Research Paper

Advances in Structural Engineering

2022, Vol. 25(9) 1851–1865
Study on mechanical properties of seawater © The Author(s) 2022
Article reuse guidelines:
sea-sand coral aggregate concrete-filled sagepub.com/journals-permissions
DOI: 10.1177/13694332221080597
journals.sagepub.com/home/ase
BFRP tubular arches

Zhiqiang Dong1,2, Ziqing Liu1,2 , Gang Wu1,2 , Hong Zhu1,2 , Fei Lu1, Zeyu Yan1,2 and
XinXing Shao2

Abstract
In this paper, a novel arch-shaped BFRP tube with longitudinal BFRP bars pre-bonded on the inner wall is proposed and
fabricated. It can provide strength in both circumferential and longitudinal directions and is suitable for the rapid con-
struction of arch bridges in island environments by filling with seawater sea-sand coral aggregate concrete (SWCC) in situ.
Ten SWCC-filled BFRP tubular arches were fabricated and tested under monotonic loading. Test variables include the wall
thickness of the BFRP tube, the radius angles of the arch, and with or without pre-bonded BFRP bars. Test results showed
that the failure modes of arches without BFRP bars were bending failure under the loading point. The ultimate bearing
capacities were increased by 33.1% and 71.7% as the wall thickness increased from 2 mm to 3 mm and 4 mm, respectively.
The failure modes of arches containing pre-bonded BFRP bars were bending-shear failure close to the loading plate. The
ultimate bearing capacity of arches, which containing BFRP bars, with a wall thickness of 2 mm, 3 mm, and 4 mm increased
by 215%, 248%, and 189%, respectively, compared with the specimens without BFRP bars. The integrity and ultimate
bearing capacities of arches decreased with the increase of the radius angle and their failure modes exhibited obvious brittle
damage characteristics. The proposed new type of arch has good prospects for application in the marine environment.

Keywords
BFRP tubular arches, coral aggregates, seawater sea-sand concrete, pre-bonded bfrp bars, bearing capacity

Introduction aggregates, the strength and long-term performance of the

coral aggregate concrete in the marine environment is also
There are many challenges to efficiently carry out infra- relatively low (Li et al., 2013), making it difficult to be used
structure construction in harsh island environments. First, reliably as a structural member (Wang et al., 2021; Cheng
the service environment is extremely corrosive and the et al., 2018; Niu et al., 2020; Lyu et al., 2019; Da et al.,
construction window is short; second, there is a lack of 2016; Su et al., 2018).
human resources, large-scale equipment, and construction Fiber-reinforced polymers (FRPs) have good resistance
materials, in particular, raw materials for concrete need to to chloride ions attack and superior durability (Cromwell
be transported from far away inland, which is costly and et al., 2011; El-Hacha et al., 2010; Teng et al., 2018). It is
time-consuming. The use of locally available seawater, sea-
sand, or even coral aggregates for concrete preparation (i.e.,
1
seawater sea-sand coral aggregate concrete, SWCC) can Key Laboratory of Concrete and Prestressed Concrete Structures of
significantly reduce the construction cost and shorten the Ministry of Education, Southeast University, Nanjing, China.
2
National and Local Joint Engineering Research Center for Intelligent
construction period. The study of SWCC has again re-
Construction and Maintenance, Nanjing, China.
ceived wide attention from scholars in recent years (Xiao
et al., 2017; Dong et al., 2020b; Zhou et al., 2021). Corresponding author:
However, the application of SWCC faces a serious bot- Hong Zhu, Key Laboratory of Concrete and Prestressed Concrete
tleneck problem, that is, the excessive chloride ion content, Structures of Ministry of Education, National and Local Joint Engineering
Research Center for Intelligent Construction and Maintenance,
which will lead to accelerated corrosion of steel and thus Southeast University, NO.2, Dongnandaxue Road, Nanjing 211189,
seriously shortens the service life of infrastructure. China.
Moreover, due to the brittle and porous nature of coral Email: [email protected]
1852 Advances in Structural Engineering 25(9)

