Assessment of the Army Plan for the Pine Bluff Non-Stockpile Facility (2004)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

17 2 The Pine Bluff Non-Stockpile Facility In this chapter, the committee describes the basic de- sign configuration of the Pine Bluff Non-Stockpile Facil- ity (PBNSF), outlines the intended operation, and dis- cusses a number of issues related to facility design and operation. In conducting its review, the committee exam- ined the initial design documents for the facility (35 per- cent design) and the permit application submitted to the Arkansas Department of Environmental Quality and re- ceived briefings and updates from the Army. The com- mittee also had follow-up meetings with the Army and its design contractors, and members had many technical dis- cussions among themselves. The committee relied on generally accepted construction practices and benchmarks, such as the Construction Industry Institute Best Capital Practices. However, a project like PBNSF cannot easily be reviewed using such practices, be- cause the entire project has been heavily driven by schedule and the need to use alternative technologies that are fairly new. The U.S. Army Corps of Engineers, which is construct- ing this project, is active in organizations such as Construc- tion Industry Institute and says it will incorporate as far as practicable such best practices in the design and construc- tion phases. However, owing to the nature of the PBNSF project, the committee could not prepare a detailed compari- son of the project’s stages against generally accepted prac- tices and benchmarks. As discussed below, the main factor driving this construction project—the treaty deadline—is by and large beyond the control of the Corps and the non-stock- pile program. BUILDING AND SITE LAYOUT The PBNSF site will occupy approximately 25 acres that previously were used for disposing of construction fill. As currently configured, the main process facility will be a 40,000 ft2 building (see Figure 2-1) surrounded by an 8-ft chain-link security fence (U.S. Army, 2003b). The process facility will comprise the following: • receiving dock • munitions warming and storage room • unpack area • fill extraction preparation area • fill extraction area • agent treatment area containing: —two explosive containment chambers (ECC-1 and a larger ECC-2) for removing the chemical agent fill from an item that has an energetic component at- tached and from which the agent can be drained —a chemical process trailer (CPT) with two neutraliza- tion reactors for destroying the chemical agent fill emptied from the items in the ECC-1 and ECC-2 —a detonation chamber (DET) for the destruction of energetic components that do not contain agent and are not contaminated with agent —a projectile washout system (PWS) for removing chemical agent from nonexplosively configured munitions • decontamination room • repacking room • storage room and associated handling areas The PBNSF relies mostly on legacy equipment from the abandoned Munitions Management Device (MMD) project, including the ECC units, the CPT, and the DET. This equip- ment consists of trailer-mounted units that can be disas- sembled, transported by road or by air, and reassembled at new locations. The decision to reuse this equipment has ne- cessitated continuing modifications, particularly to the ECC units, as well as accessibility constraints in the CPT. All processing areas will operate under negative pressure to provide chemical vapor containment in the event of a re- lease. The exhaust air from all spaces that could contain agent will be passed through high-efficiency particulate air and carbon filter systems for purification before being passed to the atmosphere. All areas where chemical materiel is handled and processed will be sealed to prevent the migra-

18 ASSESSMENT OF THE ARMY PLAN FOR THE PINE BLUFF NON-STOCKPILE FACILITY tion of agent or vapor to or from other areas in the event of a release. The Product Manager for Non-Stockpile Chemical Mate- riel (PMNSCM) has defined the maximum credible event (MCE) for the PBNSF design as the detonation of a fully configured German Traktor rocket (GTR) motor and war- head combination while being processed in the PBNSF. While it is difficult to estimate the likelihood of the occur- rence of the MCE, it is important to review such a low-prob- ability event and investigate designs that protect against it when consequences may be severe. The committee recognizes that the PBNSF design calls for the MCE to be completely contained, but only when the GTR is being processed within the ECC-2. However, should the MCE occur within the PBNSF but outside the ECC-2, there could be a release of fragments and agent to the imme- diate area outside the PBNSF building; there are similar con- cerns should a fully configured GTR be detonated during FIGURE 2-1 Diagram of the PBNSF processing area layout. SOURCE: U.S. Army (2003b).

THE PINE BLUFF NON-STOCKPILE FACILITY 19 transit. However, nothing in this report should be construed as expressing the view that such a release is likely. The Army, citing the containment requirements in De- partment of the Army Pamphlet 385-61, Section 6.6 (U.S. Army, 2002a), has reported that this condition (less than to- tal containment) is the required level of protection for both the stockpile and the non-stockpile disposal programs. Should the MCE occur outside the ECC-2, it would almost certainly result in severe worker injuries or fatalities and trig- ger a public reaction and regulatory review. Such a review would seriously delay completion of the PBNSF task regard- less of the impact on workers or the environment. For this reason the committee’s judgment is that this less-than-total- containment for the PBNSF building is not satisfactory. This containment situation bears out the committee’s recommen- dation to develop a system that would decouple the GTR motor/warhead combinations in a separate facility designed to contain both explosions and releases of lethal chemicals and minimize transportation and handling. The Army is already evaluating the possibility of decoupling fully configured GTRs in existing igloos prior to their treatment in PBNSF. If this effort is unsuccessful, the committee urges the Army to (1) develop a transport system that would contain the explosion in the event of GTR deto- nation in transit and (2) revisit the PBNSF design and/or engineering controls to ensure the safety of workers who might be outside the building if a GTR detonated inside PBNSF as well as the safety of the general public. Ancillary features and areas that would be located on the proposed site include these: • storage tanks for neutralent and the spent decontami- nation solution, including spent wastewater and neu- tralized wastes • chemical supply tanks • process chillers • two standby generators • a minibunker for storage of explosive charges • a waste storage and handling area • an administrative building • a gatehouse The process facility building, the fence, and support struc- tures and utilities are being designed and built by contractors to the U.S. Army Corps of Engineers’ Little Rock District based on the criteria provided by the PMNSCM. The processing steps that a munition or other item will undergo as it is transferred into the system are discussed in the following sections. CONFIRMATION OF MUNITION CONTENTS The contents of recovered chemical weapon materiel items at PBNSF will be confirmed in two steps. The first step in assessment of the munitions is performed in the Pine Bluff Munitions Assessment System (PBMAS), a facility separate from the PBNSF. The munitions will be unpacked, examined, analyzed, classified, and repacked into overpacks, each containing only a single munition, for storage and sub- sequent transportation to PBNSF. Adjacent to PBMAS is an explosive destruction system (EDS) that will destroy muni- tions considered too hazardous to move into storage for fu- ture destruction in PBNSF. The second step of the assess- ment occurs when the munition is first unpacked in PBNSF prior to disposal. The details of these two steps are described in the following sections. Characterization in the Pine Bluff Munitions Assessment System Each munition to be processed in PBNSF will be charac- terized individually in PBMAS by nonintrusive methods, primarily x-rays and portable isotopic neutron spectroscopy (PINS). The x-ray scan provides two kinds of information. It reg- isters the presence of energetic materials such as a fuze, a burster, explosives, or propellant and detects whether the munition contains a chemical agent. The PINS measurement provides qualitative information to assist in the identification of the chemical agent contained in the item based on the presence of elements such as sulfur, arsenic, chlorine, and nitrogen. The PINS characterization is generally accurate, but interpretation of the results may be affected by the presence of corrosion products and other effects. The PBMAS operations allow the munitions to be segre- gated into those containing energetics only, chemical agent only, or both (with the agent assessed as drainable or gelled). This characterization within PBMAS allows the appropriate processing steps within PBNSF to be selected. The muni- tions are also classified by agent type because PBNSF pro- cessing will be performed in campaigns based on the nature of the agents contained in a set of munitions. Following char- acterization in PBMAS, the items will be packed into indi- vidual overpacks that are color coded to denote the risk asso- ciated with future handling and transferred to storage until they can be processed in PBNSF. Characterization in the Pine Bluff Non-Stockpile Facility When the overpacked munitions are received in PBNSF, they are again assessed by x-ray in the receiving room and are checked for agent leakage by sampling the air in the over- pack (U.S. Army, 2003a). Subsequently, once the agent chamber in the munition has been accessed, a headspace vapor sample is withdrawn to confirm the identity of the agent by coupled gas chromatography-mass spectroscopy. This reconfirmation step is important for ensuring that the chemical agent undergoes the appropriate neutralization treatment and that suitable monitoring devices are in place for worker protection. The munition is then sent to the next

