materials

Article
Numerical Simulation of Dynamic Mechanical
Properties of Concrete under Uniaxial Compression
Yijiang Peng 1 , Qing Wang 1 , Liping Ying 1, *, Mahmoud M. A. Kamel 1,2 and
Hongtao Peng 3
1 Key Laboratory of Urban Security and Disaster Engineering, Ministry of Education,
Beijing University of Technology, Beijing 100124, China; pengyijiang@bjut.edu.cn (Y.P.);
wangq@emails.bjut.edu.cn (Q.W.); mahmoud.kamel@fayoum.edu.eg (M.M.A.K.)
2 Department of Civil Engineering, Faculty of Engineering, Fayoum University, 63514 Fayoum, Egypt
3 College of Water Resources and Civil Engineering, China Agricultural University, Beijing 100083, China;
pwb@cau.edu.cn
* Correspondence: qiaoliang@rcees.ac.cn; Tel.: +86-188-1142-3517




Received: 29 January 2019; Accepted: 15 February 2019; Published: 20 February 2019


Abstract: Based on the base force element method (BFEM), the dynamic mechanical behavior of
concrete under uniaxial compression loading at different strain rates is investigated. The concrete
can be considered as a three-phase composite material composed of aggregate, cement mortar, and
interfacial transition zone (ITZ) on the meso-level. A two-dimensional random aggregate model is
generated by the Monte Carlo method. A multi-linear two-dimensional damage model is applied
to describe the damage properties of each phase in the concrete. The strain-softening behavior,
strain-rate effect, and failure patterns of the concrete are studied. The numerical results find that the
peaks of compressive stress and compressive strain of concrete show the rate-sensitivity in various
degrees under different strain rates. The calculated results of the dynamic enhancement factors are in
a good agreement with the formula given by the Comité Euro-International du Béton (CEB) and other
experimental results. The failure diagram of the specimen clearly describes the compressive failure
process of the concrete specimen. This failure’s characteristics are similar to the experimental results.

Keywords: concrete; base force element method; strain-rate effect; meso-damage; dynamic behavior;
numerical simulation

1. Introduction
In practice, the concrete used in infrastructure is usually subjected to dynamic loading, including
impact loading and sustained loading. The previous obtained results show that the mechanical
properties and damage characteristics of concrete under dynamic loading are very different from those
under static loading [1,2]. Therefore, it is of interest to investigate the dynamic behavior of concrete
under dynamic loading at high strain rates.
Hitherto, much research based on the traditional experimental test has been conducted. Abram [3]
was the first to find out the compressive strength of concrete and show the strain rate sensitivity under
dynamic loading. Bischoff et al. [4] concluded that the strain rate plays a significant role in both the
dynamic ultimate strength and the dynamic deformation behavior of plain concrete at high strain
rates. Ross et al. [5] and John et al. [6] respectively carried out Split-Hopkinson pressure bar (SHPB)
tests to study the dynamic behavior of concrete under tensile stress. Malvar et al. [7] and Williams [8]
respectively conducted literature reviews to investigate the strain-rate effects on the dynamic increase
factor (DIF) of strength and elastic moduli; thus, the empirical equations were given correspondingly.
In addition, the mechanical properties of the interfacial transition zone (ITZ) and the effect of the ITZ

Materials 2019, 12, 643; doi:10.3390/ma12040643 www.mdpi.com/journal/materials

Materials 2019, 12, 643 2 of 15

on the properties of rock-concrete materials were investigated [9]. Erzar et al. [10] developed direct
tensile tests, spalling tests, and edge-on impact tests at high strain rates. The influence of the ITZ
between the aggregates and the cement mortar on the uniaxial dynamic tensile strength of concrete
was also studied. The SHPB apparatus was used to study the mechanical properties of cement-based
materials. Deformation and stress distribution in the specimens are non-uniform due to the composite
microstructure of the materials. The strength and failure patterns were also studied in both quasi-static
and dynamic loadings [11–17]. Li et al. [18] used both cylinder and cube concrete specimens to conduct
compressive tests and the effects of the specimen shape and size on concrete strength subjected to
different loading rates were investigated. Jurowski and Grzeszczyk [19] concluded that the stabilized
dynamic elastic modulus of concrete is proportional to the initial static elastic modulus, and the
coefficient of proportionality is affected by the type of aggregates. Using various inner diameters of
specimens, Zhang et al. [20] investigated the inertial effect on the tensile strength of concrete materials
under dynamic loading.
In recent years, numerical analysis with computer modeling has become an efficient method to
study the properties of concrete. On the basis of a damage parameter, Simo and Ju [21,22] proposed
a continuum isotropic model and an anisotropic elastoplastic-damage model, and the results are
consistent with the existing experimental data. The mechanical properties of small-eccentric loaded
reinforced concrete (RC) columns and the dynamic behavior of steel fiber-reinforced concrete (SFRC)
beams under impact loading were investigated using Abaqus software [23,24]. Furthermore, finite
element theory is one of the major methods used to analyze the mechanical properties of cement-based
composite material [25–28]. In the meso-scale, concrete can be deemed as a three-phase composite
which comprises aggregate and cement mortar, with an ITZ between both. Regarding concrete studies,
several scholars around the world have emphasized the relationship between microstructure and
macroscopic mechanical properties of concrete under dynamic loading with a numerical method.
Georgin et al. [29] exploited a viscous plastic model to simulate the SHPB test and studied the
influence of inertial force and strain rates on concrete dynamic behavior. Snozzi et al. [30,31] and
Gatuingt et al. [32,33] proposed a computational model to investigate the mechanical properties
of concrete, composed of aggregate and mortar paste matrix, under dynamic loading of tension
and compression. Also, Park et al. [34] analyzed the influence of impact loading at high strain
rates on concrete-like materials using a dynamic finite element simulation. Zhou et al. [35,36] and
Hao et al. [37,38] adopted numerical methods to analyze the influence of the ITZ on the dynamic failure
patterns of the concrete, which is considered as a three-phased composite consisting of aggregate,
mortar, and ITZ. Additionally, with the energy theory and the micro prestressed-solidification theory
applied, Cusatis [39] utilized the developed Confinement Shear Lattice (CSL) model to analyze
the effect of strain rates on the strength and failure behavior of concrete. However, Wu et al. [40]
established a new experimental method of numerical simulation to identify the rate sensitive to the
concrete dynamic tensile behavior. In this regard, in order to simulate the dynamic behavior of concrete
under tensile, Zhou et al. [41] presented a two-dimensional meso-scale finite element model, validated
by comparing with the experimental data from spall tests. Chen et al. [42] formulated a new dynamic
compressive constitutive model applied to investigations of strain-rate effects and damage effects
within the specimen.
To date, the relationship between the failure mechanism of concrete and the stress distribution at
high strain rate is not clear; however, the mesoscopic components remarkably affect the macroscopic
mechanical properties of plain concrete. Therefore, on the meso-level, it is important to know the
dynamic behavior of the concrete.
In this paper, according to the investigations of concrete-like materials, under static loading with
the base force element method (BFEM) proposed by Peng et al. [43–45], a dynamic base force element
model is developed. In addition, the failure process of concrete under uniaxial dynamic loading at high
strain rates is simulated. Strain-softening behavior and the failure process of modeled concrete under
Materials 2019, 12, 643 3 of 15
Materials 2019, 11, x FOR PEER REVIEW 3 of 16

