Engineering Structures 164 (2018) 109–118

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Design of anchor reinforcement for seismic tension loads T

a b c,⁎
Derek Petersen , Zhibin Lin , Jian Zhao
a
Osmose Railroad Services Inc., Madison, WI 57716, USA
b
Department of Civil and Environmental Engineering, North Dakota State University, Fargo, ND 58108, USA
c
Department of Civil and Environmental Engineering, University of Wisconsin, Milwaukee, WI 53201, USA

A R T I C L E I N F O A B S T R A C T

Keywords: This paper presents a study of cast-in anchors in reinforced concrete subjected to both monotonic and cyclic
Cast-in anchors tension. Laboratory tests using 25 mm [1 in.] diameter anchor bolts showed that the code-conforming anchor
Headed studs reinforcement may not ensure anchor steel failure though it effectively prevented concrete breakout. Based on
Anchor connections the observations and other tests in the literature, recommendations for the design of anchor tension reinforce-
Anchor reinforcement
ment were proposed, which consists of (1) load-carrying reinforcement in the direction of the anchors; (2) crack-
Fastening to concrete
controlling reinforcement in all directions that have a limited edge distance; and (3) local confining re-
Composite construction
And seismic design inforcement near the anchor head if side-face blowout may control the failure. The proposed anchor re-
inforcement was proven effective using additional laboratory tests. Further studies are needed to verify the
reinforcement design for anchor connections.

1. Introduction
Anchor connections are a critical component of load transfer be-
tween steel and concrete members affecting structural performance
during earthquake events [1,2]. Cast-in anchors may experience steel
fracture and/or concrete breakout failure when subjected to a tension
force [3–5]. The failure modes are mainly dependent upon the em-
bedment depth (hef) when the anchor bolt is placed in plain concrete.
Breakout failure is brittle and thus not a preferred failure mode for the
anchors in seismic zones [6]. Engineers are recommended to consider
steel reinforcement by building codes [7] and design guides [8,9] to
improve the behavior of anchors likely controlled by concrete breakout.
It has been well documented that the potential failure region near
headed stud anchors be confined using steel reinforcement [10]. The
existing reinforcement for concrete anchors in the technical literature is
summarized in Fig. 1. Two types of reinforcement are allowed in the
building codes [7] and design guides [8]: specifically designed re-
inforcement (known as anchor reinforcement) for cast-in anchors and
the existing reinforcement near the anchors in a reinforced concrete
member (known as supplementary reinforcement). With a goal of re-
straining potential concrete breakout cracks, supplementary reinforce-
ment is not designed to transfer the full tensile capacity of anchors into
concrete. The effect of supplementary reinforcement has been well
studied [11,12], and the corresponding analysis methods have been
proposed. For example, the existing design methods assume that the
concrete breakout, shown in shaded areas in Fig. 1a, c, and d, occurs
before the reinforcement takes effect. With the cracks, the tensile ca-
pacity of the anchors is provided by the reinforcement in the same di-
rection the anchors, and the tensile capacity of the reinforcement can be
affected by available development lengths [10,11].
Anchor reinforcement in this study is that specifically designed and
detailed such that the reinforced anchors achieve their full design ca-
pacities. U-shaped hairpins are specified in the current building codes
and design guides. In addition, reinforcing bars placed near anchor
bolts have been expected to create loading transferring paths through a
splice mechanism (Fig. 1c, e, and f), and explained using strut-and-tie
models (STMs). Few experimental tests have been available to verify
the effectiveness of these anchor reinforcement, especially under si-
mulated seismic loading.
2. Literature review
Very limited studies in the literature have focused on the behavior
of anchor reinforcement and reinforced anchors. However, the litera-
ture review below includes a variety of related studies because concrete
anchors are widely used in many applications in construction.
Specifically, conceptual designs in the literature and the studies, in
which headed anchors were embedded in reinforced concrete members,
in which some reinforcing bars may have improved the anchor beha-
vior/capacities. Observations from these tests are used to complement
those made in this study, based which, an improved design method is
proposed.

⁎
Corresponding author.
E-mail address: jzhao@uwm.edu (J. Zhao).

