基于BP神经网络的光伏阵列智能故障诊断  

The photovoltaic array intelligent fault diagnosis based on the BP neural network

作　　者：张文瑾 葛强 黄澄扬 周晨 葛中强 

机构地区：[1]扬州大学水利与能源动力工程学院,江苏扬州225127

出　　处：《扬州大学学报（自然科学版）》2015年第1期32-35,共4页Journal of Yangzhou University：Natural Science Edition

基　　金：江苏省科技支撑计划资助项目(BE2010699);扬州大学参照211工程建设专项基金;扬州大学学术科技创新基金资助项目(b13294)

摘　　要：针对光伏阵列的故障特点以及传统故障诊断方法的局限性,提出一种基于BP神经网络的光伏阵列智能故障诊断策略.分析了电池裂化、老化短路和阴影故障的后果,通过采集合适的故障样本训练神经网络,并在Matlab环境中进行了组件在线故障诊断的仿真和实验研究,证明了所提出故障诊断系统的准确性、有效性和环境适应性.In consideration of the PV array fault characteristics and the limitations of traditional fault diagnosis methods an intelligent fault diagnosis strategy for PV array is proposed based on BP neural network. The failure consequences of battery cracking, aging short circuit and the shadow are analyzed. Through collecting appropriate fault samples to train neural network, the simulation and experimental researches on online fault diagnosis of PV modules in the Matlab environment are performed, and the results prove the accuracy, effectiveness and environmental adaptability of the proposed system.

关 键 词：光伏阵列 故障诊断 BP神经网络 网络训练 MATLAB仿真 

分 类 号：TM615.2[电气工程—电力系统及自动化] TP183[自动化与计算机技术—控制理论与控制工程]