臺灣博碩士論文加值系統

磷酸鐵鋰電池將會越來越廣泛的運用在所有動力或儲能產品上，但電池並非可永久使用的產品，經過不同的使用方式，不同的環境下，都會加速電池的老化程度。如何在電池壽命耗盡前能得知這顆電池剩餘使用壽命還剩多少並做出相對應的應用，將會是未來電池產業最重要的議題。
此篇研究方向為，如何套用磷酸鐵鋰電池在已適用於三元鋰電池的長短期記憶法，來預測磷酸鐵鋰電池的剩餘使用壽命。本研究先是使用公開的三元鋰電池數據帶入長短期記憶法，確認方法的擬合程度與所需要的超參數，如損失函數的極值以及模型的最佳深度層數，而在特徵維度選取上，為了配合能導入產品中，故僅選用過去容量數據預測未來容量數據。
以此為基礎，將公開數據集的磷酸鐵鋰電池帶入相同的長短期記憶法模型，在資料處理的部份，移除資料數不足與數據不適合使用的電池組，並且壽命結束值設定在電池容量的30%，此稱之為電池完全耗盡。
實驗結果發現，當電芯達到壽命結束值時，實際已使用容量值與預測已使用容量值，兩者最佳誤差值約為1%，模型評估函數MAE為0.0324，RMSE為0.0489，而損失函數為3.23E-05。並且實驗過程中發現，儘管在小批次的數據量或是突然發生離群值的情況下，模型預測結果並不會出現偏移，故推估長短期記憶法可以直接應用在實際產品使用上。

The lithium-iron phosphate (LiFePO4, LFP) batteries will be used in a variety of automotive or energy storage systems. Each LFP batteries have its own cycle life. The cycle life will differ depends on the ways the users use it. The way how to predict the remaining useful life (RUL) will be the most important thing in the future.
The thesis is mainly about how to apply “long short term memory method” (LSTM) which is being used on predict the lithium-ion (Ni-Co-Al, NCA) batteries’ RUL to the RUL prediction of the LFP batteries.
First of all, this thesis uses to the data in public to ensure LSTM is suitable for Lithium-ion batteries to be predict the RUL, and regards the discharge-capacity data before as the data feature to predict discharge-capacity data in the future.
Base on the above, the thesis applies LSTM to LFP batteries to predict the RUL. Meanwhile, the unsuitable data is excluded from the prediction. The experiment sets 30% of full capacity to be the end-of-life of LFP batteries, which the 30% is called un-functional point.
The experiment found that the optimal error rate between prediction of the capacity data and the actual of the capacity data as a same EOL is about 1%. The mean absolute error (MAE) is 0.0324, the root mean square error (RMSE) is 0.0489, and the loss function is 3.23E-05 in the evaluation function.
By the experiments, small amount of data will not let the prediction data become inaccurate, so I think LSTM method can be usage in online products even if there is only a small amount of data.

摘要 iii
Abstract iv
誌 謝 v
目錄 vi
表目錄 vii
圖目錄 viii
第一章 序論 1
1.1 研究動機與目的 1
1.2 研究問題 2
第二章 文獻探討 6
2.1使用數據模型預測 RUL的文獻探討 6
2.2使用 Nerual Network 預測 RUL的文獻探討 7
第三章 LSTM方法 11
3.1 RNN 架構與流程 11
3.2 LSTM 架構與流程 13
第四章 運用LSTM建立電池RUL預測模型 17
4.1 數據集轉換 17
4.2 建立電池預測模型 20
第五章 實驗數據結果 22
5.1 實驗環境 22
5.2 實驗結果 22
5.2.1 預測三元鋰電池之RUL 22
5.2.2 預測LFP電池之RUL 24
第六章 結論 32
參考文獻 34
附錄 37

