近年因電力產業蓬勃發展，使電力需求增加，為此電力品質顯得更為重要，尤其輸配電系統的可靠度與穩定性又與變壓器有極大關係，故需經常追蹤變壓器的健康狀態，才得以確保輸電之穩定。現階段變壓器油中氣體檢測方法大多採取氣相層析法，主要分析出溶解於絕緣油中的氣體含量，並藉由氣體濃度或續增量作為判定故障的依據。此法除可兼顧經濟及效益面，更可從中取得最佳平衡。
本論文應用長短期記憶模型(long short-term memory, LSTM)與時間卷積網路(temporal convolution network, TCN)進行預測結果比較，及設計風險評估表，以利給定變壓器的健康狀態。在取樣油中可分析出各項氣體含量，包含氫氣(H2)、氧氣(O2)、氮氣(N2)、甲烷(CH4)、乙烷(C2H6)、乙炔(C2H2)、乙烯(C2H4)、一氧化碳(CO)、二氧化碳(CO2)等。藉由量測到的氣體濃度經過數據預處理後可降低取樣時的不確定性，並將得出的數據做為LSTM 及TCN 預測模型的輸入元素，而將需要預測的氣體做為輸出元素。再經由預測模型的評估指標進行比較，可發現各項氣體的預測性能皆不同。最後以本文所訂定的風險評估表做為最終審斷標準，從而給定變壓器的氣體在未來的健康狀態，且對於未來監測上可提早有效掌管運轉時可能發生的故障風險。

In recent years, due to the vigorous development of the power industry, the demand for power has increased, so the power quality has become more important, especially the reliability and stability of the power transmission and distribution system has a great relationship with the transformers, so it is necessary to track the health status of the transformers frequently to ensure the stability of power transmission. At present, gas detection methods in insulating oil mostly adopt gas chromatography (GC), which analyzes the gas dissolved in insulating oil, and uses the gas concentration or continuous increment to infer the potential fault. In addition to taking into account both economic and benefit aspects, this method can also achieve the best balance.
In this thesis, the long-short term memory model (LSTM) and temporal convolutional network (TCN) are used to compare the prediction results. Design a risk evaluation table to consult the health status of a given transformer. Various gas contents can be analyzed in the sampled oil, such as hydrogen, oxygen, nitrogen, methane, ethane, acetylene, ethylene, carbon monoxide, carbon dioxide, etc. In order to reduce the uncertainty of sampling, the gas data must be preprocessed, via preprocessed data is used as the input element of the LSTM and TCN prediction models, and the output element is the gas item that needs to be predicted. By comparing the evaluation indicators of the prediction models, it can be found that the prediction performance of each gas is different. Finally, the risk evaluation table established in this paper is used as the consult manner, so that the future health status of transformers is given, and the potential risks can be effectively controlled in advance for future monitoring.

[1] T. Stenkovski, N. Mojsoska, and B. Arapinoski, “Methods of analysis of dissolved gasses in transformer oil,” in Proc. 2022 57th International Scientific Conference on Information, Communication and Energy Systems and Technologies (ICEST), Ohrid, North Macedonia, Jun. 2022.
[2] M. S. Ali, A. H. A. Bakar, A. Omar, A. S. A. Jaafar, and S. H. Mohamed, “Conventional methods of dissolved gas analysis using oil-immersed power transformer for fault diagnosis: A review,” Electric Power Systems Research, vol. 216, Mar. 2023.
[3] IEEE Guide for the Interpretation of Gases Generated in Mineral Oil-Immersed Transformers, IEEE Standard C57.104-2019, Jun. 2019.
[4] Mineral Oil-Filled Electrical Equipment in Service–Guidance on the Interpretation of Dissolved and Free Gases Analysis, IEC Standard 60599, May 2022.
[5] Interpretation of Gas Analysis of in Service Transformers, ABNT Standard NBR 7274, May 2012.
[6] M. Chakraborty, N. Baruah, R. Sangineni, S. Kumar Nayak, and P. Kumar Maiti, “Dissolved gas analysis (DGA) of thermally aged blended transformer oil,” in Proc. 2020 IEEE Conference on Electrical Insulation and Dielectric Phenomena (CEIDP), East Rutherford, NJ, USA, Oct. 2020.
