Theoretical and Applied Fracture Mechanics 108 (2020) 102610

Contents lists available at ScienceDirect

Theoretical and Applied Fracture Mechanics

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

Mechanical properties and crack evolution of double-layer composite rock- T

like specimens with two parallel fissures under uniaxial compression
⁎
Jianhua Hu, Guanping Wen, Qibin Lin , Ping Cao, Su Li
School of Resources and Safety Engineering, Central South University, Changsha, China

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

Keywords: A large number of two-layer rock-like specimens composed of different combinations of materials and containing
Rock-like material two fissures with 30°, 45° or 60° fissure angles were manufactured and carried out a series of uniaxial com-
Two parallel fissures pression tests. A special parameter R related to the upper rock layer strength and lower rock layer strength was
Strength defined, and it was regarded to characterize the behavior of the different double-layer composite rock-like
Deformation
specimens. The axial strength of the double-layer composite rock-like specimens containing two parallel fissures
Crack evolution
Failure mode
was greatly weakened by the presence of the fissures and the degree of weakening was related to the fissure
angle and R, while the elastic and deformation modulus (Es and E50) were mainly affected by R. The axial strain
ε1c fluctuates between 9‰ and 12‰ and gradually stabilizes and was not directly related to the fissure angle or
R. The digital image correlation method was used to analyze crack evolution behavior and it showed that the
rock failure characteristics of the double-layer composite rock-like specimens were obviously different from
those of ordinary single-layer rock specimens. The ultimate failure modes could be divided into two major
categories of coalescence mode and no-coalescence mode, and each mode could be subdivided into three types,
which were mainly influenced by R and secondly influenced by the fissure angle. The results also indicate that
the interface between the upper rock layer and the lower rock layer can play a role in linking cracks.

1. Introduction Great deal of experimental research [11–15] or numerical simula-

tion [16–20] has been conducted on mechanical properties and crack
Layered composite rock is one of the most common transversely evolution behavior. Wong and Einstein [21,22] performed uniaxial
isotropic rock masses encountered in geotechnical engineering and compression experiments on gypsum and Carrara marble samples with
there have been many studies done on it using theoretical analysis, two open fissures and compared the combined effects of multiple geo-
laboratory tests, and numerical simulations [1–5]. Tien et al. [6] carried metrical factors on the cracking process and coalescence pattern of two
out uniaxial compression experiments on simulated transversely iso- different rocks. Wong and Chau [23] studied the influence of fissure
tropic rocks, and analysed and classified the failure mechanism on the angle, bridge angle, and frictional coefficient on crack coalescence
basis of a circumferential surface image of the cylindrical specimen patterns and the strength of rock-like specimens with two parallel fis-
recorded by a rotary scanner. Park and Min [7] adopted bonded-par- sures and identified three typical crack coalescence modes. Park and
ticle DEM models with embedded smooth joints to simulate transver- Bobet [24,25] compared the fracturing process in open and closed flaws
sely isotropic rock and realized that the weak planes with different and concluded that friction along the flaws increases the initiation and
inclination angles play a significant role in strength, deformation and coalescence stress and favors crack linkage through shear cracks. Zhao
failure patterns. Yang et al. [8] performed triaxial compression tests on et al. [26] measured local tensile and shear strain around fissures and
transversely isotropic composite rock-like specimens that consist of two showed that the wing and anti-wing crack propagation was initiated
different rock-like material, and analyze the mechanical properties and with a compressive strain jump and then further propagated gradually
the ultimate failure characteristics. Nevertheless, these studies are only with few jumps; in addition, a shear crack was abruptly initiated and
focused on the intact transversely isotropic rock, without considering then further unstable propagated. Yang et al. [27,28] carried out uni-
the internal defects such as pores, micro-cracks and fissures that sig- axial compression experiment on sandstone samples with two oval
nificantly affect mechanical properties and failure behavior of rock flaws and analyzed mechanical properties and failure behavior of
masses [9,10]. samples with different flaw angle α and ligament angle β through

⁎
Corresponding author.
E-mail address: [email protected] (Q. Lin).

https://doi.org/10.1016/j.tafmec.2020.102610
Received 14 January 2020; Received in revised form 8 March 2020; Accepted 13 April 2020
Available online 18 April 2020
0167-8442/ © 2020 Elsevier Ltd. All rights reserved.
J. Hu, et al. Theoretical and Applied Fracture Mechanics 108 (2020) 102610

