一维全卷积神经网络在变压器故障诊断中的应用研究

大坝与安全 ›› 2025 ›› Issue (1): 61-.

一维全卷积神经网络在变压器故障诊断中的应用研究

谭佳文，张文韬，王雅琪

中国长江电力股份有限公司三峡水力发电厂，湖北宜昌，443133

收稿日期:2024-05-24 出版日期:2025-02-10 发布日期:2025-04-14
作者简介:谭佳文（1991— ），男，湖北宜昌人，工程师，主要从事水电站电力设备自动化运维工作。

Application of one-dimensional all convolutional neural network in fault diagnosis of power transformer#br#

TAN Jiawen, ZHANG Wentao and WANG Yaqi

China Yangtze Power Co., Ltd.

Received:2024-05-24 Online:2025-02-10 Published:2025-04-14

摘要/Abstract

摘要： 为提高变压器故障诊断的准确性，提出了一种基于一维全卷积神经网络的油浸变压器故障诊断模型。首先对变压器油中溶解气体数据进行标准化处理和编码；然后构建上述诊断模型的网络，网络结构运用全卷积网络和卷积核的特点，能够有效提取输入数据中的故障特征并用于分类；最后对模型进行了训练和测试。试验结果及对比分析表明，上述方法应用于电力变压器故障诊断时具有较高的准确性。

关键词: 卷积神经网络, 深度学习, 变压器, 故障诊断, 溶解气体分析

Abstract: To improve the accuracy of fault diagnosis for transformer, a fault diagnosis model for oil immersed transformer based on one-dimensional all convolution neural network is proposed. Firstly, the data of gases dissolved in the transformer oil is standardized and coded. Then the network of the proposed diagnosis model is established. By using the characteristics of all convolutional network and convolution kernel, the network structure can effectively extract fault features from input data and then apply it for classification. Finally, the model is trained and tested. The test result and contrastive analysis show that superior performance is achieved when the model is applied to fault diagnosis of power transformer.

Key words: convolutional neural network, deep learning, transformer, fault diagnosis, dissolved gas , analysis

中图分类号:

TM411

谭佳文, 张文韬, 王雅琪. 一维全卷积神经网络在变压器故障诊断中的应用研究[J]. 大坝与安全, 2025, 0(1): 61-.

TAN Jiawen, ZHANG Wentao and WANG Yaqi. Application of one-dimensional all convolutional neural network in fault diagnosis of power transformer#br#[J]. Dam & Safety, 2025, 0(1): 61-.

参考文献

[1] 禹建丽,卞帅.基于BP神经网络的变压器故障诊断模型[J].系统仿真学报,2014,26(6):1343-1349.
[2] Bustamante S,Manana M,Arroyo A,et al.Dissolved gas analysis equipment for online monitoring of transformer oil: A review[J].Sensors,2019,19(19):4057.
[3] 龚瑞昆,李昊.主元分析优化量子神经网络的变压器故障诊断[J].现代电子技术,2019,42(17):119-123,128.
[4] 田凤兰,张恩泽,潘思蓉,等.基于特征量优选与ICA-SVM的变压器故障诊断模型[J].电力系统保护与控制,2019,47(17):163-170.
[5] 公茂法,张言攀,柳岩妮,等.基于BP网络算法优化模糊Petri网的电力变压器故障诊断[J].电力系统保护与控制,2015,43(3):113-117.

[6] 贾京龙,余涛,吴子杰,等.基于卷积神经网络的变压器故障诊断方法[J].电测与仪表,2017,54(13):62-67.

[7] 石鑫,朱永利.深度学习神经网络在电力变压器故障诊断中的应用[J].电力建设,2015,36(12):116-122.
[8] 张玉振,吉兴全,彭立岩,等.基于栈式自编码器和Softmax分类器的电力变压器故障诊断[J].中国科技论文,2018,13(23):2694-2699.
[9] 石鑫.基于深度学习的变压器故障诊断技术研究[D].保定:华北电力大学,2016.
[10] Dai J J,Song H,Sheng G H,et al.Dissolved gas analysis of insulating oil for power transformer fault diagnosis with deep belief network[J].IEEE Transactions on Dielectrics and Electrical Insulation,2017,24(5):2828-2835.
[11] Li J L,Li X Y,He D,et al.A directed acyclic graph network combined with CNN and LSTM for remaining useful life prediction[J].IEEE Access,2019,7:75464-75475.
[12] Lu S X,Feng J,Zhang H G,et al.An estimation method of defect size from MFL image using visual transformation convolutional neural network[J].IEEE Transactions on Industrial Informatics,2019,15(1):213-224.
[13] Peng D D,Liu Z L,Wang H,et al.A novel deeper one-dimensional CNN with residual learning for fault diagnosis of wheelset bearings in high-speed trains[J].IEEE Access,2019,7: 10278-10293.
[14] Srivastava N,Hinton G,Krizhevsky A,et al.Dropout: A simple way to prevent neural networks from overfitting[J].Journal of Machine Learning Research,2014,15(1):1929-1958.
[15] Springenberg J T,Dosovitskiy A,Brox T,et al.Striving for simplicity: the all convolutional net[C].International Conference on Learning Representations,2015.