Physics-Informed Features Improve Melt Pool Geometry Prediction in LPBF
November 20, 2026
Using normalized enthalpy and Peclet number as dimensionless features improves melt pool width, depth, and mode predictions in laser powder bed fusion. This method outperforms models using only raw process parameters, facilitating better cross-material transfer in additive manufacturing.
HOW THIS AFFECTS YOU
●
researcherYou can use physics-based dimensionless features to make your manufacturing ML models generalize better across different materials.