Int J Coal Sci Technol (2018) 5(2):126–141

https://doi.org/10.1007/s40789-018-0194-5

Numerical analysis of the failure process of soil–rock mixtures

through computed tomography and PFC3D models
Yang Ju1,2 • Huafei Sun3 • Mingxu Xing3 • Xiaofei Wang4 • Jiangtao Zheng2

Received: 10 December 2017 / Revised: 28 December 2017 / Accepted: 1 January 2018 / Published online: 25 January 2018
 The Author(s) 2018. This article is an open access publication

Abstract Soil–rock mixture (SRM) is a unique type of geomaterial characterized by a heterogeneous composition and a
complicated structure. It is intractable for the continuum-based soil and rock mechanics theories to accurately characterize
and predict the SRM’s mechanical properties. This study reports a novel numerical method incorporating microfocus
computed tomography and PFC3D codes to probe the deformation and failure processes of SRM. The three-dimensional
(3D) PFC models that represent the SRM’s complex structures were built. By simulating the entire failure process in
PFC3D, the SRM’s strength, elastic modulus and crack growth were obtained. The influence of rock ratios on the SRM’s
strength, deformation and failure processes, as well as its internal mesoscale mechanism, were analyzed. By comparing
simulation results with experimental data, it was verified that the 3D PFC models were in good agreement with SRM’s real
structure and the SRM’s compression process, deformation and failure patterns; its intrinsic mesomechanism can be
effectively analyzed based on such 3D PFC models.

Keywords Soil–rock mixture (SRM)
PFC3D model
Three-dimensional structure
Microfocus computed tomography
(lCT)
Failure mechanism
Crack growth

1 Introduction combination of a continuous microstructure between soil

particles and a discontinuous macrostructure between soil
Soil–rock mixture (SRM) is a unique type of heterogeneous particles and rock gravels, making the physical and
and discontinuous geomaterial that is composed of fine soil mechanical properties of SRM distinct from those of soil
and coarse rock gravels. Structurally, SRM is a and rock gravels (Zhong 1994; Yin 1998; Li et al. 2004;
Liao et al. 2006). This unique geomaterial has captured
attention in civil and infrastructure engineering applica-
& Yang Ju
tions, such as hydropower construction (Casagli et al.
juy@cumtb.edu.cn
2003), landslide monitoring (Li et al. 2007; Xu and Hu
1
State Key Laboratory for Geomechanics and Deep 2008), backfill engineering (Börgesson et al. 2003), and
Underground Engineering, China University of Mining and transportation engineering (Zhao et al. 2006), among other
Technology, No. 1 University Avenue, Xuzhou 221116,
applications. Especially, geological hazards caused by
China
2
landslide and debris flow occurred frequently and resulted
State Key Laboratory of Coal Resources and Safe Mining,
in serious casualty and great economic losses (Heim 1882;
China University of Mining and Technology, D11 Xueyuan
Road, Beijing 100083, China Shang et al. 2001; Xu et al. 2009). Thus, it is important to
3 well understand and accurately describe the strength,
School of Mechanics and Civil Engineering, China
University of Mining and Technology, D11 Xueyuan Road, deformation and failure behavior of SRM to ensure the
Beijing 100083, China safety and stability of engineering applications.
4
State Intellectual Property Bureau of China, No. 6 Xitucheng To disclose the failure behavior and governing mecha-
Road, Beijing 100088, China nisms of SRM, pioneer investigations have been conducted

