Hi Paul, Thank you very much for your reply. I'm doing DNS of turbulent pipe flow and the plots were from a pipe flow with Re_D=5300 (Re_tau=180). Basically, I used the same grid resolution as specified in the paper "Direct numerical simulation of turbulent pipe flow at moderately high Reynolds numbers" which I think should be reasonably good for the given low Reynolds number. And I'm also using PnPn-2, the same as used in the papers which use nek5000. In this case, shall I use even better grid resolution for this Reynolds number in order to get rid of the wiggles? For Lambda2, I'll have a look at the filtering subroutines you suggested. Best regards, Tony ________________________________________ From: [email protected] <[email protected]> on behalf of [email protected] <[email protected]> Sent: 10 December 2014 18:00 To: [email protected] Subject: Nek5000-users Digest, Vol 70, Issue 6 Send Nek5000-users mailing list submissions to [email protected] To subscribe or unsubscribe via the World Wide Web, visit https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users or, via email, send a message with subject or body 'help' to [email protected] You can reach the person managing the list at [email protected] When replying, please edit your Subject line so it is more specific than "Re: Contents of Nek5000-users digest..." Today's Topics: 1. Re: Unrealistic values at element edges ([email protected]) ---------------------------------------------------------------------- Message: 1 Date: Wed, 10 Dec 2014 17:34:21 +0000 From: [email protected] To: "[email protected]" <[email protected]> Subject: Re: [Nek5000-users] Unrealistic values at element edges Message-ID: <[email protected]> Content-Type: text/plain; charset="us-ascii" Tony, The wiggles in the quantities your plotting are generally indicative of marginal resolution for the given Reynolds number. They will go away with increased resolution. Lambda2 is not continuous a priori, but converges to a continuous function with increased resolution (exponentially fast). Are you using PnPn or PnPn-2 ? Pressure will generally look better with PnPn (and be better, in the marginally resolved, i.e., LES, case). For lambda2 it might be possible to do some filtering to get a smoother plot...(same for other quantities). I would suggest going that route --- there are several filtering strategies but I'm not certain which would be best for your particular case. You might start with the g_filter() routine in navier5.f with a diagonal scaling of the wavenumbers like: ( 1 1 1 1 1 ... 1 .99 .95 .90 ) say... which reduces the energy in the top 3 modes. There is also filter_d2() that you can experiment with. Paul ________________________________________ From: [email protected] [[email protected]] on behalf of [email protected] [[email protected]] Sent: Wednesday, December 10, 2014 10:28 AM To: [email protected] Subject: Re: [Nek5000-users] Unrealistic values at element edges Dear Neks, Sorry that the image wasn't shown. I've attached it here. Hope someone can help. Thanks and regards, Tony ________________________________________ From: [email protected] <[email protected]> on behalf of [email protected] <[email protected]> Sent: 10 December 2014 16:23 To: [email protected] Subject: Nek5000-users Digest, Vol 70, Issue 4 Send Nek5000-users mailing list submissions to [email protected] To subscribe or unsubscribe via the World Wide Web, visit https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users or, via email, send a message with subject or body 'help' to [email protected] You can reach the person managing the list at [email protected] When replying, please edit your Subject line so it is more specific than "Re: Contents of Nek5000-users digest..." Today's Topics: 1. Unrealistic values at element edges ([email protected]) ---------------------------------------------------------------------- Message: 1 Date: Wed, 10 Dec 2014 16:23:02 +0000 From: [email protected] To: "[email protected]" <[email protected]> Subject: [Nek5000-users] Unrealistic values at element edges Message-ID: <[email protected]> Content-Type: text/plain; charset="iso-8859-1" Dear Neks, I've been running turbulent channel and pipe flows using nek5000 for a while. However, recently when I created movies of lambda2 structures and pressure iso-surfaces in channel and pipe flows I observed some "strange" values at the edge between elements. [cid:cec5092f-41be-4ae9-a468-a5fa9ab9f7ff] This is lambda2 structures in turbulent pipe flow. It is clearly visible from the picture that at some slices in streamwise direction, the values don't seem to be realistic and it feels like the flow is passing through some filters. Interestingly, when I plotted the mesh together with the variables, the vertical noise lines happened to be at the edge between adjacent elements, which is particularly clear in lambda2 plot. Same behaviour was also found in 2D pressure plots and in turbulent channel flow cases. Has anyone observed similar things before in visualisation? I used both Paraview and VisIt for visualisation and the same thing was observed, so it shouldn't be a problem of software. In this case, I'm just wondering why the result is like this or is it because I set something wrong in my simulations? Any comments will be appreciated and thank you very much in advance. Best regards, Tony