0

ASME Conference Presenter Attendance Policy and Archival Proceedings

2014;():V04BT00A001. doi:10.1115/GT2014-NS4B.
FREE TO VIEW

This online compilation of papers from the ASME Turbo Expo 2014: Turbine Technical Conference and Exposition (GT2014) represents the archival version of the Conference Proceedings. According to ASME’s conference presenter attendance policy, if a paper is not presented at the Conference, the paper will not be published in the official archival Proceedings, which are registered with the Library of Congress and are submitted for abstracting and indexing. The paper also will not be published in The ASME Digital Collection and may not be cited as a published paper.

Commentary by Dr. Valentin Fuster

Combustion, Fuels and Emissions

2014;():V04BT04A001. doi:10.1115/GT2014-25919.

With increasing computational power, Large Eddy Simulation (LES) is being widely used to study and develop a better understanding of turbulent combustion. A variety of subgrid combustion models have been proposed to investigate premixed combustion in LES. One of the physical aspects that can be exploited, are the fractal characteristics of premixed flames which have been confirmed in several experimental works. In this work the performance of a simplified version of an already established sub-grid flame surface density combustion model, which is based on the fractal characteristics of the flame surface is investigated. The original model was derived on the basis of theoretical models, experimental and direct numerical simulations databases and its performance was validated with data from the available literature. The simplifications to the established flame surface density model are discussed, and its performance is validated in comparison to the original model. Secondly numerical simulations with both models at conditions typical for spark-ignition engines and industrial gas turbines are validated against experimental data. It is found that both original and simplified models are suitable for LES of low to high turbulent premixed combustion in ambient and elevated pressure conditions.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A002. doi:10.1115/GT2014-25988.

In the TechCLEAN project of JAXA, experimental research had been conducted to develop a combustor for a small aircraft engine. The combustor was tuned to show the behavior of the Rich-Lean combustion through tests under atmospheric and practical conditions. Finally, through full annular combustion experiments under practical conditions, the combustor was tuned to reduce NOx emissions to almost 40% of the ICAO CAEP4 standard, also sustaining low CO and THC emissions. In the developing process of above combustors, to simplify the combustor system, air blast type fuel nozzles with single fuel injection and dual swirlers were applied. Successively, in this report, the fuel nozzle is modified to dual fuel injection type with triple swirlers, aiming to control combustion performance under varying load conditions. Fuel is injected from inner and outer injection circuits, and the injection ratio between them is treated as one of the parameters.

The combination of swirl direction of the three swirlers is selected at first through ignition and blowout tests. Secondly, spray patterns of the selected fuel nozzle are observed with different fuel injection ratios. Thirdly, the nozzle is applied to a rectangular single-sector combustor, and tested under atmospheric pressure with inlet temperature of 500K. NOx, CO, CO2, THC and O2 compositions in the exhaust gas are measured, and correlation among measured emissions data and fuel injection ratio is estimated to examine the influence of the injection ratio on combustion characteristics of the Rich-Lean type aero engine combustor.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A003. doi:10.1115/GT2014-26003.

The ever increasing consumption of non-renewable fossil fuels for global economic development leads to serious energy crisis and environmental pollution. Consequently, new alternative fuels and high-efficiency combustion are required to aid the sustainable development of human society. The present paper took the RP-3 aviation kerosene and coal-to-liquid synthetic aviation fuel (manufactured through the Fischer Tropsch process., FT) for object, and experimentally investigated the influences of pressure, inlet temperature and equivalence ratio on the productions of NOx and CO in a jet stirred combustion reactor. The tests were performed under the pressures of 2bar and 3bar, and inlet air temperatures of 550K and 650K, respectively. The equivalence ratio ranged from 0.5 to 1.2. The mean residence time was approximately 8ms. Probe sampling followed by on-line emissions analyzer permitted to measure the concentration of the products. The experimental results show that these two fuels obey the same law with the variations of pressures, inlet temperatures and equivalence ratios. The NOx production increases with the pressure and inlet temperature increasing. The CO decreases with the pressure increasing, while slightly increases with the inlet temperature increasing. Numerical simulations were also performed to investigate the combustion products of these two fuels in the jet stirred combustion reactor. Two PSRs were introduced to simulate the jet flame region and post flame in the recirculation region, respectively. The combustion products of second PSR (PSR2) agreed well with the experimental results by regulating the volume ratio of first PSR (PSR1). Based on the reaction pathway analysis of NO production in present state, it is considered that for these two fuels the NOx production is led by the thermal NO above the equivalence ratio of 0.65, while by the N2O at lower equivalence ratios. With the application of the present alternative fuel and its reaction mechanism, the experimental results of aviation kerosene and Coal-to-Liquid synthetic aviation fuel can be predicted well within a certain state, which requires a further verification in a wider range. Furthermore, the numerical results show that the NO release is insensitive to the reaction components within present experimental states.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A004. doi:10.1115/GT2014-26018.

The dynamic transitions preceding combustion instability and lean blowout were investigated experimentally in a laboratory scale turbulent combustor by systematically varying the flow Reynolds number. We observe that the onset of combustion-driven oscillations is always presaged by intermittent bursts of high-amplitude periodic oscillations that appear in a near random fashion amidst regions of aperiodic, low-amplitude fluctuations. The onset of high-amplitude, combustion-driven oscillations in turbulent combustors thus corresponds to a transition in dynamics from chaos to limit cycle oscillations through a state characterized as intermittency in dynamical systems theory. These excursions to periodic oscillations become last longer in time as operating conditions approach instability and finally the system transitions completely into periodic oscillations. Such intermittent oscillations emerge through the establishment of homoclinic orbits in the phase space of the global system which is composed of hydrodynamic and acoustic subsystems that operate over different time scales. Such intermittent burst oscillations are also observed in the combustor on increasing the Reynolds number further past conditions of combustion instability towards the lean blowout limit. High-speed flame images reveal that the intermittent states observed prior to lean blowout correspond to aperiodic detachment of the flame from the bluff-body lip. These intermittent oscillations are thus of prognostic value and can be utilized to provide early warning signals to combustion instability as well as lean blowout.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A005. doi:10.1115/GT2014-26020.

The present study aims at arming an operator of fielded turbulent combustors with a repertoire of mathematical measures for real-time monitoring to forewarn the onset of impending combustion instability. In turbulent combustors, the route to high-amplitude, periodic, combustion-driven oscillations from conditions of low-amplitude, chaotic, combustion noise happens through an intermediate regime in flow conditions where the measured pressure fluctuations display bursts of intermittent, high-amplitude, periodic oscillations that appear in a near-random manner amidst chaotic fluctuations. This loss of chaos from combustion noise to combustion instability can be quantified to serve as a precursor to impending instability. The recurrence properties of intermittent burst oscillations can be quantified using dynamical systems theory by tracking the distribution of the aperiodic segments in the measured signals. Several statistical measures may be constructed through such recurrence quantification that provide robust early warning signals to an impending instability. Further, the transition to combustion driven oscillations leads to a collapse of the number of relevant time scales involved in the problem. This collapse in time scales can be quantified using generalized Hurst exponents which serve as an additional measure that captures the onset of an impending combustion instability well in advance. The various patent pending measures illustrated in this paper serve as precursors due to the inevitable presence of an intermittent regime of burst oscillations in turbulent combustors.

Topics: Combustion
Commentary by Dr. Valentin Fuster
2014;():V04BT04A006. doi:10.1115/GT2014-26023.

Siemens Industrial Turbomachinery AB (SIT) Finspong is increasingly asked by customers to consider if a wider range of gas compositions may be operated in its gas turbines. A large part of these fuel compositions contain high concentrations of highly reactive components like hydrogen and ethane. Such fuel compositions are characterized by higher flame speeds than the standard natural gas which introduce increased risk for flashback and premature combustor and/or fuel nozzle distress.

The SIT approach for fuel flexibility testing, Experimental Burner In Test engine (EBIT) uses a separate feed for testing a selected fuel composition in one burner in a standard gas turbine installation where the other burners use fuel from the standard fuel system. The separate feed is operated as a slave to engine governor heat demand, but can also be controlled independently.

This paper describes how EBIT was used to test the capability of the SIT 3rd generation DLE burner to satisfactorily operate using highly reactive fuel components such as ethane and hydrogen. Combustion monitoring techniques and measurements to check flame behavior and assess flashback potentials of the tested fuel compositions are described. Results from testing show that the SIT 3rd generation DLE burner allows extending the levels of reactive fuel components accepted for operation on the SGT-700 33MW engine and SGT-800 50MW engine.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A007. doi:10.1115/GT2014-26046.

Analysis of the Oil & Gas market segment showed that potential MS5002E customers could benefit from firing the gas turbine with distillate oil as a back-up fuel, mainly to provide power when the fuel gas is not available (e.g. when the plant itself is being commissioned). To address this customer need, the design of a dual fuel system for such mission should target simplicity, reliability and minimize the additional cost with respect to the single gas version. To achieve these targets, the development of the dual fuel system for the MS5002E leveraged the efforts made by GE for the design of a liquid fuel system for Frame 9F-1 series with no need of atomization air. Moreover, the emission capability during liquid fuel operation was enhanced allowing the mixing of water and fuel before injection in the combustion chamber and using of improved injection technology, thus improving the efficiency of water injection with a significant reduction in the required water flow rates; the importance of this achievement is related to both the increasingly stringent regulation on this subject and the often poor availability of water in the Oil & Gas market segment. The system is capable of continuous operation without water injection for applications where emissions are not critical; in these cases a small amount of demineralized water is employed occasionally for fuel line cooling and flushing, thus helping to guarantee constant performances of the injectors, and to maintain liquid fuel start-up capability over time.

This paper presents the expected performance, in terms of ignition capability, emissions, operability and expected hardware durability on LF/water-fuel emulsion operation, based on a single can rig test campaign. The new liquid fuel cartridges were tested from ignition to base load at ISO and extreme simulated ambient conditions, both with and without water injection, showing promising performance in terms of combustor operability and emissions. All the combustor components were instrumented with thermocouples to assess variations in the hardware thermal levels with respect to the single gas conditions, and identify possible issues related to the transient and steady-state liquid fuel operation. Further development and testing will be carried out in the next phases of the development, and the performance will be confirmed by a dedicated engine test at the first commercial opportunity.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A008. doi:10.1115/GT2014-26053.

