Journal of Constructional Steel Research 70 (2012) 358–367

Contents lists available at SciVerse ScienceDirect

Journal of Constructional Steel Research

Hysteretic model for steel energy dissipation devices and evaluation of a

minimal-damage seismic design approach for steel buildings
Theodore L. Karavasilis a,⁎, Sanaya Kerawala b, Emma Hale b
a
School of Engineering, University of Warwick, Coventry, CV4 7AL, UK
b
Engineering Consultancy, Mott McDonald, London, UK

a r t i c l e i n f o a b s t r a c t

Article history: This paper evaluates an alternative seismic design approach for steel structures that concentrates damage in
Received 4 April 2011 easy-to-replace steel energy dissipation devices and protects the main structural members from yielding with
Accepted 9 October 2011 capacity design rules. This approach can reduce damage repair costs and downtime, and, can be further enhanced
Available online 4 November 2011
by using rate-dependent dampers in parallel to steel devices to achieve drift reduction and protection of drift-
sensitive non-structural elements. A model for steel energy dissipation devices is proposed and calibrated against
Keywords:
Hysteretic model
experimental results. In particular, the Bouc-Wen model is modified to capture the combined kinematic and
Isotropic hardening isotropic hardening in the hysteresis of steel devices. The model is found able to accurately predict the experi-
Energy dissipation devices mentally obtained hysteresis and is implemented in the OpenSees software for use in seismic response analysis.
Viscous dampers A simplified seismic design procedure is proposed and used to design a prototype steel building equipped with
Minimal-damage seismic design steel devices and viscous dampers according to explicitly defined minimal-damage performance objectives.
Steel frames Seismic analyses results indicate the accuracy of the design procedure and confirm that the building is able to
Residual storey drift achieve immediate occupancy under the design seismic action and rapid return to occupancy under the maxi-
mum considered seismic action. The same building is designed as a conventional steel MRF according to EC8.
Results of seismic analyses show that repair of damage in the main structural members of the conventional
MRF may not be financially viable in the aftermath of the design and maximum considered earthquakes.
© 2011 Elsevier Ltd. All rights reserved.

1. Introduction
Conventional seismic-resistant systems, such as steel moment-
resisting frames (MRFs), are currently designed to develop significant
inelastic deformations in main structural members (i.e., beams and
columns) under strong earthquakes [1]. Such a design approach has cer-
tain advantages, including acceptable seismic performance in terms of
life-safety, and, economy since designing a structure to behave elastical-
ly under a strong earthquake would require very large structural mem-
bers. However, inelastic deformations in main structural members
result in difficult to repair damage and residual drifts, and hence, in
high repair costs and costly downtime while the building is repaired
and cannot be used or occupied. In addition, a recent study reported
that repairing conventional systems that had experienced residual
storey drifts greater than 0.5% after the Hyogoken-Nanbu earthquake
was not financially viable [2].
An alternative seismic design approach is to concentrate damage in
accessible and easily replaceable steel energy dissipation devices and
protect the main structural members from yielding with capacity
design rules. This design approach can significantly reduce damage repair
⁎ Corresponding author. Tel.: + 44 1865 2 73144.
E-mail address: theodore.karavasilis@eng.ox.ac.uk (T.L. Karavasilis).
costs and downtime. Extensive research has been conducted to develop
robust steel energy dissipation devices, such as the ADAS [3], TADAS
[4], shear panels [5–7] or devices based on yielding of the web of standard
steel sections [8–10]. These devices can be attached to the frame (e.g.,
bottom flange of the beam) using supporting braces. Alternatively, steel
energy dissipation devices can be realized as yielding braces able to
avoid buckling such as the buckling-restrained brace (BRB) [11–15].
Earthquake reconnaissance reports highlight that economic losses
related to damage of non-structural components far exceed those related
to damage of structural members [16]. Non-structural damage measures
are related to storey drifts and total floor accelerations. Conventional sys-
tems or systems with steel energy dissipation devices cannot provide
harmonization of structural and non-structural damage since drift (and
plastic deformation) reduction and total floor acceleration reduction
are competing objectives, i.e., adding strength and stiffness to the system
reduces drifts and plastic deformations but increases total accelerations
[17]. Recent works showed that drifts and total accelerations can be si-
multaneously controlled by inserting rate-dependent dampers (viscous,
viscoelastic or elastomeric dampers) into steel MRFs with strength
lower than that of conventional steel MRFs [18–20].
An effective seismic design approach is to use steel energy dissipa-
tion devices and rate-dependent dampers in parallel [21]. In that case,
steel devices provide hysteretic energy dissipation and protect the
main structural members from yielding, while rate-dependent

0143-974X/$ – see front matter © 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jcsr.2011.10.010
 T.L. Karavasilis et al. / Journal of Constructional Steel Research 70 (2012) 358–367
dampers provide supplemental viscous damping and drift response
reduction. However, more work is needed to apply this design
approach to realistic building cases and to investigate the applicability
of cost-effective energy dissipation devices [8–10]. In addition, the de-
sign of the steel devices, rate-dependent dampers and steel frame
should be integrated to provide effective combinations of stiffness,
strength and damping, so that drifts, residual drifts, plastic deforma-
tions and total floor accelerations can be simultaneously controlled [17].
To evaluate the aforementioned seismic design approach with non-
linear dynamic analysis, robust models for steel devices and rate-
dependent dampers are needed. Accurate models have been proposed
for rate-dependent fluid viscous, viscoelastic and elastomeric dampers
[22]. The simple bilinear elastoplastic and the smooth Ramberg–Osgood
model have been modified to capture the combined kinematic and iso-
tropic hardening of steel devices [6,13]. The smooth Bouc–Wen model
[23] has been extensively used to model steel components under cyclic
loading, since it can be modified to capture highly asymmetric hystere-
sis, stiffness deterioration, strength deterioration and pinching [24–27].
However, the Bouc–Wen model should be further enhanced to capture
the pronounced combined kinematic and isotropic hardening in the
hysteresis of steel energy dissipation devices [3–15] in order to
accurately predict the peak device force under cyclic seismic loading.
This is particularly important when evaluating seismic design methods
that use capacity design rules to avoid yielding in structural members in
the force path of steel devices.
This paper evaluates a minimal-damage seismic design approach
using steel energy dissipation devices and viscous dampers in parallel.
A model for steel energy dissipation devices is proposed and calibrated
against experimental results. In particular, the Bouc–Wen model is
modified to capture the combined kinematic and isotropic hardening
seen in the hysteresis of steel devices. The model is found able to accu-
rately predict the experimentally obtained hysteresis and is implemen-
ted in the OpenSees [27] software for use in seismic response analysis. A
simplified seismic design procedure is proposed and used to design a
prototype steel building equipped with steel devices and viscous
dampers according to explicitly defined minimal-damage performance
objectives. Results of seismic analyses indicate the accuracy of the
design procedure and confirm that the building is able to achieve imme-
diate occupancy (IO) under the design seismic action and rapid return
to occupancy (RRO) under the maximum considered seismic action.
The same building is designed as a conventional steel MRF according
to EC8 [1]. Seismic analyses results indicate that repair of damage in
main structural members of the conventional MRF may not be financial-
ly viable in the aftermath of the design and maximum considered
earthquakes.
2. Proposed model for steel energy dissipation devices
The standard Bouc–Wen model [23] results from the parallel combi-
nation of an elastic component and an elastic-perfectly plastic compo-
nent. The force output F of the model is
F ¼ pku þ ð1−pÞF y z
where u is the deformation across the model, Fy the yield force, k the
elastic stiffness, p the post-yield stiffness ratio, and z a dimensionless
hysteretic parameter governed by
k

