CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   Dns (http://www.cfd-online.com/Forums/main/102917-dns.html)

vinerm June 6, 2012 06:51

Dns
 
Hello

This discussion is about DNS and is not pertaining to any particular tool.

Usually, DNS is carried out in spectral space, especially when it comes to homogeneous turbulence. However, I would like to have your opinion about the "Implications that could arise out of carrying out DNS in physical space".

Regards,
Vinerm Andrews

FMDenaro June 6, 2012 07:21

Quote:

Originally Posted by vinerm (Post 365003)
Hello

This discussion is about DNS and is not pertaining to any particular tool.

Usually, DNS is carried out in spectral space, especially when it comes to homogeneous turbulence. However, I would like to have your opinion about the "Implications that could arise out of carrying out DNS in physical space".

Regards,
Vinerm Andrews

no implication at all ... Assuming that DNS means that you must solve for all the characteristic flow scales, you can use SM, FD, FV, FEM ....

cfdnewbie June 6, 2012 14:30

Quote:

Originally Posted by vinerm (Post 365003)
Hello

This discussion is about DNS and is not pertaining to any particular tool.

Usually, DNS is carried out in spectral space, especially when it comes to homogeneous turbulence. However, I would like to have your opinion about the "Implications that could arise out of carrying out DNS in physical space".

Regards,
Vinerm Andrews


As long as you can guarantee that you are doing a real DNS with all scales resolved, there's no difference at all. Just note that with non-spectral methods you will probably need more DOF to be numerics-free than with a spectral approach!

sbaffini June 6, 2012 16:09

DNS in pipes is most commonly (assumed to be) done with second order spatial schemes.

FMDenaro June 6, 2012 16:55

just to add that DNS performed in physical space is practically the only way to face non-simple geometry

vinerm June 10, 2012 03:47

Homogeneous and Non-homogeneous
 
Hello Everyone

Thank you all for your replies. As far as I have been able to resolve by studying papers and books, spectral methods have issue in resolving non-homogeneous turbulence but whether geometry is complex or not does not really matter. However, I am still not even a beginner with DNS. I have just read a few papers about it. So, thank you all.

Regards,
Vinerm Andrews


All times are GMT -4. The time now is 07:27.