You are on page 1of 15

Engineering Structures 248 (2021) 113266

Contents lists available at ScienceDirect

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

Flexural and cracking behaviors of reinforced UHPC beams with various


reinforcement ratios and fiber contents
Zheng Feng a, b, c, Chuanxi Li a, b, *, Doo-Yeol Yoo c, *, Rensheng Pan b, Jun He b, Lu Ke d
a
School of Civil Engineering, Changsha University of Science and Technology, Changsha, China
b
Key Laboratory of Bridge Engineering Safety Control by Department of Education, Changsha University of Science and Technology, Changsha, China
c
Department of Architectural Engineering, Hanyang University, 222 Wangsimni-ro, Seongdong-gu, Seoul 04763, Republic of Korea
d
School of Civil Engineering and Architecture, Guangxi University, Nanning 530004, China

A R T I C L E I N F O A B S T R A C T

Keywords: This research is aimed at investigating the flexural and cracking behaviors of ultra-high-performance concrete
Ultra-high-performance concrete (UHPC) (UHPC) beams. Nine UHPC beams with different reinforcement ratios (0, 1.0%, 2.9%, 4.8%, and 7.1%) and fiber
Flexural behavior volume fractions (2.0% and 3.0%) were considered under flexure. A section analysis was also performed to
Cracking behavior
predict the flexural and cracking behaviors of UHPC beams and was verified by experimental results. The test
Sectional analysis
Reinforcement ratio
results showed that the reinforcement and steel fibers can play a significant role in limiting crack development.
Fiber content Considering the performance of the steel fibers, a high-precision equation for predicting the average crack
spacing of UHPC beams was developed. The formulas in the French standard NF P 18–710 overestimated the
maximum crack width by a significant margin for the UHPC beams with a high reinforcement ratio (>4%). The
flexural stiffness of the reinforced UHPC beams increased as the reinforcement ratio increased, whereas their
initial stiffness without reinforcement was larger than that with a low reinforcement ratio of 1.0% because of the
weak bond interface between the UHPC and the reinforcement. The flexural capacity of the UHPC beams
basically increased linearly with the reinforcement ratios. The result of the sectional analysis indicated that the
contribution of steel fibers to the flexural capacity decreased significantly with the increase in the reinforcement
ratio. In the ultimate limit states design, the contribution of UHPC tensile capacity could be considered only as a
safety reserve when the reinforcement ratio is greater than 4.0%, whereas its contribution needs to be considered
for the UHPC beams with a reinforcement ratio of less than 2.9%.

1. Introduction guidelines or codes [10–14].


To further improve the design theory of UHPC beams, researchers
Ultra-high-performance concrete (UHPC) is a type of cement-based have conducted numerous studies on the flexural performance of UHPC
composite that has been increasingly used in bridge engineering [1–3] [7–9,15–27]; however, most of them focused on the flexural behaviors
owing to its high strength, high toughness, and excellent durability of the UHPC beams only in the member scale, in which the steel re­
[4–6]. The superior characteristics of UHPC can be achieved by opti­ inforcements haven’t been considered [7–9,18], and limited experi­
mizing the size and distribution of the constituent ingredients and mental results are available on the flexural and cracking behaviors of
incorporating a high volume of high-strength steel fibers. As is known, reinforced UHPC (R-UHPC) beams at the structural member scale.
the flexural behavior of beams is one of the most important mechanical Graybeal and Chen [15,16] carried out a series of prestressing UHPC I-
properties of bridge structures. Thus, research on the flexural perfor­ girder bending tests and experimentally verified that the use of UHPC
mance of UHPC beams is a key factor directly related to the application can significantly enhance the flexural capacity of concrete beams.
of long-span UHPC girder bridges. In addition, recent studies have Additionally, numerous studies have shown that steel fibers can improve
indicated that the flexural performance of UHPC beams is closely related the bearing capacity, post-cracking stiffness, and cracking behavior of
to the direct tensile strength of UHPC [7–9], which is one of the most UHPC beams [17,18]. Zhen et al. [19] conducted flexural tests on UHPC
important mechanical properties of UHPC structures in existing design that exhibits strain-softening behavior with different reinforcement

* Corresponding authors at: School of Civil Engineering, Changsha University of Science and Technology, Changsha, China (C. Li).
E-mail addresses: lichx@csust.edu.cn (C. Li), dyyoo@hanyang.ac.kr (D.-Y. Yoo).

https://doi.org/10.1016/j.engstruct.2021.113266
Received 7 May 2021; Received in revised form 5 August 2021; Accepted 20 September 2021
Available online 27 September 2021
0141-0296/© 2021 Elsevier Ltd. All rights reserved.
Z. Feng et al. Engineering Structures 248 (2021) 113266

Fig. 1. Cross-section and reinforcement arrangement of the R-UHPC beams.

