[rishik686]

Hi everyone,
I am trying to run a fluid flow simulation in an artery using pimple Foam in open Foam.
But the courant number is increasing abruptly.
geometry in complicated

Earlier when I tried the same for simple tube it run perfectly with stable Courant No.

Please help me to stabilize the courant no. so that my simulation can converge.
__________________
Rishi K.

[DLuo]

Easiest things to tackle are
1. reduce your timestep
2. decrease your element size (finer mesh)

If you reduce your timestep by 1/2 and the solution becomes stable. that may give you a quick idea if your current mesh is suitable.

Also you may try ramping your timestep. Start with a finer timestep and see if you can make it past the point of instability before increasing your timestep up to your desired step size.

Lastly you could just have some mesh quality issues. That's always a pain to troubleshoot and improve so I'd suggest starting with suggestion 1.

[FMDenaro]

Originally Posted by rishik686

Hi everyone,
I am trying to run a fluid flow simulation in an artery using pimple Foam in open Foam.
But the courant number is increasing abruptly.
geometry in complicated

Earlier when I tried the same for simple tube it run perfectly with stable Courant No.

Please help me to stabilize the courant no. so that my simulation can converge.
__________________
Rishi K.

You are not clear. You see a numerical instability? The what about your numerical procedure?
Have you tried to reduce the time step?

[rishik686]

Originally Posted by DLuo

Easiest things to tackle are
1. reduce your timestep
2. decrease your element size (finer mesh)

If you reduce your timestep by 1/2 and the solution becomes stable. that may give you a quick idea if your current mesh is suitable.

Also you may try ramping your timestep. Start with a finer timestep and see if you can make it past the point of instability before increasing your timestep up to your desired step size.

Lastly you could just have some mesh quality issues. That's always a pain to troubleshoot and improve so I'd suggest starting with suggestion 1.

Thanks for the reply Dlou.
I have tried the same problem with the finer mesh and the stability is there but the Max CFL is still >1. and the Average is >0.7

Time step reduction is not helping.
The Mesh quality improvement may be required.

[DLuo]

So you're saying the finer mesh is stable using the original timestep? Your concern is that you think the solution is inaccurate due to the max Co number being reported as >1?
That's fine, it's just telling you how much of your small time-scales you're able to resolve. as Max CO goes up you start to approach more of a URANS solution but that still has value in helping you understand how the mean flow varies with time.
Personally it sounds like you've already fixed the issue especially if that Max CO value is still in the neighborhood of 10 or less. I'm not sure how much compute resources you have available but consider the cost of wading into the unknown trying to fix your mesh with what limited options you have available in SnappyHexMesh vs just letting your computer chomp away on the finer mesh. I would bet that you will spend overall less physical/mental energy running the finer mesh.

[rishik686]

Originally Posted by DLuo

So you're saying the finer mesh is stable using the original timestep? Your concern is that you think the solution is inaccurate due to the max Co number being reported as >1?
That's fine, it's just telling you how much of your small time-scales you're able to resolve. as Max CO goes up you start to approach more of a URANS solution but that still has value in helping you understand how the mean flow varies with time.
Personally it sounds like you've already fixed the issue especially if that Max CO value is still in the neighborhood of 10 or less. I'm not sure how much compute resources you have available but consider the cost of wading into the unknown trying to fix your mesh with what limited options you have available in SnappyHexMesh vs just letting your computer chomp away on the finer mesh. I would bet that you will spend overall less physical/mental energy running the finer mesh.

Thanks for the reply DLuo.

regarding about mesh, I have not generate my mesh with SnappyHexMesh i just import it from ansys mesh modular. while checking the quality in openFOAM it says that the mesh is OK.