Prediction of Cutting Forces Generated During Turning Operations of C45E Material Using Machine Learning Models: Performance Comparison of DNN and XGBoost Models

Abstract

It is well known that significant developments have occurred in the fields of artificial intelligence and machine learning alongside advancements in computer and software technology. The cutting force prediction capabilities of deep learning (DNN) and XGBoost machine learning models were compared in this study using a dataset prepared by Vladimir Mitrović and associates. This dataset contains the cutting forces generated during cylindrical turning operations. The DNN and XGBoost models exhibited highly comparable prediction capabilities when the tangential cutting force (Ft) was considered, with R² values of 99.71% and 99.89%, respectively. Similarly, the DNN model demonstrated a prediction capability of R² = 98.65% in axial force (Fa) predictions, while the XGBoost model achieved R² = 99.60%. The XGBoost model could make 5.94% more accurate predictions than the DNN model when radial force (Fr) was examined. In general, the XGBoost model outperformed the DNN model. Additionally, it was determined that the XGBoost model could perform calculations more quickly than the DNN model due to the absence of an iteration count.

Keywords

• Cutting tool
• Cutting force
• Deep learning
• XGBoost
• Machine learning

C45E malzemenin Tornalama İşleminde Oluşan Kesme Kuvvetlerinin Makine Öğrenmesi Modelleriyle Tahmini: DNN ve XGBoost Modellerinin Performans Karşılaştırması

Öz

Günümüzde bilgisayar ve yazılım alanındaki gelişmelerle birlikte yapay zeka ve makine öğrenmesi alanında önemli gelişmeler olduğu bilinmektedir. Bu çalışmada, Vladimir Mitrović ve arkadaşlarının oluşturduğu, silindirik tornalama işlemi sırasında oluşan kesme kuvvetlerini barındıran veri seti kullanılarak derin öğrenme (DNN) ve XGBoost makine öğrenmesi modellerinin kesme kuvveti tahmin yetenekleri karşılaştırılmıştır. Teğetsel kesme kuvveti (Ft) ele alındığında DNN ve XGBoost modelin sırasıyla %99,71 ve %99,89 R2 değeriyle birbirlerine oldukça yakın tahmin yetenekleri olduğu görülmüştür. Benzer şekilde eksenel kuvvet (Fa) tahminlerinde ise DNN modelin R2 = %98,65, XGBoost modelin R2 = %99,60 oranında tahmin yeteneğine sahip olduğu tespit edilmiştir. Radyal kuvvet (Fr) incelendiğinde XGBoost modelin, DNN modele göre %5,94 oranında daha iyi tahmin yapabildiği sonucuna varılmıştır. Genel olarak XGBoost modelin, DNN modele göre daha performanslı olduğu belirlenmiştir. Ayrıca XGBoost modelde iterasyon sayısı olmadığından DNN modele göre hızlı hesap yapabildiği görülmüştür.

