Professional Documents
Culture Documents
Akguzel 2012
Akguzel 2012
Abstract: In this study, an analytical procedure for the evaluation of the expected performance of existing reinforced concrete (RC) beam-
column joints before and after being retrofitted using fiber-reinforced polymer (FRP) composite materials is presented. Focus is given on the
evaluation of the shear-strength versus deformation properties of the panel zone region either in the as-built or FRP-retrofitted configuration.
Based on experimental and numerical evidence as well as on physical models representing the mechanics of the joint region, principal tensile
stresses versus joint shear deformation relationships are adopted and preferred to more traditional nominal shear-strength rules, to evaluate,
within a step-by-step iterative procedure, the combination of the joint shear contribution provided by the FRP composite material and that
provided by the concrete core alone. The use of principal stresses allows one to directly account for the effects of variation of axial load,
typically neglected in the assessment and retrofit of beam-column joints. The hierarchy of strength and sequence of events (damage mech-
anisms) expected within a beam-column subsystem are visualized via M-N interaction performance domains, used as a basis for a
performance-based retrofit philosophy. Specific limit states or design objectives are targeted, with attention given to both strength and de-
formation limits. The proposed analytical procedure is validated on the results of a set of experimental tests available in the literature. With
the intention to provide a simple design tool that can be easily implemented by practicing engineers, a worked example for the evaluation of
the expected performance of an FRP retrofitted beam-column joint is provided and used as a basis for a parametric study to illustrate the
effects of different strengthening schemes on the behavior of strengthened exterior joint panels under various axial load levels. DOI: 10.1061/
(ASCE)CC.1943-5614.0000242. © 2012 American Society of Civil Engineers.
CE Database subject headings: Rehabilitation; Fiber reinforced polymer; Analytical techniques; Beam columns; Reinforced concrete;
Shear failures; Composite materials.
Author keywords: Rehabilitation; Fiber reinforced polymers; Analytical techniques; Beam columns; Concrete; Reinforced; Shear failure.
of the planar region, formulations were derived to represent differ- specified limit states (e.g., beam or column hinging, “cracking,”
ent damage-limit states associated with the crushing of the concrete equivalent “yielding” or extensive damage of the joint core, as well
or the debonding or failure of the FRP. Almusallam and Al-Salloum as FRP debonding or failure) and written in terms of equivalent
(2007b) more recently applied the aforementioned model on FRP- moment in the column corresponding to the occurrence of that
strengthened “interior joints” and further extended it to predict specific limit state. This can be easily achieved following equilib-
diagonal tensile stresses in the joint. In both applications, computer rium considerations within the beam-column-joint system. At a
programs were developed to track the state of stress and strain in later stage, in order to evaluate and control the governing mecha-
strengthened RC joints. It should be noted that the formulation nisms in the joint subassembly under different demand conditions
given in the Pantazopoulou and Bonacci (1992) model was origi- of a retrofitted beam-column joint, the actual sequence of events
nally derived for interior beam-column joints designed according to should be determined by comparing demand and capacities. This
current code provisions and thus was based on many assumptions can be done within an M-N (moment-axial load) diagram, also re-
such as a good bond of the beam-column reinforcements, which is
ferred to by the authors as performance domain. An appropriate
harder to assume when dealing with joints designed according to
rehabilitation solution would aim to rearrange the existing sequence
preseismic codes.
of events according to capacity design considerations.
Research Significance
design. Hence, brittle local mechanisms, such as shear failure of the phase, both shear and flexural capacity of the elements need to
joint and column, and global mechanisms, such as weak column, be evaluated and compared.
