Co-Axial Turbulent Diffusion Flames with Directed Oxygen Injection

There are practical and safety advantages in adding oxygen to as few input streams as possible. Additionally, oxy-coal combustion permits the concentration of oxygen in each burner stream to be controlled as an independent variable. Input concentrations of oxygen and the manner in which it is injected can have significant effects on the turbulent mixing and flame stability in oxy-coal burners. In this research, a pure oxygen stream was dedicated to the burner in order to investigate the effects of directed pure oxygen injection. This stream might contain a specific fraction of or all the oxygen required for combustion. The consequences of segregating all the input oxygen into one stream composed of 100% oxygen were determined using the co-axial burners with different oxygen stream configurations. Flame stability measurements were taken to evaluate the differences. Flame stability was quantified through flame probability density functions (PDF) of the stand-off distance (determined using photo-imaging techniques). The results led to determination of mechanistic concepts of flame stability in burners with directed oxygen injection. The PDFs obtained from these simplified prototype configurations led to physical insight into flame stabilization mechanisms. Flames with a pure oxygen stream in the center annulus are more stable than those with oxygen in the center pipe. Further investigation of this study added to the general knowledge of coal jet attachment mechanisms and the significant effects of fine coal particles and their radial transportation by large eddies on flame stability.

[1] CO-AXIAL TURBULENT DIFFUSION FLAMES WITH DIRECTED OXYGEN INJECTION Dadmehr Rezaei [e-mail], Jost O.L. Wendt [e-mail] Department of Chemical Engineering and Institute for Clean and Secure Energy, University of Utah, Salt Lake City, UT 84112, USA Abstract: There are practical and safety advantages in adding oxygen to as few input streams as possible. Additionally, oxy-coal combustion permits the concentration of oxygen in each burner stream to be controlled as an independent variable. Input concentrations of oxygen and the manner in which it is injected can have significant effects on the turbulent mixing and flame stability in oxy-coal burners. In this research, a pure oxygen stream was dedicated to the burner in order to investigate the effects of directed pure oxygen injection. This stream might contain a specific fraction of or all the oxygen required for combustion. The consequences of segregating all the input oxygen into one stream composed of 100% oxygen were determined using the co-axial burners with different oxygen stream configurations. Flame stability measurements were taken to evaluate the differences. Flame stability was quantified through flame probability density functions (PDF) of the stand-off distance (determined using photo-imaging techniques). The results led to determination of mechanistic concepts of flame stability in burners with directed oxygen injection. The PDFs obtained from these simplified prototype configurations led to physical insight into flame stabilization mechanisms. Flames with a pure oxygen stream in the center annulus are more stable than those with oxygen in the center pipe. Further investigation of this study added to the general knowledge of coal jet attachment mechanisms and the significant effects of fine coal particles and their radial transportation by large eddies on flame stability. [2] Keywords: Oxy-coal Combustion, Pure Oxygen Injection, Oxygen stream configuration, Flame Stability, Co-axial Turbulent Jets 1. Introduction Oxy-coal combustion, in which coal is burned with oxygen rather than air with flue gas recycle to modulate flame temperatures, has the potential for producing concentrated CO2 flue gas streams suitable for compression and sequestration. Combustion of pulverized coal in mixtures of oxygen and recycled flue gas is different from that in air, since under oxy-coal conditions the input oxygen concentration can be varied independently. Also oxygen need not be distributed evenly across all input streams, but can be introduced in at different concentrations in each stream. There are practical and safety advantages in adding the oxygen to as few input streams as possible, with special safety benefits accrued from having all the oxygen required in a single pure oxygen stream. Zhang [1] quantified flame stand-off distance in co-axial pulverized coal turbulent diffusion flames as a function of partial pressure of oxygen in the primary jet. He also showed that at high enough PO2 in the secondary jet it was possible to stabilize those flames with zero oxygen in the primary jet. The present paper builds on this previous work, but now focuses on the near field aerodynamic impacts of directed pure oxygen in a segregated stream. As in previous work [1], this study focuses on