20 ASSESSMENT OF THE ARMY PLAN FOR THE PINE BLUFF NON-STOCKPILE FACILITY processing module, which is selected based on munition con- figuration (agent and explosive content) and condition (clean, corroded, and so forth). The nature of the non-stockpile disposal program is such that a wide variety of materials must be dealt with, and there is considerable uncertainty surrounding the characteristics of these materials. For example, originally it was assumed that the overwhelming majority of the 4.2-in. mortar rounds containing mustard agent did not contain an explosive com- ponent (fuze, burster tube, etc.). This situation would have allowed these mortars to be processed in the PWS after the bottom of the munition was cut off to allow access for power washing the agent fill out of the munition body. However, more recent test data indicate that there is so much uncer- tainty about the presence of a fuze or explosive in a burster tube that it is prudent to assume that all of these 4.2-in. mor- tar rounds contain fuzes and/or bursters.1 Because it is not possible to definitively determine whether the burster tubes in the mortar rounds contain energetic material or inert ma- terial or whether they are empty, it must be assumed that any mortar round containing a burster tube may also contain en- ergetics. In addition, experience from the stockpile disposal facilities has shown that in many munitions mustard agent has gelled or solidified, which prevents the agent from drain- ing and leads to formation of heels. This combination of agent heel in an explosively configured item could make it impossible to achieve the anticipated destruction rate using the PBNSF equipment as presently configured.2 It should be noted however, that in a recent PINS assess- ment of the contents of the munitions to be processed at the PBNSF (Verrill and Salcedo, 2001), 597 of 733 4.2-in. mor- tar rounds were empty and 218 of the 399 GTRs that did not contain propellant were also empty. Of a total of 1,231 mu- nitions evaluated, 865 were empty. According to communi- cations from the Army, “empty” in this context does not mean that an item does not have trace or residual contamina- tion. For classification purposes—that is, to select a process- ing campaign—“empty” means no liquid is seen in the x-ray. Since some munitions that are “empty” may indeed contain residual or trace quantities of agent that are not detectable by PINS, the munitions will have to be processed in the PBNSF (or in the EDS) as if they do contain agent. If they contain explosives, they will be drilled and drained in an ECC; if they are inert, they will be cut open and washed out in the PWS. In either case, any residual agent will be treated with a reagent in the neutralization reactor. PROCESS DESCRIPTION Accessing the Chemical Agent The first step in processing a munition at PBNSF is to temperature-condition the contents by letting the munition stay in the warming room for at least several hours. This step is necessary because HD agent freezes at 58oF and would not drain from a cold mortar or rocket. The warmed muni- tion will then be x-rayed to ascertain the amount and condi- tion (liquid, solid, or gel) of the agent fill. After this charac- terization, the munition may be stored in the warming and storage room until it can be accepted in the unpack and fill extraction areas. Based on the presence or absence of energetic materials and on an assessment of the drainability of the agent, two options exist for gaining access to the chemical agent con- tained within a munition. In general, munitions without en- ergetics will go to the PWS for cutting, draining, and wash- out. For example, nonexplosively configured 4.2-in. mortar rounds and GTR warheads might be sent to the PWS for accessing the agent. Alternatively, the same inert mortar round could be sent to the ECC-1 for drilling and draining. Similarly, an inert GTR could be sent to the larger ECC-2 for drilling and draining. Rockets or mortar rounds that contain energetics will ordi- narily be processed in one of the two ECC units, which are designed to contain the force of an explosion should one occur during the drilling and draining operations. However, such an explosion would severely damage the internals of the ECC, and the operation of the PBNSF would be severely impacted by the loss of capacity and by an incident investigation. Processing via Explosive Containment Chamber Units Energetically configured munitions containing agent will be processed in the ECC units (the ECC-1 and the larger ECC-2). These contain an auxiliary processing vessel (APV)—a small, movable pressure vessel that contains drills, agent extracting devices, and neutralent injection and drainage capabilities—into which the munition is loaded. This drill-and-drain assembly (containing the munition) is then loaded into an ECC unit, which is essentially a large pressure vessel that will contain any explosion up to the de- sign loading (see Figure 2-2). As stated previously, these ECC units were developed as part of the MMD program and are currently being modified to improve their accessibility and operability (see Figure 2-3). The procedure for handling an energetically configured munition is to manually load it into the APV. The APV is highly complex and consists of numerous hoses, motors, and movable parts, all of which must work as required within the ECC to drill and drain the munition. The APV containing the munition is then moved into the ECC unit for drilling, sam- pling, and draining. All processes are performed within the 1Darryl Palmer, Office of the Product Manager for Non-Stockpile Chemical Material, “Multi-pack 4.2-in. mortars,” e-mail to the committee on June 5, 2003. 2Chapter 6 addresses possible options for the modification of the facility to maintain schedule in view of the reassessment that many, if not all, 4.2-in. mortar rounds should be assumed to be explosively configured.

THE PINE BLUFF NON-STOCKPILE FACILITY 21 FIGURE 2-2 Auxiliary processing vessel removed from explosive containment chamber. Note the number of hoses, connections, etc. associated with the requirement to drill and drain munitions remotely when the APV (containing the munition) is inside the ECC containment vessel. sealed ECC. The first hole is drilled into the top of the muni- tion to allow the collection of a vapor sample to confirm the identity of the agent. Another hole is then drilled in the bot- tom to drain the agent. The drained agent will be piped di- rectly to a neutralization reactor in the CPT or to a holding tank if neither of the two neutralization reactors is available at the time. If the agent is gelled there is no provision in the current design to enable the agent to be removed from the munition within the ECC. The drained munition is rinsed with a neutralization re- agent appropriate for the particular chemical agent.3 The neutralent is also sent to the CPT or a holding tank to be processed along with the neat agent. The operation of the ECC assumes that the agent will drain from the munition when the drilling is completed. This assumption is currently being reassessed by PMNSCM based on the experience of the stockpile disposal facilities with heels of gelled mustard agent (see Finding 2-2). This reas- sessment may require a redesign of the ECC units to include a high-pressure water wash system. The Army is assessing various methods of washing out any gelled or solidified agent that will not drain, but this work was not available for review by the committee. However, the problems associated with ensuring removal of the gelled agent through the small ac- cess holes available, or even smaller holes if a probe has to be inserted to maintain a seal, are significant. At the time this report was prepared, no practicable design for this system had been developed. A completely different approach may be required to overcome the problem of gelled mustard agent, which might be exacerbated by the construction of the 4.2- in. mortar rounds. These rounds have internal baffle plates welded to the sides of the agent cavity. This feature could make it difficult to wash out the additional surface areas upon 3The rinsing with neutralization reagent is performed in the ECC and not in the PWS, as the ability to apply large amounts of flushing water into the munition is limited in the ECC by the small size and fixed location of the nozzle. This limitation does not apply to the PWS, where full access and a large amount of water are available.

