Você está na página 1de 9

ACI STRUCTURAL JOURNAL

TECHNICAL PAPER

Title no. 106-S48

Anchorage of Large-Diameter Reinforcing Bars in Ducts


by Kyle P. Steuck, Marc O. Eberhard, and John F. Stanton
In precast concrete systems, connections are often made by
grouting bars that project from one member into ducts embedded
in another. For bridge bents, bar-duct systems can be assembled
rapidly if a few large bars and ducts are used to connect the
column and cap beam. In some cases, the required anchorage
lengths for the large bars exceed the length available. To evaluate
the anchorage requirements for this situation, 14 pullout tests were
performed on bars with sizes up to No. 18. The tests and a
nonlinear finite element model showed that, under conditions
similar to those tested, large bars can develop their yield and
fracture strengths in as few as six and 10 bar diameters,
respectively. The effects of the bar size were found to be small
compared to the scatter among the test results. Parallel tests with
polypropylene fibers showed that fibers generally decreased
pullout resistance, although this is likely the result of reduced
grout strength.
Keywords: anchorage; bond; duct; joint; precast; reinforcement.

INTRODUCTION
Precast components have often been used in reinforced
concrete structures to accelerate the on-site construction
process (AASHTO and FHWA 2004; Hieber et al. 2005a;
FHWA 2007). Precast components have been used mainly
for nonseismic components, such as girders in bridges and
gravity frames, and architectural elements in buildings
(Billington et al. 1999; PCI 2004, 2007). More recently,
precast components have been used to resist seismic
loads in buildings (Nakaki et al. 1999) and bridges (Fig. 1).
The main challenge of designing earthquake-resistant
systems with precast members has been the development of
connections that have good seismic performance and that are
easy to construct (Sritharan 2005; Hieber et al. 2005b;
Stanton et al. 2006).
In bridge bents, one method of connecting a column to a
cap beam is to precast the column with projecting bars,
which are then grouted into ducts in the cap beam, as shown
in Fig. 1. If a few large bars are grouted into a few large
ducts, the construction process is particularly simple because
of the large tolerances available and the small number of barduct alignments that are needed (Hieber et al. 2005b).
In this system, grouted bars carry tensile forces across the
interface between the column and cap beam. The need for
ductility in the system dictates that the bars must undergo
ductile yielding before other failure mechanisms, including
anchorage failure, occur. A challenge to implementing this
system in practice is that the development lengths required
by current specifications (AASHTO 2005; ACI Committee 318
2005) can exceed the available length in the cap beam.
Many researchers have investigated the development of
reinforcing bars in concrete (Orangun et al. 1977; Jirsa et al.
1979; Eligehausen et al. 1983), but only a few researchers
have considered the development of bars grouted into ducts.
Darwin and Salamizavaregh (1993) and Moosavi and
Bawden (2003) studied the behavior of grouted connections
506

with unlined holes, in which there was no duct. Raynor


(2000) and Brenes et al. (2006) studied ducts with relatively
small duct/bar diameter ratios (3.1 and 2.9, respectively).
Brenes et al. (2006) reported that, in some cases, the ducts
pulled out of the surrounding concrete. This failure mechanism
would be unlikely for larger duct/bar ratios, because the
shear stress on the curved duct-concrete interface is inversely
proportional to the duct diameter. For the tests in the current
study, the duct/bar ratio was 3.6.
The most relevant research to the application considered in
this paper was conducted by Raynor et al. (2002), who
measured the bond-deformation relationships for smaller
bars grouted into ducts. They concluded that high bond
stresses could be developed at the bar-grout interface,
making it possible to reduce anchorage lengths from those
required by current specifications (AASHTO 2005; ACI
Committee 318 2005). Raynor et al. (2002) did not consider
whether these results could be extrapolated to large bars,
whether fibers improved the anchorage conditions, or the
effects of the free grout surfaces.
To investigate these effects, 14 pullout tests were conducted
on large-diameter reinforcing bars grouted into corrugated
steel ducts (Steuck et al. 2008). These tests provided measurements of the anchorage capacity of bars of various sizes and
embedment lengths, and with and without polypropylene fiber
reinforcement. The tests also provided a basis for developing
a nonlinear bond-slip model and for recommending new
anchorage requirements for bars grouted in ducts.

Fig. 1Precast cap beam erection.


ACI Structural Journal, V. 106, No. 4, July-August 2009.
MS No. S-2007-414 received December 14, 2007, and reviewed under Institute
publication policies. Copyright 2009, American Concrete Institute. All rights reserved,
including the making of copies unless permission is obtained from the copyright proprietors.
Pertinent discussion including authors closure, if any, will be published in the May-June
2010 ACI Structural Journal if the discussion is received by January 1, 2010.