Large Eddy Simulations (LES) of natural gas ignition and combustion in turbulent flows are performed using a novel combustion model based on a composite progress variable, a tabulated chemistry ansatz and the stochastic-fields turbulence-chemistry interaction model. It is a significant advantage of this approach that it can be applied to industrial configurations with multi-stream mixing at relatively low computational cost and modeling complexity. The computational cost is independent of the chemical mechanism or the type of fuel, but increases linearly with the number of streams. The model is validated successfully against the Cabra methane flame and Delft Jet in Hot Coflow (DJFC) flame. Both cases constitute fuel jets in a vitiated coflow. The DJFC flame coflow has a non-uniform mixture of air and hot gases. The model considers this non-uniformity by an additional mixture fraction dimension, emulating a ternary mixing case. The model not only predicts flame location, but also the temperature distribution quantitatively. The LES combustion model is further extended to consider four stream mixing. It has been successfully validated for ALSTOM’s reheat combustor at atmospheric conditions. Compared to the past steady-state RANS (Reynolds Averaged Navier-Stokes) simulations [1], the LES simulations provide an even better understanding of the turbulent flame characteristics, which helps in the burner optimization.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A009. doi:10.1115/GT2014-26058.

Jet fuel flowing through the fuel injector is atomized and then mixed with high temperature compressed air flowing through the swirler to create a combustible mixture inside a gas turbine combustor. Individual geometric and flow features are carefully tuned at a component level to deliver optimum combustion performance. In a critical interface such as the fuel injector and swirler, manufacturing tolerances not only have an impact on combustor performance and operability but also on durability, as the relative position of the fuel injector to the swirler significantly impacts the swirler temperature. This paper studies the influence of manufacturing tolerances on component assembly and the resulting impact on swirler temperature. The oxidation damage mechanism of the swirler is used as a measure to assess swirler durability. A Pareto chart of the effect of manufacturing tolerances on metal temperature is used to highlight the key influencing parameters. Probability distribution associated with manufacturing tolerances is gathered with Monte Carlo simulation to guide the design.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A010. doi:10.1115/GT2014-26076.

Temperature profile variations in gas turbine combustors are important from the considerations of thermal stresses and material fatigue. The specific profile being addressed in this study is the combustor exit gas temperature profile in radial direction at first stage nozzle entry (also called the combustor transition-piece (TP) exit profile). Normally, in multi-can combustor configurations, this profile is assumed to be constant along the circumferential direction or from one can to another. However, field test on one of the GE-MS5002D class machine revealed that the shape of the combustor TP exit temperature profile is varying across the different cans. It is important to assess the reason of this behavior in order to define thermal input for stage 1 nozzle thermal design and define an average temperature profile for turbine bucket verification.

For investigating the reasons of varying TP exit profiles across different cans, a reacting flow CFD study is performed for a combined multiple combustor-cans geometry. This is a challenging attempt considering that mesh for a single can liner is itself typically quite large (∼30 million) for capturing all flow features. The present multi-can study was made feasible with judicious simplification of combustor geometry, retaining only important flow features and using adequate mesh to capture system physics. Results indicate that the varying flame shape across different cans is indeed captured in the CFD. Hence, this effect could be something associated with the combustor design. Subsequent detailed post-processing of CFD results revealed the root cause to be associated with the presence of unsymmetrical arrangement of struts in the compressor discharge casing region. This effect is a slight flow-recirculation created much upstream due to the struts, which eventually results in asymmetric distribution of the flow across the combustor dilution holes. This leads to the flame shifting in different orientation for different cans with a systematic reference to the struts position. In conclusion, this paper describes the approach used for multi-can CFD analysis of the combustor, flow behavior in presence of unsymmetrical strut and its impact on the combustor exit temperature profile much downstream.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A011. doi:10.1115/GT2014-26077.

Large Eddy Simulations (LES) is increasingly becoming a feasible tool for industrial design purposes on account of ongoing advancements in computational power. It is a promising arena in the field of computational fluid dynamics where more details of flow-turbulence are explicitly captured and lesser are modeled as compared to the traditional Reynolds-average (RANS) approaches. For the gas turbine combustors particularly, it is a promising tool for better predictions of reactants mixing and hence the combustion, flame shape and temperature profiles. Also, as inherent unsteady nature of the flow is captured, it can predict combustion dynamics due to heat-release (and hence pressure) fluctuations. The main factor for performing a successful and reliable LES is to find an appropriate filter size for different regions of the CFD domain. This filter size is typically same as the CFD mesh size and turbulent scales larger than this are explicitly solved in LES. In industrial gas turbine combustors, due to complex geometry and numerous small cooling flow passages, unnecessary mesh refinement may make the mesh size prohibitive for a time-marching LES simulation. Hence, judicious selection of important flow features and geometry is important. Still not much experience is available on the quantification of LES meshing requirements for practical gas turbine combustors. In this study, two different LES meshing approaches, namely one based on Taylor length scales and other based on theoretical turbulence energy spectrum are compared for various medium scale gas turbine combustors. While the former approach requires a prior RANS simulation and provides a spatial distribution of the grid size, the latter just requires mean flow properties and global length scale at various inlets but produces only a global mesh value. It is found for all combustor designs under study that the two approaches agree well with each other for predicting mesh size requirements for LES where 85–90% of turbulent length scales are captured. This helps towards standardizing LES meshing procedure in industrial scenarios and helps a user to choose meshing option based on the level of details needed and time-resource constraints.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A012. doi:10.1115/GT2014-26096.

The continuous interest in reducing pollutions and developing both an efficient and clean combustion system require large attention in the design requirements, especially when related to industrial gas turbine application. Although in recent years the advancements in modelling have increased dramatically, combustion still needs a huge computational effort. The Flamelet-Generated Manifolds (FGM) method is considered a suitable solution with an accuracy that can be comparable with detailed chemistry simulations results. The full combustion system can be described by few controlling variables while the chemical details are stored in a database (manifold) as function of controlling variables. Transport equations are solved for the Navier-Stokes system and the controlling variables. The detailed chemistry code Chem1D is used to create the manifolds. Turbulence can be modeled using a PDF approach for the subgrid modeling of the chemistry terms. The OpenFOAM open source CFD package is used as CFD tool for the simulations. The aim of this work is to demonstrate the usage of FGM with OpenFOAM and figure out the status of the implementation. To achieve this goal, the work employs as test case a confined lean jet flame is used. For the case presented, an extensive experimental data set exist, including PIV and Raman data. Results are further compared with traditional methods, while FGM method might be easily extended to other scenarios.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A013. doi:10.1115/GT2014-26124.

All key combustor performance & operability characteristics like emissions, exit profile, durability, LBO etc. have a dependence on spray quality. Hence it is important to accurately predict spray characteristics for accurate combustor modeling. In this paper, a CFD based liquid jet in cross flow spray modeling approach adopted at GE Aviation is presented. Liquid jet in cross flow is a complex phenomenon that broadly involves jet trajectory evolution, surface breakup, column fracture and dispersion of secondary droplet particles. A two-phase steady state Volume of Fluid (VOF) approach is used to predict the liquid jet trajectory. A combination of output from VOF and empirical correlations (Sallam et. al; Oda et. al) is used to predict droplet distribution that includes diameter, velocity components and mass flow rate. Surface breakup is modeled by injecting droplets along the leeward surface of the liquid jet with spanwise perturbation to capture the transverse spread. Jet column breakup is modeled by injecting droplets including effects of unsteady fluctuations empirically to mimic the column fracture behavior. Discrete particles are then transported in a lagrangian frame coupled with secondary breakup of droplets. This approach has been validated on a benchmark quality dataset with an average SMD (Sauter Mean Diameter) error of ∼6 microns and is being used on Gas Turbine combustor fuel-air mixing devices employing liquid jet in cross flow atomizers.

Topics: Modeling , Cross-flow
Commentary by Dr. Valentin Fuster
2014;():V04BT04A014. doi:10.1115/GT2014-26136.

Conceptual design of a hydrogen fueled dry low NOx combustor for heavy duty gas turbine is presented in this paper, including a complete experimental validation, with focus on both NOx emissions and operability. Effort was first provided in the identification of viable conceptual solutions: a technology screening has been carried out, balancing both innovation content and proven experience of each concept. A look to alternative solutions coming from literature has been given too. Three burner concepts have been selected, designed and procured to be tested into a reduced scale rig, arranged to mimic main features of a small size gas turbine combustor, in terms of combustion air inlet temperature, hot gases residence times and amount of cooling: atmospheric pressure operation was considered a proper approximation to actual operating conditions for a conceptual design phase. The three solutions have been first characterized in terms of emissions against equivalence ratio, pilot percentage and burner pressure drop. At the same time, safe operation margins to both flashback and combustion instabilities onset have been identified for both pure hydrogen and pure natural gas feeding options. Results, while recommending different development paths for each of the investigated concepts, clearly indicate the most mature among them, allowing authors to address specific operability detailed investigations on it: flashback and flame holding resistance tests were thus performed, demonstrating that such a solution is mature for a preliminary full scale arrangement design and experimental characterization.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A015. doi:10.1115/GT2014-26137.

In this work, a series of tests have been conducted to study the aerodynamics of two typical gas turbine combustion swirlers, using Laser Doppler Velocimetry (LDV). The first swirler used is a coaxial counter-rotating, radial-radial swirler, designated as the R-R swirler. The primary swirler accounts for 40% of the total effective area of the R-R swirler. The second swirler employed consists of an inner dual-axial swirler assembly with only around 20% of total swirler effective area and an outer coaxial radial swirler, and is designated as the A-R swirler. The aim of this work is to compare the aerodynamics performance of these two different types of swirlers under similar boundary conditions, such as confinement effect, downstream exhaust nozzle, and chamber length.

The R-R swirler has a very high radial dispersing rate, and for the unconfined case, the swirling jet remained attached to the dome plate resulting in a very weak, unbound CTRZ. The swirling jet for the A-R swirler exited at a relatively low expansion angle. For confined cases, the swirling jet attached to the walls of the confinement at a certain distance downstream of the exit plane for both swirlers.

When the exhaust nozzle was applied as a downstream boundary condition, a spinning vortex core with high turbulent intensity was detected along the swirler axis for the A-R swirler, whereas the only difference for the R-R swirler was a slight reduction in the expansion angle of the swirling jet. The proximity of the exhaust nozzle to the swirlers was changed by changing the chamber length. For the R-R swirler, the effect of chamber length was only detected in the near-field region. For the A-R swirler, a reduction in chamber length caused the spinning vortex core to become more intense. Additionally, an annular CTRZ was observed for the shorter chamber lengths.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A016. doi:10.1115/GT2014-26139.

The NOx emissions of a premixed jet in hot cross flow configuration have been studied experimentally and with simple Cantera models. Staged combustion is realized by a first reaction zone providing a flow of fully combusted products and a secondary combustion stage downstream which consists of a premixed jet injected into the hot cross flow. The focus of the study lies on the overall NOx emission reduction potential of this generic configuration with respect to single stage premixed combustion employed in the high load regime of most large gas turbines for power generation. Cantera benchmark calculations were made to address that question. The second stage reaction was modeled using two different configurations: The worst case scenario being the combustion of the mixture in the jet without any interaction with the hot cross flow, and the most favorable case being the jet and the cross flows perfectly mixed and then burned. All calculations were performed under atmospheric and high pressure conditions to illustrate the influence of pressure on the potential of staging on the reduction of NOx emissions. In parallel to the modeling the NOx formation was studied in an atmospheric test rig. For that purpose emission and temperature data were collected with single orifice probe traverses four jet diameters downstream. The thermocouple probe was calibrated against CO-equilibrium concentration data from thermodynamic modeling. Furthermore, a planar laser diagnostic technique using Mie scattering was applied for determining the mixing of jet and cross flow. The mixture field was statistically analyzed to detect differences in spatial and temporal mixing. Finally all data were used to study NOx formation under atmospheric pressure and to transfer the results to gas turbine combustor conditions.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A017. doi:10.1115/GT2014-26161.

