Calculation on vorticity and other derivative quantities
Hello everyone, I am using the incompressible solver in Nektar++ for simulating vortex flows around bluff bodies or slender bodies. The results obtained look fine, and the velocity and pressure fields are smooth, but as I exported the results into post-process software, like Tecplot, for calculating vorticity, the vorticity obtained is not smooth. I looked into the reason and found that the vorticity inside each element was calculated separately, so the vorticty values at the boundary between elements are discontinuous. I guess that this should be a special issue existing in spectral element grids, and no this kind of problem exists in finite volume grids. This is also applicable to calculating other derivative quantities, like in Q or lambda2 methods for vortex identification. I'd like to know if there is some approaches to deal with this problem. Otherwise, it is difficult to visualize vortex structures. Many thanks. Baofeng Baofeng Ma, PH.D Associate Professor Institute of Fluid Mechanics School of Aeronautical Science and Engineering Beihang University Email: bf-ma@buaa.edu.cn Tel: 0086 (0)10 82338344
Hi, FieldConvert utility has methods to calculate q cryterion, velocity gradients etc. Run FieldConvert -l for a list of modules. Best, Pozdrawiam, Stanisław Gepner -- Sent from my Android device with K-9 Mail. Please excuse my brevity. Dnia 30 czerwca 2017 15:45:34 CEST, BF MA <bf-ma@buaa.edu.cn> napisał(a):
Hello everyone,
I am using the incompressible solver in Nektar++ for simulating vortex flows around bluff bodies or slender bodies. The results obtained look fine, and the velocity and pressure fields are smooth, but as I exported the results into post-process software, like Tecplot, for calculating vorticity, the vorticity obtained is not smooth. I looked into the reason and found that the vorticity inside each element was calculated separately, so the vorticty values at the boundary between elements are discontinuous. I guess that this should be a special issue existing in spectral element grids, and no this kind of problem exists in finite volume grids. This is also applicable to calculating other derivative quantities, like in Q or lambda2 methods for vortex identification. I'd like to know if there is some approaches to deal with this problem. Otherwise, it is difficult to visualize vortex structures.
Many thanks.
Baofeng
Baofeng Ma, PH.D Associate Professor Institute of Fluid Mechanics School of Aeronautical Science and Engineering Beihang University Email: bf-ma@buaa.edu.cn Tel: 0086 (0)10 82338344
------------------------------------------------------------------------
_______________________________________________ Nektar-users mailing list Nektar-users@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/nektar-users
Hi Baofeng, Just to follow up on Stanislaw's answer, the reason your results from Tecplot probably don't look great is that the .dat or .plt produced by FieldConvert will be a linearised version of the high-order element, and derivative quantities will be using this linearisation, which inherently loses a lot of the high-order information. As also mentioned, you should use FieldConvert to calculate these derivatives, which should yield much more accurate results. We have support for Q criterion and voriticty, but not lambda2 if I recall correctly. Thanks, Dave
On 1 Jul 2017, at 19:44, Stanisław Gepner <sgepner@meil.pw.edu.pl> wrote:
Hi,
FieldConvert utility has methods to calculate q cryterion, velocity gradients etc. Run FieldConvert -l for a list of modules.
Best, Pozdrawiam, Stanisław Gepner
-- Sent from my Android device with K-9 Mail. Please excuse my brevity.
Dnia 30 czerwca 2017 15:45:34 CEST, BF MA <bf-ma@buaa.edu.cn> napisał(a): Hello everyone,
I am using the incompressible solver in Nektar++ for simulating vortex flows around bluff bodies or slender bodies. The results obtained look fine, and the velocity and pressure fields are smooth, but as I exported the results into post-process software, like Tecplot, for calculating vorticity, the vorticity obtained is not smooth. I looked into the reason and found that the vorticity inside each element was calculated separately, so the vorticty values at the boundary between elements are discontinuous. I guess that this should be a special issue existing in spectral element grids, and no this kind of problem exists in finite volume grids. This is also applicable to calculating other derivative quantities, like in Q or lambda2 methods for vortex identification. I’d like to know if there is some approaches to deal with this problem. Otherwise, it is difficult to visualize vortex structures.
