3
Selected Pueblo Chemical Agent Destruction Pilot Plant Design Issues

This chapter presents a discussion of selected design issues that the committee has identified and that it believes require further study and possible changes.1 In actuality, the Army and its contractors may already be making modifications and pursuing changes based partially on questions that the committee asked during meetings with the Army and its contractors. Also, it must be emphasized again that this review is based largely on the initial design for the Pueblo Chemical Agent Destruction Pilot Plant (PCAPP). Any information included on the intermediate design was obtained from notes taken during attendance by committee members at the May 19–21, 2004, intermediate design review meeting held at the offices of the contractor, Bechtel National, Inc., in San Francisco.

SIZING OF THE FACILITY

The Bechtel Pueblo team used iGrafx™ software to estimate the material flow in the PCAPP design and to determine the number of units and capacity of each unit operation within the total plant design. The model assists the design team in better understanding and evaluating design options and determining the rate-limiting steps. For example, this event-driven model can vary the input and output rates of items (e.g., munitions) in each step in the process, including the probabilities of encountering problems. Calculations include excursions in boxed munitions handling, leakers and rejects, munitions dismantling, agent washout, agent hydrolysis, energetics hydrolysis, metal parts treatment, biological treatment, and brine reduction. Because several of these processes are batch processes while others are continuous, the model can be used to evaluate buffer storage requirements between various operations. Buffer storage is needed for situations in which single-unit failure would limit throughput or shut down the plant completely, and to ensure that the desired throughput can be maintained.

The analysis indicates that the rate-limiting activity in the process design is the munitions disassembly step. As a result, in order to maintain the desired availability of munitions, the front-end process has a separate disassembly station for each of the three types of munitions to be treated. There is an unequal number of each type, with the number of 4.2-inch mortars being about one-third the number of either the 105-mm or the 155-mm projectiles. Therefore, after completion of the disassembly of all 4.2-inch mortars, including leakers, the 4.2-inch mortar line will be modified to disassemble the remaining types of leaking projectiles by retooling the projectile/mortar disassembly (PMD) station. However, Army safety regulations may require that all three PMD machines be shut down if maintenance is being performed on one, and thus the effective throughput capacity for the disassembly step would be reduced.

Finding 3-1. The rules for personnel safety during equipment repairs in the explosion containment rooms

1  

Unless otherwise indicated, background material in this chapter is drawn from U.S. Army, 2004b, and information on the intermediate design is from the PCAPP intermediate design review in San Francisco, May 19–21, 2004.



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 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant 3 Selected Pueblo Chemical Agent Destruction Pilot Plant Design Issues This chapter presents a discussion of selected design issues that the committee has identified and that it believes require further study and possible changes.1 In actuality, the Army and its contractors may already be making modifications and pursuing changes based partially on questions that the committee asked during meetings with the Army and its contractors. Also, it must be emphasized again that this review is based largely on the initial design for the Pueblo Chemical Agent Destruction Pilot Plant (PCAPP). Any information included on the intermediate design was obtained from notes taken during attendance by committee members at the May 19–21, 2004, intermediate design review meeting held at the offices of the contractor, Bechtel National, Inc., in San Francisco. SIZING OF THE FACILITY The Bechtel Pueblo team used iGrafx™ software to estimate the material flow in the PCAPP design and to determine the number of units and capacity of each unit operation within the total plant design. The model assists the design team in better understanding and evaluating design options and determining the rate-limiting steps. For example, this event-driven model can vary the input and output rates of items (e.g., munitions) in each step in the process, including the probabilities of encountering problems. Calculations include excursions in boxed munitions handling, leakers and rejects, munitions dismantling, agent washout, agent hydrolysis, energetics hydrolysis, metal parts treatment, biological treatment, and brine reduction. Because several of these processes are batch processes while others are continuous, the model can be used to evaluate buffer storage requirements between various operations. Buffer storage is needed for situations in which single-unit failure would limit throughput or shut down the plant completely, and to ensure that the desired throughput can be maintained. The analysis indicates that the rate-limiting activity in the process design is the munitions disassembly step. As a result, in order to maintain the desired availability of munitions, the front-end process has a separate disassembly station for each of the three types of munitions to be treated. There is an unequal number of each type, with the number of 4.2-inch mortars being about one-third the number of either the 105-mm or the 155-mm projectiles. Therefore, after completion of the disassembly of all 4.2-inch mortars, including leakers, the 4.2-inch mortar line will be modified to disassemble the remaining types of leaking projectiles by retooling the projectile/mortar disassembly (PMD) station. However, Army safety regulations may require that all three PMD machines be shut down if maintenance is being performed on one, and thus the effective throughput capacity for the disassembly step would be reduced. Finding 3-1. The rules for personnel safety during equipment repairs in the explosion containment rooms 1   Unless otherwise indicated, background material in this chapter is drawn from U.S. Army, 2004b, and information on the intermediate design is from the PCAPP intermediate design review in San Francisco, May 19–21, 2004.

OCR for page 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant may require that all three stations must be shut down if maintenance is being performed on one station. In a worst case, the unavailability caused by the maintenance of a station would be nearly three times what would ordinarily be expected. Recommendation 3-1. The Bechtel Pueblo team designing the Pueblo Chemical Agent Destruction Pilot Plant should seek clarification on the regulations for the safety of maintenance personnel during the repair of equipment in the explosion containment rooms (ECRs). If the regulations call for operations in all three ECRs to stop during maintenance on one station, the effects of such shutdowns should be examined by the event-driven iGrafx model, and performance and schedule impacts should be incorporated into the design. Design changes should be sought whereby operations in the other two projectile/mortar disassembly ECRs can continue while planned and unplanned maintenance is performed in the third station. TECHNICAL RISK REDUCTION ISSUES The core processes to be used at PCAPP for destroying chemical agent stored at Pueblo Chemical Depot (PCD) are neutralization (hydrolysis) followed by biotreatment. Neutralization is well proven for the destruction of neat mustard agent (HD or HT) and is currently being used successfully at Aberdeen Proving Ground, Maryland. As discussed in Chapter 2, however, multiple steps are required to separate the mustard agent from the projectiles and mortar rounds. Other materials obtained from the separation process include energetics; metal parts from bursters, fuzes, and munition casings; and dunnage, which includes wooden pallets, packing materials, personal protective equipment (PPE), rags, and other organic wastes (solid and liquid). Any of these other materials may be contaminated with agent. The practice in other chemical agent destruction facilities has been to consider these materials contaminated and to process them in a manner that ensures the destruction of the chemical agent. This practice also is being implemented in the PCAPP design. The biotreatment process was originally planned to be used at Aberdeen Proving Ground for the secondary treatment of the hydrolysate, but after the events of September 11, 2001, the hydrolysate was sent off-site to accelerate the schedule. However, the biotreatment of the hydrolysate was demonstrated as a viable means to treat this secondary waste earlier in the Assembled Chemical Weapons Assessment program and further tested in a technical risk reduction program study (NRC, 2001a; U.S. Army, 2003a). While these core processes being used for destroying mustard agent are well proven, many of the treatment processes for the other materials (noted above) are novel. Moreover, the PCAPP design is a first-of-a-kind pilot plant, and, consequently, overall integration of the unit processes presents additional challenges. In recognition of these technological challenges, the Bechtel Pueblo team assembled an integrated product team (IPT) to assess and select combinations of unit operations that would meet PCAPP requirements. The IPT also performed a technical risk assessment (TRA) of the proposed design concept to identify problem areas in meeting performance objectives. The IPT consisted of a team of recognized experts in the design and operation of chemical agent disposal facilities. Although the selection process was subjective, it drew on lessons learned from the chemical agent disposal facilities at Johnston Island (in the Pacific Ocean); Aberdeen, Maryland; and Newport, Indiana; and on expertise from earlier ACWA engineering design studies. After the unit operations were selected, the IPT initiated a TRA of the total design concept embodying these operations. The TRA process, thus initiated, will be continued and refined throughout the life of the project. The ultimate goal of the TRA, as stated in the PCAPP Design-Build Plan, is to “maximize the safety of workers and the public, minimize any adverse effects on the environment, and ensure a smooth process for obtaining the necessary regulatory permits” (U.S. Army, 2003a, p. 137). The Design-Build Plan also notes that the ongoing TRA process will include “measures for reducing risk for cost overruns and minimizing the overall project schedule” (U.S. Army, 2003a, p. 137). The committee observes that the initial TRA focused almost exclusively on risks stemming from cost and schedule overruns. Based largely on engineering judgment, the IPT identified 90 different risks relating to major unit operations. These risks are listed in Appendix P of the PCAPP Design-Build Plan and are reproduced in this report as Appendix C (U.S. Army, 2003a). To prioritize the risks, the IPT adopted a semiquantitative approach, assigning two weighting factors, one to “probability of occurrence” and one to “technical, schedule, and cost consequence of occurrence.” The total risk for each scenario was calculated by multiplying the two weighting factors together.

