本研究探討機器學習技術(Machine Learning Techniques)建立太陽光電系統直流側之故障分類器，所需之訓練資料取得不易之問題，應用MATLAB/Simulink產生故障分類器所需之大量訓練資料，克服在實際場域無法收集到大量之故障資料困難。因此，本研究首先利用MATLAB/Simulink模擬軟體，根據實際案場之佈置、太陽光電模組參數及變流器規格，建立完整之太陽光電系統模擬環境。其次，為驗證模擬資料之有效性，進一步設計正常運轉、遮陰故障、開路故障與短路故障四種不同運轉案例，並於實際場域進行實驗，量測實際之運轉資料，並與模擬資料進行比對分析，結果顯示模擬波形與實際量測資料波形樣態類似，且其穩定運轉之絕對平均誤差值與絕對平均誤差率落在工程可接受誤差範圍內。再者，利用模擬系統產生訓練資料，提供本研究選擇之基於卷積神經網路(Convolutional Neural Network)故障分類器訓練使用，故障分類器模擬測試準確率為87.29 %。最後，為評估運用模擬資料進行訓練之故障分類器實際性能，以實際之正常運轉、輕微遮陰故障、嚴重遮陰故障與短路故障四種案例進行比較分析。測試結果顯示實際故障分類器準確率為80.0 %，僅略低於模擬測試準確率，證實應用MATLAB/Simulink產生故障分類器所需訓練資料之可行性。

The study aimed to address the difficulty in obtaining the training data required for building a fault classifier for the DC side of a photovoltaic system through machine learning. MATLAB/Simulink was employed to generate a large quantity of training data for the classifier, thereby resolving the problem of failing to obtain sufficient fault data on site. First, MATLAB /Simulink was applied to construct a comprehensive simulation environment for the photovoltaic system according to the layout of the actual site of application, the module parameters of the system, and the inverter’s specifications. Subsequently, 4 operating scenarios, namely normal operation, shading fault, open circuit fault, and short circuit fault, were designed for the experiment at the site of application to verify the effectiveness of the simulation data. The actual operating data were measured and compared with the simulation data. The results revealed that the simulated waveforms were similar to the measured waveforms, and the mean absolute error and mean absolute percentage error of the steady state operation of 4 scenarios were within the acceptable range in engineering practice. The simulated system was then applied to generate training data for training the convolutional neural network-based fault classifier selected in this study, yielding a test accuracy of 87.29%. Finally, to evaluate the feasibility of using the simulation data to train the fault classifier performance, the data in 4 scenarios of actual measuring were used, namely normal operation, mild shading fault, severe shading fault, and short circuit fault, were compared. The results revealed a total accuracy of 80.0%, which was only slightly lower than the simulation data test accuracy. This confirmed MATLAB/Simulink as applicable for generating training data required for building a fault classifier.

[1] 台電系統歷年再生能源裝置容量，台灣電力股份有限公司，台北市，民國110年12月21日。取自：https://www.taipower.com.tw/tc/chart_m/b25_發電資訊_購入電力概況_歷年再生能源裝置容量.html
[2] M. K. Alam, F. Khan, J. Johnson and J. Flicker, “A Comprehensive Review of Catastrophic Faults in PV Arrays: Types, Detection, and Mitigation Techniques,” in IEEE Journal of Photovoltaics, Vol. 5, No. 3, pp. 982-997, May 2015.
[3] Vivienne Sze, Yu-Hsin Chen, Tien-Ju Yang, ahd Joel S. Emer, Efficient Processing of Deep Neural Networks, Morgan & Claypool, 2020.
[4] D. S. Pillai, F. Blaabjerg and N. Rajasekar, “A Comparative Evaluation of Advanced Fault Detection Approaches for PV Systems,” in IEEE Journal of Photovoltaics, Vol. 9, No. 2, pp. 513-527, March 2019.
[5] Z. Wang, S. McConnell, R. S. Balog and J. Johnson, “Arc fault signal detection - Fourier transformation vs. wavelet decomposition techniques using synthesized data,” 2014 IEEE 40th Photovoltaic Specialist Conference (PVSC), Denver, CO, USA, June 8-13, 2014, pp. 3239-3244.
[6] Katherine A. Kim, Gab-Su Seo, Bo-Hyung Cho, and Philip T. Krein, “Photovoltaic Hot-Spot Detection for Solar Panel Substrings Using AC Parameter Characterization,” IEEE Transactions on Power Electronics, Vol. 31, No. 2, pp. 1121-1130, February 2016.
[7] Wojciech Kreft, Estera Przenzak and Mariusz Filipowicz, “Photovoltaic chain operation analysis in condition of partial shading for systems with and without bypass diodes,” International Journal for Light and Electron Optics, Vol. 247, 167840, December 2021.
