On-Line Residual Compensation in Robust Fault Diagnosis of Dynamic Systems

Abstract This paper deals with the robust fault diagnosis problem for systems with unstructured uncertainties, for which disturbance de-coupling methods are not applicable. A technique for compensating the residual is proposed. By compensation, the effects of modelling errors on the residual will not adversely affect the detection and isolation of faults. The idea is to estimate approximately the bias term in the residuals due to modelling errors, then compensate it on-line. These estimates are used to form a compensated residual to decrease the effect of modelling errors on the residuals. The compensated residuals are then used to make the fault diagnosis decision. Simulation results illustrate the effect of incipient faults in sensors of a complex jet engine can be reliably detected with the method developed.