See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/276432966

Unified Model of Local Bond between Deformed Steel Rebar and Concrete:
Indentation Analogy Theory and Validation

Article  in  Journal of Engineering Mechanics · May 2015

DOI: 10.1061/(ASCE)EM.1943-7889.0000945

CITATIONS READS

3 134

2 authors:

Chenglin Wu Genda Chen

Missouri University of Science and Technology Missouri University of Science and Technology
37 PUBLICATIONS   185 CITATIONS    264 PUBLICATIONS   2,183 CITATIONS   

SEE PROFILE SEE PROFILE

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

Bond, fracture, and failure in civil infrastructure View project

All content following this page was uploaded by Chenglin Wu on 10 July 2015.

The user has requested enhancement of the downloaded file.

 Unified Model of Local Bond between Deformed
Steel Rebar and Concrete: Indentation Analogy
Theory and Validation
Chenglin Wu 1 and Genda Chen, F.ASCE 2
Downloaded from ascelibrary.org by University of Texas At Austin on 07/10/15. Copyright ASCE. For personal use only; all rights reserved.

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.

Introduction according to a softening concrete constitutive relation without the

Since the 1950s, several rebar–concrete bond equations have been
developed both statistically and mechanistically, taking into ac-
count rebar surface characteristics, coating conditions, and concrete
properties. In general, the regression equation based on a specific
set of test data cannot accurately predict the bond strengths in other
test cases (Ferguson and Breen 1965; Ferguson and Briceno 1969;
Orangun et al. 1977; Losberg and Olsson 1979; Eligehausen et al.
1983; Kemp and Wilhelm 1979; Darwin et al. 1992, 1996; Harajli
et al. 1995; Zuo and Darwin 2000). Owing to the inherited draw-
back of any data-driven approach, the mechanics-based approach
was taken along two main directions: (1) hydraulic pressure anal-
ogy and (2) rebar–concrete interfacial behavior.
Hydraulic Pressure Analogy
Tepfers (1973) developed one of the earliest analytical solutions for
the rebar pullout bond strength in a thick-walled concrete cylinder
with a hydraulic pressure analogy. As rebar was pulled out of the
concrete cylinder, the rapidly increased radial pressure induced
tensile stress in a tangential direction, causing radial cracks that
initiated from the inner face of the thick-walled cylinder and propa-
gated outward in a radial direction. The rebar–concrete bond
strength was reached when the radial cracks exceeded a critical
length and suddenly penetrated the remaining wall thickness. The
tangential stress in the cracked zone was evaluated from the as-
sumed crack opening displacement as a function of radial distance
1
Graduate Research Assistant, Dept. of Civil, Architectural, and Envir-
onmental Engineering, Missouri Univ. of Science and Technology, Rolla,
MO 65409-0030. E-mail: cwmqc@mst.edu
2
Professor and Abbett Distinguished Chair in Civil Engineering, Dept.
of Civil, Architectural, and Environmental Engineering, Missouri Univ. of
Science and Technology, Rolla, MO 65409-0030 (corresponding author).
E-mail: gchen@mst.edu
Note. This manuscript was submitted on October 20, 2014; approved on
February 2, 2015; published online on May 7, 2015. Discussion period
open until October 7, 2015; separate discussions must be submitted
for individual papers. This paper is part of the Journal of Engineering
Mechanics, © ASCE, ISSN 0733-9399/04015038(12)/$25.00.
Poisson effect. Reinhardt (1992) and Van der Veen (1990) assumed
a linearly distributed tangential displacement in the cracked zone
(constant tangential stress), and introduced a nonlinear softening
model as a product of exponential and power functions. The soft-
ening model involved several parameters, such as the fracture en-
ergy and softening rate of concrete. To represent nonlinear fracture
processing more accurately, Rosati and Schumm (1992) introduced
a parabolic tangential displacement in the cracked zone (linear tan-
gential stress). In addition to the softening model’s complexity, the
aggregate size effect on the fracture process of concrete was over-
estimated because of the maximum aggregate size. Olofsson and
Ohlsson (1995) and Noghabai (1995) simplified the cracked con-
crete model with a linear softening formulation for easy implemen-
tation in the numerical simulation of concrete structures with spiral
reinforcement.
Wang and Liu (2003) implemented an elastocohesive model
with bilinear concrete softening criteria. The splitting damage in
the rebar direction and the tensile damage in the hoop direction
were averaged and smeared into the cylinder model. In comparison
with Nielsen and Bicanic (2002), the elastocohesive model pro-
vided comparable bond strength when a significant number of
small cracks occurred, but overestimated the bond strength as the
ratio between the concrete cover and rebar diameter increased,
resulting in a few cracks.
Once the bursting/radial pressure applied on the thick-walled
cylinder had been determined, the rebar–concrete bond strength
was evaluated by simply assuming a 45° bearing angle or equal to
the radial pressure between rebar and concrete (Tepfers 1973). In
comparison with experimental data, the models developed with the
hydraulic pressure analogy overestimated the rebar–concrete bond
strength up to 100%. Eligehausen et al. (1983) back-calculated an
effective bearing angle of 26.5 to 45° by considering a plastic zone
of concrete over the rebar length and 1.5 times the rebar diameter
from the rebar surface, and letting the rebar–concrete bond strength
equal to 0.5–1.0 times the bursting pressure, as evaluated by
Tepfers (1973). The hydraulic pressure analogy was also chal-
lenged by Reynolds and Beeby (1982) for a spliced joint of two
bars in contact, since it would predict twice as much as the bond

© ASCE 04015038-1 J. Eng. Mech.

