Professional Documents
Culture Documents
4 123 PDF
4 123 PDF
Scientific paper
Abstract
Experimental results from tests on seven 650 mm deep large-scale reactive powder concrete (RPC) I-section girders
failing in shear are reported herein. The girders were cast using 150-170 MPa steel fiber RPC and were designed to as-
sess the capacity to carry shear stresses in thin webbed prestressed beams without shear reinforcement. The tests
showed that the quantity and types of fibers in the concrete mix did not significantly affect the initial shear cracking
load but increasing the volume of fibers increased the failure load. A design model is developed to calculate the strength
of the RPC beams tested in this study. The model is based on crack sliding and uses plasticity theory combined with
observations from the variable engagement model for mode I failure of fiber reinforced concrete. The results of the
model are compared with test data and show a good correlation.
Prestressing
Beam Force (kN) σtop σbot σave
Fiber
No. MPa MPa MPa Volume (%)
Top Bottom
Flange Flange
Type I- 1.5
SB5 225 450 -2.36 -13.6 -7.15
Type II- 1.0
Fig.2 Pouring of RPC in the steel forms.
SB6 225 450 -2.36 -13.6 -7.15 Type II- 2.5
3.3 Test setup and instrumentation
Details of the specimens, experimental variables, includ- Type I- 1.88
SB7 225 450 -2.36 -13.6 -7.15
ing levels of prestress, and the test setup and instrumen- Type II- 0.62
tation are given in Table 2 and Fig. 3, respectively.
Notes: σtop and σbot are the extreme fiber stresses of
Specimen SB3 was set as the reference specimen with
concrete at transfer and σave is the average prestress on
2.5 percent of Type I fibers and with 15 percent prestress
the section.
(i.e. the strands were stressed to a load corresponding to
15 percent of the guaranteed ultimate tensile strength of
the strand, GUTS). Specimens SB3 to SB7 were
prestressed to 15 percent of the GUTS of the strands to ensure a shear failure under load. While the prestress
giving an average prestress in the section of 7.15 MPa. per strand is relatively low, the total prestress is high
Each strand in specimens SB1 and SB2 was prestressed compared to that used in common practice for conven-
to 0 and 30 percent of the GUTS, respectively. Specimen tional prestressed concrete girders.
SB4 was similar to specimen SB3 except it had half the All the specimens had a similar experimental setup
amount of Type I fiber in its mix. Specimen SB5 and and instrumental gauging (see Fig. 3). Each specimen
SB7 contained varying quantities of Type I and II fibers was loaded by a single concentrated load applied at the
and beam SB6 contained only 2.5 percent of Type II mid-span of the specimens. All specimens were tested in
fibers. The relatively low prestress adopted, per strand, a 5000 kN capacity stiff testing frame and tested under
was so the beams maintained sufficient tensile capacity ram displacement control.
hydraulic jack
load cell
spreader beam
spherical seat
East 200 West
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
200
Demec gauge no. pin and roller
(gauge length = 250 mm) support
LVDT
pin support
2000 2000
One end of each specimen was a pinned support and Table 3 Mechanical properties.
the other end had a pin and roller support. The pins and
rollers were greased to minimize friction and to give free Specimen SB1 SB2 SB3 SB4 SB5 SB6 SB7
rotation and horizontal translation, as required.
Eo, GPa 44 45 43 43 49 40 46
The instrumentation used for each specimen is shown
in Fig. 3 and includes 99 sets of Demec gauges and a
ν 0.15 0.15 0.14 0.14 0.14 0.13 0.13
linear variable displacement transducer (LVDT). Demec
gauges 1 to 16 were located at the top (compressive) fcm (MPa) 161 160 149 164 171 157 169
flange of the specimen with Demec gauges 17 to 32 lo-
cated at the bottom (tensile) flange. Demec gauges 33 to fcu (MPa) 176 178 166 180 187 168 185
99 (not shown) were located in the web regions of the
specimens to measure longitudinal and shear strains in fsp (MPa) 19.2 20.9 21.9 18.0 22.4 18.3 23.5
the web (refer Voo et al., 2003).