Anahtar Kelimeler

• Kesici takım
• Kesme kuvveti
• Derin öğrenme
• XGBoost
• Makine öğrenmesi

Kaynakça

1. [1] M. Soori, B. Arezoo, and R. Dastres, “Machine learning and artificial intelligence in CNC machine tools, A review,” Sustainable Manufacturing and Service Economics, vol. 2, no. October 2022, p. 100009, 2023, doi: 10.1016/j.smse.2023.100009.
2. [2] M. Q. Tran, H. P. Doan, V. Q. Vu, and L. T. Vu, “Machine learning and IoT-based approach for tool condition monitoring: A review and future prospects,” Measurement (Lond), vol. 207, no. December 2022, p. 112351, 2023, doi: 10.1016/j.measurement.2022.112351.
3. [3] A. Przybyś-Małaczek, I. Antoniuk, K. Szymanowski, M. Kruk, and J. Kurek, “Application of Machine Learning Algorithms for Tool Condition Monitoring in Milling Chipboard Process,” Sensors, vol. 23, no. 13, 2023, doi: 10.3390/s23135850.
4. [4] H. Mostaghimi, S. S. Park, D. Y. Lee, S. Nam, and E. Nam, “Prediction of Tool Tip Dynamics Through Machine Learning and Inverse Receptance Coupling,” International Journal of Precision Engineering and Manufacturing, vol. 24, no. 10, pp. 1739–1752, 2023, doi: 10.1007/s12541-023-00831-6.
5. [5] A. Schueller and C. Saldaña, “Indirect Tool Condition Monitoring Using Ensemble Machine Learning Techniques,” J Manuf Sci Eng, vol. 145, no. 1, pp. 1–10, 2023, doi: 10.1115/1.4055822.
6. [6] M. E. Korkmaz et al., “Prediction and classification of tool wear and its state in sustainable machining of Bohler steel with different machine learning models,” Measurement (Lond), vol. 223, no. November 2023, doi: 10.1016/j.measurement.2023.113825.
7. [7] M. E. Korkmaz, M. K. Gupta, E. Çelik, N. S. Ross, and M. Günay, “Tool wear and its mechanism in turning aluminum alloys with image processing and machine learning methods,” Tribol Int, vol. 191, no. November 2023, p. 109207, 2024, doi: 10.1016/j.triboint.2023.109207.
8. [8] T. T. Truong et al., “Data-driven prediction of tool wear using Bayesian regularized artificial neural networks,” Measurement, vol. 238, no. February, p. 115303, 2024, doi: 10.1016/j.measurement.2024.115303.

