Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct
a r t i c l e
i n f o
Article history:
Received 13 May 2014
Revised 13 August 2014
Accepted 13 August 2014
Keywords:
Cable-net faade
Air blast load
Viscoelastic devices
Frictional devices
Energy dissipation
SDOF approach
Advanced numerical modelling
Dynamic nonlinear simulations
a b s t r a c t
Based on numerical and analytical results of previous literature contributions, the paper investigates the
dynamic behaviour of a structural two-way straight cable-net faade subjected to medium-level blast
loads. Numerical studies are dedicated, specically, to a cable-net prototype already experimentally
investigated under seismic loads. Several numerical nite-element (FE) models are developed in
ABAQUS/Standard. A geometrically simplied, lumped-mass FE-model (M01), a detailed cable-glass model
(M02) and a further simplied but computationally efcient cable-glass model (M03) are presented. The
models are assessed and calibrated to test data of literature, as well as to simplied SDOF analytical formulations, in order to investigate the dynamic response of the faade under high-rate impulsive loads.
Since axial forces in the cables and maximum stresses in the glass panels abruptly increase when the
explosion occurs, specic viscoelastic and frictional devices are applied at the connection between glass
and cables, as well as at the end of the cables. The main advantage of the presented technological solution
is given by the partial dissipation of the incoming energy. In the rst case, viscoelastic connectors add
exibility to the point-supported glass elements, hence reducing the amount of incoming blast energy.
In the latter case, frictional devices dissipate through friction mechanisms the strain energy stored by
the bearing cable-net. A proper combination of these multiple devices, in conclusion, can manifest a
marked increase of structural efciency for the examined structural typology.
2014 Elsevier Ltd. All rights reserved.
1. Introduction
The rising desire of transparency in modern buildings markedly
increased the use of glass as a construction material able to carryon loads in the form of innovative structural members, roofs, curtain walls. As a result, during the last decades traditional design
concepts rapidly evolved, in conjunction with ongoing manufacturing developments, in order to provide appropriate technological
solutions to continuous advanced architectural demands [18].
Glass faades, specically, highlighted several technique
improvements and optimizations, aiming to cover wide surface
often characterized by irregular shapes.
While the traditional unitized curtain wall or stick systems,
for example, typically consists of modular units in which the glass
panels are sustained by a metallic frame composed of aluminium
transoms and mullions rigidly connected to the structural backup
(e.g. concrete slab), cable-supported faades are especially used
when large volumes and surfaces must be enclosed. Although
rstly built about 20 years ago, due to their transparency, easy
Corresponding author. Tel.: +39 040 558 3842.
E-mail addresses: bedon@dicar.units.it, c.bedon@libero.it (C. Bedon).
http://dx.doi.org/10.1016/j.engstruct.2014.08.023
0141-0296/ 2014 Elsevier Ltd. All rights reserved.
277
Fig. 1. Examples of two-way straight cable-net faades. (a) New Beijing Poly Plaza Cable-Net Wall [9] and (b) Kempinski Hotel at Munich Airport [10].
Careful attention was paid especially for the study of the interaction between the glass panels and metal connectors [11,12]. Feng
et al. [13,14] deeply investigated the behavioural trends of cablenet faades subjected to seismic events. In [15], they also highlighted the effects of glass panels stiffness on the global dynamic
response of steel-glass faades, providing critical discussion and
comparison of experimental and numerical predictions. Studies
presented in [16,17] have been dedicated to the nonlinear behaviour of cable-supported faades under wind loads. Main features
and comparative discussion of cable-truss glass faades are proposed in [18].
While extended rigorous investigations are available for steelglass faades under ordinary loads, further knowledge is currently
required to understand and optimize their structural behaviour
under exceptional loading conditions. The dynamic behaviour of
glazing systems under high-strain rate and impulsive loads such
as explosion, for example, still represents a research topic of great
interests and a challenge for designers. Wide series of shock tube
experiments, analytical investigations and FE numerical studies,
in this context, have been dedicated to single laminated glass
panels subjected to air blast loads [1925]. Weggel et al. [26]
investigated the dynamic behaviour of unitized curtain walls under
low-level explosions. Numerical parametric simulations have been
discussed in [27], in order to provide pressure-impulse (PI) design
curves for the blast verication of laminated glass curtain walls.
Analytical and numerical studies have been recently proposed for
the dynamic buckling verication of monolithic and laminated
glass beam-like and columns under blast loads [28].
For the specic topic of cable-supported faades, Teich et al.
