open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

we are doing sediment transport simulations with Telemac-3D and Sisyphe internally coupled.
Some parameters:
- coupling period = 1
- law of bottom friction: Strickler
- horizontal turb. model: constant eddy viscosity (0.001)
- vertical turb. model: Prandtl
- or k-e turbulence model

Usually we prepare a steady flow field with Telemac3-D and use this for the further computations with Telemac-3D & Sisyphe.
Sisyphe calculated deep streamwise furrows in some cases.
By comparing the total bed shear stresses calculated with Telemac-3D (last time step from steady file) and Telemac-3D coupled with Sisyphe (first time steps), see figure, we noticed that in the sediment transport simulation the bed shear stress looks quite different and not physical. The pattern would explain also the calculated furrows.

What is the reason for this difference?
We would be thankful for any advice.

Best regards,
Gabi & Clemens

Hello,

The bed shear stress is recalculated in Sisyphe based on the hydrodynamic friction velocity. Results should be perfectly identical, except if the two simulations are different (one assuming fixed bed, the second one, with bed evolution). My guess is that bed evoulutions explain the differences in the hydrodynamic/sediment transport model results. For a given flow rate, an increase of bed shear stress is expected when sediment gets deposited due to reduced water depth.


Best regards, Cath

Hi Gabi & Clemens,
to complete Cath's remarks, I'd like to add that T3D computes U* (friction velocity) as a function of the velocity at the first vertical point above the bottom and the distance from the bottom to this point, therefore the calcul of U* could be (also) strongly influenced by the vertical discretization of the mesh.

All the best,

Pablo

Hello Catherine and Pablo,

thank you for your replies and sorry..it was our mistake.

In the first picture we used a constant roughness and in the second picture we used a variable roughness (channel / banks).
The rougher banks produce higher shear stresses, so in the transition from the bank to the channel we got this discontinous shear stress pattern.

In the meantime I did some steady simulations with a idealized 180° bend and noticed that the number of layers influences strongly especially along the symmetry axis in the bend the shear stress distribution and shear stress magnitudes.


Best regards,
Clemens