Flame configurations in an annular swirl burner

This paper describes an experimental investigation of flame configurations in a swirling, lean premixed combustor. The flame configuration or stabilization location within the combustor has important influences upon a variety of combustor performance metrics, such as combustion instability limits and pollutant emissions. Flame configurations were mapped for preheat temperatures ranging from 200 to 500 deg F while varying equivalence ratio from 0.40-0.70. Within this operational space, the flame exhibits four basic configurations. We show that the transitions from one flame configuration to the next is essentially a flame extinction problem, and that the sensitivity of these transition points to fuel/air ratio and preheat temperature can be reasonably captured with extinction strain rate calculations.

Flame Configurations in an Annular Swirl Burner C.W. Foley1, I. Chterev1, J. Seitzman and T.Lieuwen 1School o f Aerospace and Mechanical Engineering, Georgia Institute o f Technology, Atlanta, Georgia 30332-0150, USA This paper describes an experimental investigation of flame configurations in a swirling, lean premixed combustor. The flame configuration or stabilization location within the combustor has important influences upon a variety of combustor performance metrics, such as combustion instability limits and pollutant emissions. Flame configurations were mapped for preheat temperatures ranging from 200 to 500 deg F while varying equivalence ratio from 0.40-0.70. Within this operational space, the flame exhibits four basic configurations. We show that the transitions from one flame configuration to the next is essentially a flame extinction problem, and that the sensitivity of these transition points to fuel/air ratio and preheat temperature can be reasonably captured with extinction strain rate calculations. Introduction The objective of this paper is to consider the factors influencing heat release distribution, or more fundamentally, flame position and stabilization locations, in premixed combustors. In particular, we focus on swirling flows with a centerbody, such as shown in Figure 1, a common geometry for both commercial low NOx combustor hardware [1,2] and in fundamental studies of swirling flows [3,4]. The flow field of this geometry, in a time-averaged sense, consists of four main regions [5-7], see Figure 1: (1) the outer recirculation zone (ORZ), a toroidal recirculating regime generated by the rapid expansion of the nozzle into the combustor, (2) the inner recirculation zone (IRZ, also referred to here as the vortex breakdown bubble), due to vortex breakdown accompanying the swirling flow and the bluff body wake, (3) the high velocity, annular fluid jet that divides these regions, and (4) two annular shear layers that divide the ORZ and annular jet (denoted here as the outer shear layer, OSL) and the IRZ and the annular jet (inner shear layer, ISL). Figure 1. Illustration of the key fluid mechanical features in the geometry of interest Figure 1 illustrates several flame configurations which have been observed for this geometry. Note that there are three basic flame holding locations: (1) outer shear layer, OSL (see configuration b and d), (2) inner shear layer, ISL (configurations c and d), and (3) vortex breakdown bubble, VBB (configurations a and b). Figure 2. Illustration of basic flame configurations possible for geometry of interest The location/spatial distribution of the flame in a combustion chamber is a fundamental problem that has important ramifications on combustor operability, durability, efficiency, and emissions. It can be seen that flame stabilization location has very strong influences upon flame shape that, in turn influences heat loadings to combustor hardware (e.g., centerbody, walls, dome plate). For example, the heat transfer to the centerbody is fundamentally different in configurations (c) and (d) than in (a) and (b). This, in turn, has implications on centerbody design and life. Similarly, the degree of flame spreading to combustor walls varies between, for example, configurations (b) and (c). This in turn would influence the location at which the flame impinges on the wall and drive the design of the combustor lining at that location. Next, flame location has important influence on combustion instability boundaries [8]. It is known that combustor stability limits are controlled by the time delay between when a fuel/air ratio disturbance or vortex is created and when it reaches the flame. This time delay will certainly vary between, for example, configurations (a) and (d). This also illustrates that discontinuous changes in combustor stability behavior may occur when the flame abruptly bifurcates from one