J. Eng. Mech.
 strength of each bar, which would significantly exceed the test
results.
Rebar–Concrete Interfacial Behavior
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
series of detailed stress analyses taking into account the rib defor-
mation and surface condition of rebar for each experimentally ob-
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
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-
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
(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
condition of the rebar. How the fictitious rib-face angle or concrete-
bearing angle affects the bond strength has yet to be investigated,
particularly when the rib spacing-to-height ratio of rebar increases.
Cairns and Jones (1995) indicated that the hydraulic pressure
analogy and rebar–concrete interface studies generally overesti-
mated and underestimated the rebar–concrete bond strength, re-
spectively. In fact, CEB-FIP Task Group (2000) stated that the
core of the bond is a balance of the confinement strength, pro-
vided by concrete cover or transverse reinforcement, and the shear
strength in the vicinity of rebar ribs. Most of the existing models
included a parameter of relative rib area that is not well supported
by the test results (Rehm 1957, 1961; Darwin and Graham 1993).
In addition, Darwin et al. (1992) showed a slight variation of the
effective bearing angle along the rebar length. The experimental
contradiction to the fixed bearing angle theory was previously
noted but never explained in the literature.
In this study, the two traditionally distinct research directions are
integrated into a unified rebar–concrete bond model, including both
concrete confinement and rebar–concrete interface behavior. The
process of a rebar rib pressing against concrete is simulated by
the indentation of a steel cone into concrete so that the stress field
near the rebar rib, and thus the effective bearing angle of the con-
crete, can be evaluated. The cone indentation analogy is modified
with clearly stated assumptions to consider the effects of multiple
ribs, rib height-to-spacing ratios, and failure surface characteristics
of steel/concrete. The analytical solution for bond equations is
validated with test data obtained from uncoated rebar in concrete
by various researchers around the world.
An Indentation Analogy for Deformed
Rebar-in-Concrete Analysis
Fig. 1 shows half a cross section of deformed rebar with sr and hr
representing the spacing and height of periodical ribs, respectively.
When the rebar embedded along the center line of a concrete cyl-
inder is pulled out of its surrounding concrete (downward in Fig. 1),
the steel ribs act as a series of conical shape indenters pressed
against and separated from concrete in their front and back faces,
respectively. As a result, concrete is often crushed along the
© ASCE 04015038-2
J. Eng. Mech.
f'v
f'c
Hydrostatic
Core
f'c
sr sr0
f'v
sflat
hr
Fig. 1. Indentation analogy
rib-front face as circled in Fig. 1. This observation indicates the
existence of a hydrostatic pressure zone near the rib-front area.
Therefore, the normal component of the rebar–concrete interfacial
force corresponds to the concrete compressive strength f c0 , as
shown in Fig. 1. The concrete crushing zone forms a new
wedge of the indenter and acts as the core zone mentioned by
Johnson (1985).
A classical indentation process differs from the debonding pro-
cess in that:
• There is no rebar–concrete interface in the flat portion and no
stress-free back face of the ribs, as indicated in Fig. 2; and
• It occurs in a semi-infinite elastic body instead of the finite body
with balanced concrete confinement and near-rib stress condi-
tion in the debonding process.
Despite the this difference, the stress field in the indentation
problem closely resembles the near-rib stress distribution in the
debonding problem, as schematically indicated by the indenter-
induced displacement field in Fig. 1. To minimize their differences,
the stress in interested areas such as the back face of ribs can be
modified to meet the stress-free condition, as indicated by the free
surface Γ in Fig. 2. In this case, the stress at the Γ surface is first
calculated from the indentation solution and then cancelled by
introducing a counter-stress vector equal in magnitude and opposite
in direction. The key line between ribs is displayed in Fig. 2.
Fig. 3 shows a conical-shaped indenter pressed into a semi-
infinite space in the Cartesian (x, y, z) and cylindrical (r, θ, z) co-
ordinate systems. As shown in Fig. 3, the radius of the indenter is
represented by a (a ≈ hr ). The normal and shear tractions along
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.
Stress Induced by Normal Traction
A semi-infinite space subjected to a conical shape indenter under
uniformly distributed normal traction pm along the side surface of
the indenter is an axis-symmetrical problem studied by Sneddon
(1948). The induced elastic stress field under the normal traction
is thus independent of the coordinate θ. The three nonzero stress
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.

