Journal of Mechanical Science and Technology 26 (11) (2012) 3723~3731 www.springerlink.com/content/1738-494x
DOI 10.1007/s12206-012-0848-y
Numerical investigation of NOx reduction in a sudden-expansion combustor with inclined turbulent air jet† S. A. Hashemi, A. Fattahi*, G. A. Sheikhzadeh, N. Hajialigol and M. Nikfar Department of Mechanical Engineering and Energy Research Institute, University of Kashan, Kashan, 87317-51167, Iran (Manuscript Received December 16, 2011; Revised July 1, 2012; Accepted July 7, 2012) ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Abstract Axisymmetric sudden-expansion geometry of a co-flowing methane-air diffusion flame is considered to investigate the effect of air inlet conditions on NOx formation, flow field and temperature distribution using the k-ε turbulence and β-PDF combustion model. The predicted results are in acceptable agreement with the published experimental and numerical data. The obtained results show that increasing air turbulence intensity results in considerable decrease in NO formation. Increasing the inlet angle of the air causes the NO formation to decrease due to raising vorticity strength. As a new index, the mass-averaged integral of vorticity magnitude is introduced to investigate the effect of altering inlet angle of the air on the flow field. Keywords: Co-flowing methane flame; Turbulent diffusion flames; Turbulence intensity; Inclination angle of air inlet; NO reduction ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
1. Introduction Combustion processes are mainly governed by coupled physical and chemical phenomena. In non-premixed combustion systems, air conditions can be considered as a tool for control and better performance, such as higher efficiency and lower emissions. In order to achieve the optimum conditions, some detailed studies on the above cases are necessary. Such studies can be done using a CFD-based analysis providing perfect recognition and in-depth surveys of the flow in the reacting processes. These analyses, however, are more economical than the experiments. Control of NOx emissions in the combustion process has become an important criterion that is achieved by changing air conditions in a combustor. Chen and Chang [1] modeled NO formation in a turbulent, non-premixed jet flame. They used the joint scalar probability density function (PDF) approach and the traditional flamelet model. They studied various effects, such as flame interaction and radiative heat transfer on NO formation by comparing the predictions and the experimental data. They reported that radiative heat transfer loss affects NO predictions in the far field. Lallemant et al. [2] showed that NOx emission is increased when increasing stoichiometric boundary conditions. They also indicated that increasing N2 in the fuel and in the air by 0-19%, NOx emission is increased. Increasing of N2 in the air by 19-47%, decreases *
Corresponding author. Tel.: +983615912443, Fax.: +98 3615912475 E-mail address:
[email protected] † Recommended by Associate Editor Oh Chae Kwon © KSME & Springer 2012
NOx. Furthermore, they also observed that NOx emission is decreased when increasing the momentum of injection. Ilbas et al. [3] simulated a diffusion turbulent flame with methane and some types of methane-hydrogen mixture as fuel. They showed that the addition of methane to hydrogen decreases the flame temperature and thus decreases NOx emissions significantly. They demonstrated that air staging causes rich and lean combustion regions and results in lower NOx emissions at the combustor exit. Dally et al. [4] investigated a variety of fuel mixtures: methane, ethylene, and propane. They found that dilution of fuel with CO2 causes reduction in NOx emission and made the flame inside the furnace invisible. This dilution caused the stoichiometric mixture fraction to shift toward the rich side where the highest scalar dissipation is present. Lopez-Parra [5, 6] numerically simulated NOx prediction in gaseous flames. They used several numerical approaches to model the formation of thermal NOx and prompt NOx of various diffusion flames. Bousheki et al. [7] experimentally showed that separation of air jets leads to better flame stability and lower production of NOx. When the distance of the air jets decreases, the production of the pollutants such as NOx is considerably reduced. Lopez-Parra and Turan [8] investigated the effect of fuel inlet perturbations on NOx emissions using k-ε turbulence and EDM models. They concluded that the NOx emissions are reduced by imposing such perturbations. Kim et al. [9] experimentally showed that when the air inlet velocity in an oxy-fuel combustor is increased, the flame zone and NO level are decreased. Saqr et al. [10] studied the effect of free stream turbulence on the structure of
3724
S. A. Hashemi et al. / Journal of Mechanical Science and Technology 26 (11) (2012) 3723~3731
CH4-air flame. They presented some information on the dependency of the flame surface on the free stream turbulence compared to its dependence on the reaction-induced turbulence. Recently, numerical investigations of NOx formation and reduction have been carried out in different model combustors [11-13]. The current study aims to investigate the effect of inlet air conditions including air stream turbulence and inclined flow on the NOx, flow field and temperature distribution in a sudden-expansion geometry of a co-flowing methane-air diffusion flame using the k-ε turbulence and β-PDF combustion models. The NOx prediction models employed in this study have been thoroughly validated for the same type of flame in previous works [5, 6]. The results of this study can provide good knowledge to better control the pollutants in such combustors.
