Recent Case Studies and Advancements in Probabilistic Risk Assessment

Abstract

During the period from 1977 to 1984, Pickard, Lowe and Garrick, Inc., had the lead in preparing several full scope probabilistic risk assessments for electric utilities. Five of those studies are discussed from the point of view of advancements and lessons learned. The objective and trend of these studies is toward utilization of the risk models by the plant owners as risk management tools. Advancements that have been made are in presentation and documentation of the PRAs, generation of more understandable plant level information, and improvements in methodology to facilitate technology transfer. Specific areas of advancement are in the treatment of such issues as dependent failures, human interaction, and the uncertainty in the source term. Lessons learned cover a wide spectrum and include the importance of plant specific models for meaningful risk management, the role of external events in risk, the sensitivity of contributors to choice of risk index, and the very important finding that the public risk is extremely small. The future direction of PRA is to establish less dependence on experts for in-plant application. Computerizing the PRAs such that they can be accessed on line and interactively is the key.

A Bayesian paired comparison approach for relative accident probability assessment with covariate information

“…An accident is not a single event, but can be considered to be the culmination of a series of cascading events (see Garrick, 1984) starting with a triggering incident. In the maritime accident probability model in Merrick et al (2000) and Van Dorp et al (2001), triggering incidents have been further categorized as mechanical failures and human errors.…”

Risk analysis and port security: some contextual observations and considerations

“…This so-called "triplet" of questions (Kaplan and Garrick 1981;Garrick 1984) can be written as follows:…”

The Prince William Sound Risk Assessment

“…In performing a PRA, one identifies the series of events leading to an accident, estimates the probabilities of these events, and evaluates the consequences of the accident. Garrick (1984) noted that an accident is not a single event but the culmination of a series of events. A triggering incident is defined to be the immediate precursor of an accident.…”