You are on page 1of 13

Materials and Design 65 (2015) 662674

Contents lists available at ScienceDirect

Materials and Design


journal homepage: www.elsevier.com/locate/matdes

Numerical analysis of prestressed reinforced concrete beam subjected


to blast loading
Wensu Chen , Hong Hao, Shuyang Chen
Tianjin University and Curtin University Joint Research Center of Structural Monitoring and Protection, School of Civil and Mechanical Engineering, Curtin University, Kent
Street, Bentley, WA 6102, Australia

a r t i c l e

i n f o

Article history:
Received 1 July 2014
Accepted 14 September 2014
Available online 2 October 2014
Keywords:
Prestressed reinforced concrete beam
Blast resistance
Dynamic response

a b s t r a c t
Prestressing technique has been widely used in civilian and military constructions. The prestressed
reinforced concrete (RC) structural components such as beams and columns usually outperform the
non-prestressed RC components because prestressing not only increases the structural stiffness and load
carrying capacity, but also has higher crack resistance than non-prestressed component. As a result, it
usually leads to light structures. The investigation of non-prestressed RC components subjected to blast
loadings has been reported in the literature. However, the study on the blast-resistant capacity of prestressed RC components is very limited. In this study, the dynamic response of a simply-supported prestressed RC beam with rectangular section under blast loadings is numerically investigated by using nite
element codes LS-DYNA. The prestress is pre-applied on the RC beam in an analytical approach. The reliability of the numerical model is calibrated with testing results available in the literature. With the calibrated model, numerical simulations on four groups of prestressed RC beams to blast loadings are
carried out to investigate the inuences of prestressing level and concreted compressive strength on
beam blast loading resistance capacity. The structural responses such as mid-span maximum deection,
residual deection, cracking, stress of rebars and shear stress of concrete near the supports are extracted
from the numerical results. The effectiveness of prestressing on blast-resistant capacity of RC beam is
demonstrated through comparing the results with the bench marking non-prestressed RC beam under
the same blast loadings.
2014 Elsevier Ltd. All rights reserved.

1. Introduction
Blast loadings due to terrorist bombing or accidental gas
explosion may cause signicant structural damage, casualty and
economic loss. Reinforced concrete is commonly used in the building industry. Conventional RC column, beam and panel, as major
load carrying components, are often damaged when subjected to
blast loading, which might lead to partial or total collapse of building structures. In 1968, an internal gas explosion seriously damaged the Ronan Point residential apartment building in the UK,
owing to the failure of some structural components that triggered
progressive collapse [1]. Engineering solutions for structure protection need to be developed and improved to ensure the safety
of structures. To overcome concretes natural weakness in tension
and the growth of cracks, prestressed technique is employed in
both civilian and military constructions. Prestressing concrete
Corresponding author.
E-mail
(W. Chen).

addresses:

wensu.chen@curtin.edu.au,

http://dx.doi.org/10.1016/j.matdes.2014.09.033
0261-3069/ 2014 Elsevier Ltd. All rights reserved.

wensu.chen@hotmail.com

may increase structural stiffness and crack-control performance.


As a result it reduces structural member size, self-weight and construction cost. Prestressing can be achieved in three ways, i.e., pretension, bonded post-tension and unbonded post-tension by using
tendons [2]. Prestressed tendons are made of high tensile steel
cables or rods, which provide compressive stress on concrete.
The compressive stress balances the tensile stress of concrete
when subjected to external loads. Consequently, the appearance
and growth of cracks in concrete might be reduced and delayed
in the tension zone of concrete, as shown in Fig. 1. Various standards [3,4] provide guides on designing the prestressed RC beams
to resist static loads, but the effects of prestressing concrete on RC
structures blast loading resistance capacities are relatively less
studied.
The research on the non-prestressed concrete components
subjected to blast and impact loadings has attracted intensive
attentions over the years. The exural damage is usually observed
in the RC structures under the blast loadings. Fang and Wu [5]
found that the exural damage transferred to the brittle shearing
damage with increasing the loading rate, height of cross section,

W. Chen et al. / Materials and Design 65 (2015) 662674

longitudinal reinforcement ratio or decreasing the strength of


concrete. Lin et al. [6] numerically investigated the response of
reinforced concrete panels under blast loadings by using LS-DYNA.
The effect of blast intensity and panel specication on the panels
blast resistance performance was studied. Magnusson and Hallgren [7] carried out the tests on the high strength reinforced concrete beams to air blast loading, and observed that the failure
modes of beams were different when subjected to static and
dynamic loads. All beams subjected to static loading failed in exure mode and some of the beams subjected to air blast loadings
failed in shear mode. The concrete beams showed an increased
load carrying capacity when subjected to blast loadings compared
to static loading. Jiang et al. [8] numerically investigated reinforced
concrete beams subjected to impact loadings by using LS-DYNA.
The elasto-plastic damage cap (EPDC) model for the concrete and
the elasto-plastic model for the reinforcement are incorporated
into the numerical model. Fang et al. [9] reported that strain rate
effect has a signicant effect on blast-resistant capacities of reinforced concrete beam. Yi et al. [10] experimentally investigated
the blast-resistant capacities of ultra-high performance concrete
and reactive powder concrete in concrete structures, which have
been proved having better blast-resistant capacity than normal
strength concrete structure.
Most of the previous studies on high-rate dynamic structural
responses of the prestressed RC components are subjected to
impact loadings. Very limited study on the blast-resistant capacity
of prestressed RC components has been found in the literature. Li
et al. [11] tested three partially-prestressed concrete (PPC) beams
by using drop hammer. The impact-resistant performances of the
PPC beams were demonstrated in the study. Ishikawa et al.
[12,13] experimentally and analytically investigated the performance of bonded and unbonded prestressed concrete beams subjected to impact loadings. It was found that the prestressed
tendon was more vulnerable to break under higher loading rate.
Iskhakov and Ribakov [14] proposed a design method for two-layered bending prestressed beams which consist of steel bered high
strength concrete in compressed zone and normal strength concrete in tensile. Cramsey and Naito [15] reported the performance
of a 30-ft partially prestressed concrete panel subjected to different
blast loadings. It was found that the panel could sustain rotation of
2.7 without failure under the highest blast load considered in the
study. Three ultra-high strength concrete panels prestressed by
using high-strength steel tendons were tested to investigate the
blast-resistant performance under blast loadings. The results demonstrated the good performances of these panels, which all survived the blast tests with only minor cracks even with
considerable deections [16]. An analytical approach of nite layered section combing with rate-sensitive model was proposed to
investigate the dynamic responses of the partially prestressed
concrete beams subjected to blast loadings [17]. Cofer et al. [18]

Fig. 1. Schematic diagrams of non-prestressed and prestressed beams.

