aaa

ACI Structural Journal/March-April 2013 239

Title no. 110-S21

ACI STRUCTURAL JOURNAL TECHNICAL PAPER

ACI Structural Journal, V. 110, No. 2, March-April 2013.MS No. S-2011-112.R2 received October 11, 2011, and reviewed under Institute

publication policies. Copyright © 2013, American Concrete Institute. All rights reserved, including the making of copies unless permission is obtained from the copyright proprietors. Pertinent discussion including author’s closure, if any, will be published in the January-February 2014 ACI Structural Journal if the discussion is received by September 1, 2013.

Tensile Resistance of Steel-Reinforced Anchorages: Experimental Evaluationby José Henriques, José Maria Raposo, Luís Simões da Silva, and Luís Costa Neves

One of the research topics related to composite steel-concrete structures is the connection between steel and concrete members—namely, the design of the anchorages. Because the design of steel-reinforced anchorages using the available methodologies (namely, from the CEB Design Guide, CEN/TS 1992-4, and ACI 318-08) may lead to excessively conservative results, experimental studies were conducted at the Civil Engineering Department of the University of Coimbra dealing with cast-in-place headed anchors. The aim of this study is the evaluation of the effect of hanger reinforcement on the behavior of steel-headed anchors embedded in concrete subjected to tension loads. A total of 40 laboratory tests, divided into two groups, were performed, dealing with plain concrete; hanger and usual grid surface reinforcement; and the variation of the member thickness, embedment depth, and proximity of the concrete member edges. Some conclusions from the tests are analyzed and discussed.

Keywords: anchorage failure; hanger reinforcement; headed anchors (fasteners); tension tests.

INTRODUCTIONThe connections between steel and concrete have been the

subject of several studies—with notable achievements both in the understanding of connection behavior and connec-tion design.1-12 Different types of anchors may be used to transfer shear and/or tension between structural members and the correct assessment of their behavior in terms of stiff-ness and resistance is a key step for the characterization of connections employing such anchors. With advancements in the understanding of the failure mechanisms, a method for the prediction of the concrete failure modes known as the Concrete Capacity Design Method (CC Method) was proposed by Fuchs et al.8 Farrow and Klingner7 verified the adequacy of the method with experimental tests.

Based on experimental investigations, this paper contributes to the understanding of the behavior of anchorages—using headed anchors—subjected to tension loading and reinforced with steel hanger reinforcement bars. It is shown that the presence of steel reinforcement in the concrete—either as hooked bars (hanger reinforcement) or as a grid surface reinforcement—enhances the resistance and ductility of the connection. In addition, the influence of the anchor embed-ment depth/member thickness ratio was also investigated.

RESEARCH SIGNIFICANCEThe aim of this study is to properly understand the behavior

of headed anchors subjected to tension in the presence of reinforcement designed to increase the anchor capacity. Experimental tests were carried out with the objective of obtaining data for the calibration of numerical13 and analyt-ical models and the improvement of current design methods. The performed tests enhanced the contribution of the hanger reinforcement to improving the resistance and deformation capacity of the anchorage in concrete. The comparison with

current design methods demonstrated that the use of this type of reinforcement is conservatively taken into account.

DESIGN PROCEDURESThe first design code to incorporate a method for deter-

mining the strength of fasteners was ACI 349-7614 in the United States. Later, the CC Method was adopted by ACI 318-08.3 It was also incorporated in design-oriented documents, such as the CEB Design Guide,1 the EOTA code,15 the recent CEN/TS 1992-4,2 and the Hilti design manual.16 It has also been incorporated in other design codes worldwide, such as the German code (DIBt)17 and the Swiss code (SIA 179).18 In the scope of the Eurocodes,19-21 the design of fasteners has not yet been completely addressed. In 1987, CEB created the working group Fastenings to Reinforced Concrete and Masonry Structures, which produced the publication of a state-of-the-art report4 and the design manual “Design of Fastenings in Concrete,”1 based on the partial factors methodology. Recently, the Euro-pean Committee for Standardization (CEN) published the technical specification “Design of Fastenings for Use in Concrete.”2 These guides provide methods for the design of anchors according to the CC Method, including anchorages with hanger reinforcement.

Failure modes of tensioned headed anchorsThe five possible failure modes in a nonreinforced

tensioned anchorage are illustrated in Fig. 1. These consist of the following: 1) steel anchor failure (Fig. 1(a)) limited by the steel resistance of the fastener; 2) pullout failure (Fig. 1(b)) due to a progressive crushing of the concrete over the anchor head; 3) concrete cone failure (Fig. 1(c)), where the concrete cone-shaped failure surface propagates from the head of the anchor; 4) blowout failure (Fig. 1(d)), which involves the blowing out of the concrete surface adjacent to the anchor head; and 5) splitting failure (Fig. 1(e)), charac-terized by the formation of cracks vertically along the length of the anchor(s).

Concrete cone failureThe CC Method is based on a model corresponding to the

formation of a concrete cone with the shape of a pyramid with a quadrangular base with an inclination from the surface of approximately 35 degrees (ACI 349-7614 considers 45 degrees). The characteristic resistance (5% value fractile)

240 ACI Structural Journal/March-April 2013

José Henriques is a PhD Student at the University of Coimbra, Coimbra, Portugal, where he received his civil engineering degree and his master’s degree in structures. His research interests include steel and composite structures.

José Maria Raposo is a Design Engineer at Struplano, Ltd., Coimbra, Portugal. He received his civil engineering degree from the Instituto Superior Técnico, Lisbon, Portugal, and his master’s degree in structures from the University of Coimbra.

