Experimental Investigation of Rectangular Air-Cured

Geopolymer Concrete Columns Reinforced with

GFRP Bars and Stirrups
M. Elchalakani 1; M. Dong 2; A. Karrech 3; G. Li 4; M. S. Mohamed Ali 5; and B. Yang 6
Downloaded from ascelibrary.org by East Carolina University on 03/02/19. Copyright ASCE. For personal use only; all rights reserved.

Abstract: Glass fiber-reinforced polymer (GFRP) bars in construction are increasing in popularity due to their excellent corrosion resistance,
high tensile strength to weight ratio, and low maintenance. Geopolymer is a modern cementitious material that is known for its corrosion
resistance and low carbon footprint. Combining the two could produce a green yet durable composite material that can be applied to
aggressive environments such as Australia’s coastal zones. This paper experimentally investigates the load-moment interaction of GFRP-
reinforced air-cured geopolymer concrete columns. The behavior of reinforced geopolymer concrete under combined loading were studied
with 11 half-scale specimens. Three different stirrup spacings (75, 150, and 250 mm) were examined. Effective confinement was achieved
by reducing the stirrup spacing such that high strains were measured in the concentric columns with closely spaced (75 mm) stirrups. A
comparison between the experimental data and international design codes showed that such codes were conservative when ignoring the
compressive strengths of the longitudinal GFRP bars. The experimental results were better represented when the compressive strengths
of the bars were included; on average, the GFRP-reinforced geopolymer concrete columns exhibited 10.8% increase in strength with respect
to unreinforced concrete sections. DOI: 10.1061/(ASCE)CC.1943-5614.0000938. © 2019 American Society of Civil Engineers.
Author keywords: Geopolymer; Glass fiber-reinforced polymer (GFRP); Column; Interaction diagram.

Introduction concrete (Davidovits 1994; Duxson et al. 2007b). Research (Palomo

et al. 1999) has shown that geopolymers remain stable for up to
The term geopolymer was introduced by Davidovits (1994) to de- 270 days in various types of aggressive fluids, due to the formation
scribe inorganic polymers that form by combining alumina and of an amorphous zeolitic material during the geopolymerization pro-
silica materials with alkaline silicate solutions. Other researchers cess. Resistance to aggressive fluids provides better protection to
(Duxson et al. 2007a, b) used the same concept to describe alkali embedded reinforcing bars, which is especially advantageous for
activated materials that form when fly ash and ground granulated construction projects in coastal zones.
blast-furnace slag (GGBS) are combined with alkaline substances Glass fiber-reinforced polymer (GFRP) bars have advantages
that strengthen with time under various temperatures (including such as high corrosion resistance, high tensile strength, and low
room temperature). The polymerization process in geopolymers is specific gravity as compared to conventional steel reinforcement.
different from the formation of calcium-silicate-hydrate (C-S-H) However, due to a lower elastic modulus than steel reinforcement,
gels that can be found in ordinary portland cement (OPC) and does the longitudinal GFRP bars require greater lateral restraint to pre-
not require the highly carbon emissive and energy intensive calci- vent buckling failure. The use of GFRP bars as OPC concrete col-
nation process. Utilizing waste residuals such as fly ash, GGBS, umn reinforcement has been explored experimentally. In general,
and contaminated sand and aggregates prevents raw material ex- the behaviors of GFRP-RC columns can be reasonably predicted
traction and reduces waste disposal. Therefore, geopolymer con- using existing design guides for steel reinforcements (Amer et al.
crete (GPC) is deemed a green and sustainable alternative to OPC 1996; Hadi and Youssef 2016). However, Elchalakani et al. (2017)
concluded no distinct balanced failure point can be found on the
1
interaction diagram of GFRP RC columns with longitudinal rein-
Senior Lecturer, School of Civil, Environmental and Mining Engineer- forcement ratios above 3%. It was also found that GFRP-RC mem-
ing, Univ. of Western Australia, Crawley, WA 6009, Australia.
2 bers are required to be overly reinforced due to the brittle nature of
Ph.D. Student, School of Civil, Environmental and Mining Engineer-
ing, Univ. of Western Australia, Crawley, WA 6009, Australia. ORCID: the bars (Choo et al. 2006). In terms of load capacity, GFRP-RC
https://orcid.org/0000-0002-3342-5588. columns showed an average 3%–4% increase in load carrying
3
Associate Professor, School of Civil, Environmental and Mining capacity than the corresponding plain OPC concrete when the
Engineering, Univ. of Western Australia, Crawley, WA 6009, Australia. stress block depth coefficient α2 was assumed to 1.0; the improve-
4
Senior Lecturer, Dept. of Chemical Engineering, Melbourne Univ., ment was greater when better confinement through closely spaced
Melbourne, VIC 3000, Australia. stirrups was achieved (Elchalakani and Ma 2017; Karim et al. 2016;
5
Senior Lecturer, Dept. of Civil Engineering, Univ. of Adelaide, Maranan et al. 2016). Mirmiran et al. (2001) concluded FRP-
Adelaide, SA 5000, Australia. reinforced concrete columns are more sensitive to length effects
6
Professor, Ministry of Education, Chongqing Univ., Chongqing
due to the low elastic modulus of the bars and proposed a reduction
400045, China (corresponding author). Email: yang0206@cqu.edu.cn
Note. This manuscript was submitted on April 23, 2018; approved on
in slenderness limit from 22 for steel-reinforced columns to 17 for
October 17, 2018; published online on February 23, 2019. Discussion per- FRP-reinforced columns.
iod open until July 23, 2019; separate discussions must be submitted for Rahman and Sarker (2011) studied steel-reinforced GPC col-
individual papers. This paper is part of the Journal of Composites for Con- umns under combined axial compression and biaxial bending.
struction, © ASCE, ISSN 1090-0268. The failure mechanism they observed was spalling of cover material

© ASCE 04019011-1 J. Compos. Constr.

J. Compos. Constr., 2019, 23(3): 04019011

 Table 1. Chemical compositions (% by weight) of fly ash and GGBS
Material Al2 O3 BaO CaO Cr2 O3 Fe2 O3 K2 O MgO MnO Na2 O P2 O 5 SO3 SiO2 TiO2 Total
Fly ash 25.31 0.07 12.25 0.01 9.9 0.64 1.14 0.14 0.32 0.51 0.57 46.04 1.46 98.36
GGBS 13.8 0.06 42.13 0 0.58 0.32 5.76 0.27 0.2 0.034 3.33 32.92 0.57 99.97

followed by concrete crushing at approximately midheight of the silicate solution (SiO2 =Na2 O modulus ratio ¼ 3.2) at a ratio by
columns. Sudden and explosive failure with short postpeak behavior weight of 1:2.5. It was designed based on the conclusions drawn
was observed in columns with small load eccentricity and high by Hamidi et al. (2016) to maximize its efficiency.
concrete strength, indicating a low ductility. In GFRP-GPC beams, All specimens were 1,200 mm long with a rectangular cross sec-
nonlinear responses in the load-deflection curves were observed tion of 260 × 160 mm in the main body of the specimen. The di-
due to cracking and crushing of concrete (Maranan et al. 2015). mensions were chosen to ensure the specimens were representable
Downloaded from ascelibrary.org by East Carolina University on 03/02/19. Copyright ASCE. For personal use only; all rights reserved.

