基于SF6分解組分的負(fù)極性直流局部放電故障診斷
楊旭1,2,3,黃勤清1,2,張曉星3,文豪1,2,聶德鑫1,2,周文1,2,江翼1,2
(1 南瑞集團(tuán)(國網(wǎng)電力科學(xué)研究院)有限公司,江蘇 南京 211006;2 國網(wǎng)電力科學(xué)研究院武漢南瑞有限責(zé)任公司,湖北 武漢 430074;
3 武漢大學(xué) 電氣與自動化學(xué)院,湖北 武漢 430072)
摘 要:為了利用SF6局部放電(PD)分解特性開展直流氣體絕緣設(shè)備(GIE)故障診斷研究,以直流GIE中最為常見的4 種絕緣缺陷為例,研究了缺陷從起始放電發(fā)展至臨近擊穿整個過程的PD特性,選擇q v、n v 和Δt v作為表征PD狀態(tài)的特征量,并將PD嚴(yán)重程度劃分為3 個等級;在每種缺陷的3 個PD等級下開展了大量SF6分解實驗,獲取了SF6分解特性。實驗結(jié)果表明,SF6分解生成了CF4、CO2、SO2F2、SOF2 和SO2 5 種穩(wěn)定組分,其中SOF2是最主要的分解產(chǎn)物,且含硫組分的生成量高于含碳組分的生成量;構(gòu)建了由21 個濃度比值組成的特征集合,運用最大相關(guān)最小冗余準(zhǔn)則進(jìn)行特征量選擇,并基于BP神經(jīng)網(wǎng)絡(luò)和支持向量機(jī)進(jìn)行了故障診斷,準(zhǔn)確率超過88%。
關(guān)鍵詞:局部放電;SF6 分解特性;最大相關(guān)最小冗余;故障診斷
中圖分類號:TM21 文獻(xiàn)標(biāo)識碼:A 文章編號:1007-3175(2020)09-0001-07
Type Identification of Negative DC Partial Discharge Based on SF6 Decomposed Components
YANG Xu1, 2, 3, HUANG Qin-qing1, 2, ZHANG Xiao-xing3, WEN Hao1, 2, NIE De-xin1, 2, ZHOU Wen1, 2, JIANG Yi1, 2
(1 Nanjing NARI Group Corp(State Grid Electric Power Research Institute), Nanjing 211 006, China;
2 Wuhan NARI Co., Ltd, State Grid Electric Power Research Institute, Wuhan 430074,China;
3 School of Electrical Engineering and Automation, Wuhan University, Wuhan 430072,China)
Abstract: In order to use SF6 decomposition characteristics to identify faults of DC gas-insulated equipment (GIE) under partial discharge (PD), this paper studied the PD characteristics of the whole process from the initial discharge to near breakdown of the four most common insulation defects in DC-GIE. qv, nv, and Δtv are selected as the feature quantities for characterizing the PD state, and the PD severity is divided into three levels. Then, a large number of SF6 decomposition experiments were carried out under the three PD level of each defect, and the decomposition characteristics of SF6 were obtained. The experimental results show that SF6 decomposition produces include five stable components of CF4, CO2, SO2F2, SOF2 and SO2, among which SOF2 is the most important decomposition product, and the formation amount of sulfur components are higher than that of carbonaceous components. Finally, a feature set consisting of 21 concentration ratios is constructed, and the maximum relevance minimum redundancy criterion is used for feature quantity selection. BP neural network and support vector were used for fault identification, and the accuracy rate was higher than 88%.
Key words: partial discharge; SF6 decomposition characteristics; maximum relevance minimum redundancy; fault identification
參考文獻(xiàn)
[1] YANG D, ZENG F P, YANG X, et al.Comparison of SF6 decomposition characteristics under negative DC partial discharge initiated by two kinds of insulation defects[J].IEEE Transactions on Dielectrics and Electrical Insulation,2018,25(3):863-872.
[2] 李乃一,彭宗仁,劉鵬.1 100 kV直流SF6氣體絕緣穿墻套管電場仿真分析[J]. 高電壓技術(shù),2020,46(1):205-214.