Luís Simões da Silva is a Professor at the University of Coimbra and Head of the Steel and Composite Construction Technologies Group of the Institute for Sustainability and Innovation in Structural Engineering (ISISE). He received his civil engineering degree from the University of Coimbra and his MSc and PhD from Imperial College London, London, UK. His research interests include steel and composite structures.

Luís Costa Neves is an Assistant Professor at the University of Coimbra, where he received his civil engineering degree, his master’s degree in structures, and his PhD. His research interests include steel and composite structures.

of a headed anchor in tension (N0Rk,c) without the influence of

member edges or other adjacent fasteners, according to CEN/TS 1992-4,2 is

0 1.5, , (N)Rk c ck cube efN k f h= (1)

The resistance was found to be a function of the embed-ment depth hef (mm) and the concrete compressive strength fck,cube (N/mm2). The factor k varies as a function of the concrete state—cracked (8.5) or uncracked (11.9). In ACI 318-08,3 the characteristic resistance is calculated using the concrete compressive cylinder strength instead of the cube strength and the factor k is equal to 10 independent of the concrete state.

The other effects contributing to the resistance of the connection are taken into account by multiplying N0

Rk,c by a series of factors yi. According to CEN/TS 1992-4,2 the resis-tance of an anchor or a group of anchors is given by

0, , , , , , , (N)Rk c Rk c A N s N ec N re N ucr NN N= y y y y y (2)

where the factors yi correspond to the following effects:

• yA,N: factor for effects resulting from the proximity of other fasteners or the concrete free edges

,, 0

,

c NA N

c N

AA

y = (3)

• A0c,N = 9h2

ef : area of the concrete cone base (at the surface) corresponding to an individual fastener with a distance to other fasteners or to the concrete free edges such that these effects are negligible.

• Ac,N: area of the concrete cone base considering closely spaced anchors (superposition of cones) and/or concrete member edges.

• ys,N: factor accounting for the influence of proximate edges. This factor is relevant only when c < 1.5hef, where c (mm) is the smallest distance to the edges

, 0.7 0.3 11.5s N

ef

ch

y = + ≤ (4)

• yec,N: factor accounting for the eccentricity en (mm) between the resultant tensile force and the geometrical centroid of the tension-loaded anchors. In the case of eccentricity in two directions, the eccentricity factor should be calculated for each direction and both factors included in Eq. (2)

,

,

1 11 2

ec Nn

cr n

es

y = ≤+

(5)

• scr,n (mm): characteristic anchor spacing representing the minimum distance between anchors to form indi-vidual concrete cones (=3hef).

• yre,N: factor accounting for a local blowout of the concrete at the surface observed when hef < 100 mm (3.94 in.) and the grid surface reinforcement is closely spaced

Fig. 1—Failure modes in anchorages with headed anchors.


, 0.5 1200

efre N

hy = + ≤ (6)

• yucr,N: factor accounting for the effect of the concrete cracking within the anchorage zone. It equals 1.4 (1.25 according to ACI 318-083) for uncracked and 1.0 for cracked concrete at the serviceability limit state.

Contribution from steel hanger reinforcementBode and Hanenkamp22 verified experimentally that

adding hanger reinforcement leads to a considerable gain in strength and ductility. According to Rehm et al.,23 the most efficient reinforcement layout is stirrups or hangers (Fig. 2) close to the fastener, conveniently anchored, and involving any grid surface reinforcement. The guides1,2 present a methodology for the design of anchorages reinforced with hook-shaped reinforcing bars known as “hanger reinforce-ment” (Fig. 2). However, recent studies10,11,24 point out the fact that the contribution of this reinforcement is taken into account too conservatively.

The presence of hanger reinforcement prevents or delays the formation of a concrete cone, and therefore two new modes of failure may become dominant: 1) yielding of the steel hanger reinforcement (Fig. 2(a)); or 2) anchorage failure of the hanger reinforcement (Fig. 2(b)). In the first case, the steel yields after mobilizing the yield stress fsyk (N/mm2), corresponding to a characteristic load value of

, (N)Rk shr shr shr sykN n A f= (7)

where nshr is the number of vertical steel reinforcement bars; and Ashr (mm2) is the cross-sectional area of each reinforcing bar. In the second case, anchorage failure of the hanger reinforcement occurs when, for the actual anchorage length, the maximum bond of the hanger reinforcement is reached. This bond resistance can be predicted using the following expression given in CEN/TS 1992-42

1, , ,1, (N)shr hr i shr i bkn

iRk a

l d fN =

p= ∑

a(8)

where l1,hr,i (mm) is the anchorage length of reinforcing bar i inside the assumed concrete cone, measured verti-cally from the cone surface to the end of the vertical part of the reinforcing bar; dshr,i (mm) is the diameter of the reinforcing bar i; fbk (N/mm2) is the characteristic bond strength; and a is a factor that takes the effect of the shape of the reinforcing bar and the concrete confinement within the anchorage length into account. According to CEN/

TS 1992-42 for hooked bars, a should be taken equal to 0.7. In Eq. (8), the contribution of the bond and hook is consid-ered simultaneously; however, the hook effect is indepen-dent of the bond conditions and the anchorage length of the reinforcing bar l1,hr,i; therefore, as proposed by Kuhl-mann and Imminger,9 these two effects should be treated as summed contributions.

In ACI 318-08,3 the anchorage resistance of the hanger reinforcement is not evaluated; however, the hanger reinforcement may be considered effective if it developed on both sides of the breakout surface, accomplishing the detail requirements given in the code.

The characteristic bond strength, according to CEN/TS 1992-4,2 is obtained using the characteristic tensile resis-tance fctk (N/mm2) as follows

21 22.25 (N/mm )bk ctkf f= h h (9)

where h1 and h2 are factors that take into account the bond conditions and the diameter of the reinforcing bars, respec-tively. In all test specimens, both factors are equal to 1, as the bond conditions are assumed good and the diameter of the hanger reinforcement bars is smaller than 32 mm (1.26 in.).

