Construction and Building Materials 95 (2015) 124–132

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Behavior of post-installed large-diameter anchors in concrete

foundations
Dongpo Wang a, Dongsheng Wu a, Siming He b,c,⇑, Jun Zhou d, Chaojun Ouyang b
a
State Key Laboratory of Geohazard Prevention and Geoenvironment Protection, Chengdu University of Technology, Chengdu 610059, China
b
Key Laboratory of Mountain Hazards and Surface Process, Chinese Academy of Science, 610041 Chengdu, China
c
Center for Excellence in Tibetan Plateau Earth Sciences, Chinese Academy of Sciences, Beijing 100101, China
d
China 19th Metallurgical Corporation, China

h i g h l i g h t s

 Prototype experiments of post-installed large-diameter anchors are studied.

 The classic failure modes of post-installed large-diameter anchor systems are given.
 Enrich the literature and empirical data on post-installed large-diameter anchors.

a r t i c l e i n f o a b s t r a c t

Article history: The pull-out force for post-installed large-diameter anchors embedded in concrete foundations was
Received 8 February 2015 investigated. The main aim is to determine the optimum bonding force and ultimate tensile force of
Received in revised form 29 June 2015 the anchor bar. Anchor bars with various diameters were tested, and different embedment depths were
Accepted 14 July 2015
used. The results obtained indicate that the pull-out force improved marginally with an increase in the
bar diameter. The primary failure modes observed for the large-diameter anchors included steel bar pull-
out, concrete annulus damage and combined cone damage. Increasing the bar diameter gradually chan-
Keywords:
ged the failure mode from steel bar pullout to a combination of cone damage and concrete annulus
Post-installed anchor
Large-diameter bar
damage. Finally, grooved bars were found to be more suitable for use in post-installed anchor systems.
Pull-out test Ó 2015 Elsevier Ltd. All rights reserved.
Failure modes

1. Introduction
After decades of growth and development, industrial and civil
buildings all over the world are entering a phase of gradual deteri-
oration and aging, the demand for improved safety of existing
infrastructure has rapidly increased. This is particularly the case
in China. The result increasing in the threat to safety as well as con-
cerns about reliability has led to such building structures requiring
continuous maintenance and repair. In seismic regions, such as
south-western China, the structural deficiency of existing con-
structions is a particularly acute problem, because of lacunae in
construction knowledge and the design code. In addition, as con-
crete structures become more complex, the systems used to con-
nect steel to concrete and concrete to concrete add to the
problem [1]. This had made it necessary to demolish aging build-
ings and to reconstruct them in a structurally satisfactory manner.
⇑ Corresponding author at: Key Laboratory of Mountain Hazards and Surface
Process, Chinese Academy of Science, 610041 Chengdu, China.
E-mail address: hsm@imde.ac.cn (S. He).
However, the direct and indirect costs involved in the demolition
and reconstruction of structurally deficient constructions are often
prohibitive [2]. Therefore, structural retrofitting gets higher prior-
ity from engineers and is becoming increasingly the first choice of
action. A number of remedial methods can be implemented in the
case of existing buildings, in order to improve their mechanical
characteristics. Among the various methods available, the use of
anchor systems is one of the most effective ones.
Anchor systems are commonly used in plain or reinforced con-
crete, as well as for structurally strengthening and retrofitting
existing constructions and for connecting new structural elements
to existing ones. Using post-installed anchor systems is a very effi-
cient connection technique for structural reconstruction and
expansion, and there have been a large number of studies on the
behavior of adhesive anchors. Cook et al. [3] studied the database
including 888 European and American tests and developed a better
and more user-friendly design model than those existing previ-
ously. Eligehausen et al. [4] proposed a new failure mode and ver-
ified its validity using a worldwide database containing 415 tests.
Further, Eligehausen has [5] reviewed the development of adhesive

