ACI STRUCTURAL JOURNAL TECHNICAL PAPER

Title No. 119-S119

Group Behavior and Concrete Breakout Strength of 16 mm

Deformed Wire Anchors in Tension
by Mun-Gil Kim and Sung-Chul Chun

Due to lower carbon equivalents than those in steel reinforcing evaluate the strength of DWAs under tension, in each spec-
bars, deformed wires are easy to weld. Embedment steel plates imen, 25 DWAs with a specified yield strength of 515 MPa
welded with a straight deformed wire anchor (DWA) are commonly (75 ksi) conforming to ASTM A1064 (2018) were placed in
used to connect steel structural members to concrete members. a 5 x 5 form, and the 25 DWAs were simultaneously pulled
Because there is no anchor plate at the end of the DWA, the fail-
out. Experimental variables include the embedment depth
ures of concrete breakout and side-face blowout usually caused by
and spacing of the DWAs and the absence of cracking across
the anchor plate have not been considered. Only the development
length of deformed bars governs the anchorage design of the DWA. DWAs, and 13 specimens were tested. The measured loads
However, a concrete breakout is thought to be a possible failure. of specimens and the possibility of using Chapter 17 of ACI
To investigate the failure modes and anchorage strength of DWAs 318-19 to design DWAs were investigated by comparing test
under tension, 13 pullout tests of DWAs embedded in concrete of results with the CCD method (Fuchs et al. 1995).
fck = 42 MPa (6000 psi) were performed. In each specimen, 25
DWAs of fy = 515 MPa (75 ksi) were placed in a 5 x 5 form. The RESEARCH SIGNIFICANCE
experimental variables included the embedment depth and spacing For DWAs under tension, the tensile load is transferred
of the DWAs and the absence of cracking across the DWAs. Failure from the DWA to the concrete by bond along the embed-
modes and maximum loads were investigated and compared to ment depth, and thus, concrete breakout failure and concrete
predictions by the concrete capacity design (CCD) method (Fuchs
side-face blowout failures are not considered. In this study,
et al. 1995). All specimens failed due to concrete breakout. Eval-
13 pullout tests of DWAs without anchor plates arranged in
uating the experimental results with a 5% fractile safety factor
(Fuchs et al. 1995; Natrella 1966) shows that the concrete breakout a 5 x 5 form were performed. Test results show that even
strength of the DWAs is equivalent to the strength of a headed though DWAs have no anchor plate, concrete breakout
anchor and can be safely designed according to Chapter 17 of ACI failure occurred in DWAs, and the failure mode and tensile
318-19. behavior were very similar to concrete breakout of the group
anchor consisting of headed anchors or headed bolts. A
Keywords: anchorage; bond; concrete breakout failure; deformed wire comparison of the test results with the CCD method (Fuchs
anchors.
et al. 1995) shows that the concrete breakout strength of
DWAs is equivalent to the strength of a headed anchor, and
INTRODUCTION can be safety designed using Chapter 17 of ACI 318-19.
To connect a steel member into a concrete member, a
straight deformed wire anchor (DWA) without an anchor CODE PROVISIONS AND PREVIOUS RESEARCH
plate is allowed in ACI 349-13 Appendix D (ACI Committee ACI 318-19 and ACI 349-13
349 2013) and is considered as an alternative to a headed ACI 318-19 and ACI 349-13 allow straight anchors
stud or a headed bolt (Chicchi et al. 2020). Lee et al. (1989) without heads, such as Fig. 1 (ACI 349-13). The strength of
investigated the pullout behavior of a fabric made of the anchor under tension is calculated as the minimum value
deformed wires and Delhomme et al. (2016) and Engström among: a) steel strength; b) concrete breakout strength; c)
et al. (1998) evaluated the pullout strength of a ribbed bar. pullout strength; and d) concrete side-face blowout strength,
Because the tensile load is transferred from the DWA to according to Chapter 17 of ACI 318-19.
the concrete by bond along the embedment depth, concrete Nominal steel strength of an anchor in tension, Nsa, shall
breakout and concrete side-face blowout failures are not be calculated by Eq. (1)
considered. However, Chen (2020) and Chicchi et al. (2020)
reported the risk of concrete breakout failure even in DWAs (1)
N sa = Ase , N f uta
without anchor plates. If concrete breakout failure occurs
in DWAs, the concrete breakout prism may be smaller than where Ase,N is the effective cross-sectional area of an anchor
that of headed studs or headed bolts because DWAs have and futa is the specified tensile strength of anchor steel.
no head. Consequently, the concrete breakout strength of
DWAs may be less than predictions by the concrete capacity ACI Structural Journal, V. 119, No. 5, September 2022.
design (CCD) method (Fuchs et al. 1995), from which MS No. S-2021-329.R1, doi: 10.14359/51734669, received January 27, 2022, and
reviewed under Institute publication policies. Copyright © 2022, American Concrete
the design provisions of Chapter 17 of ACI 318-19 were Institute. All rights reserved, including the making of copies unless permission is
obtained from the copyright proprietors. Pertinent discussion including author’s
developed (ACI Committee 318 2019). In this study, to closure, if any, will be published ten months from this journal’s date if the discussion
is received within four months of the paper’s print publication.

