Home Comparing classic time series models and state-of-the-art time series neural networks for forecasting as-fired liquor properties
Article
Licensed
Unlicensed Requires Authentication

Comparing classic time series models and state-of-the-art time series neural networks for forecasting as-fired liquor properties

  • Jerry Ng ORCID logo EMAIL logo , Yuri Lawryshyn and Nikolai DeMartini
Published/Copyright: November 29, 2024
Become an author with De Gruyter Brill
Submit Manuscript Author Information
Nordic Pulp & Paper Research Journal
From the journal Nordic Pulp & Paper Research Journal Volume 40 Issue 1

Abstract

The properties of as-fired black liquor dictate kraft recovery boiler operation. If these properties could be forecasted, operations could be adjusted to optimize boiler performance. Here, we compare the performances of classic time series models and two state-of-the-art time series neural networks for forecasting as-fired liquor heating value, viscosity, and boiling point rise at a Canadian mill. Additionally, we show that, like classic time series models, autoregressive neural networks can be regarded as functions of unknown disturbances, which is useful in comparing model complexities. Our results show that classic time series models can accurately forecast as-fired liquor properties and that classic time series models perform comparably to state-of-the-art time series neural networks. We suspect this is due to the high autocorrelation of mill data that results from frequent measurements relative to long residence times. This autocorrelation is suspected to attenuate the cross-correlations between upstream disturbances and as-fired liquor properties. As a result, neural networks, which are useful for accommodating non-linear cross-correlations and dynamics, struggle to outperform classic time series models and may not always be appropriate for forecasting chemical process parameters.

Keywords: black liquor; time series; neural networks; recovery boiler; machine learning
Corresponding author: Jerry Ng, Department of Chemical Engineering and Applied Chemistry, University of Toronto Faculty of Applied Science & Engineering, Toronto, Ontario, Canada, E-mail:

Funding source: Natural Sciences and Engineering Research Council of Canada

Award Identifier / Grant number: IRCPJ 517715 -16

  1. Research ethics: Not applicable.

  2. Informed consent: Not applicable.

  3. Author contributions: The authors have accepted responsibility for the entire content of this manuscript and approved its submission. Jerry Ng: methodology, data acquisition, analysis, writing. Yuri Lawryshyn: supervision, editing. Nikolai DeMartini: supervision, editing, funding.

  4. Conflict of interest: The authors state no conflict of interest.

  5. Use of Large Language Models, AI and Machine Learning Tools: None declared.

  6. Research funding: This work is part of the industrial research consortium project “Effective Energy and Chemical Recovery in Pulp and Paper Mills – III” at the University of Toronto. Additionally, this work was supported by the Natural Sciences and Engineering Research Council of Canada.

  7. Data availability: The raw data can be obtained on request from the corresponding author.

References

Adams, T.N. (1997). Kraft recovery boilers. TAPPI Press, Atlanta, GA.Search in Google Scholar

Aguiar, H.C. and Filho, R.M. (2001). Neural network and hybrid model: a discussion about different modeling techniques to predict pulping degree with industrial data. Chem. Eng. Sci. 56: 6–570, https://doi.org/10.1016/s0009-2509(00)00261-x.Search in Google Scholar

Alabi, S.B. and Williamson, C.J. (2011). Centrifugal pump-based predictive models for kraft black liquor viscosity: an artificial neural network approach. Ind. Eng. Chem. Res. 50: 10320–10328, https://doi.org/10.1021/ie200684n.Search in Google Scholar

Alabi, S. and Williamson, C. (2015). Neural network-based model for joint prediction of the Newtonian and non-Newtonian viscosities of black liquor. Int. J. Chem. Eng. Appl. 6: 195–200, https://doi.org/10.7763/IJCEA.2015.V6.480.Search in Google Scholar

Alabi, S.B., Williamson, C.J., and Lee, J. (2010). Viscosity models for New Zealand black liquor at low solids concentrations. Asia-Pac. J. Chem. Eng. 5: 619–625, https://doi.org/10.1002/apj.444.Search in Google Scholar

Amazouz, M. and Platon, R. (2011) Soft-sensors for real-time monitoring and control of a black liquor concentration process. In: Proceedings of the international conference on data science. The Steering Committee of The World Congress in Computer Science, Computer Engineering, & Applied Computing.Search in Google Scholar

Åström, K.J. and Murray, R.M. (2008). Feedback systems: an introduction for scientists and engineers. Princeton University Press, Princeton, NJ.10.1515/9781400828739Search in Google Scholar

