基於機器學習之微弱交流電弧故障診斷設計｜國立臺灣科技大學博碩士論文系統

簡易檢索 / 詳目顯示

回結果列表

研究生：	沈育龍 Yu-Long Shen
論文名稱：	基於機器學習之微弱交流電弧故障診斷設計 Machine Learning-Based Design for Weak AC Arc Fault Detection
指導教授：	魏榮宗 Rong-Jong Wai
口試委員:	林法正 Faa-Jeng Lin 呂政修 Jenq-Shiou Leu 阮聖彰 Shanq-Jang Ruan
學位類別：	碩士 Master
系所名稱：	電資學院 - 電子工程系 Department of Electronic and Computer Engineering
論文出版年：	2021
畢業學年度：	109
語文別：	中文
論文頁數：	106
中文關鍵詞：	電弧、小波分析、奇異值分解、快速傅立葉變換、熵、支持向量機
外文關鍵詞：	Arc Fault, Wavelet Analysis (WA), Singular Value Decomposition (SVD), Fast Fourier Transform (FFT), Entropy, Support Vector Machine (SVM)
相關次數：	點閱：303 下載：0
分享至:	分享至facebook 分享至twitter

查詢本校圖書館目錄查詢臺灣博碩士論文知識加值系統勘誤回報

上一筆

本文提出了兩種基於機器學習的微弱交流電弧故障檢測設計。針對微弱電弧故障檢測的挑戰，本文首先提出通過電流值標準化(Current Amplitude Normalization)增強基於小波分析(Wavelet Analysis)的奇異值分解(Singular Value Decomposition)演算法。首先，採集從實驗平台產生的正常運行和微弱電弧故障的母線電流數據，並對採集的訊號進行電流值標準化和小波分析。其次，利用各層的小波係數構造相應的漢克爾(Hankel)矩陣，並通過奇異值分解獲得相關係數，繼而通過對小波分析奇異值分解係數的熵指數計算，確定交直流電流分量的座標，實現對無關分量的自動過濾。當負載、電流值或工作環境發生劇烈變化，依賴於手動設置的傳統電弧故障檢測演算法的效能將嚴重下降。本文所提演算法通過從同一系統的所有訊號中減去一個任意處理的正常訊號來降低信號底噪，而非手動調整，從而避免傳統電弧故障檢測演算法的困境。之後，再對訊號進行重構並提取特徵，用支持向量機(Support Vector Machine)做出最終診斷。
為尋求快速準確的交流弱弧故障檢測的有效方法，本文進一步提出基於漸進式奇異值分解(Progressive Singular-Value Decomposition)的非週期性電弧故障特徵提取演算法，並結合快速傅立葉轉換(Fast Fourier Transform)提取週期性電弧故障特徵。首先，採集正常狀態和電弧故障狀態的母線電流訊號並進行電流值標準化處理，然後由漸進式奇異值分解進一步分析，以檢測由奇異值表示的電弧故障較強的非週期性分量帶來的差異。為了提高檢測精度，採用快速傅立葉轉換來積累電弧故障引起的週期性變化，以提供更全面的特徵提取。另外，微弱電弧故障在剛發生時很難與正常訊號區分，導致假陰性是微弱電弧故障檢測中最常見的錯誤類型。為了解決這個難題，本文額外設計雙診斷窗口框架，對在大視窗中滑動的小窗口進行特徵提取和支持向量機診斷。如果在任何小視窗中檢測到發生電弧故障，則相應的大窗口便被診斷為處於電弧故障狀態，該框架在減少假陰性錯誤方面的有效性得到比較測試結果的支持。本文所提出的基於小波分析的電流值標準化增強奇異值分解演算法和基於漸進式奇異值分解特徵提取結合快速傅立葉轉換演算法的有效性通過各種實驗進行了驗證，實驗包括與有著不同工作電流值的單一負載或數個並聯負載分別進行串聯安裝的電弧產生裝置，產生串聯(拉動式電弧產生裝置)或並聯電弧故障(碳化路徑電弧產生裝置)。

