Strange seismic response on high frame building

[felipe_cordero]

Dear Frank,

My name is Felipe Cordero and I'm modeling wall buildings in OpenSees for my engineering thesis under the guidance of prof. Leonardo Massone, who has contributed to Opensees before. I've been modeling several 2D frame buildings, trying to explain the seismic behavior and the damage observed on some building after the Mw 8.8 earthquake that affected Chile on 27th February. I've sent you an email very early today.

I'll appreciate any help you can provide.

Felipe Cordero O.

[felipe_cordero]

Dear Vesna,

could I send the files to you and could you please take a look of them?

thanks!

[fmk]

one would ask why you think the results of the earthuake analysis response would lie on the pushover curve as you add more floors? for the pushover curve you are applying a single load pattern, the reference loads (probably linear) do not vary w.r.t each other over time.

i would suggest that the pushover analysis is similar for smaller models because these lateral loads you are apply most closely resemble the first mode shape. As the building gets higher, other modes come into play and the response becomes more affected by the other modes, which is why as you are increasing the number of floors the difference from the static pushover increases.

i suggest you look at the elastic case for your model first and do a modal transient analysis to see what i am talking about.

[felipe_cordero]

Thanks frank

[civilengr_tahir]

hi dear Vesna and felipe_cordero..
im using using the dispBeamColumnInt for modeling of walls..My problem is that when I use concrete 01 materal then sloution get converged for cyclic pushover anaylsis but its not working with cocncrete 06 material for cyclic pushover...Although my monotonic pushover results are ok with concrete06..do u people have any idea why its like that...im pretty confused...plz help me

[vesna]

I never tested Concrete06 so I really do not know what could be the reason. Do you have a loop in your code that changes the algorithm in the case there is a convergence problem?

[civilengr_tahir]

hi dear vesna...
thanx for ur urgent reply...Yes im using loop in cyclic pushover command but now its get converged..I have change a bit code regarding step size..however I have an other problem now...The results of cyclic pushover model with concrete06 are very strange infact far away from experimental results...is there somthing wrong with concrete06 material for cyclic behaviour....plz if anybody else can answer...im posting my model
wipe; model basic -ndm 2 -ndf 3

#unit KN,m,sec


#concrete parameters

set fc -49000; set e0 -0.002;

set n 2;
set r 1;

set fcr 4900;

set ecr 0.00008;
set b 0.4;



set alphaC 0.32;set alphaT 0.08;



#steel parameters

set fyRed 455000; set by 0.02; set Ry 20;

set fyBRed 455000; set byB 0.02;set RyB 20.0;



#concrete (confined and unconfined)

#concrete (confined and unconfined)
uniaxialMaterial Concrete06 1 $fc $e0 $n $r $alphaC $fcr $ecr $b $alphaT;



# steel (bound., web and horiz. reinforcement)

set E 199948000.013;

uniaxialMaterial Steel02 1003 $fyBRed $E $byB $RyB 0.9 0.1 0 0.1 0 0.1 ;

uniaxialMaterial Steel02 1004 $fyRed $E $by $Ry 0.9 0.1 0 0.1 0 0.1 ;





# Define cross-section
set t1 0.380; set NStrip1 2; # thickness 1
set t2 0.15; set NStrip2 4; # thickness 2
set t3 0.380; set NStrip3 2; # thickness 3

geomTransf LinearInt 1
set np 1; # int. points
set C 0.4; # center of rotation


#section definition
section FiberInt 2 -NStrip $NStrip1 $t1 $NStrip2 $t2 $NStrip3 $t3 {

#vertical fibers
fiber -0.76175 0 0.03857 1; fiber -0.76175 0 0.0003 1004;
fiber -0.66025 0 0.03857 1; fiber -0.66025 0 0.0002 1004;
fiber -0.45712 0 0.116 1; fiber -0.45712 0 0.0001 1004;
fiber -0.1523 0 0.116 1; fiber -0.1523 0 0.0001 1004;
fiber 0.1523 0 0.116 1; fiber 0.1523 0 0.0001 1004;
fiber 0.45712 0 0.116 1; fiber 0.45712 0 0.0001 1004;
fiber 0.66025 0 0.03857 1; fiber 0.66025 0 0.0002 1004;
fiber 0.76175 0 0.03857 1; fiber 0.76175 0 0.0003 1004;





#horiz. reinf.
Hfiber 0 0 0.000748 1004;

}



#nodes
node 1 0 0
node 2 0 2.03
node 3 0 4.06
node 4 0 6.09
node 5 0 8.12
node 6 0 10.15
node 7 0 12.18



puts "Generated - NODES"

#element definition
element dispBeamColumnInt 1 1 2 $np 2 1 $C
element dispBeamColumnInt 2 2 3 $np 2 1 $C
element dispBeamColumnInt 3 3 4 $np 2 1 $C
element dispBeamColumnInt 4 4 5 $np 2 1 $C
element dispBeamColumnInt 5 5 6 $np 2 1 $C
element dispBeamColumnInt 6 6 7 $np 2 1 $C

puts "Generated - ELEMENTS"
fix 1 1 1 1
# Define Mass


puts "Generated - MASS"
# Set axial load

pattern Plain 1 Constant {

load 7 0.0 -1500 0.0


}



constraints Transformation;
numberer RCM;
system UmfPack;
test NormUnbalance 1.0e-1 200 5
algorithm Newton
integrator LoadControl 0.1
puts "START : Gravity Load Analysis"
analysis Static
for {set i 1} {$i <= 9} {incr i} {
puts "Step = $i/10"
analyze 1

