behavior of high-performance fiber-reinforced cement ...earthquake.hanyang.ac.kr/journal/2016/2016,...

Construction and Building Materials 106 (2016) 89–101

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier .com/locate /conbui ldmat

Behavior of high-performance fiber-reinforced cement compositecolumns subjected to horizontal biaxial and axial loads

http://dx.doi.org/10.1016/j.conbuildmat.2015.12.0870950-0618/� 2015 Elsevier Ltd. All rights reserved.

⇑ Corresponding author.E-mail address: [email protected] (K. Lee).

Khai Mai Quang a, V.B. Phuoc Dang a, Sang Whan Han b, Myoungsu Shin c, Kihak Lee a,⇑aArchitectural Engineering Department, Sejong University, Seoul, Republic of KoreabArchitectural Engineering Department, Hanyang University, Seoul, Republic of KoreacUrban and Environmental Engineering School, Ulsan National Institute of Science and Technology, Ulsan, Republic of Korea

h i g h l i g h t s

� The strength and ductility of HPFRCC specimens were remarkable improved.� The capacity of the FC and SF columns adding PVA showed very little difference.� The ductility and energy dissipation are inversely proportional to axial load.

a r t i c l e i n f o

Article history:Received 20 June 2015Received in revised form 13 November 2015Accepted 14 December 2015

Keywords:PVA fibersHPFRCCBiaxial testingHysteresis loopEnergy dissipation

a b s t r a c t

Current design in the ACI building code for reinforced concrete columns under seismic load combinationsrequires reinforced detailing which causes reinforcement congestion and construction difficulties. As analternative solution, the use of high-performance fiber-reinforced cement composites (HPFRCC) with aneconomical type of Poly Vinyl Alcohol (PVA) fibers in column elements was investigated in this paper. Sixcolumn specimens with 2/3 scale including three standard reinforced columns and three columns usingPVA fibers were tested. In the research work herein presented, biaxial cyclic lateral loads were applied tospecimens subjected to either 10% or 30% constant axial loads. The experimental results are presentedand the global behaviors of the tested columns are discussed, particularly focusing on stiffness andstrength degradations due to increasing cyclic demand. Test results indicated superior damage toleranceand stable inelastic load–displacement responses up to 5% or 9% drift for the HPFRCC columns, eventhough they suffered severe shear cracks. Specimens without PVA all showed very limited ductilityand low strengths.

� 2015 Elsevier Ltd. All rights reserved.

1. Introduction

One of the most fundamental observations in research projectson past earthquakes is that biaxial bending moment damagecaused to reinforced concrete (RC) elements by earthquake loadingin two directions is much greater than that caused by earthquakeloading in one direction. This is because the application of biaxialbending-moment cyclic demands to RC columns tends to reducetheir capacity and stiffness and strength degradation occurs duringsuccessive load reversals. These factors indicate the importance ofinvestigating the inelastic response of structural elements whensubjected to biaxial or bidirectional cyclic loading. While previousstudies have been mostly focused on the structural performance of

members under constant axial loading, very few investigations areavailable on their structural behavior under multi-dimensionalearthquake conditions.

Among those very limited researches, Qiu et al. [8] tested sevenspecimens of RC column subjected to biaxial loading with differentload paths and concluded that the interactions of biaxial deforma-tion, under biaxial loads, were found to weaken biaxial strengthand hysteretic energy dissipation capacity. Tsuno and Park [9] per-formed cyclic bi-directional tests on two RC columns with rectan-gular cross-section and Bechtoula et al. [10] tested eight large-scale and eight small-scale RC columns under various verticaland horizontal loading patterns. From the test observations, bi-directional loading had a significant influence on the envelopecurves as well as on damage progress. It is therefore importantto implement additional studies on structural behavior undermulti-dimensional loading, and apply the results to improve theseismic capacity of columns.

http://crossmark.crossref.org/dialog/?doi=10.1016/j.conbuildmat.2015.12.087&domain=pdf

http://dx.doi.org/10.1016/j.conbuildmat.2015.12.087

mailto:[email protected]

http://dx.doi.org/10.1016/j.conbuildmat.2015.12.087

http://www.sciencedirect.com/science/journal/09500618

http://www.elsevier.com/locate/conbuildmat

Table 2.3Properties of fibers included in PVA.

Tensile strength(MPa)

Elasticmodulus (GPa)

Diameter(mm)

Length(mm)

Volumefraction (%)

1600 25 0.039 12 2.0

90 K.M. Quang et al. / Construction and Building Materials 106 (2016) 89–101

Researchers have recently looked at the applications of high-performance fiber-reinforced cement composites (HPFRCC), suchas coupling beams Afshin Canbolat et al. [11], low-rise walls Kimand Parra-montesinos [12], the cyclic behavior of precast post-tensioned segmental concrete columns with ECC Billington andYoon [13], effectiveness of low-cost fiber-reinforced cement com-posite in hollow columns under cyclic loading Shin et al. [14].Bengi Arisoy and Hwai-Chung Wu [15] investigated mixing PVAfibers into reinforced lightweight concrete. From these experimen-tal investigations it was found that the use of HPFRCC has a favor-able effect on the resistance of the column member. For instance,HPFRCC exhibited increased strength, displacement capacity anddamage tolerance in members subjected to larger deformations[1–6,16,17]. In short, HPFRCC may be effective when used in RCelements with the main aim of improving the seismic behaviorof structural members.

In view of the above, tests were conducted on six 2/3 scale col-umn specimens, including three columns using fibers and threestandard RC column specimens, subjected to biaxial loads. It isworthwhile to note that this study considered that a column con-sisted of an upper and lower part divided at the point of inflectionand specimens in this paper represent for upper part of columns.The main purpose of the experimental investigations presentedin this paper to estimate the increase in strength, ductility, energydissipation, cracking and failure mode of column specimens withadded PVA (2% of volume), as well as to observe deformation ofspecimens under biaxial loading.

