Case Studies in Construction Materials 20 (2024) e03199

Contents lists available at ScienceDirect

Case Studies in Construction Materials

journal homepage: www.elsevier.com/locate/cscm

Experimental study and numerical simulation analysis of shear

behavior of coral aggregate reinforced concrete beam
Bo Da a, b, d, Kai Sun a, Yan Chen a, Bo Yu d, Zhangyu Wu c, Chengjun Yue c,
Da Chen a, b, *
a
College of Harbour, Coastal and Offshore Engineering, Hohai University, Nanjing 210098, China
b
Key Laboratory of Coastal Disaster and Defence of Ministry of Education, Hohai University, Nanjing 210098, China
c
School of Materials Science and Engineering, Southeast University, Nanjing 211189, China
d
Guangxi Key Laboratory of Disaster Prevention and Engineering Safety, 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: In order to investigate the shear behavior of coral aggregate reinforced concrete beam (CARCB),
Coral aggregate reinforced concrete beam the different concrete strength, steel types affect the shear behavior of CARCB was studied, the
Shear behavior calculation formula of shear capacity (Vcs) and numerical analysis model of CARCB was proposed.
Stirrups corrosion
The results show that: All kinds of CARCB have inclined section suitable reinforcement failure, the
Shear capacity
Calculation formula
failure law of CARCB and ordinary aggregate reinforced concrete beam (OARCB) was basically
Numerical simulation the same. Considering stirrups corrosion and nonlinear mechanical properties characteristic of
coral aggregate concrete (CAC), the Vcs calculation formula of CARCB was proposed, and its
applicability in C25 ~ C60 CARCB was verified, its accuracy was 26% and 34% higher than
GB50010–2020 and JGJ/T 12–2019, respectively. A numerical analysis model suitable for
describing the shear behavior of CARCB was proposed based on K&C (Karagozian & Case) theory,
and its applicability in C25 ~ C60 CARCB was verified, it was found that this model can well
describe the whole process of CARCB inclined section failure, and the errors between simulated
and measured values of Vcr, Vcs and midspan deflection are 0~18%. In addition, the Vcs obtained
by the numerical model is 3% more accurate than the calculation formula, indicating that the
numerical model can effectively distinguish the variation law of shear behavior of CARCB.

1. Introduction

In the context of the Belt and Road Initiative, a large amount of building materials are urgently needed for infrastructure con­
struction in tropical island areas, where traditional building materials are scarce but coral and seawater resources are abundant [1].
Therefore, it’s significant that use coral, coral sand, seawater as the main raw materials to prepare coral aggregate reinforced concrete
beam (CARCB), and applied in engineering construction of port, wharf, bridges and other island [2]. However, the “defects” of coral’s
natural porous structure and the large amount of Cl- contained in seawater and coral, easily lead to the corrosion of CARCB internal
steel bars [3]. As is known to all, corrosion of steel bars can reduce the bonding property between steel bars and concrete, and degrades
the structural mechanical property, which affecting engineering safety [4]. At the same time, shear failure of reinforced concrete
beams belongs to brittle failure, and its consequences are much more serious than bending failure. Moreover, with the rapid

* Corresponding author at: College of Harbour, Coastal and Offshore Engineering, Hohai University, Nanjing 210098, China.
E-mail address: [email protected] (D. Chen).

https://doi.org/10.1016/j.cscm.2024.e03199
Received 15 January 2024; Received in revised form 10 April 2024; Accepted 21 April 2024
Available online 25 April 2024
2214-5095/© 2024 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).
B. Da et al. Case Studies in Construction Materials 20 (2024) e03199

development of science technology, numerical simulation method has been widely used in civil engineering field in recent years. It has
the advantages of simple operation, low cost, fast solving speed, and guiding and complementing the experimental research [5–7].
Therefore, the shear behavior of CARCB was studied based on experiment and numerical simulation technology, and it has important
scientific significance and engineering value for the construction and restoration of tropical islands.
At present, the research of coral aggregate concrete (CAC) at home and abroad mainly focuses on the preparation technology,
mechanical properties and steel corrosion. Chu et al. [8] innovatively cleaned coral aggregate with citric acid, coated coral aggregate
with basic magnesium sulfate cement (BMSC), and prepared CAC, but the strength still needs to be improved. Liu et al. [9,10] studied
the shear strength of carbon fiber reinforced polymer (CFRP) CAC under compressive-shear composite stress, and found that the
addition of CFRP could reduce the spalling of concrete and improve the shear strength. Lu et al. [11] based on the corrosion mechanism
of steel bars, added Cr and Al into steel bars to study the corrosion deterioration performance of low-alloy steel bars in CAC, and
determined the improvement value of passivation performance of low-alloy steel bars in CAC, however, there is a lack of research on
the mechanism of steel corrosion and deterioration. Wu et al. [12] studied the influence of seawater desalination on the service life of
reinforced CAC structure based on the modified chloride ion diffusion theoretical model. Da et al. [13] tested the flexural behavior of
CARCB with different concrete strengths, and proposed the calculation model of ultimate bending moment and crack width of CARCB.
To sum up, the strength of CAC is generally not high, and the consideration of stirrups corrosion and high brittleness of CAC is not
comprehensive in the analysis of bearing capacity at present. Therefore, in this paper, by preparing high-strength CAC, considering the
influence of high brittleness of CAC and stirrups corrosion, the shear behavior of CARCB were studied experimentally.
In recent years, with the rapid development of computer technology, micromechanics calculation methods emerge in an endless
stream, among which finite element software is the most representative. Because the finite element method adopts the matrix
expression form, it can describe the problem very simply, so that the method of solving the problem is normalized. Therefore, finite
element method is favored by most scholars. Malvar et al. [14] established the Karagozian & Case (K&C) model which can analyze the
strengthening and damage effects of concrete, and proposed the conclusion that the K&C model is applicable under complex stress
loading conditions (tension, shear, pressure) and high strain rate dynamic loading. Chen et al. [15] used K&C model to analyze the
mechanical behavior of concrete, which is conducive to further understanding the characteristics of material model, and laying a
theoretical foundation for the finite element analysis of concrete structure. Hawileh et al. [16] explored the influences of concrete
strength, FRP material (type, size) and reinforcement ratio on the performance of ordinary aggregate concrete (OAC) beam by finite
element model (FE). French scholar Joseph et al. [17] proposed finite element nonlinear model (FEA) applied to the element analysis of
OAC beams with glass fiber reinforced polymer (GFRP) bars, and numerically analyzed the shear properties of GFRP-OAC beams with
this model. Linh et al. [18] used finite element method to simulate the mechanical properties of steel fiber OAC beams, and found that
the efficiency of finite element simulation depends on the constitutive model and shear crack transfer coefficient, and established a
numerical model suitable for simulating the mechanical properties of steel fiber OAC beams. Jin et al. [19] established a
three-dimensional numerical analysis model of reinforced concrete beam shear failure analysis, and investigated the effect of corrosion
on mechanical behavior of structure by multi - stage analysis method. To sum up, scholars at home and abroad have carried out a large
number of numerical simulation studies on OAC and its components, due to the advantages of K&C model in analyzing the
strengthening effect, tensile and compression damage effect, deformation damage effect and strain rate effect of concrete, it is one of
the models that can fully describe the influences of concrete plasticity, fracture zone width, failure pattern and strain rate on concrete
properties. Therefore, it is necessary to use K&C model to study the shear behavior of CARCB in this paper.
By combining experiments and numerical simulation, the influences of concrete types, concrete strength grades and reinforcement
types on failure mode, mid-span deformation and shear capacity (Vcs) of CARCB has been studied, the load-deflection, load-rein­
forcement strain relationships has been established, and the Vcs calculation model of CARCB has been proposed. Then, based on the test
results, the parameters of the constitutive model of CARCB have been determined, and the finite element model has been established to
analyze its failure mode, crack development, mid-span deformation and Vcs, the variation law of shear behavior of CARCB has been
revealed.