[7] Z. Ayalew, K. Kobayashi, S. Matsumoto, and M. Kato, “Dissolved gas analysis (DGA) of arc discharge fault in transformer insulation oils (ester and mineral oils),” in Proc. 2018 IEEE Electrical Insulation Conference (EIC), San Antonio, TX, USA, Jun. 2018.
[8] M. Duval and L. Lamarre, “The new duval pentagons available for DGA diagnosis in transformers filled with mineral and ester oils,” in Proc. 2017 IEEE Electrical Insulation Conference (EIC), Baltimore, MD, USA, Jun. 2017.
[9] N. S. Suhaimi, M. T. Ishak, M. F. M. Din, M. M. Ariffin, N. A. M. Amin, and M. H. A. Hamid, “Dissolved gases analysis of rice bran oil under thermal fault for transformer application,” in Proc. 2022 IEEE International Conference on Power and Energy (PECon), Langkawi, Kedah, Malaysia, Dec. 2022.
[10] L. Ouyang, F. Wang, X. Chen, and C. Song, “Research progress of the dissolved gas analysis in synthetic ester insulating oil under electrical faults,” in Proc. 2022 IEEE International Conference on High Voltage Engineering and Applications (ICHVE), Chongqing, China, Sept. 2022.
[11] O. E. Gouda, S. H. El‐Hoshy, and H. H. El‐Tamaly, “Proposed heptagon graph for DGA interpretation of oil transformers,” IET Generation, Transmission & Distribution, vol. 12, no.2, pp. 490–498, Jan. 2018.
[12] H. Sutikno, R. A. Prasojo, and Suwarno, “Integration of duval pentagon to the multi-method interpretation to improve the accuracy of dissolved gas analysis technique,” in Proc. 2021 IEEE International Conference on the Properties and Applications of Dielectric Materials (ICPADM), Johor Bahru, Malaysia, Jul. 2021.
[13] Z. A. Latiff, and M. F. M. Yousof, “Development of software for duval triangle and pentagon interpretation on transformer oil dissolved gas analysis,” Evolution in Electrical and Electronic Engineering, vol. 2, no. 2, pp. 467–473, No. 2021.
[14] K. Bacha, S. Souahlia, and M. Gossa, “Power transformer fault diagnosis based on dissolved gas analysis by support vector machine,” Electric Power Systems Research, vol. 83, no. 1, pp. 73–79, Feb. 2012.
[15] L. Chao and M. Lin, “Health assessment model of power transformer based on dissolved gas analysis by support vector machine,” in Proc. 2013 6th International Conference on Information Management, Innovation Management and Industrial Engineering, Xi’an, China, Nov. 2013.
[16] F. R. Souza and B. Ramachandran, “Dissolved gas analysis to identify faults and improve reliability in transformers using support vector machines,” in Proc. 2016 Clemson University Power Systems Conference (PSC), Clemson, SC, USA, Mar. 2016.
[17] A. Abu-Siada, S. Hmood, and S. Islam, “A new fuzzy logic approach for consistent interpretation of dissolved gas-in-oil analysis,” IEEE Transactions on Dielectrics and Electrical Insulation, vol. 20, no. 6, pp. 2343–2349, Dec. 2013.
[18] R. Palke and P. Korde, “Dissolved gas analysis (DGA) to diagnose the internal faults of power transformer by using fuzzy logic method,” in Proc. 2020 International Conference on Communication and Signal Processing (ICCSP), Chennai, India, Jul. 2020.
[19] E. J. Kadim, C. F. Hee, N. Azis, J. Jasni, S. A. Ahmad, and M. Z. A. A. Kadir, “Dissolved gas analysis of transformers based on rough set and fuzzy logic methods,” in Proc. 2015 IEEE Conference on Energy Conversion (CENCON), Johor Bahru, Malaysia, Oct. 2015.
[20] S. Seifeddine, B. Khmais, and C. Abdelkader, “Power transformer fault diagnosis based on dissolved gas analysis by artificial neural network,” in Proc. 2012 First International Conference on Renewable Energies and Vehicular Technology, Nabeul, Tunisia, Mar. 2012.
