- Email: [email protected]
Combustion of hydrogen-enriched methane in a lean premixed swirl-stabilized burner
Proceedings of the Combustion Institute, Volume 29, 2002/pp. 843–851
COMBUSTION OF HYDROGEN-ENRICHED METHANE IN A LEAN PREMIXED SWIRL-STABILIZED BURNER R. W. SCHEFER,1 D. M. WICKSALL2 and A. K. AGRAWAL2 1
Combustion Research Facility Sandia National Laboratories Livermore, CA 94551-0969, USA 2 School of Aerospace and Mechanical Engineering University of Oklahoma Norman, OK 73019, USA
The combustion characteristics of a premixed, swirl-stabilized flame were studied to determine the effects of enriching methane with hydrogen under fuel-lean conditions. The burner consisted of a centerbody with an annular, premixed fuel-air jet. Swirl was introduced to the flow using 45-degree swirl vanes. The combustion occurred within an air-cooled quartz chamber at atmospheric pressure. Flame stability and blowout maps were obtained for different amounts of hydrogen addition at several fuel-air flow rates. Gas probe measurements were obtained to demonstrate reductions in CO concentration with hydrogen addition, without adversely affecting the NOx emissions. The flame structure near the lean stability limit was described by direct luminous photographs and planar laser-induced fluorescence measurements of the OH radical. Results show that the addition of a moderate amount of hydrogen to the methane/air mixture increased the peak OH concentration. Hydrogen addition resulted in a significant change in the flame structure, indicated by a shorter and more robust appearing flame. The observed trends concur with the strained opposed premixed flame analysis using RUN-1DL. The computations revealed that enriching the methane with hydrogen increased the strain resistance of the flame as well as the OH levels in the flame.
Introduction The development of advanced combustion capabilities for gaseous hydrogen and hydrogen-blended hydrocarbon fuels in gas turbine applications is an area of much current interest encompassing several needs. One need is the cost-effective utilization of alternative fuels with a wide range of heating values [1]. Low- and medium-heating-value fuels containing hydrogen are often produced as by-products in Coal-Gasification Combined Cycle and Fluidized Bed Combustion installations. These product gases could provide a significant source of cost-effective fuels for power-generating gas turbines. A second need is related to the recognition that ultralean premixed combustion is effective in reducing the NOx emissions. Hydrogen blended with traditional hydrocarbon fuels could improve flame stability and allow stable lean combustion at the lower temperatures needed to minimize the NOx production [2]. A longer-term need is the desire to minimize and eventually eliminate unburned hydrocarbon (UHC) and CO2 emissions. The use of hydrogen-blended fuels provides both a solution to the immediate need for NOx reduction, and a transition strategy to a carbon-free energy system in the future. Fuels containing hydrogen have been used for some time in conventional non-premixed combustion systems [3]. Premixed combustion studies in
spark-ignition engines operating on hythane, a blend of hydrogen and natural gas, show lower exhaust emissions and a leaner operating limit [4,5]. An early study on lean premixed (LPM) combustion of propane with air showed improved combustion efficiency and lower temperature at lean blowout with hydrogen addition [6]. Using hydrogen as a pilot fuel was shown to improve the lean-stability limit in LPM combustion systems for gas turbines [7]. In a recent study, fuels containing up to 10% hydrogen were utilized in a commercial gas turbine [8]. Conceptual studies of large-scale gas turbine-based power plants have concluded that the hydrogen addition can be used to improve the combustor performance [9]. At a more fundamental level, research on hydrogen-enriched fuels has been conducted using an opposed jet flow configuration [10–12]. Based on the results of strained-flame calculations at an elevated pressure (30 atm), the improved lean flame stability was attributed to hydrogen’s higher flame speed and increased resistance to strain [10]. The presence of hydrogen substantially increases the extinction strain rate, an effect which is more pronounced in leaner mixtures [11]. Sensitivity analyses show that the chain-branching reaction H Ⳮ O2 } OH Ⳮ O is important with respect to extinction [2,11]. Furthermore, the OH and O radicals contribute to the pyrolysis of the carbon species. The reduction in CO
843
844
ENERGY PRODUCTION—Swirl-Stabilized Combustion
emissions with hydrogen addition was also attributed to the increased radical pool. Higher OH concentrations are likely to promote completion of CO oxidation to CO2 via the OH radical. The primary objective of this study is to delineate the role of hydrogen addition in a LPM combustor representative of advanced gas turbines. In the remainder of the paper, the experimental system and diagnostics will be described. Results characterizing the combustion stability and emissions characteristics of a laboratory-scale, swirl-stabilized LPM combustor will be presented. The fuel will consist of methane or natural gas enriched with hydrogen. The flame structure will be characterized using direct luminosity photographs and OH planar laser-induced fluorescence (PLIF) images. In particular, the flame behavior will be characterized under fuel lean conditions near blowout using measurements of instantaneous and time-averaged OH concentrations. Strained premixed flame analysis in an opposed-flow configuration will be used to explain the observed trends. Experimental System Burner Description and Flow Conditions The test apparatus is a laboratory-scale, swirlingflow dump combustor that consists of a premixer and a combustion chamber. The inside diameter of the premixer tube is 4.1 cm and the outer diameter of the fuel centerbody is 2.5 cm. Premixed conditions were achieved by injecting fuel at 40.0 cm upstream of the combustor inlet. The flame is stabilized on a centerbody with seven 45-degree swirler vanes, located at the end of the mixing chamber. The average axial velocity at the swirler exit varied from 8.0 to 22.0 m/s for the fuel/air mixture flow rate of 400 to 1100 standard liter per minute. Although the detailed flow measurements were not obtained, turbulence intensity levels of up to 150% were reported by Ji and Gore [13] in a similar setup. The estimated strain rate based on the bulk-averaged inlet velocity and the combustor inlet length scale ranged between 400 and 2500 sⳮ1. The combustion chamber is a 30.5-cm-long quartz tube of 8.3-cm inside diameter. Combustion air was provided by an air compressor and metered upstream of the burner using either a mass flowmeter or an orifice meter. The fuels were metered using calibrated mass flowmeters. All experiments except those for emissions measurements were conducted using high-purity methane. Emissions measurements were obtained in natural gas (NG) flames containing greater than 90% methane. The exhaust-gas samples were collected through an uncooled quartz glass quenching probe. The concentrations of CO and NOx were measured using gas analyzers, and they are reported in this paper on an
