航空发动机神经网络自适应控制研究

本文研究神经网络自适应控制方法及其在航空发动机控制中应用。结合某型航空涡喷发动机 ,首先研究采用神经网络进行非线性动态系统辨识 ,包括神经网络模型辨识的格式、输入信号形式等问题。然后 ,提出了一种神经网络自适应控制方法 ,阐明了该方法基本结构、原理。最后 ,在选定的设计点处进行发动机控制系统设计 ,当偏离设计点时 ,利用神经网络很强的学习、适应能力 ,通过在线修正神经网络参数 ,使控制系统仍保持良好性能。

A New Neural Network Adaptive Control Scheme Research

A new neural network adaptive control scheme is presented and its applications to aero-engine control system are researched.In accordance with a turbojet engine,the identification problem of nonlinear dynamical system is studied by using a neural network firstly,including the pattern of identification and the type of input signal.Then a new neural network adaptive control scheme is presented,which is composed of two neural networks.One of them is used as identifier and the other as controller.Finally the aero-engine control system is designed at the design point.When the work of the aero-engine deviates from the design point,the parameters of the identifier and the controller are modified with the help of the learning and adaptive capabilitles of neural networks so that the control system always maintains fine behaviors.