Affordable Cleanup?: Opportunities for Cost Reduction in the Decontamination and Decommissioning of the Nation's Uranium Enrichment Facilities (1996)

This appendix summarizes information from other decontamination and decommissioning (D&D) efforts that the committee examined and identifies lessons learned. The only gaseous diffusion plant (GDP) that has undergone D&D is the Capenhurst GDP in the United Kingdom (see, for example, Baxter and Bradbury, 1991; Clements, 1994a,b; Spencer, 1988). The committee examined this program in detail as discussed below. The committee also benefited from briefings on the D&D of sites that were used to produce nuclear materials and components for the Manhattan Project and to fabricate nuclear fuel. Two D&D projects were reviewed in particular to glean any insights for the planning and execution of the D&D of the U.S. GDPs: the Formerly Utilized Sites Remedial Action Program and the Apollo Decommissioning Project (Kingsley, 1994). The first is an ongoing project to decontaminate and decommission facilities associated with the Manhattan Project (Hovey, 1994). The Apollo project involved decommissioning a fuel-fabrication complex that handled both enriched uranium and plutonium. Because the scope of these projects varied substantially from that of the D&D of the U.S. GDPs (these projects included substantial remedial action, such as cleanup of soils and groundwater), their costs were not useful for evaluating the GDP D&D costs.

The committee believes that the greatest opportunity to minimize the cost of D&D occurs during planning. Timely decisions on key issues are essential to execute the work efficiently within budget and schedule.

The Capenhurst GDP enrichment cascade consisted of 4,808 stages in series, each containing a converter housing the diffusion barrier material that separates uranium-235 (²³⁵U) and uranium-238 (²³⁸U) isotopes, a compressor and associated drive motor, a cooler, and interstage piping and valves. The cascade components were housed in a single process building 1,200 m long by 150 m wide. The stages were on the ground floor, and auxiliary equipment, such as electrical and heating, ventilation, and air conditioning (HVAC) systems, were located on the second floor. There were seven different sizes of converters and compressor drive motors, the latter ranging in size up to 300 hp. The cascade equipment was arranged in process cells containing 8 to 12 stages each.

Uranium feed from reactor returns was introduced into the cascade in 1962, resulting in contamination of the low-enrichment stages with technetium-99 (⁹⁹Tc) and neptunium-237 (²³⁷Np). The presence of ⁹⁹Tc in particular made the decontamination effort more difficult.

Before shutting down the plant, radiological and criticality data were gathered for use in planning and executing the dismantling, decontamination, and disposal operations. Gaseous decontamination using chlorine trifluoride (ClF₃), a fluorinating agent, was used to remove the bulk of the residual uranium deposits. Detailed radiological surveys were performed to locate these deposits.

A detailed D&D plan was developed. The initial phase involved cutout, removal, sealing, and outdoor storage of the cascade components. This removal allowed a part of the process building housing the cascade to be demolished and the land returned to greenfield status. Other parts of the process building were reused to house equipment for size reduction of components, chemical decontamination, and melting of metal components and pieces that were difficult to decontaminate.

A great deal of effort went into researching and developing cost-effective techniques for decommissioning. The key principle underlying this development work was to look outside the nuclear industry for off-the-shelf equipment that, with or without modification, would meet the D&D program needs.

The development activities performed during this period addressed the following technical issues:

• selection of a cost-effective and safe means of disassembling the plant;
• suitable size-reduction techniques and compatible ventilation and filtration systems;
• decontamination processes to deal specifically with transuranic and fission products, copper, and other metals;
• engineering safety in process equipment; and
• ensuring compatibility of waste streams with regulatory requirements.

One of the main objectives of these development activities was to minimize D&D waste by maximizing recycle for unrestricted reuse.

The decontamination and disassembly process consisted of the following activities:

• gaseous decontamination, prior to plant shutdown, to convert solid uranium deposits, primarily uranyl fluoride (UO₂F₂), to volatile fluoride compounds;
• plant characterization to identify and quantify residual deposits of radioactive materials;
• removal of nonradioactive hazardous materials such as asbestos and polychlorinated biphenyls (PCBs);
• removal and interim storage of plant equipment and removal of cell structures;
• size reduction of components;
• aqueous chemical decontamination;
• melting difficult-to-decontaminate metal parts;
• removal of process and ancillary building structures; and
• disposal of radioactive and hazardous waste.

Many items, such as structural steel and concrete, required only minimal decontamination.

