It is very important to find out the type of partial discharge signal of high-voltage cable as soon as possible and take corresponding measures to effectively avoid damage to the insulation system. Aiming at the partial discharge signals generated by four kinds of typical defects in high-voltage cables, this paper proposes a classification method of partial discharge signals in high-voltage cables based on a convolutional neural network. Firstly, the defect model is constructed, and the partial discharge signals of four types of defects are collected as samples, and the voltage amplitude-phase spectrum is used as the input data set. Then, the multi-layer convolution kernel is used for feature extraction, and multiple classifiers are used for classification regression. Finally, the trained convolution neural network is obtained. In comparison, the classification method based on back propagation (BP) neural network and autoencoder neural network have unstable classification effects and poor overall accuracy. The feature extraction ability of the method proposed in this paper is stronger, the classification effect of each kind of signal is better, and the application prospect is much broader.
现在常用的PD检测方法主要有超高频法[7]、超声波法[8]、高频电流传感器法[9]等。超高频法是通过检测PD现象产生的超高频电磁波来获取相关信息。超声波法是利用超声波信号的改变来检测电缆中存在的缺陷。高频电流传感器(high frequency current transformer, HFCT)法的原理是:在接地线上安装传感器,PD现象产生的高频电流会沿接地线传播,从而在传感器内产生感应电压,实现对PD信号的检测。这种方法不仅易操作、灵敏度高、抑制噪声性能好,而且在检测过程中,对电缆传输无影响,因而被广泛应用[10]。
CAOJ P, JIANGY K, WANGS H, et al. Analysis and prospect of defect detection key technology for XLPE power cable joints [J]. High Voltage Apparatus, 2018, 54(7): 87-97. DOI:10.13296/j.1001-1609.hva.2018.07.010(Ch ).
LIU G, CHEN Z J. Experiment of main insulation with air gap of 10 kV cross linked polyethylene cable terminal[J]. High Voltage Engineering, 2012,38(3):678-683 (Ch).
GAOJ C, ZHUY L, JIAY F, et al. Pattern recognition of unknown types of partial discharge based on improved SVDD algorithm and mahalanobis distance [J]. Transactions of China Electrotechnical Society, 2018,33(15):3510-3517. DOI:10.19595/j.cnki.1000-6753.tces.180089(Ch ).
CHENR L. The Design of Partial Discharge On-Line Monitoring System for XLPE Power Cable [D]. Shanghai: Shanghai Jiaotong University, 2014(Ch). 10.3969/j.issn.1000-3886.2014.02.034
[9]
YANGJ J, ZHANGD M. Partial discharge phenomena due to electrical treeing in XLPE [C]//2006 IEEE Conference on Industrial Electronics and Applications. New York: IEEE Press, 2006: 1-6. DOI:10.1109/ICIEA.2006.257228 .
ZHOUL J, QIUQ P, CHENGR, et al. Propagation and partial discharge characteristics of electrical trees in XLPE cable under high temperatures and high gas pressures[J]. High Voltage Engineering, 2018, 44(5): 1421-1427. DOI:10.13336/j.1003-6520.hve.20180430004(Ch ).
GUO C X, ZHANG L, QIAN Y, et al. Current status of partial discharge detection and location techniques in XLPE power cable[J]. High Voltage Apparatus, 2009, 45(3): 56-60 (Ch).
LIUY S, CHENGD F, XIAL Z, et al. Partial discharge ultra-high frequency signal de-noising method based on single-channel blind source separation algorithm [J]. Transactions of China Electrotechnical Society,2018,33(23):5625-5636. DOI:10.19595/j.cnki.1000-6753.tces.180016(Ch ).
ZHOUK, HUANGY L, HEM, et al. Test method for electrical parameters of HFCT based on transmission coefficient[J]. High Voltage Apparatus, 2018,54(11):301-306. DOI:10.13296/j.1001-1609.hva.2018.11.044(Ch ).
