open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

You should also check that the upper layers are not in the influence of the upstream conditions, in which case the turbulence that you see is just what you entered and not the result of the k-epsilon model, it sometimes takes long distances to establish.

Regards,

JMH

Hello,

I checked the result for both cases.

For the non-initialized case, the k and epsilon in the upper layers (Layer 4 to Layer 10) remain the minimum value along the whole channel (2km long).

For the initialized case, the k and epsilon evolve through the whole water column including the upper layers and reach the steady state after certain period.



Best regards,
Qilong

Hello,

I'm testing our model(with very thin bottom layer 2mm) with the new release 6.2. But unfortunately I still can't get the good logarithmic profile in k-epsilon model. You said you had a good logarithmic k-epsilon profile with a first layer of 2 mm. Is it possible to know what did your colleagues do in their model? Is there any rules about the vertical layer distribution? Thanks in advance!

Best regards,
Qilong

Hello Qilong,

Do you use tidal flats in your computation?
In v6p2 you should then check
THRESHOLD FOR VISCOSITY CORRECTION ON TIDAL FLATS
The value used to be fixed in visclip.f in v6p1. The default is 0.2 (meter). but for flumes we always lower this value down to a few centimetres or millimetres.
This might improve you computation.

Cheers
Annalena

Thank you Annalena! I have updated to v6p2 since this summer but I didn't use this keyword before. In fact I even didn't notice it exsits. Indeed it sounds very helpful for the computaion in the tidal flats and I do use tidal flats in my model. I'll test it in my case.

Thanks again!

Best regards,
Qilong