uncorrected dry basis. The uncertainty of the emissions measurements is Ⳳ 2 ppm and the uncertainty of the calculated adiabatic flame temperature based on the measured flow rates is Ⳳ 20 ⬚C. OH Fluorescence Imaging A frequency-doubled, Nd:YAG-pumped dye laser was used to pump the Q1(8) line of the (1,0) band of the OH A2R R X2P electronic transition at 283.556 nm. The OH fluorescence signal was collected using a 105-mm focal length, f/4.5 UV Nikkor lens, passed through a Schott WG305 colored glass filter, and focused onto an intensified CCD camera. With a magnification of 0.15, the 512 ⳯ 512 pixel format provides a field-of-view of 81.9 ⳯ 81.9 mm with a spatial resolution of 0.16 mm/pixel. One-dimensional, premixed flame calculations were carried out to quantify the relationship between fluorescence signal and OH concentration at the experimental conditions investigated [14]. Calculated species concentration and temperature variations were used to simulate changes in the fluorescence signal across the flame because of variations in quenching cross section and temperature. Collisional quenching cross sections for OH based on the work of Paul [15] were used. The results showed that the OH PLIF images represent OH mole fraction to within 10%. Numerical Computations The equilibrium adiabatic flame temperature was calculated for each fuel mixture approximately at 200 different lean equivalence ratios using subroutines of CHEMKIN-II [16]. Fifteen C–H–O–N species were used to obtain accurate results. Strainedflame calculations were performed using the RUN-1DL code (version 12.2) [17]. This simulation was used to model strained combustion between two opposing planar premixed jets. The strain rate was defined as the axial velocity gradient upstream of the flame front. The effects of the opposing jets were modeled by imposing zero-gradient boundary conditions at the stagnation plane, since this is a line of symmetry for identical jets. RUN-1DL contains an adaptive grid routine, so that the calculations were optimized with the fewest grid points necessary to achieve the required accuracy. RUN-1DL solves equations using a steady-state Newton Solver, but will time step as necessary to bring poor initial guess solutions into convergence. A detailed kinetic mechanism consisting of 50 reactions and 17 species was used to model the flame chemistry. Experimental Results Flame Stability Flame stability characteristics were obtained by maintaining a constant volumetric flow rate of the
LPM COMBUSTION OF HYDROGEN-ENRICHED METHANE
845
Fig. 1. Flame stability and lean blowout limits of hydrogen-enriched methane: (a) XH2 ⳱ 0.0, (b) XH2 ⳱ 0.12, (c) XH2 ⳱ 0.22, and (d) XH2 ⳱ 0.29.
fuel/air mixture (Q) through the combustor and incrementally decreasing the equilibrium adiabatic flame temperature (Tad) until instability and blowout occurred. The flame instability refers to a weakening of the flame anchoring, which produced a tornadoshaped flame in the center of the combustor. In this study, the lean stability limits of different methane/ hydrogen mixtures are compared for a given equilibrium adiabatic flame temperature. Note that the actual flame temperature will be lower than the equilibrium temperature because of the drop in the combustion efficiency as lean blowout is approached. Fig. 1 shows the lean stability maps for hydrogen mole fraction XH2 ⳱ 0.0, 0.12, 0.22, and 0.29. The upper data points refer to the flame instability and the lower data points represent the flame blowout. The results show that flames at higher fuelair flow rates stabilized at higher adiabatic temperatures. This effect is attributed to the higher strain
rate in the flame at larger fuel-air flow rates. At Q ⳱ 500 slm, the methane flame blows out at Tad ⳱ 1175 ⬚C. The corresponding value for XH2 ⳱ 0.12, 0.22, and 0.29 flames is, respectively, 1090, 1080, and 1070 ⬚C. The results demonstrate that the addition of hydrogen to methane extends the lean stability and blowout limits in the present combustor configuration. The ability of hydrogen to improve lean stability in a methane flame was further explored using one-dimensional strained premixed flame calculations in RUN1-DL. The computed results in Fig. 2 show that the strain rate at extinction increases as the flame temperature is raised, and as the amount of hydrogen in the fuel is increased. At an equilibrium adiabatic flame temperature of 1400 ⬚C, the computed extinction strain rate is 600, 1100, and 2000 sⳮ1, respectively, for flames of methane, methane with 20% hydrogen, and methane with 40% hydrogen. The
846
ENERGY PRODUCTION—Swirl-Stabilized Combustion
computational results support the experimental finding that the fuels containing hydrogen withstand higher strain before extinction. The comparison between experiments and computations is, however, not direct. In experiments, the premixed flame stabilizes by orienting itself to the local flow velocity and strain fields that are subject to the turbulent fluctuations. Hence, the computations simulate the local phenomena rather than the global characteristics of the flame. Pollutant Emissions
Fig. 2. Extinction curves obtained from strained premixed planar opposed-flow calculations.
To gain an understanding of CO and NOx emissions produced in the flame, probe measurements were taken in the combustor using NG for XH2 ⳱ 0.0 and 0.45. The combustor was operated near the lean stability limit of NG at Q ⳱ 450 slm and Tad ⳱ 1240 ⬚C for both fuels. Fig. 3 shows radial profiles of CO and NOx concentrations at two axial planes (z) measured from the swirler exit to depict
Fig. 3. Radial profiles of CO and NOx concentrations at different axial planes: (a) CO at z ⳱ 5.1 cm, (b) CO at z ⳱ 20.3 cm, (c) NOx at z ⳱ 5.1 cm, (d) NOx at z ⳱ 20.3 cm.
LPM COMBUSTION OF HYDROGEN-ENRICHED METHANE
847
enriched fuel produced CO values below 10 ppm at z ⳱ 20.3 cm, where the combustion process was nearly complete. The corresponding CO values for natural gas ranged from 15 to 100 ppm in the central region and several hundreds of parts per million in the wall-boundary region. The NOx concentration shown in Fig. 3, c and d, was always below 10 ppm because of the modest adiabatic flame temperature of the experiment. The NOx formed close to the combustor inlet and reached nearly steady conditions within z ⳱ 5.1 cm, evidently because of the fast combustion with hydrogen. Natural gas flame required a longer residence distance to reach steady NOx concentrations at z ⳱ 20.3 cm. Fig 3d shows similar NOx concentration profiles for both fuels at z ⳱ 20.3 cm. This result suggests that the NOx production was not affected by hydrogen addition and that it was determined mainly by the adiabatic flame temperature. These results demonstrate that a significant reduction in CO emissions is realized by hydrogen addition without adversely affecting the NOx emissions. Luminous Flame Shape Fig. 4. Direct flame luminosity photographs for Q ⳱ 700 slm and Tad ⳱ 1290 Ⳳ 15 ⬚C: (a) XH2 ⳱ 0.0, (b) XH2 ⳱ 0.12, (c) XH2 ⳱ 0.22, and (d) XH2 ⳱ 0.29.
