基于BP神经网络的绝缘子污秽度预测研究  

Research on Prediction of Insulator Contamination Based on BP Neural Network

作　　者：刘力歌 武剑 肖飞 Liu Li-ge;Wu Jian;Xiao Fei

机构地区：[1]国网重庆市电力公司检修分公司,重庆400039

出　　处：《电力系统装备》2020年第13期185-187,共3页Electric Power System Equipment

摘　　要：复杂地区绝缘子污闪问题日益严峻,为了解决由于污闪造成的输电安全问题,文章基于BP神经网络来预测复杂地区绝缘子污秽程度.首先对BP神经网络进行详细分析,然后基于网络模型分析建立人工神经网络应用模型,并且设计有关参数,并运用模型对数据进行研究.结果显示:在泄漏电流不小于5 mA和盐度不小于0.16 mg/cm2的情况下,能够得到正确的预测结果;BP人工神经网络改进算法在有较大泄露电流的情况下,可以正确有效地预测绝缘子表面污秽.The pollution flashover of insulator in complex area is becoming more and more serious.In order to solve the transmission safety problem caused by pollution fl ashover,this paper uses BP neural network to predict the pollution degree of insulator in complex area.Firstly,the BP neural network is analyzed in detail,then the application model of the artificial neural network is established based on the analysis of the network model,and the relevant parameters are designed,and the data are studied by using the model.The results show that the correct prediction results can be obtained when the leakage current is not less than 5mA and the salinity is not less than 0.16mg/cm2;the improved BP artifi cial neural network algorithm can correctly and effectively predict the surface pollution of insulator when there is a large leakage current.

关 键 词：绝缘子 等值附盐密度 BP神经网络 

分 类 号：TM216[一般工业技术—材料科学与工程]