Prediction of Biodiesel Properties from Fatty Acid Composition using Linear Regression and ANN Techniques

Biodiesel is currently the most widely accepted alternative fuel for diesel engines due to its various advantages. The fatty acid composition of vegetable oils affects the fuel properties of biodiesel, such as viscosity, flash point, fire point, cloud point, pour point, iodine value and saponification value. In the present work, biodiesel was prepared from different vegetable oils and its physical properties measured. Artificial neural networks (ANNs) with 6-4-1 and 6-5-1 architecture were used to determine the relative importance of the fatty acid (palmitic acid, stearic acid, oleic acid, linoleic acid, linolenic acid and erucic acid) composition of vegetable oils affecting biodiesel properties. The results show that the ANNs precisely predict the properties of biodiesel derived from certain vegetable oils better than a linear regression model. The relative importance of the input variables shows quantitatively the effect on biodiesel properties.