open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi all

(I tried to post before the problem but something went wrong, I hope to not double post).

For my thesis project, I have to study the hydrodynamic of a lagoon with Telemac 2D.
The model is forced at the boundary with conditions obtained from the CMEMS hydrodynamic model of the Mediterranean Sea.

The geometry is shown in the picture attached (see Bottom, Geometry and Boudnary).

As you can see from the pictures I'm forcing the model with sea level and fluxes at the three "sea" boundaries, while the river boundary is forced with imposed flux.

I managed to run the model, but the results are not good.

The problems are (excepting after some "normal" wiggles due to the starting shock):

1)after many time steps some instabilities occur and remain till the end of the simulation (see picture attached). I think that these are due to the option of the boundary condition (maybe thompson for the lateral boundaries would help?) or it's something difference?

2)It seems to me that at certain points of the coastline it happens something like a 'mass-adding', this is due to the treatment of the negative depth/treatment of tidal flats or it's a matter of wiggles due to the instabilities?

3) the rough results in the open sea area are due to the mesh size in that zone or it's something else? (maybe a quasi-bubble discretization would help?)

I tried different scheme of advection (for velocity and k-epsilon):

13 ,5 and 14 (for both of them, the latter seems to be the best, but is very far away from a stable solution).

I used only these schemes because I setted TIDAL FLATS =YES (actually in my area of studies the tidal has not a great effect, but there is the possibility that some mesh get dry)

I have some field measure inside the lagoon, but seeing the results obtained till now I didn't compare yet.

I attached the results file, zipped cause is very big (15 GB, sorry for thath but I set GRAPHIC OUTPUT = 20 t.step, tstep=1 sec trying to understand the problem).

I appreciate any help or advice that you could give me and sorry for the long post.

Regards,
Andrea

I cannot attach the zip file of the results, if someone could help me I will share it somehow

Hi
In my opinion, as you're model is forced with sea level, you shouldn't have 3 different boundaries for the sea and moreover, you should rather impose water level and velocity instead of water level and discharge...
Hope this helps

Hi christophe thank a lot for the reply

For what concerning the three boundaries, I have had the same problem just using only the OFFSHORE boundary with imposed levels and fluxes (the two others were set as free with thompson option). Moreover, I'm not so able with fortran and for adding a time-spatially varying boundary condition I must modified the subroutines.

Concerning the imposing of the fluxes instead the velocity, that is something that my lecturer almost "impose" to me.

Do you think that switch to Finite Volume scheme could help? (seeing that the courant number during all the simulation is always less than 0.8).

Thanksagain fot the attention.

Regards,
Andrea

Hi
This is clearly a boundary condition effect with overconstrained impositions...
There is no freedom on boundary nodes and small inconsistencies leads to instabilities...
Imposing velocity and water level means to impose a flux. So I don't understand what your lecturer means by imposing fluxes rather than velocities...

I didn't look at the different values but did you try to impose only discharge on 2 boundaries and only elevation on the 3rd one (for the sea part of course)

Finally, switching to finite volume will not solve the problem (in my opinion) as it's linked to the imposed value itself, not the resolution method...

Regards

Thanks Christophe,

I will try run the simulation imposing only velocity and sea level for the three sea boundaries and then I will try your second suggest (just discharge on 2 boundaries and sea level on the third).

Looking at my CAS file, did you find something wrong?


Again, thanks a lot for the help you are giving me.

Regards,
Andrea

Hi Christophe,

I would thank you for the advice, indeed the instabilities were due to the wrong boundary condition imposed.

I tried running the model forcing with H and Velocities on all the three sea boundaries, but the results was the same.

So, I tried running the model with two different type of forcing:

1) Sea level imposed on the offshore boundary and fluxes imposed on the others two
2) Fluxes imposed on the offshore boundary and Sea level imposed on the the others two.

The first one is clearly better(compared to the previous caases), without the strange wiggles that worried me.

I didn't look at the results of the second case yet, but during the simulation it shows "ILL POSED PROBLEM, ENTERING FREE VELOCITY AT THE BOUNDARY POINT 58" (that is the point closer to the coast of the southern sea-to-land boundary), the simulation reach the end anyway, probably that could be fixed setting thompson boundary option on it.

Thanks a lot for the attention you gave me, I really appreciate it.

Kind regards,
Andrea