ratios and concluded that 25% of the tensile strength can be considered (≥4.0%) have been seldom considered in existing R-UHPC flexural tests
in the flexural design. Li et al. [20] noted that the tensile strength of [15–27]. For instance, Qi et al. [21] analyzed the flexural impact of
UHPC contributes minimally to the flexural capacity and can be ignored UHPC beams with different fiber volume fractions (0.1%–2.0%) and
in cases where the flexural strength of UHPC is only 21.7 MPa. From the longitudinal reinforcement ratios (0, 0.7%, and 1.57%), and Shafieifar
above analysis, researchers have different understandings of the et al. [23] carried out a bending performance test on UHPC beams with a
contribution of UHPC tensile strength. This difference is largely because 2.0% fiber volume fraction and 0%–2.6% longitudinal reinforcement
few tests have considered the effect of reinforcement ratio and steel fiber ratios. Yoo et al. [17,24] experimentally studied 14 UHPC beams with
content simultaneously. In fact, the contribution of UHPC tensile different reinforcement ratios (0%–1.71%) and fiber types, where their
strength to the flexural behavior of R-UHPC beams is significantly test results indicated that the number of cracks and average crack
affected by the reinforcement ratio. spacings were marginally affected by the reinforcement ratio. The stress
Owning to the excellent mechanical properties of the UHPC, the response of R-UHPC beams with a low reinforcement ratio compared
large-span UHPC box-girder bridge can decrease the superstructure with that of R-UHPC beams with a high reinforcement ratio may be
weight by approximately 50% compared with that of the conventional different under flexural loads. Besides, in the existing researches
prestressed concrete bridge [28]. For the long-span UHPC girder bridge, [15–27], the average crack spacing, crack propagation, and crack width
limiting the height and size of the cross-section to reduce the super­ prediction of UHPC beams have not been well researched. To sum up, it
structure weight is necessary, which can significantly reduce the UHPC is important to study the flexural and cracking behaviors of R-UHPC
material consumption as well. In addition to long-span UHPC bridges, beams by considering different reinforcement ratios and fiber volume
for the urban UHPC flyover bridges, it is also necessary to control the fractions simultaneously.
sectional height of the beams. As a result, it is indispensable to use a high The objective of this research is to investigate the flexural and
reinforcement ratio for these UHPC girder bridges to ensure an excellent cracking behaviors of R-UHPC beams by considering the UHPC tensile
mechanical performance. Importantly, thanks to the high compression strength. For this purpose, eight R-UHPC beams with different rein­
strength of UHPC, the possibility of crushing failure in the compression forcement ratios (1.0%, 2.9%, 4.8%, and 7.1%) and fiber volume frac­
zone of UHPC beam can be much smaller than that of the normal con­ tions (2.0% and 3.0%) were tested under a four-point bending test setup.
crete beam with the same condition, which makes the UHPC beam can A reference beam made of UHPC without steel reinforcement was also
use higher reinforcement ratio. For these reasons, economical and tested for comparison. The flexural and cracking behaviors were dis­
practical structure sizes may be obtained by using a high reinforcement cussed in terms of crack pattern, crack spacing, crack width, cracking
ratio in UHPC beams. Therefore, the study of the flexural behavior of the moment, flexural moment, moment–deflection relationship, and failure
UHPC beams with a high reinforcement ratio is essential. In addition, mode. In addition, section analysis for predicting the flexural and
the flexural performance and fracture energy of UHPC increased almost cracking behaviors of the UHPC beams was performed and verified by
linearly with the fiber content [29], and thus the higher fiber contents in experimental results.
R-UHPC beams may have excellent mechanical properties. However, at
present, high fiber contents (≥3.0%) and high reinforcement ratios

Table 1
Parameters and test results of the UHPC beams.
Name ρf ρs As (mm2) cs (mm) a(mm) Pcr Py Pp Mtcr Mtu Failure mode Average crack spacing lcr,exp (mm)
(kN) (kN) (kN) (kN.m) (kN.m)

B20R1 2.0% 1.0% 402 26 34 30 108 155 14.25 73.63 Flexural a 105
B20R3 2.0% 2.9% 1,140 26 38 39 275 326 18.53 154.85 Flexural b 75
B20R5 2.0% 4.8% 1,742 26 60 45 410 477 21.38 226.58 Flexural b 63
B20R7 2.0% 7.1% 2,463 26 68 50 611 662 23.75 314.45 Flexural b 66
B30N 3.0% \ \ \ 35 \ 78 16.63 37.05 Flexural c \
B30R1 3.0% 1.0% 402 26 34 35 115 195 16.63 92.63 Flexural a 92
B30R3 3.0% 2.9% 1,140 26 38 40 290 358 19.00 170.05 Flexural b 68
B30R5 3.0% 4.8% 1,742 26 60 55 413 485 26.13 230.38 Flexural b 65
B30R7 3.0% 7.1% 2,463 26 68 65 564 636 30.88 302.10 Flexural b 61

Note: ρf is the steel fiber volume fraction; ρs is the tensioned reinforcement ratio; As is the total area of the tensioned reinforcement; cs is the thickness of the concrete
cover; a is the centroid point of all tensioned reinforcements to the bottom edge of the cross section; Pcr is the first cracking load; Py refers to the yielding load of the
t
tensioned reinforcement; Pp is the peak load; Mcr and Mut refer to the first cracking and ultimate flexural moments of UHPC beam, respectively; and lcr,exp is the average
crack spacing on the bottom of the surface in the pure bending region.
a
Flexural failure occurs with tensioned reinforcement rupture.
b
Flexural failure occurs with UHPC crushing at the compression zone.
c
Flexural failure occurs with UHPC rupture at the tension zone.

