open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi,

I have a problem about the thin dam or dike.
As having been described in Telemac 2D manual that
it can handle weirs, whose width is comparable
to its length and general grid size.
And each weir is represented by 2 lines along width.

But what if the dam or dike is quite long and thin, say hunderds
meters long but several meters wide, in a large model domain with
a typical grid size of several hunderd meters.
If we still use 2 lines to represent the thin dike, then grids along
the width of the dike will be tiny, which makes simulation of large model
rather expensive...

So I am wondering that can we just neglect the width of these thin dikes,
and treat them as single lines with solid BC assigned ???

Any help is deeply appreciated.

Best regard !

Hello,

Actually a dam or a dyke can be simply represented by the topography, or by an obstacle like an island (solid boundary) in the case of a non submersible dam or dyke (if there is a flow in turbines in the dam I would treat it as a culvert). The weir facility in Telemac-2D is used for thin weirs or dykes that are submerged and that cannot be represented by the mesh.

It seems to me that, according to your description, what you want is simply a non-submerged dyke, and it effectively can be treated by a solid boundary, unless you want to have dry zones in your model, in which case you could represent the slope of the dyke with the topography (if it is to have only a thin layer of uncovered elements, due to a very steep slope, I do not recommend it, and this is not in the normal range of shallow water equations which are normally meant for gentle slopes).

With best regards,

Jean-Michel Hervouet

Hello,

Thanks a lot for your great help.

I understand your meaning, but there's still
something confusing. Actually, I want to know
if I can treat the dike/groin with a single line
because it's too thin compared with an ordinary
grid size. But in the manual of FUDAA, this kind
of treatment seems unavavilable.
Instead, it takes the width of the dike/groin
into consideration(as shown in the fig.)
As far as I know, Delft3D and FVCOM could deal with
dikes in this way.

Could please give me some advice...Thanks again.

Best regard!

Hello,

The dyke may be arbitrarily thin but topologically it must be as in your figure, with distinct points on either sides (except at the extremity where it can be a single point). A single line with single points would not give a finite element mesh, as you would not know on what side the points are. So the dyke may be like a razor cut in the mesh, provided that topologically it is done with 2 different lines (if the mesh generator allows that two points are superimposed and this may be a problem... in this case you could do as in your figure, except that the tip of the dyke would be a single point, so that no edge of element is too thin).

With best regards,

Jean-Michel Hervouet

Hello.

Thank you very much for you quick reply.
Learning from your detailed explainations, I think I have to give up my idea which comes from FVCOM/Delft3D.

Actually,FVCOM simply deals with dikes/groins as sinle lines which proves to be quite effective. Maybe it is not the case for models using FEM like Telemac.
Maybe Telemac could be improved in this part.

Best regard.

Hello,

Thank you for your interesting suggestions.
Actually, it is not a question of finite element or finite volume. The idea could be interesting especially for ocean modelling where the mesh element size is very big (in the order of km) and where we need to modelize a groyne or a dike (submerged or not) with a size of several meters. This allows keeping a reasonable mesh size by using only "large" elements.
For Telemac (which is used not only for coastal applications but most often for river in inland applications) the modelization of groyne or dike is achieved using the procedure described by Jean Michel his previous post. This way is more efficient since the element size used for dike is of the same order than the one used elsewhere.

To be continued ...

with my best regards.

Riadh Ata