Curved distributed resistance as volume part possible?

[Anonymous]

I have a problem in the flow behaviour through a curved volume part which I assigned a distributed resistance to. If I use shells (which are extremely unconvenient since I have to select all the surfaces, which can be several hundreds) I come to the right results as calculated using Darcy or I have measured. If I want to use a volume part it works as a flat sheet (Assign resistance and define the flow direction and have the normal directions 5 dimensions higher) it is simplay a case of mesh refinement to generate the right result.

Due to the nature of our products we use pleated (folded) layers which shall be simuated. If I do this in a very simple model the shell method works (but is not very practicable) but the usage of a volume part comes to strange results, due to my opinion since the flow direction changes in the different fold parts. As illustration I have attached a picture showing the constant pressure loss in flow direction whcih is about 4 times too small.

Does anybody have an idea or a solution to this problem. For my sim it is very important to show the right flow behaior over the resistance.

[Jon.Wilde]

Hi,

Could you share a little more about the real world geometry here?

As an aside, you can box select surfaces to quickly select all of them in one go 🙂

Kind regards,

Jon

[Anonymous]

Hi Jon,

yes I will: In the attached pic the red outline is a 0.8 mm thick porous medium which is described by its permeability. The model itself is a 10 by 10 cm square tube with this medium as a barrier in the middle just to show my team how they can handle our special designs.

Permeability is given by the K-factor input with the constant set to zero in flow direction and to 10e+12 in the two normal directions.

Yes you are right the box select should be the method of choice, but it becomes complicated when you have a more complex model (in reality the curved area is a folded medium, rounded and welded to give (seen from the top) give a star shape with 40 to 60 pleats. Then you have to rotate and select since it is not possible to select all right surfaces at one time. So I have to be patient and select the surfaces by a single click and to add them into groups ...

A nice feature would be like this: Select a volume and swith to surface with selects all surfaces of this volume (nice) or opening a dialog which surfaces you want (outside, inside, top, bottom - nicer).

[Jon.Wilde]

Thanks for the information - a 3D image might also help although I think I've got it 🙂

As we cannot have the proeprties of a 3D resistance change with the dimensions of the part it seems to me that the only method here is a surface resistanceas you tried previously.

If you share the model we could try to come up with a neater way of selecting the surfaces if possible? Do they not all have the same properties?

Thanks,

Jon

[Anonymous]

I have a .sat file of a simple inline filter with a pleated element. I do not want to share this with the complete community 😉 how do I get it to you?

[Jon.Wilde]

OK, I've sent a shared folder to your email. Although I would like to continue the discussion here, so the community might learn too if anyone reads this thread. Hopefully we can find some images to share but I won't post anything.

Thanks,

Jon

[Anonymous]

I have no objection in sharing the result, I was only referring to the total product which is included in the .sat file.

[Jon.Wilde]

Hi,

Thanks for sharing the file. Do you need all the detail on the full model or would it be sufficient to use a 2D model to get an understanding of how the fluid flows over this and then apply it to a 3D model with a simplified resistance?

If you need to run the full 3D detail you could at least cut this in half and use symmetry, which would help with speed/memory. I think the only way to do this would be surfaces - you could have one on either side (with half the resistance), provided you had enough mesh to give you a mid-node through the thickness of the medium (and between the pleats where the surfaces are close to touching). Perhaps that might be a little more realistic then. One point to know with surfaces though is that the flow will always exit normal to the surface - would this be representative of the medium?

Kind regards,

Jon

[Anonymous]

Thanks Jon, this is helpful. Yes we need the full 3D model, but I never tried to use to resistances with half of the value. I will try to do so. The shell behaviour is very realistic for filter media, since they are very thinn compared to all other parts and due to our measurements the is neglectable flow in the normal directions. Do you have an additional tipp on how simply to select all the surfaces?

[Jon.Wilde · Accepted solution]

Hi,

No problem. Yes I am not sure why you cannot just box select them all? If you hide the other volumes and just leave the filter on screen, then 'box it' and you should be able to assign all surfaces in one go? Use the view cube to orient the view in a useful manner. As of yet I am not sure what the stumbling block was here but I may well be missing something.

Thanks,

Jon

[Anonymous]

Jon,

I think I was a little slow here. If I do a box select for all surfaces a deselect only top and bottom, I think I have the solution when I use half of the resistance out and half of it in. I never came on that idea and will definitely try it.

Thanks for your help!

[Jon.Wilde]

No problem.

As an addition, once you have them selected right click to put them in a group - helps with quick selection later. For mesh refinement maybe 😉

[Anonymous]

Obviously I still make a mistake with the resistances. In the attached file, I created a very simple model: A square tube 100 x 100 mm with 500 mm length and a 0.8 mm thick flat barrier in the middle. If I define this part in the middle as volume distributed resistance with an open area of 32,65% (perforated sheet) and create a volume flow of 200 l/min Water through it it gives the right result of about 6 mbar differential pressure between inle and outlet.

If I define the volume as water and assign the front surface as a resistance of 0.8 mm with the same open area I come to 1.8 bars of pressure. If I use as suggested two shells with 0.4 mm thickness each on the front and back with again the same open area I get 37 bars.

What is wrong with this?

And: If I go to a material described by its permeability the right result comes from the single shell resistance? Is there any wrong understanding on my side?