The turbulent eddy viscosity is introduced as:
µt = ρ Cµ
k2
(4)
ε
where Cµ is defined as: Cµ =
1 A0 + As (kU * ε )
.
(5)
Finally, the constants are: C1 = 1.44, C2 = 1.9, σk = 1.0 and σε = 1.2. The value of the constants A0 and As and more details are presented in Ref. [15]. Turbulence intensity is expressed as:
2. Governing equations Conservative equations for a steady state reacting flow are used here. A generalized equation describes overall mass, momentum, energy and chemical species concentration and is written as [14]:
∑ i
∂ ( ρViφ ) = ∂xi
∑ i
∂ ∂φ (Γφ ) + Sφ ∂xi ∂xi
(1)
3. Turbulence model The realizable k-ε turbulence model [15] is applied in this study. In comparison with the standard k-ε, this model hinders the negation of values of the normal Reynolds stresses by employing a new turbulent eddy viscosity formulation including a variable Cµ. It also utilizes a new transport equation for dissipation rate, which hires different sink and source. Because of the disadvantages of the standard k-ε turbulence model, some researchers proposed the realizable model for this type of flame [16, 8]. The k and ε transport equations of this model are written as follows:
(2)
+Gk + Gb − ρε − YM
ε2 ε − ρ C2 + C1ε C3ε Gb k k + νε
Because of the fluctuating characteristics of the turbulent mixing process, the probability density function is a skilled method for the cases including combustion process and turbulent flow. In this study, we employ the presumed PDF model. In this model, two parameters of the mean and its variance of scalar quantity define the PDF. Due to better results for the turbulent reacting flow in comparison with the other PDF models, the β-PDF model is used to calculate the thermodynamic properties [17-19]. In the presumed β-PDF model, because of difficulty in solving the transport equation for each species, the mixture fraction, f, is defined in terms of mass fraction of specie i, Yi: f =
Yi − Yi ,ox Yi , f − Yi,ox
.
(3)
(7)
where the subscripts “f” and “ox” define the fuel and oxidant (air) species, respectively. The transport equations of mean mixture fraction, f and its variance, f ′2 , are: ∂ ∂ ∂ µt ∂f (ρ f ) + ( ρu j f ) = ( ) ∂t ∂x j ∂x j σ t ∂x j
µ ∂k ∂ ∂ ( ρ kui ) = [( µ + t ) ] σ k ∂xi ∂xi ∂xi
(6)
4. Turbulence-combustion interaction
where φ is the general term. Eq. (1) yields conservation equation of mass, momentum, energy and species mass fraction when φ is one, velocity component, enthalpy and mass fraction, respectively. Sφ is the source term in the conservation equations. For brevity, the individual equations are not presented here.
µ ∂ε ∂ ∂ ( ρε ui ) = [( µ + t ) ] ∂xi ∂xi σ ε ∂xi
2 k V′ 3 I= = . 2 Vavr u + v2
(8)
∂ ∂ ( ρ f ′2 ) + ( ρ u j f ′2 ) ∂t ∂x j ∂ µt ∂ f ′2 ∂f ε ( ) + C g µt ( ) − Cd ρ f ′ 2 = ∂x j σ t ∂x j ∂x j k
(9)
in which the constant values of σt, Cg and Cd are 0.85, 2.86 and
S. A. Hashemi et al. / Journal of Mechanical Science and Technology 26 (11) (2012) 3723~3731
3725
in which I, L, Uref are defined as turbulence intensity, characteristic length, and inlet velocity, respectively. ℓ is predicted by [14]: l = 0.07L . Fig. 1. Combustor configuration.
