open TELEMAC-MASCARET The mathematically superior suite of solvers

Dear friends.

I am trying to simulate a hypothetical urban area with a flood protection wall on the right river bank (attachment). The flood wall is represented as a long 'island' in the mesh. All other islands represent buildings (the island crossing the river channel is a bridge, with flow under the bridge through 'tubes').

I would like to model what happens when the flood wall is overtopped, so I created a 'weir' file and used the nodes on both sides of the island as the nodes for the weir.

Strangely, I do not get any water to the right of the weir even tuough the water level in the river is much higher than the weir crest.

Any ideas?

Thanks!

Roberto

Hello,

Just a hint after a quick glance at your files: the weirs file seems to have a number of empty lines, I am not sure it is the right format and do not know what happens in this case (maybe it works and the program jumps to the next line if the current line is empty, and you are lucky, but maybe something goes wrong).

With best regards,

Jean-Michel Hervouet

Dear Jean-Michel,

Thanks a lot for your response.

I copied the weirs file from the exmple (v6p3) and simply replaced the data with my model's information, so I think the format should be correct.

I followed your suggestion and deleted all empty lines, but I still get no flow over the flood wall.

In the manual (v7p0), it says that "in principle, these boundaries should be sufficiently far apart, upstream and downstream of the weir": my wall is relatively thin (60 cm. approx.). Could this be the problem?

Similarly, the manual says "the distance between two points should be the same on both sides". My flood wall is not straight. Could this also be a problem?

Best regards and thanks again!

Roberto

Hello Roberto,

Actually the documentation refers to the old algorithm and it has been changed by Artelia, so that any series of points will work, hence none of these reasons could explain that there is no flow. It would be helpful is someone from Artelia could have a look at this problem... It could be also that mixing weirs and tubes makes problems for some reason.

Regards,

JMH

Hi Roberto,

if you are modelling a wall, why are you not just implementing it in your mesh with the right hight and let it overtop by the flood wave?
I mean, you can try to implement the structure fully in the mesh.

Best regards
Gabi

Hi Roberto

We changed a part of the management of weirs but the previous method remain so this should work in your case.
Nevertheless, and maybe Jean-Michel could confirm or correct me, it seems to me that the management of weirs with the old method never works if the downstream side of the weir is dry!
In such case, you could do like Gabi wrote.

Regards

Hello,

Actually I do not confirm, it seems to me that subroutine loiden.f contains a formula for the discharge over a dry weir, and loinoy.f for a flooded weir. The problem can only be debugged by adding messages in these subroutines, to see what are the data they receive and how the points are coupled.

With best regards,

Jean-Michel Hervouet

Hi
loiden.f and loinoy.f are OK!
The computation of the discharge is well done event the downstream side of the weir is dry. For me the problem is located on the steps after when telemac wants to impose the positive discharge on the downstream boundary which is dry...
In clsing.f unorm is set to 0 in such case and then, in clhuvt.f the boundary is set to solid wall.

Regards

Hello,

Yes, Christophe is perfectly right, line 181 in clsing.f cancels the velocity when there is no water downstream. The conclusion is that we cannot flood an initially dry weir. In this case we would have to decide which depth we should have downstream. It is certainly not an easy development.

With best regards,

Jean-Michel Hervouet