open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello everyone,
I'm simulating a river with relatively high slope(around 4%). though i run the program for quite longer time but cant achieve steady state. the flux is oscillating a lot in outflow boundary condition. moreover i get negative depths too. Im attaching here my Cas file,geometry file,boundary condition file and result file. any help will be highly appreciated!
PS: sorry couldn't upload the geometry and result file

Hello,

You have tuned everything to avoid negative depths, except :

TIDAL FLATS that should be = YES

Outflow 4 4 4 is for a supercritical exit, is it what you want ?

With best regards,

Jean-Michel Hervouet

Hello Mr.Hervouet,
thank you for the reply,actually this outflow boundary seems to work better than others for me, as for 544 I get infinite velocity. but with velocity and water depth free,i have avoided that. what could you suggest for my downstream boundary condition?
Regards
Najib

Hello,

At your exit, when there is no water the option 4 4 4 (all free) can work, though truncation errors deling with zeros may diverge after many iterations. When the water reaches the exit, it probably becomes a fluvial flow, so 5 4 4 is recommended. My recommendation: once you have water at the exit, make a restart with 5 4 4. However if the elevation is too low compared to your prescribed elevation at the time of restart, it may trigger entering velocities that may diverge. Other solution: Thompson's boundary conditions for this exit.

Regards,

Jean-Michel Hervouet

Hello,
Thank you very much for your suggestions,
I restarted the simulation with a file from previous boundary condition and changed the boundary condition back to normal boundary condition for rivers(455 upstream &544 downstream). luckily there is no infinite flow velocity at either of my boundaries, but still I can't reach steady state, even at Inflow with Prescribed flow, I see oscillation in Flux. I have attached the snapshot from the water depth propagation which shows somehow a continues wave.I dont know what else can help me to reach a steady state condition.

Hello,

I should have asked first the question : do you really want to start from a dry river ?

I see that your prescribed elevation at the exit is 2112, and you say:

INITIAL CONDITIONS : 'CONSTANT ELEVATION'

but /INITIAL ELEVATION : 1842

is commented, so you start from the default value which is 0. and you get a dry river (and your results are really strange, with seemingly negative depths and strange waves).

Why don't you start at least from :

INITIAL ELEVATION : 2112.

How can the water flow even if the initial elevation is 1842 and the elevation at the exit 270 m higher ?

Another potential problem is that your coordinates are extremely large, I see:

7680840 and 4097300

this must be very challenging for numerical accuracy, you should do a translation (subroutine corrxy in library bief) so that the origin is in the domain.

With best regards,

Jean-Michel Hervouet

Hello,
sorry for creating confusion,1842 was from my second model.I thought when I have a restart file,then the model will initialize from my previous computation file.
well, there is no negative depth in my model after imposing TIDAL FLATS= YES, but yes strange waves(but some areas look better,no such strange waves).
I have also tried with initial condition,but still no improvement.so now when I wanted to use corrxy.f,an error occurs. Please have a look at the error.
I thank you very much for your help!
Regards
Najib

Hello,

This result looks better, but why not looking at the free surface instead of the depth that vary also with bottom ?
You Fortran error is strange because you get this when bief is not compiled.
Is 1000 s long enough and how did you choose a time step of 1 s ?

Regards,

JMH

Hello,
well, free surface looks also wavy,here im attaching a close snapshot.
Its about 2Km reach,with a restart file 1000s seems to be Ok.
I chose the time step according to Currant Number, to keep it as lower as possible(still I have 6-10). I'll try to recompile the program, hopefully then will work.
Regards
Najib

Hello,

Another possible problem: if the slope is 4%, it could be that the entrance of your domain is supercritical, hence the depth or free surface should be prescribed also at the entrance, otherwise we get this kind of oscillations which seem to start immediately at the entrance.

Regards,

JMH