[29] focused on the analysis of the dynamic response of these systems under explosive events, emphasizing the inuence of the negative phase of a blast wave pressure on their global structural
behaviour. Wellershoff et al. [30] highlighted, through application
to the case study of the World Trade Center building, the typical
behaviour of a single-way straight cable supported faade under
blast loads, suggesting possible mitigation techniques for the limitation of damages and improvement of the expected dynamic
performances.
As also highlighted in [31], the design of a blast resisting cablesupported faade strongly differs from the conventional design of a
glazing system subjected to ordinary loads. When an explosion
occurs, the structural components of a cable-supported faade
are in fact affected by exceptional stresses both compressive
and tensile, due to mainly exural deformations involved by the
acting impulsive pressures and extreme deections. While glass
presents typical high nominal compressive strength, however,
maximum tensile stresses should be properly checked and limited,
to preserve the faade integrity and avoid brittle failure mechanisms in it.
In that earlier contribution [31], appropriately designed elastoplastic devices were proposed to be installed at the ends of the
vertical bearing cables, in order to prevent their possible brittle collapse under high-strain impulsive loads. The same cable-supported
faade markedly exible due to the presence of a single layer of
pre-stressed straight cables has been successively further
improved [32] by implementation of combined multiple devices,
namely consisting in elastoplastic devices at the cables ends and
opportune viscoelastic connectors introduced at the connection
between the glass panels and the cables [33]. Certainly, under
exceptional loading scenarios, the structural dynamic performance
of these structural systems should be optimized in order to avoid,
or at least minimize, possible injuries and structural damages, hence
preserving the activities within the building.
In this paper, based on extended numerical and analytical
results of previous literature contributions [3133], the behavioural trends of a two-way straight cable-net faade subjected to
medium-level blast loads are numerically and analytically investigated. In particular, studies are dedicated to the dynamic behaviour of a cable-net faade prototype already analyzed in [15]
under seismic events.
Based on [3133], multiple typologies of dissipative devices are
introduced both at the connection between the glass panels and
the pretensioned cable-net, as well as at the cable ends.
Modal analyses and nonlinear dynamic incremental simulations
are performed on a lumped-mass model (M01), a geometrically
detailed cable-glass model (M02) and a further simplied but computationally efcient cable-glass model (M03), able to take into
account the bending stiffness contribution of glass panes in the
form of equivalent beam elements. Parametric dynamic simulations are then performed, in order to highlight the efciency and
the criticalities of each possible solution.
As shown, by means of appropriate energy dissipating components, the increase of maximum axial forces in the bearing cables
and the maximum tensile stresses in the glass panels can be markedly reduced, hence the structural stability of the faade system
can be prevented.
278
Fig. 3. Schematic view and cross-section of the typical stainless steel clamping joint
[15].
t
kp
t
pt p0 1 e td ;
td
0 t td ;
p0) and t t
d (e.g. end of the positive phase, with p(t d ) = 0).
Three main parameters can affect the exponential pressure
function given by Eq. (1), namely (i) the distance between the
explosion source and the invested building (stand-off distance,
dTNT), (ii) the height of the explosion source from ground (hTNT)
and (iii) the quantity of explosive (equivalent mass of TNT, MTNT).
At the same time, the overpressure peak p0 is dened in accordance
with recommendations provided by the TM 5-1300 standard [37].
kp 5:2777Z 1:1975 ;
dTNT/MTNT1/3
2
1/3
being Z =
[m/Kg ] the scaled distance.
While the positive phase namely represents the ambient pressure increase and modication due to the occurring explosion,
the negative phase of an ideal blast event is characterized by a
pressure amount which is lower than the reference ambient pressure. This negative phase is typically described in the form of a triangular shape associated to a negative under-pressure peak pneg
and a total duration ttot [36].
In this paper, a medium-level air blast load (Level C-GSA [34])
was taken into account to study the behavioural trends of the
investigated faade prototype. The main parameters of the mentioned design blast load, based on recommendations of [34] and
[37], are the impulse per unit of surface (i = 272 kPa ms), the overpressure peak (p0 = 42.9 kPa), the duration of the positive phase
(t
d 0:0165 s) and the total impulse duration (ttot = 0.0748 s).
The resulting time-varying pressure function is proposed in Fig. 4.
2.3. Dissipative devices
A structural system subjected to an explosive load should be
generally able to be as much as possible exible and dissipative,
Fig. 2. Front and side elevations of the studied cable-net faade [15].
