open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi,

I am simulating a row of Culverts, There are 16 culverts represented by 16 contiguous nodes. It’s all with same parameter. When gates of culverts opened and water go through it, there are some phenomenon that I don’t understand happened:

1.The water flow through the middle one is much higher than other culverts. (I don’t think bathymetry result this).
2.The adjacent node in the downstream of culverts show a small velocity, it’s like a recirculation zone but shouldn’t appear in here.

In my mind, the velocity distribution after culverts should be uniform. Why this happened, is there anyone who could help me?

Kind Regards,
Bin

Hi
It's usually advised to avoid placing entry/output nodes on a boundary.
Try do move them and see what's happen

regards

Hi，

Many thanks for your reply. I have tried that method. The Problems are still exist. But, the water flow through the middle one is smaller than other culverts.

Could you please explain to me why should avoid placing entry/output nodes on a boundary, because on the boundary is seems more close to real. Moreover, is there any method to deactivate the Element between entry and the boundary?

I think I will try to adjust the grid and bathymetry to see the result. I will keep update this post.

Regards,
Bin

Hi
There is specific treatment on boundaries which could be incompatible with the management of sources/sinks...
However it works so maybe it's not so problematic...

Nevertheless, culverts exists Telemac in order to ensure the function (let the water flows from upstream to downstream) but necessarily without a high resolution of the local velocity field as the size is often smaller than the local mesh size.
It's of course possible to use weir eventually with adaptation to compute the right discharge but the local floww will also be "theoretical".

If you wan to focus on the local flow, you need a finer mesh and probably in case of submerged flow another code like: Code Saturne, Open Foam, ...

Regards

Hello

Culverts represent the intriduction of a sink/source at a fixed node. This means that the velocity at this node will be artifically computed in order to retreive the prescribed dischage at this source. Thant's why, it is well known that the velocity destribution will not physic around these nodes.

In your case, I think that this could be better to represent this structure by a weir rather by a set of culvert. Maybe Christophe can give his expert opinion about this suggestion.

with my kind regards

Riadh

Hi, I have done several test since then and 'MASS-LUMPING ON H =1' could improve this problem a lot.

Yet some downstream adjacent nodes still show a much smaller velocity(it's the downstream node of culvert,not the culvert extremity itself). See the Figure below. I have checked my mesh, setting and code, couldn't find the problem.

Do you have any idea about this?

Regards,
Bin

Could we have an idea of the water depth downstream the culvert?

Many thanks for your reply.

Hi Mr.coulet, The Water depth and velocity downstream are show in figures.

Hi Leo, I have checked CFL, Froude-number and discharge, everything seems good. ST as 2S, turbulence model as K-epsilon.

Thanks for all your help.

Thnaks for these info.
Could we also have a map of the bottom and the free surface at the same time than the velocity field.

As a test, it could be interesting to try to use two liquid boundary with imposed discharge and try to enter an approximation of the discharge on one tide just to see the shape of the velocity field...

Regards

Hi,

I've never used culverts. Have you checked the CLF-Number and Froude-number, etc. What time step size do you use? Which scheme?

Is the discharge (inflow/outflow) correct?


Regards

Leo