Construction and Building Materials 188 (2018) 656–675

Contents lists available at ScienceDirect

Construction and Building Materials

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

Study of induced prestress on deformation and energy absorption

characteristics of concrete slabs under drop impact loading
Vimal Kumar a, M.A. Iqbal a,⇑, A.K. Mittal b
a
Department of Civil Engineering, Indian Institute of Technology Roorkee, Roorkee 247667, India
b
Structural Engineering Group, CSIR-Central Building Research Institute, Roorkee 247667, India

h i g h l i g h t s g r a p h i c a l a b s t r a c t

Prestressed and reinforced concrete

slabs studied under drop impact.

Prestressed concrete offered better
resistance due to improved stiffness.

Punching cracks were observed with
the increase in drop height in
reinforced concrete slabs.

Numerical results accurately
reproduced the impact and reaction
time response of slabs.

Numerical simulations also enabled
to develop insight into the mechanics
of reinforcement and prestressing
wires.

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

Article history: The drop impact resistance of concrete slabs with induced initial stress was investigated through detailed
Received 11 February 2018 experimentation and the results thus obtained were reproduced through the numerical simulations per-
Received in revised form 3 June 2018 formed on ABAQUS/Explicit code. Concrete slabs of span 800 mm and thickness 100 mm pre-tensioned to
Accepted 17 August 2018
10% and 20% of the characteristic compressive strength of concrete (40 N/mm2) were subjected to impact
by a freely falling hammer (243 kg) from predefined heights. The impact force, reaction-time response
and deflection of these slabs were measured. Influence of induced prestress was studied on the damage
Keywords:
resistance and energy absorption characteristics of prestressed concrete slabs and the results were com-
Prestressed concrete
Reinforced concrete
pared with the equivalent reinforced concrete slabs. The induced initial stress has been found to enhance
Drop impact the stiffness and thus reduced the resultant deflection of slabs but increased the peak impact and reaction
Energy absorption forces. The reinforced concrete slabs experienced flexural cracks leading to punching of concrete.
Damage mechanism Prestressed concrete slabs developed low density cracks without any visible punching action.
Finite element analysis However, splitting was observed with increase in prestressing force. Numerical results predicted the peak
impact and reaction forces within 15% deviation and reproduced the damage with reasonable accuracy.
Simulations also described the mechanics of prestressing wires and the reinforcement.
Ó 2018 Elsevier Ltd. All rights reserved.

1. Introduction

The prestressing of concrete elements is primarily carried out to

reduce crack width and thickness in strategic, pressure retaining
and large span structures such as nuclear containment, rock shel-
⇑ Corresponding author. ter, long span bridge, community hall and industrial floor. Such
E-mail address: [email protected] (M.A. Iqbal).

https://doi.org/10.1016/j.conbuildmat.2018.08.113
0950-0618/Ó 2018 Elsevier Ltd. All rights reserved.
 V. Kumar et al. / Construction and Building Materials 188 (2018) 656–675 657

prestressed concrete members are often subjected to impact loads There is limited understanding of the behaviour of prestressed
generated due to internal plant accidents in nuclear power plants, concrete under impact loading despite significant utility of the
automobile accidents on bridges, toppling of rocks over rock shel- material in various applications. Authors could not find any study
ters and falling of equipment on industrial floors [1]. These struc- in the available literature on the drop impact response of pre-
tures are generally designed for static loading conditions and the stressed concrete slabs.
uncertainty in the loading is accounted for by considering a certain The present study attempts to investigate the response of pre-
constant safety/load factor. The design based on the assumed con- stressed concrete (PC) slabs against a 243 kg steel hammer falling
stant factor of safety although enhances the dynamic response to a freely from 500 and 1000 mm height. The square slabs of span
certain extent, however, it is incapable of addressing the localized 800 mm  800 mm, induced with an initial stress of approximately
deformation and the actual structural behaviour under impact 10% and 20% of characteristic compressive strength (40 N/mm2)
loads [2]. were subjected to impact at the center. The dynamic responses such
Structural elements of plain, reinforced and fiber reinforced as impact force, support reactions and resultant displacement have
concrete have been studied under impact loading to investigate been studied and compared with the equivalent non-prestressed
the energy absorption capacity and the characteristics of the concrete slabs. The damage mechanisms observed in the pre-
induced damage [3–6]. The performance of concrete filled steel stressed and non-prestressed concrete have also been compared
tubes has also been explored under impact loading [7–9]. The con- and discussed. Numerical simulations of the actual problem have
crete filling of steel tubes has demonstrated in precluding local also been performed on ABAQUS/Explicit finite element code for a
deformation under impact loading and increasing the moment car- detailed investigation of induced prestressing on the response of
rying capacity of a given section. The presence of steel surface slabs and to further explore the damage mechanics of prestressed
around the concrete core has been found to increase the lateral and reinforced concrete slabs. A simplified analytical model based
strength of the column and thus enhance its load carrying capacity. on the principle of mechanics has been introduced for calculating
In another experimental study [10], the reinforced concrete slabs the peak displacement in the slabs and it has thus facilitated a
subjected to impact by freely falling steel ball described little influ- direct comparison with the actual and computational results.
ence of the percentage of reinforcement on the resultant dynamic
deflection. Othman and Marzouk [11], on the other hand, reported
2. Concrete mix design
an inverse trend on peak displacement, residual displacement and
scabbing of slabs due to an increase in the reinforcement. The
The concrete mix was designed as per the requirement of Indian
deflection at the center of the slab has been seen to have reduced
Standard IS 456 (2000) and IS 10262 (2009). The Ordinary Portland
due to increase in its thickness [12]. Thicker slabs with relatively
Cement (OPC) 43 grade, the coarse aggregate of nominal size
higher stiffness have experienced lesser deflection and relatively
10 mm and river sand were used for the preparation of the con-
lower damage at front and back surfaces under impulsive loads.
crete mix. The aggregate passing from 10 mm and 4.75 mm sieve
The ballistic performance of the double layered target with
was considered for the mix design, respectively, as coarse and fine
front reinforced concrete and the backing steel plate has been
aggregate. The quantity of required material was calculated using
studied against non-deformable steel projectiles [13]. The resul-
the weigh-batching method of concrete mix proportioning. The
tant magnitude of damage in the reinforced concrete target
ratio of the cement, fine and coarse aggregate was considered to
obtained experimentally at different incidence velocities was com-
be 1:1.74:1.68, respectively, with water-cement ratio limited to
pared with the probability of failure of the structure computed
0.35 for obtaining the characteristic compressive strength of
based on Monte-Carlo approach, and a close correlation between
40 N/mm2 (IS 456; 2000). A small quantity of water-reducing
the experimental and probabilistic assessment was found. In
agent, 0.4%, was also added to the concrete mix. The 28 days aver-
another study [14], the double layer reinforced concrete-steel con-
age target strength of the tested concrete cube (150 mm) samples
tainment was found safe against the hard steel missile impact if
at loading rate approximately 140 kg/sq.cm/min (IS 516:1959) was
the missile velocity was lesser than 65% of the containment outer
found to be 48 N/mm2 with a standard deviation of 1.60 N/mm2.
(reinforced concrete) wall’s ballistic limit. However, if the impact
velocity was more than 90% of the outer wall’s ballistic limit, the
probability of failure of double wall containment was very high. 3. Preparation of reinforced and prestressed concrete slabs
The available studies on prestressed concrete elements under
impact loading are limited to the testing and design of railway The reinforced and prestressed concrete slabs of identical span
sleepers [2,15]. The cracks in the sleepers have transformed from (800 mm  800 mm) and thickness (100 mm) were casted with
flexure to combined flexure-shear and finally into the shear mode the help of specific moulds designed for each concrete. Each rein-
of failure due to increase in the loading rate. The rate of crack prop- forced concrete (RC) slab was casted in separate mild-steel moulds
agation in a sleeper was governed by the stiffness of the track sys- with internal dimension, 800 mm  800 mm. High strength (Fe
tem such that hard track had relatively faster crack propagation 500) deformed reinforcing bars of / 8 mm @ 140 mm c/c were pro-
and experienced longer cracks [16]. In another study, the model vided across both spans with a clear cover of 10 mm from the bot-
and steady-state analysis of prestressed concrete beams were per- tom of slab, see Fig. 1. The reinforcing bars were tested under
formed using ABAQUS finite element code [17]. The natural fre- tension on a Universal Testing Machine (UTM) and their average
quency of beam has been found to have reduced due to the yield and ultimate strengths at loading rate 1 mm s1 were found
provision of an opening at the shear zone. Jing et al. [18] investi- to be 609 and 745 N/mm2, respectively, see Fig. 2.
gated the influence of lateral vehicle impact on the behaviour of The procedure of casting the prestressed concrete slabs was
full-scale prestressed concrete bridge girder and observed localized almost same to that of the reinforced concrete slabs except that
failure at the impact location due to crushing and punching of con- the initial pretension was introduced in these slabs with the help
crete. The ballistic resistance of plain, reinforced and prestressed of high-strength indented steel wires of diameter /4 mm (Fig. 3
concrete plates studied by carrying out the experiments and finite and Fig. 4) used in accordance with IS 6003 (2010). The yield and
element simulations described best performance of prestressed ultimate tensile strength of these wires tested on Olsen H75KS uni-
concrete due to globalization in the induced damage as a result versal testing machine were approximately 1600 and 1770 N/mm2,
of initial prestressing of concrete [19,20]. respectively, at the strain rate 1.2  105 s1. The stress-strain
658 V. Kumar et al. / Construction and Building Materials 188 (2018) 656–675