OCR for page 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant The probability of occurrence was scored on a scale of 1 to 5, where 1 is “remote,” 2 is “unlikely,” 3 is “likely,” 4 is “not defined,” and 5 is “near certainty.” The consequence of occurrence was scored on a scale of 0.2 to 1.0, where 0.2 is “minimal or no impact,” and 1.0 is “unacceptable.” Multiplying probability and consequence provided the overall risk weight. Each of the risks was then assigned to one of three overall risk-weight categories: low (overall weight equal to or less than 1.0); medium (overall weight between 1.0 and 3.0); and high (overall risk weight above 3.0). No risks with an overall weight equal to or greater than 3.0 were identified. This analysis identified the 10 distinct areas requiring either paper studies or testing that are listed in Table 3-1. Also, the PCAPP team identified other issues requiring testing or trade (paper) studies to resolve key design, construction, operation, or closure issues. Both the TRA issues and these other issues are listed in Table 3-2 (U.S. Army, 2003a). The committee observes that the criteria in the TRA for assigning probability and consequence values could have been defined more precisely, but the overall risk-weight results seem reasonable. For example, whereas health, safety, and environmental impacts were considered in the description of some of the risk scenarios, the probability and consequence weightings ascribed to scenarios with health, safety, and environmental impacts are not always consistent with the scenario description. Cost and schedule impacts appear to have been the primary drivers of the probability and consequence scoring scheme. Some examples of such inconsistencies are as follows (U.S. Army, 2003a): Under “baseline/reconfiguration operations” (see Appendix C in this report), the probability of “inadequate design considerations for explosives handling” is considered to be remote, with a probability weighting of 1 out of 5. Under “energetics treatment processes,” the probability of “inadequate explosive considerations” is considered to be unlikely, with a higher probability weighting of 2. The consequence weighting, however, is 0.8 for the former and only 0.4 for the latter, even though both could result in explosion. It would seem that the consequences of explosion in both instances or locations should be the same, since the impact of investigation delays would be similar. Under “metal parts treatment processes,” one of the listed risks with a low overall rating is “explosive gas build-up/purging potential for explosion.” The probability is given as unlikely (weighting of 2 out of 5). The consequences are described as follows: “explosion occurs, requires downtime, repairs, and minimal to moderate impact to cost and schedule.” The relatively low consequence weighting of 0.45 out of 1.0 may stem from lumping the potentially catastrophic effects of an explosion with the minor impacts on cost and schedule. While it is quite possible that equipment could be repaired fairly quickly following an explosion, the project could be shut down for months while the explosion was investigated. Under “biotreatment of hydrolysate processes,” one of the listed risks with a low overall rating is “odor not adequately controlled.” The probability is unlikely (weighting of 2). The consequences are stated as “potential regulatory violation, resulting in fines and potential shutdown” and assigned a weighting of 0.3. The IPT may be optimistic in anticipating “minimal cost and schedule impacts.” The same risk is also included under “environmental risks.” There, the probability is weighted as likely (rating of 3), but the consequence remains at 0.3. The committee believes that, despite the shortcomings in the rigorousness of the implementation of the TRA process, most of the major technical roadblocks to a successful design were identified. Tests are under way to acquire design data for unit operations with insufficient prior testing or operating experience, and studies have been undertaken to evaluate promising alternatives or to resolve design decisions for areas not requiring testing. As noted earlier, these tests and studies are listed in Table 3-2, along with reference to the TRA risk areas/scenarios that have been identified in Appendix P of the PCAPP Design-Build Plan. They are discussed in more detail in the subsequent sections of Chapter 3, along with other aspects of the design important to its success. Finding 3-2. Although the implementation of the risk assessment methodology in the preliminary technical risk assessment for the Pueblo Chemical Agent Destruction Pilot Plant is sometimes inconsistent, the committee judges that the major technical risk issues have been identified and are being addressed by tests and studies.

OCR for page 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant TABLE 3-1 Major Potential Risks and Proposed Mitigation Measures for the Pueblo Chemical Agent Destruction Pilot Plant Identified in the Technical Risk Assessment Potential Risk Proposed Mitigation Measure Integration of material-handling units with process equipment may lead to more maintenance, increased operating time, and moderate increase in schedule. Design in adequate buffer capacity. Develop and maintain interface control diagrams (ICDs) early in program, use three-dimensional computer-aided design and drafting models to assess impacts, and perform systems engineering, using iGrafx to model flows beginning early in the program to mitigate rework. Include repair/access, maintainability, construction tolerances, turnover, etc. in three-dimensional model. Unknown agent and energetic characteristics impacting performance may lead to more maintenance, increased operating time, and moderate increase in schedule. Use a broad design range for feed characteristics based on lessons learned from other chemical demilitarization programs. Delays in obtaining the Certificate of Designation,a which is a new permitting requirement, may lead to delays in start of construction and all subsequent operations, with significant schedule impact. Prepare a research, development, and demonstration permitting strategy using multiphases. Involve regulators in all phases of the program, and get their buy-in to the design. Form integrated product teams to support and resolve issues. Unreliability of entire water-supply system requires facility to shut down, pending resupply, with moderate schedule impact. Perform trade studies to enhance recovery and possible alternate sources (sanitary sewer, pink water, new wells). Determine overall system reliability and availability for water supply; perform well testing and upgrade pumping system, if found inadequate to meet demands. Energetics rotary hydrolyzer (ERH)/heated discharge conveyer (HDC) throughput rates less than required, and minimal fullscale data may result in plant operational schedule not being maintained, causing impacts on other operations and moderate schedule delays. Design for surge capacity and use a conservative rate (test data indicate higher achievable). Fabricate and perform extensive tests at fabrication shop prior to shipment to site. Cotton fibers in propellants and bags impacting performance of ERH and energetics neutralization reactors from excess material plugging recirculation pumps, causing malfunction and requirement for maintenance, with moderate schedule impact. Conduct trade study to consider separate reactor for propellants, with additional testing identify alternatives to adding bags to ERH. Closure criteria not adequately defined, causing minimal to moderate impacts on closure costs and schedule due to extra time and equipment requirements. Add experienced closure expert from Johnston Atoll Chemical Agent Disposal System to design-build team to participate and provide lessons-learned input into design, develop closure criteria early, and maintain the “design to close” approach throughout the entire program. Add closure in design reviews, and add closure data to the engineering procurement and construction design tool for the closure package. Inability to deliver munitions within the facility at night, causing operations to stop when munition buffer inventory (4 hours) is depleted, with moderate schedule impact. Design for munitions night transportation, using adequate lighting and covered areas. Perform trade study to determine transportation alternatives with covered passageways, ensure safety analysis in limiting conditions of operation. Work with customer to obtain permission to perform such night operations. Aluminum dissolution in caustic causing downstream problems in immobilized cell bioreactors (ICBs) by aluminum hydroxide reducing the surface area of biomass and potential sluffing, which results in poor performance and moderate schedule impact. Conduct trade study to determine impact on downstream equipment items. Consider adding pH adjustment and filter press upstream of the ICBs to remove precipitation, if warranted. Process used in industrial applications. Verification of heat transfer for Pueblo tray configuration to validate metal parts treater throughput. If less effective heat transfer occurs, plant average operational schedule is not maintained; impacts other operations and results in moderate schedule delays. Perform additional testing using prototype munition trays in the ACWA test unit to obtain additional data to confirm the heat-transfer model prior to scale-up. Design using model to full-scale unit, fabricate full-scale unit, perform tests using approved test plan with acceptance criteria and contingencies for failure. aA Certificate of Designation (CD) is a document issued by the local (county or municipality) governing body authorizing the siting of land for a solid waste disposal site or facility. The CD is issued if it has been determined that the standards are met and after local issues are satisfied. SOURCE: U.S. Army, 2003a.

OCR for page 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant TABLE 3-2 PCAPP Risk Issues Identified for Testing or Trade Studies Test or Study Design-Build Plan Appendix P, Risk Areas/Scenariosa Testing Prototype Test (Design-Build Plan (DBP) Sect. 2.11.3) Prototype Continuous Steam Treater (CST) 43–45 Robotic Performance Coupled with Munitions Washout System (MWS) 23–27 Prototype Metal Parts Treater (MPT) 34, 35 Prototype Energetics Rotary Hydrolyzer (ERH)/Heated Discharge Conveyor (HDC) Interface 17 Laboratory/Bench-Scale Tests (DBP Sect. 2.11.4) Scale Testing of ERH 17 HT and Explosives Biotreatment 55 Propellant Bag and M8 Thread Processing (Laboratory Testing of Propellant Reaction) 22 Aluminum Hydroxide Solids 21 Other Identified Tests (Presentation, November 6–7, 2003) Personal Protective Equipment (PPE) Certification for HT Agent None identified Trade Studies (DBP App. J) Projectile/Mortar Disassembly (PMD) Machine 23–27 Explosion Containment Room (ECR) None identified Conveyor (for energetics and agent-filled munitions) 15 Munitions Refrigeration (leaking munitions) None identified On-site Munitions Transportation Alternatives (from igloos to unpack area) 1–7 Tetrytol Exudates Presence 16 Enhanced Water Recovery 49, 50 Optimize Process Modeling Various aAppendix P of the PCAPP Design-Build Plan can be found in Appendix C of this report. SOURCES: U.S. Army, 2003a; PCAPP Design Overview Briefing by Craig Myler, PCAPP Chief Scientist, to the ACWA Design Committee, Aberdeen Proving Ground, November 6, 2003. Recommendation 3-2. If use of the technical risk assessment (TRA) scoring process for the Pueblo Chemical Agent Destruction Pilot Plant is continued, the Army and the Bechtel Pueblo team should more clearly define each of the weighting factors used in the initial TRA for probability and consequences. Consideration should be given to separating cost and schedule, health and safety, and environmental impacts, if that is necessary to ensure consistency. Furthermore, the methodology for assigning risk reduction factors should be clarified. Additionally, a process should exist to verify that the implemented mitigation measures result in the same level of risk reduction that was assumed during the TRA. The lay public or qualified representatives selected by the lay public were not involved in the initial TRA, even though the lay public often perceives risks differently from how the technical analysts perceive them. For example, higher risks in the public’s perception may relate mainly to worker and public safety, whereas the Bechtel Pueblo team may see those risks as manageable and less of a challenge than other technical risks with greater probability of major cost and schedule impacts. Involving the public early in the technical risk assessment activity can help alleviate overreaction to unpleasant surprises later on. The committee believes that the IPT may have been optimistic about the maturation rate for the new technologies. Moreover, apparently small problems can cause extensive delay and cost overruns—for example, the occurrence of crystals in the sarin in M55 rockets being processed at the Anniston baseline incineration facility caused extensive concern among both the technical and public communities.

