永磁同步直线电机的小波神经网络控制  

Control for PMLSM based on wavelet neural network

作　　者：党选举[1] 徐小平[1] 于晓明[2] 姜辉[1] 

机构地区：[1]桂林电子科技大学电子工程与自动化学院,广西桂林541004 [2]陕西科技大学电气与信息工程学院,陕西西安710021

出　　处：《电机与控制学报》2013年第1期43-50,共8页Electric Machines and Control

基　　金：国家自然科学基金(60964001;61263013);广西自然科学基金重点项目(桂科自0991019Z)

摘　　要：针对永磁同步直线电机跟踪性能易受推力波动、摩擦力等干扰严重影响的问题,提出了基于小波神经网络的控制方法。分析证明了所构建的小波神经网络能以任意精度逼近主要干扰:非线性推力波动干扰。并且在复合前馈PID控制方法的基础上,利用小波神经网络实现了对永磁同步直线电机干扰的在线估计补偿。小波神经网络的控制实验的跟踪误差为0.15 mm,精确度为0.75%,实验结果表明,与复合前馈PID控制方法和神经网络自适应逆模型控制方法相比,基于小波神经网络的控制方法有效地提高了系统的跟踪性和鲁棒性,并能有效消除干扰对系统的影响。For the tracking performance of permanent magnet linear synchronous motor （ PMLSM ） serious- ly influenced by ripple force, friction force and other disturbances, a method of control based on wavelet neural network （WNN） is proposed. By theoretical analysis, it was proved that the selected wavelet func- tion approximated nonlinear ripple force with desired accuracy. On the basis of the combined feedforward plus PID control, a wavelet neural network was used to estimate the disturbances voltage. The experimen- tal tracking error of the control based on wavelet neural network is 0. 15 mm, and the precision is 0. 75 percent. The experimental results show the efficiency of the proposed method, compared with combined feedforward plus PID control and neural network adaptive inverse control. The proposed method can not only effectively improves tracking precision and robustness, but also preferably achieves effective disturb- ance compensation.

关 键 词：永磁同步直线电机 小波神经网络 干扰补偿 复合控制 

分 类 号：TM351[电气工程—电机]