open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi everybody,

I want to model a channel with sediment traps along its length, this, using flow deflectors (obstacles) that redirect the flow, getting lower velocities close of these deflectors, making fall the sediment.

Before doing this, i am modeling a channel with a flat and non erodible bottom, just studying the effect of a deflector on suspended load transport, not considering bed load.

In this simple channel that i am using to see the behaviour with just one deflector, i have gotten some results that do not look good.

I am indicating the settling velocity in my cas file (using the equation in vitchu_sisyphe.f), this value is lower than the velocity in almost all the domain, trying to make fall the suspended sediment on the zones close to the deflector, where velocity is lower than the fall velocity, despite of this, a really small quantity fall on zones where velocity is higher than fall velocity (e.g. at inlet), so, i would like to know how to prevent this (why does it happen?).

Also, there are zones where velocity is lower and does not fall sediment as i think should.

I know you could help with this.

PS: The computation requires using a small time step (0.01), does anybody know how i could get shorter computation times with that small time step?

Regards,

Jose Morales

Hi Jose,
Please find below some tips/suggestions:
1. check if your mesh if fine enough in the vicinity/around the deflectors
2. is the value of the imposed settling velocity consistent with your sediment size? how did you estimate this value?
3. I'd suggest to use an advection scheme different from the default, e.g. the numbers 14 or 15 (take a look to the user's manual)
4. take a look to the shear velocity u*. By doing some postprocessing you can plot u*/ws (ws = settling velocity). This will give you an overview of the zones prone to deposit sediment (the good old' trick by Bagnold...).

cheers,

pablo

Hi Pablo,

About settling velocity, the velocity which I have to compare with to know the zones prone to deposit sediment, is the friction velocity? I thought the flow velocity (its magnitude) was the velocity to compare with, could you give me an explanation about why is the US velocity instead of UV velocity?

Thanks in advance,

Cheers,

José

Dear José,
Good question! The shear velocity is defined as (I'm using Latex notation here):
\rho u_*^2 = \tau_b, with \tau_b the boundary shear stress.
The criterion for suspension is that the vertical turbulent velocities must be at least as large as the settling velocities of the sediment particles. From observations, it was assumed that the rootmean-square value of the vertical turbulent velocities is a good measure to quantify these velocities.
Measurements (done in the 70's) in turbulent boundary-layer flows past both smooth and rough boundaries have shown that there is a maximum close to the bed and that the maximum values reached are proportional to the shear velocity.
As \tau_b is function of the gradient of the velocity component (with respect to z) and the fluctuating components of the velocity, this can also be shown theoretically (too long to show here, good exercise to your fluid mechanics course )

Cheers,

Pablo

Hi Pablo,

Thanks in advance for the book you recommended me, that is what i was looking for.

Also, I am working with a d50=2.E-4m in my model, which has a ws=0.26 approximately, but despite of the friction velocity at upstream is larger than the settling velocity (U* is about ws) the suspended sediment fall on the inlet, it is like the velocity is not large enough to keep the sediment suspended.

What i want to ask is, what is a really good proportion of the ws to be ensured that the sediment won't fall for the friction velocity, i mean, if U*=k.ws, which one would be an appropriate value of k to stay sure that the sediment won¿t fall?

I hope being clear about the above,

Thanks in advance,

PS: I will send you an email with the files of my problem (I am using FV method).

Cheers,

José

Sorry, the settling velocity for d50=0.026 approximately.

Cheers,

José

Hi José,
It depends on the authors/data available. In the literature I've seen values of 0.80 (I couldn't find the reference).
Probably you can take a look.

Best wishes,

Pablo

Dear Pablo,

I would like to set bed load transport and suspended load transport (both simultaneously), with a different grain diameter each one (e.g. 0.4mm for suspended and 1mm for the bed), is this possible? I see in the manual that for non-uniform sediment distribution for suspended sediment transport i have to set BED LOAD =NO. So, could i set my model as i mentioned?

Thanks in advance,

Cheers,

José

Hi Pablo,

I already could set both evolution and suspended with different size each one. Also, I wanted to ask you about the maximal concentration in my model (see the picture), despite of i set 1.E-6 as a prescribed value at the inlet i am getting lower values during the simulation. Could be this happening because of two different grain sizes?

As usual, thanks in advance,

Cheers,

José

The files were not attached. Find attached Sisyphe steering file and the image i commented.