- Email: [email protected]
Thermal radiation in oxy-fuel flames
International Journal of Greenhouse Gas Control 5S (2011) S58–S65
Contents lists available at ScienceDirect
International Journal of Greenhouse Gas Control journal homepage: www.elsevier.com/locate/ijggc
Thermal radiation in oxy-fuel flames K. Andersson ∗ , R. Johansson, F. Johnsson Department of Energy and Environment, Division of Energy Technology, Chalmers University of Technology, SE-412 96 Göteborg, Sweden
a r t i c l e
i n f o
Article history: Received 5 November 2010 Received in revised form 9 May 2011 Accepted 13 May 2011 Available online 12 June 2011 Keywords: Oxy-fuel combustion Radiation Particle Gas Scattering Soot
a b s t r a c t This work investigates thermal radiation in oxy-fuel flames, based on experiments and modelling. Experiments were conducted in a 100 kW test facility in air and oxy-fuel combustion atmospheres, using two different types of fuels, lignite and propane. In-flame measurements of gas composition, temperature and total radiation intensity, were performed and used as input to radiation modelling to examine the influence of oxy-fuel conditions on gas and particle radiation characteristics. In the modelling, the spectral properties of CO2 and H2 O are treated by means of a statistical narrow band model and particle radiation is modelled for both scattering and non-scattering particles. Experiments on the propane flame show that the flame radiation conditions are drastically influenced by the recycling conditions. With OF 27 conditions (27% oxygen in the feed gas) and dry recycling, the temperature is slightly lower compared to air-fired conditions, but the emitted intensity is significantly increased. Modelling shows that this is mainly caused by a significantly increased soot radiation. Propane flame images show that the presence of soot in oxy-fuel conditions varies strongly with recycling conditions. The contribution due to an increased emission by CO2 is of minor importance. In the lignite experiments similar flame temperatures were kept during air and oxy-fuel combustion (OF 25 conditions with dry recycling). The measurements show that the intensity levels in both flames are similar which is due to a strong particle radiation in both environments. The modelling reveals that the dominance by particle radiation contra gas radiation is closely related to whether the particles are scattering or non-scattering. © 2011 Elsevier Ltd. All rights reserved.
1. Introduction The design of oxy-fuel boilers will be based on similar criteria as modern air-fired units, where fuel conversion, emission control, material and corrosion issues and heat transfer are key parameters for the design of the furnace. In oxy-fuel firing, the recycling conditions will determine flue gas flow rates, residence time, mixing conditions, gas composition and combustion temperature, which have implications on all of the aforementioned design criteria. The radiative heat transfer is of particular interest with its strong dependence on level and distribution of temperature as well as composition and concentration of combustion gases and particles. High concentrations of CO2 will result from flue gas recycling during oxy-fuel operation, and, if the flue gases are to be kept above the dew point, the H2 O concentrations will also increase compared to air-fired conditions. Unlike N2 , which is transparent to thermal radiation, both CO2 and H2 O emit and absorb thermal radiation. High concentrations of these gases therefore increase the emissivity of the flue gas with subsequent effects on the radiative heat
∗ Corresponding author. Tel.: +46 31 772 5242; fax: +46 31 772 3592. E-mail address: [email protected] (K. Andersson). 1750-5836/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijggc.2011.05.018
transfer. The relative importance of gaseous radiation compared to particle radiation depends on the type of fuel, but also on the combustion conditions. Even when gaseous fuels are burnt there can be a significant influence of soot radiation in the high temperature flame region. Pulverized fuels result in flames dominated by particle radiation, but other regions in the furnace, especially in large-scale facilities, can have different properties, where gas radiation is more important. Gaseous radiation of H2 O and CO2 is characterized by a strong spectral dependence where radiation is emitted and absorbed in specific spectral bands. These bands consist of a large number of spectral lines, which tend to overlap depending on pressure and temperature. At high temperatures even more lines become important and a complete description of high temperature spectrums is complex. Since the properties of the gases are well defined, detailed modelling methods to account for their radiative behaviour are available. Radiative properties of combustion particles are, compared to the properties of the gas species, more difficult to predict since they are highly dependent on the fuel being used and the conditions of the fuel conversion. Particle radiation is composed by contributions from different particle types: coal, char, ash and soot particles. Coal and char particles are mostly present in the flame zone and
K. Andersson et al. / International Journal of Greenhouse Gas Control 5S (2011) S58–S65
S59
Fig. 1. Lignite-fired air and OF 27 flames. Three high resolution photographs, taken consecutively in time, represent the conditions of each flame.
their residence time is coupled to the rate at which those particles are dried, devolatilized and combusted. All three processes: drying, devolatilization and combustion is either directly or in-directly related to oxidant composition and temperature (recycling conditions). Fig. 1 illustrates the differences in ignition and in the early stages of combustion, between OF 27 and air conditions for the lignite flames studied in the present work. In the oxy-fuel flame, a fuel transport gas composition of 30% O2 and 70% CO2 is used instead of air, which results in particles igniting closer to the burner. The faster ignition is related to differences in both devolatilization and particle combustion and hence also to possible differences in particle radiation. The small size in combination with large numbers makes soot particles important contributors to thermal radiation (Ahluwalia and Im, 1994; Fletcher et al., 1997). There are several papers dealing with the formation of soot in gas-fired flames and how soot formation is affected by for example gaseous additives. Experimental studies of the influence on soot formation by CO2 and O2 addition in lab-scale counter-flow diffusion flames (Du et al., 1989) identified three paths through which the soot formation was proposed to be affected by additives: thermal effects caused by a change in flame temperature, dilution effects due to reduction of the concentration of reactive species, and chemical effects caused by the participation of the additive in chemical reactions that interfere with the formation of soot. In order to study the above suggested effects in a flame of technical size, propane-fired tests were carried out in Chalmers 100 kW test unit (Andersson, 2007; Andersson and Johnsson, 2007; Andersson et al., 2008a). Significant differences between the air-fired and oxy-fuel fired flames were observed: Fig. 2 displays photographs of the propane-flames 215 mm from the burner inlet. Both the air-fired flame (a) and the OF 27 flame (c) contain soot, but it appears that the volume fraction of the soot particles increases from air to OF 27 conditions. When the oxygen concentration is increased to 45% the presence of soot is clearly increased. Furthermore, the soot volume fraction is drastically reduced from the air and OF 27 flames to the OF 21 flame with the latter being dominated by its blue/violet color rather than the characteristic soot emission (yellow/orange in color). In this paper, measured intensity data from Andersson et al. (2008a) of the air and OF 27 flames, which have similar temperature conditions, will be examined to study the change in radiative contribution by soot in the two environments (Fig. 3). Coal, char and ash particles are fairly large in size (although the particle size distribution is wide) whereas soot particles are typically nano-sized particles. As larger particles cause scattering, in
contrast to gaseous radiation, calculation of the radiative intensity along one direction must involve the intensity along all other directions, which means that the radiative transfer equations of the different directions have to be solved simultaneously. Due to this vast range of particle types and properties represented in coal combustion, some assumptions usually have to be made regarding particle size, shape and refractive index. As a result, uncertainties in modelling of particle radiation are larger and experiments play a more prominent role for validation of models treating radiation from soot, coal, char and other particles. Therefore, in addition to the separate contribution of gaseous and particle radiation respectively, the effect of different particle types on the radiative intensity in the coal-flames will be examined here. The analysis is performed by means of modelling of both scattering and non-scattering combustion particles including a comparison with experimental data from previous work (Andersson et al., 2008b). The aim of this work is to discuss the particle and gaseous radiation in air and oxy-fuel flames. The paper compares measured profiles of total intensity with modelled gas and particle radiation, for air-fired and oxy-fired cases with dry flue gas recycling for propane and lignite as fuels. Emphasis will be put on the relative contribution of particle radiation compared to gaseous radiation and differences between the fuels and air- respectively oxy-fired conditions. In addition, calculations are performed for both wet and dry recycling conditions to analyse the effect of large fractions of water vapour in atmospheres dominated by particle radiation. 2. Method The combined experimental and modelling method is described below. The experimental data is based on measurements presented in some of our previous work (Andersson et al., 2008a,b) in which further details on the test facility and instrumentation can be found together with more details on the implementation of the gas radiation modelling. 2.1. Experiments Measurements of in-furnace gas composition, temperature and radiation intensity are used both as input to, and, comparison against, the radiation modelling results in this study. All experiments were carried out in the Chalmers 100 kW test unit, which has a down-fired cylindrical refractory-lined reactor which measures 0.8 m in inner diameter and 2.4 m in inner height. The oxygen is mixed into the recycled flue gas up-streams the induced draft
S60
K. Andersson et al. / International Journal of Greenhouse Gas Control 5S (2011) S58–S65
Fig. 2. Photographs of the of the Chalmers flame with propane as a fuel. (a) Air flame, (b) oxy-fuel flame, 21% oxygen in oxidizer and (c) oxy-fuel flame, 27% oxygen in oxidizer, (d) oxy-fuel flame 45% oxygen in the oxidizer, 215 mm from the burner inlet. Images (a)–(c) were previously published in (Andersson et al., 2008a).
