Engineering Structures 110 (2016) 116–126

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Development and application of a simplified model for the design

of a super-tall mega-braced frame-core tube building
Xinzheng Lu a,⇑, Linlin Xie a, Cheng Yu b, Xiao Lu c
a
Department of Civil Engineering, Key Laboratory of Civil Engineering Safety and Durability of China Education Ministry, Tsinghua University, Beijing 100084, China
b
Department of Engineering Technology, University of North Texas, Denton, TX 76210, USA
c
School of Civil Engineering, Beijing Jiaotong University, Beijing 100044, China

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

Article history: Resilience-based earthquake design for next-generation super-tall buildings has become an important
Received 16 December 2014 trend in earthquake engineering. Due to the complex structural system in super-tall buildings and the
Revised 14 November 2015 extreme computational workload produced when using refined finite element (FE) models to design such
Accepted 19 November 2015
buildings, it is rather difficult to efficiently perform a comparison of different design schemes of super-
tall buildings and to investigate the advantages and disadvantages of different designs. Here, a simplified
nonlinear model is developed and applied to compare two design schemes (i.e., the fully braced scheme
Keywords:
and half-braced scheme) of a super-tall mega-braced frame-core tube building, which is an actual engi-
Super-tall building
Simplified model
neering project with a total height of approximately 540 m. The accuracy of the simplified model is val-
Resilience idated through a comparison of the results of modal analyses, static analyses and dynamic time history
Plastic energy dissipation analyses using the refined FE models. Subsequently, the plastic energy dissipation of different compo-
Plastic energy distribution nents and the distribution of the total plastic energy dissipation over the height of the two design
schemes are compared using the proposed simplified model. The analyses indicate that the fully braced
scheme is superior because of its more uniform energy distribution along the building height and the
large amount of energy dissipated in the replaceable coupling beams, which enables rapid repair and
re-occupancy after an earthquake. In contrast, the potential damage in the half-braced scheme is more
concentrated and more severe, and the damage in the core tubes is difficult to repair after an earthquake.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction Tower (H = 420.5 m) [14], were conducted using refined FE models,

which were established using various general-purpose FE software
In recent years, studies on the resilience-based seismic design packages (e.g., ANSYS [15], Perform 3D [16], LS-DYNA [17] and
of super-tall buildings have become increasingly popular [1–5]. ABAQUS [18–23]) and open-source software packages (e.g., Open-
To design an earthquake-resilient super-tall building, the perfor- Sees [24]). The seismic performances of these super-tall buildings
mance of super-tall buildings subjected to various earthquake subjected to various seismic intensities were predicted to optimize
intensities should be accurately simulated; such simulations are the seismic designs. More recently, collapse simulations of super-
used to predict the seismic energy dissipated in replaceable and tall buildings subjected to extreme earthquakes were successfully
repairable components as well as the structural damage in key performed by Lu et al. [25,26] using MSC.Marc [27]. The potential
components that are difficult to repair. collapse modes of these super-tall buildings were predicted, and
Numerous studies have been conducted using three- the critical zones that might induce collapse were identified, which
dimensional (3D) refined FE models to investigate the nonlinear could serve as a reference for future improved designs.
seismic performance and predict the potential collapse modes of As described above, the refined FE model has been widely
super-tall buildings. Nonlinear time-history analyses of various applied to investigate the seismic performance and reveal the
super-tall buildings, including the Taipei Financial Center potential collapse modes of tall and super-tall buildings with var-
(H = 508 m) [6,7], the Shanghai Tower (H = 632 m) [8–12], the ious structural systems [6–14,25,26,28–31]. However, such simu-
Republic Plaza (H = 280 m) [13], and the Shanghai World Financial lations have several drawbacks: first, the refined FE model
cannot be accurately established without specific structural design
details, which are typically unavailable at the preliminary design
⇑ Corresponding author.
stage, thus restricting the applications of this type of model at this
E-mail address: [email protected] (X. Lu).

http://dx.doi.org/10.1016/j.engstruct.2015.11.039
0141-0296/Ó 2015 Elsevier Ltd. All rights reserved.
 X. Lu et al. / Engineering Structures 110 (2016) 116–126 117

