OCR Text |
Show In addition to incorporating the effect of CO into the SNCR model, the effect of initial NOx levels was also considered. The effect of inlet NO levels on the effectiveness of SNCR has been demonstrated previously.22 In general, as the inlet NO levels increase the effectiveness of SNCR, as measured by reduction efficiency or any comparable measure, increases. This is due to the self-sustaining nature of SNCR chemistry which cannot be accounted for in a simplified, two-step model. However, the overall effect of inlet NOx levels can be calculated based upon the fully detailed chemical kinetic set and has been incorporated into the simplified chemical kinetic model as an additional parameter (an estimate of the initial NO). In the turbulent flow model, the overall SNCR rates are modified slightly depending upon this estimate of initial NO. Results Experiments were perfonned by injecting the reagent stream in various manners and into various ports of the furnace, presented in Figure 1. In all cases the product stream entering the SNCR section contained 2.5% O2, 8.4% CO2, 16.7% H20, and 500 ppm NO in a balance of nitrogen. The data presented in Figure 4 were obtained by radially injecting ammonia (rapid mixing) into an approximately isothennal region of the furnace with the product stream temperature above 1100K for more than 400 msec. As reported previously, the exhaust NO level depends on the inlet NH3INO ratio. Under rapid mixing and almost isothennal conditions low exhaust NO levels are possible for NH3INO ratios in excess of 1.5 and exhaust NO is almost independent of the initial NO concentration. The full chemical kinetic model (SENKIN), applied for homogeneous isothennal conditions, gives this same trend. The magnitude of NO reduction is over-predicted as would be expected due to mixing and thennal quenching limitations in the experiment. For these conditions, NO levels are independent of inlet NO at high NH3INO due to initiation reactions that build up the radical pool which are dependent on initial concentrations. The chemical kinetics capture this effect but are limited in their application because of the assumption of homogeneity. However, the obvious influence of chemical kinetics on the data implies that chemical kinetic considerations cannot be neglected even under mixing limited situations. The data of Figure 4 indicate that if a combustion system were designed specifically for the application of SNCR with an integral isothennal cavity at the opt~um temperature for selective reduction, it might not be necessary to optimize the main heat release for minimum NO fonnation in order to achieve low exhaust levels. Since fmal NO levels are independent of the initial concentration the only benefit for reduced inlet NO is a reduction in reagent costs. As indicated earlier the application of SNCR to existing equipment requires skillful design of the injection system to mix the reagent with the entire product stream and to take account of thennal quenching and the presence of trace species in the combustion products. Some groups have . developed models that attempt to predict the effects of turbulent mixing and chemical reaction separately 21. As will be shown below these effects cannot be treated separately as there are significant interactions. Thenna) Quenching The need to inject the selective reducing agent at the optimum temperature to account for mixing under the applicable thennal quench rate has long been recognized. Figure 5 quantifies Page 5 |