Академический Документы
Профессиональный Документы
Культура Документы
net/publication/276432966
Unified Model of Local Bond between Deformed Steel Rebar and Concrete:
Indentation Analogy Theory and Validation
CITATIONS READS
3 134
2 authors:
Some of the authors of this publication are also working on these related projects:
Design and Evaluation of Bridges with Improved Mechanical Performance and Sustianability View project
All content following this page was uploaded by Chenglin Wu on 10 July 2015.
Abstract: Two distinct groups of bond models were developed on the basis of dominant concrete splitting and rebar–concrete interfacial
failures, respectively. Their accuracy highly depended on the specific test data sets selected in analysis and validation. In this study, a unified
model of local bond is proposed and developed on the basis of the indentation analogy at the rib front of the deformed rebar and on the
concrete plastic deformation around the rebar. Unlike existing models, the unified model is completely validated with all sets of test data
available in the literature. It can accurately predict the transition mechanism of failure modes from concrete splitting to rebar pullout.
The predicted bond strength is in excellent agreement with experimental results as the rib spacing-to-height ratio changes in practical range.
DOI: 10.1061/(ASCE)EM.1943-7889.0000945. © 2015 American Society of Civil Engineers.
Author keywords: Rebar-concrete bond; Indentation analogy; Unified model; Interfacial mechanics; Deformed rebar; Concrete splitting.
J. Eng. Mech.
strength of each bar, which would significantly exceed the test f'v
results. f'c
Hydrostatic
Rebar–Concrete Interfacial Behavior Core
Cairns (1979) questioned what factors other than concrete splitting
significantly contributed to the bond strength. To this end, Cairns
and Jones (1995) and Cairns and Abdullah (1996) conducted a f'c
series of detailed stress analyses taking into account the rib defor-
sr sr0
mation and surface condition of rebar for each experimentally ob- f'v
served failure mode. Using numerical analysis, they concluded that
the concrete bearing angle against uncoated steel rebar was mainly
related to the cohesive force of concrete. However, other test results sflat
Downloaded from ascelibrary.org by University of Texas At Austin on 07/10/15. Copyright ASCE. For personal use only; all rights reserved.
by Choi and Lee (2002) indicated that both the deformation pattern
and surface condition of uncoated rebar affected the effective bear-
ing angle, and thus the rebar–concrete bond strength for a given
failure mode of concrete crushing at the rib-front area. Here, the
failure plane was assumed to occur in concrete so that the coeffi- hr
cient of friction between concrete and concrete was considered.
Even so, Choi and Lee (2002) still considered a constant bearing
angle of 30° in their bond modeling. In his local bond model, Wang Fig. 1. Indentation analogy
(2009) introduced an imaginary bar by repeatedly connecting the
tip of one rib to the toe of the following rib of rebar, corresponding
to the minimum rib-face angle possible in the original rebar, and
accounted for partial effects of the deformation pattern and surface rib-front face as circled in Fig. 1. This observation indicates the
condition of the rebar. How the fictitious rib-face angle or concrete- existence of a hydrostatic pressure zone near the rib-front area.
bearing angle affects the bond strength has yet to be investigated, Therefore, the normal component of the rebar–concrete interfacial
particularly when the rib spacing-to-height ratio of rebar increases. force corresponds to the concrete compressive strength f c0 , as
Cairns and Jones (1995) indicated that the hydraulic pressure shown in Fig. 1. The concrete crushing zone forms a new
analogy and rebar–concrete interface studies generally overesti- wedge of the indenter and acts as the core zone mentioned by
mated and underestimated the rebar–concrete bond strength, re- Johnson (1985).
spectively. In fact, CEB-FIP Task Group (2000) stated that the A classical indentation process differs from the debonding pro-
core of the bond is a balance of the confinement strength, pro- cess in that:
vided by concrete cover or transverse reinforcement, and the shear • There is no rebar–concrete interface in the flat portion and no
strength in the vicinity of rebar ribs. Most of the existing models stress-free back face of the ribs, as indicated in Fig. 2; and
included a parameter of relative rib area that is not well supported • It occurs in a semi-infinite elastic body instead of the finite body
by the test results (Rehm 1957, 1961; Darwin and Graham 1993). with balanced concrete confinement and near-rib stress condi-
In addition, Darwin et al. (1992) showed a slight variation of the tion in the debonding process.