Belisario, A.B., Edberg, A., Bjork, M., Almeida, G.M., and Vakkilainen, E. (2023). On the diagnosis of a fouling condition in a kraft recovery boiler: combining process knowledge and data-based insights. TAPPI J. 22: 162–171, https://doi.org/10.32964/tj22.3.162.Search in Google Scholar

Bennington, C.P.J. (1983). The effect of temperature on drop size of black liquor sprays, Master’s thesis. University of British Columbia, Available at: https://open.library.ubc.ca/soa/cIRcle/collections/ubctheses/831/items/1.0058907.Search in Google Scholar

Bikmukhametov, T. and Jäschke, J. (2020). Combining machine learning and process engineering physics towards enhanced accuracy and explainability of data-driven models. Comput. Chem. Eng. 138: 106834, https://doi.org/10.1016/j.compchemeng.2020.106834.Search in Google Scholar

Box, G.E.P., Jenkins, G.M., Reinsel, G.C., and Ljung, G.M. (2015). Time series analysis: forecasting and control, 5th ed John Wiley & Sons, Hoboken, NJ.Search in Google Scholar

Bradley, W., Kim, J., Kilwein, Z., Blakely, L., Eydenberg, M., Jalvin, J., Laird, C., and Boukouvala, F. (2022). Perspectives on the integration between first-principles and data-driven modeling. Comput. Chem. Eng. 166: 107898, https://doi.org/10.1016/j.compchemeng.2022.107898.Search in Google Scholar

Bucolo, M., Fortuna, L., Nelke, M., Rizzo, A., and Sciacca, T. (2002). Prediction models for the corrosion phenomena in pulp & paper plant. Control Eng. Pract. 10: 227–237, https://doi.org/10.1016/S0967-0661(01)00126-5.Search in Google Scholar

Correia, F.M., d’Angelo, J.V.H., Almeida, G.M., and Mingoti, S.A. (2018). Predicting kappa number in a kraft pulp continuous digester: a comparison of forecasting methods. Braz. J. Chem. Eng. 35: 1081–1094, https://doi.org/10.1590/0104-6632.20180353s20160678.Search in Google Scholar

De Carli, F. and Mejia, R.I.G. (2019). Advanced analytics applied to problem solving technique on a pulp mill. Latin Am. Appl. Res.: 6.10.52292/j.laar.2019.355Search in Google Scholar

Dürr, O., Sick, B., and Murina, E. (2020). Probabilistic deep learning: with python, keras and tensorflow probability. Manning Publications, Shelter Island, NY.Search in Google Scholar

Gäärd, P. (2004). Kappa control with online analyzer using samples from the digester’s mid-phase, Master’s thesis. Linköping University, Available at: http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-2209.Search in Google Scholar

Galicia, H.J., Peter He, Q., and Wang, J. (2011). A reduced order soft sensor approach and its application to a continuous digester. J. Process Control 21: 489–500, https://doi.org/10.1016/j.jprocont.2011.02.001.Search in Google Scholar

Goodfellow, I., Bengio, Y., and Courville, A. (2016). Deep learning. MIT Press, Available at: http://www.deeplearningbook.org.Search in Google Scholar

Hajiha, H. (2009). Multivariate analysis of variables affecting thermal performance of black liquor evaporators, Master’s thesis. University of Toronto.Search in Google Scholar

Hart, P.W. (2022). Chemical recovery in the alkaline pulping processes, 4th ed TAPPI, Peachtree Corners, GA.Search in Google Scholar

Heikkinen, M., Heikkinen, T., and Hiltunen, Y. (2008) Modelling of activated sludge treatment process in a pulp mill using neural networks. In: Proceedings of the 6th international conference on computing, communications and control technologies, pp. 5.Search in Google Scholar

Hüllen, G., Zhai, J., Kim, S.H., Sinha, A., Realff, M.J., and Boukouvala, F. (2020). Managing uncertainty in data-driven simulation-based optimization. Comput. Chem. Eng. 136: 106519, https://doi.org/10.1016/j.compchemeng.2019.106519.Search in Google Scholar

Jones, A., Wagoner, J., Michaelson, T., and Tran, H. (2014) Use of multivariate analysis to understand the root causes of premature smelt spout failures on recovery boilers. In: Proceedings of the international chemical recovery conference, Tampere, Finland, pp. 10.Search in Google Scholar

Karsimus, M. (2021). Applying machine learning based soft sensors to a pulp refining process, Master’s thesis. Aalto University.Search in Google Scholar

