Reliability of LBB Evaluation Considering Randomness in Design Parameters

Abstract

In US, definition of the Leak-Before-Break (LBB) approach and criteria for its use are provided in NUREG-1061. Volume 3 of NUREG-1061 defines LBB as “…the application of fracture mechanics technology to demonstrate that high energy fluid piping is very unlikely to experience double-ended ruptures or their equivalent as longitudinal or diagonal splits.” Current LBB evaluation uses a factor of safety of two (2) on critical flaw size and a factor of safety of ten (10) on detectable leakage to deterministically analyze, that for a given set of input those factors are achieved. Typical input for LBB evaluation consists of pipe geometry, material properties (both elastic and plastic), crack morphology, loads, and operating pressure and temperature. Since LBB has recently been applied for pipes with weld overlays (WOL), thickness, material properties, and crack morphology of WOL also becomes important. However, in real structure all the design parameters (input) for LBB evaluation are inherently random in nature. The current work includes randomness in the critical design input parameters for LBB evaluation. Based on the result of this study reliability (or its compliment, probability of failure) curves are obtained based on the randomness in the critical input parameters. A piping system is considered to fail the LBB evaluation if the actual leakage through the pipe is less than the required leak rate which is calculated as ten times the plant minimum leak detection capability. Separate reliability curves are obtained for various minimum plant leak detection capability piping (e.g.,…, 1, 0.5,…, 0.1 GPMs) and for various piping systems (large diameter pipes such as reactor coolant loop hot leg and cold leg; and small diameter pipes such as pressurizer surge line, etc.). The reliability curves give an insight into the likelihood for a deterministic design input based LBB evaluation to remain valid in view of the in-situ variations.