Engineering Structures 167 (2018) 108–120

Contents lists available at ScienceDirect

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

Assessment of the flexural behavior of reinforced concrete beams T

strengthened with concrete jackets
⁎
M. Monir A. Alhadid, Maged A. Youssef
Department of Civil and Environmental Engineering, Western University, London, ON N6A5B9, Canada

A R T I C LE I N FO A B S T R A C T

Keywords: Analysis of continuous jacketed Reinforced Concrete (RC) beams requires accounting for the nonlinear behavior
Reinforced concrete of the interface and the materials as well as redistribution of moments. This kind of analysis is complex and
Continuous require an advanced level of knowledge and experience to perform. Engineers need simplified yet robust tools to
Jacketing practically predict the actual behavior of jacketed RC beams. In the current practice, slip is neglected in the
Slip
analysis and monolithic behavior is assumed for the jacketed section, which result in higher estimates of stiffness
Interface
Monolithic factors
and/or capacity. This paper provides a simplified method to analyze continuous jacketed RC beams taking into
Inelasticity account the interfacial slip distribution and the actual nonlinear behavior of both concrete and steel. An iterative
Flexure calculation algorithm is developed to determine the moment–curvature curves of a jacketed beam at different
sections. The developed method allows the evaluation of interfacial slip and shear stress distributions in ductile
reinforced concrete beams. The developed method is utilized to conduct an extensive parametric study, which
resulted into modification factors that can be used to calculate the capacity and deformations of a strengthened
beam considering the interfacial slip.

1. Introduction investigations that have been performed to address the influence of

jacketing schemes, geometrical characteristics, mechanical properties
The need to strengthen Reinforced Concrete (RC) structures emerges and interfacial treatment on the flexural behavior of determinate
from various reasons, such as new safety requirements, a change of jacketed RC structural members. For instance, Altun [2] and Bousias
structure occupancy, an incorrect design calculations and/or degrada- et al. [3] examined the effect of RC jacketing on the mechanical per-
tion of materials with time. Jacketing is one of the widely spread pro- formance of statically-determinate RC beams considering the load-dis-
cedures to strengthen and repair RC beams. It comprises the addition of placement behavior, ultimate load, ductility and toughness. Other re-
concrete layers that are usually reinforced with longitudinal steel bars, searchers [1,4] investigated the significance of surface preparation of
stirrups, welded wire mesh or various kinds of fibrous materials. concrete members before applying the new concrete jacket. The use of
In the current practice, monolithic action is assumed between the fiber reinforced cementitious composites as an alternative to adding
original beam and the attached jacket. This implies that the internal steel reinforcement within the jacket has been addressed by other stu-
stresses developed in both substrates due to the applied loads are dis- dies [5–10]. In addition, the impact of using shear studs to further at-
tributed among them assuming infinite interfacial slip stiffness. This tach the existing beam with the additional concrete layers has been
assumption may result in higher estimates of stiffness and/or capacity investigated by Shehata et al. [11]. Furthermore, the influence of
depending on the geometrical properties and interfacial surface treat- varying the method of applying the jacket on site, such as shotcrete or
ment. The actual behavior of typical jacketed beams is partially com- cast-in-place concrete, have been considered by many researchers
posite and depends on the frictional resistance between the surfaces and [12–14].
the presence of steel anchors connecting the two substrates [1]. This Experimental and numerical studies related to strengthening in-
implies that the analysis of jacketed beams requires a knowledge of the determinate RC beams using concrete jackets is scarce in literature. At
nonlinear behavior of the interface as well as the nonlinear properties the time of writing, the only available relevant experimental work was
of both concrete and the embedded steel bars at each loading step along performed by Cheong and MacAlevey [15]. The rather extensive use of
the beam. indeterminate RC beams in building structures and bridges requires
Literature is ample with experimental programs and numerical further research regarding the influence of partial composite action on

⁎
Corresponding author.
E-mail addresses: majjanal@uwo.ca (M.M.A. Alhadid), youssef@uwo.ca (M.A. Youssef).

https://doi.org/10.1016/j.engstruct.2018.04.026
Received 9 February 2017; Received in revised form 5 March 2018; Accepted 9 April 2018
0141-0296/ © 2018 Elsevier Ltd. All rights reserved.
M.M.A. Alhadid, M.A. Youssef Engineering Structures 167 (2018) 108–120

Fig. 1. Continuous beam loading scheme and reinforcement configuration.

their flexural performance. the anchors crossing the interface. The former mechanism represents
This paper is a continuation of an ongoing research [16], which the concrete contribution; whereas the second case represents the in-
aims at proposing a simplified method to capture the influence of in- fluence of dowel action. The concrete contribution (vc) is determined in
terfacial slip on the moment-curvature (M-φ) and load-deflection (P-Δ) view of Tassios and Vintzeleou [1] empirical model as a function of the
relationships of jacketed continuous RC beams. This is achieved by lateral slip (S), ultimate slip value at the onset of frictional mechanism
performing nonlinear analysis in view of the material properties and failure (Scu) and ultimate frictional capacity of the interface (vcu). The
interfacial behavior. A calculation algorithm is proposed to determine overall interfacial shear stress (τ) corresponding to any slip (S) value
the slip distribution along the beam length and to obtain the corre- can be obtained as the summation of concrete contribution and dowel
sponding M-φ diagram at both the sagging and hogging moment re- action contribution for given material properties and interfacial surface
gions. This analysis procedure is sensitive to the bending moment dis- condition. A detailed description of this calculation procedure con-
tribution along the beam; therefore, the concept of moment sidering simply supported beams is provided by Alhadid and Youssef
redistribution in indeterminate beams is illustrated and considered in [16].
the analysis. The validated model is used to perform a parametric study
aiming at examining the flexural behavior of the strengthened beams.
4. Assumptions
Finally, a regression analysis is performed to propose slip modification
factors that can be used to obtain the actual M-φ diagram of continuous
Assumptions considered in the current study encompasses the fol-
RC beams considering interfacial slip. The scope of the proposed work is
lowing:
limited to ductile RC beams by considering sufficient reinforcement to
prevent brittle modes of failure.
(1) Plane sections remain plane after deformation, implying that shear
deformations are small relative to bending deformations.
2. Material models (2) Perfect bond exists between steel reinforcement and the sur-
rounding concrete material. Thus, strain in both concrete and steel
The stress-strain relationship of concrete in compression is con- bars at the same location is identical.
sidered in view of Scott et al.’s model [17] due to its simplicity and (3) The failure criterion of the composite beam is defined by crushing
robustness. The tensile capacity of concrete is assumed to drop after of the extreme compression fiber at a concrete ultimate strain (εcu)
reaching the cracking point. of 0.0035 [21].
The constitutive relationship of the embedded steel bars is expressed (4) The original RC beam and the added concrete layer are considered
according to the model reported by Karthik and Mander [18] that was to deform by the same curvature through the beam length [20,23].
derived in view of the general formula proposed by Ramberg and Os- (5) The interfacial shear stress distribution within each region is as-
good [19]. It conveniently combines the initial elastic response, yield sumed to vary as a cubic function with the distance from the zero
plateau and strain hardening stages in a single rigorous form to model moment section [22].
the actual behavior of steel reinforcement. The value of the strain
hardening strain (εsh) is set equal to the yield strain (εy) and the strain
5. Typical jacketed section
hardening modulus (Esh) is taken as 1% of the Young's modulus of
elasticity (Es).
The developed model is applicable to analyze symmetric continuous
RC beams subjected to either uniform or concentrate loads. Fig. 1 shows
3. Interfacial shear stress (τ) and slip (S) relationship the geometry and reinforcement details of a typical continuous beam
that will be used for discussion throughout the chapter. The main steel
The shear transfer mechanism is activated by the frictional re- reinforcement in the positive and negative moment regions are assumed
sistance between the contact surfaces and the axial forces developed in to be 20% and 40% of the balanced steel reinforcement ratio,

