Professional Documents
Culture Documents
Composite Structures
journal homepage: www.elsevier.com/locate/compstruct
A R T I C L E I N F O A B S T R A C T
Keywords: The application of fiber‐reinforced‐polymer (FRP) bars to reinforce concrete structures could mitigate the
BFRP bars corrosion‐induced damage of steel reinforcements. No study has been reported in open literature on flexure‐
Concrete beams critical or shear‐critical concrete beams reinforced with Basalt FRP (BFRP) bars under impact loads. In this
Flexure‐critical study, six BFRP bars reinforced concrete beams were tested under quasi‐static and impact loads. The test results
Shear‐critical
showed the flexure‐critical beams experienced the failure mode changing from flexure‐governed under quasi‐
Impact loads
Numerical simulation
static loads to flexure‐shear combined under impact loads. The shear‐critical beams still failed in diagonal shear
under impact loads, but experienced severer concrete spalling and more critical diagonal cracks on both sides
of the beams. The impact performance of concrete beams with higher strength concrete may not be necessarily
superior to that of beams with normal strength concrete due to the increased brittleness. Moreover, a numerical
model of the tested beams under impact loads was developed and calibrated in LS‐DYNA. Numerical results
showed increasing tension reinforcement ratio could change the failure mode from flexure‐governed to
flexure‐shear combined along with the reduced maximum midspan deflection. The BFRP bars reinforced con-
crete beams had comparable impact resistant performance with the conventional steel reinforced concrete
beams
⇑ Corresponding authors.
E-mail addresses: wensu.chen@curtin.edu.au (W. Chen), hong.hao@curtin.edu.au (H. Hao).
https://doi.org/10.1016/j.compstruct.2021.113648
Received 4 November 2020; Revised 4 November 2020; Accepted 21 January 2021
Available online 29 January 2021
0263-8223/© 2021 Elsevier Ltd. All rights reserved.
Z. Huang et al. Composite Structures 263 (2021) 113648
Fig. 1. Critical cracks and failure modes of (a) flexure-critical beams [16,28] and (b) shear-critical beams [20,30,31] with increasing impact velocity.
impact velocity increased, the beams were prone to fail in flexure‐ In open literature, very limited studies investigated the perfor-
shear combined or shear‐governed manner. Under high impact veloc- mance of flexure‐critical concrete beams reinforced with Glass FRP
ity, the beams failed in a punching shear‐governed mode with shear (GFRP) bars under impact loading [33–37]. The beams exhibited an
plug at an angle of approximately 45° initiated from the impact point average 15% higher moment capacity under impact loads (with the
to the bottom of beams. For shear‐critical beams, they failed in shear drop mass 110 kg and height 1.2 m) than that under static loads
under static loads and impact loads with low impact velocity. The [33]. Reducing stirrup spacing could mitigate the damage level of
beams were prone to fail in combined shear‐punching shear mode with the beams with lower residual midspan deflection and higher post‐
the increased impact velocity. Punching shear damage was evident impact residual load‐carrying capacity [36,37]. It is worth noting that
under the impact force profile with high magnitude and short duration the performance of concrete beams reinforced with FRP bars under
induced by high impact velocity [21,24]. FRP reinforcements and steel impact loads is influenced by many factors, e.g., concrete material
reinforcements have different mechanical properties, e.g., FRP rein- properties, types of FRP bars, longitudinal and transverse reinforce-
forcements exhibit linear‐elastic and brittle stress–strain behaviour, ment ratios (related to types of beams, e.g., flexure‐critical and
and have high tensile strength but relatively low strength under com- shear‐critical beams), impact velocity (or drop height), strain rate
pression and shear [32], whereas steel reinforcements behave in an effect, drop weight, shape of drop hammer, and boundary condition.
