materials

Article
Nonlinear Strength Reduction Method of Rock Mass in Slope
Stability Evaluation
Yifan Chen 1 , Yizhou Chen 1 , Hang Lin 1, * and Huihua Hu 1,2, *

1 School of Resources and Safety Engineering, Central South University, Changsha 410083, China
2 Hunan Provincial Communications Planning, Survey and Design Institute, Changsha 410200, China
* Correspondence: hanglin@csu.edu.cn (H.L.); huhuihuahnjg@126.com (H.H.)

Abstract: As the strength parameters of rock mass degrade differently during slope instability,
different factors should be considered in the strength reduction method. Previous nonlinear reduction
methods were essentially implemented based on the Mohr–Coulomb criterion, which was reported
not to reflect the nonlinear performance of rock mass. To address this deficiency, in this study,
the Hoek–Brown criterion was combined with a nonlinear reduction technique for slope stability
evaluation. Firstly, based on the classical definition of safety factors, the relationships that should be
satisfied by each parameter of the critical slope were derived. The critical curve of the slope regarding
the Hoek–Brown constant mb and the uniaxial compressive strength of rock mass σcmass was then
obtained. On the assumption that the slope parameter deterioration conforms to the shortest path
theory, the reduction ratio of σcmass to mb was determined. The more objective k-means algorithm was
employed to automatically search the potential sliding surface, on which the slope safety factor was
calculated as the ratio of sliding resistance to sliding force. Finally, the slopes in published literature
were adopted for verification, and the calculated safety factors were compared with those by other
methods, which showed better efficacy.

Keywords: slope stability; nonlinear criterion; strength reduction method; cluster algorithm

1. Introduction
Citation: Chen, Y.; Chen, Y.; Lin, H.;
Landslides are one of the most dangerous geological disasters, causing heavy casualties
Hu, H. Nonlinear Strength Reduction
and economic losses around the world [1–4]. In rock projects such as transportation,
Method of Rock Mass in Slope
hydropower and mining, the original stable slopes are easily affected by the disturbance
Stability Evaluation. Materials 2023,
of engineering excavations [5–8], leading to more destructive disasters such as landslides
16, 2793. https://doi.org/10.3390/
ma16072793
and collapses [9], such as landslides occurred on the expressway [10] (Figure 1). The
bearing capacity estimation has always been an important issue in the field of geotechnical
Academic Editor: John D. Clayton engineering [11,12], which is also the basis for project design. In recent years, various
Received: 15 February 2023
numerical, analytical, as well as laboratory methods have been applied to evaluate the
Revised: 25 March 2023 bearing capacity of slope rock mass [13–15], while slope stability analysis is an effective
Accepted: 27 March 2023 guarantee for the successful construction and safe operation of a project. In other words,
Published: 31 March 2023 the accurate assessment of slope stability is of great significance to reduce landslide risks,
avoid geological disasters and reduce project investment [16–18].
Currently, both the limit equilibrium method and the strength reduction method
are commonly used in slope stability calculations [19–24]. The former carries out static
Copyright: © 2023 by the authors.
equilibrium analysis on rock mass slices based on the preassumed sliding surface to
Licensee MDPI, Basel, Switzerland.
determine the minimum safety factor [25]. The strength reduction method, on the other
This article is an open access article
hand, simultaneously reduces the cohesion c and the internal friction angle ϕ of the slope
distributed under the terms and
and regards the reduction factor corresponding to the instability as the safety factor of the
conditions of the Creative Commons
slope [26–29], which is expressed in Equation (1).
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).

Materials 2023, 16, 2793. https://doi.org/10.3390/ma16072793 https://www.mdpi.com/journal/materials

Materials 2023, 16, 2793 2 of 16

cinitial tan ϕinitial

Fs = critical
= (1)
c tan ϕcritical

where Fs is the safety factor of slope, ccritical and ϕcritical are respectively the cohesion and
internal friction angle of critical slope, while cinitial and ϕinitial are respectively the initial
Materials 2023, 16, x FOR PEER REVIEW 2 of 17
values of cohesion and internal friction angle of slope.

Figure1.1.Historical
Figure Historicallandslide
landslideininexpressway
expressway[10].
[10].

Currently, with
Compared both the limit
limit equilibrium
equilibriummethod method,and thethe strength
strength reduction
reduction methodhas
method are
commonly
wider used in and
applicability slopepracticability
stability calculations [19–24].analysis
in the stability The former carries out
of projects suchstatic
as mineequi-
libriumhighway
slopes, analysissubgrades,
on rock mass deepslices
buried based
tunnels, on theandpreassumed
water conservancy sliding dams,
surfacesince to deter-
it is
mine the
simple minimum safety
in expression, factor during
convenient [25]. The strength reduction
calculation, and applicative method,toon the other
various hand,
geologi-
cal conditions [30–32].
simultaneously reduces It the
wascohesion
reported,c however,
and the internal that thefriction traditional
anglestrength
φ of thereduction
slope and
method
regardsdismisses
the reduction the different degradations of
factor corresponding tocthe
andinstability
ϕ, and theasreductionthe safety scheme
factorisofstill the
controversial [33–37]. Tang and Zheng
slope [26–29], which is expressed in Equation (1). proposed the double safety factor method for slope
stability analysis [38,39]. Pantelidis and Griffiths [40] also analyzed the influence of differ-
ent reduction strategies on slope stability.c Xue, Dang
initial
tan[41]ϕ initialderived the expression between
Fs = critical = (1)
reduction coefficients through linear attenuation c tan ϕ critical Jiang, Wang [42] suggested
assumption.
that the reduction ratio of cohesion to internal friction should be 1.75 for homogeneous
where
soil Fs isYuan,
slopes. the safety factor
Bai [43] of the
fitted slope, ccritical and
reduction ratioφcritical are respectively
of slope model withthe cohesion
different and
slope
internal
angles andfriction
determinedangle the
of critical
specificslope, whileratio
reduction c by
initial and φinitial are respectively
interpolation calculation.the Oninitial
the
values of cohesion
assumption that there andareinternal
countless friction angle of slope.
combinations of the reduction ratio of cohesion and
Compared
internal with Isakov
friction angle, the limit and equilibrium
Moryachkov method,
[44] proposedthe strength reduction
the shortest method
reduction path has
wider to
theory applicability
calculate the and practicability safety
comprehensive in the factor
stability analysis
of slope, of projects
which was then such as mine
improved
by Yuan,highway
slopes, Li [45] based subgrades, deepc-tan
on critical buriedϕ curves.
tunnels,Fang, and water Chen conservancy
[46] discovered a nonlinear
dams, since it is
instability criterion according
simple in expression, convenient to the concept
during of critical
calculation, andslope.
applicativeTheseto promote
various the rapid
geological
development of nonlinear
conditions [30–32]. It was strength reduction
reported, however, method thatforthe slope stability analysis,
traditional strength and all of
reduction
them
methodare dismisses
implemented by linearly
the different reducing the
degradations oftwo
c and Mohr–Coulomb
φ, and the reduction parameters
scheme [47–50].
is still
Meanwhile,
controversial [33–37]. Tang and Zheng proposed the double safety factor methodthe
rock mechanics gradually realized that the linear criterion cannot reflect for
nonlinear failure
slope stability characteristics
analysis of rock mass
[38,39]. Pantelidis and [51–53],
Griffithsin which
[40] also case reduction
analyzed methodsof
the influence
based on the
different nonlinear
reduction Hoek–Brown
strategies on slopestrength
stability.criterion
Xue,are Dang constantly proposed
[41] derived [54].
the expression
For the
between description
reduction of the nonlinear
coefficients through linear performance
attenuation of rock mass, Hoek
assumption. Jiang,and Wang Brown [42]
proposed
suggested thethat
Hoek–Brown
the reduction strength
ratiocriterion
of cohesion in 1980 based onfriction
to internal abundant triaxial
should test results
be 1.75 for ho-
and engineering
mogeneous soilpractice
slopes. [55–57].
Yuan, Bai In 1992, Hoek et
[43] fitted theal.reduction
corrected ratio the Hoek–Brown
of slope model strengthwith
envelope of rock mass at low stress levels. Subsequently,
different slope angles and determined the specific reduction ratio by interpolation cal- the engineering application
showed
culation.that
On itthewas too conservative
assumption that there in are
terms of rockcombinations
countless mass with good of thequality.
reduction In this
ratio
case, Hoek et al. [58,59] further improved the original results
of cohesion and internal friction angle, Isakov and Moryachkov [44] proposed the short- and proposed generalized
Hoek–Brown
est reduction strength
path theory criteria for rockthe
to calculate mass of different safety
comprehensive quality, which
factor is presented
of slope, which was in
Equation (2). In 2002, Hoek et al. [5–8] comprehensively
then improved by Yuan, Li [45] based on critical c-tanφ curves. Fang, Chen [46] discov- discussed the Hoek–Brown
parameter relations,instability
ered a nonlinear and built criterion
a new determination
according tomethod the concept of criticalmbslope.
of parameters , s andThesea by
introducing geological strength index GSI and disturbance
promote the rapid development of nonlinear strength reduction method for slope stabil- factor D (see Equations (3)–(5)).
ity analysis, and all of them are implemented by linearly reducing the two Mohr–
Coulomb parameters [47–50]. Meanwhile, rock mechanics gradually realized that the
linear criterion cannot reflect the nonlinear failure characteristics of rock mass [51–53], in
which case reduction methods based on the nonlinear Hoek–Brown strength criterion are
constantly proposed [54].
For the description of the nonlinear performance of rock mass, Hoek and Brown
Materials 2023, 16, 2793 3 of 16

At present, the Hoek–Brown strength criterion has been widely used in rock and rock mass
mechanics analysis, rock slope stability analysis, rock tunnel design and other fields [60].

