open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello, so I'm the first one in this category. I'm quite astonished, how good Telemac performs in difficult cases.

Enclosed a nice picture of the results of Telemac (v5.9) modelling (Finite Volume method with Roe Scheme) of a weir sill with stilling basin.
Like in the physical model the half widths of the abutment piers were also modelled to account for contraction losses.
The water level at the outlet is constant, but depending on the chosen roughness the main energy dissipation happens in the stilling basin or not. These are steady solutions. The grid has 6489 nodes, the node spacing is approximately 1,5 cm * 1,5 cm. The calculated spillway capacity, in other words the upstream water level, is in quite good agreement with the physical model.
Using the Finite element method - in the few cases of a steady solution - the water surface isn't so smooth as with the Finite Volume method.

Best regards,
Clemens

Now enclosed the picture!

Clemens

konsonaut,

This is great work. Would it be possible to add the input files (geo, cas, steady state boundaries, etc.).zip so we can include your test in the official benchmarking document ?

Thank you in advance.

Sébastien.

Sébastien,

thank you for the reply.

However these are only preliminary results. It would be necessary to define different roughness zones according to the physical model, so the actual results are not really accurate. Another interesting idea would be to blow up the model to prototype scale and compare the results. Maybe I have time in the near future and so I can test some cases. Enclosed you will find the geometry, the steering file and the boundary file.

Clemens

Clemens,

These are some very interesting water profiles. When you get a chance to resize the model to scale, it would be very valuable to evaluate a rating curve (Head vs. Discharge) and compare it with measured and theoretical values.

cheers

j

Nice work ! However, I'm quite skeptical with the results. I've done several models for spillways and other hydraulic structures, some are 3D and others vertical 2D using CFD codes since the flow is multiphase (water/air mixture) but never with a 2D FE/FV code.
Can you please give more details concerning how you manage the change of the flow viscosity in the stilling basin and vertical accelerations?

Aissa,

Hello Aissa,

I'm also quite sceptical in doing numerical simulations of such cases. Therefore we are still doing and maybe for another long time physical model tests for these investigations (to give reliable informations to clients). I didn't never try a 2-phase model for spillway modelling and so I don't know if the results from these models are realiable.

Since I used a depth averaged model, I could not account for vertical accelerations (which are important of course in this case). I think that a imposed or a calculated Eddy Viscosity along the horizontal does not have a large role to play in this case (much more the momentum transfer in the vertical).

However I got quite good results regarding the calculated upstream water level compared to the physical model. The upstream water level doesn't depend much on the given roughness value (as in nature). On the other side the location of the hydraulic jump depends highly on the given roughness value (as seen in the picture) which is a big problem.

In the book - Hydrodynamics of free surface flows: Modelling with the finite element method (2007) - from Jean-Michel Hervouet, there is a nice example of a depth averaged simulation of a hydraulic jump and the comparison with the theoretical solution. There is also 3-D non hydrostatic calculation of this case. You can find the input files of this case in the training material of Telemac 2D.

Regards,
Clemens