Gravity Analysis

[solmaz_moradpoor]

hi
i modeled a building with 20 storey(sac20).i used 2 analysis.gravity and dynamic analysis.the second analysis run.but gravity analysis didnt run.
i used below scripts:

pattern Plain 1 Linear {
load 71 0.0 -$Weight1 0.0
load 72 0.0 -$Weight1 0.0
load 73 0.0 -$Weight1 0.0
load 74 0.0 -$Weight2 0.0
load 75 0.0 -$Weight2 0.0
load 76 0.0 -$Weight2 0.0
load 77 0.0 -$Weight2 0.0
load 78 0.0 -$Weight2 0.0
load 79 0.0 -$Weight2 0.0
load 80 0.0 -$Weight2 0.0
load 81 0.0 -$Weight2 0.0
load 82 0.0 -$Weight2 0.0
load 83 0.0 -$Weight2 0.0
load 84 0.0 -$Weight2 0.0
load 85 0.0 -$Weight2 0.0
load 86 0.0 -$Weight2 0.0
load 87 0.0 -$Weight2 0.0
load 88 0.0 -$Weight2 0.0
load 89 0.0 -$Weight2 0.0
load 90 0.0 -$Weight2 0.0
load 91 0.0 -$Weight2 0.0
load 92 0.0 -$Weight3 0.0
}

set Tol 1.0e-6;

system UmfPack;

constraints Transformation;

test NormDispIncr $Tol 100 0;

algorithm Newton;

numberer RCM;

integrator LoadControl 1;

analysis Static;

analyze 10

loadConst -time 0.0;

after running it gave below warning:

Warning: CTestNormDispIncr ::test() -failed converge

after: 100 iterations

NewtnRaphson :: solve current step()

-the convergenceTest object failed in test()

StaticAnalysis ::analyze() -the Algorithm failed at iteration :0

with domain at load factor 1

thanks

[fmk]

well either cut down the LoadControl step size and do more analysis or modify the analysis script to do things if it fails .. also cut down the num of steps in the convergence test, 100 is way too many (10 os so should be fine) .. you can also put a 1 at end of test command to see what's happening to the two norms.

[solmaz_moradpoor]

hi
very thanks about your help
i changed loadcontrol step size until 0.01 and finally gravity analysis run.



set Tol 1.0e-6;

system UmfPack;

constraints Plain;

test NormDispIncr $Tol 10 1;

algorithm Newton;

numberer RCM;

integrator LoadControl 0.01;

analysis Static;

analyze 10

loadConst -time 0.0;

but now dynamic analysis that befor run,at time 18.92 gives error and opensees close.
my script is:

set Nsteps [expr int($TmaxAnalysis/$DtAnalysis)];
set ok 0;
set controlTime [getTime];
set nModes 20
set i 0
for {set i 0} {$i<$Nsteps} {incr i 1} {

while {$controlTime < $TmaxAnalysis && $ok == 0} {

set ok [analyze 1 $DtAnalysis]

set controlTime [getTime]

set ok [analyze 1 $DtAnalysis]

if {$ok != 0} {

puts "Trying NewtonWithLineSearch .."

test NormDispIncr $Tol 10 0

algorithm NewtonLineSearch .8

set ok [analyze 1 $DtAnalysis]

test $TestType $Tol $maxNumIter 0

algorithm NewtonLineSearch .8
}
if {$ok != 0} {

puts "Trying Broyden .."

algorithm Broyden 8

set ok [analyze 1 $DtAnalysis]

algorithm Broyden 8
}
if {$ok != 0} {

puts "Trying Newton with Initial Tangent .."

algorithm Newton -initial


set ok [analyze 1 $DtAnalysis]

algorithm Newton -initial
}

if {$ok == 0} {
eigen $nModes
}
}

}
puts "Ground Motion Done. End Time: [getTime]"

you think where is problem?
do my model is wrong?and shoud verify my model?
thankyou.

[fmk]

your model probably has a problem .. as do your analysis scripts!

1) do you relize that with a step size of 0.01 inload control you probably need 100 steps to get the full gravity load.
2) in your analysis routine where you do something if a failure to converge has occured, after you issue the analyze command you should set the algorithm back to Newton and not the new algorithm you are trying.
3) Netwon -initial usually works best, however it needs more iterations than the regular newton and after you have used it you need to set the test back to being

here is a simple example:

[url]
http://opensees.berkeley.edu/wiki/index ... ple3.3.tcl
[/url]

[solmaz_moradpoor]

thank you very much.