You are on page 1of 16

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/327039791

Concentrated and Distributed Plasticity Models for Seismic Repair of Damaged


RC Bridge Columns

Article in Journal of Composites for Construction · October 2018


DOI: 10.1061/(asce)cc.1943-5614.0000879

CITATIONS READS
14 1,725

2 authors:

Ruoyang Wu Chris Pantelides


University of Utah University of Utah
25 PUBLICATIONS 113 CITATIONS 179 PUBLICATIONS 3,619 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Impact Loading and Shear Brittle Failure Analysis of Galvanized Steel Light Pole View project

PhD Research and Dissertation View project

All content following this page was uploaded by Ruoyang Wu on 15 August 2018.

The user has requested enhancement of the downloaded file.


Concentrated and Distributed Plasticity Models for
Seismic Repair of Damaged RC Bridge Columns
Ruo-Yang Wu, S.M.ASCE 1; and Chris P. Pantelides, M.ASCE 2

Abstract: Two models, Model Fiber and Model Rotational Spring (RS), simulating the seismic performance of repaired column-to-cap
beam/footing connections using a carbon fiber-reinforced polymer (CFRP) donut are presented in this paper. In Model Fiber, distributed
Downloaded from ascelibrary.org by Marriott Lib-Univ Of UT on 08/15/18. Copyright ASCE. For personal use only; all rights reserved.

plasticity was assumed over a plastic hinge length of the nonlinear beam-column element. In Model RS, concentrated plasticity was
considered using a nonlinear moment rotational spring located at the repaired cross section. Previous concrete damage and low-cycle fatigue
of longitudinal steel bars as well as bond-slip between the damaged steel bars and surrounding concrete were included in the proposed
numerical models. Numerical simulations show that the results are in good agreement with the experiments in terms of structural response,
cumulative hysteretic energy, and moment-rotation capacity. Model Fiber can also predict the local response and low-cycle fatigue of the
longitudinal steel bars. Using appropriate pinching parameters, Model RS can model accurate pinching behavior for both the repaired cast-in-
place (CIP) and precast concrete specimens. DOI: 10.1061/(ASCE)CC.1943-5614.0000879. © 2018 American Society of Civil Engineers.
Author keywords: Numerical model; Fiber element analysis; Fiber-reinforced polymer composites; Plastic hinge relocation;
Seismic repair; Low-cycle fatigue; Bond slip; Concrete; Damage.

Introduction column longitudinal bar buckling and fracture and severe concrete
crushing in the column plastic hinge. In addition, the method has
An extensive amount of research has been developed regarding the been used for two severely damaged precast concrete (PC) column-
seismic retrofit of reinforced concrete (RC) structures using con- to-cap beam connections; in one of the specimens, the column was
crete jackets, steel jackets, prestressed steel jackets, and fiber rein- repaired for the second time, and for the second specimen the col-
forced polymer (FRP) composite jackets (Priestley et al. 1994; umn was totally separated from the cap beam before the repair.
Saadatmanesh et al. 1997; Seible et al. 1997; Xiao and Ma 1997; Based on the experimental results, the seismic performance in
Gergely et al. 1998; Pantelides and Moran 2013); however, rapid terms of lateral force capacity, displacement capacity, energy dis-
seismic repair of severely damaged columns has not been studied sipation, and displacement ductility were successfully restored
widely (He et al. 2013; Rutledge et al. 2014). Longitudinal steel (Wu and Pantelides 2017a, b).
bars yield, buckle, and fracture in bridge column plastic hinge re- Bond-slip effects must be considered to accurately determine
gions during strong earthquakes; this leads to the loss of flexural the structural response because bond slip failure occurs in most
and shear strength of RC columns (He et al. 2015). Most of the RC concrete structures, especially poorly confined concrete joints
existing repair methods are not applicable to RC columns with (Harajli 2009). When repairing a severely damaged structure, the
such severe damage; in addition, the seismic repair should be con- bond-slip between damaged longitudinal steel bars and surround-
structed rapidly to enhance seismic resilience of communities. ing concrete is critical for accurately determining the structural re-
The CFRP donut was originally developed for repairing precast sponse using numerical models. Damaged steel properties could be
RC columns connected with splice sleeves (Parks et al. 2016). The implemented with a reduction in the elastic modulus according to
method includes shape modification of the original column cross the maximum steel bar strain experienced by the original members
section by post-tensioning an FRP shell using expansive cement (Vosooghi and Saiidi 2013). Several models with modified steel
concrete (Yan et al. 2007). A seismic repair method utilizing a properties have been developed to consider bond-slip behavior in
carbon fiber-reinforced polymer shell and epoxy anchored headed the analysis (Zhao and Sritharan 2007; D’Amato et al. 2012; Ameli
steel bars has been implemented to repair severely damaged con- and Pantelides 2016). One method is to modify the steel properties
crete columns by relocating the column plastic hinge (Wu and based on the assumed bond-slip relationship and use a modified
Pantelides 2017a, b, c, d). The repair method was improved and steel constitutive curve in the numerical model.
implemented for two severely damaged cast-in-place (CIP) RC col- There is little research regarding numerical models considering
umn-to-cap beam/footing connections, with damage including bond-slip, longitudinal bar fracture/buckling, and damaged con-
crete for repaired RC column-to-cap beam/footing connections
1
Ph.D. Candidate, Dept. of Civil and Environmental Engineering, Univ. (He et al. 2016). In addition, low-cycle fatigue of damaged longi-
of Utah, Salt Lake City, UT 84112 (corresponding author). ORCID: tudinal steel bars should be considered in the analysis because the
https://orcid.org/0000-0003-0257-9768. Email: ruoyang.wu@utah.edu; longitudinal steel bars of the original members experienced signifi-
rywuandrew@gmail.com cant yielding. He et al. (2016) developed numerical models for
2
Professor, Dept. of Civil and Environmental Engineering, Univ. of
original and repaired columns using shear and bond-slip springs
Utah, Salt Lake City, UT 84112. Email: c.pantelides@utah.edu
Note. This manuscript was submitted on December 20, 2017; approved
to consider strength deterioration, chronology of loading, damage,
on June 8, 2018; published online on August 15, 2018. Discussion period and repair sequence. In the present study, two types of models,
open until January 15, 2019; separate discussions must be submitted Model Fiber and Model Rotational Spring (RS) are proposed to
for individual papers. This paper is part of the Journal of Composites address the aforementioned factors for the repaired RC connec-
for Construction, © ASCE, ISSN 1090-0268. tions. The bond-slip effect was considered in both the distributed

© ASCE 04018044-1 J. Compos. Constr.

J. Compos. Constr., 2018, 22(5): 04018044


plasticity element and the concentrated rotational spring element by two of the specimens are cast-in-place reinforced concrete column-
considering bond stress-slip relationships of the longitudinal steel to-cap beam and column-to-footing connections, referred to as
bars. The main differences between the models in the present paper CB-CIP-O and F-CIP-O, respectively; the other two specimens are
and the model presented by He et al. (2016) are as follows: (1) in precast concrete column-to-cap beam connections with grouted
the present paper shear deformation of the column was considered splice sleeves (GSS) located in the plastic hinge region of the col-
by using the command of section aggregator in OpenSees 2.5.0 umn, referred to as PC1-O and PC2-O, respectively (Ameli et al.
(Mazzoni et al. 2007) for both Model Fiber and Model RS, whereas 2015, 2016). Details of the four specimens are shown in Fig. 1.
He et al. (2016) used a shear spring; and (2) in the present paper Notation CB means column-to-cap beam connection and F stands
bond-slip effects in Model Fiber were simulated using a one- for column-to-footing connection; letter O stands for original and R
dimensional bond-slip model, developed by extending the model for repaired. The repaired specimens of CB-CIP-O and F-CIP-O are
proposed by Ameli and Pantelides (2016). For Model RS bond-slip referred to as CB-CIP-R, and F-CIP-R, respectively; specimen
effects were determined using the equations developed by Wehbe PC1-O was repaired twice, and the first and second repaired spec-
et al. (1999). Damaged instead of original steel properties were imens are referred to as PC1-R1 and PC1-R2, respectively; finally,
Downloaded from ascelibrary.org by Marriott Lib-Univ Of UT on 08/15/18. Copyright ASCE. For personal use only; all rights reserved.