Effect of concrete member thickness hThe concrete member thickness h (mm) could affect the

resistance to splitting failure. This failure mode may be disregarded if: 1) anchors are installed with edge distances greater than characteristic edge distances (ccr,sp [mm]); and 2) confining reinforcement is provided to control crack widths to 0.3 mm (0.012 in.). According to CEN/TS 1992-4,2 the resistance to splitting failure is determined similarly to Eq. (2) with inclusion of the factor yh,N.

2/32/3

,

2 efh N

min min

hhh h

y = ≤

(10)

where hmin is the minimum concrete member thickness equal to the sum of embedment depth hef (mm), the anchor head thickness th (mm), and the required concrete cover according to national regulations (ca [mm]).

EXPERIMENTAL INVESTIGATIONDescription of test specimens

The experimental work performed consisted of two groups of tests on tensioned headed anchors, tested in different periods and divided into five series each. The first group (G1) considered a total of 25 specimens, while in the second (G2), 15 specimens were tested. The test specimens’ geom-

Fig. 2—Failure modes for steel hanger reinforcement.


etry is schematically represented in Fig. 3 and 4, and the material properties are given in Table 1.

In all tests, design resistances associated with the following failure modes were considered: 1) steel anchor failure; 2) pullout failure; 3) steel hanger reinforcement bar failure; 4) anchorage failure of the steel hanger reinforcement; and 5) concrete cone failure—theoretical because this mode is prevented by the presence of reinforcement. Note that the failure modes are listed such that the associated predicted resistance decreases from 1 to 5.

Experimental tests in Group G1—The 25 tested specimens of this group were divided into five series (S1 to S5), where the geometrical properties—varied between each test series—consisted of the embedded depth hef of the headed anchor,

the number of anchors, and the edge distance c. Within each test series, a subdivision into five types of specimens (N1 to N5) was performed, varying the use of hanger reinforcement, the diameter of the hanger reinforcement dshr, and the use of top grid surface reinforcement. Table 2 summarizes all geometrical properties. In all test series except S4, only one anchor was installed. In Test Series S4, two anchors spaced at 200 mm (7.87 in.) were used. In Test Series S5, a rectan-gular block was considered to analyze the edge effect. Both S4 and S5 are particular cases of Test Series S2; therefore, the latter is used as a reference in the analysis of the results.

For all test specimens, the ratio of the anchor embedment depth hef to member thickness h was set at 0.5 to avoid any influence of the member thickness on the concrete failure

Fig. 3—Definition of test variables of each group.

Fig. 4—Geometry of concrete blocks.

Table 1—Material properties of Groups G1 and G2

Test series

ConcreteReinforcement (including

hanger reinforcement) Headed anchors Anchor plate

fcm,cub, N/mm2 (ksi)

fctm, N/mm2 (ksi)

Ecm, N/mm2 (ksi) × 103

fsyk, N/mm2 (ksi)

Es, N/mm2 (ksi) × 103

fsyk, N/mm2 (ksi)

Es, N/mm2 (ksi) × 103

fsyk, N/mm2 (ksi)

Es, N/mm2 (ksi) × 103

G1S129.5

(4.23)2.29

(0.33)30.38(4.41)

500 (72.52)

210 (30.46)

640 (92.82)

210 (30.46)

— —

G1S230.59(4.44)

2.40(0.35)

31.00(4.50)

G1S346.68(6.77)

3.43(0.50)

34.13(4.95)

G1S429.51(4.28)

2.32(0.48)

30.80(4.47)

G1S525.77(3.74)

2.04(0.44)

29.91(4.34)

G2S6 to G2S10

24.68 (3.58)

1.96 (0.28)

28.85 (4.18)

651 (94.42)

218.88 (31.75)

1008.33 (146.25)

213.31 (30.94)

355 (51.49)

210 (30.46)


modes.4 The minimum anchor embedment depth hef was taken as 150 mm (5.91 in.) because, according to the CEB Design Guide,1 this is the minimum value allowing the effi-ciency of the hanger reinforcement. The disposition of the hanger reinforcement consisted of two hooked reinforce-ment bars (four legs) (Fig. 3(a)). In the test specimens of Test Series S5, the hanger reinforcement was disposed in the longitudinal direction of the concrete block. The concrete cover for all types of reinforcement was 30 mm (1.81 in.). In Test Series S5, the concrete block geometry was rectangular (similar to Fig. 4(b)), adopting an edge distance c smaller than the cone radius at the surface (1.5hef) in one direction. In all other test series, the concrete block geometry was hexagonal (Fig. 4(a)) with edge distances higher than this value. Fastener diameters d were chosen so as to prevent their failure, and their head diameters dh were selected to avoid pullout failure.

Experimental tests in Group G2—The 15 tested speci-mens of this group were divided into five series (S6 to S10), where the geometrical properties—varied between each test series—consisted of the embedded depth hef and the member thickness h. Within each test series, a subdivision into three types of specimens (N6 to N8) was performed, varying the use of hanger reinforcement and the distance of the hanger reinforcement to the anchors. Table 3 summarizes all geometrical properties. Test Series S6 is a reference, as it represents the case where the member thickness should not affect the resistance.

In all test series, two closely spaced anchors welded to a plate were used (Fig. 3(b)). The ratio between member embedment depth hef and member thickness h was varied. The concrete block geometry was rectangular (Fig. 4(b)) and in the transverse direction (smaller dimension), the edge distance c was greater than 1.5hef. The hanger reinforcement used consisted of eight legs installed parallel to the trans-verse direction (Fig. 3(b)). The concrete cover for all types of reinforcement was 30 mm (1.81 in.). The anchor shaft and head diameters (d and dh) were chosen to avoid anchor steel failure and pullout failure, respectively.

