









摘" 要: 針對變壓器故障診斷存在的精度低、魯棒性不強等問題,提出一種基于卷積神經網絡(CNN)和支持向量機(SVM)的故障診斷方法。首先,基于油中溶解氣體分析(DGA)法,以5種特征量作為輸入,利用CNN提取數據的特征信息;然后導入SVM中進行分類,實現變壓器的故障診斷。基于336組油氣數據對所提模型的性能進行驗證,并將其與其他方法進行對比。實驗結果表明:所構建的CNN?SVM診斷模型與CNN?BiLSTM網絡、LSTM網絡和CNN相比,綜合故障診斷精度分別提高了8.9%、12.5%和19.6%,并且CNN?SVM模型有著更快的運行速度,運行時間約為3.11 s;當修改輸入數據或減少輸入的氣體特征量時,CNN?SVM模型的診斷精度相比于其他方法下降最少,說明CNN?SVM模型具有更好的魯棒性和特征提取能力。
關鍵詞: 變壓器; 故障診斷; 卷積神經網絡; 支持向量機; 特征提取; 診斷精度
中圖分類號: TN911.22?34; TM411" " " " " " " "文獻標識碼: A" " nbsp; " " " " " " " "文章編號: 1004?373X(2025)06?0073?05
Method of transformer fault diagnosis based on CNN?SVM
LI Zhou, WANG Fanrong
(School of Electrical and Electronic Engineering, Hubei University of Technology, Wuhan 430074, China)
Abstract: In allusion to the problems of low accuracy and low robustness of transformer fault diagnosis, a method of fault diagnosis based on convolutional neural networks (CNN) and support vector machine (SVM) is proposed. Based on the dissolved gas analysis (DGA) method in oil, five feature quantities are used as inputs, and the CNN is used to extract the feature information of the data, which is then imported into SVM for classification, so as to realize the fault diagnosis of the transformer. Based on 336 sets of oil and gas data, the performance of the proposed model is verified and compared with other methods. The experimental results show that in comparison with CNN?BiLSTM network, LSTM network and CNN, the comprehensive fault diagnosis accuracy of the constructed CNN?SVM diagnostic model can be improved by 8.9%, 12.5% and 19.6% respectively, and the CNN?SVM model has a much faster running speed, with a running time of about 3.11 s. When modifying the input data or reducing the amount of input gas features, the diagnostic accuracy of the CNN?SVM model decreases the least compared to other methods, which indicates that the CNN?SVM model has better robustness and feature extraction capability.
Keywords: transformer; fault diagnosis; convolutional neural network; support vector machine; feature extraction; diagnostic accuracy
0" 引" 言
變壓器是電力系統中最重要的設備之一,直接影響著電力系統運行的可靠性與安全性,當變壓器發生故障時,很有可能造成惡劣的影響,甚至發生安全事故[1]。因此,進行變壓器的故障診斷研究尤為重要。
變壓器中含有大量的絕緣油與絕緣材料,由于受到電和熱的影響,絕緣材料會逐漸老化分解,從而產生大量的氣體,這些氣體會溶于絕緣油中,當油中氣體的含量逐漸增多時,變壓器內部就會產生故障。目前,變壓器的故障診斷方法最常用的是DGA技術[2],該方法主要是根據特征氣體的比值及含量,得出變壓器的故障類型。……