本文應用二維小波轉換與類神經網路於24kV級模鑄式比壓器局部放電之圖譜辨識。主要包括比壓器局部放電訊號實際量測及應用圖譜辨識兩部分。實際量測之4種比壓器實驗模型，係由廠商配合製造有缺陷的模鑄式比壓器，而比壓器局部放電訊號之量測部分，乃是於屏蔽實驗室內，應用商用之局部放電檢測儀，測量上述4種實驗模型的三度空間之局部放電圖譜。圖譜辨識部分則由所量測之實驗數據，透過二維小波轉換進行特徵擷取，以當做倒傳遞類神經網路之訓練集，完整建構比壓器局部放電圖譜辨識系統。最後，再以原先量測之訓練集重疊加入隨機雜訊後，作為測試資料集，以驗證本文所提方法之正確性。研究結果顯示，在外加雜訊量為20pC時，系統之辨識成功率仍高達八成五，故本文值得鼓舞的結果，應足以提供電業預防設備絶緣裂化診斷之參考。

This thesis deals with a 2-D wavelet transform combined with artificial neural network to the partial discharge pattern recognition of 24kV cast-resin potential transformers. The research includes practical measurement from potential transformers and pattern recognition using partial discharge (PD) signals. Four experiment models of cast-resin potential transformers with artificial insulation defects were purposely manufactured by an electrical factory. Then, using a commercial PD detector, practical measurements of 3D patterns for the afore-mentioned four models were performed in a magnetically shielded laboratory. For increasing the recognition rate, the features of the 3D patterns obtained from partial discharge measurement are extracted by a 2-D wavelet transform. These features are subsequently used as the training data of a back-propagation neural network, which is used as the partial discharge based defect-diagnosis system. Lastly, this study randomly selects different levels of noise to distort the original measurements. These distorted data sets are then entered to the recongnition system for testing. Simulation results show that, at least 85% successful recognition rate can be achieved with 20pC noise added. These encouraging results present the effectiveness of the proposed approach to prevent system breakdown of utilities by early diagnosis of insulation deterioration.

[1] 成永紅，「電力設備絕緣檢測與診斷」，中國電力出版社，第1-5 頁。
[2] 吳瑞南、林育勳，「大型電力變壓器預防診斷量測技術介紹」，量測資訊，第83期，第12~14頁，2002年1月。
[3] 楊正光、林育勳、吳瑞南與張宏展，「高壓變壓器局部放電現場查驗方式之研究」，91年度經濟部能源研究發展基金計畫之期末報告。
[4] V. K. Agarwal, “The Mysteries of Multifactor Ageing,” IEEE Electrical Insulation Magazine, Vol. 11, No. 3, pp.37~43, May/June 1995.
[5] “High-Voltage Test Techniques-Partial Discharge Measurements,” IEC 60270, 2001.
[6] E. Gulski, “Digital Analysis of Partial Discharges,” IEEE Transactions on Dielectrics and Electrical Insulation, Vol. 2, No. 5, pp. 822-837, Oct. 1995.
[7] F. H. Kreuger, E. Gulski, and A. Krivda, “Classificaction of Partial Discharge,” IEEE Transactions on Electrical Insulation, Vol. 28, No. 6, pp. 917-931, 1993.
[8] A. Krivda, “Automated Recognition of Partial Discharge,” IEEE Transactions on Dielectrics and Electrical Insulation, Vol. 2, No. 5, October, 1995.
[9] Yuming Tu, Z. D. Wang, and P. A. Crossley, “Partial Discharge Pattern Recognition Based on 2-D Wavelet Transform and Neural Network Techniques,’’ IEEE Power Engineering Society Summer Meeting, vol.1, pp. 411-416, 2002.
[10] K. G. Kexiong, T. F. Li, and C. Wu, “PD Pattern Recognition for Stator Bar Models with Six Kinds of Characteristic Vectors Using BP Network,” IEEE Transactions on Electrical Insulation, Vol. 9, No. 3, pp. 381-389, 2002.
