- Email: [email protected]
Ambient pressure effect on non-premixed turbulent combustion of CH4–H2 mixture
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e6
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Ambient pressure effect on non-premixed turbulent combustion of CH4eH2 mixture L. Ziani a,b,*, A. Chaker b Centre de Developpement des Energies Renouvelables, BP. 62 Route de l’Observatoire, 16340, Bouzareah, Alger, Algerie b Laboratoire de Physique Energetique, Universite Mentouri Constantine, 25000, Algerie a
article info
abstract
Article history:
The present work is devoted to the study of the effect of ambient pressure on non-
Received 27 September 2015
premixed turbulent combustion of a mixture of 20% of hydrogen and 80% of methane.
Accepted 18 November 2015
The simulations were conducted using the PDF approach and modified k-ε turbulence
Available online xxx
model. The flamelet model was used with the GRI 2.11 detailed kinetic mechanism. The flame consists of two coaxial jets of air and a jet of CH4eH2 mixture with pilot flame. After
Keywords:
validating the model used we have varied the ambient pressure from 1 to 10 atm. We
Pressure effect
observed that the increase of ambient pressure leads to a slight increase of the flame
Numerical simulations
temperature and the mass fraction of NO and a decrease in the radial and axial expansions.
Non-premixed turbulent
Copyright © 2015, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights
combustion
reserved.
CH4eH2 mixtures
Introduction A big part of energy conversion systems use the fossil fuels as a source. This causes emissions of pollutants such as greenhouse gases. In addition, the threat of the fuel depletion makes urgent to find solutions. To deal with this problem, renewable energies seem like an ideal solution but their adoption in the short term poses technological and financial difficulties. An intermediate solution would be the introduction of hydrogen produced from renewable energy sources such as a fuel additive to improve combustion properties of fossil fuels. Indeed, the addition of hydrogen improves the combustion properties of hydrocarbons as reported in previous studies [1e4]. Thus, it is necessary to master the various parameters of combustion of these mixtures such as the ambient pressure to achieve the improvement of combustion systems.
The choice of a numerical model to simulate turbulent diffusion flame passes through the choice of a combustion model, a turbulence model and a kinetic scheme of combustion. For the combustion model we opted for the flamelet model. This model has been widely used and its effectiveness has been demonstrated in the literature [5e7]. The turbulence model used here is the modified k-ε model withCε1 ¼ 1.6. This model has been the subject of a comparative study between the RSM model and the standard k-ε model using the PDF approach, published recently by Ziani et al. [8]. The study shows that the modified k-ε model is the most accurate. Other studies confirm this result. Indeed, Frassoldati et al. [9] compared between RSM model, the standard k-ε and the modified k-ε using the EDC concept and shown that the modified k-ε turbulence model is the more accurate. Yilmaz and Onbasioglu [6] compared the LES model with the two k-ε models (standard and modified) and obtained the same
veloppement des Energies Renouvelables, BP. 62 Route de l’Observatoire 16340 Bouzare ah, Alger, * Corresponding author. Centre de De rie. Tel.: þ213 21 90 14 46; fax: þ213 21 90 15 60. Alge E-mail address: [email protected] (L. Ziani). http://dx.doi.org/10.1016/j.ijhydene.2015.11.167 0360-3199/Copyright © 2015, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. Please cite this article in press as: Ziani L, Chaker A, Ambient pressure effect on non-premixed turbulent combustion of CH4eH2 mixture, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2015.11.167
2
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e6
conclusion about the accuracy of the modified k-ε turbulence model. For the combustion chemistry, the reaction mechanism GRI 2.11 is chosen. When coupled with flamelet model, this mechanism appears to be most accurate and the convergence of calculation is relatively fast [10,11]. Moreover, Ravikanti and Malalasekera [7] compared this detailed kinetic scheme with three mechanisms and reported that the GRI 2.11 gives better results especially for the prediction of NO.
laminar flamelet are determined by solving governing equations for a one-dimensional laminar counter flow diffusion flame. The kinetic mechanism used here is the GRI 2.11 detailed mechanism which involves 49 species in 277 reactions and includes the nitrogen chemistry. We used a probability density function to calculate the average scalar properties as follow: e¼ f
the determination of the effect of pressure on the combustion temperature; the determination of the effect of pressure on the production of NO; the determination of the effect of pressure on the production of CO. Firstly, the validation of the physical and numerical modeling (based on the modified k-ε turbulence model coupled with the flamelet model using the PDF approach) is performed by comparing the obtained results with the experimental data of Brookes and Moss [12] considering the turbulent diffusion flame of pure methane. Secondly, we simulated the combustion of a mixture of 20% of hydrogen and 80% of methane, with the same operating conditions but with new jet velocities to avoid effects of buoyancy. The ambient pressure is varied from 1 to 10 atm to study the pressure effect on this type of turbulent diffusion flame configuration.