Test layout and instrumentationA steel structure with a square shape composed of four

beams of the HE 140 B cross section was used to anchor the specimens to a reaction slab, incorporating four prestressed high-strength steel bars. The distance between the high-strength steel bars’ axis was 1 m (39.37 in.). The anchorages were tensioned by a hydraulic jack with a maximum stroke of 300 mm (11.81 in.) connected through a set of bars and plates to a 100 ton (220 kip) load cell. In all tests, an initial load step up to 5 tons (11 kips) with complete discharge with load control was applied to eliminate any slack, then the test continued with displacement control at a speed of 0.02 mm/s (7.87 × 10–4 in./s)—increased at later stages up to 0.10 mm/s (3.94 × 10–3 in./s). Linear variable differential transformers (LVDTs) were used to evaluate the displacements directly on the fastener and the concrete surface. In some tests, strain

Table 2—Geometric properties of Group G1

Series hef, mm (in.) d, mm (in.) dh, mm (in.) th, mm (in.) c, mm (in.) s, mm (in.) h, mm (in.) beq, mm (in.) lhex, mm (in.) l1,hr, mm (in.)

G1S1 150 (5.90) 25 (0.98) 45 (1.77) 20 (0.79) 625 (24.61) — 300 (11.81) 1250 (49.21) 520 (20.39) 77 (3.03)

G1S2 200 (7.87) 30 (1.18) 60 (2.36) 25 (0.98) 625 (24.61) — 400 (15.75) 1250 (49.21) 520 (20.39) 130 (5.12)

G1S3 260 (10.24) 40 (1.57) 75 (2.95) 35 (1.38) 625 (24.61) — 520 (20.47) 1250 (49.21) 520 (20.39) 195 (7.68)

G1S4 200 (7.87) 30 (1.18) 60 (2.36) 25 (0.98) 525 (20.67) 200 (7.87) 400 (15.75) 1250 (49.21) 520 (20.39) 130 (5.12)

G1S5* 200 (7.87) 30 (1.18) 60 (2.36) 25 (0.98) 200 (7.87) — 400 (15.75) 1250 (49.21) 400 (15.75) 130 (5.12)

Type Hanger reinforcement? No. of legs dshr, mm (in.) Top surface reinforcement? dsr, mm (in.) #sr, mm (in.)

N1 No — — No — —

N2

Yes 4

10 (0.394)No — —

N3 Yes 10 (0.394) 150 (5.90)

N412 (0.472)

No — —

N5 Yes 10 (0.394) 150 (5.90)*Concrete block is rectangular.

Table 3—Geometric properties of Group G2

Series d; dh, mm (in.) th, mm (in.) c; s, mm (in.) b; l, mm (in.) tap; bap; lap, mm (in.) hef, mm (in.) h, mm (in.)

G2S6

22 (0.87); 35 (1.38) 10 (0.39)370 (14.57); 110 (4.33)

850 (33.46); 1200 (47.24)

20 (0.79); 100 (3.94); 200 (7.87)

150 (5.91) 300 (11.81)

G2S7 200 (7.87) 300 (11.81)

G2S8 240 (9.44) 300 (11.81)

G2S9 160 (6.29) 200 (7.87)

G2S10 200 (7.87) 250 (9.84)

Type Hanger reinforcement? No. of legs dshr, mm (in.) shr Top surface reinforcement? dsr, mm (in.) #sr, mm (in.)

N6 No — — —

Yes 12 (0.472) 150 (5.90)N7 Yes8 10 (0.394)

≈0.0

N8 Yes 0.75hef


gauges were placed on the steel reinforcing bars, both on the hanger and grid surface steel reinforcement. More detailed information on the experimental procedure may be found in Reference 10.

EXPERIMENTAL RESULTS AND DISCUSSIONAll tests with steel hanger reinforcement failure involved

the anchorage of the steel hanger reinforcement, followed by the formation of a concrete cone (Fig. 5(a)). The pres-ence of hanger reinforcement leads to an initial crushing of the concrete at the head of the anchor. The concrete cone is then formed at a smaller embedment depth and the anchorage length of the hanger reinforcement is reduced. Subsequently, anchorage failure of the hanger reinforcement occurs and a smaller concrete cone is observed. Specimens without hanger reinforcement (Types N1 or N6) were associated with a pure concrete cone failure (Fig. 5(b)). In Group G2, due to the geometry of the concrete block and because of the higher ratio hef /h, a transverse crack was observed before failure.

Figures 6 and 7 summarize the results of both groups of tests in terms of force-displacement curves. These curves compare different test series and types. Tables 4 and 5 present the maximum loads obtained in Test Groups G1 and G2, respectively. Independent of the test group, series, and type, the following is common: the initial stiffness is similar; the onset of the radial cracks (Fig. 5(a)) corresponds to the maximum force achieved in the force-displacement curve; during the descending branch, radial cracks develop up to the formation of circumferential cracking involving the formation of the concrete cone (Fig. 5(b)); and, at this stage, the load stabilizes and the remaining resistance is provided by the friction between surfaces in the cracks and the grid surface reinforcement (when available). The influence of the test variables is discussed in the following.

Embedment depth hefIn both groups of tests, with the increase of the embed-

ment depth hef, an increase of the load capacity was verified.

Fig. 5—Failure modes in Test Group G1.

Fig. 6—Summary of force-deformation curves obtained in Test Group G1.

Fig. 7—Summary of force-deformation curves obtained in Test Group G2.


For Group G1, by comparing Test Series S2 with S1 and S3 with S1, the increase was, on average, approximately 46% and 161%, respectively. In Group G2, by comparing Test Series S7 with S6 and S8 with S6, the load capacity was increased by approximately 29% and 74%, respectively.