stage. Moreover, super-tall buildings are typically composed of are compared for both schemes, thereby conclusively providing a
many different components, thereby leading to an extremely large reference for the selection of a better option among the various
computational workload and low efficiency when using the refined considered schemes. The analytical results indicate that the fully
FE models. Such models restrict the implementation of parametric braced scheme induces a more uniform plastic energy dissipation
analyses or incremental dynamic analyses (IDAs). In particular, distribution than the half-braced scheme. Furthermore, the fully
several different design schemes are typically proposed at the pre- braced scheme effectively enables the energy dissipation to be
liminary design stage. Due to the lack of specific design details and located in the readily replaceable components (e.g., coupling
the large computational workload, the comparison between vari- beams and perimetric trusses) instead of the key components that
ous design schemes, which is essential for the design of super- are difficult to repair (e.g., mega columns, core tubes and mega
tall buildings, cannot be easily performed using the refined FE braces). As a result, the fully braced scheme provides a better
models. seismic resilient performance than the half-braced scheme. The
In contrast, a simplified model that represents the key nonlinear outcome of this study serves as a guideline for a method to reliably
and dynamic characteristics of super-tall buildings and effectively and efficiently understand the seismic performance of different
reduces the computational effort has the potential to facilitate the preliminary design schemes of super-tall buildings, which can pro-
comparison of different preliminary design schemes. Moreover, if vide guidance and serve as a reference for the performance-based
the engineering demand parameters are available through the and resilience-based earthquake design of super-tall buildings.
analysis of the simplified model during the preliminary design
stage, such a model can also be used to guide and optimize the pre-
liminary design. 2. Introduction of two design schemes and the associated
Although limited research has been reported on establishing a refined FE models
simplified model of super-tall buildings, many researchers have
conducted studies on establishing simplified models for conven- The project studied in this research is a multi-functional super-
tional tall buildings. For example, a simplified model for the tall office building with a total height of approximately 540 m. The
framed-tube structure proposed by Connor and Pouangare [32] building adopts a hybrid lateral-force resisting system named as
was applied to analyze its elastic response subjected to static lat- ‘‘mega-braced frame-core tube” [26]. Two design schemes are pro-
eral loads and subsequently used to guide the preliminary design. posed at the preliminary design stage, which are referred to as the
Luco and Cornell [33] developed a simplified model involving the ‘‘fully braced scheme” and ‘‘half-braced scheme”. The fully braced
interconnection of two shear beams to predict the seismic perfor- scheme involves the use of mega columns, mega braces within
mance of tall buildings. Meftah et al. [34] presented a simplified the full height of the structure (i.e., Zones 1–8), perimetric trusses,
approach for the seismic analysis of a tall building braced by shear concrete core tubes and secondary frames, as shown in Fig. 1. In
walls and thin-walled open section structures, and a simplified for- contrast, the half-braced scheme involves the use of mega col-
mulation for the vibrational frequencies and internal forces sub- umns, mega braces in the lower four zones of the structure (i.e.,
jected to earthquakes was obtained based on D’Alembert’s Zones 1–4), outer frame tubes in the higher four zones (i.e., Zones
principle. An important achievement in the simplified modeling 5–8), perimetric trusses, outriggers, concrete core tubes and sec-
of super-tall buildings was accomplished by Lu et al. [35]; specifi- ondary frames, as shown in Fig. 2. Further details of the half-
cally, a two-dimensional (2D) simplified model encompassing non- braced scheme are presented in Lu et al. [26]. The differences
linear beam-column elements and nonlinear spring elements for between these two schemes are listed in Table 1.
the Shanghai Tower (H = 632 m) was proposed. The reliability of This super-tall building is located in Beijing, a relatively high
this model was validated by comparing the results of the simplified seismic region in China [36] (with a maximum spectrum accelera-
model with those of the refined FE model. The analyses of the plas- tion of 0.9g for the Maximum Considered Earthquake (MCE) level,
tic energy dissipation indicated that the outrigger was the primary where g is the acceleration of gravity); both the wind and seismic
plastic energy dissipation component, and the total plastic energy loads play important roles in the structural design. An elastic anal-
distribution along the height of the building subjected to three ysis of the building indicates that the maximum drift ratios when
seismic intensities was identified. Despite these efforts, the simpli-
fied model has only been used for the Shanghai Tower (which is a
mega-column/core-tube/outrigger system), in a study by Lu et al. Zone 8
[35]. Additional validation of the reliability of this model is Hat truss
required for other types of super-tall buildings. In addition, further
Zone 7
studies should also be performed on the application of the simpli-
fied model at the preliminary design stage and the comparison of
Conversion
different design schemes. Zone 6
truss
Therefore, based on the simplified model and associated
parameter determination approaches proposed by Lu et al. [35], a
Zone 5
simplified model is developed for the seismic analysis of an actual
Perimetric truss
super-tall mega-braced frame-core tube building. In addition, this
simplified model is used to perform the comparison of two Zone 4
preliminary design schemes for this building in terms of its resili- Mega brace
ent performance. The studies indicate that this simplified model is Zone 3
also capable of efficiently and reliably predicting the key seismic
Mega column
characteristics of this building, thereby laying a foundation for Zone 2
the further comparison of different design schemes. Subsequently, RC core tube
the energy dissipation characteristics of these two structural
Zone 1
schemes are investigated and discussed through nonlinear Arch truss
time-history analyses using the simplified models. The plastic
energy dissipation contribution of each component as well as the
total plastic energy distribution along the height of the building Fig. 1. Fully braced scheme.
118 X. Lu et al. / Engineering Structures 110 (2016) 116–126

this super-tall building, the controls of the degree and location of

Hat truss damage are critical issues; ideally, the damage should be uniformly
Zone 8 distributed in the replaceable components along the height of the
building, and severe damage inside the key components and dam-
Outer age concentration along the height of the building should be
Zone 7
frame avoided.
tube The 3D refined FE model of the half-braced scheme was estab-
Zone 6
lished by Lu et al. [26] using MSC.Marc. In this work, the 3D refined
FE model of the fully braced scheme is also established using the
Zone 5
Perimetric identical methodology proposed by Lu et al. [26]. The refined FE
truss models of two structural schemes are illustrated in Fig. 3, including
Zone 4
78,099 elements for the fully braced scheme and 64,875 elements
Mega brace for the half-braced scheme.
Zone 3

Mega column 3. Establishment of the simplified model

Zone 2
3.1. General provisions of the simplified model
Zone 1 RC core tube

Simplified models of both schemes are established based on the

simplification approaches proposed by Lu et al. [35]. Because the
Fig. 2. Half-braced scheme. major objective of establishing these simplified models was to
compare the dynamic properties and characterize the plastic
energy dissipation of the two schemes, the following three princi-
subjected to the designed seismic load (i.e., 63% probabilities of ples are followed when developing the simplified models.
exceedance in 50 years) and the designed wind load are approxi-
mately 1/940 and 1/570, respectively, both of which meet the max- (1) The models of the 3D building are simplified as planar mod-
imum allowable story drift ratio of 1/500 at the design level. els. The fundamental periods in the x and y directions of the
Nevertheless, if this building is incapable of being functionally 3D refined FE model of the fully braced scheme are approx-
recovered immediately after an earthquake due to severe damage imately 7.38 and 7.33 s, respectively. Similarly, the funda-
that could occur in key components, great economic losses will mental periods in the x and y directions of the half-braced
occur, and there will be great negative social impacts. Hence, for scheme are 7.44 and 7.69 s, respectively. Therefore, the

Table 1
Primary differences between the fully braced scheme and half-braced scheme.