The response of soot temperature to unsteady inlet airflow is characterized using pyrometry. The unsteady inlet airflow is achieved by either modulating inlet air or naturally-occurring unstable flame, running on a jet fuel at fuel-rich conditions. The inlet air is modulated by a siren device running at frequencies between 150 and 250 Hz and up to 60% of modulation level (u’/um) is achieved. Also, the combustor can be run naturally unstable at the same inlet operating condition by changing the combustor length. For the pyrometry, the emission from whole flame at 660 nm, 730 nm and 800 nm is recorded and the three-color pyrometry is used to measure soot temperature. The effect of non-isothermal distribution of soot in flame on the measured temperature is also considered. The level of overall temperature fluctuation under inlet flow modulation (Trms/Tmean) is about an order of magnitude lower than that of flame emission fluctuation (Irms/Imean). Under naturally occurring instability the measured soot temperature is in phase with the pressure measured in the combustor, indicating that the measured soot temperature can be used as a quantity related to combustion dynamics for fuel-rich sooty flames.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A018. doi:10.1115/GT2014-26162.

The breakup, penetration, droplet size and size distribution of a Jet A-1 fuel in air crossflow has been investigated with focus given to the impact of surrounding air pressure. Data has been collected by Particle Doppler Phased Analyzer (PDPA), Mie-Scattering with high speed photography augmented by laser sheet, and Mie-Scattering with ICCD Camera augmented by nano-pulse lamp. Nozzle orifice diameter, do, was 0.508 mm and nozzle orifice length to diameter ratio, lo/do, was 5.5. Air crossflow velocities ranged from 29.57 to 137.15 m/s, air pressures from 2.07 to 9.65 bar and temperature held constant at 294.26 K. Fuel flow was governed to provide a range of fuel/air momentum flux ratio q from 5 to 25 and Weber number from 250 to 1000. From the results, adjusted correlation of the mean drop size has been suggested using drop size data measured by PDPA as follows; Display Formula

(1)
D0D32=0.267Wea0.44q0.08ρlρa0.30μlμa-0.16
This correlation agrees well and shows roles of aerodynamic Weber number, Wea, momentum flux ratio, q, and density ratio, ρla. Change of the breakup regime map with respect with surrounding air pressure has been observed and revealed that the boundary between each breakup modes can be predicted by a transformed correlation induced from above correlation. In addition, the spray trajectory for the maximum Mie-scattering intensity at each axial location downstream of injector was extracted from averaged Mie-scattering images. From these results correlations with the relevant parameters including q, x/do, density ratio, viscosity ratio, and Weber number are made over a range of conditions. According to spray trajectory at the maximum Mie-scattering intensity, the effect of surrounding air pressure becomes more important in the farfield. On the other hand, effect of aerodynamic Weber number is more important in the nearfield.

Topics: Pressure , Jets
Commentary by Dr. Valentin Fuster
2014;():V04BT04A019. doi:10.1115/GT2014-26164.

Periodic behavior in the reaction zone of a multiple nozzle combustor undergoing self-sustaining combustion oscillations is examined. This combustor has three stages: a high-swirl pilot stage, a lower-swirl intermediate stage, and a low-swirl outer stage. The high-power conditions reviewed in this paper have fuel supplied to all three stages. Four conditions are examined in which thermoacoustic coupling is observed at a well-defined frequency. The highest overall equivalence ratio case displays a significantly higher oscillation amplitude than the other cases. Changes to the fuel distribution (with constant overall equivalence ratio) results in smaller effects on the oscillation strength. Phase averaged images of the OH* chemiluminescence emission show dramatic changes in the OH* distribution and intensity over an average period of the oscillation. This variation in chemiluminescence is dominated by recurrent quenching and reignition downstream of the intermediate and outer fuel stages. The pilot stage reaction zone also displays periodic variation in intensity which is out-of-phase and precedes the intermediate and outer fuel stages. Proper orthogonal decomposition is used to extract the most energetic spatial components which form the periodic behavior in the OH* distribution. The POD modes allow direct field visualization of fluctuation location and magnitude. For all four cases, the phase and location of the OH* emission variations are generally similar with small differences in the location and rates of periodic changes in the reaction oscillation.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A020. doi:10.1115/GT2014-26174.

Flame interaction between neighboring burners in a gas turbine combustor is investigated numerically for pursuit of its effect on NOx emission from the burners. In a model chamber or liner, EV burners with double cone are installed. Two burners with the same rotating direction of air stream are adopted and the distance between them is variable from 74.2 mm to 222.6 mm by the step size of 37.1 mm. Gaseous methane and air are adopted as fuel and oxidizer, respectively. From steady-state numerical analyses, flow, temperature, and NO concentration fields are calculated in all computational cases to find their correlation with NOx formation. NOx emission is evaluated at the exit of the model chamber with two burners as a function of burner distance and compared with that from a single burner. In all cases of two-burner calculations, NOx emission is higher than that of a single burner, which results from flow interactions between neighboring burners as well as between a burner and a liner wall. NOx emission is affected significantly by flow and flame interactions between them and strongly depends on burner distance. Burner interaction is divided into two regimes of a burner-burner interaction and a burner-wall interaction depending on the distance. In the former regime, NOx emission is reduced as flame interaction between burners is enhanced and in the latter regime, it is also reduced as interaction between the burner and the liner wall is enhanced.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A021. doi:10.1115/GT2014-26200.

A thermoacoustic instability observed within a lean swirl-stabilized burner is analyzed by means of numerical calculations. As extension to previously reported numerical data for the considered test case, the numerical domain is extended to include the fuel plenum and the acoustic outlet impedance is modeled by time domain impedance boundary conditions.

The predicted thermoacoustic frequency is in excellent agreement with experimental findings. Furthermore, the thermoacoustic feedback loop is discussed by means of phase resolved data which show that a variation of the equivalence ratio within the swirler is the driving mechanism of the instability. Finally, the Rayleigh criterion is shown to be satisfied within the combustion chamber. Hence, a closed thermoacoustic feedback loop is observed.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A022. doi:10.1115/GT2014-26210.

CFD predictions of flame position, stability and emissions are essential in order to obtain optimized combustor designs in a cost efficient way. However, the numerical modeling of practical combustion systems is a very challenging task. As a matter of fact, the use of detailed reaction mechanisms is necessary for such reliable predictions. Unfortunately, the modeling of the full detail of practical combustion equipment is currently prohibited by the limitations in computing power, given the large number of species and reactions involved. The Flamelet-Generated Manifold (FGM) method reduces these computational costs by several orders of magnitude without loosing too much accuracy. Hereby FGM enables the application of reliable chemistry mechanisms in CFD simulations of combustion processes. In the present paper a computational analysis of partially premixed non-adiabatic flames is presented. In this scope, chemistry is reduced by the use of the FGM method. In the FGM technique the progress of the flame is generally described by a few control variables. For each control variable a transport equation is solved during run-time. The flamelet system is computed in a pre-processing stage, and a manifold with all the information about combustion is stored in a tabulated form.

This research applies the FGM chemistry reduction method to describe partially premixed flames in combination with heat loss, which is a relevant condition for stationary gas turbine combustors. In order to take this into account, in the present implementation the reaction evolution is described by the reaction progress variable, the heat loss is described by the enthalpy and the local equivalence ratio effect on the reaction is represented by the mixture fraction. A series of test simulations is performed for a two dimensional geometry, characterized by a distinctive stratified methane/air inlet, and compared with detailed chemistry simulations. The results indicate that detailed simulations are reproduced in an excellent way with FGM.

Topics: Modeling , Flames , Manifolds
Commentary by Dr. Valentin Fuster
2014;():V04BT04A023. doi:10.1115/GT2014-26222.

This paper reports proper orthogonal decomposition (POD) analyses for the velocity fields measured in a test burner. The Cambridge/Sandia Stratified Swirl Burner has been used in various studies as a benchmark for high resolution scalar and velocity measurements, for comparison with numerical model prediction. Flow field data was collected for a series of bluff-body stabilized premixed and stratified methane/air flames at turbulent, globally lean conditions (ϕ = 0.75) using high speed stereoscopic particle image velocimetry (HS-SPIV). In this paper, a modal analysis was performed to identify the large scale flow structures and their impact on the flame dynamics. The high speed PIV system was operated at 3 kHz to acquire a series of 4096 sequential flow field images both for reactive and non-reactive cases, sufficient to follow the large-scale spatial and temporal evolution of flame and flow dynamics. The POD analysis allows identification of vortical structures, created by the bluff body, and in the shear layers surrounding the stabilization point. In addition, the analysis reveals that dominant structures are a strong function of the mixture stratification in the flow field. The dominant energetic modes of reactive and non-reactive flows are very different, as the expansion of gases and the high temperatures alter the unstable modes and their survival in the flow.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A024. doi:10.1115/GT2014-26236.

Multiple swirlers arranged in an annular fashion are used in modern day gas turbine engines. A section of this annulus can be considered as a straight line or what is referred to in the paper as a linear arrangement of swirlers. Three such linear arrangements are computationally analyzed and results are presented through this study. Study of linear arrangements is crucial and novel to the swirler aerodynamics research as it lays a foundation in understanding the flow physics when swirlers are arranged at a fixed distance next to each other. Swirling flows are complicated and when slight modifications are introduced in physical arrangements the flow is impacted drastically. In the present study observations have been presented on effect of changing the offset of exit plane of swirler from the base wall of confinement when there is a single swirler or a linear arrangement of swirlers.

Computational simulations of flow through single and multi-swirler array have been carried out to understand the effect of the distance of exit plane of swirler from the base wall of confinement on the swirler aerodynamics. The swirlers used in this study are radial-radial swirlers with counter rotating vanes. The computational domain extended from the inlet manifold to 12 D downstream from the swirler where D is the diameter of swirler exit. Realizable k-ε turbulence model is used and the computational grid is about 4 million points for a single swirler arrangement, about 12 million points for a three swirler array and up to 22 million for the five swirler arrangement. The computational model is validated by comparing the results with velocity measurements carried out at three different planes downstream of the swirler exit using LDV technique.

