A Review of Data-Driven Remaining Useful Life Prediction Methods for Closed-Loop Control System
Data-driven remaining useful life(RUL)prognostics technique has been the research frontier of reliability and automation,with the rapid development of advanced sensing and monitoring technology.It has been widely applied in system maintenance decision-making,which improves system operation safety,reliability and economy.The most reviews about data-driven RUL prediction methods mainly focus on component-level RUL prediction,but do not pay attention to the RUL prediction technology for closed-loop control system.The influence of controller and feedback control on system degradation and RUL pre-diction need to be considered when the data-driven method is explored to predict the RUL of closed-loop control system,which is different from component-level RUL prediction.Therefore,the development trend of data-driven RUL prediction methods for closed-loop control system and life extending control(LEC)methods is reviewed based on the predictive information.Meanwhile,the principles,characteris-tics and limitations are dissected about the methods based on Poisson process,Gamma process,Wiener process and hybrid model,respectively.Finally,the future research directions are discussed such as charac-terization of health status and construction of comprehensive health indicator,system RUL prediction with multi-degraded components,theory and verification application of LEC based on predictive information.
remaining useful life(RUL)predictionclosed-loop control systemdata-drivendegradation processlife extending control
国家科技期刊平台
NSTL
万方数据
维普