Fully braced scheme Half-braced scheme

Total height 536.7 m 545.6 m
Planar dimensions Top: 54 m  54 m Top: 60 m  60 m
Waist: 69 m  69 m Waist: 50 m  50 m
Bottom: 71 m  71 m Bottom: 67 m  67 m
Height of the minimal cross section of the structure H  380 m H  420 m
Setup of the mega braces Zones 1–8 Zones 1–4
Setup of the perimetric trusses 8 sets 9 sets
Setup of the outriggers – Zones 5–8
Maximum dimensions of the mega column 5.2 m  5.2 m 6.5 m  6.5 m

Zone 8 Zone 8
Zone 7
Zone 7 Zone 8 Zone 7 Zone 7 Zone 8
Zone 6
Zone 6
Zone 5
Zone 5
Zone 5 Zone 6 Zone 5 Zone 6
Zone 4 Zone 4

Zone 3 Zone 3
Zone 2 Zone 3 Zone 4 Zone 3 Zone 4
Zone 2

Zone 1
Zone 1
z z
Zone 1 Zone 2 Zone 2
x y x y Zone 1

(a) (b)
Fig. 3. Refined FE models of two structural schemes: (a) The fully braced scheme. (b) The half-braced scheme.
 X. Lu et al. / Engineering Structures 110 (2016) 116–126 119

Table 2
Reduction in period of the two models (%) when the component stiffness is reduced by 50%.

Change in period (%)

Mega column Core tube Mega brace Trusses Secondary frame
Fully Half Fully Half Fully Half Fully Half Fully Half
T1 15.52 20.07 8.57 6.68 4.71 4.95 2.73 2.96 1.19 0.81
T2 15.40 19.16 8.86 7.96 4.60 4.03 2.71 3.07 1.26 0.83
T3 3.79 1.21 15.47 18.86 6.93 7.62 3.52 1.95 0.44 1.42
T4 4.65 7.98 17.80 12.43 5.93 7.55 2.15 3.04 0.87 0.82
T5 4.64 8.08 17.79 13.31 5.88 7.25 2.12 3.18 0.91 1.15
T6 3.42 1.06 17.24 17.24 7.42 6.87 3.92 1.41 0.75 1.08
T7 2.48 3.70 21.25 19.07 4.24 4.29 1.89 2.79 0.77 0.72
T8 2.48 3.86 20.10 19.96 4.08 4.10 1.83 3.10 0.80 1.00
T9 3.28 1.98 17.29 21.65 7.25 5.45 3.89 2.35 1.24 2.07

lateral stiffness values in the two orthogonal directions are

highly similar for both schemes. The first torsion period is Zone 8
Zone 8
2.77 s for the fully braced scheme and 3.42 s for the half- Rigid link
braced scheme. Both torsion-translational period ratios Zone 7 Zone 7 Coupling beam
(0.38 and 0.44) are considerably smaller than the upper lim-
itation of 0.85 specified in the Chinese design code (i.e., the
Technical Specification for Concrete Structures for Tall Build- Zone 6 Zone 6 Rigid beam
ings, JGJ 3-2010 [37]). Thus, the torsion effects in both
schemes are not significant. As a result, it is feasible to sim- Zone 5 Zone 5 Perimetric
plify the models using planar models instead of 3D models. truss
Based on the above discussion, the models for the fully
Zone 4 Zone 4 Mega brace
braced scheme and half-braced scheme are simplified into
planar models in the fundamental transitional vibration
plane. Zone 3 Zone 3
(2) The simplified models only consider the mega columns, core
tubes (including the shear walls and coupling beams), mega Zone 2 Zone 2 Mega column
braces, and trusses (including the outriggers and perimetric
trusses). Other components, such as secondary frames, are
Zone 1 Zone 1 Core tube
not included in these simplified models. To validate the
rationality of such simplifications, the stiffness of each type
of component is reduced by 50% in the refined FE models,
whereas the total mass and the properties of other compo-
(a) (b)
nents remain unchanged, to investigate the stiffness contri- Fig. 4. Refined FE model and simplified model of the fully braced scheme: (a) The
bution of various components [35]. The first 9 vibrational refined FE model. (b) The simplified model.
periods of the models with reduced stiffness and those of
the original models are compared in Table 2. A 50% reduc-
tion in the stiffness of the mega columns results in a clear (3) The nonlinear beam-column elements and nonlinear spring
change in the fundamental translational periods (approxi- elements proposed by Lu et al. [35] are adopted to simulate
mately 15% for the fully braced scheme and 20% for the the above primary components in the simplified model. The
half-braced scheme) but a relatively small change in tor- hysteretic model and corresponding parameter calibration
sional periods, indicating that the mega columns make sub- method proposed by Lu et al. [35] are also adopted. The
stantial contributions to the lateral stiffness of the building. details of these models will be discussed in the following
When the stiffness of the core tubes is reduced, significant section.
changes in the higher-order vibration periods are observed,
with a maximum change of 21.15% for the fully braced Given the above-mentioned assumptions, the final simplified
scheme and 21.65% for the half-braced scheme. Moreover, models of both schemes are established. Comparisons between
the reduction in the core tube stiffness also has a great the refined FE models and corresponding simplified models are
impact on the torsional periods: 15.47% and 18.86% presented in Figs. 4 and 5. Note that the number of elements in
increases for the fully braced scheme and half-braced the simplified model decreases to approximately 1.68% for the fully
scheme, respectively. These results indicate that the core braced scheme (i.e., 1309 elements) and 2.80% for the half-braced
tube stiffness significantly contributes to both the lateral scheme (i.e., 1817 elements); thus, the computational workload
and torsional stiffness of the building. When the stiffness will be significantly reduced.
of the mega braces and trusses is reduced, the change in per-
iod is a moderate change of less than or equal to 7.42%. A 3.2. Simplified models for the primary components
50% reduction in the stiffness of the secondary frame results
in a small change of less than or equal to 2.07%. Thus, the A nonlinear beam-column element is adopted to simulate the
stiffness contribution of the secondary frame is negligible mega columns, mega braces, outriggers and perimetric trusses. A
and can be ignored in the simplified model. Given the above typical layout of the core tubes with coupling beams and shear
comparisons, the mega columns, core tubes, mega braces walls is shown in Fig. 6. The distributions of the coupling beams
and trusses are regarded as primary components that must of both schemes are regular and mainly distributed symmetrically
be properly considered in the simplified model. along the two orthogonal axes of the core tubes. Hence, the core
120 X. Lu et al. / Engineering Structures 110 (2016) 116–126

Rigid link (a)

Zone 8 Zone 8
Coupling beam
Zone 7 Zone 7
Rigid beam
Zone 6 Zone 6
Outrigger/peri
Zone 5 Zone 5 -metric truss

