| OCR Text |
Show present unsteady simulations. However, the further preservation of this zero flux on an instantaneous basis is a well-known limitation of any axisymmetric simulation. The nature of this approximation is consistent with the axisymmetric swirl flow representation of the tangentially-fired boiler, but must be-kept in mind when interpreting results from the simulations. When many instantaneous coal gas mixture fraction fields C,(\,t) of the type in Fig. 6 are averaged over several boiler residence times, they produce an average field that can be compared with that suggested by the 02 fields obtained from ESA's furnace probing measurements. The results of such a process show that the LIM simulations predict relatively high 02 concentrations along the boiler walls, with significantly lower and relatively uniform 02 concentrations throughout the interior of the furnace. This can be seen from the relatively dark regions near the boiler walls in Fig. 6. As noted above, the high oxygen concentrations indicated by the dark region near the furnace centerline is a manifestation of the zero flux condition. The observed presence of high oxygen concentrations near the boiler walls is in good agreement with ESA's probing measurements and agrees with ESA's conventional modeling results. Based on the role of oxygen in the gas reburn kinetics as summarized in §1, this suggests that one constraint on the gas injection system is that it should be designed to avoid placing gas near the boiler walls. 4.3 Identification of injection elevations Simulations were conducted to determine the emissions performance expected from gas injection at various elevations in the boiler, and from these to select one or more elevations for the FLGR system design. Based on ESA's previous field experience with FLGR technology and the elevations at which the boiler could be readily accessed, LIM simulations of the gas injection and mixing processes were conducted by NGB Technologies for gas injection at 704-ft, 714-ft, and 722-ft boiler elevations. In these simulations, additional vorticity surfaces are introduced at the edges of the gas injectors, and scalar surfaces at the center of the injectors, with the vortex sheet strengths and initial scalar values set by the injector size and inflow conditions (see §4.1). The resulting instantaneous injected gas mixture fraction fields C,(\,t) and mixing rate fields V^V^(x,r)at six typical instants of time are shown in Figs. 8-10. The injected gas stream is rapidly broken up by its own turbulence, as well as by the turbulence produced from the coal and air inflows to the boiler (see Figs. 6 and 7). For the mixture fraction fields in each of these figures, colors ranging from dark blue through red again denote linearly increasing mixture fraction values. For the mixing rate fields, colors ranging from dark blue through red denote logarithmically increasing mixing rates. The axisymmetric nature of the simulations must be kept in mind when interpreting these results. Among the key points to note in Figs. 8-10 are the extent of mixing between the injected gas and the furnace gas, and the gas mixture fraction values entering the convective section, since high values correspond to relatively large 00 concentrations in the reburn chemistry. Figure 11 shows the resulting NOx and 00 emissions performance obtained for each of these injection locations at various gas input levels. The results shown correspond to full boiler load (140 M W ) with 2.5% boiler 02, and six injectors sized to give 10% maximum gas heat input at 35 psig gas |
