|
[Sponsors] |
January 31, 2017, 00:30 |
PIMPLE algorithm unstability
|
#1 |
New Member
Ram
Join Date: Aug 2016
Posts: 2
Rep Power: 0 |
Hello ,
I'm trying to simulate a particulate flow in an air filter with MPPICFoam solver, K-epsilon turbulence model but the solution crashes with increase in time step continuity errors after some time, coupled with rapid increase in courant number. I'm unable to figure out where the problem lies in my setup. Checkmesh looks good and I'm running case with adjustable time step. I've attached relevant geometry, mesh and simulation files. Please have a look at it and let me know if I'm doing anything wrong and if the setup is compatible to simulate the flow I want to. boundary_conditions.txt mesh.png fvSchemes and fvSolution.txt checkMesh.log.txt solutionlog.txt |
|
January 31, 2017, 04:27 |
|
#2 |
Senior Member
Oskar
Join Date: Nov 2015
Location: Poland
Posts: 184
Rep Power: 10 |
Hello ram_7.
I can see in Your checkmesh.log that max non-orthogonality is pretty high: Mesh non-orthogonality Max: 67.1147. Also mas skewness looks little high to me. Check out this topic: Max Courant Number exploded |
|
February 1, 2017, 10:58 |
|
#3 |
New Member
Ram
Join Date: Aug 2016
Posts: 2
Rep Power: 0 |
Thanks Sheaker for a quick reply.
Yes, indeed the max values of orthogonality and skewnes are higher than ideal values , but only very few elements(<30) in the more complex region of geometry had these values. Also mean skewness is 0.18 with standard deviation around 0.08 as per the data in ANSYS meshing tool. The cutCell meshing method I used is having trouble generating good quality elements in those more complex regions but overall, the quality was good. 1.Do you think the problem is with the mesh, or are the fvSchemes and fvSolution I chose are causing non convergence of solution.Does those few elements with less quality affect the solution convergence in such a way? 2.Also I'm unable to understand why courant number is reaching such unrealistic values. With Co = (U* dt / dx), such high courant number means velocity calculated is orders of magnitude greater than inlet velocity, which is not possible. Your insight in these aspect would be helpful. Thanks. |
|
February 1, 2017, 14:06 |
|
#4 |
Senior Member
Oskar
Join Date: Nov 2015
Location: Poland
Posts: 184
Rep Power: 10 |
Hello ram_7.
1. According to CFL law I think even one bad mesh cell could affects on Your simulation. 2. I am not an expert of CFD but I think that velocity on those few non-orthogonal mesh cell could be calculated wrong so the Courant Numer increase. When CO reach 1 it affects of velocity with feedback and everything explode. Keep in mind that bad mesh is just my first try! |
|
February 8, 2018, 09:43 |
|
#5 | |
Senior Member
Join Date: Aug 2014
Location: Germany
Posts: 292
Rep Power: 14 |
Quote:
Hi, have you made progress with your problem? Have you tested the case without langrangian injection? |
||
Tags |
courant number increasing, mppicfoam, openfoam 3.0.1, pimple. openfoam |
Thread Tools | Search this Thread |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
PIMPLE algorithm | sahmed | OpenFOAM Running, Solving & CFD | 5 | March 21, 2019 01:19 |
pimple algorithm details for transient, incompressible flow | cfdopenfoam | Main CFD Forum | 0 | November 15, 2015 23:31 |
Floating point exception error | lpz_michele | OpenFOAM Running, Solving & CFD | 53 | October 19, 2015 02:50 |
A question on the PIMPLE algorithm | GerhardHolzinger | OpenFOAM Running, Solving & CFD | 4 | February 13, 2015 06:49 |
On the PIMPLE compressible algorithm | ssss | OpenFOAM Programming & Development | 1 | September 13, 2014 15:15 |