Pressure [Pa]
6x10 4
BLAST LOAD
MTNT= 50Kg
dTNT= 30m
4x10 4
2x10 4
-2x10 4
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
Time [s]
Fig. 4. Blast load time-varying pressure function (Level C-GSA [34]).
kd
G0 AVE
hVE
gkd
;
x1
and
cd
279
Fig. 5. (a) Example of viscoelastic (VE) spider connector, cross-section; (b) example of frictional (RP) device, rendering.
280
The same analytical approach, based on analogous energy considerations, is extended in this work to steel-glass faades sustained by a planar net of pretensioned straight cables (Fig. 6).
The MDOF cable-net object of study, specically, is described in
the form of an equivalent SDOF system whose dynamic behaviour
is characterized in terms of equivalent mass m* and the equivalent
stiffness k*, being rationally negligible, under high-rate impulsive
loads such as explosions, possible structural and aerolastic damping contributions. Differing from [31], moreover, the presence of
a two-way cable-net rather than a single-way bracing system is
properly taken into account. As a result, the equivalent mass of
the cable-net faade is calculated as [40]:
M
m
4
2
4Htot
p
2 sin
lcable
ncables 1
being lcable the length of cables and ncables signifying the number of
cables arranged in each main direction of the net (or the average
number of vertical ncables,v and horizontal ncables,h cables, when
ncables,v ncables,h). In Eq. (6), nally, Htot represents the total initial
pretension applied to the bearing system and should be calculated
as:
Htot H0 ncables;v ncables;h
r
m
T 1 2p
k
T 1
being
strictly related to the level of pretension force in the cablenet.
While Eq. (8) provides the fundamental period of vibration of
the faade affected by the initial pretension force Htot H0,tot only,
umax
s
r
I2
1
I
Mk
Mk
and
v max
s
I2
2I
Mm M
10
Hblast;tot
8 Ecable Acable 2
umax
2
3
lcable
11
with Ecable the Youngs modulus of steel constituting the cables and
umax the maximum deection given by Eq. (9). As a result, the maximum total pretension force affecting the cable-net after the explosion results:
12
281
Hmax
H0;tot Hblast;tot
Hmax;tot
13
By replacing Hmax in Eq. (6), the corresponding increased equivalent stiffness k*blast can be consequently calculated and taken into
account in Eq. (8), hence resulting in a modied fundamental per*
iod of vibration T1,blast
markedly lower than T1*.
Main results obtained by application of the discussed analytical
procedure to the studied faade subjected to a Level C air blast load
are collected in Table 1.
The assigned design explosion involves large deections (1/20
the total span), as well as high maximum velocities (15.4 m/s)
and axial forces in the cables (1.02 HR, with HR = 160 kN the characteristic strength of the adopted cables).
In the same table, additional analytical calculations are proposed for the same faade under a Level B-GSA (low-level explosion) and a Level D-GSA blast load (high-level explosion),
respectively. As shown, independently on the intensity of the
assigned explosive loads, due to their impulsive and high-rate nature compared to ordinary loads extreme deformations and
forces are exerted on the examined structural system. Further
advanced investigations are consequently required.
Table 1
Analytical estimation of maximum dynamic effects on the studied cable-net faade subjected to air blast loads. i: impulse per unit of surface; p0: overpressure peak.
Parameter
Description
Level B (low)
i = 168 kPa ms
p0 = 30.4 kPa
Level C (medium)
i = 272 kPa ms
p0 = 42.9 kPa
Level D (high)
i = 439 kPa ms
p0 = 62.6 kPa
T1 [s]
umax [m]
vmax [m/s]
Hmax [kN]
T1,blast [s]
0.578
0.526
9.47
109
0.350
0.578
0.699
15.33
163
0.286
0.578
0.912
24.74
249
0.231
282
pattern and a free meshing technique were used. In this case, the
minimum element size was set equal to lmesh 0.003 m (Fig. 7).
The nal choice for the adopted mesh pattern globally derived
from an appropriate balance between mesh renement, corresponding predictions in the glass panels and main purpose of this
research work. The investigation of possible damage mechanisms
and post-cracked behavioural trends for the same faade under
blast loads, hence requiring the implementation of specic damage
mechanical models for glass (e.g. [2327,41]), would in fact require
an extremely rened mesh pattern, being those material models
(e.g. the brittle cracking ABAQUS option [39]) generally strictly
dependent on the characteristic size of elements [42].