2
Z. Feng et al. Engineering Structures 248 (2021) 113266

Table 2 100 mm × 100 mm), six prism compression specimens (100 mm × 100
Mechanical properties of HRB400 reinforcement. mm × 300 mm), three flexural specimens (100 mm × 100 mm × 400
Diameter Area Yield Yield Elastic Ultimate mm), and nine direct tensile dog-bone-shaped specimens with a cross-
ds (mm) As1 strength strain modulus strength sectional area of 50 mm × 50 mm and a length of 368 mm were pre­
(mm2) fy (MPa) εy (με) Es (GPa) σul (MPa) pared to obtain the UHPC material properties [31]. All the specimens
10 78.5 483.6 2,348 206 640.2 were cast and cured under the same conditions as those of the beams.
12 113.1 481.4 2,337 206 572.9 First, the UHPC beams were cured in a natural environment for 48 h.
16 201.1 485.5 2,357 206 622.6 After demolding, the samples were cured in a high-temperature steam at
22 380.1 494.1 2,423 206 616.8 a temperature of 90 ± 3 ◦ C for 48 h. Subsequently, the UHPC specimens
25 490.9 495.5 2,405 206 608.4
28 615.8 499.3 2,395 206 598.2
were continuously placed for 14 days. Based on the experimental test
results, the material properties were obtained, as shown in Table 4.
Fig. 2 shows the tensile stress–strain curves of the UHPC dog-bone-
shaped specimens with different fiber contents. According to Table 4
Table 3
and Fig. 2, the ultimate tensile strength σpc of the UHPC with a 2.0%
Mixture proportion of UHPC (by weight ratio to cement).
fiber volume fraction is 7.9 MPa, which is greater than the initial
Cementitious material Quartz sand Steel fibre Water cracking strength σcc (6.8 MPa). Similarly, for the UHPC with a 3.0%
Cement Silica fume Mineral powder fiber volume fraction, its ultimate tensile strength (10.5 MPa) is also
1.00 0.37 0.25 1.10 0.36 0.30 greater than the initial cracking strength (7.4 MPa). Thus, the UHPC
material in this study exhibited strain-hardening characteristics under
tension.
2. Experimental program
2.3. Flexural tests setup of the UHPC beams
2.1. Test parameters and specimen details
A four-point bending test with a pure bending length of 800 mm was
In this test, rectangular beams with section sizes of 300 mm × 150
performed on the UHPC beams, as shown in Fig. 3. To obtain the
mm × 3,000 mm were used to perform the experimental study. To
deformation of the UHPC beams, five linear variable displacement
reflect the effect of fiber content on the tensile and flexural behaviors of
transducers were arranged at the mid-span, loading points, and sup­
UHPC, steel fiber volume fractions of 2.0% and 3.0% were selected. The
porting positions. The strain gauges attached to the side surfaces of the
longitudinal reinforcement ratios ρs used were 0, 1.0%, 2.9%, 4.8%, and
beams were used to measure the strains. A crack observation instrument
7.1%. The stirrup spacing of the test beam was 200 mm. The cross
with a measurement accuracy of 0.01 mm was used to measure the crack
section and the reinforcement arrangement of the R-UHPC beams are
width. In the initial loading stage, 5 kN was used as the first loading
shown in Fig. 1, and the design parameters are summarized in Table 1.
interval until the cracking load was reached. Then, 10–30 kN was used
The thickness of the concrete cover of the reinforced UHPC beam was 26
as the second loading interval (adjusted according to the predicted
mm. In Table 1, the names of the UHPC beams are indicated by the fiber
bearing capacity). When the applied load reached 90% of the predicted
contents and reinforcement ratios. For example, B20 stands for the fiber
ultimate load, the displacement in the mid-span was used to control the
volume fraction of 2.0%; B30 stands for the fiber volume fraction of
loading rate. The test ended once the load value was lower than 85% of
3.0%; N represents the UHPC beam without reinforcement; and R1, R3,
the ultimate load. During the loading process, every loading stage was
R5 and R7 represent the UHPC beams with the reinforcement ratio of
held for 5 min and the number of cracks, crack width, load, deflections,
1.0%, 2.9%, 4.8% and 7.1%, respectively. The steel reinforcements in
and strains were recorded simultaneously.
the test had a strength grade of HRB400, which is a typical steel rein­
forcement used for bridges in China. The mechanical properties of the
3. Experimental results and analysis
steel reinforcement are shown in Table 2.
3.1. Cracking behaviors and failure patterns
2.2. Materials properties
When the applied load reached the cracking load, the first visible
The UHPC dry mixture used in this study was composed of cemen­ crack appeared at the bottom of the beam in the pure bending region,
titious materials (a mix of Portland cement, silica fume, and mineral whose crack width was approximately 0.02 mm. With the increase in the
powder) and quartz sand [30]. The detailed proportion of the mixture is applied load, crack initiation occurred in the form of sparsely distributed
summarized in Table 3. The water-to-cementitious material ratio of the microcracks, and cracks gradually developed from the bottom to the top
UHPC dry mixture was 0.18, and a hooked steel fiber was adopted with a of the beam. This stage is known as the dispersion-crack stage, where the
diameter of 0.2 mm, length of 16 mm, and tensile strength of 2800 MPa. crack-bridging behavior provided by the steel fibers plays an important
To improve the workability during casting, a type of polycarboxylate role in resisting cracking. As shown in Fig. 4, with the increase in the
superplasticizer with a water reducing efficiency of 30% and a mass applied load, the crack width almost linearly increased in this stage,
fraction of cement of 4% was used. All the test beams were cast from the where steel fibers were continuously being pulled out from the matrix
middle of the span to the ends. accompanied by an audible sound. It should be mentioned that the crack
For the UHPC beams, six cube compression specimens (100 mm × width of the R-UHPC beams with a 3.0% fiber volume fraction was

Table 4
Mechanical properties of the UHPC materials (average value).
Fibre volume fractionρf fcu fcf Ec σcc σpc εpc fc εcu
(MPa) (MPa) (GPa) (MPa) (MPa) (με) (MPa) (με)

2.0% 146.8 32.6 45.1 6.8 7.9 2,705 125.4 4,088


3.0% 154.6 37.9 47.4 7.4 10.5 3,048 128.4 4,313

Note: fcu = cubic compression strength of UHPC; fcf = flexural strength of UHPC; Ec = elastic modulus of UHPC; σ cc = initial cracking strength;σpc = ultimate tensile
strength; εpc = tensile strain capacity; f c = ultimate prism compressive strength; εcu = ultimate compressive strain.

3
Z. Feng et al. Engineering Structures 248 (2021) 113266

Fig. 2. Tensile stress–strain curves of the UHPC with different fiber contents: (a) 2.0% fiber volume fraction, and (b) 3.0% fiber volume fraction.

Fig. 3. Details of the UHPC beams and test setup: (a) test setup and (b) photo of UHPC bending test.

smaller than that of the R-UHPC beams with a 2.0% fiber volume frac­ a few microcracks started to be localized, entering the reinforcement-
tion under the same load. Thus, increasing the fiber content can enhance yield stage. In this stage, the width of the localized cracks continu­
the resistance to crack width increase in UHPC beams. Upon further ously increased with the applied load, whereas the crack number and
increasing the applied load, the longitudinal reinforcement yielded and crack spacing changed slightly.

4
Z. Feng et al. Engineering Structures 248 (2021) 113266

Fig. 4. Load–crack width curves of the UHPC beams: (a) all members and (b) B30R1.