Many thanks.
Baofeng
Baofeng Ma, PH.D Associate Professor Institute of Fluid Mechanics School of Aeronautical Science and Engineering Beihang University Email: bf-ma@buaa.edu.cn Tel: 0086 (0)10 82338344
Nektar-users mailing list Nektar-users@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/nektar-users _______________________________________________ Nektar-users mailing list Nektar-users@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/nektar-users
HI Baofeng, You should be aware there are also C0 projections that can be performed on discontinuous field to make the continuous. You also mention this is a special issue for spectral elements (SEM) but not finite volume methods (FVM). This is not true since FVM are typically also only locally continuous and so a linear FVM would have a piecewise continuous derivative. However when you typically post process a FVM you also typically smooth the result and so if you want the same result with SEM methods you also have to smooth the result. Cheers, Spencer. On 3 Jul 2017, at 22:46, David Moxey <d.moxey@imperial.ac.uk<mailto:d.moxey@imperial.ac.uk>> wrote: Hi Baofeng, Just to follow up on Stanislaw's answer, the reason your results from Tecplot probably don't look great is that the .dat or .plt produced by FieldConvert will be a linearised version of the high-order element, and derivative quantities will be using this linearisation, which inherently loses a lot of the high-order information. As also mentioned, you should use FieldConvert to calculate these derivatives, which should yield much more accurate results. We have support for Q criterion and voriticty, but not lambda2 if I recall correctly. Thanks, Dave On 1 Jul 2017, at 19:44, Stanisław Gepner <sgepner@meil.pw.edu.pl<mailto:sgepner@meil.pw.edu.pl>> wrote: Hi, FieldConvert utility has methods to calculate q cryterion, velocity gradients etc. Run FieldConvert -l for a list of modules. Best, Pozdrawiam, Stanisław Gepner -- Sent from my Android device with K-9 Mail. Please excuse my brevity. Dnia 30 czerwca 2017 15:45:34 CEST, BF MA <bf-ma@buaa.edu.cn<mailto:bf-ma@buaa.edu.cn>> napisał(a): Hello everyone, I am using the incompressible solver in Nektar++ for simulating vortex flows around bluff bodies or slender bodies. The results obtained look fine, and the velocity and pressure fields are smooth, but as I exported the results into post-process software, like Tecplot, for calculating vorticity, the vorticity obtained is not smooth. I looked into the reason and found that the vorticity inside each element was calculated separately, so the vorticty values at the boundary between elements are discontinuous. I guess that this should be a special issue existing in spectral element grids, and no this kind of problem exists in finite volume grids. This is also applicable to calculating other derivative quantities, like in Q or lambda2 methods for vortex identification. I’d like to know if there is some approaches to deal with this problem. Otherwise, it is difficult to visualize vortex structures. Many thanks. Baofeng Baofeng Ma, PH.D Associate Professor Institute of Fluid Mechanics School of Aeronautical Science and Engineering Beihang University Email: bf-ma@buaa.edu.cn<mailto:bf-ma@buaa.edu.cn> Tel: 0086 (0)10 82338344 Nektar-users mailing list Nektar-users@imperial.ac.uk<mailto:Nektar-users@imperial.ac.uk> https://mailman.ic.ac.uk/mailman/listinfo/nektar-users _______________________________________________ Nektar-users mailing list Nektar-users@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/nektar-users _______________________________________________ Nektar-users mailing list Nektar-users@imperial.ac.uk<mailto:Nektar-users@imperial.ac.uk> https://mailman.ic.ac.uk/mailman/listinfo/nektar-users Spencer Sherwin McLaren Racing/Royal Academy of Engineering Research Chair, Professor of Computational Fluid Mechanics, Department of Aeronautics, Imperial College London South Kensington Campus London SW7 2AZ s.sherwin@imperial.ac.uk<mailto:s.sherwin@imperial.ac.uk> +44 (0) 20 759 45052
participants (4)
-
BF MA
-
David Moxey
-
Sherwin, Spencer J
-
Stanisław Gepner