109
M.M.A. Alhadid, M.A. Youssef Engineering Structures 167 (2018) 108–120

(a) Actual beam

(b) Structure coordinate numbers

Fig. 2. Propped cantilever analytical model.

respectively. The compression steel reinforcement is 2–10 M bars. The Assuming a propped cantilever model for each span, the analyzed
amount of jacket reinforcement is assumed as 10M bars placed in one segment within the positive moment zone is taken from the pinned
layer at the maximum spacing provided by CSA 23.3-14 [21]. One half support to the point of maximum bending; whereas, for the negative
of the core and jacket steel bars from the hogging moment region are moment zone, this segment is taken from the point of contraflexure to
assumed to extend throughout the beam. the point of maximum negative bending moment at the fixed end.
Geometry and loading scheme of the continuous beam are assumed The proposed analysis method comprises two main stages. In the
to be symmetric about the intermediate support. Thus, one span of the first one, an iterative sectional analysis procedure is performed at dif-
beam can be modeled as a propped cantilever as shown in Fig. 2(a). ferent load levels only at the maximum sagging and hogging moment
This span is assumed to be composed of several members rigidly con- sections. This results in determining the maximum slip strain (Δεmax) at
nected at their ends as illustrated in Fig. 2(b). Each segment has a de- these locations and the corresponding slip strain (Δε) and slip (S) at the
fined length (Li) and a distinct flexural rigidity (EIi). The segment length other segments along the span. In the second stage, sectional analysis is
is set at about 50 mm, which was found to enhance the accuracy based conducted directly at the remaining segments taking into account the
on a preliminary sensitivity analysis. Δε distribution evaluated from the first analysis step for each segment.
The expected trends of the moment-curvature diagrams in both the The slip distribution is obtained while satisfying the equilibrium and
positive and negative moment regions are shown in Fig. 3. The trend for compatibility conditions at each segment. Details about the mentioned
the positive moment section is characterized by three points; namely, steps are given below.
the yielding of jacket reinforcement, yielding of the core reinforcement The primary challenges for the proposed calculation algorithm are
and crushing of concrete. The trend of the negative moment section is prediction of the slip distribution along the interface and determination
defined by yielding of the core reinforcement and crushing of concrete. of the moment-curvature relationships for the beam segments shown in
Fig. 2. Alhadid and Youssef [16] have proposed a calculation algorithm
to determine these relationships in jacketed RC simply supported beams
6. Proposed calculation algorithm considering slip effect. A summary of the procedure is provided in
Section 6.1 showing the main changes used for analyzing continuous
Slip, and consequently shear stress, reach their maximum value at beams. Sectional analysis procedure to determine the equilibrium
the point of zero moment and fade away as they approach the max- conditions is described in Section 6.2. The influence of moment redis-
imum bending moment section. In continuous RC beams, each span can tribution becomes substantial in the prediction of slip distribution along
be divided into positive and negative moment zones as indicated in continuous beams and is discussed in Section 6.3. An equivalent cur-
Fig. 4. To obtain the complete slip distribution along the span, the vature distribution is then obtained based on the load-deflection re-
analysis procedure is carried out individually for each of the two zones. lationship of the actual curvature distribution considering slip effect as
illustrated in Section 6.4.

Yielding of Core Tensile Steel

6.1. Moment-curvature at maximum moment sections

Concrete Crushing
For each moment zone, the average value of interfacial shear stress
Yielding of Core Tensile Steel (τavg) at any load level can be calculated assuming a direct relationship
with the maximum slip strain (Δεmax) located at the maximum moment
Yielding of Jacket Tensile Steel
Moment

section [22–24]. Therefore, the average shear stress can be given ac-
cording to the expression (τavg = γ1 γ2 ks Δεmax L′) in terms of secant in-
terfacial stiffness, ks (N/mm3); the ratios (γ1 = τavg/τmax) and
(γ2 = Δεavg/Δεmax); the average slip strain (Δεavg) from point of zero
moment to maximum positive or negative moment; and the corre-
sponding length, L' (m).
For each of the two moment zones, the analysis procedure to de-
termine interfacial slip distribution is carried out at each applied load
Curvature level (i.e. assumed applied curvature value) until failure occurs. Firstly,
Positive Moment Section Negative Moment Section initial values of the secant interfacial stiffness (ks) and the shear stress
distribution ratios (γ1 and γ2 ) are assumed. Then, for the total curvature
Fig. 3. Moment-curvature diagrams for positive and negative moment sections. (φ) value of the current load increment, two equilibrium conditions are

110
M.M.A. Alhadid, M.A. Youssef Engineering Structures 167 (2018) 108–120

(a) propped cantilever idealization

(b) bending moment diagram showing

the point of zero moment

(c) bending moment diagram showing

the point of zero moment

(d) anticipated deflection shape of the

propped cantilever

Fig. 4. Bending moment and deflection profile of the propped cantilever model.

applied at the maximum moment sections: (1) equilibrium between the stress distribution ratios (γ1 and γ2 ) are calculated and compared to the
internal forces; and (2) equilibrium between the resultant axial forces at initially assumed values. The analysis continues if they are equal with a
one side of the interface and the resultant shear force acting along the tolerance of 1%, otherwise the whole procedure is repeated with the
interface. Hence, the moment (M) and maximum slip strain (Δεmax) at new calculated values.
the maximum moment sections corresponding to the current curvature If the beams are subjected to initial loading prior to jacketing, then a
value (φ) are obtained. After that, bending moment diagram is con- preliminary sectional analysis on the unjacketed sections has to be
structed along the span assuming uniform load and considering the carried out first to obtain the resulting moment-curvature curve and
obtained maximum moment values. Next, the slip strain (Δε) distribu- strain profile at each beam segment. These diagrams are then included
tion is determined along the span with respect to the location of each as an input in the calculation algorithm of the jacketed beam to obtain
segment as shown in the proposed Eq. (1). its full behavior at different loading stages before and after jacketing.
xj
Δε(i,j) = Δε(m,1) ⎛ ⎞
⎝ L′ ⎠ (1)
6.2. Sectional analysis in jacketed sections
where i is the load step number, j is the segment number and m is the
load step number that produces a bending moment in the mid-span The sectional analysis procedure [25] is implemented to analyze the
segment equals to the moment applied at segment j. Once the slip strain jacketed sections. The upper limit for the thickness of each layer is
(Δε) distribution along the interface is established, both the slip (S) and taken as 0.5 mm as it results in a better accuracy. At every loading step,
the shear stress (τ) in each segment is obtained from Eqs. (2) and (3), an incremental curvature is applied and the strain at each strip in both
respectively. the concrete core and the jacket is calculated based on its location from
n=j the centroid (yi) of the jacketed section. The kinematic and compat-
S(i,j) = ∑ [(Δε(i,n) )(x j )] ibility conditions are considered in view of the corresponding material
n=1 (2) stress-strain relationships and Eq. (4), which relates the incremental
applied moment (ΔM) and axial load (ΔP) to the incremental curvature
τ(i,j) = ks S(i,j) (3)
(Δφ) and axial strain (Δεa) by a defined stiffness matrix. In this equa-
Having obtained the slip distribution for both moment zones, con- tion, n represents the number of discrete layers, Ei is the elastic modulus
tinuity conditions is checked at the point of contraflexure to ensure it is of layer i, Ai is the area of layer i, subscript (c) represents concrete core
satisfied by calculating the error between the obtained slip (S) from the and subscript (J) represents concrete jacket.
sagging moment zone and the hogging moment zone. The procedure is
repeated if the error is more than 1% by adjusting the slip strain (Δε) at
all segments and repeating the analysis to check equilibrium and
compatibility conditions. Finally, based on the obtained slip and shear
stress distributions, the secant interfacial stiffness (ks) and the shear

