臺灣博碩士論文加值系統

在諧振接地配電網的接地故障中，間歇性電弧故障(intermittent arc faults, IAFs)的故障電流幅值小，且特徵不明顯。傳統方法先人工提取波形特徵，然後輸入到分類器中進行分類。通用故障信號的特徵提取一直以來是一個難點，分類器在現場故障波形的檢測上抗干擾性不夠強。為解決這些問題，提出了兩種實現間歇性電弧故障識別的方法。一種為Mallat觸發演算法和特徵融合的RNN(feature fusion multi-recurrent neural networks, FFM-RNN)新型識別方法，另一種為基於遴選網路(“for IAFs neuron elaboration net”, FINET)和主從因數(principal-subordinate factor, PSF)的新型檢測方法。第一種方法中，Mallat觸發演算法採用了簡化的小波變換演算法可以即時對電流故障波形進行觸發，而FFM-RNN則利用了機器學習強提取特徵的性能，通過增加的特徵融合層來取代傳統在時頻域上提取波形特徵的方式，改善了RNN的識別能力。第二種方法中，FINET彌補了傳統分類器的缺點，無需對波形在時域或頻域上的分解就可以提取故障信號特徵。PSF方法通過時間序列對FINET的分類結果作累加統計後作出最終決策，不僅提高了識別準確率還增加了抗擾能力。通過即時偵測零序電流波形，兩種方法都實現了對間歇性電弧故障的準確識別，第二種方法還使得間歇性電弧故障的選線、定位成為可能。結果表明，該方法具有良好的適應性，模擬和實測資料都能被準確檢測。

In the ground faults of the resonant grounding distribution network, the fault current amplitude of intermittent arc faults (IAFs) is small, and its characteristics are not obvious and thus difficult to identify. Therefore, the conventional method needs to extract the handcrafted waveform features before inputting them into the classifier. However, the feature extraction of the general fault signal is a challenging task, and the traditional classifier does not have strong anti-interference capability in the detection of field fault waveform feature. To solve these problems, two methods for identifying intermittent arc faults are proposed. One is the new RNN identification method for Mallat triggering algorithm and feature fusion multi-RNN(FFM-RNN), and the other is a novel detection method based on the combination of FINET (“For IAFs, Neuron Elaboration Net”) and principal-subordinate factor (PSF). In the first one, the Mallat triggering algorithm uses a simplified wavelet transform algorithm to trigger the fault waveform in real time, and the FFM-RNN utilizes the feature fusion layer to replace the way of handcrafted waveform features. This method improves the recognition ability of RNN. In the second method, without decomposition of the waveform in the time domain or frequency domain, the fault signal features extracted and recognized through the proposed FINET and PSF method significantly improves the recognition accuracy by statistically analyzing the classification results of FINET according to the time series. By detecting the zero-sequence current waveform in real time, both methods achieve accurate identification of intermittent arc faults, and the second method also makes it possible to select and locate intermittent arc faults. The results show that the method has good adaptability in real-world environment since both of the simulated and measured data can be accurately detected.

摘 要 i
ABSTRACT ii
誌 謝 iii
Table of Contents iv
List of Tables vi
List of Figures 7
Chapter 1. Introduction 9
Chapter 2. Fundamental Theories 13
2.1 單相接地故障暫態電流分析 13
2.2 單相接地故障穩態電流分析 15
2.3 間歇性電弧故障分析 17
Chapter 3. Software, physical simulation and field systems 20
3.1 軟體模擬配電網路不同架構及其詳細參數 20
3.2 軟體模擬事件模型 22
3.3 物理實驗平臺及其參數 23
3.4 軟體模擬、物理模擬與實際樣本 25
Chapter 4. Intermittent arc faults identification based on Mallat triggering algorithm and feature fusion multi-RNN 30
4.1 訓練樣本與測試樣本預處理 31
4.1.1 觸發演算法概述 31
4.1.2 VMD、EMD、VMDBF演算法概述 31
4.1.3 資料預處理過程 31
4.2 不同RNN架構及其分類過程 33
4.2.1 Simple-RNN、LSTM、GRU架構 33
4.2.2 M-RNN與Feature fusion layers 34
4.3 測試結果及其架構對比分析 35
Chapter 5. Intermittent arc faults identification based on FINET and principal-subordinate factor 38
5.1 訓練樣本與測試樣本預處理 38
5.2 FINET與PSF方法 40
5.2.1 FINET的設計原則和規則 40
5.2.2 FINET的訓練 41
5.2.3 PSF方法 42
5.3 結果與實踐 44
5.3.1 與其他方法的對比與分析 44
5.3.2 工程實踐與實施 45
Chapter 6. Conclusion 47
Reference 48
附 錄 一 56

[1] J. Q. Zhang, F. Yang, and M. F. Guo, “High impedance ground fault detection based on time-frequency characteristics and svm,” Electrical Engineering, vol. 19, no. 3, pp. 37–43, 2018.
