location: Current position: Home >> Scientific Research >> Paper Publications

Multiscale eigenfrequency optimization of multimaterial lattice structures based on the asymptotic homogenization method

Hits:

Indexed by:Journal Papers

Date of Publication:2020-03-01

Journal:STRUCTURAL AND MULTIDISCIPLINARY OPTIMIZATION

Included Journals:EI、SCIE

Volume:61

Issue:3

Page Number:983-998

ISSN No.:1615-147X

Key Words:Multiscale topology optimization; Multimaterial optimization; Lattice structure; Fundamental frequency; Asymptotic homogenization

Abstract:Ultralight lattice structures exhibit excellent mechanical performance and have been used widely. In structural design, the fundamental frequency is highly important. Therefore, a multiscale topology optimization method was utilized to optimize the fundamental frequency of multimaterial lattice structures in this study. Two types of optimization problems were studied, namely, maximizing the natural fundamental frequency with mass constraints and minimizing compliance with frequency constraints. The Heaviside-penalty-based discrete material optimization method was adopted for the optimal selection of candidate materials. The asymptotic homogenization method was used to evaluate the equivalent macroscale properties according to the microstructure of the lattice material. To enable gradient optimization, sensitivities were outlined in detail. A density filter with a volume-preserving Heaviside projection was used to eliminate the risk of a checkerboard pattern and reduce the number of gray elements. A polynomial penalization scheme was employed to eliminate localized spurious eigenmodes in the low-density region. Finally, several numerical examples were performed to validate the proposed method. These numerical examples resulted in novel microstructural configurations with remarkably improved vibration resistance.

Pre One:A new formulation of continuous transverse shear stress field for static and dynamic analysis of sandwich beams with soft core

Next One:An Efficient and Robust Mixed-Gauge Scheme Based on Current-Splitting for Nonlinear Magnetostatic Field Computation