open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

I did a suspended load transport simulation using Telemac-2D coupled with Sisyphe, v7p0r1 by prescribing suspended load inflow using the liquid boundary file. Simulation runs well and reasonable results, however I don't understand the listing printout:

*** FINAL BALANCE FOR TRACER 1 ***
INITIAL QUANTITY OF 1 : 0.000000 M3
FINAL QUANTITY : 25.78517 M3
QUANTITY ENTERED THROUGH LIQ. BND. : 140.7582 M3
TOTAL MASS OF DEPOSIT : 100.1235 M3
TOTAL QUANTITY LOST : 14.84955 M3
...
SUM OF THE CUMULATED EVOLUTIONS : 166.8724
VOLUME THAT ENTERED THE DOMAIN : 0.000000 M3 ( IF <0 EXIT )
VOLUME DEPOSITED ON THE BOTTOM : 166.8724 M3 ( IF <0 ERODED )
LOST VOLUME : 0.4121148E-11 M3 ( IF <0 EXIT )

What means TOTAL QUANTITY LOST in the final balance of tracer? It is to worry about when it gives values in the order of m³ for a two day simulation? Beacause on the other side below the LOST VOLUME gives a value in the order of machine accuracy..

I would be glad for any hints!
Clemens

Hello,

The first mass balance is from Telemac-2D, the second is from Sisyphe. Only the message from Telemac-2D seems to report an error. Some hints:

1) Did you ask for : CONTINUITY CORRECTION = YES

2) Do you have a mass conservative advection scheme for tracers, i.e. not characteristics (1).

3) Do you have tidal flats ? (in which case only advection scheme 13 or 14 (the same) would work ?

4) Do you have a large diffusion coefficient ? (if yes try with coefficient =0.)

5) Did you ask for OPTION FOR FLUXES AT THE BOUNDARIES = 2;2;2... (not sure of the name of this keyword, I'm on holidays with (and exceptionnally today) little Internet connection).

With best regards,

Jean-Michel

Hello Jean-Michel,

thank you for the answer and your time.

In Telemac-2D I used MOC for the velocities and NERD for the tracer as well as for suspension in Sisyphe. In Telemac I used the Elder model (because of very agreement with ADCP measurements), in Sisyphe the dispersion coefficients are taken from the computed coefficients by Telemac.

I did some tests with different diffusion as well as dispersion coefficients, Treatment of fluxes at the boundaries, higher solver accuracies and other parameters but the results are more or less the same.


Best regards and I wish you a nice holiday!
Clemens

Hello Jean-Michel,

for the previous post I did only serial testing of different options.

The combination of the keyword TREATMENT OF FLUXES AT THE BOUNDARIES = 2 for the open boundaries AND setting the dispersion coefficients to very low values does the trick.


Thank you,
Clemens