Probabilistic life models for steel structures subject to Creep-Fatigue damage

Abstract

When metal structures are subjected to long-term cyclic loading at high temperature, simultaneous creep and fatigue damage may occur. In this paper probabilistic life models, described by hold times in tension and total strain range at elevated temperature have been derived based on the creeprupture behavior of 316FR austenitic stainless steel, which is one of the candidate structural materials for fast reactors and future Generation IV nuclear power plants operating at high temperatures. The parameters of the proposed creepfatigue model were estimated using a standard Bayesian regression approach. This approach has been performed using the WinBUGS software tool, which is an open source Bayesian analysis software tool that uses the Markov Chain Monte Carlo sampling method. The results have shown a reasonable fit between the experimental data and the proposed probabilistic creep-fatigue life assessment models. The models are useful for predicting expended life of the critical structures in advanced high temperature reactors when performing structural health management.