[Nek5000-users] In consistent periodic BC after prex (CAUSE TO THE PROBLEM FOUND)

[nek5000-users at lists.mcs.anl.gov]

Thanks Paul. 

Yes you were right in your first thought. 

I was just debugging the problem. So I just tried some meshes. In reality I would be using bigger mesh. And then I will refine with GLL points. 

Thanks for your replay and it will help me to move on the right way. 

For the second thing, I use mshconvert.py to convert meshes from ICEM and this script only works for HEXA8 elements. So I prefer to use prenek 
for periodic bc. 

And for my final objective, my domain can not be extruded in that way. 

Regards 

Sijo GEORGE 

De: "nek5000-users" <nek5000-users at lists.mcs.anl.gov> 
À: "nek5000-users" <nek5000-users at lists.mcs.anl.gov> 
Envoyé: Mardi 19 Juin 2018 21:21:59 
Objet: Re: [Nek5000-users] In consistent periodic BC after prex (CAUSE TO THE PROBLEM FOUND) 






Dear Sijo, 




Prenek is a 32-bit code, so it would be difficult to resolve 

your problem if you have element sizes at positions that have order unity magnitude and 1.e-6 size. 




I have two thoughts about this. 




First, I wonder why you would need elements of that size (when there will 

be yet another order of magnitude reduction in spacing coming from the GLL 

points interior to these very small elements). 




Second, assuming you can represent the 2D mesh, I would use n2to3. 

It will generate the correct periodicity and you can output the mesh file 

in the binary .re2 format, which I believe actually stores all coordinates 

in 64-bit floating point. 




Best, 




Paul 



From: Nek5000-users <nek5000-users-bounces at lists.mcs.anl.gov> on behalf of nek5000-users at lists.mcs.anl.gov <nek5000-users at lists.mcs.anl.gov> 
Sent: Tuesday, June 19, 2018 12:50:48 PM 
To: nek5000-users 
Subject: Re: [Nek5000-users] In consistent periodic BC after prex (CAUSE TO THE PROBLEM FOUND) 
Hello Neks, 

I think I found the exact problem. 

I made two meshes for NACA0012 from same block in ICEM. 

1. Mesh with very fine refinement close to the wall. (The distance between the wall to first element in the order of -1 x 1e6 or lets say the distance between the element faces is too small) 

2. Mesh with uniform distribution.(Mesh with uniform distance between faces and in the order of -1 x 1e2) 

The mesh 1 throws the errror (when I give Auto periodic option in Prenek, the number of NFAIL is greater than zero and it never comes to zero) 

and Mesh 2 works fine. 

So I think in prenek they have set some limits to take care of this because I have seen some variables called tol. (tolerences). 

Is there anything that I can change in Prenek to take care of this problem? 

br, 

Sijo GEORGE 


De: "Sijo George" <sijo.george at eleves.ec-nantes.fr> 
À: "nek5000-users" <nek5000-users at lists.mcs.anl.gov> 
Envoyé: Lundi 18 Juin 2018 18:25:43 
Objet: In consistent periodic BC after prex 

Hi Neks, 

I did a mesh conversion from ICEM to NEK with ' ' (3 spaces) as boundary conditions instead of periodic. (It is a 3D naca profile). I tried n2to3 method for extrusion but I am not confident with that method since I already have a 3D mesh and my future works will be like that. 

So my problem is, I am trying to set periodic boundary conditions on both sides of the naca wing. I use prex for that. 

When I use prex for simple gemetries like cube and cylinder, the periodic conditions are set well. 

But when I try on NACA the periodic BCs become in consistent and it runs a loop to check the consistency and not generating rea file anymore. 

I commented some lines in source code (tools/prenek/bound.f) and tried to figure whats going on, then I generated an rea file and checked inside. 

I saw that some elements are not set with Periodic bcs. 

I understand that the loop helps to make the BC consitent. 

But could you please tell me why it happens to NACA and not for simple gemoteries. 

I am really stuck here.. 

Thank you 

Sijo GEORGE 



_______________________________________________ 
Nek5000-users mailing list 
Nek5000-users at lists.mcs.anl.gov 
https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/nek5000-users/attachments/20180619/e1e95fe1/attachment.html>