Coal from pneumatic feed system Dry, pressurized flue gas as carrier gas Direct O injection
Propane
Primary/ secondary register
air/O /CO fan
Pilot burner
mixing point O /flue gas
Air inlet pre-heater
Cooling tube 1/4
Wet flue gas recycle
Dry flue gas recycle
Measurement ports R1, R2...R7
2400 mm
Cylindrical furnace
Flue gas cooler
Fabric Filter
800 mm
Flue gas condenser
Stack gas
C3H8
O2
Dry, pressurized flue gas for dust control and fuel carrier gas
Cooling water
Fig. 3. Schematic of the experimental unit. All radiation measurements of the present work were performed in measurement port R3, 384 mm down-stream of the burner inlet.
High broadband reflectance mirror
Electronic shutter Collimating tube
Cooling water PT-100
Thermopile
Cooling water
Fig. 4. Schematic of the narrow angle radiometer.
K. Andersson et al. / International Journal of Greenhouse Gas Control 5S (2011) S58–S65
S61
Table 1 Measured gas composition at the inlet and outlet of the combustor. Fuel Lignite Propane
Test case Air OF 25 Air OF 27
O2 [vol.%, dry] In
Out
CO2 [vol.%, dry] In
Out
21 25 21 27
3.1 3.7 3.0 3.8
0 72 0 71
17 94 12 94
fan in the recycle loop. The resulting feed gas mixture then passes the recycle fan after which it is divided into a primary and a secondary stream and can be fed into the burner. Table 1 lists the test conditions. All measurements were performed during dry recycling conditions. The analysis of the lignite is given in Table 2. In the propane combustion experiments the oxygen content in the feed gas was 27% (OF 27) and in lignite combustion it was 25% (OF 25). The reason for choosing these test conditions for the respective fuel, was to obtain similar radial temperature distributions in the flame. For the flames investigated, this occurs at a distance of 384 mm from the burner (the outer burner diameter is 100 mm), a distance for which all data in the present work are presented. The in-furnace gas composition (O2 , CO, CO2 , NO, SO2 and total hydrocarbon concentrations) were measured by means of a gas suction probe inserted in the measurement ports and at the furnace exit (NO and SO2 were not included during combustion of propane). The combustor temperature measurements were performed with two different water-cooled suction pyrometers, one used in the propane tests and the other one used in the lignite tests. The principle is the same for both probes, although they are equipped with different thermocouple types due to that the fuels give different maximum flame temperature. A narrow angle radiometer, Fig. 4, is used for the radiation measurements. The probe measures the line-of-sight radiation intensity by a thermopile in the detector end of the probe. The thermopile is kept at constant temperature by a separate cooling circuit. In order to enable the measurement of the net radiation intensity of the gas layer, background radiation from the furnace wall was avoided either using a non-reflecting quartz window or a cooled bluff body covered with high-temperature resistant black paint as backgrounds. A flow of argon was applied as purge gas through the collimating tube to prevent fouling of the tube and absorption of radiation from combustion products inside the tube. The instrument was calibrated by a black body radiation source of high precision.
Fuel input [kW]
S.R. ()
76 76 80 80
1.18 1.18 1.15 1.15
For gas radiation a statistical narrow-band model has been applied. The parameters used in the model were presented by Soufiani and Taine (1997) and more details regarding the implementation are found in the work by Andersson et al. (2008a,b). Particle radiation is modelled as non-scattering soot in the propane flames, while in the lignite flames it is modelled as either soot or as scattering particles. Particle properties are calculated for each narrow band and the total transmissivity is given by the product of the gas transmissivity and the particle transmissivity. For soot, the transmissivity of band k is calculated as
k ,s = exp
−
36nk 2
(n2 − k2 + 2) + 4n2 k2
k f s
(1)
as given by Modest (2003) where fv is the soot volume fraction and n and k the parameters of the complex index of refraction taken from the polynomials given by Chang and Charalampopoulos (1990). For the scattering particles, correlations presented by Buckius and Hwang (1980) are used to calculate scattering albedo and extinction coefficients. The correlations are expressed with a dimensionless extinction and absorption coefficient, defined as ˇ,p fA
∗ ˇ,p =
∗ ,p =
,p fA
(2)
fA is the projected surface area of the particles and as x → 0 the dimensionless properties are ∗ ∗ = ,0 =− ˇ,0
8rp,av Im
m2 − 1 m2 + 2
(3)
rp,av is the average particle radius, 40 m is used here (the average particle diameter of the experiment coal), and m the refractive index of the particles. The refractive index used in the calculations is based on data of coal samples presented by Foster and Howarth (1968) to which approximate functions have been fitted. Both the correlation describing the extinction coefficient and the correlation of the scattering coefficient are written on the form, 1 1 1 = z + z yz y0 y∞
2.2. Radiation modelling The radiation modelling is based on measured profiles of temperature and concentrations of H2 O and CO2 . The amount of particles is obtained by fitting modelled intensity profiles against the measured intensity profiles. The geometry considered is an infinitely long cylinder where the distribution of species and temperature is assumed to have a radial symmetry. The discrete transfer method with the S6 scheme is used to calculate the radiative intensity field and the weights and ordinates are taken from the work of Tsai and Özis¸ik (1990).
