EJSE Special Issue: Loading on Structures (2007)

Blast Loading and Blast Effects on Structures – An Overview

T. Ngo, P. Mendis, A. Gupta & J. Ramsay
The University of Melbourne, Australia

ABSTRACT: The use of vehicle bombs to attack city centers has been a feature of campaigns by terrorist or-
ganizations around the world. A bomb explosion within or immediately nearby a building can cause catastro-
phic damage on the building's external and internal structural frames, collapsing of walls, blowing out of large
expanses of windows, and shutting down of critical life-safety systems. Loss of life and injuries to occupants
can result from many causes, including direct blast-effects, structural collapse, debris impact, fire, and smoke.
The indirect effects can combine to inhibit or prevent timely evacuation, thereby contributing to additional
casualties. In addition, major catastrophes resulting from gas-chemical explosions result in large dynamic
loads, greater than the original design loads, of many structures. Due to the threat from such extreme loading
conditions, efforts have been made during the past three decades to develop methods of structural analysis
and design to resist blast loads. The analysis and design of structures subjected to blast loads require a de-
tailed understanding of blast phenomena and the dynamic response of various structural elements. This paper
presents a comprehensive overview of the effects of explosion on structures. An explanation of the nature of
explosions and the mechanism of blast waves in free air is given. This paper also introduces different methods
to estimate blast loads and structural response. check this
understanding
against the book
1 EXPLOSIONS AND BLAST PHENOMENON The detonation of a condensed high explosive
generates hot gases under pressure up to 300 kilo bar
An explosion is defined as a large-scale, rapid and and a temperature of about 3000-4000C°. The hot
sudden release of energy. Explosions can be catego- gas expands forcing out the volume it occupies. As a
rized on the basis of their nature as physical, nuclear consequence, a layer of compressed air (blast wave)
or chemical events. In physical explosions, energy forms in front of this gas volume containing most of
may be released from the catastrophic failure of a the energy released by the explosion. Blast wave in-
cylinder of compressed gas, volcanic eruptions or stantaneously increases to a value of pressure above
even mixing of two liquids at different temperatures. the ambient atmospheric pressure. This is referred to
In a nuclear explosion, energy is released from the as the side-on overpressure that decays as the shock
formation of different atomic nuclei by the redistri- wave expands outward from the explosion source.
bution of the protons and neutrons within the inter- After a short time, the pressure behind the front may
acting nuclei, whereas the rapid oxidation of fuel drop below the ambient pressure (Figure 1). During
elements (carbon and hydrogen atoms) is the main such a negative phase, a partial vacuum is created
source of energy in the case of chemical explosions. and air is sucked in. This is also accompanied by
Explosive materials can be classified according to high suction winds that carry the debris for long dis-
their physical state as solids, liquids or gases. Solid tances away from the explosion source.
explosives are mainly high explosives for which
blast effects are best known. They can also be classi- can this region
fied on the basis of their sensitivity to ignition as exp negative?
secondary or primary explosive. The latter is one Yes, this graph
that can be easily detonated by simple ignition from not nicely drawn
Pressure

a spark, flame or impact. Materials such as mercury

fulminate and lead azide are primary explosives.
Secondary explosives when detonated create blast
Shock velocity
(shock) waves which can result in widespread dam-
age to the surroundings. Examples include trinitro-
toluene (TNT) and ANFO. Distance from explosion
Figure 1: Blast wave propagation
secondary can be controlled, and
tho it says carry debris away experience high pressure, then
more damaging
from the 76
source...the source slowly subsides... at some point
should exp neg earlier... can subside below ambient
 EJSE Special Issue: Loading on Structures (2007)

P(t) source. Reflection factors depend on the intensity of

the shock wave, and for large explosives at normal
incidence these reflection factors may enhance the
Pso incident pressures by as much as an order of magni-
tude.
Positive Specific
Impulse Throughout the pressure-time profile, two main
Negative Specific phases can be observed; portion above ambient is
Impulse called positive phase of duration td, while that be-
t
Po
tA tA+td low ambient is called negative phase of duration,
td-. The negative phase is of a longer duration and
Pso-
a lower intensity than the positive duration. As the
stand-off distance increases, the duration of the
Positive Negative positive-phase blast wave increases resulting in a
duration td duration td- lower-amplitude, longer-duration shock pulse .
Figure 2: Blast wave pressure – Time history Charges situated extremely close to a target structure
impose a highly impulsive, high intensity pressure
load over a localized region of the structure; charges
2 EXPLOSIVE AIR BLAST LOADING situated further away produce a lower-intensity,
longer-duration uniform pressure distribution over
The threat for a conventional bomb is defined by the entire structure. Eventually, the entire structure
two equally important elements, the bomb size, or is engulfed in the shock wave, with reflection and
charge weight W, and the standoff distance R be- diffraction effects creating focusing and shadow
tween the blast source and the target (Figure 3). For zones in a complex pattern around the structure.
example, the blast occurred at the basement of During the negative phase, the weakened structure
World Trade Centre in 1993 has the charge weight may be subjected to impact by debris that may cause
of 816.5 kg TNT. The Oklahoma bomb in 1995 has additional damage.
a charge weight of 1814 kg at a stand off of 4.5m
(Longinow, 1996). As terrorist attacks may range bomb shape,
from the small letter bomb to the gigantic truck orientation also Over-pressure

bomb as experienced in Oklahoma City, the me-

chanics of a conventional explosion and their effects
on a target must be addressed.

The observed characteristics of air blast waves Over-pressure

(side-on)
are found to be affected by the physical properties Reflected
of the explosion source. Figure 2 shows a typical Pressure

blast pressure profile. At the arrival time tA,

following the explosion, pressure at that position
suddenly increases to a peak value of overpres- Blast wave

sure, Pso, over the ambient pressure, Po. The pres- Stand-off distance
sure then decays to ambient level at time td, then
decays further to an under pressure Pso- (creating a
partial vacumn) before eventually returning to am- Figure 3: Blast loads on a building
bient conditions at time td + td-. The quantity Pso is
usually referred to as the peak side-on overpres- If the exterior building walls are capable of
sure, incident peak overpressure or merely peak resisting the blast load, the shock front penetrates
overpressure (TM 5-1300, 1990). through window and door openings, subjecting the
floors, ceilings, walls, contents, and people to
The incident peak over pressures Pso are ampli- sudden pressures and fragments from shattered
fied by a reflection factor as the shock wave encoun- windows, doors, etc. Building components not
ters an object or structure in its path. Except for spe- capable of resisting the blast wave will fracture and
cific focusing of high intensity shock waves at near be further fragmented and moved by the dynamic
45° incidence, these reflection factors are typically pressure that immediately follows the shock front.
greatest for normal incidence (a surface adjacent and Building contents and people will be displaced and
perpendicular to the source) and diminish with the tumbled in the direction of blast wave propagation.
angle of obliquity or angular position relative to the In this manner the blast will propagate through the
building.
dynamic pressure is the wind? the
77 forward pressure that decays
 EJSE Special Issue: Loading on Structures (2007)