OCR for page 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant The National Research Council previously reported the importance of involving members of the public, stakeholder group representatives, and local governmental officials, in addition to technical experts, in decision making about risks that are nonroutine and highly controversial, such as is the case with the PCAPP facility (NRC, 1996). The lay public and local officials have perspectives on hazards and risk issues that are both legitimate and important to consider. Involvement of participants from the lay public and local official communities should not diminish the scientific integrity of decisions, but should bring diverse concerns and considerations to bear in risk decisions. Past NRC reports have elaborated on the importance of involving all parties in every step of risk decision making, beginning with the definition of the problem.2 This involvement is particularly important because there are public considerations of policy assumptions that are embedded in risk analysis activities. Finding 3-3. The integrated product team (IPT) that initiated the technical risk assessment of the design concept for the Pueblo Chemical Agent Destruction Pilot Plant (PCAPP) included only experts in chemical demilitarization selected by the PCAPP contractor. No involvement by the lay public or qualified representatives of the lay public was included in developing the technical risk assessment of the IPT, even though the assessment process was conducive to such involvement and would ensure that process efficacy, safety, and environmental concerns were addressed from the public’s perspective. Recommendation 3-3. Qualified representatives selected by the lay public should be included in any future technical risk evaluations for the Pueblo Chemical Agent Destruction Pilot Plant, not necessarily to identify the risks, but to provide an independent perspective on the rankings of probabilities and consequences. DISASSEMBLY AND TRANSFER PROCESSES On-site Munitions Transportation Alternatives (from Igloos to Unpack Area) Palletized munitions are to be loaded by forklift onto ammunition transport vehicles from the storage igloos at Pueblo Chemical Depot, where the chemical munitions are stored. They will be transported to the PCAPP energetics processing building and then offloaded into a receiving vestibule at the unpack area of the EPB. The transfer of munitions from the storage igloos to the EPB unpack area is managed under the control of the PCD commander. The general intent is to have sufficient munitions available to provide for the plant throughput 24 hours a day, 7 days a week, while minimizing cost and ensuring safety. The technical risk reduction program (TRRP) for PCAPP determined that the following alternatives would optimize the munitions delivery flow: Transport vehicles. Largely for reasons of depot familiarity, PCD intends to continue to use the current transport vehicles. New vehicles of this type purchased for this program will likely be a larger (18-ft bed) model that increases the capacity by about 50 percent. These new vehicles will be sufficient to provide the 24-hour per day feed requirements for plant operations. In a typical workday (daytime only), one vehicle can be expected to deliver four loads. Previously overpacked pallets in the igloos, which by virtue of their bulk will decrease the delivery capacity, were not considered. Daytime operations. PCD has required that all movement of munitions from storage to PCAPP be conducted during daylight hours. Weather. Historical data on average inclement weather events in the Pueblo area (extreme temperatures, precipitation, snow, wind, electrical storms) and on the frequency of extreme adverse weather were evaluated to determine limitations to the delivery of munitions. High or gusty winds (causing 43 percent of outages) were determined to have the major impact on weather-related downtime. A probability analysis for weather outage was then applied to the various scenarios for deliveries to determine the impact on throughput. Storage limitations at the unpack area. Army regulations restrict the quantity of munitions that can be stored to what can be processed during half of a work shift (U.S. Army, 1999). Recommendations from the TRRP indicated the best and least costly scenario—that a larger staging capacity be provided at the EPB unpack area. Deliveries will be made 7 days a week. 2   For example, see NRC, 1996; 1999b; 1999c.

OCR for page 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant The presumption is that an Army waiver is possible and will be granted. Finding 3-4. The Bechtel Pueblo design team presumes that a waiver will be granted for storage of a larger operational buffer capacity in the unpack area of the energetics processing building at the Pueblo Chemical Agent Destruction Pilot Plant. This waiver is pivotal to the implementation of the optimal scenario, and if not granted, will necessitate significant changes to the design. Recommendation 3-4. The Bechtel Pueblo team should make immediate application for a waiver to obtain additional munitions storage in the unpack area of the energetics processing building at the Pueblo Chemical Agent Destruction Pilot Plant. In a presentation of the transportation simulation of the munitions transfer process, the transport vehicles were shown moving one way throughout the PCD storage area. However, after passing through the security gate, the vehicles moved both ways on the roads in front of the EPB. This two-way traffic pattern introduces the possibility of collisions between the transport vehicles as they maneuver into position to unload the munitions and to return to the security gate to retrieve another load of munitions. Finding 3-5. The two-way traffic pattern in the vicinity of the energetics processing building at the Pueblo Chemical Agent Destruction Pilot Plant introduces hazards associated with the movement of large vehicles in a limited area. Recommendation 3-5. The Bechtel Pueblo team should evaluate the hazards associated with the two-way traffic pattern within the restricted area in the vicinity of the energetics processing building at the Pueblo Chemical Agent Destruction Pilot Plant and should consider revising this pattern to maintain a one-way flow of traffic throughout the site, or it should provide suitable separation barriers for traffic on the two-way portions of the munitions transport system. Reconfiguration Room Munitions stored at Pueblo Chemical Depot are either in their original packing or have been reconfigured. Reconfiguration involves removing propellant charges, igniters, associated packing materials, and mortar round fins that are usually attached before firing. Currently, 28,376 of the 105-mm projectiles and all of the 4.2-in. mortars (97,106 rounds) are still boxed with propellant and must be reconfigured before further processing. The PCAPP initial design includes a reconfiguration room located east of the ECRs in the EPB unpack area. The boxed munitions (see Table 1-1 in Chapter 1) are moved to the reconfiguration room on carts. The boxes are then opened, and the contents are removed and placed on a conveyor, where they are disassembled manually. They are then palletized for later processing or sent directly to the appropriate ECR for disassembly. A key step in the reconfiguration is the removal of igniters (cartridges about the size of shotgun shells) from the mortars by using a pulling machine. Currently, Bechtel Pueblo team designers are developing methods and equipment for accessing the energetics contents of the igniters because the plastic casings of the energetics do not dissolve. All metal parts that are not part of the reconfigured munition, such as fins and metal strapping, are collected in bins for transport to the metal parts treater. All energetic materials except the bursters are collected and placed in trays for transfer to one of the two energetics rotary hydrolyzers (ERHs). Bursters are removed later when the reconfigured munitions are processed through the projectile/mortar disassembly machines in the ECRs, where the appropriate mechanism for burster removal is available. Currently, the PCAPP design team is seeking permission to process uncontaminated propellant from the reconfiguration process off-site. Igniters would be processed in the ERHs. All nonmetallic packing materials, including box filler material, propellant cardboard cases, and the boxes, are collected in bins for transport to the dunnage shredding and handling system and continuous steam treater. The reconfiguration room is designed so that personnel in the room can safely open and reconfigure the munitions. The operations conducted in the reconfiguration room would normally be performed in the field without further protection. Although the presence of propellant represents a potential flammability hazard to workers, an explosive hazard does not exist because the propellant is not confined (i.e., it is uncontained). Finding 3-6. The reconfiguration process for munitions at the Pueblo Chemical Agent Destruction Pilot Plant does not contain positive controls to prevent the

OCR for page 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant manual mixing of energetics and metallic and nonmetallic waste or dunnage in a manner that ensures that all energetics are properly collected for further processing. The bins for collecting the nonmetallic and metallic wastes that are sent to the dunnage shredding and handling system and metal parts treater, respectively, are located next to one another, thus increasing the likelihood of commingling the two types of wastes and up-setting the downstream processes. Furthermore, the method of ensuring that all energetic materials are collected and retained for further processing is not clearly defined at this time. Conceivably, these materials could end up in the bins. Recommendation 3-6. Positive controls should be incorporated to prevent the mixing of waste streams for the metal parts treater and dunnage shredding and handling system/continuous steam treater processes during reconfiguration room operations at the Pueblo Chemical Agent Destruction Pilot Plant. For example, means should be provided to ensure that energetics removed in the manual disassembly will only be placed in the proper trays. Robotic Performance Coupled with Munitions Washout System The design for the munitions washout system (MWS) for accessing and removing agent from the mustard agent projectiles at Pueblo Chemical Depot is based on lessons learned from processing similar munitions at the Johnston Atoll Chemical Agent Disposal System (JACADS). In tests of the MWS, the agent cavities of mortars and projectiles have been cleaned to a bright and shiny metallic surface. Based on preliminary testing, a high water temperature minimizes water usage and is more effective for the removal of agent heels from the projectiles. The design must optimize water wash temperature and volume for both the washout process itself and the downstream processing. Based on current design, the flow rates for washout of PCAPP munitions are the following: 3 gallons per minute (gal/min) at 140 seconds for 155-mm projectiles, 3 gal/min at 35 seconds for 105-mm projectiles, and 4.5 gal/min at 70 seconds for 4.2-inch mortars. These flow rates will be updated based on the TRRP test data. Several issues are of concern with the design for the MWS, most of which have been identified by the Bechtel Pueblo design team. A key issue is the potential incompatibility of the settling process with downstream processing—that is, premature neutralization in the settling tank could lead to varying agent concentration in the feed. The design team may determine that additional testing, including the redesign of the settling process, may be warranted to adjust for variations in the feed concentrate and to avoid, to the extent possible, hydrolysis upstream of the reactors and the formation of difficult-to-hydrolyze sulfonium compounds from premature hydrolysis of the MWS agent concentrate in the settling tank (NRC, 2001b). Another concern is the close mechanical tolerances for munitions placement by the robot. The MWS is a highly mechanized and somewhat complex design with narrow tolerances, having as a key element a multiaxis, floor-mounted robot specific for each of the three lines. The committee believes that it is not adequate to simply identify the need for precision, but to also articulate compensating methods in the event of misplacement or misalignment of munitions in trays. Also, the committee did not see a plan for what to do if one munition, for any reason, failed to show a sufficient difference in weight before and after washout. Finding 3-7. The committee believes the munitions washout system (MWS) design of the Pueblo Chemical Agent Destruction Pilot Plant is an effective and reliable approach to accessing and removing the chemical agent, including agent heels, from the munitions bodies. This MWS design also promises the nearly complete removal of the mustard agent and the residual heels, thereby lowering the agent loading on munitions going to the metal parts treater. Recommendation 3-7. Future testing and integration efforts for the munitions washout system design of the Pueblo Chemical Agent Destruction Pilot Plant should ensure that the design is forgiving of misalignment and misplacement of munitions in trays, and that procedures are in place to effectively deal with off-normal situations (such as when a munition fails to show a sufficient difference in weight before and after washout). Treatment of Leaking Munitions At Pueblo Chemical Depot, approximately one-tenth of 1 percent of the munitions may be leakers. The munitions are stored on pallets containing from 30 to 50 munitions each, and locating individual leakers can be