https://doi.org/10.1016/j.engstruct.2018.02.015
Received 11 October 2016; Received in revised form 23 December 2017; Accepted 1 February 2018
0141-0296/ © 2018 Elsevier Ltd. All rights reserved.
D. Petersen et al.
Nomenclature
ca1 front edge distance of anchor
da anchor diameter
db reinforcement diameter
ds stirrup diameter
fc measured concrete compressive strength
fc′ specified concrete compressive strength (27.5 MP
[4000 psi] in the study)
fy specified yield strength of steel (414 MPa [60 ksi] in the
study)
hef embedment depth of anchor bolts
Fig. 1. Schematics of existing anchor tension reinforcement.
2.1. Existing design recommendations
The recommendations on the design of anchor tension reinforce-
ment are summarized in Fig. 1. Cannon et al. [13] suggested simple
hairpins for preventing potential breakout cones as shown in Fig. 1a. To
achieve direct force transfer (similar to rebar splicing), hairpins are
required to be placed symmetrically within 1/3hef , measured from the
edge of the anchor head, where hef is the embedded depth. A similar
recommendation was adopted by CEB (1997) and ACI (2008) with a
larger effective range (0.5hef ), measured from the center of anchors, as
shown in Fig. 1b. The European committee for standardization (2009)
further increased this effective range to 0.75hef . Critical to this study,
stirrups in beams have been allowed by fib (2008) to transfer tensile
forces from anchors installed at the bottom of the beam to the com-
pression zone of the beam, as illustrated in Fig. 1c.
Cannon et al. [13] also recommended spiral reinforcement for each
anchor in tension to prevent “lateral bursting,” now known as side face
blowout failure. In an application of U-shaped hairpins to the connec-
tions between steel columns and their concrete foundations, shown in
Fig. 1d, Shipp and Haninger [14] recommended closed ties at a small
spacing (e.g. 150 mm [6 in.]) to encase the anchor bolts. This critical
measure may have been recommended after the study by Hasselwander
et al. [15], in which many specimens, without confining reinforcement,
failed in concrete side-face blowout.
Strut-and-tie models (STMs) have been used in the analysis and
design of reinforcement for concrete anchors. For example, STMs are
used to demonstrate the effectiveness of hairpins, and the needs for
crack-controlling reinforcement [4,5,7,8]. Widianto and Patel [16] ex-
plained the load transfer from headed anchors embedded in a column
110
Engineering Structures 164 (2018) 109–118
foundation to the column reinforcement, as illustrated in Fig. 1e. The
same model was also used to determine the lateral bursting force that
leads to side face blowout failure. In the tests of column base connec-
tions in foundation beams, Baba et al. [17] treated closed stirrups near
the headed anchors as anchor reinforcement, and applied STMs in ex-
plaining the load transfer. More importantly, a separate strut-and-tie
model was used to visualize the load transfer from the bottom of the
closed stirrups to the rest of the structure: column base plate in com-
pression, as illustrated in Fig. 1f.
2.2. Existing experimental studies
There are limited tests available in the literature focusing on the
verification of the above recommended anchor reinforcement. Focusing
on the concrete breakout capacity of large diameter anchors, the tests
by Lee et al. [18] included two test groups with four or eight No. 8 U-
shaped hairpins, located within 0.2hef and the other 0.35hef from the
test anchors. The No. 8 hairpins were assumed fully developed below
the anchor head through a length of 14db with hooked ends. Note that
these specimens were referred as “reinforced specimens,” which leads
to the term “reinforced anchors” in this paper. The hairpins can be
treated as supplementary reinforcement because they were not pro-
portioned to carry the anchor steel capacity in tension. The tests in-
dicate that hairpins close to the test anchor are more effective than
those located farther. The authors also indicated that the reinforcement
should be proportioned to carry 100% of the applied load, and the
design strength for the reinforcement should consider bond failure of
reinforcement in the breakout cone. Berger [12] further studied the
effect of supplementary reinforcement, and proposed a new capacity
model, which considers contribution from both concrete and steel re-
inforcement. The tensile capacity of the reinforcement was calculated
based on its bond capacity. Sharma et al. [11,19] extended the in-
vestigation to groups of four anchors. Instead of hairpins, closed stir-
rups were used as supplementary reinforcement in this latest study.
There are tests in the literature with headed studs placed in re-
inforced concrete members. The existing reinforcement in the members
was not specifically designed as anchor reinforcement; however, ob-
servations from these tests provide relevant and valuable information to
the subject of this study. For example, closed stirrups were provided in
one of the two tension tests of headed studs for use in composite con-
struction by Saari et al. [20]. The No. 3 stirrups at a spacing of 89 mm
[3.5 in.] and longitudinal bars at all corners formed a reinforcing cage
D. Petersen et al.
around the studs. Each stud thus had two No. 3 stirrups (four legs as
anchor reinforcement), which were able to resist a tension force higher
than the stud capacity in tension. Concrete failure occurred to the un-
reinforced anchor while steel fracture occurred to the reinforced anchor
with capacity increase about 100 percent. Similar observation was also
made by Baran et al. [21] in their tests, in which reinforcing cages,
consisting of No. 4 stirrups at 150 mm [6 in.] and corner bars at 75 mm
[3 in.], were provided to the studs. Compared with unreinforced spe-
cimens, reinforced studs achieved higher capacities though their failure
was not controlled by stud fracture.
In addition, Shahrooz et al. [22] conducted six 1/3-scale tests of
shear connections between concrete walls and steel outrigger beams for
use in hybrid structures. The 4-stud and 6-stud connections with 13-mm
[0.5 in.] diameter headed studs were installed in concrete representing
the boundary elements of a concrete wall. Specimen No. 4 was used to
simulate the boundary elements at lower levels, in which D2.9 stirrups
were provided with a spacing of 38 mm [1.5 in.] near the connection.
The shear tab in Specimen No. 4 yielded while concrete breakout was
observed in all other fours tests. The benefit of closely spaced stirrups
was also observed in later tests of two ¼-scale girder-wall connections
in concrete experiencing damage resulted from simulated seismic ac-
tions [23].
This study focuses on specially designed and detailed reinforcement
for anchors to achieve their full capacities corresponding to steel frac-
ture in tension. Tests of anchors with code-complying anchor re-
inforcement conducted to evaluate the easing design methods. An im-
proved reinforcement design, based on observations in this study and
those in the literature, was proposed and experimentally verified to
ensure anchor steel failure in addition to restraining concrete breakout
cracks.
3. Experimental investigation
3.1. Group 1 specimen design
A total of 24 single anchors placed in six concrete blocks, as illu-
strated in Fig. 2, were first tested, referred as Group 1 tests hereafter, to
evaluate the code-conforming anchor reinforcement. The cast-in an-
chors consisted of a 25 mm [1 in.] diameter ASTM A193 Grade B7 rod
with a heavy hex nut. Plate washer was not used in this group of spe-
cimens because pullout failure was initially not expected to control the
behavior. All anchors were embedded 152 mm [6 in.] in concrete while
two edge distances were used: 152 mm [6 in.] and 254 mm [10 in.].
Although such headed anchors usually have a larger embedment depth
in practice, this small embedment was selected in order to examine the
effectiveness of anchor reinforcement in an adverse condition. The
block length and anchor spacing were determined assuming a concrete
Fig. 2. Specimens for Group 1 tests (1 in. = 25.4 mm).
111
Engineering Structures 164 (2018) 109–118
breakout cone with a radius of two times the embedment depth (2hef)
from each anchor plus an additional 100 mm [4 in.] on both sides. The
block height was 432 mm [17 in.] similar to all other anchor tests in this
study, which was limited by the position of the vertical actuator [6].
Five types of reinforcement patterns were tested first as shown in
Fig. 2, and listed in Table 1. From the measured strength of the
threaded rods shown later, the ultimate tensile capacity of the anchor
bolts was 358 kN [80.4 kips]. The anchor reinforcement using Grade 60
steel was proportioned for this load assuming a nominal yield strength
of 414 MPa [60 ksi]. The required anchor reinforcement was found to
be 865 mm2 [1.34 in.2], which was provided using four No. 6 bars
(Types B and D), six No. 4 and two No. 3 bars (Types C and E), or eight
No. 4 bars (Type F) as shown in Fig. 2.
The anchor reinforcement is assumed to develop strains beyond the
proportional limit; therefore, the bars need to be fully developed at
both sides of the assumed breakout crack. All stirrups in Group 1 spe-
cimens were 279 mm [11 in.] high measured from the outside edges.
Concrete cover of 38 mm [1.5 in.] was used, thus stirrups were em-
bedded about 102 mm [4 in.] above the anchor head. This short de-
velopment length was deemed sufficient to develop the vertical legs of
the closed stirrups through the interaction with the corner bars [24].
Similarly, the development length below the anchor head was 152 mm
[6 in.].
3.2. Materials
The measured stress-strain relationship for a coupon made from
ASTM A193 Grade B7 rods is shown in Fig. 3. The sudden slope change
near 655 MPa [95 ksi] may have been due to a slip of the coupon out of
the grips of the loading frame. No additional tests were conducted be-
cause the new specimens had to be made from the materials of a dif-
ferent heat. The yield strength was then estimated as 724 MPa [105 ksi]
corresponding to a 0.2 percent residual strain. The measured ultimate
strength was 914 MPa [132.6 ksi]. The measured strengths are close to
the typical values reported by the manufacturer.
Ready-mixed concrete with a targeted strength of 27.6 MPa
[4000 psi] was used. Tests at 28 days showed an average strength of
26.2 MPa [3800 psi] using three 100 × 200 mm [4 × 8 in.] cylinders.
Additional cylinder tests during the anchor tests (roughly 58 days)
showed an average compressive strength of 37.2 MPa [5400 psi]. Note
that this strength is higher than the design assumption; hence the ad-
verse impact of the under-designed anchor head was greatly reduced.
3.3. Test setup
The loading frame, actuator placement, and instrumentation setup
used for Group 1 tests are illustrated in Fig. 4. Similar to most previous
D. Petersen et al. Engineering Structures 164 (2018) 109–118