regarded as a promising material for replacing ordinary possibility of its application in underground protection
steel reinforcements. By combining FRP bars/tubes with engineering. Wang et al. (2019) investigated the dynamic
the seawater sea-sand concrete, which is rich in chloride response of CFFT arches under blast loading with two
ions, good comprehensive performance can be achieved types of Carbon-FRP (CFRP) wall thicknesses and two
(Dong et al., 2020a, 2021; Gao et al., 2020; Xiong et al., types of concrete. The test results showed that the blast
2021; Zhang et al., 2020). To date, numerous ways have resistance performance of CFFT arches could be improved
been proposed in the last three decades to promote the by increasing the CFRP wall thickness and concrete
application of FRP in civil engineering, mainly include strength. In addition, De Waal et al. (2018) designed a
using it as internal reinforcements (De Luca et al., 2010; novel double-skin tubular arch bridge system, which
Tobbi et al., 2014; Afifi et al., 2014; Hadi et al., 2016; consists of an outer FRP tube, an inner steel tube, and a
Pantelides et al., 2013) and using it as external layer of concrete sandwiched between them.
strengthening/confinements (Zeng et al., 2020; Wang et al., Unlike isotropic steel tubes, FRP tubes prepared at this
2020; Hadi et al., 2015; Yu et al., 2017; Ozbakkaloglu et al., stage are linear and anisotropic and can mainly provide
2013). Among them, concrete-filled FRP tube (CFFT) circumferential tension strength (FRP tubes prepared by the
structures are an efficient structural element, since the FRP pultrusion process mainly provide longitudinal strength).
tube can not only be used as permanent formwork to fa- However, in the actual arch structure, there are not only
cilitate the construction but also can improve the me- axial forces but also certain bending moments. Therefore,
chanical properties of the core concrete by offering lateral not only circumferential confinement but also longitudinal
confinement (Wei et al., 2021; Bazli et al., 2020; Wang reinforcements are required. According to the need for
et al., 2017). What is more, the FRP tube can also protect rapid construction of high durability arch bridges in an
the core concrete from a harsh service environment. island environment, a new type of integrated basalt-FRP
Therefore, the combination of the SWCC and the FRP tube (BFRP) tube arch with longitudinal BFRP bars pre-bonded
has good prospects for application in harsh marine in the inner wall and with the SWCC filled inside was
environments. fabricated and tested. The effects of parameters such as the
At present, the research of CFFT structures adopts linear wall thickness of the BFRP tube, radius angles, and with or
members, while the research and application for curved without longitudinal BFRP bars were analyzed through the
CFFT structures are relatively raw and are still in the initial bearing capacity test under monotonic loading. The re-
stage (Meng and Wang, 2018). Dagher et al. (2012) in- search results showed that the proposed new arch bridge
vestigated the bending performance of six CFFT arches by has good bearing performance, and the pre-bonded BFRP
three-point bending test. The test results found that the bars in the inner wall can significantly improve the bearing
tensile rupture of the FRP tube occurred first at the bottom performance of the arch, which shows a broad application
side of the top of the arch, and the arch changed to be a prospect in the infrastructure construction in distant
three-hinged arch. As the load continues to be applied, the islands.
tensile rupture of FRP also occurred at about 1/4 span and
the arch lost its bearing capacity. The study proved that the
CFFT arch has good strength, stability, and fatigue resis- Materials
tance. Nonlinear finite element (FE) analyses on CFFT
Basalt fiber sheets and resin
arches under static and dynamic loads were carried out by
Wen et al. (2012) and Walton et al. (2016), respectively. The BFRP tube used in this paper was prepared by the wet
Based on the developed technology for the preparation of lay-up process in the laboratory. As shown in Figure 1(a),
curved FRP tubes (Tomblin, 2006), and also supported by the basalt fiber sheets used for winding and preparing the
the aforementioned laboratory test data and FE analysis BFRP tube were provided by Zhejiang GBF Basalt Co.,
data, Dr. Dagher, together with relevant experts and in- LTD. As shown in Figure 1(b), the resin matrix used was
vestors, established the company named Advanced Infra- provided by Shanghai Sanyu Resin Co., LTD. with a
structure Technologies (AIT) Bridges, specializing in the product model of L-500AS/BS. The mass ratio of the
consulting, design, and construction services for the above- BFRP fiber sheet and the resin was 7:3. The properties
mentioned CFFT arch bridges. CFFT arches can not only be provided by the manufacturer are listed in Table 1.
applied to the above-mentioned culvert arch bridges but
also the underground protection project. Wang and Chen
BFRP bars and tubes
(2018) conducted a series of studies on the blast resistance
performance of arch-shaped CFFT underground protection The BFRP bars bonded to the inner wall of the BFRP
engineering. The test results found that after suffering two tubular arch are produced and supplied by Jiangsu Green
explosions, the CFFT arches could remain intact and show Materials Valley New Material T&D Co., Ltd. (GMV), and
good blast resistance performance, demonstrating the its nominal diameter is 8.0 mm. According to the ASTM
Dong et al. 1853

Figure 1. Raw materials for preparing the BFRP: (a) basalt fiber sheets and (b) resin matrix.

Table 1. Physical and mechanical properties of basalt fibers and the resin.

Breaking strength Elongation at

Product Fiber diameter (μm) Linear density (tex) (N/tex) Elastic modulus (GPa) fracture

GBF 13 300 ≥0.6 ≥91 ≥3.1%

Product Curing time (23°C) Tensile strength (MPa) Flexural strength (MPa) Tensile shear strength (MPa) Compression
modulus (MPa)
L500A/B ≤18 hours ≥30 ≥40 ≥10 ≥1.0 × 103

Figure 2. Raw materials for the production of the SWCC: (a) sea-sands; (b) coral aggregates; and (c) particle gradation curves.

D7205-11 (2011), the ultimate tensile strength was mea- composition mass ratio is CaCl 2: KBr: MgCl 2: KCl:
sured to be 1177 MPa, the modulus of elasticity was Na2SO4: NaHCO3: NaCl = 1: 0.089: 4.483: 0.599:
46.9 GPa, and the ultimate tensile strain was 0.025. 3.526: 0.173: 21.147. As shown in Figure 2(a), the
Following the ASTM D2290-12 (2012), the “disk-split” adopted natural sea-sand with a fineness modulus of
method was used to test the hoop tensile properties of about 2.4 was mined in the coastal city named
BFRP tubes. The measured tensile strength of the BFRP Zhangzhou in Fujian Province, China. It belongs to the
tube was about 278.4 MP to 283.9 MPa, and the corre- medium sand class. As shown in Figure 2(b), the natural
sponding elastic modulus was 12.7 GPa to 12.8 GPa. coral aggregates mined in the South China Sea with a
particle size of 0.5 cm–2 cm were used. The tested
particle gradation curves of the sea-sands and the coral
Seawater sea-sand coral aggregate concrete aggregates are shown in Figure 2(c). In addition, or-
The artificial seawater was prepared according to the dinary silicate cement with a grade of 42.5 in the
Testing and Materials (2013), and the chemical Chinese standard was used.
1854 Advances in Structural Engineering 25(9)

Figure 3. Preparation technology of BFRP tubular arches: (a) wrapping the BFRP sheets and (b) foam arches with near-surface-mounted
BFRP bars.

