基于梯度下降法的异步电机离线参数辨识方法  

Off-line Parameter Identification Method of Induction Motor Based on Gradient Descent Method

作　　者：孙国栋 张凌云 张珍睿 李庆卓 范宏伟 SUN Guodong;ZHANG Lingyun;ZHANG Zhenrui;LI Qingzhuo;FAN Hongwei(Academy of Science and Technology Information Innovation Strategy Research Center Shandong Jinan 250101,China;Qilu University of Technology(Shandong Academy of Sciences)Jining Shandong 272073,China)

机构地区：[1]山东省科学技术情报研究院创新战略研究中心,济南250101 [2]齐鲁工业大学(山东省科学院),山东济宁272073

出　　处：《微电机》2023年第11期66-73,共8页Micromotors

基　　金：齐鲁工业大学(山东省科学院)德州中试项目(2022DZZS02);齐鲁工业大学(山东省科学院)科教产融合创新试点工程(2022GH016)项目。

摘　　要：针对异步推进电机辨识过程中电机电压和电流幅值和相位计算不准确而导致的电机参数辨识精度低的问题,提出了一种基于梯度下降法的异步电机参数辨识算法;通过向定子电阻中通入不同电压,得到能够辨识电机的定子电阻、转子电阻以及互感的相关电机采样数据。通过三次函数拟合的方式对逆变器的理论电压进行修正以提高转子电阻的辨识精度;通过梯度下降法对电机交流电压和交流电流的波形进行数值分析,得到准确的正弦数据幅值和相位并求解得到功率因数,进而求得更加准确的电机转子电阻、漏感以及互感。在基于TMS320F28379D的自主研发变频器中对一台7.5 kW的电机进行了辨识实验,证明了该离线辨识算法的辨识精度。In order to address the problem of inaccurate identification of electric motor parameters resulting from the calculation of incorrect magnitude and phase of motor voltage and current during the identification process of the asynchronous propulsion motor,a gradient descent-based asynchronous motor parameter identification algorithm was proposed.By passing different voltages into the stator resistance,relevant motor sampling data that can be used to identify the stator resistance,rotor resistance,and mutual inductance of the motor are obtained.The theoretical voltage of the inverter was modified by a third-order curve fitting method,ultimately increasing the accuracy of the rotor resistance identification.By numerically fitting the waveforms of motor AC voltage and current using the gradient descent method,accurate waveform magnitude and phase were estimated,and the power factor was calculated,allowing the precise identification of motor rotor resistance,leakage inductance,and mutual inductance.An identification experiment was conducted on a 7.5 kW motor using a self-developed inverter based on TMS320F28379D,which demonstrated the effectiveness of the proposed algorithm in improving identification accuracy.

关 键 词：异步电机 离线参数辨识算法 梯度下降法 

分 类 号：TM343[电气工程—电机] TP273[自动化与计算机技术—检测技术与自动化装置]