Construction and assessment of reduced oxidation mechanisms using global sensitivity analysis and uncertainty analysis

The global sensitivity analysis is an effective method to assess the performance of a chemical reaction mechanism. In the present study, the uncertainty analysis and the global sensitivity analysis of a detailed chemical mechanism of n-heptane are first performed based on the Monte Carlo method. The source of the prediction uncertainties of the C₀-C₇ sub-mechanisms and the reaction classes in the fuel-specific sub-mechanism of the detailed mechanism are determined by the global sensitivity analysis. Then, based on the results, a reduced mechanism for n-heptane oxidation is developed. To assess the performance of the reduced mechanism, the nominal model prediction and the frequency distribution of the ignition delay times using the present reduced mechanism are compared with those of the detailed mechanism and two additional reduced mechanisms obtained by the directed relation graph with error propagation (DRGEP) method. The results indicate that the predictions from the present reduced mechanism and the reduced mechanism with 305 species satisfactorily agree that of the detailed mechanism. Furthermore, the discrepancy of the predictions among these mechanisms is discussed based on the Spearman Rank Correlation analysis. It is found that, for the reduced mechanism with 120 species, the optimization of the reaction rate constants significantly improves the nominal model prediction of the ignition delay time, whereas its influence on the range and profile of the frequency distribution is rather weak due to the intrinsic relationship among the reactions being broken in the reduced mechanism.