Sei sulla pagina 1di 7

Impact Testing of Concrete Using a

Drop-weight Impact Machine


by N. Banthia, S. Mindess, A. Bentur and M. Pigeon

ABSTRACT--A detailed description of the instrumented dropweight impact machine is presented. The instrumentation, the
calibration, the inertial loading correction, and the dynamic
analysis of a concrete beam specimen undergoing three-point
impact flexural loading are described. Some results, using
such an impact testing machine, obtained from tests done on
plain concrete, fiber-reinforced concrete, and conventionally
reinforced concrete are presented. It is concluded that the use
of such a testing machine may be successfully made in order
to test cementitious materials under )mpact.

Introduction
The low strains associated with concrete failure place it
in the category of brittle materials. Like other ceramics,
concrete also exhibits stress-rate sensitivity in all the three
loading configurations, viz. compression, '-3 tension',' and
flexure.': This implies that the statically determined
properties of concrete in the laboratory may not be used
to predict the behavior of concrete Subjected to high
stress rates, those associated with impact, blast, or earthquake, Since the conventional testing machines may not
be used to generate such high rates of loading, special
apparatus are required. Unfortunately, a standard technique for testing concrete under impact does not exist.
Although various investigators '-7 have used various testing
techniques, results often cannot be compared. The main
reasons behind the incomparable nature of these testing
techniques are the different methods of loading, the
different energy-loss mechanisms and the different ways
of analyzing the results. Consequently, little general agreement exists over the magnitude of the observed effects.
Nevertheless, a general agreement exists over the necessity
of a standard testing technique for testing concrete under
high stress rates associated with impact. In this paper, a
drop-weight impact machine, its construction, instrumentation and calibration, analysis of the results, and the
problems associated with its use in generating impact
flexural loading are outlined. Some results obtained with
normal-strength plain, high-strength plain, fiber-reinforced,
and conventionally reinforced concrete are also presented.

Experiment
The Drop-weight Impact-testing Machine
The drop-weight impact machine is shown in Fig. I. It
has a frame 3.5-m tall mounted on a reinforced-concrete
pedestal 1.5 m 1.5 m in area and 0.9-m high. The
frame is rigidly secured on top of the pedestal using 37ram bolts. A hammer weighing 3.38 kN slides up and
down the vertical posts upon being attached to a hoist.
The hammer has pneumatic brakes in its body by which
it can 'grab on' to the vertical posts. Once the brakes are
applied, the hoist may be detached from the hammer.
Upon releasing the brakes, the hammer falls freely on a
beam specimen supported on two support anvils as shown
in Fig. 1. The striking end of the hammer (called the 'tup')
is shown in Fig. 2. The hammer may be raised to heights
of up to 2.4 m above the specimen. By dropping the
hammer through different heights, the applied stress rate
may be varied.

Instrumentation
THE TUP
As the hammer strikes the beam, the contact load
between the hammer and the beam develops. Load
measurements are made by the eight bonded strain gages
placed in two 25-mm diameter holes (Fig. 2). This
procedure resulted in an amplification (by a factor of
three in this case) in the signals by making use o f the
stress concentration at the boundaries of the holes. 8'9 The
circuit of the tup is shown in Fig. 2(b).
THE SUPPORT ANVIL
The support anvil [Fig. 3(a)] is capable of reading the
vertical as well as the horizontal support reaction. These
two reactions are read separately by the imbalance generated
in two separate Wheatstone bridges. The vertical reaction
is read from the strain gages mounted in the circular holes
[Fig. 3(b)], while the horizontal reaction is read from the
strain gages mounted in between the two holes [Fig. 3(c)].
The independent nature of the horizontal and the vertical
reaction channels in the support anvil should be noted.
ACCELEROMETERS

N. Banthia is Attache de Recherche, Department of Civil Engineering,


Laval University, Ste-Foy. Quebec, G1K7P4, Canada. S. Mindess is
Professor, Department of Civil Engineering, University of British Columbia, Vancouver, British Columbia V6T1 WS, Canada. A. Bentur is Professor, Building Research Station, Technion-Israel Institute of Technology,
Technion City, Haifa 32000, Israel. M. Pigeon is Professor, Department
of Civil Engineering, Laval University, Ste-Foy, Quebec, GIK7P4, Canada.
Original manuscript submitted: September 25, 198Z Final manuscript
received: June 2, 1988.

