Fatigue design of Additive Manufacturing components through Topology Optimization: Comparison of methodologies based on the defect distribution and on the stress gradient

In this paper, the fatigue design of AM components is addressed through Topology Optimization. Two defect-driven methodologies, which model the fatigue response in the finite life range and account for the experimental scatter, are compared. The first approach is based on the defect distribution and the marginal P-S-N curves. The second method models the size-effect by accounting for the stress distribution within the loaded material. The two approaches are used to assess the maximum allowable first principal stress, which causes crack propagation. As the fatigue strength depends on the component volume or the stress distribution, an iterative procedure is necessary to determine the optimal design. The optimal topologies of a cantilever beam and of an engine rod are determined and compared, showing that the defect-distribution method is the most conservative and the stress-gradient method can also be reliably applied in the case of uniform stress distribution, as in the engine rod. Furthermore, the choice of the volume to retain for the calculation of the fatigue strength in the defect-distribution method is critical.