2.1 Blast Wave Scaling Laws q s = 5 p so2 / 2( p so + 7 po ) (4)

If the blast wave encounters an obstacle perpen-
All blast parameters are primarily dependent on dicular to the direction of propagation, reflection
the amount of energy released by a detonation in increases the overpressure to a maximum reflected
the form of a blast wave and the distance from the pressure Pr as:
explosion. A universal normalized description of
the blast effects can be given by scaling distance ⎧ 7 P + 4 Pso ⎫
Pr = 2 Pso ⎨ o ⎬ (5)
relative to (E/Po)1/3 and scaling pressure relative to ⎩ 7 Po + Pso ⎭
Po, where E is the energy release (kJ) and Po the A full discussion and extensive charts for pre-
ambient pressure (typically 100 kN/m2). For con- dicting blast pressures and blast durations are given
venience, however, it is general practice to express by Mays and Smith (1995) and TM5-1300 (1990).
the basic explosive input or charge weight W as an Some representative numerical values of peak re-
equivalent mass of TNT. Results are then given as flected overpressure are given in Table 1.
a function of the dimensional distance parameter
(scaled distance) Z = R/W1/3, where R is the actual Table 1. Peak reflected overpressures Pr (in MPa) with differ-
effective distance from the explosion. W is gener- ent W-R combinations
ally expressed in kilograms. Scaling laws provide
parametric correlations between a particular explo-
sion and a standard charge of the same substance. W 100 kg 500 kg 1000 kg 2000 kg
R TNT TNT TNT TNT
1m 165.8 354.5 464.5 602.9
2.2 Prediction of Blast Pressure 2.5m 34.2 89.4 130.8 188.4
5m 6.65 24.8 39.5 60.19
Blast wave parameters for conventional high 10m 0.85 4.25 8.15 14.7
explosive materials have been the focus of a num- 15m 0.27 1.25 2.53 5.01
ber of studies during the 1950’s and 1960’s. Esti- 20m 0.14 0.54 1.06 2.13
mations of peak overpressure due to spherical blast 25m 0.09 0.29 0.55 1.08
based on scaled distance Z = R/W1/3 were intro- 30m 0.06 0.19 0.33 0.63
duced by Brode (1955) as:
6.7 For design purposes, reflected overpressure can
Pso = 3 + 1 bar ( Pso > 10 bar) be idealized by an equivalent triangular pulse of
Z
maximum peak pressure Pr and time duration td,
0.975 1.455 5.85 (1) which yields the reflected impulse ir
Pso = + 2
+ 3 − 0.019 bar
Z Z Z 1
(0.1 bar < Pso < 10 bar) ir = Pr t d (6)
2
Newmark and Hansen (1961) introduced a rela- Duration td is related directly to the time taken
tionship to calculate the maximum blast overpres- for the overpressure to be dissipated. Overpressure
sure, Pso, in bars, for a high explosive charge deto- arising from wave reflection dissipates as the per-
nates at the ground surface as: turbation propagates to the edges of the obstacle at
so means side on 1
a velocity related to the speed of sound (Us) in the
W ⎛ W ⎞2 (2) compressed and heated air behind the wave front.
Pso = 6784 3 + 93⎜ 3 ⎟
R ⎝R ⎠ Denoting the maximum distance from an edge as S
Another expression of the peak overpressure in (for example, the lesser of the height or half the
kPa is introduced by Mills (1987), in which W is width of a conventional building), the additional
expressed as the equivalent charge weight in kilo- pressure due to reflection is considered to reduce
grams of TNT, and Z is the scaled distance: from Pr – Pso to zero in time 3S/Us. Conserva-
1772 114 108 tively, Us can be taken as the normal speed of
Pso = 3 − 2 + (3) sound, which is about 340 m/s, and the additional
Z Z Z
As the blast wave propagates through the at- impulse to the structure evaluated on the assump-
mosphere, the air behind the shock front is moving tion of a linear decay.
outward at lower velocity. The velocity of the air
particles, and hence the wind pressure, depends on After the blast wave has passed the rear corner
the peak overpressure of the blast wave. This later of a prismatic obstacle, the pressure similarly
velocity of the air is associated with the dynamic propagates on to the rear face; linear build-up over
pressure, q(t). The maximum value, qs, say, is duration 5S/Us has been suggested. For skeletal
given by structures the effective duration of the net over-
pressure load is thus small, and the drag loading

78
 EJSE Special Issue: Loading on Structures (2007)

based on the dynamic pressure is then likely to be sions on passage of the blast through doorways, but
dominant. Conventional wind-loading pressure co- can also be created by obstacles closer to the re-
efficients may be used, with the conservative as- lease of the gas. They may be presumed to occur on
sumption of instantaneous build-up when the wave release of gas by failure of industrial pressure ves-
passes the plane of the relevant face of the build- sels or pipelines.
ing, the loads on the front and rear faces being
numerically cumulative for the overall load effect
on the structure. Various formulations have been 4 STRUCTURAL RESPONSE TO BLAST
put forward for the rate of decay of the dynamic LOADING
pressure loading; a parabolic decay (i.e. corre-
sponding to a linear decay of equivalent wind ve- Complexity in analyzing the dynamic response of
locity) over a time equal to the total duration of blast-loaded structures involves the effect of high
positive overpressure is a practical approximation. strain rates, the non-linear inelastic material behav-
ior, the uncertainties of blast load calculations and
the time-dependent deformations. Therefore, to sim-
3 GAS EXPLOSION LOADING AND EFFECT plify the analysis, a number of assumptions related
OF INTERNAL EXPLOSIONS to the response of structures and the loads has been
In the circumstances of progressive build-up of proposed and widely accepted. To establish the prin-
fuel in a low-turbulence environment, typical of ciples of this analysis, the structure is idealized as a
domestic gas explosions, flame propagation on ig- single degree of freedom (SDOF) system and the
nition is slow and the resulting pressure pulse is link between the positive duration of the blast load
correspondingly extended. The specific energy of and the natural period of vibration of the structure is
combustion of a hydrocarbon fuel is very high established. This leads to blast load idealization and
(46000 kJ/kg for propane, compared to 4520 kJ/kg simplifies the classification of the blast loading re-
for TNT) but widely differing effects are possible gimes.
according to the conditions at ignition.

