LAUSR

Abstract Process optimization has always been a crucial step for effective usage of metal additive manufacturing (AM) processes: it consists in establishing quantitative relations between final part's characteristics and process parameters to find their optimal combination and obtain a fully functional mechanical component. Experimental investigation techniques are usually employed for this purpose but they can be extremely expensive and time-consuming, especially when the output of the process depends on a large number of parameters, like for AM. Numerical simulation could represent an alternative solution: by reproducing the real process characteristics, a simulation could provide useful insights, allowing to evaluate the performance of the process for different parameter combinations without relying exclusively on expensive experimental campaigns. In this work, a finite element AM simulation based on the inherent strain (IS) method was developed and the prediction performance in terms of part's residual deformation was evaluated by comparing the numerical results with the measurements carried out on an experimental campaign. A new model calibration approach for prediction improvement was also implemented and it allowed to discover an unexpected behaviour of the model that strongly affects the validity of this method for AM simulation.