open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

Simple request: can anyone give a quick look at my mesh for any quality issues?

I keep getting the "exceeding max iterations" message and I can't find out why. It definitely seems mesh-related, however.

I've adjusted the nodes using BlueKenue, checked for holes and any random asperities. I'm using the same .cas file for previous successful runs. The main change is that this mesh is bigger, and has some slightly awkward banks geometry.

I dropped the time-step to 0.1s from 0.4 in case it was due to the greater node count, but it still occurs. Below 0.1 would take a very long time and I don't want to mask any potential issues in the mesh.

Could someone please take a look?

Kind regards.

Ah, the file is still too large even when in a .zip folder. I've attached as .rar.

Hi,

In my experience the "exceeding max iterations" message is more related to the numerical schemes/solvers you choose. I did a quick run of your case in T3D and didn't get the warning message, so I think the mesh quality is fine. You can refer to the steering file settings of T3D example "Wesel" cause its bathymetry is quite similar to your case. Hope it helps!

Regards,
Yunhao

Thankyou. I tested my mesh on the example you suggested, however I still received the exceeding maximum iterations error, which is odd. I'd be interested to know if you did anything different: the key variables I changed were the .slf and .cli file (i.e. I used LargeChannel.slf etc).

I took a mesh with lower banks that worked previously and increased its bank elevation to match the one above and I received the same error, hence I think it still might be a mesh issue i.e. the triangles at the bank are overstretched and I need to rethink bank refinement strategies.

Hi,

Please check my test files on your case attached, not sure if it's the best solution but at least no 'exceeding maximum iterations' error occurred and the result looked pretty good. The original mesh was used and time step is 0.2s, more details can be found in the steering file. Could you try the settings to see if the error would still be there?

Best regards,
Yunhao

I ran it with your test files and it works, as you said.

However, it still fails when I try and use my fortran file - the file has worked with all my previous runs. Testing it in part, the run fails when I use the CALCOT routine to distribute layers and when I use SOURCE to apply head losses at certain layers.

What's odd is that even if I remove code to something simple .e.g

Which makes me think it's one of the numerical options in the *.cas file. I played around with the preconditionings but haven't had much luck. I additionally tested it with refined banks and had the same issue.

I'm still receiving the exceeding maximum iterations error.

It occurs when I leave my calcot subroutine in my fortran file. My user programmed routine builds four rigid layers at the bottom of the channel, creating an artificial bottom. This works for previous channels, however the key difference is the bank height - my new mesh has steeper edges with long, stretched triangles. I think the error occurs as Telemac tries to stretch my new bottom along these banks. Hopefully, refining the banks might help.

Otherwise, using Debugger I notice the the error occurs in CVDF3D for W velocities. What is it about CVDF3D for W that could cause it to crash, whilst the other velocities U & V don't crash?

If the other velocities don't crash, then hopefully there's a way to prevent W from crashing too...