State of Research on Seismic Retrofit of RC Beam-Column

Joints with Externally Bonded FRP

Abdelhak Bousselham1

Abstract: A considerable amount of research has been directed recently toward understanding and promoting the use of externally
Downloaded from ascelibrary.org by University of California, San Diego on 04/22/17. Copyright ASCE. For personal use only; all rights reserved.

applied fiber-reinforced polymer 共FRP兲 for the seismic retrofit of reinforced concrete 共RC兲 structures. In this paper, a comprehensive
review and synthesis of published experimental studies on the seismic rehabilitation of RC frame beam-column joints with FRP is
presented, and the issues that need to be addressed for further research are discussed. In addition, the paper presents a simple design model
for predicting the contribution of the FRP to the shear strength of retrofitted joints. The key element in the model is the derivation of an
expression for the effective FRP strain, based on the calibration of test data reported in the literature. A total of 54 tests carried out
worldwide were considered in the review, and a database of the published studies, encompassing all relevant design parameters, was
assembled. The reported test results confirm the structural effectiveness of the FRP strengthening technique for the seismic retrofit of RC
joints. However, there are some gaps which need to be addressed. For instance, there is a lack of a rationale explanation of the resistance
mechanisms involved in the beam-column joints retrofitted with FRP. Such a rational explanation is a prerequisite for the development of
more comprehensive and rigorous design procedure.
DOI: 10.1061/共ASCE兲CC.1943-5614.0000049
CE Database subject headings: Beam columns; Concrete structures; Fiber composites; Joints; Models; Retrofitting; Seismic effects;
Shear strength.
Author keywords: Beam columns; Concrete structures; Joints; Rehabilitation; Seismic effects; Shear strength; fiber-reinforced
polymers; Bonding.

Introduction ings 共e.g., beam-column joints兲 and developing retrofit techniques

for such elements represent critical issues 关American Concrete
Seismic performance of existing buildings and bridges, in North Institute 共ACI兲 2002兴.
America and elsewhere in the world, is becoming a growing mat- As for RC beam-column joints, several deficiencies have been
ter of concern, after the devastating earthquakes worldwide expe- identified 共Beres et al. 1996兲. Little or no transverse steel rein-
rienced in recent years, such as the 1995 Kobe 共Japan兲, the 1999 forcement 共hereafter called transverse reinforcement兲 in the joint
Kocaeli 共Turkey兲, and the 2003 Boumerdes 共Algeria兲 earth- cores, inadequate anchorage of beam bottom reinforcement, and
quakes. These events repeatedly demonstrated the vulnerability of columns with bending capacity less than that of beams are ex-
those structures that were designed and built based on the pre- amples of such deficiencies. Also, many techniques have been
1970s design codes. Therefore, they constitute a high seismic risk, proposed recently for improving the seismic performance of RC
especially in large metropolitan areas. beam-column joints 共Engindeniz et al. 2005兲. Some of these tech-
Particularly vulnerable are pre-1970s designed reinforced con- niques, already used in the past 共e.g., concrete jackets and bolted
crete 共RC兲 moment-resisting frames. In such structures, the fail- steel plates兲, were revised and developed in the light of the new
ure of beam-column joints, especially exterior joints, is often the seismic code requirements while some others are based on new
cause of the collapse of buildings, as reported by several materials, such as fiber-reinforced polymer 共FRP兲 共Thermou and
postearthquake studies 关Moehle 2006; Japanese Reconnaissance Elnashai 2006兲.
Team 共JRT兲 2004; Saatcioglu et al. 2001兴. This structural type Particularly with regard to the use of externally applied FRP,
共i.e., RC moment-resisting frames兲 comprises a large share of the this technique offers many advantages: FRP has high strength
existing buildings stock, and it is not conceivable to replace sys- over weight ratio, high corrosion resistance of bonded material,
tematically all the deficient structures by new ones. Therefore, no significant increase in member size, easy applicability and lim-
identifying vulnerable elements of nonseismically designed build- ited disruption to building occupancy, and the tailorability which
allows adjusting the orientation in each ply depending on the
1
Research Associate, Dept. of Construction Engineering, Univ. specific strengthening objectives 共Engindeniz et al. 2005兲. The
of Quebec, École de technologie supérieure, 1100 Notre-Dame wide use of this retrofitting technique for various structures, in-
West St., Montreal, PQ, Canada H3C 1K3. E-mail: abdelhak. cluding buildings and bridges, has demonstrated its efficiency and
[email protected] its convenience 共Bakis et al. 2002兲. This success is the result of an
Note. This manuscript was submitted on November 18, 2008; ap-
extensive research effort carried out over the last 2 decades,
proved on May 6, 2009; published online on May 11, 2009. Discussion
period open until July 1, 2010; separate discussions must be submitted for which encompassed numerous aspects and provided valuable
individual papers. This paper is part of the Journal of Composites for findings that were synthesized and gathered in the format of de-
Construction, Vol. 14, No. 1, February 1, 2010. ©ASCE, ISSN 1090- sign guidelines for the practicing engineers. ACI Committee 440
0268/2010/1-49–61/$25.00. 关American Concrete Institute 共ACI兲 2008兴, Canadian Standards

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2010 / 49

J. Compos. Constr., 2010, 14(1): 49-61

 S806 关Canadian Standards Association 共CSA兲 2002兴, and Euro- then loaded again. The specimens were strengthened using two
pean recommendations FIB-TG9.3 关Fédération Internationale du layers of 0°/90° composite laminates consisted either of glass
Béton 共FIB兲 2001兴 are examples of such new guidelines. Con- 共GFRP: glass FRP兲 or carbon fibers. In addition, two unidirec-
cerning the seismic retrofit specifically, the use of FRP for the tional diagonal straps were placed at the surface of the laminate to
confinement of columns 共Saatcioglu 2003; Seible et al. 1997兲 and minimize the possibility of premature delamination failure. All
the retrofitting on infill masonry walls 共Saatcioglu 2006; El- tested specimens exhibited joint shear 共JS兲 failure. However, the
Gawady et al. 2005兲 are well documented. In comparison, studies geometry of the tested beam-column joints and the protocol of
on beam-column joints are few. loading are questionable.
It must be recognized that the behavior of beam-column joints Four beam-column joints were tested by Ghobarah and Said
is by its nature a complex issue, and probably the most contro- 共2002兲. The specimens were made without transverse reinforce-
versial aspect of the seismic design of RC moment-resisting ment in the joint, and in accordance with the strong-beam weak-
frames at present 共Park 2003; Priestley and Calvi 1991兲. It in- column concept. The specimens were strengthened using four
Downloaded from ascelibrary.org by University of California, San Diego on 04/22/17. Copyright ASCE. For personal use only; all rights reserved.