Fig. 1. Reinforcement detail of reinforced concrete (RC) slab.

2000 provided in the slab with an eccentricity of 25 mm from the center

of slab thickness for inducing 10% prestress, see Fig. 4(a). For induc-
1750 ing 20% prestress, however, 19 wires @ 40 mm c/c spacing were
Reinforcement bar provided in the slab at an eccentricity of 25 mm. However, since
1500 Prestressing wire the required level of prestress could not be achieved with 19 wires,
Stress (MPa)

and there was no further space available for increasing the number
1250
of wires with 25 mm eccentricity, 2 additional wires were provided
1000 at the center of slab thickness, see Fig. 4(b). These wires were
stretched with the help of hydraulic jack of 50 Ton capacity
750 through the holes (/8 mm) provided in prestressing bed at
40 mm c/c spacing, see Fig. 3(a). The stressed wires were held in
500 position with the help of high strength steel anchors, see Fig. 3
(a) and (b). The average elongation in the prestressing wires at
250 the time of pre-tensioning was 28.75 and 27.80 mm for 10%
(level-1) and 20% (level-2) pre-stress, respectively. The total pre-
0
0.00 0.03 0.06 0.09 0.12 0.15 stressing losses were assumed to be 20% of that of the magnitude
of initial pre-tension. The available studies on prestressed concrete
Strain
have considered the losses in pre-tensioning of concrete to be 15%
Fig. 2. Stress-strain curve for reinforcement bar and prestressing wire. of the magnitude of initial prestress [21,22]. Lin and Burns [23] on
the other hand described that the summation of the losses due to
elastic shortening and bending of concrete, creep, shrinkage and
relationship for prestressing wire has been presented in Fig. 2. The steel relaxation be considered to be 25 and 20%, respectively, for
initial slope of the stress-strain curve of prestressing wire has some pre-tensioned and post-tensioned concrete. The provision of rein-
small deviation at the stress of about 300 N/mm2. The resultant forcement in prestressed concrete slabs was identical to that of
deviation in the slope could be due to small adjustment of the jaws the reinforced concrete slabs, see Fig. 4(a) and (b). The concrete
of the tensile testing machine. was subsequently casted, compacted through needle vibrator
A specially designed robust mild-steel prestressing bed was (dia. 25 mm) and cured through wet gunny bags. The prestressing
employed for casting multiple prestressed concrete specimens, wires were cut after 14 days of casting, Fig. 3(c), but the curing was
see Fig. 3(a). A total number of 10 wires @ 80 mm c/c spacing were continued for 28 days.
 V. Kumar et al. / Construction and Building Materials 188 (2018) 656–675 659

Fig. 3. Prestressing of slabs showing (a) slab moulds in prestressing bed, (b) anchoring devices, and (c) typical prestressed concrete slab.

Fig. 4. Reinforcement detail and arrangement of prestressing wires in prestressed concrete slabs.

4. Experimental procedure considered to be 500 and 1000 mm, respectively. Similarly, the suf-
fix ‘1’ and ‘2’ with term PC is used to indicate the prestressing
All the PC and RC slabs were favoured identical geometrical and level-1 (initial stress 10%) and level-2 (initial stress 20%) respec-
material characteristic besides the initial stresses induced in the tively, see column ‘c’ of Table 1. The free fall velocity (V) of drop
p
concrete by pre-tensioning. A total twelve number of such slabs mass (M) under gravitational force was obtained as V ¼ ð2gHÞ
2
were casted and tested under drop weight impact, as explained where, g is the gravitational acceleration, 9.81 m s . However,
in Table 1. The term H1 and H2 in column ‘c’ of Table 1 are the actual velocity of the hammer, measured with the laser-
representing the drop height (H) of the impactor that has been velocity-meter, was approximately 98% of the free fall velocity,
660 V. Kumar et al. / Construction and Building Materials 188 (2018) 656–675

Table 1
Detail of drop impact testing.

Type of Specimen Prestressing Specimen ID Induced Initial Stresses Drop Height, Drop Mass, Impact Velocity, Experiment
Levels (% of concrete strength) H (mm) M (kg) 0.98 V (m s1) No.
a b c d e f g h
Reinforced Concrete Slabs – RC-H1 – 500 242.85 3.069 1
RC-H1 500 242.85 3.069 2
RC-H2 1000 242.85 4.341 3
RC-H2 1000 242.85 4.341 4
Prestressed Concrete Slabs Level 1 PC1-H1 10 500 242.85 3.069 5
PC1-H1 500 242.85 3.069 6
PC1-H2 1000 242.85 4.341 7
PC1-H2 1000 242.85 4.341 8
Level 2 PC2-H1 20 500 242.85 3.069 9
PC2-H1 500 242.85 3.069 10
PC2-H2 1000 242.85 4.341 11
PC2-H2 1000 242.85 4.341 12

