open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,
Apologies in advance for a long post here. I am using Sisyphe uncoupled, reading in an existing Telemac2D file and calculating bed stress internally from the depth mean velocity and a drag coefficient Cd.

I am specifying in the Sisyphe steering file
1) LAW OF BOTTOM FRICTION = 5 (Nikuradse bottom friction law)
2) SKIN FRICTION CORRECTION =1 (calculate the skin friction factor mu from the grain diameter)
3) FRICTION COEFFICIENT set to a range of values (see below)
4) Single sediment class of 0.25 mm
5) Bed-load transport only (Van Rijn option =7)

From the Sisyphe documentation and looking through the code, it would seem that the resulting calculation should be independent of the value specified for FRICTION COEFFICIENT. This is because the skin friction correction mu is calculated as the ratio of a new grain related friction coefficient divided by the old coefficient (calculated from the value specified for FRICTION COEFFICIENT). When mu is applied to the total stress to calculate the bed load the old friction coefficient should cancel out, whatever its value. Therefore the rseults should not depend on the value of FRICTION COEFFICIENT.

I am not finding this to be the case. The effect it quite small but shows up in the net erosion and deposition magnitude when FRICTION COEFFICIENT is varied by an order of magnitude.

From a grep on the source code it appear that the total stress TOB without the mu modification is used in a number of places. I don’t understand enough about the code to know what these are. But it could be introducing an unexpected dependence of the total stress so that the results are not independent of the value of the FRICTION COEFFICIENT value.

I wondered if someone with more knowledge of the workings of Sisyphe could save me a lot of time in trying to understand this.

Thanks
John

Dear John,
thanks for your post. Probably, your confusion comes by the fact that when we run sisyphe internally coupled to telemac2d, the values of friction coefficient(s) are provided by the hydrodynamics module.
But, if you are running sisyphe in stand-alone mode, you must specify the keyword FRICTION COEFFICIENT in order to provide the friction value (as you did).
The SKIN FRICTION CORRECTION represents the ratio of skin and total friction (in general mu < 1 ), allows to take into account the effects of resistance by bedforms, for example.

I hope this helps,

all the best,

Pablo

Hi Pablo,
Yes, I am aware of everything you say but that is not my question. My question is that when SKIN FRICTION CORRECTION =1 is set the results should then be independent of the supplied FRICTION COEFFICIENT but they appear not to be.

I wondered if somebody who has a better undersainfo of the internal workings of the code than I, could explain why.

Thanks
John

Hi John, you are right... the skin friction is the main parameter for sediment transport applications.
However the total friction may still modify the transport rate, especially if you are working with suspended load. Then the Rouse number is based on total friction, and the relation between bottom concentration and depth aveeraged concentration depends on the total friction.
So are you working with suspended load?
Catherine

Hello Catherine,
Thanks for your reply. I am using only bedload transport.

I did a linux grep command on the code and there are a few places where the total stress is used - one is in the suspended load calculation as you say. The other appears to be in the wave current interaction routine. As I have no wave inputs I am assuming this is not playing a role.

There are a few other occurrences but I don’t understand what these do. Any guidance would be appreciated. This would help me understand why the total stress and not just the skin friction component is influencing the solution.

Thanks
John

Hi John, sorry for this long delay due to holidays...
I am still not sure where the problem comes from, I think that this should be a second order effect.
Are you using secundary currents, sloping bed effects?

Best luck, Catherine

Hello Catherine,
Thanks for your reply.

When running sisyphe in stand-alone mode with

1) the nikuradse formulation and
2) SKIN FRICTION CORRECTION = 1,
3) bedload only (SUSPENSION = NO )

my understanding is that the friction coefficient used in the skin friction calculation for the bedload formulae is calculated from the median grain size. Therefore the user supplied value (FRICTION COEFFICIENT) should be irrelevant.

However in my calculations I find that the magnitude of bed evolution differs in magnitude (but not position) by about 50% if the user supplied friction coefficient changes from 0.006 to 0.12. I expected there to be no difference.

The user supplied FRICTION COEFFICIENT value does affect the total stress. I wondered if the total stress enters into the bedload calculations in some way that I am not aware of ?

I am including the slope effect with

FORMULA FOR SLOPE EFFECT = 2

FORMULA FOR DEVIATION = 2

Any advice would be appreciated.
Thanks
John