[21] J. F. Vidal and A. R. G. Castro, “Diagnosing faults in power transformers with variational autoencoder, genetic programming, and neural network,” IEEE Access, vol. 11, pp. 30529–30545, Mar. 2023.
[22] Z. Xing and Y. He, “Multimodal mutual neural network for health assessment of power transformer,” IEEE Systems Journal, early access, Jan. 2023.
[23] H. Schnittker, P. Werle, T. Münster, and M. Lottner, “Neural network for estimating the technical age of power transformers,” in Proc. 2022 9th International Conference on Condition Monitoring and Diagnosis (CMD), Kitakyushu, Japan, Nov. 2022.
[24] A. R. E. Soto, S. L. Lima, and O. R. Saavedra, “Incipient fault diagnosis in power transformers by DGA using a machine learning ANN - mean shift approach,” in Proc. 2019 IEEE International Autumn Meeting on Power, Electronics and Computing (ROPEC), Ixtapa, Mexico, Nov. 2019.
[25] M. U. Farooque, S. A. Wani, and S. A. Khan, “Artificial neural network (ANN) based implementation of Duval Pentagon,” in Proc. 2015 International Conference on Condition Assessment Techniques in Electrical Systems (CATCON), Bangalore, India, 2015, pp. 46–50
[26] A. M. Aciu, C. I. Nicola, M. Nicola, and M. C. Nițu, “Complementary analysis for DGA based on Duval methods and furan compounds using artificial neural networks,” Energies, vol. 14, no. 3, Jan. 2021.
[27] N. Verma and A. K. Chandel, “Incipient fault diagnosis of power transformer based on Duval Pentagon using medium neural network,” in Proc. 2022 1st International Conference on Sustainable Technology for Power and Energy Systems (STPES), SRINAGAR, India, Jul. 2022.
[28] Z. Jiayu, H. Huijuan, and S. Gehao, “Cleaning method of equipment fault database based on ISOMAP equidistant mapping clustering,” in Proc. 2022 9th International Conference on Condition Monitoring and Diagnosis (CMD), Kitakyushu, Japan, Nov. 2022.
[29] N. L. Z. Msomi and B. A. Thango, “Development of dissolved gas analysis-based fault identification system using machine learning with Google Colab,” in Proc. 2023 31st Southern African Universities Power Engineering Conference (SAUPEC), Johannesburg, South Africa, Jan. 2023.
[30] F. Ru, L. Zhang, X. Yang, H. Zou, Y. Lu, and X. Xu, “Transformer fault diagnosis based on dissolved gas analysis in oil and ensemble learning,” in Proc. 2022 IEEE 5th International Electrical and Energy Conference (CIEEC), Nangjing, China, May 2022.
[31] P. Chanchotisatien and C. Vong, “Feature engineering and feature selection for fault type classification from dissolved gas values in transformer oil,” in Proc. 2021 25th International Computer Science and Engineering Conference (ICSEC), Chiang Rai, Thailand, Nov. 2021.
[32] M. Duval and A. dePabla, “Interpretation of gas-in-oil analysis using new IEC publication 60599 and IEC TC 10 databases,” IEEE Electrical Insulation Magazine, vol. 17, no. 2, pp. 31–41, Mar.–Apr. 2001.
[33] S. Das, A. Paramane, S. Chatterjee, and U. M. Rao, “Accurate identification of transformer faults from dissolved gas data using recursive feature elimination method,” IEEE Transactions on Dielectrics and Electrical Insulation, vol. 30, no. 1, pp. 466–473, Feb. 2023.
[34] M. Badawi, S. A. Ibrahim, A. EL-Faraskoury, D. A. Mansour, and S. A. Ward, “A novel DGA oil interpretation approach based on combined techniques,” in Proc. 2022 23rd International Middle East Power Systems Conference (MEPCON), Cairo, Egypt, Dec. 2022.
[35] C. Hu, Y. Zhong, Y. Lu, X. Luo, and S. Wang, “A prediction model for time series of dissolved gas content in transformer oil based on LSTM,” in Proc. 2020 International Conference on Ubiquitous Power Internet of Things (UPIOT 2020) & 4th International Symposium on Green Energy and Smart Grid (SGESG 2020), Xi’an, China, Aug. 2020.
