Laboratory Studies of the Flow Field Characteristics of Low-Swirl Injectors for Adaptation to Fuel-Flexible Turbines

Author(s):  
R. K. Cheng ◽  
D. Littlejohn ◽  
W. A. Nazeer ◽  
K. O. Smith

The low-swirl injector (LSI) is a simple and cost-effective lean premixed combustion method for natural-gas turbines to achieve ultralow emissions (<5 ppm NOx and CO) without invoking tight control of mixture stoichiometry, elaborate active tip cooling, or costly materials and catalysts. To gain an understanding of how this flame stabilization mechanism remains robust throughout a large range of Reynolds numbers, laboratory experiments were performed to characterize the flowfield of natural-gas flames at simulated partial load conditions. Also studied was a flame using simulated landfill gas of 50% natural gas and 50% CO2. Using particle image velocimetry, the nonreacting and reacting flowfields were measured at five bulk flow velocities. The results show that the LSI flowfield exhibits similarity features. From the velocity data, an analytical expression for the flame position as function of the flowfield characteristics and turbulent flame speed has been deduced. It shows that the similarity feature coupled with a linear dependency of the turbulent flame speed with bulk flow velocity enables the flame to remain relatively stationary throughout the load range. This expression can be the basis for an analytical model for designing LSIs that operate on alternate gaseous fuels such as slower burning biomass gases or faster burning coal-based syngases.

Author(s):  
R. K. Cheng ◽  
D. Littlejohn ◽  
W. A. Nazeer ◽  
K. O. Smith

The low-swirl injector (LSI) is a simple and cost-effective lean premixed combustion method for natural-gas turbines to achieve ultra-low emissions (&lt; 5 ppm NOx and CO) without invoking tight control of mixture stoichiometry, elaborate active tip cooling or costly materials and catalysis. To gain an understanding of how this flame stabilization mechanism remains robust throughout a large range of Reynolds numbers, laboratory experiments were performed to characterize the flowfield of natural gas flames at simulated partial load conditions. Also studied was a flame using simulated landfill gas of 50% natural gas and 50% CO2. Using Particle Image Velocimetry (PIV), the non-reacting and reacting flowfields were measured at five bulk flow velocities. The results show that the LSI flowfield exhibits similarity features. From the velocity data an analytical expression for the flame position as function of the flowfield characteristics and turbulent flame speed have been deduced. It shows that the similarity feature coupled with a linear dependency of the turbulent flame speed with bulk flow velocity enable the flame to remain relatively stationary throughout the load range. This expression can be the basis for an analytical model for designing LSIs that operate on alternate gaseous fuels such as slower burning biomass gases or faster burning coal-based syngases.


Author(s):  
S. Ravi ◽  
A. Morones ◽  
E. L. Petersen ◽  
F. Güthe

Natural gas is the primary fuel for stationary, powergeneration gas turbines, and it is necessary to understand its combustion characteristics under engine-relevant (turbulent) conditions. Since its composition varies depending on the fuel source, a natural gas surrogate (NG 18% C2+) and admixtures with H2 have been utilized recently by the authors to aid chemical kinetics modeling using ignition delay times and laminar flame speed experiments. The present study focused on measuring turbulent flame speeds (displacement speeds) of natural gas (NG2) and methane with H2 using a fan-stirred flame bomb. The apparatus is a closed, cylindrical chamber fitted with four radial impellers that generate a central spherical volume of homogeneous and isotropic turbulence with negligible mean flow. Schlieren imaging was used to visually track the growth of the spherically expanding turbulent kernels during the constant-pressure period. The turbulence levels were fixed at an average RMS intensity level of 1.5 m/s and at an integral length scale of 27 mm. Turbulent flame speeds (ST,0.1) of NG2 blends were measured over a wide range of equivalence ratios between 0.7 and 1.3. ST,0.1 for the natural gas surrogate closely matched with those of methane for near-stoichiometric mixtures. However, preferential-diffusion effects (fuel effects) were observed under turbulent conditions for off-stoichiometric cases. The effects of hydrogen addition on the turbulent flame speeds of NG2 (25/75 and 50/50 (by volume) blends of H2/NG2) were also investigated and were compared with the flame speeds reported in a recent paper by the authors (ASME GT2014-26742) on the effects of hydrogen addition to turbulent flame speeds of methane. The effect of the hydrogen addition was to increase the turbulent flame speed (by about a factor of two for 50% H2 addition), although this effect was much more pronounced for the lean and stoichiometric mixtures. Interestingly, the flame speeds (both laminar and turbulent) of the CH4 blends with H2 were slightly larger than those for the NG2 blend at equivalent conditions, or about 10–20% larger at 50% H2 addition. This behavior can be explained kinetically by the increased importance of the inhibiting reaction CH3 + H (+M) ↔ CH4 (+M), where ethane oxidation produces more CH3 radicals than methane at similar conditions.


