Investigation of Finite Element Analysis Techniques for Direct Energy Deposition Additive Manufacturing

Two of the biggest problems faced by additive manufacturing (AM) are high residual stresses and part distortion. Due to the high temperature gradients generated during the process, large residual stresses are generated. These stresses can cause a part to distort out of spec or worsen fatigue properties, creating points of failure in service. To determine the impact of process inputs on residual stresses, modeling can offer an alternative to expensive and time-consuming trial and error and measurement techniques. This thesis explores the two most common finite element analysis (FEA) approaches to modeling AM: the weakly-coupled thermo-mechanical method and the inherent strain method. There is a lack of literature on thermo-mechanical AM models at a large part scale (> .5 m build height), with most large-scale works only considering steady state, symmetrical, long geometries. As the scale of parts modeled increases, the typical resolution assumptions become computationally infeasible. This thesis aims to address the overall need to verify the impact of resolution on FEA results (temperature, residual stress, distortion) by comparing the outputs of the same analysis as mesh and increment size are varied. The other main component of the thesis is to investigate the ability to utilize digital image correlation (DIC) data to inform the computationally inexpensive inherent strain method of FEA, which could open up the possibility of rapid in-situ modeling techniques for distortion correction in the future. Final part displacement measured using DIC is compared to inherent strain method results and further recommendations for building the approach are made.