摘要
为实现电力电容器故障诊断和识别的高精度预测,将云计算技术引入深度学习,提出一种基于Map Reduce的分布式DBN的电力电容器故障诊断和识别方法。通过MR_DBN和DBN、SVM、BP的对比,研究结果表明,MR_DBN可提高电力电容器故障诊断和识别的精度,精度可达99.41%,从而为电力电容器故障的研究和应用提供了新的方法。
For achieving high accuracy prediction of fault diagnosis and identification of power capacitor,and introducing cloud computing technology into depth study,a kind of fault diagnosis and identification method of power capacitor based on Map Reduce distribution DBN is proposed. It is shown by the study result by way of comparison of MRDBN and DBN,SVM and BP that MRDBN can improve the accuracy of fault diagnosis and identification of power capacitor up to 99.41%,thus providing new method for the study and application of the fault of power capacitor.
作者
黄予春
曹成涛
顾海
HUANG Yuehun;CAO Chengtao;GU Hai(State Grid Henan Electric Power Company Luohe Power Supply Company,Henan Luohe 462000,China;South China University of Teehnology,Guangzhou 510640,China;Harbin Institute of Teehnology,Harbin 150001,China)
出处
《电力电容器与无功补偿》
北大核心
2018年第4期71-75,共5页
Power Capacitor & Reactive Power Compensation
关键词
云计算
深度学习
故障诊断
电力电容器
支持向量机
eloud eomputing
depth study
fault diagnosis
power eapaeitors
support veetor maehine
