Engineering Structures 100 (2015) 645–655

Contents lists available at ScienceDirect

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

Design of post-installed reinforcing bars as end anchorage or as bonded

anchor
Christoph Mahrenholtz ⇑, Rolf Eligehausen, Hans-Wolf Reinhardt
University of Stuttgart, Pfaffenwaldring 4, 70569 Stuttgart, Germany

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

Article history: Two different design methods for anchorages with post-installed reinforcing bars are available:
Received 11 February 2015 Anchorages can be designed either as end anchorage equivalent to cast-in reinforcing bars or as bonded
Revised 8 June 2015 anchors. The two design methods are very different in their fundamental approach and range of applica-
Accepted 15 June 2015
tion. This contribution presents both design methods including the technical background and addresses
Available online 7 July 2015
the qualification of systems for post-installed reinforcing bars. The interrelations between material
strength parameters required for the design are systematically illustrated. The design of anchorages
Keywords:
according to both design methods will be explained and their outcome compared.
Reinforcing bar
Bonded anchor
Ó 2015 Elsevier Ltd. All rights reserved.
End anchorage
Bond
Post-installation
Failure mode
Anchorage

1. Introduction design of post-installed anchors can be carried out according to

CEN/TS 1992-4 [4], in future EN 1992-4 (Eurocode 2 Part 4), or
Connections between reinforced concrete elements, such as ACI 318 Appendix D [3]. The following discussion only refers to
foundation-to-column connections (Fig. 1a), represent discontinu- the European regulations which, however, are to a great extend
ity regions and are designed using truss models [1]. Here, the similar to the regulations in the US.
starter bars of the element concreted in a second step are cast in The approach to design end anchorages of post-installed rein-
the previously constructed concrete element, for which the forcing bars as bonded anchors was studied in the past and proven
anchorage length of the reinforcing bars is calculated in Europe feasible for specific cases [e.g. 5,6]. However, the requirements
or in the US according to the structural concrete design codes valid for the anchor design method are not always met in case of
EN 1992-1 [2] (Eurocode 2 Part 1) or ACI 318 Section 12 [3], anchorages with post-installed reinforcing bars, and the anchor
respectively. For the sake of simplicity, it is in general assumed that design method then potentially results in unsafe design solutions.
the anchorage length ‘b of the reinforcing bars loaded in tension Furthermore, the coexistence of two different qualification regula-
begins at the concrete joint of the two elements (Fig. 1b). tions and design methods may lead to significantly different solu-
Reinforced concrete elements may also be connected using tions which is not plausible.
post-installed reinforcing bars (Fig. 1c). The design of these con- While the components used for structural concrete are in gen-
nections according to EN 1992-1 or ACI 318 is possible if the eral codified according to relevant standards, post-installed rein-
post-installed reinforcing bar system has been qualified. forcing bars and anchors are not. The therefore required
Post-installed reinforcing bars are increasingly popular, e.g. for qualification according to EOTA TR 023 [7] and ETAG 001 [8],
the retrofit of reinforced concrete structures or for the optimiza- respectively, results in European Technical Approvals or (since
tion of the construction processes. the Construction Product Directive (CPD) [9] has been replaced
Reinforcing bars are also used as steel elements for bonded by the Construction Product Regulation (CPR) [10] in June 2013)
anchors provided the anchoring system has been qualified. The European Technical Assessments, both of which abbreviated as
ETAs. In this context, European Approval Guidelines ETAG 001
and EOTA TR 023 are also currently converted into the European
⇑ Corresponding author at: Buesingstr. 6, 12161 Berlin, Germany. Tel.: +49 172
Assessment Documents EADs. The substitute for EOTA TR 023 is
6577516.
E-mail address: [email protected] (C. Mahrenholtz).
currently drafted [11] and expected to be published soon. The

http://dx.doi.org/10.1016/j.engstruct.2015.06.028
0141-0296/Ó 2015 Elsevier Ltd. All rights reserved.
646 C. Mahrenholtz et al. / Engineering Structures 100 (2015) 645–655

(a) (b) Column (c)

Starter bar
Concrete-
joint

ℓb Anchorage
φ

Foundation

Fig. 1. Column-to-foundation connection: (a) with cast-in starter bars (after concreting); (b) scheme; (c) with post-installed starter bars (before concreting).

content basically remains identical with some smaller modifica- anchorages, starter bars represent ties which form a framework
tions, e.g. with regard to the required spacing and concrete cover. together with global struts (Fig. 3b). For the design of reinforced
The most important amendment is the assessment of the resis- concrete structures, it is assumed that reinforcing bars do not
tance of the post-installation system to fire. transfer shear loads for which shear reinforcement is required. In
The aim of this contribution is to explain the design of connec- this contribution, the verification of the struts is not addressed.
tions with post-installed reinforcing bars according to the conven- Furthermore, it is assumed that the concrete surface at a joint
tional structural concrete design rules for end anchorages, and between two reinforced concrete elements is roughened to ensure
according to the design rules for bonded anchors. For this reason, a reliable transmission of shear forces.
the design method for end anchorages and bonded anchors accord- When designing end anchorages according to EN 1992-1
ing to EN 1992-1 and CEN/TS 1992-4, respectively, are discussed (Eurocode 2 Part 1), sufficient safety margin is required against
and compared. The various concepts and key terms are laid out steel failure due to exceedance of the steel yield strength (yielding,
in Fig. 2. Note that for clarity, in this contribution strengths are Fig. 3c), bond failure due to exceedance of the bond strength (pull-
indicated with the unit MPa and stresses with the unit N/mm2, out, Fig. 3d), and splitting due to exceedance of the concrete tensile
and ultimate strengths in the sense of forces given in kN are strength (cracking of concrete cover, Fig. 3e). The verification is not
denominated as capacities. carried out for each failure mode separately because the provisions
require an anchorage length which ensures adequate safety margin
2. Explanation of both design methods against all failure modes. In the following, the rules for the deter-
mination of the anchorage length of tensioned reinforcement bars
2.1. Design of end anchorages according to EN 1992-1 [2] (Eurocode 2 according to EN 1992-1 [2] are explained. The coefficients which
Part 1) are irrelevant for the comparison with bonded anchors in
Section 3 are omitted. The design value of bond strength is:
In reinforced concrete structures, reinforcing bars carry the ten- f bd ¼ 2:25  g1  g2  f ctd ð1Þ
sile forces developing in the elements. At the end of the reinforcing
bar, the tension load is transferred via bond into the concrete. At where fctd is the design concrete tensile strength, the factor g1
overlap splices, the bond forces are in equilibrium with local struts takes into account the position of the reinforcement during
developing from an adjoining reinforcing bar (Fig. 3a). At end concreting (g1 = 1.0 or g1 = 0.7 for good or poor bond conditions,

