OCR Text |
Show (ii) Atmospheric pressure. (i ii) Two burning modes were considered, namely the freely propagating adiabatic flame (FPA) and the burner stabilised flame (BSF) using an imposed profile of temperature following the one obtained for the freely propagating adiabatic flame until the temperature ceiling found for maximum in the CFO commercial package results: 1541 K, 1757 K and 1930 K respectively for the initial methane mol fractions 0.075, 0.085 and 0.095. (iv) Two kinetic schemes were adopted. Firstly the reaction mechanism and kinetic data of Miller and Bowman [15], designated M&B in Table 2, which involves 24 species and secondly the reaction mechanism and kinetic data modifications on the Miller and Bowman mechanism made by Westmoreland and co-workers [14] (designated W in Table 2 which i nvo 1 ves 31 spec i es, inc 1 ud i ng the OeD) thermal NOx route and the nitrous oxide route to NO. The results obtained by applying the different combinations of the above conditions together with those using the global model in the CFD code (CFD (turb)) are summarised in Table 2. We may conclude that: (i) The CFD model involving turbulent flow and our global kinetic expression gives good agreement with the experimental data for modes A and C. (ii) The freely propagating laminar flames equivalent to modes A, Band C - were found to give slightly different values for total-NO depending on whether the Miller and Bowman or Westmoreland kinetics are used. The burner stabilised flames for which the Miller and Bowman mechanism was employed gave good agreement in the lean cases but a lower value than the CFD results in the stoichiometric case (26.6 ppm against 110 ppm). The difference between the two is partly due tp the neglect of the effect of the flame thickness (i.e. reaction time) in comparing the laminar I-D PREMIX model with the actual turbulent flame. 12 |