Zone 4 Zone 4 Mega brace

Zone 3 Zone 3
Core tube
Zone 2 Zone 2 Mega (b)
column
Zone 1 Zone 1

(a) (b)
Fig. 5. Refined FE model and simplified model of the half-braced scheme: (a) The
refined FE model. (b) The simplified model.

tubes are divided into two identical sub-tubes by the coupling

beams. As proposed by Lu et al. [35], the coupling beams (i.e.,
the coupling beams A and B in Fig. 6(a) and the coupling beams
A to F in Fig. 6(b)) are simulated as one shear spring element.
The two identical sub-tubes are simulated as two nonlinear Fig. 7. Hysteretic component model used in this research: (a) With pinching effect.
beam-column elements. The rigid arms are used to connect the (b) Without pinching effect.
coupling beams and sub-tubes to consider the actual size of the
core tube.
The hysteretic model proposed by Lu et al. [35,38] (shown in proposed by Lu et al. [35]. For the steel components, such as the
Fig. 7) is adopted to simulate different components in the simpli- mega braces and outriggers, all of the parameters in Fig. 7 are cal-
fied model. The parameters in this model can be classified into ibrated through the analysis of the corresponding refined FE mod-
two groups: (1) parameters for the backbone curve and (2) param- els. In regard to other components, the backbone curve parameters
eters for the hysteretic rule. The first group of parameters includes are also obtained through the analysis of the corresponding refined
K0 (the initial stiffness), Fy (the generalized yield strength, e.g. axial FE models. The hysteretic parameters are calibrated through the
force, shear force or bending moment), g (the hardening ratio), gsoft experimental data from similar specimens tested under cyclic
(the softening ratio), a (the ratio of peak strength and yield loads, as proposed by Lu et al. [35], by considering the difficulties
strength) and b (the ratio of reversed yield strength and yield in simulating the hysteretic behavior of concrete components.
strength) as shown in Fig. 7. The second group of parameters The typical values of the hysteretic parameters for different com-
includes c (representing the pinching effect), x (representing the ponents are summarized in Table 3.
position of the ending point of slip), C (dimensionless accumulated
hysteretic energy dissipation parameter, which reflects the capac- 4. Validation of the simplified model
ity of resisting strength degradation caused by the cyclic loading)
and ak (defining unloading stiffness) as shown in Fig. 7. The Based on the analytical results of the refined FE MSC.Marc mod-
corresponding parameters are also determined using the methods els for both schemes, modal analyses, static analyses and nonlinear

W1 W1
W2
A F
W2
W3 E
B W1
W2 W1
W3 D W2
W1 B W2
W2 C W3

A W2 W1
B W3
A
W1 W2
W1
y x
x y
(a) (b)
Fig. 6. Distribution of the coupling beams in typical RC core tubes: (a) The fully braced scheme. (b) The half-braced scheme.
 X. Lu et al. / Engineering Structures 110 (2016) 116–126 121

Table 3
Typical values of the hysteretic parameters of different components.

g a b gsoft C c ak x
Mega-column (flexural) 0.130 1.76 0.9 0.033 10,000 1.0 0.020 0.40
Coupling beam (shear) 0.014 1.085 1.0 0.070 12.0 0.9 0.400 0.50
Core tube (flexural) 0.010 1.40 1.0 0.001 100 0.7 0.550 0.10
Mega-brace (tension–compression) 0.010 1.00 1.0 0.001 500 1.0 0.000 0.10
Outrigger and perimetric truss (shear) 0.150 1.00 1.0 0.200 150 1.0 0.040 0.10
Outrigger and perimetric truss (flexural) 0.050 1.20 1.0 0.001 15.0 0.6 0.001 0.10

Table 4 validating the consistency of the mass distribution between these

Comparison of the first six translational periods (in seconds) from the simplified and two types of models.
refined FE MSC.Marc models.
To further validate the reliability and rationality of the simpli-
Fully braced scheme Half-braced scheme fied models, time history analyses are performed for the simplified
Refined Simplified Relative Refined Simplified Relative and refined FE models. The widely used El Centro EW 1940 record
FE model model error (%) FE model model error (%) is selected as the typical ground motion input. Because ETABS is
1st 7.380 7.424 0.59 7.4405 7.5873 1.97 incapable of conducting nonlinear analysis for super-tall buildings,
2nd 2.391 2.257 5.59 2.5707 2.5323 1.49 linear time history analyses are first conducted here. The peak
3rd 1.210 1.191 1.60 1.3387 1.3286 0.76 ground acceleration (PGA) is scaled to 70 gal, which is the intensity
4th 0.762 0.776 1.81 0.8306 0.7868 5.27
of service level earthquake (SLE) for the site of this building. The
5th 0.537 0.553 3.05 0.6053 0.5917 2.25
6th 0.411 0.421 2.49 0.4519 0.4355 3.62 comparisons of story drift ratio envelopes predicted by the simpli-
fied model and refined FE models for the fully braced scheme are
compared in Fig. 9(a). Good agreement is observed, which vali-
dates the reliability of the proposed simplified model at the elastic
time history analyses are performed to validate the reliability of stage. To validate the reliability of the proposed simplified model
the simplified models. In addition, analyses using ETABS [39], a at the nonlinear stage, nonlinear time history analyses using sim-
commercial software widely used for the design of structures, are plified model and refined FE MSC.Marc model are conducted. The
also conducted here to further validate the reliability of the pro- PGA is scaled to 400 gal, which is the intensity of the MCE for
posed simplified models. the site of this building. The comparisons of the time history anal-
The first six translational periods of the refined FE MSC.Marc ysis results are illustrated in Fig. 9(b–d). Fig. 9(b) shows the roof
models compared to those of the simplified models exhibit close displacement histories of the fully braced scheme, which are nearly
agreement for both schemes as shown in Table 4, with a maximum identical. The story displacement envelopes of the fully braced
relative error of approximately 5.57%. In addition, the first three scheme shown in Fig. 9(c) are also in good agreement, with some
translational periods of the refined FE ETABS models, which are negligible difference in the central zones. The story drift ratio
usually used for the design of structures, also agree well with those envelopes of the fully braced scheme are also found to be in good
of the simplified models. Specifically, the relative errors for the agreement with the refined FE MSC.Marc model, as shown in Fig. 9
fully braced model are 2.15%, 4.72% and 2.85%, respectively, (d). As a result, the time history analysis results predicted by the
and the relative errors for the half-braced model are 0.85%, 4.12% 2D simplified model are shown to be in good agreement with those
and 4.53%, respectively. The above comparisons indicate that the of refined FE MSC.Marc models for the fully braced scheme. Similar
simplified model is capable of capturing the basic dynamic findings are also obtained for the half-braced scheme. The story
characteristics. drift ratio envelopes of the half-braced scheme shown in Fig. 9(e)
Subjected to gravity load only, the total axial forces in the mega are found to be in good agreement. In particular, the simplified
columns and core tubes at the bottom of each zone are calculated model of the half-braced scheme is capable of representing the
and presented in Fig. 8. The axial forces in the mega columns and inter-story drift sudden change due to the outriggers. The slight
core tubes in the simplified models are shown to be in good agree- discrepancy between the predicted results of the simplified model
ment with those in the refined FE models for both schemes, thus and refined model found at the top zone of the building is considered
Axial force in refined model ( h106N)