used in Model Fiber and Model RS to determine deformation the repaired specimen for PC2-O is referred to as PC2-R. For each
and rotation due to bond-slip; on the other hand, He et al. specimen, the column was 2.59 m (102 in.) high with a 533 mm
(2016) used the bond-slip material model Bond SP01 in OpenSees (21 in.) wide octagonal cross section. The longitudinal reinforce-
developed by Zhao and Sritharan (2007), which does not consider ment consisted of six No. 25 (#8) bars arranged in a circular pattern.
reduced steel properties. A No. 13 (#4) spiral at a 64 mm (2.5 in.) pitch was provided for
transverse column reinforcement. The cap beam was 2.74 m (9 ft)
long, 610 mm (2 ft) deep, and 610 mm (2 ft) wide; the footing was
Review of Experiments for Original and Repaired 1.83 m (6 ft) long, 610 mm (2 ft) deep, and 914 mm (3 ft) wide. The
Specimens calculated longitudinal steel ratio and volumetric spiral ratio were
1.3% and 2.0%, respectively. For specimen PC1-O, a 203 mm
(8 in.) portion of the longitudinal steel bars was debonded inside
Details of Original Specimens
the cap beam to provide a better strain distribution along the critical
Four RC specimens designed based on current seismic bridge region that was outside the steel sleeves. Details of the original
design standards (AASHTO 2011) are considered in this paper: specimens are provided elsewhere (Ameli et al. 2015, 2016),

Fig. 1. Original specimen details.

© ASCE 04018044-2 J. Compos. Constr.

J. Compos. Constr., 2018, 22(5): 04018044


Table 1. Original and repaired specimen test results
Specimen
Parameter CB-CIP-O CB-CIP-R F-CIP-O F-CIP-R PC1-O PC1-R1 PC1-R2 PC2-O PC2-R
Lateral load capacity 168 (37.8) 203 (45.6) 160 (36.0) 195 (43.8) 182 (41.0) 222 (49.9) 238 (53.4) 177 (39.7) 217 (48.8)
[kN (kip)]
Ultimate drift ratio (%) 9.3 8.1 8.8 8.4 6.7 5.6 7.8 5.5 7.6
Failure mode East and west Severe concrete East and west Severe concrete GSS bar CFRP jacket East bar GSS bar West and east
bar fracture crushing bar fracture crushing pullout crack fracture pullout bar fracture
Displacement ductility 9.9 6.8 8.9 6.0 3.1 5.1 5.6 4.9 7.1
Note: GSS = grouted splice sleeve.
Downloaded from ascelibrary.org by Marriott Lib-Univ Of UT on 08/15/18. Copyright ASCE. For personal use only; all rights reserved.

and details of repaired specimen PC1-R1 are given in Parks et al. fractured and buckled, as shown in Figs. 2(a and b). For the two
(2016). precast original specimens, PC1-O and PC2-O, the column con-
crete crushed and the longitudinal steel bars pulled out from the
Experimental Results of Original Specimens sleeves, as shown in Figs. 2(c and e) (Ameli et al. 2015, 2016).
Specimen PC1-O was repaired once previously (Parks et al. 2016)
The test results regarding the four original and one repaired speci- and was repaired in this study for the second time; after removing
men PC1-R1 are summarized in Table 1 in terms of maximum the repair concrete from repaired specimen PC1-R1, severe con-
lateral load, ultimate drift ratio, failure mode, and displacement crete crushing and extensive cracks in the original column concrete
ductility. The specimens were tested under quasi-static cyclic lat- were exposed as shown in Fig. 2(d). Damage of the precast speci-
eral loading, which was applied at the top of the column until fail- men PC2-O is shown in Fig. 2(e); damage to concrete was rela-
ure. Directions of West and East indicate the lateral loading in pull tively light, and there was no damage to the core concrete.
and push directions, respectively, as shown in Fig. 1. The maximum
lateral load is defined as the peak lateral load obtained from the
Repair Design
experimental hysteretic response. The ultimate drift ratio corre-
sponds to the drift ratio at which the lateral load dropped to 80% The repair design was intended to strengthen the severely damaged
of the maximum lateral load obtained from the backbone curve, and specimens and to restore the lateral load and displacement capacity.
the displacement ductility is defined as the ratio of the ultimate dis- Considering the previous damage and an evaluation of the theoreti-
placement to the yield displacement from the idealized backbone cal plastic hinge length, a 483 mm (19 in.) high CFRP shell with a
curve (Wu and Pantelides 2017a, b). 762 mm (30 in.) diameter consisting of seven CFRP layers in the
Damage of the specimens is shown in Fig. 2. For the two hoop direction and two CFRP layers in the longitudinal direction
CIP original specimens, CB-CIP-O and F-CIP-O, the column con- was designed and implemented for CB-CIP-R, F-CIP-R, PC1-R2,
crete crushed heavily and the two extreme longitudinal steel bars and PC2-R specimens (Wu and Pantelides 2017a, b). Fig. 3 shows

Fig. 2. Condition of specimens before repair: (a) CB-CIP-O; (b) F-CIP-O; (c) PC1-O; (d) PC1-R1 (after removal of repair concrete); and (e) PC2-O.

© ASCE 04018044-3 J. Compos. Constr.

J. Compos. Constr., 2018, 22(5): 04018044


Downloaded from ascelibrary.org by Marriott Lib-Univ Of UT on 08/15/18. Copyright ASCE. For personal use only; all rights reserved.

Fig. 3. Repair design: (a) CFRP donut; and (b) steel collar (for PC2-R).

the details of the repair; six headed steel bars were provided to 2017a, b). A lateral displacement-controlled, cyclic quasi-static
satisfy the flexural moment demand at the repaired cross section, load was applied at the top of the column using a hydraulic actuator
as shown in Fig. 3(a). Nonshrink concrete was subsequently cast with a loading protocol, which consisted of increasing amplitudes
inside the CFRP shell to form a CFRP donut. Since the column and with two cycles at each drift ratio until failure. The axial load was
cap beam were totally separated by cutting all column longitudinal applied using a hydraulic cylinder; the axial load index was set to
steel bars for specimen PC2-R, a steel collar with steel studs 6% of the column axial compression capacity.
was provided around the column to increase the bond between The experimental results of the repaired specimens are summa-
the column and repair concrete, as shown in Fig. 3(b). The design rized in Table 1. The load capacity of the repaired specimens was
of the seismic repair can be found elsewhere (Wu and Pantelides improved, while the displacement capacity and displacement duc-
2017a, b). Table 2 includes the properties of the materials used for tility were successfully restored. Moreover, the plastic hinge was
the repaired specimens. successfully relocated above the repaired cross section, as shown
in Fig. 4, since one of the objectives was to avoid further damage
inside the previously damaged region. Specimens CB-CIP-R and
Experimental Results of Repaired Specimens
F-CIP-R failed due to severe concrete crushing which extended up
The test setup and loading protocol used for the repaired specimens to 546 mm (21.5 in.) above the repaired section and 76 mm (3.0 in.)
was the same used for the original specimens (Wu and Pantelides inside the CFRP shell; no column longitudinal steel bars fractured.

Table 2. Material properties


Property Variable, unit CB-CIP-R F-CIP-R PC1-R2 PC2-R
Column concrete compressive strength fc0 , MPa (ksi) 51 (7.4) 52 (7.5) 70 (10.2) 48 (6.9)
CFRP donut concrete compressive strength 0 , MPa (ksi)
f cR 76 (11.0) 76 (11.0) 76 (11.0) 76 (11.0)
Headed steel bars fy , MPa (ksi) 427 (62) 427 (62) 427 (62) 427 (62)
f u , MPa (ksi) 593 (86) 593 (86) 593 (86) 593 (86)
Longitudinal steel bars fy , MPa (ksi) 469 (68) 469 (68) 531 (77) 531 (77)
f u , MPa (ksi) 641 (93) 641 (93) 703 (102) 703 (102)
Steel stirrups fy , MPa (ksi) 434 (63) 434 (63) 434 (63) 434 (63)
f u , MPa (ksi) 710 (103) 710 (103) 710 (103) 710 (103)
CFRP composite f fu , MPa (ksi) 780 (113) 780 (113) 780 (113) 780 (113)
Ef , GPa (ksi) 65 (9,403) 65 (9,403) 65 (9,403) 65 (9,403)

© ASCE 04018044-4 J. Compos. Constr.

