DOI: https://doi.org/10.9767/jcerp.20485

Hybrid Approaches for Steam Demand Forecasting: Combining First Principles, Box and Jenkins, and Neural Network Models

1Data Analytics and Numerics in Control Engineering Unit, Department of Chemical and Metallurgical Engineering, University of the Witwatersrand, Johannesburg, South Africa

2Department of Chemical Engineering, University of Pretoria, Pretoria, South Africa

Received: 9 Sep 2025; Revised: 20 Sep 2025; Accepted: 23 Sep 2025; Available online: 26 Sep 2025; Published: 26 Dec 2025.
Editor(s): Istadi Istadi
Open Access Copyright (c) 2025 by Authors, Published by Universitas Diponegoro and BCREC Publishing Group
Creative Commons License This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Fulltext View|Download

Citation Format:
Cover Image
Abstract

The pulp and paper industry relies heavily on batch sulphite digesters for chemical cellulose production, where steam is a critical utility – lack of steam prediction results in venting losses, increased costs, and a negative environmental impact. Accurate prediction of steam demand is therefore essential for optimising digester cooking cycles and resource allocation. This study aims to develop and compare predictive models for steam demand in batch pulp digesters using magnesium bisulphite cooking liquor. Three years of production data were pre-processed to extract digester temperature profiles and batch steam demands. Seven modelling approaches were evaluated: a mechanistic first-principles energy balance model, Box–Jenkins ARIMA, two neural network models (LSTM and CNN), and three hybrid models combining first-principles with ARIMA, LSTM, and CNN. The hybrid frameworks employed dimensionless parameters from the mechanistic model as exogenous variables to compensate for unavailable process data. Model accuracy was assessed using RMSE and MAE metrics. The results show that hybrid models consistently outperformed their standalone counterparts. In particular, the hybrid first-principles–CNN model achieved the highest predictive accuracy, demonstrating the CNN’s ability to extract features and capture nonlinear temporal dependencies in steam demand. The hybrid first-principles–ARIMA model also surpassed both the standalone ARIMA and mechanistic models. Integrating mechanistic insights with data-driven methods significantly enhances prediction accuracy in complex batch processes. The findings highlight the value of hybrid modelling strategies for improving steam demand forecasting, with potential benefits for process optimisation, energy efficiency, and batch scheduling in the pulp and paper industry. Copyright © 2025 by Authors, Published by Universitas Diponegoro and BCREC Publishing Group. This is an open access article under the CC BY-SA License (https://creativecommons.org/licenses/by-sa/4.0).

Supporting Information (SI) PDF
Keywords: Steam demand forecast; Batch pulp digester; Hybrid Modelling; Neural Networks; First-principles models
Funding: University of the Witwatersrand

Article Metrics:

