Numerical Study of the Effects of Porous Burner Parameters on Combustion and Pollutants Formation

Transcription

1 Numerical Study of the Effects of Porous Burner Parameters on Combustion and Pollutants Formation Siama Hossainpour 1, Bahman Haddadi Abstract In recent years, more attention has been focused on the use of porous materials to enhance the efficiency of combustion systems and to reduce the emission of pollutants. This is because combustion in inert porous media offers an interestin and promisin route towards burner with hih-power density, hih-power dynamic rane, and very low emission of pollutants such as NO x and CO. This wor reports one-dimensional combustion in a porous burner usin three combustion models: GRI 3.0, GRI 1., seletal mechanism. We conclude that seletal mechanism has a ood areement with GRI 3.0 and it costs less. At first, we present a numerical study which shows the effects of these models on temperature, species and pollutant emissions. Then, we investiate the effects of volumetric heat transfer and emissivity coefficient and porosity on combustion and pollutions. It was concluded that NO and CO emission depend mainly on the volumetric and emissivity coefficient. When volumetric heat transfer decreased, the difference between as and solid temperature reduced, therewith NO formation noticeably decreased whereas CO emission didn t chane sensible. On the other hand, the flame pea temperature is reduced with the reduction of the solid emissivity coefficient. This important conclusion means that NO and CO emission and velocity increases. Also as and solid temperature increase and vice versa. The other parameter is Porosity. Increasin in porosity of burner resulted in decreasin as and solid temperature and subsequently NO and CO emission decreased sensible. Porosity has effected on velocity, too. As porosity decreased, velocity increased. Emissivity effects on the rate of heat flux which issue from burner. As the emissivity increased the efficiency of burner arose. So these parameters have important roles in decreasin the emission especially on No emission because it has more depend on temperature. In addition the resulted as and solid temperatures were compared with reported measurements of center line temperature in a cylindrical porous burner. The ood areement with experimental observation upholds that the numerical model is a perfect tool to investiate combustion and pollutants formation in porous media. Keywords Emissivity Coefficient, Porosity, Porous burner, Volumetric Heat Transfer I. INTRODUCTION Over the past 30 years, our understandin of combustion in porous media has increased substantially [1] []. Development of porous burners has been encouraed by lower emission standards as well as the advantaes these Manuscript submission April, 1, Department of Mechanical Enineerin, Sahand University of Technoloy, Tabriz, Iran hossainpour@sut.ac.ir Mechanical Enineerin MS student, Sahand University of Technoloy, Tabriz, Iran Tel : b_haddadi@sut.ac.ir, Haddadi60@mail.com burners offer such as fuel flexibility, hih-power density, hih-power dynamic rane and also the ability to operate at low equivalence ratios, and effective flame speeds reater than the laminar flame speed [3] [5]. This is due to the hih heat capacity, hih conductivity and hih emissivity of the solid matrix in comparison to a as. In a porous burner, the porous matrix re-circulates heat from the post-flame to the pre-flame zone throuh solid-to-solid radiation and conduction leadin to excess enthalpy flames [5] [7]. This reenerative internal heat feedbac mechanism results in several interestin characteristics relative to a free-burnin flame, namely hiher burnin speeds, extension of the lean flammability limit, low emission of pollutants and the ability to burn fuels that have a low enery content. Also, inert porous media combustors may offer hih compact very small scale sizes, which correspond to the desired characteristics of industrial applications or household heatin combustion. Development and optimization desin of advanced combustion systems which meet pollutant emission standards and maintain or increase productivity will require mathematical modellin of the systems. The reduction in time between a system concept and commercialization and the increase in the cost of testin will demand a reater reliance on mathematical models to simulate the systems and reduce the cost and time to develop a product. Thus, models with varyin derees of sophistication have been recently developed and applied to the problem of predictin flame speeds, temperature and concentration profiles and radiative efficiency of combustion within porous media. A detailed review of modellin combustion in porous media can be found in Howel et al. (1996). An early analysis was completed by Echio (198) to investiate the ability of convertin some of the enthalpy of a non-reactin hot as for radiative transfer from a porous medium throuh which the as is flowin. The Echio and co-worers (Echio et al., 