effective bearing angle along the rebar length. The experimental Despite the this difference, the stress field in the indentation
contradiction to the fixed bearing angle theory was previously problem closely resembles the near-rib stress distribution in the
noted but never explained in the literature. debonding problem, as schematically indicated by the indenter-
In this study, the two traditionally distinct research directions are induced displacement field in Fig. 1. To minimize their differences,
integrated into a unified rebar–concrete bond model, including both the stress in interested areas such as the back face of ribs can be
concrete confinement and rebar–concrete interface behavior. The modified to meet the stress-free condition, as indicated by the free
process of a rebar rib pressing against concrete is simulated by surface Γ in Fig. 2. In this case, the stress at the Γ surface is first
the indentation of a steel cone into concrete so that the stress field calculated from the indentation solution and then cancelled by
near the rebar rib, and thus the effective bearing angle of the con- introducing a counter-stress vector equal in magnitude and opposite
crete, can be evaluated. The cone indentation analogy is modified in direction. The key line between ribs is displayed in Fig. 2.
with clearly stated assumptions to consider the effects of multiple Fig. 3 shows a conical-shaped indenter pressed into a semi-
ribs, rib height-to-spacing ratios, and failure surface characteristics infinite space in the Cartesian (x, y, z) and cylindrical (r, θ, z) co-
of steel/concrete. The analytical solution for bond equations is ordinate systems. As shown in Fig. 3, the radius of the indenter is
validated with test data obtained from uncoated rebar in concrete represented by a (a ≈ hr ). The normal and shear tractions along
by various researchers around the world. the side surface of the indenter are denoted as Pm and T, where
pm ¼ F=πh2r and T ¼ c1 pm , respectively. Here, c1 represents
the ratio between the shear and normal traction.
An Indentation Analogy for Deformed
Rebar-in-Concrete Analysis
Stress Induced by Normal Traction
Fig. 1 shows half a cross section of deformed rebar with sr and hr
representing the spacing and height of periodical ribs, respectively. A semi-infinite space subjected to a conical shape indenter under
When the rebar embedded along the center line of a concrete cyl- uniformly distributed normal traction pm along the side surface of
inder is pulled out of its surrounding concrete (downward in Fig. 1), the indenter is an axis-symmetrical problem studied by Sneddon
the steel ribs act as a series of conical shape indenters pressed (1948). The induced elastic stress field under the normal traction
against and separated from concrete in their front and back faces, is thus independent of the coordinate θ. The three nonzero stress
respectively. As a result, concrete is often crushed along the components (σz , σr , τ rz ) normalized by pm can be expressed as
J. Eng. Mech.
Interface
Free Surface:
Key-line: k
Traction Induced by Reversely Applied
Downloaded from ascelibrary.org by University of Texas At Austin on 07/10/15. Copyright ASCE. For personal use only; all rights reserved.
ðnÞ
σz z 0 2ð1 − ν 2 Þpm
0
¼ − J1 þ J2 ð1Þ cot α ¼ ð4Þ
pm hr E
ðsÞ pffiffiffi
τ rz 2
¼ c fz½f10 ðzÞ − f20 ðzÞ − ½f1 ðzÞ þ f 2 ðzÞg ð7Þ
O Pm x pm 4 1
J. Eng. Mech.
Z
ðtÞ 12ðz − s2r0 Þ sr0 components at the back-rib surface due to normal and shear trac-
σr jr¼hr ¼ σr jr¼hr þ
s3r0 cos α sr0 −hr cot α tions. Fig. 4(c) shows the combined effect of normal and shear
tractions as well as three counter stress components at the back-
ðtÞ s ðtÞ
× σr jΓ z − r0 − σz jΓ × ðsr0 − zÞ tan α rib surface. Under the normal and shear tractions as indicated in
2 Fig. 4(a, b), the tension radial stress at the rebar–concrete interface
ðtÞ s starts from zero at approximately the rib-height depth, reaches its
þ τ zr jΓ r0 sin α dz ð8Þ maximum at approximately twice the rib height, and decays slowly.
2
The compression radial stress at the key line is the maximum at
where dz represents an infinitesimal length on the rib-back face the top surface and decays rapidly to near zero at approximately
along the vertical direction, sr0 ¼ sr − sflat , and sflat =hr ¼ 1 is the three times the rib height. Overall, the radial stresses at the rebar–
length of the tip of ribs. concrete interface and at the key line are significant only near the
rib-front area in the debonding process. Therefore, the effect of the
stresses at the free surface is expected to be small, as indicated by
Downloaded from ascelibrary.org by University of Texas At Austin on 07/10/15. Copyright ASCE. For personal use only; all rights reserved.
