Voxel-based evolutionary topological optimization of connected structures for natural frequency optimization

Abstract

The topology optimization methodology is widely utilized in industrial engineering for designing lightweight and efficient components. In this framework, considering natural frequencies is crucial for adequately designing components and structures exposed to dynamic loads, as in aerospace or automotive applications. The scientific community has shown the efficiency of Bi-directional Evolutionary Structural Optimization (BESO), showcasing its ability to converge towards optimal solid-void or bi-material solutions for a wide range of frequency optimization problems in continuum structures. However, these methods show limits when the complexity of the domain volume increases; thus, they are well-suited for academic case studies but may fail when dealing with industrial applications that require more complex shapes. The connectivity of the structures resulting from the optimization also plays a fundamental role in choosing the best optimization approach, as some available commercial and open-source codes nowadays return unfeasible sparse structures. An improved voxel-based BESO algorithm has been developed in this work to cope with current limits in lightweight structure optimization. A significant case study has been developed to evaluate the performances of the new methodology and compare it with existing algorithms. In contrast to previous studies, the method we developed guarantees that the final structure respects constraints on the initial design volume and that the structure’s connection is preserved, thus enabling the manufacturing of the component with Additive Manufacturing technologies. The proposed approach can be complemented by smoothing algorithms to obtain a structure with externally appealing surfaces.