2. Experiment

2.1. Raw materials

The coarse (fine) aggregate used is coral (sand), which is produced from a certain island in the South China Sea [2] (Fig. 1). Fig. 2

Fig. 1. Coral and coral sand in South China Sea.

2
B. Da et al. Case Studies in Construction Materials 20 (2024) e03199

shows the grading curve and performance index of coral aggregate. Ordinary aggregates: gravel with an apparent density of
2820 kg/m3, a packing density of 1435 kg/m3, a crush indicator of 6%. River sand with an apparent density of 2500 kg/m3, a packing
density of 1615 kg/m3, a fineness modulus of 2.54. The cementation materials used are P⋅ II 52.5 type Portland cement, I grade fly ash
and S95 grade slag. The admixtures used are PCA-I polycarboxylic acid superplasticizer and calcium nitrite rust inhibitor. The steel
bars used are ordinary steel bars (A), coated steel bars (organic newly coated steel bars (B), zinc chrome coated steel bars (C)), 316
stainless steel bars (D), 2205 duplex stainless steel bars (E). The mixing/curing water used is artificial seawater, the proportion of each
material is NaCl: Na2SO4: MgCl2⋅6 H2O: KCl: CaCl2 = 24.5: 4.1: 11.1: 0.7: 1.2 [20].Table 1

2.2. Design method

The 12 CARCBs with different strength grades and reinforcement types and 2 ordinary aggregate reinforced concrete beams
(OARCB) were made. The mix ratio of CAC was shown in Table 2. The size of CARCB was 150 mm × 200 mm × 1500 mm, and the
shear span ratio (λ) was 1.23. In order to make the occur shear failure of CARCB, 2 flexural steel bars with Φ16 and two spiral steel bars
with Φ10 are set at the lower part of beam. The stirrup is made of Φ6 round steel bars (double leg hoop), and the space between the
stirrup is 100 mm.
According to GB50010–2020 Code for Design of Concrete Structures [21], the design of CARCB shear bearing capacity was verified by
the Eq. (1). Because the actual value of the shear capacity in this paper is less than the design value, it meets the requirements of the
code. Therefore, the designed CARCB can be used to conduct experimental research on the shear performance.

V ≤ 0.7× ft× b× h0+fyv× Asv× h0/s=0.7×4.3×150×140+440×452.16×140/100=341.74×103 kN (1)

Where: b is the beam width, b=150 mm; h0 is the beam height minus concrete protective layer thickness, h0=140 mm; ft, fyv is the
design value of concrete tensile strength and stirrup yield strength, respectively, ft=4.3 MPa, fyv=440 MPa; Asv is the region of stirrup,
Asv=452.16 mm2; s is the distance between stirrups, s=100 mm.
Using self-falling stirring equipment. Firstly, the coral aggregate, cement, SG, FA should be put into the mixer for dry mixing for
1 min, and then the mixture of seawater, water reducer, inhibitor should be added for wet mixing for 3 min. After discharging, the
slump was measured, and then pouring and vibration molding was carried out. The beam is manually vibrated with a vibrator and the
cube specimen is densified with a shaking table. After preparing the beam/specimen, the mold was cured for 24 h, then the mold was
disassembled, covered with straw and sprayed with artificial seawater for curing, and the shear behavior was tested 90 d later.

2.3. Testing method

Fig. 3 shows the schematic diagram of the loading device for the shear performance test of CARCB. According to GB50010–2020
Code for Design of Concrete Structures [21], the 4-point loading method is used for the test. The measuring instruments and their uses are
shown in Fig. 3. In addition, the strain variation of concrete in tensile/compression area during shear failure was monitored by laying
concrete strain gauges, the strain variation characteristics of corresponding parts were monitored by longitudinal reinforcement
(stirrup) strain gauges, and the principal stress distribution in shear bending section of CARCB was monitored by setting strain flowers.
In addition, before the formal loading, 10 kN pressure was applied in advance to ensure the contact of each component and test the
connection status of each instrument. During formal loading: all beams shall be loading at 1 kN per stage before cracking, loading at
5 kN per stage after cracking until all cracks appear, loading at 2 kN per stage at all cracks appeared until the end of the test.

Fig. 2. Grading curve and performance index of coral aggregate.

3
B. Da et al. Case Studies in Construction Materials 20 (2024) e03199

Table 1
Basic parameter of CARCB and OARCB.
No. Type of component Concrete strength fcu fc ft Type of fy fyv
/MPa /MPa /MPa steel bar /MPa /MPa

BJ1–1 OARCB C35 51.9 43.8 5.0 B 457 440

BJ1–2 OARCB C35 51.3 43.4 5.0 B 457 440
BJ2–1 CARCB C25 42.7 37.1 4.3 B 457 440
BJ2–2 CARCB C25 42.4 36.9 4.3 B 457 440
BJ3–1 CARCB C50 66.7 54.6 6.2 B 457 440
BJ3–2 CARCB C50 66.6 54.5 6.2 B 457 440
BJ4–1 CARCB C60 75.9 61.3 6.9 B 457 440
BJ4–2 CARCB C60 76.4 61.6 7.0 B 457 440
BJ5–1 CARCB C60 76.2 61.4 6.9 A 457 440
BJ5–2 CARCB C60 76.2 61.4 6.9 A 457 440
BJ6–1 CARCB C60 75.4 60.9 6.9 C 457 440
BJ7–1 CARCB C60 75.8 61.2 6.9 D 811 805
BJ7–2 CARCB C60 76.2 61.5 6.9 D 811 805
BJ8–1 CARCB C60 76.3 61.5 6.9 E 591 588

Note: The mix ratio of C35 OARCB was consistent with that of C25 CARCB. fy and fyv are yield strength of longitudinal steel and stirrup, respectively.
fcu, fc and ft are cube compressive, axial compressive and axial tensile strength of concrete, respectively.

Table 2
The mix ratio and the properties of CARCB and OARCB.
No. Material usage per unit volume/(kg⋅m− 3) Net W/ Total W/ Slump Apparent
B B /mm density
Cement SG FA Coral Coral Total Water Rust
/(kg⋅m− 3)
sand water reducer inhibitor

C25 275 150 75 873 582 250 0.0 15.0 0.30 0.50 130 2170
C50 429 83 38 1020 680 200 3.3 16.5 0.25 0.36 70 2188
C60 780 150 70 700 300 264 6.0 30.0 0.20 0.26 255 2267

Note: the mix ratio of OARCB is consistent with C25 CARCB, except that coral, coral sand are replaced by gravel, river sand.

Fig. 3. Loading device for shear behavior test of CARCB (unit: mm).

3. Numerical simulation

3.1. Geometric model

Based on ANSYS nonlinear finite element software, the numerical analysis model was established (Fig. 4) to explore the variation
law of shear behavior of CARCB. The model size shall be established according to the actual size of CARCB (150 mm × 200 mm ×

4
B. Da et al. Case Studies in Construction Materials 20 (2024) e03199

1500 mm). To avoid stress concentration, steel rods are placed at the loading point. The elastic modulus of steel rods is 2 × 10 MPa, so
as to ensure that the steel rods will not have large deformation during the model loading process. Meanwhile, the key word * CONTACT
AUTOMATIC SURFACE_TO SURFACE (ID/TITLE/MPP)_(THERMAL) (6) is used to define the contact between the steel rod and
CARCB. In the loading test of CARCB, the restraint at the support is depend on support type. Therefore, in the analysis of the numerical
model, the restriction of the CARCB is depend on at both ends, that is, one end of the beam is allowed to rotate and displace around the
Z axis, while the other end is only permit to rotate around the Z axis. In order to make the calculation consequence converge,
displacement control is used to load.