see column ‘g’ of Table 1. The laser-velocity-meter consisted of two the stiffness degradation, stiffness recovery and shear reduction.
sets of signal emitter and signal receiver with a fixed spacing The tensile damage occurring in the concrete is also not considered
between each set. In order to measure the velocity, the emitter the primary damage parameter in the model due to the fact that in
and receiver were fixed on opposite sides of the falling hammer the impact problems the compressive stress waves reduce the
in front of each other, just above the impact point. When the ham- cracking caused by the tensile stresses by closing these cracks.
mer comes in between the emitter and receiver, the signal was The theoretical formulation of the model is relatively simple and
obstructed. The obstruction in the signal was recorded as a sudden it takes into account most of the important issues of concrete beha-
drop in the signal-time response. Knowing the distance between viour such as pressure dependency, strain-rate effect, damage and
each set of emitter and receiver, and the time duration of the material failure [19,25]. The normalized equivalent stress, r ; of
obstruction of the signals, the actual velocity was calculated. The the model is defined as a function of pressure, P, and strain-rate,
average of the 10 velocities was calculated (before starting the e ; see Eq. (1).
experiment) to be 98% of the free fall velocity. The reduction in h i
the impact velocity occurred due to some friction between the r ¼ Að1  DÞ þ BPN ½1 þ C ln e  ð1Þ
guide and the hammer.
Where r =r/f c ; r is actual equivalent stress (von-Mises); f c ; is uni-
All the four edges of the target slab were encased between thin 0 0

steel angles (5 mm thickness) to avoid the possibility of damage to axial compressive strength of concrete; A, is the cohesive strength;
the underlying expensive sensors owing to any accidental/sudden D, is the damage parameter (0  D  1); B, is the normalized pres-
failure of slab, see Fig. 5(a) and (b). After encasing between the 0
sure hardening coefficient; P = P/f c ; is normalized actual pressure;
steel angles, two adjacent corners of the slab were rested over N, is the pressure hardening exponent; C, is the strain-rate sensitiv-
the load-cells (each 150 Ton capacity) and the two remaining cor- ity coefficient; e ¼ eeq =e0 ; is the normalized dimensionless strain-
ners were supported over the steel pedestals. It should be noted
rate; eeq ; is the equivalent strain-rate and e0 ; is the reference
that the steel angles were not stiff enough to provide fixed bound-
strain-rate.
ary condition at the edge. The angles experienced bending and ver-
The damage in the HJC material model accumulated in a similar
tical deformation which could not be neglected. However, they
manner to that of the Johnson-Cook (JC) fracture model [26]. The
provided some stiffness to the edges of the slab. The angles were
damage parameter, D, is based on scaler damage formulation gov-
straightened after each experiment and reused for the next exper-
erned by both equivalent plastic strain increment, Dep ; and plastic
iment. A 242.85 kg hammer consisting of three steel blocks
attached to an impactor load-cell of 50 Ton capacity (Fig. 5(a)) volumetric strain increment, Dlp . The damage evolution of HJC
was lifted by an electric hoist and released from the specified model is expressed in Eq. (2).
height. The hammer was guided by two rigid steel posts to hit X Dep þ Dlp
the target slab at the center (Fig. 5(a)). The load-cells were con- DD ¼ ð2Þ
nected to a data-logger for recording the impact force and reaction epf þ lpf
response at an excitation of 10 Volts. The dynamic displacement of
the slab at the center and quarter spans was measured with help of epf þ lpf ¼ D1ðP þ T  ÞD2 ð3Þ
two Laser Displacement Sensors (LDS) connected to the same data-
acquisition system. All the analog data obtained from various sen- Where, epf þ lpf is plastic strain to fracture under a constant pres-
sors were recorded at a uniform sampling rate of 10 k s1 in the sure, P. D1 and D2 are material constants also known as damage
same data acquisition system. The post-impact damage in target 0
parameters. T  = T/f c is the normalized maximum tensile hydro-
slabs was captured using a digital photography camera.
static pressure and T is actual maximum tensile hydrostatic pres-
sure the concrete can experience prior to tensile failure. The
5. Constitutive modelling damage due to plastic volumetric strain is incorporated in Eqs. (2)
and (3), however, the possibility of the damage induced by equiva-
The constitutive modelling of concrete under drop weight lent plastic strain is relatively high [24,25]. An element is assumed
impact was simulated numerically by incorporating the to be completely degraded once the stress reaches the above limit-
Holmquist-Johnson-Cook (HJC) material model [24]. The HJC mate- ing criteria at an integration point. The corresponding element is
rial model is an elasto-viscoplastic material model coupled with then assumed to have lost its load carrying capacity and no longer
isotropic damage [25]. The HJC model is quite attractive to simu- takes part in to stiffness formulation. These elements are then
late the impact related problems despite its limitations to address removed from the analysis for further stiffness formulation [19].
 V. Kumar et al. / Construction and Building Materials 188 (2018) 656–675 661

Fig. 5. Drop weight impact setup showing (a) experimental setup and (b) clamping arrangement of slab (half-model).

The material parameters such as density of concrete, compres- deformation of the reinforcement and the prestressing wires. The
0
sive strength (f c ) and tensile strength (T) were evaluated by per- yield surface is well defined by the uniaxial yield stress as a func-
forming the laboratory tests on the material. All the other tion of plastic strain, temperature and field variables. The strength
material parameters for 48 MPa concrete were taken from the for both the materials was defined as a function of accumulated
study of Holmquist-Johnson-Cook [24] due to limitations of the plastic strain. The strain-rate and temperature effects have not
available testing facilities. The HJC material parameters considered been taken into account for modelling the behaviour of the steel
in the present study are enlisted in Table 2. reinforcement and the prestressing wires, due to non-availability
The constitutive modelling of the reinforcing bars and the pre- of the material parameters at high strain rate and high
stressing wires was carried out by employing the metal plasticity temperature.
model in conjunction with isotropic-linear-elasticity in ABAQUS
finite element code. The isotropic hardening behaviour was incor- 6. Numerical procedure
porated to define the material behaviour once it reaches the elastic
limit. The metal plasticity model uses von-Mises yield surface with The objective of the finite element simulations was to explore
the plastic flow to define isotropic yielding and allows the plastic numerically the mechanics of transfer and development of pre-
stress in concrete and its resultant influence on the induced dam-
age under drop impact loading. The simulation of the problem was
carried out using ABAQUS/Explicit finite element code in two
Table 2
HJC material parameters for 48 MPa concrete. steps; in the first step the transfer of prestress in the concrete
was analyzed by introducing initial tension in the wires while in
Material Parameters Values Material Parameters Values
the second step, the simulation for drop impact was performed.
q (kg/m3) 2430 D1 0.04
A 0.79 D2 1
B 1.6 Pcrush (GPa) 0.016 6.1. Modelling of geometry
N 0.61 Plock (GPa) 0.8
C 0.007 mcrush 0.001 In the present impact problem, only half of the geometry was
0
f c (MPa) 48 mlock 0.1
modelled taking advantage of symmetry. The concrete slab was
SMAX 7.0 K1 (GPa) 85
modelled as three-dimensional deformable continuum and the
G (MPa) 14,803 K2 (GPa) 171
T (MPa) 4.80 K3 (GPa) 208
reinforcement and prestressing wires as three-dimensional
deformable wire. For concrete slab, three-dimensional reduced
q is density; A is cohesive strength; B is pressure hardening coefficient; 0
N is pres- integration deformable 8-node brick and 6-node wedge shape solid
sure hardening exponent; C is strain-rate sensitivity coefficient; f c is uniaxial
compressive strength; SMAX is normalized maximum strength of concrete; G is
elements were used. On the other hand, the reinforcement bars
shear modulus; T is max. tensile hydrostatic pressure; D1 and D2 are damage and prestressing wires were modelled using three-dimensional
parameters; Pcrush, Plock, mcrush, mlock, K1, K2 and K3 are pressure constants. deformable 2-node truss elements. For all the types of selected
662 V. Kumar et al. / Construction and Building Materials 188 (2018) 656–675