Gaseous ClF₃ was circulated through the cascade prior to shutdown and dismantlement to convert solid deposits to volatile fluorides (e.g., converting UO₂F₂ to uranium hexafluoride, UF₆) prior to opening up the system. Gaseous decontamination removed an estimated 80 percent of the UO₂F₂ deposits, and substantially reduced the probability of a criticality accident during subsequent D&D operations. Both HVAC and physical methods were used to protect the workers and the environment.

Following gaseous decontamination, further cleanup and pretreatment operations were carried out on the static plant to locate and deal with any significant pockets of contamination that remained to permit safe and cost-effective intrusions into the plant during the dismantling campaigns. Cleanup techniques included vacuuming, ridding, and machining.

The initial characterization to identify and quantify residual radioactive contaminants was performed following gaseous decontamination. Gamma spectroscopy and neutron activation were

used to characterize uranium deposits. Counters and scintillation monitors were used to identify ⁹⁹Tc and ²³⁷Np deposits.

The characterization provided data on the magnitude and location of alpha (uranium) and soft beta (technetium) radionuclides throughout the plant. Nonintrusive gamma spectroscopy and neutron activation measurements provided the necessary data on ²³⁷Np and ²³⁵U.

Hazardous materials, such as asbestos, PCBs, lubricants, and laboratory chemicals, were removed and disposed of using conventional technologies, including land burial for asbestos and incineration for PCBs.

The initial phase of dismantling and disassembly consisted of cutting out, removing, and storing compressors, coolers, valves, large-diameter pipe, and large-process stage units.

Specialized workshops were built for component stripping and dry cleaning. Protection of personnel was achieved by effective ventilation and extensive alpha-in-air monitoring throughout the facility. A criticality detection system was installed, and strict criticality control procedures were applied at each stage of the dismantling process.

The low-enrichment stages of the cascade had been fed with reactor recycled UF₆ during operations for civilian purposes. This material included small but significant quantities of transuranic elements and fission products. Safe handling of contamination such as ²³⁷Np and ⁹⁹Tc had to be ensured during dismantling activities.

Following process equipment removal, the remaining cell enclosures were demolished. The materials were sold as clean scrap in the commercial metals market. The building shell was removed from about one-half of the total structure. The floors were scabbled and removed, returning that part of the structure to greenfield site status.

Before the decontamination plant was available, a large section of the process building (including the building structure and floor slab) was completely cleared to make way for construction of a new centrifuge enrichment facility. The diffusion plant cascade equipment removed was stored outdoors for up to 9 years until the new decontamination facility was available. Approximately 6,000 metric tons of contaminated components were stored, including 700 large stages weighing up to 7 metric tons each.

Large components, such as converter shells, piping, and compressors were reduced in size and weight to meet requirements of the decontamination plant and the melter. Cold cutting was preferred over hot cutting for aluminum components, because cold cutting does not generate

fumes or airborne aluminum oxide fines, thereby reducing the need for costly HVAC systems. Robotic plasma cutting was used for size reduction of large aluminum converter shells, and remotely controlled oxyacetylene methods were used for cutting steel converter shells and other steel components. A total of 1,400 seam-welded steel shells were cut using the latter method.

The HVAC system was divided into two stages. The first stage consisted of a self-cleaning filter unit in which intermittent reverse air pulses dislodged the dust that was collected in bags at the base of the unit. The balance of particulate matter was captured in a high-efficiency particulate air filter system. The effluent air stream was monitored by stack monitors before being released to the atmosphere. Overall filtration efficiency was greater than 99.997 percent.

A wet decontamination process was used to remove uranium contamination down to free-release levels. While chemical treatment for the removal of uranium and its daughter products is a well-established process, technetium is difficult to remove effectively. A means of removing ⁹⁹Tc had to be developed before effective disposal routes could be determined. Following extensive laboratory and pilot plant investigation, a full-scale decontamination plant was built in 1989. The flowsheet was based on achieving plant discharges having a negligible impact on the environment, and on satisfying the United Kingdom statutory regulations for recycling scrap metals to the open market. Most of the uranium was removed with citric acid, followed by sulfuric acid, disodium citrate, and a hot water wash. The majority of the ⁹⁹Tc and ²³⁷Np ended up in the citric acid (Anderson and Faulkner, 1989).

