[FMDenaro]

Originally Posted by juliom

Dear Mr Sridar. I read you post from 1999. I am just starting my PhD work in LES but after reading a pile of paper I still have more questions than answers. First of all I am struggleing with the filter operation, although I understand the physics behind the filtering I do not visualize how to implement it from the numerical point of vie (code). I still see the LES like something very difficult to code, but after reading some papers from Moin I have the feeling that this is like a URANS with a different in the way I compute the SGS or its equivalent if I compaer it with URANS (turbulent viscosity).
Secondly, I have read about DNS and LES and everythime I read something about LES vs DNS I see that the solution from DNS is filters... WHYYYY????.. This question also complement my following question. Can I compare experimental data with raw data from LES. From the physics point of view the LES requires a filtering to define a kind of 'cut-off'. However, the SGS models models the eddies below the cut -off widht. It means that I am solving explicitly one portion of the energy spectrum and modeling another portion of the spectrum. It means, that my final flow field (pressure, velocity and so forth) represent the entire solution. From this point of view I could comapre LES solution with any CFD solution and even with experimental data withouth performing any filtering at all...
How lost am I?
I really appreciate your help!!!

Hello,
I strongly suggest to read some topics in this forume about filtering,LES, DNS, many if not all answers are here...
good search

[juliom]

Thanks professor, I have already done. However, the more I read the less I understand...

[FMDenaro]

Originally Posted by juliom

Thanks professor, I have already done. However, the more I read the less I understand...

ok, when you have focused on the more obscure points open a thread

[juliom]

Originally Posted by FMDenaro

ok, when you have focused on the more obscure points open a thread

Is it difficult to make a contribution to my question????

I expected a more possitive contribution !!!! thank you very much

[FMDenaro]

Originally Posted by juliom

Is it difficult to make a contribution to my question????

I expected a more possitive contribution !!!! thank you very much

no but you are using a thread with different title, not focusing on your requirement

[Mirage]

I went through all the replies and other threads and papers and I am not seeing any clear answer about the requirement of LES grid.
Generally we can say that the grid resolution is depending on the physical problem and it is for me clear that the filtering and accuracy of the results is depending on the the smallest cell, used as parameter to specify the range of the modeled energy spectrum.
However I am not finding any paper, book, reference, which gives a clear and simple explanation or correlation about the mesh requirement for LES !
I was wondering, if someone could give an answer to this question

Thanks

[FMDenaro]

Originally Posted by Mirage

I went through all the replies and other threads and papers and I am seeing any clear answer about the requirement of LES grid.
Generally we can say that the grid resolution is depending on the physical problem and it is for me clear that the filtering and accuracy of the results is depending on the the smallest cell, used as parameter to specify the range of the modeled energy spectrum.
However I am not finding any paper, book, reference, which gives a clear and simple explanation or correlation about the mesh requirement for LES !
I was wondering, if someone could give an answer to this question

Thanks

what is not clear? the mesh introduces a projective filter at a frequency corresponding to the Nyquist cut-off. In LES one must ensure that this frequency lies in the intertial range of the energy cascade

[juliom]

Professor Denaro has answered your question. My opinion is that the literature shows more information about the mesh requirement for DNS than LES. In LES your need to make sure that the cut-off frequency imposed by the mesh lies in the inertial su range. Pope's book is a good reference.

[Mirage]

Originally Posted by FMDenaro

what is not clear? the mesh introduces a projective filter at a frequency corresponding to the Nyquist cut-off. In LES one must ensure that this frequency lies in the intertial range of the energy cascade

Thank you for quick reply I totally agree with you.
Let me try to specify what I do not understand:

I was wondering how to verify that my grid is appropriate for a LES Simulation. Many references mentioned that y⁺ should be small than 1 , if the wall modeling is not used which I can totally understand. Do you know if other rules for Δx⁺and Δz⁺ exist?

Could you please share with us, which steps are you taking to be sure that the mesh is good enough for a LES?

Which grid parameters should I take in account when i generate my mesh and how can verify after the solving that my results are acceptable in the framework of LES results.

Thank you so much !

[FMDenaro]

you should consider the flow problem.... if you are working in fully confined flows, no matter of the directions, x+,y+ and z+ are measured from the walls and you have to work with a DNS-like grid near the walls for a resolved BL. That means 3-4 nodes at least within x+,y+,z+=O(1).