Physical and numerical modeling Physical modeling approaches We have used the flamelet model for the combustion simulation. In this model the turbulent diffusion flame is considered as an ensemble of laminar flamelet. The properties of the
fðZ; cÞPðZ; cÞdZdc 0
Specific objectives Brooks et al. [12] made an experimental study of the effect of ambient pressure on soot production in a turbulent diffusion flame of methane in 1998 and they published a theoretical study on the same subject in 2002. In both cases, they consider a pressure up to 3 atm. In 2005, Kobayashi et al. [13] published a study dealing with the experimental effect of pressure on the turbulent combustion velocity of the methane-air mixture in a benzene-type burner with a pressure ranging from 0.1 to 1 MPa. Conditions that approach the conditions of combustion et al. [14] made for the same type in gas turbines. In 2007, Cohe of burner a study of the effect of pressure on the structure of a premixed turbulent flame of methane enriched with hydrogen up to 20% for a pressure from 0.1 to 0.9 MPa. Griebel et al. [15] published in the same year a study on the effect of pressure on premixed turbulent combustion of methane for pressures up to 1.44 MPa with an axisymmetric burner configuration. In the present work, the effect of pressure on turbulent diffusion flame of a mixture of 20% of hydrogen and 80% of methane is studied numerically by means of a CFD code. Our objectives are:
Z∞ Z1 (1)
0
Where f is a thermochemical property (i.e. temperature, mass e its Favre average, Z is the Mixture fraction and c fraction …), f is a scalar dissipation. Considering that the scalar dissipation and mixture fraction are statistically independent, we can write: PðZ; cÞ ¼ PðZÞPðcÞ
(2)
Where P(Z) and P(c) are respectively the Probability Density Function of the mixture fraction and the scalar dissipation. The mean value of scalar dissipation rate is modeled by eε 00 c ¼ Cc Z 2 ke
(3)
where Cc is a constant equal to 2.0 and ke and eε are respectively the mean turbulent kinetic energy and the mean dissipation rate. The integration of the probability density function is made before the CFD calculation during the flamelet library generation. The values of this library are directly used by the CFD solver. The probability density function (PDF) gives the distribution probability of the stochastic quantities [16]. In turbulent flow, the probability density function P is a function of position in space x and time t. Then P(u,x,t)dU means the probability of finding at the position x and the time t the value u in the interval U u U þ dU. If the probability density P is known, the average value of a quantity is defined by the following formula [6,7]: u¼
Zþ∞ 2 u u lim Pðu; x; tÞdU x/∞
(4)
∞
The RANS approach (Reynolds-Averaged NaviereStokes equations) is used in the present study. We have used the modified k-ε with Cε1 ¼ 1.6 for the turbulence treatment in the numerical modeling.
Computational domain and conditions We have adopted the same geometrical configuration (Fig. 1) as the one adopted in in the experimental study of Brookes and Moss [12]. The flame consists of two axisymmetric jets of non-premixed air and fuel. The diameters of fuel and air jets are respectively D ¼ 4.07 mm and d ¼ 155 mm. The burner include annular premixed burner of 0.16 mm width. We used the Fluent, a CFD code which uses the finite volume discritization method to perform the simulation. We considered a two dimensions axisymmetric computational domain. The NaviereStokes equations are solved in such geometrical configuration. We used the centered scheme for the diffusion terms and first order upwind scheme for
Please cite this article in press as: Ziani L, Chaker A, Ambient pressure effect on non-premixed turbulent combustion of CH4eH2 mixture, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2015.11.167
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e6
3
Fig. 1 e Scheme of the geometrical configuration.
Table 1 e Operating conditions. Absolute pressure (atm) Fuel mass flow (g min1) Air mass flow (g min1) Fuel temperature (K) Air temperature (K) Exit Reynolds number
1 10.3 708 290 290 5000
convection terms. The SIMPLE algorithm is used for the velocity-pressure coupling. The imposed boundaries conditions are the mass flow inlet at the air and fuel inlets, wall boundary for the burner wall, axis condition at the center line of the configuration and pressure outlet condition at the exit.