Anchorages in concrete with bond as load transfer mechanism

Cast-in Post-installed Post-installed anchor,

reinforcing bar reinforcing bar e.g. reinforcing bar

Qualification – ETA ETA

(all components codified) (basis: EOTA TR 023) (basis: ETAG 001)

Design method End anchorage Bonded anchor

CEN/TS 1992-4
EN 1992-1
Design standard in future EN 1992-4
(Eurocode 2 Part 1)
(Eurocode 2 Part 4)

Fig. 2. Methods und standards for the design of anchorages in concrete.

 C. Mahrenholtz et al. / Engineering Structures 100 (2015) 645–655 647

(a) (b) (c) (d) (e)

Fig. 3. Transfer of bond forces (a) at splices via local struts and (b) at end anchorages via global struts [12,13]; failure modes of an end anchorage; (c) steel failure (yielding);
(d) Bond failure (pullout); (e) Splitting.

respectively) and the factor g2 the diameter of the reinforcing bar strength fbm of cast-in reinforcing bars with short bond length
(g2 = (132 – / [mm])/100 P 1.0). The design bond strength includes (5/ 6 ‘v 6 10/) and large concrete cover in concrete C20/25 is
a safety factor cc = 1.5 for concrete failure. It is valid for reinforcing about 10 MPa and the corresponding slip s(fbm) is less than 0.1/.
bars with minimum concrete cover (c = 1/) and minimum spacing The bond strength of post-installed reinforcing bars is influenced
of adjacent reinforcing bars (a = 1/). Under these conditions by the properties of the mortar, which is product dependent and
anchorages fail due to splitting of the concrete. The basic value of therefore must be evaluated by approval tests on the basis of
the required anchorage length ‘b,rqd is: EOTA TR 23 [7]. For the determination of bond stress-slip curves
(s–s curves), pullout tests on reinforcing bars post-installed in con-
‘b;rqd ¼ /  rsd =ð4  f bd Þ ð2Þ
crete C20/25 and C50/60 with short bond length (5/ 6 ‘v 6 10/)
where / is the diameter of the reinforcing bar and rsd the design and large concrete cover are carried out. Currently, qualification
steel stress at the beginning of the anchorage. The design anchorage according to EOTA TR 23 requires up to 11 test series to assess
length is: the influence of the relevant parameters on the bond performance.
Note that the evaluation of pullout tests on reinforcing bars takes
‘bd ¼ a1  a2  ‘b;rqd P ‘b;min ð3Þ
into account the influence of the related rib area fR and compres-
where a1 and a2 take into account the type of anchorage (a1 = 1.0 sive strength fc of the tested reinforcing bars and concrete because
for straight bars) and the concrete cover (a2 = 1  0.15(cd  /)// these significantly influence the bond strength of cast-in reinforc-
for straight bars, where 0.7 6 a2 6 1.0), respectively. The term cd ing bars [14]: For the normalization, the tested bond strength is
is defined as the minimum of the concrete cover and half of the multiplied by the factors (0.08/fR)0.4 and (20 MPa/fc)0.5 where the
clear spacing between adjacent reinforcing bars. The minimum concrete compressive strength fc is determined using cylinders
anchorage length ‘b,min is equivalent to the maximum of 0.3‘b,req, with a diameter of 150 mm and a length of 300 mm. Equations
10/ and 100 mm. The coefficient a2 takes into account that the to calculate the related rib area fR, i.e. the ratio of projected rib area
bond strength increases with increasing concrete cover and there- and diameter of reinforcing bar, are provided in ISO 15630-1 [15].
fore reduces the required anchorage length. For a large concrete On the basis of the reference values, a function of the required
cover (cd P 3/) a2 = 0.7 applies which is equivalent to an increase mean bond strengths fbm,req was defined (Fig. 5a). The mean bond
of the design value of bond strength by a factor 1/a2 = 1/0.7 = 1.4. strength fbm = sRm determined by pullout tests on post-installed
The anchorage length according to Eq. (3) allows the anchoring reinforcing bars with large concrete cover, short bond length (usu-
of the design tensile force Td: ally ‘v = 7/) and confined test setup in uncracked concrete (Fig. 5b)
have to achieve the required mean bond strengths fbm,req in order to
T d ¼ rsd  p  /2 =4 6 f yd  p  /2 =4 ð4Þ
demonstrate equivalency. The value for the required mean bond
The design of end anchorages of post-installed reinforcing bars strength fbm,req = 10 MPa for concrete C20/25 may also be deduced
can be carried out as for cast-in reinforcing bars if the as follows: The design bond strength according to EN 1992-1 [2] is
post-installed reinforcing bars perform at least as good as cast-in fbd = 2.3 MPa for concrete C20/25. Assuming that the characteristic
reinforcing bars. This refers to both, the bond strength fb = sR, and bond strength is 0.75 times the mean value [16] and taking the
the corresponding slip s(fb) (Fig. 4a). To determine reference values partial safety factor cc = 1.5 into account one obtains
for post-installed reinforcing bars, pullout tests on cast-in reinforc- 2.3/0.75  1.5 = 4.6 MPa. This value applies to the minimum con-
ing bars were analyzed [12]. It was found that the mean bond crete cover, cracked concrete and standard anchorage lengths
Bond strength τR = fb [MPa]

14 14
Large concrete cover
Bond stress τ [N/mm ]

