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

net/publication/339568756

Axial behavior of reinforced concrete column with ultra-high performance

concrete stay-in-place formwork

Article  in  Engineering Structures · May 2020

DOI: 10.1016/j.engstruct.2020.110403

CITATIONS READS

0 75

4 authors, including:

Huiwen Tian Zhen Zhou

Southeast University (China) Southeast University (China)
4 PUBLICATIONS   8 CITATIONS    48 PUBLICATIONS   160 CITATIONS   

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Huiwen Tian on 29 February 2020.

The user has requested enhancement of the downloaded file.

 Engineering Structures 210 (2020) 110403

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Axial behavior of reinforced concrete column with ultra-high performance T

concrete stay-in-place formwork
⁎
Huiwen Tiana, Zhen Zhoua, , Yi Zhangb, Yang Weic
a
Key Laboratory of Concrete and Prestressed Concrete Structures of the Ministry of Education, Southeast University, Nanjing 210096, China
b
BAOYING Country Garden Real Estate Development Co., Ltd, Yangzhou 225800, China
c
College of Civil Engineering, Nanjing Forestry University, Nanjing 210037, China

A R T I C LE I N FO A B S T R A C T

Keywords: This paper presents the experimental and numerical study on the axial compressive behavior of a novel com-
Composite columns posite column that consists of a prefabricated grid-reinforced ultra-high performance concrete (UHPC) stay-in-
Ultra-high performance concrete (UHPC) place (SIP) formwork and post-cast concrete. Eight specimens, including six composite columns and two re-
Stay-in-place formwork inforced concrete columns were tested. The test variables included the type and number of grid layers as well as
Grid reinforcement
the thickness of UHPC formwork. According to the test results, the composite columns showed both higher axial
Axial compression
load carrying capacity and elastic modulus, whereas the brittle nature of UHPC caused a decrease in ductility
and toughness. The use of carbon fiber reinforced polymer (CFRP) grid could provide a more sufficient lateral
confinement than stainless steel (SS) grid to achieve a strain hardening behavior and thus increased the ductility
and toughness of the specimens. The use of more layers of SS grids could also have a positive effects on brit-
tleness reduction. Furthermore, an analytical model was developed using a finite element (FE) program to si-
mulate axial compressive response of the column. The whole failure process of the composite column and almost
all the key characteristics of the load-strain curves were reasonably captured, and a good agreement was found
between the calculated and measured load-strain curves, indicating the capability of the proposed FE model to
predict the axial compressive behavior of the composite columns.

1. Introduction the strength and ductility of the columns [5–7]. Even though CFST
possesses a number of benefits, disadvantages need to be mentioned as
Concrete is the primary engineering material for construction well: connections with beams are fairly complex and additional mea-
during the past century. Almost all of the cast-in-situ construction of surements are required to protect against fire or corrosion [8]. Another
concrete structures need the formwork system that are made up of rigid widely studied SIP formwork system for columns is concrete filled fiber
materials such as steel, aluminum and wood. However, the cost of reinforced polymer (FRP) tubes (CFFTs) [9,10]. These tubes provide a
formwork material and the labor to produce, assemble and re-use the lightweight and non-corroding permanent formwork system due to
formwork is very high, often accounting for more than 50% of the total their high strength-to-weight ratio and excellent corrosion resistance
cost of the concrete structures [1]. Also, the quality of the onsite con- [11]. The confining fibers in the hoop direction can result in profound
struction of formwork often varies a lot from site to site. Therefore, the enhancements in the compressive strength and ultimate strain of con-
stay-in-place (SIP) formwork, a prefabricated structural element that crete [12–17]. However, FRP can’t resist high temperature so the poor
modules fresh concrete to the designed dimensions and remains in place fire safety and the possible toxic gases release should be taken into
to bear loads throughout the structure lifecycle, is a promising alter- consideration [18]. Besides, plenty of literature has reported the fea-
native to the traditional formwork system [2,3]. sibility on structural SIP formworks made of novel cementitious mate-
Several SIP formwork systems for columns have been proposed and rials, such as textile reinforced concrete (TRC) [19–22] and engineered
investigated in previous literature. Concrete filled steel tube (CFST) is a cementitious composites (ECC) [23–26]. Several test results, including
well-known example of permanent column formwork [4]. The steel results from Papanicolaou et al. [19] and Qiao Z et al. [23] etc., showed
tube not only functions as the necessary reinforcement for concrete that the prefabricated TRC and ECC SIP formwork is an attractive al-
structures, but also provides confinement to inner concrete to enhance ternative to the traditional wood/steel formwork systems and the

⁎
Corresponding author.
E-mail address: [email protected] (Z. Zhou).

https://doi.org/10.1016/j.engstruct.2020.110403
Received 15 December 2019; Received in revised form 16 February 2020; Accepted 19 February 2020
0141-0296/ © 2020 Elsevier Ltd. All rights reserved.
H. Tian, et al. Engineering Structures 210 (2020) 110403