9. [9] X. Zhang, B. Shi, B. Feng, L. Liu, and Z. Gao, “A hybrid method for cutting tool RUL prediction based on CNN and multistage Wiener process using small sample data,” Measurement (Lond), vol. 213, no. October 2022, p. 112739, 2023, doi: 10.1016/j.measurement.2023.112739.
10. [10] X. Zhou et al., “Tool wear classification based on convolutional neural network and time series images during high precision turning of copper,” Wear, vol. 522, no. December 2022, p. 204692, 2023, doi: 10.1016/j.wear.2023.204692.
11. [11] I. Asiltürk, M. Kuntoğlu, R. Binali, H. Akkuş, and E. Salur, “A Comprehensive Analysis of Surface Roughness, Vibration” Metals, vol. 13, p. 437, 2023, doi: 10.3390/met13020437.
12. [12] N. P. Greis, M. L. Nogueira, S. Bhattacharya, C. Spooner, and T. Schmitz, “Stability modeling for chatter avoidance in self-aware machining: an application of physics-guided machine learning,” J Intell Manuf, vol. 34, no. 1, pp. 387–413, 2023, doi: 10.1007/s10845-022-01999-w.
13. [13] K. Biruk-Urban, P. Bere, and J. Józwik, “Machine Learning Models in Drilling of Different Types of Glass-Fiber-Reinforced Polymer Composites,” Polymers (Basel), vol. 15, no. 23, 2023, doi: 10.3390/polym15234609.
14. [14] E. Ekici, İ. Pazarkaya, and G. Uzun, “Delamination and thrust force analysis in GLARE: Influence of tool geometry and prediction with machine learning models,” J Compos Mater, vol. 59, no. 10, pp. 1233–1249, 2025, doi: 10.1177/00219983241305706.
15. [15] R. Sekhar, T. P. Singh, and P. Shah, “Machine learning based predictive modeling and control of surface roughness generation while machining micro boron carbide and carbon nanotube particle reinforced Al-Mg matrix composites,” Particulate Science and Technology, vol. 40, no. 3, pp. 355–372, 2022, doi: 10.1080/02726351.2021.1933282.
16. [16] R. Kumar et al., “Exploring the intricacies of machine learning-based optimization of electric discharge machining on squeeze cast TiB2/AA6061 composites: Insights from morphological, and microstructural aspects in the surface structure analysis of recast layer formation and worn-out analysis,” Journal of Materials Research and Technology, vol. 26, pp. 8569–8603, 2023, doi: 10.1016/j.jmrt.2023.09.127.
17. [17] İ. Pazarkaya, “Prediction of cutting tool insert parameters using machine learning and its application to industry: Performance of DNN and LSTM models,” Expert Syst Appl, vol. 279, no. April, pp. 0–2, 2025, doi: 10.1016/j.eswa.2025.127471.
18. [18] B. Li, T. Liu, J. Liao, C. Feng, L. Yao, and J. Zhang, “Non-invasive milling force monitoring through spindle vibration with LSTM and DNN in CNC machine tools,” Measurement (Lond), vol. 210, Mar. 2023, doi: 10.1016/j.measurement.2023.112554.
19. [19] Z. Li and C. Ni, “Prediction of Cutting Force for Different Tools Based on Transfer Learning and Neural Networks,” International Journal of Precision Engineering and Manufacturing, vol. 26, no. 7, pp. 1567–1586, Jul. 2025, doi: 10.1007/s12541-025-01216-7.
20. [20] H. Xie et al., “Analysis of the performance of LSTM-DNN models with the consideration of signal complexity in milling processes,” J Intell Manuf, 2025, doi: 10.1007/s10845-025-02646-w.
21. [21] V. Kumar, V. Dubey, and A. Kumar Sharma, “Comparative analysis of different machine learning algorithms in prediction of cutting force using hybrid nanofluid enriched cutting fluid in turning operation,” Mater Today Proc, May 2023, doi: 10.1016/j.matpr.2023.05.216.
22. [22] C. C. Wang, P. H. Kuo, and G. Y. Chen, “Machine Learning Prediction of Turning Precision Using Optimized XGBoost Model,” Applied Sciences (Switzerland), vol. 12, no. 15, Aug. 2022, doi: 10.3390/app12157739.
23. [23] Vladimir Mitrović, Milan Trifunović, Predrag Janković, Miloš Madić, and Milan Zdravković, “Sound and cutting force components in longitudinal turning of C45 steel at different cutting regimes,” mendeley data, May 2025.
24. [24] V. Mitrović, M. Zdravković, M. Trifunović, M. Madić, and P. Janković, “Dataset for exploring relation between sound and cutting forces components in longitudinal turning of C45E steel,” Data Brief, vol. 62, Oct. 2025, doi: 10.1016/j.dib.2025.112051.
25. [25] P. Seensattayawong, R. Brodmann, G. Stelzer, J. Seewig, and E. Kerscher, “The study of surface in early stages of fatigue failure on carbon steels using a confocal microscope and a scattered light sensor,” Eng Fail Anal, vol. 178, Sep. 2025, doi: 10.1016/j.engfailanal.2025.109720.
26. [26] M. Huang et al., “ReScConv-xLSTM: An improved xLSTM model with spatiotemporal feature extraction capability for remaining useful life prediction of Aero-engine,” Results in Engineering, vol. 26, no. March, p. 105513, 2025, doi: 10.1016/j.rineng.2025.105513.
27. [27] S. Karasu, A. Altan, Z. Sarac, and R. Hacioglu, “PREDICTION OF SOLAR RADIATION BASED ON MACHINE LEARNING METHODS,” 2017. [Online]. Available: www.dergipark.gov.tr/jcs
28. [28] S. Uğuz, Makine Öğrenmesi Teorik Yönleri ve Python Uygulamaları ile Bir Yapay Zeka Ekolü, 3rd ed. Ankara: Nobel Akademik Yayıncılık, 2023.
29. [29] A. Duran, A. Acir, G. Üniversitesi, T. E. Fakültesi, and M. E. Bölümü, “HSS Torna Kalemindeki Talaş Açısının Kesme Kuvvetlerine Etkisi,” Politeknik Dergisi Journal of Polytechnic, vol. 7, no. 3, pp. 211–215, 2004.