You are on page 1of 12

Seismic Design and Performance Evaluation of

Dual-Fused H-Frame System


Hamidreza Etebarian 1; T. Y. Yang 2; and Dorian P. Tung 3

Abstract: A dual-fused H-frame (DFHF) is an efficient structural system that combines damped H-frame (DHF) modules with welded
wide flange fuses (WWFFs) to create a structural solution that is efficient in construction and more seismically resilient. Each DHF module
Downloaded from ascelibrary.org by SUNY at Stony Brook on 10/01/19. Copyright ASCE. For personal use only; all rights reserved.

consists of two columns pin connected to a beam with two buckling restrained knee braces (BRKBs). Each DHF module can be prefabricated
at the factory, shipped to the site, and connected vertically using simple bolt connections. The connections between the DHF modules have
relatively small moment demand, which makes the design, fabrication, and construction of the DHF modules very efficient. Once the DHF
modules are assembled vertically, the bays of the DHF can be connected using WWFFs. WWFFs are simple shear connectors that can stably
dissipate earthquake energy. In this paper, two prototype DFHF buildings of varying heights (three and nine stories) are designed using the
equivalent energy design procedure (EEDP). The EEDP is a novel design method that was developed to design innovative systems, where the
structural system can satisfy different performance objectives under different earthquake shaking intensities. To verify the performance of
the DFHF, advanced finite-element models are developed using OpenSees and subjected to an extensive array of time history analyses. The
results show that the proposed EEDP designed DFHF can meet the targeted performance objectives under different seismic shaking inten-
sities. In addition, DFHF has a sufficient margin of safety against collapse. Hence, the proposed DFHF can be used as an efficient structural
system in high seismic zones. DOI: 10.1061/(ASCE)ST.1943-541X.0002445. © 2019 American Society of Civil Engineers.

Introduction (2007) and Malakoutian et al. (2012) proposed a new SFRS called a
linked column frame (LCF), which incorporates easily replaceable
The current seismic design philosophy, followed by many building links placed between closely spaced columns. Wongpakdee et al.
codes (ASCE 2016; NRCC 2015), focuses on designing the seismic (2014) and Yang et al. (2013) combined buckling restrained knee
force resisting systems (SFRSs) to meet collapse prevention perfor- braces (BRKBs) with trusses to enhance the ductility and repara-
mance objectives, where the structures can suffer from extensive bility of a truss structure after an earthquake. Their research shows
inelastic deformation and residual drifts after strong earthquake that BRKBs can be used as an efficient structural fuse to dissipate
shaking. The extensive inelastic deformation and residual drifts earthquake energy.
can lead to loss of functionality and result in prolonged downtime. To further improve the seismic performance and construction
The newer generation of SFRSs focuses on concentrating energy efficiency of fused SFRSs, a novel earthquake-resilient fused
dissipation in carefully designed and replaceable structural ele- SFRS, called the dual-fused H-frame (DFHF), is proposed in this
ments, typically known as structural fuses, where they can be re- paper. The DFHF, as shown in Fig. 1, combines damped H-frames
placed without affecting the functionality of the structure after a (DHFs) (Etebarian and Yang 2018) with welded wide flange fuses
strong earthquake. Many fused SFRSs have been developed and (WWFFs) (Yang et al. 2018) to create a structural solution that is
implemented. Balut and Gioncu (2003) proposed the dismountable efficient in construction and more seismically resilient. Each DHF
dog-bone connection in moment frames (MFs), which allows MFs module consists of two columns pin connected to a beam with two
to be repaired without replacing the entire steel beam. Similarly, BRKBs. Fig. 2(a) shows the detailed DHF module. As shown in
Nikoukalam and Dolatshahi (2015) proposed a replaceable shear Fig. 2(a), at the beam-to-column connection, a simple pin connec-
fuse for MFs, where earthquake energy can be stably dissipated tion allows the beam and column to rotate without yielding. The
using shear deformation in the middle of beams. Dusicka and Iwai BRKBs are added to provide stable energy dissipation when the
1
DHF moves laterally during earthquakes. Fig. 2(b) shows detailed
Graduate Research Assistant, Dept. of Civil Engineering, Univ. DHF connections. Each DHF module can be prefabricated at the
of British Columbia, Vancouver, BC, Canada V6T 1Z4; Structural Engi-
factory, shipped to the site, and connected vertically using simple
neer, Glotman Simpson, Vancouver, BC, Canada V6J 1N5. Email:
hamidretebarian@gmail.com bolt connections. The connections between the DHF modules have
2
Executive Director, International Joint Research Laboratory of relatively small moment demands, which makes the design, fabri-
Earthquake Engineering, Tongji Univ., Shanghai 200092, China; Professor, cation, and construction of the DHF modules very efficient. Once
Dept. of Civil Engineering, Univ. of British Columbia, Vancouver, BC, the DHF modules have been assembled vertically, the DHF bays
Canada V6T 1Z4 (corresponding author). ORCID: https://orcid.org can be connected using WWFFs. A WWFF, which was originally
/0000-0002-5751-1560. Email: yang@civil.ubc.ca proposed by Yang et al. (2018), uses a thin steel plate to dissipate
3
Research Associate, Dept. of Civil Engineering, Univ. of British earthquake energy through shear yielding. Fig. 2(c) shows a close-
Columbia, Vancouver, BC, Canada V6T 1Z4. ORCID: https://orcid.org up view of a WWFF.
/0000-0002-4450-5110
In this paper, two prototype office buildings, one three-story
Note. This manuscript was submitted on July 30, 2018; approved on
April 16, 2019; published online on September 30, 2019. Discussion and one nine-story, are designed using DFHFs for locations in San
period open until February 29, 2020; separate discussions must be sub- Francisco, California (CA), and Seattle, Washington (WA), respec-
mitted for individual papers. This paper is part of the Journal of Struc- tively. To ensure the DFHF can achieve high performance at differ-
tural Engineering, © ASCE, ISSN 0733-9445. ent levels of earthquake shaking intensities, the equivalent energy

© ASCE 04019158-1 J. Struct. Eng.