Internal explosions likely produce complex pres- 4.1 Elastic SDOF Systems
sure loading profiles as a result of the resulting two
loading phases. The first results from the blast over- The simplest discretization of transient problems
pressure reflection and, due to the confinement pro- is by means of the SDOF approach. The actual struc-
vided by the structure, re-reflection will occur. De- ture can be replaced by an equivalent system of one
pending on the degree of confinement of the concentrated mass and one weightless spring repre-
structure, the confined effects of the resulting pres- senting the resistance of the structure against defor-
sures may cause different degrees of damage to the mation. Such an idealized system is illustrated in
structure. On the basis of the confinement effect, tar- Figure 4. The structural mass, M, is under the effect
get structures can be described as either vented or of an external force, F(t), and the structural resis-
un-vented. The latter must be stronger to resist a tance, R, is expressed in terms of the vertical dis-
specific explosion yield than a vented structure placement, y, and the spring constant, K.
where some of the explosion energy would be dissi-
pated by breaking of window glass or fragile parti- The blast load can also be idealized as a triangu-
tions. lar pulse having a peak force Fm and positive phase
duration td (see Figure 4). The forcing function is
Venting following the failure of windows (at given as
typically 7 kN/m2) generally greatly reduces the ⎛ t ⎞
peak values of internal pressures. Study of this i skip the F (t ) = Fm ⎜⎜1 − ⎟⎟ (7)
problem at the Building Research Establishment math ⎝ td ⎠
(Ellis and Crowhurst, 1991) showed that an explo- The blast impulse is approximated as the area un-
sion fuelled by a 200 ml aerosol canister in a typi- der the force-time curve, and is given by
cal domestic room produced a peak pressure of 9 1 (8)
kN/m2 with a pulse duration over 0.1s. This is long I = Fm t d
2
by comparison with the natural frequency of wall The equation of motion of the un-damped elastic
panels in conventional building construction and a SDOF system for a time ranging from 0 to the posi-
quasi-static design pressure is commonly advo- tive phase duration, td, is given by Biggs (1964) as
cated. Much higher pressures with a shorter time-
⎛ t ⎞ (9)
scale are generated in turbulent conditions. Suitable My + Ky = Fm ⎜⎜1 − ⎟⎟
conditions arise in buildings in multi-room explo- ⎝ t d ⎠

79
 EJSE Special Issue: Loading on Structures (2007)

The general solution can be expressed as: merical solution requiring a nonlinear dynamic fi-
Displacement nite-element software. However, the degree of
(10)
Fm Fm ⎛ sin ωt ⎞ uncertainty in both the determination of the loading
y (t ) = (1 − cos ωt ) + ⎜ −t⎟ and the interpretation of acceptability of the result-
K Kt d ⎝ ω ⎠ ing deformation is such that solution of a postu-
Velocity lated equivalent ideal elasto-plastic SDOF system
dy Fm ⎡ 1 ⎤ (Biggs, 1964) is commonly used. Interpretation is
y (t ) = = ⎢ω sin ωt + (cos ωt − 1)⎥ based on the required ductility factor μ = ym/ye
dt K ⎣ td ⎦ (Figure 5).
in which ω is the natural circular frequency of vi-
bration of the structure and T is the natural period of Resistance
vibration of the structure which is given by equation
11. Ru
2π K (11)
ω= =
T M

Force
F(t) ye ym Deflection
F(t)
Fm
Figure 5: Simplified resistance function of an elasto-plastic
M
SDOF system
Displacement
Stiffness, K
y(t)

Time
For example, a uniform simply supported beam
has first mode shape φ(x) = sin πx/L and the
td
(a) (b)

Figure 4: (a) SDOF system and (b) blast loading equivalent mass M = (1/2)mL, where L is the span
of the beam and m is mass per unit length. The
The maximum response is defined by the maximum equivalent force corresponding to a uniformly dis-
dynamic deflection ym which occurs at time tm. The tributed load of intensity p is F = (2/π)pL. The re-
sponse of the ideal bilinear elasto-plastic system
maximum dynamic deflection ym can be evaluated
can be evaluated in closed form for the triangular
by setting dy/dt in Equation 10 equal to zero, i.e. load pulse comprising rapid rise and linear decay,
when the structural velocity is zero. The dynamic with maximum value Fm and duration td. The result
load factor, DLF, is defined as the ratio of the for the maximum displacement is generally pre-
maximum dynamic deflection ym to the static deflec- sented in chart form (TM 5-1300), as a family of
tion yst which would have resulted from the static curves for selected values of Ru/Fm showing the re-
application of the peak load Fm, which is shown as quired ductility μ as a function of td/T, in which Ru
follows: is the structural resistance of the beam and T is the
natural period (Figure 6).
(12)
y max y max ⎛ td ⎞
DLF = = = ψ (ωt d ) = Ψ ⎜ ⎟ 100
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
y st Fm ⎝T ⎠
K 50 0.9

The structural response to blast loading is signifi-

cantly influenced by the ratio td/T or ωt d (td/T
= ωt d / 2π ). Three loading regimes are categorized 10 1.0

as follows:
Xm / Xe

5
ym/ye

- ωt d < 0.4 : impulsive loading regime. 1.2

1.5
- ωt d < 0.4 : quasi-static loading regime. 1 2.0

- 0.4 < ωt d < 40 : dynamic loading regime. 0.5

Numbers
Numbersnext
next to
to curves are Ru/Fm
curves are Ru/Fm
0.1
4.2 Elasto-Plastic SDOF Systems 0.1 0.5 1 5 10 20
/T/ T
tdtd
Structural elements are expected to undergo Figure 6: Maximum response of elasto-plastic SDF system to
large inelastic deformation under blast load or high a triangular load.
velocity impact. Exact analysis of dynamic re-
sponse is then only possible by step-by-step nu-
80
 EJSE Special Issue: Loading on Structures (2007)

5 MATERIAL BEHAVIORS AT HIGH STRAIN- CEB-FIP (1990) model (Figure 9) for strain-rate en-
RATE hancement of concrete as follows:
1.026α
⎛ ε ⎞
Blast loads typically produce very high strain DIF = ⎜⎜ ⎟⎟ for ε ≤ 30s −1 (13)
rates in the range of 102 - 104 s-1. This high straining ⎝ ε s ⎠
(loading) rate would alter the dynamic mechanical 1/ 3
⎛ ε ⎞
properties of target structures and, accordingly, the DIF = γ ⎜⎜ ⎟⎟ for ε > 30s −1 (14)
expected damage mechanisms for various structural ε
⎝ s⎠
elements. For reinforced concrete structures sub- where:
jected to blast effects the strength of concrete and ε = strain rate
steel reinforcing bars can increase significantly due εs = 30×10-6 s-1 (quasi-static strain rate)
to strain rate effects. Figure 7 shows the approxi- log γ = 6.156 α - 2
mate ranges of the expected strain rates for different α = 1/(5 + 9 f’c/fco)
loading conditions. It can be seen that ordinary static fco = 10 MPa = 1450 psi
strain rate is located in the range : 10-6-10-5 s-1, while 8
blast pressures normally yield loads associated with

Dynamic Factor
strain rates in the range : 102-104 s-1. 6

Quasi-static Earthquake Impact Blast

4
-6 -5 -4 -3 -2 -1 0 1 2 2 3 4
10 10 10 10 10 10 10 10 10 10 10 10 2
Strain rate (s-1)
0
Figure 7: Strain rates associated with different types of loading 1.E-04 1.E-02 1.E+00 1.E+02 1.E+04
Strain rate (s-1)
Figure 9: Dynamic Increase Factor for peak stress of concrete
5.1 Dynamic Properties of Concrete under High-
Strain Rates
5.2 Dynamic Properties of Reinforcing Steel under
The mechanical properties of concrete under dy-
High-Strain Rates
namic loading conditions can be quite different from
that under static loading. While the dynamic stiff-
Due to the isotropic properties of metallic materi-
ness does not vary a great deal from the static stiff-
als, their elastic and inelastic response to dynamic
ness, the stresses that are sustained for a certain pe-
loading can easily be monitored and assessed. Norris
riod of time under dynamic conditions may gain
et al. (1959) tested steel with two different static
values that are remarkably higher than the static
yield strength of 330 and 278 MPa under tension at
compressive strength (Figure 8). Strength magnifica-
strain rates ranging from 10-5 to 0.1 s-1. Strength in-
tion factors as high as 4 in compression and up to 6
crease of 9 - 21% and 10 - 23 % were observed for
in tension for strain rates in the range : 102–103 /sec
the two steel types, respectively. Dowling and Hard-
have been reported (Grote et al., 2001).
ing (1967) conducted tensile experiments using the
tensile version of Split Hopkinton's Pressure Bar
250
ε = 264 (SHPB) on mild steel using strain rates varying be-
tween 10-3 s-1 and 2000 s-1. It was concluded from
200 ε = 233 this test series that materials of body-centered cubic
(BCC) structure (such as mild steel) showed the
ε = 97 greatest strain rate sensitivity. It has been found that
Stress (MPa)