If your problem has some directions of homogeneity the LES grid has a typical dimension of h+ = 20-30

[FMDenaro]

https://www.researchgate.net/publica...ates_revisited

[Mirage]

Thank you very much !

[medaouarwalid]

Originally Posted by D.Reynolds
;3452

DNS directly solves the Navier-Stokes equations capturing all eddies from the length scale of the grid geometry right down to the Kolmogorov length scales (relating to the smallest eddies). The dx,dy,dz (=dL) of the mesh needs to be small enough to capture eddies down to the Kolmogorov length scale.

The argument for the cell resolution, (and thus dL) goes something like this:

computational box of length L

number of grid points in ONE direction, N

grid spacing dL

Kolmogorov length scale, eta

Molecular viscosity, mu

Energy dissipation rate, epsilon

rms turbulent velocity scale, u'

----------

For a box of length L, the number of points depends on dL:

N = L / dL --------------(1)

dL must be small enough to resolve the smallest eddies, which have the length scale eta. Thus dL=eta is the maximum value for dL to capture the smallest eddies without them `dropping through' the grid (idealy dL= 0.5 * eta for better resolution).

Thus (1) becomes:

N(min) = L / dL(max) = L / eta --------(2)

Eta is defined as: eta = ( mu^3 / epsilon )^(1/4) --------(3)

Epsilon defined as: epsilon = u'^3 / L --------(4)

Substituting (3) & (4) into (2) gives:

N = ( u' * L / mu )^(3/4)

Noting that u'L/mu is a form of Reynolds Number this gives: N^3 = Re^(9/4)

Knowing the Reynolds number, Re and the geometry size, L enables you to make a rough estimate of dL.

---------------------------

The papers I've seen are concerned with making the cell size near the wall small enough to get the first few mesh points (this is related to pi*eta (or pi*dL) thus the first 3 points) in the viscous sub-layer. This is why they are always mentioning the y+ values of the near wall cells.

(3)&(4) See standard turbulence textbooks for definitions.

For papers discussing this matter:

Eggels et al JOURNAL OF FLUID MECH. VOL 268 pp175-209 (page 179 specifically). A paper on DNS of turbulent pipe flow

Kim, Moin, Moser JOURNAL OF FLUID MECH. VOL 177 pp133-166 (p135, there is also a reference to Moser&Moin 1984 (internal Stanford report)). A paper on DNS of channel flow

An LES computational grid only needs a dL small enough to resolve the large scale flow structures (e.g a recirculation bubble). Any structures smaller than this are passed on to the subgrid scale (SGS) model.

Hope this helps, Denver

How do we could have an idea about the size of the the large scale flow structures especially if experimental devices can not show it ? Does exist a formula for LES to calculate N like that one mentionned above for the DNS ?

[saip]

Hi Sridhar,

Can you send me your email address...

[Gang Wang]

Originally Posted by D.Reynolds
;3452

DNS directly solves the Navier-Stokes equations capturing all eddies from the length scale of the grid geometry right down to the Kolmogorov length scales (relating to the smallest eddies). The dx,dy,dz (=dL) of the mesh needs to be small enough to capture eddies down to the Kolmogorov length scale.

The argument for the cell resolution, (and thus dL) goes something like this:

computational box of length L

number of grid points in ONE direction, N

grid spacing dL

Kolmogorov length scale, eta

Molecular viscosity, mu

Energy dissipation rate, epsilon

rms turbulent velocity scale, u'

-------------------

For a box of length L, the number of points depends on dL:

N = L / dL --------------(1)

dL must be small enough to resolve the smallest eddies, which have the length scale eta. Thus dL=eta is the maximum value for dL to capture the smallest eddies without them `dropping through' the grid (idealy dL= 0.5 * eta for better resolution).

Thus (1) becomes:

N(min) = L / dL(max) = L / eta --------(2)

Eta is defined as: eta = ( mu^3 / epsilon )^(1/4) --------(3)

Epsilon defined as: epsilon = u'^3 / L --------(4)

Substituting (3) & (4) into (2) gives:

N = ( u' * L / mu )^(3/4)

Noting that u'L/mu is a form of Reynolds Number this gives: N^3 = Re^(9/4)

Knowing the Reynolds number, Re and the geometry size, L enables you to make a rough estimate of dL.