1986; Yoshizawa et al., 1988; Echio, 1991) provided a riorous model for multi-mode heat transfer, Arrhenius-type one-step reaction inetics and exact solution for radiative transfer in the absorbin/emittin medium. However, because few data were available at the time of this wor on thermo physical or transport properties in these systems, they made various simplifyin assumptions and choices of property values. A sinle-step mechanism was used for methane combustion, so that multiple species equations could be dispensed. And it was assumed that the burner could be divided into three reions. an upstream reion where no reactions occur, so that the as/solid temperature were constant; a combustion zone, where the one-step combustion reaction oes to completion; and an exit zone, where the ases leavin the combustion zone aain undero

2 no further reaction. Based on these assumptions, temperature profiles in the as were predicted. Chen et al. (1987) applied the enery and species equations to model porous medium burners. Notin the deficiency of one-step inetics in the Yoshizawa et al. (1988) analysis, a multi-step mechanism for methane combustion was used in the model, based on the reaction set from the code CHEMKIN by Kee et al. (1980) which includes 17 species and 55 reactions. Parametric variations of the thermal conductivity of the solid, volumetric heat transfer coefficient and radiative properties were carried out to determine their effect on flame speed and temperature profiles. In 1991, Ton and co-worers surveyed application of porous material in the burner structure [8]. They supposed one dimensional system and usin Spherical harmonic method in calculatin of radiation flux. They found that for maximum radiation outlet flux, they should use from porous layer with hih thicness. They simulate the combustion phenomena with a source of steady heat eneration in the porous media. Hsu et al. (Hsu et al., 1993 and Hsu and Matthews, 1993) extended Chen s model to include the Zeldovich mechanism (3 reactions and two additional species) for No chemistry, and experimental values for thermal conductivity and radiative extinction coefficient that were unavailable to Chen (1988). In addition, a two-reion burner with a small-pore size upstream section and lare-pore downstream section was modeled. Hsu compared his modellin results with experimental data athered on two-reion porous media burners made of partially stabilized zirconia of various pores size. The model was accurate in predictin the maximum flame speeds sustainable within the burner (blow-off limit); the minimum equivalence ratio for sustainable combustion; the trends of flame speed with pore diameter and equivalence ratio and the measured emissions of CO, CO and NO. The model couldn't, however, predict the minimum sustainable flame speeds in the burner. Lee et al. (1996) had experimentally and numerically investiated the combustion of premixed propane-air mixture inside a honeycomb ceramic. They used the one-dimensional flame structure model and a one-step reaction mechanism. They obtained the upstream and downstream solutions correspondin to each upper and lower solution. These are correspondent with the experimental results. Visanta and Gore [] performed a numerical study usin cordierite with 6 pores per centimeter (ppc) in the upstream section and cordierite LS- (4 ppc) in the downstream section. The total burner lenth was approximately.5 cm, includin both the upstream and downstream layer, and the firin rates varied from 36 (W/m ) to 394 (W/m ) at an equivalence ratio of 0.9 for a methane/air mixture. They used the experimentally observed flame location to predict the temperatures and radiant output. Visanta and Gore s results show that a larer heat transfer coefficient results in a hiher pea solid temperature, which promoted hiher radiative flux from the hih temperature zone, but did not sinificantly affect the maximum as temperature. In addition, increasin conductivity in the downstream section results in decreasin in solid temperature. Kularni and Pec [9] modeled a 5 cm lon two section burner. They determined the effect of porosity, lenth, radiation extinction coefficient, and albedo on radiant output from the burner. They concluded that the upstream layer should be of lower porosity, shorter lenth, and hiher optical thicness (product of extinction coefficient and lenth) than the downstream layer and that the upstream layer should be scatterin and the downstream non-scatterin in order to maximize radiant output. In 000, Brenner and co-worers assumed two dimensional systems and solve the overnin equations for combustion but they did not use any model for radiation [10]. In 003, Taludar and co-worers studied porous burner in two conditions, steady state and transient state, in the two dimensional system [11]. In this study, the effect of radiation diffusion to surroundin was considered. They used Collapsed Dimension method to specify radiation flux inert porous burner [1]. The purpose of the present wor is to develop