How to define Neuman BC for N-S Solver

[Moritz]

Hallo,

I want to set Neuman BCs for the velocities on the outer boundary of the mesh in the N-S solver to mimic an infinite volume with the same gas and pressure as inside the mesh.
Is setting

Flow Force BC = True
pressure 1 = 0
pressure 2 = 0
Slip Coefficient 2 = 0
Slip Coefficient 2 = 0

the right one for this intention?

Thanks in advance

Moritz Nadler

[raback]

Hi Moritz

I think that these definitions will not contribute to the linear system i.e. they result to the natural boundary condition with vanishing normal traction. This is probably what you also want so you don't actually need to define anything. Depending on the configuration you might in addition fix the tangential velocity component to zero. The main problem with this kind of outlet boundary condition is that with transient problems the pressure will reflect from the boundary. For this there is, unfortunately, no simple remedy.

-Peter

[Moritz]

So this would be ok?

Boundary Condition 4
Target Boundaries = 5
Flow Force BC = True
Temperature = 293
End
Or should I even delete the "Flow Force BC = True" statement?

Hi Moritz
The main problem with this kind of outlet boundary condition is that with transient problems the pressure will reflect from the boundary.
-Peter

Is this still a problem if I am only interested in steady state solutions?

Thanks for your quick reply!

[raback]

Nothing is fine. For s-s you should have no problems with the reflecting pressure pulses either.

However, if you have alternating velocity direction on the boundary defining the temperature may lead to minor problems if you have outflow. You could just define the temperature in case of inflow. For this you could use the following syntax:

Code: Select all

  Temperature = 293
  Temperature Condition = Variable Velocity 1
    Real MATC "tx"

This would set the temperature only for v_x>0 i.e. it would be good for the left wall. Use different sign for the condition for the right wall, and similar logic for other walls.

-Peter

[N. Takashi]

Hi Moritz and Elmer team!

I want to simulate the gravity field around the walls,
for that, I must take Neuman conditions of pressure;

dp/dz=rho*g, eq (1)

where p is pressure, z is the vertical direction axis, rho is density and g is gravity acceleration.

How shoul I do to set eq (1) in Sif of Elmer?
If you know their solutions, please let me know.

Thank you,
Takashi

[raback]

Hi, Gravity affects primarily as a body force. So you might rather take it into account by setting "Hyrdostatic Pressure = True" in bf section and defining gravity in the Constants section. -Peter