基于神经滑模观测器的永磁同步电机无传感器矢量控制  

PMSM sensorless vector control based on neural sliding mode observer

作　　者：苗敬利[1] 秦王毓 MIAO Jingli;QIN Wangyu(School of Information and Electrical Engineering,Hebei University of Engineering,Handan 056038,China)

机构地区：[1]河北工程大学信息与电气工程学院,河北邯郸056038

出　　处：《现代电子技术》2021年第11期110-114,共5页Modern Electronics Technique

基　　金：河北省自然科学基金重点项目(E2019402436)。

摘　　要：针对传统滑模观测器下永磁同步电机的强抖振问题,提出一种神经滑模观测器。通过分析同步旋转坐标系下锁相环转子位置估计系统的结构特点,将PI调节器和径向基函数神经网络相结合,对传统的转子位置估计方法进行改进,提高对转子位置和转子速度的追踪精度。同时,在使用饱和函数取代传统切换函数的基础上,采用指数趋近律对滑模观测器的结构进行改进。仿真结果表明,所提算法能够有效提高转子位置估算精度,削弱抖振,并且有较快的动态响应速度。In view of the strong chattering of permanent magnet synchronous motor(PMSM)under the traditional sliding mode observer,a neural sliding mode observer is designed.By analyzing the structural characteristics of the phase-locked loop rotor position estimation system in synchronous rotating coordinate system,the PI regulator and radial basis function(RBF)neural network are combined to improve the traditional rotor position estimation method,so as to improve the tracking accuracy for rotor position and rotor speed.On the basis of replacing the traditional switching function with saturation function,the structure of the sliding mode observer is improved by exponential reaching law.The simulation results show that the proposed algorithm can effectively improve the estimation accuracy of rotor position,weaken the chattering,and has faster dynamic response speed.

关 键 词：永磁同步电机 同步旋转坐标系 滑模观测器 径向基函数 神经网络 锁相环 指数趋近律 PI控制 矢量控制 

分 类 号：TN876-34[电子电信—信息与通信工程] TP2[自动化与计算机技术—检测技术与自动化装置]