Click for next page ( 16


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



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

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

OCR for page 15
Processing of M55 Rockets at JACADS and TOCDF PROCESS DESIGN FOR JACADS AND TOCDF The baseline incineration system was first operated at the Johnston Atoll Chemical Agent Disposal System (JACADS) on Johnston Island in 1990. A series of four operational verification testing (OVTs) campaigns was conducted from 1990 to 1993 using various agents in munitions and containers to make certain the baseline incineration system was safe and effective. After the OVT program was completed, MITRE Corporation, an Army contractor, and the National Research Council (NRC) concluded that the system could operate safely and effectively (MITRE, 1991; NRC, 1994b). The baseline incineration system at JACADS was conse- quently authorized to complete the destruction of chemical agent and munitions stockpiles at Johnston Island. Subsequently, a second-generation facility at Tooele, Utah, began agent disposal operations in 1996, following a period of systemization (preoperational testing). More than 7,500 GB M55 rockets on Johnston Is- land were processed during the first OVT campaign, OVT 1. All of the M55 rockets on Johnston Island con- taining VX were processed in the second campaign, OVT 2. The first disposal campaign at TOCDF was also directed at the destruction of the entire GB-filled rocket stockpile at Deseret Chemical Depot in Utah. As noted in previous chapters, some of these rockets contained gelled GB and required special processing. This chapter describes how rockets are processed in the baseline incineration system. It also reviews the results of processing rockets and lessons learned dur- 15 ing OVT 1 and OVT 2 at JACADS and discusses the TOCDF operations with both gelled and ungelled GB rockets. loading, Transport, and Unpacking The delivery of rockets from the storage areas to the disposal facility is the first step in the disposal process. Pallets, each containing 15 rockets in individual ship- ping tubes secured by steel bands, are removed from storage igloos by forklifts and loaded into a transport container that is delivered by truck or tractor trailer to the disposal facility. At JACADS, each pallet was loaded into a sealed, metal vacuum box for transport (two at a time) on a flatbed truck to the facility (MITRE, 1991~. At TOCDF, where the transport dis- tance was much longer (almost 2 miles), a larger cylin- drical vacuum container (8.5 ft diameter by 11 ft long), termed an on-site container (ONC), was developed and used for transport of multiple pallets towed by a tractor-trailer (U.S. Army, 1996b). The transport containers are unloaded at the muni- tions demilitarization building (MDB) dock. The at- mosphere of each container, maintained at subatmo- spheric pressure to prevent leakage to the environs, is checked for the presence of agent leaking from the rockets. Those containers in which no leaking rockets are detected are elevated to the unpack area on the sec- ond floor of the MDB. The pallets are removed from the container and the rockets are manually loaded into the rocket handling system (RHS), whose main com- ponent is the rocket shear machine (RSM). Empty con-

OCR for page 15
16 tainers are returned to a second dock of the MDB for return to the storage area. A limit is placed on the num- ber of containers in the unpack area. Transport contain- ers stored there are periodically checked to ascertain that no agent has leaked into them from the rockets or their shipping tubes. Containers in which leaking agent is detected are directed to the explosion containment vestibule of the MDB for special handling by personnel in demilitari- zation protective ensemble (DPE) suits. Leaking rock- ets that have been overpacked are delivered to this same area for special handling and feeding into the RSM. At JACADS and TOCDF, no safety or environmental problems and no rate limitations in processing were attributable to GB M55 rocket loading, transport, or unpacking systems and operations. Rocket Handling System The RHSs that are installed at JACADS, TOCDF, and the other baseline facilities are virtually identical and are as shown in Figure 3-1. As noted in the figure, the first part of the RHS comprises the following: The rocket metering table. - The conveyor system that carries the rocket from the metering table through gates into and out of Explosive Containment Vestibule Explosion Containment Room 1/~ Rocket Shear Pallet of IS M55 Rockets in Firing Tubes Rotary Kiln Heated Discharge Conveyor FIGURE 3-1 Rocket handling system. Source: SAIC (2002b). ASSESSMENT OF PROCESSING GELLED GB M55 ROCKETS AT ANNISTON the explosion containment vestibule, into the ex- plosion containment room (ECR), to the RSM. - The RSM punch-and-drain station for removing agent from the rocket, where agent is drained from the rocket into the agent quantification sys- tem a pump, filter, measuring station, and stor- age vessel that allows process operators to deter- mine how much of the original 10.7 lb of GB has been drained. Agent from the storage vessel is subsequently metered to the liquid incinerator (LIC) for disposal processing. The RSM shear station in which a single blade, cooled and cleaned by a flow of decontamination solution, sequentially shears the rocket into eight segments: the fuze section; the agent section and its burster into three segments; the rocket propel- lant into three segments; and the rocket nozzle and tail fin section. Figure 3-2 shows the location of the cuts made in the RSM to shear the rocket into eight segments. The fig- ure also presents information on the process in which the segments are dropped from the RSM through an angled chute into the deactivation furnace system (DFS). This process occurs in a sequence of three dumps. The volume in the chute between the two gates is water-spray cooled to minimize premature vaporiza- To . . Pollution Liquid it /1 Abatement Inc ner ~ ~ System Chamber AL ~' ~ Abatement Allerburner /scrap Bin Blast , / Attenuation Duct