150
Static the lower yield strength of mild steel can almost be
ε = 49
doubled; the ultimate tensile strength can be in-
100
creased by about 50%; and the upper yield strength
can be considerably higher. In contrast, the ultimate
50 tensile strain decreases with increasing strain rate.
Malvar (1998) also studied strength enhancement
0 of steel reinforcing bars under the effect of high
0 0.002 0.004 0.006 0.008 0.01 strain rates. This was described in terms of the dy-
Strain
namic increase factor (DIF), which can be evaluated
Figure 8: Stress-strain curves of concrete at different strain- for different steel grades and for yield stresses, fy,
rates (Ngo et al., 2004a) ranging from 290 to 710 MPa as represented by
For the increase in peak compressive stress (f’c), a equation 15.
dynamic increase factor (DIF) is introduced in the
81
 EJSE Special Issue: Loading on Structures (2007)

⎛ ε ⎞
α ated shear force is many times higher than the shear
DIF = ⎜ −4 ⎟ (15)
force associated with flexural failure modes. The
⎝ 10 ⎠ high shear stresses may lead to direct global shear
where for calculating yield stress α = α fy , failure and it may occur very early (within a few
α fy = 0.074 − 0.04( f y / 414) (16) milliseconds of shock wave arrival to the frontal sur-
face of the structure) which can be prior to any oc-
for ultimate stress calculation α = α fu currence of significant bending deformations.
α fu = 0.019 − 0.009( f y / 414) (17)
6.2 Localized Structural Behavior

The close-in effect of explosion may cause local-

6 FAILURE MODES OF BLAST-LOADED ized shear or flexural failure in the closest struc-
STRUCTURES tural elements. This depends mainly on the distance
between the source of the explosion and the target,
Blast loading effects on structural members may and the relative strength/ductility of the structural
produce both local and global responses associated elements. The localized shear failure takes place in
with different failure modes. The type of structural the form of localized punching and spalling, which
response depends mainly on the loading rate, the ori- produces low and high-speed fragments. The
entation of the target with respect to the direction of punching effect is frequently referred to as bleach-
the blast wave propagation and boundary conditions. ing, which is well known in high velocity impact
The general failure modes associated with blast applications and the case of explosions close to the
loading can be flexure, direct shear or punching surface of structural members. Bleaching failures
shear. Local responses are characterized by localized are typically accompanied by spalling and scabbing
bleaching and spalling, and generally result from the of concrete covers as well as fragments and debris
close-in effects of explosions, while global re- (Figure 10).
sponses are typically manifested as flexural failure.
???
6.1 Global Structural Behavior

The global response of structural elements is gen-

erally a consequence of transverse (out-of-plane)
loads with long exposure time (quasi-static loading),
and is usually associated with global membrane
(bending) and shear responses. Therefore, the global
response of above-ground reinforced concrete struc-
tures subjected to blast loading is referred to as
membrane/bending failure. membrane? plate? Figure 10: Breaching failure due to a close-in explosion of
6000kg TNT equivalent
The second global failure mode to be considered
is shear failure. It has been found that under the ef- 6.3 Pressure-Impulse (P-I) Diagrams
fect of both static and dynamic loading, four types of
shear failure can be identified: diagonal tension, di- The pressure-impulse (P-I) diagram is an easy
agonal compression, punching shear, and direct (dy- way to mathematically relate a specific damage level
namic) shear (Woodson, 1993). The first two types to a combination of blast pressures and impulses
are common in reinforced concrete elements under imposes on a particular structural element. An ex-
static loading while punching shear is associated ample of a P-I diagram is shown in Figure 11 to
with local shear failure, the familiar example of this show levels of damage of a structural member.
is column punching through a flat slab. These shear Region (I) corresponds to severe structural damage
response mechanisms have relatively minor struc- and region (II) refers to no or minor damage. There
tural effect in case of blast loading and can be ne- are other P-I diagrams that concern with human re-
glected. The fourth type of shear failure is direct sponse to blast in which case there are three catego-
(dynamic) shear. This failure mode is primarily as- ries of blast-induced injury, namely : primary, sec-
sociated with transient short duration dynamic loads ondary, and tertiary injury (Baker et al., 1983).
that result from blast effects, and it depends mainly
on the intensity of the pressure waves. The associ-
i have a feeling this direct shear this is the kind of shear they
82
is not our normal shear (proftan n youxin)talked about
 EJSE Special Issue: Loading on Structures (2007)

with depth (TM 5-1300, 1990). The direct shock re-

Impulse is (kPa.sec)

101 sults from the direct transmission of explosive en-

ergy through the ground. For a point of interest on
the ground surface, the net experienced ground
100 shock results from a combination of both the air-
induced and direct shocks.
(I) – Severe Damage

10 -1 7.1 Loads from Air-induced Ground Shock

(II) - No damage / minor damage To overcome complications of predicting actual
ground motion, one-dimensional wave propagation
10-2
theory has been employed to quantify the maximum
100 101 102 103
Pressure Ps (kPa)
displacement, velocity and acceleration in terms of
the already known blast wave parameters (TM 5-
Figure 11: Typical pressure-impulse (P-I) diagram
1300). The maximum vertical velocity at the ground
surface, Vv , is expressed in terms of the peak inci-
dent overpressure, Pso , as:
7 BLAST WAVE-STRUCTURE INTERACTION P (18)
Vv = so
ρC p
The structural behavior of an object or structure ex- where ρ and C p are, respectively, the mass density
posed to such blast wave may be analyzed by deal- and the wave seismic velocity in the soil.
ing with two main issues. Firstly, blast-loading ef- By integrating the vertical velocity in Equation
fects, i.e., forces that are resulted directly from the 18 with time, the maximum vertical displacement at
action of the blast pressure; secondly, the structural the ground surface, Dv , can be obtained as:
response, or the expected damage criteria associated
is (19)
with such loading effects. It is important to consider Dv =
the interaction of the blast waves with the target 1000 ρC p
structures. This might be quite complicated in the
case of complex structural configurations. However, Accounting for the depth of soil layers, an em-
it is possible to consider some equivalent simplified pirical formula is given by (TM 5-1300) to estimate
geometry. Accordingly, in analyzing the dynamic the vertical displacement in meters so that
response to blast loading, two types of target struc- 1 2
Dv = 0.09W (H / 50) (Pso ) 3 (20)
0.6
tures can be considered: diffraction-type and drag- 6

