Sie sind auf Seite 1von 14

Engineering Structures 156 (2018) 210–223

Contents lists available at ScienceDirect

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

Detailing of concrete-to-concrete interfaces for improved ductility T


a,⁎ b
Eduardo Cavaco , Ilton Pacheco, José Camara
a
CEris, FCT – Universidade NOVA de Lisboa, Almada, Portugal
b
CEris, IST – Universidade de Lisboa, Lisboa, Portugal

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

Keywords: Recent research has shown that reinforced concrete (RC) beams with concrete-to-concrete casting interfaces
Concrete-to-Concrete where plastic hinges are likely to develop, may experience reduced ductility in comparison to similar structural
Interface elements casted at once, due to a potential shear slippage along the casting interfaces. Although of relevant
Shear-friction importance for both precast and cast in situ RC structures, this problem is still not addressed in current codes and
Precast construction
standards, which limit the safety check of casting interfaces to the verification of their strength based on im-
Plastic hinge
proved expressions of the “shear friction theory”, the latter proposed in the 60’s. However, recent research has
Ductility
shown that friction strength of casting interfaces depends on interface width opening, and it is significantly
reduced after the yield of the bending reinforcement. During the formation of plastic hinges, shear stresses run
preferentially across the compressed zones of the interfaces, reducing their strength, and ultimately the speci-
mens’ ductility.
In this paper, different and alternative details for interfaces are proposed to improve global behaviour, and in
particular, ductility of RC beams with casting interfaces located on plastic hinges regions. An experimental
campaign was carried out to study the effect of: (i) epoxy and latex based adhesion promoters’ usage between
castings; (ii) web reinforcement; (iii) geometry of interfaces; (iv) and shear level.
Results show that both epoxy and latex based adhesion promoters, currently used in construction, hardly
improve the tensile strength of casting interfaces, to a point that the interface presence has negligible impact on
the cracking pattern. A much better result was observed from the use of a web reinforcement crossing the
interface perpendicularly. Although this solution revealed itself also incapable to avoid preferential cracking
along the interfaces, it proved to be efficient in limiting shear slippages. The adoption of inclined interfaces
either perpendicular or parallel to the expected direction of shear cracks proved also to be an efficient solution.
Finally, the likelihood of experiencing a shear slippage along the interface is strongly dependent on the existing
shear level after the formation of a plastic hinge.

1. Introduction between cast-in-place and precasted elements.


The safety verification of these interfaces has been traditionally
In the design and construction of reinforced concrete (RC) struc- carried out using the shear “shear-friction theory”, originally proposed
tures, every time two concretes are cast against each other at different by Birkeland and Birkeland [1] for the connections of the precasted
moments, and the hardening process of the older concrete is already construction. Later on, it was also adopted, with subsequent improve-
finished, a concrete-to-concrete interface is created and an additional ments, by the generality of codes and, in particular, by the Eurocode 2
safety check is required, corresponding to the verification of the in- [2] and the ACI 318–14 [3] or the very recent Model Code 2010 [4].
terface capacity to transfer loads across. The situation is recurrent in the However, the “shear-friction theory”, including the following changes,
construction of RC structures. did not address the safety check of interfaces subjected to a combina-
For cast-in-place structures, the concrete-to-concrete interfaces re- tion of shear and bending moment as it was developed having in mind a
sult essentially from the limited production resources or from the shear failure as a slippage along a mode II crack, according to the
casting plans, which do not allow always a single cast per structure, but classification of the mechanics of fracture [5], subjected to shear and
also from unforeseen events leading to interruptions in the erection normal forces. However, interfaces subjected to shear and bending
process. Concrete-to-concrete interfaces are also recurrent in precasted moment, and where a shear slippage along a mode I crack can poten-
RC construction in the connections between precasted elements or tially occur, are also frequent in real practice, in both cast in place or


Corresponding author.
E-mail address: e.cavaco@fct.unl.pt (E. Cavaco).

https://doi.org/10.1016/j.engstruct.2017.10.058
Received 15 May 2017; Received in revised form 17 October 2017; Accepted 23 October 2017
0141-0296/ © 2017 Elsevier Ltd. All rights reserved.
E. Cavaco et al. Engineering Structures 156 (2018) 210–223

Nomenclature ν strength reduction factor


νu ultimate shear friction strength
α angle between shear reinforcement and the shear plane ρ reinforcement ratio
β coefficient allowing for angle of concrete diagonal strut σn normal stress acting on interface due to external loading
c coefficient of cohesion Asl- negative longitudinal reinforcement area
cr coefficient of interlocking effect Asl+ positive longitudinal reinforcement area
fc concrete compressive strength Asw transversal reinforcement area
fcm mean value of concrete compressive strength MRm bending strength
fctd design value of concrete tensile strength PRmM expected ultimate testing load due to a bending failure
fctm mean value of concrete tensile strength PRmS expected ultimate testing load due to a shear friction
fy yield strength of reinforcement failure
fym mean value of yield strength of reinforcement PRmV expected ultimate testing load due to a shear failure
k1 Randl’s coefficient of efficiency of reinforcement VRmM shear load at the time of a bending failure
k2 Randl’s dowel action coefficient VRmS shear friction strength
μ coefficient of friction VRmV shear strength