σ1 − σ3 = σci (mb σ3 /σci + s) a (2)

GSI −100
mb = mi e 28−14D (3)

GSI −100
s=e 9−3D (4)

 
a = 0.5 + e−GSI/15 − e−20/3 /6 (5)

where σ1 and σ3 are, respectively, the major and minor principal stresses at rock mass
failure; mb , s, and a are all the rock mass material constants; mi is a material constant, GSI is
the geological strength index, and D is the disturbance factor ranging 0~1.
Current reduction methods based on the nonlinear Hoek–Brown criterion can be
broadly classified into three categories: direct reduction, strength envelope lowering, and
indirect reduction, which are summarized in Table 1. In direct reduction methods, Wu,
Jin [61] believed that only the uniaxial compressive strength σci and material constant mi
should be reduced when the nonlinear strength reduction method was used to calculate
the safety factor of rock slope. Song, Yan [62] discussed seven direct reduction cases of
stability calculation, and found that only by directly reducing the values of σci and GSI can
the reasonable safety factor of a slope be obtained. Yang, Wang [63] selected Hoek–Brown
parameters σci and mi as reducing objects to carry out a stability analysis of surrounding
rock in a tunnel. For the second category, Hammah, Yacoub [64] lowered the envelope of
the Hoek–Brown criterion to achieve global reduction. Lastly, Li, Merifield [65] adopted
the equivalent Mohr–Coulomb parameters to achieve the indirect reduction of the Hoek–
Brown criterion, and established the stability analysis chart of rock slope. Shen et al. [66,67]
deduced the instantaneous equivalent Mohr–Coulomb parameters from the nonlinear
Hoek–Brown criterion to implement safety factor calculation.

Table 1. Review of the previous reduction methods based on Hoek–Brown criterion.

Category Description Advantage Disadvantage

Simultaneous reducing all or part
Direct reduction method
of the Hoek–Brown parameters.
Overall reduction of the
Strength envelope lowering
Hoek–Brown envelope by a factor.
Reduction of equivalent
Mohr–Coulomb parameters or
Indirect reduction method
instantaneous equivalent
Mohr–Coulomb parameters.
Simple and easy to use
Relatively accurate, and satisfying
strength reduction concept
Compatible with Mohr–Coulomb
parameter-based criteria
or software
Insufficient theoretical
basis for
parameter selection.
Inefficient calculations
Large errors in global
equivalent parameters
and complex calculation
of instantaneous
equivalent parameters

Each of these studies has contributed to the advancement of strength reduction meth-
ods based on the Hoek–Brown criterion. However, the above studies all use the same
coefficients for the Hoek–Brown parameter reduction, which conflicts with the experimen-
tal evidence. To account for the degradation of the nonlinear strength parameter of the rock
mass, this paper proposes a new nonlinear strength reduction method for slope stability
evaluation. The selection of the Hoek–Brown parameter was achieved by establishing
an expression for the safety factor of the Hoek–Brown parameter. The respective reduc-
tion coefficients were then derived based on the critical strength curve of the slope and
the shortest path theory. Based on the horizontal displacement difference of the critical
slope, the k-means algorithm was used to search for potential slip surfaces in order to
Materials 2023, 16, 2793 4 of 16

calculate a more objective safety factor. As a result, a new nonlinear reduction method for
Hoek–Brown material slopes was established and confirmed by comparison with other
reduction methods.

2. Methods
2.1. Critical Strength Curve of Slope Satisfying Hoek–Brown Criterion
Slope stability is usually controlled by the geometrical parameters of the rock mass
(such as slope height H, slope angle θ) and physical parameters (unit weight γ), as well as all
parameters involved in the rock failure criterion. Taking the Hoek–Brown strength criterion
as an example, the parameters include σci , mb , s and a. All of these should satisfy a certain
functional relationship at the point of slope failure. According to the classical definition of
the slope safety factor, it is computed by the sliding resistance and the sliding motion on
the potential sliding surface. In another word, the slope safety factor is a function of the
anti-sliding force and the sliding force on the potential sliding surface (see Equation (6)).


Fs = f 1 τ, τg (6)

where τ is the anti-sliding stress, τ g is the sliding stress, and f 1 is an unknown function.
Generally, the sliding force on the potential sliding surface of the slope is provided by
the weight of sliding mass (Hγ), thereby the slope safety factor Fs can be expressed as a
function of slope height H, unit weight γ and slope angle θ (see Equation (7)).

Fs = f 2 (τ, θ, Hγ) (7)

The expression form of normal stress–shear stress (σn -τ) on the failure surface of the
Hoek–Brown strength criterion [68,69] is
 a
σci mb σσ3 + s
ci
σn = σ3 +   a −1 (8)
2 + amb mb σσ3 + s
ci

s
  a −1
σ3
τ = (σn − σ3 ) 1 + amb mb +s (9)
σci
It can be seen from Equation (9), τ is closely related to Hoek–Brown strength parameter
σci , mb , s and a, and normal stress σn . Thus, τ can be expressed as Equation (10).

τ = f 3 (σci , σn , mb , a, s) (10)

While the normal stress σn on the potential sliding surface of the slope is also provided
by the weight of sliding mass, which means

τ = f 4 (σci , mb , a, s, θ, Hγ) (11)

By substituting Equation (11) into the slope safety factor expression Equation (7), the
relationship between the safety factor Fs and slope parameters (see Equation (12)) can
be obtained.
Fs = f 5 (σci , mb , a, s, θ, Hγ) (12)
According to the expression of the Hoek–Brown strength criterion
(see Equation (2)), the uniaxial compressive strength of rock mass σcmass can be estimated
if σ3 = 0 (see Equation (13)). Therefore, the influences of parameters a and s on the slope
stability are incorporated into that of parameter σcmass .

σcmass = σci · s a (13)

 5 ( ci b ) (12)

According to the expression of the Hoek–Brown strength criterion (see Equation (2)),
the uniaxial compressive strength of rock mass σcmass can be estimated if σ3 = 0 (see Equa-
tion (13)). Therefore, the influences of parameters a and s on the slope stability are in-
Materials 2023, 16, 2793 corporated into that of parameter σcmass. 5 of 16

σcmass = σci ⋅ sa (13)

Combined with
Combined with Equation
Equation (12),
(12), the
the slope
slope safety
safety factor Fs can
factor Fs can be
be transformed
transformed into
into aa
function of the parameters σ cmass and m as well as the physical and geometric parameters
function of the parameters σcmass and mb as well as the physical and geometric parameters
b
of the
of the slope,
slope, as
as presented
presented in
in Equation
Equation (14).
(14).

= ff 6 ((σσcmass , ,mm
b ,θ , H γ )
Fs =
Fs b , θ, Hγ) (14)
(14)
6 cmass

For aagiven
For givenslope
slopeinin critical
critical state,
state, all all of the
of the above
above parameters
parameters shouldshould satisfy
satisfy the
the rela-
tionship of Fsof= Fs
relationship f6 (σ f 6 (σ
= cmass ,mcmass , mb , θ,= Hγ)
b, θ, Hγ) 1.0. = 1.0. Referring
Referring to the to the research
research resultsresults
by Yuanby et
Yuanal.
et al. [45], there are numerous combinations of ccritical and tan ϕcritical for a critical slope
[45], there are numerous combinations of c critical and tanφ critical for a critical slope when it
when it obeys the Mohr–Coulomb failure criterion [70–72], and the ccritical and tan ϕcritical
obeys the Mohr–Coulomb failure criterion [70–72], and the ccritical and tanφcritical obtained
obtained by traditional strength reduction is only one potential possibility. In the same way,
by traditional strength reduction is only one potential possibility. In the same way, for a
for a rock slope whose slope height H, weight γ and slope angle β have been determined,
rock slope whose slope height H, weight γ and slope angle β have been determined, there
there are also numerous critical combinations of σcmass and mb , which can also be plotted in
are also numerous critical combinations of σcmass and mb, which can also be plotted in the
the critical strength curve σcmass -mb . The critical strength curve of the rock slope is mainly
critical strength curve σcmass-m b. The critical strength curve of the rock slope is mainly de-
determined by numerical simulation of stability analysis on different slope models. Specific
termined by numerical simulation of stability analysis on different slope models. Specific
steps include:
steps include:
(1) A series of homogeneous slopes with different slope angles θ and weight γ were
(1) A series of homogeneous slopes with different slope angles θ and weight γ were
established by FLAC3D3D , and the geometric model is shown in Figure 2. The slope
established by FLAC , and the geometric model is shown in Figure 2. The slope
height H = 20 m remained unchanged, the variation range in slope angle was set
height◦ H =◦ 20 m remained unchanged, the variation range in slope angle was set to
to 30 –75 with a gradient of 15◦ , and the variation range of unit weight was set to
30°–75° with a3 gradient of 15°, and the variation range of unit weight was set to 20–
20–26 kN/m with a gradient of 2 kN/m3 .
26 kN/m3 with a gradient of 2 kN/m3.

