Scatter in fatigue crack growth behavior of a Ni-base superalloy at high temperature.

Highlights • There is a certain degree of scatter in the crack growth behavior of GH420Li. • The crack growth rates of GH4720Li under identical Δ K fit lognormal distribution. • A stochastic FCG model based on a lognormally distributed variable is derived. Abstract In order to conduct structural integrity assessment based on probabilistic fracture mechanics, the scatter in fatigue crack growth behavior of materials should be characterized accurately. In this work, ten groups of superalloy crack growth tests under the identical condition are conducted. Based on the statistic results of experimental data, there is a considerable degree of scatter in the fatigue crack growth behavior of GH420Li. The crack growth rates (da/dN) under the identical Δ K fit lognormal distribution. According to the random number generation principle of lognormally distributed variable, a stochastic crack growth model depending on the mean value and the coefficient of variation (COV) is derived. As the mean value and COV of ln(da/dN) can be expressed as functions of Δ K , the stochastic crack growth rate is determined by Δ K and a standard normally distributed random variable. At last, the stochastic model is adopted to calculate the fatigue crack growth of CT specimen. The comparison of the predicted results with the experimental results proves the effectiveness of the proposed model. [ABSTRACT FROM AUTHOR]