elastic‐plastic and ductile manner, and have the same strength in ten- Although a few studies have been carried out to investigate the impact
sion and compression. Therefore, the performance of FRP bars rein- performance of flexure‐critical GFRP bars reinforced concrete beams,
forced concrete beams and conventional steel bars reinforced no study on the impact performance of shear‐critical FRP bars rein-
concrete beams subjected to impact loads might be different from each forced concrete beams can be found yet. In addition, neither flexure‐
other, which is worthy of studying. critical nor shear‐critical concrete beams reinforced with Basalt FRP
2
Z. Huang et al. Composite Structures 263 (2021) 113648
(BFRP) bars under impact loads have been reported in open literature. beams, based on their flexural‐shear capacity ratios. For each group,
Since BFRP bars had higher tensile strength and modulus of elasticity a beam was tested under quasi‐static loads as reference and two beams
than GFRP bars [38,39], the beams reinforced with BFRP bars and were prepared for impact tests. The width (b), height (h), and total
GFPR bars may behave differently under impact loads. Therefore, fur- length (Lt) of the beams were respectively 150 mm, 200 mm, and
ther study on the impact performance of flexure‐critical and shear‐ 1250 mm, as shown in Fig. 3. The diameters of both longitudinal BFRP
critical concrete beams reinforced with BFRP bars is deemed essential. bars and stirrups (100‐mm spacing) in flexure‐critical beams were
It should be noted that the investigation presented in this study is a 10 mm while the diameters of longitudinal BFRP reinforcements and
part of a project which focuses on the structural performance of nor- stirrups (100‐mm spacing) in shear‐critical beams were 16 mm and
mal concrete (Ordinary Portland cement Concrete, also called OPC) 4 mm, respectively. ACI 440.1R‐15 [14] was followed for the beam
vs geopolymer concrete (GPC) beams reinforced with BFRP bars. design in this study. Table 1 lists the detailed information of the tested
The studies on flexural and shear behaviour of ambient cured GPC beams. For better understanding, the terminology of the beams con-
beams reinforced with BFRP bars under static and impact loads sists of four components: the first one is the concrete type (OPC); the
[40,41] have been carried out recently. Since GPC and OPC behave second one represents loading condition (‘S’ means quasi‐static loads
differently, i.e., GPC is more brittle than OPC [42,43], it is worth and ‘I’ denotes impact loads); the third one with the letters of ‘FL’
studying and reporting the behaviour of flexure‐critical and shear‐ and ‘SH’ represents the types of the beams, namely flexure‐critical
0
critical OPC beams reinforced with BFRP bars under static and impact and shear‐critical; the last one is the compressive strength f c of con-
loads and comparing the performance between GPC beams and OPC crete. For example, Beam OPC‐I‐FL‐44 denotes a flexure‐critical OPC
beams. beam (concrete strength 44 MPa) under impact loads, which was also
In this study, a total of six BFRP bars reinforced concrete beams used for a reference beam to compare the impact performance among
were prepared. One flexure‐critical beam and one shear‐critical beam the GPC beams, OPC beams, and fibre reinforced GPC beams in the
as reference beams were tested under quasi‐static loads whilst four previous study [45].
beams with varying concrete strength were tested under drop‐
hammer impact. The effect of compressive strength of concrete on 2.3. Test setup
the impact performance of the beams was experimentally investigated.
The quasi‐static and impact responses of the beams were recorded and 2.3.1. Quasi-static test
analysed. Moreover, a numerical model was developed in LS‐DYNA The quasi‐static three‐point bending test setup is shown in Fig. 4.
and the numerical model was calibrated based on the impact test The beam was simply supported by a pin and a roller, and the clear
results. The influence of tension reinforcement ratio and reinforce- span (L) of the beam was 1,100 mm. The load was exerted onto the
ments material on the impact performance of the beams was further midspan of the beams through a hydraulic jack (the loading rate
numerically investigated. around 3 mm/min). The applied load and the midspan deflection were
captured by a load cell and linear variable differential transformers
2. Experimental details (LVDTs), respectively. Two strain gauges (SGs) in the flexure‐critical
beams, i.e. top strain gauge (TSG) and bottom strain gauge (BSG),
2.1. Materials and four strain gauges in the shear‐critical beams, i.e. longitudinal‐
bar strain gauges (LSG1 and LSG2) and stirrup strain gauge (SSG1
In this study, commercial concrete with the compressive strength of and SSG2), were bonded onto the reinforcements to capture their
67 MPa was used for four beams. To study the effect of the compres- strain values as shown in Fig. 3. The static results in terms of failure
sive strength of concrete on the impact performance of the beams, mode, peak load, load‐midspan deflection curve, and load‐strain curve
other two beams were cast separately with the compressive strength of reinforcements were analysed.
