Peter Gerlinger

Scalar and joint scalar-velocity-frequency Monte Carlo PDF simulation of supersonic combustion

Approaches based on probability density functions (PDF) are a natural choice for the simulation of high-speed and supersonic turbulent reacting flows due to their ability to represent chemical sources in closed form. This paper aims in particular at the application of scalar and joint scalar-velocity-turbulent frequency PDF methods to supersonic combustion. Because transport equation PDF simulations of supersonic combustion are still rare, emphasis is placed on the peculiarities of this kind of flow. The paper reports detailed results for two supersonic hydrogen-air non-premixed flames and makes comparison with available experimental data. A reaction scheme consisting of seven elementary reactions is considered, which provides an adequate description of chemical kinetics. As mixing models still are the weakest point of transport equation PDF methods, the three most frequently applied formulations have been investigated. Only minor differences are observed for the supersonic test cases. Moreover, these simulations represent the first published application of joint scalar-velocity-turbulent frequency PDF calculations to supersonic flows with complex geometry and hydrogen chemistry. It is shown that significant improvements are achieved by the joint PDF approach for one of two high speed reacting flows investigated. Results are somewhat unconclusive in the other case, partly due to uncertainties in the experimental setup.

Numerical Simulations of Confined, Turbulent, Lean, Premixed Flames Using a Detailed Chemistry Combustion Model

In this paper numerical simulations of a confined, high strained jet flame employing a detailed chemistry combustion model are presented. Unlike other configurations available in literature, the geometry under investigation presents the jet axis shifted one side of the confining chamber in order to get non-symmetric recirculation zones and a flame stabilization mechanism based on the recirculation of a high percentage of hot combustion products. Fully three-dimensional unsteady simulations are carried out with finite-rate chemistry effects included by means of a detailed reaction scheme. Turbulence-chemistry interaction is taken into account by employing a presumed PDF approach, which is able to close species source terms by solving two additional transport equations. The use of the hybrid RANS/LES SST-SAS turbulence model is able to include large unsteady turbulent structures according to the local grid size and flow conditions. The approach presented here allows an in-depth investigation of flame stabilization mechanisms, ignition phenomena and influence of recirculation regions on flame stability. Additional simulations adopting simpler combustion models (i.e. Eddy-dissipation Concept) are also presented in order to assess the prediction capabilities of methods widely used in design environments. The paper also includes experimental data while comparison in terms of radial profiles at different heights above the burner are provided.Copyright

An Improved Lobed Strut Injector Concept for Supersonic Combustion

A major problem in supersonic combustion is the short residence time of the fluid inside the combustor due to high flow velocities. Thus techniques for mixing enhancement have to be used to achieve a fast and efficient fuel-air mixing. In the present paper, different types of strut fuel injectors are investigated numerically. All strut injectors use lobed structures to create streamwise vorticity to achieve a faster mixing. A new concept for a single wedge lobed strut injector (SWLS) will be presented which is compared with an older double wedge strut injector concept. It will be shown, that the SWLS injector improves mixing and combustion efficiency significantly. Additionally the new injector is able to achieve a complete combustion at higher equivalence ratios.

Numerical Investigation of the HyShot Supersonic Combustion Configuration

A multivariate assumed PDF approach together with finite-rate chemistry is used for the simulation of the HyShot supersonic combustion configuration. Because the combustor entrance conditions of scramjets at low flight Mach numbers (Ma 8) are close to the ignition limit of hydrogen air mixtures, detailed kinetic schemes are required and an accurate simulation of temperature and temperature fluctuations is essential. In the present paper, experiments of the HyShot supersonic combustion configuration performed at the High

Modelling of soot and NOx in a full scale turbine engine combustor with detailed chemistry

A numerical simulation of spray atomization, combustion and soot formation in a full scale turbine engine combustor is presented and discussed in this work. Turbulence has been captured by a two equation turbulence model, chemistry by a detailed kinetic mechanism and turbulence chemistry interactions have been accounted for by an APDF-approach (assumed probability density function approach). For the kerosene fuel a surrogate has been used, consisting of n-decane, isooctane and toluene. The injection of the liquid fuel spray has been included by coupling the gas field CFD code with a spray code. Apart from the n-decane and isooctane reaction paths, the chemical kinetic reaction mechanism accounts for species as heavy as toluene. Polycyclic aromatic hydrocarbons heavier than toluene are represented by a sectional approach, while soot is calculated via soot volume fraction and particle number density. Main flow features are investigated and good agreement with the experimental measurements can be reported.Copyright

Numerical Simulation of the Internal and External Flowfields of a Scramjet Fuel Strut Injector Including Conjugate Heat Transfer

