A Study based on a bank of Kalman Filters design for Gas Turbine Accessory Fault Detection

This paper proposed a novel method integrating system identification of experimental data from hydraulic actuator rig with Kalman filters for fault detection and isolation of a gas turbine accessory. Application of experimental data from hydraulic actuator rig for fault detection and isolation (FDI) of a gas turbine accessory which is of great interest to industry is proposed. The hydraulic actuator rig is designed to simulate different gas turbine accessory fault. The faults that typically occur in practice injected into real hardware subsystems to record the precise response. Through using this very realistic data the fault models are then established. Then these models are incorporated into the gas turbine engine simulation. In order to realize this, the transfer function based on the experimental data from the rig has been setup and integrated into the diagnostic algorithms for the aircraft gas turbine engine. Integration of actuator fault diagnostic algorithms was investigated and the injected faults can be detected successfully. The approach proposed in this paper can be used for online fault detection and isolation accurately, which is of practical significance for engine system.