Actual Case Indentation on Free Traction

Surface

Fig. 2. Stress modification at the back face of ribs: free surface Γ


ðnÞ
σz z 0 2ð1 − ν 2 Þpm
0
¼ − J1 þ J2 ð1Þ cot α ¼ ð4Þ
pm hr E

The indentation angle α is equal to or smaller than the rib–face

ðnÞ ðnÞ angle β, as the crushing concrete in front of the rib face serves as
σr 2ð1 − ν 2 Þ 0 σz part of the indentation wedge.
¼− J −
pm ð1 − νÞ 1 pm
  
h z
− 2νJ 01 þ r ð1 − 2νÞJ 10 − J 11 ð2Þ Stress Induced by Shear Traction
r hr
In Sneddon’s analysis, the influence of shear traction was ignored.
The elastic stress field induced by shear traction along the side
ðnÞ surface of the indenter was solved by Hanson (1992) using the
τ rz z
¼ − J 12 ð3Þ potential theory for transversely isotropic materials. That is,
pm hr
ðsÞ
1=2
σz λν
where the superscript n for each stress component signifies the ¼ c1 zg 0 ðzÞ ð5Þ
pm λ þ 2ν
cause of normal traction, and ν = Poisson ratio of concrete.
Eqs. (1–3) include several functions as defined in the Appendix.
The normal traction pm is related to the indentation angle α by ðsÞ
−1=4
σr λν ð1 þ 2νÞgðzÞ þ zg 0 ðzÞ
¼− c1 ð6Þ
pm λ þ 2ν 2ð1 − νÞ

ðsÞ pffiffiffi
τ rz 2
¼ c fz½f10 ðzÞ − f20 ðzÞ − ½f1 ðzÞ þ f 2 ðzÞg ð7Þ
O Pm x pm 4 1

θ where the superscript s for each stress component signifies the

cause of shear traction, and λ ¼ Eν=½ð1 þ νÞð1 − 2νÞ is the
T r
Lame’s constant. The superscript ð·Þ 0 represents the first derivative
of the function inside the bracket with respect to its argument.
Eqs. (5–7) also include several functions as given in the Appendix.

y Radial Stresses along Key Line

Along the key line, however, the total radial stress due to the com-
bined normal traction and shear traction, Eqs. (2) and (6), must be
modified to achieve a zero-stress boundary at the back face of the
rebar rib or at the free surface Γ shown in Fig. 2. To this end, the
total stress along the free surface due to indentation is first calcu-
lated and then cancelled by applying a counter stress along the free
surface, which is equal to the total stress with opposite sign. The
stress caused by the counter stress can be approximately evaluated
z using Timoshenko’s theory of the concrete beam between two
adjacent ribs. As such, the final radial stress along the key line
Fig. 3. Cartesian and cylindrical coordinate systems
(tension positive) can be calculated by

© ASCE 04015038-3 J. Eng. Mech.

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

© ASCE 04015038-4 J. Eng. Mech.

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. Illustrative stress distributions with various Sr0 =hr ratios

Fig. 5 presents the stress distributions at various boundaries
when sr0 =hr varies (other parameters remain the same as used
in Fig. 4). The stress state near the rib areas can be divided into
three groups: sr0 =hr ≤ 7, 7 < sr0 =hr ≤ 10, and sr0 =hr > 10, as
discussed below:
• For sr0 =hr ≤ 7, the radial stress generated along the key line is
in tension near the top surface. The tensile stress along the
rebar–concrete interface is approximately zero in the flat por-
tion of the rebar;
• For 7 < sr0 =hr ≤ 10, the radial stress generated along the key
line is in compression. The integration of the radial stress along
the vertical axis over a distance of hr constitutes approximately
77% of the total integration over the entire vertical axis. The
tensile stress along the rebar–concrete interface indicates no
contact between the rebar and concrete in the flat portion of the
rebar. In this case, the influence of the vertical stress at the free
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
concrete remains in contact in that area. Therefore, the flat por-
tion should be considered in the evaluation of bond strength;
• 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
equations are detailed below.
Low Rib Spacing-to-Height Ratio: s r 0 =h r ≤ 7
failure can initiate and propagate through the concrete key line.
At the ultimate state immediately prior to the plow through failure,
the bearing stress on the rib front face reaches the concrete com-
pressive strength fc0 . Let c2 ¼ fv0 =f c0 be the ratio between the shear
stress on the rib front face and the concrete compressive strength.
Here, the uniaxial concrete strength is used, As the triaxial confine-
ment is impossible to form because of early radial crack growth.
For a low rib spacing-to-height ratio, the radial pressure in Eq. (8)
is insignificant and can be represented by its uniform distribution
over the considered length pn , as shown in Fig. 6. In applications,
pn can be approximated as uniformly distributed radial stress at the
key line due to the effect of concrete cover. Additionally, there is no
contact force on the flat portion of the rebar. At the imminent plow
through failure, the horizontal forces per linear thickness of the
concrete key are in equilibrium, resulting in the critical rib-face
angle β cr
Tension
Opening
f'c
pn
β
f'v

Failure Mechanism and Corresponding Experimental

Findings
For a low rib spacing-to-height ratio, the concrete at the rib front
underwent gradual crushing with a rib-face angle exceeding 40° fb
(Rehm 1957, 1961). The early observation was confirmed by
Darwin and Graham (1993) with a critical rib-face angle of 60°.
This experimental observation can be explained by the possible Key Line
concrete tearing-off along the key line as illustrated in Fig. 6. The
torn part was gradually crushed, mainly under the action of inter-
face forces on the rib front face and radial pressure at the key line.
Shear Force
Critical Rib-Face Angle
Fig. 6. Failure mechanism and free body diagram at low rib spacing-
The critical rib-face angle exists because the force along the rebar
to-height ratio
direction must be sufficiently large to ensure that the plow through

© ASCE 04015038-5 J. Eng. Mech.