663

validated numerical model of prestressed concrete girder with


bulb-tee section subjected to explosions above and below the girder. A simple-span bridge with prestressed girder and slab was also
modeled to examine the damage when subjected to different blast
scenarios. Unfortunately, the above studies only reported limited
test data and general observations. Some important information
such as detonation weight is not revealed. Therefore they cannot
be used to quantitatively assess the effectiveness of prestressing
concrete structures on their blast load-carrying capacities.
All the above studies indicated the effectiveness of prestressing
concrete in increasing RC structures blast-load or impact-load
carrying capacities. Those studies concentrate on analyzing the
particular tested structural components and verifying the specic
numerical models. There is no systematic study in the literature
that devotes to examining the effectiveness of prestressing levels
on enhancing the blast-load carrying capacities of RC beams. In
the present paper, the dynamic response of simply-supported prestressed RC beams subjected to blast loadings is numerically investigated. The reliability of the numerical model is calibrated with
some testing results available in the literature. With the calibrated
model, numerical simulations of prestressed RC beams to blast
loadings are carried out by considering varied prestressing levels,
different concreted compressive strengths and blast intensities.
The mid-span maximum deection, residual deection, shear
stress and cracking of RC beams without prestressing or with
different levels of prestressing are extracted from numerical simulations. The numerical results of prestressed and non-prestressed
RC beams are compared to investigate the effectiveness of prestressing on blast-resistant performance of RC beams.

2. Numerical model calibrations


The numerical simulation is carried out by using commercial
software LS-DYNA 971 [19]. LS-DYNA is based on explicit numerical
methods and has been widely employed to analyze the problems
associated with large deformation structure response to high velocity impact and blast load, and high strain rate behavior of materials.
It has been proven yielding reliable numerical predictions of structural response and damage to blast loadings.

2.1. Experimental descriptions


To verify the accuracy and reliability of the nite element
model, numerical model is calibrated with the experimental and
numerical results available in the literature. It should be noted that
no blasting test on prestressed RC structures with sufcient information such as charge weight, dimension, and prestressing level,
etc. can be found in the open literature. Therefore, in the present
study, testing data conducted on RC structures without prestressing is adopted to verify the numerical model. It is believed that
the numerical model calibrated is applicable to simulating blast
responses of prestressed RC beams because, as will be described
later, the simulation of the responses of the prestressed RC beam
to blast loads is performed in two steps, i.e., applying the static
prestressing in the rst step, and applying the blast loading in
the second step, and the two steps are not coupled. Prestressing
only affects the initial conditions of the RC beam when blast load
is applied.
A 1/4-scale RC frame model was tested by Baylot and Bevins
[20] to study the dynamic response of the central column of a
two-storey RC structure. A 7.1 kg C4 high explosive of hemisphere
shape (with equivalent TNT weight of 8 kg) at a standoff distance
of 1.07 m was detonated in the test. The experimental setup and
results are shown in Fig. 2.

664

W. Chen et al. / Materials and Design 65 (2015) 662674

Fig. 2. Tested column (L) before blast loading; (R) after blast loadings [21].

2.2. Finite element model and boundary condition


In the previous studies, Shi et al. [22] and Mutalib and Hao [23]
carried out the numerical analysis of the tested RC column and the
results were compared with the test results [20]. The cross section
of the tested exterior central column is 85 mm by 85 mm and the
height is 900 mm. As shown in Fig. 3, the longitudinal reinforcements and the cross tie reinforcements are 3.2 mm and 1.6 mm,
respectively. The material properties of concrete and steel reinforcement are given in Table 1.
A numerical model is developed in LS-DYNA in this study to
simulate the blasting test of this column. 8-nodes constant stress
solid element (SOLID_164) with 1-point quadrature integration
and 3-nodes beam element (BEAM_161) with 2  2 Gauss quadrature integration are employed to model concrete and steel reinforcement, respectively. Perfect bond between reinforcing steel
and surrounding concrete is assumed. To avoid penetration
between concrete and steel reinforcements, the penalty-based contact algorithm Contact Automatic Single Surface is dened.
Convergence test is carried out by halving the mesh size. It
shows that the simulation converged when the mesh size is
5 mm. Further decrease in the element size only has insignicant
inuence on the numerical results but increases the computing

time and the risk of computer memory overow. Therefore,


5 mm mesh is used in the numerical model.
For boundary condition, the nodes of the head and footing are
constrained in horizontal directions, i.e., X and Y directions. The
footing is further constrained against vertical displacement, i.e., Z
direction. An axial pressure of 2.1 MPa is applied to the head of
the column before the blast loading is applied on the column surface, which is determined based on the static weight carried by the
column, and is also equivalent to applying a uniform pre-compressive stress in the column.
2.3. Material model
A number of material models available in LS-DYNA can be used
to model the behavior of concrete material such as Mat Brittle
Damage (MAT_96), Mat Johnson Holmquist Concrete (MAT_111),

Mat Pseudo Tensor (MAT_16), Mat CSCM Concrete (MAT_159)


and Mat Concrete Damage Rel3 (MAT_72_REL3) etc. In this study,
the material model MAT_72_REL3 is used to simulate the concrete
material. This material model considers strain-rate effect, plasticity
and damage softening after failure. The reliability of the material
model in predicting the response of reinforced concrete structure
subjected to blast loadings has been demonstrated by [22,24,25].