Fig. 5. Single-shot, OH PLIF images for Q ⳱ 850 slm and Tad ⳱ 1240 Ⳳ 15 ⬚C: (a,b) XH2 ⳱ 0; (c,d) XH2 ⳱ 0.12.
the major trends. The results reveal significantly lower CO concentrations at both axial planes when hydrogen is added to the natural gas. The hydrogen-
Direct luminous photographs of the flame were obtained to gain an understanding of the flame location and size. The results are shown in Fig. 4 for methane with XH2 ⳱ 0.0, 0.12, 0.22, and 0.29. Total air-fuel flow rate of 700 slm and equilibrium adiabatic flame temperature of 1290 Ⳳ 15 ⬚C were conditions near the lean stability limit of methane. The flow in a swirl-stabilized combustor consists of two flow separation zones; one is located downstream of the centerbody and another, referred to as the corner torroidal zone, is located at the intersection of the vertical combustor wall and the horizontal inlet plate. Fig. 4a shows no visible flame in the corner torroidal zone for XH2 ⳱ 0.0. The flame extended past the downstream end of the combustor tube. Hydrogen addition (XH2 ⳱ 0.12) produced a continuous stable flame in the corner torroidal zone and a significantly shorter flame, indicating a more rapid combustion. The flame size and shape were qualitatively similar to that obtained with a methane flame (XH2 ⳱ 0.0) operating at a higher equilibrium adiabatic flame temperature of 1385 ⬚C (not shown). A further increase in hydrogen mole fraction decreased the flame length, although the change was modest. The results show that flames with reactions in the corner torroidal region are shorter and more stable. Because luminous photographs provide only a qualitative description, the OH measurements using PLIF are discussed next to develop a quantitative understanding of the flame structure. PLIF OH Measurements Although the OH persists in the post flame region, the OH levels in the primary reaction front are
848
ENERGY PRODUCTION—Swirl-Stabilized Combustion
Fig. 6. Time-averaged OH PLIF images for Q ⳱ 700 slm and Tad ⳱ 1290 Ⳳ 15 ⬚C. (a) XH2 ⳱ 0.0, (b) XH2 ⳱ 0.12, (c) XH2 ⳱ 0.29.
typically a factor of two or more higher than those in the post flame gases [18]. Thus, the signal intensity in PLIF images provides a direct measure of OH concentrations to identify the flame reaction zones. The instantaneous flame structure is represented first by the single-shot OH PLIF images shown in Fig. 5 for XH2 ⳱ 0.0 and 0.12. The operating conditions of Q ⳱ 850 slm and Tad ⳱ 1240 Ⳳ 15 ⬚C are near the lean stability limit of methane. The re-
sults show that the OH distributions are irregular, probably because of the turbulent fluctuations in the local flow field. The flame without the hydrogen (XH2 ⳱ 0.0) is highly shredded and intermittent, and it appears to be nearly extinguished in Fig. 5b. Large CO emissions can be expected to form in this highly strained flame resulting in poor combustion efficiency, as evidenced by the measurements shown in Fig. 4. The flame with hydrogen addition, however,
Fig. 7. RMS OH PLIF images for Q ⳱ 700 slm and Tad ⳱ 1290 Ⳳ 15 ⬚C. (a) XH2 ⳱ 0.0, (b) XH2 ⳱ 0.12, (c) XH2 ⳱ 0.29.
LPM COMBUSTION OF HYDROGEN-ENRICHED METHANE
Fig. 8. Effect of hydrogen addition on OH concentration. (a) Experimental for Q ⳱ 850 slm and Tad ⳱ 1275 ⬚C; (b) calculations for strain rate of 645 sⳮ1 and Tad ⳱ 1400 ⬚C.
reveals a wider region of high OH concentrations. Furthermore, the OH distributions are more continuous, the peak OH levels along the outer edges of the flame are higher, and the flame does not experience local extinction. These results show that the OH levels increase in a hydrogen-enriched flame, and suggest that the elevated radial pool may be responsible for the improved flame stability observed. Although single-shot OH measurements are important, a global understanding of the flame structure is gained from statistical data composed of timeaveraged mean and root-mean-square (RMS) images of OH concentrations. A set of 600 singleshot images was used to obtain the statistical data. Fig. 6 shows time-averaged OH PLIF images for XH2 ⳱ 0.0, 0.12, and 0.29. The operating conditions
849
were the same as those in Fig. 4, that is, Q ⳱ 700 slm and Tad ⳱ 1240 ⬚C. Fig. 6 shows only a modest change in the peak OH levels with increased hydrogen addition. The location of peak OH levels was unaffected by hydrogen addition. However, the flame structure is significantly altered with the addition of hydrogen. At XH2 ⳱ 0.0, the OH level is negligible suggesting no flame in the corner torroidal region. This result is consistent with the luminous photograph of the flame in Fig. 4a. At XH2 ⳱ 0.12, the OH extends into the corner torroidal zone along a layer adjacent to the inner wall of the combustor. Further increase in hydrogen from XH2 ⳱ 0.12 to 0.29 results in high OH concentrations extending across a wider portion of the corner torroidal zone. The corresponding RMS OH PLIF images are shown in Fig. 7. Evidently, the highest fluctuation levels occur in the layer located between the incoming reactant stream and the centerbody torroidal zone. Large fluctuations in OH levels reflect the movement of the flame as it responds to the local flow turbulence. Note that the peak RMS OH levels are comparable to the peak mean OH levels. At XH2 ⳱ 0.0, the RMS OH values are near zero in the corner torroidal zone because there is no combustion. At XH2 ⳱ 0.12, the flame is present intermittently in the corner torroidal zone, resulting in large OH fluctuations between zero OH and high flame values. Finally, at XH2 ⳱ 0.29, the combustion is continuous and stable in this zone, resulting in reduced RMS OH levels. To facilitate a direct comparison of the OH levels, flow conditions were determined such that the hydrogen addition could be varied without altering the global flame structure. It was found that for Q ⳱ 850 slm and Tad ⳱ 1275 ⬚C, the flame structure was similar to that shown in Fig. 6a, that is, no flame in the corner torroidal zone and a flame that has moved somewhat away from the combustor walls, as XH2 was varied from 0.0 to 0.12. Time-averaged radial profiles of OH-signal strength corresponding to these conditions are presented in Fig. 8a for z ⳱ 3.0 cm. Evidently, there is a noticeable increase in the OH levels when a moderate amount of hydrogen is added to methane. However, the addition of larger amounts of hydrogen does not proportionally increase the peak OH concentration. For example, an increase in XH2 from 0.06 to 0.12 caused only about a 3% increase in the maximum OH level. The OH mole fraction profiles from strained premixed flame calculations are shown in Fig. 8b for different fuels at adiabatic flame temperature of 1400 ⬚C and strain rate of 645 sⳮ1. The axial coordinate in Fig. 8b is the distance from the stagnation plane. The computed OH mole fraction profiles support the experimental finding that the peak OH values increase nonlinearly with increasing hydrogen content in the methane fuel. The similarity in shape of the experimental and computed profiles in Fig. 8
850
ENERGY PRODUCTION—Swirl-Stabilized Combustion
a, and b, is coincidental, however, because the calculations simulate the flame characteristics at the local flow conditions. Conclusions Experiments were conducted in a lean premixed combustor to obtain data on flame stability/blowout and on emissions of CO and NOx using hydrogenenriched methane or natural gas. The flame structure was identified qualitatively by luminous photographs and quantitatively by OH PLIF measurements. Results show that the lean stability limit was lowered by the addition of hydrogen to the fuel. A significant reduction in CO emissions was realized by hydrogen addition as the lean stability limit of natural gas was approached. The NOx emissions for an equilibrium adiabatic flame temperature were not affected by the hydrogen addition. The flame size and shape of hydrogen-enriched fuel were qualitatively similar to those obtained with methane flame at a higher equilibrium adiabatic flame temperature. The improved stability with hydrogen enrichment was postulated to be a direct result of higher OH, H, and O radical concentrations, which increase several key reaction rates. The OH PLIF measurements show that moderate amounts of hydrogen enrichment led to significant increases in the OH radical concentrations. Increasing amounts of hydrogen enrichment to the highest levels increased the radical concentrations only marginally for the conditions tested in this study. Hydrogen addition produced a broader region of high OH concentrations. Flame characteristics were affected most significantly by reactions in the corner torroidal zone. Acknowledgments This research was supported in part by the U.S. Department of Energy, Hydrogen Program and Department of Energy, Office of Basic Energy Sciences. REFERENCES 1. Richards, G. A., McMillian, M. M., Gemmen, R. S., Rogers, W. A., and Cully, S. R., Prog. Energy Combust. Sci. 27:141 (2001).
