OCR Text |
Show O2, dry). This case was selected to assess its usefulness to evaluate model sensitivity. The conditions reduce combustion air flow by ten percent, and result in a roughly 22% increase in exit NOx. However, although this condition does show sensitivity in NOx levels, most other combustion quantities are largely unaffected, making the case useful primarily for evaluation of detailed nitrogen kinetics mechanisms. Traverses at this condition were made at the 0+,27,65, 131, 150,343, and 510mm locations. Similar data as collected for the baseline case - velocity components, gas temperature, and major species - were collected for this condition. Cold Flow Measurements In addition to the baseline flame measurements, data was also collected for non-reacting flow conditions. Seven traverses were completed for the non-reacting case (0+, 27, 65, 131,208,343, and 510mm) to provide turbulent flow validation data for a representative swirl number. This data was of considerable interest, since turbulence plays a critical role in diffusion flame combustion, and these measurements provided detailed flow validation data for the same burner configuration without the added difficulty of combustion chemistry and heat transfer. Burner air flows identical to those used for the baseline combustion case were used, here, with no change to the burner swirl setting. For these conditions, the internal recirculation zone (IRZ) observed for the baseline flame does not close, producing a large secondary recirculation zone downstream along the burner centerline. Although these conditions produce a somewhat different flow pattern than that observed with combustion, the result is a conservative test case for evaluation of turbulence models and numerically predicted swirling flow. Auxiliary Measurements In addition to the described in-flame measurements for the three cases, a significant amount of auxiliary data was also collected to provide a thorough set of boundary condition information and to provide qualitative visualization of the flame character. Burner operating conditions such as air and fuel flow and temperatures, as well as fuel composition, were monitored and documented throughout the campaign, as shown in Tables I and 2. Furnace exit conditions were recorded and furnace massflow balance was monitored with pressure drop measured across an orifice installed in the exit duct. This orifice was duplicated in some modeling to successfully confirm minimal furnace leakage through pressure drop comparisons. A summary of furnace exit conditions over several test days is provided in Table 4. Temperatures in the furnace refractory floor and on the outside of the furnace hood were recorded using thermocouples. Radiative heat flux at the furnace wall was measured with an ellipsoidal radiometer at 12 elevations. Images of burner flow were obtained using Mie scattering from fine particles of titanium dioxide formed by reaction of water vapor with titanium tetrachloride vapor added to the combustion air. The 532 nm beam from a frequency-doubled Nd:Y AG laser was formed into a sheet approximately Imm thick using cylindrical lenses to expand the beam in one dimension. Introduction of titanium tetrachloride was started impulsively to seed the combustion air. An image was then obtained of the seeded flow entering and mixing with the unseeded furnace flue gas by proper timing of the camera shutter with respect to the impulse. A pulse of 6ns duration from the laser effectively freezes the flow. A narrow bandpass filter centered at 532nm was used to minimize flame and background radiation. The illuminated sheet of particles was focused using a 105mm, fl4.5 UV lens. An electronic shutter with a I microsecond gate time was used with a high resolution Photometrics 14-bit AT200, 1024xl024 pixel CCD detector coupled with an electronically gated image intensifier. The digitized signal was process using PMIS software for Microsoft Windows. Mie scattering images were collected for the non-reacting measurements as well as both the baseline and low excess air flames. The relative time lapse between introduction of the titanium tetrachloride and the camera shutter allows the mixing to be visualized at several stages. Although these images represent nearly instantaneous snapshots of the flame structure, they are very useful in defining overall flame character and shape; for example the closed internal recirculation zone of the baseline flame is readily visible in Figure 6, compared with the open central recirculation zones formed by the non-reacting flow. The images are also useful in "defining key mixing layers and characterizing larger-scale turbulent structures. 5 |