OCR Text |
Show AFRC - 1996 Int. Symposium Baltimore September 30 - October 2 1996 Page 13 of 14 experimental data from a 2.25 MWt natural gas flame. This flame, and the modeling of it has been described in detail by Weber and Peters [1995]. The flame temperature, needed in this source term for the NOx production was calculated in each volume element assuming that natural gas is burnt to H2 and CO. This procedure is similar to the one used by Weber and Peters. The results showed that the predicted NOx emISSIon, and the distribution of the NOx concentrations at the various locations in the flame are very similar to those measured and to those obtained with the NOx model presented in the paper cited above (Breithaupt and Weber, 1995). More validation work should follow. 5. CONCLUSIONS In this paper, the stability of flames fed by a hot mixture of natural gas, air and flue-gas has been examined. The data show that, maintaining constant flame stability, there is an almost linear relation between mixture temperature and fuel concentration. In the flames under consideration (n > 1.2), the initial oxygen concentration (12 - 19%) was of minor importance with respect to flame stability. It is suggested that the effect of flue-gas admixture on flame stability is close to the effect of a decrease of the fuel concentration only. It has been shown that a lower flame temperature may be achieved when the initial temperature of the natural gas/air/flue-gas mixture is higher. Since NOx correlates strongly with the flame temperature, this implies that, maintaining the same flame stability, it is possible to obtain a lower NOx production for mixtures with a higher initial temperature. This suggests that the technique of internal (hot) flue-gas recirculation, can give lower NOx emissions than the technique of external (cold) flue-gas recirculation, because more fluegases can be admixed. A relation between NOx production and adiabatic flame temperature has been presented. Preliminary tests using this relation as a NOx production (source) term in a CFD-code showed encouraging results. More work should follow. An important advantage of this source term may be that it allows for a decoupling of the NOx production from the residence time. The temperature dependence agrees with previous work on premixed and non-premixed flames. Acknowledgements We thank J. Wamatz for the use of his numerical codes. The experiments have been carried out by B. v/d Vegt, J. Zanting, M. Lazonder and A. Sagarzazu from the Hanze-Hogeschool in Groningen, The Netherlands. |