interactions between coal particle ignition and turbulent mixing in co-axial jets, rather than on ignition chemistry or turbulent mixing by themselves. The directed pure oxygen in this study is contained either 1. In an annulus located between a central coal/CO2 circular jet and an outer O2/CO2 secondary oxidant stream annulus (Configuration A), or 2. In a central pipe surrounded by a coal/CO2 middle annulus, surrounded again by an O2/CO2 secondary oxidant stream annulus (Configuration B). [3] Shaddix and Molina [2] claimed the presence of CO2 retards single particle coal ignition. Particle devolatilization proceeds faster with higher O2 concentration, and decreases with the use of CO2 because of the influence of these two species on the mass diffusion of O2 and volatiles. Qiao [3] and his colleagues proved the impact of thermal conductivity of CO2 as another factor for the lower ignition of coal in O2/CO2 environments. Wall and Buhre [4] in their review present the impact of CO2 on coal ignition retardation as well as improvement of flame stability due to pure O2 Injection. Molina [5] found oxygen concentration decreases ignition delay in a meaning that could be used to counteract the retardant effect of CO2. Kimura [6] showed that the flame propagation speed is much smaller in O2/CO2 atmosphere than in O2/N2 and O2/Ar, and that in O2/Ar was the highest. The propagation speed increases as the oxygen concentration increases in all cases as expected. Nozaki injected 15% to 20% of the total oxygen into the furnace directly [7]. His group at IHI found HCN formation in near burner zone increases in O2/CO2 with oxygen injection, and he claimed devolatilization becomes more active at high gas temperature realized near the burner by oxygen injection. Forstall [8] studied the impact of velocity ratio on the turbulent mixing in a coaxial jet containing central and annular streams. He concluded material diffuses more rapidly than momentum that the principle independent variable determining the shape of the mixing regime is the velocity ratio. Chigier and Beer focused on the region near the nozzle in double concentric jets. They considered two streams as two different central cores and described how two streams emerge based on their fluid entrainment due to velocity difference [9-11] . In addition, they provided plots that define the length of each stream before they emerge as a function of velocity ratio for co-axial jets. In another study, Beer [12] reported when the secondary velocity is low and the density of the secondary fluid is considerably higher than that of the recirculating fluid, more recirculation and less secondary air is entrained in the early part of the primary jet. This means coal particles will be heated faster, and it has a large effect on the stability of the [4] flame. Durano [13] measured mean velocity of a co-axial jet for three different velocity ratios. He reported co-axial jets obtain a self-preserving state faster than single round jets. The current work builds upon these pioneering contributions on coal ignition and coaxial jet structure referenced above, but with the special focus on how directed pure oxygen, according to Configuration A and Configuration B above, affects flame stand-off distance and flame length. 2. Experimental 2.1. Coal selection and sample preparation The coal applied for this study was Illinois #6 Bituminous coal. The ultimate and proximate analysis of the coal as received (AR) is provided in Table 4. According to particle size distribution calculation of the coal used in this experiment, maximum particle size was in the range of 150μm to 200μm, and the mass average diameter was 62μm. 2.2. Oxy-Coal Combustion Furnace The Oxy-Fuel Combustor (OFC) at the University of Utah is a100kW down-fired furnace [1] with the capability of both air- and oxy-combustion with either once through CO2 or recycled flue gas ongoing various flue gas cleaning options. For the present study only, once through, pure CO2 from a tank was used. The OFC consists of a near field aerodynamic (burner zone) or the main chamber followed by a radiant zone located below the main chamber; Burner zone dimensions are 0.61m I.D, 0.91m O.D. with 1.22m as the height. In order to reduce the heat lost, the burner zone is insulated by 76mm thick [5] fiberboards that tolerate high temperatures up to 1700K. In addition, twenty-four 840W electric heaters embedded in fiberboard allow us to control the wall temperature at any desired temperature up to 1111K. In this study, the wall temperature was kept at 1283K. There are K-Type thermocouples all along the furnace to monitor the temperature. In order to have an optical access to the flame regarding optical diagnostic, four quartz windows have been provided on the quadrants of the cylindrical chamber. Additional details regarding the OFC is provided elsewhere[1]. 