z_ ¼
n
_ Þ þ γÞ
u_ 1−jzj ðβsgnðuz
Fy
where β and γ are parameters controlling the shape of the hysteresis, n
is a parameter that controls the sharpness of the smooth transition from
the elastic to the inelastic region of the hysteresis, sgn() is the signum
function, and the overdot denotes derivative with respect to time.
359
Eq. (1) shows that the Bouc–Wen model accounts for kinematic
hardening (i.e., post-yield force increases with increasing deformation)
due to the post-yield stiffness ratio p. However, the model does not
account for the isotropic hardening (i.e., yield force Fy increases due to
cyclic inelastic deformation) in the hysteresis of steel energy dissipation
devices [3–15].
2.1. Mathematical formulation
To incorporate isotropic hardening in the Bouc–Wen model, the
yield force Fy needs to be updated by considering the history of the
imposed cyclic deformation u. Examination of the constitutive Eqs. (1)
and (2) reveals that a change in the yield force Fy can be achieved by ap-
propriately including a third shape control parameter Ф in Eq.(2):
k
n 
z_ ¼ _ Þ þ γ− Φsgnðu_ ÞðsgnðzÞ þ sgnðu_ ÞÞ :
u_ 1−jzj βsgnðuz ð3Þ
Fy
The parameter Ф quantifies isotropic hardening and is calculated
using functions that cause Ф to increase exponentially with increasing
cumulative plastic deformation upl,c, i.e.,
"  !#
u 
 pl;c 
Φp ¼ Φmax;p 1−exp −pΦ;p   ð4:aÞ
 uy 
or
"  !#
u 
 pl;c 
Φn ¼ Φma;sn 1−exp −pΦ;n   ð4:bÞ
 uy 
where uy (=Fy / k) is the yield deformation, pФ,p and pФ,n parameters
that control the isotropic hardening rate due to cumulative plastic
deformation, and, Φmax,p and Φmax,n the maximum possible values of
Φ for the fully saturated isotropic hardening condition, i.e., for upl.
c → ∞, Φp → Φmax,p and Φn → Φmax,n. On the other hand, when upl.
c = 0.0, Φp = 0.0 and Φn = 0.0.
Φp and Φn are used to independently capture isotropic hardening in
different loading directions (positive and negative). Typically, yielding
devices exhibit the same isotropic hardening in different loading direc-
tions [3–10], and therefore, Φmax,p = Φmax,n and pФ,p = pФ,n. However,
the model can simulate different isotropic hardening in different load-
ing directions (e.g., compressive and tensile loading in BRB hysteresis
[11–15]) by using different parameter values in Eqs. (4.a) and (4.b).
To understand the effect of Φ, consider the proposed Bouc–Wen
model with p = 0.0 (i.e., without kinematic hardening), n = 1 and
β + γ = 1, under a positive deformation increment, i.e., sgnðu_ Þ ¼ 1.
Assume that in the previous deformation increment, z has reached its
positive ultimate value zu, and therefore, sgn(z) = 1 and ż = 0.0.
For this case, Eq. (3) yields zu = 1 / (1–2Φ) and Eq. (1) yields
F = Fyzu = Fy / (1–2Φ). When Φ = 0.0 (i.e., without isotropic hardening),
F is equal to Fy / (1− 0.0)= Fy. When Φ ≠ 0.0 (e.g., Φ = 0.1), F is equal to
ð1Þ Fy / (1–2Φ) = Fy / (1–2 · 0.1) = 1.25Fy. In that case, Φ reflects a 25%
percent increase in the initial yield strength Fy due to isotropic harden-
ing. The term sgnðu_ Þ after the parameter Φ in Eq. (3) ensures that the
above calculations apply to the case of a negative deformation incre-
ment and a negative ultimate value of z.
The state determination procedure of the model requires as an input
the previous force and deformation, the previous z value and the
ð2Þ current deformation. The current value of the parameter Φ is then
calculated based on the following rules: Eq. (4.a) is used to update Φp
when the deformation increment changes from negative to positive
within the plastic region of the hysteresis; Eq. (4.b) is used to update
Φn when the deformation increment changes from positive to negative
within the plastic region of the hysteresis; Φ equals Φp when a positive
deformation increment occurs; and Φ equals to Φn when a negative
360 T.L. Karavasilis et al. / Journal of Constructional Steel Research 70 (2012) 358–367

