Automatic identification and lumping of high-temperature fuel decomposition pathways for chemical kinetics mechanism reduction

Abstract The predictive capabilities of Computational Fluid Dynamics (CFD) for combustion systems rely on a proper description of the fuel chemistry. The growing interest in accurately capturing the combustion behavior of multi-component fuel mixtures creates additional challenges in developing reduced-order chemical kinetics mechanisms small enough to be used in CFD. Among the suite of chemistry reduction approaches available, lumping techniques appear especially suited to handle the complex nature of multi-component combustion chemistry. In particular, published literature provides very strong evidence that the lumping of non-rate-limiting pathways, and more specifically, the high-temperature fuel decomposition reactions, is a powerful avenue for multi-component mechanism reduction. In this work, we present a novel algorithm to identify and lump high temperature fuel decomposition reactions from detailed kinetic mechanisms. The lumping strategy is fully automatic, and relies exclusively on information available in the detailed mechanism. The performance of the technique is assessed for both a single-component fuel, n-dodecane, and its mixture with iso-octane. Results show that replacing the fuel decomposition sub-mechanism by a small number of reactions involving a single equivalent fuel component introduces very limited changes in the prediction of laminar flame speeds, ignition delay curves, and species profiles. This establishes a clear potential for the proposed algorithm to become a valuable addition to existing multi-stage mechanism reduction software.