composite columns showed a better deformability. Table 1

As an innovative cement-based material, ultra-high performance Details of tested specimens.
fiber reinforced concrete (UHPFRC, UHPC for short hereafter) exhibits Group Specimen Formwork Grid type Number of grid
superior mechanical properties, including ultra-high compressive thickness/mm layer
strength coupled with high flexural and tensile strength, high ductility
Control group RC20 0 – 0
and toughness, and high durability [27–30], which makes it possible to
RC40 0 – 0
create a lightweight, durable, and fire-proof SIP formwork. Firstly, the T20 group T20L0 20 – 0
high compressive and tensile strength of UHPC can not only lead to a T20S1 20 SS 1
thinner formwork that is beneficial for faster and cheaper transporta- T20C1 20 CFRP 1
tion and construction, but also smaller cross section of the columns with T40 group T40L0 40 – 0
T40S1 40 SS 1
UHPC SIP formwork than the traditional RC columns. Secondly, as a
T40S2 40 SS 2
cementitious material, the UHPC formwork is fire-proof [31,32]. The
excellent durability of UHPC are able to reduce maintenance costs for
cover to resist weathering effects [33,34]. Thirdly, test results of
strengthening of reinforced concrete (RC) members using UHPC
showed a good bonding between UHPC and concrete [35–38], which
UHPC formwork
means that the additional measurements to transfer possible shear force
may not be needed in such a composite column. Lastly, the UHPC Grid reinforcement
permanent formworks can be produced in specialized factories to en-
sure a high quality and then transported to the construction sites for Cast-in-place concrete
assembling. In this way, the total cost of the new columns may be even
less than traditional RC columns as a result of saving formwork mate- Steel reinforcement
rials and construction time and labors. In addition, the partial appli-
cations of UHPC in structural members are beneficial for the promotion
of this expensive but high-qualified material.
Based on the aforementioned discussions, a novel composite
column, i.e., RC column with UHPC SIP formwork, is proposed and D
presented in Fig. 1. The precast UHPC formwork is reinforced with FRP
or SS grids to enhance its tensile strength and provide additional con- Fig. 2. Cross-section of the composite columns with UHPC SIP formwork.
finement for the inner concrete. Compared with basalt FRP (BFRP) grid
sand carbon-glass hybrid FRP (HFRP) grids, the author’s previous ex- 2. Test program
perimental study on tensile mechanical property of grid reinforced
UHPC plates with carbon FRP (CFRP) and stainless steel (SS) grids 2.1. Design of the experiment
showed a more significant reinforcing effect, including tensile strength
and ductility on UHPC plates [39]. Therefore, CFRP and SS grids were To investigate the axial compressive behavior of composite col-
used in the paper to strengthen the prefabricated UHPC formworks. umns, 3 groups of specimens including Control group, T20 group and
Axial compression tests are conducted on the specimens to investigate T40 group, as shown in Table 1, were prepared and tested under axial
the effects of grid material, number of grid layers and formwork compression. The cross-section of the composite columns is shown in
thickness on the compressive behavior of the composite columns. A Fig. 2. All specimens were 240 mm in diameter and 600 mm in height.
finite element (FE) model is also established to simulate the whole CFRP and SS grids were used in the paper to strengthen the pre-
failure process. Test and simulation results highlight the significant fabricated UHPC formworks. In addition, the single layer and double
potential of the use of UHPC in the development of high performance layers of SS grids were used to clarify the effects of amount of grids. The
SIP formworks. post-cast normal strength concrete (NSC) core was reinforced trans-
versely with a spiral (Ø8 mm) at a 45 mm and 60 mm pitch in 20 mm
and 40 mm thick UHPC formwork respectively, resulting in a same
transverse steel ratios of 2.6%. Moreover, 6 longitudinal bars (Ø
10 mm) were evenly distributed to ensure a proper spacing of the spiral.
The specimens in Table 1 are labeled as follows: (a) RC20 and RC40
are the corresponding control specimens for T20 and T40 groups, and
the numbers 20 and 40 in the control specimens refer to the concrete
cover thickness. (b) the number after “T” represents formwork thick-
ness (mm); (c) the number after “L”, “S” or “C” respectively represents
the number of grid layer. For example, specimen T40S1 is the one with
40-mm-thick formwork and single layer of SS grid.

2.2. Material property

The UHPC used in the paper is an easily available commercial

product, SBT®-HDC. According to the preliminary study of the influence
of UHPC mix proportion on its physical properties, mechanical prop-
erties and constructability, the UHPC with a 1% volume fraction of steel
fibers and water binder ratio of 0.18 was adopted in this study. The
detailed proportions of the UHPC are presented in Table 2. The me-
chanical properties and dimensions of the steel fibers are presented in
Fig. 1. Reinforced concrete column with UHPC SIP formwork. Table 3.

2
H. Tian, et al. Engineering Structures 210 (2020) 110403

Table 2
Mix proportion of UHPC with 1% volume fraction steel fibers.
52.5 cement High-efficient admixture Fine sand Water Superplasticizer Steel fiber

−3
proportion/(kg·m ) 791.0 372.0 930.4 202.0 11.63 80.0

The nominal UHPC and post-cast NSC compressive strength was

determined by cubic specimens according to the Chinese standard [40].
The cubic specimens were cured in the same condition as the experi-
mental columns and the test was completed within two days to ensure
the same age of all specimens. Because the difference between all test
values and the corresponding intermediate value didn’t exceed 15% of
the intermediate value, the average compressive strength was chosen to
be the nominal concrete compressive strength, i.e., 139.3 MPa for
UHPC and 32.4 MPa for NSC. The yield strength, ultimate strength and
elongation of steel reinforcement were 417.2 MPa, 563.7 MPa and
24.9%, respectively.
The CFRP and SS grids used in the paper were the same product
used in the author’s previous study on tensile mechanical property of (a) CFRP grid (b) SS grid
grid reinforced UHPC plates [39]. As is shown in Fig. 3, the CFRP grid is
Fig. 3. Grids used to reinforcing UHPC formwork.
a flexible, open-mesh grid with a cross-sectional area of 0.812 mm2 of
each element and a center to center distance of 16 mm between the
cross elements both in the longitudinal and the lateral direction, while
the SS grid is an inflexible, rigid-jointed grid with a diameter of 1.2 mm
of each element and a center to center distance of 12 mm in both di-
rections. A group of single SS grid element and single epoxy resin im-
pregnated CFRP grid element were tested under axial tension (Fig. 4) to
obtain the stress-strain relationship. As is shown in Fig. 5, the CFRP grid
showed a linear elastic stress-strain relationship, with a maximum
tensile strength, ultimate strain and modulus elasticity of almost
3459.3 MPa, 2.09% and 167 GPa, respectively. The SS grid showed a
strain hardening stress-strain relationship, with a yield strength, max-
imum tensile strength, ultimate strain and modulus elasticity of almost
442.6 MPa, 737.4 MPa, 21.7% and 210 GPa, respectively.

2.3. Specimen preparations

In order to maintain a cylindrical shape during casting UHPC, the

flexible CFRP grid was impregnated by epoxy resin with two steel rings
at both ends, and cured for 24 h. Fine sand was evenly sprayed on the
Fig. 4. Set up for grid tensile test.
CFRP grid to increase the bond strength to UHPC. The SS grids were
formed by directly rolling up into a cylinder and fixing the overlap
zone, as is shown in Fig. 6. 2.4. Instrumentation and test procedures
To produce the experimental columns, the UHPC formworks were
firstly manufactured by casting fresh UHPC into the PVC molds with As is shown in Fig. 7, for each specimen, two linear variable dis-
embedded grids. After being cured in room temperature for 24 h, the placement transformers (LVDTs) were mounted onto two aluminum
UHPC formworks were demolded and then cured in hot water at 90 °C brackets that were attached to the middle region (gauge length equals
for 48 h. Afterwards, the reinforcement cage was fixed and the concrete 300 mm) of the specimen to obtain the axial deformation of this region.
was poured in the UHPC tube. It should be noted that the construction Another two LVDTs were installed to measure the total displacement of
time can be saved only when UHPC formworks are able to be pre- the loading plate. Four pairs of axial and hoop strain gauges with
fabricated in specialized factories. The construction of such columns 50 mm gauge length were installed at mid-height of the UHPC surface.
can be simplified to assembling UHPC formworks in site and then The tests were carried out using a 5000 kN universal testing machine
casting concrete. with a displacement control rate of 1 mm/min until failure.
All specimens were moved into the concrete standard curing room
with a temperature of 20 ± 2 °C and humidity over 95% until testing.
Moreover, to prevent the premature failure of column ends, two layers
of 100 mm width CFRP strips were used to strengthen the ends of each
specimen before testing.

