[Abhinav_Nagarajan]

Dear Foamers,
I am new to openFoam and am having troubling setting optimum under-relaxation factors for my case. My problem statement is as such:

I want to simulate a reactor (see pic attached) with laminar incompressible flow, using SimpleFoam. The Reynold's number suggests that the flow within the reactor is well within 2000, hence is laminar through out the reactor.

Problem:
I am using the SimpleC algorith for my solver, and when I set the under-relaxation factors for the velocity to a high value such as 0.9 it leads to an oscillatory behaviour and the solution does not converge.
I changed the under relaxation factor to a low value such as 0.001, and the solution converges, however I am unsure of the solution. I have tried a list of relaxation factors in between, and all exhibited a non-convergent behaviour.

I need to know if I am going wrong somewhere with the low values for the under-relaxation factors. Could you guys please guide me on this?

I have attached the fvSchemes, fvSolution files as well for reference. The dimensions mentioned in the jpeg image of the reactor are in mm.

This is a bit urgent, since its a part of my master thesis, please help me out guys.

Thank you for your time and consideration.
Sincerely,
Abhinav

[SHUBHAM9595]

Well fvSchemes looks good, now coming to the relaxation factor, 0.001 is too much for sure. Go for something like

relaxationFactors
{
    fields
    {
        p               0.3;
    }
    equations
    {
        U               0.7;
    }
}

Also considering the geometry, you can use 2/3 nonorthogonal correctors.


Regarding this,

I have tried a list of relaxation factors in between, and all exhibited a non-convergent behaviour.

have you scaled your mesh ? If not then you can execute

convertToMeters    0.001;

after reading the mesh as OF works in meters only. It's also nice to once have a look at the output of checkMesh for aspect ratio and non-orthogonality.

[Abhinav_Nagarajan]

Thank you Shubam for your prompt reply. I have attached the checkMesh file for your reference. I have already tried with the orthogonal correctors and the higher relaxation factors as well. Nothing seems to work.

Originally Posted by SHUBHAM9595

Well fvSchemes looks good, now coming to the relaxation factor, 0.001 is too much for sure. Go for something like

relaxationFactors
{
    fields
    {
        p               0.3;
    }
    equations
    {
        U               0.7;
    }
}

Also considering the geometry, you can use 2/3 nonorthogonal correctors.

Sure Shubam, I will try it out.

Regarding this,


have you scaled your mesh ? If not then you can execute

convertToMeters    0.001;

after reading the mesh as OF works in meters only. It's also nice to once have a look at the output of checkMesh for aspect ratio and non-orthogonality.

I have already scaled the mesh.

[SHUBHAM9595]

As you can clearly see your checkMesh has failed bcz of concave cells. For 2D geometry, it should not affect that much. But for your case, I'll suggest you to have a look at the 3rd post of the following link.

About OpenFOAM and concave faces or cells

After this, u can import the case in paraView and can easily visualize the local cells where u need to perform further refinement to avoid their concave shape.