http://dx.doi.org/10.1016/j.conbuildmat.2015.07.129
0950-0618/Ó 2015 Elsevier Ltd. All rights reserved.
 D. Wang et al. / Construction and Building Materials 95 (2015) 124–132
anchors in the United States. Finally, abundant design principles
for engineers referring in ACI318 [6].
In recent years, numerous studies have been performed to
investigate the behavior of adhesive anchors and the effects of
the embedment length, the type of anchoring agent used, the edge
distance, and the spacing of anchors in concrete substrates.
Pull-out tests were performed by Shahi et al. [7] on steel bars
anchored at two different embedment lengths using materials
from two different manufacturers. Makoto et al. [8] studies the
effects of a free edge on the pull-out strength of anchor systems
both experimentally and analytically. Researchers have also inves-
tigated the behavior of post-installed anchors in both natural stone
[2,9] and masonry [10]. At the same time, a few studies have inves-
tigated the behavior of adhesive anchors under external loading.
The tensile behavior of post-installed chemical anchors embedded
in low-strength concrete has also been studied [11], and so has the
shear behavior of epoxy anchors embedded in low-strength con-
crete [12]. Kwon et al. studied the behavior of post-installed shear
connectors under static and fatigue loading [13]. Further, a com-
prehensive description of the effects of the loading rate on the
load-bearing behavior of anchors with different failure modes
has also been reported [14]. Several failure modes [14–18] have
been proposed in the literature for determining the failure load
of adhesive anchors in concrete. In most cases, these modes are
related to specific products, and the diameter of the test bars used
is usually less than 40 mm.
Unfortunately, anchor bars with diameters of less than 40 mm
do not satisfy the requirements for industrial buildings, especially
in the metallurgical industry, and very limited information is avail-
able on the behavior of post-installed large-diameter anchors in
concrete foundations. In this study, pull-out tests were performed
on large-diameter anchors (diameter (U) = 36 mm, 48 mm, 90 mm,
or 150 mm) for different embedment depths (8d and 12d) using
different anchoring agents (organic and inorganic agents) and steel
bars with different types of surfaces (plain and grooved), in order
to investigate the behavior of adhesive anchors embedded in con-
crete under tensile loads. The aim was to establish the relationship
between the pull-out force and the characteristics of the
post-installed large-diameter anchors, as well as establish design
guidelines for retrofitting of old equipment in large steel plants.
2. Background of anchor systems
2.1. Types of anchors
Post-installed anchors are embedded in holes that are drilled
into existing concrete foundations. Thus, any load applied to the
supported structure is passed by the anchoring system to the pri-
mary system through the frictional forces that develop between
the sides of the holes and the anchor wedges and sleeves or any
other mechanical locking devices attached. The mechanism of load
transfer depends on the type of anchor system installed.
Traditionally, on the basis of the load-transfer mechanism
involved, post-installed anchors are classified into two categories:
(1) Mechanical anchors: these anchors transfer load through
friction and mechanical interlocking.
(2) Adhesive anchors: these anchors rely on adhesion between
the anchor and the adhesive or that between the adhesive
and the concrete to transfer loads [2].
2.2. Adhesive anchor system
An adhesive anchor system consists of a steel bar or threaded
rod that is inserted into a hole drilled in concrete, with a structural
125
Epoxy
Organic
Polyester
agent
Vinyl ester
Adhesive
anchors
Cement
Inorganic
Mortar
agent
Cementitious
Fig. 1. Types of adhesive anchor systems and the adhesives used in them.
adhesive acting as the anchoring agent between the concrete and
the steel. Structural adhesives for adhesive anchor systems include
both organic and inorganic anchoring agents. Organic anchoring
agents are available pre-packaged in glass capsules or as
two-component systems requiring proportioning and mixing by
the user [2], while inorganic anchor agents are grouted into the
hole using grouting tools. A schematic describing the typical adhe-
sive anchor systems is shown in Fig. 1.
2.3. Modes of failure of adhesive anchors in concrete
Anchor systems may be subjected to tensile or shear loads or a
combination of the two. In the past decades, the theoretical modes
of failure of adhesive anchors have been examined; however, all of
these studies have involved bars with small diameters (usually less
than 36 mm). And the basic failure modes were verified by many
studies. Table 1 shows the failure modes usually observed in
anchor systems.
As mentioned above, most previous studies have focused on
small-diameter anchor systems that use bars with a diameter of
less than 40 mm. Thus, there is a lack of data on large-diameter
anchor systems that use different anchoring agents and are
anchored in concrete. This is an urgent issue with respect to rein-
forcement projects in industrial plants. Thus, in this study, we
attempted to understand the behavior of post-installed
large-diameter anchors, so that safe and reliable design guidelines
can be developed for structural reinforcement.
3. Materials and methods
We performed pull-out tests to investigate the effects of large-diameter bars on
the pull-out strength of anchor systems. During these tests, three parameters were
taken into account. These parameters are listed in Table 2, while the results of the
tests performed on the large-diameter anchor systems are listed in Table 5. (See
Figs. 2 and 3).
3.1. Test apparatus
The tests consisted of applying a static pull-out force to bars embedded in a
concrete foundation. Owing to the large scale of the experiment, ready-made appa-
ratus was not available for the tests. Thus, a customized loading device was devel-
oped. The test apparatus, shown in Fig. 4, consisted of 4 sets of individual hydraulic
jacks (QF320T, maximum pressure of 320 tons), an ultrahigh-pressure oil pump
(ZB4-500), a displacement meter (JCQ), and a static stress test and data acquisition
system (DH3815N).
3.2. Test materials
The test materials were selected carefully on the basis of purpose of the exper-
iment. Plain concrete was used for the concrete foundation. Steel bars with two
types of surfaces (plain and grooved) and four different diameters were used.
Finally, flowing grout was used as the anchoring agent.
126 D. Wang et al. / Construction and Building Materials 95 (2015) 124–132

