基于机器学习的风机发电机绕组温度故障诊断与预警分析  

Fault diagnosis and early warning analysis of wind turbine generator winding temperature based on machine learning

作　　者：付恩强 黎天双 杨佳林 杨金声 宋美 马亚杰 Fu Enqiang;Li Tianshuang;Yang Jiain;Yang Jinsheng;Song Mei;Ma Yajie(School of mathematics and statistical science,Ludong University,Yantai,Shandong 264000,China)

机构地区：[1]鲁东大学数学与统计科学学院,山东烟台264000

出　　处：《计算机时代》2023年第1期4-7,共4页Computer Era

基　　金：国家级大学生创新创业训练计划项目(202110451161);鲁东大学专创融合课程建设重点项目(202114)。

摘　　要：为减少风电机组因故障停机造成的经济效益损失,围绕风机发电机绕组温度的故障预警分析和故障原因诊断进行研究,通过对比XGBoost等多种机器学习模型预测效果,最终选用XBGoost算法建立故障预警模型对发电机绕组进行实时监测,利用MSE、R^(2)等多个指标评价XGBoost模型,结果显示其准确率良好(R^(2)=0.9949)。本文提出的模型预警系统可以实时监测风机发电机绕组温度变化趋势,提前发出预警,最大限度的减少风机因停机造成的损失。In order to reduce the loss of economic benefits caused by wind turbine shutdown, in this paper, the fault early warning analysis and fault cause diagnosis of wind turbine generator winding temperature are studied. By comparing the prediction effects of various machine learning models, the XBGoost algorithm is finally selected to establish a fault early warning model for realtime monitoring of generator windings. Several indicators such as MSE and R^(2) are used to evaluate the XGBoost model, and the results show that its accuracy is good(R^(2)=0.9949). The early warning system proposed in this paper can monitor the change trend of wind turbine generator winding temperature in real time, give early warning in advance, and minimize the loss caused by wind turbine shutdown.

关 键 词：XGBoost 风电机组 绕组温度 故障诊断 预警 

分 类 号：TM315[电气工程—电机]