基于Kohonen神经网络SVPWM的PMSM控制系统仿真研究  

Simulation Research on PMSM Control System Based on SVPWM with Kohonen Neural Network

作　　者：蒋正友 

机构地区：[1]上海大郡动力控制技术有限公司

出　　处：《变频器世界》2018年第1期70-72,共3页The World of Inverters

摘　　要：永磁同步电动机(PMSM)是一个多变量、非线性、强耦合的复杂系统,其伺服系统的控制策略直接影响PMSM的性能指标。文章给出了PMSM在d-q坐标系下的数学模型,分析了空间矢量脉宽调制(SVPWM)的基本原理,针对传统的复杂SVPWM算法,分析了一种基于Kohonen神经网络的SVPWM算法。该算法计算简单,避免了大量的三角函数和求根运算,从而可节省处理器的计算时间。采用的控制方式,在Matlab/Simulink环境下建立PMSM伺服系统三闭环控制仿真模型,仿真结果验证了算法的有效性。Permanent magnet synchronous motor （PMSM） is a multivariable nonlinear system with strong coupling. The control method of its servo system affects performance of PMSM. In This paper, the mathematical model of PMSM in d-q coordinate is built. The basic principle of SVPWM and SVPWM based on Kohonen Neural Network are analyzed.The algorithm is simple for trigonometric function and roots computing, which greatly saves the processor computing time. The three loops controlled system simulation model was established, based on the principle of vector control, with control manner, in Matlab/Simulink environment, the results verify the validity of the algorithm.

关 键 词：永磁同步电机 矢量控制 空间矢量脉宽调制 KOHONEN神经网络 

分 类 号：TM341[电气工程—电机]