A pragmatic part scale model for residual stress and distortion prediction in powder bed fusion

Abstract

Parts manufactured by laser powder bed fusion contain significant residual stress. This stress causes failures during the build process, distorts parts and limits in-service performance. A pragmatic finite element model of the build process is introduced here to predict residual stress in a computationally efficient manner. The part is divided into coarse sections which activate at the melting temperature in an order that imitates the build process. Temperature and stress in the part are calculated using a sequentially coupled thermomechanical analysis with temperature dependent material properties. The model is validated against two sets of experimental measurements: the first from a bridge component made from 316L stainless steel and the second from a cuboidal component made from Inconel 718. For the bridge component the simulated distortion is within 5% of the experimental measurement when modelled with a section height of 0.8 mm. This is 16 times larger than the 50 μm layer height in the experimental part. For the cuboid component the simulated distortion is within 10% of experimental measurement with a section height 10 times larger than the experiment layer height. These results show that simulation of every layer in the build process is not required to obtain accurate results, reducing computational effort and enabling the prediction of residual stress in larger components.