Received: 16 July 2018 Revised: 13 March 2020 Accepted: 12 April 2020

DOI: 10.1002/nag.3079

RESEARCH ARTICLE

A solution to estimate the total and effective stresses in

backfilled stopes with an impervious base during the filling
operation of cohesionless backfill

Jian Zheng1,2 | Li Li1,2 | Yu-Chao Li3

1
Research Institute on Mines and the
Environment (RIME UQAT-
Summary
Polytechnique), Canada Mining backfill is commonly used in underground mines. A critical concern of
2
Department of Civil, Geological and this practice is to evaluate the pressures and total stresses in backfilled stopes
Mining Engineering, École Polytechnique
to ensure a safe and economic design of barricades, constructed to retain the
de Montréal, C.P. 6079, succursale Centre-
Ville, Montréal, QC, H3C 3A7, Canada backfill. When a slurried backfill is placed in a mine stope, excess pore water
3
MOE Key Laboratory of Soft Soils and pressure (PWP) can instantaneously generate and progressively dissipate. The
Geoenvironmental Engineering, dissipation of the excess PWP and consolidation lead to the development of
Department of Civil Engineering,
Zhejiang University, Hangzhou, 310058,
effective stresses, which in turn lead to an arching effect in the backfilled
China stope. Until now, arching effect has been largely considered for stress estima-
tion in dry or submerged backfill. The former corresponds to the final state at
Correspondence
Jian Zheng, Research Institute on Mines the end process of the drainage and consolidation of the backfill with a pervi-
and the Environment (RIME UQAT- ous while the latter with an impervious barricade. However, previous studies
Polytechnique), , École Polytechnique de
have shown that the most critical moment for the stability of barricades is dur-
Montréal, C.P. 6079, succursale Centre-
Ville, Montréal, QC H3C 3A7, Canada. ing the stope filling. Therefore, the design of barricades requires a proper esti-
Email: [email protected] mation of the pressure and total stresses during the filling operation. This in

Funding information
turn needs joint consideration of the arching effect and consolidation of the
Research Institute on Mines and the backfill. In this paper, a new solution is developed to evaluate the pressures
Environment; Mitacs, Grant/Award and stresses in backfilled stopes during the filling operation of cohesionless
Number: IT12573; Fonds de recherche du
Québec - Nature et Technologies, Grant/ backfill by considering the self-weight consolidation and arching effect. The
Award Number: 2015-MI-191676; Natural proposed solution is validated by numerical modeling with Plaxis2D. It can
Sciences and Engineering Research
thus be used to evaluate the pressures and stresses in backfilled stopes during
Council of Canada, Grant/Award
Number: 402318 the stope filling with an impervious barricade.

KEYWORDS
arching effect, cohesionless backfill, effective and total stresses, excess pore water pressure
(PWP), impervious base, self-weight consolidation

1 | INTRODUCTION

Mining backfill is widely used in underground mines to fill underground openings. Several advantages can be attributed
to the application of mine backfill, including increased ore recovery, reduced mineral dilution, and improved ground
stability.1–4 Environment improvement through the reduction of the surface disposal of mine wastes is another impor-
tant benefit of the mine backfilling.5–7

Int J Numer Anal Methods Geomech. 2020;1–17. wileyonlinelibrary.com/journal/nag © 2020 John Wiley & Sons, Ltd. 1
2 ZHENG ET AL.

