You are, therefore, in a good
position to decide which facilities should be the focus of pre-
vention activities.
You can help your local facilities prevent releases of chemicals
by engaging them in a dialogue about the hazards of their
chemicals and processes, and about ways to reduce the
chance of an accident. While some safety techniques are
technical, the Section 305(b) final report to Congress, Review
of Emergency Systems, indicated that plant management
practices are critical to preventing releases. In your dialogue
with industry, you can ask questions that address how the
facility identifies and deals with hazards and the level of man-
agement's commitment to safety. The information in this
bulletin can help you by identifying common problem areas
-------
(Ill,
chemicals that are frequently involved in accidents, common
causes of accidents, and actions facilities have found useful to
prevent further accidents.
ARIP: INVESTIGATING THE CAUSES OF
CHEMICAL ACCIDENTS
A number of national databases on chemical accidents exist
(see the Appendix for a summary description of these data-
bases). Most were not designed specifically to address the
causes of accidents; rather, they focus on information needed
to determine an immediate response on what happened
rather than on why it happened. Recognizing the shortcom-
ings of existing information sources, EPA developed a ques-
tionnaire on the practices facilities use to prevent releases and
on the techniques they use to assess hazards. Questionnaires
are sent to facilities that report significant or frequent releases
to the National Response Center; facilities are required to
report any releases of certain chemicals if the releases exceed
specified quantities. ARIP is now being fully implemented as a
national program, but the Agency felt it was important to
share lessons learned from tiie pilot with those who are talking
with industry. (More specific information on ARIP and the
pilot program findings can be found in the Appendix.) The
findings that may be of use and interest to you include the
following:
The most frequently released chemicals in the ARIP
database have been chlorine, methyl chloride, ammo-
nia, sulfuric acid, and sodium hydroxide ~ all large-
volume, industrial chemicals.
Most of the releases occurred at facilities that manufac-
ture chemicals or other products.
The quantity of chemicals released varies greatly.
Although accidents commonly have more than one
cause, the most commonly cited causes are equipment
failure and operator error.
About a quarter of the releases were from storage ves-
sels, and a similar number from piping and process
vessels. Valves and other equipment contributed to a
smaller fraction of releases.
Most releases occurred during routine processing of
chemicals; loading, unloading, and maintenance played
a lesser, but significant role.
-------
These general findings suggest that LEPCs may want to focus
first on facilities that use large volumes of hazardous chemi-
cals. Manufacturers may need special attention because they
use many hazardous chemicals and because they have many
complicated processes where releases can occur. You should
note, however, that some non-manufacturers also have large
volumes of hazardous chemicals for example, public drink-
ing water systems, non-residential swimming pools, and food
storage facilities.
Brief summaries of three accidents reported in ARIP question-
naires, followed by a discussion of specific concerns raised by
the ARIP findings, are presented below. Questions you may be
able to use as you talk with facilities in your community are
included for each area of concern.
ARIP ACCIDENT EXAMPLES
These examples are drawn from the more than 330 ARIP
questionnaires received as part of the initial pilot. While these
accidents are not the worst, nor are they typical (there is no
such thing as a typical accident), they serve as useful illustra-
tions of what can go wrong.
A truck driver incorrectly identified the contents of his
truck and began to unload hydrochloric acid into a
facility tank containing nitric acid. The liquid mixture
ate away the walls and bottom of the stainless steel
tank. Although the spill was confined in a concrete
containment, some of the mixed acid sprayed and
splashed beyond the containment and a yellow cloud
rose from the released liquid. Lime was applied to
neutralize the acid mixture and a water fog was used to
prevent further air contamination. To prevent a recur-
rence, the facility is installing new fittings on the nitric
acid fill pipes that are incompatible with the hydrochlo-
ric acid fittings.
Given the circumstances surrounding the incident, the
facility could also have required, as part of a revised
standard operating procedure, that an operator be pres-
ent during all chemical deliveries to assure that the
chemicals are properly identified and transferred and to
keep unloading valves locked to prevent unauthorized
use. This approach might have prevented such an error
from occurring or halted the unloading at an earlier point.