volves several mechanisms, and numerous parameters often different rehabilitation schemes, some specimens with and others
acting simultaneously, which are not fully understood 共Panta- without mechanical anchorage, including steel plates and
zopoulou and Bonacci 1992兲: shear, bond, and confinement. As a threaded rods core-drilled through the joint. It may be worth men-
result, conflicting views, clearly reflected in code provisions, tioning that the applied composite material was extended above
about joint mechanics and design recommendations, are still and below the joint and wrapped around the column, and the
under investigation. Besides, there is a lack of comprehensive flexural strength of the column increased as a result. The beams
experimental evidence of the behavior of joints with little to no with no anchorage ruptured by JS failure whereas those with an-
transverse reinforcement, which is typical of those designed be- chorage failed by flexural hinging of the beam. This result clearly
fore the 1970s, and often the case of seismically retrofitted ones indicated the importance of anchoring for the FRP joint strength-
共Hakuto et al. 2000兲. This explains the important amount of re- ening techniques.
search carried out recently in order to better quantify the perfor- Antonopoulos and Triantafillou 共2003兲 conducted a large ex-
mance of substandard joints 共Park 2002; Pantelides et al. 2002; perimental program, involving 18 tests carried out on joints with-
Hakuto et al. 2000兲. In the case of beam-column joints retrofitted out transverse reinforcement, strengthened in shear with various
with FRP, others mechanisms, with additional parameters, come FRP schemes. The main objective was to evaluate the effect of a
into play, and new failure modes are to be considered, which series of parameters, mostly related to FRP, on the seismic per-
increase the complexity of the problem. From the practical point formance of the strengthened joints. Among the parameters stud-
of view, the presence of a floor slab and transverse beams in real ied were: 共1兲 the FRP ratio; 共2兲 the presence of anchoring; and 共3兲
RC frames, and the difficulties to develop retrofit schemes within the presence of transverse reinforcement in the joint. All the
the small area of the joint, clearly show the formidable challenge tested joints failed by shear, and were preceded by partial or com-
posed by the seismic retrofitting of this critical element 共En- plete debonding of FRP, either at the unanchored ends or near the
genidiz et al. 2005兲. joint corners. The main conclusions were: 共1兲 both the strength
The objectives of this study are as follows: 共1兲 to carry out a and the energy dissipation increase with, but not proportionally, to
comprehensive synthesis of reported experimental investigations the number of FRP layers; 共2兲 mechanical anchorages increase the
performed on the use of FRP for the seismic retrofit of RC build- effectiveness of FRP rehabilitation technique; and 共3兲 the effec-
ing joints; 共2兲 to use the experimental data obtained from these tiveness of FRP increases as the transverse reinforcement in the
investigations to develop a simple design method for the calcula- joint decreases.
tion of the FRP contribution to the shear strength of beam-column In the study performed by Prota et al. 共2004兲, 11 beam-column
joints externally retrofitted with FRP; and 共3兲 to elaborate on the joints were tested. The specimens were made without transverse
issues that require further research. reinforcement in the joint and with dimensions yielding to strong-
beam weak-column. The study targeted the following parameters:
共1兲 the level of the axial loading on the column; 共2兲 the type of
Literature Review on RC Joints Retrofitted with FRP FRP; and 共3兲 the FRP ratio. The innovative aspect of the proposed
technique of strengthening was the combined use of FRP lami-
nates and FRP bars. Several retrofit schemes were investigated.
Description of Previous Work
The test results showed that column failure 共CF兲 was the pre-
In the following, a brief review of the reported laboratory work dominant mode of failure, accompanied in some cases by JS fail-
performed on beam-column joints in building frames is presented. ure. However, ductile beam failures were not achieved as
Studies devoted to the seismic retrofit of RC bridge joints using intended.
FRP are not included, due essentially to the substantial difference Ghobarah and El-Amoury 共2005兲 investigated the performance
between the desired response of building and bridge structures to of beam-column joints that had anchorage strength deficiency,
earthquake loading. combined in some cases, to JS deficiency 共no transverse rein-
Geng et al. 共1998兲 carried out a series of 19 tests with the aim forcement兲. The anchorage deficiency consisted of inadequate an-
to improve, by means of carbon FRP 共CFRP兲 sheets, the ductility chorage of beam bottom bars. The experimental investigation
and the capacity of connection containing insufficient develop- involved six test specimens and tested four FRP retrofit schemes
ment length. However, the shape of the tested specimen is not using either carbon or glass fibers, and using different anchorage
fully representative of real structures. Indeed, the absence of the systems. The specimens failed according to different modes, in-
beam element in the tested specimen may generate mechanisms cluding debonding of FRP in specimen with no mechanical an-
of resistance substantially different from those encountered in chorage, which once more showed the necessity to provide
beam-column subassemblages. adequate anchorage system for FRP strengthening methods.
Mosallam 共2000兲 tested four beam-column joints, two of Mukherjee and Joshi 共2005兲 presented results of tests carried
which were tested as a control until yielding, then retrofitted, and out on 13 interior beam-column joints. Two series of tests were

50 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2010

J. Compos. Constr., 2010, 14(1): 49-61

 performed, one series with and the other without adequate anchor- ratio 共4–5%兲 before a ductile failure occurred compared to the
age of the beam bars within the joint core. The parameters studied control specimens which failed in a brittle manner at a 3% drift
were: 共1兲 the anchorage strength deficiency; 共2兲 the FRP ratio; 共3兲 ratio.
the type of FRP fibers 共glass versus carbon兲; and 共4兲 the FRP
retrofit scheme. All the retrofitted specimens failed by flexural Matrix Synthesis of Previous Work
hinging at the beam-joint interface. The test results clearly
showed the effectiveness of the FRP system regardless of FRP The synthesis of the experimental work described in the previous
schemes, and indicated that specimens strengthened using CFRP section is presented in a convenient matrix form 共Table 1兲, which
exhibited stiffer behavior than GFRP strengthened specimens. includes the relevant information on the investigations in terms of
However, the fact that all the specimens tested were made without the geometric characteristics of the beam-column joints tested, the
transverse reinforcement, and no FRP was placed within the joint deficiencies studied, the type of fiber and the FRP systems used,
Downloaded from ascelibrary.org by University of California, San Diego on 04/22/17. Copyright ASCE. For personal use only; all rights reserved.

core to address this deficiency, was rather surprising. and the modes of failure observed. This synthesis allows for
Al-Salloum and Almusallam 共2007兲 examined the perfor- qualitative identification of the parameters that may influence the
mance of interior beam-column joints retrofitted in shear using joint behavior and were not sufficiently covered in previous stud-
two different schemes of externally bonded CFRP, some test ies. From Table 1, the following observations can be drawn:
specimens with and others without mechanical anchorage. No • Geometric characteristics of the joints tested: Most of the tests
were performed on exterior joints. On the other hand, the tech-
transverse reinforcement was provided in the joints. The experi-
nical complexity which characterizes the seismic retrofitting of
mental program involved six tests performed on four subassem-
beam-column joints in presence of transverse beams and/or
blages, two of which were tested as control specimens, then
slab may explain why not a single study on seismic retrofitting
repaired and retrofitted, and then loaded again. The test results
of beam-column joints using FRP, except that by Tsonos
showed that using any of the two schemes, shear failure of the 共2008兲, has yet specifically targeted this major aspect. In terms
joint was delayed substantially. The specimens with no anchorage of the test specimen size, the dimensions used are, in general,
experienced JS failure, preceded by debonding of FRP. By con- representative of real building frames.
trast, the specimens with anchorage failed by flexural hinging of • Parameters investigated: The absence of transverse reinforce-
the beam. ment in the joint represents the most investigated parameter. In
Tsonos 共2008兲 presented results of an experimental investiga- some studies, this deficiency is coupled to that resulting from
tion involving six tests carried out on four identical exterior slab- the strong-beam weak-column design and/or inadequate an-
beam-column subassemblages with transverse stub beams. The chorage of the steel beam bars.
specimens were made with less column transverse reinforcement • FRP retrofit schemes: Carbon fibers are used in almost all
than that required by the modern Codes. The objective of this cases although more costly than the other types of fibers. The
study was to evaluate the effectiveness of two retrofit methods, choice of carbon fibers is probably motivated by the many
namely, the CFRP-jackets and RC-jackets. The control specimens performance advantages offered by these fibers in terms of
ruptured by premature shear failure whereas those retrofitted ex- strength and stiffness, compared to fibers such as glass or ara-
hibited a more ductile failure mode with the development of flex- mid. Additionally, the fact that most experimental data on
ural hinges into the beam. The test results demonstrated the strengthening of RC elements are related to carbon fibers will
effectiveness of the CFRP system. Also, the study proposed a probably continue to favor the use of CFRP composites for
practical method to predict the ultimate shear strength in the RC future investigations. The various retrofit schemes proposed
joint retrofitted with FRP, based on the confinement conceptual generally to combine the use of composites fabrics and me-
approach. The proposed method will be discussed later in this chanical anchorages systems at the end of FRP sheets by
paper. means of steel plates.
More recently, Karayannis and Sirkelis 共2008兲 performed a • Loading: So far, only unidirectional loading has been used.
series of tests on 12 external beam-column joints repaired or/and Some of the test specimens were preloaded to some level, in
strengthened with a combination of epoxy resin injections and general corresponding to the yielding point of the control
CFRP. Two series of tests were considered, depending on the specimen, unloaded, then retrofitted and loaded again. This
procedure presents the advantage to be more representative of
presence of steel reinforcement in the joint core, i.e., with or
structures retrofitted in practice which have experienced some
without. With the exception of the control specimen with no steel
cracks and/or damage. What is the influence of the precracking
reinforcement in the joint core which failed in shear, all the other
on the behavior of the retrofitted joints is however the question
specimens ruptured by flexural beam hinging 共BH兲. The test re-
never discussed in the analysis of the test results.
sults showed that the CFRP contribution to the shear capacity With regards to the findings of these laboratory works, the
depends on whether the strengthened joint is reinforced in shear following observations can be drawn:
with steel reinforcement. • The test results indicate substantial enhancements due to FRP
Finally, Pantelides et al. 共2008兲 conducted an experimental in- in terms of strength, ductility, and energy dissipation. For in-
vestigation involving eight interior beam-column joints that had stance, Antonopoulos and Triantafillou 共2003兲 obtained a
inadequate anchorage of beam bottom bars, combined to JS defi- strength increase up to about 84%. In terms of energy dissipa-
ciency 共no transverse reinforcement兲. Two series were considered, tion, specimens tested by Ghobarah and El Amoury 共2005兲
depending on whether or not the specimens were made in accor- recorded gains up to 45%. On the other hand, the test results
dance with the strong-column weak-beam concept. The CFRP showed that the stiffness degradation is significantly reduced
composites from three manufacturers were used in the strength- in the presence of FRP. At a story drift of 2%, the stiffness of
ening of the test specimens. The experimental results clearly the strengthened joints was up 125% higher than that of the
showed the effectiveness of the FRP system regardless of FRP control joint. It must be noted, however, that these figures are
type. Indeed, all the retrofitted specimens achieved a higher drift given herein only as an illustration of the effectiveness of the

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2010 / 51

J. Compos. Constr., 2010, 14(1): 49-61

 Table 1. Summary of Research on the Use of FRP for Seismic Retrofitting of RC Building Joints
52 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2010

Ghobarah and Mukherjee Al-Salloum and Karayannis Pantelides

Geng Mosallam Ghobarah El-Amoury and Antonopoulos and Prota El-Amoury and Joshi Almusallam Tsonos and Sirkelis et al.
Downloaded from ascelibrary.org by University of California, San Diego on 04/22/17. Copyright ASCE. For personal use only; all rights reserved.