111
M.M.A. Alhadid, M.A. Youssef Engineering Structures 167 (2018) 108–120

decrease gradually with the applied load until it equates the stiffness at
the positive moment section. In this case, the bending moment diagram
obtained from stiffness analysis will be identical to that obtained from
elastic structural analysis. As the load keeps increasing, the hogging-to-
sagging stiffness ratio further decreases resulting in a shift of the point
of zero moment towards the intermediate support as more proportion of
the additional load is carried by the sagging moment region. However,
since the length of each element, and consequently the reinforcement,
is assumed to be fixed up to failure, part of the assumed hogging mo-
ment region will start to resist small amount of positive moment as
shown in Fig. 4(c). The influence of this overlap is insignificant since
the moment values adjacent to the point of contraflexure are relatively
Fig. 5. Element forces and displacements. low. Failure of the beam is activated by crushing of the extreme con-
crete fibers at the intermediate support where the maximum moment is
n n anticipated. The expected load-deflection curve of the modeled propped
⎛
∑ (Ei,c Ai,c + Ei,J Ai,J ) yi2 − ∑ (Ei,c Ai,c + Ei,J Ai,J ) yi ⎞⎟
⎜ i=1 cantilever is presented in Fig. 4(d). It shows both the point of maximum
⎛ Δφ ⎞
( ΔM
ΔP )
=⎜ n
i=1
n ⎟ Δε
⎜ ⎟ deflection and the inflection point that is determined at the initial
loading steps and fixed throughout the analysis.
∑ (Ei,c Ai,c + Ei,J Ai,J ) ⎟⎟ ⎝ ⎠
a
⎜− ∑ (Ei,c Ai,c + Ei,J Ai,J ) yi
⎜
⎝ i=1 i=1 ⎠
(4) 6.4. Load-deflection relationship and equivalent curvature distribution

Once the slip effect is incorporated in a unique M-φ diagram for

6.3. Moment redistribution in continuous beams each segment, the area-moment method is carried out to determine the
deflection at distance of 0.4215 of the span away from the edge support.
Matrix stiffness analysis is carried out to account for moment re- This distance defines the location of maximum deflection for symmetric
distribution caused by the difference in stiffness between the hogging typical continuous beam supporting a uniformly distributed load [27].
and sagging moment zones. Fig. 5 represents an arbitrary element of the Having obtained the load-deflection curve of the jacketed beam
propped cantilever model subjected to external static uniformly dis- including slip effect, the actual curvature distribution of the propped
tributed load. The distorted shape of the element can be described in cantilever is determined at different loading steps for each segment.
terms of a translational displacement (di) and in-plane rotation (θi) at its These values are obtained from the corresponding M-φ diagram taking
ends. The element stiffness is used in Eq. (5) to express the joint internal into account the partial composite action according to the jacketing
−
forces (i.e. Pi and Mi) as functions of the corresponding displacements scheme used. After that, positive (φ+eq ) and negative (φ eq) equivalent
(i.e. di and θi) and fixed-end forces due to the applied loads (i.e. pi and curvatures are obtained by assuming that the curvature distribution
mi) [26]. along the beam is similar to the monolithic behavior of jacketed beams.
Therefore, at each loading value, and consequently deflection, equiva-
12 6 −12 6
⎡ L3 Li2 Li3 Li2
⎤ lent maximum positive and negative curvatures corresponding to the
⎢ i ⎥ d
⎡ Pi−p ⎤ ⎢ 62 4 −6 2 ⎥⎡ i ⎤ applied moment can be obtained. Hence, an equivalent M-φ curve can
⎢ Mi−mi ⎥ ⎢ Li Li Li2 Li ⎥ ⎢ θi ⎥ be obtained for the jacketed beam taking into consideration slip effect.
⎢ P −p ⎥ = (EI )i,j ⎢ −12 −6 12 −6 ⎥ ⎢ d ⎥
i+1 i + 1 i + 1⎥ The load-deflection curve can be determined at any point using the
⎢ ⎥ ⎢ Li3 Li2 Li3 Li2 ⎥ ⎢
⎣ Mi + 1−mi + 1⎦ ⎢ 6 2 −6 4 ⎥⎣
⎢ θi + 1 ⎥
⎦ moment-area theorem and the anticipated deflection shape.
⎢ L2 Li Li2 Li
⎥
⎣ i ⎦ (5)
7. Validation
The proposed method modifies the matrix analysis procedure by
incorporating the influence of slip. The M-φ diagram for each section is The capability of the proposed model to capture the flexural beha-
first calculated while accounting for slip as explained in Section 6.1. vior of simply supported jacketed RC beams was previously validated
The secant stiffness is then evaluated for a given moment. For each [16]. For continuous jacketed beams, the experimental program per-
loading step, the relationship in Eq. (14) is carried out for each segment formed by Cheong and MacAlevey [15] is considered. Fig. 6 shows the
(i) considering the secant flexural stiffness (j) obtained from the cor- longitudinal and cross-sectional views of the jacketed continuous beam.
responding M-φ diagram at the specified load level. The equilibrium Initially, the T-section concrete core was cast according to the cross-
and compatibility conditions obtained from the matrix structural ana- sectional dimensions and reinforcement distribution shown. After
lysis and the slip calculation algorithm must be verified simultaneously. 28 days of curing, the contact surfaces were roughened prior to ap-
Hence, nested iterations are required for each load step to satisfy plying the jacket. The concrete compressive strength was reported as
equilibrium and continuity for each segment along the beam. 30 MPa for the core and 60 MPa for the jacketing material. The tensile
The moment redistribution along the beam is dictated by the flex- yield strength of bar size Φ16, Φ25, Φ6, Φ10 and Φ8 were 583 MPa,
ural stiffness ratio between the hogging and sagging moment regions 567 MPa, 290 MPa, 321 MPa and 407 MPa, respectively. The tensile
[27]. Fig. 3 shows the M-φ relationships for the positive and negative ultimate strength for the same sequence of bars were 652 MPa,
moment sections of an arbitrary continuous beam. Because of the 670 MPa, 394 MPa, 424 MPa and 477 MPa, respectively. The jacketed
higher initial stiffness of the negative moment section, the point of zero beam was subjected to two-point loading scheme at one span only as
moment is shifted away from the intermediate support towards the mid- shown in Fig. 6(a).
span. A sketch of the bending moment diagram and the flexural rigid- The proposed calculation method is carried out to determine the
ities within the elastic loading stage for both the hogging and sagging flexural behavior of the jacketed continuous beam in terms of the load-
regions are illustrated in Fig. 4(b). In this case, the flexural rigidity is deflection curve at the center of the loaded span. The load-deflection
constant within each region but vary between the positive and negative curves assuming monolithic and partially composite behaviors are then
zones. Bending moment diagram is obtained based on the stiffness plotted and compared with the ones obtained experimentally by
distribution along the span. Cheong and MacAlevey [15]. Fig. 7 shows that the percent errors in
Once the negative moment section yields, its secant stiffness will initial stiffness between the experimental results and the proposed

112
M.M.A. Alhadid, M.A. Youssef Engineering Structures 167 (2018) 108–120

(a) jacketed beam longitudinal view and location of the applied loads

Sec A-A Sec B-B Sec C-C

(b) cross-sectional views.
Fig. 6. Longitudinal and cross-sectional views of the beams tested experimentally [15].