[8] W. Rezgui, N. K. Mouss, L. Mouss, M. D. Mouss, Y. Amirat and M. Benbouzid, “Smart diagnosis algorithm of the open-circuit fault in a photovoltaic generator,” 2015 3rd International Conference on Control, Engineering & Information Technology (CEIT), Tlemcen, Algeria, May 25-27, 2015, pp. 1-5.
[9] Y. Zhao, J. de Palma, J. Mosesian, R. Lyons and B. Lehman, “Line–Line Fault Analysis and Protection Challenges in Solar Photovoltaic Arrays,” in IEEE Transactions on Industrial Electronics, Vol. 60, No. 9, pp. 3784-3795, September 2013.
[10] Z. Yi and A. H. Etemadi, “Line-to-Line Fault Detection for Photovoltaic Arrays Based on Multiresolution Signal Decomposition and Two-Stage Support Vector Machine,” in IEEE Transactions on Industrial Electronics, Vol. 64, No. 11, pp. 8546-8556, November 2017.
[11] 邱勳宗，「基於啟發式之太陽能系統線對線故障診斷法」，國立臺灣科技大學碩士論文，民國109年7月。
[12] P. B. Quater, F. Grimaccia, S. Leva, M. Mussetta and M. Aghaei, “Light Unmanned Aerial Vehicles (UAVs) for Cooperative Inspection of PV Plants,” in IEEE Journal of Photovoltaics, Vol. 4, No. 4, pp. 1107-1113, July 2014.
[13] S. Roy, M. K. Alam, F. Khan, J. Johnson and J. Flicker, “An Irradiance-Independent, Robust Ground-Fault Detection Scheme for PV Arrays Based on Spread Spectrum Time-Domain Reflectometry (SSTDR),” in IEEE Transactions on Power Electronics, Vol. 33, No. 8, pp. 7046-7057, August 2018.
[14] Takumi Takashima, Junji Yamaguchi, Kenji Otani, Takashi Oozeki, Kazuhiko Kato, and Masayoshi Ishida, “Experimental studies of fault location in PV module strings,” Solar Energy Materials and Solar Cells, Vol. 93, No. 6-7, pp. 1079-1082, June 2009.
[15] Fauzan Hanif Jufri, Seongmun Oh, and Jaesung Jung, “Development of Photovoltaic abnormal condition detection system using combined regression and Support Vector Machine,” Energy, Vol. 176, pp. 457- 467, June 2019.
[16] Ye Zhao, Ling Yang, Brad Lehman, Jean-François de Palma, Jerry Mosesian, and Robert Lyons, “Decision tree-based fault detection and classification in solar photovoltaic arrays,” IEEE Applied Power Electronics Conference and Exposition (APEC), Orlando, USA, Feb. 5-9, 2012, pp. 93-99.
[17] Zhicong Chen, Fuchang Han, Lijun Wu, Jinling Yu, Shuying Cheng, Peijie Lin, and Huihuang Chen, “Random forest based intelligent fault diagnosis for PV arrays using array voltage and string currents,” Energy Conversion and Management, Vol. 178, pp. 250-264, December 2018.
[18] Hyeonah Park and Hyosung Kim, “PV cell modeling on single-diode equivalent circuit,” IECON 2013 - 39th Annual Conference of the IEEE Industrial Electronics Society, Vienna, Austria, Nov. 10-13, 2013, pp. 1845-1849.
[19] G. Farivar and B. Asaei, “Photovoltaic module single diode model parameters extraction based on manufacturer datasheet parameters,” 2010 IEEE International Conference on Power and Energy, Kuala Lumpur, Malaysia, Nov. 29 - Dec. 1, 2010, pp. 929-934.
[20] M. Hejri, H. Mokhtari, M. R. Azizian, M. Ghandhari and L. Söder, “On the Parameter Extraction of a Five-Parameter Double-Diode Model of Photovoltaic Cells and Modules,” IEEE Journal of Photovoltaics, Vol. 4, No. 3, pp. 915-923, May 2014.
[21] Norazlan Hashim, Zainal Salam, Dalina Johari and Nik Fasdi Nik Ismail, “DC-DC Boost Converter Design for Fast and Accurate MPPT Algorithms in Stand-Alone Photovoltaic System,” International Journal of Power Electronics and Drive System, Vol. 9, No. 3, pp. 1038-1050, September 2018.
[22] TEXAS INSTRUMENTS, “Basic Calculation of a Boost Converter's Power Stage,” Application Report - SLVA327C, November 2009, Revised January 2014.