photographic and AE monitoring technology. Zhao et al. [29] observed black or white cement, sand, and water as follows: 8:8:5 (Black), 2:1:1
the crack evolution of rock-like specimens with fissures within the (Black), 1:2:1 (White), 8:8:5 (White) and 2:1:1 (White). The mold was
whole loading process. Asadizadeh et al. [30] adopted 3D printing to divided into two almost-equal parts by a division plate to enable the
manufacture rough joints inside artificial specimens and studied the joining together of the two different mixtures that were added in the
influence of multiple geometrical factors including joint roughness two parts. After 30 min of the setting of the concrete specimens, the
coefficient on mechanical properties and crack evolution behavior of division plate was pulled out.
specimens. These studies on the influence of two fissures on mechanical The parallel close fissures were created by inserting mica sheets
properties and crack evolution of single rock/rock-like materials pro- with a length and thickness of 15 mm and 0.4 mm, respectively. To
vides a good baseline for studying the mechanical properties and crack ensure the accuracy of the position and inclination of the fissures,
evolution of layered composite rock with two parallel fissures. covering the specimen-making mold with a plastic board, which had a
In recent years, the digital image correlation method (DIC) based on slot cut into them for inserting the mica sheet. After 24 h of the initial
the mesoscopic perspective has been widely used in the study of the setting of the concrete, the molds were released and the specimens were
crack propagation process in rocks. Compared with traditional marked. Each specimen was soaked in clear water for 7 days and then
methods, it has the following advantages: environmental adaptability, maintained in a maintenance box for 21 days. To obtain the speckles
greater experimental scope and improved experimental accuracy [31]. that meet the requirements of DIC calculations, first sprayed the surface
The DIC method is an effective way to study the crack initiation and of the cured sample uniformly with white paint. After the surface of the
propagation process. Pan et al. [32] used the DIC technology to analyze sample was air-dried, black paint was randomly sprayed cover upon
the displacement field and failure process of the jointed specimen under them. The experimental results clearly show the axial stress dropped
different filling conditions. Aliabadian et al. [33,34] used the DIC rapidly to near zero shortly after peak stress and a huge failure sound
method to determine the crack initiation location of transversely iso- was heard during the stress drop, which is in good agreement with the
tropic rocks in the Brazilian split test. Sharafisafa et al. [35,36] used the characteristics of brittle rocks stress-strain curve. According to ISRM
DIC technology to observe and analyze the deformation and failure recommendations [43], cylinder samples of Φ50 mm × 100 mm and
process of 3D printed rock-like specimens. Lin et al. [37] used the DIC Φ50 mm × 25 mm were made, and uniaxial compression test and
technology to explore the crack coalescence process of jointed rock-like Brazilian split test were performed respectively to determine the basic
materials and obtained the failure modes. The above research results physical properties of the rock-like materials. Uniaxial compressive
show that the DIC technology can be applied to analyze the fracture strength and elastic modulus were obtained through the uniaxial
behavior of rocks or rock-like materials. compression test. The Brazilian split test was conducted to obtain the
There has been plenty of work done respectively in the anisotropic tensile strength of the material. The mechanical parameters of different
behavior of transversely isotropic rock and the influence of fissures on rock-like material are illustrated in Table 1, and σc represents the
mechanical parameters and failure mode of intact rock [38–42]. uniaxial compressive strength and εc represents the peak axial strain
However, the relationship between the crack evolution behavior of value. ES represents the elastic modulus and E50 represents the de-
composite rock under the influence of various factors, including fissure formation modulus.
angle and material type, is still limited, and further research is needed.
On one hand, studies have shown that rock or rock-like materials have
an influence on crack evolution and failure modes, such as in marble 2.2. Experiment equipment and procedure
and gypsum, the conditions for tensile crack and shear crack are dif-
ferent and hence leading to different coalescence modes [21]. On an- The testing system consists of a loading system and a DIC system, as
other hand, the crack initiation load was determined by the plastic zone shown in Fig. 2. The loading system comprised a new SANS electro-
size of the crack initiation direction reaching to the critical plastic zone hydraulic servo-controlled rigidity testing machine and a loading con-
size, which was related to rock material properties [15]. Therefore, the trol system DCS-200. The displacement control mode was adopted as
mechanical properties of double-layer composite rock-like specimens the load method in this uniaxial compression test and the displacement
made from a different combination of materials and containing two loading rate was set at 0.2 mm/min. The DIC system consists of a CCD
parallel fissures are firstly studied in this paper. Then, the crack de- camera (Basler / piA2400-17 gm) with a resolution of 2456 × 2058
veloping process under uniaxial loading are recorded by a CCD camera. pixels, two white light sources and a computer for image collection.
Meanwhile, the DIC method is used to study the initiation, propagation During loading, the image collection frequency is set to 15 frames / s. In
and coalescence of a specimen’s surface cracks. Based on the analysis addition, applying a certain amount of butter on the top and bottom of
results of the DIC method, the crack evolution process and the ultimate the specimen for reducing the friction between the specimens and the
failure modes are analysed, classified, summarized and reviewed. machine.
The DIC method has been used increasingly more in the research of
2. Specimen manufacture and experiment procedure crack evolution in brittle materials, especially in the study of rock
fractures. The basic principle of the DIC method is to obtain the strain
2.1. Specimen manufacture and displacement of the sample by comparing the image of the sample
surface before and after deformation. In this method, the dense and
Molded concrete specimens were manufactured from different vo- random speckles on a specimen’s surface can be used to obtain de-
lume ratios of cement, sand, and water. In order to facilitate easier formation information. In this method, a very thin layer of white lac-
observation, two different types of cement, ordinary Portland cement quer is sprayed on the observed surface of the specimen and then the
(black cement) and white Portland cement (white cement), were used surface is sprayed with black lacquer to obtain a large number of dense
because their colors are easy to distinguish. Referring to the previous and random black dots above the white lacquer. After the DIC experi-
research, the main consideration of specimen dimensions was to ments were completed, the images were analyzed by the GOM correlate
minimize the boundary effect as much as possible, and the second was software. Crack initiation, extension, and coalescence, as well as failure
the thickness should not be too large or it would easily lead to crack mode, crack type, and rock strength were analyzed by combining the
rotation in the specimen. Thus, the dimensions of the specimens were strain field, stress-strain curve, and other experimental results.
fixed to 140 × 70 × 30 mm3 as shown in Fig. 1.
To generate two different rock characteristics in a specimen, a
special mold for making specimens was designed, as described in Fig. 1.
Specimens were created using five different volume mixture ratios of

2
J. Hu, et al. Theoretical and Applied Fracture Mechanics 108 (2020) 102610

a. Specimen-making mold b. Specimen geometry c. Specimen object

Fig. 1. Mold and specimen with two parallel fissures, in which α is fissure angle, β is bridge angle, 2a is fissure length, and 2b is ligament length.