ACI Structural Journal/July-August 2009

Kyle P. Steuck is an Engineer with Degenkolb Engineers, Seattle, WA. He received


his BS in civil engineering from the University of Texas at Austin, Austin, TX, in 2004,
and his MS in civil engineering from the University of Washington, Seattle, WA, in 2007.
Marc O. Eberhard, FACI, is a Professor at the University of Washington. He
received his BS in civil engineering, and materials science and engineering from the
University of California-Berkeley, Berkeley, CA, and his MSCE and PhD from the
University of Illinois at Urbana-Champaign, Urbana, IL. He is a member of Joint
ACI-ASCE Committee 445, Shear and Torsion.
John F. Stanton, FACI, is a Professor at the University of Washington. He received
his BA and MA from Cambridge University, Cambridge, UK; his MSCE from Cornell
University, Ithaca, NY; and his PhD from the University of California-Berkeley. He is
a member of ACI Committee 318E, Shear and Torsion (Structural Concrete Building
Code), and Joint ACI-ASCE Committee 550, Precast Concrete Structures.

RESEARCH SIGNIFICANCE
The deployment of easily constructible precast systems in
seismic regions is hampered by the requirement that large
bars have long anchorage lengths. This paper describes
14 pullout tests of large bars grouted into ducts. Based on the
results of these tests, an analytical model is proposed and
calibrated that reproduces the observed behavior well. The
tests and analyses provide a basis for greatly reducing the
required anchorage lengths of large bars grouted into ducts.
EXPERIMENTAL PROGRAM
Test specimens and setup
An experimental study was conducted to evaluate the
effects of the bar size, the embedment length, and the presence
of fibers on the pullout resistance of bars grouted into ducts.
The test program included bar sizes ranging from No. 10 to
No. 18 and embedment lengths ranging from two to 14 bar
diameters. To evaluate the effects of fiber reinforcement, a
parallel series of tests with polypropylene fibers was
conducted. As shown in Table 1, each test was denoted by a
five-digit code (nnAll) to indicate the bar size (nn = No. 10,
14, or 18), type of fiber reinforcement (A = F for tests with
fibers or N for tests without fibers), and nominal embedment
length (ll = 2, 4, 6, 8, or 14 bar diameters).
The pullout test setup is shown in Fig. 2. The test specimen
consisted of a 36 in. (915 mm) diameter, 42 in. (1067 mm)
high concrete cylinder in which a central duct was
embedded. A reinforcing bar was grouted into the duct, and
after approximately 5 days, a center-hole ram was used to
pull the bar out of the duct. A reaction block, consisting of an

18 in. (457 mm) high cylinder with a conical internal void,


distributed the load to a 3 in. (76 mm) wide annulus on the
outer edge of the top surface of the concrete cylinder. This
distribution of load was selected to avoid inducing compression
stresses in the bonded region, which is known to affect the
results of pullout tests (ACI Committee 408 2003).
With the exception of the test that had an embedment
length of 14 bar diameters (Test 18N14), the embedment
lengths were less than the full length of the duct. Several
tests were performed with each of the other three concrete
cylinders. Tests were performed in the following sequences:
Specimen 1: Tests 18N02, 18N04, 18N03, and 18N08;
Specimen 2: Tests 18F02, 18F04, 18F03, and 18N06; and
Specimen 3: Tests 10N04, 10F04, 10N06, 10N08, and 14N08.
After each test, loose grout was removed from the duct and a
smooth, level surface was formed by pouring gypsum cement
into the duct on top of the grout that remained from the
previous test. The grout for the new test was then placed on top
of the hardened gypsum cement, so later tests occurred closer
to the top surface of the concrete. Further details on the location
of each test are described in Steuck et al. (2008).

Fig. 2Anchorage test setup.

Table 1Test parameters and peak loads


avg f g ,
Displacement at
, ksi (MPa) avg, psi (MPa) psi/psi (MPa/MPa) Pmax, in. (mm)

Test

Bar size

db , in. (mm)

le , in. (mm)

le/db fg, psi (MPa)

P, kip (kN)

10N04

No. 10

1.27 (32.3)

4.25 (108)

3.35

7980 (55.0)

40.2 (18.2)

31.7 (219)

2370 (16.3)

26.3 (2.19)

10F04

No. 10

1.27 (32.3)

6.31 (160)

4.97

7630 (52.6)

33.1 (15.0)

26.1 (180)

1310 (9.03)

19.2 (1.59)

0.098 (2.5)

10N06

No. 10

1.27 (32.3)

7.63 (194)

6.00

8270 (57.0)

88.0 (39.9)

69.3 (478)

2890 (19.9)

32.8 (2.72)

0.052 (1.3)