elements, a linear order formulation was considered for determi-

nation of stiffness matrix. A uniform mesh of size 1 mm was used
for truss elements, however, the element size for the continuum
elements (for concrete) was varied from 1.5 mm at the center to
10 mm close to periphery. The drop mass impactor was modelled
as discrete rigid body since it was made up of tempered high-
strength steel and no deformation of the impactor was noticed
during experimentation. The surface to surface contact was defined
between the hammer and the concrete slab considering the ham-
mer as the master and the target slab as the slave surface. The con-
tact of the reinforcement bars and prestressing wires with that of
the concrete slab was modelled as the perfect bond by embedding
truss elements in the continuum elements. The deformation
response of the truss elements was therefore governed by the
deformation of the continuum elements. This is due to the fact that
the translational degrees of freedom (DOF) of the truss element
were eliminated by implying perfect bonding. Therefore, the trans-
lation DOFs of the truss elements were governed by some interpo- Fig. 6. Mesh sensitivity analysis results.
lated values of the continuum elements. The rotational and
translational degrees of freedom were arrested at the corner zone
of size 15 mm  15 mm from the edges of the slab. The elements at
the entire thickness of the slab in this zone were fixed with respect total number of elements in the slab increased to approximately
to translation and rotation. The dimensions of the corner zone 662000. The numerical results thus converged corresponding to
(15 mm  15 mm) were identified based on the several trials and the mesh configuration with the element size 1.5 mm 
the comparison of the numerical results with the corresponding 1.5 mm  1.5 mm in the central region and 10 mm 
experiments. The identified corner zone provided an appropriate 10 mm  10 mm at the periphery of the slab. The finite element
and true representation of the experimental conditions. The impac- model of the slab and the impactor is shown in Fig. 7.
tor was assigned an initial impact velocity at the centroid of its
mass to hit the slab at the center. 6.3. Numerical modelling of prestress

6.2. Mesh-sensitivity analysis The initial stress in the concrete was incorporated by perform-
ing a pre-impact simulation in ABAQUS/Standard. At first, the pre-
One main source of the erroneous result in the numerical sim- stressing wires embedded in the concrete were stretched up to the
ulation is the discretization error and improper element formula- required level of pre-tension then the quasi-static analysis was
tion. The primary purpose of discretization is to present a fully performed in ABAQUS finite element code in order to transfer the
continuous field into a piecewise continuous field. The size of such prestress in the concrete. The behaviour of concrete was modelled
a small piece (element) is very important such that it influences using the HJC constitutive model and the behaviour of reinforce-
the accuracy of the result and the time required for the simulation. ment and prestressing wires using the metal plasticity model for
The excessively fine mesh could give relatively accurate results but the numerical simulation for inducing prestress in concrete, see
may require very large simulation time which is undesirable. In Section 5. The resultant induced stress in the body of concrete cor-
order to obtain an optimum size of element in view of achieving responding to 10% and 20% prestress is plotted with the help of
the accuracy of the solution within reasonable computational time, contours in Fig. 8(a) and (d), respectively. The contour plots show
the mesh sensitivity analysis was performed. The size of contin- that the stress induced numerically in the concrete specimen with
uum elements in the impact region was gradually reduced from 10% prestress is 4 N/mm2 at the central cross-section (Section B-B)
10 mm  10 mm  10 mm to 1.5 mm  1.5 mm  1.5 mm keeping and its magnitude is almost same across majority of target span,
the size of the truss elements unchanged. The corresponding num- see Fig. 8(a). Away from the central region and at the bottom sur-
ber of the continuum elements in the slab therefore increased from face however the induced stress varied between 0 and 7.9 N/mm2.
approximately 52,500 to 662,000. The influence of the reduction in In the specimen with 20% induced stress, the stress developed
the mesh size was observed on the predicted peak impact force and numerically has been found to be approximately 8.4 N/mm2 that
the results were compared with the actual impact force. Initially, is very close to the required magnitude of prestress, Fig. 8(d).
the size of the element was reduced from The contour plots show however that the magnitude of the devel-
10 mm  10 mm  10 mm to 3 mm  3 mm  3 mm and the cor- oped stress varies in some region from 4 to 12 N/mm2. The stresses
responding total number of elements in the slab model increased in the prestressing wires after the quasi-static analysis have been
from 52,500 to 144,000 (see Fig. 6). This exercise eliminated the plotted in Fig. 8(c) and (f), respectively, for the concrete with 10%
computational error up to a large extent and reduced the differ- and 20% induced stress. The resultant stress in the wires has been
ence between the actual and computed impact force from 25% to found to be 1114 and 1115 N/mm2, respectively. The magnitude of
0.5%, Fig. 6. Although, the impact force was correctly predicted cor- stress in the wires corresponding to 10% and 20% induced stress
responding to the element size 3 mm  3 mm  3 mm, however, has been found to be almost same due to the fact that the initial
unrealistic punching of concrete was noticed in the computed tension given to the wire was same (14 kN) in both the cases. Nev-
damage pattern. The simulation results described that the hammer ertheless, the stress in concrete developed as per the required mag-
has penetrated the concrete upto more than half of the slab nitude, (10% and 20% of the characteristic strength) by varying the
thickness. However, the experiments did not witness any such number of wires from 10 to 21. As a result of induction of prestress
penetration of hammer in the concrete. Therefore, the size of in the concrete, the reinforcement provided in the concrete also
the continuum elements in the impact region was further reduced developed some nominal stresses with a maximum magnitude of
from 3 mm  3 mm  3 mm to 1.5 mm  1.5 mm  1.5 mm in 15.2 and 27.9 N/mm2, respectively, for the specimen with 10%
order to eliminate the problem of punching. The corresponding and 20% induced stress, see Fig. 8(b) and (e), respectively. The
 V. Kumar et al. / Construction and Building Materials 188 (2018) 656–675 663

Fig. 7. Meshing details in (a) the impactor, (b) the prestressing wires and the reinforcement bars in slab PC1, and (c) the reinforced concrete slab.

stress thus developed numerically in the concrete has been 7. Analytical study for obtaining deflection
compared with the experimentally induced stress for both the
levels of prestress (Fig. 9) and a close correlation between the A simplified method for estimating the peak displacement of
actual and predicted prestress has been found. the target slabs has been derived assuming mass, momentum
664 V. Kumar et al. / Construction and Building Materials 188 (2018) 656–675

Fig. 8. Induced initial stress in the prestressed concrete slab for 10% and 20% prestress.

and energy conservation. The mass of the slab during experi- kinetic energy. However, due to the development of friction
ments remained constant since there was neither visible spalling between the hammer and the guide rails, the actual impact veloc-
nor scabbing of concrete. The momentum conservation has been ity of the hammer would be lesser than the free fall velocity. The
applied for obtaining equivalent static force (Peq-s) from the ratio of the actual to free fall velocity is denoted by g. It should also
impact force versus time response, p(t), Fig. 10. According to be noted that due to the presence of friction, the effective acceler-
Newton’s second law of motion, the transfer of momentum ation of the hammer would also be lesser than that of the gravita-
R
(DMV) in the slab would be equal to the impulse, pðtÞdt. Thus, tional acceleration (g). When the hammer strikes the target slab,
the impulse under drop weight impact was converted into there will be a transfer of momentum and energy. Thus, both
equivalent static force by taking the average of the area under energy and momentum of the hammer will decrease, on the other
the curve, p(t). hand, the energy and momentum of the target slab would increase.
When the hammer (of mass M) is dropped from a given height The energy absorbed (Eab ) by the target slabs under drop weight
(H), the corresponding potential energy will be converted to the impact was calculated using Eq. (4) [27,28].
 V. Kumar et al. / Construction and Building Materials 188 (2018) 656–675 665

10 finite element code. A simplistic analytical study has also been car-
Pre-stress induced (MPa)
Experimental
ried out for calculating the displacement and the energy absorbed
8 Numerical by the slab specimens.