Results Validation of the physical and numerical modeling Before studying the effect of the ambient pressure, it is important to validate the physical and numerical approaches
used in this work. We compared the results obtained by numerical simulation of a pure methane flame to those given by Brookes and Moss [12]. The operating conditions of the jet are provided in the Table 1: The pilot flame is a riche premixed flame of methane and oxygen, having a methane flow rate of 2% of the main fuel flow rate. The results of this comparison are shown in Fig. 2. This Figure illustrates the axial and radial distribution at 150 mm from the jet nozzle of temperature and mean mixture fraction for a pressure p ¼ 1 atm and the radial distribution of temperature at 150 mm for p ¼ 3 atm. We can notice that the numerical results of radial temperature at p ¼ 1 atm and p ¼ 3 atm are slightly higher than the experimental data except the radius range from 15 to 22 mm at atmospheric pressure. The same tendency is observed for the axial temperature at p ¼ 1 atm for axial distances higher than 100 mm. Concerning the mean mixture fraction, Fig. 2 exhibits a good agreement between the computed data and the experimental ones. From these comparisons, we can say that the physical and numerical approaches used to simulate non-premixed turbulent combustion of CH4eH2 mixture
Fig. 2 e (a) Axial temperature and mean mixture fraction for 1atm, (b)Radial distribution of temperature and mean mixture fraction for p ¼ 1 atm at 150 mm, (c) Radial distribution of temperature for p ¼ 3 atm at 150 mm. Please cite this article in press as: Ziani L, Chaker A, Ambient pressure effect on non-premixed turbulent combustion of CH4eH2 mixture, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2015.11.167
4
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e6
Table 2 e Operating conditions. Absolute pressure (atm) Fuel mass flow (g min1) Air mass flow (g min1) Fuel temperature (K) Air temperature (K) Exit Reynolds number
Varies from 1 to 10 30 1800 290 290 14,563
have a good behavior and can be used to study the effect of pressure.
Ambient pressure effect In this section, we focus on the effect of ambient pressure on the turbulent diffusion flame of a mixture of 20% of hydrogen and 80% of methane. The ambient pressure will be varied from 1 to 10 atm. We took the same conditions as previous
section case except for air intake flow rate and CH4eH2 mixture flow rate which will be increased to avoid buoyancy effects. The Table 2 shows the operating conditions. Fig. 3 shows the temperature contours for the cases studied. It can be note that as we increase the pressure the flame length starts to shorten. In other hands the increase of pressure makes the flame thinner. This difference in the geometry of the flame is more important between the case of the atmospheric pressure and the pressure of 3 atm. Fig. 4 shows the axial distribution and the radial distribution at 150 mm of temperature and mean mixture fraction for different values of ambient pressure. We can see from Fig. 4(a) a slight increase of temperature with the increase of ambient pressure for axial distances from 200 to 400 mm. After 400 mm, it decreases with the increase of pressure. The same behavior is observed for the radial distribution of temperature. Indeed, its increase with the pressure increase is in the radius range from 0 to 15 mm and the decrease phase is observed in
Fig. 3 e Temperature contours of the five cases with a pressure from 1 to 10 atm.
Fig. 4 e (a) Axial temperature and mean mixture fraction, (b) Radial distribution of temperature and mean mixture fraction at 150 mm. Please cite this article in press as: Ziani L, Chaker A, Ambient pressure effect on non-premixed turbulent combustion of CH4eH2 mixture, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2015.11.167
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e6
5
Fig. 5 e (a) Axial distribution of NO and CO mass fraction, (b)Radial distribution of NO and CO mass fraction at 150 mm.
the radius range from 15 to 25 mm. Concerning the mean mixture fraction, Fig. 4 shows that the increase of pressure lead to a quick decrease in the axial distribution. The radial distribution decreases with the increase of pressure. This means that the fuel is consumed more rapidly with increasing pressure and that the reaction occurs in a smaller distance with the increase of pressure. In Fig. 5 we can see the axial distribution and the radial distribution at 150 mm of CO and NO mass fractions for different values of the ambient pressure. We can note from this figure that there is not a significant difference in the mass fractions of CO. but we note that the increase of pressure result in an increase of NO mass fraction and a decrease of the radial and axial expansion which means an increase in the gradient. Similar results were observed, concerning the reduction of radial and axial expansion with the increase of et al. [14] and pressure. In fact, Kobayashi et al. [13], Cohe Griebel et al. [15] noted that the flame become shorter as the pressure increase. Fig. 6 shows the axial and radial distribution of the velocity magnitude in the flame. It is noted that the increase in pressure reduce the velocity magnitude near the axis and the burner nozzle and reduces the gradient of velocity magnitude in the radial and axial direction. This reduction in the velocity magnitude is due to the fact that the increase in pressure increases the density of the gas. For the same mass flow, with a bigger density the velocity magnitude is reduced. This reduced
speed at high pressure allows the reaction to take place in shorter distance which partly explains the flame shrinkage with increasing pressure.