2. Ren, J.-Y., Qin, W., Egolfopoulos, F. N., Mak, H., and Tsotsis, T. T., Chem. Eng. Sci. 56:1541 (2001). 3. Meier, J. G., Hung, W. S .Y., and Sood, V. M., J. Eng. Gas Turb. Power 108:182 (1986). 4. Bell, S. R., and Gupta, M., Combust. Sci. Technol. 123:23 (1997). 5. Larsen, J. F., and Wallace, J. S., J. Eng. Gas Turb. Power 119:218 (1997). 6. Anderson, D. N., Effect of Hydrogen Injection on Stability and Emissions of an Experimental Premixed PreVaporized Propane Burner, NASA TM X-3301, National Technical Information Service, Springfield, VA, 1975. 7. Nguyen, O., and Samuelsen, S., ASME paper 99-GT359, 1999. 8. Morris, J. D., Symonds, R. A., Ballard, F. L., and Banti, A., ASME paper 98-GT-359, 1998. 9. Phillips, J. N., and Roby, R. J., ASME paper 99-GT115, 1999. 10. Gauducheau, J. L., Denet, B., and Searby, G., Combust. Sci. Technol. 137:81 (1998). 11. Ren, J.-Y., Qin, W., Egolfopoulos, F. N., and Tsotsis, T. T., Combust. Flame 124:717 (2001). 12. Cong, Y., and Jackson, G. S., ‘‘Impact of Hydrogen Doping on Lean-Premixed Methane Flame Stability,’’ paper 85, Fall Technical Meeting of the Eastern States Section of the Combustion Institute, Raleigh, NC, October 13–15, 1999. 13. Ji, J., and Gore, J. P., ‘‘Flow Structure in Lean Premixed Swirling Combustion,’’ paper 1331, Spring Technical Meeting of Central States Section of the Combustion Institute, Knoxville, TN, April 7–9, 2002. 14. Kee, R. J., Grcar, J. F., Smooke, M. D., and Miller, J. A., Premix: A FORTRAN Program for Modeling Steady Laminar One-Dimensional Premixed Flames, Sandia report SAND85-8240. 15. Paul, P. H., J. Quant. Spectrosc. Radiat. Transfer 51:511–524 (1994). 16. Kee, R. J., Rupley, F. M., and Miller, J. A., ChemkinII: A FORTRAN Chemical Kinetics Package for the Analysis of Gas Phase Chemical Kinetics, Sandia report SAND89-809B-UC-706. 17. Rogg, B., and Wang, W., RUN-1DL: The Laminar Flame and Flamelet Code User Manual, Rurh University, Bochum, Germany, 1995. 18. Cattolica, R. J., Combust. Flame 44:43–59 (1982).
COMMENTS D. P. Mishra, Indian Institute of Technology, Kaupur, India. It appears that you have carried out research on two different configurations, the swirl burner and the opposed jet flame. How can one be mapped into the other? Since flammability is a fundamental property, I am afraid it cannot be characterized by your swirl burner. Rather, the
standard flammability tube can be used to characterize the flammability limit. Please comment on this. Author’s Reply. In this study, we have focused on the swirl burner. As discussed in the paper, the opposed jet flame is used to model what is occurring at a local level in
LPM COMBUSTION OF HYDROGEN-ENRICHED METHANE the swirl-stabilized flame due to its highly turbulent nature. The extinction characteristics presented in this paper are not flammability limits. This data shows the performance enhancement with the addition of hydrogen to the lean blow-off limit, which is a characteristic of this particular burner. ● Wolfgang Meier, DLR Stuttgart, Germany. What is the noise level of the flames? Does the level and frequency change with H2 enrichment? Author’s Reply. The noise level of these flames is audibly loud in narrow bands of equivalence ratio. The peak noise level is observed to be constant irrespective of the amount of hydrogen enrichment, as is the frequency. The major difference observed is that the region in which the combustor exhibits thermoacoustic instabilities is shifted to a lower equivalence ratio with the addition of hydrogen. ● Robert K. Cheng, Lawrence Berkeley National Laboratory, USA. Your reference to the lean blow-off limit of your apparatus as a flammability limit is misleading. Flamma-
851
bility limit is generally defined as a property of the flame and not of the system. Additionally, can you comment on the role of the corner recirculation zone on noise and instability? Author’s Reply. Regarding the flammability limit, see the previous comment and reply by Mishra. As to the dependence of combustion noise on the presence or stability of the corner recirculation zone it has not been observed. ● Yasir M. Al-Abdeli, University of Sydney, Australia. For the conditions investigated, you specified a swirl number (S) of 0.75 and the Reynolds number (Re) in the range 10,000–20,000. Since you are applying a vane type swirl generator, did you verify that changing Re does not effectively alter S? Author’s Reply. The swirl number presented in this work was calculated assuming the reactants were injected into the combustor at the same angle as the swirl vanes. The flow in general does not perfectly follow the vanes, and therefore the actual swirl number will be different from the calculated swirl number and possibly dependent on Reynolds’s number as well.
