最小二乘支持向量回归的光伏组件故障检测  

Fault Detection for Photovoltaic Modules Using Least-square Support Vector Regression

作　　者：李红涛[1] 李春来 杨立滨 丁明昌[1] LI Hong-tao;LI Chun-lai;YANG Li-bin;DING Ming-chang(State Key Laboratory of Operation and Control of Renewable Energy ＆ Storage Systems（China Electric Power Research Institute）,Beijing 100192,China)

机构地区：[1]新能源与储能运行控制国家重点实验室(中国电力科学研究院),北京100192 [2]青海省光伏发电并网技术重点实验室(国网青海省电力公司电力科学研究院),青海西宁810000

出　　处：《电力电子技术》2018年第9期97-100,共4页Power Electronics

基　　金：青海省重点实验室建设专项(2015-Z-Y24);中国电科院青年基金(5342NY160008)~~

摘　　要：光伏组件是光伏发电系统的核心之一,对光伏组件开展及时有效的故障检测有助于提升系统发电效率。在光伏组件机理模型的基础上,采用最小二乘支持向量回归(Ls-SVR)方法对测试样本的I-U曲线进行学习,拟合出光伏组件四大参数在正常运行情况下的内在函数关系,并通过分析函数输出与实际输出之间的残差,实现光伏组件的故障检测。最后采用现有实验数据对所提方法进行验证,效果良好。Photovoltaic modules are considered as one core component of the photovoltaic generation system.To im- prove the generating efficiency, faults of these photovoltaic modules are required to be detected with high accuracy and quick response.A least-square support vector regression（Ls-SVR） based approach is proposed to solve the diag- nostic problem.By studying the mechanism of photovohaic modules,this approach firstly extracts informative features from I-U curves,then implements Ls-SVR to approximate the internal function between four key parameters of the modules.By analyzing the residuals between function outputs and real outputs,the faults of photovoltaic modules can be detected.The oroposed approach is validated on exoerimental datasets, and proved to be satisfactory.

关 键 词：光伏组件 故障检测 最小二乘 支持向量回归 

分 类 号：TM615[电气工程—电力系统及自动化]