Fig. 3. (L) Finite element model; (R) Cross section of test column [20].

665

W. Chen et al. / Materials and Design 65 (2015) 662674


Table 1
Material properties of steel reinforcement and concrete [22].
Material

LS-DYNA Model

Input parameter

Magnitude

Longitudinal steel reinforcement

MAT_PLASTIC_KINEMATIC (MAT_24)

Mass density
Youngs modulus
Poissons ratio
Yield stress
Failure strain

7800
207 GPa
0.3
450 MPa
0.18

Cross-tie and hoop Reinforcement

MAT_PLASTIC_KINEMATIC (MAT_24)

Mass density
Youngs modulus
Poissons ratio
Yield stress
Failure strain

7800
207 GPa
0.3
400 MPa
0.18

Concrete

MAT_CONCRETE_DAMAGE_REL3 (MAT_72)

Mass density
Unconned strength
Failure strain

2400
42 MPa
0.1

The material model is based on unconned compressive strength


input parameter, which is 42 MPa in this study.
The material model named Mat Piecewise Linear Plasticity
(MAT_024) is used to model longitudinal and cross-ties and hoop
reinforcements. This material model considers isotropic and kinematic hardening plasticity. It allows the denition of arbitrary
stress versus stain curve and strain rate curve. The material properties of the reinforcements are given in Table 1.
To avoid computation overow due to large deformation of concrete, the card Mat Add Erosion is utilized to eliminate concrete
elements that do not contribute to resist blast load. In the present
study, the failure criteria are based on the unconned tensile
strength and principal strain. The unconned tensile strength of
concrete is calculated according to the formula proposed by CEB
Code [26].

" 0 #1=3
fc
b0

f t 1:58

where f t is the unconned tensile strength of concrete, b0 is a unit


0
conversion factor and f c is the unconned compressive strength. In
this study, the unconned compressive strength of concrete is
42 MPa and the unconned tensile strength is calculated as
3.5 MPa. The concrete element will be deleted when it either
reaches the unconned tensile stress of 3.5 MPa or the principal
strain of 0.1.
2.4. Strain rate effect
Under high-speed impacts or blast events, the inertia and
strain-rate effects become important factors for structural
response, which is different from the quasi-static scenario. The
concrete and steel are strain rate dependent materials. The
strengths of concrete and steel materials are enhanced under high
strain rate response. The strength increment is dened by the
dynamic increase factor (DIF) at given strain rate, i.e., the ratio of
dynamic-to-static strength versus strain rate. Many empirical
equations have been proposed to determine the strain rate effect
of concrete and steel based on experimental data. The strain rate
effect is considered in the concrete and steel material models by
incorporating the DIF relationships. The DIFs of concrete compressive strength and tensile strength are dened by CEB Code 1990
[26] and Malvar and Ross [27], respectively.
The compressive DIFs of the concrete are given by the following
equations [26]

CDIF

fc

f cs

e_
e_ cs

1:026a

for e_  30 s1

CDIF

fc
1=3
ce_ for e_ > 30 s1
f cs

where fc is the dynamic compressive strength at strain rate e_ ; fcs is


the static compressive strength at e_ cs ; log c = 6.156a  0.49; a = 1/
(5 + 3fcu/4) and fcu is the static cube compressive strength in
MPa.The tensile DIFs of the concrete are given by the following
equations [27]

TDIF
TDIF

ft

f ts

e_
e_ ts

ft
b
f ts

d


for e_ td 6 1 s1

e_
e_ ts

1=3

for e_ td > 1 s1

4
5

where ft is the dynamic tensile strength at strain rate e_ in the range


of 106  s1 160 s1; fts is the static tensile strength at
e_ ts ; log b 6d  2; d 1=1 8f 0c =f 0co ; f 0c is the static uniaxial com0
pressive strength of concrete and f co is taken as 10 MPa. The DIFs of
concrete used in this study are shown in Fig. 4 (L).
The DIF relationship for steel reinforcement dened by Malvar
[28] is utilized, as follows.


DIF

e_

a

104
a 0:074  0:040f y =414

6
7

where fy is the steel yield strength in MPa. The DIF of steel used in
this study is shown in Fig. 4 (R). It should be noted that to avoid
overestimation of the DIF of steel material at high strain rate, DIF
is taken as a constant when strain rate is larger than 200 1/s as
shown in Fig. 4.
2.5. Blast load modeling
In LS-DYNA, the function Load_Blast_Enhanced has been proven in simulating blast loads on structures and it has been used
in numerical simulations of structural responses to blast loads
[29,30]. The amplitudes of blast loadings are determined by the
scaled distance (Z) and the incident angle. The scaled distance is
dened as Z = R/W1/3, where R is the standoff distance in meters
and W is the amount of TNT equivalent in kilograms [31]. In this
study, the blast loadings with positive and negative phases can
be idealized as triangular pressure time history and applied on
the structure. For a stiff structure, the inuence by the negative
phase blast pressure is relatively small compared to that of the
positive phase. Therefore, the negative pulse is usually neglected
to save computational cost. The positive pressure time history is
modeled by a steep rise to peak pressure Pr. and then decays to

666

W. Chen et al. / Materials and Design 65 (2015) 662674


8

1.4

TDIF (concrete in tension)


CDIF (concrete in compression)

Steel
1.3

DIF

DIF

1.2

4
1.1
3
1.0

2
1
-4

-3

10

-2

10

-1

10

10

10

10

10

0.9
-4
10

-3

10

-2

-1

10

10

10

10

10

10

-1

-1

Strain rate (S )

Strain rate (S )
Fig. 4. DIFs versus strain rate (L) Concrete; (R) steel.