(4)
with parameters taken from Modest (2003). The particle transmissivity and scattering albedo are given by
¯ k ,p = exp −ˇ,p s ωk =
(5)
k ,p
(6)
ˇk ,p + log(¯ k ,g )/scell
The soot volume fraction fv and the particle surface area fA are given by fitting a cosine profile to obtain a good agreement between modelled radiation and the measured profiles of total intensity in each flame. The radiation intensity is given by solving a radiative
Table 2 Analysis of the lausitz lignite. Hi [MJ/kg] (as received) Fuel analysis 20.9
Proximate [wt%, as received] M 10.2
A 5.0
Combustibles 84.8
[%d.a.f]
Ultimate [wt%, d.a.f.]
VM 59.4
C 69.9
H 5.4
N 0.6
S 1.0
O 23.1
S62
K. Andersson et al. / International Journal of Greenhouse Gas Control 5S (2011) S58–S65
Fig. 5. Discretization of a line of sight.
transfer equation (RTE) for each narrow band using the correlated formulation of the RTE. The intensity for the non-scattering soot particles of band k is I¯k ,n = I¯k ,0 ¯ k ,0→n +
I¯bk ,i+1/2
¯ k ,i+1→n − ¯ k ,i→n
(7)
i
and for the scattering particles it is I¯k ,n = I¯k ,0 ¯ k ,0→n +
i
+
ωk ,i+1/2 4
Qscatk ,i+1/2
1 − ωk ,i+1/2 I¯bk ,i+1/2
¯ k ,i+1→n − ¯ k ,i→n
(8)
Index i refers to the spatial discretization of the path as illustrated in Fig. 5, where 0 indicates the starting point of the path and Qscat is the intensity integrated over all directions. 3. Results Fig. 6 presents radiation intensity measurements by the narrow angle radiometer together with modelled intensities in the air and OF 27 propane flames, 384 mm from the burner. The modelled profiles include • gas radiation based on measured temperatures and gas concentrations, • total radiation based on measured temperatures and gas concentrations together with fitted concentrations of soot, and, • profiles of soot radiation based on the fitted soot concentrations and measured temperatures. There is a significant increase in the radiation emitted by the oxy-fuel flame. The temperature measurements, previously reported by Andersson et al. (2008a) show that the temperatures, corresponding to the radiation data in Fig. 6, are slightly lower in the OF 27 case compared to the air-fired case. Thus, the increased intensity level in the OF 27 flame has other explanations. The increase in CO2 concentration, around 90 vol % in oxy-fuel compared to 15 vol % in air-firing, both on dry basis, results in an increased emission of gas radiation according to modelled profiles of gas radiation. By comparison of the results in air and oxy-fired conditions, it is evident that the radiation emitted by soot particles increases drastically in the OF 27 flame. Thus, there is a significant increase in the volume fraction of soot in the OF 27 flame, which alters the radiative conditions compared to the air-fired flame. The obvious difference between the OF 27 and the air case, except for the change from N2 -base to CO2 -base in the oxidant, is that the volumetric flow through the burner is about 20% lower in the OF 27 case. As a consequence, the inlet fuel concentration increases which influences the initial mixing conditions. The
Fig. 6. Measured total radiation intensity with the narrow angle probe, and, modelled total, gas and soot radiation intensity in (a) air-fired conditions and (b) OF 27 conditions. The fuel is propane and the measurements were performed at a distance of 384 mm from the burner.
concentration and the residence time of soot precursors in the nearburner zone may therefore be increased during OF 27 conditions compared to air-firing, which would promote the formation of soot. In general, however, changes in soot formation in oxy-fuel combustion may be due to both chemical, thermal and dilution effects, as discussed. The oxy-fuel experiments presented in Fig. 6 were conducted for dry recycling conditions. Measurements have not been performed for wet recycling conditions. Therefore, to illustrate the influence introduced by wet recycling conditions, an additional set of calculations are presented in Fig. 7 where the ratio between CO2 and H2 O is changed from around 4:1 in dry recycling conditions to 1:1 in wet conditions, while keeping temperature and soot concentration the same as in the case with dry recycling. It is seen that the inclusion of water in the recycle loop has a significant influence on both the peak intensity in the flame as well as the radiation that reaches the combustor wall. It should be noted that the large fraction of H2 O in wet recycling conditions may influence the soot formation, but no such effects are considered in Fig. 7. Fig. 8 presents previously reported data (Andersson et al., 2008b) including total intensity measurements together with radiation modelling results of the lignite flame for both air and OF 25 conditions at a distance of 384 mm down-stream the burner inlet. The modelled profiles are based on the corresponding setup of data
K. Andersson et al. / International Journal of Greenhouse Gas Control 5S (2011) S58–S65
S63
Fig. 7. Modelled total, gas, and soot radiation under wet recycling conditions in OF 27 propane flame. Temperature and soot volume fraction are kept the same as in the OF 27 propane flame with dry recycling conditions. Grey lines are the corresponding lines of the dry recycling OF 27 case.
Fig. 9. Comparison of radiation intensity measurements (384 mm from the burner) and modelling results in (a) air-fired and (b) OF 25 conditions, during combustion of lignite. The extension of the radiative path by the opposite probe port is indicated by the line (−380 mm to 0 mm) and the furnace walls are located at the radial distances 0 and 800 mm. The modelling results include total intensity, with fitted particle concentrations as well as gas radiation and particle radiation from scattering particles.
Fig. 8. Comparison of radiation intensity measurements and modelling results in (a) air-fired and (b) OF 25 conditions, during combustion of lignite. The measurements are performed at a distance of 384 mm from the burner. The extension of the radiative path by the opposite probe port is indicated by the line (−380 mm to 0 mm) and the furnace walls are located at the radial distances 0 and 800 mm. The modelling results include total intensity, with fitted soot concentrations as well as gas and non-scattering soot radiation respectively.
as used for the propane cases. Similar to the propane experiments, the aim was to examine the influence of the CO2 on the radiative conditions. Hence, dry recycling conditions were applied. The measurements were conducted using a quartz window as background (situated at a distance −0.4 m in Fig. 8) to avoid background radiation. In the previous work (Andersson et al., 2008a,b), where a more comprehensive set of experimental data was presented, it was observed that the recycle rate is an effective parameter to control the radiative heat transfer in oxy-fuel combustion and thereby also flame and combustor wall temperature conditions. It was also concluded that the gas temperature profiles are similar for the OF 25 and the air flames. In fact, the radial temperature profiles of air and OF 25, measured in the same probe port as the radiation data in Fig. 8 (384 mm form the burner) almost coincided, differences are less than 20 K in most positions. In Fig. 8, it is therefore interesting to note that the corresponding total radiation intensities of OF 25 and air conditions are similar as well. The gas radiation calculations reveal that the additional CO2 content in the OF 25 flame give a noticeable contribution to the calculated gaseous radiation compared to air-fired conditions, but, also, that this has little or no influence on the measured total intensity. The explanation is found in the radiative contribution by particles. In Fig. 8, the modelled particle radiation is assumed to be radiation emitted by non-scattering
S64
K. Andersson et al. / International Journal of Greenhouse Gas Control 5S (2011) S58–S65
but in this case the relative contribution by the gaseous radiation becomes more important. It is obvious that further work is required in this area to increase the knowledge on coal-derived soot as well as particle radiation characteristics in oxy-fuel flames. The observations made in Figs. 8 and 9 are for dry recycling conditions where it is shown that the CO2 only has a small influence on the radiation emitted by the flame. In Fig. 10, the effect from wet recycling of flue gases is illustrated by modelling results based on a moisture fraction which is increased from about 10% (for dry recycling conditions) to 45% on molar basis. The temperature conditions are kept the same as in the dry recycling tests. In Fig. 10(a) the particle radiation is treated as in Fig. 8b (all particles are approximated as soot) whereas in Fig. 10(b), the particles are scattering similar to the calculations shown in Fig. 9(b). The high moisture content has a stronger impact on the intensity profiles than the CO2 , especially at peak temperature conditions, but there is also a clear effect on the radiation that reaches the combustor wall. In calculations both with non-scattering and scattering particles the contribution by the water vapour is important. In fact, in the case with scattering particles the contribution by particles and gas are of the same magnitude.