Fig. 5 shows the cracking and failure patterns of all UHPC beams in 3.3. Load–strain relationship
the pure bending region. As shown in the figure, the localized main
cracks of all the beams were located in the pure bending region. For the Fig. 8 shows the load–strain curves along the side surface of the
beam B30N, only one main crack occurred when it was damaged. Other UHPC beams with a 3.0% steel fiber content. Before the load reached the
microcracks were distributed near this localized crack. For the beams peak value, the strain of the UHPC beams was almost linearly distributed
B20R1 and B30R1, the average spacing of the cracks at the bottom of the along the height of the beam, which shows that the cross section of the
UHPC beams in the pure bending region was more than 90 mm (see UHPC beams always remains plane during the bending test, as was also
Table 1). However, for the remaining R-UHPC beams, the average crack reported in other studies [19–26].
spacing was less than 80 mm. With the increase in the reinforcement
ratio, the cracks became increasingly finer and denser, and the number 3.4. Cracking and ultimate moments
of main cracks increased gradually, showing the typical characteristics
of multiple cracks. The cracking, yield, peak load, and failure mode of the UHPC beams
The failure modes of the UHPC beams with different reinforcement are summarized in Table 1. Fig. 9 shows the comparison results of the
ratios are given in Table 1. As shown in the table, all the UHPC beams cracking moment and ultimate moment of the UHPC beams with
exhibited flexural failure. For the UHPC beams without reinforcement, different fiber volume fractions and reinforcement ratios. The calculated
flexural failure occurred with the fracture of the UHPC in the tension equation of the flexural moment is shown as follows:
zone. The R-UHPC beams with a 1.0% reinforcement ratio failed because
P
of the tensioned reinforcement rupture. For the R-UHPC beams with M = la (1)
2
reinforcement ratios from 2.9% to 7.1%, after the tensioned re­
inforcements yielded, flexural failure occurred, with the UHPC crushing where P is the applied load and la is the length between the support
in the compression zone. This indicates that improving the compressive and the loading points (=950 mm).
strength of UHPC can effectively enhance the structural performance of As shown in Table 1 and Fig. 9a, for the UHPC beam with a 3.0% steel
the R-UHPC beam with a high reinforcement ratio. fiber content, the cracking moment increased by 14.3% when the rein­
forcement ratio increased from 1.0% to 2.9%, and the cracking moment
3.2. Load mid-span deflection relationship increased by 37.5% when the reinforcement ratio increased from 2.9%
to 4.8%. However, the cracking moment remained unchanged when the
Figs. 6 and 7 show the load versus mid-span deflection curves of the reinforcement ratio increased from 0% to 1.0%. This indicates that
UHPC beams. According to the results in Fig. 6, the flexural stiffness of increasing the reinforcement ratio can effectively increase the cracking
the R-UHPC beams increased with the increase in the reinforcement moment of the R-UHPC beam and that the cracking moment is related to
ratio. In addition, the flexural stiffness of the R-UHPC beams can be the stiffness of the UHPC beam. In addition, the increase in steel fiber
improved by increasing the steel fiber content at low reinforcement content can also improve the cracking moment of the UHPC beams.
ratios. However, when the reinforcement ratio was up to 2.9%, the steel As shown in Table 1 and Fig. 9b, the ultimate moment of the UHPC
fiber content had no significant effect on the flexural stiffness of the R- beams increased linearly with the reinforcement ratio. The ultimate
UHPC beams. It is interesting to note that the initial flexural stiffness of moment of beam B30R1 was 150% higher than that of beam B30N. For
the UHPC beam without reinforcement (B30N) was larger than that of the UHPC beam with a 1.0% reinforcement ratio, the ultimate moment
B30R1 (see Fig. 7), as was also reported in another study [21]. The increased by 25.8% when the fiber volume fraction increased from 2.0%
reason is that, although the elastic modulus of the steel reinforcement is to 3.0%. For the UHPC beam with a 2.9% reinforcement ratio, the ul­
much greater than that of the UHPC matrix, the bond interface between timate moment increased by 9.8% as the fiber volume fraction
the reinforcement and the UHPC reduces the stiffness of the UHPC increased. However, compared with the ultimate moment of beam
beams. For the R-UHPC beam B30R1, the weakened effect of the bond B20R5, that of beam B30R5 increased by only 1.7%. Therefore, it is
interface was greater than that of the stiffness enhancement of the steel concluded that the fiber volume fraction has a significant impact on the
reinforcement itself. flexural capacity of the UHPC beams with a low reinforcement ratio (less
than2.9%), whereas the enhancement of the steel fiber is not apparent
for the UHPC beams with a high reinforcement ratio.
As we know, the price of UHPC is much higher than that of normal
concrete. Thus, ensuring the economy of UHPC structure is an extremely

5
Z. Feng et al. Engineering Structures 248 (2021) 113266

Fig. 5. Crack distribution of the UHPC beams.

6
Z. Feng et al. Engineering Structures 248 (2021) 113266

elastic–plastic constitutive law with an elastic modulus of 206 GPa.


ftk
εcr = (2)
Ec

fc
εce = (3)
Ec

Lf
εlim = (4)
4lc

2
lc = h (5)
3
where εcr and εce are the factored elastic tensile and compression
strain, respectively; ftk is the average value of the ultimate tensile stress
σ pc and the elastic stress σcc ; fc is the compression strength; εlim is the
limited tensile strain at a stress value of 0; lc is the characteristic length; h
is the height of the UHPC beam; and Lf is the steel fiber length.

Fig. 6. Load mid-span deflection curve of the R-UHPC beams.


4.2. Analytical model

important goal. Based on the test results, the high tensile strength of To predict the flexural behavior of the R-UHPC beams, a sectional
UHPC has a significant positive impact on the flexural capacity of the analysis was performed. The UHPC beam without steel reinforcement
UHPC beams, and thus using a low reinforcement ratio in small-span was also modeled using sectional analysis. Assuming that the cross
UHPC girder bridge is appropriate. And as mentioned above, section of the UHPC beam always remains in the plane during the
increasing the reinforcement ratio can improve the flexural stiffness, bending test, the compressive and tensile strain at the edge of the cross
cracking and ultimate moments and cracking resistance of the UHPC section can be calculated, as illustrated in Eqs. (6) and (7). The stress in
beams. Therefore, the use of high reinforcement ratios in long-span the steel reinforcement was calculated using a predefined elastic–plastic
UHPC girder bridges can result in more economical section sizes, constitutive model. The strain in the steel reinforcement was considered
which is an important factor to ensure the lightweight of the bridge. as the strain in the surrounding concrete, which was calculated using
Eqs. (8) and (9). According to the constitutive model under tension and
4. Numerical analysis for predicting the flexural capacity of compression, the stress and strain distributions in the cross section of the
UHPC beams UHPC beam can be obtained, as shown in Fig. 11. The cross section of
the UHPC beam was divided into a number of layers along the height, as
4.1. Constitutive model was also done in [17].