Table 1
Summary of reinforced anchor tests in tension.

Specimen ID Reinf. type Edge distance ca1 Load type Concrete strength fc Breakout capacity Peak load Peak disp. Bearing stress Specimen name in Ref.
(in.) (psi) (kips) (kips) (in.) (×fc) [25]

Group 1 specimens
UR10-m – 10 M 5400 32.40 36.56 0.05 – 5132011-A1
UR10-c – 10 C 5400 32.40 35.22 0.10 – 5162011-A2
UR06-m – 6 M 5400 24.30 30.17 0.11 – 5162011-A3
UR06-c – 6 C 5400 24.30 33.66 0.12 – 5162011-A4
RB10-m B (Fig. 2) 10 M 5400 32.40 49.65 0.12 6.13 5172011-B1
RB10-c1 B 10 C 5400 32.40 43.58 0.08 5.38 5172011-B2
RB10-c2 B 10 C 5400 32.40 44.53 0.14 5.49 5202011-C1
RB10-c3 B 10 C 5400 32.40 50.11 0.20 6.18 5202011-C2
RC06-m C (Fig. 2) 6 M 5400 24.30 63.00 0.34 7.77 5182011-B3
RC06-c1 C 6 C 5400 24.30 62.40 0.24 7.70 5182011-B4
RC06-c2 C 6 C 5400 24.30 67.31 0.74 8.30 5232011-C3
RC06-c3 C 6 C 5400 24.30 69.99 0.50 8.63 5232011-C4
RD10-m D (Fig. 2) 10 M 5400 32.40 49.07 0.19 6.05 5252011-D1
RD10-c1 D 10 C 5400 32.40 50.96 0.14 6.29 5252011-D2
RD10-c2 D 10 C 5400 32.40 52.63 0.18 6.49 5272011-E1
RD10-c3 D 10 C 5400 32.40 49.69 0.11 6.13 5272011-E2
RE06-m1 E (Fig. 2) 6 M 5400 24.30 48.24 0.09 5.95 4252011-D4
RE06-m2 E 6 M 5400 24.30 52.28 0.08 6.45 5252011-D3
RE06-c1 E 6 C 5400 24.30 53.01 0.18 6.54 5262011-E3
RE06-c2 E 6 C 5400 24.30 54.90 0.12 6.77 5252011-E4
RF06-m F (Fig. 2) 6 M 5400 24.30 57.46 0.21 7.09 5282011-F1
RF06-c1 F 6 M 5400 24.30 56.38 0.50 6.96 5282011-F4
RF06-c2 F 6 C 5400 24.30 53.65 0.18 6.62 5282011-F2
RF06-c3 F 6 C 5400 24.30 55.05 0.27 6.79 5282011-F3

Group 2 specimens
RG10-m G (Fig. 7) 10 M 6900 36.62 79.91 0.41 3.62 5122012-G1
RG10-c1 G 10 C* 6900 36.62 80.85 NA 3.66 5152012-G2
RG10-c2 G 10 C* 6900 36.62 80.86 0.45 3.66 5152012-G3
RG10-c3 G 10 C* 6900 36.62 80.54 0.49 3.65 5162012-G4

Note: M: Monotonic loading; C: Cyclic loading in displacement control; C*: cyclic loading in load control; 1 kip = 4.45 kN; 1 in. = 25.4 mm; 1 psi = 6.9 kPa.

[0.08 in./min] in the monotonic tests. The cyclic displacement steps for
each 3-cycle group were chosen as 2, 4 (failure for typical unreinforced
anchors), 8, 16, and 32 mm [0.08, 0.16, 0.32, 0.64, and 1.28 in.]. The
loading rate for displacement cycles at or below 8 mm [0.32 in.] were
kept at 2 mm/min [0.08 in./min], same as that in the monotonic tests,
while the load rate was increased to 10 mm/min [0.4 in./min] for the
last two cycle groups cycles in order to reduce test time. The impact of
the higher loading rate was found negligible [23].

