OCR Text |
Show · EXPERIMENTAL VARIABLES AND MEASUREMENTS In the following, experiments are reported which were carried out in the MIT-CRF with natural gas. The heat input in each case was about 1.0 MW thermal, and the combustion air was preheated to 450°F. Two series of experiments were made: parametric study of the effects of fluid dynamic parameters upon NOx and CO emissions, and detailed investigation of a fluid dynamically optimized flame. Numerical modeling (Fluent) of the flow, temperature and major stable species concentration distributions provided guidance for the experiments. Input variables such as the fuel and air flows, and the air preheat were maintained by automatic control at their set levels during the experiments. The distribution of the air flow, and the swirl degree in the individual burner nozzles, were hand controlled. The gas temperature distribution in the flames was measured by suction pyrometer and the CO, CO2, O2 and NOx concentrations of the gas, sampled at several points in the flame and in the exhaust, were determined by NOIR, paramagnetic and chemiluminescence continuous analyzers. The experimental ranges of burner operating variables were: Fuel jet velocity (natural gas) 15 - 200 m/s Fuel jet angle (natural gas) 0 - 25° Fuel gun position 0 to -0.45cm (retracted) Primary. secondary. tertiary air flow rates 0 - 1 00% Swirl numbers in primary. secondary and tertiary air S = 0 to 2.9 RESEARCH RESULTS Parametric studies of natural gas flames (aerodynamic variables) The primary air flow as a fraction of the total air flow rate, the radial distribution of the swirl velocity at the burner exit, and the axial position of the fuel gas introduction were found to be the significant input parameters affecting the NOx and CO emissions from the RSFC burner. Figure 3 shows a monotonic increase in NOx emission with increasing primary air fraction. This result is not unexpected beca~se an increased flow rate of primary air can be seen to promote early fuel-air mixing and NOx formation in the flame. It is noteworthy, however, that the reduction in primary air flow did not increase CO emission from the flame. The effect of the total air swirl, characterized by the swirl number, S, is shown in Figure 4. Of the vortex flow types produced with the variation of the radial distribution of the swirl velocity, the "Rankine" vortex was found to be the most favorable. In the Rankine vortex the core of the rotating flow rotates as a solid body, with the swirl velocity increasing from the center linearly with radial distance to a maximum at the core boundary, from where it decreases hyperbolically with further increase of the radial distance. With this type of swirl distribution the NOx emission drops to a low value of 82ppm for S - 0.6, which is the critical swirl number for the onset of the IRZ. 4 |