2. Experimental program

2.1. Material properties

Table 2.1 summarizes the test results of the compressive strength of concretespecimens, and the tensile strength of reinforced specimens. The compressive testsfollowed ASTM C39, ‘‘Standard Test Method for Compressive Strength of CylindricalConcrete Specimens”. The compressive strength of the concrete was determined asthe average of at least three cylinder specimens with sizes of 200 mm � 100 mm(height � diameter). The top and bottom of the cylinder specimens were properlyground and capped with neoprene pad caps (ASTM C1231) to ensure a uniform loaddistribution.

Tensile tests were conducted in order to determine the material properties ofthe reinforced specimens. Based on the test results, the tensile strength of the steelwas determined as the average of three specimens. The Young’s modulus of elastic-ity of reinforcement was calculated based on a stress–strain curve in the elastic lim-

Table 2.1Test results of reinforcing bars and concrete (all units are MPa).

Longitudinal bar (D19) Stirrup (D10)

Tensilestrength

Young’smodulus

Tensilestrength

Young’smodulus

Specified 400.0 200,000 400.0 200,000Measured 484.7 190,870 528.3 183,697Difference (%) 21.2 �4.6 32.1 �8.2

Table 2.2HPFRCC mix proportions for the detailed material specimens.

Material No.01 No.02 No.0

Binder (wt.%) 72.3 67.4 67.1Filler (wt.%) Silica sand 25.0 30.0 10.0

CaCO3 – – 20.0CW150 – – –

CA (wt.%) 2.7 2.6 2.9PVA fiber (Vol. %) 2.0 2.0 2.0W/PCM (wt.%) 20.0 20.0 20.0

1. Binder: cement, fly ash, and powdered slag, silica fume, consisting expandable.2. PCM: Premixed Cement Mortar (dry mortar), Binder + Filler + CA.

itations. The compressive strength of concrete at 28 days is 27.3 MPa for normalconcrete and 48.3 MPa for HPFRCC specimens and the tensile strength (fy) of rein-forcement for D19 and D10 are 484.7 MPa and 528.3 MPa respectively.

Table 2.2 shows the HPFRCC mix proportions for the specimen’s detailed mate-rials, and the physical properties of the PVA fibers used are shown in Table 2.3. Atotal of six HPFRCC mix proportions were examined, in which the volumetric ratioof PVA fibers were approximately 2.0%, and the water/PCM ratio was kept roughly20–22%. The direct tensile test used at least two dog-bone shaped specimens foreach of HPFRCC mixture type, as shown in Fig. 2.1. Two LVDTs were mounted alongthe sides of the specimen in the loading direction, in which the gage length wasequal to 76 mm (3 in.). One or two layers of steel wire mesh were used to reinforceeach end of the specimen to avoid failure outside the LVDT gage length. The directtensile tests were displacement-controlled, with an actuator travel velocity ofroughly 0.5 mm/min based on the JSCE recommendations [18].

The stress–strain responses of the six types of HPFRCC specimens are compared,in which two or three similar results were acquired, and one of them is selected forthe comparison. In general, the specimen No.6 presented the highest ductility,developing numerous well-distributed micro-cracks (Fig. 2.1); the maximum ten-sile strain exceeded 5%, and the tensile strength was approximately 7 MPa. There-fore, specimen No.6 mixture was used for the column specimens.

2.2. Specimen description

All of the column specimens in this experiment have all details in common butone with the only difference being in stirrup spacing. Transverse reinforcement isgenerally considered to serve three main functions, of confining the concrete core,restraining buckling of longitudinal bars and avoiding shear failure. Hence, differentstirrup spacing in the column specimens can lead to serious effects correspondingto the three above mentioned behaviors. More specifically, the wider the stirrupspacing is, the less confinement it provides to the core concrete, and the main rein-forcements are also poorly supported to prevent buckling as well, resulting in non-ductile behavior and the sudden brittle failure of the columns. That feature wasexploited to intentionally control the failure modes of column specimens:flexure-controlled (FC) and shear–flexure controlled (SF).

In this study, six 2/3-scale column specimens with sections of300 mm � 300 mm and 900 mm in height were tested to investigate the behaviorof concrete columns under seismic load. Three columns, which were flexure-control (FC) specimens, were tested to assess flexural behavior and the threeremaining specimens called shear–flexure (SF) specimens were tested to considerthe shear and flexure behavior simultaneously. Though there is different betweenfull and scaled models in the maximum shear strength, but the crack pattern andhysteresis loops were quite similar. Also, the flexural and shear deformation

Concrete

Compressive strength (normalconcrete)

Compressive strength(HPFRCC)

27.3 48.3

3 No.04 No.05 No.06 Note

64.2 57.8 56.213.0 – –20.0 40.0 41.0– 0.5 1.02.8 1.7 1.8 Chemical admixture2.0 2.1 2.120.5 22.0 22.0

8540

8040

85

6030

30

60

330

Fig. 2.1. Dog-bone specimen configuration for direct tensile tests of No. 6 binding (units: mm).

K.M. Quang et al. / Construction and Building Materials 106 (2016) 89–101 91

components seemed to be consistently scaled and hence no scale effect on the col-umn deformation was found. Therefore, the overall behaviors of the small-scalemodels were quite similar to the full-scale unit during the test [20].

Fig. 2.2 illustrates the dimensions and reinforcement details of the columnspecimens. Each column was arranged with eight longitudinal reinforcementsD19 (diameter of 19.1 mm), so that the longitudinal reinforcement ratio (q)exceeded a minimum longitudinal reinforcement ratio as required by ACI 318-11[7] (2.55% and 1% respectively). Transverse reinforcement D10 (diameter of9.53 mm) with 200 mm spacing was placed throughout the column in the shear–flexure specimens and the first tie reinforcement was placed 100 mm from the slabor footing surface according to the requirement of the first tie location specified inSection 7.10.5 in ACI 318-11. The tie spacing of the flexure specimens, however, washalf of the tie spacing in the shear–flexure columns with 100 mm.