Conservative results were obtained using the calculations in ACI at half-scale and their capacity would be lower than the 2,000 kN
440.1R (ACI 2015) and CAN/CSA S806-12 (CAN/CSA 2017). nominal capacity of the Amsler machine in the University of
Maranan et al. (2016) discovered that GPC circular columns rein- Western Australia Structures Laboratory. For a short unbraced col-
forced fully with GFRP bars and stirrups have higher capacities umn, the slenderness ratio (Le =r, where Le = effective length of the
than their OPC concrete counterparts. However, the difference in column; and r = radius of gyration) should be less than or equal to
ductility and confinement efficiency was small. 22 for steel RC structures [AS 3600 (AS 2009)] or 17 for FRP-
The literature review highlighted the lack of experimental stud- reinforced concrete structures (Mirmiran et al. 2001). When Le ¼
ies on GFRP fully reinforced GPC columns. The experimental in- 1,200 mm (pinned connections, full height of the columns), the
vestigations are warranted to study their properties, especially when columns were deemed to be short (Le =rx ¼ 16) about the major
the design standards do not fully address the structural design of axis, and slender (Le =ry ¼ 26) about the minor axis. More realis-
concrete members with GFRP bars. The aim of this experiment tically, when Le ¼ 800 mm (clear distance between the end
is to investigate GFRP-reinforced GPC as a sustainable and durable caps) the columns were short about both axes (Le =rx ¼ 11 and
alternative to conventional reinforced OPC concrete. In this paper, Le =ry ¼ 17). It was decided that eccentrically loaded columns
11 half-scale rectangular GPC columns were constructed with would be tested about the minor axis to prevent premature buckling
GFRP longitudinal bars and stirrups and tested under combined about the major axis.
loading. The strains in the longitudinal GFRP bars were quantified
and the different failure modes of the members were examined.
The improvements by reducing the GFRP stirrup spacing were Column Construction
quantified. In addition, comparisons against AS 3600 (AS All specimens were constructed using GFRP longitudinal bars and
2009), ACI 440.1R (ACI 2015), CAN/CSA S806-12 (CAN/ stirrups. Mateenbar GFRP bars were donated by Pultron Compo-
CSA 2017), and the Geopolymer Handbook (Berndt 2017) were sites in New Zealand. The four corner longitudinal bars were
made. 14 mm in diameter, the stirrups were 8 mm in diameter, and the
cover was 20 mm for all the columns. Longitudinal reinforcement
ratio (ρc ) of all the specimens was 2.22%, which satisfies the rec-
Experimental Program ommended reinforcement ratio by Choo et al. (2006). The proper-
ties of the Mateenbar bars and stirrups (bends) are provided in
Table 2. Various stirrup spacings were adopted for columns to test
Materials, Mix Design, and Specimens
the effect of ligature spacing on core confinement and ultimate
This study involved the construction and testing of 11 rectangular capacities. For the concentrically loaded columns, stirrup spacings
concrete specimens; nine specimens tested as columns under of 75, 150, and 250 mm were used. The shear reinforcement volu-
various eccentricities while two specimens were tested as beams metric ratios were 2.14%, 1.07%, and 0.64% for stirrups spacing of
in a four-point bending test rig. Six 100 × 200 mm cylinders were 75, 150, and 250 mm, respectively. The eccentrically loaded col-
poured alongside each column specimen to obtain the compressive umns had two variations in spacing (75 and 150 mm). Three eccen-
strengths. The column specimens and cylinders were air cured tricities (25, 50, and 75 mm) were investigated for each stirrup
in open areas. Six additional cylinders of the same size were poured spacing. Both the concentric and eccentric columns had increased
as a control group and cured in an ambient moist curing room with sections (top and bottom caps) with denser stirrup distributions at
a constant temperature of 21°C and a humidity of 95%. their ends to prevent premature failures. The beams were con-
A fly ash and GGBS-based GPC mix with a nominal density structed with two spacings between the stirrups of 40 and 80 mm.
of 2,400 kg=m3 consisted of 15% binder, 6.5% alkali solution, Fig. 1 shows the design details of the concentrically and eccen-
24.9% fine aggregates, and 47.3% coarse aggregates. To prevent trically loaded columns with 75 mm stirrup spacing and a beam
flash setting and ensure satisfactory workability, 6.1% water and with 40 mm stirrup spacing. Strain gauges were attached to the
0.1% superplasticizer was added. The binder had 50% of ASTM
Class C (2015) fly ash and 50% GGBS. Table 1 shows the oxide Table 2. Physical properties of GFRP Mateenbars and stirrups as provided
analysis test results of the two materials. The basicity coefficient by Pultron Composites
[K b ¼ ðCaO þ MgOÞ=ðSiO2 þ Al2 O3 Þ (Leong et al. 2016)] was
Tensile Ultimate
0.19 and 1.03 for fly ash and GGBS, and the hydration modulus
Outside Root elastic tensile Ultimate
[HM ¼ ðCaO þ MgO þ Al2 O3 Þ=SiO2 (Leong et al. 2016)] was diameter diameter Area modulus strength strain in
0.84 and 1.87, respectively. The coefficients indicated higher cal- Bar (mm) (mm) (mm2 ) (GPa) (MPa) tension
cium content, which was beneficial for the early age hardening of
the geopolymer. The alkali solution was a mix of 12 M sodium Longitudinal bars 14.0 13.2 136.8 59 930 0.017
Stirrups 8.2 7.4 40.3 55 650 0.012
silicate solution prepared 24 h in advance and N-grade sodium

© ASCE 04019011-2 J. Compos. Constr.

J. Compos. Constr., 2019, 23(3): 04019011

 260 260

N10 N10
Steel Steel
Stirrups

200
200 Stirrups
@ 60mm @ 60mm

1200

8mm

800
N10

1200
8mm
1200
800

GFRP Steel

112
160
GFRP Stirrups
Stirrups Stirrups
@50mm @ 60mm
@50mm
Downloaded from ascelibrary.org by East Carolina University on 03/02/19. Copyright ASCE. For personal use only; all rights reserved.