J. Compos. Constr., 2018, 22(5): 04018044


Downloaded from ascelibrary.org by Marriott Lib-Univ Of UT on 08/15/18. Copyright ASCE. For personal use only; all rights reserved.

Fig. 4. Plastic hinge relocation of repaired specimens: (a) CB-CIP-R; (b) F-CIP-R; (c) PC1-R2; and (d) PC2-R.

Testing of specimen PC1-R2 was terminated at a drift ratio equal to with the CFRP-confined concrete, which reduced the effect of FRP
9.7% with fracture of the east longitudinal bar at 152 mm (6 in.) confinement on the bond-slip of steel bars compared to previous
above the donut. Severe concrete crushing occurred in the column research (Harajli 2009). In the analysis, the effects of the CFRP
up to 343 mm (14 in.) above the donut and 25 mm (1.0 in.) below composite on the bond-slip of the steel bars were considered ac-
the donut top. Specimen PC2-R failed due to longitudinal column cording to the corresponding damage of steel bars and surrounding
steel bar fracture at both the west and east side of the column at a concrete.
drift ratio of 8.7%; concrete spalling and crushing, including con- The stress distribution in the CFRP shell was not uniform,
crete damage in the column core, reached up to 610 mm (24 in.) which can be observed in the related experimental study [Fig. 3(a)
above the CFRP donut on the east side and 229 mm (9.0 in.) on the in Wu and Pantelides (2017b)]. The hoop stresses of repaired speci-
west side. men CB-CIP-R obtained from a static pushover analysis using
Abaqus 6.14 clearly demonstrate the presence of a stress gradient.
However, due to the complexity of simulating CFRP composites
Consideration of Bond-Slip Effects and Existing and the limitations present in OpenSees, this behavior was not
Damage directly implemented in this paper; instead, an equivalent CFRP-
confined concrete stress-strain curve was used in the analysis.
To simulate the experimental results of the repaired specimens, two
numerical models, Model Fiber and Model RS, were developed in
OpenSees considering the bond-slip effect, damaged steel proper- Consideration of Existing Damage
ties, and low-cycle fatigue of steel bars. Based on previous research, five levels of damage states (DS-1 to
DS-5) were used to evaluate the damage state of the original spec-
imens (Vosooghi and Saiidi 2013). The previously tested specimens
Bond-Slip Effects
CB-CIP-O and F-CIP-O had reached a damage state of DS-5 due to
Research on bond-slip between steel bars and concrete has resulted severe concrete crushing, and longitudinal bar fracture and buck-
in several bond-slip models (Eligehausen et al. 1982; Harajli 2009). ling; specimen PC1-R1 reached DS-5 due to severe concrete crush-
Zhao and Sritharan (2007) proposed a modified steel stress-slip ing and extensive circumferential concrete cracks; even though
bar model incorporating bond-slip in OpenSees, which can be damage to the original specimen PC2-O was light, the damage state
implemented through the Zero Length Section Element. Braga et al. was considered as DS-5 due to intentional cutting of all longitudi-
(2012) developed a modified steel bar model considering bond-slip nal steel bars; this was done to simulate column replacement using
relationships provided in model codes (CEB-FIP 1993). Only the the repair technique (Wu and Pantelides 2017a).
residual bond strength, instead of a full bond stress-slip curve, was Considering the measured strain in the longitudinal column steel
considered in the modified steel bar model (Braga et al. 2012). bars of damaged original specimens, modified steel properties were
After the plastic hinge length is determined, the modified steel used to account for the Bauschinger effect. The slope of the first
stress-slip curve is converted to a modified steel stress-strain curve, branch was taken as a fraction of the original steel modulus of elas-
which could be used as input in the sectional properties (Wu and ticity; a modification factor equal to 0.4 was used for specimens
Pantelides 2018a, b). CB-CIP-R and F-CIP-R, and a factor of 0.6 was used for specimens
In the repaired specimens, CFRP composites were used to form PC1-R2 and PC2-R. The steel constitutive curve considering the
the cylindrical shell to provide continuous column confinement; it existing damage is denoted as damaged steel and is used in the
is noted that the effects of the CFRP composite shell on concrete numerical models.
confinement and steel bar slipping should be appropriately incor-
porated in the numerical models. Harajli (2009) developed several
empirical equations of the bond stress-slip response of steel bars Model Fiber
embedded in FRP-confined concrete under cyclic loading, which
were also considered in this study. However, it should be noted that A schematic of Model Fiber for the repaired column is shown in
the column longitudinal steel bars were still embedded in the pre- Fig. 5. The octagonal column was approximated by a circular sec-
viously damaged concrete; these bars were not directly interacting tion of equal cross-sectional area to simplify the discretization

© ASCE 04018044-5 J. Compos. Constr.

J. Compos. Constr., 2018, 22(5): 04018044


Downloaded from ascelibrary.org by Marriott Lib-Univ Of UT on 08/15/18. Copyright ASCE. For personal use only; all rights reserved.

Fig. 5. Model Fiber.

process. There were 40 circumferential subdivisions for the core defined column shear capacity, the model detects shear failure
concrete, cover concrete, and CFRP confined concrete, and there of the column.
were 20, 5, and 10 radial subdivisions for the core concrete, cover
concrete, and CFRP confined concrete, respectively.
The Concrete04 material model was used to model steel- Bond-Slip Model for Modified Steel
confined concrete and cover concrete based on Mander’s steel- To consider the bond-slip effects in Model Fiber, a one-dimensional
confined concrete model (Mander et al. 1989). The CFRP confined bond-slip model, as shown in Fig. 6, was developed by extending
concrete model was implemented in the Concrete01 material the model proposed by Ameli and Pantelides (2016). The latter was
model, which was calculated based on the recommendations of developed for precast concrete columns connected with grouted
ACI Committee 440 [ACI 440 (ACI 2015)]. splice sleeve connections. In this bond-slip model, the longitudinal
steel bar with the damaged steel properties was discretized and
Distributed Plasticity over Plastic Hinge Length connected to the bond-slip springs, which were modeled using
the ZeroLength Element. The bond stress-slip relationships in the
In Model Fiber, distributed plasticity considering bond-slip is
bond-slip springs were obtained from the CEB-FIP Code in the
assumed to be concentrated in the plastic hinge length of the non-
case of splitting mode failure (CEB-FIP 1993, 2012). For the length
linear beam-column element instead of the total length of the
of the steel bar embedded in the previously damaged region, the
element; this was implemented using the BeamWithHinges element
bond stress-slip relationship for unconfined concrete was used
(Scott and Fenves 2006). In this element, the incorporation of a
for the bond-slip spring elements; for the remaining length of
plastic hinge length in the element integration method ensures
the steel bar outside the previously damaged region, the bond
an objective element and sectional response, which is important for
stress-slip relationship for confined concrete was used for the
strain-softening behavior in RC structures. The two-point Gauss–
bond-slip spring elements. Bond stress-slip relationships for con-
Radau integration rule was applied over a length equal to four times
fined and unconfined concrete used for specimen CB-CIP-R are
the plastic hinge length at the element ends.
shown in Fig. 7.
One of the benefits of this element is that the plastic hinge length
In the one-dimensional bond-slip model, one end of the steel
can be defined by the user. The plastic hinge length, Lpl , can be
bar was pulled to get the total deformation, including both steel bar
determined based on the damaged regions observed from tests
elongation and slip. The total deformation of the steel bar, Δs, is the
or from empirical relationships in the literature (Panagiotakos
sum of the corresponding total deformation from both the column
and Fardis 2001). The numerical model based on this element is
part and the cap beam or footing part, as shown in Eq. (1)
referred to as Model Fiber in this paper.
(D’Amato et al. 2012):
Although the columns in this study were well-confined with
enough transverse steel reinforcement, the command of section Δs ¼ ΔsA þ ΔsB ð1Þ
aggregator available in OpenSees was used to consider shear
deformation and shear failure of the column in Model Fiber where Δs = total deformation of steel bar; ΔsA = total deformation
(Mazzoni et al. 2007). Once the column shear force reaches the of steel bar in the column; and ΔsB = total deformation of steel bar

© ASCE 04018044-6 J. Compos. Constr.

