40th International Conference on Production Engineering of Serbia
ICPES 2025
Nis, Serbia, 18-19th september 2025
APPLICATION OF RANDOM FOREST REGRESSION FOR PREDICTION OF SURFACE ROUGHNESS IN FIBER LASER CUTTING
Dragan Rodic, Milenko Sekulic, Borislav Savkovic, Andelko Aleksic, Aleksandra Kosanovic
DOI: 10.46793/ICPES25.091R
Accurate prediction of surface roughness (Ra) in laser cutting is essential for quality control and process optimisation. This study presents a Random Forest Regression (RFR) model for the prediction of Ra in fibre laser cutting of EN 10130 steel sheets. The model was trained using 14 experimental samples obtained from a Box-Behnken design with cutting speed, abrasive concentration and gas pressure as input parameters. Three additional samples were used for testing. The RFR model was implemented in MATLAB with 500 regression trees. On the test set, it achieved an RMSE of 0.204 µm and an R² of 0.55, with an average absolute error of 2.11%. The average error across all 17 samples remained below 10%. These results confirm that Random Forest is a reliable and interpretable method for modelling surface roughness in laser cutting, especially when working with limited and non-linear experimental data
Surface Roughness, Laser Cutting, Random Forest Regression, Machine Learning, Predictive Modeling
[1] A. F. Courtier, M. McDonnell, M. Praeger, J. A. Grant-Jacob, C. Codemard, P. Harrison, et al.: Modelling of fibre laser cutting via deep learning, Optics Express, vol. 29, pp. 36487-36502, 2021.
[2] B. Mills and J. A. Grant-Jacob: Lasers that learn: The interface of laser machining and machine learning, IET Optoelectronics, vol. 15, pp. 207-224, 2021.
[3] N. E. Sizemore, M. L. Nogueira, N. P. Greis, and M. A. Davies: Application of machine learning to the prediction of surface roughness in diamond machining, Procedia Manufacturing, vol. 48, pp. 1029-1040, 2020.
[4] N. Levichev, A. T. García, M. Kardan, D. Cattrysse, and J. R. Duflou: Experimental validation of a machine learning algorithm for roughness quantification in laser cutting, Procedia CIRP, vol. 113, pp. 564-569, 2022.
[5] T. Steege, G. Bernard, P. Darm, T. Kunze, and A. F. Lasagni, "Prediction of surface roughness in functional laser surface texturing utilizing machine learning," in Photonics, 2023, p. 361.
[6] J. D. Kechagias, A. Tsiolikas, M. Petousis, K. Ninikas, N. Vidakis, and L. Tzounis: A robust methodology for optimizing the topology and the learning parameters of an ANN for accurate predictions of laser-cut edges surface roughness, Simulation Modelling Practice and Theory, vol. 114, p. 102414, 2022.
[7] M. Arunadevi, S. Saravanan, G. Mahesh, and S. Chethan: Machine Learning Based Surface Finish Prediction and Optimization of Process Parameters in Pulsed CO2 Laser Cutting of Particle (TiC) Reinforced Al6061 Composite Using KNN & ANN, Journal of The Institution of Engineers (India): Series D, pp. 1-10, 2024.
[8] T. H. Nguyen, C.-K. Lin, P.-C. Tung, C. Nguyen-Van, and J.-R. Ho: An extreme learning machine for predicting kerf waviness and heat affected zone in pulsed laser cutting of thin non-oriented silicon steel, Optics and Lasers in Engineering, vol. 134, p. 106244, 2020.
[9] L. Breiman: Random forests, Machine learning, vol. 45, pp. 5-32, 2001.
[10] D. Rodic, M. Sekulic, B. Savkovic, M. Madic, and M. Trifunovic: Integration of RSM and Machine Learning for Accurate Prediction of Surface Roughness in Laser Processing, Applied Sciences, vol. 15, p. 7064, 2025.