6 8.1. Impact force-time history

4
The actual and predicted impact force versus time history
curves attained a sharp peak load initially due to inertia of slab,
2
Fig. 11. However, the force suddenly dropped due to localized
0 deformation at the contact surface of the slab. Subsequent nominal
0% 5% 10% 15% 20% 25% increase in the impact force is the resultant of the resistance
Prestressing Level offered by the reinforcing bars. The duration of impact phe-
nomenon was observed to be 12.9–15 ms for the drop height
Fig. 9. Stresses induced in the concrete slab for 10 and 20% prestress. 500 mm, and 13.7–18.4 ms for the drop height 1000 mm, Table 3.
The contact time for prestressed concrete slabs has been found to
be relatively smaller than the reinforced concrete slabs. The pre-
dicted duration of the phenomenon of the slabs has been found
to have slightly reduced primarily due to simplicity incorporated
by modelling the fixed boundary at the slab corners and omitting
the modelling of supporting frame structure.
For drop height 500 mm, the magnitude of peak impact force
against level-2 prestressed concrete slabs has increased by 11%
while against level-1 prestressed concrete slabs by 12.2% in com-
parison to the reinforced concrete slabs, see Table 3. For drop
height 1000 mm, the magnitude of peak impact force although
increased in the prestressed concrete slab, however, the rate of
increase was found to be relatively lesser in comparison to the
drop height 500 mm.
Also, for a given concrete, the peak impact force increased with
increase in the drop height. The increase was noticed to be 37% in
the reinforced and 29% in the prestressed concrete slabs (level-1
and level-2). Therefore, the reinforced concrete slabs were found
to be relatively more loading rate sensitive. The numerical results
Fig. 10. Idealization of dynamic impact force p(t) into equivalent static force, Peq-s.
predicted the peak impact force with 10 and 15% deviation for drop
height 500 and 1000 mm, respectively.

8.2. Reaction-time response of slabs

 pffiffiffiffiffiffiffiffiffi Z 2 !
1 1
Eab ¼ M 2gHg  g 2gH 
2
pðtÞdt ð4Þ The reaction offered by the slab was measured only at two sup-
2 M
ports. The total reaction offered by the four supports was obtained
The ratio of actual velocity to free fall velocity, g, of the drop as twice the total measured reaction, Fig. 12. The magnitude of
impact setup was 0.98. The suddenly developed high stresses in reaction force offered by the slab was relatively smaller in compar-
the slab will then propagate towards the slab boundaries, leading ison to the peak impact force for both reinforced and prestressed
to deformation and cracking of concrete. concrete slabs, see Table 3. A similar behaviour has been noticed
It should also be noted that according to energy conservation by Othman and Marzouk [11] and Kishi and Bhatti [29]. The mag-
principle, work done by the equivalent static force [in causing dis- nitude of peak reaction force was measured to be 57.8–77.8% of the
placement (us), Eq. (5)] should be equal to the energy absorbed peak impact force, see column ‘g’ of Table 3. This observed differ-
obtained from Eq. (4); ence between the impact and reaction force is responsible for bal-
ancing the inertia force. The peak support reaction was also seen to
Eab be influenced due to the induced pre-tensioning in concrete as well
us ¼ ð5Þ
Peqs as the kinetic energy of hammer. For a given concrete, the magni-
tude of peak reaction increased with the increase in drop height,
The displacement obtained from Eq. (5) has been compared
see column ‘d’ of Table 3. For a given drop height, the magnitude
with the measured displacement using Laser Displacement Sensor
of peak reaction was higher in prestressed concrete possibly due
(LDS).
to the increase in the stiffness of concrete (column ‘d’ of Table 3).
With increase in the level of prestress however the magnitude of
8. Results and discussion peak reaction was found to be same. The numerical results repro-
duced the magnitude of reaction force with 14 and 7% deviation,
The drop impact experiments have been performed wherein a respectively, for drop height 500 and 1000 mm. It should be noted
242.85 kg steel hammer has been dropped on reinforced and pre- that the simulated reaction-time curve was smoother than the
stressed concrete slabs of thickness 100 mm from 500 and measured curve. This is due to the fact that the supporting frame
1000 mm heights. The magnitude of prestress in the slab has been of target slab has induced some vibrations in the measured
considered to be 10% (level-1) and 20% (level-2) of the characteris- reaction-time curve.
tic strength of concrete. The impact force, reaction force and dis- The measured reaction of the target slab was noticed to have
placement of the slab have been measured and the results thus some time lag in comparison to the measured impact force. This
obtained have been reproduced numerically on ABAQUS/Explicit implies that the reaction response of the slab started sometime
666 V. Kumar et al. / Construction and Building Materials 188 (2018) 656–675

500 500
Experimental Experimental

Impact Force (kN)

Numerical
Impact Force (kN)

400 400 Numerical

300 300

200 200

100 100

0 0
0 5 10 15 20 0 5 10 15 20
Time (ms) Time (ms)
(a) Slab RC-H1 (b) Slab RC-H2
500 500
Experimental
Experimental
Impact Force (kN)

Impact Force (kN)

400 Numerical 400 Numerical

300 300

200 200

100 100

0 0
0 5 10 15 20 0 5 10 15 20
Time (ms) Time (ms)
(c) Slab PC1-H1 (d) Slab PC1-H2
500 500
Experimental
Numerical Experimental
Impact Force (kN)

Impact Force (kN)

400 400
Numerical
300 300

200 200

100 100

0 0
0 5 10 15 20 0 5 10 15 20
Time (ms) Time (ms)
(e) Slab PC2-H1 (f) Slab PC2-H2

Fig. 11. Impact force response for prestressed and reinforced concrete slabs.

Table 3
Impact force and reaction for prestressed and reinforced concrete slabs.

Specimen Impact Force Reaction Force Ratio Of Reaction Force To Impact

ID Force (%)
Experimental Peak, Impact Duration Numerical Peak Experimental Peak, Time-lag Numerical Peak
PI (kN) (ms) (kN) R (kN) (ms) (kN)
a b c d e f g = d/a (%)
RC-H1 259.10 15.0 254.22 194.81 0.2 201.53 75.19
PC1-H1 290.80 12.9 318.79 223.33 0.4 254.05 76.80
PC2-H1 287.78 13.4 259.09 223.95 0.4 255.25 77.82
RC-H2 355.46 18.4 382.53 205.48 0.4 212.12 57.81
PC1-H2 374.90 15.8 420.73 270.61 0.2 252.08 72.18
PC2-H2 371.24 13.7 427.84 269.17 0.3 265.64 72.51

after the establishment of contact between the hammer and the 8.3. Displacement-time response of slabs
slab, see Fig. 12. The time-lag in the reaction-time response repre-
sented the duration of time required to travel the stress wave from The actual displacement-time response for the prestressed and
point of impact to the support. For the tested concrete slabs, the reinforced concrete slabs has been provided in Fig. 13(a). The mag-
p
stress wave propagation velocity ðmw ¼ ðG=qÞÞ calculated with nitude of displacement increased suddenly and reached a maxi-
the help of shear modulus (G) and density (q) of the target slab mum value. The displacement subsequently decreased up to
was 2468 m/s, which took approximately 0.23 ms to reach the sup- some extent due to elastic recovery of slab and retained some per-
port from the point of impact [30]. The measured time-lag has manent value, see Fig. 13(b). The peak displacement in the pre-
been found to be 0.2–0.4 ms, see column ‘e’ of Table 3. stressed concrete slab was smaller than in the corresponding
 V. Kumar et al. / Construction and Building Materials 188 (2018) 656–675 667

