The United States has signed and ratified the Chemical Weapons Convention, which outlaws the production and possession of chemical weapons and a number of related chemicals. To date, the United States has destroyed about 90 percent of its stockpile, mostly using incineration.
As part of the U.S. effort to destroy its remaining stockpile of chemical munitions, the Department of Defense is building the Blue Grass Chemical Agent Destruction Pilot Plant (BGCAPP) on the Blue Grass Army Depot (BGAD), near Richmond, Kentucky. The stockpile stored at BGAD consists of rockets and projectiles containing the nerve agents GB and VX and the blister agent mustard. Continued storage poses a risk to the BGAD workforce and the surrounding community because these munitions are several decades old and are developing leaks. The projectiles containing mustard agent will be destroyed using a Static Detonation Chamber being built adjacent to BGCAPP. BGCAPP will destroy the rockets and projectiles containing GB and VX. The variety of munition and agent types, and the degrading agent they contain, poses a variety of challenges to their destruction.
Due to public opposition to the use of incineration to destroy the BGAD stockpile, Congress mandated that non-incineration technologies be identified for use at BGCAPP.1 As a result, BGCAPP will destroy the GB and VX by hydrolysis using hot caustic solution (sodium hydroxide). To comply with the Chemical Weapons Convention requirements for the destruction of chemical weapons,2 the resulting hydrolysates must be further treated. At BGCAPP, this will be accomplished using supercritical water oxidation.
The original BGCAPP design called for munitions to be drained of agent and then for the munition bodies to be washed out using high-pressure hot water. However, during the course of committee discussions related to the systemization of BGCAPP, several concerns emerged that held the potential to compromise safe operations and impede agent processing throughput in the plant. Much of the concern focused on the mixture of agent and wash water that was produced during agent drain and water washout operations. The mixing water and VX has the potential to cause an autocatalytic exothermic reaction that can lead to frothing and overflow in storage tanks upstream of the agent neutralization reactors. The storage tanks are not designed to contain the reaction, in contrast to the neutralization reactors, which are designed to operate at high temperature. Water and VX mixtures can also produce agent gels that could impact agent destruction processes. Mixing water and GB can produce significant hydrofluoric acid, which can be damaging to the steel transfer lines between tanks.
As a result, as part of a larger package of modifications called Engineering Change Proposal 87 (ECP-87), the munition washout step was eliminated. However, implementing this solution will cause larger quantities of agent—more than originally planned—to be partitioned into different BGCAPP processes where agent destruction for these larger quantities is unproven because those processes were designed to treat only small amounts of residual agent. This could have the unintended effect of compromising the ability of the plant to achieve and demonstrate the Kentucky statutory requirement of 99.9999 percent destruction efficiency (DE).
The Program Executive Officer for Assembled Chemical Weapons Alternatives asked that an ad hoc study committee be formed to look into the effects of deleting the water washout step. The statement of task of the Committee on Effects of the Deletion of Chemical Agent Washout on Operations at the Blue Grass Chemical Agent Destruction Pilot Plant was as follows:
1 A similar neutralization plant is also being completed at Pueblo Chemical Depot in Pueblo, Colorado.
2Destruction of chemical weapons means a process by which chemicals are converted in an essentially irreversible way to a form unsuitable for production of chemical weapons and which, in an irreversible manner, renders munitions and other devises unusable as such (Chemical Weapons Convention, Annex on Implementation and Verification, Part IV (A), Destruction of Chemical Weapons and Its Verification Pursuant to Article IV).
- Assess the impact of the design change on plant operations and the impacts to plant throughput, taking into account revised rocket and projectile drain times, strainer change-out frequency, and metal parts treater throughput;
- Review and assess the calculations associated with the ability of the metal parts treater and thermal oxidizer to effectively process additional residual agent GB and VX contained in the drained rocket and projectile munition bodies;
- Review and assess the contractor’s approach to the destruction efficiency (DE) calculations and provide any suggestions that support the DE confirmation process; and,
- Assess the validity of process modeling conducted to date and recommend where additional modeling may be of benefit for understanding likely plant operation performance.