Table 2.4 summarizes the characteristics of the column specimens. Three spec-imens using PVA fibers were named with ‘‘HP” at the end, while three other spec-imens were named with ‘‘RC” at the end to indicate that no PVA fibers were used.The notation FC was used to classify flexure-control specimens and the SF notationwas assigned for shear–flexure specimens. The numbers 10 and 30 represent thepercent of axial load, and notation B indicates bi-axial load. Three column speci-mens, FC-10B-RC, SF-10B-RC and SF-10B-HP, were tested with a constant axial loadof 243 kN (10% f0cAg; Ag is the cross area of the column section) and three remainingspecimens, FC-30B-RC, FC-30B-HP, SF-30B-HP, were tested with a constant axialload of 729 kN (30% f0cAg).

(a) Shear -flexure column (SF)

Biaxial lateral load

Axial load

950

600

900

220

300100

120

200

170

100

6575

7565

60

D19

D10D10

D10 @200mm

8-D19

D19

Ø70mm PVC

Ø28mm PVCM24Bolt

100

100

ELEVATION

1

A

Temperature

D19

D10

D10 @

D19

Ø70mm P

Ø28m

D10Temperature

Biaxial lateral loa

Fig. 2.2. Details and cross-sect

2.3. Test set-up

Fig. 2.3 shows the setup adopted for the experimental testing of the RC columns.The system includes two independent horizontal actuators of 500 kN to apply thelateral loads to the column specimens and a vertical actuator of 1000 kN to applythe axial load. A steel reaction frame and a concrete reaction wall were used tofix three actuators. The column specimens were cast in square shape concrete foun-dation blocks which were fixed to the floor of the laboratory to avoid sliding andoverturning of the specimens during the test. Because the axial load actuatorremains in the same position during the test while the column specimen deflectslaterally, a sliding device was placed between the top of the column and theactuator.

In order to measure the longitudinal and transverse bar strain, four layers ofstrain gauges were installed into the column specimens, and each layer includedthree strain gauges. Also, twelve LVDTs were installed to measure vertical displace-ment and three LVDTs were installed to measure horizontal displacement. The loca-tions of the strain gauges and LVDTs for the specimens are shown in Fig. 2.4.

To characterize the response of the column specimens, cyclic lateral displace-ments were imposed at the top of the column with steadily increasing force.Fig. 2.5 illustrates the biaxial loading applied to the columns. Three cycles wereapplied for each of the loading cases from 0.25% to 1% of drift ratio, while onlytwo cycles were applied for each loading case higher than 1.5% of drift ratio. Repeat-ing the cycles for each displacement level demand allowed the capturing of

(b) Flexure -control column (FC)

Axial load

950

600

900

220

300100

120

200

170

100

100mm

8-D19

VC

m PVCM24Bolt

300

300

8-D19D10 @200mm

26.6

7

300

300

100

1001

0010

0

100 100 Bolt M24

SECTION 1

ELEVATION

VIEW A

300

300

8-D19 D10 @100mm

26.6

7

SECTION 2

2

A

d 6575

7565

60

ion of the test specimens.

Table 2.4Summary of test specimens.

Specimen Failure PVA volume (%) Column height L (mm) Longitudinal reinforcement Stirrup (mm) Axial load (%Agf0c)

FC-10B-RC Flexure 0 900 8-D19 D10@100 10% (243 kN)FC-30B-RC Flexure 0 900 8-D19 D10@100 30% (729 kN)SF-10B-RC Shear–flexure 0 900 8-D19 D10@200 10% (243 kN)FC-30B-HP Flexure 2 900 8-D19 D10@100 30% (729 kN)SF-10B-HP Shear–flexure 2 900 8-D19 D10@200 10% (243 kN)SF-30B-HP Shear–flexure 2 900 8-D19 D10@200 30% (729 kN)

(a) Test setup in the laboratory (b) Plan view

Actuator for perpendicular

direction

Actuator for main direction

Fig. 2.3. Test setup for biaxial loading.


information to better understand the stiffness and strength degradation of the col-umn. Rhombus path loading was applied with a drift ratio of 100% in the maindirection and 30% in the perpendicular direction in order to produce the most crit-ical load effect based on ASCE 7–10 (see Fig. 2.5).

3. Analysis of column test results

Based on visual observations and recorded data, the crack pat-tern and failure mode, load–displacement response, displacementductility, energy dissipation, strength evaluation, and stiffnessdegradation of each column specimen can be discussed. The effectsof the test variables on the performance measures are highlightedbelow.

(a) Strain gauges attached position

Axial load

950

600

900

220

300

D10@100mm

ELEVATION

1

100

950

600

900

220

300

rod

ELEVATIO

2LVDT

750.

25D

150

0.5D

225

0.75

D

D5 Thread

Biaxial lateral load

Strain Gauges

200

200

200

Fig. 2.4. Arrangement of str

3.1. Hysteretic comparison, cracking and failure mode

Fig. 3.1 plots the cyclic hysteretic load–displacement of sixspecimens. The hysteretic curve loops of specimens with PVA com-posites and specimens without PVA are similar at drifts less than1%. However, the hysteretic curve loops of specimens with PVAcomposites are higher than specimens without PVA. The differenceoccurs because of the loading paths adopted in the experiment: thefirst cycle of each displacement level in the biaxial tests alwaysoccurs in the same direction. This induces a different response inthe first cycle of each displacement amplitude. It means that thebehavior of the column specimens with PVA composites and

(b) LVTD attached positionN

300

300

3 Strain Gauges

SECTION 1

LVDT

300

300

SECTION 2

9012

090

L>600

D5 Thread

String Potentiometer

Reference columnper layer

rod

ain gauges and LVDTs.

