基于XGBoost的SRD功率变换器故障定位研究  

Research on Fault Location of SRD Power Converter based on XGBoost

作　　者：唐静 TANG Jing(School of Automation and Electrical Engineering,Chengdu Technological University,Chengdu 611730,China)

机构地区：[1]成都工业学院自动化与电气工程学院,成都611730

出　　处：《成都工业学院学报》2023年第5期43-48,共6页Journal of Chengdu Technological University

基　　金：四川省大学生创新创业训练计划(202111116003);成都工业学院科研项目(2021ZR003)。

摘　　要：为准确定位开关磁阻电机调速系统(SRD)功率变换器开关管故障相,设计一种基于极限梯度提升(XGBoost)算法的功率变换器故障定位方法。通过对功率变换器故障类型的分析,综合考虑ABC三相电流、ABC三相输出转矩、ABC三相位置信息、电机转速等影响因子,并在Matlab/Simulink环境下搭建开关磁阻电机调速系统,设置功率变换器故障,从而获取故障数据。其仿真结果表明,采用XGBoost算法训练分类器模型,可根据分类结果确定故障相和故障类型。并使用准确率、召回率、f1-score对模型进行评估,评估结果良好,验证了该方法的可行性。In order to accurately locate the fault phase of the switch tube of the switched reluctance driver(SRD)power converter,a fault location method for power converter based on the extreme gradient boosting(XGBoost)was proposed in this paper.Through the analysis of the fault types of power converter,and comprehensively considering influencing factors such as ABC three-phase current,ABC three-phase output torque,ABC three-phase position information and motor speed,the switching resistance motor speed regulation system was built in Matlab/Simulink environment,and the power converter fault was set to obtain the fault data.The simulation results show that using the XGBoost algorithm to train the classifier model,the fault phase and fault type can be determined according to the classification results,and the accuracy rate,recall rate and f1-score are used to evaluate the model,and the evaluation results are good.The good evaluation results verify the feasibility of the method.

关 键 词：开关磁阻电机调速系统 功率变换器 极限梯度提升 故障定位 

分 类 号：TM352[电气工程—电机]