OCR Text |
Show the efficiency of an axisymmetric solution, the axisymmetric models remain a viable alternative for qualitative analysis of a flame. DISCUSSION There are two potential culprits in producing the "qualitative" results in the IRZ of Figures 10 and 11: the k-e turbulence model and the two-step chemistry. The lack of IRZ agreement is largely an outcome of the turbulence model and/or the 2-step chemistry assumed for the NO combustion. The simplistic turbulence model continues to thwart development of accurate predictive schemes. Is there a need to use second-order closure? Work is underway to improve the NO chemistry by using a reduced kinetic mechanism. It is not likely that NO combustion can be adequately characterized by a 2-step combustion mechanism or an isotropic turbulence assumption, a principal assumption of the k-e model. Future work will compare the predictions from a variety of chemistry models. CONCLUSIONS A comprehensive model for natural gas combustion has been compared to data with good results. The model captures most of the trends well including the velocity, temperature, and species profiles. The largest differences between the model and data occur in the IRZ. The 3- D models reproduce the flame better, but axisymmetric models are able to qualitatively capture key trends in the flame making them useful for scoping studies. Prediction of turbulent combustion, even in simple geometries, remains a significant challenge. The models for turbulence and combustion which are used in today's practical codes are not adequate to achieve good quantitative agreement with data. With care they can be used to qualitatively assess new burner and furnace designs, but the limitations of the models should be recognized. Future work should focus on improving the turbulence and combustion models. Despite significant developments in stochastic PDF methods, much work is still needed in this area before these models may be applied on a daily basis to practical combustion systems. Studies using simplified kinetics are now underway to improve predictive capability. 8 |