Table 1
Possible failure modes.

Failure mode Concrete cone Combined cone mode Steel bar pullout
Adhesive/cone Steel/adhesive Adhesive/cone and steel/
interface interface adhesive interface
Characteristics Embedment depth is less Embedment depth is greater than a certain depth Embedment depth is adequate
than a certain depth
Sketch of failure mode [2,3]

dimensions (b  d  s) of the annular grooves were 30  5  100 mm, while those

Table 2
of the vertical grooves (b  d) were 30  5 mm; here, b is the groove breadth, d is
Characteristics of the test specimens.
the groove depth, and s is the groove space.
Surface type
Diameter (mm) Plain Grooved 3.2.3. Anchoring agent
Flowing grout was used as the anchoring agent in this study. 42.5R P.O (ordi-
36 8d 12d 8d 12d nary Portland cement) and sand from local quarries were used for the aggregate.
48 8d 12d 8d 12d An S.F. admixture produced by the Building Research Institute of China 19th
90 8d 12d 8d 12d Metallurgical Corporation was used as the inorganic anchoring agent. The composi-
150 8d 12d 8d 12d tion of the flowing grout used is listed in Table 3.

3.3. Testing
Table 3
Composition of flowing grout used. 3.3.1. Installation of post-installed anchors
The installation of anchors is a critical process that has a significant effect on
Material Concentration (kg/m3) Compressive their safety. The parameter of the hole diameter and effective hole depth are listed
Strength (MPa) in Table 4, and the installation process used in the present study could be divided
into the following phases (see Fig. 5).
7d 14d
42.5R P.O 850 44 54
3.3.2. Loading procedure
Sand 850
In this study, we used an equivalent gradient loading system. Each load level
S.F. Admixture 93.5
was set at 10% of the estimated value, and the design load was applied in 10 steps
Water 400
per load in 2-min intervals. The pull-out stress was recorded once every 5 s using a
static stress tester (DH-3812N). Further, the displacement of each point was mea-
sured using a displacement meter.
The following were the conditions under which the test was terminated:
Table 4
(1) The anchor bar is being pulled up; the volume continues to grow; and there
Hole diameters and depths used in this study; these were in keeping with the HILTI
are no visible signs of stabilization for over 1 h.
standard.
(2) Efforts to apply a load fail or the load does not stay stable after being
Serial Diameter of Hole diameter Effective hole applied.
number bars (mm) (mm) depth (mm) (3) The anchor bar is pulled at the ultimate yield strength.

