open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

we are still modelling reservoir flushing using Sisyphe coupled internally with Telemac3D. Because of the calculation routine of the skin friction correction we would like to use a smaller active layer thickness and not the default value to take the changes in the skin friction due to sediment transport in the flushing process into account.
The simulations are generally running well by using the default active layer thickness but when we switch to CONSTANT ACTIVE LAYER THICKNESS = NO, then we get the error message "ERROR IN LAYER CASE 1" until approx. 3600 time steps (0.5s).
If we change the conditions to a constant active layer thickness of 0.1m, we get the same error message after approx. 13h of simulation.
By changing the active layer thickness to 0.8m, I got the following error message after 45900 iterations (1s):
TOO MUCH EROSION AT POINT J= 698
DECREASE DT OR INCREASE ELAY0
EVOL= -0.800000023995342 ELAY0= 0.800000000000000
I checked the node and the total erosion in the 45900 iterations were -0.04986m. So I am wondering what happened in this special time step were the error message occurred….
The hydrograph is very smooth, the time step is small, how can the erosion rate in this time step increase at this magnitude?

Many thanks,
Gabi

Dear Gabi,
In my opinion, you should probably try with some non-constant active layers, as the one proposed by Van Niekerk et al. (see, e.g., Journal of Hydraulic Research Vol. 48, No. 6, 2010), and computed as a function of the shear stress. Van Rijn (1983), Wong (2006), etc. also proposed similar active layers formulae based on the same criterion.

I hope this helps,

all the best,
Pablo

Dear Pablo,
thank you for the answer. Do you know if somebody has already coded this function?

But there is still a problem in my case. The skin friction correction is computed taking into account the active layer thickness. In my case I have lateral distributed graded sediments in a reservoir with a lot of fine sediments in front of the weir. When in case of flood coarser sediment is transported through the reservoir, the skin friction correction plays an important role. If I would use a non-constant active layer with is linked to the bed shear stress now, the skin friction correction would also be connected to the bed shear stress.

And that would mean in the case of coarser sediment entering the domain, the skin friction correction would be higher if the shear stress is higher. I think this would not be correct in terms of skin friction correction.

In our case the skin friction correction is between 0.8 and 0.16 – that means this factor has a very big impact.
This leads me to my main question – why is Telemac stopping the computation if the EVOL is bigger than ELAY? Why is it not updated with ESTRAT in this case?

Best regards
Gabi

Hello again,
> thank you for the answer. Do you know if somebody has already coded this function?

for the new release of telemac (v6p2), a new vertical grain sorting model (developed at BAW) will be incorporated. This model compute the active layer in several ways. At this moment, this functionality will be only available for the new model, but you can get some inspiration from there (but you have to be a little patient)

Regarding the second part of your post, I'll take a closer look.

best,

Pablo

Hello Pablo,

I'm also trying to use the vertical grain sorting model in our model. As you mentioned above, is this new vertical grain sorting model available now? How to use it in version 6.2? Because I didn't find any information about it in the manuals.

Hi Qilong,
the new vertical grain sorting model will be presented tomorrow at the TUC by the BAW. A documentation and step-by-step example will be available soon.

all the best,

pablo

Hello Pablo,

Thank you for the good news!
I'm looking forward to it!

Best regards,
Qilong