Concerning the mechanical characterization of toughened glass,
in accordance with [3133], an isotropic, linear elastic characteristic behaviour was in fact dened, with Eglass = 70 GPa the Youngs
modulus, qglass = 2490 kg/m3 the density and mglass = 0.23 the Poissons ratio. As a result, possible brittle cracking of glass panes
was fully neglected. From a numerical point of view, this assumption directly manifested in geometrical and mechanical simplied
modelling hypotheses, but anyway in reasonable FE-predictions for
the studied faade, as discussed in Sections 45. Glass is in fact a
material characterized by typical high resistance in compression
and brittle behaviour in tension, being its nominal quasi-static
characteristic strength generally comprised between 45 MPa
(annealed glass) and 120 MPa (fully tempered glass) often representative of the main inuencing parameter for designers [43].
Under high-strain loads, however, conventional quasi-static tensile
strengths are usually replaced with magnied corresponding values able to take into account strain rate effects. Under explosions,
having a typical duration of 0.0050.025 s, the magnied tensile
strength recommended by design manuals and standards for
toughened or thermally tempered glass as for the studied faade
is 160 MPa [44]. Further references are available for glass under
blast loads, but certainly the estimation of the actual tensile
strength under high-strain rates is rather difcult to exactly predict. Blast design strength values for tempered glass can be in fact
found comprised between 145 MPa and 209 MPa [45,46], thus a
magnifying dynamic coefcient at least equal to 1.2 is generally
accepted for FT glass. Comparative discussion of technical documents recommendations and assumptions is provided in [47].
Experimental studies have been performed also in [48], to investigate the dynamic increase of both compressive and tensile
strengths for glass under high-strain loads. Data measurements
and high-speed camera acquisitions obtained from tests performed
on small cylindrical glass specimens in the range from 103 to
103 s1 highlighted that the compressive strength and Young modulus of glass are slightly sensitive to strain-rate effects. Conversely,
the quasi-static tensile strength manifested marked dependency
on strain-rate effects and an appreciably increased dynamic
resistance.
In these hypotheses, the occurring of possible failure mechanisms in the M02 FE-model was indirectly prevented by continuously monitoring the maximum tensile stresses in glass, hence
ensuring the exceeding of the reference magnied resistance (Sections 45).
Careful consideration was then dedicated to the modelling of
the clamping joints, since constituting the key component of
numerical simulations. The typical metallic joint consisted of four
rigidly connected stainless steel beams (B31 elements), representative of the nominal geometry of the metal fasteners. A weld connector, able to provide a fully bonded connection between the
relative displacements (ux, uy, uz) and rotations (rx, ry, rz) of these
four beams was used, hence ensuring the desired restraining effect
between them. Concerning the glass panel-to-joint and joint-tocable structural interaction, each clamping joints was attached to
the adjacent glass panels by means of appropriate join and slideplane connectors (Fig. 8). The rst typology of connectors, able to
prohibit possible relative displacements (ux, uy, uz) in the interested
nodes, was inserted at the top-left corner of each glass panel. Slideplane connectors able to guarantee small relative displacements in
the plane of glass sheets were used, conversely, for the three other
corners of each panel. In this manner, the attainment of peaks of
stress at the connection points of the glass panels was prevented,
especially when applying the initial pretension force in the cables.
Finally, the structural interaction between the so modelled
clamped joints and the pretensioned cable-net was guaranteed by
means of a series of two additional connectors (Fig. 8, Connector
1 and Connector 2), able to provide in presence of external
pressures orthogonal to the surface the faade the same out-ofplane deection for the glass panels and the cable-net in the interested nodes.
As also discussed in [15,31,32], the mass of clamping joints was
neglected. At the same time, since characterized by a negligible
stiffness [26], the layers of structural silicone sealant interposed
between the adjacent glass panels were not taken into account.
With reference to possible structural and aerolastic damping
effects [15], shock excitation experiments performed on the
cable-net prototype with the glass panels highlighted as expected
a marked increase of dissipative capabilities for the studied system (nexp 3.4%), compared to the cable-net alone.
In this context, it should be noticed that the obtained experimental value is rather in agreement with simple analytical calculations derived from literature. Assuming the total damping ratio
nTOT = nstructural + naerolastic 3.4% as the sum of structural and aerolastic damping terms, the test result can be in fact rationally
assessed, being nstructural 1% [49,50] the conventional structural
contribution and naerolastic 2.43% the aerolastic term, with:
naerolastic
14
283
p 3
2l
1:244l
beq
2
2:4 1 m2glass l
15
p
The t beq 2l diagonal beams were then linked together at
their mid-span node by means of a tie constraint, able to provide
full coupling and interaction between them. At the same time,
the self-weight of glass was described in the form of lumped
masses mi.
