Recent onset of several problems

I recently started having a number of problems with Caedium, which I have been using for several months now. I am using the latest version.

One problem is with streamlines going forward from a seed face coincident with an inlet face; they simply stay straight and do nothing, even after hundreds of iterations, whereas streamlines (in the same simulation) attached to inlet/outlet seed faces will move about, reflecting the changing fluid flow. On an outlet, I also have to have the streamlines’ direction set to “forward” in order to visualize them, even though “backward” (in order to see how they were getting to the outlet) seems like it should be the correct setting.

I have also had problems with velocity vector arrows pointing upstream from outlets, which seems backwards. I have verified that the outlets were, in fact, outlets.

Also, I recently added in a seed face and had to translate it to get it to the correct position. That worked fine, but I noticed that with subsequent updates of the simulation display, its position is translated a little bit in the same direction with every update. I have worked around this by translating it back in the direction from which it moved.

Most recently, after every simulation display update, the display has started reorienting the viewing angle of the simulation to an angle along the Y-axis, and it is greatly zoomed-in. It never used to do this. Usually, the display would just update the simulation, without moving it, regardless of how zoomed-in I was on the display or not, and regardless of the angle of view of the display. This makes it very difficult to compare simulation changes between updates because the instantaneous display update is not easy to see with the display zoomed in so much.

I have had these problems with different files, and trying different settings, rebooting, etc., but the problems remain.

Any ideas what might cause this? I am using this on a 2013 MacBook Pro that is fully updated.

Thanks very much in advance for any help you can offer. I am happy to share my files via DropBox.

No votes yet

Animation Transforms

You'll need to share your simulation for me to diagnose the streamline issues you are seeing. If your outlet is too close to a recirculation zone then flow will enter through it.

The geometry moving with simulation updates is most likely because you transformed the geometry when the simulation was not at the beginning of time. A transform applied at a simulation update point will animate (move) the geometry when you run the simulation.

Your view is likely also animated, i.e., you have applied a transform to the background (simulation).

In both cases you can confirm whether an animation transform is in place by selecting the geometry or sim and then in the Properties Panel select the Transform tab. If you see an entry your geometry/simulation is animated. To remove it select the entry, right-click, and select Delete.

If you don't want to reset a simulation but you want to transform some unrelated geometry, then you can:

  1. Toggle Home->Lock to decouple the simulation controls from control of the solver
  2. Click Home->Start to set the simulation time to the beginning, transform the geometry
  3. Toggle Home->Unlock to return to the original simulation time and re-enable solver control

Use incompressible solver to diagnose problems

Thanks for sharing your simulation.

My previous comment is correct concerning the animation transforms.

While diagnosing problems I would advise against using streamlines and instead use arrows. Streamlines with the high Accuracy you've specified causes the simulation to respond very slowly. Using arrows you can quickly see that the outlets do appear to be reversed. However, this I believe is caused by conflicting boundary conditions - the combination of inlets, outlets, and inlet-outlets. Also using the compressible solver with such low Relaxations Factors for p and e will likely prevent any meaningful convergence.

I suggest you use the incompressible solver (it being faster and more robust) to diagnose issues with the boundary conditions, then once you are satisfied that you can achieve a well converged solution try the compressible solver.

Strictly speaking the two faces you have specified as Symmetry Local should be coupled using a Cyclic boundary condition.

Thanks! That gives me a lot

Thanks! That gives me a lot to work with. I greatly appreciate the help.