Dipl.-Ing. R. Ambs, M.Sc. W. Calver, Dr.-Ing. M. Matthes, Dr.-Ing. S. Cucuz
Visteon Corporation, Kerpen/Germany und Plymouth/Michigan, USA
In this paper a method will be presented that enables a detailed transient analysis of the thermally coupled airflow through the vehicle compartment. Focus is set on the prediction of time-dependent passenger thermal sensation and on the prediction of defrost and demist performance. The geometry model of the vehicle interior is based on a generic parametric cabin model, which needs to be adapted to the investigated car. Due to this approach the method can be used in the preprototype stage before detailed CAD data is available. The generation of the unstructured tetrahedral mesh is fully automatic. A normally time consuming semi-automatic meshing of the boundary surfaces does not need to be carried out. The heart of the simulation process is a highly efficient finite element solver for the transient CFD analysis. It is based on the Galerkin/least-squares (GLS) finite element formulation of the Navier-Stokes equations. For transient analyses a second order time discretization scheme is used and the time step is automatically controlled. Natural convection is taken into account by employing the Boussinesq approximation. For an adequate consideration of the important heat transfer through the different wall surfaces thermal shells are added to the boundaries. The shells can be subdivided into several layers with different material properties. In A/C conditions the heat load by solar irradiation is of special importance and is included by the use of a ray tracing algorithm. Since temperature differences inside the cabin can reach 40 °C and more in cooldown conditions as well as in warmup conditions, the inclusion of radiation heat transfer becomes necessary. Hence the gray body enclosure radiation equation is added to the system of equations. Cabin humidity is accounted for by calculating the scalar transport of water vapor. Once the CFD analysis has been completed, the objective results are used by a special comfort model to calculate subjective ratings of thermal sensation for each passenger. The comfort model takes into account temperature, velocity, solar radiation, humidity, type of clothing and activity of the occupants. To determine detailed time-dependent passenger thermal comfort virtual manikins can be added to the cabin model. For the prediction of the defrost performance a layer of ice is added on top of the glass surfaces and the phase change of melting ice is captured with a latent heat model. Demist predictions are made by computing the local saturation of air on glass surfaces.