the difference between DES and WMLES?
 

Hi friends,
I got some confusion between DES and WMLES, I don`t know whether the WMLES belongs to DES or they are two different method of hybird RANS/LES. I think the WMLES is derivated from DES97 and DDES(somebody call it DES2006), so it is also one kind of DES. Because WMLES just modelled the boundary layer by RANS and this is same to DES.
Am I right?:)

you should take a look at this paper:

A hybrid RANS-LES approach with delayed-DES and wall-modelled LES capabilities
Mikhail L. Shur a, Philippe R. Spalart b, Mikhail Kh. Strelets a,*, Andrey K. Travin a
a New Technologies and Services, 14, Dobrolyubov Avenue, 197198 St. Petersburg, Russia
b Boeing Commercial Airplanes, P.O. Box 3707, Seattle, WA 98124, USA

Good luck!

Dear all,
I have a question regarding this old topic, or better I would like a clarification to understand if I what I have in mind is correct.

In fluent user guide, both IDDES and WMLES are referring to the same paper of Shur, where the IDDES formulation is specified and which consists of two branches, the DDES and the WMLES.

Anyway, from what I understood the two methods, DDES and WMLES, are both hybrid methods, coupling the RANS and the LES equation.

Now my clarification is about the difference between them which I will try to specify below, and hopefully someone can comment it:)

From what I understood the difference is that the DES (or DDES) method is trying to use the RANS equation in the entire BL, while the WMLES is using RANS only in the very near wall.
This is performed changing the way the characteristic length is defined.

Moreover I cannot find in the WMLES method which kind of RANS equations are applied. Also in Fluent the WMLES is written that it is solved using an algebraic method, which should be the Prandtl's mixing layer.
Is it correct?

I am very interested on the answer of your questions. I hope someone could answer them. I have compared DES (with k-w SST) and WMLES and they show very different characteristics. DES is much more similar to RANS, and WMLES is much more similar to other LES approaches such as the smagorinski and the dynamic ones.

How do you compare the cases to state that?

For reference to the future Googlers:



If you are talking about the commonly used term of WMLES, I believe you can summarize the difference as this:


DES has different governing equations, LES and RANS, within simulation itself. Like a mosaic cake if you will. Over here it is a LES, over there it is a RANS. (Of course this is a super-simplification, sadly I do not work with DES type methods so I cannot give a much clearer explanation)



WMLES, as the way it is most commonly understood in the field, has ONLY LES region. There is no region in the flow domain that is RANS, it is all LES. BUT, it does not use your good-old no-slip boundary condition on the wall because it is too coarse, instead it calls a magical wall-model near the wall, feeds it the flow information, and the wall-model returns a wall-shear stress value as the boundary condition to the LES equations. And, it is seen that, a very good candidate for this magical wall-model is a SEPARATE AUXILIARY RANS solution that covers only the distance between the wall and the wall-model location.



A note on terminology, WMLES literally is wall-modeled LES. Thus, you can also see people referring DES type methods as WMLES, because what they do is also model the wall, so they are not wrong with their language. A clearer distinction can be made between Hybrid RANS-LES methods and wall-stress models (which is what I called WMLES here).