4000 5000
Axial force in refined model ( h106N)

Mega column Zone 1 Mega column Zone 1

Core tube 4000 Core tube
3000

3000
2000
2000

1000
1000

Zone 8 Zone 8
0 0
0 1000 2000 3000 4000 0 1000 2000 3000 4000 5000
Axial force in simplified model (h106N) Axial force in simplified model (h106N)
(a) (b)
Fig. 8. Comparison of the axial forces in mega columns and core tubes of the simplified and refined models: (a) The fully braced scheme. (b) The half-braced scheme.
122 X. Lu et al. / Engineering Structures 110 (2016) 116–126

600

500

Building height (m)

400

300

200
MSC.Marc model
100 ETABS model
Simplified model
0
0 0.0003 0.0006 0.0009 0.0012 0.0015
Inter-story drift ratio
(a)
1.5 600
MSC.Marc model
1 500
Top displacement (m)

Simplified model

Building height (m)

0.5 400

0 300

-0.5 200

-1 100 MSC.Marc model

Simplified model
-1.5 0
0 10 20 30 40 0 0.2 0.4 0.6 0.8
Time (s) Displacement (m)
(b) (c)
600 600

500 500
Building height (m)

Building height (m)

400 400

300 300

200 200

100 MSC.Marc model 100 MSC.Marc model

Simplified model Simplified model
0 0
0 0.002 0.004 0.006 0.008 0.01 0 0.002 0.004 0.006 0.008 0.01
Inter-story drift ratio Inter-story drift ratio
(d) (e)
Fig. 9. Comparison of the time history analysis results between the simplified and refined FE models: (a) The story drift ratio envelopes of the fully braced scheme under
70 gal. (b) The roof displacement of the fully braced scheme under 400 gal. (c) The story displacement envelopes of the fully braced scheme under 400 gal. (d) The story drift
ratio envelopes of the fully braced scheme under 400 gal. (e) The story drift ratio envelopes of the half-braced scheme under 400 gal.

to be acceptable, which conclusively validate the reliability of the simplified model. At the preliminary design stage, it is rather diffi-
proposed simplified model even at the nonlinear stage. Such a time cult and time consuming to create a refined FE model. Therefore, it
history analysis using the simplified model takes only 18 min, is almost impossible to use the refined FE model to assist in design-
whereas the analysis using the refined FE MSC.Marc model ing at the preliminary design stage. In contrast, the simplified mod-
requires more than 36 h on the same computer, which is a els developed in this study can be readily established based on the
2.20 GHz Intel(R) Xeon(R) CPU E5-2430 and 48 GB of memory. preliminary design information of the structural system and the
The computational time is reduced by more than 100 times. primary components. As a result, the seismic behavior of the pre-
The above-described validations confirm that the simplified liminary design schemes can be efficiently predicted through the
models of both schemes are capable of predicting the nonlinear linear and nonlinear analysis of this simplified model, thus facili-
story displacements and story drift ratio as well as the critical tating the development of an improved design.
mechanical characteristics. This capability of the simplified models
will reliably and effectively assist researchers and designers in 5. Comparison of the plastic energy dissipation
understanding the seismic behaviors of different design schemes.
Note that the above-mentioned validation of the simplified To satisfy the demands for resilience, the earthquake-induced
models based on a comparison with the refined FE models is damage should be uniformly distributed in the replaceable compo-
mainly for research purposes of validating the accuracy of the nents along the height of the building. Severe damage inside the
 X. Lu et al. / Engineering Structures 110 (2016) 116–126 123