2.0, respectively. The relation between the obtained timeaveraged values from the above equations and the instantaneous mixture fraction is established by a PDF. This function is written as p(f), which demonstrates the probability that the fluid is in the vicinity of state f. The method applies the mean values of species concentration and temperature. The mean mass fraction of species and temperature, φi , is calculated from:
φi =
∫
1
0
p ( f )φi ( f ) df
where p ( f ) =
(10)
f α −1(1 − f ) β −1
∫
f α −1 (1 − f ) β −1 df
; α and β are defined as
follows:
α= f[
f (1 − f ) f ′2
β = (1 − f )[
− 1]
f (1 − f ) f ′2
(11) − 1] .
(12)
For the non-adiabatic case, the mean enthalpy transport equation is described as: r k ∂ ( ρ h) + ∇.( ρ v h) = ∇.( t ∇ h) . ∂t cp
(13)
Chemical equilibrium is used for determining product mole fractions.
5. Geometry and boundary conditions Fig. 1 shows the axisymmetric geometry of the current combustor model. According to the experimental setup [20], the fuel (pure methane) radius (Rfuel) and the air radius (Rair) of the coaxial burner are 29.5 mm and 46.5 mm, respectively. The length, L, of the combustor is 1.7 m and the diameter, D is 122.3 mm. The mass flow rate is 7.2 g/s in fuel stream and 137 g/s in air stream. The inlet turbulence kinetic energy and its dissipation rate are computed as [14]: k=
k1.5 3 (U ref I )2 ε = 0.16 l 2
(15)
The flow leaves the combustor with absolute pressure of 1 bar. No slip condition is adopted at the solid wall, while the standard wall function is applied to compute the tangential velocity near the wall. At the combustor centerline, radial velocity component is zero, and the radial gradients of the other quantities become zero due to symmetry conditions. For the radiation condition, the walls are assumed as a gray heat sink of emissivity 0.7.
6. Numerical procedure The governing nonlinear equations together with the boundary conditions were solved by iterative numerical approach using the finite volume method [14] and a second order upwind scheme to discretize the equation terms. In order to couple the velocity field and pressure in the momentum equations, the well-known SIMPLE1 algorithm was adopted. In order to prevent divergence, appropriate under relaxation factors are employed. The convergence criterion of the numerical method was chosen as the total normalized residuals being less than 10-6. The grid is denser near the annular inlet zone due to the mixing and reaction process. In the present study, it is found that the grid size of 37000 cells for the geometry ensures a grid independent solution.
7. NOx formation NOx formation is an important topic in combustion because of its considerable contribution to air pollution. NO is the most important species in the NOx emission for many types of flames [21]. Due to smallness of the NO mole fraction, the NOx formation process does not have an important effect on the flow field; hence, it is post-processed from the simulation [22]. To include NO formation for this type of flame, thermal and prompt mechanisms are employed [23] and calculated by finite rate chemistry. For these two mechanisms, only the following NO species transport equation is required. r ∂ ( ρYNO ) + ∇.( ρ vYNO ) ∂t = ∇.( ρ D∇YNO ) + S NO
(16)
in which YNO, D and SNO are mass fraction, effective diffusion and the source term, respectively. SNO can be calculated as follows:
(14) 1
Semi-implicit method for pressure-linked equations
3726
S. A. Hashemi et al. / Journal of Mechanical Science and Technology 26 (11) (2012) 3723~3731 30
d NO S NO = M w, NO ( ) dt
(17)
Nunerical results Experimental data [20]
25
where Mw,NO is the molecular weight of NO and d NO / dt is calculated by both thermal and prompt mechanisms. The thermal NO formation rate is determined according to the highly temperature-dependent reactions referred as the extended Zeldovich mechanism:
Axial velocity (m/s)
20 15 10 5 0 -5
N 2 + O ⇔ NO + N O2 + N ⇔ NO + O N + OH ⇔ NO + H .
(18)
-10
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
2r/D
Fig. 2. Axial velocity distribution at z/D = 0.052: Comparison of the present results with the experimental data [20].
