OCR Text |
Show NOx PORT REQUIREMENTS As a first step in the system design, the maximum limitation on throat stoichiometry required to obtain the NSPS of .3 Ib/MMBTU total NOx was determined. This limitation is calculated by first arriving at a "baseline" NOx estimate, which is the sum total of the thermal and fuel NOx components obtained under nonstaged conditions. Percent NOx reduction from baseline as a function of reduced throat stoichiometry is then empirically correlated based on available inhouse data. Approximation of the thermal NOx component is calculated as a function of adiabatic flame temperature and the rate of temperature weighted heat release into the furnace heat absorbing surfaces. Fuel NOx is approximated as a function of fuel bound Nitrogen content. For the given fuel analysis, furnace geometry, and operating conditions specified for this application, it was estimated that a maximum throat stoichiometry of between .9 and .95 would be required to meet NSPS. With a total specified excess air level of 7 percent, approximately 13 percent of the total combustion air stream would enter the furnace as tertiary air, with mixing being delayed until after fuel oil droplet evaporation. In order to ensure adequate fuel burnout without sacrifice to staged combustion NOx control, the downstream location of the ports with-respect-to the burner guns became a critical design parameter: The "optimum" distance would be long enough to delay mixing just past the point of oil droplet evaporation and flammable air entrainment, but would be compact enough to minimize the corresponding reduction in high temperature residence time required for CO burnout. Droplet burnout time was initially calculated based on a 150 micron particle, which corresponds with the + 95 percent less point on a rosin-rammler curve for an Coen MV gun operating at .15 #stm/#fuel. Based on a droplet temperature of 2500 F, a burnout time of under .1 seconds was calculated. This information, when used in conjunction with an estimated droplet velocity profile, dictated a port location of 9 feet downstream of the burner throat exit (roughly half of the furnace length). This initial location was then used as an input in the Base Case CFD model, discussed below. COMPUTER MODELLING In order to help optimize system design parameters compatible with maximum fuel burnout, a commercially available computational fluid dynamics code was utilized: Initially, a Base Case computer model was established, converged, and analyzed. Sensitivity cases, whereby specific parameters were altered, were then developed off of the Base Case. By analyzing the converged sensitivity cases, the relative effect of specific design parameters on system performance was quantified. This information was then used in the actual equipment design. OVERVIEW OF CODE All computer modelling was done with Version 2.99 of the FLUENT code. The code was run on a Sun Microsystems 4/110 minicomputer, utilizing Version 4.0 of the SUNOS operating system (UNIX based). Although the theoretical basis of the code is well documented elsewhere (Ref.1-3), a brief overview is given below for completion: The code employs a finite number (user specified) of discrete control volumes over the computational grid to obtain a 3-D steady-state solution of the governing equations of continuity, momentum, and energy (Navier-Stokes). After integrating the governing equations across the control volumes, velocity & pressure coupling are resolved via the SIMPLE algorithm. For this application, the standard k-epsilon model was used to account for the effects of turbulence (an algebraic stress model can also be used, although this tends to increase the time of convergence). |