Suzhou Electric Appliance Research Institute
期刊號: CN32-1800/TM| ISSN1007-3175

Article retrieval

文章檢索

首頁 >> 文章檢索 >> 往年索引

基于ICEEMDAN-SVM算法的復(fù)合絕緣子缺陷識別研究

來源:電工電氣發(fā)布時間:2023-10-18 08:18 瀏覽次數(shù):192

基于ICEEMDAN-SVM算法的復(fù)合絕緣子缺陷識別研究

池小佳1,肖建華1,肖曉江1,邢文忠1,吳慰東1,張建峰2,馮浩文3
(1 廣東電網(wǎng)有限責任公司揭陽供電局,廣東 揭陽 522000;
2 廣東電網(wǎng)有限責任公司梅州供電局,廣東 梅州 514021;
3 廣東工業(yè)大學 自動化學院,廣東 廣州 510006)
 
    摘 要:為了對復(fù)合絕緣子進行快速、有效檢測,提出了基于改進的自適應(yīng)白噪聲完備集合經(jīng)驗?zāi)?span style="font-size: 12px;">態(tài)分解 (ICEEMDAN) 和支持向量機 (SVM) 相結(jié)合的缺陷信號識別方法, 該方法將克服傳統(tǒng)經(jīng)驗?zāi)B(tài)分解的模態(tài)混疊缺點,在對復(fù)合絕緣子進行超聲導(dǎo)波檢測時,可準確、快速識別回波信號,保障電力系統(tǒng)穩(wěn)定運行。對絕緣子進行無缺陷、中部斷面缺陷、中部氣孔缺陷的有限元仿真,運用 ICEEMDAN 對絕緣子各缺陷類型的超聲回波數(shù)據(jù)進行分解;計算出各模態(tài)下的樣本熵、排列熵,并通過 SVM 進行復(fù)合絕緣子的缺陷類型識別。研究結(jié)果表明,基于 ICEEMDAN 與 SVM 的信號識別方法能夠較好地提取復(fù)合絕緣子的
故障特征并進行缺陷識別分類。 
    關(guān)鍵詞: 復(fù)合絕緣子;超聲導(dǎo)波;缺陷識別;改進的自適應(yīng)白噪聲完備集合經(jīng)驗?zāi)B(tài)分解;支持向量機
    中圖分類號:TM216     文獻標識碼:A     文章編號:1007-3175(2023)09-0001-07
 
Research on Defect Identification of Composite Insulators
Based on ICEEMDAN-SVM Algorithm
 
CHI Xiao-jia1, XIAO Jian-hua1, XIAO Xiao-jiang1, XING Wen-zhong1, WU Wei-dong1, ZHANG Jian-feng2, FENG Hao-wen3
(1 Guangdong Power Grid Co., Ltd. Jieyang Power Supply Bureau, Jieyang 522000, China;
2 Guangdong Power Grid Co., Ltd. Meizhou Power Supply Bureau, Meizhou 514021, China;
3 School of Automation, Guangdong University of Technology, Guangzhou 510006, China)
 
    Abstract: In order to detect composite insulators quickly and effectively, a defect signal recognition method based on ICEEMDAN (Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise) and SVM (Support Vector Machine) is proposed in this paper. It not only can overcome mode mixing of the traditional empirical mode decomposition, but also can identify echo signals accurately and quickly to ensure the stable operation of power systems when conducting ultrasonic guided wave detection on composite insulators.First, finite element simulation of no defect insulators, central section defect insulators and central pore defect insulators is carried out. Then,ICEEMDAN is used to decompose ultrasonic echo data of all the detect types of insulators. Third, sample entropies and permutation entropies of all the modes are calculated and SVM is employed to identify the defect types of composite insulators. According to the results, the signal recognition method based on ICEEMDAN and SVM can extract the fault characteristics of composite insulators and can identify and classify these defects well.
    Key words: composite insulator; ultrasonic guided wave; defect identification; improved complete ensemble empirical mode decomposition with adaptive noise; support vector machine
 