Edge effectIn Test Series S5 of Group G1, the reduction of the edge

distance produced a reduction of load capacity of approxi-mately 10% for the test specimen without hanger reinforce-ment (Type N1). When hanger reinforcement was used without grid surface reinforcement, an increase was observed. Again, the presence of hanger reinforcement modifies the failure mechanism to an anchorage failure of the hanger reinforce-ment, which is not affected by the edge distance.

Group effectIn Test Group G1 (Test Series S4), the use of two anchors

with a spacing smaller than 3hef demonstrates that the

Table 4—Maximum anchorage resistance obtained in tests of Test Group G1, kN (kips)

Test series

Test type S1 S2 S3 S4 S5

N1173.00(38.89)

268.50(60.36)

463.50(104.20)

299.30(67.29)

159.10(35.77)

N2186.40(41.90)

266.60(59.93)

476.60(107.14)

298.50(67.11)

204.70(46.02)

N3240.10(53.98)

337.50(75.87)

587.80(132.14)

299.20(67.26)

228.80(51.43)

N4186.10(41.84)

284.90(64.05)

539.40(121.62)

273.80(61.55)

209.60(47.12)

N5256.80(57.73)

349.40(78.55)

636.30(143.05)

408.30(91.79)

256.40(57.64)

Table 5—Maximum anchorage resistance obtained in tests of Test Group G2, kN (kips)

Test series

Test type S6 S7 S8 S9 S10

N6144.64(32.52)

189.90(42.69)

291.84(65.61)

147.81(33.23)

243.56(54.75)

N7226.18(50.85)

246.86(55.50)

322.76(72.56)

177.83(39.98)

227.33(51.10)

N8161.43(36.29)

234.46(52.71)

284.53(63.96)

172.78(38.84)

284.94(64.06)

Fig. 8—Effect of presence of hanger reinforcement on anchorage response.

increase in load capacity is not directly proportional to the number of anchors, confirming the effect of the superposi-tion stress fields. Because the same amount of reinforce-ment was used in specimens with one and two anchors, the increase in load capacity demonstrates that, although the failure mechanism involves anchorage failure of the hanger reinforcement, the concrete breakout resistance contributes to the final resistance.

Hanger reinforcementThe average increase of resistance due to the contribution

of the hanger reinforcement is higher in Test Group G2, as eight legs are used instead of four (30 to 10%). Addition-ally, an increase in deformation capacity is also noticed in the force-displacement curves (Fig. 6 and 7). A higher displacement of the anchors is observed because the tensile strain is controlled by a compressive force arising between the anchor head and the steel hanger reinforcement that produces a compressive strain (Fig. 8(a)). Experimentally, this compression was observed by comparing the relative displacements between the concrete surface near the anchor and the anchor itself in reinforced tests (Fig. 8(b)). This effect may be significant and may lead to the onset of some pullout failure, as observed in some tests.

Position of hanger reinforcementIn Test Group G2, the results show that the smaller the

distance from the anchors to the hanger reinforcement, the higher the resistance of the anchorage. The average increase of load capacity in test Type N7 is approximately 30%, while in Type N8, it is approximately 13%.

Grid surface reinforcementTest Group G1 demonstrates that grid surface reinforce-

ment increases the efficiency of the hanger reinforcement. The average increase of resistance, from test Type N1 to Types N3 and N5, was approximately 44% and 61%, respec-tively. Without grid surface reinforcement, the increase was much smaller—8% and 11% (Types N2 to N1 and N4 to N1, respectively). The activation of the grid surface reinforce-ment is noted in the force-displacement curve as the load capacity recovers after the first peak (refer to Fig. 6(a)) (Test Specimens S3N3 and S3N5). This is due to the fact that the grid surface reinforcement needs some displacement to be mobilized and to start to contribute to the overall resis-tance. The increase of ductility is also observed.


Embedment-depth-to-member-thickness ratio hef /hThe comparison between Test Series S7 and S10

(Group G2) demonstrates that the higher the ratio hef/h, the higher the resistance of anchorage. This may be justified by the fact that the bottom grid surface reinforcement is acti-vated to a higher level in the case of a higher hef/h ratio. However, the limited number of tests does not allow for conclusive results.

COMPARISON OF PREDICTIONS AND EXPERIMENTAL RESULTS

Resistance assessmentIn this section, the performed comparison uses the analyt-

ical formulation proposed by CEN/TS 1992-4.2 The failure loads, as governed by yielding of the hanger reinforcement (Eq. (7)) and anchorage failure of the hanger reinforce-ment (Eq. (8)), were computed using the mean values of the material properties instead of the characteristic values (refer to Table 1).

Tables 6 and 7 present the analytical predictions and the ratio between the Min(NRm,shr, NRm,a) and the test result Nu. It is observed that the analytical values are clearly smaller than the experimental ones; however, for Specimens S9N7 and S10N7 of Group G2, this was not the case. The result of Specimen S10N7 is justified by an invalid test result as the loading system failed. Note that for Test Group G2, the predictions were calculated considering the cracked state of the concrete, as a transverse crack formed well before the failure load.

As far as the steel hanger reinforcement failure is concerned, it may be concluded that the available analytical formulations1,2,7 are quite conservative because the total resistance of the system is not exclusively provided by the steel in tension Min(NRm,shr, NRm,a). A summed contribu-tion of the concrete in tension and the hanger reinforcement exists but is not accounted for in the calculations.

