臺灣博碩士論文加值系統

在科技的飛速進步下，人們對於能源的需求與日俱增，應對這種不斷增長的需求的最有效的方法便是增加發電量和節約能源。但增加發電量並不是那麼簡單和符合效益的，再者臺灣的發電結構中，燃油與燃煤站比高達74.4%，貿然提升發電量只會與「2050年全球淨零碳排」的目標漸行漸遠。對此若能將節能減碳、節約能源的意識帶入人們的生活中，想必對永續發展將會是不小的幫助。故本研究提出利用非侵入式負載監控技術配合深度學習演算法對主電源進行分析，透過此方法即可在不必破壞家中迴路以及安裝成本較低的情況下，提高人們對家庭能源管理系統的使用意願，進而理解自身用電模式，改善自身用電習慣，促進永續發展的目標。
本研究將運用長短期記憶網路技術配合非侵入負載監控技術，在僅監測主電源用電資訊的情況下，使長短期記憶網路辨識出正在使用中的電器特徵，以及推測出電器用電情形，落實節能減碳的目標。

With the rapid progress of science and technology, people's demand for energy is increasing day by day, which indirectly causes the world to drift away from the original goal of "2050 global net zero carbon emissions". In this regard, if the awareness of energy conservation and carbon reduction can be brought into people's lives, it will certainly be of great help to sustainable development. Therefore, this study proposes to use non-intrusive load monitoring technology with deep learning algorithm to analyze the main power supply. Through this method, people can improve the use of home energy management system without destroying the home circuit and the installation cost is low. Willingness to promote the goal of sustainable development.
This research will use long-term short-term memory network technology combined with non-intrusive load monitoring technology. In the case of only monitoring the power consumption information of the main power supply, the long-term short-term memory network can identify the characteristics of the electrical appliances in use and infer the power consumption of the electrical appliances situation, and implement the goal of energy saving and carbon reduction.

中文摘要 ......i
Abstract ......ii
誌謝 ......iii
目錄 ......iv
表目錄 ......vi
圖目錄 ......vii
符號說明 ......viii
第一章 緒論......1
1.1研究背景與動機......1
1.2研究目的......1
1.3研究架構與流程......2
第二章 文獻探討......3
2.1 家庭能源管理系統......3
2.2 電力解析問題......3
2.3 電力解析問題之深度學習應用......5
第三章 研究方法......7
3.1 用電資訊來源......7
3.2 資料預處理......7
3.3 類神經網路模型......8
3.4 評量標準......13
3.4.1 平均絕對誤差(Mean Absolute Error, MAE)......13
3.4.2 信號聚合誤差(Signal Aggregate Error,SAE)......13
第四章 實驗設計與結果......14
4.1 使用單一房屋用電資訊訓練模型......17
4.1.1 House 1測試結果......17
4.1.2 House 2測試結果......18
4.1.3 House 3測試結果......19
4.1.4 House 5測試結果......20
4.1.5 House 6測試結果......21
4.2 使用多重房屋用電資訊整合訓練模型......22
4.2.1 House 1測試結果......22
4.2.2 House 2測試結果......23
4.2.3 House 3測試結果......24
4.2.4 House 5測試結果......25
4.2.5 House 6測試結果......26
4.3 實驗結果分析......27
第五章 結論與未來方向......28
5.1 研究結論......28
5.2 未來方向......28
參考文獻......30
Extended Absstract ......32

