4
Implications of Past Chemical Events for Ongoing and Future Chemical Demilitarization Activities

Chapters 1 through 3 of this report are based on an examination of activities at Johnston Atoll Chemical Agent Disposal System (JACADS) and Tooele Chemical Disposal Facility (TOCDF), both of which employ baseline incineration systems to destroy chemical agents. Third-generation incineration facilities are scheduled to begin operation in 2002 or 2003 at Anniston, Alabama, Umatilla, Oregon, and Pine Bluff, Arkansas. The committee believes that many of the observations and recommendations made in this report are applicable to all demilitarization facilities, including those that may not use incineration.

Evidence indicates that chemical demilitarization incineration facilities are safe as designed if they are operated properly and if the appropriate operating procedures and protocols are in place (NRC, 1996). The avoidance of risk during any type of process upset depends on having the necessary engineering controls in place and on the operator’s skill and training in using them to advantage. This level of preparedness requires in turn that a thorough hazard risk analysis be performed and that all personnel be thoroughly trained and given refresher courses at appropriate intervals. At both JACADS and TOCDF, extensive written procedures are in place for normal operations as well as for startup and shutdown, and operators receive systematic refresher training in these procedures. It can be expected that future chemical demilitarization facilities will also operate this way. Key factors for minimizing—if not eliminating—chemical events in the future include:

  • Sound risk and change management programs and procedures;

  • Effective safety programs that are focused on continuous improvement, and have the full visible support of all levels of management; and

  • Systems for efficient and timely program-wide dissemination of information and communication.

RISK AND MANAGEMENT OF CHANGE PROGRAMS ALREADY IN PLACE

This section describes the procedures that are in place for evaluation of change, including the risk associated with a change. The current Chemical Stockpile Disposal Program (CSDP) risk management program is fully described in Risk Assessment and Management at Deseret Chemical Depot and the Tooele Chemical Agent Disposal Facility (NRC, 1997). It is a multilevel program that defines policy, sets requirements, provides guidance on implementation, and, at the facility level, defines specific requirements the facility must meet and specific management processes that must be implemented. The CSDP risk management program is based on a long history of safety and hazard analysis and regulation by the Army. An informal risk management process was developed at the TOCDF in parallel with the site-specific quantitative risk analysis (QRA). This process was described in the NRC report Review of Systemization of the Tooele Chemical Agent Disposal Facility (NRC, 1996), which summarized a number of plant and operational changes that had been implemented as a result of accident scenarios identified in preliminary work on the QRA. As part of the risk management process, the following risk-monitoring activities have been introduced:

  • Performance evaluation (based on feedback from activities and incidents);

  • Emergency response exercises (periodic exercises on site, with Chemical Stockpile Emergency Preparedness Program (CSEPP) personnel);

  • Risk tracking (as new data become available, as risk models are improved, and when changes occur in the facility, the related changes in risk related to safety, environmental protection, and emergency preparedness will be calculated and tracked); and



