Innovative strategies for seismic retrofitting of steel and composite structures

Innovative strategies for seismicretrofitting of steel andcomposite structuresL Di Sarno1 and A S Elnashai2

1 University of Sannio, Benevento, Italy2 University of Illinois, Urbana USA

SummaryThis paper reviews briefly traditionalrehabilitation methods and provides a detaileddiscussion of design issues along with theadvantages and the disadvantages of innovativestrategies for seismic retrofitting of steel andcomposite structures. Special emphasis is placedon base isolation and supplemental damping

devices. The viability and cost-effectiveness ofthese strategies are assessed on the basis ofmultiple limit states within the framework ofperformance-based design. A number ofparameters which govern the selection of theintervention devices are investigated.

Key words: retrofitting; steel structures; composite structures; base isolation; dampers

Prog. Struct. Engng Mater. (in press)

Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/pse.195

Introduction

Recent earthquakes, such as the Northridge (January17, 1994, California, USA), Hyogoken-Nanbu (January17, 1995, Japan) and Chi-Chi (September 21, 1999,Taiwan) caused unexpectedly high levels of damageof steel and composite structures. Damage was variedand widespread, ranging from residential buildings tohighway bridges and lifelines[1]. In California themost severe effects were imposed on low-rise steelbuildings, typically three-storey apartments with aground level garage, second-storey living quartersand third-storey bedrooms[2]. In Japan, columns,braces, column bases and beam-to-columnconnections of old and new steel buildings sustaineddamage[3], and 1067 old buildings}constructed with

bundled light-gauged sections for columns andtrusses for beams}collapsed or were damagedbeyond repair[4]. A further 100 new structures,consisting of two to five storeys, collapsed, and 332were severely damaged. In Taichung, (Taiwan) somehigh-rise dual systems and welded MRFs sufferedlocalized yielding and damage to architecturalfeatures[5].

Observation of these earthquake effects prompted anumber of research projects in seismic retrofitting(see, for example Di Sarno & Elnashai[6]). The workcarried out has shown that structural deficiencies inexisting steel and composite buildings may beremedied by the application of several retrofittingstrategies, either traditional or innovative. While localand global interventions are effective for enhancing

Review Article

Terminology

A b = bonded areaBD = damping-dependent coefficient at

design displacementBD = damping-dependent coefficient at

maximum displacementC = damper coefficientD = displacementEloop = energy dissipated for cycle of

loadingED= energy dissipated for cycle at de-

sign displacementEM = energy dissipated for cycle at

maximum displacement

F = forceFy = damper yield forceFE = elastic force in the structureg = acceleration due to gravityG0 = shear storage modulusG00 = shear loss moduluskeff = effective stiffnessk0eff = effective stiffness at peak displace-

mentN = brace forceNb = brace buckling forceNg = brace tensile force at slipNl = damper force at slip

SD = design 5% damped spectral accel-eration for 475-year return periodearthquake

SM = design 5% damped spectral accel-eration for 2475-year return per-iod earthquake

R = force reduction factorTD = effective period of a seismically

isolated structure at designdisplacement

TM = effective period of a seismicallyisolated structure at maximumdisplacement

V = base shear

Vb = design base shearW = seismic weight of the structureb = equivalent viscous dampingbeff = effective equivalent viscous damp-

ingbD = effective equivalent viscous damping

at design displacementbM = effective equivalent viscous damp-

ing at maximum displacementD = displacement;Dp = peak displacementD’ = velocityo = natural circular frequency.

Copyright & 2005 John Wiley & Sons, Ltd. Prog. Struct. Engng Mater. (in press)

the energy dissipation capacity of framed structures,novel materials (e.g. special metals) and technologies(such as base isolation and supplemental damping)are also available for structural retrofitting. This paperreviews briefly conventional rehabilitation methodsand discusses in details non-conventional strategiesfor steel and composite steel and concrete frames.

Overview of traditional strategies

Strategies for seismic retrofitting of steel andcomposite structures are generally based onstrengthening or weakening of the existing structureor its components. Rehabilitation interventions areaimed at modifying structural strength, stiffness,ductility and, consequently or additionally, dampingto improve seismic performance. Two approachesmay be followed[7–9]: local modification of structuralcomponents and connections and global modificationof the structural system. Requirements for thefulfilment of these modifications are detailed belowfor each structural component and for the framesystem as a whole.

Effective local rehabilitation approaches consistof interventions (Fig. 1) aimed at providing: (i) baseand parent materials with adequate mechanicalproperties, (ii) sections and members with sufficientductility and (iii) connections with adequatestiffness, strength and ductility. It is important toavoid premature local and/or global (latero-torsional)

buckling which can significantly reduce theenergy dissipation capacity of steel structuralmembers. However, based on the nature and extentof damage, several alternative local approaches torepair are available and may be summarized asfollows[9]:

* replacement of damaged portions of base metal,e.g. beam-column sections;

* replacement of damaged connection elements;* replacement of connection welds;* repairs to portions of any of the aforementioned

components.

Effective global rehabilitation approaches (Fig. 1)provide lateral force resisting systems with enhancedductility. The improvement of seismic performance maybe achieved in different ways, depending on thestructural deficiencies. For example, structures withirregularities in plan and/or elevation should bemodified in order to regularize the system becauseregular frames exhibit enhanced seismic performance.Mass reduction or lessening is benign because of thelower inertial forces attracted during earthquake loading.

Increasing global stiffness, strength and dampinghelps to reduce storey drifts and shears. Higher globaldamping enhances the energy dissipation capacity ofthe structure. As a result, structural damage isminimized and important activities}in the case ofoffice buildings, the business hosted in the structure–are not interrupted; thus, reducing economic lossesresulting from earthquakes.

Fig. 1 Retrofitting strategies for steel and composite buildings

EARTHQUAKE ENGINEERING AND STRUCTURAL DYNAMICS


Enhanced global ductility may be achieved bydirecting inelastic deformation towards dissipativezones and ensuring that the rest of the structure behavesalmost linearly. In such cases it is essential to secure:

* regularity of mass;* regularity of stiffness distribution;* regularity of strength distribution;* continuity and redundancy between members.

Retrofitting of steel and composite structures mayalso be effected by modifying one of the mechanicalparameters characterizing the dynamic response,namely mass distribution, stiffness distribution,strength distribution and structural damping.

Innovative strategies: vibration control

Recent developments have shown that the structuralcontrol of vibrations may be achieved by modifyingmasses, stiffness and damping, and allowing passiveor active counter forces to be generated[10]. Innovativeprotection strategies have been rapidly adopted inAsia, Europe and the US, where several approacheshave been assessed, optimized and used successfullyin practical applications[11]. A classification of themost common control systems is provided in Fig. 2.Four groups are identified, (i) passive, (ii) active, (iii)hybrid, and (iv) semi-active control systems, incompliance with the basic principles behind thecontrol strategies[12]. Relevant assumptions andspecific mechanisms characterizing each category(Fig. 3) are summarized hereafter.

ACTIVE CONTROL SYSTEMS

Such systems possess external sources poweringcontrol actuator(s) that apply forces to the structurein a predefined manner. These forces can be used bothto add and dissipate energy in the structure. In anactive feedback control system, the signals sent to thecontrol actuators depend upon the dynamic responseof the system that is measured through physicalsensors, i.e., optical, mechanical, electrical or chemicalsensors.

PASSIVE CONTROL SYSTEMS

Unlike devices for active control, these systems do notrequire any external power source; the vibrationcauses them to impart forces which protect thestructure. The energy in passively controlledstructural systems cannot be increased by passivecontrol devices; in contrast, high damping levels maybe beneficially achieved in the main structure.

HYBRID CONTROL SYSTEMS

Hybrid systems combine both active and passivecontrol approaches. To maximize the system

efficiency, it is common practice to employ activedevices that may redress drawbacks exhibited bycertain passive dampers, and vice versa.