-------
(In,
A steel manufacturer released 5,000 gallons of chromic
acid plating solution from a storage tank. Although all
valves to the tank and transfer pipe were in the proper
position, the chromic acid leaked out over a three-day
period. Upon further investigation, it was discovered
that one of the valves was defective and another valve
was badly corroded by the acid. Both valves have since
been replaced with lined plug valves more suitable for
contact with acid.
In. this case, the facility could also have run corrosion
tests to determine the optimum material of construction
for the valve, instituted aperiodic corrosion inspection
and valve maintenance program, and trained operators
in the use of procedures or checklists to inspect for leaks
and obviously corroded equipment
An aluminum sheet manufacturer with 3,000 employees
released 1,700 pounds of chlorine gas for 13 minutes.
An operator attempted to transfer liquid chlorine from a
rail car to an evaporator building and the transfer hose
failed. An investigation of the incident revealed that the
hose was Inappropriate for chlorine transfer operations.
Consequently, all chlorine handling hoses are now
specifically identified "For Chlorine Use Only." Also, the
facility has installed a remote actuated chlorine shut-off
valve and has conducted a hazard evaluation on the
entire chlorine handling system.
Conducting a hazard evaluation is a very positive step.
In these circumstances, the facility should also stress
that operators must always make sure the proper equip-
ment is being used. This message could be reinforced
through revised standard operating procedures and
through a training course. The facility should consider
ensuring that excess Jlow shut-off valves are used on all
rail car transfers and investigate the possibility of using
sensors, alarms, and automatic shut-off equipment
These examples begin to indicate the complexity of the acci-
dent process; In many accidents, the release is the result of a
number of factors and determining the cause as opposed to
contributing factors can be difficult. The examples also illus-
trate the kinds of steps facilities can take to prevent
recurrences.
PREVENTION INSIGHTS BASED ON ARIP DATA
The following sections present several prevention-related
insights you can use when you talk with facilities. Each
-------
insight is supported by several ARIP statistics, descriptions of
a few ARIP releases, and a list of specific questions suggested
for LEPCs to use in discussions with local facility managers.
Operator Training: A Key to Preventing Human
Error
Operator error was a primary or contributing cause in 31
percent of ARIP releases. Operator training can be a key to
preventing human error: almost half of the facilities surveyed
said they have a training program and consider it a significant
release prevention measure. At facilities where operator error
was the primary cause of the release, over 59 percent of the
facilities noted that they considered upgrading or have already
upgraded their training program. ARIP data also highlight the
important role of standard operating procedures in preventing
operator error. The following examples illustrate the value of
training and standard operating procedures in preventing
releases due to human error.
A chlorine-caustic manufacturer released 500 pounds of
chlorine gas for 20 minutes. An operator had failed to
follow procedures, and the chlorine valves in the system
were improperly set prior to the start-up of a pump. A
temperature sensor for liquid chlorine, which could
have provided some advance warning of the release, was
not properly connected to the control room. Following
the release, the facility properly connected the tempera-
ture sensor, conducted weekly tests of all control room
alarms, developed an additional written procedure to
assure that the valves are in the proper position, and
investigated the incident using the plant's formal inci-
dent investigation process.
A pulp and paper manufacturer released 100 pounds of
sulfur dioxide due to operator error. The operators were
instructed not to move certain chemical-filled containers
in wet weather, possibly because the containers that
were used were open or were not waterproof. Nonethe-
less, the chemicals were moved during a rainstorm. As
a result of the incident, the facility has re-emphasized
through additional training and practical exercises the
standard operating procedures that prohibit the trans-
port of certain chemical-filled containers in wet weather
so operators understand the rationale for the procedure.
Making sure that operators understand why a condition
is unsafe is important, but it is also important to train
workers to recognize when those hazards exist. Man-
agement should investigate alternative containers or
-------
Ilil,
enclosures to protect the chemicals from exposure to
water. Without these options, management forces opera-
tors to judge what are "wet" conditions rather than
solves the problem.
A facility with 100 employees released vinyl chloride to
the air because the operator mistakenly directed the
chemical to an incinerator that was being repaired. In
response, the facility now trains and requires the opera-
tor to follow a process checklist.
The facility could have Jurther guarded against operator
error by programming the values to close automatically,
or manually closing, lodging, and labelling them, with "Do
Not Operate" tags when any equipment is inoperable.
