Engineering Geology 260 (2019) 105245

Contents lists available at ScienceDirect

Engineering Geology
journal homepage: www.elsevier.com/locate/enggeo

Coupled calculation model for anchoring force loss in a slope reinforced by a T

frame beam and anchor cables
Keyou Shia,b, Xiaoping Wua,b, , Ze Liuc, Shenglan Daia,b
⁎

a
School of Civil Engineering, Central South University, Changsha 410075, Hunan, People's Republic of China
b
National Engineering Laboratory for High Speed Railway Construction, Changsha 410075, Hunan, People's Republic of China
c
School of Civil Engineering, Hunan University of Science and Technology, Xiangtan 411201, Hunan, People's Republic of China

ARTICLE INFO ABSTRACT

Keywords: Structures consisting of anchor cables and frame beams are structurally safe and reliable and are widely used in
Slope reinforced by frame beam and anchor slope support engineering. However, the loss of anchoring force of the anchor cables is also very common, and
cables this phenomenon directly affects the reinforcement effect of the anchoring structure. To predict the law and
Anchoring force loss degree of the anchoring force loss accurately, a new coupled calculation model is established in this work. This
On-site monitoring test
new model is especially suitable for slopes reinforced with frame beams and anchor cables, and its accuracy is
Coupled calculation model
Rock mass creep
verified with on-site monitoring tests. The primary application of this new model is to provide theoretical
guidance for late reinforcements of the anchoring forces of cables, which has great significance for safe operation
in slope engineering.

1. Introduction force loss in actual slope engineering projects accurately.

Many scholars have monitored the long-term anchoring force of
Prestressed anchor cables have been widely used in slope support anchor cables by installing intelligent sensors, and based on the ob-
engineering owing to their unique advantages of high strength, simple served change in the anchoring force, prediction models for the change
construction, and economic benefits. The anchor cables with frame in the anchoring force of the anchor cables have been established (Jiang
beam structure is a commonly used support structure form (Danziger et al., 2014; Chen et al., 2018; Fan et al., 2015; Fu et al., 2017; Sung
et al., 2006; Yang et al., 2015; Fu et al., 2018; Yan et al., 2019). The et al., 2017). However, most of these prediction models are empirical,
principle of using anchor cables and a frame beam to reinforce a slope is and the error of these empirical equations is very large under different
to make the anchor cables fully exert their high strength through ten- slope engineering conditions. Considering that the long-term loss of
sioning and to connect the weak structural plane of the rock mass with anchoring force of the anchor cables is closely related to the creep of the
the stable rock stratum and the frame beam to improve the stability of rock mass, many coupled creep models have been established to cal-
the slope. Although a large number of engineering projects have de- culate the long-term anchoring force of the anchor cables, and the re-
monstrated that the anchor cable and frame beam structure is struc- liability of these models has been verified through model tests (Wang
turally safe and reliable, loss of anchoring force of the anchor cables is et al., 2014a; Chen et al., 2013; Xie et al., 2017).Unlike the empirical
also very common. In the multi-anchor reinforcement slope, for the models, the calculation parameters for the coupled creep models are
anchor cables acting at different positions on the slope, the degree of obtained through on-site creep tests of the rock mass (Tan et al., 2018;
loss of the anchoring force is also different. When the anchoring force of Wu et al., 2017; Mansouri and Ajalloeian, 2018; Paraskevopoulou et al.,
each anchor cable has different degrees of loss, it is easy to cause the 2017), which greatly improves the applicability and accuracy of these
stress of the frame beam to be unbalanced, which may cause cracking of coupled creep models. However, the existing coupled creep models only
the reinforced concrete structure. Further development of cracks will consider the coupling between the anchor cable and the rock mass. For
not only cause structural deformation, but also directly affect the re- anchor cable and frame beam structures, the influence of the frame
inforcement effect of the anchoring structure (Zhu et al., 2005; Zheng beam on the anchor cables is not reflected in the existing coupled creep
et al., 2019; Bi et al., 2019; Li et al., 2016; Shi et al., 2018). Therefore, it models.
is of great importance to establish a calculation model for the loss of The premise for establishing a coupled creep model is to assume that
anchoring force of the anchor cables to predict the amount of anchoring the rock mass is uniformly stressed (Wang et al., 2014b; Wang et al.,

⁎
Corresponding author at: School of Civil Engineering, Central South University, Changsha 410075, Hunan, People's Republic of China.
E-mail address: [email protected] (X. Wu).

https://doi.org/10.1016/j.enggeo.2019.105245
Received 28 February 2019; Received in revised form 21 July 2019; Accepted 1 August 2019
Available online 06 August 2019
0013-7952/ © 2019 Published by Elsevier B.V.
K. Shi, et al. Engineering Geology 260 (2019) 105245