J. Eng. Mech.


p s
β cr ¼ arccot c2 þ n0 r0 ≤ β ð9Þ
f c hr
The critical rib-face angle must be equal to or smaller than β,
because the steel rib remains elastic. When c2 ¼ 0.83 (Rehm 1957,
1961), sr0 =hr ¼ 6, sflat ≅ hr and pn =fc0 ¼ 0.06, β cr ¼ 40°, which
is less than the rib angle β ¼ 60° widely used in representative
deformed rebar.
To ensure that the considered balancing pressure pn is realistic
in application, a thick-walled hollow cylinder with inner and outer
diameters of db and db þ 2c (db = rebar diameter and c = clear
concrete cover) is analyzed under internal pressure pn . When db ¼
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-
sure corresponding to a hoop stress equal to the tensile strength
of concrete is pn ¼ 1.1ft (ft is the uniaxial tensile strength of
concrete). When f t ≈ 0.1f c0 , pn =fc0 ≈ 0.11. When the stress con-
centration near the rib is considered, the pn =fc0 ¼ 0.06 value seems
reasonable.
On the basis of various tests (Idun and Darwin 1999), the aver-
age c2 value of uncoated rebar in normal-strength concrete is ap-
proximately 0.6. According to Eq. (9), its corresponding critical
rib-face angle β cr is 59° when pn is neglected. For their specimens,
db ¼ 25.2 mm and c ¼ 50.4 mm were also used.
To facilitate the understanding of the plow through failure mode,
the initial tearing-off condition at the key line must be evaluated.
At the beginning of the plow through failure, the concrete key re-
mains intact and can be used as a cantilever beam to estimate the
near-rib stress conditions under the interface forces, as shown in
Fig. 6. Specifically, the average shear stress σrz over the key line
and the flexural tension stress σrr at the extreme fiber of the key line
section can be evaluated by
hr
σrz ¼ fc0 ð1 þ c2 cot βÞ ð10aÞ
sr0

2
hr
σrr ¼ 3fc0 ð1 þ c2 cot βÞ ð10bÞ
sr0
The shear-to-flexural stress ratio can then be expressed into a
function of sr0 =hr by
σrz s
¼ r0 ð11Þ
σrr 3hr
The shear strength is equal to 0.5ft for a uniaxial stress state and
f t for a pure-shear stress state. For a less conservative estimate, the
shear strength equal to f t can be considered. Therefore, the shear
failure governs when sr0 =hr ≥ 3 in applications and the plow
through failure mode is accompanied by the shear failure along
the key line. If the lower shear strength 0.5ft were used, the shear
failure along the key line would always govern.
Bond Strength
Before the imminent plow through failure, the total vertical force
per linear thickness of the concrete key by neglecting the pn effect
can be determined and used to evaluate the average bond strength
f̄ b over the rib spacing sr
hr
f̄ b ¼ fc0 ð1 þ c2 cot βÞ ð12Þ
sr0
© ASCE 04015038-6
J. Eng. Mech.
Medium Rib Spacing-to-Height Ratio: 7 < s r =h r 0 ≤ 10
Failure Mechanism and Corresponding Experimental
Findings
As the rib spacing-to-height ratio falls into a medium range of 7
to 10, the shear mechanics dominates the bond behavior of rebar
in normal strength concrete. In this case, the force equilibriums in
two orthogonal directions are used to determine the effective bear-
ing angle. Depending upon the shear strength and confinement
level, the effective bearing angle varies and leads to different failure
patterns.
Previous research has shown that the concrete at the rib-front
area becomes compact powder because of high stress concentration
(Lutz and Gergely 1967; Esfahani and Rangan 1998). It was
found that the effective bearing zone plays a critical role in trans-
ferring the bearing component of bond forces to their surrounding
concrete.
Existence of Crushing Zone
The cross section of half a reinforcing bar with a crushing zone
at the rib-front area is presented in Fig. 7. Possible failures for
an intermediate rib spacing-to-height ratio include three cases:
(1) rib sliding for weak rebar–concrete interface, (2) concrete
crushing for relatively weak rebar–concrete interface, and (3) con-
crete shear-off for strong rebar–weak interface. As illustrated in
Fig. 7, Case (1) can be considered as a special case of (2) when
the crushing angle α is equal to the rib-face angle β. Case (3) is
similar to Case (2) except that the rebar–concrete interface strength
and concrete strength are in different magnitude order. Therefore,
the general condition for Case (2) is discussed in detail.
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
with α angle are fn ¼ f c0 and fv ¼ c2 f c0 . The radial and vertical
force equilibriums of the uncrushed part of the rib-front area in
Fig. 7 (unit thickness out of plane) yield
f c0 t1 cos α − c2 f c0 t1 sin α − pn t2 ¼ 0 ð13aÞ
−fc0 t1 sin α − c2 fc0 t1 cos α − f b t2 ¼ 0 ð13bÞ
where t1 and t2 represent the lengths of the crushed plane and its
vertical projection along the key line, t2 ¼ t1 cos α, f b is the
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
relation between the calculated α and the β value. When α < β, the
crushing zone is present. For the three cases in Fig. 7, failure can
occur inside the concrete c2 ¼ cc related to the cohesion and in-
ternal friction angle and at the rebar–concrete interface c2 ¼ ci re-
lated to the coefficient of friction. For example, ci ¼ 0.53 and
cc ¼ 0.25 þ tan 30° ¼ 0.83 (Cairns and Abdullah 1996).
In this way, the interfacial strength is quantified by the param-
eter c2, and it is bounded between ci and cc . The confining effect
is then quantified by the parameter c0 ∶ ¼ ffnc0 . Therefore, the effective
bearing angle can be evaluated with these two parameters for differ-
ent failure cases.
Following is a brief summary of the effective bearing angle α
and the average bond strength f̄b in three cases:
1. Rib sliding [α ≥ β; c2 ¼ ci ≤ ð1 − c0 Þ cot β]
The pull-out specimen fails at the rebar–concrete interface.
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