123
Numerical analysis of the failure process of soil–rock mixtures through computed tomography… 127

in both theory and engineering practice. Regarding in situ (DFM), and the discrete element method (DEM), were used
investigations, the deformation and failure properties of to analyze the deformation and failure properties of SRM.
SRM were analyzed by the horizontal push-shear in situ For example, based on FEM, a mathematical model for the
tests, and the strength was reported to have a close relation rain infiltration in the rock–soil slop has been established
with the rock shape and soil cohesiveness (He 2004). The and solved (Liu et al. 2005). The large-scale direct shear
concept of the average sliding surface, which is used to tests of inhomogeneous SRM were simulated through
calculate the strength parameters, and the concept of the FEM, and three types of failure models during the failure
key diameter, which is important for determining the process were proposed (Xu et al. 2007). Based on DFM,
cohesion value, were put forward (Xu et al. 2006). The the control function of the shape and distribution charac-
mechanical properties of SRM were found to be closely teristics of a rock block on SRM deformation failure was
related to the rock content, mainly predominated by the explored by FLAC3D, and the relationship between elastic
staggered structure of the rock blocks (Xu et al. 2011). To modulus and rock content was presented (You 2002). The
probe the deformation and failure properties of SRM under thermo-hydromechanical behavior of Callovo–Oxfordian
cyclic loads, an in situ horizontal push-shear cycle test was claystone was investigated by FLAC3D through a three-
performed (Xu et al. 2008a). The shear-strength calculating dimensional (3D) simulation of an in situ heating experi-
methods were proposed, and the liquefaction resistance and ment, and a set of basic parameters involved in thermo-
permeability properties of SRM were studied (Lin et al. hydro-mechanical processes were discussed (Bian et al.
2004; Chen et al. 2012). 2012). Based on DEM, a stochastic PFC3D model of SRM
Studies involving laboratory tests were mainly focused was established, and a numerical experiment was con-
on the impact of SRM components such as the composition ducted to verify the structural effect on the shear-push
and distribution of particle size, rock content and moisture deformation damage (He et al. 2009). The failure process
condition on the macro-mechanical properties of SRM. For of SRM during a biaxial test was simulated using PFC3D
instance, a direct shear test was performed on rock–soil (Cheng et al. 2010). A complete integration of the 3D
mixtures with different porosities, and the shear strength fracture system, which linked the 3D fracture system into a
was found to depend largely on the aggregation degree of 2D PFC model, was provided, and a 3D SRM model was
the rocks and soil (Vallejo and Mawby 2000). The rainfall- subsequently formed to analyze the stress–structure inter-
induced SRM landslide was reproduced by utilizing the action on the stability of vertical excavations in hard rock
self-developed laboratory experimental apparatus, and the (Hadjigeorgiou et al. 2009). The relationship between
observed failure phenomena revealed that the landslide moisture content and the process of loosening deposit
failure mode depended greatly on the grain size and fine forming in the mud-rock flow was revealed by PFC2D (Hu
particle content (Wang and Sassa 2003). The effect of et al. 2010).
gravel content on SRM shear properties was also investi- It is noteworthy that the geometry and distribution of
gated (Kuenza et al. 2004; Xu et al. 2008b; Ouyang et al. rock gravels inside SRM are complex, and the most
2010). Considering the nonlinear properties of SRM, the essential attribute is that SRM’s composition and structure
fractal theory was applied to study the strength properties are highly heterogeneous and discontinuous, which causes
of SRM (Shu et al. 2009). The needle penetration resis- not only the particularity of its deformation and failure
tance, uniaxial compressive strength, tensile strength and mechanism but also the significant difference between
elasticity modulus of clay-bearing rocks with different analysis methods for SRM and those for conventional soil
water contents were explored by laboratory tests, and a mechanics or rock mechanics. Conventional continuum-
method for estimating the clay-bearing rock strength and based theories of soil mechanics or rock mechanics are
deformability was proposed (Erguler and Ulusay 2009). difficultly applied to accurately describe the deformation
The effect of saturation on the elastic wave velocity of and failure behaviors of this special geomaterial whose
clay–rock samples was studied, and the applicability of properties are between those of homogeneous soil and
classical models was verified (Ghorbani et al. 2009). On those of rock fragments (Xu and Zhang 2013; Sun et al.
the basis of high-resolution microfocus computed tomog- 2014c). For those theories or analysis methods that were
raphy (lCT) techniques, the generation, evolution and not based on the essential attribute of SRM, it would
coalescence of internal fractures of SRM specimens with inevitably be difficult to provide scientific bases and
four different rock contents were explored, and the internal effective ways to evaluate the stability of the geomaterial.
damage evolution mechanism and the influence of the rock Regarding SRM experimental research, restraints exist in
content on SRM mechanical properties were analyzed (Sun both in situ field tests and laboratory tests. For instance,
et al. 2012, 2014a, b). in situ field tests greatly rely on the experimental apparatus
Different methods of numerical research, such as the and environmental conditions, for which the test durations
finite element method (FEM), the finite difference method are limited. In most cases, field tests can only obtain the

123
128 Y. Ju et al.

macro mechanical properties, making it difficult to reveal established a DEM particle library with representative
the intrinsic meso-mechanism of the SRM deformation and particles that were generated by clumping tens of balls
failure process. Laboratory tests focusing on SRM’s together in PFC3D to match the scanning image obtained
internal structures have been performed using laser scan- from the tests (Yu and Shen 2012, 2013). Subsequently,
ning (Lanaro and Tolppanen 2002), digital photography particles of different sizes could be selected from this
(Yue et al. 2003) and CT techniques (Sun et al. 2012; Sun particle library to generate the specimen model. Although
et al. 2014a). As a nondestructive technique, CT scanning great achievements have been made in modeling the real
is more preferable for studying SRM’s complex structure. 3D materials containing blocks with irregular shapes, dif-
However, CT image processing is a challenging, time- ferences still exist between the internal structure of the
consuming work, the quality of which has a large influence generated model and the real sample, and it is difficult to
on the subsequent analytical results. In addition, it is determine the corresponding prototype tests of the real
expensive to obtain CT images of SRM’s entire deforma- sample for comparison with simulations of these models.
tion and failure process through advanced 3D CT mea- Therefore, generating PFC3D models that can reflect
surement techniques (Spanne et al. 1994; Coker et al. SRM’s real heterogeneity and irregular structure, as well as
1996). conduct a numerical simulation of SRM’s deformation,
Despite that numerical simulation has become a pow- failure and crack propagation process based on such a real
erful tool to study SRM’s deformation and failure proper- 3D model, remains a great challenge.
ties, some challenges need to be tackled. Generally, most To solve the remaining problems in numerically char-
of SRM’s numerical models cannot accurately represent acterizing the mechanical properties and failure mecha-
the real, complex 3D structure of SRM. The problems of nisms of SRM, we propose using a lCT technique to
SRM’s discontinuous structures and crack propagation also acquire the real structural information of SRM samples and
pose certain limitations to SRM numerical research (Sun constructing a 3D SRM model in PFC3D to represent
et al. 2014b). Recently, PFC3D (Itasca_Consulting_Group SRM’s real heterogeneity and irregular structure. By
1999a, c) has been used to model complicated problems in applying PFC3D code to simulate the entire failure pro-
solid mechanics and granular flow. Compared with other cesses of SRM cylinders under uniaxial compressive loads,
simulation programs, PFC is markedly superior for mod- we attained and analyzed the mechanical properties,
eling rock size gradation and distribution, as well as the including strength, elastic modulus, and deformability, and
interaction of rock blocks inside SRM (Zhou et al. 2000; Fu the failure behaviors, such as 3D crack initiation, growth
et al. 2004; Luo et al. 2008; Fakhimi 2009; Ding et al. and distribution, gravel movement, of SRM. The influences
2010, 2012). However, problems remain with respect to of rock ratios (0%, 10%, 20%, 30% and 50%) on the
building PFC models that can represent SRM’s real strength, deformation and failure modes, as well as its
structure. At present, the most SRM models based on PFC internal meso-mechanism, of SRM were discussed. In
are the stochastic structure models in which the rock blocks contrast to building a stochastic structural model, the pro-
are generated according to the particles’ self-organization posed approach enables conducting numerical simulation
on the macroscopic statistical level (You and Tang 2002; and evaluation of SRM’s properties based on real 3D
He 2004; Gan and Kamlah 2010; Liu et al. 2013). In some structural models, which are undoubtedly closer to the
models, the rock blocks are created based on the two-di- actual situation.
mensional (2D) representation of real SRM structure (Xu This paper is organized as follows: Sect. 2 outlines the
et al. 2007; Ding et al. 2010; Zhou et al. 2013). Few 3D specimen preparation, CT scanning procedures and iden-
PFC models with real rock geometry of SRM are available. tification methods for acquiring structural information of
However, in other research fields, PFC has also been uti- SRM. Section 3 presents the SRM modeling method in
lized to simulate 3D blocks with irregular geometry. Lu PFC3D and the setup for numerical simulation of SRM’s
et al. proposed a modeling method in which overlapping deformation and failure process under uniaxial compres-
balls were used to form clumps using a simple procedure sion tests. Section 4 discusses the simulation results from
that controls the sphericity, angularity and surface texture different aspects, including failure modes, crack growth
of the clump to model the real ballast particles (Lu and and distribution, rock translation and rotation, and strength
McDowell 2007). Using a high-resolution scanner, You and modulus. To verify the simulation, the numerical
et al. (2008) captured the 2D microstructure of the asphalt results are compared with the CT experimental results.
concrete mixture, and a 3D model at a depth of 32 mm was Conclusions are presented in Sect. 5.
generated by using four 2D models, each of which was
duplicated 12 times, approximately 8 mm deep. To simu-
late the real asphalt mixtures of different aggregate gra-
dations, shapes and angularity distributions, Yu and Shen