BGCAPP is legally required to achieve what is termed “six-nines” destruction of GB and VX, which means that it must be demonstrated that the fraction of agent destroyed be greater than 0.999999, or alternatively, the fraction remaining must be less than 1 × 10−6 of what was originally present in the munitions. In the original plant design, it was intended that almost all of the agent would be processed by caustic hydrolysis through the agent neutralization system (ANS). Of course, this was never strictly the case because some vapors from the agent draining operations will be vented into the munitions demilitarization building (MDB) heating, ventilation, and air conditioning (HVAC) filtration system, and there was always likely to be traces of agent remaining on the projectiles and rocket warheads after the agent washout. But these quantities were deemed to be negligible, and thus DE could be demonstrated by merely measuring agent in the ANS hydrolysate to a concentration equivalent to or less than 1 × 10−6 times that of the concentration of agent fed into the unit.
The deletion of the water washout step will now result in significant quantities of agent being partitioned into other process streams of BGCAPP. The rocket warhead pieces (after the warheads are drained and sheared) will contain more residual agent than originally planned when they are sent to the energetics batch hydrolyzer (EBH) units. These units, designed to hydrolyze the energetic materials in the rocket warhead bursters, also contain caustic (sodium hydroxide) at elevated temperature. Calculations by BGCAPP predict that there will be sufficient excess caustic present to ensure complete destruction of any agent that is partitioned to the EBHs, and this expectation is supported by prior experience with caustic hydrolysis of GB and VX. GB, on account of its solubility, will very likely be completely eliminated from the caustic solution. However, VX may survive as a result of incomplete mixing or as a result of being sequestered in cracks in the metal parts. Destruction efficiency of agent in the EBHs is not explicitly known and will need to be demonstrated in order to provide a defensible calculation of DE. In addition, there is also a chance that a fraction of the agent, particularly GB, which has a lower boiling point, will be partitioned into the off-gas stream from the EBHs into the EBH off-gas treatment system (OTE). Significantly, the OTE is not designed to destroy agent, which means that any agent that is volatilized in the EBHs will instead be captured on the carbon filters of the MDB HVAC. While extensive experience in the broader chemical demilitarization program indicates that this outcome would be protective of the public and the environment, BGCAPP believes that it is not likely to be allowed to take credit for the removal of agent vapor by the carbon bank adsorption prior to release of exhaust into the atmosphere when DE is calculated per the Kentucky Revised Statutes or the BGCAPP operating permit.3
The drained projectile bodies will also contain more residual agent than originally planned. Because these items do not have energetics components, they will be processed through the metal parts treater (MPT). The MPT thermally decontaminates agent-contaminated items by ensuring that they are exposed to 1,000°F for at least 15 minutes. Extensive operational experience and calculations by BGCAPP indicate that processing time and temperature of the MPT should be sufficient to destroy any extra agent partitioned into that unit, but, as in the case of the EBHs, this will not be known until the unit is actually operated. The gaseous effluent from the MPT flows into the off-gas treatment system (OTM), which is equipped with a thermal oxidizer (TOX), and the committee believes that the TOX will destroy any fugitive agent vapors escaping from the MPT.
Demonstrating six-nines destruction after deletion of the washout step will be significantly more difficult than originally planned due to the change in the amount of agent now partitioned outside of the ANS. BGCAPP personnel have considered two alternative methodologies to determine DE, but these entail much more measurement, and, in many cases, good analytical methods do not currently exist. With some streams, like the caustic in the EBHs, it may be difficult to measure agent concentration down to a level that would demonstrate achievement of the DE criteria.