12 τR = fb 12
2

10 10
ℓv = 10φ
8 s(fb) 8 Cast-in
6 6 φ = 8 mm
φ = 16 mm φ = 12 mm
4 4 φ = 16 mm
ℓv = 160 mm
2 Small concrete cover Post-installed 2 φ = 20 mm
(c = 35 mm) Cast-in φ = 25 mm
0 0
0 1 2 3 4 5 6 7 8 9 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
Slip s [mm] Slip at ultimate bond stress s(fb) [mm]
(a) (b)
Fig. 4. Example (reinforcing bar B500 in concrete C20/25): (a) Bond stress–slip curve of cast-in and post-installed reinforcing bars for small and large concrete cover; (b)
correlation of bond strengths and corresponding slip of cast-in reinforcing bars [12].
648 C. Mahrenholtz et al. / Engineering Structures 100 (2015) 645–655

25

Mean bond strength required fbm,req [MPa]

fbm tested Example 2:

Mean bond strength tested fbm [MPa]

Equivalence shown up Load cell
Interpolation
20 to concrete C50/60
fbm,req = 10 / 202/3 · fck2/3
Displacement
15
transducer

10
Example 1:
Equivalence shown only up
5
to concrete C25/30 and bond
strength fbd is limited to 2.7 MPa ℓv
0
C12/15 C16/20 C20/25 C25/30 C30/37 C35/45 C40/50 C45/55 C50/60
1.6 2.0 2.3 2.7 3.0 3.4 3.7 4.0 4.3
Design value of bond strength fbd [MPa]
(a) (b)
Fig. 5. (a) Required mean bond strength of post-installed reinforcing bars according to EOTA TR 023 [7]; (b) pullout test with short bond length and confined test setup
according to ETAG 001 [8].

which are required to develop the yield stress. The pullout tests as smaller stress peaks and splitting forces [19]. However, further
part of approval procedures are carried out with large concrete research is needed in this respect.
cover and short bond length and confined test setup in uncracked For the assessment of the results of the pullout tests, the bond
concrete (Fig. 5a). As pointed out above, the bond strength for a strength is taken into account only up to the slip d1 (/ < 25 mm:
large concrete cover (cd P 3/) may be taken as 1.4 times the value d1 = 1.5 mm; 25 mm 6 / 6 40 mm: d1 = 2.0 mm; / > 40 mm:
valid for the minimum concrete cover (cd = 1/). The bond strength d1 = 3.0 mm) to ensure a slip behavior similar to cast-in reinforcing
is reduced by about 10% for standard anchorage lengths if com- bars [12]. Post-installed reinforcing bars installed with a high
pared to short bond lengths [12]. Since the test are carried out strength mortar may show a much stiffer bond behavior than
under confined test setup conditions, the stress state equals to that cast-in reinforcing bars. This may negatively influence the failure
in reinforced concrete elements in which the tensioned end load of tensioned reinforcing bars with a small concrete cover
anchorages of reinforcing bars prop against concrete struts. In because of a premature zipper type splitting failure starting from
cracked concrete, approximately 30% lower bond strengths are the loaded end. A corresponding test to check if the bond stiffness
tested if compared to uncracked concrete [17]. Thus, for the of post-installed reinforcing bars is acceptable is not given in
approval test conditions, one receives for concrete C20/25 EOTA TR 023 [7] because such high strength mortars were not on
fbm,req = 4.6/(0.7  0.9  0.7)  10 MPa. According to EN 1992-1 [2], the market when EOTA TR 023 was published. Such tests and
fbd increases with increasing concrete strength proportional to acceptance criteria are given in AC 308 [20] which was published
fck2/3. Therefore, also the mean bond strength fbm,req required by recently.
EOTA TR 023 develops in proportion with fck2/3 [7]. For the sake For product qualification, also the practicability of the installa-
of completeness it is noted that other regulations, e.g. the Model tion technique is checked and the influence of the relevant
Code 2010 [18] assumes that the bond strength increases with fc0.5. parameters on the bond performance (e.g. concrete compressive
As for cast-in reinforcing bars, the stressed area is used for the strength, bar diameter, drilling method, wet concrete, sustained
determination of the bond strength of post-installed reinforcing load, freeze/thaw cycles) is investigated to determine potential
bars on the basis of the tested ultimate load Fu, i.e. reductions of the allowable bond strength. Drilling of holes may
fb = Fu/(‘v  p  /). Therefore, it is not differentiated whether the fail- damage the concrete, this is why in EOTA TR 023 [7] the mini-
ure occurs between reinforcing bar and mortar or mortar and con- mum concrete cover cmin and the minimum spacing a of
crete. For each test series 5 pullout tests are carried out in the post-installed reinforcing bars is increased (hammer drilling:
course of the approval procedure. If the determined mean bond cmin = 30 mm + 0.06‘b P 2/, a = 40 mm P 4/; compressed air dril-
strength of post-installed reinforcing bars is smaller than the ling: cmin = 50 mm + 0.08‘b P 2/, a = 40 mm P 4/). These values
required mean bond strength, the approved design bond strength may be reduced if special drilling aids are used. Furthermore, it
is reduced (example 1 in Fig. 5a) which leads to increased anchor- is checked in the course of product qualification if the mortar
age lengths if compared to cast-in reinforcing bars. The mean bond provides the required protection against corrosion of the reinforc-
strength must be at least 7.1 MPa since this is the value corre- ing bar.
sponding to the lowest concrete strength class C12/15 otherwise In EOTA TR 023 [7], the characteristic steel yield strength of
the system is not approved. The utilization of mean bond strengths reinforcing bars and the characteristic compressive strength of
which are higher than the required values is not allowed (exam- concrete is limited to fyk = 500 MPa and fck = 50 MPa, respectively.
ple 2 in Fig. 5a). For this reason, all current high performance mor- Within these limitations, post-installed reinforcing bars allow a
tar products are approved for the same design bond strengths monolithic connection between existing and new reinforced con-
which are equal to the design bond strengths of cast-in reinforcing crete elements, provided suitable mortars are used and the manu-
bars. Research on post-installed reinforcing bars, however, showed facturer’s printed installation instructions (MPII) are followed.
that for a concrete cover larger than 3/ and a sufficiently strong Furthermore, qualified installers are required to drill and clean
mortar an increase of the design bond strength compared to the relatively long boreholes, inject the mortar and install the rein-
post-installed reinforcing bars might be possible, because the load forcing bars since no additional factors are taken into account for
transfer takes place at an enlarged bond area of the mortar with the installation safety.
 C. Mahrenholtz et al. / Engineering Structures 100 (2015) 645–655 649

