open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello!

I chime in here as it is a topic I am interested in.
Unless this has been solved recently, I think that in some configurations it can be hard to combine stratification (salt and or temperature) with tidal flats. When some elements are drying out, the small water depths combined with the local temp and/or salt values could affect the density calculation and hence impacting other vaariables during the solving step. It seems that the 2nd and 3rd issues could be caused by such a phenomenon?

What do you say Chi-Tuan and Costas?

Kind regards
PL

Hello,
Looking up through the thread the bed friction in the model appears to be very unrealistic. A Manning friction coefficient of 0.12 is being used. I wonder if this is a mistake? A value of 0.012 would be more sensible. This could be causing instability on the intertidal areas.
Regards
Tom Benson

Hello all,

Valuable answers from the experienced users Costas, Pierre-Louis and Tom.

Tracers and tidal flats is still an issue with TELEMAC-3D which may give bad results.

Anyway, you have changed some parameters since 1st sending:
- the FRICTION COEFFICIENT FOR THE BOTTOM for a Manning law was more realistic with a 0.04 value rather than 0.12 as Tom wrote,
- I would recommend to use OPTION FOR THE TREATMENT OF TIDAL FLATS = 1 rather than 2 (frozen values). I did not even know that it was possible to set this value with TREATMENT OF NEGATIVE DEPTHS = 2.

Before trying to optimise the options to decrease CPU time, I would suggest to investigate to avoid or decrease your issues.

Needing MAXIMUM NUMBER OF ITERATIONS FOR ADVECTION SCHEMES = 1500 and reaching this value seem to be the main issue. 1500 is quite big. Using a value greater than 200 is not so usual (only one test case in the examples database), and 200 is already quite big. It often means that you should decrease the time step...

Can you try to change the advection schemes for velocities + k-epsilon : with characteristics first, then try LIPS (SCHEME FOR ADVECTION OF... = 5 + SCHEME FOR ADVECTION OF... = 4) for velocities (and characteristics for k-epsilon).
Then try the same but with LIPS for tracers?
MAXIMUM NUMBER OF ITERATIONS FOR ADVECTION SCHEMES is used for schemes 13 and 14 only.

What about the boundary where you can see big velocities? Is it always wet or sometimes dry? What kind of boundary? Maritime? River? Shoreline?

Same question for your 2nd point: is it at the boundary or inside the computational domain?

Hope this helps,

Chi-Tuan

PS: The 3 keywords TREATMENT ON TIDAL FLATS FOR VELOCITIES/K-EPSILON/TRACERS should not be useful as only used with OPTION FOR THE TREATMENT OF TIDAL FLATS = 2 or ELEMENTS MASKED BY USER = YES, what is not recommended. You can delete these keywords from you steering file.

Hello Chi-Tuan!

Sorry, I go back to the stratification / tidal flat issue. If the problem is generated by erroneous densities, could that be easily fixed by simply bounding the densities with physical min/max values where water depth becomes small? Or can the problem appear elsewhere (ie. even with passive tracers)?

Kind regards
PL

Hello Pierre-Louis,

To be honnest, I do not know. But clipping variables may change other properties that are searched. It should be investigated.
If you have any experience or tries, do not hesitate to share.

Best regards,

Chi-Tuan