Volumetric mesh with cell volumes smaller than machine precision

- November 8, 2024 at 7:20 pm
  zoelle.wong
  Subscriber
  My y+ =10, gives me a cell height that is very small 2x10^-6. This is due to my fluid speed and choice of fluid.
  When making a volumetric mesh, this gives me a cell volume that is well below machine precision (see cut out below of cell volume in mesh). The mesh has a 200 million point cell count, and the conservation variables are exponentially increasing. I suspect this is due to the volumetric size since the flow field "jumps" in areas that look like my grid (flow field below).
  My questions are as follows:
  - Does ANSYS have a way to handle these cell volumes? If not, is there a scaling function that users can employ such that the boundary-conditions are also scaled?
  - How well can ANSYS solve a volumetric domain that is comprised of structured and unstructured grids?
  thank you so much!
- November 8, 2024 at 7:23 pm
  
  zoelle.wong
  Subscriber
  
  sorry the image is discontinuuous - the file size was too big
- November 11, 2024 at 1:32 pm
  
  Federico
  Ansys Employee
  
  What length units are you using here?
- November 11, 2024 at 3:12 pm
  
  zoelle.wong
  Subscriber
  
  Meters!
  - November 11, 2024 at 3:34 pm
    
    Federico
    Ansys Employee
    
    You can try setting your case in milimeters to see if it helps. Also, make sure you use double precision
- November 12, 2024 at 12:12 pm
  
  Rob
  Forum Moderator
  
  Fluent should be OK with that scale (we run micro fluidic models) but check the near wall resolution and aspect ratio too.
  Hybrid meshes should be fine, and were routine up until we moved to poly meshes for many applications. However, check the cell size change at the boundary between the cell zones as the pyramid caps can cause issues.
  As an aside, the length unit is metre.... ;)