J. Struct. Eng., 2019, 145(12): 04019158


Downloaded from ascelibrary.org by SUNY at Stony Brook on 10/01/19. Copyright ASCE. For personal use only; all rights reserved.

DHF WWFF Column splice Pinned Column Base

Fig. 1. DFHF configuration.

Fig. 2. Detailing of DFHF: (a) detailed DHF module; (b) DHF connection; and (c) WWFF close-up view.

© ASCE 04019158-2 J. Struct. Eng.

J. Struct. Eng., 2019, 145(12): 04019158


design procedure (EEDP), developed by Yang et al. (2017), is uti-
lized to design DFHF prototypes. The EEDP allows engineers to
design innovative fused SFRSs to meet multiple performance ob-
jectives under different earthquake shaking intensities, where the
structural members can be selected using simple hand calculations
without iterations. The EEDP has been applied to several innova-
tive fused SFRSs (Yang et al. 2017, forthcoming; Li et al. 2018;
Etebarian and Yang 2018). To validate the performance of an
EEDP-designed DFHF, robust finite-element models of the DFHF
prototypes are created in the finite-element software OpenSees
(PEER 2000). The seismic performances of the prototypes are in-
vestigated through nonlinear time history analysis (NTHA) and in-
cremental dynamic analysis (IDA) (Vamvatsikos and Cornell 2002)
using the FEMA-P695 methodology (FEMA 2009). The results
Downloaded from ascelibrary.org by SUNY at Stony Brook on 10/01/19. Copyright ASCE. For personal use only; all rights reserved.

substantiate that the proposed DFHF prototypes designed in accor- Fig. 3. Performance objectives and force-deformation relationship of
dance with the EEDP can meet different performance objectives set DFHF.
by the engineer at different shaking intensities. In addition, the
DFHF has an acceptable margin of safety against collapse. Hence,
DFHF can be used as a robust SFRS in high seismic zones.

Energy Dissipation Mechanism of DFHF

As presented, DFHF utilizes two structural fuses (WWFFs and


BRKBs) to dissipate earthquake energy, while the remaining beams
and columns are capacity designed to remain elastic. To further
improve the earthquake resiliency of a DFHF, the WWFFs and
BRKBs are decoupled from the gravity loads, where they can be
quickly inspected, repaired, or replaced after strong earthquake
shaking. Since the inelastic deformation is concentrated only in
structural fuses, the beams and columns that are designed to remain
elastic are likely to be restored by replacing structural fuses after a
strong earthquake shaking. This makes the proposed DFHF even
more resilient to future earthquakes.
Fig. 3 shows the expected trilinear force-deformation relation-
ship of the DFHF. After a service level earthquake (SLE) shaking,
the structure is expected to remain elastic without repairs. Hence,
the structure is expected to meet the immediate occupancy (IO)
performance objective. After a design-based earthquake (DBE)
shaking, WWFFs (primary structural fuse) are designed to yield Fig. 4. EEDP energy balance concept.
in shear and dissipate earthquake energy, while BRKBs (secondary
structural fuse) are designed to remain elastic. The WWFFs are
designed to be easily repaired or replaced immediately or shortly
after a DBE shaking. Hence the structure is expected to meet the ELSDOF system (elastic energy, Ea ) to the energy dissipated by
rapid return (RR) performance objective. After a maximum credi- the ENLSDOF system (elastic strain energy, Es , and hysteretic
ble earthquake (MCE) shaking, both WWFFs and BRKBs are de- energy, Eh ) via energy modification factors.
signed to yield to dissipate earthquake energy and prevent structures The goal of the EEDP is to calculate Fy ð¼ ðSa Þy mÞ and
from collapsing. Hence, structures are expected to meet the collapse Fp ð¼ ðSa Þp mÞ, which are the yielding and plastic base shears of
prevention (CP) performance objective. the ENLSDOF system, respectively. Δy and Δp are the design
RDRs of the ENLSDOF system and correspond to the base shears
of Fy and Fp , respectively. Δu is the ultimate RDR. The following
DFHF Design Procedure steps describe how to design the proposed DFHF using the EEDP.
To practically design the proposed DFHF to meet the stated per-
formance objectives under different earthquake shaking intensities, Select Seismic Hazard Intensities
the EEDP proposed by Yang et al. (2017) is used in this paper.
The concept of the EEDP is illustrated in Fig. 4. In this figure, the The first step of the EEDP is to select the three target hazard
vertical axis represents base shear, F, which is calculated using intensities (SLE, DBE, and MCE). Different hazard intensities
pseudo acceleration, Sa , multiplied by the structural mass, m. The can be selected arbitrarily by the stakeholders and engineers based
horizontal axis represents the roof drift ratio (RDR), Δ, which is the on performance requirements.
roof displacement normalized by structural height, H. The dashed
line represents the backbone of the equivalent nonlinear single-
Determine Δy to Calculate F y and T
degree-of-freedom (ENLSDOF) system. The solid line represents
the backbone of the equivalent linear single-degree-of-freedom In this step, Δy , which is the RDR where WWFFs start to yield, can
(ELSDOF) system. The EEDP relates the energy stored in the be calculated using Eq. (1):

© ASCE 04019158-3 J. Struct. Eng.