2.3. Methodology of flame stability measurement Flame stability measurement employed similar methods as those used in Ref. [1]. Stand-off distance and flame length are defined by the position of the luminous zone, as observed by the human eye, but recorded on 6000 photo images at 30fps with an exposure of 8.3ms. Each run had up to 5 replicates taken. Therefore, flame stand-off distance is defined as the distance starting from the tip of the burner to the luminous and visible part of the flame and it is described in detail elsewhere [1]. 2.4. Burners and combustion operating conditions In the previous studies, coaxial double annulus burners with only two streams have been used [1]. However, in this research, a pure oxygen stream was dedicated to the burner in order to investigate the effects of directed pure oxygen injection. This stream might contain a specific fraction or 100% of all the oxygen required for combustion for the combustion operating condition. The primary stream is a transport medium carrying pulverized coal particles in the chamber and for the experiments here contained only CO2, with no oxygen. Directed oxygen and transport streams were at room temperature. . The secondary stream is a mixture of O2 and CO2, and plays a significant role on moderating the [6] combustion temperature as well as entrainment of the flame jet. Secondary stream temperature is controlled at 489K by a gas preheater. Coal feed rate, primary stream CO2 (30% of the total CO2 required), overall volume fraction O2 (40%), stoichiometric ratio (1.15), and wall temperature (1283K) were all kept constant. These and other variables are shown on Table 3. 2.5. Experimental methodology for Configuration A: Burners with oxygen stream in middle annular stream In this set, six burners were used to allow stream velocities to remain constant, as shown on Table 1. As case numbers increase, O2 is added from the secondary stream to the pure oxygen stream while the overall O2 is maintained constant. It is important to notice that in burner number 6 all of the oxygen required is being introduced in the pure directed oxygen stream. The primary stream (Coal + CO2) is located in the center of the burner. The annulus around the primary stream is the directed oxygen stream, and the furthest outer annulus contains the secondary stream. The velocity of the pure oxygen stream and the primary stream are kept approximately equal in order to delay mixing, leading potentially to longer flames with a more uniform heat distribution. The secondary stream velocity is 2.5 times higher than that of the other two streams, creating external recirculation in the upper chamber and maintaining jet entrainment (IFRF Type-0 flame). Multiple burners with various appropriate dimensions allowed the stream velocities to be maintained constant such that aerodynamics mixing effects were unchanged as much as possible, as shown as Table 1. 2.6. Experimental methodology for Configuration B: Burners with oxygen stream in centrally located pipe. [7] For Configuration B, burners are also triple concentric burners but now with a pure oxygen stream that is located in the center pipe. The primary stream that carries the coal particles is located in the inner (middle) annulus, and the outer annulus is the secondary (O2/CO2) stream. As for Configuration A, six cases cover different fractions of total oxygen in the oxygen stream located in the center from 0% to 100%. In the very last case, all the oxygen is once again in one segregated stream (now at the center). Velocity ratios of the stream are similar as for Configuration A, to minimize the aerodynamic differences as shown in Table 2. 3. Results Figure 1 shows the PDF of flame stand-off distance and also that for the flame length for Configuration A with no directed oxygen in the middle annulus, and all the oxygen required for combustion mixed in with the secondary CO2 in the outer annulus (Case 1 in Table 1). FO2 is the fraction of the total oxygen entering as the directed pure oxygen stream, with (1- FO2) being the fraction mixed in with secondary CO2 in the outer annulus. These data can be used to validate simulations of a limiting case with no directed oxygen injection, but with the appropriate hardware in place. . The plot on the left of Figure 1 represents the probability of the distance from the tip of the burner until the luminous part of the flame. Error bars represent the standard deviation of the replicates. The plot on the right shows the PDF of flame length. Flame length in this paper is defined as the luminous length of the flame. It may not include the char burn out zone. Two main effects of Oxygen configuration and oxygen fraction on flame stability were investigated in this study. The contribution of oxygen stream configuration to flame stability was explored by situating the oxygen stream in different locations namely Configuration A and Configuration B. Ultimately, the effects of injection of all the oxygen in a segregated stream were studied. Figure 2 