Research On Aero-Engine Control System Sensor Fault Diagnosis And Fault Tolerant Control

In aero-engine control system, actuators are the bridge between controller and engine, and sensors as measuring elements are the basis of the control system operation. Whether the actuators and sensors signals are normal will directly affect the performance of the aero-engine control system. Therefore, in order to improve the reliability of control system, it is necessary to establish the actuator, gas path sensor fault diagnosis and fault-tolerant control systems.First of all, the actuator fault diagnosis technology is studied. Fault of actuator is diagnosed and located according to the actuator model and the inverse engine model. The actuator model is built based on offline training neural networks. The network parameters are trained according to the test data of a semi-physical simulation test bed. The inverse engine model is built respectively using the offline training neural network and online training least squares support vector regression. The training samples are selected using the threshold discrimination which can reduce the online training time and improve the real-time property of inverse model during the online training process. The simulation results on certain engine fuel system actuator show that the proposed system can diagnose and locate the faults of actuator or its sensor accurately in steady and dynamic conditions. The real-time property and adaptive capability are satisfied.Secondly, gas path sensor fault diagnosis technology based on online extreme learning machine(OS-ELM) is studied. In order to realize the fault diagnosis of double sensors, fault discrimination and reset function of diagnosis system are designed. The signals of fault sensor are removed from the inputs of prediction model when reset the fault diagnosis system, which improves the prediction ability. The simulation results show that the algorithm can effectively diagnose and isolate single sensor fault and some double sensor faults. Prediction accuracy and real-time performance are satisfied.Finally, on the foundation of sensor fault diagnosis module, an active fault tolerant control system is designed based on augmented LQR, whicn realizes the effective separation of fault sensors and avoids the failure of the controller through active switching between multi-group controllers.