450 further validation. The connection between the concrete core and the
400 Finite Element surrounding jacket is modeled assuming full bond between the adjacent
(Monolithic) nodes. Geometrical details of both the concrete core and the jacket are
350 modeled using SOLID65, which is an 8–node solid element. This ele-
ment is capable of cracking under tensile stresses and crushing when
300
Applied Load (kN)

Experimental subjected to excessive compressive stresses. The element is defined by 8

250 nodes each having three translational degrees of freedom. The steel
main and secondary steel bars in both the core and the jacket are
200 modeled using LINK180 element. This element is a uniaxial tension-
Analytical compression element with three degrees of freedom at each node cor-
150
(Slip)
responding to the X, Y and Z coordinates. The supporting plates used at
100 Analytical the location of applied concentrated loads or at the supports are mod-
(Monolithic) eled using SOLID185 element assuming elastic behavior. This element
50
is also defined by 8 nodes each having 3 translational degrees of
0 freedom in each direction. A sensitivity analysis is performed to de-
0 10 20 30 40 50 60 70 80 termine the optimum mesh size as shown in Fig. 8. The stiffness and
Deflection at Center of Loaded Span (mm)
yield stress obtained from the finite element model are 8.7% and 10.2%
Fig. 7. Validation of the proposed model in view of the experimetal study by higher than the corresponding experimental values as shown in Fig. 7,
Cheong and MacAlevey [15]. respectively. This variation is attributed to neglecting the influence of
slip between the concrete core and the attached jacket. Fig. 7 also
analytical ones are 7.9% and 2.9% assuming both monolithic and shows that the load–deflection curve obtained from the proposed ana-
partial interaction, respectively. lytical model assuming monolithic action follows the same trend as that
Cheong and MacAlevey [15] reported that a slip between the con-
crete core and the surrounding jacket was detected without presenting
any more data about the slip distribution along the interface. The re-
latively close variations in the flexural stiffness in the elastic range
indicates that the proposed model is capable of predicting the load-
deflection behavior prior to steel yielding. Introducing the slip effect in
the analysis further improves the predictions by lowering the stiffness
to approach the experimental trend. The value of friction coefficient
chosen in the analysis is 0.8 to account for surface treatment using
electric chisel used in the experiment (i.e. roughened surface) [28].
Regarding the ultimate load, the percent error between the experi-
mental and proposed analytical results is 6.2% and 3.8% by ignoring
and including the slip effect, respectively. Cheong and MacAlevey [15]
reported that the observed failure is brittle caused by the excessive
tensile stresses at the narrow bearing supports, which was not ac-
counted for in the proposed model.
Due to the lack of experimental results, a finite element model is
developed assuming monolithic behavior using ANSYS Software for Fig. 8. Meshing of the jacketed beam in ANSYS.

113
M.M.A. Alhadid, M.A. Youssef Engineering Structures 167 (2018) 108–120

obtained from the developed finite element model with a percent dif- 18.0 Monolithic
4.2%
ference of 5.2% and 6.4% for stiffness and yield stress, respectively. Slip
16.0 3.3%

14.0

Initial Stiffness (×103 kN.m2)

8. Parametric study
4.1%
12.0 4.3%
A parametric study is carried out to investigate the influence of
different design parameters on the performance of jacketed continuous 10.0
RC beams. The concrete compressive strength is taken as 25 MPa, 4.0% 3.3%
8.0 6.4%
30 MPa and 35 MPa; and the steel yield strength is taken as 300 MPa, 7.1% 10.3%
7.9% 11.9% 5.1%
400 MPa and 500 MPa. For each of the analyzed sections, the me- 6.0 5.8%
chanical properties are assumed to be the same for the concrete core
4.0
and its jacket. The coefficient of friction ranges according to ACI [28]
between 0.4 for smooth concrete surface to 1.4 for intentionally highly 2.0
roughened concrete in increments of 0.2. The beam’s cross-sectional
0.0
dimensions are defined with reference to the existing beam height
B-1 B-2 B-3 B-4 B-5 B-6 B-7 B-8 B-9 B-10 B-11 B-12 B-13
(300 mm, 450 mm and 600 mm), jacket thickness (100 mm, 150 mm Section
and 200 mm), existing beam width (200 mm, 300 mm and 400 mm),
and beam span (3 m, 4 m and 5 m). The steel reinforcement distribution Fig. 9. Percent difference of initial stiffness with and without slip effect (sag-
along the beam is shown in Fig. 1 in which the balanced steel re- ging).
inforcement ratio is determined with regard to CSA A23.3 [21]. Jack-
eting from one side at the soffit of all beams is adopted in the analysis. 18.0 Monolithic
Each section is analyzed 63 times to account for the considered vari- 3.8%
Slip
16.0 5.1%
ables. Therefore, a total of 5103 different cases are considered in the
6.4%
current parametric study. 14.0
Initial Stiffness (×103 kN.m2) 6.2%
12.0
9. Moment-curvature behavior
10.0
The following discussion refers to the beam sections whose geo- 8.2% 4.4% 2.5%
8.0
metrical and mechanical properties are listed in Table 1. These sections 10.1% 7.8% 12.5%
11.4%
are considered to examine the influence of slip on flexural behavior of 6.0
13.3%
jacketed RC beams due to the variation of jacket thickness, beam width, 7.7%
4.0
beam height, span, concrete compressive strength and steel grade.
Figs. 9 and 10 show the initial stiffness values for each section assuming 2.0
full and partial composite actions (assuming a friction coefficient of
0.4) under both sagging and hogging moments, respectively. The re- 0.0
B-1 B-2 B-3 B-4 B-5 B-6 B-7 B-8 B-9 B-10 B-11 B-12 B-13
duction in initial stiffness caused by slip is indicated as a percentage in Section
the corresponding figures. Fig. 11 describes the variation of the re-
duction rate in stiffness (as percentage) due the variation of each of the Fig. 10. Percent difference of initial stiffness with and without slip effect
aforementioned parameters. Reference to Table 1 and Figs. 9–11 is (hogging).
made throughout the following discussion.
jacket reinforcement ensued by yielding of core reinforcement and a
9.1. Effect of jacket thickness (hJ) yielding plateau until failure by concrete crushing. Regarding the
hogging moment region, the yielding plateau occurs immediately after
Beams B-1, B-2 and B-3 are considered for comparison. Figs. 12 and yielding of the tension steel bars located in the original beam. The same
13 illustrate the influence of varying the jacket thickness on the flexural behavior is found for the remaining parameters; therefore, only the
behavior of continuous beams in view of the M-φ relationships along stiffness values are included in the discussion.
the sagging and hogging moment regions, respectively. The flexural The stiffness reduction rate in both the sagging and hogging mo-
behavior in the sagging moment region is characterized by yielding of ment zones slightly decreases with increasing the jacket thickness. The

Table 1
Geometry of the discussed jacketed beams.
Section Studied parameters Span (m) bc (mm) hc (mm) hJ (mm) fc' (MPa) fy (MPa)

'
B-1 hJ, fc , fy 3 200 300 100 30 400
B-2 hJ 3 200 300 150 30 400
B-3 hJ, bc, hc, Span 3 200 300 200 30 400
B-4 hc 3 200 450 200 30 400
B-5 hc 3 200 600 200 30 400
B-6 bc 3 300 300 200 30 400
B-7 bc 3 400 300 200 30 400
B-8 Span 4 200 300 200 30 400
B-9 Span 5 200 300 200 30 400
B-10 fc' 3 200 300 100 25 400
B-11 fc' 3 200 300 100 35 400
B-12 fy 3 200 300 100 30 300
B-13 fy 3 200 300 100 30 500
B-14 Δε, S, τ 3 200 450 150 30 300

114
M.M.A. Alhadid, M.A. Youssef Engineering Structures 167 (2018) 108–120

16 Sagging 16 Sagging

Change in Initial Stiffness (%)

Change in Initial Stiffness (%)

14 Hogging 14 Hogging
12 12
10 10
8 8
6 6
4 4
2 2
0 0
B-1 B-2 B-3 B-3 B-6 B-7
Section Section
(a) jacket thickness (b) section width
16 Sagging 16 Sagging
Change in Initial Stiffness (%)

Change in Initial Stiffness (%)

14 Hogging 14 Hogging
12 12
10 10
8 8
6 6
4 4
2 2
0 0
B-3 B-4 B-5 B-3 B-8 B-9
Section Section
(c) section height (d) span
16 Sagging 16 Sagging
Change in Initial Stiffness (%)

Change in Initial Stiffness (%)

14 Hogging 14 Hogging
12 12
10 10
8 8
6 6
4 4
2 2
0 0
B-10 B-1 B-11 B-12 B-1 B-13
Section Section
(e) concrete compressive strength (f) steel grade
Fig. 11. Variation of sagging and hogging initial stiffness with various parameters.