2017). In a slope supported by a frame beam and anchor cables, the

frame beam allows each anchor cable to play a synergistic role, and the
actual stress situation can thus be closer to the assumption. Therefore,
Fig. 1. Maxwell model.
to reflect the role of the frame beam in the coupled creep model, the
relationship between the anchor cables, rock mass, and frame beam was
considered in this work to establish a coupled calculation model for the characteristics of elasticity, plasticity, and creep deformation under
anchoring force loss of a slope reinforced by a frame beam and anchor different stress conditions. Thus, in order to describe the complex force
cables, and the accuracy of the model was verified with on-site mon- characteristics of the rock mass, creep models of the rock mass have
itoring. always been a focus of geotechnical engineering research. At present,
the creep models commonly used to describe the rock mass include the
2. Factors influencing the anchoring force loss of anchor cables following (Zhang et al., 2019; Wu et al., 2019; Pramthawee et al., 2017;
Liu et al., 2017).
After the installation of prestressed anchor cables is completed, the
anchoring force of the anchor cables will be lost to varying degrees. A 3.1.1. Maxwell model
large amount of research and engineering practice has demonstrated The Maxwell model (Fig. 1) consists of a Hooke body (H-body) and a
that the change in the anchoring force of the anchor cables after con- Newton body (N-body) in series. The Maxwell model can describe the
struction is complete can be divided into three phases: the rapid decline creep properties of a rock mass under constant stress and can also re-
phase, fluctuation phase, and stable phase (Zhang et al., 2004).There flect the relaxation characteristics of the rock mass when the stress is
are several reasons for the loss of anchoring force of the anchor cables. gradually reduced. The Maxwell model is suitable for describing the
(1) Influence of the deformation of the anchor cable structure itself. second stage of rock mass creep. The constitutive equation for this
After the stretching of the anchor cables is completed, the rebound model is:
deformation of the anchorage and the slack in the steel strands will
cause a loss of anchoring force of the anchor cables. This loss is un- = +
E (1)
avoidable, but can be controlled by improving the construction
methods and construction quality. where σ is the stress on the rock mass,σ′ is the first derivative of the
(2) Influence of rock mass compression deformation. The anchor cable stress on the rock mass,ε′ is the first derivative of the strain of the rock
passes through the structural surface inside the rock mass, and cracks in mass,E is the elastic modulus of the H-body, and η is the viscosity
the rock body are compressed under the anchoring force, thus causing coefficient of the N-body.
the loss of anchoring force of the anchor cable. In the initial stage, when
the anchor cable is tensioned, the compressive deformation of the rock 3.1.2. Kelvin model
mass has an obvious influence on the anchorage force of the anchor The Kelvin model (Fig. 2) consists of an H-body and an N-body in
cables, which not only relates closely to the degree of crack develop- parallel. The Kelvin model is suitable for describing the first stage of
ment in the rock mass but also decreases gradually with time. rock mass creep, but it cannot reflect the transient stress relaxation and
(3) Influence of creep deformation of the rock mass. Under the action of elastic deformation characteristics of the rock mass. The constitutive
stress, the rock mass has mechanical properties such as plasticity, equation for this model is as follows:
elastoplasticity, and rheology. The rock mass is thus prone to creep
= +E (2)
deformation under the long-term effect of the anchoring force. Owing
to the coupled effect of the anchor cable and the rock mass, the anchor where ε is the strain generated by the rock mass under stress.
cable relaxes with creep deformation of the rock mass, resulting in a
loss of anchoring force of the anchor cable. In general, the worse the 3.1.3. General Kelvin model
mechanical properties of the rock mass are, the greater the creep de- The General Kelvin model (Fig. 3) consists of an H-body and a
formation, and the greater the anchoring force loss of the anchor cables Kelvin model in series. The General Kelvin model can describe the
will be. various stages of rock mass creep and can comprehensively describe the
(4) Influence of construction disturbance. Blasting excavation of the viscoelastic–plastic characteristics of the rock mass at different stages. It
lower slope is often accompanied by a strong unloading effect. This is widely used in geotechnical theoretical analyses. The constitutive
unloading effect will change the stress state of the supported slope, equation of this model is as follows:
causing the cracks inside the supported slope to develop further. At the
K K EH EH EK
same time, the anchorage cable produces retraction deformation, re- + = +
EH + EK EH + EK EH + EK (3)
sulting in a loss of anchoring force of the anchor cable.
(5) Influence of environmental changes. Changes in the temperature where EH is the instantaneous elastic modulus,EK is the viscoelastic
will alter the stress state of the anchor cable structure, causing expan- modulus, and ηK is the viscosity coefficient.
sion and contraction of the anchor cable. With heavy rainfall, rainwater
will seep into the rock mass, which reduces the shear strength of the 3.2. Selection of rock mass creep model
rock mass. These environmental factors can cause changes in the an-
choring force of the anchor cable. The above creep models commonly used for rock masses all consist
It can be seen that the loss of anchoring force of the anchor cable is of a combination of basic components. The basic components that
mainly determined by its own material properties and the mechanical comprise rock mass creep models mainly include elastic components
properties of the rock. Thus, the loss of anchoring force is generally
established through a reasonable combination of anchor cable and rock
mass calculation models.