Although the Grashofs method represents an approximate
approach, preliminary static, nonlinear geometry numerical analyses performed on a single point-supported shell glass panel and
Tshell = 0.07916 s
(a)
Tbeam = 0.07938 s
(b)
Fig. 9. Fundamental modal shape of a point-supported glass panel. (a) shell model
and (b) equivalent diagonal beam model (ABAQUS/Standard).
1.0
Modal amplitude
284
0.8
M01
0.6
M02
M03
0.4
TEST
1
Point of measure
Fig. 11. Assessment of FE-models M01, M02 and M03 (ABAQUS/Standard) to
experimental modal amplitudes [15] (rst vibration mode).
Fig. 10. Fundamental modal shapes for FE-models (a) M01 (800 DOFs), (b) M02 (70,000 DOFs) and (c) M03 (7000 DOFs). ABAQUS/Standard.
285
20
16
12
8
Velocity [m/s]
rupture would cause the collapse of the entire suspended curtainwall. At the same time, the possible cracking of glass panels should
be possibly prevented.
Main results obtained from dynamic simulations performed on
the M01 model are proposed in Figs. 1214 respectively, in the
form of maximum deections, maximum velocities and maximum
axial forces in the bearing cables.
As shown, the positive phase of blast alone involves maximum
effects that are markedly onerous for the studied system. The maximum deection (uM01
max 0:78 m; 1/15 the total span), velocity
0
-4
-8
No devices
Pos.
-12
(HM01
max
-16
-20
0
Time [s]
being
HM01
av g
200
175
No devices
Pos.
HR
150
125
100
75
H0
50
25
0
0
10 12 14 16 18 20 22 24 26 28 30 32
Time [s]
Fig. 14. Maximum axial force in the cables (ABAQUS/Standard, M01 FE-model, no
devices).
and HM01
max 110 kN respectively (with
v M01
max 8:2 m=s
1.2
0.9
0.6
Displacement [m]
M01
Eq. (8), a vibration period T M01
well agreeing
1;blast 0:261 s 0:45T 1
with the estimated nal vibration period T 1;blast = 0.286 s proposed
in Table 1.
Although the numerical results collected in Figs. 1214 conrm
the effectiveness of the SDOF analytical procedure for the preliminary assessment of the cable-net faade dynamic response under
the positive phase of blast only, it should be anyway highlighted
that both the SDOF procedure and the M01 model do not take into
account the possible bending stiffening contribution of the glass
panels, as well the possible benecial effects due to the negative
phase of the assigned blast load.
The effect of the total impulsive pressure function depicted in
Fig. 4 can be easily noticed in the same Figs. 1214. For the studied
faade, as expected, the negative phase of the design explosion
reduces in fact up to 30% the maximum effects of the positive
phase alone. As a result, the maximum deection is uM01
max 0:53 m,
that is 1/20 the total span of the faade. The maximum velocity
of the oscillating glazing system and the corresponding maximum
axial forces in the cables are also markedly reduced, compared
to the same faade under the positive phase only of blast,
10 12 14 16 18 20 22 24 26 28 30 32
0.3
0.0
-0.3
-0.6
No devices
Pos.
-0.9
of the faade, thus sensibly lower than the corresponding M01 predictions (Fig. 15a).
In terms of maximum tensile stresses in the glass panels, simulations performed on M02 and M03 models conrmed that maximum peaks are located near the point-supports. The typical
distribution of maximum in-plane stresses obtained from the M02
and M03 FE-models is proposed in Fig. 15b for the central panel of
the faade (e.g. the panel subjected to maximum overall deections,
due to blast pressure), in the form of blue (min)-to-red (max) contour plot and vectorial representation, respectively. Despite the
approximate modelling assumptions of M03 model, acceptable correlations were generally found with the corresponding M02 model.
In Fig. 15c, for example, tensile stresses attained close to the corner
of the same glass panel depicted in Fig. 15b are shown.
At the same time, numerical predictions obtained from both
the FE-models conrmed the assumption of an indenitely linear
elastic material for glass, being the obtained maximum tensile
stresses markedly lower than the reference magnied resistance
(Section 4.1.2).
Based on these statements and comparisons proposed in Fig. 15,
parametric numerical simulations discussed in Section 5 were performed on further M03 models only.
