open TELEMAC-MASCARET The mathematically superior suite of solvers

Dear Users,

Could anyone please advise on how to keep stable stratification in closed domain. I have two stratified layers (layers 1 to 7 with profile Salinity - 35 and Temperature - 4C; layers 8 to 10 with profile Salinity -25 and Temperature 1C). The domain is closed, no open boundaries. At the beginning of simulation the stratification is stable, but after day 6 the layers profile is smoothing - f.e. Salinity becomes 33.33 everywhere. I have run it for one year,but really 6 days is enough to break it. In this case I have very steep bathymetry. However, I have run similar case with same steering file for closed rectangular tank and stratification was stable for whole year. Please advise, maybe I have used wrong diffusion/advections schemes. Thanks in advance. Please see two pics attached, at beginning and end of run, also results for stratification in a tank after one year run.

Kind Regards!

Violeta

Hi Violeta

In the tank case, the planes of yoru mesh are automatically horizontal.
In your real case, I suppose it's not necessary the case (depends on the choice you made on the vertical discretization). you can see this effect at time 0.
Maybe you should look at this particular point and try to use fix planes to avoid spurious diffusion of tracer

Hope this helps

Hi Christophe,

Thank you very much for your reply. I had adaptable mesh, Mesh transformation = 5. I have changed it to 1, it is running now. If it won't work then I will have to change it to 3. I remember I had tried to use 3 once, but code did not work. I will wait for results first. Thanks again.

Kind Regards!

Violeta

Hi Christophe,

I have run the case with Mesh transformation=1 for 10 days, but with same results as earlier, the Stratification is not stable. The surface was not distorted as I was afraid. Could it be there something else to adjust to improve results? Thanks in advance.

Kind Regards!

Violeta

Hello,

Did you try:

HYDROSTATIC INCONSISTENCY FILTER : YES

it cancels the tracer gradients when they are artificial and due to the mesh distortion. If it has an effect then you can try to look at subroutine trisou.f, where you have FORMUL(6:6)='2', this is a hardcoded option of the filter, you can also try 3 and 4.

Regards,

JMH

Thank you Jean-Michel. Yes, I already have Hydrostatic Inconsistency Filter. I have just added subroutine you recommended to the main fortan file and changed Formul to 4 for start. I hope it will do the trick. Thanks again. I will let you know if any problems.

Kind Regards!

Violeta

Dear Jean-Michel/Christophe,

Thanks again for helping me with this case. I have tried Formul=3 and 4 in TRISOU subroutine, and 3 gives better results, the stratification is more stable at the beginning but then it still mixes. I have also tried Mesh Transformation=1 but it did not solve the problem and now I am trying Mesh Transformation=3, same example as in Telemac3d Operating Manual. I am trying to recreate same mesh as on Figure 4, page 30. I have used the code below for the start, jut to check. The total number of layers in my case is 10, but for some reasons below code gives only 5. As per below code, all 5 layers are unevenly spaced but it ignores the rest. See the picture of my mesh. I have specified Mesh transformation as 3, but in code below it is actually combination of 1,2 and 3. Looks like it is reading only option 2. Please advise how to amend the code so all 10 layers would apply. Thank you very much.

DO IPLAN = 1,5
TRANSF_PLANE%I(IPLAN)=2
ENDDO
ZSTAR%R(1)=0.D0
ZSTAR%R(2)=0.2D0
ZSTAR%R(3)=0.5D0
ZSTAR%R(4)=0.7D0
ZSTAR%R(5)=0.8D0
ZSTAR%R(6)=1.D0
DO IPLAN = 7,NPLAN
TRANSF_PLANE%I(IPLAN)=1
ENDDO
TRANSF_PLANE%I(6)=3
ZPLANE%R(6)=0.D0

Kind Regards!

Violeta

Jean-Michele/Christope,

Please ignore my last post. I have used another code for Mesh Transformation =3 and salinity/temperature works fine, well for 2 days so far! I will re-run it for longer and fingers crossed it will be fine. Thanks again for looking into it.

Kind Regards!

Violeta