Fuel Composition Effects on Forced Ignition of Liquid Fuel Sprays

In gas turbine combustors, ignition is achieved by using sparks from igniters to start a flame. The process of sparks interacting with fuel/air mixture and creating self-sustained flames is termed forced ignition. Physical and chemical properties of a liquid fuel can influence forced ignition. The physical effects manifest through processes such as droplet atomization, spray distribution, and vaporization rate. The chemical effects impact reaction rates and heat release. This study focuses on the effect of fuel composition on forced ignition of fuel sprays in a well-controlled flow with a commercial style igniter. A facility previously used to examine prevaporized, premixed liquid fuel-air mixtures is modified and employed to study forced ignition of liquid fuel sprays. In our experiments, a wall-mounted, high energy, recessed cavity discharge igniter operating at 15 Hz with average spark energy of 1.25 J is used to ignite liquid fuel spray produced by a pressure atomizer located in a uniform air coflow. The successful outcome of each ignition events is characterized by the (continued) presence of chemiluminescence 2 ms after spark discharge, as detected by a high-speed camera. The ignition probability is defined as the fraction of successful sparks at a fixed condition, with the number of events evaluated for each fuel typically in the range 600–1200. Ten fuels were tested, including standard distillate jet fuels (e.g., JP-8 and Jet-A), as well as many distillate and alternative fuel blends, technical grade n-dodecane, and surrogates composed of a small number of components. During the experiments, the air temperature is controlled at 27 C and the fuel temperature is controlled at 21 C. Experiments are conducted at a global equivalence ratio of 0.55. Results show that ignition probabilities correlate strongly to liquid fuel viscosity (presumably through droplet atomization) and vapor pressure (or recovery temperature), as smaller droplets of a more volatile fuel would lead to increased vaporization rates. This allows the kernel to transition to a self-sustained flame before entrainment reduces its temperature to a point where chemical rates are too slow. Chemical properties of the fuel showed little influence, except when the fuels had similar physical properties. This result demonstrates that physical properties of liquid fuels have dominating effects on forced ignition of liquid fuel spray in coflow air.

Analysis of the Mixing and Emission Characteristics in a Model Combustor

Modern gas turbines in power systems employ lean premixed combustion to lower flame temperature and thus achieve low NOx emissions. The fuel/air mixing process and its impacts on emissions are of paramount importance to combustor performance. In this study, the mixing process in a methane-fired model combustor was studied through an integrated experimental and numerical study. The experimental results show that at the dump location, the time-averaged fuel/air unmixedness is less than 10% over a wide range of testing conditions, demonstrating the good mixing performance of the specific premixer on the time-averaged level. A study of the effects of turbulent Schmidt number on the unmixedness prediction shows that for the complex flow field involved, it is challenging for Reynolds-Averaged Navier-Stokes (RANS) simulations with constant turbulent Schmidt number to accurately predict the mixing process throughout the combustor. Further analysis reveals that the production and scalar dissipation are the key physical processes controlling the fuel/air mixing. Finally, the NOx formation in this model combustor was analyzed and modelled through a flamelet-based approach, in which NOx formation is characterized through flame-front NOx and its post-flame formation rate obtained from one-dimensional laminar premixed flames. The effect of fuel/air unmixedness on NOx formation is accounted for through the presumed probability density functions (PDF) of mixture fraction. Results show that the measured NOx in the model combustor are bounded by the model predictions with the fuel/air unmixedness being 3% and 5% of the maximum unmixedness. In the context of RANS, the accuracy in NOx prediction depends on the unmixedness prediction which is sensitive to turbulent Schmidt number.

