About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Accelerated Discovery and Insertion of Next Generation Structural Materials
|
| Presentation Title |
Novel Approach to Rapid Material Characterization and Multi-Material Design Optimization |
| Author(s) |
Sergio Dos Santos E Lucato, Akshat Agha, Janet Davis, Alex Wagner, Derek Eidum, Saketh Sridhara, Krishnan Suresh |
| On-Site Speaker (Planned) |
Sergio Dos Santos E Lucato |
| Abstract Scope |
The Multi-Alloy Pixelated Structures (MAPS) program offers a novel pathway to design single, multi-material components replacing conventional assemblies of single-material parts with superior performance at reduced cost. The approach couples rapid acquisition of temperature-dependent, structural materials properties data to a highly efficient multi-material topology optimization framework. A unique test geometry has been developed to simultaneously test multiple tension and compression loads which are evaluated using full-field optical strain and temperature measurements. Novel machine learning (ML) algorithms use the resulting data to compute the corresponding full-field stress map providing a means to predict the properties of new alloys efficiently. A corresponding paradigm change in design optimization is enabled by a novel ML based toolset that expands upon conventional topology optimization through simultaneous materials selection. The MAPS optimizer uses a continuous latent space representation of the materials properties space to enable optimal selection of different materials for each sub-volume of the structure. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Mechanical Properties, Machine Learning, Modeling and Simulation |