J. Struct. Eng., 2019, 145(12): 04019158


D0 modification factor from SLE to DBE. A detailed description of
Δy ¼ δy;WWFF ð1Þ
L ΔEE1 and values of γ a can be found in Yang et al. (2017).
where D 0 = length of the WWFFs; L = distance between column
centerlines; and δ y;WWFF = WWFF yielding drift ratio. It should be Calculate Δu
noted that based on the experimental results presented in Yang et al. At the MCE hazard intensity, a DFHF shall withstand earthquake
(2018), δy;WWFF is quite consistent for WWFFs with different loading without collapse. This is achieved by designing WWFFs
geometries, which can be approximated using 0.23%. Fig. 5 shows and BRKBs to maintain their yielding strengths until the structure
the yielding mechanism of the primary system. reaches its ultimate RDR, Δu , which can be calculated using
With Δy calculated using Eq. (1), Fy can be identified from the Eq. (4):
intersection of the SLE hazard curve and Δy shown in Fig. 4. With
Δy and Fy identified, the structural period of the building, T, can be ΔEE2
Δu ¼ þ Δp ð4Þ
calculated using Eq. (2) γ b Fp H
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Downloaded from ascelibrary.org by SUNY at Stony Brook on 10/01/19. Copyright ASCE. For personal use only; all rights reserved.

Δy H=C0 where ΔEE2 ¼ mc2 0 ½ðSa ÞMCE þ ðSa ÞDBE ½ðSd ÞMCE − ðSd ÞDBE  = in-
T ¼ 2π ð2Þ cremental input energy from earthquake when shaking intensity
Fy =m
increases from DBE to MCE; and γ b = incremental energy modi-
fication factor from DBE to MCE. A detailed description of ΔEE2
where C0 = coefficient to account for difference between roof and values of γ b can be found in Yang et al. (2017).
displacement of multi-degree-of-freedom (MDOF) building and
displacement of equivalent single-degree-of-freedom (SDOF) sys-
tem (Table 7-5 of ASCE 2013). Distribute Base Shear between Primary and Secondary
SFRSs

Select Δp to Calculate F p The combination of WWFFs and DHFs creates a trilinear force-
deformation response of the proposed DFHF, as shown in Fig. 3.
At this stage, designers should select a plastic RDR, Δp , which To design the structural fuses (WWFFs and BRKBs), this re-
is the RDR where BRKBs start to yield. Since Δy is calculated sponse needs to be distributed between WWFF (primary) and DHF
based on a building’s geometry [Eq. (1)], Δp can be selected by (secondary) systems, as shown in Fig. 6. Using the equilibrium
the designer to ensure the system has the desirable ductility for relationships presented in Yang et al. (2017), the yielding strengths
the proposed DFHF. Once Δp is selected, the plastic base shear, of the primary system, FPR , and secondary system, FSE , are calcu-
Fp , can be calculated using Eq. (3): lated by Eqs. (5) and (6), respectively:

2ΔEE1 μ−λ
Fp ¼ − Fy ð3Þ FPR ¼ Fy ð5Þ
γ a HðΔp − Δy Þ μ−1

where ΔEE1 ¼ mc2 0 ½ðSa ÞDBE þ ðSa ÞSLE ½ðSd ÞDBE − ðSd ÞSLE  = in- λ−1
FSE ¼ Fy μ ð6Þ
cremental input energy from an earthquake when the shaking in- μ−1
tensity increases from SLE to DBE; and γ a = incremental energy
where λ is defined as the ratio of Fp to Fy , and μ is defined as the
ratio of Δp to Δy .

Distribute Base Shear over Building Height


Once the yielding strengths of the primary and secondary systems
are established, the calculated design base shears in Eqs. (5) and (6)

Fig. 6. Distribution of design base shear to primary and secondary


Fig. 5. Primary system yielding mechanism. systems.

© ASCE 04019158-4 J. Struct. Eng.

J. Struct. Eng., 2019, 145(12): 04019158


are distributed vertically over the height of the primary and secon- X
n

dary systems using the distribution proposed by Chao et al. (2007) W ext;SE ¼ ðFi hi ÞΔp ð13Þ
i
shown in Eq. (7):
Pn 0.75T−0.2 X
n
Vi j¼i wj hj W int;SE ¼ 2ðβ i N BRKB;n Þδ y;BRKB ð14Þ
βi ¼ ¼ ð7Þ
Vn wn hn i

Eq. (7) represents a normalized story shear distribution with where N BRKB;n = total yielding strength of BRKBs on top story
respect to the story shear in the top story, V n , where wn and wj (level n); and δ y;BRKB = yielding axial deformation of BRKBs
are the seismic weights at the roof (level n) and level j, respectively, as shown in Eq. (15):
and hn and hj are the heights from the ground to the roof (level n)
Dx Dy
and level j, respectively. Using the given normalized story shear δ y;BRKB ¼ ðΔp Þ ¼ lb sin α cos αðΔp Þ ð15Þ
distribution, the vertical distribution of the design lateral forces lb
for primary and secondary systems are obtained using Eq. (8):
Downloaded from ascelibrary.org by SUNY at Stony Brook on 10/01/19. Copyright ASCE. For personal use only; all rights reserved.

where lb = length of BRKBs; α = BRKB vertical inclination


Fi ¼ λi V y ð8Þ angle; and Dx and Dy = horizontal and vertical lengths of BRKBs,
respectively.
where It should be noted that by substituting δ y;BRKB ¼ εy lb , where εy
 0.75T −0.2 is the BRKB yielding strain, into Eq. (15), α can be determined as
w h
λi ¼ ðβ i − β iþ1 Þ Pn n n shown in Eq. (16). Since εy is defined by the BRKB properties,
j¼i wj hj when Δp is selected by the engineer, α can be determined using
Eq. (16). In other words, α and Δp cannot be independently
selected:
Design Structural Fuses
2εy
The lateral design forces obtained from the section “Distribute Base α ¼ 0.5sin−1 ð16Þ
Shear over Building Height” are then used to design the primary Δp
and secondary structural fuses. Design details are given in the
following. Once α is determined using Eq. (16), δ y;BRKB can be calculated
from Eq. (15). Having Eq. (13) equal to Eq. (14) and substituting
Primary Structural Fuse—WWFF
Using the yielding mechanism shown in Fig. 5, the external work
shown in Eq. (9) will be equal to the internal work shown in
Eq. (10):
X
n
W ext;PR ¼ ðFi hi ÞΔy ð9Þ
i