a mathematical model for fluid flow, combustion and heat transfer in porous media via development of the PREMIX code which is used for modelin steady laminar one-dimensional premixed flames. In this paper, we modeled one- dimensional combustion and heat transfer of methane/air fuel in a two-reion burner with a small-pore size upstream section and lare-pore downstream section. Three mechanisms for the methane/air fuel combustion were used in this model. First Mechanism is GRI-3.0 which includes 77 species and 7 reactions [13]. Second mechanism is GRI-1. that includes 3 species and 177 reactions. Third mechanism is Seletal mechanism. This Seletal mechanism was developed by Glarbor et al. [14]. It is derived from the full reaction mechanism throuh sensitivity analysis and rate of production analysis of perfectly stirred reactor calculations coverin the rane of interest. This Seletal mechanism comprises 6 species and 77 elemental reactions. The purpose of this seletal mechanism is to provide a ood description of methane oxidation includin nitroen chemistry at hih temperature (T>1500K) and not excessively fuel-rich conditions. Compared to the full mechanism, the most important simplification of the Seletal model is the exclusion of C-hydrocarbon chemistry. However, the most important reaction steps for methane oxidation and nitroen chemistry are retained in the seletal scheme. We investiated the effects of volumetric heat transfer, emissivity coefficient and porosity on the temperature profiles and emission of Co and No in a porous burner. In addition, the predicted as and solid temperatures were compared with experimental observation. The focus of this paper is on the predicted pollutants formation and their comparison with experiments. The ood areement with experimental observations suests that usin this developed PREMIX code is an excellent tool to investiate the different as fuel combustion and pollutants formation in porous media. Also this method costs low prices.

3 II. NUMERICAL METHOD The problem considered in the present study is a one-dimensional methane/air fuel combustion process within a porous burner. The model has been adapted from that described in Zhou and Pereira [15] and Hennee and Ellzey [16] for two sections of porous media. The present modellin of combustion within porous media has made the followin principal assumptions: (1) The flame structure and heat transfer mechanism are one-dimensional. () Potential catalytic effects of the hih temperature solid are neliible. (3) The Dofour, Soret, 'bul' viscosity and body forces are neliible. (4) The flow speed is sufficiently low that the process is isobaric. (5) The mixture as is non radiatin. (6) Compressibility effects are nelected Based upon the above assumptions, the flame can be assumed to be one-dimensional and solved by usin adiabatic laminar flame theory. The one-dimensional laminar flame PREMIX code [17] was modified for current use. This code allows the use of three mechanism detailed chemical inetics [18] and use of the TRANFIT subroutine [19] for accurate determination of the transport properties of the as. We made some chanes in this code to solve this problem. These chanes include these items: 1- Addin porous material s properties and calculatin the coefficient of radiation heat transfer. - Addin radiative equation of porous media in overnin equations 3- Addin enery conservation equation of solid phase 4- Addin porosity in overnin equations and term of convection heat transfer in as and solid phase equations 5- Chanin in boundary conditions A. Burner eometry _ ρ φ c (3) p, ρ φ c s T ρφ uc p, T T φ K T ( 1 φ ) hv ( T Ts ) φ ρ YVc p, φ ω hw = 0. = 1 K = 1 Ts T dq p, s s = v s Y ρφ ρφ u ( = 1,..., K) s n ( 1 φ) ( 1 φ) h ( T T ) 0 Y ( ρ φy V ) φω W = 0 The radiation equation calculates from two-flux model [0]: dq 4 = σ aq bε q σ aσ Ts bε q dq 4 = σ aq bε q σ aσ Ts bε q dq n () (4) (5) (6) dq dq = (7) Convection was included by solvin separate enery equations for the solid and the as and couplin them throuh a convective heat transfer coefficient. The enery equation for the as does not include radiation terms and the enery equation for the solid does not include enery liberation (reaction) terms. Gas phase thermo chemical and transport properties are obtained from the CHEMKIN [17] and TRANFIT [18] pacaes. All simulations are conducted for methane/air mixtures. The GRI 1. chemical inetics mechanism [1] [] with 3 species and 177 reactions, GRI 3.0 [13] with 77 species and 7 reactions and Seletal mechanism [14] with 6 species and 77 elemental reactions is used to represent the chemistry. C. Boundary conditions B. Governin equations Fiure 1 Porous burner schematic The followin conservation equations for mass, as enery, solid enery, and as species are solved: ( ρ φ) ( ρ u φ) 4 cm 1 cm = 0 (1) The conditions specified at the inlet are the temperature and mass fraction of methane/air fuel and two below equations: At the inlet: dt mc p, ( Ti, T ) = at x= dts ( 1 φ ) [ hi ( Ti, Ts ) σε ( Ti, Ts )] = s T = Tin Y = Y, in at x= 0 (8)