Despite the numerous advantages, successful application of backfill depends on the stability of a confining structure
constructed at the stope base to retain the slurried backfill in place. This structure is called barricade. Several cases of
barricade failure have been reported over the years, and the consequences were usually very serious.8–13 It is thus criti-
cal to ensure a safe and economic barricade design, which requires a good understanding of the pressures and stresses
in the backfilled stopes.
When a slurried backfill is instantaneously placed in a mine stope, the solid particles tend to settle down under their
own gravity. The backfill then tends to become denser, while the pore water tends to be expelled out of the backfill. As
the backfill, especially paste backfill, usually contains a large portion of fine particles, the permeability is low and the
drainage can be slow. This results in an instantaneous generation of excess pore water pressure (pore water pressure
(PWP), whose dissipation depends on the permeability and drainage time.14–17
The problem becomes more complicated when the stope filling operation is considered. On the one hand, the excess
PWP can be continuously generated and increased with the addition of new backfill layers. On the other hand, the dis-
sipation of excess PWP takes place with time during the filling operation. The resulting excess PWP depends on the
rates of the generation (increase) and dissipation (decrease) of the excess PWP during the filling operation, which in
turn depends on the hydraulic conductivity and filling rate of the backfill. This problem is known as self-weight consol-
idation of accreting deposition (constant increase in backfill thickness), which has been treated by Gibson,18 who con-
sidered one-dimensional consolidation in the vertical direction and proposed a solution for estimating the distribution
and evolution of the excess PWP. The application of the Gibson18 model in mine backfilled stopes has been shown by
several researchers.19–25 It is noted that the Gibson18 model and the ensuing solutions did not consider any effective
stresses during the filling operation.
However, field pressure and stress measurements showed that effective stresses develop during the filling opera-
tion.26,27 Subsequently, shear strength and shear stresses can develop along the fill–wall contact interfaces, leading to
the generation of arching effect.
Over the years, the consideration of arching effect in stress estimation in backfilled stopes has been largely
reported.28–35 Until now, arching effect has only been considered for dry and submerged (with a hydrostatic pressure;
e.g., Li and Aubertin36,37) backfills, which correspond to the final states of uncemented backfill at the end of the drain-
age and self-weight consolidation process with pervious and impervious barricades, respectively.
However, the previous studies have shown that the most critical moment for the stability of barricades is during the
filling operation of stopes, not at the end of the drainage and self-weight consolidation process.12,13 Therefore, the
design of barricades requires a proper estimation of the pressure and total stresses during the filling operation, which in
turn needs joint consideration of the arching effect and the drainage and consolidation of the backfill. This task can be
realized by numerical modeling. However, a simple solution is preferred to allow parameter sensitivity analyses during
the preliminary stage of projects.
In this paper, a solution is for the first time proposed to evaluate the pressures and stresses in stopes backfilled with
cohesionless backfill by considering the combined effects of arching and self-weight consolidation. The self-weight con-
solidation is first evaluated using an analytical solution proposed by the authors25, based on the Gibson18 model. The
dissipation of the excess PWP is then introduced into the arching model of Li and Aubertin.36 The proposed solution is
validated by numerical results obtained with Plaxis2D. A few sample applications of the proposed solution are shown.

2 | P R O P O S ED S O L U T I O N

2.1 | Self-weight consolidation

Figure 1 shows a typical mine stope filled with a saturated cohesionless backfill at a constant filling rate m (m/h). In
the figure, B (m) is the width of the stope; h (m) is the current height at the filing time t (h), that is, h = mt; H (m) is the
final height at the end of filling operation; y (m) is the elevation of the calculation point from the base of the stope,
ranging from 0 to h; and l (m) is the depth of the calculation point from the current top surface of the backfill, that is,
l = h − y.
The governing equation in terms of excess PWP is given by Gibson,18 as follows:

∂ 2 u ∂u 0 dh
cv = −γ , ð1Þ
∂y2 ∂t dt
ZHENG ET AL. 3

F I G U R E 1 A vertical
backfilled stope with saturated
cohesionless backfill and an isolated
horizontal layer element [Colour
figure can be viewed at
wileyonlinelibrary.com]

where u (kPa) is the excess PWP defined as the difference between the current total PWP and the hydrostatic water
pressure,38,39 cv (m2/h) is the consolidation coefficient, γ 0 (kN/m3) is the submerged unit weight of the backfill
(γ 0 = γ sat − γ w, where γ sat [kN/m3] is the saturated unit weight of the backfill and γ w [kN/m3] is the unit weight of
water). The excess PWP was further given by Gibson,18 as follows:


0 0 −y2
− 12
u = γ mt −γ E ðπcv t Þ exp , ð2Þ
4cv t

with
ð∞



mξ yξ ξ2
E= ξtanh × cosh × exp − dξ, ð3Þ
0 2cv 2cv t 4cv t

where ξ (m) is an arbitrary integration variable in the range of zero to infinity (∞).
Equations 1 and 2 are the solution of Gibson18 for the variation and evolution of the excess PWP during the filling
operation. However, the integral (Equation 3) of the Gibson18 solution cannot be solved by hand calculation or with
simple and commonly available calculation tools. The following transformation has been done by the authors25 to
obtain an analytical solution of the Gibson18 solution.Equation 3 can be represented by a function, as follows:
ð∞
 
G ðz Þ = gðzÞexp −z2 dz, ð4Þ
0

where

pffiffi

mz t yz
gðzÞ = 4cv t × z × tanh pffiffiffiffi × cosh pffiffiffiffiffiffi ð5Þ
cv cv t

and
ξ2
z2 = ð6Þ
4cv t
4 ZHENG ET AL.

Because g(z) is an even function of z, Equation 4 can be rewritten as

ð∞
1  
GðzÞ = gðzÞexp −z2 dz ð7Þ
2 −∞

Goodwin40 proposed the following transformation, which can give a good approximation of the integral with a high
accuracy by using a value of h0 in the range of 0 to 1 as long as the number n is taken large enough:
ð∞ X
∞
 
gðnh0 Þe − n h0 :
2 2
gðzÞ × exp −z2 dz = h0 ð8Þ
−∞ n = −∞

Comparing Equations 7 and 8 yields the following equation for E:

pffiffi

h0 X ∞
mnh0 t ynh0  
E= 4cv t × ðnh0 Þ × tanh pffiffiffiffi × cosh pffiffiffiffiffiffi × exp −n2 h20 : ð9Þ
2 n = −∞ cv cv t

Introducing Equation 9 into Equation 2 yields


−y2
u = γ mt − γ ðπcv t Þ − 2 exp
0 0 1
×
4cv t

pffiffi
 ð10Þ
h0 X ∞
mnh0 t ynh0  
4cv t × ðnh0 Þ × tanh pffiffiffiffi × cosh p ffiffiffiffiffi
ffi × exp −n2 h20 :
2 n = −∞ cv cv t

Equation 10 consists of the analytical solution of Gibson18 model, proposed by Zheng et al,25 for estimating the excess
PWP during the filling operation with an impervious base. In order to obtain stable and accurate results, the value of h0
should be small enough, while the range of n should be large enough. For most cases, stable results can be obtained by
taking a value of 0.3 for h0 and n in the range of −91 to 91.24,25
The derivation of the excess PWP can lead to

2  
 
du 0 − 12 −y −y
= −γ ðπcv t Þ exp × ×E+F , ð11Þ
dy 4cv t 2cv t

where F is expressed as
ð∞




mξ yξ ξ ξ2
F= ξtanh × sinh × × exp − dξ: ð12Þ
0 2cv 2cv t 2cv t 4cv t

Performing the same process to function E leads to the following expression for F:

pffiffi


h0 X ∞
mnh0 t ynh0 nh0  
F= 4cv t × ðnh0 Þ × tanh pffiffiffiffi × sinh pffiffiffiffiffiffi × pffiffiffiffiffiffi × exp − n2 h20 : ð13Þ
2 n = −∞ cv cv t cv t

Introducing Equations 9 and 13 into Equation 11 leads to

2

pffiffi
 3
−y h0 X ∞
mnh0 t ynh0  2 2

2  6 2c t × 2 4cv t × ðnh0 Þ × tanh pffiffiffiffi × cosh pffiffiffiffiffiffi × exp −n h0 + 7
du −y 6 v n = −∞
cv cv t 7
0 − 12
= −γ ðπcv t Þ exp 6
×6
p ffiffi 

 7:
dy 4cv t X∞   7
4 h0 mnh0 t ynh0 nh0 5
4cv t × ðnh0 Þ × tanh pffiffiffiffi × sinh pffiffiffiffiffiffi × pffiffiffiffiffiffi × exp −n h0
2 2
2 n = −∞ cv cv t cv t
ð14Þ
ZHENG ET AL. 5

Equation 14 will be used in the next section to consider the arching effect solution for the stresses during the placement
of backfill in mine stopes.