type structures. As these names imply, the former where W is the explosion yield in 109 kg, and H is
would be affected mainly by diffraction (engulfing) the depth of the soil layer in meters.
loading and the latter by drag loading. It should be
emphasized that actual buildings will respond to
both types of loading and the distinction is made 7.2 Loads from Direct Ground Shock
primarily to simplify the analysis. The structural re-
sponse will depend upon the size, shape and weight As a result of the direct transmission of the ex-
of the target, how firmly it is attached to the ground, plosion energy, the ground surface experiences ver-
and also on the existence of openings in each face of tical and horizontal motions. Some empirical equa-
the structure. tions were derived (TM 5-1300) to predict the
direct-induced ground motions in three different
Above ground or shallow-buried structures can be ground media; dry soil, saturated soil and rock me-
subjected to ground shock resulting from the detona- dia. The peak vertical displacement in m/s at the
tion of explosive charges that are on/or close to ground surface for rock, DVrock and dry soil, DVsoil are
ground surface. The energy imparted to the ground given as
by the explosion is the main source of ground shock.
A part of this energy is directly transmitted through 1 1
the ground as directly-induced ground shock, while 0.25 R W3 3
(21)
part is transmitted through the air as air-induced DVrock = 1
ground shock. Air-induced ground shock results Z 3

when the air-blast wave compresses the ground sur- 1 1

face and sends a stress pulse into the ground under- 0.17 R 3W 3 (22)
DVsoil =
layers. Generally, motion due to air-induced ground Z 2.3
is maximum at the ground surface and attenuates
83
 EJSE Special Issue: Loading on Structures (2007)

The maximum vertical acceleration, Av, in m/s2 structures subjected to the effects of conventional
for all ground media is given by weapons. It is intended for use by engineers in-
1000 volved in designing hardened facilities.
Av = 1 (23) The Design and Analysis of Hardened Struc-
W 8Z 2 tures to Conventional Weapons Effects (DAHS
CWE, 1998). This new Joint Services manual, writ-
ten by a team of more than 200 experts in conven-
8 TECHNICAL DESIGN MANUALS FOR tional weapons and protective structures engineer-
BLAST-RESISTANT DESIGN ing, supersedes U.S. Department of the Army TM 5-
855-1, Fundamentals of Protective Design for Con-
This section summarizes applicable military design ventional Weapons (1986), and Air Force Engineer-
manuals and computational approaches to predicting ing and Services Centre ESL-TR-87-57, Protective
blast loads and the responses of structural systems. Construction Design Manual (1989).
Although the majority of these design guidelines
were focused on military applications these knowl- Structural Design for Physical Security—State
of the Practice Report (ASCE, 1995). This report
edge are relevant for civil design practice.
is intended to be a comprehensive guide for civilian
designers and planners who wish to incorporate
Structures to Resist the Effects of Accidental
physical security considerations into their designs or
Explosions, TM 5-1300 (U.S. Departments of the
building retrofit efforts.
Army, Navy, and Air Force, 1990). This manual
appears to be the most widely used publication by
both military and civilian organizations for design-
9 COMPUTER PROGRAMS FOR BLAST AND
ing structures to prevent the propagation of explo-
SHOCK EFFECTS
sion and to provide protection for personnel and
valuable equipment. It includes step-by-step analysis
Computational methods in the area of blast-
and design procedures, including information on
effects mitigation are generally divided into those
such items as (1) blast, fragment, and shock-loading;
used for prediction of blast loads on the structure
(2) principles of dynamic analysis; (3) reinforced
and those for calculation of structural response to the
and structural steel design; and (4) a number of spe-
loads. Computational programs for blast prediction
cial design considerations, including information on
and structural response use both first-principle and
tolerances and fragility, as well as shock isolation.
semi-empirical methods. Programs using the first-
Guidance is provided for selection and design of se-
principle method can be categorized into uncouple
curity windows, doors, utility openings, and other
and couple analyses. The uncouple analysis calcu-
components that must resist blast and forced-entry
lates blast loads as if the structure (and its compo-
effects.
nents) were rigid and then applying these loads to a
responding model of the structure. The shortcoming
A Manual for the Prediction of Blast and
of this procedure is that when the blast field is ob-
Fragment Loadings on Structures, DOE/TIC-
tained with a rigid model of the structure, the loads
11268 (U.S. Department of Energy, 1992). This
on the structure are often over-predicted, particularly
manual provides guidance to the designers of facili-
if significant motion or failure of the structure oc-
ties subject to accidental explosions and aids in the
curs during the loading period.
assessment of the explosion-resistant capabilities of
existing buildings.
For a coupled analysis, the blast simulation mod-
ule is linked with the structural response module. In
Protective Construction Design Manual, ESL-
this type of analysis the CFD (computational fluid
TR-87-57 (Air Force Engineering and Services
mechanics) model for blast-load prediction is solved
Center, 1989). This manual provides procedures for
simultaneously with the CSM (computational solid
the analysis and design of protective structures ex-
mechanics) model for structural response. By ac-
posed to the effects of conventional (non-nuclear)
counting for the motion of the structure while the
weapons and is intended for use by engineers with
blast calculation proceeds, the pressures that arise
basic knowledge of weapons effects, structural dy-
due to motion and failure of the structure can be pre-
namics, and hardened protective structures.
dicted more accurately. Examples of this type of
computer codes are AUTODYN, DYNA3D, LS-
Fundamentals of Protective Design for Con-
DYNA and ABAQUS. Table 2 summarizes a listing
ventional Weapons, TM 5-855-1 (U.S. Depart-
of computer programs that are currently being used
ment of the Army, 1986). This manual provides
to model blast-effects on structures.
procedures for the design and analysis of protective
84
 EJSE Special Issue: Loading on Structures (2007)

HSC column) and spacing of ligatures (400mm for

Table 2. Examples of computer programs used to simulate ordinary detailing-OMRF and 100mm for special
blast effects and structural response seismic detailing-SMRF). It has been found that
with increasing concrete compressive strength, the
Name Purpose and type of Author/Vendor column size can be effectively reduced. In this case
analysis the column size was reduced from 500 x 900 mm for
BLASTX Blast prediction, CFD SAIC the NSC column down to 350 x 750 for the HSC
code column (Table 2) while the axial load capacities of
CTH Blast prediction, CFD Sandia National the two columns are still the same.
code Laboratories
The blast load was calculated based on data from
FEFLO Blast prediction, CFD SAIC
code
the Oklahoma bombing report (ASCE 1996) with a
FOIL Blast prediction, CFD Applied Research As-
stand off distance of 11.2m. The simplified triangle
code sociates, Waterways shape of the blast load profile was used (see Fig.
Experiment Station 13). The duration of the positive phase of the blast is
SHARC Blast prediction, CFD Applied Research As- 1.3 milliseconds.
code sociates, Inc. The 3D model of the column (see Fig. 14) was
DYNA3D Structural response + CFD Lawrence Livermore analysed using the nonlinear explicit code LS-Dyna
(Couple analysis) National Laboratory 3D (2002) which takes into account both material
(LLNL) nonlinearity and geometric nonlinearity. The strain-
ALE3D Coupled analysis Lawrence Livermore rate-dependent constitutive model proposed in the
National Laboratory previous section was adopted. The effects of the
(LLNL)
blast loading were modelled in the dynamic analysis
LS- Structural response + CFD Livermore Software
DYNA (Couple analysis) Technology Corpora-
to obtain the deflection time history of the column.
tion (LSTC)
Table 3. Concrete grades and member sizes
___________________________________________
Air3D Blast prediction, CFD Royal Military of Sci-
code ence College, Cran- Column Sizes f’c (MPa) Ligature Spacing
___________________________________________
field University NSC 500x900 40 400mm and 100mm
CONWEP Blast prediction (empiri- US Army Waterways HSC 350x750 80 400mm and 100mm
___________________________________________
cal) Experiment Station
AUTO- Structural response + CFD Century Dynamics 900 mm
DYN (Couple analysis)
ABAQUS Structural response + CFD ABAQUS Inc.
(Couple analysis)

