ODB file too big

Discussion:

ODB file too big

Kamalesh

2004-04-05 13:26:19 UTC

Dear all,

There are few ODBs which are very big in size because of unnecessary time
point results stored. Is there any way to delete those unimportant time
point results and store the new ODB which occupies less space?

Lately, I learned how to restore only final time point results. Rerun with
this option of all those files will take weeks. Kindly let me know if there
is any way.

Thanks
Kamalesh Vasudeva
Structural Engr.
Infotech Enterprises Limited, Infocity,
Hyderabad,
India

Shiladitya Basu

2004-04-05 14:00:07 UTC

Permalink

Post by Kamalesh
point results stored. Is there any way to delete those unimportant time
point results and store the new ODB which occupies less space?

...as far as I know, you can't delete info from an existing ODB and store
the smaller version. May Mr. Keene could shed some light on this.

On the other hand, for future analysis you could use

*NODE OUTPUT and *ELEMENT OUTPUT commands to store information only for
those output sets which you will need for further analysis.

Post by Kamalesh
Lately, I learned how to restore only final time point results. Rerun with
this option of all those files will take weeks. Kindly let me know if there
is any way.

Damian Kavanagh

2004-04-05 17:17:54 UTC

Permalink

hey all

I'm want to model an elastic-perfectly-plastic material. Am I correct in thinking I only need to input

1) the youngs modulus and poissons in the 'elastic' option and

2) the yield point i.e. 300MPa with a corresponding strain of zero in the 'plastic' option???

Thanks

Damian Kavanagh

Dept. of Mechanical & Manufacturing Engineering,
Parsons Building,
Trinity College Dublin,
Dublin 2,
Ireland.

---------------------------------
Do you Yahoo!?
Yahoo! Small Business $15K Web Design Giveaway - Enter today

[Non-text portions of this message have been removed]

------------------------ Yahoo! Groups Sponsor ---------------------~-->
Buy Ink Cartridges or Refill Kits for your HP, Epson, Canon or Lexmark
Printer at MyInks.com. Free s/h on orders $50 or more to the US & Canada.
http://www.c1tracking.com/l.asp?cid=5511
http://us.click.yahoo.com/mOAaAA/3exGAA/qnsNAA/PMYolB/TM
---------------------------------------------------------------------~->

Rod Pinna

2004-04-06 02:40:42 UTC

Permalink

Yes, that's all you need. It often helps numerical convergence to give a
small slope to the yield part of the curve. Between E/1000 to E/10000 is
often used.

Rod Pinna
Research Associate
School of Oil & Gas Engineering
The University of Western Australia
35 Stirling Highway
Crawley, 6009, WA

-----Original Message-----
Sent: Tuesday, 6 April 2004 1:18 AM
Subject: [ABAQUS] elastic-perfectly-plastic material
hey all
I'm want to model an elastic-perfectly-plastic material. Am I
correct in thinking I only need to input
1) the youngs modulus and poissons in the 'elastic' option and
2) the yield point i.e. 300MPa with a corresponding strain of
zero in the 'plastic' option???
Thanks
Damian Kavanagh
Dept. of Mechanical & Manufacturing Engineering,
Parsons Building,
Trinity College Dublin,
Dublin 2,
Ireland.
---------------------------------
Do you Yahoo!?
Yahoo! Small Business $15K Web Design Giveaway - Enter today
[Non-text portions of this message have been removed]
------------------------ Yahoo! Groups Sponsor
---------------------~--> Buy Ink Cartridges or Refill Kits
for your HP, Epson, Canon or Lexmark Printer at MyInks.com.
Free s/h on orders $50 or more to the US & Canada.
http://www.c1tracking.com/l.asp?cid=5511
http://us.click.yahoo.com/mOAaAA/3exGAA/qnsNAA/PMYolB/TM
--------------------------------------------------------------
-------~->
http://groups.yahoo.com/group/abaqus
Yahoo! Groups Links

Damian Kavanagh

2004-04-06 11:12:17 UTC

Permalink

hey rod
thanks for the reply
when you say 'Between E/1000 to E/10000',do you mean add a second row in the 'plastic' materials option?
If so what would the corresponding strain value be
Damian K

Rod Pinna <***@uwa.edu.au> wrote:
Yes, that's all you need. It often helps numerical convergence to give a
small slope to the yield part of the curve. Between E/1000 to E/10000 is
often used.

Rod Pinna
Research Associate
School of Oil & Gas Engineering
The University of Western Australia
35 Stirling Highway
Crawley, 6009, WA

Rod Pinna

2004-04-07 01:20:21 UTC

Permalink

Yep, if you have a perfectly flat plasticity plateau, it can sometimes cause
numerical problems. Usually you specify the second point with a strain much
larger than any you expect.

Saying that, I have never actually had a convergence problem due to not
using a slope, but it is a very common recommendation in FE texts.