Table 2. Test program.

Wall thickness of Pre-bonded with

Specimen no the BFRP tube (mm) Span (mm) Rise of the arch (mm) Radius angles BFRP bars or not

Arch-3-180-Y 3 2000 925 180o Yes

Arch-3-150-Y 3 2000 692 150o Yes
Arch-3-120-Y 3 2000 502 120o Yes
Arch-3-180-N 3 2000 925 180o No
Arch-3-150-N 3 2000 692 150o No
Arch-3-120-N 3 2000 502 120o No
Arch-2-150-Y 2 2000 692 150o Yes
Arch-4-150-Y 4 2000 692 150o Yes
Arch-2-150-N 2 2000 692 150o No
Arch-4-150-N 4 2000 692 150o No

The detailed mass mix ratio of the SWCC is Portland axis of the foam arch. The inner foam was removed after
cement: artificial seawater: sea-sands: coral aggregates: the resin was cured. Considering that the construction of a
water reducer = 1.00: 0.55: 1.50: 2.50: 0.012. The water curved BFRP reinforcement cage in situ is time-consuming
reducer used is a polycarboxylate superplasticizer, which and is not easy to remain curvilinear shape due to the linear
has a water reduction of 26%. The SWCC was prepared by elastic characteristic of FRP bars, grooves were cut on the
mechanical mixing in the laboratory. Cylindrical speci- surface of the curved foam arch in advance, and the BFRP
mens with 150 mm in diameter and 300 mm in height were bars were near-surface-mounted (NSM) into the grooves
made and the cylindrical compressive strength at 28 days and were temporary fixed as shown in Figure 3(b). It should
was measured to be 18.58 MPa. be noted that the curved BFRP bars are bent from straight
bars, which causes a certain amount of prestressing in the
bars themselves. According to equation (1) (Zhu et al.,
Experimental program
2017), the corresponding prestress of BFRP bars with
Production of curved BFRP tubes different rise of specimens (Table 2) were 198.4 MPa,
159.7 MPa and 114.5 MPa, respectively. The prestress
As shown in Figure 3(a), the curved BFRP tube was made accounted for 16.8%, 13.5% and 9.7% of the ultimate
by manually winding basalt fiber sheets on the surface of a strength (1177 MPa), respectively.
customized curved foam arch. The diameter of the foam
arch was 150 mm, and the resin-impregnated basalt fiber 2Ep rθ
σ¼ ± (1)
sheets were wrapped at an approximate angle of 67.5° to the πR
Dong et al. 1855

where σ is the additional stress caused by bending; Ep is the displacement of the arches at different locations during the
elastic modulus of FRP bars; θ is the bending angle of FRP loading process (i.e., the middle span and the quarter
bars; R is the radius of curvature; and r is the radius of FRP spans). Sixteen strain gauges were attached to each
bars. specimen to measure the axial and circumferential strain at
After that, the impregnated basalt fiber sheets were the top, side, and bottom of the arch. Since the top surface
wrapped following the aforementioned same process on the at the middle span is in contact with the loading steel block,
outer side of the foam. After the resin was cured and the the strain gauges are arranged only on the side and bottom
foam was removed, the new integrated BFRP tube arch of the arch. All measurements, including load, displace-
with longitudinal BFRP bars well pre-bonded was ment, and strain, were recorded with a TDS530 data ac-
obtained. quisition system. In addition, as shown in Figures 5(b) and
(a) non-contact full-field digital image correlation (DIC)
technique was used to monitor the strain field during the
Casting of the SWCC-filled BFRP tubular arches
test. The DIC method is used to obtain the displacement
As shown in Figure 4(a), the prepared curved BFRP field on the specimen surface by tracking the pixel motion
tubes were erected upside down on wooden supports. of the scattered spots on the specimen surface during the
Then, as shown in Figure 4(b), the prepared fresh test (Wang et al., 2015). The displacement field is then post-
SWCC was directly cast into the inverted BFRP tube processed to obtain the strain field.
from both ends and was compacted with a vibration rod
from the two ends. After 28 days of curing in the
laboratory, as shown in Figure 4(c), the CFFT arch was Test results
turned over and the foot foundations were cast with C50
Failure modes
concrete.
In this paper, the effects of three parameters, that is, the wall
thickness of the BFRP tube, the radius angles of the arch,
Test program and instrumentations and the containing of BFRP bars or not, were investigated.
The number of the prepared ten SWCC-filled BFRP tubular The failure modes under different variables are shown in
arches is shown in Table 2. Each specimen is named in the the following Figure 6.
following format: Arch-x-y-N(/Y), where “Arch” denotes Figure 6(a)–(c) shows the failure modes of the arches
the arch specimen; the number “x” denotes the nominal without BFRP bars under different wall thicknesses at the
wall thickness of the BFRP tube (in mm); the number “y” same radius angle (150°). In the process of loading, cracks
denotes specimens with pre-bonded BFRP bars. first appeared at the bottom of the loading point, and then
The loading device and the layout of the instrumenta- new cracks appeared at the outside of the left and right
tions are shown in Figure 5(a). The test was conducted on a quarter section in turn. As the loading continues, the cracks
servo actuator for monotonic loading. A load sensor was at these points became wider and wider. The bearing ca-
connected to the loading head of the actuator to accurately pacity of the specimen was rapidly reduced and the
and synchronously record the loads. The test was initially specimen was damaged when the crack width exceeds a
loaded at a rate of 1 kN/s, and when the load reached 10 kN, certain limit. Figure 6(d)–(f) shows the failure modes of the
it was changed to displacement control mode with a arches containing BFRP bars under different wall thick-
loading rate of 0.5 mm/min. Three linear variable differ- nesses at the same radius angle (150°). Compared to the
ential transformers (LVDTs) were used to measure the failure modes of the Arch-2(3/4)-150-N specimens as

