THE USE OF PROBABILISTIC RISK ASSESSMENT TECHNOLOGY FOR EVALUATING MC&A EFFECTIVENESS AND RELATIVE RISK CONTRIBUTIONS

Probabilistic Risk Assessment (PRA) technology is routinely employed to evaluate the relative risks associated with operation of chemical facilities, space operations, and the operation of nuclear power plants (NPPs). The United States and Western Europe require the use of PRA in evaluating nuclear power plant operations, modifications and special activities. Under the auspices and guidance of the International Atomic Energy Agency (IAEA) many earlier Soviet designed and built nuclear plants are also conducting such evaluations. Evaluation of PRA technology indicated that PRA could be used for assessment of risks associated with material control and accountability (MC&A) processes in facilities in addition to nuclear power plants. An examination of the functional elements needed for good nuclear MC&A practices, leads to the conclusion that applying PRA to the evaluation of risks associated with MC&A results in a focus on institutional behavior more than on the reliability of specific hardware. A fault tree for the basic reliability of the MC&A process at a nuclear facility has been developed. This fault tree can be used to assess the increase in risk associated with a number of “initiating events” consisting of adversarial actions. Quantification of the basic events of the fault tree is done by MC&A specialists using a “Delphi” process. This process is guided by a set of metrics, relating the fault tree elements to the output of a survey which examines MC&A at the facility under consideration. The results of the fault tree quantification provide relative risk values associated with the facility being examined. They also provide quantitative evaluations of the dominant contributors to that risk; the worth, in terms of risk, of reducing those dominant elements to insignificance; and the impact of the various proposed initiating events on the risk picture. In this latter context, the model allows the user to identify the “weak links” contributing most to the relative risk for each “initiating event” and to assess the worth of corrective actions.