Conclusions We first validate the physical and numerical modeling by simulating the non-premixed turbulent combustion of pure methane with the same condition used in the experimental work of Brooks et al. [12]. Then we made a study of the effect of ambient pressure variation on non-premixed turbulent combustion of a mixture of 80% methane and 20% hydrogen using the physical and numerical modeling validated in the first part. The conclusions that can be drawn from this section are: 1. The increase of ambient pressure results in a more quick decrease in the radial distribution of mean mixture fraction. 2. The increase of ambient pressure causes an increase in the NO mass fraction produced. 3. The increase of ambient pressure reduces the axial and radial expansion of flame. 4. The increase of ambient pressure has not a significant effect on the CO mass fraction produced.
Fig. 6 e (a) Axial distribution of velocity magnitude, (b) Radial distribution of velocity magnitude at 150 mm. Please cite this article in press as: Ziani L, Chaker A, Ambient pressure effect on non-premixed turbulent combustion of CH4eH2 mixture, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2015.11.167
6
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e6
references
[1] Miao H, Jiao Q, Huang Z, Jiang D. Measurement of laminar burning velocities and Markstein lengths of diluted hydrogen-enriched natural gas. Int J Hydrogen Energy 2009;34:507e18. [2] Miao H, Ji M, Jiao Q, Huang Q, Huang Z. Laminar burning velocity and Markstein length of nitrogen diluted natural gas/hydrogen/air mixtures at normal, reduced and elevated pressures. Int J Hydrogen Energy 2009;34:3145e55. [3] Di Sarli V, Di Benedetto A. Laminar burning velocity of hydrogenemethane/air premixed flames. Int J Hydrogen Energy 2007;32:637e46. [4] Di Sarli V, Di Benedettob A, Long EJ, Hargrave GK. Timeresolved particle image velocimetry of dynamic interactions between hydrogen-enriched methane/air premixed flames and toroidal vortex structures. Int J Hydrogen Energy 2012;37:16201e13. [5] Hossain M, Malalasekera W. Numerical study of bluff-body flame structures using laminar flamelet model. P I Mech Eng A J Pow 2005;219:361e70. [6] Yilmaz D, Onbasioglu SU. Computational modeling of hydrogen enriched non premixed turbulent methane air flames. In: Proceedings of European combustion meeting; 2005. [7] Ravikanti MHM, Malalasekera W. Laminar flamelet model prediction of NO formation in a turbulent bluff-body combustor. P I Mech Eng A J Pow 2009;223:41e54.
[8] Ziani L, Chaker A, Chetehouna K, Malek A, Mahmah B. Numerical simulations of non-premixed turbulent combustion of CH4eH2 mixtures using the PDF approach. Int J Hydrogen Energy 2013;38:8597e603. [9] Frassoldati A, Sharma P, Cuoci A, Faravelli T, Ranzi E. Kinetic and fluid dynamic modeling of methane/hydrogen jet flames in diluted coflow. Appl Therm Eng 2010;30:376e83. [10] Mardani A, Tabejamaat S. Effect of hydrogen on hydrogenemethane turbulent non-premixed flame under MILD condition. Int J Hydrogen Energy 2010;35:11324e31. [11] Rohani B, Saqr KM. Effects of hydrogen addition on the structure and pollutant emissions of a turbulent unconfined swirling flame. Int Commun Heat Mass 2012;39:681e8. [12] Brookes SJ, Moss JB. Measurements of soot production and thermal radiation from confined turbulent jet diffusion flames of methane. Combust flame 1999;116:49e61. [13] Kobayashi H, Seyama K, Hagiwara H, Ogami Y. Burning velocity correlation of methane/air turbulent premixed flames at high pressure and high temperature. Proc Combust Inst 2005;30:827e34. € Fractal C, Halter F, Chauveau C, Go € kalp I, Gu¨lder OL. [14] Cohe characterisation of high-pressure and hydrogen-enriched CH4eair turbulent premixed flames. Proc Combust Inst 2007;31:1345e52. [15] Griebel P, Siewert P, Jansohn P. Flame characteristics of turbulent lean premixed methane/air flames at high pressure: turbulent flame speed and flame brush thickness. Proc Combust Inst 2007;31:3083e90. [16] Poinsot T, Veynante D. Theoretical and numerical combustion. In: Edwards RT, editor. 2nd ed. 2005. Philadelphia.