Table 3
Mechanical properties and dimensions of steel fibers.
Length/(mm) Diameter/(mm) Aspect ratio Density/(g·cm−3) Tensile strength/(MPa) Elastic modulus/(GPa)

13.0 0.2 65 7.9 > 2900 200

3
H. Tian, et al. Engineering Structures 210 (2020) 110403

4000 800

C1
C2
3000 600
Stress (MPa) C3

Stress (MPa)
2000 400

1000 200 S1
S2
S3
0 0
0.0 0.5 1.0 1.5 2.0 2.5 0 5 10 15 20 25
Strain (%) Strain (%)
(a) CFRP grid (b) SS grid
Fig. 5. Stress-strain relationships of grids.

strength of the column, the cracks grew more rapidly and the concrete
cover began to spall, as shown in Fig. 8a and e.
The failure modes of two composite specimens without grid re-
inforcement, i.e., T20L0 and T40L0 were similar (Fig. 8b and f). Slight
intermittent cracking noises could be heard after the load reached
about 85% of the peak value. Several vertical cracks on the surface
occurred and increased sharply when the load reached the compressive
strength. Although steel fibers around the cracks were pulled out, the
integrity of the specimens were well maintained because of the good
bond strength between steel fibers and matrix. Compared with the
composite specimens without grid reinforcement, the failure of the
reinforced counterparts was marked by fully developed inclined cracks.

3.2. Axial load-axial strain behavior

Table 4 summarize the major test results of all specimens. The load
Pmax is the maximum axial load of the columns. The nominal stress fcc′ is
(a) CFRP grid (b) SS grid obtained by dividing Pmax by the cross-sectional area of the column.
Fig. 6. The formed grid reinforcement. Axial strains εc , ε80 and ε50 , corresponding to the peak load, the re-
maining capacity of the column dropping to 80% and 50% of the peak
load, respectively, are also presented. These strains provide indicators
of the rate of strength degradation during the testing. Elastic modulus E
is the slop of the straight line from the origin through 0.45Pmax .
The ductility index, I10 [41], and toughness index, T . I . [42], were
used to evaluate the influence of selected variables on the post-peak
ductility and toughness of the composite columns, respectively. The
definitions of I10 and T . I . are shown in Fig. 9. The ductility index I10 is
determined by dividing the area AF by the area AY , where AF is the area
under the axial load-strain curve up to a strain of 5.5 times the ‘‘yield
strain” and AY is the area under the curve up to the ‘‘yield strain”. The
“yield strain” εy , is obtained by extrapolating a straight line from the
origin through 0.75Pmax until it reaches Pmax (see Fig. 9a). The toughness
index T . I . is the ratio of area under the load–strain curve up to εAu to
the area of an equivalent rigid material, as shown in Fig. 9b.
Fig. 10 presents the axial load versus strain responses of the com-
posite columns, and Fig. 11 illustrates the primary turning points of the
Fig. 7. Test set up. representative load-stain curves. Based on the test results, four key
characteristics can be concluded:
3. Test results and discussion
(1) Almost linear elastic response until initial peak load. All composite
3.1. Observed failure modes columns exhibited almost linear response in the ascending part of
the load-strain relationship until reaching the initial peak load. The
The typical failure modes of all specimens are shown in Fig. 8. The typical softening response of ascending part of the load-strain
control specimens, RC20 and RC40 exhibited several slight inclined curves in NSC specimens was not found (RC20 and RC40 in Fig. 10),
cracks when the axial load reached about 75% of the peak value. As the which basically because UHPC members behave elastically up to
loading proceeded, especially when approaching the compressive approximately 90%–95% of their compressive strength [43].
(2) Sudden strength loss after the initial peak load. All composite

4
H. Tian, et al. Engineering Structures 210 (2020) 110403

(a) RC20 (b) T20L0 (c) T20S1 (d) T20C1

(e) RC40 (f) T40L0 (g) T40S1 (h) T40S2

Fig. 8. Observed failure modes.

Table 4 to the post-peak strength between the composite columns and the
Summary of test results. corresponding control specimens. However, because of the con-
Specimen Pmax /kN ′ /MPa
fcc εc /% ε80 /% ε50 /% E /GPa I10 T. I. finement provided by the steel fibers and the unfractured grids, the
UHPC formwork did not lose the bearing capacity completely,
RC20 1325.4 29.3 0.264 2.010 -b 28.9 8.81 0.87 which results in a larger post-peak strength than their control
T20L0 2252.0 49.6 0.228 0.249 -b 33.0 7.35 0.61 counterparts.
T20S1 2061.9 45.2 0.217 0.914 -b 32.0 8.59 0.75
T20C1 2436.8 53.6 0.579 0.620 2.811 32.4 7.89 0.68
RC40 1326.9 29.3 0.241 0.292 1.07 27.8 6.67 0.45a
It should be noted that the failure of the composite column in this
T40L0 3813.6 84.1 0.298 0.303 0.303 37.6 3.53a 0.26a research was marked by the sudden strength loss shortly after the peak
T40S1 3825.8 85.7 0.280 0.294 0.294 37.6 4.36 0.35 load because of the fact that the remaining capacity after this strength
T40S2 4027.2 88.8 0.321 0.322 0.322 37.7 4.83 0.40 loss was less than 85% of the peak load. However, the amplitude of the
a strength loss could be significantly reduced to a relatively low level by
The test was terminated before reaching the axial strain of 0.02 and the
decreasing the proportion of UHPC on the cross-section of the compo-
T . I . was calculated by extending the curve using a linear fitting function.
b
The test was terminated before the remaining capacity of the column drop site column, which could be concluded by making a comparison of the
to 50% of the peak load. axial load-strain curves between T20 and T40 series specimens. Under
such circumstance, the failure of the composite column could be de-
columns experienced a sudden strength loss shortly after the peak fined by the remaining capacity of the column dropping to 85% of the
load owing to the brittle nature of UHPC material, except for spe- peak load.
cimen with sufficient lateral confinement (e.g. T20C1) that re-
strained the rapid development of UHPC cracks and reduced the 3.3. Influence of formwork thickness
strength loss to a negligible level.
(3) A strength recovery stage. A strength recovery stage after the It can be observed from Table 4 and Fig. 10 that the composite
sudden strength loss was observed owing to the good bond between columns exhibited a higher maximum load and elastic modulus com-
steel fibers and UHPC matrix as well as the confinement provided pared with the corresponding RC columns. Specifically, the specimen
by the grid reinforcements. This strength recovery stage was ter- T20L0 gained an increase of 69.3% and 14.2% in maximum load and
minated by the pull-out of steel fibers and fracture of grids and elastic modulus respectively. While for that of T40L0, substantial in-
followed by another strength loss. creases of 186.6% and 35.3% were found. However, the thicker the
(4) A strain softening stage. The post-peak strength was mainly con- UHPC formwork was, the more significant sudden strength loss shortly
tributed by the confined NSC core, which is evident when compared after the initial peak load was observed. Although a strength recovery