Figure
Figure 2.
2. Geometric
Geometric model
model of
of homogeneous
homogeneous slope.
slope.

(2) Referring
Referringtotothe
thebuilt-in
built-inslope
slopeexample
exampleininFLAC
FLAC 3D , other parameters are presented
(2) 3D, other parameters are presented in
in Table
Table 2. 2.
Table 2. Parameters of built-in slope in Hoek–Brown material.

E/(MPa) ν mb s a σ ci /(MPa)
5000 0.3 0.067 0.000025 0.619 30

(3) Different values of mb were assigned to slope models. In this study, it was linearly
changed according to Equation (15).

mnew
b = mb × k (15)

where mb new is the value of the Hoek–Brown parameter mb for each slope model, k
represents the variation coefficient, and k = 1, 5, 10, 15, 20, . . . , 100.
(4) The left and right boundaries were normally constrained, and model bottom was
fixed in all directions. The initial stress field was generated by the gravity of the model
 mbnew = mb × k
where mbnew is the value of the Hoek–Brown parameter mb for each slope model, k
sents the variation coefficient, and k = 1, 5, 10, 15, 20, …, 100.
Materials 2023, 16, 2793
(4) The left and right boundaries were normally constrained, and6 of 16
model bottom
fixed in all directions. The initial stress field was generated by the gravity
model and the default convergence condition was used as the indicator of
failure. Thecondition
and the default convergence value of σ ci was
was continuously
used reduced
as the indicator during
of slope the numerical
failure. The simu
of stability analysis until the slope model reached
value of σci was continuously reduced during the numerical simulation of stabilitythe critical state for each
model.
analysis until the slope model reached the critical state for each slope model.
(5) of
(5) The critical value Theσcicritical
and the value of σci and theHoek–Brown
corresponding corresponding Hoek–Brown
parameters parameters
of the slope of the
were substituted towere substituted
Equation (13) totocalculate
Equation the(13) to calculate
specific value ofthe
thespecific
uniaxialvalue
com-of the u
pressive strength of compressive
rock mass σstrength
cmass .
of rock mass σcmass.
(6) Taking m b as the horizontal axis and σcmass as the vertical axis, the critical st
(6) Taking mb as the horizontal axis and σcmass as the vertical axis, the critical strength
curve of slope at curve of slope
any slope at any
angle can slope angle can
be drawn. be drawn.
Finally, Finally,expression
the general the general express
slope critical strength curve at given slope angle can
of slope critical strength curve at given slope angle can be obtained by nonlinear be obtained by nonlinea
data fitting. fitting.
The critical strength curves of each slope under different unit weight and slope
The critical strength curves
are shown of each3.slope under different unit weight and slope angle
in Figure
are shown in Figure 3.

0.07 0.07
θ = 30° θ = 30°
θ = 45° θ = 45°
0.06 0.06
θ = 60° θ = 60°
θ = 75° θ = 75°
0.05 0.05
σcmass(MPa)

σcmass(MPa)
0.04 0.04

0.03 0.03

0.02 0.02

0.01 0.01

0.00
Materials 2023, 16, x FOR PEER REVIEW 0.00
0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
mb mb
(a) (b)
0.08
0.08
0.07 θ = 30° θ = 30°
θ = 45° 0.07 θ = 45°
0.06 θ = 60° θ = 60°
θ = 75° 0.06 θ = 75°
σcmass(MPa)

0.05
σcmass(MPa)

0.05
0.04
0.04
0.03 0.03
0.02 0.02
0.01 0.01

0.00 0.00
0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
mb mb
(c) (d)

Figure 3. Critical Figure 3. Critical Hoek–Brown

curvesstrength curves slopes.
of different
(a)slopes. (a) γkN/m
= 20 3kN/m3; (b)
Hoek–Brown strength of different γ = 20 ;
γ = 24; (c)
(b) γ = 22 kN/m3 ; (c)kN/m γ = 24 kN/mγ =; 26
(d)kN/m
γ = 26 3kN/m .
3 3 3
kN/m 3 ; (d) .

For all the critical For all thecurves

strength critical in
strength
Figurecurves
3, theinvalue
Figureof 3,σcmass
the value
drops of with
σcmass the
drops with
crease in m b, and the dropping speed is decreasing. In principle, the slope stability
increase in mb , and the dropping speed is decreasing. In principle, the slope stability is
improved when theprovedvalue of when the value and
mb increases, of mthe
b increases, and the required value of σci is lowered at
required value of σci is lowered at slope
critical state. As a result, the value
critical state. As a result, the value of σcmass drops according of σcmass drops according
to Equation (13).toUnder
Equation the(13). Und
conditions of the same unit weight, greater strength is required to maintain slope stability slope st
conditions of the same unit weight, greater strength is required to maintain
at a greater slope angle. Therefore, the reduction factor of σci is less, and the critica
at a greater slope angle. Therefore, the reduction factor of σci is less, and the critical value
of σcmass is greater. Reflected in Figure 3, the larger the slope angle, the higher the
of σcmass is greater. Reflected in Figure 3, the larger the slope angle, the higher the slope
critical strength curve.
critical strength curve.
Since the slope critical strength curve is decreasing more and more slowly, i
sumed that all the curves in this study conform to the law of power function va
(see Equation (16)).

σcmass = AmbB
where both parameters A and B are undetermined coefficients, and can be determin
Materials 2023, 16, 2793 7 of 16

Since the slope critical strength curve is decreasing more and more slowly, it is as-
sumed that all the curves in this study conform to the law of power function variation
(see Equation (16)).
σcmass = AmbB (16)
where both parameters A and B are undetermined coefficients, and can be determined by
data fitting.
The fitting results and parameter values of each curve in Figure 3 are shown in Table 3,
and the variation laws of each parameter are shown in Figures 4 and 5. From the value of
R2 corresponding to each curve, the fitting curves show good matching with the critical
data, which explains the reliability of the fitting results to a certain extent. In addition, it
is obvious that the fitting values of slope critical curve parameters A and B are roughly
the same under different slope unit weight, with slight changes. The slope angle has a
stronger influence on the critical curve parameters, both of which increase with slope
inclination. Also, it can be seen from Figures 4 and 5 that the values of parameters A and B
are approximately linearly positively correlated with the slope angle.

Table 3. Fitting results of slope critical curves.

Unit Weight/kN/m3 Slope Angle/◦ A B R2

30 0.013 −0.374 0.995
45 0.019 −0.310 0.999
20
60 0.029 −0.267 0.993
75 0.039 −0.228 0.989
30 0.013 −0.373 0.996
45 0.020 −0.310 0.999
22
60 0.030 −0.275 0.996
75 0.040 −0.232 0.984
30 0.013 −0.377 0.995
45 0.020 −0.316 0.999
24
60 0.031 −0.279 0.995
75 0.041 −0.234 0.987
30 0.014 −0.381 0.995
Materials 2023, 16, x FOR PEER REVIEW 8 of 17
45 0.021 −0.325 0.999
26
60 0.031 −0.287 0.997
75 0.042 −0.239 0.988

0.045
20kN/m3
0.040 22kN/m3
23kN/m3
0.035 24kN/m3
parameter A

0.030

0.025

0.020

0.015

0.010
30 40 50 60 70 80
slope angle (°)

Figure 4. The
Figure4. The variation
variation in parameter A
in parameter A with
with slope
slope angle.
angle.

-0.22
20kN/m3
-0.24
22kN/m3
-0.26 23kN/m3
24kN/m3
-0.28
eter B

-0.30
 0.015

0.010
30 40 50 60 70 80
slope angle (°)
Materials 2023, 16, 2793 Figure 4. The variation in parameter A with slope angle. 8 of 16

-0.22
20kN/m3
-0.24
22kN/m3
-0.26 23kN/m3
24kN/m3
-0.28

parameter B
-0.30
-0.32
-0.34
-0.36
-0.38
-0.40
30 40 50 60 70 80
slope angle (°)

Figure 5. The variation in parameter B with slope angle.