The LVDT was used to measure the midspan dis- fdp (MPa) 11.9 11.2 10.6 10.3 13.6 10.2 11.1
placement of the beam. For each specimen, the test was
paused at preset load points (for the increasing load part fcf (MPa) 29.8 26.4 23.2 14.8 26.3 25.2 23.8
of the test) and the displacement held constant while
Demec gauge data were recorded and the crack pattern Gf (N/mm) 27.7 24.7 21.0 14.3 15.5 12.4 18.6
traced.
Flow (mm) 170 180 210 170 150 210 180
specimens SB1, SB2 and SB3 (the specimens with 2.5 ual tensile capacity across the crack sufficient to control
percent of 13 mm, straight fiber), the ratios of tensile the splitting and induce a bearing failure beneath the
strength to compressive strength are similar. A similar loading strips (Voo, 2004). In Table 4 the experimental
variation is also seen in the flexural tension strengths results are compared to that for specimens SB3 corrected
and fracture energies but, curiously, not in the split cyl- for the variation in compressive strength.
inder tension results. It is thought that the split cylinder In terms of the prestressing levels, the comparison
test does not give a reliable measure of the tensile presented in Table 4 indicates a 15 percent variation in
strength due to the effect of the fibres providing a resid- strength due to the effect of prestressing. In terms of the
WEST EAST
700 700
500 400 500 700 600 600 600 600 600 500 700 600 600 650
300 500 400 500 500
400 400 500 600
600 400
500 400
300 400 600 600 600
600
400 500 500 300 400 400
300 500 300
SB1
WEST EAST
800 700 900 800 800
900 700 900 900800 750 900 900
700 600 600 750 800 900
800 750
800 700 900 600 900
700 600 600
600 700 500 600 500 750
750 700 700
900 600 700 900 600
SB2
WEST EAST
600 700 760 600 600 500 500
700 500 600
600 600 600 500 600 500 600 700 600 500 400 400 600
500
700 400 300 300 500 600 500 500
600 400 400 400
500 300 300 300 400 500 600
300
400 400 400
SB3
WEST EAST
600 500 500 500
600 600 600 400500 500 500 400 400400
400 400400 300 400
450 400 400
400 500 500
500 400 300 500
600 300
500 500500 400
400
500 400400 400
400
600 600 300
400 300 300
SB4
WEST EAST
700
800 800 800 700 700 700 500 700 700 700 700 700 700 800 800 700
700 700 700 600
600 700
500
600 500 400
600 600 450
600 600 500
SB5
WEST EAST
600 450 450500500 400 400 500 500 500 500
500 400 400 450 600 400 400 400 400 400 400
400 400 450400 200 300 300 400 450 500
400 500 450 450
450
200 450 400
600 500 600
400 300 300 600 450 450400
300
450 400 300 300
SB6
WEST EAST
790 790 700 700 600 700700 700 700 700 700 700 700 700 700 700 700
700
790
600 600 600600600 600 600 600 600 600 700 600 600 600 600
600 600 600 500
500 600
500 500 500 500 400
500 400 500 500 500
500 400 400 400 500
400 400
400 400 500 500
SB7
1200
Distance from Centre Line (mm)
SB 2 -1875 -1125 -375 375 1125 1875
1000 -800
SB 1
SB 3
800 Load(kN)
CL 0
-600
200
Strain (µε)
600 SB 4 400
)
500
-400 600
Load (kN)
L
400 700
800
d (kN)
Strain (
850
200 -200
0 0
0 20 40 60 80 100 120 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Midspan Displacement (mm) Demec Position
(a) (a)
1200 Distance from Centre Line (mm)
-1875 -1125 -375 375 1125 1875
1000 4000
SB 3 SB 5
SB 7 3000 Load (kN)
800 CL
0
Load (kN)
SB 6 200
Strain (µε)
500
600
400 1000 700
800
Strain (
850
200 0
0 -1000
0 20 40 60 80 100 120 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
Midspan Displacement (mm) Demec Position
(b) (b)
Fig. 8 Strains in beam SB3 (including prestrain): (a) top
Fig. 7 Load versus midspan deflection of (a) SB1 to SB4 flange (Demecs 1 to 16); (b) strain in bottom flange
and (b) SB3, and SB5 to SB7. (Demecs 17 to 32).