ACI Structural Journal/September 2022 299

 Eq. (4) contains a safety factor of the 5% fractile (Fuchs
et al. 1995), it is necessary to exclude the safety factor to
predict the concrete breakout strengths. The kc values (Elige-
hausen et al. 2006; Fuchs et al. 1995) excluding the safety
factor for cast-in anchors are 13.3 (32 in U.S. Customary
units) and 16.74 (40 in U.S. Customary units) for cracked
and uncracked concrete, respectively.
The equations of concrete breakout strength are based on
the CCD method (Fuchs et al. 1995) for cast-in anchors with
heads. Therefore, the concrete breakout failure is not consid-
ered in straight deformed wires that transfer the tensile load
through deformations on the surface, such as indentations or
ribs. Side-face blowout strength need not be considered for
anchors without a head.

Fig. 1—Typical embedment for tension loads from ACI Seo et al. (2017)
349-13. Seo et al. (2017) compared the design values of steel
If the anchor has sufficient embedment length, pullout strength of anchors and concrete breakout strength of
failure does not occur. For deformed wires, ld shall be calcu- anchors. The concrete breakout strength of a group DWA
lated by Eq. (2) according to Chapter 25 of ACI 318-19 has not been considered when designing a group DWA in
accordance with ACI 349-13. The compressive strength
   
of concrete is 34.5 MPa (5000 psi), and yield strength of
    DWA is 448.2 MPa (65 ksi). It was reported that the concrete
fy t e  s  g  3 f y t e  s  g 
ld   d mm  = 
 1.1 f   c  K   b   40  f   c  K   b  
d in. breakout failure occurred because in the case of 16 (4 x 4)


c

b tr

 

c

b tr

 DWAs, the concrete breakout strength of anchors was lower
  db     db   than the steel strength of anchors. Therefore, Seo et al.
(2) (2017) insisted that verification through experiments and
analytic models was required, and design procedures and
40 Atr provisions were suggested.
K tr = (3)
sn
Chicchi et al. (2020)
where fy is the specified yield strength for non-prestressed
Chicchi et al. (2020) conducted a pullout test of group
reinforcement; fc′ is the specified compressive strength of
DWAs and group reinforcing bars (hereafter DRAs) under
concrete; ψt, ψe, ψs, and ψg are factors for casting location,
tension. Test results showed that the concrete breakout
reinforcement coating, reinforcement size, and grade of rein-
failure occurred in both DWAs and DRAs. It was reported
forcement, respectively; and λ is the modification factor for
that the failure load for the DRAs could be properly
lightweight concrete, in which the confinement term (cb +
predicted by Eq. (4) and (5) with the kc values excluding the
Ktr)/db shall not exceed 2.5. For a detailed definition, see
safety factor equation, and further studies would be required
Chapter 25 of ACI 318-19.
to verify various design conditions, such as diameters,
Nominal concrete breakout strength in tension Ncbg of an
spacing, and boundary conditions. The anchorage strength
anchor group shall be calculated by Eq. (4) and (5)
calculated based on the development length of ACI 318-19
overestimated the loads of DWAs. Although the DWAs also
ANc had concrete breakout failure, pullout failure occurred at the
N cbg   ec , N  ed , N  c , N  cp , N N b (4)
ANco same time, because the DWAs had lower bond capacity than
the DRAs. The starting point of concrete breakout failure
N b  kc f chef 1.5 (5) varies depending on the anchor’s bond strength, which
affects the interaction with adjacent anchors. Further studies
where ANc is the projected concrete failure area of an anchor are required to calculate accurately the strength of DWAs.
group; ANco (=9hef2) is the projected concrete failure area
of a single anchor with an edge distance of at least 1.5hef; Chen (2020)
the factors ψec,N, ψed,N, ψc,N, and ψcp,N represent modifica- The failure mode and strength of DWAs under tension
tion factors for eccentric loading, edge effects, uncracked with the embedment depth equal to the development length
concrete, and critical spacing of post-installed anchors to of reinforcing bars according to ACI 318-19 were eval-
control splitting, respectively; hef is the effective embedment uated through finite element analysis. It was found that
depth of anchor; and kc is the coefficient for basic concrete pullout failure and concrete breakout failure simultaneously
breakout strength in tension. occurred. For headed anchor groups with the same projected
In ACI 318-19, the value of kc for cast-in anchors is 10 (24 concrete failure area, ANc, the concrete breakout failure
in U.S. Customary units) for cracked concrete, and 12.5 (30 strengths of the groups were not affected by the total number
in U.S. Customary units) for uncracked concrete. Because of anchors. However, the pullout strengths were affected by

