Bauman Moscow State Technical University (Moscow, Russia)

A. G. Zinyagin, Cand. Eng., Associate Prof., e-mail: ziniagin_ag@bmstu.ru
D. A. Brayko, Postgraduate Student, e-mail: brayko@bmstu.ru
A. V. Muntin, Cand. Eng., Associate Prof., e-mail: muntin_av@bmstu.ru
P. Yu. Zhikharev, Senior Lecturer, e-mail: zhikharev@bmstu.ru

This paper presents an approach to predicting Continuous Cooling Transformation (CCT) diagrams of low-carbon steels using a mathematical model based on regression and classification. The method of digitizing CCT diagrams and its application taken from atlases and current scientific articles are given. The digitization method is based on reading the color value from the CCT diagram image and then converting the coordinates of the color position in accordance with the scale of the CCT diagram axes. It was assumed that CCT diagram consists of zones defined by the beginning and end of ferrite transformation, the beginning of pearlite transformation, the end of bainite transformation, and the beginning and end of martensite transformation. When developing a predictive mathematical model, an optimization algorithm was used to find a model with the best hyperparameters among classical machine learning models (k-Nearest Neighbors, Support Vector Machine, Linear/Logistic Regression) and based on decision trees (LightGBM, CatBoost). The model solves regression (temperature prediction) and classification (binary mask prediction) problems. The superimposition of a binary mask on the temperature vector made it possible to constrain the resulting phase transformation curve along the time axis. To build test CCT diagrams, a number of dilatometric studies of four steel grades were carried out. The new predictive approach made it possible to achieve satisfactory values of metrics on test CCT diagrams. The average absolute error did not exceed 20°C; the coefficient of determination was in the range of 0.55–0.86, but for the martensite transformation it took negative values, which can be explained by the initial approximation of the transformation by a polygonal chain; ROC AUC metric was at least 0.80.

The research was conducted within the framework of the Russian Federation’s strategic academic leadership program «Priority-2030,» aimed at supporting the development programs of higher education institutions. The scientific project PRIOR/SN/NU/22/SP5/26, titled «Creating Innovative Digital Tools for Applying Applied Artificial Intelligence and Advanced Statistical Analysis of Big Data in Technological Processes of Metallurgical Production», was also part of this initiative.

1. Efron L. I. Metallovedenie v bolshoi metallurgii. Trubnye stali. Moscow: Metallurgizdat. 2012. 694 p.
2. Ringinen D. A. Formation of a homogeneous structure during thermomechanical treatment under conditions of mill 5000 and stability of impact toughness and cold resistance of pipe steels of strength classes X80 and X100. Dissertation … of Candidate of Technical Sciences: 05.16.01. Moscow, 2016. 141 p.
3. Zinyagin A. G., Muntin A. V., Tynchenko V. S., Zhikharev P. I., Borisenko N. R., Malashin I. Recurrent Neural Network (RNN) – Based Approach to Predict Mean Flow Stress in Industrial Rolling. Metals. 2024. Vol. 14 (12). 1329.
4. Zhikharev P. Y., Muntin A. V., Brayko D. A., Kryuchkova M. O. Artificial Intelligence and Machine Learning In Metallurgy. Part 2. Application Examples. Metallurgist. 2024. Vol. 67. pp. 1545–1560.
5. Wróbel J., Kulawik A. Algorithm for determining time series of phase transformations in the solid state using long-short-term memory neural network. Materials. 2022. Vol. 15. No. 11. 3792.
6. Lindemann B. et al. A survey on long short-term memory networks for time series prediction. Procedia CIRP. 2021. Vol. 99. pp. 650–655.
7. Hernández-Flores L. et al. Determination of TTT Diagrams of Ni–Al Binary Using Neural Networks. Materials. 2022. Vol. 15 (24). 8767.
8. Luukkonen J. et al. Gradient Boosted Regression Trees for Modelling Onset of Austenite Decomposition During Cooling of Steels. Metallurgical and Materials Transactions B. 2023. Vol. 54. No. 4. pp. 1705–1724.
9. Xiaoxiao G., Wang H., Xue W., Song X., Huang H., Li M., Guang M. Modeling of CCT diagrams for tool steels using different machine learning techniques. Computational Materials Science. 2020. Vol. 171. 109235.
10. Ustimenko A., Beliakov A., Prokhorenkova L. Gradient boosting performs gaussian process inference. arXiv. 2023. arXiv: 2206.05608.
11. Shi Y., Ke G., Chen Z., Zheng S., Liu. T.-Y. Quantized Training of Gradient Boosting Decision Trees. Advances in Neural Information Processing Systems. 2022. pp. 18822–18833.
12. Lundberg S. M., Lee S. I. A unified approach to interpreting model predictions. CoRR. 2017. 1705.07874.
13. Lundh F. Pillow. Pypi.org. 2024. Available at: https://pypi.org/project/pillow/
14. Zhang Y. et al. Phase Transformation Temperature Prediction in Steels via Machine Learning. Materials. 2024. Vol. 27. No. 5. 1117.
15. Teplukhina I. V., Golod V. M., Tsvetkov A. S. CCT diagram plotting based on the numerical analysis of dilatometric tests results. Letters on Materials. 2018. Vol. 8. No. 1. pp. 37–41.
16. Bräutigam–Matus K. et al. Experimental determination of continuous cooling transformation (CCT) diagrams for dual-phase steels from the intercritical temperature range. Metals. 2018. Vol. 8. No. 9. p. 674.
17. Grajcar A., Morawiec M., Zalecki W. Austenite decomposition and precipitation behavior of plastically deformed low-Si microalloyed steel. Metals. 2018. Vol. 8. No. 12. 1028.
18. Schindler I. et al. Effects of austenitization temperature and predeformation on CCT diagrams of 23MnNiCrMo5-3 steel. Materials. 2020. No. 13 (22). 5116.
19. Zurutuza I. et al. Effect of Quenching Strategy and Nb-Mo Additions on Phase Transformations and Quenchability of High-Strength Boron Steels. JOM. 2021. Vol. 73. pp. 3158–3168.
20. Krbata M. et al. Austenite decomposition of a lean medium Mn steel suitable for quenching and partitioning process: comparison of CCT and DCCT diagram and their microstructural changes. Materials. 2022. Vol. 15. No. 5. 1753.
21. Wang Z. et al. Effect of tungsten addition on continuous cooling transformation and precipitation behavior of a high titanium microalloyed steel. Metals. 2022. Vol. 12 (10). 1649.
22. Baak M., Koopman R., Snoek H., Klous S. A new correlation coefficient between categorical, ordinal and interval variables with Pearson characteristics. Computational Statistics & Data Analysis. 2020. Vol. 152. 107043.