Database - Saving and Restarting Analysis

Forum for OpenSees users to post questions, comments, etc. on the use of the OpenSees interpreter, OpenSees.exe

Moderators: silvia, selimgunay, Moderators

Post Reply
motalaat

Database - Saving and Restarting Analysis

Post by motalaat » Wed Jul 21, 2004 11:07 am

It seems that the methods for the database command are not fullt implemented. The command is not supported for Corotational formulation, and apparently for some advanced material models (one or both of Concrete02 and Steel02)

fmk
Site Admin
Posts: 5883
Joined: Fri Jun 11, 2004 2:33 pm
Location: UC Berkeley
Contact:

Post by fmk » Mon Aug 09, 2004 7:25 pm

i have fixed the code for these .. avail with 1.6.0 and higher.

if anybody notices any other things that fail when the save/restore commands are used please post them in this thread.

thanks.
Last edited by fmk on Thu Jun 02, 2005 9:13 am, edited 1 time in total.

Paolo Franchin

Use of database related commands save and restore

Post by Paolo Franchin » Wed Sep 22, 2004 4:38 am

I have tried to use the database command with -type set to other than File (MySQL and BerkeleyDB): I receive a message stating that no such types exist (I triple checked the spelling) and telling me the available type, i.e. MySQL and BerkeleyDB...

Also, I have tried to open the database files produced by the only option I could make to work (File) but they are not among the recognised file types of Access (I must admit that I am not that conversant with databases). They look binaries when opened in a text editors, can someone tell me their format and how to open them?

Thank you

Paolo Franchin

Follow-up on the use of database: geotech problem

Post by Paolo Franchin » Wed Sep 22, 2004 4:46 am

A further explanation on what I am trying to do, just in case someone nows how to circumvent my problem without recourse to databases.

I am analysing a model with soil and a concrete pile wall (as a 2D problem). Since I am modelling the soil nonlinearly I would like to have a realistic state of stress in it at the end of the gravity loads phase, which implies removal of part of the soil on one side of th wall. This amounts not only to removing the corresponding load but also the stiffness etc. It is a problem which requires the capability of removing elements from the mesh. As far as I see I can't do that yet in OS. I thought about saving the model in DB after gravity loads, "setting to zero the stiffness" of some elements in the DB, restoring the model and perform time history analysis.

Any suggestion or comment is welcome

fmk
Site Admin
Posts: 5883
Joined: Fri Jun 11, 2004 2:33 pm
Location: UC Berkeley
Contact:

Post by fmk » Wed Sep 29, 2004 10:43 am

Paolo,

you can remove elements in OpenSees .. use the remove command .. just make sure you leave no hanging nodes and loads.

the reason that MySQL and BerkeleyDB fail is that we do not by default link to them on windows machines (as requires users to go off and install the packages and then set up a
server on their machine to handle the requests).

frank.

Paolo Franchin

Removing elements

Post by Paolo Franchin » Mon Oct 04, 2004 6:45 am

Frank,

thank you! I suppose this further question should go to the documentation section (I'll post it there too): what is the syntax of the remove command? It is not in the manual version on the website (or am I wrong?).

Thank you in advance

Paolo

fmk
Site Admin
Posts: 5883
Joined: Fri Jun 11, 2004 2:33 pm
Location: UC Berkeley
Contact:

Post by fmk » Mon Oct 04, 2004 10:31 am

i have posted a response to this last question in the Documentation forum.

Rocky
Posts: 38
Joined: Tue May 17, 2005 7:57 pm
Location: IEM

Post by Rocky » Thu Jun 02, 2005 3:01 am

I also want to know if there exist any command for adding some elements for constructure analysis???And if exist, what's the syntax??/
Thank you in advance.

Rocky
Posts: 38
Joined: Tue May 17, 2005 7:57 pm
Location: IEM

Post by Rocky » Thu Jun 02, 2005 3:25 am

Hello.
I make a small example for checking restart capability of OpenSees.
But error occurs, I wonder if someone can tell me what's the problem.
Thanks in advance.

I have three steps of linear analysis, after the second step, I hope to
restart from the end of the first step analysis. But some error occurs.

The main Tcl file (including first & second steps)
------------------------------------------------------------
### The test of restart capability of OpenSees.
### Units: kN, mm
### by WangTao 2005.6.1

model BasicBuilder -ndm 2 -ndf 3

# Geometry definition
node 1 0.0 0.0
node 2 0.0 1000.0
fix 1 1 1 1

geomTransf Linear 1
element elasticBeamColumn 1 1 2 100.0 25000.0 40000.0 1

# Define recorder and database
recorder Node -file Node2disp.txt -time -node 2 -dof 1 disp
recorder Element -file Ele1force.txt -time -ele 1 force
set dbtype "File"
set dbname "restart"
database $dbtype $dbname

# Define analysis 1
pattern Plain 1 Linear {
load 2 3.0 0.0 0.0
}

constraints Plain
numberer RCM
system BandGeneral
algorithm Linear
integrator LoadControl 0.1
analysis Static
analyze 10
save 1

# Define analysis 2
wipeAnalysis
loadConst -time 0.0
pattern Plain 2 Linear {
load 2 6.0 0.0 0.0
}

constraints Plain
numberer RCM
system BandGeneral
algorithm Linear
integrator LoadControl 0.1
analysis Static
analyze 10
save 2

The restart Tcl file (including the third step)
------------------------------------------------------------
set dbtype "File"
set dbname "restart"
database $dbtype $dbname

# Define analysis 3
restore 1

pattern Plain 3 Linear {
load 2 9.0 0.0 0.0
}

constraints Plain
numberer RCM
system BandGeneral
algorithm Linear
integrator LoadControl 0.1
analysis Static
analyze 10
save 3


--------------------------------------------------

berktaftali
Posts: 68
Joined: Fri Jul 02, 2004 6:10 am
Location: Computers and Structures, Inc.