3.2. Element types and mesh division

The element type greatly affects the results of ANSYS finite element numerical analysis [22]. In this paper, BEAM161 beam element
is used for simulation analysis of CARCB and steel rod support, which can accurately describe the displacement changes of the model,
and ensure the analysis accuracy even under the condition of mesh distortion. The steel bars in the CARCB were simulated using
LINK160 bar units. Adopt *MA T_CONCRETE_DAMAGE_REL3_(072R3) (K&C model) to simulate concrete structure, use *MA
T_PLASTIC_KINEMA TIC_(003) to simulated steel bar, and define the loading steel rod as a rigid body, the loading rate to 2×10− 6/s.
Create parts for concrete, steel bar and loading steel rod, respectively, a total of 8 parts, contact between parts is defined by *CON­
TACT_AUTOMA TIC_SURFACE_TO_SURFACE.
In order to better describe the damage changes of CARCB, after the previous study, the research group determined that the mesh
size of CARCB in the model was 10 mm [23], and the mesh size of steel rod was 30 mm due to the loaded steel rod was not the main
analysis object. Therefore, a total of 1472 solid elements are obtained by using 1 mm and 30 mm spatial hexahedral elements for finite
element meshes of concrete structure and steel bar structure, respectively. In addition, in order to describe the failure patterns of
CARCB, the key word * MAT_ADD_EROSION is used to calculate failure units of concrete structure under load. The failure criterion is
set to 0.05 times of the maximum principal strain, that is, when the maximum principal strain of unit exceeds the preset failure cri­
terion, failure of the unit is determined. Fig. 4 shows the finite element model and mesh division of CARCB.

3.3. Contact relationship

There are three simulation methods for steel bar in ANSYS software: (1) Define the steel bar layer, and embed it directly into the
element. (2) The steel bar is embedded into the solid unit as surface unit or membrane unit. (3) Establish the independent geometric
properties of steel bar, and then assemble them into the solid unit. Compared with the method (1) ~ (2), the independent modeling
process of steel bar (method (3)) is more complicated, but it can simulate the steel cage more accurately, stress of steel bar analysis
results are also easy to read. Therefore, in order to improve the simulation analysis method accuracy, the independent modeling of
concrete and steel bar is selected.
In addition, in the separation modeling, the key to realize the combined action is to establish the interaction relationship between
steel bar and concrete, because they are independent from each other. There are usually two methods to consider the contact between
steel bar and concrete: (1) The connection between steel bar and concrete is established through spring elements. It is necessary to
manually modify K file to realize the definition of spring elements, which is not only difficult to operate, but also difficult to accurately
define spring constitutive relation due to the complex bond and slip mechanism between steel bar and concrete. (2) The interaction
between steel bar and concrete was considered using embedded element, which showed that there was no mutual slip between the two.
The displacements of the steel bar were obtained using linear interpolation, and the mechanical relationship between the two was
established through displacement coordination. Therefore, due to the strong bond between CAC and steel bar, considering the
complexity of modeling and solving efficiency, embedded element was used to simulate the interaction between CAC and steel bar.
Fig. 5 shows the concrete unit.

Fig. 4. Finite element model and mesh division of CARCB.

5
B. Da et al. Case Studies in Construction Materials 20 (2024) e03199

3.4. Model building

Fig. 6 shows the failure surface and constitutive relationship of concrete [24]. Fig. 6(a) represents the failure surface of the K&C
model, which the curves 1, 2 and 3 represent the maximum, yield and residual failure surface, respectively. The failure process of
concrete goes through three stages: (1) Elastic stage, from the origin to the initial yield strength Pt.1. (2) Plastic strengthening stage,
Pt.1 to the ultimate strength Pt.2. (3) Softening stage, Pt.2 to the residual strength Pt.3, the size of Pt.3 is related to the constraints.
Fig. 6(b) represents the stress-strain relationship of concrete, and points 4, 5 and 6 represent the ultimate strength, initial yield strength
and residual strength, respectively. Points 4, 5 and 6 correspond to points Pt.1, Pt.2 and Pt.3 in Fig. 6(a), respectively.
The following are three mathematical functions of failure surface:
p
Yieldfailuresurface(Curve1) : Δσy = a0y + (2)
a1y + a2yp

p
Maximumfailuresurface(Curve2) : Δσ m = a0 + (3)
a1 + a2p

p
Residualfailuresurface(Curve3) : Δσr = a0 + (4)
a1f + a2fp

Where: Δσ y, Δσ m and Δσ r are the effective stress deviation of three failure surfaces (Δσ =√3J2), respectively, among them, J2 is the
second variable of deviatoric stress, J2=(S21+S32+S23)/2, S1, S2, S3 are the first, second and third deviatoric stresses, respectively. p is the
hydrostatic pressure, p = -(σ1+σ2+σ 3)/3, among them, σ 1、σ2、σ 3 are the first, second and third principal stresses, respectively. a0y,
a1y, a2y, a0, a1 f, a2 f are constants of concrete materials.
In the K&C model, the failure surface of concrete under different stress states is represented by the strengthening surface and
softening surface, respectively. The formula is as follows:
Strengtheningsurface : Δσf = ηΔσ m + (1 − η)Δσ y λ > λm (5)

Softeningsurface : Δσh = ηΔσ m + (1 − η)Δσr λ > λm (6)

Where: Δσ f、Δσh are the effective stress deviation of the strengthened surface and softened surface, respectively. η is the reduction
coefficient. λ is the shear span ratio.

3.5. Parameter determination

The K&C numerical model mainly involves physical, strength and unit conversion parameters. The physical parameters of concrete
include density (ρ), Poisson’s ratio (μ). The strength parameters of concrete include cube compressive/splitting tensile strength (fcu/ft),
strain rate curve coefficient (l). Unit conversion parameters include length unit conversion coefficient (r), when the length unit is m,
39.37 m is taken, stress unit conversion coefficient (u), when the stress unit is MPa, 145 MPa is taken [25]. When using the numerical
simulation method to study the mechanical properties of CARCB, fcu of different concrete strength should be input, and the finite
element software can automatically generate other parameters to complete the calculation.
The ideal elastic-plastic model is used to simulate the steel bar, it is defined by keyword *MA T_PLASTIC KINEMA TIC_(TITLE)
(003), including density (ρ), modulus of elasticity (E) [26], poisson’s ratio (μ) [27], yield stress of steel bar (fy) [28], yield strain of steel
bar (ε). Table 3 ~ Table 4 shows the concrete and steel bar parameters of CARCB.

Fig. 5. Concrete unit.

6
B. Da et al. Case Studies in Construction Materials 20 (2024) e03199

Fig. 6. The failure criterion of concrete [24].

4. Results and analysis

4.1. Failure pattern

Fig. 7 shows the comparison of measured and numerical simulated failure patterns of CARCB. As can be seen from the figure, both
ordinary aggregate reinforced concrete beam (OARCB) and CARCB have inclined section failure, and the failure pattern is basically the
same. The numerical simulation results of failure pattern for different strength grade CARCB agree well with the test result, and the
crack development process, width and number of the side and bottom of beams are in good agreement. Moreover, the CARCB shows
obvious ductility of suitable reinforcement failure. With the higher of concrete strength, the destruction degree of CARCB increases
gradually. In addition, the crack development whole process of CARCB is relatively stable, the concrete cracking in the tensile area, and
the steel yielding and concrete crushing in the compression area, the deformation of CARCB before failure is large, there are obvious
signs of failure. In addition, due to the low shear strength of coral aggregate, when CARCB cracks under shear, coral aggregate has been
clipped from all cracks, which makes it difficult to maintain the integrity of the particles.

