OCR Text |
Show In th~s. case k = 7.14 and. Xo = -15 mm. This demonstrates that the jet similarity hypothesis in reacting condItions, can be establIshed and the radial concentration profiles can all be expressed as a function of 1]. When. the downstream distance is greater than 150 mm however, this hypothesis breaks down because the Jet becomes asymmetric. This indicates that the effect of buoyancy or the cross-flowing air stream has become significant. On the centerline, where g( 17) equals 1, equation (3) becomes: f = C(x,r) =_1 Co cz (6) Since the TSL model is based on a single jet mixing constant, c, defined in equation (6), It IS important to compare our experimental results with the constant c used in the model. Because of assumptions used to derive equation (6) and the wide range of experimental values existing in the literature for c, it is crucial to estimate this constant experimentally. Figure 9 represents the centerline seed concentration profile obtained from the averaged Mie scattering images for all three conditions. The dashed line represents the analytical expression (6) for the constant c = 0.16 typically used in the TSL model. In general, the agreement between our measurements and previous measurements is quite good. Specifically, the agreement between the model and the experiment is excellent for a downstream distance region of 10 < xld < 20 (where d is the jet nozzle diameter) in all three cases. Where 20 < x/d < 60, the model overestimates the amount of surrounding gas entrained by the fuel jet, while at xld > 60, the jet begins to bend upward and this analysis no longer applies. In order to discuss these three jet regions, it is important to note the model assumptions. First, the model assumes that the jet is highly turbulent and enters a quiescent atmosphere. Our observations indicate that an axisymmetric lifted flame is present downstream of the fuel jet nozzle at around xld = 20. Because heat release changes the fluid viscosity, flames often act as mixing shields. This observation is consistent with the discrepancies observed between the model and the experiment where 20 < xld < 60. The difference between the model and the experiment in the third region (xld> 60) can be attributed to the jet bending upward due to buoyancy or entrainment by the combustion air flow or both. Residence Time Distribution Mixing was also considered at the opposite end of the range of spatial scales, using a residence time distribution to characterize the mixedness of furnace gas. A possible division of the furnace for modeling purposes, proposed by C. F. Edwards (unpublished work, 1995), is shown in Figure 4. In this picture the near-burner region is divided into three parts, the in-quarl, near-field, and recirculation zones, followed by a plug flow in the post-flame region of the furnace. An experimentally-determined residence time distribution. is not ex~ected t~ res.olve details associated with the near-burner region, but may be able to provl.de the resld~nce t~me In the plug ?ow and a composite mean residence time for the in-quarl, near-fIeld, and recIrculatIon zones. SpeCIfically, the 6 |