Reduction of large-scale chemical mechanisms using global sensitivity analysis on reaction class/sub-mechanism

Abstract

The reliable chemical mechanisms with compact scale play an important role in engine modeling. In this research, a new method was presented to reduce the large-scale reaction kinetic mechanisms through the global sensitivity analysis on the reaction classes/sub-mechanisms in detailed mechanisms. In the present method, the importance of each reaction sub-mechanism in a detailed mechanism was first assessed using the Morris method. Then, the crucial reaction classes in the fuel-specific sub-mechanism were identified based on their significance on the detailed mechanism predictions and their coupling relationship using the Morris method and the path sensitivity analysis. Third, the reactions in the unimportant sub-mechanisms and the isomers in the dominant reaction classes of the fuel-specific sub-mechanism were lumped. The scale of the final mechanism was reduced further by removing the unimportant reactions in the important small-molecule sub-mechanism, and the initial reduced mechanism was obtained. Final, the optimization of the reaction rate constants in the fuel-specific sub-mechanism was performed under their uncertainties to improve the performance of the reduced mechanism over a wide temperature range. With the proposed method, a detailed iso-cetane mechanism including 1107 species and 4469 reactions was reduced. The reduced mechanism contains only 56 species and 131 reactions. Good agreements between the reduced mechanism and the detailed mechanism were achieved on the predicted results of ignition delay times and species concentrations over wide conditions.