OCR for page 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant very labor-intensive and time-consuming.3 It can take as long as a week to locate a leaker in an igloo. Subsequent overpacking of the leakers is also a labor-intensive and hazardous process. To enable safer and more efficient handling of leakers, several options employing refrigeration were considered, as reported in the PCAPP refrigeration study (U.S. Army, 2003b). These options were derived from consideration of JACADS experience with leakers and “frothing” while processing munitions containing mustard agent. Frothing of agent when accessing the agent cavity is similar to the foaming that can occur when opening a carbonated beverage bottle. The frothing results from dissolution of pressurized gases such as hydrogen that are produced by chemical degradation of the agent over time.4 The operators at JACADS anticipated that this problem would be most severe with leakers and other munitions that were rejected because of difficulty in accessing their agent cavity during disassembly. Thus, they decided to freeze the leakers and the rejected munitions as a means of addressing the frothing issue. The PCAPP refrigeration study, anticipating that the design and testing of the cavity access machines (CAMS) would be successful in demonstrating control of frothing without the use of refrigeration, recommended using a single refrigeration unit similar to the one used at JACADS for leakers and rejected munitions only. In subsequent development of the initial design for PCAPP, the Bechtel Pueblo team proposed that the normal practice for location and isolation of individual leakers at PCD be replaced with a procedure in which the entire pallet containing a leaker would be overpacked by using a container that is still to be designed. These overpacked pallets would later be refrigerated and processed during the campaigns for processing leakers and rejected munitions, which will occur after all nonleaking 4.2-inch mortars have been processed. The committee understands that the PCAPP refrigeration unit, if used, would be installed in the unpack area of the EPB before the start of the campaigns for processing leakers and rejected munitions. The refrigeration unit would be used to freeze pallets containing known leakers. The pallets would be overpacked at the storage igloos. The pallet overpack would be a specially designed container, and after delivery to the EPB, it would be placed in a freezer to lower the agent temperature in the munitions to a point that would prevent thawing until after the agent cavity was accessed. Currently, no thawing would be expected for approximately 3 hours. The freezer unit would be a commercially available modular unit of about 640 cubic feet. Refrigeration was applied at JACADS only to eliminate problems with frothing when accessing the munition cavity. Contamination associated with these munitions prior to processing was not addressed, but the control of frothing when accessing the agent cavity was. The control of frothing appears to be addressed successfully by the new MWS/CAM design to be used at PCAPP. Testing to date provides assurance that this design also should be able to handle munitions that contain frothing agent without the need for refrigeration because the nose of the munition is sealed to the receiving vessel as described in Chapter 2 (FOCIS, 2003a; 2003b). Finding 3-8. Without resorting to refrigeration, the new munitions washout system/cavity access machine (MWS/CAM) design to be used at the Pueblo Chemical Agent Destruction Pilot Plant appears to satisfactorily address the problem of frothing mustard agent from munitions. On the basis of completed and planned testing (FOCIS, 2003a) and observation by committee members of the prototype mortar MWS, the committee believes that the CAM used to remove the base from 4.2-inch mortar rounds will be an effective means of accessing agent in these munitions. Recommendation 3-8. If the tests of the munitions washout system/cavity access machine to be used at the Pueblo Chemical Agent Destruction Pilot Plant still indicate problems with frothing agent from munitions, the committee recommends solving these problems without the application of refrigeration for leakers and rejected munitions. Specifically, refrigeration should not be applied to leakers and rejected munitions. The pallet overpack units, if used, would introduce additional contaminated material for processing in the continuous steam treaters (CSTs). This is material that has not been identified for inclusion in the CST test program. Furthermore, overpacking pallets means that 3   ACWA Design Committee meeting with Army and Bechtel National, Inc., participants at Irvine, Calif., February 11–13, 2004. 4   This frothing of mustard agent, or “champagning,” as it is sometimes called, resulted in increased maintenance of disassembly equipment, equipment modifications, and Resource Conservation and Recovery Act permit modifications during the processing of mustard agent munitions at JACADS (NRC, 2004).

OCR for page 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant most of the munitions on a pallet that are not leakers, and which could have been processed during the nonleaker campaign, might now have surface contamination from the leaker munitions as a consequence of the overpacking. There is little evidence to show that the existing procedure for leaker overpacking presents unique safety or logistical problems. Furthermore, there are currently only 28 known leakers and 490 “suspected” SUPLECAM (Surveillance Program for Lethal Chemical Agents and Munitions) leakers at PCD (U.S. Army, 2003b).5 This number would not be expected to increase significantly before the time when PCAPP begins to process leakers (NRC, 2004). Finding 3-9. The proposed overpacking of entire pallets containing leakers adds to the number of munitions that must be handled during the leaker campaigns. There appear to be little advantage and some disadvantages with this approach and an added processing burden for the continuous steam treaters. Recommendation 3-9. The procedure currently used to locate and overpack individual leakers at Pueblo Chemical Depot and other storage sites should be considered for continued use during operations of the Pueblo Chemical Agent Destruction Pilot Plant. Agent and Energetics Transfer Systems As noted in Chapter 2, the placement of both the agent transfer system and energetics transfer system in the transfer corridor of the energetics processing building requires structural blast protection elements to be added in the transfer corridor to prevent possible energetics explosions from dispersing agent from the munitions being transported by the ATS. However, the PCAPP design team found that the need for blast resistance complicated the design of the ETS overhead monorail transfer system. Because the ATS is in the same area as the ETS, the design of the ATS may also be impacted. At the time that this report was prepared, a pneumatic conveyor system was under consideration for transferring energetics in capsules placed in pneumatic tubes. Two pneumatic tubes would be provided for each explosion containment room and for the munitions reconfiguration area to allow each of the ECRs and the munitions reconfiguration area to feed either of the energetics rotary hydrolyzers. Thus, eight tubes would cross the transfer corridor. The committee has identified several concerns with the pneumatic tube design that must be addressed by the PCAPP design team, including the following: Requiring blast/missile protection at both ends of each tube, Addressing the mechanics of loading and unloading and catching the transfer capsules without severe deceleration loads, Preventing static electricity discharges in the tubes from movement of the capsules and air through the tubes, Ensuring a sufficiently large radius of curvature in the axial direction to permit capsules long enough to carry bursters, and Achieving a rate of energetics transfer equivalent to the design rate established for the overhead monorail conveyor system. Finally, regardless of the design chosen for the ETS, the ATS also will be affected because it uses the same corridor. The committee is concerned that the lack of firm choices for the ETS and ATS designs at this point may impact other design choices for interfacing systems and for the building footprint. The final design must ensure that problems with the ETS or ATS do not result in frequent shutdowns of both the ETS and ATS for repairs. This potential problem arises because the current layout places both the energetics and agent in the same space after effectively separating them in the ECRs. Finding 3-10. The choice of design for the energetics transfer system (ETS) has a significant impact on the interfacing systems, and it is not obvious that desired processing throughputs can be achieved with the current design because the transfer corridor in the energetics processing building is used by both the agent transfer system and the ETS. The energetics transfer system contained in the initial design poses reliability and maintenance problems and may require additional design changes in order to address explosive safety issues—for example, limits on the amount of energetics 5   SUPLECAM, a program conducted in the 1980s and early 1990s, involved intrusive sampling of the agent cavity of selected munitions to investigate the physical and chemical condition of the agents.

