Prediction of cetane number of biodiesel from its fatty acid ester composition using Artificial Neural Networks

Abstract

Models for estimation of the cetane number of biodiesel from their methyl ester composition using artificial neural networks were obtained in this work. An experimental data that covers 48 and 15 biodiesels in the modeling step and validation step respectively were taken. The selection of biodiesel samples took into account a wide range of ester compounds with different unsaturation characteristics and number of carbon atoms. A model to predict cetane number using artificial neural networks was obtained with better accuracy than 95 %. The best neural network for predict the cetane number was a backpropagationnetwork(11:4:1)using a Conjugate Gradient Descend algorithm for the second training step and showing 96.3 % of correlation for the validation data and a mean absolute error of 1.5.The proposed network is useful for prediction of the cetane number of biodiesel in a wide range of composition but keeping the percent of total unsaturations lower than 80 %. The use of the artificial neural networks in this case let to study and understands the effect of individual fatty acids in the cetane number. Therefore they can be used for improving some biofuel properties related to the combustion process and its efficiency.