本研究旨在結合深度學習演算法與CFD模擬，建立套管式熱交換器之預測性保養策略。首先，利用Ansys Fluent建立沉積模擬模型，利用離散相模型(Discrete Phase Model, DPM)模擬沉積粒子在熱交換器內部的運動和沉積行為，進一步分析潛在結垢區域及沉積分布。接著，結合模擬結果與台聚關係企業（USI Corporation）提供之歷史運行資料，以構建深度學習預測模型，並導入長短期記憶網路（LSTM）與門控循環單元（GRU）等時間序列演算法，預測未來熱傳性能與關鍵運行指標。為提升模型效能，本研究設計多組變數組合，探討結合歷史資料與模擬資料對預測準確度的影響。結果顯示，融合模擬與實測資料能有效降低雜訊干擾與測量誤差，顯著提升模型收斂速度與預測能力。其中，GRU模型表現最佳，其RMSE為0.0035，MAE為0.0027，決定係數R²達0.9979，顯示其結構簡潔且學習效率優於LSTM。最終，基於訓練完成之模型建構警示系統，能準確預測未來12小時的熱交換效能，並對各級警示值發出即時警報。本研究可成功應用於實廠熱交換設備，協助操作人員掌握最佳保養時機，避免不必要的維修與突發性損壞，提升整體設備運行效率與可靠性。

This study aims to integrate deep learning algorithms with CFD simulations to establish a predictive maintenance strategy for double-pipe heat exchangers. First, a fouling simulation model was developed using Ansys Fluent, where the Discrete Phase Model (DPM) was applied to simulate the motion and deposition behavior of particles within the heat exchanger, enabling the identification of potential fouling zones and deposition patterns. The simulation results were then combined with historical operating data provided by USI Corporation to construct a deep learning predictive model. Time-series algorithms, including Long Short-Term Memory and Gated Recurrent Unit networks, were implemented to forecast future heat transfer performance and key operational indicators. To enhance model performance, multiple variable combinations were designed to investigate the impact of integrating historical and simulation data on prediction accuracy. The results indicate that fusing simulated and measured data effectively reduces noise interference and measurement errors, significantly improving model convergence speed and prediction capability. Finally, based on the trained model, a warning system was developed to accurately predict the heat exchanger’s performance over the next 12 hours. This approach can be effectively applied to industrial heat exchangers, enabling operators to determine optimal maintenance timing, prevent unnecessary repairs and unexpected failures, and enhance overall operational efficiency and reliability.

[1] M. Hermann, T. Pentek, and B. Otto, "Design principles for industrie 4.0 scenarios," in 2016 49th Hawaii international conference on system sciences (HICSS), 2016: IEEE, pp. 3928-3937.
[2] M. Compare, P. Baraldi, and E. Zio, "Challenges to IoT-enabled predictive maintenance for industry 4.0," IEEE Internet of things journal, vol. 7, no. 5, pp. 4585-4597, 2019.
[3] M. Molęda, B. Małysiak-Mrozek, W. Ding, V. Sunderam, and D. Mrozek, "From corrective to predictive maintenance—A review of maintenance approaches for the power industry," Sensors, vol. 23, no. 13, p. 5970, 2023.
[4] M. Awais and A. A. Bhuiyan, "Recent advancements in impedance of fouling resistance and particulate depositions in heat exchangers," International Journal of Heat and Mass Transfer, vol. 141, pp. 580-603, 2019/10/01/ 2019, doi: https://doi.org/10.1016/j.ijheatmasstransfer.2019.07.011.
[5] C. Groba, S. Cech, F. Rosenthal, and A. Gossling, "Architecture of a predictive maintenance framework," in 6th International Conference on Computer Information Systems and Industrial Management Applications (CISIM'07), 2007: IEEE, pp. 59-64.
[6] D. K. Mohanty and P. M. Singru, "Fouling analysis of a shell and tube heat exchanger using local linear wavelet neural network," International Journal of Heat and Mass Transfer, vol. 77, pp. 946-955, 2014/10/01/ 2014, doi: https://doi.org/10.1016/j.ijheatmasstransfer.2014.06.007.
[7] K.-T. Yang, "Artificial neural networks (ANNs): a new paradigm for thermal science and engineering," 2008.
[8] R. Jradi, A. Fguiri, C. Marvillet, and M. R. Jeday, "Tubular heat exchanger fouling in phosphoric acid concentration process," Heat Exchangers [Working Title], 2019.
[9] D. Kern, "A theoretical analysis of thermal surface fouling," Br. Chem. Eng., vol. 4, pp. 258-262, 1959.
[10] G. C. Diaz, Simulation and control of heat exchangers using artificial neural networks. University of Notre Dame, 2000.
[11] R. Jradi, C. Marvillet, and M. R. Jeday, "Modeling and comparative study of heat exchangers fouling in phosphoric acid concentration plant using experimental data," Heat and Mass Transfer, vol. 56, no. 9, pp. 2653-2666, 2020.