deformation increment occurs. With the current Φ value known, Eq. (3)
is numerically integrated to obtain the current value of z which is simply
used in Eq. (1) to provide the current force output F of the model.
The state determination procedure is verified using the following
model parameters: Fy = 220 kN, k = 440 kN/mm, uy = Fy / k = 0.5 mm,
p = 0.0047, β = 0.59, γ = 0.41, n = 1, Φmax,p = Φmax,n = 0.2844 and pФ,
p = pФ,n = 0.0135. The imposed cyclic deformation history u (in mm)
is: 0.0 → +6.0 → −6.0 → +9.5. Fig. 1a and b show the z–u and F–u
hysteresis from state determination along with numerical integration
of Eq. (3) using a Newton–Raphson scheme. The plastic deformations
(upl,1 and upl,2 in Fig. 1b) are calculated with respect to the yielding
deformations uy,1⁎ (=uy = 0.5 mm) and uy,2⁎(=4.9 mm) where the
extrapolated lines of the elastic and the plastic region of the F–u hyster-
esis meet. To verify the hysteresis in Fig. 1a and b, Eq. (3) is analytically
solved for the different cases shown in Fig. 1a as follows:
Case 1 with z > 0 and u>0 _ (points 1 to 2): Φ = Φp = 0.0; the initial
conditions are uinit =0.0 mm and zinit =0.0; the analytical solution of
Eq. (3) is z(u)=(1/(1–2Φ))(1−−-exp(−(1–2Φ) (u−uinit)/uy));
and z(6)=1.0.
Case 2 with z > 0 and ub0 _ (points 2 to 3): uy,1⁎ = uy = 0.5 mm; upl,
c = upl,1 = 6− 0.5 = 5.5 mm; Φ = Φn = 0.0392; the initial conditions
are uinit =6 mm and zinit =1.0; the analytical solution of Eq. (3) is z
(u)=−5.56+6.56exp(0.18(u−uinit)/uy)); and z(u)=0.0 at
u=5.54 mm.
Case 3 with z b 0 andub0_ (points 3 to 4): Φ = Φn = 0.0392; the ini-
tial conditions are uinit = 5.54 mm and zinit = 0.0; the analytical
solution of Eq.(3) is z(u)=(1/(1–2Φ))(−1+exp((1–2Φ)
(u−uinit)/uy)); and z(−6)=−1.09.
_
Case 4 with z b 0 andu>0 (points 4 to 5): uy,2⁎ = 4.9 mm; upl,c = upl,1
+ upl,2 = 5.5 + (4.9 + 6) = 16.4 mm; Φ = Φp = 0.1017; the initial
conditions are uinit = −6 mm and zinit = −1.09; the analytical solu-
tion of Eq. (3) is z(u) = 5.56–6.65exp(−0.18(u − uinit) / uy); and z
(u) = 0.0 at u = −5.5 mm.
Case 5 with z > 0 andu>0_ (points 5 to 6): Φ = Φp = 0.1017; the ini-
tial conditions are uinit = −5.5 mm and zinit = 0.0; the analytical
solution of Eq. (3) is the same with that of Case 1; and z(9.5) = 1.26.
The above analytical solutions can be used to verify the z–u and F–u
hysteresis. As an example, consider point A in Fig 1a and b. Inputting
a 2
1 2
u =6.0
u =-5.5
z =0.0
z =1.0
0 5 1
z
u=−5.8 mm in the analytical z(u) solution for Case 4 gives z(−5.8)=
−0.63, and hence, Eq. (1) gives F=−149 kN for u=−5.8 and z=
−0.63.
2.2. Model calibration against experimental results
The parameters of the proposed model are determined from avail-
able characterization test data on different steel energy dissipation de-
vices. An unconstrained nonlinear minimization method embedded in
MATLAB [28] is used to minimize the root mean square error (RMS) be-
tween the measured experimental force (Fexp) in the test and the force F
from the model. The RMS is calculated as
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
uN  2
uP
u
ui¼1 F i −F exp;i
RMS ¼ u
u ð5Þ
t PN
F 2i
i¼1
where N is the number of the experimental data points for a specific
characterization test. Eq. (3) is numerically integrated using a New-
ton–Raphson scheme. Fig. 2 shows characterization test data and results
from the model for low-yield steel shear panels [5], slit steel devices [8],
and a BRB [15]. Table 1 provides information for the test specimens, the
optimum model parameters and the corresponding RMS values that
indicate the accuracy of the model. The difference in the hysteretic
behaviors of the various steel devices is reflected in the values of the
model parameters in Table 1. For example, low-yield shear panels
show significant isotropic hardening (Фmax = 0.284), while BRBs exhib-
it lower isotropic hardening in tension (Фmax,p = 0.145) than in com-
pression (Фmax,n = 0.195).
3. Cost-effective steel energy dissipation devices
BRBs [11–15] or the ADAS [3] and TADAS [4] energy dissipation
devices require careful manufacturing processes compared to that of
cost-effective devices such as the shear panels [5–7] or the devices pro-
posed in Refs. [8–10]. This work examines the applicability of cost-
effective steel slit devices [8,9] for designing a steel building according
to minimal-damage seismic performance objectives. However, the
same design procedure and concept applies to any type of steel energy
dissipation device.
The slit device [8,9] is fabricated from a short length of a standard I
steel section with a number of slits cut from the web, leaving a number
u=9.5
of strips between the two flanges to deform in flexure and dissipate
z=1.26 6 energy by forming plastic hinges at their ends (Fig. 3). As shown in
Fig. 3, the variables involved in the design of the device are the strip
length l0, strip width b and web thickness t. Based on the analysis

A u =-5.8
u =0.0
z =0.0
3 presented in Ref. [8], the yield strength Py of the device is equal to
z =-0.63 u =5.54
-1 4 z =0.0
u =-6.0
z =-1.09 nst σ y tb2
-2
-10 -5 0 5 10
P y ¼ cy ð6Þ
2l0
u (mm)
where nst is the number of strips in the device, σy is the yield strength of
b upl,1=5.5 the material and cy is a correction factor to be determined from experi-
400
*
mental results. In addition, the elastic stiffness ke of the device is calcu-
200 uy,1 =0.5
lated as
F (kN)

0
A 3
u =-5.8
nst Etb
-200 F=-150 *
uy,2 =4.9 ke ¼ ck ð7Þ
upl,2 =10.9
l30
-400
-10 -5 0 5 10
u (mm) where E is the Young's modulus and ck is a stiffness correction factor to
be determined from experimental results.
Fig. 1. Analytical verification of the (a) z–u; and (b) F–u hysteresis of the proposed By equating the calibrated values of k and Fy of the proposed model
model under cyclic deformation. (Table 1) with ke and Py calculated from Eqs. (6) and (7) (the geometry
 T.L. Karavasilis et al. / Journal of Constructional Steel Research 70 (2012) 358–367 361

a b 600
400
Test Test
Model Model
200 300

F (kN)
F (kN)
0 0

-200 -300
SP-6 SP-9
-400 -600
-15 -10 -5 0 5 10 15 -20 -10 0 10 20
u (mm) u (mm)

c 800
d 40
Test Test
Model Model
400 20
F (kN)

F (kN)
0 0

-400 -20
SL-3
SP-12
-800 -40
-20 -10 0 10 20 -15 -10 -5 0 5 10 15
u (mm) u (mm)

e 40
f 40
Test Test
Model
20 Model 20
F (kN)
F (kN)

0 0

-20 -20
SL-8 SL-9
-40 -40
-15 -10 -5 0 5 10 15 -15 -10 -5 0 5 10 15
u (mm) u (mm)

g 2000
Test
1000 Model
F (kN)

-1000
BRB
-2000
-100 -50 0 50 100
u (mm)

Fig. 2. Results from the proposed model and characterization test data for: (a–c): low-yield steel shear panels (SP-6, SP-9 and SP-12 in Ref. [5]); (d–f) slit steel devices (SL-3, SL-
8 and SL-9 in Ref. [8]); and (g): a buckling-restrained brace (BRB in Ref. [15]).

of the devices is provided in Ref. [8]), the mean values of the correction
factors ck and cy were found equal to 0.22 and 1.45 respectively.
Eq. (6) provides the force level at which the slit device yields. How-
ever, the ultimate strength, Fu, of the device is needed to enable reliable
capacity design of the main structural members in the force path of the
slit device. The mean value of the ratio of the ultimate device strength Pu
from experimental results to the yield strength Py from Eq. (6), Pu/Py,
was found equal to 1.32. Another design parameter is the ultimate
deformation capacity uult of the device against fracture under cyclic
loading. Based on the test results presented in Ref. [8], uult is approxi-
mately and conservatively considered equal to 35uy.
4. Prototype building
Fig. 4a shows the plan view of the 5-storey, 3-bay by 3-bay proto-
type steel office building used for the study. The building has two iden-
tical perimeter 3-bay steel MRFs (one at each side) to resist lateral
forces in the N–S direction. The design study focuses on one perimeter

Table 1
Model parameters calibrated from experimental results.