This thesis presents two machine learning-based designs for weak AC arc fault detection. A wavelet-analysis-based singular-value-decomposition (WASVD) algorithm augmented by the current amplitude normalization (CAN) is proposed firstly to tackle the challenge of weak arc-fault detection. First, an experimental platform is prepared for producing bus-current data from normal operations and weak arc faults. Then, the CAN and the wavelet analysis (WA) are performed on collected signals. Moreover, the coefficients in each layer of the WA are exploited to construct a correspondent Hankel matrix for acquiring the WASVD coefficients by the singular value decomposition (SVD). In order to filter the irrelevant components automatically, the coordinates of AC/DC current components are located by the entropy index calculation on WASVD coefficients. The performances of conventional arc-fault detection algorithms with dependence on manual setup will be severely degraded if loads, current amplitudes, or working environments undergo drastic changes. The proposed algorithm avoids this dilemma by subtracting one arbitrary processed normal signal from all signals of the same system to reduce the noise floor instead of manual adjustment. In addition, signals are reconstructed to have features extracted for the support vector machine (SVM) to make the final diagnostic judgment.
To pursue a competent method for fast and accurate AC weak arc fault detection, a novel algorithm based on the progressive singular-value decomposition (PSVD) algorithm for obtaining singular values from a given signal recursively in order to extract non-periodic arc-fault features combined with the fast Fourier transform (FFT) to extract periodic arc-fault features is further proposed in this thesis. First, bus-current signals of the normal state and the arc-fault state are collected and normalized before being processed by the PSVD to detect the discrepancy brought by comparatively stronger arc-fault non-periodic components expressed by singular values. To enhance the detection accuracy, the FFT is incorporated for accumulating periodic variations caused by arc faults with the aim of providing more comprehensive feature extraction. Furthermore, weak arc faults are hard to distinguish from normal signals when they just started, therefore, false negative is the most common type of error in weak arc fault detection. In order to address this dilemma, a double diagnostic window frame is designed. A small window sliding within a big window undergoes feature extraction and the SVM diagnosis. If an arc fault is reported to have occurred in any of the small windows, the correspondent big window is diagnosed as in the arc-fault state. The effectiveness of this frame in reducing false-negative errors is supported by the comparison tests. The effectiveness of the proposed WASVD-CAN and PSVD-FFT algorithms is verified by various experiments including single load and parallel loads with different operating current amplitudes and an arc generator mounted in series with each load respectively generating series (rods-pulling arc generator) or parallel arc faults (carbonized-path arc generator).

中文摘要              I
Abstract    III
致謝             VI
Contents    VII
List of Figures     IX
List of Tables    XI
Chapter 1    Introduction    1
Chapter 2    Arc Fault and Experimental Platform     15
2.1     Arc Fault    15
2.2     Experimental Platform    18
2.3    Standard for Arc-Fault Circuit-Interrupters    20
2.4    Overview of Experimental Conditions    21
Chapter 3    Wavelet-Analysis-Based Singular-Value-Decomposition Algorithm via Current Amplitude Normalization    23
3.1     Current Amplitude Normalization for WASVD-CAN    23
3.2     Frequency Bands Overlap of Wavelet Analysis (WA)    24
3.3     Singular Value Decomposition (SVD) and Hankel Matrix Analysis of
Signals    26
3.4     WASVD Coefficients Processing    32
3.5    Components Filtering    36
3.6    Signal Reconstruction and Diagnosis    40
Chapter 4    Fast-Fourier-Transform Enhanced Singular-Value -Decomposition Algorithm in Double Diagnostic Window Frame                                    46
4.1     Current Amplitude Normalization for PSVD-FFT    46
4.2     Progressive Singular Value Decomposition (PSVD)    46
4.3     Enhancement with Fast Fourier Transform (FFT)    50
4.4     Double Diagnostic Window Frame and Overview    51
4.5     Support Vector Machine (SVM) for Small Diagnostic Window Labeling    54
4.6     SVM Kernel Verification    55
4.7     Design and Implementation Procedure    57
Chapter 5    Experiments and Comparisons    61
5.1     Results of Experiments and Comparison Tests for WASVD-CAN    61
5.1.1 Algorithms for Comparisons    61
5.1.2 Experiments    64
5.2     Results of Experiments and Comparison Tests for PSVD-FFT    68
5.2.1 Experimental Setup    68
5.2.2 Algorithms for Performance Comparisons    71
5.3     Research Direction of Feature Extraction Modifications    74
Chapter 6    Analytic Conclusions and Future Researches    77
6.1     Results Analysis for WASVD-CAN    77
6.2     Conclusion of WASVD-CAN and Future Research    80
6.3     Results Analysis for PSVD-FFT    81
6.4     Conclusion of PSVD-FFT and Future Research    84
References    86
Biography    93

