Internal flow meshing

Hi @pfernandez,

I will try so now!

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

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

@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

1 Like

@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.

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,

3 Likes

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

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

Hey! I started working on a different geometry with similar boundary conditions. attempting to do an incompressible flow first.

I’d like some input on simulation control, how much should my end time be ?
ideally a steeady state in my case would be good enough for 10 seconds of real time (the time for which my condition would persist in real life )

also, how much should my relative tolerance aim should be , i did a sim successfully with a relative convergence of 0.01 and 0.3 as relaxation factors. my absoulte residue target for p, u, k etc were at or less than 1e-5 but my post processor values dont make sense…

should i play around a lot with factors such as gauss limited corrected etc ?

Hi @jeettrivedi and sorry for the delay!

Have you got any results so far or are you still struggling to get your simulation done? Looking forward to hear about your progress!

All the best!

Jousef

No. i have simulated once but the results were wrong. I recently found an air intake simulation done by milad mafi and amd attempting to replicate it .

I am unnsuccesful with that as well.

@Milad_Mafi Hi. Could you please help me out here .

Hi @jeettrivedi!

Could it be that your project is still set to private? If so please change it so we can have a look at it :slight_smile:

Thanks!

Jousef

Hi. Yes. I have shared with support. let me know if you can browse it .

Hi @jeettrivedi!

Would be good to share it for others users as well.

Best,

Jousef

done.

Hi @jeettrivedi,

Could you give more details on what was wrong and what is to be expected from the results as well as which simulation you are referring to?

Cheers.

Regards,
Barry

Hi @Get_Barried & @jeettrivedi!

Jeet already gave me some more details: “sim in first link. basic run, run 6, run 7. the values converge but expecting a plenum pressure ranging from 98kpa to 70kpa …”

Best,

Jousef

1 Like

Hi Jousef!

Ah sorry completely overlooked that, will take a look into the project.

Cheers.

Regards,
Barry

1 Like