Please cite this article in press as: Ziani L, Chaker A, Ambient pressure effect on non-premixed turbulent combustion of CH4eH2 mixture, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2015.11.167
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Ambient pressure effect on non-premixed turbulent combustion of CH4eH2 mixture L. Ziani a,b,*, A. Chaker b Centre de Developpement des Energies Renouvelables, BP. 62 Route de l’Observatoire, 16340, Bouzareah, Alger, Algerie b Laboratoire de Physique Energetique, Universite Mentouri Constantine, 25000, Algerie a
article info
abstract
Article history:
The present work is devoted to the study of the effect of ambient pressure on non-
Received 27 September 2015
premixed turbulent combustion of a mixture of 20% of hydrogen and 80% of methane.
Accepted 18 November 2015
The simulations were conducted using the PDF approach and modified k-ε turbulence
Available online xxx
model. The flamelet model was used with the GRI 2.11 detailed kinetic mechanism. The flame consists of two coaxial jets of air and a jet of CH4eH2 mixture with pilot flame. After
Keywords:
validating the model used we have varied the ambient pressure from 1 to 10 atm. We
Pressure effect
observed that the increase of ambient pressure leads to a slight increase of the flame
Numerical simulations
temperature and the mass fraction of NO and a decrease in the radial and axial expansions.
Non-premixed turbulent
Copyright © 2015, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights
combustion
reserved.
CH4eH2 mixtures
Introduction A big part of energy conversion systems use the fossil fuels as a source. This causes emissions of pollutants such as greenhouse gases. In addition, the threat of the fuel depletion makes urgent to find solutions. To deal with this problem, renewable energies seem like an ideal solution but their adoption in the short term poses technological and financial difficulties. An intermediate solution would be the introduction of hydrogen produced from renewable energy sources such as a fuel additive to improve combustion properties of fossil fuels. Indeed, the addition of hydrogen improves the combustion properties of hydrocarbons as reported in previous studies [1e4]. Thus, it is necessary to master the various parameters of combustion of these mixtures such as the ambient pressure to achieve the improvement of combustion systems.
The choice of a numerical model to simulate turbulent diffusion flame passes through the choice of a combustion model, a turbulence model and a kinetic scheme of combustion. For the combustion model we opted for the flamelet model. This model has been widely used and its effectiveness has been demonstrated in the literature [5e7]. The turbulence model used here is the modified k-ε model withCε1 ¼ 1.6. This model has been the subject of a comparative study between the RSM model and the standard k-ε model using the PDF approach, published recently by Ziani et al. [8]. The study shows that the modified k-ε model is the most accurate. Other studies confirm this result. Indeed, Frassoldati et al. [9] compared between RSM model, the standard k-ε and the modified k-ε using the EDC concept and shown that the modified k-ε turbulence model is the more accurate. Yilmaz and Onbasioglu [6] compared the LES model with the two k-ε models (standard and modified) and obtained the same
veloppement des Energies Renouvelables, BP. 62 Route de l’Observatoire 16340 Bouzare ah, Alger, * Corresponding author. Centre de De rie. Tel.: þ213 21 90 14 46; fax: þ213 21 90 15 60. Alge E-mail address: [email protected] (L. Ziani). http://dx.doi.org/10.1016/j.ijhydene.2015.11.167 0360-3199/Copyright © 2015, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. Please cite this article in press as: Ziani L, Chaker A, Ambient pressure effect on non-premixed turbulent combustion of CH4eH2 mixture, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2015.11.167
2
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e6
conclusion about the accuracy of the modified k-ε turbulence model. For the combustion chemistry, the reaction mechanism GRI 2.11 is chosen. When coupled with flamelet model, this mechanism appears to be most accurate and the convergence of calculation is relatively fast [10,11]. Moreover, Ravikanti and Malalasekera [7] compared this detailed kinetic scheme with three mechanisms and reported that the GRI 2.11 gives better results especially for the prediction of NO.
laminar flamelet are determined by solving governing equations for a one-dimensional laminar counter flow diffusion flame. The kinetic mechanism used here is the GRI 2.11 detailed mechanism which involves 49 species in 277 reactions and includes the nitrogen chemistry. We used a probability density function to calculate the average scalar properties as follow: e¼ f
the determination of the effect of pressure on the combustion temperature; the determination of the effect of pressure on the production of NO; the determination of the effect of pressure on the production of CO. Firstly, the validation of the physical and numerical modeling (based on the modified k-ε turbulence model coupled with the flamelet model using the PDF approach) is performed by comparing the obtained results with the experimental data of Brookes and Moss [12] considering the turbulent diffusion flame of pure methane. Secondly, we simulated the combustion of a mixture of 20% of hydrogen and 80% of methane, with the same operating conditions but with new jet velocities to avoid effects of buoyancy. The ambient pressure is varied from 1 to 10 atm to study the pressure effect on this type of turbulent diffusion flame configuration.
