A Phenomenological Homogenization Model Considering Direct Fuel Injection and EGR for SI Engines 2020-01-0576
As a consequence of reduced fuel consumption, direct injection gasoline engines have already prevailed against port fuel injection. However, in-cylinder fuel homogenization strongly depends on charge motion and injection strategies and can be challenging due to the reduced available time for mixture formation. An insufficient homogenization has generally a negative impact on the combustion and therefore also on efficiency and emissions.
In order to reach the targets of the intensified CO2 legislation, further increase in efficiency of SI engines is essential. In this connection, 0D/1D simulation is a fundamental tool due to its applica-tion area in an early stage of development and its relatively low computational costs. Certainly, inhomogeneities are still not considered in quasi dimensional combustion models because the prediction of mixture formation is not included in the state of the art 0D/1D simulation. Therefore, a phenomenological homogenization model has been developed in this work based on 3D CFD simulations.
The model is based on a finite volume approximation of the 2D convection diffusion equation rep-resenting the in-cylinder flow field. The velocity field is modelled as a Taylor Green vortex accord-ing to an existing charge motion and turbulence model. Turbulence effect and the influence of piston motion on homogenization are also considered. The latter one demands the solution of a moving boundary flow problem that is described by means of immersed boundary method. As a result, the model is able to predict the progress of fuel homogenization as well as residual gas distribution, accounting for different cam profiles, injection strategies, engine loads and speeds. Furthermore, a characteristic number representing the level of inhomogeneities is linked to the laminar flame speed of a quasi-dimensional combustion model to predict the impact of inhomogeneities on the burn rate. The predictive capability was validated with measurement data of two different engines.
Sebastian Fritsch, Michael Grill, Michael Bargende, Oliver Dingel
University of Stuttgart, FKFS, IAV GmbH