Prediction and validation of melt pool dimensions and geometric distortions of additively manufactured AlSi10Mg

Abstract

Abstract A finite-element (FE) based thermomechanical modeling approach is developed in this study to provide a prediction of the mesoscale melt-pool behavior and part-scale properties for AlSi10Mg alloy. On the mesoscale, the widely adopted Goldak heat source model is used to predict melt pool formed by laser during powder bed fusion process (PBF), which, however, requires the determination of certain parameters as they control temperature distribution and hence melt pool boundaries. An approach based on a systematic parametric study is proposed in the study to determine these parameters, such as absorption coefficient and transient temperature evolution compared with the morphology of melt pool from experiments. Focusing on the part-scale domain, there is increasing demand for predicting geometric distortions and analyzing underlying residual stresses, which are highly influenced by the mesh size and initial temperature (Tinitial) setup. This study aims to propose a strategy for the correlation between the mesh size and the initial temperature to provide correct residual stresses when scaling up the model for efficiency. Results revealed that the predicted melt pool (MP) error produced by optimal Goldak function parameters is between 5–12%. On the part scale, according to the findings, the FE model is less sensitive to mesh size for distortion prediction and layer-lumping can be used to speed up simulation. The effect of large time increments, and layer lumping can be compensated by appropriate Tinitial value for AlSi10Mg.