Physical and numerical modeling Physical modeling approaches We have used the flamelet model for the combustion simulation. In this model the turbulent diffusion flame is considered as an ensemble of laminar flamelet. The properties of the
fðZ; cÞPðZ; cÞdZdc 0
Specific objectives Brooks et al. [12] made an experimental study of the effect of ambient pressure on soot production in a turbulent diffusion flame of methane in 1998 and they published a theoretical study on the same subject in 2002. In both cases, they consider a pressure up to 3 atm. In 2005, Kobayashi et al. [13] published a study dealing with the experimental effect of pressure on the turbulent combustion velocity of the methane-air mixture in a benzene-type burner with a pressure ranging from 0.1 to 1 MPa. Conditions that approach the conditions of combustion et al. [14] made for the same type in gas turbines. In 2007, Cohe of burner a study of the effect of pressure on the structure of a premixed turbulent flame of methane enriched with hydrogen up to 20% for a pressure from 0.1 to 0.9 MPa. Griebel et al. [15] published in the same year a study on the effect of pressure on premixed turbulent combustion of methane for pressures up to 1.44 MPa with an axisymmetric burner configuration. In the present work, the effect of pressure on turbulent diffusion flame of a mixture of 20% of hydrogen and 80% of methane is studied numerically by means of a CFD code. Our objectives are:
Z∞ Z1 (1)
0
Where f is a thermochemical property (i.e. temperature, mass e its Favre average, Z is the Mixture fraction and c fraction …), f is a scalar dissipation. Considering that the scalar dissipation and mixture fraction are statistically independent, we can write: PðZ; cÞ ¼ PðZÞPðcÞ
(2)
Where P(Z) and P(c) are respectively the Probability Density Function of the mixture fraction and the scalar dissipation. The mean value of scalar dissipation rate is modeled by eε 00 c ¼ Cc Z 2 ke
(3)
where Cc is a constant equal to 2.0 and ke and eε are respectively the mean turbulent kinetic energy and the mean dissipation rate. The integration of the probability density function is made before the CFD calculation during the flamelet library generation. The values of this library are directly used by the CFD solver. The probability density function (PDF) gives the distribution probability of the stochastic quantities [16]. In turbulent flow, the probability density function P is a function of position in space x and time t. Then P(u,x,t)dU means the probability of finding at the position x and the time t the value u in the interval U u U þ dU. If the probability density P is known, the average value of a quantity is defined by the following formula [6,7]: u¼
Zþ∞ 2 u u lim Pðu; x; tÞdU x/∞
(4)
∞
The RANS approach (Reynolds-Averaged NaviereStokes equations) is used in the present study. We have used the modified k-ε with Cε1 ¼ 1.6 for the turbulence treatment in the numerical modeling.
Computational domain and conditions We have adopted the same geometrical configuration (Fig. 1) as the one adopted in in the experimental study of Brookes and Moss [12]. The flame consists of two axisymmetric jets of non-premixed air and fuel. The diameters of fuel and air jets are respectively D ¼ 4.07 mm and d ¼ 155 mm. The burner include annular premixed burner of 0.16 mm width. We used the Fluent, a CFD code which uses the finite volume discritization method to perform the simulation. We considered a two dimensions axisymmetric computational domain. The NaviereStokes equations are solved in such geometrical configuration. We used the centered scheme for the diffusion terms and first order upwind scheme for
Please cite this article in press as: Ziani L, Chaker A, Ambient pressure effect on non-premixed turbulent combustion of CH4eH2 mixture, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2015.11.167
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e6
3
Fig. 1 e Scheme of the geometrical configuration.
Table 1 e Operating conditions. Absolute pressure (atm) Fuel mass flow (g min1) Air mass flow (g min1) Fuel temperature (K) Air temperature (K) Exit Reynolds number
1 10.3 708 290 290 5000
convection terms. The SIMPLE algorithm is used for the velocity-pressure coupling. The imposed boundaries conditions are the mass flow inlet at the air and fuel inlets, wall boundary for the burner wall, axis condition at the center line of the configuration and pressure outlet condition at the exit.
