|
[Sponsors] |
November 17, 2005, 11:50 |
Symmetry in Gridgen
|
#1 |
Guest
Posts: n/a
|
After creating a block in Gridgen I want to mirror this block. When I use the "mirror blocks for export" button, the resulting file cannot be read by Fluent. If I try to mirror the block manually by copy and mirror, the resulting grid can be read by fluent, but the gridcheck fails. The mirror plane that I use is the y=0 plane. I think the problem is that the face that is supposed to be in the y=0 plane is actually not located there, but it has a very small offset (about -5*10e-10). I tried to get rid of this offset by projecting the domain onto database entities, but that did not work. How can I solve this problem?
|
|
November 17, 2005, 11:53 |
Re: Symmetry in Gridgen
|
#2 |
Guest
Posts: n/a
|
Did you check to make sure that the mirrored block is not left-handed?
|
|
November 18, 2005, 04:08 |
Re: Symmetry in Gridgen
|
#3 |
Guest
Posts: n/a
|
I have experienced this problem in the past and still haven't found a satisfactory solution. I adopt one of three strategies:
(i) Build the entire grid by hand without using the mirror or rotate block functionality (preferred, but can be tedious) (ii) Use the merge entity function (its in the connectors menu). This will ensure the two grid planes of adjacent blocks match, but it won't neccesarily be the plane of symmetry. (iii) Where the plane of symmetry lies on x=0, y=0 or z=0 you could scale the domain using scaling factors of 0 1 1 (for an x=0 plane). |
|
November 25, 2005, 06:16 |
Re: Symmetry in Gridgen
|
#4 |
Guest
Posts: n/a
|
I mirrored all domains manually and created new blocks, that worked well
|
|
November 25, 2005, 12:51 |
Re: Symmetry in Gridgen
|
#5 |
Guest
Posts: n/a
|
I'm going to assume that you're using a database (geometry model) on which the grid is being built. Often, the symmetry plane as defined by the database is not accurate to within CFD tolerance - it wiggles and wanders around the z=0 (for example) plane.
There is a Glyph script on the Glyph Script Exchange web site (www.pointwise.com/glyph) that's specifically designed to fix this problem. I think the script is called SymmetryPlane.glf. Once you use it, the mirrored blocks should be fully connected on export and that should take care of the Fluent import issue. As for left-handedness, that should be automatically handled by mirroring. If your mesh has singularities (e.g. axis of rotation) this can result in hex cells that get collapsed to prisms and sometimes, depending on tolerance issues, Fluent doesn't like those. |
|
November 26, 2005, 06:18 |
Re: Symmetry in Gridgen
|
#6 |
Guest
Posts: n/a
|
I'm using a database for my grid, so I will try the glyph. What does it mean if a block is left or right handed and how I can I check this?
|
|
November 26, 2005, 15:01 |
Re: Symmetry in Gridgen
|
#7 |
Guest
Posts: n/a
|
Left and right handedness refers to the computational coordinate system (i,j,k) of your structured blocks. (If you're not generating structured grids it doesn't matter.) A right handed (i,j,k) or (xi,eta,zeta) coordinate system obeys the right hand rule - crossing the i direction with the j-direction points in the k-direction. If your CFD solver requires a right-handed coordinate system and you give it one that's left handed, it messes up the coordinate transformations and you get negative Jacobians (makes the solver thinks the cell volumes are negative). Believe it or not, I know of at least one CFD solver that requires a left-handed coordinate system.
|
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
How to import NACA 0012 Profile in Gridgen | maplepink | Main CFD Forum | 17 | May 7, 2013 14:01 |
[ICEM] Blocking and Symmetry | BrolY | ANSYS Meshing & Geometry | 32 | August 24, 2012 04:13 |
Gridgen Symmetry layer Height | harishameed33 | Pointwise & Gridgen | 3 | July 23, 2010 10:33 |
Properly using symmetry with both CFX and ANSYS workbench for a FSI analysis, help! | Cirion0000 | CFX | 0 | July 6, 2009 15:26 |
Forces in Symmetry | Matt | CFX | 2 | February 25, 2009 06:41 |