About this Abstract |
Meeting |
MS&T23: Materials Science & Technology
|
Symposium
|
Materials Processing and Fundamental Understanding Based on Machine Learning and Data Informatics
|
Presentation Title |
Machine Learning-assisted Exploration of the Chemistry-processing Design Space Under Additive Manufacturing: Application to an FCC HEA Space |
Author(s) |
Raymundo Arroyave |
On-Site Speaker (Planned) |
Raymundo Arroyave |
Abstract Scope |
In this talk, we present recent work by our group in which we employ a synergistic combination of efficient physics-based models, exhaustive hand-curated databases, and state-of-the-art machine learning (ML) approaches to explore vast alloy spaces. Our focus is on High Entropy Alloys (HEAs), and we present a novel intrinsic metric for printability that we believe can be used as an alloy indicator when designing for performance and processability. We have verified our approach by comparing our predictions with available experimental data extracted from an in-house database containing hundreds of features for thousands of alloy-processing combinations. The framework has then been applied to estimate the printability of thousands of alloys within the 'Cantor' HEA system. Our work demonstrates that our methodology can significantly reduce the time and cost associated with the development of new HEAs for AM, while providing a unique metric for the intrinsic printability of an alloy. |