2.2.
TableNonlinear
3. Fitting Reduction Method
results of slope critical curves.
As previously described, there are countless possibilities for a stable rock slope that
Unit Weight/kN/m
causes
3 Slope
the initial strength of rockAngle/°
mass to decay toAcritical strength. B Without considering
R2
other external factors, the faster the30 strength of natural
0.013rock slope−0.374 decays, the more 0.995
likely it
45 0.019 −0.310
is to fail. Therefore, based on the position relationship between the initial strength of rock 0.999
20
mass and the σcmass -mb critical curve60 in the coordinate 0.029system, the−0.267 corresponding0.993 reduction
principle of parameters σcmass and75 mb can be established
0.039 by assuming
−0.228 that σ cmass and
0.989mb are
reduced along the shortest path from 30 the critical curve
0.013 during slope
−0.373 strength attenuation.
0.996
That is the shortest reduction path theory.
45 0.020 −0.310 0.999
The distance
22 from the initial state of slope to the critical strength curve can be
60 0.030 −0.275 0.996
calculated as q
75 2 0.040 −0.232 2 0.984
minitial − mcritical



D= initial − σ critical
+ σcmass (17)
b b cmass
30 0.013 −0.377 0.995
45 and mb are known,
In the case that the initial σcmass 0.020 the critical values of σcmass
−0.316 and mb
0.999
24 by substituting Equation (16) into Equation (17) and minimizing D. Then,
can be obtained
60 0.031 −0.279 0.995
the reduction ratio ε of σcmass and mb can be solved by Equation (18).
75 0.041 −0.234 0.987
30 σinitial 0.014 −0.381 0.995
! !
mbinitial
cmass
ε =45 critical / 0.021critical −0.325 0.999 (18)
26 σcmass mb
60 0.031 −0.287 0.997
75
From Figure 3, the slope strength 0.042
critical curve is not unique,−0.239 0.988 ratio
so the reduction
ε is not consistent either for different slopes. One of the cores of the nonlinear strength
reduction method based on the Hoek–Brown criterion studied in this paper is just the
determination of the reduction ratio ε. By ensuring that the reduction coefficients of σcmass
and mb meet the ratio relation ε at each reduction step, the numerical calculation of rock
slope stability analysis conforms to the shortest path hypothesis.
As Equation (19) indicated, reducing σcmass by coefficient kσ can be realized by
reducing σci .
σcmass σ
= ci · s a (19)
kσ kσ
While according to Equation (20), the reduction coefficient kmb of parameter mb in each
iteration can be calculated (see Equation (22)).

kσ
k mb = (20)
ε

The reduction strategy of Equation (20) can be implemented in FLAC3D by program-

ming a custom rock slope strength reduction program in Fish language. Furthermore, it
is inevitable to consider the method of determining the safety factor of slope because the
Materials 2023, 16, 2793 9 of 16

parameter reduction is not synchronous. The classic definition in Equation (6) points out
that the safety factor of slope Fs can be evaluated by the ratio of sliding resistance to sliding
movement on the potential sliding surface, while the sliding resistance approximates to
sliding movement at slope failure, which signifies that Fs is also be expressed by the ratio
of initial sliding resistance τ initial to critical sliding resistance τ critical (see Equation (21)).
R initial
τ dl
Fs = R l critical (21)
Materials 2023, 16, x FOR PEER REVIEW
lτ dl 10 of 17

The assessment of Equation (21) strongly relies on the positioning of the potential
sliding surface, which can be located through various methods, such as displacement
contour, maximum shear strain and maximum shear strain increment [73,74]. In this
contour, maximum shear strain and maximum shear strain increment [73,74]. In this study,
study, the horizontal displacement of the node of the slope model was adopted as the
the horizontal displacement of the node of the slope model was adopted as the index to
index to judge the position of slope potential sliding surface. Once the potential sliding
judge the position of slope potential sliding surface. Once the potential sliding surface is
surface is
located, thelocated, the information
coordinate coordinate information
of each nodeof oneach node on
the sliding the sliding
surface can be surface canand
extracted, be
extracted, and the corresponding minor principal stress can be output.
the corresponding minor principal stress can be output. Then, the antisliding force on the Then, the an-
tisliding force
potential slidingonsurface
the potential sliding surface
can be calculated can be calculated
by Equations (8) and (9), byand
Equations (8)factor
the safety and (9),
of
and the safety factor of slope can be solved by Equation (21). The specific
slope can be solved by Equation (21). The specific steps are as follows: (1) traversing all grid steps are as
follows:
nodes of (1)
the traversing
slope model allafter
gridstability
nodes of the slopeto
calculation model
deriveafter stability calculation
the coordinates to de-
and horizontal
rive the coordinates and horizontal displacements of each node; (2)
displacements of each node; (2) importing the position and displacement information of importing the posi-
the
tion and displacement information of the derived nodes into MATLAB
derived nodes into MATLAB computing software, and processing by k-means clustering computing soft-
ware, andtoprocessing
algorithm find those by
nodesk-means clustering
consisting algorithm
of potential slidingto surfaces;
find those and nodes consistingthe
(3) positioning of
potential on
elements sliding surfaces;sliding
the potential and (3)surface
positioning the the
to derive elements on the potential
corresponding minimum sliding sur-
principal
face toinformation,
stress derive the corresponding minimum
which is successively principal
substituted stress
into information,
Equations (8), (9) which
and (21)isto
succes-
solve
sively
the substituted
slope into Equations (8), (9) and (21) to solve the slope safety factor.
safety factor.

3. Results and Discussion

To
To verify
verify the
the correctness
correctness of of the
the Hoek–Brown
Hoek–Brown criterion-based
criterion-based nonlinear
nonlinear reduction
reduction
method proposed in this paper,
paper, itit is
is necessary
necessary to employ the slope example whose safety
factor has been
been confirmed
confirmed for
for stability
stability analysis.
analysis. The
The built-in
built-in example
example in FLAC3D
inFLAC 3D, simple

slope in Hoek–Brown
Hoek–Brown material,
material,was wasselected
selectedand
andisisshown
shownininFigure
Figure6.6. The uniformity
The uniformity of
of calculated
calculated safety
safety factors
factors of slope
of slope exampleexample by various
by various methods methods
in Tablein4 Table 4 explains
explains that the
that the calculation
calculation result byresult by Hoek–Brown
Hoek–Brown local linearization
local linearization is relatively
is relatively correct,
correct, which which
was
was therefore
therefore employed
employed as indicator
as the the indicator to evaluate
to evaluate the efficacy
the efficacy of proposed
of proposed method
method in
in this
this study.
study.

Hoek–Brown material.
Figure 6. The simple slope model in Hoek–Brown material.
 Materials 2023, 16, 2793 10 of 16

Table 4. Factors of safety results for Hoek–Brown slope.

Resource Method Safety Factor

FLAC3D Hoek–Brown local linearization 1.15
Hoek–Brown envelop lowering 1.15
Equivalent Mohr–Coulomb reduction 1.15
Hammah et al. [64]
Bishop simplified limit equilibrium 1.153
Spencer limit equilibrium 1.152

The same slope model was established by FLAC3D , and corresponding constitutive
model and parameters were assigned. The slope strength was adjusted to reach the
Materials 2023, 16, x FOR PEER REVIEW
critical state so as to obtain the critical strength curve and its expression 11 of 17
of this slope
example. There are two methods to determine the critical strength curve expression of
slope. One is to repeat the above operations to obtain a large number of data points
Table 4.about the
Factors critical
of safety parameters
results and mb , and to fit the critical strength curve expression
σcmass slope.
for Hoek–Brown
of the model according to Equation (16). Secondly, the values of parameters A and B
Resource
can also be theoretically obtained Method by solving the equations with Safety Factor
only two sets of critical
FLAC
parameters
3D
σcmass and mHoek–Brown
, local linearization
approximately determining the 1.15
critical curve expression. The
b
Hoek–Brown envelop lowering 1.15
latter is simple and less time-consuming, but curve accuracy cannot be guaranteed. For the
sake of
Hammah et precision,
al. [64] theEquivalent
first method Mohr–Coulomb
was used toreduction
obtain multiple sets1.15of σ
cmass and mb , and
Bishop simplified limit equilibrium 1.153
the expression was fitted, as shown in Figure 7. Note that, the unit of σcmass is set as kPa to
Spencer limit equilibrium 1.152
avoid gaining values too small for the parameters A and B by fitting.
The initial state of slope example (σcmass , mb ) is (42.5056, 0.067). According to
The same slope model was established by FLAC3D , and corresponding constitutive
Equation (17), the distance from the initial state to the critical curve is
model and parameters were assigned. The slope strength was adjusted to reach the crit-
ical state so as to obtain the r critical strength curve andits expression of this slope exam-
ple. There are two