Assuming a quasi-steady-state for concentration of nitrogen atoms, the thermal NO formation rate becomes:
30 Nunerical results Experimental data [20]
25
2
dt
(1 − = 2k1 O N 2
(1 +
k−1k−2 NO ) k1 N 2 k2 O2 k−1 NO
k2 O2 + k3 OH
20
(19) )
Axial velocity (m/s)
d NO
15 10 5 0
in which the reaction rate coefficients are: k1 = 1.8∗108exp(-38370/T)[m3/kmol-s], k-1 = 3.8*107exp (-425/T), k2 = 1.8*104Texp(-4680/T), k-2 = 3.81*103Texp (-20820/T), and k3 = 7.1*107exp(-450/T) [23]. According to partial equilibrium calculation, the concentration of O atoms is achieved as follows [24]: 0.5
0.5 O = 36.64T O2
exp( −27123/ T ) .
(20)
The prompt NO formation rate is depicted by the following equation [25]: d NO
dt
Ea α = fc k prompt O2 N 2 fuel exp( RT )
(21)
in which α is the order of reaction and fc is a correction factor which depends on the fuel type and fuel air ratio. The values of kprompt and E are 6.4*106 and 72500 cal/gr, respectively.
8. Model validation In order to validate the numerical model, the present solution is compared to the experimental data [20] and numerical results of Eq. [26]. Fig. 2 and Fig. 3 show the comparison for axial velocity at Z/D = 0.052 and 0.146, respectively. The figures show that the present models predict velocity very precisely at lower r/D. The maximum deviation from the reference data is seen in the vicinity of the walls. Altogether, the present predicted values and the experimental data are in satisfactory agreement both qualitatively and quantitatively. At Z/D = 1.73, the predicted, experimental and numerical results demonstrate that the location of maximum temperature are 46
-5 -10
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
2r/D
Fig. 3. Axial velocity distribution at z/D = 0.146: Comparison of the present results with the experimental data [20].
mm, 37 mm and 67 mm, respectively. While the predicted, experimental and numerical values of maximum temperature are 2013K, 2009K and 2007K, respectively. These values for Z/D = 1.99 are 2045K, 2036K and 2059K while the location of maximum temperature are 48 mm, 38 mm and 68 mm, respectively. Fig. 4 and Fig. 5 represent the comparison between three sets of temperature results at Z/D = 1.73 and 1.99, respectively. The present models predict the temperature trend, the maximum temperature and its location more precisely than that of Ref. [26]. The models predict the temperature behavior more precisely in the vicinity of the centerline for lower Z. Furthermore, the models predict temperature field more accurately in the vicinity of the wall for larger Z. As can be seen, the present results are in closer agreement with the experimental data in comparison with results of Ref. [26]. For further comparison, the predicted average temperature and NOx emissions at the exit of the combustor are compared to the numerical results of Ilbas et al. [27] in Table 1. The fuel used in Ref. [27] was pure methane and the combustor geometry was similar to the current combustor except in dimensions (L = 2 m, radius of combustor Rc = 0.3 m, radius of fuel inlet Rf = 0.005 m, radius of air inlet, ra = 0.15 m and temperature of both fuel and air are 300K). The two sets of results are in close agreement. The differences are due to using different models for turbulence and combustion.
3727
S. A. Hashemi et al. / Journal of Mechanical Science and Technology 26 (11) (2012) 3723~3731 120
Table 1. Average temperature and NOx emissions: Comparison between the present results and those of Ref. [27]. Present results
Temperature (K)
1311
1387
NOx (ppm)
553
601
2100
ATI=15% ATI=30% ATI=45%
100
NO Concentration (ppm)
Results of Ref. [27]
110
90 80 70 60 50 40 30 20
2000
10
1900 0
Temperature (K)
1800
0.1
0.3
0.5
0.7
0.9
1.1
1.3
1.5
1. 7
1.1
1.3
1.5
1.7
1.1
1.3
1.5
1. 7
1.1
1.3
1.5
1.7
z (m)
1700
(a)
1600 1500 Present results Experimental data [20] Numerical results [26]
1400 1300
120 110
ATI=15% ATI=30% ATI=45%
1100 1000 900
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
2r/D
Fig. 4. Temperature distribution at z/D = 1.73: Comparison of the present results with the experimental data [20] and numerical results [26].
