Probabilistic assessment of creep crack growth rate for Gr. 91 steel

Highlights: → We use the three methods, a least square fitting method (LSFM), mean value method (MVM) and probabilistic distribution method (PDM), to assess creep crack growth rate (CCGR) for Gr. 91. → The PDM is found to be the most useful. → The data of both the B and q coefficients follow lognormal distribution well. A great number of random variables for the B and q are generated by Monte Carlo simulation (MCS) technique. → We predict the probability CCGR lines for 10% and 90% using the MCS results. - Abstract: This paper focuses on a probabilistic assessment of creep crack growth rate (CCGR) for Gr. 91 steel which is regarded as one of major structural materials of Gen-IV reactors. A series of creep creak growth (CCG) data was obtained from the CCG tests under various applied loads at 600 deg. C. Using the experimental CCG data, four methods such as a least square fitting method (LSFM), mean value method (MVM), probabilistic distribution method (PDM), and the Monte Carlo method (MCM) were used to determine the parameters B and q for a power law equation between CCGR and C* integral. The commonly used LSFM revealed a considerable difference in the CCGR lines compared with the MVM and PDM. The PDM was found to be more useful than the LSFM, because it can assess the CCGR lines from the probabilistic viewpoints. It was verified that the two parameters B and q followed a lognormal distribution well. From the lognormal distribution, a number of random variables for the B and q parameters were successfully generated by the Monte Carlo Simulation (MCS) technique. The CCGR lines for the 10% and 90% probabilities were predicted by the PDM and MCM, and the MCM result was compared with the PDM one.