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 37
4 Implications of Past Chemical Events for Ongoing and Future Chemical Demilitarization Activities RISK AND MANAGEMENT OF CHANGE PROGRAMS Chapters 1 through 3 of this report are based on an ex- ALREADY IN PLACE amination of activities at Johnston Atoll Chemical Agent Disposal System (JACADS) and Tooele Chemical Disposal Facility (TOCDF), both of which employ baseline incinera- This section describes the procedures that are in place tion systems to destroy chemical agents. Third-generation for evaluation of change, including the risk associated with a incineration facilities are scheduled to begin operation in change. The current Chemical Stockpile Disposal Program 2002 or 2003 at Anniston, Alabama, Umatilla, Oregon, and (CSDP) risk management program is fully described in Risk Pine Bluff, Arkansas. The committee believes that many of Assessment and Management at Deseret Chemical Depot the observations and recommendations made in this report and the Tooele Chemical Agent Disposal Facility (NRC, are applicable to all demilitarization facilities, including 1997). It is a multilevel program that defines policy, sets those that may not use incineration. requirements, provides guidance on implementation, and, at Evidence indicates that chemical demilitarization incin- the facility level, defines specific requirements the facility eration facilities are safe as designed if they are operated must meet and specific management processes that must be properly and if the appropriate operating procedures and pro- implemented. The CSDP risk management program is based tocols are in place (NRC, 1996). The avoidance of risk dur- on a long history of safety and hazard analysis and regula- ing any type of process upset depends on having the neces- tion by the Army. An informal risk management process was sary engineering controls in place and on the operator’s skill developed at the TOCDF in parallel with the site-specific and training in using them to advantage. This level of pre- quantitative risk analysis (QRA). This process was described paredness requires in turn that a thorough hazard risk analy- in the NRC report Review of Systemization of the Tooele sis be performed and that all personnel be thoroughly trained Chemical Agent Disposal Facility (NRC, 1996), which sum- and given refresher courses at appropriate intervals. At both marized a number of plant and operational changes that had JACADS and TOCDF, extensive written procedures are in been implemented as a result of accident scenarios identified place for normal operations as well as for startup and shut- in preliminary work on the QRA. As part of the risk manage- down, and operators receive systematic refresher training in ment process, the following risk-monitoring activities have these procedures. It can be expected that future chemical been introduced: demilitarization facilities will also operate this way. Key factors for minimizing—if not eliminating—chemical events • Performance evaluation (based on feedback from in the future include: activities and incidents); • Emergency response exercises (periodic exercises on • Sound risk and change management programs and site, with Chemical Stockpile Emergency Prepared- procedures; ness Program (CSEPP) personnel); • Effective safety programs that are focused on con- • Risk tracking (as new data become available, as risk tinuous improvement, and have the full visible sup- models are improved, and when changes occur in port of all levels of management; and the facility, the related changes in risk related to • Systems for efficient and timely program-wide dis- safety, environmental protection, and emergency semination of information and communication. preparedness will be calculated and tracked); and 37

OCR for page 37
38 EVALUATION OF CHEMICAL EVENTS AT ARMY CHEMICAL AGENT DISPOSAL FACILITIES • As required by the Program Manager for Chemical training, or other aspects of the program. This established Demilitarization (PMCD) now for essentially all fa- configuration is based on the initial design of the facility and cilities, participation in meetings and/or teleconfer- incorporates changes that have been approved and imple- ences about design lessons learned and program- mented. The established configuration is the basis for the matic lessons learned. plant’s up-to-date health risk assessment (HRA) and QRA. If a proposed change is significant, assessing its value is The Army’s formal risk management process is de- acknowledged to be both a policy question and a factual scribed in a program-wide document, Chemical Agent Dis- question. Structured discussions focus attention on all fac- posal Facility Risk Management Program Requirements tors that affect the decision, and information on the impact (U.S. Army, 1996c), which provides a basis for the CSDP of the proposed change in significant cases should be made risk management program. The risk management program is available to the public, to the CAC, and to state regulators, a framework for understanding and controlling all elements and public comments should be solicited when appropriate of risk within the disposal facility and the stockpile storage to the change contemplated. For the most significant changes area. It links risk management needs to other specific re- (Resource Conservation and Recovery Act (RCRA) Class 3) quirements of the Army and other parties at top levels of the Army, with the assistance of the controlling regulatory management and identifies specific documents and refer- body, must schedule a public hearing. The definition of a ences that apply to all CSDP facilities. RCRA Class 3 change is embodied in the existing federal regu- In January 1997, the Army issued its draft, A Guide to lations. The Army’s decision will take into account commu- Risk Management Policy and Activities (the Guide) (U.S. nity desires (where appropriate to the complexity of the Army, 1997b). This draft provides an overview of the pro- change) and needs as well as important facts and intangible cesses for managing risks associated with Program Manager factors, which are summarized in Table 4-1. Note that factor for Chemical Demilitarization (PMCD) activities and de- 6 in Table 4-1, “comparison to previous decisions,” ensures scribes a process for managing changes that may affect the either that decisions are consistent or that the reasons for in- risk associated with PMCD activities. It defines issues that consistencies are clearly stated. A thorough consideration of are matters of risk assessment and issues that are matters uncertainties is also required. The Army is tasked to prepare involving policy (value judgments) and attempts to establish an approach to integrating them and to involving the public in that integration. TABLE 4-1 Issues and Factors in Assessing the Value of The PMCD policy indicates that risk management is Change Options integrated into the normal functioning of the organization: 1 Public Input • Operations are now based on the risk management program requirements document (U.S. Army, 1996c). 2 QRA Risk • The Risk Management and Quality Assurance Office a. All available QRA risk measures, including expected fatalities, cancer incidence, fatalities at a one-in-a-billion probability, and has been assigned the task of integrating risk manage- probability of one or more fatalities ment for operations, design, and construction. b. Risk trade-offs: public versus worker, individual versus societal, • The Environmental and Monitoring Office has been processing versus storage assigned the task of assessing hazards to the envi- c. Uncertainties in the technical assessment of risk ronment, the populace, and biota in terms of regula- d. Insights from sensitivity studies tory requirements. 3 Hazard Evaluations • The CSEPP has the task of planning for potential emergencies and providing liaisons with other emer- 4 HRA Risk gency preparedness organizations. Note that this pro- a. Insight from sensitivity studies gram is not a part of PMCD. 5 Programmatic • The Public Affairs Office is charged with providing a. Cost of the change relative to other proposals and program liaisons among the public, the Citizens Advisory objectives Commission (CAC), state authorities, and the Army b. Schedule for implementation to facilitate public involvement. c. Uncertainties in estimates d. Impact of implementation on overall objectives and schedule for disposal of the weapons and chemical agent Another significant element in risk management is the e. Consideration of the improvement anticipated by this change management of change. Although changes are presumably with other proposed improvements made for good reasons, the overall safety of the facility could be compromised if the effects of change on risk estimates 6 Comparison to previous decisions are not evaluated or understood. Changes need to be docu- mented and analyzed to determine if they affect procedures, SOURCE: Reprinted from U.S. Army (1997b), p. 53.

