[huangxianbei]

Hi,all:
I read the paper "Numerical Modeling Studies of Wake Vortices: Real Case Simulations" 1999 by Shen et,al. And find an extention of the Richardson number for the curvature or rotation effect. However, the expression "can't discriminate between a shear flow and flow with coherent rotation, a discriminator function is applied...". Can anyone tell me something about how to discriminate this ?

Thanks

Xianbei

[huangxianbei]

Hi, I just have a idea to distinguish them. As in pure plane shear flow, the vorticity should be 0, while in flow with coherent rotation, the vorticity should be non-zero. Is it can be the discriminator?

[Jan.Östh]

It sounds like you're dealing with vortex definitions which is a bit tricky.
Check out the work of George Haller, for instance:

"An objective definition of a vortex", Haller, G. J. Fluid Mech. (2005), vol. 525, pp. 1–26

That publication will also lead you to earlier work, such as the work of Jeong, Hussain and Robinson. It's great stuff.

[huangxianbei]

Originally Posted by Jan.Östh

It sounds like you're dealing with vortex definitions which is a bit tricky.
Check out the work of George Haller, for instance:

"An objective definition of a vortex", Haller, G. J. Fluid Mech. (2005), vol. 525, pp. 1–26

That publication will also lead you to earlier work, such as the work of Jeong, Hussain and Robinson. It's great stuff.

Hi,Jan.Östh:
Thank you very much for the reference. Vortex definition is a little bit complex which has no good criteria to define it. My purpose is to distinguish the flow region between plane shear flow and flow with coherent rotation to implement a curvature correction of the flow field. So what I need is a discriminator function to fulfill this. Do you know something about it?

Xianbei

[FMDenaro]

Originally Posted by huangxianbei

Hi,Jan.Östh:
Thank you very much for the reference. Vortex definition is a little bit complex which has no good criteria to define it. My purpose is to distinguish the flow region between plane shear flow and flow with coherent rotation to implement a curvature correction of the flow field. So what I need is a discriminator function to fulfill this. Do you know something about it?

Xianbei

you can identify a vortex core, for example using the method of
http://journals.cambridge.org/action...22112095000462

that helps you distinguish between rotational and shear flows.
However, "coherence" is something different, implying statistical analysis

[agd]

Are you looking for something like the Kato-Launder modification

http://www.cfd-online.com/Wiki/Kato-...r_modification

[huangxianbei]

Originally Posted by agd

Are you looking for something like the Kato-Launder modification

http://www.cfd-online.com/Wiki/Kato-...r_modification

Hi,agd:
Thank you for the link. Kato-Launder modification is an interesting modification for k-e model, while I'd like to implement a curvature/rotation correction to LES model, so the main problem is to define the flow region: whether it's rotational or shear flow.

Xianbei

Originally Posted by FMDenaro

you can identify a vortex core, for example using the method of
http://journals.cambridge.org/action...22112095000462

that helps you distinguish between rotational and shear flows.
However, "coherence" is something different, implying statistical analysis

Hi,Filippo:
Thank you for your reply. So in your opinion, the way to implement this is to identify a vortex, right? Then the discriminator becomes a function for identifying vortex: when a vortex appears, the flow is defined as rotational, otherwise, it's shear flow.

Xianbei

[FMDenaro]

Originally Posted by huangxianbei

Hi,agd:
Thank you for the link. Kato-Launder modification is an interesting modification for k-e model, while I'd like to implement a curvature/rotation correction to LES model, so the main problem is to define the flow region: whether it's rotational or shear flow.

Xianbei

I don't think your problem is well posed for LES ....
The sgs model does not need to act differently in rotational and shear zones as it is supposed to model only small energetic isotropic scales of the motion

[huangxianbei]

Originally Posted by FMDenaro

I don't think your problem is well posed for LES ....
The sgs model does not need to act differently in rotational and shear zones as it is supposed to model only small energetic isotropic scales of the motion

Hi,Filippo:
Yes, the sgs model is to model small scales of the motion. However,it would be better if additional effects are taken into account. This would improve the complexity and accuracy of the model. In fact, there are many works published focusing on the rotational effect in the sgs model according to the SGS decomposition proposed by Novikov,1992.

Xianbei

[FMDenaro]

Originally Posted by huangxianbei

Hi,Filippo:
Yes, the sgs model is to model small scales of the motion. However,it would be better if additional effects are taken into account. This would improve the complexity and accuracy of the model. In fact, there are many works published focusing on the rotational effect in the sgs model according to the SGS decomposition proposed by Novikov,1992.

Xianbei

The model of Lund and Novikov is a non-linear SGS model that is very computational expensive showing no real improvement in practical computational cases...Nowday, I have no news about a practical use in present LES applications. However, it is based on the [S, Omega] tensors that discriminate between shear and rotationale part of the flow.

[huangxianbei]

Originally Posted by FMDenaro

The model of Lund and Novikov is a non-linear SGS model that is very computational expensive showing no real improvement in practical computational cases...Nowday, I have no news about a practical use in present LES applications. However, it is based on the [S, Omega] tensors that discriminate between shear and rotationale part of the flow.

Z.X Yang, et al. Large eddy simulation of rotating turbulent channel flow with a new dynamic global-coefficient nonlinear subgrid stress model,2012. This is a combination of non-linear SGS model and the model proposed by Vreman.

Well,let's go back to the original problem. I'd like to do a modification of SGS viscosity considering the rotation/curvature effect, so the modification should be only applied to the rotational flow. As you suggested in the former reply, a vortex core would be used to define the rotating part. So the method I should use is the method of defining vortex core to judge whether it's rotational.