shows the flame [8] stability results for the both oxygen configurations A (left) and B (right) at different fractions (FO2) of the total inlet O2 being placed in the directed pure oxygen stream. In both configurations, placing increased amounts of the total oxygen in the directed pure oxygen stream amplifies the stability of the flame; however, this effect is considerably more significant when oxygen is located in the middle annular stream (Configuration A), rather than in the central pipe (Configuration B). For Configuration A (Left side) the stand-off distance of the flame does not change appreciably until FO2 = 64%. At FO2 = 75%, the first indications of a stable attached flame can be appear. The flame attachment is in a transient status at this point. However, at higher fractions of oxygen 85% and 100%, extremely stable and attached flames were generated. It is important to notice that in the last condition, all the oxygen was located in a single stream. The PDF plots on the right of the Figure 2 present the data for Configuration B. As with Configuration A, flame stand-off distance diminishes with increasing FO2 in the central oxygen stream. However, the flame never became fully attached even when FO2 = 100%. These data result that injection of oxygen in the inner annulus creates more flame stability that when it is injected in the center of the burner. Figure 3 shows results depicting flame length, defined by the length of the luminous zone. The PDF's on the left present the data of Configuration A and the PDF's on the right provides flame length data of the Configuration B. In both types of Configurations, flame length increased as the fraction of total oxygen placed in the directed pure oxygen stream increased. This increase of flame length was more significant with burners containing directed oxygen injection in the inner annulus (Configuration A). The red bar in the plots represents the probability of the flames that were longer than the length of the window, and they were not possible to be measured. It was found, although the stand-off distance PDF's of the cases FO2 = 85% and FO2 = 100% do not manifest a difference; the flame length PDF's show [9] significant changes in the flame length of these two cases. By increasing oxygen fraction, the probability of the flames with a length longer than the window increased. 4. Discussion and Conclusions The study was carried out to help understanding the effect of directed pure oxygen and the manner in which it is injected on the length and stability of pulverized oxy-coal flames stabilized on triple concentric co-axial burners with no swirl. Two configurations of burners were used in this research; Configuration A with directed pure O2 in a middle annulus and Configuration B with directed pure O2 in the center pipe. In both types, increase of FO2, the fraction of total oxygen in the directed oxygen stream, increased the stability of the flame; however, it was more significant in Configuration A burners containing the pure oxygen stream in the inner annulus such that fully attached and stable fames were obtained at FO2 = 85% and higher. In Configuration B burners, even at FO2 = 100% the flame was detached and unstable. We propose two hypotheses that might account for the fact the directed oxygen in the middle annulus is more effective in attaching Type 0 oxy-coal flames than directed oxygen in the center pipe. There are two explanations for the flame behavior. The first is concerned with coal particle dispersion effects on flame stability. Budilarto [14] used Laser Doppler Velocimetry to show that smaller particles migrated to the edge of co-axial two phase turbulent jets. Preliminary high performance computer simulations [15] using Large Eddy Simulation (LES) suggest that coal particle ignition occurs at clusters of small particles many of which have been transported by large eddies radially outwards. The experimental results of Zhang et al [16] suggest that [10] coal jet ignition in coaxial oxy-coal flames similar to those investigated here is determined by a mechanism through which large eddies project clusters of particles and their surrounding fluid radially outwards. The new results on flame attachment for the two configurations are consistent with these mechanisms. With directed oxygen in an annulus surrounding the coal jet (Configuration A), (smaller) coal particles migrate from the center towards and through the pure oxygen stream, thus providing a suitable environment for early ignition and flame attachment. With directed oxygen in the center (Configuration B), the (smaller) coal particles are now directed away from the high oxygen concentration regions, thus delaying ignition and consequent flame attachment. Crowe recommended evaluation of Stokes Number [17] in turbulent two phase jets is: = ! ! where τp is response time of the particle and τf is defined as fluid time scale that is a function of characteristic