COMBUSTION OF HYDROGEN-ENRICHED METHANE IN A LEAN PREMIXED SWIRL-STABILIZED BURNER R. W. SCHEFER,1 D. M. WICKSALL2 and A. K. AGRAWAL2 1
Combustion Research Facility Sandia National Laboratories Livermore, CA 94551-0969, USA 2 School of Aerospace and Mechanical Engineering University of Oklahoma Norman, OK 73019, USA
The combustion characteristics of a premixed, swirl-stabilized flame were studied to determine the effects of enriching methane with hydrogen under fuel-lean conditions. The burner consisted of a centerbody with an annular, premixed fuel-air jet. Swirl was introduced to the flow using 45-degree swirl vanes. The combustion occurred within an air-cooled quartz chamber at atmospheric pressure. Flame stability and blowout maps were obtained for different amounts of hydrogen addition at several fuel-air flow rates. Gas probe measurements were obtained to demonstrate reductions in CO concentration with hydrogen addition, without adversely affecting the NOx emissions. The flame structure near the lean stability limit was described by direct luminous photographs and planar laser-induced fluorescence measurements of the OH radical. Results show that the addition of a moderate amount of hydrogen to the methane/air mixture increased the peak OH concentration. Hydrogen addition resulted in a significant change in the flame structure, indicated by a shorter and more robust appearing flame. The observed trends concur with the strained opposed premixed flame analysis using RUN-1DL. The computations revealed that enriching the methane with hydrogen increased the strain resistance of the flame as well as the OH levels in the flame.
Introduction The development of advanced combustion capabilities for gaseous hydrogen and hydrogen-blended hydrocarbon fuels in gas turbine applications is an area of much current interest encompassing several needs. One need is the cost-effective utilization of alternative fuels with a wide range of heating values [1]. Low- and medium-heating-value fuels containing hydrogen are often produced as by-products in Coal-Gasification Combined Cycle and Fluidized Bed Combustion installations. These product gases could provide a significant source of cost-effective fuels for power-generating gas turbines. A second need is related to the recognition that ultralean premixed combustion is effective in reducing the NOx emissions. Hydrogen blended with traditional hydrocarbon fuels could improve flame stability and allow stable lean combustion at the lower temperatures needed to minimize the NOx production [2]. A longer-term need is the desire to minimize and eventually eliminate unburned hydrocarbon (UHC) and CO2 emissions. The use of hydrogen-blended fuels provides both a solution to the immediate need for NOx reduction, and a transition strategy to a carbon-free energy system in the future. Fuels containing hydrogen have been used for some time in conventional non-premixed combustion systems [3]. Premixed combustion studies in
spark-ignition engines operating on hythane, a blend of hydrogen and natural gas, show lower exhaust emissions and a leaner operating limit [4,5]. An early study on lean premixed (LPM) combustion of propane with air showed improved combustion efficiency and lower temperature at lean blowout with hydrogen addition [6]. Using hydrogen as a pilot fuel was shown to improve the lean-stability limit in LPM combustion systems for gas turbines [7]. In a recent study, fuels containing up to 10% hydrogen were utilized in a commercial gas turbine [8]. Conceptual studies of large-scale gas turbine-based power plants have concluded that the hydrogen addition can be used to improve the combustor performance [9]. At a more fundamental level, research on hydrogen-enriched fuels has been conducted using an opposed jet flow configuration [10–12]. Based on the results of strained-flame calculations at an elevated pressure (30 atm), the improved lean flame stability was attributed to hydrogen’s higher flame speed and increased resistance to strain [10]. The presence of hydrogen substantially increases the extinction strain rate, an effect which is more pronounced in leaner mixtures [11]. Sensitivity analyses show that the chain-branching reaction H Ⳮ O2 } OH Ⳮ O is important with respect to extinction [2,11]. Furthermore, the OH and O radicals contribute to the pyrolysis of the carbon species. The reduction in CO
843
844
ENERGY PRODUCTION—Swirl-Stabilized Combustion
emissions with hydrogen addition was also attributed to the increased radical pool. Higher OH concentrations are likely to promote completion of CO oxidation to CO2 via the OH radical. The primary objective of this study is to delineate the role of hydrogen addition in a LPM combustor representative of advanced gas turbines. In the remainder of the paper, the experimental system and diagnostics will be described. Results characterizing the combustion stability and emissions characteristics of a laboratory-scale, swirl-stabilized LPM combustor will be presented. The fuel will consist of methane or natural gas enriched with hydrogen. The flame structure will be characterized using direct luminosity photographs and OH planar laser-induced fluorescence (PLIF) images. In particular, the flame behavior will be characterized under fuel lean conditions near blowout using measurements of instantaneous and time-averaged OH concentrations. Strained premixed flame analysis in an opposed-flow configuration will be used to explain the observed trends. Experimental System Burner Description and Flow Conditions The test apparatus is a laboratory-scale, swirlingflow dump combustor that consists of a premixer and a combustion chamber. The inside diameter of the premixer tube is 4.1 cm and the outer diameter of the fuel centerbody is 2.5 cm. Premixed conditions were achieved by injecting fuel at 40.0 cm upstream of the combustor inlet. The flame is stabilized on a centerbody with seven 45-degree swirler vanes, located at the end of the mixing chamber. The average axial velocity at the swirler exit varied from 8.0 to 22.0 m/s for the fuel/air mixture flow rate of 400 to 1100 standard liter per minute. Although the detailed flow measurements were not obtained, turbulence intensity levels of up to 150% were reported by Ji and Gore [13] in a similar setup. The estimated strain rate based on the bulk-averaged inlet velocity and the combustor inlet length scale ranged between 400 and 2500 sⳮ1. The combustion chamber is a 30.5-cm-long quartz tube of 8.3-cm inside diameter. Combustion air was provided by an air compressor and metered upstream of the burner using either a mass flowmeter or an orifice meter. The fuels were metered using calibrated mass flowmeters. All experiments except those for emissions measurements were conducted using high-purity methane. Emissions measurements were obtained in natural gas (NG) flames containing greater than 90% methane. The exhaust-gas samples were collected through an uncooled quartz glass quenching probe. The concentrations of CO and NOx were measured using gas analyzers, and they are reported in this paper on an