3. Mechanical properties analysis the cracks mainly occur in the weak rock layers.
The strength of the intact double-layer specimens is related to the
3.1. Mechanical properties of intact double-layer composite rock-like material properties of the upper and lower rock layers according to
specimens made from different combinations of materials some rules. As shown in Fig. 4, when the strength difference between
the two rock materials is about two times greater than that of the
To explore how different material combinations of rock-like mate- weaker rock layer, the strength of the two-rock layer is slightly smaller
rials affect the mechanical properties of double-layer composite rock- than that of the weaker rock layer alone, which indicates that when the
like samples and to use these results as a control group for samples with strength difference is large, the strength of the stronger rock layer does
fissures, the mechanical properties of intact double-layer composite not enhance the strength of the entire two-rock layer, but weakens it
rock-like samples were obtained, as shown in Table 2. The σc, εc, ES, and slightly (S-I-B1-W1, S-I-B2-W1); when the strength difference between
E50 are as defined as above; SR(U) is the strength ratio of the double- the two rock layers and the strength ratio of the weaker rock layer is
layer composite rock-like specimen to the intact specimen made by the close to zero, the strength of the entire rock layer at this time becomes
upper rock-like material and SR(L) is the strength ratio of the double- slightly weaker. When the strength difference between the two rock
layer composite rock-like specimen to the intact specimen made by the layers and the strength ratio of the weaker rock layer is between 0.2 and
lower rock-like materials. R is defined as shown in Eq(1), and can 2.0, it shows a certain linear relationship and decreases linearly with an
characterize the mechanical behavior of the double-layer composite increase in the R value (S-I-B1-W2, S-I-B2-W3, S-I-B2-W2, S-I-B1-W1).
rock-like specimens made from different combinations of materials The relationship between the peak axial strain and R (ε1c-R), the elastic
based on the relationship between the upper and lower rock layer modulus and R (Es-R), and the deformation modulus and R (E50-R) is
strength. almost fully consistent with the relationship between the uniaxial
compressive strength and R (σc-R).
SR (U ) − SR (L)
R=
SR (U ) (1)
3.2. Mechanical properties of double-layer composite rock-like specimens
As shown in Fig. 3, three typical failure modes were identified based made from a different combination of materials and containing two parallel
on their damage position by analyzing the ultimate failure modes. It is fissures
mainly used as a control group and reflects the ultimate failure modes
of intact double-layer composite rock-like specimens made from a dif- In this section, the influence of the geometry of the two parallel
ferent combination of materials. The experimental results show that the fissures on the mechanical properties of the double-layer composite
law of crack development of crack is not obvious, mainly irregular local rock-like specimens made from different combinations of materials is
cracks. However, from the distribution of cracks produced during the presented. The details are listed in Table 3.
failure, it still reflects the characteristics of the composite rock. Gen- In order to better demonstrate the influence of the upper rock-like
erally, when the properties of the two combinations of rock-like ma- materials and fissure angle on the mechanical properties of the speci-
terials are near, the cracks penetrate the upper and lower rock layers, or mens, three groups of experiments were separately conducted for

Table 1
Parameters of the specimens containing different volumes of cement, water, and sand subjected to uniaxial compression.
Specimens Type of cement Volume mixture ratios σc /MPa εc/10−3 σt /MPa Es /GPa E50 /GPa

B1 black cement 8:8:5 41.771 14.806 5.671 4.576 2.533

B2 2:1:1 65.206 16.607 7.600 5.741 3.730
W1 white cement 1:2:1 13.502 11.026 3.695 1.527 1.418
W2 8:8:5 35.295 14.478 7.171 4.330 2.403
W3 2:1:1 48.946 15.752 7.176 5.632 3.114

3
J. Hu, et al. Theoretical and Applied Fracture Mechanics 108 (2020) 102610

Fig. 2. Schematic diagram of the loading system and the DIC system.

Table 2
Mechanical parameters obtained from uniaxial compression of the double-layer
composite rock-like specimens made from different combinations of materials.
Specimens σc/Mpa εc Es/Gpa E50/Gpa SR(U) SR(L) R

S-I-B1-W1 12.333 8.971 2.035 1.191 0.913 0.295 2.094

S-I-B1-W2 36.501 14.343 4.233 2.227 1.034 0.874 0.183
S-I-B1-W3 50.844 15.662 5.867 2.973 1.039 1.217 −0.147
S-I-B2-W1 12.419 10.972 1.238 1.142 0.920 0.190 3.829
S-I-B2-W2 25.365 11.588 3.943 2.161 0.719 0.389 0.847
S-I-B2-W3 33.492 13.361 4.556 2.143 0.684 0.513 0.332

comparison. The complete stress-strain process during uniaxial com-

pression of the specimens generally includes four typical stages: micro-
crack compaction, elastic deformation, micro-crack branching and
propagation, and post-failure. Fig. 5a shows the comparison of the
stress-strain curves when the upper rock-like material is changed and
when the other factors are kept unchanged (keeping α = 45° and the
lower material of the double-layer composite rock-like specimens was Fig. 4. Effect of R on the mechanical properties of the intact double-layer
material B1). From Fig. 5a, we can see that there is a great difference composite rock-like specimens under uniaxial compression.
among the stress-strain curves of the double-layer composite rock-like
specimens made with different combinations of materials except during
the micro-crack compaction stage. The stress-strain curves of the

Fig. 3. Three typical failure modes of intact double-layer composite rock-like specimens under uniaxial compression.

4
J. Hu, et al. Theoretical and Applied Fracture Mechanics 108 (2020) 102610

Table 3
Mechanical parameters obtained from uniaxial compression of the double-layer composite rock-like specimens made from different combinations of materials.
Sample α/° β/° 2a/mm 2b/mm R σc/MPa εc/10−3 Es/GPa E50/GPa

