open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi,

I am trying to run a simulation of a open channel flume which is 1m wide and the depth of flow is 10cm. I am trying to match the verticle velocity profile from the simulation with the experimental data. The results from the simulation provides over-estimation of the velocity profile when compared to experimental data.

I am using following values:

VERTICAL and HORIZONTAL TURBULENCE MODEL =3
LAW OF BOTTOM FRICTION = 5
SCHEME FOR ADVECTION .....(ALL) = 5
SCHEME OPTION FOR ADVECTION ......(ALL) = 4
PRECONDITIONING FOR DIFFUSION OF VELOCITIES : 34
PRECONDITIONING FOR DIFFUSION OF K-EPSILON : 34
FREE SURFACE GRADIENT COMPATIBILITY : 0.9

What could be the possible reasons in this case affecting my solution? Can anyone help me with the problem?

Also, I have few questions that could help me understand the issue effectively:

1. What value of FRICTION COEFFICIENT FOR THE BOTTOM for Nikuradse law would be good assumption for mean sediment dia of 1.5mm?

2. How many verticle planes (NUMBER OF HORIZONTAL LEVELS) would be enough to simulate this problem with depth 10cm and averge velocity of 0.5m/s?

3. I have provided a MESH TRANSFORMATION = 2; with closely spaced layers at the bottom and gradually increasing spacing towards the surface.In this case is there as minimum spacing to be taken into consideration?

4. As this is a representation of a fluvial case with sediment transport; while coupling this with GAIA, will the bedforms have effect on the friction factor as well?

5. If there is anything else that could help me achieve a good and accurate solution please suggest me.

I would really appreciate your cooperation and suggetions.

Thanks.

Hello Ravi,

One important thing to have in mind when modelling scale experiments with TELEMAC-3D is that there is a minimum volume that may have to be changed depending on the size of the elements (along x, y and z). It can be changed with the keyword MINIMUM VOLUME OF 3D ELEMENTS. Default value = 1.E-6 m3 = 1 cm3.
What are the minimum sizes of your prisms along x, y and z?

Another tip is to keep dz lower than dx and dy otherwise the results may be questionable.

1. please carefully read the TELEMAC-3D user manual. There is advice how to choose FRICTION COEFFICIENT FOR THE BOTTOM for Nikuradse law, in particular in the E appendix (recommendations)/

2. hard to answer without knowing what is expected along the vertical. You have experiments, they should guide you.

3. it should be consistent with MINIMUM VOLUME OF 3D ELEMENTS.

4. please read the GAIA user manual, in particular the keywords COMPUTE BED ROUGHNESS AT SEDIMENT SCALE and BED ROUGHNESS PREDICTOR OPTION.

Hope this helps,

Chi-Tuan

Thank you for your prompt resposne.


The minimum size of the element near the bed is 2cm * 2cm and the first plane is at 0.16cm above the bottom.

But I have simulated with these values without changing the MINIMUM VOLUME OF 3D ELEMENTS value and didn't get any errors. Is it possible that the Telemac 3D automatically increases the thickness of layer as per the fefault value = 1.E-6 m3 = 1 cm3?

2*2*0.16 = 0.64 < 1
I would suggest to switch from 1.E-6 (default) to 1.E-7 for keyword MINIMUM VOLUME OF 3D ELEMENTS.

There are divisions by volumes in TELEMAC-3D, that is why this minimum volume exists otherwise you may have division by 0 which is not good. It is a threshold below which the division is done with this minimum value. If you are curious, you can have a look at the MESH_PROP subroutine.

With MESH TRANSFORMATION = 2, there is no change in the vertical distribution if the thickness of layer is not big enough.

Having no errors does not mean the results are good!

Hope this helps,

Chi-Tuan

Thank you again.

Does the aspect ratio of the elements also play a role while deciding the distance between the planes near the bottom?

As told before, do not have planes further than the horizontal minimum distance. Elements should be as classical environmental flows you try to model: wider than higher.

Chi-Tuan

Hi,

Can you also tell me if the diameter of the bed material has any impact while deciding the minimum distacne of the verticle layers near the bottom?


Can the distance the the first two layers be smaller than the diameters of the sediment particles?

Hello Ravi,

My experience with sediment transport modelling is quite low. I would suggest you post this question in the GAIA/SISYPHE section. You may have an answer there.

Chi-Tuan