Partial Discharge Pattern Recognition of Transformer Based on OS-ELM

被引：0
|
作者：
Zhang Q. [1 ]
Song H. [1 ]
Jiang Y. [2 ]
Chen Y. [3 ]
Sheng G. [1 ]
Jiang X. [1 ]
机构：
[1] Department of Electrical Engineering, Shanghai Jiao Tong University, Shanghai
[2] Electric Power Research Institute, Shanghai Municipal Electric Power Company of State Grid, Shanghai
[3] Electric Power Research Institute, Shandong Electric Power Company of State Grid, Jinan
来源：
Gaodianya Jishu/High Voltage Engineering | 2018年 / 44卷 / 04期
基金：
中国国家自然科学基金;
关键词：
OS-ELM; Pattern recognition; PD; Transformer; UHF;
D O I：
10.13336/j.1003-6520.hve.20180329011
中图分类号：
学科分类号：
摘要：
The traditional partial discharge (PD) pattern recognition algorithms have low recognition accuracies and slow recognition speed in practical engineering applications because of their limitations, including a large number of parameters to tune, the difficulty in optimizing parameters and low learning rates. Therefore, we proposed a PD pattern recognition method for transformer based on the Online Sequential - Extreme Learning Machine (OS-ELM) algorithm. OS-ELM is an online-learning and improved algorithm of Extreme Learning Machine (ELM), and a new type of Single-hidden Layer Feed-forward neural network (SLFN). Meanwhile, a lot of experimental data were obtained from real transformer in the high voltage laboratory on the PD experiments based on Ultra High Frequency (UHF) detection method. In addition, OS-ELM is analyzed and compared with ELM, Support Vector Machine (SVM) and BP neural network (BPNN) in both pattern recognition effect and performance aspects. The results show that the accuracy of OS-ELM is 5.2% and 23.2% higher than that of SVM and BPNN, respectively. When reducing the size of the training samples, the fluctuation of OS-ELM recognition results is significantly smaller than that of SVM and BPNN, showing better generalization ability. Besides, the training time of OS-ELM is only 0.031 2 s, far less than that of SVM and BPNN. Therefore, OS-ELM is more suitable for engineering applications of large data volume samples. © 2018, High Voltage Engineering Editorial Department of CEPRI. All right reserved.
引用
收藏
页码：1122 / 1130
页数：8
相关论文
共 22 条
  • [1] Shen Y., Ruan L., Xie Q., Et al., On-site application of partial discharge detection of transformer using very wide bandwidth pulse current method, High Voltage Engineering, 37, 4, pp. 937-943, (2011)
  • [2] Dai W., Tang X., High-frequency pulse current method applied in partial discharge of online detection of power transformer, Auto-Apply, 12, 1, pp. 94-96, (2014)
  • [3] Van-Bolhuis J.P., Gulski E., Smit J.J., Et al., Interpretation of recovery voltage measurements on power transformers, Eleventh International Symposium on High Voltage Engineering, pp. 90-93, (1999)
  • [4] Judd M.D., Cleary G.P., Bennoch C.J., Applying UHF partial discharge detection to power transformers, IEEE Power Engineering Review, 22, 8, pp. 57-59, (2002)
  • [5] Li J., Han X., Liu Z., Et al., Review on partial discharge measurement technology of electrical equipment, High Voltage Engineering, 41, 8, pp. 2583-2601, (2015)
  • [6] Wang G., Hao Y., Li Y., Present and furute development of partial discharge detection in power transformers, Advanced Technology of Electrical Engineering and Energy, 12, 2, pp. 52-57, (2001)
  • [7] Tang L., Hu Y., Wang H., Et al., Time-delay estimation of partial discharge UHF pulse signals based on interpolation cross-relation algorithm, High Voltage Engineering, 41, 10, pp. 3320-3325, (2015)
  • [8] Zhang G., Xu Y., Liu W., Et al., Sensitivity of the internal UHF coupler used for partial discharge detection in GIS, High Voltage Engineering, 42, 11, pp. 3683-3688, (2016)
  • [9] Chen J., The attenuation of UHF signals when passing through transformer oil-valve, Power System Technology, 38, 3, pp. 817-822, (2014)
  • [10] Wang C., Tang Z., Chang W., Et al., A method for anti-interference and multi-source discharge signal separation in ultra high-frequency partial discharge de-tection, Power System Technology, 36, 3, pp. 46-50, (2012)
123