DTC系统中定子磁通观测器的神经网络实现

这篇文章提出了一种基于人工神经网络 (ANN)的直接转矩控制(DTC)系统定子磁通观测器。为了使得定子磁通神经网络具有动态性能 ,引入了输入与输出的延迟量 ;为了拓宽控制范围 ,在用于训练的输出数据中叠加了随机扰动分量。利用Matlab/Simulink和神经网络工具箱对应用此模型的直接转矩控制系统进行了仿真 ,仿真结果表明建立的神经网络磁通观测器是正确的。

Implementation of Stator Flux Observer in DTC System with Neural Network

This paper presents an artificial-neural-network-based stator flux observer for the direct torque control (DTC) system. In order to attain a dynamic performance and expand the applicable range of the flux neural network, the delayed data of the input and output and the strategy of adding random disturbance to the delayed output are adopted. A computer simulation of direct torque control system with flux neural network is developed using Matlab/Simulink together with the Neural Network Toolbox. The simulated results demonstrate the feasibility of the stator flux model based on artificial neural network.