8d 12d
Comply with the above load level before termination of test conditions, which
1 36 46 288 435 was the ultimate pull-out force of anchor bar test values.
2 48 58 384 576
3 90 100 720 1080
4 150 160 1200 1800 4. Results and discussion

In this study, a total of 48 tests were performed. In keeping with

3.2.1. Concrete foundation
In keeping with the industrial standards for most existing industrial and civil
constructions, a concrete with a characteristic compressive strength was 25 MPa
was used in this study. To meet the requirements of the large-scale in situ tests,
the concrete foundations used were shaped as squares and were poured and cured
in the open.
3.2.2. Post-installed anchor bars
Post-installed anchor bars can be either of steel or a glass-fiber-reinforced plas-
tic (GFRP). Bars of steel, which is a traditional building material, are used more
widely in engineering, including in the industrial construction industry; however,
GFRP bars are also being used in concrete structures under special conditions.
The steel bars used in this study were unmarked and of HPB345 steel (tensile
strength of 490 N/mm2 and yield strength of 345 N/mm2). Two types of steel bars
were employed: plain bars and grooved bars. For each type, bars with four nominal
diameters, U, were considered: 36 mm, 48 mm, 90 mm, and 150 mm. The
a previously reported sample-numbering scheme [2,9], each test
sample was labeled using a 5-field alphanumeric code. The mean-
ing of each field in this code is shown in Fig. 6.
For example, F-G-36-8d-1 indicates that flowing grout was used
as the anchoring agent and that a grooved bar with a diameter, U,
of 36 mm was installed in concrete as an anchor at an embedment
depth, L, of 8d mm. The last numeric field represents the number of
the sample in this series.
4.1. Displacement–pull-out force curves
The displacement–pull-out force curves for the bars with differ-
ent diameters, as determined from the flowing grout tests, are
shown in the Fig. 7. The pull-out forces for different conditions
 D. Wang et al. / Construction and Building Materials 95 (2015) 124–132 127

Table 5
Failure modes of the large-diameter anchor systems.

Diameter Anchoring Hole diameter Hole depth Surface Crack load Max. load Cone height Failure mode
(mm) agent (mm) (mm) treatment (kN) (kN) (mm)
36 Flowing 56 288 Plain bar 243 242 / Annulus crack, steel anchor pullout
grout 288 Plain bar 242 242 / Steel anchor pullout
288 Plain bar 273 273 / Steel anchor pullout
288 Grooved bar 400 467 / Annulus crack, steel bar is not pulled out
288 Grooved bar 467 467 / Steel anchor pullout
288 Grooved bar 424 424 / Annulus crack, u = 450 mm, steel anchor is not
pulled out
432 Plain bar 300 300 / Steel anchor pullout
432 Plain bar 324 324 / Steel anchor pullout, flowing grout is damaged
432 Plain bar 318 318 / Steel anchor pullout, flowing grout is damaged
432 Grooved bar 400 400 / Annulus crack flowing grout is damaged, steel
anchor is not pulled out
432 Grooved bar 400 570 35 Combined cone damage
432 Grooved bar 515 565 / Annulus crack, u = 390 mm, steel anchor is not
pulled out
48 Flowing 68 384 Plain bar 322 322 / Annulus crack, u = 240–370 mm, bar/adhesive
grout interface damage
384 Plain bar — — / Steel anchor pullout, bar/adhesive interface
damage,
384 Plain bar / Arched surface, steel anchor is pulled out, flowing
grout unset
384 Grooved bar 650 716 145 Combined cone damage, u = 1150–1280 mm
384 Grooved bar 650 716 / No damage
384 Grooved bar — — 120 Combined cone damage, u = 800–950 mm
576 Plain bar 315 315 / Annulus crack and arched surface, no cracking
576 Plain bar 249 315 130 Combined cone damage, u = 1120 mm
576 Plain bar 295 315 / Annulus crack, u = 60–120 mm
576 Grooved bar / 800 100 Combined cone damage, u = 760 mm
576 Grooved bar 500 550 80 Combined cone damage, u = 250 mm
576 Grooved bar 300 585 / Annulus crack, u = 50–100 mm
90 Flowing 130 720 Plain bar 1080 1080 / Steel anchor pullout, bar/adhesive interface
grout damage, u = 480–620 mm
720 Plain bar 1060 1060 150 Combined cone damage, u = 650–1050 mm
720 Plain bar 800 900 / Annulus crack, u = 550–740 mm
720 Grooved bar 1600 1800 / Annulus crack, u = 1450–1540 mm
720 Grooved bar 1800 1800 112 Combined cone damage, u = 620–930 mm
720 Grooved bar 2000 2000 130 Combined cone damage, u = 820–990 mm
1080 Plain bar 1400 1400 / Steel anchor pullout, bar/adhesive interface
damage, u = 250–480 mm
1080 Plain bar 900 900 / Annulus crack, u = 400–600 mm
1080 Plain bar 1300 1400 / Annulus crack, u = 600–700 mm
1080 Grooved bar 2000 2000 75 Combined cone damage, u = 420–490 mm
1080 Grooved bar 2200 2250 100 Combined cone damage, u = 1100–1200 mm
1080 Grooved bar 2000 2000 / Annulus crack, u = 1200 mm
150 Flowing 210 1200 Plain bar 2200 2400 / Annulus crack, u = 1240–1700 mm
grout 1200 Plain bar 2400 2400 250 Annulus crack, u = 1300–2300 mm
1200 Plain bar 2220 2220 / Annulus crack, u = 1800–2000 mm
1200 Grooved bar 4800 4800 450 Combined cone damage, u = 2700–3400 mm
1200 Grooved bar 2800 5200 450 Combined cone damage, u = 1400–2100 mm
1200 Grooved bar 4400 5800 350 Combined cone damage, u = 1200–1900 mm
1800 Plain bar 1600 2720 / Annulus crack, u = 80–160 mm, Grouting
segregation
1800 Plain bar 2800 3200 300 Combined cone damage, u = 1250–2350 mm,
adhesive/concrete damage
1800 Plain bar 1600 3000 / Annulus crack, u = 1240–1550 mm, steel anchor
is pulled out
1800 Grooved bar 2400 6400 / Annulus crack, not conical, height = 200 mm
1800 Grooved bar 6000 7200 260 Combined cone damage
1800 Grooved bar 5600 6400 220 Combined cone damage, u = 2000 mm