4. Discussion of group one tests

The monotonic tests of reinforced anchors are compared with the

Fig. 3. Stress vs. strain relationship of anchor steel (ASTM A193 Grade B7).
anchor tests in the literature, a self-balanced load frame was used.
Different from other reaction systems, the reaction beams in this test
setup had small lateral stiffness, thus providing negligible lateral re-
straint for crack development in concrete during a test. The reactions
were provided 356 mm [14 in.] (2.3hef) from the test anchor to mini-
mize the impact. The tensile force in the test anchor thus would transfer
to the reinforcement, before reaching the reaction points. The load
transfer from the closed stirrups to the reaction points may be visua-
lized using a strut-and-tie model, similar to the tests by Baba et al. [9].
An MTS Model 244.41, 489-kN [110-kip] actuator was used to
apply tension loading to the anchor bolt through a loading plate. One
monotonic test and three cyclic tests were conducted for each type of
reinforced anchor specimens, as shown in Table 1, expect for the group
of four anchors with Type E reinforcement. The cracks formed during
the test of RE06-m1 extended to the region near RE06-m2, which could
impact the cyclic behavior and provide false comparison. Hence RE06-
m2 was tested with monotonic loading.
The displacement-controlled loading rates were kept at 2 mm/min
tests of unreinforced anchors (shown in dark solid lines) in Fig. 5. In
general, the capacities of reinforced anchors are 34 to 100 percent
higher than that of unreinforced anchors though the desired steel
fracture was not achieved. The design of anchor reinforcement in this
test group focused on the number of reinforcing bars and their positions
as required by ACI 318-08 Appendix D. However, concrete developed
cracks in multiple directions as shown in Fig. 6. Following the cracking,
the concrete above the anchor heads lost its confinement, and crushed,
resulting in pullout failure in most tests. The maximum concrete
bearing stress, calculated by the observed ultimate loads divided by the
bearing area ( Abrg ) and the measured concrete strength as listed in
Table 1, was as low as 5.4fc . This indicates that anchors with code-
conforming head sizes, which usually are proportioned assuming a
concrete bearing strength of 8.0fc′, might still have experienced the
pullout failure.
Anchor bolts in plain concrete had a brittle failure after the concrete
breakout as shown in Fig. 5. The measured peak load was 163 kN
[36.6 kips] for the anchor with an edge distance of 254 mm [10 in.].
This is higher than the code predicted capacity (144 kN [32.4 kips])
calculated with the actual concrete strength, but close to the average
breakout capacity (161 kN [36.3 kips]) from the tests of headed

