Sensor Fault Diagnosis of Maglev Train Based on Kalman Filter Group

Sensor fault diagnosis of maglev train is studied based on Kalman filtering theory. Usually, a single Kalman filter of a control system can only detect faults, but can not locate fault parts. Therefore, Kalman filter group is introduced in. Further more, a single sensor fault will cause the system matrix of a close loop feedback control system to change which is a disadvantage to fault location. In order to solve this problem, this paper treats the feedback signals as external inputs to make the close loop system be equivalent to an open loop system, and the feedback signals are introduced into the Kalman filter group at the same time. Then, a special fault location decision criterion for maglev suspension control system is proposed according to the statistic characteristics of the residual error signals of the Kalman filter group. At last, the fault diagnosis strategy presented is proved to be effective by simulation and experimental results.