open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

I'm performing some dam-break simulations with Telemac-2D & Sisyphe.
The main configuration parameters are as follows:
dam breach is beeing initiated by an initial breach channel, only non-cohesive bed load transport, Meyer-Peter and Müller formula, one grain size, skin frction correction on, sediment slide on.

Analysing the time dependent breach process or plotting the breach hydrographs for different grain sizes gives me some unexpected results. The breach process is faster with a larger grain size, see enclosed the picture.
I don't have an explanation for this behaviour and cannot figure out what may be the reason.
Anyone has some ideas? It is the skin friction correction or something with the sediment slide algorithm?



I would be very happy for any suggestions.
Clemens

Hello,

Partial answer on sand slide side: the sand slide algorithm first assumes that an equilibrium is reached instantaneously in every element, then "mass-conservatively" suppresses the discontinuities thus created at interfaces between elements and at points, all this regardless of the grain size, so it is doubtful that it can explain what you find. I would rather look at skin friction. Anyway you can remove the sand slide effect to be sure.

With best regards,

Jean-Michel

Hello,

yes, indeed one reason is the skin friction correction. Another reason is the critical Shields parameter calculation which for very low grain Reynolds numbers or dimensionless grain diameters gives very high critical Shields values.
However, it is interesting that in this case the variation of some parameters has such a massive impact on the dam breach development.

Best regards,
Clemens

Hello dear Clemens,

I am also working on dam breaching by coupling 2D Telemac and Sisyphe. I tried with different bed load methods, I found that the overtopping hydrograph result from Van Rijn is twice higher than the MPM method. Another interesting things which I tried is when using the same parameters but different method, for instance, when I did simulation when using FEM is totally different from FVM.

Thanks,
AMANJ