300 ACI Structural Journal/September 2022

Fig. 2—Details of specimen.
the number of anchors. Consequently, for two specimens (18.9 in.), and three embedment depths of 330, 457, and
with the same ANc, the specimen with 25 DWAs had higher 591 mm (13, 18, and 23 in.) were used. The spacing between
failure load than the specimen with 16 DWAs. DWAs was 152 mm (6.0 in.) except for H23-S4 specimens,
which had 101 mm (4 in.) spacing between the DWAs. To
EXPERIMENTAL PROGRAM control the occurrence of cracks across the DWAs, three
Test variables and specimen design specimen heights of 800, 850, and 1000 mm (31.5, 33.5,
The design compressive strength of concrete is 42 MPa and 39.4 in.) were used. To evaluate the possibility that
(6000 psi). Considering the maximum capacity of 3000 kN reinforcing bars placed perpendicularly to the anchor could
(674 kip) of the loading apparatus, the number of DWAs was enhance the concrete breakout capacity, the reinforcement
determined to be 25. The diameter, cross-sectional area, and ratio was added as a test variable. In Table 1, three iden-
design yield strength of the DWAs are 16 mm (0.63 in.), tical specimens were constructed for H23, H23-S4, and H18,
200 mm2 (0.31 in.2), and 515 MPa (75 ksi), respectively. and two identical specimens were constructed for H13 and
To stiffen the embedment plate, 32 mm (1.3 in.) thick ribs H18-UR. A total of 13 specimens were tested.
were welded onto the top surface of the embedment plate,
as shown in Fig. 2. Test procedure
Table 1 shows the test matrix, while Fig. 2 shows the details Figure 3 shows the test setup. The reaction frames were
of the specimens. The main parameters used in Eq. (4) and placed on both ends of the specimen, and the simple beam-
(5)—embedment depth, anchor spacing, and the presence type loading frame was located on the reaction frames. A
or absence of cracking—were included in the test variables. center-hole-type hydraulic cylinder was placed on the center
The embedment depth significantly affects the basic concrete of the loading frame. The force was applied through five
breakout strength Nb and the projected concrete failure area loading bars, of which the tensile strength and cross-sec-
ANc. The spacing between anchors also affects the projected tional area are 1050 MPa (152 ksi) and 1257 mm2 (1.95 in.2),
concrete failure area ANc. The development length of DWAs respectively. One loading bar was installed in the middle
in accordance with Eq. (2) was calculated as 480.8 mm of the embedment plate, while the other four loading bars

ACI Structural Journal/September 2022 301

Table 1—Test matrix
Specimen Effective embedment DWA spacing s, Steel ratio Absence of Specimen dimensions Spacing of loading bars
ID* depth hef, mm (in.) mm (in.) ρ, % cracking L x W x D, m (ft) sl1 x sl2, m (ft)
152 3.3 x 2.8 x 0.8 0.4 x 0.4
H23 1.06
(6.0) (10.8 x 9.2 x 2.6) (1.3 x 1.3)
591 (23.25) Cracked
102 3.2 x 2.75 x 0.85 0.3 x 0.4
H23-S4 1.00
(4.0) (10.5 x 9.0 x 2.8) (1.0 x 1.3)
H18 457 (18.00) 0.85
152 3.0 x 2.5 x 1.0 0.4 x 0.4
H13 330 (13.00) 0.85 Uncracked
(6.0) (9.85 x 8.2 x 3.3) (1.3 x 1.3)
H18-UR 457 (18.00) 0.51
*
H-S or -UR:  is embedment depth, in inches.;  is spacing between DWAs, in inches.; and ‘UR’ means that specimen had lower reinforcement ratio than H18 specimen.
Note: Refer to Fig. 2 for notations of hef, s, L, W, D, sl1, and sl2; and ρ is reinforcement ratio of bars perpendicular to DWAs.

Fig. 3—Test setup.