X
n
W int;PR ¼ ðβ i ÞV n δ y;WWFF D 0 ð10Þ
i

Having Eq. (9) equal to Eq. (10) and substituting δ y;WWFF from
Eq. (1) into Eq. (10), the total yielding strength for WWFFs at the
roof, V n , can be calculated as shown in Eq. (11):
Pn
Fh
V n ¼ Pn i i i ð11Þ
i i×L
β

Using the β factor calculated from Eq. (7), the total yielding
strength of the WWFFs at the other floors can be calculated using
Eq. (12):
V i ¼ βiV n ð12Þ

It should be noted that V i is the total yielding strength shared by


multiple WWFFs on each floor. To calculate the yielding strength
for each WWFF, V i should be divided by the total number of
WWFFs used on each floor.

Secondary Structural Fuse—BRKB


Fig. 7(a) shows the yielding mechanism of the secondary structural
fuse. Figs. 7(b and c) show the undeformed and deformed configu-
rations of BRKBs, respectively. Using the energy balance concept,
Fig. 7. (a) Secondary system yielding mechanism; (b) undeformed;
the external work shown in Eq. (13) will be equal to the internal
and (c) deformed geometry of BRKBs.
work shown in Eq. (14):

© ASCE 04019158-5 J. Struct. Eng.

J. Struct. Eng., 2019, 145(12): 04019158


δ y;BRKB from Eq. (15) into Eq. (14), N BRKB;n can be calculated as (Table 4). It should be noted that it is not necessary to use WWFFs
shown in Eq. (17): on every floor. WWFFs are tunable structural fuses against the de-
Pn sired seismic hazard and performance objective. Since WWFFs are
Fh designed through shear yielding within the space between columns
N BRKB;n ¼ P i i Di x Dy ð17Þ
2 ni β i × lb (Fig. 2), the total dissipated energy by WWFFs, in the event of an
earthquake, is the sum of the energy dissipated by each individual
Using the β factor calculated from Eq. (7), the total yielding WWFF. Therefore, the distribution of WWFFs within the elevation
strength of BRKBs on other floors can be calculated using Eq. (18): is less important. In this study, WWFFs are placed on all floors of
the three-story prototype building [Fig. 8(a)]. However, for the
N BRKB;i ¼ β i N BRKB;n ð18Þ nine-story prototype building, WWFFs are placed only on the
lower floors and pin-connected rigid links are utilized to connect
It should be noted that N BRKB;i is the total yielding strength the DHFs on the upper floors [Fig. 8(b)]. This is a design choice
shared by multiple bays of BRKBs on each floor. To calculate that engineers can make to tune structural fuses and does not
the yielding strength for each BRKB, N BRKB;i should be divided change the design approach/procedure presented in the manuscript.
Downloaded from ascelibrary.org by SUNY at Stony Brook on 10/01/19. Copyright ASCE. For personal use only; all rights reserved.

by the number of bays of BRKBs.