of 44 MPa and 52 MPa. The BFRP bars used in this study are shown
in Fig. 2. The specific properties of BFRP bars provided by the manu- 2.3.2. Impact test
facturer [44] are given below: the ultimate tensile strength ffu is Fig. 5 shows the drop hammer impact test setup. The beams were
1200 MPa; the elastic modulus Ef is 55 GPa; the break strain εfu is 2%. supported by using two steel plates on the supports and the clear span
(L) of the beam was kept as 1,100 mm. The impact force was recorded
2.2. Details of beam specimens by a load cell. The load cell was connected with an adaptor and placed
on the midspan of the beam. Four strain gauges (i.e., TSG, BSG, SSG1,
A total of six beams were prepared and cast, which were divided and SSG2 in the flexure‐critical beams, LSG1, LSG2, SSG1, and SSG2 in
into two groups, namely, flexure‐critical beams and shear‐critical the shear‐critical beams) were attached to the BFRP bars as shown in
3
Z. Huang et al. Composite Structures 263 (2021) 113648
Table 1
Details of the tested beams.
Beam group Beam Loading Concrete strength Tension reinforcement Stirrup ratio Designed flexural Designed shear Designed flexural-shear
name condition 0
f c (MPa) ratio ρf (%) ρf v (%) capacity (kN) [14] capacity (kN) [14] capacity ratio
fps. The data of impact force and reinforcement strain was collected
by a data acquisition system with the sampling rate of 50 kHz.
4
Z. Huang et al. Composite Structures 263 (2021) 113648
5
Z. Huang et al. Composite Structures 263 (2021) 113648
6
Z. Huang et al. Composite Structures 263 (2021) 113648
well as wider distribution area of cracks, as compared to Beam OPC‐S‐ on the beam from 2 ms to 3 ms. At 5 ms, very wide longitudinal cracks
SH‐67 under quasi‐static loads. were noticed, which led to subsequent concrete spalling at the bottom.
Fig. 11 presents the progressive failure of the tested beams. For As can be seen at 7 ms, the cracks extended to the compression zone
Beam OPC‐I‐FL‐44, two flexural cracks and a very short longitudinal underneath the load cell adaptor and the concrete subsequently started
crack at the bottom appeared at the instance of 1 ms. At 2 ms, another crushing for both Beams OPC‐I‐FL‐44 and OPC‐I‐FL‐67 due to the glo-
flexural crack near the existing left one and two new flexure‐shear bal bending deflection of the beams, which was about 24 mm as shown
cracks close to the supports were observed. These cracks then further in Fig. 10. For Beam OPC‐I‐SH‐52, a shear crack, two flexural cracks,
extended and became wider and nearly no new crack appeared after and some short longitudinal cracks were observed at 1 ms. More
7 ms. For Beam OPC‐I‐FL‐67, two flexural cracks were observed at flexure‐shear cracks appeared on the beam at 2 ms and there was no
1 ms. Some flexure‐shear, shear, and longitudinal cracks developed obvious change of the main crack pattern of the beam after 3 ms.
7
Z. Huang et al. Composite Structures 263 (2021) 113648
8
Z. Huang et al. Composite Structures 263 (2021) 113648
Fig. 11. Failure progress of the tested beams under impact loads.
some indices [36,57] such as the ratio of the post‐impact residual 5. Numerical simulations
load‐carrying capacity to the load‐carrying capacity of the beams
without impact testing. Since Beams OPC‐I‐FL‐67, OPC‐I‐SH‐52, and 5.1. Numerical model
OPC‐I‐SH‐67 experienced very severe damage with relatively large
residual deflection under impact loads, only Beam OPC‐I‐FL‐44 was In this section, the numerical simulations were conducted by using
further tested under monotonic quasi‐static loads to determine the commercial software LS‐DYNA [58], which has been widely used to
post‐impact residual load‐carrying capacity. The failure modes of simulate RC structures under impact or blast loads. It shows the
Beam OPC‐I‐FL‐44 before and after the residual load‐carrying capacity numerical simulations could achieve reliable predictions [24,59].