Author(s):  
Martin Zajadatz ◽  
Nikolaos Zarzalis ◽  
Wolfgang Leuckel

In gas turbine combustion application, there is a strong tendency towards high volumetric heat release rates without compromising ignition stability and the requirement of low emission concentrations of NOx, CO and unburned hydrocarbons. In order to meet these demands for industrial gas turbines the lean premixed combustion concept has been developed. In the scope of this paper fundamental experimental work, which has been carried out in order to analyze the important topic of turbulence/chemistry interaction on a semi-technical scale, will be reported. The turbulent intensity and length scales have been varied by a generic burner system, which consists of four geometrically scaled burners. At atmospheric pressure conditions more than 700 Bunsen type flames in a Reynolds number range from 21000 to 128000 have been investigated. Gas/air mixture preheating has been included in the tests as a typical boundary condition for combustion in gas turbines. The natural gas was blended with 25 vol. % and 50 vol. % hydrogen in order to alter the kinetics of the fuel gas. The influence of the aforementioned parameters on the turbulent flame speed were assessed and compared with existing correlations for the turbulent flame speed. Special emphasis has been taken on the influence of gas/air mixture preheating and kinetics.


Fluids ◽  
2019 ◽  
Vol 4 (3) ◽  
pp. 146 ◽  
Author(s):  
Aaron Endres ◽  
Thomas Sattelmayer

Boundary layer flashback from the combustion chamber into the premixing section is a threat associated with the premixed combustion of hydrogen-containing fuels in gas turbines. In this study, the effect of pressure on the confined flashback behaviour of hydrogen-air flames was investigated numerically. This was done by means of large eddy simulations with finite rate chemistry as well as detailed chemical kinetics and diffusion models at pressures between 0 . 5 and 3 . It was found that the flashback propensity increases with increasing pressure. The separation zone size and the turbulent flame speed at flashback conditions decrease with increasing pressure, which decreases flashback propensity. At the same time the quenching distance decreases with increasing pressure, which increases flashback propensity. It is not possible to predict the occurrence of boundary layer flashback based on the turbulent flame speed or the ratio of separation zone size to quenching distance alone. Instead the interaction of all effects has to be accounted for when modelling boundary layer flashback. It was further found that the pressure rise ahead of the flame cannot be approximated by one-dimensional analyses and that the assumptions of the boundary layer theory are not satisfied during confined boundary layer flashback.


Author(s):  
P. Griebel ◽  
R. Bombach ◽  
A. Inauen ◽  
R. Scha¨ren ◽  
S. Schenker ◽  
...  

The present experimental study focuses on flame characteristics and turbulent flame speeds of lean premixed flames typical for stationary gas turbines. Measurements were performed in a generic combustor at a preheating temperature of 673 K, pressures up to 14.4 bars (absolute), a bulk velocity of 40 m/s, and an equivalence ratio in the range of 0.43–0.56. Turbulence intensities and integral length scales were measured in an isothermal flow field with Particle Image Velocimetry (PIV). The turbulence intensity (u′) and the integral length scale (LT) at the combustor inlet were varied using turbulence grids with different blockage ratios and different hole diameters. The position, shape, and fluctuation of the flame front were characterized by a statistical analysis of Planar Laser Induced Fluorescence images of the OH radical (OH-PLIF). Turbulent flame speeds were calculated and their dependence on operating conditions (p, φ) and turbulence quantities (u′, LT) are discussed and compared to correlations from literature. No influence of pressure on the most probable flame front position or on the turbulent flame speed was observed. As expected, the equivalence ratio had a strong influence on the most probable flame front position, the spatial flame front fluctuation, and the turbulent flame speed. Decreasing the equivalence ratio results in a shift of the flame front position farther downstream due to the lower fuel concentration and the lower adiabatic flame temperature and subsequently lower turbulent flame speed. Flames operated at leaner equivalence ratios show a broader spatial fluctuation as the lean blow-out limit is approached and therefore are more susceptible to flow disturbances. In addition, because of a lower turbulent flame speed these flames stabilize farther downstream in a region with higher velocity fluctuations. This increases the fluctuation of the flame front. Flames with higher turbulence quantities (u′, LT) in the vicinity of the combustor inlet exhibited a shorter length and a higher calculated flame speed. An enhanced turbulent heat and mass transport from the recirculation zone to the flame root location due to an intensified mixing which might increase the preheating temperature or the radical concentration is believed to be the reason for that.