Various forms of uniaxial compressive and tension stress–strain re­ εc = ϕ⋅x (6)
lationships for UHPC have been established in various codes or guide­
εt = ϕ⋅(h − x) (7)
lines [10–14]. In this study, the compression stress–strain relationships
of UHPC complied with the constitutive model recommended by the
ε’s = ϕ⋅(x − d) (8)
specification of the Japan Society of Civil Engineers (JSCE) [12], as
shown in Fig. 10. The ultimate tensile strain εpc and the ultimate
εs = ϕ⋅(h − x − a) (9)
compressive strain εcu shown in the figure were obtained experimen­
tally, and the other strain values, such as εcr , εce , and εlim , were calcu­ where εt and εc are the tensile strain at the bottom edge and the
lated using Eqs. (2)–(5). The values of the UHPC constitutive model compressive strain at the top edge of the cross section, respectively; εs is
under different fiber volume fractions are shown in Table 5. The the strain of the tensioned reinforcement; εs is the strain of the steel

constitutive model for the steel reinforcements was defined using an reinforcement in the compression zone; x is the neutral axis depth; ϕ is

Fig. 7. Load mid-span deflection curves of the UHPC beams with 3.0% fibre volume fractions.

7
Z. Feng et al. Engineering Structures 248 (2021) 113266

Fig. 8. Load-strain curves along the mid-span section of the UHPC beams with 3.0% steel fibre content: (a) B30N, (b) B30R1, (c) B30R3, (d) B30R5, and (e) B30R7.

8
Z. Feng et al. Engineering Structures 248 (2021) 113266

Fig. 9. Comparative analysis of UHPC beams; (a) cracking moment, and (b) ultimate moment.

calculation was repeated until the ultimate compression strain εcu or the
limited tensile strain εlim of the UHPC was reached. The algorithm for
the sectional analysis of the UHPC beams is illustrated in Fig. 12.
∫ ∫
fu dAc + fs dAs = 0 (10)
Au As

∫ ∫
Mu = fu ydAc + fs ydAs (11)
Au As

where fu is the stress at each layer of the UHPC, fs is the stress in the
steel reinforcement, and y is the distance from the neutral axis.
To convert the calculated moment–curvature curves obtained from
the sectional analysis to the moment–deflection curves, the following
Eq. (12) was simply adopted [17,32] on the basis of linear elastic theory.
M 24
ϕ= = Δ (12)
Ec I 3L2 − 4l2a

where I is the inertia moment of the UHPC beam, L is the span length
(=2700 mm), and Δ is the mid-span deflection.
Fig. 10. Constitutive model of UHPC in compression and tension.

4.3. Experimental and numerical results from the sectional analysis


the curvature; and d is the centroid point of the compression rein­
forcement to the top edge of the cross section.
Fig. 13 shows the experimental and numerical ultimate moment re­
In Fig. 11, Cu is the compression force of the UHPC, Cs is the
sults of the UHPC beams. As shown in this figure, for the R-UHPC beams,
compression force of the steel reinforcement in the compression zone, Tu
the predicted ultimate moment obtained by the proposed analysis model
is the tension force of the UHPC, and Ts is the force of the tensioned
has an excellent prediction accuracy, in which the average difference
reinforcement.
between the predicted and the experimental ultimate moment was only
After determining the strain distribution, the stress at each layer in
0.9% and the coefficient of variation was 0.087. Fig. 14 plots the
the cross section was calculated using the UHPC stress–strain constitu­
experimental and numerical moment–deflection curves of the UHPC
tive model. On the basis of the strain distribution and cross-section force
beams. On the whole, the numerical moment–deflection curves exhibi­
equilibrium condition, the sectional compressive and tensile forces can
ted good agreement with the experimental results for the R-UHPC
be calculated as Eq. (10). To satisfy Eq. (10), iterative calculations were
beams. However, it should be mentioned that the flexural stiffness of the
performed to obtain a perfect neutral axis depth x and then the flexural
UHPC beams in the numerical model was slightly more than the
moment Mu at a given curvature was calculated by Eq. (11). When εc =
experimental results, particularly for the UHPC beams with a low rein­
εcu and εt = εlim are reached at the same time, the characteristic area of
forcement ratio or without reinforcement. The reason for this discrep­
the tensioned reinforcement [As ] can be calculated using Eq. (10). The
ancy is that the UHPC material is assumed to be homogeneous and has

Table 5
Basic parameters of the UHPC constitutive model.
Fibre volume fraction Ec fc εce εcu ftk εcr εpc εlim
(GPa) (MPa) (με) (με) (MPa) (με) (με) (mm)

2.0% 45.1 125.4 2,780 4,088 7.4 156 2,705 0.01


3.0% 47.4 128.4 2,709 4,313 9.0 189 3,048 0.01

9
Z. Feng et al. Engineering Structures 248 (2021) 113266

Fig. 11. Stress and strain distribution along the depth of the R-UHPC beams.

equal strength anywhere in the numerical model. However, in the actual In contrast, the contribution of the UHPC tensile strength needs to be
test model, several flaws will appear during the casting and curing pe­ considered for the UHPC beams with a reinforcement ratio of less than
riods, which may be randomly distributed in the UHPC. Moreover, the 2.9%.
dispersion of the steel fibers is inhomogeneous. Therefore, for the UHPC
beams with a low reinforcement ratio or without reinforcement, the 5. Calculation method of the crack width and average crack
effect of the UHPC materials on the test results is significant, which spacing
results in an overestimation of the numerical predictions. Nevertheless,
for the R-UHPC beams with high reinforcement ratios, the steel rein­ 5.1. Average crack spacing of the UHPC beams
forcement can work collaboratively with UHPC and can reduce the ef­
fects of inhomogeneous dispersion and random orientation of steel fibers The average crack spacing lcr for normal concrete can be calculated
in the UHPC materials. In addition, as mentioned previously, the bond using the Chinese concrete structural code GB 50010–2010 [33], and the
interface between the reinforcement and the UHPC can decrease the related equations are shown in Eqs. (13)–(15).
stiffness of the R-UHPC beams. For the R-UHPC beam with a low rein­
deq
forcement ratio, the weakened effect of the bond interface is significant. lcr = 1.9cs + 0.08 (13)
pte
But in the numerical model, perfect interfacial bond properties between
the UHPC and steel reinforcement are considered, which also results in ∑ 2
ni d
an overestimation of the numerical predictions. deq = ∑ i (14)
ni di