Prediction of the blast-induced pressure field on a 20-N32 500 mm

structure and its response involves highly nonlinear
behavior. Computational methods for blast-response
prediction must therefore be validated by comparing
calculations to experiments. Considerable skill is re-
quired to evaluate the output of the computer code,
N12 @400
both as to its correctness and its appropriateness to
the situation modeled; without such judgment, it is Figure 12. Cross section of the NSC column – Ordinary detail-
possible through a combination of modeling errors ing (400 mm ligature spacing).
and poor interpretation to obtain erroneous or mean-
ingless results. Therefore, successful computational
modeling of specific blast scenarios by engineers
unfamiliar with these programs is difficult, if not Blast
impossible. Pressure

10 CASE STUDY – RC COLUMN SUBJECTED 10 MPa

TO BLAST LOADING

A ground floor column (6.4m high) of a multi-storey

building (modified from a typical building designed
in Australia) was analysed in this case study (see
Fig. 12).
The parameters considered were the concrete 1.3 ms Time
strength (40MPa for NSC column and 80 MPa for Figure 13. Blast loading

85
 EJSE Special Issue: Loading on Structures (2007)

0.08

0.07 40MPa Column

0.06 80MPa Column

Deflection (m)
0.05

0.04

0.03

0.02 Point of column

0.01 failure in shear

0
0 0.002 0.004 0.006 0.008
Time (ms)
Figure 14. 3D model of the column using Explicit code LS- Figure 16. Lateral Deflection -Time history at mid point of
Dyna column with 100mm ligature spacing (SMRF).
The lateral deflection at mid point versus time 50

history of two columns made of NSC and HSC are Point of column failure

shown in Figs. 15 and 16. The graphs clearly show 40

Strain Energy (kNm)

the lateral resistance of the columns. It can be seen 30
that under this close-range bomb blast both columns
failed in shear. However, the 80MPa columns with 20
reduced cross section have a higher lateral deflec-
tion, which shows a better energy absorption capac- 10
40MPa Column

ity compared to that of the 40 MPa columns (see 80MPa Column

Fig. 17 and Table 4). 0

It can be seen from Figs. 15 and 16 that the effect 0 0.002 0.004 0.006 0.008
Time (ms)
of shear reinforcement is also significant. The ulti-
mate lateral displacements at failure increase from Figure 17. Comparison of energy absorption capacities
45mm (400 mm ligature spacing) to 63mm (100mm (100mm ligature spacing).
ligature spacing) for the HSC column. Those values Table 4. Energy absorptions at failure of HSC and NSC col-
for the NSC column are 20mm and 32mm, respec- umns
___________________________________________
tively. Column 400mm spacing 100mm spacing
___________________________________________
0.08 NSC 12.0 kNm 33.9 kNm
HSC 27.6 kNm 43.5 kNm
___________________________________________
0.07 40MPa Column

0.06 80MPa Column

Deflection (m)

0.05

0.04 10.1 Effect of strain-rate on ductility

0.03

0.02 It is evident that increasing the rate of loading will

0.01
Point of column result in increases in strength and stiffness of con-
failure in shear
crete, yield strength of steel and load-carrying ca-
0
0 0.002 0.004 0.006 0.008
pacity of reinforced concrete flexural members. A
Time (ms) parametric study has been carried out to investigate
the effects of high strain-rate on the ductility of rein-
Figure 15. Lateral Deflection -Time history at mid point of forced concrete members, and on their flexural and
column with 400mm ligature spacing (OMRF).
shear capacities. The proposed strain-rate dependent
model for concrete is adopted in this study. As
shown in Fig. 18 the flexural capacity and the ductil-
ity of a reinforced concrete column were signifi-
cantly increased due to the increase in yield strength
of steel and compressive strength of concrete at high
strain rate. The shear capacity of the column was
calculated using the Modified Compression Field
theory (Vecchio and Collins, 1986).

86
 EJSE Special Issue: Loading on Structures (2007)

1200
the building are shown in Fig. 20. The typical story
700/s height is 3.85m. Perimeter columns are spaced at
1000 typical 8.4m centers and are connected by spandrel
beams to support the facade. The lateral loads are re-
800
300/s
sisted by 6 core boxes located at the centre of the
Moment (kNm)

structural plan. The building is designed to resist lat-

600 eral loads due to wind and seismic ground motion
Static specified by Australian Loading Standards
400 AS1170.2 and AS1170.4. The slab, columns and
core walls are all cast-in-place concrete. The lateral
200 load resistance system (LLRS) of the building relies
mainly on the lateral load capacity of the core walls
0
0 0.02 0.04 0.06 0.08 0.1 0.12
which account for about 80% of the overall capacity.
Curvature (rad/m)

Figure 18. M-φ curves of a cross-section of a column at differ- 1 2 3 4 5 6 7

ent strain rates
Fig. 19 shows the increased ratio of flexural ca-
pacity (Mu.dyn/Mu.stat) and shear capacity
(Vu.dyn/Vu.stat) at high strain rate compared to those
capacities under static loading. It can be observed
from Fig. 19 that the increase in flexural strength
was greater than that of shear strength. Thus, the in-
crease in the material strengths under dynamic con-
ditions may lead to a shift from a ductile flexural
failure to a brittle shear failure mode. 8.4m typical Direction of attack

2
Increase in Blast pressure
Dynamic Increase Factor

flexural strength
1.5

1
Increase in
shear strength
0.5

0 Pmax=4.1MPa
0 200 400 600 800
Strain rate (1/s)
Fig. 20 Structural configuration
Figure 19. DIFs for flexural strength and shear strength of a
column at different strain rates In this study which is based on the local damage
assessment due to bomb blast at ground level, pro-
11 CASE STUDY - PROGRESSIVE COLLAPSE gressive collapse analyses was performed on the ex-
ANALYSIS ample building. The structural stability and integrity
Design recommendations on progressive collapse of the building were assessed by considering the ef-
analysis have been introduced in British Standards fects of the failure of some perimeter columns, span-
since 1968, after the collapse of 22-storey Ronan drel beams and floor slabs due to blast overpressure
Point apartment building. In recognition of this is- or aircraft impact. The main purpose of the analysis
sue, a number of European countries, USA and Can- is to check if failure of any primary structural mem-
ada have incorporated progressive collapse provi- ber will cause progressive collapse propagating be-
sions in their building codes. The American National yond one story level above or below the affected
Standards Institute (ANSI) Standard A58.1-1982, member vertically, or to the next vertical structural
“Minimum Design Loads for Buildings and other member.
Structures” recommends the alternative path
method, in which the local failure is allowed to oc-
cur but an alternative path must be provided around
the failed structural elements.
A 52 storey building (modified from a typical tall
building designed in Australia) was analyzed in this
study. The plan view and structural configuration of
87
 EJSE Special Issue: Loading on Structures (2007)

