基于BP神经网络的电能表软件故障分类研究  

Research for Power Meter Software Faults Classification Based on BP Neural Network

作　　者：郑蒙蒙[1] 李新利[1] 巨汉基[2] 庞富宽 李斯琪[2] 

机构地区：[1]华北电力大学,北京102206 [2]国网冀北电力有限公司电力科学研究院(华北电力科学研究院有限责任公司),北京100045

出　　处：《华北电力技术》2016年第8期8-12,共5页North China Electric Power

摘　　要：BP神经网络有良好的非线性映射能力,因此可广泛应用于故障分类算法中。针对智能电能表的软件故障,设计了一种基于BP神经网络的分类算法,并应用该算法将智能电能表的8种常见软件故障准确分为4类。最后,使用Matlab软件对该实验进行了计算机仿真实验,仿真的结果验证了所设计的基于BP神经网络的智能电能表故障分类器在实际应用中具有显著的效果。Because of its nonlinear mapping ability, BP neural network is used in an algorithm of fault classification. In order to solve the problem of smart meters＇ software fault classification,a kind of classification algorithms based on BP neural network is designed. And 8 common software faults are classified into 4 classes with the algorithm. Sim- ulation with MATLAB is conducted. The result shows that the classification algorithms based on BP neural network for smart meters is effective.

关 键 词：BP神经网络 智能电能表 软件故障模式 分类 

分 类 号：TM763[电气工程—电力系统及自动化]