Figure 4. Casting of the SWCC-filled BFRP tubular arches: (a) prepared curved BFRP tubes; (b) casting the SWCC; and (c) casting the
foot foundations.
1856 Advances in Structural Engineering 25(9)

Figure 5. Test setup and instrumentations of arch specimens: (a) schematic diagram and (b) photograph.

shown in Figure 6(a)–(c), the location of the first crack in loading stages are shown in Figure 7 and Figure 8. For the
the Arch-2(/3/4)-150-Y specimens shifted from the bottom Arch-3-120-N specimen, when loaded to 71.8 kN
of the mid-span to the edge of the loading steel block. This (Figure 7a), obvious tensile strains appeared at the bottom
is because the Arch-2(3/4)-150-N specimen has the largest part of the arch under the loading plate. As the load in-
bending moment in the middle span, so that is where the creased to 97.9 kN, new obvious concentrated tensile
cracks first appeared. While, for the Arch-2(/3/4)-150-Y strains appeared at the top part of the arch near the quarter
specimens containing BFRP bars, the bending moment can span (Figure 7b). When the load was further increased to
be well withstood by the BFRP bars so that the tube at the 108.5 kN, which was the ultimate load (Figure 7c), the
mid-span was not prematurely damaged. According to the tensile strains (which represents obvious cracks) at the
location of the initial cracks observed in the Arch-2(/3/4)- above two locations developed further. Then, as the loading
150-Y specimens, it was believed that the shear force continued, the bearing capacity of the arch gradually de-
generated at the edge of the steel loading plate was the main creased. When the load dropped to 102.7 kN (Figure 7d),
reason for the damage. some of the scattered spots began to be unrecognizable.
Figure 6(g)–(h) shows the failure modes of specimens Finally, the arch was destroyed with a sudden fracture at the
with the radius angle of 180°, the failure mode of Arch-3- mid-span and the quarter spans.
180-N was similar to Arch-2(3/4)-150-Y specimens. Al- For the Arch-3-120-Y specimen, as shown in Figure 8a,
though there were no longitudinal BFRP bars pre-bonded, when loaded to 104.7 kN, there was a significant strain
the larger radius angle led to a relatively lower tensile stress concentration at the edge of the loading steel plate, which
at the tube bottom of the mid-span section, causing the was caused by the large bending-shear force at this loca-
failure to be still controlled by the shear force generated at tion. As the load increased to 146.8 kN (Figure 8b), cracks
the edge of the loading steel plate. As to the Arch-3-180-Y appeared at the right edge of the loading plate, where the
specimen, it exhibited a similar failure mode to Arch-2(3/ strain increased significantly. When the load was reduced to
4)-150-Y specimens due to the existence of the internal 142.3 kN (Figure 8c), the arch section close to the edge of
longitudinal BFRP bars. the loading steel plate was significantly cracked and the
The strain contours of the Arch-3-120-N and Arch-3- strain contours could not be obtained. When the load was
120-Y specimens obtained by the DIC method at different further reduced to 139.4 kN (Figure 8d), the right span of
Dong et al. 1857

Figure 6. Failure modes of specimens: (a) Arch-2-150-N; (b) Arch-3-150-N; (c) Arch-4-150-N; (d) Arch-2-150-Y; (e) Arch-3-150-Y; (f)
Arch-4-150-Y; (g) Arch-3-180-N; and (h) Arch-3-180-Y.

Figure 7. Strain contours of the Arch-3-120-N specimen.

1858 Advances in Structural Engineering 25(9)

Figure 8. Strain contours of the Arch-3-120-Y specimen.

the arch was unrecognizable with the DIC method due to arch, and the rigid connection turned into hinge con-
the oversized cracks. nection at the crack location, and the load could still be
maintained and continued to increase by a small amount.
Finally, the arch fractured at the quarter of the span be-
Load vs. displacement curves cause the cracks developed too large and formed a huge
The load–displacement (P-Δ) curves of the arch specimens gap. In addition, it can be seen from Figure 9(a) that
are shown in Figure 9, Figure 10 and Figure 11. The ul- increasing the wall thickness of the BFRP tube could
timate load and corresponding ultimate mid-span dis- delay the initial damage and increase the ultimate bearing
placement of each specimen are shown in Table 3, where capacity of the specimen. Compared with the Arch-2-150-
the Pu represents the ultimate load and the Δu represents the N specimen, the ultimate bearing capacities of Arch-3-
ultimate mid-span displacement. The radial displacements 150-N and Arch-4-150-N specimens increased by 33.1%
at the left and right quarter span are defined as Δl and Δr, and 71.7%, respectively.
respectively. As can be seen from Figure 9(a) and Figure 9(b), the pre-
bonded longitudinal BFRP bars significantly increased the
Load vs. mid-span displacement curves. The effects of the ultimate bearing capacity of the arches. For the specimens
wall thickness of the BFRP tube and pre-bonded BFRP with a wall thickness of 2 mm, 3 mm, and 4 mm, the pre-
bars on the P-Δ curves are shown in Figure 9. As shown in bonded BFRP bars increased the ultimate bearing capacity
Figure 9(a), the P-Δ curves were approximately linear in by 215%, 248%, and 189%, respectively. Besides, as can
the early stage of loading. With the increase of load, the be seen from the curves in Figure 9(b), the P-Δ curves of
specimens entered the cracking damage stage, the tensile Arch-2-150-Y, Arch-3-150-Y, and Arch-4-150-Y were
stress at the bottom of the arch gradually increases, and nearly coincident at the early stage of loading, which in-
when the axial tensile stress in the BFRP tube exceeded its dicated that the wall thickness of the BFRP tube had little
ultimate tensile strength, it was ruptured, and the load was effect on the global rigidity of the specimens. The Arch-3-
reduced. With the development of cracks at the bottom of 150-Y and Arch-4-150-Y specimens nearly entered the
the mid-span arch and the outer side of the quarter span plastic stage at the same time, and their P-Δ curves were
Dong et al. 1859