NO Concentration (ppm)
100
1200
90 80 70 60 50 40 30 20 10
2100
0
0.1
0.3
0.5
0.7
0.9
z (m)
2000
(b)
1900
100
1700
1500
Present results Experimental data [20] Numerical results [26]
1400
80
1300 1200 1100 1000 900
ATI=15% ATI=30% ATI=45%
90
1600
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
NO Concentration (ppm)
Temperature (K)
1800
70 60 50 40 30 20
2r/D
10
Fig. 5. Temperature distribution at z/D = 1.99: Comparison of the present results with the experimental data [20] and numerical results [26].
0
0.1
0.3
0.5
0.7
0.9
z (m)
(c)
9. Results and discussion NO Concentration (ppm)
In this section, the effect of changing inlet oxidant condition (i.e. turbulence intensity and inclination of inlet air) on the combustor characteristics is studied. Although the experiments are widely used in studies pertaining to oxy-fuel flame combustion, they cannot or can rarely obtain all the flow aspects, whereas a validated numerical simulation can appropriately predict all needed aspects of an air-fuel flame.
80 ATI=15% ATI=30% ATI=45%
70 60 50 40 30 20 10 0
0.1
0.3
0.5
0.7
0.9
z (m)
9.1 Effect of air turbulence intensity The axial NO concentration at different radial distances is reported, as shown in Fig. 6(a)-(d). It is obvious that NO concentration significantly decreases with increasing air turbulence intensity (ATI). The decrease in NO concentration is noticeable when ATI increases from 15% to 30%. The maximum NO concentration decreases more than 70% for r/R = 0 and more than 60% for r/R = 0.6. The least decrease occurs when ATI increases from 30% to 45%. NO
(d) Fig. 6. Axial NO concentration for various ATIs at (a) r/R = 0.0; (b) r/R = 0.1; (c) r/R = 0.3; (d) r/R = 0.6.
reduction approaches to about 60% for both r/R = 0 and 0.6 locations. The location of maximum NO concentration tends to the combustor entrance with increasing r/R. At r/R = 0.6, the maximum NO locates at z = 0. It can be concluded from Fig. 6(a)-(d) that reduction in NO concentration is propor-
3728
S. A. Hashemi et al. / Journal of Mechanical Science and Technology 26 (11) (2012) 3723~3731
9.2 Effect of inclination angle of air inlet
(a)
(b) Fig. 7. Streamline plots for (a) ATI = 15%; (b) ATI = 45%.
(a)
Effect of inclination angle of air inlet on flow field is studied in this section. The inclination angle is considered relative to the centerline. The inclination angle of air inlet (IAA) is varied between 0 and 30 degrees. The NO concentration is calculated for various values of r/R as shown in Fig. 9(a)-(d). It is observed that the NO concentration is decreased by increasing IAA. The decrease is significant when IAA is increased from 15º to 30º. For r/R = 0.0, reduction of maximum NO is 63% when IAA is changed from 15º to 30º, while the reduction is only 40% when IAA is increased from 0º to 15º. In the combustor geometry with a sudden expansion, a recirculation zone that extends to the entrance of the combustor is created above the combustion zone [26]. Due to the recirculation zone, burnt gases come near the unburnt gases in the combustor entrance. Fig. 10 shows the streamlines for IAA = 30º. By comparing this figure with Fig. 7(a) (where IAA = 0º), it is clear that an increase in IAA results in extension of the recirculation zone. The recirculation zone is extended to X = 0.06m for IAA = 0º, while it is extended to X = 0.11m for IAA = 30º. Extension of this zone aids to more blending burnt and unburnt gases. Thus, the temperature of hot products decreases and accordingly, NO formation reduces. Fig. 11 shows the isotherms when IAA = 30º. By comparing Fig. 11 with Fig. 8(a) (where IAA = 0º), it is clear that extent of the combustion zone and the temperature of hot gases decreases as the result of recirculation zone development (the maximum temperature for IAA = 0º is 1688K, while this becomes 1526K for IAA = 30º). As a new index, to clarify the role of inclination angle of air inlet for strength of the recirculation zone, the mass-averaged
∫
integral of vorticity magnitude ( (1/ m) × ω dm ) is calculated. (b) Fig. 8. Isotherms for (a) ATI = 15%; (b) ATI = 45%.