were installed at 400 x 400 mm (15.75 x 15.75 in.) to mini- the construction process of specimens. The specimens
mize eccentricity by evenly distributing the tensile forces simulating a wall were vertically constructed, as shown in
to the DWAs. Because the H23-S4 specimen had narrow Fig. 5(b).
DWA spacing, loading bars were installed at 400 x 300 mm
(15.75 x 11.8 in.). The loading bars were connected to the Expected failure mode
embedment plate using nuts; four nuts on every corner Because a DWA does not have a head or a hook, concrete
were placed beneath the embedment plate, and one nut at breakout mixed with pullout at the ends of embedment
the center was welded onto the embedment plate because depth, as shown in Fig. 6, is possible. Figure 7 shows the
a DWA was welded beneath the center of the embedment expected load for the H23 specimen on the assumed failure
plate, as shown in Fig. 3. of concrete breakout mixed with pullout along the portion
The applied load was measured using a hydraulic pres- of c as shown in Fig. 6, where the x-axis represents the ratio
sure gauge installed on the pump connected to the hydraulic of pullout length c to the embedment depth hef. The pullout
cylinder. Figure 4 shows the locations of displacement trans- contribution is obtained by using Eq. (2) with c of pullout
ducers and strain gauges. The relative vertical displacements length, and the safety factor of 1.25 (Orangun et al. 1977) is
of the embedment plate and concrete were measured at excluded. The concrete breakout contribution is obtained by
two points as shown in Fig. 4(a), and the tensile deforma- using Eq. (4) with kc of 13.3 (32 in U.S. Customary units)
tions, including crack width, of the concrete surface were (Fuchs et al. 1995) for cracked concrete.
measured at three regions of the concrete surface adjacent As the pullout length c increases, the increase in pullout
to the embedment plate, as shown in Fig. 4(b). In H18, H13, strength is greater than the decrease in concrete breakout
and H18-UR specimens, two strain gauges per DWA were strength of DWAs with reduced embedment depth of hef – c.
attached to five DWAs as shown in Fig. 4(d) to measure the Consequently, when only the concrete breakout failure
tensile strains of the DWAs at the 13 mm (0.5 in.) distance occurs, c/hef = 0, the expected load is the lowest. For all spec-
beneath the embedment plate. The strains of the five loading imens, the same failure of concrete breakout is expected.
bars were also measured, as shown Fig. 4(c). Figure 5 shows

302 ACI Structural Journal/September 2022

Fig. 4—Layout of instrumentation.

Fig. 5—Construction of specimens.

TEST RESULTS strength are 566.8 and 620.3 MPa (82.2 and 90.0 ksi),
Result of material test respectively, satisfying ASTM A1064 Grade 75.
The average compressive strength of concrete is 44.4 MPa
(6444 psi) for H23 and H23-S4 specimens, and 41.3 MPa Failure modes
(5994 psi) for H18, H13, and H18-UR specimens. Figure 8 Figure 9 shows typical failures, and concrete breakout
shows the measured stress-strain curve of DWAs. Because failure occurred in all 13 specimens. Circular and radiation-type
the DWA is a cold-formed product, the yield point is not cracks occurred on the top of the concrete around the embed-
clear, and therefore the yield stress is determined from a ment plate. For all specimens, the failure area originating
0.2% offset method. The average yield strength and tensile from the end of the DWAs formed a cone-shaped failure

ACI Structural Journal/September 2022 303

Fig. 6—Expected failure surface of concrete breakout mixed with pullout.

Fig. 7—Expected strength of concrete breakout mixed with Fig. 8—Stress-strain relationships of DWA.
pullout (H23).

Fig. 9—Typical specimen failures (grids spaced at 200 mm [7.9 in.]).

304 ACI Structural Journal/September 2022
Fig. 10—Load-vertical displacement relationships.
prism, and the slopes of the failure surface were greater Load-vertical displacement relations
than 1:1.5. Although H23 and H23-S4 specimens with the Figure 10 shows the relations of load-vertical displace-
longest embedment depth of 591 mm (23 in.) had some pull- ment for all specimens. The horizontal lines represent the
outs at the end of the DWAs, their final failure mode was expected strength by Eq. (4) with the measured concrete
identical to concrete breakout. Headed anchors or headed strength and kc values of 13.3 and 16.74 (32 and 40 in U.S.
bolts transfer a tensile force to concrete through the bearing Customary units) for cracked and uncracked concrete,
of heads, but DWAs without a head transfer a tensile force respectively. The specimens of H23 and H23-S4 were
to concrete by the bond along the embedment depth. Test expected to be cracked concrete, while the specimens of
results show concrete breakout failures for DWAs that were H18, H13, and H18-UR were expected to be uncracked
similar to those for headed anchors or headed bolts. Because concrete. The measured maximum loads of the H23 spec-
there is no head at the end of DWAs, it is considered that imens exceed the expected strength, and the average load
after the concrete breakout failure occurred, the concrete at of H23-S4 specimens is almost the same as the expected
the end of DWAs was relaxed. strength. The maximum loads measured in H18, H13, and
H18-UR specimens are a little less than the expected values,

ACI Structural Journal/September 2022 305

Table 2—Comparison of predicted strength and test result
Specimen ID ψec,N1 ψec,N2 Ncbg,p1, kN (kip) Ncbg,p2, kN (kip) Ncbg,t, kN (kip) Ncbg,t/Ncbg,p1 Ncbg,t/Ncbg,p2
H23-1 0.994 N.A. 2185 (491) N.A. 2245 (504) 1.03 N.A.
H23-2 0.994 N.A. 2205 (496) N.A. 2515 (565) 1.14 N.A.
H23-3 0.995 N.A. 2226 (500) N.A. 2555 (574) 1.15 N.A.
H23-S4-1 0.993 N.A. 1941 (436) N.A. 2220 (499) 1.14 N.A.
H23-S4-2 0.992 N.A. 1939 (436) N.A. 2002 (450) 1.03 N.A.
H23-S4-3 0.993 N.A. 1832 (412) N.A. 1725 (388) 0.94 N.A.
H18-1 0.997 0.963 2185 (491) 2111 (474) 2136 (480) 0.98 1.01
H18-2 0.997 0.977 2185 (491) 2142 (481) 2132 (479) 0.98 1.00
H18-3 0.993 0.985 2177 (489) 2159 (485) 2143 (482) 0.98 0.99
H13-1 0.991 0.989 1668 (375) 1665 (374) 1649 (371) 0.99 0.99
H13-2 0.986 0.990 1660 (373) 1667 (375) 1732 (389) 1.04 1.04
H18-UR-1 0.983 0.967 2155 (484) 2120 (476) 2066 (464) 0.96 0.97
H18-UR-2 0.997 0.958 2185 (491) 2101 (472) 2081 (468) 0.95 0.99
Avg. 1.02 1.00
SD 0.075 0.021
COV 7.3% 2.1%