Note: (1) The ‘‘—’’ symbol indicates that these data were not recorded. (2) The box ‘‘ ’’ means that these data are considered anomalous.

(i.e., different surface treatments, bar diameters, and embedment
depths) can be seen clearly from these curves. In each diagram,
i.e., Fig. 7d shows that before Pr = 300 kN, compared a grooved
bar with a plain bar, the displacement–pull-out force curves of
all the samples were almost identical, besides, a linear relationship
was observed between the increase in the pull-out force and the
displacement, and the steel bar remains in the elastic deformation
stage, the bond stress at the interface plays a critical role in deter-
mining the pull-out force. For 300 kN < Pr < 400 kN, in the case of
the plain bars, the elastic deformation stage was extended. Here,
the pull-out force consists mainly of a weak bonding force and a
strong frictional force. On the other hand, for the grooved bars, this
phase lasted till Pr was 500 kN. For Pr values greater 400 kN, for the
plain bars, accelerated fracturing occurred. However, the pull-out
force for the grooved bars continued to increase and eventually
became nearly twice that for the plain bars. In the case of the
grooved bar-based anchors, the grooves increase the contact area
between the bar and the anchoring agent. In addition, they also
128 D. Wang et al. / Construction and Building Materials 95 (2015) 124–132

embedded in an epoxy resin at a depth of 12d, for U of 48 mm,

the pull-out force was 685 kN. Further, for U of 90 mm, the
pull-out force increased by 195% to 2020 kN. Finally, for U of
150 mm, the pull-out force increased by 951% to 7200 kN.
As can be seen from the curves, the pull-out force of the
anchors, which depends on the embedment depth, improved to
some extent for all the bars, but the increase was not as sharp as
expected. However, with an increase in the diameter, the effect
of the embedment depth on the pull-out force increased too.
The aim of these tests was to investigate the mechanical behav-
ior of large-diameter anchor bars (U > 40 mm) embedded in a con-
crete foundation. Thus, the curves of the bars with U of 36 mm are
not shown in Fig. 7. Further, the data for two samples with U of
48 mm have not been included, owing problems with the data
acquisition devices.