SEMI-ACTIVE CONTROL SYSTEMS

Similar in principle to active control, these systemsrequire lower external energy sources. Typically, theydo not add energy to the structural system. Therefore,bounded-input as well as bounded-output stability isguaranteed. Semi-active control devices are oftenviewed as controllable passive devices.

Detailed descriptions and applications of variousvibration control devices may be found in Soong &Costantinou[13], Housner et al.[12]; Soong & Dargush[14]

and Naeim & Kelly[15]. The present paper focuses onlyon passive devices, such as base isolation andsupplemental damping devices for seismic retrofittingof steel and composite structures.

Base isolation and supplementaldamping devices

The effectiveness of seismically isolated structures orstructures with supplemental damping has beenrecently codified in international seismic regulationsand recommendations[9,16,17]. In Europe, provisionsonly for base isolated systems have been suggested. Itis recognized that passive energy dissipation devicescan absorb a portion of the earthquake-inducedenergy in structures and reduce the demand onprimary structural members such as beams, columns,beam–column joints, and walls; thus, structural safetymay be guaranteed. The conceptual design of seismicisolation and systems with energy dissipation isoutlined in Fig. 4; their range of application anddegree of maturity is compared with other strategiesof seismic control in Table 1.

Seismic isolation and supplemental dampingsystems are viable retrofitting strategies to enhanceearthquake performance in building structures and/or whenever owners can afford the costs of design,fabrication, and installation of these special devices.These costs are offset by the reduced need forstiffening and strengthening. The efficacy of suchstrategies is summarized in Table 2; US provisions[18]

are contrasted to a proposal for European standards,the latter are compliant with limit states adopted inEC8 draft for retrofitting[12].

Base isolation is a viable strategy for retrofittingsteel and composite buildings because of its (i)functionality; (ii) contents protection; (iii) investmentprotection and (iv) construction economy. Compositestructures are generally used for buildings incommercial and financial areas which containsensitive and valuable equipment vital for businessand emergency use; their disruption after anearthquake can have a devastating socioeconomic

SEISMIC RETROFITTING OF STEEL


impact. Fixed-base buildings may, in fact, undergolarge storey drifts (flexible structures) or high flooraccelerations (rigid structures) causing structuraland/or non-structural damage. In these cases,retrofitting via seismic isolation is a cost-effectiveoption[20]; drifts and accelerations may be reduced bya factor of 2–6.

Buildings retrofitted with seismic isolation systemsconsist of three distinct parts: the structure above theisolation system, the isolation system itself and thefoundation and other structural elements below theisolation system. Each should be properly assessedand detailed to ensure the effectiveness of thisrehabilitation strategy. Transient design situations

Fig. 2 Common vibration control systems



such as lifting the superstructure, cutting structuralelements, placing the isolators and giving back theload to the columns, should be adequately checked atthe design stage. Isolators are generally located insub-basements, at the top of basement columns, or atthe bottom or top of first-storey columns. Therefore,the working site is limited to garages or warehouseswith no interruption to activities within the buildingand no damage to finishes and equipment.

The choice of bearing location within a building isnot straightforward[21]; the advantages and disadvan-tages of bearing locations are outlined in Table 3.

From a mechanical viewpoint, the use of isolationdevices in seismic applications, particularly forretrofitting, relies upon three fundamental mechanicalproperties: (i) horizontal flexibility to increasestructural period and reduce the transfer of seismicenergy to the superstructure (except for very softsites); (ii) energy dissipation (damping) to reducedisplacements; and (iii) sufficient stiffness at smalldisplacements to provide adequate rigidity for servicelevel environmental loads[22]. Isolators should exhibitsignificant energy dissipation capacity and/or re-centring capability[23]. This target can be achievedeither by choosing devices with intrinsic dissipativeand re-centring capacities or by providing ad hocadditional elements.

Seismic isolators are divided into two groups,according to the type of energy dissipationmechanism (Fig. 5):

* Elastomeric isolators: usually made of layers ofrubber separated by steel shims, which constrainlateral deformation of the rubber under verticalloads. Elastomers manufactured with special fillersensure elevated hysteretic energy dissipation.Elastomeric devices are high-damping rubberbearings (HDRBs), low-damping rubber bearings(LDRBs) or low-damping rubber bearings with alead core (LDRB-LC). These devices have beenused worldwide in new and existing structuresbecause of their high efficiency[13].

* Sliding isolators: their main source of energydissipation is friction. They have twodisadvantages: friction is usually difficult toquantify and permanent offset between the slidingparts may occur after an earthquake (non re-centringsystems)[15]. Sliding devices are usually flatassemblies, such as sliding poly-tetrafluoro-ethylene (PTFE), or have a curved surface, i.e. afriction–pendulum system (FPS). The latterovercomes the above disadvantages by employinga curved rather than a flat surface. Rolling isolatorsconsist of flat assemblies or have a curved orconical surface. A typical rolling system is the balland a cone system.

The choice of isolators should comply with specificacceptance criteria. In particular, it should remainstable for design displacements and provideincreasing resistance with increasing displacement(no degradation) under repeated cyclic load.

Fig. 3 Application of control schemes to mitigate structural vibrations[10]. PED¼ passive energy dissipation



Advantages and disadvantages of common baseisolators are outlined in Table 4.

The mechanical response of isolation systems maybe characterized via four types of relationships (asshown in Fig. 6) in which idealized curves refer to thesame design displacement DD:

* Linear system: the effective period does not changewith earthquake loading. Forces applied to thesuperstructure are proportional to thedisplacement across the isolation device. Somewind-restraining is required.

* Hardening system: the stiffness is initially lower(long effective period) and increases (short effective

period) according to earthquake loading. The super-structure is subjected to higher forces than in alinear system. However, the displacements of theisolation system are smaller. A wind-restrainingmechanism is required.

* Softening system: the stiffness is initially higher(short effective period) and decreases (long effectiveperiod) as earthquake loading increases. The super-structure is subjected to lower forces than in alinear system; however, the displacements of theisolation system are higher.

* Sliding system: the effective period increasesaccording to earthquake loading. The super-structure is subjected to constant loads.

Fig. 4 Conceptual design for base isolation and supplemental damping systems



Wind-restraining mechanisms are required for lowfriction values.

Isolated buildings are designed using eitherstatic analysis (equivalent lateral force procedure) ordynamic analysis (dynamic lateral response procedure);however, the latter is usually preferred because thelateral force design has several limitations whichdepend upon the seismic input (site location withrespect to active faults and local soil conditions) andthe structural system (geometrical and mechanicalproperties of the building as well as the isolationsystem). Dynamic analyses are carried out usingeither response spectra or time-history analyses.International standards and recommendedprovisions[17,24,25] and relevant commentaries[26] maybe used as reference for the formulation of theaforementioned methods of analysis. The use of UBC1997 should be avoided because it is unnecessarilycomplicated and conservative[15].

Minimum design displacement and/or lateralforces of seismically isolated structures arebased on the deformation characteristics of theisolation system, which can be represented

either by an equivalent linearly elastic model orvia nonlinear models (nonlinear time-historyanalyses). The use of equivalent linear modelsis advised at this stage because they are morereliable.

Design parameters of linearly elastic isolationsystems include effective stiffness keff anddamping beff. These mechanical parameters maybe determined via force–deflection curvesobtained from hysteresis loops of device prototypesin cyclic load tests (Fig. 7). The effective stiffnessis as follows:

keff ¼Fþj j þ F�j jDþ�� þ D�j j

(1)

where Fþ and F� are the positive and negative forcesat D+ and D�, respectively.

The effective damping is computed via therelationship:

beff ¼2

pEloop

k0eff Dþp�� þ D�p

�� 2

264

375 (2)

.......................................................................................................................................................

.......................................................................................................................................................

Table 1 Range of application and maturity of seismic control of structures.

