Research Overview
A combustor is a race between mixing and residence time. In high-speed flows the air is through the duct in milliseconds; in compact designs the duct itself is short. Either way, the chemistry must finish before the flow leaves. We study flameholding strategies that buy back this margin — cavity-wire hybrids for supersonic combustors, trapped-vortex schemes for compact ones — and the diagnostics needed to see whether they work, from laser absorption to X-ray imaging of sprays.
For investigating high-speed propulsion systems, one of the experimental challenges is to reproduce the flight condition. Since such vehicles fly at high altitudes (low pressure/density) with high Mach numbers (high total enthalpy), a test apparatus should be able to produce low backpressure conditions with high-enthalpy plenum (freestream) conditions reaching thousands of kelvins.
Arc-heated wind tunnels are good examples of such rigs. Other approaches include resistively heated and combustion-vitiated wind tunnels.
The backpressure requirement is usually hard to meet and the greater the backpressure, the shorter the test time will be with a given plenum pressure. Also, with a full flow path, the combustor analysis may be biased due to the upstream inlet conditions.
To avoid these issues, high-speed combustion research is often conducted via direct-connect types. This configuration directly connects the convergent-divergent (C/D) nozzle of the wind tunnel to the inlet of the combustor, isolating the inlet effect. The top left figure presents a typical flow path of a supersonic cavity-stabilized combustor.
The resulting flows can be diagnosed with various techniques, including flame imaging, wall pressure measurements, and laser absorption spectroscopy (LAS).
Two common practices of supersonic flameholders are cavities and struts. Cavity flameholders induce a recirculating flow inside, increasing the flow residence time with minimal losses. On the other hand, strut flameholders produce strong shocks to elevate the pressure/temperature of the flow and create vorticity to enhance mixing at a cost of total pressure.
To combine the benefits of the two, mesh wires are inserted in the flow path of a cavity-stabilized combustor. This attempts to create a distributed deceleration region for enhanced mixing and recirculating zones for longer residence times.
The mesh insert combustors were tested at an equivalence ratio of 0.2.
The blank configuration (top) required an air throttling downstream to force flow choking for additional flow deceleration.
The mesh-coarse (middle) and mesh-fine (bottom) cases showed well-stabilized flames near the cavity even without the air throttle.
The combustion efficiency, defined as a fraction of CO2 over the sum of CO and CO2, is estimated. The respective mole fractions were measured via the tunable diode laser absorption spectroscopy (TDLAS) method.
Compared to the blank insert (black), the mesh insert cases exhibited a comparable combustion efficiency at half the equivalence ratio. At a comparable equivalence ratio (at around 0.2), the mesh inserts exceeded the blank case by nearly 40% in combustion efficiency.
This shows the hybrid method (cavity + intrusive structure) can significantly improve the combustor performance.
Compacting a combustor is not a trivial task. If one simply downsizes the combustor, the residence time will decrease by the geometric scale while the combustion scales (such as ignition time delay) remain comparable. Therefore, compact combustors necessitate a dedicated flameholding scheme for the given operating conditions.
A trapped-vortex combustor (TVC) is one such approach. This aims to enhance fuel-air mixing via vortex flow structures created from different streams. In principle, the enhanced mixing and increased residence time inside the vortex flow can significantly improve the performance of compact combustors.
To verify the vortex structures, particle image velocimetry (PIV) was performed to estimate the velocity field. This technique requires seeding particles (TiO2 in this study) and a double-pulsed laser sheet with a short pulse separation to visualize the particles' movement in sequence. By tracking particles' displacement in a given period, the velocity components can be estimated in each interrogation window.
With the PIV technique, the vortex structure of the TVC combustor can be assessed across different operating pressures and mass ratios of the streams.
By investigating and optimizing the vortex strength, the TVC combustor can deliver the maximum residence time and mixing performance required for compacting the combustors.
Propulsion systems often prefer liquid fuels due to their superior energy densities. Since liquid fuels must be vaporized for the combustion processes, improving the liquid vaporization, i.e., spray atomization, is critical for any liquid-fueled combustors.
To deliver a reference performance, DEVCOM Army Research Laboratory (ARL) developed a referee combustor M1 for comprehensive liquid-fuel combustor studies.
Spray atomization, or liquid break-up, is a chaotic process in nature. Consequently, an imaging technique must deliver high-speed and high-resolution images of liquid particles to allow such studies.
One of the options is to employ X-ray phase contrast imaging (PCI), which provides strong contrast between the gas phase and the liquid phase at greater frame rates than conventional CMOS cameras.
Several machine learning techniques were employed for PCI images.
A self-learned denoising algorithm.
Boundary detection algorithm for assessing the particle size.
Unsupervised velocity prediction algorithm.
The PCI images along with the ML techniques allowed simultaneous sampling of several quantitative parameters of spray flows:
Size distribution along the longitudinal direction from which the spray break-up length is inferred.
Size distribution along the radial direction, which indicates the atomizer performance.
Shape distribution of particles.
Velocity distribution of particles.