Vous êtes sur la page 1sur 10

ATC & SEI 2015 585

Seismic Response of a Steel SMF Building: Comparison between Conventional


Design and Damper Options

H. Kit Miyamoto1 and Amir S. J. Gilani2


1
President, Miyamoto International, Inc., 1450 Halyard Dr., West Sacramento, CA
Downloaded from ascelibrary.org by New York University on 02/19/16. Copyright ASCE. For personal use only; all rights reserved.

95691. E-mail: kit@miyamotointernational.com


2
Manager Earthquake Engineering, Miyamoto International, Inc., 1450 Halyard Dr.,
West Sacramento, CA 95691. E-mail: agilani@miyamotointernational.com

Abstract

Performance based design with seismic protection devices such as viscous dampers
have fundamentally altered the landscape of earthquake engineering and design.
Structures designed and built without such devices typically use a code-prescribed
design that implies extensive structural damage, loss of operation, and likely
replacement at design-level earthquake. In contrast, performance based design
incorporating earthquake protection devices leads to a combination of best
engineering practice and reducing life-cycle costs. These devices are robust, cost-
effective, and have a proven exceptional performance record in past earthquakes. In
most cases, initial cost of their utilization is neutralized by reduction in cost of other
structural members. The long-term performance is the key parameter used for
evaluation. Structures properly designed with these devices will likely only require
minimum post-earthquake inspection and can be fully operational within hours of a
seismic event. When utilized for critical structures, such performance reduces the
need for use of natural resources by eliminating post-earthquake repair or
reconstruction and thus improving the community resiliency. Example cases are
presented.

INTRODUCTION

Fluid viscous dampers (FVDs) were originally developed as shock absorbers for
the defense and aerospace industries. In re-cent years, they have been used
extensively for seismic application for both new and retrofit construction. During
seismic events, the devices become active and the seismic input energy is used to heat
the fluid and is thusly dissipated. Subsequent to installation, the dampers require
minimal maintenance. They have been shown to possess stable and dependable
properties for design earthquakes.

FVDs consist of a cylinder and a stainless steel piston. The cylinder is filled with
incompressible silicone fluid. The damper is activated by the flow of silicone fluid
between chambers at opposite ends of the unit, through small orifices. Figure 2 shows
the damper cross section.

ASCE and ATC 2015

Improving the Seismic Performance of Existing Buildings and Other Structures 2015
ATC & SEI 2015 586
Downloaded from ascelibrary.org by New York University on 02/19/16. Copyright ASCE. For personal use only; all rights reserved.

Fiigure 1. Steel SMF


F with damp
pers Figu
ure 2. D
Damper crooss section

EEL MOM
STE MENT FRAM
ME APPLIC
CATION

Ovverview. Provisions of
o ASCE/SE EI 7-10 (AS SCE 2010) were used tto design a
new
w steel frammed multi-sto g in the Los Angeles arrea. The steeel members
ory building
werre sized usin
ng conventio onal code dessign proceduures. Dampeers were sized to control
the story driftss. The dam mpers were placed
p only at the grouund floor w with pinned
coluumn bases where
w the maximum
m veelocity is exxpected to occcur. A paraallel design
wass carried outt using the co
onventional design methhodology. Thhis model was designed
folllowing the conventional code proced dure for bothh strength annd drift.

The four-sttory commerrcial buildin ng is 18.5 m tall and haas a total flooor space of
8,000 m2. The seismic masss of the buiilding was aapproximately 9 MN. A Architectural
ren
ndering of thee building iss presented in
n Figure 3. FFor the dampped model, tthe bases of
all columns were
w modeled d as pinnedd. For conveentional dessign model, the fixity,
ovided by th
pro he grade beaams, was assumed at tthe base of all columns. Figure 4
deppicts the matthematical model
m of the building.
b Sixxteen nonlinnear FVDs wwere used.

Fiigure 3. Architectu
ural renditio
on Figu
ure 4. M
Mathematiccal model

Nonlinear response hiistory analysis was perrformed to evaluate peerformance.


Commponents of o the groun nd motion were alignned with buuilding prinncipal axes.
Maaximum resp ponse quantitties, such as,, building flooor displacem
ment and acccelerations,
storry shears, FV
VD forces, and
a member stresses, weere extractedd for evaluatiion.