The accelerometers (Fig. 1) mounted along the length of


the beam were piezoelectric sensors with a resonant frequency of 45 kHz. With a resolution of 0.01 g, the
accelerometers can read up to 500 g and have an overload protection of up to 5000 g (where g is Earth's gravitational acceleration). The calibration for the accelerometers
was supplied by the manufacturer.

E x p e r i m e n t a l M e c h a n i c s ~ 63

PHOTOCELL ASSEMBLY
The photocell assembly consists of a strip of metal
with holes punched in it, that runs parallel to the
columns of the machine (Fig. I) and a photocell
mounted on the hammer that slides along the strip.
As soon as the photocell reaches a hole in the strip
(Fig. 4), the beam of light falls on the photocell
through the hole registering an output. The use of
the photocell assembly is made for two purposes:
first, for the triggering of the data-acquisition system and second, for the determination of the hammer
acceleration as it fails under gravity. The hammer, once
released, passes a hole in the strip before hitting the
specimen. This interception of the hole triggers the dataacquisition system (Fig. I). The use of the photocell for
the determination of the hammer acceleration is described
below.

Calibration
CALIBRATION OF THE TUP AND THE SUPPORT ANVIL
The tup and the vertical-load channel of the support
anvil were calibrated b y subjecting them to static compres-

sion in a hydraulically loaded universal testing machine.


The horizontal load channel of the support anvil was
calibrated by applying a horizontal static load on the
anvil using a lower range universal testing machine, Load
was applied in small steps up to about 70 percent of the
elastic capacity followed by an unloading in steps to check
for any hysteresis. Although, the steel in the tup and the
support anvil was to be loaded dynamically in an actual
test, the static calibration was assumed to be reasonable.
CALIBRATION OF THE HAMMER ACCELERATION
AS mentioned earlier, the photocell assembly sent out a
voltage signal, in the form of a spike, whenever it intercepted a hole in the metal strip. A typical output from the
photoccU assembly is shown in Fig. 5. The data from the
photocell assembly indicated the time required by the
hammer to travel the distances between the successive
holes.
If it can be assumed that the downward acceleration of
the hammer (ah) is constant, and if we know the time
required by the hammer under a free fall to travel two
adjacent segments of length S, and $2 (Fig. 5) then from
the laws of motion, the acceleration of the hammer (t/h)
may be obtained as follows. Between the first and second
hole,
(1)

S, = V, Atl + lAahAt~

Between the second and third hole,


(2)

$2 = V2At2 + 89

Also

;:ooo= o

~ ~ ~

o.ao

oo

'

Pedestol

Fig. 1--A schematic layout of the impact testing setup

L
(a)

_J

@
(a)

T2

+JISTRAIN GAUGIrS
TYPE: SON DEO

RESISTANCE : Sw
_e 0 . 3 %
SAGE FACTOR: 2.OF ~" 0.5%
TEMPERATURE COEFFIClENT:~ O.t%

Fig. 2--The tup and its circuit

64 9 March 1989

EXCITATION

(b)

[.m
Excilotion

[xcitotion

(b)

Ic)

Fig. 3--The support anvil and its circuit

(3)

II2 = II1 + ahAt~

pillers after repeated use yielded accelerations as low as


8.60 m/s 2.

