基于SVM-BP混合网络的输电线路安全评价模型研究  

Research on Transmission Line Safety Evaluation Model Based on SVM-BP Hybrid Network

作　　者：黄锋 崔志美 黄志都 蒋圣超 俸波 HUANG Feng;CUI Zhimei;HUANG Zhidu;JIANG Shengchao;FENG Bo(Electric Power Research Institute of Guangxi Power Grid Co.,Ltd.,Nanning 530023,Guangxi,China)

机构地区：[1]广西电网有限责任公司电力科学研究院,广西南宁530023

出　　处：《电气传动》2023年第8期77-83,共7页Electric Drive

摘　　要：为了能够科学、客观地评价输电线路安全状态,提出了基于SVM-BP混合网络的输电线路安全分析模型。首先,对天气数据、输电线路标准文件和缺陷数据等进行处理和分析,利用关联规则和主成分分析法对输电线路缺陷数据进行整合,建立输电线路安全状态评价体系;然后,提出了SVM-BP混合网络模型挖掘各因素与缺陷状态间的关联,评价输电线路的安全状态;最后,以某地区输电线路为数据集进行验证,该评价模型的准确率达到97.4%,可以快速准确地评价输电的安全状态,并在灾害性天气下提前做好应对措施,保证输电线路的稳定运行。In order to evaluate the safety status of transmission lines scientifically and objectively,a safety analysis model of transmission lines based on support vector machine-back propagation(SVM-BP)hybrid network was proposed.Firstly,the weather data,transmission line standard documents and defect data were processed and analyzed,and the defect data of transmission lines were integrated by association rules and principal component analysis,and the safety status evaluation system of transmission lines was established.Then,the SVM-BP hybrid network model was proposed to mine the correlation between each factor and the defect state,and evaluate the security state of the transmission line.Finally,the accuracy of the evaluation model is 97.4% when the transmission line in a certain region was used as the data set to verify.It can evaluate the safety state of transmission quickly and accurately,and take countermeasures in advance under disastrous weather to ensure the stable operation of transmission line.

关 键 词：输电线路 关联规则 主成分分析 支持向量机-多层前馈神经网络 

分 类 号：TM769[电气工程—电力系统及自动化]