First, single swirler with the exit plane of swirler with an offset of 0.04 D and 0.02D with the base wall of confinement and that with no offset (swirler exit in-line with base wall of confinement) are analyzed. It is observed that flow development in region close to the swirler exit is highly sensitive to the offset condition. In case of 0.04D and 0.02D offset a strong jet is formed as soon as the air exits the swirler. The flow tends to progress vertically forming recirculation zones in the vicinity of corners of the horizontal and vertical walls. When there is no offset, the flow exiting the swirler tends to align with the base wall and then progresses vertically. Thus for no offset case a jet formation is not observed.

Next, multi-swirler arrangements with 0.04D, 0.02D offset as well as no offset configurations are simulated. All the swirlers tend to show similar pattern as single swirler arrangements with a slight difference in intensity of the flow field. For swirlers with offset of 0.04D and 0.02D there is formation of a strong jet exiting the swirler and recirculation zones are formed in corners of the base and vertical walls of the confinement as was observed for the single swirler arrangement. Recirculation zones are also formed in areas between each swirler assembly in the multi swirler arrangement. For the no offset condition it is again observed that flow aligns with the horizontal base wall for each of the swirler assembly. The axial velocity of the flow in this arrangement tends to be lower than the offset case in regions between each swirler.

An interesting phenomenon of multi swirler arrangement is an asymmetrical flow pattern that is observed at each swirler. While each swirler geometry is identical, the flow pattern as well as the strength of recirculation zone developed from each individual swirler differs significantly. Results show that alternate swirlers tend to exhibit similar flow characteristics.

Topics: Aerodynamics
Commentary by Dr. Valentin Fuster
2014;():V04BT04A025. doi:10.1115/GT2014-26259.

A simulation study of high-pressure lifted flames in a constant-volume chamber has been conducted using detailed reaction mechanisms in CFD to investigate ignition times, flame lift-off lengths, soot production, and fuel effects. The fuels considered include n-heptane, a two-component surrogate fuel (SR), conventional U.S. No. 2 diesel (D2), and world-averaged jet fuel (Jet-A). Conditions for the flames are those of the experiments performed at Sandia National Laboratories; the n-heptane flame is labeled Spray H [Idicheria and Pickett, SAE 2005-01-3834], and the conditions for all other fuels studied are labeled Spray A [Kook and Pickett, SAE 2012-01-0678]. 3D CFD simulations have been performed using the FORTÉ CFD package. Complex fuels D2 and Jet-A have been modeled using multi-component surrogates. Detailed reaction mechanisms for fuel combustion and emissions formation have been used in the simulations. The size of the fuel mechanisms varied from 326 species to 1000 species for the different fuels. For soot predictions, two different models were used in the simulations: a detailed soot-surface mechanism and a seven-step phenomenological soot model. Both soot models were coupled with the fuel mechanism precursor predictions that included aromatics from benzene to pyrene. While using the detailed soot-surface mechanism, particulate (PM) size and number density were determined using the Method of Moments, which is implemented in the CFD software to calculate particle size distribution characteristics.

Results show excellent prediction of flame location and ignition for all fuels. Location and magnitude of soot fractions in the various flames show good agreement with the published data. Both the phenomenological soot model and the detailed soot-surface mechanism estimated comparable soot fractions in all flames. In addition, PM size information was predicted using the detailed soot-surface mechanism. Impacts of fuel, temperature, pressure, and oxygen concentrations on combustion and soot fractions have been captured by the simulations.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A026. doi:10.1115/GT2014-26260.

The local equivalence ratio distribution in a flame affects its shape and response under velocity perturbations. The forced heat release response of stratified lean-premixed flames to acoustic velocity fluctuations are investigated via chemiluminescence measurements and spatial Fourier transfer analysis. A laboratory scale burner and its boundary conditions were designed to generate high-amplitude acoustic velocity fluctuations in flames. These flames are subject to inlet radial equivalence ratio distributions created via a split annular fuel delivery system outfitted with a swirling stabilizer. Simultaneous measurements on the oscillations of inlet velocity and heat release rate were carried out via a two-microphone technique, and OH* chemiluminescence. The measurements show that, for a given mean total power and equivalence ratio (ϕg = 0.60), the flame responses vary significantly depending on forcing frequency, equivalence ratio split and velocity fluctuation amplitude, showing significant non-linearities with respect to forcing amplitude and stratification ratio. Furthermore, the spatial Fourier transfer analysis shows the underlying changes in the rate of heat release, including the direction and speed of the perturbation within the flame.

Topics: Heat , Turbulence , Flames
Commentary by Dr. Valentin Fuster
2014;():V04BT04A027. doi:10.1115/GT2014-26271.

Lean blow-out (LBO) is critical to operational performance of combustion systems in propulsion and power generation. Current predictive tools for LBO are based on decades-old empirical correlations that have limited applicability for modern combustor designs. Based on Lefebvre’s model for LBO and flame volume concept, an FV (Flame Volume) model was proposed by Authors in early study. The FV model adds two key parameters of α and β that represent the fraction of dome air and dimensionless flame volume defined as the ratio of flame volume and combustor volume. Due to the flame volume is obtained from the experimental image, FV model could only be used in LBO analysis instead of predictions. In the present study, a hybrid FV model is proposed that combines the FV model with numerical simulation for LBO predictions. In the hybrid FV model, α and β are estimated from the numerical simulation result of the non-reacting flow in the combustor. Comparing with the experimental data for 11 combustors, the LBO fuel/air ratio obtained by hybrid FV model shows better agreement than that obtained by Lefebvre’s model. The maximum prediction uncertainties of hybrid FV model and Lefebvre’s model are about ±16% and ±48%, respectively. Moreover, the time cost of the LBO prediction using hybrid FV model for each case is about 6 hours with the computer equipment of CPU×12 and 24G memory, showing that the hybrid FV model is reliable and efficient to be used for the performance evaluation of the combustor, even the so called “paper combustors” in the primary design stage.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A028. doi:10.1115/GT2014-26272.

This paper describes an experimental investigation of a transversely forced, swirl stabilized combustor. Its objective is to compare the unsteady flow structures in single and triple nozzle combustors and determine how well a single nozzle configuration emulates the characteristics of a multi-nozzle one. The experiment consists of a series of velocity field measurements captured on planes normal to the jet axis. As expected, there are differences between the single and triple-nozzle flow fields, but the differences are not large in the regions upstream of the jet merging zone. Direct comparisons of the time averaged flow fields reveal a higher degree of non-axisymmetry for the flowfields of nozzles in a multi-nozzle configuration. Azimuthal decompositions of the velocity fields show that the transverse acoustic forcing has an important influence on the dynamics, but that the single and multi-nozzle configurations have similar forced response dynamics near the dump plane. Specifically, the axial dependence of the amplitude in the highest energy axisymmetric and helical flow structures is quite similar in the two configurations. This result suggests that the hydrodynamic influence of one swirling jet on the other is minimal and, as such, that jet-jet interactions in this configuration do not have a significant influence on the unsteady flow structures.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A029. doi:10.1115/GT2014-26293.

The effect of hydrogen enrichment to natural gas flames was experimentally investigated at atmospheric pressure conditions using flame chemiluminescence imaging, planar laser-induced fluorescence of hydroxyl radicals (OH PLIF) and dynamic pressure monitoring. The experiments were performed using a 3rd generation dry low emission (DLE) burner used in both SGT-700 and SGT-800 industrial gas turbines from Siemens. The burner was mounted in an atmospheric combustion test rig at Siemens with optical access in the flame region. Four different hydrogen enriched natural gas flames were investigated; 0 vol.%, 30 vol.%, 60 vol.% and 80 vol.% of hydrogen. The results from flame chemiluminescence imaging and OH PLIF show that the size and shape of the flame was clearly affected by hydrogen addition. The flame becomes shorter and narrower when the amount of hydrogen is increased. For the 60 vol.% and 80 vol.% hydrogen flames the flame has moved upstream and the central recirculation zone that anchors the flame has moved upstream the burner exit. Furthermore, the position of the flame front fluctuated more for the full premixed flame with only natural gas as fuel than for the hydrogen enriched flames. Measurements of pressure drop over the burner show an increase with increased hydrogen in the natural gas despite same air flow thus confirming the observation that the flame front moves upstream towards the burner exit and thereby increasing the blockage of the exit. Dynamic pressure measurements in the combustion chamber wall confirms that small amounts of hydrogen in natural gas changes the amplitude of the dynamic pressure fluctuations and initially dampens the axial mode but at higher levels of hydrogen an enhancement of a transversal mode in the combustion chamber at higher frequencies could occur.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A030. doi:10.1115/GT2014-26361.

This paper addresses the numerical prediction of NOx emissions from a micro turbine generator (MTG) using a chemical reactor network (CRN). In particular, the effect of fuel composition on NOx formation is analyzed. The numerical predictions are compared with available experimental results. The results obtained with the reactor network analysis (RNA) for the mixtures with heavier alkanes are in good agreement with the experimental results; however the methodology indicates that the prediction is sensitive to the selection of the reaction mechanism. Also, it was possible to predict measured trends for the effect of the dilution with CO2, but the slope of the trend differs from the experiments. The analysis of the results indicates that most of the NOx measured at the turbine exit for the conditions of this study is formed through the N2O route. This is the dominant pathway for this system, regardless of the fuel composition or the reaction mechanism used.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A031. doi:10.1115/GT2014-26376.

In this contribution, comprehensive optical and laser based measurements in a generic multi-jet combustor at gas turbine relevant conditions are presented. The flame position and shape, flow field, temperatures and species concentrations of turbulent premixed natural gas and hydrogen flames were investigated in a high-pressure test rig with optical access.

The needs of modern highly efficient gas turbine combustion systems, i.e., fuel flexibility, load flexibility with increased part load capability, and high turbine inlet temperatures, have to be addressed by novel or improved burner concepts. One promising design is the enhanced FLOX® burner, which can achieve low pollutant emissions in a very wide range of operating conditions. In principle, this kind of gas turbine combustor consists of several nozzles without swirl, which discharge axial high momentum jets through orifices arranged on a circle. The geometry provides a pronounced inner recirculation zone in the combustion chamber. Flame stabilization takes place in a shear layer around the jet flow, where fresh gas is mixed with hot exhaust gas. Flashback resistance is obtained through the absence of low velocity zones, which favors this concept for multi-fuel applications, e.g. fuels with medium to high hydrogen content.

The understanding of flame stabilization mechanisms of jet flames for different fuels is the key to identify and control the main parameters in the design process of combustors based on an enhanced FLOX® burner concept. Both experimental analysis and numerical simulations can contribute and complement each other in this task. They need a detailed and relevant data base, with well-known boundary conditions. For this purpose, a high-pressure burner assembly was designed with a generic 3-nozzle combustor in a rectangular combustion chamber with optical access. The nozzles are linearly arranged in z direction to allow for jet-jet interaction of the middle jet. This line is off-centered in y direction to develop a distinct recirculation zone. This arrangement approximates a sector of a full FLOX® gas turbine burner. The experiments were conducted at a pressure of 8 bar with preheated and premixed natural gas/air and hydrogen/air flows and jet velocities of 120 m/s.