8mm GFRP Stirrups @ 40mm

120
200 120
200

212 212
260
260
14mm 14mm 212
GFRP GFRP 14mm GFRP bar
bar bar
360
302

260
212

160
112
N12
N12
Steel
Steel
Rebar 112 260
Rebar
160
160
(a) (b) (c)

Fig. 1. Specimen design details of (a) concentrically loaded columns; (b) eccentrically loaded columns; and (c) beams. Distance is measured between
the centerlines of stirrups (in millimeters).

outside midpoint of all longitudinal bars in each specimen to mea- spacing inside its formwork. Fig. 2(c) shows the dimensions of the
sure the maximum strain reached by the bars. Figs. 2(a and b) show GFRP closed stirrups used in this study.
the GFRP cage for the concentrically loaded specimen G250-C and
the GFRP cage of a typical eccentric column with 75 mm stirrup
Testing Method
All the specimens were loaded to failure using a 2,000 kN Amsler
machine with postpeak behavior observed where possible. Sche-
matic drawings of the experimental setup are shown in Fig. 3.
The hydraulic base plate displaces upward to compress the spec-
imens. The loading rate was kept constant at 20 kN=min for
columns and 2 kN=min for beams under a load control regime. Dis-
placement of the hydraulic base plate, the magnitude of compres-
sive force, strain gauge readings, lateral deflection (for eccentric
columns), and midpoint deflection (for beams) were recorded. Note
that a displacement control regime is more appropriate for analyz-
ing the ductility of the columns from the load-displacement curves.
The analysis method was adjusted accordingly, which is explained
in “Column Ductility.”
Steel plates were inserted on both ends to ensure even distribu-
tion of force over the surface of the concentrically loaded column.
To effectively test columns with a given eccentricity, pin supports
were replicated with a flat steel plate welded to a cylindrical steel
roller at the base and the top. The flat plate was allowed to rotate
with the deflection of the column as required throughout the load-
ing process. A string potentiometer was attached to each column at
Fig. 2. Specimen construction: (a) GFRP cage for G250-C; (b) GFRP
midheight to measure the lateral deflection. Two beam specimens
cage of a typical eccentric column inside its formwork; and (c) closed
were tested under four-point bending as per AS 1012.11-00 (AS
stirrups used in this study.
2014). The beam span was 1,100 mm with a 50 mm overhang

© ASCE 04019011-3 J. Compos. Constr.

J. Compos. Constr., 2019, 23(3): 04019011

 P

Loading pin
String pot h=160 mm
Eccentricity
(e)
Original shape

Compression face
Deformed shape Original shape

b=260 mm
Tension face
P
Mid-height Deformed shape
Compression face
Tension face

Compression face
Downloaded from ascelibrary.org by East Carolina University on 03/02/19. Copyright ASCE. For personal use only; all rights reserved.

Lateral Axial 367 mm 367 mm 367 mm

displacement displacement Plan View at Tension face
( ) Mid-height

P Laser
* : strain gauges triangulation
(a) (b)

Fig. 3. Experimental setup for (a) eccentric columns; and (b) beams.

Fig. 4. Specimens after testing.

at each end. Each beam specimen was positioned on two rollers concentric, eccentric, and flexural loading. The specimens were
spaced on a larger steel I-beam. Two steel loading rollers were categorized in terms of stirrup spacing and load eccentricity. For
placed at one-third of the span, 367 mm apart, on the concrete example, G75-50 indicates the specimen had a stirrup spacing
specimen with a spreader bar placed across the top. A triangulation of 75 mm and was loaded at an eccentricity of 50 mm. The “C”
laser was placed under the midpoint of each specimen to measure represents concentric loading while “F” represents flexural loading.
the deflection. The ultimate compression capacity of the corresponding plain con-
crete section would be 1,082 kN (α2 ¼ 1.0) based on the average

Experimental Results
Table 3. Test results of the GFRP-reinforced GPC specimens
Load Carrying Capacity Volumetric
ratio of the
The average compressive strength was 26.0 MPa at 28 days for all Stirrup transverse Peak Bending
the cylinders poured and air cured in the same manner as the spacing reinforcement, Eccentricity load moment Ductility
columns. The control group of cylinders that was moist cured in Specimen (mm) ρs (%) (mm) (kN) (kNm) index
the ambient curing room achieved 39.7 MPa at 28 days. The curing
G75-C 75 0.94 Concentric 1,357 0 2.95
of geopolymer is sensitive to factors such as temperature and hu- G75-25 25 804 20.1 2.19
midity. The temperature and humidity were collected throughout G75-50 50 454 22.7 3.37
the curing process on an hourly basis. As the curing happened pri- G75-75 75 244 18.3 3.78
marily in September in Perth, Australia, the average temperature G150-C 150 0.47 Concentric 1,194 0 2.37
over 28 days was 14.9°C while the average humidity was 71.5%. G150-25 25 657 16.4 —
The substantial difference of 6.1°C in temperature and 23.5% in G150-50 50 353 17.6 3.33
humidity contributed to a 34.5% reduction in strength found in the G150-75 75 234 17.5 3.44
air-cured cylinders. G250-C 250 0.28 Concentric 1,041 0 2.15
Fig. 4 shows the GFRP-GPC specimens after testing. Table 3 G80-F 80 0.88 Flexure 124 22.8 4.72
G40-F 40 1.76 Flexure 172 31.6 —
shows the experimental results of all the specimens tested under

© ASCE 04019011-4 J. Compos. Constr.

J. Compos. Constr., 2019, 23(3): 04019011

 Table 4. Column reinforcement details of the OPC-GFRP columns used by Elchalakani and Ma (2017) and the GPC-GFRP columns in the current study
OPC columns GPC columns
Key parameters of the specimens (Elchalakani and Ma 2017) (current study)
Concrete column dimensions 160 × 260 × 1,200 160 × 260 × 1,200
Stirrup spacings (mm) 75, 150, 250 75, 150, 250
Average cylinder compressive strength, fc0 (MPa) 32.8 26.8
GFRP long bars, diameter (mm) 12a 14b
GFRP stirrups, diameter (mm) 6a 8b
Volumetric ratio of the transverse reinforcement, ρs at 75 mm spacing 1.19% 2.14%
Volumetric ratio of the longitudinal reinforcement, ρc 1.63% 2.22%
Percent increase in load carrying capacity over the corresponding plain concrete (α2 ¼ 1) 3.2% 10.8%
a
V-rod GFRP bars and stirrups from Pultrall, Canada.
b
Mateenbar GFRP bars and stirrups from Pultron Composites, New Zealand.
Downloaded from ascelibrary.org by East Carolina University on 03/02/19. Copyright ASCE. For personal use only; all rights reserved.

concrete cylinder compressive strength of 26.0 MPa and a cross Phase I (A-B): Elastic range
section of 260 × 160 mm. The concentrically loaded columns Phase II (B-C): Inelastic tensile
cracking
(G75-C, G150-C, and G250-C) achieved an average increase of C Phase III (C-E): Sudden collapse and
Pu
10.8% in the load carrying capacity over the plain concrete section. loss of capacity
The improvement would become 30.4% if α2 was taken as 0.85 as 0.85Pu D Phase IV (E-F): Steady failure and
opening of the stirrups
per ACI 440 (ACI 2015). The α2 value of 1.0 was used as a con- 0.75Pu B
servative calculation and a fair comparison against other studies.