Specimen Fy (kN) k (kN/mm) p β γ n Φmax pΦ RMS

Slit device SL-3 in Ref. [8] 19.4 9.8 0.040 0.90 0.10 1.0 0.1094 0.0130 0.07
Slit device SL-8 in Ref. [8] 21.5 8.7 0.036 0.83 0.17 1.0 0.0573 0.0698 0.11
Slit device SL-9 in Ref. [8] 25.0 8.5 0.017 0.90 0.10 1.0 0.0555 0.0705 0.10
Average for SL-3,SL-8 and SL-9 – – 0.031 0.88 0.12 1.0 0.0741 0.0511 –
Shear panel SP-6 in Ref. [5] 126.1 338.1 0.008 0.57 0.43 1.0 0.2843 0.0136 0.12
Shear panel SP-9 in Ref. [5] 224.7 441.9 0.005 0.56 0.44 1.0 0.2844 0.0135 0.08
Shear panel SP-12 in Ref. [5] 293.2 471.3 0.006 0.62 0.38 1.0 0.2838 0.0104 0.08
BRB in Ref. [15] 1050.0 93.5 0.017 0.84 0.16 1.0 0.1449 ⁎0.1949 0.1000 ⁎0.1500 0.13
⁎ Parameter value for Eq. (3.b).
362 T.L. Karavasilis et al. / Journal of Constructional Steel Research 70 (2012) 358–367
Loading direction
t b l0
Fig. 3. Slit steel device designed and tested in Ref. [8].
MRF. This MRF is designed either as a conventional MRF or as an MRF
with slit devices and viscous dampers in order to compare their seismic
response. The slit devices are supported by braces and connected to the
bottom flange of the beam of the steel MRF. The viscous dampers are
inserted in an interior gravity frame (with pin connections) of the build-
ing. The MRF with slit devices and the gravity frame with viscous
dampers are coupled due to the floor diaphragm and form a hybrid
lateral-load resisting system, referred to herein as the steel MRF with
slit devices and viscous dampers, which is shown in Fig. 4b.
The dead and live gravity loads considered in the design are selected
according to EC1 [29]. The design seismic action, referred to herein as
design basis earthquake (DBE), has a return period equal to 475 years
and is expressed by the Type 1 elastic response spectrum of the EC8
[1] with peak ground acceleration equal to 0.3 g and ground type B (av-
erage shear wave velocity between 360 and 800 m/s.). The program
SAP2000 [30] is utilized for designing the steel MRFs according to EC3
[31] and EC8 [1]. S275 and S235 [31] steel grades are used for the
main structural members and slit devices respectively.
5. Design of conventional steel MRF
The SAP2000 model used for design is based on the centerline di-
mensions of the perimeter MRF without accounting for the finite
panel zone dimensions. A “lean-on” column is included in the
SAP2000 model to account for the P–Δ effects of the vertical loads acting
on the gravity columns in the tributary plan area (half of the total plan
area) assigned to the perimeter MRF. This column is pinned at its base,
continuous over the height of the building, and carries the vertical loads
acting on the gravity columns. The cross-section area and flexural stiff-
ness of the lean-on column is based on the summation of the areas and
flexural stiffnesses of the gravity columns. The column base and beam-
to-column connections of the MRF are assumed to be fully rigid. A rigid
diaphragm constraint is imposed at the nodes of each floor level.
The perimeter MRF is designed as a conventional MRF using the
modal response spectrum analysis procedure of EC8. The MRF satisfies
Ductility Class High [1] by using compact Class 1 cross-sections [31]
and therefore, the behavior (or “strength reduction”) factor is equal to
q = 5αu / α1 = 6.5, where αu/α1 is the overstrength factor with a recom-
mended value of 1.3 for multi-bay multi-storey steel MRFs. The
displacement behavior (or “displacement amplification”) factor is
Interior gravity frame
with viscous dampers
a b
MRF with slit devices
3@8m
Slit device
N
S
3@8 m
Fig. 4. Prototype building structure: (a) plain view and (b) perimeter steel MRF with slit devices coupled with the interior steel gravity frame with viscous dampers (referred to as
steel MRF with slit devices and viscous dampers).
equal to q, i.e., EC8 adopts the equal displacement rule to estimate
peak inelastic drifts. These drifts are then used to check second order
(P–Δ) effects. Additionally, EC8 imposes a serviceability limit on the
peak storey drift, θmax, under the frequently occurring earthquake
(FOE) with a return period equal to 95 years. The FOE has intensity of
40% (reduction factor v = 0.4 [1]) the intensity of the DBE and the asso-
ciated θmax is equal to 0.75%, assuming ductile non-structural compo-
nents. The MRF satisfies the strong column–weak beam capacity
design rule of EC8, the beam-to-column connections are designed to
be fully rigid and the panel zones are strengthened with doubler plates
to avoid yielding. A strength-based design with q = 6.5 under the DBE
was first performed. However, beams and columns from strength-
based design had to be increased to satisfy the serviceability drift re-
quirements under the FOE. The final sections were found iteratively
by decreasing the value of q, designing the MRF for increased strength
(and hence increased stiffness) and then checking drifts under the FOE.
6. Design of steel MRF with slit steel devices and viscous dampers
6.1. Definition of minimal-damage performance objectives
The following minimal-damage performance objectives along with
explicit structural and non-structural limit states are defined for design-
ing the steel MRF with slit devices and viscous dampers:
6.1.1. Immediate occupancy (IO) under the DBE
The target θmax is set equal to 1% so that drift-sensitive non-structural
elements have little or no damage. According to EC8, drift-sensitive
non-structural elements designed not to interfere with structural defor-
mations can avoid damage for θmax =1%. There is no consensus for defin-
ing a residual story drift, θr, target value associated with IO. However, the
target θr is set equal to the global sway imperfections defined in EC3 [31],
equal to 0.264% for the geometry of the prototype building in Fig. 4. Based
on the results presented in Ref. [17], the residual storey drift of the steel
MRF with slit devices and viscous dampers is conservatively estimated
equal to 0.15% of θmax, i.e., θr =0.15∙1%=0.15% that is lower than the
target 0.264% value and can be easily straightened in the aftermath of
the DBE by replacing damaged slit devices. In addition, beams and
columns of the MRF can be damage-free for θmax =1.0%. In particular,
previous research on steel frames with steel devices or rate-dependent
dampers has shown that beams and columns can be designed to be essen-
tially elastic for θmax values of the order of 1.5% [13,18,19].
6.1.2. Rapid return to occupancy under the maximum considered earthquake
(MCE)
The MCE has intensity of 150% the intensity of the DBE, and hence,
θmax = 1.5∙1%= 1.5%. Main structural members can be designed to be-
have elastically for θmax = 1.5% [13,18,19]. However, drift-sensitive
non-structural elements will sustain controlled damage that should be
repaired for θmax = 1.5%. The residual storey drift is estimated equal to
Floor diaphragm
4@3.2m
Viscous
damper
4m
3@8 m 8m
 T.L. Karavasilis et al. / Journal of Constructional Steel Research 70 (2012) 358–367 363

Table 2
Design details of the conventional steel MRF and the steel MRF with slit devices and viscous dampers.

