open TELEMAC-MASCARET The mathematically superior suite of solvers

Dear Telemac-forum,

I have a question regarding morphodynamics at a tidal system using the online coupling between T2D and Sisyphe(v7p0r1). I use a diurnal tide at the eastern boundary and the hydrodynamics, including wetting and drying works quite well.

Regarding morphodynamics, I use Engelund-Hansen, option=3, (1 sediment class, 1 bed layer)for details see attached SIS-steering file.

The main problem I see is that although I see already formation of channels(see picture), across these channels, instead of getting a continuous eroded area, cells shifting between erosion and deposition from one cell to the other (marked by the blue and yellow cells) develop?

I tried it with and without TYPE OF ADVECTION = 14, with no notable changes.

Do you maybe know what could cause this problem and how to fix it?

Best Regards,
Christian

Hi Christian,

I have been having this problem for a while. I suspect that it is caused by the manner in which SISYPHE calculates the bed changes from the sediment entering the elements.

At the moment this is distributed based on the shape of the elements only, which causes some uphill transport ( an node which has nil bedload transport is still gaining material).

In the new release (I think it is from 7.2 onwards), this transport is stopped for dry elements, reducing the issue, but not solving it completely. Reducing your morphological factor will reduce the issue as well, but will not prevent it.

We are not the only model to have this issue. In general, the solution is to add a diffusion that creates downhill transport:

1. Some use an additional bedload transport S_bed,slope =-alpha*dz/dx that causes sediment to go down hill even if there is no bedload transport (the slope correction is insufficient, as this has no effect on nodes with no bedload transport.
2. Others use a bed smoothing algorithm, creating some diffusion that stops it.
3. I haven't tested it yet, but applying some (small) wave stirring should have the same diffusive impact.

The proper solution, however, would be to change the bedload material over the elements. But I haven't figured out how that can be done. Distribution based on the water volume distribution rather than the surface area of the elements? Or perhaps a distribution based on the ratios of the sediment transport?

Regards,

Michiel

Dear Michiel,

Thank you for the explanation.

I will try stopping the transport for dry elements and see if it improves the results.

Could you maybe point me in the right direction on how to apply the from you suggested "wave stirring" to diffuse the impact ??

Best Regards,
Christian

With wave stirring, I mean adding wave effects to the sediment transport, but that will be impossible with Engelund Hansen.

Hi Christian,
In the trunk version of the TMS v7p2, a correction of the subroutine solving bedload has been pushed. This modification enhances the treatment of wetting and drying processes when the solution of the Exner equation is performed.

Please consider to use also the new keyword to fix a threshold value: MINIMUM DEPTH FOR BEDLOAD. The default value for this keyword is 1.0E-2 m (new variable HMIN_BEDLOAD). This value can be modified by the user: as a guideline and after some discussions and testing, we can suggest a value in the range [2-3]d_50, being d_50 the sediment diameter.

Best wishes,

Pablo

Hey Pablo,

Thanks a lot, I will test it and keep you updated on the progress.

Best Regards,
Christian

Dear Pablo,

Thank you again for making the prerelease available.

I ran some tests to see whether various settings of MINIMUM DEPTH of BEDLOAD,would fix the problem, however unfortunately the problem presists. I also encountered that this variable might become problematic at channel edges(steep gradients).

Regarding the circular patches shown in the previous post, I think it might be linked to the way sisphye distributes sediment, i.e. not on an element base but on a "kind of" a varonoi neighborhood around a node. This seems to cause this cicular-erosion/sedimentation islands.

I am not sure whether there is a short-term solution, but maybe it could be something for future releases.

Best Regards,
Christian