[2] T. J. Yuan, F. Yin, D. Jian et al., Single phase arc grounding overvoltage in modern distribution network. China Electric Power Press, 2017.
[3] M. Adamiak, B. Russell, C. Benner, R. Dempsey, J. Parsick, and A. Depew, “Field experience with high-impedance fault detection relays,” in 2005/2006 IEEE/PES Transmission and Distribution Conference and Exhibition. IEEE, 2006, pp. 868–873.
[4] A. Ghaderi, H. L. Ginn III, and H. A. Mohammadpour, “High impedance fault detection: A review,” Electric Power Systems Research, vol. 143, pp. 376–388, 2017.
[5] M. Sarlak and S. Shahrtash, “High impedance fault detection using combination of multi-layer perceptron neural networks based on multi-resolution morphological gradient features of current waveform,” IET generation, transmission & distribution, vol. 5, no. 5, pp. 588–595, 2011.
[6] F. B. Costa, “Fault-induced transient detection based on real-time analysis of the wavelet coefficient energy,” IEEE transactions on power delivery, vol. 29, no. 1, pp. 140–153, 2013.
[7] A. Sedighi, “Classification of transient phenomena in distribution system using wavelet transform,” Journal of Electrical Engineering, vol. 65, no. 3, pp. 144–150, 2014.
[8] M.-R. Haghifam, A.-R. Sedighi, and O. Malik, “Development of a fuzzy inference system based on genetic algorithm for high-impedance fault detection,” IEE Proceedings-Generation, Transmission and Distribution, vol. 153, no. 3, pp. 359–367, 2006.
[9] W. Santos, F. Lopes, N. Brito, and B. Souza, “Highimpedance fault identification on distribution networks,” IEEE Transactions on Power Delivery, vol. 32, no. 1, pp. 23–32, 2016.
[10] M. Michalik, W. Rebizant, M. Lukowicz, S.-J. Lee, and S.-H. Kang, “High-impedance fault detection in distribution networks with use of wavelet-based algorithm,” IEEE Transactions on Power Delivery, vol. 21, no. 4, pp. 1793–1802, 2006.
[11] S. H. Mortazavi, Z. Moravej, and S. M. Shahrtash, “A hybrid method for arcing faults detection in large distribution networks,” International Journal of Electrical Power & Energy Systems, vol. 94, pp. 141–150, 2018.
[12] I. Baqui, I. Zamora, J. Mazon, and G. Buigues, “High ´ impedance fault detection methodology using wavelet transform and artificial neural networks,” Electric Power Systems Research, vol. 81, no. 7, pp. 1325–1333, 2011.
[13] H. Livani and C. Y. Evrenosoglu, “A fault classification ˘ and localization method for three-terminal circuits using machine learning,” IEEE Transactions on Power Delivery, vol. 28, no. 4, pp. 2282–2290, 2013.
[14] M. Mishra, P. Routray, and P. kumar Rout, “A universal high impedance fault detection technique for distribution system using s-transform and pattern recognition,” Technology and Economics of Smart Grids and Sustainable Energy, vol. 1, no. 1, p. 9, 2016.
[15] W. Sima, R. Ran, T. Yuan, A. Yao, L. Ma, and D. Yang, “Identification of arc grounding over-voltage using mathematical morphology transform,” Gaodianya Jishu/ High Voltage Engineering, vol. 36, no. 4, pp. 835–841, 2010.
[16] S. Gautam and S. M. Brahma, “Detection of high impedance fault in power distribution systems using mathematical morphology,” IEEE Transactions on Power Systems, vol. 28, no. 2, pp. 1226–1234, 2012.
[17] S. Samantaray, P. Dash, and S. Upadhyay, “Adaptive kalman filter and neural network based high impedance fault detection in power distribution networks,” International Journal of Electrical Power & Energy Systems, vol. 31, no. 4, pp. 167–172, 2009.
[18] M. F. Guo, X. D. Zeng, D. Y. Chen, and N. C. Yang, “Deep-learning-based earth fault detection using continuous wavelet transform and convolutional neural network in resonant grounding distribution systems,” IEEE Sensors Journal, vol. 18, no. 3, pp. 1291–1300, 2017.
