Fault detection and isolation of a hybrid synchronous machine using parity equations

Abstract

In this paper, the parity space approach, a modelbased fault detection and isolation scheme, is applied to a hybrid synchronous machine. The aim is to detect additive faults, i.e., faults in the actuators and sensors. The inconsistency between the actual behavior of the process and the expected behavior of the model is represented by means of residuals. For the isolation of faults, a limit checking criterion based on the energy content of a residual is employed. Furthermore, the load torque acting as an external disturbance is estimated by using a reduced-order observer. A simulation analysis shows the effectiveness of the parity space approach for the detection and the isolation of additive faults in a hybrid synchronous machine.