[12] M. Pan, I. Bulatov, and R. Smith, "Improving heat recovery in retrofitting heat exchanger networks with heat transfer intensification, pressure drop constraint and fouling mitigation," Applied Energy, vol. 161, pp. 611-626, 2016/01/01/ 2016, doi: https://doi.org/10.1016/j.apenergy.2015.09.073.
[13] M. Bayat, J. Aminian, M. Bazmi, S. Shahhosseini, and K. Sharifi, "CFD modeling of fouling in crude oil pre-heaters," Energy Conversion and Management, vol. 64, pp. 344-350, 2012/12/01/ 2012, doi: https://doi.org/10.1016/j.enconman.2012.05.003.
[14] Z. Han, Z. Xu, and X. Yu, "CFD modeling for prediction of particulate fouling of heat transfer surface in turbulent flow," International Journal of Heat and Mass Transfer, vol. 144, p. 118428, 2019/12/01/ 2019, doi: https://doi.org/10.1016/j.ijheatmasstransfer.2019.07.078.
[15] M. Nadda, S. K. Shah, S. Roy, and A. Yadav, "CFD-based deep neural networks (DNN) model for predicting the hydrodynamics of fluidized beds," Digital Chemical Engineering, vol. 8, p. 100113, 2023/09/01/ 2023, doi: https://doi.org/10.1016/j.dche.2023.100113.
[16] R. K. Pasupuleti, M. Bedhapudi, S. R. Jonnala, and A. R. Kandimalla, "Computational Analysis of Conventional and Helical Finned Shell and Tube Heat Exchanger Using ANSYS-CFD," International Journal of Heat & Technology, vol. 39, no. 6, 2021.
[17] B. E. Launder and D. B. Spalding, "Lectures in mathematical models of turbulence," (No Title), 1972.
[18] M. Holt, Numerical methods in fluid dynamics. Springer Science & Business Media, 2012.
[19] J. P. Van Doormaal and G. D. Raithby, "Enhancements of the SIMPLE method for predicting incompressible fluid flows," Numerical heat transfer, vol. 7, no. 2, pp. 147-163, 1984.
[20] P. Nunes, J. Santos, and E. Rocha, "Challenges in predictive maintenance – A review," CIRP Journal of Manufacturing Science and Technology, vol. 40, pp. 53-67, 2023/02/01/ 2023, doi: https://doi.org/10.1016/j.cirpj.2022.11.004.
[21] M. Pérez, "An Investigation of ADAM: A Stochastic Optimization Method," in Proceedings of the 39th International Conference on Machine Learning, Baltimore, MD, USA, 2022, pp. 17-23.
[22] S. Hochreiter and J. Schmidhuber, "Long Short-Term Memory," Neural Computation, vol. 9, no. 8, pp. 1735-1780, 1997, doi: 10.1162/neco.1997.9.8.1735.
[23] C. Olah. "Understanding LSTM Networks." https://colah.github.io/posts/2015-08-Understanding-LSTMs/ (accessed May 20, 2025).
[24] J. Chung, Ç. Gülçehre, K. Cho, and Y. Bengio, "Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling," ArXiv, vol. abs/1412.3555, 2014.
[25] P. A. Banu, D. R. Lohith, M. P. Kalyan, D. S. Vempati, and B. H. Sai, "Simulation of fin and tube heat exchanger and validation with CFD analysis," Materials Today: Proceedings, vol. 66, pp. 1471-1476, 2022.
[26] K. P. Meurer, G. Sylvester, C. Ohm, H. Zimmermann, and H. Bartl, "Process for the preparation of ethylene/vinyl acetate copolymers with reduced tendency to tackiness, new ethylene/vinyl acetate copolymers and their use," ed: Google Patents, 1992.
[27] M. H. P. Swari, I. P. S. Handika, and I. K. S. Satwika, "Comparison of simple moving average, single and modified single exponential smoothing," in 2021 IEEE 7th Information Technology International Seminar (ITIS), 2021: IEEE, pp. 1-5.
[28] Y. Dai, Y. Wang, M. Leng, X. Yang, and Q. Zhou, "LOWESS smoothing and Random Forest based GRU model: A short-term photovoltaic power generation forecasting method," Energy, vol. 256, p. 124661, 2022/10/01/ 2022, doi: https://doi.org/10.1016/j.energy.2022.124661.
[29] L. B. V. de Amorim, G. D. C. Cavalcanti, and R. M. O. Cruz, "The choice of scaling technique matters for classification performance," Applied Soft Computing, vol. 133, p. 109924, 2023/01/01/ 2023, doi: https://doi.org/10.1016/j.asoc.2022.109924.
[30] P. Geladi and B. R. Kowalski, "Partial least-squares regression: a tutorial," Analytica chimica acta, vol. 185, pp. 1-17, 1986.