22 ASSESSMENT OF THE ARMY PLAN FOR THE PINE BLUFF NON-STOCKPILE FACILITY which heel can be deposited. An Army test report states, “the interstices between the baffle central core and the burster well walls could present areas difficult to cleanse even by high-pressure water” (U.S. Army, 2002c). In addition to these conceptual design issues, the ECC drill-and-drain assembly is already a highly complex sys- tem. Adding complexity in the form of a new wash system may affect the reliability of the system. Processing via the Projectile Washout System The PWS is an equipment assembly acquired from the Assembled Chemical Weapons Assessment (ACWA) pro- gram. The ACWA program demonstrated the ability of this equipment to effectively cut open and wash out 4.2-in. mor- tar rounds that do not contain any energetic material (fuze, burster, etc.).4 Eighty-five such rounds were processed in PWS testing; the test report (U.S. Army, 2002c) said, “the agent cavity of all 85 munitions was inspected after wash- out. All 85 were visibly clean and metal bright.” The test report also said, “The minute amount of agent heel in the munitions after washout did, however, affect observation of the system’s ability to destroy HD.” The system referred to is a thermal metal parts treater (not part of PBNSF) that uses superheated steam to bring the munition body to a 5X condi- tion following washout in the PWS.5 The point, however, is that regardless of the metal parts treater’s ability to decon- taminate the munition body, some agent heel may remain in the munition following washout in the PWS (U.S. Army, 2002c). FIGURE 2-3 Internal layout of the chemical process trailer. Note the complexity and congestion of the piping and instruments. 4The ability of a high-volume, high-pressure water wash to effectively clean out the munition in the PWS eliminates the requirement to use a neutralent wash, which is used for cleaning munitions in the ECC. 55X refers to a level of decontamination at which solids may be released for general use or sold (e.g., as scrap metal) to the general public in accordance with applicable federal, state, and local regulations. A common misconception is that 5X means simply that the solid has been placed in a temperature zone of 1000°F or higher for 15 min. In fact, a 5X condition indicates that the solid has been completely heated to and then held at a temperature of at least 1000°F for 15 min.

THE PINE BLUFF NON-STOCKPILE FACILITY 23 The PWS is a glove box system containing equipment to drill, cut, drain, and pressure-wash projectiles (no neutralent rinsing is performed owing to the proven effectiveness of the pressure wash in the PWS). It is not designed to handle ener- getic material. After the munition is drilled to obtain a sample of the agent vapor for analysis, the base of the munition is cut off with a pinch roller, which allows full access for wash- ing (unlike the ECC, which allows very limited access). The drained liquid agent is then piped to the CPT for neutraliza- tion, as is done with drained agent from the ECC. After the agent is drained, the munition body is pressure-washed with water to flush out residual agent and solid residues. The washings are piped to the spent decontamination solution (SDS) reactors for neutralization along with decontamina- tion solutions from other operations. These reactors are new units and are to be located in a separate section of PBNSF. The original design intent was for inert mortar rounds (i.e., those with no energetic material) and GTRs to be pro- cessed in the PWS (U.S. Army, 2002b). The original assess- ment of the materiel to be disposed of at PBNSF was that the overwhelming majority of the 4.2-in. mortar rounds contain- ing mustard agent were inert. Recent analysis suggests that this assumption may be incorrect and that it may instead have to be assumed that almost all of the 4.2-in. mortar rounds contain energetic material. This change will signifi- cantly impact the ability of the PWS to process munitions. If it is concluded that the most recent assessment is correct, then the PWS (as currently configured) will not be able to process the large number of 4.2-in. mortar rounds specified in the current schedule. To maintain the current schedule, one of two actions would be needed: • Agent-containing 4.2-in. mortar rounds would have to be processed in another item of equipment (e.g., in a modified ECC with a power washing capability added, or in an EDS unit in addition to the one attached to PBMAS); the 597 rounds that do not contain agent can be processed as planned in unmodified ECC units fol- lowed by heel-washing and detonation, or • The PWS would have to be modified to allow it to be remotely operated within a pressure vessel able to con- tain an explosion of the energetic material—that is, the PWS becomes a modified version of the ECC. The effect on schedule and cost of redesigning either the ECC or the PWS is unknown but would probably be signifi- cant because of the associated development time, the cost of demonstrating the effectiveness of the modifications, and the need to conduct process safety studies and hazard analyses. Neutralizing the Agents The CPT unit was also obtained from the discontinued MMD project. The equipment is contained in a trailer so that it is transportable by road. The requirement that the unit be transportable has led to congestion in the interior of the trailer (see Figure 2-3). Equipment located behind the reactors and on the ceiling may be difficult to access for repair or mainte- nance, particularly when the operating staff must wear Level A personal protective equipment (PPE) inside the CPT (see Figure 2-4). The CPT contains two continuously stirred tank reactors, one large and one small (123 gal and 66 gal, respectively). Each is agitated by stirring as well as by circulating the reac- tor contents through an external loop. The neutralization is carried out in much the same way as in the mobile MMD-1; in fact, much of the MMD-1 neutralization hardware is to be used in the PBNSF CPT. The MMD-1 system was demon- strated extensively in Utah (Cash et al., 2001). The MMD-1 testing was performed using phosgene6 as the agent and di- lute sodium hydroxide as the reagent/neutralent. While phos- gene should not be encountered in the munitions to be pro- cessed through PBNSF, extensive bench-scale testing with mustard agent was performed at Edgewood Arsenal under conditions applicable to the MMD-1 and PBNSF neutraliza- tion processes. Neutralization of mustard agent with MEA containing 10 percent water produced a neutralent solution containing 67-89 percent MEA, 9-10 percent water, varying amounts of MEA hydrochloride, and a sulfur compound de- rived from reaction of the mustard agent with MEA (see the chemical equation later in this section). No residual mustard agent was detectable by an analytical procedure with a de- tection limit of 50 parts per billion (ppb). Small amounts of organic sulfur compounds and chlorinated hydrocarbons were detected. These compounds probably arose from impu- rities in the HD agent or from high-explosive degradation products. The results of these neutralizations are summa- rized in a National Research Council report (NRC, 2001b). The drained agent and the reagent neutralent rinses are accumulated in one of the two neutralization reactors until the volume is sufficient for the agitator to function effi- ciently. Additional neutralization reagent is added as needed to ensure destruction of the agent. Agitation and circulation through the external loop are begun, and the reactor is warmed to about 50oC to initiate the reaction (U.S. Army, 2001a). Since the neutralization reaction is exothermic, no heat input is required once the reaction begins. In the MMD-1, neutralization of HD resulted in peak temperatures of about 80oC. Cooling via the external circulation loop may be required, so this is provided for in the equipment design. The reactor contents are held at the desired temperature for 90 min after adding the contents of the agent fill from the last munition of the batch in order to complete the neutral- ization reaction. The reactor contents are cooled and sampled 6Phosgene, an asphyxiating gas, was used extensively as an antiperson- nel agent in World War I and is found in some recovered weapons of World War II vintage.