This indicates that crushing of the top flanges was not an of the cement-fine aggregate matrix with high tensile
issue. The maximum tensile strains at the bottom flange strengths maintained for large crack openings. Thus,
were 2500 µε to 4500 µε indicating that the strands re- plasticity is an appropriate approach for the design of
mained in the elastic range and yielding of the tension such structural elements.
reinforcement was also not an issue (strand yield strain, In the crack sliding model, concrete is treated as a
εpy = 8970 µε). The strains in the top and bottom flanges Mohr-Coulomb material with a zero tension cut-off. A
of SB3 are given in Fig. 8 and are typical of the data more comprehensive description of the theory can be
obtained. found in Johansen (1958), Sandbye (1965) and Nielsen
(1967). According to the crack sliding model, the crack-
6. Shear strength – plasticity model ing of concrete introduces a potential yield-slip line
which, due to a reduced sliding resistance, forms the
Plasticity approaches came to the forefront of reinforced critical failure mechanism.
concrete design with work published by Nielsen (1963, For both non-prestressed and prestressed simply sup-
1967). Based on the upper bound theory of plasticity, ported rectangular beams with rectangular cross-sections
Zhang (1994) developed the crack sliding model to cal- and loaded with two symmetrically located point loads,
culate the shear strength of flexurally over-reinforced the ultimate shear strength of the section can be deter-
rectangular beams without shear reinforcement. In the mined by
development of the model that follows the flange out-
stands of the I-girders are ignored for the purpose of ⎛ 2 ⎞
1 * ⎜ ⎛x⎞ x⎟
assessing the shear capacity. The crack sliding model is Vu = fc b h ⎜ 1 + ⎜ ⎟ − ⎟ (1)
selected in this study as the use of high quantities of fi- 2 ⎜ ⎝h⎠ h⎟
⎝ ⎠
bers leads to a relatively plastic response after cracking
130 Y. L. Voo, S. J. Foster and R. I. Gilbert / Journal of Advanced Concrete Technology Vol. 4, No. 1, 123-132, 2006
Vcr =
1 * h +x
ft b +
2 2 ∑ Pe d p
i (2)
(b)(b)
2 a a
Fig. 9 Simply supported beam with critical diagonal
where dpi is the distance of the effective prestressing crack: (a) yield line; (b) cracking load.
force (Pe) at the ith level from the top surface of the
beam, a is the shear span and ft* is the effective tensile
strength of the concrete.
The solution for x is obtained by equating Vu given by V
Eq. 1 to Vcr given in Eq. 2. That is, the solution of x is
Vu (Eq. 1) Prestresssed
the intersection of Eqs. 1 and 2 (Fig. 10) giving Beams
⎛ 2 ⎞
⎜ ⎛ x⎞ x⎟
f c* ⎜ 1 + ⎜ ⎟ − ⎟ =
⎜ ⎝h⎠ h⎟
⎝ ⎠ Non-Prestresssed
Beams
(x2 + h2 ) + 2∑ Pi d pi
Pe d pi / a
f t * = ν t f tf (4) 8 σ (MPa)
Range of COD at the
f approximate peak load
where νt is the tensile effectiveness factor and ftf is the tf for FR-RPC girders
6
maximum tensile stress of the fiber component and can in this study
be calculated using a model such as variable engagement
model (VEM) of Voo and Foster (2003, 2004), for ex- 4
ample.
From the experimental data it was observed that at the
peak load the crack widths for the SFR-RPC experimen- 2
tal girders tested in this study were in the range of 0.5 to
1.5 mm. From the VEM the effective tensile strengths w (mm)
0
(ft*) in this range for the 13 mm long by 0.2 mm diame- 0.5 1.5 6.5
ter fibers used were in the range of 70 to 90 percent of
the peak fiber contribution to the tensile strength (Fig. Fig. 11 Range of tensile stress across failure surface for
11). In the analyses that follow, the effective tensile SB1 to SB3.
strength of the fiber reinforced composite is calculated
using Eq. 4 with νt = 0.8.