2.2 | Arching model

In order to take into account the arching effect in the process of self-weight consolidation, the forces acting on a layer
element shown in Figure 1 will be considered. W (kN) is the weight of the layer element; C (kN) and S (kN) are the
compressive and shear forces on each side wall, respectively; V (kN) and V + dV (kN) are the vertical forces on the top
and at the base, respectively.
Considering the equilibrium of the layer element in the vertical direction yields:

dV + 2S = W : ð15Þ

The weight of the layer element is calculated as

W = γ sat B dl, ð16Þ

where dl (m) is the thickness of the layer element.

The vertical force can be expressed as

V = σ v B, ð17Þ

where σ v (kPa) is the vertical total stress, which is assumed to be uniformly distributed across the width of the stope.
By assuming the effective friction angle δ0 along the fill–wall interfaces to be equal to the effective internal friction
angle of the backfill ϕ0 , the application of the Mohr–Coulomb criterion leads to the following expression for shearing
force S:

0 0 0 0
S = σ h tanδ dl = Kσ v tanϕ dl, ð18Þ

where σ 0 h (kPa) and σ 0 v (kPa) are the horizontal and vertical effective stresses, respectively; K (=σ 0 h/σ 0 v) is the earth
pressure coefficient of the saturated cohesionless backfill. In this study, the Rankine active earth pressure coefficient
(Ka) and Jaky's at-rest earth pressure coefficient (K0) are considered for the earth pressure coefficient K. The former is
expressed as follows (e.g., CGS and Das41,42):

0
K a = tan2 45 −ϕ =2 , ð19Þ

while the latter is given as follows (e.g., Jaky43):

0
K 0 = 1 −sinϕ : ð20Þ

The vertical and horizontal total stresses can be calculated as follows:

0
σ v = σ v + u + γ w l, ð21Þ

0 0
σ h = σ h + u + γ w l = Kσ v + u + γ w l: ð22Þ

From Equations 17 and 21, one can obtain the following expression:
6 ZHENG ET AL.

h 0 i
dV = B dσ v + du + γ w dl : ð23Þ

Substituting Equations 16, 18, and 23 into Equation 15 results in

h 0 i 0 0
Bd σ v + u + γ w l + 2Kσ v tanϕ dl = γ sat Bdl: ð24Þ

Equation 24 can be rewritten as follows:

0 0
dσ v 2Ktanϕ 0 0 du
+ σv = γ − : ð25Þ
dl B dl

Solving Equation 25 leads to

0
ð
 
0 − 2Ktanϕ du 2Ktanϕ0 l
0
σv = e B l
γ− e B dl + A , ð26Þ
dl

where A is an arbitrary constant.

Considering the boundary condition σ 0 v = 0 at l = 0, one obtains the vertical effective stress as follows (see Appen-
dix A for the details of derivation):

0 ðl

0 − 2Ktanϕ 0 du 2Ktanϕ0 l
σv = e B l
γ− e B dl: ð27Þ
0 dl

As l = h − y, one thus has

du du
=− : ð28Þ
dl dy

Equation 27 can then be rewritten as

0 ðl

0 2Ktanϕ 0 du 2Ktanϕ0 l
σv = e − B l
γ + e B dl, ð29Þ
0 dy

where du/dy can be obtained from Equation 14.

The horizontal effective stress can then be obtained by the following expression:

0 ðl

0 2Ktanϕ 0 du 2Ktanϕ0 l
σ h = Ke − B l
γ + e B dl: ð30Þ
0 dy

Adding the total PWP (equaling to the summation of excess PWP u and hydrostatic pressure) leads to the following
equations for the vertical (σ v) and horizontal (σ h) total stresses:

0 ðl

− 2Ktanϕ du 2Ktanϕ0 l
0
σv = e B l
γ + e B dl + u + γ w l, ð31Þ
0 dy

0 ðl

− 2Ktanϕ du 2Ktanϕ0 l
0
σ h = Ke B l
γ + e B dl + u + γ w l: ð32Þ
0 dy
ZHENG ET AL. 7

