open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi everybody,

I was wandering if anybody had some experinece with using salinity in a 3D model as an active tracer AND use the sedi3d to model mud transport?
Can you use both at the same time in your model?


Best regards,

Sven

Hi Sven,

I am also interested in a similar set-up. I tried with a simple flume test model with a sediment release just below the water surface and the results seem weird in terms of BED EVOLUTION.

The bed evolution is almost constant in the whole domain while I do observe a clear deposition locally around my source if I do not include salinity. In other words, the bed evolution is a lot more diffused than with sediment only.

It would be good to have feedback from developpers and advanced T3D users.

This test was made with T3D only, I will try to couple it with Sisyphe.

Best regards
PL

Hi PL,

thanks for the info. I will sure we look into that.
Actually i'm working together with Qilong on this project. We're building a mud model for the Scheldt estuary and we start from the 3D model we already have. We are configuring the SEDI3D but i'm also coupling with SISYPHE so we can compare the results and speed of both with each other.

I'm looking at the code of the density law now to see if it is only applied on the salinity or not. I think there might be a problem.

I will report back later this week if we have some results.

Best regards,

Sven

Hi Sven,

Thanks for the info. What is Qilong's experience with salinity and sediment in 3D?

I also had a look at the density laws and none of them include the effect of suspended sediment. However, it is not hard to amend the available laws so that sediment concentration can be taken into account, it exists several simple equations / models to do it.

It is great if you can keep us informed about your progress. I will do it too.

Best regards
PL

Hi PL and Sven,

I'm currently working on SEDI-3D subroutines inside TELEMAC-3D.

I check the code and got the similar conclusion as yours.

The sediment is mainly treated as passive tracer except for one place. In drsurr.f near the end, sediment concentration affects the DELTAR, which is the RELATIVE DENSITY ((RHO-RHO0)/RHO0). Then this DELTAR will be used to calculate the buoyancy term in momentum equations and the turbulence models, such as in k-epsilon equations. But it actually doesn't change water density itself as I can tell from the code.

On the other hand, the density laws (1, 2 or 3) in the code only considers salinity and/or temperature. It recognize the right tracer by reading the name of the tracers. And it changes water density directly, therefore will change DELTAR, too.

Density law 4 seems only considers sediment concentrations.

If we combine salinity (density law 2) and sediment (SEDI=YES), it seems like the salinity contributes to the change of water density RHO, and then leads to a new DELTAR. Later in the same subroutine, this new DELTAR is being overwritten, based on RHO0, which is average water density defined as a constant in steering file. In this sense, I think the salinity effect has been left out.

If we want to keep both contributions to DELTAR, maybe we should use RHO instead of RHO0 in the following code?

Best regards,
Qilong

Hello,

I look through the code again and I found even for salinity (density law 2), it changes water density but this water density RHO doesn't replace the constant value RHO0 defined in the steering file. RHO is only used in calculating DELTAR and then discarded.

I suppose the water density doesn't change throughout the whole simulation. The water density is only modified within the subroutine drsurr.f and then used for computing DELTAR.

Best regards,
Qilong

Hello Qilong,

Thank you for your reply.
It will have a closer look and drsurr.f and to your comments.

I made some tests with:
1/ only sediment DENSITY LAW = 0
2/ sediment + salinity with 0 g/l with DENSITY LAW = 2

I was expecting to get similar results but in case 2 the sediment release (source just below water surface) in plunging quicker towards the bottom than case 1. The desposition is also much more diffused in the domain.





Please let me know if there is something I am missing. I am attaching the steering file of the second case. I will have a look at drsurr.f to check the density treatment with and without salinity + sediments.

Best regards
PL

Hi PL,

The steering file seems ok to me. I use similar settings for sediment. I think the problem again is in the subroutine drsuurr.f.

I made a calculation, assuming the sea water has defined density about 1025 kg/m3 and sediment concentration is 0.5 g/L. The new water density affected by salinity is 1030.

According to the equations in that subroutine:

1. Only consider sediment, DELTAR=0.5*(2650-1025)/(1025*2650)=0.000299
2. Sediment + salinity, DELTAR=(1030-1025)/1025+0.5*(2650-1025)/(1025*2650)=0.0052

There are big different here. The relative density increases about 17.38 times.

But I'm not sure if this is the real cause, since buoyancy term depends on the gradient of density.

I'll keep look at it and hope there are other opinions can be shared here.

Best,
Qilong

Hi all,

I had to leave this topic for a while but will be coming back to this very soon.
Any feedback from one of you? Sven, did you make tests?

All the best,
PL

Hi there,

I'm not sure if what I'll say here will help but here are some remarks about the way TELEMAC handles active tracers:

- the Boussinesq approximation is used, which is why everywhere in the equations the density is constant, equal to the reference density, but there is an additional buoyancy term in the equations which involves the dilatation/contraction coefficient, the relative tracer concentration (C-C_0) with C_0 the reference concentration and the relative density (rho -rho0)/rho_0.

- you can use the density law 4 to combine the effects of several scalars on the density. The equation of state linking the density to the scalar used in TELEMAC is linear. Any more complex behaviour has to be implemented by the user (e.g. non-linear density variation).

- in your case file, you need to specify the reference density and reference values for the tracers (e.g. water at 10 degrees C with salinity 35psu will have a reference density of about 1025 kg/m3) and the corresponding dilatation/contraction coefficients (beta=2e-4K^-1 for the temperature, -2.e-4psu^-1 for the salinity). I don't know what these values are for the sediment though.

I hope this helps.
Agnès