S-F-B1-W1-30 45 30 15 30 2.094 12.327 10.204 1.507 1.197

S-F-B1-W1-45 45 45 15 30 2.094 11.743 10.436 1.60 1.263
S-F-B1-W1-60 45 60 15 30 2.094 11.933 10.226 1.432 1.171
S-F-B1-W2-30 45 30 15 30 0.183 26.431 9.699 4.316 2.245
S-F-B1-W2-45 45 45 15 30 0.183 26.962 9.974 4.242 2.188
S-F-B1-W2-60 45 60 15 30 0.183 31.460 10.365 4.506 2.521
S-F-B1-W3-30 45 30 15 30 −0.147 31.679 10.097 5.044 2.408
S-F-B1-W3-45 45 45 15 30 −0.147 32.286 9.664 5.079 2.508
S-F-B1-W3-60 45 60 15 30 −0.147 40.098 11.458 5.677 3.030
S-F-B2-W1-30 45 30 15 30 3.829 13.365 9.139 2.152 1.285
S-F-B2-W1-45 45 45 15 30 3.829 12.569 9.369 1.814 1.335
S-F-B2-W1-60 45 60 15 30 3.829 11.714 9.051 1.773 1.253
S-F-B2-W2-30 45 30 15 30 0.847 21.435 9.785 3.195 1.921
S-F-B2-W2-45 45 45 15 30 0.847 21.911 9.172 3.903 1.996
S-F-B2-W2-60 45 60 15 30 0.847 25.171 10.612 3.923 2.183
S-F-B2-W3-30 45 30 15 30 0.332 24.822 10.098 3.776 1.962
S-F-B2-W3-45 45 45 15 30 0.332 25.120 10.881 3.786 1.967
S-F-B2-W3-60 45 60 15 30 0.332 29.850 11.767 4.236 2.337

Fig. 5. Axial stress-axial strain curves of double-layer composite rock-like specimens with two parallel fissures, in which Fig (a) and (b), respectively, show the effect
of upper rock materials and fissure angle on the axial stress-axial strain curves.

double-layer composite specimen with the lowest strength of the upper tips of the fissures and cause the specimens to fail faster, which shortens
rock material (W1) rises most slowly and fails first; on the contrary, the the elastic deformation stage and quickens the micro-crack branching
stress-strain curves with the highest strength of the upper rock material and propagation stage. Third, the post-failure stage is more complex
(W3) rises fastest and fails last. The stress-strain curves of the two and shows an irregular stepped descent instead of a rapid drop. There is
specimens with the lower rock-like materials of W2 and W3 are quite almost no difference among the stress-strain curves of the intact double-
different even if their absolute value of R is very close. The strength of layer composite rock-like specimens with two parallel fissures both at
the double-layer specimen can effectively be improved by slightly en- the micro-crack compaction and elastic deformation stages. During the
hancing the strength of the upper rock-like materials when the lower micro-crack branching and propagation stage, the initiation and pro-
rock-like materials remain unchanged. pagation of new micro-cracks close to the tips of fissures makes it easier
Fig. 5b compares the stress-strain curve when the fissure angles are for the specimens to reach the critical failure state, especially when the
changed and when the other factors remain unchanged (the double- fissure angle is small.
layer composite rock-like specimens made with the combinations of Fig. 6 shows the effect of R on the mechanical properties of the
material B1 and material W2). From Fig. 5b, we can see that the fissure double-layer composite rock-like specimens with two parallel fissures
angles have a remarkable influence on the stress-strain relation. Com- subjected to uniaxial compression. The figure shows that the presence
pared with the stress-strain curve of the intact double-layer composite of the fissures greatly weakens the strength of the specimens, especially
rock-like specimens, the stress-strain curves of the intact double-layer when R is small. Furthermore, the fissure angles also have a significant
composite rock-like specimens with two parallel fissures show some influence on the strength of the specimens. The angles of 30° and 45°
differences. First, the micro-crack compaction stage is shortened. weaken the strength of the specimens significantly and there is little
Second, due to the stress concentration close to the tips of the two difference between the two weakening effects. The strength of the
fissures, a considerable number of new micro-cracks develop from the specimens with fissure angles of 60° is close to that of the intact

5
J. Hu, et al. Theoretical and Applied Fracture Mechanics 108 (2020) 102610

a. Peak strength (Mpa) b. Peak strain (10-3)

c. Elastic modulus (Gpa) d. Deformation modulus (Gpa)

Fig. 6. Effect of R on the mechanical properties of the double-layer composite rock-like specimens with two parallel fissures under uniaxial compression.

specimens. Additionally, the elastic modulus Es and deformation consistent with the results of Wong and Einstein [21] who analyzed
modulus E50 are also only slightly affected by the presence of the fis- gypsum and Carrara marble specimens. Furthermore, the tensile wing
sures and their angle, and are almost exactly the same as the elastic cracks always initiate from the two fissures of the specimen, but the
modulus Es and deformation modulus E50 of the intact double-layer propagation of the cracks around the upper fissure and the lower fissure
specimens. The fissures and their angles have a greater influence on the is not synchronous in time, and is closely related to the relative strength
axial strain ε1c of the double-layer specimens. The axial strain ε1c of the of the upper and lower rock materials. Fig. 7 is the strain strains contour
double-layer composite rock-like specimens with two parallel fissures plot obtained by the DIC method when the residual strength after the
fluctuates between 9‰ and 12‰ and gradually flattens with the in- peak is 50%. It presents the ultimate failure modes of the double-layer
crease in the R value, while the axial strain ε1c of the intact double-layer composite rock-like specimens with parallel fissures subjected to uni-
specimens decreases and then flattens with an increase in the R value. axial compression. It shows that the failure of all the specimens was due
to the propagation of cracks from the fissures. The failure modes of the
4. The ultimate failure modes and crack evolution behavior double-layer composite rock-like specimens are different from those of
analysis ordinary single-layer rock specimens [21–29].

4.1. Failure modes 4.1.1. No-coalescence mode

In this mode, vertical or inclined cracks developed and propagated
Using the DIC method, crack evolution (initiation, propagation, and from the two fissures, but did not connect, and there were significant
coalescence) in the double-layer composite rock-like specimens made differences in the crack behavior in the double-layer composite rock-
from different combinations of materials and containing two parallel like specimens. This is mainly due to the difference in composition
fissures were identified and analyzed. The coalescence modes of fissures between the upper and lower layers where the fissures were located.
obtained in the experimental results were compared against the crack When the composition of the two layers were significantly different,
patterns of the fissures in the specimens during compression. The DIC cracks developed and propagated in the weaker layers and penetrated
results show that tensile wing cracks were always the earliest cracks to the upper and lower ends of the specimen, while the cracks around the
develop in the double-layer composite rock-like specimens, which is fissures in the stronger layer only just started to develop or did not

