Multi-variable neural network adaptive control for aeroengine

A method of multi-variable neural network adaptive control based on dynamical recurrent neural network was put forward to control aeroengine with strong nonlinearity and time-varying uncertainty. In the method, nonlinear model of engine was real-time identified by dynamical recurrent neural network, and system sensitivity information was real-time feed back to neural network controller so that controller could exactly control the engine. Through simulation of some turbofan engine in the full flight envelope, the results show that the proposed method doesnpsilat depend on the aeroengine precise model, it can effectively realize the multi-variable adaptive control for aeroengine, and the controlled plant has good dynamic and static performances.