AI-Driven Prediction and Optimization of Surface Roughness and Residual Stresses in Turning of 42CrMo4 + QT Steel

Shreyas Chakravarthy; Shyam Kaushik B M; Soma Partha Sai; Prashant Doranahalli; Sujan R; Gajanan M Naik

doi:10.70454/IJMRE.2026.60103

Authors

Shreyas Chakravarthy Dept. of Computer Science & Engineering, RVITM, Bengaluru, India Author
Shyam Kaushik B M Dept. of Computer Science & Engineering, RVITM, Bengaluru, India Author
Soma Partha Sai Dept. of Computer Science & Engineering, RVITM, Bengaluru, India Author
Prashant Doranahalli Dept. of Computer Science & Engineering, RVITM, Bengaluru, India Author
Sujan R Dept. of Computer Science & Engineering, RVITM, Bengaluru, India Author
Gajanan M Naik Dept. of Mechanical Engineering, RVITM, Bengaluru, India Author

DOI:

https://doi.org/10.70454/IJMRE.2026.60103

Keywords:

Surface Roughness Prediction, Residual Stress, Random Forest Regression, Gaussian Process Regression, 42CrMo4 + QT Steel, Multi-objective Optimization, Machining Analytics, Digital Twin

Abstract

Accurately predicting surface roughness and residual stress is important for keeping machined parts the right size and making sure they hold up well over time. This study introduces a hybrid approach that combines Random Forest Regression and Gaussian Process Regression to predict and improve outcomes when turning 42CrMo4 + QT steel. The MaRoReS dataset, which is publicly available, includes machining parameters along with vibration and cutting force data.

The RFR model is used to find the main process parameters and measure how they affect surface integrity. At the same time, the GPR model offers predictions that take uncertainty into account for optimizing multiple objectives. We trained and validated the model by using grid search along with 10-fold cross-validation. The hybrid approach they tried managed to get an R² value above zero. The proposed hybrid approach achieved R² > 0.98 and RMSE < 0.12 μm for surface roughness prediction, outperforming traditional regression and neural models reported in literature.

A Pareto-based optimization strategy identified the optimal parameter window (Vc = 160 m/min, f = 0.15 mm/rev, ap = 0.6 mm), resulting in a 28% improvement in surface finish and a 35% reduction in tensile residual stresses. The results highlight that the proposed RF–GPR framework offers a practical and transparent approach to data-driven process optimization. It also shows strong potential for integration into digital twin–based machining environments.

References

] D. Díaz-Salamanca, D. I. L. Pacheco, and F. J. Trujillo, “Turning of 42CrMo4 + QT under different scenarios: Dataset of machining, roughness and residual stress,” Data in Brief, vol. 54, p. 111445, 2024.

[2] D. Díaz-Salamanca, “Influence of turning parameters on residual stresses and roughness of 42CrMo4 + QT steel,” Int. J. Adv. Manuf. Technol., vol. 130, no. 7–8, pp. 3217–3231, 2024.

[3] N. E. Sizemore, R. Venkatesh, and M. S. Kumar, “Application of machine learning to the prediction of surface roughness,” Procedia Manuf., vol. 48, pp. 890–897, 2020.

[4] B. A. Beatrice and K. Ramesh, “Surface roughness prediction using artificial neural networks,” Procedia Eng., vol. 97, pp. 205–214, 2014.

[5] J. Chen, Z. Wang, and Y. Li, “Predicting surface roughness in turning complex workpieces using Gaussian process regression,” Sensors, vol. 24, no. 2, pp. 812–826, 2024.

[6] N. Masmiati and M. H. El-Axir, “Optimization of cutting conditions for minimum residual stress using response surface methodology,” Int. J. Mach. Tools Manuf., vol. 106, pp. 36–45, 2016.

[7] M. H. El-Axir, “Modeling and parameter optimization for surface roughness in machining processes,” Eng. Technol. Appl. Sci. Res., vol. 7, no. 1, pp. 1426–1431, 2017.

[8] T. Reeber, P. Walker, and C. J. Davis, “A data-driven approach for cutting force prediction in metal cutting,” Manufacturing, vol. 4, pp. 112–126, 2024.

[9] Z. Wang, Y. Zhou, and S. Zhang, “Prediction of surface residual stress in multi-axis milling using radial basis function neural networks,” Materials, vol. 15, no. 4, pp. 2101–2114, 2022.

[10] S. H. Wu, J. Chen, and H. Xu, “Experimental and machine learning-driven prediction of residual stresses in manufacturing processes,” Materials, vol. 17, no. 8, pp. 1205–1218, 2024.

[11] R. Zhou, Y. Zhao, and T. Yang, “Modeling and simulation of residual stress in metal cutting: A computational and machine learning approach,” Int. J. Mech. Sci., vol. 250, p. 108092, 2024.

[12] Y. Zhou, M. Chen, and L. Wang, “Machine learning-based prediction of residual stresses in thermal and laser-assisted processes,” Appl. Surf. Sci., vol. 632, p. 157934, 2024.

[13] B. Peng, L. Zhong, and J. Liu, “Physics-informed neural networks for thermal simulation in manufacturing,” Comput. Methods Appl. Mech. Eng., vol. 421, pp. 116754–116768, 2024.

[14] Z. Yao, X. Zhang, and P. Sun, “Position-dependent process monitoring and surface roughness prediction in milling using machine learning,” CIRP J. Manuf. Sci. Technol., vol. 41, pp. 70–82, 2023.

[15] A. Balasuadhakar, S. Raj, and K. Suresh, “Machine learning prediction of surface roughness in turning: Recent advances,” J. Manuf. Process., vol. 96, pp. 312–329, 2025.

[16] R. P. Shetty, P. Joshi, and A. Rao, “Surface roughness modeling and prediction in high-speed turning using hybrid learning models,” Sci. Rep., vol. 15, p. 19284, 2025.

[17] A. Tzotzis, D. Papazoglou, and C. Alexopoulos, “A dynamic surface roughness prediction system based on machine learning,” J. Manuf. Syst., vol. 74, pp. 451–465, 2025.

[18] M. M. Hasan, “Machining-induced residual stress modeling using physics-based and machine learning methods,” Ph.D. dissertation, Univ. of Kentucky, Lexington, KY, USA, 2023.

[19] F. Farias, R. Almeida, and D. Martins, “FEM and artificial neural network surrogate modeling for thermal prediction in metal cutting,” Addit. Manuf., vol. 48, p. 102387, 2021.

[20] D. Díaz-Salamanca, “MaRoReS: Machining, Roughness and Residual Stresses generated in turning of 42CrMo4 + QT steel,” Mendeley Data, V1, 2024.