Solving for ah we have


ah

2($2Atl - S, Atz)
A t x A t 2 ( A t l + At2)

(4)

If the holes are equally spaced (as in the case of this


study), i.e., $1 = $2 = S, then
ah =

2S(At~ - AG)
At, At2(At~ + At2)

(5)

It is worth mentioning here that the acceleration of the


hammer, as obtained by using eq (5), was always found
to be less than Earth's gravitational acceleration of 9.81
m/s 2. The friction between the columns of the machine
and the hammer was thought to be the reason behind this
discrepancy. An acceleration test done right after cleaning
the pillers with acetone resulted in a value of hammer
acceleration of 9.60 m/s 2. On the other hand, unclean

Acquisition Storage and Retrieval of Data


The five-channel data-acquisition system was triggered
by the freely falling hammer itself during each test. Once
triggered, the data-acquisition system acquired the data
from the tup, the support anvil (only one reaction channel connected at a time), and the three accelerometers.
The data-acquisition system, based on an IBM PC,
deposited the time-base data thus acquired on a floppy
disk which was eventually transferred to a mainframe
computer and analyzed. During each test the time-base
data were acquired for a preselected length of time
depending on the expected duration of the impact. While
the time required to fail a conventionally reinforced concrete beam was about 150 millisecond, the corresponding
time for a plain concrete beam was only about 15 millisecond. The voltage signals were converted to load and
acceleration values using the calibration curves described
above.

Analysis of the Results


~

The flexural impact tests on the cementitious materials


are performed with two objectives in mind: first, the
determination of the peak bending load, and second, the
determination of energy that is consumed by the specimen
either up to the peak load or up to failure. The energy
consumed up to failure is also sometimes called the
'toughness' or simply the fracture energy. The computation of the two may be accomplished as follows.

Metol strip

fqJ

,,o,o

Z>--7 Q~

cell

Light

source

Holes

:R

THE PEAK BENDING LOAD


?Excitation

Output

Fig. 4--The photocell assembly

i~ Output in volts
Hammer
velocity = VI

-- .~

Hammer
veloctty=V 2

(~

i
-'-~ . . . .
N

Hammer
velocity--V3

(~

Fig. 5--Determination of hammer acceleration using


the photocell

It is now a well-known fact that the load recorded by


the tup in the initial part of impact is not the true stressing
or the true bending load owing to the specimen inertial
effects. '~ The acceleration of the specimen gives rise to
d'Alembert's forces which are recorded by the tup along
with the stressing or the bending load. Since the cementitious materials are usually very brittle, the entire impact
event may occur while the specimen is still being accelerated.
The mechanical response may thus be entirely overshadowed by the inertial response, and as such, the true
bending load may be only a fraction of the recorded tup
load. Proper dynamic analysis of the system is therefore essential.
The recorded tup load, P , ( t ) , (Fig. 6) is a point load at
the midspan of the beam whereas the inertial reaction of
the beam is a body force distributed throughout the body
of the beam. This distributed inertial load should therefore be replaced by a generalized point inertial load,
P i ( t ) , which can then be subtracted from the tup load in
order to obtain the generalized bending load, P b ( t ) ,
acting at the center. As will be shown later, this generalized
bending load can then be assumed to act on the beam at
the midspan by itself, and will predict the correct energies,
midspan moments, and stresses. If the acceleration in
between the accelerometers may be obtained by linear
interpolationfl the acceleration at the center may be
obtained by linear extrapolation, and finally, if the
accelerations may be assumed to be symmetric about the
midspan, then the acceleration at every point along the
length of the beam is known [Fig. 6(b)]. If the beam is
given a virtual displacement, 6Uo, compatible with its
constraints [Fig. 6(0], then from the virtual work expression we may write

Experimental Mechanics 9 65

+ 2 J~ o A i t ' ( y , t ) $ u ( y ) d y

(6)

where / i ( x , t ) or ~/(y,t) is the acceleration at a certain


location along the beam, ~u(x) or $u(y) is the virtual
displacement at a certain location along the beam, 0 is the
mass density of the beam material and A is the area of the
cross s~'tion of the beam. On breaking the first integral into
different segments of the beam between the accelerometers,
expressing accelerations at various points along the segments in terms of the accelerations at the accelerometer
locations (recorded), and assuming that the virtual displacement at a location is proportional to the acceleration
there, we may write after a few steps 7

