APPROACHES FOR ADDRESSING PARAMETRIC AND MODELING UNCERTAINTIES IN A REFERNCE PWR PRA

摘要

In this paper, approaches used in a reference PWR PRA for addressing parametric and modeling uncertainties were discussed. A challenge in performing parametric uncertainty analysis is to properly treat the state-of-knowledge correlations among basic event probabilities. An approach was developed and applied successfully for treating the state-of- knowledge correlations effectively in CAFTA and UNCERT codes environment. The basic strategy for reducing modeling uncertainties in the reference PWR PRA was to perform accident analysis as many as possible using MAAP code from the early stage of the PRA modeling and use the results and insights from such MAAP analyses in PRA modeling , especially in determining success criteria, event progresses, timings for operator actions, and timings for recoveries. In some cases, sensitivity studies were performed to address uncertainties. Insights from uncertainty analyses included a potentially significant under estimation of interfacing system loss of coolant accident risk if the-state-of-knowledge correlations are ignored, significant difference in plant responses to a different break sizes in a same loss of coolant accident category or steam generator tube rupture initiating event, and the significant impacts of steam generator tube condition on large early release frequency. Since steam generator tube condition affects large early release frequency significantly, it is needed to re-evaluate the steam generator tube condition during the future updates of the reference PWR PRA to reflect such impacts properly. Also, even though much efforts had been made beforehand to reduce modeling uncertainties, when it is required to evaluate the risk associated with a very specific case, like a loss of coolant accident with a known break size, it may be desirable to perform additional case specific accident analysis and PRA modeling in order to evaluate the associated risk more accurately and to support a proper risk informed decision making.