Influence of test parameters on experimental results

In Test Group G1, the concrete resistance obtained for each test series was different (Table 1). To have an equivalent comparison, the experimental results were normalized with respect to the concrete resistance fcm (N/mm2) of the refer-ence test series (S2). The normalization varied according to the use of hanger reinforcement. Without hanger reinforce-ment, the prediction of the resistance has the proportion Nu ~ fcm

0.5. This leads to the normalized resistance as follows

1/2, 2 (N)cm S

u ucm

fN N

f

=′ (11)

When using hanger reinforcement, the results are expected to be governed by the anchorage of the hanger reinforcement. Thus, according to Eq. (8) and (9), the resistance depends on mean bond strength fbm (N/mm2) and mean tensile strength fctm (N/mm2) which, according to CEN/TS 1992-4,2 has the proportion fctm~fcm

2/3.

2/3, 2 (N)cm S

u ucm

fN N

f

=′ (12)

For Test Group G2, no normalization is performed on the experimental results.

Effect of embedment depth hef—The variation of hef results in a change in the anchorage length l1,hr,i (Eq. (8)); therefore, the resistance of both types of test specimens—with and without hanger reinforcement—is expected to be affected. Figure 9(a) shows the ratios of normalized resis-tances (N′u,Si/Nu,S2 or S6) as a function of the ratio between embedment depths (hef,Si/hef,S2 or S6)1.5. The power 1.5 is

Table 6—Prediction loads for Test Group G1

S1 S2 S3 S4 S5

NRk,c, kN (kips) N1 165.25 (37.15) 260.62 (58.59) 477.20 (107.28) 341.31 (76.73) 143.53 (32.27)

NRm,shr, kN (kips)N2 and N3 157.08 (35.31) 157.08 (35.31) 157.08 (35.31) 157.08 (35.31) 157.08 (35.31)

N4 and N5 226.19 (50.85) 226.19 (50.85) 226.19 (50.85) 226.19 (50.85) 226.19 (50.85)

NRm,a, kN (kips)N2 and N3 71.22 (16.01) 126.02 (28.33) 270.16 (60.73) 121.82 (27.39) 107.12 (24.08)

N4 and N5 85.47 (19.21) 151.23 (33.00) 324.19 (72.88) 146.19 (32.86) 125.54 (28.90)

min(NRm,shr, NRm,a)/Nu,test

N2 0.38 0.47 0.33 0.41 0.52

N3 0.30 0.37 0.27 0.41 0.47

N4 0.46 0.53 0.42 0.53 0.61

N5 0.33 0.43 0.36 0.36 0.50

Table 7—Prediction loads for Test Group G2

S6 S7 S8 S9 S10

NRk,c, kN (kips) N6 96.54 (21.70) 141.33 (31.77) 180.99 (40.69) 105.05 (23.62) 141.33 (31.77)

NRm,shr, kN (kips) N7 and N8 409.03 (91.95) 409.03 (91.95) 409.03 (91.95) 409.03 (91.95) 409.03 (91.95)

NRm,a, kN (kips)N7 158.34 (35.60) 237.50 (53.39) 308.76 (69.41) 182.09 (40.93) 245.45 (55.17)

N8 71.25 (16.02) 102.92 (23.14) 158.34 (35.59) 79.17 (17.80) 110.84 (24.92)

min(NRm,shr, NRm,a)/Nu,test

N7 0.70 0.96 0.96 1.02 1.08

N8 0.44 0.44 0.56 0.46 0.39


used, as in Eq. (1), the resistance is proportional to hef1.5.

The ratios were separately calculated for each group of tests. The results show a linear correlation between resis-tance and hef

1.5. A worse approximation was obtained for the tests of Group G2, which may be justified by the fact that the ratio between embedment depth hef and member thick-ness h was varied simultaneously. For the tests with hanger reinforcement, failure was linked to the anchorage failure of the hanger reinforcement. Therefore, according to Eq. (8), the resistance follows a linear variation with l1,i. Figure 9(b) illustrates the ratio between resistances as a function of the ratio between anchorage lengths (l1,hr,Si/l1,hr,S2 or S6). A similar trend is observed as for the embedment depth hef.

Effect of proximity of other headed anchors and concrete member free edges—Using the variable y = yA,Nys,N, the influence of both effects is analyzed simultaneously. The normalized values of the failure loads N′u,Si/Nu,S2 should follow the normalized variable ySi/yS2. Figure 10 presents the evolution of the normalized resistance and normalized variable for all types of specimens in Test Series S2, S4, and S5 of Test Group G1. For the unreinforced tests (Type N1), Fig. 10 shows a reasonable agreement with the concrete capacity method (where NRk,c ~ yi). According to the CEB Design Guide1 and CEN/TS 1992-4,2 the failure load involving the steel hanger reinforcement does not depend on the effects accounted for in the factor y as long as the failure load of the concrete cone NRk,c is lower than the failure load of the anchorage of the steel hanger reinforcement NRd,a. This should correspond to ratios N′u,Si/Nu,S2 equal to 1. As shown in Fig. 10, the results from the tests show that N′u,Si/Nu,S2 increase with y, suggesting that even when failure is linked to the hanger reinforcement, the shape of the concrete cone affects the resistance of the anchorage.

Effect of hanger reinforcement—The values presented in Tables 6 and 7 show that, for the majority of the cases, the ratios between the analytical predictions and test results are clearly smaller than 1. According to the guides,1-3 the hanger reinforcement to be used in an anchorage should at least resist the pure concrete cone load capacity. Consequently, in the cases where this condition is not guaranteed, the hanger reinforcement is considered ineffective. However, these experimental investigations demonstrate that even in the cases of low hanger reinforcement resistance—either limited by anchorage or steel failure—there is a combined contribu-tion (hanger reinforcement and concrete) to the anchorage resistance, which increases its load capacity in comparison to the cases with a pure concrete cone.