Equations 29 and 30 along with Equations 31 and 32 consist the proposed solution to evaluate the variation and evolu-
tion of the vertical and horizontal effective and total stresses in a backfilled stope during the filling operation of cohe-
sionless backfill.
A MATLAB program based on Equation 32 to estimate the distribution of the horizontal total stress (very important
for barricade design) within the backfilled stopes has been given in Appendix B. In the program, the backfill final
thickness H is divided into nmax elements. Once again, the value of nmax should be large enough to obtain stable and
reliable results of σ h. The sensitivity analysis should be conducted case by case for each project to find the optimum
value of nmax.
Figure 2 shows the variation of the horizontal total stress (σ h) at the bottom of the stope (l = 40 m) as the value of
nmax varies from 11 to 151 calculated with the MATLAB program by considering one case with the parameters of
H = 40 m, B = 6 m, m = 0.2 m/h, cv = 5 m2/h, γ sat = 20 kN/m3, γ w = 10 kN/m3, ϕ0 = 20 , and K = Ka. One sees that
stable results of σ h can be obtained when nmax is larger than 101 by using the given parameters (σ h = 600 kPa with
nmax = 101 compared with σ h = 599.3 kPa with nmax = 10 001).
As a special case when the filling operation is very slow (i.e., m à 0) or/and the consolidation coefficient is very
large (i.e., cv à ∞), the excess PWP (u = 0) and hydraulic gradient approach zero (i.e., u à 0, du/dy à 0). The proposed
solution (Equation 29) then reduces to the analytical solution of Li and Aubertin36 as follows:

0 ðl 0 0 
0 
0 2Ktanφ 0 2Ktanφ Bγ 2Kltanφ
σv = e − B l
γ ×e B l
dl = 1 −exp − : ð33Þ
0 2Ktanφ0 B

F I G U R E 2 Sensitivity analysis of nmax to obtain stable and

reliable results of σ h by using the MATLAB program (plotted using
Microsoft Excel); calculation made with H = 40 m, B = 6 m,
m = 0.2 m/h, cv = 5 m2/h, γ sat = 20 kN/m3, γ w = 10 kN/m3,
ϕ0 = 20 , and K = Ka

F I G U R E 3 Variation of the (A) vertical and (B) horizontal total stresses along the vertical center line (VCL) of the stope at the end of
the stope filling for different consolidation coefficients, calculated with the Li and Aubertin36 solution (Equation 33) and the proposed
solution (Equations 31 and 32); calculation made with H = 40 m, B = 6 m, m = 0.2 m/h, γ sat = 20 kN/m3, γ w = 10 kN/m3, ϕ0 = 20 , and
K = Ka
8 ZHENG ET AL.

Again, the solution of Li and Aubertin36 only considers the final state of the backfill when the drainage and consolida-
tion end and the water reaches a hydraulic equilibrium. This state can be reached when the filling rate is very small
or/and the consolidation coefficient is very large.
Figure 3 shows the variation of the vertical (Figure 3A) and horizontal (Figure 3B) total stresses along the height of
a stope 40 m high and 6 m wide at the end of filling operation, calculated with the solution of Li and Aubertin36 and
the proposed solution (Equations 31 and 32) by considering a filling rate of m = 0.2 m/h and friction angle of ϕ0 = 20
when the consolidation coefficient cv changes from 4 × 102 to 4 × 10−1 m2/h. The results calculated with the iso-
geostatic overburden pressure have also been plotted in the figure. One can see that the total stresses calculated with
the proposed solution are very close to those based on the iso-geostatic overburden pressure when the consolidation
coefficient is very small but approach the vertical and horizontal total stresses calculated with Li and Aubertin36 solu-
tion as long as the consolidation coefficient reaches a value of 4 × 101 m2/h.