6
J. Hu, et al. Theoretical and Applied Fracture Mechanics 108 (2020) 102610

Fig. 7. The ultimate failure modes of the double-layer composite rock-like specimens made from different combinations of materials and containing two parallel
fissures.

develop at all. Therefore, in the no-coalescence mode in the double- Type 1: The T1 crack initiates from the inner tip of the upper fissure
layer composite rock-like specimens, one fissure begins to propagate and the T2 crack initiates from the outer tip of the upper fissure. In
while the other just begins to grow or does not begin to grow at all. addition, a shear crack propagates obliquely downward from the
Observation of the surface cracks in the failed specimens captured top-left corner of the specimen and connects with the inner tip of the
by the CCD camera shows that there is no obvious crack development upper fissure.
around the lower fissure of the specimens, while the cracks around the Type 2: The T1 crack initiates from the inner tip of the upper fissure
upper fissure of the specimens penetrate the top and bottom of the and the T2 crack initiates from the outer tip of the upper fissure,
specimens. The results of the DIC method show the initiation of wing which is the same as Mode 1. What is different is the shear crack,
cracks around the upper fissure. Furthermore, the secondary coplanar which propagates obliquely downward from the top-left corner of
crack propagated from the outer tip of the upper fissure to the top of the the specimen and connects with the S2 crack initiated from the inner
specimen, while the tensile wing crack propagated downward along the tip of the upper fissure. In particular, the point of connection is lo-
direction of the maximum stress from the inner tip of the upper fissure cated at the boundary between the upper and lower rock layer.
and the direction of propagation changed sharply at the interface be- Type 3: The T1 crack initiates from the inner tip of the upper fissure
tween the two rock layers; the anti-wing crack propagated from the and the shear crack propagates obliquely downward from the top-
inner tip of the upper fissure to the top of the specimen with a tortuous left corner of the specimen and connects with the inner tip of the
and random propagation path. upper fissure.
This mode can be classified into three types according to the strain
contour obtained by the DIC method and verified with the image taken
by the camera, which is shown in Fig. 8. The value of R and the fissure 4.1.2. Coalescence mode
inclination angle have a significant effect on the crack propagation According to Wong and Einstein’s study [21], the coalescence mode
types in the specimens. These three types are similar in crack propa- can be divided into nine categories. However, only a few coalescence
gation forms and their main difference lies in two minor details. One modes were observed in our study, which is mainly because the rock
detail is the location of the shear crack initiated from the top-left corner bridge angle was fixed at 45°, the ligament length was fixed at twice the
and the manner in which it connects with the other cracks. The other is length of the prefabricated joints, and there were only three fissure
the crack initiation location through the interface of the two-rock layer. inclination angles (30°, 45° and 60°). In addition, there were some
differences between the double-layer rock and single layer rock speci-
mens, which were mainly caused by the influence of the interface

7
J. Hu, et al. Theoretical and Applied Fracture Mechanics 108 (2020) 102610

Fig. 8. Three types of the no-coalescence mode in the double-layer composite rock-like specimens with two parallel fissures.

between the two rock layers. the upper and lower fissures.
There are cracks seen around the upper and lower fissures of the This mode can be simply divided into an indirect type and a direct
specimens as captured by the CCD camera and the DIC method. The coalescence type, and the indirect type also can be subdivided into two
secondary coplanar crack propagated from the outer tip of the upper types, as shown in Fig. 9. Though the value of R is small, the Type III
fissure to the top of the specimen or from the outer tip of the lower still differs from the modes of the single layer under the same condi-
fissure to the bottom of the specimen, which is consistent with the tions. This indicates that the interface between the upper rock layer and
failure modes of a single rock layer specimen. However, the crack lower rock layer can play a role in linking cracks.
propagation around the upper and lower fissures is complex and quite
different from that seen in previous studies [21–29] using only a single Type 1: The T1 tensile crack initiates from the inner tip of the upper
rock layer. The tensile wing crack propagates downward along the di- fissure and the mixed tensile-shear crack initiates from the inner and
rection of maximum stress from the inner tip of the upper fissure or outer tip of the upper fissure. In addition, a shear crack propagates
propagates upward along the direction of the maximum stress from the obliquely downward from the top-left corner of the specimen and
inner tip of the lower fissure, and depends on the rock strength between connects with the mixed tensile-shear crack initiated from the inner

Fig. 9. Three types of the coalescence mode of the double-layer composite rock-like specimens with two parallel fissures.

8
J. Hu, et al. Theoretical and Applied Fracture Mechanics 108 (2020) 102610

Fig. 10. Comparison of crack evolution process in the specimens made from different combinations of materials and containing two parallel fissures.