2. Establishment of the Dynamic Base Force Element Equilibrium Equation

uniaxial compressive loading is also investigated. Finally, the effects of different loading velocities on
compressive strength
Generally, and the of
the behavior stress
realdistribution in the specimens
physical structures aretostudied.
subjected loadings or displacements is
almost dynamic. According to the D'Alembert's principle, considering the actions of inertial force
2. Establishment
and damping force,of the
the Dynamic Base Force
balance equation of aElement Equilibrium Equation
single-degree-of-freedom system can be obtained as
follows:
Generally, the behavior of real physical structures subjected to loadings or displacements is
almost dynamic. According to the D’Alembert’s principle, considering the actions of inertial force and
[ M ]{u(t )} + [C ]{u ( t )} + [ K d ]{u ( t )} = { P ( t )} (1)
damping force, the balance equation of a single-degree-of-freedom system can be obtained as follows:
in which {u(t )} , {u (t )} ,
and {..u (t )} respectively
 . indicate the acceleration vector quantity, the
[ M] u(t) + [C ] u(t) + [Kd ]{u(t)} = { P(t)} (1)
velocity vector quantity, and the displacement vector quantity of each node in structure; [ M ]
 ..  .
means
in whichglobal
u(t)mass t) , and [{Cu](t)}
, u(matrix; means the damping
respectively indicatematrix; [ K d ] is the
the acceleration global
vector stiffness
quantity, matrix
the velocity
vector on
based quantity, andforce
the base the displacement
as proposed in vector quantity
the paper; and {P ( t )}node
of each dynamic [load
in structure;
is the M] means global
array of the
mass matrix; [C ] means the damping matrix; [Kd ] is the global stiffness matrix based on the base force
structure.
as proposed in the paper; and { P(t)} is the dynamic load array of the structure.
Furthermore, the full variable form of the Newmark-β method is adopted to solve the dynamic
Furthermore, the full variable form of the Newmark-β method is adopted to solve the dynamic
equilibrium equation in the study.
equilibrium equation in the study.
2.1. Base Force Element Stiffness Matrix
2.1. Base Force Element Stiffness Matrix
According to the BFEM, which is based on the potential energy principle, a plane triangular
According to IJthe BFEM, which is based on the potential energy principle, a plane triangular
element matrix KKI J [46],
element matrix [46],expressed
expressedbyby base
base force,
force, can
can bebe obtained
obtained asas follows
follows (shown
(shown inin Figure
Figure 1):1):
E  2ν  
K IJ = E 2ν m I ⊗ m J + m IJ U + m J ⊗ m I  I
K I J = 2 A(1 + ν ) 1 − 2ν m I ⊗ m J + m I J U + m J ⊗ m 
(2)
(2)
2A(1 + ν) 1 − 2ν
where
where E,E ,ν, and AAmean
ν ,and meanthe theYoung’s
Young’smodulus,modulus,thethePoisson’s
Poisson’s ratio,
ratio, and
and the
the area
area of
of an
an element,
element,
I J K
respectively; U is the unit tensor; I u J, uK , u represent the displacements of the triangular
respectively; U is the unit tensor; u , u , u represent the displacements of the triangular element element
vertex.
vertex.
uJ

uI

uK
Figure 1. A triangular element.
Figure 1. A triangular element.
For plane strain problems, x and y represent Cartesian coordinate system, and the element matrix
KI JFor
can beplane strainas
expressed problems,
follows: x and y represent Cartesian coordinate system, and the element
IJ
matrix K can be expressed as follows: " 2−2ν I J J J J #
E m x m x + myI my 1−2ν2ν m xI my + myI m x
 2 − 2ν I J
− 2ν
 e
J 
1 2ν
KI J = mxI mxJ + myI m J I J I (3)
2A(1E+ ν)  2ν m 2−2ν mxIm y J + m y m
2ν2ν my my + m x mx
I y J Ix J
1 −1−2ν2ν y m x + m x my 1 −
[ K IJ ] =
e 1−
  (3)
2 A(1 + ν )  2ν I J I J 2 − 2ν I J I J 
I J I J I
where m = m · m , m and m can be J m m + m m m m + m m
 1described
− 2ν
y x as follows
x y
1and
y y x x 
− 2νare shown in Figure 
2:

1 I I J m J can 1 1
IJ I
where mm I I
==mm ⋅i m J
ei =, m( Land
IJn + LKI nKI be ( L I J niI J eas
) =described i+ LKI niKIand
follows ei ) =are (shown
I J in Figure
L I J ni + LKI niKI )2: ei , (4)
2 2 2
1 1 1
m IJ = m I e1 = ( LJK IJ
IJ n+ L + LKIInJ KI ) =1 ( LIJ niIJJKei + LKI niKI eI iJ ) = ( L1IJ niIJ + LJKKI niKI )ei , I J (4)
m J = mi ei =i i ( L JK n n ) = ( L n e + L n e ) = ( L JK ni + L I J ni )ei , (5)
2 2 22
IJ JK i i IJ i i 2
2
( ) ( ) ( )
I J
m xI 1 nx nKIx
J J 1 JK = IJ ( L I J1 I J JK + L LI IJ KI 1 ), JK (6)
m = mi ei = ( LJK m n y + LIJ n2 ) = ( Ln
I
n
JKy i e i + L n
IJ i e i y=
n) ( LJK ni + LIJ niIJ )ei , (5)
2 2 2
 Materials 2019, 12, 643 4 of 15

( ) ( ) ( )
J JK IJ
mx 1 nx nx
J = ( L JK JK + LI J IJ ). (7)
my 2 ny ny
in which I, J, K represent the vertexes of the triangular element; L I J , LKI , L JK are the lengths of the
Materials 2019, 11, x FOR PEER REVIEW 4 of 16
element boundary lines; n , n , n JK are the normal vectors of the element boundary lines. Average
IJ KI

strain components, which could take the place of the real strain in the case of small deformation,
are given as follows:
mxI  1 nnxIJ  nxKI 
 I  = εx (=LIJ ∑IJ (u+Ix m
1 LLIxI ), KI ) , (6)(8)
m y  2 AIn=y1  n y 
1 n
A I∑
εy = (u Iy myI ), (9)
m  1 nx  n 
J JK
=1 IJ
x x
  = (1LJKn  JK  + L  ) .
J IJ IJ (7)
m  =2 A ∑ (unIxy myI + n 
yγ xy u Iy m xIy). (10)
I =1

n JK J

mJ n IJ
mK
K

mI
I
n KI
Constructionofofaatriangular
Figure2.2.Construction
Figure triangularelement
elementstiffness.
stiffness.

