| Abstract Scope |
Powder bed fusion is a key process in metal additive manufacturing, where precise control of melt pool dynamics is essential for achieving desired microstructures and mechanical performance. The melt pool shape and evolution are governed by complex thermal fluid phenomena occurring during laser irradiation. Direct experimental observation of temperature, flow fields, and melt pool shape is extremely difficult due to optical inaccessibility and limited spatial and temporal resolution. Machine learning has emerged as a promising approach for predicting melt pool behavior; however, its effectiveness relies on large, high-quality datasets that are challenging to obtain experimentally. To address this issue, we develop a high-performance phase-field lattice Boltzmann model that enables high-fidelity, high-resolution simulations of melt pool dynamics. The proposed framework accurately reproduces melt pool shape, temperature distribution, and fluid flow, providing a valuable data source for training machine learning models and advancing predictive process design in metal additive manufacturing. |