precast structures. Some examples include: (i) the interfaces defined by or normal forces perpendicular to the interface, or a combination of
the connection of precast beams elements to execute a continuous beam both, where a shear slippage is possible to occur, has been the object of
on the hogging region; (ii) the casting interface between columns and several research works published in the latter 50 years [7–22]. The
the foundations; (iii) and every casting interface, created accidentally “shear-friction theory”, originally proposed in [1] and given by Eq. (1),
or intentionally, on beams and columns of cast in place RC structures has been used to predict the shear strength of concrete interfaces.
matching sections subjected simultaneously to shear and non-negligible
vu = μρf y (1)
bending moment.
Recent research work [6] has shown that the “shear-friction theory”, It has seen major improvements over the years, as a result of several
including the design expressions of the Eurocode 2 [2] and the ACI research works, to take into account several different effects such as
318–14 [3], are incapable to describe the physical mechanism behind adhesion, aggregate interlock, dowel action, the weakest concrete, in-
the load transfer across an interface subjected to a normal stress gra- terface roughness, among others. It has been adopted in most design
dient resulting from the combination of a shear force and a bending codes, in particular in the ACI 318–14 [3] (see Eq. (2)), Eurocode 2 [2]
moment. Moreover, it was found that the presence of a vertical inter- (see Eq. (3)) or the very recent Model Code 2010 [4] (see Eq. (4)).
face between two differently aged concretes is responsible for a ducti-
lity reduction in bending of a RC beam, in comparison to that of a beam vu = ρf y (μsinα + cosα ) (2)
produced in a single cast, thus without interfaces.
Following the research presented in [6], in particular the results vu = cfctd + μσn + ρf y (μsinα + cosα ) ⩽ 0.5νfc (3)
suggesting the ductility reduction of a RC beam with vertical casting
interfaces, in this paper different and improved design strategies are vu = cr fc1/3 + μ (ρk1 f y + σn ) + k2 ρ f y fc ⩽ βνfc (4)
proposed and the corresponding behaviours are experimentally in-
However, and according to the crack classification of the fracture me-
vestigated..
chanics [5], the original theory has been developed for the hypothesis
of having a shear slippage along a mode II crack, and not a mode I
2. Literature review crack, resulting from tension stresses. Moreover, this fact has rarely
been addressed in the subsequent improvements made to the original
The behaviour of concrete-to-concrete interfaces subjected to shear theory. The application of the “shear-friction theory”, including the

Fig. 1. Standard beam specimen: (a) side view; (b) cross section.

211
E. Cavaco et al. Engineering Structures 156 (2018) 210–223

Table 1
Specimens details, interface design and reinforcement detailin.

Specimen Interface Longitudinal reinforcement Transverse


design reinforcement

J1 and J2 J3 and J4 Compression Tension


zone zone

V1 Vertical Interface width Vertical Interface width 4ϕ8 4ϕ20 + 4ϕ12 + 4ϕ10 ϕ8//7cm
Sikalatex adhesive Icosit K101 adhesive (2.0 cm2) (20.2 cm2) (14.3 cm2/m)
V2 Vertical Interface Vertical Interface width
web reinforcement
V3 Interface perpendicular Interface parallel
to shear cracks to shear cracks
V4 Vertical Interface Vertical Interface width 4ϕ16 + 4ϕ12 + 4ϕ10
web reinforcement (15.7 cm2)

Fig. 2. Beam specimens: (a) standard procedure; (b) V1; (c) V2 and V4; (d) V3.

Fig. 3. Adhesion promotors application in specimen B1 interfaces: (a) J1 and J2 (Sika Sikalatex); (b) J2 and J3 (Sika Icosit K101).

expressions adopted in the design codes, to predict the shear strength to transfer strength; (ii) tension across the shear plane results in a re-
slippage of an interface subjected to a combination of shear force and duction in shear transfer strength equal to that which would result from
bending moment is at least questionable, as a tension crack is likely to a reduction in the reinforcement ratio, ρf y , by an amount equal to the
develop partially along the interface, due to the bending moment effect. tension stress. Thus, in the case of the bending moment acting across
The research conducted by Mattock et al. [23,24] was pioneering by the shear plane, it is assumed that opposite and self-compensated ef-
addressing explicitly the effect of the bending moment on the concrete- fects to the shear strength are provided by both the auto balanced
to-concrete interfaces strength. A set of corbel type push-off specimens flexural tension and compression stresses.
with moment or tension acting across the shear plane was tested. It was Recently Sørensen et al. [25] addressed the problem of ductility on a
concluded that: (i) moments in the shear plane less than or equal to the keyed shear connections for RC shear walls. The authors proposed a
flexural ultimate moment of the shear plane do not reduce the shear new design for improved ductility, but the experimental tests were

212
E. Cavaco et al. Engineering Structures 156 (2018) 210–223

Fig. 4. Casting interfaces of specimen V3, leaved perpendicular and parallel to the expected shear-cracking pattern.

Fig. 5. Interface preparation for the second casting.

