Assessment of the transesterification stage of biodiesel production ii: optimisation of process variables using a boxbehnken design

Abstract

Renewable energy policies, in tandem with favorable market conditions, have culminated in large increases in biodiesel use globally. With increasing volumes of biodiesel and a broader range of suitable feedstocks, including used oils, stabilization is becoming ever more important. The quality of biodiesel depends on the feedstock, purity of the milled oils, the storage conditions and the transesterification process. In this study, process variables in transesterification were optimized using a Box-Behnken design. The results of linear and non-linear (quadratic) functions and interaction terms (catalyst-FFA (Free fatty acid), stirring-catalyst, FFA-stirring pairs) indicate complex relationships between the set of factors and biodiesel yield. At the 95% significance level, all the linear terms were statistically significant unlike the quadratic term in which only the catalyst was significant. Furthermore, the joint effects of the FFA*Catalyst and the Stir*Catalyst interaction terms were statistically significant whereas the interaction between stirring and FFA had no influence on the yield of biodiesel. The yield distributions for varying combinations of the factors show that the optimum factor combinations for a percentage yield greater than 98 is low (-1/0.5 g) FFA, low (-1/400 rpm) stirring rate and high catalyst amount (1/4 g). Overall, the multivariate model accounted for 86.4% of the total variance in the data