LLC Nornickel Sputnik, Moscow, Russia

Е. V. Nasvishchuk, Chief Manager, Internal Development and Attraction of External Contractors of the Mineral Resources Complex, e-mail: NasvischukEV@nornik.ru

Polar Branch of PJSC MMC Norilsk Nickel, Norilsk, Russia

R. V. Ushakov, Head of the Production Automation Service of the Talnakh Concentrator, e-mail: ushakovrv@nornik.ru
D. N. Demydko, Head of the X-ray Spectral Laboratory of the Production Automation Service of the Talnakh Concentrator, e-mail: demydkodn@nornik.ru
М. I. Prikazchikova, First Category Engineer of the X-ray Spectral Laboratory of the Production Automation Service of the Talnakh Concentrator, e-mail: PrikazchikovaMI@nornik.ru

Modern production requires operational control of key parameters. These data directly affect the process management and ensure that the specified quality standards of the finished product are achieved. Any processing plant contains a variety of monitoring devices that provide operational monitoring. Enrichment is a technological process aimed at increasing the content of target metals in concentrates, exceeding their initial content in the processed ore. Using X-ray fluorescence analyzers, the contents of metals and sulfur in accumulative balance and other powder samples are determined. The cumulative nature of the samples does not allow the use of such data in operational management. An alternative is rapid analysis on X-ray multichannel pulp analyzers, which provide information about the intensity of the spectra of substances interaction. These data make it possible to reconstruct the contents of metals and sulfur through the coupling equations. It is necessary to maintain the relevance of these equations, i.e. periodic calibration of pulp analyzers. An additional complexity factor in this case is the variability of the composition of the material (ore or pulp), which leads to the need for constant monitoring of the qua lity of the contents recovery. Increasing the frequency of oscillation and quality control of calibration models requires the introduction of automated tools, in particular, based on machine learning, which can be used by personnel of X-ray spectral analysis laboratories. The article provides an example of automatic calibration implemented at the Talnakh Concentrator. It is based on the enumeration and evaluation of models of a known structure (coupling equations for recovery of the nickel and copper content with regressors of a given type). The situation of recovery of the sulfur content is considered separately: the complexity of enhancement (technological improvements) and modeling results. Using machine learning methods, the possibility of estimating the sulfur content based on the intensities of the bands of the nickel and copper spectrum has been studied. The error in restoring the operational values of the sulfur content for the method used is 3% (MAPE).
Specialists of PJSC MMC Norilsk Nickel took part in the work: M. S. Datsiev, A. G. Aryshtaev, D. I. Ivashechkin, I. F. Zaporozhtsev, V. Yu. Ivanov, E. M. Strelet skaya and others.

1. Abrarov A. D., Datsiev M. S., Chikildin D. E., Fedotov D. N. Optimization of bulk flotation process at Talnakh Concentrator based on machine learning algorithms. Tsvetnye Metally. 2022. No. 2. pp. 87–93.
2. Datsiev M. S., Lesnikova L. S., Dzansolov I. V., Ananko I. A. Comprehensive use of raw materials through optimization of the concentration and metallurgical complex of Norilsk division. Tsvetnye Metally. 2022. No. 2. pp. 12–17.
3. Krupnov L. V., Midyukov D. O., Datsiev M. S., Ilin V. B. Change in mineral resource supplies in production of heavy nonferrous metals in terms of copper and nickel. Gornyi Zhurnal. 2024. No. 3. pp. 10–16.
4. Krupnov L. V., Midyukov D. O., Malakhov P. V. Ways to cover the raw material demand in the copper-nickel sector. Obogashchenie Rud. 2022. No. 2. pp. 37–41.
5. Krupnov L. V., Malakhov P. V., Ozerov S. S., Pakhomov R. A. Analyzing Russian cobalt metallurgy and ways to raise recovery. Tsvetnye Metally. 2023. No. 7. pp. 25–33.
6. Osipova N. V. Review of projects and solutions for digital twins for processing plants. Avtomatizatsiya v promyshlennosti. 2023. No. 7. pp. 37–42.
7. Oosthuizen D., Williams B., Van der Spuy D. Barriers to integrated flotation control solutions: lessons from an industrial implementation. Proceedings of Internationsl Federation of Automatic Control Conference. 2023. Vol 56. pp. 2335–2340.
8. Klokov A., Abrarov A., Danilov P. Flotation froth monitoring using unsupervised multiple object tracking methods. Journal of Mineral and Material Science. 2023. Vol. 4, Iss. 1. pp. 1–4.
9. Rudolph M., Kurz S., Rakitsch B. Hybrid modeling design patterns. Journal of Mathematics in Industry. 2024. Vol. 4. 3.
10. Wang Z. Physics-based linear regression for high-dimensional forward uncertainty quantification. Journal of Computational Physics. 2025. DOI: 10.1016/j.jcp.2024.113668
11. Dinç E. Linear regression analysis and its application to the multivariate spectral calibrations for the multiresolution of a ternary mixture of caffeine, paracetamol and metamizol in tablets. Journal of Pharmaceutical and Biomedical Analysis. 2003. No. 33. pp. 605–615.
12. Westad F., Schmidt A., Kermit M. Incorporating chemical band-assignment in near infrared spectroscopy regression models. Journal of Near Infrared Spectroscopy. 2008. Vol. 16, Iss. 3. pp. 265–273.
13. Kosyanov P. M. X-ray analysis of inorganic substances of complex chemical composition. Tyumen : TIU, 2016. 195 p
14. Diaz V. et al. Monitoring of multivariate calibration models in the absence of new reference values: the regression case. Chemometrics and Intelligent Laboratory Systems. 2023. Vol 240, Iss. 10. 104884.