J. Compos. Constr., 2018, 22(5): 04018044


100 689
Original Steel
Damaged Steel
80 Modified Steel 552

Stress (MPa)
Stress (ksi)
60 414

40 276

20 138

0 0
0.00 0.02 0.04 0.06 0.08 0.10
Strain
Downloaded from ascelibrary.org by Marriott Lib-Univ Of UT on 08/15/18. Copyright ASCE. For personal use only; all rights reserved.

Fig. 8. Steel stress-strain curves for specimens CB-CIP-R and


F-CIP-R.

defined plastic hinge region and CFRP donut cross section of


Model Fiber. The elastic modulus of modified steel was 0.33
times the original steel elastic modulus for specimens CB-CIP-R
and F-CIP-R; the elastic modulus of modified steel was 0.29
Fig. 6. Bond-slip springs model. and 0.40 times the original steel elastic modulus for specimens
PC1-R2 and PC2-R, respectively. A comparison of the different
steel bar properties for specimens CB-CIP-R and F-CIP-R is shown
in Fig. 8.
Slip (in.) The Hysteretic Material available in OpenSees was used to
0.00 0.16 0.31 0.47 0.63 0.79 0.94 model longitudinal steel bars in Model Fiber. In the defined
1.0 6.9
Unconfined concrete modified steel stress-strain curve, although the columns are well-
Confined concrete confined, buckling of the steel bars in compression was also
0.8 5.5
considered using the method proposed by Dhakal and Maekawa
Bond stress (MPa)
Bond stress (ksi)

(2002).
0.6 4.1

0.4 2.8 Low-Cycle Fatigue


In Model Fiber, fatigue material in OpenSees was used to consider
0.2 1.4
the accumulation of low-cycle fatigue damage in the longitudinal
steel bars (Uriz and Mahin 2008). Since the column longitudinal
0.0 0.0
0.0 4.0 8.0 12.0 16.0 20.0 24.0
steel bars had experienced high strains in the original specimen
Slip (mm) tests, it was necessary to consider the cumulative strain in the analy-
sis of the repaired specimens to obtain the accumulated damage
Fig. 7. Bond stress-slip curves used in bond-slip springs model for of the longitudinal steel bars. A cumulative linear damage rule
CB-CIP-R. from the Coffin–Manson expression was incorporated in the fatigue
material, as shown in Eqs. (3) and (4), respectively (Kunnath et al.
2009):
in the cap beam or footing, as indicated in Fig. 6. For specimen εi ¼ ε0 ð2N f Þ−m ð3Þ
PC2-R, only the deformation of the steel bar in the column was
considered because all the longitudinal steel bars were intentionally X
n

cut at the column-to-cap beam interface. Df ¼ 1= ð2N f Þi ð4Þ


The total deformation of the steel bar, Δs, including both elon- i¼1

gation and slip was obtained. The steel strain, ε, was calculated where εi = total strain range (εmax –εmin ) at cycle i; ε0 = value of
based on Eq. (2) strain at which one cycle will cause failure; 2N f = number of half-
cycles to failure; m = material constant related to the slope of the
ε ¼ Δs=Lpl ð2Þ Coffin–Manson curve in log-log space; and Df = accumulated
fatigue damage. In this study, a value ε0 of 0.25 was used for spec-
where ε = steel strain; and Lpl = defined plastic hinge length above imens CB-CIP-R, F-CIP-R and PC1-R2; a value of 0.191 was used
the repaired section. In Model Fiber, 356 mm (14 in.), or 67% of the for specimen PC2-R. In addition, a value of m equal to 0.36 was
column width, was used for the defined plastic hinge length of cast- used for all repaired specimens, based on recommendations from
in-place specimens CB-CIP-R and F-CIP-R; a defined plastic hinge the literature (Kunnath and Brown 2004; Uriz and Mahin 2008).
length of 305 mm (12 in.), or 57% of the column width, was used A large value of ε0 implies less accumulated fatigue damage; for
for precast concrete specimens PC1-R2 and PC2-R. The modified specimen PC2-R with a steel collar that was used to improve the
steel stress-strain curve with consideration of initial damage and bond between column and repair concrete, a smaller value of ε0 was
bond-slip is thus obtained, which was used for steel bars in the used compared to the other three repaired specimens.

© ASCE 04018044-7 J. Compos. Constr.

J. Compos. Constr., 2018, 22(5): 04018044


Downloaded from ascelibrary.org by Marriott Lib-Univ Of UT on 08/15/18. Copyright ASCE. For personal use only; all rights reserved.

Fig. 9. Model RS.

Model Rotational Spring Concentrated Plasticity in Moment-Rotation Spring

The model with a rotational spring, referred to as Model RS, is In Model RS, a moment-rotation curve is considered as the input, as
shown in Fig. 9. In Model RS, concentrated plasticity was consid- shown in Fig. 10 (Ibarra and Krawinkler 2005). The total rotation
ered using a nonlinear moment rotational spring located at the re- includes the deformations that result from elastic deformations
and plastic hinge rotation. The moment-rotation curve was defined
paired cross section. Damaged steel properties were used for the
as a trilinear curve, including an initial elastic part, a hardening part,
column cross section, and the moment-rotation spring was modeled
and a softening part, to represent structural behavior considering
using a ZeroLength Element.
material deterioration. Shear failure of the column was also consid-
The Hysteretic Material was also used to define moment-
ered for Model RS in a similar manner as described for Model Fiber.
rotation relationships in Model RS. There are five parameters in-
corporated in the material to control the cyclic behavior of the
model, including pinching of force and deformation during reload- Calibration of Moment-Rotation Curve
ing (px , py ), damage due to ductility and energy (D1 , D2 ), and un-
loading stiffness degradation based on ductility (β). Considering To calculate the nonlinear moment-rotational spring properties, a
sectional analysis was performed considering damaged steel prop-
the pinching behavior of repaired CIP specimens CB-CIP-R and
erties and bond-slip, as shown in Fig. 11. The bond stress was as-
F-CIP-R, pinching factors px and py were less than 1.0 in Model
sumed as a constant residual bond stress (D’Amato et al. 2012), as
RS; pinching factors px and py were equal to 1.0 for repaired pre-
shown in Eqs. (5) and (6); Eq. (5) is applicable for CIP repaired
cast specimens PC1-R2 and PC2-R because no pinching behavior
specimens CB-CIP-R and F-CIP-R; Eq. (6) applies to precast
was observed in these tests. The values of the five parameters used
repaired specimens PC1-R2 and PC2-R with a relatively good
in the model are listed in Table 3 and were further validated
bond condition between the longitudinal steel bars and surrounding
by comparing the calculated and experimental results as presented
concrete:
in the following sections. The remaining material and geometry pffiffiffiffiffi
information of Model RS was the same as for Model Fiber. τ ¼ 0.15 fc0 ðin MPaÞ ð5Þ

Table 3. Parameters used in hysteretic material of Model RS


Pinching factor Pinching factor Damage due to Damage due to Unloading stiffness
Specimen for force, px for deformation, py ductility, D1 energy, D2 degradation factor, β
CB-CIP-R 0.45 0.20 0.004 0.02 0.30
F-CIP-R 0.40 0.15 0.001 0.02 0.40
PC1-R2 1.00 1.00 0.001 0.02 0.40
PC2-R 1.00 1.00 0.000 0.01 0.35

© ASCE 04018044-8 J. Compos. Constr.

