open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello community, best wishes for you.

I am developing a flood model of a small river that crosses an urban area.
In the steady model I have set the boundary conditions at inlet with Q+H and the outflow boundary condition with free open condition (444). The time step was 0.06 (s) and the number of time steps was 60000 (resulting a 3600 seconds of total simulation). This model worked very well.

Now I am looking for an unsteady simulation with an inflow hydrograph but maintaining the free open boundary in the outlet.

In the unsteady cas file I have used the COMPUTATION CONTINUED : YES with the selafin file of steady model as PREVIOUS COMPUTATION FILE. I didn't use the RESTART FROM ZERO TIME option.

During the simulation, around the 1H48m, the simulation crashes showing a message that the time (6600.000006) is out of range.

Could someone help with this trouble?

Best regards

Hello,

When you write "RESTART FROM ZERO TIME", do you mean the keyword INITIAL TIME SET TO ZERO? If yes and you do not use it, this is let to default value = NO. In that case, your 2nd computation starts at time 3,600 s.

Do you have an input file dealing with time like a LIQUID BOUNDARIES FILE?
Do you wish to run your computation during more than 3,000 s?
Be sure every input data file dealing with time to be defined during a period of time including the duration of the simulation (a little bit more if too close to the ending time of simulation).

Hope this helps,

Chi-Tuan

Hello, it's me again

During my unsteady simulation I have solved the problem of time out of ranges by mean starting the simulation from zero time.

But now I have another problem. I have a one hour hydrograph (attached image) and when the simulation reaches the iteration 15400 (TIME: 15 MN 24.0 S) the next message appears:

LIQUID BOUNDARY: Q(2) = 1.2239999999996432
ILL-POSED PROBLEM, FREE DEPTH
ON BOUNDARY WITH ENTERING VELOCITY
AND SUPERCRITICAL FLOW
FOR EXAMPLE AT BOUNDARY POINT NUMBER 556

And after that the simulations crashes with the message:
DEBIMP: PROBLEM ON BOUNDARY NUMBER 2
GIVE A VELOCITY PROFILE
IN THE BOUNDARY CONDITIONS FILE
OR CHECK THE WATER DEPTHS.
OTHER POSSIBLE CAUSE:
SUPERCRITICAL ENTRY WITH FREE DEPTH

This simulation is coupled with GAIA to analyze sediment transport.

Could someone help me with ideas of possible solutions for this issue?

Note: in my unsteady cas file I am using VELOCITY PROFILES : 1 (Downstream open boundary 444); 4 (Upstream boundary with prescribed Q)

Best regards
Ricardo

Hi Ricardo,

What do the depths look like at the upstream boundary? When coupling with GAIA, I've had numerical instabilities cause depths to go to 0.0000000001 (approximately) at certain nodes, which causes the velocity to spike and thus erosion to skyrocket.

See #42143 to collect data from the last recorded timestep, if you're using multithread.

Hope this helps,
André Renault

I have another doubt. I have run the model in steady conditions with boundaries (inlet Q+H and outlet free 444) with 2 m3/s during 30 minutes and the simulations ended well.

This file I have put as PREVIOUS COMPUTATION FILE for the unsteady model; the last frame shows a 0.19 m WATER DEPTH at the inlet. But when the unsteady model starts, in this case the first iteration shows:

---

ITERATION 0 TIME: 0.0000 S

GAIA COUPLED WITH: TELEMAC2D
TELEMAC2D INITIALIZED
THE LIQUID BOUNDARIES FILE CONTAINS
7 LINES WITH:
Q(2)
USING STREAMLINE VERSION V8P4 FOR CHARACTERISTICS
DEBIMP: PROBLEM ON BOUNDARY NUMBER 2
GIVE A VELOCITY PROFILE
IN THE BOUNDARY CONDITIONS FILE
OR CHECK THE WATER DEPTHS.
OTHER POSSIBLE CAUSE:
SUPERCRITICAL ENTRY WITH FREE DEPTH
IN THIS CASE GIVE A POSITIVE DEPTH

If the steady model worked, why the unsteady can't start?

Regards

DId you activate the computation continuation, there is a dedicated keyword ...

Thank you for all your responses, the problem was at the inlflow, where supercritical flow was developing. I use an artifice to reduce the velocity just at the nearest section in the inlet, and then the model in the computation continued worked.

Regards