4.2. Failure process

Fig. 8 shows the whole process failure pattern of CARCB (BJ6–1). As can be seen from the figure, the failure process of CARCB can
be divided into 4 stages: (1) when loading to 30~40% shear capacity (Vcs) (point A), the first obvious vertical crack appears in the
bending area, which is located at about 60% beam height, the bearing capacity of CARCB continues to increase. (2) Continued loading
to 50~60% Vcs (point B), the number of cracks increased in the flexural area, but the existing fine cracks extended slowly. At this time,
the bearing capacity continued to increase, but the rate gradually slowed down. (3) Near 80~90% Vcs (point C), oblique cracks appear
at the connection between the left/right supports and loading point of CARCB, the stirrup begins to yield. At the same time, the
propagation speed of cracks increases, but the number of cracks no longer increases, and the bearing capacity is basically stable. (4)
Continue loading to Vcs (point D), CARCB is obviously bending deformation, cracks extend rapidly, the CARCB was damaged along the
oblique crack through, and the bearing capacity decreases rapidly. This show that the whole process failure characteristics of CARCB
and OARCB are basically the same [29].
The failure pattern of CARCB (Fig. 9) is different from that of OARCB. During the failure of OARCB, cracks penetrate along the
interface layer between mortar and aggregate, while cracks of CARCB directly penetrate the coral aggregate, forming a flat fracture
surface without developing around the aggregate. This is because coral aggregate does not have shear failure along the maximum shear
stress plane like other aggregates, but tensile failure along the growth line of coral aggregate, forming a number of broken strips [30].
Therefore, when cracks occur, cracks can run through coral aggregate, and the appearance of cracks is sudden and shows obvious
brittleness. At the same time, the shear strength of coral aggregate is low, and when CARCB cracks under shear, the coral aggregate
itself on all cracks is cut, and it is difficult to maintain the integrity of the particles [1]. After the crack section of the shear beam and the
compression test block is opened, it is also found that the situation is the same, that is, the tensile crack section is relatively flat, and the
coral aggregate itself is all pulled apart.

Table 3
The concrete parameters of CARCB.
Concrete strength ρ (kg/m3) μ fcu (MPa) l [25] r [25] u [25]

C25 2167 0.25 38.3 39.37 1.45E-4 1

C35 2070 0.26 42.4 39.37 1.45E-4 1
C50 2183 0.28 49.0 39.37 1.45E-4 1
C60 2267 0.27 60.4 39.37 1.45E-4 1

7
B. Da et al. Case Studies in Construction Materials 20 (2024) e03199

Table 4
The steel bar parameters of CARCB.
ρ (kg/m3) E (GPa) μ fy (MPa) ε
7800 210 0.3 400 10,000

Note: Table 4 is the value of steel bar without corrosion, and the value after corrosion is corrected according to the mass loss rate.

Fig. 7. Comparison of measured and numerical simulated failure patterns of CARCB.

Fig. 8. The whole process failure pattern of CARCB.

8
B. Da et al. Case Studies in Construction Materials 20 (2024) e03199

Fig. 9. Fracture diagram of CARCB.

4.3. Load-midspan deflection curve

Fig. 10 shows the load-midspan deflection measured curve of CARCB. As can be seen from the figure, the bending moment-midspan
deflection curve of concrete beam is divided into two stages (concrete normal section before cracking, after cracking to failure stage).
With the increase of concrete strength, Vcs of CARCB gradually increases, because Vcs of CARCB consists of concrete bearing capacity
(Vc) and stirrup bearing capacity (Vs). When Vs is constant, Vc increases, and Vcs increases. In addition, the stiffness of BJ4–2 and BJ6–1
is significantly higher than that of BJ5–1 and BJ5–2, indicating that containing a large amount of Cl- in CARCB, the ordinary steel bar
(BJ5–1 and BJ5–2) has been seriously corroded, resulting in a certain degree degradation of the stiffness of CARCB. The organic new-
coating (BJ4–2) and zinc-chromium coating (BJ6–1) inhibit the corrosion of steel bars due to their excellent anti-rust properties.
Fig. 11 shows the comparison of measured and numerical simulated load-midspan deflection curves of CARCB. As can be seen from
the figure, the load-midspan deflection simulation curve of CARCB is in excellent agreement with the measured curve. When the load is
small, the two curves almost coincide completely, and the curves are close to a straight line. At this time, concrete plays the main role of
bearing the load, and the CARCB is in a linear elastic state. When the load increases to concrete cracking, there is a relatively large
deviation between the two curves. The peak point and the descending section of the numerical simulated curve are both slightly larger
than the measured results, this is because of the relative slip gradually occurs between the steel bar and concrete with the increase of
the load. When CARCB enters the elastic-plastic stage, due to the large load, the tensile force on the steel bar gradually exceeds the
friction force and mechanical bite force between the steel bar and concrete, resulting in relative slipping, accelerating load capacity
reduction of CARCB [31].

4.4. Bearing capacity

Table 5 shows the comparison of measured and numerically simulated shear capacity and midspan deflection of CARCB. The table
shows that, when the concrete strength is same, the normal section cracking load (Vcr) and Vcs of CARCB with different steel bars are as
follows: ordinary steel bars < organic new-coated steel bars < zinc-chromium coated steel bars < 316 stainless steel bars. This is
mainly due to the following reasons: (1) before cracking, stirrups will share a part of shear stress, and the yield strength of 316 stainless
steel bars (fyv = 805 MPa) is significantly greater than that of ordinary steel bars (fyv = 440 MPa), that is, the shear behavior of 316
stainless steel stirrups is strong, so the normal section Vcr of 316 stainless steel bar CARCB is the largest. (2) Vcs of CARCB consists of

Fig. 10. The load-midspan deflection measured curve of CARCB.

9
B. Da et al. Case Studies in Construction Materials 20 (2024) e03199

Fig. 11. Comparison of measured and numerical simulated load-midspan deflection curves of CARCB.

Table 5
Comparison of measured and numerically simulated shear capacity and mid-span deflection of CARCB.
NO. Concrete strength Normal section Vcr/kN RE Oblique section Vcr/kN RE Vcs/kN RE Midspan deflection/mm RE

Vm
crn Vscrn Vm
cro Vscro Vm
cs Vscs fm s
f

BJ2–1 C25 17 19 11% 57 66 15% 117 106 10% 3.34 3.40 1%

BJ2–2 C25 21 19 9% 65 66 1% 113 109 4% 3.59 3.40 5%
BJ3–1 C50 37 43 16% 63 72 14% 132 130 1% 4.84 4.43 8%
BJ3–2 C50 36 41 13% 80 72 15% 130 125 4% 3.78 4.43 17%
BJ4–1 C60 34 36 5% 90 81 10% 137 132 4% 3.98 4.72 18%
BJ4–2 C60 38 44 15% 69 81 17% 136 126 7% 4.24 4.72 11%
BJ5–1 C60 44 44 0% 76 62 18% 116 114 1% 2.97 3.47 16%
BJ5–2 C60 46 44 4% 54 62 14% 98 99 1% 3.35 3.47 3%
BJ6–1 C60 37 43 16% 65 67 3% 184 200 10% 3.22 3.14 2%
BJ7–1 C60 54 44 18% 70 67 4% 214 210 2% 3.12 3.14 2%
BJ7–2 C60 52 44 15% 62 67 8% 208 216 5% 3.05 3.14 3%
BJ8–1 C60 30 35 17% 90 102 13% 200 211 5% 2.60 2.82 8%

Note: Vm s m s
crn and Vcrn are the measured value and simulated value of normal section, Vcro and Vcro are the measured value and simulated value of oblique
section, Vmcs and V s
cs are the measured value and simulated value of shear capacity, f m
and f s
are the measured value and simulated value of mid-span
deflection, RE is the relative errors.