OCR for page 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant and the mix of energetics transferred in each tray. At the present time, any equipment problem in the transfer corridor may require shutdown of all disassembly processing. Recommendation 3-10a. The choice of agent transfer system (ATS) and energetics transfer system (ETS) designs should be resolved as quickly as possible in order to minimize schedule and cost impacts on the design and construction of the Pueblo Chemical Agent Destruction Pilot Plant (PCAPP). The PCAPP design team should accelerate its efforts to resolve both the ETS and ATS design selection while ensuring an acceptable level of reliability and explosive safety. Recommendation 3-10b. As the Bechtel Pueblo team considers design alternatives for the Pueblo Chemical Agent Destruction Pilot Plant, reconfiguration of the building layout should be considered to allow the transfer of agent-filled munitions and of energetics through entirely separate pathways in order to minimize the synergistic impact of agent transfer system and energetics transfer system failures on processing throughput. CORE PROCESSES Scale Testing of Energetics Rotary Hydrolyzer Energetics Hydrolysis System Hydrolysis tests were performed in 2003 at Deseret Chemical Depot (DCD) in Tooele, Utah (U.S. Army, 2003c). The final report was expected to be available at the beginning of May 2004. Because testing prior to 2003 had been done at or below 105°C, the DCD tests were planned to verify that the bulk of the reaction of energetics would be completed during the residence time in the energetics rotary hydrolyzer. One of the significant tests at DCD was the destruction of energetics in M14 boosters, which are long, narrow burster tubes that are closed at one end. The test results showed that in 45 to 55 minutes at 118°C to 124°C in 35 percent caustic, all of the tetrytol had been removed from the burster tubes. This was verified by visual inspection of the burster tubes. It was necessary to have excess caustic present (more than is required to react with the energetics) to prevent the hydrolysate from becoming too viscous. At one time, using twice the stoichiometric amount of caustic was thought to be sufficient, but a higher ratio of caustic to energetic is necessary to keep the hydrolysate sufficiently fluid. The DCD tests were run for 1 hour. These tests showed that the burster explosives and the propellants (provided they are exposed to the hot caustic) are easily hydrolyzed in less than 1 hour in 35 percent caustic at 114°C to 120°C. Caustic readily penetrates the propellant bags, and the cotton threads holding the bags together decompose in the ERH, spilling the propellant out into the solution, where it is hydrolyzed. Tests were performed with propellant that was sewn together in two-sheet bundles. It was unclear whether this represented the most sheets that could be encountered in a sewn stack. According to MIL-I-48086, the sheets may be packed in five-sheet bundles, so the tests do not represent the worst case for hydrolysis of the propellant (U.S. Army, 1972). Because the surface area will be smaller per unit mass of propellant, it should take longer to hydrolyze the propellant in a five-sheet bundle than in a two-sheet bundle.6 The design anticipates that hydrolysis of the energetics will be completed in the energetics neutralization reactors (ENRs), which are monitored for the presence of energetics by differential scanning calorimetry (DSC).7 However, the bench-scale hydrolysis data indicate that the hydrolysis of the energetics in the projectiles stored at the Pueblo site is likely to be completed in the two energetics rotary hydrolyzers.8 This likelihood reduces the need to include four ENRs in the design. It also reduces the need for a long residence time of the hydrolysate in the ENRs and the two holding tanks. Finding 3-11. The hydrolysis of energetic materials at the Pueblo Chemical Agent Destruction Pilot Plant is expected to be substantially completed in the energetics rotary hydrolyzers. Recommendation 3-11. The Bechtel Pueblo design team should review the number and sizing of the post-energetics rotary hydrolyzer components of the energetics hydrolysis system for the Pueblo Chemical Agent Destruction Pilot Plant. 6   ACWA Design Committee site visit to General Atomics, San Diego, Calif., April 6, 2004. 7   The test protocol is still being developed. 8   ACWA Design Committee site visit to Battelle Memorial Institute, Columbus, Ohio, March 19, 2004.

OCR for page 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant A caustic solution of 35 percent NaOH at 120°C is to be used in the ERH, and the potential exists for further concentration of the solution by splashing and evaporation. The effects of this environment on the equipment warrant consideration. For instance, there is the possibility of stress corrosion cracking (also often called caustic embrittlement) in such environments. Austenitic and ferritic stainless steels are often used in what could be termed intermediate ranges of caustic service for reasons of economy and utility. Austenitic stainless steels, primarily types 304 and 316, are very resistant to caustic in concentrations up to 50 percent and temperatures to about 95°C (200°F). Stress corrosion cracking can occur in 304 or 316 stainless steel at temperatures as low as 120°C for 35 percent caustic solutions. As this is the approximate operating range of the ERH, the use of 304 or 316 stainless steel must be carefully evaluated (Nelson, 1987). Another possible hazard is the presence of mercury as a contaminant. Mercury has been found in some mustard agent, and this may have contaminated some of the energetics. This contamination constitutes a potential hazard, since it can contribute to cracking or pitting of austenitic stainless steel with mercury concentrations as low as a few parts per million (Nelson, 1987). Nickel or nickel-base alloys, although more expensive than 304 or 316 stainless steel, are extensively used in more severe caustic applications. The very low corrosion rates also ensure low metal-ion contamination. Nickel has the lowest corrosion rates—even in molten anhydrous NaOH up to 538°C (1000°F)—and is essentially immune to caustic stress corrosion cracking. Inconel 600, although excellent in caustic service, has higher corrosion rates than the nickel 200 and 201 metals (Hoxie, 1975). Nickel-clad vessels and equipment are frequently fabricated to minimize the need for expensive high-nickel alloys. Typically, the nickel-clad thickness represents 20 percent or less of the base steel thickness. However, mercury contamination as previously discussed will also affect nickel and nickel alloys. One way to address this latter issue is to determine whether mercury contamination of the energetics to be processed through the ERH is a real issue. If it is determined that mercury could be present in parts per million quantities in the ERHs, then steps could be taken to ensure that this situation will not cause excessive downtime. Steps that might be taken include identifying the mercury-contaminated munitions and processing them in such a manner that the mercury concentration in an ERH does not exceed a certain level. Alternatively, the mercury might be removed from the munitions prior to processing, or inspection and repair intervals for the ERHs could take the possibility of the effects of mercury contamination into account. Finding 3-12. The use of 35 percent caustic at 120°C in the energetics rotary hydrolyzer of the Pueblo Chemical Agent Destruction Pilot Plant could cause corrosion issues. Recommendation 3-12. Design decisions for the Pueblo Chemical Agent Destruction Pilot Plant regarding the appropriate material for use in caustic service should take into account temperatures, concentrations of caustics, and contaminants. Consideration should be given to the possibility of lowering the operating temperature and concentration of the caustic in the energetics rotary hydrolyzer (ERH) reactor to guard against stress corrosion. Another alternative would be to consider the used of nickel-clad ERH vessels. Biological Treatment of Hydrolysates After neutralization, the agent and energetics hydrolysates will be combined for secondary processing via biotreatment. Hydrolysis is a well-known process for destroying energetic materials and can reliably convert these materials to nonenergetic by-products (Bonnett and Elmasri, 2001). Hydrolysis likewise converts agent into by-products that are less toxic. The by-products can be biotreated after adjusting pH and establishing appropriate conditions. The success of biological treatment depends on knowing the hydrolysis by-products, their degradability and toxicity, and the mass rates at which they are produced. The PCAPP design calls for the use of immobilized cell bioreactors (ICBs) (e.g., a fixed-film reactor). The bioreactors are a proprietary technology, patented by Honeywell. Previous piloting studies for a chemical demilitarization application were conducted on a full-scale unit during engineering design studies earlier in the ACWA program to determine operating parameters and throughput expectations. More recently, testing to optimize the performance of the ICBs was being performed at Aberdeen Proving Ground, Maryland, while this report was being prepared; the testing used a 4-liter-scale test unit on a blend of hydrolysates of

OCR for page 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant HD/HT and tetryl and tetrytol (70 percent tetryl and 30 percent TNT). This testing was scheduled for completion in mid-2004. This study should reaffirm the suitability of the process for PCAPP and identify problems and areas for improvement. The following subsections comment on specific aspects of the biological treatment system. Energetic Materials The complexity of hydrolyzing energetic materials depends on the chemical composition of the explosive (Heilmann et al., 1996). Propellant components such as nitrocellulose and nitroglycerine are easily hydrolysable (Newman, 1999). Destruction of energetics components of bursters such as tetryl and TNT by hydrolysis is complicated by the aromatic ring structure of these substances. It is well established that TNT loses its energetic properties when it undergoes base hydrolysis (Earley et al., undated). However, the organic by-products are less well known and quite complicated. Recently, Thorn et al. (2004) investigated base hydrolysis of TNT, and their findings illustrate the complexity of identifying the products. They concluded that the biodegradability of the products is still unknown. Several investigators have treated mixed hydrolysis products, including TNT hydrolysis products, in biological reactors. Earley et al. (undated) have summarized previous work. The disappearance of TNT by-products was not documented by rigorous methods, but by the removal of chemical oxygen demand (COD) and total organic carbon (TOC). The processes appear successful. Finding 3-13. While it has been demonstrated that TNT hydrolysate can be treated in a bioreactor, the nature of the decomposition products and the toxicity of the residual organic carbon in the bioreactor effluent from treatment of tetryl and TNT hydrolysate have not been established. Until there is conclusive evidence that the effluent from the bioreactors is not toxic, no final decision about the disposal of the sludge from the bioreactor can be made. The determination of the toxicity of the sludge is an important issue to the public and to state regulators. Recommendation 3-13. In designing the Pueblo Chemical Agent Destruction Pilot Plant, the Bechtel Pueblo team should establish the toxicity of the effluent from biotreatment of TNT and tetryl hydrolysates, including any carcinogenic or mutagenic properties, so that an acceptable disposal plan for the sludge can be designed. The ammonia produced during the hydrolysis of energetic materials is easily volatilized at the high pH. The residual ammonia can be nitrified to nitrate in the biological process. The nitrate is less toxic than ammonia or nitrite and does not express an oxygen demand. Removing nitrogen from wastewaters is a well-established technology that can be addressed by the Bechtel Pueblo design team should it become necessary. Various denitrification processes could be used. And, the processes that separate salts during the reclamation of biological reactor effluents will also separate nitrate. Mustard Agent Mustard agent is effectively destroyed by hydrolysis, as noted previously, and the by-products are well known and degradable. Earley et al. (undated) have reviewed previous studies which document bioreactors that successfully treated the main hydrolysis by-products, thiodiglycol and dithiane. Degradation was documented by the disappearance of COD and TOC. Other Contaminants In the process of hydrolyzing the mustard agent and energetic materials from the munitions stored at Pueblo, other substances will also come into contact with the high-pH, high-temperature hydrolyzing solutions. Materials such as aluminum are expected to completely dissolve. Low-carbon steel and stainless steels will be unaffected, although iron particles in agent heels, possibly from corrosion, have been noted. Some plastics will dissolve. Additional contaminants may be present in small quantities. Earley et al. (undated) note that various volatile compounds were found in previous studies. These compounds may have been laboratory contaminants or contaminants in manufacturing or disassembly of the weapons (e.g., methylene chloride used in organic extractions, organics associated with lubricants, residual cleaning agents, and so on.). The design for the biological treatment process should be capable of handling these contaminants. As the pH of the hydrolysates is reduced to ranges suitable for biotreatment (i.e., pH 6 to 9), some compounds will precipitate, as