-60 -40 -20 0 20 40 60-15

-10

-5

0

5

10

15

Dis

plac

emen

t (m

m)

Perp

endi

cula

r di

rect

ion

Displacement (mm)Main direction

Fig. 2.5. Loading history for biaxial loading.


specimens without PVA is similar under low demand levels. But athigh demand levels, the specimens with PVA have better strength.Thus, the appearance and development of cracks in specimens

-90 -60 -30 0 30 60 90-250

-200

-150

-100

-50

0

50

100

150

200

250

Loa

d (k

N)

Displacement (mm)

Maximum strength

0.75 maximum strength

Δy(+)

Δy(-)

Δmax(+)

Δmax(-)

Vmax

: 147.3kNΔmax: 18.42mm

Vmax: -128.1kNΔ

max: -18mm

(a) Specimen FC-30B-RC

-90 -60 -30 0 30 60 90-250

-200

-150

-100

-50

0

50

100

150

200

250

Δy(-)

Δy(+)

Loa

d (k

N)

Displacement (mm)

SF-10B-RC

Maximum strength


Δmax(+)

Δmax(-)

Vmax: 154.5kNΔ

max: 26.55mm

Vmax: -138.0kNΔ

max: -24.72mm

(c) SpecimenSF-10B-RC

-90 -60 -30 0 30 60 90-250

-200

-150

-100

-50

0

50

100

150

200

250


Loa

d (k

N)

Displacement (mm)

FC-30B-HPMaximum strength Δy(+)

Δy(-)

Δmax(+)

Δmax(-)

Vmax: 212.6kNΔmax: 33.96mm

Vmax: -183.0kNΔmax: -26.89mm

(e) Specimen FC-30B-HP

Fig. 3.1. Hysteretic load–displa

without PVA were earlier and faster than specimens with PVAcomposites.

As shown in Fig. 3.1(b), (c), (e), and (f), the strength of FC and SFcolumn was relatively the same for both case of RC and added PVAfibers, irrespective of the tie spacing. This may be because of thesame section and longitudinal reinforcement ratio for both casesof FC and SF specimens, which mostly affect to the lateral strengthof column. However, the strength of specimen FC-30B-HP wasremarkably improved as compared with the FC-30B-RC column,because of the added PVA fibers. To be more specific, with 2% vol-umetric ratio of PVA fibers, the compressive strength of specimenFC-30B-HP is one and a half times as much as that of specimenFC-30B-RC (48.3 MPa and 27.3 MPa respectively). And the shearcapacity of FC-30B-HP (212.6 kN) is approximately one and a halftimes higher than the shear capacity of specimen FC-30B-RC(147.3 kN). Similarly, the specimen SF-10B-HP (using 2% of PVAfibers) had strength superior to the specimen SF-10B-RC. As

-90 -60 -30 0 30 60 90

250

-250

-200

-150

-100

-50

0

50

100

150

200



Δmax(-)

Δmax(+)

Loa

d (k

N)

Displacement (mm)

FC-10B-RC

Maximum strength0.75 maximum strength

Δy(+)

Δy(-)

(b) Specimen FC-10B-RC

-90 -60 -30 0 30 60 90-250

-200

-150

-100

-50

0

50

100

150

200

250

Loa

d (k

N)

Displacement (mm)

SF-10B-HPMaximum strength


Δy(+)

Δy(-)

Δmax(+)

Δmax(-)



(d) Specimen SF-10B-HP

-90 -60 -30 0 30 60 90-250

-200

-150

-100

-50

0

50

100

150

200

250

Loa

d (k

N)

Displacement (mm)

SF-30B-HPMaximum strength


Δy(+)

Δy(-)

Δmax(+)

Δmax(-)



(f) Specimen SF-30B-HP

cements of the specimens.

0.25% drift 1% drift 3% drift at failure 5% (a) FC-10B-RC

0.25% drift 1% drift 2% drift at failure 3%(b) FC-30B-RC

0.25% drift 1% drift 3% drift at failure 4% (c) SF-10B-RC

Fig. 3.2. Cracking patterns of specimens.


indicated in Fig. 3.1(c) and (d), the hysteretic curve of specimen SF-10B-HP is fuller and more stable than the specimen SF-10B-RC.

After reaching the maximum load (at the last cycle before a crit-ical strength drop, as well as at the completion of each test), thecracking patterns were observed in the six specimens and areshown in Fig. 3.2. For all six specimens, inclined cracks started inthe region near the middle of the column. Fig. 3.2(a) and (b) showsthe cracking patterns of specimens FC-10B-RC and FC-30B-RC. In adrift less than 2%, specimen FC-30B-RC had fewer cracks than spec-imen FC-10B-RC. However, the cracks on specimen FC-30B-RCdeveloped faster than specimen FC-10B-RC for drifts higher than2%. This should explain why, when the horizontal displacementof the column is small, the effect of the axial load was not signifi-cant. Nevertheless, when the horizontal displacement of the col-umn was increased significantly, it led to eccentricity betweenthe axial load and the center line of the column. As a result, spec-

imen FC-30B-RC was subjected to higher flexural moment com-pared to FC-10B-RC, because it suffered a higher axial load levelthan specimen FC-10B-RC. Hence, the specimen FC-30B-RCreached a critical stage faster than specimen FC-10B-RC and failedat a drift of 3%, compared with the 5% drift of the FC-10B-RCspecimen.

The cracking patterns of specimen FC-10B-RC and SF-10B-RCwere quite similar at drifts less than 2% (see Fig. 3.2(a) and (c)).However, the crack pattern and crack width of specimen FC-10B-RC are totally different from the SF-10B-RC specimen at the 3% driftcycle. A relatively small number of severe inclined cracks wereobserved in the former, while several hairline cracks were widelyopened in the SF-10B-RC specimen. This is due to the differencein tie spacing between the two specimens. By the end of testing,the shear–flexure specimen (SF-10B-RC) suffered severe concretecracking around the middle of the column web at 4% drift. In

0.25% drift 1% drift 3% drift at failure 5%(d) FC-30B-HP

0.25% drift 1% drift 5% drift at failure 9%(e) SF-10B-HP

0.25% drift 1% drift 3% drift at failure 5%

(f) SF-30B-HP

Fig. 3.2. (continued)

Table 3.1Displacement ductility factor.