List of Figures
Fig. 1.1        Design template of arc-fault detection methods.    12
Fig. 1.2        Mainframe of proposed methods.    12
Fig. 2.1        Demonstration of arc fault.    15
Fig. 2.2        VI lifecycle of general arc fault based on air discharge alone.    16
Fig. 2.3        Schematic diagram of experimental platform.    18
Fig. 2.4        Waveform demonstration of weak arc fault.    20
Fig. 2.5        Experiments categorization.    22
Fig. 3.1        Demonstration of current amplitude normalization (CAN).    23
Fig. 3.2        Band-pass filtering characteristics of db4 wavelets at different scales.    25
Fig. 3.3        Components of simulating signal.    31
Fig. 3.4        Entropy index calculation.    35
Fig. 3.5        WASVD coefficients of normal current before filtering.    38
Fig. 3.6        WASVD coefficients of arc-fault current before filtering.    38
Fig. 3.7        WASVD coefficients of normal current after filtering.    39
Fig. 3.8        WASVD coefficients of arc-fault current after filtering.    39
Fig. 3.9        Current waveforms of normal operations and arc faults after noise reduction.    40
Fig. 3.10    Design process of proposed WASVD-CAN algorithm.    41
Fig. 3.11    Single-case detection process of proposed WASVD-CAN algorithm.    41
Fig. 3.12    Comparative experiments on wavelet bases.    42
Fig. 3.13    Comparative experiments on WA decomposition layer numbers.    43
Fig. 3.14    Detection accuracy and time for different data window sizes.    43
Fig. 3.15    Implementation flowchart of proposed WASVD-CAN algorithm.    45
Fig. 4.1        Comparison experiments of PSVD layers.    50
Fig. 4.2        Demonstration of double diagnostic window frame.    51
Fig. 4.3        Accuracy comparison on big window sizes.    53
Fig. 4.4        Comparison tests on double diagnostic window frame.    54
Fig. 4.5    Accuracy comparison of SVM kernels.    56
Fig. 4.6    Algorithm design of proposed PSVD-FFT algorithm.    57
Fig. 4.7        Single-case detection process of proposed PSVD-FFT algorithm.    58
Fig. 4.8        Implementation flowchart of proposed PSVD-FFT algorithm.    60
Fig. 5.1        Voltage/current waveforms and data collection demonstration for every experiment.    67
Fig. 5.2        Display of abnormal arc-fault distortion from experiment 3.    75
Fig. 6.1        Summary of overall accuracy comparisons for different methods.    77
Fig. 6.2        Summary of overall accuracy comparisons.    82

List of Tables
Table 1.1        Research Comparisons of Previous Detection Methods.    7
Table 1.2        Research Comparisons of Feature Extraction Methods.    14
Table 3.1        Detection Accuracies at Different Sliding Window Sizes and Entropy Index Values.    36
Table 4.1        Accuracy Comparison of SVM Kernels.    55
Table 5.1        Pearson Correlation Analysis of SVM Features.    63
Table 5.2        Summary of Experiment Conditions.    66
Table 5.3      Comparisons of Detection Accuracies of Different Methods for Various
Load Combinations.    68
Table 5.4      Comparisons of Detection Time for Different Methods.    69
Table 5.5        Summary of Experiment Conditions.    70
Table 5.6      Comparisons of Detection Accuracies of Different Methods for Various
Load Combinations.    73
Table 5.7            Comparisons of Detection Time for Different Methods.    73

                                