24 ASSESSMENT OF THE ARMY PLAN FOR THE PINE BLUFF NON-STOCKPILE FACILITY to ensure that agent concentration has been reduced to a level below the release standard—50 parts per million (ppm) for blister agents such as HD, HN-3 (nitrogen mustard), or phenyldichloroarsine (PD). The reactor contents are drained into a waste retention tank for storage until the agent con- centration has been established. If the concentration exceeds the release standard, the batch is returned to the neutraliza- tion reactor for additional treatment. When the concentra- tion is determined to be less than the release standard, the neutralent is transferred to a neutralized waste storage tank to await shipment to a treatment, storage, and disposal facil- ity for final disposal. The reagent used to neutralize the sulfur (HD) and nitro- gen mustard (HN-3) agents in the continuously stirred tank reactor is 90 percent MEA and 10 percent water, the same solution as that chosen for the MMD project and the EDS mobile destruction systems. This reagent is miscible with water, is a good solvent for the agents (even better than a 100 percent water wash), and reacts readily with them. It also does not produce significant quantities of products listed as Schedule 2 compounds under the Chemical Weapons Con- vention (CWC). From an engineering viewpoint, it has low corrosivity with stainless steel under the chosen operating conditions and has low flammability.7 The main chemical reaction with sulfur mustard is this: FIGURE 2-4 Reactor vessel in the chemical process trailer. In this view note the proximity of the reactor vessel to the wall of the trailer and the consequent restricted access for inspection, maintenance, or repair of the pumps and instrumentation. 7There exists a broad base of experience with MEA in the chemical indus- try. In addition to its industrial uses, aqueous MEA has been extensively studied for chemical demilitarization applications by both the U.S. (Durst et al., 1988) and Russian (Petrov et al., 1998) demilitarization programs.

THE PINE BLUFF NON-STOCKPILE FACILITY 25 S(CH2CH2Cl)2 + 2H2NCH2CH2OH → S(CH2CH2)2NCH2CH2OH• HCl + MEA•HCl HD MEA The choice of neutralization reagent for the arsenic-con- taining agents in the GTRs is more complex. Some rockets may contain Winterlost, a solution of HD, PD, and diphenyl- chloroarsine (DA). The planned destruction of Winterlost at PBNSF is based on the operating protocol for the EDS (Verrill and Salcedo, 2001); it uses 90 percent MEA as when treating HD alone. Other GTRs are reported to contain a mixture of PD, DA, arsenic trichloride, and triphenylarsine. Although the initial plan for treating such mixtures involved reaction with 10 percent aqueous sodium hydroxide (NaOH), it is not clear that this reagent will completely destroy the phenylarsenic compounds in the mixture. Current research indicates that treatment with the proprietary reagent HPO2 8 can destroy the phenylarsenic compounds, but this method may require conditions that are unacceptably severe from an engineering viewpoint.9 However, neutralization of the ar- senic-containing agents generates a liquid waste containing arsenic, the toxicity of which cannot be mitigated by chemi- cal treatments. The small amounts of mustard agent washed from the PWS, the ECCs, the heel-dissolving tanks, and the metal decontamination units (MDUs) using hot water or alkaline decontamination solution will give rise to some thiodiglycol. This compound has relatively low toxicity but is a Schedule 2 precursor under the CWC (cf. Chapter 5) because it can be reconverted to mustard agent under some circumstances. The thiodiglycol concentration in the effluent from the SDS re- actor is expected to be quite variable depending on the source of the washings but always less than 1 percent. Rinsing the Munition Bodies Once the chemical agent has been drained from the muni- tions in the ECC, the munition bodies are rinsed with neu- tralizing reagent to remove gross quantities of agent adher- ing to the interior surfaces. (As described previously, rinsing munitions with neutralizing reagent is not performed in the PWS.) As noted earlier, these first rinses are combined with drained agent for treatment in the CPT. The munition bodies are then cleaned to remove trace amounts of agent as well as residues of corrosion products and agent decomposition products. The latter include heels—the tarry or solid agent material that has been found in many HD-containing muni- tions in the stockpile disposal program. The intent of the rinses and washes is to prepare the metal parts for further handling and disposal. Munitions Treated in the Explosive Containment Chamber Once the munition has been drained and rinsed with neu- tralizing reagent in the ECC, it is rinsed with three portions of a decontamination solution, typically dilute NaOH solu- tion injected via the drill assembly. The rinse water from this final rinse of the munition before its removal from the ECC is piped to an SDS reactor for treatment. The rinsed muni- tion is sent to one of five heel-dissolving tanks, where any solid residues that remain are dissolved with hot, circulating 10 percent NaOH solution during an overnight soak. The HD solids dissolve in the hot caustic solution, enabling the residuals to pass through the small drainage channels cre- ated by the drilling. The nearly agent-free munition body is then rinsed with water and sent to the DET, where an explo- sive charge is used to destroy the energetics remaining in the munition. The contents of the SDS reactor will be heated, recircu- lated, sampled, and analyzed for chemical agent (U.S. Army, 2003a). If the agent exceeds the 50 ppm standard set for blister agent, 20 percent NaOH solution will be added to the reactor and the treatment will be repeated until the agent concentration falls below the release standard. The solution will then be held in an SDS storage tank pending shipment to a treatment, storage, and disposal facility for disposal. Munitions Treated in the Projectile Washout System In the PWS, the drained mortar round or rocket warhead will be pressure-washed with hot water only, which has been demonstrated to be fully effective because the internals of the munition are fully accessible in the PWS. The agent- contaminated water will be sent to an SDS reactor, where it will be treated along with the aqueous rinses from the ECC processing. Solids Handling After rinsing (for munitions processed in the ECC) and washing to remove chemical agent and other residues, the munition bodies undergo several steps to prepare them for shipment off-site for ultimate disposal. The exact sequence of steps depends on whether the munition bodies contain energetics such as bursters or propellants. Munitions Containing Energetics Rocket or mortar bodies containing energetics that have been washed in tanks to dissolve any agent heels will be sent to the DET, where an explosive charge will detonate the energetics and fragment the munition body. The DET is an existing item of equipment from the discontinued MMD program. It is not designed to destroy munitions that con- tain a significant quantity of agent, although the tempera- ture generated within it by the explosion will destroy agent 8HPO2 is formulated from Oxone (a formulation of potassium monopersulfate) and aqueous hydrogen peroxide. 9Lucy Forrest, Non-Stockpile Chemical Materiel Product, “Traktor Rocket Sampling,” briefing to the committee on April 22, 2003.