2.2. Design of bonded anchors according to CEN/TS 1992-4 [4], in The factor k was empirically determined to kcr = 7.7 for cracked
future EN 1992-4 (Eurocode 2 Part 4) concrete and kucr = kcr/0.7 = 11.0 for uncracked concrete, if SI units
are used and the characteristic concrete compressive strength fck is
Bonded anchors consist of a steel element, e.g. a threaded rod or based on tests using cylinders with a diameter of 150 mm and a
a reinforcing bar, which is post-installed in a hole drilled into a length of 300 mm. The reduction of the capacity for edge distances
base material, e.g. a reinforced concrete element. Bonded anchors smaller than the critical value ccr,N is taken into account by the fac-
enable the transfer of tension and shear loads acting on the steel tors Ac,N/Ac,N0 and ws,N [4].
element into the reinforced concrete element. Since shear load
capacities are irrelevant in the context of this contribution, only  Bond capacity (Fig. 6d) – The anchorage fails when the bond
tension load capacities are discussed in the following. For bonded strength is exceeded and the anchor is pulled out. When assum-
anchors loaded in tension, steel failure (rupture or yielding of steel, ing a constant bond strength along the embedment depth, the
Fig. 6a and b), concrete cone failure (concrete breakout, Fig. 6c), design capacity equals to [16]:
bond failure (pullout, Fig. 6d) and splitting failure (cracking of con-
crete cover, Fig. 6e) may occur.
NRd;p ¼ p  /  ‘b  sRk  Ap;N =A0p;N  Ws;Np =cMp ðMode P;cMp ¼ cMc Þ ð8Þ
Currently, the design of concrete anchors is carried out accord-
ing to ETAG 001 Annex C [8] and in addition for bonded anchors The characteristic bond strength sRk has to be determined
according to EOTA TR 029 [21]. The content of both documents is experimentally and is different for cracked concrete (sRk,cr) and
included in the prestandard CEN/TS 1992-4 [4] which is uncracked concrete (sRk,ucr). The capacity reduction for edge dis-
referred to in the following. It is noted that the transition of tances smaller than the critical value ccr,Np is captured by the fac-
CEN/TS 1992-4 into the harmonized standard EN 1992-4 tors Ac,Np/Ac,Np0 and ws,Np [4].
(Eurocode 2 Part 4) is currently in progress and due for publication
in 2016. In the following, the equations for the calculation of the  Splitting of concrete (Fig. 6e) – Due to the small radial splitting
basic design capacities of bonded anchors are compiled. For the forces of bonded anchors, this failure mode is commonly gov-
sake of consistency with the structural concrete design variables, erning for a small member thickness which is not assumed in
the diameter and the length of bonded anchors are in this contribu- this contribution.
tion denominated by / and ‘b, respectively (instead of ds and hef).
However, the anchor design specific term edge distance but not The parameters Ac,N0 = (ccr,N)2 and Ac,N represent the full and
concrete cover is used. Furthermore, the bond strength in the con- actually developing surface of the conical breakout body; the
text of anchor design is denominated sR which value is not codified disturbance of the stress distribution close to edges is taken
(in contrast to fb). In general, the equations given in the following into account by ws,N = 0.7 + 0.3c/ccr,N 6 1.0. The parameters
for the calculation of the capacities are consistent with Ap,N0 = (ccr,Np)2 and Ap,N represent the full and actually available
CEN/TS 1992-4 [4]. To comply with the future EN 1992-4, however, cross sections of the activated volume; while the disturbance of
the used k-factors are those valid for cylinder (instead of cube) the stress distribution close to edges is taken into account by
compressive concrete strengths. ws,Np = 0.7 + 0.3c/ccr,Np 6 1.0. The edge distances critical for the con-
crete and bond capacities are ccr,N = 1.5‘b and ccr,Np = 10/
 Steel capacity (Fig. 6a and b) – For anchors, the design capacity is (sRk,ucr/7.5 MPa)0.5, respectively. The characteristic bond strength
calculated using the ultimate steel strength fu (Eq. (5)). To com- sRk is product specific and is deduced from the mean bond
ply with the design strength equation for reinforcing bars used strengths sRm determined by pullout tests with a short bond length
in reinforced concrete structures, however, the ultimate steel and large edge distance, in general with confined test setup.
strength is substituted by the steel yield strength fy (Eq. (6)): However, for the design of anchorages the bond strength as the
result of an unconfined test setup is needed. Therefore, the mea-
NRd;s ¼ As  f uk =cMs ðMode S;cMs ¼ 1:2  f uk =f yk  1:4Þ ð5Þ sured mean bond strength sRm is multiplied by asetup (=0.75 for
uncracked concrete). The calculated design capacity of a single
NRd;y ¼ As  f yk =cs ðMode Y;cs ¼ 1:15Þ ð6Þ bonded anchor equals to the minimum of the capacities deter-
mined for the individual failure modes:
2
The stressed area As is p  (//2) .
NRd ¼ minfNRd;s ; N Rd;c ; NRd;p g ð9Þ
 Concrete cone capacity (Fig. 6c) – The design capacity of a con-
ical concrete breakout which develops from the embedded end When designing connections of reinforced concrete structures,
of the anchorage is calculated as follows [16]: which is discussed in Section 4.2 below, the calculated design
capacity is:
0:5 0
NRd;c ¼ k  f ck  ‘1:5
b  Ac;N =Ac;N  Ws;N =cMc ðMode C;cMc ¼ cc  cinst Þ ð7Þ NRd ¼ minfNRd;y ; NRd;c ; NRd;p g ð10Þ

(a) (b) (c) (d) (e)

Fig. 6. Failure modes of a bonded anchor under tensile loading: (a and b) rupture and yielding of steel (steel failure); (c) concrete breakout (concrete cone failure); (d) pullout
(bond failure); (e) splitting.
650 C. Mahrenholtz et al. / Engineering Structures 100 (2015) 645–655

600 600

Failure load NR [kN]

Failure load NR [kN]

500 Governing failure Mode S 500 Mode S
S Y P C mode Governing
S Y Pfailure
C mode
400 Mode Y 400 Mode Y Mode C
300 Mode C
φ = 32 mm 300 φ = 32 mm.
Mode P τR = fb = 10.0 MPa Mode P τR = fb = 17.5 MPa

210
200 fy = 540 MPa 200 fy = 540 MPa
fu = 676 MPa fu = 676 MPa
100 fc = 21 MPa 100 fc = 21 MPa
uncracked cracked
0 0
0 5 10 15 20 25 0 5 10 15 20 25
Anchorage length ℓ b / φ [-] Anchorage length ℓb / φ [-]
(a) (b)
Fig. 7. Schematic representation of capacities corresponding to the individual failure modes; example cases for large edge distances (c P 1.5‘b > 3/): (a) single bonded
anchor; (b) bonded anchor group; (dotted line represents minimum capacity for the controlling failure modes) [22].

