A practical approach to treating time dependencies between diesel generator failures and offsite power recovery in multi-unit PSA

Treating time dependencies between the failures of nuclear power plant (NPP) components and offsite power recovery has been regarded as an important issue to evaluate risk practically. For some US NPPs, such time dependencies have been estimated through multiple integrals, but this approach has a limitation for application to a multi-unit probabilistic safety assessment (PSA) for a large number of units due to long calculation times. This study proposes a practical approach to deal with time dependencies between diesel generator (DG) failures and offsite power recovery in multi-unit PSA based on the post-processing of minimal cut sets. First, the types of risk-significant minimal cut sets are defined according to DG failure modes. The DG mission times are divided into several time intervals, and the representative failure times of the DGs with the available offsite power recovery times are determined for each time interval. The cut set frequencies are then recalculated using the probabilities of updated offsite power (non-)recovery and time fractions. Case studies were conducted that demonstrate the applicability of the approach to single-unit and multi-unit PSA models, and validity was confirmed by comparing the recovery factors and core damage frequency between the proposed approach and the integral approach.