Unified Bond-Strength Formulation comparing Figs. 4(a and c). Furthermore, the vertical stress com-
ponent σz at the back-rib surface is approximately 10 times as large
When β ¼ 60°, sr0 =hr ¼ 10, for representative deformed rebar, as the radial component σr and the shear component τ zr . In com-
numerical analysis was conducted to evaluate the effects of normal bination with other geometry parameters in Eq. (8), the effect of the
traction, shear traction, and counter stresses at the free surface. vertical stress component at the free surface Γ, σz , is dominant.
Fig. 4(a) presents the radial stress components at the rebar–concrete Therefore, the following analysis takes into account the effect of
interface and at the key line. Fig. 4(b) represents three stress the vertical stress component at the traction-free surface only.
0 0.005
1 r
/pm-keyline
0 hr /p
/pm-near rebar Comp. r m
r
2
r /p
z m
3 hr -0.005 Tension
Comp.
r
4
Tension -0.01
r:
“+”
m
r
z/h
/p
5 10hr
r:
“-”
“+” -0.015
6 r:
10hr
r:
“-”
7 -0.02
z z:
“-”
8
-0.025
9 z z:
“-”
10 -0.03
-0.25 -0.2 -0.15 -0.1 -0.05 0 0.05 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
/p Distance from Bar along Surface in Ratio
(a) r m (b)
0
r
/pm-keyline
2
r
/pm-near rebar
3
hr
Comp.
4 r
Tension
z/hr
r:
“+”
6
10hr
r:
“-”
7
8
z z:
“-”
9
10
-0.25 -0.2 -0.15 -0.1 -0.05 0 0.05
(c) /p
r m
Fig. 4. Stress distributions along various boundaries: (a) normal pressure and shear traction effect; (b) stress distribution along free surface; (c) effect
of normal pressure, shear traction, and the counter force at the back-rib surface
J. Eng. Mech.
Downloaded from ascelibrary.org by University of Texas At Austin on 07/10/15. Copyright ASCE. For personal use only; all rights reserved.
Fig. 5 presents the stress distributions at various boundaries failure can initiate and propagate through the concrete key line.
when sr0 =hr varies (other parameters remain the same as used At the ultimate state immediately prior to the plow through failure,
in Fig. 4). The stress state near the rib areas can be divided into the bearing stress on the rib front face reaches the concrete com-
three groups: sr0 =hr ≤ 7, 7 < sr0 =hr ≤ 10, and sr0 =hr > 10, as pressive strength fc0 . Let c2 ¼ fv0 =f c0 be the ratio between the shear
discussed below: stress on the rib front face and the concrete compressive strength.
• For sr0 =hr ≤ 7, the radial stress generated along the key line is Here, the uniaxial concrete strength is used, As the triaxial confine-
in tension near the top surface. The tensile stress along the ment is impossible to form because of early radial crack growth.
rebar–concrete interface is approximately zero in the flat por- For a low rib spacing-to-height ratio, the radial pressure in Eq. (8)
tion of the rebar; is insignificant and can be represented by its uniform distribution
• For 7 < sr0 =hr ≤ 10, the radial stress generated along the key over the considered length pn , as shown in Fig. 6. In applications,
line is in compression. The integration of the radial stress along pn can be approximated as uniformly distributed radial stress at the
the vertical axis over a distance of hr constitutes approximately key line due to the effect of concrete cover. Additionally, there is no
77% of the total integration over the entire vertical axis. The contact force on the flat portion of the rebar. At the imminent plow
tensile stress along the rebar–concrete interface indicates no through failure, the horizontal forces per linear thickness of the
contact between the rebar and concrete in the flat portion of the concrete key are in equilibrium, resulting in the critical rib-face
rebar. In this case, the influence of the vertical stress at the free angle β cr
surface Γ becomes insignificant;