Fig. 7. Failure mechanism at medium rib spacing-to-height ratio

f̄b h 1 þ c2 cot β was underestimated as experimentally demonstrated by Darwin and

α¼β and ¼ r ð14Þ Graham (1993).
pm sr0 c0

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

Failure Mechanism and Corresponding Experimental µe pn Rib Back Face

Findings
As previously discussed, the flat portion of rebar can be divided
into noncontact and contact areas. The effect of the contact area
on the bond strength must be considered. Otherwise, the bond
Fig. 8. Failure mechanism at high rib spacing-to-height ratio
strength by considering the near-rib effect in noncontact areas only

© ASCE 04015038-7 J. Eng. Mech.

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.
(a) (b)
Fig. 9. Softening behavior of concrete cover: (a) axis-symmetric
problem; (b) stress–strain relation in tangential direction
estimated from the case of medium rib spacing-to-height ratio with
a rib spacing equal to 10hr and ci ¼ 0.53. Because the rib back face
is stress-free, the flat portion effect f̄b−flat will not be associated
with concrete compressive failure and can be calculated from
the Mohr–Coulomb failure criterion at the rebar–concrete contact
surface. That is, the average bond strength of the sliding flat portion
associated with radial concrete splitting is
f̄ b−flat ¼ ci pn ð17Þ
Maximum Radial Pressure
The average bond strengths in the three cases discussed in previ-
ously are all related to the radial pressure pn within a depth of
hr along the key line (Losberg and Olsson 1979; Soretz and
Holzenbein 1979). The radial pressure can be approximately evalu-
ated by the maximum stress p0 , as shown in Fig. 9(a). To evaluate
the radial pressure associated with concrete cover, a plane-strain
axisymmetric problem with a hollow cylinder is considered, as
shown in Fig. 9(a). The cylinder can be divided into two parts: inner
inelastic and outer elastic regions. The outer elastic region is a hol-
low cylinder under internal pressure pe . The inner and outer radii
of the hollow cylinder are re and rc , respectively. In this case, the
classical Lame’s solution is directly used to relate the internal pres-
sure pe to the maximum tangential stress f tu at r ¼ re
r2c − r2e
pe ¼ f tu ð18Þ
r2c þ r2e
The inner inelastic region included many axis-symmetrically
distributed radial cracks that cause tension softening in concrete
and is accounted for in this study by the smeared crack model
(Hillerborg et al. 1976). The tangential/hoop stress–strain
(σt − εt ) relation, as shown in Fig. 9(b), includes low-strain elastic
and high-strain inelastic behaviors. It can be mathematically ex-
pressed into
 These parameters correspond to cc ¼ 0.83. For the interface be-
Ec εt ; εt ≤ εt0
σt ðεt Þ ¼ ð19Þ
ftu exp½−ðεt − εt0 Þ=ðεtu − εt0 Þ; εt0 ≤ εt ≤ εtu
where f tu = ultimate tensile stress corresponding to the maximum
elastic strain in concrete εt0 ð¼ 0.0001Þ, and εtu ð¼ 0.002Þ is the
δt
ultimate tangential strain. Let a tangential strain of εt0 ¼ 2πr e
at
the crack front in the smeared crack model be equivalent to the
effect of radial cracks in the smeared crack model (Hillerborg
et al. 1976). If the radial displacement is assumed as a constant in
© ASCE 04015038-8
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
after Eq. (18) has been introduced.
Hence, the maximum p0 is achieved when the first derivative of
Eq. (20) with respect to re is set to zero. The corresponding re value
is referred to as the critical radial crack length, which can be found
from the following characteristic equation:
Z 
r4c − 5r4e 1 re ∂σ
t
þ dr þ σ ðr Þ ¼0 ð21Þ
ðr2c þ r2e Þ2 ftu rs ∂re
t e
The critical radial crack length can be numerically solved from
Eq. (21) and substituted back into Eq. (20) to evaluate pn .
Model Validation and Comparison
Test Database and Competing Methods
Experimental data from beam-end and cylinder specimens are used
to validate the unified model and compare its performance with
existing models. Although most of the beam-end specimens have
an embedment length of over five times the diameter of rebar
and thus may result in nonuniform bond behavior, they are still
included in the database because of wide applications and adop-
tions by many researchers. More importantly, prior research find-
ings (Tepfers 1973; Esfahani and Kianoush 2005) indicate the
average bond strength of long-embedment specimens can be in-
creased by only 1.5%.
The selected experimental data (Choi et al. 1991; Darwin and
Graham 1993; Idun and Darwin 1999; Miller et al. 2003; De Anda
et al. 2004; Wu et al. 2012) were obtained mainly from local
bond tests with concrete cover splitting as a primary failure mode.
Overall, the collected database resulted from a total of 284 tests
with uncoated steel rebar, each repeated with at least two samples.
The database covers various test parameters such as rebar size, rib
geometry, concrete strength, and concrete cover.
Both empirical approaches (Orangun et al. 1977; Zuo and
Darwin 2000) and theoretical approaches (Wang 2009; Wang and
Liu 2003) are considered for comparison. As the state-of-the-art
methodologies in bond-strength determination, Wang (2009) and
Wang and Liu (2003) was focused on the rebar–concrete interface
and hydraulic pressure analogy studies, respectively.
Test-over-Prediction Ratio of Bond Strength
According to Cairns and Abdullah (1996), the cohesion of concrete
was estimated to be 0.25fc0 , and the internal frictional angle is 30°.
tween uncoated rebar and concrete, ci ¼ 0.6 with a cohesion of
0.11f c0 and a frictional angle of 28°.
The test-over-prediction bond-strength ratios of uncoated rebar
in concrete by various researchers are compared in Table 1 and
Fig. 10. Both the mean and the coefficient of variation (COV)
of various bond models using individual data sets and the overall
database are presented in Table 1.
When all the test data sets were used, this study resulted in
a mean ratio of 1.019, closest to one, with the smallest COV value,
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.

