Study on temperature and hardness behaviors of Al-6060 alloy during magnetic abrasive finishing process using artificial neural networks

Abstract

Magnetic Abrasive Finishing (MAF) revolutionizes surface finishing by utilizing magnetic fields to apply pressure, ensuring exceptionally high-quality surfaces. Its potential for ultra-precise finishing with fine magnetic abrasives is particularly noteworthy. This study analyzes subsurface temperature dynamics during the MAF process on Al-6060, recognizing temperature's pivotal role in shaping surface texture and mechanical-chemical properties. Unbonded magnetic abrasives containing SiC were used without lubrication or coolants in experimentation. Experimental parameters, including machining gap, abrasive weight, voltage, and rotational speed, were systematically varied using a Box-Behnken Design of Experiments. An artificial neural network facilitated comprehensive data modelling, enabling a detailed parametric study. Post-processing surface characterization provided crucial insights into the impact of surface temperature rise on surface finish and hardness. This holistic approach enhances understanding of temperature's influence on MAF efficacy and its outcomes on Al-6060 surfaces. Experimental findings demonstrate MAF's effectiveness in achieving low-temperature finishing of Al-6060, with a maximum surface temperature of 36 °C. ANOVA analysis quantitatively determined that voltage has greatest influence (55.56%), followed by abrasive weight (20.00%) and machining gap (13.89%), with rotational speed having the least impact (2.22%). Qualitatively, the developed ANN model accurately predicted changes in surface roughness (ΔRa), temperature (ΔT), and hardness (ΔH), with maximum errors of 4.107%, 5.588%, and 6.680%, respectively. XRD analysis provided quantitative evidence of SiC diffusion into the surface, resulting in a hardness increase from 2.7 HV to 5.6 HV. Additionally, chemical substrates deposition, such as silicon carbide, kaolinite, iron silicide, and cristobalite, elucidated chemical-mechanical interactions during the process at such low temperatures.

Keywords

• Magnetic abrasive finishing
• Artificial neural network
• Temperature
• Hardness
• Abrasive particle deposition
• Process innovation