The Research On Robust Adaptive Control With Neural Network Compensation For Aeroengines

Advanced control strategy is one of the dominant directions in the current aero-engine control systems. A robust model reference adaptive control with error modification and neural network compensation for an aeroengines has been studied by digital simulation and hardware in loop experiment, which is hoped to be applied in engineering.In the adaptive control, a robust model reference adaptive control with error modification is proposed. The proposed control strategy is applied to design the rotor speed control loop and pressure ratio control loop of the aeroengine. Simulation results show the designed control algorithm has good real time characteristic and robustness in the full flight envelope and is effective to overcome the stability problem caused by outside disturbances.In the neural network control, a neural network with BP algorithm is proposed. Neural network is used to compensate the shortcomings of adaptive controller using the nonlinear approximation ability of neural networks. Parameters of the adaptive control algorithm and the weights of the neural network are tuned on line. System stability of the on-line tuning algorithm is guaranteed based on the Lyapunov stability theory. Simulation results demonstrate that the developed system can achieve favorable tracking performance in flight envelope.The above algorithm is applied to the rotor speed control system of aeroengine. Hardware in loop experiment is studied on the integrated flight/propulsion control simulation platform. The results show that the proposed control algorithm for aero-engine has a satisfactory tracking performance.