A new optimal design synthesis method for flexure-based mechanism: recent advance of metaheuristic-based artificial intelligence for precision micropositioning system

This article presents a new design synthesis method to solve the optimization problem for a flexure-based mechanism in precision micropositioning system. In the proposed design synthesis method, structural optimization, neural-fuzzy system, finite element method, metaheuristic algorithm, and statistical technique is hybridized. Firstly, the topology procedure is performed via combination of solid isotropic material with method and optimality criteria to produce a new flexure-based mechanism with lightweight and good compliance. The topological mechanism is then redesigned into a few variants of mechanisms, and then the best flexible mechanism is chosen. Secondly, nonlinear simulations are carried out the outcomes of the mechanism. Thirdly, the adaptive neuro fuzzy inference with three different GENFIS types of grid partition, subtractive clustering, and fuzzy c-means clustering is formulated to build the regression models for the cost function and constrain function. The results revealed that the GENFIS2 is an appropriate type for training the strain energy while the GENFIS3 type is the best for both the displacement and the fatigue life. The sensitivity of geometrical dimensions is investigated to determine their influential levels. Finally, the neural network algorithm is utilized to find the optimal dimensions of the mechanism. The optimal results for the mechanism found that the displacement is about 25.63 mm, the strain energy is 42.57 mJ, and the fatigue life is approximately 161,833.77 cycles. The discrepancies of the predicted results against the verification results for the displacement, the strain energy, and the fatigue life are 1.58%, 3.17%, and 1.14%, respectively. By performing the statistical tests, the proposed method outperforms the other metaheuristics. The achieved results provide a useful optimizer for related microsystems.