---------------------------

The papers I've seen are concerned with making the cell size near the wall small enough to get the first few mesh points (this is related to pi*eta (or pi*dL) thus the first 3 points) in the viscous sub-layer. This is why they are always mentioning the y+ values of the near wall cells.

(3)&(4) See standard turbulence textbooks for definitions.

For papers discussing this matter:

Eggels et al JOURNAL OF FLUID MECH. VOL 268 pp175-209 (page 179 specifically). A paper on DNS of turbulent pipe flow

Kim, Moin, Moser JOURNAL OF FLUID MECH. VOL 177 pp133-166 (p135, there is also a reference to Moser&Moin 1984 (internal Stanford report)). A paper on DNS of channel flow

An LES computational grid only needs a dL small enough to resolve the large scale flow structures (e.g a recirculation bubble). Any structures smaller than this are passed on to the subgrid scale (SGS) model.

Hope this helps, Denver

Hi every former!

I know it's an old thread but I'm still quite interested in this topic about mesh size, Kolmogrove scale and DNS. Like Mr. D. Reynolds said, epsilon is associated with the eta. But through the epsilon expression Mr. D. Reynolds post, epsilon = u'^3 / L. I think this epsilon is a resolved part of epsilon, right? It's closely related to velocity flunctuation, but what about L? Is L the computational length? Can anyone give me some suggestions about how to get epilson in DNS or LES? I don't think it's a modelled one.

Best,
Gang Wang

[Gang Wang]

Originally Posted by Gang Wang

Hi every former!

I know it's an old thread but I'm still quite interested in this topic about mesh size, Kolmogrove scale and DNS. Like Mr. D. Reynolds said, epsilon is associated with the eta. But through the epsilon expression Mr. D. Reynolds post, epsilon = u'^3 / L. I think this epsilon is a resolved part of epsilon, right? It's closely related to velocity flunctuation, but what about L? Is L the computational length? Can anyone give me some suggestions about how to get epilson in DNS or LES? I don't think it's a modelled one.

Best,
Gang Wang

In addition, there are guys goes that like this expression, totally different with the Mr. D. Reynolds said.
https://physics.stackexchange.com/qu...v-length-scale

"From the quotes poem, you can anticipate that everything that is dissipated at the smallest scales, has to be present at larger scale first. Therefor, as a very crude estimate, for a system of length 𝐿 and size 𝑈 (and dimensional grounds, on this scale viscosity does not play a role!),"

Which one is right?

Best,
Gang

[FMDenaro]

The viscosity of the fluid plays a fundamental role in defining the Kolmogorov lenght scale.
Just as a rapid estimation, at this scale corresponds a Re number of O(1).

[Gang Wang]

Originally Posted by FMDenaro

The viscosity of the fluid plays a fundamental role in defining the Kolmogorov lenght scale.
Just as a rapid estimation, at this scale corresponds a Re number of O(1).

Hi!

Yes, I think we could also estimate is roughly using Re^(3/4)* (The maximum eddy scale). But I was also wondering what's the maximum eddy scale in my simulation? If I used a box computational domain, is this the largest side length of my domain?

Best,
Gang

[FMDenaro]

Originally Posted by Gang Wang

Hi!

Yes, I think we could also estimate is roughly using Re^(3/4)* (The maximum eddy scale). But I was also wondering what's the maximum eddy scale in my simulation? If I used a box computational domain, is this the largest side length of my domain?

Best,
Gang

What you want is the integral length scale L that depends on the specific flow problem. That means you should ensure that your domain has a dimension such to describe properly the structure of dimension L.

[Gang Wang]

Originally Posted by FMDenaro

What you want is the integral length scale L that depends on the specific flow problem. That means you should ensure that your domain has a dimension such to describe properly the structure of dimension L.

Thanks for your quick reply!

My flow problem is quite esay to understand: 3D simulation of Flow around the circular cylinder at Re=3900, just want to carry out a benchmark test to validate my DNS code is feasible. I think the largest integral length scale is not the length of my domain, maybe width or height? Any suggestions?

Best,
Gang