open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi,

I am a bit puzzled at my model result and hope that someone may have an idea on what causes it.

Below are the velocity distributions from the model on the bottom plane (Layer 1) and one plane (Layer 2) above it. The vectors are about 50 m apart and the black contour lines are shown are for every 2 m. The deepest elevation in the reach shown is at about -18 m. I have 14 planes in the model with Layer 2 prescribed at 4% of the water depth.

The velocity distribution in Layer 2 looks reasonable. But in Layer 1, I noted neighbouring nodes with similar elevation (within 50 cm) but one node consists of velocity of nearly 0 m/s and the other one has velocity of 30 cm/s. It seems that the bed slope/gradient has something to do with this. Any suggestion?

Thanks.

Edwin

Layer 1 (bottom plane)


Layer 2

Hello,

An important detail is whether you use a Strickler-like friction law (all laws except Nikuradse, which are based on 2D reasoning and local depth) or Nikuradse, which is more local and looks at plane 1 and 2. Another detail is the turbulence model.

Anyway you may have recirculations after slope gradients, could this explain your observations ?

REgards,

JMH

Hi,

Following your suggestions, I have tried the following:

1) Nikuradse with Smagorinski (horizontal) and Prandtl (vertical)
2) Nikuradse with constant viscosity of 1.0 (horizontal and vertical)
3) Manning with Smagorinski (horizontal) and Prandtl (vertical)

Black arrows -> Case 1
Red arrows -> Case 2
Blue arrows -> Case 3

Layer02




Layer01



Unfortunately, similar patterns still persist in Layer1.

Edwin

Hello,

OK. The other parameters to check are :

TURBULENCE MODEL FOR THE BOTTOM : 2 (but this is the default)

and you can try :

VERICAL VELOCITY DERIVATIVES : 2 (default is 1) A log derivative fits well the logarithmic law so we get a better accuracy.

Another important aspect is the mesh size on the vertical and whether you refine on the bottom or not.

With best regards,

Jean-Michel Hervouet