View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0000750 | OpenFOAM | Bug | public | 2013-02-18 15:34 | 2013-02-18 22:17 |
| Reporter | chegdan | Assigned To | henry | ||
| Priority | normal | Severity | minor | Reproducibility | always |
| Status | closed | Resolution | no change required | ||
| Platform | Linux | OS | Ubuntu | OS Version | 12.04 |
| Summary | 0000750: localMin does not report correct value of local min between owner and neighbor cells in certain cases | ||||
| Description | localMin does not report correct value of local min between owner and neighbor cells of a face, if values are equal in magnitude but opposite sign. I have attached a snippet of code and test case to reproduce the error. Both numbers reported in the snippet should be equal to 40, which is found by inspecting paraview for the L field and counting the faces straddling the two regions of opposite sign. | ||||
| Steps To Reproduce | compile the code with wmake. Run the "box" test case with Allrun, where the script runs blockMesh, setFields, and localMinTest executables. It should report two numbers that should be equal. | ||||
| Additional Information | in localMin.H, line ~158 forAll(vff, facei) { vff[facei] = minMod(vf[own[facei]], vf[nei[facei]]); } where the line might want to read forAll(vff, facei) { vff[facei] = min(vf[own[facei]], vf[nei[facei]]); } After this change, i get the expected number of faces. When checking the min of owner neighbor cells on coupled boundaries, the method uses min instead of minMod as expected. Unless I'm missing something, the function min should be used on internal faces instead of minMod. | ||||
| Tags | No tags attached. | ||||
|
|
|
|
|
Would it help if we called it localMinMod or localMinMag rather than localMin? |
|
|
why not have another scheme for localMinMod and alter the behavior of localMin to be the expected behavior? Was there a particular reason to have it this way (reaching back years ago when it was programmed)? |
|
|
Yes the reason is that the minimum value of interest in this scheme is the minimum of the magnitude, I am not sure if the minimum of the signed field would be useful in the context of surfaceInterpolation. |
|
|
the only situation I can think of is when you would want to identify the face that straddles an interface in a level-set field. each side of the face could be equal in value but opposite in sign. Clearly, localMin would not be able to identify some of the faces. Thanks for your comments. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2013-02-18 15:34 | chegdan | New Issue | |
| 2013-02-18 15:34 | chegdan | File Added: localMinTest.tar.gz | |
| 2013-02-18 15:44 | henry | Note Added: 0001923 | |
| 2013-02-18 15:57 | chegdan | Note Added: 0001924 | |
| 2013-02-18 16:23 | henry | Note Added: 0001925 | |
| 2013-02-18 20:58 | chegdan | Note Added: 0001926 | |
| 2013-02-18 22:17 | henry | Status | new => closed |
| 2013-02-18 22:17 | henry | Assigned To | => henry |
| 2013-02-18 22:17 | henry | Resolution | open => no change required |