Rod Pinna
Research Associate
School of Oil & Gas Engineering
The University of Western Australia
35 Stirling Highway
Crawley, 6009, WA
Tel: +618 6488 2954
Fax: +618 6488 1964
Email: ***@uwa.edu.au
CRICOS code: 00126G

-----Original Message-----
Sent: Tuesday, 6 April 2004 7:12 PM
Subject: RE: [ABAQUS] elastic-perfectly-plastic material
hey rod
thanks for the reply
when you say 'Between E/1000 to E/10000',do you mean add a
second row in the 'plastic' materials option? If so what
would the corresponding strain value be Damian K
Yes, that's all you need. It often helps numerical
convergence to give a small slope to the yield part of the
curve. Between E/1000 to E/10000 is often used.
Rod Pinna
Research Associate
School of Oil & Gas Engineering
The University of Western Australia
35 Stirling Highway
Crawley, 6009, WA

http://groups.yahoo.com/group/abaqus
---------------------------------
Yahoo! Groups Links
http://groups.yahoo.com/group/ABAQUS/
Your use of Yahoo! Groups is subject to the Yahoo! Terms
of Service.
Dept. of Mechanical & Manufacturing Engineering,
Parsons Building,
Trinity College Dublin,
Dublin 2,
Ireland.
---------------------------------
Do you Yahoo!?
Yahoo! Small Business $15K Web Design Giveaway - Enter today
[Non-text portions of this message have been removed]
------------------------ Yahoo! Groups Sponsor
---------------------~--> Buy Ink Cartridges or Refill Kits
for your HP, Epson, Canon or Lexmark Printer at MyInks.com.
Free s/h on orders $50 or more to the US & Canada.
http://www.c1tracking.com/l.asp?cid=5511
http://us.click.yahoo.com/mOAaAA/3exGAA/qnsNAA/PMYolB/TM
--------------------------------------------------------------
-------~->
http://groups.yahoo.com/group/abaqus
Yahoo! Groups Links

------------------------ Yahoo! Groups Sponsor ---------------------~-->
Buy Ink Cartridges or Refill Kits for your HP, Epson, Canon or Lexmark
Printer at MyInks.com. Free s/h on orders $50 or more to the US & Canada.
http://www.c1tracking.com/l.asp?cid=5511
http://us.click.yahoo.com/mOAaAA/3exGAA/qnsNAA/PMYolB/TM
---------------------------------------------------------------------~->

Utpal Borah

2004-04-07 07:23:54 UTC

Permalink

I would like to get some guidelines on the following issues:

Problem Description:

Room temperature bending of a 4209 mm X 3928 mm X 3000 mm stainless steel
plate of 40 mm thickness to radii of curvature of 6020 mm on 4206 mm side,
4515 mm on 3928 mm side and 6450 mm on the 3000 mm side using specially
machined die and punch. This is one sector of a large axisymmetric vessel.
The aim is to calculate the bending load necessary and to assess the
springback after removal of punch. I am using reduced integration shell
elements and rigid die and punch.

I have the following doubts:

1. What should be the starting mesh size for such kind of problems? Is there
any thumb rule for this? I started with a 15 mm mesh, but the model size
becomes huge and takes a long time to complete.

2. Is it possible to use very fine solid elements in lieu of shell elements?
It is predominantly a bending situation with some stretching. Is it correct
to solid elements for such kind of problems?

3. For shell elements, is there any relation between element length and
thickness of the shell? I mean, how the shell thickness affects the element
edge length selection?

With kind regards to all

Utpal
07.04.2004

------------------------ Yahoo! Groups Sponsor ---------------------~-->
Buy Ink Cartridges or Refill Kits for your HP, Epson, Canon or Lexmark
Printer at MyInks.com. Free s/h on orders $50 or more to the US & Canada.
http://www.c1tracking.com/l.asp?cid=5511
http://us.click.yahoo.com/mOAaAA/3exGAA/qnsNAA/PMYolB/TM
---------------------------------------------------------------------~->

benjamin_hagege

2004-04-07 21:34:18 UTC

Permalink

1/It depends on element type. If solid elements are used, put at
least 4 elements in the thickness.
2/Yes (why not ?), but do NOT use C3D8 elements because of the shear
locking. Use C3D8R (a lot) or C3D20R (or C3D20).
3/NO. It is often said in the industry that the size of the shell
elements cannot be as small as the thickness... This pure bullshit
because finite elements are supposed to discretize a continuous
theory !!! You can take the size you want.

C ya.

benZ.

Post by Utpal Borah
Room temperature bending of a 4209 mm X 3928 mm X 3000 mm stainless steel
plate of 40 mm thickness to radii of curvature of 6020 mm on 4206 mm side,
4515 mm on 3928 mm side and 6450 mm on the 3000 mm side using

specially

Post by Utpal Borah
machined die and punch. This is one sector of a large axisymmetric vessel.
The aim is to calculate the bending load necessary and to assess the
springback after removal of punch. I am using reduced integration shell
elements and rigid die and punch.
1. What should be the starting mesh size for such kind of problems? Is there
any thumb rule for this? I started with a 15 mm mesh, but the model size
becomes huge and takes a long time to complete.
2. Is it possible to use very fine solid elements in lieu of shell elements?
It is predominantly a bending situation with some stretching. Is it correct
to solid elements for such kind of problems?
3. For shell elements, is there any relation between element length and
thickness of the shell? I mean, how the shell thickness affects the element
edge length selection?
With kind regards to all
Utpal
07.04.2004

Utpal Borah

2004-04-09 04:25:34 UTC

Permalink

dear benZ,

Thanks for your valuable suggestions.

With kind regards

Utpal

-----Original Message-----
From: benjamin_hagege [mailto:***@yahoo.fr]
Sent: Thursday, April 08, 2004 3:04 AM
To: ***@yahoogroups.com
Subject: [ABAQUS] Re: Meshing guidelines for plate bending simulations

1/It depends on element type. If solid elements are used, put at
least 4 elements in the thickness.
2/Yes (why not ?), but do NOT use C3D8 elements because of the shear
locking. Use C3D8R (a lot) or C3D20R (or C3D20).
3/NO. It is often said in the industry that the size of the shell
elements cannot be as small as the thickness... This pure bullshit
because finite elements are supposed to discretize a continuous
theory !!! You can take the size you want.

C ya.

benZ.

specially