Lessons Learned from First Generation Nuclear Plant Probabilistic Risk Assessments

Abstract

Direct experience with some 10 probabilistic risk assessments (PRAs) has provided numerous insights into nuclear plant risk and risk analysis. Much has been learned about the effectiveness of systems safeguards, and procedures. Risk levels are much more in focus and contributors to risk have been quantified. The risk levels have been found to be extremely small in terms of health effects. On the other hand, in some instances, core damage frequencies are only marginally low. Risk analysis appears to be the most attractive tool for defining options to control and, if appropriate, reduce core damage frequencies. The extremely small public risk is very much less than the utility risk.

Risk studies have indicated the importance of treating external events, including earthquakes, fires, floods, and storms. In many cases, they are the major contributors to health effects risk. Support systems to key safety systems have been identified as important contributors to system unreliability. Corrective measures are apparent. Containment systems are very effective in preventing major releases following accidents that result in damaged cores.

Risk analysis methods, while still evolving, have matured greatly in recent years. No new major breakthroughs in methods are believed necessary for PRA to assume an important position as a technical discipline. The new advances are expected to derive from better understanding of physical processes and broader application and use of the results from PRA. The PRAs can become the key input to meaningful decision analysis to enhance the management of risk.