-1.2
0
10 12 14 16 18 20 22 24 26 28 30 32
Time [s]
Fig. 12. Maximum deection in the cable-net (ABAQUS/Standard, M01 FE-model,
no devices).
60
0.4
50
Displacement [m]
286
0.2
0.0
-0.2
M01
M02
M03
-0.4
M02
M03
40
30
20
10
-0.6
2.00
2.25
2.50
2.75
3.00
3.25
3.50
2.00
2.25
2.50
2.75
Time [s]
Time [s]
(a)
(c)
M02 FE-model
(b)
3.00
3.25
3.50
M03 FE-model
Fig. 15. (a) Maximum deections and (c) tensile stresses in the central glass panel (ABAQUS/Standard, M02 and M03 FE-models, no devices). (b) Typical distribution of
maximum in-plane stresses due to the applied blast load (ABAQUS/Standard; M02 (blue (min)-to-red (max) contour plot) and M03 (vectorial representation) FE-models).
Example for the central glass panel. (For interpretation of the references to colour in this gure legend, the reader is referred to the web version of this article.)
faade, both at the ends of the bearing cables and at the pointsupports of the glass panels, in order to analyze their effectiveness
and to identify the most advantageous solution.
Studies were dedicated to faades equipped by (i) viscoelastic
devices only, (ii) frictional devices at the cable ends, (iii) combined
viscoelastic and frictional devices.
Based on previous works [3133], the reference M03 model
was properly modied. Viscoelastic connectors, specically, were
described in the form of equivalent axial springs having elastic
stiffness kd and damping ratio cd, and directly connected to the
glass corners. Frictional devices, indeed, were applied at the cable
ends in the form of rigid-plastic springs with sliding force Fs. Parametric studies, partly discussed in Section 5, were then carried-out
in order to assess the possible structural efciency of the proposed
devices for the investigated cable-net faade.
5.1. Numerical assessment of VE and RP devices structural capabilities
5.1.1. Viscoelastic devices
In general, comparative calculations highlighted that VE devices
can have important and efcient mitigating role in the global
dynamic behaviour of the studied faade, if properly calibrated.
VE curves collected in Fig. 16, for example, are proposed for a
287
0.6
120
0.4
100
0.2
80
Displacement [m]
0.0
-0.2
No devices
VE
HRP
VE+HRP
-0.4
60
40
No devices
VE
HRP
VE+HRP
H0
20
0
-0.6
2.00
2.25
2.50
3.00
2.75
3.25
3.50
2.00
2.25
2.50
2.75
Time [s]
Time [s]
(a)
(b)
3.00
3.25
3.50
50
No devices
VE
HRP
VE+HRP
40
30
20
10
0
2.00
2.25
2.50
2.75
3.00
3.25
3.50
Time [s]
(c)
Fig. 16. (a) Maximum deections, (b) axial forces in the cables and (c) tensile stresses at the corner of the central glass panel (ABAQUS/Standard, M03 FE-model).
288
1.2x105
1.2x10 5
Total incoming energy
VE, viscous energy
HRP, friction energy
VE+HRP, viscous energy
VE+HRP, friction energy
8.0x10 4
1.0x105
1.0x10 5
6.0x10 4
4.0x10 4
2.0x10 4
8.0x104
6.0x104
4.0x104
2.0x104
0.0x10 0
0.0x100
2.00
2.25
2.50
3.00
2.75
3.25
3.5
2.00
2.25
2.50
2.75
Time [s]
Time [s]
(a)
(b)
3.00
3.25
3.5
5.0x10 4
4.0x10
No devices
VE
HRP
VE+HRP
3.0x10 4
2.0x10 4
1.0x10 4
0.0x10 0
2.00
2.25
2.50
2.75
3.00
3.25
3.50
Time [s]
(c)
Fig. 17. (a) Dissipated energy; (b) kinetic energy in the faade and (c) strain energy in the cable-net (ABAQUS/Standard, M03 FE-model).
tions, extended parametric dynamic analyses have been consequently performed on M03 FE-models.
Three typologies of devices have been introduced in the investigated faade, namely at the connection between glass panels
and cables, as well as at the cable ends.
The structural efciency under explosive loads of possible (i)
viscoelastic connectors, (ii) frictional devices at the cable ends
and (iii) combined multiple (i) and (ii) devices has been highlighted by means of advanced parametric dynamic nonlinear analyses. Although further extended studies are required, exploratory
investigations emphasized the high potentiality of the proposed
dissipative systems, hence suggesting possible successful applications in glazing faades sustained by two-way straight cable-nets
and subjected to high-strain blast pressures.
