Mass flow boundary condition

Renato.

Hi all,

I'm writing to ask about how to prescribe a mass flux boundary condition in an incompressible x transport flow solver based on FEM (Finite Element Method). First of all, let us consider that this boundary condition has been applied in only one node -- it's common when we can't easily model the injection point or it is in a smaller/larger scale than the mesh.

I know that we must respect the mass/momentum conservation of the "real injection" and in this sense 1 m/s with unitary mass fraction is different from applying 2 m/s with 0.5 mass fraction. We can't say if the momentum is being preserved or not since the mass of the mixture would depend on the mass fraction of each specie and the inertia would be different accordingly.

Other tricky problem is that we can't know, a priori, the normal injection area since the boundary condition will be applied in only one node with null area (the injection area will be "virtual"). Has anybody got this kind of problem? Am I missing anything or doing any wrong assumption?

Regards

Renato.