G. Kuenne

Evaluation of the flame propagation within an SI engine using flame imaging and LES

This work shows experiments and simulations of the fired operation of a spark ignition engine with port-fuelled injection. The test rig considered is an optically accessible single cylinder engine specifically designed at TU Darmstadt for the detailed investigation of in-cylinder processes and model validation. The engine was operated under lean conditions using iso-octane as a substitute for gasoline. Experiments have been conducted to provide a sound database of the combustion process. A planar flame imaging technique has been applied within the swirl- and tumble-planes to provide statistical information on the combustion process to complement a pressure-based comparison between simulation and experiments. This data is then analysed and used to assess the large eddy simulation performed within this work. For the simulation, the engine code KIVA has been extended by the dynamically thickened flame model combined with chemistry reduction by means of pressure dependent tabulation. Sixty cycles have been simulated to perform a statistical evaluation. Based on a detailed comparison with the experimental data, a systematic study has been conducted to obtain insight into the most crucial modelling uncertainties.

A Numerical Study of the Flame Stabilization Mechanism Being Determined by Chemical Reaction Rates Submitted to Heat Transfer Processes

Abstract Large Eddy Simulations of a turbulent lean premixed stratified burner are conducted in order to determine the physical mechanisms that dominate the flame stabilization close to burner walls. The purpose of this work is both to provide insight into the underlying physics as well as to check whether the deficiencies found in previous simulations are related to an inappropriate heat transfer treatment. The simulation utilizes a three-dimensional detailed chemistry database in order to capture the chemical reaction rates based on local mixing and thermal conditions. The study is supplemented by very accurate wall temperature measurements to remove the large uncertainty revealed in the past for this configuration. The results obtained from the simulations are evaluated by means of a qualitative illustration of the different flame stabilizations and comparisons with experimental data.

High performance computing of the Darmstadt stratified burner by means of large eddy simulation and a joint ATF-FGM approach

Current trend in design and operation of industrial gas turbines or internal combustion engines implies using the lean-fuel and stratified conditions aiming at the reduction of the harmful emissions and efficiency improvement. This has led to an increasing use of computational methodology, which allows detailed insight into combustion physics and processes controlling the emission formation. In the present work, the Darmstadt stratified burner is investigated by means of Large Eddy Simulation, implemented into the in-house, finite-volume-based numerical code FASTEST. The code solves the incompressible, variable-density Navier–Stokes equations coupled with the species transport equations. It is parallelized via domain decomposition technique using message passing interface (MPI). The complex chemical mechanisms are described by tabulated detailed chemistry utilizing the Flamelet Generated Manifolds (FGM) approach combined with the Artificially Thickened Flame model (ATF). The results obtained are comparatively assessed along with the complementary measurements. In-depth analysis of the flow field is conducted based on numerical simulations. Further studies have been carried out with respect to grid resolution and scalability.

Large Eddy Simulation of Combustion Systems at Gas Turbine Conditions

Three different aspects of Large Eddy Simulation (LES) of combustion processes are covered in this chapter. All three are based on using tabulated chemistry models to cover the chemical reactions occurring in gas turbines. The chosen approaches are all based on flamelet models. One part of the work deals with the investigation of subgrid scale models using Transported Eulerian Monte Carlo Probability Density Function (PDF) methods. The chemistry is dealt with using non-premixed flamelets and premixed Flamelet Generated Manifolds (FGM). The second aspect covered is the extension of the FGM method towards premixed flames where unresolved subgrid flame structures need to be handled. Therefore, the FGM approach was coupled with the Artificially Thickened Flames (ATF) model. The third aspect of combustion LES discussed here deals with the inclusion of more detailed reaction kinetics in the FGM approach in order to better predict minor species like nitric oxides or carbon monoxide, which are important development goals in today’s gas turbine industry. As all three aspects discussed in this chapter are located on the smallest scales in a combustion system, either the flow or flame subgrid structures, they are closely related to each other.