References
[1] Br L, Geiler B. MULTITERM-Fassade ein modulares Stahlfassadensystem
fr kombinierte Anwendungsflle. Stahlbau 2003;72(12):8338.
[2] SCI Steel Knowledge P396. New Beijing Poly Plaza Cable-Net Wall Structural
Stainless Steel Case Study 0.9; 2011 <www.steel-sci.org>.
[3] Sobek W, Straub W, Schmid A. Horizon Serono Konstruktion des weltweit
grten zu ffnenden Glasdaches und der darunterliegenden Forumfassade.
Stahlbau 2009;78(1):110.
[4] Kirchhoff J, Vaglio J, Patterson M. Double-Skin cable-net faade case study:
loyola information commons. In: Proceedings of BESS2010: high performance
building enclosures practical sustainability symposium at cal-poly Pomona;
2010.
[5] Sobek W, Weiss I. Eine Stahl-Glashalle als neuer zentraler Eingangsbereich der
Universitt Bremen. Stahlbau 2001;70(10):30913.
[6] Eekhout M, van de Rotten P. Cable stayed glass faade in the market hall
Rotterdam. In: Proceedings of challenging glass 4 & COST action TU0905 nal
conference, Lausanne, Switzerland; 2014.
[7] Schlaich J, Schober H, Moschner T. Prestressed cable net faades. Struct Eng Int
2005;15(1):369.
[8] Schobe H, Scheider J. Developments in structural glass and glass structures.
Struct Eng Int 2004;14(2):846.
[9] SOM-Skidmore Owings & Merrill (www.som.com). Rocker Faade Support
System, Poly Corporation Headquarters, Beijing. <http://www.ctbuh.org/
TallBuildings/FeaturedTallBuildings/FeaturedTallBuildingArchive2013>.
[10] Schlaich Bergermann und Partner, Screenwall Hotel Kempinski. <http://
www.sbp.de/en/build/show/9-Screenwall_Hotel_Kempinski>.
[11] Vyzantiadou MA, Avdelas AV. Point xed glazing systems: technological and
morphological aspects. J Constr Steel Res 2004;60(8):122740.
[12] Saitoh M, Okada A, Imamura R. Study on glass supporting system pinched at
corner structural characteristics and structural design method. In: Proceedings
of international symposium on theory design and realization of shell and
spatial structures; 2001.
[13] Feng RQ, Wu Y, Shen SZ. Working mechanism of single-layer cable net
supported glass curtain walls. Adv Struct Eng 2007;10(2):18395.
[14] Feng RQ, Ye JH, Wu Y, Shen SZ. Nonlinear response spectra of cable net faades.
Soil Dyn Earthquake Eng 2011;32:7186.
[15] Feng RQ, Zhang LL, Wu Y, Shen SZ. Dynamic performance of cable net facades. J
Constr Steel Res 2009;65(12):221727.
[16] Chan SL, Liu Y, Lee A. Nonlinear analysis of pre-tensioned glass wall faade by
stability function with initial imperfection. Front Arch Eng China
2010;4(3):37682.
[17] Weng ZJ, Yin LF, Shan J, Tang G, Huang W. Static analytical method of singlelayer plane canle net structures under bidirectional uniform load: basic
control equations and logarithmic solution. Eng Mech 2012;29(12):721.
[18] Mesda Y. Analytical study of the cable-truss systems on the glass curtain walls
with vertical uses. Engineering 2013;5:81926.
[19] Wedding WC, Lusk BT. Novel method to determine blast resistant glazing
system response to explosive loading. Measurement 2012;45:14719.
[20] Zhao S, Dharani LR, Liang X. Analysis of damage in laminated architectural
glazing subjected to blast loading. Adv Struct Eng 2008;11:12934.
[21] Dharani LR, Wei J. Dynamic response of laminated glass panels subjected to
blast loading. Structures under shock and impact VIII, Crete. Greece
2004;15:18190.
289
[22] Morison C. The resistance of laminated glass to blast pressure loading and the
coefcients for single degree of freedom analysis of laminated glass. PhD
Thesis, Craneld University; 2007.
[23] Larcher M, Solomos G. Laminated glass loaded by air blast waves
experiments and numerical simulations. Joint Research Centre Technical
Note, Pubsy JRC57559; 2010.
[24] Larcher M, Solomos G, Casadei F, Gebbeken N. Experimental and numerical
investigations of laminated glass subjected to blast loading. Int J Impact Eng
2012;39:4250.