stirrup shear capacity (Vs) and concrete shear capacity (Vc), when the concrete strength is same, the Vcs of CARCB gradually increases
with the stirrup strength increase, so the Vcs of 316 stainless steel bar CARCB is the highest. At the same time, Vcs of ordinary steel bar
(BJ5–1) is slightly smaller than that of organic new-coated steel bar (BJ4–2) and zinc-chromium coated steel bar (BJ6–1). The main
reason is that there is a large amount of Cl- in CARCB, which causes serious corrosion of ordinary steel bar, and reduces the effective
section of stirrup [32], thus reducing Vcs of BJ5–1.
In addition, the type of steel bar (BJ4~BJ8) has little influence on the oblique section Vcr of CARCB, this is due to the shear stress of
oblique section is mainly borne by CAC before the oblique section cracks appear. At this point, the stirrup stress is very low, and the
restraint of the stirrup on the concrete has not yet taken effect. At the same time, it can be seen from the table that the absolute errors
(AE) of the simulated and measured normal section Vcr of CARCB are 2 kN (BJ2–1), 6 kN (BJ3–1), 2 kN (BJ4–1), 0 kN (BJ5–1) and
6 kN (BJ6–1), respectively, and the relative errors (RE) is between 0~18%. The absolute error of the simulated and measured oblique
section Vcr of CARCB are 9 kN (BJ2–1), 9 kN (BJ3–1), 9 kN (BJ4–1), 6 kN (BJ5–1) and 2 kN (BJ6–1), respectively, and the relative
error is between 1~18%. The absolute errors of the simulated and measured Vcs of CARCB are 11 kN (BJ2–1), 1 kN (BJ3–1), 6 kN
(BJ4–1), 1 kN (BJ5–1) and 20 kN (BJ6–1), respectively, and the relative errors is between 1~10%. The absolute error of the simulated
and measured midspan deflection of CARCB are 0.06 mm (BJ2–1), 0.41 mm (BJ3–1), 0.74 mm (BJ4–1), 0.50 mm (BJ5–1) and
0.08 mm (BJ6–1), respectively, and the relative error is between 1~18%. In summary, the analysis shows that the relative errors (RE)
between the simulated and measured values of different indexes of CARCB are between 0~18 %, the numerical simulation model
proposed has high accuracy.

10
B. Da et al. Case Studies in Construction Materials 20 (2024) e03199

4.5. Steel bar strain

Fig. 12 shows the load-longitudinal steel bar strain curve of CARCB. It can be seen from the figure that: (1) When the load is small,
the steel bar strain does not change significantly, when the load is greater than 50 kN, the steel bar strain suddenly increases, it is
basically consistent with the cracking load of concrete (Table 5), indicating that the sudden change value of steel bar strain can also
judge the cracking load of concrete. (2) When the shear failure occurred in different types of CARCB, the longitudinal steel bar did not
reach its yield strain, and all showed shear-compressive failure. (3) When CARCB shear failure occurs, the strain law of longitudinal
steel bar is 2205/316 stainless steel bars, organic new-coated steel bars is small, followed by zinc-chromium coated steel bars, and
ordinary steel bars is the highest, indicating that coated/stainless steel bars can effectively anti-corrosion, while ordinary steel bars
have serious corrosion. This is mainly due to the following reasons: (1) The organic new-coating does not react with acid and base, and
has high chemical stability. It can effectively insulate the surface of the steel bar, and block the transmission path of chemical sub­
stances such as Cl- in the concrete, thus prolonging the time of steel bar rusting and effectively protecting the steel bar [29]. (2) 316
stainless steel bars (00Cr17Ni14Mo2) contains Cr and Mo, which will form a dense (Cr2O3) passivation film during corrosion, thus
preventing its erosion [33]. (3) Due to the superposition of zinc powder and aluminum powder, the porosity of the coating decreases,
thus effectively preventing the intrusion of Cl- and other harmful ions, and thus preventing the corrosion of steel bar [34].
Fig. 13 shows the comparison of measured and numerically simulated load-longitudinal steel bar strain of CARCB. It can be seen
from the figure that: (1) at the initial stage of loading, the two curves are consistent. (2) With the external load exceeding 30 ~ 50 kN,
concrete enters the plastic deformation stage, cracks occur in the tension zone of CARCB, and steel bars continue to bear the load
instead of the damaged concrete. The variation trend of the two curves begins to difference, and the simulation curve of longitudinal
steel bar strain increases faster, and the curve changes more smooth and steady. This is because the strain gauge is difficult to collect
the strain of the longitudinal steel bar which increases sharply, while the numerical simulation method can describe the strain value at
each time point conveniently. In addition, steel bar corrosion in the actual process can reduce the cross-section and yield strength of
stirrup, but the influence of steel bar corrosion is not considered in the numerical simulation. Therefore, there is a certain deviation
between the measured and simulated load-longitudinal steel bar strain curve after cracking of CARCB. (3) The load continues to in­
crease until the steel bar yields, the load-longitudinal steel bar strain curve of CARCB gradually decreases, and the beam loses its
bearing capacity.

4.6. Concrete strain

Fig. 14 shows the concrete strain distribution of the midspan section of the CARCB. The concrete strain of CARCB midspan section
can better conform to the assumption of plain section when the external load is not very large. When the load increases, the strain of the
whole section no longer satisfies the assumption of plain section.

4.7. Shear capacity

According to GB 50010–2020 Code for Design of Concrete Structures [21] and JGJ/T 12–2019 Technical Standard for Application of
Lightweight Aggregate Concrete [35], Vcs calculation formula of reinforced concrete beams is
1.75 Asv
GB50010 − 2020 : · Vcs = ft bh0 + fyv h0 (7)
λ+1 s

Fig. 12. Load-longitudinal steel bar strain curve of CARCB.

11
B. Da et al. Case Studies in Construction Materials 20 (2024) e03199

Fig. 13. Comparison between measured and numerically simulated load-longitudinal steel bar strain of CARCB.

Fig. 14. Strain distribution diagram of midspan section of CARCB.

/
0.175 Asv
JGJ T12 − 2019 : · Vcs = fc bh0 + 1.25fyv h0 (8)
λ + 1.5 s
Where: Vcs is shear capacity. λ is shear span ratio, λ=a/h0, a is the distance between the calculated section and support section/node
edge. For GB 50010–2020, when λ < 1.5, λ = 1.5, when λ > 3, λ = 3. For JGJ/T 12–2019, when λ < 1.4, λ = 1.4, when λ > 3, λ = 3. The
meanings of other parameters are the same as in Eq. (8) and Table 1.
According to Eq. (7) and Eq. (8), the Vcs of 10 CARCB in this experiment were calculated. The average ratio of measured (Vm cs) to
calculated value (Vccs) of Vcs was 1.672 and 1.861, the standard deviation was 0.174 and 0.381, and the coefficient of variation was
0.104 and 0.204, respectively (Table 6 and Fig. 15), the calculated values of Eqs. (7); (8) are quite different from the measured values.
Therefore, Vcs calculation formulas of GB 50010–2020 and JGJ/T 12–2019 are not applicable to CARCB. The main reasons are: (1) The
strength of traditional lightweight aggregate concrete (LAC) is lower, while the strength of CARCB in this test is higher, in addition,
CAC has nonlinear mechanical properties different from other concrete (ordinary concrete, lightweight aggregate concrete, etc.). (2)
The natural porous structure “defect” of coral, and the large amount of Cl- contained in seawater and coral are prone to corrosion of
steel bars, thus reducing the section of stirrups and the yield strength of stirrups, seriously affecting Vcs of CARCB [36]. However, the
influence of steel bar corrosion is not considered in Eq. (7) and Eq. (8). Therefore, it is necessary to comprehensively consider the
effects of concrete properties on Vc and stirrups corrosion on Vs, revise the coefficients of Vc and Vs, and put forward a Vcs calculation
model suitable for CARCB.

