open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi all,

I keep receiving the error:
"EXCEEDING MAXIMUM ITERATIONS: 200 RELATIVE PRECISION NaN"

As far as I can tell this is because I'm using both the K-epsilon option and tidal flats.

Initially I reduced the time step, but the error still occurs even at very small values.

I also altered the keywords in the steering file (e.g. SOLVER; IMPLICITATION FOR DEPTH; SOLVER FOR VERTICAL VELOCITY; SCHEME FOR ADVECTION OF VELOCITIES - see *.cas file attached.) Playing around with this didn't resolve the issue.

I've also changed EMIN and KMIN values in cstkep.f. Despite playing with the values (e.g. I've used 1.D-10, and 1.D-7) nothing has changed.

Does anyone know of a way to get around this?

Thanks,
Simon

Hello,

try to use another advection scheme for k and epsilon, like this on:
SCHEME FOR ADVECTION OF K-EPSILON =14

Furthermore depending on your expected averaged water depths in the domain (laboratory scale or field scale) you should consider to lower the THRESHOLD FOR VISCOSITY CORRECTION ON TIDAL FLATS which by default is 0.2.

Hope this helps,
Clemens

Hi,

did you manage to solve this? I am running into the same problem - model crashing when using k-epsilon with tidal flat.

thanks.

Ed

Hi,

It was a while ago, so I can't remember the exact solution. I'm sure it was something numerical in the *.cas file that I was using.
It may be due to the advection scheme you're using (I tend to use LIPS - scheme 5), the preconditioning (PRECONDITIONING FOR K-EPSILON = 17 seems to be ok), and whether you have tidal flats i.e. very shallow or dry areas (TIDAL FLATS = YES; TREATMENT OF NEGATIVE DEPTHS = 2; OPTION FOR THE TREATMENT OF TIDAL FLATS = 2 solved some initial problems for me too).
I often consult the example *.cas files or others people have shared on the forum for inspiration in tricky cases.

Otherwise it may be mesh related. CHECKING THE MESH = YES is a good place to start to see if there are any issues.

Regards

okay. I will look into these parameters. thanks.

Ed