Optimizing biodiesel production from abundant waste oils through empirical method and grey wolf optimizer

The failure of classical techniques and algorithms have triggered researchers to search for stochastic tools capably of exploring the search space with constant convergence speed. Grey wolf optimizer (GWO) is a moderately novel stochastic algorithm with only a few parameters to regulate that can be easily employed for global optimization. For the first time, this study explored GWO to model biodiesel yield. It is worthy to note that, kerosene (KS) popularly known as paraffin oil has gained global attention in the aviation and biodiesel industries to improve cold flow properties and can be mixed with diesel in different proportions. Therefore, in this study, (1) Response Surface Methodology (RSM) and GWO were explored to model the waste sunflower oil methyl ester (WSOME)/biodiesel (BD) production from waste sunflower oil (WSO) and (2) Least square regression method was accosted to correlate the density of (0%KS + 100%BD), (5PKS + 95%BD), (20%KS + 80% BD), (50%KS + 50%BD) and (100%KS + BD) blends. The yield of WSOME (96.70%) was optimum at the methanol/oil molar ratio of 5.99/1, catalyst amount of 1.1 wt.% and reaction time of 77.6 min. The GWO model displayed a higher coefficient of determination, and a lower value of root mean squared errors compared to the RSM model. GWO predicted values, as compared to RSM, predicted yield indicates its reliability and usefulness for prediction without trial and error of conventional experimentation. The fuel properties concurred with the ranges of the ASTMD6751 and EN 14214 specifications. The quadratic relation with high regression coefficient (R2) was detected for the densities of (0%KS+100%BD) and (20%KS +80% BD) while the linear was found suitable for the densities of (5KS + 95%BD), (50%KS +50BD) and (100%KS+0%BD). The model and correlations can find application in biodiesel and aviation industries.