Reliability Modeling and Estimation of Electro-mechanical Systems under Alternating Loads Based on Current Signal
Reliability estimation and failure prediction are important for the security and stability of electromechanical systems.Using motor current signature analysis to model and estimate the reliability of mechanical components of electromechanical systems can minimize the damage to the system and save cost.However,most of existing studies focus on the systems under constant load or single operating condition,while some systems with multi-stage cyclic tasks and varying loads,such as industrial robots,suffer from distinguishing loads at different stages of specific tasks,(For example,the grinding task includes taking,grinding,refueling and other stages),resulting in the diversity of amplitude of the motor current in a single task,and it appears in cycles with all the tasks during service.As a result,the traditional reliability estimation method with a single failure threshold is no longer applicable.For electromechanical systems under alternating load conditions,a reliability modeling and evaluation method based on current signals is proposed.Considering the differences in current amplitude caused by varying loads in different task stages,a unified performance degradation model for each task stage is constructed,and an analytical expression of reliability function is derived by combining the failure thresholds corresponding to different stages.An improved two-step likelihood method is proposed for parameter estimation of single or multiple samples.The proposed method is verified by the experiment of alternating load of harmonic reducer of industrial robot.The results show that the reliability estimation results based on the proposed model are more accurate and reasonable,and the degradation model with a single failure threshold may overestimate or underestimate the reliability.
万方数据
维普
