Savonius Turbine AMI Rotating Zone


I am trying to determine the cut-in speed for a Savonius Turbine by doing multiple simulations with different windspeed and a fixed rotating speed for the turbine. My idea is to see at which windspeed the moment on the blades is close to 0. This is the link to my project: SimScale

I believe I have the right CAD setup, and I have looked extensively at the tutorials regarding rotating zones (since I have lateral flow, I saw that AMI rotating zones are most appropriate). However, I keep getting velocity divergence errors at a wall boundary.

“Velocity field started diverging. Please check the mesh quality near the reported location and try refining the mesh. If the problem occured near a boundary, please check the boundary conditions. In case of doubt, please ask for assistance via our support chat. Velocity = 1.43894e+07 at position: (0.02182 m, 0.1447 m, 0.0959 m).”

This is what I have tried so far:

  • Refining the mesh of the rotating region and looked at isovolumes
  • Introduced 4 non-orthogonal correctors
  • Checking my boundary conditions. I only have a velocity inlet, a pressure outlet, and some non-slip walls for the turbine
  • Allowing the flow to be fully developed before entering the rotating zone

I would really appreciate any help. Thanks,

Hi there, I see that you are running a simulation with a brand new mesh now, so let me know how this goes - I can not inspect the previous results, as the mesh is now changed-.

However I noticed that some steps described in the following page under the AMI section were left out: Rotating Zones | Advanced Concepts | SimScale

Specifically, you should set a new cell zone, as described in this forum post as well:

Also, always use these guides when divergence occurs:

Best regards,

Thanks for the suggestions Fillia, I disabled the physics-based meshing and added the cell zone. I also eliminated the no-slip wall condition. Lastly, I added an inflate boundary later to the blades. While inspecting the mesh log, my mesh almost satisfies the general requirements of the documentation you linked, but there are some extreme anomalies like a node with a volume ratio of more than 300 and a node with aspect ratio of 123. However, the divergence always occurs at the same point and that point is not a node with high volume ratio when analyzed with isovolumes. What else should I be doing?

I suggest you refine the area where the divergence appears (or the local element size if it happens on a surface). Alternatively, you can increase the uniform fineness by applying a higher Level, which makes sense if there are more problematic areas that will probably create divergence later on too.

Fillia, I am having trouble understanding your suggestion. By refining one area wouldn’t that increase the aspect and volume ratio, which is undesirable? How important is the mesh uniformity between the large flow region and the rotating one? I have absolutely no knowledge of this, so I would appreciate any explanation. This project (SimScale) is very similar to mine and does not get any errors while having way coarser mesh, so is it mesh coarseness that leads to divergence? or lack of mesh uniformity?