Topology optimization of linkage mechanisms using spring-connected link model

Quang Dat Tran,Gang-Won Jang
DOI: https://doi.org/10.1007/s00158-024-03857-0
IF: 4.279
2024-08-18
Structural and Multidisciplinary Optimization
Abstract:A novel optimization-based approach for automated synthesis of linkage mechanisms is introduced. The method involves using moving morphable links (MMLs) as design components in topology optimization. These links float in the design space, connecting to create a linkage mechanism with a target output path for a given input motion. The design change of each link is represented by four design variables, corresponding to the positions of its endpoints. This results in a significant decrease in the number of design variables when compared to a model employing spring-connected rigid blocks (SBM), which is one of the most effective design models for topology optimization of linkage mechanisms. Additionally, the use of MML allows for the synthesis of mechanism layouts with cross-links, which cannot be achieved using the SBM. The proposed approach's efficiency is further enhanced due to the use of global optimizers to solve the optimization problem. For this investigation, the Bayesian optimization method is employed. To improve the effectiveness of the global optimizer by finding optimal hyperparameter settings, the study of the design space's nonlinearity is conducted. The proposed method's validity is demonstrated by successfully solving synthesis problems involving four-bar mechanisms and six-bar mechanisms including cross-link layouts.
mechanics,computer science, interdisciplinary applications,engineering, multidisciplinary
What problem does this paper attempt to address?