Lapas attēli
PDF
ePub

veillance elements of safeguards for existing facilities.

Heretofore, the development of instruments has been focused on the need to measure uranium and plutonium generated in fabrication processes that keep these elements isolated from one another and are free of fission product contamination.

Current studies and conceptual process and facility designs involve both reprocessing and coprocessed uranium and plutonium fuels. These material forms and compositions pose a whole new set of measurement problems, which we have only begun to address. In the international arena, the fact that fuel reprocessing and development of the breeder are proceeding now makes solving the associated measurement and accounting problems an urgent need. Some of the materials for which NDA measurement techniques should be developed, tested, and evaluated are: leached spent-fuel hulls and other "hot" scrap and waste, mixed uranium and plutonium solutions with the various levels of fission product contamination characteristic of reprocessing, "cold" mixed plutonium and uranium solutions. and solids characteristic of co-converson (from nitrate to oxide), and final breeder and (perhaps plutonium recycle) fuels. Futhermore, until commercial (LWR) spent fuels are reprocessed, they must be stored at the reactor sites or at special away from-reactor (AFR) facilities. The spent fuels should be verified and measured quantitatively, if possible, to establish their economic value and to control batch make up for future reprocessing, as well as for safeguards.

Concerns are increasing about safeguards for uranium enrichment, the front end of the fuel cycle, as advanced isotope separation processes based on centrifuges, lasers, or plasma devices are implemented. Because of their inherently large isotope separation factors, fewer separation stages are required to produce high-enriched uranium than are needed for classical gaseous diffusion separators. Hence the advanced isotope separation techniques could be more easily used for covert production of bomb materials by the facility operator. New safeguards measurement problems associated with these advanced methods must be addressed.

Process lines in future facilities will use advanced processing and control technology, including remote operation and maintenance, to minimize personnel radiation exposure, assure material containment, and handle high throughput efficiently. Almost certainly more, not less, measurement instrumentation will be needed in process lines. The instru ments must function very reliably, even in such high-radiation environments as reprocessing canyons, because of the restricted access. Individual measurement stations must be designed to meet the well ordered requirements of costeffective integrated, automated systems of materials accounting and control. At present, LASL and the Hanford Engineering Development Laboratory are planning NDA instrumentation to be incorporated at DOE's new Fuel Materials Examination Facility at Richland in an automated mixed uranium and plutonium oxide fuel fabrication line. With the DOE/ORNL

Consolidated Fuel Reprocessing Program, LASL also is working on the conceptual safeguards design for an advanced fuel-reprocessing plant.

Implementation of advanced automated materials accounting and control systems may constitute an enigma for safeguards inspectors. The question is how an inspector can independently verify that a large complex integrated system has been operated as declared. Providing system integrity with sufficient transparency for verification will undoubtedly bring changes in the design of individual instruments and their measurement controls. Inspectors will need improved portable instruments. to confirm the results of the large inplant systems.

Safeguards technology, like waste management, was once viewed by many scientists and planners as a supportive, ancillary factor in the development of nuclear power for the generation of electricity. These factors, together with reactor safety, have now become dominant in the determination of the future of nuclear energy. Thus levels of activity in future safeguards instrumentation. development just described will depend to a large extent on the acceptance and use of nuclear energy, which in turn will be strongly influenced by public perception of the safety of reactors, the capabilities for safeguarding these fuels for peaceful uses, and safe management of the associated nuclear waste materials. In this chicken-and-egg cycle, we hope that the significant advances in safeguards technology, as well as in reactor safety and waste management, will be given due consideration.

[ocr errors]

Howard O. Menlove, Group Leader of the International Safeguards Group, has worked with the Nuclear Safeguards Program at LASL for 12 years. He has been active in research and development of advanced techniques for nondestructive assay of fissionable materials. At present, his work is in inspector instrumentation development and implementation, nondestructive assay standards and calibration, spent-fuel verification techniques, training, and technology transfer. Before joining LASL, he had considerable experience in neutron and fission physics and in gamma-ray spectroscopy. After earning his Ph.D. in nuclear engineering at Stanford University, he spent a year at the Kernforschungszentrum in Karlsruhe, FRG, supported by a Fullbright Award.

Roddy B. Walton earned his bachelor of science degree in physics at Texas A&M University, and his Ph.D. in nuclear physics at the University of Wisconsin in 1957. From 1959 to 1967, he was with the General Atomic Company, where he was involved in electron linac instrumentation and experiments in neutron thermalization, neutron capture cross sections, delayed gamma rays from fission, and photonuclear reactions. In 1967, he joined LASL, and has been working in development and implementation of methods for the measurement of fissionable materials for the nuclear safeguards program. He has been instrumental in the initiation of new projects, including the Mobile Nondestructive Assay Laboratory and its operation in the field, DYMAC (dynamic materials accounting), and systems analysis. He is a Fellow of the American Nuclear Society.

[ocr errors]
[graphic][graphic]
[graphic]

Systems analysis suggests that near-real-time materials accounting systems designed for future large-throughput nuclear facilities can meet high performance standards.

move through the various processing stages and must keep track of them so well that the absence of even small amounts can be detected. The uncertainties inherent in any measurement process and the difficulties of measuring in high-radiation fields behind heavy shielding complicate this task.