26 ASSESSMENT OF THE ARMY PLAN FOR THE PINE BLUFF NON-STOCKPILE FACILITY to some extent. The efficacy of the detonation procedure will be assessed by visual inspection of the fragments and residues. Large metal pieces will be sent to a metal cutting station, where they are reduced in size with a remotely operated saw in a controlled atmosphere enclosure. Small metal fragments from the DET, along with size-reduced pieces from the cut- ting station, are sent to the MDUs. In the MDUs, the metal parts will again be washed with 10 percent NaOH solution and rinsed with water. (The de- contamination solution and rinse water will be sent to the SDS reactors for decontamination.) After the metal parts have been dried in the MDU, the atmosphere in the unit will be tested. If no agent is detected by headspace sampling with a MINICAMS10 unit after 4 hours at 70°F, the metal pieces are designated 3X11 and can be shipped under Army control to another site. Typically, they would go to the Rock Island Arsenal for thermal decontamination and smelting to recycle the metal. Thermal treatment at 1000°F for 15 minutes is defined as leading to a 5X condition, which means that the metal can be released from Army control. If 5X thermal treat- ment can be done in an existing furnace at the Pine Bluff Arsenal, the scrap metal could be released directly to a recycler. Munitions Not Containing Energetics Rocket warheads and mortar rounds not containing ener- getics that have been drained and washed with water in the PWS will be sent to the metal cutting stations for size reduc- tion. Munition bodies that are heavily corroded will undergo an additional step to remove rust and scale that might entrap chemical agent. The corroded parts will be sent to a station in which they are blasted with solid carbon dioxide (dry ice) pellets. The impact of the CO2 pellets will flake off the cor- rosion products to leave a relatively clean metal surface. The clean metal parts will be sent to the metal cutting station or directly to the MDUs, depending on size. The fine-grain cor- rosion debris deposited after evaporation of the CO2 will be transferred to a drum of 10 percent NaOH decontamination solution to destroy any residual agent. The resulting slurry of debris in decontamination solution will be tested for the presence of agent before being sent off-site for further treat- ment (if necessary) and disposal. PROCESS INTEGRATION This section considers how the existing design for PBNSF is intended to operate in terms of material flow, the indi- vidual equipment, and the interaction of personnel with the system. Material Flow As indicated previously, PBNSF contains a variety of trailer-housed processing equipment obtained from the dis- continued MMD program and from the ACWA program. Because the processing equipment and the PBNSF building were not designed at the same time, compromises were made that could adversely affect operations. Examples of such compromises include these: • The location of the reactor vessels in the CPT (origi- nally for purposes of transportability) has resulted in limited workspace behind these vessels that may im- pede maintenance activities (see Figure 2-4). • Space constraints in the trailers that carry the ECC units limit the lengths of hoses and lines and constrain the extent to which the APV can be removed from the containment vessel (see Figure 2-2). They also limit access to the containment vessel and may increase the time and effort needed to maintain equipment in the ECC, particularly if problems arise with the APV’s drill/drain/wash assembly while it is within the con- tainment vessel. PBNSF differs from a stockpile disposal facility in sev- eral ways: • Munitions characteristics. The munitions to be pro- cessed in PBNSF are in variable condition, with a large proportion being old, corroded munitions containing a variety of fills, some of which are not encountered in the stockpile disposal program (e.g., arsenicals and nitrogen mustard). In some cases, the fill constituents will not be known with certainty until intrusive sam- pling of the munition contents is carried out. There- fore, each munition has the potential to present unique processing challenges, especially when explosive com- ponents are also involved. The effect of the variability of munition characteristics was demonstrated in the recent reassessment of the 4.2-in. mortar rounds. The original basis for the design of the equipment was that the great majority of these munitions would contain mustard agent that would drain and have no energetics (fuzes, bursters, etc.). The most recent assessment is that the great majority of the 4.2-in. mortar rounds might be energetically configured. For example, in one Army assessment, 128 of 130 mortar round x-rays that showed visible burster wells also showed that an ex- 10MINICAMS is a low-level, near-real-time monitor typically used to provide early warning of airborne exposure hazards. According to the U.S. Army (2000), the “basic operation of the MINICAMS involves collection of analytes onto a solid sorbent held in a preconcentrator tube (PCT). Analytes collected onto the PCT are then heat-desorbed into the gas chro- matographic column for separation before passing through a halogen selec- tive detector (XSD).” The XSD detects the chlorine present in all the blister agents. 113X refers to a level of decontamination at which solids are suitable for transport for further processing.

THE PINE BLUFF NON-STOCKPILE FACILITY 27 plosive burster tube was present, indicating that many more mortar rounds than expected may be explosively configured.12 If this is the case, throughput could be limited since in a single 10-hour shift, no more than five energetically configured munitions would be pro- cessed (U.S. Army, 2003c). The operating goal of pro- cessing five inert and five energetic munitions per shift may not be achievable if there are an insufficient num- ber of inert mortar rounds to allow this mix. As a re- sult, PBNSF might be limited to processing only five energetically configured mortar rounds per day, which would have adverse effects on project schedule and cost. In another recent assessment of munition agent contents, it was found that 597 of 733 4.2-in. mortar rounds were nominally empty—that is, they did not have a detectable agent fill—and that for GTRs, 263 of 477 rockets were also nominally empty (Verrill and Salcedo, 2001). Such uncertainties about the quantity of agent in the munitions will require conservative as- sumptions when deciding on the processing steps re- quired to ensure safe operation of the facility. • Throughput rates. While the stockpile facilities have been designed to process hundreds of essentially iden- tical munitions per day, the PBNSF is designed to pro- cess only 10 or so munitions per day. This limitation is due to the munitions’ characteristics and to the equip- ment selected to process the munitions. • Processing sequence. In stockpile facilities the se- quence of processing steps for a given munition type rarely varies. In PBNSF, seemingly identical muni- tions may be processed in a completely different man- ner owing to uncertainties about whether agent is present, what type of agent is in the munition when it is accessed, and whether the munition contains, or is thought to contain, an energetic component. • Munition handling. In PBNSF, as currently config- ured, there are a large number of manual handling tasks associated with moving a munition within the facility. This includes unpacking, placing the munition on a cart, removing the munition from the cart, placing the munition in a device for draining/explosion, etc. An average munition will be handled approximately five times. Systems Integration and Facility Operations PBNSF contains various pieces of equipment that will be used in the processing steps. The processing steps will be tailored to the configuration of the item being processed. The possible configurations are these: • Item containing, or assumed to contain, agent that can be drained but known to have no energetic component. • Item containing, or assumed to contain, agent that can- not be assumed to be drainable but known to have no energetic component. • Item containing, or assumed to contain, agent that can be drained but known to have, or suspected of having, an associated energetic component. • Item containing, or assumed to contain, agent that can- not be assumed to be drainable but known to have, or suspected of having, an energetic component. Each configuration may require a different set of equip- ment to safely access and neutralize the agent. Because of the known variability of the munitions to be processed at PBNSF, the process design attempts to provide considerable flexibility in the sequence of unit operations. This flexibility is intended to allow any munition/energetic combination to be safely processed while maintaining the overall schedule for disposal operations. The nature and con- dition of the agent and the presence or absence of energetics would determine the equipment to be used for processing an individual munition. For example, 4.2-in. mortar rounds that do not contain ener- getics can be processed in one of two ways. They can be sent to the PWS for cutting and a high-pressure washout or they can be sent to one of the two ECCs for drilling and draining (assuming the agent will drain freely). In either case, the munition would finally be sent to the cutting station for size reduction. The recent assessment that a large majority of the 4.2-in. mortar rounds should be considered to be explosively con- figured and that some will contain agent heels could, if cor- rect, significantly increase the time required to process these munitions. An extension of the schedule can be avoided only if some significant modifications are made to the current configuration of PBNSF. For munitions containing energet- ics, an attempt could be made to dissolve any heels by plac- ing these munitions in one of the five tanks used to dissolve agent heels. However, this would increase processing time since only one munition containing energetics can be placed in a tank, thus limiting the number being soaked to five per night. It might be possible to process explosively configured munitions in the single EDS associated with PBMAS, but the schedule is likely to slip. It would also be possible for the munition to be placed in the DET and explosively destroyed. However, this would introduce agent into the DET, necessi- tating the use of decontamination solution in the DET and the collection, treatment, and disposal of the SDS. At present, this option is not being considered in the design of the DET, although it is being considered by the Army. Even if this option were selected, it would slow down operations as a result of the need to manually decontaminate the DET. For inert munitions containing a solid fill, the Army could treat the munition at the PWS and spray rinse the munition at 300 lb/ft2 gauge (psig) (low pressure) or at high pressure 12Darryl Palmer, Office of the Product Manager for Non-Stockpile Chemical Material, “Multi-pack 4.2-in. mortars,” e-mail to the committee on June 5, 2003.