OCR for page 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant discussed below, and may potentially create problems in the bioreactors. The aluminum and suspended solids are of particular concern. Aluminum will not be toxic to the biological process, but will produce a voluminous aluminum hydroxide floc at the point of neutralization (the various aluminum hydroxide polymers are least soluble at pH 5.5). This floc will tie up suspended solids that are present in the hydrolysates. Any heavy metals that precipitate or that exist in particulate form will probably be incorporated in the aluminum hydroxide sludge. Aluminum also reacts with phosphorus. Aluminum sulfate (alum) is typically used at wastewater treatment plants to precipitate phosphorus. Both phosphorus and nitrogen are essential nutrients for biological reactors. Nitrogen will be available in sufficient quantities from the explosives, but there is no natural source of phosphorus, and any that exists in the wastewater streams will be precipitated by the aluminum. The Bechtel Pueblo design team should demonstrate how they can avoid depletion of phosphorus in the PCAPP immobilized cell bioreactors. Finding 3-14. Phosphorus is a required nutrient for the bacteria in the immobilized cell bioreactors (ICBs). Because of the precipitation of phosphorus by the aluminum, the wastewaters will be devoid of phosphorus, and the bioreactors will be phosphorus-limited. The bench-scale testing of the ICBs at the Battelle Memorial Institute in Columbus, Ohio, is not evaluating the effects of aluminum and will not observe the effects of phosphorus depletion. Recommendation 3-14. The Bechtel Pueblo team should provide for phosphorus addition for the immobilized cell bioreactors in the Pueblo Chemical Agent Destruction Pilot Plant. When aluminum hydroxide floc has been allowed to enter a biological process such as the activated sludge process, it has become enmeshed in the activated sludge flocs, which in turn overloaded the clarifier/ thickener and required operation at reduced sludge age (mean cell retention time [MCRT] or solids retention time [SRT]). The PCAPP process design calls for the use of a fixed-film growth (immobilized cells), and the bulk of the aluminum hydroxide floc will pass through the process and become part of the suspended solids in the effluent or, at very high concentrations, may coat or clog the bioreactor packing. In either case, it will be necessary to remove the aluminum floc from the bioreactors, or by a clarifier after the bioreactors. Aluminum hydroxide floc is difficult to thicken and rarely can be thickened beyond 1 percent in a conventional gravity clarifier. The mass can be predicted from the mass of aluminum to be destroyed in the munitions (the mass of floc will be much greater owing to the hydroxide and waters of hydration). All of the aluminum can be expected to precipitate. Aluminum hydroxide flocs are common in water treatment processes, since many if not most water treatment plants use aluminum sulfate as a primary coagulant. There are also examples of treatment of waste aluminum chloride solutions in wastewaters from chemicals manufacturing. In developing a design for PCAPP, the Bechtel Pueblo team must anticipate the problem presented by the aluminum hydroxide floc. A screening for other dissolution or precipitation problems, such as the issue with phosphorus, should also be made part of this analysis. The flow diagrams presented at an earlier meeting (see Figure 2-1 in Chapter 2) showed a hydrolysate neutralization tank prior to biological treatment.9 It may be necessary to add a clarifier/thickener after this tank, depending on the mass of precipitates. Finding 3-15. There is a significant potential problem with aluminum hydroxide precipitates and other precipitates and their effect on the biotreatment process at the Pueblo Chemical Agent Destruction Pilot Plant. The Bechtel Pueblo team is addressing the problem, but has yet to provide a satisfactory solution. Recommendation 3-15. The Bechtel Pueblo team, in designing the Pueblo Chemical Agent Destruction Pilot Plant, should include a process to remove the aluminum hydroxide precipitates prior to biological treatment, or demonstrate that the biological process can be operated successfully in spite of the precipitates. Controls At the committee meeting in April 2004, the Bechtel Pueblo team seemed to be knowledgeable of biological treatment technology, such as start-up time 9   ACWA Design Committee site visit to Battelle Memorial Institute, Columbus, Ohio, March 19, 2004; and PCAPP briefing by Craig Myler, PCAPP Chief Scientist, to the committee, Aberdeen Proving Ground, Md., April 13, 2004.

OCR for page 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant and acclimation and nutrient requirements.10 Biological processes generally require a minimum of three cell retention times to approach steady state. During this acclimation period, loading rates need to be reduced, and efficiencies will be low. These issues are not unique to the treatment of hydrolysates but are concerns for any industrial biological wastewater treatment. RESIDUALS TREATMENT PROCESSES Prototype Metal Parts Treater As part of the technical risk reduction program (TRRP), the Bechtel Pueblo team is fabricating a two-thirds-scale metal parts treater (MPT) at the Parsons Fabrication Facility, located in Pasco, Washington. The Parsons facility will also build the actual MPTs to be used for PCAPP. This TRRP activity will not test the offgas treatment system because the Bechtel Pueblo team believes that it was successfully tested as part of the earlier engineering design study (EDS) phase of the ACWA program for PCAPP. According to Parsons, the scope of the TRRP for the MPT is to develop “system design and fabrication data for a full-scale, fully automated treatment system using modified ACWA water hydrolysis of explosives and agent technology (WHEAT) EDS equipment to mitigate TRA risks.”11 The scope encompasses activities to design, fabricate, and test a mock-up prototype unit (single train), validate the heat-transfer model and test the configuration for the three sizes of projectiles to be processed through PCAPP, validate the throughput rates, and develop timing and availability data on tray handling and the induction heating system. The targeted heat-up rates for all three munitions types is less than 90 minutes with a throughput rate of 1 tray per hour. Part of the test objective is also to verify that typical maintenance activities can be performed within a reasonable time period and within the allowable period (about 2 hours) for a worker to be in a demilitarization protective ensemble (DPE) suit. The main thermal treatment chamber of the prototype MPT is 11 ft long and 4 ft 8 in. in diameter and will be inductively heated to 1350°F ± 50°F. The power supply is a 600-kW unit running at 3 kHz. The fabricators at the Parsons facility planned to use only 200 kW of the available power and may shift to a 10-kHz power supply if coil noise becomes excessive. The chamber of the prototype MPT is about 3 ft shorter than the 14-ft-long chamber of the PCAPP MPT design because the fabricators used a shell for the inductively heated chamber that was available from earlier ACWA engineering design study tests. This shell is made from 316L stainless steel, whereas the shells for the PCAPP MPTs will be manufactured from Hastelloy C276. The differences between the planned design for the MPT and the design used for the TRRP test performed at Pasco could allow some potential design deficiencies to be untested. For instance, since the planned MPT furnace shell is larger than the one being used for the TRRP, there may be some problems discovered at full scale that were not found at the lower scale of the TRRP. Some of these factors may include (1) the time required to reach 1000°F at certain munitions locations; (2) high-temperature creep, distortion, or sag of the MPT furnace shell and other components subjected to large cyclic thermal gradients; and (3) the effect of a superheated steam temperature on the soak time required for each batch. The committee members who observed the prototype MPT had some concerns about the temperature drop within the unit and the heating of adjacent components when the doors were opened for inserting and pulling the trays through. Although munitions trays have been enhanced over earlier versions to reduce the risk of cracking and misalignment, the opening and closing of the MPT chamber doors generate cyclic heating of the munition trays and other MPT components located near the doors. In addition, the conveyer used to move the trays will also be cyclically heated and cooled. However, the testing campaign is probably not long enough to detect any mechanical or thermal fatigue problems. Finding 3-16. Repeated thermal cycling of the trays and other components of the metal parts treater (MPT) in the egress chamber may cause distortion and thermal fatigue cracking, resulting in excessive needs for repair or replacement of parts. The testing planned at Pasco regarding the technical risk reduction program for the MPT will be of insufficient duration to determine if the proposed design will suffer from distortion 10   PCAPP briefing by Craig Myler, PCAPP Chief Scientist, to the ACWA Design Committee, Aberdeen Proving Ground, Md., April 13, 2004. 11   Personal communication of committee member with Mark Rieb, Parsons Project Manager, Pasco, Wash., site visit, May 7, 2004.

