'CHT Heat Exchanger' simulation project by CbsTau


#1

I created a new simulation project called 'CHT Heat Exchanger':

A Test of the CHT Capability on a heat exchanger


More of my public projects can be found here.


#2

@CbsTau - neat geometry. Currently taking a look at it as well!


#3

Thanks. I am trying to build a CHT test model. My first one was very simple but failed to solve with OpenFOAM. Your support team said there is a bug in the coupling on that model which they are looking into. They said they got a “similar” model to work.

So i built this model in OnShape. The first version included a fillet / blend between the coil and the inner pipe, to simulate the material of the weld. But this model imported into SimScale with mssng faces, and so did not form a closed solid.

Then I removed the “weld” and it now imports correclty, but wont mesh.I have opened a support tag on this too.

Having come from 20 years in traditional FEA, CFD and DEM - i am very excited about SimScale and what you guys are doing. The problem i have at the moment is that with traditional tools, if you hit a problem, youcan usually find a way around it, whereas with simscale being cloud based and alot is removed from the user, when you encounter a problem - you have to wait for support to look into it. I do support what you are doing and you have made excellent ground, but Robustness of meshing and solving (with ribust default settings) is going to have to be alot higher that traditional CFD / FEA if you remove from the user the ability to fix / tweak things himself.

For example - if the model is a valid solid in OnShape, and then imports with missing faces into Simscasle - there is nothing the user can do to fix this.

I look forward to helping to make this work better going forward!

Let me know if you want access to the OnShape model.

Kind regrads,
Clinton


#4

Hi @CbsTau,

great feedback - thanks for elaborating!

Regarding the missing faces: Do you still have the model somewhere that was imported with missing faces? Ideally the original Onshape model + the imported one? Would be eager to look into this.

It’s interesting what you’re saying about “Removing the ability from the user to fix things”. I agree that you can never have enough robustness in a simulation tool and we’re aggressively working on it, but probably there will be always situations where something unplanned / unforeseen happens to the user. Here SimScale needs to provide good error logs + hints how to fix it. Currently the main tool for this is the “Event Log”, where for each operation all info, warning and error is logged with hints on how to fix it.

It’s far from being complete, but we’re adding hints here on a daily basis. So generally, I don’t think that being cloud based does necessarily mean that the user is limited in terms of tweaking/fixing problems. What do you think?

I’ll take a closer look at your heat exchanger now. CHT is in fact quite a new feature, so there are not as many public projects yet available. Generally that’s the best approach to start with SimScale - using a copy of an already completed project.

Looking forward to see this heat being exchanged :slight_smile:

Best,

David


#5

Hi David,

Many thanks for your reply. I appreciate what you have shared and do think this is on the right track!

Here is the SmScale model which shows the problems importing the Geometry directly from Onshape:
https://www.simscale.com/projects/CbsTau/cht_heat_exchanger_with_weld/

Here is the Onshape Model:
https://cad.onshape.com/documents/1dedae700401b877344c10db/w/92d17224bea80f3e6e8e64e1/e/29a615ff4af4447b992991d8

You will see two assemblies:

  • Assembly with Fillet
  • Assembly with Chamfer

If I model the weld with a fillet, the entire solid body is missing in Simscale. Only the Fluid bodies come through.
If I model the weld with a chamfer, then the soldi body comes through but is missing a face which gets assigned to one of the fluid boides.

Also, inside this Onshape document, you will find is the simple CHT model which caused OpenFOAM to give faulty results. It is simply a rectangle box with two rectangular channels for hot and cold water to flow. Your support team found that this geometry caused the thermal coupling between the domains to fail in OpenFoam.

I hope this helps. I look forward to your response.
Regards,
Clinton.


#6

Hello David,

I have reduced the Chamfer size of the weld in Onshape slightly, and now the model seems to import correctly, and gives me a reasonable mesh.

The other Geometry without the Chamfer for the weld does not mesh. It fills the “BaseMeshBox” with cells, instead of meshing the solid. The Geometry with the rounded fillet/blend for the weld does not import correctly in SimScale at all.

I will now proceed to see if it solves :slight_smile:

Regards,
Clinton.


#7

Hi Clinton (@CbsTau),

that’s great - the mesh indeed looks fine. However to be sure, one of our application engineers is right now looking into your case in more detail + working on a proof-of-concept simulation to make sure this works end to end as the import+meshing indeed looked strange for this geometry.

Best,

David


#8

Thanks David,

I found another small “bug” in the meshing… If both the inlet surface of the fluid and the outlet surface lie on the same face of the BaseMeshBox, then the mesher creates a single “zone” region for both faces, so you are unable to select one face as an inlet, and the other as an outlet. this is the case with the top two fluid surfaces of the “coil” Fluid.

I worked around this adding a bend to the inlet pipe, so that the face is in a different plane to the outlet.

I was then able to mesh and set up the model, and generate a solution, but i am getting the same error with the coupling that i had with my simple box model, called CHT_Test

https://www.simscale.com/projects/CbsTau/cht_test/

Your support team said the coupling wasn’t working on that model, and they would get back to me once they found a solution.

This model is doing the same thing… The coupling does not seem to transfer the heat from the hot fluid to the cold fluid.

Perhaps, i am making the same mistake each time? But i have followed the steps as outlined in your CHT Webinar from two weeks ago, so this is quite strange.

Please could you ask your support engineer to check the model in its current state. It has completed results which show strange temperatures in the fluids, and almost no temperature gradient in the Solid - pointing again to a problem with the coupling.

Kind regards,
Clinton.


#9

Hi Clinton (@CbsTau),

got it! Yes, will forward this to them. They are aggressively working on providing a solution for that. You should hear soon from them.

Best,

David