基于自联想核回归算法的数据驱动电机控制  

Data Driven Motor Control Based on Auto Associative Kernel Regression Algorithm

作　　者：黎卫国[1] 邓渊 张长虹[1] 李明洋 杨旭[1] 徐航 LI Weiguo;DENG Yuan;ZHANG Changhong;LI Mingyang;YANG Xu;XU Hang(China Southern Power Grid Co.,LTD.,Ultra High Voltage Transmission Company Electric Power Research Institute,Guangzhou 510000,China;Pinggao Group Co.,LTD.,Pingdingshan Henan 467001,China;Department of Electrical Engineering,Tsinghua University,Beijing 100084,China)

机构地区：[1]中国南方电网有限责任公司超高压输电公司电力科研院,广州510000 [2]河南平高电气股份有限公司,河南平顶山467001 [3]清华大学电机工程与应用电子技术系,北京100084

出　　处：《微电机》2024年第5期17-22,共6页Micromotors

摘　　要：在电机控制领域,基于模型驱动的控制方法已经广为研究,随着计算机运算能力的提升,给大量数据计算带来了可能性。为进一步提升电机控制性能,研究了一种基于自联想核回归法算法的控制补偿策略,基于运行过程中的大量数据来校正补偿电机控制过程,以提升电机控制性能,平抑运行过程中的波动。通过理论以及仿真分析,该方法能较好完成电机运行控制,加上补偿后的控制策略能减小电流波动以及转矩脉动,有一定的控制补偿效果。此控制算法同样适用于其他电机控制策略的优化过程,有一定的参考价值。In the field of motor control,the model-driven control method has been widely studied.With the improvement of computer computing power,it brings the possibility to calculate a large number of data.In order to further improve the motor control performance,a control compensation strategy based on Auto Associative Kernel Regression(AAKR)algorithm was studied to correct and compensate the motor control process based on a large number of data during operation,so as to improve the motor control performance,smooth out fluctuations during operation.Through theoretical and simulation analysis,the method can complete the motor operation control well,and the compensated control strategy can reduce the current fluctuation and torque ripple,and has a certain control compensation effect.This control algorithm is also applicable to the optimization process of other motor control strategies,and has a certain reference value.

关 键 词：自联想核回归 电机控制 补偿校正 数据驱动 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]