7000

Pressure (kPa)

6000
5000
4000
3000
2000
1000
0
0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

Time (ms)
Fig. 5. Idealized air explosion pressure time history.

ambient pressure. It can be dened by the arrival time ta, the ctitious duration t0f and the positive reected peak pressure Pr. The
ctitious duration is determined by the positive reected impulse
Ir and the positive reected peak pressure Pr [31] as

t of

2Ir
Pr

In this study, the peak pressure and positive reected impulse Ir


of the blast load corresponding to the eld blasting test are
6100 kPa and 1007 kPa ms, respectively [22]. The duration of positive phase (t0f) is calculated as around 0.33 ms. The idealized pressure time history is shown in Fig. 5. The blast loadings are assumed
to be uniformly distributed on the front surface of the target. It
should be noted that the assumption of uniform blast loading distribution is only valid when the charge is not too close to the object.

Fig. 6. Plastic strain contours (a) maximum deection (t = 3 ms); (b) residual
deection.

indicating the numerical model gives reliable predictions of RC column response to blast loadings.

2.6. Results and comparisons

3. Numerical model of prestressed RC beam

Figs. 6 and 7 show the plastic strain and Von Mises stress distributions along the column, respectively. High plastic strain occurs
near the supports and the mid-height of the inner vertical face,
where the cracks appear and grow. The damage mode is the combination of exural and shear damage.
Fig. 8 shows that the mid-height displacement time histories of
the column. The maximum deection and residual deection at
mid height of column are around 11.6 mm and 6.9 mm,
respectively. Fig. 9 shows the numerical results agree well with
the testing results reported by Baylot and Bevins [20], as well as
the numerical results by Shi et al. [22] and Mutalib and Hao [23],

The above calibrated numerical model is used to perform a series of simulations of RC beams with different prestressing levels
subjected to blast loadings. The structural responses such as the
mid-span maximum deection, residual deection and shear stress
in concrete elements near supports are compared to examine the
effectiveness of prestressing RC beam on blast-resistant capacities.
3.1. Beam specications
The specication of a prestressed RC beam in [11] is used in this
simulation. The simply supported prestressed RC beam of 220 mm

W. Chen et al. / Materials and Design 65 (2015) 662674

667

(width) by 160 mm (depth) and 2600 mm (span) is considered. The


nite element model is shown in Fig. 10. The diameters of longitudinal reinforcements at top and bottom sides are 8 mm and
12 mm, respectively. The hoop reinforcements of diameter 6 mm
are spaced at 100 mm along the beam. A prestressed tendon with
a diameter of 15.2 mm is utilized as shown in Fig. 11. Without losing the generality, the beam is assumed to subject to a uniformly
distributed triangular blast load of peak reected pressure
6100 kPa, positive reected impulse 3050 kPa ms and duration of
the positive phase 1.0 ms.
3.2. Method of prestressing

Fig. 7. Von Mises stress contours (a) maximum deection (t = 3 ms); (b) residual
deection.

0.012

Peak deflection=11.6mm

Deflection (m)

0.010

0.008

Residual deflection=6.9 mm

0.006

0.004

0.002

0.000
0.00

0.01

0.02

0.03

0.04

0.05

Two methods for applying prestressing in numerical modeling


have been found in the literature. Ngo et al. [16] modeled the
prestress by setting initial tensile stresses in steel reinforcement
elements and initial compressive stresses in concrete elements.
This approach is straightforward in applying the prestress in RC
elements. The drawback is that the prestresses in concrete elements are not necessarily uniformly distributed along the cross
section and it is not straightforward to determine the prestress distributions. Moreover this approach neglects the deformation
induced by applying prestress to the RC member. This small initial
deformation might not be critical to beam responses when the
response mode is primarily exural. When the response mode is
governed by shear under large amplitude short duration blast loading, this initial small hogging deformation may signicantly affect
beam responses. To overcome the above problems, Bi and Hao [32]
established the relationship of the preload applied and the initial
hogging deection of RC beam through analytical method. In the
latter study, the prestress in the beam is generated by applying initial hogging deformation of RC beam. This method is adopted in the
present study. For the beam under consideration, the calculations
of the equivalent bending moment and the applications of the prestress into the beam are presented below.
Fig. 12 illustrates a RC beam with a prestressing tendon. Assuming the prestressed tendon is placed in the concrete at a certain distance below the neutral axis of the cross section, prestressing the
tendon is equivalent to applying an eccentric point load on the
beam cross section, which results in an uniform bending moment
acting on the beam, besides a compressive point load. The likely
stress block is shown in Fig. 12. The initial hogging deformation
at the mid-span of the beam caused by prestressing the tendon
can then be obtained by,

Time (s)

Fig. 8. Mid-height deection time history.

Present study
Baylot & Bevins (2007) measured deflection
Mutalib & Hao (2010)
Shi et al. (2008)

Deflection (m)

0.014

M F  d r p  As  d

0.012
0.010
0.008
0.006
0.004
0.002
0.000
0.00

0.01

0.02

0.03

0.04

Time (s)
Fig. 9. Comparisons of deection time histories.

where M is the bending moment; l is the span length of the beam; E


is the Youngs modulus of concrete; I is the moment of inertia of
cross section.
The bending moment at neutral axis (M) is given by

0.018
0.016

Ml
8EI

0.05

10

where rp is the prestress applied in the tendon; As is the cross


sectional area of the prestressed bar; and d is the distance between
the neutral axis and the prestressed bar.
In this study, the parameters are assumed as follows.
Length of beam (l) = 2.6 m;
Prestressed tendon diameter r = 15.2 mm; Prestressed stress:
rp = 315 MPa;
Cross sectional area of the prestressed bar:
As = pr2 = p(0.015/2)2 = 1.767  104 m2
Prestressed force: F = rp  As = 315  1.767  104 = 55.7 kN;
Bending moment at neutral axis: M = F  d = 55.7  0.04 =
2.23 kN  m;

668

W. Chen et al. / Materials and Design 65 (2015) 662674

Fig. 10. Model of concrete and reinforcement cage.

