PimpleFoam: High Courant and Instability Problem
Good morning,
I'm trying to study the porosity problem with OpenFoam because I have to simulate the porosity in a real geometry of thoracic aorta. So, I've started with a circular cylinder made of two parts. The first part of fluid, and the second part solid, which is the porous part of the cylinder. I've set the simulation with a Re=50. A constant velocity profile in z-direction (the axis direction of the cylinder) and it becomes unstable after a few steps. At the beginning the Courant is very low and the deltaT is that one I set, but after a few steps the deltaT becomes infinitesimal (e-5,e-6 and so on) and the Co reaches values of 256!!! How can I try to solve this problem and reach a convergence? If you need (and you probably do) I can post my setup, the geometry, mesh and so on. Thanks Alessandro |
My first thought: what happens when you disable adjustTimeStep in your contolDict?
|
Hi,
Can you post your... 1. checkMesh output 2. fvSchemes 3. fvSolution (surely it'll be easier if you just post case files) |
1.checkMesh output
Code:
Create time Code:
Code:
@HanSolo123: I'll try and I'll tell you the result. |
Quote:
and then: Code:
DILUPBiCG: Solving for Ux, Initial residual = 0.99612, Final residual = 0.284478, No Iterations 1001 |
Try changing:
1. Code:
div(phi,U) Gauss linear; Code:
div(phi,U) Gauss upwind; Code:
p Code:
p 3. Either increase number of outer iterations or switch to residualControl for exiting PIMPLE loop. 4. Further you can try switching from Code:
gradSchemes Code:
gradSchemes (and if you disable time step adjustment, sure your time step won't be reducing, but it won't lead to convergence). |
1. Done.
2. Done. 3. Which value of outer iterations should I set? Which files, referring to BCs, do you need? The "0" folder? Excuse me but it's my first time with OpenFoam and I'm still learning, so, thank you for your patience. |
With the 1. and 2. corrections, the problem reaches the convergence almost instantly and I can post-processing the results.
But I don't really know if they're correct or not :confused: |
Well, something greater than 1 ;) Start with 5. Or just switch to residualControl, i.e. you change PIMPLE dictionary to something like:
Code:
PIMPLE Yes, initial and boundary conditions are described in the files in 0 folder. It'll be easier if you just compress it and attach to the message. |
Well,
Correction no. 1 is just a switch to a lower order discretisation scheme. To be sure about results you have to use residualControl to check convergence. |
1 Attachment(s)
In the .zip file there are all the files of my simulation ("0", "constant" and "system" folders).
|
ICs and BCs seem to be reasonable. I lost one bracket it my previous post, so the correct PIMPLE dictionary (which is in fvSolution file) is
Code:
PIMPLE |
The solutions seems to be reasonable for the velocity. The flux becomes instable almost instantly. The problem is with pression, I think. Because the result shows values of 1e14 for the pressure inside the first fluid part of the body.
So, I think there's something wrong. My purpose - in doing this simple simulation with the cylinder - is to investigate and try to understand the potential of the porous media in order to increase the Re number in my main simulation with the aortic vessel. But now, with your suggestions, the simulation converges. |
Well, there could be a problem with connection between two regions. I still can't get why do you need mesh with two regions. You can create single region mesh (if it's just a cylinder, you can even have fully hexagonal mesh), then you create cellZone with topoSet utility and that's it.
|
Quote:
And in the main simulation I have the real geometry of thoracic aorta (that must be meshed with an unstructured mesh) and 3 porous cylindrical media that must be structured. So I've followed this way. This preliminary study is only made in order to know if I can set an higher Re number in the real simulation. And I can't use the TopoSet utility at all, so the simpliest way was that. |
Quote:
|
All times are GMT -4. The time now is 07:26. |