Development of a Model For Predicting the Yield of Biodiesel During Biodiesel Production

The production of methyl esters using vegetable oil or animal fat requires selective catalysis and controlled process conditions to meet biodiesel specifications and optimum yield. In this work, a suitable model for the optimization of biodiesel yield as a function of live independent process parameters namely: reaction time, reaction temperature, stir speed, catalyst concentration and methanol-oil ratio was developed using the central composite design and response surface methodology. Alkali transesterification was used for the conversion of palm olein virgin oil to biodiesel with 99.8% pure methanol and sodium hydroxide (NaOH) as catalyst, methanol being in excess to oil ratio (4:1 to 9:1). The designed experiment was carried out using a four-level-five factor central composite design model and response surface methodology to study the interaction of the independent variables; reaction time (1-5 hours), temperature (40-90oC), stir speed (200-400 rpm), catalyst concentration (1-2 wt%) and methanol-oil ratio (4:1 to 9:1) on the biodiesel yield with the response in terms of percentage yield. A predictive model was formulated which correlates the yield of biodiesel to the five process variables. The regression model was found to be highly significant at 95% confidence level as correlation coefficient R (0.985), R-Squared (0.9700), adjusted R-Squared (0.9686) and predicted R-Squared (0.9653) was very close to 1. This is an indication that very small deviation exists between the experimental and predicted values. Results obtained from the model output was in good agreement with experimental values hence the developed model can be employed to predict the yield of biodiesel.