[1] Angelis, G. F., Timplalexis, C., Krinidis, S., Ioannidis, D., & Tzovaras, D. (2022). NILM Applications: Literature review of learning approaches, recent developments and challenges. Energy and Buildings, 111951.
[2] Moen, R. L. (1979, December). Solar energy management system. In 1979 18th IEEE Conference on Decision and Control including the Symposium on Adaptive Processes (Vol. 2, pp. 917-919). IEEE.
[3] Prince, B., Sasson, A. M., Wynne, W. T., Trefny, F., Cleveland, F., Robinson, J. T., & Horton, J. S. (1986). Advances in energy management systems. IEEE transactions on power systems, 1(3), 226-234.
[4] Hart, G. W., Prototype nonintrusive appliance load monitor, Sept. 1985.
[5] LeCun, Y. (1989). Generalization and network design strategies. Connectionism in perspective, 19(143-155), 18.
[6] LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), 2278-2324.
[7] Chellapilla, K., Puri, S., & Simard, P. (2006, October). High performance convolutional neural networks for document processing. In Tenth international workshop on frontiers in handwriting recognition. Suvisoft.
[8] LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. nature, 521(7553), 436-444.
[9] Hopfield, J. J. (1982). Neural networks and physical systems with emergent collective computational abilities. Proceedings of the national academy of sciences, 79(8), 2554-2558.
[10] Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural computation, 9(8), 1735-1780.
[11] Schuster, M., & Paliwal, K. K. (1997). Bidirectional recurrent neural networks. IEEE transactions on Signal Processing, 45(11), 2673-2681.
[12] Olah, C. (2015). Understanding lstm networks. Retrieved from https://colah.github.io/posts/2015-08-Understanding-LSTMs/
[13] Bonfigli, R., Felicetti, A., Principi, E., Fagiani, M., Squartini, S., & Piazza, F. (2018). Denoising autoencoders for non-intrusive load monitoring: improvements and comparative evaluation. Energy and Buildings, 158, 1461-1474.
[14] Mauch, L., & Yang, B. (2015, December). A new approach for supervised power disaggregation by using a deep recurrent LSTM network. In 2015 IEEE Global Conference on Signal and Information Processing (GlobalSIP) (pp. 63-67). IEEE.
[15] 王越先（民110）。基於聯邦選擇學習法暨最佳化演算法於非侵入式負載分析。（碩士論文。國立中興大學﹞。
[16] Kolter, J. Z., & Johnson, M. J. (2011, August). REDD: A public data set for energy disaggregation research. In Workshop on data mining applications in sustainability (SIGKDD), San Diego, CA (Vol. 25, No. Citeseer, pp. 59-62).
[17] Kelly, J., & Knottenbelt, W. (2015). The UK-DALE dataset, domestic appliance-level electricity demand and whole-house demand from five UK homes. Scientific data, 2(1), 1-14.
[18] Murray, D., Stankovic, L., & Stankovic, V. (2017). An electrical load measurements dataset of United Kingdom households from a two-year longitudinal study. Scientific data, 4(1), 1-12.
[19] Iqbal, H. K., Malik, F. H., Muhammad, A., Qureshi, M. A., Abbasi, M. N., & Chishti, A. R. (2021). A critical review of state-of-the-art non-intrusive load monitoring datasets. Electric Power Systems Research, 192, 106921.
[20] Bhanja, S., & Das, A. (2018). Impact of data normalization on deep neural network for time series forecasting. arXiv preprint arXiv:1812.05519.
[21] Kelly, J., & Knottenbelt, W. (2015, November). Neural nilm: Deep neural networks applied to energy disaggregation. In Proceedings of the 2nd ACM international conference on embedded systems for energy-efficient built environments (pp. 55-64).
[22] He, W., & Chai, Y. (2016, November). An empirical study on energy disaggregation via deep learning. In 2016 2nd International Conference on Artificial Intelligence and Industrial Engineering (AIIE 2016) (pp. 338-342). Atlantis Press.
[23] Zhong, M., Goddard, N., & Sutton, C. (2014). Signal aggregate constraints in additive factorial HMMs, with application to energy disaggregation. Advances in Neural Information Processing Systems, 27.
[24] Jia, Z., Yang, L., Zhang, Z., Liu, H., & Kong, F. (2021). Sequence to point learning based on bidirectional dilated residual network for non-intrusive load monitoring. International Journal of Electrical Power & Energy Systems, 129, 106837.
[25] Song, J., Wang, H., Du, M., Peng, L., Zhang, S., & Xu, G. (2021). Non-intrusive load identification method based on improved long short term memory network. Energies, 14(3), 684.
[26] Rumelhart, D. E., Hinton, G. E., & Williams, R. J. (1986). Learning representations by back-propagating errors. nature, 323(6088), 533-536.
[27] Mauch, L., & Yang, B. (2016, March). A novel DNN-HMM-based approach for extracting single loads from aggregate power signals. In 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (pp. 2384-2388). IEEE.
[28] Abbasimehr, H., Shabani, M., & Yousefi, M. (2020). An optimized model using LSTM network for demand forecasting. Computers & industrial engineering, 143, 106435.