Probabilistic basis and assessment methodology for effectiveness of protecting nuclear materials

Safeguards and security systems for nuclear facilities include material control and accounting (MC&A) and a physical protection system (PPS) to protect against theft, sabotage and other malevolent human acts. The insider threat is most often addressed as part of the evaluation of a facility´s PPS. A PPS is evaluated using probabilistic analysis of adversary paths on the basis of detection, delay, and response timelines to determine timely detection. Because insider adversaries have access to, knowledge of, and authority for facility operations, the PPS actually provides minimal protection against the insider threat. By monitoring and tracking critical materials, MC&A activities are an important protection element against inside adversaries. Timely detection for MC&A activities, however, has been difficult to determine so that for the most part, the effectiveness of these activities has not been explicitly incorporated in the insider threat evaluation of a PPS. This paper presents research on a new approach to incorporate MC&A protection elements explicitly within the existing probabilistic path analysis methodology. MC&A activities, from monitoring to inventory measurements, provide many, often recurring opportunities to determine the status of critical items, including detection of missing materials. Human reliability analysis methods for nuclear power plant operations are used to determine human error probabilities to characterize the detection capabilities of MC&A activities. An object-based state machine paradigm was developed to characterize the path elements and timing of an insider theft scenario as a race against MC&A detection that can move a facility from a normal state to an alert state having additional detection opportunities. Event sequence diagrams describe insider paths through the PPS and also incorporate MC&A activities as path elements. To address the insider threat, this work establishes a probabilistic b- sis for timely MC&A detection and methods to evaluate the effectiveness of MC&A activities explicitly within the existing path analysis methodology.