Solutions of the steady, two-dimensional Navier-Stokes, thermal energy and species conservation equations have been computed for a low-speed dump combustor geometry with a downstream constriction. The equations are solved in the conservative finite-difference form on a nonuniform rectilinear grid of sufficient resolution to accurately capture the momentum/thermal boundary layers and, with somewhat lower accuracy, the flame structure. The chemistry is represented by a finite-rate reduced methane-air mechanism involving seven species. The computed flame shapes for different cavity lengths and mixture equivalence ratios are compared to those determined experimentally by OH (A) chemiluminescence. Except for the leanest mixture studied, the computed and experimental flame lengths agree to within a few percent. In general, the computed flame lengths depend more strongly on equivalence ratio than does the experimental. The dependence of computed flame shape on cavity length generally agrees with that determined from experiment. Computations indicate that the structure of the recirculation zone is similar to that of a nonreacting flow for short cavities, but is qualitatively different for longer cavities. These differences are a consequence of the heat release in the flame front. Simulations with additional fuel (representing a hydrocarbon waste)injected into the recirculation zone show two stages of heat release. The first is associated with premixed burning of the primary fuel-air mixture, while the second is associated with a diffusion flame at the interface between the oxygen-starved recirculation zone and the comparatively oxygen-rich combustion products from the primary flame. These simulations indicate that excess oxygen from a lean primary flame can be effectively utilized for waste destruction, even for very lean core flames resulting from oxygen enrichment. Oxygen utilization is seen to be somewhat better for longer cavities.
展开▼