[23] 蘇俊彥，「適用於太陽能發電系統之快速最大功率追蹤演算法研究」，國立臺灣科技大學碩士論文，民國111年1月。
[24] N. Femia, G. Petrone, G. Spagnuolo and M. Vitelli, “Optimization of perturb and observe maximum power point tracking method,” in IEEE Transactions on Power Electronics, Vol. 20, No. 4, pp. 963-973, July 2005.
[25] F. Liu, S. Duan, F. Liu, B. Liu and Y. Kang, “A Variable Step Size INC MPPT Method for PV Systems,” in IEEE Transactions on Industrial Electronics, Vol. 55, No. 7, pp. 2622-2628, July 2008.
[26] K. Ishaque, Z. Salam, M. Amjad and S. Mekhilef, “An Improved Particle Swarm Optimization (PSO)–Based MPPT for PV With Reduced Steady-State Oscillation,” in IEEE Transactions on Power Electronics, Vol. 27, No. 8, pp. 3627-3638, August 2012.
[27] D. A. Nugraha, K. L. Lian and Suwarno, “A Novel MPPT Method Based on Cuckoo Search Algorithm and Golden Section Search Algorithm for Partially Shaded PV System,” in Canadian Journal of Electrical and Computer Engineering, Vol. 42, No. 3, pp. 173-182, Summer 2019.
[28] Mukesh Kumar, Dr S.R.Kapoor, Rajkumar Nagar, Amit Verma, “Comparison between IC and Fuzzy Logic MPPT Algorithm Based Solar PV System using Boost Converter,” International Journal of Advanced Research in Electrical, Electronics and Instrumentation Engineering (IJAREEIE), Vol. 4, Issue 6, pp. 4927-4939, June 2015
[29] R. Ramaprabha, V. Gothandaraman, K. Kanimozhi, R. Divya and B. L. Mathur, “Maximum power point tracking using GA-optimized artificial neural network for Solar PV system,” 2011 1st International Conference on Electrical Energy Systems, Chennai, India, Jan. 3-5, 2011, pp. 264-268.
[30] N. Güler and E. Irmak, “MPPT Based Model Predictive Control of Grid Connected Inverter for PV Systems,” 2019 8th International Conference on Renewable Energy Research and Applications (ICRERA), Brasov, Romania, Nov. 3-6, 2019, pp. 982-986.
[31] A. M. Haque, S. Sharma and D. Nagal, “Proposed inverter topology and control circuit for the proposed grid connected photovoltaic system: Description and design specification,” 2016 International Conference on Electrical, Electronics, and Optimization Techniques (ICEEOT), Chennai, India, Mar. 3-5, 2016, pp. 549-554.
[32] Raef Aboelsaud, A. Ibrahim and Alexander G. Garganeev3, “Review of three-phase inverters control for unbalanced load compensation,” International Journal of Power Electronics and Drive System (IJPEDS), Vol. 10, No. 1, pp. 242-255, March 2019.
[33] Ahmed Abdalrahman, Abdalhalim Zekry and Ahmed Alshazly, “Simulation and Implementation of Grid-connected Inverters,” International Journal of Computer Applications, Vol. 60, No. 4, pp. 41-49, December 2012.
[34] 黃仲欽，「電機機械理論講義」，上課教材，國立臺灣科技大學，2019 年。
[35] 台灣電力股份有限公司，「台灣電力股份有限公司再生能源發電系統併聯技術要點 」，民國107年11月20日。
[36] 台灣電力股份有限公司，「電 力 系 統 諧 波 管 制 暫 行 標 準」，民國82年6月4日。
[37] Santiago Silvestre, Mario Aires da Silva, Aissa Chouder, Daniel Guasch, and Engin Karatepe, “New procedure for fault detection in grid connected PV systems based on the evaluation of current and voltage indicators,” Energy Conversion and Management, Vol. 86, pp. 241-249, May 2014.
[38] Tingting Pei, and Xiaohong Hao, “A Fault Detection Method for Photovoltaic Systems Based on Voltage and Current Observation and Evaluation,” Energies, Vol. 12, No. 9, pp. 1-16, May 2019.
[39] S. Lu, B. T. Phung and D. Zhang, “Study on DC series arc fault in photovoltaic systems for condition monitoring purpose,” 2017 Australasian Universities Power Engineering Conference (AUPEC), Melbourne, VIC, Australia, Nov. 19-22, 2017, pp. 1-6.
[40] Meetarani Tripathy, Manish Kumar and P.K. Sadhu, “Photovoltaic system using Lambert W function-based technique,” Solar Energy, Vol. 158, pp. 432-439, December 2017.
[41] Efstratios I. BatzelisEmail, Georgios Anagnostou, Chandan Chakraborty and Bikash C. Pal, “Computation of the Lambert W Function in Photovoltaic Modeling,” ELECTRIMACS 2019, Salerno, Italy, May. 21-23, 2019, pp. 583-595.