~ Hammer

t'

'

D1

P,(t) = -~ oA [ ~

P~(t)6Uo = Jo OAli (x,t) ~u(x)dx

Aceelerometers

D2

(U2o(t) + U~(t) + Uo(t)i/,(t))

(ii~(t)+ii~(t)+ii,(t)iiz(t))

Da
+ ~

h
[~](t) ]
no(t)

U~(t)

(7)

Once the generalized inertial load is obtained, the beam


can be modeled as a single degree of freedom system, and
the generalized bending load [Pb(t)] may be obtained
simply by using the equation of dynamic equilibrium,

Pb(t) = P+(t) - Pi(t)

(8)

Considerable simplification is possible in the above


treatment of inertial loading if some simple mathematical
function may be chosen to define the acceleration distri -+
bution along the length of the beam. Many tests conducted
to study this distribution indicated that the accelerations
along the beam are indeed simple mathematical functions;
the distribution being linear for plain, and sinusoidal for
conventionally reinforced beams. 7 Equation (7) can then
be appropriately simplified to include only the acceleration
at the midspan.
rP
8 h3
Pi(t) = oAiio(t) [ 3 + 3 - t-~]
(9)

(a)
(for the linear case) and
r~
P~(t) = oA~io(t) t~- + 2 7r2h3 ]

(10)

(for the sinusoidal case) where the second term on the


right-hand side accounts for the linear accelerations along
the overhangs of the beam. 8
Once the generalized bending load Pb(t) is known, the
tup load and the distributed inertial load [Fig. 7(a)] in the
actual dynamic system may be replaced by Pb(t) alone to
form the equivalent static system of Fig. 7(b).
For the case of linear accelerations (plain-concrete
beams), it may be shown using dynamic equilibrium
[Fig. 7(a)] that

Assumed~ ]~Lineorly extrapolated


(b)

Pt (t)

R,(t) = 2 P , ( t ) - o A ~ i o ( t ) [ e _

_0~]

(11)

and

Mo(t) = Pt(t) ~ - oAiio(t) [1E-~2+ - ~ g3 ]

where Mo(t) is the moment at the center in the dynamic


system of Fig. 7(a). The generalized bending load Pb(t)
acting on the equivalent static system may be obtained
using eqs (8) and (9) for the case of plain concrete.

(c)
Fig. 6--Computation of the generalized
inertial load

8h3 j]
Pb(t) = P t ( t ) - A o i i o ( t ) re
t--~- + 3t
-j
2

inertialload ~pt (t)

(a)
Fig. 7(a)--The dynamic loading and (b) the
equivalent static loading

66 9 March 1989

(12)

Pb(t)

(b)

+,+j

(13)

If Me o(t) is the midspan moment in the equivalent static


system, then,
e

e2

M.o(t) = P,(t)--~ = P , ( t ) - i - - o A i i o ( t ) [-~ + . .

]
(14)

A comparison between eq (12) and eq (14) indicates that

Mo(t) and Me o(t) are the same. Or, in other words, the
placement of Pb(t) on the beam [Fig. 7(b)] results in the
true equivalent static system. The peak bending load,
therefore, is the peak value of the generalized load.
A CHECK
The instrumentation in the support may be used to
check the validity of the above analysis. Figure 8 presents
a comparison between the experimental support load and
the evaluated support reaction as computed using eq (11).
It can be noted that they reasonably agree with each
other. The other significant feature of Fig. 8 is the lag of
about 0.4 millisecond between the evaluated and the
measured reaction. The finite time taken by the stress
waves to travel the distance from the point of impact to
the support may, to some extent, explain this lag.
Figure 9 presents the measured horizontal reaction. It
can be seen that the horizontal reaction is close to zero at
all times indicating the simply supported nature of the
beam.
THE FRACTURE ENERGY
As soon as the hammer hits the beam, a sudden transfer
of energy from the hammer to the beam occurs. The
hammer velocity decreases due to the obstacle in its path.
At any time t during the impact, if I P , ( t ) d t represents
the impulse acting against the hammer, the kinetic energy
lost by the hammer AE(t) may be obtained from the
impulse-momentum relationships,7
A E ( t ) = "~-mh
1
[2ahh - ( 2x/~d~hh-