12
B. Da et al. Case Studies in Construction Materials 20 (2024) e03199

Table 6
Measured and calculated shear capacity of CARCB under different calculation methods.
No. ω Vm
cs Eq. (7) Eq. (8) Eq. (9) Eq. (13) Numerical simulation
/%
Vccs Vm c
cs/Vcs Vccs Vm c
cs/Vcs Vccs Vm c
cs/Vcs Vcc Vcs Vccs Vm c
cs/Vcs Vcc Vcs Vccs Vm c
cs/Vcs

BJ2–1 0.66 117 68 1.508 60 1.701 87 1.344 28 64 92 1.271 40 72 112 1.044

BJ2–2 0.66 113 67 1.760 59 1.985 79 1.424 36 73 109 1.036 34 86 120 0.941
BJ3–1 0.35 132 87 1.391 74 1.626 83 1.578 41 74 115 1.051 44 82 126 1.047
BJ3–2 0.35 130 80 1.475 68 1.723 98 1.319 39 75 114 1.140 40 83 123 1.056
BJ4–1 0.33 137 77 1.617 66 1.908 100 1.369 44 90 134 1.022 36 89 127 1.078
BJ4–2 0.66 136 81 1.508 72 1.701 177 0.769 46 85 131 0.934 43 83 126 1.079
BJ5–1 0.94 116 74 1.805 55 2.425 76 1.534 46 79 125 0.928 37 84 121 0.958
BJ5–2 0.94 98 70 1.932 54 2.521 68 1.446 56 93 119 0.823 31 86 117 0.837
BJ6–1 0.11 184 94 1.963 76 2.412 109 1.689 74 89 163 1.128 59 145 204 0.902
BJ7–1 0.04 214 122 1.725 155 1.356 157 1.364 89 105 194 1.103 68 132 200 1.070
BJ7–2 0.04 208 128 1.625 149 1.396 161 1.292 86 106 192 1.083 71 125 196 1.061
BJ8–1 0.05 200 114 1.754 126 1.587 168 1.200 79 105 184 1.086 81 99 180 1.111

Eq. (7): the average value of Vm c

cs/Vcs was 1.672, the standard deviation was 0.174, and the coefficient of variation was 0.104, respectively.
Eq. (8): the average value of Vm c
cs/Vcs was 1.861, the standard deviation was 0.381, and the coefficient of variation was 0.204, respectively.
Eq. (9): the average value of Vm cs /Vc
cs was 1.360, the standard deviation was 0.220, and the coefficient of variation was 0.161, respectively.
Eq. (13): the average value of Vm c
cs/Vcs was 1.050, the standard deviation was 0.111, and the coefficient of variation was 0.106, respectively.
Numerical simulation: the average value of Vm c
cs/Vcs was 1.015, the standard deviation was 0.081, and the coefficient of variation was 0.080,
respectively.
Note: Vm c
cs, Vcs is the measured, calculated shear capacity, respectively. The meanings of other parameters are the same as in Eq. (7), Eq. (8) and Eq.
(12). Among them λ=1.23, h0=162 mm, b=150 mm, s=100 mm, fyv=440 MPa (BJ1 to BJ6), fyv=805 MPa (BJ7), fyv=588 MPa (BJ8), the ft and fc are
shown in Table 1, the ω was obtained through the experiments. Vcc, Vcs is the calculated concrete shear capacity, stirrup shear capacity, respectively.

Fig. 15. Comparison of measured and calculated shear capacity of CARCB.

4.8. Coefficient correction

4.8.1. Vs coefficient
After the test of CARCB with different types of steel bars, the steel surface state is shown in Fig. 16. As can be seen from the figure, in
addition to 316 stainless steel bars, serious pitting occurred in ordinary steel bars, while organic new-coated steel bars and zinc-
chromium coated steel bars mainly suffered local corrosion. This is due to the inclusions on the coated steel bar surface, which will
destroy the coated uniformity and the passivation film formed on the steel bar surface, so as to easily cause local corrosion. Therefore,
it is recommended that the organic new-coating and zinc-chromium coating should ensure that the steel bar surface is clean when
brushing, and the missing coating and uneven coating parts should be recoated, so as to effectively avoid local corrosion.
Yuan et al. [37] showed that the effect of steel bar corrosion on Vcs of beams is mainly the reduction of stirrup section and stirrup
yield strength. For the reduction of stirrup section, the relationship between it and stirrup corrosion rate can be directly considered. As
for the relationship between the yield strength of stirrup [38] and stirrup corrosion, Yuan et al. [39] studied it in detail, which can be
directly used. Therefore, considering the influence of stirrup corrosion of CARCB, the calculation formula of Vcs and comprehensive
reduction coefficient (α) is expressed as:

13
B. Da et al. Case Studies in Construction Materials 20 (2024) e03199

Fig. 16. Surface state of steel bars in CARCB with different types of steel bars (90 days).

0.175 Asv
Vcs = fc bh0 + 1.25αfyv h0 (9)
λ + 1.5 s
{
(1 − ωs )(1 − 1.608ωs ) 0 < ωs ≤ 5%
α= (10)
(1 − ωs )(0.962 − 0.848ωs ) ωs > 5%

Where: α is the comprehensive reduction coefficient considering stirrup cross-section loss and stirrup yield strength reduction caused
by steel bar corrosion. ωs is the average section loss rate of stirrup. The meanings of other parameters are the same as in Eq. (7).
Fig. 17 shows the relationship between the maximum section loss rate (ωsm) and mass loss rate (ω) of steel bar under chloride salt
erosion environment. It can be seen from the figure that there is a good linear relationship between ωsm and ω.
ωsm = 1.344ω (11)
In the marine environment, the steel bar occurs different degrees of pitting corrosion. Therefore, the smallest section of the steel bar
(in ωsm area) is the most easily broken. Therefore, it is more accurate to use ωsm to calculate the comprehensive reduction coefficient,

Fig. 17. The relationship between the ωsm and ω of steel bar under chloride salt erosion environment.

14
B. Da et al. Case Studies in Construction Materials 20 (2024) e03199

which considering the loss of steel bar cross section and the reduction of steel bar yield strength caused by steel bar corrosion. In
addition, a large number of studies show that ω and ωs are roughly equal [40,41]. Therefore, based on rationality and practicality, α is
expressed as:
{
(1 − 1.344ω)(1 − 1.608ω) 0 < ω ≤ 5%
α= (12)
(1 − 1.344ω)(0.962 − 0848ω) ω > 5%

4.8.2. Vc coefficient
The shear behavior of beams without web steel bar was not tested in this experiment. So, by reference to Eq. (9), Vc is approximately
equal to Vcs-1.25αfyvh0Asv/s. Fig. 18 shows the relationship between Vc/(fcbh0) and λ of CARCB. As can be seen from the figure, among
the 10 CARCBs under concentrated load, the measured Vc of 4 CARCBs are less than the calculated result of Eq. (9). Therefore, it can be
considered that the value of 0.175/(λ+1.5) is too high, that is, it is not safe enough to design Vc of CARCB according to JGJ/T12–2019.
This is consistent with what was observed in the experiment.
In consideration of Vcs calculation of CARCB, the principle of Vc coefficient determination is: within the range of shear-compressive
failure section (λ = 1 ~ 3), this coefficient represents the “lower envelope” of all test points, λ = 1.23 in this experiment, λ = 0.84 ~
2.60 in Zhang [42] experiment. According to the analysis in Fig. 18, when λ ≤ 1.3, the Vc coefficient is 0.104. When λ ≥ 2.8, Vc
coefficient can only be 0.022. According to the above conditions can be calculated Vc coefficient of 0.1/2.4(λ－0.9), that is, 0.042/(λ－
0.9).
In addition, the calculation formula of Vcs in JGJ/T 12–2019 is the lower limit taken from the test data after statistical analysis. In
order to calculate Vcs of CARCB more accurately, measured values of test materials should be used in calculation. Therefore,
considering stirrup corrosion and CAC high strength characteristics comprehensively, Vcs calculation formula of CARCB is
0.042 Asv
Vcs = Vc + Vs = fc bh0 + 1.25αfyv h0 (13)
λ − 0.9 s

Where: Vcs, Vc and Vs are the shear capacity of CARCB, concrete and stirrup, respectively. fc is the axis compressive strength of
concrete. fyv is the tensile strength of stirrup. The meanings of other parameters are the same as Eq. (7) and Eq. (10).
The Vcs of 10 CARCBs were calculated according to Eq. (13) were shown in Table 6. The average value of Vm c
cs/Vcs was 1.050, the
standard deviation was 0.111, and the coefficient of variation was 0.106, respectively. Compared with the calculated results of GB
50010–2020 and JGJ/T 12–2019, the average value of Vm c
cs/Vcs is closer to 1, the standard deviation and coefficient of variation were
90.0 % and 16.9 % lower than JGJ 12–2006 respectively, and the coefficient of variation is reduced by 0.8 % and 7.4 %, respectively.