[1]程薇, "BP 公司在 2019 年世界能源統計年鑑中指出世界正走在一條不可持續的道路上," 石油煉製與化工, vol. 50, no. 9, p. 96, 2019.
[2]C. Grey and J. Tarascon, "Sustainability and in situ monitoring in battery development," Nature materials, vol. 16, no. 1, pp. 45-56, 2017.
[3]孟祥飛、龐秀嵐、崇鋒、侯少攀、祁斌, "電化學儲能在電網中的應用分析及展望," 儲能科學與技術, vol. 8, no. S1, p. 38, 2019.
[4]Y. Xing, E. W. Ma, K. L. Tsui, and M. Pecht, "Battery management systems in electric and hybrid vehicles," Energies, vol. 4, no. 11, pp. 1840-1857, 2011.
[5]M. S. H. Lipu, M. A. Hannan, A. Hussain, M. M. Hoque, P. J. Ker, M. H. M. Saad, and A. Ayob," A review of state of health and remaining useful life estimation methods for lithium-ion battery in electric vehicles: Challenges and recommendations," Journal of cleaner production, vol. 205, pp. 115-133, 2018.
[6]S. Park, J. Ahn, T. Kang, S. Park, Y. Kim, I. Cho, and J. Kim, "Review of state-of-the-art battery state estimation technologies for battery management systems of stationary energy storage systems," Journal of Power Electronics, vol. 20, no. 6, pp. 1526-1540, 2020.
[7]G. Pil'atowicz, A. Marongiu, J. Drillkens, P. Sinhuber, and D. U. Sauer, "A critical overview of definitions and determination techniques of the internal resistance using lithium-ion, lead-acid, nickel metal-hydride batteries and electrochemical double-layer capacitors as examples," Journal of Power Sources, vol. 296, pp. 365-376, 2015.
[8]莊全超、徐守東、邱祥雲、崔永麗、方亮、孫世剛, "鋰離子電池的電化學阻抗譜分析," 化學進展, vol. 22, no. 06, p. 1044, 2010.
[9]S. Montoya-Bedoya, L. A. Sabogal-Moncada, E. Garcia-Tamayo, and H. V. Martínez-Tejada, "A circular economy of electrochemical energy storage systems: Critical review of SOH/RUL estimation methods for second-life batteries," Green Energy and Environment, p. 67, 2020.
[10]Y. Li, K. Liu, A. M. Foley, A. Zülke, M. Berecibar, E. Nanini-Maury, J. Van Mierlo, and H. E. Hoster, "Data-driven health estimation and lifetime prediction of lithium-ion batteries: A review," Renewable and Sustainable Energy Reviews, vol. 113, p. 109254, 2019.
[11]S. Zhang, "A new method for lithium-ion battery's SOH estimation and RUL prediction," IEEE Conference on Industrial Electronics and Applications (ICIEA), pp. 2693-2697, 2018.
[12]F. Geng, Y. Kang, Z. Li, C. Zhang, and B. Duan, "Remaining useful life prediction of LiFePO4 battery based on particle filter," Chinese Automation Congress (CAC), pp. 1149-1153, 2017.
[13]Y. Zhang, R. Xiong, H. He, and Z. Liu, "A LSTM-RNN method for the lithuim-ion battery remaining useful life prediction," Prognostics and System Health Management Conference (PHM-Harbin), IEEE, pp. 1-4, 2017
[14]Y. Zhang, R. Xiong, H. He, and M. G. Pecht, "Long short-term memory recurrent neural network for remaining useful life prediction of lithium-ion batteries," IEEE Transactions on Vehicular Technology, vol. 67, no. 7, pp. 5695-5705, 2018.
[15]C. Wang, N. Lu, S. Wang, Y. Cheng, and B. Jiang, "Dynamic long short-term memory neural-network-based indirect remaining-useful-life prognosis for satellite lithium-ion battery," Applied Sciences, vol. 8, no. 11, p. 2078, 2018.
[16]K. A. Severson, P. M. Attia, N. Jin, N. Perkins, B. Jiang, Z. Yang, M. H. Chen, M. Aykol, P. K. Herring, D. Fraggedakis, M. Z. Bazant, S. J. Harris, W. C. Chueh, and R. D. Braatz, "Data-driven prediction of battery cycle life before capacity degradation," Nature Energy, vol. 4, no. 5, pp. 383-391, 2019.
[17]I. Sanz-Gorrachategui, P. Pastor-Flores, M. Pajovic, Y. Wang, P. V. Orlik, C. Bernal-Ruiz, A. Bono-Nuez, and J. S. Artal-Sevil, "Remaining Useful Life Estimation for LFP Cells in Second-Life Applications," IEEE Transactions on Instrumentation and Measurement, vol. 70, pp. 1-10, 2021.
[18]S. Lu, F. Wang, C. Piao, and Y. Ma, "Health State Prediction of Lithium Ion Battery Based On Deep Learning Method," in IOP Conference Series: Materials Science and Engineering, vol. 782, 2020.
[19]C. Olah, "Understanding LSTM networks", available at: https://colah.github.io/posts/2015-08-Understanding-LSTMs/ (Accessed 20 November 2021), 2015.
[20]S. Hochreiter and J. Schmidhuber, "Long Short-Term Memory," Neural Computation, vol. 9, no. 8, pp. 1735-1780, 1997.
[21]G. Dos Reis, C. Strange, M. Yadav, and S. Li, "Lithium-ion battery data and where to find it," Energy and AI, p. 100081, 2021.
[22]B. Saha and K. Goebel. Battery data set. NASA Ames Prognostics Data Repository; NASA Ames: Moffett Field, CA, USA, 2007. Available online: https://ti.arc.nasa.gov/tech/dash/groups/pcoe/prognostic-data-repository/#battery (accessed on 21 September 2021), 2007.
[23]D. P. Abraham, D. W. Dees, J. Knuth, E. Reynolds, R. Gerald, Y-E. Hyung, I. Belharouak, M. Stoll, E. Sammann, S. MacLaren, R. Haasch, R. Twesten, M. Sardela, V. Battaglia, E. Cairns, J. Kerr, M. Kerlau, R. Kostecki, J. Lei, K. McCarthy, F. McLarnon, J. Reimer, T. Richardson, P. Ross, S. Sloop, X. Song, V. Zhuang, M. Balasubramanian, J. McBreen, K-Y. Chung, X. Q. Yang, W.-S. Yoon, and L. Norin, "Diagnostic examination of Generation 2 lithium-ion cells and assessment ofperformance degradation mechanisms," in Argonne National Lab.(ANL), Argonne, IL (United States), no. ANL-05/21, 2005
[24]D. Liu, J. Pang, J. Zhou, and Y. Peng, "Data-driven prognostics for lithium-ion battery based on Gaussian process regression," in Proceedings of the IEEE 2012 Prognostics and System Health Management Conference (PHM-2012 Beijing), pp. 1-5, 2012.