Wu et al. (2012) 12 Mean 1.384 0.724 1.148 0.723 1.041

— COV 0.232 0.262 0.156 0.114 0.036
All 133 Mean 1.481 1.121 1.053 0.729 1.019
— COV 0.186 0.201 0.114 0.098 0.069
a
Only applicable with rib spacing-to-height ratio ranging between 10 and 12, including 13 tests.

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)

Fig. 10. Test-over-predicted ratios for uncoated rebar

and was thus the most accurate prediction in bond strength of
uncoated rebar in concrete. The theoretical approach taken by
Wang (2009) yielded the second most accurate prediction. The em-
pirical approach by Orangun et al. (1977) led to the least accurate
results. Overall, the theoretical approaches included in this study
are more accurate than the empirical approaches, because the em-
pirical models lack a solid mechanical basis and are thus less ver-
satile in predicting bond behaviors under the different conditions
investigated by various researchers. The theoretical approach by
Wang and Liu (2003) resulted in a nonconservative prediction
as they neglected the effect of rebar surface characteristics such
as rib geometries and interface bond strength.
When individual data sets were used, the bond-strength predic-
tion accuracies by various groups were also compared. With the
data sets developed by others, Orangun et al. (1977) among all
the models predicted the least accurate bond strength against
every data set. On one hand, the two most accurate predictions
by Zuo and Darwin (2000) corresponded to the data sets developed
by the same group (Idun and Darwin 1999; Miller et al. 2003).
On the other hand, the least accurate prediction by Zuo and Darwin
(2000) was also for the data set generated by the same group
(Darwin and Graham 1993). Therefore, the prediction accuracy by
empirical approaches most likely depended on the test conditions
and data sets used in the regression analysis.
Among the three theoretical approaches, the proposed model in
this study is most flexible and applicable to various conditions such
as for coated rebar (Table 2 and Fig. 11). Wang (2009) did not take
full consideration of plastic behaviors during concrete splitting, re-
sulting in a prediction that is insensitive to the variation in confine-
ment extent (Wu et al. 2012). Wang and Liu (2003) neglected the
effects of rebar surface characteristics such as rib geometries and
interface bond strength. Furthermore, most theoretical approaches
failed to distinguish various failure modes associated with different
deformation patterns.

© ASCE 04015038-9 J. Eng. Mech.

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.

All 151 Mean 1.618 1.058 1.031 0.763 1.030

— COV 0.287 0.209 0.228 0.101 0.071
a
Rebar coated with fusion-bonded epoxy coating.
b
Rebar coated with enamel coating.