10N08

No. 10

1.27 (32.3)

9.75 (248)

7.68

8000 (55.2)

112.3 (50.9)

88.4 (610)

2890 (19.9)

33.3 (2.77)

0.150 (3.8)

14N08

No. 14

1.693 (43.0)

13.13 (334)

7.75

8220 (56.7)

178.7 (81.1)

79.4 (547)

2560 (17.7)

28.2 (2.34)

18N02

No. 18

2.257 (57.3)

4.38 (111)

1.94

8370 (57.7)

29.4 (13.3)

7.3 (50)

950 (6.55)

10.4 (0.87)

0.052 (1.3)

0.074 (1.9)

18F02

No. 18

2.257 (57.3)

4.63 (118)

2.05

7270 (50.1)

37.5 (17.0)

9.4 (65)

1140 (7.86)

13.5 (1.12)

0.048 (1.2)

18N03

No. 18

2.257 (57.3)

6.19 (157)

2.74

7770 (53.6)

142.5 (64.6)

35.6 (245)

3250 (22.4)

34.5 (2.87)

0.120 (3.0)

18F03

No. 18

2.257 (57.3)

7.19 (183)

3.18

7520 (51.9)

119.7 (54.3)

29.9 (206)

2350 (16.2)

27.6 (2.29)

0.045 (1.1)

18N04

No. 18

2.257 (57.3)

9.50 (241)

4.21

8510 (58.7)

186.2 (84.5)

46.6 (321)

2760 (19.0)

30.9 (2.56)

0.112 (2.8)

18F04
18N06

No. 18
No. 18

2.257 (57.3)
2.257 (57.3)

9.38 (238)
13.81 (351)

4.15
6.12

6800 (46.9) 167.0 (75.8)


9890 (68.2) 301.0 (136.5)

41.8 (288)
75.3 (519)

2510 (17.3)
3070 (21.2)

31.0 (2.57)
31.0 (2.57)

0.149 (3.8)
*

18N08

No. 18

2.257 (57.3)

18.00 (457)

7.98

8900 (61.4) 358.9 (162.8)

89.7 (618)

2810 (19.4)

28.2 (2.34)

0.803 (20.4)

18N14

No. 18

2.257 (57.3)

31.50 (800) 13.96 10,300 (71.0) 420.0 (190.5) 105.0 (724)

1880 (13.0)

18.8 (1.57)

Peak load reached after displacement readings became unreliable.

ACI Structural Journal/July-August 2009

507

The concrete remained undamaged in all tests except Tests


18N08 and 18N14, which were conducted near the surface of
the specimen. The large forces applied during these tests
caused minor cracking of the concrete surrounding the duct.
The conditions outside the duct were therefore assumed to
have no influence on the anchorage resistance.
The reinforcing bars were ASTM A706 Grade 60, with
bamboo-style deformations. For tests in which the bar
remained elastic, the section of the bar that had been
embedded in grout was cut off, and the remainder of the bar
was used for the next test. Yielded bars were not reused.
Previous researchers have considered anchorage of
reinforcing bars within post-tensioning (PT) ducts
(Raynor et al. 2002; Brenes et al. 2006). PT duct is available up
to an outer diameter of 6 in. (152 mm), which is too small
to accommodate a No. 18 bar and the desired construction
tolerance of 3 in. (76 mm). This study used 8 in. (203 mm)
nominal diameter, corrugated, galvanized steel pipe, with
a nominal thickness of 0.064 in. (1.6 mm), and corrugations
with a peak-to-peak amplitude of 0.25 in. (6.3 mm).
Corrugated pipes are available in diameters from 6 in.
(152 mm) to 12 ft (3.7 m). The helical corrugations of

Fig. 3Stress-displacement plots: effect of embedment


length. (Note: Displacement measurements became
unreliable for Tests 18N06 and 18N14 before peak loads
were reached.)