All employers covered by the Occupational Safety and Health
Administration (OSHA) Hazard Communication Standard
(HCS) are required to train their employees to work safely with
various classes of chemical hazards. Training includes meth-
ods and observations used to detect releases of hazardous
chemicals, the physical and health hazards of the chemicals,
and methods to protect themselves. Facility records that
document OSHA training should be available. Keep in mind
as you talk with the facility that the type of training program
used will vary with the facility. Individual training may be
more feasible or effective than group training programs for
some situations. The format of the training program is less
important than the content and rigor of the program. Training
is only effective if the operators understand what they are
being taught. Therefore, a training course must include meth-
ods of testing operators to ensure they understand both the
procedures and the reasons for the procedures. When opera-
tors understand the hazards of unsafe conditions, they will be
more likely to work to avoid those conditions.
A last caution is that training alone cannot ensure error-free
operations. In the practical world of industrial operations,
there are limits on what even a well-trained operator can be
expected to do. Part of management's responsibility is to
structure tasks so operators can perform them safely during
normal operating conditions and under stress. Managers need
to be aware of the accident prevention role of "human factors
technology" such as the design, layout, and locations of oper-
ating switches, alarms, and controls. For example, the switch
for the hazard flashers in your car has become easier to use
over the years because auto manufacturers recognized that it
is important to find and operate the switch quickly in an
emergency.
-------
TRAINING QUESTIONS
The following questions may be used to start a dialogue
with facilities in your community about facility training
programs:
Describe the training provided to workers. Which
workers receive it? Are refresher courses offered?
How often and for whom?
M Who gives the training? What are their qualifications?
How are workers tested to see if they understand what
they have been taught?
Are workers retrained when processes or operating
procedures are changed?
Equipment Failure:
Cope With It
How to Limit It and How to
Equipment failure was a primary cause of approximately 56
percent of the ARIP surveyed releases. This large percentage is
not surprising because many factors contribute to equipment
failure: lack of equipment inspections, lack of preventive
maintenance, improper operation of equipment, poor equip-
ment design, defective equipment, and age. The following
incidents resulting from equipment failure were reported
through the ARIP questionnaire.
Seventeen hundred pounds of benzene overflowed a
storage vessel during a loading operation. An alarm
system designed to alert operators to high levels within
the vessel on the tank malfunctioned. The operator had
noted and logged the tank level before loading; however,
he failed to appreciate the significance of the tank level
measurement. As a result of the incident, the facility
repaired and tested the high-level alarm system and
connected this system to the control room. The facility
has also installed safety valves that automatically shut
down the benzene unloading pump on a high-level
indication. For further assurance against a recurrence,
the tank was refitted with an overflow pipeline which
leads directly to an on-site recovery system.
A facility producing nitrogen-based fertilizers and
chemicals released ammonia vapors when a pipeline
failed. The failed pipeline was repaired and reinforced
-------
ltd,
supports were added to reduce the vibration suspected
of causing the failure.
The facility should also consider a periodic inspection
program to identify trouble spots before failures.
A petroleum refinery with over 1,500 employees re-
leased almost 200 pounds of hydrogen sulfide into the
air. The release was the result of a power failure and a
resulting over-pressurization in an operating unit,
possibly causing a relief valve or rupture disk to open.
The facility installed a new alarm system to alert opera-
tors that a back-up power supply (batteries) has been
activated when a power failure occurs. The alarm
warns that operations personnel must secure the unit
within 30 minutes to avoid atmospheric releases.
All facilities should ensure that equipment is fail safe"
in the event of a power failure and recognize the impact
and consequences on a process unit when power failures
occur. In addition, facilities need to recognize and evalu-
ate the consequences and impacts of controls used to
prevent a hazard. In this example, relief valves or rup-
ture dislcs may have been used to prevent the unit from
rupturing if over-pressurization occurred. However,
safeguarding the unit this way leads to emergency
venting to the air. The facility then needs to evaluate if a
scrubber or other methods of pressure reduction could
eliminate releases of the substance to air.
Hazard Evaluation: Identifying Problems Before
They Happen
No one likes to think about accidents, but one way to prevent
them is to analyze what would happen if one piece of equip-
ment fails, or a chain of equipment failures and human errors
occurs. This process is called hazard evaluation. The Ameri-
can Institute of Chemical Engineers has published guidelines
for hazard evaluation techniques; these guidelines explain 11
hazard evaluation techniques ~ how they are conducted, when
they are appropriate, how long they may take, and who needs
to be involved.