1 2 3 4 5 6 7
th
6 Level
th
5 Level
th
4 Level
rd
3 Level
nd
2 Level
st
1 Level
Extreme bending
and shear under Ground Flr.
direct pressure
Damaged columns,
beams and slabs Point of detonation (20m dist.)
Fig. 21 Direct column loading (Blast pressure)

Net upward Critical beams

pressure on slab
Fig. 23 Progressive collapse analysis of perimeter frame (dam-
aged by blast load)
Fig. 22 Uplifting of floor slabs (Blast pressure)

Figs. 21 & 22 show the effect of direct blast pres- As seen from Figs. 23, the alternative load paths
sure on perimeter columns, beams and floor slabs. go through columns surrounding the damaged area
The concrete slabs in this example building are where the vertical loads are transferred. Beams and
125mm thick supported by prestressed wide band floor slabs above that area become critical due to the
beams. The portion of floor slabs in close proximity loss of the supporting columns. The overall stability
to the blast were directly hit by the blast overpres- of the structure will rely on continuity and ductility
sure. The normal glazing façade offers insignificant of these elements to redistribute forces within the
resistance to the blast wave so after the failure of structure. The falling debris of the collapsed mem-
glazing, the blast fills the structural bay above and bers also imposes severe loading on the floors be-
below each floor slab. The pressure below the slab is low. It is essential to check whether that overload
greater than the pressure above and causes the net can be carried without causing further collapse.
upward load on each slab (Fig. 22). Parametric studies were also undertaken to inves-
To detect local damage, the blast analysis was tigate the impact resistance of the floor slab, assum-
carried out for perimeter columns, beams and floor ing a floor above had collapsed onto it. The collaps-
slabs based on the actual blast pressures on each ing floor was treated as falling debris, i.e. the energy
element. Results plotted in Fig. 23 show column applied to the floor below was the weight of the col-
lines 4, 5 of the ground and 1st levels failed due to lapsing floor multiplied by the height through which
the direct impact of the blast wave. Slabs and beams it fell. To obtain an estimate of the impact
from column line 3 to 6 also collapsed. Member as- load-bearing capacity of the floor slab, the structure
sessment was carried out using program RE- was analyzed using program LSDYNA. In addition
SPONSE (2001) based on the Modified Compres- to material and geometric nonlinearities, the analy-
sion Field theory and LSDYNA (see section 10). ses considered membrane action, inertia effects, and
More details are given elsewhere (Mendis & Ngo, other influencing factors. The results show that the
2002). The calculation also showed that if the col- ultimate capacity of the floor slab is approximately
umns were detailed using requirements for special 16.5kPa which is 2.75 times the total floor load
moment resisting frames (SMRS) as given in ACI- (dead load plus 0.4 live load). Therefore in this case
318 Section 21, the shear capacity and the ductility study if more than two floor collapsed, the falling
would be improved significantly, thus improving the debris of the collapsed members may impose an
blast and impact resistance of the member. The overload for the floor below and trigger a progres-
damaged model of perimeter frame, in which failed sive collapse of the example building.
elements were removed (Fig. 23), was analyzed to
check if progressive collapse would propagate be-
yond one story level above or below. 12 BLAST EFFECTS ON BUILDING’S FAÇADE

When a terrorist bomb explodes in an urban area, air

blast pressure typically fractures lightweight façades
88
 EJSE Special Issue: Loading on Structures (2007)

such as windows, causing catastrophic results. As a 1 2 3 4 5 6 7

consequence of façade failure, the blast enters build- 8kPa 7th Level
ings, causing relatively superficial structural dam- 11kPa 6th Level
age, but great disruption to the working environment
24kPa 5th Level
and the mechanical and electrical services. Falling
window glass shards cause injuries in addition to 4th Level
73kPa
those resulting from blast pressure. Office fixtures 3rd Level
and fittings are destroyed, and suspended ceiling and 0.5MPa 2nd Level
partitions are disturbed. Broken glass is projected
everywhere, with particularly severe consequences 1.6MPa 1st Level

for air conditioning systems.

Past events have shown that even relatively small 4.1MPa
Ground Flr.
explosions can cause significant window glass
breakage, requiring as a minimum, window glass re- Point of detonation (20m dist.)
placement and significant cleanup. Ideally, properly Fig. 25 Distribution of blast pressure on building façade
designed blast-resistant glazing should minimise, if (Mendis & Ngo, 2002)
not eliminate, flying and falling glass shards in any
explosion. In addition, under air blast pressure load- There are only few guidelines developed to de-
ing, properly designed blast-resistant glazing should sign blast-resistant façades. One of those is TM 5-
maintain closure of its fenestration, reducing 1300, Structures to Resist the Effects of Accidental
cleanup costs and reducing pressure-related injuries. Explosions. In Section 6 of TM 5-1300, guidance is
Even with blast-resistant glazing, air blast pressure provided for the design, evaluation, and certification
will fracture windows, necessitating replacement. of windows to safely survive a prescribed blast envi-
However, blast-resistant glazing should remain in its ronment described by a triangular-shaped pressure-
openings and reduce the urgency for immediate re- time curve (Fig. 25). Window designs using mono-
placement. lithic (un-laminated) thermally tempered glass based
Advanced computer codes such as CFD have on these guidelines can be expected to provide a
been used to simulate the blast effects in the urban probability of failure equivalent to that provided by
environment (Figs. 23, 24). A typical tall building current safety standards for safely resisting wind
subjected to a bomb blast detonated at different loads.
stand-off distances from the ground level was ana- An example of a glazing system is illustrated in
lysed in this study. The peak overpressure is 4.1MPa Fig. 25. The glazing is a rectangular, fully thermally
at the ground level and reduces rapidly up the height tempered glazing panel having a long dimension, a;
of the building. The average duration of loading was a short dimension, b; and thickness, t. The glass
adopted as 15 milliseconds. Façade damage at dif- pane is simply supported along all four edges, with
ferent levels was assessed based on the blast pres- no in-plane and rotational restraints at the edges.
sure distribution and also using other in-house pro- The bending stiffness of the supporting elements of
grams. the panel is assumed to be an order of magnitude
higher than that of the glass pane. Recent static and
blast load tests indicate for a case in which the sup-
porting frame deflects by up to 1/264 of the span
distance, the glass pane resistance will not be sig-
nificantly different to the ideal conditions of infi-
nitely rigid supports.
The blast pressure loading is described by a peak
triangular-shaped pressure-time curve as shown in
Fig. 25. The blast pressure rises instantaneously to a
peak blast pressure, B, and then decays linearly with
a blast pressure duration, T. The pressure is uni-
formly distributed over the surface of the pane and
applied normal to the pane.
Fig. 24 CFD modelling of blast pressure on building structures The resistance function, r(X) (static uniform load,
(Mendis & Ngo, 2002) r as a function of centre deflection, X) for the plate
accounts for both bending and membrane stresses.
The effects of membrane stresses produce a nonlin-
ear stiffness of the resistance-deflection function
(Fig. 25). The design deflection, Xu is defined as the

