Dear all,
We should not be using the RadauMAlpha1Beta0 with Hex elements. This stands for Radau distribution at minus 1 (i.e. not including the point at +1) but it also involves the weight 0.25 (1-x)^alpha (1+x)^beta. So the Alpha1 means you are adding a weight function of 0.5(1-x)^alpha into the integral which is naturally part of the collapsed coordinate transform but this is not required in a Hex.
So in theory one could use RadauMAlpha0Beta0 (not sure if this actually exists) but would be best just to use GaussLobattoLegendre in all directions.
Cheers,
Spencer.
From: nektar-users-bounces@imperial.ac.uk <nektar-users-bounces@imperial.ac.uk> on behalf of Boyang Xia <boyang.xia@kcl.ac.uk>
Date: Monday, 15 May 2023 at 16:19
To: Ehsan Asgari <eh.asgari@gmail.com>
Cc: nektar-users <nektar-users@imperial.ac.uk>
Subject: Re: [Nektar-users] potential bug with detailed expansionHi,
GaussRadauMAlpha1Beta0 seems to be often used in Tri and Tet with Modified_B. I am not sure if it works with Modified_A(Ideally it should). So, what is your consideration of using GaussRadauMAlpha1Beta0 on the 3rd direction of hex?
Are you using the incompressible Navier-Stokes solver and VCS? Then maybe you should follow chapter 11.1.1.6 to set different number of modes and points for velocity and pressure field.
Hope it helps, or correct me if it's wrong.
Thanks,
Boyang
From: nektar-users-bounces@imperial.ac.uk <nektar-users-bounces@imperial.ac.uk> on behalf of Ehsan Asgari <eh.asgari@gmail.com>
Sent: Monday, May 15, 2023 11:53
To: Jacques Xing <jacques.xing@kcl.ac.uk>
Cc: nektar-users <nektar-users@imperial.ac.uk>
Subject: Re: [Nektar-users] potential bug with detailed expansion
Hi Jacques
I have a fully hex mesh. My perception is that for hex elements, modified_a should be used.
Please correct me if I'm wrong.
syavash
On Mon, May 15, 2023 at 3:11 PM Jacques Xing <jacques.xing@kcl.ac.uk> wrote:
Hi Ehsan
I believe that you should use Modified_B basis from GaussRadauMAlpha1Beta0 quadrature. I am also assuming that you are using prism elements for the mesh?
Best,
Jacques
From: nektar-users-bounces@imperial.ac.uk <nektar-users-bounces@imperial.ac.uk> on behalf of Ehsan Asgari <eh.asgari@gmail.com>
Sent: Monday, May 15, 2023 9:29 AM
To: nektar-users <nektar-users@imperial.ac.uk>
Subject: [Nektar-users] potential bug with detailed expansion
This email from eh.asgari@gmail.com originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list to disable email stamping for this address.
Dear All,
When I use a "Modfied" type for expansions everything works fine. But if I specify the expansions in detailed mode:
<E COMPOSITE="C[30011]"
BASISTYPE="Modified_A,Modified_A,Modified_A"
NUMMODES="7,7,7" POINTSTYPE="GaussLobattoLegendre,GaussLobattoLegendre,GaussRadauMAlpha1Beta0"
NUMPOINTS="8,8,7"
FIELDS="u,v,w" />
<E COMPOSITE="C[30011]"
BASISTYPE="Modified_A,Modified_A,Modified_A"
NUMMODES="6,6,6"
POINTSTYPE="GaussLobattoLegendre,GaussLobattoLegendre,GaussRadauMAlpha1Beta0"
NUMPOINTS="7,7,6"
FIELDS="p" />
Then I get a bunch of errors:
[xxx518:29773] *** Process received signal ***
[xxx518:29773] Signal: Segmentation fault (11)
[xxx518:29773] Signal code: Address not mapped (1)
[xxx518:29773] Failing at address: 0x27b2000
I have a fully hex mesh.
Could you kindly point out if something is wrong with the way I specify the expansions?
Kind regards
syavash