| With the development of technology, the control system is more and more complicated. So it becomes more and more importance, diagnosing accurately fault in time and ensuring that system runs reliably. And it is the basic of tolerant control system. To increase the reliability and the security of control system, the fault tolerant control theory is developed step by step, and gradually became a primary research subject in the modern control theory.Aeroengines are complicated, non-linear and multi-variable control systems. Neural networks are good candidates for performing a variety of reflexive functions in aeroengine control system because they are nonlinear, can be fault-tolerant themselves, and can learn from experience. So the neural network is used as system controller in this paper. And we detected and isolated sensor failures with Kalman filtering residual and WSSR(Weighted Sun Squared Residual). We designed a fault tolerant controller for turbofan aeroengine with output reconstruction method.In the condition of feedback sensor failure, the fault tolerant control is also deeply researched in this paper. When feedback sensor is in failure, the system could do the real-time estimation through the detected data of the fault sensors and feedback the signals, thus fault tolerance is achieved.The simulation experiment has been done on a turbofan aeroengine. Simulation results indicate that the neural network controller has quick response and has good performances in the full envelope. Conclusion and proof were given which ensured that the tolerant control system is insensitive to the class of failures. Simulation results identified the validity of the method. |
