Estimation of the γ and γ' Lattice Parameters in Nickel-base Superalloys Using Neural Network Analysis

Abstract

The lattice constants of the γ and γ' phases of nickel base superalloys have been modelled using a neural network within a Bayesian framework. The analysis is based on datasets compiled from new experiments and the published literature, the parameters being expressed as a non-linear function of some eighteen variables which include the chemical composition and temperature. The analysis permits the estimation of error bars whose magnitude depends on their position in the input space. Of the many models possible, a "committee of models" is found to give the most reliable estimate. The method is demonstrated to be consistent with known metallurgical trends and has been applied towards the study of some experimental alloys.