Penalized anisotropy: Controlling anisotropy growth in concurrent optimization of topology and fiber orientation for orthotropic composite materials

Abstract

The topology and fiber orientation of fiber-reinforced composite structures must be designed computationally to fully demonstrate their anisotropic mechanical properties. Optimization of the topology and fiber orientation is performed either sequentially or concurrently. In the sequential method, topology optimization is performed first, followed by fiber orientation optimization. This separated optimization process does not consider material anisotropy during the topology optimization process. The concurrent method has the potential to realize better results by considering the anisotropic properties during topology optimization. However, the concurrent method often obtains locally optimal solutions because of material anisotropy in the initial optimization process. A continuation approach for anisotropy, which starts with weak anisotropy and then improves the anisotropy gradually, resolves this issue. This paper proposes a new concurrent optimization that can control the growth of anisotropy. The growth of anisotropy affects the topology, which is used to determine a better structure. Optimization results were obtained serially by varying the growth of anisotropy, which allowed a highly optimal topology and fiber orientation to be explored by considering their anisotropic nature.