[1] S. H. Motazarvi, Z. Moravej, and S. M. Shahrtash, “A hybrid method for arcing faults detection in large distribution networks”, Int. J. Electr. Power Energy Syst., vol. 94, pp. 141-150, January 2018.
[2] S. H. Mortazavi, Z. Moravej, and S. M. Shahrtash, “A searching based method for locating high impedance arcing fault in distribution networks”, IEEE Trans. Power Deliv., vol. 34, no. 2, pp, 438-447, April 2019.
[3] Z. Wang and R. S. Balog, “Arc-fault and flash signal analysis in DC distribution systems using wavelet transformation”, IEEE Trans. Smart Grid, vol. 6, no. 4, pp. 1955-1963, July 2015.
[4] X. Han, W. Sheng, S. Du, J. Su, and G. Liu, “Novel protection scheme for residual current device-based electric fault time detection and touch current identification”, IET Gener. Transm. Distrib., vol. 11, no. 10, pp. 2478-2488, July 2017.
[5] X. Luo, X. Wang, and M. Chen, “Tripping characteristics of residual current devices under nonsinusoidal currents”, IEEE Trans. Ind. Appl., vol. 47, no. 3, pp. 1515-1521, June 2011.
[6] J. Lezama, P. Schweitzer, E. Tisserand, J. B. Humbert, S. Weber, and P. Joyeux, “An embedded system for AC series arc detection by inter-period correlations of current”, Electr. Power Syst. Res., no. 129, pp. 227-234, September 2015.
[7] Y. Wang, Q. Li, F. Zhou, Y. Zhou, and X. Mu, “A new method with Hilbert transform and slip-SVD-based noise-suppression algorithm for noisy power quality monitoring”, IEEE Trans. Instrum. Meas., vol. 68, no. 4, pp. 987-1001, April 2019.
[8] Z. Zhang, Y. Nie, and W. J. Lee, “Approach of voltage characteristics modeling for medium-low-voltage arc-fault in short gaps”, IEEE Trans. Ind. Appl., vol. 55, no. 3, pp. 2281-2289, June 2019.
[9] Z. Radojevic, V. Terzija, G. Preston, S. Padmanabhan, and D. Novosel, “Smart overhead lines autoreclosure algorithm based on detailed fault analysis”, IEEE Trans. Power Deliv., vol. 4, no. 4, pp. 1829-1838, December 2013.
[10] V. V. Terzija and Z. M. Radojevic, “Numerical algorithm for adaptive autoreclosure and protection of medium-voltage overhead lines”, IEEE Trans. Power Deliv., vol. 19, no. 2, pp. 554-559, April 2004.
[11] M. Sedighizadeh, A. Rezazadeh, and N. I. Elkalashy, “Approaches in high impedance fault detection-a chronological review”, Advances in Electrical and Computer Engineering, vol. 10, no. 3, pp. 114-128, November 2010.
[12] G. W. Roscoe and T. F. Papallo, “Arc flash detection system, apparatus and method”, U.S. Patent 7,791, September 2017.
[13] A. Dong, Q. Yu, H. Lv, Q. Yu, and D. Zheng, “Study on electric arc's on-line detection and warning system for low voltage distribution box”, in Proc. 2nd Int. Conf. Advanced Computer Control (ICACC), Shenyang, China, 2010, pp. 253-256.
[14] G. Bao, R. Jiang, and X. Gao, “Novel series arc-fault detector using high-frequency coupling analysis and multi-indicator algorithm”, IEEE Access, vol. 7, pp. 92161-92170, July 2019.
[15] G. Liu, Y. Cao, Y. Liu, and Z. Liu, “A survey on arc-fault detection and wire fault location for aircraft wiring system”, SAE Int. J. Aerosp., vol. 2008-01-2870, pp. 903-914, January 2008.
[16] T. S. Sidhu and Z. Xu, “Detection of incipient faults in distribution underground cables”, IEEE Trans. Power Deliv., vol. 25, no. 3, pp. 1363-1371, July 2010.
[17] G. Artale, A. Cataliotti, V. Cosentino, and G. Privitera, “Experimental characterization of series arc faults in AC and DC electrical circuits”, in Proc. Int. Instrum. Meas. Technol. Conf. (I2MTC), Montevideo, Uruguay, 2014, pp. 1015-1020.
