open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello, i have some velocity distribution and stability problem on entrance of my river model.
However, all seems to be ok:
- i have a flat bathymetry on the first range mesh
- i think i have enough liquid boundary nodes
- flow on entrance is only 1 m3/s
- liquid boundary option is set to 1 (not better with 2)
- velocity profile is set to 1
- Strickler coefficient is set to 10 on first range mesh

As someone an idea of what i should do in order to have inflow more uniformely distributed and more stable??
Thank you in advance!

Hello Mathieu,

What about the bathymetry close to your inlet? Do you have steep slope? If yes, you may have to smooth the gradient of bathymetry in your geometry file.
Do you have different friction coefficients in different areas? What about them close to your inlet? You only mention a "first range mesh".
Have you tried to decrease the time step?
What have implemented in the Fortran file?

Hope this helps,

Chi-Tuan

Hello, thank you for your answer.
- Friction coefficient is set to 30 in bed and 20 on overbanks
- bathymetrie is flat in the bed (see illustration t100)
- not better with less time step or without fortran file

z instability at the entry of the model :


velocity instability :





I dont remember same kind of problems with older telemac versions!?

Thank you in advance!

Same problem on my second entrance :




Velocity is not uniformely distributed on boundary entrance

Hello Mathieu,

Is the profile of bathymetry exactly flat along your 2 liquid boundaries (constant bottom elevation)?
It seems that water depth may be below your last threshold (below 1 cm?) in your figures. It is quite shallow.

I can see uniform profiles of velocity when water depth is bigger than 1 cm along the 2 liquid boundaries.
You can try to change the velocity profile to 4 or 5 and see if inside the domain, the results are improved.

Chi-Tuan

Hello,
Yes bathymetry is exactly flat along liquid boundaries.
So the best thing i done to improve velocity distribution is to set Strickler friction coefficient to 2 on entrance!
Thank you for your help

Hello Mathieu,

Using 2 m^1/3/s for Strickler coefficient is quite low when using TELEMAC-2D, there may also be other issues with your model. But if it is OK for you to solve your issue.

Chi-Tuan