Specimen Vmax (kN) 0.75Vmax (kN) Maximum displacement Dmax (mm) Yield displacement Dy (mm) Displacement ductility factor l = Du/Dy

FC-10B-RC + 154.0 115.5 34.56 6.4 5.4� 143.0 107.3 35.91 6.2 5.8

SF-10B-RC + 154.5 115.9 26.55 6.7 3.9� 138.0 103.5 24.72 6.2 4.0

FC-30B-RC + 147.3 110.5 18.42 4.4 4.2� 128.1 96.1 18 4.0 4.5

SF-10B-HP + 185.1 138.8 48.6 8.8 5.5� 188.2 141.2 49.58 8.8 5.7

FC-30B-HP + 212.6 159.5 33.96 7.8 4.3� 183.0 137.3 26.89 6.6 4.1

SF-30B-HP + 206.5 154.9 31.98 6.9 4.7� 185.9 139.4 25.48 6.1 4.2


-10 -8 -6 -4 -2 0 2 4 6 8

-200

-150

-100

-50

0

50

100

150

200

App

lied

Load

(kN

)

Driff (%)

FC-10B-RC FC-30B-RC FC-30B-HP

(a) Three specimens FC in comparison

10

-200

-150

-100

-50

0

50

100

150

200

App

lied

Load

(kN

)

Driff (%)

SF-10B-RC SF-10B-HP SF-30B-HP

(b) Three specimens SF in comparison

-10 -8 -6 -4 -2 0 2 4 6 8 10

Fig. 3.3. Load-drift envelope curves of the specimens in comparison.

0.0000 0.0025 0.0050 0.0075 0.0100 0.0125 0.0150 0.0000 0.0025 0.0050 0.0075 0.0100 0.0125 0.01500

75

150

225

300

375

450

525

Col

umn'

s hei

ght (

mm

)

FC-10B-RC Longitudinal bar strain

0.75% 1% 1.5% 2% 3% 4% 5% 6%yield line

0

75

150

225

300

375

450

525

Col

umn'

s hei

ght (

mm

)

SF-10B-RC Longitudinal bar strain

0.75% 1% 1.5% 2% 3% 4%yield line

(a) FC-10B-RC (b) SF-10B-RC

0.0000 0.0025 0.0050 0.0075 0.0100 0.0125 0.01500.0000 0.0025 0.0050 0.0075 0.0100 0.0125 0.01500

75

150

225

300

375

450

525

Col

umn'

s hei

ght (

mm

)

FC-30B-RC Longitudinal bar strain

0.75% 1% 1.5% 2% 3%yield line

0

75

150

225

300

375

450

525

Col

umn'

s hei

ght (

mm

)

SF-10B-HP Longitudinal bar strain

0.75% 1% 1.5% 2% 3% 4% 5% 6%yield line

(c) FC-30B-RC (d) SF-10B-HP

0.0000 0.0025 0.0050 0.0075 0.0100 0.0125 0.0150 0.0000 0.0025 0.0050 0.0075 0.0100 0.0125 0.01500

75

150

225

300

375

450

525

Col

umn'

s hei

ght (

mm

)

SF-30B-HP Longitudinal bar strain

0.75% 1% 1.5% 2% 3% 4% 5%yield line

0

75

150

225

300

375

450

525

Col

umn'

s hei

ght (

mm

)

FC-30B-HP Longitudinal bar strain

0.75% 1% 1.5% 2% 3% 4% 5%yield line

(e) SF-30B-HP (f) FC-30B-HP

Fig. 3.4. Longitudinal bar strain distribution.


0.0000 0.0005 0.0010 0.0015 0.0020 0.0025 0.0030 0.0000 0.0005 0.0010 0.0015 0.0020 0.0025 0.0030100

200

300

400

500

600

700

Col

umn'

s hei

ght (

mm

)

FC-10B-RC transverse bar strain

0.75% 1% 1.5% 2% 3% 4% 5% 6%yield line

100

200

300

400

500

600

700

Col

umn'

s hei

ght (

mm

)

FC-30B-RC transverse bar strain

0.75% 1% 1.5% 2% 3%yield line

(a) FC-10B-RC (b) FC-30B-RC

0.0000 0.0005 0.0010 0.0015 0.0020 0.0025 0.00300.0000 0.0005 0.0010 0.0015 0.0020 0.0025 0.0030

0.0000 0.0005 0.0010 0.0015 0.0020 0.0025 0.0030 0.0000 0.0005 0.0010 0.0015 0.0020 0.0025 0.0030

100

200

300

400

500

600

700

Col

umn'

s hei

ght (

mm

)

SF-10B-RC transverse bar strain

0.75% 1% 1.5% 2% 3% 4%yield line

100

200

300

400

500

600

700

Col

umn'

s hei

ght (

mm

)

SF-10B-HP transverse bar strain

0.75% 1% 1.5% 2% 3% 4% 5% 6%yield line

(c) SF-10B-RC (d) SF-10B-HP

100

200

300

400

500

600

700

Col

umn'

s hei

ght (

mm

)

FC-30B-HP transverse bar strain

0.75% 1% 1.5% 2% 3% 4% 5%yield line

100

200

300

400

500

600

700

Col

umn'

s hei

ght (

mm

)

SF-30B-HP transverse bar strain

0.75% 1% 1.5% 2% 3% 4% 5%yield line

(e) FC-30B-HP (f) SF-30B-HP

Fig. 3.5. Transverse bar strain distribution.

0 2 4 6 80

4

8

12

16

20

24

Ene

rgy

diss

ipat

ion

per

cycl

e (k

N.m

)

Drift ratio (%)

FC-10B-RC SF-10B-RC FC-30B-RC SF-10B-HP FC-30B-HP SF-30B-HP

(a) Energy dissipationDrift ratio (%)

0 2 4 6 80

20

40

60

80

100

120

140

160

180

Ene

rgy

diss

ipat

ion

per

cycl

e (k

N.m

)

FC-10B-RC SF-10B-RC FC-30B-RC SF-10B-HP FC-30B-HP SF-30B-HP

(b) Cumulated amount of dissipated energy

Fig. 3.6. Comparison of energy dissipation of all specimens.