123
Numerical analysis of the failure process of soil–rock mixtures through computed tomography… 129

2 Computed tomography identification fractured specimens that experienced various loads. During
and acquisition the CT investigation, six loading stages of the uniaxial
compression (i.e., unload, 40% pre-peak, 70% pre- peak,
2.1 Specimen preparation peak, 70% post-peak, and 40% post-peak) were applied
simultaneously with CT scanning using a self-designed
Thirty cylindrical specimens of 33 mm in diameter and loading device, as shown in Fig. 1. At every loading stage,
66 mm in height were prepared. Fifteen of the specimens 1024 CT horizontal cross-section images of 550 9 550
were tested to acquire the uniaxial compressive strengths, pixels with a resolution of 72 lm per pixel were obtained.
elastic modulus and deformation characteristics, ten were Figure 2, as an example, displays one original CT image of
scanned to identify the internal fractures of SRM when a 50% rock ratio specimen.
loaded, and the remaining five were prepared as standbys. The quality of the original CT images is far from being
Five different rock ratios (i.e., 0%, 10%, 20%, 30%, and able to meet the requirements of precise image analysis and
50%) of SRM were tested. For every rock ratio, there were data acquisition. To improve image quality, we applied the
three specimens for uniaxial compressive tests and two special image processing measures as follows:
specimens for CT tests. The rock ratio refers to the pro-
(1) Noise removal. There were several noises randomly
portion of gravel mass to overall mass of SRM. Table 1
distributed in the original CT images and some slight
lists the ingredients of SRM. The soil component consists
ring artifacts, which caused great difficulty in
of 10% sand (0.5–1 mm) and 90% loess (\ 0.5 mm). The
recognizing detailed SRM information. Therefore,
rock component comprises small-size gravels (1–4 mm)
a Gaussian filter was used to remove noises and ring
and large-size gravels (4–7 mm). The water content refers
artifacts first.
to the ratio of water mass to soil mass, as shown in Table 1.
(2) Multiple-threshold segmentation. Every component
of SRM, considered homogeneous, is supposed to
2.2 CT image processing and acquisition
have the same gray value. Through multiple-thresh-
of structural information
old segmentation, images were redrawn with three
colors (black, gray and white). The black portions
To acquire the failure characteristics of SRM when loaded,
represent the fracture and outer space of the spec-
we employed a CT machine to scan the entire structures of
imen; the gray portions indicate soil, and the white
portions refer to rock gravels.
Table 1 SRM ingredients
(3) Gray value compensation. Another problem was that
Rock ratio (%) 0 10 20 30 50 the gray value in the CT image was found to be
Soil ratio (%) 100 90 80 70 50 greater at the specimen edge. In other words, as the
Small-size rock ratio (%) 0 7 14 21 35 distance between a certain pixel point and the image
Large-size rock ratio (%) 0 3 6 9 15 center becomes larger, the gray value of that pixel
Initial water content (%) 15 15 15 15 15 tends to be large. This phenomenon was verified by
Final water content upon testing (%) 6 6 6 6 6 the statistics obtained from the specimen at a rock
ratio of 0%, whose gray values were supposed to be
the same in the image but were not. To relieve this
negative impact, based on the statistics, we calcu-
lated the pixel-value difference that each pixel
requires to be compensated. A computer program
was developed to automatically determine compen-
sation for correcting gray-value of each point. The
images were improved through the gray-value
correction.
(4) Partial smoothing. A partial smoothing program was
applied to smooth the rock edge without affecting
detailed information on the fractures inside the SRM.
This program was accomplished by applying the
median filter to the image, which only includes the
rock information that was extracted from the mul-
tiple-threshold segmented image. Then, the
Fig. 1 The loading apparatus mounted on the CT objective stage

123
130 Y. Ju et al.

preprocessed images. Define L as a length threshold.