89
 EJSE Special Issue: Loading on Structures (2007)

centre deflection where the maximum principle ten- calculation is also provided in TM5-1300. This can
sile stress at any point in the glass first reaches the be used to evaluate the blast capacity of glass when
design stress, fun. Typically, as the deflection of the interpolation between charts is unadvisable, when
plate exceeds a third of its thickness, the points of design parameters are outside the limits of the chart.
maximum stress will migrate from the centre and Peak blast pressure capacity
towards the corners of the plate. for tempered glass panes:
a/b = 1.00, Tg = 1/4 inch
The model, illustrated in Fig. 25 uses a single-
20
degree-of-freedom system to simulate the dynamic Numbers next to curves indicate
response of the plate. The model calculates the peak pane dimension b, inches

blast pressures required to exceed the prescribed 10 12

Peak Pressure, B (psi)

probability of failure. The model assumes that fail-
ure occurs when the maximum deflection exceeds 5 18

ten times the glazing thickness. More details of dy- 24

namic analysis of the glazing system can be found in 30
TM5-1300. 36

1 42
48
54
60
0.5
2 5 10 50 100 500 1,000
Duration of Blast Pressure, T (msec)
Figure 27 - A design chart for tempered glass panel (TM5-
1300)

13 CONCLUSIONS

For high-risks facilities such as public and com-

mercial tall buildings, design considerations against
extreme events (bomb blast, high velocity impact) is
very important. It is recommended that guidelines on
abnormal load cases and provisions on progressive
collapse prevention should be included in the current
Building Regulations and Design Standards. Re-
quirements on ductility levels also help improve the
building performance under severe load conditions.

14 REFERENCES
Figure 26 – Blast analysis for a glazing system

1. Baker, W.E., Cox, P.A., Westine, P.S., Kulesz, J.J. and

Design Charts Strehlow, R.A., Explosion hazards and evaluation, El-
sevier Scientific Publishing Company, New York, 1983.
In TM5-1300, guidelines are also provided in the 2. Biggs, J.M, Introduction to Structural Dynamics,
form of load criteria for the design of both the glass McGraw-Hill, New York, 1964.
3. Brode, H.L., “Numerical solution of spherical blast
panes and framing system for the window. The crite- waves”, Journal of Applied Physics, American Institute of
ria account for both the bending and membrane Physics, Ney York, 1955.
stresses and their effect on maximum principle 4. Comité Euro-International du Béton, CEB-FIP Model
stresses and the nonlinear behaviour of glass panes. Code 1990, Redwood Books, Trowbridge, Wiltshire, UK,
1990.
The criteria cover a broad range of design parame- 5. DAHS CWE, Technical Manual - Design and Analysis of
ters for rectangular-shaped glass panes. Design Hardened Structures to Conventional Weapons Effects,
charts (see Fig. 26 as an example) are presented for U.S. Army Corps of Engineers (CEMP-ET), Washington
monolithic thermally tempered glazing with blast DC, September, 1998.
6. DOE/TIC-11268. A Manual for the Prediction of Blast
overpressure capacity up to 690 KPa, an aspect ratio and Fragment Loadings on Structures, U.S. Department
of 1.00 ≤ a/b ≤ 4.00, pane area 0.1 ≤ ab ≤ 2.3 m2, of Energy, Washington DC, 1992.
and nominal glass thickness 6mm ≤ t ≤ 20mm. An 7. Dowling, A.R. and Harding, J., “Tensile properties of
mild steel under high strain rates”, Proceedings of the 1st
alternative method for blast capacity evaluation by HERF Conf., University of Denver, Colorado, 1967.

90
 EJSE Special Issue: Loading on Structures (2007)

8. Ellis, B.R. and Crowhurst, D., “The response of several

LPS maisonettes to small gas explosions”. IStructE/BRE
Seminar: Structural design for Hazardous Loads: The
Role of Physical Tests, Construction Press, New York,
1991.
9. ESL-TR-87-57, Protective Construction Design Manual,
U.S. Air Force Engineering and Services Center, Tyndall
Air Force Base, Florida, 1989.
10. Grote, D., Park, S., and Zhou, M., “Dynamic behaviour of
concrete at high strain rates and pressures”, Journal of
Impact Engineering, Vol. 25, Pergamon Press, New
York, pp. 869-886, 2001.
11. Longinow A, and Mniszewski KR., “Protecting buildings
against vehicle bomb attacks”, Practice Periodical on
Structural Design and Construction , ASCE, New York,
pp. 51-54, 1996.
12. LS-DYNA Version 960. 2002. Livermore Software Tech-
nology Corporation.
13. Malvar LJ., “Review of Static and Dynamic Properties of
Steel Reinforcing Bars”, ACI Materials Journal, 95(5),
ACI, Detroit, Michigan, pp. 609-616, 1998.
14. Mendis, P.A., and Ngo, T. “Assessment of tall buildings
under blast loading and aircraft impact”, Toward a better
Built Environment, Innovation, Sustainability and Infor-
mation Technology, International Association of Bridge
and Structural Engineering, Australia, pp. 375-376,
2002.
15. Mills, C.A., “ The design of concrete structure to resist
explosions and weapon effects,” Proceedings of the 1st
Int. Conference on concrete for hazard protections, Edin-
burgh, UK, pp. 61-73, 1987.
16. Milne-Thompson, L.M., Theoretical Hydrodynamics,
McMillan, New York, 1968.
17. Newmark, N.M. and Hansen, R.J., “Design of blast resis-
tant structures,” Shock and Vibration Handbook, Vol. 3,
Eds. Harris and Crede. McGraw-Hill, New York, USA.
1961.
18. Ngo, T., Mendis, P., & Itoh, A., “Modelling reinforced
concrete structures subjected to impulsive loading using
the Modified Continuum Lattice Model”, Proc. of 18th
Australasian Conference on the Mechanics of Structures
and Materials, Perth, Australia. 2004b.
19. Ngo, T., Mendis, P., Hongwei, M. & Mak, S., “High
strain rate behaviour of concrete cylinders subjected to
uniaxial compressive impact loading”, Proc. of 18th Aus-
tralasian Conference on the Mechanics of Structures and
Materials, Perth, Australia. 2004a.
20. Norris, G.H., Hansen, R.J., Holly, M.J., Biggs, J.M., Na-
myet, S. and Minami, J.K., Structural design for dynamic
loads, McGraw-Hill, New York, USA. 1959.
21. TM 5-1300, The Design of Structures to Resist the Effects
of Accidental Explosions, Technical Manual, US Depart-
ment of the Army, Navy, and Air Force, Washington DC,
1990.
22. TM 5-855-1, Fundamentals of Protective Design for
Conventional Weapons, U.S. Department of the Army,
Washington DC, 1986.
23. Vecchio, F.J., & Collins, M.P. 1986. The Modified Com-
pression Field Theory for Reinforced Concrete Elements
Subjected to Shear, ACI Structural Journal, V83, 2, pp.
219-231.

91