such pipe are deeper and the bond properties are potentially
better than those of standard PT duct.
The same nonshrink grout was used for each test (Steuck et al.
2008). The grout was proportioned by converting the
manufacturers volumetric instructions to a water-to-grout
weight ratio of 0.15. The researchers tested each pullout
specimen at 5 days, when the compressive strength of the
grout without fibers was approximately 8500 psi (59 MPa).
The four fiber-reinforced specimens had a lower average
compressive strength of 7300 psi (50 MPa) at 5 days. The
0.5 in. (13 mm) long polypropylene fibers were added to the
mixture at a dosage of 3 lb/yd3 (1.8 kg/m3), which was twice
the manufacturers suggested value. This dosage was
selected to be consistent with previous research (Cheok and
Stone 1994; Raynor et al. 2002).
The reinforcing bars were loaded monotonically at a rate of
approximately 0.1 in. (2.5 mm) per minute (with unloading
when the ram stroke was reached) to failure. The post-peak
response was recorded to the limits of the measuring devices,
approximately 1.2 in. (30 mm). Such monotonic tests have been
found to approximate the envelope curve of cyclic tests
(Eligehausen 1983).
Test results
The stress-displacement responses of tests on No. 18 bars
in grout without fibers are compared in Fig. 3. The axial bar
stress is based on the nominal bar area. The reported
displacement is the movement of the bar at the free grout
surface relative to a nominally fixed point on the concrete
cylinder surface 10 in. (254 mm) from the axis of the bar.
During the test, the grout surface was inaccessible, so the bar
displacement there was calculated by subtracting the elongation
of the exposed part of the bar (calculated from strain
measurements) from the bar displacement measured 2 in.
(51 mm) above the surface of the concrete cylinder. Steuck
et al. (2008) provides detailed results for all of the tests.
Elastic analysis of the shear deformation of the grout and
concrete under the applied loads showed that slip of the bar
through the grout is the primary mechanism for bar displacement.
For the tests with short embedment lengths (2, 3, and 4 bar
diameters), a typical stress-displacement curve consisted of
four regions: 1) an approximately linear region; 2) a region
of approximately constant stress, near the peak stress; 3) a
region of rapid drop in stress between displacements of
approximately 0.1 and 0.2 in. (2.5 and 5 mm); and 4) a region in
which the stress dropped more slowly until the anchorage
provided no further resistance. Because the bar remained elastic
in these tests, the response was dominated by the bond properties.
By contrast, the behavior of the longer embedment specimens (6,
8, and 14 bar diameters) was dominated by bar yielding.
The peak resistances are reported in Table 1. At an
anchorage length of six bar diameters, the bars reached a
tensile stress of 75 ksi (519 MPa), which exceeded the bars
yield strength of 65 ksi (450 MPa). At an anchorage length
of 14 bar diameters, the bar fractured.
Equilibrium provides a relationship between bar tensile
stress and average bond stress.
l
= 4 ave ----edb

Fig. 4Effect of embedment length on peak bar stress.


508

(1)

If the peak average bond stress were constant among the


tests, the peak bar stress would vary linearly with the normalized
embedment length le /db. As shown in Fig. 4, most of the
ACI Structural Journal/July-August 2009

results followed this trend, with the exception of Test


18N14, in which the capacity was limited by the bar strength.
The trend line does not pass through the origin, however,
which implies that the local bond resistance near the free
surface of the grout is lower than elsewhere.
Damage observations
The damage to the grout and bar was observed for each test
after the reaction block had been removed. In each test, a
cone of grout formed at the free surface, as shown in Fig. 5.
Below this cone, a cylinder of grout with the same diameter
as the reinforcing bar lugs pulled through the surrounding
grout. The existence of these separate mechanisms was
attributed to the difference in the confinement provided by
the duct and surrounding concrete. These two mechanisms
have also been reported for pullout tests on bars embedded
directly in concrete confined by transverse bar reinforcement
(Viwathanatepa et al. 1979; Eligehausen et al. 1983).
The observed semi-angle of the cone varied between
approximately 45 and 60 degrees to the bar axis. The angle
of the cone was similar for each test regardless of bar size,
indicating that the thickness of the grout medium, rather than
the size of the bar, controlled the length of the cone.
The shear-cylinder mechanism has been reported by other
researchers (for example, Moosavi and Bawden 2003),
although most sources refer to crushing of the concrete
around the lugs as the failure mechanism. Crushing of the grout
was not visible in these tests, but may have been masked by the
subsequent shearing along the cylindrical surface.
DISCUSSION OF TEST RESULTS
Bar size
The effect of the bar size is illustrated in Fig. 6, which
shows stress-displacement responses from tests with No. 10,
14, and 18 bars with the same normalized embedment length
of eight bar diameters. To compare results from tests
conducted with different bar sizes, the displacement was
normalized by the bar diameter. The peak bar stress in
Test 14N08 was 11% lower than those in Tests 18N08 or
10N08, which in turn differed by only 1%. This conflicting
evidence suggests that the effects of bar size, if present, are
small within the range of bar sizes tested and compared with
the scatter among test results.
Fibers
The presence of fibers in the grout reduced the compressive
strength of the grout for each pair of parallel tests conducted
with and without fibers. Fibers also reduced the peak average
bond stress for three of the four pairs. When the bond
strengths were normalized by the square root of the grout

Fig. 5Cone and strut diagram.