conducted without the effect of the bending moment. shear and bending moment.
Cavaco and Camara [6], based on the experimental testing of con- Tested specimens consisted into four “I” beams (V1 to V4) with
tinuous RC beams with casting interfaces on the hogging and sagging 3.50 m long and 0.50 m high, simply supported with 3.0 m span and
regions, concluded that the “shear-friction theory”, including the ex- loaded at the central section, as shown in Fig. 1. Beams were cast in two
pressions based thereon, do not faithfully describe the physical me- stages, following such a convenient procedure that four casting joints
chanism behind the strength to slippage of concrete-to-concrete inter- per specimen were created: J2 and J3 subjected to shear and bending
faces subjected to both shear and bending moment. Although neither moment; and J1 and J4 subjected to identical shear load but negligible
the shear strength nor the bending strength are affected by the presence bending moment.
of a casting interface, the reduced tension strength between the two The research conducted in [6] suggested that the interface between
differently aged concretes, below that of the weakest concrete, leads to two differently aged concretes affects the specimens ductility, due to a
a premature and pronounce cracking along the interface, in the tension reduction in the plastic hinge rotation capacity, rather than specimens
zone, reducing the local friction strength. This reduction is most sig- strength. Thus, the new specimens (V1 to V4) were detailed with the
nificant after the yield of the bending reinforcement due to un- transverse reinforcement overdesigned in relation to the longitudinal
controlled interface width opening. The shear stresses are therefore reinforcement (see Table 1), to ensure yielding of the latter to occur
deviated to the compressed zone of the interface, reducing its strength first.
and ultimately the plastic hinge rotation and the specimen ductility. A Different and alternative design strategies were used for the casting
similar behaviour has already been documented in reinforced concrete interfaces of the specimens. According to Cavaco and Camara [6], a
members without shear reinforcement [26]. premature and pronounced cracking along the interface in the tension
The capacity of concrete-to-concrete interfaces to transfer shear zone was observed due to the bending moment effect and reduced
across is not purely of frictional nature, and it is strongly dependent on tensile strength of the interface, smaller than that of the weakest con-
the interface cracking pattern and opening. These details are neither crete. This abnormality, in comparison with the single casted specimen,
considerer in current codes nor in the improved “shear-friction theory” conditioned the cracking pattern in the interface vicinity, and led to a
based expressions. reduced capacity to transfer loads across the interface towards the
support. Taking this into account, in the present research work, some
adhesion promoters were applied to the interfaces in the V1 specimen,
3. Experimental research in an attempt to increase their tensile strength and avoid premature
cracking. A latex water based dispersion (Sika Sikalatex) was used in
The experimental work presented herein was planned following the interfaces J1 and J2 (see Fig. 3(a)), and a two component epoxy based
research and results presented in [6] suggesting that the tensile strength structural adhesive (Sika Icosit K101) was applied to interfaces J3 and
of the casting interfaces is smaller than that of the weakest concrete. J4 (see Fig. 3(b)). The application was completed just prior to the
This leads to a pronounced and premature cracking along the interface second casting and following the procedures recommended in the
restricting its ability to transfer shear loads across, in particular after manufacturer technical datasheets.
the longitudinal reinforcement yields. Thus, this second experimental In the specimen V2, a dedicated web reinforcement (see Fig. 2(c)),
work described in this paper aimed to investigate the effect of: (i) epoxy consisting into three double rebars with 8 mm diameters (3.0 cm2), was
and latex based adhesion promoters; (ii) web reinforcement transverse used in interfaces J3 and J4, in an attempt to improve structural be-
to the interface; (iii) interface geometry; (iv) and shear level, on the haviour by controlling cracking pattern development and pronounced
strength and ductility of beam specimens with casting interfaces under

213
E. Cavaco et al. Engineering Structures 156 (2018) 210–223

Fig. 6. Reinforcement detailing of specimens: (a) lateral view of specimen V1, V2 and V4; (b) lateral view of specimen V3; (c) Cross section 1–1; (d) Cross section 2–2.

Table 2 particularly inclined (θ = 45°, see Fig. 2(d)). Concrete-to-concrete in-


Mechanical properties of steel bars. terfaces J1 and J2, and, J3 and J4, resulted perpendicular and parallel
to the expected shear-cracking pattern, respectively (see Fig. 4). The
Reinforcement fym εym fum εum
(MPa) (%) (MPa) (%)
shear-cracking pattern inclination was estimated based on a developed
strut-and-tie model of the specimen. The idea behind this alternative
Transverse 556 27.8 639 107.4 geometry was: (i) on one hand to force shear cracking to cross the
Longitudinal 566 28.9 664 241.2 casting interfaces (J1 and J2), which was not observed in the tests
conducted in the previous work [6]; (ii) and on the other hand (in in-
terfaces J3 and J4), to do not disrupt the traditional shear-cracking
width opening along the interface, as it was observed in the previous
pattern, which was observed by [6] due to the presence of a vertical
work by Cavaco and Camara [6]. For a direct comparison, simple ver-
interface with reduce tensile strength.
tical interfaces were used in casting surfaces J3 and J4.
The specimen V4 was produced equal to the specimen V2, in what
In specimen V3, casting surfaces were left intentionally and
concerns to interface design, however with reduced bending

Table 3
Mean values of the concrete compressive and tensile strengths.

Specimen Compressive strength tests fcm,28 (MPa) fctm,28 (MPa)

1st casting Age (days) 2nd casting Age (days) 1st casting 2nd casting 1st casting 2nd casting
fc,cm (MPa) fc,cm (MPa)