4.8.3. Equation verification

Eq. (13) and numerical model are used to analyze the other scholars’ [43] CARCB, and the results are shown in Table 7. It can be
seen from the table that the Vm c
cs/Vcs in original article are 1.554, 1.612, 1.244, 0.961, 0.758, 0.693, 0.942, 1.102, 1.070, 1.081,
m c
respectively. The Vcs/Vcs (using Eq. (13)) are 1.232, 1.279, 1.107, 0.979, 0.883, 0.810, 0.977, 1.092, 0.929, 0.939, respectively. The
Vm c
cs/Vcs (using numerical model) are 1.193, 1.238, 1.097, 0.847, 0.917, 0.841, 0.951, 1.111, 1.057, 1.069, respectively. It can be seen
that the accuracy of Eq. (13) is 20 % higher than that of the original article, while the numerical simulation results are further 3%
higher than that of Eq. (13). Therefore, it can be seen that the use of Eq. (13) and numerical simulation can better reflect the experiment
results.

Fig. 18. The relationship between Vc/(fcbh0) and λ of CARCB [44].

15
B. Da et al. Case Studies in Construction Materials 20 (2024) e03199

Table 7
Measured and simulated results of the Vcs of CARCB under different calculation models.
NO. λ Vm
cs/kN Original article Eq. (13) Numerical simulation

Vccs/kN Vm c
cs/Vcs Vccs/kN Vm c
cs/Vcs Vccs/kN Vm c
cs/Vcs

WL1[43] 0.95 300 193 1.554 244 1.232 252 1.193

WL2[43] 0.95 312 193 1.612 244 1.279 252 1.238
WL3[43] 1.50 237 190 1.244 214 1.107 216 1.097
WL4[43] 1.50 183 190 0.961 214 0.855 216 0.847
WL5[43] 2.00 144 190 0.758 163 0.883 157 0.917
WL6[43] 2.00 132 190 0.693 163 0.810 157 0.841
WL7[43] 1.50 213 226 0.942 228 0.977 224 0.951
WL8[43] 1.50 249 226 1.102 228 1.092 224 1.111
WL9[43] 1.50 184 172 1.070 198 0.929 174 1.057
WL10[43] 1.50 186 172 1.081 198 0.939 174 1.069

Original article: the average value of Vm c

cs/Vcs was 1.101, the standard deviation was 0.286, and the coefficient of variation was 0.259, respectively.
Eq. (13): the average value of Vm c
cs/Vcs was 1.011, the standard deviation was 0.152, and the coefficient of variation was 0.151, respectively.
Numerical simulation: the average value of Vm c
cs/Vcs was 1.032, the standard deviation was 0.131, and the coefficient of variation was 0.127,
respectively.

5. Conclusion

(1) Experimental and numerical simulation results shows that the failure of inclined section suitable reinforcement occurs of
CARCB, and the law of CARCB is consistent with that of OARCB. The load-midspan deflection simulation curve of CARCB closely
matches the measured curve. The two curves almost coincide completely, and the curve is close to a straight line. At this point,
the concrete bears most of the load, and there is no relative slip between steel bar and concrete, resulting in a linear elastic state.
As the load increases and the concrete cracks, the two curves deviate significantly.
(2) The normal section cracking load (Vcr) and shear capacity (Vcs) law of CARCB are as follows: 316 stainless steel bars > ordinary
steel bars > zinc-chromium coated steel bars > organic new-coated steel bars, steel bar type has little influence on the oblique
section Vcr of CARCB, mainly because before the oblique section cracks, the shear stress on the oblique section is mainly borne
by the concrete. At this point, the stirrup stress is very low, and the restraint of the stirrup on the concrete has not yet taken
effect.
(3) Considering the stirrups corrosion and nonlinear mechanical properties of CAC, Vcs calculation formula of CARCB was pro­
posed. The applicability of Vcs calculation formula in C25 ~ C60 CARCB was verified. It was found that Vcs calculated by this
formula increased by 26% and 34% compared with GB 50010–2020 and JGJ/T 12–2019, respectively.
(4) A numerical analysis model suitable for describing the shear behavior of CARCB was proposed based on K&C theory, and the
applicability of the numerical model in C25 ~ C60 CARCB was verified. The model can well describe the whole process of
CARCB inclined section failure mode. The error between the simulated value and the measured value of Vcr, Vcs and midspan
deflection is about 0~18%. The numerical model proposed in this paper increases the accuracy of the shear capacity by 3%
compared with the calculation formula.

CRediT authorship contribution statement

Bo Yu: Methodology. Kai Sun: Writing – original draft, Investigation. Yan Chen: Writing – original draft. Bo Da: Writing – review
& editing, Funding acquisition, Conceptualization. Da Chen: Supervision, Funding acquisition, Conceptualization. Zhangyu Wu:
Methodology. Chengjun Yue: Formal analysis.

Declaration of Competing Interest

The authors declare that they have no conflict of interest. The ethics responsibility is informed and acknowledged to all the authors
as a part of content of right of informed consent.

Data Availability

Data will be made available on request.

Acknowledgements

The authors gratefully acknowledge the financial sup-ports from the National Key Research and Development Program of China
(No. 2023YFB3711800), the Fundamental Research Funds for the Central Universities (No. B230201045), the National Natural Sci­
ence Foundation of China (No. 52208241), the Chunhui Project of Ministry of Education of the People’s Republic of China (No.
HZKY20220123), the Systematic Project of Guangxi Key Laboratory of Disaster Prevention and Structural Safety (No. 2019ZDK006).