J. Compos. Constr., 2018, 22(5): 04018044


δs
θs ¼ ð9Þ
d−c
where θs = bond-slip rotation; d = effective cross-sectional depth;
and c = compression height of the cross section. Thus, the peak
point (M c , θc ) in Fig. 10 was obtained.
Haselton and Deierlein (2008) proposed several empirical equa-
tions to simulate the response of RC beam-columns to collapse.
These equations are used in this research to determine the remain-
ing two characteristic points, namely the yield point (M y , θy ) and
the residual point (M r , θr ) (Haselton et al. 2016). The ratio between
M c and My is defined in Eq. (10), and thus M y can be determined:
0
M c =M y ¼ 1.25ð0.89Þv ð0.91Þ0.01kfc ð10Þ
Downloaded from ascelibrary.org by Marriott Lib-Univ Of UT on 08/15/18. Copyright ASCE. For personal use only; all rights reserved.

where Mc = maximum moment; M y = yield moment; v = axial load


Fig. 10. Backbone curve for hysteretic rotational spring. ratio, P=ðAg fc0 Þ; P = axial load; fc0 = concrete compressive
strength; Ag = column cross-sectional area; and k = unit conversion
ratio (1.0 when fc0 is in MPa and 6.9 when fc0 is in ksi).
Plastic rotation capacity for a structure subjected to cyclic load-
ing between the yield and peak moment can be determined from
Eqs. (11) and (12) (Haselton et al. 2016):

θcap;pl ¼ 0.12ð1 þ 0.55αsl Þ0.16v ð0.02 þ 40ρsh Þ0.43


0
× ð0.54Þ0.01kfc ð0.66Þ0.1sn ð2.27Þ10ρ ð11Þ

θcap;pl;cyclic ¼ 0.7θcap;pl ð12Þ

where θcap;pl = monotonic plastic rotation capacity; θcap;pl;cyclic =


cyclic plastic rotation capacity; αsl = bond-slip coefficient
(1.0 when bond-slip is considered and 0.0 when bond-slip can
be ignored); ρsh = transverse reinforcement ratio, Ash =ðsbÞ; Ash =
cross-sectional area of transverse reinforcement; s = transverse
reinforcement spacing; b = width of column cross section; sn =
Fig. 11. Bond-slip rotation at column base. longitudinal steel bar slenderness ratio of tie spacing to longitudinal
bar diameter, s=db ; ρ = longitudinal steel reinforcement ratio,
As =ðbdÞ; and As = area of longitudinal steel reinforcement. In this
pffiffiffiffiffi study, bond-slip was considered (αsl ¼ 1) in Eq. (11) for specimens
τ ¼ 0.30 fc0 ðin MPaÞ ð6Þ
CB-CIP-R, F-CIP-R, and PC1-R2; bond-slip was ignored (αsl ¼ 0)
where τ = bond stress (MPa); and fc0 = concrete compressive for specimen PC2-R due to the presence of the steel collar, which
strength (MPa). The coefficients in Eqs. (5) and (6) are consider- improved the bond between original column and repair concrete.
ably smaller than the ones commonly used for undamaged concrete The yield rotation, θy , can be obtained by subtracting the plastic
members (Ozcan et al. 2008). rotation from the peak rotation, as shown in Fig. 10 and Eq. (13):
According to the longitudinal steel strain and stress, the bond-
θy ¼ θc − θcap;pl;cyclic ð13Þ
slip development length and bond-slip deformation at the end of the
column were determined from Eqs. (7) and (8), respectively The post-capping rotation capacity, θpc , can be determined from
(Wehbe et al. 1999): Eqs. (14) and (15). Owing to the lack of reliable data for elements
8 subjected to large deformations, the post-capping rotation is limited
>
> fd
> s b
< εs < ε y to 0.10 rad (Haselton et al. 2016):

Ld ¼ ð7Þ
>
> fy db ðf s − f y Þdb θpc ¼ 0.76ð0.031Þv ð0.02 þ 40ρsh Þ1.02 ≤ 0.10 ð14Þ
>
: þ ε s > ε y
4τ 4τ
θpc;cyclic ¼ 0.5θpc ð15Þ
8
> f s εs d b
>
< ε s < εy where θpc = monotonic post-capping rotation capacity; and
8τ θpc;cyclic = cyclic post-capping rotation capacity.
δs ¼ ð8Þ
>
> f εy db ðf s − fy Þðεs þ εy Þdb The residual rotation, θr , is obtained by adding the post-capping
: y
þ ε s > εy
8τ 8τ rotation to the peak rotation, as shown in Fig. 10 and Eq. (16):
where Ld = bond-slip development length; δ s = bond-slip deforma- θr ¼ θc þ θpc;cyclic ð16Þ
tion; f s = steel stress; εs = steel strain; fy = steel yield stress;
εy = steel yield strain; and db = diameter of longitudinal steel bar. To determine the residual moment, M r , the ratio of residual
Once the bond-slip deformation at the end of the column was moment to peak moment is assumed as 0.75 in this study based
determined, the bond-slip rotation at the column base was obtained, on the hysteretic response results from the tests, as shown in
as shown in Eq. (9): Fig. 10 and Eq. (17):

© ASCE 04018044-9 J. Compos. Constr.

J. Compos. Constr., 2018, 22(5): 04018044


M r ¼ 0.75 M c ð17Þ column and repair system, the selected pinching factors of Table 3,
for Model RS, were appropriate to capture the pinching behavior of
The procedure for determining the moment-rotation relationship the hysteresis, as shown in Figs. 12(b and d).
is briefly summarized as follows: (1) determine the peak point For repaired precast concrete specimens PC1-R2 and PC2-R
(Mc , θc ) from sectional analysis and Eqs. (6)–(9); (2) calculate with the column core concrete in good condition and no significant
the yield point (My , θy ) based on Eqs. (10)–(15); and (3) determine bond-slip effects or debonding between the column and repair sys-
the residual point (M r , θr ) based on Eqs. (16) and (17). tem, Model RS with a pinching factor equal to 1.0 was also able to
match the hysteresis curves observed in the experiments, as shown
in Figs. 13(b and d).
Analysis Results and Comparison to Experiments

Two types of models were used in the analysis of the four repaired Cumulative Hysteretic Energy
specimens. The numerical results from Model Fiber and Model Cumulative hysteretic energy versus drift ratio from numerical
Downloaded from ascelibrary.org by Marriott Lib-Univ Of UT on 08/15/18. Copyright ASCE. For personal use only; all rights reserved.

RS compared to the experiments are shown in Figs. 12–15, in terms models was superimposed with the experimental results, as shown
of hysteretic response, cumulative hysteretic energy, and moment- in Fig. 14. The numerical results from both Model Fiber and Model
rotation relationships, and they are discussed in the following RS are in good agreement with the corresponding experimental
sections. results. The error in cumulative hysteretic energy between models
and experiments at each drift ratio was less than 5%.
Hysteretic Response
Moment-Rotation Response
Hysteretic response results from the numerical analysis were com-
pared to experimental results for the repaired specimens, as shown One of the differences between the two models was that the
in Figs. 12 and 13. moment-rotation relationship was considered as input for Model
Both Model Fiber and Model RS predicted the backbone curve RS; moment-rotation results for Model Fiber were obtained as out-
very well, capturing important characteristics including peak lateral put. The moment-rotation relationships at the repaired cross section
force and hysteretic energy, as shown in Figs. 12–15. Model from Model Fiber and Model RS were compared to the experimen-
Fiber was able to model pinching behavior and matched the exper- tal results measured using linear variable differential transformers
imental results for both repaired CIP and repaired precast concrete (LVDTs) for the repaired specimens, as shown in Fig. 15.
specimens. The LVDTs were removed at the 6% drift ratio in the tests,
For repaired CIP specimens, CB-CIP-R and F-CIP-R, with which explains the missing data shortly after the peak moment of
pinching and severe bond-slip due to debonding between the the experimental moment-rotation response. The moment-rotation

60 267 60 267
CB-CIP-R (test) CB-CIP-R (test)
Model Fiber Model RS
40 178 40 178
Lateral force (kip)

Lateral force (kN)


Lateral force (kip)

Lateral force (kN)

20 89 20 89

0 0 0 0

-20 -89 -20 -89

-40 -178 -40 -178

-60 -267 -60 -267


-10 -8 -6 -4 -2 0 2 4 6 8 10 -10 -8 -6 -4 -2 0 2 4 6 8 10
(a) Drift ratio (%) (b) Drift ratio (%)

60 267 60 267
F-CIP-R (test) F-CIP-R (test)
40 Model Fiber 178 40 Model RS 178
Lateral force ( kN)

Lateral force (kN)


Lateral force (kip)

Lateral force (kip)

20 89 20 89

0 0 0 0

-20 -89 -20 -89

-40 -178 -40 -178

-60 -267 -60 -267


-10 -8 -6 -4 -2 0 2 4 6 8 10 -10 -8 -6 -4 -2 0 2 4 6 8 10
(c) Drift ratio (%) (d) Drift ratio (%)

Fig. 12. Hysteretic response of repaired CIP specimens: (a) Model Fiber and test for CB-CIP-R; (b) Model RS and test for CB-CIP-R; (c) Model
Fiber and test for F-CIP-R; and (d) Model RS and test for F-CIP-R.