OCR for page 15
PROCESSING OF M55 ROCKETS AT JACADS AND TOCDF ~ ~ o cn Z o o o I I o #2 DUMP #3 DUMP CURRENT ROCKET SHEAR SEQUENCE 9) PUSH ROCKET & 4th CHOP 10) PUSH ROCKET & 5th CHOP 11) 2nd DUMP (4 PIECES) 12) PUSH ROCKET & 6th CHOP 13) PUSH ROCKET & 7th CHOP 14) 3rd DUMP (2 PIECES) 15) PUSH TAILPIECE 16) REPEAT STARTING AT (1) 1) ADVANCE ROCKET TO RSM FROM RDS 2) RAISE ROCKET ENTRY BLAST GATE 3) ADVANCE NEXT ROCKET TO RDS 4) LOWER ROCKET ENTRY BLAST GATE 5) 1st CHOP (FUSE) 6) 1st DUMP (FUSE & TAILPIECE IF PRESENT) 7) PUSH ROCKET AND 2nd CHOP 8) PUSH ROCKET & 3rd CHOP 17 CN Cal ~ UP o g o g In o I 4.69 8.80 , , INFUSE ~M36 BURSTER 4.69 #1 DUMP 11.88 / ~ [AGENT ~ IGNITER / M34 BURSTER BURSTER WELL 39.67 ~7 to o IIJ cn IL In co o I Cut _ _ _ CO o 4t 8.95 cat In rid I C) o ~ ~ Cal CN C~ o o o o o o Cal Cal Cal Cat Cal ~ cn on co cn On u' Cal I red ~ '_ ,_ us I I I ~ O I 11.88 11.88 9.57 16.20 ~ ~ _ _ _ _ _ _ _ 10.04 11.88 11.88 9.57 _ _ _ _ _ _ _ 17~ ~ ~ :~ 1 U~ ~ -- ~ ,] ma- at-. ~.~ - ~[~ ~ -, _ 1 =_= ~ . _ PROPELLANT . ~ ELFIN ASSEMBLY MOTOR 2145 DUMP WITH FUSE FROM NEXT ROCKET 5 _ ~ FIGURE 3-2 Chopping sequence for 115-mm M55 rocket. Source: General Physics Corporation (20001. Note: The numbers above the doubleheaded arrows are the length of the rocket sections in inches. tion of the agent and ignition of the energetic fuze, burster, and propellant components of the rocket. In the first dump, the fuze is admitted into the DFS through the gates of the chute along with the tail fin sec- tion of the previously processed rocket. In the second dump, the four sheared rocket segments comprising the burster, the agent cavity, and a portion of the propellant are dropped into the DFS. In the third dump, the two re- maining sheared rocket segments containing the rest of the propellant are dropped. Separating the energetics into three separate dumps or drops avoids detonating the burster and propellant by the fuze and avoids the simulta- neous heat release and pressure rise that would result from the combustion of the burster with all of the propellant. Not considered in the testing described above (and the modeling discussed in Chapter 4) is the possibility of dumping a rocket's parts into the DFS in as many as seven separate dumps. Reducing the size and Btu content of the feed packages could result in more uniform combustion within the kiln. Doing so might produce more uniform control and reduce or eliminate automatic waste feed cut- offs, but this would have to be demonstrated during the agent trial burn (ATE), when the effect of more cycling of the chute gates would also be evaluated. Two identical, independent, parallel RSMs are in- stalled in each of the facilities, although for simplifica- tion only one is shown in Figure 3-1. They discharge through separate chutes into a common DFS. For the most part, rocket processing at JACADS and TOCDF made use of only one of the installed RSMs at any given time (EG&G, 2002a). The operator can set the RSMs to operate between 10 and 50 rockets per hour. The original RSM design was based on an average process- ing rate of 32 rockets per hour, with a peak capacity of 60 per hour. Punch-and-drain time was intended to be 50 to 75 s under normal conditions. Agent Disposal, Decontamination of Metal Paris, and Destruction of Energetics and Shipping Tubes The original designs for the JACADS and TODCF baseline incineration system facilities were based on the

OCR for page 15
18 assumption that 95 percent or more of the GB agent in the M55 rockets would be drained during RSM opera- tions, stored, and subsequently processed in the LIC. The agent heel of 5 percent or less would be destroyed in the DFS along with the energetics, metal parts, and shipping tube fragments produced in the RSM operation. The LIC has two combustion chambers. In the pri- mary chamber, liquid agent is atomized with air and burned at 2700F. The secondary chamber is provided with a separate burner system set for a chamber tem- perature of 2000F to ensure complete combustion and agent destruction (U.S. Army, 1999a). The principal component of the DFS is a rotating kiln about 33 ft long and 5 ft in diameter. Internal flights ensure the movement of the metal parts and ash residues through the kiln. At a planned kiln rotation of ~1.85 rpm, the solids residence time is 6.5 min. The rocket and shipping tube segments produced in the RSM operation are dumped into a feed chute and slide through two blast gates into the DFS. The interlocked gates prevent the injection of rocket segments until the materials currently being processed in the DFS have moved out of the way. The gates are interlocked so that one is always closed when the other is open. This ar- rangement minimizes the possibility of a backflow of gases into the ECRs containing the conveyor and the RSM equipment as a result of overpressurization in the DFS. The DFS is designed to operate at between 1000 and 1500F and processed up to 38 drained rockets per hour at JACADS (U.S. Army, 1993~. Water sprays in the gas exhaust piping and the feed chute prevent ex- cessive temperatures. The DFS kiln has an outer shroud through which the combustion air is drawn to lower the temperature of the kiln shell. Noncombustible solids pass out of the DFS kiln onto a heated discharge con- veyer (HDC) that is designed to complete the