An attractive solution would be to count the agent trapped in the MDB HVAC carbon beds as destroyed. However, the Kentucky Department for Environmental Protection has stated that BGCAPP may not take credit for measurements downstream of the carbon filtration system without revision to the statute/guidelines; thus, agent trapped in the MDB HVAC carbon beds cannot be counted as destruction in the calculation of DE, even though the design and build of this system is consistent with the capture/removal devices approved for use in other (incineration) demilitarization
3 John McArthur, environmental manager, BPBG, “Destruction Efficiency Considerations,” presentation to the committee on September 9, 2015.
facilities where removal was considered in the calculation of “destruction and/or removal efficiency” at those sites.4
These overarching assessments are summarized in a series of findings and recommendations, which provide a summary of all of the committee’s work. Chapter 2 assesses the process impacts of washout deletion. Specific findings and recommendations are focused on the effects of agent degradation over decades in storage on the physical state of the agent and, hence, agent drain times. A second issue is that the change-out of the filter socks used to capture agent solids may be time consuming, even though new socks with greater capacity have been introduced, and, with washout deletion, the change-outs will demand additional time, which has not been included in the process modeling. Additional attention was focused on the potential for the impact of increased agent loading to the OTM, although there are approaches for ensuring a high DE. Similarly, the EBHs and units situated serially downstream will be subjected to additional agent loading. The effect of the washout deletion on agent loading to the MDB HVAC system is uncertain at this time. It is likely that there will be additional agent loading from the EBH-OTE process stream. In addition, there may be a change in the amount of agent directed to the MDB HVAC from the munitions-draining operations.
Finding 2-1. Uncertainty in the number of munitions containing degraded agent and the degree of agent degradation is compounded by a lack of knowledge of the physico-chemical characteristics of degraded agent as they relate to drain times and amounts of residual agent retained in munitions at the end of the drain process. Better data are needed to properly estimate the time that will be required to process the nerve agent munitions through BGCAPP.
Recommendation 2-1. BGCAPP should gather data, such as mass drained, drain time, and any available information on physical state, for each individual munition during operations ramp up to assess the state of the agent fills and thus expected variability in drain times for each agent lot and type of munition. The acquisition of these data should continue throughout operations to continuously improve the quality of estimates as an aid toward planning of plant operations and to estimate completion times.
Finding 2-2. Even with the change in filter sock capacity, the change-out frequency could become the rate-determining step in the processing of rockets and projectiles.
Finding 2-3. Agent processed through the MPT and the off-gas treatment system will constitute a significant fraction of the agent destroyed at BGCAPP. This is a departure from the original design where almost the entire agent volume was being treated by hydrolysis.
Finding 2-4. Multiple mechanisms exist for controlling the MPT throughput rate to reduce instantaneous agent loading in the MPT and the off-gas treatment system. These include, but are not limited to, approaches such as increasing the residence time in zone 1 of the MPT, reducing the number of projectiles on each tray being processed and increasing the steam addition rate to the MPT.
Recommendation 2-2. BGCAPP should evaluate whether higher agent vaporization rates in the metal parts treater (MPT) can be accommodated by optimizing the operating parameters of the MPT, the off-gas treatment system, and associated systems.
Finding 2-5. With the deletion of munitions washout, some of the chemical agent from the rocket warheads will be sent to the EBHs. Some fraction of the agent introduced into the EBHs will be volatilized and then flow into the EBH OTE. The OTE does not have a TOX, so some of the agent transported from the EBH to the OTE may penetrate to the MDB HVAC.
Recommendation 2-3. BGCAPP should conduct modeling and experimental studies to bound the quantity of agent present in the OTE vent stream (stream #8517).
Finding 2-6. During punch and drain operations, vapors are released directly to the room air and are exhausted through the MDB HVAC system. The primary mode of capture of these vapors is the carbon filter bank. This function is part of the original plant process; however, the washout deletion may affect agent concentrations in the gas phase that will be transferred to the MDB HVAC system.