By detecting each portion with the same gray value
in the image, the projected length (L1 and L2) in the
X and Y directions of the portion were obtained. If
both L1 and L2 were less than L, the color of that
portion was replaced by its surrounding color.
(6) Background extraction. Using multi- threshold seg-
mentation is not enough to separate all of the
specimen components clearly. When the fracture and
outer space of the specimen have almost the same
performance when scanned by the CT technique,
their gray values are close to each other, sometimes
even mixed, which are difficult to distinguish from
each other. Thus, a computer program of regional
morphological segmentations was developed to
Fig. 2 The original CT image identify and extract the fractures. As a result, the
black portions represent the fractures, and the outer
smoothed rock image and multiple-threshold seg- space was given the color of light gray.
mented image were combined to obtain the result.
(5) Noisy-points elimination. Noisy points still existed Figure 3 illustrates the results of CT images processing
in a disorderly fashion in the image after applied and acquisition, including noise removal, multiple-thresh-
multi-threshold segmentation. To remove the noisy old segmentation, gray value compensation, partial
points, we applied the following measures to the

Fig. 3 Illustration of CT image processing, identification and acquisition (a–f represent the imaging results of noise removal, multiple-threshold
segmentation, gray value compensation, partial smoothing, noisy-particle elimination and background extraction, respectively)

123
Numerical analysis of the failure process of soil–rock mixtures through computed tomography… 131

smoothing, noisy- point elimination, and fracture

extraction.

3 PFC3D modeling and simulation

3.1 PFC3D modeling

The SRM model can be considered the combination of a

Fig. 5 Volume mesh information of a rock gravel
soil model and a rock model. The basic PFC3D modeling
procedure is as follows: First, we build a soil model with
balls randomly distributed; next, we select a certain group
of balls that belong to the rock according to the real rock
information; and finally, we combine this group of balls to
represent the rock gravels. We will subsequently take one
rock as an example to illustrate the specific modeling
methods in PFC3D, as follows:
Step 1: Generate walls in PFC3D to build a closed
cylindrical space, which has the same size as the actual
specimen. Then, generate randomly distributed balls inside
the space using the GENERATE command to create the
pure soil model. The ball radius in our model is set at
Fig. 6 Single rock model
0.5 mm; consequently, the soil model has 37,925 balls in
total. This step is shown in Fig. 4.
Step 3: Using FISH programming, the centroid infor-
Step 2: Input the processed CT images into the MIMICS
mation of every ball in the soil model is read, and the
platform to reconstruct the CT model, which has the same
distance between the ball centroid and the rock mesh node
structure as the prototype specimen. Then, draw the vol-
is calculated. If this distance is sufficiently small so that the
ume mesh of a certain rock gravel to obtain the mesh nodes
centroid is close enough to the node, this ball will be
information, as shown in Fig. 5. By conducting the sec-
considered to belong to the rock. Using this method, a
ondary development with FISH programming, the rock
group of balls belonging to a certain rock will be selected.
volume mesh is input into PFC3D.
Step 4: Use the CLUMP command to combine that
group of balls to represent a certain rock gravel inside
SRM, as shown in Fig. 6.
The above steps demonstrate how to build the single
rock model. Using this method, we can draw the volume
mesh of all the rock inside SRM together and then build the
whole SRM model. Figure 7 shows the rock volume mesh
of SRM with a 30% rock ratio, and Fig. 8 displays the final
PFC3D model of SRM with a 30% rock ratio. In Fig. 8, the
white color represents soil particles, and the other colors
represent rock particles. In this SRM model, all the rock
gravels are generated in one step, according to the CT
reconstruction model. Each rock clump has its own unique
shape resembling the counterpart rocks in the real SRM
specimen. The biggest advantage is that simulation results
conducted with this model can be compared with the
experimental results produced by the corresponding spec-
imen. Furthermore, the generating procedure is very simple
because all the manipulating steps have been integrated
into one program.
To verify the effectiveness of SRM PFC3D models, the
cross section of the PFC3D model was cut out for
Fig. 4 Soil particle model

123
132 Y. Ju et al.

comparison with the CT image. Figure 9a illustrates one

cross-section of the PFC3D model. Figure 9b displays the
original CT image. In addition, Fig. 9c shows the com-
parison between the cross section of the PFC3D model and
the prototype CT image. It is shown that the PFC3D model
can reflect the real heterogeneity and irregular structure of
SRM.
Some investigations indicate that SRM can be consid-
ered a system that comprises soil, rock and a soil/rock
interface. The interface strength has a large influence on
Fig. 7 Volume meshes for rocks SRM’s deformation and failure properties (Wang and Li
2014; Wang et al. 2014). Consequently, it would be rea-
sonable to consider the soil/rock interface in the model. In
PFC3D, there will be a ‘‘contact’’ between one ball and its
neighboring ball. If two balls contact each other, the
‘‘contact’’ will be active. Otherwise, the ‘‘contact’’ will be
inactive. By giving the ‘‘contact’’ a certain bond, the whole
model will perform as an entirety. Alternatively, the balls
in the model will perform discretely. Consequently, if a
different contact type can be separated, the soil/rock
interface will be modeled. Using FISH programming, the
active ‘‘contacts’’ between soil particle and rock particle or
between rock particle and rock particle were selected.
Figure 10 demonstrates the 2D soil/rock interface and
contacts, and Fig. 11 illustrates the 3D soil/rock interface
in which panel (a) is the 3D rock geometry and panel (b) is
the soil/rock interfaces and contacts surrounding the rock.
By giving different bond strengths for the different contact
types, the soil/rock interface will be modeled.

