Active fault diagnosis of a switched reluctance motor using sliding mode observer and average torque estimator for light electric vehicle applications

Abstract

In this study, an active fault diagnosis of a switched reluctance motor (SRM) using a sliding mode observer (SMO) and an average torque estimator for light electric vehicle (LEV) is proposed. The residual based fault detection using the SMO is used to diagnose a fault in a position sensor. When the fault is diagnosed, the corrupted sensor outputs are detected and then substituted with the estimated speed and position signals for sensorless operation in the sensor fault-tolerance mode. Moreover, all possible stator winding faults are monitored and can be diagnosed using an average torque estimation as soon as the completion of one electric cycle. The parameter uncertainties in the motor model are considered and analyzed for robustness of the operation. Therefore, any deviation of the actual system parameters from the estimated parameters can be effectively detected and diagnosing a type of fault in a sensor or stator windings. Finally, to illustrate the applications of the proposed scheme, simulation studies of several fault scenarios are discussed and conducted for a 3.5 kW 6/4 three-phase SRM.