基于K-means算法的光伏路灯蓄电池故障识别研究  

Research on fault identification for photovoltaic street light battery based on K-means algorithm

作　　者：侯林涛 阎俏[1,2] 张桂青 金东毅[1] 刘业春 张公飞 Hou Lintao;Yan Qiao;Zhang Guiqing;Jin Dongyi;Liu Yechun;Zhang Gongfei(School of Information and Electrical Engineering,Shandong Jianzhu University,Jinan,Shandong 250101,China;Shandong Key Laboratory of Intelligent Buildings Technology)

机构地区：[1]山东建筑大学信息与电气工程学院,山东济南250101 [2]山东省智能建筑技术重点实验室

出　　处：《计算机时代》2023年第6期114-118,共5页Computer Era

基　　金：山东省重大科技创新工程项目(2019JZZY010115)。

摘　　要：针对乡村光伏路灯存在维修维护成本高、维护不及时等问题,在分析蓄电池故障数据特征的基础上,提出一种基于K-means算法的蓄电池故障识别方法。给出算法流程以及参数K的循环寻优选取方式,在故障聚类基础上分析并设计了故障类型的识别方法。通过对210盏太阳能路灯18个月的电流、电压等实时数据进行验证,该故障识别算法具有较高的故障识别精度。在光伏路灯实现物联网管控的基础上,该算法具有成本低、操作性强的特点。There are problems such as high maintenance cost and untimely maintenance of rural photovoltaic street lights.On the basis of analyzing the data characteristics of battery fault,an identification method of battery fault based on K-means clustering algorithm is proposed.The algorithm process and the cyclic optimization selection method for parameter K are given in detail,and the identification method of fault types is analyzed and designed on the basis of fault clustering.The fault identification algorithm is verified with the real-time data such as current and voltage of 210 solar street lights over 18 months,which shows that the algorithm has high fault identification accuracy.With IoT control for photovoltaic street lights,this algorithm has the advantages of low cost and high operability.

关 键 词：光伏路灯 故障识别 K-MEANS聚类算法 异常数据检测 

分 类 号：TP311.1[自动化与计算机技术—计算机软件与理论]