Figure 9. P-Δ curves of specimens with different BFRP tube thicknesses: (a) Arch-2(3/4)-150-N and (b) Arch-2(3/4)-150-Y.

Figure 10. P-Δ curves of specimens with different radius angles: (a) Arch-3-150(120/180)-N and (b) Arch-3-150(120/180)-Y.

Figure 11. Representative load-radial displacement curves at quarter spans: (a) Arch-3-150-Y and (b) Arch-4-150-Y.

coinciding when the value of Δ was less than 10 mm. After had thicker wall thickness, could still increase. The load
that, the Arch-3-150-Y specimen approached the ultimate values of the Arch-3-150-Y and Arch-4-150-Y specimens
load, while the load of the Arch-4-150-Y specimen, which declined rapidly after reaching the peak point, which
1860 Advances in Structural Engineering 25(9)

Table 3. Ultimate load and ultimate displacement of specimens.

Specimen no Wall thickness of the BFRP tube (mm) Radius angles Pu (kN) Δu (mm)

Arch-2-150-Y 2 150° 128.8 19.17

Arch-3-150-Y 3 150° 189.3 13.31
Arch-4-150-Y 4 150° 203.2 22.09
Arch-3-120-Y 3 120° 151.3 16.05
Arch-3-180-Y 3 180° 79.2 10.48
Arch-2-150-N 2 150° 40.9 15.76
Arch-3-150-N 3 150° 54.4 36.76
Arch-4-150-N 4 150° 70.2 33.39
Arch-3-120-N 3 180° 108.5 9.75
Arch-3-180-N 3 120° 46.7 3.79

showed a brittle failure mode. While the curve of the Arch- damage in the later stage of the failure is more concentrated
2-150-Y specimen decreased relatively slowly after in a small area of the middle span. At the angle of 120°,
reaching the peak point. Combined with the failure mode as when the failure occurs, the triangular diagonal bracing
shown in Figure 6(a), it is believed that this was due to the effect of the arch cannot be exerted due to the small arc, and
premature damage at the left half of the span, resulting in a the failure is controlled by bending.
lower bearing capacity; meanwhile, the relatively low
damaged arch body on the right side provided a continuous Load vs. radial displacement curves at the quarter
bearing capacity in the later stage of loading. spans. Figure 11 shows the typical load vs. radial dis-
The effect of the radius angles on the P-Δ curve is shown placement curves at the quarter spans (the mid-span dis-
in Figure 10. As shown in Figure 10(a), for arches without placement is also given for comparison). Here, the Arch-3-
pre-bonded BFRP bars, the ultimate bearing capacities 150-Y and Arch-4-150-Y specimens are taken as examples.
decreased with the increase of the radius angle (i.e., the rise- As can be seen, in the whole test process, the displacements
to-span ratio). The change of the rise-to-span ratio resulted in at the two quarter points were all negative, and the two
a change in the stress distribution within the arch body. The curves were close to or nearly coincident before ap-
P-Δ curves of Arch-3120-N and Arch-3-180-N specimens proaching the ultimate load. This phenomenon indicated
did not show a decreasing and then increasing trend as that that the deformation of the Arch-3-150-Y and Arch-4-150-
of the Arch-3-150-N specimen. The Arch-3-120-N speci- Y specimens was symmetrical, and the stresses generated
men was fractured at the mid-span and the two quarter spans in the arch body were also symmetrical.
simultaneously. while the fracture section of Arch-3-180-N
was not centered (close to the edge of loading steel block),
which led to poor integrity of the specimen, and the bearing
Load vs. strain curves
capacity decreased rapidly after reaching the ultimate load, Load vs. longitudinal strain curves. Figure 12(a) shows the
showing obvious brittle characteristics. relationships between the load and longitudinal strain of the
It can be seen from the comparison of Figure10(a) and top and bottom surfaces of the representative Arch-3-150-
Figure 10(b) that the enhancing effect of adding BFRP bars Y specimen. The strain increased linearly with the load at
on bearing capacity was related to the radius angle. For the initial loading stage. The arch bottom at the quarter
example, after pre-bonded with BFRP bars, the ultimate points was compressed, and the bottom of the arch at the
bearing capacity of the specimen with a 150° radius angle mid-span and the top of the arch at the quarter points were
was increased by 248%, while the ultimate bearing ca- tensioned. Among them, the longitudinal strain of the
pacities of the specimens with a 120° and a 180° radius bottom of the arch at the mid-span had the fastest in-
angle were only increased by 39% and 70%, respectively. It creasing rate. With the cracking of the BFRP tube at the
can be seen that the reinforcement form proposed in this bottom of the mid-span, the strain gauge at this location
paper has the best enhancement effect on the specimens was failed. When the bottom of the arch in the mid-span
with a 150° radius angle. The internal force distribution in was cracked and destroyed, the internal stress in the arch
the arch body will change due to formation of concentrated body was continuously redistributed. The strain growth
cracks, and the bending moment distribution of the arch rate on the tension side and the compression side at the
body is relatively more uniform at an angle of 150°. At the quadrant point gradually accelerated, and the growth rate
angle of 180° because the main cracks at the arch shoulder of the strain on the tension side was greater than that of the
are too close to the main crack in the middle span, the compression side. Figure 12(b) shows the relationships
Dong et al. 1861

Figure 12. Longitudinal strain of the Arch-3-150-Y specimen: (a) top and bottom surfaces and (b) central axis.

