Probabilistic analysis of the creep crack growth rate of type 316LN stainless steel by the Monte Carlo simulation

Abstract

This paper presents a probabilistic analysis for evaluating the creep crack growth rate (CCGR) of type 316LN stainless steel. The CCGR data was obtained from the creep crack growth tests, which are conducted under various applied loads at 600°C. The crack growth rate was characterized as a function of the C* fracture parameter. In order to logically obtain the B and q values in the CCGR equation of a = B(C*)q, three methods of the least square fitting method (LSFM), a mean value method (MVM), and a probabilistic distribution method (PDM) were adopted. Also, using the Monte Carlo simulation, a number of random variables was generated, and the CCGR lines were predicted probabilistically. The three methods did not show a large difference in the CCGR lines, but the PDM was most useful because the CCGR line can be evaluated with a probabilistic reliability. Both the B and q coefficients followed a lognormal distribution, even though the B ones were a little scattered for the points of the data. In the case of a standard deviation of 1 σ for the probability variables, P (B,q), the results of the MCS and the PDM for the distribution ranges of the CCGR lines were compared and discussed. Copyright © 2005 by ASTM International.