open TELEMAC-MASCARET The mathematically superior suite of solvers

Dear all,

I am trying to simulate propagation of suspended sediment over a hump in telemac 2D coupled with sisyphe. The bed bathymetry is given in the Geom_GaHump.slf file attached. The ultimate goal is to apply pseudo-periodical conditions in the lateral boundaries for the suspended sediment, so that the hump propagates the same way along the channel’s width. This is easily achieved for the flow by applying no friction in the lateral boundaries. The boundary conditions for the flow are prescribed Q in the inflow and prescribed H in the outflow. The side-walls are solid boundaries (2), however, I ask for no friction at the lateral boundaries (0) which for the flow it seems that it works. For the boundary conditions file of the sisyphe, I only restrict the inflow with using no bed evolution (5) there. For the side walls I tried either 2 or 4 but there was no different effects. The SUSPENSION switch in sisyphe is activated and the equilibrium concentration is computed using the model by VanRijn:
SUSPENSION : YES
REFERENCE CONCENTRATION FORMULA : 3
EQUILIBRIUM INFLOW CONCENTRATION: YES
The solution should not give any different behaviour in the spanwise direction (perpendicularly to the flow direction). However it does. I was trying to figure out which of my choices may cause this effect and how can I correct this. I am attatching the resources files with the .cas and .cli files for the simulation I attempted.

Thank you in advance,
Efstratios

Hello,
I'm out-of-office now but I'll try to give some (fast) hints:

1) your BC files for T2D and SIS are different. Why? Furthermore, for your outflow boundary 5 4 4 you've set a wall bc for the sediment transport. Also, to avoid inconsistencies, your wall BC (for T2D) must be also set = 2 for the 8th column.

2) Use another TYPE OF ADVECTION scheme, e.g. the numbers 14 (EDGE-BASED N-SCHEME) or the 15 (ERIA scheme)

3) Take off the keyword DISPERSION ACROSS THE FLOW. Then play around with that.

4) Start your simulation from a hotstart initial condition (e.g. solve your hydrodynamics until steady state, then use the solution as initial condition for your hydro-sediment transport problem.

Hope it helps,

Pablo

Dear Pablo,

Thank you so much for the directives you provided.

1. The reason I chose to have different boundary condition file for sisyphe is because I tried different boundary conditions on solid boundaries to see if the previously mentioned issue would be tackled. Which could not solve the case.

2. different advection schemes have been tested including 14 that gave an quite unstable result and 15 for which the simulation crashed. The most stable and logical solution is taken when the advection scheme is set to 1 (method of characteristics).

3. I did set the DISPERSION ACROSS THE FLOW equal to zero but still didn't change the result.

4. I am certainly initiating my simulation from a converged fluid flow field. This along with the results I get lead to the conclusion that it may be boundary condition related issue.

If you see from the screenshots submitted, it seems like on the walls there is no bed morphology computed even if the converged flow field gives same velocity as in the middle of the channel (since friction on lateral boundaries is 0). The behavior seems like the inflow where I requested zero bed evolution by assigning value 5 in the 8th column of the .cli file.

Any additional information on this matter would definitely help.

Stratos

Hi Efstratios,
Please try a simulation with only one BC file (e.g. the one you used for T2D), modified accordingly to account for sediments also. See then what's happening...
As I don't have you T2D file, which time step are you using?

If you still have problems, please include ALL the files and I'll take a closer look.

Best wishes,

Pablo

Dear Pablo,

I have tried different options for the proposed parameters but the issue is still here. My timestep is ensuring a CFL = 0.3. I am attatching the source files along with the cas file for telemac2d (T2d_GHump) I forgot to add initially.
Thank you in advance

Stratos

Hello,
Please try :
TURBULENCE MODEL = 1
DIFFUSION = NO
CORRECTION ON CONVECTION VELOCITY = NO
TYPE OF ADVECTION = 14
SLOPE EFFECT = NO

and set the wall BCs for sisyphe as follows (for all nodes on the lateral walls):
2 2 2 0.000 0.000 0.000 0.000 2

See how the model behaves with this setup. Then you can modify them, as well as the mesh density.

Hope it helps,

cheers

pablo

Hi,
I am trying to run the bedload and suspended load together. I have my bedload size class is 0.0004 m, 0.0005, and 0.0006 meters and the suspended load solid boundary size is 0.0001 meters. My question is how can define the different sizes in the same cas where the suspended load boundary will be one size 0.0001 meter and the bedload will be three sizes, all are silt?
Should I use one-grain size which works for both suspended and bedload as my grain size is all silt?

Thanks,