open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

I did some flood inundation modelling with Telemac 2D and I used mainly the Finite Element method together with the wave equation and the Characteristic method for the velocities.
I used 2 time dependent inflow boundaries and a stage-discharge curve for the outlet. I got quite good and reasonable results.

For comparison purpose I tried to use the Finite Volume method togehter with the Kinetic Order 1 scheme. Besides the much longer computation time, at the outlet boundary the water level is much lower comparing with the results of the Finite Element computation. Moreover sometimes the water level is lower than the minimum imposed value in the stage-discharge file.
Surprisingly the maximum inundation areas with FV and FE are quite similar, and also the water depths at this time step.
What happens?

Following the new manual I tried also the numerical scheme 14 for the velocities. However after some time steps the calculation stopped without an alert. Decreasing the time step to a very low value didn't help. Do I need to consider some other parameters when using this type of advection scheme?

Best regards,
Clemens

Hello Clemens,

On stage-discharge curves with finite volumes and kinetic schemes, I'll let my colleague in charge of Telemac-2D answer.

On scheme 14, it must be used with strictly positive depths. So you need to have the same keywords as e.g. file caspos.txt in the malpasset test case in directoru test.gb, namely :

TREATMENT OF NEGATIVE DEPTHS : 2
CONTINUITY CORRECTION : YES
MASS-LUMPING ON H : 1.
SUPG OPTION : 0;0

Would you have inf-sup oscillations of the free surface, you can use :

COMPATIBILITY OF FREE SURFACE GRADIENT : 0.9

I have to put a stop in the next version to prevent what you did and give the above explanations.

With best regards,

Jean-Michel Hervouet

Hello Jean-Michel,

thank you very much!
Yes, for the TREATMENT OF NEGATIVE DEPTHS I used option 1. This was the mistake.

Best regards,

Clemens

Dear Clemens,

Until release V6P0 of Telemac, finite volume kernel is not abe to consider a stage-discharge boundary condition. I have already done the necessary to overcome this problem, and these corrections will be included in the next version.

Kind regards.
Riadh.