mh

the form of elastic strains and vibrations. Since the strain


energy or the vibrational energy in the machine may never
be determined, eq (15) may not be used to determine the
energy received by t h e beam from the hammer at any
time t. Moreover, a major portion of the energy received
by the beam appears as kinetic energy which is of little
concern to us. The energy that does concern us is the
bending energy [Eb(t)], or the energy given by the area
under the generalized bending load [Pb(t)] versus the
load-point deflection [Uo(t)] curve (Fig. 10).
t

Eb(t) = IO Pb(t)duo

The deflection at the load point, uo(t), may be obtained by double integration of the extrapolated acceleration at the load point [Uo(t)].
t t
Uo(t) = Io Io tio(t) dtdt
(17)
At the point of failure (Fig. 10), the beam stops receiving
energy from the hammer (tup load falls to zero) and the
area under the Pb(t) versus the LPD [Uo(t)] curve [eq
(16)] represents the fracture energy or the energy required
to create two new fracture surfaces. At this point, a plain-

2,0

1,5

I P ' ( t ) d t ) 2]

(15)

(16)

1,0

where m h is the mass of the hammer and h is the height


of its drop.
The energy lost by the hammer [eq (15)] is partly transferred to the beam and partly stays within the machine in

_l

0,5

J
0

Time, ms

Fig. 9 - - H o r i z o n t a l support reaction


(recorded)
Experimental support load

af

Evaluated support reaction


" 4

.J

tea = fracture energy

t
z

'/i

I
I

(.9
Fodure

6
8
Time, ms
Fig. 8 - - A check on the analysis

I0

12

Load point deflection U o ( t )

Fig. I O - - A typical load (Pb) versus loadpoint-deflection (Uo) plot

Experimental Mechanics ~ 67

concrete beam breaks into two halves and the two broken
halves swing about their support points away from the
tup. It may be assumed that the beam halves, although
having considerable kinetic energy, have no bending or
strain energy and all the energy given by the area under
the generalized-bending-load versus load-point-deflection
plot has been used up in creating new fracture surfaces.
It is shown in Ref. 13 that at the point of failure, the
fracture energy and the kinetic energy of the broken
halves together account for most of the energy lost by the
hammer as given by eq (15).

Results and Discussion


Detailed results of the extensive testing carried out
using the above described impact machine appear in Refs.
14-18. However, some of the results are presented in the
form of generalized-bending-load versus load-point-deflection plots of Figs. ll(a), (b), (c), (d) and (e) and also in
the form of Table 1. These plots present a comparison
between the static and the dynamic results obtained for
normal-strength plain concrete, high-strength plain concrete, normal-strength polypropylene-fiber-reinforced concrete, normal-strength steel-fiber-reinforced concrete, and
conventionally reinforced normal-strength concrete,
respectively. (The static tests correspond to a x-head
movement rate of 4.2 x 10-7 m/s. The dynamic results

TABLE 1 --RESULTS FROM STATIC AND IMPACT TESTS


Static I
Impact =
Peak Bending Fracture Peak Bending Fracture
Load 3
Energy ~
Load 3
Energy3
N
Nm
N
Nm

Normal-Strength 6344 (306)4

5.5 (1,5)

16932 (428)

90,1 (6.5)

2.8 (0,6)

18760 (446) 74.9 (18.6)

14.0 (4.4)

17300 (821)

Plane Concrete s

High-Strength

9720 (1809)

Plain Concrete 6

Normal-Strength 7302 (99)


PolypropyleneFiberReinforced

119.4 (8.1)

Concrete 7

Normal-Strength 11500 (670)


Steel-Fi ber-

44.8 (19)