tional to ATI, which is in agreement with the previous published results [8]. However, the geometry and the boundary conditions of fuel and air are different. With increasing ATI, more blending of the burnt and unburnt gases takes place [28]. Then, the temperature of hot gases decreases due to the high heat capacity of the burnt gases, and this is the main reason of NO reduction. This mechanism is similar to EGR mechanism used for NOx reduction [28]. To clarify, Figs. 7 and 8 present streamlines and isotherms with increasing ATI, respectively. As can be seen from Fig. 7, variation of ATI does not remain a dramatic effect on streamlines and recirculation zone extension. But there are some changes in isotherms. Fig. 8 shows that by increasing ATI, combustion begins in areas nearer to the inlet because more reactants blend together. On the other hand, the temperature of hot gases decreases by increasing ATI (for ATI = 15%, the maximum temperature is 1688K, while for ATI = 45% this becomes 1515K).
From Fig. 12, it is obvious that the increase in IAA can increase the vorticity magnitude in the combustor and hence, the recirculation zone can be extended. The value is increased by 25% when IAA increases from 0 to 30. Indeed, the reduction in NO concentration with inclination angle of air inlet is because of the extension of the recirculation zone. This recirculation dilutes the reactants, which decreases the temperature in the flame zone and hence results in reduction of NO concentration.
10. Conclusions The effects of air inlet conditions including air stream turbulence and inclined flow in an axisymmetric sudden-expansion geometry of a co-flowing methane-air diffusion flame on the NOx, flow field, and temperature concentration using the k-ε turbulence and β-PDF combustion model have been investigated. The obtained results were validated when compared with established published data. The analysis shows that the increasing air turbulence intensity results in decreasing NO formation. In addition, the effect of increasing inlet air angle
3729
S. A. Hashemi et al. / Journal of Mechanical Science and Technology 26 (11) (2012) 3723~3731 120 110
o
IAA=0 o IAA=15 o IAA=30
NO Concentration (ppm)
100 90 80 70 60 50 40 30 20
Fig. 10. Streamline plot for IAA = 30°.
10 0
0.1
0.3
0.5
0.7
0.9
1.1
1.3
1.5
1.7
z (m)
(a) 120 110
o
IAA=0 o IAA=15 o IAA=30
90 80 70 60
Fig. 11. Isotherms for IAA = 30°.
50 40
1100
30 20 10 0
1050
0.1
0.3
0.5
0.7
0.9
1.1
1.3
1.5
1.7
z (m)
1000
(b) 120 110
IAA=0 o IAA=15 o o IAA=30
100
NO Concentration (ppm)
Mass-averaged integral of vorticity magnitude
NO Concentration (ppm)
100
90
950
900
80 70
850
60
0
5
10
15
20
25
30
IAA (degree)
50
Fig. 12. Mass-averaged integral of vorticity magnitude in terms of IAA.
40 30 20 10 0
Acknowledgment 0.1
0.3
0.5
0.7
0.9
1.1
1.3
1.5
1.7
z (m)
The authors would like to thank University of Kashan for the support regarding this research (Grant No. 65477).
(c) 100 IAA=0 o IAA=15 o IAA=30 o
90
NO Concentration (ppm)
80
Nomenclature------------------------------------------------------------------------
70 60 50 40 30 20 10 0
0.1
0.3
0.5
0.7
0.9
1.1
1.3
1.5
1.7
z (m)
(d) Fig. 9. Axial NO concentration for various IAAs at (a) r/R = 0.0; (b) r/R = 0.1; (c) r/R = 0.3; (d) r/R = 0.6.
on NO formation was investigated by introducing a new index as the mass-averaged integral of vorticity magnitude. It was found that increasing the inclination angle of air inlet results in decreasing NO formation and increasing the mass-averaged integral of vorticity magnitude.