3. Research on coupled calculation models of anchoring force loss

3.1. Common creep model of rock mass

The rock mass has complex mechanical properties, with different Fig. 2. Kelvin model.

2
K. Shi, et al. Engineering Geology 260 (2019) 105245

K considering the relationship between the anchor cables, the rock mass,
EH and the frame beam. The established coupled calculation model is
composed of three parts: the first part is the General Kelvin model,
which is used to simulate the rock mass; the second part is an H-body,
which is used to simulate the anchor cable; and the third part is the
EK Kelvin model, which is used to simulate the frame beam. For a slope
reinforced by a frame beam and anchor cables, the rebound deforma-
Fig. 3. General Kelvin model.
tion of the anchor cable and the creep of the frame beam can be con-
sidered independent of each other. Owing to the coupling effect be-
(such as H-bodies), viscous components (such as N-bodies), and plastic tween the anchor cable and the rock mass, the rebound deformation of
components. Different creep models have different characteristics and the anchor cable and the deformation of the rock mass can be con-
application scopes, and their emphasis of the creep characteristics of sidered to be synchronized. Therefore, the coupled calculation model is
the rock mass reaction also varies. Therefore, the selection of a suitable created by connecting the first and the second parts in parallel, and
rock mass creep model is key for establishing a coupled calculation then connecting that unit in series with the third part.
model for the anchorage force loss of anchor cables. Considering the In Fig. 4, EC is the equivalent elastic modulus of the anchor body
factors influencing the anchoring force loss of the anchor cables and the within the anchorage range, EL is the elastic modulus of the frame
variation in the anchoring force loss at different stages, the established beam, and ηL is the viscosity coefficient of the frame beam. The
coupled calculation model should reflect the following characteristics equivalent elastic modulus of the anchor body can be obtained as fol-
of the coupling system: lows:
A0
(1) the transient stress relaxation and elastic deformation character- EC = E0 ×
istics of the rock mass; AC (4)
(2) the viscoelastic–plastic characteristics of the rock mass at different where E0 is the actual elastic modulus of the anchor cable, A0 is the
stages; and. actual cross-sectional area of the anchor cable, and AC is the cross-
(3) the long-term nature of the anchoring force loss of the anchor cable. sectional area of the anchor body.

According to the structural characteristics of the coupling system,

4. Derivation of the basic equations for the coupled calculation
the retraction deformation of the anchor cable and the deformation of
model
the rock mass are synchronous, i.e., when the rock mass is deformed,
the anchoring force of the anchor cable changes accordingly. Therefore,
It can be seen from the coupled calculation model in Fig. 4 that the
the coupled calculation model to be established should fully reflect the
stress and strain between the anchor cable, rock mass, and frame beam
aging deformation characteristics of the rock mass.
satisfy the following relationships:
Ananalysis of the commonly used rock mass creep models reveals
that the Maxwell model is only suitable to describe the characteristics of = L = N + C (5)
stress relaxation and elastic deformation of the rock mass, and thus = + = + (6)
L N L C
cannot meet the requirements of conditions (2) and (3) listed above.
Although the Kelvin model can reflect the viscoelastic–plastic char- where σ is the total stress of the coupled calculation model; ε is the total
acteristics of the rock mass, it has no transient elastic deformation or strain of the coupled calculation model; σL and εL are the stress and
viscous flow characteristics, and thus cannot meet the requirements of strain of the frame beam, respectively; σN and εN are the stress and
the conditions (1) and (3) listed above. Compared with the Maxwell strain of the rock mass, respectively; and σC and εC are the stress and
model and the Kelvin model, the General Kelvin model can compre- strain of the anchor cable, respectively.
hensively describe the viscoelastic–plastic characteristics of the rock By converting Eqs. (5) and (6), the following can be obtained:
mass at different stages. Therefore, the General Kelvin model is selected (EC + EL )
to simulate the creep behavior of the rock mass. N = C = ( L ) EC = EC
EL (7)

3.3. Establishment of a coupled calculation model for the loss of anchoring N = L =

EL (8)
force
where EL is the elastic modulus of the frame beam, and EC is the elastic
As shown in Fig. 4, a new coupled calculation model was established modulus of the anchor cable.
Substituting Eqs. (7) and (8) into Eq. (3), the constitutive equation
Part 2 for the coupled calculation model can be obtained as follows:
+A =B +C (9)
EC Part 3 where
(EC + EL )(EH + EK ) + EH EK
A=
K EC + K EL + K EH

