Bayesian neural network analysis of fatigue crack growth rate in nickel base superalloys

Abstract

The fatigue crack growth rate of nickel base superalloys has been modelled using a neural network model within a Bayesian framework. A 'committee' model was also introduced to increase the accuracy of the predictions. The rate was modelled as a function of some 51 variables, including stress intensity range AK, log ΔK, chemical composition, temperature, grain size, heat treatment, frequency, load waveform, atmosphere, R-ratio, the distinction between short crack growth and long crack growth, sample thickness and yield strength. The Bayesian method puts error bars on the predicted value of the rate and allows the significance of each individual factor to be estimated. In addition, it was possible to estimate the isolated effect of particular variables such as the grain size, which cannot in practice be varied independently. This demonstrates the ability of the method to investigate new phenomena in cases where the information cannot be accessed experimentally.