uncorrected dry basis. The uncertainty of the emissions measurements is Ⳳ 2 ppm and the uncertainty of the calculated adiabatic flame temperature based on the measured flow rates is Ⳳ 20 ⬚C. OH Fluorescence Imaging A frequency-doubled, Nd:YAG-pumped dye laser was used to pump the Q1(8) line of the (1,0) band of the OH A2R R X2P electronic transition at 283.556 nm. The OH fluorescence signal was collected using a 105-mm focal length, f/4.5 UV Nikkor lens, passed through a Schott WG305 colored glass filter, and focused onto an intensified CCD camera. With a magnification of 0.15, the 512 ⳯ 512 pixel format provides a field-of-view of 81.9 ⳯ 81.9 mm with a spatial resolution of 0.16 mm/pixel. One-dimensional, premixed flame calculations were carried out to quantify the relationship between fluorescence signal and OH concentration at the experimental conditions investigated [14]. Calculated species concentration and temperature variations were used to simulate changes in the fluorescence signal across the flame because of variations in quenching cross section and temperature. Collisional quenching cross sections for OH based on the work of Paul [15] were used. The results showed that the OH PLIF images represent OH mole fraction to within 10%. Numerical Computations The equilibrium adiabatic flame temperature was calculated for each fuel mixture approximately at 200 different lean equivalence ratios using subroutines of CHEMKIN-II [16]. Fifteen C–H–O–N species were used to obtain accurate results. Strainedflame calculations were performed using the RUN-1DL code (version 12.2) [17]. This simulation was used to model strained combustion between two opposing planar premixed jets. The strain rate was defined as the axial velocity gradient upstream of the flame front. The effects of the opposing jets were modeled by imposing zero-gradient boundary conditions at the stagnation plane, since this is a line of symmetry for identical jets. RUN-1DL contains an adaptive grid routine, so that the calculations were optimized with the fewest grid points necessary to achieve the required accuracy. RUN-1DL solves equations using a steady-state Newton Solver, but will time step as necessary to bring poor initial guess solutions into convergence. A detailed kinetic mechanism consisting of 50 reactions and 17 species was used to model the flame chemistry. Experimental Results Flame Stability Flame stability characteristics were obtained by maintaining a constant volumetric flow rate of the
LPM COMBUSTION OF HYDROGEN-ENRICHED METHANE
845
Fig. 1. Flame stability and lean blowout limits of hydrogen-enriched methane: (a) XH2 ⳱ 0.0, (b) XH2 ⳱ 0.12, (c) XH2 ⳱ 0.22, and (d) XH2 ⳱ 0.29.
fuel/air mixture (Q) through the combustor and incrementally decreasing the equilibrium adiabatic flame temperature (Tad) until instability and blowout occurred. The flame instability refers to a weakening of the flame anchoring, which produced a tornadoshaped flame in the center of the combustor. In this study, the lean stability limits of different methane/ hydrogen mixtures are compared for a given equilibrium adiabatic flame temperature. Note that the actual flame temperature will be lower than the equilibrium temperature because of the drop in the combustion efficiency as lean blowout is approached. Fig. 1 shows the lean stability maps for hydrogen mole fraction XH2 ⳱ 0.0, 0.12, 0.22, and 0.29. The upper data points refer to the flame instability and the lower data points represent the flame blowout. The results show that flames at higher fuelair flow rates stabilized at higher adiabatic temperatures. This effect is attributed to the higher strain
rate in the flame at larger fuel-air flow rates. At Q ⳱ 500 slm, the methane flame blows out at Tad ⳱ 1175 ⬚C. The corresponding value for XH2 ⳱ 0.12, 0.22, and 0.29 flames is, respectively, 1090, 1080, and 1070 ⬚C. The results demonstrate that the addition of hydrogen to methane extends the lean stability and blowout limits in the present combustor configuration. The ability of hydrogen to improve lean stability in a methane flame was further explored using one-dimensional strained premixed flame calculations in RUN1-DL. The computed results in Fig. 2 show that the strain rate at extinction increases as the flame temperature is raised, and as the amount of hydrogen in the fuel is increased. At an equilibrium adiabatic flame temperature of 1400 ⬚C, the computed extinction strain rate is 600, 1100, and 2000 sⳮ1, respectively, for flames of methane, methane with 20% hydrogen, and methane with 40% hydrogen. The
846
ENERGY PRODUCTION—Swirl-Stabilized Combustion
computational results support the experimental finding that the fuels containing hydrogen withstand higher strain before extinction. The comparison between experiments and computations is, however, not direct. In experiments, the premixed flame stabilizes by orienting itself to the local flow velocity and strain fields that are subject to the turbulent fluctuations. Hence, the computations simulate the local phenomena rather than the global characteristics of the flame. Pollutant Emissions
Fig. 2. Extinction curves obtained from strained premixed planar opposed-flow calculations.
To gain an understanding of CO and NOx emissions produced in the flame, probe measurements were taken in the combustor using NG for XH2 ⳱ 0.0 and 0.45. The combustor was operated near the lean stability limit of NG at Q ⳱ 450 slm and Tad ⳱ 1240 ⬚C for both fuels. Fig. 3 shows radial profiles of CO and NOx concentrations at two axial planes (z) measured from the swirler exit to depict
Fig. 3. Radial profiles of CO and NOx concentrations at different axial planes: (a) CO at z ⳱ 5.1 cm, (b) CO at z ⳱ 20.3 cm, (c) NOx at z ⳱ 5.1 cm, (d) NOx at z ⳱ 20.3 cm.
LPM COMBUSTION OF HYDROGEN-ENRICHED METHANE
847
enriched fuel produced CO values below 10 ppm at z ⳱ 20.3 cm, where the combustion process was nearly complete. The corresponding CO values for natural gas ranged from 15 to 100 ppm in the central region and several hundreds of parts per million in the wall-boundary region. The NOx concentration shown in Fig. 3, c and d, was always below 10 ppm because of the modest adiabatic flame temperature of the experiment. The NOx formed close to the combustor inlet and reached nearly steady conditions within z ⳱ 5.1 cm, evidently because of the fast combustion with hydrogen. Natural gas flame required a longer residence distance to reach steady NOx concentrations at z ⳱ 20.3 cm. Fig 3d shows similar NOx concentration profiles for both fuels at z ⳱ 20.3 cm. This result suggests that the NOx production was not affected by hydrogen addition and that it was determined mainly by the adiabatic flame temperature. These results demonstrate that a significant reduction in CO emissions is realized by hydrogen addition without adversely affecting the NOx emissions. Luminous Flame Shape Fig. 4. Direct flame luminosity photographs for Q ⳱ 700 slm and Tad ⳱ 1290 Ⳳ 15 ⬚C: (a) XH2 ⳱ 0.0, (b) XH2 ⳱ 0.12, (c) XH2 ⳱ 0.22, and (d) XH2 ⳱ 0.29.