decon- tamination of the solids to a 5X level.2 This conveyer lifts the residues to another chute, from which they drop through gates into a residue collection bin. Drums of cooled residue decontaminated to a 5X condition are shipped off-site. iThe term "flight" refers to helical plates attached to the kiln shell to convey the feed materials horizontally through the rotating kiln. 2Solids are treated to a SX decontamination level by holding the material at 1000F for 15 min. This treatment results in completely decontaminated materials that can be released for general use or sold to the public in accordance with applicable federal, state, and local regulations. ASSESSMENT OF PROCESSING GELLED GB M55 ROCKETS AT ANNISTON Exhaust gas from the DFS kiln goes first to a blast attenuation duct, then to a cyclone in which larger particu- lates are separated from the gas stream, and then to an afterburner operating at 2200F for a residence time of at least 2 s. The afterburner and pollution abatement system ensure that agent destruction meets the required 99.9999 percent destruction and removal efficiency (DRE). The combustion flue gases from both the LIC and DFS go to identical, parallel pollution abatement sys- tems (PAS ), in which the flue gas is first quenched with process water via a spray system that reduces the gas temperature. The gas then passes through a venturi scrubber, where 18 percent caustic solution is injected, and combines with the acid components of the gas, forming sodium salts. The salts are removed as brine in a downstream water scrubber tower and either stored for subsequent processing on-site in a brine reduction area (BRA) or shipped off-site for processing. The BRA is a set of evaporators to crystallize the brine. Although both JACADS and TOCDF had BRAs, the off-site shipping approach proved cheaper and was used at both sites. At JACADS and TOCDF, after passing through the water scrubber tower of the PAS, the flue gas went to the stack, where it was discharged to the atmosphere. In the newer designs for other baseline incineration systems employed at Anniston, Umatilla, and Pine Bluff, the flue gas, after passing through the PAS scrubber tower, goes to a series of high-efficiency par- ticulate (HEPA) and carbon filters, known as the PAS filter system (PFS), before going to the stack. The PFS acts as an additional safeguard by removing any remaining traces of agent and products of incomplete combustion, giving the surrounding community addi- tional assurance that harmful emissions have been suitably controlled in a manner that protects public health. GB M55 ROCKET DISPOSAL: ACTUAL VERSUS DESIGN RATE JACADS Rocke! Disposal Operations During OVT ~ In 1990 and 1991, the entire Johnston Island stock- pile of 7,490 GB M55 rockets was processed at the newly commissioned JACADS facility over a 7-month period in its very first operation, OVT 1 (MITRE, 1991~. The destruction of GB M55 rockets at JACADS took longer than originally planned (MITRE, 1991; NRC, 1994b). The RSM performed fairly well, with

OCR for page 15
PROCESSING OF M55 ROCKETS AT JACADS AND TOCDF about 94 percent availability. A peak rate of 32 rockets per hour was demonstrated, although the average rate of 7 per hour over the campaign was well below the design rate of 32 per hour for the baseline incineration system. The lower rate was primarily due to problems with the DFS/HDC (MITRE, 1991~. The "best shift" goal is the process designer's average intended (design) throughput rate. "Full rate" goals and results are com- puted as about two-thirds of the design throughput rate. At JACADS, the single best shift rate achieved for GB M55 rockets was 27 rockets per hour, achieved over a 4-hour period in OVT 1 (NRC, 1994b). The full rate goal for extended periods of operation was 24 rockets per hour. Actual results over an extended period came to 15.3 drained rockets per hour, or less than half of the designer's intended rate. The disposal rate shortfall at JACADS was attrib- uted in general to problems associated with the start- up and shakedown of a complex, new industrial pro- totype facility whose associated processes had never before been operated together as a system. The goals may have been set too high (MITRE, 1991~. The lim- ited number of rockets in the GB campaign at JACADS allowed too little time to correct initial pro- cess problems and to achieve an improved, steady- state production rate. Numerous short-duration, un- documented interruptions and downtime significantly degraded the processing rate. System component failures are to be expected dur- ing any start-up operation. Lessons learned from JACADS were used to improve the performance at baseline facilities (MITRE, 1992; NRC, 1994b). TOCDF Rocket Disposal Operations Although TOCDF benefited from lessons learned at JACADS and its throughput of rockets containing liq- uid GB was marginally better than that of JACADS, it still fell short of the design rate. The GB rocket cam- paign at the TOCDF processed 28,945 M55 rockets (EG&G, 2002a) and 1,057 M56 (EG&G, 2002b) war- heads from August 22, 1996, through March 24, 1997, and from October 26, 1998, through August 14, 2001. Processing at TOCDF was subject to interruptions from the gelled (thickened and crystallized) agent that was encountered in approximately one-sixth of the GB- filled rockets processed. The gelled agent clogged the agent handling system. Removal of gelled agent at the punch-and-drain station was slowed, the removal of 95 percent called for by the design and specified in the 19 original TOCDF (and JACADS) Resource Conserva- tion and Recovery Act (RCRA) operating permits- was never achieved, and attempts at agent removal were continually frustrated. Also interrupting