3 | V A L I D A T I O N O F T H E P R O P O S E D S O L U T IO N AG A I N S T N U M ER I C A L
MODELING

To verify the validity of the proposed solution, numerical modeling has been performed using Plaxis2D, a numerical
code based on finite element method. 44
Figure 4A shows a backfilled stope 4 m large and 12.4 m high. The backfill is filled at a filling rate of m = 0.5 m/h
to a height of 12 m with a void space of 0.4 m left between the top surface of the backfill and the roof of the stope. The
backfill and rock mass are elasto-plastic and obey the Mohr–Coulomb criterion.
The backfill is cohesionless and characterized as γ sat = 20 kN/m3 (saturated unit weight), k = 4.176 × 10−2 m/h
(hydraulic conductivity), μ = 0.2 (Poisson's ratio), E' = 864 kPa (drained Young's modulus), ϕ0 = 10 (effective friction
angle), and ψ 0 = 0 (effective dilation angle). The drained constrained modulus of the backfill, M', can be calculated as
960 kPa from M' = E'(1 − μ)/[(1 + μ)(1 − 2 μ)] with μ = 0.2 and E' = 864 kPa. The consolidation coefficient (cv) of the
backfill can be calculated by the following equation:

F I G U R E 4 A, A backfilled stope. B, The

numerical model of Plaxis2D by taking into
account the plane of symmetry at x = 0 [Colour
figure can be viewed at wileyonlinelibrary.com]
ZHENG ET AL. 9

kM 0
cv = , ð34Þ
γw

and one obtains cv = 4 m2/h.

The surrounding rock mass is impermeable and characterized as γ r = 27 kN/m3 (unit weight), μr = 0.25 (Poisson's
ratio), Er = 42 GPa (Young's modulus), cr = 9400 kPa (cohesion), ϕr = 38 (friction angle), and ψ r = 0 (dilation angle).
The in situ stress ratio of rock mass is taken as Kr = 2, a typical value for the stress regime of Canadian Shield.45
This value does not have any influence on the stresses in the backfill because the backfill is placed in the stope after the
elastic deformation.
Figure 4B shows the numerical model of the backfilled stope, built with Plaxis2D. By considering the vertical sym-
metry plane, only half of the full model is simulated. The upper boundary of the model is free, while the lower bound-
ary is fixed in all directions. The two side boundaries (rock mass and backfill) of the model are fixed in the horizontal
direction but free for displacement in the vertical direction. In the stope, the vertical symmetry plane and the fill-rock
contacts are impermeable, while the top of each newly added backfill layer is simulated as permeable with zero PWP. It
should be noted that this zero PWP boundary condition cannot reflect the field hydraulic condition owing to the neglect
of the accumulated ponding water on top of the backfill. This will be further addressed in Section 5.
The numerical simulations have been carried out in three steps. The first step is to obtain an initial stress state in
the rock mass prior to any excavation. The stope excavation is simulated in the second step. The third step consists of
filling the stope layer by layer from the bottom to the top. The waiting delay for the addition of a new backfill layer is
calculated by considering the filling rate m. If each of the backfill layer has a thickness of dh, the waiting delay should
be dh/m. In order to simulate and approach as much as possible the continuous filling, the layer thickness dh should be
enough small. In this study, the stope of 12 m high was filled in 30 layers with each layer having a thickness of 0.4 m.
The waiting delay is 0.8 h for a filling rate of 0.5 m/h.
Figure 5 shows the variation of the excess PWP (Figure 5A), total and effective stresses (Figure 5B) along the vertical
center line (VCL) of the backfilled stope at the end of filling operation (at t = 24 h) obtained by numerical modeling
with Plaxis2D and calculated with the proposed solution (Equations 10 and 29 to 32) by considering H = 12 m,
B = 4 m, m = 0.5 m/h, cv = 4 m2/h, γ sat = 20 kN/m3, γ w = 10 kN/m3, and ϕ0 = 10 . The excess PWP (=γ'h) and the total
stress (=γ sath) based on the iso-geostatic overburden pressure are also plotted in the figure. One sees that the calculated
vertical and horizontal total stresses are lower than the overburden pressure, indicating the occurrence of arching effect
within the backfill. The σ 0 v and σ 0 h calculated with the proposed solutions using Ka are more consistent with the results
in the Plaxis2D. The excellent agreement between the numerical and proposed solution results indicates that the pro-
posed solution using Ka can be used to evaluate the pressures and stresses in backfilled stope during and at the end of
filling operation.

