Detailed mechanism of CH4/O-2/Ar flames and modelling in fuel-rich conditions

Experimental concentration profiles of species occurring in five methane-oxygen-argon flames burning at lon pressure and at equivalence ratios varying from 0.92 to 1.94 have been simulated by using the CHEMKIN package. The numerical modeling has been performed by including as input parameters the initial composition, the initial total mass flux and the experimental temperature profile measured in similar conditions as the concentration ones. Several reaction mechanisms available in the literature have been tested (GRI, Sanogo, Hennessy, Warnatz, Fukutani, Purl, Senkan and Miler) and the calculated mole fraction profiles have been compared with the experimental ones and this over the whole equivalence ratio range. The eight tested mechanisms differ from one another by the number and the type of the elementary reactions selected, as well as by the rate coefficients. The calculated mole fractions of different species and the location of the maximum values depend drastically on the mechanism chosen. The best agreement between experimental data and simulated results is reached when the GRI mechanism is used to simulate the flame burning close to the stoichiometry. The correlation becomes less satisfactory when the equivalence ratio increases. The other mechanisms lead to large discrepancies from the experimental data. The position of the maximum concentrations of the intermediates and the location of the largest gradient of the mole fractions of H, O and OH species vary strongly with the mechanism tested and likely induce large changes on the mole fraction profiles of the other species.\nA new mechanism elaborated with data from Baulch's evaluation for C-1 species, with the H-2/O-2 reaction mechanism from Westbrook and for C-2, C-3 and C-4 species from the Miller mechanism has been evaluated against the experimental results. The shape of the profiles and the maximum concentrations of major stable species like H-2, CH4, H2O, CO, O-2, Ar and CO2 are well simulated in all flames within the equivalence ratio range. Calculated mole fractions of methyl radicals are generally overestimated by about a factor of two irrespective to the equivalence ratio. For C-2-species, the experimental mole fraction profiles agree in some extent with the simulated ones. Finally, poor predictions are obtained for C-3- and C-4- species. The simulation predicts 2 to 3 times more C3H3, 25 times less C3H4 and 2 times more C4H2 than experimental values. It reflects the lack of knowledge of the rate coefficients of the reactions involving these compounds.