strong beam, were prevented by upgrading the internal hierarchy A layer-by-layer iterative procedure (sort of fiber-section
of strength such that the beam plastic hinging (thus insuring a analysis) is utilized to obtain the moment-curvature response of
beam-sway mechanism) became the first predicted “event.” The ex- the critical sections by idealizing the cross section as a series of
pected hierarchy of strength and sequence of events for the as-built rectangular layers to evaluate the sectional forces corresponding
and retrofitted joints were evaluated using the concept of M-N per- to a given strain distribution. The Bernoulli-Navier hypothesis
formance domain (Pampanin et al. 2007), as shown in Fig. 1(c). of “plane sections remain plane” after bending is assumed. The po-
sition of the neutral axis is first estimated, and a series of iterations
Effects of Varying Axial Load are carried out until both equilibrium and compatibility conditions
are satisfied [Fig. 2, Eq. (1)]. Using the nomenclature given in Fig. 2
Two different demand conditions corresponding to either a constant
for axial force N h , force equilibrium at any level of response can be
axial load or a varying axial load are illustrated, in order to dem-
written as
onstrate the influence of the axial load variation in the sequence Z
of events formation. It is, in fact, important to note that most of X X
N h ¼ f c ðyÞbw ts þ f si Asi þ f f i Af i ð1Þ
the experimental work regarding the retrofitting of exterior beam-
column joints is typically conducted under constant axial load.
The moment capacity of the retrofitted section can be found by
Because of the effects of lateral loads on a frame as well as the
taking moments of the stress resultants about the section centroid
vertical component of ground motions, a high level of axial load
Z X X
variation in the exterior beam-column joint can be experienced.
As shown in Fig. 1(c) for the as-built condition, the sequence of M b;Rij ¼ f c ðyÞbw t s yi þ f si Asi yi þ f f i Af i yi ð2Þ
events or damage (e.g., 1 ¼ joint, 2 ¼ beam, 3 ¼ column) are,
in this case, the same under both demand curve conditions, regard- The material constitutive relationships include refined models
less of the assumption of the axial load (see the numbers in circles). for the confined and unconfined concrete (Mander et al. 1988), for
Two different retrofit schemes (in this study referred as Scheme 1 the reinforcing steel (Dodd et al. 1995) and a simplified linear-
and Scheme 2) on the same joint may produce totally different se- elastic stress-strain relationship for the FRP composite material
quences of events (see the numbers in rectangular boxes) when provided by the supplier. For the sake of simplicity, instead of using
considering the actual variation of axial load demand. a smeared concrete model (Mander et al. 1988), a “segmented core
If assuming, for the sake of simplicity, a constant axial load, the model” (Allington 2003) approach, along with the simplified
joint-strengthening Scheme 1 may seem sufficient and adequate to method for discretization proposed by Manfredi et al. (2001), has
protect the joint region while inducing a ductile flexural hinge in been adopted in this study in the section analysis. The cross section
the beam. However, due to the reduction of axial load during the is thus divided into two regions: unconfined and effectively con-
frame sway, the actual strength of the joint would be lower (in the fined. The effectively confined area is subjected to the full lateral
direction of reduction of axial load) of that of the beam, thus confining pressure, due to yielding of the transverse reinforcement
leading to a premature damage of the joint even after the retrofit and confining pressure provided by an equivalent FRP jacket
intervention. (Wang and Restrepo 2001). According to the segmented core
Hence, an alternative retrofit Scheme 2 would be required to model and the simplified form proposed by Manfredi et al. (2001),
guarantee an appropriate ductile failure mode, with the formation the confined region is determined. The confinement effects of the
of flexural hinges in the beam in all realistically predicted demand FRP on the section can be taken into account following procedures
conditions (see number 3 in rectangular box). Recent experimental available in the literature (e.g., Spoelstra and Monti 1999).
studies conducted by the authors have confirmed the importance of
accounting for the variation of axial load (Akguzel and Pampanin
2010) and will be used in the following paragraphs as a validation
of the analytical assessment and retrofit procedure.
Analytical tools to assess the capacity of each element within
the beam-column-joint system are given in the following sections.
As suggested in many FRP design guidelines [Federation Inter-
national du Beton (FIB) 2001, 2006; American Concrete Institute
(ACI) Committee 440.2, 2007], additional (material-based or
section-related) strength reduction factors may be used to reflect
higher uncertainty (or better lesser existing knowledge) on the FRP
Fig. 2. Layer-by-layer sectional analysis of strengthened RC beam
material and system behavior, when compared to more traditional
section with n layers of rebar
reinforced and prestressed concrete solutions.
Fig. 4. Exterior RC joint shear resisting mechanisms: (a) forces acting on as-built joint; (b) kinematics of as-built joint; (c) forces acting on a
retrofitted joint; (d) horizontal force equilibrium; (e) vertical force equilibrium
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
p2t þ pt hNc bvc bj hc ðlc hb Þ
Mj ¼ ð11Þ
½lc ðlb hc =2Þ=0:9dlb 1 2
2l0b lc 1:8dlb
ω¼ ð13Þ
0:9dlb Ae ðlc hb Þ
where l0b = half clear span length; and Ae = effective area in the joint.