3.2 Calculation setup and parameters

Fig. 8 A complete PFC3D model of SRM with a 30% rock ratio. The
The boundary condition was set according to the SRM
white color represents soil particles, and the other colors represent
rock particles uniaxial compressive tests. The lateral walls were deleted
to eliminate lateral confinement. In addition, the bottom

Fig. 9 Comparison between the selected cross section of the PFC3D model and the prototype CT image, a the cross section truncated from the
constructed PFC3D model, b the prototype CT image of the selected cross section, c comparing the rock gravels that are represented in PFC3D
model with the real ones that are shown in the CT image

123
Numerical analysis of the failure process of soil–rock mixtures through computed tomography… 133

Table 2 Meso-level parameters of the soil model

Particle parameters Contact bond parameters

Particle density (kg/m3) 1776 Mean normal strength 0.41

(MPa)
Particle contact module 0.12 SD of the normal strength 0.06
(GPa) (MPa)
Particle normal/shear 3.5 Mean shear strength (MPa) 0.82
stiffness
Particle friction 0.6 SD of the shear strength 0.12
coefficient (MPa)

particles are combined by the CLUMP command making

every rock model a rigid body that will not break apart. The
Fig. 10 Illustration of 2D soil/rock interfaces and contacts ‘‘contact bond’’ mode is set to define the bond between two
balls contacting each other. The meso-level parameters of
the soil model applied in our simulation are shown in
Table 2. In the natural state, different conditions may yield
different interface types. Some interfaces have weak bonds
with low strength, while others have strong bonds with
high strength (Wang and Li 2014; Wang et al. 2014). In our
simulation, it is difficult to determine the interface strength
before specific investigation. Therefore, currently, we only
set the parameters of the soil/rock contact bond to be the
same as those of the soil/soil contact bond. Figure 12
illustrates the initial PFC3D models.

4 Simulation results and analyses

4.1 Failure processes and patterns

Figure 13 shows the failure patterns of SRM models with

Fig. 11 Illustration of 3D soil/rock interfaces and contacts, a 3D rock
different rock ratios, in which blue and green disks repre-
geometry, b soil/rock interfaces and contacts surrounding the rock
sent tensile and shear fractures or cracks. The SRM models
wall was fixed, while the top wall had a velocity to apply a exhibit lateral expansion owing to the vertical displacement
displacement load. loading. The cracks tend to develop longitudinally and
In PFC3D, it is considerably difficult to determine the obliquely. Furthermore, in some models, shear zones form.
appropriate values of meso-level parameters of particles Comparing the SRM failure patterns in PFC3D with those
because the values cannot be directly obtained by labora- obtained through CT tests, as shown in Fig. 14, some
tory tests, field measurements or theoretical calculations. similarities were observed in the failure patterns, thereby
Our solution is to numerically test the models by a trial and verifying the effectiveness of the simulation results.
to adjust the values of the meso-level parameters of parti-
cles until the test results are in accord with those of real 4.2 Crack growth
samples. In this study, the meso-level parameters of the
SRM model in PFC3D were tested and obtained according In PFC3D, the bond between two particles will fail due to
to the method recommended by the Itasca Corporation its normal or shear strength being exceeded. Then, the
Consulting Group and other scholars to save adjusting time microcrack initiates, the geometry and location of which
(Itasca_Consulting_Group 1999b; Kulatilake et al. 2001; are determined by the sizes and current locations of the two
Potyondy and Cundall 2004; Cho et al. 2007; Guo et al. parent particles from which the microcrack originated. This
2013). Considering that the stiffness of rock gravels in our means that once two particles break, a microcrack initiates.
test is much larger than that of soil, in the simulation, rock However, due to the resolution limitation, CT scanning
cannot reveal these tiny cracks. Therefore, to maintain

123
134 Y. Ju et al.

Fig. 12 Initial PFC models (Light yellow and gray represent the soil particles and rock particles, respectively; a–e represent rock ratios of 0%,
10%, 20%, 30% and 50%, respectively)

Fig. 13 Failure patterns of the PFC3D models with various rock ratios (images a–e represent rock ratios of 0%, 10%, 20%, 30% and 50%,
respectively)

Fig. 14 Failure patterns of the prototype models acquired through CT scan (images a–e represent rock ratios of 0%, 10%, 20%, 30% and 50%,
respectively), in which dark blue color indicates the fractures or cracks when loaded

123
Numerical analysis of the failure process of soil–rock mixtures through computed tomography… 135

consistency with our CT tests, tiny cracks in PFC3D were Fig. 16 PFC3D simulation results of 3D crack growth in SRM c
eliminated by FISH programming. This means that if the (Transparent light gray represents soil particles, and opaque dark gray
indicates rock particles; the blue and green disks represent tensile
distance between two broken particles in the PFC model is cracks and shear cracks, respectively. The rows from a–e represent
smaller than 72 lm (the resolution of our CT scanning), the five different rock ratios of 0%, 10%, 20%, 30% and 50%,
microcrack will not appear in the simulation results. respectively. The images from left to right of each row show the
To investigate the properties of crack growth, six load- crack distribution and growth when the applied load reached the
levels of null, peak load, 70% post-peak load, 40% post-peak load and
ing stages were simulated, which are 40% pre-peak, 70% 20% post-peak load; in addition, the rock particles of the null load are
pre-peak, peak, 70% post-peak, 40% post-peak, and 20% marked according to the CT images.)
post-peak, as shown in Fig. 15. According to the simula-
tion results, there were few cracks before the peak load was
reached; thus, we only selected certain simulation stages to
influence, the cracks mainly had a short length and were
show. Figures 16 and 17 illustrate the 3D crack growth in
concentrated in certain portions near the model edge. In
the PFC3D and CT scan tests, respectively. Moreover,
addition, there were no obvious shear zones.
similarities were found between these two groups of ima-
ges. To further study the behavior of crack growth, a group
4.3 Rock translation and rotation
of vertical sections of SRM models were attained, as
illustrated in Fig. 18. By comparing cracks in different
The simulation results revealed that rock gravels inside the
loading stages, it was found that before the peak load was
SRM had translated and rotated during the loading process.
reached, there were few cracks; however, once the peak
To clearly present and analyze the particle movement
load was reached, a few cracks emerged. As the loading
inside SRM, a partial region of the vertical section of the
continued, some big, scattered cracks occurred; then, some
unloading 50% rock ratio SRM was enlarged, and different
portions of the cracks connected with each other, forming
colors were used to represent different rocks in that region,
the penetrating cracks. A comparison of cracks in SRM
as shown in Fig. 19. Figure 20 demonstrates the rock
models with different rock ratios revealed that when the
movement within the region (panel) of concern during the
rock ratio was low (0% and 10%), the SRMs were mainly
loading process. Panels (a)–(c) reveal that before the peak
controlled by soil. Obvious shear zones and large con-
load was reached, the rock gravels mostly moved down-
nected cracks were inside the models. When the rock ratio
ward, indicating that the SRM was compressed and there
was high (20%, 30% and 50%), because of the rock
were few rotations and horizontal translations. From panel

Fig. 15 Illustration of the loading and simulation stages

123
136 Y. Ju et al.