共1998兲 共2000兲 共2002兲 Ghobarah 共2002兲 Triantafillou 共2003兲 共2004兲 共2005兲 共2005兲 共2007兲 共2008兲 共2008兲 共2008兲
Characteristics
and parameters
Number of Total number 19 4 4 2 18 11 6 12 4 2 4 8
specimens Control specimen共s兲 5 2 2 1 4 2 2 2 2 1 2 3
Retrofitted specimen 14 4 4 2 14 9 4 11 4 2 4 5
Type of joints Interior 3 3 3 3 3 3
Exterior 3 3 3 3 3 3
Corner
With slab 3
With transverse beam 3 3
Type of Joint transverse steel 3 NAa 3 3 3 3 3 3 3 3 3 3
deficiencies Beam bar anchorage NAa 3 3 3 3 3
Space column ties NAa 3 3
Lapped splices column 3 NAa
Weak-column/strong-beam NAa 3 3 3 3
FRP type Carbon 3 3 3 3 3 3 3 3 3 3
Glass 3 3 3 3 3 3
Aramid
FRP systems Sheets 3 3 3 3 3 3 3 3 3 3 3 3
Strips 3 3
NSM bars 3
Surface prepar 3 3 3 3
Amount of FRP 3 3
Angle of fiber 3 3 3
Mechan. anchors 3 3 3 3 3 3 3
Loading Bidirectionnel
Preloading 3 3 3 3 3 3 3 3
Influence of axial load 3 3 3 3 3 3
Modes of JS 3 3 3 3 3 3 3 3
failure Bond-slip failure 3 3
CF 3
BH 3 3 3 3 3 3 3
FD 3 3 3 3
FRP fracture
a
NA⫽not available.

J. Compos. Constr., 2010, 14(1): 49-61

 FRP retrofit technique. In other words, it is not judicious to provisions, about the joint mechanics and the role of transverse
directly compare, on the basis of these figures, the efficiency reinforcement, whether it is required to equilibrate the truss
of the various schemes proposed, partly because the joint be- mechanism or only necessary to confine the joint core, the pro-
havior involves several parameters, such as the ratio of the posed method has the following two shortcomings: 共1兲 the con-
flexural strength of column to that of the beam, the concrete finement of joint by means of FRP jackets is hard to justify since
strength, the joint reinforcement ratio, the bond conditions, the strengthening scheme typically used in seismic retrofit of
that were not necessary similar in all the reported tests. joints is not fully wrapped around the joint core 共besides, it is an
• Most studies showed that the failure due to, or initiated by, experimentally established fact that a large part of square or rect-
debonding of the FRP sheets, including delamination, repre- angular section, as it is generally the case of joints cross section,
sents a potential scenario of rupture. On the other hand, the remains unconfined even if it is fully wrapped兲 and 共2兲 the equa-
test results clearly demonstrate the important role of mechani- tion used to predict the increased joint concrete compressive
cal anchorage systems in limiting such undesirable mode of strength due to confining seems to be derived from test data ob-
Downloaded from ascelibrary.org by University of California, San Diego on 04/22/17. Copyright ASCE. For personal use only; all rights reserved.

failure. tained on columns retrofitted with FRP. However, the use of the
• Concerning the influence of geometric properties of FRP, the column analogy for the prediction of JS stress is not true because
tests results confirmed what investigations on other types of the mechanisms of resistance in joints and those encountered in
RC elements strengthened with FRP, such as beams strength- columns are substantially different 共Cheung et al. 1993兲.
ened in shear, had indicated: the performance increases signifi-
cantly with, but not proportionally, to the FRP ratio, and the Experimental Database
FRP sheets are more effective than strips. As for the effective-
ness of carbon versus glass fibers, there is only one study A database of all available studies was compiled by the writer,
reported in the literature dealing with this parameter 共Anto- containing a total of 54 tests carried out worldwide, excluding
nopoulos and Triantafillou 2003兲. Based on the result of one those tests which were irrelevant for our research study, incom-
test only, this study concluded that, for the same axial stiff- plete or dubious. A summary of this constructed database is
ness, glass fiber sheets proved marginally more effective than provided in Table 2. Of those tests, 36 simulated exterior beam-
carbon fiber sheets. column joints whereas the remaining 18 tests were models of
• Other parameters covered, here and there, also have influence interior ones. All included test specimens were subassemblages of
on the performance of the joint behavior retrofitted with FRP, RC moment-resisting frames at or above one-third scale. All
for instance: the positive effect of the column axial load as its specimens were statically determinate subassemblages and sub-
intensity increases, and the decrease in the gain in strength due jected to cyclic loading by displacing vertically the end of the
to FRP as the transverse reinforcement ratio increases. beam while restraining translation of the column ends, except
specimens tested by Prota et al. 共2004兲, and more recently those
tested by Al-Salloum and Almusallam 共2007兲, which were loaded
Existing Design Proposals by displacing laterally the end of the column while restraining the
The guidelines provided by codes and standards for the design of beam ends from translation.
structures strengthened with externally bonded FRP 关American This database gathers all relevant data, such as the geometric
Concrete Institute 共ACI兲 2008; Canadian Standards Association properties of the test specimens and of the retrofitting scheme, the
共CSA兲 2002; Fédération Internationale du Béton 共FIB兲 2001兴 do mechanical properties of the materials used, the maximum tip
not include clauses related to beam-column joints. Also, no ana- load 共taken as the average in the push and pull direction兲, and the
lytical model has been yet proposed with the exception of the mode of failure. Concrete with compressive strengths ranging be-
work of Antonopoulos and Triantafillou 共2002兲, in which the tween 19 and 43 MPa was used. On total of 54 tests reviewed,
model developed initially by Pantazopoulou and Bonacci 共1992兲 only five test specimens had transverse reinforcement in the
for the analysis of conventional RC joints has been extended to joints. In terms of mode of failure, 37 specimens failed by JS
account for the effect of externally bonded FRP. Using stress preceded by or combined with, FRP debonding 共FD兲. The other
equilibrium and strain compatibility, and from a compressive reported failure modes include: BH, CF, and bond slip 共BS兲.
stress-strain behavior of concrete based on modified compressive Where combinations of these symbols appear in the database, the
field test for concrete, the proposed model seems to offer the respective failure modes have been reported to have occurred in
potential for parametric study in order to assess the factors that the sequence given.
influence most the behavior of joint retrofitted with FRP. How- The assembled experimental data will help for use in simulat-
ever, due to the complexity of the equations derived, the practical ing the response of joints retrofitted in shear using FRP, with a
use of this model needs an important computational effort, with wide range of design parameters, and helpfully will be a valuable
the result that its implementation in a computer program is im- tool for developing rational and effective specifications for the
peratively required. On the other hand, it may be worth mention- practicing engineer. In this research study, these experimental
ing that the initial model, which was developed for conventional data are used to calibrate a simple design model to determine the
beam-column joints, assumes that the joint is properly detailed FRP contribution to the JS strength, as shown in the following.
and bond deterioration is not significant 共Pantazopoulou and
Bonacci 1992兲. Under these circumstances, the validity of these Simple Design Method of Beam-Column Joints
assumptions may be questionable for joints with poor reinforce- Retrofitted in Shear with FRP
ment details, which is typical of older beam-column subassem-
blages in need of seismic retrofit. Background
More recently, Tsonos 共2008兲 proposed a design method to
predict the JS stress based on the assumption that the role of the Horizontal Shear Stress of Joint Core
externally bonded FRP is to provide confinement to the joint core Fig. 1共a兲 shows the free-body of an exterior beam-column subas-
concrete. Beyond the conflicting views, clearly reflected in codes semblage between its points of contraflexure, typical of test speci-

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2010 / 53

J. Compos. Constr., 2010, 14(1): 49-61

 Table 2. Summary of Experimental Database on Building Joints Retrofitted with FRP
Geometry and dimensions

Subassemblage Beam Column Joint reinforcement details Results

Concrete
lc lb hb bw hc bc fc Ef f fu ␳f ␳s Pmax
Specimen IDa Typeb 共m兲 共m兲 共mm兲 共mm兲 共mm兲 共mm兲 共MPa兲 Fiber 共GPa兲 共MPa兲 FRP schemec ␤° 共%兲 共%兲 Failure 共kN兲
GHO1共T1兲 Ext. 2.75 1.75 400 250 400 250 30.8 None — — — — — 0 JS 109.0
GHO1共T1R兲 Ext. 2.75 1.75 400 250 400 250 30.6 GFRP 27.6 552 U-W+ anchor 45 0.88 0 FD, JS 120.4
GHO1共T2R兲 Ext. 2.75 1.75 400 250 400 250 30.6 GFRP 27.6 552 U-W+ anchor 45 1.76 0 BH 125.7
GHO1共T9兲 Ext. 2.75 1.75 400 250 400 250 38.0 GFRP 27.6 552 T-W 45 2.64 0 JS 127.2
EL-AM共T0兲 Ext. 2.85 1.75 400 250 400 250 30.6 None — — — — — 0 JS, BS 73.0
Downloaded from ascelibrary.org by University of California, San Diego on 04/22/17. Copyright ASCE. For personal use only; all rights reserved.