processing were occasional DFS feed chute jams, which required removal by personnel in DPE suits, reducing the availability of the DFS. Thermal stressing of the DFS kiln led to cracks that were observed during maintenance and then repaired (Vaughn, 2002~. The HDC was another source of system downtime. Conveyor link deformation associated with high-tem- perature operation allowed extra slack in the system, causing rollers to disengage from the track. Molten alu- minum from the rocket bodies exiting the DFS spilled and caused additional jams. Solid debris sometimes failed to dump as intended, choking the conveyor. Processing was also slowed somewhat by the need to handle the 419 overpacked (leaking) rockets stored at TOCDF (EG&G, 2002a). The RSMs enjoyed a very high availability rate, bet- ter than 95 percent (EG&G, 2002a). However, average production rates were restrained by other system and regulatory limitations, so that the two RSMs had mean production rates over the entire campaign of only 2.28 and 1.38 rockets per hour, respectively. The maximum daily production rates for RSM 1, RSM 2, and the two RSMs combined were 312 (13.0 rockets per hour), 334 (13.9 rockets per hour), and 448 (18.7 rockets per hour), respectively (EG&G, 2002a). Gelled rockets numbering 5,287 from three specific munition lots were processed without draining. The permitted rate of only 1.0 rocket per hour delayed campaign completion, but a decision to coprocess multiple types of GB-filled munitions shortened the time that would have been needed for overall destruc- tion of GB munitions if processing had been accom- plished sequentially. Factors Affecting Operational Experience The very substantial difference between design and experienced disposal production rates for GB-filled M55 rockets at both JACADS and TOCDF suggests a need for careful analysis of cause and effect, including the possibility that the design production rate was set unreasonably high. The multiple causes of delay were unexpected, such as the discovery of gelled and leak- ing agent, equipment failures, faulty operations, and stringent regulatory limitations. The reaction of opera-

OCR for page 15
20 tions management personnel to each of these circum- stances deserves review in light of subsequent events. This is especially true regarding regulatory limitations. Since interruption and delay at any step in a sequential process necessarily affect throughput, an analysis of the total system is required in addition to analyses of individual components. Regu/atory Limitations DFS operation is subject to compliance with both RCRA and Toxic Substances Control Act (TSCA) regu- lations. TOCDF had a RCRA permit to process 33 liquid- filled rockets per hour (EG&G, 2002a). A 5 percent heel of the original 10.7 lb of GB agent was assumed to be present in each of the drained M55 rockets fed to the DFS; the corresponding flow of agent to the DFS is 17.66 lb/in (10.7 lb per rocket x 0.05 x 33 rockets per hour). While there is no indication that the agent feed rate to the DFS is limiting in terms of the 99.9999 percent DRE require- ment, the emission of products of incomplete combus- tion, or the thermal input to the DFS, the revised RCRA permit limited the processing of gelled (undrained) GB M55 rockets to 1.6 per hour to avoid agent flows to the DFS higher than 17.66 lb/h.3 The processing rate was fur- ther reduced to 1.0 rocket per hour when coprocessing was undertaken. Such slow processing of gelled GB M55 rockets at TOCDF significantly extended the operating schedule and slowed the reduction in storage risk. The TSCA permit was required because a polychlori- nated biphenyl (PCB) material had been used as a lubricant when some of the rockets were inserted into their firing tubes, although the quantity was very small.4 During trial burns at JACADS, when the plant was operated at the pro- posed throughput rate, PCBs and controlled PCB products of combustion were found to be below permissible emis- sion limits (NRC, 1994b). The allowable TSCA through- put for JACADS was set at 40 rockets per hour and for TOCDF at 36 per hour. The planned rate for JACADS (32) and the RCRA-established rate for TOCDF (33), which were lower, were the controlling rates (MITRE, 1993~. impact of Leaker Processing While the special handling required for overpacked leaking rockets is no doubt burdensome, there is no 3From the notes of a meeting between a fact-finding group from the Stockpile Committee and the Army, September 25, 2002. 4Ibid. ASSESSMENT OF PROCESSING GELLED GB M55 ROCKETS AT ANNISTON indication in the end-of-campaign reports for JACADS or TOCDF that processing leakers slowed the overall processing throughput appreciably (MITRE, 1991; EG&G, 2002a; U.S. Army, 2002~. A key reason for this view is that the number of leaking rockets was rela- tively small, a total of about 800 at both sites (out of more than 60,000 GB M55 rockets processed). impact of Ge//ed Agent Processing Gelled rockets were not in evidence at JACADS so there was no impact during disposal processing op- erations there. At TOCDF, the story was different (EG&G, 2002a). Three munition lots (1033-55-1076, 1033-55-1077, and 1033-51-1086) totaling 5,287 M55 rockets or 18 percent of the Deseret Chemical Depot stockpile were identified as likely to contain gelled GB material. These munition lots were all pro- cessed through the DFS at 1.0 rocket per hour, as de- scribed earlier in this chapter. The average produc- tion rates for the three lots were 0.5, 0.5, and 0.7 per hour, respectively, for a 24-hour period (U.S. Army, 2002e). These are very much less than the "full rate" of 15.3 per hour achieved at JACADS for processing ungelled rockets. The full rate was developed by tak- ing the