Overview
The development of gasoline direct injection (GDI) engines is still in an early stage although, currently, there are considerable activities in this area. Different combustion strategies are under discussion, the wall-guided system, the air-guided system and the spray-guided system. Each of these concepts has its own demands on injector, spark-plug, combustion chamber design and in-cylinder flow pattern, and exhibits a different combustion characteristics with respect to fuel economy and emissions.
Therefore, evaluation and optimisation of all these concepts is a tedious task, requiring detailed knowledge about all the processes occurring in the combustion chamber. Consequently, CFD simulation could strongly support this development process. Modelling of the in-cylinder flow can already be done by modern engine CFD codes with considerable accuracy and, accordingly, is integrated tightly in the development process of most car manufacturers.
For an efficient support of the combustion process development, however, a predictive simulation of mixture preparation, heat release and pollutant formation would be desirable. This cannot yet be achieved by current CFD tools. Even the current spray models still suffer from some weaknesses, also with respect to numeric, thereby jeopardizing the validity of the calculations. One of these weaknesses is the problem of spray formation at the nozzle (which is, in the case of GDI, usually a swirl nozzle), i.e. the determination of the starting conditions of the spray simulation.
There are still many more questions open concerning the process of stratified combustion itself which is not yet completely understood. Hence, a detailed analysis of the combustion process has to proceed the 3D CFD modelling. This shall be achieved by experiments in a high pressure vessel and by direct numerical simulation (DNS) of combustion and of spray. The usage of DNS for model- development and validation for 3D engine codes is a very strong and innovative but, so far, not often employed approach.
Since, in a lot of cases, the physical questions refer to a microscopic level, where experimental validation would be very expensive if not impossible, and the eventual results would be very questionable. On the other hand, the relevant physics and chemistry on this scale are quite often believed to be understood as separate phenomena, (Navier-Stokes equations, detailed chemical kinetics), the
The main objective of the project is to develop a predictive simulation tool for the process of GDI mixture formation and combustion, to be obtained by the following single steps:
- Analysis and modelling of gasoline spray formation,
- Analysis and modelling of partially premixed combustion,
- Generation of an experimental database,
- Model implementation and final validation.
The work within the project was aiming to generate advanced simulation and diagnostic tools for the GDI engine development.
The focus has been on the characterisation of the process of stratified combustion by theoretical, numerical and diagnostic means, resulting in predictive numerical models to be used within standard CFD engine codes. These models are able to address the whole range of possible GDI combustion regimes from premixed combustion to the strongly stratified case including a diffusion controlled combustion regime.
Two alternative combustion models have been developed. The models are able to simulate the complete combustion phase, from ignition to emission prediction. The development has been supported by direct numerical simulation in order to attain a better understanding of the physically and chemically complex phenomena.
The generated database enabled the quantification of the effect of stratification on the global flame behaviour and the development of several modelling approaches. Validation has been done by experiments in an engine with optical access equipped with state-of-the-art diagnostics. High speed visualisation of the injection and combustion phase within a single engine cycle has been performed, thus enabling study of complex events which is impossible with conventional single shot images captured in different cycles. Since the whole combustion process depends crucially on the mixture prediction, also the spray model has been revisited, assisted by spray diagnostics. An out-of-nozzle flow model for swirl-injectors has been developed, validated by measurements of spray formation and combustion under GDI conditions.
With DNS the behaviour of evaporating droplets injected into turbulent flow has been investigated, which resulted in the developed of improved models for droplet-turbulence interaction. The respective sub-models have been coded by the model developers with a well-defined interface to basic CFD codes.
The work to implement and apply models in their in-house engine CFD codes was carried out by the participating automotive companies themselves. Also the final validation of the complete simulation tools was done by the automotive companies by comparing results from calculations to experiments using their own engines and operating conditions. All experimental data and DNS results have been compiled to valuable databases, for future mode
Funding
Results
Normal
0
false
false
false
MicrosoftInternetExplorer4
/* Style Definitions */
table.MsoNormalTable
{mso-style-name:"Normální tabulka";
mso-tstyle-rowband-size:0;
mso-tstyle-colband-size:0;
mso-style-noshow:yes;
mso-style-parent:"";
mso-padding-alt:0in 5.4pt 0in 5.4pt;
mso-para-margin:0in;
mso-para-margin-bottom:.0001pt;
mso-pagination:widow-orphan;
font-size:10.0pt;
font-family:"Times New Roman";
mso-ansi-language:#0400;
mso-fareast-language:#0400;
mso-bidi-language:#0400;}
The work within
the project was aiming to generate
advanced simulation and diagnostic tools for the GDI engine development. The
focus has been on the characterisation of the process of stratified combustion
by theoretical, numerical and diagnostic means, resulting in predictive
numerical models to be used within standard CFD engine codes. These models are
able to address the whole range of possible GDI combustion regimes from
premixed combustion to the strongly stratified case including a diffusion
controlled combustion regime. Two alternative combustion models have been
developed. The models are able to simulate the complete combustion phase, from
ignition to emission prediction. The development has been supported by direct numerical
simulation in order to attain a better understanding of the physically and
chemically complex phenomena. The generated database enabled the quantification
of the effect of stratificati