Although martensitic stainless steels (MSS) are widely used due to their superior mechanical and chemical properties, their poor machinability can lead to surface quality degradation. Therefore, in addition to improving the surface quality of the material using advanced machining methods, it is important to predict the surface quality to increase efficiency, reduce waste, and prevent faulty functionality. In this study, surface roughness was predicted for the first time using various machine learning models such as, polynomial regression, support vector regression, decision tree regression, random forest regression, and k-nearest neighbors regression, depending on the cutting condition, minimum quantity lubrication (MQL) flow rate, nanoparticle concentration, cutting tool coating, and cutting temperature in sustainable milling of AISI 420 MSS using the MQL method with multi-walled carbon nanotube (MWCNT)-reinforced nanofluid (NMQL). Titanium nitride (TiN)-coated tungsten carbide (WC) cutting tool, NMQL method containing 0.15 wt.% MWCNT and 40 mL/h MQL flow rate improved the cutting temperature and surface roughness by 25.3% and 49.9%, respectively, compared to dry cutting condition with the uncoated WC cutting tool. he K-nearest neighbors regression model achieved the highest predictive accuracy for surface roughness, with a coefficient of determination (R²) of 0.9789, outperforming the other evaluated machine learning approaches.
Yapan et al. (Wed,) studied this question.