伴隨著人工智能、互聯網、大數據分析日漸增長，大幅加速了各領域對預測剩餘使用壽命精準度之重視。在鋰離子電池中，未定期保養及更換容易造成電池失效，恐造成生產線及公司停滯，損失難以估計。由於鋰離子電池使用到一定次數後，電量及穩定度會迅速下降，傳統解決方式只能定期保養及故障後再進行更換電池，造成了工廠許多產品之延遲；因此進行剩餘使用壽命(RUL, Remaining useful life)之預測，當我們選擇出準確的模型方法時，能有效且即時更換電池及保持電池中的健康程度，以即能監控電池的穩定度。本研究主要為預測鋰離子電池之RUL，因此本研究主要著重為類神經網路-長短期記憶之注意力模型及長短期記憶力模型為主要預測方法。本研究主要是以Severson et al. (2019)所發表之資料集，資料中分為“2018-04-12”、“2017-05-12”、“2017-06-30”，本研究所選擇之資料為“2017-05-12”及“2017-06-30”，並考量電池之狀態及電量，將相同之電池狀態及電量之資料整合在一起，運用計算模型進行學習測試資料，再使用轉移學習(Transfer learning)將已經學習之模型參數應用於新資料中進行RUL之預測。最終將長短期記憶之注意力模型(LSTM-Attention, Long short-term memory with attention)及長短期注意力模型(LSTM, Long short-term memory)之預測結果做相互比較，得到之RUL之評估結果分別為： LSTM-Attention之絕對誤差(AE, Absolute error)為23、3、9、11；LSTM之AE為57、23、16、32。LSTM-Attention之相對誤差(RE, relative error)為0.6682、0.3610、1.205、1.242；LSTM之RE為3.174、2.768、2.1419、3.6117。得到之四種試驗結果中，四組結果皆顯示LSTM-Attention優於LSTM

With the increasing growth of artificial intelligence, the Internet, and big data analysis, all fields have greatly accelerated the importance of predicting the accuracy of Remaining useful life. In Lithium-ion batteries, not regularly maintain or change them can easily lead to battery failure, which may cause production lines and companies to stagnate, and these losses are difficult to estimate. Since the Lithium-ion battery is used for a certain number of times, the power and stability will decrease rapidly. The traditional solution can only change the battery after regular maintenance and failure, causing delays for many products in the factory; therefore, the Remaining useful life (RUL) prediction can effectively and instantly change the battery and maintain the health of the battery when we choose an accurate model method, so that the stability of the battery can be monitored. This research is mainly to predict the RUL of lithium-ion batteries. Therefore, this research mainly focuses on the long short-term memory with attention (LSTM-attention) and LSTM models as the main prediction methods. This research is mainly based on the data set published by Severson et al. (2019). The data is divided into "2018-04-12", "2017-05-12", and "2017-06-30". This study chose "2017-05-12" and "2017-06-30" data, and consider the battery status and power, then integrate the same battery status and power data; moreover, using the calculation model to learn and test the data, and the transfer learning applies parameters to the new data for RUL prediction. The RUL evaluation results of four cases are: LSTM-Attention's absolute error (AE) are 23, 3, 9, and 11; LSTM's AE is 57, 23, 16, and 32. The relative error (RE) values of LSTM-Attention are 0.6682, 0.3610, 1.205, and 1.242; the RE values of LSTM are 3.174, 2.768,2 .1419, and 3.6117. The results show that LSTM-Attention is better than LSTM.

Bahdanau, D., Cho, K., & Bengio, Y. (2014). Neural machine translation by jointly
learning to align and translate. arXiv preprint arXiv:1409.0473.
Bengio, Y. (2012). Deep learning of representations for unsupervised and
transfer learning. In Proceedings of ICML Workshop on Unsupervised and Transfer Learning, 17-36.
Burgueño, L., Cabot, J., & Gérard, S. (2019). An LSTM-based
neural network architecture for model transformations. In 2019 ACM/IEEE 22nd International Conference on Model Driven Engineering Languages and Systems (MODELS), 294-299.
Caruana, R. (1997). Multitask learning. Machine learning, 28(1), 41-75.
Cho, K., Van Merriënboer, B., Gulcehre, C., Bahdanau, D., Bougares, F., Schwenk, H., & Bengio, Y. (2014). Learning phrase representations using RNN encoder-decoder for statistical machine translation. arXiv preprint arXiv:1406.1078.
Chung, S. Y., Bloking, J. T., & Chiang, Y. M. (2002). Electronically conductive phospho-olivines as lithium storage electrodes. Nature materials, 1(2), 123-128.
Dong, G., Chen, Z., Wei, J., & Ling, Q. (2018). Battery health prognosis using
Brownian motion modeling and particle filtering. IEEE Transactions on Industrial Electronics, 65(11), 8646-8655.
Gao, L., Guo, Z., Zhang, H., Xu, X., & Shen, H. T. (2017). Video captioning
with attention-based LSTM and semantic consistency. IEEE Transactions on Multimedia, 19(9), 2045-2055.
Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural
Computation, 9(8), 1735-1780.
How, D. N., Hannan, M. A., Lipu, M. H., & Ker, P. J. (2019). State of charge
estimation for lithium-ion batteries using model-based and data-driven methods: A review. IEEE Access, 7, 136116-136136.
Huang, C. G., Huang, H. Z., & Li, Y. F. (2019). A Bidirectional LSTM
prognostics method under multiple operational conditions. IEEE Transactions on Industrial Electronics, 66(11), 8792-8802.
Ion, L. (2003). Technical handbook: Lithium Ion catalog. Gold Peak Industries (Taiwan) Ltd.
Liu, C., Wang, Y., & Chen, Z. (2019). Degradation model and cycle life
prediction for lithium-ion battery used in hybrid energy storage system. Energy, 166, 796-806.

