[f0208secretx]

Hi,

This is rather a chat but not a technical question. And BTW the virtuoso guy wasn't me, so leave this if you supposed to GRADUATE this year. (just a joke)

I have a frustrating meeting in Dec. 31 with my supervisor. Aiming at the simulation of a HAWT, I was learning the numerical models by delivering a classic flow over cylinder problem. It should be a happy experience since you are learning new things with salary (full-time RA). However, my boss kept ask me to reproduce the same result of the literature. That's right, EXACT SAME results. Including the contour lines, the vectors and their pointing directions, and of course with all the plots and the data points should be identical with the presented results. As you can see, I'm kind of frustrated since I can't even do the 2D LES with Fluent right.... (I got k-e with good agreement)

I mean, it's that at all possible to reproduce the identical results? You guys can get really great agreement with the reference? Or that his was being pompous and exacting....

[julien.decharentenay]

From a strict scientific point of view, one should be able to reproduce the results from the literature - provided that all parameters are identical... In my experience, it is always a challenge as sometimes software do not give identical solution on different computer architecture (but the differences are minor).

Just pointing out that you are referring to 2D LES. There is no such thing as 2D LES (at least not within the framework of wind turbines)...

I personally recommend to build confidence using building block: start using simple (relevant) cases and increase complexity.

Good luck.
Julien

[f0208secretx]

Since you have mentioned, I was wondering why the 2D LES isn't existed? Is it because it resolves the big votices and in nature there are no 2D vortices? I would very like to know the physical and mathematical implication of this problem. Thank you.

Yeah I think building block is a good way to learn the possibilities and difficulties of the turbulence models. Thanks for your advices.

[julien.decharentenay]

In a nutshell, the behavior of turbulence is very different in 2D and 3D. You could look at turbulent cascade of energy (isotropic homogeneous turbulence would be good keywords) and you will find that in 3D the energy cascades from the large structure down to the small structures. In 2D there are 2 cascades of energy: (a) one from medium structures down to small ones and (b) one from medium to large structures.

LES (properly resolved, and with proper numerics) will have very different 2D and 3D behaviors since it will capture the above differences in transfer of turbulent energy.

2D LES is actually used for meteorological applications (think cyclones) and shallow water applications.

Just in case you may not be aware have a look at the Ercoftac classic database (and QNET) there are quite a number of reference test cases available and I am sure that at least a couple would be relevant to your application.

Good luck. Julien

[f0208secretx]

Thanks for your information. They were very useful.