Abstract
The management of ferromanganese production requires the monitoring and analysis of a large amount of data. This case study explores how much information and value can be extracted from the process monitoring data using data-driven methods. The study describes data preparation, cleaning and the application of unsupervised learning methods to eventually summarize over 140 monitored variables in just two statistically independent time series. Although this analysis helped to condense the information to humanly consumable form, to further extract the value from this information requires the input of a specialist with the domain-specific knowledge and experience. As a result, it is suggested to use machine learning methods outputs as virtual advisors to aid decision-making done by domain experts.
Similar content being viewed by others
Notes
Sadly, there is no published paper that describes the algorithm.
References
J.D. Steenkamp, D. Chetty, A. Singh, S.A.C. Hockaday and G.M. Denton, J. Metall. 72, 3422 (2020)
R.T. Jones, Fundamental Aspects of Alloy Smelting in a DC Arc Furnace, PhD Thesis (University of the Witwatersrand, 2015), p. 45.
S.E. Olsen, M. Tangstad and T. Lindstad, Production of Manganese Ferroalloys (Tapir Akademisk Forlag, Trondheim, Norway, 2007), p. 218
ASTM-A99, Standard Specification for Ferromanganese (West Conshohocken, PA, 2009), https://www.astm.org/. Accessed on 26 July 2021.
ASTM-A483, Standard Specification for Silicomanganese (West Conshohocken, PA, 2010). https://www.astm.org/. Accessed on 26 July 2021.
M. Tangstad, Handbook of Ferroalloys, ed. M. Gasik (Elsevier, 2013), p. 221.
T. Lindstad, S.E. Olsen, G. Tranell, T. Færden and J. Lubetsky, INFACON XI, Proceedings of the Innovation Ferro-Alloy Industry (2007), p. 18.
E. Cairncross, Assessment of Eskom coal fired power stations for compliance with their 1 April 2015 AELs over the period 1 April 2015 to 31 March 2016; and ranking of their pollutant and \({\rm CO}_2\) emission intensities (2017), https://cer.org.za/wp-content/uploads/2016/07/AEL-Compliance-Assessment-of-Eskom-CFPSs-final-19-May-2017_final.pdf. Accessed on 14 June 2021.
J.E. Davidsen and M. Honstad, J. South. Afr. Inst. Min. Metall. 119, 545 (2019)
J.D. Steenkamp, P. Maphutha, O. Makwarela, W.K. Banda, I. Thobadi, M. Sitefane, J. Gous and J.J. Sutherland, J. South. Afr. Inst. Min. Metall. 118, 309 (2018)
G. James, D. Witten, T. Hastie and R. Tibshirani, An Introduction to Statistical Learning with Applications in R (Springer, 2017), pp. 1–17, 165, 374–385, 210–213.
G. Meyer, G. Adomavicius, P.E. Johnson, M. Elidrisi, W.A. Rush, J.M. Sperl-Hillen and P.J. O’Connor, ISR 25, 239 (2014)
J.T. McCoy and L. Auret, Miner. Eng. 132, 95 (2019)
J. McNiff, All You Need to Know About Action Research (SAGE, Thousand Oaks, CA, 2006), pp. 7–15
M.S. Rennie, W.D. Carew, R. Terblanche, A.L. Moolman and N. Anthony, Information management in the ferroalloy business (Mintek, 2001), https://www.pyro.co.za/InfaconIX/108-Rennie.pdf. Accessed on 26 July 2021.
G.W. Miligan and M.C. Cooper, J. Classif. 5, 181 (1988)
N. Marwan, M. Carmen Romano, M. Thiel and J. Kurths, Phys. Rep. 438, 237 (2007)
J.-G. Wang, Y. Wang, Y. Yao, B.-H. Yang and S.-W. Ma, Control Eng. Pract. 88, 110 (2019)
A. Hyvärinen and E. Oja, Neural Netw. 13, 411 (2000)
N.E. Helwig, ICA: Independent Component Analysis. R package version 1.0-2 (2018), https://CRAN.R-project.org/package=ica. Accessed on 26 July 2021.
C.-C.M. Yeh, Y. Zhu, L. Ulanova, N. Begum,Y. Ding, H.A. Dau, D. Furtado Silva, A. Mueen and E. Keogh, Conference Data Mining (IEEE, 2016), p. 1317.
F. Bischoff, Matrixprofiler: Matrix Profile for R. R Package Version 0.1.5 (2021), https://CRAN.R-project.org/package=matrixprofiler. Accessed on 26 July 2021.
C. Rackauckas, Y. Ma, J. Martensen, C. Warner, K. Zubov, R. Supekar, D. Skinner, A. Ramadhan and A. Edelman, Universal Differential Equations for Scientific Machine Learning (arXiv:2001.04385 [cs.LG], 2020). arXiv:2001.04385v3. Accessed on 26 July 2021.
Acknowledgements
The financial assistance of the National Research Foundation (NRF) towards this research is hereby acknowledged. Opinions expressed and conclusions arrived at are those of the author and are not necessarily to be attributed to the NRF.
This paper is published by permission of Mintek and Transalloys.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
On behalf of all authors, the corresponding author states that there is no conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Cherkaev, A.V., Reynolds, Q.G. & Steenkamp, J.D. Towards Application of Machine Learning Methods in Pyrometallurgy: A Case Study of an Exploratory Data Analysis for Ferromanganese Production. JOM 74, 47–52 (2022). https://doi.org/10.1007/s11837-021-05023-z
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11837-021-05023-z