Axial Load
E
For example, Table 4 shows the column reinforcement imple-
Pres F
mented by a similar study on the GFRP fully reinforced OPC col-
umns with the same column dimensions (160 × 260 × 1,200 mm)
by Elchalakani and Ma (2017). The GFRP-OPC concrete columns
exhibited an average increment of 3.2% (α2 ¼ 1.0) with respect to
the plain concrete section, which was significantly lower than the ABG
value obtained in this study. Among other variables, there was a
noticeable difference of 80% in the volumetric ratio of the trans- G H
verse reinforcement (ρs ¼ 1.19% for OPC concrete and 2.14%
A 75 85 res
for GPC) between the two types of columns, implying that the more
effective core confinement in the concentric GPC columns was Axial deformation
attained by increasing the amount of transverse reinforcement. In Fig. 5. Ductility index (I).
addition, the volumetric ratio of the longitudinal reinforcement (ρc )
in the GPC specimens was slightly higher at 2.22% as opposed to
1.63% for the OPC concrete specimens.
Reducing stirrup spacing had a positive effect on the load capac- ductility indices for those two specimens were not calculated.
ity for the GFRP-GPC columns. G75-C had a 13.7% increase over As expected, the ductility index increases as the stirrup spacing
G150-C. Similarly, the eccentrically loaded GPC columns with decreases for the concentrically loaded columns, i.e., G75-C,
75 mm stirrup spacing had an 18.4% improvement over those with G150-C, and G250-C. Specimen G75-C had the least concrete
150 mm stirrup spacing. The GPC beam G40-F had a 38.7% higher spalling at failure of the three concentric columns. The effectively
moment capacity than did G80-F due to the more closely spaced confined concrete and well restrained longitudinal bars increased
shear reinforcement. the ability of the column to absorb energy. It was also observed
that ductility index increased as the eccentricity of the load
increased (compare G75-25, G75-50, and G75-75 in Table 3). Col-
Column Ductility umns with smaller stirrup spacings were still found to be more
Column ductility can be measured by the ability to absorb energy ductile. The displacement used for the beam G80-F was in the
after the peak load. The ductility index was determined as shown in transverse direction, which was not directly comparable to the col-
Eq. (1), based on previous studies (Elchalakani and Ma 2017; Hadi umns. However, a high energy absorption ability was observed in
et al. 2016; Hadi and Youssef 2016). The method is presented the beams with two-stage response (with two peaks) as will be
graphically in Fig. 5 discussed in “Beam Behavior.”

ABCDH
I¼ ð1Þ Concentrically Loaded Columns
ABG
The three concentric columns tested (G75-C, G150-C, and G250-C)
where ABCDH is the area under the axial load-axial deflection showed two different failure modes. Fig. 6 shows the specimens
curve up to the point on the postpeak collapse curve where the post- after testing. It can be seen from the figure that as the stirrup spac-
peak load equals 0.85 Pu (Point D in Fig. 5), with Pu representing ing decreases, the core confinement of the concrete increases.
the peak load; and ABG is the area under the axial load-axial de- Specimen G250-C had little of the core concrete remaining after
flection curve up to the first point (Point B in Fig. 5) where the load testing, whereas G150-C and G75-C had increasingly more core
reaches 0.75 Pu . The results are reported in Table 3. The load- concrete remaining intact. For both G250-C and G150-C, vertical
displacement data for G150-25 and G40-F was not captured due cracking was observed on the concrete surface before spalling was
to the malfunction of the data acquisition system, therefore the observed at midheight of the column. Once the cover had spalled,

© ASCE 04019011-5 J. Compos. Constr.

J. Compos. Constr., 2019, 23(3): 04019011

 1600

1400

1200

Axial Load (kN)

1000

Rupture of 800
longitudinal
GFRP bars at Buckling of
longitudinal 600
high strain
GFRP bars
400
G75-C
G150-C Elastic range and
200
microcracking
G250-C
0
-0.012 -0.01 -0.008 -0.006 -0.004 -0.002 0
Strain
Downloaded from ascelibrary.org by East Carolina University on 03/02/19. Copyright ASCE. For personal use only; all rights reserved.

Fig. 8. Load versus strain graphs for concentrically loaded columns

G75-C, G150-C, and G250-C.

Fig. 6. Concentric columns after failure: (a) longitudinal GFRP bar

ruptured in G75-C; (b) stirrups opened and longitudinal bars buckled
the longitudinal bars of G75-C rupturing at εfrp;c ¼ 0.009; the lon-
in G150-C; and (c) stirrups opened and longitudinal bars buckled in
gitudinal bars in G150-C and G250-C buckled upon failure, achiev-
G250-C.
ing strains of approximately 0.006 and 0.004, respectively. Choo
et al. (2006) recommended use of FRP bars with ultimate compres-
sion strains much greater than the concrete compression strain to
1600 prevent premature compression failure. It was discovered that this
G75-C
Steady loss of requirement can be satisfied by providing effective lateral restraint
1400 capacity as G150-C
Spalling of
stirrups opened G250-C
to the longitudinal GFRP bars.
concrete cover
1200
Axial Load (kN)

1000
Eccentrically Loaded Columns
800 Explosive The expected four-phase loading of eccentric columns is illustrated
failure of
600 concrete core in Fig. 5. Phase I (Points A to B) is the elastic range with micro-
cracking in the concrete. Phase II (Points B to C) is the inelastic
400 range with tensile cracks developing in the tension face. Phase III
200 Elastic range and (Point C to E) consists of column failure and loss in load carrying
microcracking capacity. For the columns with uniform crack distribution, the fail-
0
0 2 4 6 8 10 12 14
ure mode at Point C is expected to be spalling of the concrete cover
Axial Displacement (mm)
and crushing of concrete in the compression face. The failure mode
shifts to the slippage of longitudinal GFRP bars after localized
Fig. 7. Load versus displacement graphs for concentrically loaded spalling and crushing of concrete if a plastic hinge has formed.
columns G75-C, G150-C, and G250-C. Plastic bending occurs at the plastic hinge due to the excessive de-
formation of the member (Murray 1984), which will generally
cause a larger loss of capacity. Phase IV (Points E to F) is the steady
loss of load carrying capacity as the stirrups open before the even-
the GFRP stirrups were no longer adequately restrained and there- tual rupture or buckling in the longitudinal bars.
fore began to open up. The longitudinal GFRP bars subsequently Two columns with 75 and 150 mm stirrup spacing were tested
buckled and ruptured. The explosive failures due to core shear fail- with a 25 mm eccentric load applied through a steel roller. Hori-
ure took place shortly after the initial spalling of the cover. A sim- zontal tension cracks appeared on the backside of the column
ilar trend was found in the study by Elchalakani and Ma (2017). without any loss of load capacity. Both columns failed due to con-
Fig. 7 shows the axial load versus axial displacement for the crete crushing on the compression face of the column. Specimen
three concentric columns. Again, the two distinct failure modes G75-25 was seen to maintain a high proportion of the core concrete
can be seen. As expected, columns with smaller stirrup spacing throughout the test, whereas G150-25 was seen with heavy spalling
reached higher peak axial loads. Specimen G75-C had an identical near failure. The load-displacement graphs of G75-25 and G150-25
elastic response to G150-C and G250-C but a different postpeak are shown in Fig. 9. The four phases (from Fig. 5) of the loading
response due to the ductile behaviors of the confined concrete. behavior are clearly seen. The loading of G75-25 resembled that of
The release of energy was more gradual for G75-C. The opening G75-C with failure by crushing of concrete followed by a steady
of the stirrups was also relatively slower and the collapse less rapid. loss of capacity as stirrups opened. The reduction in capacity at
Also, due to the greater number of stirrups in G75-C, failure of an peak load was more abrupt in G150-25. The load reduced from
individual stirrup was less critical and only led to a small drop in 804 to 510 kN (36% reduction) after the failure of the concrete
load. The same explosive failure of the concrete core occurred at a in G75-25, and from 657 to 280 kN (56% reduction) in G150-25.
much later stage, implying a higher ductility of G75-C, which The strains of G75-25 and G150-25 are shown in Fig. 10. At
reached Pu ¼ 1,357 kN or 13.7% and 30.4% greater than G150-C low eccentricity (e ¼ 25 mm), the columns initially exhibited
and G250-C, respectively. responses similar to the columns with concentric loading. Initially,
Fig. 8 shows the strains developed in the GFRP bars. Columns compressive strains developed in the tension bars. As the lateral
with smaller stirrup spacing were able to reach higher strains, with displacement increased, tensile strains developed in the bars.

