; YO2)( S =)P( 

)×P(YO2)d 

 dYO2: (32)For the methane–air turbulent combustion, an axi-symmetrical density of the gas mixture is calculated accord-ing to the ideal gas-state equation. The boundary conditionsare that a uniform distribution condition, a no-slip condition,a fully developed 4ow condition and an axi-symmetricalcondition are taken at the inlet, wall, exit and axis,respectively. and 0:8932 m3h−1, respectively. In order to simulate fuelNO formation in methane–air swirling combusting 4ows,NH3 at a 4ux that is 4.91% of methane is injected intothe methane. The computational domain is a half of thecombustor and 35 × 25 × 8 grid nodes are used inside thedomain. The geometrical sizes of the burner and combustorare given in Table 3.As can be seen from Fig. 6, the calculated temperaturesare close to experimental data. In most of the pro9les, thecalculated results are in good agreement with the measure-ments. Fig. 7 gives the fuel NO concentration pro9les. It canbe seen that the trend in the calculated results is the sameas that measured. However, quantitatively the calculated re-sults are a little higher thanmeasurements in the pro9les withx=D1 640. The inaccuracy is probably caused by measure-ment error or the simpli9ed NO model. Due to the in4uenceof swirling, NO concentration is high in the near-wall re-gion. Therefore, the computational program based on SOMNO formation model can be used to simulate swirling com-bustion and fuel NO formation.Table 4Geometrical sizes (Unit: mm)D1 D2 D3 D4 D5 D6 L14 15 22 23 25 100 4003.2. Simulation of pulverized-coal swirling combustionFig. 8 gives the geometrical con9guration of the burnerand pulverized-coal concentrator for calculated swirlingcombustion. It has three inlets: primary-air inlet, swirlingsecondary-air inlet and non-swirling secondary-air inlet.When a pulverized-coal concentrator designed by authors isinstalled in the primary-air tube, the burner is called a biascoal swirl burner (Case 1).When there is no pulverized-coalconcentrator in the primary-air tube, the burner is called adouble air register swirl burner (Case 2). The computationaldomain is a quarter of the burner and 16 × 14 × 11 gridnodes are used inside the domain. The geometrical sizesand 4ow parameters of the burner are given in Tables 4 and5. The coal proximate analysis is given in Table 6. For theboundary condition, except that a non-uniform distributionpulverized-coal concentration is used at the primary-air in-let, uniform distribution conditions for the other variablesare used at the inlet.
The fully developed 4ow condition fortwo phases is adopted as the outlet condition. At the wall,no-slip condition is used for the gas-phase. Zero normalvelocity, zero mass 4ux, zero gradients of longitudinal andtangential velocities are used for the particle phase. At theaxis, symmetrical conditions are adopted for both phases.The wall temperature is 1270 K.Fig. 9 gives the particle temperature maps. It can be seenthat the particle temperature in Case 1 is higher than in Case2. So, Case 1 is favorable for igniting of coal. It impliesthat rapid devolatilization occurs at the exit of the burnerin Case 1. As seen from Fig. 10, the particle concentrationdistribution during coal combustion in Case 1 is lower thanthat everywhere in the combustor in Case 2. This is becausea high combusting temperature in Case 1 is favorable for thecombustion of pulverized coal. The low particle temperaturein Case 2 is because only a small portion of the char beginsto burn in the combustor.Figs. 11 and 12 show CH4 and O2 concentration dis-tributions respectively. The concentrations of CH4 and O2in Case 1 are much lower than those everywhere in Case2. This indicates that the combustion in Case 1 is moreintensive and complete than that in Case 2. The fast ignitionand complete combustion of volatile can lead to good 4amestabilization.The calculated thermal NO concentration maps given inFig. 13 show that the thermal NO concentration in Case1 is a little higher than that in Case 2. As a whole, ther-mal NO concentration is low in the combustor where thelargest value is 93 ppm in Case 1. Fig. 14 gives the total NO concentration. It can be seen that the trend for total NO for-mation in Case 1 is an increase 9rst and then a decrease,while in Case 2 it is increasing monotonically. This indi-cates that the coal combustion is not complete and fuel NOformation is not completed in Case 2. At the outlet of thecombustor, the total NO concentration in Case 1 is muchlower than in Case 2. Also, the total NO formation impliesthat the fuel NO formation is larger than thermal NO for-mation, and NOX formation during the coal combustion iscontrolled by the mechanism of fuel NO formation. Thisconclusion is in agreement with some previous researches.Through comparison of the simulations, it can be seen thata bias coal swirl burner (Case 1) is favorable for increasingthe combustion rate and decreasing fuel NOX formation.This is because the fuel-rich mode is realized in the central recirculation zone at the exit of the burner by the action ofthe pulverized-coal concentrator (Li et al., 1999).4. Conclusions1. The modi9ed k–  model proposed in this paper can givemore reasonable 4ow 9eld prediction than the standardk −   model.2. A second-order-moment (SOM) reaction rate model forNOX formation in turbulent 4ows is proposed. It cangive more reasonable results of thermal NOX than thepresumed PDF 9nite-reaction-rate model. It also givesreasonable results of fuel NOX . And the SOM modelneeds less computation time than the PDF model.3. The calculated results of swirling pulverized-coal com-bustion indicate that the pulverized-coal concentrator de-signed by authors has a strong e>ect on coal combustionand NOX formation. It increases the combustion rate ofcoal and decreases fuel NO formation. Its validation stillneeds to be checked up by hot experiment. 煤粉燃烧和NO形成的数值模拟英文文献和中文翻译(5):http://www.751com.cn/fanyi/lunwen_43992.html