EL-AM共TR1兲 Ext. 2.85 1.75 400 250 400 250 30.6 GFRP 19 279 U-W 共1 bidir. 45 0.69 0 FD, BS 110.0
+共1 unidir.兲+
71 1,700 anchor 0 0.28
EL-AM共TR2兲 Ext. 2.85 1.75 400 250 400 250 39.5 GFRP 19 279 U-W 共1 bidir. 45 0.69 0 FD, JS 132.5
+共1 unidir.兲+
71 1,700 anchor 0 0.28
ANT共C兲 Ext. 1.30 1.00 300 200 200 200 21.6 None — — — — — 0 JS 30.1
ANT共S-C兲 Ext. 1.30 1.00 300 200 200 200 19.3 None — — — — — 0.23 JS 32.7
ANT共T-C兲 Ext. 1.30 1.00 300 200 200 200 24.6 None — — — — — 0 JS 34.9
ANT共S33兲 Ext. 1.30 1.00 300 200 200 200 26.0 CFRP 150 2,400 Strip 0 0.35 0 FD, JS 35.0
ANT共S63兲 Ext. 1.30 1.00 300 200 200 200 24.2 CFRP 150 2,400 Strip 0 1.05 0 FD, JS 39.8
ANT共S33L兲 Ext. 1.30 1.00 300 200 200 200 26.3 CFRP 150 2,400 Strip+ anchor 0 0.35 0 FD, JS 42.5
ANT共F11兲 Ext. 1.30 1.00 300 200 200 200 22.8 CFRP 230 3,450 U-W 0 0.13 0 FD, JS 42.6
ANT共F22兲 Ext. 1.30 1.00 300 200 200 200 27.2 CFRP 230 3,450 U-W 0 0.26 0 FD, JS 49.6
ANT共F21兲 Ext. 1.30 1.00 300 200 200 200 27.0 CFRP 230 3,450 U-W 0 0.26 0 FD, JS 50.7
ANT共F12兲 Ext. 1.30 1.00 300 200 200 200 29.5 CFRP 230 3,450 U-W 0 0.13 0 FD, JS 44.4
ANT共F22A兲 Ext. 1.30 1.00 300 200 200 200 27.8 CFRP 230 3,450 U-W 0 0.26 0 FD, JS 55.0
ANT共F22W兲 Ext. 1.30 1.00 300 200 200 200 29.2 CFRP 230 3,450 U-W+ anchor 0 0.26 0 JS 55.4
ANT共F22in兲 Ext. 1.30 1.00 300 200 200 200 21.0 CFRP 230 3,450 U-W 0 0.26 0 FD, JS 41.8
ANT共GL兲 Ext. 1.30 1.00 300 200 200 200 19.5 GFRP 70 2,170 U-W 0 0.42 0 FD, JS 43.6
ANT共S-F22兲 Ext. 1.30 1.00 300 200 200 200 19.0 CFRP 230 3,450 U-W 0 0.26 0.23 FD, JS 43.7
ANT共T-F33兲 Ext. 1.30 1.00 300 200 200 200 26.0 CFRP 230 3,450 L-Shaped 0 0.20 0 FD, JS 44.4
ANT共T-F22S2兲 Ext. 1.30 1.00 300 200 200 200 22.0 CFRP 230 3,450 L-Shaped+ 0 0.26 0 FD, JS 39.9
150 2,400 strips 0 0.16
PRO共L1兲 Int. 2.64 3.05 355 200 200 200 38.9 None — — — — — 0 CF 35.6
PRO共H1兲 Int. 2.64 3.05 355 200 200 200 38.5 None — — — — — 0 CF 33.3
PRO共L4兲 Int. 2.64 3.05 355 200 200 200 36.5 CFRP 264 4,323 Laminate+ 0 0.17 0 CF 49.0
108 2,014 NSM bars 0 1.17
PRO共H4兲 Int. 2.64 3.05 355 200 200 200 39.8 CFRP 264 4,323 Laminate+ 0 0.17 0 CF 61.0
108 2,014 NSM bars 0 1.17
GHO2共T-SB3兲 Ext. 2.85 1.75 400 250 400 250 30.6 None — — — — — 0 JS, BS 76.8
GHO2共T-SB8兲 Ext. 2.85 1.75 400 250 400 250 22.5 GFRP 19 279 U-W 共1 bidir. 45 0.69 0 BS 139.8
+共1 unidir.兲+
71 1,700 anchor 0 0.28
GHO2共T-SB7兲 Ext. 2.85 1.75 400 250 400 250 22.5 GFRP 19 279 U-W 共1 bidir. 45 0.69 0 JS 126.3
+共1 unidir.兲+
71 1,700 anchor 0 0.28
AL-SA共IC1兲 Int. 1.80 2.00 350 160 300 160 30.0 None — — — — — 0 JS 71.0
AL-SA共IR1兲 Int. 1.80 2.00 350 160 300 160 30.0 CFRP 61.5 738 Sheet/side 0 1.25 0 FD, JS 64.1
AL-SA共IS1兲 Int. 1.80 2.00 350 160 300 160 30.0 CFRP 61.5 738 Sheet/side 0 1.25 0 FD, BH 85.2
AL-SA共IC2兲 Int. 1.80 2.00 350 160 300 160 25.0 None — — — — — 0 JS 59.1
AL-SA共IR2兲 Int. 1.80 2.00 350 160 300 160 25.0 CFRP 61.5 738 Sheet/ side+ anchor 0 1.25 0 FD, JS 66.8
AL-SA共IS2兲 Int. 1.80 2.00 350 160 300 160 25.0 CFRP 61.5 738 Sheet/ side+ anchor 0 1.25 0 FD, BH 63.3
TSO共F1兲 Ext. 1.4 1.05 300 200 200 200 22.0 None — — — — — 0 JS NAd
TSO共FRPF1兲 Ext. 1.4 1.05 300 200 200 200 22.0 CFRP 230 3450 UW+ anchor 0 1.65 0 BH NA
TSO共FRPS1兲 Ext. 1.4 1.05 300 200 200 200 21.8 CFRP 230 3450 UW+ anchor 0 1.65 0 BH NA
KAR共A2兲 Ext. 1.80 1.10 300 200 200 200 36.4 None — — — — — 0 JS 21.0
KAR共B2兲 Ext. 1.80 1.10 300 200 200 200 36.4 None — — — — — 0.84 BH 23.3

54 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2010

J. Compos. Constr., 2010, 14(1): 49-61

 Table 2. 共Continued.兲
Geometry and dimensions

Subassemblage Beam Column Joint reinforcement details Results

Concrete
lc lb hb bw hc bc fc Ef f fu ␳f ␳s Pmax
Specimen IDa Typeb 共m兲 共m兲 共mm兲 共mm兲 共mm兲 共mm兲 共MPa兲 Fiber 共GPa兲 共MPa兲 FRP schemec ␤° 共%兲 共%兲 Failure 共kN兲
KAR共A2R兲 Ext. 1.80 1.10 300 200 200 200 36.4 CFRP 230 4,100 U-W 0 0.12 0 BH 40.3
KAR共A3兲 Ext. 1.80 1.10 300 200 200 200 36.4 CFRP 230 4,100 U-W 0 0.12 0 BH 39.5
KAR共B2R兲 Ext. 1.80 1.10 300 200 200 200 36.4 CFRP 230 4,100 U-W 0 0.12 0.84 BH 39.3
KAR共B3兲 Ext. 1.80 1.10 300 200 200 200 36.4 CFRP 230 4,100 U-W 0 0.12 0.84 BH 39.5
PAN共24–1兲 Int. 3.20 3.76 610 406 406 406 43.0 None — — — — — 0 JS 144
Downloaded from ascelibrary.org by University of California, San Diego on 04/22/17. Copyright ASCE. For personal use only; all rights reserved.