Type of control Range of application Degree of maturity

Seismic isolation Low-to-medium rise buildings (eithernew or existing)

Mature technique

Bridges and subways Many theoretical and practical resultsEquipment or facilities Many applications world-wide

Energy dissipation Medium-to-high rise buildings (eithernew or existing)

Mature technique

Towers, stacks and chimneys Many theoretical and practical resultsMedium-to-long span bridges Many applications world-wideLifelines

Passive control Medium-to-high rise buildings Relatively mature techniqueTowers, stacks and chimneys Several theoretical and practical resultsMedium-to-long span bridges Several applications world-wideLifelines

Active, semi-active and hybrid control High rise buildings Ongoing research stageTowers, stacks and chimneys Several theoretical resultsMedium-to-long span bridges Few applications world-wide

...........................................................................................

...........................................................................................

...........................................................................................

............................................................................................

...........................................................................................

...........................................................................................

Table 2 Efficacy of isolation system and energy dissipation for retrofitting: FEMA274[18] (A) and proposed scheme (B).

A. FEMA 1997

Performance level Isolation system Energy dissipation

Operational Very likely LimitedImmediate occupancy Likely LikelyLife safety Limited LikelyCollapse prevention Not practical Limited

B. Proposed scheme

Limit State Isolation system Energy dissipation

Damage limitation Very likely LimitedSignificant damage Likely LikelyCollapse Limited Likely



where Eloop is the energy dissipated for cycle ofloading; and k0eff is the effective stiffness based onpeak negative and positive test displacements, D�p andDþp respectively.

In order to determine design displacements, lowesteffective stiffness and damping should be used. Bycontrast, design forces correspond to greatest stiffnessvalues and lowest damping[26].

Static analysis may be employed to establish aminimum level of design displacements and forcesand for preliminary designs, as detailed below.

Four distinct displacements should be computedaccording to the seismic hazard level, i.e. a 475-yearreturn period earthquake (corresponding, forexample, to LS of significant damage) or 2475-yearreturn period earthquake (LS of collapse). These

return periods are relative to residential buildings;higher values should be used for structures ofprimary importance for public safety, e.g. hospitals,fire stations.

These displacements are as follows:

* Design displacement (DD): displacement at the centreof rigidity of the isolation system at the LS ofsignificant damage. It is computed as follows:

DD ¼g

4p2

� �SDTD

BD(3)

where g is the acceleration due to gravity; SD isthe design 5% damped spectral acceleration for475-year return period earthquake; TD is theeffective period of a seismically isolated structureat the design displacement; and BD is the coefficient

........................................................................................................................................................

........................................................................................................................................................

Table 3 Advantages and disadvantages for different layout of base isolator locations (adapted from Naeim, 2001).

Bearing Location Advantages Disadvantages

Sub-basement No special detailing required for separationof internal services, e.g., elevator and stairways

Added structural costs unless sub-basementrequired for other purposes

No special cladding separation details Requires a separate (independent) retainingwall

Base of columns connected by diaphragmat isolation levelSimple to incorporate back-up system ofvertical loads

Top of basement columns No sub-basement requirement May require cantilevered elevator shaft belowfirst floor level

Minimal added structural costs Special treatment required for internalstairways below first floor level

Base of columns connected by diaphragmat isolation levelBackup system for vertical loads providedby columns

Bottom of first-storey columns Minimal added structural costs May require cantilever pitSeparation at level of base isolation is simpleto incorporateBase of columns may be connected bydiaphragmEasy to incorporate backup system forvertical loads

Top first-storey columns Minimal added structural costs Special detail required for elevators andstairs

Economic if first level is for parking Special cladding details required if first level isnot open

Backup system for vertical loads providedby columns

Special details required for vertical services

Fig. 5 Common isolation devices: steel reinforced elastomeric bearing with central lead core (left) and friction pendulum device (right)



expressed as a function of the effective damping ofthe isolation system at the design displacement bD.The latter may either be expressed in tabular

form[17] or in a close-form[15]; a graphicalcomparison is provided in Fig. 8. Theapproximation is very accurate for the values ofeffective damping of interest for practicalapplications, i.e. less than 50%. The damping bD inthe system is as follows:

bD ¼1

2p

PED

kD;maxD2D

� �(4)

whereP

ED is the sum of the energy dissipated percycle in all isolator units measured at DD.

* Maximum displacement (DM): displacement at thecentre of rigidity of the isolation system at the LS ofcollapse. Its expression is given by:

DM ¼g

4p2

� �SMTM

BM(5)

where SM is the design 5% damped spectralacceleration for 2475-year return periodearthquake; TM is the effective period of aseismically isolated structure at the maximumdisplacement; and BM is the coefficient expressedas a function of the effective damping of theisolation system at the maximum

Fig. 6 Idealized force–displacement relationships for isolationsystems

.......................................................................................................................................

.......................................................................................................................................

Table 4 Advantages and disadvantages of common base isolator devices (adapted from Naeim, 2001)

Isolator type Advantages Disadvantages

Elastomeric Low in-structure accelerations. High displacements.Low cost. Low damping.

No resistance to service load.P-D moments top and bottom.

High damping rubber Moderate in-structure acceleration. Strain dependent stiffness and damping.Resistance to service loads. Complex analysis.Moderate-to-high damping. Limited choice of stiffness and damping.

P-D moments top and bottom.Lead rubber Moderate in-structure accelerations. May require cantilever pit.

Wide choice of stiffness and damping. P-D moments top and bottom.Flat sliders Low profile. High in-structure accelerations.

Resistance to service loads. Properties depend on pressure and velocity.High damping. Sticking.P-D moments can be top and bottom. No restoring forces.

Curved sliders Low profile. High in-structure accelerations.Resistance to service loads. Properties depend on pressure and velocity.Moderate-to-high damping. Sticking.P-D moments can be top and bottom.Reduced torsion response.

Fig. 7 Design mechanical parameters for isolation systems

1.00.8 0.8

1.21.5

1.72.02.01.9

0.0

1.0

0 10 20 30 40 50 60 70 80 10090

2.0

3.0

4.0

(Naeim and Kelly, 1999)(FEMA 368, 2001)

Effective Damping β (Percentage of Critical)

Dam

ping

Coe

ffic

ient

(B

D o

r B

M)

Fig. 8 Damping coefficients as a function of effective damping



displacement (bM). The damping bM in the systemis as follows:

bM ¼1

2p

PEM

kM;maxD2M

� �(6)

whereP

EM is the sum of the energy dissipatedper cycle in all isolator units measured at DM.Values of equivalent viscous damping for commonseismic isolators vary between 10 and 20%, as afunction of the rubber compound. The numericalvalue of BM can be computed from Fig. 8.

* Total design displacement DTD: displacement of abearing at a corner of the building, including thecomponent of the torsional displacement in thedirection of the design displacement. Therefore, itshould be computed at the LS of significantdamage. Its expression is as follows:

DTD ¼ DD 1þ y12e

b2 þ d2

� �� (7)

where y is the distance between the centre ofrigidity of the isolation system and the element ofinterest measured perpendicular to the direction ofloading; e is the actual eccentricity between thecentre of mass of the super-structure and the centreof rigidity of the isolation system plus accidentaleccentricity equal to 5% of the longest plan dimen-sion of the structure perpendicular to the directionof loading; b is the shortest plane dimension of thestructure measured perpendicular to d; and d is thelongest plan dimension of the structure.

* Total maximum displacement DTM: displacement of abearing at a corner of the building, including thecomponent of the torsional displacement in thedirection of the maximum displacement. Thereforeit is computed at the LS of collapse. It may beevaluated through:

DTD ¼ DM 1þ y12e

b2 þ d2

� �� (8)

The effective isolated periods at designdisplacement (TD) and maximum displacement(TM) are computed from:

TD ¼ 2p

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiW

kD;ming

s(9)

and

TM ¼ 2p

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiW

kM;ming

s(10)

where W is the weight of the building; g is theacceleration due to gravity; kD,min is the minimumeffective horizontal stiffness of the isolation systemat the LS of significant damage; and kM,min is theminimum effective horizontal stiffness of theisolation system at the LS of collapse. The values ofTD are generally assumed equal to 2.5–3.0 s.However, in Japan, values of TD¼ 4.0 s are also

used because of the characteristics of the designearthquake spectra.