5
H. Tian, et al. Engineering Structures 210 (2020) 110403

Fig. 9. Definitions of ductility index I10 and toughness index T . I .

could be achieved in specimens with thinner formwork (e.g. T20L0)

Initial peak load Pull-out of teel fibers
owing to the bond strength between steel fibers and UHPC matrix, this
phenomenon was not found in specimens with thicker formwork (e.g. and fracture of grids
Pmax
T40L0). In fact, the steel fibers in these specimens were pulled put once
the UHPC cracked because of the instantaneous release of a greater

Axial load (kN)

accumulated strain energy. As a result, the ductility index I10 , and
toughness index T . I . decreased due to the brittle nature of the UHPC
material. In other words, the thicker the UHPC formwork, the more Initiation of
Initiation of
brittle the composite columns were, if no sufficient lateral confinement softening stage
strength recovery
for UHPC formwork were provided.
To sum up, although the severe strength drop after peak load and
the brittle behavior of the specimens could be controlled by reducing
the thickness of the UHPC formwork, the beneficial effect on peak load
enhancement could become more insignificant. Additionally, in con-
sideration of the requirement of strength and stiffness for UHPC form-
works and the function to act as a protective jacket, a minimum
thickness of 20 mm for UHPC formworks is necessary in actual struc- εc Axial strain
tures. In addition, compared with the full-scale columns constructed
with the same-thickness UHPC formworks in the actual structures, the Fig. 11. Key characteristics of typical axial load-strain curve
proportion of UHPC material in the total cross section is higher for the
experimental columns. Therefore, the severe strength loss after peak
3.4. Influence of grid type
load caused by UHPC could be reduced to an acceptable level when
applied in actual structures.
It can be seen from the comparison of the axial load–strain re-
sponses (Fig. 10a) of composite columns with plain UHPC formworks
(T20L0) and CFRP grid and SS grid reinforced UHPC formworks (T20C1
and T20S1), the type of reinforcing grids showed a significant effect on

2500
RC20 4000 RC40
T20L0 T40L0
2000
T20S1 T40S1
T20C1 3000
Axial load (kN)
Axial load (kN)

T40S2
1500

2000
1000

1000
500

0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Axial strain (%) Axial strain (%)

(a)T20 series and RC20 specimen (b) T40 series and RC40 specimen
Fig. 10. Axial load–strain responses of the specimens

6
H. Tian, et al. Engineering Structures 210 (2020) 110403

Fig. 12. The numerical model: (a) Overall model; (b) NSC; (c) Steel reinforcement in NSC; (d) UHPC matrix; (e) Grid in UHPC matrix; (f) randomly distributed steel
fibers; (g) Load plate and strengthening FRP strips.

the axial compressive performance of the specimens. the maximum load of T20S1 should be slightly larger than that of
The axial load-strain relationship of T20S1 was similar to that of T20L0, theoretically.
T20L0, i.e., a linear elastic increase to the peak load followed with a The axial load-strain relationship of T20C1 was different from that
sudden strength loss to a residual bearing capacity. This phenomenon of T20L0 and T20S1, where a strain hardening stage rather than a
could be explained that the yield strength of the SS grid was too small to sudden strength loss shortly after the linear elastic stage was found.
provide a sufficient confinement to restrain the occurrence and devel- This difference was caused by a more sufficient lateral confinement
opment of the cracks in the early stage of loading. However, the ex- provided by the CFRP grid that effectively restrained the occurrence
cellent elongation of the SS grid, as illustrated in Fig. 5, permitted a and development of the cracks on UHPC. The gradual fracture of the
continuous lateral confinement even when the cracking widths were CFRP grid around the major cracks caused several step-shaped decrease
very large, which resulted in a greater post-peak strength in T20S1. As in the axial load-strain curve. Owing to the lower ultimate tensile strain
mentioned previously, the step-shaped decrease in the strain softening of CFRP grid than SS grid, the ductility and toughness of T20C1 was
stage could be attributed to the gradual fracture of the grids around slightly lower than that of T20S1. However, a positive increase of 7.3%
major cracks. The ductility and toughness of the composite columns and 11.5% were found in I10 and T . I .when compared with T20L0.
were substantially enhanced with the introduction of SS grids. Speci- In conclusion, to achieve a strain hardening behavior in the com-
fically, specimen T20S1 showed positive increases of 16.9% and 23.0% posite column, adequate amounts of lateral confinement are required
in ductility index I10 , and toughness index T . I ., respectively, when for UHPC formworks. Compared with SS grids with a high elongation,
compared with T20L0. It should be noted that the maximum load of CFRP grids with high strength is more effective to achieve a strain
T20S1 was slightly lower than that of T20L0, which was due to the hardening behavior. What’s more, CFRP grids with larger cross-section
premature cracks caused by an unexpected hole around the protruding area or more layers are needed to realize a more ductile behavior by
wires of inner strain gauges. As was further illustrated in Section 4.2, extending the strain hardening stage.

7
H. Tian, et al. Engineering Structures 210 (2020) 110403

Table 5
Adopted strength surfaces parameters for the K&C model.
Yield surface Maximum surface Residual surface

a0y a1y a2y a0 a1 a2 a0f a1f a2f

NSC 6.026 0.625 0.00954 2.707 0.446 0.00299 0.00 0.442 0.00438
UHPC 29.02 0.625 0.00198 38.43 0.446 0.00622 0.00 0.442 0.00091

Start

Define the space of a specimen and input

the number of steel fibers

Randomly generate the starting point for

a fiber within the defined space
Yes
Calculate the end point of the fiber by
randomly rotating with fixed length No

Is the end point located

within the defined space?

Yes
Yes

Does the fiber intersect or

Is this the first fiber? No
coincide with existing fibers?

Yes

Record the fiber location No

Is there any more fibers?

No

Output the records.

End

Fig. 13. Generation of steel fibers.