The use of fuel strut injectors is a promising concept to achieve good mixing and combustion performances in supersonic combustors. In the present paper a lobed strut injector is investigated. The chosen geometry creates streamwise vorticity to enhance the mixing of fuel and air compared to a planar strut. The injector is located in the center of the hot supersonic flow field at the entrance of the combustor. To avoid burning, the strut has to be cooled which is done by the injected fuel. Especially at high flow Mach numbers, the thermal load for the strut may be severe. In the present paper, the strut surface heat fluxes and the temperature distribution in the strut are calculated. Conjugate heat transfer to the strut surface is considered both, from the simulated external air flow to the strut and from the strut to the cold internal hydrogen flow. An all-Mach number preconditioning is used in the numerical code to compute supersonic as well as subsonic flows. The strut investigated is assumed to be out of copper. The coupled simulations provide a survey of the internal strut temperature distribution and delivers accurate boundary and injection conditions (for hydrogen) in cases, where the interior fuel flow and heat transfer are neglected.

Numerical Simulations of Soot and NOx Distributions in a Full Scale Aero-Engine Combustor at Two Different Flight Altitudes

Numerical simulations of a full-scale aero-engine combustor at two different flight altitudes are presented. Turbulence is treated by the SST model. A finite-rate chemistry model, where a separate transport equation is solved for each species, is applied in order to treat combustion accurately. Chemistry-turbulence interactions are modeled by an assumed probability density function model (APDF). A sectional model is applied for polycyclic aromatic hydrocarbons, while soot is modeled by a highly efficient two-equation model. The liquid phase of the fuel is calculated by a Lagrangian spray code which is coupled to the CFD code. Comparatively good agreement between experiment and simulation is observed at the reference operating point. At a reduced flight altitude, lower temperatures and lower NOx emissions are predicted, which result from a higher air-fuel ratio. Soot emissions at the combustor exit are due to cold regions in the vicinity of the combustor walls.

Soot Modeling in a Turbulent Unconfined C2H4/Air Jet Flame

A turbulent unconfined C2H4/air jet flame has recently been investigated using contactfree state of the art laser measurement techniques. Aiming to reproduce the experiment, calculations have now been performed with the DLR in-house code THETA using a two equation turbulence model, detailed gas phase chemistry, and a recently published soot model. Heat radiation is accounted for by a heat sink approach assuming an optically thin medium. The influence of turbulent temperature fluctuations on heat radiation as well as turbulence-chemistry interactions have been included by an APDF-approach. Good agreement between experimental data and the simulation can be reported. In addition, calculated soot size distributions and the mean particle diameter field give important insight to the evolution of soot.

Numerical Investigation of Soot Formation in Laminar Ethylene-Air Diffusion Flames

In this work a new soot formation model is used to predict temperature, species and soot concentrations in laminar ethylene-air diffusion flames. The gas-phase chemistry is described by elementary reactions with transport equations solved for any species. The chemical paths yielding to soot are modeled by a sectional approach for Polycyclic Aromatic Hydrocarbons (PAHs). Soot dynamics is described by a two-equation model for soot mass fraction and particle number density. Phenomena like nucleation, growth and oxidation have been included both for PAHs and soot. Moreover, PAH-PAH and PAH-soot collisions are taken into account. Species, PAH and soot transport equations are implemented in the in-house DLR-THETA CFD code. The laminar, ethylene-air diffusion flame investigated experimentally by McEnally and coworkers (2000) is simulated in order to validate the model. An analysis of the main flame’s features as well as the interaction between them and the soot chemistry will be given. A qualitative correlation between local stoichiometric values and soot formation rate is assessed. In order to study the sensitivity of the combustion model to simulation parameters like the inlet temperature and kinetic mechanism, additional simulations are performed. Results are also compared with experimental data in terms of temperature, species mole fractions and soot volume fraction axial profiles.Copyright

Investigation of Hybrid RANS/LES Approaches for Compressible High Speed Flows

In the present paper two different hybrid RANS/LES approaches are investigated to predict compressible high speed flows the DDES (Delayed Detached Eddy Simulation) and the SAS approach (Scale Adaptive Simulation). All simulations are performed with the scientific code TASCOM3D (Turbulent All Speed Combustion Multigrid Solver). For the reconstruction of the interface values of the finite-volume scheme, a fourth/fifth-order upwind biased scheme combined with an improved multi-dimensional limiting process (MLP) is employed. The inviscid fluxes are calculated using the AUSM−up flux vector splitting. In order to prove the capability of the solver to perform LES, a fully developed turbulent channel flow is simulated and compared with experimental data and results obtained by Direct Numerical Simulation (DNS). Thereby discretization schemes of different accuracies and effects of a shock detection sensor are investigated. To compare the capabilities and limits of DDES and SAS in compressible high speed flows a supersonic turbulent mixing layer is investigated. The variation of modeling parameters, grid spacing, and domain widths in spanwise direction are used to asses the different turbulence modeling techniques.