OCR for page 37
39 IMPLICATIONS OF PAST CHEMICAL EVENTS FOR ONGOING AND FUTURE CHEMICAL DEMILITARIZATION ACTIVITIES Monthly Rate 12 month average 10.0 9.0 8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0 Dec-96 Jun-97 Dec-97 Jun-98 Dec-98 Jun-99 Dec-99 Jun-00 Dec-00 Jun-01 Dec-01 FIGURE 4-1 TOCDF recordable injury rate 12-month rolling average, August 1996 (the start of agent operations) through December 2001. SOURCE: Data up to 1998 from NRC (1999a); data for 1999 to 2001 from U.S. Army (2002b). responses to all public comments and inform regulators and occupational health and safety programs and performance, the CAC of decisions and their rationale. and it has urged plant management to lead the operating sites If all issues are considered in an appropriate and timely toward a “safety culture” (NRC, 1999a). manner, general consensus may be possible. But even if con- At JACADS, significant progress was made in develop- sensus is not reached, the Army, as decision maker, will ing a safety culture, and during the latter phases of demilita- provide a “synopsis of the considerations and a summary of rization operations the plant was consistently achieving ex- the overall decision basis, listing the rationale for each fac- cellent safety performance. This does not appear to have tor” (U.S. Army, 1997b). In this way, interested parties can been the case at TOCDF (NRC, 1999a). see if their concerns were considered and what effect they Although traditional performance indicators such as re- had on the decision. cordable injury rates (RIRs) at TOCDF are comparable to all- industry averages, there has been very little improvement in these metrics since operations began (Figure 4-1). Nor is there SAFETY PROGRAMS an indication that TOCDF has moved toward a safety culture The safety of the public, the environment, and workers at any appreciable rate, even though management has devel- is a very significant part of a congressional mandate for the oped a TOCDF safety culture plan and has implemented sev- conduct of the chemical demilitarization program. The eral programs aimed at achieving the safety plan’s goals (NRC, NRC’s Stockpile Committee previously expressed concern 1999a). No additional findings or observations resulted from over production (agent destruction) having a higher priority this study. The Committee on Evaluation of Chemical Events than safety—at least from the standpoint of the contractors’ at Army Chemical Agent Disposal Facilities concurs with the award fee criteria (NRC, 1999a and 2002). Responding to Stockpile Committee in observing that the TOCDF is being this observation, the Army revised the criteria to emphasize operated in a safe manner, but that it can and should be con- safety and production equally. An additional concern ex- tinuously improving its safety programs and performance. pressed repeatedly by the Stockpile Committee is a preoc- The Committee on Evaluation of Chemical Events also cupation with agent safety, to the detriment of traditional concurs with the Stockpile Committee in its belief that future

