open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello everyone,

I am doing my internship and I have to model flow through a river and when comparing the results with the real data and I find that my water depth is low in comparison with the real data. I know it can be due to many factors but first I want to check that my simulation file is correct and that I am not using the wrong value somewhere.

Hope someone can take a look at my file.

I assume your two open boundaries correspond to the period of your observed data.

Are your first boundary flow and second boundary level?

The Manning friction coefficient is 0.064, which seems a bit high for a river channel. I would expect it to be around 0.035. However, reducing it will further lower the level.

You are using constant viscosity (Turbulence model = 1). Your velocity diffusivity should be the actual eddy viscosity. Try a value 1. Higher viscosity is likely to raise the level.

Manning n and viscosity are normally calibrated.

Hailiang

Hello Shenh, thank you for your answer.

Yes, my two boundaries represents the observed data and I have impose discharge, inflow and outflow, because I need to obtain information close to the outlet so I don't want the measurements to be influenced by the imposed level. I was wondering if a totally free boundary (444) would make any difference.

As for the Manning, I tried with values around 0.03/0.04 the first time but since the level was to low I increase it, since it is a more or less coarse river bed I figured this value was not so far up from the reality. But it did not changed that much the level.

I will try 1 for the viscosity and see how it change.

Hello Shenh,

I changed the Diffusivity to 1 and yes my water level raised but I encounter some other problems. I observed an unexpected pic on the graph and to try to solve it I used the wave equation and now I see really high water levels. Any other suggestions to try and obtained a better result?

I am trying to model a small shallow stream in order to simulated sediment transport afterwards.

Attached my last 2 .cas files and their respective resultant water level graph.

Thank you in advance for your answers.

Can you zip all your model files and attach it? I will find some time to test run it.

Hailiang

Thank you so much, that would be very helpful.

*My initial conditions file is to big so I can not send it.

Is your model in a global coordinate system? The x and y values look like lat/long.

I'm sorry but i actually don't know. How would it change my results? I think is in global coordinates buy i'm not sure.

By measuring the width of the river, it is only three (3) metres. Is it real?

When you were creating the mesh, were you using some shapefiles and are they in a projected coordinate system or geographic coordinate system? You can google the difference between the two. I always use a projected coordinate system. If you were using geographic (used for modeling large waterbody), I believe you need to add key/value lines SPHERICAL COORDINATES = True and LATITUDE OF ORIGIN POINT. Others may be able to share their experience using such a system.

Hailiang

I will look into this. Thank you for your answer.

I created the mesh from a set of .xyz points that were handed to my by my internship responsible.