Note: Avg., SD, and COV are mean, standard deviation, and coefficient of variation of Ncbg,t/Ncbg,p1 ratio, respectively.

but are almost the same as the expected values. For the H23
specimens, vertical displacements at the maximum loads are
larger than the displacements of the other specimens. The
large vertical displacement of H23 specimen resulted from
the elongation of DWAs at the maximum load. Figure 8
shows that the yield point of DWA is unclear, and the strain
at the yield strength is much higher than fy/Es. Even though
the maximum loads of H23 specimens are less than the yield
force of 2834 kN (636.9 kip) of the 25 DWAs, the vertical
displacement is significantly large.

Eccentricity of applied load

The five loading bars shown in Fig. 3 and 4(c) were used. Fig. 11—Load-strain relationships of DWAs (H18-3).
Because the loading bars did not yield at maximum loads, the
tensile forces applied to the loading bars are linearly propor- where sr1 and sr2 are the x- and y-axis distances from the
tional to the measured strains. With the strains of the loading center of the embedment plate to the loading bar, respec-
bars, the eccentricity of the applied load can be calculated tively; and ε1 to ε5 are the strain of each loading bar.
by Eq. (6) and (7). Table 2 lists modification factors for
the eccentricity ψec,N1 calculated by Eq. (8) according to Stress-strain relationships of DWA
Chapter 17 of ACI 318-19. For all specimens, the minimum, Figure 11 shows the load-strain relation for H18-3 that
maximum, and average values of ψec,N1 are 0.983, 0.997, and failed by concrete breakout before DWAs reached the yield
0.993, respectively, so there was little effect of eccentricity strain. In all specimens of H18, H13, and H18-UR, the
strains of DWAs did not reach the yield strain. H18 and
sr1  1  3   2   4  sr 2  1   2  3   4  H18-UR specimens had an embedment depth of 457 mm
ex  5
, ey  5 (6) (18 in.), which is approximately similar to the development
 i  i length of 480.8 mm (18.9 in.) calculated by Eq. (2), but the
i 1 i 1
concrete breakout failure occurred before DWAs yielded.
eN  ex2  ey2 (7) In anchor groups, the resistances of individual anchors are
not uniform, even with no eccentricity in the applied load,
due to the inhomogeneity of concrete. Table 2 shows the
1 modification factors for the eccentricity ψec,N2 of the H18,
 ec , N 1  (8) H13, and H18-UR specimens that are calculated using the
 2eN  strains of DWAs in the same manner as the eccentricity
1  
 3hef  of loading bars. Most of the modification factors based on
the DWA strains are slightly lower than those based on the

306 ACI Structural Journal/September 2022

 Table 3—Measured crack width at maximum load,
mm
Specimen ID Crack width 1 Crack width 2 Crack width 3
H23-1 0.48 (0.019) 0.16 (0.006) 0.13 (0.005)
H23-2 0.58 (0.023) 0.33 (0.013) 0.24 (0.009)
H23-3 0.39 (0.015) 0.31 (0.012) 0.27 (0.011)
H23-S4-1 0.37 (0.015) 0.04 (0.002) 0.11 (0.004)
H23-S4-2 0.18 (0.007) 0.10 (0.004) 0.22 (0.009)
H23-S4-3 0.34 (0.013) 0.04 (0.002) 0.16 (0.006)
H18-1 0.19 (0.007) 0.03 (0.001) 0.13 (0.005)
H18-2 0.10 (0.004) 0.04 (0.002) 0.13 (0.005)
H18-3 0.11 (0.004) 0.05 (0.002) 0.06 (0.002)
H13-1 0.20 (0.008) 0.04 (0.002) 0.23 (0.009)
H13-2 0.06 (0.002) 0.09 (0.004) 0.03 (0.001)
H18-UR-1 0.18 (0.007) 0.17 (0.007) —
H18-UR-2 0.16 (0.006) 0.05 (0.002) 0.16 (0.006)

Note: 1 mm = 0.04 in.