Luong, M. T., Pham, H., & Manning, C. D. (2015). Effective approaches to
attention-based neural machine translation. arXiv preprint arXiv:1508.04025.
Mayer, D. G., Kinghorn, B. P., & Archer, A. A. (2005). Differential evolution–an easy and efficient evolutionary algorithm for model optimisation. Agricultural Systems, 83(3), 315-328.
Okoh, C., Roy, R., Mehnen, J., & Redding, L. E. (2014). Overview of remaining useful life prediction techniques in through-life engineering services. Procedia
CIRP, 16, 158-163.
Pan, S. J., & Yang, Q. (2009). A survey on transfer learning. IEEE Transactions on
knowledge and data engineering, 22(10), 1345-1359.
Qu, J., Liu, F., Ma, Y., & Fan, J. (2019). A neural-network-based method for RUL prediction and SOH monitoring of lithium-ion battery. IEEE Access, 7, 87178-87191.
Rahnamayan, S., Tizhoosh, H. R., & Salama, M. M. (2008). Opposition-based
differential evolution. IEEE Transactions on Evolutionary computation, 12(1), 64-79.
Ran, X., Shan, Z., Fang, Y., & Lin, C. (2019). An LSTM-based method with
attention mechanism for travel time prediction. Sensors, 19(4), 861.
Rodrigues, L. R. (2017). Remaining useful life prediction for multiple-component
systems based on a system-level performance indicator. IEEE/ASME Transactions on Mechatronics, 23(1), 141-150.
Schmidhuber, J. (2015). Deep learning in neural networks: An overview. Neural
Networks, 61, 85-117.
Severson, K. A., Attia, P. M., Jin, N., Perkins, N., Jiang, B., Yang, Z., Chen, M.,
Aykol, M., Herring, P. K., Fraggedakis, D., Harris, S. J., Chueh, W. C., Braatz,
R. D,& Bazant, M. Z. (2019). Data-driven prediction of battery cycle life before
capacity degradation. Nature Energy, 4(5), 383-391.
Shen, S., Sadoughi, M., & Hu, C. (2019). Online estimation of lithium-ion
battery capacity using transfer learning. In 2019 IEEE Transportation Electrification Conference and Expo (ITEC), 1-4.
Storn, R. (1996). On the usage of differential evolution for function optimization. In
Proceedings of North American Fuzzy Information Processing, 519-523.
Storn, R., & Price, K. (1997). Differential evolution–a simple and efficient heuristic
for global optimization over continuous spaces. Journal of Global Optimization, 11(4), 341-359.
Tang, S., Yu, C., Wang, X., Guo, X., & Si, X. (2014). Remaining useful life
prediction of lithium-ion batteries based on the wiener process with measurement error. Energies, 7(2), 520-547.
Voelcker, J. (2007). Lithium batteries take to the road. IEEE Spectrum, 44(9), 26-31
Wang, Y., Huang, M., Zhu, X., & Zhao, L. (2016). Attention-based LSTM for
aspect-level sentiment classification. In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, 606-615.
Wang, Y., Ni, Y., Li, N., Lu, S., Zhang, S., Feng, Z., & Wang, J. (2019). A
method based on improved ant lion optimization and support vector regression for remaining useful life estimation of lithium‐ion batteries. Energy Science & Engineering, 7(6), 2797-2813.
Wang, Y., Ni, Y., Lu, S., Wang, J., & Zhang, X. (2019). Remaining useful life
prediction of lithium-ion batteries using support vector regression optimized by artificial bee colony. IEEE Transactions on Vehicular Technology, 68(10), 9543-9553.
Wei, J., Dong, G., & Chen, Z. (2017). Remaining useful life prediction and state
of health diagnosis for lithium-ion batteries using particle filter and support vector regression. IEEE Transactions on Industrial Electronics, 65(7), 5634-5643.
Wu, J., Zhang, C., & Chen, Z. (2016). An online method for lithium-ion battery
remaining useful life estimation using importance sampling and neural networks. Applied Energy, 173, 134-140.
Yu, Y., Si, X., Hu, C., & Zhang, J. (2019). A review of recurrent neural
networks: LSTM cells and network architectures. Neural Computation, 31(7), 1235-1270.
Zhang, J., & Zulkernine, M. (2006). Anomaly based network intrusion detection with
unsupervised outlier detection. In 2006 IEEE International Conference on Communications, 5, 2388-2393.
Zhang, Y., Xiong, R., He, H., & Pecht, M. G. (2018). Long short-term memory
recurrent neural network for remaining useful life prediction of lithium-ion batteries. IEEE Transactions on Vehicular Technology, 67(7), 5695-5705.
Zheng, Y., Ouyang, M., Han, X., Lu, L., & Li, J. (2018). Investigating the error
sources of the online state of charge estimation methods for lithium-ion batteries in electric vehicles. Journal of Power Sources, 377, 161-188.
楊凱傑(2018). Implementing support vector regression to predict the remaining.
useful lifetime of mechanical components with health indicator. 國立臺灣科技大學碩士論文。
楊熙文(2018). Different methods of the anomaly detection of ball bearings on remaining useful lifetime prediction. 國立臺灣科技大學碩士論文。
蕭凱駿(2019). Application ensemble method for wafer classification in semiconductor etching process. 國立臺灣科技大學碩士論文。