For the visualization of the flame, OH* chemiluminescence imaging was performed. 1D laser Raman scattering was applied and evaluated on an average and single shot basis in order to simultaneously and quantitatively determine the major species concentrations, the mixture fraction and the temperature. Flow velocities were measured using particle image velocimetry at different section planes through the combustion chamber.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A032. doi:10.1115/GT2014-26420.

Fuel composition has a strong influence on the turbulent flame speed, even at very high turbulence intensities. An important implication of this result is that the turbulent flame speed cannot be extrapolated from one fuel to the next using only the laminar flame speed and turbulence intensity as scaling variables. This paper presents curvature and tangential strain rate statistics of premixed turbulent flames for high hydrogen content fuels. Global (unconditioned) stretch statistics are presented as well as measurements conditioned on the leading points of the flame front. These measurements are motivated by previous experimental and theoretical work that suggests the turbulent flame speed is controlled by the flame front characteristics at these points. The data were acquired with high speed particle image velocimetry (PIV) in a low swirl burner (LSB). We attained measurements for several H2:CO mixtures over a range of mean flow velocities and turbulence intensities. The results show that fuel composition has a systematic, yet weak effect on curvatures and tangential strain rates at the leading points. Instead, stretch statistics at the leading points are more strongly influenced by mean flow velocity and turbulence level. It has been argued that the increased turbulent flame speeds seen with increasing hydrogen content are the result of increasing flame stretch rates, and therefore SL,max values, at the flame leading points. However, the differences observed with changing fuel compositions are not significant enough to support this hypothesis. Additional analysis is needed to understand the physical mechanisms through which the turbulent flame speed is altered by fuel composition effects.

Topics: Fuels , Flames , Hydrogen
Commentary by Dr. Valentin Fuster
2014;():V04BT04A033. doi:10.1115/GT2014-26435.

Hydrodynamic instabilities of the flow field in lean premixed gas turbine combustors can generate velocity perturbations that wrinkle and distort the flame sheet over length scales that are smaller than the flame length. The resultant heat release oscillations can then potentially result in combustion instability. Thus, it is essential to understand the hydrodynamic instability characteristics of the combustor flow field in order to understand its overall influence on combustion instability characteristics. To this end, this paper elucidates the role of fluctuating vorticity production from a linear hydrodynamic stability analysis as the key mechanism promoting absolute/convective instability transitions in shear layers occurring in the flow behind a backward facing step. These results are obtained within the framework of an inviscid, incompressible, local temporal and spatio-temporal stability analysis. Vorticity fluctuations in this limit result from interaction between two competing mechanisms — (1) production from interaction between velocity perturbations and the base flow vorticity gradient and (2) baroclinic torque in the presence of base flow density gradients. This interaction has a significant effect on hydrodynamic instability characteristics when the base flow density and velocity gradients are co-located. Regions in the space of parameters characterizing the base flow velocity profile, i.e. shear layer thickness and ratio of forward to reverse flow velocity, corresponding to convective and absolute instability are identified. The implications of the present results on prior observations of flow instability in other flows such as heated jets and bluff-body stabilized flames is discussed.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A034. doi:10.1115/GT2014-26615.

The present study focuses on identifying and resolving large-scale energy containing structures and turbulent eddies in a typical gas turbine combustor single nozzle rig, using particle image velocimetry in cold flow. A generic fuel-air nozzle through a swirler is integrated with a sudden expansion square duct with optical access to perform laser diagnostics. Experiments are conducted to analyze the swirl flow field under starting and operating flow conditions. Three-component velocities are obtained in cross-sectional planes of Z/D = 0, 1.25, and 2.5 (normalized by the nozzle diameter), and two-component velocities are obtained in the mid-plane along the longitudinal (Z-) axis from Z/D = 0 to 2.5D. Velocity splitting is performed using spatial Gaussian smoothing with a kernel with filter width equal to integral scale is performed over the velocity fields to resolve the field of large-scale energy containing eddies. Proper orthogonal decomposition is performed over the large-scale velocity field, and the modes obtained indicate the existence of the precessing vortex core (PVC), formation of small scales Kelvin-Helmholtz (K-H) vortices for Z/D < 1.25D, and large-scale growing K-H structures in 1.25D < Z/D < 2.5D. Turbulent kinetic energy (TKE) is obtained from the turbulent velocity fluctuations below the integral length scale and is observed to be higher at the interface of the corner recirculation zone (CRZ) and central toroidal recirculation zone (CTRZ). Resolving the swirl velocity field obtained in the above manner into large-scale structures formed by the PVC, CTRZ, K-H vortices, CRZ, and small-scale turbulence field, indicates the clear distinction in rapid mixing zones and unsteady convective zones. The length-scales and zones of these structures within the swirl combustor are identified.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A035. doi:10.1115/GT2014-26652.

This paper presents numerical simulation of a combustor diffusion system, mainly focusing on the effects of the cowling geometry, the area ratio of the pre-diffuser and the axial length of the dump gap. The diffusion system of a full-scale single annular combustor is analyzed, using commercial software Fluent. Six different geometries are designed and numerically analyzed. Case 1 is a baseline, for which low emission technology burning in a lean mode is adopted. Cases 2, 3, and 4 are simulated to study the influence of the cowling geometry, especially the area of cowl capture plane. The study of cowling geometry shows that the best layout is case 4, owing to its least spillage of the air from the dome region into the inner and outer annuli of combustor. The normalized total pressure loss over the combustor is 3.49%. The total pressure distribution at the inlet of the main stage is more uniform than the baseline case 1. Cases 4 and 5 are also analyzed to investigate the influences of area ratio of the pre-diffuser, which varies from 1.4 in case 4 to 1.5 in case 5. The normalized combustor total pressure loss decreased to 3.30%, whose total pressure loss of pre-diffuser decreased while the static pressure recovery coefficient of the pre-diffuser increased from 0.43 to 0.52. Cases 4 and 6 are examined for the axial length of the diffuser dump gap, which is increased by 20% from case 4 to case 6. The combustor total pressure loss increases slightly, with little impact on the pressure loss of the pre-diffuser and dump gap region.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A036. doi:10.1115/GT2014-26667.

The evolution of a spark kernel ejected by a sunken fire igniter into a turbulent, fuel-air stratified cross-flow was studied both experimentally and using a model in a configuration that is similar to the conditions found in turbine engine combustors. This study allows for variations in the transit time of the kernel across a uniform non-flammable region, before entering a second stream containing a flammable fuel-air mixture. High speed schlieren and emission imaging systems are used to visualize the evolution of the kernel and determine the probability of ignition based on measurements over many spark events. Experiments are performed for a range of mean velocities, transit times, inlet (preheat) temperatures, flammable zone equivalence ratios, and non-flammable zone equivalence ratios. In addition to the typical dependence of ignition on the equivalence ratio of the flammable mixture, the results indicate the strong influence of the kernel transit time and the inlet flow temperature on the probability of ignition. The entrainment between the kernel and the surrounding flow appears to be primarily controlled by the kernel ejection-induced flowfield. Reduced-order modeling suggests that the lowering of the kernel temperature associated with entrainment of the non-flammable mixture significantly reduces the ignition probability, and leads to the conclusion that the presence of fuel close to the igniter is necessary to ensure reliable ignition under adverse conditions.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A037. doi:10.1115/GT2014-26711.

Staged combustion is a promising technology for gas turbines to achieve load flexibility and low NOx emission levels at the same time. Therefore, a large scale atmospheric test rig has been set up at the Institute of Thermodynamics, Technical University of Munich to study NOx emission characteristics of a reacting jet in hot cross flow. The premixed primary combustion stage is operated at ϕ = 0.5 and provides the hot cross flow. In the second stage a premixed jet at ϕ = 0.77 is injected perpendicular to the first stage. In both stages natural gas is used as fuel and air as oxidant. This paper presents a reactor model approach for the computation of the resulting NOx concentrations. The mixing and ignition process along the jet streamline of maximum NOx formation is simulated using a perfectly stirred reactor with Cantera 1.8. The reactor model is validated for the ambient pressure case using experimental data. Afterwards, a high pressure simulation is performed in order to investigate the NOx emission characteristics under gas turbine conditions.

The NOx formation is divided into flame NOx and post flame NOx. The reactor model reveals that the formation of post flame NOx in the second combustion stage can be efficiently suppressed due to fast mixing with cross flow material and the corresponding temperature reduction. Compared to single stage combustion with the same power output, no NOx reduction was observed in the experiment. However, the results from the reactor model suggest a NOx reduction potential at gas turbine conditions caused by the increased influence of post flame NOx production at high pressure.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A038. doi:10.1115/GT2014-26736.

This paper reports about a high pressure combustion test rig which was designed and erected inside the laboratory of the Institute of Steam and Gas Turbine (IDG, RWTH Aachen University, Univ.-Prof. Dr.-Ing. Dieter Bohn) in the time period from 1992. The first cold start-up was done to test and adjust the complex infrastructure including monitoring and control. The first hot start-up followed in May 2009 with a can-type combustor for a 7 MWel gas turbine. Thereafter, different gas turbine combustors have been tested and optimized with regard to NOx emission and combustor stability (see Tanaka et al. [1]).

This test rig is designed to conduct combustion tests with unscaled gas turbine combustors with a thermal power of up to 10 MW and exhaust gas temperatures of up to 1350°C. The test rig is capable of achieving air inlet conditions of up to 24 bar, 550°C and 12 kg/s.

After a successful phase of operation, the test rig has been continuously modified and upgraded at the Institute for Power Plant Technology, Steam and Gas Turbines (IKDG, RWTH Aachen University, Univ.-Prof. Dr.-Ing. habil. Manfred Wirsum).

This paper introduces the current test rig. First, the test rig is classified in relation to similar test rigs by relevant literature. Thereafter, the test rig design and operation mode is presented in detail including a quality evaluation of the combustor inlet conditions. Furthermore, a steady-state simulation is set up. Based on its results, the theoretical operating ranges and limitations are identified and discussed.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A039. doi:10.1115/GT2014-26742.

Hydrogen-based fuels have become a primary interest in the gas turbine market. To better predict the reactivity of mixtures containing different levels of hydrogen, laminar and turbulent flame speed experiments have been conducted. The laminar flame speed measurements were performed for various methane and natural gas surrogate blends with significant amounts of hydrogen at elevated pressures (up to 5 atm) and temperatures (up to 450 K) using a heated, high-pressure, cylindrical, constant-volume vessel. The hydrogen content ranged from 50% to 90% by volume. All measurements were compared to a chemical kinetic model, and good agreement within experimental measurement uncertainty was observed over most conditions. Turbulent combustion experiments were also performed for pure H2 and 50:50 H2:CH4 mixtures using a fan-stirred flame speed vessel. All tests were made with a fixed integral length scale of 27 mm and with a turbulent intensity level of 1.5 m/s at 1 atm initial pressure. Most of the turbulent flame speed results were in either the corrugated or thin reaction zones when plotted on a Borghi diagram, with Damköhler numbers up to 100 and turbulent Reynolds numbers between about 100 and 450. Flame speeds for a 50:50 blend of H2:CH4 for both laminar and turbulent cases were about a factor of 1.8 higher than for pure methane.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A040. doi:10.1115/GT2014-26788.

