I have a simple 2D code (solving the NS equations) that works well with total conditions and flow direction + exit static pressure. I need to specify an inlet velocity profile and the only way I can see is to modify the boundary condition for the flow to work with inlet conditions + exit outflow.

Supposing I have a collocated variable arrangements, what I do is this: - specify everything at the inlet (velocity, pressure and density) - zero gradients at the exit (would mean setting vx(ni,j)=vx(ni-1,j) and vy(ni,j)=vy(ni-1,j)).

I can't manage to make the solver work. Is there anybody who can help me?

Thanks

Tiz

Hi Tiz

To give you some suggestions I would like to know te type of grid are you using, What type of flow is and general aspects about the simulation that you want to perform.

Best Regards

Hi thanks for replying first of all. It is a subsonic flow (around 0.5). I have both a grid with collocated variable arrangement and a staggered one. (both structured). The flow is viscous but this doesn't affect the boundary conditions.

I think this is all.

Tiz

I guess it is not possible to set both velocity and pressure at the inlet together and not set any pressure at the outlet (using basically an outflow ...). Anybody knows if this is the case?

I read somewhere that this is possible for incompressible flows? Anybody can give me some suggestions on this?

Thanks

Tiz

Hi Tiz,

For the Inflow case, You are trying to impose all the primitive quantities which looks like wrong. According to the sign of characteristic waves(for subsonic or Incompressible), one information must be taken from inside (adjacent cell.. usually pressure). In outlet, with the same idea , one quantity should be specified and remaining required information should be extrapolated from the interior.

Since all the characteristic waves are positive in supersonic, Your implementation may work for Supersonic flows.

Regards, Mubassh

Thanks Mubassh,

in fact, it does work in supersonic flows. But, would he be correct to specify pressure and velocity at inlet and assume zero gradients at outlet? This is the same approach used in incompressible flows in a number of commercial codes.

Tiz

For Compressible 2D flows, the characteristic wave speeds are u, u+c, u-c, u.. for subsonic flows u-c = -ve... so 3 informations are moving from L-right and one information is moving from R-Left. So You should take one information from the interior(for Inlet)... and also Only one information imposed for outlet. hope what you are doing at outlet is right

Regards, Mubassh