According to the demand of enterprises which produce UPS,a condition based maintenance(CBM)management system of UPS based on extended Kalman filter(EKF)-Markov is designed.Under the permission of users,the status data of online position and real-time operation of the equipment is visualized by using the geographic information system.Compared with the traditional post-maintenance scheme,the weighted method is used in data preprocessing to model the CBM of the data-driven collected information and reduce differences caused by different types of data.The EKF is used to eliminate the influence of noise on the sampling results,and the average error of state-of-charge(SOC)predicted using the algorithm is 0.434 3％.Combined with the Markov decision process,the UPS battery state is analyzed,the health management and CBM strategy in charge-change mode is implemented,and the maintenance time is reduced by 57.12％on average.Results show that compared with the traditional maintenance,the state prediction and health management system can improve the maintenance efficiency and accelerate the transformation from traditional planned maintenance to CBM mode.
关键词
不间断电源/状态维修/荷电状态预测/扩展卡尔曼滤波/马尔可夫决策/状态预测与健康管理
Key words
Uninterruptible power supply(UPS)/condition based maintenance(CBM)/state-of-charge(SOC)prediction/extended Kalman filter(EKF)/Markov decision/state prediction and health management
维普
万方数据