基于最小二乘支持向量機的變壓器故障診斷
王逸萍
(江蘇省電力公司檢修分公司,江蘇 無錫 214001)
摘 要:介紹了一種基于最小二乘支持向量機(LS-SVM) 的電力變壓器故障診斷方法,將樣本數據進行歸一化處理,以絕緣油中特征氣體種類及其含量為依據建立變壓器故障診斷LS-SVM 模型,對模型中的核參數σ 與懲罰參數C 進行優(yōu)化,并將測試樣本輸入訓練好的LS-SVM 模型,得到診斷結果。實例結果分析表明,LS-SVM 將原先的非線性問題轉化為求解線性問題,即使在小訓練樣本的前提下,也能獲得更為準確的診斷結果。
關鍵詞:電力變壓器;故障診斷;最小二乘支持向量機;核函數;氣體分析
中圖分類號:TM41 文獻標識碼:A 文章編號:1007-3175(2016)06-0024-04
Fault Diagnosis of Power Transformers Based on
Least Squares Support Vector Machine
WANG Yi-ping
(Jiangsu Electric Power Maintenance Branch Company, Wuxi 214001, China)
Abstract: Introduction was made to a kind of power transformer fault diagnosis method based on least squares support vector machine (LS-SVM). The sample data was carried out normalization processing. On the basis of the characteristic gas type and its content of the insulating oil, this paper established the LS-SVM model of transformer fault diagnose and optimized the nuclear parameter σ and penalty parameter C in the model, putting the test sample into the trained LS-SVM model to obtain the diagnosis results. Experimental results analysis shows that LS-SVM changes the original nonlinear problem into the solution of linear problem and the more accurate diagnosis result could be obtained even under the conditions of small training sample.
Key words: power transformer; fault diagnosis; least squares support vector machine; kernel function; gas analysis
參考文獻
[1] 周婧婧. 基于故障樹分析的電力變壓器可靠性評估方法研究[D]. 重慶:重慶大學,2009.
[2] 楊志超,張成龍,吳奕,等. 基于粗糙集和RBF神經網絡的變壓器故障診斷方法研究[J]. 電測與儀表,2014,51(21):34-39.
[3] 張啟清. 電力變壓器故障診斷專家系統的研究[D].重慶:重慶大學,2002.
[4] NELLO Cristianimi, JOHN Shawe-Taylor.An Introduction to Support Vector Machines and Other Kernel-based Learning Methods[M].New York:Cambridge University Press,2000.
[5] 武中利. 電力變壓器故障診斷新方法研究[D]. 北京:華北電力大學,2013.
[6] 尹金良,朱永利. 支持向量機參數優(yōu)化及其在變壓器故障診斷中的應用[J]. 電測與儀表,2012,49(5):11-16.
[7] 鄭含博. 電力變壓器狀態(tài)評估及故障診斷方法研究[D]. 重慶:重慶大學,2012.
[8] 孫才新,陳偉根,李檢,等. 電氣設備油中氣體在線監(jiān)測與故障診斷技術[M]. 北京:科學出版社,2003.