[11] 李建明、朱康，「高壓電氣設備試驗方法」，中國電力出版社，第二版，第298~322頁，2000年2月。
[12] 吳瑞南、林育勳，「局部放電電氣信號量測技術介紹」，量測資訊，第85期，第35~38頁，2002年5月。.
[13] 楊正光、林育勳，「局部放電技術應用於電力設備之預防診斷」，電機月刊，第134期，第200~206頁，2002年2月。
[14] 「局部放電測量及超聲波故障定位技術資料彙編(一)」，電力供應部武漢高壓研究所，1997年11月。
[15] D. A. Nattrass, ”Partial Discharge Measurement and Interpretaion,” IEEE Electrical Insulation Magazine May/June Vol.4, No.3, pp10-23, 1988.
[16] 台灣電力公司材料標準規範C034（92-02）高壓計費用變比器。
[17] 林育勳，「數位局部放電量測應用於模鑄式比流器」，博士論文，國立台灣科技大學電機所，民國93年。
[18] 「變比器」，中國國家標準CNS11437 C4435，第51~52頁，2001年。
[19] “Instrument Transformers Part1: Current Transformers,” IEC60044-1, pp. 29-31, 1996.
[20] “Operating Instructions,” Partial Discharge Detector TE571.
[21] M. C. Hu and I. S. Tsai, “ The inspection of fabric defects by using wavelet transform,” J. Text. Inst., Part 1,Vol. 91, No. 3,pp. 420-433,1999.
[22] 葉怡成，「類神經網路模式應用與實作」，儒林圖書，第五版，1997年10月。
[23] 焦成李，「神經網路系統理論」，儒林圖書，1991年。
[24] J. M. Zurada, “Introduction to Artificial Neural Systems,” 1992.
[25] 葉怡成，「應用類神經網路」，儒林圖書，第三版，2002年6月。
[26] 林昇甫、洪成安，「神經網路入門與圖樣辨識」，全華科技圖書。
[27] M. T. Hagan, H. B. Demuth, and M. Beale, “Neural Network Design,” 1996.
[28] M. Riedmiller and H. Braun, “A Direct Adaptive Method for Faster Backpropagation Learning: The RPROP Algorithm,” IEEE Conference on Electrical Insulation, Vol.1, pp. 586-591, 1993.
[29] 羅華強，「類神經網路－Matlab的應用」，清蔚科技，2001年9月。
[30] 胡昌華、張軍波、夏軍及張偉等，「基於MATLAB的系統分析與設計—小波分析」，西安電子科技大學出版社，2000年。
[31] 連國珍，「數位影像技術處理」，儒林圖書，2001年。
[32] C. M. Pun and M. C. Lee, “Rotation-invariant texture classification using a two-stage wavelet packet feature approach,” IEE Proceedings on Vision, Image and Signal Processing, Vol. 184, pp. 422-428, 2001.
[33] S. G. Mallat, “Wavelets for a Vision,” Proceeding of IEEE on Image Analysis and Interpretation, Vol. 84, No. 4, pp. 604-614, 1996.
[34] C. Yim, W.N. Klarquist, and A. C. Bovik, “Multiresolution feature extraction based on multifrequency decomposition,” IEEE International Conference on Systems, Man, and Cybernetics, Vol. 3, pp. 2384-2389, 1994.
[35] A. Grossman and J. Morlet. Decomposition of Hardy functions into square integrable wavelet of constant shape. SIAM J. of Math. Anal. 15(4):723-736, July 1984.
[36] 廖紹綱，「數位影像處理活動Matlab」，全華科技圖書，1999年。
[37] 「輔助小分析與應用」，電子工業出版社，2003年。
[38] 許晉維，「應用類神經網路於比流氣局部放電圖譜之辨識」，碩士論文，國立台灣科技大學電機所，民國九十二年。
[39] Z. Wang, C. Li, P. Peng, L. Ding, Y. Jia, W. Wang, and J. C. Wang, “Partial Discharge Recognition of Stator Winding Insulation Based on Artificial Neural Network,” IEEE Conference on Electrical Insulation, pp. 9-12, 2000.