As conclusion, we can say that our model is not a correct one. This is firstly due to problems of definition of the 3D mesh. Some boundary conditions could not be used and the solution of the porous outlet could not be applied.
The other problem is a lack of data. Although it was quite easy to find information on aluminium or on glass, the evaluation of the physical parameters of the seats was impossible without specific documentation, which is not officially available and kept secret by the automotive constructors.
The greatest problem concerned the outlet and the losses of air in the cabin. Although we could find an approximative solution with the 2D problem, it was not accurate and not physical. The losses of fluid need thus a much better modelization.
Finally, we could not compare our numerical results with other official real simulations
This work show the difficulty to modelize a phenomenon which is not an academic case and to create a complex mesh.