2 strength curve expression critical
−0.283 2
Dexample = to determine
methods 0.067 − mtheb
critical
critical + 42.5056 − 17.339 × m b
of slope. (22)
One is to repeat the above operations to obtain a large number of data points about the
critical parameters σcmass and mb, and to fit the critical strength curve expression of the
The calculation results show that during the range of mb (0, 25), the minimum Dexample
model according to Equation (16). Secondly, the values of parameters A and B can also be
theoretically obtainedto
approximates by0.0267
solvingwhen mb critical is
the equations 0.0421,
with only and the corresponding
two sets of critical parametersvalue of σcmass critical
is 42.4964
σcmass and kPa. By determining
mb, approximately substituting thethe critical
critical curvevalues into The
expression. Equation
latter is(18),
simpleε = 0.6285 were
obtained, which indicates that when
and less time-consuming, but curve accuracy cannot σ and m are
ci be guaranteed.
b reduced in each iterative
For the sake of pre- calculation
cision,oftheslope stability numerical analysis, the reduction ratio remains 0.6285. Afterwards, the
first method was used to obtain multiple sets of σcmass and m b, and the expres-
sion wasFLAC 3D as
fitted, shown inreduction
nonlinear Figure 7. Note that, the
algorithm wasunitcustomized
of σcmass is setbyasFish
kPa to avoid to solve the
language
gaining values too small for the parameters A and B by fitting.
slope stability according to this reduction ratio, and the horizontal displacement is shown
in Figure 8.

40

35 σ cmass = 17.339mb−0.283
R 2 = 0.999
30
σcmass (kPa)

25

20

15

10

5
0 5 10 15 20
mb

Figure 7. The critical points of σcmass and mb for slope example.

Figure 7. The critical points of σcmass and mb for slope example.

The initial state of slope example (σcmass, mb) is (42.5056, 0.067). According to Equation
(17), the distance from the initial state to the critical curve is

-0.283 2
( 0.067 − mbcritical ) + 42.5056 −17.339 × ( mbcritical )
2
Dexample = (22)
 tained, which indicates that when σci and mb are reduced in each iterative calculat
slope stability numerical analysis, the reduction ratio remains 0.6285. Afterward
tained, which FLAC
indicates that whenreduction
3D nonlinear σci and malgorithm
b are reduced
wasincustomized
each iterative
by calculation
Fish language of to solv
slope stabilityslope
numerical analysis,
stability thetoreduction
according ratio ratio,
this reduction remainsand0.6285. Afterwards,
the horizontal the
displacement is s
FLAC3D nonlinear reduction
in Figure 8. algorithm was customized by Fish language to solve the
Materials 2023, 16, 2793 11 of 16
slope stability according to this reduction ratio, and the horizontal displacement is shown
in Figure 8.

Figure 8. The stability calculation result and the horizontal displacement.

Figure 8. The stabilityThe

calculation
coordinates and the
result and displacement displacement.
data of each node in Figure 8 were derive
horizontal displacement.
then imported into the MATLB calculation program. Then, the k-means clustering
The coordinates
coordinates
rithm wasand
anddisplacement
displacement
used to classifydata
dataof of
these eacheach
nodesnodenode in Figure
in Figure
according 8 were
to derived
8 were
the and
derived
horizontal then
and
displacement d
imported
then importedinto ences,
the
intoMATLB
the calculation
MATLB program.
calculation program.Then, the
Then, k-means
the clustering
k-means
and corresponding nodes of sliding mass and stable mass were obtained, as s algorithm
clustering algo-
was
rithmused
wasto classify
usedin to these
classify
Figure nodes
these
9. The according
nodes
nodes tomass
according
of sliding thetohorizontal
the horizontal
were displacement
isolated analysis,differences,
fordisplacement anddiffer-
the rightmost
and corresponding
ences, and corresponding nodes
were considered of sliding
nodesthe mass
of sliding
nodes on and
mass stable
theand mass
stablesliding
potential were
mass were obtained,
obtained,
surface as shown
(dark as shown
blue in in Figu
nodes
Figure 9. The
in Figure 9. The nodes
Thenodes of sliding
of sliding
potential mass were
masssurface
sliding isolated
were isolated for
of slopefor analysis,
can analysis, and the rightmost
and thebyrightmost
be obtained smoothing nodes
nodesthese node
were
were considered
considered the
the nodes
nodes on
on the
the potential
potential sliding surface
sliding surface (dark(dark blue
blue nodes
nodes in in Figure
Figure 9). 9).
cording to the node coordinate information, the minimum principal stress of the
The potential
The potential wassliding
sliding surface
surfaceof slope
of the can
slope be obtained
can3Dbeslope
obtainedby smoothing
byand these
smoothing nodes. According
these nodes. Ac- of the
derived from FLAC model, the calculated safety factor
to the node coordinate
cording to theexample information,
node coordinate the minimum
information, the principal stress of the node was derived
is 1.314. Compared
3D slope model, with theminimum
safety factor principal stress
calculated byofother
the methods
node in Ta
from the FLAC
was derived from the FLAC 3D and
slope themodel,
calculated safety
andproves factor of the
the calculated slope
safety example
factor is slope
ofmethod
the 1.314.
the error is about 14.26%, which the correctness of this to some e
Compared
example is with
1.314. the safety factor calculated by other methodsby in Table 3, the error in is about
TheCompared
reason forwith the the safety
14.26% factor
error can calculated
be attributed other methods
to the fact that Table 3,
the potential s
14.26%, which
the error is about proves the
14.26%, correctness
which of
proves this
the method
correctness to some
of extent.
this method The toreason
some for the
extent.
surface is determined by node displacement, while the safety factor must be solv
14.26% errorfor
The reason canthebe 14.26%
attributed to the fact that the potential sliding surface is determined
cording to theerror
nearbycanstress-state
be attributed to the That
elements. fact that
is tothesay,potential slidingsafety fac
the calculated
by node displacement, while the safety factor must be solved according to the nearby
surface is determined
the slopeby node displacement,
example is susceptiblewhile to thethe gridsafety
densityfactor must
of the be solved
model. ac-
The shear–stress
stress-state elements. That is to say, the calculated safety factor of the slope example is
cording to thelation
nearbyofstress-state
nodes mayelements.
make theThat safetyis to say, the
factor calculated
by this methodsafety morefactor of howe
accurate;
susceptible to the grid density of the model. The shear–stress calculation of nodes may
the slope example seems is to
susceptible
be difficult totothe grid density
realize in finite of the model.
element The shear–stress calcu-
software.
make the safety factor by this method more accurate; however, it seems to be difficult to
lation of nodes may make the safety factor by this method more accurate; however, it
realize in finite element software.
seems to be difficult to realize in finite element software.
20
stable body
18 sliding body
20 sliding surface
stable body
16
18 sliding body
14
sliding surface
16
12
14
10
z

12
8
10
z

6
8
4
6
2
4
0
2 0 5 10 15 20 25 30 35 40

0 x
0 5 10 15 20 25 30 35 40
Figure 9. The clustering results and nodes on the potential sliding surface.
x
Figure 9. The clustering results and nodes on the potential sliding surface.
Similarly, the slope examples in [75] (Case 1) and [65] (Case 2) were selected to further
demonstrate the feasibility
Figure 9. The clustering and
results applicability
and nodes on theofpotential
the proposed
slidingmethod.
surface. The slope parameters
are presented in Table 5, and the same steps were implemented for these slope models, the
results of which are presented in Table 6.
Materials 2023, 16, 2793 12 of 16

Materials 2023, 16, x FOR PEER REVIEW 13 of 17

Materials 2023, 16, x FOR PEER REVIEW 13 of 17

Table 5. The parameters of other reported slope cases.