Recommendation 2-4. BGCAPP should complete modeling to estimate the agent load to the carbon beds in the absence of a munition washout step to ensure that the lifetime of these beds is known.
As noted above, the washout deletion will have a pronounced effect on the calculation of DE that is mandated by Kentucky Revised Statutes; this is the subject of Chapter 3. Although some findings and recommendations in Chapter 3 overlap with those offered in Chapter 2, the focus in Chapter 3 is on the potential for washout deletion to complicate the calculation of DE and supporting measurements. Specifically, because an increased fraction of agent will now be partitioned into the EBHs, a possible pathway for agent would be any agent residual on the metal parts following processing by the EBHs, which would then be transferred into the MPT. Given the harsh treatment conditions in the EBH and MPT, it is not likely that agent would survive these
4 NRC Washout Deletion Committee Questions and Responses 151029, received via e-mail on December 8, 2015.
units; however, this has not been demonstrated. Additionally, the EBHs generate energetics hydrolysate. This hydrolysate is then sent to the energetics neutralization system (ENS), which operates at a higher temperature and pressure than the EBHs. BGCAPP calculations indicate that any residual agent in the hydrolysate should be destroyed by the ENS, but, again, this has not been demonstrated for the additional loading that could result from washout deletion. The off-gas from the EBHs may also contain some agent. This off-gas will be treated by the OTE, which is not designed to destroy agent, and its capacity to do so is unknown.
After washout deletion, the facility OTM will need to accommodate more agent than was originally intended. It is probably capable of doing this, but BGCAPP will need to demonstrate this in order to provide assurance of DE. The particles formed in the TOX unit of the OTM will very likely be free of agent, as a consequence of the high temperature and residence time of the agent in the TOX. However, BGCAPP will need to demonstrate DE for this solid waste stream from the TOX.
To account for the possibility of agent in these non-ANS effluent streams, BGCAPP has considered two possible alternative approaches to the original approach (called Approach 1) for demonstrating DRE. The first alternative, Approach 2, would measure the difference in agent quantities in the feed and effluent streams from specific individual units. The second alternative, Approach 3, would assume a bulk quantity in the munitions input lines, while measuring agent quantities in the effluent streams from all treatment units. As stated in the findings below, Approach 2 is not viable for demonstrating DE criteria because it does not include all possible agent-contaminated streams. Approach 3 is more achievable but would present significant challenges in developing measurement methodologies for the different forms of effluent.
One recommendation related to the physical plant did emerge from the committee’s extended DE discussion. It was noted that because the OTE was not designed to destroy organics (such as agent), a possible modification might be to send the off-gas from the OTE into the OTM, which is designed to destroy organics and is expected to be able to destroy agent.
Finding 3-1. With washout deletion, the current methodology for calculating whether BGCAPP meets the statutory and regulatory requirement for a DE of 99.9999 percent will no longer be applicable. This is because the drained and washed out agent will no longer go only to the ANS. Instead, a greater amount of residual agent remaining in the rocket warhead pieces will now be processed in the EBHs and the ENS, and residual agent in the projectiles and possibly on rocket warhead pieces will now be processed through the MPT.
Finding 3-2. The partitioning of agent across additional process streams will introduce monitoring challenges that BGCAPP needs to be aware of. It may be difficult to identify monitoring technologies or strategies of sufficient sensitivity to measure what are expected to be low concentrations of agent in some streams. Additionally, the identification of new technologies or strategies carries the risk of a negative impact on the BGCAPP schedule.
Finding 3-3. It is likely that the combination of the EBH and MPT treatment conditions will be sufficient to destroy all residual agent residing on rocket warhead pieces. This, however, needs to be shown through either calculations or demonstration.
Recommendation 3-1. BGCAPP should calculate or otherwise demonstrate a 99.9999 percent (“six-nines”) destruction efficiency (DE) for residual agent residing on rocket warhead pieces exiting the metals parts treatment unit (MPT). This would provide assurance that the solid effluent from the MPT (stream #7652) generated during rocket campaigns is free of agent to ensure compliance with DE requirements.
