Elmer Discussion Forum

Hej,

I am trying to simulate the pressure waves propagating in some mm thick cavitiy. I defined Navier-Stokes solver on a body and a Pressure wave on a a boundary. The rest I defined a non-slip condition.

Pressure profile is shown on the .SIF file. It has a 25 Hz frequency.

To achieve stationary I already tried:

- Gradp Gradient = Logical True
- Refined the mesh.
- Bubbles = Logical True
- Decrease time step size.
- Decrease average and amplitude of the pressure wave deinfed on the boundary.
- Decrease relaxation factor.

There is no way I can overcome the transient, it has always crashed first. Nothing seems to work.

Thanks in advance,

/Xavi

Hej!

Something funny...

If a null velocity,

Velocity 1 = 0
Velocity 2 = 0

is set at the same boundary where the pressure profile is defined (Boundary 1);

the simulation perfectly converges.

I am not that much into N-S. However, since the frequency of the pressure profile is 25 Hz shouldn't this affect that much the results right?

I am thinking on estimating somehow the velcoity profile at the entrance so I do not apply the restrictive null velocity there and still have convergent results.

/Xavi

I attach a snap to show the results

PARTICLE DYNAMICS

Hej,

from the case explained above I tried to add some particles to check if there is actually a movement of gas/oil/dust... when having a 25Hz pressure wave on a boundary.

I have been able to run the Model 29: Particle Dynamic however I cannot understand how the particles are intially placed. I am using the "random box" however the particles are only located on one part of my model. The particles migrates till other parts of the model with time however I am not able to place them where I want them to be.

Could someone experienced on the Particle Dynamic Solver explain me how the "Initial Position" command work? I attach a snap showing the problem I have (paticles just on one side).