Note that in general, the steel capacity of qualified anchors is 3. Comparison of the design of anchorages as end anchorage
given in the ETA. To account for specific design situations and to and as bonded anchor
design bonded anchor groups, the basic values of the concrete cone
and bond capacities (Eqs. (7) and (8)) have to be multiplied by 3.1. General comparison of both design methods
additional factors wi. The discussion of the additional factors is
beyond the scope of this contribution, however, bonded anchor Design methods aim to model complex mechanical problems
groups are addressed in examples given below. realistically but as general and simple as possible. In principle,
Comparing the equations for the calculation of capacities corre- there are various models possible which application may lead to
sponding to the different failure modes (Eqs. (5)–(8)) allows the significantly different results. This is the case for the design meth-
determination of that anchorage length for which the governing ods of end anchorages and bonded anchors discussed in this contri-
failure mode changes from one to another. E.g. the anchorage bution. Both design methods have advantages and disadvantages.
length ‘b = ((p  /  ‘b  sRk)/(k  fck0.5))2/3 which corresponds to the Table 1 provides a general comparison of both design methods,
transition between concrete breakout and pullout may be calcu- taking also into account the critical range of application.
lated on the basis of Eqs. (7) and (8). In schematic representations,
these points are indicated by the intersection of the curves corre- 3.2. Summary of both design methods
sponding to the failure modes (Fig. 7). The lower bond envelope
of the curves represent the capacity of the anchorage as a function The design of reinforcing bar anchorages as end anchorage
of the anchorage length. according to EN 1992-1 [2] is relatively simple. The required

Table 1
Design of anchorages as end anchorage and as bonded anchor.

Design method End anchorage Bonded anchor

Relevant design standard EN 1992-1 (Eurocode 2, Part 1) CEN/TS 1992-4 in future EN 1992-4 (Eurocode 2, Part 4)
Load direction Tension (shear requires additional reinforcement Tension, shear, combination of both
in general)
Load transfer mechanism Equilibrium with local or global concrete struts Utilization of concrete tensile strength
Failure modes Steel yielding, bond failure (pullout) and splitting Steel failure, concrete breakout, pullout, splitting
Capacity Anchorage length ‘bd allows transfer of steel stress rsd NRk = min{NRk,s; NRk,c; NRk,p; NRk,sp}a
Basic design value of bond Deduced by calculation: fbd = 2.25  g1  g2  fctk/cc Tested and approved: sRd = sRk/cc
strength
Safety factors cc = 1.5 for concrete cMc = cMp = cc  cinst = 1.5  cinstb for concrete breakout, pullout, splitting
cs = 1.15 for reinforcing bars cMs = 1.2fuk/fyk P 1.4 for steel failure
Design steps 1. Calculation of required steel cross section of 1. Calculation of all characteristic capacities
reinforcement
2. Calculation of required anchorage length 2. Determination of minimum capacity controlling failure
anchorage
Limit states and assumed ULSanchorage at ULSstructure w > 0.3 mm ULSanchorage at SLSstructure w 6 0.3 mm
mean crack width (at SLSstructure w 6 0.3 mm) (condition base material)
Minimum spacing a = min{/; dg + 5 mm; 20 mm}c smin according to approval (ETA)
(dg largest aggregate size) (P5/ in general)
Minimum concrete cover cmine P /d depending on environment cminf according to approval (ETA)
(edge distance) (>5/ in general)
Allowable anchorage length ‘b P max{0.3‘b,req; 10/; 100 mm}g 20/ P ‘b P 4/h
a
Only failure modes under tension load given.
b
cinst = 1.0, 1.2 or 1.4 for systems with high, normal or low installation safety (factor depends in particular on the sensitivity of the bond strength with regard to the
cleaning of the drilled hole).
c
Minimum bar spacing of post-installed reinforcing bars for hammer and compressed air drilling method 40 mm P 4/.
d
Minimum concrete cover of post-installed reinforcing bars for hammer and compressed air drilling method 30 mm + 0.06‘b P 2/ and 50 mm + 0.08‘b P 2/, respectively.
e
With reference to outer edge of reinforcing bar.
f
With reference to axis of bonded anchor.
g
Without further assessment, minimum anchorage length for post-installed reinforcing bars to be increased by factor 1.5.
h
Without further assessment 5/.
 C. Mahrenholtz et al. / Engineering Structures 100 (2015) 645–655 651