V1 58.5 48 55.7 38 45.8 44.1 3.4 3.3


V2 47.7 47 55.2 35 38.3 44.0 2.9 3.3
V3 51.7 41 64.8 31 41.1 54.2 3.1 3.9
V4 61.0 46 58.3 32 48.4 47.6 3.5 3.5

214
E. Cavaco et al. Engineering Structures 156 (2018) 210–223

Table 4 compressive strength. The concrete tensile strength was estimated


Theoretical bending strength - MRm. based on the Eurocode 2 [2]. The mean values of the concrete com-
pressive strength at the testing days, 28 days, and the respective tensile
Specimen Section Asl+ (cm2) fcm (MPa) MRm (kN m) VRmM/PRmM (kN)
strengths are presented in Table 3. The second cast of specimens was
V1 J1 and J4 10.1 44.1 311 370/740 executed 11.5 days, in average, after the first cast, simulating the con-
J2 and J3 20.2 555 struction of a real precast structure.
Mid-span
Table 4 summarizes the theoretical bending strength of specimens
V2 J1 and J4 10.1 38.3 311 370/740 assessed using mean values for both steel and concrete strength and
J2 and J3 20.2 450 according to Eurocode 2 [2] and considering perfect monolithic beha-
Mid-span
viour. The compressive strength of the weakest concrete was used in the
V3 J1 and J4 10.1 41.1 312 375/750 calculations of the resisting bending moment of interfaces J1 to J4 for
J2 and J3 20.2
the four specimens. Although some differences were found between
Mid-span
concrete compressive strengths of different specimens and castings, as
V4 J1 and J4 7.8 47.6 241 293/585
expected, resisting bending moments for similar sections resulted al-
J2 and J3 15.7 439
Mid-span most the same. According to Table 4, the ultimate load of specimens V1
to V4 ranged from 585 kN to 750 kN and it was conditioned by the
bending failure of the mid-span section. The corresponding acting shear
Table 5 at failure ranged from 293 kN to 375 kN.
Theoretical shear strength - VRmV. Table 5 summarizes the specimens theoretical shear strength con-
trolled by the failure of the stirrups in the web, or concrete crushing at
Specimen Asw (cm2/m) VRmV (kN) PRmV (kN)
the compression struts, computed according to the Eurocode 2 [2] and
V1 14.3 426 725a 852 assuming perfect monolithic behaviour. An angle of 42°, estimated
V2 647a based on a developed strut-and-tie model, was considered for the failure
V3 685a surface and the weakest concrete was considered in the calculations. As
V4 769a
shown, failure of stirrups was decisive. A similar expected load bearing
a
Crushing at the compression struts. capacity resulted for all the specimens and was equal to 852 kN.
If the possibility of a shear friction failure along the casting inter-
reinforcement in the tension zone. The objective herein was to compare faces was neglected, it was possible to conclude, by comparing the re-
the importance of the shear level in the potential ductility reduction of sults presented in Tables 4 and 5, that the expected ultimate load
the specimen. bearing capacity of all the four specimens was controlled by the failure
With the exception of specimen V1, whose interfaces J1 to J4 were in bending moment of the mid-span section. It is worth to mention that
treated with adhesion promoters, the first casting of remaining speci- one of the objectives of the tests herein presented is to study the effect
mens was completed against a piece of styrofoam. Preceding the second of concrete-to-concrete interfaces on the bending ductility of beam
cast, the hardened concrete surfaces were saturated and scarified with specimens, since previous research [6] has already shown that these
an air impact hammer with the goal of removing attached styrofoam, interfaces have negligible effect on the shear strength of similar speci-
remove the laitance and expose the coarse aggregate (see Fig. 5). The mens.
idea was to simulate real practice. The surface was finally classified as: On Table 6, the estimated theoretical shear-friction resistance of the
(i) free of laitance and not intentionally roughened accordingly to the specimens casted in different stages is presented, according to Eqs.
ACI 318–14 [3]; (ii) “smooth” accordingly to the Eurocode 2 [2] and (2)–(4). For the interfaces subjected to both shear and bending moment,
the Model Code 2010 [4]. However, it must be noted that a quantitative the following hypothesis were considered in the calculations: (i) tensile
measure of the interface mean roughness was not carried out, and thus and compressive strength of concrete were considered those of the
a conservative classification was adopted. weakest material; (ii) where moment acts on a shear plane, the ACI
With the exception of the interfaces J3 and J4 of the specimens V1 318–14 [3] based on the work of Mattock [24] suggests the flexural
and V4, with dedicated web reinforcement, in the remaining specimens, compression and tension forces are in equilibrium and do not change
the longitudinal reinforcement crossing the interfaces J1 to J4, was the resultant compression acting across the shear plane or the shear-
strictly the necessary to resist the bending tension forces. As suggested friction resistance. It is therefore not necessary to provide additional
by the ACI 318–14 [3], this reinforcement can be used simultaneously reinforcement to resist the flexural tension stresses, unless the required
for bending and shear friction purposes, without interaction between flexural tension reinforcement exceeds the amount of shear-transfer
both. It is worth to mention that this situation is addressed neither in reinforcement provided in the flexural tension zone; (iii) coefficients
the Eurocode 2 [2] nor in the Mode Code 2010 [4]. In the latter, a depending on interface roughness were obtained based on the re-
bending, tensile and shear interaction in the reinforcing bars crossing spective standards [2–4] and considering the interface classification
the interface is considered to distress both the friction and dowel previously described. Exception was made for the interfaces of spe-
strengths. These are reduced by using coefficients k1 and k2 in Eq. (4), cimen V1, treated with epoxy and latex adhesives. In this case, and
respectively, which translate the impact of local bending of reinforcing considering that the product manufacturer suggest a monolithic beha-
bars, caused by the shear slip of the interface, on its tensile and shear viour resulting from these products usage, the maximum friction coef-
strengths. However, it must be noted that these coefficients do not in- ficient (1.4) was considered in the calculations, when the ACI 318–14
tend to reflect the effect of an external tensile force and/or bending [3] design expression was used. On the other hand, maximum value for
moment. the adhesion coefficient was considered when the Eurocode 2 [2] de-
Finally, 4ϕ8 mm steel bars were adopted in the compression zone, to sign expressions was used. Finally, when applying Model Code 2010 [4]
allow reinforcements assemblage during the construction stage. expression, c coefficient in Eq. (4), even for the epoxy an latex treated
Reinforcement detailing of specimens V1 to V4 is depicted in Fig. 6. surfaces, was considered similar to that of “smooth” interfaces, as in
Three samples of each steel bar size were subjected to tensile tests. this expression, the referred coefficient is related to the aggregate in-
The mechanical properties of the reinforcement are presented in terlock effect and thus it cannot be increased by the adhesives use; (iv)
Table 2. potential friction failure surfaces of specimen V3 were considered with
Three cubic concrete specimens per casting were tested for the an angle of 45°. A percentage of the transverse reinforcement was
considered for frictional proposes. This percentage corresponds to that

215
E. Cavaco et al. Engineering Structures 156 (2018) 210–223

Table 6
Theoretical shear friction strength – VRmS.

Expression Specimen c μ k1 k2 α (°) fcmmin (MPa) fctmmin (MPa) fym (MPa) Interface ρ (%) σn (MPa) VRmS (kN) PRmS (kN)

ACI 318–14 V1 – 1.4 – – 90 44.1 3.3 566 J1 0.63 0.0 959 1918
[3] J2 1.16 1759
Eq. (2) J3 1.16 1759
J4 0.63 959
V2 – 0.6 – – 90 38.3 2.9 566 J1 0.79 0.0 513 822
J2 1.32 856
J3 1.16 754
J4 0.63 411
V3 – 0.6 – – 45 41.1 3.1 566 J1 0.48 0.98 1401 1903
J2 0.85 0.98 2316
J3 0.85 −0.98 1866
J4 0.48 −0.98 951
V4 – 0.6 – – 90 47.6 3.5 566 J1 0.67 0.0 435 666
J2 1.08 703
J3 0.93 601
J4 0.51 333

