About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
2026 Technical Division Student Poster Contest
|
| Presentation Title |
SPG-50: Multimodal Learning for Material Property Prediction in Additive Manufacturing |
| Author(s) |
Nafiz Imtiaz, Maitreyee Sharma Priyadarshini |
| On-Site Speaker (Planned) |
Nafiz Imtiaz |
| Abstract Scope |
Laser powder bed fusion is a type of additive manufacturing technique capable of producing complex materials whose properties depend on process parameters and microstructure. Accurately predicting these properties remains challenging due to the heterogeneous and high-dimensional nature of manufacturing data. Here, we propose a multimodal machine learning framework that integrates image-based microstructural information with processing parameters to predict mechanical properties of additively manufactured materials. Convolutional autoencoders are used to encode diverse spatial characteristics from microstructure images. These encoded features are then combined with laser power and scan speed data to construct a supervised predictive model. Our approach is evaluated for predicting two key material properties, hardness and yield strength. Results demonstrate that incorporating image-derived features with process parameters improves prediction accuracy compared to models relying solely on process data. This study highlights the potential of multimodal learning in determining process-structure-property relationships in additive manufacturing. |
| Proceedings Inclusion? |
Undecided |
| Keywords |
Additive Manufacturing, Machine Learning, Mechanical Properties |