anchorage length is calculated for the design steel stress rsd, where the required anchorage length for straight reinforcing bars by
rsd is limited by the design steel yield strength fyd. It is assumed 30%. However, post-installed reinforcing bars can only be installed
that the reinforced concrete elements are at the ultimate limit with straight ends. When post-installed reinforcing bars are
state (ULS). It is plausible to calculate the design bond strength spliced with existing reinforcement as shown in Fig. 8b, the design
fbd on the basis of the concrete tensile or compressive strength situation becomes worse because the anchorage length has to be
(Eq. (1)) because all components (concrete and reinforcement) increased by a factor of 1.5 if all reinforcing bars of the connection
are codified. In general, the shear load in the region of anchorages are spliced (EN 1992-1).
has to be carried by shear reinforcement. The minimum spacing as
well as the minimum concrete cover of 1/ are small. The anchor-
age length is in general P25/ to allow full utilization of the rein- 3.4. Comparison of design capacities
forcing steel.
For the design of reinforcing bar anchorages as bonded Fig. 9 shows the relationship of mean values (left), characteris-
anchor according to CEN/TS 1992-4 [4], the design capacities corre- tic values (middle) and design values (right) required by research-
sponding to all failure modes have to be calculated to determine ers to evaluate test results correctly and by practitioners to
the minimum capacity. The underlying characteristic bond calculate the capacities of end anchorages (top) according to
strength is product specific and may be substantially higher than EN 1992-1 [2] and of bonded anchors (bottom) according to
the bond strength of the same product used for the design of rein- CEN/TS 1992-4 [4]. Furthermore, the diagram demonstrates how
forcing bar anchorages according to the end anchorage design to derive design values used for the calculation of design capacities
method. The partial material safety factor for concrete cMc is equiv- on the basis of actual mean bond strengths which is required for
alent to cc = 1.5 given in EN 1992-1, multiplied by the installation the comparisons below. The shaded boxes indicate with which
safety factor 1.0 6 cinst 6 1.4 taking into account potential installa- values to start and the arrows how to arrive at other values.
tion inaccuracies. The partial material safety factor for steel Solid box frames indicate values which are codified and dashed
cMs = 1.4fyk/fuk P 1.2 is slightly larger than cs = 1.15 according to box frames represent values which may be estimated. The only
EN 1992-1. It is assumed that at the ultimate limit state (ULS) of value which definitely has to be tested is the bond strength of
the bonded anchor the concrete element serving as a base material post-installation systems.
is still at the service limit state (SLS), corresponding to a character- The scatter and overstrength of the steel strength may be
istic crack widths of 0.3 mm. Typically, the minimum spacing and taken into account by a factor multiplying the characteristic val-
edge distances are P5/ and therefore substantially larger than for ues to estimate mean values on the basis of codified characteris-
reinforcing bar anchorages designed as end anchorage. The appli- tic values. This factor is given in standards as e.g. 1.3 [18]; lower
cation of the bonded anchor design method is limited to an anchor- values may be found in other references, e.g. 1.25 [23]. The scat-
age length between 4/ and 20/ and elaborative, however, allows ter of the concrete compressive strength may be taken into
full utilization of the bond strength of the mortar. account by an allowance which is added to the characteristic
values to estimate mean values on the basis of codified charac-
3.3. Comparison of range of application teristic values. The allowance is defined in standards as e.g.
8.0 MPa [2]; lower values are common for studies under labora-
Fig. 8a illustrates the potential range of applications of the tory conditions, e.g. 4.0 MPa [24]. Design values are derived by
design methods for bonded anchors and end anchorages in regard dividing the characteristic values by material safety factors.
to concrete cover (edge distance) and anchorage length. Note that The safety factors cc = cMc for concrete, cs and cMs for steel are
EOTA TR 023 requires to increase the minimum anchorage length given in Table 1.
‘b,min (stipulated for straight cast-in reinforcing bars in In the following, the design capacities calculated for bonded
EN 1992-1) for post-installed reinforcing bars by a the factor of anchors and end anchorages are compared for the example with
1.5 unless it can be shown that the bond strength of the selected a single reinforcing bar / = 16 mm B500 (fyk = 500 MPa;
post-installed reinforcing bars and of cast-in reinforcing bars in fuk = 550 MPa) in uncracked concrete C20/25 (fck = 20 MPa)
cracked concrete (w = 0.3 mm) is at least equal. The required (Fig. 10). The design capacities Td for the design as end anchorage
anchorage length to develop the steel yield strength ‘b,req is typi- according to Eq. (4) (bold line) and NRd for the design as bonded
cally around 30/ to 50/ (depending on yield and bond strength, anchor according to Eq. (9) (dotted line) are plotted as a function
refer to Eq. (2)). If the actual design anchorage length ‘bd is shorter of the normalized anchorage length ‘b//. Note that these design
than ‘b,req, the utilization of the reinforcement is reduced by the capacities Td and NRd are indicated only for the range of application,
factor ‘bd/‘b,req. End anchorages with hooks allow a reduction of i.e. Td for ‘b P 10/ and NRd for 4/ 6 ‘b 6 20/.

Bonded anchor End anchorage

Normalized concrete cover c / φ [-]

20

15
Concrete cover
bonded anchor Concrete cover post-installed
10 rarely < 5φ reinforcing bars according to
ETA, cmin ≥ 2φ
5

0
0 5 10 15 20 25 30 35 40 45
Normalized anchorage length ℓb / φ [-]
(a) (b)

Fig. 8. (a) Range of application of bonded anchors and end anchorages with regard to concrete cover (edge distance) and anchorage length; (b) column-to-foundation
connection with post-installed reinforcing bars suggested in EOTA TR 023 [7].
652 C. Mahrenholtz et al. / Engineering Structures 100 (2015) 645–655

Fig. 9. Definition of steel, concrete and bond strengths required for end anchorages and bonded anchors according to the relevant design standards.

Fig. 10a to d shows the result for a low mean bond strength  The bond capacity is critical for bonded anchors installed with
fbm = sRm = 10 MPa valid for bonded anchors installed with low low strength mortar (sRm = fbm = 10 MPa). Compared to end
strength mortar and cast-in end anchorages. A high mean bond anchorages, the design capacity of bonded anchors with a
strength fbm = sRm = 30 MPa is assumed in Fig. 10e to h, which is large edge distance and a high installation safety
representative for epoxy based mortars used for bonded anchors (c P 1.5‘b > 3/; cinst = 1.0) is almost identical, otherwise
as well as for post-installed end anchorages. In Fig. 10a, b, e (c = 2/; cinst = 1.4) it is substantially smaller. This is due to
and f cinst = 1.0 and in Fig. 10c, d, g and h cinst = 1.4 are adopted the fact that the bonded anchor and end anchorage design
to show the pronounced effect of the installation safety factor on methods use different approaches to take the edge influence
the design capacity of bonded anchors NRd. Fig. 10a, c, e and g into account.
are valid for a large concrete cover (edge distance) of c P 1.5‘b > 3/  The concrete cone capacity controls the design capacity of
and Fig. 10b, d, f and h for a concrete cover (edge distance) of c = 2/ bonded anchors with high strength mortar (sRm = fbm =
which is required for post-installed reinforcing bars. As explained 30 MPa). For utilization of the steel strength, bonded anchors
above, the design capacity of the reinforcing bar Td is not a function require an anchorage length of ‘b  10/ in case of a large edge
of the actual bond strength and no factor to take the installation distance and high installation safety (c P 1.5‘b > 3/; cinst = 1.0).
safety into account is applied: Td solely depends on the design The anchorage length is increased to about 20/ for a small
bond strength corresponding to the concrete strength class (in this edge distance (c = 2/; cinst = 1.0). For a low installation safety
example fbd = 2.3 MPa) and ai (in this example only a2 = 1  (cinst = 1.4), the anchorage length must be increased by
0.15(cd – /)// for concrete cover). 40% compared to high installation safety (cinst = 1.0) to
The diagrams show the following with regard to the design achieve the same design capacity. In contrast, the required
methods: anchorage length of end anchorages is significantly larger in
all cases.
 For the design method of bonded anchors the steel capacity is
lower than for the design method of end anchorages. This is As discussed above, it is not surprising that different design
owed to the higher safety factor for steel failure of bonded methods have different outcomes. Furthermore, the different con-
anchors compared to end anchorages. ditions of use have to be taken into account. Notwithstanding, the
 C. Mahrenholtz et al. / Engineering Structures 100 (2015) 645–655 653