Fig. 2. The experimental field. 4.2. Axial stress–embedment depth curve

The relationships between the axial stress of the anchor bars

Grooved bar
Plain bar
Fig. 3. Steel bars used.
produce a mechanical interlocking force when a pull-out force is
applied. Owing to this dual mechanism, this anchor system is more
stable.
The bar diameter also has a significant influence on the anchor
pull-out force; the pull-out force increased with the increase in the
bar diameter (see Fig. 7a–f). For example, for the grooved bars
and the embedment depth for different load ratings are shown in
Fig. 8. The axial stress decreased with the increase in the embed-
ment depth. However, the closer to the loading end, the faster
the stress increased. When the load was small and the interface
was not decoupled, the axial stress at the end of the embedded
bar was the highest, with its value being equal to the pull-out
force. Moreover, the axial force along the embedment depth exhib-
ited a negative exponent distribution, with the increase in loading,
the axial stress is increasing. Once the pull-out force had reached a
certain threshold, the end of the bar near the interface began to
decouple, resulting in an increase in the maximum axial stress.
For example, as can be seen from Fig. 8a, when Pr was 500 kN,
the maximum axial stress at the end of the bar increased signifi-
cantly. When Pr was less than 500 kN, the distribution of the axial
stress of the bar exhibited a negative exponential distribution.
Further, with an increase in the load applied, the
stress-distribution curve corresponding to the area near the end
of the anchor bar became concave; this was because decoupling
occurred at the interface at the front end. In this case, the
pull-out force consisted only of the frictional force and the
mechanical interlocking force at the decoupled interface. The same
phenomenon can be observed in Fig. 8b and c; the only difference
is in the Pr value. When the embedment depth is 10–12 times the
diameter, the axial stress decreased, it then plateaued for diame-
ters greater than 10 times the embedment depth.

Fig. 4. Test setup.

 D. Wang et al. / Construction and Building Materials 95 (2015) 124–132
Curing of concrete
foundation
Drilling of holes
Cleaning of holes
Grouting the bars
Grouting and curing
Fig. 5. Scheme of the process for installing the post-installed anchors.
Fig. 6. Labeling of the test samples.
The effect of the anchor system depends primarily on the trans-
mission of stress between the adhesive/concrete and the
steel/adhesive interfaces. Therefore, the distribution of the axial
stress and the type of anchor system used are the main factors that
determine the anchoring effect.
4.3. Failure modes of anchors
As mentioned in Section 2, five basic failure modes are observed
in adhesive anchors in concrete. In the tests performed in the pre-
sent study, however, three different failure modes were observed.
These were steel anchor pullout, concrete annulus damage, and
combined cone damage. In the cases the steel anchor/substrate
system did not fail, damage to the surface concrete material was
observed. That is to say, either the spalling of the surface concrete
material occurred around the anchor; or the partial pulling out of
the steel anchor from the concrete foundation was observed,
accompanied by the formation of a conical wedge of the anchoring
agent, which came from the foundation; or the cracking of the
129
1. Vibrating the concrete to prevent segregation
2. Curing temperature different between 25°C
3. Curing time of 28 days
4. Standard compressive strength of more than 23 MPa
1. Drilled with a JX-1 geological drill
2. Holes of different diameters for different bars
1. Flushing the holes with clean water
2. Brushing the hole surfaces to remove dust
3. Blowing the holes with compressed air
1. Measuring diameter of holes with electronic station
(values listed in Table 4)
2. Grouting the steel bar
1. Grouting the anchoring agent
2. Curing the anchoring agent at 25°C for 28 days
concrete foundation was observed around the steel anchor. These
failure modes are different from those described in Section 2.
Fig. 9 shows the failure modes observed during the experimental
tests, while the characteristics of these failure modes are listed in
Table 5.
The steel anchor pullout mode was mainly observed for U of
36 mm, as shown in Fig. 9a. In this case, the steel bar was pulled
out in a straight manner, and the concrete foundation did not fail.
Further, spalling occurred only on the surface concrete material
around the anchor. In addition, the interface between the bar and
the anchoring agent failed, because the flowing grout that was
used as the anchoring agent had physical characteristics similar
to those of the concrete foundation. This was also because the
holes for the anchor bars had been cleaned, in order to ensure that
the bonding between the anchoring agent and concrete foundation
was strong.
The concrete annulus damage mode is a type of combined dam-
age mode. In such modes, the concrete foundation cracks, a con-
crete cone is formed in the concrete foundation, and the steel
anchor rod is pulled out from the substrate. The concrete annulus
damage mode results in the formation of a shallow concrete cone
on the surface of the concrete foundation and this cone is similar
to the one observed in the case of other combined damage modes.
Further, the cracking diameter of the concrete surface is larger and
the cone height is smaller than those for the other combined dam-
age modes. Finally, with an increase in the diameter of the steel
bar, the failure mode gradually changed from the steel anchor pull-
out mode to the combined cone damage and concrete annulus
damage modes.
Thus, for small-diameter bars (U of 36 mm), the plain-bar
anchors exhibited the steel anchor pullout mode, while the
grooved-bar anchors exhibited the concrete annulus damage
mode. With an increase in the bar diameter, the failure mode of
the plain-bar anchors changed gradually to the concrete annulus
damage mode; in contrast, the combined cone damage mode was
observed in the case of the grooved-bar anchors. The reason for
is that the large-diameter bars had a larger area in contact with
the anchoring agent. This made the pull-out force concentrate near
the free face and increased the scope of impact. Furthermore, the
130 D. Wang et al. / Construction and Building Materials 95 (2015) 124–132

