Introduction to Computational Fluid Dynamics - Turbulence - 6 - DNS and LES

[@pbandj7764]

I have come back to turbulence after 50 years and see some things have changed but the problems remain the same! Surprised that I understood your lecture and presented so much better than other introductory talks because it has the right amount of the theory and practicality. Challenges oh! Onward and upward

[@pbandj7764]

Nice well organized and clear talk that I am able to ask a question!!!!

[@sagga18]

Nice and useful explanation ,,,,thanks so much

[@pbandj7764]

So my question…. If you are using smagorinky/kolmogrov model, K -5/3 is based on equilibrium cascade of energy, does this apply in inhomogeneous time varying (which implies non-equilibrium or dynamic unbalanced cascade, am I correct?) flow/turbulence? Is there an explicitly or implicit assumption here that the dissipation scales are steady state, homogeneous or in equilibrium? Is this studied via lab, wind tunnel or real experiments? If I recall there is intermittency in the dissipation…etc etc. Thanks

[@SAEMiller]

These are simple questions, and really good questions, and probably a comment on KZbin cannot answer them accurately. Intermittancy in theory has been included approx. 20 years after K41 theory. That's not an issue. All turbulence is unsteady. Inhomogeneous flow-fields do have cascades, and often look like the spectra of homogeneous isotropic turbulence. Turbulence has a universality and a 'memory.' Studies have and are being conducted in every type of facility you can imagine - and likely ones not easily thought of today.

[@pbandj7764]

@@SAEMiller Thank you for your reply’s. They point me in a direction to explore/dig.

[@comrade786z]

Really helpful thanks

[@SAEMiller]

You're welcome. This class is of course no substitute into a full study of DNS / LES. Good for just a quick overview / knowledge of the subject.

[@ishwarchandragulgule5900]

Hello Miller, a very nice explanation and I worked with DNS in MIPT University, Russia. I am trying to solve it by commercial software. Can you suggest some DNS coding books for Ansys software?

[@SAEMiller]

What a wonderful university. I'm sure that you are enjoying it very much. If you want to do DNS I suggest using an open source code online or creating your own. Usually DNS simulations are very computationally expensive and the commercial packages aren't really designed for them. I suggest you read the overview paper DIRECT NUMERICAL SIMULATION: A Tool in Turbulence Research, Parviz Moin and and Krishnan Mahesh, Annual Review of Fluid Mechanics, Vol. 30, 1998

[@madhurjyamahanta6145]

Sir, how can I learn dns simulation? From where to start can you guide?

[@Jatinderchahal19]

Hello sir , I am studying Meteorology in Leibniz university Germany. i want to learn more about LES Modelling and also have subject. from where I can learn and understanding better.

[@SAEMiller]

There are some great articles on LES from Fluent and also on Annual Review of Fluid Mechanics. You should be able to find those articles via google or from the journal.

[@pbandj7764]

To further…. Is LES incomplete simply because of the transition to dissipation (i.e. heat or nothingness) is not modeled. Numerical dissipation seems lazy. In my application, atmosphere wind gusts are intermittent and we do not know how much. :-)

[@SAEMiller]

An incomplete turbulence model, in my view, is one that needs more information than just the equations and physics to make a prediction. For example, some RANS based models would need distance to the wall. They are incomplete. Implicit LES, in my view, is incomplete because it relies on min. grid point spacing for a filter. LES with a sub grid scale model would be complete, because they model (though with some empirical functions / constants), is considered complete because it would not rely on additional information from the user such as wall distance etc. Filters are dependent on grid point spacing either way. Some people for this reason would consider all LES to be incomplete. I can see both viewpoints.

[@yassinekaroum6449]

Hello Mr S.Miller i want to thank you for this interesting video, i want to learn more about DNS and i found some problem to find resource and reference if you can to help me about that

[@SAEMiller]

I would recommend examining the following article DIRECT NUMERICAL SIMULATION: A Tool in Turbulence Research Annual Review of Fluid Mechanics Vol. 30:539-578 (Volume publication date January 1998) doi.org/10.1146/annurev.fluid.30.1.539