ACI Structural Journal/July-August 2009

compressive strength, the peak resistance of the specimens


with and without fibers was similar on average (Table 1).
In the post-peak region, the fibers had been expected to
improve the behavior by bridging cracks in the grout. This
enhancement was modest and was observed in only two of
the four pairs of tests (Fig. 7).
Yielding
The bar yielded in the tests with embedment lengths of 6db
or more. Inelastic elongation is accompanied by a reduction
in bar diameter, which causes the lugs to partially disengage
from the surrounding grout, thereby reducing the bond
capacity. Others (Eligehausen et al. 1983; Raynor 2000)
have observed this behavior and have attempted to quantify
it. Whereas such behavior may have occurred in these tests,
the data associated with yielding bars were insufficient to
quantify it.
Confined and unconfined regions
The trend of bar stress with embedment length (Fig. 4)
suggests that the bond stress was lower near the free grout
surface than at depth. Such an observation is consistent with
the strut mechanism illustrated in Fig. 5. Prior to cracking,
the grout can carry shear and tensile stresses. After formation

Fig. 6Stress-displacement plots: effect of bar size. (Note:


Displacement measurements became unreliable for Test
14N08 before peak load was reached.)

Fig. 7Stress-displacement plots: effect of fibers.


509

of the crack that separates the end cone, the grout within the
cone can provide resistance only by aggregate interlock
across the crack. By contrast, at greater depths, the grout is
better confined and a conical pattern of struts can form. The
radial component of the strut force is resisted by hoop stress
in the duct. Potential failure modes include tie (that is,
circumferential duct) yielding, strut crushing, and shear
failure in the grout at the nodes between the bar and the
grout. With the exception of the one failure by bar fracture,
grout shear failure was observed in each test. After the shear
failure initiated, some residual resistance was provided by
shear friction and aggregate interlock.
Duct
The duct appears to benefit the anchorage resistance, but
the precise mechanism by which it does so is unclear. Strain
gauge data showed that the duct did not yield during
Test 18N14. Raynor (2000) used much smaller concrete
blocks and thinner ducts, but also found low hoop stresses in
the duct and negligible cracking of the block. Furthermore,
an elastic analysis of concentric thick-walled cylinders
shows that even the relatively thick-walled ducts used herein
contribute negligibly to the total hoop stiffness compared to
the concrete cylinder. The value of the duct may lie less in its
hoop stiffness than in some other role, such as an arrestor of
radial cracks in the grout.
The duct did not slip relative to the concrete. Due to the
large duct diameter, the highest average bond stress applied
at the bar-duct interface during the tests was 500 psi (3.4 MPa).
FINITE ELEMENT ANALYSIS
Model development
A nonlinear finite element model was developed to simulate
the observed force-displacement behavior. The model
consists of a series of one-dimensional, nonlinear bar elements
attached to one-dimensional, nonlinear bond springs, as
shown schematically in Fig. 8. The model was based on the
assumption that slip and consequent damage to the bond
interface at one location do not affect the stress-slip behavior
at another location. This assumption would be violated if, for
example, the grout suffered a longitudinal splitting failure.
The steel constitutive model, based on one developed by
Raynor et al. (2002), is shown in Fig. 9. The material parameters
for the steel model were taken from the average results of
two tests of bare bars. The bond springs in the confined and
unconfined regions were defined by two different bond-slip
models. The unconfined model was applied to a length at the
loaded end of the bar equivalent to a 45-degree cone, and the
confined model was applied to the remainder of the
embedded length.
For a given set of specimen properties and a specified
unloaded-end slip, the stress, strain, and slip at each node can
be determined without iteration. Each test was simulated by
applying a displacement to the unloaded end of the bar, then
calculating the bond stress, stress, strain, and displacement at
each node along the bar, including at the loaded end. By

Fig. 8Schematic of nonlinear model.


510

incrementing the slip at the unloaded end of the bar, the


complete force-displacement curve at the loaded end
was developed.
The properties of the bond models were calibrated by
minimizing the peak stress error, Epeak, and envelope error,
Eenv, between the measured and computed results. The peak
stress error was defined as
N tests

( i

TEST

iFEM )

i=1
E peak = ---------------------------------------------------N tests f y

(2)

The envelope error was defined as


Npoints
N tests

E env

( i, jTEST i, jFEM )

j=1
--------------------------------------------------------N points
i
=
1
= ---------------------------------------------------------------------------N tests f y

(3)

