Intelligent modeling and optimization of the liquid CO₂-MQL burnishing operation for friction and environmental emission

Abstract

In this study, a novel liquid CO₂-MQL burnishing (LCMB) process is introduced to simultaneously improve tribological and environmental outcomes during the machining of AISI H13 steel. The process considers key input parameters, including pressure of the compressed air (P), the oil flow rate (L), and the CO₂ flow rate (C), while the outputs are the coefficient of friction (COF), the maximum wear depth (WD), the PM10, and the vibrational amplitude along the X-axis (VX). An advanced modeling-optimization framework is employed, combining a Bayesian-tuned support vector machine (OSVM) with entropy-based weighting to characterize the responses, followed by a multi-objective grasshopper optimization algorithm (MOGOA) to generate candidate solutions. The weighted aggregated sum product assessment (WASPAS) is utilized to find the proper solution. The results demonstrated that the COF, WD, PM10, and VX were decreased by 5.9%, 3.5%, 22.3%, and 10.4%, respectively, as compared to user-defined values. The developed burnishing process significantly enhanced subsurface integrity by producing a deeper and more uniform plastically deformed layer compared to MQL and dry conditions, which directly contributes to improved hardness, wear resistance, and tribological performances.