Fig. 10. Design capacity of a single anchorage according to EN 1992-1 and CEN/TS 1992-4 for the example / = 16 mm reinforcing bar B500 in concrete C20/25 (concrete
cover c equals axial edge distance minus //2).

bonded anchor design method draws the attention on a potential  If compared to the end anchorage design method, the bonded
weakness of the end anchorage design method. Fig. 11 shows for anchor design method potentially allows smaller anchorage
an example (diameter / = 16 mm reinforcing bar grade B500 and lengths to anchor the design steel yield capacity of single rein-
concrete strength class C20/25, clear distance a = 80 mm, concrete forcing bars (as pointed out above) or small groups of reinforc-
cover c  1.5 ‘b) the calculated anchorage length ‘b// which is ing bars, i.e. ‘b < ‘bd according to EN 1992-1.
needed to fully utilize the design steel yield capacity: An anchorage  The bonded anchor design method indicates that in particular
length ‘b = ‘bd = 33/ (Eq. (2)) is required for end anchorages, inde- large groups of reinforcing bars may fail due to pullout or con-
pendent of the number of anchored reinforcing bars. In contrast, crete breakout if the anchorage length is determined following
the anchorage length required for bonded anchors (Eq. (10)) the end anchorage design method, i.e. ‘b = ‘bd according to
depends on the number of anchored reinforcing bars and whether EN 1992-1. To utilize the design steel yield capacity, a signifi-
the reinforcing bars are cast-in or post-installed with low strength cant larger anchorage length may be required. This is plausible
mortar (e.g. sRm = 10 MPa) or high strength mortar (e.g. because the end anchorage design method does not take into
sRm = 30 MPa). In the diagram, also the failure modes calculated account concrete cone failure but requires supplementary rein-
according to the bonded anchor design method corresponding to forcement to take up the force introduced by the anchored rein-
the anchorage length ‘b = ‘bd = 33/ are indicated. The diagram forcing bars into the concrete. Note, however, that this
demonstrates: supplementary reinforcement may not be detailed in practice.
654 C. Mahrenholtz et al. / Engineering Structures 100 (2015) 645–655

stresses and the adverse effect of cyclic loading [6,28]. Whereas

fbm = 10 MPa designed as end anchorage
Calculated anchorage length ℓb / φ [-]

the anchorage length may be too small for increasing number of

needed to utilize steel yield capacity

τRm = 10 MPa designed as bonded Y Y anchored starter bars if these are designed as end anchorage
according to EN 1992-1 and no supplementary reinforcement is
τRm = 30 MPa anchor, γinst = 1.0 provided to take up the forces introduced by the anchored starter
Y bars into the concrete. More research is needed (e.g. studies on
Y Y beam-column joints) to allow the harmonization the design meth-
ods of bonded anchors and end anchorages for cast-in and
P PC CC post-installed reinforcing bars. In particular, the cracking behavior
Y Y must be studied because of its critical influence on the corrosion
protection of the reinforcement.