3.5. Influence of grid layers fracture of two layers of grids caused by the different hoop strain ex-
isted in the two layers, i.e., a larger hoop strain existed in the outer
The influence of number of grid layers on the axial behavior of the layer of SS grid than that in the inner layer. The ductility and toughness
specimens was illustrated in Fig. 10b, where the composite columns of the specimens enhanced with the increase of number of grid layers.
with different number of grid layers (T40L0, T40S1 and T40S2) are Compared with T40L0, significant increases of 25.3% and 34.6% of
presented. It is evident that the axial load-strain response in the linear I10 were found in T40S1 and T40S2 respectively, and the corresponding
elastic stage was identical while the maximum load of specimens with increases of T . I . were 36.8% and 53.8%.
more number of grid layers was slightly greater. Specimens with two
layers of SS grid also experienced a sudden strength loss shortly after 4. Numerical simulation
the peak load, indicating that just increasing the amount of SS grids
with relatively low strength is not capable to achieve a strain hardening 4.1. Development of model
behavior.
With the increase of number of grid layers, a more significant As is shown in Fig. 12, the FE model consisting of UHPC matrix,
strength recovery was found. In addition, the specimens with two layers randomly distributed steel fibers, grid reinforcement, NSC and steel
of SS grid showed a second sudden strength loss owing to the gradual reinforcement, was established in LS-DYNA software on the basis of

8
H. Tian, et al. Engineering Structures 210 (2020) 110403

Table 6 actual dimension and boundary conditions to simulate the axial com-
Summary of comparison between experimental and numerical results. pressive behavior of the specimens. Based on the observations from
Specimen Pmax (kN) εc (%) ε80 (%) ε50 (%) previous literature [44–47], UHPC could be assumed to be a two-phase
model composed of UHPC matrix and steel fibers, and the explicitly
NUM EXP/ NUM EXP/ NUM EXP/ NUM EXP/ modelled steel fibers could give beneficial insights on the fiber re-
NUM NUM NUM NUM inforcement effect. Therefore, in this research, the UHPC matrix was
T20L0 2275.8 0.99 0.241 0.95 0.284 0.88 – – simulated with homogeneous material and the straight steel fibers were
T20S1 2401.6 0.86 0.308 0.70 0.372 2.46 – – presumed to be dispersed in the UHPC matrix with random locations
T20C1 2474.6 0.98 0.335 1.73 0.406 1.53 2.988 0.94 and orientations. Details of the FE model development are as follows.
T40L0 3815.3 1.00 0.294 1.01 0.331 0.92 0.343 0.88
T40S1 3887.1 0.98 0.325 0.86 0.348 0.84 0.359 0.82
T40S2 3957.4 1.02 0.344 0.93 0.359 0.90 0.387 0.83 4.1.1. Element description
The concrete and UHPC matrix was modelled using a three-di-
mensional eight-node constant stress element with a single point

2500 2500
Test Test
Simulation Simulation
2000 2000
Axial load (kN)

Axial load (kN)

1500 1500

1000 1000

500 500
T20L0 T20S1
0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Axial strain (%) Axial strain (%)
2500
Test 4000 Test
Simulation Simulation
2000
3000
Axial load (kN)

Axial load (kN)

1500

2000
1000

1000
500
T20C1 T40L0
0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Axial strain (%) Axial strain (%)

4000 Test 4000 Test

Simulation Simulation

3000 3000
Axial load (kN)

Axial load (kN)

2000 2000

1000 1000

T40S1 T40S2
0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Axial strain (%) Axial strain (%)
Fig. 14. Comparison of experimental and numerical axial load-strain curves.

9
H. Tian, et al. Engineering Structures 210 (2020) 110403

(a) NSC core (b) UHPC formwork (c) steel spiral (d) SS grids (e) steel fibers
Fig. 15. Stress distribution of each component of specimen T40S1 at ε = 0.1% .

(a) NSC core (b) UHPC formwork (c) steel spiral (d) SS grids (e) steel fibers
Fig. 16. Stress distrbution of each component of specimen T40S1 at ε = εc .

(a) NSC core (b) UHPC formwork (c) steel spiral (d) SS grids (e) steel fibers
Fig. 17. Stress distrbution of each component of specimen T40S1 at ε = ε80 .

integration formulation, which could effectively overcome the compu- element size of 5 mm. What’s more, the element size of 15 mm was
tational difficulties caused by the material nonlinearity and large ele- selected for concrete and 5 mm for steel reinforcement and grid.
ment deformation. An hourglass coefficient of 0.06 was adopted to
resist the zero-energy deformation modes, i.e., the hourglass modes.
4.1.2. Material model
The Hughes-Liu beam element was used to model the steel reinforce-
The reasonable constitutive models for NSC and UHPC are the key
ment, grid and steel fibers. The strengthening FRP strips on each end of
to predict the axial performance of composite columns. The Karagozian
the specimen were modeled using a four-node shell element with six
and Case (K&C) model in LS-DYNA [48] is widely used in the nonlinear
degrees of freedom at each node that included membrane, bending and
simulation of concrete structures and has proven itself capable of re-
shear deformation capabilities. The Belytschko–Tsay element formula-
flecting most of the key characteristics of concrete, such as hardening,
tion was adopted for the shell element. The loading plates were pre-
post-peak softening, confinement and shear dilatancy, etc. This model
sented by rigid elements.
uses three pressure-sensitive, independent strength surfaces, i.e., yield
The element size of UHPC matrix need to be compatible with that of
surface, maximum surface and residual surface, to represent the beha-
steel fibers to reflect the interaction of each other. Therefore, con-
vior of concrete material. What’s more, this model provides an auto-
sidering computational efficiency and accuracy, each 13 mm long steel
matic model parameter generation program based solely on the un-
fibers was divided into two elements, and the UHPC matrix had an
confined compression strength of the concrete. The automatic

10
H. Tian, et al. Engineering Structures 210 (2020) 110403

(a) NSC core (b) UHPC formwork (c) steel spiral (d) SS grids (e) steel fibers
Fig. 18. Stress distrbution of each component of specimen T40S1 at ε = 2.5% .