OCR for page 37
40 EVALUATION OF CHEMICAL EVENTS AT ARMY CHEMICAL AGENT DISPOSAL FACILITIES PLL PROGRAM DATABASE demilitarization facilities should be safer at start-up, as evi- denced by performance metrics, than their predecessors. The PLL program is a mechanism for developing and Such performance should not be difficult to achieve, given maintaining information associated with lessons related to an effective programmatic lessons learned (PLL) program preconstruction and construction activities, systemization, and the fact that several managers with chemical demilita- operations, and closure of the chemical demilitarization fa- rization experience will be working at the newer sites. cilities. The majority of lessons learned are captured in the Management and employees at new sites must begin the PLL database, which contains considerable information process of establishing a safety culture before operations and is potentially an excellent resource for helping to main- commence. tain a high level of operational safety and security. How- ever, so much information is present that plant personnel PROGRAMMATIC LESSONS LEARNED PROGRAM believe it is hard to identify what will be helpful in any given situation. The PLL program—the principal means of communi- The information in the PLL computerized database is cating lessons learned both within and among the various available to all participants in the PLL program. The data- chemical demilitarization facilities—is the PMCD’s only base is searchable using both Boolean logic (and key significant vehicle for communicating and coordinating word(s)) and a decision tree. Although other means of com- risk, design, and operational issues among sites. The PLL munication exist for discussing the operational and safety program until recently was administered by PMCD with issues arising at the demilitarization facilities (described support from Science Applications International Corpo- below), essentially all the information is contained in the ration (SAIC). The program manager at SAIC was hired database. The data are continuously updated and include specifically because of his background in and extensive information from workshops since 1994 and document re- familiarity with detailed operating procedures, training, views before that date. Not all PLL program components and quality control in a hazardous and demanding envi- that lead to data included in the database were in place in ronment. 1994, and some have been improved since their inception. Dr. Mario Fiori, Assistant Secretary of the Army for For example, workshops, critical document reviews, quick Installations and Environment, presented his vision of some reacts, and the PLL oversight board were initiated in 1994; changes in the management and operating philosophy for the technical bulletin, in 1995; and operational assessments, the chemical demilitarization program to the Stockpile Com- in 1996. mittee on June 29, 2002. A major thrust of his presentation The issues database was first provided to the chemical was that the contractors need to take “ownership” of the vari- demilitarization sites in 1997, the programmatic planning ous aspects of the program for which they are responsible. documents became available in 1997, the site documents Included in this change is the concept that PMCD would no comparison began in 1998, the directed action philosophy longer be directly responsible for the PLL program, but that was revised in 1999, the engineering change proposal (ECP) a contractor (yet to be selected from the two operating con- review process (which began in 1987) was integrated with tractors) would instead be responsible for it. PLL in 1999, and the lessons learned database (a different The philosophy and purpose of the PLL program are: way of sorting and accessing the information) was started in 2000-2001. The PLL team has also developed a help line to • to capture lessons learned during construction, facilitate easier use of the information. equipment installation, systemization, operations, Data in the PLL database are accessible to the following and closure, i.e., all phases of the operation staff: • to provide assistance to the sites and PMCD in as- sessing and utilizing these lessons and experiences 1. PMCD home office, which includes stockpile dis- • to support PMCD’s emphasis on safety and environ- posal, alternative technologies, non-stockpile mate- mental compliance riel, cooperative threat reduction, support offices, • to reduce cost and schedule and contractors. • to provide information to decision makers. 2. Project Manager Chemical Stockpile Disposal (PMCSD) and Project Manager for Alternative Tech- The PLL program is a comprehensive, multicomponent nologies and Approaches (PMATA) sites, which in- activity that is distributed across all PMCD demilitarization clude field offices and site systems contractors. sites and includes workshops, assessments, technical bulle- 3. Other stakeholders, including operations support tins, directed actions and updates, programmatic planning command, U.S. Army Corps of Engineers, U.S. De- documents, site document comparisons, critical document partment of Health and Human Services, U.S. Army reviews, and a “quick react” feature (Box 4-1). Center for Health Promotion and Preventive Medi-