Separate processing plants were used to clean up the spent citric acid, sulfuric acid, and disodium citrate decontamination liquors. Ion exchange removed contaminants from the process solutions, substantially reducing the volume of waste. The ion exchange resins were encapsulated in concrete and sent to the low-level waste burial ground at Drigg.

Strict criticality control was maintained, with detectors placed at key points in the decontamination facility. The activity of each individual piece was monitored after decontamination to ensure it met the applicable release criteria.

A melter was used to handle metallic components that were difficult or impossible to decontaminate cost-effectively by chemical means. The melter had several functions:

• removing impurities from aluminum, steel, and other metals to increase resale value;
• homogenizing the radioactivity in materials with varying degrees of contamination, shapes, and sizes; and
• reducing waste volume.

A number of ancillary buildings and structures were demolished, including 11 large natural draft cooling towers, their pump houses, and an electrical substation. Including the floor slab, this operation produced 46,000 metric tons of clean concrete rubble for off-site disposal.

Metallic materials recovered from the plant were categorized according to their potential for sale to the commercial metals market as follows:

• clean scrap;
• contaminated scrap economical to decontaminate to de minimis level; and
• contaminated scrap uneconomical to decontaminate to de minimis level.

Clean scrap, such as cell cubicle structures, base plates, and some motors, was sold directly to the metals market. Scrap that was economical to decontaminate to de minimis levels was reduced in size, decontaminated and/or melted to homogenize the contamination, and sold. This approach was used for the bulk of the steel, copper, and aluminum components. Scrap that was uneconomical to recover, such as small-bore pipe and instruments, was dispatched to the low-level radioactive waste site at Drigg.

Approximately 99 percent of the material removed from the Capenhurst plant was recycled to the commercial markets, including bulk concrete as well as metals.

Personnel protection was achieved through multiple methods:

• strict criticality control with criticality detectors on selected operations;
• extensive alpha-in-air monitoring; and
• special HVAC systems with high-efficiency particulate air filters.

Criticality control was achieved by a number of actions:

• removing as much uranic contamination as possible during the size reduction, decontamination, and preparation stages;
• designing the plant to minimize the likelihood of criticality incidents; and

• using batch metering techniques to control spent citric acid movements and concentrations of ²³⁵U.

Static personnel air samples and film badges were used throughout the project. Whole body monitoring was performed twice yearly for each D&D worker. No special dispensation or relaxation of exposure limits was given for this work.

Very low-levels of exposure were experienced by the work force. The mean total dose for 1993 was 0.03 mSv. These low-levels were achieved by sound engineering design, safe operational practices, and a methodical approach to safety.

Although the Capenhurst plant was substantially smaller than the U.S. GDPs, they have many similarities:

• similar process flowsheets and cascade arrangement;
• multistory, steel-frame and concrete buildings with transite siding;
• stages grouped into cells;
• Freon-cooled stages;
• same species of radiological contamination, including uranium (from depleted to highly enriched), ⁹⁹Tc, and ²³⁷Np;
• large quantities of hazardous materials, such as asbestos, PCBs, and Freon®;
• mixture of aluminum and nickel-plated stage components;
• steam-heated autoclaves for feed vaporization; and
• purge cascade for removal of light gases.

The principal differences between Capenhurst and the U.S. GDPs are the following:

• Physical size and separative work capacity of the U.S. plants are substantially larger.
• Most of the large interstage piping at Capenhurst was aluminum, whereas U.S. GDPs use nickel-plated steel exclusively.

• U.S. GDPs have a larger number of support facilities to decontaminate and decommission than Capenhurst had. For example, the D&D scope at the Oak Ridge and Portsmouth GDPs includes centrifuge enrichment facilities. However, these were not included in the cost estimates under review by the committee.
• Capenhurst cascade equipment was located on the first floor of the process building; this equipment is located on the second floor in the U.S. plants.

Since the technology employed in the Capenhurst plant was very similar to that in the U.S. enrichment plants, a comparison of various quantities associated with them provides a means to estimate the relative cost of cleaning up the U.S. facilities. Selected quantities and design features of the Capenhurst and the Oak Ridge GDPs are shown in Table H-1. A breakdown of the quantities of various metals at Capenhurst and the three U.S. GDPs is presented in Table H-2. Table H-3 presents the ratio of Oak Ridge to Capenhurst GDP for total metals contained in the cascade, building footprint, total area under roof, weight of largest converter, and peak electric power.