Table 5 building is subjected to the service load; as a result, they can be

Plastic energy dissipation contribution of different structural components under conveniently repaired or replaced after an earthquake [42,43].
different earthquake intensities.
The plastic energy dissipation contribution of the outriggers and
PGA (gal) Plastic energy dissipation of different structural components perimetric trusses in the half-braced scheme is considerably
Coupling Trusses Core- Mega Mega higher, with a value of up to 28.39%, because the outriggers bare
beam (%) (%) tube (%) brace (%) column (%) significant shear forces due to the different patterns of deformation
200 Fully 97.71 2.29 0.00 0.00 0.00 between the mega columns and core tubes. Both the outriggers and
Half 67.15 32.77 0.00 0.08 0.00 perimetric trusses can be designed to be repairable or replaceable
310 Fully 96.15 3.80 0.05 0.00 0.00 after an earthquake to satisfy the requirement for resiliency.
Half 64.8 32.05 0.42 2.67 0.00 The braces are normally regarded as replaceable components in
400 Fully 93.80 5.07 0.56 0.55 0.02 conventional buildings [44]. However, the mega braces in this
Half 63.25 28.39 2.92 5.44 0.00 super-tall building are difficult to be replaced due to its extremely
large weight (>150 tons). The mega braces in the fully braced
scheme basically remains elastic and dissipates little plastic energy
key components and damage concentration along the height of the (60.55%) when subjected to all three considered intensities. In con-
building should be avoided. The cumulative hysteretic energy dis- trast, the mega braces in the half-braced scheme dissipate 5.44% of
sipated by various structural components is commonly considered the plastic energy when PGA = 400 gal. This result indicates that
as a good indicator of earthquake-induced damage. Hence, the hys- the mega braces in the half-braced scheme suffer moderate dam-
teretic energy dissipated by different components and the hys- age and should be repaired or replaced after an earthquake, which
teretic energy distribution along the height of the two schemes is rather difficult to implement.
are compared using the simplified models validated above. The mega columns and core tubes are key components of the
To investigate the influence of the seismic intensity on the plas- building that are difficult to repair or replace. The plastic energy
tic energy dissipation distribution, three different seismic intensi- dissipation contribution of these components in both schemes
ties are adopted: PGA = 220 gal (i.e., the MCE in the Intensity 7 increase with increasing seismic intensities. The plastic energy dis-
Region), PGA = 310 gal (i.e., the MCE in the Intensity 7.5 Region) sipation percentage of the mega columns and core tubes are 0.02%
and PGA = 400 gal (i.e., the MCE in the Intensity 8 Region, which and 0.56%, respectively, for the fully braced scheme when
is the design intensity of this super-tall building) [36]. Moreover, PGA = 400 gal. This result indicates that these key components in
due to the significant randomness in the input ground motion this design scheme remain elastic and dissipate little plastic
records, the seismic response obtained from a single specified energy, thereby resulting in minor damage and enabling immedi-
ground motion record may result in a large deviation. Hence, 22 ate re-occupancy of this building without any repair after earth-
far-field ground motion records recommended by FEMA P695 quakes. In contrast, the plastic energy dissipation ratio of the
[40] are adopted as the basic ground motion set. The plastic energy core tubes is up to 2.92% for the half-braced scheme when
dissipation behaviors are discussed based on the mean value of the PGA = 400 gal, which means that a certain extent of damage occurs
response parameters obtained from the 22 ground motions. Note inside these core tubes, thus requiring repair, which will delay the
that the simplified models offer notable advantage in terms of functional recovery of the building.
computational efficiency. Although both the simplified models As described above, the fully braced scheme is proven to be
and refined FE models are established and used in this work, the more effective in focusing the damage to occur at the readily
nonlinear dynamic analyses of 22 ground motions under three dif- replaceable components (i.e., coupling beams and perimetric
ferent seismic intensities using the refined FE models will require trusses) instead of the key components that are more difficult to
more than 5000 h, which is unacceptable for the design of a repair (i.e., mega columns, core tubes and mega braces), thus
super-tall building. In contrast, using the simplified models ensuring the functional resilience of super-tall buildings after
requires only approximately 40 h; such a reduced computational strong earthquakes. Therefore, the fully braced scheme is a better
burden for the simulations will be beneficial for the comparison choice than the half-braced scheme for the design of this super-
of different design schemes. tall building.

5.2. Total plastic energy dissipation distribution along the building

5.1. Plastic energy dissipation analysis of different components height

Nonlinear time history analyses of the three above-mentioned The total plastic hysteretic energy dissipation distribution along
seismic intensities and 22 ground motion records are performed the height of the building based on the half-braced scheme is illus-
for both schemes. The mean percentages of the plastic energy dis- trated in Fig. 10 and is named the ‘‘Total” curve. Because the
sipation contribution of different structural components are pre- lateral-force resisting system significantly changes at Zone 5 (i.e.,
sented in Table 5. the mega braces installed in the lower four zones no longer exist
The coupling beams dissipate most of the plastic energy in both in the upper four zones), large deformations and higher plastic
schemes (no less than 93.80% in the fully braced scheme and no energy dissipation are present in the upper 4 zones of the building
less than 63.25% in the half-braced scheme), although the percent- as opposed to the lower 4 zones. Furthermore, the dimensions of
ages slightly decrease with increased seismic intensity. Thus, the the cross section of each floor gradually decrease from Zone 5 to
coupling beams are the primary plastic energy dissipation compo- Zone 6 and subsequently gradually increase from Zone 6 to Zone
nent in both schemes. Because the coupling beams are replaceable 7, thus leading to more plastic energy dissipation in these zones
components, they can be designed to be readily replaced so that compared to the lower 4 zones. As a result, these zones have the
the super-tall building can be easily repaired after an earthquake potential to suffer relatively severe damage and are rather difficult
[41], which satisfies the requirement for resiliency. to repair after an earthquake.
The perimetric trusses in the fully braced scheme dissipate The total plastic hysteretic energy dissipation distribution along
approximately 5.07% of the total plastic energy when PGA = 400 gal, the height of the fully braced scheme is illustrated in Fig. 11 and is
which means that moderate damage occurs in these elements. also called the ‘‘Total” curve. Because there are only 6 stories in
Note that the perimetric trusses bear a limited load when the Zone 8, the plastic hysteretic energy dissipated in Zone 8 is
124 X. Lu et al. / Engineering Structures 110 (2016) 116–126

9 8
8 7
7
6
6
5
5

Zone
Zone

4
4
Mega column 3 Mega column
3 Core tube Core tube
2 Mega brace 2 Mega brace
Truss Truss
1 Coupling beam 1 Coupling beam
Total Total
0 0
0 10 20 30 40 0 10 20 30 40
Energy dissipation (h106N.m) Energy dissipation (h106N.m)
(a) (a)
9 8
8 7
7
6
6
5
5
Zone

Zone
4
4
Mega column 3 Mega column
3 Core tube Core tube
2 Mega brace 2 Mega brace
Truss Truss
1 Coupling beam 1 Coupling beam
Total Total
0 0
0 15 30 45 60 0 15 30 45 60
Energy dissipation (h106N.m) Energy dissipation (h106N.m)
(b) (b)
9 8
8
7
7
6
6
5
5
Zone