112
D. Petersen et al.
a)
Fig. 4. Experimental test setup for Group 1 Specimens. (a) Schematic; (b) picture.
(a)
b)
(b)
Fig. 5. Monotonic tension test results of reinforced anchors. (a) ca1 = 254 mm [10 in.];
(b) ca1 = 152 mm [6 in.].
anchors/studs in the literature [6]. The code prediction becomes
108 kN [24.3 kips] for the unreinforced anchor with an edge distance of
152 mm [6 in.], which is a bit lower than the measured capacity 134 kN
[30.2 kips], as shown in Fig. 5. Instead of forming an idealized breakout
cone, the concrete developed radial splitting cracks, and the cracks
propagated below the reaction plates, as shown in Fig. 6a.
Anchor bolts reinforced with two No. 6 stirrups (Specimen RB10-m)
had brittle failure as shown in dotted lines in Fig. 5a. Before the sudden
capacity drop near 4 mm [0.16 in.], the test anchor experienced a 2-mm
[0.08-in.] slip. This is similar to the observations by Shahrooz et al. [22]
in their tests of anchor groups in concrete walls without closely spaced
stirrups. The two narrow stirrups did not provide effective restraints
113
Engineering Structures 164 (2018) 109–118
b)
such that the front concrete was pushed out right after a splitting crack
was observed at about 169 kN [38 kips]. In addition, the corner bars
were actually 76 mm [3 in.] below the top surface because of the large
bending radius of No. 6 bars. These factors contributed to the poor
confinement for the concrete above the anchor head. As a result, a small
breakout cone (76 mm [3 in.] deep) formed mostly above the corner
bars as shown in Fig. 6b, leading to the sudden capacity drop shown in
Fig. 6a.
The corner bars were better distributed in Specimen RD10-m with
Type D anchor reinforcement, in which the anchor bolt was reinforced
by four No. 6 bars from four stirrups. The two wider stirrups were next
to the anchor bolt as shown in Fig. 2, and extended towards the front
face of the concrete block. The distributed bars along the top surface
provided better restraints for concrete as shown in Fig. 6c. As a result,
the load capacity was maintained for a large displacement as shown in
dashed lines in Fig. 5a. On the other hand, the ultimate capacity
(controlled by pullout failure) was similar to that of Specimen RB10-m
because the left side concrete was pushed away after a crack formed
outside the No. 6 stirrup (Fig. 6c) at about 200 kN [45 kips], and the
confinement condition for the concrete near the anchor was thus partly
destructed.
The concrete was better confined in Specimen RC06-m with Type C
reinforcement. Four No. 4 stirrups were extended to the front face with
a cover of 38 mm [1.5 in.], and the corner bars were close to the anchor
along the top surface as shown in Fig. 2. Unlike the specimen with Type
D reinforcement, smaller size stirrups were distributed within a wider
range, providing better confinement for the concrete. The specimen
achieved a load of 280 kN (63 kips), above the yield capacity of the
anchor bolt, and the load capacity was maintained over a large dis-
placement (i.e., 15 mm [0.6 in.]) as shown in dotted lines in Fig. 5b. In
addition to better confinement, the four bars placed near the anchor
bolt may have provided better load transfer. Although pullout failure
controlled the test as shown in Fig. 6d, the calculated concrete bearing
strength was around 7.7fc. A larger anchor head may increase the
pullout capacity beyond that needed to rupture the anchor bolt, which
was the expected behavior of reinforced anchors.
Specimen RE06-m1 with Type E anchor reinforcement was similar
to Specimen RC06-m except that the legs of four narrow stirrups, as
anchor tension reinforcement, were placed farther to provide sym-
metric tensile resistance as suggested by existing design guidelines [7].
However, the behavior of Specimens RE06-m1 was surprisingly dif-
ferent from that of Specimen RC06-m as shown in Fig. 5b. Concrete
crack initiated above the anchor head and propagated in two different
angles: the crack towards the front face was flat similar to that observed
in the test of unreinforced anchors, while the crack towards the back
D. Petersen et al.
a) A1-unreinforced
c) D1-4#6’s
e) D3-6#4’s+2#3’s
Fig. 6. Typical failure of anchor bolts in Group 1 tests with various patterns of reinforcement.
bent upwards. As a result, the vertical legs of the four larger stirrups
bridged the crack and acted as anchor reinforcement, but the four legs
on the back was bypassed by the crack as shown in Fig. 6e. The crack on
the back caused a sudden capacity drop as shown in dashed lines in
Fig. 5b. The load capacity was not completely lost because the breakout
cone was held back by the net of reinforcing bars on the top surface and
restrained by the four vertical bars in the front.
The importance of closely spaced bars to mobilize farther bars in
transferring the tensile force from the anchor was demonstrated by
Specimen RF06-m. The anchor bolt was reinforced by eight No. 4 bars
from eight stirrups with a spacing of 51 mm [2 in.] as shown in Fig. 2. If
a strut-and-tie model is used to explain the load transfer, the strut be-
tween the anchor head and the node formed by the outmost stirrup and
the corner bar had an angle of 41° with respect to the horizontal axis,
indicating that all the bars were effective. Unfortunately, side-face
blowout occurred, and the concrete above the anchor head crushed at a
load of 222 kN [50 kips]. The corresponding bearing strength was found
to be 7.1fc . This A breakout crack was identified as shown by the marked
lines in Fig. 6f. The crack was almost flat within the range of anchor
reinforcement, and bent up towards the reaction plates on the top
surface. This observation indicates that closely spaced bars can change
the direction of crack propagation, which was also observed in the tests
of reinforced anchors in shear [24].
The observed pullout failure was also observed in the cyclic tests,
and the detailed behavioral data can be found elsewhere [25]. Com-
pared with the lack of proper confinement, the impact of under-
114
Engineering Structures 164 (2018) 109–118
b) B1-2#6 (2 legs)
d) B3-6#4’s+2#3’s
f) F1-8#4’s
designed anchor head was deemed insignificant when taking into
consideration the tests by Hasselwander et al. [15], which included a
series of tension tests for 44-mm [1.75-in.] high-strength anchor bolts.
