Flame-propagation in inhomogeneous mixtures resulting from direct injection – large-eddy-simulation, model validation and analysis

点击翻译按钮获取中文摘要

In this project, detailed simulations of flow, mixture formation and combustion in the hydrogen engine with direct injection are to be carried out. These large-eddy simulations (LES) take into account the unsteady turbulence fields, the mixture inhomogeneities and the propagation of the strongly convoluted flame with high spatial and temporal resolution, so that many successive cycles can be simulated, which differ from each other by cyclic variations. The results of the simulations are first checked (validated) for correctness using experimental data and then analysed. A backward analysis is used which, with various statistical methods and logged particle trajectories, makes it possible to identify special features in early flow and mixing fields which lead to particularly strong or particularly weak flame propagation at the (later) ignition time and thereafter. In this context, the particularities of hydrogen engines have to be taken into account, especially (i) the strong influence of the (supersonic) direct injection on the flow field and (ii) the effect of the thermodiffusive instability, which stretches, folds and thus enlarges the flame (surface area), leading to a strongly accel-erated combustion. This project aims to answer the following questions: a) How do cyclic fluctuations occur in the engine during hydrogen direct injection and how can they be efficiently predicted? b) What effects does the thermodiffusive instability of hydrogen combustion have on the cyclic fluctuations and how can it be modelled? c) What is the effect of fuel (and flue gas) inhomogeneities in the combustion chamber? The simulation uses detailed boundary conditions for hydrogen injection, reaction and flame propagation models, as well as experimental validation data. Conversely, the simulation allows the verification, analysis and interpretation of the data meas-ured in the experiment, the testing of the models as well as the tracing back into the injection jet and the injector. The experiment provides fast and cost-efficient measurement points from a large number of cycles, whereas the simulation is limited to fewer cycles and is subject to model assumptions, but allows detailed access to all states at any time and at any location in the combustion chamber. By combining experiment and simulation, relevant, validated and complete data sets are created as a basis for the detailed backward analysis of the chains of effects that lead to cyclic fluctuations in the hydrogen engine with internal mixture formation.