基于小波神经网络的PMSM最小损耗反推控制方法  

Wavelet neural network based loss minimization backstepping control of PMSM

作　　者：郝娜[1] 詹志坤[2] HAO Na;ZHAN Zhikun(Electrical and Electronic Engineering Department,Shijiazhuang University of Applied Technology, Shijiazhuang,Hebei 050081,China;Key Lab of Industrial Computer Control Engineering of Hebei Province, Yanshan University,Qinhuangdao,Hebei 066004,China)

机构地区：[1]石家庄职业技术学院电气与电子工程系,河北石家庄050081 [2]燕山大学工业计算机控制工程河北省重点实验室,河北秦皇岛066004

出　　处：《燕山大学学报》2018年第5期451-457,共7页Journal of Yanshan University

基　　金：国家自然科学基金青年科学基金资助项目(61503322)

摘　　要：适应负载的时变特性和低损耗运行是永磁同步电机(PMSM)控制中遇到的两大难题,本文针对这些问题提出了一种基于小波神经网络的PMSM最小损耗反推控制方法,实现了负载特性时变情形下的PMSM最小损耗速度控制。首先,在建立PMSM损耗模型的基础上,针对负载转矩与转动惯量的同时变化情况,采用合适的小波神经网络对时变负载的转矩与转动惯量进行了统一辨识;然后,利用负载转矩与转动惯量的估计值代替实际值,设计了基于最小损耗的PMSM反推控制器,并从理论上证明了该控制器的稳定性;最后,利用MATLAB软件对控制方法的正确性和有效性进行了仿真验证。仿真结果表明,该方法对于时变负载特性具有较好的辨识效果,与反推控制和i_d=0的PI控制策略相比,该方法对于参考信号的跟踪更加快速和精确,可提高运行效率约4. 3%。Two issues encountered in permanent magnet synchronous motor(PMSM)control are adapting to the time varying characteristics of load and pursuing the low cost in running,which are independently solved currently.An approach of wavelet neural network(WNN)based loss minimization backstepping control is proposed and loss minimum speed control under the condition of time-varying load characteristics is fulfilled.Firstly,on the basis of construction of loss model for PMSM,a reasonable WNN is used to identify the torque and inertia of time-varying load unifiedly.Then,the actual values of load torque and inertia is replaced by the estimations and the backstepping controller based on loss minimum is designed.The stability of the designed controller is theoretically proved.Ultimately,the correctness and effectiveness of the controller is verified in MATLAB.The simulation results demonstrate that the control approach has a good identification effect to time-varying load.Compared with the backstepping control and PI control,this algorithm enables to track the reference signal more rapidly and precisely,and the running efficiency of the motor is increased by about 4.3%.

关 键 词：永磁同步电机 时变负载 小波神经网络 最小损耗 反推控制方法 

分 类 号：TM351[电气工程—电机]