© ASCE 04019011-6 J. Compos. Constr.

J. Compos. Constr., 2019, 23(3): 04019011

 900
Phase II: G75-25 than G150-C (εfrp;c ¼ 0.004–0.005), resulting in change in failure
800 Tensile G150-25 mode to rupturing of the compression bars.
700 cracks in Two distinct failure modes were observed for the two columns
Phase III: Spalling of concrete
the tension
cover and crushing of loaded at a 50 mm eccentricity, as seen in Fig. 11. Failure of speci-
Axial Load (kN)

600 face
concrete in the compression
face men G75-50 was due to concrete crushing on the compressive face
500
of the column. As the column was loaded, cracks formed in the
400 tension and compression faces of the column and were well distrib-
300 Phase IV: uted along the height of the column [Fig. 11(a)]. The cracks were
Steady loss of
200 Phase I: Elastic capacity as uniformly spread along the column height. Concrete crushing on
range and stirrups the compression face of the column led to the reduction in column
100 microcracking opened
compressive capacity and eventual collapse with rupture of the
0 GFRP bars. Unlike G75-50, the cracks were concentrated close
0 5 10 15 20
to the top of G150-50. The crushing of concrete and cracks on
Axial Displacement (mm)
Downloaded from ascelibrary.org by East Carolina University on 03/02/19. Copyright ASCE. For personal use only; all rights reserved.

the tension side were localized due to the formation of a plastic

Fig. 9. Load versus displacement graphs for columns loaded at 25 mm hinge near the top cap with concrete crushing and a widely opened
eccentricity, G75-25 and G150-25. diagonal shear crack as shown in Fig. 11(b). The longitudinal bars
slipped as a result of the large deformation. Fig. 12 shows the in-
creased compressive capacity (22.2%) of G75-50 over G150-50.
The postpeak capacity of G75-50 was also significantly higher than
900
Tensile cracks in the
that of G150-50, indicating a higher ductility. The four different
800 tension face response phases from Fig. 5 were again clearly seen in G75-50.
700 However, the load reduction after concrete crushing was 42%,
600 which was 6% greater than the load reduction in G75-25. However,
Axial Load (kN)

500 the steady loss of capacity after the sudden drop was at a reduced
rate compared with G75-C and G75-25. The failure of G150-50
400
Crushing of was caused by the slippage of the longitudinal GFRP bars in
300 concrete in the
compression the tension zone that was initiated at peak load. The formation
200 face
of the plastic hinge resulted in a significant drop in the column’s
100 G75-25 Initial compressive load carrying capacity. A second peak was observed after the initial
strains
G150-25
0 failure of G150-50. Fig. 13 shows the strains recorded in the strain
-0.012 -0.01 -0.008 -0.006 -0.004 -0.002 0 0.002 0.004 0.006 0.008
gauges. The longitudinal GFRP bars on the tension face of G150-50
Strain
slipped and unloaded upon failure, which was not observed in
Fig. 10. Load versus strain graphs for columns loaded at 25 mm G75-50. The strain in the tension bars of G75-50 reached εfrp;t ¼
eccentricity, G75-25 and G150-25. 0.013 while the strains reached 0.006 in G150-50, showing the pre-
mature slip failure of the latter column. The GFRP bars on the com-
pression face of G75-50 failed at εfrp;c ¼ 0.008; bar buckling in
G150-50 occurred within the range of εfrp;c ¼ 0.005 to 0.006.
The tensile strains at peak load were less than εfrp;t ¼ 0.007 for The two columns loaded at high eccentricity failed in the same
both columns. The compressive strain in the bars of G75-25 at peak manner as G150-50 with the four phases of response from Fig. 5.
load reached the same level as the concentrically loaded G75-C Instead of the tension cracks and concrete crushing propagating
before rupturing of the bars (εfrp;c ¼ 0.01), indicating effective from the midpoint of the column, failure initiated close to the head
confinement was achieved. The compressive strain at peak load of the columns by forming a plastic hinge with concrete crushing
in the bars of G150-25 (εfrp;c ¼ 0.008) were moderately higher and a widely opened diagonal shear crack. The longitudinal GFRP

Fig. 11. Columns loaded at 50 mm eccentricity after testing: (a) G75-50; and (b) G150-50.

© ASCE 04019011-7 J. Compos. Constr.

J. Compos. Constr., 2019, 23(3): 04019011

 300
500 Phase II: Tensile cracks in
the tension face Formation of plastic
G75-50 hinge beneath the
450 G150-50 250
Phase III: Crushing of top cap
400 concrete on the
compression face 200

Axial Load (kN)

350
Axial Load (kN)

300 150
250
Phase IV: Steady 100
200 Phase III: Sudden
loss of capacity due loss of capacity
150 as stirrups opened
to slip failure 50
100 G75-75
Phase I: Elastic range G150-75
50 and microcracking 0
-0.008 -0.006 -0.004 -0.002 0 0.002 0.004 0.006 0.008 0.01
0 Strain
0 2 4 6 8 10 12 14 16 18 20
Downloaded from ascelibrary.org by East Carolina University on 03/02/19. Copyright ASCE. For personal use only; all rights reserved.