2.2. Damping Matrix and Mass Matrix

in which I , J , K represent the vertexes of the triangular element; LIJ , LKI , LJK are the lengths of the
In the finite element dynamic
IJ analysis, lumped mass matrix, one of the most common forms of the
element boundary lines; n , nKI , n JK are the normal vectors of the element boundary lines. Average
element mass matrix, is applied to simplify the calculation and reduce the storage space. According to
strain components, which could take the place of the real strain in the case of small deformation, are
the equivalent static principle, the mass of the element can be evenly distributed among three vertexes
given as follows:
of a triangular element, according to the hypothesis that there is no interaction between the inertial
forces of each vertex. n


1
x) ,
In dynamic response problems it εisx =generally (uIx m I
assumed that the viscous damping force
(8) is
A
proportional to the velocity of particle motion. I =1
Therefore, the effect of damping in the dynamic structural analysis at a high loading rate cannot be
ignored. As a widely used orthogonal damping model, the Rayleigh damping model can be expressed
n


as follows: 1
ε β[yK) d, ],
I
y =
[C ] = Aα[ M](u+Iy m (9)
(11)
I =1
2ζω1 ω2
α= , (12)
ω1 + ω2
n


1 2ζ
γ xy = β = (uIx myI + u,Iy mxI ) . (13)
(10)
A ω1 + ω2
I =1
 
1 0 0 0 0 0
 0 1 0 0 0 0 
 
2.2. Damping Matrix and Mass Matrixe ρb TA  0 0 1 0 0 0 

[ M] = , (14)
3g lumped
 0 0 0 1 0 0 

In the finite element dynamic analysis,  mass matrix, one
 of the most common forms of
 0 0 0 0 1 0 
the element mass matrix, is applied to simplify the calculation and reduce the storage space.
According to the equivalent static principle, the mass0 of0 the
0 0 0 element
1 can be evenly distributed
among three vertexes of a triangular element, according to the hypothesis that there is no interaction
between the inertial forces of each vertex.
In dynamic response problems it is generally assumed that the viscous damping force is
proportional to the velocity of particle motion.
Therefore, the effect of damping in the dynamic structural analysis at a high loading rate cannot
be ignored. As a widely used orthogonal damping model, the Rayleigh damping model can be
  
ρbTA 0 0 1 0 0 0 
[M ]
e
=   , (14)
3g 0 0 0 1 0 0
0 0 0 0 1 0
 
Materials 2019, 12, 643  0 0 0 0 0 1  5 of 15

whichα αandand
ininwhich β are
β are the the
mean coefficients
mean of proportionality,
coefficients of proportionality,which cancan
which bebe
calculated
calculatedfrom
from
Equations (6) and (7); means the damping ratio;
Equations (6) and (7); ζ means the damping ratio;
ζ ω 1 ω1 2and ω2 are the first and the second order
and ω are the first and the second order angular
frequencies of the specimen.
angular frequencies of the specimen.
It should be noted that, in the process of the meso-mechanical failure of concrete, there are few
It should be noted that, in the process of the meso-mechanical failure of concrete, there are few
research works on damping models. The damping theory is a methodology which aims to describe
research works on damping models. The damping theory is a methodology which aims to describe
the microscopic mechanism of damping in a macroscopic way. For the general structural analysis,
the microscopic mechanism of damping in a macroscopic way. For the general structural analysis,
Rayleigh damping is usually used to approximately describe the damping characteristics. Normally,
Rayleigh damping is usually used to approximately describe the damping characteristics. Normally,
the damping ratio of the engineering structure ranges from 0.01 to 0.1. The average value 0.05 is
the damping ratio of the engineering structure ranges from 0.01 to 0.1. The average value 0.05 is
applied in this paper.
applied in this paper.
In this paper, to simplify the model, the influence of damage development on damping matrices
In this paper, to simplify the model, the influence of damage development on damping matrices
is not considered. Consequently, the initial stiffness matrix of the specimen is adopted in the present
is not considered. Consequently, the initial stiffness matrix of the specimen is adopted in the present
dynamic analysis and the damping ratio is regarded as a constant. It is assumed that the mass matrix
dynamic analysis and the damping ratio is regarded as a constant. It is assumed that the mass matrix
is independent from the damage state due to the conservation of mass.
is independent from the damage state due to the conservation of mass.
3. Meso-Structure of Concrete
3. Meso-Structure of Concrete
It is well known that concrete is a multi-phase heterogeneous brittle material. The mechanical
It is well known that concrete is a multi-phase heterogeneous brittle material. The mechanical
properties of concrete are determined by components in its mixture. In this paper, concrete is described
properties of concrete are determined by components in its mixture. In this paper, concrete is
as a three-phase composite material at the meso-level, composed of coarse aggregates, cement mortar,
described as a three-phase composite material at the meso-level, composed of coarse aggregates,
and an ITZ between both phases. With the spherical aggregate applied, the microstructure in concrete
cement mortar, and an ITZ between both phases. With the spherical aggregate applied, the
can be depicted as in Figure 3. In the model only coarse aggregates with a particle size greater
microstructure in concrete can be depicted as in Figure 3. In the model only coarse aggregates with a
than 5 mm are represented clearly, while the other smaller aggregates are mixed up in the cement
particle size greater than 5 mm are represented clearly, while the other smaller aggregates are mixed
paste matrix.
up in the cement paste matrix.

Figure 3. Three-phase sphere model of concrete.

3.1. Random Aggregate Model

In this paper, in order to simulate the real distribution of aggregate as possible, the Monte
Carlo method is used to establish the two-dimensional random aggregate model. Based on the
Fuller three-dimensional aggregate gradation curve, Walraven [47] developed the two-dimensional
cross-section aggregate gradation curve. The cumulative probability of aggregate particles satisfied
the condition of D < D0 (herein D means the diameter of the aggregate) is calculated using the
following formula:
"
D0 1/2 D0 4 D0 6
     
Pc ( D − D0 ) = Pk 1.065 − 0.053 − 0.012
Dmax Dmax Dmax
# (15)
D0 8 D0 10
   
−0.0045 + 0.0025
Dmax Dmax
 following formula:

  D 
1/ 2
 D 
4
 D 
6

Pc ( D − D0 ) = Pk 1.065  0  − 0.053  0  − 0.012  0 

  Dmax   Dmax   Dmax 
(15)
 D0  
8 10
Materials 2019, 12, 643  D0  6 of 15
−0.0045   + 0.0025   
 Dmax   Dmax  

in which Pk represents
in which the percentage
Pk represents the percentageof the volume
of the volumeofofaggregate
aggregate in
in the wholeconcrete
the whole specimen; Pc
concretespecimen;
represents of aggregate particles with a size smaller than D Dmaxisis the
Pc represents the cumulative probability of aggregate particles with a size smaller than D0 0; ; D
the cumulative probability
max
maximum diameter of the aggregate particles.
the maximum diameter of the aggregate particles.
The number
The numberof aggregate particles
of aggregate meeting
particles meetingthethe requirementofofD1D1<< D
requirement D <<DD 2 can be calculated by
2 can be calculated
the following equation:
by the following equation:
n = [ Pc ( D < D2 ) − Pc ( D < D1 )] × A/Ai (16)

where A is the cross-section area of  Pc (concrete

n =the D < D2 ) − Pc ( D < D1 ) 
specimen × A / AAi is the area of the representative
 and (16)
i
particle diameter of the aggregate.
where A is to
According thethe
cross-section
theory ofarea of the maximum
Fuller’s concrete specimen
density and Ai is the
curve, area
three of the representative
representative diameters
particle diameter of the aggregate.
(10 mm, 20 mm, and 32.5 mm) of aggregates are selected to calculate number of aggregate particles.
According
The numbers to the theory66,
are, respectively, of 11,
Fuller's
and maximum density
3, as obtained fromcurve, three representative
Equation diameters
(20). The aggregates (10
calculated
mm, 20 mm, and 32.5 mm) of aggregates are selected to calculate number of aggregate particles.
are put into the two-dimensional region, the same size as concrete specimen, by using the Monte
The numbers are, respectively, 66, 11, and 3, as obtained from Equation (20). The aggregates
Carlo method with three sets of different random numbers. Each aggregate should be in line with
calculated are put into the two-dimensional region, the same size as concrete specimen, by using
the boundary conditions and not overlap with the other aggregates before it is put on. The modeled
aggregates are shown in Figure 4.

