基于深度信念网络的变压器故障诊断  

Transformer Fault Diagnosis Based on Deep Belief Network

作　　者：刘冰尧 庞腾 雷菊阳[1] LIU Bing-yao;PANG Teng;LEI Ju-yang(College of Mechanical Engineering,Shanghai University of Engineering Science;Zhejiang ECO Environmental Technology Co.,Ltd.)

机构地区：[1]上海工程技术大学机械工程学院 [2]浙江宜可欧环保科技有限公司

出　　处：《化工自动化及仪表》2018年第11期873-878,共6页Control and Instruments in Chemical Industry

摘　　要：提高变压器故障诊断精度的关键是构建具有强大数据处理能力和故障特征提取能力的诊断模型,而深度信念网络(Deep Belief Networks,DBNs)能够突显样本数据的特征性。本项目整个深度网络使用自适应算法进行优化,实现学习速率的自适应变化,并对变压器油色谱中的监测数据进行深入分析,实现对故障情况的迅速准确判断。Constructing a fault diagnosis model to be powerful in data-handling capacity and in extracting fault characteristics is crucial to improving the fault diagnosis accuracy of the transformer, and the deep belief networks （DBNs） can highlight the characteristic of sample data. In this paper, having adaptive algorithm adopted to optimize whole deep belief networks to achieve adaptive learning rate changes and analyze monitoring data in the transtormer oil chromatographic analysis so as to decide the fault conditions rapidly and accurately.

关 键 词：油浸式变压器 故障诊断 深度信念网络 自适应学习速率 

分 类 号：TH865[机械工程—仪器科学与技术] TM411[机械工程—精密仪器及机械]