[GiDlist] Mesh to nurbs with GiD

[Arturo di Gioia]

Hi,
I would like to ask a couple of question about the possibility of
performing a surface reconstruction in GiD (version 7, from 7.2 to
latest beta) starting from a surface mesh.
I know that GiD is not a reverse engineering tool, but I cannot afford
the expense of a tool like Geomagic or Paraform.

I'm experimenting with the following procedure.

1. Create a surfmesh geometry by a combination of "Create Surfmesh"
commands in combination with "Select by tolerance" and "Tolangle = x".

2. Clean the surfmesh entities, tweak the line entities and create
4-sided nonplanar nurbs surfaces. At this point the topology in GiD is
correct (no cracks/gaps between contiguous surfaces)

3. Export the resulting model in IGES format.

The issues I am experiencing are of bad surface tolerance (when I import
the IGES in other programs, I must set the import tolerance value at a
very high level).

My question is: are there any documented/undocumented methods in GiD to
check the NURBS quality? I can try to edit the nurbs, raise their degree
and reparametrize them, but I have no control on how this affects the
geometrical gap between the nurbs. Additionally, based on your
experience, how would you rate GiD's ability to export the geometry in
IGES format?

Moreover, does the "Import tolerance" value affect the geometry
created/modified in GiD, and exported from it? If yes, how?

I repeat: I know that this is not GiD's playfield, and I am really
astonished by what GiD is able to do with the SurfMesh command
(expecially compared to more expensive alternatives).

Thanks in advance for any possible answer,

Arturo di Gioia

[Thorsten Steinmetz]

Dear GiD-Team,

I am wondering, if there is any possiblity to get a finite-element mesh of
the following geometry:
I have one cylinder with radius 3m on whose top is another with radius 2m
which has a cylindrical hole (coaxial), which has the radius 1m, see the
attached picutre. My problem is that I cannot generate a mesh at the
touching interface of the big cylinder and the cylinder with the hole that
uses the same nodes for both surface meshes. I am using GiD 7.5.8b
professional version. Thanks a lot in advance for any help.

Yours,
Thorsten Steinmetz.

-------------------------------------------------------------------
Dipl.-Ing. Thorsten Steinmetz
Helmut-Schmidt-University
University of the Federal Armed Forces Hamburg
School of Electrical Engineering
Theory of Electrical Engineering and Computational Electromagnetics
Holstenhofweg 85
D-22043 Hamburg
Tel.: +49 (0)40 / 6541-2149
Fax : +49 (0)40 / 6541-3764
eMail: thorsten.steinmetz at hsu-hamburg.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: geo.jpg
Type: image/jpeg
Size: 22810 bytes
Desc: not available
Url : http://listas.cimne.upc.edu/pipermail/gidlist/attachments/20050525/7a1f2c44/attachment.jpg

[escolano]

You must split the top (radius 3) surface, to create 3 parts: two rings and
a circle
menu: Geometry-Edit-Divide-Surfaces-Split
and select the surface, and the lines to split in two parts

The middle ring must be shared between the two volumes (must delete a
duplicated)
(for example wiht Utilities-Collapse-Model)
or delete this common surface to create a single volume.

I attach a sample

Regards

Enrique Escolano
----- Original Message -----
From: "Thorsten Steinmetz" thorsten.steinmetz at hsu-hh.de
To: gidlist at gatxan.cimne.upc.es
Sent: Wednesday, May 25, 2005 7:09 PM
Subject: {SPAM?} [GiDlist] {SPAM?} WG: Mesh generation problems


Dear GiD-Team,

I am wondering, if there is any possiblity to get a finite-element mesh of
the following geometry:
I have one cylinder with radius 3m on whose top is another with radius 2m
which has a cylindrical hole (coaxial), which has the radius 1m, see the
attached picutre. My problem is that I cannot generate a mesh at the
touching interface of the big cylinder and the cylinder with the hole that
uses the same nodes for both surface meshes. I am using GiD 7.5.8b
professional version. Thanks a lot in advance for any help.

