open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi,

I run the Telemac3D model with one tracer salinity.

I try the k-e model but it always stop with the error:
"GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS
Note: The following floating-point exceptions are signalling: IEEE_UNDERFLOW_FLAG IEEE_DENORMAL"

I try the Mixing length and Smagorinsky vertical turbulence model, then I got two results.

The tide(water level) and current(velocity) are verified good with the observed data.

But the vertical salinity is not stratificated in the estuary area.
The value of surface salinity is the same as the bottom`s.

In general, the surface salinity should be 10-20 psu smaller than the bottom`s.

How to model the salinity stratification in the estuary area ?
By the way, how to output the vertical diffusion of tracer ?

Thank you in advance!

Best regards

He Jiawei

(PS：希望更多在中国使用TELEMAC的朋友进QQ群219764703互相交流学习)

Hello He Jiawei

The solution here is, at first, quite simple. The problem is that the solver (GRACJG, in this case) is exceeding the maximum number of iterations before reaching the desired precision. To solve this just increase this number of iterations or reduce the precision, whichever best suit your needs.

But I cannot tell at which solution step this is happening. If it is, for example, at the diffusion of the tracers (since you mentioned the salinity), you have to either increase the MAXIMUM NUMBER OF ITERATIONS FOR DIFFUSION OF TRACERS (default = 60), or decrease the ACCURACY FOR DIFFUSION OF TRACERS (default = 1e-6), until the solver converges.

From what I have already seen, bigger meshes (such as yours) tent to take more iterations of the solver to converge. However, keep in mind that if even these modifications don't work, maybe you'll have to check the overall quality of your mesh, because then this may be what is spoiling the solution for your simulation.

Hope this helps.

Best regards,

Phelype

Hello,

In the k-e model, I test it and find out that the GRACJG error comes from DIFFUSION OF VELOCITIES.

When I use the Mixing length model or Smagorinsky model, I could get the good water lever and u-v velocity results.
But the salinity is well-mixed.

I want to model the salinity stratification with the Mixing length model or Smagorinsky model.
Please help me to modify my steering file.

By the way, how to output the vertical diffusion of tracer into the result file ?

Thank you in advance!

Best regards

He Jiawei

Well, if you are sure the error arises at the diffusion of velocities step, then just add the following to your steering file:

This value of 120 you modify as you need. If the number of iterations is getting too high, then you could change the accuracy:

Again, if that doesn't work either (i.e. Even if you change these values, the solver don't converge in a reasonable amount of iterations), then you should check the quality of your mesh and the density and distribution of the vertical layers.

I don't know if I understand your question here, but if you want to save the concentration field of the tracer you are using, just add TAx (x is the number of the tracer) to the VARIABLES FOR 3D GRAPHIC PRINTOUTS. As I can see in your steering file, you are only using one tracer, Salinity, and already have TA1 in the above keyword, so the salinity should be present in your results file.

Hope this helps.

Best regards,

Phelype