HAN Y Z, GUO X Q, CHU Q. Research on failure detection method of high voltage cable terminal based on HFCT test technology of partial discharge [J]. Electrical Measurement & Instrumentation, 2016, 53(17): 60-64 (Ch).
JING X P, PENG X S, JIANG W, et al. Phase resolved partial discharge pattern recognition method for on-line cable condition monitoring system based on K-Means clustering[J]. High Voltage Engineering, 2012, 38(9): 2437-2446 (Ch).
LAIL Y, XUW J. Recognition of partial discharge faults using zero phase error filtering [J]. Electric Power Engineering Technology, 2018, 37(3): 67-71. DOI:10.19464/j.cnki.cn32-1541/tm.2018.03.012(Ch ).
ZHANGX B, TANGJ, PANC, et al. Research of partial discharge recognition based on deep belief nets[J]. Power System Technology, 2016, 40(10): 3272-3278.DOI:10.13335/j.1000-3673.pst.2016.10.048(Ch ).
JIAY Y, DENGM, LIY J, et al. Research on GIS partial discharge pattern recognition based on deep residual network[J]. High Voltage Apparatus, 2018,54(11):123-129. DOI:10.13296/j.1001-1609.hva.2018.11.017(Ch ).
ZHOUK, WUK, WANL, et al. Validity of inorganic nano-particles on the suppression of partial discharge in air gap defects for medium voltage cable terminations[J]. Transactions of China Electrotechnical Society, 2016, 31(22): 230-238. DOI:10.19595/j.cnki.1000-6753.tces.2016.22.027(Ch ).
LIUR. Study on Measurement and Diagnosis Technology of XLPE Cable Insulation Defects Based on Ultrasonic Method [D]. Xi’an: Northwestern Polytechnical University, 2015(Ch). 10.1142/9789814719391_0008
XUC H, CHENK X, MAJ, et al. Recognition of power loads based on deep belief network [J]. Transactions of China Electrotechnical Society,2019,34(19):4135-4142.DOI:10.19595/j.cnki.1000-6753.tces.181385(Ch ).
[34]
WILSOND R, MARTINEZT R. The general inefficiency of batch training for gradient descent learning [J]. Neural Networks,2003,16(10):1429-1451. DOI:10.1016/S0893-6080(03)00138-2 .
[35]
KONECNYJ, LIUJ, RICHTARIKP, et al. Mini-batch semi-stochastic gradient descent in the proximal setting [J]. IEEE Journal of Selected Topics in Signal Processing,2016,10(2):242-255. DOI:10.1109/JSTSP.2015.2505682 .
[36]
QIANQ, JINR, YIJ F, et al. Efficient distance metric learning by adaptive sampling and mini-batch stochastic gradient descent (SGD) [J]. Machine Learning,2015,99(3):353-372. DOI:10.1007/s10994-014-5456-x .
YANGF, WANGG J, PENGX S, et al. Partial discharge pattern recognition of high-voltage cables based on convolutional neural network [J]. Electric Power Automation Equipment, 2018, 38(5):123-128. DOI:10.16081/j.issn.1006-6047.2018.05.018(Ch ).
LIF. Partial Discharge Pattern Recognition of Mining Cable Based on Convolutional Neural Network [D]. Xi’an: Xi’an University of Science and Technology, 2020(Ch). 10.21436/inbom.18882202
WANGX R, YANGM Z, YANZ. Review and prospect of rejecting interference in partial discharge measurement of power equipment [J]. Power System Technology,2000,24(6):41-45. DOI:10.13335/j.1000-3673.pst.2000.06.010(Ch ).
YANG Q, ZENG H R, HUANG H,et al. Study on predischarge recognition of hidden dangers in transmission lines based on BP neural network [J]. Power Systems and Big Data,2020,23(3):47-54 (Ch).
[45]
VINCENTP, LAROCHELLEH, LAJOIEI, et al. Stacked denoising autoencoders: Learning useful represent ations in a deep network with a local denoising criterion [J]. Journal of Machine Learning Research,2010,11(12):3371-3408. DOI:10.1016/j.mechatronics.2010.09.004 .