PAN共24–2兲 Int. 3.20 3.76 610 406 406 406 43.0 None — — — — — 0 JS 144
PAN共R24-3兲 Int. 3.20 3.76 610 406 406 406 43.0 CFRP 43 868 Sheet/side 60 1.33 0 FD, JS 216
PAN共R24-4兲 Int. 3.20 3.76 610 406 406 406 43.0 CFRP 106 1,184 Sheet/side 60 1.00 0 FD, JS 200
PAN共16–1兲 Int. 3.20 3.76 406 406 406 406 43.0 None — — — — — 0 BS 84
PAN共R16-2兲 Int. 3.20 3.76 406 406 406 406 43.0 CFRP 80 1,163 Sheet/side 60 0.75 0 FD, JS 115
PAN共R16-3兲 Int. 3.20 3.76 406 406 406 406 43.0 CFRP 43 868 Sheet/side 60 1.33 0 FD, JS 125
PAN共R16-4兲 Int. 3.20 3.76 406 406 406 406 43.0 CFRP 106 1,184 Sheet/side 60 1.00 0 FD, JS 130
a
GHO1= Ghoborah and Said 共2002兲, EL-AM= El-Amoury and Ghobarah 共2002兲, ANT= Antonopoulos and Triantafillou 共2003兲, PRO= Prota et al. 共2004兲,
GHO2= Ghobarah and Al-Amoury 共2005兲, AL-SA= Al-Salloum and Almusallam 共2007兲, TSO= Tsonos 共2008兲, KAR= Karayannis and Sirkelis 共2008兲,
and PAN= Pantelides et al. 共2008兲.
b
Ext. = exterior joint and Int. = interior joint.
c
U-W= U-wrap, T-W= total wrap, unidir. = unidirectional sheet, and bidir. = bidirectional sheet.
d
NA= not available.

mens used in laboratory work to idealize the RC frame Plb

connections. The beam is subjected to reversed cyclic load 共P兲 T= 共2兲
jd
applied at the beam tip. The resulting internal forces acting on the
face of the joint core under the simulated seismic forces can be where jd = distance between the tension and compression result-
depicted as illustrated in Fig. 1共b兲. Using the conditions of equi- ants in the beam at the beam-joint interface, assumed equal to
librium, the horizontal shear force across the mid-depth of the 0.75hb 共hb = height of the beam兲. Note that the stress in the tension
joint core 共V jh兲 can be expressed by steel calculated using the above equation is assumed here to be
limited to 1.25 times the specified yield strength of the tension
V jh = T − Vcol 共1兲 steel.
The horizontal column shear force can be obtained based on
where T = the force in the tension steel reinforcement of the beam equilibrium conditions 关Fig. 1共a兲兴
and Vcol = is the shear force in the column.
The tensile force in steel of the beam may be calculated by
P共lb + 0.5hc兲
Vcol = 共3兲
lc
Substituting Eq. 共2兲 and Eq. 共3兲 into Eq. 共1兲 gives

V jh = P 冋 lb lb 0.5hc
− −
jd lc lc
册 共4兲

The nominal horizontal shear stress at mid-depth of the joint core

can be written as

v jh = V jh/b jhc 共5兲

where b j = effective width of the joint core, taken hereafter as the
minimum out-of-plane dimension of the beam or column and hc
= depth of the column.

Shear Strength of Joints

To prevent diagonal compression failure, the shear stress com-
puted with Eq. 共5兲 should be limited to the nominal shear strength
of the joint, which is expressed, in all codes and standards guide-
lines, as a function of concrete strength. The ACI Committee
Fig. 1. JS of exterior connections: 共a兲 loading condition; 共b兲 free- 352-02 provisions 关American Concrete Institute 共ACI兲 2002兴, for
body diagram at the midheight of joint core instance, states

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2010 / 55

J. Compos. Constr., 2010, 14(1): 49-61

Downloaded from ascelibrary.org by University of California, San Diego on 04/22/17. Copyright ASCE. For personal use only; all rights reserved.

Fig. 2. Exterior beam-column joint subjected to seismic loading 共a兲 shear strengthening of joint core using FRP strips; 共b兲 tension force in FRP
after diagonal tension cracking initiates in joint core

v jh ⱕ ␥冑 f ⬘c 共6兲 Proposed Design Model

In this research study, a simple design model is proposed for
where ␥ = empirical constant reflecting confinement of joint by the predicting the contribution of the FRP to the shear strength of
adjoining members. exterior joint lacking transverse reinforcement. It is based on the
As for the beam-column joint cores without transverse rein- assumption that the FRP crossing a potential shear crack in the
forcement, which is typically the case of the retrofitted specimens joint core will carry that JS force in excess of the concrete con-
tested in laboratory work, Priestley 共1997兲 suggests to assume tribution to the joint capacity. In accordance with the criterion of
that the shear strength is reached at the stage of initial diagonal initial diagonal tension cracking 共Priestley 1997兲, the concrete
tension cracking of the joint core. This approach for the assess- contribution to the joint strength is assumed to be equal to the JS
ment of the shear strength of substandard joints has the advantage force which causes the first diagonal crack of the joint core, and
to take into account the influence of axial load. From Mohr’s can be evaluated by using either of the Eq. 共10兲 or Eq. 共11兲,
circle analysis, the principal tensile stress at the mid-depth of the depending on the type of joint 共i.e., interior versus exterior兲. As
joint core 共f 1兲 is given by for the FRP contribution to the JS strength, the linear behavior of

f1 =
− ␴c
2
+ 冑冉 冊 ␴c
2
2
+ v2jh 共7兲
the FRP up to rupture makes difficult the theoretical prediction of
the effective strain attained by the FRP at failure. As a result, the
ultimate FRP tensile strain, which is never reached under shear, is
from which replaced by the so-called effective strain 共␧ f,e兲, which is substan-

冑
tially lesser than the ultimate tensile strain.
␴c This approach has been used by Pantelides and Gergely 共2002兲
v jh = f 1 1+ 共8兲
f1 in order to design the thickness of the FRP composite in the joint
region of an RC bridge bent retrofitted using FRP. However, the
where ␴c = the nominal axial compressive stress on the column at FRP strain of the composite in the fiber direction was taken as a
the mid-depth of the joint core, given by fixed value 共0.002兲, based on tests results performed by the au-
N thors. Recent work by Karayannis and Sirkelis 共2008兲 also used
␴c = 共9兲 the principal tensile stress as criterion to assess the damage level
b jhc
in the joint region. In this research study, the key element is
where N = axial compressive load on the column. actually the derivation of an expression for the effective FRP
Based on the results of laboratory tests, joint failure of interior strain, based on the calibration of test data reported in the litera-
joint with no transverse reinforcement is suggested to initiate at ture. The following describes the proposed design model.
共Priestley 1997兲
Expression for Effective FRP Strain
f 1 = 0.29冑 f ⬘c 共10兲 Fig. 2共a兲 shows a diagonal tension crack in exterior joint with no
As for exterior joint, in which beam bars bent down across the transverse reinforcement. The joint core is retrofitted with FRP
back of the joint, higher principal tension stresses are expected strips inclined at an angle ␤ with respect to the beam axis. The
共Priestley 1997兲 increase of lateral load due to FRP will result in increase in the
joint principal tensile stress 共⌬f 1兲. Equilibrium conditions require
f 1 = 0.42冑 f ⬘c 共11兲 that

56 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2010

J. Compos. Constr., 2010, 14(1): 49-61

 冉
⌬f 1 b j
hc
cos ␪
冊
= V j,f sin共␪ + ␤兲 共12兲

where V j,f = the tension force in the externally bonded FRP com-
posite crossing the crack and ␪ = the angle corresponding to the
direction of the principal stresses.
The tension force in the FRP can be expressed by

V j,f = A f f f,e 共13兲

where A f = the cross-sectional area of the FRP and f f,e = the effec-
tive tensile strength of the FRP.
Downloaded from ascelibrary.org by University of California, San Diego on 04/22/17. Copyright ASCE. For personal use only; all rights reserved.

Using the following notation, the cross-sectional area of FRP

can be expressed by 关see Fig. 2共b兲兴

A f = t f 共2nw f 兲 共14兲
where

hc共tg␪ + tg␤兲
n= 共15兲
sf
and t f = nominal thickness of the FRP; n = number of FRP strips;
w f = width of a FRP strip; and s f = spacing of the FRP strips.
Let ␳ f denote the FRP reinforcement ratio

2t f w f
␳f = 共16兲
bj sf
Substituting Eqs. 共15兲 and 共16兲 into Eq. 共14兲 and rearranging it
gives

A f = ␳ f hcb j共tg␪ + tg␤兲 共17兲

As for the effective tensile strength of the FRP, it can be Fig. 3. Effective FRP strain in terms of E f ␳ f / f ⬘c 2/3: 共a兲 exterior joints;
expressed by 共b兲 interior joints

f f,e = E f ␧ f,e 共18兲

where E f is the modulus of elasticity of the FRP.
Substituting Eqs. 共17兲 and 共18兲 into Eq. 共13兲 yields ␴c,exp

