Application

Abstract

The thermal fatigue damaging case from 1998, when a longitudinal crack was discovered at the outer edge of an elbow in a mixing zone of the Residual Heat Removal System (RHRS) of the Civaux NPP unit 1 has been chosen to verify the stochastic model from the book. The variability in statistical properties of material parameters is usually accounted by the statistical properties of Paris law parameters C and n. Also, the original crack depth of flaws has a certain probability density function, which is more related to probability of detection based on experimental in-service inspection (ISI) results. A lognormal probability density function of C scaling parameter and an exponential one for initial crack depth are used to provide a probabilistic input for solving the integral giving the crack depth as a function of time. The results of the stochastic approach to modeling of thermal fatigue crack growth in mixing tee, completed with probabilistic input to account variability in material characteristics, are given as the probability of failure as a function of the time reference period.