Design forces are calculated at or below theisolation system (Vb) and for the superstructure(Vs); the relevant formulae are:

ðat or below isolation systemÞVb ¼ kD;maxDD (11:1)

ðregular superstructureÞ

Vs ¼kD;maxDD

R0(11:2)

ðirregular superstructureÞVs ¼ kD;maxDDR0 (11:3)

in which R0 is the force reduction factor for isolatedstructures, which is related to the counterpart forfixed base buildings via:

1:04R0 ¼ 3=8ð ÞR½ �4 2:0 (12)

In the Eurocode 8[19], the value of the behaviourfactor q is equal to 1.5; this value of q is used to checkthe members and connections of the superstructure.The seismic isolation theory suggests a uniformdistribution of forces over the height of thesuperstructure; however, to account for higher modeeffects, the inverted triangular distribution may beconservately adopted[20]. In the European code[19], thedistribution of force is assumed proportional to thestorey masses.

By performing dynamic analysis, lower designdisplacements and forces can be used; a comparisonof the design parameters used in static and dynamicanalysis is provided in Table 5.

Retrofitting of steel and composite buildingsmay also be performed by employing dampers.Recently, their use has been allowed by design andassessment guidance documents[9,17] which alsoprovide comprehensive design rules. These devices,like base isolation, reduce the demand on thestructure by enhancing global damping; this limitsdamage to structural and/or non-structuralcomponents.

Dampers may be grouped as a function of theirmechanical response as follows:

* Displacement-dependent dampers: force–displacementresponse characteristics depend primarily onthe relative displacements. They includehysteretic (metallic), friction based, and SMAdampers.

* Velocity-dependent dampers: force-displacementresponse characteristics depend primarily on therelative velocity or the frequency of the motion.They include primarily viscous dampers.Viscoelastic devices are displacement and velocitydependent; they exhibit an elastic stiffness alongwith a viscous component.



All devices possess a similar feature, i.e. theyconvert external kinetic energy into heat; however, thelatter may be performed in different ways. Commondampers are summarized in Table 6, together withrelevant energy dissipation mechanisms.

Dampers may assume different configurations withrespect to the structural system that resists lateralforces[26]. They can be placed either externally or sharecommon elements with the structural systems;intermediate layouts may also be effective (Fig. 9).These layouts point to a fundamental differencebetween structures with dampers and base isolation(Table 7): the latter forms a series system (structure andisolators) while the former is a parallel system(structure and dampers). Isolators dissipate the inputenergy before it is transferred to the superstructure.By contrast, dampers receive and dissipate seismicenergy in combination with the lateral force-resistingstructure. The amount of dissipation depends uponthe dynamic characteristics of both components. As aresult, the damping should be tuned for optimumperformance of the overall system; this is usually a

cumbersome iterative design procedure. Designissues of common dampers are addressed below.

Hysteretic dampers

Hysteretic dampers are metal devices that candissipate energy from an earthquake through inelasticdeformation of metals. These dampers may yieldeither in bending, torsion and/or axially (mild steel)or shear (mild steel or lead). They usually have specialshapes; e.g. triangular or X-shapes, so that yieldingspreads uniformly throughout the material across thesection[13]. Typical added damping and addedstiffness (ADAS) metallic dampers with X-shape aredisplayed in Fig. 10.

Pioneering applications of ADAS were made inNew Zealand and Japan; recently they have also beenused for seismic rehabilitation of steel and compositestructures in the USA and Mexico[10,27]. K-bracings areusually combined with ADAS devices (Fig. 10); ADASdampers, located between the end of cross braces and

...................................................

..............................................................................................................................................

..............................................................................................................................................

Table 5 Comparison between static and dynamic analysis limit.

Dynamic analysis

Design Parameter Static Analysis Response Spectrum Time History

Design displacement (DD)

DD ¼g

4p2

SDTD

BD

� �

Maximum displacement (DM)

DM ¼g

4p2

SMTM

BM

� �

Total design displacement (DTD)DTD51:1DD DTD50:9DD DTD50:9DD

Total maximum displacement (DTM)DTM51:1DM DTD50:8DD DTD50:8DD

Design shear (Vb) (at or below isolation system)Vb ¼ kDmaxDD 50:9Vb 50:9Vb

Design shear (Vs) (Regular Super-structure)

Vb ¼kDmaxDD

R050:8Vs 50:6Vs

I Design shear Vs (Irregular Super-structure)Vs ¼ kDmaxDDR0 51:0Vs 50:8Vs

Story drift (%) 1.5 1.5 2.0

.............................................................................................................

.............................................................................................................

Table 6 Common dampers and energy dissipation mechanisms.

Damper type Response dependence Energy Dissipation Mechanism

Metallic Displacement Inelastic deformation of metal platesFriction Displacement Solid sliding frictionSMA Displacement Solid state phase transformationViscous Velocity Deformation of highly viscous fluidVisco-elastic Displacement and velocity Shear deformation of fluid



beam mid-span, are activated by storey drifts. Thesedampers should be designed in such a way that attheir yielding, axial loads in the braces are lower thanthe buckling values. The design is thereforeuneconomical because the tensile plastic capacity ofdiagonals is not fully exploited. The performance ofthese dampers depends upon the elastic stiffness andyield force of the damper and the elastic stiffness ofthe structure to which it is applied. To achievemaximum effectiveness the device should have highstiffness and high yield strength[28–30]. In practicalapplications (damping of the device varying between10 and 15%), it is difficult to separate the effects ofadded stiffness from the effects of added damping onresponse; both tend to reduce the displacementresponse. The higher the device-to-structure stiffnessis, the higher the damping will be. As a consequence,hysteretic dampers do not simply add damping, butmodify significantly all dynamic characteristics of thestructure. Typically, they reduce the fundamentalperiod, thus increasing the base shear. However, thesesystems are particularly attractive for retrofitting ofsteel buildings that are vulnerable to resonantresponse with the ground[31].

By assuming that the response of a hystereticdamper is elasto-perfectly plastic and provided that

the structure behaves elastically, the equivalentviscous damping (Fig. 11) is related to the ratio of thedamper yield force FY to structure elastic force FE andthe ratio of damper elastic stiffness kD to structureelastic stiffness kS as follows:

b ¼2

FY

FE1� FY

FE

kD

kS

� �

p 1þ FY

FE

� � (13)

The damping increases as a function of the damperrelative stiffness. There is an optimum value of thebrace strength to the elastic structure force; it varieswith the stiffness ratio.

The unbounded brace is based upon the samemetallic yielding principle of ADAS, i.e. tension/compression yielding brace; this device has becomepopular in Japan and recently also in the US[12]. Itconsists of a core steel plate (Fig. 12) encased in aconcrete filled steel tube. Yielding of the interiorcomponent under reversal axial loads provides stableenergy dissipation; the exterior concrete-filled steeltube prevents local and member buckling. A specialcoating is applied to reduce friction. Severallaboratory tests are being conducted to investigate

Fig. 9 Different configuration of dampers and seismic force resisting system (SFRS)[26]

........................................................................................................................................................

........................................................................................................................................................

Table 7 Comparison between seismic isolation and damper devices.

System characteristics Base isolation Dampers

Location of devices Concentrated at base Distributed along heightEffect on mechanic response Lengthening period (greater) and increasing

damping (lesser) increasing damping (greater)Shortening period (lesser) and

Structural benefits Reduction of forces, displacements andaccelerations (factors range 2 to 6) andaccelerations (factor range: (1.5 to 2)

Reduction of forces, displacements

Mechanism of energy dissipation Series system (structure and isolators) Parallel system (structure and dampers)Candidate buildings Rigid buildings with stiff lateral load

resisting structure (low-to-medium rise)Flexible buildings with slender lateralload resisting structure (medium-to-high rise)

Buildings on firm soil sites. Buildings on soft soil sites.Device installation Intrusive Not intrusiveCost Expensive Relatively cheap



the performance of unbonded braces and steel framesretrofitted with such devices[32].

