Application of Grey Relational Approach and Artificial Neural Network to Optimise Design Parameters of Bridge-Type Compliant Mechanism Flexure Hinge

Abstract

The investigation proposed a hybrid Grey-artificial neural network to optimise the design parameters of a two degree of freedom (2-DOF) bridge-type compliant mechanism flexure hinge (BTCMFH). The design variables play a vital role in determining the deformation and stress of the mechanism. The investigation is different from the previous studies where the hybrid method is a combination of grey relational analysis and artificial neural network based on finite element method (FEM) in ANSYS to maximise output displacement (DI) and minimise the stress (ST) of the mechanism. The simulation and ANOVA results identified the design variables have significantly affected the output displacement and stress by their contribution. The grey relational analysis and artificial neural network predicted values are in agreement with the simulation results at optimal combination parameters with a deviation error displacement and stress being 0.57% and 2.1%, respectively. The optimal combination parameters with a deviation error of displacement and stress of 0.52% and 2.1%, respectively. The optimal values of DI and ST were obtained as 0.957 mm and 104.74 MPa, respectively. The optimal value of displacement amplification ratio gained is 95.7.