Design Nonyielding Members Numerical Model


After WWFFs and BRKBs are designed, the beams and columns To evaluate the seismic performance of the DFHF under different
are capacity designed to remain elastic under the expected gravity levels of earthquake shaking intensities, a robust nonlinear finite-
loads and maximum probable forces created by the structural fuses. element model for each prototype building is constructed using
To determine the maximum probable force from WWFFs, an over- OpenSees (PEER 2000). The columns and beams are capacity de-
strength value of 1.4, as observed from the experimental test results signed to remain elastic during an earthquake. Hence, they are
obtained by Yang et al. (2018), is used. For the maximum probable modeled using elastic beam-column elements and checked during
force from BRKBs, strain hardening, material overstrength, com- postprocessing to ensure that they do not yield. The beams are pin
pression overstrength, and resistance factor as summarized in AISC connected to the columns, and the columns are pinned at the base.
341 (AISC 2016) are employed. Note that the columns are pinned These connections allow for the rotation necessary to yield the
at the base to allow the necessary rotation to yield the structural structural fuses without the requirement to repair the connections.
fuses. This also minimizes damage and repairs at the column base To properly account for the actual geometry, rigid end offsets of the
after the earthquake shaking. beam-to-column and WWFF-to-column intersections are explicitly
modeled. The BRKBs are modeled using nonlinear truss elements
with Giuffré-Menegotto-Pinto material (Filippou et al. 1983). The
Overview of Prototype Designs BRKB material force-deformation relationship is calibrated against
the experimental data obtained by Black et al. (2004) and shown in
Two prototype office buildings, one three-story and one nine-story, Fig. 10(a). The results illustrate that the calibrated model is capable
are designed using DFHF for locations in San Francisco, CA, and of modeling the kinematic and isotropic hardening of the BRKBs.
Seattle, WA, respectively. The overall building dimensions, story The WWFFs are modeled using nonlinear two-node-link elements
heights, story masses, and gravity loads are adopted from the three- with Pinching4 material developed by Mitra and Lowes (2007).
and nine-story SAC buildings (Gupta and Krawinkler 2000). To The WWFF material force-deformation relationship is calibrated
maintain the overall plan dimensions of the SAC buildings, the against the experimental data obtained by Yang et al. (2018)
bay width of the original SAC building is decreased to accommo- and shown in Fig. 10(b). The results show that the calibrated model
date WWFF link bays. The three-story building has four bays of is capable of simulating the force-deformation response of the
DHFs, while the nine-story building has five bays of DHFs. Fig. 8 WWFFs. Masses are assigned as lump masses at the nodes based
shows the elevation and plan of the prototype buildings. Because on the tributary area. Similarly, gravity loads are uniformly distrib-
the prototype buildings are symmetric about both the north–south uted within the floors and assigned to the nodes based on their
(N–S) and east–west (E–W) directions, only the response in the tributary area. P-Δ effects are modeled using the P-Δ transforma-
N–S direction is presented in this paper. To avoid the design of tion in OpenSees (PEER 2000), and 2.5% Rayleigh stiffness and
corner columns in the nine-story building, additional columns are mass proportional damping is calculated based on the first and third
added at the corners of the nine-story building, where the bay width modes of each prototype building. The first three modal periods are
is reduced by 305 mm (1 ft) at the end bays. The prototypes are 0.55, 0.11, 0.10 s and 1.43, 0.24, 0.08 s for the three- and nine-story
assumed to be located on Type C soil according to ASCE/SEI prototype buildings, respectively.
7-16 (ASCE 2016). The MCE and DBE hazard intensities are
selected as 2% and 10% probability of exceedance in 50 years,
respectively (ASCE 2016). The SLE hazard intensity is selected Nonlinear Dynamic Analysis
as one-fifth of the MCE intensity. Figs. 9(a and b) show the three
design hazard intensities for the locations in San Francisco and The finite-element models are subjected to a range of earth-
Seattle, respectively. quake shaking intensities to assess the seismic performances of
Table 1 presents the EEDP design parameters for the two proto- the DFHFs. The suite of the 22 far-field ground motion records
type buildings. Using the EEDP presented in the previous section identified by the FEMA P695 project (FEMA 2009) is used. As
and input parameters from Table 1, the design for the DFHF is presented in Fig. 11, the records are amplitude scaled such that the
summarized in Tables 2–4. As shown in Table 3, typical rolled mean spectrum reasonably matches the target spectra of SLE, DBE,
W-sections are adopted for beams in this study. However, owing and MCE hazard intensities within the period range of 0.2 and
to the large resulting axial demands in the columns, typical rolled 1.5 T, where T is the fundamental period of the building. The
W-sections are not efficient. Thus, to efficiently resist axial de- shorter periods account for higher modes and the longer periods
mands, built-up I-shape sections are used for columns in this study for period elongation due to fuse yielding.

© ASCE 04019158-6 J. Struct. Eng.

J. Struct. Eng., 2019, 145(12): 04019158


Downloaded from ascelibrary.org by SUNY at Stony Brook on 10/01/19. Copyright ASCE. For personal use only; all rights reserved.

Fig. 8. Elevation and plan of prototype buildings.

2 2
MCE MCE
1.5 DBE 1.5 DBE
SLE SLE
Sa [g]

Sa [g]

1 1

0.5 0.5

0 0
0 1 2 3 0 1 2 3
(a) Period [s] (b) Period [s]

Fig. 9. Design spectra for (a) San Francisco, California; and (b) Seattle, Washington.

© ASCE 04019158-7 J. Struct. Eng.

J. Struct. Eng., 2019, 145(12): 04019158


Table 1. Summary of EEDP design parameters the nine-story building. These median RDRs are very close to the
Building Δy (%) Δp ð%) Δu (%) C0 γa γb target RDRs presented in Table 1. This shows that a DFHF can be
effectively designed using the EEDP methodology.
3-story 0.15 0.9 1.8 1.3 1.25 1.15
Fig. 13 presents the median of the maximum demand-capacity
9-story 0.15 0.45 1.2 1.5 1.51 2.18
ratio (DCR) for the WWFFs and BRKBs on each floor when the
prototype buildings are subjected to the scaled ground motions as
presented in Fig. 11. The results show that when the ground mo-
tions are scaled to the SLE shaking intensity, both the WWFF and
Fig. 12 presents the distribution of the maximum RDR of the BRKB demands are under the yielding capacities. This means that
DFHF for each prototype building when the DFHF is subjected the DFHFs are damage-free and, hence, met the IO performance
to all the scaled ground motions presented in Fig. 11. The results objective. When ground motions are scaled to the DBE shaking
show that the median RDRs (50% probability) are 0.2%, 1.0%, and intensity, the WWFFs yield, while the BRKBs remain elastic. This
1.6% for the three-story building and 0.17%, 0.50%, and 1.07% for means that emergency responders can enter buildings to repair or
Downloaded from ascelibrary.org by SUNY at Stony Brook on 10/01/19. Copyright ASCE. For personal use only; all rights reserved.

Table 2. Summary of structural fuses


Building Floor 1 Floor 2 Floor 3 Floor 4 Floor 5 Floor 6 Floor 7 Floor 8 Floor 9
WWFF yielding force [kN (kips)]
3-story 3,292 (740) 2,816 (633) 1,833 (412) — — — — — —
9-story 2,963 (666) 2,905 (653) 2,798 (629) — — — — — —
BRKB yielding force [kN (kips)]
3-story 1,112 (250) 952 (214) 618 (139) — — — — — —
9-story 1,695 (381) 1,659 (373) 1,601 (360) 1,517 (341) 1,401 (315) 1,254 (282) 1,072 (241) 845 (190) 547 (123)

Table 3. Summary of beams


Floors 1–3 Floors 4–6 Floors 7–9
Building Beam [SI (US)] Beam [SI (US)] Beam [SI (US)]
3-story W610 × 238 ðW24 × 146Þ — —
9-story W533 × 181 ðW21 × 111Þ W533 × 165 ðW21 × 101Þ W533 × 135 ðW21 × 83Þ

Table 4. Summary of columns


Floors 1–3 Floors 4–6 Floors 7–9
Column [mm (in.)] Column [mm (in.)] Column [mm (in.)]
Building h b tf tw h b tf tw h b tf tw
3-story 838 (33) 559 (22) 76 (3) 51 (2) — — — — — — — —
9-story 838 (33) 559 (22) 76 (3) 51 (2) 838 (33) 457 (18) 51 (2) 38 (1.5) 686 (27) 381 (15) 38 (1.5) 25 (1)
Note: h = column height; b = column width; tf = flange thickness; and tw = web thickness.