• For sr0 =hr > 10, the radial stress generated along the key line is
in compression and concentrated in the near rib area. Part of the
flat rebar region is in compression, indicating that the rebar and Tension
concrete remains in contact in that area. Therefore, the flat por- Opening
tion should be considered in the evaluation of bond strength; f'c
• On the basis of these findings, approximations were made to
formulate the bond-strength equations for the three cases cover-
ing the entire application range in practice. These bond-strength pn
equations are detailed below. β
f'v
Low Rib Spacing-to-Height Ratio: s r 0 =h r ≤ 7
J. Eng. Mech.
p s Medium Rib Spacing-to-Height Ratio: 7 < s r =h r 0 ≤ 10
β cr ¼ arccot c2 þ n0 r0 ≤ β ð9Þ
f c hr
Failure Mechanism and Corresponding Experimental
The critical rib-face angle must be equal to or smaller than β, Findings
because the steel rib remains elastic. When c2 ¼ 0.83 (Rehm 1957, As the rib spacing-to-height ratio falls into a medium range of 7
1961), sr0 =hr ¼ 6, sflat ≅ hr and pn =fc0 ¼ 0.06, β cr ¼ 40°, which to 10, the shear mechanics dominates the bond behavior of rebar
is less than the rib angle β ¼ 60° widely used in representative in normal strength concrete. In this case, the force equilibriums in
deformed rebar. two orthogonal directions are used to determine the effective bear-
To ensure that the considered balancing pressure pn is realistic ing angle. Depending upon the shear strength and confinement
in application, a thick-walled hollow cylinder with inner and outer level, the effective bearing angle varies and leads to different failure
diameters of db and db þ 2c (db = rebar diameter and c = clear patterns.
concrete cover) is analyzed under internal pressure pn . When db ¼ Previous research has shown that the concrete at the rib-front
Downloaded from ascelibrary.org by University of Texas At Austin on 07/10/15. Copyright ASCE. For personal use only; all rights reserved.
25.2 mm and c ¼ 50.4 mm (Rehm 1957, 1961), the internal pres- area becomes compact powder because of high stress concentration
sure corresponding to a hoop stress equal to the tensile strength (Lutz and Gergely 1967; Esfahani and Rangan 1998). It was
of concrete is pn ¼ 1.1ft (ft is the uniaxial tensile strength of found that the effective bearing zone plays a critical role in trans-
concrete). When f t ≈ 0.1f c0 , pn =fc0 ≈ 0.11. When the stress con- ferring the bearing component of bond forces to their surrounding
centration near the rib is considered, the pn =fc0 ¼ 0.06 value seems concrete.
reasonable.
On the basis of various tests (Idun and Darwin 1999), the aver- Existence of Crushing Zone
age c2 value of uncoated rebar in normal-strength concrete is ap- The cross section of half a reinforcing bar with a crushing zone
proximately 0.6. According to Eq. (9), its corresponding critical at the rib-front area is presented in Fig. 7. Possible failures for
rib-face angle β cr is 59° when pn is neglected. For their specimens, an intermediate rib spacing-to-height ratio include three cases:
db ¼ 25.2 mm and c ¼ 50.4 mm were also used. (1) rib sliding for weak rebar–concrete interface, (2) concrete
To facilitate the understanding of the plow through failure mode, crushing for relatively weak rebar–concrete interface, and (3) con-
the initial tearing-off condition at the key line must be evaluated. crete shear-off for strong rebar–weak interface. As illustrated in
At the beginning of the plow through failure, the concrete key re- Fig. 7, Case (1) can be considered as a special case of (2) when
mains intact and can be used as a cantilever beam to estimate the the crushing angle α is equal to the rib-face angle β. Case (3) is
near-rib stress conditions under the interface forces, as shown in similar to Case (2) except that the rebar–concrete interface strength
Fig. 6. Specifically, the average shear stress σrz over the key line and concrete strength are in different magnitude order. Therefore,
and the flexural tension stress σrr at the extreme fiber of the key line the general condition for Case (2) is discussed in detail.
section can be evaluated by As shown in Fig. 7, when fc1 ¼ fc2 ¼ fc0 and fv1 ¼ fv2 ¼
c2 f c0 , the shaded crushed zone provides a hydrostatic pressure
condition. Therefore, the stress components on the crushed plane
hr
σrz ¼ fc0 ð1 þ c2 cot βÞ ð10aÞ with α angle are fn ¼ f c0 and fv ¼ c2 f c0 . The radial and vertical
sr0 force equilibriums of the uncrushed part of the rib-front area in
Fig. 7 (unit thickness out of plane) yield
2
hr f c0 t1 cos α − c2 f c0 t1 sin α − pn t2 ¼ 0 ð13aÞ
σrr ¼ 3fc0 ð1 þ c2 cot βÞ ð10bÞ
sr0
The shear-to-flexural stress ratio can then be expressed into a −fc0 t1 sin α − c2 fc0 t1 cos α − f b t2 ¼ 0 ð13bÞ
function of sr0 =hr by
where t1 and t2 represent the lengths of the crushed plane and its
σrz s vertical projection along the key line, t2 ¼ t1 cos α, f b is the
¼ r0 ð11Þ
σrr 3hr strength distributed over the projected length, and f̄b represents
the average bond strength over the rib spacing sr , and f̄ b sr0 ¼ f b t2 .
