Geometric scaling studies of an inlet-fuelled axisymmetric radical farming scramjet engine

Abstract

Scramjet engine technology has the potential to transform commercial flight and mankinds’ability to access space. Major technology hurdles remain which must be investigated and re-solved before the potential of the scramjet engine can be realised. The ability to scale a smallgeometric scale scientific test model to a large practical vehicle size with similar performancecharacteristics is one such challenge. In this study, the small scale geometric scaling character-istics of a scramjet engine, one such tool to aid in future experimental based investigations andoverall scramjet development, is presented. The study specifically concentrates on investigatingthe similarity effects for a new type of scramjet engine, the axisymmetric inlet-fuelled radicalfarming scramjet engine, and is concerned with the Pressure-Length scaling laws in both plainair mediums and hypersonic shock tunnel facility test mediums. The Pressure-Length scaling laws are a type of scaling law in which all inflow parametersbetween a base model and a small size model are maintained with the exception of pressure.In the PL method, pressure is inversely proportional to length scale (i.e. PL = constant), thusmaintaining both Reynolds number and Damkohler number based on ignition delay timescales.In the PnL method, PnL is maintained constant between the two models where n is selected sothat the Damkohler number based on three-body heat releasing timescales remains identical. This study employs an experimental and numerical approach. Experiments are performedin the HEG and HIEST free-piston reflected shock tunnel facilities in order to generate a datasetfor an axisymmetric inlet-fuelled radical farming scramjet engine. CFD reconstructions of theexperimental flowfields are used to validate and gain confidence in the approach of the nu-merical study in addition to developing a comprehensive understanding of the flow physics,chemistry and their coupling. The numerical study is performed to conduct the small scaledscaling studies. Supersonic combustion is achieved over four nominal dynamic pressures of 27.9kPa, 48.1kPa,55kPa and 75.5kPa and equivalence ratios ranging 0.0 – 0.98. Wall cp and Stanton number pro-files are found to be consistent with a radical farming scramjet engine. The computational fluiddynamics (CFD) solver CFD++ is employed to reconstruct the experimental flowfields. TheCFD reconstructions illustrate the complex coupled fluid dynamic – chemistry interactions inthe engine. These flowfields confirm the radical farming nature of the engine and illustrate thecomplex combustion processes occurring within the radical farm. In addition the simulationestablishes that the initial formation of radicals and H2O occurs in the inlet and the existence ofa counter rotating vortex pair, which is crucial in the distribution/mixing of fuel in the engineand combustion characteristics. The CFD solver is validated with experimental data. The plain air small scaled length scaling study demonstrates that quasi-scaling, scaling inwhich identical flowfields and all non-dimensional numbers are not achieved but strong similari-ties are present, is achievable under premixed air-fuel inflow conditions, within length scale lim-its (0.5 length scale), for both the PL approach and the PnL approach. In the three-dimensionalstudy which includes the effects of discrete fuel injection and complex air-fuel mixing phe-nomena, quasi-scaling is only found to be achievable with the PL approach. Although the PLapproach could produce the correct general flowfield structures in the flow including the lo-cation of ignition, excessive heat release due to increased Damkohler numbers based on heatreleasing timescales is evident through the upstream shift of flowfield structures from the pointof heat release, higher normalised pressures and enhanced engine performance. The failure ofthe PnL approach is attributed to the extended ignition delay length scales and the location ofthe point of ignition in the base scale flowfield.A contaminated air small scaled length scaling study is also performed in order to examinethe effects of common shock tunnel contaminants on small size engine scaling. This studydemonstrates that quasi-scaling is achievable under both premixed conditions and discrete fuelinjection only with the PnL approach. The action of atomic oxygen reduces the ignition delaytimescales and the relative location of ignition in the base scale simulation. The offset observeddue to the extended ignition delay length scales in the plain air simulations is reduced due tothe action of atomic oxygen. This allows for three-body heat releasing reaction to initiate in asimilar normalised pressure and temperature environment and scaled heat release is achieved.Normalised pressures to within 5% and normalised shock structure locations within 2.5% of thebase simulation downstream of the radical farm are achieved in both the axisymmetric and three-dimensional simulations. The PnL approach is also shown to have the capability of generatingin a small length scale engine with YO = 0.01% present a quasi-scaled flowfield sharing manyfeatures with a full scale model in a plain air medium. It is concluded that quasi-scaled heatrelease between the small scaled and base scale model is necessary to achieve a quasi-scaledflowfield.