[3] YANG D, TANG J, ZENG F P, et al.Correlation characteristics between SF6 decomposition process and partial discharge quantity under negative DC condition initiated by free metal particle defect[J].IEEE Transactions on Dielectrics and Electrical Insulation,2018,25(2):574-583.
[4] TANG J, YANG X, YAO Q, et al.Correlation analysis between SF6 decomposed components and negative DC partial discharge strength initiated by needle-plate defect[J].IEEJ Transactions on Electrical & Electronic Engineering,2018,13(3):382-389.
[5] 趙科, 王靜君, 劉通, 等. 直流G I L 絕緣設(shè)計及局部放電檢測研究進(jìn)展[J]. 電力工程技術(shù),2017(5):98-102.
[6] VANBRUNT R J, MISAKIAN M.Mechanisms for Inception of DC and 60 Hz AC Corona in SF6[J]. IEEE Transactions on Electrical Insulation,1982,17(2):106-120.
[7] CHU F Y.SF6 decomposition in gas-insulated equipment[J].IEEE Transactions on Electrical Insulation,1986,21(5):693-725.
[8] 唐炬,任曉龍,張曉星,等. 氣隙缺陷下不同局部放電強(qiáng)度的SF6分解特性[J]. 電網(wǎng)技術(shù),2012,36(3):40-45.
[9] VANBRUNT R J, HERRON J T. Fundamental processes of SF6 decomposition and oxidation in glow and corona discharges [J]. I E E E Transactions on Electrical Insulation,1990,25(1):75-94.
[10] TANG J, LIU F, MENG Q H, et al.Partial discharge recognition through an analysis o f SF6 decomposition products part 2: Feature extraction and decision tree-based pattern recognition[J].IEEE Transactions on Dielectrics and Electrical Insulation,2012,19(1):37-44.
[11] 陳俊. 基于氣體分析的SF6電氣設(shè)備潛伏性缺陷辨識技術(shù)研究及應(yīng)用[D]. 武漢:武漢大學(xué),2014.
[12] 齊波,李成榕,駱立實,等. GIS中局部放電與氣體分解產(chǎn)物關(guān)系的試驗[J]. 高電壓技術(shù),2010,36(4):957-963.
[13] 趙永寧,葉林. 區(qū)域風(fēng)電場短期風(fēng)電功率預(yù)測的最大相關(guān)- 最小冗余數(shù)值天氣預(yù)報特征選取策略[J].中國電機(jī)工程學(xué)報,2015,35(23):5985-5994.
[14] 盧詩華,孫密,謝景海,等. 一種基于最大相關(guān)-最小冗余算法的輸電線路[J]. 電測與儀表,2020,57(3):79-85.
[15] LENG X, WANG J, JI H, et al.Prediction of size-fractionated airborne particle-bound metals using MLR, BP-ANN and SVM analyses[J]. Chemosphere,2017,180:513-522.
[16] 孟衛(wèi)東,劉楊,張偉,等. 基于B P 神經(jīng)網(wǎng)絡(luò)的變壓器故障診斷[J]. 通信電源技術(shù),2020,37(2):84-86.
[17] 朱梓倩,劉蓉,付瑜,等. 基于振動云圖HOG和SVM的變壓器繞組松動故障診斷方法[J]. 高壓電器,2019,55(11):227-231.
[18] 張文雅,范雨強(qiáng),韓華,等. 基于交叉驗證網(wǎng)格尋優(yōu)支持向量機(jī)的產(chǎn)品銷售預(yù)測[J]. 計算機(jī)系統(tǒng)應(yīng)用,2019,28(5):1-9.
[19] 楊廷方,劉沛,李景祿,等. FCM結(jié)合IEC三比值法診斷變壓器故障[J]. 高電壓技術(shù),2007,33(8):66-70.
[20] 楊鑫,張家洪,李英娜,等. 基于FCM聚類的小電流接地系統(tǒng)故障區(qū)段定位[J]. 電力科學(xué)與工程,2019,35(12):49-55.