Learning Forum

[tatimaes]

Hi everyone!

I am a beginner in Ansys (and CFD), I am doing my MSc. research and dealing now with something that seems inconsistent :

I am running VOF+Open channel submodel (free surface) / BC: Inlet/outlet: pressure in/out --- Top domain: pressure outlet --- wall: no-slip conditions (I am assigning a ks value)

Geometry: L=800m / width: 360m / total height 6m (I assigned values of free surface in/out in BC setting)

As shown in the screenshot, when checking fluxes - mass flow rate as one of my controls it seems weird that the top domain represents a negative value, which means water is leaving through the top domain (?) why is it happening?

*I am using Ansys student, I've been already playing with the mesh by surface trying to refine some areas of my domain without exceeding the limits of this license.

Thank you very much!

[Nikhil Narale]

Hello,

Can you try computing for the water phase? (Change the phase from mixture to water in the flux reports)
Let us know the results.

Nikhil

[tatimaes]

Hello,
oh yes of course, here it is, basically the same thing... (?)

[Nikhil Narale]

Hello,

Is the solution converged? From the values, it looks like there is mass imbalance.

Nikhil

[tatimaes]

Yes, it is converging (I think without problems). *for residuals criteria: 1e^-3
When solution converged and I plot contour the top surface is shown as 0% of volume fraction of water... so, I do not know the fluxes report as if it were leaving through it.
Does something occur to you about this problem?

[Rob]

There look to be a couple of red stripes on the image, could that be the reason? How large is the domain, and how well resolved is the mesh?

[tatimaes]

Those red ones are simulating two solid structures perpendicular to the flow, I set it up as a wall (with the specified value of ks)...In contours or vector plots it does not show as if the water were leaving through there, so seems to do its job.
My domain is: longitude= 800m / widht= 360m
regarding the mesh, I use the multizone - Hexa method and I am doing face sizing so each surface has a different value, to keep the limits of the student license some faces are coarser others finer, I am playing with these values but basically, I could get 11 vertical cells, refining in the lateral side where the structures are placed (orthogonal quality is good, the minimum value I got is 0.85). Is it what you mean about how well resolved is the mesh?
Thank you!

[Rob]

I suspect the liquid is diffusing out the top. Increase the space above the water and make sure you're resolving the free surface.

[tatimaes]

Thank you, I've worked on it. I extend my geometry to have a higher free board and I had to do the same with my inlet. The problem of water "escaping" through the top domain is gone.
Can you please tell me what do you mean by 'resolving the free surface'? Thank you

[Rob]

The free surface (where you switch from gas to liquid) can require a more refined mesh than the bulk of the domain. If the mesh is too coarse the interface will become diffuse and solver accuracy is not as good as it should be.

[tatimaes]

Hello,
well, I was just trying to refine it from - adaptive mesh - (first time), and when I define it for a cell register on this area of the -ISO- surface, it crash and hit me out from fluent. My first guess is that probably this is about the limits of the student license (?).
How can I do this refinement in this free surface area? is that the way -adaptive mesh-? or I am looking in the wrong direction?
Thank you very much

[Rob]

Adaptive mesh is correct. In 2021Rx and onwards use the predefined VOF criterion but limit the number of adaption levels to 2: 6 is overkill here. Can't confirm on the error, but with only 512k cells available you can very quickly exceed this with adaption.