open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi,

I've build a simple straight, flat bottomed channel with solid boundaries along both longitudinal sides (sitting at the same elevation as that of the channel bottom) - to give straight walls. Both the bottom surface and lateral boundaries are to have a friction value of manning's n = 0.025. LAW OF BOTTOM FRICTION and appropriate coefficient was included in the steering file.

I ran 2 simulations. The following was included in just one.

LAW OF FRICTION ON LATERAL BOUNDARIES : 4
ROUGHNESS COEFFICIENT OF BOUNDARIES : 0.025

When excluded, it returned a normal depth significantly below that calculated in HEC RAS and ISIS. It gave a smooth water surface profile along the channel length (see image attached) but with no 2D velocity profile.

When included, it returned a normal depth close to that calculated in HEC RAS and ISIS and a 2D velocity profile was observed However, the water surface profile fluctuated along the channel length - by roughly 15mm. This can be seen in the below image.

Is the water profile expected to fluctuate like this? Or could the problem be due to a sparse mesh/short simulation duration?

Many thanks,
Brian

Hi

If you need to simulate a straight channel using a real Saint Venant model, you have to model both sides of the channel with elements with high bathymetries (higher than the free surface) and to impose a perfect slipping condition (i.e. "2 2 2" condition in the boundary condition file).

There are specific cases where we need to impose friction at the side walls. In these cases, we use key words LAW OF FRICTION ON LATERAL BOUNDARIES and
ROUGHNESS COEFFICIENT OF BOUNDARIES. However, you need always to use the same kind of model I described before. Anyway, it seems to not be your main aim here.

To be continued ...
Kind regards.
Riadh