Mode Shape

jwaugh · Post by **jwaugh** » Thu Apr 12, 2007 8:23 am

Is there a script someone has written so I can get the mode shapes for a structure? Thanks.

Jon

silvia · Post by **silvia** » Thu Apr 12, 2007 8:38 am

if you just want to view them, use the usual viewing script, with the following:

set nEigen 3
display -$nEigen 0 $dAmp; # display mode shape for mode $nEigen

shows you the mode shape for mode 3. the - is important!

otherwise, you need to put a recorder at each node and use the eigen option -- see the manual.

jwaugh · Post by **jwaugh** » Thu Apr 12, 2007 8:51 am

What if I want the shape normalized to the top level displacement? I would like the displacements at the floor level output to a file.

silvia · Post by **silvia** » Thu Apr 12, 2007 11:09 am

i don't know how they are normalized, you'd have to do that in post-processing.
yes, you get the eigen-mode displacements put into a file by the recorder.

jwaugh · Post by **jwaugh** » Thu Apr 12, 2007 12:55 pm

Silvia,

I tried to use the eigen vector recorder and so I wrote in my code like this:

# recorders for the getting the mode shapes
recorder Node -file Mode1.out -time -node 3 5 7 9 11 13 15 17 -dof 3 "eigen 1"
recorder Node -file Mode2.out -time -node 3 5 7 9 11 13 15 17 -dof 3 "eigen 2"
recorder Node -file Mode3.out -time -node 3 5 7 9 11 13 15 17 -dof 3 "eigen 3"
recorder Node -file Mode4.out -time -node 3 5 7 9 11 13 15 17 -dof 3 "eigen 4"
recorder Node -file Mode5.out -time -node 3 5 7 9 11 13 15 17 -dof 3 "eigen 5"

source getPeriod.tcl

procGetTperiod 5 on; # get Natural Periods and print to screen (on)

I either get fblank files or an error saying the eigen vectors don't exist. What am I doing wrong?

endryus · Post by **endryus** » Thu Apr 12, 2007 10:38 pm

I was not able to make it working too.
You can use

Code: Select all

print nodes.out node

andrea

Libra23 · Post by **Libra23** » Thu Apr 12, 2007 10:52 pm

Code: Select all

I either get fblank files or an error saying the eigen vectors don't exist. What am I doing wrong?

Are you sure dof 3 is free?

jwaugh · Post by **jwaugh** » Fri Apr 13, 2007 8:08 am

Yes, that is the only DOF with Masses associated with it.

Libra23 · Post by **Libra23** » Fri Apr 13, 2007 9:27 am

Try to make run one step of transient analysis,it works to me .

jwaugh · Post by **jwaugh** » Fri Apr 13, 2007 1:45 pm

I still get:

Node::getEigenvectors<> = eigenvectors have not been set.

What is going wrong?

silvia · Post by **silvia** » Fri Apr 13, 2007 6:33 pm

i think you may have to do an eigen analysis first, using the eigen command.....

jwaugh · Post by **jwaugh** » Sun Apr 15, 2007 9:47 am

I have tried editing the script you gave me to find the period so it reads as following:

proc procGetTperiod {Neigen {PrintScreen "off"}} {
set fmt1 "Mode=%.1i: Tperiod=%.3f "
set PI [expr 2*asin(1.0)]; # define constants

set iTperiod ""
set lambdaN [eigen $Neigen]
for {set i 1} {$i <= $Neigen} {incr i 1} {; # zero to one
set lambda [lindex $lambdaN [expr $i-1]];
set omega [expr pow($lambda,0.5)]
set Tperiod [expr 2*$PI/$omega]; # period (sec.)
lappend iTperiod $Tperiod
if {$PrintScreen == "on" } {puts [format $fmt1 $i $Tperiod ]};
}

# recorder Node -file Mode1.out -time -node 5 7 9 11 13 15 17 -dof 3 "eigen 1"
# recorder Node -file Mode2.out -time -node 5 7 9 11 13 15 17 -dof 3 "eigen 2"
# recorder Node -file Mode3.out -time -node 5 7 9 11 13 15 17 -dof 3 "eigen 3"
# recorder Node -file Mode4.out -time -node 5 7 9 11 13 15 17 -dof 3 "eigen 4"
# recorder Node -file Mode5.out -time -node 5 7 9 11 13 15 17 -dof 3 "eigen 5"
print nodes.out -node 3 5 7 9 11 13 15 17
return $iTperiod
};

neither the recorder nor the pront command gives me the displacement at the nodes. THanks for you help

silvia · Post by **silvia** » Sun Apr 15, 2007 1:59 pm

try getDisp, or something like that.

jwaugh · Post by **jwaugh** » Sun Apr 15, 2007 7:18 pm

I updated the proc to read like this:

proc procGetTperiod {Neigen {PrintScreen "off"}} {
set fmt1 "Mode=%.1i: Tperiod=%.3f "
set PI [expr 2*asin(1.0)]; # define constants

set iTperiod ""
set lambdaN [eigen $Neigen]
for {set i 1} {$i <= $Neigen} {incr i 1} {; # zero to one
set lambda [lindex $lambdaN [expr $i-1]];
set omega [expr pow($lambda,0.5)]
set Tperiod [expr 2*$PI/$omega]; # period (sec.)
lappend iTperiod $Tperiod
if {$PrintScreen == "on" } {puts [format $fmt1 $i $Tperiod ]};
puts "Node 5: [nodeDisp 5 3]"
puts "Node 7: [nodeDisp 7 3]"
puts "Node 9: [nodeDisp 9 3]"
puts "Node 11: [nodeDisp 11 3]"
puts "Node 13: [nodeDisp 13 3]"
puts "Node 15: [nodeDisp 15 3]"
puts "Node 17: [nodeDisp 17 3]"

}

# recorder Node -file Mode1.out -time -node 5 7 9 11 13 15 17 -dof 3 "eigen 1"
# recorder Node -file Mode2.out -time -node 5 7 9 11 13 15 17 -dof 3 "eigen 2"
# recorder Node -file Mode3.out -time -node 5 7 9 11 13 15 17 -dof 3 "eigen 3"
# recorder Node -file Mode4.out -time -node 5 7 9 11 13 15 17 -dof 3 "eigen 4"
# recorder Node -file Mode5.out -time -node 5 7 9 11 13 15 17 -dof 3 "eigen 5"
# print nodes.out -node 3 5 7 9 11 13 15 17
return $iTperiod
};

However, all I get on the ouput is zero displacement for each of the nodes. nodeDisp is the only command I could find similar to the getDisp you mentioned. THanks

silvia · Post by **silvia** » Sun Apr 15, 2007 9:30 pm

J-
please send me your scripts tomorrow: opensees-support @berkeley.edu

also, whenever you guys modify procs, especially mine, you should give them a new name (you can put in a comment that this is a modification of such and suc proc by so and so). This way we don't get confused.