To eliminate bias, stresses were obtained by interpolation


at 100 uniformly spaced displacement values for both
measured and simulated data.
The resulting calibrated bond models are plotted in Fig. 10.
The unconfined bond model is triangular. The bond stress
increases linearly with increasing normalized slip to a peak
stress of 9 fc psi (0.8 fc MPa) at a normalized slip of 0.02.
The bond stress then decreases linearly to a stress of zero at
a normalized slip of 0.04 and is zero thereafter.
The confined bond model is multilinear. The bond stress
increases linearly to a plateau at 41 fc psi (3.4 fc MPa).
The plateau extends from normalized slips of 0.025 to 0.055.
The normalized bond stress then decreases to 19fc psi
(1.6fc MPa) at a normalized slip of 0.1, then to 1.6fc psi
(0.13fc MPa) at a normalized slip of 0.5. Beyond that, it
remains constant.
Results
Measured and simulated stress-displacement plots for
Tests 18N04 and 18N14 are shown in Fig. 11. The full stressdisplacement curves for both tests were well reproduced.
Figure 12 shows that the measured and calculated peak bar

Fig. 9Assumed steel constitutive model.


ACI Structural Journal/July-August 2009

stresses were similar. The root mean square error is 9.6 ksi
(66 MPa) for all tests from this experimental program and
5.7 ksi (39 MPa) for the tests on No.18 bars.
To determine the anchorage length needed to achieve yield
and fracture, the model was implemented with the nominal
properties of a No. 18 bar and 8000 psi (55 MPa) grout.
These analyses showed that the bar can achieve nominal
yield with an embedment length of six bar diameters and
fracture (100 ksi [690 MPa]) in 10 bar diameters. These
embedment lengths include the effect of the unconfined cone
near the grout surface. In some connections, the bars are
deliberately debonded over a short length near the interface
to reduce the strain concentration there. If the debonding is
at least as long as the thickness of the grout annulus, no cone
forms, and yield can be achieved in four bar diameters and
fracture in eight (Fig. 13).
COMPARISON WITH PREVIOUS RESEARCH
Comparisons of bond models are complicated by variations in
test configurations and approaches to measuring bar
properties and normalizing results. Figure 10 compares the

calibrated bond models from this research to the bond


models proposed by Eligehausen et al. (1983) and Raynor et
al. (2002) for confined embedment, and by Viwathantepa et
al. (1979) for unconfined concrete embedment. The models
are plotted as normalized bond stress against normalized
displacement. Although these variables were not explicitly
calculated by each researcher, the results have been
converted to that common basis for comparison. The bond
model developed herein predicts higher strengths than does
Eligehausen et al.s (1983) model for bars embedded in
concrete but lower strengths than Raynors (2002) model for
No. 8 bars embedded in grouted ducts.
Comparisons with Raynor et al.s (2002) tests of No. 6,
No. 8, and No. 10 bars make it possible to broaden the range
of bar sizes considered. Raynors (2002) test setup, however,
did not include a free grout surface, so the formation of the
cone was suppressed. To compare the tests consistently, Fig. 13
plots the bar stress versus the confined embedment length.
This approach implicitly assumes that the cone (unconfined)
region did not contribute significantly to the pullout resistance.
From Fig. 13, it appears that Raynors (2002) tests on No. 6,
No. 8, and No. 10 bars and the current tests on No. 10 bars
have higher pullout resistance than the No. 18 bars. The
differences, however, are approximately the same magnitude as the scatter among the tests.

Fig. 10Bond model comparison.

Fig. 12Peak stress comparison: simulated versus measured.

Fig. 11Model and test comparison: Tests 18N04 and 18N14.


(Note: Displacement measurements became unreliable for
Test 18N14 before peak load was reached.)
ACI Structural Journal/July-August 2009

Fig. 13Effect of confined embedment length on peak bar stress.


511

IMPLEMENTATION
A design equation was developed by assuming fg =
8000 psi (55 MPa) and fitting a lower bound curve to the test
data. It shares with the ACI development length equation the
same dependencies on steel strength, bar diameter, and
concrete or grout strength. In psi units, it is
fy
d duct d b
l d = ---------------------- d b + --------------------

2
130 f g

(4)