300 300
Experimental Experimental
Numerical Numerical

Reaction (kN)
200 200
Reaction (kN)

100 100

0 0

-100 -100

0 5 10 15 20 0 5 10 15 20
Time (ms) Time (ms)
(a) Slab RC-H1 (b) Slab RC-H2
300 300
Experimental Experimental
200 Numerical 200 Numerical
Reaction (kN)

Reaction (kN)
100 100

0 0

-100 -100

0 5 10 15 20 0 5 10 15 20
Time (ms) Time (ms)
(c) Slab PC1-H1 (d) Slab PC1-H2
300 300
Experimental Experimental
Numerical Numerical
200 200
Reaction (kN)

Reaction (kN)

100 100

0 0

-100 -100

0 5 10 15 20 0 5 10 15 20
Time (ms) Time (ms)
(e) Slab PC2-H1 (f) Slab PC2-H2

Fig. 12. Reaction-time response for prestressed and reinforced concrete slabs.

reinforced concrete slab for a given drop height, Table 4. For The displacement obtained using the analytical approach
1000 mm drop height, the measured peak displacement in slab (Table 5) has been found to have close correlation with the actual
RC-H2, PC1-H2 and PC2-H2 was 19.9, 16.1 and 13.4 mm, respec- displacements measured through Laser Displacement Sensor,
tively (see Fig. 14). The residual displacement (at quarter span) Fig. 14(a). The maximum deviation between the measured and cal-
was also found to be highest in the RC-H2 slab (7 mm) followed culated displacement was 2.9 mm. Further, the displacement
by PC1-H2 (3.2 mm) and PC2-H2 (2.3 mm) slab, respectively obtained analytically was found to be closer to the experimental
(Fig. 13(b)). It may therefore be concluded that inducing prestress results than that predicted by numerical study (Fig. 14(b) and
enhanced the slab stiffness and thus improved the deformation (c)). In the present study, the maximum displacement experienced
resistance. by reinforced concrete slab corresponding to impact energy
The magnitude of displacement reproduced numerically for 2382.4 J (for 1000 mm drop height) was 19.96 mm. Thus, the ratio
reinforced and prestressed concrete slabs was relatively smaller of peak displacement to slab span and peak displacement to slab
than the actual displacement. For drop height 1000 mm, the pre- thickness was 0.025 and 0.20, respectively. In a similar experimen-
dicted peak displacement was 14.3, 7 and 6.5 mm, respectively tal investigation [32], wherein the reinforced concrete slabs (with
(Table 4). The primary reason behind the difference in the pre- and without FRP strengthened) were impacted by 300 kg mass at
dicted and experimental results has been attributed to the simplic- 4 m/s velocity with almost same impact energy (2400 J), the ratio
ity exercised in modelling the boundary conditions of the slab. of peak displacement to slab span and peak displacement to slab
Jones [31] discussed complexities and limitations involved in sim- thickness was 0.009 and 0.1, respectively.
ulating the physical behaviour of the structure through finite ele-
ment analysis. It was emphasized that simplifications exercised 8.4. Energy absorption capacity
in modelling the dynamic material behaviour, application of load-
ing, structural detailing, boundary conditions and scaling of geom- The energy absorption capacity of the prestressed and rein-
etry in structural modelling are some potential reasons behind the forced concrete slabs has been obtained by calculating the area
difference between the actual and numerical response. under the load versus mid-span deflection plots, see Fig. 15. For
668 V. Kumar et al. / Construction and Building Materials 188 (2018) 656–675

Fig. 13. Displacement-time response and residual displacement of the slab specimens obtained experimentally.

the drop height 500 mm, no significant difference of prestressing shown highest energy absorption (column ‘a’ of Table 6). The loss
force was observed on the energy absorption capacity. For drop of energy during the impact primarily occurred due to the friction
height 1000 mm, the prestressed concrete slab with level-1 has between the hammer and the guides, vibrations of slab, vibrations
 V. Kumar et al. / Construction and Building Materials 188 (2018) 656–675 669

Table 4
Peak and residual displacement in prestressed and reinforced concrete slabs.

Specimen Displacement
ID
At center of span At quarter of span
Exp. Peak, Resi. Disp. Resi. disp./Peak Num. Peak Exp. Peak, Resi. Disp. Resi. disp./Peak Num. Peak
(mm) (mm) disp. (mm) (mm) (mm) disp. (mm)
a b c = b/a (%) d e f g = f/e (%) h
RC-H1 9.76 3.30 33.81 9.02 6.79 2.50 36.82 7.02
PC1-H1 8.20 2.67 32.56 6.17 5.22 1.87 35.82 5.40
PC2-H1 8.39 3.16 37.66 4.47 5.02 1.63 32.47 3.47
RC-H2 19.96 12.14 60.82 14.25 12.67 7.01 55.33 9.74
PC1-H2 16.09 6.96 43.26 7.0 8.97 3.18 35.45 5.67
PC2-H2 13.41 4.86 36.24 6.5 7.84 2.31 29.46 5.09

Resi. = Residual; Disp. = Displacement; Exp. = Experimental; Num. = Numerical.

Fig. 14. Comparison of slab displacements obtained by different approaches.

Table 5
Verification of peak displacements obtained using the proposed method.

Specimen Equivalent Static Peak Value of Displacements

ID Force, Peq-s (N)
Calculated from the energy absorption Eq. (4) Obtained by area under the load–displacement Measured with the help of
and Eq. (5), us1 (mm) curve and Eq. (5), us2 (mm) LDS, uI (mm)
a b c d
RC-H1 104636.5 8.85 9.21 9.76
PC1-H1 119047.2 7.82 7.92 8.20
PC2-H1 113005.5 8.42 8.73 8.39
RC-H2 120642.9 17.03 17.64 19.96
PC1-H2 144051.2 13.67 15.91 16.09
PC2-H2 149728.7 13.14 14.15 13.41
670 V. Kumar et al. / Construction and Building Materials 188 (2018) 656–675

400 400

300 300

Load (kN)

Load (kN)
200 200

100 100

0 0
0 5 10 15 20 0 5 10 15 20
Displacement (mm) Displacement (mm)
(a) Slab RC-H1 (b) Slab RC-H2
400 400

300 300
Load (kN)

Load (kN)
200 200

100 100

0 0
0 5 10 15 20 0 5 10 15 20
Displacement (mm) Displacement (mm)
(c) Slab PC1-H1 (d) Slab PC1-H2

400 400

300 300
Load (kN)

Load (kN)

200 200

100 100

0 0
0 5 10 15 20 0 5 10 15 20
Displacement (mm) Displacement (mm)
(e) Slab PC2-H1 (f) Slab PC2-H2

Fig. 15. Load-displacement curve for prestressed and reinforced concrete slabs.

Table 6
Energy absorption capacity of prestressed and reinforced concrete slabs.