Table 2
Initial stress in concrete with respect to prestress level (Group #1).

Fig. 11. Schematic diagram of cross section of prestressed RC beams.

1
1
Moment
of
Inertia:
I 12
bh 12
 0:22  0:163 7:5
105 m4 ;
Therefore, the initial hogging deformation of the beam at
mid-span for the prestress of 315 MPa is calculated as
2

1:25 mm
d Ml
8EI
The calculation includes two steps. In the rst step, the initial
hogging deformation distribution along the beam is estimated with
the calculated mid-span deformation. The nodes along the beam
are dened as displacement restraints in ANSYS. The initial geometry and the initial prestress of the beam can be obtained through
the implicit analysis. The bending moment can be applied through

MPa

Initial midspan
displacement
mm

Maximum initial
principal stress
(in tension)
MPa

Maximum initial
principal stress (in
compression)
MPa

0
315
472
630

0
1.25
2.0
2.5

0
2.289
3.653
4.067

0
2.356
3.734
4.193

Beam

Prestress

No.
#11
#12
#13
#14

the initial hogging deformation. In the second step, the conguration and initial stress state of the beam resulted from the rst-step
implicit analysis are taken as the initial condition for the subsequent explicit analysis under blast loading in LS-DYNA through
its implicit to explicit option.
As shown in Fig. 13, the geometry of prestressed RC beam is created according to the above calculations. To study the inuence of
prestressing level on beam responses, a total of four beams, categorized as Group #1 in this study, with increasing prestress are considered. Table 2 lists the corresponding initial hogging
deformations and initial principal stress in tension and compression, respectively. The prestress distributions are generated by
applying the prestress in the beam in the present approach, it is
obtained by applying bending moment to induce the initial deformation of the beam. Fig. 14 shows the typical initial principal
stress contour in the beam. The maximum initial principal stress

Fig. 12. Schematic diagram of prestressed rebar for theoretical calculation.

Fig. 13. (a) Initial geometry of the original beam; (b) geometry of the RC beam after prestressing (not to scale).

669

W. Chen et al. / Materials and Design 65 (2015) 662674

Fig. 14. Initial principal stress contour of beam #13 after prestressing.

Table 3
Axial stress and strain in longitudinal reinforcement with respect to prestress level (Group #1).
Beam

Prestress

Top rebar (in tension)

No.

MPa

Axial stress (MPa)

Axial strain (%)

Axial stress (MPa)

Axial strain (%)

#11
#12
#13
#14

0
315
472
630

0
18.9
30.3
38.2

0
0.0095
0.0152
0.0191

0
17.2
27.4
34.2

0
0.0086
0.0137
0.0171

in tension and the maximum initial principal stress in compression


under different prestresses are given in Table 2. The maximum
initial principal stress in tension of beam #14 is 4.067 MPa, which
is less than the concrete principal stress of around 4.9 MPa.
Table 3 and Fig. 16 show the axial stress and strain in the
longitudinal reinforcement bars corresponding to the different prestressing levels. As shown, increasing the prestress, both the stress
and strain in the longitudinal steel reinforcement bars in both the
compression and tension zones increase.

Bottom rebar (in compression)

Table 4
Mid-span deection time history (Group #1).
Beam

Prestress

Concrete

Maximum deection

Residual deection

No.

MPa

MPa

mm

Reduction

mm

Reduction

#11
#12
#13
#14

0
315
472
630

42
42
42
42

112.0
111.6
104.0
88.9

0.4%
7.2%
20.6%

86.5
81.0
60.5
50.0

6.4%
30.1%
42.2%

Note: benchmark.

4. Numerical simulations of prestressed RC beam to blast


loading
With the above numerical model of prestressed RC beam, intensive numerical simulations are carried out to investigate the inuences of prestressing levels and the concrete compressive strength
on blast-resistance capacities of prestressed RC beams.
4.1. Effect of prestressing levels
Four RC beams in Group #1 are considered with different
prestressing levels. The mid-span maximum deection and residual deection are extracted from numerical simulation results to

200

Maximum deflection
Residual deflection

Deflection (mm)

160

120

80

40

0
#11

#12

#13

#14

Beam
Fig. 15. Maximum and residual deection of beam with respect to prestress level
(Group #1).

examine the exural capacity. The prestressed RC beam experiences smaller maximum and residual deections, and the response
decreases with the increased prestress level, as shown in Fig. 15.
Taking Beam #11 without prestress as the benchmark, as shown
in Table 4, when the prestress of 630 MPa is applied, the beam
#14 experiences the maximum deection of 88.9 mm, which is
20.6% less than that of the benchmark beam. The residual deection is 50 mm, which is 42.2% lower than that of the benchmark
beam. The mid-span deection time histories are shown in
Fig. 16. These results demonstrate that prestressing the RC beam
is benecial to its blast loading resistance capacity.
As shown in Fig. 17, the crack patterns of beams in Group #1 are
different at t = 1.5 ms after the blast event. The primary damage of
beam #11 without prestress is exural cracks developed at the
bottom side of the mid-span of the beam. The damage patterns
of beam #12 to beam #14 with increasing prestress change from
exural damage dominant to more prominent shear damages.
Large cracks move from the mid span towards the supports. This
is because prestressing the beam increases its exural capacity
and therefore decreases the concrete tensile damage at the midspan. However, it should be noted that prestressing would slightly
increase the diagonal shear stress in the beam because of the compressive prestressing over the cross section, therefore it results in
damage towards the supports. To demonstrate this, the stress on
the longitudinal rebars near the mid-span and the stress of hoop
rebars near the supports of beam #11 and beam #12 are extracted
and compared from the numerical results. The time histories of
stresses of longitudinal bar and hoop rebar are shown in Fig. 18.
As summarized in Table 5, the maximum stresses on the longitudinal rebars are 547.2 MPa and 487.2 MPa for the non-prestressed
and prestressed beam #12, respectively. The maximum stresses
of hoop rebars near the supports are 476.7 MPa and 511.2 MPa

670

W. Chen et al. / Materials and Design 65 (2015) 662674


Table 5
Maximum stresses of longitudinal bar and hoop rebar at supports (#11 and #12).