stabilization location to another. Additionally, stabilization locations influence the blowoff limits of the system. In reality, shifts in flame location can be thought of as a sequence of local blowoff events; e.g., a flame nominally looking like configuration (d) will bifurcate to configuration (b) due to local blowoff of the flame from the centerbody shear layer. How the flame is stabilized plays a crucial role in which physical processes control its dynamics. For example, the configuration shown in Figure 2(d) is clearly affected by the dynamics of both shear layers, while that in Figure 2(b) may be lesser affected by the centerbody shear layer. Similarly, the dynamics of the central portion of the flame is strongly influenced by vortex breakdown bubble dynamics in configurations Figure 2(a) and (b), while configurations (d) and (c) are presumably less affected. Furthermore, the time averaged stabilization location can vary with perturbation amplitude during combustion dynamics, implying that one set of fluid mechanic processes is important at low amplitudes, and another at higher amplitudes. The focus of this paper is on the flame stabilization locations specifically. Flames can hold in either low velocity regimes in the shear layer, or near the stagnation region ahead of the vortex breakdown bubble. Because of the centerbody, there is a low velocity region in the separating shear layer where the flame can stabilize. However, it is known that shear introduces aerodynamic straining on the flame [9], which alters the local temperature and burning rate [10]. If the shear rate and consequent flame strain rate is too large, the flame will locally extinguish and either blow out of the combustor completely, or stabilize at another location, such as transitioning from configuration (d) to (c) or from configuration (c) to (a). . The significance of aerodynamic straining of the flame sheet in the shear layer near the attachment point was apparently first discussed by Karlovitz et a/. [11]. They noted that holes appeared in the side of the flame as flow velocity increased, apparently due to local extinction. Similar observations of such holes in flames near blowoff are detailed in a review by Shanbogue et al. [12] and by Khosla and Smith [13]. A discussion of flame stabilization in shear layers can be found in Prof. Law's text [10]. Heat losses are also important to consider as it has a significant effect on the flame consumption speed and therefore on the ability of the flame to stabilize in a shear layer near combustor boundaries [14]. This paper presents an experimental investigation of the parametric dependence of how these different flame stabilization locations are influenced by fuel/air ratio and preheat temperature. It also describes extinction strain rate calculations of the transition points, and show that transitions between each flame stabilization point occurs at a roughly constant Damkohler number. These results show that flame heat release distribution is largely an extinction controlled problem, as the points where the flame can and cannot exist determine its possible configurations. Experimental Facility The basic experimental facility is sketched in Fig. 3 and can be divided into a reactant supply system, flow conditioning and fuel/air mixing (A), premixer (B), combustor (C), and exhaust (D) sections. C Figure 3. Schematic of test facility (not to scale) The air and natural gas supplied are regulated to a pressure of ~60 psi and ~25 psi respectively. Before entering the preconditioning section, the fuel and air mass flow rates are measured using calibrated orifice plates. The air supply then enters a heater allowing for the air to be preheated to temperatures ranging from 200 to 500 deg F. The air enters the test section as shown on the left side in Figure 3 from two sides. It then passes through a perforated plate with a blockage ratio ~50% (1) entering the preconditioning section (A) after which fuel is injected into the flow through eight equally spaced radial fuel injectors (2). There is a settling length of ~40 inches after fuel injection to allow for the fuel and air to mix adequately before passing through a honeycomb flow straightener (3) and a wire mesh (4) to straighten the flow and reduce turbulence in the flow. Upon entering the premixer section (B) the outer and inner diameter of the test section are smoothly transitioned to the desired dimensions of 2.44" for the outer diameter and 1.42" for the inner or centerbody diameter (6). The flow then passes through the aerodynamically designed swirler (5) which has a swirl number of ~0.7 before entering the combustor section (C) which consists of a ~5.3" diameter by 7.9" long quartz tube (Fig. 4). The exit of the combustor