open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello dear TELEMAC users,

I am working on a source that can occur (depending on the tides) in dry areas. I noticed that when my simulation starts at low tide (with the source located in a dry area), nothing happens—there is no discharge at all. However, when the simulation begins with the source submerged, it behaves completely normally, even if the sea level later decreases and the source ends up in a dry area. Moreover, when starting at low tide, the source begins functioning as soon as it becomes submerged.

I searched the forum to see if anyone else has encountered this issue but couldn’t find much information. I also tried running my model with different advection schemes and tidal flat treatments, but the problem remains the same.

As a workaround, I’ve managed to bypass the issue by starting the simulation with a constant elevation higher than the source height. While this causes instabilities during the first few hours of the simulation, it eventually converges.

If anyone has any insights or suggestions on how to solve this problem, I’d greatly appreciate your input!

Have a great day,
Thomas

Hello,

Which release do you use?

What is the bottom elevation of the node where you want to put the punctual source? Is it 1.9 m? It is not recommended to put a punctual source on the bottom (bottom should be impervious).
Rather than this feature, you should try OPEN BOUNDARY CONDITIONS ON THE BED and PRESCRIBED FLOWRATES ON THE BED keywords (but not available for tracers).

Chi-Tuan

Hi Chi-Tuan,

Thank you very much for your answer and I apologize for my late response.
I am currently working on v8p5r1, and the bottom at the node of the punctual source is 1.88... m. I tried to modify this value (the elevation of the source) (above and under) but it doesnt seem to change anything to my problem.

I then followed your advices putting an open boundary conditions on the bottom, using the user_limi3d.f from the bottom, and there is some improvements, but It seems that I am now having some stability issues.
When using the MURD + PSI sheme for velocities the max number of iteration is reached at the first time step. Similar issues when using the characteristics.

As I am not far from my lateral boundary, a solution that's also possible in my case is to put a new lateral boundary with an input flow + tracers. But I would still like to solve this issue and find what is going on.

Have a great day!

Thomas

UPDATES

I managed to create a minimal test case (a simple slope in a rectangular domain), and the method using sources exhibited the same issue: no flow from the sources when the source point is dry.

However, the method using an open boundary at the bottom worked well. That said, if this method cannot be applied to tracers, it won’t be a viable solution for me. Additionally, based on the few test cases I ran, it seems that using a bottom boundary significantly increases the computation time.

Finally, I adapted my initial 3D test case to 2D, and in 2D I didn’t encounter any issues with sources in dry areas.

If you have any new insights into this issue, please let me know. Otherwise, I will continue bypassing the problem as described in my first message.

Thomas