The type of hazard evaluation technique used depends on the
facility and its processes. A small facility with, for example,
two vessels, a pump, and four employees can work through
one of the simpler techniques fairly quickly. A larger facility
may need several months to identify all the possible failures
8
-------
and their consequences. The important thing is that the
analysis is thorough and that possible problem areas, such as
inappropriate equipment, are addressed once the analysis is
complete.
What should be of concern to you is how the facilities identify
equipment that is likely to fail or has failed. Many accidents
occur because one piece of equipment such as a pressure
gauge fails, no one notices, and pressure builds, leading to a
release. Installing the best equipment will not solve the re-
lease problem by itself; the facility must have preventive main-
tenance procedures, make use of reliability information for
equipment, and consider what they will do if the equipment
does fail. All of these factors must be addressed in a holistic
approach to preventing accidental releases.
HAZARD EVALUATIONS AND EQUIPMENT
FAILURE QUESTIONS
Has the facility conducted a hazard evaluation? What
kind and when was it performed? What recommenda-
tions resulted from the evaluation? Were they imple-
mented?
What procedures are in place to address equipment
failure? Are workers aware of and trained in design
and operating specifications?
What does the facility consider to be critical equip-
ment?
How do they know when critical equipment has failed?
What happens if critical equipment fails?
Does the facility investigate all accidental releases?
What actions have been taken as a result of accident
investigations? Will the facility share the results with
other facilities, the LEPC, and other organizations?
Inspection and Maintenance: A Key to Preventing
Equipment Problems
Equipment inspections can spot potential releases before they
happen (e.g., thin pipe wall, loose valve). Regular maintenance
-------
Illl,
of process equipment helps to prevent equipment failure and
accidental releases. Over 25 percent of the facilities in the
ARIP database indicated that prior to the release, they main-
tained an inspections program on the equipment or systems
that caused the release. Only four percent indicated that they
performed preventive maintenance on the failed equipment or
system; preventive maintenance checks or replaces equipment
before it fails. The following examples illustrate uses of in-
spection and maintenance on process equipment to help
prevent releases.
A petroleum refinery released 2,000 pounds of hy-
drofluoric acid gas because of a pipe leak. As a result,
the facility started an annual pipe inspection program
using radiography (X-rays) to identify actual or potential
piping failures without having to disassemble the pip-
ing. The program uses computer records to prompt
inspections and gives special attention to critical piping.
A manufacturer of agricultural chemicals released 13
pounds of chlorine gas due to a leak in piping. Opera-
tions and maintenance personnel at the facility regu-
larly conduct piping inspections to identify lines and
fittings that show significant metal loss or other charac-
teristics that might lead to a failure. However, this pro-
cedure failed to prevent the leak. As a result of the
release, the piping inspections program has been up-
graded to include more frequent inspections and im-
proved documentation.
Hydrochloric acid leaking from a small crack in a pipe
elbow corroded the leg supports of a nearby drain pipe-
line that also contained hydrochloric acid. The eventual
collapse of the supports caused a one-inch opening in
the connections in the drain pipeline. Consequently,
almost 200,000 pounds of hydrochloric acid were
spilled. Henceforth, the acid pipelines will be visually
inspected once a week. Also, the facility installed a new
process line constructed of more corrosion resistant
material.
Maintenance and inspection programs are key elements of
accident prevention. Facilities should have programs for
inspecting all their equipment. For critical elements such as
pressure gauges, preventive maintenance is needed. Where
manufacturer's information or facility experience indicates that
equipment has a limited useful life, the equipment should be
replaced before it fails. Gauges and monitors should be cali-
brated periodically. The frequency of needed calibration varies
with the instrument. Some monitors hold their calibration for
months; other must be calibrated weekly. Facility experience
10
-------
or information from the manufacturer probably indicates how
frequently instruments need recalibration to ensure accuracy.
Inspection and maintenance schedules should be written and
tracked, and all maintenance should be logged so a manager
can check the written record to know what has been done and
what needs to be done. One aspect of maintenance that
LEPCs should not overlook is emergency response equipment.
