A functional-group-based approach to modeling real-fuel combustion chemistry – III: Application to biodiesels

Abstract

Real biodiesel fuels are mixtures comprising many high molecular weight components, making it a challenge to predict their combustion chemistry with detailed kinetic models. Our group previously proposed a functional-group approach (FGMech) to model the combustion chemistry of real gasoline and jet fuels; this approach has now been extended to model real biodiesel combustion and mixtures with petroleum fuels. As in our previous work, a decoupling philosophy is adopted for construction of the model. A lumped reaction mechanism describes the (oxidative) pyrolysis of fuels, while a detailed base chemistry model represents the oxidation of key pyrolysis intermediates. However, due to the presence of the ester group, several oxygenated species are identified as additional primary products and incorporated into the lumped reaction steps. In addition to the lumped reactions initiated by unimolecular decomposition and H-atom abstraction reactions, a lumped H-atom addition-elimination reaction is also incorporated as a new reaction class to account for the presence of double bonds. Stoichiometric parameters are obtained based on a multiple linear regression (MLR) model, which establishes relationships between the fuel's functional group distributions and the stoichiometric parameters of the lumped reactions. Global rate constants are developed from consistent rate rules obtained from pure fuels. New pyrolysis experimental data for methyl pentanoate/methyl nonanoate and methyl heptanoate/n-heptane mixtures (50%/50% in mol) are obtained in a jet-stirred reactor at atmospheric pressure. In general, kinetic models developed using the FGMech approach can reasonably reproduce all the validation targets obtained in this work, as well as those in the literature, confirming that functional-group-modeling is a promising approach to simulate combustion behavior of diesel/biodiesel surrogate fuels and real biodiesels.