Axial Displacement (mm) Fig. 15. Load versus strain graphs for columns loaded at 75 mm
eccentricity, G75-75 and G150-75.
Fig. 12. Load versus displacement graphs for columns loaded at
50 mm eccentricity, G75-50 and G150-50.
face buckled at strain within εfrp;c ¼ 0.004 to 0.006. Similarly to
G150-50, the longitudinal bars in the tension face in G75-75 and
500 G150-75 experienced slip failures. The two columns loaded at high
Crushing Formation
450 of
concrete
of plastic eccentricity both failed prematurely at similar strains, indicating the
hinge
400 in the beneath the limiting factor was the interfacial bond between the GFRP bar and
compress
350 ion face top cap the concrete.
Axial Load (kN)

300 Tensile cracks

250 in the tension
face Beam Behavior
200
150
The two beam specimens G80-F and G40-F had the same tension
reinforcement ratio (ρt ) of 1.18% at 80 and 40 mm stirrup spacing.
100
G75-50
Figs. 16(a and b) shows the similar failure modes of G80-F and
50
G150-50 G40-F, respectively. The two beams experienced concrete crushing
0
-0.01 -0.005 0 0.005 0.01 0.015 in the compression face (the first peak) followed by the formation
Strain of critical diagonal cracks (the second peak). Fig. 16(a) shows the
mild crushing in the compression face, diagonal shear crack and the
Fig. 13. Load versus strain graphs for columns loaded at 50 mm
tensile cracks in the tension face, and the overall bending of G80-F
eccentricity, G75-50 and G150-50.
at failure. Fig. 16(b) shows the severe crushing of the concrete on
the compression face, spalling of concrete on the tension face and
the widely opened critical diagonal shear crack formed after testing
300 G40-F. Fig. 17 shows the load-displacement curve for G40-F. A
Phase II: Tensile cracks in
the tension face
G75-75 clear two-stage response with two peaks can be observed. The
Phase III: Crushing of
250 concrete on the compression G150-75 load-deflection data for G80-F was not captured due to a problem
face
with the data acquisition system. However, an ultimate load (Pu ) of
200 123.7 kN was recorded, which corresponded to a 38.7% reduction
Axial Load (kN)

Phase III: Sudden from G40-F, implying a composite failure mode through bending
150 loss of capacity and shear occurred. Fig. 18 shows the strains in the GFRP bars in
due to slip failure
Phase IV: Steady loss G80-F and G40-F. Specimen G80-F collapsed due to the formation
of capacity as
100
stirrups opened
of a large diagonal shear crack and the localized bending of the bars
in the tension face at a tensile strain of approximately 0.01. Speci-
50 Phase I: Elastic range men G40-F showed a greatly improved strain capacity as indicated
and microcracking by the compressive strains (the strain gauges on the tension face
0 detached at high strains), again indicating bending and shear failure
0 5 10 15 20 25 30 35 40
Axial Displacement (mm)
occurred. The failure mode could be optimized by reducing the
stirrup spacing.
Fig. 14. Load versus displacement graphs for columns loaded at
75 mm eccentricity, G75-75 and G150-75. Interaction Diagrams
Three international design standards, namely AS 3600 (AS 2009),
ACI 440.1R (ACI 2015), and CAN/CSA S806-12 (CAN/CSA
bars in the tension face slipped at failure. Specimen G75-75 2017), were compared against the experimental results. AS 3600
reached a higher peak axial load than G150-75. Fig. 14 shows (AS 2009) applies to steel-reinforced OPC concrete designs. The
the load-displacement behaviors of the two specimens. The failure mechanical properties of GFRP bars in Table 2 were used in the
mode and postpeak behavior were almost identical with the same calculations. The GFRP bars were assumed to have the same elastic
level of reduction in axial capacity postcollapse followed by a sim- modulus in compression and to fail at half of the ultimate strain in
ilar postpeak response. The strains recorded in the GFRP bars are tension. ACI 440.1R (ACI 2015) and CAN/CSA S806-12(CAN/
shown in Fig. 15. The longitudinal GFRP bars in the compression CSA 2017) are design codes for GFRP-reinforced structures.

© ASCE 04019011-8 J. Compos. Constr.

J. Compos. Constr., 2019, 23(3): 04019011

 Fig. 16. GPC beams: (a) G80-F at the formation of diagonal shear crack and crushing of the concrete; and (b) G40-F crushing of the concrete and the
Downloaded from ascelibrary.org by East Carolina University on 03/02/19. Copyright ASCE. For personal use only; all rights reserved.

formation of diagonal shear crack and spalling of concrete cover.

250 b=260 mm u 2 f’c

Shear failure by critical diagonal crack. G40-F frp,c ffrp,c

d=133 mm
Local bending of GFRP bars on the tension dn
200 dn
face and spalling of concrete cover
Axial Load (kN)

Cracks on frp,t
150 ffrp,t
the tension
face
Sudden crushing Fig. 19. Parameters α2 and γ from the rectangular stress block method
100
of concrete on the in AS 3600 (AS 2009). In this figure εfrp;c = strain in the GFRP bars in
compression face the compression face; εfrp;t = strain in the GFRP bars in the tension
50 face; εu = ultimate strain of the unconfined concrete; f frp;c = compres-
First tensile flexure cracking sive stress in the bars in the compression face; ffrp;t = tensile stress in
in the tension face
the bars in the tension face; d = distance from the compressive concrete
0
surface to the center of the tensile GFRP bars; and dn = distance from
0 10 20 30 40 50 60
compressive concrete surface to the neutral axis.
Displacement (mm)

Fig. 17. Load versus displacement graph for G40-F.

Table 5. Comparison of the rectangular stress block parameters equivalent

to α2 and γ in AS 3600 in the four design curves
200
Design guidelines α2 γ
180 Shear failure by critical diagonal Strain gauges
crack. Crushing and cracking of detached from Geopolymer Handbook (Berndt 2017) 0.85 0.69
160 concrete on the compression face the bars AS 3600 (AS 2009) 0.85 0.83
140 ACI 440.1R (ACI 2015) 0.85 0.82
Axial Load (kN)

120 CAN/CSA S806-12 (CAN/CSA 2017) 0.80 0.89

100
80
60
40 2017). Table 5 compares the values of γ and α2 in the four design
20 G40-F Tensile cracks in the tension face curves
0 G80-F
-0.01 -0.005 0 0.005 0.01 0.015 γ ¼ −0.0041f c0 þ 0.818ð0.46 ≤ γ ≤ 0.7Þ ð2Þ
Strain

Fig. 18. Load versus strain graphs for beams G40-F and G80-F. α2 ¼ 1.0 − 0.003fc0 ð0.67 ≤ α2 ≤ 0.85Þ ð3Þ

Fig. 20 shows the normalized interaction diagrams obtained

with test data from this experiment and from the previous study
However, neither allows the inclusion of longitudinal GFRP bars by Elchalakani and Ma (2017) carried out on OPC reinforced con-
in compression. To account for the difference in mechanical proper- crete columns using GFRP bars and stirrups. The axial load was
ties between GPC and OPC concrete, the recommended modifica- normalized by N n (N n ¼ bhf c0 , where b = width; and h = height
tion to AS 3600-09 (2009) in the research report by Berndt of the specimen cross section) while the bending moment was nor-
(2017)—referred to as Geopolymer Handbook—was adopted as malized by M n (M n ¼ bh2 fc0 ). The solid curves in Fig. 20 represent
a separate curve. The two parameters in the rectangular stress block design curves when the compression capacity of the GFRP bars is
method (γ and α2 , shown in Fig. 19) were calculated based on not considered, as recommended by the design codes (f frp;c ¼ 0).
Eqs. (2) and (3), respectively. Equivalent parameters are found The model in the Geopolymer Handbook [Fig. 20(d)] was the most
in ACI 440.1R (ACI 2015) and CAN/CSA S806-12 (CAN/CSA conservative model in terms of bending capacity. The standard

© ASCE 04019011-9 J. Compos. Constr.