Post by berktaftali » Thu Jun 02, 2005 6:15 am

fmk wrote:i have fixed the code for these .. as i don't yet have permission to check them into cvs you will have to wait till the next release v1.6.0 due any day now.
If Frank does not have permission to check code into CVS, who has ??? :o
Berk Taftali
Georgia Institute of Technology
Ph.D. Candidate, Structural Engineering, Mechanics, and Materials
School of Civil and Environmental Engineering
Atlanta, GA 30332 USA
Email: gte994y@mail.gatech.edu

fmk
Site Admin
Posts: 5883
Joined: Fri Jun 11, 2004 2:33 pm
Location: UC Berkeley
Contact:

Post by fmk » Thu Jun 02, 2005 9:13 am

try redifing the modelbuilder in the second file .. when you start opensees you
automatically get all analysis commands .. you do not get any model generation commands, hence the interpreter will fail when it gets to the pattern command .. the model generation only comes after the model command .. this is to allow people to develop their own modelbuilders .. ithe script with this mod works on linux .. post again if it fails on windows.

fmk
Site Admin
Posts: 5883
Joined: Fri Jun 11, 2004 2:33 pm
Location: UC Berkeley
Contact:

Post by fmk » Thu Jun 02, 2005 9:20 am

as for construction sequencing .. you can add elements and new nodes at any stage .. if you add any of the elements that use a coordinate transformation the initial disp at the nodes resulting from previous displacement during construction are subtracted off the current displacement when computing the element deformation.

Rocky
Posts: 38
Joined: Tue May 17, 2005 7:57 pm
Location: IEM

Post by Rocky » Thu Jun 02, 2005 10:38 pm

Well, after adding "model" command in the second, I can do one step restart now, but result seems not so good. And more critical problem
is: when I want to restart again, it seems impossible.

The revised file list below:

Main file:
---------------------------------------------------------
### The test of restart capability of OpenSees.
### Units: kN, mm
### by WangTao 2005.6.1

model BasicBuilder -ndm 2 -ndf 3

# Geometry definition
node 1 0.0 0.0
node 2 0.0 1000.0
fix 1 1 1 1

geomTransf Linear 1
element elasticBeamColumn 1 1 2 100.0 25000.0 40000.0 1

# Define recorder and database
recorder Node -file Node2disp.txt -time -node 2 -dof 1 disp
recorder Element -file Ele1force.txt -time -ele 1 force
set dbtype "File"
set dbname "restart"
database $dbtype $dbname

# Define analysis 1
pattern Plain 1 Linear {
load 2 3.0 0.0 0.0
}

constraints Plain
numberer RCM
system BandGeneral
algorithm Linear
integrator LoadControl 0.1
analysis Static
analyze 10
puts "Result of Step 1:"
print node
save 1

# Define analysis 2
wipeAnalysis
loadConst -time 0.0
pattern Plain 2 Linear {
load 2 6.0 0.0 0.0
}

constraints Plain
numberer RCM
system BandGeneral
algorithm Linear
integrator LoadControl 0.1
analysis Static
analyze 10
puts "Result of Step 2:"
print node
save 2


Restart file:
---------------------------------------------------
model BasicBuilder -ndm 2 -ndf 3

set dbtype "File"
set dbname "restart"
database $dbtype $dbname

# Define analysis 3
restore 1
loadConst -time 0.0

pattern Plain 3 Linear {
load 2 9.0 0.0 0.0
}

constraints Plain
numberer RCM
system BandGeneral
algorithm Linear
integrator LoadControl 0.1
analysis Static
analyze 10
puts "Result of Step 3:"
print node
save 3
----------------------------------------------------

What I do is like this:

source Main.tcl
source Restart.tcl

The disp of Node 2 after the third step should be 4, but it's 3. That's the problem No.1.
If I restore from the end of the second step, the result should be 6, and the result is correct. I don't know why?

The second problem is: when I do like this:

source Main.tcl
wipe
source Restart.tcl

A fatal error occurs. That's the problem No.2.

The third problem is: when I do like this:

source Main.tcl
source Restart.tcl
source Restart.tcl

An error also ocurrs. Cann't OpenSees do a restart analysis repeatedly?
That's the problem No.3.

The third one is critical, if it cannot, I don't think it can do any restart, am I right? I hope it's only because of my ignorance about OpenSees.

Thank you very much to every guys on the forum. And especially to Mr.
fmk.


Best regards
WangTao

Rocky
Posts: 38
Joined: Tue May 17, 2005 7:57 pm
Location: IEM

Post by Rocky » Wed Jun 08, 2005 8:35 pm

Isn't :( it a big problem?

Post Reply