Fig. 5. Single-shot, OH PLIF images for Q ⳱ 850 slm and Tad ⳱ 1240 Ⳳ 15 ⬚C: (a,b) XH2 ⳱ 0; (c,d) XH2 ⳱ 0.12.
the major trends. The results reveal significantly lower CO concentrations at both axial planes when hydrogen is added to the natural gas. The hydrogen-
Direct luminous photographs of the flame were obtained to gain an understanding of the flame location and size. The results are shown in Fig. 4 for methane with XH2 ⳱ 0.0, 0.12, 0.22, and 0.29. Total air-fuel flow rate of 700 slm and equilibrium adiabatic flame temperature of 1290 Ⳳ 15 ⬚C were conditions near the lean stability limit of methane. The flow in a swirl-stabilized combustor consists of two flow separation zones; one is located downstream of the centerbody and another, referred to as the corner torroidal zone, is located at the intersection of the vertical combustor wall and the horizontal inlet plate. Fig. 4a shows no visible flame in the corner torroidal zone for XH2 ⳱ 0.0. The flame extended past the downstream end of the combustor tube. Hydrogen addition (XH2 ⳱ 0.12) produced a continuous stable flame in the corner torroidal zone and a significantly shorter flame, indicating a more rapid combustion. The flame size and shape were qualitatively similar to that obtained with a methane flame (XH2 ⳱ 0.0) operating at a higher equilibrium adiabatic flame temperature of 1385 ⬚C (not shown). A further increase in hydrogen mole fraction decreased the flame length, although the change was modest. The results show that flames with reactions in the corner torroidal region are shorter and more stable. Because luminous photographs provide only a qualitative description, the OH measurements using PLIF are discussed next to develop a quantitative understanding of the flame structure. PLIF OH Measurements Although the OH persists in the post flame region, the OH levels in the primary reaction front are
848
ENERGY PRODUCTION—Swirl-Stabilized Combustion
Fig. 6. Time-averaged OH PLIF images for Q ⳱ 700 slm and Tad ⳱ 1290 Ⳳ 15 ⬚C. (a) XH2 ⳱ 0.0, (b) XH2 ⳱ 0.12, (c) XH2 ⳱ 0.29.
typically a factor of two or more higher than those in the post flame gases [18]. Thus, the signal intensity in PLIF images provides a direct measure of OH concentrations to identify the flame reaction zones. The instantaneous flame structure is represented first by the single-shot OH PLIF images shown in Fig. 5 for XH2 ⳱ 0.0 and 0.12. The operating conditions of Q ⳱ 850 slm and Tad ⳱ 1240 Ⳳ 15 ⬚C are near the lean stability limit of methane. The re-
sults show that the OH distributions are irregular, probably because of the turbulent fluctuations in the local flow field. The flame without the hydrogen (XH2 ⳱ 0.0) is highly shredded and intermittent, and it appears to be nearly extinguished in Fig. 5b. Large CO emissions can be expected to form in this highly strained flame resulting in poor combustion efficiency, as evidenced by the measurements shown in Fig. 4. The flame with hydrogen addition, however,
Fig. 7. RMS OH PLIF images for Q ⳱ 700 slm and Tad ⳱ 1290 Ⳳ 15 ⬚C. (a) XH2 ⳱ 0.0, (b) XH2 ⳱ 0.12, (c) XH2 ⳱ 0.29.
LPM COMBUSTION OF HYDROGEN-ENRICHED METHANE
Fig. 8. Effect of hydrogen addition on OH concentration. (a) Experimental for Q ⳱ 850 slm and Tad ⳱ 1275 ⬚C; (b) calculations for strain rate of 645 sⳮ1 and Tad ⳱ 1400 ⬚C.
reveals a wider region of high OH concentrations. Furthermore, the OH distributions are more continuous, the peak OH levels along the outer edges of the flame are higher, and the flame does not experience local extinction. These results show that the OH levels increase in a hydrogen-enriched flame, and suggest that the elevated radial pool may be responsible for the improved flame stability observed. Although single-shot OH measurements are important, a global understanding of the flame structure is gained from statistical data composed of timeaveraged mean and root-mean-square (RMS) images of OH concentrations. A set of 600 singleshot images was used to obtain the statistical data. Fig. 6 shows time-averaged OH PLIF images for XH2 ⳱ 0.0, 0.12, and 0.29. The operating conditions
849
were the same as those in Fig. 4, that is, Q ⳱ 700 slm and Tad ⳱ 1240 ⬚C. Fig. 6 shows only a modest change in the peak OH levels with increased hydrogen addition. The location of peak OH levels was unaffected by hydrogen addition. However, the flame structure is significantly altered with the addition of hydrogen. At XH2 ⳱ 0.0, the OH level is negligible suggesting no flame in the corner torroidal region. This result is consistent with the luminous photograph of the flame in Fig. 4a. At XH2 ⳱ 0.12, the OH extends into the corner torroidal zone along a layer adjacent to the inner wall of the combustor. Further increase in hydrogen from XH2 ⳱ 0.12 to 0.29 results in high OH concentrations extending across a wider portion of the corner torroidal zone. The corresponding RMS OH PLIF images are shown in Fig. 7. Evidently, the highest fluctuation levels occur in the layer located between the incoming reactant stream and the centerbody torroidal zone. Large fluctuations in OH levels reflect the movement of the flame as it responds to the local flow turbulence. Note that the peak RMS OH levels are comparable to the peak mean OH levels. At XH2 ⳱ 0.0, the RMS OH values are near zero in the corner torroidal zone because there is no combustion. At XH2 ⳱ 0.12, the flame is present intermittently in the corner torroidal zone, resulting in large OH fluctuations between zero OH and high flame values. Finally, at XH2 ⳱ 0.29, the combustion is continuous and stable in this zone, resulting in reduced RMS OH levels. To facilitate a direct comparison of the OH levels, flow conditions were determined such that the hydrogen addition could be varied without altering the global flame structure. It was found that for Q ⳱ 850 slm and Tad ⳱ 1275 ⬚C, the flame structure was similar to that shown in Fig. 6a, that is, no flame in the corner torroidal zone and a flame that has moved somewhat away from the combustor walls, as XH2 was varied from 0.0 to 0.12. Time-averaged radial profiles of OH-signal strength corresponding to these conditions are presented in Fig. 8a for z ⳱ 3.0 cm. Evidently, there is a noticeable increase in the OH levels when a moderate amount of hydrogen is added to methane. However, the addition of larger amounts of hydrogen does not proportionally increase the peak OH concentration. For example, an increase in XH2 from 0.06 to 0.12 caused only about a 3% increase in the maximum OH level. The OH mole fraction profiles from strained premixed flame calculations are shown in Fig. 8b for different fuels at adiabatic flame temperature of 1400 ⬚C and strain rate of 645 sⳮ1. The axial coordinate in Fig. 8b is the distance from the stagnation plane. The computed OH mole fraction profiles support the experimental finding that the peak OH values increase nonlinearly with increasing hydrogen content in the methane fuel. The similarity in shape of the experimental and computed profiles in Fig. 8
850
ENERGY PRODUCTION—Swirl-Stabilized Combustion
a, and b, is coincidental, however, because the calculations simulate the flame characteristics at the local flow conditions. Conclusions Experiments were conducted in a lean premixed combustor to obtain data on flame stability/blowout and on emissions of CO and NOx using hydrogenenriched methane or natural gas. The flame structure was identified qualitatively by luminous photographs and quantitatively by OH PLIF measurements. Results show that the lean stability limit was lowered by the addition of hydrogen to the fuel. A significant reduction in CO emissions was realized by hydrogen addition as the lean stability limit of natural gas was approached. The NOx emissions for an equilibrium adiabatic flame temperature were not affected by the hydrogen addition. The flame size and shape of hydrogen-enriched fuel were qualitatively similar to those obtained with methane flame at a higher equilibrium adiabatic flame temperature. The improved stability with hydrogen enrichment was postulated to be a direct result of higher OH, H, and O radical concentrations, which increase several key reaction rates. The OH PLIF measurements show that moderate amounts of hydrogen enrichment led to significant increases in the OH radical concentrations. Increasing amounts of hydrogen enrichment to the highest levels increased the radical concentrations only marginally for the conditions tested in this study. Hydrogen addition produced a broader region of high OH concentrations. Flame characteristics were affected most significantly by reactions in the corner torroidal zone. Acknowledgments This research was supported in part by the U.S. Department of Energy, Hydrogen Program and Department of Energy, Office of Basic Energy Sciences. REFERENCES 1. Richards, G. A., McMillian, M. M., Gemmen, R. S., Rogers, W. A., and Cully, S. R., Prog. Energy Combust. Sci. 27:141 (2001).