K EH EL + EC EL
B=
EH EL EC + EL + EH

EH EK EL + EC EL (EH + EK )
C=
EK K EC + K EL + K EH

Part 1 Assuming that when t = 0,the stress and strain of the coupled body
are σ0 and ε0, respectively, then Eq. (9) can be transformed into the
Fig. 4. New coupled calculation model. stress relaxation equation for the coupled calculation model as follows:

3
K. Shi, et al. Engineering Geology 260 (2019) 105245

Elevation (m)
815
Strong weathered
quartz schist 810

Unit (m)
Frame beam 805

2 2 2 2 2 2 2 2 2 2
20° Mid-weathered
1 Cable8 quartz schist
1:
10
Cable7 800
Cable6
Frame beam Cable5 Stage2 795

1 Cable4
1:
10

Cable3 790
Cable2
20°
Cable1
F r ee Stage1 785
len
12m gth
Anc
hore
d 780
10m length

775

Fig. 5. Reinforcement scheme of the deep cutting slope.

+A =C 0 (10) Table 1
Main performance indicators of the sensor used for on-site monitoring.
Solving the differential Eq. (10), the following can be obtained:
Sensor model Type Installation Range (kN) Sensitivity (kN)
C method
= C1 exp( At ) + 0
A (11)
JMZX-3108HAT Four-string heart- 800 0.1
intelligent piercing
When the stress initially acts on the rock mass, the rock mass pri-
sensor
marily undergoes elastic deformation, and thus the strain of the coupled
body is 0 = E +0 E + E0 at this time. In addition, the initial condition
C H L
indicates that when t = 0, the stress of the coupled body is σ = σ0. The 5. On-site monitoring test verification
value of coefficient C1 can be obtained by substituting the initial con-
dition into Eq. (11) as follows: 5.1. Project overview
C
C1 = 0 0
(12) The EK1 + 555–EK1 + 706.93 section of the Zhengxi (Zhengzhou
A
to Xixia) highway is located in Miaozi Town, Luoyang City, China. This
Eq. (11) is the relaxation equation of the coupled calculation model area is marked by canyons, with dangerous and undulating terrain, and
for the anchoring force loss of an anchor cable. Under the condition that there are a large number of deep excavation slopes. The on-site mon-
the initial strain of the coupling body, ε0, is known, the long-term an- itoring in this study is based on the deep cutting slope in the EK1 + 640
choring force of the anchor cable can be reasonably predicted. As- section. The maximum excavation depth of the slope is 70.6 m. The rock
suming that the anchoring force is evenly distributed in the rock mass
within the length of the anchor cable, the equation for calculating the
anchoring force of the anchor cable is as follows:

F = AC (13)

where F is the anchoring force of the anchor cable, and is also a function
of time, t.

Frame beam

Sensor

Steel strand

Slope
Anchorage

Fig. 7. Deep cutting slope after tensioning of the anchor cables and installation
Fig. 6. Diagram of sensor installation. of the sensors.