4.4. Analysis of the contributions of the UHPC tensile strength As As


pte = = (15)
Ate hte b
According to Fig. 11, the flexural capacity of UHPC beams is
hte = min[2.5a, 0.5h] (16)
composed of compressive concrete (MC–UHPC), compressive steel bar
(MC–Steel bar), tension steel bar (MT–Steel bar), and tension concrete where lcr is the average crack spacing of the concrete; cs is the
(MT–UHPC). Based on the numerical model, the contribution of the thickness of the concrete cover; deq is the equivalent diameter of the
flexural strength of the R-UHPC beams with different reinforcement tensioned reinforcement; di is the diameter of the i type of tensioned
ratios is shown in Fig. 15. As shown in Fig. 15a, for the R-UHPC beams reinforcement; ni is the amount of the i type of tensioned reinforcement;
with 2.0 vol% of steel fibers, the contribution of the UHPC tensile ca­ and hte and Ate are the effective cross-sectional height and area around
pacity to the flexural capacity decreases from 67.4% to 5.7%, whereas the tensioned reinforcement, respectively [13].
the contribution of the tensioned reinforcement increases from 0 to However, compared with normal concrete, UHPC doped with high-
48.3% with the increase in the reinforcement ratio (0 to 7.1%). Simi­ strength steel fibers can restrain crack from propagating. In addition,
larly, the contribution of the UHPC tensile capacity to the flexural UHPC is a compact material that can bond well with the reinforcements,
strength decreases from 70.5% to 7.6% for the R-UHPC beams with 3.0 which can reduce the slip between the UHPC and the reinforcement and
vol% of steel fibers (see Fig. 15b). Therefore, with the increase in the shorten the stress transfer length. The experiment result of Hasgul et al.
reinforcement ratio, the contribution of the tensioned reinforcement to [27] indicated that the maximum crack widths in UHPC beams
the flexural capacity increases, whereas the contribution of the UHPC decreased by 41% at the serviceability limit states compared with that of
tensile capacity decreases significantly. non-fiber beams. Therefore, the actual crack spacing of UHPC will be
Fig. 16 shows the contribution ratio of the UHPC tensile capacity to smaller than the results calculated by Eq. (13), owing to the fiber
the flexural strength of the UHPC beams. It can be summarized that the bridging capabilities between the cracks. On the basis of our test out­
contribution proportion of the UHPC tensile capacity increases with the comes and the test results presented in [25,26], the calculation and
increase in steel fiber content. For the R-UHPC beam with a 1.0% experimental results of the average crack spacing of the UHPC beams are
reinforcement ratio, the proportion of the UHPC tensile capacity illustrated in Fig. 17. As shown in Fig. 17a, the difference between the
contribution is up to 33.7%. For the R-UHPC beam with a 2.9% rein­ experimental and the predicted results calculated by Eq. (13) was up to
forcement ratio, the proportion of the UHPC tensile capacity contribu­ 12.8%, which was consistent with the qualitative analysis results.
tion is more than 10%. Nevertheless, when the reinforcement ratio is Therefore, to correctly predict the average crack spacing lcr of the UHPC
more than 2.9%, the contribution of the UHPC tensile strength is basi­ beam under flexural load, the Eq. (13) was modified by considering the
cally less than 10%. As a result, in the ultimate limit states design, the enhancement of the steel fibers. The steel fiber characteristic parameter
contribution of the UHPC tensile capacity can be considered as a safety λf was used to describe the enhancement effect of the steel fibers.
reserve when the reinforcement ratio is greater than approximately 4%.

10
Z. Feng et al. Engineering Structures 248 (2021) 113266

Fig. 12. Algorithm for sectional analysis.

Furthermore, the contribution of steel fiber was corrected by consid­ fiber diameter.
ering the fiber influence coefficient ψ . Thus, the original Eq. (13) can be For minimizing the differences between the calculated value (lcr,cal )
modified into Eq. (17). and the experimental value (lcr,exp ), on the basis of the existing results in
Fig. 17a and the equation (17), the value of ψ of 0.07 can be obtained.
deq
lcr = (1.9cs + 0.08 ) × (1 − ψ λf ) (17) However, in this way, the coefficient of variation (COV) of lcr,cal /lcr,exp
pte
was up to 0.135, which is more than the results calculated by Eq. (13)
ρf lf (see Fig. 17a). Thus, the influences of discrepancy and COV on the
λf = (18) predicted results should be considered comprehensively, as shown in
df
Fig. 18, and eventually ψ of 0.05 was selected in Eq. (17) for predicting
where ψ is the fiber influence coefficient, λf is the characteristic the average crack spacing of UHPC beams more accurately. Fig. 17b lists
parameter of the steel fiber content, lf is the fiber length, and df is the the calculation and experimental results of the average crack spacing

11
Z. Feng et al. Engineering Structures 248 (2021) 113266

5.2. Maximum crack width of UHPC beams

In accordance with the French standard NF P 18–710 [13], at the


serviceability limit states, the maximum crack width at the bottom
surface of the UHPC beams (wu,max ) can be obtained as follows:

h− x− x
wu,max = ws,max (19)
h0 − x − x′

ws,max = lcr (εsm,f − εcm,f ) (20)


( ( ( ))
σ s σpc 1 σpc ) 1 Es
εsm,f − εcm,f = − − kt σcc − × + (21)
Es K⋅Ec Es K pte Ec
where K is the orientation factor of the fibers (=1.25), kt is a factor
that depends on the loading duration (=0.5 in this paper), x is the

height of the uncracked tensioned region, h0 is the effective height of the


Fig. 13. Comparison between the experimental and predicted ulti­ UHPC beam (=h − a), and (εsm,f − εcm,f ) is the average strain difference
mate moment. between the UHPC and the tensioned reinforcement.
On the basis of the sectional analysis in Fig. 11, Eqs. (19)–(21), and
based on the modified equation (17) with the ψ of 0.05. As shown in the the calculation procedure (changed from Fig. 12), the maximum crack
figure, there was a relatively good agreement between the calculation width at the bottom surface of the UHPC beams at the serviceability
results and the experimental results, in which the discrepancy was only limit states can be obtained. Table 6 shows the calculation parameters
3.5% and the COV was 0.127. Therefore, the modified equation (Eq. and the calculation results of the maximum crack width. As shown in the
(17)) for predicting the average crack spacing of UHPC beams is feasible. table, the calculation results matched well with the experimental results
It is worth mentioning that the performance of the steel fiber has been for the R-UHPC beams with reinforcement ratios of 1.0% and 2.9%. In
considered in Eq. (17), which can well reflect the performance of the contrast, there was a large gap between the calculation and the exper­
UHPC beams. imental results for the R-UHPC beams with reinforcement ratios of 4.8%
and 7.1%. This indicates that the formulas in NF P 18–710 [13]

Fig. 14. Comparison on the experimental and predicted moment–deflection curves.