Displacement (m)

0.05

Beam
No.

0.00

-0.05

#11
#12

Prestress
MPa

Maximum stress of
longitudinal rebar
MPa

Maximum stress of hoop


rebar near the supports
MPa

0
315

547.2
487.2

476.7
511.2

-0.10

-0.15

#11
#13

-0.20
0.00

0.02

0.04

0.06

#12
#14

0.08

0.10

Time (s)
Fig. 16. Mid-span deection time histories (Group #1).

to blast loading resistance capacity if damage is governed primarily


by exural mode. However, prestressing may have adverse effect if
shear damage induced by large amplitude short duration blast load
is the dominant damage mode.
4.2. Effect of concrete compressive strength

for the non-prestressed and prestressed beam #12, respectively.


The results indicate that prestressing in the beam might reduce
the exural damage near the mid-span but increase the shear damage near the beam supports.
When subjected to impulsive loading, most of the RC beams
experience the combined damage mode of exural and shear damage. Besides the exural damage discussed above, the RC beam also
experiences shear damage near the supports as observed in Fig. 17.
To examine the inuences of prestressing on the shear damage,
shear stress time history at a concrete element near the support
(i.e. element 38,889) as shown in Fig. 19 is extracted. As shown
in Fig. 20, increasing the prestress, the shear stress in the element
also increases. These results indicate that prestressing is benecial

To investigate the effects of concrete compressive strength,


three more groups of beams (i.e. Group #2, #3 and #4) having different concrete compressive strength (i.e. 50 MPa, 60 Mpa and
70 MPa) are compared with the beams of Group #1 with concrete
compressive strength of 42 MPa. Four levels of prestress (i.e. 0,
315 MPa, 472 MPa and 630 MPa) are applied for each group. All
groups of beams are subjected to the same blast loadings with peak
pressure of 6100 kPa, positive reected impulse of 3050 kPa ms
and duration of positive phase of 1.0 ms.
As shown in Figs. 21 and 22, the maximum and residual deections at mid-span decrease with the increase of concrete compressive strength. Figs. 2325 show the deection time histories of
beams. As given in Table 6, using the beam #13 as a benchmark,
the maximum deections of the beam #23, beam #33 and beam

600

600

500

500

400

400

Stress (MPa)

Stress (MPa)

Fig. 17. Crack patterns of beam #11, #12 and #14 at t = 1.5 ms.

300
200
100
0
0.00

Beam #12
Beam #11
0.01

0.02

Time (s)

0.03

300
200
100
0
0.00

Beam #12
Beam #11
0.01

0.02

Time (s)

Fig. 18. (L) Stress time histories of longitudinal rebar; (R) stress time histories of hoop rebar at supports.

0.03

671

W. Chen et al. / Materials and Design 65 (2015) 662674

Fig. 19. Schematic diagram of selected concrete element near support.

0.05

#11
#12
#13
#14

15

Displacement (m)

Maximum shear stress (MPa)

20

10

0.00

-0.05

-0.10

-0.15

#21
#23

-0.20
0.00
0
0.00

0.02

0.04

0.06

0.02

0.06

0.08

0.10

Time (s)

0.08

Time (s)

0.04

#22
#24

Fig. 23. Mid-span deection time histories for Group #2.

Fig. 20. Maximum shear stress time histories at supports (Group #1).

#43 decrease 1.9%, 23.5% and 37.3%, their residual deections


decrease 45.4%, 50.6% and 60.0%, respectively, with the increase
of concrete compressive strength from 42 to 50, 60 and 70 MPa.
The beam #44 with prestress of 630 MPa and concrete compressive strength of 70 MPa experiences the maximum deection of
54.0 mm and residual deection of 28.0 mm while the beam #22
with prestress of 315 MPa and concrete compressive strength of
42 MPa has larger maximum deection of 111.6 mm and residual
deection of 81.0 mm. Therefore, it can be concluded that the
increase of concrete compressive strength and prestress has obvious benets to blast-resistant capacity of prestressed RC beams.

160

Maximum deflection-Group#1(fc'=42MPa)
Maximum deflection-Group#2(fc'=50MPa)
Maximum deflection-Group#3(fc'=60MPa)
Maximum deflection-Group#4(fc'=70MPa)

Deflection (mm)

140

120

100

80

4.3. Effect of blast intensity

60
-100

100

200

300

400

500

600

700

Prestress (MPa)
Fig. 21. Maximum deections versus prestress for different concrete strength.

120

Residual deflection-Group#1 (fc'=42MPa)


Residual deflection-Group#2 (fc'=50MPa)
Residual deflection-Group#3 (fc'=60MPa)
Residual deflection-Group#4 (fc'=70MPa)

Deflection (mm)

100

80

60

40

-100

100

200

300

400

500

600

700

Prestress (MPa)
Fig. 22. Residual deections versus prestress for different concrete strength.

