电力信息与通信技术2025，Vol.23Issue(11)：8-17,10.DOI:10.16543/j.2095-641x.electric.power.ict.2025.11.02
基于协方差矩阵和模糊神经网络的光伏电站光伏板异常状态检测
Abnormal State Detection of Photovoltaic Panels in Photovoltaic Power Plants Based on Covariance Matrix and Fuzzy Neural Network
摘要
Abstract
At present,the intelligent image detection method for abnormal states of photovoltaic panels lacks the process of analyzing the correlation between key sample features,and the recognition process mainly relies on neural networks.When faced with similar nonlinear features,it will fall into an iterative loop capture process,resulting in a high rate of missed and false detections of abnormal states of photovoltaic panels.Design an abnormal state detection method for photovoltaic panels in photovoltaic power plants based on covariance matrix and fuzzy neural network.Firstly,the surface image and internal working data of the photovoltaic panel are collected,and the initial image of the photovoltaic panel is processed through distortion correction,image filtering,and image enhancement to obtain the processed image;Secondly,the covariance matrix is used to analyze the correlation between different parameters of the data,extract the internal working data features of the photovoltaic panel,construct a fuzzy neural network to obtain the changes in the operation data of the photovoltaic panel,capture surface image features,and solve nonlinear infection problems;Finally,based on the detection criteria for abnormal states,the abnormal state and type of the current photovoltaic panel are determined through feature matching,achieving the detection of abnormal states in the photovoltaic panel.The conclusion drawn from performance testing experiments is that compared with traditional detection methods,the optimized design method reduces the detection errors of abnormal current and surface abnormal area of photovoltaic panels by about 0.35 A and 0.03 m2,respectively.At the same time,the missed detection rate and false detection rate are significantly reduced.This method has obvious advantages in photovoltaic panel performance testing.关键词
协方差矩阵/模糊神经网络/光伏电站/光伏板/异常状态检测Key words
covariance matrix/fuzzy neural network/photovoltaic power plants/photovoltaic panels/abnormal state detection分类
计算机与自动化引用本文复制引用
樊腾飞,彭超逸,聂涌泉,黄俊聪,刘杰..基于协方差矩阵和模糊神经网络的光伏电站光伏板异常状态检测[J].电力信息与通信技术,2025,23(11):8-17,10.基金项目
南方电网公司重点科技项目资助"分布式能源提供电力辅助服务的可信感知与调控"(000005KC22120026). （000005KC22120026）
