Lean premixed laminar methanol flames: a computational study

Three experimental premixed laminar methanol flames at 40 Torr, measured in detail by Vandooren, Balakhin, and Van Tiggelen, were studied computationally. The equivalence ratios of the flames were 0.89 (1), 0.36 (2), and 0.21 (3). Flame 3 seemed to be the flame with the highest experimental accuracy. A time-dependent flame code with an implicit description for both chemistry and transport allowed extensive computations. Two kinetic mechanisms for methanol combustion were used to compute the species profiles. The first was based on the mechanism compiled by Westbrook, Dryer, and Schugh (WDS), but with 5 times the rate constant for the reaction HCO + M. The second mechanism was compiled by Dove and Warnatz (DW). Strikingly, the experimental CH/sub 2/OH maxima were much smaller than the corresponding computational maxima. The smallest differences were found in flame 3. In this flame the experimental CH/sub 2/OH maxima were about 70 and 20 times smaller than computational maxima found by using the WDS and DW mechanisms, respectively. The sensitivity analysis gave analogous results for the three flames. It was found that the CH/sub 2/OH profiles in the three flames were dominated by the reaction CH/sub 2/OH + O/sub 2/. An increase in the rate constant of reaction CH/sub 2/OH + O/sub 2/ decreased the CH/sub 2/OH maxima nearly linearly, leaving other species profiles almost unchanged. For this reaction a rate constant (WDS) of about 1.0 x 10/sup 14/ exp(-3019/T) cm/sup 3/ mol/sup 1/ s/sup -1/ gave good agreement between the experimental and computational maxima for CH/sub 2/OH in flame 3.