A numerical study of highly-diluted, burner-stabilised dimethyl etherflames

Mayer D, Moshammer K, Cai L, Pitsch H, Kohse-Höinghaus K. A numerical study of highly-diluted, burner-stabilised dimethyl etherflames. Combustion Theory and Modelling. 2015;19(2):238-259.Recently, a new burner was designed by Zhang et al. (Proc. Combust. Inst. 34 [2013], 763-770) to enable the investigation of 1D, premixed flames at atmospheric pressure with a temperature in the burnt gases near 1500 K. It consists of a matrix burner plate with alternating fuel and oxidiser feeds that, because of small-scale nozzles, mix quite rapidly. Flames at high dilution and reduced temperatures such as realised here are of relevance for the understanding of low-temperature combustion strategies. In this work, we examine the burner with regard to the validity of the 1D assumption for the investigated flames. Experimental measurements are conducted and 1D and 3D simulations are performed in which the chemistry is described by a detailed chemistry approach based on a reduced reaction scheme derived from the mechanism of Fischer et al. (Int. J. Chem. Kinetics 32 [2000], 713-740). The experimental results are compared to 1D simulations with two different temperature treatments. First, the unburnt temperature is set to the measured temperature closest to the burner surface; second, the experimental temperature profile is prescribed in the whole simulation domain without solving the energy equation. The comparison shows that the 1D simulation predicts the experimental results reasonably well, if the experimentally obtained temperature profile is prescribed in the simulation domain. Differences are found in the mole fractions of methyl and formaldehyde. Further comparisons of the experimental data with 3D simulation results and comparisons of 3D and 1D simulation results indicate that the differences between measured and computed mole fractions of these species are not a result of the 3D nature of the experimental flame and might be attributed to the chemical mechanism. The conclusion is that the measurement data can be used for validation purposes with the 1D simulation setup shown here if the measured temperature profiles are prescribed in the 1D simulation domain