Conventional steel MRF Steel MRF with slit devices and viscous dampers

Storey Column Beam T Steel θmax FOE Column Beam Damper constant Slit device Slit device T Steel θmax FOE
(s.) weight (kN) DBE MCE c {kN(s./mm)0.5} geometry: t/b/l0 (mm) number of strips nst (s.) weight (kN) DBE MCE
1 HEB400 IPE450 1.70 180 HEB280 IPE270 33.2 15/66/440 13 1.50 124
2 HEB400 IPE450 0.70% HEB280 IPE270 38.0 15/53/350 15 0.40%
3 HEB400 IPE400 1.75% HEB280 IPE270 34.0 15/53/350 13 1.00%
4 HEB360 IPE400 2.63% HEB240 IPE240 25.1 15/53/350 10 1.50%
5 HEB360 IPE360 HEB240 IPE240 19.1 15/53/350 8

θr = 0.15∙1.5%= 0.225% [17]. This value is lower than the previously
defined target θr value associated with IO (i.e., 0.264%) and can be
straightened by replacing damaged slit devices. In addition, slit devices
can be designed to avoid fracture under the deformations associated
with θmax = 1.5%.
Peak total floor accelerations should also be controlled to avoid
having toppled or dropped non-structural elements and dysfunctional
acceleration-sensitive equipment. The results presented in Ref. [17]
showed that drifts and total accelerations can be simultaneously
controlled by providing supplemental damping to systems with low
strength. Therefore, the steel MRF with slit devices should be designed
for low strength, and then, viscous dampers should be designed to
reduce drifts at the target θmax, equal to 1.0% under the DBE.
They also satisfy the strong column–weak beam capacity design rule of
EC8. Beam-to-column connections are designed to be fully rigid and
panel zones are strengthened with doubler plates [1]. Under the DBE,
the MRF with slit devices has θmax = 1.5%.
With the period of vibration of the steel MRF with slit devices
known, nonlinear viscous dampers can be designed for the interior
braced frame to provide the desired drift reduction. In particular,
nonlinear viscous dampers are designed to provide supplemental
viscous damping ratio ξs equal to 18% according to
 1þa
∑j ð2πÞa T 2−a λj cj ura−1 ϕj −ϕj−1
ξs ¼ ð8Þ
8π3 ∑j mj φ2j

6.2. Structural analysis and design
The steel MRF with slit devices was first designed without consider-
ing the effect of viscous dampers. The SAP2000 model used for design
includes the members of the steel MRF, the supporting braces (pinned
connected) and a horizontal spring representing the stiffness of the
slit device. The “lean-on” column (see discussion in Section 5) includes
the increased stiffness of the columns of the gravity frame with
dampers. These columns are designed to sustain the peak viscous
damper force and their effect is considered iteratively since viscous
dampers are designed at the end of the design process.
The steel MRF with slit devices is designed using a q factor equal to
6.5 and without any drift requirements so that it has less strength
than the conventional MRF. The sections of the latter had to be signifi-
cantly increased compared to those obtained using q = 6.5 to satisfy
the FOE drift requirements. The design starts by selecting values of the
spring stiffnesses at each storey, so that they produce a uniform distri-
bution of storey drifts based on modal response spectrum analysis.
The slit devices at each storey are designed using Eqs. (6) and (7) to pro-
vide stiffness ke equal to the spring stiffness and ultimate strength Pu
(=1.32Py; see Section 3) equal to the force in the spring. Braces are
sized to be stiff enough so that storey drift produces slit device deforma-
tion rather than brace deformation. For all stories, a ratio of 10 is
adopted for the total brace horizontal stiffness per storey to slit device
stiffness. Beams and columns are designed according to capacity design
rules so that they do not yield under the slit device ultimate strength Pu.
0.2
Conventional MRF
MRF with slit devices
0.1
V/W
where j denotes a specific storey of the MRF, a is the velocity exponent
of the nonlinear viscous dampers (selected to equal 0.5 to control the
peak damper force [22]), Tis the fundamental period of vibration, λjis
a dimensionless parameter, cjis the damper constant, uris the amplitude
of the roof displacement, mjis the storey mass, and φjis the modal coor-
dinate of the first mode shape [32]. Adding 18% damping to the inherent
2% damping of the MRF provides a damping reduction factor equal to
1.5 [33], and therefore, the θmax under the DBE is reduced to 1.5%/
1.5 = 1.0%, which is equal to the target θmax under the DBE (Section 6.1).
The above simple design procedure used to size nonlinear viscous
dampers is based on previous works showing that (a) the use of
undamped frequencies and mode shapes along with an energy-based
calculation of the supplemental damping ratio can provide good esti-
mates of the peak displacement response of highly damped structures
[32]; and (b) the equal displacement rule is valid for highly damped in-
elastic structures [17].
Different cj distributions along the height of the frame can provide
the same supplemental damping ratio according to Eq. (8). Viscous
dampers are designed to satisfy cj = εKo,i, where Ko,i is the horizontal
storey stiffness of the steel MRF with slit devices and ε is a constant
that is directly obtained by substituting cj = εKo,i into Eq. (8). Stiff braces
are used to support the viscous dampers, while beams and columns of
the gravity frame are designed so that they do not buckle under the
peak viscous damper force. The latter is approximately estimated
using the procedure described in Refs. [32,33].
7. Design details and nonlinear cyclic static analysis
Table 2 compares properties of the conventional MRF and the MRF
with slit devices and viscous dampers. The table lists the column
cross-sections, beam cross-sections, steel weight, fundamental period
of vibration (T1) and estimates of θmax under the FOE, DBE and MCE.

0
The steel MRF with slit devices has less steel weight and significantly
lower θmax than the conventional steel MRF. Table 2 also provides the
-0.1 First yielding in beams
of the steelMRFwith slit properties of the designed slit devices and viscous dampers that have
devices
cost-effective practical sizes.
-0.2
-0.03 -0.02 -0.01 0 0.01 0.02 0.03 Fig. 5 compares the base shear coefficient (V/W)—roof drift (θrf)
θrf behavior (pushover curves) of the conventional MRF and the MRF
with slit devices from nonlinear cyclic (push–pull) static analysis
Fig. 5. Base shear coefficient-roof drift behavior from nonlinear cyclic (push–pull) static using analytical models described in Section 8.2. V is the base shear
analysis. force and W is the seismic weight. The first cycle of the static analysis
364 T.L. Karavasilis et al. / Journal of Constructional Steel Research 70 (2012) 358–367

Table 3
Properties of the ground motions used for nonlinear dynamic analyses.