Similarly, the slope examples in [75] (Case 1) and [65] (Case 2) were selected to fur-
Similarly, the slope examples in [75] (Case 1) and [65] (Case 2) were selected to fur-
ther demonstrate
ther demonstrate the
the feasibility
feasibility and
and applicability
applicability of
of the
the proposed
proposed method.
method. The
The slope
slope pa-
pa-
Slope Case H/m θ/◦ γ/kN/m3 E/MPa rameters are presented
rameters are
models,the
ν presented in
theresults
resultsof
in Table
σ
ofwhich
Table 5,
/MPa 5, and
and the
the same
same steps
m
ci are presented in Table
which
steps were
b 6.
were implemented
implemented for
s for these
these slope
slope
a
models, are presented in Table 6.
Case 1 [75] 32 75 25 5000 0.3 40
Table 5. The parameters of other reported slope cases.
Table 5. The parameters of other reported slope cases.
0.281 2 × 10−4 0.508
Case 2 [65] 50 60 SlopeCase
23Slope Case 5000
H/m
H/m θ/°
θ/° γ/kN/m
0.3
γ/kN/m 3
3 E/MPa
10
E/MPa νν 0.657
σσcici/MPa
/MPa mbb 10−4ss
4 m× aa
0.522
Case11[75]
Case [75] 32
32 75
75 25
25 5000
5000 0.3
0.3 40
40 0.281
0.281 22××10
10−4−4 0.508
0.508
Case22[65]
Case [65] 50
50 60
60 23
23 5000
5000 0.3
0.3 10
10 0.657
0.657 4 × 10−4−4
4 × 10 0.522
0.522
Table 6. The calculation results of other
Tableslope cases.results of other slope cases.
6. The calculation
Table 6. The calculation results of other slope cases.
Expressionof
Expression ofCritical
Critical
Case Ratioof
ofReduction
ReductionCoefficients
Coefficients Potential Sliding Surface
Case Expression of Critical Curve Case Ratio Curve
of Reduction
Curve
Ratio
Coefficients Potential Sliding Surface
Potential Sliding Surface
80
80 sliding body
70 sliding
stable body
body
70 stable body
sliding surface
60 sliding surface
60
50
50
−0.096

σσcmass
40
11[75]
[75] = 0.303m−0.096 1.399
σcmass = 0.303m− 0.096 40
cmass = 0.303m b
1 [75] b
b 1.399 1.399 30
30
20
20
10
10
0
0
0 20 40 60 80 100 120 140
0 20 40 60 80 100 120 140

120
120 stable body
stable
slidingbody
body
100 sliding
slidingbody
surface
100 sliding surface
80
80

σcmass = 0.139m− 0.213 −0.213

σσcmass
60
2 [65] 22[65]
[65] = 0.139m−0.213
cmass = 0.139m b 1.104 1.104
1.104
60
b b 40
40
20
20
0
0
0 40 80 120 160 200
0 40 80 120 160 200

Meanwhile, except
Meanwhile, except for
for Hoek–Brown
Hoek–Brown local
local linearization
linearization reduction,
reduction, the
the other
other Hoek–
Hoek–
Brown nonlinear strength reduction methods, such as direct reduction of the involved
Meanwhile, except for Hoek–Brown Brown nonlinearlocal
parameterslinearization
Hoek–Brown parameters
strength reduction methods, such
(σ and
and m m ),), equivalent
equivalent
ci
reduction,
Mohr–Coulombthe
b
as direct reduction of other
parameter Hoek–
the involved
reduction
Hoek–Brown (σ ci bMohr–Coulomb parameter reduction
Brown nonlinear strength reduction was methods,
was utilized
utilized to
Brown-based strength
calculatesuch
to calculate the
strength nonlinear
as
the safety
safety
nonlinear reduction
direct
factor
factor for
reduction method,
reduction
for proving
proving
method, and
the
and the
of
the superiority
superiority
the calculated
the
of
calculated safety
involved
of this
this Hoek–
Hoek–
safety factors
factors are
are
Hoek–Brown parameters (σci andBrown-based mb ), equivalent
presented
presented in Table
in Table 7. 7. Among
Among Mohr–Coulomb
these, the
these, the safety factorsparameter
safety factors calculated by
calculated thereduction
by the proposedmethod
proposed method was
are the closest to those by Hoek–Brown local linearization. The errors for slope cases are
utilized to calculate the safety factor are thefor proving
only8.22%
closest
8.22%
to
and8.18%,
those bythe superiority of this Hoek–Brown-based
Hoek–Brown
8.18%,respectively.
respectively.
local linearization. The errors for slope cases are
only and
strength nonlinear reduction method, and the calculated safety factors are presented in
Table 7. The comparison between factor of safety by different methods.
Table 7. The comparison between factor of safety by different methods.
Table 7. Among these, the safety factors calculated Factorof
by the
ofSafety
proposed
SafetyCalculation
Calculation
method are the closest
Factor
to those by Hoek–Brown
Method
Method local linearization.
Hoek–Brown Local
Hoek–Brown Local Direct The
Reduction
Direct Reduction of
errors
of for slope
Equivalent cases
Mohr–Coulomb
Equivalent Mohr–Coulomb
are only 8.22%
ProposedMethod
Method
and
Linearization Proposed
8.18%, respectively. Linearization σσ and
andm m ci
ci b
b Reduction
Reduction
Case 1 [75] 1.5501
Case 1 [75] 1.5501 1.9659 1.9659 1.9201 1.9201 1.6775 1.6775
Case22[65]
Case [65] 1.0090
1.0090 0.9900
0.9900 1.5070
1.5070 1.0915
1.0915
Table 7. The comparison between factor of safety by different methods.

Factor of Safety Calculation

Method Hoek–Brown Local Direct Reduction of Equivalent Mohr–Coulomb Proposed
Linearization σ ci and mb Reduction Method
Case 1 [75] 1.5501 1.9659 1.9201 1.6775
Case 2 [65] 1.0090 0.9900 1.5070 1.0915

4. Conclusions
In response to the previous strength reduction methods on the Hoek–Brown criterion,
which all use the same reduction factor for the parameters, this paper deduced a new re-
duction method that can reflect the nonlinear deterioration of the Hoek–Brown parameters
during slope instability. The main findings of this paper are as follows.
(1) Critical strength curves for slopes show that uniaxial compressive strength of rock
mass σcmass decreases with the increase in the Hoek–Brown parameter mb under
different unit weight or slope angle conditions. On this basis, the general expression
of slope critical strength curve was fitted by power function. Combined with the
shortest reduction path theory, the ratio of the reduction coefficients of σcmass and
mb can be determined.
(2) There are significant differences in the critical horizontal displacements between
sliding and stable nodes of slope model. The k-means clustering algorithm was used
to separate the sliding nodes and stable nodes according to such differences, so as
Materials 2023, 16, 2793 13 of 16

to automatically identify the potential sliding surface. Then, the ratio of the sliding
resistance to the sliding force was solved.
(3) Based on the proposed nonlinear reduction method and other known methods, sta-
bility calculations and safety factor comparisons were carried out on slope examples.
Compared to other methods, the safety factors calculated by the proposed method
differ less from the reference safety factor, which justifies the correctness and feasibility
of this method to a certain extent.