4
K. Shi, et al. Engineering Geology 260 (2019) 105245

Fig. 8. Time–deformation curves of the rock mass. Fig. 9. Measured variation in the anchoring force of the anchor cables during
the monitoring period.
mass of the slope can be divided into three geological engineering
layers. From top to bottom, it is divided into fully weathered quartz cutting slope after the anchor cables were tensioned and the sensors
schist, strongly weathered quartz schist, and mid-weathered quartz were installed.
schist. The upper rock mass was excavated during the slope construc-
tion, and the rock mass within the excavation depth is fully weathered 5.3. Comparative analysis of the monitoring and model calculation results
rock. The rock stratum supported by the anchor cables is mainly com-
posed of strongly weathered and mid-weathered rock. According to The designed anchoring load of the anchor cables monitored in this
field shear tests and laboratory uniaxial compression tests, the density study is 400kN.The cross section of each slope stage consists of four
of the rock mass is 20.79kN/m3, the cohesion is 29.3 kPa, the internal anchor cables, and the anchor cables are arranged at equal intervals.
friction angle is14.6°, the elastic modulus is 33.1 MPa, and the Poisson's The anchor hole diameter is 130 mm, the grouting hole diameter is
ratio is 0.24. 25 mm, and the grouting uses 1:1 cement mortar. Each anchor cable
The monitoring slope is divided into seven stages for excavation. consists of four steel strands. The diameter of each steel strand is
The excavation depth of each stage is 10 m, and the slope rate is 1:1. As 15 mm, the nominal area is 140mm2, the tensile strength is 1860 MPa,
shown in Fig. 5, the support structure for the slope is frame beams and and the elastic modulus is 1.95 × 105 MPa. The frame beam is poured
prestressed anchor cables. Each anchor cable is 22 m long, the an- using C25 concrete. The interface dimensions of the frame beam are
choring section is 10 m long, and the vertical spacing between anchor 0.4 m × 0.4 m, and the elastic modulus is 2.8 × 104 MPa.
cables is 2 m. The angle between the anchor cables and the horizontal To obtain the creep parameters of the rock mass, a creep test of the
direction is 20°. rock mass was carried out through the rigid plate center hole method at
the engineering site, and the time–deformation curve of the rock mass
5.2. Monitoring test method was obtained, as shown in Fig. 8. The test was divided into three groups
with constant applied loads of 5 MPa, 10 MPa, and 15 MPa, and the
For this test, two stages near the foot of the slope were selected, and creep test duration was 30 d.
the anchoring force of eight groups of anchor cables was monitored. At the same time, Eq. (3)) is transformed and solved to obtain the
The anchor cables are numbered 1–8 from the foot to the top of the creep equation for the rock mass as follows:
slope (Fig. 5). Before the anchor cables were tensioned, an anchoring
force sensor was installed under the anchorage of each anchor cable =
EH + EK
exp
EK
t +
EH + EK
0 m m
(Fig. 6). The sensors used for monitoring are intelligent string digital EH EK K EH EK (14)
heart-piercing force sensors, and the main performance indicators of
the sensors are listed in Table 1. The monitoring data acquisition where ε0 is the instantaneous strain of the rock mass, and σm is the
methods are as follows: (1) installation of the DSC (Data-operation constant load applied during the creep test.
Supervise Center) wireless data acquisition system on site; (2) using the Equation (14) indicates that when t = 0, the rock mass is subject to
internet function of the GPRS (General Packet Radio Service), establish an instantaneous elastic strain. Based on the instantaneous strain of the
a live internet mobile phone module by agreeing on the domain name; rock mass under each constant load in Fig. 8, the instantaneous elastic
(3) receive data through the wireless network using a computer with modulus of the rock mass can be obtained as follows:
professional software installed. EH = m
Before monitoring the anchoring force of the anchor cables, the 0 (15)
computer receiving software was set to an automatic monitoring mode.
When t → ∞, the creep of the rock mass tends to become stable.
After the setup was completed, data was collected and recorded auto-
According to the stable strain values for the rock mass under each
matically at 6:00 every morning. Monitoring of the anchoring force of
constant load in Fig. 8, the viscoelastic modulus of the rock mass can be
the anchor cables was performed for 120 d. Fig. 7 shows the deep
obtained as follows:
Table 2 m EH m EH
EK =
Parameters of the coupled calculation model. EH (t ) m EH (t = 30) m (16)
2
EC(MPa) EL(MPa) EH(MPa) EK(MPa) ηK(MPa·d) AC(m )
After obtaining the creep parameters of the rock mass,EH and EK, by
8.23 × 10 3
2.8 × 10 4
4.19 × 10 4
5.75 × 10 4
5.97 × 10 5
0.01327 defining a point (ε(t > 0),t) on the creep curve shown in Fig. 8, the
viscosity coefficient,ηK, of the rock mass can be obtained. The

5
K. Shi, et al. Engineering Geology 260 (2019) 105245

Fig. 10. Comparison of anchoring forces of the anchor cables acting on the upper middle of the slope.

parameters for the coupled calculation model were obtained by para- cables with a large degree of anchoring force loss are located at the top
meter inversion and are listed in Table 2. of each stage slope. In addition, the anchoring force of the anchor
As shown in Fig. 9, It can be seen from the monitoring data that the cables at the top of each slope stage exhibit obvious fluctuations, which
anchoring force of the anchor cables experiences varying degrees of loss indicates that those cables are greatly affected by environmental
after the cable tension is completed; the loss of anchoring force is changes and construction disturbances. Considering that the anchoring
greatest for cable 4 and cable 8. Combined with the position of each force loss of cable 4 and cable 8 exceeds 50%, and the environmental
anchor cable shown in Fig. 5, it can be determined that the anchor changes and construction disturbances have a randomness and

Fig. 11. Comparison of the anchoring forces of cables acting on the lower middle of the slope.

6
K. Shi, et al. Engineering Geology 260 (2019) 105245

Table 3
Comparison of the monitoring and model calculation results for the anchoring force of the cables acting at the upper middle part of the slope.
Cable number Data category Anchoring force stabilization time (d) Anchoring force stability value (kN) Anchoring force loss Wang et al.,

