隨著全球科技進步和能源轉型的持續推進，變壓器是現今能源系統中的關鍵角色之一，特別是城市化和基礎設施擴張以及再生能源裝機增加，都使得變壓器市場需求不斷增長，為了擴大產能、提高生產效率和降低成本，製造變壓器的傳統工廠勢必需要智慧化轉型。而人工智慧（Artificial Intelligence, AI）也已經成為當今工業生產中不可或缺的關鍵技術，在人工智慧的各個分支中，機器學習（Machine Learning, ML）更為重要，它可以透過演算法對收集到的數據進行分類、預測和模型訓練，在多個方面對生產流程進行優化，從而提高生產連續性和效率還能有效降低整體運營成本。
　　本論文以某工廠在進行變壓器製程為例，用繞線時所收集的張力值數據進行分析與處理，在假設變壓器結構、材料、製造方式都一樣的條件下，分析張力值對阻抗電壓的影響關係，並且在數據量少和特徵不明顯的情況中導入預測模型，利用機器學習中的多元線性迴歸（Multiple Linear Regression, MLR），以及深度學習中的人工神經網路(Artificial Neural Network, ANN)、一維卷積神經網路（1D-Convolutional Neural Network, 1D-CNN），比較各模型在此研究中的表現，再藉由表現最好的模型建置預測系統，將有助於減少不良品的發生，同時降低生產成本並提升生產效率之目的。

With the global advancement of technology and energy transition, transformers are critical in today's energy systems. The rising demand due to urbanization, infrastructure expansion, and renewable energy growth necessitates the intelligent transformation of traditional transformer manufacturing plants to increase capacity, efficiency, and cost-effectiveness. Artificial Intelligence (AI), especially Machine Learning (ML), plays a key role by optimizing production processes through data classification, prediction, and model training.
　　This thesis uses data collected from the winding tension values during the manufacturing process of distribution transformers in a specific factory as an example. Assuming the same transformer architecture, materials, and manufacturing methods, the relationship between tension values and impedance voltage is analyzed. In situations with limited data and unclear features, predictive models are introduced. The study employs Multiple Linear Regression (MLR) from machine learning and Artificial Neural Networks (ANN) and 1D-Convolutional Neural Networks (1D-CNN) from deep learning to compare the performance of each model in this research. The best-performing model is then used to build a predictive system, which will help reduce defects and rework, lower production costs, and improve production efficiency.

[1] F. T. Y. Wang, Y. Zuo, M. Zhang, Q. Qi, "Digital-Twin-Enhanced Quality Prediction for the Composite Materials," Engineering pp. pp. 23-33, 22 (2023).
[2] C.-H. Lee, C.-L. Liu, A. J. C. Trappey, J. P. T. Mo, and K. C. Desouza, "Understanding digital transformation in advanced manufacturing and engineering: A bibliometric analysis, topic modeling and research trend discovery," Advanced Engineering Informatics, vol. 50, p. 101428, 2021/10/01/ 2021, doi: https://doi.org/10.1016/j.aei.2021.101428.
[3] M. I. S. P. Ltd, Global Power Transformer Market Size, Share & Industry Trends Analysis Report By Insulation, By Application, By Rating, By Phase, By Core, By Regional Outlook and Forecast, 2022 - 2028. February 2023.
[4] 任浩楠, 「变压器绕线机张力控制系统的设计与研究」, 哈尔滨理工大学硕士学位论文, 2021年6月.
[5] R. Bevans. "Multiple Linear Regression | A Quick Guide (Examples)." https://www.scribbr.com/statistics/multiple-linear-regression/ (accessed.
[6] D. O’Leary and J. Kubby, "Feature Selection and ANN Solar Power Prediction," Journal of Renewable Energy, vol. 2017, p. 2437387, 2017/11/08 2017, doi: 10.1155/2017/2437387.
[7] X. Yang, Y. Ye, X. Li, R. Lau, X. Zhang, and X. Huang, "Hyperspectral Image Classification With Deep Learning Models," IEEE Transactions on Geoscience and Remote Sensing, vol. PP, pp. 1-16, 04/17 2018, doi: 10.1109/TGRS.2018.2815613.
[8] H. C. J. e. al., "Prediction for Energy Demand Using 1D-CNN and Bidirectional LSTM in Internet of Energy," Journal of IKEEE, vol. vol. 23, no. 1, pp. pp. 134–142, Mar. 2019.
[9] S. W. T. P. Limited. https://www.synthesiswinding.com/ (accessed.
[10] 運泰電機有限公司. https://www.euntay.com.tw/motion-control_system00.html#a (accessed.
[11] 紹凱動能科技股份有限公司, "產品介紹." [Online]. Available: http://www.sktc-tech.com.tw/.
[12] 上海今景电气科技发展有限公司. "变压器绕线机适用线径与安全张力." https://bbs.co188.com/thread-1417520-1-1.html (accessed.
[13] A. Semiconductor. "Transformer vs Converter vs Inverter: What are Differences." https://avaqsemi.medium.com/transformer-vs-converter-vs-inverter-what-are-differences-d09d0bc83fe1 (accessed.
[14] 明明科技論. "一文介紹變壓器器身結構（鐵心、繞組、引線篇）." https://read01.com/Nok83e0.html (accessed.
[15] E. F. M. W. LLC. "Formvar – AL." https://essexfurukawa.com/products/formvar-al/ (accessed.
[16] 華城電機股份有限公司. https://www.fortune.com.tw/tw/ (accessed.
[17] 三江電機企業股份有限公司. https://www.sanjiang-electric.com/ (accessed.
[18] 士林電機廠股份有限公司. https://www.seec.com.tw/index.aspx?SiteID=10 (accessed.
[19] S. Wang, S. Wang, N. Zhang, D. Yuan, and H. Qiu, "Calculation and Analysis of Mechanical Characteristics of Transformer Windings Under Short-Circuit Condition," IEEE Transactions on Magnetics, vol. 55, no. 7, pp. 1-4, 2019, doi: 10.1109/TMAG.2019.2898183.
[20] W. W. Shao, L. J. Chen, C. L. Pan, Y. B. Liu, and Z. H. Feng, "Power density of piezoelectric transformers improved using a contact heat transfer structure," IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 59, no. 1, pp. 73-81, 2012, doi: 10.1109/TUFFC.2012.2157.
[21] KevinLuo. "深度學習-各種新優化器介紹-LookAhead&Ranger&LARS." https://kilong31442.medium.com/%E6%B7%B1%E5%BA%A6%E5%AD%B8%E7%BF%92-%E5%90%84%E7%A8%AE%E6%96%B0%E5%84%AA%E5%8C%96%E5%99%A8%E4%BB%8B%E7%B4%B9-lookahead-ranger-lars-830ca2250dd5 (accessed.
[22] D. J. Santry, "Demystifying Deep Learning: An Introduction to the Mathematics of Neural Networks," 2023-12-12.
[23] X. Chen, F. Kopsaftopoulos, Q. Wu, H. Ren, and F.-K. Chang, "A Self-Adaptive 1D Convolutional Neural Network for Flight-State Identification," Sensors, vol. 19, no. 2, doi: 10.3390/s19020275.
[24] S. Hansun, "A new approach of moving average method in time series analysis," in 2013 Conference on New Media Studies (CoNMedia), 27-28 Nov. 2013, pp. 1-4, doi: 10.1109/CoNMedia.2013.6708545.