open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

when using a input flow hydrograph and an outlet stage-discharge file in which by increasing discharge the water level goes down, for example:
#
# STAGE-DISCHARGE CURVE BOUNDARY 2
#
Q(2) SL(2)
m3/s m
0.0 533.6
150.0 533.6
170.0 533.6
220.0 531.8
#

than, after the discharge at the outlet reached the value of 170.0, the water level in the simulations is jumping between 533.6 and 531.8m and is not correctly interpolated.
There seems to be a bug, where Telemac interpolates between the given values, so probably Telemac calculates in this case a negative water level and uses then the minimum value.
Is there a way to correct this?

At the moment we added the water level in the hydrograph file. This solution is working well, but it is not really correct.

Many thanks,

Gabriele

Hello,

The subroutine that does the interpolation is in library telemac2d and is called sta_dis_cur.f

Maybe you can add it in your fortran file and trace what happens, because a priori I don't see any problem there (though obviously you have one). In this routine I expect that the discharges monotonically grow in the curve, which is your case, but the elevation may be random. Note however that there is a relaxation (but this should prevent jumping too fast).

With best regards,

Jean-Michel Hervouet

Hello Jean-Michel,

thank you very much for the answer.
I had a look in the subroutine sta_dis_cur.f and tested some examples and there I could not find any problem as you said. However, I tried some modifications, but it didn't help. Every time when I impose a drop down of the water level the water elevations at the outlet seeme to be random and sometimes even the water depth at the outlet becomes negative. So also the limiters in the if-statements in the sta_dis_cur.f are ignored: if discharge is lower or greater than min or max Q -> use W.

It is really strange for me what could be the difference between an imposed increasing water level and a decreasing water level.

Best regards,

Clemens

Hello,

Yes it could be that in such a situation this way of using stage-discharge curves happens to be unstable. It is probably due to the fact that it can be only a transcritical situation, is it ? In this case there should be no influence of the downstream boundary condition, so trying to change anything at the exit is againt the theory of characteristics.

With best regards,

Jean-Michel Hervouet

Hello,

yes this could be a good explanation but unfortunately in my test case the Froude number is below 0.5 during the whole simulation. So I don't have any ideas.
Enclosed there are the input files for the tests: SDC_tests.zip
Maybe anyone has some ideas?
By the way: in the simulation until to a steady state with the cas_steady steering file, I get no real steady flow field. The flow field oscillates a little bit even after several time steps and the continuity at the outlet is not really satisfied (1-2% deviation). Why this? With such an geometry I would expect a good steady flow field with the given parameters.


Best regards,

Clemens

Hello,

What I do not understand then is how you can have elevation decreasing with discharge increasing in your stage discharge curve, isn't that strange ?

For what concerns the flow field, you can have some kind of seiches in the domain, so that it may take some time to become stable. To have an exact mass conservation, use keyword : CONTINUITY CORRECTION : YES, that will give you the exact flux at the exit where the elevation is imposed.

With best regards,

Jean-Michel Hervouet

Hello,

in the case of floods with opening of the gates at hydro power plants (to prevent flooding upstream) or in the case of reservoir flushing during floods (increasing the shear stress by a draw down of the water level at the outlet), such an outlet boundary condition occurs.

In the meantime I did some additional tests but without success. Even if I don't use an unsteady inflow boundary file but a constant discharge in the steering file with a flux dependent draw down of the water level at the outlet, I get oscillations in the outlet water surface. It seems to me, that the enforced draw down of the water level causes an additional outlet flux which in the next time step results in a too low calculated water level and so the oscillations become bigger and bigger.

Best regards,

Clemens