KEY INDICATORS OF STEEL QUALITY OF CAST PRODUCTS FOR RAILWAY TRANSPORT

The audit of the technological process of 20GFL steel melting in arc furnaces with the capacity of six tons in Rubtsovsk branch office of JSC “Altaivagon” has revealed some problems connected with defects of mechanical properties of steel. The chemical composition of the researched samples is within the established State Standards. To confirm this fact and to reveal the cause of the defect problem the authors of the article have conducted the statistical analysis of the influence of chemical composition on the quality characteristics of mechanical properties of metal. To conduct the regression analysis and to build the model of forecasting the following algorithms there have been used: Linear Regression (LR), Random Forest (RF) and Support Vector Machine (SVM). The solution of the problem connected with the steel defects by mechanical properties has been offered with the help of forecasting models. As the result it has been established that there are some possibilities to forecast and to control mechanical properties of steel along its melting. When receiving the forecast of defects by one of the parameters, assuming the current values of the content of chemical composition and the final conditions, one can calculate the chemical composition of steel, which can be reached on the considering melting and which will assure the absence of defects. It has been established that for the analysis of tendencies of steel quality of cast products one can use two key indicators out of two groups of mechanical properties – yield point stress σт and impact hardness КСV. Schematic control of the quality can be fulfilled by timing diagram of mean removal chart of the controlled parameters by the defect from its border value in standard defections. The indicator of good quality and the risk minimization is the output of the timebase diagram from the two standard defections.

steel, forecasting, mechanical properties, chemical composition, modeling, melting, control, castings

1. Knyazev S.V., Skopich D.V., Usol’tsev A.A., Fat’yanova E.A. Forecasting of quality characteristics of 20GFL steel. Vestnik Sibirskogo gosudarstvennogo industrial’nogo universiteta. 2015, no. 2 (12), pp. 31–33. (In Russ.).

2. Knyazev S.V., Usoltsev A.A., Skopich D.V., Fatyanova E.A., Dolgopolov A.E. Automated system of control and diagnostics of caststeel defects in the mass production. IOP Conference Series: Materials Science and Engineering. 2016, vol. 150, рр. 1–5 (012039).

3. Cheprasov A.I., Knyazev S.V., Usoltsev A.A., Dolgopolov A.E., Mamedov R.O. Detection of cold cracks in the cast-steels by the methods of ultrasonic and eddy-current infrared thermography. IOP Conference Series: Materials Science and Engineering. 2016, vol.  150, рр. 1–5 (012026).

4. Hastie T., Tibshirani R., Friedman J. The elements of statistical learning: Data mining, inference and prediction. 2nd ed. SpringerVerlag, 2009.

5. Ho Tin Kam. Random decision forests. Proceedings of the 3rd International Conference on Document Analysis and Recognition, Montreal, QC, 14–16 August 1995, pp. 278–282.

6. Prinzie A., Van den Poel D. Random multiclass classification: generalizing random forests to random MNL and random NB. In: Lecture Notes in Computer Science. Dexa, 2007, рр. 349–358.

7. Statnikov A., Aliferis C.F., Hardin D.P. A gentle introduction to support vector machines in biomedicine: Theory and methods. World Scientific, 2011.

8. Vapnik V.N. Vosstanovlenie zavisimostei po empiricheskim dannym [Renewal of dependences according to empiric data]. Moscow: Nauka, 1979, 448 p. (In Russ.).

9. Cristianini N., Shawe-Taylor J. An introduction to support vector machines and other kernel-based learning methods. Cambridge University Press, 2000.

10. Antipenko V.I., Knyazev S.V. Diagnostics of steel castings production with the aid of technological pilot samples. Soviet Castings Technology (English Translation of Liteinoe Proizvodstvo), 1987, no. 7, p. 34.

11. Il’inskii V.A., Kostyleva L.V., Grebnev Yu.V. Optimization of composition of cast medium-carbon steel. Stal’. 1985, no. 1, pp. 24–26. (In Russ.).

12. Kostyleva L.V., Grebnev Yu.V., Il’inskii V.A. Dendritic segregation in castings of carbon steel. Liteinoe proizvodstvo. 2000, no. 4, pp.  13–15. (In Russ.).

13. Bolyukh V.A., Shinskii I.O. Increase of mechanical properties of large steel castings by their reinforcement at consumable patterns casting. Protsessy lit’ya. 2010, no. 2, pp. 57–62. (In Russ.).

14. Kul’bovskii I.K., Soldatov V.G., Manuev M.S. Improvement of mechanical properties of low-alloyed steel for chargeable castings of railway transport. Zagotovitel’nye proizvodstva v mashinostroenii. 2006, no. 4, pp. 3–6. (In Russ.).

15. Tylkin M.A. Spravochnik termista remontnoi sluzhby [Reference book of heat treatment operator of repair service]. Moscow: Metallurgiya, 1981, 63 p. (In Russ.).

16. Sorokin V.G., Volosnikova A.V., Vyatkin S.A. etc. Marochnik stalei i splavov [Grade guide of steels and alloys]. Moscow: Mashinostroenie, 1989, 640 p. (In Russ.).

17. Golyandina N.E. Metod “Gusenitsa’’-SSA: analiz vremennykh ryadov: uchebnoe posobie [Method “Caterpillar’’-SSA: analysis of timing series: Manual]. St. Petersburg: VVM, 2004, 76 p. (In Russ.).

18. Aleksandrov F.I., Golyandina N.E. Automation of formation of trend and periodical components of timing series in the frameworks of “Caterpillar” – SSA method. ExponentaPro. Matematika v prilozheniyakh. 2004, no. 3–4, pp. 54–61. (In Russ.).

19. Golyandina N. Osipov E. The ‘Caterpillar’-SSA method for analysis of time series with missing values. J. Stat. Plan. Inference. 2007, no. 137 (8), рр. 642– 653.

20. Golyandina N. Usevich K. (2010): 2D-extension of Singular Spectrum Analysis: algorithm and elements of theory. In: Matrix Methods: Theory, Algorithms and Applications. Olshevsky V., Tyrtyshnikov E. eds. World Scientific Publishing, рр. 449–473.