open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi all,
how can i change the coefficient for diffusion of tracer for TELEMAC2D with FUDAA?

Best regards

Massi

Hello,

I suppose that you just change the :

COEFFICIENT FOR DIFFUSION OF TRACERS

is there a problem and you do not see it in Fudaa ? Anyway in this case just add it in your steering file, with the value you want.

Regards,

JMH

Yes,i don't see the keyword.
Where exactly should i put it in steering file? And what is default value?

Best regards,

Massi

Hello,

Please have a look at the user manual, you can put the keyword where you want in the steering file, even 2 keywords on the same line, and the default value is 1.E-6.

Regards,

JMH

Thanks,i didn't know.

Hi all,
when i use in my tracer test default value for COEFFICIENT OF DIFFUSION FOR TRACERS, it work ok. If i change this value to 1.0E-2, it don't work and show this message:

GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 60
RELATIVE PRECISION: NaN

where is the problem?

Best regards,

Massi

Hello,

Strange, generally the diffusion linear system is easy to solve. You can try :

MAXIMUM NUMBER OF ITERATIONS FOR DIFFUSION OF TRACERS : 200

but little chance. Another possibility is that you use SUPG for the advection of tracers :

TYPE OF ADVECTION : *;*;2 (third number equal to 2)

then the diffusion linear system is not symmetric and you must have :

SOLVER FOR DIFFUSION OF TRACERS : 3 or 7 (a solver that works for non symmetric matrices), or try another schme for advection: 3 instead of 2 (or 13 if you have tidal flats).

If nothing works look at the results before it crashes, is it a problem on one point ? is it near an obtuse angle in a triangle ? (with obtuse angles diffusion becomes anti-diffusion).

Yet another (unlikely) possibility: your problem is a small physical model and 1.E-2 is too high.

I hope something works in this mist...

Jean-Michel Hervouet