OCR for page 37
41 IMPLICATIONS OF PAST CHEMICAL EVENTS FOR ONGOING AND FUTURE CHEMICAL DEMILITARIZATION ACTIVITIES BOX 4-1 Additional PLL Program Components In addition to the computerized PLL database, the PLL program tion management plan, guide to systemization planning, has several other significant components, including workshops, as- guide to emergency response planning, and guide to clo- sessments, technical bulletins, directed actions and updates, program- sure planning (the last in draft form as of October 2001). matic planning documents, site document comparisons, critical docu- The PLL team incorporates into these documents the re- ment reviews, and a “quick react” feature. A brief description of these sults of lessons learned, adds new requirements (includ- components follows. ing applicable regulatory requirements), and comments on the cost/benefit considerations of producing a new ver- • Workshops enable communication between and among sion of any document. • Site documents comparison involves PLL team review of PMCD personnel (including sites) and are the basis for information that ultimately is included in the PLL program new documents prepared at a site and comparison to pre- database. Facilitated by a person knowledgeable about viously approved documents from an earlier site and to the issues (but not a decision maker) and usually from programmatic guidance. The PLL team provides com- either PMCD or SAIC, the workshops are essentially fo- ments and recommendations, but implementing them is cused technical meetings held in person or via teleconfer- not mandatory at the site level. • Critical document review is done by the PLL team for docu- ence or videoconference. Sample workshop topics include incinerators and secondary treatment support systems; ments provided by the government, including reports of general operations maintenance and training; personal unusual occurrences or events, safety reviews (including protective systems; environmental, laboratory, and moni- near-miss advisories), reports of nonconformance or non- toring procedures; safety, surety and security; quality as- compliance, reports of test results, audit surveillance in- surance/quality control (QA/QC); construction; systemiza- spection reports, daily and weekly operating reports, and tion; public outreach; trial burns; and information man- campaign reports. The purpose of these reviews is to iden- agement systems. tify lessons that will be added to the PLL database, update • Assessments, relatively detailed studies of an issue such the database as needed, introduce appropriate topics for as management or very technical topics, are intended as a discussion at workshops and, if needed, recommend di- means of rapidly starting an effort. A topic is developed by rect actions to secure further information. • Quick react i nvolves passing critical information — the government, and SAIC follows up in planning and ex- ecution with appropriate teaming partners (approved by “changes to processes or equipment that affect operational the government). safety, [or] environmental protection, or have the possibil- • Technical bulletins are published quarterly and about 1000 ity to cause substantial equipment damage” quickly to the copies are sent to various stakeholders. Each site makes other affected parties. The site project manager or the chief additional copies as needed. The bulletins contain infor- of the operations division is responsible for and empow- mation that is not in the “quick react” category (discussed ered to designate a lesson as critical. The time frame for a below) but requires attention before the next workshop, quick react is 24 hours. The quick react process consists information that is of general interest but is not likely to be of the following: a workshop topic, or in some cases information that 1. The site project manager (or the chief of the opera- supplements the workshop discussions. tions division) designates an issue as quick react. • Directed actions and updates transfer information or re- 2. The site faxes the information to PMCD and the PLL quest that it be sent from or to the chemical demilitariza- team (using a specific, designated form). tion sites. Originating primarily in the workshops and/or 3. The site calls the chief of the operations division and quick reacts, directed actions and updates can also come the PLL contractor staff (who have 24-hour pagers). from the PLL oversight board, critical document reviews, 4. The PLL team then conducts a data search and ob- and other similar activities. The directed actions are as- tains any needed backup data, provides recom- signed by PMCD managers and tracked by the PLL team mendations(s), faxes the government decision, con- until they are acted on. The responses are reviewed by the firms receipt, puts the data into the database, and PLL team and incorporated into the PLL database. tracks the directed actions. • Programmatic planning documents that are maintained by the PLL team include a chemical demilitarization opera- The actual course of action is determined by PMCD opera- tions manual, PMCD management plan, PMCD informa- tions management.