The above results indicate that prestressing will increase the


exural resistance capacity of the beam, but may reduce the shear
resistance capacity. It is well understood that when the blast loading duration is longer than the fundamental vibration period of the
beam, the beam response and damage are governed by exural
response mode, whereas they are governed by the shear response
and damage mode if the blast loading duration is substantially
shorter than the beam fundamental vibration period. To further
demonstrate the inuences of prestressing on beams exural
and shear resistance capacities, in this section, the responses of
the non-prestressed beam #11 and the prestressed beam #12
under three different levels of blast loadings are considered. The
prestressed beam #12 has prestress of 315 MPa and concrete compressive strength of 42 MPa. For comparison, the three blast loads
considered have the same impulse of 3050 kPa ms but different
amplitudes and duration. The peak pressures of the three loads
are respectively 6100 kPa, 13,000 kPa and 20,000 kPa, and the corresponding positive phase duration 1.0 ms, 0.47 ms and 0.31 ms.
The blast load parameters are given in Table 7 and the idealized triangular blast loadings are shown in Fig. 26. The simulation cases
are denoted as #11-1 to #11-3 for the unprestressed beam
subjected to the three blast loadings, and #12-1 to #12-3 for the
prestressed beam subjected to the three blast loadings.
The damage modes under the three blast loadings at
time = 4.5 ms are shown in Figs. 2729. The performance of the
beam #11-1 and the prestressed beam #12-1 under the blast load
with peak pressure 6100 kPa have been discussed in Section 4.1. As

672

W. Chen et al. / Materials and Design 65 (2015) 662674


Table 7
Parameters of blast loadings.

0.05

Beam
No.

Displacement (m)

0.00

Prestress
(MPa)

Case

Peak
pressure
(kPa)

Positive
reected
impulse
(kPa ms)

Duration of
positive
phase (ms)

-0.05

#11

#11-1
#11-2
#11-3

6100
13,000
20,000

3050
3050
3050

1.0
0.47
0.31

#12

315

#12-1
#12-2
#12-3

6100
13,000
20,000

3050
3050
3050

1.0
0.47
0.31

-0.10

-0.15

#31
#33
-0.20
0.00

0.02

0.04

0.06

#32
#34

0.08

0.10

Time (s)
Fig. 24. Mid-span deection time histories for Group #3.
25000
0.05

Peak pressure (kPa)

0.00

Displacement (m)

#11-1 #12-1
#11-2 #12-2
#11-3 #12-3

20000

-0.05

-0.10

#41
#43

-0.15

0.02

0.04

0.06

0.08

10000

5000

#42
#44

-0.20
0.00

15000

0
0.0

0.10

0.1

0.2

0.3

Time (s)

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Time (ms)

Fig. 25. Mid-span deection time histories for Group #4.

Fig. 26. Idealized blast loadings of #11-1(2/3) and #12-1(2/3).

shown in Fig. 27, both beams suffer primarily exural damage with
most of the cracks at the bottom side of the mid-span. The beam
#12-1 experiences less damage at the mid-span region than the
beam #11-1 because the prestressing reduces the exural damage
as discussed above. When the beams are subjected to the blast load
of peak pressure 13,000 kPa, the prestressed beam #12-2 experiences the combined pattern of shear and exural damage as shown
in Fig. 28. The damage moves from the mid-span towards the supports where longitudinal rebars fail. The non-prestressed beam
#11-2 has a similar failure mode with more cracks developed in
the mid-span and longitudinal rebars also failed. However, it is

found that the longitudinal rebar of the prestressed beam #12-2


fails at around 13 ms, earlier than the unprestressed beam #11-2,
which fails at 21 ms. When subjected to the largest blast load considered in this study, the prestressed beam #12-3 fails primarily by
diagonal shear failure close to the support as shown in Fig. 29. It is
observed that one longitudinal rebar and some hoop rebars fail
near the supports. The non-prestressed beam #11-3 also experiences similar shear failure with longitudinal rebar and hoop rebars
near the supports failed. The simulations reveal again that the longitudinal rebar of prestressed beam fails at around 10 ms, while
the unprestressed beam fails at 14 ms, implying the prestressed

Table 6
Mid-span deections of beams (Group #1 #4).
Group

Beam

Prestress

Concrete

Maximum Deection

No.

No.

MPa

MPa

mm

Reduction

mm

Reduction

#1
#2
#3
#4
#1
#2
#3
#4
#1
#2
#3
#4
#1
#2
#3
#4

#11
#21
#31
#41
#12
#22
#32
#42
#13
#23
#33
#43
#14
#24
#34
#44

0
0
0
0
315
315
315
315
472
472
472
472
630
630
630
630

42
50
60
70
42
50
60
70
42
50
60
70
42
50
60
70

112.0
111.9
104.5
94.6
111.6
111.5
91.3
66.9
104.0
102.0
79.6
65.3
88.9
76.6
70.0
54.0

0.1%
6.7%
15.5%

0.2%
18.2%
40.0%

1.9%
23.5%
37.3%

13.8%
21.3%
39.2%

86.5
83.6
80.5
62.4
81.0
75.8
69.5
44.6
60.5
56.8
51.4
41.6
50.0
48.1
44.4
28.0

3.4%
6.9%
27.9%

6.4%
14.2%
45.0%

45.4%
50.6%
60.0%

45.9%
50.1%
68.5%

Note: Group #1 as benchmark.

Residual Deection

W. Chen et al. / Materials and Design 65 (2015) 662674

673

#11-1

#12-1
Fig. 27. Damage modes of beams #11-1 and #12-1 (at time = 4.5 ms).

#11-2

#12-2
Fig. 28. Damage modes of beams #11-2 and #12-2 (at time = 4.5 ms).

