Skeletal mechanism reduction through species-targeted sensitivity analysis

The use of simplifying techniques to obtain skeletal kinetic mechanisms with the required accuracy is often a necessary step when computationally demanding simulations are concerned. In this work, a novel approach for an automatic mechanism reduction, aimed at retaining accuracy on specific target species, is proposed. Starting from the consolidated coupling between flux analysis and sensitivity analysis, a methodology based on curve matching and functional data analysis was developed, through which the importance of a species in the target accuracy is assessed via a proper metric. The error associated with the removal of uncertain species from the detailed mechanism is quantified in terms of distance and similarity indices before and after such a removal, within a Species-Targeted Sensitivity Analysis (STSA) framework. A species ranking is then generated, and the original mechanism is progressively reduced. The whole algorithm also implements several improvements to enhance a faster convergence, and adds a novel criterion to remove unimportant reactions, based on sensitivity analysis to kinetic parameters.The capability of this algorithm was tested through two case studies in this work. A kinetic mechanism for a Toluene Reference Fuel (TRF) was first obtained, with the overall reactivity as reduction target. The numerical procedure allowed to obtain a compact skeletal mechanism (115 species and 856 reactions), able to retain good accuracy in ignition delay time and laminar flame speed predictions of both fuel mixture and pure compounds. More important, two skeletal mechanisms for methane combustion, including chemistry of nitrogen oxides (NOx), were developed, with different degrees of reduction. The agreement between the original and the skeletal mechanisms in terms of NO formation was successfully assessed with satisfactory results. Attention was also dedicated to the choice of the type of reactor where undertaking reduction, which turned out to play a major role in the overall process.