OCR for page 37
42 EVALUATION OF CHEMICAL EVENTS AT ARMY CHEMICAL AGENT DISPOSAL FACILITIES cine, U.S. Army Materiel Systems Analysis Activ- operation. The committee found no accurate means for as- ity, Edgewood Chemical and Biological Center, and sessing this assertion other than the use of anecdotal infor- regulators. mation. When queried, some operators were unaware of the database and its uses. As of August 2001 the database was organized as an SAIC is in the process of prioritizing the data so that the “issues” database and included about 3200 items from which highest-priority issues will require a response from the site. users can choose to determine lessons applicable to their At present, a site does not have to respond, since there are particular problem. Currently the PLL team (SAIC) is devel- too many issues in the database relative to staffing levels at oping a new way to present the data and estimates that 5000 the site. Additionally each ECP approved by any site is dis- lessons will ultimately be available from the issues repre- cussed at a biweekly ECP review teleconference. At a sub- sented to date. sequent teleconference the sites inform the PLL team of Although not specifically categorized as such, a signifi- what action will be taken regarding the ECP. These appear cant number of laboratory issues are included in the data- to be among the few issues that are handled in this more base. Until recently JACADS provided most of the cases, structured manner. The ECP review process consists of the but TOCDF is now providing most of the issues. The data- following steps: base also includes lessons from Anniston, Pine Bluff, Aber- deen, Newport, and Umatilla, all of which are currently un- 1. The sites approve the ECPs and forward approved der construction or undergoing systemization. The PLL team ECPs to the PLL team. categorized these lessons as design, 341; systemization, 687; 2. The PLL team researches related issues and ECPs operations, 843; and closure, 241. Of these, 196 are catego- (using the database) and sends the ECPs and accom- rized as maintenance and 202 as training lessons. Prior to panying information to the other sites. 1999, the ECPs were handled in a separate manner, but all 3. The ECP review team, which includes representa- ECPs have now been captured in the database. Permitting tives of the PMCD office, demilitarization sites, issues are also included in the database. Army Corps of Engineers, and the PLL team, con- When the PLL program began in 1994-1995, the major ducts biweekly teleconferences and puts the deci- source of issues was the review of documents (event reports, sion documentation into the database. end-of-campaign reports, inspection reports, and so on.). Now most of the information comes from the facilitated The PLL database and PLL concept reflect a systematic workshops run by the PLL program, which allow input and effort to take advantage of lessons learned in one chemical peer review by multiple program personnel with expertise in demilitarization facility and use the information at another the subjects under discussion. The initiators of the informa- facility. At present only one facility is operating (TOCDF), tion (subjects) are primarily the chemical demilitarization and one is undergoing closure (JACADS). As more facilities sites, but some issues come from other program participants. come on line it will be more difficult to track the data and As currently operating, the decision process used to deter- ensure that the most important issues are addressed at all sites. mine the ultimate content of the PLL database is as follows: PMCD will need to strengthen the communication and imple- menting mechanisms in the near future. PMCD (SAIC) is 1. PMCD approves the list of topics (subjects) used at currently developing a set of criteria for prioritizing the infor- a facilitated workshop. mation in the PLL database. The intent is to create a few 2. Twice a year the PLL team holds workshops for en- categories of issues (lessons), sorted according to relative vironmental and environmental oversight topics. importance. For instance, those with the highest priority (for 3. The minutes from the workshop are prepared, re- example an important, operational safety directive) should viewed, and tentatively approved by SAIC. These probably be available and implemented at all facilities. minutes are sent to PMCD for its review and approval. 4. PMCD then makes the final decision before the min- RESULTANT CHANGES utes and lessons learned are entered into the data- base. At JACADS and TOCDF, operations personnel did not 5. The database is distributed as a CD-ROM to each appear to generalize lessons learned beyond the immediate chemical demilitarization site to be loaded onto its equipment and task in the original incidents. There is room local area network. It is not available on the Internet for making much wider use of these valuable lessons, such or on a wide-area network. as by “mining” the information in the PLL database to de- tect patterns that may underlie several incidents. The effort There is no mechanism to track the use of the data, but to prioritize the data is a good start toward increasing the SAIC stated to the committee that use of the data is exten- information’s usefulness. PMCD could also make better sive at the engineering change proposal (ECP) level, as well use of information available from industries such as the as at the chemical demilitarization sites during start-up and chemical and petroleum manufacturing sectors. Both have