total number of rockets processed during the five best production weeks and dropping the highest and lowest weeks.5 At this rate, it would take about 367 full days of operation to process the gelled rock- ets 5,287/0.6 x 24) = 3671 and 67 full days of opera- tion to process the nongelled rockets t(30,000 - 5,287~/~15.3 x 24) = 671. Processing large numbers of gelled rockets is a much more serious impediment to production than processing large numbers of liquid- filled rockets a few of which are leakers. COPROCESSING In an effort to mitigate the impact of slowdowns experienced during the processing of gelled rockets, TOCDF managers conceived techniques for coprocessing munitions. Coprocessing and comple- mentary processing have been defined for planning purposes as follows:6 sInformation from Armv answers to Questions from the Stock- pile Committee as a follow-up to the September 25, 2002, fact- finding meeting with the Army. 6Ibid. ~ 1

OCR for page 15
PROCESSING OF M55 ROCKETS AT JACADS AND TOCDF Co-Processing. Co-processing refers to the concurrent processing of two munition types that use different footprints of the facility and different equipment. For example, bulk items and rockets can be co- processed since they do not utilize the same handling or dem~litari- zation processing equipment. In addition, rockets can be co-pro- cessed with non-explosively configured projectiles. Complementary Processing. Complementary processing involves the processing of two munition/bulk types that utilize a common footprint of the facility or the same equipment. For example, [the Anniston Chemical Agent Disposal Facility] ANCDF is consider- ing complementary processing of explosively configured projectiles with gelled rockets. One ECR will be configured for projectiles, and the second one for rockets. Only one type of munition will be pro- cessed at a given time arid processing of rockets and projectiles will alternate. Also, projectiles would be processed during down periods of the rocket line and vice versa. Thus, gelled rockets could be processed through the RSM and the DFS concurrently with non-explosively- configured projectiles being processed through projec- tile/mortar disassembly machines, multipurpose de- militarization machines, the metal parts furnace (MPF), and the other LIC. A safety-driven operational limitation is that the quantity of munitions in the unpack area must be con- trolled to limit the total energetics load in that space at any given time.7 This was achieved by processing pro- jectiles through the area while rocket processing was suspended for maintenance. Utah regulators granted a Class 1 RCRA permit modification to permit coprocessing, but in so doing, they limited rocket throughput to the DFS to 1.0 rocket per hour while al- lowing the coprocessing of 88 non-explosively-config- ured M360 105-mm projectiles per hour (U.S. Army, 2002f). The time required for rocket destruction was extended as a result of a cut in the disposal processing rate from 1.6 rockets per hour to 1.0 per hour, but the duration of the overall GB disposal campaign schedule was reduced as a result of coprocessing (EG&G, 2002a). PROCESS CHANGES FROM LESSONS LEARNED The lessons learned from the pioneering experience in processing M55 rockets at JACADS (NRC, 1994b) were adopted and built upon at TOCDF (EG&G, 2002a), which contributed uniquely because gelled rockets had not been encountered at JACADS. Lessons 7From the notes of a meeting between a fact-finding group from the Stockpile Committee and the Army, September 25, 2002. 21 learned at the two facilities and applied, later on, to processing or to facility design are discussed next. Lessons from JACADS The DFS kiln wall must be able to withstand po- tential detonation of energetics. It was redesigned and increased in thickness from 0.5 in. to 2 in. for TOCDF and facilities at other sites. The DFS kiln flange bolts failed. The DFS kiln is constructed in five sections that are bolted to- gether to form a single continuous shell. During the GB M55 rocket testing, the bolts holding the kiln sections failed on three occasions. The fail- ures of kiln bolts accounted for 120 hours (or 18 percent) of DFS downtime, the second largest contributor to total downtime. The DFS kiln bolts were replaced with bolts of improved de- sign and different materials of construction. The replacement bolts were larger in diameter, stron- ger, and had a coefficient of thermal expansion that was similar to that of the kiln flanges. There were no failures of these bolts during the VX rocket campaign. The HDC was jammed by slag, pieces of the rocket body, and molten aluminum. The HDC was the largest contributor to JACADS down- time during GB rocket testing, accounting for 248 hours of the 929 hours total downtime. This was 27 percent of the downtime for JACADS and 38 percent of the downtime for the DFS. The HDC mesh conveyor was replaced by a bucket conveyor. The initial testing of the bucket conveyor indicated the drive chain assembly was inadequate for the HDC operating temperature. The chain design was then modified and the con- veyor reassembled with a drive chain assembly that was identical to the one on the mesh con- veyor. This modification was successful, and the only downtime associated with the HDC con- veyor during the OVT 2 VX rocket testing oc- curred when a rocket piece jammed between the conveyor and the HDC housing. All other down- time attributed to the HDC was caused by the heater elements. The system was redesigned and additional preventive maintenance undertaken to avoid breakdown maintenance. The LIC flame detector malfunctioned at high feed rates. First, the feed rate was reduced, and then the flame scanner was properly adjusted.

