Moving Morphable Multi Components Introducing Intent of Designer in Topology Optimization

Abstract

Topology optimization based on moving morphable components efficiently generates a topology that is expressed by a few geometrical design variables. However, conventional moving morphable components have three problems: lack of [Formula: see text] continuity between components, difficulty in describing a smooth rollup shape, and difficulty in generating a rigid joint to an optimized topology. In this study, a novel topology optimization framework was developed by introducing theories devised for multibody analysis. First, an adaptive moving morphable component based on absolute nodal coordinate formulation was proposed. Because both the position and gradient are used as design variables, [Formula: see text] continuity is ensured. Second, a position and gradient connection algorithm leveraging the linear constraint of the absolute nodal coordinate formulation was proposed to describe the smooth rollup shape. Third, a rigid joint was generated by introducing the gradient constraint equation in an optimizer. The developed framework exhibited superior convergence as compared with the conventional one in the benchmark short beam problem. It successfully generated an optimal topology with the intent of a designer (that is, designer-selected topology continuity and rigid joints), which facilitated the assembly and manufacturing of topologically optimized structural members to construct an entire aerospace structure.

Moving morphable component (MMC) topology optimization with different void structure scaling factors

“…Based on the descriptions of areas, Yang and Huang [ 26 ] discussed the geometric descriptions of different components, and an extended form of the multi-section component was constructed to improve the deformation ability of components. To solve the shortage of C 1 continuity, Otsuka et al [ 27 ] proposed an adaptive moving morphable component (ANCF) component with two nodes based on absolute nodal coordinate formulation. Because both the position and gradient are used as design variables, C 1 continuity is ensured.…”