Fast and accurate laser powder bed fusion metamodels predicting melt pool dimensions, effective laser absorptivity and lack of fusion defects

Abstract

The significant computational expenses associated with simulating the Laser Powder Bed Fusion (LPBF) process often restrict the insights gained from modeling endeavors to specific combinations of process parameters, hindering broader conclusions. In this study, we employ a classical Design of Experiments approach on results obtained by multiphase Finite Element simulations. Utilizing this framework, we derive quadratic metamodels for the dimensions of the melt pool, enabling predictions of melt pool width, depth, and length across a wide spectrum of processing conditions. Notably, our findings indicate that as few as 25 simulations can suffice to predict melt pool dimensions in conduction mode LPBF across varying laser power, velocity, initial temperature, and spot size parameters. Among other insights, the metamodels uncover and quantify the substantial influence of initial temperature (the local temperature of the volume preceding the laser interaction). Additionally, rare insights regarding the melt pool sensitivity towards the laser spot size are provided. Furthermore, our investigation delves into laser interactions with different phases (powder, liquid, solid) across diverse processing conditions to establish a net global absorption coefficient. These analyses underscore that, under conventional process conditions, most of the incident laser intensity falls onto the liquid phase during conduction mode LPBF simulations of 316L stainless steel and Ti-6Al-4V. However, the laser spot size significantly affects the laser intensity interacting with the liquid phase, warranting consideration of laser spot size dependent absorptivity values in part-scale models. Lastly, employing straightforward geometric simulations, we derive a full processing map predicting the occurrence of Lack of Fusion defects, based on calculated melt pool dimensions and the associated scanning strategy.