open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi There,

Just wondering if there is any user documentation regarding the SEDI-3D module in TELEMAC-3D? I noticed there is a French example called 'rouse' but is this it?

Regards,
Toby

Hi Toby,

Actually, there is no user documentation regarding (specifically) SEDI-3D.

Regarding "Rouse" example, it is a test that validates the modeling of the hydrodynamics and non-cohesive suspended sediment
transport (in 3D). You can find it into validation folder of Telemac3D.

Cheers,
José Andrés.

Hi Toby,

I had the same problem when I started to work with SEDI-3D. Then I figured it out by looking at the main subroutine telemac3d.f.

Inside this file, I searched for the keyword SEDI and I found a lot of places where this keyword was implemented.

By looking at those lines of code, you could have an idea how it works and the structure of SEDI-3D module.

Kind regards,
Qilong

Hi Qilong! Since you seem to have some knowledge within the area, I'll have to ask you two questions:

1. Is the unit of 'SEDIMENT' as a tracer equal to kg/m3?
2. Is it possible to release sediment with two different characteristics?

/schmirre

Hi Schmirre,

In SEDI-3D, the unit of sediment concentration is indeed kg/m3. We don't have to convert it from volume concentration.

As far as I understand, in the v7p2r1, it can only deal one class of sediment, either cohesive or non-cohesive in 'non-mixed' case. The last tracer will be used as sediment.

If you set the keyword mixed-sediment to yes, it can deal with one class of sand and one class of mud. In this case the last two tracers will be recognized as sediment.

It cannot deal with multiple classes at this moment but it is not so difficult to expand the code. Generally speaking, if you copy and paste the part of the code in SEDI-3D, it will solve the advection-diffusion equation (in CVDF3D) for more sediment classes. What you have to do is to give the right boundary (in CLSEDI) and initial conditions to them.

It sounds easy but requires some efforts...

Kind regards,
Qilong

Wow, thanks! That helps a lot, and explains my extreme bed evolution, hadn't realised that the last tracer will be sediment.

I am simply modelling the area of influence from a polluted sediment source of two different grain sizes, so perhaps a more modest approach would simply be to perform two identical runs with different sediments.

Thanks again!

/Schmirre

Your solution sounds good.

If you run two identical runs with different sediments, it seems no problem for looking at the transport of suspended sediment since they don't interact with each other, and the influence to the flow is negligible if the concentration (vertical concentration gradient may contribute to the buoyancy effects in turbulence model, depends on the turbulence model you use) is not very high.

But for the bed evolution, you'll lose a fraction of sediment that should deposited to the bed. And this will influence the bathymetry, hence, the hydrodynamics. This could also be negligible if the bed changes is not dramatic.

Good luck with your model!

Qilong

All of the above it a great help getting started with this.

I am having a bit of an issue with getting convergence on the sediment diffusion. I believe that this is related to the fact that I am modelling it as part of an outfall source which is located on a shallow shelf. As I have a model with 7 layers and water depths can get become very low at times, compressing the layers and causing non-compliance with the CFL of the fall velocities (as mentioned in previous forum posts)

Would like to hear your thoughts on the best way to tackle this or whether its best to just move the source to deeper waters.

Could it have to do with the following keyword (from telemac3d.dico):

NOM = 'SEUIL LIMITE POUR EROSION SUR BANCS DECOUVRANTS'
NOM1 = 'THRESHOLD FOR SEDIMENT FLUX CORRECTION ON TIDAL FLATS'
TYPE = REEL
INDEX = 36
MNEMO = 'HSED'
TAILLE = 0
DEFAUT = 0.2
DEFAUT1 = 0.2
RUBRIQUE = 'PARAMETRES NUMERIQUES'
RUBRIQUE1 = 'NUMERICAL PARAMETERS'
NIVEAU = 1
AIDE = 'Pour les profondeurs inferieures a cette valeur limite, la viscosite sera
progressivement reduite. Voir le sous-programme FLUSED'
AIDE1 = 'Below this limiting depth, all sediment erosion rates are set to zero.
See subroutine FLUSED'

Hi Toby,

When the water depth gets very low, you might get very high velocity on the shallow areas, with the high velocity (it exerts high shear stress on the bed), you might encounter instability in solving sediment advection-diffusion equation. But I'm not sure if it is your problem...

In my model, I use

TIDAL FLATS = YES
OPTION FOR THE TREATMENT OF TIDAL FLATS = 1
TREATMENT OF NEGATIVE DEPTHS = 2
MINIMAL VALUE FOR DEPTH = 0.1

to deal with tidal flats. The MINIMAL VALUE FOR DEPTH = 0.1 ensures the location where the water depth is below 0.1m could get corrections.

As suggested by Schmirre, I also use THRESHOLD FOR SEDIMENT FLUX CORRECTION ON TIDAL FLATS = 0.1. So it means no erosion happening when the water depth is below 0.1m.

If you put the outfall sources in the model, I remember I read it somewhere, they should not be on the bottom plane since the bottom plane is imposed as boundary conditions. In your case, indeed it can cause such problem.

I don't know how to tackle this problem at this moment. Did you include TREATMENT OF TIDAL FLATS? Maybe try the OPTION FOR THE TREATMENT OF TIDAL FLATS = 2? This will remove the dry locations from the computational domain, but I'm not sure if it will help...