The effective compressive strength (fc*) takes into ac-
count the cohesion and friction along the crack during f c * = ν c f cm (5)
sliding (refer Zhang, 1994) and is taken as where fcm is the mean cylinder compressive strength and
Y. L. Voo, S. J. Foster and R. I. Gilbert / Journal of Advanced Concrete Technology Vol. 4, No. 1, 123-132, 2006 131
the compression effectiveness factor is taken as νc = 0.8. and the coefficient of variation is 0.095.
Further justification of this value, however, is needed for While it is shown that the rectangular section model is
concretes containing relatively high fiber volumes and capable of calculating the shear strength of the SFR-RPC
this is a matter for further research. prestressed girders tested, the angle of the failure crack
In the calculation of x the limits are 0 ≤ x ≤ a. The so- is, generally, flatter than that of the experiments. Voo et
lution to Eq. 3 is not readily obtained analytically. In- al. (2003) show that a crack sliding model refined to
stead, a numerical method is used to find the critical include the flanges can improve the calculation of the
value of x/h. In this way the starting position of the criti- crack sliding angle. The model is, however, substantially
cal diagonal crack for a given a/h is found and, conse- more complex and gives only small differences in the
quently, the ultimate load and the shear capacity can be calculated shear strengths, as shown in Table 5.
determined from Eq. 1. Finally, the ultimate shear
strength is written as 7. Conclusions
τ u = Vu (bh ) (6) Seven reactive powder concrete prestressed girders
without stirrups were tested to study the capacity of fiber
The shear strengths calculated using the plasticity
reinforced RPC beams in shear. The test variables were
model described above for the SFR-RPC shear beams
the quantity and type of fibers and the prestress. The
tested in this study are compared with the experimental
steel fibers used in the tests consisted of either 13 mm
data and the results are summarized in Table 5. The
straight fibers and/or 30 mm end-hooked fibers. All the
prestressing force used in the calculations is that given in
tested specimens had the same cross-section and were
Table 2 without losses. Although the level of prestress,
subjected to mid-point loading over a shear span of 2
in absolute terms was relatively high, the prestress in
meters. The shear span to effective depth ratio for the
each strand was low and below the threshold for relaxa-
beams was 3.33.
tion of the tendons. Also, while the average compressive
From the experimental study the following conclu-
stresses on the section were high relative to current de-
sions are drawn:
sign practice, they are not high relative to the compres-
1. The quantity of fibers and type of fibers used in the
sive strength of the RPC. For these reasons the losses are
concrete mix do not significantly affect the cracking
small and not significant to the strength calculations. The
load but have a significant influence on the rate of
mean model to experimental shear capacity ratio is 0.89
Table 5 Comparison of the model shear strength calculations with the experimental data.
Beam
kN deg. MPa MPa MPa Vu,theo θtheo Model Vu,theo θtheo Model
kN deg. Exp. kN deg. Exp.
Notes: # fiber contribution to the peak tensile strength (ftf ) is calculated using the variable engagement model of Voo and
Foster (2003, 2004).
132 Y. L. Voo, S. J. Foster and R. I. Gilbert / Journal of Advanced Concrete Technology Vol. 4, No. 1, 123-132, 2006
crack propagation and on the failure loads. Johansen, K.W. (1958). “Brudbetingelse for sten og
2. At the peak load, many fine cracks had formed in the beton (Failure criteria for rock and concrete).”
web, with the cracks well distributed through the Bygningsstatiske Meddelelser, 29 (2), 25-44.
shear spans. The failure loads were more than twice Nielsen, M. P. (1963). “Yield condition for reinfroced
the cracking loads. concrete shells in the membrane state.” Non-classical
A design model is also developed to calculate the Shell Problems (ed. Olszak and Sawczuk), Proc. IASS
strength of steel fiber reinforced reactive powder con- Symposium, Warsaw, Amsterdam, 1030-1040.
