MEK-Mining (St. Petersburg, Russia):
Beriashvili A. T., Production Design Engineer, aleksandr.beriashvili@mekgroup.ru
Pikulina V. M., Production Design Engineer

A study of cupriferous sandstones of the Zhezkazgan ore field is used to demonstrate the drawbacks of applying the existing methodology, based on the calculation of multiple regression equations, for solving the variability problem for respective process indicators. For the purpose of developing a financial and performance model of the Zhezkazgan concentrator, the task was set to assess the variability for the critical areas of the Zhezkazgan ore field. All ores, except for one site, are sulfide type ores. The content of iron oxides and hydroxides in the ores is almost the same for all samples studied and is generally ignored by the specialists. However, iron hydroxides contained in the ores, especially the gothite types of iron oxides, are involved in the formation of [Fe(OH)]⁺ type compounds that bind the xanthate introduced into the slurry with the formation of the [Fe(OH)X]X^– complex and shift the equilibrium towards desorption of the collector from the surface of sulfide minerals. This results in significant variations in respective process indicators for copper and even in a copper concentrate grade reduction to 28% in monthly terms. This may be explained both by a certain imperfection of the processing technology and by the specific material composition fluctuations in the ore, in particular by higher sericitization factors. When calculating the correlation matrix, it is shown that there is no formal relationship between the copper content in the ore and the respective process indicators for copper, which naturally complicates the use of regression equations for solving the variability problem. In order to obtain a more adequate object model, a generalized regression neural network was designed. The adequacy assessment results obtained are not completely satisfactory, however, the neural network model designed enables a generalized assessment of the functional relationship between copper and silver recovery values and the content of these metals in the ore.

1. Reznikov A. P. Processing the accumulated information in difficult conditions. Moscow: Nauka, 1976. 236 p.
2. Eliseeva I. I., Yuzbashev M. M. General theory of statistics: a textbook. Edited by I. I. Eliseeva. 4^th ed., revised. and enlarged Moscow: Finansy i Statistika, 2002. 480 p.
3. Mashevsky G. N., Petrov A. V., Romanenko S. A., Sufyanov F. S., Balmanova A. Zh. A new approach to regulating the process of sulfide minerals selective flotation separation from pyrite in lime medium. Obogashchenie Rud. 2012. No. 1. pp. 12–16.
4. Neural networks. STATISTICA Neural Networks: Methodology and technologies of modern data analysis. Edited by V. P. Borovikov. 2nd ed., revised and enlarged. Moscow: Telekom, 2008. 392 p.
5. Romanenko S. A. Effectiveness of multi-sensor ionometry systems and neural network modeling methods application in flotation processes laboratory studies. Obogashchenie Rud. 2013. No. 1. pp. 18–22.
6. Mashevsky G. N., Petrov A. V., Romanenko S. A., Sufyanov F. S. Development of ore types processing classification principles on the basis of flotation process parameters control and neural network modeling. Obogashchenie Rud. 2012. No. 4. pp. 36–42.
7. Arustamyan A. M., Sanakulov K. S. Harmonic analysis of random functions in flotation studies. Gornyi Zhurnal. 2017. No. 10. pp. 33–40. DOI: 10.17580/gzh.2017.10.07.
8. Napier-Munn T. J. Statistical methods for mineral engineers — How to design experiments and analyse data. Queensland, Australia: Julius Kruttschnitt Mineral Research Centre, 2014. 627 p.
9. Nakhaei F., Mosavi M. R., Sam A., Vaghei Y. Recovery and grade accurate prediction of pilot plant flotation column concentrate: Neural network and statistical techniques. International Journal of Mineral Processing. 2012. Vol. 110–111. pp. 140–154.
10. Li X., McKee D. J., Horbery T., Powell M. S. The control room operator: The forgotten element in mineral process control. Miner. Eng. 2011. 24, No. 8. pp. 894–902.
11. Shean B. J., Cilliers J. J. A review of froth flotation control. International Journal of Mineral Processing. 2011. Vol. 100, Iss. 3–4. pp. 57–71.
12. Arustamyan K. M., Romanenko S. A., Arustamyan A. M. Using potentiograms in assessment of process properties of copper ore in terms of Zhezkazgan deposit. Gornyi Zhurnal. 2016. No. 11. pp. 65–70. DOI: 10.17580/gzh.2016.11.12.
13. Mashevsky G. N., Kokorin A. M. Development of the scientific basis for improving the technology and creating an algorithmic base for computerized control of ore flotation at the Erdenetin Ovoo deposit in order to improve the profitability of production. Collection of reports of NTS of the JV «Erdenet». Erdenet Concern, 1998. pp. 87–101.
14. Mashevsky G. N., Petrov А. V., Moilainen Ya., Timpere Yu., Saloheimo K., Kemppinen H. Principles of flotation process computer control on the basis of FrothMaster video system. Obogashchenie Rud. 2010. No. 3. pp. 39–43.