We also illustrate the potential benefits of these systems by describing the development and expected performance of a materials accounting system we have designed for the Allied-General Nuclear Services (AGNS) spent-fuel reprocessing plant at Barnwell, South Carolina (Fig. 1). This plant was designed to process large amounts of irradiated fuel from power reactors. The accounting system was designed after the plant was built and with simulated data because the plant is not yet operating.

The potential of system performance is based on projected measurement capabilities of instruments, some of which are still under development. These projections cannot be tested without access to an operating facility. However, our preliminary evaluations suggest that we can design dynamic materials accounting systems for large bulkprocessing facilities that meet detection. standards close to those recommended by the IAEA.

[blocks in formation]

ques are used to decide whether a balance indicates diversion of material.

At present, materials balances are drawn around an entire plant or a major portion of it after the facility has been shut down and cleaned out to inventory the material present. Although such accounting methods are essential to safeguards control of nuclear materials, they have inherent limitations in sensitivity and timeliness. The sensitivity is limited by measurement uncertainties that may conceal losses of significant quantities of nuclear material in large plants. The timeliness is limited by the frequency of physical inventories; that is, the practical limits on how often a facility can be shut down for inventory and still remain productive.

Both sensitivity and timeliness can be improved by implementation of dynamic materials accounting. This approach combines conventional chemical analysis, weighing, and volume measurements with the on-line measurement capability of NDA (nondestructive assay) instrumentation to provide rapid and accurate assessment of the locations and amounts of nuclear material in a facility. Materials balances are drawn without shutting down the plant: inprocess inventories are measured, or otherwise estimated, while the process is operating.

To implement the approach, the facility is partitioned into several discrete accounting areas. Each accounting area contains one or more chemical or physical processes and is chosen on the basis of process logic and the ability to draw a materials balance, rather than on geography, custodianship, or regulatory requirements. By measuring all material flows in each area separately, quantities of material much smaller than the total plant inventory can be controlled on a timely basis and any discrepancies can be localized to the portion of the process contained in the accounting area.

Control by dynamic materials accounting is rigorous. It forces a potential divertor to steal nuclear material in quantities small enough to be masked by measurement uncertainties. Thus, to obtain a significant quantity of material, the divertor must commit many thefts and run the concomitant high risk of detection by the accounting system, surveillance instruments, and physical protection system.

Designing a Materials Accounting System

The performance, or diversion detection sensitivity, of a materials accounting system depends on the details of the measurement system, which in turn depend on the details of the process. Because these details vary from one plant to another, the Los Alamos safeguards systems studies focus on specific designs of existing or planned nuclear facilities.

The first step in the development of a facility's accounting system is to determine the flows of nuclear materials through the facility from design data and operator experience. Then, the facility is partitioned into logical accounting areas, and an appropriate measurement system is postulated for each area. Wherever

[blocks in formation]

To develop preliminary designs of materials accounting systems, we model and simulate the in-plant processes and measurement systems by computer because no large fuel-cycle plants are yet in operation. Detailed dynamic models of material flows are based on actual process design data. They include bulk flow rates, concentrations of nuclear materials, holdup of materials in the process line, and the variability of all these quantities. Design concepts for the accounting systems are evolved by identifying key measurement points and appropriate measurement techniques, comparing possible materials accounting strategies, developing and testing appropriate data-analysis algorithms, and quantitatively evaluating the proposed system's capability to detect losses. The use of modeling and simulation allows us to study the effects of process and measurement variations over long operating periods and for various operating modes in a short time.

Computer codes simulate the operation of each model process using standard Monte Carlo techniques. Input data include initial values for all process variables and values of statistical parameters that describe each independent process variable. These data are best estimates obtained from process designers and operators. Each accounting area is modeled separately. When a process event occurs in a particular area, the values of the flows and in-process inventories associated with that part of the process are computed

*See "Nondestructive Assay for Nuclear Safeguards.

and stored in data matrix. These data are available for further processing and as input to computer codes that simulate accounting measurements and materials balances.

The flow and inventory quantities from a simulated process model are converted to measured values by applying simulated measurements. Each measurement type is modeled separately; measurement errors are assumed to be normally distributed (Gaussian), and provisions are made for both additive (absolute) and multiplicative (relative) errors. Significant measurement correlations are included explicitly. In most cases the measurement models are derived from the performance of similar instrumentation that has been used and characterized in laboratory and field applications involving similar materials. Simulated measurements are combined to form dynamic materials balances under various accounting strategies.

Data Analysis

We combine the most promising measurement and accounting strategies with statistical techniques in comparative studies of loss-detection sensitivities. One of the major functions of the materials accounting system is to indicate loss, or possible diversion. Diversion may occur in two basic patterns: abrupt diversion (the single theft of a relatively large amount of nuclear material) and protracted diversion (repeated thefts of nuclear material on a scale too small to be detected in a single materials balance because of measurement uncertainties). Protracted diversion usually is the most difficult to detect.

The use of dynamic materials accounting enhances the ability to detect both diversion patterns, but it results in the rapid accumulation of relatively large quantities of materials accounting data. For example, if an area's materials

« iepriekšējāTurpināt »