Hi all,

I am modeling heat transfer between solid and liquid with phase change. The geometry is quite simple. They are two cubic and i meshed it by CFX. When i run it in transisent, even though i set the time step very small (2e-10 second, total time for my simulation=1e-6 second), RMS is strongly fluctuation. RMS is below 1e-4 but it's up and down which create a row of color. Domain imbalance is very good (below 0.1%).

So, if the RMS is fluctuation but domain imbalance is good, the result is applicaple? Anybody know how to fix it??? Pls. advise me.

Thank in advance.

Well, for most of the transient analysis, it is only normal that the RMS would fluctuate for each substep in the iteration!

How good is your solution can only be verified by comparing it with a real case.

And you need to be really sure if 1e-4 does really give you a convergent result. If you have a monitor point of some kind when you run the steady state solution, you would get to know when a good amount of convergence is reached.

What RMS are you using? It could be due to the way you set up the problem too.

Thank you Omer and TB.

I set RMS=1e-4. My model is quite sensitive with mesh and tiem step. I tried to reduce mesh size and time step but it still oscillation.

Anybody know how to mesh by hex element? My current mesh is tet and prism then i think it's not good enough.

Thanks in advanced

If you are using CFX Mesh , you wouldn't get a Hex mesh.

Depends on what kinda meshing tool you would be putting to use.