Whether the crushing zone exists or not mainly depends on the
The shear strength is equal to 0.5ft for a uniaxial stress state and
relation between the calculated α and the β value. When α < β, the
f t for a pure-shear stress state. For a less conservative estimate, the
crushing zone is present. For the three cases in Fig. 7, failure can
shear strength equal to f t can be considered. Therefore, the shear
occur inside the concrete c2 ¼ cc related to the cohesion and in-
failure governs when sr0 =hr ≥ 3 in applications and the plow
ternal friction angle and at the rebar–concrete interface c2 ¼ ci re-
through failure mode is accompanied by the shear failure along
lated to the coefficient of friction. For example, ci ¼ 0.53 and
the key line. If the lower shear strength 0.5ft were used, the shear
cc ¼ 0.25 þ tan 30° ¼ 0.83 (Cairns and Abdullah 1996).
failure along the key line would always govern.
In this way, the interfacial strength is quantified by the param-
eter c2, and it is bounded between ci and cc . The confining effect
Bond Strength
is then quantified by the parameter c0 ∶ ¼ ffnc0 . Therefore, the effective
Before the imminent plow through failure, the total vertical force
bearing angle can be evaluated with these two parameters for differ-
per linear thickness of the concrete key by neglecting the pn effect
ent failure cases.
can be determined and used to evaluate the average bond strength
Following is a brief summary of the effective bearing angle α
f̄ b over the rib spacing sr
and the average bond strength f̄b in three cases:
1. Rib sliding [α ≥ β; c2 ¼ ci ≤ ð1 − c0 Þ cot β]
hr
f̄ b ¼ fc0 ð1 þ c2 cot βÞ ð12Þ The pull-out specimen fails at the rebar–concrete interface.
sr0 Letting α ¼ β, Eq. (13a) results in
J. Eng. Mech.
fn
fv fb
fv fb
Failure Plane
β fn
Concrete Crushing hr
Case 1: Rib Sliding Case 3: Concrete Shearing-off
sr
fc1
fv1 fn fv
fv fb
β
Downloaded from ascelibrary.org by University of Texas At Austin on 07/10/15. Copyright ASCE. For personal use only; all rights reserved.
sflat
fc2 α
fv2
fn
hr
α
hr
Case 2: Concrete crushing
For a weak rebar–concrete interface, the bond strength is Bond Strength of the Flat Portion
governed by the interface contact strength prior to concrete Fig. 8(a) shows the stress components on various boundaries of a
splitting. The effective bearing angle is equal to the rib- concrete key. The flat portion near the rib front face (sr0−1 ¼ 10hr )
face angle. was dealt with in the same way as considered in the case of medium
2. Concrete crushing [α < β; ð1 − c0 Þ cot β < c2 ¼ ci < cc ] rib spacing-to-height ratio. The flat portion near the rib back face
By introducing c0 ¼ pn =fc0 , Eq. (13a) leads to (sr0−2 ¼ sr0 − sr0−1 ) was considered as smooth rebar. The average
bond strength can then be obtained from a weighted summation
1 − c0 f̄b h
α ¼ arctan ; ¼ r ð1 þ c2 cot αÞ ð15Þ 1
c2 fc0 sr0 f̄b ¼ ½f̄ s þ sr0−2 f̄b−flat ð16Þ
sr b−rib r0−1
For a relatively weak interface, the concrete near the rib where f̄ b−rib and f̄b−flat represent the average bond strengths due to
face was crushed prior to concrete-splitting failure. The bond the rib and flat effects, respectively. The rib effect f̄ b−rib can be
strength is governed by the confinement effect. The higher
interface bond corresponds to a lower effective bearing
angle. f'c
3. Concrete shear-off [α < β; ð1 − c0 Þ cot β < c2 ¼ cc < ci ]
Eq. (15) is applicable for this case. For a strong rebar–
concrete interface, the bond strength is governed first by fv
the concrete shear along part of the key line, like a block-shear pn
failure, and then concrete splitting. As the concrete confine- f'a
c
ment increases, the effective bearing angle tends to decrease sr0-1= 10hr
Rib Portion
to zero and the concrete key starts to shear-off. In the extreme
case when α ¼ 0, c0 ¼ 1, or pn approaches f c0 , a transforma- sr0= sr- sflat
tion occurs from an overall concrete splitting to a pure pull-
out failure.
pn pn
High Rib Spacing-to-Height Ratio: 10 < s r 0 =h r sr0-2= sr0 -sr0-1
J. Eng. Mech.
the cracked concrete, the tangential strain is inversely proportional
to the radial distance r (rs ≤ r ≤ re ). That is, εt ¼ εt0 re =r. For an
infinitesimal element, the force equilibrium in the radial
direction gives the following equation:
Z
re r2c − r2e re dr
pn ¼ f tu þ σt ð20Þ
rs r2c þ r2e rs rs
(a) (b) Z
r4c − 5r4e 1 re ∂σ
t
þ dr þ σ ðr Þ ¼0 ð21Þ
ðr2c þ r2e Þ2 ftu rs ∂re
t e
Fig. 9. Softening behavior of concrete cover: (a) axis-symmetric
problem; (b) stress–strain relation in tangential direction
The critical radial crack length can be numerically solved from
Eq. (21) and substituted back into Eq. (20) to evaluate pn .
