About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Additive Manufacturing Modeling, Simulation and Artificial Intelligence
|
| Presentation Title |
A Statistics-Informed Efficient Micromechanical Model for Additively Manufactured Ti-6Al-4V |
| Author(s) |
Lucas Prata Ferreira, Nolan Strauss, Prajwal Arunachala, George Weber, Somnath Ghosh |
| On-Site Speaker (Planned) |
Lucas Prata Ferreira |
| Abstract Scope |
This work proposes an efficient statistics-informed micromechanical model for additively manufactured Ti-6Al-4V. Traditional crystal plasticity models explicitly resolve the complex Widmanstätten microstructure but are prohibitively expensive for part-scale simulations. The proposed solution integrates image-based microstructure statistics with interpretable machine learning, using Genetic Programming Symbolic Regression (GPSR) to derive closed-form constitutive relationships from data. Statistical descriptors including the 2-point auto- and cross-correlations are used as model inputs, while variance prediction models capture uncertainty in material responses. This approach reduces mesh sensitivity and highly accelerates computational efficiency while maintaining high accuracy. Validation is performed using Statistically Equivalent Representative Volume Elements (SERVEs) and experimental data. This framework enables efficient and reliable part-scale qualification for demanding industrial applications. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Additive Manufacturing, Modeling and Simulation, Machine Learning |