100 100
B-3 (hJ = 200 mm) B-3 (hJ = 200 mm)

75 75
Moment (kN.m)

Moment (kN.m)

B-2 (hJ = 150 mm) B-1 (hJ = 100 mm)

50 50 B-2 (hJ = 150 mm)
B-1 (hJ = 100 mm)

25 25

0 0
0 25 50 75 100 0 25 50 75 100
Curvature (Rad/km) Curvature (Rad/km)

Monolithic Slip (ȝ= 0.4) Monolithic Slip (ȝ = 0.4)

Fig. 12. Effect of varying hJ on the M-φ relationship (sagging). Fig. 13. Effect of varying hJ on the M-φ relationship (hogging).

ductility increase is insignificant when slip is considered for the sagging identical beyond the yielding point of the jacket steel bars regardless of
moment region indicating that the compressive strains at the extreme the slip strain. However, in the hogging moment region, as the jacket
compression fibers reach the concrete crushing strain value at the same thickness increases, the contribution of the concrete material and the
curvature. This happens since the axial stress in the jacketing layer compression steel bars located in the jacket layer becomes more pre-
assuming both monolithic and partially composite actions become valent relative to the entire section. Therefore, slip strain reduces the

115
M.M.A. Alhadid, M.A. Youssef Engineering Structures 167 (2018) 108–120

generated compressive stresses within the jacket layer at the same bending moment, this higher stress at the tension steel bars is translated
curvature value. This results in delaying the concrete crushing and into higher compressive stresses at the compression face of the jacketed
consequently increasing the ductility as the jacket height increases re- beam causing the concrete to reach its crushing strain at lower curva-
lative to the monolithic beams. ture values. Regarding the slip influence on the flexural behavior of
these beams, the initial stiffness reduction rate decreases as the original
9.2. Effect of beam width (bc) section height increases for both the sagging and hogging moment
cases. This decrease is a result of the higher slip strain required to
Increasing the beam width results in a consequent increase in both equilibrate the axial force within the jacket with the horizontal shear
the initial stiffness and capacity with minor influence on the flexural force along the interface.
ductility. Regarding the slip influence, increasing the beam width re-
sults in decreasing the reduction rate of the initial stiffness in both 9.4. Effect of beam span (L)
sections. This is justified by the larger contact area between the con-
crete core and the jacket that is provided by the additional beam width. The M-φ curve assuming monolithic interaction between the core
Two main differences arise from changing the location of the contact and the jacket are identical regardless of the span as they depend
surface with respect to the neutral axis. When the interface is located at merely on the cross-sectional properties. However, including the slip
the tension side (i.e. sagging moment section), the reduction in the effect activates the partially composite action and consequently the
elastic stiffness is relatively smaller than the case of hogging moment. horizontal shear distribution along the interface becomes a major
This variation in stiffness reduction is attributed to the contribution of player in determining the flexural behavior of any section along the
both concrete and steel in determining the slip strain (Δε) at each beam. In both the sagging and hogging moment cases, as the span in-
section. For the sagging moment region, the bending stresses at the creases, the elastic stiffness reduction rate decreases proportionally.
tension side are resisted by both the core steel bars and the jacket steel This observation is justified by the higher contact area provided by the
bars especially after concrete cracking takes place. This means that the larger span and consequently the increased frictional forces along the
steel in the jacketing layer sustains part of the generated tensile stresses jacketed beam. For the positive moment section, the partially composite
and the remaining part is resisted by the steel bars in the original beam. flexural behavior becomes identical to the monolithic counterpart once
Thus, the slip strain required to achieve equilibrium at any section jacket steel bars yield. This happens due to the small variation in the
along the jacketed beam is governed by a portion of the total tensile axial stresses governed by the strain hardening modulus of jacket steel
stress generated at a given applied load. A different situation is ob- bars after yielding occurs. Thus, at the same curvature value, the stress
served along the hogging moment region where the jacketing layer is at in these steel bars is almost identical to the ones in the monolithic case.
the compression side. In this case, the entire concrete material is uti- Although it still exists, the influence of slip strain diminishes even more
lized along with the jacket steel bars to resist the same applied load. at higher loading values due to the higher contribution of compression
This indicates that a larger portion of bending is carried by the jacket concrete and tension core steel bars while the stresses in the jacket steel
part causing an increase in the slip strain required to achieve equili- bars remain almost constant. Regarding the hogging moment region,
brium at any segment along the hogging moment region. the variation between the partially composite scenario and monolithic
The other difference that prevails from changing the location of the behaviors persists within a portion of the inelastic region. This occurs
interface with respect to the neutral axis is the point which the M-φ since the jacketing layer is governed by the compressive stresses de-
curves ignoring and including slip effects follow the same path. For the veloped in concrete and the embedded steel bars rather than the tensile
sagging moment region, the major difference in the moment-curvature stresses generated merely in the steel bars. Thus, even after yielding of
diagram is within the elastic region before yielding of the jacket steel the section takes place, the axial force within the jacket at any section
bars. This is justified by knowing that the axial force at any section is remains different from the monolithic case due to the influence of slip
determined by the jacket steel bars. So, once these bars yield, the tensile strain which decreases the jacket stresses at any curvature value. At
forces in the jacket steel bars becomes almost constant and any increase higher loading values, the slip strain becomes less pronounced relative
depends on the strain hardening modulus. Thus, after jacket yielding is to the higher curvature values and consequently its influence becomes
reached, the influence of slip strain becomes negligible in changing the less substantial.
behavior of the M-φ diagram compared to its monolithic counterpart.
Regarding the hogging moment region, yielding point is dictated by the 9.5. Effect of concrete compressive strength ( fc′)
tensile steel reinforcement in the concrete core. Therefore, the yielding
point considering slip occurs at a larger curvature value compared to Increasing the concrete compressive strength results in a consequent
the monolithic case. Since the tensile stresses in the jacketing layer are increase in the beam capacity as it resists higher stresses for the same
governed by the compression behavior of both the concrete material peak strain value. Also, increasing the concrete grade rises the concrete
and the embedded jacket steel bars, the influence of slip strain remains modulus resulting in a higher elastic stiffness value. Regarding the slip
considerable in reducing the moment carrying capacity at a given effect, increasing the concrete compressive strength decreases the
curvature. As the load increases, the effect of slip strain diminishes until stiffness reduction rate indirectly through increasing the friction be-
the moment-curvature behavior of the partially composite section be- tween the two surfaces. This is inferred by examining the change in
comes identical to the monolithic one. flexural behavior when slip is considered in both the sagging and
hogging moment regions.
9.3. Effect of existing beam depth (hc)
9.6. Effect of steel grade ( f y )
In both the sagging and hogging moment cases, increasing the ex-
isting section height increases both the elastic stiffness and capacity of Increasing the steel yield strength has a negligible influence on the
the jacketed beams. This is justified by the larger concrete material initial stiffness of the jacketed beams but a substantial enhancement to
available in the compression side and the longer lever arm the tension its capacity. The main reduction in stiffness will be in the elastic zone in
steel bars have. The ductility, on the other hand, decreases as the sec- which the steel elastic modulus plays the major role. Considering slip in
tion height increases and becomes even more pronounced if the inter- the analysis shows that as the steel grade increases, the drop in flexural
face is at the compression side. This drop in ductility is related to the stiffness also increases for both the sagging and hogging moment sec-
higher stresses developed in the tension steel bars as the original beam tions. This happens since the steel bars with higher grade within the
height rises at any curvature level. Therefore, at the same applied jacket resist larger axial forces before yielding and consequently result