Eurocode 2 V1 0.5 0.5 – – 90 44.1 3.3 566 J1 0.63 0.0 656 1311
[2] J2 1.16 942
Eq. (3) J3 1.16 942
J4 0.63 656
V2 0.35 0.5 – – 90 38.3 2.9 566 J1 0.79 0.0 622 1075
J2 1.32 908
J3 1.16 823
J4 0.63 538
V3 0.35 0.5 – – 45 41.1 3.1 566 J1 0.48 0.98 1704 1831
J2 0.85 0.98 2562
0 J3 0.85 −0.98 1773
J4 0.48 −0.98 915
V4 0.35 0.5 – – 90 47.6 3.5 566 J1 0.67 0.0 595 1021
J2 1.08 819
J3 0.93 734
J4 0.51 511

Model Code 2010 V1 0 0.6 0.5 1.1 90 44.1 3.3 566 J1 0.63 0.0 416 831
[4] J2 1.16 763
Eq. (4) J3 1.16 763
J4 0.63 416
V2 0 0.6 0.5 1.1 90 38.3 2.9 566 J1 0.79 0.0 501 803
J2 1.32 836
J3 1.16 737
J4 0.63 401
V3 0 0.6 0.5 1.1 45 41.1 3.1 566 J1 0.48 0.98 845 790
J2 0.85 0.98 1327
J3 0.85 −0.98 877
J4 0.48 −0.98 395
V4 0 0.6 0.5 1.1 90 47.6 3.5 566 J1 0.67 0.0 448 687
J2 1.08 725
J3 0.93 620
J4 0.51 343

exceeding the minimal transverse reinforcement demanded to resist the displacement and using a 100 ton hydraulic jack attached to a reaction
maximum possible shear load (375 kN, see Table 4). On these inter- steel frame with a maximum capacity of 100 ton (see Fig. 7). A single
faces, additional stresses perpendicular to the interface plane were load cell (100 kN capacity) and a displacement transducer (100 mm
considered. These stresses were defined considering the shear load capacity) were used, above and below the loading section, respectively.
vector decomposition into two components, the first parallel to the Between 20 and 27 strain gages per specimen were used to record
interface and the second perpendicular to it. The normal component strains on both the longitudinal and the transverse reinforcements.
generates compression stresses on interfaces J1 and J2 and tensile
stresses on interfaces J3 and J4. 4. Results and discussion
Table 6 shows: (i) the Model Code 2010 [4] to provide the most
conservative approach since it considers the reduction of the friction Fig. 8 shows the load-displacement diagram obtained for tested
and dowel action strengths due to the local bending of the reinforcing specimens V1 to V4. Observed global behaviours were very similar with
bars; (ii) specimens strength is controlled by the interfaces J1 and J4 the exception of that respecting to specimen V2, showing significant
friction strength; (iii) comparison with the estimated bending strength reduced ductility. Specimen V4 has shown comparable ductility to that
(MRm; see Table 4) shows the latter to be critical. However, [6] suggests of specimens V1 and V3, although with reduced load bearing capacity,
that friction strength starts reducing with the interface opening, parti- explained by the inferior amount of bending reinforcement.
cularly after the yield of the bending reinforcement, and therefore a The initial stage of the load-displacement diagram shows a reduc-
potential friction failure of interfaces J2 and J3, should not be dis- tion of specimens’ stiffness for a testing load of approximately 120 kN,
regarded. due to the initiation of cracking. However, first cracks were firstly
The specimen tests were performed controlling imposed visible for testing loads between 90 kN and 100 kN for specimens V4

216
E. Cavaco et al. Engineering Structures 156 (2018) 210–223

Fig. 7. Experimental Setup.

Fig. 8. Load-displacement diagram of specimens V1 to V4.

Table 7
Loading test results of specimens.

Specimen Cracking load/displacement Load/displacement Failure

J2 and J3 Mid-span Yield Maximum Ultimate Mechanism Ductility factor

V1 100 kN/0.8 mm 120 kN/1.0 mm 650 kN/15 mm 745 kN/81 mm 725 kN/88 mm J2 slip 5.9
V2 100 kN/0.9 mm 150 kN/1.5 mm 685 kN/16 mm 720 kN/31 mm 710 kN/42 mm J3 slip 2.6
V3 100 kN/0.8 mm 150 kN/1.5 mm 685 kN/16 mm 807 kN/89 mm 804 kN/95 mm Bendinga > 6.3
V4 90 kN/0.7 mm 115 kN/1.0 mm 560 kN/15 mm 660 kN/87 mm 645 kN/95 mm Bendingb 6.3

a
Crushing of concrete at the compressed flange.
b
Failure of the longitudinal reinforcement.

and the remaining, respectively (see Table 7). Moreover, and as de- [6], which suggested that this was the reason for an atypical cracking
picted in Fig. 9, cracking was initiated simultaneously on interface J2 pattern development in the interfaces vicinity, leading ultimately to a
and J3, rather than on the mid-span section, subjected to the highest reduced ability to transfer loads across. The pronounced cracking along
bending moments. This was observed on all the specimens, including the interfaces can have also some impact on service limit states. How-
V1 treated with both epoxy and latex based adhesion promoters, ever, interface opening was not accurately measured during the tests,
showing that, even in this situation, the tensile strength of the interfaces and a relation with applied and expected service loads was not estab-
is smaller than that of the weakest concrete. The presence of a casting lished. Thus, conclusions about the service behaviour of beam speci-
interface in regions of non-negligible shear and bending moment leads mens with casting interfaces are hardly possible.
to premature and pronounced cracking along the interface which is not Figs. 10 and 11 show the cracking patterns and failure mechanisms
consistent with the cracking pattern developed in the specimen man- of specimens V1 to V4. With the exception of V3, remaining specimens
ufactured in a single cast. This phenomenon was already described by showed pronounced vertical cracking, in particular along interfaces J2

217
E. Cavaco et al. Engineering Structures 156 (2018) 210–223

Fig. 9. Crack initiation at interfaces J2 and J3: (a) V1; (b) V2; (c) V3; (d) V4.

Fig. 10. Crack pattern at failure of specimen V1.

and J3, subjected to higher bending moments. Both epoxy and latex casting interfaces, the direction of shear cracks in the specimens’ web,
adhesion promoters seem to have nearly insignificant effect preventing between casting interfaces ranged between 38° and 42°, accordingly to
interfaces preferential cracking in specimen V1 (see Fig. 10), which is the assumption considered for the shear strength calculations. Thus,
consistent with the results present in [27], at least for the former agent. and as expected, in all cases, failure mechanism was significantly con-
However, Table 7 shows specimen V1 presenting the lower difference strained by the yield of bending reinforcement, rather than the trans-
between loads and displacements corresponding to cracking of inter- verse reinforcement. Table 7 presents the loads and displacement cor-
faces J2 and J3 and cracking of mid-span section. Thus, a minimal effect responding to the yield of the bending reinforcement, which marks the
of these products in improving interfaces tensile strength can poten- entry on the ultimate state. The maximum and ultimate loads and re-
tially be expected. spective displacements are also shown, including a ductility factor ob-
Less pronounced cracking is observed from the use of dedicated web tained by the ratio between the yielding and the failure displacements.
reinforcement in specimens V2 and V4 (see Figs. 11 and 13). Even so, The Specimen V1 yielded for an applied load/displacement equal to
Fig. 11 shows deep vertical cracking penetration along interface J3, 650 kN/15 mm. However, experimental test could be continued up to
almost similar to interface J2, without web reinforcement. Fig. 12 maximum load of 745 kN by taking advantage of the reinforcement
shows that, the presence of inclined casting interfaces in specimen V3 strain hardening, and increasing displacement up to 81 mm. Although
has had influence on the initial cracking development. In particular, in an identical reinforcement detailing to that of the specimens V2 and V3
interfaces J1 and J2, perpendicular to the expected shear cracks, initial was used in specimen V1, the latter yielded for an applied load corre-
cracks develop along the interface and contrary to the expected direc- sponding to 95% of the formers, probably due to a smaller reinforce-
tion. However, the final cracking pattern seems compatible with a ment lever arm. The specimen presented relevant plastic deformations,
single casted specimen, i.e., shear cracks cross interfaces J1 and J2 in a corresponding to an ultimate load/displacement equal to 725 kN/
perpendicular direction, and the opposite is found in interfaces J3 and 88 mm and a ductility factor equal to 5.9. However, the specimen
J4, these leaved parallel to the expected shear cracks. failure was caused by a friction failure at interface J3, depicted by a
Despite, the local disturbances resulting from the presence of sudden drop of the load displacement diagram (see Fig. 8),

218
E. Cavaco et al. Engineering Structures 156 (2018) 210–223

Fig. 11. Crack pattern at failure of specimen V2.

Fig. 12. Crack pattern at failure of specimen V3.

accompanied by a slip along the interface in the tensioned flange and attached to a stirrup affected by the failure surface, show strains clearly
on the section web, and the formation of an inclined crack towards the beyond the yielding strain. Fig. 14 also shows that friction failure of the
loading plate in the compressed flange (see Fig. 14). This failure mode interface J2 was also imminent, as shown by the observation of some
has already been described in [6]. vertical displacements and the yield of stirrup where strain gage E5 was
Fig. 15 shows the strain records of transverse reinforcement for the attached (see Fig. 15).
specimen V1 on interface J3 vicinity. The records of strain gage E9, As shown in Table 7, the specimen V2 yielded for an applied load/

219
E. Cavaco et al. Engineering Structures 156 (2018) 210–223

Fig. 13. Crack pattern at failure of specimen V4.

Fig. 14. Failure mode of specimen V1.

displacement equal to 685 kN/16 mm. A reduced maximum/yield load research on the basic resisting friction mechanisms [28–30], the in-
ratio of 1.05 was observed in comparison to a ratio of 1.15 observed for terface separation causes a significant reduction of the aggregate in-
the specimen V1. No significant improved global behaviour was ob- terlock and dowel action effects, which is more severe with the for-
served resulting from the strain hardening effect of the reinforcement. mation of a plastic hinge due to an uncontrolled interface opening.
The specimen collapsed for a vertical displacement of 42 mm, thus less Thus, the shear stresses are transferred to the compressed zones of the
than a half of that of V1. The reason for such a reduced ductility (equal interface, causing its failure.
to 2.6) was the friction failure of interface J2 (see Fig. 16), that without Fig. 17 shows the records obtained from strain gages E4 to E9 at-
web reinforcement. The failure surface was identical to that observed in tached to stirrups on the interface J2 and J3 - with the web reinforce-
the specimen V1, including the formation of an inclined tension crack in ment - for the specimen V2. Again, and similarly to what was observed
the compressed flange towards the loading plate. The reason for such an in the specimen V1, the records for the strain gage near to the interface
impaired behaviour can be explained based on the reduced ability of J2, E5, show strains clearly above the strain of yield of the stirrup. The
the interface to transfer loads across it. According to fundamental opposite was found for all the stirrups in interface J3 vicinity, proving

220
E. Cavaco et al. Engineering Structures 156 (2018) 210–223

Fig. 15. Specimen V1 strain records of transverse re-


inforcement at interface J2 and J3.

