View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0003505 | OpenFOAM | Bug | public | 2020-06-04 05:51 | 2020-06-09 09:00 |
| Reporter | SamMallinson | Assigned To | will | ||
| Priority | normal | Severity | minor | Reproducibility | always |
| Status | closed | Resolution | no change required | ||
| Platform | Linux | OS | Ubuntu and Debian | OS Version | 18.04 and sid |
| Summary | 0003505: lineCell sampling changes detected cells in version 7 | ||||
| Description | This is a side issue from https://bugs.openfoam.org/view.php?id=3504, which had the example of a planar 2D channel with one wall attached to a heated plate When going from version 6 to version 7, the number of points detected using the lineCell method changes. The mesh used is very simple: a rectangular prism with 100x15x1 cells, with the domain split in the y-direction with the upper 10 cells being the fluid (liquid) and the lower 5 the solid (heater). In version 6, 15 y-locations are output, whereas in version 7, the cell in the heater layer adjacent to the liquid is not detected. See https://bugs.openfoam.org/file_download.php?file_id=2987&type=bug and https://bugs.openfoam.org/file_download.php?file_id=2989&type=bug I think this might have arisen due to changes in the lineFace code, which is #include'd in lineCell.H, but I could not identify the commit where this change happened. | ||||
| Steps To Reproduce | For version 6: - download chtPlate_of6.tar, see https://bugs.openfoam.org/file_download.php?file_id=2991&type=bug, un-tar it, call foam-bash-6 (. ~/pathTo/OpenFoam-6/etc/bashrc) then ./Allrun For version 7: - download chtPlate_of7.tar, see https://bugs.openfoam.org/file_download.php?file_id=2990&type=bug, un-tar it, call foam-bash-7 (. ~/pathTo/OpenFoam-6/etc/bashrc) then ./Allrun NB: the plotting routine in the folder plots requires python3 with numpy pylab os and sys installed. You can always just inspect the values in postProcessing/singleGraph0/liquid/<iteration>/line_T.xy and postProcessing/singleGraph1/heater/<iteration>/line_T.xy, where <iteration> is the number of iterations to converge which may vary from system to system. | ||||
| Tags | No tags attached. | ||||
|
|
The heater graph line is exactly on a boundary, and the start and end points are exactly on an edge. Sampling lines have to unambiguously intersect the cells, or what points you get out aren't well defined. The fact that there's a difference between versions is just a fluke. A start and end of (0.9999 -0.1001 5e-3) and (0.9999 -0.0499 5e-3) works fine. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2020-06-04 05:51 | SamMallinson | New Issue | |
| 2020-06-09 09:00 | will | Assigned To | => will |
| 2020-06-09 09:00 | will | Status | new => closed |
| 2020-06-09 09:00 | will | Resolution | open => no change required |
| 2020-06-09 09:00 | will | Note Added: 0011388 |
