An Investigation of Ultra Low NOx Natural Gas Burners

Page 9

food (e.g. malt drying) and the other two cases illustrate the overall effects of stoichiometry on NO formation for this burner. x In the model the combustion of methane was described by the two-step reaction mechanism, reactions (i) and (ii), which had the advantage of shorten i ng the comput i ng time wh i ch was already exten s i ve due to the 3 -D treatment of the model. The choi ce of the gl oba 1 k i net i c express ions is not s t ra i ght forward and there is some variation due to changes in stoichiometry and temperature range. Thus the use of the Dryer and Glassman [3] data in the present computations failed to allow the fuel-air mixtures to ignite due to the high activation energies demanding high temperatures and high initial concentrations of fuel. Furthermore, the time in which the highest concentrations of fuel were in contact with the air was reduced by the high velocities involved. Clearly such kinetic data which were suitable for laminar fuel rich mixture [1] could not apply for turbulent/semi-turbulent, fuel-lean/stoichiometric reacting flows as is the case here. Thus after a sensitivity analysis of the kinetic data at ignition the following data was used[2]: dXCH4 1010 exp(-120l9/T)XCH4 X02 dt (6) dXC02 = 1010 exp(-120l9/T)XCO X02 dt (7) where Xi is the mol fraction of the species i . Computat ions of the fl ame behavi our were undertaken us i ng the CFD code assumi ng turbu1 ent combust i on and i nformat ion obta i ned on ve 1 oc i ty, composition and temperature fields. Figure 4 gives the velocity in the x direction in the centre of the recirculation zone and at the exit. Figure 5 shows in its top half the methane concentration and on its bottom hal f the oxygen concentration predicted by the computer model for the reacting flow in the fuel-lean case (mode A). The initial gas 9