基于BP神经网络的智能台区识别方法研究  

Research on the intelligent transformer area recognition method based on BP neural network

作　　者：李亚[1] 樊汝森[2] 蒋伟[1] 杨俊杰[1] 宋涛[1] 赵勤学[1] 

机构地区：[1]上海电力学院电子与信息工程学院,上海200090 [2]国网上海市电力公司青浦供电公司,上海200122

出　　处：《电测与仪表》2017年第3期25-30,共6页Electrical Measurement & Instrumentation

基　　金：国家自然科学基金资助项目(61202369);上海市自然科学基金自主项目(14ZR1417400);上海市科技创新行动计划地方院校能力建设项目(13160500900);上海市教育委员会科研创新项项(12ZZ176;13YZ102)

摘　　要：为方便普查用户台区和相位信息,特别是解决跨台区用户信息识别难题,提出一种基于BP神经网络的智能台区用户信息识别方法并研制了该系统。系统由识别器和手持器两部分组成,通信方式采用电力载波通信技术,对于垮台区用户,依据系统和已识别用户之间的通信信号品质,选取隐藏层节点数为6的前向BP神经网络作为跨台区用户识别模型进行识别。MATLAB仿真和实际测试结果表明:该方法可有效解决跨台区通信串扰难题,能够智能识别用户台区和相位信息,同时具有识别准确性高、容差性能较好的优点,对提高台区用户信息识别准确性,减少工作量降低成本,具有重要意义。For the convenience of census the transformer area and the phase information,especially to solve the identifying problems of user information across the transformer area,we put forward a kind of intelligent transformer area of the users information recognition method based on BP neural network and developed the system.This system consists of identifier and handheld devices,the communication adopted power line carrier communication technology.For the across transformer area,according to the communication signal quality between the system and the recognized users,we selected the forward BP neural network with the hidden layer nodes is 6 as the across the transformer area model to recognize.The MATLAB simulation and actual test results show that this method can not only effectively solve the cross transformer area communication crosstalk problems,but also intelligently identify the user transformer area and the phase information,and also with the advantages of high identification accuracy,higher tolerance performance,which has a great significance to improve the accuracy of the transformer area and the phase information,and reduce the workload and cost.

关 键 词：台区识别 电力载波 信号品质 BP神经网络 

分 类 号：TM934[电气工程—电力电子与电力传动]