[36] L. Wang, T. Littler, and X. Liu, “Dynamic incipient fault forecasting for power transformers using an LSTM model,” IEEE Transactions on Dielectrics and Electrical Insulation, early access, Mar. 2023.
[37] J. Dai, H. Song, G. Sheng, and X. Jiang, “LSTM networks for the trend prediction of gases dissolved in power transformer insulation oil,” in Proc. 2018 12th International Conference on the Properties and Applications of Dielectric Materials (ICPADM), Xi’an, China, May 2018.
[38] Z. Xin, S. Wang, Y. Jiang, F. Wu, and C. Sun, “Prediction of dissolved gas in power transformer oil based on LSTM-GA,” in Proc. The Fifth International Conference on Energy Engineering and Environmental Protection, Xiamen, China, Nov. 2020.
[39] Y. Zhang, D. Liu, H. Liu, Y. Wang, Y. Wang, and Q. Zhu, “Prediction of dissolved gas in transformer oil based on SSA-LSTM model,” in Proc. 2022 9th International Conference on Condition Monitoring and Diagnosis (CMD), Kitakyushu, Japan, Nov. 2022.
[40] A. W. Mahrukh, G. X. Lian, and S. S. Bin, “Prediction of power transformer oil chromatography based on LSTM and RF model,” in Proc. 2020 IEEE International Conference on High Voltage Engineering and Application (ICHVE), Beijing, China, Sept. 2020.
[41] S. Das, A. Paramane, S. Chatterjee, and U. M. Rao, “Sensing incipient faults in power transformers using bi-directional long short-term memory network,” IEEE Sensors Letters, vol. 7, no. 1, Jan. 2023.
[42] D. Luo, J. Fang, H. He, W. J. Lee, Z. Zhang, H. Zai, W. Chen, and K. Zhang, “Prediction for dissolved gas in power transformer oil based on TCN and GCN,” IEEE Transactions on Industry Applications, vol. 58, no. 6, pp. 7818–7826, Nov.-Dec. 2022.
[43] P. Liu, C. Li, Z. He, D. Yu, Z. Xu, and M. Lei, “Probabilistic forecasting for dissolved gas concentrations in transformer oil based on the Bayesian temporal convolutional network,” in Proc. 2022 4th International Conference on Electrical Engineering and Control Technologies (CEECT), Shanghai, China, Dec. 2022.
[44] X. Zhou, T. Tian, N. He, Y. Ma, W. Liu, Z. Yan, Y. Luo, X. Li, and H. Ni, “Prediction method of dissolved gas in transformer oil based on firefly algorithm - random forest,” in Proc. 2022 Asia Power and Electrical Technology Conference (APET), Shanghai, China, Nov. 2022.
[45] J. Liu, Z. Zhao, Y. Zhong, C. Zhao, and G. Zhang, “Prediction of the dissolved gas concentration in power transformer oil based on SARIMA model,” Energy Reports, vol. 8, no. 5, pp. 1360–1367, Aug. 2022.
[46] G. Nandagopan, K. S. Beevi, A. S. K. Lekshmi, K. J. Vishnupriya, D. S. Kumar, S. Abhijith, and K. K. Rishika, “Online prediction of DGA results for intelligent condition monitoring of power transformers,” in Proc. 2022 IEEE International Conference on Power Electronics, Smart Grid, and Renewable Energy (PESGRE), Trivandrum, India, Jan. 2022.
[47] W. Zhang, Y. Zeng, Y. Li, and Z. Zhang “Prediction of dissolved gas concentration in transformer oil considering data loss scenarios in power system,” Energy Reports, vol. 9, no. 1, pp. 186–193, Mar. 2023.
[48] Q. Yang, Y. Cheng, N. He, X. Wu, W. Sha, and X. Zhou, “Trend analysis of dissolved gas in oil for transformer,” in Proc. 2021 International Conference on Advanced Electrical Equipment and Reliable Operation (AEERO), Beijing, China, Oct. 2021.
