About this Abstract |
| Meeting |
2026 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2026)
|
| Symposium
|
2026 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2026)
|
| Presentation Title |
Machine Learning-Based Predictive Modeling for Geometric Distortion in Metal Material Extrusion |
| Author(s) |
Saisruthi Kotagiri, Andrea Gonzalez Martinez, Jesus Diaz, Chaitanya Mahajan, Venkata Sirimuvva Chirala |
| On-Site Speaker (Planned) |
Saisruthi Kotagiri |
| Abstract Scope |
Additive Manufacturing of metals via Material Extrusion (MEX) enables cost-effective production of complex geometries; however, post-processing steps, debinding and sintering,cause significant non-linear shrinkage and distortion, complicating the expected final-shape. This research introduces a machine learning method to predict the final sintered component from its initial Computer Aided Design (CAD) from laser-scanned meshes. Fabricated benchmark partsare printed, sintered, and scanned into meshes to create paired datasets for an artificial neural network which will be trained to model shrinkage and distortion, allowing accurate prediction of final part shapes. This approach aims to provide compensation during the design phase, enabling designers to adjust CAD models for manufacturing-induced changes. Ultimately, the method seeks to improve dimensional accuracy in metal MEX, reducing trial-and-error while achieving tolerances comparable to traditional manufacturing—all within a concise, data-driven workflow. |
| Proceedings Inclusion? |
Undecided |