Zone

4
4
Mega column 3 Mega column
3 Core tube Core tube
2 Mega brace 2 Mega brace
Truss Truss
1 Coupling beam 1 Coupling beam
Total Total
0 0
0 25 50 75 100 0 25 50 75 100
Energy dissipation (h106N.m) Energy dissipation (h106N.m)
(c) (c)
Fig. 10. Total plastic energy dissipation distribution and plastic energy dissipation Fig. 11. Total plastic energy dissipation distribution and plastic energy dissipation
in different components along the structural height of the half-braced scheme for in different components along the structural height of the fully braced scheme
different seismic intensities: (a) PGA = 220 gal. (b) PGA = 310 gal. (c) PGA = 400 gal. under different seismic intensities: (a) PGA = 220 gal. (b) PGA = 310 gal. (c)
PGA = 400 gal.

combined with that in Zone 7 as a single zone (i.e., Zone 7). Compared
to the half-braced scheme, a relatively uniform distribution of the plastic energy in a single zone of the fully braced is only 50.84%
total plastic energy is observed for the fully braced scheme. The of that of the half-braced scheme when PGA = 400 gal.
following two characteristics of fully braced scheme account for The above discussions illustrate that the fully braced scheme
this difference: (1) the mega braces are installed along the height induces a more uniform plastic energy dissipation distribution
of the entire building, which leads to a more uniform distribution compared to the half-braced scheme. In addition, the fully braced
of the structural stiffness compared to the half-braced scheme; (2) scheme avoids severe damage concentration in a single zone. All
the strength-based design of the fully braced scheme is more of these plastic energy dissipation characteristics improve the per-
reasonable; as a result, the degree of nonlinearity in each zone formance of the building in terms of resiliency and enable rapid
is approximately identical. Moreover, the maximum dissipated recovery in the aftermath of a strong earthquake.
 X. Lu et al. / Engineering Structures 110 (2016) 116–126 125

5.3. Plastic energy dissipation contribution of each component along 8142024) and the National Key Technology R&D Program (No.
the building height 2015BAK17B03).

The average plastic energy dissipations in different components References

along the height of the half-braced scheme are presented in Fig. 10
and are referred to as ‘‘Mega column”, ‘‘Core tube”, ‘‘Mega brace”, [1] Almufti I, Willford M. Resilience-based earthquake design initiative for the
next generation of building. Rating system for the next generation of buildings;
‘‘Trusses”, and ‘‘Coupling beam” curves. Because the mega braces 2014.
are only installed in the lower 4 zones of the building, the plastic [2] Takewaki I, Moustafa A, Fujita K. Improving the earthquake resilience of
energy dissipated in the mega braces is located in the lower 4 buildings. Springer; 2013.
[3] Christopoulos C, Montgomery M. Viscoelastic coupling dampers (VCDs) for
zones. In contrast, most of the plastic energy is dissipated by the
enhanced wind and seismic performance of high-rise buildings. Earthquake
core tube, trusses and coupling beams located in the upper 4 zones. Eng Struct Dynam 2013;42(15):2217–33.
Significant damage concentration in the coupling beams is [4] Resilient Washington State. A framework for minimizing loss and improving
statewide recovery after an earthquake. Washington State Seismic Safety
observed in Zones 4, 5 and 7, which means that the half-braced
Committee Emergency Management Council; 2012.
scheme does not make full use of the energy dissipation capacity [5] Giovinazzi S, Wilson T, Davis C, Bristow D, Gallagher M, Schofield A. Lifelines
of all of the coupling beams along the height of the building. performance and management following the 22 February 2011 Christchurch
The average plastic energy dissipation contribution in different earthquake, New Zealand: highlights of resilience. Bull NZ Soc Earthquake Eng
2011;44(4).
components along the height of the fully braced scheme building is [6] Fan H, Li QS, Tuan Alex Y, Xu LH. Seismic analysis of the world’s tallest
presented in Fig. 11. The trusses participated in the energy dissipa- building. J Constr Steel Res 2009;65(5):1206–15.
tion at the upper 3 zones of the building. The results clearly indi- [7] Li QS, Zhi LH, Tuan AY, Kao CS, Su SC, Wu CF. Dynamic behavior of Taipei 101
Tower: field measurement and numerical analysis. J Struct Eng 2010;137
cate that the primary plastic energy dissipation component is the (1):143–55.
coupling beams. In addition, the plastic energy dissipated by cou- [8] Poon DCK, Hsiao LE, Zhu Y, Joseph L, Zuo S, Fu GY, Ihtiyar O. Non-linear time
pling beams is uniformly distributed along the height of the build- history analysis for the performance based design of Shanghai Tower. Struct
Congr 2011:541–51.
ing, which avoids the concentration of damage. [9] Lu XL, Su NF, Zhou Y. Nonlinear time history analysis of a super-tall building
As described above, the fully braced scheme induces a more with setbacks in elevation. Struct Des Tall Spec Build 2013;22(7):593–614.
uniform plastic energy dissipation in the coupling beams along [10] Jiang HJ, Fu B, Liu LE, Yin XW. Study on seismic performance of a super-tall
steel–concrete hybrid structure. Struct Des Tall Spec Build 2014;23(5):334–49.
the building height compared to the plastic energy dissipation of [11] Li MK, Lu X, Lu XZ, Ye LP. Influence of soil–structure interaction on seismic
the half-braced scheme. The fully braced scheme makes full use collapse resistance of super-tall buildings. J Rock Mech Geotech Eng 2014;6
of the energy dissipation capacity of the coupling beams in each (5):477–85.
[12] Lu XZ, Xie LL, Guan H, Huang YL, Lu X. A shear wall element for nonlinear
zone of the building.
seismic analysis of super-tall buildings using OpenSees. Finite Elem Anal Des
2015;98:14–25.
[13] Pan TC, Brownjohn JMW, You XT. Correlating measured and simulated
dynamic responses of a tall building to long-distance earthquakes.
6. Conclusions
Earthquake Eng Struct Dynam 2004;33(5):611–32.
[14] Lu XL, Zhu JJ, Zou Y. Study on performance-based seismic design of Shanghai
A 2D simplified model is developed for the seismic analysis of a World Financial Center Tower. J Earthquake Tsunami 2009;3(4):273–84.
[15] ANSYS theory reference. Ansys Inc; 1999.
super-tall mega-braced frame-core tube building and applied to
[16] Perform-3D: nonlinear analysis and performance assessment for 3D
compare two design schemes of this building in terms of resilience. structures: version 4. Computers and Structures Inc; 2006.
The following conclusions can be drawn: [17] LS-DYNA theory manual, vol. 3. Livermore Software Technology Corporation;
2006.
[18] ABAQUS/standard: user’s manual. Hibbett, Karlsson & Sorensen Corporation;
(1) Compared to the refined FE model, the proposed simplified 1998.
model reduces the computational time by a factor of 100 [19] Fu F. Progressive collapse of high-rise building with 3-D finite element
while still providing a reliable accuracy in predicting the modeling method. J Constr Steel Res 2009;65(6):1269–78.
[20] Lange D, Röben C, Usmani A. Tall building collapse mechanisms initiated by
seismic performances of the two considered design schemes. fire: mechanisms and design methodology. Eng Struct 2012;36:90–103.
(2) The plastic energy dissipation characteristics of both [21] Kotsovinos P, Usmani A. The World Trade Center 9/11 disaster and progressive
schemes are predicted and compared using the proposed collapse of tall buildings. Fire Technol 2013;49(3):741–65.
[22] Jiang HJ, Lu XL, Liu XJ, He LS. Performance-based seismic design principles and
simplified model. The fully braced scheme is found to induce structural analysis of Shanghai Tower. Adv Struct Eng 2014;17(4):513–28.
a uniform plastic energy dissipation distribution and effec- [23] Jiang J, Chen LZ, Jiang SC, Li GQ, Usmani A. Fire safety assessment of super tall
tively enables the energy dissipation to be located in readily buildings: a case study on Shanghai Tower. Case Stud Fire Saf 2015;4:28–38.
[24] Mazzoni S, McKenna F, Scott MH, Fenves GL. OpenSees command language
replaceable components. In contrast, significant plastic
manual. Pacific Earthquake Engineering Research (PEER) Center; 2006.
energy dissipation concentration is observed at the upper [25] Lu X, Lu XZ, Zhang WK, Ye LP. Collapse simulation of a super high-rise building
4 zones of the half-braced scheme, and the core tubes are subjected to extremely strong earthquakes. Sci China Technol Sci 2011;54
(10):2549–60.
found to suffer significant damage; as a result, the functional
[26] Lu XZ, Lu X, Guan H, Zhang WK, Ye LP. Earthquake-induced collapse simulation
recovery of the building will be delayed. Overall, the fully of a super-tall mega-braced frame-core tube building. J Constr Steel Res
braced scheme provides a better seismic resilient perfor- 2013;82:59–71.
mance compared to the half-braced scheme. [27] MSC.Marc User’s manual. Santa Ana (CA): MSC Software Corporation; 2007.
[28] Lu XL, Gong ZG, Weng DG, Ren XS. The application of a new structural control
concept for tall building with large podium structure. Eng Struct 2007;29
The outcome of this study will provide guidance and act as a (8):1833–44.
reference for the resilience-based earthquake design of super-tall [29] Carpinteri A, Lacidogna G, Cammarano S. Structural analysis of high-rise
buildings under horizontal loads: a study on the Intesa Sanpaolo Tower in
buildings. The proposed simplified model can be used to compare Turin. Eng Struct 2013;56:1362–71.
various design schemes at the preliminary design stage. [30] Lu X, Lu XZ, Guan H, Ye LP. Collapse simulation of reinforced concrete high-rise
building induced by extreme earthquakes. Earthquake Eng Struct Dynam
2013;42:705–23.
[31] Lee S, Tovar A. Outrigger placement in tall buildings using topology
Acknowledgements optimization. Eng Struct 2014;74:122–9.
[32] Connor JJ, Pouangare CC. Simple-model for design of framed-tube structures. J
The authors are grateful for the financial support received from Struct Eng 1991;117(12):3623–44.
[33] Luco N, Cornell CA. Structure-specific scalar intensity measures for near-
the National Natural Science Foundation of China (Nos. 51578320 source and ordinary earthquake ground motions. Earthquake Spectra 2007;23
and 51261120377), the Beijing Natural Science Foundation (No. (2):357–92.
126 X. Lu et al. / Engineering Structures 110 (2016) 116–126