Analysis resultts. The maximum computed


c sttory drift raatio was appproximately
1.5% for both models. Baase fixity and
a larger m
member sizees control ddrift for the

ASCE and ATC 2015

Improving the Seismic Performance of Existing Buildings and Other Structures 2015
ATC & SEI 2015 587

con
nventional model.
m FVDss provide succh control fo
for the dampped model. T The damped
model has smaaller base shear and floo
or acceleratioons; see Figgure 5; becauuse it has a
lon
nger period and
a larger efffective damp
ping ratio.

100 3

.
2
50

Roof acceleration, g
1
Cs %g

0 0
Downloaded from ascelibrary.org by New York University on 02/19/16. Copyright ASCE. For personal use only; all rights reserved.

Cs,

-1
-50
-2

-100 -3
0 5 10 15 20 25 0 5 10 15 20 25
Time, sec Time, sec

Base shear coefficcients Roof aaccelerations


Figure
F 5. Responsee histories (ssolid: damp
per model, d
dashed: codee design)

Figure 6 shows
s the snap
s shot ofo the convventional annd damped models at
max ximum defo ormation fo or a spectraally-matchedd event. Booth models meet their
perrformance goal
g of life safety. How wever, the damped moodel meets the higher
essentially elasstic performaance goal; th
he columns oof the dampeer model rem main elastic
d, as listed in
and n Table 1, the plastic rotaations are sm
maller for thee damped m
model.

Conv
ventional dessign Model w
with damperss
Fig
gure 6. MC
CE plastic h
hinge rotatioons

Tab
ble 1. Maximum MCE
M plastiic hinge rotaations, % raadian
Co onventionall Damped
Beam
B 1.7 1.3
Co
olumn 2.6 --

ASCE and ATC 2015

Improving the Seismic Performance of Existing Buildings and Other Structures 2015
ATC & SEI 2015 588

Figure 7 presents the damper hysteresis loop and the components of seismic
energy computed from analysis. In the absence of dampers, yielding in ductile beam
members would substitute for such energy dissipation.

2,000 3,000
Input
Inherent
1,000
Axial force, kN

Damper
2,000

Energy, kJ
0
Downloaded from ascelibrary.org by New York University on 02/19/16. Copyright ASCE. For personal use only; all rights reserved.

1,000
-1,000

-2,000
-
-120 -60 0 60 120
0 5 10 15 20 25
Axial displacement, mm Time, sec

FVD hysteretic behaviour Components of seismic energy


Figure 7. Energy dissipated by dampers

Cost-Benefit Analysis. The damped structure has superior long-term


performance and lower maintenance costs. Following a design earthquake, the
conventional building should provide life safety, but will sustain significant damage
and could require replacement. The additional initial cost of the dampers is offset by
the savings in steel tonnage and foundation concrete volume. Sample data is
presented in Table 2.

Table 2. Cost comparison for conventional and damped structures


Item Conventional Damped Differential cost
Moment Frames 274 Ton 223 Ton - $150,000

Foundation RC grade beams No grade beams - $200,000


Dampers None $200,000 + $200,000
Net -$150,000

MODELING OF VISCOUS DAMPERS

In most applications, the dampers are modeled as simple Maxwell model of


Figure 8. The viscous damper itself is modeled as a dashpot in series with the elastic
driver brace member. Such model is adequate for most design applications, but is not
adequate for large earthquake analysis

Damper limit states are governed by a few elements. The dampers bottoms out,
once the piston motion reaches its available stroke. This is the stroke limit and results
in transition from viscous damper to a steel brace with stiffness equal to that of the
cylinder wall. The force limit states in compression and tension are governed by the
buckling capacity of the driver brace and the tensile capacity of the piston rod,
respectively. Figure 9 presents the proposed limit-state model for viscous dampers.

ASCE and ATC 2015

Improving the Seismic Performance of Existing Buildings and Other Structures 2015
ATC & SEI 2015 589

Thiis model is developed


d too incorporate the pertineent limit stattes and conssists of five
com
mponents. TheT damper components
c are modeleed: a) the drriver used too attach the
dam
mper to the beams
b and columns
c is modeled
m as a nonlinear sspring, b) thee piston rod
and
d undercut iss modeled ass a nonlinearr spring. In ttension, the undercut section of the
pistton can yieldd and fractu
ure, c) dashp pot is used too model the viscous com mponent, d)
gap
p element an nd linear spriings, e) hookk elements aand linear sprrings are useed to model
the limit state when
w the pistton extensio on reaches thhe damper strroke.
Linear
Downloaded from ascelibrary.org by New York University on 02/19/16. Copyright ASCE. For personal use only; all rights reserved.

