Chip cooling ==> new solver available

[Greg66]

I fully agree with your suggestions and thank you for the offer. However unfortunately this is not something I personally have time for at the moment, but I will happily review any pull requests containing such enhancements.

In order to do that, i certainly would need some guidance from you.
I do have experience with Visual Basic (Visual Studio), but neither with python, programming FreeCad workbenches nor internals of CFDOF-workbench.

[julieng]

Hello,

Good to see that buyantSimpleFoam solver is now supported!
It is on the road to have CHT heat transfer solver available some day. Thank you for your work

Julien

[julieng]

Hello,

I test the compressible solver.

What is the recommended value for y+ ? With the wll function SST I know that y+ need to be between 30 - 100.
But to resolve the energy equation normaly the value of y + should be decrease close to 1 ?

Other question on BC.
When I choose in the solver compressible, it open new choices in BC. When double click it open the window but with simple click
in the table we have more choices than in the window with the line "thermal boundary type".
"thermal boundary type" can be : "fixed value" = Dirichlet condition
"zero gradient" = thermal insulation for example
"fixed gradient" = Heat flux or Heat transfer coefficient

First problem, in the window by double click I cannot chose heat transfert coefficient, only heat flux.
Outlet should be "zero gradient" ? Only I have the option of "fixed value" by double click

I would test the following BC: Inlet = 500 K
Outlet = zeroGradient
Wall = Heat transfer coefficient = 50 W/(m2.K) with external temp = 290K
Symmetric plane = zeroGradient


How to do this with the settings of cfdOF?

My first try give:

Code: Select all

Calculating approximate pressure field
20:52:33  

--> FOAM FATAL ERROR: 
updateCoeffs(const scalarField& snGradp) MUST be called before updateCoeffs() or evaluate() to set the boundary gradient.

20:52:34      From function virtual void Foam::fixedFluxPressureFvPatchScalarField::updateCoeffs()
    in file fields/fvPatchFields/derived/fixedFluxPressure/fixedFluxPressureFvPatchScalarField.C at line 157.

FOAM exiting

[julieng]

I try to run the test case chipCooling but it doesn't converge. I use bluecfdcore with FC0.19 on Windows 10



Julien

[thschrader]

I try to run the test case chipCooling but it doesn't converge.

My residuals look the same.
I would see if the solver runs, not computing a physical correct result.
You can try to rebuilt one of the OF-tutorials in FC-cfdof for comparison.

[oliveroxtoby]

Other question on BC.
When I choose in the solver compressible, it open new choices in BC. When double click it open the window but with simple click
in the table we have more choices than in the window with the line "thermal boundary type".
"thermal boundary type" can be : "fixed value" = Dirichlet condition
"zero gradient" = thermal insulation for example
"fixed gradient" = Heat flux or Heat transfer coefficient

First problem, in the window by double click I cannot chose heat transfert coefficient, only heat flux.

I have added this option to the GUI.

Outlet should be "zero gradient" ? Only I have the option of "fixed value" by double click

It is treated as zero gradient. The temperature specified is a reference value needed by the boundary conditions used with HiSA. Needs to be removed for buoyantSimpleFoam, but for now it can be ignored.

My first try give:

Code: Select all

Calculating approximate pressure field
20:52:33  

--> FOAM FATAL ERROR: 
updateCoeffs(const scalarField& snGradp) MUST be called before updateCoeffs() or evaluate() to set the boundary gradient.

20:52:34      From function virtual void Foam::fixedFluxPressureFvPatchScalarField::updateCoeffs()
    in file fields/fvPatchFields/derived/fixedFluxPressure/fixedFluxPressureFvPatchScalarField.C at line 157.

FOAM exiting

The "potential flow" initialisaition type does not seem to work in this case. I am not sure why. You can initialise with a uniform pressure.

Pressures for the compressible solvers need to be specified as absolute (e.g. 100kPa).

[julieng]

Hello,

I have update the cfdOF WB and now I can impose heat transfer coef by double click. But I think that there is a problem again on this because when click OK and I open again the window the heat transfer coef is not record.
I try without potential flow initialization. Imposing 10 Pa for the pressure. The simulation runs but the results is not realistic for the temperature, the temperature is constant everywhere in the volume (= Temp of fixedValue of the inlet BC). For flow variables (U p etc) it is realistic. I think the error comes from the non record parameter of heat transfer coef.
The simulations finishes with success but the convergence criterion is not reaches...



Julien

[oliveroxtoby]

Hello,

I have update the cfdOF WB and now I can impose heat transfer coef by double click. But I think that there is a problem again on this because when click OK and I open again the window the heat transfer coef is not record.

I think you just need to delete and re-add the boundary condition since it was created before the update to the code.

[julieng]

Yes, now it works!

But the solution is not converged (I am in steady case) criteron for Energy > 10^-2 and the solver return that it is finished.

Do you know which y+ is required with this solver?

Best regards

[oliveroxtoby]

Yes, now it works!

But the solution is not converged (I am in steady case) criteron for Energy > 10^-2 and the solver return that it is finished.

Do you know which y+ is required with this solver?

Best regards

Please update - I think I fixed that. Not sure about y+ off the top of my head, sorry.