[ralphn]

Hi everyone,

I am trying to simulate something I read from a paper. The paper uses a 2D axisymmetric simulation of a droplet breaking up in an airflow, with velocity and gravity acting in the +x-direction. The bottom axis (with the half-droplet being adjacent to that surface) having an axis boundary condition, and the top axis being a symmetry boundary condition. Is the symmetry boundary condition possible in that way? I can't seem to get my head around it, what the author is trying to do is not have a wall there as if affects the breakup of the droplet. I tried the part with symmetry and I was getting issues with my residuals, does anyone know what it is?

[CFDKareem]

Originally Posted by ralphn

Hi everyone,

I am trying to simulate something I read from a paper. The paper uses a 2D axisymmetric simulation of a droplet breaking up in an airflow, with velocity and gravity acting in the +x-direction. The bottom axis (with the half-droplet being adjacent to that surface) having an axis boundary condition, and the top axis being a symmetry boundary condition. Is the symmetry boundary condition possible in that way? I can't seem to get my head around it, what the author is trying to do is not have a wall there as if affects the breakup of the droplet. I tried the part with symmetry and I was getting issues with my residuals, does anyone know what it is?

The symmetry condition is a bit of a misnomer, as it doesn't have to be a symmetric face in the domain. mathematically the symmetric condition assumes that all values have a gradient=0 at the interface. It is often used similar to a pressure=0 condition acting like an ambient condition.

If you can provide some pictures of the set-up I can get a better sense of what may be affecting your convergence.

[ralphn]

Hey Kareem,
Thanks for your reply. Here's a screenshot of what my domain looks like. I've been trying to replace the top boundary condition (here marked down as slip wall) with other BCs like symmetry (similar to what other papers have done) and pressure outlet.
I've also got an adaptive mesh running at the interface to make it easier to compute.

PS: I've been referring to some of your videos and they've been really helpful, keep it up!

[CFDKareem]

Originally Posted by ralphn

Hey Kareem,
Thanks for your reply. Here's a screenshot of what my domain looks like. I've been trying to replace the top boundary condition (here marked down as slip wall) with other BCs like symmetry (similar to what other papers have done) and pressure outlet.
I've also got an adaptive mesh running at the interface to make it easier to compute.

PS: I've been referring to some of your videos and they've been really helpful, keep it up!

A symmetry condition should be totally acceptable for this interface. I doubt that it is causing the convergence challenges.

Are you using VOF? If so, what is your timestep?

[ralphn]

I'm using a VOF scheme with a global Courant number of 0.5, so it's an adaptive time step. I'm going for really low time steps to start with, and they increase afterwards. I tried a global Courant number of 0.25 and get similar results.