OCR for page 15
22 ASSESSMENT OF PROCESSING GELLED GB M55 ROCKETS AT ANNISTON . . . . . Glassiiced salt and slag accumulated in the sec- ondary chamber of the LIC. The refractory brick in the secondary chamber of the LIC was replaced with a spell-resistant brick. This brick was more resistant to the corrosive conditions in the sec- ondary chamber. The original bricks were com- posed of alumina and silica, which reacted at high temperature with sodium from the decontamina- tion solutions and with phosphorus from the agent to form slag, which degraded the brick. The sys- tem was redesigned to provide combustion air for each LIC to allow independent operation during slag removal. A hot tap withdrawal system was installed to drain slag from the secondary com- bustion chamber. The DFS feed chute experienced material crack- ing. The chute failed on four separate occasions, accounting for 39 hours of the downtime (6 per- cent of the total) for the DFS. The feed chute was replaced with one of a different design. The rede- sign proved to be inadequate and the chute was replaced by one of a still different design follow- ing the fourth (last) OVT test. Problems with the HDC discharge gates stopped rocket processing on 1I separate occasions. A blast enclosure was installed at the discharge end of the HDC, and the HDC discharge gates were replaced with thicker, ceramic-coated units. This reduced the number of times the gates jammed and reduced the amount of gate warping. Thrust bearings on the DFS failed. The bearings were replaced and relocated to enhance cooling and to facilitate future replacement. The fuse segregator conveyor system malfunc- tioned. It was removed and the rocket cutting se- quence was revised to ensure separation of the fuze from other energetic components. The JA CADS facility was shut down by order of the Environmental Protection Agency (EPA) for 1I days during OVT I after the Army informed the EPA that record-keeping practices at JA CADS were inadequate (NRC, 1994b). During the downtime, better systems were installed. For this reason, and to ensure compliance with all environmental requirements, environmental de- partment staffing was increased. The BRA as originally built at JACADS did not have a PAS associated with it because it was assumed that any emissions would contain only small amounts of nontoxic salts. However, dur- ing OVT 1, particulate emissions exceeded the 30 mg/dscm regulatory limit, and the BRA was shut down (MITRE, 1991~. While OVT 1 was proceeding, a PAS was constructed for the BRA. In October 1991 it was tested with brine from OVT 1 operations spiked with heavy metals. Although the emissions were within regulatory limits, the test was not successful. The tempera- ture of the gas stream into the PAS for the BRA was below the dew point, which caused conden- sation of entrained moisture in the inlet duct. This moisture saturated the salt particulates and caused them to be deposited in the duct instead of entering the baghouse for collection (MITRE, 1993~. This situation was corrected by additional heating of the gas stream. Brine produced dur- ing OVT 1 and OVT 2 was shipped off-island for disposal. During most of the OVT programs, the BRA did not operate satisfactorily. However, after modifications, the BRA did process the brines generated during OVT 3 and OVT 4, al- though some operating problems remained and the required BRA PAS compliance test had not yet been performed. After the OVT program had concluded, the PAS passed the test and the BRA operated satisfactorily until the closure of JACADS. At TOCDF, the BRA was installed but never used because it was cheaper to send brine off-site for processing. Lessons from TOCDF Gelled GB rockets were encountered that could not be drained of agent in the RSM as intended. Thickened or crystallized agent plugged filters, agent collection system components, and the agent quantification system. A modification to the RCRA permit was obtained to allow rockets with a full agent fill to be processed through the DFS at a rate of 1.6 rockets per hour. Addition- ally, coprocessing of GB-filled munitions was undertaken. Although this reduced the allowable processing rate for gelled GB rockets to 1.0 per hour, it improved a disposal schedule that had been adversely affected by munitions contain- ing gelled agent. Overpacked leaking rockets required special handling and delayed the processing rate. For- tunately, there were not very many of them, as noted earlier.

