Predicting Additive Manufacturing Build Cracks Using XFEM

This paper was produced for the 2019 NAFEMS World Congress in Quebec Canada

Resource Abstract

Laser powder bed fusion additive manufacturing (LPBF-AM) technology is nowadays being used for complex, near net shape in service metal parts in various industries like aerospace, defense and automotive. While manufacturing such parts, users strive to choose the best print orientation considering build stability and minimizing support structure. Although support structures provide stiffness to the build and facilitate heat transfer from part to build plate, adapting support structures reduces the benefits of LPBF-AM in that it increases build time, manufacturing cost and can adversely affect the surface finish of a part. The biggest concern amongst all is that residual stresses are shown to fracture support structures leading to further exaggerated part distortion. It is therefore essential to understand how support strategy affects the build quality and design intelligent supports with added values such as heat sinks, fixtures and datums instead of thoughtless secondary supports.



In this paper, a first stage high pressure turbine is built using a hybrid support structure including solid and column supports. Column supports are designed with a reduced cross-sectional area when they come into contact with overhanging part regions for easy removal during post processing and to reduce stress build up. The LPBF-AM printer is used to print this turbine blade with a designed support strategy and the Finite Element Method (FEM) is used to simulate the print process to check distortion, structural integrity and residual stress. The Extended Finite Element Method (XFEM) which can simulate the crack initiation and propagation along an arbitrary, solution dependent path has been enhanced to support Additive Manufacturing process simulation and is used to model potential build cracks during the LPBF-AM process. First order tetrahedron elements are enriched to model the turbine blade and first order brick elements are enriched to model the support structures. A stress based criterion is used to determine crack initiation, followed by crack evolution governed by critical fracture energy, leading to eventual failure in the enriched elements. Build cracks are observed at the junction between the turbine blade and its column supports both in physical print and in simulation. The results show that finite element simulations effectively capture the physics involved with the LPBF-AM process and accurately predict not only distortion and residual stresses but also support structure crack mechanisms and part separation from support during print. The simulation result is then used to guide the development of improved support strategy.