(a) (b) (c)

Figure
Figure 4. Random
4. Random aggregate
aggregate modelsofofconcrete.
models concrete. (a)
(a)Specimen
Specimen1;1;
(b)(b)
specimen 2; (c)
specimen 2;specimen 3. 3.
(c) specimen

3.2. Mesh Generation

3.2. Mesh Method
Generation and
Method Element
and ElementAttributes
Attributes

In theInfinite
the finite element
element analysis,
analysis, twotwo common
common methodsare
methods areadopted
adoptedto todeal
deal with
with the thickness
thicknessofof ITZ
ITZto
in order insimplify
order to the
simplify the calculation.
calculation. One considers
One considers the thickness
the thickness of the
of the ITZ ITZ0.5
from from
mm0.5 to mm
2 mm to[48,49],
2
mm [48,49], and the other does not consider the thickness of the ITZ [50–53].
and the other does not consider the thickness of the ITZ [50–53]. In this paper, the dynamic behavior In this paper, the
of concrete is studied
Materials 2019, 11, xby
FORusing
PEER the former method and the size of the mesh element is deemed7 as
REVIEW the
of 16

ITZ thickness to simplify the mesoscopic model.

dynamic behavior of concrete is studied by using the former method and the size of the mesh
The finite element mesh is shadowed on the random aggregate model. The different mechanical
element is deemed as the ITZ thickness to simplify the mesoscopic model.
properties areThe assigned to the corresponding elements. A linear elastic triangular finite element grid is
finite element mesh is shadowed on the random aggregate model. The different mechanical
applied in the paper,
properties areshown
assigned intoFigure 5. The type elements.
the corresponding of element A is determined
linear by the position
elastic triangular of element
finite element grid
nodes. The element
is applied could
in the be deemed
paper, shown inas an aggregate
Figure 5. The type (orofcement
elementmortar) element
is determined by when the three
the position of
nodes ofelement
the element
nodes. are allelement
The locatedcould
at thebeaggregate
deemed as(orancement
aggregate mortar) region;
(or cement otherwise,
mortar) elementthe element
when the
is the ITZthree
when nodes
the of the element
element are allin
is located located at the
the both theaggregate
aggregate (orand
cement
cementmortar) region; otherwise, the
mortar.
element is the ITZ when the element is located in the both the aggregate and cement mortar.

Aggregate
ITZ

Cement mortars

5. The5.diagram
FigureFigure of mesh
The diagram generation.
of mesh generation.

4. Dynamic Behavior for Concrete Meso-Components

4.1. Concrete Dynamic Damaged Model

As known, the mechanical properties of microscopic components have a great influence on the
Materials 2019, 12, 643 7 of 15

4. Dynamic Behavior for Concrete Meso-Components

4.1. Concrete Dynamic Damaged Model

As known, the mechanical properties of microscopic components have a great influence on the
fracture damage behavior of concrete. Therefore, the heterogeneity of concrete should be taken into
account in this numerical simulation. In this paper, concrete can be treated as a three-phase composite
material composed of coarse aggregate, cement mortar, and the ITZ. Each phase material in concrete
is assumed to be homogeneous and isotropic. The constitutive relation is presented in Figure 6.
The maximum principal stress criterion is applied as the failure criteria in the study. The reduction of
the elastic modulus E of the material in the damage processes can be expressed as follows:

E = E0 (1 − D ) (17)
Materials 2019, 11, x FOR PEER REVIEW 8 of 16

in which E0 is the initial modulus of elasticity and D is the damage factor, defined as follows:
 β
0 1− ε maxε max≤ ε≤t0λε c 0
γ


1 − ε t0 + ε max −ε t0 ε t0 (1 − α) ε < ε max ≤ ηt ε



Dt = 1 − 1ε−max β ε maxηt−ε t0λε
−εco t0 ε max ε co
t0 t0
 ε − η ε − β λε c 0 < ε max ≤ ε c 0
(18)
1 −1ξ−t −ληt εε max
α max t t0
ε
αε
+ ε maxt0
ηt ε t0 < ε max ≤ ξ t ε t0


max

 max

Dc =1 1 − γ ε max − ε co ε co ε max > ξ t ε t0 (19)
 1 − 1 −η − ε c 0 < ε max ≤ ηc ε c 0
ε ε
 β
c max max

 1− γ γε ε max ≤ λε c0



1− β1ε−
c 0
ηc ε c 0 < ε max ≤ ξ c ε c 0
ε −λε co
 ε co


 1 − 1− λ
max
ε max
max
− β ε max
λε c0 < ε max ≤ ε c0

Dc = 1 − 1−η ε 1−γ ε max 1 − ε co ε
− ε maxco
ε c0 < εε max
max>≤ ξ cηε c0ε c0 (19)
 c max
 γε c0
1 − ε max ηc ε c0 < ε max ≤ ξ c ε c0





1 ε max > ξ c ε c0

in which ε0 means the principle strain; η denotes the residual strain coefficient; ξ represents the
in which ε 0 means the principle strain; η denotes the residual strain coefficient; ξ represents the
ultimate strain coefficient.
ultimate strain coefficient.
In Figure 6, fc and ft stand for compressive and tensile strength; furthermore, the subscripts
In Figure 6, f c and f t stand for compressive and tensile strength; furthermore, the subscripts t
t and
and c respectively
c respectively symbolize
symbolize tensiletensile and compressive
and compressive features.features.
The material The material parameters
parameters are given are
in
given in Table 1, which is obtained after
Table 1, which is obtained after calculations. calculations.

fc

β fc

γ fc
ξ t ε t 0 ηt ε t 0 ε t 0 0
λε c 0 ε c0 ηc ε c 0 ξcε c 0 ε
α ft

ft

Figure 6. Mechanical constitutive model of materials.

Figure 6. Mechanical constitutive model of materials.

Table 1. Mechanical parameters of materials.

Interfacial
transition
Mechanical parameters Cement mortar
zone
(ITZ) Aggregate
Density ρ (kg/m ) 3
Materials 2019, 12, 643 8 of 15

Table 1. Mechanical parameters of materials.

Interfacial Transition
Mechanical Parameters Cement Mortar Aggregate
Zone (ITZ)
Density ρ (kg/m3 ) 2100 1700 2700
Poisson’s ratio ν 0.22 0.2 0.16
Strength (tensile/compressive) σ (MPa) 3.2/32 2.5/25 7/70
λ 0.25 0.25 0.80
β 0.85 0.65 0.90
γ 0.35 0.35 0.35
α 0.3 0.3 0.3
ηt /ηc 4/4 3/3 5/5
ξ t /ξ c 10/10 10/10 10/10

4.2. Dynamic Increase Factor (DIF) for Concrete

In practice, damage patterns and mechanical properties of concrete under dynamic loading
present different forms, which is called the strain-rate effect and is characterized by the dynamic
increase factor (DIF). The dynamic increase factor for the compressive strength is recommended by
Comité Euro-International du Béton [54] as the following:
  
. 1.026α .
 .ε ε ≤ 30s−1

DIFc = f cd / f c 0 = εs
 . 1/3 .
(20)
 γ .ε

ε
ε > 30s−1
s

in which f cd is the dynamic uniaxial compressive strength and f c 0 is the quasi-static uniaxial
. .
compressive strength; ε is the quasi-dynamic strain rate and εs is the quasi-static strain rate;
f cs is the quasi-static uniaxial compression strength and f c0 = 10 MPa. It should be noted that
α = 1/(5 + 9 f cs / f c0 ) and lgγ = 6.156α − 2. As known from the empirical formula, the value of
.
DIFc is 1 when the strain rate ε is 30 × 10−6 s−1 , which is called the quasi-static load modal. In this
paper, the minimum strain rate 10−3 s−1 is deemed as the quasi-static strain rate for comparison.
According to the previous study, the strain-rate sensitivity of concrete characteristics, such as the
Poisson’s ratio, elasticity modulus, energy dissipation capacity, and so forth, are much lower than the
tensile and compressive strength of concrete [55].