Finding 3-4. Complete destruction of augmented agent loadings passing through the EBH/ENS system has not been demonstrated.
Recommendation 3-2. BGCAPP should demonstrate satisfactory destruction efficiency for agent serially treated with caustic under the same conditions as those present in the energetics batch hydrolyzers (EBHs) and the energetics neutralization system at agent loadings equivalent to the highest quantities anticipated to be treated by the EBHs without washout.
Finding 3-5. An unknown fraction of agent entering the EBHs during the rocket campaigns may undergo volatilization instead of hydrolysis. Volatilized agent will be processed through the OTE system (stream #8517), which is not designed to destroy agent. Agent escaping the OTE will be removed to the MDB HVAC carbon filter banks, together with fugitive agent emissions from the munition drain processes. Because agent partitioned into these pathways cannot be counted as destroyed, and because BGCAPP believes it is not likely to be allowed to take credit for removal of agent vapor by carbon bank adsorption in the MDB HVAC prior to release of exhaust to the atmosphere, the implementation of washout deletion will require significant permit modifications and has the potential to prevent BGCAPP from achieving DE criteria.
Recommendation 3-3. For all of the gaseous process streams, BGCAPP should rigorously demonstrate that negligible agent is partitioned into the munitions demilitarization building (MDB) heating, ventilation, and air conditioning (HVAC) carbon filter banks under all conditions that could arise during the rocket campaign. BGCAPP should provide
for monitoring of the OTE effluent stream (#8517) with analytical sensitivity sufficient to ensure that destruction efficiency criteria are achieved before they enter the MDB HVAC system.
Recommendation 3-4. BGCAPP should examine the possibility of routing the gaseous effluent from the OTE (energetics batch hydrolyzer off-gas treatment system) into the OTM (off-gas treatment system). This would eliminate the biggest uncertainties in MOut exiting the munitions demilitarization building (MDB), because it is likely that any agent surviving the OTE would be destroyed in the OTM. The number of gaseous streams from processing units exiting the MDB would be reduced to a single stream—namely, the off-gas from the OTM—and would be less likely to contain significant agent as a result of off-gas passing through the thermal oxidizer.
Finding 3-6. It is likely that the combination of the MPT and the OTM will completely destroy any agent entering the MPT. However, after washout deletion, the OTM will receive gaseous streams from other sources that may contain more agent than originally planned. It is currently unknown whether the OTM can adequately treat the combined load of all streams after washout deletion.
Recommendation 3-5. BGCAPP should measure solid, gaseous, and liquid effluents from the OTM (off-gas treatment system) during initial projectile campaigns to ensure that these effluents meet the destruction efficiency criteria.
Finding 3-7. The solid waste stream from the OTM should be agent-free. This conclusion will need to be demonstrated to the Kentucky Department for Environmental Protection based on validated process controls and statistical testing.
Finding 3-8. Approach 2 is not an appropriate option for the calculation of DE. It is incomplete because it does not include the gaseous emissions from the OTE—which, under the new configuration, may contain agent—and because it is not operationally practical to measure agent quantities in the feed to, and effluent from, the individual process units.
Finding 3-9. Approach 3 could conceivably be used for a defendable DE determination, because it accounts for the OTE gaseous process stream #8517, provided it is modified to include the fugitive releases of agent vapor directed to the MDB HVAC system. However, Approach 3 would require development of additional methodologies for measuring masses of agent partitioned into the two gaseous waste streams entering the MDB HVAC system.
Recommendation 3-6. If Approach 3 is adopted, then BGCAPP should evaluate the concentrations of agent liable to be present in all gaseous process streams and develop measurement approaches with sufficient sensitivity to ensure that destruction efficiency criteria are being achieved.