[34] Meftah SA, Tounsi A, El-Abbas AB. A simplified approach for seismic [40] FEMA P695: quantification of building seismic performance factors. Redwood
calculation of a tall building braced by shear walls and thin-walled open City (CA): Applied Technology Council; 2009.
section structures. Eng Struct 2007;29(10):2576–85. [41] Fortney PJ, Shahrooz BM, Rassati GA. Large-scale testing of a replaceable ‘‘fuse”
[35] Lu X, Lu XZ, Sezen H, Ye LP. Development of a simplified model and seismic steel coupling beam. J Struct Eng 2007;133(12):1801–7.
energy dissipation in a super-tall building. Eng Struct 2014;67:109–22. [42] Chen Y, McFarland DM, Wang Z, Spencer BF, Bergman LA. Analysis of tall
[36] GB50011-2010. Code for seismic design of buildings. Beijing: Ministry of buildings with damped outriggers. J Struct Eng 2010;136(11):1435–43.
Housing and Urban-Rural Development of the People’s Republic of China; [43] Zhou Y, Zhang CQ, Lu XL. Earthquake resilience of a 632-meter super-tall
2010. building with energy dissipation outriggers. In: Proceeding of the 10th
[37] JGJ 3-2010. The technical specification for concrete structures for tall national conference on earthquake engineering. Anchorage (AK): Earthquake
buildings. Beijing: Ministry of Housing and Urban-Rural Development of the Engineering Research Institute; 2014.
People’s Republic of China; 2010. [44] Chen CH. Performance-based seismic demand assessment of concentrically
[38] Lu XZ, Ye LP, Miao ZW. Elasto-plastic analysis of buildings against braced steel frame buildings. Ph.D. Dissertation, University of California,
earthquake. Beijing: China Architecture & Building Press; 2009 [in Chinese]. Berkeley, USA; 2010.
[39] Habibullah A. ETABS-Three dimensional analysis of building systems, user
manual. Computers and Structures Inc; 1997.