Korsbakke, A. and Lidmark, J. (2021). Forecasting alarms using machine learning: predicting tall oil production at södra cell, Master’s thesis. Blekinge Institute of Technology, Available at: http://urn.kb.se/resolve?urn=urn:nbn:se:bth-21939.Search in Google Scholar

Ljung, L. (1998). System identification, 2nd ed Springer, London.10.1002/047134608X.W1046Search in Google Scholar

Meyer, T., Barquin, A., Gionest, P., and Tran, H. (2020). Multivariate data analysis of full-scale sludge dewatering. J. Environ. Chem. Eng. 8: 103648, https://doi.org/10.1016/j.jece.2019.103648.Search in Google Scholar

Miikkulainen, P. (2006). Spray formation of high dry solids black liquor in recovery boiler furnaces, Ph.D. thesis. Helsinki University of Technology.Search in Google Scholar

Miikkulainen, P., Kankkunen, A., Järvinen, M.P., and Fogelholm, C.-J. (2005). Predicting droplet size from black liquor spray characteristics. TAPPI J. 8: 36–40, https://doi.org/10.32964/tj8.1.36.Search in Google Scholar

Montgomery, D.C. (2009). Introduction to statistical quality control, 7th ed John Wiley & Sons, Hoboken, NJ.Search in Google Scholar

Nykyri, M., Kuisma, M., Kärkkäinen, T.J., Junkkari, T., Kerkelä, K., Puustinen, J., Myrberg, J., and Hallikas, J. (2019) Predictive analytics in a pulp mill using factory automation data—hidden potential. In: Proceedings of the 17th international conference on industrial informatics, Vol. 1. IEEE, pp. 1014–1020.10.1109/INDIN41052.2019.8972070Search in Google Scholar

Orang, N., Tran, H., Jones, A., and Donald Jones, F. (2017). Operating parameters affecting the thermal performance of biomass boilers. TAPPI J. 16: 453–461, https://doi.org/10.32964/TJ16.8.453.Search in Google Scholar

Oreshkin, B.N., Dudek, G., Pełka, P., and Turkina, E. (2021). N-BEATS neural network for mid-term electricity load forecasting. Appl. Energy 293: 116918, https://doi.org/10.1016/j.apenergy.2021.116918.Search in Google Scholar

Pereira Parente, A., de Souza, M.B.Jr., Valdman, A., and Folly, R.O.M. (2019). Data augmentation applied to machine learning-based monitoring of a pulp and paper process. Processes 7: 958, https://doi.org/10.3390/pr7120958.Search in Google Scholar

Phatwong, A. and Koolpiruck, D. (2019) Kappa number prediction of pulp digester using LSTM neural network. In: 2019 16th international conference on electrical engineering/electronics, computer, telecommunications and information technology (ECTI-CON), pp. 151–154.10.1109/ECTI-CON47248.2019.8955373Search in Google Scholar

Rahman, M., Dahlquist, E., and Kyprianidis, K. (2022). Modelling and diagnostics of process faults in continuous pulp digesters. Comput. Chem. Eng. 157: 107589, https://doi.org/10.1016/j.compchemeng.2021.107589.Search in Google Scholar

Rogerson, A.M. (2022). Dynamic modelling of process chemistry in kraft pulp mills, Master’s thesis. University of Toronto, Available at: https://tspace.library.utoronto.ca/handle/1807/124895.Accepted:2022-11-11T16:03:01Z.Search in Google Scholar

Sainlez, M. and Heyen, G. (2010) Performance monitoring of an industrial boiler: classification of relevant variables with random forests. In: Pierucci, S. and Buzzi Ferraris, G. (Eds.). Computer aided chemical engineering, volume 28 of 20 European symposium on computer aided process engineering. Elsevier, pp. 403–408, Available at: http://www.sciencedirect.com/science/article/pii/S1570794610280689.10.1016/S1570-7946(10)28068-9Search in Google Scholar

Sainlez, M. and Heyen, G. (2011) Recurrent neural network prediction of steam production in a kraft recovery boiler. In: Computer aided chemical engineering, Vol. 29. Elsevier, pp. 1784–1788, Available at: https://linkinghub.elsevier.com/retrieve/pii/B9780444542984501355.10.1016/B978-0-444-54298-4.50135-5Search in Google Scholar

Sainlez, M. and Heyen, G. (2013). Comparison of supervised learning techniques for atmospheric pollutant monitoring in a kraft pulp mill. J. Comput. Appl. Math. 246: 329–334, https://doi.org/10.1016/j.cam.2012.06.026.Search in Google Scholar

