Reliability Modeling and Estimation for Electromechanical System Under Periodic Varying Loads

Using motor current signals for reliability modeling and estimation can minimize the damage to the mechanical transmission mechanism in electromechanical systems and save costs. In some cases, the transmission system suffered from varying loads, which may lead to a substantial difference in the amplitude of current signals across the entire life. For instance, joint reducers in industrial robots are subjected to varying loads at different stages in a single task, with the amplitude of the current have distinguishing differences; therefore, we cannot predefine one failure threshold for the reliability estimation of such systems. In this paper, we propose a periodic multistage task modeling approach based on the Wiener process, which considers varying loads and failure thresholds at different stages of a periodic task. A closed-form reliability function is derived when the stages are independent. The developed method is verified by a periodic varying load test on an industrial robot's harmonic reducer. Results show that the developed model is more accurate and reasonable for systems under periodic varying loads.