Reduction of Detailed Chemical Mechanisms by Entropy Production Analysis in the Presence of Irreversible Reactions

The chemical composition of renewable fuels can be very complex and chemical mechanisms composed by hundreds of species and several thousands of reactions can be required for an accurate prediction of their combustion behavior [1, 2]. An important step for their computationally efficient use in simulations is, therefore, the possibility to obtain reduced mechanisms with an assessed range of validity. Kooshkbaghi et al. [3, 4] proposed an easy to implement approach for mechanism reduction that is based on the contribution of the entropy generated by the elementary chemical reactions as a measure of their importance. A simple analysis of thermodynamic states during the chemical transformation of reactants to products allows for the identification of the most relevant reactions of the detailed mechanism, thus deriving skeletal mechanisms which only include the most significant reactions for a particular application. Being based on thermodynamic law, the degree of approximation can be selected and verified. In the original formulation [3], the evaluation of the entropy generation was based on the principle of detailed balance, a property that does not hold in the case of irreversible reactions, which are extensively adopted in the development of several detailed mechanisms for complex fuels (e.g. [6]). Therefore, It is advisable to have a more general formulation which also applies in presence of irreversible reactions. In this work, an alternative formulation is derived starting from first principles of chemical reaction thermodynamics. A formulation valid for both reversible and irreversible reactions is derived that can be adopted as the basis for the reduction of a wide set of detailed mechanisms. The new approach is then exploited to obtain a skeletal mechanism for n-dodecane starting from the detailed reaction scheme developed by the CRECK group [6], a mechanism composed mainly by irreversible reactions.