OCR for page 15
PROCESSING OF M55 ROCKETS AT JACADS AND TOCDF VX M55 ROCKET DISPOSAL AT JACADS: ACTUAL VERSUS DESIGN RATE In addition to meeting the DRE for agent destruc- tion and other requirements of the RCRA and TSCA permits, there were two additional process objectives during the OVT 2 with VX M55 rockets (MITRE, 1992): Destroy all 13,889 VX rockets stored on Johnston . Island safely and expeditiously. Determine the effectiveness of the equipment modifications made following the GB rocket OVT 1 testing. Three of the four TSCA DFS trial burns in OVT 2 met the 99.9999 percent DRE requirement for PCBs. The fourth just missed (99.999896 percent) (NRC, 1994b). EPA accepted this result, and nothing was done to remedy it in OVT 2. However, the trial burn results led to a rethinking of the design and operation of the DFS afterburner. In TOCDF and the other mainland baseline facilities, the residence time in the afterburner has been increased from 1 s to 2 s and the temperature has been increased from 2000F to 2200F. The JACADS throughput rate and availability ex- ceeded the goals established for the total duration of the VX M55 rocket testing during OVT 2. All of the 13,889 VX rockets were destroyed during 19 weeks of operation, which commenced on November 15, 1991, and terminated on March 31, 1992 (MITRE, 1992~. The JACADS daily average rocket throughput rate during the full rate part of OVT 2 was 20.6 rockets per hour, which was below the goal of 24.0 rockets per hour. The average throughput rate for the entire test period was 19.6 rockets per hour, which exceeded the throughput goal of 14.7 rockets per hour for the full OVT 2. JACADS was able to maintain a throughput rate of 25.3 rockets per hour for the last 10 days of OVT 2. The throughput rate was 32.0 rockets per hour during the first 10 hours of operation on March 23, 1992, which met the single shift throughput goal of 32 rockets per hour for a 10-hour shift. The integrated system availability for JACADS was 43.4 percent for the duration of VX rocket testing in OVT 2. The integrated system availability for JACADS was 55.4 percent during the full rate portion of the test and 68.9 percent during the last 10 days (MITRE, 1992~. 23 COMPARISON OF GB AND VX M55 ROCKET DISPOSAL CAMPAIGNS Stack Emissions In 1988, Congress mandated that an OVT program be undertaken at JACADS to assess the readiness of the baseline incineration system to process agent safely and effectively. The ability of the technology to meet the emission standards required under TSCA and RCRA was an important criterion in this assessment. One of the four OVT campaigns (OVT 1) destroyed GB M55 rock- ets, and the second (OVT 2) processed VX M55 rockets. The air emissions for all metals and organic compounds from trial burns conducted in these OVT operations met the then-current RCRA requirements with one excep- tion (U.S. Army, 1998a): The mercury level in the MPF stack gas from GB operations was 66 ,ug/m3, somewhat higher than the standard of 50 ,ug/m3. Of particular note is the very low concentration of dioxin and furan in emis- sions from the OVT trial burns at JACADS. The mea- sured result was 0 to 0.16 ng/m3, which is well below the standard of 30 ng/m3 (NRC, 1994b). Trial burns were also conducted at TOCDF during the systemization (preoperational testing) of the facil- ity. Lead levels in the DFS emissions from the GB M55 rocket trial burn were extremely high, 1,101 ,ug/m3, well over the standard of 270 ,ug/m3 (U.S. Army, 1998a). The propellant in each rocket contains 0.4 lb of lead stearate. The fuze has a lead rotor, and the detona- tor contains lead styphnate and lead azide. These are likely contributors to the high lead emissions, but the Army has not developed a precise rationale for why lead emissions were so much higher at TOCDF than at JACADS. As in the JACADS tests, dioxin and furan emissions in the TOCDF tests were extremely low and well below the 30 ng/m3 standard. Two final points are important. First, the HEPA and carbon filters that make up the PFS being incorporated into the PAS in baseline facility designs for the Anniston, Umatilla, and Pine Bluff sites should reduce emissions at these facilities below those reported for JACADS and TOCDF. Second, none of the agent trial burns conducted to date at JACADS and TOCDF have included destruction of gelled GB. Throughput Rates A comparison of throughput rates for the OVT tests with GB and VX rockets at JACADS reveals that the

OCR for page 15
24 average full rate throughput increased from 15.3 rock- ets per hour during OVT 1 (GB rockets) to 20.6 rockets per hour during OVT 2 (VX rockets), a 42 percent in- crease. The maximum throughput rate demonstrated during the GB rocket testing was 27 rockets per hour for 4 hours. During VX rocket testing, the maximum throughput rate was 32 rockets per hour. This rate matched the design throughput rate and was sustained for one complete 10-hour shift (MITRE, 1992~. The integrated system availability of JACADS to process rockets increased from 33 percent during OVT 1 to 46.8 percent during OVT 2, after adjusting for the downtimes caused by the weather and the fuze segregator conveyor of the RSM (MITRE, 1992~. This reflected the benefits of a more experienced workforce and learning experiences during OVT 1 that resulted in major process improvements. Some of the improvements created another set of problems. For example, because JACADS processed rockets more rapidly during OVT 2, the DFS furnace room operated at a higher temperature. This caused the HDC heater element fuses to blow, resulting in ap- proximately 200 h JACADS downtime. After OVT 2, the heater