Table 3.2Energy dissipation calculation (kN-mm).

Drift % Cycle FC-10B-RC SF-10B-RC FC-30B-RC SF-10B-HP FC-30B-HP SF-30B-HP

0.25 1 57 38 98 58 79 732 42 27 85 49 66 563 39 24 83 51 61 53

0.50 1 157 129 299 173 213 1852 106 77 283 142 189 1483 93 73 223 133 184 136

0.75 1 273 224 482 316 399 3372 214 151 425 262 345 2823 186 138 403 244 336 258

1.00 1 387 341 839 466 623 5192 327 253 664 413 553 4613 311 234 632 390 531 444

1.50 1 992 859 1694 1018 1375 12172 706 624 1339 849 1134 975

2.00 1 1758 1570 2620 1830 2408 22272 1328 1150 2256 1431 1990 1763

3.00 1 4220 4052 6175 4866 6123 63112 3486 2778 3874 5309 5105

4.00 1 6485 3426 7760 10293 102182 5982 6920 9362 9487

5.00 1 9074 10,938 14,654 14,8832 6550 10,245 13,932 14,653

6.00 1 14,4902 13,912

7.00 1 18,2182 17,632

8.00 1 21,9562 21,320

Total energy dissipation 42,772 16,168 18,601 159,956 70,160 69,791


contrast, at 5% drift, specimen FC-10B-RC experienced only minorspalling damage.

Based on the test results, it was concluded that the three PVAspecimens showed no apparent failure by the end of the test whilethe normal concrete specimens displayed serious spalling damage.In other words, the ductility and load-carrying capacity of speci-mens was remarkably improved after adding PVA fibers.

3.2. Ductility and displacement comparison

In the dynamic analysis of structures subjected to seismic load-ing, ductility factors, which are defined as the maximum deforma-tion divided by the corresponding deformation when yieldingoccurs, were used to express the various response parametersrelated to displacements, rotations, and curvatures. The displace-ment ductility factor is determined by the following formula:

l ¼ Dmax

Dyð1Þ

where Dmax is the maximum displacement and Dy is the displace-ment at yield.

Park [19] proposed a procedure to determine the value of max-imum displacement and yield displacement based on testingresults for a column subjected to a cyclic biaxial load. In this study,the maximum displacement is defined as the point that the lateralstrength of specimen was reduced by 20% of maximum strength.The adopted methodology can be described in four steps. The firststep was an evaluation of the maximum strength of the specimenin both directions. In the next step, the cycle, which had strengthlower than three-quarters of the evaluated maximum strength ofthe specimen, was identified. From this cycle, calculating thesecant stiffness, and then the intersection points of secant stiffnessand maximum strength for each direction were determined, givingthe yield displacement in each direction. The reference yield dis-placement is the average of the yield displacements obtained inthe two directions.

The displacement ductility factors of each specimen were calcu-lated and are shown in Table 3.1. In general, without PVA fibers,the FC specimens had a superior ductility capacity compared tothe SF specimens. The maximum drift ratio of FC-10B-RC is 5%,which is much higher than that of SF-10B-RC specimen with only4 % of drift ratio. While there are no clear differences in ductilitycapacity between FC and SF specimens added PVA fibers. To bemore specific, the maximum drift ratio of FC-30B-HP is 4.5%, whichis the same with SF-30B-HP. In addition, the ductility capacity ofspecimen FC-30B-RC and FC-10B-RC was judged directly fromthe load and deformation relationship. Fig. 3.1(a) and (b) show thatthe ductility of specimen FC-10B-RC was greater than that of spec-imen FC-30B-RC because of the lower axial load. The same ten-dency was seen in the PVA specimens with 9% for SF-10B-HP and5% for SF-30B-HP.

Similarly, specimen SF-10B-HP (with 2% PVA fibers) had a supe-rior ductility capacity compared to specimen SF-10B-RC. The max-imum displacement of specimen SF-10B-HP is 48.6 mm for thepositive direction, which is greater than the 26.55 mm of specimenSF-10B-RC. The strengths of column specimens FC-30B-HP and SF-30B-HP were relatively similar irrespective of the tie spacing. How-ever, because specimen FC-30B-HP had a tie reinforcement spacing(@100 mm) which was smaller than that of specimen SF-30B-HP(@200 mm), the lateral load capacity of specimen FC-30B-HP wasbetter than specimen SF-30B-HP. Hence, column specimen FC-30B-HP exhibited a better deformation capacity than column spec-imen SF-30B-HP.

To evaluate the effect of the biaxial loading load paths on thestiffness and strength degradation of the RC column, the maximumenvelope of the applied load-drift hysteresis curves were analyzed.The envelopes of the hysteretic curves for the columns test areplotted in Fig. 3.3. The graph shows that there is a significant dropof initial stiffness during the early cycles in three specimens with-out PVA fibers. This is because the opening of large diagonal cracksin these specimens caused a sudden stiffness drop. In contrast, thethree specimens with added PVA fibers display a relatively gradual

2 4 6 8 10 12 14 16 18 20 22 24 2 4 6 8 10 12 14 16 18 20 22 240

5

10

15

20

25

Cum

ulat

ive

diss

ipat

ed e

nerg

y ( k

Nm

)

Dis

sipa

ted

ener

gy fo

r ea

ch c

ycle

( kN

m)

Cyclic number of displacement applitude

Dissipated energyCumulative energy

FC-10B-RC

0

30

60

90

120

150

5%

3%

4%

2%1.5%1%

Drift ratio (%)

0

5

10

15

20

25

Dis

sipa

ted

ener

gy fo

r ea

ch c

ycle

( kN

m)



0

30

60

90

120

150

Drift ratio (%)

FC-30B-RC

Cum

ulat

ive

disi

pate

d en

ergy

(kN

m)

3%

2%1.5%

1%

(a) Specimen FC-10B-RC (b) Specimen FC-30B-RC

2 4 6 8 10 12 14 16 18 20 22 242 4 6 8 10 12 14 16 18 20 22 240

5

10

15

20

25

Dis

sipa

ted

ener

gy fo

r ea

ch c

ycle

( kN

m)



