Problem with the pattern "MultipleSupport"

[aminrahmaani]

Hi all,

I modeled a cantilever with a lumped mass on its top (very simple model). I want to excite the column at the base by using pattern "MultipleSupport" instead of pattern "UniformExcitation". Using "MultipleSupport", I get unreasonable acceleration at the top of the column( e.g a=170 m/s^2); however if I use "UniformExcitation" pattern I get reasonable acceleration. Has anybody seen this problem, before?

Thanks,
Amin

[vesna]

Are you assigning appropriate displacement history to your column when using "MultipleSupport" excitation?

[aminrahmaani]

What I do is that, I have sinusoidal acceleration time history; then I use SeismoSignal software to convert it to displacement.

[vesna]

would you post your example so I can test it?

[aminrahmaani]

For sure! The acceleration time history that I'm using is sinusoidal with amplitude of 0.2g and period of 0.5 sec.

# Units: m, sec, ton, kN

wipe all


model basic -ndm 3 -ndf 6
node 1 0 0 0
node 2 0 0 1
node 3 0 0 2
node 4 0 0 3
node 5 0 0 4.4


fix 1 0 1 1 1 1 1

geomTransf Linear 101 1 0 0

set D 0.5
set A [expr 3.1416*($D/2.0)*($D/2.0)]
set E 200.e+6
set G [expr $E/2.0/(1+0.4)]
set Iz [expr (1.0/4.0)*(3.1416)*(pow($D/2 , 4))]
set Iy [expr (1.0/4.0)*(3.1416)*(pow($D/2 , 4))]
set J [expr $Iz+$Iy]

for { set k 1 } { $k <= 4 } { incr k 1 } {
element elasticBeamColumn $k $k [expr $k+1] $A $E $G $J $Iy $Iz 101
}

mass 5 4.534 4.534 4.534 0 0 0

loadConst -time 0.0

set FACT 1
timeSeries Path 500 -fileTime SineTIME.txt -filePath SineACCEL.txt -factor [expr $FACT]
pattern MultipleSupport 2000 {

groundMotion 2000 Plain -disp 500

imposedSupportMotion 1 1 2000

}

recorder Node -file accel.out -time -nodeRange 1 5 -dof 1 accel
recorder Element -file pileforcesGlob.out -time -eleRange 1 4 globalForce
recorder Node -file disp.out -time -nodeRange 1 5 -dof 1 disp


set gamma 0.500000
set compdT 0.01


integrator Newmark $gamma [expr pow($gamma+0.5,2)/4]
numberer RCM
constraints Penalty 1.0e12 1.0e12
test NormDispIncr 1e-3 50 1
algorithm KrylovNewton
system BandGeneral
analysis VariableTransient

for {set numIncr 1} {$numIncr <= 100 } {incr numIncr 1} {
puts "##### Seismic excitation step : $numIncr #####";
analyze 1 $compdT [expr $compdT/100.0] $compdT 20
}

[vesna]

timeSeries 500 is acceleration (based on the name of your input file SineACCEL.txt) and you assign it at node 1 in dof 1 as a displacement.

To correct it generate SineDISP.txt based on SineACCEL.txt and define a timeSeries 500 using SineDISP.txt.

[aminrahmaani]

Yes, I know that. That file contains displacements. That's not my problem. Can you see any other problem with my code?

[vesna]

Your file looks OK to me.

When looking at the output are you aware that the results for the responses at the nodes are the ABSOLUTE values, and not relative values as in the case of a UniformExciatation? To get relative acceleration subtract ground motion from recorded acceleration.