Specimen ID Exp. Energy Absorption, Eab (J) Total Energy of Impactor, MgH (J) Energy Loss (J) % Energy Loss Energy Absorption, using Eq. (4), Eab (J)
a b c = b-a d = c/b (%) e
RC-H1 963.47 1191.18 227.71 19.12 926.50
PC1-H1 942.52 1191.18 248.66 20.88 931.31
PC2-H1 940.12 1191.18 251.06 21.08 951.47
RC-H2 2128.48 2382.36 253.88 10.66 2055.10
PC1-H2 2291.61 2382.36 90.75 3.81 1968.54
PC2-H2 2117.98 2382.36 264.38 11.1 1967.32

of experimental setup, deformation of setup components, increase actual and predicted results described localized indentation at the
in the temperature and energy loss in the form of sound (column front surface (Figs. 16–18). The size of front surface indentation
‘d’ of Table 6). was found to be 35 and 70 mm in reinforced concrete slab, 30
The energy absorption capacity of the slabs computed using and 38 mm in prestressed concrete level-1 and 30 and 49 mm in
analytical approach discussed in Section 7 predicted the absorbed prestressed concrete level-2 for drop height 500 and 1000 mm,
energy with a maximum 14% deviation from the actual results (col- respectively. It may be concluded here that the size of indentation
umn ‘e’ of Table 6). did increase with increase in drop height, however, for prestressed
concrete no significant influence of drop height was observed.
8.5. Damage pattern The crack density and crack-width in target slab has increased
with the increase in drop height. For drop height 500 mm, the rein-
The target slabs have suffered the damage under impact loading forced concrete slab (RC-H1) experienced flexure cracks (at rear
leading to dominant flexural and some nominal shear cracks. Both surface) which traversed across both the spans as well as in
 V. Kumar et al. / Construction and Building Materials 188 (2018) 656–675 671

Fig. 16. Damage in reinforced concrete slabs subjected to drop impact from 500 mm.

diagonal directions. The longitudinal cracks were observed to be increased significantly. Further, the punching shear cracks were
wider than the others cracks (Fig. 16). For drop height 1000 mm, also noticed in this case (Fig. 17). The cracks in the punching zone
the flexural cracks developed in the reinforced concrete slab were observed to be wider. Although the development of cracks
672 V. Kumar et al. / Construction and Building Materials 188 (2018) 656–675

Fig. 17. Damage in reinforced concrete slabs subjected to drop impact from 1000 mm.

could not be predicted through numerical simulations, however, respectively for drop height 500 and 1000 mm. Simulations also
an increase in the stress intensity with increase in drop height is predicted the punching action in concrete corresponding to drop
clearly visible through the stress contours, see Figs. 16 and 17 height 1000 mm, Fig. 17.
 V. Kumar et al. / Construction and Building Materials 188 (2018) 656–675 673

Fig. 18. Damage in prestressed concrete slabs (20% prestress) subjected to drop impact from 1000 mm.

The cracks developed in prestressed concrete slabs (at rear sur- For a given prestressing level, the crack width as well as crack pop-
face) were significantly lesser in the population and smaller in ulation increased with the increase in drop height. Further, there
thickness in comparison to the reinforced concrete slabs (Fig. 18). was no visible punching action in the prestressed concrete. With
674 V. Kumar et al. / Construction and Building Materials 188 (2018) 656–675