}
# Set the gravity loads to be constant & reset the time in the domain

loadConst -time 0.0

remove recorders


system BandGeneral; constraints Transformation; numberer Plain;


# Create a recorder to monitor nodal displacement and element forces

recorder Node -file nodeTop.out -node 7 -dof 1 disp

recorder Element -file elesX.out -time -ele 1 globalForce


# recorder for element1 section1 steel stress/strain and section force-def.

recorder Element -file Sect_FandD.out -ele 1 section 1 forceAndDeformation

recorder Element -file Sect_eX.out -ele 1 section 1 eX

recorder Element -file Sect_eY.out -ele 1 section 1 eY

recorder Element -file Sect_sX.out -ele 1 section 1 sX

recorder Element -file Sect_sY.out -ele 1 section 1 sY
puts "Step = 10/10"
analyze 1
puts "COMPLETED : Gravity Load Analysis"


puts "Model Built"

set ControlNode 7;
set H 12.2;
set dU 0.09;
set cycle 2;

constraints Transformation;
numberer RCM;
system UmfPack;
test NormDispIncr 1.0e-1 100 5
algorithm KrylovNewton
pattern Plain 2 Linear {
load $ControlNode 1000 0 0 0 0 0
}

foreach Drift {0.13 0.25 0.50 0.75 1 1.5 2 2.5 3} {
for {set i 1} {$i <= $cycle} {incr i} {
integrator DisplacementControl $ControlNode 1 [expr (+1) * $dU * $H / 100];
for {set j 1} {$j <= [expr int($Drift/$dU)]} {incr j} {
puts "drift (%) = $Drift, cycle = $i, step = [expr $j + int($Drift/$dU*0)] / [expr int($Drift/$dU*4)]";
analyze 1;
}
integrator DisplacementControl $ControlNode 1 [expr (-1) * $dU * $H / 100];
for {set j 1} {$j <= [expr int($Drift/$dU)]} {incr j} {
puts "drift (%) = $Drift, cycle = $i, step = [expr $j + int($Drift/$dU*1)] / [expr int($Drift/$dU*4)]";
analyze 1;
}
integrator DisplacementControl $ControlNode 1 [expr (-1) * $dU * $H / 100]
for {set j 1} {$j <= [expr int($Drift/$dU)]} {incr j} {
puts "drift (%) = $Drift, cycle = $i, step = [expr $j + int($Drift/$dU*2)] / [expr int($Drift/$dU*4)]";
analyze 1;
}
integrator DisplacementControl $ControlNode 1 [expr (+1) * $dU * $H / 100]
for {set j 1} {$j <= [expr int($Drift/$dU)]} {incr j} {
puts "drift (%) = $Drift, cycle = $i, step = [expr $j + int($Drift/$dU*3)] / [expr int($Drift/$dU*4)]";
analyze 1;
}
}
}

puts "Completed: Cyclic Push"

[nmp]

Dear all...

I’m having similar problems with concrete 06. In cyclic behavior, particularly after an unloading branch, the monitored fiber defined in the model presents positive strains and negative stress. I defined my parameters in SI units. Do you think that this could be the problem and i need to adjust? i made stress strain analysis with a monotonic profile and the analytical response matched the theoretical one. Changing it by introducing an unloading/reloading point, the response DIVERGES. did you found a solution to the model problem? if yes can you help me with that?

[civilengr_tahir]

dear nmp
I think there is nothing wrong with defining units in SI. I have contacted Dr.Massone regarding this issue here is his response about cyclic analysis by using displacement basedInt element
"The model has been validated to monotonic response so far. Cyclic analysis requires defining damage location and progression (for further discussion you can look into Vecchio and others publications were cyclic response for panels is described).Thus, even though cyclic analysis can be done (because materials are cyclic) the problem need to be studied further, and the actual formulation is not meant yet for that."
So it is clear that there is something wrong with element formulation for cyclic analysis. Have u seen its formulation.They have used NewtonRaphson alogrithim.I think problem can be solved if we change the NewtonRaphson Alogrithim to KreloveNewton Alogrithim.
If you have further information please share with us.

[ksecond]

Dear civilengr_tahir,
I have some problem with using the Flexure-Shear Interaction Displacement-Based Beam-Column Element, it failed to get convergence even applied the axial load on it. I have no idea of what's wrong with the model for I have checked it many times. Could I send you my tcl script and have a look on it? Thanks!

[civilengr_tahir]

hi ksecond ..yeah sure send it,I cant comment without looking at tcl script

[civilengr_tahir]

hi Ksecond r u doining cyclic analysis or monotonic?

[ksecond]

In fact, I'm doing the cylic analysis. But it failed to get convergence before the cylic analysis, in other words, when I applied axial load on it, the program displayed "it failed to get convergence.....". I've checked the data time after time, but did not find some relevant errors. Can you just leave me your email that I can send you my program? Thanks for attention!

[civilengr_tahir]

Hi..This element is not for cyclic analysis. Element formulation is not meant for that yet. Anywas u can send me your script at civilengr_tahir@hotmail.com. I will recommend you to read the Peer report about the Modeling of shear walls considering flexure shear interaction by Wallace and L.Massone. hope so this will help

[ksecond]

Do you mean the paper "Shear-Flexure Interaction for structural walls" by L.M. Massone, K. Orakcal, and J.W. Wallace? Prof.Wallace sent me that file several days ago.Or could you send me that PEER report to my email of chuangkunlau@gmail.com ? Is it "Analytical Modeling of Reinforced Concrete Walls for Predicting Flexural and Coupled-Shear-Flexural Responses"? Thanks~