冑冉 冊
cos 2␪ = 共23兲
␴c,exp 2

V j,f = ␳ f E f hcb j␧ f,e共tg␪ + tg␤兲 共19兲 2 + 共v jh,exp兲 2

2
Substituting finally Eq. 共19兲 into Eq. 共12兲 and rearranging gives
where ␴c,exp and ␯ jh,exp are, respectively, the compressive stress on
⌬f 1 the column, due to the axial force applied to the column during
␧ f,e = 共20兲 the test 关see Eq. 共9兲兴, and the experimentally measured horizontal
␳ f E f 共sin ␪ + cos ␪tg␤兲sin共␪ + ␤兲
JS 关see Eq. 共5兲兴.
For FRP fibers at an angle ␤ = 0, Eq. 共20兲 becomes Figs. 3共a and b兲 show for exterior and interior joints, respec-
tively, the variation of effective FRP strain 共␧ f,e兲 as a function of
⌬f 1 FRP stiffness, expressed with respect to the compressive strength
␧ f,e = 共21兲
␳ f E f 共sin ␪兲2 of concrete 共E f ␳ f / f ⬘c 2/3兲. This term simultaneously includes the
effects of: 共1兲 FRP thickness expressed in terms of the FRP ratio
共␳ f 兲; 共2兲 fiber type expressed in terms of the modulus of elasticity
Calibration Model
of FRP 共E f 兲; and 共3兲 tensile strength of concrete, which is a major
As mentioned previously, the key of the model is to propose a
factor influencing the behavior of the FRP-concrete interface. The
formula for the determination of the effective strain of the FRP.
latter is expressed in terms of 共f ⬘c 兲2/3 to be coherent with current
The latter is obtained by calibration of test data reported in the
guidelines with regard to the shear strengthening 关American Con-
literature on beam-column subassemblages exhibited exclusively
crete Institute 共ACI兲 2008; Canadian Standards Association
JS failure 共see Table 2兲. The effective strain of the FRP is deter-
共CSA兲 2002; Fédération Internationale du Béton 共FIB兲 2001兴.
mined using Eq. 共20兲, with a slight modification that the term ⌬f 1,
From Fig. 3共a兲, it can be observed that the effective FRP strain
is replaced by the experimentally measured increase in the joint
decreases as FRP stiffness increases. However, this figure exhibits
principal tensile stress due to FRP 共⌬f 1,exp兲; thus
a relatively high degree of data point scatter, which appears to
⌬f 1,exp reflect the large number of parameters, some of them coupled,
␧ f, e = 共22兲 that can influence to various extents the behavior of joints
␳ f E f 共sin ␪ + cos ␪tg␤兲sin共␪ + ␤兲
strengthened in shear with externally bonded FRP. In the case of
The magnitude of ␪ is calculated by the following: exterior joints, the best-fit power-type expression for the effective

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2010 / 57

J. Compos. Constr., 2010, 14(1): 49-61

 strain ␧ f,e is given by Eq. 共24兲 below. This expression should be shear resisted by the concrete and that resisted by the FRP; 共2兲 the
improved in the future when more test results become available concrete contribution, which refers to the JS force corresponding

冉 冊
to the first diagonal cracking, is conservatively assumed un-
f⬘c 2/3 1.11
changed whether or not the joint is retrofitted with FRP; 共3兲 the
␧fe = 0.61 共f ⬘c in MPa, E f in GPa兲 共24兲
E f␳ f angle of inclination of the principal tensile stresses 共␪兲, given by
As for interior joints, it can be seen that, unlike the exterior Eq. 共31兲, is assumed to be fixed; and 共4兲 no transverse steel rein-
joints, there is not enough test data available to derive a reliable forcement is provided in the joint core, which is typical of RC
expression of the effective strain of FRP 关see Fig. 3共b兲兴. This frames that were designed and built based on the pre-1970s de-
makes evident that additional experimental investigations on in- sign codes. Besides, the proposed expression of the effective
terior joints covering a wide range of FRP stiffness values are strain of FRP has been derived based on test results obtained on
required. On the other hand, this result explains why the proposed specimens with no transverse reinforcement. The existence of an
Downloaded from ascelibrary.org by University of California, San Diego on 04/22/17. Copyright ASCE. For personal use only; all rights reserved.

design model is applicable to exterior joints only. interaction between the transverse steel and the externally bonded
FRP, in which an increase of the steel ratio would result in a
decrease in the FRP contribution to the shear strength, may be
Summary of the Design Method and Validation expected.
The proposed equation to predict the shear strength of exterior To evaluate the validity of the proposed design method, the
building joints retrofitted with FRP is given by available test data of control 共nonretrofitted兲 specimens are first

冑
compared with the prescribed limiting values predicted by the
␴c ACI Committee 352-02 关American Concrete Institute 共ACI兲
v jh = f 1 1+ 共25兲
f1 2002兴, and with those given by the criterion of initial diagonal
tension cracking 共Priestley 1997兲. Then, the total experimentally
The principal tensile stress is expressed as follows:
obtained shear stresses of joint core are compared to those pre-
f 1 = f 1,c + ⌬f 1 共26兲 dicted by the proposed design method.
As for the control specimens 共see Table 3兲, the comparison is
The first term 共f 1,c兲 in this equation refers to the concrete contri- conducted by means of 11 test results extracted from different
bution to the shear strength whereas the second one 共⌬f 1兲 refers experimental investigations reported in the literature. The selected
to that of the FRP. tests correspond to both exterior and interior joints, with no trans-
According to Priestley 共1997兲, the principal tensile stress due
verse reinforcement, in which failure occurred by shear only.
to concrete for exterior joint 共for instance, with adequately an-
Table 3 also provides the ratio of column flexural strength to that
chored bars of the beam兲 can be determined as follows:
of beam as well as the axial load, expressed as a percentage of the
f 1,c = 0.42冑 f ⬘c 共27兲 product of the gross column area 共Ag兲 and the concrete uniaxial
compressive strength. Fig. 4 illustrates the results from compari-
As for the principal tensile stress corresponding to the FRP con- son. The variation of the ratio v jh,test / v jh,theo is plotted against the
tribution to the shear strength, it is given by square root of the concrete compressive strength 共冑 f ⬘c 兲. It is ob-
⌬f 1 = ␳ f E f ␧ f,e共sin ␪ + cos ␪tg␤兲sin共␪ + ␤兲 共28兲 served that the shear strengths predicted by the ACI 352-02 are, in
most cases, overly optimistic compared to the test data, whereas
For FRP fibers at an angle ␤ = 0, Eq. 共28兲 becomes the shear strengths predictions based on the criterion of initial
⌬f 1 = ␳ f E f ␧ f,e共sin ␪兲2 共29兲 diagonal tension cracking generally form a reasonable lower
bound on the test data. As a result, Eq. 共11兲 seems to be much
The effective strain ␧ f,e is determined using Eq. 共24兲. In order appropriate to assess the shear strength of joints with no trans-
to maintain the integrity of the concrete in the joint core, an upper verse reinforcement.
limit of the FRP strain equal to 0.004 is however suggested. The With regard to the total JS capacity, including the contribution
proposed value for ␧ f max is equal to that recommended by some of concrete and that of the FRP, Fig. 5 compares shear strengths
researchers 共Priestley and Seible 1995兲 for columns and also obtained from tests to the corresponding nominal values as pre-
adopted by current guidelines for FRP shear strengthening appli- dicted by the proposed design method 共see also Table 4兲. It is
cations 关American Concrete Institute 共ACI兲 2008; Fédération In- observed that the latter shows satisfactory agreement with test
ternationale du Béton 共FIB兲 2001兴 to secure activation of the data. In fact, this should not be surprising since the basis of com-
aggregate interlock mechanism. This value could be revised when parison has been data already used for calibration of the model
experimental data specific to beam-column joints will be avail- predicting the effective strain of FRP. But these are all available
able. Thus, Eq. 共24兲 becomes test data. Besides, no analytical model has been yet proposed.

␧ fe = 0.61 冉 冊 f ⬘c 2/3
Ef␳f
1.11
ⱕ ␧f max = 0.004 共30兲
Conclusions and Suggestions for Further Research
For practical purposes, a simple expression, based on the sug-
gestions made by Parra-Montesinos and Wight 共2002兲, can be The research study described in this paper is an effort to synthe-
used to estimate the angle of inclination of the principal tensile size the findings of the experimental studies carried out in the last
stresses few years on the use of externally bonded FRP for the seismic

␪ = a tan 冉冊
jd
hc
共31兲
retrofit of RC building joints. A database of these studies was also
constructed, containing a total of 54 tests carried out worldwide.
The review of these studies shows that the FRP strengthening
The proposed design method is based on the following as- technique is a promising alternative to the traditional techniques.
sumptions: 共1兲 the shear strength of the joint is the sum of the Indeed, the reported test results indicate substantial enhancements

58 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2010

J. Compos. Constr., 2010, 14(1): 49-61

 Table 3. Experimental versus Predicted Shear of Unstrengthened 共Control兲 Specimens
Experimental ACI 352R-02 Criterion of initial diagonal cracking
␴c
P / 共Ag f c⬘兲 共MPa兲 v jh 共MPa兲 f 1 共MPa兲 ␥ v jh,ACI 共MPa兲 k f 1 c 共MPa兲 v jh,CIDC 共MPa兲
a
Specimen ID Mc / Mb % Eq. 共9兲 Eq. 共5兲 Eq. 共7兲 Eq. 共6兲 Eq. 共6兲 Eqs. 共10兲 and 共11兲 Eqs. 共10兲 and 共11兲 Eq. 共8兲
GHO1共T1兲 2.89 20 6.00 5.10 2.92 0.996 5.53 0.42 2.33 4.41
El-AM共TO兲 2.89 20 6.00 3.42 1.55 0.996 5.51 0.29 1.60 3.49
ANT共C兲 0.91 5 1.15 3.20 2.68 0.996 4.63 0.42 1.95 2.46
ANT共S-C兲 0.91 5 1.15 3.48 2.95 0.996 4.38 0.42 1.85 2.35
ANT共T-C兲 0.91 5 1.15 3.71 3.18 0.996 4.94 0.42 2.08 2.60
Downloaded from ascelibrary.org by University of California, San Diego on 04/22/17. Copyright ASCE. For personal use only; all rights reserved.