Cable 3 Monitoring data 90 275.5 16.8

Calculation data (new model) 30 277.3 16.3
Calculation data (General Kelvin) 30 191.2 42.1
Calculation data (Wang et al., 2014a) 30 256.7 22.5
Cable 7 Monitoring data 85 261.4 16.2
Calculation data (new model) 30 261.3 16.3
Calculation data (General Kelvin) 30 180.6 42.2
Calculation data (Wang et al., 2014a) 30 241.9 22.5

uncertainty and thus cannot be reflected in the coupled calculation 3 and cable 7 are compared with the model calculation results. The
model, therefore, to compare and analyze the monitoring results and results show that the final stable value of the anchoring force and the
model calculation results accurately, the coupled calculation model can loss degree of the anchoring force are consistent in the two sets of re-
be verified using only the monitoring results for cables 1–3 and 5–7. sults. However, when the coupled calculation model is used to calculate
Furthermore, to evaluate the prediction improvement achieved with the anchoring force of the anchor cables, the anchoring forces of cable 3
the new model, curves representing the calculation results obtained and cable 7 reach a stable phase after approximately 30 d, while in the
with the General Kelvin model and the GK-H model (Wang et al., field monitoring results, the anchoring forces of cable 3 and cable 7
2014a) are also included in Figs. 10 and 11. reach a stable phase after approximately 90 and 85 d, respectively.
During the construction of the project, the designed anchoring load Therefore, for the anchor cables acting on the upper middle part of each
of the anchor cables is 400KN, but It can be seen from the monitoring slope stage, the coupled calculation model can be used to calculate the
data in Fig. 9 that the anchoring force of the anchor cables will drop degree of loss of the anchoring force and the final stable value of the
rapidly within 1–2 days after the construction is completed, the actual anchoring force, but it cannot be used to predict the time to reach
anchorage load of each anchor cable is made smaller than their de- stability or the fluctuation amplitude of the anchoring force for cables
signed anchoring load. in this area.
The rapid decline of the anchoring force of the anchor cables is In addition, comparing the results of the coupled calculation model
mainly caused by the springback deformation of the anchorage and the with the results of the other two calculation models, it can be seen that
compression of the joint cracks in the rock mass, rather than the creep because the General Kelvin model is mainly used to calculate the creep
of the anchoring body. The coupled calculation model established in of the rock mass, the anchor cable and the frame beam are not con-
this paper is based on the creep of the anchoring body, therefore, the sidered. The GK-H model (Wang et al., 2014a) is a parallel connection
effective anchoring force value of the anchor cable after rapid decline is between the General Kelvin model and an H-body; the H-body is used to
used as the initial condition of the model. For the anchor cables acting simulate the anchor cable, but the model does not consider the role of
at different positions on the slope, the initial conditions (the initial the frame beam. Therefore, the calculation results obtained with the
stress and strain of the anchoring body at the start of creep) of the two existing models are both lower than the monitoring results.
coupled calculation model is also different, which cause to the different The results in Table 3 indicate that when the General Kelvin model
calculation results in Figs. 10 and 11 and Tables 3 and 4. is used to calculate the anchoring force loss of the anchor cables, the
The comparison of the anchoring force of the anchor cables acting resulting loss is 42.1%. When the GK-H model (Wang et al., 2014a) is
on the upper middle part of the slope is shown in Fig. 10. It can be seen used to calculate the anchoring force loss of the anchor cables, the loss
from the curves in Fig. 10 that both the monitoring results and model is 22.5%. Thus, the degree of loss of the anchoring force predicted with
calculation results for cable 3 and cable 7exhibit a rapidly declining these models is far greater than that in the monitoring results and the
trend. However, because anchor cable 3 and anchor cable 7 are acting calculation results with the new model.
on the upper middle part of each slope stage, the anchor cables in this The comparison of the anchoring forces for the anchor cables acting
area are also sensitive to environmental changes and construction dis- on the lower middle section of the slope is shown in Fig. 11. It can be
turbances, and thus their monitoring results exhibit obvious fluctua- seen that for cables 1, 2, 5, and 6, the trends in the monitoring results
tions. In Table 3, the monitoring results for the anchoring force of cable for the anchoring forces are almost identical to the calculation results

Table 4
Comparison of the monitoring and model calculation results for the anchoring force of cables acting at the lower middle part of the slope.
Cable number Data category Anchoring force stabilization time (d) Anchoring force stability value (kN) Anchoring force loss Wang et al.,

Cable 1 Monitoring data 35 318 16.9

Calculation data (new model) 30 320.3 16.3
Calculation data (General Kelvin) 30 221.4 42.1
Calculation data (Wang et al., 2014a) 30 296.6 22.5
Cable 2 Monitoring data 45 327 16.0
Calculation data (new model) 30 326.2 16.3
Calculation data (General Kelvin) 30 225.5 42.1
Calculation data (Wang et al., 2014a) 30 302 22.5
Cable 5 Monitoring data 30 306 16.1
Calculation data (new model) 30 305.5 16.3
Calculation data (General Kelvin) 30 211.3 42.1
Calculation data (Wang et al., 2014a) 30 282.9 22.5
Cable 6 Monitoring data 25 311 16.2
Calculation data (new model) 30 310.6 16.3
Calculation data (General Kelvin) 30 214.7 42.1
Calculation data (Wang et al., 2014a) 30 287.6 22.5