generating model parameters for NSC in this paper was adopted based simulate this linear elastic material with consideration of the fiber
on its actual compressive strength. It should be noted that the com- fracture. The parameters of these materials used the measured values in
pressive strength of 26.6 MPa was used in consideration of a conversion Section 2.2.
coefficient of 0.82 from cube strength into cylinder strength [49]. The
corresponding strength surfaces parameters for NSC are listed in 4.1.3. Contact algorism
Table 5. The default values for other input parameters of K&C model for As was illustrated in Section 3.1, the steel reinforcement was
NSC were used, such as exponents b1 and b2 respectively governing the bonded to NSC firmly, so was reinforcing grids to UHPC, thus the steel
compressive and tensile softening damage evolutions, and equation of and grid reinforcement elements were embedded into the corre-
state controlling the volumetric response and so on. sponding solid elements. The experimental results also revealed that the
Owing to the above mentioned advantages, and especially the steel fibers around the cracks were generally pulled out as the enlarging
capability to capture the implicit confinement of reinforcement that is of the crack width, which means a certain bond-slip model may be
essential to reflect the fiber effect on UHPC, the K&C model was also required for a precise simulation. The research from C Wu et al. [44]
adopted for UHPC matrix in this study. However, this material model indicated that slide line contact in LS-DYNA had the capacity to truly
was developed based on the extensive calibration of generic concrete reflect the interfacial behavior of fiber and matrix, and steel fiber ma-
(45.4 MPa NSC). Thus the automatic parameter may not effectively terial could be explicitly modelled by this way and achieved high si-
represent the response of UHPC. Research from Xu [50] validated the mulation fidelity in small size models. However, the specimen’s size in
feasibility of K&C model to simulate the mechanical behavior of UHPC this paper were much larger. On account of the massive amount of the
and proposed a set of parameters for UHPC (150 MPa). Then the steel fiber elements, the computational cost and difficulties in solution
parameters for UHPC with other similar strength could be determined convergence will dramatically increase if the bond-slip between fiber
by and UHPC matrix are considered. Moreover, several existing literature
[45–47] reported that the model with perfect bonding behavior could
a0new
i = ϕa0old
i a1new
i = a1old
i a2new
i = a2old
i ϕ (1)
also realize fairly good simulation results. Therefore, the explicitly
with modelled steel fibers in this study were also embedded into the UHPC
matrix. The interfacial behavior between NSC and UHPC was re-
ϕ = f c'new f c'old (2) presented by the surface to surface contact, where the pressure could be
where a0i, a1i and a2i are the strength surfaces parameters; and f c'old
and transformed in the normal direction, and any shear sliding on the in-
f c'new are the given and undetermined unconfined compressive strength terface was avoided by adopting a value of 1.0 for friction coefficient.
of UHPC, respectively. In order to simulate the displacement boundary conditions, two
As was established in previous studies [51], a conversion coefficient rigid plates were established. The surface to surface contact was also
of 0.94 from cube strength into cylinder strength need to be considered used for the corresponding surface of the column and rigid plates. A
in this paper. Therefore, the compressive strength of 130.9 MPa was value of 0.6 was adopted for friction coefficient to consider the lateral
used to calculate the strength surfaces parameters for UHPC, which are confinement effect on specimen end. The rigid plates in the base was
listed in Table 5. The equation of state suggested by Xu [50] were also applied for nodal translation constraints in all directions, while rigid
adopted in this study and default values were used for other input plates on the top was applied in the global X and Y directions and the
parameters. nodes were free to translate in the z-direction to apply displacement
In addition, LS-DYNA provides several numerical erosion algo- controlled axial compression to the column.
rithms, such as strain based, stress based damage based erosion criteria.
Although element erosions could be used to represent physical phe- 4.1.4. Random distributed steel fibers
nomena, it is an effective numerical solution to avoid element distortion In order to explicitly analyze the influence of steel fibers on the
during the simulation and is beneficial in the visualization of structural compressive behavior of UHPC, the random distributed steel fibers
component failure. In this study, the erosion criteria based on the were established in the model. The random fibers are generated in
maximum principal strain was selected, and the value of 0.15 and 0.025 Matlab software and the flowchart of program is given in Fig. 13.
were adopted for NSC and UHPC respectively.
The bilinear elastic-perfect plastic model (Mat_Plastic_Kinematic) 4.2. Comparison with test results
were used to simulate the steel reinforcement, steel fiber and grid. This
model is very cost-effective and is capable to reflect the isotropic and A summary of comparison between experimental and numerical
kinematic hardening plasticity of steel materials. What’s more, a limit results are presented in Table 6, where the peak axial load Pmax , and
of effective plastic strain could be defined for eroding elements. corresponding axial strain εc , the axial strains ε80 and ε50 respectively
Therefore, a very small value of 10-6 was adopt for CFRP grid to corresponding to the remaining capacity dropping to 80% and 50% of