16
B. Da et al. Case Studies in Construction Materials 20 (2024) e03199

References

[1] Z.Y. Wu, J.H. Zhang, H.F. Yu, et al., 3D mesoscopic investigation of the specimen aspect ratio effect on the compressive behavior of coral aggregate concrete,
Compos. Part B Eng. 198 (2020) 108025.
[2] L.J. Ma, Z. Li, J.G. Liu, et al., Mechanical properties of coral concrete subjected to uniaxial dynamic compression, Constr. Build. Mater. 199 (2019) 244–255.
[3] B. Da, H.F. Yu, H.Y. Ma, et al., Research on compression behavior of coral aggregate reinforced concrete columns under large eccentric compression loading,
Ocean Eng. 155 (2018) 251–260.
[4] B. Da, Y. Chen, H.F. Yu, et al., Preparation technology, mechanical properties and durability of coral aggregate seawater concrete in the island-reef environment,
J. Clean. Prod. 339 (2022) 130572.
[5] F. Du, W. Zhu, R. Yang, et al., Bioinspired super thermal lnsulating, strong and low carbon cement aerogel for building envelope, Adv. Sci. 10 (2023) e2300340.
[6] C. Liu, Z.Y. Liu, Y.S. Zhang, A multi–scale framework for modelling effective gas diffusivity in dry cement paste: Combined effects of surface, Knudsen and
molecular diffusion, Cem. Concr. Res. 131 (2020) 106035.
[7] R.J. Mi, G.H. Pan, Slowing down CO2 effective diffusion speeds in recycled aggregate concrete by using carbon capture technology and high-quality recycled
aggregate, J. Build. Eng. 45 (2022) 103628.
[8] Y.J. Chu, A.G. Wang, Y.C. Zhu, et al., Enhancing the performance of basic magnesium sulfate cement-based coral aggregate concrete through gradient composite
design technology, Compos. Part B Eng. 227 (2021) 109382.
[9] B. Liu, J.P. Ye, X.Y. Liu, et al., Shear strength and failure criterion of carbon fiber reinforced coral concrete under combined compression-shear stresses, Constr.
Build. Mater. 325 (2022) 126728.
[10] B. Liu, J.K. Zhou, X.Y. Wen, et al., Mechanical properties and constitutive model of carbon fiber reinforced coral concrete under uniaxial compression, Constr.
Build. Mater. 263 (2020) 120649.
[11] Z. Lu, T.N. Di, W. Bo, et al., Corrosion behavior of low alloy steel bars containing Cr and Al in coral concrete for ocean construction, Constr. Build. Mater. 258
(2020) 119564.
[12] Z.Y. Wu, H.F. Yu, H.Y. Ma, et al., Effect of seawater desalination on service life of reinforced coral concrete structures, J. Mater. Sci. Eng. 39 (01) (2021) 82–88.
[13] B. Da, H.F. Yu, H.Y. Ma, et al., Influence of steel corrosion to flexural behavior of coral aggregate concrete beam, J. Cent. South Univ. 27 (05) (2020) 1530–1542.
[14] L.J. Malvar, J.E. Crawford, J.W. Wesevich, et al., A plasticity concrete material model for DYNA3D, Int. J. Impact Eng. 19 (1997) 847–873.
[15] L. Chen, Z.F. Yan, Analysis of basic mechanical behavior of concrete damage model (K&C) in LS-DYNA, J. Hunan Inst. Eng. (Nat. Sci. Ed. ) 27 (02) (2017) 67–70.
[16] R.A. Hawileh, H.A. Musto, J.A. Abdalla, et al., Finite element modeling of reinforced concrete beams externally strengthened in flexure with side-bonded FRP
laminates, Compos. Part B Eng. 173 (2019) 106952.
[17] J.G. Stoner, M.A. Polak, Finite element modelling of GFRP reinforced concrete beams, Comput. Concr. 25 (04) (2020) 369–382.
[18] V. Linh, J. Pitcha, Simulation of shear behavior of corroded reinforced concrete beams flexurally repaired with steel fiber-reinforced concrete, Structures 34
(2021) 1545–1559.
[19] L. Jin, H.S. Yang, R.B. Zhang, et al., Mesoscopic numerical study on shear performance of corroded longitudinal reinforcement concrete beams without
abdominal reinforcement, Chin. J. Eng. 45 (01) (2023) 117–127.
[20] Standard practice for the preparation of substitute ocean water: ASTM D1141—2003 [S]. Philadelphia: American Society for Testing and Materials. 2003.
[21] Code for design of concrete structures: GB 50010-2020 [S]. Beijing: China Architecture ＆ Building Press. 2020.
[22] Thanh Quang Khai Lam, Thi My. Dung Do, The behavior of RC beams strengthened with steel fiber concrete layer by ANSYS simulation, Adv. Civ. Eng. 2023
(2023) 4711699.
[23] H.Y. Ma, B. Da, H.F. Yu, et al., Research on flexural behavior of coral aggregate reinforced concrete beams, China Ocean Eng. 32 (05) (2018) 593–604.
[24] J.X. Ma, L. Yu, B. Li, et al., Stress-strain model for confined concrete in rectangular columns with corroded transverse reinforcement, Eng. Struct. 267 (2022)
114710.
[25] B. Da, H.F. Yu, H.Y. Ma, et al., Experimental investigation of whole stress-strain curves of coral concrete, Constr. Build. Mater. 122 (2016) 81–89.
[26] E.F. Felix, E. Possan, R. Carrazedo, A new formulation to estimate the elastic modulus of recycled concrete based on regression and ANN, Sustainability 13 (15)
(2021) 8561.
[27] A. Lamis, Dynamic measurements for determining poisson’s ratio of young concrete, Nord. Concr. Res. 58 (01)) (2018) 95–106.
[28] A. Brencich, F. Campeggio, Leeb hardness for yielding stress assessment of steel bars in existing reinforced structures, Constr. Build. Mater. 227 (2019) 116570.
[29] B. Da, H.F. Yu, H.Y. Ma, et al., Effect of concrete strength grade on shear performance of reinforced concrete beams in coral seawater, Sci. China Technol. Sci. 49
(02) (2019) 212–222.
[30] B. Da, H.F. Yu, H.Y. Ma, et al., Experimental research on whole stress-strain curves of coral aggregate seawater concrete under uniaxial compression, J. Build.
Struct. 38 (01) (2017) 144–151.
[31] H.F. Yu, B. Da, H.Y. Ma, et al., Durability of concrete structures in tropical atoll environment, Ocean Eng. 135 (2017) 1–10.
[32] H. Karolina, Š. Vít, J. Libor, et al., Prediction of reinforcement corrosion due to chloride ingress and its effects on serviceability, Eng. Struct. 174 (2018)
768–777.
[33] D.H. Du, K. Chen, L.F. Zhang, et al., Effect of zinc on stress corrosion crack growth rate of type 316 stainless steel, Nucl. Power Eng. 38 (02) (2017) 78–83.
[34] B. Da, H.F. Yu, H.Y. Ma, et al., Reinforcement corrosion research based on the linear polarization resistance method for coral aggregate seawater concrete in a
marine environment, Anti-Corros. Methods Mater. 65 (05) (2018) 458–470.
[35] Specification for design of lightweight aggregate concrete structures: JGJ12-2019 [S]. Beijing: China Architecture and Building Press. 2019.
[36] B. Da, H.F. Yu, H.Y. Ma, et al., Durability influencing factors of coral concrete structures in the South China Sea, J. Chin. Ceram. Soc. 44 (02) (2016) 253–260.
[37] Y.Z. Yuan, F.P. Jia, Y. Cai, Structural degradation model of corroded reinforced concrete beams, J. Civ. Eng. 34 (03) (2001) 47–52, 96.
[38] L. Wang, W. Li, S. Chen, et al., Effect of seawater immersion on bonding properties of FRP bars and coral concrete, Acta Mater. Compos. Sin. 35 (12) (2018)
3458–3465.
[39] Y.Z. Yuan, F.P. Jia, Study on mechanical properties degradation of corroded steel bar, J. ]. Ind. Build. 30 (01) (2000) 43–46.
[40] Lin X.C. Study on the Constitutive Relationship of Corroded Steel Bar under Monotonic and Repeated Loads [D]. Xi’an: Xi’an University of Architecture and
Technology, 2014.
[41] W.P. Zhang, C.K. Li, X.L. Gu, et al., Study on shear behavior of corroded reinforced concrete simple supported beams with inclined section, J. Build. Mater. 17
(05) (2014) 920–926.
[42] Zhang W. Experimental study on reinforced coral concrete component [D]. Nanjing: Hohai University. 1995.
[43] J.W. Zhang, Experimental Study On Shear Performance Of Coral Concrete Beams With CFRP bars [D], Guilin University of Technology, Guilin, 2020.
[44] Wang Y.A. Shear behaviour of seawater sea-sand coral aggregate concrete beams reinforced with CFRP strip stirrups [D]. Nanchang: East China Jiaotong
University. 2023.

17