4 The conditions specified at the outlet are the temperature and mass fraction of methane/air fuel and one below equation: At the outlet: dt = 0 at x = L 4 4 dts ( 1 φ ) [ ho ( To, Ts ) σε ( To, sur Ts )] = s at x = L (9) dy = 0 temperature of solid and as within porous media is not correct. As it is obvious in Fi. 3, which shows CO profiles toward distance (x), the maximum of CO was produced at the flame front and it decreased in downstream aain. Fi. 4 shows the CO profiles toward distance (x). It exhibits that production of CO is depends on the temperature. As the temperature increased the concentration of CO increased sinificantly. D. Solution method The overnin equations were solved with the damped Newton scheme used by modified PREMIX code. The number of rid bean with 6 and continued to 00 rids. The run time is very dependent on initial uess. Typically -0 minutes is needed in a Pentium 4 computer for cases started with ood initial uesses. In some cases, the problem taes several hours to reach convered solution. For example, to solve this problem with GRI 3.0 mechanism, it taes 1 hour. For all cases, a relative converence of l0-5 was specified, which corresponds to four sinificant diits in the results. The absolute converence was l0-9. The rid effects on the solutions were examined by increasin the number of rid points after initial solution until the results chaned no loner in a specified tolerance. Fiure 3 Emission of CO for different reaction mechanism with stoichiometric combustion in 5000 w/m power III. RESULT AND DISCUSSION Fi. displays the as temperature profiles in porous media with stoichiometric and power 5000 w/m. As this Fiure shows, GRI 1. has a ood areement with GRI 3.0. These results have a ood areement with that described in Zhou and Pereira [15] and Hennee and Ellzey [16] for two sections of porous media. Fiure 4 Emission of CO for different reaction mechanism with stoichiometric combustion in 5000 w/m power Fi. 5 shows NO profiles. Most of NO was produced at the flame front. It is obvious that temperature effects the production of NO considerably. From fiure, GRI 3.0 and Seletal mechanism have a little difference in estimatin of temperatures profiles. Fiure Temperature profiles for stoichiometric combustion and 5000 w/m power Fi. displays the temperature profiles of as phase and solid matrix in stoichiometric and power 5000 w/m. The temperature for as and solid at inlet and outlet of burner is almost the same, but the difference at the flame front is evident. This shows that the assumption of the same This little variance mae a reatly difference in NO production. For Seletal mechanism, the pic temperature is 143 K, the exit NO is , for GRI 3.0 mechanism the pic temperature is 80 K, and the exit NO is

5 Fiure 5 Emission of NO for stoichiometric combustion and 5000 w/m power The effect of volumetric heat transfer on temperature profiles are shown in Fi. 6. GRI 3.0 reaction mechanism used in volumetric heat transfer effected on the difference between as and solid temperature. As the volumetric heat transfer increased, the temperature of solid and as ot closely toether. It can be explained by this fact that hih volumetric heat transfer can cause more heat transfer between as and solid, so their temperature became closely toether. The other effect is on the as temperature. By increasin the volumetric heat transfer the as temperature decreased slowly. The effect of emissivity on temperature profile was shown in Fi. 7. This picture shows that when emissivity decreased the temperature of as and solid increased considerably. It is because of by decreasin the emissivity factor the heat transfer to surroundin decreases. So the heat remains in the porous burner and increases the temperature. On the other hand, increasin in the temperature causes increasin the emission of NO and CO rapidly. Fiure 7 Temperature profiles for different emissivity Fi. 8 expresses the temperature profiles with different porosity. Porosity has important effect on the temperature. As the porosity decreased the pea of temperature increased sinificantly and the temperature of the outlet as increased, too. Fiure 8 Temperature profiles for different porosity This can explain that some characteristics of porous media such as hih emissivity mae a ood heat transfer with surroundin and removin the heat which produces from combustion. This is an important parameter for decreasin emissions. ACKNOWLEDGMENT This wor was supported by Office of Research of Sahand University of Technoloy and I would lie to than the head of this office. REFERENCES Fiure 6 Temperature profiles for different volumetric heat transfer [1] J.R. Howell, M.J. Hall, J.L. Ellzey, Pro. Enery Combust. Sci. () (1996) [] R. Visanta, J.P. Gore, Environ. Combust. Technol. 1 (000) [3] J.R. Howell, M.J. Hall, J.L. Ellzey, Pro. Enery Combust. Sci. (1996) 11. [4] R. Visanta, J.P. Gore, Environ. Combust. Technol. 1 (000) 167. [5] T. Taeno, K. Sato, Combust. Science Technol. 0 (1979) 73.