test are shown in Fig. 14. It can be seen that the beam experienced The test results of Beam OPC‐I‐FL‐44 were utilized for the calibration
more concrete crushing and spalling after the residual load‐carrying against the numerical results since this beam had more complete data
capacity test. Fig. 15 displays the post‐impact load‐midspan deflection than other three beams. The developed numerical model is shown in
curve of the beam. The beam had a post‐impact residual capacity of Fig. 16. Eight‐node solid elements were utilized for all parts except
84.1 kN and experienced a brittle failure with the applied load the BFRP reinforcements. The longitudinal BFRP reinforcements and
decreased suddenly after the applied load reached the residual load‐ stirrups were modelled using Hughes‐Liu beam elements with cross
carrying capacity. section integration. The supports were simplified without considering
9
Z. Huang et al. Composite Structures 263 (2021) 113648
Fig. 11 (continued)
Table 3
Typical impact testing results.
Beam Impact velocity (m/s) Peak impact force (kN) Peak deflection (mm) Residual deflection (mm) Residual load-carrying capacity (kN)
Note: ‘*’: data lost due to malfunction of the data acquisition system; ‘-’: not tested due to severe damage of the beams after the impact tests.
10
Z. Huang et al. Composite Structures 263 (2021) 113648
Fig. 13. (a) Reinforcements strain time histories and (b) schematic diagram of neutral axis at midspan section of Beam OPC-I-FL-44 at various instants.
threaded rods between the steel plates so that the boundaries were ance the efficiency and accuracy. Reinforcements were embedded into
simulated by constraining the outer plates and the steel rollers in all concrete by using the keyword *Constrained_Beam_in_Solid. The con-
directions. After conducting mesh sensitivity analysis, concrete beams tacts among the drop hammer, load cell cap, load cell, load cell adap-
and reinforcements were simulated by using 7.5 mm‐size elements tor, concrete, steel plates, and steel rollers were modelled by using the
while other parts were modelled by using 10 mm‐size elements to bal- keyword *Automatic_Surface_to_Surface, while the keyword
11
Z. Huang et al. Composite Structures 263 (2021) 113648
Table 4
Material models and related parameters used in the simulation.
12
Z. Huang et al. Composite Structures 263 (2021) 113648
parts were simulated by using *Mat_Piecewise_Linear_Plasticity equivalent Young’s modulus of the modelled load cell (i.e., 150 GPa)
(*Mat_024). The parameters of material models are listed in Table 4. was also adopted as 75% of the actual modulus of steel material.
The load cell was simplified into solid mass due to the complex inter- The strain rate effect was considered in the present study. The
nal structures, and its equivalent density (actual mass divided by the strength increment with strain rate could be specified by defining
model volume in simulation, about 5850 kg/m3) used for the simula- dynamic increase factor (DIF) for both material models *Mat_72R3
tion is about 75% of the density of steel material. For simplicity, the and *Mat_024. The DIFs of concrete, BFRP composites, and steel
13
Z. Huang et al. Composite Structures 263 (2021) 113648
Fig. 17 and Fig. 18 compares the numerical results with the test
results regarding failure progress and impact responses of Beam
OPC‐I‐FL‐44, respectively. The numerical results are in good agree-
ment with the test results in general. As shown in Fig. 17, the effective
plastic strain contour from the numerical simulation can reflect the
crack patterns of the tested beam, including the local damage such
as concrete crushing and spalling. The comparisons of impact force,
midspan deflection, and the strain of TSG are shown in Fig. 18. The
contact force of the interface between the load cell cap and the load
cell extracted from numerical results can match well with the impact
force time history of the test results as shown in Fig. 18(a). Numerical
simulation gives very close prediction of peak impact force to the test
results (413.3 kN vs 418.0 kN). Besides, the maximum and residual
displacements (29.6 mm and 5.0 mm, respectively) from the numeri-
cal simulation agree well with those (30.0 mm and 8.9 mm, respec-
tively) from the test results. It should be noted that the numerical
simulation could well capture the trend (i.e. tension and compression)
of strain time history of TSG. However, the numerical simulation over
predicted TSG strain although similar displacement was predicted,
which might be due to the element erosion. In the testing, the concrete
near the top longitudinal BFRP bars experienced severe crushing dam-
age but could still sustain load. However, in the numerical simulation,
once reaching the defined erosion criteria the concrete elements near
the top longitudinal reinforcements were eroded and could not resist
load, resulting in a higher strain in top reinforcements as compared
to the testing results. Overall, the time histories of impact force, mid-
span deflection, and the trend of strain time history of TSG were rea-
sonably predicted by the numerical simulation.
