Effect of Machining Parameters on Surface Roughness in Mild-Steel Turning and Computer-Vision-Based Tool Wear Prediction
DOI:
https://doi.org/10.37255/jme.v21i2pp058-063Keywords:
Machining parameters, Surface Roughness (Ra), Tool Wear, Computer Vision, Semi-supervised learning, OpenVINO, Mild steel, TurningAbstract
In this paper, we present a combined experimental and data-driven study of surface integrity and tool condition in mild-steel turning. Controlled experiments were conducted on a semi-automatic lathe to analyze the effects of spindle speed, feed rate, depth of cut, and coolant use on surface roughness (Ra), with surface finish measured using a Mitutoyo SJ-210 profilometer and high-resolution images of machined surfaces and cutting tools acquired with a Nikon D-3000 imaging setup. Experimental findings showed that coolant application reduced Ra by an average of 22.69%, with the improvement more pronounced at higher spindle speeds due to enhanced thermal control, lubrication, and chip evacuation. To complement these results, a computer vision–based framework was developed for tool health monitoring using semi-supervised anomaly detection, wherein a PaDiM-based model trained exclusively on healthy tool images successfully identified worn-tool states and Grad-CAM heatmaps localized wear regions on tool edges, thereby enhancing explainability. The trained model was further optimized with OpenVINO for edge deployment, enabling real-time inference on CPU-based industrial devices with minimal latency. The contributions of this work are threefold: (i) systematic analysis of machining parameters on surface roughness, (ii) development of a computer-vision–driven framework for tool wear classification and prediction under limited labeled data, and (iii) demonstration of edge-optimized deployment for real-time shop-floor feasibility. This integrated approach addresses critical gaps in machining research by jointly considering machining parameters, measured roughness, and visual tool/surface features, paving the way for intelligent, explainable, and deployable monitoring systems in smart manufacturing.
Downloads
References
1. S. Kalpakjian and S. Schmid, Manufacturing Engineering and Technology, 7th ed. Pearson, 2014.
2. E. M. Trent and P. K. Wright, Metal Cutting, 4th ed. Oxford, U.K.: Butterworth-Heinemann, 2000.
3. S. Y. Liang, R. L. Hecker, and R. G. Landers, “Machining process monitoring and control: The state of the art,” Journal of Manufacturing Science and Engineering, vol. 126, no. 2, pp. 297–310, 2004.
4. Y. Altintas, Manufacturing Automation: Metal Cutting Mechanics, Machine Tool Vibrations, and CNC Design. Cambridge, U.K.: Cambridge University Press, 2012.
5. Jemielniak, “Commercial tool condition monitoring systems,” The International Journal of Advanced Manufacturing Technology, vol. 15, pp. 711–721, 1999.
6. J. Denkena and F. Biermann, “Cutting edge geometries,” CIRP Annals, vol. 63, no. 2, pp. 631–653, 2014.
7. M. A. Siddique, A. N. Khan, A. A. D. Sarhan, and A. K. Ariffin, “Optimization of cutting conditions for sustainable machining of Inconel 718 considering tool wear and surface roughness,” Materials, vol. 15, no. 4359, pp. 1–18, 2022.
8. K. Huang and W.-S. Lee, “Prediction of tool wear using sensor fusion and 1D CNN,” Machines, vol. 11, no. 1008, pp. 1–18, 2023.
9. M. Tabaszewski, M. Musial, and A. Chwalczuk, “Prediction of tool wear in the turning process of gray cast iron using machine learning,” Sensors, vol. 21, no. 5338, pp. 1–20, 2021.
10. D. V. Baykasoglu, O. Korkmaz, and E. Sari, “Gaussian data augmentation for surface roughness prediction using LSSVM and MGGP models,” Technologies, vol. 13, no. 243, pp. 1–14, 2025.
11. S. Pang, M. Liu, and J. Xie, “Semi-supervised anomaly detection in manufacturing visual inspection: A survey,” IEEE Transactions on Industrial Informatics, vol. 18, no. 10, pp. 6784–6796, 2022.