C1ε, C2, C3ε, σk, σε : Turbulence model constants D : Diffusion coefficient f : Mean mixture fraction f '2 : Variance mixture fraction h : Enthalpy Gk, Gb : Generation of turbulent kinetic Energy I : Turbulence intensity K : Turbulence kinetic energy ℓ : Characteristic length Mw,i : Molecular weight of species i MR : Initial momentum ratio NOx : Nitrogen oxides p(f) : Probability density function R : Universal gas constant Sφ1, Sφ2 : Source and sink terms Yi : Mass fraction of species i Vi : Velocity components
3730
S. A. Hashemi et al. / Journal of Mechanical Science and Technology 26 (11) (2012) 3723~3731
Greek symbols Γφ ε φ
: Generalized effective transport coefficient : Dissipation rate of turbulence kinetic energy : Generalized variable
Subscripts f I ox
: Fuel : Species : Oxidant
References [1] J. Y. Chen and W. C. Chang, Flamelet and pdf modeling of CO and NOx emissions from a turbulent, methane hydrogen jet non-premixed flame, Symposium (International) on Combustion, 26 (N2) (1996) 2207-2214. [2] N. Lallemant, F. Breussin, R. Weber, T. Ekman, J. Dugue and J. Samaniego, et al., Heat transfer and pollutant emissions characteristics of oxy-natural gas flames in the 0.71MW thermal input range, J. Inst. Energ. 73 (2000) 169-182. [3] M. Ilbas, I. Yilmaz, T. N. Veziroglu and Y. Kaplan, Hydrogen as burner fuel: modelling of hydrogen-hydrocarbon composite fuel combustion and NOx formation in a small burner, Int. J. Energ. Res. 29 (2004) 973-990. [4] B. B. Dally, E. Riesmeier and N. Peters. Effect of fuel mixture on moderate and intense low oxygen dilution combust, Combust. Flame. 137 (4) (2004) 418-430. [5] F. A. T. Lopez-Parra, Computational study on the effect of pulse characteristics on the soot and NOx formation and combustion in diffusion flames, (Ed.) Proc. European Combustion Meeting Louvain-la-Neuve, Belgium (2005). [6] F. A. T. Lopez-Parra, Computational study on the effect of turbulence intensity in soot formation and depletion in an acetylene diffusion flame. (Ed.) Proc. European Combustion Meeting, Louvain-la-Neuve, Belgium (2005). [7] T. Boushaki, M. A. Mergheni, J. C. Sautet and B. Labegorre, Effects of inclined jets on turbulent oxy-flame characteristics in a triple jet burner, Exp. Therm. Fluid. Sci. 32 (2008) 1363-1370. [8] F. Lopez-Parra and A. Turan, Computational study on the effects of non periodic flow perturbations on the emissions of soot and NOx in a confined turbulent methane/air diffusion flame, Combust. Sci. Tech. 179 (2007) 1361-1384. [9] H. K. Kim, Y. Kim, S. M. Lee and K. Y. Ahn, Emission characteristics of the 0.2 MW oxy-fuel combustor, Energ. Fuel. 23 (2009) 5331-5337. [10] K. M. Saqr, M. M. Sies and M. A. Wahid, Numerical investigation of the turbulence-combustion interaction in nonpremixed CH4/air flames, Int. J. Appl. Math. Mech. 5(8) (2009) 69-79. [11] J. Ahn, H. J Kim and K. S. Choi, Oxy-fuel combustion boiler for CO2 capturing: 50 kW-class model test and numerical simulation, J. Mech. Sci. Tech. 24 (10) (2010) 2135-2141. [12] W. Kim, D. J. Lee and S. W Park, Experimental study on
optimization of over-fire air in modified combustion condition with selective catalytic reduction, J. Mech. Sci. Tech. 24 (4) (2010) 901-909. [13] D. Lee, J. Park, J. Jin and M. Lee, A simulation for prediction of nitrogen oxide emissions in lean premixed combustor, J. Mech. Sci. Tech. 25 (7) (2011) 1871-1878. [14] H. K. Versteeg and W. Malalasekera, An introduction to computational fluid dynamics: the finite volume method, Addison Wesley-Longman (1995). [15] T. H. Shih, W. W. Lion, A. Shabbir, Z. Yang and J. Zhu, A new k-ε eddy-viscosity model for high Reynolds numerical turbulent flows-Model development and validation, Comput. Fluids. 