Dear Enrique Escolano,
Thanks for you help.
I used GID 9.0.x to generate the mesh and it is OK.
I’m wonder that why the two additional boxes are so important to generate the mesh.
And if we meet the problem “38 cound’t mesh at this location”, how can we deal with it?
And the other problem is after i hit the generate, there is a parameter I can modify, this parameter is
also important to the problem. How can I choose this parameter?
I will do the structured mesh on the other steel. Thanks again for your nice help!
Best Regards，
CHEN Ningning
2009-03-03
Dear Chen,
I’ve modified your file 'Team13_Last’creating two additional boxes that
connect the two parts filling the ‘thin air space’
This auxiliary boxes are in the layer aux1 and aux2 and are interesting to
aid to mesh. I set as structured only the thinner one, with number of
divisions 3x7x11. I was only interested to have 3 divisions in the small
gap, the other numbers are to have reasonable tetrahedra sizes
Note: I worked with only a quarter of the piece to do less operations, and
at the end I created the rest by copy with 180 degrees of rotation and with
copy with simmetry.
I attach my model zipped including the mesh (about 100.000 elements and
17.000 nodes).
It was meshed with GiD 9.x. If you are unable to mesh it with GiD 8.x can
download this new version and use it with a trial one month password
obtained from our web. There are also special prices to upgrade from GiD 8.x
to 9.x (60 discount), you can see all this information in the web
www.gidhome.com
Regards
Enrique
----- Original Message -----
From: “CHEN Ningning” xjtucnn at 163.com
To: gidlist at gid.cimne.upc.edu
Cc: 08901607r at polyu.edu.hk
Sent: Thursday, February 26, 2009 12:00 AM
Subject: Re: Re: [GiDlist] Hi,there is a question about the unstructured
mesh on thin plate
Dear Enrique,
Thanks very much for your nice help!
I really want to know in the sample, when the air is meshed with
unstructured elements, and the length and width are meshed with 40 interval
while the thickness with 4 interval. But I can’t generate it.
If the lenth and width are meshed with 200 interval, the problem will be
very large for the Finite element analysis.
My problem is more complicated than the sample I attached.
Here the attachment is my problem. And in this attachment
model(team13all.geo),I enlarged the air gap from 0.0005 to 0.0032 compare to
the real model(which is also in the attachment,PDF file,and team13last.geo).
if I didn’t enlarge it ,I can’t generate the mesh. (Or I can generate the
mesh all with unstructured elements, but air gap is only with one element.I
want split the air gap for 3 or 4 interval.)
My ideal was that after I generate the mesh, I would modify the coordinate
of the node in the air gap to the real model.
So after this idear, I enlarge the air gap and the thick of steel,too. (
team_13_OK).But the problem is after I
modify the coordinate, there are some elements(6,7) are bad( their volume is
little than 0).
So the core problem is that I don’t want to modify the real model, and I
want to spit the air gap and the thickness of steel to 3or4 intervals.
Because the air gap is very important for the FEM result. And I used this
mesh, my result is 1/3 of the experiment result.
I work on the model for 2 months, and there is rare progress. So I’m so
appreciate for your help and suggestion.
By the way, our group bought 6 license of GID8.0.2. how can we update to
GID9? If GID9 will be help in my problem?
Best regards,
CHEN Ningning
I have been able to generate a mesh without any problem using GiD 9.1.0b
Rfast, Automatic correct sizes=Normal, transition=0.7
internal volume structured 200x200x4 cells (your applied number of
divisions)
General mesh size=0.1
I’ve obtained about 1.6 Milion of elements
The geometry is very simple, if you are unable to mesh unstructured with
the
number of divisions that you want, then you can split your volumes into 27
structured volumes (it’s easy to rebuid it using the copy window)
I attach this kind of splitted model, you can generate the mesh with the
current number of assigned divissions to obtain 9.36 Milion of hexahedra
(or
assign less division to avoid so big mesh)
Enrique
----- Original Message -----
From: “CHEN Ningning” xjtucnn at 163.com
To: gidlist at gid.cimne.upc.edu
Sent: Thursday, February 19, 2009 6:18 AM
Subject: [GiDlist] Hi,there is a question about the unstructured mesh on
thin plate
Dear gidlist，！
There is a question about the unstructured mesh.
When I model a thin plate surrounded by air in 3D(the long and width
are
100 times than the thickness), I want to using unstructured tetrahedral
elements to mesh the air.
The problem is I want to mesh 4,5 intervals on the thickness direction to
assure the accuracy of the solution, so I need to using structured mesh on
the thin plate.
But when I generate the unstructured mesh on the air, I failed.
I must guarantee the size of edge on the triangle element on the
intersaction of air and plate is almost the same. Just like if I mesh 4,5
intervals on the thickness direction, I must mesh 400-500 intervals on the
direction of the longth or width.
I want to only mesh 40,50 intervals on the direction of the longth or
width, but I can’t generate that mesh.
Are there any parameters which allow me to generate the mesh?
I have set the unstructure size transitions to any value I can, but it
failed also.
My major is computational electromagnetics and there are many problems
about
very thick air gap in between the iron.
So it is the problem disturbed me most.
The attachment is my model. Thanks,
Best Regards，
_ CHEN Ningning_
_ 2009-02-19_
Best Regards，
_ CHEN Ningning_
_ 2009-02-20_