Influence of parameter uncertainty upon complexity of creep behavior of P91 steel

With the Wilshire-Cano-Stewart (WCS) model and probability theory, this study dealt with the reliable prediction of creep rupture time for P91 heat-resistant steel. Firstly, the WCS model parameters were calibrated using creep test data of P91 steel, and the probability distribution functions (PDFs) of which were determined. Then, the sources of uncertainties were incorporated into the model through calibrated the PDFs by the Markov Chain Monte Carlo (MCMC) method with the Metropolis-Hastings (MH) algorithm. Consequently, the short-term and long-term creep rupture times of P91 steel were predicted using the probabilistic WCS model, respectively. The obtained results suggested that due to the complex uncertainties, the proposition of the probabilistic model offers an effective tool to predict the bounds of long-term creep rupture time for steels in particular.