2. Ren, J.-Y., Qin, W., Egolfopoulos, F. N., Mak, H., and Tsotsis, T. T., Chem. Eng. Sci. 56:1541 (2001). 3. Meier, J. G., Hung, W. S .Y., and Sood, V. M., J. Eng. Gas Turb. Power 108:182 (1986). 4. Bell, S. R., and Gupta, M., Combust. Sci. Technol. 123:23 (1997). 5. Larsen, J. F., and Wallace, J. S., J. Eng. Gas Turb. Power 119:218 (1997). 6. Anderson, D. N., Effect of Hydrogen Injection on Stability and Emissions of an Experimental Premixed PreVaporized Propane Burner, NASA TM X-3301, National Technical Information Service, Springfield, VA, 1975. 7. Nguyen, O., and Samuelsen, S., ASME paper 99-GT359, 1999. 8. Morris, J. D., Symonds, R. A., Ballard, F. L., and Banti, A., ASME paper 98-GT-359, 1998. 9. Phillips, J. N., and Roby, R. J., ASME paper 99-GT115, 1999. 10. Gauducheau, J. L., Denet, B., and Searby, G., Combust. Sci. Technol. 137:81 (1998). 11. Ren, J.-Y., Qin, W., Egolfopoulos, F. N., and Tsotsis, T. T., Combust. Flame 124:717 (2001). 12. Cong, Y., and Jackson, G. S., ‘‘Impact of Hydrogen Doping on Lean-Premixed Methane Flame Stability,’’ paper 85, Fall Technical Meeting of the Eastern States Section of the Combustion Institute, Raleigh, NC, October 13–15, 1999. 13. Ji, J., and Gore, J. P., ‘‘Flow Structure in Lean Premixed Swirling Combustion,’’ paper 1331, Spring Technical Meeting of Central States Section of the Combustion Institute, Knoxville, TN, April 7–9, 2002. 14. Kee, R. J., Grcar, J. F., Smooke, M. D., and Miller, J. A., Premix: A FORTRAN Program for Modeling Steady Laminar One-Dimensional Premixed Flames, Sandia report SAND85-8240. 15. Paul, P. H., J. Quant. Spectrosc. Radiat. Transfer 51:511–524 (1994). 16. Kee, R. J., Rupley, F. M., and Miller, J. A., ChemkinII: A FORTRAN Chemical Kinetics Package for the Analysis of Gas Phase Chemical Kinetics, Sandia report SAND89-809B-UC-706. 17. Rogg, B., and Wang, W., RUN-1DL: The Laminar Flame and Flamelet Code User Manual, Rurh University, Bochum, Germany, 1995. 18. Cattolica, R. J., Combust. Flame 44:43–59 (1982).
COMMENTS D. P. Mishra, Indian Institute of Technology, Kaupur, India. It appears that you have carried out research on two different configurations, the swirl burner and the opposed jet flame. How can one be mapped into the other? Since flammability is a fundamental property, I am afraid it cannot be characterized by your swirl burner. Rather, the
standard flammability tube can be used to characterize the flammability limit. Please comment on this. Author’s Reply. In this study, we have focused on the swirl burner. As discussed in the paper, the opposed jet flame is used to model what is occurring at a local level in
LPM COMBUSTION OF HYDROGEN-ENRICHED METHANE the swirl-stabilized flame due to its highly turbulent nature. The extinction characteristics presented in this paper are not flammability limits. This data shows the performance enhancement with the addition of hydrogen to the lean blow-off limit, which is a characteristic of this particular burner. ● Wolfgang Meier, DLR Stuttgart, Germany. What is the noise level of the flames? Does the level and frequency change with H2 enrichment? Author’s Reply. The noise level of these flames is audibly loud in narrow bands of equivalence ratio. The peak noise level is observed to be constant irrespective of the amount of hydrogen enrichment, as is the frequency. The major difference observed is that the region in which the combustor exhibits thermoacoustic instabilities is shifted to a lower equivalence ratio with the addition of hydrogen. ● Robert K. Cheng, Lawrence Berkeley National Laboratory, USA. Your reference to the lean blow-off limit of your apparatus as a flammability limit is misleading. Flamma-
851
bility limit is generally defined as a property of the flame and not of the system. Additionally, can you comment on the role of the corner recirculation zone on noise and instability? Author’s Reply. Regarding the flammability limit, see the previous comment and reply by Mishra. As to the dependence of combustion noise on the presence or stability of the corner recirculation zone it has not been observed. ● Yasir M. Al-Abdeli, University of Sydney, Australia. For the conditions investigated, you specified a swirl number (S) of 0.75 and the Reynolds number (Re) in the range 10,000–20,000. Since you are applying a vane type swirl generator, did you verify that changing Re does not effectively alter S? Author’s Reply. The swirl number presented in this work was calculated assuming the reactants were injected into the combustor at the same angle as the swirl vanes. The flow in general does not perfectly follow the vanes, and therefore the actual swirl number will be different from the calculated swirl number and possibly dependent on Reynolds’s number as well.