[19] Y. Bengio, A. Courville, and P. Vincent, “Representation learning: A review and new perspectives,” IEEE transactions on pattern analysis and machine intelligence, vol. 35, no. 8, pp. 1798–1828, 2013.
[20] I. V. Tetko, D. J. Livingstone, and A. I. Luik, “Neural network studies. 1. comparison of overfitting and overtraining,” Journal of chemical information and computer sciences, vol. 35, no. 5, pp. 826–833, 1995.
[21] S. Hochreiter, Y. Bengio, P. Frasconi, J. Schmidhuber et al., “Gradient flow in recurrent nets: the difficulty of learning long-term dependencies,” 2001.
[22] C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed, D. Anguelov, D. Erhan, V. Vanhoucke, and A. Rabinovich, “Going deeper with convolutions,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2015, pp. 1–9.
[23] K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 770–778.
[24] S. Jegou, M. Drozdzal, D. Vazquez, A. Romero, and ´ Y. Bengio, “The one hundred layers tiramisu: Fully convolutional densenets for semantic segmentation,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, 2017, pp. 11–19.
[25] B. Lim and K. M. Lee, “Deep recurrent resnet for video super-resolution,” in 2017 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC). IEEE, 2017, pp. 1452–1455.
[26] J. Ma, X. D. Zhang, N. Yang, and Y. L. Tian, “Gesture recognition method integrating dense convolution and space conversion network,” Journal of Electronics and Information, vol. 40, no. 4, pp. 951–956, 2018.
[27] H. Qassim, A. Verma, and D. Feinzimer, “Compressed residual-vgg16 cnn model for big data places image recognition,” in 2018 IEEE 8th Annual Computing and Communication Workshop and Conference (CCWC). IEEE TRANSACTIONS ON , VOL. *, NO. *, * * 9 IEEE, 2018, pp. 169–175.
[28] L. N. Smith and N. Topin, “Deep convolutional neural network design patterns,” arXiv preprint arXiv:1611.00847, 2016.
[29] R. J. Xu, F. Yan, and Y. L. Pei, “Research on single-phase ground fault location method for 10kv distribution network,” Power Science and Engineering, vol. 26, no. 11, pp. 14–17, 2010.
[30] J. Liu, H. G. Li, and Z. G. Wang, “10kv single phase ground fault analysis,” Engineering construction and design, no. z1, pp. 47–49, 2017.
[31] T. Cui, X. Dong, Z. Bo, and A. Juszczyk, “Hilberttransform-based transient/intermittent earth fault detection in noneffectively grounded distribution systems,” IEEE Transactions on Power Delivery, vol. 26, no. 1, pp. 143–151, 2010.
[32] B. Wang, J. Z. Gen, and X. Z. Dong, “Analysis and detection of volt-ampere characteristics of high-resistance grounding faults in distribution network,” Chinese Society for Electrical Engineering, vol. 34, no. 22, pp. 3815–3823, 2014.
[33] Z. H. San, “Research on new flexible ground compensation line selection device,” Master’s thesis, China University of Mining, 2014.
[34] S. W. Shu, D. C. Huang, and J. J. Ruan, “Research progress on numerical simulation method of vacuum switch arc breaking process,” High voltage electrical appliance, vol. 50, no. 2, pp. 131–138, 2014.
[35] W. Tang, H. Zhao, and Y. J. Yue, “Simulation research on reactive power compensation of electric arc furnace based on tcr tsc,” Power, vol. 28, no. 2, 2012.
[36] L. J. Wang, S. L. Jia, Z. Q. Shi, and M. Z. Rong, “Dynamic model and simulation of vacuum arc magnetic fluid,” Chinese Society for Electrical Engineering, vol. 25, no. 4, pp. 113–118, 2005.
[37] H. W. Jin, L. S. Luo et al., “Fault arc modeling and arc grounding analysis of neutral point via high resistance grounding grid,” Electric Automation, vol. 36, no. 4, pp. 54–55, 2014.
[38] A. M. Cassie, “Arc rupture and circuit severity: a new theory,” CIGRE report, 1939.
[39] O. Mayr, “Beitrage zur theorie des statischen und des ¨ dynamischen lichtbogens,” Archiv fur Elektrotechnik ¨ , vol. 37, no. 12, pp. 588–608, 1943.