OCR for page 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant and thermal fatigue cracking that could affect time to repair or maintain, which in turn could affect overall unit availability. Recommendation 3-16. Thermal stress modeling, using a computational fluid dynamics model developed by Bechtel for the metal parts treater (MPT) of the Pueblo Chemical Agent Destruction Pilot Plant, should be used to determine if repeated cycles of parking hot trays on the MPT exit hardware will cause excessive creep fatigue and cracking of the trays. Similar modeling should evaluate the potential for creep and warpage of MPT chamber door seals and the Hastelloy C276 shell. Prototype Continuous Steam Treater The use of a continuous steam treater is proposed for 5X decontamination of potentially contaminated nonmetallic wastes at PCAPP. These wastes include shredded wood from pallets; shredded plastic from DPE suits; waste lubrication and hydraulic oils; mixed glass, plastic, wood, metal, and paper packaging materials; spent activated carbon; and other trash and debris. The 5X designation “indicates an item that has been decontaminated completely of the indicated agent and the material may be released for general use or sold to the general public…” (U.S. Army, 2002a, Section 5-1. c. (3)). From the time that the request for proposal for PCAPP was issued to the end of data gathering for this report, the only approved method for decontamination to 5X was heating the item to 538°C (1000°F) for 15 minutes. On June 18, 2004, following the advice of the Centers for Disease Control and Prevention (CDC), the Army issued revised health-based criteria for exposure to airborne agents (U.S. Army, 2004a). The Army also eliminated the use of 0, X, 3X, and 5X decontamination terminology to be in conformance with existing laws and regulations.12 The concept of a material that has been completely decontaminated of agent is obsolete, because it is impossible to verify that every molecule of agent has been removed. The Army has retained the fundamental criterion of 5X, which is that the material is sufficiently decontaminated that it may be released for general use or sold to the general public. Under the updated rules, a material can be released unconditionally under any of the following three circumstances (U.S. Army, 2004a): Documented evidence is available to prove that the material has never contacted liquid agent and has never been exposed to airborne agent at concentrations exceeding the short-term exposure limit (STEL) of 3 × 10−3 µg/m3 time-weighted average over a 15-minute period. The material is heated to a surface temperature of 538°C (1000°F) for at least 15 minutes. The material is cleaned in accordance with an approved equipment decontamination plan and certified by the mission commander to the selected health-based criteria for the reasonably anticipated use environment of the public owner. The selected health-based criterion may be the worker population limit, the STEL, the immediately dangerous to life and health limit, or the general population limit, depending on how the decontaminated material will be used. Finding 3-17. The decision by the Army to adopt the recommendations of the Centers for Disease Control and Prevention on airborne exposure limits to chemical agents provides possible new options for the treatment of dunnage, such as the following: (1) maintaining good documentation to prove that the material has never contacted liquid agent and has never been exposed to airborne agent at concentrations exceeding the short-term exposure limit so that it can be released to a commercial disposal facility without further treatment, and (2) low-temperature treatment that meets one of the new health-based criteria for unrestricted release and is acceptable to the regulatory agency and to the intended disposal facility. Recommendation 3-17. The Army and the Bechtel Pueblo team designing the Pueblo Chemical Agent Destruction Pilot Plant should develop alternative conceptual designs for the decontamination of dunnage on the basis of the three allowable criteria and should en- 12   At the 3X decontamination level, solids are decontaminated to the point that agent concentration in the headspace above the encapsulated solid does not exceed the health-based, 8-hour, time-weighted average limit for worker exposure. The level for mustard agent is 3.0 mg/m3 in air. Materials classified as 3X may be handled by qualified plant workers using appropriate procedures but are not releasable to the environment or for general public reuse. In specific cases in which approval has been granted, a 3X material may be shipped to an approved hazardous waste treatment facility for disposal in a landfill or for further treatment; 0 and X designate lesser degrees of decontamination (U.S. Army, 2002a).

OCR for page 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant gage the local public, interested stakeholder groups, and state and local government officials in a dialogue about the most appropriate decontamination approach. Prior to issuance of the new guidance by the Army, the Bechtel Pueblo team was constrained to treat dunnage to 5X by heating the material to 1000°F for 15 minutes. Accordingly, the team redesigned the continuous steam treater for PCAPP to correct problems identified with an earlier design during the engineering design study phase of the ACWA program. The Bechtel Pueblo team planned to test a full-scale prototype of the redesigned unit at the Parsons Fabrication Facility in Pasco, Washington, during the latter half of 2004. Three similar units are to be installed at PCAPP.13 Heating of agent-contaminated metals to 1000°F in the absence of air or oxygen would only attack the agent and any other organic material—for example, paint associated with the metal. The treated material would then be suitable for sale to commercial scrap dealers. In contrast, heating of agent-contaminated nonmetallic wastes to 1000°F would attack both the agent and the substrate in a complex set of reactions including pyrolytic decomposition. The treated material would be totally different in form and composition from the feed and would only be suitable for disposal in a commercial incinerator or landfill. The problems of treating wood are illustrative. In the presence of air or oxygen, wood catches fire and burns when it reaches a temperature of about 400°C to 500°C (752°F to 932°F). Since this effectively would be incineration, the wood would have had to be heated to 538°C (1000°F) for 15 minutes in the absence of air in order to be releasable to a commercial facility under the old rules. When wood is heated in the absence of air, moisture is driven off first. The wood temperature remains at about 100°C to 110°C (212°F to 230°F) until drying is complete. After that, the temperature of the wood rises. At about 270°C (518°F), the wood begins to decompose with evolution of gaseous by-products. Evolution is complete at about 410°C (770°F). The solid residue remaining is charcoal, with about 70 percent carbon and small amounts of tars. To drive off or decompose the tars, the temperature must be raised above about 600°C (1112°F) (FAO, 1983). Tars are sticky, difficult to handle, and can plug equipment. The presence of chlorine in the plastics from DPE suits may create additional complications with respect to the chemical composition of the gases—for example, corrosion and the possibility of dioxin and furan formation in the CST offgas treatment system (OTS). The Bechtel Pueblo team plans to use superheated steam as a sweep gas in the CST for removing gases generated when decontaminating nonmetallic waste to 5X. Chemical reactions between the steam and the gases and hot solids are anticipated in the complete absence of oxygen. The Bechtel Pueblo team refers to the reactions as steam reforming. However, the reaction products of agent-contaminated dunnage with steam have never been measured and are not amenable to modeling. The committee is waiting for the results of the TRRP testing planned for the equipment. The key reactions are as follows: (Reforming) (Water gas shift) The first reaction is highly endothermic and will cause the temperature to drop in the CST. The second is slightly exothermic. The reforming reaction is generally conducted at 700°C to 1000°C at 3 to 25 bar in the presence of a catalyst. Side reactions that can result in carbon formation are as follows: The committee notes that the prior demonstration test of the CST was, in fact, not a test of steam reforming, because when air leaked into the system some exothermic oxidation reactions took place, and traces of dioxin were found in the offgas. The literature describes many pyrolysis and gasification systems for the conversion of coal, biomass, and municipal solid waste into fuel gas, bio-oil, and char. None is designed to maintain the feed materials at 1000°F for 15 minutes. All focus on the generation of clean-burning fuel for power production. Successful design and operation of the CST and its associated OTS depends on the thoroughness of the TRRP testing planned for this equipment. Tests on a prior design during the ACWA engineering design 13   ACWA Design Committee meeting with Army and Bechtel National, Inc., participants, Irvine, Calif., February 11–13, 2004.

OCR for page 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant study phase of the program failed to yield completely satisfactory results. Furthermore, the TRRP tests as planned are not likely to test all actual operating conditions, making it highly probable that unexpected problems in operation and maintenance of these systems will occur. Finally, the CST OTS configuration as designed is an entirely new application of old technologies in which the technologies are being asked to meet very stringent performance goals. The new CST design is expected to allow the solid feed material to reach 1000°F in the first chamber quickly, while the second chamber provides assurance of an adequate residence time. Agitator speeds will be set to give a residence time in each cylinder of about 15 minutes. Discussions of the CST design at the May 19–21, 2004, design review meeting indicated that the composition of the material resulting from CST operation may include tarry materials as well as char and ash. The TRRP tests are essential to determine the characteristics of the solids as they pass through the CST and the impact of those characteristics on operational performance and ease of maintenance, especially when processing potentially contaminated materials. These tests must include all materials in the planned feed ratios, not just the three major waste streams, to ensure that the designers have anticipated all problems that may be encountered in operation and maintenance. All materials currently planned for processing in the CSTs during PCAPP operations are listed in Table 2-1 in Chapter 2. Also, the TRRP testing should be used to demonstrate the ease of maintenance using the proposed auger shell withdrawal mechanisms. The auger shells in both chambers are being designed to be inserted into the chamber shells on rails to permit easy removal for maintenance. Potential jams in the material flow would be located inside the withdrawn auger shells. Finding 3-18. The proposed two-chamber design for the continuous steam treater with auger shells may be more prone to jamming than the original concept, especially when all of the different feed materials listed in Table 2-1 in Chapter 2 of this report are processed. Moreover, it is unclear how any jammed material would be removed from the withdrawn auger shell. Further disassembly of the auger shell would be required. Furthermore, the condition of the material in the withdrawn auger shell that caused a jam may be very stiff and make extraction of the auger very difficult. Recommendation 3-18. Technical risk reduction program testing of the design of the two-chamber continuous steam treater (CST) for the Pueblo Chemical Agent Destruction Pilot Plant should include the testing of all feed mixes at the design rates in order to characterize the composition of solids in the CST and the sweep gas flow characteristics through the CST. Finding 3-19. The continuous steam treaters for the Pueblo Chemical Agent Destruction Pilot Plant will be required to treat a wide range of feed materials, including organic liquids. Recommendation 3-19. The continuous steam treater (CST) testing should include all of the feed streams anticipated over the life of the Pueblo Chemical Agent Destruction Pilot Plant so as to identify those waste streams that may require alternate means of treatment for acceptable levels of decontamination. Moreover, consideration should be given to other means of treating these wastes if some prove to be the cause of low reliability and high maintenance requirements for the CST. Wood shredded to 3/8-inch particle size, shredded DPE suit material (metal parts removed), and granulated activated carbon (GAC) are the three major feeds in the CST. But there are other materials as well (see Table 2-1 in Chapter 2). Currently, none of the other materials listed in Table 2-1 are included in the TRRP CST testing. Also, the PCAPP design criteria assume that the wood from the pallets and boxes is oak; however, the wood feed for the TRRP tests will be 10 percent plywood, with the remainder as oak. Copper naphthenate and copper arsenate solutions have been known to be used as wood preservatives in plywood. Copper is known to catalyze dioxin formation, and copper and arsenic may be poisonous to the catalytic oxidation catalyst. Finding 3-20. The inclusion of plywood in the feed to the continuous steam treater (CST) may result in a feed stream containing constituents detrimental to the catalytic oxidation unit of the CST offgas treatment system. Recommendation 3-20. Sampling of the pallet and wood box composition at Pueblo Chemical Depot should be performed to verify the presence or absence of preservatives that may contain heavy metals. If

