open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi everybody,

I'm modelling marine submersion along a coast. the model is forced by the tide. There in an estuary on my liquid boundary and I have met an issue here. The evolution of the elevation surface, the current number and the froude number are very strange as you can see in attachment :



One file of elevation surface is coming from the liquid boundary and it is ok. The other file of elavation surface is quite different and it should not because the two points are very close (less than 100 meters).
The evolution of current number is strange for me because it varies very sharply.
Then, the Froude number evolution is also disturbing because it is higher than 1, and I think the flow should not be turbulent.
The different graphics are coming from points on the liquid boundary or very close to it.

I would like to improve the numeric parameters of my case file (in attachment) to avoid "the error".



Is somebody have an idea?

Thanks in advance

Nicolas

Hi
As I understand, the liquid boundary of your estuary is located in the tidal area.
Are you sure that the liquid boundaries condition are compatible?
If you impose the water level, how are the velocities? probably free?
So when tide arrive near the boundary where you impose the free surface (which is increasing and by consequence generate entering velocities) there could be a sort of shock...

The node results doesn't allow us to well understand the problem. A 2D view of the local aspect of free surface could give us more information.

Regards

You well understand.
I don't know what are you talking about "compatible condition" but I impose the flow of the tide and only the tide. There is no water coming from the river. I impose the water level and the velocities are free, you are right.
You can see in attachment 2D views.



I have add the bathymetry, it will be clearer for you. The liquid boundary is coming inside a part of the estuary.

Regards

Nicolas

It's not clear for me.
Where did you impose what?

When I talk about compatibility, I would say if you impose water level (and let the velocities free) on the river limit, and if a tidal wave arrive in this limit, I suppose you integrate this phenomenon on the impose water level. The problem is located in the way you obtain this water level... If this is taken from a tide gauge, this water level is the consequence of the tide which arrive. In you model, if the propagation of the tide wave is not exactly the same as in the reality, that means you try to impose something which is not real (physically).
So this could generate some problems...

Hope it's clearer for you now.

regards

I impose the water level all along the boundary you see on the pictures 2.
I integrate the flow coming from the river in the water level. The data I impose on the boundary don't come from a tide gauge but from another model.
When you say "not exactly the same", that means approximatively how many centimeters of difference?
So if it is not really physical, what can I do?

Regards

Nicolas

Hi
Tide gauge or another model it's the same phenomenon.
It's difficult to give you a value of difference because in theory to avoid local problems, this could be 0.
For the solution, it's hard to answer. In fact, your model is badly built (in term of the cover area).
You could try thomson method but I'm not sure this will solve the problem.
Another solution could be to program a sort of data assimilation...

Good luck

I can change the boundary and the cover area. I mean to simulate the flow in all the estuary but at the new boundary it could be the same : the data imposed and the physical reallity could also be different... Or you think to place the new boundary somewhere else?
You advise me to use Thomson method because you think the model is under forced? It could be an idea, I'm going to test it.
And about data assimilation, can you develop?

Regards

Nicolas

No if you place your limit outside the tide influence and impose the river discharge...
Your model is underconstraint on this boundary. For me, Thomson method is more relevant fore overconstraint problem in order to relax it. I'm not sure this will have an effect here.

For data assimilation, as the name indicate it suppose to use the result model to adapt the boundary condition in order to limit the problem of uncompatibility. Probably not easy ...

Regards

I can't impose the river discharge separately from the tide because I just know the water level of both in few points but not of each one...
Ok for Thompson method.
Data assimilation seems not easy as you say. I will try if I have no solution.

Thank you for your advice.

Regards

Nicolas

I don't well understand your first comment. On the river, if you go upstream after the limit of tide influence, you could find a discharge...
You talk about another model but what king of model?

keep in mind that in your case an oscillating volume in your river exist and this volume could be important for the result. With your method, there is no verification on this volume (as you impose only water level)

Good luck