Fig. 7. Displacement–pull-out force curves for the bars with different diameters, as determined from the flowing grout tests.

Fig. 8. Change in the axial stress with the embedment depth for the different loading ratings.

grooves resulted in a mechanical interlocking force. Thus, these
two coupling effects resulted in the three failure modes observed.
5. Conclusions
In this study, pull-out tests were performed on post-installed
large-diameter anchor systems embedded in concrete, in order to
elucidate their performance under static loading conditions.
The pull-out force increased with an increase in the bar diame-
ter; this was owing to the increase in the contact area between the
bar and the anchoring agent and that between the anchoring agent
and the concrete. This, in turn, increased the frictional resistance,
as expected, because the cementing force is the primary factor
determining the resistance.
Grooved bars had a greater anchoring effect than did plain ones.
The pull-out force for the grooved bars was larger and nearly twice
that for the plain bars. In case of the grooved-bar anchors, the
grooves increase the contact area between the bar and the anchor-
ing agent. In addition, they also produce a mechanical interlocking
force when the pull-out force is applied. This dual mechanism is
what makes this anchoring system more secure.
The pull-out force of the anchors, which is related to the
embedment depth, improved to some extent; however, the
improvement was not as marked as expected. Further, the axial
stress of the anchor bars decreased with the increase in the embed-
ment depth. The closer to the loading end, the faster the stress
increased. The axial force along the depth direction exhibited a
negative exponent distribution under low loads. Once the
pull-out force reached a certain value, the bar end near the inter-
face began to decouple, and the maximum axial stress increased.
Three main failure modes, namely, steel anchor pullout, concrete
annulus damage, and combined cone damage, were observed. In
the case of the post-installed large-diameter anchors, the primary
failure modes were concrete annulus damage and combined cone
damage. With an increase in the diameter of the steel bars, the fail-
ure mode gradually changed from steel anchor pullout to com-
bined cone damage and concrete annulus damage. Further, the
surface type (plain or grooved) of the steel bars had a significant
influence on their failure mode.
These results suggested that grooved bars are better suited for
use in post-installed anchor systems, as they allow for the proper-
ties of the anchoring agent and the concrete foundation material to
be exploited fully.
In the future, we plan to study the differences in the perfor-
mances of post-installed large-diameter anchor systems that use
organic and inorganic anchoring agents. Moreover, we plan to
use numerical simulations as predictive techniques for
 D. Wang et al. / Construction and Building Materials 95 (2015) 124–132 131

Fig. 9. Failure modes observed during the anchor tests.

determining the ultimate pull-out force and steel bar failure assisted in the experimental test. C.J.O.Y. helped draft and revise
modes. In addition, the ideal characteristics of post-installed the manuscript. All authors read and approved the final
anchor structures will be determined through reliability analyses. manuscript.

Competing interests section

The authors declare that they have no competing interest.
Authors’ contributions section
D.P.W. carried out the experiment, and D.S.W. helped in the
drafting and structuring of the manuscript. S.M.H. was our mentor
and conceived the idea of the study. J.Z. provided a test site and
Acknowledgements
This research has received financial support from the 973
Program of China (Grant No. 2013CB733201), Key Projects in the
National Science & Technology Pillar Program of China
(2014BAL05B01) and the Hundred Young Talents Program of the
Institute of Mountain Hazards and Environment (Grant No.
SDSQB-2013-01). This support is greatly appreciated.
132 D. Wang et al. / Construction and Building Materials 95 (2015) 124–132