[49] arXiv. (2018). An empirical evaluation of generic convolutional and recurrent networks for sequence modeling. [Online]. Available: https://arxiv.org/pdf/1803.01271.pdf
[50] Bureau of Reclamation, Transformers: Basics, Maintenance and Diagnostics. Denver, Colorado, USA: US Department of the Interior, Apr. 2005.
[51] F. Pereira, F. Bezerra, S. Junior, J. Santos, I. Chabu, G. Souza, F. Micerino, and S. Nabeta, ‘‘Nonlinear autoregressive neural network models for prediction of transformer oil-dissolved gas concentrations,’’ Energies, vol. 11, no. 7, Jun. 2018, art. no. 1691.
[52] T. Manoj, C. Ranga, S. S. M. Ghoneim, U. Mohan Rao, and S. A. M. Abdelwahab, “Alternate and effective dissolved gas interpretation to understand the transformer incipient faults,” IEEE Transactions on Dielectrics and Electrical Insulation, early access, Jan. 2023.
[53] 供電單位變電設備維護手冊，台灣電力股份有限公司輸供電事業部，民國一百一十一年三月。
[54] 張晏承，「長短期記憶法與極限梯度提升法於太陽光電系統發電預測之比較」，國立臺灣科技大學碩士論文，2021年7月。
[55] 電力變壓器，大同股份有限公司，2014年10月。
[56] A. Savitzky and M. J. E. Golay, “Smoothing and differentiation of data by simplified least squares procedures,” Analytical Chemistry, vol. 36, no. 8, pp. 1627–1639, Jul. 1964.
[57] F. Yu and V. Koltun, “Multi-scale context aggregation by dilated convolutions,” in Proc. International Conference on Learning Representations (ICLR), San Juan, Puerto Rico, May 2016.
[58] N. Lelekakis, D. Martin, W. Guo, and J. Wijaya, “Comparison of dissolved gas-in-oil analysis methods using a dissolved gas-in-oil standard,” IEEE Electrical Insulation Magazine, vol. 27, no. 5, pp. 29–35, Sept. –Oct. 2011.
[59] A. Maher, D. E. A. Mansour, K. Helal, and R. A. A. Abd El Aal, “Dissolved gas analysis and dissipation factor measurement of mineral oil‐based nanofluids under thermal and electrical faults,” High Voltage, Mar. 2023.
[60] X. Zhou T. Tian, N. Liu, J. Bai, Y. Luo, N. He, P. Zhang, and J. Sun, “Fault diagnosis method of oil-immersed transformer based on the production pattern of free gas and dissolved gas in oil,” in Proc. 2022 Asia Power and Electrical Technology Conference (APET), Shanghai, China, Nov. 2022.
[61] A. M. Selva, N. Azis, M. F. M. Yousof, N. Sallehuddin, A. B. Tadam, and A. D. Saliang, “Off-line partial discharge measurement and localization for a 33/6.9 kV in-service transformer,” in Proc. 2022 9th International Conference on Condition Monitoring and Diagnosis (CMD), Kitakyushu, Japan, Nov. 2022.
[62] M. Wu, G. Wang, and H. Liu, “Research on transformer fault diagnosis based on smote and random forest,” in Proc. 2022 4th International Conference on Electrical Engineering and Control Technologies (CEECT), Shanghai, China, Nov. 2022.
[63] M. Duval and L. Lamarre, “The duval pentagon-a new complementary tool for the interpretation of dissolved gas analysis in transformers,” IEEE Electrical Insulation Magazine, vol. 30, no. 6, pp. 9–12, Nov. –Dec. 2014.
[64] Electrical4U. (2021). Buchholz Relay in Transformers | Buchholz Relay Operation and Principle. [Online]. Available: https://www.electrical4u.com/buchholz-relay-in-transformer-buchholz-relay-operation-and-principle/
[65] 童耀宗，「變壓器油取樣與油中氣體分析」，台電綜合研究所，2005年9月。
[66] IBM. What are recurrent neural networks? [Online]. Available: https://www.ibm.com/topics/recurrent-neural-networks
[67] Sepp Hochreiter and Jürgen Schmidhuber, “Long short-term memory,” Neural Comput, vol. 9 , no. 8, pp. 1735–1780, 1997.