Effects of Reacting Conditions on Flow Fields in a Swirl Stabilized Lean Premixed Can Combustor

PIV measurements to understand the flow differences between reacting and non-reacting conditions were conducted in an optically accessible single can combustor. An industrial fuel nozzle was installed at the inlet of the test section to generate the swirl flow for flame stabilization and simulate realistic conditions of a gas turbine combustor. Five different equivalence ratios between 0.50 and 0.75 were tested with propane as fuel. Main air flow was also varied from Reynolds number from 50000 to 110000 with respect to the fuel nozzle diameter. Effect of preheating was tested by changing inlet air temperature from 23 to 200°C. The pressure at the test section was close to atmospheric condition throughout the tests. The measurements were performed with a 2-D PIV system. Time-averaged flow velocity, vorticity and turbulent kinetic energy (TKE) were obtained from PIV data and flow structures under different conditions were compared. Swirl jet impingement location on the liner wall was determined as well to understand the impact on the liner wall. Proper orthogonal decomposition (POD) further analyzed the data to compare coherent structures in the reacting and non-reacting flows.

Feature Extraction From Time Resolved Reacting Flow Data Sets

This paper presents measurements of the simultaneous fuel distribution, flame position and flow velocity in a high pressure, liquid fueled combustor. Its objective is to develop methods to process, display and compare large quantities of instantaneous data with computations. However, time-averaged flow fields rarely represent the instantaneous, dynamical flow fields in combustion systems. It is therefore important to develop methods that can algorithmically extract dynamical flow features and be directly compared between measurements and computations. While a number of data-driven approaches have been previously presented in the literature, the purpose of this paper is to propose several approaches that are based on understanding of key physical features of the flow — for this reacting swirl flow, these include the annular jet, the swirling flow which may be precessing, the recirculating flow between the annular jets, and the helical flow structures in the shear layers. This paper demonstrates nonlinear averaging of axial and azimuthal velocity profiles, which provide insights into the structure of the recirculation zone and degree of flow precession. It also presents probability fields for the location of vortex cores that enables a convenient method for comparison of their trajectory and phasing with computations. Taken together, these methods illustrate the structure and relative locations of the annular fluid jet, recirculating flow zone, spray location, flame location, and trajectory of the helical vortices.

Investigation of the Down-Scaling Effects on the Low Swirl Burner and its Application to Microturbines

About 75% of the electric power generated by centralized power plants feeds the energy needs from the residential and commercial sectors. These power plants waste about 67% of primary energy as heat emitting 2 billion tons of CO2 per year in the process (∼ 38% of total US CO2 generated per year) [1]. A study conducted by the United States Department of Energy indicated that developing small-scale combined heat and power systems to serve the commercial and residential sectors could have a significant impact on both energy savings and CO2 emissions. However, systems of this scale historically suffer from low efficiencies for a variety of reasons. From a combustion perspective, at these small scales, few systems can achieve the balance between low emissions and high efficiencies due in part to the increasing sensitivity of the system to hydrodynamic and heat transfer effects. Addressing the hydrodynamic impact, the effects of downscaling on the flowfield evolution were studied on the low swirl burner (LSB) to understand if it could be adapted to systems at smaller scales. Utilizing particle image velocimetry (PIV), three different swirlers were studied ranging from 12 mm to 25.4 mm representing an output range of less than 1 kW to over 23 kW. Results have shown that the small-scale burners tested exhibited similar flowfield characteristics to their larger-scale counterparts in the non-reacting cases studied. Utilizing this data, as a proof of concept, a 14 mm diameter LSB with an output of 3.33 kW was developed for use in microturbine operating on a recuperated Brayton cycle. Emissions results from this burner proved the feasibility of the system at sufficiently lean mixtures. Furthermore, integration of the newly developed LSB into a can style combustor for a microturbine application was successfully completed and comfortably meet the stringent emissions targets. While the analysis of the non-reacting cases was successful, the reacting cases were less conclusive and further investigation is required to gain an understanding of the flowfield evolution which is the subject of future work.

