Development of an empirical model to represent the cetane number of biodiesel fuels, for their use instead of fossil fuels in order to reduce air pollution

Abstract

This work introduces a novel empirical, phenomenological equation specifically developed for the accurate prediction of the cetane number (CN) of both pure fatty acid methyl esters (FAMEs) and biodiesel fuels. The model comprehensively incorporates all key structural factors influencing CN, including chain length, degree of unsaturation, molecular weight, concentration (weight percentage), and, importantly, the position of double bonds. A robust and extensive dataset of 135 data points was meticulously examined to create this phenomenological model, enabling CN prediction for biodiesel derived from diverse feedstock oils. All key influencing parameters within the proposed model are thoroughly analyzed, and its predictive capabilities are rigorously compared against both experimental data and existing empirical models. The model demonstrates a high degree of accuracy, yielding Average Absolute Relative Deviation (AARD) and Average Standard Relative Deviation (ASRD) values of 2.20% and 2.69%, respectively, for predicting the CN of pure FAMEs. For a large experimental dataset of biodiesel fuels, the corresponding values are 3.70% and 5.07%, respectively. Significantly, the proposed model exhibits superior performance compared to 13 other models reported in the existing literature. This demonstrates that the newly developed predictive model can successfully and accurately estimate the CN of both pure FAMEs and biodiesel fuels, offering a valuable tool for reducing the time and cost associated with engine testing.