OCR Text |
Show 16 3. For underport firing, only 2000 SCFIh of oxygen (representing about 3% of total furnace stoichiometric requirement) was sufficient to achieve over 500/0 NOx reduction for the enriched hot air case, while 7000 SCF/h of oxygen (representing about 12% of the total furnace stoichiometric requirement) was needed in the case of straight oxygen. Figures 8 and 9 compare the NOx and CO levels measured in the stack with the respective levels measured in the top of the regenerator. The NOx data was within ±20% and does not show any trend, that is, no NOx was formed or destroyed in the regenerator. The variance was probably caused by the following: 1. The dynamic nature of the glass tank produces somewhat different NOx levels during different cycles. The NOx levels in the top of the regenerator and in the stack generally represent different cycles. 2. Point measurements made in the top of the regenerator and in the stack do not accurately represent average values; however, because temperature and emissions change continuously during the reversals, as well as between different reversals, it is neither practical nor very useful to perform cross sectional traverses for each test. Based on this information, it appears that the agreement between the two measurement locations in fairly good. The CO data in Figure 9 indicates that even a small amount of oxygen in the top of the regenerator is sufficient to burnout the CO at the relatively high exhaust gas temperatures (2750° to 2850°F) that are typical in these types of furnaces. Even 5000 vppm CO in the top of the regenerator was reduced to below 150 vppm in the stack. It must be noted, however, that high CO levels are not desirable because of the increased regenerator temperatures which could increase stack opacity and combustion air temperatures and adversely impact regenerator life. An increase in combustion air temperatures, as shown earlier, would be counter-productive in that it will increase the NOx formulation in the primary combustion zone. Figure 10 shows the effect of overall stoichiometric ratio on NOx levels for both the baseline and the staging operation. The data show that while NOx decreases with decreasing overall stoichiometric ratio for baseline operation, as would be expected, this is not true of the staging operation. This effect is also expected, however, because with staging, NOx formation would depend primarily on the primary zone stoichiometry. Figure 11 shows the impact of overall stoichiometric ratio on the CO levels in the stack. The data illustrate that the CO levels decrease with increasing overall stoichiometric ratio because of the increased oxygen availability for CO burnout. The figure also shows that CO levels remained below 150 vppm even at very low excess air levels and that 50/0 excess air in the top of the regenerator was sufficient to bring the stack CO levels to below 25 vppm. Computational Modeling This section summarizes the results of the Computational Fluid Dynamics (CFD) modeling studies conducted by Air Products and Chemicals, Inc., after the field tests for the optimization of the OEAS technology. The modeling of the combustion space was done using Fluent V.3.03. The grid used for this work consisted of about 140,000 nodes. (P-002\1194Maui .doc |