spherical coordinates do not work in TELEMAC-3D

mentioned in merge request !175 (merged)

Hi @alexander.breugem. Where is your model located ? I have a similar behaviour for a global 3D model.

In case you want to check: forum_T3D.zip

High velocities propagate from the south/north pole. I think I also had a problem with the prisms crossing the dateline (+/- 180)

For your problem it could be that the distortion caused by MERCATOR projection needs to be corrected in real meters in T3D.

I had a look on the forum and it has also been reported here with a lot of details (in november 2023)

It's been a while I haven't looked into this problem, but I had narrowed it down to the calculation of

MESH%SURDET
MESH%SURFAC
STOSEG

in inbief.f being completely wrong

Hi Thomas,

The issue that I found, was that in TELEMAC-3D the calculation of latitude or longitude (i forgot which one) was done twice, leading to nonsensical values in mesh3d. That is the bugfix that I proposed, and with that fix, our models works well. However, I don't think we have done models that cross the dateline. So there might be another issue as well. I will try to have a look at it soon.

mentioned in commit 9f751c1d

mentioned in commit c4e39fa1

Fix pushed in SHAs c4e39fa1 and 9f751c1d

closed

mentioned in commit 0e17f354

mentioned in commit c7084ce1

mentioned in merge request !190 (merged)

mentioned in commit f18e94a2

mentioned in commit c4ea0a19

mentioned in commit 76243558

Hi, I understand this issue has been closed along with a MR that had even a test case to compare between 2D and 3D!

However, I still had problems at polar latitudes (and also for triangles across the +/- 180° meridian), which was fixed by modifiying these 2 routines:

flupri.f: I think the sustainable way of doing this, would be to use XEL/YEL in FLUPRI (because I am essentially re-doing the fix in LATITU for SPHERICAL COORDINATES = YES)
inbief.f: here I am just shutting off the SURDET calculation, but I don't even think it needed in the end, is it?

(full story on the forum)

This fix might also be a element of solution for a long standing issue #321 (closed)

mentioned in commit 246579fa

spherical coordinates do not work in TELEMAC-3D

Designs

Child items ...

Activity