摘要
针对现有的大多数人工智能模型难以适用于变电站柜内电力线缆寿命预测的问题,文中设计一种基于WSN与改进SVM算法的变电站柜内线缆寿命预测模型。该模型利用无线传感网络构建包含同步相量测量单元的变电站柜内环境数据采集传输系统,并将采集到的数据上传至系统管理中心做进一步分析。同时,采用Rao算法对支持向量机算法加以改进,并将改进后的算法用于预测变电站柜内线缆的寿命。通过对比预测值与安全阈值的数据得到最终识别结果,实现了电缆状态的异常识别,从而辅助工作人员采取相应的措施。基于搭建的测试平台对所提模型进行了实验测试,分析所得到的数据可知,预测结果的ME、RMSE和MAPE值分别为78 h,52 h及2.263%,能够准确地预测变电站柜内线缆的寿命。
In allusion to the problem that most existing artificial intelligence models are difficult to apply to the life prediction of power cables in substation cabinets,a substation cabinet cables service life prediction model based on WSN and improved SVM(support vector machine)algorithm is designed.In the model,the wireless sensor network is used to build a substation cabinet environment data acquisition and transmission system including synchronous phasor measurement unit,and upload the collected data to the system management center for further analysis.Rao algorithm is used to improve SVM algorithm,and the improved algorithm is used to predict the service life of the cables in the substation cabinet.The final identification result is obtained by comparing the data of the predicted value with the safety threshold to realize the abnormal identification of the cable state,thus assisting the staff to take appropriate measures.The experimental test of proposed model was conducted based on the built testing platform.According to the analysis of the obtained data,the ME,RMSE and MAPE values of the prediction results are 78 h,52 h and 2.263%,respectively,which can accurately predict the service life of the cables in the substation cabinet.
作者
翁一潇
邹沉
余皓
鲁敏
朱俊
李挺
WENG Yixiao;ZOU Chen;YU Hao;LU Min;ZHU Jun;LI Ting(School of Electrical Engineering and Automation,Wuhan University,Wuhan 430072,China;State Grid Hubei Electric Power Co.,Ltd.,Wuhan 430050,China)
出处
《现代电子技术》
2022年第24期103-107,共5页
Modern Electronics Technique
基金
国家重点研发计划资助项目(2017YFB0902904)。
关键词
变电站
柜内线缆
寿命预测
无线传感器网络
Rao算法
支持向量机
同步相量
异常识别
substation
cabinet cable
service life prediction
wireless sensor network
Rao algorithm
SVM
synchronous phasor measurement unit
abnormal recognition