116
M.M.A. Alhadid, M.A. Youssef Engineering Structures 167 (2018) 108–120

0.60
Ultimate

0.40
Yielding of Tension Rft.

0.20

Slip, S (mm)
0.00
0 500 1,000 1,500 2,000 2,500 3,000

Elastic Region
-0.20

-0.40

-0.60
Distance from Support (mm)

w = 30 kN/m w = 90 kN/m w = 104 kN/m w = 114 kN/m

Fig. 14. Slip strain (Δε) distribution along beam B-5 (μ = 0.4). Fig. 16. Slip distribution along beam B-5 (μ = 0.4).

in higher shear stresses to achieve equilibrium. These higher stresses 0.60

result in larger slip and consequently larger slip strain that reduces the
flexural stiffness of the jacketed beams.
0.40 Ultimate

10. Interfacial slip behavior Yielding of Tension Rft.

0.20

The influence of interfacial slip between the concrete core and the
Slip, S (mm)

underlying jacket layer is investigated in view of the slip strain (Δε), slip 0.00
(S) and interfacial shear stress (τ) distribution along the continuous 0 500 1,000 1,500 2,000 2,500 3,000
beams under different loading values. Beam B-14 in Table 1 is con-
-0.20 Elastic Region
sidered for the following discussion. The coefficient of friction between
the two surfaces is taken as 0.4 and 1.0 which represent untreated
surfaces and intentionally roughened surfaces, respectively. Figs. 14–18 -0.40
represent the distribution along one span only of the continuous beam.
-0.60
Distance from Support (mm)
10.1. Slip strain (Δε) distribution
w = 30 kN/m w = 90 kN/m w = 106 kN/m w = 115 kN/m
Figs. 14 and 15 illustrate the slip strain distribution from the edge Fig. 17. Slip distribution along beam B-5 (μ = 1.0).
support towards the intermediate support for coefficient of friction of
0.4 and 1.0, respectively. Four loading values representing the elastic
2.00
range, onset of jacket yielding, onset of core yielding and maximum
Ultimate
capacity of the section at the intermediate support at the onset of 1.50
Yielding of Tension Rft.
concrete crushing. Both figures show the same trend in which the slip
Interfacial Shear Stress, IJ (MPa)

strain at any section increases with the applied load except at the points 1.00
of zero moment (i.e. the edge support and the point of contraflexure).
This increase corresponds to the rise in the axial stresses within the 0.50
jacket layer to maintain the equilibrium condition with the interfacial
0.00
0 500 1,000 1,500 2,000 2,500 3,000
-0.50
Elastic Region
-1.00

-1.50

-2.00
Distance from Support (mm)

w = 30 kN/m w = 90 kN/m w = 104 kN/m w = 114 kN/m

Fig. 18. Interfacial shear stress distribution along beam B-5 (μ = 0.4).

shear along the contact plane.

The maximum slip strain reaches the peak value at two points
corresponding to the maximum positive bending moment and the
maximum negative bending moment sections. The slip strain is always
higher at the intermediate support than the maximum positive moment
for two main reasons. The first one is that the sagging moment region
Fig. 15. Slip strain (Δε) distribution along beam B-5 (μ = 1.0). extends along a larger distance than the hogging moment zone resulting

117
M.M.A. Alhadid, M.A. Youssef Engineering Structures 167 (2018) 108–120

in a larger contact area and friction resistance and consequently less slip 2.00

strain in the former case. Also, the slip strain is proportional to the Ultimate
1.50
bending moment that develop axial stresses within the jacket. Since the

Interfacial Shear Stress, IJ (MPa)

negative moment at the middle support is always larger than the Yielding of Tension Rft.
1.00
maximum positive moment at any loading value, the slip strain follows
the same trend and becomes higher at the intermediate support. By 0.50
comparing the curves in Fig. 14 for μ = 0.4 to their counterparts in
Fig. 15 for μ = 1.0, higher slip strain values at any given load are ob- 0.00
0 500 1,000 1,500 2,000 2,500 3,000
served in the former case. This difference occurs due to the lower in-
-0.50
terfacial stiffness as the friction coefficient decreases. Thus, for the same
Elastic Region
axial stresses in the jacket, higher slip strain is required to achieve
-1.00
equilibrium with the interfacial shear stress. By roughening the con-
crete surface prior to jacketing, the slip strain at the maximum positive -1.50
moment section drops from about 0.39 to 0.16 indicating a ratio of
58.9%. This drop at the maximum negative moment section is shown to -2.00
Distance from Support (mm)
be from 1.25 to 0.66 with a ratio of 47.2%. The slip strain increasing
rate rises at higher loading values compared to the elastic region in both w = 30 kN/m w = 90 kN/m w = 106 kN/m w = 115 kN/m
the hogging and sagging moment regions. For the maximum positive Fig. 19. Interfacial shear stress distribution along beam B-5 (μ = 1.0).
moment section, increasing the uniformly applied load from 30 kN/m
to 90 kN/m along the beam results in a consequent increase of the slip
strain by just 0.07 × 10−3 for the untreated surface and by just ultimate loading value. This is justified by the higher frictional re-
0.03 × 10−3 for the roughened surface. After yielding occurs, in- sistance and consequently the higher interfacial stiffness as the original
creasing the load by about 10 kN/m results in an increase of beam surface is roughened. It is worth mentioning that at the ultimate
0.16 × 10−3 and 0.08 × 10−3 for the smooth and rough surfaces, re- load, the slip value at any section is less than the failure value defined
spectively. The same observation applies for the maximum negative in the slip model of about 2 mm.
moment section but with different increasing rate. This is explained by
the larger curvature the beam undergoes within the yielding plateau 10.3. Interfacial shear stress (τ) distribution
corresponding to any variation in the applied load relative to the elastic
range. Figs. 18 and 19 detail the horizontal shear stress distribution along
the interface between the concrete core and the attached jacket layer
10.2. Slip (S) distribution considering untreated and roughened surfaces, respectively. As shown
in both figures, the distribution follows a third order parabolic function
The slip distribution along the interface considering both smooth as initially assumed. The shear values are then determined by carrying
and rough surfaces are shown in Figs. 16 and 17, respectively. As shown out both sectional and longitudinal analyses to satisfy the equilibrium,
in these figures, the maximum slip values are obtained at the edge compatibility and constitutive conditions. Also, the figures demonstrate
support and the point of contraflexure that both correspond to the the direct relationship that relate the interfacial shear stress (τ) to the
sections of zero moments. On the contrary, relative slip between the slip (S) at any section through the interfacial stiffness (ks). Since the
two surfaces becomes negligible at the locations of maximum positive shear-slip model at the interface is non-linear, the secant interfacial
and negative moments. The same figures also show that slip values at stiffness varies depending on the slip value. For the smooth connection,
sections closer to the edge support are slightly less than those near the the interfacial shear to slip ratio at the edge support is obtained as
middle support. This reduction is due to concrete cracking in the sag- 3.4 N/mm3 for all distributions up to the yield point and 2.9 N/mm3 at
ging moment region that takes place during the initial loading stage. ultimate. The same conclusion is drawn by comparing the curves in the
When the concrete jacket is cracked, only the jacket reinforcement second figure but with the secant interfacial stiffness of 9.1 N/mm3 up
contributes in resisting the generated axial force resulting in lower level to the yield point and 7.7 N/mm3 at ultimate at the same section. As
of shear stresses transferred along the interface. expected, the interfacial stiffness at any given load is higher when the
In the hogging moment region, both concrete and the jacket steel original beam surface is roughened compared to the untreated case.
bars are active and resist the bending moment in terms of compressive Although the slip distribution along the interface is different for both
stresses. This results in larger contribution of the jacket and conse- cases, the interfacial shear stress distribution is almost identical. This is
quently higher shear stress to be transferred along the interface as justified by the variation of the interfacial stiffness between both cases
translated by the higher slip values. The slip increasing rate after the that result in equilibrium between the axial force in the jacketing layer
first yielding point is higher than the rate before yielding for both kinds and the horizontal shear force at any segment along the interface. The
of surface treatment. For example, increasing the applied load within same observations are shown in the hogging moment region.
the elastic region from 30 kN/m to 90 kN/m results in a consequent
increase in the maximum slip at the edge support from 0.04 mm to just 11. Proposed expressions for the effective stiffness
over 0.12 mm for the untreated surface case. However, after the
yielding point is reached, increasing the load by just 10 kN/m results in Having developed and verified an analytical procedure to analyze
extra relative sliding between the two surfaces of about 0.18 mm. The jacketed continuous RC beams considering the influence of interfacial
same observations are shown for the roughened surface case. This slip, a parametric study including 5103 specimens is carried out to
higher rate of slip rise is justified by the higher curvature the beam determine the contribution of various parameters on the flexural be-
exhibits after reaching the yielding point for the same amount of load havior of such beams. These parameters encompass the beams’ geo-
compared to the elastic range. Consequently, the slip strain (Δε) at the metrical properties, mechanical properties and interfacial behavior
yielded segments increases resulting in a larger increase in slip. By between the core and the RC jacket. The outcomes show that ignoring
examining both figures, the slip values considering smooth surfaces are the relative slip between the two substrates may overestimate the
higher than the ones obtained assuming roughened surfaces at any flexural stiffness causing serviceability issues such as larger deflections
section for the same loading level. For example, the slip at the edge and unexpected cracking. Therefore, the influence of slip should be
support for the former case is 138.9% higher than the second case at the considered when designing such jacketed beams. Including the