is a smoothly converging nozzle that contracts to a diameter of 4.4" providing a ~30% contraction in flow area. The flow then exits the exhaust section (D) to a centralized building exhaust system. The bulkhead and the centerbody are instrumented with thermocouples allowing for the temperatures to be monitored and recorded during testing. The test facility is operated with Reynolds numbers on the order of ~104-105. Figure 4. Photograph of combustor section of the test facility Flame Configuration Mapping Figure 5 shows images of the different flame configurations observed in this study. Configuration I is vortex breakdown bubble stabilized. Configurations II and III are both ISL stabilized. While stabilized at the same point, the resultant flame shape is quite different, as can be seen in the figures. Moreover, the transition between Configurations II and III is equally abrupt as the other transitions. This II-III transition is probably due to a bifurcation in vortex breakdown bubble characteristics, but this is still being explored. Configuration IV is stabilized in both the ISL and OSL. Again, the fact that the transitions between these various flame configurations are abrupt is of particular interest in defining the combustor's operating space both under steady state and transient operation; a flame configuration transition could result in a sudden and drastic change in the power output of the combustor amongst other things. I: VBB II: ISL III:|ISL Figure 5. Images of observed flame configurations IV: ISL/OSL These flame configuration studies were performed by fixing preheat temperature, lighting near the lean limit and increasing equivalence ratio while tracking the changes in flame configuration ( I ^ I I ^ I I I ^ I V ) . For example, at a preheat temperature of 200 deg F these transitions occurred at 0=0.59 and 0.71 for I I ^ III^ IV , respectively (Note due to high flow velocity and point of ignition, combustor stabilized in configuration II near the lean limit). The fuel/air ratio was then decreased and the transitions from I V ^ I I I ^ I I ^ I were recorded. As might be expected, significant hysteresis is observed. For example, consider configuration IV. The III^ IV transition occurs at 0.71, but once the flame is stabilized in the outer shear layer it can survive in the outer shear layer at a lower equivalence ratio of 0.59 (presumably due to heat feedback from the hot recirculating gases). These transition point measurements were repeated three times and the average results are plotted in Fig. 6. The transition equivalence ratio differed on the order of about 10% between the different test runs. These flame configuration maps were then repeated at other preheat temperatures, as shown in Fig. 6. As expected, the transitions occur at lower equivalence ratios as preheat temperature is increased. In addition to the influence of the flame location on combustor design criteria such as combustor head loadings and stability, the four different flame shapes each lead to different levels of combustion efficiency. As can be inferred from the shape of configurations I & II, the flame does not span across the entire cross section of the combustor, thus allowing fuel to bypass the flame front and leave the combustor in the form of unburned hydrocarbons. This point is highlighted by the fact that when the equivalence ratio is increased enough to cause a transition to flame configuration III from either I or II, there is a sharp increase in the exhaust temperature much larger than would be expected than due to the additional fuel entering the system alone. Again, compared to I & II, configuration III spans a larger length of the combustor transversely, thus allowing less opportunity for reactable mixture to leave the combustor without having passed through a flame region. Figure 6. Flame configuration map for premixer exit velocity of 70 m/s Extinction Strain Rate Calculations The ability of the flame to stabilize in a shear layer depends upon whether the local straining of the flame by the high shear is greater or less than the extinction strain rate, Kext. In order to obtain insight into the above data, detailed kinetics calculations of the mixture strain sensitivities were performed. It is important to note that the calculations are performed for an adiabatic, opposed flow flame with symmetric, premixed flows impinging upon each other. In reality, the flame stabilized in the shear layer is a non adiabatic, edge flame [14], with some possibility of exhaust gas dilution near the flame base. Nonetheless, as shown next, the simplified opposed flow calculations provide much insight into the observed fuel-air ratio/preheat temperature sensitivity. These calculations were performed using the Opposed