ArticleName |
The practice of applying binary logistic regression to reduce the number of negative technological events in rolling production |
Abstract |
The prospects for the next decade, both in the metallurgical industry as a whole and in rolling production, are a high level of competition in the face of falling demand for ordinary steel grades and rising energy prices. In turn, this prospect makes urgent the high need to improve the operating efficiency of rolling mills. The decisive advantage in such conditions will be the high rate of improvement in operational efficiency. Therefore, it is important to implement the most effective solutions in rolling production as quickly as possible. One of the sources for improving the operating efficiency of rolling mills is to reduce the number of negative technological events. And the best tool in the study of negative technological events is binary logistic regression, which allows you to solve classification problems. The key advantage over other methods for solving the classification problem (classification trees, random forest or gradient boosting) is the interpretability of the results obtained - the resulting mathematical formula can be easily checked for compliance with the physical meaning. The article presents a brief description of the binary logistic regression and shows how the assessment of the adequacy of the obtained mathematical model differs from the usual regression. It is shown that it is important to obtain not only an adequate and reliable mathematical model of the process, but also to correctly determine the decision threshold. Approaches to data preparation, analysis and use of results are demonstrated in order to reduce the number of negative technological events in rolling production. Practical results of using binary logistic regression to improve processes in the conditions of NLMK’s rolling facilities are reported. |
References |
1. Belsky S. M., Shopin I. I. Mathematical model of the probability of strip breakage during cold rolling. Chernye metally. 2020. No. 3. pp. 18−23. 2. Belskiy S. M., Shopin I. I. Statistical regression analysis of breakages in cold rolling. IOP Conference Series: Materials Science and Engineering. 2020. Vol. 969. 012074. 3. Belskiy S. M., Shopin I. I. On issue of strip’s break during cold rolling. IOP Conference Series: Materials Science and Engineering. 2020. Vol. 969. 012072.
4. Belskiy S. M., Shopin I. I., Safronov A. A. Improving efficiency of rolling production by predicting negative technological events. Defect and Diffusion Forum. 2021. Vol. 410. pp. 96−101. 5. Belsky S. M., Shopin I. I. Application of the saddle coefficient for assessing the quality of hotrolled strip. Chernye metally. 2019. No. 9. pp. 9−13. 6. Ibraev I. K. Quality of low-carbon thin-sheet rolled steel. Monograph. Moscow, 2021. 291 p. 7. Lim Y., Venugopal R., Ulsoy A. G. Process control for sheet-metal stamping process modeling, controller design and stop-floor implementation. Springer, 2014. 140 p. 8. Klocke F. Manufacturing processes 4. Forming. Springer, 2013. 516 p. 9. Rees D. Basic engineering plasticity. An introduction to engineering and manufacturing applications. Butterworth-Heinemann, 2006. 528 p. 10. Wilko C. E. Formability. A review of parameters and processes that control, limit or enhance the formability of sheet metal. Springer, 2011. 112 p. 11. Mazur V. L., Nogovitsyn A. V. Theory and technology of thin sheet rolling (numerical analysis and technical applications). Dnepropetrovsk : RVA “Dnipro-VAL”, 2010. 500 p. 12. Mazur V. L. Causes and methods of preventing bends (fractures) of thin-sheet steel. Stal. 2015. No. 12. pp. 21−28. 13. Belsky S. M., Shchedrin I. N., Shopin I. I. Solution to the problem of loss of stability of rolls of steel strip with a polymer coating. Proizvodstvo prokata. 2018. No. 7. pp. 5−8. 14. Shopin I. I., Belsky S. M. Simplified model of the stress-strain state of a coil on a winder. Proizvodstvo prokata. 2016. No. 5. pp. 13−17. 15. Shopin I. I., Belsky S. M. Layered model of the stress-strain state of a coil on a winder. Proizvodstvo prokata. 2016. No. 8. pp. 3−7. 16. Antonov V. I. Method for determining the initial stresses in a roll with a nonlinear relationship between stresses and strains. Vestnik MGSU. 2010. No. 4. pp. 177−180. 17. Ashikhmin G. V., Iroshnikov S. A. Formation of the roll generatrix profile during strip winding. Proizvodstvo prokata. 2002. No. 9. pp. 14−17. 18. Ashikhmin G. V., Iroshnikov S. A. The relationship between the distribution of longitudinal stresses across the width of the strip in a roll and the profile of its generatrix. Proizvodstvo prokata. 2002. No. 10. pp. 16−22. 19. Bozhkov A. I. Scientific explanation and development of quality control of sheet-rolling shop products at enterprises of iron and steel industry. Report 1. Integrated automated system for quality control of products at steel works. Proizvodstvo prokata. 2014. No. 7. pp. 27-33. 20. Bozhkov A. I., Yeryomin G. N., Ivliyev S. N., Bakhtin S. V. et al. Scientific explanation and development of quality control automation systems of plate-rolling shop products at ironworks. Report 3. Control subsystem of properties of rolled sheets. Functional structure and mathematical support of subsystem. Proizvodstvo prokata. 2015. No. 3. pp. 39-48. 21. Bel’skii S. M., Mazur S. I., Mukhin Y. A., Goncharov A. I. Influence of the cross section of hot-rolled steel on the flatness of cold-rolled strip. Steel in Translation. 2013. Vol. 43. No. 5. pp. 313−316. 22. Fukushima S., Washikita Y., Sasaki T., Nakagawa S., Buei Y., Yakita Y., Yanagimoto J. Mixed scheduled rolling of high tensile strength and mild steel using a high-accuracy profile model in hot strip finishing mill. Tetsu-to-Hagane. 2014. Vol. 10. No. 12. pp. 1499−1507. 23. Maksimov E. A., Ostsemin A. A. Study of the mechanism of flatness violation during hot sheet rolling. Stal. 2015. No. 6. pp. 27−30. 24. Reinschke J. U., Schmid F., Lamp H., Miele M. Novel profile and flatness control system for hot strip mills. Stahl und Eisen. 2004. Vol. 124. No. 2. pp. 53−58. 25. Shinkin V. N. Elastoplastic flexure of round steel beams. 1. Springback coefficient. Steel in Translation. 2018. Vol. 48. No. 3. pp. 149−153. 26. Hingole R. S. Advances in metal forming. Expert system for metal forming. Springer, 2015. 116 p. 27. Chakrabarty J. Applied plasticity. Springer, 2010. 758 p. 28. Shinkin V. N. Elastoplastic flexure of round steel beams. 2. Residual stress. Steel in Translation. 2018. Vol. 48. No. 11. pp. 718−723. 29. Kobzar A. I. Applied mathematical statistics. Moscow, 2006. 816 p. 30. Hastie T., Tibshirani R., Friedman J. The elements of statistical learning. Data mining, inference and prediction. Springer, 2017. 745 p. 31. GOST R ISO 13053-1-2015 Statistical methods. Quantitative Methods for Process Improvement Six Sigma. Part 1. DMAIC Methodology. Introduced: 01.07.2016. 32. GOST R ISO 13053-2-2015 Statistical methods. Quantitative Methods for Process Improvement Six Sigma. Part 2. Methods. Introduced: 01.07.2016. |