Noises in budget terms

[Teng, Ming]

Dear Nek users,

 

I am looking at a 2D flow over a backward-facing step. Perturbations are involved and I would like to take a look at the development of budget terms.  When I plotted these terms, I notice there are noise associated with it, and its magnitude is non-negligible. I have no clue about its cause.

 

I am positive the subroutines and the macros I used for calculation are correct; I used it previously in turbulent flow and had it verified; the results were all good. This observed noise, according to the contour, occurred at the element border.  8^th polynomial order was employed and I also tried 16^th order, but increase in polynormal order was not helpful; in fact, it got worse.

 

I had a close look at the different terms in TKE budget and the noise is mostly associated with terms involving  <u’u’>. Meanwhile, <u’v’> budget terms are all good without such noise.

 

Has anyone run into this previously?

Thanks for you guys’ help.

 

Ming

 

[Tanmoy Chatterjee]

Ming,

I have observed this phenomenon, especially if you are dealing with budget terms involving gradients of u'u' and likes, but mostly in LES computations. This is in part related to C-1 continuity of the derivative terms. In DNS, this should not be observable if you are using fine enough grids. So, for my LES/coarse grids  I have observed this phenomenon for dissipation and advection terms in the budget. I cant compare the magnitude one-one but yes they are definitely not non-negligible. 

ps: finer grids should have more elements (h refinement) than the coarser grids.

Best Regards,

Tanmoy

--
You received this message because you are subscribed to the Google Groups "Nek5000" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nek5000+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nek5000/YT1PR01MB4025232A281246064C41B946C24B0%40YT1PR01MB4025.CANPRD01.PROD.OUTLOOK.COM.

[Teng, Ming]

Hello Tanmoy,

 

Thank you for the reply.

I tried refinement by increasing polynomial order and it was not helpful; it fact, the noise become worse. The location where the noise occurred in my case is mostly in the region where the mesh is finest. This is odd.

I will try refinement by increasing number of elements.

 

Again, thank you.

 

Ming

 

 

 

From: Tanmoy Chatterjee
Sent: Thursday, August 6, 2020 8:40 AM
To: Teng, Ming
Cc: nek...@googlegroups.com
Subject: Re: [nek5000] Noises in budget terms

 

CAUTION: External to USask. Verify sender and use caution with links and attachments. Forward suspicious emails to phis...@usask.ca

CAUTION: External to USask. Verify sender and use caution with links and attachments. Forward suspicious emails to phis...@usask.ca