Earthquake Station Component Magnitude Distance Scale factor

(Mw) (km)
FOE DBE MCE

Imperial Valley 1979 Cerro Prieto H-CPE237 6.53 15.19 0.82 2.05 3.08
Loma Prieta 1989 Hollister-S & P HSP000 6.93 27.67 0.29 0.72 1.08
Loma Prieta 1989 Woodside WDS000 6.93 33.87 1.40 3.49 5.24
Loma Prieta 1989 WAHO WAH090 6.93 17.47 0.48 1.20 1.80
Manjil 1990 Abbar ABBAR—T 7.37 12.56 0.28 0.70 1.05
Cape Mendocino 1992 Fortuna—Fortuna Blvd. FOR000 7.01 15.97 0.99 2.47 3.71
Cape Mendocino 1992 Rio Del Overpass—FF RIO360 7.01 14.33 0.50 1.25 1.88
Landers 1992 Desert—Hot Springs LD-DSP000 7.30 21.78 0.95 2.37 3.56
Northridge 1994 LA—W. 15th St. W15090 6.69 25.60 1.14 2.86 4.29
Northridge 1994 Moorpark—Fire Sta. MRP180 6.69 16.92 0.78 1.94 2.91
Northridge 1994 N. Hollywood—Cw CWC270 6.69 7.89 0.53 1.33 2.00
Northridge 1994 Santa Susana Ground 5108-360 6.69 1.69 0.78 1.95 2.93
Northridge 1994 LA—Brentwood VA 0638-285 6.69 12.92 0.85 2.12 3.18
Northridge 1994 LA—Wadsworth VA 5082-235 6.69 14.55 0.62 1.54 2.31
Kobe 1995 Nishi-Akashi NIS090 6.90 7.08 0.48 1.19 1.79
Kobe 1995 Abeno ABN090 6.90 24.85 1.00 2.49 3.74
ChiChi 1999 TCU105 TCU105-E 7.62 17.18 0.96 2.39 3.59
ChiChi 1999 CHY029 CHY029-N 7.62 10.97 0.53 1.32 1.98
ChiChi 1999 CHY029 CHY041-N 7.62 19.83 0.56 1.40 2.10
Hector 1999 Hector HEC090 7.13 10.35 0.42 1.04 1.56

is performed at θrf equal to 1.5% while the next 2 cycles are performed at
θrf equal to 2.5%. For the MRF with slit devices, softening begins at a low
θrf due to early yielding of the slit devices. Beams and columns of the
MRF with slit devices deform elastically for θrf lower or equal to 1.5%
and therefore, avoid damage under both the DBE and MCE (see θmax es-
timates in Table 2); consistent with the minimal-damage performance
objectives in Section 6.1. The steel MRF with slit devices exhibits higher
hysteretic energy dissipation capacity than the conventional MRF under
both low and high drifts. Moreover, Fig. 5 shows that the MRF with slit
devices has lower strength than the conventional MRF for θrf equal to
1.0% (DBE level); consistent with the minimal-damage design approach
(Section 6.1) aiming at reduced strength to control peak total floor ac-
celerations. For roof drifts higher than 2%, both frames have the same
strength. For the steel MRF with slit devices, a 2% drift is expected
under a seismic intensity of 133% of the intensity of the MCE and may
jeopardize the deformation capacity of the slit devices. In that case,
the back-up steel MRF members would provide adequate resistance
and deformation capacity since they exhibit mild inelastic deformations
at 2% drift.
8. Nonlinear dynamic analyses
8.1. Ground motions
An ensemble of 20 earthquake ground motions recorded on ground
type B are used in 2D nonlinear dynamic analyses to evaluate the per-
formance of the conventional steel MRF and the performance of the
steel MRF with slit devices and viscous dampers. None of the ground
motions exhibit near-fault forward-directivity effects. The ground
motions are scaled to the DBE level using the scaling procedure of Som-
erville [34]. Table 3 provides the scale factors and information on the 20
ground motions. Fig. 6 compares the DBE elastic response spectrum of
EC8 with the mean (μ) and mean plus/minus one standard deviation
(μ ± σ) spectra of the DBE ground motions. The amplitudes of the DBE
ground motions are further scaled by 0.4 and 1.5 to represent FOE and
MCE ground motions.
8.2. Modeling of MRFs for nonlinear dynamic analysis
2D nonlinear analytical models of the conventional MRF and the
MRF with slit devices and viscous dampers are developed for nonlinear
dynamic analysis using OpenSees [27]. A diaphragm constraint is
imposed on the nodes of each floor level. The gravity frame with viscous
dampers is modeled and coupled (due to the floor diaphragm; as shown
in Fig. 4b) with the steel MRF with slit devices. A distributed plasticity
force-based beam column is used to model beams, columns and braces.
An elastic beam column element is used to model the lean-on column
that accounts for the P–Δ effects of the vertical loads on the interior
gravity columns of the prototype building. The panel zones of the
beam–column joints are modeled as proposed by Herrera et al. [35].
The proposed model for steel energy dissipation devices described in
Section 2 is implemented in OpenSees and used to model the slit de-
vices. The average values of the model parameters (p, β, γ, n, Фmax,
pФ) given in Table 1 are used in the analysis, while Eqs. (6) and (7)
are used to estimate the stiffness, k, and yield strength, Fy, of the slit de-
vices in Table 2. Nonlinear viscous dampers are modeled using dashpots
with a force output [33]
a
f d ¼ cjvj sgnðvÞ ð9Þ

2
where v is the velocity across the damper.
EC8 DBE The Newmark method with constant acceleration is used to inte-
1.5 μ
grate the equations of motion. The Newton method with tangent stiff-
μ+σ
Sa (g)

1
ness is used to minimize the unbalanced forces at each step of the
μ-σ
nonlinear dynamic analysis. The integration time step is selected
0.5 equal to 0.0005 s since it has been shown that a particularly small inte-
gration time step is needed to accurately predict floor accelerations
0 [36]. Smaller time steps provided similar response results. A Rayleigh
0 1 2 3 4
damping matrix is used to model the inherent 2% critical damping at
Period (s.)
the first two modes of vibration. A nonlinear load controlled static anal-
Fig. 6. Comparison of the design EC8 [1] spectrum with the spectra of the ground motions ysis under the gravity loads is first performed, and then, the nonlinear
used for nonlinear dynamic analyses. dynamic earthquake analysis is executed. Each dynamic analysis is
 T.L. Karavasilis et al. / Journal of Constructional Steel Research 70 (2012) 358–367 365

a b
0.03 0.03
MRF with slit dev./visc. damp. MRF with slit dev./visc. damp.
0.015 Conv. MRF 0.015 Conventional MRF

0 0
θrf

θrf
-0.015 -0.015

DBE MCE
-0.03 -0.03
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Time (s.) Time (s.)

Fig. 7. Comparison of roof drift time histories under the HSP ground motion scaled to the (a) DBE and (b) MCE.