Because this equipment is used infrequently, the facility
should test it as well as inspect it to ensure that the equip-
ment will function whenever it may be needed.
MAINTENANCE QUESTIONS
Does the facility have a written maintenance program,
with schedules?
Do they keep a log of when each piece of equipment is
checked and replaced?
Do they train maintenance employees on proper pro-
cedures?
Do they have a schedule for replacing critical elements
even if they are not noticeably worn?
Have they had problems with particular equipment?
If so, did they change their inspection or maintenance
schedules?
11
-------
lib.
APPENDIX
SUMMARY OF THE AMP DATABASE AND
FINDINGS FROM THE PILOT
Databases on Hazardous Chemical Accidents and
Prevention
ARSP is unique among databases concerning chemical acci-
dents because of its focus on the causes of accidents. Several
other accident databases often referred to in discussions of
accident prevention are:
The National Response Center (NRC) database, consist-
ing of release reports received under the Superfund
(CERCLA) requirement that parties report to the NRC
hazardous substance releases exceeding specified re-
portable quantities. NRC reports are used as a mecha-
nism for notifying federal On-Scene Coordinators and
determining the need for federal response.
The Acute Hazardous Events Database (AHE/DB) on the
causes and consequences of releases is compiled by
EPA from a wide variety of sources, including NRC
reports and press stories. AHE/DB provides EPA with
qualitative, largely anecdotal information about chemi-
cal accidents and does not address preventive actions.
The Emergency Response Notification System (ERNS),
compiled by EPA from reports to the NRC, the Coast
Guard, and EPA Regional Offices, is a national database
used to collect information on releases of oil and haz-
ardous substances as well as subsequent responses to
such releases. ERNS is used by EPA for enforcement
tracking and program management purposes.
The Hazardous Materials Information System (HMIS),
based on written reports transport carriers are required
to file, is the central system for hazardous materials
transportation spill data.
How Releases Are Targeted for Inclusion in the
AREP Database
The data collected in ARIP are derived from questionnaires
completed by selected facilities that have reported releases to
the National Response Center (NRC), as required by law.
12
-------
Facilities selected to receive an ARIP questionnaire have expe-
rienced a "triggered" release exhibiting one or more of the fol-
lowing characteristics:
Release quantities in excess of a multiple of the CER-
CLA reportable quantity for the chemical involved;
H Releases resulting in deaths or injuries;
Releases that are part of a trend of frequent releases
from the same facility; or
Releases involving extremely hazardous substances
designated under Title III.
The Uses of ARIP Data
The ARIP database and program have been designed to serve
multiple purposes, including:
Uncovering national trends in the chemicals and proc-
esses involved in releases, the causes of releases, and
release prevention practices used by facilities;
Identifying "accidents waiting to happen," i.e., facilities
showing a persistent pattern of small releases that may
foreshadow more severe future releases, and calling
attention of facility management to the problem;
Heightening corporate awareness and involvement in
preventing accidental releases through the conscious-
ness-raising device of the questionnaire; and
Providing LEPCs and SERCs with important information
useful both in preparing Title Ill-mandated emergency
response plans and in working with facilities to reduce
hazards through prevention.
ARIP's Limitations
Although ARIP is the best available database on the causes of,
and means of preventing, chemical accidents, it, too, has its
limitations. These limitations must be kept in mind in deriv-
ing conclusions and recommendations from ARIP. The most
important limitations of the ARIP data include:
13
-------
Illl,
ARIP is a database in its formative stages -- current
"results" are preliminary in nature because they are
based on information from only approximately 330
releases. Also, the questionnaire has been recently
expanded as a result of the Section 305[b) study. Data
from many more releases will be analyzed and included
in the database as it grows.
The screening criteria or triggers chosen to narrow the
universe of releases covered by questionnaires inten-
tionally skew the database to larger, more severe, and
more frequent releases. The designers of ARIP consid-
ered this bias sensible from an environmental and
public health standpoint.
Because the NRC database defines the universe of
releases that are subject to the ARIP screening triggers,
releases involving numerous toxic and hazardous
chemicals that are not subject to CERCLA and Title III
reporting requirements are not included in the ARIP
database.