Experimental and analytical work on hystereticdampers has been carried out in Europe during thelast decade[33,35]; several new configurations anddevices have been proposed. For example, Fig. 13shows two braces connecting an outer rectangular toan inner smaller frame. This layout enhances thestructural performance with respect to ordinary cross-braces[36]. The latter behave poorly under seismicloads because of buckling of compressed members;the proposed layout overcomes this drawback. Itenables the brace to recover the deformation due tocompression, either partially or entirely, thusimproving the dissipative capacity of the systemunder earthquakes.

To design these dampers, i.e. to compute theforces relative to the onset of their yielding, theirgeometry must be adequately defined. Yieldingof the devices should anticipate the yielding of

cross-braces. Such arrangements may show out-planevibrations at device locations. Relatively small massesare placed at the centre of bracings and can displacesideways.

Friction dampers

These dampers rely upon the mechanism of frictionbetween two solid bodies sliding relative to oneanother. Friction is an excellent mechanism of energydissipation; it has been used extensively andsuccessfully in automotive breaks to dissipate kineticenergy.

Various materials are used for the sliding surfacesuch as brake pad material on steel, steel on steel, steelon brass in slip-bolted connections, graphite-impregnated bronze on stainless steel and other metalalloys. The choice of the base metal for frictiondampers is crucial; poor corrosion resistance can often

Fig. 10 ADAS damper: location within a frame and close up (top) and typical geometry (bottom)



reduce the coefficient of friction assumed in thedesign for the intended life of the device. Low-carbonalloy steels corrode and their interface properties varywith time, while brass and bronze promote additionalcorrosion when in contact with low carbon. Bycontrast, stainless steels do not appear to sufferadditional corrosion when in contact with brass orsteel; therefore, they are suitable for such devices.

Generally, friction devices offer good seismicperformance and their response is independent ofloading amplitude, frequency and number of cycles.They combine high energy-dissipation potential andrelatively low cost, and are easy to install andmaintain. Other beneficial effects of such dampershave been investigated by Austin & Pister[37];Filiatrault & Cherry[36] and Aiken & Kelly[38]. Indesigning friction-based dampers it is essential tominimize stick-slip phenomena, thus avoiding high-frequency excitation. The ratio of initial slip load tostorey yielding shear and ratio of bracing-to-storeystiffness also significantly affect the performance ofthe device.

Friction devices usually produce a stablerectangular hysteresis, although some are configuredto produce self-centring force and providenonrectangular hysteresis shapes with loadproportional to displacement. The Coulomb model isa macroscopic hysteretic model for friction-based

dampers with a constant coefficient of friction; acomparison of hysteretic loops for the assesseddampers is given in Fig. 14.

The equivalent viscous damping of a frictiondamper may be computed as follows (Fig. 15):

b ¼2

FY

FE

p 1þ FY

FE

� � (14)

This formula may be derived from that in eq. (13)by setting the ratio of the damper stiffness kD tostructure stiffness kS to infinity. When the force of thefriction device is greater than the structure force(Fig. 15), the upper bound of the equivalent viscousdamping of the damper is 2/p, i.e. 63.99%.

Friction dampers have been commonly placedwithin diagonal braces, as with yielding metaldampers, but can also be placed horizontally betweenthe top of a wall and the beam above. These devicesinitially possess finite stiffness because they aremounted on braces; therefore, their behaviour issimilar to hysteretic damping.

A typical friction damper is shown in Fig. 16: it canbe installed at the crossing of two braces wheretension in one brace forces the joint to slip, thusactivating four links which in turn force the joint inthe other brace to slip[39]. This device is fixed underwind and moderate earthquakes but slips underintense ground motions, thus protecting primarystructural members from yielding[40].

0

10

20

30

40

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Kd / Ks = 0.1 Kd / Ks = 0.5 Kd / Ks = 1

Kd / Ks = 1.5 Kd / Ks = 2 Kd / Ks = 5

Kd / Ks = 10 Kd / Ks = 50 Kd / Ks = 100

Damper Yield / Elastic Force

Equ

ival

ent V

isco

us D

ampi

ng (

%)

Fig. 11 Equivalent viscous damping as a function of the braceproperties

Fig. 12 Unbounded brace: internal and external components (left) and hysteresis loop (right)

Fig. 13 Hysteretic braces: device (left) and location withinframes (left)



Sliding connections contribute to elevated frictionand hence, energy dissipation. Force acting within thejoints can be easily controlled by satisfying the forceequilibrium equation:

Ng ¼ 2N1 �Nb (15)

where Ng is the tensile force in the brace when slipstarts and Nl is the correspondent force in the damper.By assuming slender diagonal braces, which can

buckle elastically, the buckling load (Nb) is low andmay be neglected in eq. (15). Therefore:

Ng ’ 2N1 (16)

The device can be calibrated via the tightening forceapplied to the bolts. Diagonal braces using the Pallsystem possess enhanced dissipative capacity}theenergy dissipation is roughly two-fold}with regardto ordinary cross-bracings[41].

(a)

(a)

(b)

(c)

(d)

(e)

(b) (c)

(d) (e)

Fig. 14 Idealized force–displacement relationships for dampers

00.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

10

20

30

40

50

60

70

Damper Yield / ElasticForce

Equ

ival

ent V

isco

us D

ampi

ng (

%)

00 10 20 30 40 50 60 70 80 90 100

10

20

30

40

50

60

70

Damper Yield / Elastic Force

Equ

ival

ent V

isco

us D

ampi

ng (

%)

Fig. 15 Equivalent viscous damping of friction device: force less (left) or greater (right) than structure

Fig. 16 Pall friction device: frame location (left), device layout (middle) and hysteretic loop (right)



Recently, novel friction dampers have been testedexperimentally and many have been installed in newand existing buildings around the world[38,42,43]. Twoexamples of devices for braced connections areprovided in Fig. 17; they are a slotted boltedconnection energy dissipater[43] and a novel device forinverted V-braced connections[44], respectively.

Slotted bolted connections are becoming verypopular for braced connections because they requireonly a slight modification of standard constructionpractice, and are thus are easy to construct andimplement[45]. They also make use of materials widelyavailable on the market, and are a cost-effectivemeans of retrofitting existing steel and compositeframed buildings. These connections are designed todissipate energy through friction between the steelsurfaces along the brace in tension and compressionloading cycles[46]; alternatively, brass in contact withsteel may be used. Experimental tests have shownthat the behaviour of connections with brass on steelis more uniform[42]; moreover, they are simpler tomodel analytically than those using steel on steel, andtheir performance in braced systems is verysatisfactory[27,47]. The slotted connection in Fig. 17 hasshown excellent energy dissipation on knee-bracedframes[43]. It has slotted holes in the main connectionplate which are parallel to the line of loading. Themain plate is sandwiched between two outermembers. A friction lining pad is placed between eachouter member and the main plate. The lining padmoves with the outer member. Two bolts are used toclamp together the plates and lining pads. Upontightening the bolts, frictions develop between thecontact surfaces of lining pads and slotted plate.When either tensile or compressive forces are appliedto the connection and the friction is exceeded, theslotted plane slips relative to the lining pads andenergy is dissipated.

The effectiveness of the novel damper for invertedV-braced connections (as per Fig. 17) has beenassessed experimentally and numerically[44]. Thedamper consists of three steel plates, i.e. one central(vertical), two on the side (horizontally) and two

circular friction pad discs sandwiched between thesteel plates. The central plate is used for theconnection with the mid-span of the floor beam in theframe. This connection is pinned and thus increasesthe relative rotation between the central and the sideplates; as a result, energy dissipation is enhanced.