Figure 13. Longitudinal strain of the Arch-4-150-Y specimen: (a) top and bottom surfaces and (b) central axis.

between the load and longitudinal strain of the arch central representative specimens Arch-3-150-Yand Arch-4-150-Y.
axis of the representative specimen Arch-3-150-Y. It can Comparing Figure 14 and the above Figure 12, it can be
be seen that the increasing rate of longitudinal strain in the seen that the circumferential strains of the Arch-3-150-Y
mid-span was significantly higher than that in the quarter specimen at different positions were all smaller than the
points. longitudinal strains. This is because, in this test, the loading
The test results of the surface strains of another repre- method was single-point concentrated loading. The arch
sentative specimen Arch-4-150-Y are shown in Figure 13. It was subjected to compression and bending during the test,
can be seen that the law of the curves was similar to that of and the bending moment had a greater impact on the de-
the Arch-3-150-Y specimen, which has a slightly thinner formation and failure of the arch. The good mechanical
wall thickness. The difference was that the increasing rate of properties in the hoop direction had not been fully utilized.
the longitudinal strain of Arch-4-150-Y at the bottom of the In Figure 15, because of the increase in the wall thickness
mid-span was not significantly greater than that at top of the of the BFRP tube, the hoop strains of the Arch-4-150-Y
arch body of the quarter spans, which means that increasing specimen were lower than those of the Arch-3-150-Y
the wall thickness reduced the strain development rate. specimen under the same load levels. This phenomenon
Figure 13(b) shows the longitudinal strain of the Arch-4- showed that increasing the wall thickness of BFRP tube
150-Y specimen along the arch central axis. It can be seen could also suppress the increase of hoop strains.
that the strain at the mid-span position was greater than that
at the quarter points. Compared with the Arch-3-150-Y
specimen, the strain values at the mid-span position under
Change of the arch shape
the same load were also significantly reduced. Taking Arch-3-120-N and Arch-3-120-Y specimens as
examples, the change of arch shape with load is introduced
Load vs. circumferential strain curves. Figure 14 and based on the data tested by the DIC technology. As shown
Figure 15 show the load vs. circumferential strain curves of in Figure 16, 14 points on the arch axis were selected to
1862 Advances in Structural Engineering 25(9)

Figure 14. Circumferential strain of the Arch-3-150-Y specimen: (a) top and bottom surfaces and (b) central axis.

Figure 15. Circumferential strain of the Arch-4-150-Y specimen: (a) top and bottom surfaces and (b) central axis.

Figure 16. Change of the arch shape of representative specimens: (a) Arch-3-120-N and (b) Arch-3-120-Y.

draw the arch shape. To more clearly distinguish the arch displacements were significantly increased. After the
shape under different loads, the displacement values in the specimen reached 98.5 kN, the BFRP tube was cracked,
figures were magnified by 30 times. It can be seen from and the displacement developed rapidly within the range of
Figure 16(a) that the displacement of the Arch-3-120-N ±750 mm in the middle of the span. Figure 16(b) shows the
specimen remained basically unchanged at the initial stage displacement values of the Arch-3-120-Y specimen. The
of loading. When the load was less than 49.6 kN, only a change rule was similar to that of the Arch-3-120-N
partial region close to the loading head occurred slight specimen at the initial loading stage. However, due to
vertical displacements. When the load reached 74.3 kN, the the existence of the pre-bonded BFRP bars, the mid-span
Dong et al. 1863

