小波神经网络在绝缘子漏电量预测中的应用  

Study and application of wavelet neural network in prediction Creep quantity of insulator

作　　者：刘唯义[1] 王丽侠[2] 

机构地区：[1]大庆石油学院审计处,黑龙江大庆163318 [2]浙江师范大学信息学院,浙江金华321004

出　　处：《应用科技》2007年第6期12-14,22,共4页Applied Science and Technology

摘　　要：为了实现对绝缘子漏电量的准确预测,提出了基于小波神经网络的预测模型,分析了网络的拓扑结构,给出了网络学习方法.通过对绝缘子漏电量样本数据进行预处理,生成学习样本和测试样本,进而对预测模型进行测试,实现了对绝缘子漏电量的准确预测.将其应用于电业局的绝缘子漏电量预测中,达到了实际应用的精度要求.实验和实际应用表明,该预测模型的误差小,精确度高,能有效地预测绝缘子漏电量.A prediction model was put forward based on wavelet neural network （WNN） to accurately predict creepage quantity of insulator. The topological structure of WNN was analyzed. The learning algorithm of WNN was given. Through the preprocessing of sampled data of creepage quautity and generation of learning samples and test samples, the test was made on the prediction model, obtaining the accurate prediction of creepage quantity of insulators. The testing method was applied in the creepage test in Electricity Power Bureau, meeting the precision requirement in actual application. Experimental results and application show that the prediction model proposed here has high precision and is efficient for predicting creepage quantity of insulator.

关 键 词：小波神经网络 预测 模型 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]