#11-3

#12-3
Fig. 29. Damage modes of beams #11-3 and #12-3 (at time = 4.5 ms).

beam has weaker blast loading resistance capacity if the beam


failure mode is primarily shear failure. These observations demonstrate that prestressing is not necessarily always benecial to the
blast loading resistance capacities of RC beams. When the beam
failure mode is governed by shear failure, prestressing reduces
the beam blast loading resistance capacity. However, it always
enhances the beam exural resistance capacity.
5. Conclusions
This study presents numerical simulations of simply-supported
prestressed RC beams subjected to blast loadings by using nite
element codes LS-DYNA. Numerical results demonstrated that
prestressing RC beams increases their blast loading resistance
capacities if the failure is governed by exural responses, but
may reduce the beams blast loading resistance capacities if the
failure is governed by shear responses. Parametric studies on four
groups of prestressed RC beams have been carried out by considering different prestress, concrete compressive strength and blast
loading. It is found that the appearance and growth of exural
cracks in concrete is delayed effectively by using prestress. In
addition, the higher are the prestress and the concrete compressive
strength, the more effective is the prestressed RC beam in blastresistant capacity. However, increasing the prestress level in the
beam may increase the diagonal shear damage near the beam supports, the blast resistance performance of prestressed beam might
be disadvantaged. Therefore, proper analysis and design are

necessary to determine the optimal prestressing levels to enhance


the blast loading capacities of RC beams.

Acknowledgements
The authors acknowledge the partial nancial support from
Australian Commonwealth Scientic and Industrial Research Organization (CSIRO) to carry out this research work.

References
[1] Ellingwood BR. Mitigating risk from abnormal loads and progressive collapse. J
Perform Constr Facil 2006;20:31523.
[2] Collins MP, Mitchell D. Prestressed concrete structures. NJ: Hall Englewood
Cliffs; 1991.
[3] AS/NZS. Concrete structures. Standard Australia & Standards New Zealand;
2009.
[4] Eurocode 2: Design of concrete structures concrete bridges: design and
detailing rules. 2005.
[5] Fang Q, Wu P. Main factors affecting failure modes of blast loaded RC beams.
Chin J Comput Mech 2003;1:009.
[6] Lin X, Zhang Y, Hazell PJ. Modelling the response of reinforced concrete panels
under blast loading. Mater Des 2014;56:6208.
[7] Magnusson J, Hallgren M. High performance concrete beams subjected to
shock waves from air blast. Swedish Defence Research Agency; 2003.
[8] Jiang H, Wang X, He S. Numerical simulation of impact tests on reinforced
concrete beams. Mater Des 2012;39:11120.
[9] Fang Q, Qian Q-H, Shi Y-L. A rate-sensitive analysis of R/C beams subjected to
blast loads. In: International conference on structures under shock and impact;
1996. p. 22130.

674

W. Chen et al. / Materials and Design 65 (2015) 662674

[10] Yi N-H, Kim J-HJ, Han T-S, Cho Y-G, Lee JH. Blast-resistant characteristics of
ultra-high strength concrete and reactive powder concrete. Constr Build Mater
2012;28:694707.
[11] Li Y, Wang X, Guo X. Experimental study on anti-impact properties of a
partially prestressed concrete beam. Explos Shock Waves 2006;26:256.
[12] Ishikawa N, Enrin H, Katsuki S, Ohta T. Dynamic behavior of prestressed
concrete beams under rapid speed loading. Struct Under Shock Impact V
1998:71726.
[13] Ishikawa N, Katsuki S, Takemoto K. Incremental impact test and simulation of
prestressed concrete beam. Struct Mater 2002;11:48998.
[14] Iskhakov I, Ribakov Y. Two-layer pre-stressed beams consisting of normal and
steel bered high strength concrete. Mater Des 2008;29:161622.
[15] Cramsey N, Naito C. Analytical assessment of blast resistance of precast,
prestressed concrete components, AFRL-ML-TY-2007-4529. Air Force Research
Laboratory; 2007.
[16] Ngo T, Mendis P, Krauthammer T. Behavior of ultrahigh-strength prestressed
concrete panels subjected to blast loading. J Struct Eng 2007;133:158290.
[17] Chen L, Fang Q, Liu J-c, Zhang Y-d, Xiang H-b. Nonlinear analysis of blast
performance of partially prestressed RC beams. Int J Protective Struct
2011;2:295314.
[18] Cofer WF, Matthews DS, McLean DI. Effects of blast loading on prestressed
girder bridges. Shock Vib 2012;19:118.
[19] LSTC, LS-DYNA version 971 keyword users manual_Rev5-beta. Livermore
Software Technology Corporation; 2010.
[20] Baylot JT, Bevins TL. Effect of responding and failing structural components on
the airblast pressures and loads on and inside of the structure. Comput Struct
2007;85:891910.

[21] Woodson SC, Baylot JT. Structural collapse: quarter-scale model


experiments. DTIC Document; 1999.
[22] Shi Y, Hao H, Li Z-X. Numerical derivation of pressureimpulse diagrams for
prediction of RC column damage to blast loads. Int J Impact Eng
2008;35:121327.
[23] Mutalib AA, Hao H. Development of PI diagrams for FRP strengthened RC
columns. Int J Impact Eng 2010.
[24] Malvar LJ, Crawford JE, Wesevich JW, Simons D. A plasticity concrete material
model for DYNA3D. Int J Impact Eng 1997;19:84773.
[25] Li J, Hao H. Inuence of brittle shear damage on accuracy of the two-step
method in prediction of structural response to blast loads. Int J Impact Eng
2013;54:21731.
[26] Beton CE-Id. Concrete structures under impact and impulsive loading. CEB Bull
1990;187.
[27] Malvar LJ, Ross CA. Review of strain rate effects for concrete in tension. Aci
Mater J 1998:95.
[28] Malvar LJ. Review of static and dynamic properties of steel reinforcing bars. Aci
Mater J 1998:95.
[29] Chen W, Hao H. Numerical study of blast-resistant sandwich panels with
rotational friction dampers. Int J Struct Stab Dyn 2013:13.
[30] Chen W, Hao H. Numerical study of a new multi-arch double-layered blastresistance door panel. Int J Impact Eng 2012;43:1628.
[31] US_Department_of_Defense. Structures to resist the effects of accidental
explosions (UFC 3-340-02). Washington, DC: Unied Facilities Criteria; 2008.
[32] Bi K, Hao H. Numerical simulation of pounding damage to bridge structures
under spatially varying ground motions. Eng Struct 2013;46:6276.