5. Numerical Examples and Results

5.1. Boundary Conditions and Loading Model

In this section, the dynamic test of concrete subjected to uniaxial compressive loads is simulated
with the BFEM. A standard concrete specimen, with the dimensions 150 mm × 150 mm, is chosen
to conduct the test. The particle diameter of the coarse aggregates ranges from 5 mm to 40 mm.
The boundary conditions are shown in Figure 7. The bottom and top surfaces of the specimen are
restricted only in the y-direction; other surfaces of the specimen are free in all directions. The influence
of friction between the specimen and the loaded end on the compressive strength of concrete is ignored.
Continuous and uniform vertical displacement loading is used on the specimens. The direction of
vertical displacement loading is parallel to the y-axis. Dynamic displacement step-load is applied in
this study and the velocity of displacement loading is well controlled by the duration of each load-step.
.
Vertical strain rate εm under different loading velocity can be calculated as the following:

. 2v
εm = (21)
h
where v is the constant loading velocity and h is the height of the specimen. The loading curves of
concrete specimens are shown in Figure 8.
 Materials 2019, 11, x FOR PEER REVIEW 10 of 16

Materials
Materials2019, 12,11,
2019, 643x FOR PEER REVIEW 910ofof
15 16

Figure 7. Loading model of compression test.

Figure 7. Loading model of compression test.
Figure 7. Loading model of compression test.
2
0 ε=10 /s
7.5×10
1
-1 ε=10
ε=10/s/s
2
7.5××
7.5 10100
0
-2 εε=10
=10/s/s
1
7.5×
7.5 ×10 -1
10
)

-1
-3 ε=10
ε=10/s/s
0
7.5×
7.5 ×10 -2
10
) (

-2
-4 ε=10
ε=10/s/s
-1
7.5×
7.5 ×10 -3
10
(

-3
-5 ε=10
ε=10/s/s
-2

7.5×
7.5 ×10 -4
10
-3
ε=10 /s
7.5×100 0
-5
-3 -2 -1
5×10 5×10 5×10
t (ms)
0
0 -3
×dynamic -2
5×10 compression
-1
5×10 test.
Figure 8. Loading curve 5of 10 uniaxial
t (ms)
Figure 8. Loading curve of dynamic uniaxial compression test.
5.2. Dynamic Failure Behavior of Concrete under Uniaxial Compressive Stress
5.2. Based
Dynamic Figureof
8. Concrete
Loading curve of dynamicCompressive
uniaxial compression test.
on Failure Behavior
the meso-mechanical undersimulation
numerical Uniaxial proposed in Stress
the present study, the specimen
can beDynamic
5.2. subdivided
Based theinto
onFailure triangular
meso-mechanical
Behavior finitenumerical
of Concrete element
under Uniaxial meshes whereproposed
simulation
Compressive the Stress
meshin size
theis 1present
mm. Withstudy, eightthe
groups
specimen withcan different loading rates
be subdivided intoapplied,
triangular dynamic
finite uniaxial
elementcompression
meshes where tests aremesh
the carried
sizeout
is on the
1 mm.
With Based
concrete specimens.
eight on theThe
groups meso-mechanical
with corresponding
different loading numerical
macroscopic
rates applied,simulation
nominaldynamic proposed
strain rates arein10the
uniaxial −3 /s,present
10−2 /s,study,
compression 10−1 /s,
tests the
are
specimen
1/s, 10/s,out
carried can
30/s, be
on80/s, subdivided
and 100/s.
the concrete into triangular finite element meshes where
specimens. The corresponding macroscopic nominal strain rates are the mesh size is 1 mm.
With
-3 eight -2 groups
10 / s , 10 / s , 10 / s ,
The stress–strain -1 with different loading rates applied, dynamic uniaxial compression tests are
curves 1/ s
of 10 / s , 30/
, concrete s , 80/ s ,are
specimens andshown100 / in
s . Figure 9. The peak values of strain and
carried
stress The out
are shown on
stress–strainthe
in Figureconcrete
curves specimens.
10 atofthe different
concrete The corresponding
strain-rates.
specimens The damage
are shown macroscopic
in Figureprocess nominal
9. Theof peak
concretestrain rates
specimens
values of are
strain
10
under
-3
/
and stress s , 10
different
-2
/ s , 10
are strain-rate
-1
/ s , 1/
shown in compression s , 10 / s , 30/
Figure 10 atlevels s , 80/ s , and 100
is shownstrain-rates.
the different / s .
in Figure 11.The Thedamage
failure patterns
processatofdifferent
concrete
rates The
are stress–strain
illustrated in curves
Figure of
12. concrete
The specimens
distribution of are
the shown
maximum
specimens under different strain-rate compression levels is shown in Figure 11. The failure patterns in Figure
principal9. The peak
stress in values
the of strain
damaged
.
and
concrete stress
at different is shown are
rates inshown
areFigure in Figure
13. It in
illustrated 10
is Figure at the
to be noted different
12. The strain-rates.
thatdistribution
ε means theof strain The damage
rate, ε is the
the maximum process
strain, and
principal of concrete
σ isinthe
stress the
specimensconcrete
compressive
damaged under In
stress. different
shownstrain-rate
isFigure 13,
inthe compression
positive
Figure 13. It is to belevels
numbers is that
represent
noted shown in Figure
theεelements
means the11.strain
under The failure
tensile εpatterns
stress
rate, isand
the
theatnegative
different
strain, σrates
and numbers arerepresent
is the illustratedthe
compressive in stress.
FigureIn
elements 12.are
The
Figure distribution
under 13,compressive of thestress.
the positive maximum
numbers principal
represent stress
the in the
elements
damaged concrete is shown in Figure 13. It is to be noted
under tensile stress and the negative numbers represent the elements are under compressive stress. that ε means the strain rate, ε is the
strain, and σ is the compressive stress. In Figure 13, the positive numbers represent the elements
under tensile stress and the negative numbers represent the elements are under compressive stress.
 Materials
Materials2019,
2019,11,
11,xxFOR
FOR PEER
PEER REVIEW
REVIEW 11 16
11 of of 16

Materials 2019, 11, x FOR PEER REVIEW 11 of 16

80 specimen11
specimen
Materials 2019, 12, 643 specimen22
specimen 10 of 15
70
specimen33
specimen -1 -1
ε=100s
ε=100s
60 -1 -1
80 specimen 1 ε=80s
ε=80s
50 specimen 2
70

MPa))
specimen 3 -1

σ(MPa
ε=30s
ε=30s
-1
ε=100s
-1
40
40
60

σ(
-1
30 -1 ε=80s
30 ε=10s
ε=10s
-1
50

σ(MPa)
20
20 -1
ε=1s -1
-1
ε=30s
40 ε=1s
-1 -1
10 ε=0.01s -1 ε=0.1s -1
10 ε=0.01s -1 ε=0.1s
30 ε=0.001s -1 ε=10s
-1
0 ε=0.001s
00 1000 2000 3000 -1 4000 5000 6000
20 0 1000 2000 ε3000
=1s -6 -1 4000 5000 6000
-1 -1 ε(×10 S-6 )-1
10 ε=0.01s ε=0.1s ε(×10 S )
-1
ε=0.001s
0
Figure 9. Stress–strain curve
0 of concrete
1000 2000under dynamic
3000 4000uniaxial
5000compressive
6000 stress at different
Figure 9. Stress–strain curve of concrete under dynamic uniaxial compressive stress at different
strain rates. -6 -1
strain rates. ε(×10 S )