4. Summary and conclusions
Fig. 10. Modelled total, gas, and (a) non-scattering soot, and (b) scattering particle radiation under wet recycling conditions. Temperature and particle concentrations are kept the same as in the OF 25 lignite flame with dry recycling conditions. The measurements were done at a distance of 384 mm from the burner. The extension of the radiative path by the opposite probe port is indicated by the line (−380 mm to 0 mm) and the furnace walls are located at the radial distances 0 and 800 mm.
soot particles. The results show that the similarity in the intensity profiles between air and OF 25 is due to the fact that the particle contribution to the measured total intensity signal is dominating in high temperature parts of the flame, which reduces the influence of increased CO2 radiation in the O2 /CO2 environment. As mentioned, a significant simplification was made in the particle radiation calculations in Fig. 8 where all particles were approximated as small, non-scattering, soot particles. However, as discussed, other particle types will contribute to the radiation in addition to soot particles in coal-flames. The relative contribution will vary depending on coal type and combustion conditions. As a comparison to the non-scattering soot, calculations based on scattering particles are presented in Fig. 9. The scattering particles are approximated as coal-particles with an average diameter of 40 m, which is the average particle diameter of the fuel. As the particle projected surface area is fitted to obtain a good agreement between measured and modelled total intensity, the total intensity is not affected by a change of particle type. The aim is instead to illustrate how the contribution of particle radiation is affected by the physical properties of the particles. As seen, for scattering particles, the radiative contribution is reduced compared to non-scattering soot particles. Thus, in this case the particle radiation becomes less dominant; it is still the major contributor to the emitted flame radiation,
The present work examines the radiative conditions in oxy-fuel flames, both for gaseous and solid fuels. The focus of the investigation is to study the influence by CO2 (dry recycling conditions), H2 O (wet recycling conditions) and scattering and non-scattering particles. Experiments and modelling were carried out on lignite and propane-fired flames. Total intensity measurements are used for comparison against modelled gas and particle radiation. The experimental data was chosen to eliminate, as far as possible, the temperature effects by means of choosing the recycling conditions to achieve similar flame temperatures in air and oxy-fired conditions. In the experimental oxy-fuel flames dry flue gas recycling was applied. The propane combustion experiments show significant differences between soot emission in air compared to OF 27 conditions, whereas the increased radiation due to CO2 is pronounced only at peak temperature conditions, i.e. close to the flame boundary. The increased soot levels in the OF 27 flame, completely alters the radiative emission compared to air-fired conditions, and, images of OF 45 conditions clearly indicates a further increase, and a drastic difference, in the amount of soot present compared to airfired conditions. Thus, due to its critical role in flame radiation, soot formation in oxy-fuel combustion needs further exploitation and direct in-flame measurements in flames of technical size together with modelling of soot formation and oxidation is needed to develop an increased understanding. The lignite experiments show that the significant increase in the CO2 content of the oxy-fuel flame did not contribute much to its radiative emission compared to air-fired conditions. This, since the particle radiation dominates the high-temperature zones of both the N2 -based and the CO2 -based atmospheres. Modelling shows that, if all particles are approximated as non-scattering soot particles, the radiative emission by those particles is far stronger than the gaseous radiation, in airfiring as well as in oxy-fuel conditions. If instead all particles are approximated as scattering coal particles, their radiative contribution is reduced, but still dominates in both flames. When wet recycling is applied in oxy-fuel combustion of lignite, the gaseous emission becomes stronger and approaches the contribution by the particles. In fact, in the case with scattering particles, the emission from gas and particles are of similar magnitude, both at peak temperature conditions and at the combustor wall. The influence of H2 O on secondary effects, such as particle radiation characteristics
K. Andersson et al. / International Journal of Greenhouse Gas Control 5S (2011) S58–S65
has not been studied in the present work and requires further experimental studies. Acknowledgements The financial support provided by Vattenfall AB is acknowledged. References Ahluwalia, R.K., Im, K.H.J., 1994. Inst. Energy 67, 23–29. Andersson, K., 2007. Characterization of Oxy-fuel Flames – Their Composition Temperature and Radiation. Chalmers University of Technology: Dissertation, Göteborg, ISBN: 978-91-7385-006-3. Andersson, K., Johnsson, F., 2007. Flame and radiation characteristics of gas-fired O2 /CO2 combustion. Fuel 86 (5–6), 656–668. Andersson, K., Johansson, R., Johnsson, F., Leckner, B., 2008a. Radiation intensity of propane-fired oxy-fuel flames: implications for soot formation. Energy Fuels 22, 1535–1541.
S65
Andersson, K., Johansson, R., Hjärtstam, S., Johnsson, F., Leckner, B., 2008b. Radiation intensity of lignite-fired oxy-fuel flames. Exp. Therm. Fluid Sci. 33, 67–76. Buckius, R.O., Hwang, D.C., 1980. Radiation properties for polydispersions: application to coal. ASME J. Heat Transfer 102, 99–103. Chang, H., Charalampopoulos, T.T., 1990. Determination of the wavelength dependence of refractive indices of flame soot. Proc. R. Soc. Lond. A 430, 577–591. Du, D.X., Axelbaum, R.L., Law, C.K., 1989. Twenty-Third Symposium (Int.) on Combustion. The Combustion Institute, Pittsburgh, PA, pp. 1501–1507. Fletcher, T.H., Ma, J., Rigby, J.R., Brown, A.L., Webb, B.W., 1997. Prog. Energy Combust. Sci. 23, 283–301. Foster, P.J., Howarth, C.R., 1968. Optical constants of carbons and coals in the infrared. Carbon 6, 719–729. Modest, M.F., 2003. Radiative Heat Transfer, 2nd ed. Academic Press. Soufiani, A., Taine, J., 1997. High temperature gas radiative property parameters of statistical narrow-band model for H2 O CO2 and CO, and correlated-K model for H2 O and CO2 . Int. J. Heat Mass Transfer 40, 987–991. Tsai, J.R., Özis¸ik, M.N., 1990. Radiation in cylindrical symmetry with anisotropic scattering and variable properties. Int. J. Heat Mass Transfer 33, 2651– 2658.