123
Numerical analysis of the failure process of soil–rock mixtures through computed tomography… 137

Fig. 17 CT images of 3D crack distribution and growth in SRM (Blue, dark gray and transparent light gray represent the crack, rock and soil,
respectively. Rows a and b show the crack distribution and growth when the applied load reached the levels of peak load and 40% post-peak
load)

(d) to panel (f), when the peak load was reached, the rock interface to be identical to those of the soil/soil contact
gravels exhibited obvious horizontal translation, as well as bond could have some influence on the results. These
some rotation, which indicated the lateral expansion of problems are expected to be solved in future studies.
SRM. Taking the green rock gravel as an example, during
the loading process, its position became closer to the bot-
tom and right side of the image, which meant that the green 5 Conclusions
rock gravel was translated vertically and horizontally.
From panel (d) to panel (f), a green particle became pro- This paper proposed a novel numerical method to construct
gressively larger, which indicated that the green rock a three-dimensional SRM model in PFC3D representing
gravel was rotated or translated perpendicularly to the the real complex structure of SRM. A lCT technique was
image. Obvious rock rotation of the pink and red rock used to acquire the real structural information and to
gravels can be observed in panels (e) and (f). identify the entire deformation and failure processes of
SRM under uniaxial compressive loads. We developed an
4.4 Strength and elastic modulus image-processing program to improve CT images and to
extract the internal structure information of SRM speci-
Table 3 compares the results of strength and elastic mod- mens. By conducting the secondary development with
ulus of the numerical simulation with those of the CT tests. FISH programming in the PFC3D platform, the real 3D
It is shown that, unfortunately, a difference exists between SRM structure models were built using the attained CT
the numerical simulation results and the CT experiment information. Each rock clump in the model has its own
results. The causes for this inconsistency, from authors’ unique shape resembling the counterpart rocks in the real
points of view, are possibly due to: (1) the loading pro- SRM specimen. The comparison between the PFC3D
cesses were not identical. The load was applied piecewise model and the prototype shows good consistency. The real
in the CT tests but continuously in the simulation, and (2) heterogeneity and irregular structure of SRM can be rep-
the strength of the soil/rock interface may not be consistent resented by the PFC model.
with the real value. Setting the parameters of the soil/rock

123
138 Y. Ju et al.

Fig. 18 Vertical sectional representations of 3D crack distribution and growth in PFC3D models of SRM (Light yellow and gray represent soil
particles and rock particles, respectively, while the short lines with the other dark color refer to cracks. Rows a–e represent five different rock
ratios of 0%, 10%, 20%, 30%, and 50%, respectively. The images from left to right of each row show the crack distribution and growth when the
applied load reached the levels of null, peak load, 70% post-peak load, 40% post-peak load and 20% post-peak load. The rock geometries at the
unload stage are marked according to the CT images)

123
Numerical analysis of the failure process of soil–rock mixtures through computed tomography… 139

Fig. 19 Partial enlargement panel showing the positions of rock particles

Fig. 20 Illustration of rock translation and rotation (Panels a–f refer to 40% peak load, 70% peak load, peak load, 70% post-peak load, 40%
post-peak load, and 20% post-peak load, respectively)

Based on the PFC models, the entire compression-in- simulation and CT tests have much similarity. Based on the
duced failure process was simulated in PFC3D. The results simulation results, it can be concluded that the rock ratio is
revealed that the SRM failure patterns in the numerical an important factor that influences the SRM failure patterns

123
140 Y. Ju et al.

