issue with use of OpenSees(MP)

Post Reply
Asad005
Posts: 49
Joined: Sat Dec 07, 2019 6:50 am
Location: Italy

issue with use of OpenSees(MP)

Post by Asad005 » Tue Oct 20, 2020 9:38 am

Thanks if anyone can help in this.

When I used standard OpenSees then the analysis gets stopped (or didn't converge) after 50% of the analysis while by using the OpenSees(mp) for parallel analysis, the model unable to run and gives the following convergence error:

The file is attached herewith for ur consideration.

WARNING:SparseGenColLinSOE::setSize : vertex 0 not in graph! - size set to 0
StaticAnalysis::handle() - LinearSOE::setSize() failedStaticAnalysis::analyze() - domainChanged failed at step 0 of 1
OpenSees > analyze failed, returned: -1 error flag
ERROR: the analysis did not converge
while executing
"error "ERROR: the analysis did not converge""
("for" body line 23)
invoked from within
"for {set increment_counter 1} {$increment_counter <= $initial_num_incr} {incr increment_counter} {
if {$process_id == 0} {
puts "Increment: $increm..."
(file "analysis_steps.tcl" line 116)
invoked from within
"source analysis_steps.tcl"
(file "./main.tcl" line 16)
Process Terminating 0
WARNING:SparseGenColLinSOE::setSize : vertex 0 not in graph! - size set to 0
StaticAnalysis::handle() - LinearSOE::setSize() failedStaticAnalysis::analyze() - domainChanged failed at step 0 of 1
OpenSees > analyze failed, returned: -1 error flag
ERROR: the analysis did not converge
while executing
"error "ERROR: the analysis did not converge""
("for" body line 23)
invoked from within
"for {set increment_counter 1} {$increment_counter <= $initial_num_incr} {incr increment_counter} {
if {$process_id == 0} {
puts "Increment: $increm..."
(file "analysis_steps.tcl" line 116)
invoked from within
"source analysis_steps.tcl"
(file "./main.tcl" line 16)
Process Terminating 1
WARNING:SparseGenColLinSOE::setSize : vertex 7296 not in graph! - size set to 0
StaticAnalysis::handle() - LinearSOE::setSize() failedStaticAnalysis::analyze() - domainChanged failed at step 0 of 1
OpenSees > analyze failed, returned: -1 error flag
ERROR: the analysis did not converge
while executing
"error "ERROR: the analysis did not converge""
("for" body line 23)
invoked from within
"for {set increment_counter 1} {$increment_counter <= $initial_num_incr} {incr increment_counter} {
if {$process_id == 0} {
puts "Increment: $increm..."
(file "analysis_steps.tcl" line 116)
invoked from within
"source analysis_steps.tcl"
(file "./main.tcl" line 16)
Process Terminating 2
destroying lib handle: 0x7ffc70040000 ... 0
destroying lib handle: 0x7ffc70040000 ... 0
destroying lib handle: 0x7ffc70040000 ... 0
Plugin Library "pdm" correctly un-loaded
Plugin Library "pdm" correctly un-loaded
Plugin Library "pdm" correctly un-loaded
Plugin DamageTC3D: number of allocated instances was 8609
Plugin DamageTC3D: number of allocated instances was 13685
Plugin DamageTC3D: number of allocated instances was 13721
Press any key to continue . . .


Regards
Attachments
r2,mergedH26,AL0.2,MP.zip
(1.37 MiB) Downloaded 224 times

STKO Team
Posts: 3070
Joined: Tue Oct 29, 2019 8:45 am

Re: issue with use of OpenSees(MP)

Post by STKO Team » Tue Oct 20, 2020 1:54 pm

When using OpenSeesMP you MUST:
  • Use MUMS as system
  • Use Parallel RCM as numberer
  • Use ParallelDisplacementControl instead of DisplacementControl (only if you want to do displacement-controlled analyses)
  • partition your mesh
  • run
However, why are you using shells here? yours is a 2D problem, use plane stress element in a 2D analysis:
r2,mergedH26,AL0.2,MP.scd.zip
(1.3 MiB) Downloaded 254 times
Last edited by STKO Team on Tue Oct 20, 2020 2:22 pm, edited 1 time in total.

