基于粒子群优化小波神经网络的光伏组件故障诊断方法  

Fault diagnosis method of photovoltaic module based on particle swarm optimized wavelet neural network

作　　者：尹春雨[1] 孙凤杰[1] 范杰清[1] 王亚微[2] 

机构地区：[1]华北电力大学电气与电子工程学院,北京102206 [2]青海省电力公司信息通信公司,西宁810008

出　　处：《计算机应用》2017年第A01期106-108,114,共4页journal of Computer Applications

基　　金：中央高校基本科研业务费专项资金资助项目(2016MS06)

摘　　要：针对光伏组件故障监测在线诊断中的传统BP神经网络和小波神经网络全局搜索能力差以及小波神经网络算法高失误率的问题,引入了粒子群优化小波神经网络算法,利用测试获得的光伏组件参数,分别从运算成功率、精度和速度3个角度实现了对光伏组件故障的诊断。仿真结果表明,引入的优化算法在提高全局搜索能力和降低小波神经网络失误率方面优于传统神经网络算法,运算成功率能够达到94%,而同样条件下其他故障算法的还不到90%。Concerning the problem of the poor global search ability of traditional BP（ Back Propagation） neural network and wavelet neural network and high false positives of wavelet neural network algorithm in the online fault diagnosis of photovoltaic modules, the wavelet neural network optimized by the particle swarm was introduced, which combined with the measured photovoltaic component parameters, to diagnoze photovoltaic component faults. The simulation results show that the introduced optimization algorithm for the online fault diagnosis is superior to the traditional neural networks algorithm in improving global search ability and reducing the error rate of a wavelet neural network.

关 键 词：光伏组件 故障诊断 BP神经网络 遗传算法 小波神经网络 粒子群算法 

分 类 号：P315.69[天文地球—地震学]