7
K. Shi, et al. Engineering Geology 260 (2019) 105245

obtained with the new coupled calculation model. The results in Table 4 Therefore, for the anchor cables acting on the lower middle part of
also indicate the time for the anchoring force of these four anchor the slope, the coupled calculation model can be used to predict the
cables to reach a stable phase is also approximately the same in the various stages of the anchoring force of the anchor cables.
monitoring and new model calculation results. Therefore, for the an- (4) Comparing the model results with the on-site monitoring data and
chor cables acting on the lower middle part of each stage slope, the new calculation results obtained with other existing models shows that
coupled calculation model is not only suitable for describing the rapid although there are certain errors in the model calculation results,
decline phase and the steady phase of the anchoring force loss (i.e., the the loss of anchoring force of anchor cables can be effectively
first phase and the third phase of the anchoring force loss), but can also predicted by the coupled model, which is in line with the actual
be used to predict the time required for the anchoring force value to situation and has significant engineering application value.
reach the stable phase.
In addition, it can be seen from Fig. 11 that when the General Kelvin Acknowledgments
model and the GK-H model (Wang et al., 2014a) are used to calculate
the anchoring forces of cables 1, 2,5, and 6, the calculation results are This work is supported by the National Natural Science Foundation
also lower than the monitoring values. Therefore, compared with the of China (No. 51878672), and the Zhejiang Provincial Department of
existing models, the new coupled calculation model has higher preci- Transportation Foundation of China (No.2017028, No.2017033,
sion and is more suitable for the calculation of the anchoring force loss No.2017A01).
in a slope reinforced by a frame beam and anchor cables.
References
5.4. Discussion
Bi, J.F., Luo, X.Q., Zhang, H.T., et al., 2019. Stability analysis of complex rock slopes
Using the on-site monitoring data to verify the coupled calculation reinforced with prestressed anchor cables and anti-shear cavities. Bull. Eng. Geol.
Environ. 78 (3), 2027–2039.
model, it can be seen that the coupled calculation model established in Chen, Y.J., Yin, J., Hu, Y.F., 2013. Research on prestress quantitative loss law of soft rock
this paper can accurately predict the loss of the anchoring force of the slope anchor cable. Chin. J. Rock Mech. Eng. 32 (8), 1685–1691 (in Chinese).
anchor cables. The results are in line with the actual situation and thus Chen, G.Q., Chen, T., Chen, Y., et al., 2018. A new method of predicting the prestress
variations in anchored cables with excavation unloading destruction. Eng. Geol. 241,
the model has significant engineering application value. However, 109–120.
whether for the anchor cable acting on the upper middle part of the Danziger, F.A.B., Danziger, B.R., Pacheco, M.P., 2006. The simultaneous use of piles and
slope or the anchor cable acting on the middle and lower middle parts prestressed anchors in foundation design. Eng. Geol. 87 (3–4), 163–177.
Fan, Q.X., Zhu, H.B., Geng, J., 2015. Monitoring result analyses of high slope of five-step
of the slope, there are certain errors in the calculation results of the
ship lock in the three Gorges Project. J. Rock Mech. Geotech. Eng. 7 (2), 199–206.
model. There are several reasons for these errors. The model parameters Fu, X., Zhang, J.J., Zhou, L.R., 2017. Shaking table test of seismic response of slope re-
are calculated based on data from creep tests of the rock mass at the inforced by combination of anti-slide piles and multi-frame foundation beam with
anchor cable. Rock Soil Mech. 38 (2), 462–470.
engineering site. However, an invariable stress is applied during the
Fu, X., Ji, W.Y., Zhang, J.J., et al., 2018. Seismic response for plane sliding of slope
creep tests, which is different from the stress state of the rock mass in reinforced by anchor-chain-framed ground beams through shaking table test. Rock
engineering practice. Moreover, except for the influence of construction Soil Mech. 39 (5), 1709–1719.
disturbances such as slope excavation and blasting impact, the change Jiang, S.H., Li, D.Q., Zhang, L.M., et al., 2014. Time-dependent system reliability of an-
chored rock slopes considering rock bolt corrosion effect. Eng. Geol. 175, 1–8.
in the anchoring force of anchor cables in actual engineering is also Li, C., Wu, J., Wang, J., et al., 2016. Layout and length optimization of anchor cables for
affected by factors such as rainwater infiltration and temperature reinforcing rock wedges. Bull. Eng. Geol. Environ. 75 (4), 1399–1412.
changes, which cannot be reflected by the coupled calculation model. Liu, H.Z., Xie, H.Q., He, J.D., et al., 2017. Nonlinear creep damage constitutive model for
