Faster, Smarter Residual Stress Prediction in LPBF Using a Modified Inherent Strain Framework

Residual stress remains one of the most persistent barriers to reliable Laser Powder Bed Fusion (LPBF), driving distortion, dimensional inaccuracy, and cracking in parts. This work presents an improved multiscale finite element framework for predicting residual stress in LPBF-fabricated Inconel 625 and Inconel 718 using a Modified Inherent Strain (MIS) approach. Unlike the conventional inherent strain method originally developed for welding, the proposed framework is tailored to LPBF by explicitly accounting for layerwise mechanical constraints and process-dependent thermo-mechanical interactions that govern stress accumulation during fabrication. The framework links mesoscale thermo-mechanical simulations, used to extract anisotropic inherent strain components, with macroscale mechanical analysis for part-scale residual stress prediction. Laser power was varied from 100 to 240 W and scanning speed from 750 to 1500 mm/s, and predictions were validated using X-ray diffraction measurements from the surface to a depth of 3 mm in LPBF Inconel 625 coupons. While the original MIS approach predicted surface residual stresses with errors below 14%, it could not accurately capture the subsurface stress state measured in Inconel 625. The improved MIS framework showed promising results in overcoming this limitation and was validated for both Inconel 625 and Inconel 718, enabling rapid process optimization in LPBF applications.