Yours,
Thorsten Steinmetz.

-------------------------------------------------------------------
Dipl.-Ing. Thorsten Steinmetz
Helmut-Schmidt-University
University of the Federal Armed Forces Hamburg
School of Electrical Engineering
Theory of Electrical Engineering and Computational Electromagnetics
Holstenhofweg 85
D-22043 Hamburg
Tel.: +49 (0)40 / 6541-2149
Fax : +49 (0)40 / 6541-3764
eMail: thorsten.steinmetz at hsu-hamburg.de

-------------- next part --------------
A non-text attachment was scrubbed...
Name: coaxial.zip
Type: application/x-zip-compressed
Size: 1822 bytes
Desc: not available
Url : http://listas.cimne.upc.edu/pipermail/gidlist/attachments/20050525/dfd613ec/attachment.bin

[Thorsten Steinmetz]

Dear Enrique Escolano,

I'm still having problems with the geometry. Could you tell me please, how
did you model the geometry? I tried to create three cylinders:
1) base 0 0 0, radius 3, height 1
2) base 0 0 1, radius 2, height 1
3) base 0 0 1, radius 1, height 1
all orientated along positive z-direction and each in an own layer. Then I
created the cylinder with the hole by the volume boolean operation substract
(cyl2 - cyl3) and sent the inner surface of the resulting cylinder with hole
which was still in the layer of cylinder 3 to the layer of cylinder 2.
Trying what you told me in your last mail with that geometry did not work,
as the top surface of cyl1 disappeared after splitting it. However, if I
tried to generate a mesh with this geometry, the error
"19 Check self intersections, or bad oriented boundary"
was reported by GiD.

May it be that I do anything wrong during the geometry modelling?

Thank you for your help,

yours sincerely,
Thorsten Steinmetz.

-----Ursprüngliche Nachricht-----
Von: gidlist-admin at gatxan.cimne.upc.edu
[mailto:gidlist-admin at gatxan.cimne.upc.edu]Im Auftrag von Enrique
Escolano
Gesendet: Mittwoch, 25. Mai 2005 23:10
An: gidlist at gatxan.cimne.upc.edu
Betreff: Re: {SPAM?} [GiDlist] {SPAM?} WG: Mesh generation problems


You must split the top (radius 3) surface, to create 3 parts: two rings and
a circle
menu: Geometry-Edit-Divide-Surfaces-Split
and select the surface, and the lines to split in two parts

The middle ring must be shared between the two volumes (must delete a
duplicated)
(for example wiht Utilities-Collapse-Model)
or delete this common surface to create a single volume.

I attach a sample

Regards

Enrique Escolano
----- Original Message -----
From: "Thorsten Steinmetz" thorsten.steinmetz at hsu-hh.de
To: gidlist at gatxan.cimne.upc.es
Sent: Wednesday, May 25, 2005 7:09 PM
Subject: {SPAM?} [GiDlist] {SPAM?} WG: Mesh generation problems


Dear GiD-Team,

I am wondering, if there is any possiblity to get a finite-element mesh of
the following geometry:
I have one cylinder with radius 3m on whose top is another with radius 2m
which has a cylindrical hole (coaxial), which has the radius 1m, see the
attached picutre. My problem is that I cannot generate a mesh at the
touching interface of the big cylinder and the cylinder with the hole that
uses the same nodes for both surface meshes. I am using GiD 7.5.8b
professional version. Thanks a lot in advance for any help.

Yours,
Thorsten Steinmetz.