Flow Extinction Model of CHEMKIN using the GRI-Mech 3.0. A typical calculation of the strain sensitivity of a lean, methane/air flame is shown in Fig. 7. Because of the negative Markstein number of this flame, the flame speed initially increases with strain, K. However, beyond some strain value the flame speed decreases and then extinguishes at K=Kext. 50 0 200 400 600 800 1000 Flame Strain [1/s] Figure 7. Laminar flame speed sensitivity to strain for methane-air mixture at preheat temperature of 200 deg F and 0=0.6 Figure 8 plots the dependence of the computed extinction strain rate upon fuel/air ratio at several preheat temperature values. As expected, the extinction strain rate increases with fuel/air ratio and preheat temperature. 0 4 ____________ 1____________ 1_____________1____________ 1____________ 1_____________J '2 0 0 250 300 350 400 450 500 Preheat Temperature [cleg F] Figure 8. Calculated extinction strain rate sensitivity to O and preheat temperature for methane-air mixtures. Figure 9 replots the same flame transition data shown in Fig. 6, along with an overlay of the extinction strain rate contours as shown in Fig. 8. As can be seen in the plot, the lines of constant Kext slope downward with increasing preheat temperature capturing the same general qualitative behavior of any given transition line with respect to preheat temperature. In addition, several transition lines follow the same slope as lines of constant Kext. Even those transitions, that deviate from a constant Kext, do not vary by any large magnitude across the preheat temperature range tested. Given the exponential sensitivity of Kext to preheat temperature and fuel/air ratio, these results suggest that the transitions occur at constant Kext to within experimental uncertainty. 200 250 300 3 50 400 450 500 Preheat Temperature [deg F] Figure 9. Flame configuration transition points overlayed with isolines of extinction strain rate as a function of preheat temperature at a premixer velocity of 70 m/s Future Work Ongoing work is proceeding in several parallel paths. First, further sensitivity studies are being performed at other flow velocities and geometric configurations. Second, detailed PIV measurements are being performed to characterize the flow field characteristics. Finally, additional computations are being performed that include heat losses and effects of product-reactant mixing. Acknowledgements The authors gratefully acknowledge the financial support of Pratt & Whitney for this research. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] Lefebvre, A.H., (1999) "Gas Turbine Combustion", 2nd edition, Taylor&Francis, New York Vandervort, C.L. (2001) "9 ppm no sub x /CO combustion system for ‘F' class industrial gas turbines", J. Engineering fo r Gas Turbines and Power, Vol. 123 (2), pp. 317-321 Schefer, R.W., Wicksall, D.M., Agrawal, A.K., (2002) "Combustion of Hydrogen-Enriched Methane in a Lean Premixed Swirl-Stabilized Burner", Proc. Combust. Inst., 29, pp. 843-851 Yang, V., Huang, Y., (2004) "Bifurcation of Flame Structure in a Lean-Premixed Swirl-Stabilized Combustor: Transition from Stable to Unstable Flame", Combust. Flame, Vol. 136, pp. 383-389 Gupta, A.K., Lilley, D.J., Syred, N., (1984) "Swirl Flows", Abacus Press, Tunbrige Wells, UK Wicksall, D.M., Agrawal, A.K., Schefer, R.W., Keller, J.O., (2004) "The Interaction of Flame and Flow Field in a Lean Premixed Swirl-Stabilized Combustor Operated on H2/CH4/Air", Proc. Combust. Inst., Vol. 30, pp. 2875-2883 Lu, X., Wang, S., Sung, H., Hsieh, S., Yang, V., (1999) "Large Eddy Simulations of Turbulent Swirling Flows Injected into a Dump Chamber", J. Fluid Mech., Vol. 527, pp. 171-195 Lieuwen, T., McDonell, V., Petersen, E., Santavicca, D., (2008) "Fuel Flexibility Influences on Premixed Combustor Blowout, Flashback, Autoignition, and Stability", J. Engineering fo r Gas Turbines and Power, Vol. 130 (1), pp. 954-961 C.K. Law, C.J. Sung, (2000) Progress in Energy and Combustion Science, 26 (4) 459-505 C.K. Law, (2006) "Combustion physics", Cambridge University Press, New York Karlovitz, B., Denniston Jr., D.W., Knapschaefer, D.H., Wells, F.E., (1953) "Studies on Turbulent Flame", Proc. Combust. Inst., Vol. 4, pp. 613-620 Shanbogue, S., Hussain, S., Lieuwen, T., (2009) "Lean Blowoff of Bluff Body Stabilized Flames: Scaling and Dynamics", Progress in Energy and Combustion Sciences, Vol. 35, pp.98-120 Khosla, S., Leach, T., and Smith, C., (2007) "Flame Stabilization and Role of von Karman Vortex Shedding Behind Bluff Body Flameholders", AIAA-2007-5653 Tay Wo Chong, L., Komarek, T., Zellhuber, M., Lenz, J., Hirsch, C., Polifke, W., (2009) "Influence of strain and heat loss on flame stabilization in a non-adiabatic combustor", Proc. Euro. Combust. Meeting 2009