OCR Text |
Show B. STABILITY EXPERIMENTS For this study the natural gas fired flame stabilizer, Fig. 1, was replaced by a new end plate lid, built by Maxon specifically for this study, as shown in Fig. 11. High electrical resistance alumina refractory material cast into the lid provided electrical insulation between the high voltage torus electrode and ground and also helped in centering the torus electrode as shown in Fig. 12. These last two figures should be compared to Figs. 1 and 2 corresponding to conditions prior to the above modifications. Tests were carried out at different levels of firing rate, excess air, and nozzle velocity for different electrode configurations and locations. The torus electrode system was installed through the lid and could be easily moved axially in the combustor. High-voltage was supplied by a dc power supply rated at ± 5 kV and 200 rnA. In order to safely operate the water-cooled electrode with an applied high voltage, water was supplied through 200 ft. of plastic tubing and the same length was used on the outlet side. The total number of hours of exposure of the torus electrode system to the combustor environment was extensive and no damage was noted. Gas temperatures and compositions at port no. 4, close to the combustor exit orifice, were measured to evaluate flame stability and combustion completion in the cyclonic combustor. Flame stability was also monitored by a flame uv sensor located 4.5 in. downstream of the nozzle centerline. Since the location and characteristics of ignition and flame stabilization in the combustor depend on various parameters, data were collected at different operating conditions of firing rate, excess air, and nozzle velocity. Experiments were performed at IGT's Applied Combustion Research Facilities. C. RESULTS OF STABILITY EXPERIMENTS Flame stability was evaluated based on both flame signal and CO emission. It was found that the flame was unstable when extremely high CO (> 1000 vppm) was detected in the region 1 to 2 in. from the combustor wall. The appearance of high CO in this region indicates that the limit of flame stability is being approached. Tests were first carried out at 1 x 106 Btuth firing rate with different excess air and nozzle velocities without disks attached to the electrode. During the testing either a positive or a negative voltage was applied to the water-cooled electrode. To evaluate the effect of the dc field on flame stability, combustion characteristics were measured without applying high voltage to the torus electrode as a comparable base. The location of the electrodes was changed relative to the nozzles. Some of the 9 |