Flow Inhomogeneities in a Realistic Aeronautical Gas-Turbine Combustor: Formation, Evolution and Indirect Noise

Indirect noise generated by the acceleration of combustion inhomogeneities is an important aspect in the design of aeroengines because of its impact on the overall noise emitted by an aircraft and the possible contribution to combustion instabilities. In this study, a realistic rich-quench-lean combustor is numerically investigated, with the objective of quantitatively analyzing the formation and evolution of flow inhomogeneities and determine the level of indirect combustion noise in the nozzle guide vane (NGV). Both entropy and compositional noise are calculated in this work. A high-fidelity numerical simulation of the combustion chamber, based on the Large-Eddy Simulation (LES) approach with the Conditional Moment Closure (CMC) combustion model, is performed. The contributions of the different air streams to the formation of flow inhomogeneities are pinned down and separated with seven dedicated passive scalars. LES-CMC results are then used to determine the acoustic sources to feed an NGV aeroacoustic model, which outputs the noise generated by entropy and compositional inhomogeneities. Results show that non-negligible fluctuations of temperature and composition reach the combustor’s exit. Combustion inhomogeneities originate both from finite-rate chemistry effects and incomplete mixing. In particular, the role of mixing with dilution and liner air flows on the level of combustion inhomogeneities at the combustor’s exit is highlighted. The species that most contribute to indirect noise are identified and the transfer functions of a realistic NGV are computed. The noise level indicates that indirect noise generated by temperature fluctuations is larger that the indirect noise generated by compositional inhomogeneities, although the latter is not negligible and is expected to become louder in supersonic nozzles. It is also shown that relatively small fluctuations of the local flame structure can lead to significant variations of the nozzle transfer function, whose gain increases with the Mach number. This highlights the necessity of an on-line solution of the local flame structure, which is performed in this paper by CMC, for an accurate prediction of the level of compositional noise. This study opens new possibilities for the identification, separation and calculation of the sources of indirect combustion noise in realistic aeronautical gas turbines.

Ignition Dynamics in an Annular Combustor With Gyratory Flow Motion

In concepts of integrated design of combustor and turbine, an annular combustor model is developed and featured with multiple oblique-injecting swirling injectors to introduce gyratory flow motion in the combustion chamber. The ignition process is experimentally investigated to study the effects of introducing circumferential velocity component Uc to the light-round sequence. Experiments are carried out with premixed propane/air mixture in ambient conditions. The light-round sequence is recorded by a high-speed camera, which provides detailed flame azimuthal positions during the sequence and gives access to the light-round time τ and the circumferential flame propagation speed Sc. The results have also been compared with that obtained from a straight-injecting annular combustor. The effects of bulk velocity Ub, thermal power P and equivalence ratio Φ are also explored. Due to the gyratory flow motion induced by oblique injection, the flame fronts only propagate along the direction of circumferential flow. Both of the circumferential flame propagation speed increase with increasing bulk velocity in two injection types. It seems mainly to depend on bulk velocity, regardless of Φ, in oblique-injecting combustor when compared with the straight one. It indicates that the circumferential velocity component would play a dominant role in light-round sequence when it is sufficient higher than the displacement flame speed.

Experimental Investigation Into the High Altitude Relight of a Three-Cup Combustor Sector

The altitude relight of a gas turbine combustor is an FAA and EASA regulation which dictates the successful re-ignition of an engine and its proper spool-up after an in-flight shutdown. Combustor pressure loss, ambient pressure, ambient temperature, and equivalence ratio were all studied on a full-scale, 3-cup, single-annular aviation combustor sector to create an ignition map. The flame development process was studied through the implementation of high-speed video. Testing was conducted by placing the sector horizontally upstream of an air jet ejector in a high altitude relight testing facility. Air was maintained at room temperature for varying pressure, and then the cryogenic heat exchanger was fed with liquid nitrogen to chill the air down to a limit of −50 deg F, corresponding with an altitude of 30,000 feet. Fuel was injected at constant equivalence ratios across multiple operating conditions, giving insight into the ignition map of the combustor sector. Results of testing indicated difficulty in achieving ignition at high altitudes for pressure drops greater than 2%, while low pressure drops show adequate performance. Introducing low temperatures to simulate the ambient conditions yielded a worse outcome, with all conditions having poor results except for 1%. High-speed video of the flame development process during the relight conditions across all altitudes yielded a substantial effect of the pressure drop on ignitability of the combustor. An increase in pressure drop was associated with a decrease in the likelihood of ignition success, especially at increasing altitudes. The introduction of the reduced temperature effect exacerbated this effect, further hurting ignition. High velocity regions in the combustor were detrimental to the ignition, and high area, low velocity regions aided greatly. The flame tended to settle into the corner recirculation zone and recirculate back into the center-toroidal recirculation zone (CTRZ), spreading downstream and likewise into adjacent swirl cups. These tests demonstrate the need for new combustor designs to consider adding large recirculation zones for combustor flame stability that will aid in relight requirements.

