II. SIMULATION SET


II.1. Description of the domain
 

The domain considered can be seen as a very simple injector.
 

Box extend :

x_coordinates : x(min) = 0 m               x(max) = 0.21 m
y_coordinates : y(min) = -0.05 m        y(max) = 0.05 m

Fuel injection pipe :
The fuel is injected about 0.06 m after the air injection.
 

 

II.2.  Meshing

It is possible to use three different grids; structured, unstructured or unstructured coarse.
But because of the important consequences of the fuel entrance in the injector, as we could see later and as we already can guess, the unstructured grid is the most adapted.
Indeed, as we can see in the picture below, mesh is very refined in these two regions, so phenomenons as vortices by example could be correctly detected.
 
 

Injector Mesh









A zoom of the grid allows us to confirm what we have just said :
 
 

Zoom Injector mesh





III.3. Initials Conditions

At the beginning ( t = 0 s ), there is no fuel in the room ( the blue region of the figure II.1.).
The initial velocity (in the domain) of the air is equal to 0 m
 

Let's see the correponding part of the code:



 

III.4. Method
 

The method used is the LES method because it may allows to take in count turbulent phenomena more exactly than  others models. It must be noticed that LES resolves Naver-Stokes equations.
It is instructive to make 2D LES  but it can never capture exactly these phenomena which are typically 3D.