Finding 3-10. The performance requirements for the analytical measurement methodology for measuring agent in the off-gas process stream from the OTE (#8517) are not known, because the fraction of agent that will be partitioned into this stream is uncertain.
Recommendation 3-7. If Recommendation 3-4 is not pursued, BGCAPP should conduct research to determine what fraction of GB agent might partition into the off-gas process stream from the OTE (energetics batch hydrolyzer off-gas treatment system) and then use this information to set analytical performance requirements that can be used to identify analytical measurement methodology.
Chapter 4 addresses the modeling of munitions throughput at BGCAPP, which is done using an overall process model (rather than individual process models). Input parameters were based on point observations made by BGCAPP staff, which may well be accurate in the mean. However, the committee does not believe that the expected variability in plant operational parameters is reflected in the model.
An area of repeated concern for BGCAPP and for the committee was the accuracy of estimates of the time required to drain the munitions. Inaccurate drain-time estimates have the potential to result in inaccurate model throughput estimates. The committee believes that the best approach for estimating drain times would be to capture information on munition drain operations from individuals who have actually conducted these activities at other demilitarization sites. There may also be opportunities to collect actual operating data during systemization and as plant operations begin at BGCAPP. Another process concern is that of filter sock change-out, which is related to the issue of draining in that both processes are affected by the extent of solids and gels in the munitions. In general, BGCAPP staff have attempted to be conservative in all of their parameter estimates, but it is also clear that many of the parameters potentially have large variances (e.g., the amount of agent fill that cannot be drained). With regard to bounding the output of the model, it would be useful to run the model with input parameters equivalent to the highest and lowest levels that could be encountered. Finally, statistical quality control might have significant utility for managing operations.
Finding 4-1. While the process model explores the influence of variations in operating parameters on the performance of BGCAPP, the limited treatment of the stochastic nature of those parameters does not reflect operational experience.
Finding 4-2. The reliance on point estimates in the model data does raise concerns about the ability of the model to accurately forecast future facility operations in terms of the
length of time to complete the processing of the chemical weapons and the risks involved in operating the facility.
Finding 4-3. The stochastic nature of the gelling or crystallization of the GB agent may still be partially retrievable. A formal debriefing of individuals who have drained munitions to capture the (informal and clearly anecdotal) nature of the condition of the agent in the weapons might be useful in developing more believable assumptions as to the condition and variability of the chemical agents in the weapons.
Recommendation 4-1. BGCAPP should retrieve and document historical (informal and anecdotal) data on munition drain times and run these data, complete with ranges of uncertainty, through the BGCAPP model.
Finding 4-4. The actual filter sock change-out rate may be the most important rate-limiting factor in BGCAPP operations and may be underestimated.
Finding 4-5. Analysis of the sensitivity of the BGCAPP operations to variations in model input parameters might expose potential operational issues, allowing them to be quantified and possibly mitigated prior to operations.
Recommendation 4-2. BGCAPP should design and execute a series of modeling experiments to determine the sensitivity of operations to variations in operating parameters, reflecting the stochastic nature of some processes. Examples of parameters include maintenance and repair times, added characterization steps, retreatment for batches not meeting destruction efficiency, and compounding problems such as long munitions drain times together with very frequent filter sock change-outs. The results of these experiments should be used to prepare for potential challenges and mitigate them ahead of time as much as possible.
Finding 4-6. Point estimates of operational parameters are only a starting point. To fully understand the plant operation and, ultimately, to understand the plant timeline, one needs data on the distribution of parameter values that may be encountered during operation.
Recommendation 4-3. During start-up, and continuing through plant operations, BGCAPP should gather data for relevant model parameters with sufficient resolution to assess the probability density functions for these parameters.
Finding 4-7. Statistical quality control could be a useful management tool for understanding and identifying possible problems as they occur.
Recommendation 4-4. BGCAPP should give attention to developing analysis tools such as statistical quality control prior to actual facility start-up.