GHO2共T-SB3兲 2.07b 20 6.00 3.59 1.68 0.996 5.51 0.29 1.60 3.49
AL-SA共IC1兲 0.56 20 6.00 6.73 4.37 1.245 6.82 0.29 1.59 3.47
AL-SA共IC2兲 0.51 20 5.00 5.60 3.64 1.245 6.23 0.29 1.45 3.06
TSO共F1兲 0.95 20 3.75 4.08 2.61 0.996 4.67 0.42 1.97 3.36
PAN共24–1兲 1.10 10 4.30 3.07 1.60 1.245 8.16 0.29 1.90 3.43
PAN共24–2兲 1.10 10 4.30 3.07 1.60 1.245 8.16 0.29 1.90 3.43
a
The ratio of the flexural strength of column to that of the beam.
b
As provided by the author共s兲.
c
f 1 = k冑 f c⬘, where k, an empirically determined constant, is equal to 0.42 for exterior joint in which bars bent across the back of the joint, and equal to 0.29
for exterior joints with inadequately anchored bars of the beam and interior joints 共Priestley 1997兲.

due to FRP in terms of strength, ductility, and energy dissipation tal investigations, focusing mainly on the joint behavior and
of joint cores. specifically designed to cause shear failure in the joint, are
In addition, the paper presents a simple design model for pre- therefore needed.
dicting the contribution of the FRP to the shear capacity of exte- • There is lack of a rationale behind the resistance mechanisms
rior joints. The key element in the model is the derivation of an involved in the beam-column joints retrofitted with FRP. In-
expression for the effective FRP strain. The latter is calibrated deed, most of the reviewed studies only provide a description
with test data available to date. of overall behavior 共i.e., story shear versus story drift兲, while
The analysis of research on seismic retrofit of RC building the detailed description of the sequence in which different
joints carried out thus far revealed, however, some gaps which milestones of response, such as FD, yielding of transverse re-
need to be addressed. Following are the most salient ones: inforcement and crushing of diagonal concrete struts occur, is
• The experimental objectives of numerous studies were to not always provided. Nevertheless, rational explanations of the
achieve a shift from a brittle mode of failure 共i.e., JS failure or resistance mechanisms, and consequently the development of
column hinging兲 in the deficient specimens to a ductile BH more comprehensive and rigorous design procedure, are only
mechanism in the retrofitted ones, focusing rather on the ef- possible when sufficient experimental data on both overall and
fectiveness of the retrofit scheme with regard to the behavior local behaviors is made available. Of particular interest, in this
of the entire beam-column subassemblage, including beam, context, would be the assessment of the shear strength dete-
column, and joint. As a result, there is a lack of experimental rioration of beam-column joints retrofitted with FRP under cy-
data on retrofitted test specimens that failed by JS. Experimen- clic forces.
• The retrofit schemes proposed in the literature require different

Fig. 4. Experimental versus predicted JS strength of control speci- Fig. 5. Comparison of the proposed model predictions with the ex-
mens perimental shear strength data

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2010 / 59

J. Compos. Constr., 2010, 14(1): 49-61

 Table 4. Experimental versus Predicted Shear of Strengthened Specimens
Experimental ACI 352R-02 Proposed
␴c v jh f1 v jh f 1,c ⌬f 1 v jh
共MPa兲 共MPa兲 共MPa兲 共MPa兲 共MPa兲 共MPa兲 共MPa兲
Specimen ID Eq. 共9兲 Eq. 共5兲 Eq. 共7兲 Eq. 共6兲 Eq. 共27兲 Eq. 共28兲 Eq. 共25兲
GHO1共T1R兲 6.00 5.59 3.35 5.51 2.32 1.36 5.96
GHO1共T9兲 6.00 5.55 3.31 6.14 2.59 4.07 9.18
EL-AM共TR2兲 6.00 5.52 3.28 5.51 1.82 0.51 4.40
ANT共S33兲 1.15 3.72 3.19 5.08 2.14 1.23 3.90
ANT共S63兲 1.15 4.23 3.70 4.90 2.07 1.68 4.28
Downloaded from ascelibrary.org by University of California, San Diego on 04/22/17. Copyright ASCE. For personal use only; all rights reserved.

ANT共S33L兲 1.15 4.52 3.98 5.11 2.15 1.23 3.91

ANT共F11兲 1.15 4.53 3.99 4.75 2.01 0.70 3.23
ANT共F22兲 1.15 5.27 4.73 5.19 2.19 1.40 4.12
ANT共F21兲 1.15 5.39 4.84 5.18 2.18 1.40 4.11
ANT共F12兲 1.15 4.72 4.18 5.41 2.28 0.70 3.51
ANT共F22A兲 2.88 5.84 4.58 5.25 2.21 1.40 4.84
ANT共F22W兲 1.15 5.89 5.34 5.38 2.27 1.40 4.20
ANT共F22in兲 1.15 4.44 3.90 4.56 1.92 1.40 3.85
ANT共GL兲 1.15 4.63 4.09 4.40 1.85 0.70 3.07
ANT共S-F22兲 1.15 4.64 4.10 4.34 1.83 1.40 3.76
ANT共T-F33兲 1.15 4.72 4.18 5.08 2.14 1.05 3.72
ANT共T-F22S2兲 1.15 4.24 3.71 4.67 1.97 1.68 4.18
GHO2共T-SB8兲 6.00 6.25 3.93 4.72 1.38 0.82 4.24
GHO2共T-SB7兲 6.00 5.91 3.62 4.72 1.38 0.82 4.24

levels of intensive labor and artful detailing. Besides, all these hb ⫽ height of the beam;
schemes, with the exception of one case reported in the litera- hc ⫽ height of the column;
ture, have been tested on isolated beam-column subassem- jd ⫽ flexural lever arm in the beam at the beam-joint
blages with no floor members 共i.e., transverse beams and floor interface;
slabs兲. This limits the range of their applicability, and therefore lb ⫽ beam length;
inhibits their adoption in practice. On the other hand, most of lc ⫽ column length;
the tests were performed on exterior joints. However, corner M b ⫽ beam flexural strength;
joints might represent more critical conditions in building M c ⫽ column flexural strength;
frames because of the biaxial seismic loading 共Priestley 1997兲. N ⫽ axial compressive load on the column;
Additional investigations, with particular emphasis on these n ⫽ number of FRP strip;
aspects, are strongly recommended. P ⫽ applied load at the beam tip;
Pmax ⫽ maximum applied load given by tests;
s f ⫽ center-to-center spacing of FRP strips;
Acknowledgments
T ⫽ force in the tension steel reinforcement of the
beam;
The financial support of the Fonds Québécois de la Recherche sur
t f ⫽ nominal thickness of the FRP;
la Nature et les Technologies 共FQRNT兲 through operating grants
Vcol ⫽ shear force in the column;
is gratefully acknowledged.
V j,f ⫽ tension force in the externally bonded FRP
composite crossing the crack;
Notation V jh ⫽ horizontal shear force across the mid-depth of
the joint core;
The following symbols are used in this paper: w f ⫽ width of FRP strip;
A f ⫽ cross-sectional area of FRP; ␤ ⫽ angle of inclination of the FRP fibers with
Ag ⫽ gross column area; respect to the horizontal axis;
b j ⫽ effective width of the joint core; ␥ ⫽ empirical constant reflecting confinement of
bw ⫽ width of beam web; joint by the adjoining members;
E f ⫽ elastic modulus of FRP; ⌬f 1 ⫽ increase in the joint principal tensile stress
f ⬘c ⫽ compressive strength of the concrete; due to FRP;
f f,e ⫽ effective tensile strength of FRP; ⌬f 1,exp ⫽ experimentally measured increase in the joint
f f,u ⫽ ultimate tensile strength of FRP; principal tensile stress;
f 1 ⫽ principal tensile stress at the mid-depth of the ␧ f,e ⫽ effective tensile strain of FRP;
joint core; ␪ ⫽ angle corresponding to the direction of the
f 1,c ⫽ principal tensile stress due to concrete; principal tensile stresses;
f 2 ⫽ principal compression stress at the mid-depth ␯ jh ⫽ nominal horizontal shear stress at mid-depth
of the joint core; of the joint core;

