Statistical Analysis of the Early Phase of SBO Accident for PWR

Statistical Analysis of the Early Phase of SBO Accident for PWR

Kozmenkov, Y.; Jobst, M.; Kliem, S.; Schaefer, F.; Wilhelm, P.


A statistical approach is used to analyse the early phase of station blackout accident for generic German PWR with the best estimate system code ATHLET-CD as a computation tool. The analysis is mainly focused on the timescale uncertainties of the accident events which can be detected at the plant. The developed input deck allows variations of all input uncertainty parameters relevant to the case. The list of identified and quantified input uncertainties includes 30 parameters related to the simulated physical phenomena/processes. Time uncertainties of main events as well as the major contributors to these uncertainties are defined. A linear regression analysis is used for predicting times of future events from detected times of occurred/past events. The presented statistical approach could be helpful for assessing and improving existing or elaborating additional emergency operating procedures aimed to prevent severe damage of reactor core.

Keywords: Pressurized water reactor; best estimate simulation; station blackout; accident management measures; statistical approach; timescale uncertainties of events; sensitivity analysis; linear regression; prediction intervals