OCR Text |
Show temperature across the staging point wa s approximately 1600C. A detailed description of the experimental facility is available elsewhere. 16 Reducing agents were injected at various locat i ons (fuel rich or with the burnout air) either as gases, or solids. The solid reducing agents (ammonium sulfate, cyanuric acid, and urea) were mixed with pulverized limestone (15 percent additive by weight) to facilitate feeding. These were metered with a variable speed screw feeder and transported with nitrogen. The additive particle sizes were: ammonium sulfate - 180 microns, cyanuric acid - 130 microns and urea - 70 microns. The nitrogen carrier flow was 48 liters/minute, which resulted in a temperature drop of 1600C at the additive injection point. The ammonia gas was also mixed with an equivalent amount of nitrogen, and all reducing agents were added with the radial injector. Exhaust gas samples were withdrawn by a stainless steel, water jacketed probe at the 6000C point. These were analyzed for NO (chemiluminescence), 02 (paramagnetic), CO/C02 (NDIR), amines (NH3 electrode) and HCN (specific ion electrode). Furnace thermal profiles were characterized with a suction pyrometer. KINETIC MODELING Detailed chemical kinetic modeling of the experiments was performed using a plug-flow/stirred reactor algorithm. 17 This program has the necessary advantages of allowing arbitrary specification of temperature profile as well as the ability to accept cross stream reagent addition along the length of the reactor. The reactor was modeled by use of a 1.0 millisecond residence time stirred reactor ignition zone. This was followed by a plug-flow reactor upon which the experimentally observed 2250 C/s temperature decline was imposed. A sufficient amount of NO was introduced into the stirred reactor to provide the desired NO concentration in the post-flame gas. Additives and burnout air were introduced as sidestreams over 10.0 millisecond mixin8 times. At the highest temperature used for additive injection, 1200 C, the gas composition had reached equilibrium, with the exception of NO which was essentially frozen at the targeted value. Thus, the results of the calculations were not sensitive to the upstream reactor configuration or conditions. Table 1 lists the reactions considered. This table shows a portion of a more comprehensive mechanism that was assembled primarily to model the reburning process. Adjustments to the literature values were applied to two of the rate constants. For Reaction 26, Zellner25 recommends a rate constant in non-Arrhenius form. Since this form was not compatible with the present kinetics code, a modified Arrhenius expression was derived which agrees with Zellner's rate to within 5 percent over the temperature range of 1000-2500 K. The rate for Reaction 84 is shown as half the value reported by Glarborg. 18 This adjustment was necessary to obtain agreement with ammonig and HCN measured at fuel rich stoichiometries under reburning conditions; however, this adjustment was within the error bounds of the original assessment. 3 |