Fatigue Life Estimation of 2(3HUS+S) Parallel Manipulator for Simulation of Hip Joint Motion

Abstract

Fatigue life for robotic components is the time or disturbance load times of the fatigue crack initiation to fatigue damage. Most of the loads on robotic components are random. The amplitude and interaction of the random load have an effect on the damage quantity, which will affect the accuracy of fatigue life estimation. In this paper, a non-homogeneous Poisson process is used to simulate the random load, and the parameters is calculated by a neural network method. The associated damage theory solves the inaccuracy of fatigue life estimation because of the load interaction, and random load tests of eight levels are employed to verify the rationality of the method. Finally, the fatigue life of the ball screws for a 2(3HUS+S) parallel manipulator is accurately and reliably estimated by using the non-homogeneous Poisson process and the associated damage theory. This study provides a new means to predict the fatigue life for the parallel hip joint simulator.