displacement of the Arch-3-120-Y specimen when it en- characteristics. For specimens with radius angles of
tered the plastic phase was increased by 71.3%. In addition, 120° and 180°, the additional BFRP bars only
the displacements of the Arch-3-120-N specimen devel- increased the ultimate bearing capacity by 39% and
oped symmetrically on the left and right sides of the arch 70%, while for specimens with a radius angle of
before entering the plastic stage, while the Arch-3-120-Y 150°, the improvement was 248%, which was most
had a clear tendency to lean to one side. Due to the ex- obvious.
istence of the pre-bonded FRP bars, the main cracks at the
mid-span appeared randomly at one side of the two edges In addition, it should be noted that the preparation
of the loading steel block, resulting in the inconsistent process of BFRP tubular arches with BFRP longitudinal
failure time at the left and right quadratic points of the bars pre-bonded in this paper can be further optimized. In
specimen. This is the reason why the displacements of the subsequent industrial production, a customized steel
specimens with FRP bars developed unsymmetrically. inner formwork with grooves can be used instead, which
will be more efficient.
Conclusions
Declaration of conflicting interests
In this paper, a new type of SWCC-filled BFRP tubular
The author(s) declared no potential conflicts of interest with re-
arches that is suitable for the marine environment was
spect to the research, authorship, and/orpublication of this article.
proposed. Ten arch specimens with different BFRP tube
wall thicknesses (2 mm, 3 mm, and 4 mm), different radius
angles (120°, 150°, and 180°), and with or without pre- Funding
bonded longitudinal BFRP bars) are tested by monotonic The author(s) disclosed receipt of the following financial support
loading. Based on the research of this paper, the following for the research, authorship, and/or publication of this article: This
main conclusions are drawn: work was supported by the financial support from the Natural
Science Foundation of Jiangsu Province (BK20190369 and
(1) For the arches with a radius angle of 150° and BK20191146), the National Natural Science Foundation of China
without pre-bonded BFRP bars inside, the failure (Grant No:51908118, 51838004, and 52078127), and the
modes were the fracture of the arch body. After the “Zhishan” Scholars Programs of Southeast University.
first cracking in the middle of the span, the bearing
capacity decreased slightly. Subsequently, cracks ORCID iDs
of varying degrees appeared at the middle and the
Ziqing Liu  https://orcid.org/0000-0003-2054-987X
left and right quadratic points of specimens, and the
Gang Wu  https://orcid.org/0000-0001-9040-7303
rigid connection turned into hinge connection at
Hong Zhu  https://orcid.org/0000-0002-4870-4336
the cracking section, so the bearing capacity in-
creased slightly. In addition, as the wall thickness
of the BFRP tube increased from 2 mm to 3mm References
and 4mm, the ultimate bearing capacity increased Afifi MZ, Mohamed HM and Benmokrane B (2014) Axial Ca-
by 33.1% and 71.7%. pacity of Circular Concrete Columns Reinforced with GFRP
(2) For the arches containing pre-bonded longitudinal Bars and Spirals. Journal of Composites for Construction,
BFRP bars, the first cracking of the specimens was 18(1): 04013017.
transferred from the mid-span section to the section ASTM D7205-11 (2011) Standard Test Method for Tensile
close to the edge of the loading steel block due to Properties of Fiber Reinforced Polymer Matrix Composite
the combined bending-shear force. The existence Bars. West Conshohocken, PA: American Society for
of BFRP bars significantly improved the ultimate Testing and Materials.
bearing capacities of the arches. For the specimens ASTM D2290-12(2012) Standard Test Method for Apparent
with a radius angle of 150° and a wall thickness of Hoop Tensile Strength of Plastic or Reinforced Plastic Pipe.
2 mm, 3 mm, and 4 mm, the improvements West Conshohocken, PA: American Society for Testing and
were 215%, 248%, and 189%, respectively. Materials.
(3) As the radius angle of the arches increased, the Bazli M, Li YL, Zhao XL, et al. (2020) Durability of seawater and sea
ultimate bearing capacity gradually decreased. Due sand concrete filled filament wound FRP tubes under seawater
to the excessively large radius angle of the Arch-3- environments. Composites Part B-Engineering, 202: 108409.
180-N specimen, the failure mode changed from Cheng S, Shui Z, Sun T, et al. (2018) Durability and micro-
bending failure to bending-shear coupled failure, structure of coral sand concrete incorporating supplementary
and the ultimate load-bearing capacity decreased cementitious materials. Construction and Building Mate-
significantly and showed obvious brittleness rials, 171: 44–53.
1864 Advances in Structural Engineering 25(9)