© ASCE 04018044-10 J. Compos. Constr.

J. Compos. Constr., 2018, 22(5): 04018044


60 267 60 267
PC1-R2 (test) PC1-R2 (test)
Model Fiber Model RS
40 178 40 178

Lateral force (kip)

Lateral force (kN)

Lateral force (kip)

Lateral force (kN)


20 89 20 89

0 0 0 0

-20 -89 -20 -89

-40 Bar fracture -178 -40 -178

-60 -267 -60 -267


-10 -8 -6 -4 -2 0 2 4 6 8 10 -10 -8 -6 -4 -2 0 2 4 6 8 10
(a) Drift ratio (%) (b) Drift ratio (%)
Downloaded from ascelibrary.org by Marriott Lib-Univ Of UT on 08/15/18. Copyright ASCE. For personal use only; all rights reserved.

60 267 60 267
PC2-R (test) PC2-R (test)
Model Fiber Model RS
40 178 40 178
Lateral force (kip)

Lateral force (kip)


Lateral force (kN)

Lateral force (kN)


20 89 20 89

0 0 0 0

-20 -89 -20 -89

Bar fracture
-40 -178 -40 -178

-60 -267 -60 -267


-10 -8 -6 -4 -2 0 2 4 6 8 10 -10 -8 -6 -4 -2 0 2 4 6 8 10
(c) Drift ratio (%) (d) Drift ratio (%)

Fig. 13. Hysteretic response of repaired PC specimens: (a) Model Fiber and test for PC1-R2; (b) Model RS and test for PC1-R2; (c) Model Fiber and
test for PC2-R; and (d) Model RS and test for PC2-R.

4000 452 4000 452


Cumulative hysteretic energy (kip-in.)
Cumulative hysteretic energy (kip-in.)

CB-CIP-R (test) F-CIP-R (test)

Cumulative hysteretic energy (kN-m)


Cumulative hysteretic energy (kN-m)

3500 Model Fiber 395 3500 Model Fiber 395


Model RS Model RS
3000 339 3000 339

2500 282 2500 282

2000 226 2000 226

1500 169 1500 169

1000 113 1000 113

500 56 500 56

0 0 0 0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
(a) Drift ratio (%) (b) Drift ratio (%)

4000 452 4000 452


Cumulative hysteretic energy (kip-in.)
Cumulative hysteretic energy (kip-in.)

PC1-R2 (test) PC2-R (test)


Cumulative hysteretic energy (kN-m)

Cumulative hysteretic energy (kN-m)

3500 Model Fiber 395 3500 Model Fiber 395


Model RS Model RS
3000 339 3000 339

2500 282 2500 282

2000 226 2000 226

1500 169 1500 169

1000 113 1000 113

500 56 500 56

0 0 0 0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
(c) Drift ratio (%) (d) Drift ratio (%)

Fig. 14. Cumulative hysteretic energy comparisons: (a) CB-CIP-R; (b) F-CIP-R; (c) PC1-R2; and (d) PC2-R.

© ASCE 04018044-11 J. Compos. Constr.

J. Compos. Constr., 2018, 22(5): 04018044


5000 565 5000 565

Moment @ repaired section (kip-in.)

Moment @ repaired section (kip-in.)


Moment @ repaired section (kN-m)

Moment @ repaired section (kN-m)


4000 452 4000 452

3000 339 3000 339

2000 226 2000 226

1000 CB-CIP-R (test) 113 1000 F-CIP-R (test) 113


Model Fiber Output Model Fiber Output
Model RS Input Model RS Input
0 0 0 0
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.00 0.02 0.04 0.06 0.08 0.10 0.12
(a) Rotation (rad) (b) Rotation (rad)
Downloaded from ascelibrary.org by Marriott Lib-Univ Of UT on 08/15/18. Copyright ASCE. For personal use only; all rights reserved.

5000 565 5000 565


Moment @ repaired section (kip-in.)

Moment @ repaired section (kip-in.)


Moment @ repaired section (kN-m)

Moment @ repaired section (kN-m)


4000 452 4000 452

3000 339 3000 339

2000 226 2000 226

1000 PC1-R2 (test) 113 1000 PC2-R (test) 113


Model Fiber Output Model Fiber Output
Model RS Input Model RS Input
0 0 0 0
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.00 0.02 0.04 0.06 0.08 0.10 0.12
(c) Rotation (rad) (d) Rotation (rad)

Fig. 15. Moment-rotation response comparisons: (a) CB-CIP-R; (b) F-CIP-R; (c) PC1-R2; (d) PC2-R.

1.2 1.2
Accumulated damage of steel bar
Accumulated damage of steel bar

fracture of the steel bar fracture of the steel bar


1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4

0.2 West bar 0.2 West bar


East bar East bar
0.0 0.0
0 2 4 6 8 10 0 2 4 6 8 10
(a) Drift ratio (%) (b) Drift ratio (%)

1.2 1.2
Accumulated damage of steel bar
Accumulated damage of steel bar

fracture of the steel bar fracture of the steel bar


1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4

0.2 West bar 0.2 West bar


East bar East bar
0.0 0.0
0 2 4 6 8 10 0 2 4 6 8 10
(c) Drift ratio (%) (d) Drift ratio (%)

Fig. 16. Low-cycle fatigue of two extreme column longitudinal steel bars from Model Fiber: (a) CB-CIP-R; (b) F-CIP-R; (c) PC1-R2; and (d) PC2-R.

© ASCE 04018044-12 J. Compos. Constr.

J. Compos. Constr., 2018, 22(5): 04018044


results obtained from Model Fiber not only matched the existing effective. The proposed calibration procedure produced accurate
experimental results, but also predicted the tendency after the peak moment-rotation relationships.
moment was reached. In addition, the trilinear moment-rotation 5. Model Fiber performed well for matching the pinching behavior
curve calculated following the procedure given in the previous sec- of the structure; in addition to matching the hysteresis curves,
tion was consistent with the experimental results and also matched Model Fiber produced local responses such as moment-rotation
the full moment-rotation curve obtained from Model Fiber. relationships and could be used for prediction purposes.
6. In Model RS, appropriate pinching factors need to be considered
for matching the hysteresis curves of cast-in-place concrete
Accumulated Low-Cycle Fatigue Damage structures with relatively poor bond conditions between longi-
Low-cycle fatigue of column longitudinal steel bars was incorpo- tudinal steel bars and concrete for columns with a severely da-
rated in Model Fiber. The fatigue damage of the two extreme lon- maged column core concrete, as well as poor bond conditions
gitudinal steel bars in the loading direction was obtained from between column concrete and CFRP donut concrete.
Model Fiber for the repaired specimens, as shown in Fig. 16. Ac- 7. For future modeling, it is recommended that both models should
Downloaded from ascelibrary.org by Marriott Lib-Univ Of UT on 08/15/18. Copyright ASCE. For personal use only; all rights reserved.