In Test Group G1, the diameter of the hanger reinforce-ment was varied and its influence is shown in Fig. 11, where the chart presents the ratio between the test results of different diameters. From these results, it may be concluded that there is an increase of the failure load with the increase of dshr, but at a smaller rate than d12/d10 (1.20). This is most likely due to the fact that the steel hanger reinforcement and its anchorage are not the only contributions for the overall resistance, with contribution from other components inde-

Fig. 9—Effect of anchorage embedment depth hef.

Fig. 10—Effect of factor y on test specimens of Test Group G1.

Fig. 11—Effect of steel hanger reinforcement diameter on different series of Test Group G1 (Specimens N2vsN4 and N3vsN5). (Note: 1 mm = 0.03937 in.)


Fig. 12—Effect of providing grid surface reinforcement on test specimens of Test Group G1.

Fig. 13—Influence of parameters only varied in Test Group G2.

hanger reinforcement is only activated upon the formation of the concrete cone.

Effect of embedment-depth-to-member-thickness ratio hef/h—This parameter is taken into account by adding the factor presented in Eq. (9) to Eq. (2). The evaluation of the test results with this factor is compared in Fig. 13(b) for Test Group G2. To isolate the variable hef /h, the test values were normalized using Test Series S6 and the proportion hef

1.5 as follows.

1.5

, 6, ,

,

(N)ef Su Si u Si

ef si

hN N

h

=′

(13)

The ratio of resistances (Nu,Si/Nu,S6) does not follow the same trend as the ratio of factors (yh,si/yh,S6), which indi-cates that the specimens’ resistance is not significantly influ-enced by the hef /h ratio. In fact, according to the CEB Design Guide1 and CEN/TS 1992-4,2 this factor may be disregarded if the amount of reinforcement in the concrete member takes the splitting forces generated by the anchorage in tension into account. In addition, the limited number of tests and the fact that this variable was not completely isolated within all tests limit the conclusions.

FURTHER RESEARCHIn the future, more accurate analytical formulas to

account for the resistance of the hanger reinforcement and the correct interaction between relevant components are required. To fulfill this objective, some additional tests could be performed, considering the variation from the anchor to the steel reinforcement, as performed in Test

pendent of dshr, such as the concrete cone resistance and grid surface reinforcement.

Effect of grid surface reinforcement—Figure 12(a) shows the relative increase of resistance verified in Test Group G1. Only Test Series S1 to S3 are used. The values are normalized to the specimens with hanger reinforcement and without grid surface reinforcement (Types N2 and N4). It can be observed that there is always a resistance increase—up to 38%—and the relative gain is more significant for smaller embedment depths. Figure 12(b) shows the absolute resistances. Here, contrary to the relative values, for higher embedment depth, the absolute resistance gains are higher as well. This result may be justified by the fact that with higher embedment depth hef, the larger the concrete cone and, therefore, the greater the percentage of grid surface reinforcement mobi-lized to take the transversal component of the applied load, which develops upon the formation of the concrete cone.

Effect of position of hanger reinforcement—The posi-tion of the hanger reinforcement affects the resistance to anchorage failure as the anchorage length l1,hr,i used in Eq. (8) decreases with the increase of the distance between the anchorage and hanger reinforcement. In Test Group G2, test Type N8 considered the installation distant from the anchors; in all other test specimens with hanger reinforce-ment (Groups G1 and G2), this was installed adjacent to the anchor (refer to Fig. 4(b) and Table 3). In Fig. 13(a), the relative increase of resistance is shown as (Nu,i/Nu,N6). Although the hanger reinforcement is installed distant from the anchor (smaller anchorage length), an increase of resis-tance is still obtained. According to the current approach, the anchorage failure defines the failure load and, therefore, the resistance should be equivalent to the concrete cone, as


Group G2. Due to the limitations often related to experi-mental research, numerical work should also be performed, providing a better comprehension of the phenomena and a larger amount of results.

CONCLUSIONSThe performed experimental tests demonstrated that the

behavior of reinforced anchorages is governed by a depen-dent set of components and variables: concrete tensile resis-tance, bond of the steel hanger reinforcement in the concrete cone, the hook effect of the hanger reinforcement, the posi-tion of the hanger reinforcement, and the confining effect of the grid surface reinforcement. From the comparison of the test results with the analytical predictions (according to the design guides1-3), it was concluded that the current methods that take into account the use of steel hanger reinforcement are quite conservative.

The presented results revealed that the evaluation of the resistance of a reinforced anchorage cannot rely only on the capacity of the steel hanger reinforcement or its anchorage capacity, as is currently done. The contribution of the concrete resistance is clear and should be taken into account. It was also found that the steel hanger reinforcement provides a resistance enhancement when not associated with grid surface reinforcement and when its resistance is smaller than the pure concrete cone. Thus, in a future improvement of the current model, a summed contribution of concrete and hanger reinforcement should be considered.

Also, the test results suggest that the hook effect is inde-pendent of the embedment depth or anchorage length l1,hr; therefore, the expression proposed by the guides1,2 for computing NRk,a seems inadequate because it considers this effect by multiplying the resistance value by a factor and not adding it to the resistance. This is in agreement with the study performed by Kuhlmann and Imminger,9 which stated that these effects should be considered independently.

Regarding the concrete member thickness, the limited number of results did not allow for the attainment of conclusive answers; however, the test results indicate that the member thickness effect is minimized by the concrete member reinforcement.

REFERENCES1. Comité Euro-International du Béton (CEB), “Design of Fastenings in

Concrete,” Thomas Telford, Lausanne, Switzerland, 1997, 92 pp.2. CEN/TS 1992-4, “CEN Technical Specification (TS): Design of

Fastenings for Use in Concrete,” Final Draft, European Organization for Standardization (CEN), Brussels, Belgium, 2009, 166 pp.

