Additively manufactured metals are special in that they behave close to isotropic in elasticity and monotonic strength, however their fatigue behavior is anisotropic and the fatigue strength is degraded. Designs generated by topology optimization have commonly been used as inspiration rather than realizing the actual design, primarily due to restrictions of the manufacturing methods. However, with additive manufacturing, it is possible to manufacture the highly complex designs common in topology optimization. An increased focus is therefore placed on design for manufacture, and for this purpose a novel smoothing approach is developed. Assuming isotropic stiffness and monotonic strength, the topology optimization is formulated as an extension to existing fatigue constraint functions using density-based topology optimization. An improved formulation for the fatigue damage is proposed to achieve a good combination of accuracy and computational efficiency, which has caused problems in previously published literature. These approaches and methods are demonstrated by solving both 2D and 3D problems, and the designs are subsequently verified
using commercial finite element software.