Viscous dampers

Viscous dampers are similar to shock absorbers in acar. They consist of a piston within a damper chamberfilled with a compound of silicone oil[48,49], and maybe pressurized[50,51]. The piston rod is connected to apiston head with a number of small orifices. As thepiston moves within the cylinder the oil is forced toflow through holes in the piston head thus causingfriction. If the fluid is purely viscous (e.g. Newtonian)the output force of the damper is directly proportionalto the velocity of the piston. The linear responsecharacterizes a broad range of load frequencies anddoes not vary with the temperature. This is typical ofdampers employing a storage chamber and controlvalve (Fig. 18); such devices are calibrated withthreshold forces (cut-off frequencies) to enhance theirdissipation capacity. Moreover, these viscous dampersexhibit stiffening characteristics at higher frequenciesof deformation[12]; thus, they are used to damp higher-mode effects. Re-centring properties are typicallyfound in dampers with pressurized fluids.

In addition, damping walls, consisting of arectangular steel plate immersed in a highly viscousfluid, showed high energy dissipation when installedin steel frames[52,53]. Their damping characteristicsvary significantly with load frequency andtemperature.

Viscous dampers have low resistance todeformation when loads are applied slowly, butresistance increases as the applied deformation rateincreases. When they are installed in buildings,usually in bracings, input energy due to earthquakeloading is transformed by friction into heat. Viscousdampers have been used to control the vibrations of

Fig. 17 Novel friction dampers for braced connections: slotted bolted connection (left) and inverted V-braced connection (right)



new buildings and to retrofit either RC or steelframes, especially in Japan and the US[10].

Fluid viscous dampers provide damping as afunction of velocity; the velocity is out of phase withthe displacements. The relationship between thedamper force FD and the applied velocity ð ’DÞ is asfollows:

FD ¼ C ’D�� asgn ð ’DÞ (17)

in which a is the damper exponent and sgnð ’DÞ is thesignum function defining the sign of the relativevelocity. Therefore, the mechanical response ofviscous dampers is governed by two parameters: thecoefficient C and the power exponent a. For practicalapplications, the latter generally ranges between 0.3and 1.0[12]; a should be as high as possible, with anoptimal value of 1.0 which produces a linearly elasticrelationship between velocity and displacement.Lower limits of the exponent, say around 0.3, shouldbe avoided; in these cases the damper forces are notout of phase with the displacements and hence areadditive to the structural displacements. This

coupling effect increases with the amount of dampingprovided; the more damping provided, the smallerthe benefit of having the damper force out of phasewith the structure force.

The damper coefficient C can assume almost anyvalue: C may be increased or decreased simply byinstalling more or fewer dampers in the structure.Certain types of viscous dampers have a relief valveproviding a limited velocity. This is useful in limitingforces but reduces out-phase between velocity anddisplacements which may undermine theeffectiveness of the damper on the global structuralperformance.

Several methodologies have been given for thedesign of viscous dampers[13]. The determination ofoptimal locations to achieve maximum performancebased on given constraints, such as the number ofdampers and their capacity, is of paramountimportance. The interaction damper structure shouldbe explicitly accounted for because of the connectionflexibility. Viscous dampers are usually mounted onsteel braces and then installed in building frames

Fig. 18 Typical internal layout of viscous dampers (Courtesy of Taylor Devices, Inc.)

Fig. 19 Viscous dampers in steel framed building



(Fig. 19). Therefore, the damper is linked in serieswith the brace and in parallel with the structure.

As a consequence, the brace reduces thedeformation in the damper due to storey drift. In fact,the former behaves like a spring while the damperforms a spring–dashpot system in series (Maxwellmodel); therefore the spring deformability reduces therelative displacement of the damper. However, thiseffect is a function of connection flexibility: the higherthe connection flexibility the lower the damping force.The interaction damper structure dependssignificantly on whether the structure undergoesinelastic deformations.

Viscoelastic dampers

These devices are based on viscoelastic materials,such as copolymers or glassy substances, with highenergy dissipation due to shear deformations.Viscoelastic materials have linear response over awide range of strains provided the temperature isconstant. At large strains, e.g. 300–500%, the energydissipation produces self-heating, affecting themechanical properties of the material, which thenbecomes highly nonlinear.

Typical viscoelastic dampers, in this case developedby 3M Company, are shown in Fig. 20: they consist ofviscoelastic layers bounded with steel plates.Mounted in the structure, shear deformation andhence, energy dissipation, take place when structuralvibration induces relative motion between outer steelflanges and centre plates.

The damping force FD is a function of the velocityð ’DÞ and the displacement D and can be computed as:

FD ¼ keffDþ C ’D (18)

where C is the damping coefficient and keff is theeffective stiffness of the damper and is given as:

keff ¼G0Ab

t(19)

in which G0 is the shear storage modulus, Ab thebonded area of the device and t the total thickness of

the viscoelastic material in the device (the sum of thelayers for the devices in Fig. 20).

The damping coefficient C depends upon: (i)geometric properties of the device; (ii) mechanicalproperties of the material; and (iii) frequency of thecontent of the applied load. It may be given asfollows:

C ¼ G00Ab

ot(20)

where G00 is the shear loss modulus; and o is thefrequency of the load.

The shear storage G0 and shear loss G00 moduli arefunctions of: (i) the material used; (ii) the strain; (iii)the frequency of the load; and (iv) the servicetemperature[54]. The former influences the stiffnessand hence the frequency of the damper, while thelatter relates to the energy dissipated in each cycle.Their ratio (loss factor) usually varies between 0.8 and1.40; for practical applications it is independent ofstrain and temperature.

Several analytical expressions are provided in theliterature for G0 and G00[55]; there are four parameterformulations, based either upon experimental tests oron Boltzmann’s superposition principle.

The design of viscoelastic dampers is complex andcumbersome and is based upon iterative procedures.The proposed methodologies, common to the designof viscous dampers, may be categorized in twogroups based on the behaviour of framed structures,i.e. elastic[56–59] and inelastic with plasticdeformations[60,61]. The design steps for the design ofthese dampers for seismic retrofitting are as follows:

1. Establish the number and location of dampers. Thelocation generally depends upon the storeydisplacements. The optimum location is where thedrifts cause maximum relative displacement of thebrace with dampers. It is important to account forthe increment of axial load in the columns due tothe bracings. This is less important for viscousdampers because of the out-of-phase betweendamper and structural forces. Other criteria for theoptimal location of dampers may be found inelsewhere[62,63].

Fig. 20 Typical viscoelastic dampers



2. Define through experimental tests or designspectra the design parameters of dampers, i.e.G0 and G00 for viscoelastic and C and a forviscous[58–61].

3. Evaluate the stiffness of dissipative braces,assuming elastic response for braces and avoidingbuckling. For viscoelastic dampers, the area of thecross-section of viscoelastic material may becomputed by using eq. (20). The stiffness of thedissipative brace is computed through thefollowing relationships[59]:

bieq ¼

1

2k

i1� oi

s

ois;b

!224

35 (21:1)

ki¼

k0eff

keff

!i

(21:2)

keff

i¼ kb þ keffð Þkbkeff þ kbk0 2eff

kb þ keffð Þ2þk0 2eff

" #i

(21:3)

k0eff

i¼ k2

bk0 2eff

kb þ keff

2þk0 2eff

" #i

(21:4)

where ois and oi

s;b are the ith frequencies of thestructure without and with dampers; the ithfrequency is the frequency of the mode to mitigate(usually it is the first mode); kb is the stiffness of thebrace with the damper; keff is given in eq. (20); andk0eff has the same expression as keff, but with G00

replacing G0.4. The thickness should comply with the maximum

shear deformations, e.g. 150% (LS of significantdamage) and 250% (LS of collapse).

5. Perform dynamic analysis of the structure withdampers to check that the target equivalent viscousdamping is achieved. The iterative procedure isrequired to adjust the damper characteristics inorder to achieve the target.

Experimental and numerical tests havedemonstrated the effectiveness of these dampers for

the retrofitting of either steel or RCstructures[10,57,64,65].