All of the H23 and H23-S4 specimens are determined

to be cracked concrete from the measured crack widths.
For all H23 and H23-S4 specimens, except H23-S4-2, the
crack widths are greater than 0.3 mm (0.012 in.), while the
crack width of H23-S4-2 at the maximum load is 0.22 mm
(0.009 in.). Because the measuring point of the crack width
is over 400 mm (15.7 in.) distance from the center of the
DWAs, as shown in Fig. 4(b), the moment at the center of
the DWAs is higher than the moment at the measuring point,
Fig. 12—Load-crack width relationships. and therefore the crack width at the center may be greater
than the measured crack widths. Consequently, all H23
loading bar strains. However, the minimum, maximum, and and H23-S4 specimens are judged to be cracked concrete.
average values of ψec,N2 are 0.957, 0.990, and 0.976, so the From Table 3 and Fig. 12(b), all H13, H13, and H18-UR
effect of eccentricity is not significant. specimens are judged to be uncracked concrete, because the
crack widths are less than 0.3 mm (0.012 in.), or less at the
Crack width maximum load.
The coefficient kc for basic concrete breakout strength in
tension of ACI 318-19 was developed based on the crack Ultimate strength evaluation
width of 0.3 mm (0.012 in.) (Fuchs et al. 1995). Figure 4(b) The expected strengths are calculated by Eq. (4) with the
shows that a side of the embedment plate was divided kc values of 13.3 (32 in U.S. Customary units) and 16.74
into three sections, and linear variable displacement trans- (40 in U.S. Customary units) for cracked and uncracked
ducers (LVDTs) were horizontally installed to measure the concrete, respectively. The modification factors for the
crack width, including the tensile deformation of concrete. eccentricity of ψec,N1 and ψec,N2 calculated using the strains
Figure 12 shows the load-crack width relations measured in of the loading bars and DWAs were applied to the expected
the H23-S4-1 and H18-2 specimens. The absence or pres- strengths, and then Table 2 compares the modified strengths
ence of cracking of each specimen was determined from the of Ncbg,p1 and Ncbg,p2 with the test loads. Due to the test setup
measured crack widths. When a crack occurs in one of three as shown in Fig. 3, the loading-system weight of the embed-
sections, in general, the tensile strain of the concrete adjacent ment plate, loading bars, and connection frames is included
to the crack is reduced. In Fig. 12(a), when a crack occurred in the measured loads. The loading-system weight is 17.9 kN
in the first section at approximately 1200 kN (270 kip), the (4.0 kip) for H23-S4 specimens, and 19.6 kN (4.4 kip) for the
measured width of the third section was a little reduced. At other specimens. The test loads Ncbg,t in Table 2 are obtained
approximately 1770 kN (398 kip), the width in the second by subtracting the loading-system weight from the measured
section decreased as a crack occurred in the third section. loads in the tests.
The cracking was confirmed by the increased crack width of The mean and coefficient of variation (COV) values of the
a particular section, and the decreased crack width of adja- Ncbg,t/Ncbg,p1 ratios are 1.02 and 7.3%, respectively. The test
cent sections. Table 3 lists the crack widths at the maximum results are almost the same as the predictions by the CCD
load for each specimen. method (Fuchs et al. 1995), which is developed from the

ACI Structural Journal/September 2022 307

headed studs and headed bolts. The mean and COV values of breakout strength of DWAs can be regarded as being propor-
the Ncbg,t/Ncbg,p2 ratios are 1.00 and 2.1%, respectively. The tional to hef1.5.
expected strengths with ψec,N2 are more accurate and reliable, Figure 14(b) investigates the effects of the spacing by
but the strains of DWAs were measured in only six speci- comparing the H23 and H23-S4 specimens that have the
mens of H23 and H23-S4. Figure 13 compares the test loads same design parameters, except for the spacing between
and the expected strengths. The CCD method can accurately DWAs. The test-to-prediction ratios show no significant
predict the concrete breakout strength of DWAs without bias trend against increasing the spacing. DWAs transfer force
on the absence or presence of cracking. to the concrete through the bond on the surface of DWAs,
but failure occurs due to the concrete breakout of a group
Concrete breakout strengths of DWAs anchor, not to the splitting of concrete around the DWAs
The test results show that the failure mode of DWAs as (ACI Comittee 408 2003). The spacing between DWAs
a group anchor is concrete breakout failure. Whether the affects the size of the concrete breakout prism in the same
factors used in the design equations of concrete breakout manner as a group of headed anchors.
failure strength in ACI 318-19 affect the concrete breakout The effects of the ratios of reinforcing bars placed perpen-
strength of DWAs is investigated. In Eq. (5), the concrete dicular to DWAs on the concrete breakout strength is inves-
breakout strength is proportional to hef1.5. In Fig. 14(a), the tigated with comparing the measured loads of H18 and
measured loads and the predicted strengths of H23, H18, and H18-UR specimens that have the same design parameters
H13 specimens are compared with varying the embedment except for the reinforcement ratios of 0.85 and 0.51% for
depth hef. The test-to-prediction ratios show no particular H18 and H18-UR, respectively. According to Chapter 17
trend against the embedment depth. Therefore, the concrete of ACI 318-19, if the reinforcement ratio conforms to the
minimum requirement, the ratio does not quantitatively affect
the concrete breakout strength. If the minimum requirement
is not satisfied, a lower strength reduction factor should be
used. Although the reinforcement ratios of the specimens
differed by 1.67 times, the mean value of H18 specimens
is 2137 kN (480 kip), which is 1.03 times the mean value
of H18-UR specimens of 2074 kN (466 kip), as shown in
Table 2. This comparison shows that the ratio of reinforce-
ment perpendicular to the DWAs does not quantitatively
affect the concrete breakout strength of DWAs, similar to the
design provision of a headed anchor of ACI 318-19.

