基于RT-LAB的DFIG网侧变流器控制参数多目标分步辨识方法  

Multi-objective Step-by-step Identification Method of Control Parameters for DFIG Grid Side Converter Based on RT-LAB

作　　者：徐恒山 曾宪金 张旭军 李颜汝 薛飞 黄永章 李晨阳 XU Hengshan;ZENG Xianjin;ZHANG Xujun;LI Yanru;XUE Fei;HUANG Yongzhang;LI Chenyang(College of Electrical Engineering&New Energy,China Three Gorges University,Yichang 443002,Hubei Province,China;State Grid Gansu Electric Power Research Institute,Lanzhou 730070,Gansu Province,China;Electric Power Research Institute,State Grid Ningxia Electric Power Co.,Ltd.,Yinchuan 750001,Ningxia Hui Autonomous Region,China;State Key Laboratory for Alternate Electrical Power System With Renewable Energy Sources(North China Electric Power University),Changping District,Beijing 102206,China)

机构地区：[1]三峡大学电气与新能源学院,湖北省宜昌市443002 [2]国网甘肃省电力公司电力科学研究院,甘肃省兰州市730070 [3]国网宁夏电力有限公司电力科学研究院,宁夏回族自治区银川市750001 [4]新能源电力系统全国重点实验室(华北电力大学),北京市昌平区102206

出　　处：《电网技术》2025年第2期771-780,I0103,I0104,共12页Power System Technology

基　　金：宁夏自然科学基金项目(2023AAC03857);新能源电力系统全国重点实验室2024年开放课题(LAPS24006)。

摘　　要：为准确获取双馈风机(doubly fed induction generator,DFIG)网侧变流器低电压穿越(low voltage ride through,LVRT)控制参数以建立其精准仿真模型,采用随机森林算法(random forest,RF)和非支配排序遗传算法(non-dominated sorting genetic algorithms,NSGA)-II对影响DFIG低穿特性的LVRT控制参数进行高精度分步辨识。首先,利用RT-LAB平台通过硬件在环(hardware-in-loop,HIL)实验获得DFIG真实控制器的响应数据集;其次,利用RF算法分析控制参数与待选观测量的关联程度,筛选出高关联度观测量,并依据各LVRT控制参数灵敏度不同确定其分步辨识顺序;然后,建立多目标优化模型,利用NSGA-II算法和多目标决策方法求出最优参数解;最后,根据辨识参数结果建立DFIG仿真模型,并与HIL测试结果进行多案例对比,结果表明,所提辨识方法在20%~80%低电压穿越工况下具有良好的适应性,并能有效提高参数辨识精度。To accurately obtain the low voltage ride through(LVRT)control parameters of doubly fed induction generator(DFIG)’s grid side converter to establish the accurate simulation model,the RF-NSGA-II algorithm is proposed to identify the LVRT control parameters that affect the dynamic characteristics of DFIG in a high-precision step-by-step manner.Firstly,the response dataset of DFIG’s real controller is obtained by the Hardware-In-Loop(HIL)experiment with the RT-LAB platform.Secondly,the association degree between the control parameters and the observations was analyzed with the Random Forest(RF)algorithm,the high association degree observations were selected,and the step-by-step identifying order was determined based on the different sensitivity of each LVRT control parameter.Then,a multi-objective optimization model is established to find the optimal parameter solutions using the proposed NSGA-II algorithm and multi-objective decision-making method.Finally,the simulation model of DFIG is built according to the identified parameters,and the comparisons between the simulation and the HIL-tested results are conducted,the comparison results show that the proposed identification method has good adaptation and high identification accuracy under 20%~80%LVRT operating conditions.

关 键 词：随机森林-遗传算法 双馈风机 低电压穿越 控制参数 参数辨识 

分 类 号：TM614[电气工程—电力系统及自动化]