[18] W. Zhang, Y. Jing, and X. Xiao, “Model-based general arc fault detection in medium-voltage distribution lines”, IEEE Trans. Power Deliv., vol. 31, no. 5, pp. 2231-2241, October 2016.
[19] Q. Lu, Z. Ye, M. Su, Y. Li, Y. Sun, and H. Huang, “A DC series arc fault detection method using line current and supply voltage”, IEEE Access, vol. 8, pp. 10134-10146, January 2020.
[20] S. Liu, L. Dong, X. Liao, X. Cao, X. Wang, and B. Wang, “Application of the variational mode decomposition-based time and time–frequency domain analysis on series DC arc fault detection of photovoltaic arrays”, IEEE Access, vol. 7, pp. 126177-126190, September 2019.
[21] H. P. Park and S. Chae, “DC series arc fault detection algorithm for distributed energy resources using arc fault impedance modeling”, IEEE Access, vol. 8, pp. 179039-179046, September 2020.
[22] A. F. Ilman and Dzulkiflih, “Low voltage series arc fault detecting with discrete wavelet transform”, in Proc. Int. Conf. Applied Engineering (ICAE), Batam, Indonesia, 2018, pp. 1-5.
[23] X. Yao, L. Herrera, S. Ji, K. Zou, and J. Wang, “Characteristic study and time-domain discrete wavelet-transform based hybrid detection of series DC arc-faults”, IEEE Trans. Power Electron., vol. 29, no. 6, pp. 3103-3115, June 2014.
[24] S. Govindarajan, J. Subbaiah, A. Cavallini, K. Krithivasan, and J. Jayakumar, “Development of Hankel-SVD hybrid technique for multiple noise removal from PD signature”, IET Sci. Meas. Technol., vol. 13, no. 8, pp. 1075-1084, October 2019.
[25] J. Yao, Q. Chen, and D. Wu, “Traveling wave fault location based on wavelet and improved singular value difference spectrum”, in Proc. Int. Conf. Circuits, Devices Syst. (ICCDS), Chengdu, 2017, pp. 141-145.
[26] K. Yang, R. Zhang, J. Yang, J. Du, S. Chen, and R. Tu, “Series arc-fault diagnostic method based on fractal dimension and support vector machine”, Trans. China Electrotechnical Society, vol. 31, no. 2, pp. 10-77, January 2016.
[27] A. Ghaderi, H. A. Mohammadpour, H. Ginn, and Y. J. Shin, “High-impedance fault detection in the distribution network using the time-frequency-based algorithm”, IEEE Trans. Power Deliv., vol. 30, no. 3, pp. 1260-1268, October 2014.
[28] N. Qu, J. Zuo, J. Chen, and Z. Li, “Series arc fault detection of indoor power distribution system based on LVQ-NN and PSO-SVM”, IEEE Access, vol. 7, pp. 184020-184028, December 2019.
[29] I. Baqui, I. Zamora, J. Mazón, and G. Buigues, “High impedance fault detection methodology using wavelet transform and artificial neural network”, Int. J. Electr. Power Energy Syst., vol. 81, pp. 1325-1333, February 2011.
[30] I. M. Karmacharya and R. Gokaraju, “Fault location in ungrounded photovoltaic system using wavelets and ANN”, IEEE Trans. Power Deliv., vol. 33, no. 2, pp. 549-558, April 2018.
[31] W. Li, Y. Liu, Y. Li, and F. Guo, “Series arc fault diagnosis and line selection method based on recurrent neural network”, IEEE Access, vol. 8, pp. 177815-177822, September 2020.
[32] N. L. Georgijevic, M. V. Jankovic, S. Srdic, and Z. Radakovic, “The detection of series arc-fault in photovoltaic systems based on the arc current entropy”, IEEE Trans. Power Electron., vol. 31, no. 8, pp. 5917-5930, August 2016.
[33] W. Li, Y. Fan, W. Liu, M. Xin, H. Wang, and Q. Jin, “A self-adaptive process mining algorithm based on information entropy to deal with uncertain data”, IEEE Access, vol. 7, pp. 131681-131691, September 2019.
[34] P. Qi, S. Jovanovic, J. Lezama, and P. Schweitzer, “Discrete wavelet transform optimal parameters estimation for arc fault detection in low-voltage residential power networks”, Electric Power Systems Research, vol. 143, pp. 130-139, February 2017.