Author(s):  
Pradeep Parajuli ◽  
Tyler Paschal ◽  
Mattias A. Turner ◽  
Eric L. Petersen ◽  
Waruna D. Kulatilaka

Abstract Natural gas is a major fuel source for many industrial and power-generation applications. The primary constituent of natural gas is methane (CH4), while smaller quantities of higher order hydrocarbons such as ethane (C2H6) and propane (C3H8) can also be present. Detailed understanding of natural gas combustion is important to obtain the highest possible combustion efficiency with minimal environmental impact in devices such as gas turbines and industrial furnaces. For a better understanding the combustion performance of natural gas, several important parameters to study are the flame temperature, heat release zone, flame front evolution, and laminar flame speed as a function of flame equivalence ratio. Spectrally and temporally resolved, high-speed chemiluminescence imaging can provide direct measurements of some of these parameters under controlled laboratory conditions. A series of experiments were performed on premixed methane/ethane-air flames at different equivalence ratios inside a closed flame speed vessel that allows the direct observation of the spherically expanding flame front. The vessel was filled with the mixtures of CH4 and C2H6 along with respective partial pressures of O2 and N2, to obtain the desired equivalence ratios at 1 atm initial pressure. A high-speed camera coupled with an image intensifier system was used to capture the chemiluminescence emitted by the excited hydroxyl (OH*) and methylidyne (CH*) radicals, which are two of the most important species present in the natural gas flames. The calculated laminar flame speeds for an 80/20 methane/ethane blend based on high-speed chemiluminescence images agreed well with the previously conducted Z-type schlieren imaging-based measurements. A high-pressure test, conducted at 5 atm initial pressure, produced wrinkles in the flame and decreased flame propagation rate. In comparison to the spherically expanding laminar flames, subsequent turbulent flame studies showed the sporadic nature of the flame resulting from multiple flame fronts that were evolved discontinuously and independently with the time. This paper documents some of the first results of quantitative spherical flame speed experiments using high-speed chemiluminescence imaging.


Author(s):  
Ehsan Abbasi-Atibeh ◽  
Sandeep Jella ◽  
Jeffrey M. Bergthorson

Sensitivity to stretch and differential diffusion of chemical species are known to influence premixed flame propagation, even in the turbulent environment where mass diffusion can be greatly enhanced. In this context, it is convenient to characterize flames by their Lewis number (Le), a ratio of thermal-to-mass diffusion. The work reported in this paper describes a study of flame stabilization characteristics when the Le is varied. The test data is comprised of Le ≪ 1 (Hydrogen), Le ≈ 1 (Methane), and Le > 1 (Propane) flames stabilized at various turbulence levels. The experiments were carried out in a Hot exhaust Opposed-flow Turbulent Flame Rig (HOTFR), which consists of two axially-opposed, symmetric turbulent round jets. The stagnation plane between the two jets allows the aerodynamic stabilization of a flame, and clearly identifies fuel influences on turbulent flames. Furthermore, high-speed Particle Image Velocimetry (PIV), using oil droplet seeding, allowed simultaneous recordings of velocity (mean and rms) and flame surface position. These experiments, along with data processing tools developed through this study, illustrated that in the mixtures with Le ≪ 1, turbulent flame speed increases considerably compared to the laminar flame speed due to differential diffusion effects, where higher burning rates compensate for the steepening average velocity gradient, and keeps these flames almost stationary as bulk flow velocity increases. These experiments are suitable for validating the ability of turbulent combustion models to predict lifted, aerodynamically-stabilized flames. In the final part of this paper, we model the three fuels at two turbulence intensities using the FGM model in a RANS context. Computations reveal that the qualitative flame stabilization trends reproduce the effects of turbulence intensity, however, more accurate predictions are required to capture the influences of fuel variations and differential diffusion.


Author(s):  
Alireza Kalantari ◽  
Nicolas Auwaijan ◽  
Vincent McDonell

Abstract Lean-premixed combustion is commonly used in gas turbines to achieve low pollutant emissions, in particular nitrogen oxides. But use of hydrogen-rich fuels in premixed systems can potentially lead to flashback. Adding significant amounts of hydrogen to fuel mixtures substantially impacts the operating range of the combustor. Hence, to incorporate high hydrogen content fuels into gas turbine power generation systems, flashback limits need to be determined at relevant conditions. The present work compares two boundary layer flashback prediction methods developed for turbulent premixed jet flames. The Damköhler model was developed at University of California Irvine (UCI) and evaluated against flashback data from literature including actual engines. The second model was developed at Paul Scherrer Institut (PSI) using data obtained at gas turbine premixer conditions and is based on turbulent flame speed. Despite different overall approaches used, both models characterize flashback in terms of similar parameters. The Damköhler model takes into account the effect of thermal coupling and predicts flashback limits within a reasonable range. But the turbulent flame speed model provides a good agreement for a cooled burner, but shows less agreement for uncooled burner conditions. The impact of hydrogen addition (0 to 100% by volume) to methane or carbon monoxide is also investigated at different operating conditions and flashback prediction trends are consistent with the existing data at atmospheric pressure.


Sign in / Sign up

Export Citation Format

Share Document