隨著科技不斷的發展，我們發現停電所帶來的影響力越來越大，而主要的原因往往來自於電網的問題，這也突顯出電網的可靠度需進一步提升 。尤其在電力系統中，氣體絕緣開關與電力電纜的角色不容忽視，高壓長期工作環境可能導致設備中的絕緣材料裂化，進而可能造成局部放電。局部放電的現象充滿多樣性，若能深入分析並理解其差異性，我們就可以提前預測不同設備的局部放電，進行預防性的維護和更換，進一步提升電網的可靠度。
本研究將進行瑕疵辨識診斷，包括對氣體絕緣開關的三種不同瑕疵，以及對電纜直線接頭的三種不同瑕疵，總共採納了六種瑕疵案例 進行實驗，並將量測到的局部放電資料進行相位分析圖譜與脈衝序列分析的轉換。採用圖像卷積的方式提取放電特徵，分別為三種不同的深度學習模型，CNN、ResNet18、MobileNet， 進行訓練與評估結果，後續利用類激活可視化來解釋深度學習模型黑盒子的問題，各模型的準確度皆達到95%以上，最後再透過集成式學習的投票法，提升模型整體效能。

With the trend of technological advancement, the harm caused by power outages is substantial, mostly due to problems in the power grid. This highlights the necessity for further improvement in the reliability of the power system. In the power system, gas-insulated switches (GIS) and power cables play a crucial role. Long-term operation under high voltage can cause insulation materials in the equipment to crack, potentially leading to partial discharges. If these partial discharges can be analyzed, preventative maintenance and replacement of equipment can be carried out, there by improving the reliability of the power grid.
This research will diagnose defects by identifying three different defects in GIS and three different defects in straight cable joints, for a total of six types of defects. The partial discharge data measured will be converted through phase analysis diagrams and pulse sequence analysis. Discharge features will be extracted using convolutional image processing, and three different deep learning models, CNN, ResNet18, and MobileNet, will be used for training and evaluation. Class Activation Mapping will be utilized to interpret the black-box problem of deep learning models, with each model achieving an accuracy rate of over 95%. Lastly, the overall model performance will be enhanced through an ensemble learning voting method.

[1] 經濟部，「112年4月能源統計月報」，民國一百一十二年。
[2] 行政院，「0303興達電廠事故至全台停電專案報告」，民國一百一十一年。
[3] Tanaka, T. "Aging of polymeric and composite insulating materials. Aspects of interfacial performance in aging." IEEE Transactions on dielectrics and electrical insulation 9.5 (2002): 704-716.
[4] Peng, Xiaosheng, et al. "Random forest based optimal feature selection for partial discharge pattern recognition in HV cables." IEEE Transactions on Power Delivery 34.4 (2019): 1715-1724.
[5] Abdel-Galil, T. K., et al. "Partial discharge pattern classification using the fuzzy decision tree approach." IEEE Transactions on Instrumentation and Measurement 54.6 (2005): 2258-2263.
[6] Umamaheswari, R., and R. Sarathi. "Identification of partial discharges in gas-insulated switchgear by ultra-high-frequency technique and classification by adopting multi-class support vector machines." Electric Power Components and Systems 39.14 (2011): 1577-1595.
[7] Sun, Shengya, et al. "Partial discharge pattern recognition of transformers based on the gray-level co-occurrence matrix of optimal parameters." Ieee Access 9 (2021): 102422-102432.
[8] Barrios, Sonia, et al. "Partial discharge classification using deep learning methods—Survey of recent progress." Energies 12.13 (2019): 2485.
[9] Lu, Yao, et al. "A GIS Partial Discharge Defect Identification Method Based on YOLOv5." Applied Sciences 12.16 (2022): 8360.
[10] Zemouri, Ryad, et al. "Deep convolutional variational autoencoder as a 2d-visualization tool for partial discharge source classification in hydrogenerators." IEEE Access 8 (2019): 5438-5454.
[11] Vaswani, Ashish, et al. "Attention is all you need." Advances in neural information processing systems 30 (2017).
[12] LeCun, Yann, et al. "Gradient-based learning applied to document recognition." Proceedings of the IEEE 86.11 (1998): 2278-2324.
[13] Krizhevsky, Alex, Ilya Sutskever, and Geoffrey E. Hinton. "Imagenet classification with deep convolutional neural networks." Communications of the ACM 60.6 (2017): 84-90.
[14] Simonyan, Karen, and Andrew Zisserman. "Very deep convolutional networks for large-scale image recognition." arXiv preprint arXiv:1409.1556 (2014).
[15] Szegedy, Christian, et al. "Going deeper with convolutions." Proceedings of the IEEE conference on computer vision and pattern recognition. 2015.
[16] He, Kaiming, et al. "Deep residual learning for image recognition." Proceedings of the IEEE conference on computer vision and pattern recognition. 2016.
[17] 張軒豪，應用局部放電脈衝序列分析與卷積神經網路於電纜接頭之絕緣狀態評估，碩士論文，臺北，國立臺灣科技大學電機工程系，民國111年。
[18] Howard, Andrew G., et al. "Mobilenets: Efficient convolutional neural networks for mobile vision applications." arXiv preprint arXiv:1704.04861 (2017).
[19] Lin, Min, Qiang Chen, and Shuicheng Yan. "Network in network." arXiv preprint arXiv:1312.4400 (2013).
[20] Zhou, Bolei, et al. "Learning deep features for discriminative localization." Proceedings of the IEEE conference on computer vision and pattern recognition. 2016.
[21] Standard, IEC60270. "High-voltage test techniques: partial discharge measurements." IEC-60270 (2000): 13-31.
[22] Eigner, Alexander, and Kay Rethmeier. "An overview on the current status of partial discharge measurements on AC high voltage cable accessories." IEEE Electrical Insulation Magazine 32.2 (2016): 48-55.
[23] Zhao, Xuefeng, et al. "Research on the defect development of cable accessories under AC voltage." 2017 1st International Conference on Electrical Materials and Power Equipment (ICEMPE). IEEE, 2017.
[24] 賴佳駿，基於卷積神經網路於交連聚乙烯電纜接頭之狀態診斷及瑕疵識別研究，碩士論文，臺北，國立臺灣科技大學電機工程系，民國110年。
[25] 鐘易樺，應用適應性類神經模糊推論系統於氣體絕緣開關瑕疵辨識之研究，博士論文，臺北，國立臺灣科技大學電機工程系，民國102年。
[26] Wang, Yanxin, et al. "Optimizing GIS partial discharge pattern recognition in the ubiquitous power internet of things context: A MixNet deep learning model." International Journal of Electrical Power & Energy Systems 125 (2021): 106484.
[27] Sandler, Mark, et al. "Mobilenetv2: Inverted residuals and linear bottlenecks." Proceedings of the IEEE conference on computer vision and pattern recognition. 2018.
[28] Khan, Qasim, et al. "Partial discharge detection and diagnosis in gas insulated switchgear: State of the art." IEEE Electrical Insulation Magazine 35.4 (2019): 16-33.
[29] Wang, Yan-Bo, et al. "Separating multi-source partial discharge signals using linear prediction analysis and isolation forest algorithm." IEEE Transactions on Instrumentation and Measurement 69.6 (2019): 2734-2742.