Y
5. Summary, conclusion, and outlook

This contribution describes the design methods for reinforcing

a = 80 mm Number of reinforcing bars n [pcs] bars as cast-in or post-installed end anchorages according to
φ = 16 mm EN 1992-1 and as bonded anchors according to CEN/TS 1992-4.
Also the qualification of post-installed reinforcing bars according
B500 φ a to EOTA TR 023 is explained. All relevant aspects of both design
C20/25 methods are discussed, and the interrelations of the material
Note: Mode Y: yielding; Mode P: pullout; Mode C: concrete breakout strengths required for the calculation of the design capacities are
systematically presented.
Fig. 11. Comparison of calculated anchorage length ‘b// required to fully utilize The comparison of both design methods by means of example
steel yield capacity (‘b = ‘b,req for end anchorage design method; failure modes given design situations shows that for high strength mortars the design
for bonded anchor design method correspond to indicated anchorage length).
of post-installed reinforcing bars as bonded anchors may allow
smaller anchorage lengths than required if designed as end anchor-
4. Application of both design methods for post-installed ages where only the bond strength corresponding to cast-in rein-
connections of reinforced concrete elements forcing bars can be taken into account. Furthermore, for several
reinforcing bars at close spacing the anchorage length designed
4.1. Post-installed reinforcing bars designed as end anchorage as end anchorage may be too short to exclude concrete cone
failure.
Post-installed starter bars for the connections of reinforced con- Finally, the design of end anchorages as part of strut-and-tie
crete elements designed as end anchorage are state of the art for long models is briefly discussed and the difficulties pointed out.
and proved to be valid if the full anchorage length according to the Reference is given to the successful application of post-installed
structural concrete design codes is specified or the utilization of the reinforcing bars designed as bonded anchors which was carried
reinforcement is reduced accordingly [e.g. 25]. The strut-and-tie out for the specific design situation of column-to-foundation con-
models required for the design of reinforced concrete connections nections. Further research is required to bridge the gap between
are very useful because they show the flow of forces, ensure equilib- the two design methods for various types of connections in order
rium and indicate regions where the concrete is stressed in tension. to predict realistic and similar capacities irrespective of the design
However, the approach has certain disadvantages. Strut-and-tie methods applied.
models are not always easy to construct in particular for
three-dimensional connections, e.g. column-to-foundation connec- References
tions [26]. The design requires several assumptions of the designer
and is not a straight forward task. Therefore, the predicted design [1] Schlaich J, Schäfer K, Jennewein M. Towards a consistent design of structural
concrete. PCI J 1987;32(3):74–150.
capacities are ambiguous. [2] EN 1992-1-1. Eurocode 2: Design of concrete structures – Part 1-1: General
rules, and rules for buildings. European Committee for Standardization (CEN);
4.2. Post-installed reinforcing bars designed as bonded anchor EN 1992-1-1:2004 +AC; 2010.
[3] ACI 318. Building code requirements for structural concrete (ACI 318–11) and
commentary (ACI 318R-11). Farmington Hills, Michigan: American Concrete
Comprehensive numerical and experimental studies were car- Institute; 2011.
ried out to scrutinize whether the anchorage of starter bars may [4] CEN/TS 1992-4. Design of fastenings for use in concrete – Part 4-1 – 4–5.
Technical specification, European Committee for Standardization (CEN),
be based on the design method for bonded anchors. To date, the Brussels; 2009.
focus was put on column-to-foundation connections [22,27]. [5] Eligehausen R, Mahrenholtz C, Pampanin S, Akguzel U. Bridging the gap
Beside pullout tests on reinforcing bars anchored with short bond between design provisions for connections using anchorage or strut- and tie
models. ACI Spec Publ 2014;296:1–16.
lengths, large scale tests on column-to-foundation connections [6] Mahrenholtz C, Akguzel U, Eligehausen R, Pampanin S. New design
with cast-in and post-installed reinforcing bars under monotonic methodology for seismic column-to-foundation anchorage connections. ACI
and cyclic loading were carried out. For comparison, the starter Struct J 2014;111(5):1179–89.
[7] EOTA TR 023. Assessment of post-installed rebar connections. Technical Report
bars of the large scale test members were designed with and with- 023; 2006.
out hook while the anchorage lengths were smaller than the [8] ETAG 001. Guideline for European technical approval of metal anchors for use
required anchorage length according to the design method for in concrete, Parts 1–6. European Organization of Technical Approvals (EOTA),
Brussels; 2006.
end anchorages. The behavior of the test members with straight
[9] Directive 89/106/EEC. Council Directive of 21 December 1988 on the
cast-in and post-installed starter bars were not significantly approximation of laws, regulations and administrative provision of the
different. It was shown that the bonded anchor design method Member States relating to construction products; 1988.
leads to safe results even for seismic applications if the rules of [10] Regulation (EU) No 305/2011. Regulation (EU) No 305/2011 of the European
Parliament and The Council of 9 March 2011 laying down harmonised
CEN/TS 1992-4 are amended by additional factors take into conditions for the marketing of construction products and repealing Council
account the beneficial effect of moment induced confinement Directive 89/106/EEC; 2011.
 C. Mahrenholtz et al. / Engineering Structures 100 (2015) 645–655 655

[11] EAD 330087-00-0601. Systems for post-installed rebar connections with [22] Mahrenholtz C. Seismic bond model for concrete reinforcement and the
mortar. Draft of European Assessment Document, TAB DIBt, Berlin; 2015. application to column-to-foundation connections. Dissertation, University of
[12] Spieth H. Tragverhalten und Bemessung von eingemörtelten Stuttgart; 2012.
Bewehrungsstäben (Behaviour of post-installed reinforcing bars). [23] Bachmann H. Erdbebensicherung von Bauwerken (Earthquake Guarding of
Dissertation, University of Stuttgart (in German); 2003. Buildings). 2. überarbeitete Auflage, Birkhäuser Verlag, Basel (in German);
[13] fib Bulletin No. 58. Design of anchorages in concrete. Guide to good practice. 2002.
Bulletin No. 58, Fédération International du Béton (fib); 2011. [24] Reineck K-H, Kuchma D, Kim K-S, Marx S. Shear database for reinforced
[14] Rehm G. Über die Grundlagen des Verbundes zwischen Stahl und Beton (Basic concrete members without shear reinforcement. ACI Struct J 2003;100(2):
principles of the bond between steel and concrete). Dissertation, Technische 240–9.
Universität München, Schriftenreihe DAfStb, No. 138 (in German); 1958. [25] Spieth H, Eligehausen R. Bewehrungsanschlüsse mit nachträglich
[15] ISO 15630-1. Steel for the reinforcement and prestressing of concrete – test eingemörtelten Bewehrungsstäben (Reinforcement connections with post-
methods – Part 1: reinforcing bars, wire rod and wire; 2010. installed reinforcing bars). Beton- und Stahlbetonbau 2002;97(Heft 9):445–59
[16] Eligehausen R, Cook R, Appl J. Behavior and design of adhesive bonded (in German).
anchors. ACI Struct J 2006;103(6):822–30. [26] Bruckner M. Anwendung von Ankerstäben in Rahmenecken,
[17] Idda K. Verbundverhalten von Betonrippenstählen bei Querzug (Bond Rahmenendknoten und Stütze-Fundament-Verbindungen (Application of
behaviour of deformed reinforcing bars considering transverse tension). headed bars in external beam-column joints and column-to-foundation
Dissertation, IMB, Universität Karlsruhe (TH) (in German); 1999. connections). Dissertation, University of Stuttgart (in German); 2006.
[18] fib Model Code 2010. Design Code, Fédération Internationale du Béton; 2013. [27] Herzog M. Vereinheitlichung der Bemessung im Stahlbetonbau und in der
[19] Randl N, Kunz J. Post-installed reinforcement connections at ultimate and Befestigungstechnik (Unification of the design concept for reinforced concrete
serviceability limit states. Struct Concr 2014;15(4). and concrete anchors). Dissertation, University of Stuttgart (in German); 2014.
[20] AC 308. Acceptance criteria for post-installed adhesive anchors in concrete [28] Mahrenholtz C, Eligehausen R. Simulation of tests on cast-in and post-installed
elements. Whittier, California: International Code Council Evaluation Service, column-to-foundation connections to quantify the effect of cyclic loading.
Inc. (ICC-ES); 2013. ASCE J Struct Eng, 2015 [in press]. http://ascelibrary.org/doi/abs/10.1061/
[21] EOTA TR 029. Design of bonded anchors. Technical Report 029; 2010. (ASCE)ST.1943-541X.0001330.