-------------------------------------------------------------------
Dipl.-Ing. Thorsten Steinmetz
Helmut-Schmidt-University
University of the Federal Armed Forces Hamburg
School of Electrical Engineering
Theory of Electrical Engineering and Computational Electromagnetics
Holstenhofweg 85
D-22043 Hamburg
Tel.: +49 (0)40 / 6541-2149
Fax : +49 (0)40 / 6541-3764
eMail: thorsten.steinmetz at hsu-hamburg.de

-------------- next part --------------
A non-text attachment was scrubbed...
Name: coaxial.zip
Type: application/x-zip-compressed
Size: 1822 bytes
Desc: not available
Url : http://listas.cimne.upc.edu/pipermail/gidlist/attachments/20050526/c9205ae9/attachment.bin

[escolano]

I repeated your creation mode and it worked ok. (with a split of the circle
in three parts, and a collapse to join the midpart)

Maybe the problem while splitting the surface is a corrected bug of your GiD
version (use help-About to see your version).
Try to download the last available beta version 7.5.8b (the password is the
same for all 7.x versions)
http://www.gidhome.com/download/do07.subst

Also are possible other options, instead to use split , for example:
a ) For planar shapes, as your case, can delete the surface and create two
new surfaces selecting his boundary (and hole) lines
b) For non planar shapes it is not sufficient the boundary to reconstruct
the same internal shape, but can create a new surface trimming this shape
Geometry-Create-Nurbs surface-Trimmed
select the surface to get his shape, and the trimmed boundary (and
hole/s) lines.
Then can delete the old surface

Note: in general is not recommended to use GiD boolean operations, it can be
slow and unstable. For example in your case it is possible to create a
planar ring, and create the cylinder ring volume by extrusion with a
translacion transform (from the copy window)

Regards
Enrique Escolano

----- Original Message -----
From: "Thorsten Steinmetz" thorsten.steinmetz at hsu-hh.de
To: gidlist at gatxan.cimne.upc.edu
Sent: Thursday, May 26, 2005 9:10 AM
Subject: WG: {SPAM?} [GiDlist] {SPAM?} WG: Mesh generation problems


Dear Enrique Escolano,

I'm still having problems with the geometry. Could you tell me please, how
did you model the geometry? I tried to create three cylinders:
1) base 0 0 0, radius 3, height 1
2) base 0 0 1, radius 2, height 1
3) base 0 0 1, radius 1, height 1
all orientated along positive z-direction and each in an own layer. Then I
created the cylinder with the hole by the volume boolean operation
substract
(cyl2 - cyl3) and sent the inner surface of the resulting cylinder with
hole
which was still in the layer of cylinder 3 to the layer of cylinder 2.
Trying what you told me in your last mail with that geometry did not work,
as the top surface of cyl1 disappeared after splitting it. However, if I
tried to generate a mesh with this geometry, the error
"19 Check self intersections, or bad oriented boundary"
was reported by GiD.

May it be that I do anything wrong during the geometry modelling?

Thank you for your help,

yours sincerely,
Thorsten Steinmetz.

-----Ursprüngliche Nachricht-----
Von: gidlist-admin at gatxan.cimne.upc.edu
[mailto:gidlist-admin at gatxan.cimne.upc.edu]Im Auftrag von Enrique
Escolano
Gesendet: Mittwoch, 25. Mai 2005 23:10
An: gidlist at gatxan.cimne.upc.edu
Betreff: Re: {SPAM?} [GiDlist] {SPAM?} WG: Mesh generation problems


You must split the top (radius 3) surface, to create 3 parts: two rings
and
a circle
menu: Geometry-Edit-Divide-Surfaces-Split
and select the surface, and the lines to split in two parts

The middle ring must be shared between the two volumes (must delete a
duplicated)
(for example wiht Utilities-Collapse-Model)
or delete this common surface to create a single volume.

I attach a sample

Regards

Enrique Escolano
----- Original Message -----
From: "Thorsten Steinmetz" thorsten.steinmetz at hsu-hh.de
To: gidlist at gatxan.cimne.upc.es
Sent: Wednesday, May 25, 2005 7:09 PM
Subject: {SPAM?} [GiDlist] {SPAM?} WG: Mesh generation problems


Dear GiD-Team,

I am wondering, if there is any possiblity to get a finite-element mesh
of
the following geometry:
I have one cylinder with radius 3m on whose top is another with radius
2m
which has a cylindrical hole (coaxial), which has the radius 1m, see the
attached picutre. My problem is that I cannot generate a mesh at the
touching interface of the big cylinder and the cylinder with the hole
that
uses the same nodes for both surface meshes. I am using GiD 7.5.8b
professional version. Thanks a lot in advance for any help.

