Probabilistic Fracture Mechanics Analysis for Leak-Before-Break Evaluation of Light Water Reactor's Piping.

This paper describes Probabilistic Fracture Mechanics (PFM) analyses for quantitative evaluation of the likelihood of Leak-Before-Break (LBB) of Light Water Reactor's (LWR's) piping. The PFM analyses in general assume probabilistic distributions of initial crack size, applied stress cycles, crack growth laws, fracture criteria, leakage detection capability, defect inspection capability and so on. Referring to the deterministic procedure for LBB evaluation, most appropriate PFM models and data for LBB evaluation are discussed. Here the LBB index is newly proposed in order to quantitatively evaluate the likelihood of LBB. Through intensive sensitivity analyses, it is clarified that the LBB is more likely to occur for larger diameter pipe; the performance of leakage detection significantly affects the LBB likelihood; the LBB likelihood increases with plant's aging even conservatively assuming leak detection capability; the R6 method (Category 1, Option 1) for fracture criterion gives very conservative results; and In-Service Inspection (ISI) reduces the increase rate of failure probability than the failure probability itself.