ARIP Pilot Results
There are several tentative conclusions that can be reached
based on the information in ARE? questionnaires received to
date:
Most Frequently Released Chemicals (Exhibit 1)
The most frequently released chemicals reported in the ARIP
database are chlorine, methyl chloride, ammonia, sulfuric
add, sodium hydroxide, and hydrochloric acid. These are all
common, large-volume industrial chemicals. They are also
among the most frequent releases of hazardous substances in
the NRC and Acute Hazardous Events (AHE/DB) databases.
In fact, the six most frequently released ARIP chemicals are
the same six chemicals most frequently reported in the AHE/
DB.
Industries Most Frequently Involved in Accidental Re-
leases (Exhibit 2)
In the ARIP questionnaire, facilities identified their primary-
area of business based upon Standard Industrial Classifica-
tion (SIC) code designations. More than 89 percent of the
facilities are from the manufacturing sector (SIC codes 20-39).
Moreover, more than 69 percent of all facilities are specifically
involved in chemical manufacturing and petroleum refining
14
-------
(SIC codes 2800-2900). Consequently, ARIP results reflect the
releases of producers and industrial users rather than other,
more numerous facilities in the non-manufacturing sector.
Quantities of Chemicals Released (Exhibit 3)
Both large and small chemical releases are reported to the
NRC. The ARIP releases also vary greatly because they are
selected from a sample of the NRC releases. More than half
(177 of 332) of the reports cover releases of 1,000 pounds or
more, and over four percent of these releases were 100,000
pounds or more. On the other end of the scale, about nine
percent of the reports showed less than 10 pounds released;
those reports were for accidents selected because a death or
injury was reported or because the facility had a number of
repeat releases. The size of the release is not always a valid
measurement of the seriousness of the release. The question-
naires verified that both large and small releases can cause
extensive damage, public evacuations, injuries, and death.
Most Common Causes of Releases (Exhibit 4)
Primary causes and the whole spectrum of contributing causes
of releases are difficult to define and distinguish (e.g., an op-
erator's failure to activate the correct valve may cast doubt on
both the operator and the facility's training program). Recog-
nizing this difficulty, ARIP results show equipment failure (56
percent) and operator error (24 percent) as the most frequently
reported primary causes of releases. This seems reasonable:
causal information in the AHE/DB confirms the relative impor-
tance of these two primary causes. From a release prevention
standpoint, operator error is usually addressed with training
programs and revised standard operating procedures. Equip-
ment failure is usually addressed through regular inspections,
maintenance, equipment, design, and choice of construction
material.
Nearly 10 percent of releases were attributed to "upset condi-
tions" (e.g., overpressurization) or "bypass conditions" (e.g.,
detouring chemicals around the main chemical processing
pathway). These conditions result from a variety of causes,
including operator error, equipment failure, and poor system
design. Consequently, the measures used to prevent releases
from upset and bypass conditions depend greatly on the indi-
vidual details of the incident. A significant number (38 per-
cent) of ARIP releases were attributed to more than one cause.
Among these releases, the most commonly cited secondary
causes were equipment failure, operator error, and other
unspecified causes.
J5
-------
Illl,
In-PIant Location of Releases (Exhibit 5)
ARIP information also permits analysis of the most common
locations of releases within facilities. The data indicate that
ARIP releases occurred with roughly similar frequency in
process vessels, storage vessels, and piping (approximately 25
percent each). Releases from valves are slightly less frequent
(11 percent). Data on release location in the AHE/DB indicate
a similar distribution of releases from process vessels, storage
vessels, and piping/valves. Consequently, an effective release
prevention program must examine all the parts and equipment
involved in the facility's chemical processing.
Process Status When Releases Occur (Exhibit 6)
ARIP explores the relationship between the incidence of acci-
dental releases and the different phases of a chemical process.
Typically, chemical facilities use either a batch process or a
continuous process. In a batch process, the raw chemicals are
basically added to one or several containers (e.g., vats) or proc-
essing equipment (e.g., evaporator, distiller) and then sub-
jected to specific temperatures, pressures, and agitations for a
discrete time to produce a batch of the desired product. Paint
manufacturers use batch processes to produce different col-
ored paints. In a continuous process, the reactants continu-
ously flow through a series of processing equipment to pro-
duce a continuous flow of product. Petroleum refineries use
continuous processes. Separate from these process operations
are the procedures to load the raw chemical reactants from a
supplier (e.g., truck, railcar) to the facility's storage and to
unload the chemical products to the shipper. Also, both batch
and continuous operations require process start-up and shut-
down procedures. Releases can occur during any of these
procedures as well as during the operations mode (e.g., appli-
cation of heat, mixing) of the batch or continuous processes.