[35] J. Su and Z. Xu, “Multi-variable method for arc-fault diagnosis based on EMD and PNN”, Electric Power Automation Equipment, vol. 39, no. 4, pp. 106-113, April 2019.
[36] M. Ahmadi, H. Samet, and T. Ghanbari, “A new method for detecting series arc fault in photovoltaic systems based on the blind-source separation”, IEEE Trans. Ind. Electron., vol. 67, no. 6, pp. 5041-5049, June 2020.
[37] R. Campbell, “Home electrical fires”, NFPA Research, Quincy, MA, USA, Mar. 2019.
[38] Y. Liu et al., "Feature analysis in time-domain and fault diagnosis of series arc fault", in Proc. 2017 IEEE Holm Conf. Electrical Contacts, Denver, CO, USA, 2017, pp. 306-311.
[39] R. D. Telford, S. Galloway, B. Stephen, and I. Elders, "Diagnosis of series DC arc faults-A machine learning approach", IEEE Trans Indus. Inform., vol. 13, no. 4, pp. 1598-1609, Aug. 2017.
[40] L. Zhao, Y. Zhou, K. Chen, S. Rau and W. Lee, "High-speed arcing fault detection: using the light spectrum", IEEE Ind. Appl. Mag., vol. 26, no. 3, pp. 29-36, June 2020.
[41] G. Artale, A. Cataliotti, V. Cosentino, and G. Privitera, “Experimental characterization of series arc faults in AC and DC electrical circuits”, in Proc. Int. Instrum. Meas. Technol. Conf. (I2MTC), Montevideo, Uruguay, 2014, pp. 1015-1020.
[42] K. Zeng, L. Xing, Y. Zhang and L. Wang, "Characteristics analysis of AC arc fault in time and frequency domain," in Proc. 2017 Prognostics and System Health Management Conf. (PHM-Harbin), Harbin, China, 2017, pp. 1-5.
[43] H. Gao, Z. Wang, A. Tang, C. Han, F. Guo and B. Li, "Research on series arc fault detection and phase selection feature extraction method", IEEE Trans. Instrum. Meas., vol. 70, pp. 1-8, May 2021.
[44] Y. -L. Shen and R. -J. Wai, "Wavelet-analysis-based singular-value-decomposition algorithm for weak arc fault detection via current amplitude normalization", IEEE Access, vol. 9, pp. 71535-71552, May 2021.
[45] Y. Zhang and L. Wang, "The research on characteristics of DC arc fault based on FRFT," in Proc. 2018 20th European Conf. Power Electronics and Applications (EPE'18 ECCE Europe), Riga, Latvia, 2018, pp. 1-6.
[46] J. Huang, X.G. Hu, and F. Yang, “Support vector machine with genetic algorithm for machinery fault diagnosis of high voltage circuit breaker”, Measurement, vol. 44, no. 6, pp. 1018-1027, July 2011.
[47] N.T. Huang, H.J. Chen, S.X. Zhang, G.W. Cai, W.G. Li, D.G. Xu, and L.H. Fang, “Mechanical fault diagnosis of high voltage circuit breakers based on wavelet time-frequency entropy and one-class support vector machine”, Entropy, vol. 18, no. 1, pp. 7-23, Dec. 2015.
[48] W. Gao, S. -P. Qiao, R. -J. Wai and M. -F. Guo, "A newly designed diagnostic method for mechanical faults of high-voltage circuit breakers via SSAE and IELM", IEEE Trans. Instrum. Meas., vol. 70, pp. 1-13, July 2021.
[49] “Arc fault”. Accessed: June 8, 2021. [Online]. Available: https://en.wikipedia.org/wiki/Arc_fault
[50] Z. Xu, “Electrical arc fault theories” in Basics of Electrical Theory, 1st ed. China: Machinery Industry Press, 2014, pp. 112-181.
[51] “Arc in Circuit Breaker | Electric Arc”, Aug. 23, 2021. [Online]. Available: https://www.technicalbookspdf.com/arc-in-circuit-breaker-electric-arc
[52] “What is an AFCI: what you need to know”, April 15, 2017. [Online]. Available: https://rsbelectricalinc.com/blog/2017/4/15/what-is-an-afci-what-you-need-to-know
[53] UL Standard for Arc-Fault Circuit-Interrupters, UL 1699, 2008.
[54] GB/T Standard for Arc-Fault Detection Devices, GB/T 31143, 2014.

全文公開日期 2026/10/04 (校內網路)
全文公開日期 2026/10/04 (校外網路)
全文公開日期 2026/10/04 (國家圖書館：臺灣博碩士論文系統)

簡易檢索 / 詳目顯示

相關論文