摘要
传统的故障诊断方法难以发现计量自动化终端较为繁杂的故障逻辑关系,误诊、漏诊的概率极大,为此,提出了一种基于深度置信网络和支持向量机的故障诊断方法。该方法利用深度置信网络进行终端故障数据的特征提取,然后利用支持向量机对新的特征样本集合进行学习,并作为最后的故障分类器,从而提高了故障诊断效率。算例结果表明,所提出的方法提高了计量自动化终端故障诊断的准确率。
It is difficult to find the more complicated fault logic relations of metering automatic terminal by traditional fault diagnosis method,which leads to the very high probability of misdiagnosis.So a fault diagnosis method based on deep belief network and support vector machine is proposed.The deep belief network is used to extract features from fault data of terminal and then the support vector machine is used to learn the new feature samples set and is used as the final fault classifier which improves the fault diagnosis efficiency.Simulation results show that the proposed method improves the accuracy rate of the fault diagnosis of automatic metering terminal.
作者
陈俊
CHEN Jun(Electric Power Reearch Institute of Guangxi Power Grid Co.,Ltd.,Guangxi Nanning 530023,China)
出处
《广西电力》
2018年第5期16-19,共4页
Guangxi Electric Power
关键词
计量自动化终端
故障诊断
特征提取
深度置信网络
支持向量机
metering automatic terminal
fault diagnosis
feature extraction
deep belief network
support vector machine
作者简介
陈俊（1986），男，湖北仙桃人，工程师，工学硕士，从事计量自动化及需求侧管理研究工作。