11
H. Tian, et al. Engineering Structures 210 (2020) 110403

the peak load are evaluated. As is shown in the table, except for spe- were not able to restrain the development of the cracks on UHPC
cimen T20S1, the predicted results of the numerical model were gen- formwork due to its small cross-sectional area. As is shown in Fig. 17e,
erally in good agreement with the experimental results, especially in the steel fibers around the cracks were subjected to high tensile stress,
terms of predicted peak load of composite columns. Compared with which could provide additional confinement for UHPC formwork.
other specimens, the difference between the predicted and tested peak What’s more, a number of steel fibers entered plastic stage, while the
load of specimen T20S1 was much larger. This difference can be at- steel fibers may have already been pulled out from UHPC matrix in the
tributed to the unexpected premature cracks in the specimen end, experiment. Therefore, the reinforcing effects of steel fibers were
which has been explained in Section 3.3. The error between simulation slightly overestimated in the FE model as a result of the ignoring the
and test results became much larger in the post-peak stage of the spe- bond-slip relationship between steel fibers and UHPC matrix.
cimen, which could be observed in comparison of simulation and test in Upon the end of loading (ε = 2.5% ), the SS grids around “cracks” on
terms of ε80 and ε50 . the UHPC have fractured and most of the UHPC elements lost its
The comparison of experimental and numerical axial load–strain bearing capacity completely. Although some NSC elements remained a
curves are shown in Fig. 14. It can be seen that almost all the key limited bearing capacity, the majority, especially the elements at the
characteristics of the curves were captured by the FE model, including middle height of the column became eroding as the loading proceeded.
the linear elastic stage, the post-peak stage and the sudden strength loss According to the above analysis, the three dimensional FE model
shortly after peak load. However, the strength recovery after the sudden developed in this research could provide reasonably accurate predic-
strength loss could not be properly reflected. To explain this phenom- tions for the whole failure process of RC column with UHPC formwork
enon, it is beneficial to refer to the experimental failure modes of the under axial compression.
composite columns. For grid reinforced composite columns, a sudden
strength loss was always accompanied by the occurrence of several 5. Conclusion
major localized cracks. The SS or CFRP grid were then fully activated
and provided a relatively strong confinement to restrain the rapid de- This research reported an experimental investigation and numerical
velopment of the cracks, which resulted in a strength recovery of the simulation for reinforced concrete (RC) column with ultra-high per-
column. While for FE models, the isotropic material used for UHPC formance concrete (UHPC) stay-in-place (SIP) formwork under axial
matrix could not simulate the actual development of the cracks on compression. The test results including failure modes, axial load-stain
specimens so that the gradual fracture of the SS or CFRP grid around the relationships and influence of selected variables were firstly presented.
cracks could not be properly reflected. Therefore, the strength recovery A three-dimensional finite element (FE) model was then developed and
were not found in the FE models. What’s more, the shorter plateau calibrated based on the test results. The following conclusions were
region in the strain hardening stage of the load-strain curve of T20C1 drawn:
could also attributed to the absence of the actual macro cracks on the
FE model. (1) The failure modes of the composite columns are marked by several
Therefore, a good agreement was observed between the predicted inclined major cracks and the integrity of the specimens is well
and tested peak strength and strain, and almost all the key character- maintained due to the good bond strength between steel fibers and
istics except for the strength recovery of the load- strain curves were matrix, which is dramatically different from the failure modes
reflected, which indicated the capability of the proposed FE model to marked by large-area cover spalling in RC columns.
predict the axial compressive behavior of the composite columns. (2) The typical axial load–strain curves demonstrate four key char-
acteristics, including an almost linear elastic response until initial
4.3. Analysis of whole failure process peak load, a sudden strength loss after the initial peak load, a
strength recovery stage coupled with a strain softening stage.
Take specimen T40S1 for example, the failure processes of the (3) The load carrying capacity and elastic modulus are significantly
column are shown in Figs. 15–18, where the stress distribution of NSC improved with the increase of UHPC formwork thickness, but are
core, UHPC formwork, steel reinforcement, SS grid and steel fibers were hardly influenced by grid layers and grid types. However, the
presented. greater the UHPC formwork thickness, the more brittle the com-
In the linear elastic stage (e.g., ε = 0.1% ), the NSC and UHPC ele- posite columns are.
ments undertook a small uniform compression in Z direction, except for (4) Carbon fiber reinforced polymer (CFRP) grid can provide a more
several elements in the end of the column that subjected to slight stress sufficient lateral confinement than stainless steel (SS) grid to in-
concentration. Because of the insignificant expansion of NSC and crease the ductility and toughness of the specimens. The use of
UHPC, the tensile stress in the hoop direction of steel spiral and SS grid more layers of SS grids can also have a positive effects on brittleness
was very small, and most of the steel fibers were in the compression reduction.
status. (5) The whole failure process of the composite column and almost all
As the loading proceeded and upon reaching the ultimate bearing the key characteristics except for the strength recovery of the load-
capacity (ε = εc ), damages were firstly observed on the elements at the strain curves are reasonably reflected, and a good agreement is
middle height of UHPC and NSC. Due to the expansion of the UHPC and observed between the predicted and tested peak load and the cor-
NSC, especially at the middle height of the column, the grid and steel responding strain, which indicates the capability of the proposed FE
spiral experienced a relatively high tensile stress and the stress state of a model to predict the axial compressive behavior of the composite
number of steel fibers changed from compression to tension. columns.
Shortly after the peak load, a sudden strength loss was observed in
this stage (e.g., ε = ε80 ), accompanied by the occurrence of several vi- CRediT authorship contribution statement
sual “cracks” on the UHPC formwork. The “cracks” were represented by
the eroded elements that reached their erosion limits. According to the Huiwen Tian: Conceptualization, Methodology, Software,
comparison with the experimental failure mode in Fig. 17b, the simu- Validation, Formal analysis, Investigation, Resources, Data curation,
lation result showed a high fidelity. Although the steel spiral at the Writing - original draft, Writing - review & editing, Visualization. Zhen
middle height of the column reached its yield strength, it could provide Zhou: Conceptualization, Validation, Resources, Writing - review &
sustainable confining pressure owing to its excellent elongation, and editing, Supervision, Project administration, Funding acquisition. Yi
thus the bearing capacity and ductility of NSC were significantly im- Zhang: Methodology, Software, Data curation. Yang Wei: Validation,
proved. While for the SS grid, the insufficient confinement it provided Resources, Writing - review & editing.

