Data Analytics and Machine Learning Toolkit to Accelerate Materials Design and Processing Development

The US Navy has identified refractory high entropy alloys (RHEAs) and metal additive manufacturing (AM) as enabling technologies to meet performance and sustainability targets for shipboard and aircraft systems. Key challenges include designing RHEAs and optimizing metal AM to achieve desired material properties for Navy propulsion applications. Developing materials and processes via traditional experimentation and process optimization techniques is slow due to the large number of variables in these systems. The application of machine learning (ML) techniques is envisioned to accelerate the overall process. The Phase I R&D has successfully demonstrated that, given sufficient data, ML models can be developed for correlating a set of descriptors with mechanical property and material phases. It was also shown that ML is applicable to developing relationships between process conditions and mechanical properties. In Phase II, we will further develop robust ML models for prediction of mechanical properties, classify phases, and derive design rules for screening materials, thereby accelerating their development. We will also generate and utilize laser powder bed fusion AM data to develop process-property correlations. We will develop and deliver a user-friendly tool-kit incorporating the ML models to facilitate material property predictions in support of RHEA and metal AM process development.