Internal flow meshing


#1

Hey everyone! I am having trouble meshing my geomtery. My cad seems to be water tight but meshing fails. I have attempted hex dom para and hex dom automatic.

My geometry is for internal flow i.e my cad is of the air inside the volume.

Appreciate any help with it !
Thanks!

Project link : https://www.simscale.com/projects/jeettrivedi/plenum/


#2

Hi @jeettrivedi,

According to the log, your geometry is in fact not watertight. You will need to correct this carefully in the CAD software. Please refer below.

Cheers.

Regards,
Barry


#3

Hi. I felt there would be no problem with my geometry since it was easily meshing on other platfroms. thanks for the input buddy!


#4

https://www.simscale.com/projects/jeettrivedi/nozzle/

To understand the setup better i started working on a simple C D nozzle for simulations. m boundary conditions were 1atm inlet and 90kpa outlet.

tolerances are in order of e-5

end time is 500 secs and max run time is 100000 s .

i cant seem to run my simulation since it says max no of timesteps exceeded . could someone help me clear ot wheter the end time and max time should be same ?


#5

Hello everyone, an update

I had success in meshing and running my nozzle simulation . I did that to work on my main project
this is the project link : https://www.simscale.com/workbench/?pid=2569433344198424510#tab_0-0

my mesh in hex dom auto seems correct for all the faces except for the inlet face. i tried solving it by using parametric and surface refinment level 2 and 3 but as seen (in mesh 6) the inlet geometry is absorbed and everything is distorted. I have tried parametric on same CAD file before and it only disappears when i use a surface refinment.

any suggestions would be very helpful

Thanks!


#6

Hi @jeettrivedi,

I was about to help you mesh on my end but I realize that your “plenum stl” may of the incorrect scale as it is currently over 300m long which compared to your “plenum step” geometry (about 0.3m) is obviously absurdly large.

Do rectify this issue and I will instead mesh your “plenum step” file to see if I can get something usable.

Cheers.

Regards,
Barry


#7

Hey ! is it fine if I use the STEP file? when i upload the .stl file it gets scaled 1000 times . i saved the same file in stl format but still it gets scaled


#8

Hi @jeettrivedi,

Step is fine. I’m currently meshing your STEP file. Let you know when I get something workable.

Cheers.

Regards,
Barry


#9

Hi. I think i need to refine that edge. im trying to iterate the feature refinement included angle so the solver detects the mesh. is this the right way?


#10

Hi @jeettrivedi,

You can see my mesh here but I’m still attempting to get the ideal mesh.

Regards,
Barry


#11

Hi @jeettrivedi,

While everything other parts of the geometry is fine, my attempts to obtain a proper mesh of the inlet has not been successful.

@jousefm Do you or any @PowerUsers_CFD have any ideas on how to get a better definition at the inlet? I’ve tried higher refinement levels for that particular inlet and feature refinements as well but to no success. You can refer to the distorted mesh below.


#12

Have you tried to reduce the Min cell volume [m³] parameter?


#13

Hi @pfernandez,

I will try so now!


#14

Hi! I have a mesh where the inlet is NOT distorted.


But I believe there might be some problem in the simulation. I did a Hex Dom Auto (level 5 ) and it meshes with the bounding box (I have kept only the rear face of box visible ). In my simulation, I do not need the box and just he faces of my actual solid.

The mesh geometry is a STEP file


#15

Hi @jeettrivedi!

Fantastic. Looks like the auto solver managed to solve the issue. I’ve been trying for the past few days playing around the meshing parameters in the hex-parametric meshing option but to no avail. At least you can try simulating now.

The box is by default and you can ignore it and perform the simulation as if the box isn’t there at all.

Hi @pfernandez, I lowered down this from a default of 1e-13 to 1e-15 then to 1e-20 and that didn’t really produce better results as far as I can see. Any other possible suggestions? The automatic mesher seems to be able to solve the issue albeit at the highest refinement level. Cheers for the input however!

Regards,
Barry


#16

@Get_Barried reducing min vol to -1e+30 turns this check off completely.

@jeettrivedi it looks like the mesher has meshed the wrong region, external region instead of internal. Try Hex-dominant parametric starting from one of Barry’s meshes and ensure the material point is within the geometry.

Hope this get you unstuck,
Darren


#17

@Get_Barried

I ran a simulation with the new mesh. same problem. It always solves till the same amount of timesteps and give same values of residues before displaying an error

Okay. I shall attempt reducing the volume.


#18

I’ve reviewed your simulation again and would like to add some additional comments:

  • You should split the geometry, because now it’s just a single surface. This is necessary if you want to specify inlet and outlet boundary conditions and also add prism-layers to the walls. You can use the STL splitting operation in SimScale.
  • You might want to align the outlets of the geometry with one of main Cartesian axes in order to get a better mesh definition at that boundary.
  • You should extrude the inlet in order to get a fully developed flow. As you have it now, when adding the bell you have a near perpendicular-to-the-flow wall facing the inlet. This way you are forcing a homogeneous flow, straight into a wall, which need to be deflected and this just calls for trouble. An even better solution —if you are simulating what seems to be an air intake— is to create an artificial plenum at the inlet with boundary conditions that of the freestream (have a look at the following video to get an idea).

Kind regards,


#19

Hi ! Thanks for the input! I shall make a CAD of it .
I have tried the splitting on stl but for some reason my geometery is scaled by 1000 times in stl file


#20

Hi @jeettrivedi!

That is no problem as we have a scaling operation inside the platform. Please also have a look at that post for more information.

All the best!

Jousef