含降雨量修正的台风灾害下输电杆塔数据机理联合故障概率预测  

Data Mechanism Joint Failure Probability Prediction of Transmission Tower with Rainfall Correction Under Typhoon Disaster

作　　者：侯慧[1] 徐海峰 王少华 谷山强[3] 王振国 苏杰[3] HOU Hui;XU Haifeng;WANG Shaohua;GU Shanqiang;WANG Zhenguo;SU Jie(School of Automation,Wuhan University of Technology,Wuhan 430070,China;State Grid Zhejiang Electric Power Co.,Ltd.Research Institute,Hangzhou 310014,China;Wuhan NARI Co.,Ltd.,State Grid Electric Power Research Institute,Wuhan 430074,China)

机构地区：[1]武汉理工大学自动化学院,武汉430070 [2]国网浙江省电力有限公司电力科学研究院,杭州310014 [3]国网电力科学研究院武汉南瑞有限责任公司,武汉430074

出　　处：《高电压技术》2025年第4期1654-1662,共9页High Voltage Engineering

基　　金：国家电网公司科技项目(5200-202319382A-2-3-XG);国网浙江省电力有限公司科技项目(B311DS24001A);浙江省自然科学基金(LZJMY25D050006)。

摘　　要：针对以往研究往往侧重台风或暴雨等单一灾害下的输电杆塔故障,忽视了台风灾害携带暴雨共同威胁输电杆塔安全。为此建立含降雨量修正的台风灾害下输电杆塔数据机理联合故障概率预测模型,以准确预测台风与暴雨复合作用下输电杆塔故障概率。首先,在数据驱动部分,通过生成对抗网络(generative adversarial network,GAN)解决数据量不足、数据信息不均衡等问题,并以支持向量回归、岭回归、随机森林、K近邻、极端随机树及自适应提升算法等6种机器学习算法预测输电杆塔故障概率。其次,在机理驱动部分,考虑降雨量对输电杆塔的影响,通过降雨雨压模型,计算降雨修正系数修正输电杆塔的故障概率。最后,以2022年登陆浙江省舟山市的台风“梅花”为例进行仿真验证,算例表明所提模型与实际情况更为相符,可精准地预测输电杆塔故障概率。The previous research focuses on the failure of transmission tower under single disasters such as typhoon or rainstorm and neglects that typhoon often carry rainstorm as a common threat to the safety of transmission tower.To address the issue above,we establish a combined failure probability prediction model of data mechanism of transmission tower under typhoon disaster with rainfall correction,so as to accurately predict the failure probability of transmission tower under the combined actions of typhoon and rainstorm.Firstly,in data driving part,the generative adversarial network(GAN)is adopted to solve problems such as insufficient data volume and unbalanced data information.And the support vector regression,ridge regression,random forest,K-nearest neighbors,extra trees,and AdaBoost,are used to predict the failure probability of transmission tower.Secondly,in mechanism driving part,considering the impact of rainfall on transmission tower,the rainfall correction coefficient is calculated to correct the failure probability of transmission tower through rainfall pressure model.Finally,Typhoon“Muifa”,which landed in Zhoushan City,Zhejiang Province,is used for simulation verification.The results show that the proposed model is more consistent with the actual situation,and can be adopted to accurately predict the failure probability of transmission tower.

关 键 词：台风灾害 降雨 输电杆塔 机器学习 生成对抗网络 故障概率预测 

分 类 号：TM753[电气工程—电力系统及自动化] TM75