Figure 9. Stress–strain curve of concrete under dynamic uniaxial compressive stress at different
Figure
strain 9. Stress–strain curve of concrete under dynamic uniaxial compressive stress at different
rates.
10
strain rates.
10
stress peaks
8
strain peaks
stress peaks
factor

8
strain peaks
factor

10
6
increase

6
increase

stress peaks
8
4 strain peaks
Dynamic
Dynamic increase factor

4
Dynamic

6 2

2
4 0
0 20 40 60 80 100 120
0 -1
Strain rate (s )
2 0 20 40 60 80 100 120
-1
Strain rate (s )
Figure10.
Figure 10.The
The comparison
comparison of
of the
the peak
peak values
valuesof
ofstrain
strainand
andstress.
stress.
0
Figure 10.0The comparison
20 40 of the60peak values
-1
80 of100
strain 120
and stress.
Strain rate (s )

Figure 10. The comparison of the peak values of strain and stress.

(a) ε =0.001s
-1

(a) ε =0.001s
-1

(a) ε =0.001s
-1

(b) ε =1s
-1

Figure 11. Cont.

(b) ε =1s
-1

(b) ε =1s
-1
Materials 2019, 12, 643 11 of 15
Materials 2019, 11, x FOR PEER REVIEW 12 of 16

-1
(c) ε =10s
-1

(d) ε =100s-1
-1

Figure
Figure 11.11.The
Thedamage
damageprocess
processof ofconcrete
concrete specimen
specimen 11 under
under dynamic
dynamicuniaxial
uniaxialcompressive
compressivestress;
stress;
. -1 . -1 −1 . -1 . -1
ε=ε =0.001s
(a)(a) 0.001 s ; (b) εε=
−-1
1
=1s -1
; (c)ε =10s
1 s ; (c) ε = 10; s(d) ; ε(d)
-1 − 1  =100s -1
ε = 100 −
. s .1

-1 -1 -1 -1 -1
ε =0.001s -1 ε =0.1s -1 ε =1s -1 ε =10s -1 ε =100s -1
(a) Specimen 1

ε =0.001s-1-1 ε =0.1s-1-1 ε =1s-1-1 ε =10s-1-1 ε =100s-1-1

(b) Specimen 2

ε =0.001s -1 ε =0.1s -1 ε =1s-1 ε =10s-1 ε =100s -1

-1 -1 -1 -1 -1

(c) Specimen 3

Figure 12.The
Figure12. The failure patternofofconcrete
failure pattern concretespecimens
specimens under
under dynamic
dynamic uniaxial
uniaxial compressive
compressive stress;stress;
(a)
specimen 1; (b) specimen 2; (c) specimen 3.
(a) specimen 1; (b) specimen 2; (c) specimen 3.
 Materials 2019, 11, x FOR PEER REVIEW 13 of 16
Materials 2019, 12, 643 12 of 15

ε =0.001s -1 ε =0.1s -1 ε =1s -1 ε =10s -1 ε =100s -1

σ / MPa
Figure 13. The distribution of maximum principal stress at different strain rates of specimen 1.
Figure 13. The distribution of maximum principal stress at different strain rates of specimen 1.
The stress–strain curves are similar to the static at low strain rates from 0.01/s to 0.1/s and the
compressive strengths present a small enhancement. However, at the high strain rates from 1/s to
100/s, the curves slow down4gradually as the strain rate increases and the stress peak and strain peak
Abrams（1917）
increase significantly, and especially Malvern the ratio of dynamic compressive strength at the rate of 100/s to
＆Ross（1984,1985）
Dynamic compressive increase factor

the static compressive strength 3

is about
Takeda 2.67.
（1959It ）is also found that the increasing trend of the strain is
Evans（1942）
more prominent than that of the stress from Figure 10.
Hatano＆Tsutsumi（1960）
The meso-cracks first occur andBresler spread in the
＆Bertero （1975 ITZ,
） or the weak link, in cement mortar. Finally, a
Popp（1977）
region is formed, filled with 2a concentrative Yan（2006）
zonal crack, which results in the destruction of concrete.
At the high strain rates from
Materials 2019, 11, x FOR PEER REVIEW10/s to 100/s,
CEB the number of cracks greatly increases and the cracks are
13 of 16
diffused through the whole concretePrensent results
specimen—a few cracks even pass through the aggregate area.
Coalescent
ε =0.001scracks
-1 as a diffusion
ε =0.1s -1state can be seen ε =1s in-1 Figure 12. In εaddition,
 =10s -1 from Figures 9 and
ε =100s -1 12,
there is also a minor effect of the distribution of aggregates on the compressive stress. However, the
σ / MPa can
distribution of aggregates 1
significantly affect the initiation and propagation of cracks.
In Figure 14, it can be observed -8
10 10 10
-7
that
-6
10the10
-5
curve
-4
10 of
-3
10 the
-2
10present
-1
10 10 result
0 1 2
10 10 goes
3
up precipitously at the
-1
strain rate results obtained show Strain
good rate (s )
Figure 13. The distribution of maximum principal stress at different strain ratesEuro-International
of 0.1/s. The agreement with the Comité of specimen 1. du
Béton (CEB) standard and other experimental data at the strain rates from 10/s to 100/s. At the strain
rates from 0.01/s Figureto 10/s,
14.the
TheDIF
effectisoflower
strainthanrate onthe CEB standard.
compressive strength of plain concrete.

4
The stress–strain curves are similar
Abrams（ the）static at low strain rates from 0.01 / s to 0.1 / s and
to1917
the compressive strengths present aMalvern Ross（1984,1985）However, at the high strain rates from 1/ s
small＆enhancement.
Dynamic compressive increase factor