tip of the upper fissure. In particular, the connection point is located the specimens.
at the boundary between the upper and lower rock layers. Fig. 11 shows a comparison of the crack evolution process in the
Type 2: One form is the S1 crack that initiates from the inner tip of specimens with two parallel fissures at different fissure angles. It is
the upper fissure and links with the S3 crack initiated from the inner obvious that the weak interface between the upper rock layer and the
tip of the upper fissure. Another form is a crack lined the S1 crack lower rock layer has a significant influence on crack propagation with
initiated from the inner tip of the upper fissure with the S1 crack the increase in the fissure angle. The most obvious feature is extrusion
initiated from the inner tip of the lower fissure. and enlargement of the crushing zone between the inner tips of the
Type 3: One form is a crack around the boundary between the upper upper and lower pre-fissures. On the whole, the crack initiation and
and lower rock layer which links the S1 crack initiated from the propagation stages are almost the same and the main difference lies in
inner tip of the upper fissure with the S1 crack initiated from the the crack propagation between the inner tip of the upper fissure and
inner tip of the lower fissure. Another form is similar but the linked lower fissure. When the fissure angle (α) is 30°, the cracks initiated
crack is present above and below the boundary between the upper from the inner tip of the upper fissure and the lower fissure do not
and lower rock layers. coalesce and the main cause of specimen failure is that the cracks in-
itiated from the fissures run through both the upper and lower ends of
the specimen. When the fissure angle (α) is 45°, the cracks initiated
4.2. Crack evolution behavior from the inner tip of the upper fissure and lower fissure coalesce di-
rectly and combine with the cracks initiated from the outer tip of the
In order to show crack evolution behavior of specimens at different upper fissure and lower fissure to form a large shear surface. In addi-
loading stages, especially during the micro-crack compaction, elastic tion, anti-wing cracks initiated from the inner tip of the upper fissure
deformation, micro-crack branching and propagation, and post-failure and the lower fissure propagate obviously. When the fissure angle (α) is
stages, we present and describe some typical pictures obtained by the 60°, the cracks initiated from the inner tip of the upper fissure and
DIC system below. lower fissure also coalesce. The shear crack propagated from the inner
Fig. 10 shows a comparison of the crack evolution process of the fissures tip and along the direction of pre-fissures, but it didn't expand
specimens made from different combinations of materials and con- further at the interface between two different molding materials.
taining two parallel fissures. The crack propagation process of the During the further loading process, the shear cracks that propagated
double-layer composite rock-like specimens is completely different from the upper and lower fissures coalesced to each other along with
when there is a different upper rock material. First, the cracks leading the interface between two different molding materials. In the process of
to the failure of specimens are generally initiated from the fissure on the the crack coalescence with each other, it is necessary to overcome the
weaker rock layer and the weaker layer is relatively more broken after friction force, which is almost perpendicular to the direction of max-
the failure. Second, if the strength between the upper and lower rock imum stress. Thus the specimen can still resist large loads, and a big
layer of the specimens varies greatly, the cracks initiated from the pre- extrusion crushing zone around the interface between the upper rock
fissure on the weaker rock layer propagate into the strong layer while layer and the lower rock layer finally forms.
the cracks initiated from the pre-fissure on the stronger rock layer only
just began to form. Third, the interface between the upper rock layer
and lower rock layer can be regarded as a weak surface when there is
little difference in strength between the upper and lower rock layer of

9
J. Hu, et al. Theoretical and Applied Fracture Mechanics 108 (2020) 102610

Fig. 11. Comparison of crack evolution process in the specimens with two parallel fissures at different fissure angles.

5. Conclusions influence on crack propagation and coalescence when there is little

difference in strength between the upper and lower rock layers of
In this paper, the mechanical properties, crack evolution behavior the specimens and when the fissure angle is large.
and failure mode of different compositions rock-like material con-
taining two parallel fissures under uniaxial compression were system- CRediT authorship contribution statement
atically studied. Based on the results of the experimental and DIC
analysis, the following conclusions can be drawn. Jianhua Hu: Resources, Validation, Writing - review & editing,
Funding acquisition. Guanping Wen: Methodology, Investigation, Data
(1) To characterize the strength and deformation of double-layer curation, Writing - original draft. Qibin Lin: Conceptualization,
composite rock-like specimens, we define a special parameter R. It Methodology, Investigation, Writing - review & editing. Ping Cao:
is based on the mechanical parameters of the intact double-layer Resources, Supervision, Funding acquisition. Su Li: Investigation, Data
composite rock-like specimens made of different material combi- curation.
nations. The axial strength (σc), peak axial strain (ε1c), elastic
modulus (Es), and deformation modulus (E50) of the intact double- Declaration of Competing Interest
layer specimens gradually decreases with increasing R and even-
tually stabilizes. The authors declare that they have no known competing financial
(2) The presence of the parallel fissures significantly weakens the axial interests or personal relationships that could have appeared to influ-
strengths of the double-layer composite rock-like specimens, and ence the work reported in this paper.
the degree of weakening is related to R and the fissure angle. While
the elastic modulus Es and the deformation modulus E50 are mainly Acknowledgements
affected by R. The axial strain ε1c fluctuates between 9‰ and 12‰
and gradually stabilizes and is not directly related to R or the fissure This paper received its funding from Project (41672298, 11772358,
angle. 51174228) supported by the National Natural Science Foundation of
(3) The analysis results of the DIC method showed that the rock failure China, and the Fundamental Research Funds for the Central
characteristics of the double-layer composite rock-like specimens Universities, China (2019zzts980). The authors are very grateful for the
were different from those of ordinary single-layer rock specimens. anonymous reviewers' valuable comments.
The ultimate failure modes could be divided into two major cate-
gories of coalescence mode and no-coalescence mode, and each Appendix A. Supplementary material
mode could be subdivided into three types, which were mainly
influenced by R and secondly influenced by the fissure angle. The Supplementary data to this article can be found online at https://
results also indicate that the interface between the upper rock layer doi.org/10.1016/j.tafmec.2020.102610.
and the lower rock layer can play a role in linking cracks.
(4) Changing one rock layer of the double-layer composite rock-like References
specimens or the fissure angle has a significant effect on crack
evolution. The interface between the upper rock layer and the lower [1] R. Nova, The failure of transversely isotropic rocks in triaxial compression, Int. J.
rock layer can be regarded as a weak surface. It has a significant Rock Mech. Min. Sci. Geo. Abs. 17 (6) (1980) 325–332, https://doi.org/10.1016/

10
J. Hu, et al. Theoretical and Applied Fracture Mechanics 108 (2020) 102610

0148-9062(80)90515-X. rock, Int. J. Solids Struct. 50 (2013) 3685–3706, https://doi.org/10.1016/j.ijsolstr.