118
M.M.A. Alhadid, M.A. Youssef Engineering Structures 167 (2018) 108–120

influence of slip in the analysis is tedious and requires a sequence of monolithic (φu) and partially composite (φ∗u) actions. For the initially
nested iterations that may not be convenient for design engineers. Here loaded beams, two additional terms are introduced that define the both
comes the importance of providing the engineers with expressions that the moment (Minitial) and the equivalent curvature (φinitial) corre-
improves the accuracy of their designs with less time and effort. The sponding to the initial loading value at the onset of jacketing as in-
extent of flexural stiffness reduction as well as the point at which both dicated in Fig. 20(b).
the monolithic and partially composite curves becomes almost identical The proposed design procedure is summarized in the following
differ between the sagging and hogging moment regions. Therefore, three steps to obtain the actual load-deflection curve considering the
different expressions are provided to adjust the monolithic M-φ dia- sliding between the two surfaces:
gram of each region by considering the slip effect. Eqs. (6)–(9) provide
the expressions for αy and αu that represent the yield monolithic factor (1) Plot the M-φ diagram for the sections representing both the sagging
and ultimate monolithic factor for the hogging moment section, re- and hogging moment regions assuming monolithic interaction be-
spectively. Eqs. (10) and (11) presents the yield monolithic factor (αy) tween the original beam and the attached jacket. The hogging M-φ
for the sagging moment section. diagram is assumed bilinear and can be plotted by evaluating the
yield and ultimate points. Regarding the sagging moment section,
α y−ve = (ξ y−ve ) + [22.6645(ξ y−ve )2−46.3178(ξ y−ve ) + 23.6573] (6) only the yield point is required since concrete crushing usually
occurs at the negative moment section in continuous beams sub-
ξ y−ve = 1.15545−2.661 × 10−4fc′ + 3.229 × 10−5f y −1.266 × 10−5L
jected to static loads.
hJ2 (2) Modify these M-φ diagrams in view of Fig. 10(a) and Eqs. (6)–(11)
+ 3.30 × 10−7bc2−2.811 × 10−4bc−1.704 × 10−5hc + 5.22 × 10−6
bc for beams not subjected to initial load during jacketing. If the beam
−1.57 × 10−5hJ −0.037306 μ ⩾ 1.0 was subjected to initial loading prior to jacketing, then modify the
(7)
M-φ diagrams in view of Fig. 20(b) and Eqs. (12)–(14) taking into
αu−ve = (ξu−ve ) + [1.4756exp(138.9291ξu−ve )] (8) account the initial applied load level and the monolithic factors for
unloaded beams obtained from Eqs. (6)–(11).
ξu−ve = 1.11070−1.108 × 10−4fc′ + 3.459 × 10−5f y −1.018 × 10−5L (3) Use the equivalent M-φ diagrams obtained at the sagging and
hogging moment regions along with the moment-area theorem to
hJ2
+ 1.90 × 10−7bc2−1.784 × 10−4bc−8.39 × 10−6hc + 2.857 × 10−5 obtain the load-deflection diagram at any point along the beam.
bc
−9.06 × 10−6hJ −0.033465 μ ⩾ 1.0 (9) The expectation function of the proposed monolithic factors is de-
termined considering nonlinear regression analysis of the data.
α y+ve = (ξ y+ve ) + [20.3463(ξ y+ve )2−41.0203(ξ y+ve ) + 20.6732] (10) Fig. 21(a), (b) and (c) present the line of equality corresponding to α y−ve ,
αu−ve and α y+ve without initial loading, respectively.
ξ y+ve = 1.11354−1.108 × 10−4fc′ + 3.459 × 10−5f y −1.018 × 10−5L The line of equality plots for all factors reveal that the model pro-
+ 2.20 × 10−7bc2−2.043 × 10−4bc−8.39 × 10−6hc−2.190 × 10−5hJ vides a very good prediction of the actual behavior. Residual analysis
for the three factors clearly shows a normally distributed pattern of the
−0.033465 μ ⩾ 1.0 (11)
residuals about the mean. The small positive value of mean indicates
where fc′ is the concrete compressive strength in MPa; f y is the steel that the proposed expressions tend to slightly round up the actual factor
yield strength in MPa; L is the beam span in mm; bc is the section width resulting in higher stiffness reduction and therefore more conservative
in mm; hc is the section height in mm; hJ is the jacket thickness in mm estimates. Similar statistical analysis is carried out for the factors when
and μ is the coefficient of friction between the original beam and the initial load level is considered and a very good agreement is also found.
attached jacket. If the beams were subjected to initial loading before
jacketing, then the monolithic factors should be reduced according to 12. Summary and conclusions
the expressions given in Eqs. (12)–(14) for hogging ultimate monolithic
factor, hogging yield monolithic factor and sagging yield monolithic An investigation of the influence of RC jackets on the flexural be-
factor, respectively. havior of continuous RC beams was presented. A parametric study in-
1.327 cluding 5103 symmetric continuous beams subjected to uniformly
Minitial ⎞
(α y−ve )initial = α y−ve−⎜⎛ ⎟ (α y−ve−1.0) ⩾ 1.0 distributed loads is carried out. The jacket is applied from one side at
⎝ Mu,unjacketed ⎠ (12) the soffit of all beams. Different parameters including the geometrical
properties (i.e. original beam width, original beam depth, jacket
0.849
Minitial ⎞ thickness and beam span); mechanical properties (i.e. concrete com-
(αu−ve )initial = αu−ve−⎛⎜ ⎟ (αu−ve−1.0) ⩾ 1.0
M
⎝ ,unjacketed ⎠
u (13) pressive strength and steel yield strength); and surface treatment (i.e.
interfacial friction coefficient) are investigated. An analytical model
1.113
Minitial ⎞ encompassing sectional and interfacial analyses were developed taking
(α y+ve )initial = α y+ve−⎛⎜ ⎟ (α y+ve−1.0) ⩾ 1.0 into account that constitutive, compatibility and equilibrium conditions
⎝ Mu,unjacketed ⎠ (14)
are satisfied. The proposed model was validated in view of the ex-
where Minitial is the maximum applied moment during jacketing and perimental work performed by Cheong and MacAlevy [15]. Due to the
Mu,unjacketed is the flexural capacity of the unjacketed section. In these lack of other relevant experimental studies, a finite element model was
expressions, the section subjected to maximum negative moment is also developed for further validation. The parametric study revealed
considered to determine the hogging moment, while the section sub- that including the slip influence in the analysis results in a reduction of
jected to maximum positive moment is used in evaluating the sagging stiffness that should be considered when designing jacketed sections.
moment. Fig. 20(a) and (b) detail the variation in a typical equivalent For the beams considered in the analysis, ductile failure mode char-
moment-curvature diagrams assuming monolithic and partially com- acterized by yielding of tension steel bars followed by concrete crushing
posite sections without and with initially applied load, respectively. at the extreme compression fiber was detected. The effect of each of the
The main parameters defining the curves in Fig. 20 are the yield studied parameters on the M-φ relationship is similar for both the
moment (My) and the corresponding equivalent curvature assuming hogging and sagging moment regions but shown to be more pro-
monolithic (φy) and partially composite (φ∗y) actions; and ultimate nounced in the former zone. A design procedure and stiffness mono-
moment (Mu) and the corresponding equivalent curvature assuming lithic factors are introduced in terms of the studied parameters to obtain

