Development of optimal L-PBF process parameters using an accelerated discrete element simulation framework

Abstract

AbstractLaser Powder Bed Fusion (L-PBF) has immense potential for the production of complex, lightweight, and high-performance components. The traditional optimization of process parameters is costly and time-intensive, due to reliance on experimental approaches. Current numerical analyses often model single-line scans, while it is necessary to model multiple fully scanned layers to optimize for bulk material quality. Here, we introduce a novel approach utilizing discrete element simulations with a ray tracing-modeled laser heat source. Our approach significantly reduces the cost and time consumption compared to conventional optimization methods. GPU acceleration enables efficient simulation of multiple layers, resulting in parameters optimized for bulk material. In a case study, parameters were optimized for AlSi10Mg in just 5 days, a process that would have taken over 8 months without GPU acceleration. Experimental validation affirms the quality of the optimized process parameters, achieving an optical density of 99.91%. Graphical Abstract Optimization using the accelerated simulation yielded an optimized parameter set within 5 days. This resulted in a part with an optical density of 99.91%.