[2] L. Li, M. Aubertin, A crack-induced stress approach to describe the tensile strength 2013.07.010.
of transversely isotropic rocks, Can. Geotech. J. 39 (1) (2002) 1–13, https://doi. [23] R.H.C. Wong, K.T. Chau, Crack coalescence in a rock-like material containing two
org/10.1139/t01-069. cracks, Int. J. Rock Mech. Min. Sci. 35 (2) (1998) 147–164, https://doi.org/10.
[3] D.P. Xu, X.T. Feng, D.F. Chen, C.Q. Zhang, Q.X. Fan, Constitutive representation 1016/S0148-9062(97)00303-3.
and damage degree index for the layered rock mass excavation response in un- [24] C.H. Park, A. Bobet, Crack coalescence in specimens with open and closed flaws: a
derground openings, Tunn. Undergr. Sp. Tech. 64 (2017) 133–145, https://doi.org/ comparison, Int. J. Rock Mech. Min. Sci. 46 (5) (2009) 819–829, https://doi.org/
10.1016/j.tust.2017.01.016. 10.1016/j.ijrmms.2009.02.006.
[4] J.L. Cheng, S.Q. Yang, K. Chen, D. Ma, F.Y. Li, L.M. Wang, Uniaxial experimental [25] C.H. Park, A. Bobet, Crack initiation, propagation and coalescence from frictional
study of the acoustic emission and deformation behavior of composite rock based flaws in uniaxial compression, Eng. Fracture Mech. 77 (14) (2010) 2727–2748,
on 3d digital image correlation (DIC), Acta Mech. Sinica. 33 (6) (2017) 999–1021, https://doi.org/10.1016/j.engfracmech.2010.06.027.
https://doi.org/10.1007/s10409-017-0706-3. [26] Y.L. Zhao, L.Y. Zhang, W.J. Wang, C.Z. Pu, W. Wan, J.Z. Tang, Cracking and stress-
[5] L. Selçuk, D. Aşma, Experimental investigation of the Rock-Concrete bi-materials strain behavior of rock-like material containing two flaws under uniaxial com-
influence of inclined interface on strength and failure behavior, Int. J. Rock Mech. pression, Rock Mech Rock Eng 49 (7) (2016) 2665–2687, https://doi.org/10.1007/
Min. Sci. 123 (2019) 104–119, https://doi.org/10.1016/j.ijrmms.2019.104119. s00603-016-0932-1.
[6] Y.M. Tien, M.C. Kuo, C.H. Juang, An experimental investigation of the failure [27] S.Q. Yang, Crack coalescence behavior of brittle sandstone samples containing two
mechanism of simulated transversely isotropic rocks, Int. J. Rock Mech. Min. Sci. 43 coplanar fissures in the process of deformation failure, Eng. Fract. Mech. 78 (17)
(8) (2006) 1163–1181, https://doi.org/10.1016/j.ijrmms.2006.03.011. (2011) 3059–3081, https://doi.org/10.1016/j.engfracmech.2011.09.002.
[7] B. Park, K.B. Min, Bonded-particle discrete element modeling of mechanical be- [28] S.Q. Yang, Y.H. Huang, W.L. Tian, J.B. Zhu, An experimental investigation on
havior of transversely isotropic rock, Int. J. Rock Mech. Min. Sci. 76 (2015) strength, deformation and crack evolution behavior of sandstone containing two
243–255, https://doi.org/10.1016/j.ijrmms.2015.03.014. oval flaws under uniaxial compression, Eng. Geol. 217 (2017) 35–48, https://doi.
[8] S.Q. Yang, P.F. Yin, Y.H. Huang, J.L. Cheng, Strength, deformability and x-ray org/10.1016/j.enggeo.2016.12.004.
micro-ct observations of transversely isotropic composite rock under different [29] C. Zhao, Z. Meng, Z. Feng, B. Chong, Cracking processes and coalescence modes in
confining pressures, Eng. Fracture Mech. 214 (2019) 1–20, https://doi.org/10. rock-like specimens with two parallel cracks, Rock Mech. Rock Eng. 51 (2018)
1016/j.engfracmech.2019.04.030. 3377–3393, https://doi.org/10.1007/s00603-018-1525-y.
[9] Y.L. Gui, J.L. Shang, J.J. Ma, Z.Y. Zhao, Numerical study of the circular opening [30] M. Asadizadeh, M.F. Hossaini, M. Moosavi, H. Masoumi, P.G. Ranjith, Mechanical
effect on mechanical behavior of rock under confinement, J. Rock Mech. Geo. Eng. characterisation of jointed rock-like material with non-persistent rough joints
11 (6) (2019) 1201–1210, https://doi.org/10.1016/j.jrmge.2019.07.003. subjected to uniaxial compression, Eng. Geol. 260 (2019) 105224, https://doi.org/
[10] J. Shang, Rupture of veined granite in polyaxial compression: Insights from three- 10.1016/j.enggeo.2019.105224.
dimensional Discrete Element Method modelling, J. Geophys. Res. Sol. Ea. (2020), [31] B. Pan, K. Qian, H. Xie, A. Asundi, Two-dimensional digital image correlation for in-
https://doi.org/10.1029/2019JB019052. plane displacement and strain measurement: a review, Meas. Sci. Technol. 20 (6)
[11] M. Naderloo, M. Moosavi, M. Ahmadi, Using acoustic emission technique to (2009) 062001.
monitor damage progress around joints in brittle materials, Theor. Appl. Fract. Mec. [32] P.Z. Pan, S. Miao, Q. Jiang, Z. Wu, C. Shao, The influence of infilling conditions on
104 (2019) 102368, https://doi.org/10.1016/j.tafmec.2019.102368. flaw surface relative displacement induced cracking behavior in hard rock, Rock
[12] H. Chen, X. Fan, H. Lai, Y. Xie, Z. He, Experimental and numerical study of granite Mech. Rock Eng. 1–22 (2019), https://doi.org/10.1007/s00603-019-02033-x.
blocks containing two side flaws and a tunnel-shaped opening, Theor. Appl. Fract. [33] Z. Aliabadian, G.F. Zhao, A.R. Russell, Failure, crack initiation and the tensile
Mec. 104 (2019) 102394, https://doi.org/10.1016/j.tafmec.2019.102394. strength of transversely isotropic rock using the Brazilian test, Int. J. Rock Mech.