Salinas, D., Flunkert, V., Gasthaus, J., and Januschowski, T. (2020). DeepAR: probabilistic forecasting with autoregressive recurrent networks. Int. J. Forecast. 36: 1181–1191, https://doi.org/10.1016/j.ijforecast.2019.07.001.Search in Google Scholar

Shah, P., Choi, H.-K., and Kwon, J.S.-I. (2023). Achieving optimal paper properties: a layered multiscale kMC and LSTM-ANN-based control approach for kraft pulping. Processes 11: 809, https://doi.org/10.3390/pr11030809.Search in Google Scholar

Smook, G.A. (2016). Handbook for pulp & paper technologists, 4th ed TAPPI, Peacetree Corners, GA.Search in Google Scholar

Soualhi, M., Ragab, A., Nguyen, T.P.K., Medjaher, K., Ghezzaz, H., and Amazouz, M. (2021). Data analytics methodology for construction of fouling prognostic indicators: towards cost-effective maintenance scheduling. PHM Soc. Eur. Conf. 6: 8, https://doi.org/10.36001/phme.2021.v6i1.2886.Search in Google Scholar

Souza, T., Tran, H., and Repka, J. (2006) Effects of process variabilty on ring formation in lime kilns. In: TAPPI EPE conference proceedings. TAPPI, Atlanta, Georgia.Search in Google Scholar

Talebjedi, B., Laukkanen, T., Holmberg, H., Vakkilainen, E., and Syri, S. (2021). Energy simulation and variable analysis of refining process in thermo-mechanical pulp mill using machine learning approach. Math. Comput. Model. Dyn. Syst. 27: 562–585, https://doi.org/10.1080/13873954.2021.1990967.Search in Google Scholar

Talebjedi, B., Laukkanen, T., Holmberg, H., Vakkilainen, E., and Syri, S. (2022). Advanced energy-saving optimization strategy in thermo-mechanical pulping by machine learning approach. Nord. Pulp Pap. Res. J. 37: 434–452, https://doi.org/10.1515/npprj-2022-0013.Search in Google Scholar

Tran, H. (2018). Kraft recovery boilers, 3rd ed TAPPI, Peacetree Corners, GA.Search in Google Scholar

Versteeg, P. and Tran, H. (2009). Monitoring kraft recovery boiler fouling using principal component analysis. TAPPI J.: 9.10.32964/TJ8.11.22Search in Google Scholar

Wallinder, J., Lindgren, K., and Samuelsson, A. (2018) Dynamic modelling of a kraft pulp mill producing softwood and hardwood pulp in campaigns. In: TAPPI pulping engineering and environmental recycling sustainability conference proceedings, Portland, United States, pp. 673–683.Search in Google Scholar

Yule, G.U. (1927). VII. On a method of investigating periodicities disturbed series, with special reference to Wolfer’s sunspot numbers. Philos. Trans. R. Soc. Lond. – Ser. A Contain. Pap. a Math. or Phys. Character 226: 267–298, https://doi.org/10.1098/rsta.1927.0007.Search in Google Scholar

Zaman, A.A. and Fricke, A.L. (1996). Effect of pulping conditions and black liquor composition on Newtonian viscosity of high solids kraft black liquors. Ind. Eng. Chem. Res. 35: 590–597, https://doi.org/10.1021/ie950202h.Search in Google Scholar

Zaman, A.A., McNally, T.W., and Fricke, A.L. (1998). Vapor pressure and boiling point elevation of slash pine black liquors: predictive models with statistical approach. Ind. Eng. Chem. Res. 37: 275–283, https://doi.org/10.1021/ie9705315.Search in Google Scholar

Zhang, Y., Suzuki, G., and Shioya, H. (2022). Prediction and detection of sewage treatment process using N-BEATS autoencoder network. IEEE Access 10: 112594–112608, https://doi.org/10.1109/ACCESS.2022.3216924.Search in Google Scholar
Received: 2024-04-11
Accepted: 2024-06-24
Published Online: 2024-11-29
Published in Print: 2025-03-26