The anchor heads were made of 102 mm [4 in.] diameter washers,
which was overdesigned corresponding to a maximum bearing stress of
6.3fc′. However, the anchors had an edge distances ranging from 25 mm
[1 in.] to 152 mm [6 in.], and no closed ties were provided near the
anchor heads. The failure of the specimens was controlled by side-face
blowout with the calculated maximum bearing stresses ranging from
2.7fc′ to 5.3fc′, and the edge distance was found critical to the anchor
capacities.
5. Proposed anchor tension reinforcement
The code-recommended anchor reinforcement [5,7] can be effective
for anchors placed far from any edges such that splitting cracks would
not weaken the confinement to the concrete near the anchor bolts. An
alternative design method for anchor tension reinforcement was pro-
posed based on the observations made both in Group 1 tests and those
in the literature. The proposed anchor tension reinforcement, as illu-
strated in Fig. 7, includes: (1) load-carrying reinforcement in terms of
closed stirrups; (2) crack-controlling reinforcement in all directions that
has a limited edge distance; and (3) local confining reinforcement near
the anchor head if side-face blowout may control the failure. In addi-
tion, the tensile load on the anchors eventually needs to be transferred
to the rest of the structure. Sufficient reinforcement, such as stirrups
D. Petersen et al.
Fig. 7. Proposed anchor tension reinforcement layout.
and surface bars, is needed to ensure proper load transfer.
The load-carrying reinforcement should be proportioned to carry a
force equal to the anchor steel capacity in tension. The nominal yield
strength of steel should be used in the calculation. The load-carrying
reinforcement should be implemented using small-diameter closed
stirrups. The stirrup width should be small yet satisfy the bending ra-
dius requirements. The depth of the stirrups should be large enough as
shown in Fig. 7 such that the vertical legs are fully developed for the
tension load at both sides of the anchor head. Two stirrups should be
placed next to the anchor shaft, where the crack in concrete may in-
itiate above the anchor head under a tensile load. Rather than placing
all bars within a small distance from the anchor (e.g. 0.5hef) [5,7],
closely spaced stirrups with a center-on-center spacing of 51 mm [2 in.]
to 76 mm [3 in.] can extend further from the anchor bolt (e.g., 1.0hef).
Although the development length requirements for the anchor re-
inforcement can be satisfied through the interaction between the closed
stirrups and longitudinal bars at all four corners [25], it is re-
commended that the closed stirrups are at least 8.0ds (where ds is the
diameter of the stirrups) from the anchor head at both sides [19], as
shown in Fig. 7.
The crack-controlling reinforcement is needed to cross all possible
splitting cracks. The load transfer from the anchor head to the load-
carrying reinforcement can be represented by a strut-and-tie model
Fig. 8. Specimen design for Group 2 tests (1 in. = 25.4 mm).
115
Engineering Structures 164 (2018) 109–118
with the struts ranging from the anchor head to the outmost load-car-
rying stirrups. The crack-controlling reinforcement can thus be de-
termined considering the load transfer from the load-carrying re-
inforcement to the rest of the structure, as shown in Fig. 7. All crack-
controlling reinforcement should be proportioned using a steel stress of
0.6f y , and implemented using small diameter bars evenly distributed
with a small and practical spacing in two orthogonal directions [7]. The
crack-controlling reinforcement should also be properly developed.
In addition to determining the needed reinforcement, design using
STMs should also include capacity checks for struts and nodes. Such
capacity checks require the geometry of struts and nodes, which can be
difficult to define in the design of anchor reinforcement. For example,
all struts in the 3-D space start from the anchor head, which defines the
size of the nodal zone. If the code-specified effective compressive
strength (i.e., 0.85fc′ in ACI 318-08) is used in the capacity check of the
node, the design may result in a large head size. On the other hand,
such large head size is not necessary if anchor pull-out failure is to
control the design, in which a very different concrete strength (i.e., 8.0fc′
in ACI 318-08) is used. Concrete confinement is critical in this case to
ensure proper load transfer. In addition, distributed reinforcement is
required to restrain struts from splitting cracks in typical design using
STMs; however, such measures have been overlooked in the existing
anchor design recommendations, as illustrated in Fig. 1.
The closest four stirrups, as the load-carrying reinforcement, can be
encased by three or four closed hoops in the transverse direction near
the anchor head if the anchors are close to a free side surface. This is to
better confine the concrete that is subjected to large compressive
stresses above the head. These hoops may also help prevent the side
face blowout failure observed in this study and that by Hasselwander
et al. [15]. Further study is needed to quantify this part of the design
procedure.
6. Verification of proposed anchor tension reinforcement
6.1. Group 2 experiment design
Four tests were conducted to verify the proposed anchor tension
reinforcement. The specimen design is shown in Fig. 8. The cast-in
anchors were made of the same 1-in. diameter ASTM A193 Grade B7
rod as Group 1 specimens while the anchor head and the anchor re-
inforcement were redesigned. The head size was determined assuming
cracks existed in the concrete and the maximum concrete bearing
strength was assumed 7.0fc′, where a nominal design concrete strength
of 27.6 MPa [4 ksi] was used. The 1-in. diameter anchor could carry a
maximum tensile load of 358 kN [80.4 kips]; thus a net bearing area of
18.5 cm2 [2.9 in.2] was needed, and a 3/8-in. thick plate washer
(5.1 × 5.1 cm [2 × 2 in.]) was tack welded to the heavy hex nut, pro-
viding a net bearing area of 20.6 cm2 [3.2 in.2]. The anchors were again
embedded 152 mm [6 in.] into concrete as illustrated in Fig. 9. The
concrete beam width was 508 mm [20 in.] and the height was 432 mm
[17 in.]. The closest tie-down was 305 mm [12 in.] (2hef ) away from the
test anchor.
D. Petersen et al. Engineering Structures 164 (2018) 109–118

