Topology and size optimization for a flexure hinge using an integration of SIMP, deep artificial neural network, and water cycle algorithm

Abstract

This paper develops a multistage optimization method for designing a new flexure hinge (FH). The proposed method is a combination of the topology optimization, the deep artificial neural network (DANN)-based modeling, and the water cycle algorithm-based size optimization. Firstly, solid isotropic material with penalization is employed to topologize the FH. Then, the topological FH is modified to transform into a compliant configuration. Finite element method is used to collect the output datasets of the hinge. Subsequently, the architectures of DANN are optimized to formulate the objective functions and constraints of the hinge. The results showed that the prediction accuracy of the developed DANN is better than that of the multivariate general linear model. Lastly, the geometrical sizes of the hinge are optimized by hybridizing the optimal DANN and the water cycle algorithm. The results found that the optimal solutions found from the proposed method are greater than those obtained from the other metaheuristic algorithms. Based on the results of Wilcoxon, Friedman, and Post-hoc tests, the proposed method outperforms the other methods. Besides, the results indicated that the performances of the FH are superior to the conventional hinges. The proposed optimization framework can be considered as a systematic design method for compliant mechanisms and related engineering areas.