Table 3 Comparison between the numerical simulation results and Coker DA, Torquato S, Dunsmuir JH (1996) Morphology and
the CT experiment results physical properties of Fontainebleau sandstone via a tomo-
graphic analysis. J Geophys Res Solid Earth 101:17497–17506
Rock ratio (%) 0 10 20 30 50 Ding X, Li Y, Wang X (2010) Particle flow modeling mechanical
Strength of the CT test (MPa) 1.09 2.80 2.22 2.44 1.17 properties of soil and rock mixtures based on digital image. Chin
Strength of the simulation 1.08 1.15 1.20 1.35 1.49 J Rock Mech Eng 29:477–484
(MPa) Ding X, Zhang H, Huang S, Lu B, Zhang Q (2012) Research on
mechanical characteristics of unsaturated soil–rock mixture
Elastic modulus of the CT test 76.7 122.0 72.7 83.0 76.6 based on numerical experiments of mesostructure. Chin J Rock
(MPa) Mech Eng 31:1553–1566
Elastic modulus of the 77.9 86.7 96.1 103.0 122.2 Erguler Z, Ulusay R (2009) Water-induced variations in mechanical
simulation (MPa) properties of clay–bearing rocks. Int J Rock Mech Min Sci
46:355–370
Fakhimi A (2009) A hybrid discrete–finite element model for
numerical simulation of geomaterials. Comput Geotech
and mechanisms. Furthermore, PFC3D shows great 36:386–395
advantages in simulating the translation and rotation of Fu W, Cai J-J, Dong H, Wang A-H, Liu H-Q, Wang L-Y, Tian H
particles or blocks. (2004) Current status of numerical simulation of granular flow.
J Mater Metall 3:172–175
Gan Y, Kamlah M (2010) Discrete element modelling of pebble beds:
Acknowledgements The authors gratefully acknowledge the finan- with application to uniaxial compression tests of ceramic breeder
cial support from the State Key Research Development Program of pebble beds. J Mech Phys Solids 58:129–144
China (Grant No. 2016YFC0600705), the National Natural Science Ghorbani A, Zamora M, Cosenza P (2009) Effects of desiccation on
Foundation of China (Grant Nos. 51674251, 51727807, 51374213), the elastic wave velocities of clay–rocks. Int J Rock Mech Min
the National Natural Science Foundation for Distinguished Young Sci 46:1267–1272
Scholars of China (Grant No. 51125017), the Fund for Creative Guo J, Xu G, Jing H, Kuang T (2013) Fast determination of meso-
Research and Development Group Program of Jiangsu Province level mechanical parameters of PFC models. Int J Min Sci
(Grant No. 2014-27), and the Priority Academic Program Develop- Technol 23:157–162
ment of Jiangsu Higher Education Institutions (Grant No. Hadjigeorgiou J, Esmaieli K, Grenon M (2009) Stability analysis of
PAPD2014), and an open project sponsored by the State Key Labo- vertical excavations in hard rock by integrating a fracture system
ratory for Geomechanics and Deep Underground Engineering (Grant into a PFC model. Tunn Undergr Space Technol 24:296–308
SKLGDUE K1318) for their financial support. He J (2004) Study of deformation and failure mechanisms of rock–
soil aggregate in three gorges reservoir area. Dissertation, China
Open Access This article is distributed under the terms of the
University of Mining and Technology, Beijing
Creative Commons Attribution 4.0 International License (http://crea He J-M, Li X, Wu J-B, Cheng G-W, Xu J-F (2009) Modeling method
tivecommons.org/licenses/by/4.0/), which permits unrestricted use, of the rock–soil aggregate and its numerical test. Min Metall Eng
distribution, and reproduction in any medium, provided you give
29:1–78
appropriate credit to the original author(s) and the source, provide a
Heim A (1882) Der Bergsturz von Elm. Z Deutsch Geol Gesells
link to the Creative Commons license, and indicate if changes were 34:74–115
made. Hu M, Wang R, Chen Z, Wang Z (2010) Initiation process simulation
of debris deposit based on particle flow code. Rock Soil Mech
31:394–397
References Itasca_Consulting_Group (1999a) PFC3D—theory and background
Itasca_Consulting_Group (1999b) PFC3D—use’s guide
Bian H, Jia Y, Armand G, Duveau G, Shao J-F (2012) 3D numerical Itasca_Consulting_Group (1999c) PFC3D—verification problems and
modelling thermo-hydromechanical behaviour of underground example applications
storages in clay rock. Tunn Undergr Space Technol 30:93–109 Kuenza K, Towhata I, Orense RP, Wassan TH (2004) Undrained
Börgesson L, Johannesson L-E, Gunnarsson D (2003) Influence of torsional shear tests on gravelly soils. Landslides 1:185–194
soil structure heterogeneities on the behaviour of backfill Kulatilake P, Malama B, Wang J (2001) Physical and particle flow
materials based on mixtures of bentonite and crushed rock. modeling of jointed rock block behavior under uniaxial loading.
Appl Clay Sci 23:121–131 Int J Rock Mech Min Sci 38:641–657
Casagli N, Ermini L, Rosati G (2003) Determining grain size Lanaro F, Tolppanen P (2002) 3D characterization of coarse
distribution of the material composing landslide dams in the aggregates. Eng Geol 65:17–30
Northern Apennines: sampling and processing methods. J Eng Li X, Liao Q, He J (2004) In-situ tests and a stochastic structural
Geol 69:83–97 model of rock and soil aggregate in the three gorges reservoir
Chen Z, Chen S, Chen J, Sheng Q, Yan H, Hu W (2012) In-situ area, China. Int J Rock Mec Min Sci 41:702–707
double-ring infiltration test of soil–rock mixture. J Yangtze River Li X, Liao Q-L, He J-M, Chen J (2007) Study on in situ tests of
Sci Res Inst 29:52–56 mechanical characteristics on soil–rock aggregate. Chin J Rock
Cheng G-W, He J-M, Li X, Di B-R, Li S-D (2010) Particle flow Mech Eng 26:2377–2384
simulation for soil–rock mixtures under biaxial pressure. Min Liao Q, Li X, Hao Z, Wang S, Wu M, He J (2006) Current status and
Metall Eng 30:1–4 future trends of studies on rock and soil aggregates (RSA). J Eng
Cho NA, Martin C, Sego D (2007) A clumped particle model for rock. Geol 14:800–807
Int J Rock Mec Min Sci 44:997–1010 Lin P-S, Chang C-W, Chang W-J (2004) Characterization of
liquefaction resistance in gravelly soil: large hammer penetration
test and shear wave velocity approach. Soil Dyn Earthq Eng
24:675–687