Contents lists available at ScienceDirect
International Journal of Greenhouse Gas Control journal homepage: www.elsevier.com/locate/ijggc
Thermal radiation in oxy-fuel flames K. Andersson ∗ , R. Johansson, F. Johnsson Department of Energy and Environment, Division of Energy Technology, Chalmers University of Technology, SE-412 96 Göteborg, Sweden
a r t i c l e
i n f o
Article history: Received 5 November 2010 Received in revised form 9 May 2011 Accepted 13 May 2011 Available online 12 June 2011 Keywords: Oxy-fuel combustion Radiation Particle Gas Scattering Soot
a b s t r a c t This work investigates thermal radiation in oxy-fuel flames, based on experiments and modelling. Experiments were conducted in a 100 kW test facility in air and oxy-fuel combustion atmospheres, using two different types of fuels, lignite and propane. In-flame measurements of gas composition, temperature and total radiation intensity, were performed and used as input to radiation modelling to examine the influence of oxy-fuel conditions on gas and particle radiation characteristics. In the modelling, the spectral properties of CO2 and H2 O are treated by means of a statistical narrow band model and particle radiation is modelled for both scattering and non-scattering particles. Experiments on the propane flame show that the flame radiation conditions are drastically influenced by the recycling conditions. With OF 27 conditions (27% oxygen in the feed gas) and dry recycling, the temperature is slightly lower compared to air-fired conditions, but the emitted intensity is significantly increased. Modelling shows that this is mainly caused by a significantly increased soot radiation. Propane flame images show that the presence of soot in oxy-fuel conditions varies strongly with recycling conditions. The contribution due to an increased emission by CO2 is of minor importance. In the lignite experiments similar flame temperatures were kept during air and oxy-fuel combustion (OF 25 conditions with dry recycling). The measurements show that the intensity levels in both flames are similar which is due to a strong particle radiation in both environments. The modelling reveals that the dominance by particle radiation contra gas radiation is closely related to whether the particles are scattering or non-scattering. © 2011 Elsevier Ltd. All rights reserved.
1. Introduction The design of oxy-fuel boilers will be based on similar criteria as modern air-fired units, where fuel conversion, emission control, material and corrosion issues and heat transfer are key parameters for the design of the furnace. In oxy-fuel firing, the recycling conditions will determine flue gas flow rates, residence time, mixing conditions, gas composition and combustion temperature, which have implications on all of the aforementioned design criteria. The radiative heat transfer is of particular interest with its strong dependence on level and distribution of temperature as well as composition and concentration of combustion gases and particles. High concentrations of CO2 will result from flue gas recycling during oxy-fuel operation, and, if the flue gases are to be kept above the dew point, the H2 O concentrations will also increase compared to air-fired conditions. Unlike N2 , which is transparent to thermal radiation, both CO2 and H2 O emit and absorb thermal radiation. High concentrations of these gases therefore increase the emissivity of the flue gas with subsequent effects on the radiative heat
∗ Corresponding author. Tel.: +46 31 772 5242; fax: +46 31 772 3592. E-mail address: [email protected] (K. Andersson). 1750-5836/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijggc.2011.05.018
transfer. The relative importance of gaseous radiation compared to particle radiation depends on the type of fuel, but also on the combustion conditions. Even when gaseous fuels are burnt there can be a significant influence of soot radiation in the high temperature flame region. Pulverized fuels result in flames dominated by particle radiation, but other regions in the furnace, especially in large-scale facilities, can have different properties, where gas radiation is more important. Gaseous radiation of H2 O and CO2 is characterized by a strong spectral dependence where radiation is emitted and absorbed in specific spectral bands. These bands consist of a large number of spectral lines, which tend to overlap depending on pressure and temperature. At high temperatures even more lines become important and a complete description of high temperature spectrums is complex. Since the properties of the gases are well defined, detailed modelling methods to account for their radiative behaviour are available. Radiative properties of combustion particles are, compared to the properties of the gas species, more difficult to predict since they are highly dependent on the fuel being used and the conditions of the fuel conversion. Particle radiation is composed by contributions from different particle types: coal, char, ash and soot particles. Coal and char particles are mostly present in the flame zone and
K. Andersson et al. / International Journal of Greenhouse Gas Control 5S (2011) S58–S65
S59
Fig. 1. Lignite-fired air and OF 27 flames. Three high resolution photographs, taken consecutively in time, represent the conditions of each flame.
their residence time is coupled to the rate at which those particles are dried, devolatilized and combusted. All three processes: drying, devolatilization and combustion is either directly or in-directly related to oxidant composition and temperature (recycling conditions). Fig. 1 illustrates the differences in ignition and in the early stages of combustion, between OF 27 and air conditions for the lignite flames studied in the present work. In the oxy-fuel flame, a fuel transport gas composition of 30% O2 and 70% CO2 is used instead of air, which results in particles igniting closer to the burner. The faster ignition is related to differences in both devolatilization and particle combustion and hence also to possible differences in particle radiation. The small size in combination with large numbers makes soot particles important contributors to thermal radiation (Ahluwalia and Im, 1994; Fletcher et al., 1997). There are several papers dealing with the formation of soot in gas-fired flames and how soot formation is affected by for example gaseous additives. Experimental studies of the influence on soot formation by CO2 and O2 addition in lab-scale counter-flow diffusion flames (Du et al., 1989) identified three paths through which the soot formation was proposed to be affected by additives: thermal effects caused by a change in flame temperature, dilution effects due to reduction of the concentration of reactive species, and chemical effects caused by the participation of the additive in chemical reactions that interfere with the formation of soot. In order to study the above suggested effects in a flame of technical size, propane-fired tests were carried out in Chalmers 100 kW test unit (Andersson, 2007; Andersson and Johnsson, 2007; Andersson et al., 2008a). Significant differences between the air-fired and oxy-fuel fired flames were observed: Fig. 2 displays photographs of the propane-flames 215 mm from the burner inlet. Both the air-fired flame (a) and the OF 27 flame (c) contain soot, but it appears that the volume fraction of the soot particles increases from air to OF 27 conditions. When the oxygen concentration is increased to 45% the presence of soot is clearly increased. Furthermore, the soot volume fraction is drastically reduced from the air and OF 27 flames to the OF 21 flame with the latter being dominated by its blue/violet color rather than the characteristic soot emission (yellow/orange in color). In this paper, measured intensity data from Andersson et al. (2008a) of the air and OF 27 flames, which have similar temperature conditions, will be examined to study the change in radiative contribution by soot in the two environments (Fig. 3). Coal, char and ash particles are fairly large in size (although the particle size distribution is wide) whereas soot particles are typically nano-sized particles. As larger particles cause scattering, in
contrast to gaseous radiation, calculation of the radiative intensity along one direction must involve the intensity along all other directions, which means that the radiative transfer equations of the different directions have to be solved simultaneously. Due to this vast range of particle types and properties represented in coal combustion, some assumptions usually have to be made regarding particle size, shape and refractive index. As a result, uncertainties in modelling of particle radiation are larger and experiments play a more prominent role for validation of models treating radiation from soot, coal, char and other particles. Therefore, in addition to the separate contribution of gaseous and particle radiation respectively, the effect of different particle types on the radiative intensity in the coal-flames will be examined here. The analysis is performed by means of modelling of both scattering and non-scattering combustion particles including a comparison with experimental data from previous work (Andersson et al., 2008b). The aim of this work is to discuss the particle and gaseous radiation in air and oxy-fuel flames. The paper compares measured profiles of total intensity with modelled gas and particle radiation, for air-fired and oxy-fired cases with dry flue gas recycling for propane and lignite as fuels. Emphasis will be put on the relative contribution of particle radiation compared to gaseous radiation and differences between the fuels and air- respectively oxy-fired conditions. In addition, calculations are performed for both wet and dry recycling conditions to analyse the effect of large fractions of water vapour in atmospheres dominated by particle radiation. 2. Method The combined experimental and modelling method is described below. The experimental data is based on measurements presented in some of our previous work (Andersson et al., 2008a,b) in which further details on the test facility and instrumentation can be found together with more details on the implementation of the gas radiation modelling. 2.1. Experiments Measurements of in-furnace gas composition, temperature and radiation intensity are used both as input to, and, comparison against, the radiation modelling results in this study. All experiments were carried out in the Chalmers 100 kW test unit, which has a down-fired cylindrical refractory-lined reactor which measures 0.8 m in inner diameter and 2.4 m in inner height. The oxygen is mixed into the recycled flue gas up-streams the induced draft
S60
K. Andersson et al. / International Journal of Greenhouse Gas Control 5S (2011) S58–S65
Fig. 2. Photographs of the of the Chalmers flame with propane as a fuel. (a) Air flame, (b) oxy-fuel flame, 21% oxygen in oxidizer and (c) oxy-fuel flame, 27% oxygen in oxidizer, (d) oxy-fuel flame 45% oxygen in the oxidizer, 215 mm from the burner inlet. Images (a)–(c) were previously published in (Andersson et al., 2008a).