119
M.M.A. Alhadid, M.A. Youssef Engineering Structures 167 (2018) 108–120

Before Jacketing After Jacketing

Mu Mu
My My

Moment

Moment
Minitial

ijy ij* y iju ij* u ijinitial ijy ij*y iju ij* u
Curvature Curvature
(a) No initial load. (b) With initial load.
Fig. 20. Typical moment-curvature diagram for jacketed beams.

1.10 1.10 [8] Li VC. From micromechanics to structural engineering - the design of cementitious
Įu(-ve) (Proposed Equation)
Įy(-ve) (Proposed Equation)

1.08 1.08 composites for civil engineering applications. Doboku Gakkai Rombun-Hokokushu/
Proc Jpn Soc Civil Eng 1993;471(1–24):1–12.
1.06 1.06 [9] Rossi P. High performance multimodal fiber reinforced fiber reinforced cement
1.04 1.04 composite (HPMFRCC): the LPC experience. ACI Mater J 1997;94(6):478–83.
[10] Meda A, Minelli F, Plizzari GA, Riva P. Shear behavior of steel fiber reinforced
1.02 1.02 concrete beams. Mater Struct 2005;38(3):343–51.
[11] Shehata I, Shehata L, Santos E, Simoes M. Strengthening of reinforced concrete
1.00 1.00
1.00 1.02 1.04 1.06 1.08 1.10 beams in flexure by partial jacketing. Mater Struct 2009;42(4):495–504.
1.00 1.02 1.04 1.06 1.08 1.10
[12] Tsonos ADG. Performance enhancement of R/C building columns and beam-column
Įy(-ve) (Analytical) Įu(-ve) (Analytical)
joints through shotcrete jacketing. Eng Struct 2010;32:726–40.
(a) (b) [13] Hamilton C, Pardoen G, Navalpakkam S, Kanzanjy R. Reinforced concrete bridge
1.10 column performance enhancement through shotcrete jacketing. ACI Struct J
2004;101(3):332–40.
Įy(+ve) (Proposed Equation)

1.08 [14] Souza RHF, Appleton J. Flexural behavior of strengthened reinforced concrete
1.06 beams. Mater Struct 1997;30:154–9.
[15] Cheong HK, MacAlevey N. Experimental behavior of jacketed reinforced concrete
1.04 beams. J Struct Eng – ASCE 2000;126(6):692–9.
1.02 [16] Alhadid MM, Youssef MA. Analysis of reinforced concrete beams strengthened using
concrete jackets. Eng Struct 2017;132:172–87.
1.00 [17] Scott BD, Park R, Priestley MJN. Stress-strain behavior of concrete confined by
1.00 1.02 1.04 1.06 1.08 1.10 overlapping hoops at low and high strain rates. J Am Concr Instit
Įy(+ve) (Analytical) 1982;79(1):13–27.
(c) [18] Karthik MM, Mander JB. Stress-block parameters for unconfined and confined
concrete based on a unified stress-strain model. J Struct Eng – ASCE
Fig. 21. Statistical analysis for the proposed expressions. 2011;137(2):270–3.
[19] Ramberg W, Osgood WR. Description of stress–strain curves by three parameters
Technical Note No. 902 Washington, DC: National Advisory Committee for
the flexural behavior of the continuous RC beams. Aeronautics; 1943.
[20] Thermou GE, Pantazopoulou SJ, Elnashai AS. Flexural behavior of brittle RC
members rehabilitated with concrete jacketing. J Struct Eng ASCE
References 2007;133(10):1373–84.
[21] CSA. Design of concrete structures (CAN/CSA A23.3-14). Ottawa (ON): Cement
[1] Tassios T, Vintzeleou E. Concrete-to-concrete friction. J Struct Eng ASCE Association of Canada; 2014.
1987;113(4):832–49. [22] Tsioulou OT, Dritsos SE. A theoretical model to predict interface slip due to
[2] Altun F. An experimental study of the jacketed reinforced-concrete beams under bending. Mater Struct 2011;44:825–43.
bending. Constr Build Mater 2004;18(8):611–8. [23] Kotsira E, Dritsos S, Pilakoutas K. Effectiveness of techniques for flexural repair and
[3] Bousias S, Spathis A, Fardis M. Seismic retrofitting of columns with lap-spliced strengthening of RC members. In: Proceedings of the 5th international conference
smooth bars through FRP or concrete jackets. J Earthquake Eng 2007;11:653–74. on Structural faults and repair, Edinburgh, UK; 1993. p. 235–43.
[4] Santos P, Júlio E. A state-of-the-art review on shear-friction. Eng Struct [24] Saiidi M, Vrontinos S, Douglas B. Model for the response of reinforced concrete
2012;45:435–48. beams strengthened by concrete overlays. ACI Struct J 1990;87(6):687–95.
[5] Martinola G, Meda A, Plizzari GA, Zinaldi Z. An application of high performance [25] Youssef MA, Rahman M. Simplified seismic modeling of reinforced concrete flexural
fiber reinforced cementitious composites for RC beam strengthening. Fracture me- members. Mag Concr Res 2007;59(9):639–49.
chanics of concrete and structures – high performance concrete Taylor and Francis [26] Kassimali A. Matrix analysis of structures. 2nd ed. Thomson-Engineering; 2011.
Group; 2007. p. 1541–8. 640 pp.
[6] Shimoyama Y, Uzawa M. Properties and application of ductal. J Taiheiyo Cem Corp [27] Oehlers DJ, Haskett M, Mohamed MS, Griffith MC. Moment redistribution in re-
2002;142:55–62. inforced concrete beams. Proc Inst Civ Eng Struct Build 2010;163(June):165–76.
[7] Vicenzino E, Culhman G, Perry V, Zakariasen D, Chow T. The first use of UHPFRC in [28] ACI Committee 318. Building code requirements for structural concrete and com-
thin precast roof shell for LRT Canadian station. PCI J 2005. [September-October]. mentary (ACI 318-14). Farmington Hills (MI): American Concrete Institute; 2014.

120