Author Contributions: Conceptualization, Y.C. (Yifan Chen); Methodology, Y.C. (Yifan Chen),
H.L. and H.H.; Software, Y.C. (Yizhou Chen); Validation, Y.C. (Yifan Chen); Formal analysis,
Y.C. (Yifan Chen), H.L. and H.H.; Investigation, Y.C. (Yizhou Chen); Writing—original draft,
Y.C. (Yifan Chen), Y.C. (Yizhou Chen), H.L. and H.H.; Writing—review & editing, H.L. and H.H.; Fund-
ing acquisition, H.L. All authors have read and agreed to the published version of the manuscript.
Funding: This paper got its funding from project NRMSSHR-2022-Z08, supported by the Key
Laboratory of Natural Resources Monitoring and Supervision in Southern Hilly Region, Ministry
of Natural Resources, Hunan Provincial Key Research and Development Program (2022SK2082),
projects 42277175 and 52104110 supported by the National Natural Science Foundation of China,
and the Science and Technology Progress and Innovation Plan of Hunan Provincial Department of
Transportation (201003 and 202120). The authors wish to acknowledge this support.
Data Availability Statement: The data used to support the findings of this study are available from
the corresponding author upon request.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Alcantara-Ayala, I.; Garnica-Pena, R.J.; Dominguez-Morales, L.; Gonzalez-Huesca, A.E.; Calderon-Vega, A. The La Pintada
landslide, Guerrero, Mexico: Hints from the Pre-Classic to the disasters of modern times. Landslides 2017, 14, 1195–1205. [CrossRef]
2. Loi, D.H.; Quang, L.H.; Sassa, K.; Takara, K.; Dang, K.; Thanh, N.K.; Tien, P.V. The 28 July 2015 rapid landslide at Ha Long City,
Quang Ninh, Vietnam. Landslides 2017, 14, 1207–1215. [CrossRef]
3. Wu, G.; Cunningham, D.; Yuan, R.M.; Zhou, Q.; Zeng, X.F.; Yang, X.P. Mass-wasting effects induced by the 2015 Gorkha (Nepal)
M-w 7.8 earthquake within a large paleo-landslide site adjacent to the Tatopani Border Station, Nepal: Implications for future
development along the critical Bhote Koshi River valley transport corridor between Nepal and China. Landslides 2017, 14,
1147–1160. [CrossRef]
4. Akgun, H.; Kockar, M.K. Design of anchorage and assessment of the stability of openings in silty, sandy limestone: A case study
in Turkey. Int. J. Rock Mech. Min. Sci. 2004, 41, 37–49. [CrossRef]
5. Liu, J.; Zhao, Y.; Tan, T.; Zhang, L.; Zhu, S.; Xu, F. Evolution and modeling of mine water inflow and hazard characteristics in
southern coalfields of China: A case of Meitanba mine. Int. J. Min. Sci. Technol. 2022. [CrossRef]
6. Xie, Q.; Ren, B.; Hursthouse, A.S.; Shi, X. Effects of mining activities on the distribution, controlling factors, and sources of metals
in soils from the Xikuangshan South Mine, Hunan Province. Integr. Environ. Assess. Manag. 2021, 18, 748–756. [CrossRef]
7. Feng, T.; Chen, H.; Wang, K.; Nie, Y.; Zhang, X.; Mo, H. Assessment of underground soil loss via the tapering grikes on limestone
hillslopes. Agric. Ecosyst. Environ. 2020, 297, 106935. [CrossRef]
8. Kockar, M.K.; Akgun, H. Engineering geological investigations along the Iliksu Tunnels, Alanya, southern turkey. Eng. Geol. 2003,
68, 141–158. [CrossRef]
9. Fan, H.; Lu, Y.; Hu, Y.; Fang, J.; Lv, C.; Xu, C.; Feng, X.; Liu, Y. A Landslide Susceptibility Evaluation of Highway Disasters Based
on the Frequency Ratio Coupling Model. Sustainability 2022, 14, 7740. [CrossRef]
10. Chen, Y.; Lin, H.; Xie, S.; Cao, R.; Sun, S.; Zha, W.; Wang, Y.; Zhao, Y.; Hu, H. Fracture Closure Empirical Model and Theoretical
Damage Model of Rock under Compression. Materials. 2023, 16, 589. [CrossRef]
11. Ahmadi, S.; Kamalian, M.; Askari, F. Evaluation of the Static Bearing Capacity Coefficients of Rough Strip Footing Using the
Stress Characteristics Method. Int. J. Civ. Eng. 2021, 19, 155–165. [CrossRef]
12. Li, C.C.; Jiang, P.M. Method of Rigorous Characteristics for Seismic Bearing Capacity of Strip Footings Placed Adjacent to
Homogeneous Soil Slopes. Int. J. Geomech. 2022, 22, 04022179. [CrossRef]
13. Fang, H.W.; Xu, Y.X. New Failure Mechanism for Evaluating the Inclined Failure Load Adjacent to Undrained Soil Slope.
Sustainability 2023, 15, 2159. [CrossRef]
14. Ahmadi, S.; Kamalian, M.; Askari, F. Considerations on Bearing Capacity Factors of Rough Strip Footing Using the Stress
Characteristics Method. Iran. J. Sci. Technol. Trans. Civ. Eng. 2021, 45, 2611–2621. [CrossRef]
15. Ahmadi, S.; Kamalian, M.; Askari, F. Evaluating the N gamma coefficient for rough strip footing located adjacent to the slope
using the stress characteristics method. Comput. Geotech. 2022, 142, 104543. [CrossRef]
Materials 2023, 16, 2793 14 of 16

16. Boubazine, L.; Boumazbeur, A.; Hadji, R.; Fares, K. Slope failure characterization: A joint multi-geophysical and geotechnical
analysis, case study of Babor Mountains range, NE Algeria. Min. Miner. Depos. 2022, 16, 65–70. [CrossRef]
17. Sdvyzhkova, O.; Moldabayev, S.; Bascetin, A.; Babets, D.; Kuldeyev, E.; Sultanbekova, Z.; Amankulov, M.; Issakov, B. Probabilistic
assessment of slope stability at ore mining with steep layers in deep open pits. Min. Miner. Depos. 2022, 16, 11–18. [CrossRef]
18. Avsar, O.; Akgun, H.; Kockar, M.K. Investigation of the failure mechanism and stabilization of a landslide in weathered tuffite,
Giresun, northeastern Turkey. Environ. Earth Sci. 2014, 72, 3723–3740. [CrossRef]
19. Duncan, J.M. State of the Art: Limit Equilibrium and Finite-Element Analysis of Slopes. J. Geotech. Eng. 1996, 123, 577–596.
[CrossRef]
20. Griffiths, D.V.; Lane, P.A. Slope stability analysis by finite elements. Geotechnique 1999, 49, 387–403. [CrossRef]
21. Sun, C.W.; Chai, J.R.; Ma, B.; Luo, T.; Gao, Y.; Qiu, H.F. Stability Charts for Pseudostatic Stability Analysis of 3D Homogeneous
Soil Slopes Using Strength Reduction Finite Element Method. Adv. Civ. Eng. 2019, 2019, 6025698. [CrossRef]
22. Tang, G.P.; Zhao, L.H.; Li, L.; Yang, F. Stability charts of slopes under typical conditions developed by upper bound limit analysis.
Comput. Geotech. 2015, 65, 233–240. [CrossRef]
23. Zhang, J.; Li, J. A comparative study between infinite slope model and Bishop’s method for the shallow slope stability evaluation.
Eur. J. Environ. Civ. Eng. 2019, 1–18. [CrossRef]
24. Kockar, M.K.; Akgun, H. Methodology for tunnel and portal support design in mixed limestone, schist and phyllite conditions: A
case study in Turkey. Int. J. Rock Mech. Min. Sci. 2003, 40, 173–196. [CrossRef]
25. Azarafza, M.; Akgun, H.; Ghazifard, A.; Asghari-Kaljahi, E.; Rahnamarad, J.; Derakhshani, R. Discontinuous rock slope stability
analysis by limit equilibrium approaches—A review. Int. J. Digit. Earth 2021, 14, 1918–1941. [CrossRef]
26. Matsui, T.; San, K.C. Finite element slope stability analysis by shear strength reduction technique. Soils Found. 1992, 32, 59–70.
[CrossRef]
27. Renani, H.R.; Martin, C.D. Factor of safety of strain-softening slopes. J. Rock Mech. Geotech. Eng. 2020, 12, 473–483. [CrossRef]
28. Shen, J.; Karakus, M. Three-Dimensional Numerical Analysis for Rock Slope Stability Using Shear Strength Reduction Method.
Can. Geotech. J. 2014, 51, 164–172. [CrossRef]
29. Bai, B.; Yuan, W.; Li, X.C. A new double reduction method for slope stability analysis. J. Cent. South Univ. 2014, 21, 1158–1164.
[CrossRef]
30. Zhu, C.; Long, S.; Zhang, J.; Wu, W.; Zhang, L. Time Series Multi-Sensors of Interferometry Synthetic Aperture Radar for
Monitoring Ground Deformation. Front. Environ. Sci. 2022, 10, 929958. [CrossRef]
31. Zheng, C.; Jiang, B.; Xue, S.; Chen, Z.; Li, H. Coalbed methane emissions and drainage methods in underground mining for
mining safety and environmental benefits: A review. Process Saf. Environ. Prot. 2019, 127, 103–124. [CrossRef]
32. Huang, Z.; Zhang, L.; Yang, Z.; Zhang, J.; Gao, Y.; Zhang, Y. Preparation and properties of a rock dust suppressant for a copper
mine. Atmos. Pollut. Res. 2019, 10, 2010–2017. [CrossRef]
33. Chen, Y.; Lin, H.; Wang, Y.; Cao, R.; Zhang, C.; Zhao, Y. Modified Double-Reduction Method considering Strain Softening and
Equivalent Influence Angle. KSCE J. Civ. Eng. 2020, 24, 3257–3266. [CrossRef]
34. Chen, Y.; Lin, H.; Cao, R.; Zhang, C. Slope Stability Analysis Considering Different Contributions of Shear Strength Parameters.
Int. J. Geomech. 2020, 21, 1–9. [CrossRef]
35. Zhang, K.; Sun, B.; Li, W.; Nie, M. Study on the failure mechanism of expansive soil slope and limit equilibrium analysis method
based on fully softening strength. J. Cent. South Univ. (Sci. Technol.) 2022, 53, 75–94. [CrossRef]
36. Hu, H.; Zhang, Q.; Zhang, P. Research of improved G-B model for toppling stability analysis of plate-shaped steep slope
foundation. J. Cent. South Univ. (Sci. Technol.) 2022, 53, 4123–4131. [CrossRef]
37. Liu, X.; Tu, Y.; Zhong, Z.; Liu, Y. Slope’s failure criterion based on energy catastrophe in shear strength reduction method. J. Cent.
South Univ. (Sci. Technol.) 2016, 47, 2065–2072. [CrossRef]
38. Tang, F.; Zheng, Y. Mechanism analysis on dual reduction factors about the progressive failure of slope. Chin. J. Undergr. Space
Eng. 2008, 4, 436–441.
39. Tang, F.; Zheng, Y.; Zhao, S. Discussion on two safety factors for progressive failure of soil slope. Chin. J. Rock Mech. Eng. 2007, 26,
1402–1407. [CrossRef]
40. Pantelidis, L.; Griffiths, D.V. Stability Assessment of Slopes Using Different Factoring Strategies. J. Geotech. Geoenviron. Eng. 2012,
138, 1158–1160. [CrossRef]
41. Xue, H.; Dang, F.; Yin, X.; Yang, C.; Yan, F. Research on method of slope strength parameters non-proportional associated
reduction. Chin. J. Rock Mech. Eng. 2015, 34, 4005–4012. [CrossRef]
42. Jiang, X.; Wang, Z.; Liu, L.; Zhou, Z. The Determination of Reduction Ratio Factor in Homogeneous Soil-Slope with Finite Element
Double Strength Reduction Method. Open Civ. Eng. J. 2013, 7, 205–209. [CrossRef]
43. Yuan, W.; Bai, B.; Li, X.; Wang, H. A strength reduction method based on double-reduction parameters and its application. J. Cent.
South Univ. 2013, 20, 2555–2562. [CrossRef]
44. Isakov, A.; Moryachkov, Y. Estimation of slope stability using two-parameter criterion of stability. Int. J. Geomech. 2014, 14,
613–624. [CrossRef]
45. Yuan, W.; Li, X.C.; Wang, W.; Bai, B.; Wang, Q.Z.; Chen, X.J. Study on strength reduction method based on double reduction
parameters. Rock Soil Mech. 2016, 37, 2222–2230.
Materials 2023, 16, 2793 15 of 16