參考文獻
[1] 張鳴,陳勉.500 kV 羅北甲線合成絕緣子芯棒脆斷原因分析[J]. 電網(wǎng)技術(shù),2003,27(12) :51-53.
[2] 盧明,張中浩,李黎,等. 復(fù)合絕緣子酥朽發(fā)熱老化的原因分析[J] . 電網(wǎng)技術(shù),2018,42(4) :1335-1341.
[3] 王浩然,郭子豪,張絲鈺,等. 缺陷對特高壓交流盆式絕緣子電場分布的影響[J] . 高電壓技術(shù),2018,44(3) :982-992.
[4] 律方成,牛雷雷,王勝輝,等. 基于紫外成像和改進 YOLOv3 的瓷懸式絕緣子放電嚴重程度評估[J].高電壓技術(shù),2021,47(2) :377-386.
[5] 商俊平,李儲欣,陳亮. 基于視覺的絕緣子定位與自爆缺陷檢測[J] . 電子測量與儀器學報,2017,31(6) :844-849.
[6] 李良. 復(fù)合絕緣子超聲波探傷信號處理方法研究[D].長沙:長沙理工大學,2019.
[7] 鄧紅雷,魯強,陳力,等. 基于超聲導(dǎo)波的復(fù)合絕緣子檢測[J]. 高電壓技術(shù),2016,42(4) :1236-1244.
[8] 鄧紅雷, 陳力, 魯強, 等. 超聲導(dǎo)波檢測絕緣子用玻璃鋼芯棒缺陷[J] . 電工技術(shù)學報,2017,32(12) :268-276.
[9] 魯強.基于超聲導(dǎo)波的復(fù)合絕緣子檢測技術(shù)的研究[D].廣州:華南理工大學,2016.
[10] 陳力. 超聲導(dǎo)波在復(fù)合絕緣子中的傳播特性研究[D].廣州:華南理工大學,2017.
[11] 何戰(zhàn)峰. 基于超聲導(dǎo)波的復(fù)合絕緣子芯棒和脫粘缺陷檢測研究[D]. 廣州:華南理工大學,2018.
[12] 鄧紅雷,何戰(zhàn)峰,陳力. 復(fù)合絕緣子脫粘缺陷的超聲導(dǎo)波檢測[J] . 高電壓技術(shù),2019,45(1) :196-202.
[13] 鄧紅雷,何戰(zhàn)峰,陳力,等. 基于匹配追蹤的 L 模態(tài)超聲導(dǎo)波檢測復(fù)合絕緣子芯棒缺陷研究[J] . 電瓷避雷器,2019(2) :168-174.
[14] 張檣,周西峰,王瑾,等. 基于改進的 EMD 超聲信號降噪方法研究[J] . 南京郵電大學學報(自然科學版),2016,36(2) :49-55.
[15] 左憲章,康健,師小紅,等. 基于小波包最優(yōu)基子帶能量的裂紋特征提取[J] . 機械強度,2010,32(2) :212-217.
[16] LEE K, ESTIVILL-CASTRO V.Feature extraction and gating techniques for ultrasonic shaft signal classification[J] . Applied Soft Computing Journal,2007,7(1) :156-165.
[17] DE FENZA A , SORRENTINO A , VITIELLO P .Application of Artificial Neural Networks and Probability Ellipse Methods for Damage Detection Using Lamb Waves[J].Composite Structures,2015,133 :390-403.
[18] DOBSON J, CAWLEY P.Independent Component Analysis for Improved Defect Detection in Guided Wave Monitoring[J].Proceedings of the IEEE,2016,104(8) :1620-1631.
[19] COLOMINAS M A, SCHLOTTHAUER G, TORRES M E.Improved complete ensemble EMD: A suitable tool for biomedical signal processing[J].Biomedical Signal Processing and Control,2014,14(1) :19-29.
[20] RICHMAN J, MOORMAN J R.Physiological timeseries analysis using approximate entropy and sample entropy[J].American Journal of Physiology Heart and Circulatory Physiology,2000,278(6) :2039-2049.
[21] 陳英強,陳煜敏,蔣勁,等. 基于小波包樣本熵和 SVM 的水泵機組振動故障診斷[J]. 中國農(nóng)村水利水電,2017(3):165-168.
[22] BANDT C , POMPE B . Permutation entropy:A natural complexity measure for time series[J].Physical Review Letters,2002,88(17):174102.
[23] 王志斌,曹紅偉,劉佳佳. 基于小波包去噪與 EMD 的故障電弧檢測算法研究[J] . 電測與儀表,2019,56(6):117-121.
[24] 劉偉,韓彥華,王荊,等. 基于粒子群算法優(yōu)化支持向量機的變壓器繞組變形分類方法[J] . 高壓電器,2020,56(3):72-78.
[25] 邵鑫明,萬書亭,劉榮海,等. 基于 LMD-PCA 和樣本熵的瓷支柱絕緣子故障診斷[J] . 無損檢測,2021,43(3):69-73.