60 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2010

J. Compos. Constr., 2010, 14(1): 49-61

 ␯ jh,exp ⫽ experimentally measured horizontal JS Karayannis, C. G. and Sirkelis, G. M. 共2008兲. “Strengthening and reha-
␳f ⫽ FRP reinforcement amount; bilitation of RC beam–column joints using carbon-FRP jacketing and
␳s ⫽ transverse steel reinforcement amount; epoxy resin injection.” Earthquake Eng. Struct. Dyn., 37共5兲, 769–
␴c ⫽ nominal axial compressive stress on the 790.
Moehle, J. P. 共2006兲. “Collapse of lightly confined reinforced concrete
column at the mid-depth of the joint core; and
frames during earthquakes.” Advances in earthquake eng. for urban
␴c,exp ⫽ experimental compressive stress on the
risk reduction, S. T. Wasti and G. Ozcebe, eds., Springer, The Neth-
column.
erlands, 317–332.
Mosallam, A. S. 共2000兲. “Strength and ductility of RC moment frame
References connections strengthened with quasi-isotropic laminates.” Compos-
ites, Part B, 31共6–7兲, 481–497.
Al-Salloum, A. Y., and Almusallam, T. H. 共2007兲. “Seismic response of Mukherjee, A., and Joshi, M. 共2005兲. “FRPC reinforced concrete beam-
interior RC beam-column joints upgraded with FRP sheets. I: Experi-
Downloaded from ascelibrary.org by University of California, San Diego on 04/22/17. Copyright ASCE. For personal use only; all rights reserved.

column joints under cyclic excitation.” Compos. Struct., 70共2兲, 185–

mental study.” J. Compos. Constr., 11共6兲, 575–589. 199.
American Concrete Institute 共ACI兲. 共2002兲. “Recommendations for de- Pantazopoulou, S., and Bonacci, J. 共1992兲. “Consideration of questions
sign of beam-column connections in monolithic reinforced concrete about beam-column joints.” ACI Struct. J., 89共1兲, 27–36.
structures.” ACI 352R-02, ACI-ASCE Joint Committee 352, Farming- Pantelides, C. P., and Gergely, J. 共2002兲. “Carbon-fiber-reinforced poly-
ton Hills, Mich. mer seismic retrofit of RC bridges bent: Design and in situ valida-
American Concrete Institute 共ACI兲. 共2008兲. “Guide for the design and tion.” J. Compos. Constr., 6共1兲, 52–60.
construction of externally bonded FRP systems for strengthening con- Pantelides, C. P., Hansen, J., Nadauld, J., and Reaveley, L. D. 共2002兲.
crete structures.” ACI 440 2R, ACI Committee 440, Farmington Hills, “Assessment of reinforced concrete building exterior joints with sub-
Mich. standard details.” Pacific Earthquake Eng. Res. Center (PEERC) Rep.
Antonopoulos, C. P., and Triantafillou, T. C. 共2002兲. “Analysis of FRP No. 2002-18.
strengthened RC beam-column joints.” J. Compos. Constr., 6共1兲, 41– Pantelides, C. P., Okahashi, Y., and Reaveley, L. D. 共2008兲. “Seismic
51. rehabilitation of reinforced concrete frame interior-beam joints with
Antonopoulos, C. P., and Triantafillou, T. C. 共2003兲. “Experimental in- FRP composites.” J. Compos. Constr., 12共4兲, 435–445.
vestigation of FRP-strengthened RC beam-column joints.” J. Compos. Park, R. 共2002兲. “A summary of results of simulated seismic load tests on
Constr., 7共1兲, 39–49. reinforced concrete beam-column joints, beams and columns with
Bakis, C. E., et al. 共2002兲. “Fiber-reinforced polymer composites for substandard reinforcing details.” J. Earthquake Eng., 6共2兲, 147–174.
construction—State of the art review.” J. Compos. Constr., 6共2兲, 73– Park, R. 共2003兲. “Some controversial aspects of the seismic design of
187. reinforced concrete building structures.” New Zealand Nat. Soc.
Beres, A., Pessiki, S. P., White, R. N., and Gergely, P. 共1996兲. “Implica- Earthquake Eng. Bull, 36共3兲, 165–188.
tions of experiments on the seismic behavior of gravity load designed Parra-Montesinos, G. J. and Wight, J. K. 共2002兲. “Prediction of strength
RC beam-column connections.” Earthq. Spectra, 12共2兲, 185–198. and shear distortion in R/C beam-column joints.” ACI J., SP197-10,
Canadian Standards Association 共CSA兲. 共2002兲. “Design and construction 191–214.
of building components with fiber-reinforced polymer.” CSA S806-02, Priestley, M. J. N. 共1997兲. “Displacement based seismic assessment of
Canadian Standards Association, Mississauga, Ont. reinforced concrete buildings.” J. Earthquake Eng., 1共1兲, 157–192.
Cheung, P. C., Paulay, T., and Park, R. 共1993兲. “Behavior of beam- Priestley, M. J. N., and Calvi, G. M. 共1991兲. “Towards a capacity-design
column joints in seismically loaded RC frames.” Struct. Eng., 71共8兲, assessment procedure for reinforced concrete frames.” Earthquake
129–138. Spectra, 7共3兲, 413–437.
El-Amoury, T., and Ghobarah, A. 共2002兲. “Shear rehabilitation of beam- Priestley, M. J. N., and Seible, F. 共1995兲. “Design of seismic retrofit
column joints using GFRP sheets.” Eng. Struct., 24共11兲, 1397–1407. measures for concrete and masonry structures.” Constr. Build. Mater.,
El-Gawady, M. A., Lestuzzi, P., and Badoux, M. 共2005兲. “In-plane seis- 9共6兲, 365–377.
mic response of URM walls upgraded with FRP.” J. Compos. Constr., Prota, A., Nanni, A., Manfredi, G., and Cosenza, E. 共2004兲. “Selective
9共6兲, 524–535. upgrade of under-designed reinforced beam-column joints using car-
Engindeniz, M., Kahn, L. F., and Zurieck, A. 共2005兲. “Repair and bon fiber-reinforced concrete.” ACI Struct. J., 101共5兲, 699–707.
strengthening of reinforced concrete beam-column joints: State of the Saatcioglu, M. 共2003兲. “Research on seismic retrofit and rehabilitation of
art.” ACI Struct. J., 102共2兲, 187–197. reinforced concrete structures.” Proc., 31st Annual Conf. of the Cana-
Fédération Internationale du Béton 共FIB兲. 共2001兲. “Externally bonded dian Society for Civil Engineering, CSCE, M. Massiéra and G. Poi-
FRP reinforcement for RC structures.” FIB: Bulletin 14, Fédération tras, eds., Vol. GCU-542, Canadian Society for Civil Engineering,
Internationale du Béton, Lausanne, Switzerland. Moncton, Canada, 10.
Geng, Z.-J., Chajes, M. J., Chou, T.-W., and Yen-Cheng Pan, D. 共1998兲. Saatcioglu, M. 共2006兲. “Seismic risk mitigation through retrofitting non-
“The retrofitting of reinforced column-to-beam connections.” Com- ductile concrete frame systems.” Advances in earthquake eng. for
pos. Sci. Technol., 58共8兲, 1297–1305. urban risk reduction, S. T. Wasti and G. Ozcebe, eds., Springer, The
Ghobarah, A., and El-Amoury, T. 共2005兲. “Seismic rehabilitation of de- Netherlands, 179–194.
ficient exterior concrete frame joints.” J. Compos. Constr., 9共5兲, 408– Saatcioglu, M., et al. 共2001兲. “The August 17, 1999, Kocaeli 共Turkey兲
416. earthquake—Damage to structures.” Can. J. Civ. Eng., 28共4兲, 715–
Ghobarah, A., and Said, A. 共2002兲. “Shear strengthening of beam-column 737.
joints.” Eng. Struct., 24共7兲, 881–888. Seible, F., Priestley, M. J. N., Hegemier, G. A., and Innamorato, D.
Hakuto, S., Park, R., and Tanaka, H. 共2000兲. “Seismic load tests on inte- 共1997兲. “Seismic retrofit of RC columns with continuous carbon fiber
rior and exterior beam-column joints with substandard reinforcing de- jackets.” J. Compos. Constr., 1共2兲, 52–62.
tails.” ACI Struct. J., 97共1兲, 11–25. Thermou, G. E., and Elnashai, A. S. 共2006兲. “Seismic retrofit schemes for
Japanese Reconnaissance Team 共JRT兲. 共2004兲. Boumerdes earthquake: RC structures and local-global consequences.” Prog. Struct. Eng.
The 21st May 2003, M. Hamada and K. Ramdane, eds., Japan Assoc. Mater., 8共1兲, 1–15.
Earthquake Eng. 共JAEE兲, Japan Soc. Civ. Eng. 共JSCE兲, Architectural Tsonos, A. G. 共2008兲. “Effectiveness of CFRP-jackets and RC-jackets in
Inst. Japan 共AIJ兲, and Japan Geotechnical Soc. 共JGS兲, Algiers, Alge- post-earthquake and pre-earthquake retrofitting of beam–column sub-
ria. assemblages.” Eng. Struct., 30共3兲, 777–793.

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2010 / 61

J. Compos. Constr., 2010, 14(1): 49-61