The internal flow within a pressure swirl atomizer was numerically simulated and evaluated in the present investigation. To validate the numerical method, a large-scale atomizer with an orifice diameter 21mm has been simulated and compared with former experimental results in the literature. Then a production-scale atomizer with an orifice diameter 1mm was simulated and compared to the results of large-scale atomizer. The internal flow characteristics of the swirl chamber were evaluated mainly in terms of the film thickness at the exit of the orifice, the cone angle of the spray and the discharge coefficient of the nozzle. It was found that the numerical results of the large-scale atomizer with turbulent Reynolds Stress model yield more accurate solutions than the results with laminar flow model, which indicated that a turbulence flow has been formed within the large scale atomizer. Nevertheless, when the turbulent model was applied to a production-scale atomizer tested by Lacava (2004), its numerical results did not fit well with the experimental data any more. It was found that the Reynolds number of the flow in production-scale atomizer is about 2000, which is one order of magnitude lower than the Reynold’s number in the large-scale atomizer. As such a laminar flow model was successfully applied to its internal flow simulation and it is shown that the numerical results of the production-scale atomizer with laminar model yield more accurate solutions than the results with turbulent flow model. Finally, the effects of orifice contraction angle and mass flow rate were investigated in the production-scale pressure swirl atomizer using the laminar model. The numerical results showed that the discharge coefficient keeps almost constant with increasing orifice contraction angle, and the discharge coefficient, the film thickness at exit and the spray cone angle also almost keep constant with increasing the mass flow rate.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A041. doi:10.1115/GT2014-26808.

Lean burn aero-engine combustors usually exploit advanced prefilming airblast injection systems in order to promote the formation of highly homogeneous air-fuel mixtures with the main aim of reducing NOx emissions. The combustion process is strongly influenced by the liquid fuel preparation and a reliable prediction of pollutant emissions requires proper tools able to consider the most important aspects characterizing liquid film evolution and primary breakup. This paper presents the numerical analysis of an advanced lean burn injection system using a multi-coupled two-phase flow three-dimensional solver developed on the basis of OpenFOAM modelling and numerics. The solver allows the coupled solution of gas-phase, droplets and liquid film exploiting correlation-based and theoretical models for liquid film primary atomization. A detailed analysis of the liquid film evolution is presented together with an investigation of the influence of film modelling and primary breakup on the combustion process at different operating conditions. The combustion field is strongly influenced by the characteristics of droplet population generated by the liquid film and this study proposes a computational setup, suitable for industrial calculations, able to account for all the main physical processes that characterize advanced prefilming airblast injection systems.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A042. doi:10.1115/GT2014-26811.

Humidified gas turbines using steam generated from excess heat feature increased cycle efficiencies. Injecting the steam into the combustor reduces NOx emissions, flame temperatures and burning velocities, promising a clean and stable combustion of highly reactive fuels, such as hydrogen or hydrogen-methane blends. This study presents laminar burning velocities for methane, and hydrogen-enriched methane (10 mol% and 50 mol%) at steam contents up to 30% of the air mass flow. Experiments were conducted on prismatic Bunsen flames stabilized on a slot-burner employing OH planar laser-induced fluorescence for determining the flame front areas. The experimental burning velocities agree well with results from one dimensional simulations using the GRI 3.0 mechanism. Burning velocities are increased with hydrogen enrichment, and reduce non-linearly with ascending steam mole fractions, showing the potential of steam dilution for a stable combustion of these fuels over a wide flammability range. Additionally measured NOx and CO emissions reveal a strong reduction in NOx emissions for an increasing dilution with steam, whereas CO curves are shifted towards higher equivalence ratios.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A043. doi:10.1115/GT2014-26829.

The advent of annular combustors introduced a new facet of flame-acoustic interaction: flame coupling with standing and spinning azimuthal acoustic waves. Such coupling involves an acoustic field that is essentially transverse to the flame. Recent experiments on single burner test rigs have provided significant insight into flame interaction with transverse standing waves. However, experiments that focus on the spinning/rotating nature of azimuthal instabilities are still lacking. In this report, we demonstrate a methodology for studying spinning azimuthal instabilities on a single burner test rig. This methodology is based on analyzing flame response to a traveling acoustic wave generated in the combustor. We generate traveling acoustic waves in our transverse acoustic forcing test-rig by converting one end of the transverse extensions to a non-reflecting end. This is achieved through the implementation of the technique of impedance tuning. In the paper, we have discussed this implementation, followed by discussions on the effects of a traveling acoustic wave on a swirl-stabilized flame. The discussion is in the form of a comparison of flame oscillations for traveling wave and standing wave transverse forcing cases. Results show that the effect of transverse pressure oscillations dominates the flame response to traveling acoustic waves.

Topics: Acoustics , Waves , Flames
Commentary by Dr. Valentin Fuster
2014;():V04BT04A044. doi:10.1115/GT2014-26902.

This paper presents particulate matter (PM) size spectral measurements, analysed to determine number and mass concentration, taken using a fast response differential mobility spectrometer (DMS500). Exhaust samples from multiple commercially available large civil aviation gas turbine engines and an auxiliary power unit operating at high and low engine power conditions were studied, in addition to a simulated aviation gas turbine exhaust, which was operated to exhibit specific PM output. Results show all exhaust sources as having similar bi-modal PM size spectra with both number and mass concentrations highly dependent on the emission source and the sampling testing condition. When operating at high power levels all of the tested gas turbine emission sources, with the exception of the 2-stage combustor design, generally produced distributions of PM which exhibited larger average mean diameter particle sizes and higher number and mass concentrations.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A045. doi:10.1115/GT2014-26906.

In the present paper a numerical analysis of a low NOx partially premixed burner for industrial gas turbine applications is presented. The first part of the work is focused on the study of the premixing process inside the burner. Standard RANS CFD approach was used: k–ε turbulence model was modified and calibrated in order to find a configuration able to fit available experimental profiles of fuel/air concentration at the exit of the burner. The resulting profiles at different test points have been used to perform reactive simulations of an experimental test rig, where exhaust NOx emissions were measured. An assessment of the turbulent combustion model was carried out with a critical investigation of the expected turbulent combustion regimes in the system and taking into account the partially premixed nature of the flame due to the presence of diffusion type pilot flames. A reliable numerical setup was discovered by comparing predicted and measured NOx emissions at different operating conditions and at different split ratio between main and pilot fuel. In the investigated range, the influence of the premixer in the NOx formation rate was found to be marginal if compared with the pilot flame one. The calibrated numerical setup was then employed to explore possible modifications to fuel injection criteria and fuel split, with the aim of minimizing exhaust NOx emissions. This preliminary numerical screening of alternative fuel injection strategies allowed to define a set of advanced configurations to be investigated in future experimental tests.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A046. doi:10.1115/GT2014-26907.

This paper is concerned with the potential use of Helmholtz resonators to provide increased acoustic damping within aero gas turbine combustion systems. Experimental measurements were undertaken using a high intensity facility into which a three burner combustor sector (non-reacting) model could be incorporated. In this way the performance of various damper geometry combinations were assessed. The effect of incident noise levels was also considered along with the associated transition from linear absorption (i.e. where absorption is directly proportional to incident pressure magnitude) to nonlinear absorption (i.e. where the proportion of acoustic loss decreases with increasing noise levels). This complicates the performance comparison between different damping geometries and means care is required when relating laboratory to engine operating conditions. In addition, all the measurements were undertaken in the presence of fuel injectors and other realistic flow field features found within a combustion system and which could affect damping performance. Finally, experimental and numerical assessment was made of the noise levels at which ingestion of hot gas will occur into the resonator cavities with and without the presence of a purging flow. For the geometries investigated ingestion occurs when the fluid displacement in the neck during an acoustic cycle is approximately equal to, or greater than, the resonator neck length. The ratio of fluid displacement and neck length provides a limit for the noise levels at which hot gas is ingested into the cavity and hence the operating condition where damping performance and system mechanical integrity is significantly compromised.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A047. doi:10.1115/GT2014-26927.

In modern gas turbine engines swirl is typically imparted to the airflow as it enters the region of heat release to stabilize the flame. This swirling airstream is often highly turbulent and contains non-uniformities such as swirl vane wakes. However, it is within this environment that fuel atomization takes place. This paper is concerned with the potential effect of these airstream characteristics on the atomization process. Such a flow field is difficult to capture within simplified geometries and so measurements have been made within, and downstream of, injector representative geometries. This is experimentally challenging and required the application of a variety of techniques. The geometry considered is thought typical of an air-blast style injector, as may be used within current or future applications, whereby liquid fuel is introduced onto a pre-filming surface over which an airstream passes.

Data is presented which characterizes the atomizing airstream presented to the pre-filming region. This includes significant flow field non-uniformities and turbulence characteristics that are mainly associated with the swirling flow along with the vanes used to impart this swirl. The subsequent development of these aerodynamic features over the pre-filming surface is also captured with, for example, swirl vane wakes being evident through the injector passage and into the downstream flow field. It is argued these characteristics will be common to many injector designs. Measurements with and without fuel indicate the effect of the liquid film, on the non-dimensional aerodynamic flow field upstream of the pre-filming region, is minimal. However, the amount of airflow passing through the pre-filming passage is affected. In addition to characterization of the airstream, its impact on the liquid fuel film and its development along the pre-filming surface is visualized. Furthermore, PDA measurements downstream of the fuel injector (i.e. the injector ‘far-field) are presented and the observed spray characteristics spatially correlated with the upstream aerodynamic atomizing flow field. Hence for the first time a series of experimental techniques have been used to capture and correlate both near and far field atomization characteristics within an engine representative aerodynamic flow field.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A048. doi:10.1115/GT2014-26977.

The development process for gas turbine combustion systems includes single-burner high-pressure combustion tests as an important validation step. In these tests the performance of a combustor is investigated at realistic gas turbine conditions. Measurement techniques that are typically used in these tests include mass flow meters, thermocouples, pressure transducers, and probes for exhaust-gas composition measurements. These measurement techniques, however, do not provide direct information of the flame behavior.

Chemiluminescence measurements have proven to being a valuable and robust technique to close this gap [1,2]. This paper summarizes the results of chemiluminescence measurements performed at Siemens full-scale high-pressure single-burner combustion test rigs at the German Aerospace Center (DLR) in Cologne, Germany. To minimize the impact of the measurement system on the experiment, the optical access to the test rigs was provided by a water-cooled endoscopic probe. The probe was located in a side-wall downstream of the burner, viewing upstream towards the burner outlet. The probe was successfully operated up to full engine pressure and flame temperatures of approximately 1900 K.