extended well beyond the actual earthquake time to allow for damped
free vibration decay and correct residual drift calculation.
8.3. Seismic response results
Fig. 7 compares the roof drift time histories of the conventional steel
MRF and the MRF with slit devices and viscous dampers under the HSP
ground motion scaled to the DBE and MCE (Table 3). Near the end of the
time histories the MRF with slit devices and viscous dampers oscillates
around the origin, indicating negligible residual drift, while the conven-
tional MRF experiences large residual drifts. Fig. 8 shows the slit device
and nonlinear viscous damper hysteresis in the third storey of the build-
ing under the HSP ground motion scaled to the DBE and MCE. The non-
linear viscous dampers develop a higher peak force than that of the slit
device, especially under the MCE.
Fig. 9a shows μ and μ + σ values of the peak storey drifts, θmax, under
the earthquake ground motions of Table 3 scaled to the FOE, DBE and
MCE. Both MRFs show uniform height-wise distributions of θmax. The
MRF with slit devices and viscous dampers exhibits significantly better
performance than the conventional MRF. The mean θmax values of the
steel MRF with slit devices and viscous dampers from dynamic analyses
are close to the minimal-damage target θmax values (1.0% under the DBE
and 1.5% under the MCE; Section 6.1).
Fig. 9b shows μ and μ + σ values of the residual storey drifts, θr. The
MRF with slit devices and viscous dampers exhibits significantly better
performance than the conventional MRF. The mean θr values of the steel
MRF with slit devices and viscous dampers from dynamic analyses are
close to the expected θr values (0.15% under the DBE and 0.225%
under the MCE; Section 6.1) and lower than the target θr value
associated with IO (0.264%; Section 6.1). On the other hand, the μ + σ
values under the DBE and the μ values under the MCE of the conven-
tional steel MRF are larger than the critical 0.5% θr value proposed in
Ref. [2]. θr values higher than 0.5% indicate that repair of damage in
main structural members of the conventional steel MRF may not be fi-
nancially viable in the aftermath of the DBE and MCE.
Fig. 9c shows μ and μ + σ values of the peak total floor accelerations
amax. Along with satisfying the low target θmax and θr, the MRF with slit
devices and viscous dampers experiences reduced peak total floor ac-
celerations compared to those of the conventional steel MRF. These re-
ductions indicate less damage in acceleration-sensitive non-structural
components and confirm the effectiveness of the proposed design strat-
egy that combines reduced strength with supplemental viscous
damping.
The aforementioned statistical seismic response results confirm the
accuracy of the proposed simplified design procedure and illustrate that
the building using the steel MRF with slit devices and viscous dampers
is able to achieve IO under the DBE and RRO under the MCE.
9. Summary and conclusions
This study evaluated a seismic design approach for steel structures
that concentrates damage in easy-to-replace steel energy dissipation
devices and protects the main structural members from yielding with
capacity design rules. This approach can reduce damage repair costs
and downtime, and, can be further enhanced by using rate-dependent
dampers in parallel to steel devices to achieve drift reduction and pro-
tection of drift-sensitive non-structural elements.

a b 500
500 Slit device Slit device
DBE MCE
250 250
F (kN)
F (kN)

0 0

-250 -250

-500 -500
-0.06 -0.03 0 0.03 0.06 -0.06 -0.03 0 0.03 0.06
u (m) u (m)

c d 500
500 Nonlinear viscous Nonlinear viscous
damper - MCE
damper - DBE 250
250
Fd (kN)
Fd (kN)

0 0

-250 -250

-500 -500
-0.06 -0.03 0 0.03 0.06 -0.06 -0.03 0 0.03 0.06
u (m) u (m)

Fig. 8. (a–b) Slit device and (c–d) nonlinear viscous damper hysteresis in the third storey of the minimal-damage steel MRF under the HSP ground motion scaled to the DBE and
MCE level.
366 T.L. Karavasilis et al. / Journal of Constructional Steel Research 70 (2012) 358–367

a 5 5 5
MRF slit dev./visc. damp. ( μ )
4 MRF slit dev./visc. damp. ( μ+ σ) 4 4

Storey
Storey

Storey
3 Conv. MRF ( μ ) 3 3
Conv. MRF ( μ+ σ)
2 2 2
1 1 1
FOE DBE MCE
0 0 0
0 0.01 0.02 0.03 0.04 0 0.01 0.02 0.03 0.04 0 0.01 0.02 0.03 0.04
θ max θ max θ max

b 5 5 5
4 4 4
Storey

Storey

Storey
3 3 3
2 2 2
1 1 1
FOE DBE MCE
0 0 0
0 0.005 0.01 0.015 0.02 0 0.005 0.01 0.015 0.02 0 0.005 0.01 0.015 0.02
θr θr θr

c 5 5 5
4 4 4
Storey

Storey
Storey

3 3 3
2 2 2
1 1 1
FOE DBE MCE
0 0 0
0 2.5 5 7.5 10 12.5 15 0 2.5 5 7.5 10 12.5 15 0 2.5 5 7.5 10 12.5 15
α max(m/s.2) α max(m/s.2) α max(m/s.2)

Fig. 9. Statistics of (a) peak storey drifts; (b) residual storey drifts; and (c) peak total floor accelerations of the conventional MRF and the MRF with slit devices and viscous dampers
under 20 ground motions scaled to the FOE, DBE and MCE.