24 (1995) 227-238. [16] F. Lopez-Parra and A. Turan, Computational study on the effect of turbulence intensity and pulse frequency in soot concentration in an acetylene diffusion flame, Int. Conference on Computational Sciences, Springer-Verlag Berlin, Heidelberg (2005). [17] T. Poinsot and D. Veynante, Theoretical and numerical combustion. Philadelphia, PA: R.T. Edwards Inc. (2001). [18] S. Repp, A. Sadiki, C. Schneider, A. Hinz, T. Landenfeld and J. Janicka, Prediction of swirling confined diffusion flame with a monte carlo and a presumed-PDF model, Int. J. Heat Mass Transf. 45 (2002) 1271-1285. [19] L. Y. Jiang and I. A. Campbell, Critical evaluation of NOx modeling in a model combustor, J. Eng. Gas Turb. Power. 127 (2005) 483-491. [20] F. K. Owen, L. J. Spaddacini and C. T. Bowman, Aerodynamic phenomena of pollutant formation in combustion. technical report, EPA-600/2-76-247a, Washington (1976). [21] M. C. Drake, S. M. Correa, R. W. Pitz, W. Shyy and C. P. Fenimore, Super equilibrium and thermal nitric oxide formation in turbulent diffusion flames, Combust. Flame. 69(N3) (1987) 347-365. [22] R. K. Hanson and S. Salimian, Survey of rate constants in the N/H/O system, combustion chemistry. In Gardiner WC, editor. New York, Springer (1984). [23] J. Bin, L. Hongying, H. Guoqiang and L. Xingang, Study on NOx formation in CH4/Air jet combustion, Chinese J. Chem. Eng. 14(N6) (2006) 723-728. [24] R. R. Raine, C. R. Stone and J. Gould, Modeling of nitric oxide formation in spark ignition engines with a multizone burned gas, Combust. Flame. 102 (3) (1995) 241-255. [25] G. De Soete, Overall reaction rates of NO and N2 formation from fuel nitrogen, Proc. Combust. Inst. Pittsburgh, USA (1974) 1093-1102. [26] J. Nisbet, L. Davidson and E. Olsson, Analysis of two fastchemistry combustion models and turbulence modeling in variable density flow, Comput. Fluid Dyn. 1 (1992) 557-563. [27] M. Ilbas, I. Yilmaz and Y. Kaplan, Investigations of hydrogen and hydrogen-hydrocarbon composite fuel combustion and NOx emission characteristics in a model combustor, Int J Hydrogen Energ. 30 (2005) 1139-1147. [28] SR. Turns, An introduction to combustion: concepts and applications. 2rd ed., Ch 2. Springer (2000).
S. A. Hashemi et al. / Journal of Mechanical Science and Technology 26 (11) (2012) 3723~3731
Seyed Abdolmehdi Hashemi is an Assistant Professor of Mechanical Engineering at University of Kashan, Kashan, Iran. He received his Ph.D. in Mechanical Engineering from Tarbait Modares University. His researches focus on numerical and experimental combustion phenomenon, such as pours media, burners and furnaces. Abolfazl Fattahi is a researcher of Energy Research Institute at University of Kashan, Kashan, Iran. He received his MSc degree from University of Kashan, Kashan, Iran. His research activities are paid on combustion (numerically and empirically), computational fluid dynamics and heat transfer.
3731
Ghanbar Ali Sheikhzadeh is an Associate Professor of Mechanical Engineering Department at University of Kashan, Kashan, Iran. He received his Ph.D in Mechanical Engineering from Shahid Bahonar University of Kerman. His research works concern numerical analysis and application of heat transfer in nano-systems and other areas of thermal and fluid sciences. Dr. Sheikhzadeh has published many papers in journal and conference in his research fields. Majid Nikfar is a researcher of Energy Research Institute at University of Kashan, Kashan, Iran. He received his Master's degree from University of Kashan, Kashan, Iran. His research interest is fluid dynamics, heat transfer and combustion.