In the as-built and retrofitted conditions, the equivalent joint
Fig. 5. Exterior RC joint strengthening with FRP: schematic illustra-
moment, M j , would be referred to as M jo and M j;Rij , respectively.
tion of design dimensions
Rij, which stands for the retrofitting scheme, is explained in the
design example section.
bj ¼ minðbc ; bc þ 0:5hc Þ if bc ≥ bw ð7Þ
Evaluation of Joint Shear Capacity after FRP
bj ¼ minðbw ; bc þ 0:5hc Þ if bc ≤ bw ð8Þ Strengthening
Prota et al. (2001) also confirmed that the limit of 0:3f 0c for the
principal compression stress at failure of the joint could be a rea- retrofitted configuration. Therefore, the evaluation of hierarchy
sonable design criterion, even though it could be conservative for of strength and sequence of events are carried out to support the
high axial load ratios. By using the aforementioned relationships, definition of the most appropriate retrofit solutions to adopt.
the FRP contribution to the increase in the joint shear capacity of The specimen 2DR4 was retrofitted according to the R21
poorly detailed beam-column joints can be adequately estimated in scheme. Note that the retrofitting scheme is expressed as Rij, where
terms of principal tensile stress, ptf , as a function of the amount and i = number of sheets applied on the beam surface; and j = number of
layout of the externally bonded reinforcement. sheets on the column (e.g., R12 suggests one FRP horizontal layer
The whole procedure without iteration (as a valid option to the on the beam and two vertical layers on the column). SikaWrap®-
iterative approach) to assess the capacity of FRP retrofitted exterior
100G type high-strength E-glass unidirectional fiber polymers were
beam-column joints with plain round bars and end hooks is sum-
used for the retrofitting. More information regarding the actual FRP
marized as a flow chart in Fig. 6.
application and testing can be found in the study by Akguzel and
Pampanin (2010). The summary of geometric and material proper-
Design Example ties and FRP retrofit layout is given as follows.
• Geometric data: hb ¼ 330 mm, bw ¼ 230 mm, hc ¼ 230 mm,
A full design example of the FRP retrofit intervention, following bc ¼ 230 mm, lb ¼ 1;524 mm, lc ¼ 2;000 mm, l0b ¼ 1;409 mm,
the aforementioned analytical procedure, is in this study given d ¼ 305 mm pffiffiffiffi
with reference to the specimen 2DR4 tested at the University of • Material properties: f 0c ¼ 18:4 MPa, Ec ¼ 5;700 f 0c ¼
Canterbury. One single level of axial load, N v ¼ 50 kN, is consid- 24;450 MPa, E f ¼ 76;000 MPa, t f ¼ 0:36 mm, εf u ¼ 2:8%
ered for this design example. By varying the level of N v , a M-N • Beam and column FRP scheme: nf b ¼ 2, nsf ;b ¼ 2, d f b ¼
interaction diagram could be established for the as-built and 300 mm, nf c ¼ 1, nsf ;c ¼ 2, d f c ¼ 200 mm
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
¼ ð0:64Þ fð76;000Þ½ð0:3Þð18:4Þð2=3Þ g=½ð0:36Þð2Þ ¼ ðð50;000Þ=ð230Þð230Þð0:00272Þð0:00569Þð76;000Þ2 ð0:004ÞÞ
¼ 300:1 kN ¼ 0:0027
Fig. 7. Effects of GFRP strengthening on the principal tensile stress versus shear deformation behavior of the joint under varying axial load
Step 4. Shear deformation angle Step 7. Calculate principal tensile strength contribution due to
FRP
2ðε1 εt Þ 2ð0:007 0:004Þ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
γ¼ ¼ ¼ 0:006 ptf ¼ f v =2 þ ðf v =2Þ2 þ v2f
tan θ 0:99
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Downloaded from ascelibrary.org by University of California, San Diego on 04/22/17. Copyright ASCE. For personal use only; all rights reserved.
