Dear Spencer,

Thanks for the information. Please, if possible let me know which branch I should look at.

I tried with the ADRSolver, advecting a bump over a periodic condition through the geometry I included previously. It seemed to work without specifying rotation, and I got the bump going out on the top-right and returning through the bottom-right, since velocity vector does need to be reversed, and is given in the settings.

Emboldened by this I started going through pieces of code connected with periodicity in the IncNavierStokesSolver (and there is a quite complex logic that considers many possible configurations) in hope of finding a place where values are copied, or associated for the periodic BC to put a dirty hack, to achieve what (I think) I need. For now all I need is to either reverse the velocity components (Pi-rotation) or swap and reverse (Pi/2 rotation). My geometry is very simple, so I thought I could get away with this.

Might be there is a better solution (e.g. filtering some modes of the modes from the expansion) to what I want to do. Ultimately I would like to apply different types of symmetries. Mirror-symmetry I can do by setting a mix of Dirichlet and Neuman conditions. Rotational symmetries are more challenging. I thought it could be possible with periodic-rotation condition.

Cheers,
Stan

Hi Stan,

I think this type of boundary condition may only currently be working for scalar field. I did start a branch ( probably more than a year ago) which also mapped the velocity field when using an iterative solver. The issue is that we need to iteratively solver the velocity fields together so we can then apply the vector rotation rather than solving the velocity fields as a series of scalar problems which is what we do currently. 

If you wish I can try and find the branch and let you know? 

Cheers,

Spencer.

On 4 Nov 2021, at 21:50, Stanislaw Gepner <stanislaw.gepner@pw.edu.pl> wrote:

This email from stanislaw.gepner@pw.edu.pl 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.

 

Hi,

I am looking at the periodic BC with a rotation in a 2D flow. I managed to get NekMesh to work and align a PI-rotational boundary pair. I included a Rotated tag in the BOUNDARYCONDITIONS (it seems the tag is not used in 2D anyway), but what I get is not what I have expected. Consider attached figure, I have an outflow on top-right, rotated to bottom-right and I was expecting to see an inflow there.

I had a look at the DisConstField.ccp, and there is a small piece of code setting everything up (~4,5k lines of code), and from what I understand rotation is not accounted for in a 2D case.

My question is than:

1. Am I looking at the periodic-rotation BC in a wrong way?
2. Where are periodic BC enforced? I was unable to find the spot where I could just plug a *-1 to reverse the flow.

Cheers,
Stan

--
Pozdrawiam,
dr inż. Stanisław Gepner,

Wydział Mechaniczny Energetyki i Lotnictwa
Nowowiejska 21/25 Str.
00-665 Warszawa, Polska
tel. +48 (22) 234 51 70

<6BwuTwNjpq2SByoU.png>

<Screenshot from 2021-11-04 22-28-50.png>_______________________________________________
Nektar-users mailing list
Nektar-users@imperial.ac.uk
https://mailman.ic.ac.uk/mailman/listinfo/nektar-users

--
Pozdrawiam,
dr inż. Stanisław Gepner,

Wydział Mechaniczny Energetyki i Lotnictwa
Nowowiejska 21/25 Str.
00-665 Warszawa, Polska
tel. +48 (22) 234 51 70