24006 (1629) 237.6 (7.5)

Reinforced
Concrete 8

Conventionally

22671 (3102) 4421~ (45) 36664 (888) 8801~ (300)

Reinforced

Normal-Strength
Concrete g

1Static tests done at the cross head speed of 4.2 x 10-~ m l s


=Height of hammer drop = 0.5 m
3Average taken over six or more specimens
4Numbers in parentheses are the standard deviations
SCrushing strength = 42 MPa
SCrushing strength = 82 MPa, 16 percent (by weight of cement)
of microsilica
7Fibrillated polypropylene fibers, 37-mm long, 0.5 percent by
volume
aSteel fibers with both ends hooked, 50-mm long, 1.5 percent by
volume
9Steel Area = 1.12 percent
o Calculated up to a point when the load had dropped back to
1/3 of its peak value

68 9 March 1989

correspond to a 0.5-m hammer drop.) The significant


stress-rate sensitivity of concrete may be noticed from the
plots of Fig. 11. In general, all kinds of concretes were
found to be stronger (higher peak bending loads) and
tougher under impact than under static conditions. This
change in the mechanical behavior of concrete is in accordance with the results obtained by other investigators s''9
using totally different techniques of high-stress-rate
generation.
The use of fibers in cementitious construction materials
is gaining importance. These composites are generally
believed to be tougher than the unreinforced matrix and
as such are considered to be more impact resistant. The
higher values of fracture energies obtained for these
composites (Table 1) than the unreinforced matrix seem
to confirm this belief.
Conventionally reinforced concrete with its strategically
placed reinforcing bars was found to be the most impact
resistant of all (Table 1). This suggests that fiber-reinforced matrix along with conventional reinforcement
would produce the most suitable material for impactloading situations.

Conclusions
The use of an instrumented, drop-weight impact
machine may be successfully made in order to investigate
the impact behavior of concrete. The instrumentation
described here is sufficient to apply the inertial loading
correction and to derive useful information from the
impact testing. Concrete is a significantly stress-ratesensitive material. In general it is stronger and more energy
absorbing under impact than under static loading.

Acknowledgments
The authors are grateful to the staff of the University
of British Columbia, Department of Civil Engineering,
Vancouver, Canada where this work was carried out. In
particular, the help from Mr. G.D. Jolly, Mr. R.B. Nussbaumer and Mr. M. Nazar is thankfully acknowledged.

References
1. Abrams, D.A., "'Effect of Rate of Application of Load on the
Compressive Strength of Concrete, "" Proc. ASTM 17, Part 2, 364-367
(1917).
2. Watstein, D., "Effect of Straining Rate on the Compressive
Strength and Elastic Properties of Concrete, '" J. ACI, 49 (8), 729-756
(April 1953).
3. Atchly, B.L. and Furr, H.L., "'Strength and Energy Absorption
Capacity of Plain Concrete Under Dynamic and Static Loading, "' J. ACI,
745-756 (Nov. 1967).
4. Macneely, D.J. and Lash, S.D., "'Tensile Strength of Concrete
Under Dynamic and Static Loading, ""J. ACI, 60 (6), 751-760 (1963).
5. Zielinsky, A.J., "'Fracture of Concrete and Mortar Under Uniaxial
Impact Tensile Loading, '" PhD Thesis, De(ft Univ. of Tech. (1982).
6. Suaris, W. and Shah, S.P., "'Properties of Concrete Subjected to
Impact, "' ASCE, Struct. Div., 109 (7), 1727-1741 (July 1983).
Z Banthia, N.P., "'Impact Resistance of Concrete," PhD Thesis,
Univ. of British Columbia, Vancouver, Canada (1987).
8. Bentur, A., Mindess, S. and Banthia, N., "'The Behavior of
Concrete Under Impact Loading: Experimental Procedures and Method
of Analysis, ""Materials and Structures, 19 (113), 371-378 (1986).
9. Timoshenko, S.P. and Goodier, N., "'Theory of Elasticity,'"
McGraw-HilI Kogakusha, Ltd., 3rd Ed. (1970).
10. Vanzi, S., Priest, A. and May, M.J., "'Influence of Inertial Loads
in Instrumented Impact Tests, ""Impact Testing of Metals, ASTM STP466,
165-180 (1970).
11. Server, W.L., "'Impact Three Point Bend Testing for Notched and
Pre-Cracked Specimens, ""J. Test. and Eval., 6 (1), 29-34 (Jan. 1978).
12. Gopalaratnam, V.S., Shah, S.P. and John, R., "'A Modbqed
Instrumented Charpy Test for Cement-Based Composites," EXPERIMENTAL
MECHANICS, 24 (2), 102-111 (June 1984).

