No abstract

For scramjet applications at low flight Mach numbers, decreased autoignition temperatures and shortened ignition delay times are essential. The possibility to influence these parameters with preseeded radicals is evaluated numerically using Cantera and a reduced GRI3.0 reaction scheme. It is found that single-digit ppm concentrations of OH-radicals can significantly reduce ignition delay. Combustion experiments are conducted to evaluate the possibility of producing radicals by catalytic fuel pre-treatment in hydrogen-air mixtures. While fuel equivalence ratios below Φ = 11 show an inhibiting effect due to increased water production, very high equivalence ratios show ignition promotion. Possibilities to further increase the ignition enhancement are discussed. Nomenclature Φ [-] equivalence ratio k r [mol, l, s] reaction rate constant A r [mol, l, s] pre-exponential factor T [K] temperature n r [-] factor for extended Arrhenius law E r [cal/mol] activation energy R m [J/(mol K)] ideal gas constant ∆t ign,X [ms] ignition delay time at temperature X p reactor [kPa]overall reactor pressure (absolute) T ignition [K] temperature measured at the reactor exit

A major reason for the long simulation runtimes of large eddy simulations of wall bounded flows is the necessary spatial resolution of the turbulent boundary layer in wall parallel directions. This limitation can be reduced by a hybridization of RANS and LES, whereas RANS is utilized for near-wall regions. In the context of hypersonic airbreathing engines it is necessary that such a method correctly reproduces shock positions within an engine duct. This work presents a generic RANS/LES hybridization approach based on the k-ω-SST RANS model utilizing a wall function approach and a mixed model for the LES mode. The method rests upon a hard RANS/LES interface, the location of which is determined by a comparison of computed RANS and LES model quantities. It is shown that the favorable behavior of the k-ω-SST model to predict shock reflection is retained, while maintaining a LES behavior in the core of the flow. The method is applied to two supersonic test cases:The first case is a non-reacting supersonic flow through a model scramjet combustion chamber. The second case is a supersonic injection experiment with perpendicular carbon dioxide injection into a Mach 1.9 crossflow. RANS and hybrid RANS/LES computations are conducted, compared to each other and to experimental results. Nomenclature CModel constant e energy k turbulent kinetic energy l lenght scale p pressure P r Prandtl number r turbulence mode indicator Sc Schmidt number x, y position µ dynamic viscosity Subscripts k kinetic t turbulent