Learning Forum

[Salix Bair]

I am running a 2D axis-symmetric model and am interested in looking at particle trajectories in the fluid. The particle trajectories are crossing the axis, my understanding is that for this type of space, particles shouldn't cross the axis, but rather be reflected about it. I am not sure what is causing this issue, any insight or help is greatly appreciated.

[Rob]

That's unusual. Usually, if particles leave the domain they just keep going in the same direction. Check it is 2d-axi, and that you're using high-res tracking. How does the velocity contour & vector field look? From scale what is the minimum y coordinate. Please also zoom in on the angled edge - the mesh looks a little odd from your picture. 

[Salix Bair]

 

I am using 2d Axis symmetric and high-resolution particle tracking. I am including my mesh at the angled edge, velocity contour, vector field, and another set of particle tracks. The minimum y coordinate is ~1mm below the axis, my geometry has a maximum height of 5mm. 

[Rob]

It's the DPM tracks getting very confused by the mesh. I have no idea how you've created that, but please review the meshing training and do it again. The jump in cell size is going to cause problems, and the aspect ratio will add to those problems.

[Salix Bair]

I have re-meshed my geometry, I have 5 bodies as 1 part. The model is set up as 2-D axis-symmetric, mass-flow inlet and pressure outlet, running for 200,000 iterations, however, the convergence criterion is reached long before. Even with the updated mesh, the particle trajectories are behaving the same as before. I even tried, starting a new workbench file, in case my original had been corrupted in some way. At this point I have no idea what the problem is. 

[Rob]

What does the velocity field look like? To get particles bouncing like that takes something in the flow field or mesh. 

[Salix Bair]

Here are my velocity field as well as the velocity vectors: 

[Rob]

Can you zoom in on the interesting bit? 80m/s is borderline compressible for ambient air too. 

[zachary.gauvin]

Has this problem been resolved? I am having a similar issue and would love some insight on how to fix it.

[Rob]

The OP didn't reply so not sure. If you post images of the mesh and particle tracks it'll help. 

[zachary.gauvin]

unfortunately I am not able to post images of the mesh/geometry, but it is the exact same issue as the OP with particles crossing the symmetry axis

[Rob]

Check the mesh quality and also the trajectories near the region where the particles escape the mesh bounds.  

[zachary.gauvin]

 

Target size of the mesh is 0.020 in. and there are two different inflation layers added. One starts with 1e-4 in. first cell height with 15 layers and a growth rate of 1.5. The other starts with 2e-5 first cell height with 14 layers at 1.7 growth rate. There is no inflation applied on the axis that is the boundary of the axisymmetric mesh (which is where the particles are crossing the bounds). The Min orthogonal quality is 1.06542e-2 and the max aspect ratio is 1.68877e3. Any of this stand out to you? Also to add, the boundary layers with the inflation are the only areas that have the low quality. Thanks

 

[Rob]

Min ortho quality isn't great, nor is the aspect ratio depending on where those cells are. 

[zachary.gauvin]

yeah unfortunately I need the very small first element sizes to keep the Y+ values at the wall as low as possible. The areas with high aspect ratio and low quality are along the walls, which are not where the particles are crossing the domain, that is occuring at the center line axis

[Salix Bair]

Hi, I actually believe it is an issue with the version of ANSYS Fluent. I thoroughly troubleshot my mesh, model, and setup and found that no matter what the DPM tracker was tracking particle trajectories across the axis. I have access to an older version of the software (R2021) and when I ran my model with no changes to mesh or set up in the older version, the DPM tracker correctly tracked particles. 

[zachary.gauvin]

what version were you orginally using? I am currently using Fluent 19.2 and experiencing the issue

[Salix Bair]

I was originally using 2022. 

[zachary.gauvin]

I turned on the turbulent dispersion and this seems to have fixed my issues and no particles are crossing the centerline axis anymore, this may help someone else running into a similar issue