12
H. Tian, et al. Engineering Structures 210 (2020) 110403

Acknowledgements [24] Xu L, Pan J, Chen J. Mechanical behavior of ECC and ECC/RC composite columns
under reversed cyclic loading. J Mater Civ Eng 2017;29(9):04017097.
[25] Li QH, Huang BT, Xu SL. Development of assembled permanent formwork using
The authors gratefully acknowledge the financial support by ultra high toughness cementitious composites. Adv Struct Eng 2016;19(7):1142–52.
National Key R&D Program of China under 2017YFC0703700 and the [26] Huang BT, Li QH, Xu SL, et al. Development of reinforced ultra-high toughness
Scientific Research Foundation of Graduate School of Southeast cementitious composite permanent formwork: Experimental study and digital
image correlation analysis. Compos Struct 2017;180:892–903.
University (Grant NO. YBPY1959). The authors would also like to thank [27] Graybeal BA. Compressive behavior of ultra-high-performance fiber-reinforced
the Sobute New Materials Co. Ltd for providing the UHPC reported in concrete. ACI Mater J 2007;104(2):146.
this paper. [28] Park SH, Kim DJ, Ryu GS, et al. Tensile behavior of ultra high performance hybrid
fiber reinforced concrete. Cem Concr Compos 2012;34(2):172–84.
[29] Shi C, Wu Z, Xiao J, et al. A review on ultra high performance concrete: Part I. Raw
References materials and mixture design. Constr Build Mater 2015;101:741–51.
[30] Wang D, Shi C, Wu Z, et al. A review on ultra high performance concrete: Part II.
Hydration, microstructure and properties. Constr Build Mater 2015;96:368–77.
[1] Robert H. Think formwork-reduced cost. Struct Mag 2007:14–6.
[31] Xiong MX, Liew JYR. Spalling behavior and residual resistance of fibre reinforced
[2] Remy O, Verbruggen S, Wastiels J, et al. Cement composite stay-in-place formwork:
Ultra-High performance concrete after exposure to high temperatures. Materiales de
a concept for future building systems. Proceedings of the 18th international con-
Construcción 2015;65(320):071.
ference on composite materials. Jeju Island, Korea. 2011.
[32] Ye H, Feng N, Ling-hu Y, et al. Research on fire resistance of ultra-high-performance
[3] Wrigley RG. Permanent formwork in construction. CIRIA. 2001.
concrete. Adv Mater Sci Eng 2012;2012.
[4] Han LH, Xu CY, Tao Z. Performance of concrete filled stainless steel tubular (CFSST)
[33] Graybeal B, Tanesi J. Durability of an ultrahigh-performance concrete. J Mater Civ
columns and joints: Summary of recent research. J Constr Steel Res
Eng 2007;19(10):848–54.
2019;152:117–31.
[34] Magureanu C, Sosa I, Negrutiu C, et al. Mechanical properties and durability of
[5] Yang Wei, Cheng Jiang, Yufei Wu. Confinement effectiveness of circular concrete-
ultra-high-performance concrete. ACI Mater J 2012;109(2):177.
filled steel tubular columns under axial compression. J Construct Steel Res
[35] Al-Osta MA, Isa MN, Baluch MH, et al. Flexural behavior of reinforced concrete
2019;158(7):15–27.
beams strengthened with ultra-high performance fiber reinforced concrete. Constr
[6] Abed F, AlHamaydeh M, Abdalla S. Experimental and numerical investigations of
Build Mater 2017;134:279–96.
the compressive behavior of concrete filled steel tubes (CFSTs). J Constr Steel Res
[36] Massicotte B, Dagenais MA, Lagier F. Performance of UHPFRC jackets for the
2013;80:429–39.
seismic strengthening of bridge piers. Proceeding of the 2nd Int Symp on UHPFRC
[7] Yirui Zhang, Yang Wei, Jiawen Bai, Yongxing Zhang. Stress-strain model of an FRP-
Designing and building with UHPFRC: from innovation to large-scale realizations.
confined concrete filled steel tube under axial compression. Thin-Walled Struct
Marseille, France. 2013. p. 1–11.
2019;142(9):149–59.
[37] Farzad M, Shafieifar M, Azizinamini A. Retrofitting of bridge columns using UHPC.
[8] Wu Y-F, Liu T, Oehlers D. Fundamental principles that govern retrofitting of re-
J Bridge Eng 2019;24(12):04019121.
inforced concrete columns by steel and FRP jacketing. Adv Struct Eng
[38] Khan MI, Al-Osta MA, Ahmad S, et al. Seismic behavior of beam-column joints
2006;9(4):507–33.
strengthened with ultra-high performance fiber reinforced concrete. Compos Struct
[9] Ozbakkaloglu T, Saatcioglu M. Seismic performance of square high-strength con-
2018;200:103–19.
crete columns in FRP stay-in-place formwork. J Struct Eng 2007;133(1):44–56.
[39] Zhou Zhen, Zhang Yi, Wang Yongquan, et al. Experimental study on tensile me-
[10] Yufei Wu, Yang Wei. General stress-Strain model for steel- and FRP-confined con-
chanical property of grid reinforced UHPC plates. J Southeast Univ (Natural Science
crete. J Compos Constr 2015;19(4):040140691–0401406914.
Edition) 2019;4:1. [in Chinese].
[11] Wei Y, Wu G, Li G. Performance of circular concrete-filled fiber-reinforced polymer-
[40] Code of China. ‘‘Standard for evaluation of concrete compressive strength”. GB/
steel composite tube columns under axial compression. J Reinf Plast Compos
T50107-2010. Beijing: China Architecture & Building Press; 2010. [in Chinese].
2014;33(20):1911–28.
[41] Shin HO, Min KH, Mitchell D. Confinement of ultra-high-performance fiber re-
[12] Tian H, Zhou Z, Wei Y, et al. Experimental investigation on axial compressive be-
inforced concrete columns. Compos Struct 2017;176:124–42.
havior of ultra-high performance concrete (UHPC) filled glass FRP tubes. Constr
[42] Hosinieh MM, Aoude H, Cook WD, et al. Behavior of ultra-high performance fiber
Build Mater 2019;225:678–91.
reinforced concrete columns under pure axial loading. Eng Struct 2015;99:388–401.
[13] Wu YF, Jiang C. Effect of load eccentricity on the stress–strain relationship of FRP-
[43] Hassan AMT, Jones SW, Mahmud GH. Experimental test methods to determine the
confined concrete columns. Compos Struct 2013;98:228–41.
uniaxial tensile and compressive behaviour of ultra high performance fibre re-
[14] Tian HW, Zhou Z, Lu JP, Han FY. Effects of steel fiber content on axial compression
inforced concrete (UHPFRC). Constr Build Mater 2012;37:874–82.
performance of UHPC filled FRP tubes. J Southeast Univ (Natural Science Edition)
[44] Su Y, Li J, Wu C, et al. Mesoscale study of steel fibre-reinforced ultra-high perfor-
2019;49(3):481–7. [in Chinese].
mance concrete under static and dynamic loads. Mater Des 2017;116:340–51.
[15] Pan Y, Guo R, Li H, et al. Analysis-oriented stress–strain model for FRP-confined
[45] Liang X, Wu C. Meso-scale modelling of steel fibre reinforced concrete with high
concrete with preload. Compos Struct 2017;166:57–67.
strength. Constr Build Mater 2018;165:187–98.
[16] Wei YY, Wu YF. Unified stress–strain model of concrete for FRP-confined columns.
[46] Huiwen TIAN, Zhen ZHOU, Jiping LU, Zhen PENG. Meso-scale numerical simula-
Constr Build Mater 2012;26(1):381–92.
tion of axial compression performance of fiber reinforced polymer-confined ultra-
[17] Zhang B, Yu T, Teng JG. Behavior of concrete-filled FRP tubes under cyclic axial
high performance concrete. Acta Materiae Compositae Sinica 2019;36. [in
compression. J Compos Constr 2015;19(3):0401.
Chinese].
[18] Hu YJ, Jiang C, Liu W, et al. Degradation of the in-plane shear modulus of structural
[47] Xu S, Wu C, Liu Z, et al. Numerical study of ultra-high-performance steel fibre–r-
BFRP laminates due to high temperature. Sensors 2018;18(10):3361.
einforced concrete columns under monotonic push loading. Adv Struct Eng
[19] Papanicolaou CG, Papantoniou IC. Mechanical behavior of textile reinforced con-
2018;21(8):1234–48.
crete (TRC)/concrete composite elements. J Adv Concr Technol 2010;8(1):35–47.
[48] Malvar LJ, Crawford JE, Morrill KB. K&C Concrete model release III, automated
[20] Brameschuber W, Koster M, Hegger J, et al. Integrated formwork elements made of
generation of material model input. Karagozian & Case Technical Report TR-99-24-
textile-reinforced concrete. Special Publ 2008;250:35–48.
B1. Glendale, Karagozian & Case Structural Engineers, 2000: 1–17.
[21] Shen L, Xu S, Wang J. Mechanical behaviour of TRC thin-plates exposed to high
[49] EN B S. 206-1: 2000. Concrete, Specification, Performance, Production and
temperature: experimental study. Mag Concr Res 2015;67(21):1135–49.
Conformity, 2001.
[22] Xu S, Shen L, Wang J. The high-temperature resistance performance of TRC thin-
[50] Xu M, Wille K. Calibration of K&C concrete model for UHPC in LS-DYNA. In:
plates with different cementitious materials: Experimental study. Constr Build
Advanced materials research, vol. 1081. Trans Tech Publications, 2015, p. 254–9.
Mater 2016;115:506–19.
[51] Kusumawardaningsih Y, Fehling E, Ismail M. UHPC compressive strength test
[23] Qiao Z, Pan Z, Xue W, et al. Experimental study on flexural behavior of ECC/RC
specimens: Cylinder or cube? Procedia Eng 2015;125:1076–80.
composite beams with U-shaped ECC permanent formwork. Front Struct Civil Eng
2019;13(5):1271–87.

13

View publication stats