[13] P. Feng, F. Dai, Y. Liu, N.W. Xu, T. Zhao, Effects of strain rate on the mechanical and Min. Sci. 122 (2019) 104073, https://doi.org/10.1016/j.ijrmms.2019.104073.
fracturing behaviors of rock-like specimens containing two unparallel fissures under [34] Z. Aliabadian, G.F. Zhao, A.R. Russell, Crack development in transversely isotropic
uniaxial compression, Soil Dyn. Earthq. Eng. 110 (2018) 195–211, https://doi.org/ sandstone discs subjected to Brazilian tests observed using digital image correlation,
10.1016/j.soildyn.2018.03.026. Int. J. Rock Mech. Min. Sci. 119 (2019) 211–221, https://doi.org/10.1016/j.
[14] P. Feng, F. Dai, Y. Liu, N.W. Xu, H.B. Du, Coupled effects of static-dynamic strain ijrmms.2019.04.004.
rates on the mechanical and fracturing behaviors of rock-like specimens containing [35] M. Sharafisafa, L. Shen, Q. Xu, Characterisation of mechanical behaviour of 3D
two unparallel fissures, Eng Fract Mech. 207 (2019) 237–253, https://doi.org/10. printed rock-like material with digital image correlation, Int. J. Rock Mech. Min.
1016/j.engfracmech.2018.12.033. Sci. 112 (2018) 122–138, https://doi.org/10.1016/j.ijrmms.2018.10.012.
[15] X.P. Zhou, Y.T. Wang, J.Z. Zhang, F.N. Liu, Fracturing behavior study of three- [36] M. Sharafisafa, L. Shen, Y. Zheng, J. Xiao, The effect of flaw filling material on the
flawed specimens by uniaxial compression and 3D digital image correlation: compressive behaviour of 3D printed rock-like discs, Int. J. Rock Mech. Min. Sci.
Sensitivity to brittleness, Rock Mech. Rock Eng. 52 (3) (2019) 691–718, https://doi. 117 (2019) 105–117, https://doi.org/10.1016/j.ijrmms.2019.03.031.
org/10.1007/s00603-018-1600-4. [37] Q.B. Lin, P. Cao, R.H. Cao, H. Lin, J.J. Meng, Mechanical behavior around double
[16] X.P. Zhang, L.N.Y. Wong, Crack initiation, propagation and coalescence in rock-like circular openings in a jointed rock mass under uniaxial compression, Arch. Civ.
material containing two flaws: a numerical study based on bonded- particle model Mech. Eng. 20 (1) (2020) 1–18, https://doi.org/10.1007/s43452-020-00027-z.
approach, Rock Mech. Rock Eng. 46 (5) (2013) 1001–1021, https://doi.org/10. [38] K.Y. Kim, L. Zhuang, H. Yang, H. Kim, K.B. Min, Strength anisotropy of Berea
1007/s00603-012-0323-1. sandstone: results of X-ray computed tomography, compression tests, and discrete
[17] Y.T. Wang, X.P. Zhou, Y.D. Shou, The modeling of crack propagation and coales- modeling, Rock Mech. Rock Eng. 49 (4) (2016) 1201–1210, https://doi.org/10.
cence in rocks under uniaxial compression using the novel conjugated bond-based 1007/s00603-015-0820-0.
peridynamics, Int J Mech Sci. 128–129 (2017) 614–643, https://doi.org/10.1016/j. [39] X.X. Yang, H.W. Jing, W.G. Qiao, Numerical investigation of the failure mechanism
ijmecsci.2017.05.019. of transversely isotropic rocks with a particle flow modeling method, Processes 6
[18] J. Shang, Y. Gui, Z. Zhao, Broad-spectrum fracture toughness of an anisotropic (9) (2018) 171, https://doi.org/10.3390/pr6090171.
sandstone under mixed-mode loading, Theor. Appl. Fract. Mec. 96 (2018) 556–575, [40] J. Shang, K. Duan, Y. Gui, K. Handley, Z. Zhao, Numerical investigation of the direct
https://doi.org/10.1016/j.tafmec.2018.07.005. tensile behaviour of laminated and transversely isotropic rocks containing incipient
[19] X. Fan, K. Li, H. Lai, Y. Xie, R. Cao, J. Zheng, Internal stress distribution and bedding planes with different strengths, Comput Geotech. 104 (2018) 373–388,
cracking around flaws and openings of rock block under uniaxial compression: a https://doi.org/10.1016/j.compgeo.2017.11.007.
particle mechanics approach, Comput. Geotech. 102 (2018) 28–38, https://doi.org/ [41] J. Shang, L.B. Jayasinghe, F. Xiao, K. Duan, W. Nie, Z. Zhao, Three-dimensional
10.1016/j.compgeo.2018.06.002. DEM investigation of the fracture behaviour of thermally degraded rocks with
[20] H. Lin, H.T. Yang, Y.X. Wang, Y.L. Zhao, R.H. Cao, Determination of the stress field consideration of material anisotropy, Theor. Appl. Fract. Mec. 104 (2019) 102330, ,
and crack initiation angle of an open flaw tip under uniaxial compression, Theor. https://doi.org/10.1016/j.tafmec.2019.102330.
Appl. Fract. Mec. 104 (2019) 102358, https://doi.org/10.1016/j.tafmec.2019. [42] M. Alneasan, M. Behnia, R. Bagherpour, Analytical investigations of interface crack
102358. growth between two dissimilar rock layers under compression and tension, Eng
[21] L.N.Y. Wong, H.H. Einstein, Crack coalescence in molded gypsum and Carrara Geol. 259 (2019) 105188, , https://doi.org/10.1016/j.enggeo.2019.105188.
marble: part 1. Macroscopic observations and interpretation, Rock Mech. Rock Eng. [43] ISRM, Suggested methods for determining the uniaxial compressive strength and
42 (3) (2009) 475–511, https://doi.org/10.1007/s00603-008-0002-4. deformability of rock materials, Int. J. Rock Mech. Min. Sci. 16 (2) (1979) 138–140,
[22] L.N.Y. Wong, H.Q. Li, Numerical study on coalescence of two coplanar flaws in https://doi.org/10.1016/0148-9062(79)91451-7.

11