| OCR Text |
Show 3 Combustion Modifications As discussed below, a number of these approaches have been investigated in the past by IGT and CTI with funding support from SoCalGas and GRI, while others have been described in technical literature. Earlier investigations at IGT were carried out on a 4.5-ft-wide by 3-ft-high by 11.2-ft-Iong glass tank simulator using three different one-fourth-scale glass melter ports. The furnace load consisted of watercooled panels on the furnace hearth covered with layers of refractory and molten glass. Some of the methods were subsequently field-tested on both major types of glass tanks -- a 165-tonlday endport furnace and a 250-ton/day sideport furnace. Details of the tests and the findings have been presented (2,4,6,7,8,9). These will only be summarized here with focus on NOx emissions. Other measurements are not discussed because of space constraints. It should be noted, however, that for all the IGT simulator data presented here, CO remained below 50 vppm. Low Excess Air Firing: Lowering the excess air to reduce the 02 availability in the flame zone to the minimum necessary for high-efficiency combustion. This technique was tested on the IGT glass tank simulator, as well as on two commercial glass melters. The results presented in Figure 1, as a generalized correlation, show good agreement between the pilot- and full-scale data. The data show that decreasing the excess air from 150/0 to 50/0 reduced NOx by 35%. This is a very effective NOx reduction method -- more so because of the added benefit of improved heat transfer. Flame Shape Modifications: Adjusting the mixing of gas and air to alter the shape of the flame, with a long, lazy, luminous flame resulting from decreased fuel/air mixing and a corresponding reduction in peakflame temperature. The results from the IGT glass tank simulator and the fullscale glass melters show that fuel/air mixing has a significant impact on NOx formation. Factors such as combustion air and natural gas injection velocities, natural gas injection locations and angles, and port geometry all affect the rate as well as the pattern of fuel/air mixing, thereby influencing NOx formation. The data show that NOx formation generally decreases with decreasing fuel-injection velocity, decreasing combustion air velocity, decreasing the angle of impact between the air stream and the fuel jets (for example, switching from side-of-port to underport firing), and increasing the amount of air bypassing the fuel jets. The tests on the glass tank simulator and two commercial glass melters provided data for the impacts of-- • Port design • Firing angle • Fuel-injection velocity. These data were used to develop a generalized correlation between NOx formation and the intensity of mixing, which is defined in References 2 and 8. Using this factor and the correlation, it may be possible to estimate the impacts of fuel/air mixing modifications. Researchers have also found that sealing of the burner blocks can reduce NOx by 300/0 while substitution of cold primary air (no primary air was used in the IGT work) by hot primary air, combined with enlargement of nozzle area to maintain the same velocity, further reduced NOx by 40% (1). (P-002\ 1 1 94Maui.doc |