soft rocks. Mech. Time-Depend. Mat. 21 (1), 73–96.
Finally, the coupled calculation model was simplified when it was es- Mansouri, H., Ajalloeian, R., 2018. Mechanical behavior of salt rock under uniaxial
tablished. It considers that the rock mass is homogeneous and that the compression and creep tests. Int. J. Rock. Mech. Min. 110, 19–27.
anchoring force is evenly distributed, which is not the case in actual Paraskevopoulou, C., Perras, M., Diederichs, M., et al., 2017. The three stages of stress
relaxation - Observations for the time-dependent behaviour of brittle rocks based on
engineering. laboratory testing. Eng. Geol. 216, 56–75.
Pramthawee, P., Jongpradist, P., Sukkarak, R., 2017. Integration of creep into a modified
6. Conclusion hardening soil model for time-dependent analysis of a high rockfill dam. Comput.
Geotech. 91, 104–116.
Shi, J., Lin, P., Zhou, Y., et al., 2018. Reinforcement analysis of toe blocks and anchor
(1) In view of the coupling relationship between the frame beam, an- cables at the xiluodu super-high arch dam. Rock Mech. Rock. Eng. 51 (8),
chor cables, and rock mass, a coupled calculation model for the 2533–2554.
Sung, H.J., Tan Manh, D., Kim, J.M., et al., 2017. Long-term monitoring of ground anchor
anchoring force loss in a slope reinforced by a frame beam and
tensile forces by FBG sensors embedded tendon. Smart. Struct. Syst. 19 (3), 269–277.
anchor cables was established. The constitutive equation and re- Tan, Q.W., Tang, H.M., Fan, L., et al., 2018. In situ triaxial creep test for investigating
laxation equation of the model were deduced, and the theoretical deformational properties of gravelly sliding zone soil: example of the Huangtupo 1#
calculation equation for the anchoring loss of the anchor cables was landslide, China. Landslides. 15 (12), 2499–2508.
Wang, Q.B., Zhang, C., Wang, H., et al., 2014a. Study of coupling effect between an-
obtained. chorage force loss of prestressed anchor cable and rock and soil creep. Rock Soil
(2) The model was verified through on-site monitoring tests. Owing to Mech. 35 (8) 2150–2156 and 2162 (in Chinese).
their sensitivity to construction disturbances and environmental Wang, Dai, Harries, K.A., et al., 2014b. Prediction of prestress losses in RC beams ex-
ternally strengthened with prestressed CFRP sheets/plates. J. Reinf. Plast. Compos 33
changes, the monitoring results for the anchor cables acting on the (8), 699–713.
upper middle part of the slopes exhibit obvious fluctuations, and Wang, G.F., Li, Q., Lu, L.H., et al., 2017. A coupled model research of anchor prestress loss
the coupled calculation model cannot be used to predict the time considering the relaxation characteristics of anchor. Chin. J. Undergr. Space. Eng. 13
(6), 1585–1591 (in Chinese).
and amplitude of the fluctuations in the anchoring force of the Wu, L.Z., Li, B., Huang, R.Q., et al., 2017. Experimental study and modeling of shear
cables in this area. However, it can be used to calculate the degree rheology in sandstone with non-persistent joints. Eng. Geol. 222, 201–211.
of loss of the anchoring force and the final stable value of the an- Wu, F., Chen, J., Zou, Q.L., 2019. A nonlinear creep damage model for salt rock. Int. J.
Damage Mech. 28 (5), 758–771.
choring force. Xie, C., Li, S.C., Li, S.C., et al., 2017. Study of anchorage force loss of anchor cable under
(3) The anchor cables acting on the lower middle part of the slope are seepage flow and soil creep. Rock Soil Mech. 38 (8), 2313–2321 (in Chinese).
less affected by construction disturbance and environmental Yan, M.J., Xia, Y.Y., Liu, T.T., et al., 2019. Limit analysis under seismic conditions of a
slope reinforced with prestressed anchor cables. Comput. Geotech. 108, 226–233.
changes. Thus, not only are the trends in the monitoring results and
Yang, G.H., Zhong, Z.H., Zhang, Y.C., et al., 2015. Optimal design of anchor cables for
the model calculation results for the anchoring force almost iden- slope reinforcement based on stress and displacement fields. J. Rock Mech. Geotech.
tical, but the time required for the anchoring force of these anchor Eng. 7 (4), 411–420.
cables to reach the steady phase is also approximately the same. Zhang, F.M., Zhao, W.B., Liu, N., et al., 2004. Long-term performance and load prediction

8
K. Shi, et al. Engineering Geology 260 (2019) 105245

model of prestressed cables. Chin. J. Rock Mech. Eng. 23 (1), 39–43 (in Chinese). locally reinforced by rock bolts. Eng. Geol. 251, 228–240.
Zhang, S.G., Liu, W.B., Lv, H.M., 2019. Creep energy damage model of rock graded Zhu, D.Y., Lee, C.F., Chan, D.H., et al., 2005. Evaluation of the stability of anchor-re-
loading. Results Phys. 12, 1119–1125. inforced slopes. Can. Geotech. J. 42 (5), 1342–1349.
Zheng, Y., Chen, C., Liu, T., et al., 2019. Stability analysis of anti-dip bedding rock slopes