(Georgia Institute of Technology, 2019-06-20)
Williams, Aimee
The objective of this study is to understand the underlying mechanisms of autoignition of a polydisperse fuel spray. Understanding and predicting autoignition of fuel sprays is important to the design of modern gas turbine engines, especially in the interest of developing a flame-holder-less afterburner concept. In this system, liquid fuel is injected into a high temperature, flowing, vitiated air flow. Previous studies of fuel spray autoignition have suggested multiple mechanisms for a fuel spray to autoignite, including single droplet and droplet cloud ignition behavior. The majority of liquid-fueled autoignition studies have been parametric in nature and describe the overall effect of droplet size, equivalence ratio, turbulence intensity, etc. on ignition delay time but do not investigate the phenomena controlling the local behavior of autoignition kernel formation and growth. Autoignition studies of cold gaseous fuel jets in hot oxidizer cross flows have shown the importance of local mixture fraction. A test facility was developed that is capable of reproducing flow conditions in an aero-engine reheat combustor. Fuel is injected using a reproduction of a commercially available spray nozzle installed on an aerodynamically shaped body centered in the flow by three aerodynamic pylons. High speed chemiluminescence and UV PLIF were used to determine the dependence of the locations where autoignition kernels form, upon the flow temperature and velocity. Analysis of the scatter in the time-resolved ignition locations revealed the importance of temperature fluctuations in the vitiated flow. Specifically, the most upstream ignition locations likely correspond to the hottest and, therefore, most reactive fluid packets. The distribution of the fuel spray was found to affect the appearance of most upstream autoignition kernels. A near stationary (on average) flame was found to exist at high co-flow temperatures, being stabilized by autoignition as distinct kernels were formed upstream of the main flame region.
(Georgia Institute of Technology, 2017-01-05)
Tan, Zu Puayen
The jet-in-crossflow (JICF) fuel-injection technique is widely applied in modern jet-engine fuel-air mixers to provide rapid fuel atomization and mixing. However, the “Classical” JICF places large amounts of fuel into the initial jet/spray’s recirculation zone and the wall boundary-layer, both of which can risk flashback and fuel-coking on the wall, particularly for next-generation jet-engines that will operate at increasingly higher pressures and temperatures. Twin-Fluid (TF) JICF, where streams of air are co-injected with the fuel jet into the crossflow, is being considered as a way to mitigate the Classical-JICF’s shortcomings. However, the TF-JICF is a nascent fuel-injection technique that is not well understood, especially at the high operating pressures of jet-engines. This dissertation reports an experimental investigation of TF-JICF where liquid Jet-A fuel was co-injected with pressurized nitrogen into a crossflow of air. The developed fuel sprays were characterized using shadowgraphy. The fuel-to-crossflow momentum-flux ratios were varied from J=5-40, the air-nozzles pressure-drops were varied from dP=0-150% of crossflow pressure, and the crossflow Weber numbers were varied from Wecf=175-1050. These operating conditions allowed us to obtain a dataset that is both comparable with near-atmospheric studies of TF-JICF in the literature and applicable to jet-engines. The results show that TF-JICF can be classified into four spray-formation regimes (i.e., Classical-JICF, Air-Assist JICF, Airblast JICF and Airblast Spray-in-Crossflow), each containing a unique set of spray characteristics and mechanisms. In the Air-Assist regime that spans dP≈3-13%, the injected air formed a protective air-sheath around the initial fuel jet, which inhibited the development of Rayleigh-Taylor waves and surface-shearing (i.e., disturbances created by the crossflow), thus reducing the near-wall fuel concentrations. Applying higher levels of dP transitioned the spray into the Airblast JICF regime, where the intensified fuel-air impingement and shearing generated new disturbances on the jet. These generally caused the near-wall regions to become repopulated with fuel droplets (i.e., counter-productive towards mitigating flashback and wall-coking). When dP was higher than 100%, the jet became completely atomized by air prior to encountering the crossflow, producing an “Airblast Spray-in-Crossflow”. The resulting spray-plume’s penetration became related to the combination of the fuel and air’s momentum-fluxes, where increasing dP caused increasing separation between the spray-plume and test-channel wall. This reduces the near-wall fuel concentrations and is beneficial towards fuel-air mixer design, although the required levels of dP for this regime is likely too high for practical jet-engine operation.