6 [6] R. Echio, Y. Yoshizawa, K. Hanamura, T. Tomimura.Eihth Symposium (International) Heat Transfer Conference, Hemisphere, New Yor, [7] D.R. Hardesty, F.J. Weinber, Combust. Sci. Technol. 8 (1974) 01. [8] Ton, T. and sathe, S. heat transfer characteristics of porous radiant burners, Trans, Int. J. Heat Mass Transfer, 1990, 33, [9] P-F Hsu, W.D. Evans, J.R. Howell, Combust. Sci. Technol. 90 (1993) 149. [10] Brenner, G. Picenacer, K. Picenacer,O. Trimis, D., Wawrzine, K. and Weber, T. Numerical and experimental investiation of matrix-stabilized methane/air combustion in porous media. Combust. Flame, 000, 13, [11] Taludar, P., Mishra, S., Trimis,D and Durst, F. heat transfer characteristics of porous radiant burner under the influence of a -D radiation field. J. Quantitative spectroscopy & Radiative Transfer, 003,1-11 [1] Taludar, P., Mishra, S. Analysis of conduction radiation problem in absorbin-emittin and anisotropically scatterin media usin the collapsed dimension method, Int. J. Heat Mass Transfer, 00, 45, [13] G.P. Smith, D.M. Golden, M. Frenlach, N.W. Moriarty, B. Eiteneer, M. Goldenber, C.T. Bowman, R.K. Hanson, S. Son, W.C. Gardiner Jr., V. Lissiansi, Z. Qin, available at: [14] P. Glarbor, N. I. Lilleheie, S. Bystoyl, B. F. Manussen, P. Kilpinen and M. Hupa, 4th Symposium (International) on Combustion, The Combustion Institute, pp [15] M.R. Hennee, PhD Dissertation, The University of Texas at Austin, [16] M.R. Hennee, J.L. Ellzey, Combust. Flame 117 (4) (1999) [17] Kee, R. J., Dixon-Lewis, G., Warnatz, J., Coltrin, M. E. and Miler, J. A. (1996) A Fortran Computer Code Pacae for the Evaluation of Gas Phase Multi-component Transport Properties, Sandia National Lab. Report SAND [18] Kee, R. J., Miller, J. A. and Jefferson, T. H. (1980, 1996) CHEMKIN: A General Purpose problem Independent, Transportable, Fortran, Chemical Kinetic Proram Pacae, Sandia National Lab. Report, SAND [19] Kee, R. J., Grcar, J. F., Smooe, M. D. and Miller, J. A. (1996) A Fortran Proram for modelin study laminar one-dimensional premixed flames. Sandiq National Lab Rept., SAND85-840, Albuquerque. [0] Wan, K. Y. and Tien, C.L Thermal Insulation in Flow Systems: Combined Radiation and Convection throuh a Porous Sement J. Heat Transfer, 1984, 106, [1] M. Frenlach, H. Wan, M. Goldenber, G.P. Smith, D.M. Golden, C.T. Bowman, R.K. Hanson, W.C. Gardiner, V. Lissiansi, GRI-Mech an optimized detailed chemical reaction mechanism for methane combustion, Gas Research Institute Topic, Report GRI-95/0058, November 1, [] C.T. Bowman, R.K. Hanson, D.F. Davidson, W.C. Gardiner Jr., V. Lissiansi, G.P. Smith, D.M. Golden, M. Frenlach, M. Goldenber, available at: