8th fib PhD Symposium in Kgs.

Lyngby, Denmark June 20 – 23, 2010

Assessment of the Shear Strength between

Concrete Layers
PEDRO M. D. SANTOS (a) AND EDUARDO N. B. S. JÚLIO (b)
(a)
Research Assistant, ISISE, Dept. of Civil Engineering, Polytechnic Institute of Leiria,
School of Technology and Management, Campus 2 – Morro do Lena – Alto do Vieiro,
2411-901 Leiria, Portugal
[email protected]
(b)
Professor, ISISE, Dept. of Civil Engineering, University of Coimbra, Faculty of Sciences
and Technology, Rua Luís Reis Santos, Polo II, 3030-788 Coimbra, Portugal
[email protected]

Abstract
Concrete-to-concrete interfaces are present in new and existing RC structures. Precast
members with cast-in-place parts and the repair and rehabilitation of existing concrete
members are typical examples.
The behaviour of RC composite members is highly influenced by the surface conditions of the
interface. To improve the bond strength between the substrate and the added concrete layer it
is common to increase the roughness of the substrate surface by the means of concrete
removal techniques, such as sand-blasting or hydrodemolition. Other two parameters with a
significant influence are the curing conditions and the material strength.
Current design codes present design expressions for the assessment of the longitudinal shear
strength of concrete-to-concrete interfaces. These expressions are empirically determined and
based on the shear-friction theory, covering a large number of situations from normal density
concrete to lightweight and high strength concretes.
Some drawbacks can be pointed to these design expressions: 1) the evaluation of the surface
roughness is purely qualitative (based on a visual inspection) and, therefore, subjected to
human error; 2) the curing conditions of both concrete parts are not taken into consideration
and, therefore, the differential shrinkage is neglected; and 3) the differential stiffness, due to
the difference between Young modulus of both concrete layers, is not addressed either.
For the exposed reasons, it can be stated that current design expressions need improvements
to increase their accuracy. This paper describes a research study that aims to add a
contribution to the development of such design expressions. The influence of the surface
roughness, differential shrinkage and differential stiffness was investigated. Modifications to
the current shear-friction provisions of Eurocode 2 are proposed.

1. Introduction
The bond strength at the interface between concrete layers cast at different ages is important
to ensure the monolithic behaviour of RC composite members. Precast beams with cast-in-
place slabs and repair and strengthening of existing concrete structural members, such as
bridge decks, by adding a new concrete layer are typical examples of RC composite members.
Current design codes [1, 2, 3] present design expressions for the assessment of the
longitudinal shear strength of concrete-to-concrete interfaces. These expressions are based on
the shear-friction theory and the shear strength is evaluated considering basically four

1
8th fib PhD Symposium in Kgs. Lyngby, Denmark June 20 – 23, 2010

parameters: a) compressive strength of the weakest concrete; b) normal stress at the interface;
c) shear reinforcement crossing the interface; and d) roughness of the substrate surface.
A qualitative evaluation of the surface roughness, based on a visual inspection, is currently
adopted by all design codes. It is common to classify the surface as very smooth, smooth,
rough or very rough or simply as intentionally roughened or not intentionally roughened.
Typical finishing treatments of concrete surfaces are usually linked to this classification and
the values of two coefficients, friction and cohesion, are given to be adopted in the design
expressions. This approach is clearly inaccurate because it is highly influenced by the
technician opinion and, therefore, subjected to human error.
Since no method or device is specified by design codes to help the designer in the
roughness classification it is common to use the Sand Patch Test [4] or the Concrete Surface
Profiles [5]. Although simple, the use of both methods presents major drawbacks. The first is
only applicable to top horizontal surfaces while the second is purely qualitative.
Design codes do not take into account the curing conditions of both concrete parts.
Therefore, the differential shrinkage is neglected. The differential stiffness, due to the
difference between Young modulus of both concrete layers, is not addressed either. However,
both parameters have a significant influence because they can create additional stresses at the
interface.
For all these reasons, current design expressions need improvements to increase their
accuracy. This research study aims to add a contribution to the development of such design
expressions. The influence of the surface roughness and differential shrinkage and stiffness
was investigated. A new optical measuring device [6] was specifically developed to
characterize the roughness of concrete surfaces. A full in situ non-destructive methodology is
proposed for the assessment of the bond strength of concrete-to-concrete interfaces.
Modifications to the current shear-friction provisions of Eurocode 2 [2] are proposed.

2. Shear-friction
The shear-friction theory assumes that the shear strength of a concrete-to-concrete interface
subjected simultaneously to shear and compression forces is ensured by friction only. A
simple “saw-tooth model” is usually adopted to exemplify the basic principles of this theory,
Figure 1. This design philosophy assumes that, due to relative slippage between old and new
concrete layers, the interface crack width increases, the steel reinforcement yields in tension
thus compressing the interface and the shear forces are transmitted by friction.
s 




s
Figure 1: Shear-friction.

2
8th fib PhD Symposium in Kgs. Lyngby, Denmark June 20 – 23, 2010

Several design expressions were proposed to predict the ultimate longitudinal shear stress
at the concrete-to-concrete interface (vu). The five most significant contributions are presented
in Table 1.
Table 1: Shear-friction milestones.
Researcher(s) Year Design expression

Birkeland and Birkeland [7] 1966 vu   f y

Mattock and Hawkins [8] 1972 vu  1.38  0.8  n   f y 

Loov [9] 1978 vu  k fc  n   f y 

v u  C1   f y 
C2

Walraven et al. [10] 1987 C1  0.822 f c

0.406

C 2  0.159 f c
0.303

vu  cf c    n   kf y    f y fc
1 3
Randl [11] 1997

Birkeland and Birkeland [7] proposed the design expression currently known as the
“shear-friction expression”. These researchers suggested the following values for the
coefficient of friction: a) µ = 1.7, for monolithic concrete; b) µ = 1.4, for artificially
roughened construction joints; and c) µ = 0.8 to 1.0, for ordinary construction joints and for
concrete-to-steel interfaces. Later, Mattock and Hawkins [8] proposed an improved design
expression, known as the “modified shear-friction expression”, which includes a constant due
to cohesion. The coefficient of friction is considered constant and equal to 0.8.
Loov [9] was the first to explicitly include the concrete strength. Walraven et al. [10]
proposed a non-linear function to predict the shear strength of initially cracked interfaces. An
innovative “sphere model” was developed to analyse the interaction between the aggregates,
the binding paste and the interface zone.
Randl [11] proposed the first design expression that explicitly includes the contribution of:
cohesion, related with the interlocking between aggregates; friction, related with normal
stresses to the interface and the longitudinal relative slip between concrete parts; and dowel
action, related with the deformation of the shear reinforcement crossing the interface.
InTable 2 are also presented the design expressions of three major design codes for RC
structures, which are mainly derived from the first two expressions presented in Table 1
[7, 8].
Table 2: Shear-friction provisions of design codes.
Design Code Year Design expression

CEB-FIP Model Code 1990 [1] 1990 vu  cfctd    n   f y 

Eurocode 2 [2] 2004 vu  cf ctd   n   f y   sin   cos  

ACI 318 [3] 2008 vu   f y   sin   cos  

In these expressions (Table 1 and Table 2), µ is the coefficient of friction; ρ is the
reinforcement ratio; fy is the yield strength of the reinforcement; σn is the normal stress acting
on the interface due to external loading; k is a constant (Loov’s expression); fc is the concrete
3
8th fib PhD Symposium in Kgs. Lyngby, Denmark June 20 – 23, 2010

compressive strength; c is the coefficient of cohesion; k is a coefficient of efficiency related

with the reinforcement (Randl’s expression); α is a coefficient for dowel action (Randl’s
expression) or the angle between the shear reinforcement and the shear plane; fctd is the tensile
strength of the weakest concrete.
Besides the format of the design expressions, the main difference between codes is the
roughness classification and the proposed coefficients of friction and cohesion for the same
surface condition. This incongruence is probably the main drawback of the design expressions
and of the design codes referred to.

3. Experimental Study
A large experimental study was developed to investigate the influence of the surface
roughness, differential shrinkage and differential stiffness on the bond strength of the
concrete-to-concrete interface.
A Portland cement type I 52.5R was adopted to achieve a C50/60 class concrete. Four
different aggregates were selected: fine sand; coarse sand; fine limestone crushed aggregates;
and coarse limestone crushed aggregates. A commercial admixture was also adopted.
Two bond tests were adopted to assess the bond strength of the interface, respectively in
shear and in tension: a) the slant shear test, using prismatic specimens with 150×150×450mm3
and the shear plane at 30º to the vertical; and b) the splitting test, using cubic specimens of
150mm with the interface at middle height.
Five different situations were considered for the interface surface between the substrate
and the added concrete layer, Figure 2. Left as-cast (LAC) against steel formwork was
considered to serve as reference situation. To increase the roughness of hardened concrete, the
interface surface was subjected to the following treatments: a) wire-brushing (WB); b) sand-
blasting (SAB); and c) shot-blasting (SHB). Hand-scrubbing (HS), a technique commonly
used in the precast industry to increase the roughness of fresh concrete surfaces, also known
as “raking”, was equally considered. The 2D-LRA method [6] was adopted to characterize the
surface roughness and to correlate this with the bond strength of the interface.
Two curing conditions, with a set of specimens cured inside the laboratory and another set
cured outside, therefore directly exposed to the environmental conditions such as solar
radiation, rain and wind, were considered. The time gap between casting the substrate and the
added concrete layer was considered equal to 28, 56 and 84 days, to maximize the influence
of differential shrinkage.

a) b) c) d) e)
Figure 2: Surface conditions: a) left as-cast; b) wire-brushing; c) sand-blasting; d) shot-
blasting; and e) hand-scrubbing.

4. Conclusions
The review of the state of the art showed that the surface roughness has a significant influence
on the bond strength of concrete-to-concrete interfaces. It can be stated that in all design
4
8th fib PhD Symposium in Kgs. Lyngby, Denmark June 20 – 23, 2010

codes this parameter is qualitatively assessed, although each of these presents its own
classification. It is recognized today [11] that the load transfer mechanism at the concrete-to-
concrete interface is due to cohesion, friction and dowel action but the latter is not explicitly
considered in any design codes. However, since all these design expressions are empirically
based, dowel action is implicitly considered in cohesion and friction terms.
Common to all design codes is the absence of any provision related with the curing
conditions and, therefore, with the differential shrinkage between concrete parts. Also
neglected is the difference between the substrate and added concrete Young modules and thus
the differential stiffness between concrete parts.
In relation to roughness, it is proposed that a quantitative methodology be adopted to
avoid the subjective assessment proposed in all design codes. The authors proposed an
innovative and non-destructive method, the 2D-LRA method [6], to predict the bond strength
of concrete-to-concrete interfaces. This new method proved to be effective, since it is possible
to obtain 2D profiles of the surface texture; to compute texture parameters from these; and to
correlate the latter with the bond strength of the concrete-to-concrete interface, both in shear
and in tension, with high coefficients of correlation. Moreover, it was demonstrated that the
proposed new method presents all the advantages, with even higher accuracy, and overcomes
all the disadvantages of existing methods [4, 5, 12].
Based in recent research studies [13] and adopting the design expression of Eurocode 2
[2], the authors propose that the coefficients of cohesion and friction be predicted using the
following expressions:
1.062 Rvm
0.145

cd  (1)
 coh
1.366 Rvm
0.041

d  (2)
 fr
where cd is the design coefficient of cohesion; µd is the design coefficient of friction; Rvm is
the Mean Valley Depth [13] of the primary profile of the surface in millimetre; γcoh is the
partial safety factor for the coefficient of cohesion; and γfr is the partial safety factor for the
coefficient of friction.
The proposed expressions were obtained by adjusting a power function to the
experimental values of the coefficients of cohesion and friction, Figure 3, determined for the
five different surface conditions considered: left as-cast; wire-brushing; sand-blasting; shot-
blasting and hand-scrubbing. Based in the coefficient of variation of both coefficients, the
authors propose the values of 2.6 and 1.2 for the partial safety factors of the coefficients of
cohesion and friction, respectively.
1.6
SHB HS
1.4
WB HS
SAB
Coefficient of cohesion
Coefficient of friction

1.2
LAC SHB
1.0 Coeff. of cohesion
WB SAB Power Regression
0.8
LAC for Coeff. of cohe-
0.6 sion
Coeff. of friction
0.4 Power Regression
for Coeff. of friction
0.2

0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Mean Valley Depth, Rvm (mm)

Figure 3: Correlation between the Mean Valley Depth (Rvm) and the coefficients of cohesion
and friction.
5
 8th fib PhD Symposium in Kgs. Lyngby, Denmark June 20 – 23, 2010

It also proved that differential shrinkage and differential stiffness can have a significant
influence on the shear strength of the interface between concretes cast at different times [13].
These effects should at least be mentioned in codes and analyzed for each specific situation.

Acknowlegments
The authors acknowledge the financial support of the Portuguese Science and Technology
Foundation (FCT), PhD Grant number SFRH/BD/25510/2005. Acknowledgements are
extended to the companies Maprel – Empresa de Pavimentos e Materiais Pré-Esforçados Lda,
Sika Portugal SA, AFAssociados – Projectos de Engenharia SA, Weber Cimenfix, Cimpor –
Cimentos de Portugal, Betão-Liz, Euro-Planning – Engenharia & Gestão Lda, TrueGage and
SYCODE also for their financial support.

References
[1] CEB-FIP Model Code, Comité Euro-International du Béton, Secretariat Permanent, Case
Postale 88, CH-1015 Lausanne, Switzerland, 437 p., 1990.
[2] Eurocode 2, “Design of concrete structures - Part 1-1: General rules and rules for
buildings,” European Committee for Standardization, Avenue Marnix 17, B-1000
Brussels, Belgium, 225 p., 2004. (with corrigendum dated of 16 January 2008)
[3] ACI 318, “Building code requirements for structural concrete (ACI 318-08) and
commentary,” American Concrete Institute, PO Box 9094, Farmington Hills, MI 48333-
9094, USA, 471 p., 2008.
[4] ASTM E 965, “Standard test method for measuring pavement macrotexture depth using a
volumetric technique,” American Society for Testing Materials, 100 Barr Harbor Dr.,
West Conshohocken, PA 19428, USA, 3 p., 2001.
[5] ICRI (International Concrete Repair Institute), “Selecting and specifying concrete surface
preparation for sealers, coatings, and polymer overlays,” Technical Guideline No. 03732,
Des Plaines, Illinois, USA, 41 p., 1997.
[6] P. Santos and E. Júlio, “Development of a laser roughness analyser to predict in situ the
bond strength of concrete-to-concrete interfaces,” Magazine of Concrete Research, vol.
60, no. 5, pp. 329-337, 2008.
[7] P.W. Birkeland and H.W. Birkeland, “Connections in precast concrete construction,”
Journal of the American Concrete Institute, vol. 63, no. 3, pp. 345-368, 1966.
[8] A.H. Mattock and N.M. Hawkins, “Shear transfer in reinforced concrete – recent
research,” Journal of the Precast/Prestressed Concrete Institute, vol. 17, no. 2, pp. 55-
75, 1972.
[9] R.E. Loov, “Design of precast connections,” Paper presented at a seminar organized by
Compa International Pte, Ltd., Singapore, 8 p., 1978.
[10] J. Walraven, J. Frénay and A. Pruijssers, “Influence of concrete strength and load history
on the shear friction capacity of concrete members,” Journal of the Precast/Prestressed
Concrete Institute, vol. 32, no. 1, pp. 66-84, 1987.
[11] N. Randl, “Investigations on transfer of forces between old and new concrete at different
joint roughness,” PhD thesis, University of Innsbruck, 379 p., 1997. (in German)
[12] P. Santos, E. Júlio and V.D. Silva, “Correlation between concrete-to-concrete bond
strength and the roughness of the substrate surface,” Construction and Building
Materials, vol. 21, no. 8, pp. 1688-1695, 2007.
[13] P.M.D. Santos, “Assessment of the Shear Strength between Concrete Layers,” PhD
thesis, University of Coimbra, 338 p., October 2009.

View publication stats