J. Eng. Mech.
Table 1. Test-over-Prediction Ratios for Bond Strength of Uncoated Rebar
Number Orangun Zuo and Wang and Proposed
Database of tests Data type et al. (1977) Darwin (2000) Wang (2009) Liu (2003) model
Choi et al. (1991) 29 Mean 1.491 1.247 1.050 0.709 1.004
— COV 0.184 0.187 0.122 0.086 0.069
Darwin and Graham (1993) 33 Mean 1.479 1.432 1.024a 0.614 1.060
— COV 0.106 0.103 0.086 0.214 0.064
Idun and Darwin (1999) 14 Mean 1.432 1.090 1.150 0.715 0.963
— COV 0.221 0.173 0.146 0.093 0.096
Miller et al. (2003) 35 Mean 1.603 1.095 0.983 0.820 0.974
— COV 0.208 0.103 0.080 0.054 0.103
De Anda et al. (2004) 10 Mean 1.497 1.135 0.965 0.794 1.055
— COV 0.163 0.237 0.091 0.028 0.061
Downloaded from ascelibrary.org by University of Texas At Austin on 07/10/15. Copyright ASCE. For personal use only; all rights reserved.
2
Orangun et al. (1977) Zuo and Darwin (2000)
1.8
Wang (2009) Wang and Liu (2003)
1.6
Proposed Model
1.4
Tested Over Predicted Ratio
1.2
0.8
0.6
0.4
0.2
0
Choi et al. Dawrin and Idun and Miller et al. De Anda et al. Wu et al.
(1991) Graham (1993) Darwin (1999) (2003) (2004) (2012)
and was thus the most accurate prediction in bond strength of by Zuo and Darwin (2000) corresponded to the data sets developed
uncoated rebar in concrete. The theoretical approach taken by by the same group (Idun and Darwin 1999; Miller et al. 2003).
Wang (2009) yielded the second most accurate prediction. The em- On the other hand, the least accurate prediction by Zuo and Darwin
pirical approach by Orangun et al. (1977) led to the least accurate (2000) was also for the data set generated by the same group
results. Overall, the theoretical approaches included in this study (Darwin and Graham 1993). Therefore, the prediction accuracy by
are more accurate than the empirical approaches, because the em- empirical approaches most likely depended on the test conditions
pirical models lack a solid mechanical basis and are thus less ver- and data sets used in the regression analysis.
satile in predicting bond behaviors under the different conditions Among the three theoretical approaches, the proposed model in
investigated by various researchers. The theoretical approach by this study is most flexible and applicable to various conditions such
Wang and Liu (2003) resulted in a nonconservative prediction as for coated rebar (Table 2 and Fig. 11). Wang (2009) did not take
as they neglected the effect of rebar surface characteristics such full consideration of plastic behaviors during concrete splitting, re-
as rib geometries and interface bond strength. sulting in a prediction that is insensitive to the variation in confine-
When individual data sets were used, the bond-strength predic- ment extent (Wu et al. 2012). Wang and Liu (2003) neglected the
tion accuracies by various groups were also compared. With the effects of rebar surface characteristics such as rib geometries and
data sets developed by others, Orangun et al. (1977) among all interface bond strength. Furthermore, most theoretical approaches
the models predicted the least accurate bond strength against failed to distinguish various failure modes associated with different
every data set. On one hand, the two most accurate predictions deformation patterns.
J. Eng. Mech.
Table 2. Test-over-Prediction Ratios for Bond Strength of Coated Rebar
Number Orangun et al. Zuo and Darwin Wang and Liu Proposed
Database of tests Data type (1977) (2000) Wang (2009) (2003) model
Choi et al. (1991)a 29 Mean 1.972 1.154 1.182 0.752 1.008
— COV 0.356 0.374 0.233 0.124 0.064
Idun and Darwin (1999)a 14 Mean 1.342 1.094 1.112 0.728 1.082
— COV 0.280 0.071 0.186 0.084 0.044
Miller et al. (2003)a 35 Mean 1.802 1.095 0.903 0.821 1.040
— COV 0.180 0.103 0.152 0.054 0.090
De Anda et al. (2004)a 61 Mean 1.597 0.803 0.905 0.782 1.022
— COV 0.157 0.175 0.191 0.089 0.095
Wu et al. (2012)b 12 Mean 1.379 1.143 1.053 0.733 0.997
— COV 0.462 0.324 0.377 0.156 0.067
Downloaded from ascelibrary.org by University of Texas At Austin on 07/10/15. Copyright ASCE. For personal use only; all rights reserved.