ARIP also examines the incidence of releases when mainte-
nance is being performed on process equipment.
The largest number of ARIP releases (153 cases or 45%) occur
during continuous processes. The wide variety of continuous
processes used by the chemical industry and consequently,
the resulting diversity and integration of process equipment
may be responsible for this high incidence of accidental re-
leases. A significantly lower number of releases (34 or 10%)
occur during batch processes. More than 16 percent of re-
leases (56 cases) occur during loading or unloading proce-
dures. A similar percentage (14% or 46 cases) of releases
16
-------
occur during either process start-up or shut-down. Chemicals
released while maintenance is performed account for less than
five percent.
17
-------
PILOT ARIP RESULTS
Exhibit 1
MOST FREQUENTLY RELEASED CHEMICALS
Methyl Chloride
(31 releases)
\
Ammonia
(28 releases)
Sulfuric Acid
(26 releases)
Other Chemicals
(132 releases)
Sodium Hydroxide
(16 releases)
Hydrochloric Acid
(13 releases)
Ethylene Oxide
(13 releases)
Vinyl Chloride
(11 releases)
Toluene
(10 releases)
Creosote
(9 releases)
Exhibit 2
INDUSTRIES MOST FREQUENTLY INVOLVED
IN ACCIDENTAL RELEASES
Non-Manufacturing
Types
(36 cases)
Other
Manufacturing
(65 cases)
Chemical Manufacturing
and Petroleum Refining
SIC Codes 28 and 29
(230 cases)
-------
Number
of
Releases
90 -r
80 --
70 --
60 --
50 --
40 --
30 --
20 --
10 --
PILOT ARIP RESULTS
Exhibit 3
QUANTITIES OF CHEMICALS RELEASED
28
54
70
90
72
15
Less Than 1 I 10-99 I 1,000-9,999 | 100,000-999,999
1-9 100 - 999 10.000 - 99,999
Quantity of Release (Pounds)
Note: Quantities were not given for five releases.
-------
PILOT ARIP RESULTS
Exhibit 4
MOST COMMON CAUSES OF RELEASES
Equipment Failure
(187 cases)
Upset Condition
(26 cases)
Bypass Condition
(4 cases)
Fire
(1 case)
Exhibit 5
IN-PLANT LOCATION OF RELEASES
Storage Vessel
(88 cases)
Process Vessel
(80 cases)
Other
(49 cases)
Piping
(82 cases)
Valves
(38 cases)
-------
PILOT AMP RESULTS
Exhibit 6
PROCESS STATUS WHEN RELEASES OCCUR
Continuous Operations
(153 cases)
Other
(32 cases)
Loading
(36 cases)
Batch
Operations
(34 cases)
Start-up
(18 cases)
Unloading
(20 cases)
Process
Shut-down
(16 cases)
Maintenance
(16 cases)
-------
Ilil.
CHEMICAL EMERGENCY PREPAREDNESS AND
PREVENTION TECHNICAL ASSISTANCE
BULLETINS
Introduction to Exercises in Chemical Emergency
Preparedness Programs
A Guide to Planning and Conducting Table-Top Exercises
A Guide to Planning and Conducting Field Simulation
Exercises
Report on a Conference on Risk Communication and
Environmental Management
Successful Practices in Title III Implementation
Tort Liability in Emergency Planning
Why Accidents Occur: Insights from the Accidental
Release Information Program
For additional copies, write:
Emergency Planning and Community Right-to-Know
Information Service
U.S. Environmental Protection Agency
OS 120
401 M St., SW
Washington, DC 20460
EPA welcomes comments and suggestions for topics for future
bulletins. Please send your comments and suggestions to:
Co-ordinator, Technical Assistance Bulletins
Chemical Emergency Preparedness and Prevention Office
U.S. Environmental Protection Agency
OS 120
401 M St., SW
Washington, DC 20460
------- |