Investigation of approximation methods for mechanism reliability analysis considered nonlinear response

The joint clearance and contact phenomena are the key factors which can induce the nonlinear dynamic responses and affect the performance of the practical mechanism. It is difficult to solve the nonlinear relations between the input variables and responses when the random factors are considered. In this paper, three approximation models, i.e., Response Surface Model (RSM), Kriging model and Support Vector Machine (SVM), are selected to investigate the reliability analysis procedure by the combination of simulation technique and approximation model for the mechanism with joint clearance. The solution strategy with supplemental samples is proposed to solve the problem caused by the occurrence of the abnormal point in the iteration procedure, and to avoid the divergence in the reliability analysis caused by the nonlinear dynamic responses of the mechanism. Meanwhile, the computational consumption can be reduced as the results of the known samples are reused. Finally, the reliability analysis for a slider-crank mechanism with joint clearance is utilized to illustrate the proposed procedure.