Takeda（1959）
to 100 / s , the curves slow 3down gradually Evans（1942）
as the strain rate increases and the stress peak and strain
peak increase significantly, and especially Hatano＆Tsutsumithe ratio
（1960 of）dynamic compressive strength at the rate of
Bresler
100 / s to the static compressive strength is about 2.67. ＆ Bertero （ 1975 ） It is also found that the increasing trend of
Popp（1977）
the strain is more prominent 2 than that Yan（ of2006
the ）stress from Figure 10.
The meso-cracks first occur and CEB spread in the ITZ, or the weak link, in cement mortar. Finally, a
region is formed, filled with a concentrative Prensent resultszonal crack, which results in the destruction of concrete.
At the high strain rates from 10 / s to 100 / s , the number of cracks greatly increases and the cracks
are diffused through the whole concrete specimen—a few cracks even pass through the aggregate
area. Coalescent cracks as a diffusion state can be seen in Figure 12. In addition, from Figure 9 and
1
Figure 12, there is also a minor -8 effect
-7 -6 of -5the -4distribution
-3 -2 of aggregates
-1 0 1 2 on3 the compressive stress.
10 10 10 10 10 10 10 10 10 10 10 10
However, the distribution of aggregates canStrain significantly
-1
rate (s ) affect the initiation and propagation of
cracks.
In Figure Figure
14, it can Theobserved
14. be effect of strain
that the ratecurve
on compressive
of the present strength
resultof plain
goes concrete.
up precipitously at the
strain rate of 0.1 Figure 14. The
/ s . The effect
results of strainshow
obtained rate on compressive
good agreement strength
with theof plain concrete.
Comité Euro-International
6. Discussion
du Béton (CEB) standard and other experimental data at the strain rates from 10 / s to 100 / s . At
The
theThe stress–strain
BFEM
strain ratesfor
from curves
0.01
dynamic are similar
/ sanalysis
to 10 / sof toDIF
concrete
, the theisstatic
loweratthan
introduced lowthe strain
in the rates
CEBpresent from
standard.study0.01 / s to an
verifies 0.1efficient
/ s and
the compressive strengths present a small enhancement. However,
numerical simulation method to investigate the damage mechanism of concrete. The strain-rate effect at the high strain rates from 1/ s
to6.100
and the /failure
s , the curves
Discussion slow down
characteristics gradually
are explored as the
using strain rate increases
two-dimensional models. andSeveral
the stress peak andcan
conclusions strain
be
peak increase
obtained as follows: significantly, and especially the ratio of dynamic compressive strength at the rate of
100 / s to the static compressive strength is about 2.67. It is also found that the increasing trend of
the strain is more prominent than that of the stress from Figure 10.
The meso-cracks first occur and spread in the ITZ, or the weak link, in cement mortar. Finally, a
region is formed, filled with a concentrative zonal crack, which results in the destruction of concrete.
At the high strain rates from 10 / s to 100 / s , the number of cracks greatly increases and the cracks
are diffused through the whole concrete specimen—a few cracks even pass through the aggregate
area. Coalescent cracks as a diffusion state can be seen in Figure 12. In addition, from Figure 9 and
Materials 2019, 12, 643 13 of 15

(1) The present results prove that the failure process of concrete under dynamic compression is
simulated well. With the increase of stress, the material enters the nonlinear stage and the stress–strain
curves show a nonlinear increase relationship, simultaneously.
(2) The variation tendency of the DIF of concrete at different strain rates is consistent with the
available experimental test data and the CEB empirical formula.
(3) At high strain rates, cracks increase and, in a diffusion state, some of the elements present
fracture damage and more energy is released, which could enhance the dynamic strength of
the concrete.
(4) The strain peaks of concrete present rate-sensitivity under different strain rates similar to
stress peaks.
For future work, a three-dimensional model of dynamic problems could be addressed by the
BFEM and other conditions could be considered; for instance, considering the shape of aggregates,
complicated boundary constraints, and so on, so as to simulate real concrete as much as possible.

Author Contributions: Y.P. made the revision and the final approval of the article. Q.W. carried out the numerical
experiments, interpretation of results and article. L.Y. provided the technical support of the computer software
program and the comprehensive review. M.M.A.K. made the check and revision of language and grammar of the
article. H.P. carried out the data collection and the data processing.
Funding: This research received no external funding.
Acknowledgments: This work is supported by National Science Foundation of China (10972015, 11172015), the
Beijing Natural Science Foundation (8162008), and the Pre-exploration Project of Key Laboratory of Urban Security
and Disaster Engineering, Ministry of Education, Beijing University of Technology (USDE201404).
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Bischoff, P.H.; Perry, S.H. Compressive behavior of concrete at high strain rates. Mater. Struct. 1991, 24,
425–450. [CrossRef]
2. Evans, R.H. Effect of rate of loading on the mechanical properties of some materials. J. Ice 1942, 18, 296–306.
[CrossRef]
3. Abrams, D.A. Effect of rate of application of load on the compressive strength of concrete. ASTM J. 1917, 17,
70–78.
4. Bischoff, P.H.; Perry, S.H. Impact behavior of plain concrete loaded in uniaxial compression. J. Eng. Mech.
1995, 25, 685–693. [CrossRef]
5. Ross, C.A.; Tedesco, J.W.; Kuennen, S.T. Effects of strain rate on concrete strength. Aci Mater. J. 1995, 92,
37–47.
6. John, R.; Antoun, T.; Rajendran, A.M. Effect of strain rate and size on tensile strength of concrete.
Shock Compression Condensed Matter 1992, 1994, 501–504.
7. Malvar, L.J.; Crawford, J.E. Dynamic Increase Factors for Concrete. In Proceedings of the Twenty-Eighth
DDESB Seminar, Orlando, FL, USA, 18–20 August 1998; pp. 1–17.
8. Williams, M.S. Modeling of local impact effects on plain and reinforced concrete. Aci Struc. J. 1994, 91,
178–187.
9. Kishen, J.M.C. Fracture of Rock-Concrete Interfaces: Laboratory Tests and Applications. ACI Struct. J. 2004,
101, 325–331.
10. Erzar, B.; Forquin, P. Experiments and mesoscopic modelling of dynamic testing of concrete. Mech. Mater
2011, 43, 505–527. [CrossRef]
11. Grote, D.L.; Park, S.W.; Zhou, M. Experimental characterization of the dynamic failure behavior of mortar
under impact loading. J. Appl. Phys. 2001, 89, 2115–2123. [CrossRef]
12. Grote, D.L.; Park, S.W.; Zhou, M. Dynamic behavior of concrete at high strain rates and pressures: I.
experimental characterization. Int. J. Impact Eng. 2001, 25, 869–886. [CrossRef]
13. Jin, X.C.; Hou, C.; Fan, X. Quasi-static and dynamic experimental studies on the tensile strength and failure
pattern of concrete and mortar discs. Sci Rep. 2017, 7, 15305. [CrossRef] [PubMed]
Materials 2019, 12, 643 14 of 15