The second term represents the length of the cone. If no cone


will form, for example because of the presence of a sleeve to
debond part of the bar, this term should be taken as zero.
Equation (4) is intended to develop the nominal yield
stress (ACI Committee 408 2003). From it, the bar stress can
be computed and compared with the test results (Fig. 4).
Equation (4) is conservative compared with the test results,
but nonetheless results in much shorter embedment lengths
for a given bar stress than those provided by current codes.
For example, Eq. (4) results in a development length of 6.5db
for a 60 ksi (415 MPa) bar in 8000 psi (55 MPa) grout in a
duct with diameter 3.6db. The comparable ACI (2005) equation
(12-1) and AASHTO (2005) equation (5.11.2.1.1) require
development lengths of 20db and 33db, respectively.
The validity of Eq. (4) depends on the suppression of
several failure mechanisms not considered in this study. For
example, if the body of the concrete were to split, the duct
could pull out of it (Brenes et al. 2006). That failure mode
was not observed in these tests and is unlikely for corrugated
ducts in heavily confined joints such as those used in bridge
bent construction.
Seismic development length
Seismic modification factors to Eq. (4) are needed to
account for the reduction in bond strength due to cycling.
ACI (2005) and AASHTO (2005) require increases in development length of 30% and 25%, respectively. Raynor (2000)
proposed development length equations for cyclic loads that
are 40% longer than monotonic loads. For practice, a
seismic modification factor of 1.5 would be conservative
and would result in a seismic development length of 10 bar
diameters. Further testing would be required to determine an
accurate relationship between cyclic and monotonic
anchorage lengths for these grouted bars.
SUMMARY AND CONCLUSIONS
Fourteen pullout tests of reinforcing bars grouted into
ducts were conducted to determine the development length
of the bars, the effect of bar size, and the effect of polypropylene
fiber reinforcement. A bond model was formulated to extend
the test results. The experiments and analyses led to the
following conclusions:
1. The development length of a reinforcing bar grouted
into a corrugated steel pipe is much shorter than suggested
by current code equations. The analytical model and the test
results showed that yield and fracture can be achieved with
embedment lengths of six and 10 bar diameters, respectively,
for monotonic loads;
2. A simple equation based on this research results in
development lengths of approximately 6.5db for static loads and
10db for cyclic loads for typical bar and grout properties tested;
3. In all the tests reported herein, failure occurred either by
formation of a cone near the grout surface combined with
512

shearing along the confined bar-grout interface or by bar


fracture. Within the grout, the region near the pulled end of
the bar (defined by a 45-degree cone) is relatively unconfined,
resulting in resistance that is lower than in the remaining
embedded length;
4. The specimens did not fail by duct pullout or concrete
splitting. The suppression of these failure modes was attributed
to the size of the duct, the deep corrugations on the duct, and
the restraint provided by the surrounding concrete;
5. The effect of the bar diameter on the peak bond stress is
uncertain. No clear trend is visible from these tests.
Although there appears to be a difference in response
between the large bars (No. 18) tested in this study and small
bars (No. 8) tested by Raynor et al. (2002), this difference
may be due to differences in test setup of the two studies; and
6. Polypropylene fiber reinforcement typically lowered
the peak pullout strength of the bar and had mixed impact on
the post-peak response. The fibers also consistently lowered
the grout strength.
ACKNOWLEDGMENTS
The experiments were conducted with the valuable assistance of J. B.K.
Pang and V. Chaijaroen. Funding for this research was provided by the
Washington State Department of Transportation and the Valle Scholarship
and Scandinavian Exchange Program at the University of Washington.

NOTATION
db
Eenv
Epeak
fg
fu
fy
lc
le
le/db
P

avg

=
=
=
=
=
=
=
=
=
=
=
=

bar diameter, in. (mm)


envelope error
peak stress error
grout compressive stress, psi (MPa)
bar ultimate stress, ksi (MPa)
bar yield stress, ksi (MPa)
cone length, in. (mm)
embedment length, in. (mm)
normalized embedment length
axial force on bar, kip (kN)
axial bar stress calculated using nominal bar area, ksi (MPa)
average bond stress calculated using nominal bar perimeter,
psi (MPa)
avg /fg = normalized bond stress

REFERENCES
AASHTO and FHWA, 2004, Prefabricated Bridges 2004, Federal
Highway Administration and American Association of State Highway and
Transportation Officials, Washington, DC, 16 pp.
AASHTO, 2005, LRFD Bridge Design Specifications, fourth edition,
American Association of State Highway and Transportation Officials,
Washington, DC, 464 pp.
ACI Committee 318, 2005, Building Code Requirements for Structural
Concrete (ACI 318-05) and Commentary (318R-05), American Concrete
Institute, Farmington Hills, MI, 430 pp.
ACI Committee 408, 2003, Bond and Development of Straight Reinforcing
Bars in Tension (ACI 408R-03), American Concrete Institute, Farmington
Hills, MI, 49 pp.
Billington, S. L.; Barnes, R. W.; and Breen, J. E., 1999, Precast
Segmental Substructure System for Standard Bridges, PCI Journal, V. 44,
No. 4, pp. 56-73.
Brenes, F. J.; Wood, S. L.; and Kreger, M. E., 2006, Anchorage
Requirements for Grouted Vertical-Duct Connectors in Precast Bent Cap
Systems: A Summary, Project Summary Report 0-4176, Center for
Transportation Research, Austin, TX, pp. 1-3.
Cheok, G. S. and Stone, W. C., 1994, Performance of 1/3 Scale Model
Precast Beam-Column Connections Subjected to Cyclic Inelastic Loads:
Report No. 4, Report No. NISTIR 5436, National Institute of Standards
and Technology, Gaithersburg, MD, June, 59 pp.
Darwin, D., and Salamizavaregh, S., 1993, Bond Strength of Grouted Reinforcing Bars, Report No. SM 32, Structural Engineering and Engineering
Materials, University of Kansas Center for Research, Inc., Lawrence, KS, 139 pp.
Eligehausen, R.; Popov, E.; and Bertero, V., 1983, Local Bond Stress-Slip
Relationships of Deformed Bars Under Generalized Excitations, Report
UCB/EERC-83/23, Earthquake Engineering Research Center, University of
California-Berkeley, Berkeley, CA, 169 pp.

