Comparison of deep-learning and physics-aware surrogate models for melt-pool temperature prediction in laser welding

This dataset contains simulation output files for laser welding experiments under multiple process parameter settings. The files include temperature-related field data generated from flow3D for different combinations of laser power and welding speed. File names encode the corresponding process parameters, for example power in watts and speed in mm/s. The dataset is intended for research on thermal behavior, temperature prediction, reduced-order modeling, and data-driven methods for laser welding processes. Documentation and metadata will be completed before public release.