pffiffiffiffi pffiffiffiffiffiffiffiffiffi Step 9. Calculate the capacity of the FRP retrofitted joint in
0:15 f 0c ¼ 0:15 18:4 ¼ 0:64 MPa terms of equivalent moment
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
p2tt þ ptt f v ð1:94Þ2 þ ð1:94Þð0:95Þ
Step 6. Calculate nominal shear-strength contribution due to M j;Rij ¼ ð103 Þ ¼ ð103 Þ
FRP ω 129:86
¼ 18:2 kNm
ρf t E f εt ð0:00569Þð76;000Þð0:004Þ
vf ¼ ¼ ¼ 1:75 MPa As evident in this numerical example, the proposed assessment
tan θ ð0:99Þ procedure can be easily implemented into a worksheet/spreadsheet
Fig. 9. Debonding in specimen 2DR3: (a) comparison of hysteresis loops of 2DR2 (constant axial load) and 2DR3 (varying axial load); (b) average
strain distribution in the joint FRP during debonding initiation in the bottom of the joint; (c) joint damage under FRP sheet at the end of test (images by
U. Akguzel)
A simple analytical procedure has been proposed here for the as- The following symbols are used in this paper:
sessment of the seismic behavior of existing exterior beam-column A, B, C = quadratic equation coefficients;
joints and the design of a performance-based FRP-retrofitting Ae = effective area in the joint;
intervention. Af = FRP area;
A detailed assessment of each joint component within the beam- As = reinforcement area;
column-joint subassembly is first performed in its as-built configu- bc = width of column;
ration. The component capacities are compared within a M-N bj = effective width of joint;
(moment-axial load) performance domain in order to establish bw = width of beam;
the internal hierarchy of strength and the sequence of events before C c = compressive force provided by concrete;
the retrofit intervention, while accounting for the correct demand C s = compressive force provided by reinforcement;
Downloaded from ascelibrary.org by University of California, San Diego on 04/22/17. Copyright ASCE. For personal use only; all rights reserved.
nsf ;c = number of column sides covered with FRP; Al-Salloum, Y. A., and Almusallam, T. H. (2007a). “Seismic response of
pc = principal compression stress; interior beam-column joints upgraded with FRP sheets. I: Experimental
pt = principal tensile stress; study.” J. Compos. Constr., 11(6), 575–589.
Al-Salloum, Y. A., and Almusallam, T. H. (2007b). “Seismic response of
ptc = principal tension strength contribution due to plain
interior beam-column joints upgraded with FRP sheets. II: Analysis and
concrete; parametric study.” J. Compos. Constr., 11(6), 575–590.
ptf = principal tensile strength contribution due to FRP; Antonopoulos, C., and Triantafillou, T. C. (2002). “Analysis of FRP-
ptt = total principal tensile strength; strengthened beam-column joints.” J. Compos. Constr., 6(1), 41–51.
Rij = FRP retrofit scheme; Antonopoulos, C., and Triantafillou, T. C. (2003). “Experimental investi-
T = tensile force at the bottom beam bars; gation of FRP-strengthened RC beam-column joints.” J. Compos.
T f = tension force in FRP; Constr., 7(1), 39–49.
T s = tension force in steel; Calvi, G. M., Magenes, G., and Pampanin, S. (2002). “Relevance of beam-
tan θ = average direction of the principal stresses; column damage and collapse in RC frame assessment.” J. Earthquake
t = transverse (column) direction; Eng., 6(1), 75–100.
tf = FRP thickness per layer; Dodd, L. L., and Restrepo, J. I. (1995). “Model for predicting cyclic behav-
ts = strip thickness in the cross section; ior of reinforcing steel.” J. Struct. Eng., 121(3), 433–445.
El-Amoury, T., and Ghobarah, A. (2002). “Seismic rehabilitation of beam-
V b = beam shear force;
column joints using GFRP sheets.” Eng. Struct., 24(11), 1397–1407.
V c = horizontal column shear force;
El-Amoury, T., and Ghobarah, A. (2005). “Seismic rehabilitation of
V jh = horizontal shear force acting on the joint core; deficient exterior concrete frame joints.” J. Compos. Constr., 5(1),
y = distance from the centroidal axis; 408–416.