Cromwell JR, Harries KA and Shahrooz BM (2011) Environ- Meng XN and Wang HH (2018) Research status of new CFFT
mental durability of externally bonded FRP materials in- arch bridge and its problem analysis. China Metal Bulletin,
tended for repair of concrete structures. Construction and 07: 291–292.
Building Materials, 25: 2528–2539. Niu D, Su L, Luo Y, et al. (2020) Experimental study on me-
Da B, Yu H, Ma H, et al. (2016) Factors Influencing Durability of chanical properties and durability of basalt fiber reinforced
Coral Concrete Structure in the South China Sea. Journal of coral aggregate concrete. Construction and Building Ma-
the Chinese Ceramic Society, 44: 253–260. terials, 237: 117628.
Dagher HJ, Bannon DJ, Davids WG, et al. (2012) Bending be- Ozbakkaloglu T, Lim JC and Vincent T (2013) FRP-confined
havior of concrete-filled tubular FRP arches for bridge concrete in circular sections: review and assessment of
structures. Construction and Building Materials, 37: stress-strain models. Engineering Structures, 49: 1068–1088.
432–439. Pantelides CP, Gibbons ME and Reaveley LD (2013) Axial load
De luca A, Matta F and Nanni A (2010) Behavior of full-scale behavior of concrete columns confined with GFRP spirals.
glass fiber-reinforced polymer reinforced concrete columns Journal of Composites for Construction, 17, 305–313.
under axial load. ACI Structural Journal, 107: 589–596. Su L, Niu D and Luo D (2018) Research of coral aggregate
De waal L, Jiang S and Torres J (2018) Design and construction of concrete on mechanical property and durability. Materials
a hybrid double-skin tubular arch bridge. In: 9th Interna- Review, 32: 3387–3393.
tional Conference on Fibre-Reinforced Polymer (FRP) Teng JG, Zhang B, Zhang S, et al. (2018) Steel-free hybrid re-
Composites in Civil Engineering, CICE 2018, Paris, France, inforcing bars for concrete structures. Advances in Structural
17-19 July 2018. Engineering, 21: 2617–2621.
Dong ZQ, Sun Y, Wu G, et al. (2021) Flexural behavior of Testing ASF, Materials (2013) Standard Practice for the Prep-
seawater sea-sand concrete beams reinforced with BFRP aration of Substitute Ocean Water, West Conshohocken, PA:
bars/grids and BFRP-wrapped steel tubes. Composite ASTM International.
Structures, 268, 113956. Tobbi H, Farghaly AS and Benmokrane B (2014) Behavior of
Dong ZQ, Wu G, Zhao XL, et al. (2020a) Mechanical properties concentrically loaded fiber-reinforced polymer reinforced
of discrete BFRP needles reinforced seawater sea-sand concrete columns with varying reinforcement types and
concrete-filled GFRP tubular stub columns. Construction ratios. ACI Structural Journal, 111, 375–385.
and Building Materials, 244: 118330. Tomblin J (2006) Buried FRP-Concrete Arches. Princeton, NJ:
Dong ZQ, Wu G, Zhu H, et al. (2020b) Flexural behavior of Citeseer.
seawater sea-sand coral concrete-UHPC composite beams Walton HJ, Davids WG, Landon ME, et al. (2016) Simulation of
reinforced with BFRP bars. Construction and Building buried arch bridge response to backfilling and live loading.
Materials, 265, 120279. Journal of Bridge Engineering, 21(9): 04016052.
El-hacha R, Green MF and Wight GR (2010) Effect of Severe Wang G, Wu Q, Zhou H, et al. (2021) Diffusion of chloride ion in
Environmental Exposures on CFRP Wrapped Concrete coral aggregate seawater concrete under marine environ-
Columns. Journal of Composites for Construction, 14: ment. Construction and Building Materials, 284: 122821.
83–93. Wang HH and Chen HL (2018) Preparation and blast resistance
Gao Y, Zhou Y, Zhou J, et al. (2020) Blast responses of one-way test of carbon fiber reinforced composite pipe concrete arch.
sea-sand seawater concrete slabs reinforced with BFRP bars. Journal of Arms and Equipment Engineering, 39(10):
Construction and Building Materials, 232: 117254. 149–154.
Hadi MNS, Khan QS and Sheikh MN (2016) Axial and flexural Wang H, Chen H, Zhou Y, et al. (2019) Blast responses and
behavior of unreinforced and FRP bar reinforced circular damage evaluation of CFRP tubular arches. Construction
concrete filled FRP tube columns. Construction and and Building Materials, 196: 233–244.
Building Materials, 122: 43–53. Wang J, Feng P, Hao T, et al. 2017. Axial compressive behavior of
Hadi MNS, Wang W and Sheikh MN (2015) Axial compressive seawater coral aggregate concrete-filled FRP tubes. Con-
behavior of GFRP tube reinforced concrete columns. Con- struction and Building Materials, 147: 272–285.
struction and Building Materials, 81, 198–207. Wang W, Rans C, Alderliesten RC, et al. (2015) Predicting the
Li Y, Zhou L, Zhang Y, et al. (2013) Study on long-term per- influence of discretely notched layers on fatigue crack
formance of concrete based on seawater, sea sand and coral growth in fiber metal laminates. Engineering Fracture
sand. In: 3rd International Conference on Mechatronics and Mechanics, 145: 1–14.
Intelligent Materials (MIM 2013), Xi Shuang Ban Na, May Wang Y, Chen G, Wan B, et al. (2020) Behavior of circular ice-
18-19 2013, PEOPLES R CHINA. filled self-luminous FRP tubular stub columns under axial
Lyu B, Wang A, Zhang Z, et al. (2019) Coral aggregate concrete: compression. Construction and Building Materials, 232:
numerical description of physical, chemical and morpho- 117287.
logical properties of coral aggregate. Cement & Concrete Wei Y, Bai J, ZhanG Y, et al. (2021) Compressive performance of
Composites, 100: 25–34. high-strength seawater and sea sand concrete-filled circular
Dong et al. 1865

FRP-steel composite tube columns. Engineering Structures, Yu T., Zhang S., Huang L., et al. (2017). Compressive behavior of
240: 112357. hybrid double-skin tubular columns with a large rupture
Wen Y, Teng JG, Yu T, et al. (2012) Nonlinear Finite Element strain FRP tube. Composite Structures, 171: 10–18.
Modeling of Concrete-filled Circular FRP Arch Tubes under Zeng JJ, Gao WY, Duan ZJ, et al. (2020) Axial compressive
Monotonic Loading. 2nd International Conference on behavior of polyethylene terephthalate/carbon FRP-confined
Structures and Building Materials (ICSBM), 1 November seawater sea-sand concrete in circular columns. Construc-
2019, Eindhoven, Netherlands, Advanced Materials tion and Building Materials, 234: 117383.
Research, Zhang Y, Wei Y, Bai J, et al. (2020). A novel seawater and sea sand
Xiao J, Qiang C, Nanni A, et al. (2017) Use of sea-sand and concrete filled FRP-carbon steel composite tube column:
seawater in concrete construction: current status and future concept and behavior. Composite Structures, 246: 112421.
opportunities. Construction and Building Materials, 155: Zhou W, Feng P and Yang JQ (2021) Advances in coral aggregate
1101–1111. concrete and its combination with FRP: a state-of-the-art
Xiong Z, Zeng Y, Li L, et al. (2021) Experimental study on the review. Advances in Structural Engineering, 24: 1161–1181.
effects of glass fibers and expansive agent on the bond be- Zhu H, Dong ZQ, WU G, et al. (2017) Experimental evaluation of
havior of glass/basalt FRP bars in seawater sea-sand concrete. bent FRP tendons for strengthening by external prestressing.
Construction and Building Materials, 274: 122100. Journal of Composites for Construction, 21: 04017032.