strength (852 kN), some stirrups near the interface without web re-
inforcement - J2 - recorded strains above the yield of steel. Specimen V3
exhibited the highest ductility factor although it was not possible to
obtain a quantitative measure of it.
The effect of the reduced amount of bending reinforcement of the
specimen V4 can be noted by the lower yielding load (560 kN).
However, a maximum/yield load ratio equal to 1.18 was attained, si-
milar to that achieved by specimen V3. The specimen presented a
ductility factor equal to 6.3 and the failure mechanism consisted on the
failure of the bottom bending reinforcement at the loading section,
shown in Fig. 18. The same figure suggests relative vertical displace-
ment at the interface J2, however failure of the bending reinforcement
was prior to the formation of an inclined tension crack towards the
loading plate, in the compressed flange. Signs of potential interface J3
Fig. 16. Failure mode of specimen V2. slippage were not detected which proves the efficiency of the web re-
inforcement.
the efficiency of the dedicated web reinforcement. At this stage, it is worthy to mention that specimens V1 and V4
Even though the Table 7 and Fig. 8 show that the specimen V2 and exhibited a ductility factor within the range of specimen V3. In the first
V3 presented the same yielding and displacement values, the latter case, the interface failure was the cause of failure and, in the second,
exhibited higher load bearing capacity, corresponding to maximum/ the interface was showing signs of potential failure. On both, if shear
yield load ratio of 1.18. The experimental testing was conducted up to levels were not below those of specimens V2 and V3, probably earlier
the maximum imposed displacement of approximately 95 mm, and then failures could have been observed, compromising also specimens’
stopped because of the maximum capacity of the displacement trans- ductility.
ducer. At this stage, compressed concrete bellow the loading plate had Finally, and by comparing the behaviour of the interfaces J1 and J2
started to crush but no signs of imminent slippage at the interfaces were or J3 and J4 within the same specimen, with the exception of specimen
observed. As the maximum load (808 kN) surpassed the predicted V3 (inconclusive in this matter) it is possible to state that the bending
bending strength (750 kN), and approached the estimated shear moment has a non-negligible effect on the friction strength of concrete-

Fig. 17. Specimen V2 strain records of transverse re-


inforcement at interface J2 and J3.

221
E. Cavaco et al. Engineering Structures 156 (2018) 210–223

Fig. 18. Failure mode of specimen V4.

Table 8
Comparison between analytical and experimental results.

Specimen Analytical Experimental

Bendinga Sheara Frictionb Failure Maximum Ultimate Failure Ductility

b
V1 740 kN 852 kN 831 kN Bending 745 kN 725 kN Friction 5.9
V2b 740 kN 852 kN 803 kN Bending 720 kN 710 kN Friction 2.6
V3a 750 kN 852 kN 790 kN Bending 807 kN 804 kN Bending > 6.3
V4a 585 kN 852 kN 687 kN Bending 660 kN 645 kN Bending 6.3

a
Monolithic.
b
Non-monolithic.

to-concrete interfaces, mainly after the yielding of the bending re- at the interface section); (ii) dedicated web reinforcement transverse to
inforcements. A similar detailing was used on interfaces J1 and J2 or J3 the interface; (iii) interface geometry; (iv) and shear level, on the
and J4, but friction failures, or signs of potential friction failure, were strength and ductility of beam specimens with casting interfaces under
always detected on J2 and J3, which were simultaneously subjected to considerable levels of shear and bending moment.
considerable shear and bending moment, rather than on J1 and J4, The comparison between the behaviours of beam specimens with
under the same shear but negligible bending moment. similar casting interfaces subjected to either shear or shear plus bending
In Table 8, a comparison between specimens’ analytical and ex- moment proves that the latter has a non-negligible effect on their
perimental strengths is presented based on data of Tables 4–7. Analy- friction strength. In general, a tendency for a premature and pro-
tical friction strength respects to minimum values, thus obtained ac- nounced cracking along the interfaces exists, resulting from their re-
cording to Eq. (4). As observed, a bending failure and monolithic duced tensile strength, lower to the weakest concrete strength. The
behaviour were expected for all the specimens, since the theoretical presence of the interfaces strongly conditioned the development of the
shear strength and even the most conservative friction resistance are typical diagonal cracking pattern on the specimens’ web.
above the flexural strength. However, and as discussed, specimens V1 After the formation of a plastic hinge, on the interface section, a
and V2 showed a friction failure and non-monolithic behaviour, for a friction slippage is likely to occur on it, closely followed by the for-
loading test bellow the maximum load applied, but corresponding to a mation of a tension crack on the compression zone towards the loading
higher imposed displacement. It results that the design expressions plates, and ultimately reducing the specimens’ ductility.
(Eqs. (2)–(4)) do not take in consideration the reduction of the interface Both the epoxy and latex based adhesion promoters proved to have
capacity to transfer loads across as a function of its opening. The values a negligible effect on the tensile strength of casting interfaces corro-
presented in Table 8 show that, at least for specimens V1 and V2, they borating previous published research [27]. Although, the specimen
provide non-conservative values for the friction strength. Additionally, treated with these products did not showed relevant reduced ductility,
they do not entirely describe the physical mechanism behind the failure the friction failure of the interface treated with the epoxy based pro-
mechanism, since a full slip along all the interface length was never moter was observed. However, the test conditions and its reduced
observed. number do not allow concluding about the superior performance of the
latex based promoter in relation to the epoxy one.
On the other side, the use of web reinforcement proved to be much
5. Conclusions more efficient. Although from the use of it, the pronounced vertical
cracking along the section web could not be avoided, in particular in
Following the research presented in [6] suggesting the reduced the interface J3 of specimen V2 subjected to significant bending mo-
ductility of RC beams with casting interfaces where plastic hinges are ment, the same was effective in preventing a friction failure in com-
likely to develop, an experimental work was conducted to study the parison with an interface in the same exact conditions but not using it.
behaviour of different and alternative designs for concrete-to-concrete Even if specimen V2 test was interrupted due to the friction failure of
interfaces subjected to a combination of shear and bending moment. the interface without web reinforcement, that crossed by this re-
The main objective was that the proposed designs could result in im- inforcement was not showing apparent signs of an imminent friction
proved global behaviour, with comparable ductility to specimens pro- failure.
duced on a single cast. An almost perfect monolithic behaviour was observed from the
Four beam samples were tested to study the effect of: (i) epoxy and adoption of inclined interfaces, perpendicular (J1 and J2) and parallel
latex based adhesion promoters (to improve tensile strength of concrete