2
Orangun et al. (1977) Zuo and Darwin (2000)
1.8
Wang (2009) Wang and Liu (2003)
1.6
Proposed Model
1.4
Tested Over Predicted Ratio
1.2
0.8
0.6
0.4
0.2
0
Choi et al. (1991) Idun and Darwin Miller et al. (2003) De Anda et al. Wu et al. (2012)
(1999) (2004)
J. Eng. Mech.
1.6 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Choi et al. (1991) f1 ðzÞ ¼ ln l2 ðhr Þ þ l2 ðhr Þ2 − r2 þ ln z þ r2 þ z2 ð30Þ
1.5 Darwin and Graham (1993)
Idun and Darwin (1999)
1.4
Miller et al. (2003)
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Tested Over Predicted Ratio
1.3 De Anda et al. (2004) ½2h2r − l22 ðhr Þ h2r − l21 ðhr Þ z r2 þ z2 h2r
Wu et al. (2012) f2 ðzÞ ¼ þ − 2 ð31Þ
1.2 hr r2 r2 r
1.1 +1 × STD
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Mean 1
1 l1 ðhr Þ ¼ ðr þ hr Þ2 þ z2 − ðr − hr Þ2 þ z2 ð32Þ
-1 × STD 2
0.9
References
Appendix
Cairns, J. (1979). “Analysis of the ultimate strength of lapped joints of
The functions in Eqs. (1–3) are defined as follows: compression reinforcement.” Mag. Concr. Res., 31(106), 19–27.
2 2 −1=2 Cairns, J., and Abdullah, R. B. (1996). “Bond strength of black and epoxy-
r z cos ϕ coated reinforcement-a theoretical approach.” ACI Mater. J., 93(4),
J 02 ¼ þ − ð22Þ 362–369.
hr hr R
Cairns, J., and Jones, K. (1995). “Influence of rib geometry on strength of
lapped joints: an experimental and analytical study.” Mag. Concr. Res.,
2 2 1=2 47(172), 253–262.
h r z
J 11 ¼ r þ − R cos ϕ ð23Þ CEB-FIP Task Group Bond Models. (2000). “Bond of reinforcement
r hr hr in concrete.” Fédération International de la Précontrainte, Lausanne,
Switzerland.
2 1=2 Choi, O. C., Ghaffari, H. H., Darwin, D., and McCabe, S. L. (1991). “Bond
hr z cosðϑ − ϕÞ
J 12 ¼ 1þ of epoxy-coated reinforcement: bar parameters.” ACI Struct. J., 88(2),
r hr R 207–217.
2 2 −1=2 Choi, O. C., and Lee, W. S. (2002). “Interfacial bond analysis of deformed
z r z
− þ ð24Þ bars to concrete.” ACI Struct. J., 99(6), 750–756.
hr hr hr Darwin, D., and Graham, E. K. (1993). “Effect of deformation height and
spacing on bond strength of reinforcing bars.” ACI Struct. J., 90(6),
( 2 2) 646–657.
R þ 2Rð1 þ hz 2 Þ1=2 cosðϑ − ϕÞ þ 1 þ hz 2
2
1 Darwin, D., McCabe, S. L., Idun, E. K., and Schoenkase, S. (1992).
J 01 ¼ ln r r
ð25Þ “Development length criteria: Bars not confined by transverse
2 ½hz þ ð r 2 þ z 2 Þ1=2 2
2 2
J. Eng. Mech.
Hanson, M. T. (1992). “The elastic field for conical indentation including Rehm, G. (1957). “Fundamental law of bond.” Proc. Symp. Bond and
sliding friction for transverse isotropy.” J. Appl. Mech., 59(2), Crack Formation in Reinforced Concrete (Stockholm), Tekniska
S123–S130. Hogskolans Rotaprinttrychkeri, Stockholm, Sweden.
Harajli, M. H., Hout, M., and Jalkh, W. (1995). “Local bond stress-slip Rehm, G. (1961). “Uber die grunlagen des verbundes zwischen stahl und
behavior of reinforcing bars embedded in plain and fiber concrete.” beton [Basic principle of the bond between steel and concrete].”