α = proportionality coefficient for axial load variation; Federation International du Beton (FIB). (2001). “Externally bonded
γ = average angle of joint shear distortion; FRP reinforcement for RC structures.” FIB Bull. No. 14, Lausanne,
ε1 = maximum principal strain in the joint panel; Switzerland.
ε2 = minimum principal strain in the joint panel; Federation International du Beton (FIB). (2003). “Seismic assessment and
εc = concrete strain; retrofit of reinforced concrete buildings.” FIB Bull. No. 24, Lausanne,
εf = FRP strain; Switzerland.
εf deb = debonding FRP strain; Federation International du Beton (FIB). (2006). “Retrofitting of concrete
εf u = ultimate FRP strain; structures by externally bonded FRPs with emphasis on seismic appli-
cations.” FIB Bull. No. 35, Lausanne, Switzerland.
εl = average normal strain along normal direction l of
Gergely, J., Pantelides, C. P., and Reaveley, L. D. (2000). “Shear strength-
joint panel; ening of R.C. T-joints using CFRP composites.” J. Compos. Constr.,
εs = reinforcement strain; 4(2), 56–64.
εt = average normal strain along normal direction t of Ghannoum, W. M., Moehle, J. P., and Bozorgnia, Y. (2006) “Collapse of
joint panel; lightly confined reinforced concrete frames during earthquakes.” Proc.,
θ = direction of the principal compression stresses; 8th U.S. National Conf. on Earthquake Engineering, Earthquake Engi-
vjh = joint nominal shear stress; neering Research Institute (EERI), El Cerrito, CA.
vf = nominal shear stress contribution due to the FRP; Ghobarah, A., and Said, A. (2001). “Seismic rehabilitation of beam-column
ρb = total main beam reinforcement ratio; joints using FRP laminates.” J. Earthquake Eng., 5(1), 113–129.
ρf t = FRP reinforcement ratio in the transverse direction; Hakuto, S., Park, R., and Tanaka, H. (2000). “Seismic load tests on interior
ρf l = FRP reinforcement ratio in the longitudinal and exterior beam-column joints with substandard reinforcing details.”
ACI Struct. J., 97(1), 11–25.
direction;
Hertanto, E. (2006). “Seismic assessment of pre-1970s reinforced concrete
ρs = stirrup reinforcement ratio; beam-column joint subassemblies.” M.E. dissertation, Dept. of Civil
σ1 = maximum principal stress in the concrete; Engineering, Univ. of Canterbury, Christchurch, New Zealand.
σ2 = minimum principal stress in the concrete; Holzenkämpfer, P. (1994). “Ingenieurmodelle des verbundes geklebterbe-
σl = average in concrete along direction l; wehrung für betonbauteile.” Ph.D. dissertation, TU Braunshwieg (in
σt = average in concrete along direction t; German).
ϕ = curvature of reinforced concrete section; and Mander, J. B., Priestley, M. J. N., and Park, R. (1988). “Theoretical stress-
ω = geometric coefficient. strain model for confined concrete.” J. Struct. Eng., 114(8), 1804–1826.
Manfredi, G., and Realfonzo (2001). “Modellazione del comportamento di
elementi presso-inflessi in c.a. confinati con tessuti in materiale com-
posito.” Tech. Rep., 10 Convegno Nazionale “L’ingegneria Sismica in
References Italia,” Potenza-Matera, Italy.
Neubauer, U., Rostasy, F. S. (1997). “Design aspects of concrete structures
American Concrete Institute (ACI) Committee 440.2. (2002). “Guide for strengthened with externally bonded CFRP-plates.” Proc., 7th Int. Conf.
the design and construction of externally bonded FRP systems for on Structural Faults and Repair, London, 109–118.
bents under simulated seismic loads.” Earthquake Spectra, 17(3), composite materials (FRP).’’ Proc., 13th Hellenic Concrete Conf.,
507–530. Vol. 1, Rethymno, Greece, 455–466 (in Greek).
Paulay, T., and Priestley, M. J. N. (1992). Seismic design of reinforced con- Wang, Y. C., and Restrepo, J. I. (2001). “Investigation of concentrically
crete and masonry buildings, Wiley, New York. loaded reinforced concrete columns confined with glass fiber-reinforced
Priestley, M. J. N. (1997). “Displacement-based seismic assessment polymer jackets.” ACI Struct. J., 98(3), 377–385.