References [10] A. Braimah, R. Guilbeault, E. Contestabile, Strain rate behaviour of adhesive

anchors in masonry, Eng. Struct. 67 (2014) 96–108, http://dx.doi.org/10.1016/
j.engstruct.2014.02.018.
[1] J.S. Kim, W.Y. Jung, M.H. Kwon, B.S. Ju, Performance evaluation of the post-
[11] S. Yilmaz, M.A. Özen, Y. Yardim, Tensile behavior of post-installed chemical
installed anchor for sign structure in South Korea, Constr. Build. Mater. 44
anchors embedded to low strength concrete, Constr. Build. Mater. 47 (2013)
(2013) 496–506, http://dx.doi.org/10.1016/j.conbuildmat.2013.03.015.
861–866, http://dx.doi.org/10.1016/j.conbuildmat.2013.05.032.
[2] L. Contrafatto, R. Cosenza, Behaviour of post-installed adhesive anchors in
[12] Ö. Çalısßkan, S. Yılmaz, H. Kaplan, N. Kıraç, Shear strength of epoxy anchors
natural stone, Constr. Build. Mater. 68 (2014) 355–369, http://dx.doi.org/
embedded into low strength concrete, Constr. Build. Mater. 38 (2013) 723–
10.1016/j.conbuildmat.2014.05.099.
730, http://dx.doi.org/10.1016/j.conbuildmat.2012.09.020.
[3] R.A. Cook, J. Kunz, W. Fuchs, R.C. Konz, Behavior and design of single adhesive
[13] G. Kwon, M.D. Engelhardt, R.E. Klingner, Behavior of post-installed shear
anchors under tensile load in uncracked concrete, ACI Struct. J. 95 (1) (1998) 9–
connectors under static and fatigue loading, J. Constr. Steel Res. 66 (4) (2010)
26, http://dx.doi.org/10.14359/522.
532–541, http://dx.doi.org/10.1016/j.jcsr.2009.09.012.
[4] R.A. Cook, R.C. Konz, Factors influencing bond strength of adhesive anchors, ACI
[14] M.S. Hoehler, P. Mahrenholtz, R. Eligehausen, Behavior of anchors in concrete
Struct. J. 98 (1) (2001) 76–86, http://dx.doi.org/10.14359/10149.
at seismic-relevant loading rates, ACI Struct. J. 108 (2) (2011) 238–247, http://
[5] R. Eligehausen, J. Silva, The Assessment and Design of Adhesive Anchors in
dx.doi.org/10.14359/51664259.
Concrete for Sustained Loading, Report of Hilti, Inc., 2008.
[15] A.H. Nilson, Internal measurement of bond slip, ACI J. 69 (7) (1972) 439–441,
[6] ACI318. Building Code Requirements for Reinforced Concrete, American
http://dx.doi.org/10.14359/7170.
concrete institute, Detroit, USA, 2005.
[16] R.A. Cook, Behavior of chemically bonded anchors, J. Struct. Eng. 119 (9) (2001)
[7] A. Shah, Q. Ali, B. Alam, K. Shahzada, R. Khan, N. Ahmad, Study on Performance
2744–2762, http://dx.doi.org/10.1061/(ASCE)0733-9445(1993)119:9(2744).
Evaluation of Adhesive Anchors in Concrete, Int. J. Adv. Struct. Geotech. Eng. 01
[17] R.A. Cook, G.T. Doerr, R.E. Klingner, Bond stress model for design of adhesive
(2) (2012) 74–78.
anchors, ACI Struct. J. 90 (5) (1993) 514–524, http://dx.doi.org/10.14359/3945.
[8] M. Obata, M. Inoue, Y. Goto, The failure mechanism and the pull-out strength
[18] M. McVay, R. Cook, K. Krishnamurthy, Pullout simulation of post installed
of a bond-type anchor near a free edge, Mech. Mater. 28 (1–4) (1998) 113–122,
chemically bonded anchors, J. Struct. Eng. 122 (9) (1996) 1016–1024, http://
http://dx.doi.org/10.1016/S0167-6636(97)00052-5.
dx.doi.org/10.1061/(ASCE)0733-9445(1996)122:9(1016).
[9] L. Contrafatto, R. Cosenza, Prediction of the pull-out strength of chemical
anchors in natural stone, Frattura Integrità Strutturale 29 (2014) 196–208,
http://dx.doi.org/10.3221/IGF-ESIS.29.17.