基于小波神经网络的直线电动机控制系统性能改善

摘要

In this paper,wavelet neural network was used in improving the direct thrust control performance of linear motor,in respect that wavelet neural network theory has the wavelet characteristic of the time frequency location characteristic,the neural network properties of self-learning,self-organizing and self-adapting,and the strong approximation and fault-tolerant capability.Wavelet neural network was used to identify the primary resistance,the flux linkage observation was improved by the primary resistance accurate identification,and the performance of linear motor control system was further improved.Form the simulation results,it can be concluded that the identification system has good impact in slot the primary resistance.The direct thrust control system with resistance identification can efficiently improve the control system performances,and provide a new design method to remedy the limitation of linear motor low-speed performances.%基于小波神经网络具有小波分析的时频局部性和神经网络的自学习、自适应、自组织的能力,以及良好的容错和逼近的功能,将其用于提高直线电动机的控制系统性能.为精确观测直线电动机的磁链,利用小波神经网络准确辨识初级绕组的阻值,改善直接推力控制系统的低速性能.仿真结果表明该辨识系统跟踪初级绕组阻值的效果良好,带阻值辨识的直接推力控制系统的性能得到了改善,同时提供一种新的设计思路来弥补直线电动机低速性能的缺陷.