46. Fang, H.W.; Chen, Y.F.; Xu, G.W.; Hou, Z.K.; Wu, J.X. New Instability Criterion for Stability Analysis of Homogeneous Slopes
with Double Strength Reduction. Int. J. Geomech. 2020, 20, 1–11. [CrossRef]
47. Zhao, Y.; Luo, S.; Wang, Y.; Wang, W.; Zhang, L.; Wan, W. Numerical Analysis of Karst Water Inrush and a Criterion for
Establishing the Width of Water-resistant Rock Pillars. Mine Water Environ. 2017, 36, 508–519. [CrossRef]
48. Yuan, Z.; Zhao, J.; Li, S.; Jiang, Z.; Huang, F. A Unified Solution for Surrounding Rock of Roadway Considering Seepage, Dilatancy,
Strain-Softening and Intermediate Principal Stress. Sustainability 2022, 14, 8099. [CrossRef]
49. Liu, S.; Nie, Y.; Hu, W.; Ashiru, M.; Li, Z.; Zuo, J. The Influence of Mixing Degree between Coarse and Fine Particles on the
Strength of Offshore and Coast Foundations. Sustainability 2022, 14, 9177. [CrossRef]
50. Zhao, Y.; Tang, J.; Chen, Y.; Zhang, L.; Wang, W.; Wan, W.; Liao, J. Hydromechanical coupling tests for mechanical and permeability
characteristics of fractured limestone in complete stress-strain process. Environ. Earth Sci. 2017, 76, 24. [CrossRef]
51. Zhu, H.; Zhang, Q.; Zhang, L. Review of research progresses and applications of Hoek-Brown strength criterion. Chin. J. Rock
Mech. Eng. 2013, 32, 1945–1963.
52. Yu, W.; Li, K.; Liu, Z.; An, B.; Wang, P.; Wu, H. Mechanical characteristics and deformation control of surrounding rock in weakly
cemented siltstone. Environ. Earth Sci. 2021, 80, 337. [CrossRef]
53. Chen, W.; Wan, W.; Zhao, Y.; Peng, W. Experimental Study of the Crack Predominance of Rock-Like Material Containing Parallel
Double Fissures under Uniaxial Compression. Sustainability 2020, 12, 5188. [CrossRef]
54. Hu, S.M.; Hu, X.W. Estimation of rock mass parameters based on quantitative GSI system and Hoek-Brown criterion. Rock Soil
Mech. 2011, 32, 861–866.
55. Hoek, E.; Brown, E.T. Empirical strength criterion for rock masses. J. Geotech. Eng. Div. 1980, 106, 1013–1035. [CrossRef]
56. Su, Y.; Zou, Y. Stability analysis of support structure based on HOEK-BROWN strain-softening model. J. Cent. South Univ. (Sci.
Technol.) 2020, 51, 453–463. [CrossRef]
57. Li, L.; Zeng, Z.; Wang, Z.; Li, T. Upper bound punching failure analysis of circular foundation overlying a cave in Karst area
based on the Hoek-Brown failure criterion. J. Cent. South Univ. (Sci. Technol.) 2020, 51, 134–144. [CrossRef]
58. Hoek, E. Strength of rock and rock masses. ISRM News J. 1994, 2, 4–16.
59. Hoek, E.; Brown, E.T. Practical estimates of rock mass strength. Int. J. Rock Mech. Min. Ences 1997, 34, 1165–1186. [CrossRef]
60. Hoek, E.; Brown, E.T. The Hoek-Brown failure criterion and GSI-2018 edition. J. Rock Mech. Geotech. Eng. 2019, 11, 445–463.
[CrossRef]
61. Wu, S.C.; Jin, A.B.; Gao, Y.T. Numerical simulation analysis on strength reduction for slope of jointed rock masses based on
gereralized Hoek-Brown failure criterion. Chin. J. Geotech. Eng. 2006, 28, 1975–1980.
62. Song, K.; Yan, E.C.; Mao, W.; Zhang, T.T. Determination of shear strength reduction factor for generalized Hoek-Brown criterion.
Chin. J. Rock Mech. Eng. 2012, 31, 106–112.
63. Yang, X.L.; Wang, J.M.; Sui, Z.R. Stability analysis of tunnel surrounding rock based on Hoek-Brown yield criterion. J. Railw.
Acience Eng. 2008, 5, 37–40.
64. Hammah, R.E.; Yacoub, T.E.; Corkum, B.; Curran, J.H. The shear strength reduction method for the generalized Hoek-Brown
criterion. In Proceedings of the 40th US Symposium on Rock Mechanics: Rock Mechanics for Energy, Mineral and Infrastructure
Development in the Northern Regions, Anchorage, AK, USA, 25–29 June 2005.
65. Li, A.J.; Merifield, R.S.; Lyamin, A.V. Effect of rock mass disturbance on the stability of rock slopes using the Hoek–Brown failure
criterion. Comput. Geotech. 2011, 38, 546–558. [CrossRef]
66. Shen, J.Y.; Priest, S.D.; Karakus, M. Determination of Mohr-Coulomb Shear Strength Parameters from Generalized Hoek-Brown
Criterion for Slope Stability Analysis. Rock Mech. Rock Eng. 2012, 45, 123–129. [CrossRef]
67. Shen, J.Y.; Karakus, M.; Xu, C.S. Chart-based slope stability assessment using the Generalized Hoek-Brown criterion. Int. J. Rock
Mech. Min. Sci. 2013, 64, 210–219. [CrossRef]
68. Chen, Y.F.; Lin, H.; Li, S.; Cao, R.H.; Yong, W.X.; Wang, Y.X.; Zhao, Y.L. Shear expression derivation and parameter evaluation of
Hoek–Brown criterion. Arch. Civ. Mech. Eng. 2022, 22, 77. [CrossRef]
69. Shen, J.; Karakus, M.; Xu, C. Direct expressions for linearization of shear strength envelopes given by the Generalized Hoek–Brown
criterion using genetic programming. Comput. Geotech. 2012, 44, 139–146. [CrossRef]
70. Wang, W.; Zhang, H.; Li, X. Tension-compression damage model of rock under blasting load. J. Cent. South Univ. (Sci. Technol.)
2021, 52, 3918–3929. [CrossRef]
71. Wang, L.; Zhu, Z.; Lei, J.; Zhang, S.; Long, W.; Shu, B. Mechanical responses to wellbore instability of crystalline rocks in scientific
deep well during thermo-mechano coupling. J. Cent. South Univ. (Sci. Technol.) 2021, 52, 465–477. [CrossRef]
72. Zhang, D.; Ma, Z.; Liu, Z. Failure mode of deep underground cavity and upper bound solution of surrounding rock pressure.
J. Cent. South Univ. (Sci. Technol.) 2019, 50, 649–657. [CrossRef]
73. Zhao, Y.; Zhang, C.; Wang, Y.; Lin, H. Shear-related roughness classification and strength model of natural rock joint based on
fuzzy comprehensive evaluation. Int. J. Rock Mech. Min. Sci. 2021, 137, 104550. [CrossRef]
Materials 2023, 16, 2793 16 of 16

74. Zhao, Y.; Liu, Q.; Zhang, C.; Liao, J.; Lin, H.; Wang, Y. Coupled seepage-damage effect in fractured rock masses: Model
development and a case study. Int. J. Rock Mech. Min. Sci. 2021, 144, 104822. [CrossRef]
75. Benz, T.; Schwab, R.; Kauther, R.A.; Vermeer, P.A. A Hoek–Brown criterion with intrinsic material strength factorization. Int. J.
Rock Mech. Min. Sci. 2008, 45, 210–222. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.