Safety ratio of design equation

The nominal concrete breakout strength of ACI 318-19
was developed from the CCD method (Fuchs et al. 1995)
with a safety factor 0.75 from the mean value of tests, as
shown in Eq. (9). The safety factor 0.75 was determined
as the 5% fractile coefficient of n5% (ACI 318 2019; Elige-
hausen et al. 2006; Fuchs et al. 1995; Natrella 1966)
Fig. 13—Comparison of measured loads with predicted
strengths.

Fig. 14—Test-to-prediction ratios with varying test variables.

308 ACI Structural Journal/September 2022

  increasing the steel ratio of the reinforcing bars perpendic-
 
f n  f m  k   1  k  f m  1  k   f m  n5% f m (9) ular to the DWAs is not effective for increasing the strength.
 fm  3. A comparison of the measured loads with the strengths
where fn is the nominal strength; fm and σ are the mean predicted by the CCD method (Fuchs et al. 1995), which
strength and standard deviation of data, respectively; δ is the is the basis of the design provisions of Chapter 17 of ACI
COV; and k is the coefficient according to the number of data 318-19, shows that the mean and coefficient of variation of
(Natrella 1966) the test-to-prediction ratios are 1.02 and 7.3%, respectively.
The 5% fractile coefficient of the test-to-prediction ratios
N cbg , n   Ncbg ,t  k  Ncbg ,t  n5%  Ncbg ,t (10) of the thirteen specimens is estimated to be 0.844, which is
higher than the safety factor of 0.75 adopted in ACI 318-19
where Ncbg,n is the nominal strength; and µ Ncbg ,t and σ Ncbg ,t are for the nominal concrete breakout strength (Fuchs et al.
the mean and the standard deviation of the measured load, 1995). The concrete breakout failure should be considered
respectively. in the design of DWAs and the concrete breakout strengths
Substituting Ncbg,n and µ Ncbg ,t for fn and fm, respectively, can be obtained by Chapter 17 of ACI 318-19, the same as
gives Eq. (10), and then multiplying the right side by Ncbg,t/ the headed anchors.
Ncbg,p yields Eq. (11). The 5% fractile coefficient n5% of
Eq. (12) for the ratio of [measured load/predicted strength] AUTHOR BIOS
Mun-Gil Kim is a Postdoctoral Researcher in the Division of Architecture
is driven by Eq. (11) at Incheon National University, Incheon, Korea. He received his BS from
Mokpo National University, Muan County, Korea, in 2015, and his MS and
  PhD from Incheon National University in 2017 and 2022, respectively.
N cbg ,t