Coal from pneumatic feed system Dry, pressurized flue gas as carrier gas Direct O injection
Propane
Primary/ secondary register
air/O /CO fan
Pilot burner
mixing point O /flue gas
Air inlet pre-heater
Cooling tube 1/4
Wet flue gas recycle
Dry flue gas recycle
Measurement ports R1, R2...R7
2400 mm
Cylindrical furnace
Flue gas cooler
Fabric Filter
800 mm
Flue gas condenser
Stack gas
C3H8
O2
Dry, pressurized flue gas for dust control and fuel carrier gas
Cooling water
Fig. 3. Schematic of the experimental unit. All radiation measurements of the present work were performed in measurement port R3, 384 mm down-stream of the burner inlet.
High broadband reflectance mirror
Electronic shutter Collimating tube
Cooling water PT-100
Thermopile
Cooling water
Fig. 4. Schematic of the narrow angle radiometer.
K. Andersson et al. / International Journal of Greenhouse Gas Control 5S (2011) S58–S65
S61
Table 1 Measured gas composition at the inlet and outlet of the combustor. Fuel Lignite Propane
Test case Air OF 25 Air OF 27
O2 [vol.%, dry] In
Out
CO2 [vol.%, dry] In
Out
21 25 21 27
3.1 3.7 3.0 3.8
0 72 0 71
17 94 12 94
fan in the recycle loop. The resulting feed gas mixture then passes the recycle fan after which it is divided into a primary and a secondary stream and can be fed into the burner. Table 1 lists the test conditions. All measurements were performed during dry recycling conditions. The analysis of the lignite is given in Table 2. In the propane combustion experiments the oxygen content in the feed gas was 27% (OF 27) and in lignite combustion it was 25% (OF 25). The reason for choosing these test conditions for the respective fuel, was to obtain similar radial temperature distributions in the flame. For the flames investigated, this occurs at a distance of 384 mm from the burner (the outer burner diameter is 100 mm), a distance for which all data in the present work are presented. The in-furnace gas composition (O2 , CO, CO2 , NO, SO2 and total hydrocarbon concentrations) were measured by means of a gas suction probe inserted in the measurement ports and at the furnace exit (NO and SO2 were not included during combustion of propane). The combustor temperature measurements were performed with two different water-cooled suction pyrometers, one used in the propane tests and the other one used in the lignite tests. The principle is the same for both probes, although they are equipped with different thermocouple types due to that the fuels give different maximum flame temperature. A narrow angle radiometer, Fig. 4, is used for the radiation measurements. The probe measures the line-of-sight radiation intensity by a thermopile in the detector end of the probe. The thermopile is kept at constant temperature by a separate cooling circuit. In order to enable the measurement of the net radiation intensity of the gas layer, background radiation from the furnace wall was avoided either using a non-reflecting quartz window or a cooled bluff body covered with high-temperature resistant black paint as backgrounds. A flow of argon was applied as purge gas through the collimating tube to prevent fouling of the tube and absorption of radiation from combustion products inside the tube. The instrument was calibrated by a black body radiation source of high precision.
Fuel input [kW]
S.R. ()
76 76 80 80
1.18 1.18 1.15 1.15
For gas radiation a statistical narrow-band model has been applied. The parameters used in the model were presented by Soufiani and Taine (1997) and more details regarding the implementation are found in the work by Andersson et al. (2008a,b). Particle radiation is modelled as non-scattering soot in the propane flames, while in the lignite flames it is modelled as either soot or as scattering particles. Particle properties are calculated for each narrow band and the total transmissivity is given by the product of the gas transmissivity and the particle transmissivity. For soot, the transmissivity of band k is calculated as
k ,s = exp
−
36nk 2
(n2 − k2 + 2) + 4n2 k2
k f s
(1)
as given by Modest (2003) where fv is the soot volume fraction and n and k the parameters of the complex index of refraction taken from the polynomials given by Chang and Charalampopoulos (1990). For the scattering particles, correlations presented by Buckius and Hwang (1980) are used to calculate scattering albedo and extinction coefficients. The correlations are expressed with a dimensionless extinction and absorption coefficient, defined as ˇ,p fA
∗ ˇ,p =
∗ ,p =
,p fA
(2)
fA is the projected surface area of the particles and as x → 0 the dimensionless properties are ∗ ∗ = ,0 =− ˇ,0
8rp,av Im
m2 − 1 m2 + 2
(3)
rp,av is the average particle radius, 40 m is used here (the average particle diameter of the experiment coal), and m the refractive index of the particles. The refractive index used in the calculations is based on data of coal samples presented by Foster and Howarth (1968) to which approximate functions have been fitted. Both the correlation describing the extinction coefficient and the correlation of the scattering coefficient are written on the form, 1 1 1 = z + z yz y0 y∞
2.2. Radiation modelling The radiation modelling is based on measured profiles of temperature and concentrations of H2 O and CO2 . The amount of particles is obtained by fitting modelled intensity profiles against the measured intensity profiles. The geometry considered is an infinitely long cylinder where the distribution of species and temperature is assumed to have a radial symmetry. The discrete transfer method with the S6 scheme is used to calculate the radiative intensity field and the weights and ordinates are taken from the work of Tsai and Özis¸ik (1990).