[25] Hooper PA, Sukhram RAM, Blackman BRK, Dear JP. On the blast resistance of
laminated glass. Int J Solids Struct 2012;49:899918.
[26] Weggel DC, Zapata BJ. Laminated glass curtain walls and laminated glass plies
subjected to low-level blast loading. J Struct Eng 2008;134:46677.
[27] Zhang X, Hao H, Ma G. Parametric study of laminated glass window response
to blast loads. Eng Struct 2013;56:170717.
[28] Bedon C, Amadio C, Sinico A. Structural stability of compressed monolithic and
laminated glass elements under blast loads. In: Proceedings of glasscon global
conference innovation in glass technology, 710 July 2014, Philadelphia,
Pennsylvania Convention Center, USA, pp. 3452 (USB Flash Drive); 2014.
[29] Teich M, Warnstedt P, Gebbeken N. The inuence of negative phase loading on
cable net faade response. J Arch Eng 2011. http://dx.doi.org/10.1061/
(ASCE)AE.1943-5568.0000083 (in print).
[30] Wellershoff F, Lori G, Zobec M, Osterland K. Structural design of blast enhanced
cable net facades. In: Proceedings of COST Action TU0905 Mid-term
Conference on Structural Glass, Porec, Croatia; 2013.
[31] Amadio C, Bedon C. Elastoplastic dissipative devices for the mitigation of blast
resisting cable-supported faades. Eng Struct 2012;39:10315.
[32] Amadio C, Bedon C. Viscoelastic spider connectors for the mitigation of cablesupported faades subjected to air blast loads. Eng Struct 2012;42:190200.
[33] Amadio C, Bedon C. Multiple dissipative devices for blast-resisting cablesupported glazing faades. Model Simul Eng, vol. 2013, article ID 964910;
2013.
[34] GSA-TSO1-2003. Standard test method for glazing and window systems
subject to dynamic overpressure loadings. U.S. General Service
Administration; 2003.
[35] Goel MD, Matsagar VA, Gupta AK, Marburg S. An abridged review of blast wave
parameters. Defence Sci J 2012;62(5):3006.
[36] Bulson P. Explosive loading of engineering structures. A history of research and
a review of recent developments. London: E&F Spon; 1997.
[37] TM 5-1300. The design of structures to resist the effects of accidental
explosions. U.S. Dept. of the Army, the Navy and the Air Force, Technical
Manual, Washington DC; 1990.
[38] Larcher M, Hermann N, Stempniewsky L. Explosionssimulation leichter
Hallenhllkonstruktionen. Bauingenieur 2006;81(6):2717.
[39] Dassault Systmes Simulia Corp. ABAQUS ver. 6.9 Users Manual; Providence,
RI, USA; 2009.
[40] Schleyer FK. Tensile structures analysis of cables, cable nets and cable
structures. MIT Press; 1969.
[41] Wang Z, Gong G, Zhang Y, Bai C. Simulation of the response of glass window
under blast load. Transactions of Tianjin University, 14: 50410. Tianjin
University and Springer-Verlag; 2008.
[42] Bedon C, Louter C. Exploratory numerical analysis of SG-laminated reinforced
glass beam experiments. Eng Struct 2014;75(9):45768.
[43] EN 572-2. Glass in buildings basic soda lime silicate glass products. CEN;
2004.
[44] MoD/DE Functional Standard Design and Maintenance Guide 02. Glazing
standards for MoD buildings subject to terrorist threat, Ministry of Defence;
1996.
[45] Meyers GE. BLASTOP version 1 4, US Department of Energy; 1994.
[46] Mayor RP, Flanders R. Technical manual simplied computer model of air blast
effects on building walls. US Department of State, Ofce of Diplomatic
Security, Washington DC; 1990.
[47] Morison C. The resistance of laminated glass to blast pressure loading and the
coefcients for single degree of freedom analysis of laminated glass. PhD
Dissertation, Craneld University; 2007.
[48] Peroni M, Solomos G, Pizzinato V, Larcher M. Experimental investigation of
high strain-rate behaviour of glass. Appl Mech Mater 2011;82:638.
[49] Bachmann H, Ammann WJ, Florian D. Vibration problems in structures:
practical guidelines. Birkhuser; 1995.
[50] Buchholdt HA. An introduction to cable roof structures. 2nd ed. Thomas
Telford; 1999.
[51] Timoshenko S, Woinowsky-Krieger S. Theory of plates and shells. McGraw-Hill
Book Company; 1959.