open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi everybody,
actually I'm working over a river analisys. We have a really defined description of the bottom, because it has been measured with a photogrammetric tecnique.
I have something like millions of points, with a density of 1 point every centimeter.

With such precision, I can really see every cobble along the river.

So, here's the question: if I run a simulation with a very dense mesh, how does this precision of the bathimetry influence over the global roughness?
Normally the values of roughness are given under the assumption of a coarse definition of the bottom.
Using roughness values from literature, could I risk to overestimate energy losses?
Should I reduce the roughness value?

Thanks for any suggestions.
Best regards.
Riccardo

Hello,

In the theory there should be no influence of the mesh density over the roughness in the equations but the first question is why do you want to have such a dense mesh for computing hydrodynamics? You should start with a reasonable mesh (let's say order of magnitude of meters rather than centimeters), and then you can do sensitivity tests with a finer mesh.

With best regards,

Jean-Michel Hervouet

Probably I wasn't so clear in the explanation..
It's not simply a matter of density of the mesh, but the fact that I have a really detailed topographic survey.

With such data, I need a fine mesh in order to keep the informations about the bottom geometry. From these data, the bottom is described "as it is", with sand, cobbles, boulders.. so it's very irregular. It's not the "flat" bottom that you have with a point every meter.

Could this bottom asperity increase energy losses?

Thanks again.
Riccardo

Dear Riccardo,

you would like to model the flow over your complex irregular geometry, right?
I think that if your mesh is fine enough, you need to specify only the grain roughness.
The total roughness usually consists of bed roughness and grain roughness (and sometimes roughness of the bed forms). If your mesh is fine enough to include the bed roughness effects, you just need to specify the grain roughness.
Easiest solution: Try and check with measured values…

BUT - I think you should use TELEMAC3D to capture the flow behaviour, otherwise you can go with a much coarser mesh like suggested by JMH.

Best regards,
Gabi