Fluids

Fluids

Problem with mesh in boundary layers

    • aitor.amatriain
      Subscriber

      I want to add boundary layers in three walls (0 thickness) that separate three parts of my preliminary model

    • Keyur Kanade
      Ansys Employee
      The issue is not very clear. The images look different.
      Are you doing 2D or 3D ?
      Regards Keyur
      How to access Ansys Online Help Document
      Guidelines on the Student Community

    • aitor.amatriain
      Subscriber
      3D, Fluent Meshing. The images are taken from the opposite sides.
      The issue is that when the boundary layer approaches the walls marked in red, the number of elements reduces, and I want to have a constant number of elements.
    • aitor.amatriain
      Subscriber
      It is not possible to solve my issue?
    • Rob
      Ansys Employee
      The problem you'll then face is the poly cell at the "end" of the inflation layer. If you mesh with poly only then refine near the wall will that work for the flow field?
    • aitor.amatriain
      Subscriber
      Hi Rob, thank you for your help. I understand that what you propose is to mesh the geometry without the boundary layer, to then add the "boundary layer" by means of a local refinement. How can I do that in Fluent Meshing?
      The only solution that I can think is to create a zone near the wall and then select in that zone a BOI or body size much lower than the other one. However:
      All the elements will have the same length in all directions. I do not want this because in this case the cell count will increase dramatically.
      I will loose the possibility of increasing the cell height as the cells are further from the surface.

    • Rob
      Ansys Employee
      I was thinking of adaption in Fluent. Inflation is a good way to refine on continuous walls, it's less useful for baffles and where jet impingement/flow detachment occurs.
    • aitor.amatriain
      Subscriber
      I do not use the (non dynamic) mesh adaption in Fluent because I had bad experiences in the past, but I understand that it is the best option.
      Dou you have any recommendation to share about the implementation for this case in particular?
      Thank you
    • Rob
      Ansys Employee
      No need to use dynamic adaption. If the flow is fairly stable run a few hundred iterations and then adapt a couple of cells from the wall & it's shadow.
    • aitor.amatriain
      Subscriber
      My answer was not very clear, sorry. I meant that I have used the dynamic mesh adaption (essential for some multiphase flows), but I have reduced experience with the "static" mesh adaption.
    • Rob
      Ansys Employee
      No worries. If you use the boundary adaption in 2022R1 I don't see any issues. We switched to PUMA and have since improved it significantly over the older hanging node and early PUMA implementation.
Viewing 10 reply threads
  • You must be logged in to reply to this topic.