(4)
with parameters taken from Modest (2003). The particle transmissivity and scattering albedo are given by
¯ k ,p = exp −ˇ,p s ωk =
(5)
k ,p
(6)
ˇk ,p + log(¯ k ,g )/scell
The soot volume fraction fv and the particle surface area fA are given by fitting a cosine profile to obtain a good agreement between modelled radiation and the measured profiles of total intensity in each flame. The radiation intensity is given by solving a radiative
Table 2 Analysis of the lausitz lignite. Hi [MJ/kg] (as received) Fuel analysis 20.9
Proximate [wt%, as received] M 10.2
A 5.0
Combustibles 84.8
[%d.a.f]
Ultimate [wt%, d.a.f.]
VM 59.4
C 69.9
H 5.4
N 0.6
S 1.0
O 23.1
S62
K. Andersson et al. / International Journal of Greenhouse Gas Control 5S (2011) S58–S65
Fig. 5. Discretization of a line of sight.
transfer equation (RTE) for each narrow band using the correlated formulation of the RTE. The intensity for the non-scattering soot particles of band k is I¯k ,n = I¯k ,0 ¯ k ,0→n +
I¯bk ,i+1/2
¯ k ,i+1→n − ¯ k ,i→n
(7)
i
and for the scattering particles it is I¯k ,n = I¯k ,0 ¯ k ,0→n +
i
+
ωk ,i+1/2 4
Qscatk ,i+1/2
1 − ωk ,i+1/2 I¯bk ,i+1/2
¯ k ,i+1→n − ¯ k ,i→n
(8)
Index i refers to the spatial discretization of the path as illustrated in Fig. 5, where 0 indicates the starting point of the path and Qscat is the intensity integrated over all directions. 3. Results Fig. 6 presents radiation intensity measurements by the narrow angle radiometer together with modelled intensities in the air and OF 27 propane flames, 384 mm from the burner. The modelled profiles include • gas radiation based on measured temperatures and gas concentrations, • total radiation based on measured temperatures and gas concentrations together with fitted concentrations of soot, and, • profiles of soot radiation based on the fitted soot concentrations and measured temperatures. There is a significant increase in the radiation emitted by the oxy-fuel flame. The temperature measurements, previously reported by Andersson et al. (2008a) show that the temperatures, corresponding to the radiation data in Fig. 6, are slightly lower in the OF 27 case compared to the air-fired case. Thus, the increased intensity level in the OF 27 flame has other explanations. The increase in CO2 concentration, around 90 vol % in oxy-fuel compared to 15 vol % in air-firing, both on dry basis, results in an increased emission of gas radiation according to modelled profiles of gas radiation. By comparison of the results in air and oxy-fired conditions, it is evident that the radiation emitted by soot particles increases drastically in the OF 27 flame. Thus, there is a significant increase in the volume fraction of soot in the OF 27 flame, which alters the radiative conditions compared to the air-fired flame. The obvious difference between the OF 27 and the air case, except for the change from N2 -base to CO2 -base in the oxidant, is that the volumetric flow through the burner is about 20% lower in the OF 27 case. As a consequence, the inlet fuel concentration increases which influences the initial mixing conditions. The
Fig. 6. Measured total radiation intensity with the narrow angle probe, and, modelled total, gas and soot radiation intensity in (a) air-fired conditions and (b) OF 27 conditions. The fuel is propane and the measurements were performed at a distance of 384 mm from the burner.
concentration and the residence time of soot precursors in the nearburner zone may therefore be increased during OF 27 conditions compared to air-firing, which would promote the formation of soot. In general, however, changes in soot formation in oxy-fuel combustion may be due to both chemical, thermal and dilution effects, as discussed. The oxy-fuel experiments presented in Fig. 6 were conducted for dry recycling conditions. Measurements have not been performed for wet recycling conditions. Therefore, to illustrate the influence introduced by wet recycling conditions, an additional set of calculations are presented in Fig. 7 where the ratio between CO2 and H2 O is changed from around 4:1 in dry recycling conditions to 1:1 in wet conditions, while keeping temperature and soot concentration the same as in the case with dry recycling. It is seen that the inclusion of water in the recycle loop has a significant influence on both the peak intensity in the flame as well as the radiation that reaches the combustor wall. It should be noted that the large fraction of H2 O in wet recycling conditions may influence the soot formation, but no such effects are considered in Fig. 7. Fig. 8 presents previously reported data (Andersson et al., 2008b) including total intensity measurements together with radiation modelling results of the lignite flame for both air and OF 25 conditions at a distance of 384 mm down-stream the burner inlet. The modelled profiles are based on the corresponding setup of data
K. Andersson et al. / International Journal of Greenhouse Gas Control 5S (2011) S58–S65
S63
Fig. 7. Modelled total, gas, and soot radiation under wet recycling conditions in OF 27 propane flame. Temperature and soot volume fraction are kept the same as in the OF 27 propane flame with dry recycling conditions. Grey lines are the corresponding lines of the dry recycling OF 27 case.
Fig. 9. Comparison of radiation intensity measurements (384 mm from the burner) and modelling results in (a) air-fired and (b) OF 25 conditions, during combustion of lignite. The extension of the radiative path by the opposite probe port is indicated by the line (−380 mm to 0 mm) and the furnace walls are located at the radial distances 0 and 800 mm. The modelling results include total intensity, with fitted particle concentrations as well as gas radiation and particle radiation from scattering particles.
Fig. 8. Comparison of radiation intensity measurements and modelling results in (a) air-fired and (b) OF 25 conditions, during combustion of lignite. The measurements are performed at a distance of 384 mm from the burner. The extension of the radiative path by the opposite probe port is indicated by the line (−380 mm to 0 mm) and the furnace walls are located at the radial distances 0 and 800 mm. The modelling results include total intensity, with fitted soot concentrations as well as gas and non-scattering soot radiation respectively.
as used for the propane cases. Similar to the propane experiments, the aim was to examine the influence of the CO2 on the radiative conditions. Hence, dry recycling conditions were applied. The measurements were conducted using a quartz window as background (situated at a distance −0.4 m in Fig. 8) to avoid background radiation. In the previous work (Andersson et al., 2008a,b), where a more comprehensive set of experimental data was presented, it was observed that the recycle rate is an effective parameter to control the radiative heat transfer in oxy-fuel combustion and thereby also flame and combustor wall temperature conditions. It was also concluded that the gas temperature profiles are similar for the OF 25 and the air flames. In fact, the radial temperature profiles of air and OF 25, measured in the same probe port as the radiation data in Fig. 8 (384 mm form the burner) almost coincided, differences are less than 20 K in most positions. In Fig. 8, it is therefore interesting to note that the corresponding total radiation intensities of OF 25 and air conditions are similar as well. The gas radiation calculations reveal that the additional CO2 content in the OF 25 flame give a noticeable contribution to the calculated gaseous radiation compared to air-fired conditions, but, also, that this has little or no influence on the measured total intensity. The explanation is found in the radiative contribution by particles. In Fig. 8, the modelled particle radiation is assumed to be radiation emitted by non-scattering
S64
K. Andersson et al. / International Journal of Greenhouse Gas Control 5S (2011) S58–S65
but in this case the relative contribution by the gaseous radiation becomes more important. It is obvious that further work is required in this area to increase the knowledge on coal-derived soot as well as particle radiation characteristics in oxy-fuel flames. The observations made in Figs. 8 and 9 are for dry recycling conditions where it is shown that the CO2 only has a small influence on the radiation emitted by the flame. In Fig. 10, the effect from wet recycling of flue gases is illustrated by modelling results based on a moisture fraction which is increased from about 10% (for dry recycling conditions) to 45% on molar basis. The temperature conditions are kept the same as in the dry recycling tests. In Fig. 10(a) the particle radiation is treated as in Fig. 8b (all particles are approximated as soot) whereas in Fig. 10(b), the particles are scattering similar to the calculations shown in Fig. 9(b). The high moisture content has a stronger impact on the intensity profiles than the CO2 , especially at peak temperature conditions, but there is also a clear effect on the radiation that reaches the combustor wall. In calculations both with non-scattering and scattering particles the contribution by the water vapour is important. In fact, in the case with scattering particles the contribution by particles and gas are of the same magnitude.