1
Experimental data
Numerical model
Normalized Force [-]

0.5

-0.5

-1
-0.1 0 0.1
(a) (b) Drift ratio [rad]

Fig. 10. Force-deformation relationship comparison for (a) BRKB; and (b) WWFF between experimental result and OpenSees simulation.

© ASCE 04019158-8 J. Struct. Eng.

J. Struct. Eng., 2019, 145(12): 04019158


1 3
Target Spectrum Target Spectrum 4 Target Spectrum
Mean Spectrum Mean Spectrum Mean Spectrum
Individual Spectrum Individual Spectrum Individual Spectrum
2 3

Sa [g]

Sa [g]

Sa [g]
2
1
1
0.2T-1.5T 0.2T-1.5T 0.2T-1.5T
0 0 0
0 0.5 1 1.5 0 0.5 1 1.5 0 0.5 1 1.5
(a) Period [sec] (b) Period [sec] (c) Period [sec]

1 4
Target Spectrum Target Spectrum Target Spectrum
Downloaded from ascelibrary.org by SUNY at Stony Brook on 10/01/19. Copyright ASCE. For personal use only; all rights reserved.

Mean Spectrum Mean Spectrum Mean Spectrum


Individual Spectrum 2 Individual Spectrum 3 Individual Spectrum

0.2T-1.5T 0.2T-1.5T
Sa [g]

Sa [g]

Sa [g]
0.2T-1.5T
2
1
1

0 0 0
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
(d) Period [sec] (e) Period [sec] (f) Period [sec]

Fig. 11. Response spectra of scaled ground motion records for (a–c) three-story; and (d–f) nine-story prototype buildings: (a) SLE; (b) DBE;
(c) MCE; (d) SLE; (e) DBE; and (f) MCE.

Collapse Assessment
Probability of exeedance[-]

0.75 To assess the seismic safety of a DFHF against collapse, the proto-
type buildings are analyzed using the procedure recommended in
0.5
SLE
FEMA P695 (FEMA 2009). Accordingly, the prototype models are
0.25 DBE subjected to an extensive incremental dynamic analyses (IDA)
MCE
Individual record
(Vamvatsikos and Cornell 2002) using the aforementioned 22 far-
0 field ground motion records. To properly model force redistribution
0.1 0.2 1 1.6 6
within the DFHF when the WWFFs and BRKBs fracture, the
(a) Roof drift ratio [%]
element removal technique as presented by Yang et al. (2015) is
implemented. In this paper, the fracture strain limits for the
Probability of exeedance[-]

1
WWFFs and BRKBs are selected as 10% (Yang et al. 2018)
0.75 and 2% (López and Sabelli 2004), respectively. At the end of each
0.5
converged step, the current WWFF and BRKB strain is compared
SLE with WWFF and BRKB capacity, respectively. If the WWFF or
DBE
0.25 MCE
BRKB has a higher strain demand than its capacity, the WWFF
Individual record or BRKB is considered to be damaged and needs to be removed.
0 After the removal of the WWFF or BRKB, the structural system is
0.17 0.5 1.07 6
(b)
updated to the new stiffness and geometry. Different solver types,
Roof drift ratio [%]
convergence criteria, and other analysis options are switched iter-
Fig. 12. Distribution of roof drift ratio: (a) three-story; and atively to solve convergence problems. Fig. 14 shows the IDA re-
(b) nine-story. sults for the prototype models. The vertical axis represents the
shaking intensity of the ground motions defined using the spectral
acceleration at the fundamental period of the prototype model. The
horizontal axis represents the corresponding peak interstory drift
replace the WWFFs without affecting the occupants. This allows ratios (ISDRs). Collapse is defined when the maximum ISDR
the DFHFs to meet the RR performance objective. It should be reaches 10%. Fragility curves are also constructed from the IDA
noted that repairing or replacing WWFFs may require the removal results and shown in Fig. 15. The horizontal axis represents the
and replacement of dry wall, celling, and floor finishes in certain spectral acceleration at the fundamental period of the prototype
small areas. With proper architectural planning of space usage dur- building, and the vertical axis represents the corresponding prob-
ing design to accommodate structural fuse replacement, occupancy ability of collapse. The collapse margin ratio (CMR), which is de-
and business interruption can be minimized. When the DFHFs are fined as the ratio of the spectral acceleration at 50% collapse
subjected to a high MCE shaking intensity, both the WWFFs and probability, SCT to the spectral acceleration of the design base level
BRKBs yield to protect the gravity system. Hence the DFHF has at the fundamental period of the structure, SMT , is calculated to be
met the CP performance objective. 1.82 and 3.93 for the three- and nine-story buildings, respectively.

© ASCE 04019158-9 J. Struct. Eng.

J. Struct. Eng., 2019, 145(12): 04019158


3 3

SLE
DBE
MCE

Floor[-]

Floor[-]
2 2
SLE
DBE
MCE

1 1
0 0.5 1 1.5 0 0.5 1 1.5
(a) Median WWFF DCR [-] (b) Median WWFF DCR [-]

3 9
SLE 8
Downloaded from ascelibrary.org by SUNY at Stony Brook on 10/01/19. Copyright ASCE. For personal use only; all rights reserved.

