Understanding openfoam residual discontinuities

johannes.tophoj · April 8, 2016, 05:44

Hi

On the attached image the residuals of my flow variables are shown as a function of the iteration count. The convergence criterion is 1e-5 for all variables but p, for which it is 1e-6. The residuals of the variables are discontinuous/make jumps, but perhaps it could be interpreted like some truncation as the shape of the curves before and after the jumps seem to match.

I'm merely curious. Can anyone explain these jumps?

Johannes

wyldckat · April 16, 2016, 12:14

Quick questions:

Which solver are you using?
What relaxation values are you using?

johannes.tophoj · April 21, 2016, 03:08

Hi

The solver is simpleFoam (consistent yes/true) with relaxation factor 0.3 on all variables. I have seen these jumps in several similar cases and as this case eventually converged (at iteration ~2600) my interpretation has been that it is some kind of numerical artefact and not a sign of convergence problems. But it did make me wonder...

Johannes

wyldckat · April 21, 2016, 17:44

Hi Johannes,

I recently had a similar problem when using k-omega SST and it was related to two problems:

I was running in parallel, for which k-omega is apparently more sensitive to the values defined in the boundaries named "processor".
Turbulence fields should not be overly solved in the matrix equations.
- In other words, if you have a very small tolerance in the entries for k/epsilon/omega in "fvSolutions.solvers", it can result in these situations you observed or it could potentially crash as well, because it diverged, given that the solution was incompatible with the other equations.

Best regards,
Bruno

johannes.tophoj · April 25, 2016, 08:51

Hi again

Yes my jobs are parallel as well though I so far have not had any crashes. But thank you for the explanation

Johannes

April 16, 2016, 12:14		#2
wyldckat Retired Super Moderator Bruno Santos Join Date: Mar 2009 Location: Lisbon, Portugal Posts: 10,975 Blog Entries: 45 Rep Power: 128	Quick questions: Which solver are you using? What relaxation values are you using? __________________ OpenFOAM: FAQ \| Getting started Forum: How to get help, to post code/output and forum guide Read this before sending me PM

April 21, 2016, 17:44		#4
wyldckat Retired Super Moderator Bruno Santos Join Date: Mar 2009 Location: Lisbon, Portugal Posts: 10,975 Blog Entries: 45 Rep Power: 128	Hi Johannes, I recently had a similar problem when using k-omega SST and it was related to two problems: I was running in parallel, for which k-omega is apparently more sensitive to the values defined in the boundaries named "processor". Turbulence fields should not be overly solved in the matrix equations. In other words, if you have a very small tolerance in the entries for k/epsilon/omega in "fvSolutions.solvers", it can result in these situations you observed or it could potentially crash as well, because it diverged, given that the solution was incompatible with the other equations. Best regards, Bruno

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Problem with chtMultiregionFoam radiation boundary condition	baran_foam	OpenFOAM Running, Solving & CFD	10	December 17, 2019 17:36
Floating point exception error	lpz_michele	OpenFOAM Running, Solving & CFD	53	October 19, 2015 02:50
Wrong fluctuation of pressure in transient simulation	caitao	OpenFOAM Running, Solving & CFD	2	March 5, 2015 21:33
Unstabil Simulation with chtMultiRegionFoam	mbay101	OpenFOAM Running, Solving & CFD	13	December 28, 2013 13:12
Micro Scale Pore, icoFoam	gooya_kabir	OpenFOAM Running, Solving & CFD	2	November 2, 2013 13:58

April 21, 2016, 03:08		#3
johannes.tophoj New Member Johannes Rasmussen Join Date: Dec 2011 Posts: 9 Rep Power: 14	Hi The solver is simpleFoam (consistent yes/true) with relaxation factor 0.3 on all variables. I have seen these jumps in several similar cases and as this case eventually converged (at iteration ~2600) my interpretation has been that it is some kind of numerical artefact and not a sign of convergence problems. But it did make me wonder... Johannes

April 25, 2016, 08:51		#5
johannes.tophoj New Member Johannes Rasmussen Join Date: Dec 2011 Posts: 9 Rep Power: 14	Hi again Yes my jobs are parallel as well though I so far have not had any crashes. But thank you for the explanation Johannes