[40] Y. K. Li, S. B. Gao, and T. Wang, “Research on the principle of new out-of-phase short circuit protection for traction power supply,” Power system protection and control, no. 22, pp. 63–67, 2010.
[41] X. Ye, G. Moufa, C. Bin et al., “Modeling and simulation analysis of arc in distribution network,” Power System Protection and Control, vol. 43, no. 7, pp. 57–64, 2015.
[42] Q. Wang, Z. Liu, Y. l. Bai, and F. r. Tu, “Research on simulation model of traction power supply system based on pscad/emtdc,” Power system protection and control, no. 16, pp. 35–40, 2009.
[43] D. network automation technology, Distribution network automation technology. Mechanical Industry Press, 2012.
[44] A. L. Lin, M. F. Guo et al., “Flexibility fault generating device of physical simulation system for distribution network,” Electrical Engineering, no. 11, pp. 18–24, 2017.
[45] A. Krizhevsky, I. Sutskever, and G. E. Hinton, “Imagenet classification with deep convolutional neural networks,” in Advances in neural information processing systems, 2012, pp. 1097–1105.
[46] S. M. Chen, N. X. Qiu, H. Q. Zhao, R. A. Wang, and M. F. Guo, “Development of distributed single-phase-toground fault recorder in distribution networks,” Power System Protection and Control, no. 8, pp. 39–45, 2018.
[47] Z. Y. Zheng, X. Y. Cai et al., “Development of distributed feeder fault monitoring system in distribution network,” Electrical and Energy Efficiency Management Technology, no. 13, pp. 46–51, 2016.
[48] S. Mallat, “Zero-crossings of a wavelet transform,” IEEE Transactions on Information theory, vol. 37, no. 4, pp. 1019–1033, 1991.
[49] L. Hormander, ¨ The analysis of linear partial differential operators I: Distribution theory and Fourier analysis. Springer, 2015.
[50] M.-F. Guo, N.-C. Yang, and L.-X. You, “Wavelettransform based early detection method for short-circuit faults in power distribution networks,” International Journal of Electrical Power & Energy Systems, vol. 99, pp. 706–721, 2018.
[51] C. Lin, W. Gao, and M. Guo, “Discrete wavelet transform based triggering method for single-phase earth fault in power distribution systems,” IEEE Transactions on Power Delivery, 2019.
[52] K. Dragomiretskiy and D. Zosso, “Variational mode decomposition,” IEEE transactions on signal processing, vol. 62, no. 3, pp. 531–544, 2013.
[53] W. Xiaowei, G. Jie, W. Xiangxiang, S. Guobing, W. Lei, L. Jingwei, Z. Zhihui, and M. Kheshti, “High impedance fault detection method based on variational mode decomposition and teager-kaiser energy operators for distribution network,” IEEE Transactions on Smart Grid, 2019.
[54] Y. b. Xu and Z. x. Cai, “Application of variational mode decomposition and kl divergence in fault diagnosis of vibrating screen bearings,” Noise and vibration control, vol. 37, no. 4, pp. 160–165, 2017.
[55] N. E. Huang, Z. Shen, S. R. Long, M. C. Wu, H. H. Shih, Q. Zheng, N.-C. Yen, C. C. Tung, and H. H. Liu, “The empirical mode decomposition and the hilbert spectrum for nonlinear and non-stationary time series analysis,” Proceedings of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences, vol. 454, no. 1971, pp. 903–995, 1998.
[56] W. f. Chen, “Research on fault type identification method of distribution network feeder,” Master’s thesis, Fuzhou University, 2017.
[57] T. Mikolov, M. Karafiat, L. Burget, J. ´ Cernock ˇ y, and ` S. Khudanpur, “Recurrent neural network based language model,” in Eleventh annual conference of the international speech communication association, 2010.
[58] J. Schmidhuber, F. Gers, and D. Eck, “Learning non-regular languages: A comparison of simple recurrent networks and lstm,” Neural Computation, vol. 14, no. 9, pp. 2039–2041, 2002.
[59] K. Cho, B. van Merrienboer, C. Gulcehre, D. Bahdanau, F. Bougares, H. Schwenk, and Y. Bengio, “Learning phrase representations using rnn encoder–decoder for statistical machine translation,” in Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), 2014, pp. 1724–1734.
[60] T.-Y. Lin, P. Dollar, R. Girshick, K. He, B. Hariharan, ´ and S. Belongie, “Feature pyramid networks for object detection,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 2117– 2125.
[61] C. M. Biship, Pattern Recognition and Machine Learning (Information Science and Statistics). Springer, 2006.