DBE 7
MCE
Floor[-]

Floor [-]
6
2 5
4 SLE
3 DBE
2 MCE
1 1
0 0.5 1 0 0.5 1
(c) Median BRKB DCR [-] (d) Median BRKB DCR [-]

Fig. 13. Demand-capacity ratio of (a and b) WWFFs; and (c and d) BRKBs: (a) three-story; (b) nine-story; (c) three-story; and (d) nine-story.

7 1
Individual GM Log-Normal CDF
6 Median
Individual ground motion
Collapse Probability [-]

0.75
5
Sa T1 [g]

4
0.5
S CT =3.06
3 S =3.06g
CT

2 0.25 CMR=S /S =1.82


S MT =1.68g CT MT

1
0
0 0 1 2 3 4 5 6 7
2 4 6 8 10
(a) Peak inter-story drift ratio[%] (a) Sa T1 [g]
5
Individual GM 1
Median Log-Normal CDF
4
Individual ground motion
Collapse Probability [-]

0.75
3
Sa T1 [g]

S CT =2.04g 0.5
2 S CT =2.04

1 0.25 CMR=S /S =3.93


CT MT
S MT =0.52g

0 0
2 4 6 8 10
(b) Peak inter-story drift ratio[%]
0 1 2 3 4 5
(b) Sa T1 [g]
Fig. 14. Incremental dynamic analysis curves: (a) three-story; and
(b) nine-story. Fig. 15. Collapse fragility curve: (a) three-story; and (b) nine-story.

Once CMRs are determined, adjusted CMRs (ACMRs) should in Table 7-1 of FEMA P695 (FEMA 2009). The SSFs of the
be computed to account for the difference in spectral shape of three- and nine-story prototype buildings are selected as 1.28
the applied ground motion records. The ACMR is obtained by and 1.18, respectively. Therefore, the ACMRs are found to be
multiplying the CMR by the spectral shape factor (SSF) provided 2.32 and 4.63 for the three- and nine-story prototype buildings,

© ASCE 04019158-10 J. Struct. Eng.

J. Struct. Eng., 2019, 145(12): 04019158


respectively. For adequate structural safety against collapse, the Any opinions, findings, and conclusions or recommendations
calculated ACMR should be greater than the acceptable ACMR, expressed in this paper are those of the authors.
i.e., ACMR10% . ACMR10% is determined by evaluating the total
system collapse uncertainty due to various sources of uncertainty,
including record-to-record variability in ground motions, quality References
ratings of design requirements, test data, and numerical models.
In this paper, ACMR10% is 2.1, which is less than the calculated AISC. 2016. Specification for structural steel buildings. ANSI/AISC 341.
ACMRs for both the three- and nine-story prototype buildings. Chicago: AISC.
This shows that the proposed DFHF designed using the EEDP ASCE. 2013. Seismic evaluation and retrofit of existing buildings. ASCE/
SEI 41. Reston, VA: ASCE.
has adequate seismic resistance against collapse.
ASCE. 2016. Minimum design loads for buildings and other structures.
ASCE/SEI 7. Reston, VA: ASCE.
Balut, N., and V. Gioncu. 2003. “Suggestion for an improved ‘dog-bone’
Summary and Conclusion solution.” In Proc., 4th Int. Conf. STESSA, 129–134. Rotterdam,
Downloaded from ascelibrary.org by SUNY at Stony Brook on 10/01/19. Copyright ASCE. For personal use only; all rights reserved.

Netherlands: A.A. Balkema.