28 ASSESSMENT OF THE ARMY PLAN FOR THE PINE BLUFF NON-STOCKPILE FACILITY with a 10 percent NaOH solution. A PWS with a caustic spray washout at these pressures resulted in effective re- moval of solid mustard (HD) agent heels in the rounds tested (U.S. Army, 2002c). At the time this report was being prepared it had not been decided whether the GTRs will arrive at PBNSF with the warheads separated from the rocket motor or whether the entire munition (warhead plus motor) will have to be pro- cessed. If only the nonexplosively configured warheads (the part of the GTR that contains agent) are to be processed, they can go either to an ECC for drilling and agent draining or to a PWS for cutting and agent washout. If the warhead and rocket motor are not separated, they must go to the ECC-2. Two options for the separation of the GTR rocket motors and warheads have been proposed. The first is unscrewing the two units, the feasibility of which depends upon the con- dition of the body of the munition and the threads. The second option is water jet cutting between the motor and warhead. Neither option has been demonstrated to date, although no conceptual obstacles are foreseen by the com- mittee. The configuration of the GTRs (warhead and rocket motor or warhead only) to be processed through PBNSF could significantly affect the design basis selected for the worst credible internal explosion event for the building. Manual Materials Handling Unlike chemical disposal facilities in the stockpile dis- posal program, the munitions transport in the PBNSF pro- cess building is not automated and uses a variety of manu- ally operated equipment: • an overpack transfer cart • a munition transfer cart • a munition cradle cart • a parts transfer cart For example, energetically configured munitions travel to and from the ECC/APV on a munitions cradle cart, while a munition transfer cart is used for transport to the heel-dis- solving tanks and to the DET. Following detonation, the munition fragments are transferred to the MDUs on a parts transfer cart. The use of such manually operated equipment is acceptable because only 10 munitions per day are expected to be processed, making the use of conveyors uneconomical. In addition, the time spent moving munitions does not affect the throughput of the facility. Elimination (where practi- cable) or minimization of the number of manual tasks asso- ciated with processing munitions would nonetheless signifi- cantly reduce risk to operations personnel. Processing Sequence The processing sequence for treating munitions in sepa- rate campaigns is as follows: • sulfur mustard • nitrogen mustard • arsenicals Any leaking munitions containing the agent being pro- cessed in a particular campaign will be treated at the end of that campaign. Pine Bluff Non-Stockpile Facility Design Status, Operability, Reliability, and Accessibility by Humans Status of the Engineering Design At the time this report was prepared, several issues criti- cal to the finalization of the PBNSF design had not been resolved. First, there are the issues of whether the 4.2-in. mortar rounds are to be assumed to contain energetic mate- rial (fuze, burster, etc.) and whether the mustard agent that they contain should be considered drainable or not. These issues are important because facility equipment was selected and the schedule was developed based on the assumption that the 4.2-in. mortars were not energetically configured and that the mustard agent would drain easily. The most re- cent assessment is that the original assumptions were not appropriate, so that the existing equipment will not be able to process the 4.2-in. rounds fast enough to meet the intended schedule. This is because (1) the PWS cannot be used as currently configured owing to the anticipated presence of energetics and (2) the ECCs cannot be used as currently con- figured owing to the anticipated inability of the mustard agent to drain. If the most recent assessment of the 4.2-in. mortar rounds is correct, then meeting the schedule will require either sig- nificant modifications to the PWS and/or the ECC or an al- ternative method of processing the munitions. Modification of the PWS and/or the ECC would require significant development and testing, which might affect the schedule. No practicable design to enable the reliable re- moval of gelled/solid agent from a munition in the ECC (with the small access holes currently proposed) was presented to the committee. Chapter 6 addresses possible alternative pro- cessing options. A second processing issue is the current lack of an effec- tive neutralization process for the arsenical fills in approxi- mately 40 percent of the GTRs. Until such a process is de- veloped and any process modifications are designed and implemented, the schedule will be at risk. Another critical issue is that the maximum overpressure the process facility building should be designed to resist has not been finalized. This is a decision that drives the design of the building walls and roof and the design of the heating, ventilation, and air conditioning (HVAC) system. Initially, it was thought that the design basis should be the explosion of the energetic material in a GTR plus the GTR rocket mo- tor, but this is being reconsidered. Eliminating the potential

THE PINE BLUFF NON-STOCKPILE FACILITY 29 explosion of the rocket motor components of a GTR from the PBNSF design basis would reduce the pressure the build- ing and HVAC would have to withstand. The cost reduction that could be achieved by doing this was being examined by the Army as this report was being prepared. If complete GTRs were eliminated from processing at PBNSF, the issue of how this munition might be otherwise processed remains. Several options appear possible: • Deciding that the agent in the GTR is outside the scope of the CWC treaty requirements and addressing its destruction separately and at a later date. (This would also defer the issue of the current lack of a neutraliza- tion technology for approximately 40 percent of the GTR fills.) • Separating the propellant charge from the agent con- tainer. This would allow the agent to be handled in the PWS. • Destroying the GTR in another unit (e.g., the EDS as- sociated with PBMAS). An associated issue is whether the building is to be de- signed to prevent its penetration by the shrapnel from a pos- sible explosion or whether such a penetration (and potential agent release) will be tolerated. The maximum agent release that the various carbon-bed filtration systems must be capable of handling has also not been finalized. The final design of the carbon bed filters can- not be demonstrated until the design bases are further de- fined. The process hazard reviews have not been completed even though the piping and instrument diagrams have been issued as final. Typically, these diagrams are issued as final only after the safety reviews have been completed and all issues resolved. It is known that other issues will affect the current heat and mass balance for the process, and so this document must also be considered incomplete. The reasons these basic design criteria and activities have not been finalized are not clear. The 4.2-in. mortar rounds have been available for inspection for many years, as have the GTRs. The continuing delay in finalizing issues basic to the design of the facility and in performing the safety studies has put pressure on the design staff to meet the requirements of the schedule. On the basis of a brief review of the proposed construc- tion schedule, several general comments can be made. First, the schedule is based on working 6 days a week, 10 hours per day, for the first 6 months. Although the schedule in- cludes a 40-day margin for weather delays, working 6 10- hour days for an extended period will place a severe strain on the workforce and the supervisory staff—productivity might suffer, and there could be adverse impact on safety. It would seem that the published schedule, which is based on achieving the CWC-specified end date for disposal op- erations, is driving the project rather than realistic estimates for the completion of normal engineering tasks. The Army is constrained by CWC treaty and legislative mandates to achieve the destruction of the munitions assigned to PBNSF by April 2007. However, these constraints are at cross-pur- poses with the accepted practice for designing and building a complex industrial facility, the budget constraints imposed by Congress, and the need to develop a technology that sat- isfies regulatory requirements and the public’s desire that the weapons be destroyed in a safe manner. Given the sig- nificance and complexity of these interrelated issues, a dis- cussion of this constraint is warranted. Any large industrial construction project must balance the desired schedule for completing the project, the time neces- sary for obtaining regulatory approvals, the cost of the project, and the need for flexibility to address inevitable un- anticipated implementation issues. All other factors being equal, there is generally a trade-off between the cost of a project and the schedule for implementing the project (i.e., the time needed to develop the specifications or goals of the project, the time to design the facility, and the time to build it) (GAO, 1997). Typically, the shorter the schedule for implementation, the higher the project costs. A complex project with the potential to adversely affect public safety (such as the destruction of non-stockpile materials) simply cannot be implemented more quickly, and for less money, without potentially compromising effectiveness and, ulti- mately, safety. The Army’s PBNSF project has several unusual con- straints that differ from those of a typical industrial project; these constraints may significantly increase costs if the cur- rent design and deadline are maintained. First, because certain citizen groups expressed great con- cerns, Congress required the Army to develop technologies other than incineration for the destruction of the non-stock- pile chemical weapons material (NRC, 2002a). The task of finding a promising technology and turning it into a viable facility design has not been trivial and has consumed much of the time allowed under the CWC. The additional tasks of obtaining regulatory approval and permits and soliciting pub- lic involvement for non-stockpile technologies have also added delay and uncertainty over the ultimate technology to be used (this is discussed more completely in Chapter 5.) In turn, these delays and uncertainties have resulted in increased costs and the need for an unusually long time to develop the conceptual design and goals for PBNSF. Second, the Army’s deadline is imposed by international treaty and U.S. domestic implementing legislation, and so is more inflexible than for a typical industrial project. The de- sire to shorten implementation schedules on most industrial projects is often driven by the economics of the project. The sooner the capital invested in a commercial project begins returning revenue on the investment, the sooner the com- pany will begin to reap a profit. Thus, when a schedule slips there are economic consequences, and economic expecta- tions must be adjusted.