OCR for page 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant such constituents are found, additional testing may be required. Another issue for the CST concerns the material of construction. High-temperature reducing gaseous environments can be more damaging than oxidizing environments. In addition, the prior experience with the earlier CST design indicated extensive corrosion (NRC, 2002c). The CST will have a reducing environment that has a significant potential for carburization. Similar conditions are typically found in syngas generation, chemical reactors, furnaces, steam generators, and downstream reformer components. Under these conditions, an alloy can suffer rapid metal corrosion through a process known as metal dusting (Lai and Patriarca, 1987). This metal dusting tends to occur in a carbonaceous atmosphere (high ratios of carbon monoxide to carbon dioxide and low ratios of steam to hydrogen). When such environments are present in the process stream within the critical temperature range of 400°C to 750°C, metal dusting can be severe (Baker et al., 2004). New nickel-base alloys (Alloy 671 or 693) (Baker et al., undated) have been recently developed for increased resistance to metal dusting. These alloys appear to be more resistant to dusting-type attack than are Alloys C276 and 625. Finding 3-21. Significant corrosion occurred on an earlier prototype continuous steam treater unit used during the Assembled Chemical Weapons Assessment engineering design study testing (observed during the site visit to the Parsons Fabrication Facility in Pasco, Washington), indicating there could be significant material degradation issues. Recommendation 3-21. Since metal dusting already appears to have been encountered with austenitic stainless steels used in the prototype continuous steam treater (CST), more dusting-corrosion-resistant alloys should be considered for the current CST design. Superheated steam in the primary chamber of the CST will flow countercurrent to the solid-feed material flow. Cocurrent steam and dunnage feed material flow is not planned to be tested, although cocurrent flow was originally planned since entrainment of unprocessed fines in the offgas was experienced with countercurrent flow during earlier ACWA testing. (According to the test plan for the CST, the test system will be built so that it can also provide cocurrent flow between the steam and dunnage if so desired.) The CST offgas stream (steam, steam reaction gases, purge nitrogen and, possibly, a small amount of entrained particulates) will exit the primary auger shell near the feed end. The offgas then flows to the OTS. If the pyrolysis reactions occurring in the second (lower) auger shell are greater than expected, gas flow problems in the primary auger shell and potential seal leakage may result. Finding 3-22. It is not possible to fully understand or characterize the full range of continuous steam treater (CST) offgas contaminants from all feeds if just the three major feed streams are tested, as proposed in the current technical risk reduction program test plan for the CST (U.S. Army, 2003a). Recommendation 3-22. The test plans for the continuous steam treater (CST) at the Pueblo Chemical Agent Destruction Pilot Plant should provide for characterizing the CST offgas feed to the offgas treatment system (OTS) as well as the gas stream composition at critical locations in the OTS. OFFGAS TREATMENT SYSTEMS All process offgas streams flow through an offgas treatment system prior to release to the atmosphere through a stack. Equipment for each of the OTSs is summarized in Table 3-3. As shown by the listing of components in the table, the OTS designs vary from the most complex for the CSTs and MPTs to simple catalytic oxidation (CATOX) units for odor removal in the brine recovery system (BRS) OTS. Also, as indicated in Table 3-3, the munitions washout system/agent neutralization reactor (MWS/ANR) OTS design was not defined in the initial design, but it will be provided at the completion of the intermediate design. In addition to the process OTSs, the ventilation air from all process areas that may be contaminated with agent is discharged to the HVAC activated carbon filter farm and then released to the atmosphere through a stack. Agent monitors are provided at various points in all systems to monitor system performance. Since the CST OTS is the most complex of the OTSs and will be subjected to TRRP testing, it is reviewed here in more detail. The committee notes that experience with the CST OTS testing will be used in the design of the metal parts treater OTS as well as other OTSs.

OCR for page 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant TABLE 3-3 Equipment Summary for Offgas Treatment Systems at Pueblo Chemical Agent Destruction Pilot Plant   OTS OTS Equipment CST MPT ERH/ENR MWS/ANRa ICB BRS Trains 3 (1 per CST) 3 (1 per MPT) 2 (1 per ERH) To be determined 6 (1 per 4 ICBs) 1 Induction heater 1/train (1200°F discharge)           Cyclones 2 in series/train 2 in series/train         Thermal oxidizer 1/train           CATOX feed temperature control 1/train 1/train 1/train       CATOX 1/train (1050°F discharge) 1/train (1150°F discharge) 1/train (750°F discharge)     1 (400°F discharge) Quench scrubber 1 1 1       Particulate filter 1 1 1       Heater 1 1 1   1/train   Carbon filter prefilter         1/train   HEPA filter         2/train   Blower (induced draft) 1 1 1   1/train 1 Carbon filter 1 1 1   1/train   NOTES: A blank space indicates the absence of a component. Acronyms are spelled out in the report’s “List of Acronyms.” aThe OTS for the MWS/ANR had not been designed when this report was prepared. SOURCE: Adapted from U.S. Army, 2004b. Offgas Treatment System of the Continuous Steam Treater An effluent heater heats the offgas stream leaving the CST primary trough to 1200°F. The effluent heater is an inductive heating unit, so it allows materials of construction to be used that can withstand the presence of chloride gases generated from DPE suit material decomposition. Resistive-type heating elements were found to be unsuitable in earlier ACWA testing because of chloride corrosion. An oxygen sensor, located upstream of the effluent heater in the OTS system, is used to monitor oxygen levels and to ensure that these levels are maintained below 3 percent to prevent fires or explosion. From the effluent heater, the process gases are directed to a two-stage cyclone to remove most of the particulate matter from the offgas stream. The first cyclone removes more than 99 percent of particulates that are 100 microns and larger, while the second cyclone removes more than 99 percent of particulates that are 10 microns and larger. Each cyclone discharges collected particulates into a small drum. The gas stream flowing from the downstream cyclone is expected to contain less than 1 percent of particulates that are 10 microns or larger. Since it cannot be guaranteed that the collected particulates have undergone decontamination to a condition suitable for release to a commercial disposal facility, they will be collected and fed back to the CST or the MPT. Offgas leaving the cyclones is then mixed with heated ambient air and fed to the bulk oxidizer (a catalytic oxidizer). A flame arrester is provided as a safe-guard prior to the entry of gases into the bulk oxidizer. A differential pressure indicator is provided across the flame arrester to provide indication of imminent plugging conditions. The heated ambient air is added to the oxidizer inlet gas stream to ensure that the feed to the

OCR for page 27
Interim Design Assessment for the Pueblo Chemical Agent Destruction Pilot Plant bulk oxidizer has a hydrogen concentration below its lower explosive limit (LEL) and a gas temperature of about 800°F to prevent the formation of dioxins and furans. The LEL sensor location is after the air addition point but before the bulk oxidizer. In the bulk oxidizer, hydrogen is converted to water; methane and hydrocarbons to water and carbon dioxide; carbon monoxide to carbon dioxide; and chlorinated organics to hydrogen chloride, carbon dioxide, water, and products of incomplete combustion. This oxidizer is expected to remove at least 99 percent of both hydrogen and carbon monoxide and at least 90 percent of methane from the stream. The minimum oxygen concentration in the discharge gas will be 12 percent by volume. A water spray is provided to maintain the bulk oxidizer exit gas temperature at 800°F as the gas flows to the finishing CATOX unit. A Military Air Purification CATOX unit is used as a finishing unit to complete oxidation of residual volatile organic compounds (VOCs) and semi-VOCs in the bulk oxidizer effluent. The discharge oxygen concentration of the finishing CATOX unit is maintained at a minimum of 12 percent to ensure complete oxidation. Oxygen content will be monitored with an electrolytic diffusion cell monitor capable of a 12 percent range. The finishing CATOX unit discharge temperature is maintained below 1050°F by injecting a water mist into the unit’s inlet. According to the prototype CST test plan, if the finishing CATOX unit discharge temperature exceeds 1100°F, the CST feed will be stopped and the CST heaters will be shut down (U.S. Army, 2003d). Presumably, these steps are required to prevent damage to the catalyst and failure to properly treat material in the CST and its offgas. Other measures (operating conditions and procedures) also may be provided in the final design to protect against excessive temperature in the finishing CATOX unit discharge stream. After the gas stream exits the finishing CATOX unit, it passes through a venturi scrubber to rapidly cool the gases and minimize the formation of dioxins and furans. The venturi scrubber also removes hydrochloric acid. In the venturi section, the hot process gases are cooled and condensed using a spray of cooled condensate. The cooled gases then go to a scrubber, where a spray of water further cools and scrubs the gases. After start-up, the spray water will be generated by condensation of the water in the incoming gas stream. The liquid condensate generated in the scrubber tower is collected in a tank and then cooled to the required spray temperature. Condensate pH will be monitored and adjusted using caustic to maintain the pH in the 7 to 10 range. Excess condensate is pumped to the agent hydrolysate tanks for subsequent treatment in the ICB units. Process gas exiting the scrubber tower is directed to the filter, air heater, and carbon filter system. The first filter removes the solid particles greater than 0.5 micron. An electrical heater reduces the relative humidity to ensure better performance of the downstream carbon filter. The carbon filter has inlet and outlet dampers. After the carbon filter, the gas stream is exhausted through an induced draft blower into an exhaust duct and to the plant HVAC carbon filter farm. The carbon filter unit contains six elements. From inlet to outlet, they are a pre-filter, a high-efficiency particulate air (HEPA) filter, three carbon filters, and a final HEPA filter. Each filter element is equipped with a differential pressure gauge to monitor its corresponding pressure drop. According to the TRRP test plan for the CST, this filter unit is for testing purposes only. The production unit may use carbon filters that are different in configuration from those used in the prototype system (U.S. Army, 2003d). In addition, the production unit may use the building’s HVAC, or it may have its own filter system. Finding 3-23. The number of offgas treatment systems required for operation of the Pueblo Chemical Agent Destruction Pilot Plant (PCAPP) is large and the process elements numerous. Most of the process elements have commercial equivalents, but the operating requirements for PCAPP, especially efficiency and reliability, may not be readily achieved. Recommendation 3-23. Since only the continuous steam treater offgas treatment system (OTS) will be tested prior to plant systemization, there may be significant restriction of plant operation if the OTSs of the Pueblo Chemical Agent Destruction Pilot Plant prove to have lower-than-expected reliability. The Bechtel Pueblo team should develop limiting conditions of operation and operating workarounds that allow continued operation with some OTSs down or operating at reduced efficiency.