increase in prestressing level, the crack width and crack density stresses developed in the prestressing wires decreased with
were reduced, however, one major tensile crack was noticed in increase in the prestress level as well as drop height.
the direction of prestressing across the thickness of slab only for A simplified analytical method based on mass, momentum and
the drop height 1000 mm (Fig. 18). Numerical results also pre- energy conservation enabled calculation of the peak displacement
dicted high degree of deformation corresponding to the prestress and the absorbed energy in close correlation with the experimental
level-2. results.
The most important contribution of the finite element simula-
tions is the prediction of the stresses and deformation in the rein-
Conflict of interest
forcement and the prestressing wires which have been typically
explained through Figs. 16–18.
None.
In the reinforced concrete slabs the stresses developed in rein-
forcement were confined to the slab boundary possibly due to the
major deformation of slab in the central region. Significant defor- Acknowledgements
mation has been observed in the reinforcing bars corresponding
to drop height 1000 mm due to predominant punching action. Authors gratefully acknowledge the financial support provided
In the prestressed concrete slabs, the reinforcement normal to by Science and Engineering Research Board, Department of Science
the prestressing wires has developed tensile stresses (Fig. 18). In and Technology, India through the research Grant No. SB/S3/
general, the magnitude of stresses in reinforcing bars has increased CEE/0032/2014 for carrying out the present study.
with the increase in prestress level, and it has decreased, with the
increase in drop height. References
The prestressing wires experienced visible permanent deforma-
tion in the case of prestressing level-2 (Fig. 18). However, no frac- [1] W. Struck, W. Voggenreiter, Examples of impact and impulsive loading in the
field of civil engineering, Mater. Struct. 8 (1975) 81–87.
ture of wires was observed. The stresses developed in the [2] S. Kaewunruen, A.M. Remennikov, Impact capacity of railway prestressed
prestressing wires decreased with increase in the prestress level concrete sleepers, Eng. Fail. Anal. 16 (2009) 1520–1532, https://doi.org/
as well as drop height. 10.1016/j.engfailanal.2008.09.026.
[3] S. Mindess, N. Banthia, A. Bentur, The response of reinforced concrete beams
with a fibre concrete matrix to impact loading, Int. J. Cem. Compos. Light
Concr. 8 (1986) 165–170, https://doi.org/10.1016/0262-5075(86)90037-0.
9. Conclusions [4] F. Wittmann, P. Roelfstra, H. Mihashi, Y. Huang, X. Zhang, N. Nomura, Influence
of age of loading, water-cement ratio and rate of loading on fracture energy of
concrete, Mater. Struct. 20 (1987) 103–110, https://doi.org/10.1007/
Impact response of prestressed and non-prestressed reinforced BF02472745.
concrete has been studied by carrying out experiments and finite [5] D.Y. Yoo, N. Banthia, S.W. Kim, Y.S. Yoon, Response of ultra-high-performance
element simulations of 100 mm thick slabs subjected to impact by fiber-reinforced concrete beams with continuous steel reinforcement
subjected to low-velocity impact loading, Compos. Struct. 126 (2015) 233–
242.8 kg hammer falling freely from different predefined heights. 245, https://doi.org/10.1016/j.compstruct.2015.02.058.
The resultant peak impact force and the support reactions have [6] Y. Hao, H. Hao, G. Chen, Experimental investigation of the behaviour of spiral
been found to be higher against prestressed concrete slabs than steel fibre reinforced concrete beams subjected to drop-weight impact loads,
Mater. Struct. 49 (2016) 353–370, https://doi.org/10.1617/s11527-014-0502-
the equivalent reinforced concrete slabs. The peak impact force 5.
and the reaction increased by 12% and 32%, respectively, against [7] L.H. Han, C.C. Hou, X.L. Zhao, K.J.R. Rasmussen, Behaviour of high-strength
the slab with 10% prestress and 11% and 31%, respectively, against concrete filled steel tubes under transverse impact loading, J. Constr. Steel. Res.
92 (2014) 25–39, https://doi.org/10.1016/j.jcsr.2013.09.003.
the slab with 20% prestress. The numerical results predicted the [8] M.R. Bambach, Design of hollow and concrete filled steel and stainless steel
peak impact and reaction force within 15% and 14% deviation, tubular columns for transverse impact loads, Thin-Walled Struct. 49 (2011)
respectively. 1251–1260, https://doi.org/10.1016/j.tws.2011.05.009.
[9] M.R. Bambach, H. Jama, X.L. Zhao, R.H. Grzebieta, Hollow and concrete filled
Prestressed concrete slabs have also been found to experience steel hollow sections under transverse impact loads, Eng. Struct. 30 (2008)
reduced peak displacement in comparison with the equivalent 2859–2870, https://doi.org/10.1016/j.engstruct.2008.04.003.
non-prestressed concrete slabs. A maximum reduction in the peak [10] J. Sawan, Impact effect on R.C. slabs: experimental approach, J. Struct. Eng. 112
(1986) 2057–2065.
displacement at the center has been found to be 19.4 and 32.8% in
[11] H. Othman, H. Marzouk, An experimental investigation on the effect of steel
the prestressed concrete slab with 10% and 20% induced stress, reinforcement on impact response of reinforced concrete plates, Int. J. Impact
respectively. It may therefore be concluded that the prestressed Eng. 88 (2016) 12–21, https://doi.org/10.1016/j.ijimpeng.2015.08.015.
concrete slabs have offered higher impact resistance in comparison [12] Y. Chen, I.M. May, Reinforced concrete members under drop-weight impacts,
Proc. ICE – Struct. Build. 126 (2009) 45–56, https://doi.org/10.1680/
to the equivalent non-prestressed concrete slab. stbu.2009.162.1.45.
With the increase in drop height, the flexural cracks developed [13] N.A. Siddiqui, B.M.A. Khateeb, T.H. Almusallam, Y.A. Al-Salloum, R.A. Iqbal, H.
in the reinforced concrete slab increased and punching action Abbas, Reliability of RC shielded steel plates against the impact of sharp nose
projectiles, Int. J. Imp. Eng. 69 (2014) 122–135, https://doi.org/10.1016/j.
became more pronounced. Simulations also predicted the punch- ijimpeng.2014.03.001.
ing action in reinforced concrete corresponding to drop height [14] N.A. Siddiqui, B.M.A. Khateeb, T.H. Almusallam, H. Abbas, Reliability of double-
1000 mm. The cracks developed in prestressed concrete were sig- wall containment against the impact of hard projectiles, Nucl. Egn. Des. 270
(2014) 143–151, https://doi.org/10.1016/j.nucengdes.2014.01.003.
nificantly lesser in population and there was no visible punching [15] X. Ye, N. Wang, S. Mindess, Effect of loading rate and support conditions on the
action. With increase in the prestressing level, however, the split- mode of failure of prestressed concrete railroad ties subjected to impact
ting action was noticed in prestressed concrete slab. loading, Cem. Concr. Res. 24 (1994) 1286–1298.
[16] S. Kaewunruen, A.M. Remennikov, Dynamic crack propagations in prestressed
Simulations predicted that the stresses developed in reinforce- concrete sleepers in railway track systems subjected to severe impact loads,
ment of the reinforced concrete slab were confined to the slab ASCE J. Struct. Eng. 136 (2010) 749–754, https://doi.org/10.1061/(ASCE)
boundary due to the major deformation of slab occurred in the cen- ST.1943-541X.0000152.
[17] H. Abdalla, J.B. Kennedy, Dynamic analysis of prestressed concrete beams with
tral region. In the prestressed concrete slabs, the reinforcement
openings, J. Struct. Eng. 121 (1995) 1058–1068.
normal to the prestressing wires has developed tensile stresses. [18] Y. Jing, Z.J. Ma, D.B. Clarke, Full-scale lateral impact testing of prestressed
In general, the magnitude of stresses in reinforcing bars has concrete girder, Struct. Concr. (2016) 1–35, https://doi.org/
increased with the increase in prestress level, and it has decreased, 10.1002/suco.201500224.
[19] A. Rajput, M.A. Iqbal, Impact behavior of plain, reinforced and prestressed
with the increase in drop height. The prestressing wires experi- concrete targets, Mater. Des. 114 (2017) 459–474, https://doi.org/10.1016/
enced visible permanent deformation for 20% prestress. The j.matdes.2016.10.073.
 V. Kumar et al. / Construction and Building Materials 188 (2018) 656–675 675

[20] A. Rajput, M.A. Iqbal, Ballistic performance of plain, reinforced and pre- [27] S. Kaewunruen, A.M. Remennikov, Investigations of static and dynamic
stressed concrete slabs under normal impact by an ogival-nosed projectile, Int. performance of railway prestressed concrete sleepers, Annu. Conf. Expo. Exp.
J. Impact Eng. (2017), https://doi.org/10.1016/j.ijimpeng.2017.03.008. Appl. Mech., 2007.
[21] T. Hassan, G. Lucier, S. Rizkalla, P. Zia, Modeling of L-shaped, precast, [28] P. Sukontasukkul, S. Mindess, The shear fracture of concrete under
prestressed concrete spandrels, PCI J. 52 (2007) 78–92. impact loading using end confined beams, Mater. Struct. 36 (2003)
[22] B. Mercan, A.E. Schultz, H.K. Stolarski, Finite element modeling of prestressed 372–378.
concrete spandrel beams, Eng. Struct. 32 (2010) 2804–2813, https://doi.org/ [29] N. Kishi, A.Q. Bhatti, An equivalent fracture energy concept for nonlinear
10.1016/j.engstruct.2010.04.049. dynamic response analysis of prototype RC girders subjected to falling-weight
[23] T.Y. Lin, N.H. Burns, Design of Prestressed Concrete Structures, 3rd edition., impact loading, Int. J. Impact Eng. 37 (2010) 103–113, https://doi.org/10.1016/
John Wiley & Sons, New York, NY, 1981. j.ijimpeng.2009.07.007.
[24] T.J. Holmquist, G.R. Johnson, W.H. Cook, A computational constitutive model [30] D.M. Cotsovos, A simplified approach for assessing the load-carrying capacity
for concrete subjected to large strains, high strain rates, and high pressures, of reinforced concrete beams under concentrated load applied at high rates,
Fourteenth international symposium on ballistics, Quebec City, Canada, 1993, Int. J. Impact Eng. 37 (2010) 907–917, https://doi.org/10.1016/j.
1–10. ijimpeng.2010.01.005.
[25] M. Polanco-Loria, O.S. Hopperstad, T. Børvik, T. Berstad, Numerical predictions [31] N. Jones, The credibility of predictions for structural designs subjected to large
of ballistic limits for concrete slabs using a modified version of the HJC dynamic loadings causing inelastic behaviour, Int. J. Impact Eng. 53 (2013)
concrete model, Int. J. Imp. Eng. 35 (2008) 290–303, https://doi.org/10.1016/j. 106–114, https://doi.org/10.1016/j.ijimpeng.2011.12.008.
ijimpeng.2007.03.001. [32] A.Q. Bhatti, N. Kishi, K.H. Tan, Impact resistant behaviour of RC slab
[26] G.R. Johnson, W.H. Cook, Fracture characteristics of three metals subjected to strengthened with FRP sheet, Mater. Struct. 44 (2011) 1855–1864, https://
various strains, strain rates, temperatures and pressures, Eng. Fract. Mech. 21 doi.org/10.1617/s11527-011-9742-9.
(1985) 31–48, https://doi.org/10.1016/0013-7944(85)90052-9.s.