The Treatment of ISI Uncertainty in Extremely Low Probability of Rupture
The U.S. Nuclear Regulatory Commission (NRC) in cooperation with the nuclear industry is constructing an improved probabilistic fracture model for piping systems that in the past have not been susceptible to known degradation processes that could lead to pipe rupture. Recent operating experience with primary water stress corrosion cracking (PWSCC) has challenged this prior position of leak-before-break and which has now become known as “extremely Low Probability of Rupture” (xLPR). This paper focuses on the xLPR model’s treatment of uncertainty for in-service inspection. In the xLPR model, uncertainty is classified as either aleatory or epistemic, and both types of uncertainty are described with probability distributions. Earlier PFM models included aleatory, but ignored epistemic, uncertainty, or attempted to deal with epistemic uncertainty by use of conservative bounds. Thus, inclusion of both types of uncertainty in xLPR should produce more realistic results than the earlier models. This work shows that by including epistemic uncertainty in the xLPR ISI module, there can be a significant effect on rupture probability; however, this depends upon the specific scenarios being studied. Some simple scenarios are presented to illustrate those where there is no effect and those having a significant effect on the probability of rupture.