SMA dampers

Shape memory alloys (SMAs) are metal alloys whichcan sustain large strains (up to 10%) with no residualdeformation after unloading (super-elasticity). Thisproperty is due to reversible solid-to-solid phasetransformations (austenite to martensite), which canbe either thermal or stress induced. The chemicalinstability of the martensite which switches toaustenite at low stresses generates stable hystereticloops, thus giving rise to adequate energy dissipationcapacity[66]. Further details on the mechanicalproperties of SMAs can be found in Di Sarno &Elnashai[67]. It suffices here to say that these specialmetal alloys exhibit advantageous re-centringproperties and high low-cyclic fatigue resistance;therefore, SMAs are particularly attractive forapplications in seismic design and redesign. Recently,they have been implemented in several dissipativedevices for earthquake protection, especially diagonalbraces and base-isolation systems[23,68,69]. Thesedevices combine the re-centring and dissipativeproperties of SMAs, as displayed in Fig. 21. Thereader is referred to the companion paper by Di Sarno& Elnashai,[67] for a thorough critical review of theseismic applications of SMAs in new and existingstructures.

Conclusions

Design issues and the advantages and disadvantagesof innovative strategies for seismic retrofitting of steeland composite structures, such as base isolation anddevices for supplemental damping, have beendiscussed in this paper.

By introducing three limit states, as used in recentEurocode 8 guidelines, namely damage limitation,significant damage and collapse prevention, the

d

F

O

RE-CENTRING PROPERTY

d

F

DISSIPATIVE PROPERTY COMPLETE DEVICE

d

F

O

Fig. 21 Idealized behaviour of SMA device



viability and cost-effectiveness of the aforementionednon-conventional strategies has been assessed.

This study has shown that the selection of aparticular type of vibration control device is governedby a number of factors which include the following:

* efficiency;* compactness;* weight;* capital and operating costs;* maintenance Requirements;* safety.

Sound engineering practices enable all but thecapital and operating costs to be counteracted,depending upon the required levels of structuralresponse parameter to be reduced.

Comparison of the various types of devices for baseisolation systems and dampers is a recommendedexercise in determining the optimal solution to beadopted.

Acknowledgements

This work was supported in part by the EarthquakeEngineering Research Centers Program of theNational Science Foundation under NSF Award EEC97-01785. Any opinions, findings and conclusions orrecommendations expressed are those of the authorsand do not necessarily reflect those of the NationalScience Foundation.

References

[1] Mahin SA. Lessons from damage to steel buildings during the Northridge

earthquake. Engineering Structures 1998: 20(4–6): 261–270.

[2] Miller DK. Lessons learned from the Northridge earthquake. Engineering

Structures 1998: 20(4–6): 249–260.

[3] Watanabe E, Sugiura K, Nagata K & Kitane Y. Performances and

damages to steel structures during 1995 Hyogoken-Nanbu earthquake. Engineering

Structures 1998: 20(4–6): 282–290.

[4] Nakashima M, Inoue K & Tada M. Classification of damage to steel

buildings observed in the 1995 Hyogoken-Nanbu earthquake. Engineering Structures

1998: 20(4–6): 271–281.

[5] Federal Emergency Management Agency. State of Art Report on

past performance of steel moment frame buildings in earthquakes. Report No. FEMA

355E. Washington, D.C., USA. 2000.

[6] Di Sarno L & Elnashai AS. Seismic Retrofitting of Steel and Composite

Buildings. Mid-America Earthquake Center, CD Release 02-01, University of Illinois

at Urbana-Champaign, USA. 2002.

[7] Federal Emergency Management Agency. Interim Guidelines:

Evaluation, repair, modification and design of welded steel moment frames.

Report No. FEMA 267/SAC-95-02, SAC Joint Venture, Sacramento, California,

USA. 1995.

[8] Federal Emergency Management Agency. Recommended seismic

evaluation and upgrade criteria for existing welded steel moment-frame buildings.

Report No. FEMA 351. Washington, D.C., USA. 2000.

[9] Federal Emergency Management Agency. Pre-standard and com-

mentary for the seismic rehabilitation of buildings. Report No. FEMA 356. Washington,

D.C., USA. 2000.

[10] Soong TT & Spencer BF Jr. Supplemental energy dissipation: state-of-

the-art and state-of-practice. Engineering Structures 2002: 24(3): 243–259.

[11] Nishitani A & Inoue Y. Overview of the application of active/semiactive

control to building structures in Japan. Earthquake Engineering and Structural Dynamics

2001: 30(11): 1565–1574.

[12] Housner GW, Bergman LA, Caughey TK, Chassiakos AG, Claus

RO, Masri SF, Skelton RE, Soong TT, Spenser BF Jr & Yao TP. Structural

control: past, present and future. Journal of Engineering Mechanics (ASCE) 1997:

123(9): 897–971.

[13] Soong TT & Costantinou MC. Passive and active structural vibration

control in civil engineering. Springer-Verlag, Wien. 1994.

[14] Soong TT & Dargush GF. Passive energy dissipation systems in

structural engineering. John Wiley & Sons, Inc., New York, New York, USA. 1997.

[15] Naeim F & Kelly JM. Design of seismic isolated structures: from theory

to practice. John Wiley & Sons Inc., New York, New York, USA. 1999.

[16] Federal Emergency Management Agency. NEHRP guidelines for the

seismic rehabilitation of buildings. Report No. FEMA 273, Washington, D.C., USA.

1997.

[17] Federal Emergency Management Agency. NEHRP recommended

provisions for seismic regulations for new buildings and other structures,

Part 1}Provisions. Report No. FEMA 368, Washington, D.C., USA. 2001.

[18] Federal Emergency Management Agency. NEHRP Commentary on

the guidelines for the seismic rehabilitation of buildings. Report No. FEMA 274.

Washington, D.C., USA. 1997.

[19] Eurocode 8. Design provisions for earthquake resistance of structures.

Part 1.3: General rules. Specific rules for various materials and elements. European

Communities for Standardisation, Brussels, Belgium. 1998.

[20] Kelly JM. Earthquake-resistant design with rubber. 2nd Edition, Springer-

Verlag, Inc., New York, New York, USA. 1996.

[21] Naeim F. The seismic design handbook, 2nd Edition. Kluwer Academic

Publisher. 2001.

[22] Skinner RI, Robison WH & McVerry GH. An introduction to seismic

isolation. John Wiley & Sons, Inc., New York, New York, USA. 1993.

[23] Wilde K, Gardoni P & Fujino Y. Base isolation system with shape

memory alloy devices for elevated highway bridges. Engineering Structures 2000:

22(3): 222–229.

[24] International Code Council. International Building Code. Whittier,

California, USA. 2000.

[25] Federal Emergency Management Agency. Global topics report on

the pre-standard and commentary for the seismic rehabilitation of buildings.

Report No. FEMA 357, Washington, D.C., USA. 2000.

[26] Federal Emergency Management Agency. NEHRP recommended

provisions for seismic regulations for new buildings and other structures,

Part 2 } Commentary. Report No. FEMA 369, Washington, D.C., USA. 2001.

[27] Aiken I, Nims D, Whittaker A & Kelly JM. Testing of passive energy

dissipation systems. Earthquake Spectra 1993: 9(2): 335–370.

[28] Whittaker AS, Bertero VV, Thompson CL & Alonso LJ. Seismic

testing of steel plate energy dissipation devices. Earthquake Spectra 1991: 7(4):

563–604.

[29] Xia C & Hanson RD. Influence of ADAS element parameters n building

seismic response. Journal of Structural Engineering (ASCE) 1992: 118(7): 1903–1918.

[30] Tsai CS, Chen HW, Hong CP & Su YF. Design of steel triangular plate

energy absorbers for seismic-resistant construction. Earthquake Spectra 1993: 9(3):

505–528.

[31] Tena-Colunga A & Vergara A. Comparative study on the seismic

retrofit of a mid-rise steel building: steel bracing vs. energy dissipation. Earthquake

Engineering and Structural Dynamics 1997: 26(6): 637–655.