123
Numerical analysis of the failure process of soil–rock mixtures through computed tomography… 141

Liu Y, Liu Q, Chen H, Gong X, Zhang D, Yao D, Li L (2005) Xu W-j Hu, R-l Yue Z-q, R-j Tan (2009) Genesis and stability of the
Numerical study on rainfall infiltration in rock–soil slop. Sci Zhoujiawan landslide, Three Gorges, China. Bull Eng Geol
China Ser E: Technol Sci 48:33–46 Environ 68:47–54
Liu Z, Zhou N, Zhang J (2013) Random gravel model and particle Xu W, Zhang H (2013) Research status and development trend of
flow based numerical biaxial test of solid backfill materials. Int J soil–rock mixture. Adv Sci Technol Water Resour 33:80–88
Min Sci Technol 23:463–467 Xu W, Hu R, Tan R, Zeng R, Yu H (2006) Study on field test of rock–
Lu M, McDowell G (2007) The importance of modelling ballast soil aggregate on right bank of Longpan in Tiger-Leaping Gorge
particle shape in the discrete element method. Granul Matter area. Chin J Rock Mech Eng 25:1270–1277
9:69–80 Xu W-J, Hu R-L, Wang Y-P (2007) PFC2D model for mesostructure
Luo Y, Gong X-N, Lian F (2008) Simulation of mechanical behaviors of inhomogeneous geomaterial based on digital image process-
of granular materials by three-dimensional discrete element ing. J Chin Coal Soc 32:358–362
method based on particle flow code. Chin J Geotech Eng Xu W, Hu R, Yue Z, Zhang R, Wang G (2008a) Research on
30:292–297 relationship between rock block proportion and shear strength of
Ouyang Z-H, Li S-H, Dai Z-S (2010) On the influence factors of soil–rock mixtures based on digital image analysis and large
mechanical properties for soil–rock mixture. J Exp Mech direct shear test. J Rock Mech Geotech Eng 27:996–1007
25:61–67 Xu W-J, Yue Z-Q, Hu R-L (2008b) Study on the mesostructure and
Potyondy D, Cundall P (2004) A bonded-particle model for rock. Int J mesomechanical characteristics of the soil–rock mixture using
Rock Mech Min Sci 41:1329–1364 digital image processing based finite element method. Int J Rock
Shang Y, Yang Z, Liao Q (2001) Geological hazard distribution and Mech Min Sci 45:749–762
prevention in North of Brahmaputra Canyon, Tibet. Chin J Geol Xu W-J, Xu Q, Hu R-L (2011) Study on the shear strength of soil–
Hazard Control 12:30–40 rock mixture by large scale direct shear test. Int J Rock Mech
Shu Z, Liu X, Liu B, Guo Z (2009) Study of strength properties of Min Sci 48:1235–1247
earth-rock aggregate based on fractals. Chin J Rock Mech Eng Yin Y (1998) Research on geologic hazards of relocation construction
28:2651–2656 at the Three Gorges Reservoir, the Yangtze River. Chin J Geol
Spanne P, Thovert J, Jacquin C, Lindquist W, Jones K, Adler P (1994) Hazard Control 7:59–66
Synchrotron computed microtomography of porous media: You X (2002) Stochastic structural model of the earth–rock aggregate
topology and transports. Phys Rev Lett 73:2001–2004 and its application. Chin J Rock Mech Eng 11:1748–1748
Sun HF, Yang ZK, Xing MX, Ju Y, Yang YM CT (2012) You X-H, Tang J-S (2002) Research on horizontal push-shear in situ
Investigation of fracture mechanism of soil–rock mixtures. In: test of soil and rock–mixture. Chin J Rock Mech Eng
Applied mechanics and materials. Trans Tech Publ, pp 67–71 21:1537–1540
Sun H, Ju Y, Wang X, Xing M, Zhang Q, Yang Y (2014a) Review of You Z, Adhikari S, Dai Q (2008) Three-dimensional discrete element
the study on deformation, failure and the mesomechanisms of models for asphalt mixtures. J Eng Mech 134:1053–1063
rock–soil mixture. Sci Chin 44:172–181 Yu H, Shen S (2012) Impact of aggregate packing on dynamic
Sun H-F, Ju Y, Hang M-X, Wang X-F, Yang Y-M (2014b) 3D modulus of hot mix asphalt mixtures using three-dimensional
identification and analysis of fracture and damage in soil–rock discrete element method. Constr Build Mater 26:302–309
mixtures based on CT image processing. J China Coal Soc Yu H, Shen S (2013) A micromechanical based three-dimensional
39:452–459 DEM approach to characterize the complex modulus of asphalt
Sun S, Xu P, Wu J, Wei J, Fu W, Liu J, Kanungo DP (2014c) Strength mixtures. Constr Build Mater 38:1089–1096
parameter identification and application of soil–rock mixture for Yue Z, Chen S, Tham L (2003) Finite element modeling of
steep-walled talus slopes in southwestern China. Bull Eng Geol geomaterials using digital image processing. Comput Geotech
Env 73:123–140 30:375–397
Vallejo LE, Mawby R (2000) Porosity influence on the shear strength Zhao L, Luo H, Li L (2006) Research on application of impact
of granular material–clay mixtures. Eng Geol 58:125–136 compaction technology to high embankment of highway. Chin J
Wang Y, Li X (2014) Research on damage cracking for rock and soil Rock Mech Eng 25:4191–4197
aggregate using calculation meso-mechanics. Chin J Rock Mech Zhong L (1994) Enlightenments from the accident of Vaiont landslide
Eng 33:020–024 in Italy. Chin J Geol Hazard Control 5:77–84
Wang G, Sassa K (2003) Pore-pressure generation and movement of Zhou J, Chi Y, Chi Y, Xu J-P (2000) The method of particle flow and
rainfall-induced landslides: effects of grain size and fine-particle PFC2D code. Rock Soil Mech 21:271–273
content. Eng Geol 69:109–125 Zhou J, Zhang L-Q, Dai F-C, Min H (2013) Numerical simulation of
Wang Y, Li X, Hao J (2014) Research status and prospect of rock and direct shear tests for rock and soil mixture in a landslide based on
soil aggregate. J Eng Geol 22:112–123 bonded-particle model. Chin J Rock Mech Eng 32:650–652
Xu R, Hu W (2008) Field horizontal push shear test for mechanical
property of soil–rock mixtures under cyclic loading. J Eng Geol
16:63–69

123