Proportional Topology Optimization algorithm with virtual elements for multi-material problems considering mass and cost constraints

Abstract

This paper presents an extension of the Proportional Topology Optimization (PTO) with virtual elements for multi-material problems with mass and cost constraints. In particular, the linear virtual element method (VEM) is constructed on unstructured polygonal meshes. The linear VEM is desirable in the sense that numerical integration is not explicitly required, significantly reducing the computational effort. Furthermore, the unstructured polygonal mesh naturally eliminates the issue of one-node connections encountered by the usual quadrilateral mesh. A feature of PTO is that it does not require sensitivity information, i.e., the derivative of the objective function with respect to design variables. Instead, the amount of material distributed into each element is determined proportionally to the contribution of that element to the objective function. For multi-material problems, the Ordered Solid Isotropic Material with Penalization (Ordered SIMP) technique is integrated into the PTO framework. Compared to other techniques for problems that involve multiple materials, Ordered SIMP has the advantage that computational cost does not depend on the number of materials. Furthermore, for the first time, the PTO approach is extended to consider two types of constraints: mass and cost simultaneously. The feasibility and efficiency of the proposed method are demonstrated via several benchmark examples and comparisons with the existing approach.