30 ASSESSMENT OF THE ARMY PLAN FOR THE PINE BLUFF NON-STOCKPILE FACILITY The schedule for destroying chemical weapons cannot slip without significant international and legal consequences (among other things, giving other countries an excuse to de- lay destruction of their chemical weapons), and the benefit of destroying this component of the triad of weapons of mass destruction (chemical, biological, and nuclear weapons) is obvious. It would appear that the schedule for the design, construc- tion, systemization and destruction operations for PBNSF has been developed by taking the April 2007 CWC deadline as the completion date and compressing the normal design and implementation steps into the time until then. As a re- sult, the design process has not always followed generally accepted practices. This conclusion is supported by a letter from the Army Corps of Engineers, Little Rock District, not- ing situations that will affect the proposed schedule: • More than 40 days on which weather interferes with construction; • Approval of the Resource Conservation and Recovery Act permit later than May 2004; • The fact that the building design and the process de- sign for the facility are taking place concurrently, pos- sibly necessitating modifications to the contract; • A delay of more than 5 minutes per trip for vehicles accessing the site; • The known adverse effects that rain will have on the workability of the soil; and • The lack of project funding to cover unanticipated delays.13 Given these uncertainties, there is only one certainty— namely, that delays will occur because so many issues re- main unresolved. It would seem that the only option avail- able to the Army is to increase the cost of the project, perhaps precipitously so. Third, the Army is not a business and must operate within the budget appropriated by Congress. The uncertainties in- herent in developing alternative technologies to destroy the chemical agents and resulting neutralents have led to an even greater cost uncertainty. The federal budget allocates funds to a particular project, and sometimes the funds must be ex- pended by a particular date, hampering sound, long-term planning. There may even be pressure to keep the project on a schedule that matches the fund allocation schedule for pur- chasing equipment, site preparation and construction, and operations, whether or not the engineering status of the project warrants such adherence. Fourth, as April 2007 approaches, the more likely it is that the PBNSF project will undergo ad hoc changes to over- come obstacles (e.g., making modifications to the ECC or PWS to deal with gelled mustard in an energetically config- ured munition) rather than moving forward in a well- planned, integrated way. Avoiding an inefficient design is particularly critical when human health or the environment can be compromised by design flaws that result in accidents. Ideally, the owner of an industrial project should consider the environmental requirements applicable to the facility in the design phase (NRC, 2001b). However, in the case of PBNSF, significant regulatory requirements could be im- posed in the future. The committee therefore believes that the published schedule does not reflect the relative immaturity of the engi- neering design and is likely to be too optimistic. It reached this conclusion after examining the project documentation and communications from Army personnel and the Corps of Engineers and comparing the design process for this project with the typical design process for an industrial construction project of a similar magnitude and complexity. Fifth, as the schedule is shortened, there is a need for enhanced coordination and communication between the fa- cility design company, the construction company that will build the facility, and the owner who will use the facility. The approach that has been adopted by the Army is to ini- tiate the construction phase, continue to address outstanding issues as they arise, and refine the design up to the time systemization is started. In theory, it is possible to execute a project in this manner. However, experience has shown that attempting to refine a design, perform development activi- ties, and manage changes during construction almost always results in confusion, delay, continual fixes, and cost over- runs. This is particularly true where a defined end point, such as the CWC date, and a budget allocation schedule are ap- plying pressure to a project. This conclusion is also supported by communications from the Army Corps of Engineers.14 The fact that many of the key design criteria have not been finalized (see above) increases the uncertainty as to whether the Army can attain the April 2007 deadline with the existing PBNSF approach. If the design criteria that are finally agreed upon require modifications to the initial as- sumptions or result in delays, the pressure on the schedule will increase still further. This could result in even less time being available to perform the engineering tasks required to design, construct, and systemize the PBNSF than is shown by the present schedule. The current design of PBNSF relies on equipment located in cramped quarters (namely the ECC units and the CPT). Such space constraints may affect the ability of operating staff to perform required maintenance tasks during normal and upset conditions. The committee believes that a review 13Benjamin H. Butler, Commander, Little Rock District Corps of Engi- neers, memorandum to James Fletcher, Product Manager for Non-Stockpile Chemical Materiel, SFAE-CD-N, September 9, 2003. 14Benjamin H. Butler, Commander, Little Rock District Corps of Engi- neers, memorandum to James Fletcher, Product Manager for Non-Stockpile Chemical Materiel, SFAE-CD-N, September 9, 2003.

THE PINE BLUFF NON-STOCKPILE FACILITY 31 of the facility design by experienced operations personnel from stockpile chemical demilitarization sites would be of value. The U.S. Army Corps of Engineers (Little Rock District) is responsible for the construction of PBNSF. The Army in- tends to offer the contract for systemization and operation of PBNSF to competitive bidding by contractors other than the company performing the design. To permit the bidders to develop a realistic cost and schedule estimate, a significant amount of documentation will have to be prepared. Given the time constraints on the project as currently envisaged, this requirement could retard the construction schedule. Moreover, given all the uncertainties, contractors will find it difficult to provide a cost estimate that accurately reflects the required tasks. Where such uncertainties exist, contrac- tors typically bid low and then rely on change orders to make their profit. In addition, the time required for a new contrac- tor to become familiar with all of the issues on the project will substantially impact the already tight schedule. All of these issues would tend to support retaining the design con- tractor for the construction and operational phases of the project. Some form of cost control would be required, but any other option will probably result in higher costs and an extended schedule. In summary, the committee considers that the following factors are now contributing (or will contribute) to the in- ability of PBNSF (as currently proposed) to achieve the de- struction of the non-stockpile munitions by the CWC date: • Several important design criteria are still undefined. These include the condition of the 4.2-in. mortar shells (gelled agent/explosively configured); the basis for maximum overpressure and the design of the HVAC system; whether the building is to be designed to con- tain shrapnel from an internal explosion; an effective neutralization technology for the arsenical fills in the GTRs; and the maximum agent release that the HVAC system must handle. • Because the process hazard analyses have not been completed, no recommendations have been generated or implemented, even though the piping and instru- ment diagrams have been designated as final. • The schedule is driven by pressure to meet the CWC date rather than being objectively set by the time re- quired for the design and engineering activities. • The budget cycle assumes that the CWC date will be met and allocates monies to the project on that basis, requiring the monies to be expended or lost. There- fore, contracts are awarded and equipment is pur- chased in accordance with the budget schedule and not in accordance with the progress of the design process. • The project has implemented a “design/build” ap- proach. For a project with many basic design criteria still not finalized, this will probably result in additional schedule delays. • A new contractor is expected to be awarded the con- tract to operate the facility. Given the number of exist- ing uncertainties, attempting to hand over the project from one contractor to another will likely result in confusion and schedule delay. Finding 2-1a: The published schedule for the design and implementation of the PBNSF is driven by the April 2007 congressional and CWC deadlines for destroying the chemi- cal weapons and associated materiel. The committee consid- ers that this schedule should not be allowed to drive the start- up of the facility if the engineering design and the operational design are not mature enough. Finding 2-1b: Key design and safety criteria for the PBNSF are still undefined. The key criteria include: • Enabling the handling of gelled and/or energetic 4.2- in. mortars; • Accommodating a neutralization technology that is not yet defined for the arsenical fills in the GTRs; • Implementing the findings of the process hazards analyses; and • Defining the MCE15 for the building and the HVAC system so that it does not put personnel outside the building at risk. Recommendation 2-1: If the current design for the Pine Bluff Non-Stockpile Facility is pursued, a realistic schedule should be developed based on the time required to properly perform the engineering, construction, commissioning, and processing steps. As part of this task, the required basic de- sign criteria must be finalized. In addition, process hazard analyses must be completed and any issues raised by them resolved. Finding 2-2: The Army has attempted to ensure that lessons learned in both the non-stockpile and stockpile disposal pro- grams are shared. However, in certain areas (e.g., the recog- nition of and response to the gelling of mustard agent), this sharing has not been as effective as it could have been. In another critical area—the ability of maintenance workers in PPE to access equipment in the CPT—engineering and op- erational personnel experienced in this type of activity have not been asked for input into the design, nor have they been asked to review it. The purpose of such built-in peer review would be to re- view the facility design to ensure that engineering and op- 15The maximum credible event is defined as the worst single event that could occur at any time, with the maximum release of a chemical agent from a munition, container, or process as a result of an unintended, un- planned, or accidental occurrence (U.S. Army, 1999).