Past earthquakes have shown that structures designed using current Black, C. J., N. Makris, and I. D. Aiken. 2004. “Component testing,
seismic design philosophy, which focuses on collapse prevention seismic evaluation and characterization of buckling-restrained braces.”
performance objectives, can suffer from extensive inelastic defor- J. Struct. Eng. 130 (6): 880–894. https://doi.org/10.1061/(ASCE)0733
mation and residual drifts. Extensive inelastic deformation and -9445(2004)130:6(880).
residual drifts can lead to a loss of functionality and result in pro- Chao, S. H., S. C. Goel, and S. S. Lee. 2007. “A seismic design lateral force
distribution based on inelastic state of structures.” Earthquake Spectra
longed downtime. The newer generation of structures focuses on
23 (3): 547–569. https://doi.org/10.1193/1.2753549.
concentrating energy dissipation in carefully designed and replace- Dusicka, P., and R. Iwai. 2007. “Development of linked column frame
able structural elements. In this paper, a novel resilient seismic system for seismic lateral loads.” Struct. Eng. Res. Front. 1–13. https://
force resisting system called dual-fused H-frame (DFHF) is pro- doi.org/10.1061/40944(249)63.
posed. DFHF combines damped H-frame (DHF) modules with Etebarian, H., and T. Y. Yang. 2018. “Development and assessment of in-
welded wide flange fuses (WWFFs) to create a structural solution novative modular damped H-Frame system.” In Proc., 11th National
that is efficient in construction and more seismic resilient. Each Conf. in Earthquake Engineering, 11. Los Angeles, CA: Earthquake
DHF module is prefabricated in the factory, shipped to the site, Engineering Research Institute.
and connected vertically using simple bolt connections. The con- FEMA. 2009. Quantification of building seismic performance factors.
nections between the DHF modules have relatively small moment FEMA P-695. Washington, DC: FEMA.
Filippou, F. C., E. P. Popov, and V. V. Bertero. 1983. Effects of bond
demand, which makes the design, fabrication, and construction of
deterioration on hysteretic behavior of reinforced concrete joints. Tech-
the DHF modules very efficient. Once the DHF modules have been nical Rep. No. EERC-83/19. Berkeley, CA: Earthquake Engineering
assembled vertically, the bays of the DHF can be connected using Research Center, Univ. of California.
WWFFs. WWFFs serve as the primary structural fuses that are de- Gupta, A., and H. Krawinkler. 2000. “Behavior of ductile SMRFS at
signed to yield and dissipate earthquake energy when the seismic various seismic hazard levels.” J. Struct. Eng. 126 (1): 98–107. https://
hazard exceeds the service-level earthquake shaking intensity. doi.org/10.1061/(ASCE)0733-9445(2000)126:1(98).
Buckling restrained knee braces (BRKBs) are the secondary struc- Li, T., T. Y. Yang, G. Tong, D. P. Tung, and Y. Li. 2018. “Performance-
tural fuses that are designed to yield when the hazard exceeds the based seismic design and evaluation of fused steel diagrid frame.”
design-based earthquake shaking intensity. Both WWFFs and Earthquake Spectra 34 (4): 1869–1891. https://doi.org/10.1193
BRKBs are designed to have sufficient ductility to prevent DFHFs /121017EQS257M.
López, W. A., and R. Sabelli. 2004. Seismic design of buckling-restrained
from collapsing when the hazard exceeds the maximum credible
braced frames. Moraga, CA: Structural Steel Educational Council.
earthquake shaking intensity. In this paper, two prototype office Malakoutian, M., J. W. Berman, and P. Dusicka. 2012. “Seismic response
buildings, a three-story and a nine-story building, are designed using evaluation of the linked column frame system.” Earthquake Eng. Struct.
DFHFs for locations in San Francisco, CA, and Seattle, WA, respec- Dyn. 42 (6): 795–814. https://doi.org/10.1002/eqe.2245.
tively. To ensure the DFHF can achieve high performance at differ- Mitra, N., and L. N. Lowes. 2007. “Evaluation, calibration, and verification
ent levels of earthquake shaking intensities, the equivalent energy of a reinforced concrete beam–column joint model.” J. Struct. Eng.
design procedure (EEDP) is utilized to design the DFHF prototypes. 133 (1): 105–120. https://doi.org/10.1061/(ASCE)0733-9445(2007)
To validate the performance of the EEDP-designed DFHF, robust 133:1(105).
finite-element models of the DFHF prototypes are created in Open- Nikoukalam, M. T., and K. M. Dolatshahi. 2015. “Development of struc-
tural shear fuse in moment resisting frames.” J. Constr. Steel Res.
Sees. The finite-element models are subjected to a suite of far-field
114 (Nov): 349–361. https://doi.org/10.1016/j.jcsr.2015.08.008.
ground motions scaled to the target hazard intensities. The results
NRCC (National Research Council of Canada). 2015. National building
show that the proposed DFHF prototypes designed in accordance code of Canada 2015. Ottawa: NRCC.
with the EEDP can meet different performance objectives set by the PEER (Pacific Earthquake Engineering Research). 2000. Open system
engineer at different shaking intensities. In addition, it has an accept- for earthquake engineering simulation (OpenSees). Berkeley, CA:
able collapse safety margin. Hence, the DFHF can be used as an Pacific Earthquake Engineering Research Center, Univ. of
efficient seismic force resisting system in high seismic zones. California.
Vamvatsikos, D., and C. A. Cornell. 2002. “Incremental dynamic analysis.”
Earthquake Eng. Struct. Dyn. 31 (3): 491–514. https://doi.org/10.1002
/eqe.141.
Acknowledgments Wongpakdee, N., and S. Leelataviwat, S. C. Goel, W. C. Liao. 2014.
“Performance-based design and collapse evaluation of buckling re-
The authors would like to acknowledge the funding provided strained knee braced truss moment frames.” Eng. Struct. 60 (Feb):
by the International Joint Research Laboratory of Earthquake 23–31. https://doi.org/10.1016/j.engstruct.2013.12.014.
Engineering (ILEE) and National Science Foundation China Yang, T. Y., J. Atkinson, L. Tobber, and D. P. Tung. Forthcoming. “Seismic
(Grant No. 51778486), State Key Laboratory of Disaster Reduc- design of outrigger systems using equivalent energy design procedure.”
tion in Civil Engineering for theur support the graduate student. Earthquake Spectra.

© ASCE 04019158-11 J. Struct. Eng.

J. Struct. Eng., 2019, 145(12): 04019158


Yang, T. Y., W. Banjuradja, and L. Tobber. 2018. “Experimental test of Yang, T. Y., Y. Li, and S. Leelataviwat. 2013. “Performance-based design
welded wide flange fuses.” Key Eng. Mater. 763: 414–422. https://doi and optimization of buckling restrained knee braced truss moment
.org/10.4028/www.scientific.net/KEM.763.414. frame.” J. Perform. Constr. Facil. 28 (6): A4014007. https://doi.org/10
Yang, T. Y., Y. Li, and S. C. Goel. 2015. “Seismic performance evalu- .1061/(ASCE)CF.1943-5509.0000558.
ation of long-span conventional moment frames and buckling-restrained Yang, T. Y., D. P. Tung, and Y. Li. 2017. “Equivalent energy design pro-
knee-braced truss moment frames.” J. Struct. Eng. 142 (1): 04015081. cedure for earthquake resilient fused structures.” Earthquake Spectra.
https://doi.org/10.1061/(ASCE)ST.1943-541X.0001333. 34 (2): 795–815. https://doi.org/10.1193/122716EQS254M.
Downloaded from ascelibrary.org by SUNY at Stony Brook on 10/01/19. Copyright ASCE. For personal use only; all rights reserved.

© ASCE 04019158-12 J. Struct. Eng.

J. Struct. Eng., 2019, 145(12): 04019158

You might also like