For the detection of the chemiluminescence signal different approaches were applied:

• Spectral analyses of the chemiluminescence signal were done by using an USB spectrometer.

• For flame imaging up to two intensified CCD cameras were applied. In front of the cameras various combinations of optical filters were installed to selectively record the respective chemiluminescent species (OH*, CH*, CO2*).

• For studies with special focus on combustion dynamics an intensified high-speed CMOS camera was used. High-repetition-rate measurements were used for identifying the shapes of flame modes.

• Acoustic pressure oscillations inside the combustion chamber were recorded by pressure transducers simultaneously to the camera images. This allows the pressure oscillations to be correlated with flame fluctuations during post-processing [3,4].

Generally, the robustness of endoscopic chemiluminescence measurements was successfully demonstrated in numerous tests at realistic gas turbine conditions. The applied imaging setups provided new information about the connection between the flame position and NOx emissions as well as the correlation of flame fluctuations and pressure oscillations. Hence, they have become a valuable experimental tool to improve the evaluation and understanding of the combustor performance. Future work will focus on further improvement of quantitative evaluations by compensation of line-of-sight image integration, reabsorption of OH* by OH, and beam steering.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A049. doi:10.1115/GT2014-27002.

Since lean premixed combustion allows for fuel-efficiency and low emissions, it is nowadays state of the art in stationary gas turbines. In the long term, it is also a promising approach for aero engines, when safety issues like lean blowout (LBO) and flame flashback in the premixer can be overcome. While for the use of hydrogen the LBO limits are extended, the flashback propensity is increased. Thus, axial air injection is applied in order to eliminate flashback in a swirl-stabilized combustor burning premixed hydrogen. Axial injection constitutes a non-swirling jet on the central axis of the radial swirl generator which influences the vortex breakdown position. In the present work changes in the flow field and their impact on flashback limits of a model combustor are evaluated. First, a parametric study is conducted under isothermal test conditions in a water tunnel employing particle image velocimetry (PIV). The varied parameters are the amount of axially injected air and swirl number. Subsequently, flashback safety is evaluated in the presence of axial air injection in an atmospheric combustor test rig and a stability map is recorded. The flame structure is measured using high-speed OH* chemiluminescence imaging. Simultaneous high-speed PIV measurements of the reacting flow provide insight in the time-resolved reacting flow field and indicate the flame location by evaluating the Mie scattering of the raw PIV images by the means of the Qualitative Light Sheet (QLS) technique.

The isothermal tests identify the potential of axial air injection to overcome the axial velocity deficits at the nozzle outlet, which is considered crucial in order to provide flashback safety. This effect of axial air injection is shown to prevail in the presence of a flame. Generally, flashback safety is shown to benefit from an elevated amount of axial air injection and a lower swirl number. Note, that the latter also leads to increased NOx emissions, while axial air injection does not. Additionally, fuel momentum is indicated to positively influence flashback resistance, although based on a different mechanism, an explanation of which is suggested. In summary, flashback-proof operation of the burner with a high amount of axial air injection is achieved on the whole operating range of the test rig at inlet temperatures of 620 K and up to stoichiometric conditions while maintaining single digit NOx emissions below a flame temperature of 2000 K.

Topics: Combustion , Hydrogen
Commentary by Dr. Valentin Fuster
2014;():V04BT04A050. doi:10.1115/GT2014-27023.

The majority of recent stationary gas turbine combustors employ swirling flows for flame stabilization. The swirling flow undergoes vortex breakdown and exhibits a complex flow field including zones of recirculating fluid and regions of high shear. Often, self-excited helical flow instabilities are found in these flows that may influence the combustion process in beneficial and adverse ways. In the present study we investigate the occurrence and shape of self-excited hydrodynamic instabilities and the related heat-release fluctuations over a wide range of operating conditions. We employ high-speed stereoscopic particle image velocimetry and simultaneous OH*-chemiluminescence imaging to resolve the flow velocities and heat release distribution, respectively. The results reveal four different flame shapes: A detached annular flame, a long trumpet shaped flame, a typical V-flame, and a very short flame anchored near the combustor inlet. The flame shapes were found to closely correlate with the reactivity of the mixture. Highly steam-diluted or very lean flames cause a detachment, whereas hydrogen fuel leads to very short flames. The detached flames feature a helical instability, which in terms of frequency and shape is similar to the isothermal case. A complete suppression of the helical structure is found for the V-flame. Both, the trumpet shaped flame and the very short flame feature helical instabilities of different frequencies and appearances. The phase-averaged OH*-chemiluminescence images show that the helical instabilities cause large scale-heat release fluctuations. The helical structure of the fluctuations is verified using a tomographic reconstruction technique.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A051. doi:10.1115/GT2014-27034.

This work discusses from a system theoretic point of view the low order modeling and identification of the acoustic scattering behavior of a ducted flame.

In this context, one distinguishes between black-box and grey-box models. The former rely on time series data only and do not require any physical modeling of the system that is to be identified. The latter exploit prior knowledge of the system physics to some extent and in this sense are physically motivated.

For the case of a flame stabilized in a duct, a grey-box model is formulated that comprises an acoustic part as well as sub-models for the flame dynamics and the jump conditions for acoustic variables across the region of heat release. Each of the subsystems can be modeled with or without physical a priori knowledge, in combination they yield a model for the scattering behavior of the flame.

We demonstrate these concepts by analyzing a CFD model of a laminar conical premixed flame. The grey-box approach allows to optimize directly the scattering behavior of the identified model. Furthermore, we show that the method allows to estimate heat release rate fluctuations as well as the flame transfer function from acoustic measurements only.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A052. doi:10.1115/GT2014-27059.

Flow oscillations associated with hydrodynamic instabilities comprise a key element of the feedback loop during self-excited combustion driven oscillations. This work is motivated in particular by the question of how to scale thermoacoustic stability results from single nozzle or sector combustors to full scale systems. Specifically, this paper considers the response of non-axisymmetric flames to helical flow disturbances of the form Display Formulau^iexpimθ where m denotes the helical mode number. This work closely follows prior studies of the response of axisymmetric flames to helical disturbances. In that case, helical modes induce strong flame wrinkling, but only the axisymmetric, m = 0 mode, leads to fluctuations in overall flame surface area and, therefore, heat release. All other helical modes induce local area/heat release fluctuations with azimuthal phase variations that cancel each other out when integrated over all azimuthal angles. However, in the case of mean flame non-axisymmetries, the azimuthal deviations on the mean flame surface inhibit such cancellations and the asymmetric helical modes (m ≠ 0) cause a finite global flame response. In this paper, a theoretical framework for non-axisymmetric flames is developed and used to illustrate how the flame shape influences which helical modes lead to net flame surface area fluctuations. Example results are presented which illustrate the contributions made by these asymmetric helical modes to the global flame response and how this varies with different control parameters such as degree of asymmetry in the mean flame shape or Strouhal number. Thus, significantly different sensitivities may be observed in single and multi-nozzle flames in otherwise identical hardware in flows with strong helical disturbances, if there are significant deviations in time averaged flame shape between the two, particularly if one of the cases is nearly axisymmetric.

Topics: Flames
Commentary by Dr. Valentin Fuster
2014;():V04BT04A053. doi:10.1115/GT2014-27080.

Ignition is a problem of fundamental interest with critical practical implications. While there are many studies of ignition of single injector configurations, the transient ignition of a full annular combustor has not been extensively investigated, mainly because of the added geometrical complexity. The present investigation combines simulations and experiments on a complete annular combustor. The setup, developed at EM2C laboratory, features sixteen swirl injectors and quartz walls allowing direct visualization of the flame. High speed imaging is used to record the space time flame structure and study the dynamics of the light-round process. On the numerical side, massively parallel computations are carried out in the LES framework using the Filtered Tabulated (F-TACLES) flamelet model. Comparisons are carried out at different instants during the light-round process between experimental data and results of calculations. It is found that the simulation results are in remarkable agreement with experiments provided that the thermal effects at the walls are considered. Further post-processings indicate that the flame burning velocity and flame front geometry are close to those found in the experiment. This analysis confirms that the LES framework used for these calculations and the selected combustion model are adequate for such calculations but that further work is needed to confirm that ignition prediction can be used reliably over a range of operating parameters.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A054. doi:10.1115/GT2014-27099.

This paper presents the results of an investigation in which the fuel/air mixing process in a single slot within the radial swirler of a dry low emission (DLE) combustion system is explored using air/air mixing. Experimental studies have been carried out on an atmospheric test facility in which the test domain is a large-scale representation of a swirler slot from a Siemens DLE SGT-400 combustion system. Hot air with a temperature of 300°C is supplied to the slot, while the injected fuel gas is represented using air jets with temperatures of about 25°C. Temperature has been used as a scalar to measure the mixing of the jets with the cross-flow. The mixture temperatures were measured using thermocouples while Pitot probes were used to obtain local velocity measurements. The experimental data have been used to validate a computational fluid dynamics (CFD) mixing model.

Numerical simulations were carried out using CFD software ANSYS-CFX. Due to the complex three-dimensional flow structure inside the swirler slot, different RANS turbulence models were tested. The shear stress transport (SST) turbulence model was observed to give best agreement with the experimental data. The momentum flux ratio between the main air flow and the injected fuel jet, and the aerodynamics inside the slot, were both identified by this study as major factors in determining the mixing characteristics. It has been shown that mixing in the swirler can be significantly improved by exploiting the aerodynamic characteristics of the flow inside the slot. The validated CFD model provides a tool which will be used in future studies to explore fuel/air mixing at engine conditions.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A055. doi:10.1115/GT2014-27123.

A detailed chemistry model is necessary to simulate the effects of variations in fuel composition on soot emissions. In this work, we have developed a detailed chemistry model for the soot formation and oxidation chemistry, with a focus on the surface kinetics of the soot-particle. The model has been compared to a unique set of soot particle-size data measured in flames for several single-component fuels. Fuel components used in the experiments represent the chemical classes found in jet, gasoline, and diesel fuels, including n-heptane (representative of n-alkanes) and toluene (aromatic). Measurements were taken in burner-stabilized stagnation-flame (BSSF) experiments, which can be simulated well using the 1-dimensional BSSF flame model in CHEMKIN-PRO. Soot volume fraction and particle size distributions are modeled using the sectional method option for Particle Tracking, within CHEMKIN-PRO software. The well-characterized flow of the BSSF experiments allows the modeling to focus on the kinetics. Validated detailed reaction mechanisms for fuel combustion and PAH production, combined with the new soot surface-kinetics mechanism, were used in the simulations. Simulation results were compared to measurements for both particle size distributions and total soot volume fraction. Observed effects of fuel, temperature, pressure, equivalence ratio and residence time on the soot size distribution shape and soot quantity were reproduced by the model.