Article Info
Section: Original Research Articles
Language : EN
Recent articles
Conversion Improvement of Propylene to Acrylic Acid Process Hybrid Approaches for Steam Demand Forecasting: Combining First Principles, Box and Jenkins, and Neural Network Models Process Intensification of Hydrodealkylation (HDA) for Benzene Production through Heat Integration and Gas Recycle Optimization More recent articles
  1. Bajpai, P. (2015). Pulp and Paper Industry: Chemicals. 1st ed. Elsevier
  2. May, R., Maier, H., Dandy, G., Fernando, T. (2008). Non-linear variable selection for artificial neural networks using partial mutual information. Environmental Modelling & Software, 23(10-11). 1312-1326. DOI: 10.1016/j.envsoft.2008.03.007
  3. Khashei, M., Bijari, M. (2011). A novel hybridization of artificial neural networks and ARIMA models for time series forecasting. Applied Soft Computing, 11(2), 2664-2675. DOI: 10.1016/j.asoc.2010.10.015
  4. Kwon, Y., Kim, S., Choi, Y., Kang, S. (2022). Generative modeling to predict multiple suitable conditions for chemical reactions. Journal of Chemical Information and Modeling, 62(23), 5952-5960. DOI: 10.1021/acs.jcim.2c01085
  5. Gallegos, L., Luchini, G., St. John, P., Kim, S., Paton, R. (2021). Importance of engineered and learned molecular representations in predicting organic reactivity, selectivity, and chemical properties. Accounts of Chemical Research, 54(4), 827-836. DOI: 10.1021/acs.accounts.0c00745
  6. Sadrzadeh, M., Mohammadi, T., Ivakpour, J., Kasiri, N. (2008). Separation of lead ions from wastewater using electrodialysis: comparing mathematical and neural network modeling. Chemical Engineering Journal, 144(3), 431-441. DOI: 10.1016/j.cej.2008.02.023
  7. Lashkarbolooki, M., Vaferi, B., Rahimpour, M.R. (2011). Comparing the capability of artificial neural network (ANN) and EOS for prediction of solid solubilities in supercritical carbon dioxide. Fluid Phase Equilibria, 308(1-2), 35-43. DOI: 10.1016/j.fluid.2011.06.002
  8. Grondin-Perez, B., Roche, S., Lebreton, C., Benne, M., Damour, C., Kadjo, A.J.J. (2014). Mechanistic model versus artificial neural network model of a single-cell pemfc. Engineering, 6(8), 418-426. DOI: 10.4236/eng.2014.68044
  9. Miheţ, M., Cristea, V.M., Agachi, P.Ş. (2009). FCCU simulation based on first principle and artificial neural network models. Asia‐Pacific Journal of Chemical Engineering, 4(6), 878-884. DOI: 10.1002/apj.312
  10. Li, K., Duan, H., Liu, L., Qiu, R., van den Akker, B., Ni, BJ., Chen, T., Yin, H., Yuan, Z., Ye, L. (2022). An integrated first principal and deep learning approach for modeling nitrous oxide emissions from wastewater treatment plants. Environmental Science & Technology, 56(4), 2816-2826. DOI: 10.1021/acs.est.1c05020
  11. Faruk, D.Ö. (2010). A hybrid neural network and ARIMA model for water quality time series prediction. Engineering Applications of Artificial Intelligence, 23(4), 586-594. DOI: 10.1016/j.engappai.2009.09.015
  12. Ghahnavieh, A.E. (2019). Time series forecasting of styrene price using a hybrid ARIMA and neural network model. Independent Journal of Management & Production, 10(3), 915-933. DOI: 10.14807/ijmp.v10i3.877
  13. Phatwong, A., Koolpiruck, D. (2019). Kappa number prediction of pulp digester using LSTM neural network. 16th International Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology (ECTI-CON), 151-154
  14. Shah, P., Choi, H.K., Kwon, J.S.I. (2023). Achieving optimal paper properties: A layered multiscale kMC and LSTM-ANN-based control approach for kraft pulping. Processes, 11(3), 809. DOI: 10.3390/pr11030809
  15. Correia, F.M., d’Angelo, J.V.H., Zemp, R.J., Mingoti, S.A. (2014). Prediction of kappa number in eucalyptus kraft pulp continuous digester using the Box & Jenkins methodology. Advances in Chemical Engineering and Science, 4(4), 539-547. DOI: 10.4236/aces.2014.44055
  16. Shao, B., Hu, X., Bian, G., Zhao, Y. (2019). A multichannel LSTM‐CNN method for fault diagnosis of chemical process. Mathematical Problems in Engineering, 2019(1). DOI: 10.1155/2019/1032480
  17. Chen, H., Cen, J., Yang, Z., Si, W., Cheng, H. (2022). Fault diagnosis of the dynamic chemical process based on the optimized CNN-LSTM network. ACS Omega, 7(38), 34389-34400. DOI: 10.1021/acsomega.2c04017
  18. Chen, L., Li, S., Bai, Q., Yang, J., Jiang, S., Miao, Y. (2021). Review of image classification algorithms based on convolutional neural networks. Remote Sensing, 13(22), 4712. DOI: 10.3390/rs13224712
  19. Fogler, H.S. (2006). Elements of Chemical Reaction Engineering. 4th ed. Prentice Hall
  20. Lai, K., Mishra, S., Panda, G., Chakraborty, S., Samei, M., Ram, B. (2020). A limited memory q-BFGS algorithm for unconstrained optimization problems. Journal of Applied Mathematics and Computing, 2021(66), 183-202. DOI: 10.1007/s12190-020-01432-6
  21. Ho, S., Xie, M., Goh, T. (2002). A comparative study of neural network and Box-Jenkins ARIMA modeling in time series prediction. Computers & Industrial Engineering, 42(2-4), 371-375. DOI: 10.1016/S0360-8352(02)00036-0
  22. Vani, S., Rao, T., Naidu, C. (2019). Comparative analysis on variants of neural networks: an experimental study. 5th International Conference on Advanced Computing & Communication Systems, 429-434. DOI: 10.1109/ICACCS.2019.8728327
  23. Liu, S., Ji, H., Wang, M. (2020). Nonpooling convolutional neural network forecasting for seasonal time series with trends. IEEE Transactions on Neural Networks and Learning Systems, 31(8), 2879-2888. DOI: 10.1109/TNNLS.2019.2934110
  24. Chicco, D., Warrens, M., Jurman, G. (2021). The coefficient of determination R-squared is more informative than SMAPE, MAE, MAPE, MSE and RMSE in regression analysis evaluation. PeerJ Computer Science, 7. DOI: 10.7717/peerj-cs.623
  25. Filhoa, R., Aguiar, H., Polowskia, N. (2005). Hybrid modelling development for a continuous industrial kraft pulping digester. European symposium on computer aided process engineering: ESCAPE-15 conference paper
  26. Son, S., Choi, H., Moon, J., Kwon, J. (2022). Hybrid Koopman model predictive control of nonlinear systems using multiple EDMD models: An application to a batch pulp digester with feed fluctuation. Control Engineering Practice, 118(14), 104956. DOI: 10.1016/j.conengprac.2021.104956

Last update:

No citation recorded.

Last update:

No citation recorded.