Results Validation of the physical and numerical modeling Before studying the effect of the ambient pressure, it is important to validate the physical and numerical approaches
used in this work. We compared the results obtained by numerical simulation of a pure methane flame to those given by Brookes and Moss [12]. The operating conditions of the jet are provided in the Table 1: The pilot flame is a riche premixed flame of methane and oxygen, having a methane flow rate of 2% of the main fuel flow rate. The results of this comparison are shown in Fig. 2. This Figure illustrates the axial and radial distribution at 150 mm from the jet nozzle of temperature and mean mixture fraction for a pressure p ¼ 1 atm and the radial distribution of temperature at 150 mm for p ¼ 3 atm. We can notice that the numerical results of radial temperature at p ¼ 1 atm and p ¼ 3 atm are slightly higher than the experimental data except the radius range from 15 to 22 mm at atmospheric pressure. The same tendency is observed for the axial temperature at p ¼ 1 atm for axial distances higher than 100 mm. Concerning the mean mixture fraction, Fig. 2 exhibits a good agreement between the computed data and the experimental ones. From these comparisons, we can say that the physical and numerical approaches used to simulate non-premixed turbulent combustion of CH4eH2 mixture
Fig. 2 e (a) Axial temperature and mean mixture fraction for 1atm, (b)Radial distribution of temperature and mean mixture fraction for p ¼ 1 atm at 150 mm, (c) Radial distribution of temperature for p ¼ 3 atm at 150 mm. Please cite this article in press as: Ziani L, Chaker A, Ambient pressure effect on non-premixed turbulent combustion of CH4eH2 mixture, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2015.11.167
4
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e6
Table 2 e Operating conditions. Absolute pressure (atm) Fuel mass flow (g min1) Air mass flow (g min1) Fuel temperature (K) Air temperature (K) Exit Reynolds number
Varies from 1 to 10 30 1800 290 290 14,563
have a good behavior and can be used to study the effect of pressure.
Ambient pressure effect In this section, we focus on the effect of ambient pressure on the turbulent diffusion flame of a mixture of 20% of hydrogen and 80% of methane. The ambient pressure will be varied from 1 to 10 atm. We took the same conditions as previous
section case except for air intake flow rate and CH4eH2 mixture flow rate which will be increased to avoid buoyancy effects. The Table 2 shows the operating conditions. Fig. 3 shows the temperature contours for the cases studied. It can be note that as we increase the pressure the flame length starts to shorten. In other hands the increase of pressure makes the flame thinner. This difference in the geometry of the flame is more important between the case of the atmospheric pressure and the pressure of 3 atm. Fig. 4 shows the axial distribution and the radial distribution at 150 mm of temperature and mean mixture fraction for different values of ambient pressure. We can see from Fig. 4(a) a slight increase of temperature with the increase of ambient pressure for axial distances from 200 to 400 mm. After 400 mm, it decreases with the increase of pressure. The same behavior is observed for the radial distribution of temperature. Indeed, its increase with the pressure increase is in the radius range from 0 to 15 mm and the decrease phase is observed in
Fig. 3 e Temperature contours of the five cases with a pressure from 1 to 10 atm.
Fig. 4 e (a) Axial temperature and mean mixture fraction, (b) Radial distribution of temperature and mean mixture fraction at 150 mm. Please cite this article in press as: Ziani L, Chaker A, Ambient pressure effect on non-premixed turbulent combustion of CH4eH2 mixture, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2015.11.167
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e6
5
Fig. 5 e (a) Axial distribution of NO and CO mass fraction, (b)Radial distribution of NO and CO mass fraction at 150 mm.
the radius range from 15 to 25 mm. Concerning the mean mixture fraction, Fig. 4 shows that the increase of pressure lead to a quick decrease in the axial distribution. The radial distribution decreases with the increase of pressure. This means that the fuel is consumed more rapidly with increasing pressure and that the reaction occurs in a smaller distance with the increase of pressure. In Fig. 5 we can see the axial distribution and the radial distribution at 150 mm of CO and NO mass fractions for different values of the ambient pressure. We can note from this figure that there is not a significant difference in the mass fractions of CO. but we note that the increase of pressure result in an increase of NO mass fraction and a decrease of the radial and axial expansion which means an increase in the gradient. Similar results were observed, concerning the reduction of radial and axial expansion with the increase of et al. [14] and pressure. In fact, Kobayashi et al. [13], Cohe Griebel et al. [15] noted that the flame become shorter as the pressure increase. Fig. 6 shows the axial and radial distribution of the velocity magnitude in the flame. It is noted that the increase in pressure reduce the velocity magnitude near the axis and the burner nozzle and reduces the gradient of velocity magnitude in the radial and axial direction. This reduction in the velocity magnitude is due to the fact that the increase in pressure increases the density of the gas. For the same mass flow, with a bigger density the velocity magnitude is reduced. This reduced
speed at high pressure allows the reaction to take place in shorter distance which partly explains the flame shrinkage with increasing pressure.
