Publication: Investigating Cryogenic and Heat Treatment Effects on Hardness and Wear of Uncoated Carbide Inserts: A Microstructure and AI Approach
5
0
Issued Date
2024-01-01
Resource Type
ISSN
10599495
eISSN
15441024
Scopus ID
2-s2.0-85212837591
Journal Title
Journal of Materials Engineering and Performance
Rights Holder(s)
SCOPUS
Bibliographic Citation
Journal of Materials Engineering and Performance (2024)
Suggested Citation
Chanpariyavatevong A., Se C., Timtong A., Boongsood W., Wiangkham A. Investigating Cryogenic and Heat Treatment Effects on Hardness and Wear of Uncoated Carbide Inserts: A Microstructure and AI Approach. Journal of Materials Engineering and Performance (2024). doi:10.1007/s11665-024-10586-4 Retrieved from: https://hdl.handle.net/20.500.14740/20531
Author's Affiliation
Corresponding Author(s)
Other Contributor(s)
Abstract
Carbide cutting tools are widely employed in industrial applications involving the machining of challenging materials. Enhancing tool life and mechanical properties is a critical concern within this domain. Cryogenic treatment represents a promising approach to achieving these objectives. This study investigated the impact of cryogenic treatment on carbide cutting tools, employing various treatment parameters including soaking period, tempering temperature, and tempering time. The objective was to evaluate changes in microhardness and wear resistance before and after the treatment. To investigate the microstructural and phase changes induced by cryogenic treatment, scanning electron microscopy (SEM) and x-ray diffraction (XRD) were employed. To identify the optimal cryogenic treatment parameters for carbide cutting tools, this study employed machine learning techniques such as linear regression model (LRM), support vector machine (SVM), and extreme gradient boost (XGBoost) to estimate hardness, and SHAP value analysis to assess the significance of various treatment factors. The results show that the cryogenic treatment of carbide cutting tools results in enhanced hardness, wear resistance and microstructural changes within the tungsten carbide. Additionally, a phase transformation within the cobalt binder was observed. The machine learning analysis demonstrated that XGBoost outperformed LRM and SVM in terms of predicting microhardness, as evidenced by the R2, RMSE and MAE metrics. Additionally, SHAP value analysis identified the soaking period as the primary factor influencing microhardness, followed by tempering temperature and tempering time.