| OCR Text |
Show mp0 (2D where mp is average mass of the particle in the control volume, nipo is the initial mass of the panicle, hfg is the latent heat of volatiles evolved, T p is the particle temperature at the exit of the control volume, and Tref is the reference temperature for enthalpy. This heat exchange appears as a source or sink of heat in the continuous phase energy equation during any subsequent iteration. A similar equation governs heat exchange in surface combustion, in which heat of release in surface combustion is incorporated. 3.0 SIMULATION RESULTS AND DISCUSSION The transport equations arising from the various submodels described in section 2are discretized by the finite volume method using a hybrid scheme. The computations were performed using the first/second order accurate schemes in Fluent 5.0 3.1 COMPARISON WITH EXPERIMENTS To test the model, simulations were compared to measurements of unburned carbon, and C02 and 02 concentration profiles made while cofiring biomass and coal in the Multifuel Combustor (MFC) at Sandia National Laboratories. The M F C is a pilot-scale (30 k W ) , 420 cm-high, down-fired, turbulent flow reactor that simulates gas temperature and composition histories experienced by particles in combustion systems. The combustor has a circular cross section with an inner diameter of 15 cm. Electric heaters are used to control the temperature of the walls of the combustor. The M F C was originally designed to simulate the combustion and ash deposition in pulverized-coal-fired boilers in a well-characterized, highly-instrumented environment. A more detailed discussion of the M F C is available in the literature (Baxter and Mitchell, 1992). For these experiments, the coal was Pittsburgh#8 with the composition of 75.77 wt. % C, 5.44 wt. % H, wt. 7.96 % 0 , 1.45 wt. % N, 2.65 wt. % S, and 7.61 wt. % ash. The biomass was switch grass ( S W G ) with the composition of wt. 40.89 % C, 6.09 wt. % H, 42.09 wt. % 0 , 0.76 wt. % N, 0.11 wt. % S, and 7.64 wt. % ash. The heat content for the dry coal and biomass was 13,643 and 7,002 BTU/lb, respectively. The proximate analysis performed in accordance with A S T M gave a volatile yield for the coal and biomass to be 36 % and 6 7 % respectively. It has been well established that the volatile yields under commercially relevant conditions are typically higher than those reported by the standard A S T M test. For these simulations w e used value of 8 5 % for the volatile yield of S W G ; this value was determined in separate experiments conducted under inert conditions in the M F C . The coal particles were utility ground with a mean diameter of 60 pirn (range 20-150 nm). The switchgrass was prepared using a micropulverizer to a nominal particle size of 1 m m . Because of the fibrous nature of biomass, a typical biomass particle is not spherical. For this work, we approximated the particles as ellipsoidal with a minor axis range of 0.053 m m to 1.42 m m and a mean aspect ratio of 3. The fuels were blended and well-mixed before injection into the furnace. Comparisons of the measured and predicted unburned carbon are shown in Table 1. The burnout is expressed as a mass ratio of the collected char at the exit of the combustor to the unburned fuel. Measurements are shown for unblended fuels and for three different fuel blends. There is relatively good agreement between the measurements and the model predictions. Interestingly, despite the larger biomass particle size relative to coal, blending biomass reduces the unburned carbon. W e attribute this phenomenon to the high volatile content of the S W G . The model predicts this trend, however if the biomass particle size is too big to devolatilize inside the combustor or has a short residence time for the particle to go any significant char oxidation after the devolatilization, then the trend may reverse, which was seen in the C E R F exploratory simulations described in the next section. Q = ^ r _ +-^- m Po m Po -K +/V™/ + \c,/n |
