交流变频调速系统的神经网络自适应控制

A Neural Network Adaptive Control for AC Motor Variable Frequency Speed Control System

针对交流变频调速系统模型未知情况，提出一种改进的神经网络自适应控制方案．该方法采用双神经网络结构，通过引入一个神经网络辨识器来逼近系统的动力学特性以代替真实系统，另一神经网络作为自适应控制器以改善控制性能．系统仿真实验表明，该控制方法具有很强的鲁棒性和抗干扰能力，特别对于系统结构不确定、参数时变或含非线性因素的交流调速系统是一种有效的控制策略．

In this paper, an adaptive control method based neural networks is presented for a set of AC motor Variable Frequency Speed Control System with model unknown. Nonlinear model is obtained by using a NN identifier instead of real system dynamic characteristics, another NN is used for adaptive controller to obtain high performance. Simulation results show that the proposed control method has strong robustness and ability of rejecting disturbance, especially it is effective for the nonlinear AC speed control system with undetermined or variable parameters.