J. Compos. Constr., 2019, 23(3): 04019011

 1.4 1.4
AS 3600-09 (f'c =26.0 MPa, ffrp,c=0.5 ) ACI 440.1R-15 (f'c=26.0 MPa, ffrp,c=0.5 )

Normalised axial load, Nu /(bhf'c)

Normalised axial load, Nu /(bhf'c) AS 3600-09 (f'c =26.0 MPa, ffrp,c=0) ACI 440.1R-15 (f'c=26.0 MPa, ffrp,c=0)
1.2 1.2
AS 3600-09 (f'c =32.8 MPa, ffrp,c=0.5 ) ACI 440.1R-15 (f'c=32.8 MPa, ffrp,c=0.5 )
AS 3600-09 (f'c =32.8 MPa, ffrp,c=0) ACI 440.1R-15 (f'c=32.8 MPa, ffrp,c=0)
1 OPC columns (Elchalakani and Ma 2017) 1 OPC columns (Elchalakani and Ma 2017)
GPC columns (this study) GPC columns (this study)
0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2
Downloaded from ascelibrary.org by East Carolina University on 03/02/19. Copyright ASCE. For personal use only; all rights reserved.

0 0
0 0.05 0.1 0.15 0.2 0 0.05 0.1 0.15 0.2
(a) Normalised bending moment, Mu /(bh2f'c) (b) Normalised bending moment, Mu /(bh2f'c)

1.4 1.4
CAN/CSA S806-12 (f'c=26.0 MPa, ffrp,c=0.5 ) Geopolymer Handbook ( f'c =26.0 MPa, ffrp,c=0.5 )
CAN/CSA S806-12 (f'c=26.0 MPa, ffrp,c=0) Geopolymer Handbook ( f'c =26.0 MPa, ffrp,c=0)

Normalised axial load, Nu/(bhf'c)

Normalised axial load, Nu /(bhf'c)

1.2 1.2 Geopolymer Handbook ( f'c =32.8 MPa, ffrp,c=0.5

CAN/CSA S806-12 (f'c=32.8 MPa, ffrp,c=0.5 ) )
CAN/CSA S806-12 ( f'c=32.8 MPa, ffrp,c=0) Geopolymer Handbook ( f'c =32.8 MPa, ffrp,c =0)
1 OPC columns (Elchalakani and Ma 2017) 1 OPC columns (Elchalakani and Ma 2017)
GPC columns (this study) GPC columns (this study)
0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0
0 0.05 0.1 0.15 0.2 0 0.05 0.1 0.15 0.2
(c) Normalised bending moment, Mu /(bh2f'c) (d) Normalised bending moment, Mu /(bh2f'c)

Fig. 20. Comparisons of results from this study and Elchalakani and Ma (2017) with (a) AS 3600 (AS 2009); (b) ACI 440.1R (ACI 2015);
(c) CAN/CSA S806-12 (CAN/CSA 2017); and (d) Geopolymer Handbook (Berndt 2017). In this figure fc0 = unconfined compressive strength
of the concrete; ffrp;c = compressive strength of the longitudinal GFRP bars; and ffrp;t = tensile strength of the longitudinal GFRP bars.

CAN/CSA S806-12 (CAN/CSA 2017) [Fig. 20(c)] was more lower than that of the OPC concrete beams; however, they were still
conservative in terms of axial load capacity. It is clear that all considerably higher than the predictions of the design codes.
the design curves favorably underpredicted the capacities of the
specimens when compression in GFRP bars is neglected. However,
the omission of the GFRP bar compression capacity and the con- Conclusions
finement of the concrete led to noticeable discrepancies between
the curves and the experimental results. The dashed curves in GFRP bars and GPC have been increasingly studied over the past
Fig. 20 represent the four design curves with the compression decade. This paper investigates 11 half-scale GFRP-reinforced
capacity of the GFRP bars included as f frp;c ¼ 0.5ffrp;t as recom- GPC specimens (1,200 × 260 × 160 mm) under various axial and
mended by Elchalakani and Ma (2017). At zero eccentricity, G75-C flexural load combinations. Two stirrup spacings (75 and 150 mm)
and G150-C still reached code predictions, whereas G250-C did and three eccentricities (25, 50, and 75 mm) were examined. The
not. The capacities of the eccentrically loaded columns were pre- experimental results were then compared to AS 3600 (AS 2009),
dicted at a reasonable accuracy by the design codes. However, ACI 440.1R (ACI 2015), CAN/CSA S806-12 (CAN/CSA 2017)
the design codes underpredicted the moment capacities of the and the Geopolymer Handbook (Berndt 2017). The following con-
GPC flexure specimens (G40-F and G80-F) as well as the OPC ones. clusions were drawn based on the experimental results presented:
The concentric GPC columns on average have 4% higher load car- • The air-cured (14.9°C and 71.5% humidity on average) GPC
rying capacities than the concentric OPC concrete columns due to cylinders had a 34.5% reduction in strength than similar
the higher amount of transverse reinforcement (Table 4). Note that cylinders cured in an ambient curing room (a constant 21°C
the volumetric ratio of the transverse reinforcement was ρs ¼ 1.19% and 95% humidity), suggesting the impact of curing conditions
and 2.14% for OPC and GPC columns, respectively. However, on geopolymer hardening.
the capacities of the eccentric GPC columns were moderately lower • Higher loads and moments were achieved for specimens with
than the respective OPC columns. This was consistent with the as- smaller stirrup spacing. Reducing the stirrup spacing from
sumptions made by the Geopolymer Handbook to accommodate the 150 to 75 mm resulted in a 13.7% and 18.4% improve-
smaller rectangular stress block found in GPC flexure tests (Berndt ment of load capacity for the concentrically and eccentrically
2017). The bending capacities of the GPC beams were on average loaded columns, respectively. A 38.7% improvement in moment

© ASCE 04019011-10 J. Compos. Constr.

J. Compos. Constr., 2019, 23(3): 04019011

 capacity was observed when the stirrup spacing reduced from CAN/CSA (Canadian Standards Association/National Standard of
80 mm to 40 mm for the beam specimens. Canada). 2017. Design and construction of building components
• The concentric columns with smaller stirrup spacings were more with fibre-reinforced polymers. CAN/CSA S806-12. Mississauga,
ductile. The ductility increased when the columns were loaded ON, Canada: CAN/CSA.
Choo, C. C., I. E. Harik, and H. Gesund. 2006. “Minimum reinforcement
at an increasing eccentricity.
ratio for fiber-reinforced polymer reinforced concrete rectangular
• High strains (>0.003) were observed in the specimens, indicat- columns.” ACI Struct. J. 103 (3): 460–466.
ing that effective confinement has been achieved by reducing Davidovits, J. 1994. “High-alkali cements for 21st century concretes.” Am.
the stirrup spacing. Bar buckling was observed at strains within Concr. Inst. 144: 383–398.
the range of 0.004–0.006 whereas bar failure in compression Duxson, P., A. Fernández-Jiménez, J. L. Provis, G. C. Lukey, A. Palomo,
was observed at strains within the range of 0.008–0.01. The and J. S. J. Van Deventer. 2007a. “Geopolymer technology: The current
large compression strains compared with concrete strain capa- state of the art.” J. Mater. Sci. 42 (9): 2917–2933. https://doi.org/10
city showed that the premature compression failure of the bars .1007/s10853-006-0637-z.
could be prevented through sufficient lateral restraint from the Duxson, P., J. L. Provis, G. C. Lukey, and J. S. J. van Deventer. 2007b. “The
role of inorganic polymer technology in the development of ‘green con-
Downloaded from ascelibrary.org by East Carolina University on 03/02/19. Copyright ASCE. For personal use only; all rights reserved.