crete beams failing in shear. The model is based on the Nielsen, M. P. (1967). “Om forskydningsarmering i
crack sliding model of Zhang (1994) combined with jernbetonbjælker- On shear reinforcement in
observations from the variable engagement model of reinforced concrete beams.” Bygninsstatiske
Voo and Foster (2004) for mode I fracture of fiber rein- Meddelelser, 38 (2), November, 33-58.
forced concrete. The results of the model are compared Richard, P. and Cheyrezy, M. (1994). “Reactive powder
with the tests reported herein and a good correlation was concretes with high ductility and 200-800 MPa
observed for the strengths calculated by the design compressive strength.” ACI, SP-144(24), San
model when compared to the test data. However, the Francisco, USA, 507-518.
model, generally, calculates a flatter crack sliding plane Richard, P. and Cheyrezy, M. (1995). “Composition of
compared to that observed in the tests. To better deter- reactive powder concretes.” Cement and Concrete
mine the crack sliding angle a more refined model is Research, 25 (7), 1501-1511.
required that includes the flange effects. Sandbye, P. (1965). “A plastic theory for plain concrete.”
Bygningsstatiske Meddelelser, 36 (2), 41-62.
Acknowledgements Voo, J. Y. L. and Foster, S. J. (2003). “Variable
This project was undertaken in the Heavy Structural engagement model for fibre reinforced concrete in
Laboratory of University of New South Wales and it was tension.” UNICIV Report R-420, School of Civil and
supported by an Australian Research Council Strategic Environmental Engineering, The University of New
Partnerships with Industry Grant 2001-2003 in collabo- South Wales, June, 86 pp.
ration with VSL (Aust). This support is acknowledged Voo, J. Y. L. and Foster, S. J. (2004). “Tensile fracture of
with appreciation. fibre-reinforced concrete: Variable engagement
model.” 6th RILEM Symposium on Fibre-Reinforced
References Concretes (FRC) – BEFIB, 20-22 September, Varenna,
AS3972 (1997). Portland and Blended Cements. Italy, 875-884.
Standards Australia. Voo, J. Y. L, Foster S. J. and Gilbert R. I. (2003). “Shear
ASTM C230-03 (2003). “Standard Specification for Strength of Fibre Reinforced Reactive Powder
Flow Table for use in Tests of Hydraulic Cement.” Concrete Girders without Stirrups.” UNICIV Report
Aїtchin, P. C. (2000). “Cement of yesterday and today R-421, School of Civil and Environmental
concrete of tomorrow.” Cement and Concrete Engineering, The University of New South Wales,
Research, 30 (9), 1349-1359. November, 131 pp.
Cavill, B. and Chirgwin, G. (2003). “The worlds first Voo, J. Y L. (2004). “An investigation into the behaviour
RPC road bridge at Shepherds Gully Creek, NSW.” of prestressed reactive powder concrete girders
21st Biennial Conference of the Concrete Institute of subjects to non-flexural actions.” PhD Thesis, School
Australia (CIA), July, 89-98. of Civil and Environmental Engineering, The
Chen, W. F. and Yuan, R.L. (1980). “Tensile strength of University of New South Wales, Sydney, Australia.
cconcrete: Double-punch test.” Journal of the Zhang, J. P. (1994). “Strength of cracked concrete: Part 1
Structural Division, ASCE, 106, 1673-1693. – Shear strength of conventional reinforced concrete
Gilbert, R. I., Gowripalan, N. and Cavill, B. (2000). “On beams, deep beams, corbels, and prestressed
the design of precast, prestressed reactive powder reinforced concrete beams without shear
concrete (Ductal) girders.” Proceedings of 4th reinforcement.” Report No. 311, Technical University
Austroads Bridge Engineering Conference, Adelaide, of Denmark, Department of Structural Engineering,
Australia, Nov 30 - Dec 1, 3, 313-324. Lyngby, 106 pp