open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi everyone,

I've recently conducted a series of simulations to model the formation of braiding in a rectangular channel with a small inlet channel to soothe the inflow using Telemac coupled with GAIA. However, I am facing various problems despite following tips I found in the forum.

Here is the setup I used:
Fortran File: USER_BED_INIT.f with the variable RIGID BED in the geometry file, which avoids erosion in the inlet (Definition of the variable: Channel = -10, Inlet = Bottom Elevation).
Velocity Profile: 4;4.
Geometry: Extended inlet, resampled inlet nodes (0.007), smooth bed with 3% slope, inlet without slope, small central incision of 2 cm.
Previous Computation File: Steady state res_hydraulic.slf to start the sediment calculations.
Time Step: 0.005 s.

Trial 1: Coupling with GAIA
GAIA BC at the Inlet: 4 5 5 1.0 0 0 0 4 ... (BC_gaia_0507.cli)
Results: Strong sediment deposition at the BC nodes inlet, leading to increased velocity at the inlet due to flow from only a few non-deposited nodes.

Question: How can I reduce the sediment deposition at the inlet nodes in the coupled simulations to allow a correct flow?

Trial 2: Coupling with GAIA and Equilibrium BC
Equilibrium BC at the Inlet: 4 4 5 0 0 0 0 5 ... (BC_gaia_050702.cli)
Results: Sediment deposition at the inlet is reduced, with only a few nodes raised. However, a sediment sill forms at the end of the inlet due to deposition and no sediment outflow is observed.

Question: Has anyone encountered and fixed the issue of "no sediment outflow"? How can I address this problem?

I did attach all the files.

Thank you for your help!
Best regards.

These are all files used:



(unfortunately, the previous hydraulic computation res_hydraulic.slf is too big, so i cannot upload it here)

Hi Hannah,

for simulating lab models we usually use Nestor for the sediment input. So we use your first GAIA boundary conditions but set the input discharge in gaia-cas file to zero. The sediment is then put by nestor on the rigid bed. With the deposition polygon you can distribute the input sediments in order to avoid big sedimentations. Usually a larger input area is better but low shear stress areas should be avoided.

If you don't want to use nestor you can give the distribution of the sediments for the inlet boundary in the 4th column of cli-file. For nodes which have too much deposition you can decrease the value smaller than 1. All nodes = 1 means at each node the same sediment input is given. But as you use velocity profile =4 (the velovities distribution is dependent to the water depth) the shear stress is not the same at each node which leads to less sediment transport at nodes with small shear stress and maybe to deposition.

I cannot say something about sediment not reaching the outlet. Do you have at least a small amount of sediments at the inlet (e.g. 1mm). I am not sure if the boundary condition will work without any sediments at the boundaries. Maybe this will help.

Be aware that in your user_bed_init.f an exception is coded if you use nestor (rigid bed for all nodes with x values less than 1.75). I guess this comes from another case and you should remove it if you use nestor.

I see that you don't use skin friction correction. This means that you use the friction coming from telemac. Depending on your chosen friction coefficient this will lead to bad results. I would strongly recommend to use skin friction correction=1. I don't know why there is your comment that you should use no skin friction correction in case of small water depths. I don't know why this should be the case and would either recommend to play around with MINIMUM DEPTH FOR BEDLOAD or MINIMUM VALUE OF THE WATER HEIGHT.

Furthermore we use FINITE VOLUME in Gaia. It seems that the results are usually more stable. In that case you can use UPWINDING FOR BEDLOAD > 0.5 up to 1. For Telemac we still use finite elements.

It would be interesting to hear if some of my proposals help (or not) in your case.
Best regards,
Rebekka

Thank you very much for your answer.
I have tested both of your suggestions, but I am afraid that modifying the distribution of sediments at the inlet nodes did not achieve the desired result (still deposition at the inlet nodes).
The use of NESTOR did, however, prove an effective solution to the issue of deposition at the inlet nodes, while also enhancing the hydromechanics.
Please refer to the attached files (the geoemtry file and BC file for the hydraulic simulation & the previous calc. are unchanged). Unfortunately i cannot upload the results file.

Unfortunately, we are still experiencing an issue where no braiding is forming.
All parameters are based on a study by Rifai (2014) "Numerical Study of Braiding Channels Formation", in which this scenario is coupled with SISYPHE.

As for now there are no changes in the bed level elevation in the channel, with only some little erosion and deposition at the first few centimeters after the inlet.
We have already attempted to increase the discharge rate (up to twice the original rate) but this has not resulted in any significant changes to the bottom elevation.

Do you have any idea why this could be, and what enhancement could be made, given that this configuration did produce braiding in the Rifai study?

Thank you in advance for your assistance.

Hannah

Dear Hannah,
there are some thresholds in the code.
MINIMAL VALUE OF THE WATER HEIGHT = 1mm per default and
MINIMUM DEPTH FOR BEDLOAD=1cm per default
I would recommend to set both variables to smaller values or if you have no stability issues to zero.
I guess that the last one is responsible for your missing bedload as your initial condition is 2 mm water depth.



Best regards,
Rebekka

Thank you Rebekka!
Now it is working as intended!