A model for steel energy dissipation devices was proposed and
calibrated against existing experimental results. A simplified seismic
design procedure was proposed and used to design a prototype steel
building equipped with steel devices and viscous dampers according
to explicitly defined minimal-damage performance objectives. The
same building was designed as a conventional steel MRF according to
EC8. Extensive seismic analyses were conducted and drift, residual
drift and total floor acceleration response statistics were obtained.
Based on the results presented in this paper, the following conclu-
sions are drawn:
1) The versatile Bouc–Wen model can be modified to capture the
combined kinematic and isotropic hardening seen in the hysteresis
of steel energy dissipation devices. The model was found able to
accurately predict the experimentally obtained hysteresis of various
steel devices, including low-yield steel shear panels, slit steel devices
and buckling-restrained braces. In the context of nonlinear dynamic
or cyclic (push–pull) static analysis, this model can be used to
reliable evaluate seismic design methods using capacity design
rules to avoid yielding in structural members in the force path of
steel energy dissipation devices.
2) The accuracy of the proposed seismic design procedure for steel
buildings with energy dissipation devices and rate-dependent
dampers was confirmed with nonlinear dynamic analyses. The pro-
cedure is compatible with current seismic codes and starts with
explicitly defined minimal-damage performance objectives based
on target peak storey drift and residual storey drift values.
3) Results of seismic analyses indicated that the steel building with slit
devices and nonlinear viscous dampers is able to achieve immediate
occupancy under the design earthquake and rapid return to occupan-
cy under the maximum considered earthquake by avoiding yielding in
main structural members and by satisfying predifined minimal-
damage target peak storey drift and residual storey drift values. Resid-
ual storey drifts of the steel MRF with slit devices and viscous dampers
were found to be negligible under the DBE and lower than the global
sway imperfections defined in EC3 under the MCE.
4) Seismic analysis results indicated that repair of damage in the main
structural members of a conventional steel MRF designed according
to EC8 may not be financially viable in the aftermath of the design
and maximum considered earthquakes. Residual storey drifts of
the conventional MRF did not satisfy the critical 0.5% value that en-
sures reparability based on findings from past earthquakes.
5) Total floor accelerations can be effectively controlled while achieving
significant drift reductions by designing a structure to have reduced
strength (e.g., using low strength yielding devices) and supplemental
viscous damping. The steel MRF with slit devices and nonlinear
viscous dampers experienced lower drifts, residual drifts and total
floor accelerations than those of the conventional steel MRF.
6) Conventional seismic-resistant frames, such as a steel moment-
resisting frame, cannot provide harmonization of structural and
non-structural damage since drift (and plastic deformation) reduc-
tion and total floor acceleration reduction are competing objectives.
7) A height-wise stiffness and strength distribution of steel energy
dissipation devices based on force and drift demands from elastic
modal response spectrum analysis, combined with a storey stiffness
proportional distribution of nonlinear viscous dampers, were found
to produce uniform distributions of peak storey drifts and residual
storey drifts along the height of the prototype building under differ-
ent earthquake intensities.
8) Cost-effective steel energy dissipation devices were found to be ap-
plicable for minimal-damage seismic design of steel frames. Howev-
er, more work is needed to predict the fracture deformation capacity
of these devices. This is particularly important for evaluating the
performance of steel frames equipped with steel devices and rate-
dependent dampers under very strong earthquake ground motions
with intensity higher than the MCE.
References
[1] CEN. Eurocode 8: design of structures for earthquake resistance. Part 1: general
rules, seismic actions and rules for buildings; 2004.
 T.L. Karavasilis et al. / Journal of Constructional Steel Research 70 (2012) 358–367
[2] Mc Cormick J, Aburano H, Ikenaga M, Nakashima M. Permissible residual deforma-
tion levels for building structures considering both safety and human elements.
14th WCEE, Beijing, China; 2008.
[3] Whittaker AS, Bertero VV, Thompson CL, Alonso LJ. Seismic testing of steel plate
energy dissipation devices. Earthquake Spectra 1991;7(4):563–604.
[4] Tsai K, Chen H, Hong C, Su Y. Design of steel triangular plate energy absorbers for
seismic-resistant construction. Earthquake Spectra 1993;9(3):505–28.
[5] Nakashima M. Strain-hardening behavior of shear panels made of low-yield steel.
I: test. J Struct Eng, ASCE 1995;12:1742–9.
[6] Nakashima M, Akazawa T, Tsuji B. Strain-hardening behavior of shear panels
made of low-yield steel. II: model. J Struct Eng, ASCE 1995;121(12):1750–7.
[7] De Matteis G, Brando G, Mazzolani FM. Hysteretic behaviour of bracing-type pure
aluminium shear panels by experimental tests. Earthquake Engineering and
Structural Dynamics; 2010. available online.
[8] Chan R-WK, Albermani F. Experimental study of steel slit damper for passive energy
dissipation. Engineering Structures 2008;30:1058–66.
[9] Ghabraie K, Chan R-WK, Huang X, Xie Y-M. Shape optimization of metallic yield-
ing devices for passive mitigation of seismic energy. Engineering Structures
2010;32:2258–67.
[10] Benavent-Climent A, Morillas L, Vico JM. A study on using wide-flange section web
under out-of-plane flexure for passive energy dissipation. Earthquake Engineering
and Structural Dynamics 2011;40(5):473–90.
[11] Black CJ, Makris N, Aiken ID. Component testing, seismic evaluation and characteri-
zation of buckling-restrained braces. J Struct Eng, ASCE 2004;Vol. 130(6):880–94.
[12] Fahnestock LA, Ricles JM, Sause R. Experimental evaluation of a large-scale
buckling-restrained braced frame. J Struct Eng, ASCE 2007;133(9):1205–14.
[13] Fahnestock LA, Sause R, Ricles JM. Seismic response and performance of buckling-
restrained braced frames. J Struct Eng, ASCE 2007;133(9):1195–204.
[14] Vargas R, Bruneau M. Experimental response of buildings designed with metallic
structural fuses. II. J Struct Eng, ASCE 2007;135(4):394–403.
[15] Merritt S, Uang CM, Benzoni G. Subassemblage testing of star seismic buckling-
restrained braces. Structural Systems Research Project, Report No. TR-2003/04.
San Diego: University of California; 2003.
[16] Dolce M, Manfredi G. Research needs in earthquake engineering highlighted by
the 2009 L'Aquila earthquake. The ReLUIS-DPC Project; 2009.
[17] Karavasilis TL, Seo C-Y. Seismic structural and non-structural performance evalu-
ation of highly damped self-centering and conventional systems. Engineering
Structures 2011;33:2248–58.
[18] Karavasilis TL, Ricles JM, Sause R, Chen C. Experimental evaluation of the seismic
performance of steel MRFs with compressed elastomer dampers using large-scale
real-time hybrid simulation. Engineering Structures 2011;33(6):1859–69.
[19] Karavasilis TL, Sause R, Ricles JM. Seismic design and evaluation of steel MRFs with
compressed elastomer dampers. Earthquake Engineering and Structural Dynamics
2011, doi:10.1002/eqe.1136.
367
[20] Pavlou E, Constantinou MC. Response of nonstructural components in structures
with damping systems. J Struct Eng, ASCE 2006;132(7):1108–17.
[21] Vargas R, Bruneau M. Effect of supplemental viscous damping on the seismic
response of structural systems with metallic dampers. J Struct Eng, ASCE 2007;133
(10):1434–44.
[22] Symans MD, Charney FA, Whittaker AS, Constantinou MC, Kircher CA, Johnson
MW, McNamara RJ. Energy dissipation systems for seismic applications: current
practice and recent developments. J Struct Eng, ASCE 2008;134(1):3–21.
[23] Wen YK. Method for random vibration of hysteretic systems. J Eng Mech Div
1976;102(2):249–63.
[24] Song J, Der Kiureghian A. Generalized Bouc–Wen model for highly asymmetric
hysteresis. J Eng Mech 2006;132(6):610–8.
[25] Wang C-H, Wen Y-K. Evaluation of pre-Northridge low-rise steel buildings. I:
modeling. J Struct Eng, ASCE 2000;126(10):1160–8.
[26] Sivaselvan MV, Reinhorn AM. Hysteretic models for deteriorating inelastic structures.
J Eng Mech, ASCE 2000;126(6):633–40.
[27] Mazzoni S, McKenna F, Scott M, Fenves G. Open system for earthquake engineering
simulation (OpenSees). User Command Language Manual. University of California,
Berkeley: Pacific Earthquake Engineering Research Center 2006; 2006.
[28] MATLAB. A registered trademark of The Math Works, Inc.; 2011.
[29] CEN. EN 1991-1-1, Eurocode 1: actions on structures—part 1.1: general actions—
densities, self-weight, imposed loads for buildings. European Committee for
Standardization; 2002.
[30] SAP2000. Static and dynamic finite element analysis of structures. Version 9.1.4.
Berkeley, California, U.S.A.: Computers and Structures, Inc.; 2005
[31] CEN. EN 1993-1-1, Eurocode 3: design of steel structures—part 1.1: general rules
and rules for buildings. European Committee for Standardization; 2005.
[32] Whittaker AS, Constantinou MC, Oscar MR, Johnson MW, Chrysostomou CZ. Equiva-
lent lateral force and modal analysis procedures of the 2000 NEHRP Provisions for
buildings with damping systems. Earthquake Spectra 2003;19(4):959–80.
[33] Building Seismic Safety Council (BSSC). NEHRP recommended provisions for seismic
regulations for new buildings and other structures. Report FEMA 4502003 ed.
Washington, D.C: Federal Emergency Management Agency; 2001.
[34] Somerville P. Development of ground motion time histories for phase 2 of the
FEMA/SAC steel project; 1997. Report No. SAC/DB-97/04, Sacramento, CA.
[35] Herrera R, Ricles J, Sause R. Analytical studies of steel MRFs with CFT columns
under earthquake loading conditions. 4th International Conference on Behavior
of Steel Structures in Seismic Areas (STESSA), Naples, Italy; 2003. p. 519–25.
[36] Rodriguez MA, Restrepo JI, Blandon JJ. Shaking table tests of a four-story miniature
steel building—model validation. Earthquake Spectra 2006;22(3):755–80.