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)

Fig. 11. Test-over-predicted ratios for coated rebar

Conclusions crushing and both interface and concrete shear-off, determin-

A unified bond theory of deformed rebar in concrete has been
developed to understand local bond behavior. Its deduced bond-
strength equations for various practical scenarios have been vali-
dated with experimental data sets available to this study. Using
extensive analysis, comparison, and validation, several conclusions
can be drawn:
1. The proposed unified bond theory combines the indentation
analogy for near-rebar stress analysis and the hydraulic
pressure analogy for concrete confinement analysis. It sheds
insight on various failure-mode transitions by varying rebar–
concrete interface bond strengths and unifies two traditionally
distinct bond models based on shear-stress analysis and the
hydraulic pressure analogy.
2. At low rib spacing-to-height ratios, the potential failure modes
near rebar ribs are plow through with concrete shear-off along
the key line between ribs. At medium and high rib spacing-
to-height ratios, the likely failure modes involve concrete
ing the effective bearing angle of rebar in concrete. As the rib
spacing-to-height ratio increases, the role that the confinement
provided by concrete cover plays in the occurrence of the fail-
ure modes becomes more critical.
3. The critical concrete cover to ensure a concrete shear-off fail-
ure, the critical rib-face angle to determine concrete shear-off
and crushing, and the critical rib spacing-to-height ratio to
control different failure modes can all be explained by the uni-
fied theory and match with experimental findings from various
researchers.
4. The average predicted average bond strengths are in good
agreement with test results with less than 6% relative error.
The predication accuracy seems to be consistent with all three
different rib spacing-to-height ratio cases (Fig. 12). The pro-
posed bond equations are more accurate than at least two
theoretical and two empirical approaches available in the lit-
erature. They are demonstrated to be robust in all application
scenarios with various confinement conditions in terms of

© ASCE 04015038-10 J. Eng. Mech.

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

0.8 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

1
l2 ðhr Þ ¼ ðr þ hr Þ2 þ z2 þ ðr − hr Þ2 þ z2 ð33Þ
Downloaded from ascelibrary.org by University of Texas At Austin on 07/10/15. Copyright ASCE. For personal use only; all rights reserved.

0.7 Low Medium High 2

0.6
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Sr0 /hr
Acknowledgments
Fig. 12. Test-over-predicted ratios versus rib spacing-to-height ratio
Financial support for this study was provided in part by the
National Science Foundation under Award No. CMMI-0900159
and by the Dept. of Civil, Architectural, and Environmental Engi-
different concrete-cover thickness. The effect of external steel neering at Missouri Univ. of Science and Technology. The findings
or fiber-reinforced polymer jacket can be easily treated as ad- and opinions expressed in this paper are those of the authors only
ditional confining pressure, which will be investigated in the and do not necessarily represent those of the sponsors.
future study.

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

r hr hr reinforcement.” ACI J., 89(6), 709–720.

Darwin, D., Tholen, M. L., Idun, E. K., and Zuo, J. (1996). “Splice length
  of high relative rib area reinforcing bars.” ACI Struct. J., 93(1), 95–107.
1 r 0 hr z
J 10 ¼ J 1 þ ð1 − R sin ϕÞ − J 11 ð26Þ De Anda, L., Courtier, C., and Moehle, J. P. (2004). “Bond strength of pre-
2 hr r hr fabricated epoxy-coated reinforcement.” Final Rep., 04-01, Univ. of
California, Berkeley/Caltrans, Berkeley, CA.

2
2 2
2 1=4 Eligehausen, R., Ropov, E., and Bertero, V. V. (1983). “Local bond stress-
r z z slip relationships of deformed bars under generalized excitation.”
R¼ þ −1 þ4 ð27Þ
hr hr hr Rep. No. UCB/EERC-83/23, Univ. of California, Berkeley.
Esfahani, M. R., and Kianoush, M. R., (2005). “Development/splice length

2
2 −1 of reinforcing bars.” ACI Struct. J., 102(1), 22–30.
hr z r z Esfahani, M. R., and Rangan, B. V. (1998). “Local bond strength of
tan ϑ ¼ ; tan 2ϕ ¼ 2 þ −1 ð28Þ reinforcing bars in normal strength and high-strength concrete
z hr hr hr
(HSC).” ACI Struct. J., 95(2), 96–106.
Ferguson, P. M., and Breen, J. E. (1965). “Lapped splices for high-strength
The functions in Eqs. (3–5) are defined as follows:
reinforcing bars.” ACI J., 62(9), 1063–1078.
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffi Ferguson, P. M., and Briceno, A. (1969). “Tensile lap splices-part 1: retain-
l22 ðhr Þ − h2r − r2 þ z2
gðzÞ ¼ ð29Þ ing wall type, varying moment zone.” Research Rep. No. 113-2, Center
r for Highway Research, The Univ. of Texas at Austin, Austin, TX.

© ASCE 04015038-11 J. Eng. Mech.

