open TELEMAC-MASCARET The mathematically superior suite of solvers

Dear all,
I have a question about the initial condition setup of TELEMAC-3D for the reservoir.
The configuration of my study is shown in Fig. 1.
Due to the downstream boundary (water level is 240 m), I set the "constant elevation" to 240 m.
However, Fig. 1 shows that the upstream boundary is the dry element and the error occurs.
In the forum, I found that some suggest setting a high water level for making every element wet.
After checking the results, the water surface is reasonable and it can successfully continue to event simulation (Fig. 2).
Nevertheless, Fig. 3 reveals that with the water level decreasing, still, some water remains on the bank side, which supposes to be dry.

Overall, there are two questions for asking for help:
(1) Are there any suggestions for the initial water level setting?
In the flooding season, It just needs to drop the water level from 245 to 240.
But, In the dry season, it should drop to 220, which will cost a lot of CPU time.
(2) How to solve the shallow water at the bank side?
I have tried the "dry element frozen" but it still cannot be solved.
Could you give me some suggestions?

Hello,

What is the type of your upstream boundary? Liquid or solid boundary? Do you prescribe something at this boundary?

What MESH TRANFORMATION do you use? How many planes do you use?
What kind of TREATMENT OF NEGATIVE DEPTHS do you use?

Can you upload your steering file, Fortran file if you have one and boundary conditions file please?

Chi-Tuan

Dear Chi-Tuan,
Thank you for the quick reply.

The upstream boundary is the liquid boundary and I use "LIQUID BOUNDARIES FILE" for time serious boundary conditions.

15 layers are used and the "MESH TRANSFORMATION" is set as 2 (Zstar).


The "FLUX CONTROL" is applied for "TREATMENT OF NEGATIVE DEPTHS" in this simulation.

The steering file is attached, please check it.

Hello,

In your Fortran file with subroutine USER_MESH_TRANSF, you should write ZSTAR%R(1)=0.D0 not ZSTAR%R(1)=0.0083D0
You did right with ZSTAR%R(15)=1.D0
ZSTAR has to start from 0.D0 and end to 1.D0

Looking quickly at your mesh, you have very flat triangles at your intake (where you prescribe flowrate for boundary #2). I think you should rebuild your mesh at least at this location. Try to avoid overconstrained elements also.

After changing these 2 things, can you tell us if you still have the same issue?

Last remark, after you solve your issue: you use PSI scheme for tracers (SCHEME FOR ADVECTION OF TRACERS = 5 and SCHEME OPTION FOR ADVECTION OF TRACERS = 1). If you have dry elements, it should not work. I would rather use an advection scheme which works with tidal flats, i.e. NERD schemes (SCHEME FOR ADVECTION OF TRACERS = 13 or 14) or LIPS (SCHEME FOR ADVECTION OF TRACERS = 5 + SCHEME OPTION FOR ADVECTION OF TRACERS = 4)

Chi-Tuan

Dear Chi-Tuan,
Thank you for the kind reply.
I have revised my Fortran file and mesh (Fig. 1).
Also, the SCHEME FOR ADVECTION OF TRACERS is changed to 14.
But the shallow water at the bankside still exists.
Though I revised the mesh and it seems reasonable, I found that the water level is still unstable at the inflow boundary (Fug. 2).
Does it mean that my mesh at the upstream boundary is still of low quality?


In illustrated in Fig. 3, the shallow water, where should be dry elements, lead to the extremely high bed shear stress and sediment stuck at that element.
I believe that those elements are supposed to be dry and without any suspended sediment. So if it is possible, anyone could help me to check the missing setup?
The revised steering file is attached, if it is available, please help me to check my setup.