cording to the definition of accumulated fatigue in the fatigue be used to determine the range of possible structural responses.
material, a steel bar fractures when the fatigue damage reaches 1.0. A lower bound estimate of the load and displacement capacity
For specimens CB-CIP-R and F-CIP-R, the maximum fatigue could thus be obtained.
damage of two extreme steel bars was 0.6 and 0.7, respectively,
which is less than 1.0; this is consistent with the experimental
results for which there was no fracture of longitudinal steel bars. Acknowledgments
For specimen PC1-R2, the accumulated fatigue damage of the west
and east extreme bars reached 0.87 and 1.0, respectively, as shown The authors would like to acknowledge the financial support of
in Fig. 16(c). The east steel bar fractured according to the prediction the Mountain Plains Consortium under contract MPC-491. In ad-
of Model Fiber, which was observed in the test of specimen dition, they would like to acknowledge the support of Dr. Luis
PC1-R2 and marked in Fig. 13(a). Ibarra, MJ Ameli, Mark Bryant, Priyank Sankholkar, Trevor Nye,
For specimen PC2-R, both the west and east extreme bars Anurag Upadhyay, and Joel Parks of the Department of Civil and
reached an accumulated fatigue damage of 1.0 at the last drift ratio, Environmental Engineering at the University of Utah.
as shown in Fig. 16(d). This shows that both bars fractured, which
is in good agreement with the experimental failure mode of speci-
men PC2-R, as marked in Fig. 13(c). The marks of bar fracture Notation
shown in Figs. 13(a and c) correspond to experimental observations
The following symbols are used in this paper:
and were accompanied by a sudden drop of lateral load.
Ag = column cross-sectional area;
As = area of longitudinal steel reinforcement;
Conclusions Ash = cross-sectional area of transverse reinforcement;
b = width of column cross section;
A seismic repair method was implemented for four severely c = compression width of the cross section;
damaged bridge column-to-cap beam/footing connections, which Df = accumulated fatigue damage;
suffered bar pullout, buckling and fracture of reinforcing steel bars, D1 = damage due to ductility;
and concrete crushing inside the column core concrete. Based on D2 = damage due to energy;
the results from numerical simulations using the models developed d = effective cross-sectional depth;
in this research the following conclusions can be made: db = diameter of longitudinal steel bar;
1. The two models developed, the first considering distributed
Ef = elastic modulus of CFRP composite;
plasticity spread over a defined plastic hinge length (Model
f c0 = concrete compressive strength;
Fiber) and the second using a concentrated rotational spring 0 = repair concrete compressive strength;
(Model RS), reproduced hysteresis curves and energy dissipa- f cR
tion that matched the experimental results in a satisfactory man- f fu = ultimate tensile strength of CFRP composite;
ner. Both Model Fiber and Model RS considered bond-slip f s = steel stress;
effects, the effects of previous loading history, and cyclic degra- f u = steel ultimate stress;
dation of column longitudinal steel bars. f y = steel yield stress;
2. The modified steel stress-strain curve, which was derived from k = unit conversion ratio;
the improved one-dimensional bond-slip model, was successful Ld = bond-slip development length;
in Model Fiber for simulating the experimental response. The Lpl = defined plastic hinge length above the repaired
defined plastic hinge length ranged from 57% of the column section;
width for precast concrete specimens to 67% for cast-in-place M c = maximum moment;
specimens. M r = residual moment;
3. Model Fiber was able to consider low-cycle fatigue effects and
M y = yield moment;
demonstrated that the accumulated fatigue damage of the long-
m = material constant related to the slope of the
itudinal steel bars was consistent with the experimental results.
Coffin–Manson curve in log-log space;
Moreover, the model was able to obtain the accumulated fatigue
of the steel bars and to predict steel bar fracture for the same 2N f = number of half-cycles to failure;
drift ratio cycle as the experiment. P = axial load;
4. Model RS, using plasticity concentrated in a nonlinear spring px = pinching factor for force;
element with a simplified moment-rotation relationship to repre- py = pinching factor for deformation;
sent the structural behavior of the repaired specimens, was also s = transverse reinforcement spacing;

© ASCE 04018044-13 J. Compos. Constr.

J. Compos. Constr., 2018, 22(5): 04018044


sn = longitudinal steel bar slenderness ratio of tie In Proc., 7th European Conf. on Earthquake Engineering, 69–80.
spacing to longitudinal bar diameter; Istanbul, Turkey: European Association for Earthquake Engineering.
v = axial load ratio; Gergely, I., C. P. Pantelides, R. J. Nuismer, and L. D. Reaveley. 1998.
“Bridge pier retrofit using fiber-reinforced plastic composites.” J. Com-
αsl = bond-slip coefficient;
pos. Constr. 2 (4): 165–174. https://doi.org/10.1061/(ASCE)1090-0268
β = unloading stiffness degradation factor; (1998)2:4(165).
Δs = total deformation of the steel bar; Harajli, M. H. 2009. “Bond stress-slip model for steel bars in unconfined or
ΔsA = total deformation of the steel bar in column; steel, FRC, or FRP confined concrete under cyclic loading.” J. Struct.
ΔsB = total deformation of the steel bar in the cap beam or Eng. 135 (5): 509–518. https://doi.org/10.1061/(ASCE)0733-9445
footing; (2009)135:5(509).
δ s = bond-slip deformation; Haselton, C. B., and G. G. Deierlein. 2008. Assessing seismic collapse
safety of modern reinforced concrete moment-frame buildings. Rep.
ε = steel strain;
No. 2007/08. Berkeley, CA: Pacific Earthquake Engineering Research
εi = total strain range at cycle i; (PEER) Center.
ε0 = value of strain at which one cycle will cause failure;
Downloaded from ascelibrary.org by Marriott Lib-Univ Of UT on 08/15/18. Copyright ASCE. For personal use only; all rights reserved.

Haselton, C. B., A. B. Liel, S. C. Taylor-Lange, and G. G. Deierlein. 2016.