Asad005
Posts: 49
Joined: Sat Dec 07, 2019 6:50 am
Location: Italy

Re: issue with use of OpenSees(MP)

Post by Asad005 » Tue Oct 20, 2020 2:11 pm

Thanks for ur assistance.
But I did in the same suggested way before and also again but still, the problem occurs as shown below.

puts "Increment: $increm..."
(file "analysis_steps.tcl" line 116)
invoked from within
"source analysis_steps.tcl"
(file "./main.tcl" line 16)
Process Terminating 0
OpenSees > analyze failed, returned: -3 error flag
ERROR: the analysis did not converge
while executing
"error "ERROR: the analysis did not converge""
("for" body line 23)
invoked from within
"for {set increment_counter 1} {$increment_counter <= $initial_num_incr} {incr increment_counter} {
if {$process_id == 0} {
puts "Increment: $increm..."
(file "analysis_steps.tcl" line 116)
invoked from within
"source analysis_steps.tcl"
(file "./main.tcl" line 16)
Process Terminating 1
destroying lib handle: 0x7ffc9b480000 ... 0
destroying lib handle: 0x7ffc9b480000 ... 0
Plugin Library "pdm" correctly un-loaded
destroying lib handle: 0x7ffc9b480000 ... 0
Plugin Library "pdm" correctly un-loaded
Plugin Library "pdm" correctly un-loaded
Plugin DamageTC3D: number of allocated instances was 8609
Plugin DamageTC3D: number of allocated instances was 13685
Plugin DamageTC3D: number of allocated instances was 13721
Press any key to continue . . .

The screenshots of the steps being followed are also attached herewith for ur kind consideration.

The same model is on run with the standard OpenSees as shown below but it takes a large piece time (especially when need to revise with other algorithms in case of error) to analyze it with standard OpenSees:

Increment: 402 - Iterations: 9 - Norm: 7.19376837442025967766e-09 ( 29.088277858176546 % )
Increment: 403. time_increment = 0.000723589001447178. Current time = 0.2908827785817655
Increment: 403 - Iterations: 10 - Norm: 4.62153059805799586669e-09 ( 29.160636758321267 % )
Increment: 404. time_increment = 0.000723589001447178. Current time = 0.2916063675832127
Increment: 404 - Iterations: 8 - Norm: 1.60199873888798576254e-09 ( 29.23299565846599 % )
Increment: 405. time_increment = 0.000723589001447178. Current time = 0.2923299565846599
Increment: 405 - Iterations: 23 - Norm: 6.50425562579786025121e-09 ( 29.30535455861071 % )
Increment: 406. time_increment = 0.000723589001447178. Current time = 0.2930535455861071
Increment: 406 - Iterations: 9 - Norm: 8.44208042273660885141e-09 ( 29.377713458755427 % )
Increment: 407. time_increment = 0.000723589001447178. Current time = 0.2937771345875543
Increment: 407 - Iterations: 10 - Norm: 4.50220451196524470659e-09 ( 29.450072358900147 % )
Increment: 408. time_increment = 0.000723589001447178. Current time = 0.2945007235890015
Increment: 408 - Iterations: 8 - Norm: 1.67102907038970935532e-09 ( 29.52243125904487 % )
Increment: 409. time_increment = 0.000723589001447178. Current time = 0.2952243125904487
Increment: 409 - Iterations: 24 - Norm: 9.03789264571465778421e-09 ( 29.59479015918959 % )
Increment: 410. time_increment = 0.000723589001447178. Current time = 0.2959479015918959
Increment: 410 - Iterations: 10 - Norm: 2.27389143077854885861e-09 ( 29.667149059334307 % )
Increment: 411. time_increment = 0.000723589001447178. Current time = 0.2966714905933431
Attachments
MPpartition.png
MPpartition.png (91.05 KiB) Viewed 3322 times
MP.png
MP.png (37.65 KiB) Viewed 3322 times
MP,Anlysisi RUn.png
MP,Anlysisi RUn.png (139.35 KiB) Viewed 3322 times

STKO Team
Posts: 3070
Joined: Tue Oct 29, 2019 8:45 am

Re: issue with use of OpenSees(MP)

Post by STKO Team » Wed Oct 21, 2020 10:46 am

Try the model I sent you in the previous post:
However, why are you using shells here? yours is a 2D problem, use plane stress element in a 2D analysis:
r2,mergedH26,AL0.2,MP.scd.zip
(1.3 MiB) Not downloaded yet
When I used the 2d plane stress it worked. I'm afraid there is something wrong with that shell element

Asad005
Posts: 49
Joined: Sat Dec 07, 2019 6:50 am
Location: Italy

Re: issue with use of OpenSees(MP)

Post by Asad005 » Wed Oct 21, 2020 11:17 am

Thanks a lot for ur kind response.
Sorry, unfortunately, I did not see the last line in your previous response. I will do it as suggested.
regards

Post Reply