0

30

60

90

120

150

Cum

ulat

ive

diss

ipat

ed e

nerg

y (k

Nm

)

4%3%

2%1.5%1%

Drift ratio (%)

SF-10B-RC

0

5

10

15

20

25

Dis

sipa

ted

ener

gy fo

r ea

ch c

ycle

( kN

m)



0

30

60

90

120

150

Cum

ulat

ive

diss

ipat

ed e

nerg

y ( k

Nm

)

5%

4%

3%

2%1.5%

1%Drift ratio (%)

FC-30B-HP

(c) Specimen SF-10B-RC (d) Specimen FC-30B-HP

2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 2 4 6 8 10 12 14 16 18 20 22 24 26 28 300

5

10

15

20

25

Dis

sipa

ted

ener

gy fo

r ea

ch c

ycle

( kN

m)



0

30

60

90

120

150

Cum

ulat

ive

diss

ipat

ed e

nerg

y (k

Nm

)

8%

7%SF-10B-HP

6%

5%

4%

3%

2%1%Drift ratio (%)

0

5

10

15

20

25

Dis

sipa

ted

ener

gy fo

r ea

ch c

ycle

( kN

m)



0

30

60

90

120

150

4%

3%

Cum

ulat

ive

diss

ipat

ed e

nerg

y ( k

Nm

)

5%

2%1.5%

1%Drift ratio (%)

SF-30B-HP

(e) Specimen SF-10B-HP (f) Specimen SF-30B-HP

Fig. 3.7. Energy dissipation of six specimens.


stiffness decrease. Thus, the presence of PVA fibers in the columnsnot only delayed stiffness drop up to 4.0% drift, but also improvedthe stiffness retention at low drift values. In addition, comparingthe normalized stiffness values at 2.0% drift for the six specimens,it could be seen that the normal concrete specimens had an initialstiffness significantly smaller than the specimens with PVAfibers.

Figs. 3.4 and 3.5 plot the longitudinal and transverse reinforce-ments strain distribution of all specimens. It is noted that the yieldstrains of the longitudinal and transverse reinforcement are 0.0024and 0.0026 respectively, and are expressed by a solid verticalstraight line in the graphs. It can be seen that after the specimensyielded at 0.75% drift ratio, the longitudinal reinforcements yieldedtoo. The biaxial effect can only be clearly seen in flexure-controlledspecimens. After yielding of the longitudinal reinforcement, the FC-10B-RC and SF-10B-RC specimen’s reinforcement strain fluctuated

in a wider range compared to the SF-10B-RC specimen, but the FC-30B-HP and SF-30B-HP specimen’s reinforcement strain fluctuatedmore than the FC-30B-RC specimen. No discernible differencecould be found in the longitudinal reinforcement strain of the FCspecimens and SF specimens without HPFRCC but a very clear dif-ference can be seen between the three reinforced concrete speci-mens and the three specimens with PVA fibers. Based on thegraph, it was concluded that a higher axial load significantlyincreased the longitudinal reinforcement strains.

For transverse reinforcement strain distribution, all of the spec-imens have transverse reinforcement, which still behaved in anelastic range at failure. A biaxial effect was only found to increasetransverse reinforcement strains in flexure-controlled specimenssubjected to 30% axial load. The effect of higher axial loads on spec-imen increased the transverse bar strain of specimen, only in com-bination with the biaxial effect.


3.3. Energy dissipation

Energy dissipation capacity is considered to be fundamentalperformance criteria of RC elements subjected to seismic demands.For assessment or design, nonlinear static methods use energy dis-sipation capacity related parameters to evaluate the inelasticearthquake responses of structures and to describe the strengthand stiffness degradation of RC elements subjected to cyclic load-ing. The amount of energy dissipated during a loading cycle is con-sidered to be one of the crucial seismic performance factors, whichis taken to be the area enclosed by the corresponding load–dis-placement hysteretic curve.

Fig. 3.6 compares the energy dissipated during the first cycle toeach drift ratio and the cumulative amount of dissipated energy upto the indicated drift cycle (counting all cycles to each drift ratio)for all six specimens. According to this Figure and Table 3.2, allspecimens dissipated a similar amount of energy until 1% driftratio. Beyond that drift, every specimen dissipated a differentamount of energy.

The energy dissipation of the six specimens is shown in Fig. 3.7.The amount of energy dissipation of two specimens, FC-30B-HPand SF-30B-HP, are quite similar (70,160 kN mm and69,791 kN mm respectively) and both of them failed at 5% driftcycle, although there was a difference in their tie spacing. This indi-cates that PVA fibers affect the lateral load capacity of the columngreatly. Specimen SF-10B-HP has the largest amount of energy dis-sipation (159,956 kN mm), while the SF-10B-RC specimen has thesmallest amount of energy dissipation (16,168 kN mm). Therefore,the energy dissipation capacity of specimens is remarkablyincreased by applying PVA fibers. This may explain why the con-crete members that were connected strongly by using PVA fibershave high ductility and bonding strength. As mentioned in theforegoing Section 3.1, when the lateral load was increased, thespecimens subjected to higher axial load should reach a criticalstage faster than the specimens under lower axial load.

The amount of energy dissipation capacity of specimen (FC-10B-RC) was decreased after axial load was increased (FC-30B-RC). Similar results can be seen for the two specimens SF-10B-HPand SF-30B-HP. Moreover, the energy dissipation capacity of spec-imen SF-10B-RC is smaller than specimen FC-10B-RC because ofthe difference in tie spacing.

4. Conclusion

This study was carried out to assess the seismic performance ofnormal reinforced concrete columns, and columns with PVA com-posites, subjected to biaxial cyclic lateral loads. The main purposeof this research was to experimentally investigate the PVA fibers’contribution to the overall response of the column, particularlyfocusing on stiffness and strength degradations due to increasingcyclic demand, by making a comparison with reference RC speci-mens. In order to reach these goals, six columns elements wereconstructed and subjected to the biaxial loading for testing, includ-ing three standard RC and three PVA specimens. The following con-clusions can be drawn from the results of this investigation:

(1) The strength and ductility of specimens with HPFRCC wereremarkably improved in addition to the increase of failuredrift. The HPFRCC specimens achieved larger displacementductility and better sustained intensive cracking damagecompared to the RC columns.