a)

b)

Fig. 9. Strut-and-Tie models for the specimen design for Group 2 tests.

The loading-carrying reinforcement was proportioned to carry a

load of 358 kN [80.4 kips]; thus the required anchor reinforcement
(Grade 60 steel) was found to be 8.6 cm2 [1.34 in.2], which was pro-
vided using eight No. 4 bars from four closed stirrups. Two stirrups
were placed by the anchor shaft and the other two were 51 mm [2 in.]
away. These four stirrups were 356 mm [14 in.] high measured from the Fig. 10. Test setup for Group 2 tests. (a) Schematics; (b) picture.
outside edges. The development length below the anchor head was
254 mm [10 in.] to facilitate the fabrication of the reinforcing cage. The
degree bend on both ends. The closed stirrups and side bars also formed
load transfer from the anchor to the loading-carrying reinforcement
a cage to prevent concrete side face blowout failure; hence, the re-
may cause splitting cracks in two directions, as shown by the strut-and-
commended local confining reinforcement, in terms of closed hoops
tie models in Fig. 9. The splitting force at the location of the horizontal
perpendicular to the anchor shaft near the anchor head shown in Fig. 7,
leg of the stirrups was roughly 60 percent of the tensile force in the
was not provided.
vertical legs of the stirrups; thus the provided reinforcement (4 No. 4’s)
Ready-mixed concrete with a targeted strength of 27.6 MPa
was sufficient. Similarly, the provided reinforcement (2 No. 5’s) were
[4000 psi] was used. Tests at 28 days showed an average strength of
deemed sufficient in the perpendicular direction. The concrete stress at
39.9 MPa [5800 psi] using three 100 × 200 mm [4 × 8 in.] cylinders.
the nodal zone above the anchor head was deemed sufficient because
Additional cylinder tests during the anchor tests (roughly 85 days)
the nodal zone was confined in multiple directions.
showed an average concrete strength of 47.6 MPa [6900 psi].
The rest crack-controlling reinforcement was proportioned using the
The concrete beam was fixed to the strong wall of the UWM
second strut-and-tie model shown in Fig. 9. The direction of the beam
Structures Laboratory at the two tie-down points, as illustrated in
was the main load transfer path. From the overall equilibrium of the
Fig. 10. Two MTS Model 244.31, 245-kN [55-kip] actuators were used
beam, which was loaded at an anchor and simply supported at two tie-
to apply tension loading to the anchor bolt through a loading beam.
down locations, the tie-down closer to the anchor would carry a load of
One displacement-controlled monotonic test and three load-controlled
238.4 kN [53.6 kips]. The tensile force at the top tie was thus 216.5 kN
cyclic tests were conducted, as indicated in Table 1. The loading steps
[48.7 kips], and the required reinforcement was calculated as 8.7 cm2
were chosen as 133 kN [30 kips] (elastic behavior), 248 kN [56 kips]
[1.4 in.2] using a steel stress of 0.6fy. Six No. 5 longitudinal bars were
(first bar yielding according to the monotonic behavior), 320 kN
used on the top face, and the same number of bars were used on the
[72 kips] (90 percent of the peak load), and 378 kN [76 kips] (95 per-
bottom face to facilitate the construction. A similar analysis was con-
cent of the peak load). Three loading cycles were used for each loading
ducted in the transverse directions, and the analysis indicate that No. 3
step, and the specimen was loaded monotonically to failure after the
stirrups with a spacing of 76 mm [3 in.] throughout the beam was
cyclic loading. The loading rate for the monotonic test was kept at
sufficient to restrain the splitting cracks. The dimensions of the nodal
2 mm/min [0.08 in./min] while the loading rate for the cyclic tests was
zone at the bottom of the beam, shown by a shaded rectangular in
0.89 kN/min [200 lbs/min].
Fig. 9, were defined by the load-carrying reinforcement (4 No. 4 stir-
rups), resulting in a nodal area of 152 × 152 mm [6 × 6 in.]. This nodal
zone was sufficient according to the regulations in ACI 318-08 Ap-
6.2. Observed anchor behavior
pendix A [7]. The strut from the bottom node to the tie-down at the top
of the beam can be a 3D bottle-shaped strut. According to RA3.3 of ACI
The tensile behavior of selected anchors with the proposed anchor
318-08, such struts should be crossed by reinforcement to resist the
reinforcement is shown in Fig. 11a, and the measured capacities listed
splitting tensile force from the compression force spreading. Two
in Table 1. The desired ductile steel failure was achieved in all four
longitudinal bars (No. 5) was placed along the side faces. These bars,
specimens, and the fractured anchors ae shown in the picture in
along with the provided stirrups, were found sufficient to confine the
Fig. 11b. The monotonic behavior, shown in dashed lines in Fig. 11a for
struts according to ACI 318-08 [7]. All longitudinal bars had a 90-
Specimen RG10-m, indicated that the specimen behaved elastically