A Comparison of the Transfer Functions and Flow Fields of Flames With Increasing Swirl Number

The frequency response of premixed swirled flames is investigated by comparing their Transfer Function (FTF) between velocity and heat release rate fluctuations. The equivalence ratio and flow velocity are kept constant and four different swirling injectors are tested with increasing swirl numbers. The first injector features a vanishing low swirl number S = 0.20 and produces a flame anchored by the recirculating flow in the wake of a central bluff body. The three other swirling injectors produce highly swirled flows (S > 0.6) leading to a much larger internal recirculation region, which size increases with the swirl level. When operating the burner at S = 0.20, the FTF gain curve smoothly increases to reach a maximum and then smoothly decreases towards zero. For the highly swirled flames (S > 0.6), the FTF gain curve shows a succession of valleys and peaks attributed to interferences between axial and azimuthal velocity fluctuations at the injector outlet. The FTF phase-lag curves from the vanishing low and highly swirled flames are the same at low frequencies despite their large differences in flame length and flame aspect ratio. Deviations between the FTF phase lag curves of the different swirled flames start above the frequency corresponding to the first valley in the FTF gain of the highly swirled flames. Phase averaged images of the axial flow fields and of the flame chemiluminescence are used to interpret these features. At forcing frequencies corresponding to peak FTF gain values, the cold flow response of all flames investigated is dominated by large coherent vortical structures shed from the injector lip. At forcing frequencies corresponding to a valley in the FTF gain curve of the highly swirled flames, the formation of large coherent structures is strongly hindered in the cold flow response. These observations contrast with previous interpretations of the mechanisms associated to the low FTF response of swirled flames. It is finally found that for flames stabilized with a large swirl number, heat release rate fluctuations result both from large flame luminosity oscillations and large flame volume oscillations. For conditions leading to a small FTF gain value, both the flame luminosity and flame volume fluctuations are suppressed confirming the absence of strong perturbations within the flow at these frequencies. The experiments made in this work reveal a purely hydrodynamic mechanism at the origin of the low response of swirling flames at certain specific frequencies.

Modelling the Effect of External Flue Gas Recirculation on NOx and CO Emissions in a Premixed Gas Turbine Combustor With Chemical Reactor Networks

The increasing amount of renewable energy and emission norms challenge gas turbine power plants to operate at part-load with high efficiency, while reducing NOx and CO emissions. A novel solution to this dilemma is external Flue Gas Recirculation (FGR), in which flue gases are recirculated to the gas turbine inlet, increasing compressor inlet temperature and enabling higher part load efficiencies. FGR also alters the oxidizer composition, potentially leading to reduced NOx levels. This paper presents a kinetic model using chemical reactor networks in a lean premixed combustor to study the impact of FGR on emissions. The flame zone is split in two perfectly stirred reactors modelling the flame front and the recirculation zone. The flame reactor is determined based on a chemical time scale approach, accounting for different reaction kinetics due to FGR oxidizers. The recirculation zone is determined through empirical correlations. It is followed by a plug flow reactor. This method requires less details of the flow field, has been validated with literature data and is generally applicable for modelling premixed flames. Results show that due to less O2 concentration, NOx formation is inhibited down to 10–40% and CO levels are escalated up to 50%, for identical flame temperatures. Increasing combustor pressure leads to a rise in NOx due to thermal effects beyond 1800 K, and a drop in CO levels, due to the reduced chemical dissociation of CO2. Wet FGR reduces NOx by 5–10% and increases CO by 10–20%.