About this Abstract |
Meeting |
2025 TMS Annual Meeting & Exhibition
|
Symposium
|
Algorithms Development in Materials Science and Engineering
|
Presentation Title |
Discovering High-Performance High Entropy Alloys: A Combined Genetic Algorithm and Machine Learning Approach |
Author(s) |
Caroline Binde Stoco, Guillaume Deffrennes, Yannick Champion, Francisco Gil Coury |
On-Site Speaker (Planned) |
Caroline Binde Stoco |
Abstract Scope |
A novel approach combining genetic algorithms (GA) and machine learning (ML) techniques has been developed with the objective of identifying high entropy alloys (HEAs) with optimized mechanical properties, including high ductility and significant strength. This study focuses on single-phase face-centered cubic (FCC) alloys, with the objective of reducing the stacking fault energy (SFE) while simultaneously enhancing the Hall-Petch constant (K) and the critical resolved shear stress for TWIP/TRIP effects in combination with hardening by grain refinement and solid solution, respectively. The properties were calculated using empirical equations and machine learning models trained on databases comprising CALPHAD predictions and DFT calculations. Through iterative generations, the algorithm effectively discovered innovative alloy compositions that met the predefined criteria. The two most promising compositions were produced and subjected to thorough characterization in order to ascertain the applicability of the methodology in identifying innovative alloys. |
Proceedings Inclusion? |
Planned: |
Keywords |
High-Entropy Alloys, Computational Materials Science & Engineering, Characterization |