Nonlinerar Noninea
ar C&

Linear

Figure 8. Maxweell model Figgure 9. Limit statee model

CO
OLLAPSE ANALYSIS
A OF BUILD
DINGS WIT
TH DAMPE
ERS

Ground Motio ons The input historries used inn analysis w were based on the two
commponents off the 22 far-filed (measu ured 10 km or more froom fault ruppture) NGA
PEE ER (2009b)) records. These
T 44 reecords have been identtified by FE EMA P695
(FE
EMA 2009) for collapse evaluation analysis.
a Thee selected 222 records coorrespond to
a reelatively larg
ge sample off strong reco
orded motionns that are cconsistent wiith the code
(ASSCE/SEI 7) and are stru ucture-type and site-hazzard indepenndent. For aanalysis, the
44 records weree normalized d to remove the record-tto-record varriation in inttensity.

SRRF Connectio on. Steell SRFs with h reduced beeam sectionss (RBSs) aree one of the
prequalified co onnections fo or seismic ap pplications aand were useed in this annalysis. The
connstitutive po ost-yield relation for th he RBS plaastic hinges developed by Lignos
(20
008) was useed in this su ubject study. Those authhors used exxperimental ddata from a
database of 42 RBS connecctions tested d in laboratoories using reegression annalysis; they
idenntified the plastic hing ge propertiees as a funnction of flaange slendeerness, web
slennderness, laateral bracinng, and yield strength of beams. T The momennt-rotational
deffinitions, thee multilinearr moment-rotation constiitutive relatiion for the R RBS plastic
hin
nges was thu us defined. FEMA
F 350 (FEMA 20000) recommeends the inttroduction a
red
duction in th he flexural stiffness to account fo r the effectt of the redduced beam
flan
nges. Such reduction
r wiill result in an
a increase in the storyy drifts by 3% % to 7% in
typical applicattions. FEMA A 350 also recommends
r s increasing story drifts by 10% to
acccount conserrvatively fo or this effecct. This appproach was used to sccale up the
commputed inelaastic drifts:

Moodel Propertties and IDAA Program OpenSees ((PEER 20099a) was usedd to conduct
the nonlinear analyses
a described in th
his paper. Peertinent moddel propertiees are listed
herre.

ASCE and ATC 2015

Improving the Seismic Performance of Existing Buildings and Other Structures 2015
ATC & SEI 2015 590

Analytical models are two-dimensional


Material nonlinearity is represented by concentrated plastic hinges represented
by RBS hinges.
The damper element is represented by the refined model including the limit
states.

For collapse analysis, the normalized records are then scaled upward or
downward to obtain data points for the nonlinear incremental dynamic analysis (IDA)
Downloaded from ascelibrary.org by New York University on 02/19/16. Copyright ASCE. For personal use only; all rights reserved.

simulations (Vamvatsikos and Cornell, 2004).

APPLICATION TO STEEL BUILDINGS

To illustrate the concepts described in this paper, design and analysis of a group
of archetypes with viscous damping was conducted. Fifteen archetypes are currently
under consideration. The basic geometry and distribution of dampers for these models
are summarized in Table 3. The selected building models will be regular in plan and
elevation with a dominant first mode response. The period of tall buildings is limited
to approximately 5 sec to ensure sufficient energy is present in the input histories.
The investigations for all but the 20- and 30-story structures have been completed.
The frames were designed using the code provisions and special requirements for
SMRFs. The ASCE 7 maximum period used to compute base shear period is used for
evaluation. A typical 5-story archetype is shown in Figure 10.