N cbg , n   Ncbg ,t  k  Ncbg ,t   N cbg , p
   Ncbg ,t  k  Ncbg ,t
 N
 N cbg , p  n5% N cbg , p
 ACI member Sung-Chul Chun is a Professor in the Division of Architec-
 cbg , p N cbg , p  ture and Urban Design at Incheon National University. He received his
(11) BS, MS, and PhD from Seoul National University, Seoul, Korea, in 1994,
1996, and 2007, respectively. He is a member of Joint ACI-ASCE Commit-
tees 352, Joints and Connections in Monolithic Concrete Structures, and
n5%   Ncbg ,n  k  Ncbg ,n  1.02  2.403  0.075  0.844 (12) 408, Bond and Development of Steel Reinforcement. His research interests
N cbg , p N cbg , p include steel anchorage to concrete, composite structures, and the rehabil-
itation of reinforced concrete structures.
where µ Ncbg ,t , σ Ncbg ,t , and δ Ncbg ,t are the mean, standard devia-
N cbg , p N cbg , p N cbg , p ACKNOWLEDGMENTS
This work was supported by the Korea Hydro & Nuclear Power Co. Ltd.
tion, and COV of the [measured load/predicted strength] and KEPCO Engineering & Construction Company (Research on DWAs),
ratios, respectively; and k = 2.403, corresponding to the and a National Research Foundation of Korea (NRF) grant, supported by
number of data, n = 13 (Natrella 1966). the Korea government (MSIT) (No. 2021R1A2C2010863).
The 5% fractile coefficient of 0.844 is sufficiently higher
than the safety factor 0.75 adopted in ACI 318-19 for the REFERENCES
ACI Committee 318, 2019, “Building Code Requirements for Struc-
nominal concrete breakout strength (Fuchs et al. 1995). tural Concrete (ACI 318-19) and Commentary (ACI 318R-19),” American
Consequently, the concrete breakout strengths of DWAs Concrete Institute, Farmington Hills, MI, 623 pp.
calculated conforming to Chapter 17 of ACI 318-19 are ACI Committee 349, 2013, “Code Requirements for Nuclear Safety-Related
Concrete Structures (ACI 349-13) and Commentary,” American Concrete
acceptable, and the safety margin is not less than that of the Institute, Farmington Hills, MI, 196 pp.
headed anchors. ASTM A1064/A1064M-18, 2018, “Standard Specification for Carbon-
Steel Wire and Welded Wire Reinforcement, Plain and Deformed, for
Concrete,” ASTM International, West Conshohocken, PA, 10 pp.
CONCLUSIONS Chen, M. J., 2020, “Concrete Tensile Breakout at Fully Developed
Pullout tests of group deformed wire anchors (DWAs) Deformed Bar/Wire Anchors,” ACI Structural Journal, V. 117, No. 6, Nov.,
were conducted to investigate the failure mode and to pp. 45-57.
Chicchi, R.; Varma, A. H.; Seo, J.; Bradt, T.; and McCarty, E., 2020,
evaluate the strengths of group DWAs that do not have a “Experimental Testing of Tension-Loaded Deformed Anchors in Concrete,”
head or hook. Test variables include the embedment depth, ACI Structural Journal, V. 117, No. 5, Sept., pp. 133-146.
spacing between DWAs, and steel ratio of the reinforcing Delhomme, F.; Roure, T.; Arrieta, B.; and Limam, A., 2016, “Pullout
Behavior of Cast-in-Place Headed and Bonded Anchors with Different
bars perpendicular to the DWAs, and a total of 13 speci- Embedment Depths,” Materials and Structures, V. 49, No. 5, pp. 1843-
mens were tested. Based on the findings of this study, the 1859. doi: 10.1617/s11527-015-0616-4
following conclusions can be made: Eligehausen, R.; Mallee, R; and Silva, J. F., 2006, Anchorage in Concrete
Construction, Ernst & Sohn, 378 pp.
1. Thirteen pullout tests of DWAs show that concrete Engström, B.; Magnusson, J.; and Huang, Z., 1998, “Pull-Out Bond
breakout failure occurred, not bond failure. The failure Behavior of Ribbed Bars in Normal and High-Strength Concrete with
prism, which began at the end of the DWAs, is very similar Various Confinements, Bond and Development of Reinforcement,” Bond
and Development of Reinforcement—A Tribute to Dr. Peter Gergely,
to that of the headed anchors with an assumed failure angle SP-180, R. Leon, ed., American Concrete Institute, Farmington Hills, MI,
of 1:1.5 in Chapter 17 of ACI 318-19. pp. 215-242.
2. The measured loads were almost proportional to hef1.5, Fuchs, W.; Eligehausen, R.; and Breen, J. E., 1995, “Concrete Capacity
Design (CCD) Approach for Fastening to Concrete,” ACI Structural
as were the design provisions of ACI 318 of concrete Journal, V. 92, No. 1, Jan.-Feb., pp. 787-802.
breakout strengths. The spacing between DWAs also affects Joint ACI-ASCE Committee 408, 2003, “Bond and Development of
the measured loads in the same manner as the design provi- Straight Reinforcing Bars in Tension (ACI 408R-03),” American Concrete
Institute, Farmington Hills, MI, 49 pp.
sions of ACI 318 of concrete breakout strengths. In addition,

ACI Structural Journal/September 2022 309

 Lee, S. L.; Mansur, M. A.; Tan, K. H.; and Kasiraju, K., 1989, “Crack Orangun, C. O.; Jirsa, J. O.; and Breen, J. E., 1977, “Reevaluation of
Control in Beams Using Deformed Wire Fabric,” Journal of Struc- Test Data on Development Length and Splices,” ACI Journal Proceedings,
tural Engineering, ASCE, V. 115, No. 10, pp. 2645-2660. doi: 10.1061/ V. 74, No. 3, Mar., pp. 114-122.
(ASCE)0733-9445(1989)115:10(2645) Seo, J.; Chicchi, R.; Varma, A. H.; and Bradt, T., 2017, “Consideration of
Natrella, M. G., 1966, “Experimental Statistics,” National Bureau of Breakout Failure Modes in Design of Attachments to Concrete Structures,”
Standards Handbook, V. 91, pp. 1-14. SMiRT-24, BEXCO, Busan, Korea, Aug. 20-25.

310 ACI Structural Journal/September 2022

Copyright of ACI Structural Journal is the property of American Concrete Institute and its
content may not be copied or emailed to multiple sites or posted to a listserv without the
copyright holder's express written permission. However, users may print, download, or email
articles for individual use.