length and velocity of the fluid [17-19]. We calculate τp ≃ 7.5ms, and τf = 5.5 for Configuration B, yielding St≃1.3, suggesting that the larger coal particles will penetrate through the larger turbulent eddies, while the smaller ones will be carried radially outwards by large eddies. Kennedy in Moody [20] in their research found St has the highest value in the near field, and it becomes much smaller downstream in the jet. Thus the calculations support the fact that particle paths of the smaller particles in Configuration B are transported away from the oxygen in the center and towards the oxygen deficient zones near the outside. The second hypothesis to explain why Configuration A was more stable than Configuration B is related to heat transfer effects on coal ignition and shielding of wall radiation to the oxygen/coal free shear layer interface by the coal particles in the coal stream [11] 5. Acknowledgment This material is based upon work supported by the Department of Energy under award number DE-NT0005015. Assistance in running the experiment, and preparing the data were provided by David Wagner, Ryan Okerlund, Charles German, Michael Newton, and Travis LeGrande. 6. References [1] J. Zhang, K.E. Kelly, E.G. Eddings, J.O.L. Wendt, Proc. Combust. Inst. 33 (2011) 3375-3382. [2] C.R. Shaddix, A. Molina, Proc. Combust. Inst. 32 (2009) 2091-2098. [3] Y. Qiao, L. Zhang, E. Binner, M. Xu, C.-Z. Li, Fuel 89 (2010) 3381-3387. [4] B.J.P. Buhre, L.K. Elliott, C.D. Sheng, R.P. Gupta, T.F. Wall, Progress in Energy and Combustion Science 31 (2005) 283-307. [5] C.R. Shaddix, E.S. Hecht, A. Molina, Ignition of a Group of Coal Particles in Oxyfuel Combustion with CO2 Recirculation, Clearwater, FL, 2009. [6] T. Kiga, S. Takano, N. Kimura, K. Omata, M. Okawa, T. Mori, M. Kato, Energy Conversion and Management 38 (1997) 129-134. [7] T. Nozaki, S.-I. Takano, T. Kiga, K. Omata, N. Kimura, Energy 22 (1997) 199-205. [8] W. Forstall, A.H. Shapiro, Journal of Applied Mechanics 18 (1951) 219-228. [9] N. Chigier, J. Beer, Journal of Basic Engineering 86 (1964) 797. [10] N. Chigier, J. Beer, Journal of Basic Engineering 86 (1964) 788. [11] J. Beer, N. Chigier, Combustion and Flame 12 (1968) 575-586. [12] J. Beer, N. Chigier, K. Lee, Proc. Combust. Inst. 9 (1963) 892-900. [13] D. Durao, J. Whitelaw, Journal of Fluids Engineering 95 (1973) 467. [14] S.G. Budilarto, An Experimental Study on Effects Of Fluid Aerodynamics and Particle Size Distribution in Particle Laden Flows, Purdue University, West Lafayette, Indiana, USA, 2003. [15] J. Pedel, J. Zhang, J. Thornock, J.O.L. Wendt, P. Smith, Ignition of Coaxial Turbulent Oxy-coal Jet Flames: Experiments and Simulations Collaboration, Clearwater, Florida, 2011. [12] [16] J. Zhang, K.E. Kelly, E.G. Eddings, J.O.L. Wendt, International Journal of Greenhouse Gas Control 5 (2011) S47-S57. [17] C.T. Crowe, R.A. Gore, T.R. Troutt, Particulate Science and Technology 3 (1985) 149-158. [18] E. Longmire, J. Eaton, Journal of Fluid Mechanics 236 (1992) 217. [19] C. Crowe, J. Chung, Progress in Energy and Combustion (1988). [20] I.M. Kennedy, M.H. Moody, Experimental Thermal and Fluid Science 18 (1998) 11-26. [13] Figure 1. Stand-off distance and length of the flame for zero oxygen in the directed O2 stream for Configuration A, Case 1, Table 1 [14] Figure 2. Flame stand-off distance PDF's. Left: Configuration A, Right: Configuration B [15] Figure 3. Flame length PDF's. Left: Configuration A, Right: Configuration B [16] Table 1 Flame aerodynamic conditions for burners with inner annular oxygen stream (Configuration A) Case FO2 inner O2 annulus Prim. Vel. (m/s) O2 stream Vel. (m/s) Second. Vel. (m/s) Prim. AS (cm2) O2 Stream AS (cm2) Second. AS (cm2) 1 0.0% 6.37 0.0 14.9 1.9604 6.0744 2 23.0% 6.37 6.10 15.8 1.9604 1.0653 5.0665 3 55.0% 6.34 6.40 15.8 1.9604 2.3788 3.9442 4 75.0% 6.38 6.10 14.6 1.9478 3.2475 3.9515 5 85.0% 6.38 6.70 14.6 1.9478 3.5059 3.6361 6 100.0% 6.38 6.50 15.0 1.9478 4.2651 3.0994 [17] Table 2 Flame aerodynamic conditions for burners with central oxygen stream (Configuration B) Case FO2 center O2 stream Prim. Vel. (m/s) O2 stream Vel. (m/s) Second. Vel. (m/s) Prim. AS (cm2) O2 Stream AS (cm2) Second. AS (cm2) 1 0.0% 6.37 0.0 14.9 1.9604 6.0744 2 21.0% 6.40 6.28 16.43 1.9801 0.9066 4.9498 3 44.0% 6.32 6.22 16.27 2.0005 1.9604 4.3723 4 64.0% 6.00 6.26 15.8 1.9883 2.9267 3.9442 5 80.0% 6.26 6.27 16.21 1.9883 3.5244 3.4077 6 100.0% 6.30 6.26 16.17 1.9883 4.3825 2.8689 [18] Table 3 Constant parameter of combustion operating conditions Parameter Quantity Total S.R. 1.15 Pri. Stream Vel. 305 K Pri. Stream Temp. 6.3 (m/s) O2 Stream Vel. 305 K O2 Stream Temp. 6.3 ± 0.3 (m/s) Sec. Stream Vel. 15.4 ± 1.4 (m/s) Sec. Stream Vel. 489 K Wall Temp. 1283 K Overall O2 V.F. 40% Flame Tadb 2485.5 K [19] Table 4 Ultimate and proximate analysis of Illinois #6 Ultimate (AR) (%) Proximate (AR) (%) C H O N S Moisture Ash V.M. F.C. 64.67 4.51 8.07 1.12 3.98 9.65 7.99 36.78 45.58