12
Z. Feng et al. Engineering Structures 248 (2021) 113266

Fig. 15. Proportion of different flexural strength resistance mechanisms: (a) fibre volume fraction of 2.0%, and (b) fibre volume fraction of 3.0%.

Fig. 16. Contribution proportion of UHPC tensile strength on the flex­


ural capacity. Fig. 18. Variation curves of discrepancy and COV with fiber influence coeffi­
cient ψ .

Fig. 17. Calculation and experimental results of the average crack spacing: (a) calculated by Eq. (13), and (b) calculated by Eq. (17) (ψ = 0.05).

13
Z. Feng et al. Engineering Structures 248 (2021) 113266

Table 6
Calculation parameters as well as the results of the maximum crack width.
Name As (mm2) ρs a(mm) hte (mm) λf pte lcr,exp (mm) lcr,cal (mm) wu,max,exp (mm) wu,max,cal (mm) wu,max,cal -wu,max,exp wu,max,cal /wu,max,exp
(mm)

B20R1 402 1.0% 34 85 1.6 0.031 105 84 0.20 0.20 0.00 1.000
B20R3 1,140 2.9% 38 95 1.6 0.080 75 66 0.21 0.17 − 0.04 0.809
B20R5 1,742 4.8% 60 150 1.6 0.077 63 68 0.28 0.11 − 0.17 0.393
B20R7 2,463 7.1% 68 150 1.6 0.109 66 64 0.28 0.10 − 0.18 0.357
B30R1 402 1.0% 34 85 2.4 0.031 92 80 0.20 0.17 − 0.03 0.850
B30R3 1,140 2.9% 38 95 2.4 0.080 68 63 0.20 0.16 − 0.04 0.800
B30R5 1,742 4.8% 60 150 2.4 0.077 65 65 0.27 0.10 − 0.17 0.370
B30R7 2,463 7.1% 68 150 2.4 0.109 61 62 0.29 0.08 − 0.21 0.276

Note: the load for the serviceability limit states is conservatively taken as 60% of the peak load Pp [34]. Thus, wu,max,exp is the experimental maximum crack width on the
load of 60%Pp .

overestimated the maximum crack width by a significant margin for the the effect of steel fiber content should be considered as an
UHPC beams with high reinforcement ratios. That is to say, the formulas important factor.
in NF P 18–710 [13] overestimated the ability of the reinforcement to
limit crack propagation for the R-UHPC beams with high reinforcement Data Availability Statement
ratios. All data, models, and code generated or used during the study appear
in the submitted article.
6. Conclusion
CRediT authorship contribution statement
In this study, nine flexural tests on UHPC beams with different
reinforcement ratios and fiber volume fractions to investigate the flex­ Zheng Feng: Writing – original draft, Writing – review & editing,
ural and cracking behaviors of UHPC beams were conducted. In addi­ Methodology. Chuanxi Li: Conceptualization, Supervision. Doo-Yeol
tion, to predict the flexural and cracking behaviors of the UHPC beams, a Yoo: Writing – review & editing. Rensheng Pan: Writing – review &
section analysis was performed and verified by the experimental results. editing. Jun He: Validation. Lu Ke: Writing – review & editing.
The following conclusions are drawn:

(1) The reinforcement and steel fibers can play a significant role in Declaration of Competing Interest
limiting crack development. Considering the enhancement effect
of the steel fibers, a modified equation was developed for pre­ The authors declare that they have no known competing financial
dicting the average crack spacing of UHPC beams, and the interests or personal relationships that could have appeared to influence
calculated results were in good agreement with the experimental the work reported in this paper.
results. The formulas in NF P 18–710 overestimated the
maximum crack width by a significant margin for the UHPC
Acknowledgements
beams with a high reinforcement ratio (greater than4%).
(2) The UHPC beam without reinforcement failed owing to the UHPC
This research is supported by the National Natural Science Founda­
rupture in the tension zone. The R-UHPC beams with a 1.0%
tion Project of China (Grant NO.52078059, NO.51778069,
reinforcement ratio failed owing to the tensioned reinforcement
NO.51978081 and NO.51808055), the Horizon 2020- Marie Skło­
rupture, whereas the UHPC beams with reinforcement ratios from
dowska-Curie Individual Fellowship of European Commission (REUSE:
2.9% to 7.1% failed owing to the UHPC crushing in the
793787), and the Natural Science Foundation of Hunan Province, China
compression zone. Improving the compressive strength of UHPC
(Grant No. 2021JJ30712). In addition, the financial support from the
can improve the structural performance of the R-UHPC beams
China Scholarship Council (CSC) to the first author is gratefully
with a high reinforcement ratio.
acknowledge.
(3) The flexural stiffness and capacity of the R-UHPC beams
increased with the increase in the reinforcement ratio and steel
fiber content. The fiber volume fraction has a significant impact References
on the flexural capacity of the UHPC beams with a low rein­
[1] Blais PY, Couture M. Precast prestressed pedestrian bridge-world’s first reactive
forcement ratio (less than2.9%), whereas the enhancement of the powder concrete structure. PCI Journal 1999;44(5):60–71.
steel fiber is not apparent to flexural capacity for the UHPC beams [2] Zhou Mi, Lu W, Song J, Lee GC. Application of ultra-high performance concrete in
with a high reinforcement ratio. bridge engineering. Constr Build Mater 2018;186:1256–67.
[3] Voo YL, Foster SJ, Voo CC. Ultrahigh-performance concrete segmental bridge
(4) The proposed numerical model provided good predictions of the technology: toward sustainable bridge construction. J Bridge Eng 2015;20(08):
flexural behavior of the R-UHPC beams with desirable accuracy. B5014001–12.
With the increase in the reinforcement ratio, the contribution of [4] Shirai K, Yin H, Teo W. Flexural capacity prediction of composite RC members
strengthened with UHPC based on existing design models. Structures 2020;23:
the reinforcement to the flexural strength increases, whereas the 44–55.
contribution of the UHPC tensile capacity decreases significantly. [5] Wille K, EI-Tawil S, Naaman AE. Properties of strain hardening ultra high
In addition, the proportion of the UHPC tensile capacity contri­ performance fibre reinforced concrete (UHP-FRC) under direct tensile loading.
Cement and Concrete Composites, 2014;48: 53-66.
bution increases with the increase in steel fiber content for the R- [6] Wang Jun-Yan, Guo Jun-Yuan. Damage investigation of ultra high performance
UHPC beams. concrete under direct tensile test using acoustic emission techniques. Cem Concr
(5) The contribution of the UHPC tensile capacity can be considered Compos 2018;88:17–28.
[7] Kim DJ, Park SH, Ryu GS, et al. Comparative flexural behavior of Hybrid ultra high
as a safety reserve when the reinforcement ratio is greater than performance fibre reinforced concrete with different macro fibres. Constr Build
4.0%, whereas the contribution of the UHPC tensile capacity Mater 2011;25:4144–55.
needs to be considered for the UHPC beams with a reinforcement [8] Krahl Pablo Augusto, Carrazedo Ricardo, El Debs Mounir Khalil. Mechanical
damage evolution in UHPFRC: experimental and numerical investigation. Eng
ratio of less than 2.9%. In the serviceability limit states design,
Struct 2018;170:63–77.