6. Parametric study
14
Z. Huang et al. Composite Structures 263 (2021) 113648
Fig. 19. Failure mode comparison of the beams with varying tension reinforcement ratios under impact loads.
respect to the midspan deflection, axial stress and axial strain of ten-
sion reinforcements at midspan. Beam B‐Steel exhibited slightly larger
maximum midspan deflection (31.1 mm vs 29.6 mm) and much larger
residual midspan deflection (27.8 mm vs 5.0 mm) than Beam B‐BFRP
as shown in Fig. 22(a). The maximum tensile stress (i.e. 1096 MPa)
in the tension BFRP bars of Beam B‐BFRP as shown in Fig. 22(b) did
not even reach the static tensile strength of BFRP bars (i.e.
1200 MPa). Due to the nature of BFRP rebar which is a linear elastic
material, it could recover to its original state, leading to a very small
residual deflection of the beam, e.g. 5.0 mm as shown in Fig. 22(a).
However, the tensile stress in tension steel bars of Beam B‐Steel
reached a relatively constant value of about 660 MPa (induced by
Fig. 20. Comparison of midspan displacement time histories of the beams strain rate effect with a DIF 1.3, i.e., 660 MPa/500 MPa) from 1 ms
with varying tension reinforcement ratios under impact loads. to 15 ms as shown in Fig. 22(b), while their strain increased from
Fig. 21. Comparison of failure modes between conventional RC beams and concrete beams reinforced with BFRP bars.
15
Z. Huang et al. Composite Structures 263 (2021) 113648
7. Conclusion
Fig. 22. Dynamic responses of conventional RC beams and concrete beams Zhijie Huang: Formal analysis, Investigation, Data curation, Writ-
reinforced with BFRP bars: (a) midspan deflection, (b) axial stress of tension ing ‐ original draft. Wensu Chen: Conceptualization, Methodology,
rebars, and (c) axial strain of tension rebars at midspan. Supervision, Writing ‐ review & editing. Tung T. Tran: Investigation,
Methodology. Thong M. Pham: Writing ‐ review & editing. Hong
Hao: Funding acquisition, Supervision, Writing ‐ review & editing.
4.1 × 103 με to 1.6 × 105 με in the same time period (i.e. 1–15 ms) as Zuyu Chen: Writing ‐ review & editing. Mohamed Elchalakani:
shown in Fig. 22(c). This meant the tension steel bars in Beam B‐Steel Investigation.
yielded, thereby Beam B‐Steel experienced much larger residual mid-
span deflection (i.e. 27.8 mm) than that of Beam B‐BFRP (i.e. Declaration of Competing Interest
5.0 mm). From these observations, it could be concluded that the
impact performance of flexure‐critical concrete beams reinforced with The authors declare that they have no known competing financial
BFRP bars is comparable to that of conventional RC beams with steel interests or personal relationships that could have appeared to influ-
bars. It should be mentioned that this conclusion may not be applica- ence the work reported in this paper.
ble for flexure‐critical concrete beams with tension reinforcement ratio
less than 0.41% since the beams reinforced with BFRP bars could expe- Acknowledgements
rience rupture damage of tension BFRP bars owing to their less defor-
mation capability as compared to steel bars and thus may lead to an The support by Australian Research Council (ARC) via Australian
adverse effect as demonstrated in section 6.1. Laureate Fellowship (FL180100196) is acknowledged.
16
Z. Huang et al. Composite Structures 263 (2021) 113648
17