Conclusions We first validate the physical and numerical modeling by simulating the non-premixed turbulent combustion of pure methane with the same condition used in the experimental work of Brooks et al. [12]. Then we made a study of the effect of ambient pressure variation on non-premixed turbulent combustion of a mixture of 80% methane and 20% hydrogen using the physical and numerical modeling validated in the first part. The conclusions that can be drawn from this section are: 1. The increase of ambient pressure results in a more quick decrease in the radial distribution of mean mixture fraction. 2. The increase of ambient pressure causes an increase in the NO mass fraction produced. 3. The increase of ambient pressure reduces the axial and radial expansion of flame. 4. The increase of ambient pressure has not a significant effect on the CO mass fraction produced.
Fig. 6 e (a) Axial distribution of velocity magnitude, (b) Radial distribution of velocity magnitude at 150 mm. Please cite this article in press as: Ziani L, Chaker A, Ambient pressure effect on non-premixed turbulent combustion of CH4eH2 mixture, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2015.11.167
6
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e6
references
[1] Miao H, Jiao Q, Huang Z, Jiang D. Measurement of laminar burning velocities and Markstein lengths of diluted hydrogen-enriched natural gas. Int J Hydrogen Energy 2009;34:507e18. [2] Miao H, Ji M, Jiao Q, Huang Q, Huang Z. Laminar burning velocity and Markstein length of nitrogen diluted natural gas/hydrogen/air mixtures at normal, reduced and elevated pressures. Int J Hydrogen Energy 2009;34:3145e55. [3] Di Sarli V, Di Benedetto A. Laminar burning velocity of hydrogenemethane/air premixed flames. Int J Hydrogen Energy 2007;32:637e46. [4] Di Sarli V, Di Benedettob A, Long EJ, Hargrave GK. Timeresolved particle image velocimetry of dynamic interactions between hydrogen-enriched methane/air premixed flames and toroidal vortex structures. Int J Hydrogen Energy 2012;37:16201e13. [5] Hossain M, Malalasekera W. Numerical study of bluff-body flame structures using laminar flamelet model. P I Mech Eng A J Pow 2005;219:361e70. [6] Yilmaz D, Onbasioglu SU. Computational modeling of hydrogen enriched non premixed turbulent methane air flames. In: Proceedings of European combustion meeting; 2005. [7] Ravikanti MHM, Malalasekera W. Laminar flamelet model prediction of NO formation in a turbulent bluff-body combustor. P I Mech Eng A J Pow 2009;223:41e54.
[8] Ziani L, Chaker A, Chetehouna K, Malek A, Mahmah B. Numerical simulations of non-premixed turbulent combustion of CH4eH2 mixtures using the PDF approach. Int J Hydrogen Energy 2013;38:8597e603. [9] Frassoldati A, Sharma P, Cuoci A, Faravelli T, Ranzi E. Kinetic and fluid dynamic modeling of methane/hydrogen jet flames in diluted coflow. Appl Therm Eng 2010;30:376e83. [10] Mardani A, Tabejamaat S. Effect of hydrogen on hydrogenemethane turbulent non-premixed flame under MILD condition. Int J Hydrogen Energy 2010;35:11324e31. [11] Rohani B, Saqr KM. Effects of hydrogen addition on the structure and pollutant emissions of a turbulent unconfined swirling flame. Int Commun Heat Mass 2012;39:681e8. [12] Brookes SJ, Moss JB. Measurements of soot production and thermal radiation from confined turbulent jet diffusion flames of methane. Combust flame 1999;116:49e61. [13] Kobayashi H, Seyama K, Hagiwara H, Ogami Y. Burning velocity correlation of methane/air turbulent premixed flames at high pressure and high temperature. Proc Combust Inst 2005;30:827e34. € Fractal C, Halter F, Chauveau C, Go € kalp I, Gu¨lder OL. [14] Cohe characterisation of high-pressure and hydrogen-enriched CH4eair turbulent premixed flames. Proc Combust Inst 2007;31:1345e52. [15] Griebel P, Siewert P, Jansohn P. Flame characteristics of turbulent lean premixed methane/air flames at high pressure: turbulent flame speed and flame brush thickness. Proc Combust Inst 2007;31:3083e90. [16] Poinsot T, Veynante D. Theoretical and numerical combustion. In: Edwards RT, editor. 2nd ed. 2005. Philadelphia.
Please cite this article in press as: Ziani L, Chaker A, Ambient pressure effect on non-premixed turbulent combustion of CH4eH2 mixture, International Journal of Hydrogen Energy (2016), http://dx.doi.org/10.1016/j.ijhydene.2015.11.167