222
E. Cavaco et al. Engineering Structures 156 (2018) 210–223

(J3 and J4) to the expected shear cracks. Although a specimen pro- [11] Kahn LF, Mitchell AD. Shear friction tests with high-strength concrete. Struct J
duced in a single cast was not tested, the only difference believed to be 2002;99(1):98–103.
[12] Papanicolaou CG, Triantafillou TC. Shear transfer capacity along pumice aggregate
found was the earlier cracking of sections matching the interfaces J2 concrete and high-performance concrete interfaces. Mater Struct
and J3, prior to the cracking of the mid-span section subjected to a 2002;35(4):237–45.
higher bending moment. The initial crack development also reported [13] Gohnert M. Horizontal shear transfer across a roughened surface. Cem Concr
Compos 2003;25(3):379–85.
the interfaces existence, however the specimen global behaviour was [14] Santos PMD, Júlio ENBS, Silva VD. Correlation between concrete-to-concrete bond
similar to monolithic, as the shear cracks were able to cross interfaces strength and the roughness of the substrate surface. Constr Build Mater
J1 and J2 and to develop parallel to interfaces J3 and J4. 2007;21(8):1688–95.
[15] Mansur MA, Vinayagam T, Tan K-H. Shear transfer across a crack in reinforced
The results also showed that after the formation of a plastic hinge on high-strength concrete. J Mater Civ Eng 2008;20(4):294–302.
the interface section, the likelihood of having a friction failure at the [16] Santos PMD, Júlio ENBS. Development of a laser roughness analyser to predict in
interface strongly depends on the existing shear level. Although some situ the bond strength of concrete-to-concrete interfaces. Mag Concr Res
2008;60(5):329–37.
level of friction slip was observed along specimen V4 web, that sub-
[17] Júlio ENBS, Dias-da-Costa D, Branco FAB, Alfaiate JMV. Accuracy of design code
jected to a lower testing and shear loads, the specimen V4 test could be expressions for estimating longitudinal shear strength of strengthening concrete
carried out until the physical break of the longitudinal reinforcement, overlays. Eng Struct 2010;32(8):2387–93.
without experiencing a friction failure. Unfortunately, the experimental [18] Santos PMD, Júlio ENBS. A state-of-the-art review on shear-friction. Eng Struct
2012;45:435–48.
campaign was not extensive enough to allow a definition of a critical [19] Randl N. Design recommendations for interface shear transfer in fib Model Code
shear level to clarify this point. 2010. Struct Concr 2013;14(3):230–41.
[20] Almeida T. Ensaios Experimentais de Ligações Pilar/Viga Pré-Fabricada para
Acções Cíclicas MSc thesis Portugal: Instituto Superior Técnico, Universidade de
Acknowledgements Lisboa; 2014
[21] Ceia F, Raposo J, Guerra M, Júlio E, de Brito J. Shear strength of recycled aggregate
The authors would like to acknowledge Pavilis – Pré-fabricação S.A. concrete to natural aggregate concrete interfaces. Constr Build Mater
2016;109:139–45.
[22] Simões T, Octávio C, Valença J, Costa H, Dias-da-Costa D, Júlio E. Influence of
References concrete strength and steel fibre geometry on the fibre/matrix interface. Compos
Part B Eng 2017;122:156–64.
[23] Mattock AH, Hawkins NM. Shear transfer in reinforced concrete-recent research.
[1] Birkeland PW, Birkeland HW. Connections in Precast Concrete Construction. J Am
Pci J 1972(17).
Concr Inst 1966;63(3):345–68.
[24] Mattock AH, Johal L, Chow H. Shear transfer in reinforced concrete with moment or
[2] E. C. for S. CEN, EN 1992-1-1, Eurocode 2: Design of Concrete Structures: Part 1–1:
tension acting across the shear plane. Precast Concr Inst J 1975;20(4).
General Rules and Rules for Buildings; 2004.
[25] Sørensen JH, Hoang LC, Olesen JF, Fischer G. Test and analysis of a new ductile
[3] ACI Committee 318, Building Code Requirements for Structural Concrete and
shear connection design for RC shear walls. Struct Concr 2017;18(1):189–204.
Commentary (ACI 318–14). American Concrete Institute; 2014.
[26] Rodrigues RV, Muttoni A, Ruiz MF. Influence of shear on rotation capacity of re-
[4] Walraven Big-van, et al. Fib model code for concrete structures 2010. Ernst Sohn:
inforced concrete members without shear reinforcement. Struct J
Wiley; 2013.
2010;107(05):516–25.
[5] Anderson TL. Fracture mechanics: fundamentals and applications. CRC Press; 2017.
[27] Júlio ENBS, Branco FaB, Silva VD. Concrete-to-concrete bond strength: influence of
[6] Cavaco E, Camara J. Experimental research on the behaviour of concrete-to-con-
an epoxy-based bonding agent on a roughened substrate surface. Mag Concr Res
crete interfaces subjected to a combination of shear and bending moment. Eng
2005;57(8):463–8.
Struct 2017;132:278–87.
[28] Walraven JC. Fundamental analysis of aggregate interlock. J Struct Div
[7] Walraven J, Frenay J, Pruijssers A. Influence of concrete strength and load history
1981;107(11):2245–70.
on the shear friction capacity of concrete members. PCI J 1987;32(1):66–84.
[29] Vintzēleou E, Tassios T. Mechanisms of load transfer along interfaces in reinforced
[8] Tsoukantas S, Tassios T. Shear resistance of connections between reinforced con-
concrete: prediction of shear force vs. shear displacement curves. Studi E Ric.
crete linear precast elements. ACI Struct J 1989;86(3).
Perfez. Costr. Cem. Armato Frat. Pasenti 1985(7):121–61.
[9] Randl N. Investigations on transfer of forces between old and new concrete at dif-
[30] Vintzēleou E, Tassios T. Mathematical models for dowel action under monotonic
ferent joint roughness Ph D. thesis Austria: University of Innsbruck; 1997.
and cyclic conditions. Mag Concr Res 1986;38(134):13–22.
[10] Mattock AH. Shear friction and high-strength concrete. ACI Struct J 2001;98(1).

223

Das könnte Ihnen auch gefallen