ACI Mater. J., 92(4), 343–354. Deutscher Ausschuss fur Stahlberton, 1381(5), 59 (in German).
Hillerborg, A., Modeer, M., and Petersson, P. E. (1976). “Analysis of crack Reinhardt, H. W. (1992). “Bond of steel to strain-softening concrete taking
formation and crack growth in concrete by means of fracture mechanics account of loading rate.” Fracture Mechanics of Concrete Structures,
and finite elements.” Cem. Concr. Res., 6(6), 773–781. Z. Bazant, ed., Elsvier, London, 80–820.
Idun, E. K., and Darwin, D. (1999). “Bond of epoxy-coated reinforcement: Reynolds, G. C., and Beeby, A. W. (1982). “Bond strength of deformed
coefficient of friction and rib face angle.” ACI Struct. J., 96(4), bars.” Bond in Concrete, P. Bartos, ed., Applied Science, London,
609–616. 434–445.
Johnson, K. L. (1985). Contact mechanics, Cambridge University Press, Rosati, G., and Schumm, C. (1992). “Modeling of local bar-to-concrete
Cambridge. bond in reinforced concrete beams.” Int. Conf. Bond in Concrete—
Downloaded from ascelibrary.org by University of Texas At Austin on 07/10/15. Copyright ASCE. For personal use only; all rights reserved.
Kemp, E. L., and Wilhelm, W. J. (1979). “Investigation of the parameters From Research to Practice, A. Skudra and T. Tepfers, eds., Riga
influencing bond cracking.” ACI J., 77(3), 47–71. Technical University, Riga, Latvia, 12–43.
Losberg, A., and Olsson, P. A. (1979). “Bond failure of deformed reinforc- Sneddon, I. N. (1948). “Boussinesq’s problem for a rigid cone.” Math.
ing bars based on the longitudinal splitting effect of the bars.” ACI J., Proc. Cambridge Philos. Soc., 44(4492), 492–507.
76(1), 5–18. Soretz, S., and Holzenbein, H. (1979). “Influence of rib dimensions
Losberg, A., and Olsson, P.-A. (1979). “Bond failure of deformed reinforc- of reinforcing bars on bond and bendability.” ACI J. Proc., 76(1),
ing bars based on longitudinal splitting effect of the bars.” ACIJ., Proc., 111–128.
76(1), 5–18. Tepfers, R. (1973). “A theory of bond applied to overlapped tensile
Lutz, L. A., and Gergely, P. (1967). “Mechanics of bond and slip of de- reinforcement splices for deformed bars.” Publication No. 73:2, Div.
formed bars in concrete.” ACIJ. Proc., 64(11), 711–721. Concrete Structures, Charlmers Univ. of Technology, Goteborg,
Miller, G. G., Kepler, J. L., and Darwin, D. (2003). “Effect of epoxy coating Sweden, 328.
thickness on bond strength of reinforcing bars.” ACI Struct. J., 100(3), Van der Veen, C. (1990). “Cryogenic bond stress-slip relationship.”
1–7. Doctoral thesis, Delft Univ. of Technology, Delft, The Netherlands.
Nielsen, C. V., and Bicanic, N. (2002). “Radial fictitious cracking of thick- Wang, H. (2009). “An analytical study of bond strength associated with
walled cylinder due to bar pull-out.” Mag. Concr. Res., 54(3), 215–221. splitting of concrete cover.” Eng. Struct., 31(4), 968–975.
Noghaba, K. (1995). “Splitting of concrete in the anchoring zone of de- Wang, X., and Liu, X. (2003). “A strain-softening model for steel-concrete
formed bars-a fracture mechanics approach to bond.” Doctoral thesis, bond.” Cem. Concr. Res., 33(10), 1669–1673.
Lulea Univ. of Technology, Lulea, Sweden. Wu, C., Chen, G., Volz, J. S., Brown, R. K., and Koenigstein, M. L. (2012).
Olofsson, T., and Ohlsson, U. (1995). “A simple fracture mechanics model “Local bond strength of vitreous enamel coated rebar to concrete.”
for mixed-mode failure in concrete.” Int. Conf. Fracture Mechanics of Constr. Build. Mater., 35, 428–439.
Concrete Structures, Zurich, Switzerland, 25–28. Zuo, J., and Darwin, D. (2000). “Splice strength of conventional and high
Orangun, C. O., Jirsa, J. O., and Breen, J. E. (1977). “A reevaluation of test relative rib area bars in normal and high-strength concrete.” ACI Struct.
data on development length and splices.” ACI J., 74(3), 114–122. J., 97(4), 630–641.