116
D. Petersen et al.
a)
b)
Fig. 11. Behavior of reinforced anchors with proposed anchor reinforcement. (a) Load-
displacement behavior; (b) failure pictures.
before the anchor yielded at a displacement of 2 mm [0.08 in.]. The
measured anchor displacement, within this range of behavior, was
largely from the concrete compressive deformation above the head in
addition to the elongation of the anchor shaft. The concrete at the top
surface developed splitting cracks as shown in Fig. 11b, and the split-
ting cracks were restrained by the crack-controlling reinforcement.
The post-yielding deformation of the anchor was mainly from the
inelastic deformation of the anchor shaft. The total elongation of the
152-mm [6-in.] long anchor shaft, with a peak strain of 5 percent
(Fig. 4), was calculated as about 8 mm [0.3 in.], which is similar to the
measured anchor displacement at the peak load, as shown in Fig. 11a. A
sudden increase in the anchor displacement was captured at about
320 kN [72 kips], which was attributed to the flexural crack in the
concrete beam at the anchor location (Fig. 11b). The flexural crack was
again restrained upon further loading by the longitudinal reinforcement
in the beam, specifically designed for crack controlling purposes for
these specimens. Note that using longitudinal bars in flexural members
a) under monotonic loading;
Fig. 12. Measured strains in load-carrying reinforcement.
117
Engineering Structures 164 (2018) 109–118
as crack-controlling reinforcement in the longitudinal direction may not
be practical because the concrete members are expected to develop
inelastic behavior in an earthquake. Further tests of headed anchors
installed in the plastic hinge zone of reinforced concrete columns in-
dicated that local crack-controlling reinforcement can be effective [26].
The cyclic behavior of Specimen RG10-c2 was reliable, similar to
the monotonic behavior. The initial slope of Specimen G3 was higher
than that in Specimen G1, as shown in Fig. 11a. Before yielding, the
cyclic loading at 248 kN [56 kips] caused insignificant damage such
that the anchor displacements were almost the same at three peak
loads. Under larger loads, the concrete damage caused the anchor dis-
placement to noticeably increase at every consecutive peak load.
6.3. Observed anchor reinforcement behavior
Strain gages were placed on four legs of the stirrups, which served
as the load-carrying reinforcement. The positions of the strain gages are
shown by the inserted schematic in Fig. 12. An idealized breakout crack
was used as the reference, and the gages were placed 13 mm [0.5 in.]
below the assumed breakout cracks. Two gages, placed on the stirrups
next to the anchor, had a horizontal position of 19 mm [0.75 in.] from
the center of the anchor, as marked in Fig. 12; however the absolute
position of the stirrups legs was 79 mm [3.1 in.] from the center of the
anchor because the stirrups were 152 mm [6 in.] wide. Similarly, the
gages on the two outer stirrups were marked at 70 mm [2.75 in.] in
Fig. 12, and the absolute position of the stirrups was 103 mm [4.1 in.]
from the center of the test anchor.
The gages on the closer bars initially recorded larger strains than
those on the farther bars, which is similar to the observation by Lee
et al. [18]. In addition, the measured strains indicated that all No. 4
stirrup legs, as the load-carrying reinforcement, yielded when the an-
chor bolt fractured. Strains in the crack-controlling reinforcement were
not monitored. Further tests may be needed to verify the effectiveness
of using strut-and-tie models in the anchor reinforcement design.
7. Summary and conclusions
It is well recognized that the concrete around anchor connections
needs proper protection. The existing anchor reinforcement re-
commended in the literature, mainly U-shaped hanger steel, was eval-
uated using experimental tests in this study. Five patterns of anchor
tension reinforcement were implemented for single cast-in anchors,
made of 1-in. diameter ASTM A193 Grade B7 rods with a heavy hex
nut. The anchor reinforcement, in terms of the legs of closed stirrups,
was proportioned to carry a force equal to the tensile capacity of the
anchors, which were expected to develop ductile steel fracture.
Compared with the anchors embedded in plain concrete, a capacity
increase ranging from 40% to 130% was observed, which is similar to
the observations in the literature. However, the desired ductile steel
fracture was not achieved. Splitting cracks developed in concrete, and
the concrete around the anchor head lost its confinement and crushed
b) under cyclic loading
D. Petersen et al.
prematurely, resulting in anchor pullout failure.
Based on these observations and other tests in the literature, re-
commendations for anchor tension reinforcement were proposed. The
proposed anchor tension reinforcement includes: (1) load-carrying re-
inforcement in terms of closed stirrups; (2) crack-controlling re-
inforcement in all directions; and (3) local confining reinforcement near
the anchor head if side-face blowout may control the failure.
Specifically, the load-carrying reinforcement is recommended to be
proportioned to carry a force equal to the design tensile capacity of the
anchors, and implemented using small-diameter closed stirrups. The
closely spaced stirrups with a center-on-center spacing of 51 mm [2 in.]
to 76 mm [3 in.] can extend up to 1.0hef from the anchor bolt. Crack-
controlling reinforcement are necessary crossing all possible cracks. The
design of the crack-controlling reinforcement may use two strut-and-tie
models: one describing the load transfer from the anchor head to the
load-carrying reinforcement, and the other describing the load transfer
from the load-carrying reinforcement to the rest of the structure. All
crack-controlling reinforcement should be proportioned using a steel
stress of 0.6f y , and implemented using small diameter bars evenly dis-
tributed with a small and practical spacing in two orthogonal direc-
tions. In addition, the anchors and the load-carrying reinforcement
close to the anchor can be encased by three or four closed hoops in the
transverse direction near the anchor head if the anchors are close to a
free side surface.
The proposed anchor reinforcement was evaluated with additional
four tests. Splitting cracks initiated during the tests and were success-
fully restrained by the crack-controlling reinforcement. Reinforced with
the proposed anchor reinforcement, all four single anchors made with
1-in. diameter ASTM A193 Grade B7 rods, developed their full tensile
capacity and ductile behavior with an embedment depth of six times the
anchor diameter.
Acknowledgments
The study reported in this paper is from a project funded by the
National Science Foundation (NSF) of United States under Grant No.
0724097. The authors gratefully acknowledge the support of Dr. Joy
Pauschke, who served as the program director for this grant. The au-
thors also thank the colleagues in ACI committee 355 for their valuable
inputs. Any opinions, findings, and recommendations or conclusions
expressed in this material are those of the authors and do not ne-
cessarily reflect the views of NSF.
References
[1] Lifeline Earthquake Engineering (ASCE). Northridge earthquake: lifeline perfor-
mance and post-earthquake response. A report to U.S. department of commerce;
NIST Building and Fire Research Laboratory. Gaithersburg, MD; 1997.
118
Engineering Structures 164 (2018) 109–118
[2] Asia-Pacific Economic Cooperation. Earthquake disaster management of energy
supply system of APEC member economies. Taipei, China: Energy Commission,
Ministry of Economic Affairs; 2002.
[3] American Concrete Institute (ACI). State-of-the-art report on anchorage to concrete,
ACI 355.1R-91 (reapproved in 1997), American Concrete Institute, Farmington
Hills. MI.
[4] Comité Euro-International du Béton (CEB). Fastenings to concrete and masonry
structures: state of the art report. Thomas Telford Service Ltd., London; 1997.
[5] Federation Internationale du Beton (fib). Fastenings to concrete and masonry
structures. Special Activity Groups (SAG) 4 draft report, 2008, Obtained from Dr.
Eligehausen.
[6] Petersen D. Seismic behavior and design of cast-in-place anchors MS Thesis
Milwaukee, WI: University of Wisconsin; 2011.
[7] American Concrete Institute (ACI). Building code requirements for structural con-
crete (ACI 318-08). Farmington Hills, MI; 2008.
[8] European Committee for Standardization. Design of fastenings for use in concrete –
Part 4–2: headed fasteners. CEN/TS 1992-4-2, Brussels Belgium; 2009.
[9] Industry Handbook Committee. PCI design handbook: precast and prestressed
concrete, 6th ed. MNL-120-04. Chicago, IL: PCI; 2004.
[10] Anderson N, Meinheit D. A review of headed-stud design criteria in the sixth edition
of the PCI design handbook. PCI J 2007;52(1):82–100.
[11] Sharma A, Eligehausen R, Asmus J. Comprehensive analytical model for anchorages
with supplementary reinforcement. In: The 3rd international symposium on con-
nections between steel and concrete. Stuttgart, Germany; 2017.
[12] Berger W. Load-displacement behavior and design of anchorages with headed studs
with and without supplementary reinforcement under tension load (PhD Thesis)
NY: Institute of Construction Materials, University of Stuttgart; 2015. [in German].
[13] Cannon R, Godfrey D, Moreadith F. Guide to the design of anchor bolts and other
steel embedments. Concr Int 1981:28–41.
[14] Shipp J, Haninger E. Design of headed anchor bolts. Eng J Am Inst Steel Constr
1983;20:58–69.
[15] Hasselwander G, Jirsa J, Breen J, Lo K. Strength and behavior of anchor bolts
embedded near edges of concrete piers Research Report 29-2F TX: Center for
Transportation Research, the University of Texas at Austin; 1974.
[16] Widianto O, Patel C. Design of anchor reinforcement in concrete pedestals. In:
Proceedings of the 2010 structures congress. Orlando, FL; 2010. p. 2500–11.
[17] Baba N, Kanai S, Nishimura Y. Experimental study on stress transfer of joints
connected steel member with reinforced concrete member using anchor bolt. In:
Proceedings of the 14th world conference on earthquake engineering. Beijing,
China; 2008.
[18] Lee N, Kim K, Bang C, Park K. Tensile-headed anchors with large diameter and deep
embedment in concrete. ACI Struct J 2007;104(4):479–86.
[19] Sharma A, Eligehausen R, Asmus J. Comprehensive experimental investigations on
anchorages with supplementary reinforcement. In: The 3rd international sympo-
sium on connections between steel and concrete. Stuttgart, Germany; 2017.
[20] Saari W, Hajjar J, Schultz A, Shield C. Behavior of shear studs in steel frames with
reinforced concrete infill walls. J Constr Steel Res 2004;60(10):1453–80.
[21] Baran E, Schultz A, French C. Tension tests on cast-in-place inserts: influence of
reinforcement and prestress. PCI J 2006;51(5):88–108.
[22] Shahrooz B, Deason J, Tunc G. Outrigger beam-wall connections. I: Component
testing and development of design model. J Struct Eng 2004;130(2):253–61.
[23] Shahrooz B, Tunc G, Deason J. Outrigger beam-wall connections. II: subassembly
testing and further modeling enhancements. J Struct Eng 2004;130(2):262–70.
[24] Petersen D, Zhao J. Design of anchor reinforcement for seismic shear loads. ACI
Struct J 2013;110(1):53–62.
[25] Petersen D, Lin Z, Zhao J. Reinforcement for headed anchors Final report (volume
II) for project – behavior and design of cast-in-place anchors under simulated
seismic loading. Milwaukee, WI: University of Wisconsin; 2013.
[26] Lin Z, Luke B, Shahrooz B, Zhao J. Headed anchors in plastic hinge zone of re-
inforced concrete members Final report (volume III) for project – behavior and
design of cast-in-place anchors under simulated seismic loading. Milwaukee, WI:
University of Wisconsin; 2013.