3. ACI Committee 318, “Building Code Requirements for Structural Concrete (ACI 318-08) and Commentary,” American Concrete Institute, Farmington Hills, MI, 2008, 473 pp.

4. Comité Euro-International du Béton (CEB), “Fastenings to Concrete and Masonry Structures,” State of the Art Report, Thomas Telford, Laus-anne, Switzerland, 1994, 83 pp.

5. Eligehausen, R.; Hofacker, I.; and Lettow, S., “Fastenings Technique—Current Status and Future Trends,” International Symposium on Connec-

tions between Steel and Concrete, R. Eligehausen, ed., Stuttgart, Germany, 2001, pp. 11-27.

6. Eligehausen, R.; Malée, R.; and Silva, J. F., Anchorage in Concrete Construction, Ernst & Sohn, Berlin, Germany, 2006, 391 pp.

7. Farrow, C. B., and Klingner, R. E., “Tensile Capacity of Anchors with Partial or Overlapping Failure Surfaces: Evaluation of Existing Formulas on an LRFD Basis,” ACI Structural Journal, V. 92, No. 6, Nov.-Dec. 1995, pp. 698-710.

8. Fuchs, W.; Eligehausen, R.; and Breen, J. E., “Concrete Capacity Design: Approach for Fastenings to Concrete,” ACI Structural Journal, V. 92, No. 1, Jan.-Feb. 1995, pp. 73-94.

9. Kuhlmann, U., and Imminger, T., “Ankerplatten und Einbaudetails zur Kraftüberleitung im Stahlbau,” Schlussbericht, Forschungsvorhaben im Auftrag des Deutschen Ausschusses für Stahlbau (DASt), Institut für Konstruktion und Entwurf, Stahl, Holz und Verbundbau, Universität Stutt-gart, Stuttgart, Germany, 2003, 187 pp. (in German)

10. Raposo, J. M., “Reforço com armaduras em ligações com chumba-douros de cabeça à tracção,” Tese de Mestrado, Departamento de Engen-haria Civil, Universidade de Coimbra, Coimbra, Portugal, 2006, 217 pp. (in Portuguese)

11. Rölle, L., “Konzentrierte Lasteinleitung mittels Kopfbolzendübel in Stahlbetonstützen. Diplomarbeit,” Institut für Konstruktion und Entwurf, Universität Stuttgart, Stuttgart, Germany, 2003, 137 pp. (in German)

12. Stark, J., and Hordijk, D. A., “Where Structural Steel and Concrete Meet. International,” International Symposium on Connections between Steel and Concrete, R. Eligehausen, ed., Stuttgart, Germany, 2001, pp. 1-11.

13. Raposo, J. M.; Neves, L. F. C.; and Simões da Silva, L., “Comporta-mento de chumbadouros à tracção ancorados em peças de betão,” Actas da Conferência Métodos Numéricos en Ingeniería 2005, J. L. Pérez Aparício, J. C. Sá, R. Delgado, R. Gallego, J. Martins, M. Pasadas, and A. R. Ferran, eds., SEMNI-APMTAC, Granada, Spain, 2005, 17 pp. (in Portuguese)

14. ACI Committee 349, “Code Requirements for Nuclear Safety Related Concrete Structures (ACI 349-76),” American Concrete Institute, Farm-ington Hills, MI, 1976, 72 pp.

15. European Organisation for Technical Approvals (EOTA), “Guideline for European Technical Approval of Anchors (Metal Anchors) for Use in Concrete,” Brussels, Belgium, 1994, 166 pp.

16. Hilti Corporation, Fastening Technology Manual, Schaan, Liechten-stein, 2005, 384 pp.

17. DIBt, “Bemessungsverfahren für Dübel zur Verankerung in Beton,” Deutsches Institut für Bautechnik, Berlin, Germany, 1993. (in German)

18. SIA 179, “Befestigungen in Beton und Mauerwerk,” Schweizeri-scher Ingenieur und Architekten-Verein, Zürich, Switzerland, 1998, 64 pp. (in German)

19. EN 1992-1-1, “Eurocode 2: Design of Concrete Structures—Part 1.1: General Rules and Rules for Buildings,” European Committee for Standard-ization (CEN), Brussels, Belgium, 2004, 225 pp.

20. EN 1993-1-1, “Eurocode 3: Design of Steel Structures—Part 1.1: General Rules and Rules for Buildings,” European Committee for Standard-ization (CEN), Brussels, Belgium, 2004, 91 pp.

21. EN 1994-1-1, “Eurocode 4: Design of Composite Steel and Concrete Structures—Part 1.1: General Rules and Rules for Buildings,” European Committee for Standardization (CEN), Brussels, Belgium, 2004, 118 pp.

22. Bode, H., and Hanenkamp, W., “Zur Tragfähigkeit von Kopfbolzen bei Zugbeanspruchung (Bearing Capacity of Headed Studs under Tension),” Bauingenieur, V. 60, 1985, pp. 361-367. (in German)

23. Rehm, G.; Schlaich, J.; Schäfer, K.; and Eligehausen, R., “Fritz-Leon-hardt-Kolloquium. Forschungskolloquium des Deutschen Ausschusses für Stahlbeton,” Beton und Stahlbetonbau, No. 6 and 7, 1985, pp. 156-161 and 190-194.

24. Nogueira, J. B., “Estudo do comportamento de chumbadouros pré-instatalados de cabeça submetidos à tracção,” Tese de Mestrado, Departa-mento de Engenharia Civil, Universidade de Coimbra, Coimbra, Portugal, 2010, 103 pp. (in Portuguese)


NOTES:

aaa

Documents

design of steel

behavior of steel

behavior of headed anchors

plain concrete hanger

concrete failure modes

effect of hanger reinforcement

tensile resistance of

experimental tests