A direct comparison of plant separative work capacities was not possible because the Capenhurst capacity is considered to be proprietary information. However, enrichment plant capacity (in separative work units [SWUs]/year) is nearly directly proportional to power consumption because most plant electric power usage is to drive the UF₆ compressors. For example, specific power consumption is reported as 2,433 Kwh/SWU for U.S. enrichment plants and 2,538 Kwh/SWU for the Eurodif plant in France, although these facilities have substantially different stage designs (Kroschwitz and Howe-Grant, 1993). Other things being equal, total UF₆ flow through the cascade is proportional to power and plant capacity and would also be indicative of equipment size and its associated D&D cost. Peak power delivered to the Oak Ridge GDP facilities was 1,725 MW, compared with 300 MW for Capenhurst, a ratio of 5.7.

Another consideration is that 19.7 percent of the total Capenhurst D&D cost was for technology development, which is excluded from the Ebasco cost estimate. Subtracting the cost of technology development would reduce the total Capenhurst cost from $160 million to $128 million. The cost of planning for the Capenhurst D&D was 11.6 percent of the total cost. This should not be appreciably larger for a large plant than a smaller plant when the two plants have similar systems, structures, and contaminants. Similarly, the cost of selecting the most cost-effective D&D techniques should not be substantially different, particularly when there is a substantial experience base available from Capenhurst and other successful D&D projects.

There are other factors that may increase or decrease the cost of D&D of the U.S. GDPs relative to Capenhurst. Differences in the management and contracting approach, wage rates, labor productivity, and regulatory requirements are some of the important considerations. Although the Capenhurst D&D was a government program sponsored by the Central Electricity Generating Board and the United Kingdom's Ministry of Defense, there was apparently a very high commitment to cost control, as evidenced by the relatively small number of management

personnel (Clements, 1994b). The management and contracting approach appears to include a much larger portion of the total costs associated with performing the work rather than managing it.

Anderson, R.W., and R.L. Faulkner. 1989. British Nuclear Fuels plc Process for Decontamination and Decommissioning of Capenhurst Gaseous Diffusion Plant and Information from the International Conference on Decommissioning of Major Radioactive Facilities. K/QT-254. Washington, D.C.: Oak Ridge Gaseous Diffusion Plant for U.S. Department of Energy (DOE).

Baxter, S. G. and P. Bradbury. 1991. BNFL's [British Nuclear Fuel Limited's] Capenhurst Diffusion Plant Decommissioning. August. Chester, U.K.: BNFL.

Clements, D.W. 1993. BNFL Capenhurst Works. Minimizing Nuclear Waste by Special Recovery Techniques. Presented at the 4th International Environmental and Waste Management Conference held April 18–21, 1993 in Knoxville, Tennessee.

Clements, D. 1994a. Capenhurst Diffusion Plant Decommissioning, BNFL, Enrichment Division. Presented to the Committee on Decontamination and Decommissioning of Uranium Enrichment Facilities at the National Academy of Sciences, Washington, D.C. March 28, 1994.

Clements, D. 1994b. Personal communication to committee member Ray Sandberg, November 6, 1994.

DOE. 1991. Environmental Restoration of the Gaseous Diffusion Plants. Technical Summary Document, vol. 1. Washington, D.C.: Ebasco Environmental and Main-1893 for the DOE. October.

Hovey, G. 1994. D&D Project Management Perspectives From Formerly Utilized Sites Remedial Action Program (FUSRAP). Presented to the Committee on Decontamination and Decommissioning of the Uranium Enrichment Facilities at the National Academy of Sciences, Washington, D.C., June 16, 1994.

IMF (International Monetary Fund). 1993. International Financial Statistics Yearbook. Washington, D.C.: IMF.

Kingsley, R.S. 1994. Apollo Decommissioning Project, Decommissioning a LEU/HEU [Low Enriched Uranium/Highly Enriched Uranium] Fuel Fabrication Complex, Babcock and Wilcox Nuclear Environmental Services, Inc. Presented to the Committee on Decontamination and Decommissioning of the Uranium Enrichment Facilities, National Academy of Sciences, Washington, D.C., June 15–16, 1994

Kroschwitz, J.I., and M. Howe-Grant, eds. 1993. P. 175 in Kirk Othmer Encyclopedia of Chemical Technology, 4th Edition. New York: Wiley.

Spencer, A.N. 1988. Decommissioning and Decontamination of Enrichment Facilities. Presented at the 15th Annual Meeting of the World Nuclear Fuel Market, held in Seville, Spain, October 16–19, 1988.