Machine learning superalloy microchemistry and creep strength from physical descriptors

TL;DR

This paper introduces a novel element-agnostic modeling approach using Gaussian process regression to predict superalloy microchemistry and creep strength, enabling accurate predictions even for unseen elements.

Contribution

It develops a set of physical descriptors that are element-agnostic and a correction method to improve microchemistry predictions in multi-phase alloys.

Findings

Models accurately predict superalloy microchemistry, microstructure, and strength.

Good extrapolation performance for unseen elements in chemical space.

Element-agnostic descriptors outperform component-based methods.

Abstract

We propose an element-agnostic set of descriptors to model superalloy properties with Gaussian process regression. Furthermore, we develop a correction method to deliver the best and most physical predictions for microchemistry in multi-phase alloys. The models' performance in predictions is confirmed for superalloy microchemistry, microstructure, and strength properties. When including new, unseen elements in the test data, the models still give good predictions; such extrapolations into new chemical-space would be impossible with component-based descriptors.