ACI Structural Journal/July-August 2009

FHWA, 2007, Accelerated Bridge Construction List, United States


Department of Transportation Federal Highway Administration, http://
www.fhwa.dot.gov/bridge/accelerated/abclist.cfm.
Hieber, D. G.; Wacker, J. M.; Eberhard, M. O.; and Stanton, J. F., 2005a,
State-of-the-Art Report on Precast Concrete Systems for Rapid Construction
of Bridges, Report WA-RD 594.1, Washington State Department of
Transportation, Olympia, WA, 112 pp.
Hieber, D. G.; Wacker, J. M.; Eberhard, M. O.; and Stanton, J. F., 2005b,
Precast Concrete Pier Systems for Rapid Construction of Bridges in
Seismic Regions, Report WA-RD 611.1, Washington State Department of
Transportation, Olympia, WA, 312 pp.
Jirsa, J. O.; Lutz, L.; and Gergely, P., 1979, Rationale for Suggested
Development, Splice, and Standard Hook Provisions for Deformed Bars in
Tension, Concrete International, V. 1, No. 7, July 1979, pp. 47-61.
Moosavi, M., and Bawden, W. F., 2003, Shear Strength of Portland Cement
Grout, Cement and Concrete Composites, V. 25, No. 7, pp. 729-735.
Nakaki, S. D.; Stanton, J. F.; and Sritharan, S., 1999, Overview of the
PRESSS Five-Story Precast Test Building, PCI Journal, V. 44, No. 2,
Mar.-Apr., pp. 26-39.
Orangun, C. O.; Jirsa, J. O.; and Breen, J. E., 1977, A Reevaluation of
Test Data on Development Length and Splices, ACI JOURNAL, Proceedings
V. 74, No. 3, Mar., pp. 114-122.
PCI, 2004, PCI Design Handbook, sixth edition, Precast/Prestressed
Concrete Institute, Chicago, IL, 736 pp.
PCI, 2007, Architectural Precast Concrete, third edition, Precast/
Prestressed Concrete Institute, Chicago, IL, 588 pp.

ACI Structural Journal/July-August 2009

Raynor, D. J., 2000, Bond Assessment of Hybrid Frame Continuity


Reinforcement, masters thesis, University of Washington, Seattle,
WA, 248 pp.
Raynor, D. J.; Lehman, D. E.; and Stanton, J. F., 2002, Bond-Slip
Response of Reinforcing Bars Grouted in Ducts, ACI Structural Journal,
V. 99, No. 5, Sept.-Oct., pp. 568-576.
Sritharan, S., 2005, Strut-and-Tie Analysis of Bridge Tee Joints
Subjected to Seismic Actions, Journal of Structural Engineering, V. 131
No. 9, pp. 1321-1333.
Stanton, J. F.; Eberhard, M. O.; Gunnarsson, K.; Hieber, D. G.; and
Wacker, J. M., 2006, Rapid Construction Details for Bridges in Seismic
Zones, Proceedings of the 8th U.S. National Conference on Earthquake
Engineering, San Francisco, CA, Apr. (CD-ROM)
Steuck, K. P.; Pang, J. B. K.; Eberhard, M. O.; and Stanton, J. F., 2008,
Anchorage of Large-Diameter Reinforcing Bars Grouted into Ducts,
Report WA-RD 684.1, Washington State Department of Transportation,
Olympia, WA, 148 pp.
Viwathanatepa, S.; Popov, E. P.; and Bertero, V. V., 1979, Effects of
Generalized Loadings on Bond of Reinforcing Bars Embedded in Confined
Concrete Blocks, Report UCB/EERC-79/22, Earthquake Engineering
Research Center, University of California-Berkeley, Berkeley, CA, 304 pp.
Wacker, J. M.; Hieber, D. G.; Eberhard, M. O.; and Stanton, J. F., 2005,
Design of Precast Concrete Piers for Rapid Construction in Seismic
Regions, Report WA-RD 629, Washington State Department of Transportation,
Olympia, WA, 160 pp.

513

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Você também pode gostar