数据与模型驱动在输电杆塔风致灾损预测中的应用对比分析  

Comparative Analysis of Data-driven Method and Model-driven Method in Wind-induced Damage Prediction of Transmission Towers

作　　者：侯慧[1] 朱韶华 吴细秀[1] 魏瑞增 梁永超 何浣 HOU Hui;ZHU Shaohua;WU Xixiu;WEI Ruizeng;LIANG Yongchao;HE Huan(School of Automation,Wuhan University of Technology,Wuhan 430070,Hubei Province,China;Guangdong Key Laboratory of Electric Power Equipment Reliability(Electric Power Research Institute of Guangdong Power Grid Co.,Ltd.),Guangzhou 510080,Guangdong Province,China)

机构地区：[1]武汉理工大学自动化学院,湖北省武汉市430070 [2]广东省电力装备可靠性重点实验室(广东电网有限责任公司电力科学研究院),广东省广州市510080

出　　处：《电网技术》2023年第4期1721-1727,共7页Power System Technology

基　　金：教育部产学合作协同育人项目(201902056044);中国南方电网有限责任公司科技项目(GDKJXM20198441(036100KK52190053)、GDKJXM20210044(036100KK52210047))。

摘　　要：应对气候变化带来的电网灾损是保障新型电力系统安全运行的关键之一。该文以输电杆塔风致灾损预测为出发点,对现有两类主流方法——数据驱动与模型驱动进行对比分析。在数据驱动方面,基于多源异构数据构建样本集,利用随机森林算法对输电杆塔风致灾损概率进行预测。在模型驱动方面,将输电线路与杆塔构成的塔线系统进行建模分析,利用设备应力强度与风荷载相关关系求取杆塔损毁概率。最后,通过台风“彩虹”下我国沿海某地区案例,对两类方法的预测结果进行对比,结果表明,模型驱动方法的预测精度较高,但建模复杂度较高,难以大范围广域应用;而数据驱动方法建模复杂度较低,计算效率高,同时具有较高预测精度,可为生产实际提供指导。随着数字化、信息化及智能化等技术普及,数据驱动方法因其建模复杂度低等特点,将会在未来科学应对电网灾害中发挥关键作用。Coping with the power grid disasters caused by climate changes is one of the keys to ensure the safe operation of the new power system.Based on the wind-induced damage prediction for the transmission towers,this paper compares the existing two mainstream prediction methods:the data-driven method and the model-driven method.In the data-driven method,a data driving set is constructed based on the multi-source heterogeneous data,and the random forest algorithm is used to predict the wind-induced disaster probability on a network tower.In the model-driven method,the tower line system composed of the transmission lines and the towers is modeled and analyzed.And the damage probability of the towers is calculated by using the correlation between the equipment stress intensity and the wind load.Finally,through the case analysis of typhoon"Mujigae",the results of the two prediction methods are compared.It shows that the model-driven method has a higher prediction accuracy but it is more complex to model for it,unappliable for widespread areas.The data-driven method has lower modeling complexity,but higher computational efficiency,wider application range,lower data dependence and lower realization economy.At the same time,it has 73.5%precision rate and 70.5%recall rate,which provides guidance for production practice.With the popularization of digital,information and intelligent technologies,the data-driven method will play a more and more important role in the scientific response to the power grid disasters in the future because of its superiorities above.

关 键 词：风致灾损 输电杆塔 防灾减灾 数据驱动 模型驱动 

分 类 号：TM721[电气工程—电力系统及自动化]