open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi everybody,

I have a problem with a simulation. I have one upstream entrance with Q pescribed and 2 downstream outputs with stage-discharge curves. I gradually introduced a discharge and then I make a long enough step (40000 s) to obtain a steady state throughout my model. The problem is that in the 2 downstream sections, there is no match between the free surface and the discharge passing by the section. The stage-discharge curve is not respected.

For example, there is a discharge of 3200 m3/s at one output and results show a free surface at 4.8 m. But, in the stage-discharge curve, the free surface corresponding to this discharge is 2.7 m.

Someone have a idea?

Best regards.

Jules

Hello,

The commonest mistake consists in not putting the keywords :

PRESCRIBED ELEVATIONS and PRESCRIBED DISCHARGES in the steering file, in numbers corresponding to the number of liquid boundaries.

Then there is a delay (relaxation) to avoid triggering resonances and high oscillations, which is at the bottom of function sta_dis_cur.f in library telemac2d. This could also explain that you do not have exactly the values expected. You may change the value 0.02 which is hardcoded.

With best regards,

Jean-Michel Hervouet

Hi Jean-Michel,

Thanks for your reply.

I put these two keywords in my steering file but I also have a stage-discharge curves file and a liquid boundaries file.

What is the effect of increase or decrease this value of 0.02 in the sta_dis_cur.f

Best regards.

Jules

Hello Jules

The value 0.02 is a relaxation parameter that allows to smooth high gradient of the prescribed values. If it is your case, you can try to decrease it to 0.01 for instance.
An other case where the prescribed elevation (or discharge) can not be respected, is when you have supercritical outlet. In this case, the theory of characteristic tells us that there are 2 characteristic lines going out from the domain to the outlet, and therefore, none of the prescribed boundary conditions will be satisfied.

I hope that this helps
Riadh ATA

Hello everyone and Happy New Year,
Unfortunately we are not able to identify the source of the error: Model water levels several meters above our rating curve - water slope appears to be correct I am quite certain that my keywords are correctly entered: COTES IMPOSEES, DEBITS IMPOSES - i am attaching my .cas file here. As per Jean-Michel's suggestion, I tried setting the 0.02 value in the sta_dis_cur.f file to 0.1 and 0.01 and the results barely changed.
I am at a dead end. Any suggestions for other sources of error would be very appreciated! Thank you in advance
Jules

Hello,

There are 2 functions:

sta_dis_cur when the elevation (stage) is function of the discharge

dis_sta_cur when the discharge is function of the stage


There is a relaxation only in the first case. If you set a relaxation of 1 instead of 0.02 you should see exactly the value of your curve, so it is probably simple to try like this (if it remains stable).

In the second case (discharge function of stage) you should always see what your curve says.

So the things to check are: is the function sta_dis_cur or dis_sta_cur really called (you can add a print in it to see). If yes it becomes more complicated and we shall have to look at the boundary conditions file, and/or what happens when both liquid boundaries file and stage-discharge curves files are combined, we'll probably need to have the whole case to try it.

With best regards,

Jean-Michel Hervouet

Hi Jean-Michel,

Thanks, I will try to put the relaxation coefficient to 1 in the sta_dis_cur.f
I will keep you posted. To check if the function sta_dis_cur is really called, I just have to add "PRINT" at the end of the file? Where can I see if the function is called? In the listing printouts?

Thanks again.

Jules

Yes, but rather than a print, something like:

WRITE(LU,*) 'The function is called'

will print only once on the listing in parallel.

JMH

Hello,

I added the print at the end of the subroutine sta_dis_cur.f and it doesn't appear in the listing file.
I made a mistake. The boundary condition is respected but just at the end of the model. Here I'm attached the long profil of the free surface observed and a cross section of the boundary. There is no fall in the topography and we can see a fall of the free surface level to respect the rating curve. Why the upstream free surface level is different?

Thanks for your help.

Jules

Hello,

It looks like the flow is supercritical, so that you cannot influence your domain with what happens at the exit ?

If the entrance is supercritical you should give also the free surface elevation at the entrance ?

Regards,

JMH