J. Eng. Mech.
 Hanson, M. T. (1992). “The elastic field for conical indentation including
sliding friction for transverse isotropy.” J. Appl. Mech., 59(2),
S123–S130.
Harajli, M. H., Hout, M., and Jalkh, W. (1995). “Local bond stress-slip
behavior of reinforcing bars embedded in plain and fiber concrete.”
ACI Mater. J., 92(4), 343–354.
Hillerborg, A., Modeer, M., and Petersson, P. E. (1976). “Analysis of crack
formation and crack growth in concrete by means of fracture mechanics
and finite elements.” Cem. Concr. Res., 6(6), 773–781.
Idun, E. K., and Darwin, D. (1999). “Bond of epoxy-coated reinforcement:
coefficient of friction and rib face angle.” ACI Struct. J., 96(4),
609–616.
Johnson, K. L. (1985). Contact mechanics, Cambridge University Press,
Cambridge.
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
influencing bond cracking.” ACI J., 77(3), 47–71.
Losberg, A., and Olsson, P. A. (1979). “Bond failure of deformed reinforc-
ing bars based on the longitudinal splitting effect of the bars.” ACI J.,
76(1), 5–18.
Losberg, A., and Olsson, P.-A. (1979). “Bond failure of deformed reinforc-
ing bars based on longitudinal splitting effect of the bars.” ACIJ., Proc.,
76(1), 5–18.
Lutz, L. A., and Gergely, P. (1967). “Mechanics of bond and slip of de-
formed bars in concrete.” ACIJ. Proc., 64(11), 711–721.
Miller, G. G., Kepler, J. L., and Darwin, D. (2003). “Effect of epoxy coating
thickness on bond strength of reinforcing bars.” ACI Struct. J., 100(3),
1–7.
Nielsen, C. V., and Bicanic, N. (2002). “Radial fictitious cracking of thick-
walled cylinder due to bar pull-out.” Mag. Concr. Res., 54(3), 215–221.
Noghaba, K. (1995). “Splitting of concrete in the anchoring zone of de-
formed bars-a fracture mechanics approach to bond.” Doctoral thesis,
Lulea Univ. of Technology, Lulea, Sweden.
Olofsson, T., and Ohlsson, U. (1995). “A simple fracture mechanics model
for mixed-mode failure in concrete.” Int. Conf. Fracture Mechanics of
Concrete Structures, Zurich, Switzerland, 25–28.
Orangun, C. O., Jirsa, J. O., and Breen, J. E. (1977). “A reevaluation of test
data on development length and splices.” ACI J., 74(3), 114–122.
© ASCE 04015038-12
View publication stats J. Eng. Mech.
Rehm, G. (1957). “Fundamental law of bond.” Proc. Symp. Bond and
Crack Formation in Reinforced Concrete (Stockholm), Tekniska
Hogskolans Rotaprinttrychkeri, Stockholm, Sweden.
Rehm, G. (1961). “Uber die grunlagen des verbundes zwischen stahl und
beton [Basic principle of the bond between steel and concrete].”
Deutscher Ausschuss fur Stahlberton, 1381(5), 59 (in German).
Reinhardt, H. W. (1992). “Bond of steel to strain-softening concrete taking
account of loading rate.” Fracture Mechanics of Concrete Structures,
Z. Bazant, ed., Elsvier, London, 80–820.
Reynolds, G. C., and Beeby, A. W. (1982). “Bond strength of deformed
bars.” Bond in Concrete, P. Bartos, ed., Applied Science, London,
434–445.
Rosati, G., and Schumm, C. (1992). “Modeling of local bar-to-concrete
bond in reinforced concrete beams.” Int. Conf. Bond in Concrete—
From Research to Practice, A. Skudra and T. Tepfers, eds., Riga
Technical University, Riga, Latvia, 12–43.
Sneddon, I. N. (1948). “Boussinesq’s problem for a rigid cone.” Math.
Proc. Cambridge Philos. Soc., 44(4492), 492–507.
Soretz, S., and Holzenbein, H. (1979). “Influence of rib dimensions
of reinforcing bars on bond and bendability.” ACI J. Proc., 76(1),
111–128.
Tepfers, R. (1973). “A theory of bond applied to overlapped tensile
reinforcement splices for deformed bars.” Publication No. 73:2, Div.
Concrete Structures, Charlmers Univ. of Technology, Goteborg,
Sweden, 328.
Van der Veen, C. (1990). “Cryogenic bond stress-slip relationship.”
Doctoral thesis, Delft Univ. of Technology, Delft, The Netherlands.
Wang, H. (2009). “An analytical study of bond strength associated with
splitting of concrete cover.” Eng. Struct., 31(4), 968–975.
Wang, X., and Liu, X. (2003). “A strain-softening model for steel-concrete
bond.” Cem. Concr. Res., 33(10), 1669–1673.
Wu, C., Chen, G., Volz, J. S., Brown, R. K., and Koenigstein, M. L. (2012).
“Local bond strength of vitreous enamel coated rebar to concrete.”
Constr. Build. Mater., 35, 428–439.
Zuo, J., and Darwin, D. (2000). “Splice strength of conventional and high
relative rib area bars in normal and high-strength concrete.” ACI Struct.
J., 97(4), 630–641.
J. Eng. Mech.