Process Optimization of the Re-Refining of Spent Lubricating Oil Using Response Surface Methodology

Lubricating oils used in automotive engines become contaminated with impurities and residues. Improper management of these oils is a leading cause of pollution worldwide. Re-refining with solvent extraction is a competitive process for recycling them. To optimize this process, used engine oils were collected and homogenized representing typical feedstock for re-refining plants. The oil was characterized and heated to 260°C to remove emulsified water, gas oils, and fuel diluents. It was then mixed with ethanol at 10:1 and subjected to different temperatures, mixing times and speed conditions using Box-Behnken design-of-experiment. The mixture was centrifuged, and the extract distilled to separate and weigh the base oil. An optimal yield of 39.21 wt.% was achieved at 70°C, 20 minutes, and 200 rpm. The quadratic model fitting using ANOVA showed a coefficient of determination (R2) of 92.86% and standard deviation (S) of 1.54, indicating that the model sufficiently explained the variation in base oil yield. The p-value regression (0.021) was statistically significant, demonstrating the model's validity. There was no evidence of lack-of-fit as p-value (0.570) was statistically insignificant. The recovered base oil met acceptable SAE standards, with specific gravity (0.8924), kinematic viscosity (7.63 cSt at 100°C), flash point (194°C), and pour point (-11°C).