[32] Iwata M, Kato T & Wada A. Buckling-restrained braces as hysteretic

dampers. Proceedings of the 3rd International Conference on Behavior of Steel Structures in

Seismic Areas (STESSA 2000), Montreal, Canada, 2000: 33–38.

[33] Vulcano A. Uso di controventi metallici con dispositivi dissipativi

nell’adeguamento antisismico di strutture intelaiate. Costruzioni Metalliche 1991: 4:

244–261 (in Italian).

[34] Ciampi V. Development of passive energy dissipation techniques for

buildings. Proceedings International Post-SMIRT Conference Seminar on Isolation, Energy

Dissipation and Control of Vibrations of Structures, Capri (Italy), 1993: 495–510.

[35] Chen SJ. A simple and effective retrofit method for steel beam-to-column

connections. Seventh US-Japan Workshop on the improvement of structural design and

construction practices, ATC, Redwood City, California, USA. 1996.

[36] Filiatrault A & Cherry S. Performance evaluation of friction damped

braced frames under simulated earthquake loads. Earthquake Spectra, 1987:

3(1): 57–78.



[37] Austin MA & Pister KS. Design of seismic-resistant friction-braced

frames. Journal of Structural Engineering (ASCE) 1985: 111(12): 2751–2769.

[38] Aiken ID & Kelly JM. Earthquake simulator testing and analytical studies

of two energy-absorbing systems for multistory structures. Report No. UCB/EERC-90/

03, University of California at Berkeley, California, USA. 1990.

[39] Pall AS & Marsh C. Response of friction damped braced frames. Journal

of Structural Division (ASCE) 1982: 108(9): 1313–1323.

[40] Filiatrault A & Cherry S. Seismic design spectra for friction-damped

structures. Journal of Structural Engineering (ASCE) 1990: 116(5): 1334–1355.

[41] Filiatrault A & Cherry S. Parameters influencing the design of friction-

damped braced frames. Canadian Journal of Civil Engineering 1989: 16: 753–766.

[42] Grigorian CE, Yang TS & Popov EV. Slotted bolted connection energy

dissipators. Earthquake Spectra 1993: 9(3): 491–504.

[43] Balendra T, Yu CH & Lee FL. An economical structural system for

wind and earthquake loads. Engineering Structures 2001: 23(5): 491–501.

[44] Mualla IY & Belev B. Performance of steel frames with a new friction

damper device under earthquake excitation. Engineering Structures 2002: 24(3):

365–371

[45] Levy R, Marianchik E, Rutenberg A & Segal F. Seismic design

methodology for friction damped braced frames. Earthquake Engineering and Structural

Dynamics 2000: 29(11): 1569–1585.

[46] Popov EV, Yang TS & Grigorian CE. New directions in structural

seismic design. Earthquake Spectra 1993: 9(4): 845–875.

[47] Colajanni P & Papia M. Hysteretic characterization of friction-damped

braced frames. Journal of Structural Engineering (ASCE) 1997: 123(8): 1020–1028.

[48] Makris N, Dargush GF & Constantinou MC. Dynamic analysis of

generalized viscoelastic fluids. Journal of Engineering Mechanics (ASCE) 1993: 119:

1663–1679.

[49] Constantinou MC, Symans MD, Tsopelas P & Taylor DP. Fluid

viscous dampers in applications of seismic energy dissipation and seismic isolation.

Proceedings ATC 17-1 on Seismic Isolation, Passive and Active Control 1993: 2: 581–592.

[50] Tsopelas P & Constantinou MC. Experimental and analytical study of a

system consisting of sliding bearings and fluid restoring forces/damping devices.

Technical Report NCEER-94-0014, National Center for Earthquake Engineering

Research, Buffalo, New York, USA. 1994.

[51] Pekcan G, Mander JB & Chen SS. The seismic response of a 1:3 scale

model RC structure with elastomeric spring dampers. Earthquake Spectra 1995:

11(2): 249–267.

[52] Arima F, Miyazaki M, Tanaka H & Yamazaki Y. A study on buildings

with arge damping using viscous damping walls. Proceedings of the 9th World Conference

on Earthquake Engineering, Tokyo (Japan), August 2–9, 1988: vol. V: 821–826.

[53] Miyazaki M & Mitsusaka Y. Design of a building with 20% or greater

damping. Proceedings of the 10th World Conference on Earthquake Engineering, Madrid

(Spain), July 19–24, 1992: vol. 8: 4143–4148.

[54] Chang KC, Soong TT, Oh ST & Lai ML. Effect of ambient

temperature on viscoelastically damped structure. Journal of Structural Engineering

(ASCE) 1992: 118(7): 1955–1973.

[55] Shen KL & Soong TT. Modeling of viscoelastic dampers for structural

applications. Journal of Engineering Mechanics (ASCE) 1995: 121(6): 694–701.

[56] Chang KC, Soong TT, Lai ML & Nielsen EJ. Development of a design

procedure for structures with added viscoelastic dampers. Proceedings ATC 17-1 on

Seismic Isolation, Energy Dissipation and Active Control, 1993: 2: 473–484.

[57] Chang KC, Soong TT, Oh ST & Lai ML. Seismic behavior of steel

frame with added viscoelastic dampers. Journal of Structural Engineering (ASCE) 1995:

121(10): 1418–1426.

[58] Kasai K, Fu Y & Watanabe A. Passive control systems for

seismic damage mitigation. Journal of Structural Engineering (ASCE) 1998: 124(5):

501–512.

[59] Fu Y & Kasai K. Comparative study of frames using viscoelastic and

viscous dampers. Journal of Structural Engineering (ASCE) 1998: 124(5): 513–522.

[60] Abbas H & Kelly JM. A methodology for design of visco-elastic dampers

in earthquake-resistant structures. Report No. UCB/EERC-93/09, University of

California at Berkeley, California, USA. 1993.

[61] Ciampi V, De Angelis M & Paolacci F. Una metodologia di progetto su

base energetica per controventi dissipativi viscoelastici. Proceedings of the 8th National

Conference on Earthquake Engineering, Taormina (Italy), September 21–24, 1997: 1:

653–660 (in Italian).

[62] Zhang R-H & Soong TT. Seismic design of viscoelastic dampers for

structural applications. Journal of Structural Engineering (ASCE) 1992: 118(5):

1375–1392.

[63] Shukla AK & Datta TK. Optimal use of viscoelastic dampers in building

frames for seismic forces. Journal of Structural Engineering (ASCE) 1999: 125(4):

401–409.

[64] Foutch DA, Wood SL & Brady PA. Seismic retrofit of nonductile

reinforced concrete frames using viscoelastic dampers. Proceedings ATC 17-1 on

Seismic Isolation, Passive and Active Control, 1993: 2: 605–616.

[65] Shen KL, Soong TT, Chang KC & Lai ML. Seismic behavior of

reinforced concrete frame with added visco-elastic dampers. Engineering Structures

1995: 17(5): 372–380.

[66] Duerig TW, Melton KN, Stockel D & Wayman CM. Engineering

aspects of shape memory alloys. Butterworth-Heinemann, London, UK. 1990.

[67] Di Sarno L & Elnashai AS. Special Metal for Seismic Retrofitting of Steel

Structures. Progress in Structural Engineering and Materials, 2003: 5(2): 60–76.

[68] Dolce M, Cardone D & Marnetto R. Implementation and testing of

passive control devices based on shape memory alloys. Earthquake Engineering and

Structural Dynamics, 2000: 29(7): 945–968.

[69] Des Roches & R Delemont M. Seismic retrofit of simply supported

bridges using shape memory alloys. Engineering Structures, 2002: 24(3): 325–332.

L D Sarno

Department of Engineering,

University of Sannio, Benevento,

82100, Italy

E-mail: [email protected]

A S Elnashai

Department of Civil and Environmental Engineering,

University of Illinois,

Urbana, IL 61801, USA



Innovative strategies for seismic retrofitting of steel and composite structures

Documents

Transcript of Innovative strategies for seismic retrofitting of steel and composite structures