© 2024 Walter de Gruyter GmbH, Berlin/Boston

Articles in the same Issue

  1. Frontmatter
  2. Biorefining
  3. Fractionation methods of eucalyptus kraft lignin for application in biorefinery
  4. Pulp and paper industry side-stream materials as feed for the oleaginous yeast species Lipomyces starkeyi and Rhodotorula toruloides
  5. Chemical Pulping
  6. Comparing classic time series models and state-of-the-art time series neural networks for forecasting as-fired liquor properties
  7. Optimization of kraft pulping process for Sesbania aculeata (dhaincha) stems using RSM
  8. On the nature of the selectivity of oxygen delignification
  9. Unlocking potential: the role of chemometric modeling in pulp and paper manufacturing
  10. Effects of chemical environment on softwood kraft pulp: exploring beyond conventional washing methods
  11. Bleaching
  12. Variations in carbohydrates molar mass distribution during chemical degradation and consequences on fibre strength
  13. Mechanical Pulping
  14. Energy consumption in refiner mechanical pulping
  15. Paper Technology
  16. Australian wheat and hardwood fibers for advanced packaging materials
  17. Compression refining: the future of refining? Application to bleached kraft eucalyptus pulp
  18. The effect of nanocellulose to coated paper and recycled paper
  19. Interpreting the relationship between properties of wood and pulping & paper via machine learning algorithms combined with SHAP analysis
  20. Hybridization to prepare environmentally friendly, cost-effective superhydrophobic oleophobic coatings
  21. Paper Physics
  22. Characterising the mechanical behaviour of dry-formed cellulose fibre materials
  23. Paper Chemistry
  24. Study on the properties of ground film paper prepared from lactic acid-modified cellulose
  25. Environmental Impact
  26. Characterization of sludge from a cellulose pulp mill for its potential biovalorization
  27. The in situ green synthesis of metal organic framework (HKUST-1)/cellulose/chitosan composite aerogel (CSGA/HKUST-1) and its adsorption on tetracycline
  28. Evaluation of the potential use of powdered activated carbon in the treatment of effluents from bleached kraft pulp mills
  29. Recycling
  30. Waste newspaper activation by sodium phosphate for adsorption dynamics of methylene blue
https://doi.org/10.1515/npprj-2024-0025
Keywords for this article
black liquor; time series; neural networks; recovery boiler; machine learning

Articles in the same Issue

  1. Frontmatter
  2. Biorefining
  3. Fractionation methods of eucalyptus kraft lignin for application in biorefinery
  4. Pulp and paper industry side-stream materials as feed for the oleaginous yeast species Lipomyces starkeyi and Rhodotorula toruloides
  5. Chemical Pulping
  6. Comparing classic time series models and state-of-the-art time series neural networks for forecasting as-fired liquor properties
  7. Optimization of kraft pulping process for Sesbania aculeata (dhaincha) stems using RSM
  8. On the nature of the selectivity of oxygen delignification
  9. Unlocking potential: the role of chemometric modeling in pulp and paper manufacturing
  10. Effects of chemical environment on softwood kraft pulp: exploring beyond conventional washing methods
  11. Bleaching
  12. Variations in carbohydrates molar mass distribution during chemical degradation and consequences on fibre strength
  13. Mechanical Pulping
  14. Energy consumption in refiner mechanical pulping
  15. Paper Technology
  16. Australian wheat and hardwood fibers for advanced packaging materials
  17. Compression refining: the future of refining? Application to bleached kraft eucalyptus pulp
  18. The effect of nanocellulose to coated paper and recycled paper
  19. Interpreting the relationship between properties of wood and pulping & paper via machine learning algorithms combined with SHAP analysis
  20. Hybridization to prepare environmentally friendly, cost-effective superhydrophobic oleophobic coatings
  21. Paper Physics
  22. Characterising the mechanical behaviour of dry-formed cellulose fibre materials
  23. Paper Chemistry
  24. Study on the properties of ground film paper prepared from lactic acid-modified cellulose
  25. Environmental Impact
  26. Characterization of sludge from a cellulose pulp mill for its potential biovalorization
  27. The in situ green synthesis of metal organic framework (HKUST-1)/cellulose/chitosan composite aerogel (CSGA/HKUST-1) and its adsorption on tetracycline
  28. Evaluation of the potential use of powdered activated carbon in the treatment of effluents from bleached kraft pulp mills
  29. Recycling
  30. Waste newspaper activation by sodium phosphate for adsorption dynamics of methylene blue
Sign up now to receive a 20% welcome discount
Institutional Access
How does access work?
Have an idea on how to improve our website?
Please write us.
© 2025 De Gruyter Brill
Downloaded on 15.9.2025 from https://www.degruyterbrill.com/document/doi/10.1515/npprj-2024-0025/html
Scroll to top button