Layer-dependent inherent strain for improved distortion prediction in directed energy deposition (DED) additive manufacturing

Abstract

Optimization for large industrial parts manufactured with directed energy deposition using a thermomechanical simulation is time-intensive, due to the complex thermomechanical interactions and high deposition rates necessitating small time steps and element sizes. This work develops a layer-dependent inherent strain method using the minimum thermal strain extracted from a thermomechanical simulation for distortion prediction, employing the thermal shrinkage as the predominant driver of distortion in welding processes. The method has been tested on a wall with 40 mm height and a cylinder, achieving 94.1% accuracy, improving predictions by 15% compared to simulations using a constant inherent strain tensor. The approach enables fast distortion prediction, while the thermal strain correlation enables future parameter estimations without prior thermomechanical simulation.