The present study addresses the challenge of predicting porosity in laser-based powder bed fusion of AlSi10Mg by developing and validating a numerical two-dimensional multilayer simulation framework. Existing research lacks efficient and cost-effective methods for simulating the interplay of multiple process parameters and defect formation. To close this gap, simulations incorporating laser power, scan speed, hatch distance, layer thickness, and absorptance were conducted and compared to experimental data. Simulations and experiments showed good agreement (average deviation of 0.67 %) for samples with nearly fully dense material. However, stripe pore formation, a special form of lack of fusion, was initially underrepresented but successfully reproduced by adjusting the scan strategy in the simulations. A weighted combination of two simulations enabled replication of the experimental outcome of stripe pores with an overall good agreement between simulation and experiment (average deviation of 2.01 %) with minor outliers. Moreover, introducing a rescaled energy density enabled a unifying correlation between porosity and input energy across varying absorptance values highlighting the transferability of the model. Overall, the study demonstrates that computationally efficient 2D simulations can reliably predict porosity and support parameter optimization without extensive experimental effort, thereby contributing a validated numerical tool to accelerate process development.
Engelhardt et al. (Sun,) studied this question.