Reliability analysis of elastic link mechanism based on BP Neural Network

Abstract

The demand of kinematic accuracy is increasing for modern mechanical systems, while elastic deformation will lead to the decline in kinematic accuracy. This paper examines the application of finite element analysis (FEA) method to kinematic error analysis of elastic mechanisms based on Kineto-Elastodynamics model. The simulation of limit state equations of link mechanisms is performed with Back Propagation Neural Networks (BPNN), while the reliability analysis is carried out with the Monte Carlo simulation (MCS) method for solving the reliability, of the elastic mechanism. The numerical example shows that the proposed analysis method is feasible and effective in improving the kinematic accuracy and can be applied to reliability analysis of other mechanisms in practical engineering.