14
Z. Feng et al. Engineering Structures 248 (2021) 113266

[9] Kang ST, Lee Y, Park YD, et al. Tensile fracture properties of an Ultra High [22] Yang In Hwan, Joh Changbin, Kim Byung-Suk. Structural behavior of ultra high
Performance fibre Reinforced Concrete (UHPFRC) with steel fibre. Compos Struct performance concrete beams subjected to bending. Eng Struct 2010;32(11):
2010;92:61–71. 3478–87.
[10] AFGC-Setra (French Association of Civil Engineering-French Authorities of Civil [23] Shafieifar Mohamadreza, Farzad Mahsa, Azizinamini Atorod. A comparison of
Engineering Structure Design, and Control), Ultra high performance fibre existing analytical methods to predict the flexural capacity of Ultra High
reinforced concretes, Paris: AFGC and SETRA Working Group, 2013. Performance Concrete (UHPC) beams. Constr Build Mater 2018;172:10–8.
[11] SIA (Swiss Society of Engineers and Architects), Ultra-high performance fibre [24] Yoo DY, Banthia N, Yoon YS. Experimental and numerical study on flexural
reinforced concrete (UHPFRC)-materials, design and execution, Final Draft, Zurich, behavior of UHPFRC beams with low reinforcement ratios. Can J Civ Eng 2016;44
Switzerland, 2016. (1):18–28.
[12] JSCE (Japan Society of Civil Engineers), Recommendations for design and [25] Xu H, Deng Z. Cracking moment and crack width of ultra-high performance
construction of ultrahigh strength fibre reinforced concrete structures (draft), JSCE concrete beams. Journal of Harbin Institute of Technology 2014;46(04):87–92 (in
Guideline for Concrete No. 9, Tokyo, 2006. Chinese).
[13] Association Française de Normalisation (AFNOR — French standard institute), [26] Qiu M, Shao X, Hu W, et al. Calculation method for crack width of reinforced UHPC
National addition to Eurocode 2 — Design of concrete structures: specific rules for flexural components. China Civil Engineering Journal 2020;53(10):89–119 (in
Ultra-High Performance Fibre-Reinforced Concrete (UHPFRC), NF P 18-710, Chinese).
French Standard, 2016. [27] Hasgul Umut, Turker Kaan, Birol Tamer, Yavas Altug. Flexural behavior of ultra-
[14] Aaleti S, Petersen B, Sritharan S. Design guide for precast UHPC waffle deck panel high-performance fiber reinforced concrete beams with low and high
system, including connections. Washington DC: Federal highway administration; reinforcement ratios. Structrual Concrete 2018;19(6):1577–90.
2013. [28] Shao Xudong, Pan Rensheng, Zhan Hao, Fan Wei, Yang Zhijie, Lei Wei.
[15] Graybeal Benjamin A. Flexural behavior of an ultrahigh-performance concrete I- Experimental verification of the feasibility of a novel prestressed reactive powder
girder. J Bridge Eng 2008;13(6):602–10. concrete box-girder bridge structure. J Bridge Eng 2017;22(6):04017015. https://
[16] Chen Linfeng, Graybeal Benjamin A. Modeling structural performance of ultrahigh doi.org/10.1061/(ASCE)BE.1943-5592.0001033.
performance concrete I-girders. J Bridge Eng 2012;17(5):754–64. [29] Yoo DY, Lee JH, Yoon YS. Effect of fibre content on mechanical and fracture
[17] Yoo DY, Yoon YS. Structural performance of ultra-high performance concrete properties of ultra high performance fibre reinforced cementitious composites.
beams with different steel fibres. Eng Struct 2015;102:409–23. Compos Struct 2013;106:742–53.
[18] Yoo Doo-Yeol, Kang Su-Tae, Yoon Young-Soo. Enhancing the flexural performance [30] Feng Z, Li C, He J, et al. Static and fatigue test on lightweight UHPC-OSD
of ultra-high-performance concrete using long steel fibers. Compos Struct 2016; composite bridge deck system subjected to hogging moment. Eng Struct 2021;228:
147:220–30. 111610.
[19] Zhen W, Li L, Lu S. Experimental research on the mechanical performance of [31] T/CBMF 37-2018. Fundamental characteristics and test methods of ultra-high
normal section of reinforced reactive powder concrete beam. Journal of Building performance concrete. China Architecture & Building Press, Beijing, 2018. (in
Structures 2011;32(06):125–34 (in Chinese). Chinese).
[20] Li L, Fan X, Shi X, et al. Experimental study on flexural of large-scale prestressed [32] Gere JM. Mechanics of materials. 6th ed. Belmont: Brooks/Cole; 2003.
UHPC T-shaped beam. China Civil Engineering Journal 2018;51(05):84–102 (in [33] China Architecture and Building Press. Code for design of concrete structures.
Chinese). GB50010-2010. Chinese Standard, 2010. (in Chinese).
[21] Qi J, Wang J, Ma ZGJ. Flexural response of high-strength steel-ultra-high- [34] Xu H, Deng Z. Experimental research on flexural behavior of prestressed ultra-high
performance fibre reinforced concrete beams based on a mesoscale constitutive performance steel fibre concrete beams. Journal of Building Structures 2014;35
model: experiment and theory. Structural Concrete 2018;19(3):719–34. (12):58–64 (in Chinese).

15

You might also like