GFRP stirrups.
crete.’” Cem. Concr. Res. 37 (12): 1590–1597. https://doi.org/10.1016/j
• Slip failures were observed in the columns loaded at higher ec-
.cemconres.2007.08.018.
centricities. The formation of a localized plastic hinge with con- Elchalakani, M., and G. Ma. 2017. “Tests of glass fibre reinforced polymer
crete crushing and a widely opened diagonal shear crack beneath rectangular concrete columns subjected to concentric and eccentric
the top cap resulted in a reduction in the load carrying capacity. axial loading.” Eng. Struct. 151: 93–104. https://doi.org/10.1016/j
• A comparison between the experimental data of GPC specimens .engstruct.2017.08.023.
and international design codes showed that such codes were Elchalakani, M., G. Ma, F. Aslani, and W. Duan. 2017. “Design of GFRP-
conservative when ignoring the compressive strengths of the reinforced rectangular concrete columns under eccentric axial loading.”
longitudinal GFRP bars. The experimental results for the Mag. Concr. Res. 69 (17): 865–877. https://doi.org/10.1680/jmacr.16
GPC specimens were better represented when the compressive .00437.
strengths of the bars were included. It was found that the GFRP- Hadi, M., H. Karim, and N. Sheikh. 2016. “Experimental investigations on
circular concrete columns reinforced with GFRP bars and helices under
reinforced geopolymer concrete columns exhibited 10.8% aver-
different loading conditions.” J. Compos. Constr. 20 (4): 04016009.
age increase in strength with respect to plain concrete sections. https://doi.org/10.1061/(ASCE)CC.1943-5614.0000670.
• The study highlighted the significant effect of the geometrical Hadi, M., and J. Youssef. 2016. “Experimental investigation of GFRP-
shape of the closed stirrups. The concentrically loaded columns reinforced and GFRP-encased square concrete specimens under axial
failed when the stirrups opened. In the future, it is recommended and eccentric load, and four-point bending test.” J. Compos. Constr.
to increase the overlapping distance of such stirrups at least by a 20 (5): 04016020. https://doi.org/10.1061/(ASCE)CC.1943-5614
factor of two. .0000675.
Hamidi, R. M., Z. Man, and K. A. Azizli. 2016. “Concentration of NaOH
and the effect on the properties of fly ash based geopolymer.” Procedia
Acknowledgments Eng. 148: 189–193. https://doi.org/10.1016/j.proeng.2016.06.568.
Karim, H., M. N. Sheikh, and M. N. S. Hadi. 2016. “Axial load-axial de-
This study is supported by 111 Project of China (Grant No. B18062). formation behaviour of circular concrete columns reinforced with
The authors thank the donations and support provided by Pultron GFRP bars and helices.” Constr. Build. Mater. 112: 1147–1157.
https://doi.org/10.1016/j.conbuildmat.2016.02.219.
Composites, New Zealand. The authors acknowledge the support re-
Leong, H. Y., D. E. L. Ong, J. G. Sanjayan, and A. Nazari. 2016. “Suit-
ceived from Gary Boon and Anthony Miles from Sika in Australia. ability of Sarawak and Gladstone fly ash to produce geopolymers: A
Thanks are also given to former student Mr. Jordan Beard and la- physical, chemical, mechanical, mineralogical and microstructural
boratory technicians Mr. Jim Waters, Mr. Brad Rose, and Mr. Matt analysis.” Ceram. Int. 42 (8): 9613–9620. https://doi.org/10.1016/j
Arpin for the help with the many practical aspects of this project. .ceramint.2016.03.046.
Maranan, G. B., A. C. Manalo, B. Benmokrane, W. Karunasena, and
P. Mendis. 2015. “Evaluation of the flexural strength and serviceability
References of geopolymer concrete beams reinforced with glass-fibre-reinforced
polymer (GFRP) bars.” Eng. Struct. 101: 529–541. https://doi.org/10
ACI (American Concrete Institute). 2015. Guide for the design and .1016/j.engstruct.2015.08.003.
construction of concrete reinforced with FRP bars. ACI 440.1R. Maranan, G. B., A. C. Manalo, B. Benmokrane, W. Karunasena, and
Farmington Hills, MI: ACI. P. Mendis. 2016. “Behavior of concentrically loaded geopolymer-
Amer, A., M. Arockiasamy, and M. Shahawy. 1996. “Ultimate strength of concrete circular columns reinforced longitudinally and transversely
eccentrically loaded concrete columns reinforced with CFRP bars.” In with GFRP bars.” Eng. Struct. 117: 422–436. https://doi.org/10.1016/j
Proc., 2nd Int. Conf. on Advanced Composite Materials in Bridges and .engstruct.2016.03.036.
Structures (ACMBS-II), 209–216. Montréal, Canada: Canadian Society Mirmiran, A., W. Yuan, and X. Chen. 2001. “Design for slenderness in
for Civil Engineering. concrete columns internally reinforced with fiber-reinforced polymer
AS (Standards Australia). 2009. Concrete structures. AS 3600. Sydney, bars.” ACI Struct. J. 98 (1): 116–125.
Australia: Standards Australia. Murray, N. W. 1984. Introduction to the theory of thin-walled structures,
AS (Standards Australia). 2014. Methods of testing concrete: Determina- Oxford Engineering Science Series. Oxford: Clarendon Press.
tion of the modulus of rupture. AS 1012.11. Sydney, Australia: Palomo, A., M. T. Blanco-Varela, M. L. Granizo, F. Puertas, T. Vazquez,
Standards Australia. and M. W. Grutzeck. 1999. “Chemical stability of cementitious materi-
ASTM. 2015. Standard specification for coal fly ash and raw or als based on metakaolin.” Cem. Concr. Res. 29 (7): 997–1004. https://
calcined natural pozzolan for use in concrete. ASTM C618. West doi.org/10.1016/S0008-8846(99)00074-5.
Conshohocken, PA: ASTM. Rahman, M. M., and P. K. Sarker. 2011. “Geopolymer concrete columns
Berndt, M. 2017. Guide to specification and use of geopolymer concrete under combined axial load and biaxial bending.” In Proc., Concrete
with case studies. Sydney, Australia: Standards Australia. 2011 Conf. Perth, Australia: Concrete Institute of Australia.

© ASCE 04019011-11 J. Compos. Constr.

J. Compos. Constr., 2019, 23(3): 04019011