一种基于BP神经网络的电能表校验方法  

Watt-hour Meter Calibration Based on BP Neural Network(NN)

作　　者：高中文[1] 董红福[1] 宋伟伟[1] 

机构地区：[1]哈尔滨理工大学自动化学院,哈尔滨150080

出　　处：《电测与仪表》2006年第9期37-39,共3页Electrical Measurement & Instrumentation

摘　　要：本文运用了神经网络技术,用训练成熟的多层前向BP神经网络取代了电能表校验的标准表。实验结果表明,该方法可以很高精度逼近标准表的输入输出数据,且对不同的输入信号有较好的泛化能力,可以做为下一级电能表校验的标准表,降低了电能表校验的成本。With the NN technology ,the study poses a new calibration method for Watthour meter. The mature trained BP NN,which substitues for the standard meter checkout way.Through experiment, it indicates that, using the new way,it can approach the input and output data tested by the standard meter accurately,and for different input signals it has better generalization capability,then it can become the standard meter for the next level.Utilizing this way,it can lower the cost of energy meter checkout.

关 键 词：电能表校验 标准表 BP神经网络 数据处理 

分 类 号：TM933[电气工程—电力电子与电力传动]