(2) The deformation capacity of the flexure column specimenswas greatly higher than that of the shear–flexure specimensbecause of the difference in the stirrup spacing. But withusing 2% volumetric ratio of PVA fibers, the results showedvery little difference between FC and SF specimens.

(3) The PVA columns exhibited higher initial stiffness than theregular RC specimens, and the use of PVA fibers allowedthe column to delay stiffness drop at higher drift comparedwith the RC specimens, and improved stiffness retentiongreatly.

(4) Biaxial loading tended to reduce the maximum strengthdegradation, and the ductility of the specimen wasdecreased when increasing axial load, or, in other words,the ductility capacity is inversely proportional to axial loadlevel.

(5) The presence of PVA fibers dramatically enhanced the per-formance of the column specimens in terms of energy dissi-pation capacity. Moreover, in columns with the sameproperties, a lower level of energy dissipation capacity wasobserved after increasing axial load.

Finally, it is worth emphasizing that the RC columns’ response ishighly dependent on the loading path. In fact, the effects of biaxialloading on stiffness degradation and strength degradation aremore pronounced. Thus, in columns where demands are expectedwith large moment reversals in both directions, specific detailsshould be provided regarding their critical regions to improvethe columns’ performance and to avoid premature failure.

Acknowledgments

This research was supported by the National Research Founda-tion of Korea (NRF) funded by the Korea government (MEST) (GrantNo. 2011 – 0010384).

References

[1] D.J. Kim, A.E. Naaman, S. El-Tawil, High performance fiber reinforced cementcomposites with innovative slip hardening twisted steel fibers, Int. J. Concr.Struct. Mater. 3 (2) (2009) 119–126.

[2] V. Li, On engineered cementitious composites (ECC); a review of the materialand its application, J. Adv. Concr. Technol. 1 (3) (2003) 215–230.

[3] A.E. Naaman, Engineered steel fibers with optimal properties for reinforcementof cement composites, J. Adv. Concr. Technol. 1 (3) (2003) 241–252.

[4] A.E. Namman, H.W. Reinhardt, Proposed classification of HPFRC compositesbased on their tensile response, Mater. Struct. 39 (2006) 457–555.

[5] E. Yang, V.C. Li, Strain-hardening fiber cement optimization and componenttailoring by means of a micro mechanical model, Constr. Build. Mater. 24(2010) 130–139.

[6] G.J. Parra-Montesinos, High-performance fiber-reinforced cement composites;an alternative for seismic design of structures, ACI Struct. J. 102 (5) (2005)668–675.

[7] American Concrete Institute, Building code requirements for structuralconcrete and commentary, ACI 318M-11, 2011.

[8] F. Qiu, W. Li, P. Pan, J. Qian, Experimental tests on RC columns under biaxialquasi-static loading, Eng. Struct. 24 (2002) 419–428.

[9] K. Tsuno, R. Park, Experimental study of reinforced concrete bridge pierssubjected to bi-directional quasi-static loading, Eng. Struct. JSCE 21 (1) (2004)11–26.

[10] H. Bechtoula, S. Kono, F. Watanabe, Experimental and analytical investigationsof seismic performance of cantilever reinforced concrete columns undervarying transverse and axial loads, J. Asian Archit. Build. Eng. 4 (2005) 467–474.

[11] B. Afshin Canbolat et al., Experimental study on seismic behavior of high-performance fiber-reinforced cement composite coupling beams, ACI Struct. J.102 (2005) 159–166.

[12] K. Kim, G. Parra-montesinos, Behavior of HPFRCC low-rise walls subjected todisplacement reversals, in: International Workshop High Performance FiberReinforced Cement Composites, 2003, pp. 505–515.

[13] S.L. Billington, J. Yoon, Cyclic behavior of precast post-tensioned segmentalconcrete columns with ECC, in: Proceedings of the JCI International Workshopon Ductile Fiber Reinforced Cementitious Composites (DFRCC)—Applicationsand Evaluation (DFRCC-2002), Japan Concrete Institute, Tokyo, 2002, pp. 279–288.

[14] M. Shin, Y. Choi, I. Kim, K. Lee, Effectiveness of low-cost fiber-reinforcedcement composites in hollow columns under cyclic loading, Constr. Build.Mater. 47 (2013) 623–635.

[15] Bengi Arisoy, Wu. Hwai-Chung, Material characteristics of high performancelightweight concrete reinforced with PVA, Constr. Build. Mater. 22 (4) (2008)635–645.

http://refhub.elsevier.com/S0950-0618(15)30759-5/h0005








































[16] T. Horikoshi, et al. Properties of Polyvinyl Alcohol Fiber as ReinforcingMaterials for Cementitious Composites, Industrial materials R&Ddepartment, Kuraray Co., Ltd, Japan, 2014.

[17] K. Sirijaroonchai, S. El-Tawil, G. Parra-Montesinos, Behavior of highperformance fiber reinforced cement composites under multi-axialcompressive loading, Cem. Concr. Compos. 32 (2010) 62–72.

[18] Japan Society of Civil Engineers (JSCEs), Recommendation for design andconstruction of high performance fiber reinforced cement composite withmultiple fine cracks, Concr. Eng. Ser. 82 (2008).

[19] R. Park, Evaluation of ductility of structures and structural assemblages fromlaboratory testing, Bull. N. Z. National Soc. Earthquake Eng. 22 (3) (1989) 155–166.

[20] T. Ohtaki, An experimental study on scale effects in shear failure of reinforcedconcrete columns, Strain 500 (500) (2000). 36-D13..












behavior of high-performance fiber-reinforced cement ...earthquake.hanyang.ac.kr/journal/2016/2016,...

Documents