Table 3. Archetypes analyzed in this study


Archetype Stories Column base Drift Ratio Damper FS
O1 1 Pinned 2.5% 1.0
O2 1 Pinned 1.0% 1.3
O3 1 Fixed 2.5% 1.0
O4 1 Fixed 1.0% 1.3
A1 2 Pinned 2.5% 1.0
A2 2 Pinned 1.0% 1.3
A3 2 Fixed 2.5% 1.0
A4 2 Fixed 1.0% 1.3
B1 5 Fixed 2.0% 1.0
B2 5 Fixed 1.0% 1.3
C1 10 Fixed 2.0% 1.0
C2 10 Fixed 1.0% 1.3
D1 20 Fixed 2.0% 1.0
D2 20 Fixed 1.0% 1.3
E1 30 Fixed 1.0% 1.0

ASCE and ATC 2015

Improving the Seismic Performance of Existing Buildings and Other Structures 2015
ATC & SEI 2015 591
Downloaded from ascelibrary.org by New York University on 02/19/16. Copyright ASCE. For personal use only; all rights reserved.

Figu
ure 10. Five-story
F arrchetype B11

Damper propeerty selectio on


Following the
t design of o moment frames
f accorrding to ASC CE/SEI 7 reequirements
for strength, dampers
d werre sized to limit
l story drift ratios. ASCE/SEI 7 presents
recommendatio ons for the design
d of daampers. Thiss approach w was used to provide an
appproximate damper size, assuming stiffness prroportional tto damping. However,
beccause dampeers had a velocity coefficcient ( of 00.5 and becauuse they didd not extend
throoughout the full buildin
ng height, thhe damper cconstant waas then compputed more
acccurately by conducting nonlinear analysis
a at tthe design eearthquake (DE) level.
Thrree sets of spectrum-co ompatible records
r that matched thhe DE specctrum were
devveloped. Noorthridge, Kobe,
K and Newhall
N recoords were uused from the Pacific
Earrthquake En ngineering Research
R Center
C (PEEER) Next G Generation A Attenuation
(NGGA) databasse (PEER 20 009b). Addittionally, thee Kobe recorrd was scaleed such that
the ordinate off its responsse spectrum
m matched thhe DE specctrum at thee buildings
ndamental peeriod
fun

NALYSIS RESULTS
AN R

The analysiis results forr the five-stoory archetyppes are preseented in Figgure 11. For
the pushover curves,
c the solid and .dashed
. linees corresponnd to the cases where
dammper are exccluded and included,
i resspectively, iin analysis. As long as the damper
doees not botto om out, the plots are identical.
i Onnce the dam mper bottom ms, there is
signnificant incrrease in stiffn
fness and strength since a stiff bracee (cylinder wwall) is now
addded to the system. After A the damper
d faills, the dam mped pushoover curve
asy
ymptotically approachess the undam mmed case. The dotted line corressponds to a
biliinear approx ximation ussed to com mpute the yyield and ultimate driffts and the
corrresponding ductility (cc). The comp puted system m ductility w
was 8.0 and rresulted in a
SSF F of 1.34.

ASCE and ATC 2015

Improving the Seismic Performance of Existing Buildings and Other Structures 2015
ATC & SEI 2015 592

For the IDA A plots, the solid and daashed red linnes correspoond to the M MCE (SMT)
andd the median n collapse caapacity (SCT T), respectively. Note thhat the additiion of small
dammper factor ofo safety sign ncreases collaapse marginn. For the fraagility plots,
nificantly in
the 44 collapsee data are sttatiscally org ganized andd a lognormaal curve is ffilled to the
data (dashed lin nes in the fiigures). The plot was thhen rotated tto corresponnd to a total
unccertainty of 0.55 (solid line) per FEMA P695 . Finally thee curve wass shifted to
acccount for thee effect of thhe SSF (dark k solid liness in the figurres). The proobability of
collapse at MC w then be computed. T
CE intensity was The probability of collappse at MCE
level was reducced by a facttor of approx ximately 4 wwhen an addiitional dampper factor of
Downloaded from ascelibrary.org by New York University on 02/19/16. Copyright ASCE. For personal use only; all rights reserved.

fety of 30% is
safe i included. The
T probabiility of the d amper reachhing its limitt state at the
MCCE intensity can then bee computed from the daamper fragillity plots. Note that the
pro
obability of damper
d reacching a limitt state is siggnificantly reeduced whenn a damper
facttor of safety
y of 30% is inncluded in design.
d

Table 4 suummarizes th he analysis results. Thee collapse mmargin ratioo (CMR) is


deffined as the ratio
r of SCTT and SMT. The adjustedd collapse m margin ratio (ACMR) is
thenn computed as the prod duct of SSF and CMR. FEMA P695 specifies a minimum
ACCMR of 1.59 9 for acceptaable performaance. Both aarchetypes hhave significantly larger
collapse margin ns and thereffore pass eassily.