OCR for page 37
43 IMPLICATIONS OF PAST CHEMICAL EVENTS FOR ONGOING AND FUTURE CHEMICAL DEMILITARIZATION ACTIVITIES very active trade associations and routinely share informa- • The deactivation furnace system (DFS) cyclone is tion regarding safety procedures and good operating and contained in an enclosure that is monitored by an maintenance practices among different companies. ACAMS. The destruction of chemical weapons was first begun at • There is a carbon filter on the incinerators’ exhaust. JACADS and its design was based on equipment and proce- • As a result of the JACADS waste-bin event (dis- dures developed at the Chemical Agent Munitions Disposal cussed in detail in Chapters 2 and 3), drip trays have System (CAMDS) at Deseret Chemical Depot, Utah. Many been added to rocket and mine lines, a search is on design changes were made after operations had begun, some for combustible spill pillows, and spill pillows will in response to chemical events, but most to correct recognized generally be treated in the metal parts furnace, not problems with the original design. (Both types are included in the DFS. the PLL database.) For all operating chemical facilities, de- • The large isolation valves on the individual heating, sign changes are part of a continuing process aimed at taking ventilation, and air conditioning (HVAC) carbon fil- advantage of lessons learned from ongoing operations, new ter banks now have a small “bleed” valve, connect- technology as it is developed, or better procedures developed ing to the exhaust flow, to maintain the filter bank at at a plant or transferred from another facility. negative pressure even if it is temporarily out of ser- Many design changes have also been made to improve vice and to prevent migration of agent from the filter productivity (e.g., inclusion of the hot slag withdrawal sys- bank to the connecting vestibule. (Such migration tem on the liquid incinerator (LIC) secondary burner, and of agent has been a problem in the past.) the process, currently under review, for freezing the M2A1 projectiles at Anniston before disassembly to minimize spill- Chapter 1 discusses the systems hazards analysis (SHA) ing, and subsequent cleanup, of mustard agent). Design performed for TOCDF. A primary purpose of a standard changes to improve operating safety, however, are not as hazardous operation (HAZOP) analysis is to learn to antici- readily identified except in direct response to a chemical pate where safety may be compromised. There have been event. For instance, the airflow systems handling ventilation many changes to the original design (see footnote 1, Chapter throughout a plant as well as combustion air will have vari- 5), some identified above and all included in the PLL data- able-speed motors driving the fans, allowing improved con- base. It is not apparent that each of these design changes has trol of airflow, particularly at low rates (combustion airflow been subjected to the appropriate level of HAZOP analysis. control was a problem for the operator during the May 8-9, In view of the challenging nature of the chemical weapons 2000, incident at TOCDF). disposal program and its perceived potential for harm, this A large number of changes have been made to operating aspect of the design process needs particular and ongoing procedures and equipment in response to the PLL program attention. and based on incident reports from JACADS and TOCDF. It is common practice in industry for people who do the Of the 24 recommendations for change resulting from the design and initial HAZOP analysis to be included on the May 8, 2000, event at TOCDF, for example, all have been plant start-up team. The people who did the actual detailed examined, although not all have required action at the newer design work and participated in the HAZOP studies done as plants because of differences in the feed mix and in the plant a part of the design process should also play a strong role in designs. In the committee’s view, some of the more signifi- operator training in the use(s) of the HAZOP procedures and cant changes made in response to the PLL are as follows: information. It is also common industry practice for compa- nies to share nonproprietary information about safety issues, • Staggered automatic continuous air monitoring sys- operating procedures, HAZOP findings, and so on. PMCD tem (ACAMS) monitors are now being installed in could make better use of the experiences of other industries, exhaust ducts, to shorten the time for detection of such as the chemical and petroleum refining industries, in any release of agent. the benchmarking of its procedures and processes.