Several questions on meshing and simulation

BakhtiyarBakhtiyar Member
edited September 15 in Preprocessing

Hello, I am trying to replicate the paper with simplified human geometry. I have created a domain, which is divided in 3 and meshed separately.

My questions are:

  1. How do I do mesh verification (grid independence)? Do I just increase mesh nodes size by 50% for each region?
  2. How do I create some kind of bond between different types of meshes so the data transfered is not distorted?
  3. Can I cut the domain in half and apply symmetry boundary condition?
  4. I tried to run the simulation with CFX with residual target 10^-5, Everything goes fine till 10^-4, but after starts to fluctuate. The solver gives a warning: "there is a some percentage of wall placed at the outlet. Please consider changing to opening type boundary condition." Any suggestions?

Thank you


  • kkanadekkanade Forum Coordinator
    1. Please check gradients and then accordingly decide refinement criteria. 50% refinement would be too much. So please check areas with high gradients and refine the mesh there first.
    2. from images it is clear that you have non conformal mesh. It is ok. Please make sure that you do not have cell jump across interface by factor more than 3. If possible, you can create a conformal mesh instead of non conformal depending on your solver set up.
    3. Again depending on your solver set up, you need to decide on symmetry. Geometry looks symmetric.
    4. CFX guys can comment on last one. But I think if you have sufficient grid refinement you would be able to achieve 10e-5.

    For more details please check help manual. Also please use search on forum which would give you lot of posts regarding your questions.



    Guidelines for Posting on Ansys Learning Forum

    How to access ANSYS help links

  • Thank you for your reply

Sign In or Register to comment.