18
16

NORMAL STRENGTH CONCRETE

16
14

HIGH STRENGTH CONCRETE

14

~ t2

zl2
.lO

,,m

< 8
o,._1 6

2
0

2
"~'-- J

~ i

Stotic

I0 II 12 13
DEFLECTION , mm
6

f-----Dynamic

I0

2 .'.'.'5 4

(a)

24l

ynomic

18

. ~..-~-

Dynomic

~,.~-t
St otic

t''"s
I 2

"',t
3 4

t
5

t
6

r
7

1 t
8 9

I
I0

t
II

12 13

DEFLECTION , m m

(c)

0
0

246

8 I0 12 14 16 18 20 22 24
DEFLECTION , mm
(d)

CONVENTIONALLY REINFORCED
CONCRETE

32

<12
o
J
9

4
0

,-,15

,--.,

<

STEEL FIBRE REINFORCED


CONCRETE

/ ~

21

14
z12
"~10
o
..j

(b)

POLYPROPYLENE FIBRE
REINFORCED CONCRETE

16

5 6 7 8 9 t0 II 12
DEFLECTION, rnm

l, o,~

28
24
20

<z 16
o 12
_d

4 r

I/

"-.....
V
"-..
4

Fig. 11--Results in terms of load (Pb) versus deflection


(Ue) plots for (a) normal-strength plain concrete, (b) highstrength plain concrete, (c) normal-strength polypropylene
fiber-reinforced concrete, (d) normal-strength steel fiberreinforced concrete, and (e) conventionally reinforced
normal-strength concrete

8 12 16 20 24 28 .32 36 40 4 4 4 8
DEFLECTION , m m

(e)

13. Banthia, N., Mindess, S. and Bentur, A., "'Energy Balance in


Instrumented Impact Tests on Plain Concrete Beams, "" The SEM-RILEM
Int. Conf. on Fracture of Concrete and Rock (June 1987).
14. Banthia, N., Mindess, S. and Bentur, A., "'Impact Behaviour of
Concrete Beams, "' Materials and Structures, 20, 293-302 (1987).
15. Mindess, S., Banthia, N. and Cheng, Y., "The Fracture Toughness
of Concrete Under Impact Loading," Cement and Concrete Res., 17,
231-241 (1987).
16. Mindess, S., Banthia, N. and Bentur, A., "'The Response of
Reinforced Concrete Beams with a Fibrous Concrete Matrix to Impact

Loading, "" Int. J. Cement Composites and Light Weight Concrete (UK),
8 (3), 165-170 (1986).
17. Bentur, A., Mindess, S. and Banthia, N., "'The Behaviour of
Reinforced Concrete Under Impact: The Effect of Concrete Strength, ""
The SEM-RIL.EM Int. Conf. on Fracture of Concrete and Rock (June 1987).
18. Banthia, N., Mindess, S. and Bentur, A., "'Steel Fibre Reinforced
Concrete Under Impact, "" Int. Symp. on Fibre Reinforced Concrete,
Madras, India (Dec. 1987).
19. Hibbert, A.P., "'Impact Resistance of Fibre Concrete, "" PhD
Thesis, Univ. of Surrey, UK (1979).

ExperimentaIMechanics 9 69

Potrebbero piacerti anche