32 ASSESSMENT OF THE ARMY PLAN FOR THE PINE BLUFF NON-STOCKPILE FACILITY erational lessons from the stockpile disposal facilities are transferred appropriately. Recommendation 2-2: The Army should increase its efforts to share relevant experience between the non-stockpile and stockpile disposal programs and, where appropriate, seek outside peer review of designs. This review should include assessment of the chemical processing trailer and the explo- sive containment chamber units to determine which inspec- tion and maintenance activities are feasible for personnel wearing Level A personal protective equipment. Systematic Design Integration Review Issues Several specific issues that surfaced during the review of the design may need to be addressed as the design is final- ized and an integrated facility is constructed. The CPT is the site of a number of systematic design integration issues. Inspection of the CPT showed that extensive use has been made of screwed fittings, flanges, couplings, and compres- sion fittings. In some cases—for example, where air, water, caustic, and the like are flowing through the piping—this may be acceptable. In cases where agent or neutralized agent could be present in the piping/tubing, a careful review of the existing connections in the CPT should be performed. Wher- ever possible, welded connections should be used. At a mini- mum, only connections certified or approved for Fatal Ser- vice should be used in such services, and they should be installed and tested in accordance with the manufacturer’s recommendations. Although operators entering the CPT will be in Level A PPE, proper engineering design attempts to minimize the potential for the release of hazardous materi- als. Historical data from plants handling phosgene and simi- larly hazardous materials demonstrate that leakage usually occurs from small connections and flanges rather than from large flanges. The design of the CPT piping systems that contain agent calls for numerous fittings and screwed couplings, which are more prone to leak than welded joints. Additionally, the agent piping contains dead-legs, which could interfere with its flushing. Also, a cursory inspection of the piping layout showed areas where tripping hazards exist (particularly for personnel in Level A PPE). In one case, the outlet of a relief valve was connected to the relief valve header using a flex- ible hose. It is questionable whether such a design could bear the impact load should the relief valve open. The use of a flexible hose to connect a relief valve to the relief valve header should be carefully considered. Access to the interior of the ECC units requires moving the APV drill assembly inside the vessel before the operator (wearing Level A PPE) can enter the vessel and then with- drawing the APV drill assembly. This is a clumsy procedure and leaves the operator vulnerable if an emergency arises. In addition, the present design for moving the APV drill assem- bly has no backup if the motor fails or the drive system breaks or jams. An Army contractor has recommended al- lowing for the disconnection of the drill assembly from the motor and drive assembly (possibly by a chain link system) and adding a manual pulley for moving the APV assembly in an emergency.16 The CPT has been built but will not undergo systemiza- tion until the second half of 2005. This means that all the equipment, instrumentation, piping, and so forth will be left untested and idle until then. Several generally accepted process design approaches are typically utilized to address the issues outlined above. For example, a process hazard evaluation could be used to iden- tify scenarios for which suitable layers of protection are re- quired, but this has not been performed. Also, semiquanti- tative techniques such as layer-of-protection analysis for a smaller number of scenarios and, possibly, full quantitative studies for an even smaller number of scenarios could be performed. Since there is much human interaction with equipment units in PBNSF, a human factors analysis could also be performed. This systematic design integration review would include all of the piping, instrument connections, and vessels that handle, or might handle, agent or neutralized agent or any other material that would present a significant hazard to op- erating staff. It would also include considerations for rede- signing and specifying piping and connections to minimize the potential for leakage in the CPT or in any other equip- ment at PBNSF. This could involve replacement of piping connections with welded connections wherever possible and the use of flanges with covers, seals, or plates. The systematic design integration review should also ex- amine the elimination of dead-legs and tripping hazards (es- pecially since operators will be working in heavy suits and boots); the feasibility of eliminating flexible hoses; the suit- ability of using a flexible hose to connect a relief valve to the relief header; and whether the design adequately accounts for impact loads (e.g., relief valve opening) and thermal ex- pansion loads. As part of this systematic design integration review, the Army and its contractors should decide what equipment will have to be inspected and tested to ensure that the CPT and other equipment will be operational when required. The Army should also consider whether the CPT system should be functionally tested on a regular basis prior to becoming operational to minimize the potential for failures due to dried-out rubber or polymeric materials (e.g., seals). Addi- tionally, the Army should carefully consider the require- ments for accessing the internals of the ECC units. 16Shaw Engineering, Stone & Webster, Inc., GFE Modification Recom- mendation, sent to the committee on July 31, 2003.

THE PINE BLUFF NON-STOCKPILE FACILITY 33 Finding 2-3: The number of design, implementation, and operational issues that must still be addressed before con- struction of the PBNSF is greater than is typical for an indus- trial facility because of the unique complexity of the techni- cal problems, the need for first-time integration of the systems, and the very short deadline imposed by the CWC. Recommendation 2-3: As soon as possible, the Army should systematically review the design integration and op- eration of all the equipment in the Pine Bluff Non-Stockpile Facility (including piping, connections, and vessels) to find ways for simplifying the processing taking place there. This review should identify ways of (1) minimizing the chances for equipment or operational or human failures, using pre- ventive redesign and related measures to reduce reliance on protective clothing, and (2) optimizing the reliability of the Pine Bluff Non-Stockpile Facility processes. Finding 2-4: In keeping with its assigned task of assessing the concept of operation and engineering design plans for the PBNSF, the committee concludes that the PBNSF is unlikely to meet the goal of destroying the non-stockpile materiel by April 2007 for several reasons, including the following: • The schedule does not recognize the known uncertain- ties in munition configuration, neutralization effective- ness, and system design. • The schedule for construction of the building alone is shorter than generally accepted. • The schedule for installation of the complex piping, other chemical treatment, and instrument and control systems is overly optimistic. • The schedule does not take into account the potential damage to equipment from an unplanned detonation in the ECC. • The schedule does not take into account the potential for delays caused by contamination of the DET when a munition containing solidified mustard agent is de- stroyed. • The schedule assumes optimal operation of systems, some of which have not been designed, e.g., hot water washout of agent in the ECC. Although the committee believes that implementation of the findings and recommendations in this and subsequent chapters might increase the likelihood that PBNSF as cur- rently designed will meet the treaty deadline, in Chapter 6 the committee recommends an alternative approach that will increase safety, reduce long-term costs, and increase the like- lihood that the treaty deadline is met.