View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0001354 | OpenFOAM | Bug | public | 2014-07-21 13:32 | 2015-04-30 20:28 |
| Reporter | Assigned To | henry | |||
| Priority | normal | Severity | major | Reproducibility | always |
| Status | closed | Resolution | fixed | ||
| Platform | Linux | OS | Ubuntu | OS Version | 12.04 |
| Summary | 0001354: interFoam pressure miscalculation in 2.3 (wrt previous versions) | ||||
| Description | I have been running a benchmark case to test each new version of OpenFOAM. It is just a simple dam break case in 3D for which pressure and free surface laboratory data are available. Up to version 2.2 the numerical results match the experimental ones with a high degree of accordance. However, in version 2.3 everything is off, and especially the pressure, having differences of more than 2000Pa (200% difference). See attached file. It includes a report and the cases to run in OpenFOAM 2.1.1 and 2.3. | ||||
| Steps To Reproduce | ./runParallelCase Comparison: python plotCompPres.py python plotCompVOF.py | ||||
| Additional Information | I first thought that the new pressure boundary condition (fixedFluxPressure) introduced in 2.3 might be the cause of this problem. I compiled buoyantPressure boundary condition from 2.2.2 in 2.3 and the results did not change, so fixedFluxPressure is discarded. The other factor that has undergone heavy changes from 2.2.2 to 2.3 is the way to calculate the VOF field, and I think it might be the most probable cause of the differences. However, having the same fvSchemes should in principle bypass the new techniques and yield the same results... | ||||
| Tags | No tags attached. | ||||
|
2014-07-21 13:32
|
|
|
|
You must use the fixedFluxPressure BC in OpenFOAM-2.3.x. |
|
|
Comparing the results in presOF230.pdf and presOF211.pdf the pressure levels look very similar. Could you please clarify where the 200% difference is seen? |
|
|
There is one notable difference between your setups for 2.3.0 and 2.1.1: In 2.3.0 you have chosen nLimiterIter 1; whereas is 2.1.1 the default is 3. Do you get unboundedness in your 2.3.0 alpha.water field? |
|
|
Indeed, fixedFluxPressure is used in the 2.3 case. The 200% difference is in pressure gauge 7, figures might look similar, but check the scale on the left, as it is different. There is a large peak in the 2.3 version which was not there before. Apparently I do not get an unbounded alpha.water field. I did not notice that change, I will try changing it and report back tomorrow. |
|
|
After running the case with nLimiterIter = 3 I can confirm that the results are identical and pressure is still miscalculated. |
|
|
Can you reproduce the problem on the 2D damBreak tutorial case? |
|
2014-07-22 08:17
|
|
|
|
Indeed. Please, find attached dambreak.tar.gz |
|
|
These dam-break cases are partly chaotic and the behaviour, particularly the slamming-pressure, is sensitive to all the details of the numerics, schemes and tolerances. I played with the 2D setup you provided and found that the results are difference partly due the the differences in the schemes and settings you are using for the case. Once I had set the cases equivalently I could analyse the changes between 2.2 and 2.3 which influence the detailed behaviour of the interface, particularly as it is projected up beyond the obstacle. These are 1. Improvements to ddtPhiCorr 2. Rationalization of the weighting used in fvc::reconstruct 3. Addition mixture.correct() in the sub-cycling to update the interface-normal for each corrector. On a wide range of cases and solvers the first two have proven beneficial and the third has proved beneficial for a range of VoF case. If you need to reproduce the behaviour of 2.2.x in 2.3.x you will need to revert the above three developments but I would recommend againt it. What would be more useful would be to run a sensitivity study on your case to changes in mesh, schemes and settings as I think you will find the pressure differences seen between 2.2.x and 2.3.x are within those due to these changes. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2014-07-21 13:32 |
|
New Issue | |
| 2014-07-21 13:32 |
|
File Added: interFoamBugReport.zip | |
| 2014-07-21 14:12 | henry | Note Added: 0003166 | |
| 2014-07-21 14:29 | henry | Note Added: 0003167 | |
| 2014-07-21 14:47 | henry | Note Added: 0003168 | |
| 2014-07-21 15:04 |
|
Note Added: 0003169 | |
| 2014-07-22 07:19 |
|
Note Added: 0003170 | |
| 2014-07-22 07:50 | henry | Note Added: 0003171 | |
| 2014-07-22 08:17 |
|
File Added: dambreak.tar.gz | |
| 2014-07-22 08:17 |
|
Note Added: 0003172 | |
| 2014-07-22 21:32 | henry | Note Added: 0003174 | |
| 2014-07-22 21:32 | henry | Status | new => closed |
| 2014-07-22 21:35 | henry | Assigned To | => henry |
| 2014-07-22 21:35 | henry | Resolution | open => fixed |