14. Guo, Y.B.; Gao, G.F.; Jing, L. Response of high-strength concrete to dynamic compressive loading. Int. J.
Impact Eng. 2017, 108, 114–135. [CrossRef]
15. Levi-Hevroni, D.; Kochavi, E.; Kofman, B. Experimental and numerical investigation on the dynamic increase
factor of tensile strength in concrete. Int. J. Impact Eng. 2018, 114, 93–104. [CrossRef]
16. Zhou, J.; Chen, X.; Wu, L. Influence of free water content on the compressive mechanical behavior of cement
mortar under high strain rate. Sadhana 2011, 36, 357–369. [CrossRef]
17. Quast, M.; Curbach, M. Concrete under biaxial dynamic compressive loading. Procedia Eng. 2017, 210, 24–31.
[CrossRef]
18. Li, M.; Hao, H.; Shi, Y. Specimen shape and size effects on the concrete compressive strength under static
and dynamic tests. Constr. Build. Mater. 2018, 161, 84–93. [CrossRef]
19. Jurowski, K.; Grzeszczyk, S. Influence of Selected Factors on the Relationship between the Dynamic Elastic
Modulus and Compressive Strength of Concrete. Materials 2018, 11, 477. [CrossRef]
20. Zhang, S.; Lu, Y.; Jiang, X. Inertial effect on concrete-like materials under dynamic direct tension. Int. J.
Prot. Struct. 2018, 9, 377–396. [CrossRef]
21. Simo, J.C.; Ju, J.W. Strain- and stress-based continuum damage model—II. Comput aspects. Int. J. Solids Struct.
1989, 23, 841–869. [CrossRef]
22. Simo, J.C.; Ju, J.W. Strain- and stress-based continuum damage models—I. Formulation. Int. J. Solids Struct.
1987, 23, 821–840. [CrossRef]
23. Jin, L.; Zhang, R.; Dou, G.; Xu, J.; Du, X. Experimental and numerical study of reinforced concrete beams
with steel fibers subjected to impact loading. Int. J. Damage Mech. 2018, 27, 1058–1083. [CrossRef]
24. Jin, L.; Ding, Z.; Li, D.; Du, X. Experimental and numerical investigations on the size effect of moderate
high-strength reinforced concrete columns under small-eccentric compression. Int. J. Damage Mech. 2018, 27,
657–685. [CrossRef]
25. Li, W.S.; Wu, J.Y. A consistent and efficient localized damage model for concrete. Int. J. Damage Mech. 2018,
27, 541–567. [CrossRef]
26. Wardeh, M.A.; Toutanji, H.A. Parameter estimation of an anisotropic damage model for concrete using
genetic algorithms. Int. J. Damage Mech. 2017, 26, 801–825. [CrossRef]
27. Yoo, D.Y.; Kang, S.T.; Banthia, N.; Yoon, Y.S. Nonlinear finite element analysis of ultra-high-performance
fiber-reinforced concrete beams. Int. J. Damage Mech. 2017, 26, 735–757. [CrossRef]
28. Liang, J.; Ren, X.; Li, J. A competitive mechanism driven damage-plasticity model for fatigue behavior of
concrete. Int. J. Damage Mech. 2016, 25, 377–399. [CrossRef]
29. Georgin, J.F.; Reynouard, J.M. Modeling of structures subjected to impact: concrete behavior under high
strain rate. Cem. Concr. Compos. 2003, 25, 131–143. [CrossRef]
30. Snozzi, L.; Caballero, A.; Molinari, J.F. Influence of the meso-structure in dynamic fracture simulation of
concrete under tensile loading. Cem. Concr. Res. 2011, 41, 1130–1142. [CrossRef]
31. Snozzi, L.; Gatuingt, F.; Molinari, J.F. A meso-mechanical model for concrete under dynamic tensile and
compressive loading. Int. J. Fract. 2012, 178, 179–194. [CrossRef]
32. Gatuingt, F.; Snozzi, L.; Molinari, J.F. Numerical determination of the tensile response and the dissipated
fracture energy of concrete: role of the mesostructure and influence of the loading rate. Int. J. Numer. Anal.
Methods Geomech. 2013, 37, 3112–3130. [CrossRef]
33. Gatuingt, F.; Snozzi, L.; Molinari, J.F. Determination of the Dynamic Tensile Response and Dissipated
Fracture Energy of Concrete with a Cohesive Element Model. International Conference on Numerical
Modeling Strategies for Sustainable Concrete Structures—SSCS 2012, June 2012, Aix en Provence, France.
2017. Available online: https://hal.archives-ouvertes.fr/hal-01624529 (accessed on 15 January 2019).
34. Park, S.W.; Xia, Q.; Zhou, M. Dynamic behavior of concrete at high strain rates and pressures: II. numerical
simulation. Int. J. Impact Eng. 2001, 25, 887–910. [CrossRef]
35. Zhou, X.Q.; Hao, H. Mesoscale modelling of concrete tensile failure mechanism at high strain rates.
Comput. Struct. 2008, 86, 2013–2026. [CrossRef]
36. Zhou, X.Q.; Hao, H. Modelling of compressive behavior of concrete-like materials at high strain rate. Int. J.
Solids Struct. 2008, 45, 4648–4661. [CrossRef]
37. Hao, H.; Hao, Y.; Li, Z.X. A numerical study of factors influencing high-speed impact tests of concrete
material properties. In Proceedings of the 8th International Conference on Shock & Impact Loads on
Structures, Adelaide, Australia, 2–4 December 2009; pp. 37–52.
Materials 2019, 12, 643 15 of 15

38. Hao, H.; Zhou, X.Q. Concrete material model for high rate dynamic analysis. In Proceedings of the
Seventh International Conference on shock & impact loads on structures, Beijing, China, 17–19 October 2007;
pp. 753–768.
39. Cusatis, G. Strain-rate effects on concrete behavior. Int. J. Impact Eng. 2011, 38, 162–170. [CrossRef]
40. Wu, H.; Zhang, Q.; Huang, F. Experimental and numerical investigation on the dynamic tensile strength of
concrete. Int. J. Impact Eng. 2005, 32, 605–617. [CrossRef]
41. Zhou, W.; Tang, L.; Liu, X. Mesoscopic simulation of the dynamic tensile behavior of concrete based on a
rate-dependent cohesive model. Int. J. Impact Eng. 2016, 95, 165–175. [CrossRef]
42. Chen, X.; Wu, S.; Zhou, J. Experimental and modeling study of dynamic mechanical properties of cement
paste, mortar and concrete. Constr. Build. Mater. 2013, 47, 419–430. [CrossRef]
43. Peng, Y.J.; Liu, Y.H.; Pu, J.W.; Zhang, L. Application of base force element method to mesomechanics analysis
for recycled aggregate concrete. Math Probl. Eng. 2013, 2013, 1–8. [CrossRef]
44. Peng, Y.J.; Pu, J.W. Micromechanical investigation on size effect of tensile strength for recycled aggregate
concrete using BFEM. Int. J. Mech. Mater. Des. 2016, 12, 525–538. [CrossRef]
45. Peng, Y.J.; Chu, H.; Pu, J.W. Numerical simulation of recycled concrete using convex aggregate model and
base force element method. Adv. Mater. Sci. Eng. 2016, 2016, 1–10. [CrossRef]
46. Peng, Y.J.; Dong, Z.L.; Peng, B.; Liu, Y.H. Base force element method (BFEM) on potential energy principle
for elasticity problems. Int. J. Mech. Mater. Des. 2011, 7, 245–251. [CrossRef]
47. Walraven, J.C.; Reinhardt, H.W. Theory and experiments on the mechanical behavior of cracks in plain and
reinforced concrete subjected to shear loading. Heron. 1991, 26, 26–35.
48. Du, X.L.; Jin, L.; Ma, G. A meso-scale analysis method for the simulation of nonlinear damage and failure
behavior of reinforced concrete members. Int. J. Damage Mech. 2013, 22, 878–904. [CrossRef]
49. Jayasuriya, A.; Adams, M.P.; Bandelt, M.J. Understanding variability in recycled aggregate concrete
mechanical properties through numerical simulation and statistical evaluation. Constr. Build. Mater. 2018,
178, 301–312. [CrossRef]
50. Nilsen, A.U.; Monteiro, P.J.M. Concrete: A three phase material. Cem. Concr. Res. 1993, 23, 147–151. [CrossRef]
51. Agioutantis, Z.G.; Stiakakis, C.; Kleftakis, S. Numerical simulation of the mechanical behavior of epoxy
based mortars under compressive loads. Comput. Struct. 2002, 80, 2071–2084. [CrossRef]
52. Fakhari Tehrani, F.; Absi, J.; Allou, F. Heterogeneous numerical modeling of asphalt concrete through use of
a biphasic approach: Porous matrix/inclusions. Comput. Mater. Sci. 2013, 69, 186–196. [CrossRef]
53. Du, X.L.; Jin, L.; Ma, G. Numerical simulation of dynamic tensile-failure of concrete at meso-scale. Int. J.
Impact Eng. 2014, 66, 5–17. [CrossRef]
54. Comité Euro-International du Béton. CEB-FIP Model Code 1990; Redwood Books: Trowbridge, Wiltshire,
UK, 1993.
55. Dilger, W.H.; Koch, R.; Kowalczyk, R. Ductility of plain and confined concrete under different strain rates.
ACI J. 1984, 81, 73–81.

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).