fuse box was relocated to a cooler location, outside the DFS furnace room. Safely Performance The safety performance of JACADS personnel was better in both OVT 1 and OVT 2 than the program goal. The Army elected to use the metric "cases with days away" (CWDA) per 200,000 hours worked to monitor safety performance (NRC, 1994b). The CWDA rates realized in OVT 1 and OVT 2 were 1.2 and 2.9, respectively better than the goals of 4.1 and 3.3 that were predetermined for the operations. A metric more often used in industry is the record- able injury rate per 200,000 hours worked (RIR). The RIR covers the CWDA cases but also includes injuries where the worker goes back to work after medical treat- ment. The RIR for OVT 1 was 5.8 and for OVT 2 was 5.7. These values were very high by industrial stan- dards for ongoing operations. Army contractors subse- quently improved worker safety programs and the RIRs for JACADS. Environmental Performance The environmental performance of JACADS con- tinued to be a high priority during OVT 2. There were ASSESSMENT OF PROCESSING GELLED GB M55 ROCKETS AT ANNISTON more reported instances of environmental noncompli- ance at JACADS during OVT 2 than during OVT 1. This was primarily due to the aggressive efforts of plant personnel to identify and correct any area that was not in strict compliance with the appropriate permit or regulation. A self-audit program was implemented to identify activities that were not performed in accor- dance with permit requirements. A training program was implemented to inform the JACADS workforce of the applicable permit requirements. Whenever an ac- tivity was identified as not being in compliance with the permit, the noncompliance was documented and a corrective action program was initiated. While some of the noncompliances could not be corrected during OVT 2 because long-term solutions or permit modifi- cations were required, all instances of noncompliance were addressed. All RCRA emission limits were met. No releases of VX agent to the environment have been documented. The seawater discharge quantity and tem- perature were maintained within National Pollution Discharge Elimination System permit limits. All solid hazardous wastes were properly disposed of in an EPA- approved landfill (MITRE, 1992). The operation of JACADS during OVT 2 proved that the baseline sys- tem technology could be operated safely and in an en- vironmentally sound manner. The safety pro cram con- tinued to function adequately. SUMMARY OBSERVATIONS ON M55 ROCKET DISPOSAL EXPERIENCE The following major modifications were imple- mented after encountering problems at JACADS: The DFS kiln wall thickness was increased from 0.5 in. to 2 in. The furnace bearings were relocated to prevent overheating. The HDC was redesigned to avoid downtime as- sociated with molten aluminum problems en- countered in the original design. Notwithstanding that each site is unique with respect to the number and type of munitions stored, lot num- bers represented, regulatory climate, public affairs cli- mate, numbers and types of anomalous munitions, and to some extent, system design, a number of issues com- mon to baseline facilities are apparent from a review of the experience in processing M55 rockets at JACADS and TOCDF:

OCR for page 15
PROCESSING OF M55 ROCKETS AT JACADS AND TOCDF . At JACADS and TOCDF, processing rates for the M55 rockets in the DFS were established and sub- sequently demonstrated in trial burns based on their handling in the RSM and on the thermal loading of drained rockets containing 5 percent or less of their original agent charge. Since gelled GB could not be drained from the rockets, the processing rate was arbitrarily reduced by a fac- tor of 20, because a gelled rocket contained about 20 times as much agent as a drained ungelled rocket. Gelled GB agent will not drain as intended, ne- cessitating identification of the anomalous rock- ets and the lot number of their contents and forc- ing the modification of some process steps. RCRA permits must acknowledge the process modifications, and programmable logic control- lers must be adjusted to achieve the necessary changes in process control. Rocket handling and transportation to and through the unpack area are identical for gelled and liquid-filled rockets. The RSM must be re- programmed for gelled rockets to skip the drain station and, accordingly, the agent quantification system. Coprocessing is a proven option for expediting the completion of a disposal campaign for GB . 25 when the need to process rockets containing gelled agent reduces throughput rates. The control and sensing of internal DFS kiln tem- perature and pressure remain challenging issues. Energetics burn quickly, producing temperature and pressure spikes. Exceedance of set tempera- ture and pressure limits can cause thermal stress in the kiln wall and feed chute. Cracks that re- quired repair were found in the furnace wall dur- ing inspections. Although these are not unusual in furnace operations, they are an indicator of thermal stress. A majority of DFS operational downtime can be attributed to three causes: HDC jams (27 percent), DFS bolt failures (18 percent), and DFS feed chute jams (6 percent). Equipment has been modi- fied to address these causes, including a modified DFS feed chute design at ANCDF that is expected to mitigate jamming. Overpacked leaking rockets must be handled separately and at a somewhat slower throughput rate. PCBs do not present a problem in achieving ap- propriate DRE levels when processing M55 rock- ets. VX rockets have not shown agent gelling, and there is nothing currently known to suggest that they might.