4. Summary and conclusions
Fig. 10. Modelled total, gas, and (a) non-scattering soot, and (b) scattering particle radiation under wet recycling conditions. Temperature and particle concentrations are kept the same as in the OF 25 lignite flame with dry recycling conditions. The measurements were done at a distance of 384 mm from the burner. The extension of the radiative path by the opposite probe port is indicated by the line (−380 mm to 0 mm) and the furnace walls are located at the radial distances 0 and 800 mm.
soot particles. The results show that the similarity in the intensity profiles between air and OF 25 is due to the fact that the particle contribution to the measured total intensity signal is dominating in high temperature parts of the flame, which reduces the influence of increased CO2 radiation in the O2 /CO2 environment. As mentioned, a significant simplification was made in the particle radiation calculations in Fig. 8 where all particles were approximated as small, non-scattering, soot particles. However, as discussed, other particle types will contribute to the radiation in addition to soot particles in coal-flames. The relative contribution will vary depending on coal type and combustion conditions. As a comparison to the non-scattering soot, calculations based on scattering particles are presented in Fig. 9. The scattering particles are approximated as coal-particles with an average diameter of 40 m, which is the average particle diameter of the fuel. As the particle projected surface area is fitted to obtain a good agreement between measured and modelled total intensity, the total intensity is not affected by a change of particle type. The aim is instead to illustrate how the contribution of particle radiation is affected by the physical properties of the particles. As seen, for scattering particles, the radiative contribution is reduced compared to non-scattering soot particles. Thus, in this case the particle radiation becomes less dominant; it is still the major contributor to the emitted flame radiation,
The present work examines the radiative conditions in oxy-fuel flames, both for gaseous and solid fuels. The focus of the investigation is to study the influence by CO2 (dry recycling conditions), H2 O (wet recycling conditions) and scattering and non-scattering particles. Experiments and modelling were carried out on lignite and propane-fired flames. Total intensity measurements are used for comparison against modelled gas and particle radiation. The experimental data was chosen to eliminate, as far as possible, the temperature effects by means of choosing the recycling conditions to achieve similar flame temperatures in air and oxy-fired conditions. In the experimental oxy-fuel flames dry flue gas recycling was applied. The propane combustion experiments show significant differences between soot emission in air compared to OF 27 conditions, whereas the increased radiation due to CO2 is pronounced only at peak temperature conditions, i.e. close to the flame boundary. The increased soot levels in the OF 27 flame, completely alters the radiative emission compared to air-fired conditions, and, images of OF 45 conditions clearly indicates a further increase, and a drastic difference, in the amount of soot present compared to airfired conditions. Thus, due to its critical role in flame radiation, soot formation in oxy-fuel combustion needs further exploitation and direct in-flame measurements in flames of technical size together with modelling of soot formation and oxidation is needed to develop an increased understanding. The lignite experiments show that the significant increase in the CO2 content of the oxy-fuel flame did not contribute much to its radiative emission compared to air-fired conditions. This, since the particle radiation dominates the high-temperature zones of both the N2 -based and the CO2 -based atmospheres. Modelling shows that, if all particles are approximated as non-scattering soot particles, the radiative emission by those particles is far stronger than the gaseous radiation, in airfiring as well as in oxy-fuel conditions. If instead all particles are approximated as scattering coal particles, their radiative contribution is reduced, but still dominates in both flames. When wet recycling is applied in oxy-fuel combustion of lignite, the gaseous emission becomes stronger and approaches the contribution by the particles. In fact, in the case with scattering particles, the emission from gas and particles are of similar magnitude, both at peak temperature conditions and at the combustor wall. The influence of H2 O on secondary effects, such as particle radiation characteristics
K. Andersson et al. / International Journal of Greenhouse Gas Control 5S (2011) S58–S65
has not been studied in the present work and requires further experimental studies. Acknowledgements The financial support provided by Vattenfall AB is acknowledged. References Ahluwalia, R.K., Im, K.H.J., 1994. Inst. Energy 67, 23–29. Andersson, K., 2007. Characterization of Oxy-fuel Flames – Their Composition Temperature and Radiation. Chalmers University of Technology: Dissertation, Göteborg, ISBN: 978-91-7385-006-3. Andersson, K., Johnsson, F., 2007. Flame and radiation characteristics of gas-fired O2 /CO2 combustion. Fuel 86 (5–6), 656–668. Andersson, K., Johansson, R., Johnsson, F., Leckner, B., 2008a. Radiation intensity of propane-fired oxy-fuel flames: implications for soot formation. Energy Fuels 22, 1535–1541.
S65
Andersson, K., Johansson, R., Hjärtstam, S., Johnsson, F., Leckner, B., 2008b. Radiation intensity of lignite-fired oxy-fuel flames. Exp. Therm. Fluid Sci. 33, 67–76. Buckius, R.O., Hwang, D.C., 1980. Radiation properties for polydispersions: application to coal. ASME J. Heat Transfer 102, 99–103. Chang, H., Charalampopoulos, T.T., 1990. Determination of the wavelength dependence of refractive indices of flame soot. Proc. R. Soc. Lond. A 430, 577–591. Du, D.X., Axelbaum, R.L., Law, C.K., 1989. Twenty-Third Symposium (Int.) on Combustion. The Combustion Institute, Pittsburgh, PA, pp. 1501–1507. Fletcher, T.H., Ma, J., Rigby, J.R., Brown, A.L., Webb, B.W., 1997. Prog. Energy Combust. Sci. 23, 283–301. Foster, P.J., Howarth, C.R., 1968. Optical constants of carbons and coals in the infrared. Carbon 6, 719–729. Modest, M.F., 2003. Radiative Heat Transfer, 2nd ed. Academic Press. Soufiani, A., Taine, J., 1997. High temperature gas radiative property parameters of statistical narrow-band model for H2 O CO2 and CO, and correlated-K model for H2 O and CO2 . Int. J. Heat Mass Transfer 40, 987–991. Tsai, J.R., Özis¸ik, M.N., 1990. Radiation in cylindrical symmetry with anisotropic scattering and variable properties. Int. J. Heat Mass Transfer 33, 2651– 2658.