εs = steel strain; “Calibration of model to simulate response of reinforced concrete
εy = steel yield strain; beam-columns to collapse.” ACI Struct. J. 113 (6): 1141–1152.
θc = rotation at capping point; https://doi.org/10.14359/51689245.
θcap;pl = monotonic plastic rotation capacity; He, R., L. H. Sneed, and A. Belarbi. 2013. “Rapid repair of severely
damaged RC columns with different damage conditions: An experimen-
θcap;pl;cyclic = cyclic plastic rotation capacity;
tal study.” Int. J. Concr. Struct. Mater. 7 (1): 35–50. https://doi.org/10
θpc = monotonic post-capping rotation capacity; .1007/s40069-013-0030-7.
θpc;cyclic = cyclic post-capping rotation capacity; He, R., Y. Yang, and L. H. Sneed. 2015. “Seismic repair of reinforced
θr = residual rotation; concrete bridge columns: Review of research findings.” J. Bridge
θs = bond-slip rotation; Eng. 20 (12): 04015015. https://doi.org/10.1061/(ASCE)BE.1943
θy = yield rotation; -5592.0000760.
ρ = longitudinal steel reinforcement ratio; He, R., Y. Yang, and L. H. Sneed. 2016. “Post-repair seismic assessment
of RC bridges damaged with fractured column bars—A numerical ap-
ρsh = transverse reinforcement ratio; and proach.” Eng. Struct. 112: 100–113. https://doi.org/10.1016/j.engstruct
τ = bond stress. .2016.01.007.
Ibarra, L. F., and H. Krawinkler. 2005. Global collapse of frame structures
under seismic excitations. Rep. No. 2005/06. Berkeley, CA: Pacific
References Earthquake Engineering Research Center.
AASHTO. 2011. AASHTO guide specifications for LRFD seismic bridge Kunnath, S. K., and J. Brown. 2004. “Low-cycle fatigue failure of reinforc-
design. Washington, DC: AASHTO. ing steel bars.” ACI Mater. J. 101 (6): 457–466.
ACI (American Concrete Institute). 2015. Guide for the design and Kunnath, S. K., Y. Heo, and J. F. Mohle. 2009. “Nonlinear uniaxial material
construction of concrete reinforced with FRP bars. ACI 440.1R-15. model for reinforcing steel bars.” J. Struct. Eng. 135 (4): 335–343.
Farmington Hills, MI: ACI. https://doi.org/10.1061/(ASCE)0733-9445(2009)135:4(335).
Ameli, M. J., D. N. Brown, J. E. Parks, and C. P. Pantelides. 2016. “Seismic Mander, J. B., M. J. N. Priestley, and R. Park. 1989. “Observed stress-strain
column-to-footing connections using grouted splice sleeves.” ACI behavior of confined concrete.” J. Struct. Eng. 114 (8): 1827–1849.
Struct. J. 113 (5): 1021–1030. https://doi.org/10.1061/(ASCE)0733-9445(1988)114:8(1827).
Ameli, M. J., and C. P. Pantelides. 2016. “Seismic analysis of precast con- Mazzoni, S., F. McKenna, M. H. Scott, and G. L. Fenves. 2007. OpenSees
crete bridge columns connected with grouted splice sleeve connectors.” command language manual. Berkeley, CA: Pacific Earthquake
J. Struct. Eng. 143 (2): 1–13. https://doi.org/10.1061/(ASCE)ST.1943 Engineering Research Center.
-541X.0001678. Ozcan, O., B. Binici, and G. Ozcebe. 2008. “Improving seismic perfor-
Ameli, M. J., J. E. Parks, D. N. Brown, and C. P. Pantelides. 2015. “Seismic mance of deficient reinforced concrete columns using carbon fiber-
evaluation of grouted splice sleeve connections for reinforced precast reinforced polymers.” Eng. Struct. 30 (6): 1632–1646. https://doi.org
concrete column-to-cap beam joints in accelerated bridge construction.” /10.1016/j.engstruct.2007.10.013.
PCI J. 60 (2): 80–103. https://doi.org/10.15554/pcij.03012015.80.103. Panagiotakos, T. B., and M. N. Fardis. 2001. “Deformation of reinforced
Braga, F., R. Gigliotti, M. Laterza, M. D. Amato, and S. Kunnath. 2012. concrete members at yielding and ultimate.” ACI Struct. J. 98 (2):
“Modified steel bar model incorporating bond-slip for seismic assess- 135–148.
ment of concrete structures.” J. Struct. Eng. 138 (11): 1342–1350. Pantelides, C. P., and D. A. Moran. 2013. “Design of FRP jackets for plastic
https://doi.org/10.1061/(ASCE)ST.1943-541X.0000587. hinge confinement of RC columns.” J. Compos. Constr. 17 (4):
CEB-FIP (Comite Euro-International du Beton–Fédération Internationale 433–442. https://doi.org/10.1061/(ASCE)CC.1943-5614.0000332.
de la Précontrainte). 1993. CEB-FIP model code 1990. London:
Parks, J. E., D. N. Brown, M. J. Ameli, and C. P. Pantelides. 2016. “Seismic
Thomas Telford.
repair of severely damaged precast reinforced concrete bridge columns
CEB-FIP (Comite Euro-International du Beton–Fédération Internationale
connected with grouted splice sleeves.” ACI Struct. J. 113 (3): 615–626.
de la Précontrainte). 2012. CEB-FIP model code 2010. London:
Thomas Telford. https://doi.org/10.14359/51688756.
D’Amato, M., F. Braga, R. Gigliotti, S. Kunnath, and M. Laterza. 2012. Priestley, M. J. N., F. Seible, Y. Xiao, and R. Verma. 1994. “Steel jacket
“Validation of a modified steel bar model incorporating bond-slip retrofitting of reinforced concrete bridge columns for enhanced shear
for seismic assessment of concrete structures.” J. Struct. Eng. strength. Part 1: Theoretical considerations and test design.” ACI Struct.
138 (11): 1351–1360. https://doi.org/10.1061/(ASCE)ST.1943-541X J. 91 (4): 394–405.
.0000588. Rutledge, S. T., M. J. Kowalsky, R. Seracino, and J. M. Nau. 2014. “Repair
Dhakal, R. P., and K. Maekawa. 2002. “Modeling for postyield buckling of of reinforced concrete bridge columns containing buckled and fractured
reinforcement.” J. Struct. Eng. 128 (9): 1139–1147. https://doi.org/10 reinforcement by plastic hinge relocation.” J. Bridge Eng. 19 (8):
.1061/(ASCE)0733-9445(2002)128:9(1139). A4013001. https://doi.org/10.1061/(ASCE)BE.1943-5592.0000492.
Eligehausen, R., E. P. Popov, and V. V. Bertero. 1982. “Local bond stress- Saadatmanesh, H., M. R. Ehsani, and L. Jin. 1997. “Repair of earthquake-
slip relationships of deformed bars under generalized excitations.” damaged RC columns with FRP wraps.” ACI Struct. J. 94 (2): 206–214.

© ASCE 04018044-14 J. Compos. Constr.

J. Compos. Constr., 2018, 22(5): 04018044


Scott, M. H., and G. L. Fenves. 2006. “Plastic hinge integration methods for Wu, R.-Y., and C. P. Pantelides. 2017c. “Rapid seismic repair of severely
force-based beam-column elements.” J. Struct. Eng. 132 (2): 244–252. damaged cast-in-place reinforced concrete bridge piers.” In Proc.,
https://doi.org/10.1061/(ASCE)0733-9445(2006)132:2(244). Transportation Research Board 96th Annual Meeting. Washington,
Seible, B. F., M. J. N. Priestley, G. A. Hegemier, and D. Innamorato. 1997. DC: Transportation Research Board.
“Seismic retrofit of RC columns with continuous carbon fiber jackets.” Wu, R.-Y., and C. P. Pantelides. 2017d. “Rapid seismic repair of severely
J. Compos. Constr. 1 (2): 52–62. https://doi.org/10.1061/(ASCE)1090 damaged reinforced concrete bridge piers.” In Proc., Structures
-0268(1997)1:2(52). Congress 2017, 370–381. Reston, VA: ASCE.
Uriz, P., and S. A. Mahin. 2008. Toward earthquake-resistant design of Wu, R.-Y., and C. P. Pantelides. 2018a. “Analytical models for seismic
concentrically braced steel-frame structures. Berkeley, CA: Pacific repair of bridge columns using plastic hinge relocation.” In Proc., Struc-
Earthquake Engineering Research Center. tures Congress 2018, 296–307. Reston, VA: ASCE.
Vosooghi, A., and M. S. Saiidi. 2013. “Design guidelines for rapid repair
Wu, R.-Y., and C. P. Pantelides. 2018b. “Seismic analysis of repaired
of earthquake-damaged circular RC bridge columns using CFRP.”
RC bridge column assemblies using advanced composites.” In Proc.,
J. Bridge Eng. 18 (9): 827–836. https://doi.org/10.1061/(ASCE)BE
11th US National Conf. on Earthquake Engineering (11NCEE).
.1943-5592.0000426.
Wehbe, N. I., M. S. Saiidi, and D. H. Sanders. 1999. “Seismic performance Los Angeles: EERI.
Downloaded from ascelibrary.org by Marriott Lib-Univ Of UT on 08/15/18. Copyright ASCE. For personal use only; all rights reserved.

of rectangular bridge columns with moderate confinement.” ACI Struct. Xiao, Y., and R. Ma. 1997. “Seismic retrofit of RC circular columns using
J. 96 (2): 248–258. prefabricated composite jacketing.” J. Struct. Eng. 123 (10): 1357–1364.
Wu, R.-Y., and C. P. Pantelides. 2017a. “Rapid repair and replacement of https://doi.org/10.1061/(ASCE)0733-9445(1997)123:10(1357).
earthquake-damaged concrete columns using plastic hinge relocation.” Yan, Z., C. P. Pantelides, and L. D. Reaveley. 2007. “Posttensioned FRP
Compos. Struct. 180: 467–483. https://doi.org/10.1016/j.compstruct composite shells for concrete confinement.” J. Compos. Constr. 11 (1):
.2017.08.051. 81–90. https://doi.org/10.1061/(ASCE)1090-0268(2007)11:1(81).
Wu, R.-Y., and C. P. Pantelides. 2017b. “Rapid seismic repair of reinforced Zhao, J., and S. Sritharan. 2007. “Modeling of strain penetration effects
concrete bridge columns.” ACI Struct. J. 114 (5): 1339–1350. https://doi in fiber-based analysis of reinforced concrete structures.” ACI Struct.
.org/10.14359/51700789. J. 104 (2): 133–141.

© ASCE 04018044-15 J. Compos. Constr.

View publication stats J. Compos. Constr., 2018, 22(5): 04018044

You might also like