The chemistry in the soot surface model includes particle nucleation, growth through the HACA (hydrogen-abstraction/carbon-addition) and PAH-condensation (polycyclic aromatic hydrocarbons) pathways, as well as soot-oxidation pathways. In addition to soot chemistry, the physics of particle coagulation and aggregation were included in the model. The results demonstrate the ability of well-validated chemistry to predict both dramatic and subtle effects related to soot mass and soot particle size.

Topics: Fuels , Chemistry , Soot
Commentary by Dr. Valentin Fuster
2014;():V04BT04A056. doi:10.1115/GT2014-27138.

MILD combustion (Moderate or Intense Low Oxygen Dilution) is a well known technique that can substantially reduce high temperature regions in burners and thereby reduce thermal NOx emissions. This technology has been successfully applied to conventional furnace systems and seems to be an auspicious concept for reducing NOx and CO emissions in stationary gas turbines. To achieve a flameless combustion regime, fast mixing of recirculated burnt gases with fresh air and fuel in the combustion chamber is needed. In the present study, the combustor concept is based on the reverse flow configuration with two concentrically arranged nozzles for fuel and air injections. The present work deals with the active control of MILD combustion for gas turbine applications. For this purpose, a new concept of air flow rate pulsation is introduced. The pulsating unit offers the possibility to vary the inlet pressure conditions with a high degree of freedom: amplitude, frequency and waveform. The influence of air flow pulsation on MILD combustion is analyzed in terms of NOx and CO emissions. Results under atmospheric pressure show a drastic decrease of NOx emissions, up to 55%, when the pulsating unit is active. CO emissions are maintained at a very low level so that flame extinction is not observed. To get more insights into the effects of pulsation on combustion characteristics, velocity fields in cold flow conditions are investigated. Results show a large radial transfer of flow when pulsation is activated, hence enhancing the mixing process. The flame behavior is analyzed by using OH* chemiluminescence. Images show a larger distributed reaction region over the combustion chamber for pulsation conditions, confirming the hypothesis of a better mixing between fresh and burnt gases.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A057. doi:10.1115/GT2014-27203.

Combustion noise and thermo-acoustic instabilities are of primary importance in highly critical applications such as rocket propulsion systems, power generation, and jet propulsion engines. Mechanisms for combustion instabilities are extremely complex because they often involve interactions among several different physical phenomena such as unsteady flame propagation leading to unsteady flow field, acoustic wave propagation, natural and forced hydrodynamic instabilities, etc. In the past, we have utilized porous inert media (PIM) to mitigate combustion noise and thermo-acoustic instabilities in both lean premixed (LPM) and lean direct injection (LDI) combustion systems. While these studies demonstrated the efficacy of the PIM concept to mitigate noise and thermo-acoustic instabilities, the actual mechanisms involved have not been understood. The present study utilizes time-resolved particle image velocimetry to measure the turbulent flow field in a non-reacting swirl-stabilized combustor without and with PIM. Although the flow field inside the annulus of the PIM cannot be observed, measurements immediately downstream of the PIM provide insight into the turbulent structures. Results are analyzed using the Proper Orthogonal Decomposition (POD) method and show that the PIM alters the flow field in an advantageous manner by modifying the turbulence structures and eliminating the corner recirculation zones and precessing vortex core, which would ultimately affect the acoustic behavior in a favorable manner.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A058. doi:10.1115/GT2014-27220.

The research and development described in this paper arises from the need for modeling a realistic fuel atomization process in a complex combustor/augmentor fuel injector. In the atomization process, it is important to understand the primary breakup mechanism and to predict the resulting fuel droplets. However, the mechanism of atomization and the resulting spray formation processes in realistic complex fuel injectors are not well understood because experimental access to the atomization region is typically severely limited. A significant portion of the atomization process occurs in spatial regions adjacent to solid walls that block experimental access into the injector so that experimental studies are limited to either far field measurements of complex injectors, after most of the atomization has occurred, or to simple injector geometries such as a circular cross-section pipes injecting into crossflow channels.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A059. doi:10.1115/GT2014-27225.

A large eddy simulation (LES) of a turbulent swirl stabilized jet-A1 flame is presented. The scope of the study is to incorporate a reduced chemistry model, as well as, coupling the turbulent flow characteristics to the chemical reactions and at the same time model the local chemical non-equilibrium due to the turbulent strain. Standard Eulerian and Lagrangian approaches are used to describe both gas and liquid phases, respectively. A joint presumed probability density function (PDF) is used to model turbulent-chemistry interactions in swirling jet-A1 spray flames. A one-component fuel, n-decane, is used as a surrogate for jet-A1. The combustion chemistry of the one component is represented through a reduced chemical kinetic mechanism (CKM) which comprises 139 species and 1 045 reactions, derived from the detailed jet fuel surrogate model, JetSurf 2.0. Numerical results of the gas velocity, the gas temperature and the species mole fractions are compared with a set of published experimental data of a steady flame. In addition to the overall reasonable agreement obtained with the experimental data, it is observed that, by combining a sufficiently realistic chemistry model with LES to simulate a jet-A1 spray flame, the prediction of major species is significantly improved while pollutants such as carbon monoxide (CO) and other species involved in slow reactions, are under predicted for reasons discussed in the paper.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A060. doi:10.1115/GT2014-27258.

Stringent emissions regulations have led engine manufacturers to focus on fuel-efficient low emission technologies. Basic understanding and modeling of fundamental mechanisms governing formation and destruction of NOx, CO and UHC is essential to reduce pollutant emissions. Recent advances in turbulent combustion modeling have enabled designers to use CFD as a design tool for evaluating low emission concepts at the conceptual design phase.

Prediction of pollutant NOx for gas turbine combustors has proven successful for design validation applications. The challenge is to provide quick and accurate estimates of NOx for application to gas turbine combustor preliminary design phase, which can be characterized by multiple design changes, varying operating conditions and a variety of fuel staging concepts. NOx formation processes are typically slow compared to the fast hydrocarbon oxidation reactions. As a result, NOx predictions are typically performed as a post-processing step on thermal field obtained from reacting flow simulations. This work builds on prior work on flamelet approach [1,3] by suitably blending it with FLUENT®’s species transport. NOx production within gas turbine combustors has contributions from two major sources: flame front & post-flame thermal NO. The flame front contributions are obtained from flamelet based computations involving detailed chemistry whereas the slow evolution of post-flame NOx is tracked by explicitly solving for NO species transport. The closure of turbulence-chemistry-interactions is derived from Girimaji’s [2] assumed PDF closure using temperature-composition correlations. A Gaussian PDF shape is used with mean and variance of temperatures accounting for the first and second moments, required for PDF weighting computations. The formulation has been validated against SANDIA D flame, and then extended to GE Aviation’s fielded combustors over a wide range of operating conditions, with errors within 11% at Take-Off condition. The model has also been used for pre-test predictions on a number of combustors under development.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A061. doi:10.1115/GT2014-27266.

The prediction of dynamic instability remains an open and important issue in the development of gas turbine systems, particularly those constrained by emissions limitations. The existence and characteristics of dynamic instability are known to be functions of combustor geometry, flow conditions, and combustion parameters, but the form of dependence is not well understood. By modifying the acoustic boundary conditions, changes in flame and flow structure due to inlet parameters can be studied independent of the acoustic modes with which they couple. This paper examines the effect of equivalence ratio on the flame macrostructure — the relationship between the turbulent flame brush and the dominant flow structures — in an acoustically uncoupled environment. The flame brush is measured using CH* chemiluminescence, and the flow is interrogated using two-dimensional particle image velocimetry. We examine a range of equivalence ratios spanning three distinct macrostructures. The first macrostructure (ϕ = 0.550) is characterized by a diffuse flame brush confined to the interior of the inner recirculation zone. We observe a conical flame in the inner shear layer, continuing along the wall shear layer in the second macrostructure (ϕ = 0.600). The third macrostructure exhibits the same flame brush as the second, with an additional flame brush in the outer shear layer (ϕ = 0.650). Between the second and third macrostructures, we observe a regime in which the flame brush transitions intermittently between the two structures. We use dynamic mode decomposition on the PIV data to show that this transition event, which we call flickering, is linked to vorticity generated by the intermittent expansion of the outer recirculation zone as the flame jumps in and out of the outer shear layer. In a companion paper, we show how the macrostructures described in this paper are linked with dynamic instability [1].

Commentary by Dr. Valentin Fuster
2014;():V04BT04A062. doi:10.1115/GT2014-27293.

Models of finite rate combustion chemistry based on tabulation have been found to be efficient and accurate for a broad range of fuel and flame configurations. However, the modelling of formation of oxides of nitrogen remains a challenging issue. Even recent developments that take into account heat losses do not provide accurate prediction of NOx species despite a very good match for major species and temperature. The present paper introduces a model based on a chemical lookup-table for the NO formation rate in the preheat, initiation and fuel consumption region of a premix flame combined with algebraic relations in the post-flame area, thus reducing advantageously the tabulated information. This model is applied to a premixed laminar burner and compared to previous models based on tabulated chemistry.

Commentary by Dr. Valentin Fuster
2014;():V04BT04A063. doi:10.1115/GT2014-27316.

In this paper, we conduct an experimental investigation of a confined premixed swirl-stabilized dump combustor similar to those found in modern gas turbines. We operate the combustor with premixed methane-air in the lean range of equivalence ratio ϕ ∈ [0.5–0.75]. First, we observe different dynamic modes in the lean operating range, as the equivalence ratio is raised, confirming observations made previously in a similar combustor geometry but with a different fuel [1]. Next we examine the correspondence between dynamic mode transitions and changes in the mean flame configuration or macrostructure. We show that each dynamic mode is associated with a specific flame macrostructure. By modifying the combustor length without changing the underlying flow, the resonant frequencies of the geometry are altered allowing for decoupling the heat release fluctuations and the acoustic field, in a certain range of equivalence ratio. Mean flame configurations in the modified (short) combustor and for the same range of equivalence ratio are examined. It is found that not only the same sequence of flame configurations is observed in both combustors (long and short) but also that the set of equivalence ratio where transitions in the flame configuration occur is closely related to the onset of thermo-acoustic instabilities. For both combustor lengths, the flame structure changes at similar equivalence ratio whether thermo-acoustic coupling is allowed or not, suggesting that the flame configuration holds the key to understanding the onset of self-excited thermo-acoustic instability in this range. Finally, we focus on the flame configuration transition that was correlated with the onset of the first dynamically unstable mode ϕ ∈ [0.61–0.64]. Our analysis of this transition in the short, uncoupled combustor shows that it is associated with an intermittent appearance of a flame in the outer recirculation zone (ORZ). The spectral analysis of this “ORZ flame flickering” — based on flame chemiluminescence data — shows the presence of unsteady events occurring at two distinct frequency ranges. A broad band at low frequency in the range ∼[1 Hz – 10 Hz] and a narrow band centered around 28 Hz.

Topics: Combustion , Flames
Commentary by Dr. Valentin Fuster

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In