The OpenSees Community

Dear Frank
How may I send you my file? It is too large to paste here.

Thanks

Dear fmk,
I am developing a custom material. Since I failed to build openseesMP with the new material included, I am trying to introduce my material as a supplementary dll to OpenSeesMP version 2.4.5. The material seems to load and run successfully but after the first round of model building, analyzing, and wiping finishes successfully, the "model" command invoked at the beginning of the second round causes the OpenSeesMP to exit. The mpiexec terminates with following lines:

"job aborted:
rank: node: exit code[: error message]
0: my-pc: 123
1: my-pc: -1073740940: process 1 exited without calling finalize"

my tcl code is as follows:

"set generalFolder general

set np [getNP]
set pid [getPID]
if {$pid == 0} {
# Coordinator
foreach modelFolder {MLL MML MHL} {
for {set iRec 1 } {$iRec < $numRecs} {incr iRec} {
recv -pid ANY pidWorker
puts "pidWorker= $pidWorker\n"
send -pid $pidWorker [list $modelFolder $iRec]
}
}
for {set i 1} {$i < $np} {incr i 1} {
send -pid $i "DONE"
}
} else {
# Worker
set done NOT_DONE;
while {$done != "DONE"} {
send -pid 0 $pid
recv -pid 0 list
puts "list= $list\n"
set modelFolder [lindex $list 0]
if {$modelFolder == "DONE"} {
puts "pid:$pid: Done"
break;
}
set recn [lindex $list 1]
set outFolder $modelFolder
set saInit 0.05
set nPnt 1
set sa 0.05
set GMFile "$generalFolder/GMfiles/transformed/$recn.txt"
set deltaT 0.02
set Tmax 5.
set folder $outFolder/$recn/$sa
source $modelFolder/initialize.tcl
set fac $g
set logfileId [open $folder/log.txt w+]
source $generalFolder/model.analyze.gravity.tcl
source $generalFolder/model.analyze.GM.Uniform.tcl
close $logfileId
}
}
barrier"

The "initialize.tcl" file contains a "wipe" command followed by a "model" command and model generation commands (including definition of the new material which is loaded from a dll). As I said, the model is built and run at the first round (for recn == 1) but it exits unexpectedly at the "model" command invoked at the beginning of the second round.

I wondered if this could be a problem caused by the dll material but I am not sure and would be thankful if you could provide me with your idea.

Thanks and Regards,
SA Jalali

Hi Frank.
Thanks for your answer. Is what you mentioned correct for all types of analyses or just for the one being questioned in above post?
What do you exactly mean by "correction step"?

Hi.
You can find the related descriptions and instructions about joint2d in Altoontash dissertation which can be found on the web.

hi.
the problem is clear. Joint2d nodes must represent nodes at four edges of panel zone, they shall not fall on each other. you must use the dimensions of your panel zone.
--------*3

4*-------------*2

--------*1

Hi.
Could you post your script to see what the problem exactly is?

Hi all.
I'm trying to model a post tensioned structure in opensees. could anybody tell how i can do it?
thank you very much

the link for mentioned source is here:
opensees.berkeley.edu /OpenSees/doc/ Altoontash_Dissertation.pdf

the joint2D element is a useful element for modeling the panel zone as well as the rotational deformability of beam ends. as this element is not described in documentation neither in pdf nor in HTML format , I thought it would be helpful for other users to put the parameters of this command:

element Joint2D tag? Nd1? Nd2? Nd3? Nd4? NdC? [Mat1? Mat2? Mat3? Mat4?] MatC? LrgDspTag?

tag: unique element tag
Nd1: integer tag indicating node 1
NdC: integer tag indicating the central node. this tag is used for generating the internal node (the node is generated automatically by element command) and must not exist in the domain
Mat1 ~ Mat4 :a group of optional integer tags indicating the uniaxial materials for interface rotational springs injoint2D model. use a zero tag to indicate the case that a beam-column element is rigidly framed to the joint
MatC : uniaxial material for the shear panel behavior
LrgDsdTag: 0 for small deformations and constant geometry; 1 for large deformations and time varying geometry; 2 for large deformations ,time varying geometry and length correction

the nodes must be located such that the main chords bisect, and the node tags are required to be entered in a clockwise or counter-clockwise order

source:"simulation and damage models for performance assessment of reinforced concrete beam-column joints" by Arash Altoontash

Thank you Mr McKenna.
I will read the document and call you for further help.

Good Luck

Hi,
could you help me how I can link external routines dynamically to my code or where I can find some document about this matter to read?

Thanks

Thank you.

I could find the answer.
the formulation given in the manual for calculating the isotropic strain hardening is not correct and appears that it should be modified to the below formula which I found in literature:
SIGMAst= SIGMAy*a2*(EPSILONmax/EPSILONy-a3)

Regards
Alireza Jalali

Hi
I am trying to use steel02 material as an element force-deformation response. the element's yield strength starts to deteriorate after a deformation of .04 . I want to use steel02 material isotropic strain hardening parameters a1, a2, a3, and a4 but the mentioned decrease starts at earlier points than I desire. can you tell me how should I set the parameters to get the deterioration start exactly at the needed point?
thank you

thank you very much indeed.
You are very kind to me

ok, I looked up the web for the dissertation but I couldn't find it. Just a 1 page summary report of it could be find on Blume center web site. Can you help me where i can find it?
thank you.