The inherent strain method for laser powder bed fusion additive manufacturing (AM) allows for rapidly simulating process-induced distortion while still retaining the influence of thermomechanical physics. However, previous implementation of a single vector taken from a global average does not fully capture the known AM process results. This study aims to increase predicted accuracy in simulations by modifying the inherent strain method to better reflect experimentally defined print behavior.
This study analyzed the thermomechanical response of a small-scale AM part and applied the findings to a part-scale simulation, validated against experimental measurements of a benchmark geometry. This included the effects of the compressive core and the rapidly cooled tensile shell of an AM part and the differing in-plane and build-direction environments, captured with two strain vectors in the inherent strain method.
When compared to experimental X-ray diffraction strain measurements in the in-plane and build directions, simulation accuracy was improved from the traditional single-vector approach, reaching errors as low as 1% and lowering part median error from 82% to 17%.
This vector assignment approach better captures the process physics to improve AM simulation accuracy without sacrificing the lightweight computational cost advantage of inherent strain modeling. Part residual strain distribution was characterized, leading to a better understanding of post-print effects in AM parts.
