Probabilistic damage estimation in piping components against thermal creep and fatigue

Abstract

This study focusses on probabilistic damage estimation of piping components that carry high temperature fluids in nuclear power plants, when the fluid temperature has random temporal fluctuations. The damaging mechanisms considered are thermal creep and fatigue induced by the thermal stresses. First, a stochastic model for the random temporal fluctuations in the temperature of the fluid is fitted from measured data. The thermal multiaxial stresses generated due to the loadings are non-Gaussian random processes in time. Consequently, the creep and the fatigue damage induced in the structure are also random processes and at any specified time instant, the residual life can be quantified in a probabilistic sense. Monte Carlo simulations are carried out to investigate the random creep damage growth while an analytical approach is used to obtain estimates of the expected fatigue damage. Questions related to the accuracy of the estimates with random process models vis-a-vis random variable models for the temperature fluctuations are examined. Finally, estimates of the failure probability are obtained by considering some models available in the literature for the creep–fatigue interaction effects.