Yours,
Thorsten Steinmetz.

-------------------------------------------------------------------
Dipl.-Ing. Thorsten Steinmetz
Helmut-Schmidt-University
University of the Federal Armed Forces Hamburg
School of Electrical Engineering
Theory of Electrical Engineering and Computational Electromagnetics
Holstenhofweg 85
D-22043 Hamburg
Tel.: +49 (0)40 / 6541-2149
Fax : +49 (0)40 / 6541-3764
eMail: thorsten.steinmetz at hsu-hamburg.de

[escolano]

Sorry, but there not exists any tool to enhance the precision of a surface
(to decrease the gap distance between the surface and the trimming lines
"over" this surface).

But maybe you can try some tools before create this surfaces, to enhance his
boundary lines.
For example:
0- Create surfmeshes and delete it to have only curves as you related
1- Convert all types lines to NUBS lines (because other tools only works for
NURBS lines)
menu: Geometry-EditConvert to NURBS-Lines
2- Force some "near tangent" curves to be tangents
command: Geometry Edit TangentNurbsL
(not exists a menu entry for this command, must use the rigth buttons or
write in the lower command line)
3- Collapse entities to remove points between tangent lines, and avoid too
small lines
menu: UtilitiesCollapse-Modes
(check the value of the ImportTolerance, in the preferences window,
points close that this tolerance will be join)
5- Simplify lines, creating another approximated interpolation (can create
better surfaces after)
menu: Geometry-EditSimplify NURBS-Lines
6 -Then create surfaces by boundary. An iteresting option is to use
"Automatic" to search surfaces with "n" sides
menu: Geometry-CreateNURBS Surface-Automatic
7 - Complete manually the model adding or deleting surfaces, and export it

Note: Import tolerance is used when importing IGES, to join separated
entities, but not when exporting IGES

If you created also volumes, can try to select "B-Rep output style" in
preferences, to write also the topological relations between geometrical
entities, but some CAD's doesn't support this boundary-representation style.

Regards

Enrique Escolano

----- Original Message -----
From: "Arturo di Gioia" arturo.digioia at ing.unitn.it
To: gidlist at gatxan.cimne.upc.es
Sent: Wednesday, May 25, 2005 9:08 AM
Subject: [GiDlist] Mesh to nurbs with GiD


Hi,
I would like to ask a couple of question about the possibility of
performing a surface reconstruction in GiD (version 7, from 7.2 to
latest beta) starting from a surface mesh.
I know that GiD is not a reverse engineering tool, but I cannot afford
the expense of a tool like Geomagic or Paraform.

I'm experimenting with the following procedure.

1. Create a surfmesh geometry by a combination of "Create Surfmesh"
commands in combination with "Select by tolerance" and "Tolangle = x".

2. Clean the surfmesh entities, tweak the line entities and create
4-sided nonplanar nurbs surfaces. At this point the topology in GiD is
correct (no cracks/gaps between contiguous surfaces)

3. Export the resulting model in IGES format.

The issues I am experiencing are of bad surface tolerance (when I import
the IGES in other programs, I must set the import tolerance value at a
very high level).

My question is: are there any documented/undocumented methods in GiD to
check the NURBS quality? I can try to edit the nurbs, raise their degree
and reparametrize them, but I have no control on how this affects the
geometrical gap between the nurbs. Additionally, based on your
experience, how would you rate GiD's ability to export the geometry in
IGES format?

Moreover, does the "Import tolerance" value affect the geometry
created/modified in GiD, and exported from it? If yes, how?

I repeat: I know that this is not GiD's playfield, and I am really
astonished by what GiD is able to do with the SurfMesh command
(expecially compared to more expensive alternatives).

Thanks in advance for any possible answer,

Arturo di Gioia

_______________________________________________
GiDlist mailing list
GiDlist at gid.cimne.upc.es
http://gid.cimne.upc.es/mailman/listinfo/gidlist