Tab
ble 4. Collapse fra
agility data
M
MCE probabbility
Daamper
Arch.
A SCTT SMT CMR SSF ACMR P P/F Collappse
cap
pacity
B1 1.24 0.82 1.51 1.34 2.20 P
Pass 8.0% 22%
B2 1.81 0.82 2.25 1.34 3.10 P
Pass 2.0% 10%

Static pushover
p curvves

ASCE and ATC 2015

Improving the Seismic Performance of Existing Buildings and Other Structures 2015
ATC & SEI 2015 593
Downloaded from ascelibrary.org by New York University on 02/19/16. Copyright ASCE. For personal use only; all rights reserved.

Static pushover
p curvves

Fraagility plots
Figure 11. Analysiss results

ONCLUSIONS
CO

New steel buildings


b were designed d using perfformance baased engineeering (PBE)
and
d provisions of ASCE 7. SMRFs werre used to prrovide strenggth; damperss were used
to control story drifts. PBBE design using
u dampeers is superiior to the coonventional
dessign. The demand on bo oth structurall and nonstrructural commponents is rreduced. To
date, a model of
o viscous daampers with h limit statess has been foormulated thhat includes
mper limit states.
dam s Current research h using IDA A and limitt states of dampers is
currrently undeerway. The outcome of o this studdy will proovide a moore realistic
assessment of the
t performance of mom ment framess with damppers. All the archetypes
had
d significant margin against collapsee and thus h ad satisfactoory performaance. When
a damper factorr of safety iss included in
n design, addditional proteection for thhe structures
and
d dampers iss provided. As A one of th he research deliverabless, pertinent iinformation
willl be provided for the dessigners to assist in seism
mic design ussing this appproach

ASCE and ATC 2015

Improving the Seismic Performance of Existing Buildings and Other Structures 2015
ATC & SEI 2015 594

REFERENCES

ASCE (2005), ASCE 7-05: Minimum design load for buildings and other
structures, American Society of Civil Engineers, Reston, VA
FEMA (2000), FEMA 350: Recommended Seismic Design Criteria for New Steel
Moment Frame Buildings, Federal Emergency Management Agency,
Washington DC.
Liel, A.B.., and Deierlein G.G., (2008), Assessing The Collapse Risk Of Californias
Existing Reinforced Concrete Frame, Structures: Metrics For Seismic Safety
Downloaded from ascelibrary.org by New York University on 02/19/16. Copyright ASCE. For personal use only; all rights reserved.

Decisions, John A. Blume Earthquake Engineering Center Report No. TR


166, Department of Civil Engineering, Stanford University.
Lignos, D. G.(2008), Sidesway Collapse Of Deteriorating Structural Systems Under
Seismic Excitation, PhD dissertation, Department of Civil and Environmental
Engineering, Stanford University.
Miyamoto, H.K., and Gilani, A.SJ. (2008), Design of a new steel-framed building
using ASCE 7 damper provisions, ASCE Structures Congress, Vancouver,
BC, SEI institute.
NEHRP (2009), ATC 63, FEMA P695: Quantification of Building Seismic
Performance Factors, Federal Emergency Management Agency,
Washington, D.C.
PEER (2009a), Open System for Earthquake Engineering Simulation (OpenSees),
McKenna, F., Fenves, G., et al, Pacific Earthquake Engineering Research
center, University of California, Berkeley. Berkeley, CA.
PEER (2009b), PEER NGA, Records Pacific Earthquake Engineering Research
center, University of California, Berkeley. Berkeley, CA.
Taylor (2009), Personal Communications
Vamvatsikos, D. and Cornell, A.C. (2004) Applied Incremental Dynamic Analysis,
Earthquake Spectra, Volume 20, No. 2, pages 523553, Earthquake
Engineering Research Institute, Oakland, CA

ASCE and ATC 2015

Improving the Seismic Performance of Existing Buildings and Other Structures 2015