open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello everyone,

I hope someone can help me with this problem:

I'm modelling a bridge as a series of tubes.
I'm trying to calibrate the behaviour of the bridge (the discharge that can flow under the bridge), for instance, by changing the parameter L12 of the TUBE DATA FILE. The bridge is quite far from the boundaries. The thing that I don't understand is the following:
If I set a value of L12=4.7 the simulation run, and the results seems reasonable.
If I set a value of L12=4.5 the simulation run, and the results seems reasonable.
If I set a value of L12=4.6 the simulation does not completely run and the following error appears after a lot of iterations:

"
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
IN THE INITIAL CONDITIONS
OR PRESCRIBE THE DEPTH AT THE ENTRANCE.

PLANTE: PROGRAM STOPPED AFTER AN ERROR
"

I've checked the "partial" results:
- the velocities seem reasonable for all the "partial simulation" including the ones one step before it stops
- all the boundary is wet for all the "partial simulation".

I've checked on the forum to find out some similar problem but without lucky.

Thank you in advance,

Daniele

Hello,

I do not know where is boundary number 2 but this message occurs when the velocities or the depth on a boundary make a discharge equal to 0, whereas you precribed a non zero discharge. Then the program cannot multiply the velocities by a constant to get the required discharge. It could be that this boundary becomes supercritical. It does not explain why it works for two cases and not the case in between, but if the problem is not well posed due to a supercritical flow, results become highly unstable and you may just be lucky in 2 of the 3 cases.

With best regards,

Jean-Miche Hervouet