F I G U R E 5 Variation of the (A) excess pore water pressure (PWP), (B) total and effective stresses along the vertical center line (VCL) of
the stope at the end of filling operation (at t = 24 h and h = H = 12 m), obtained by numerical modeling with Plaxis2D and calculated with
the proposed solution (Equations 10, 29, 30, 31, and 32) using Rankine's active coefficient Ka and Jaky's at-rest coefficient K0; calculations
made with H = 12 m, B = 4 m, m = 0.5 m/h, cv = 4 m2/h, γ sat = 20 kN/m3, γ w = 10 kN/m3, and ϕ0 = 10
10 ZHENG ET AL.

4 | SAMPLE APPLICATIONS

Compared with numerical modeling that can be used to simulate diverse problems, the proposed solution can be partic-
ularly useful to perform parameter sensitivity analysis in the preliminary stage of projects. In this study, the influence
of the fill properties, stope width, and filling rate on the stresses in backfilled stopes will be evaluated by considering a
backfill having a saturated unit weight of γ sat = 20 kN/m3 and a final height of H = 40 m.
Figure 6 shows the variation of the effective (Figure 6A) and total (Figure 6B) stresses along the height of the stope,
calculated by the proposed solution at the end of stope filling as the consolidation coefficient cv varies from 0.1 to
10 m2/h. The other parameters considered are B = 6 m, m = 0.2 m/h, and ϕ0 = 20 . The Rankine active earth pressure
coefficient Ka is used for the calculations. It is seen that the effective and total stresses at the end of stope filling change
significantly with the change of cv. When cv takes a value of 0.1 m2/h, the horizontal and vertical effective stresses at
the end of filling along the upper part of the backfill are close to zero, while the vertical and horizontal total stresses are
close to the iso-geostatic overburden pressure. These results indicate the few drainage and consolidation during the fil-
ing operation in the upper part of the backfill. When cv takes a value of 10 m2/h, the horizontal and vertical effective
stresses at the end of filling are not negligible, while the vertical and horizontal total stresses are much smaller than the
iso-geostatic overburden pressure.
Figure 7 shows the variation of the effective (Figure 7A) and total (Figure 7B) stresses along the height of the stope
at the end of the stope filling for different filling rates m, obtained by applying the proposed solution (Equations 29, 30,
31, and 32) and considering a consolidation coefficient cv = 5 m2/h, an effective friction angle ϕ0 = 20 , and the Ran-
kine active earth pressure coefficient of K = Ka. One sees that an increase in the filling rate has effect to slow down the
excess PWP dissipation, resulting in low effective stresses and high total stresses at the end of stope filling.
Figure 8 shows the variation of the effective (Figure 8A) and total (Figure 8B) stresses along the height of the stope
at the end of filling for different effective friction angles of backfill ϕ0 , obtained by applying the proposed solution
(Equations 29, 30, 31, and 32) and considering a filling rate of m = 0.2 m/h, a consolidation coefficient of cv = 5 m2/h,

F I G U R E 6 Variation of the
horizontal and vertical effective (A) and
total (B) stresses along the height of the
stope at the end of filling for different
consolidation coefficients cv, calculated
with the proposed solution
(Equations 29, 30, 31, and 32) by
considering H = 40 m, B = 6 m,
m = 0.2 m/h, γ sat = 20 kN/m3, γ w = 10
kN/m3, ϕ0 = 20 , and K = Ka

F I G U R E 7 Variation of the
horizontal and vertical effective (A) and
total (B) stresses along the height of the
stope at the end of filling for different
filling rates m, calculated with the
proposed solution (Equations 29, 30, 31,
and 32) by considering H = 40 m,
B = 6 m, cv = 5 m2/h, γ sat = 20 kN/m3,
γ w = 10 kN/m3, ϕ0 = 20 , and K = Ka
ZHENG ET AL. 11

F I G U R E 8 Variation of the
horizontal and vertical effective (A) and
total (B) stresses along the height of the
stope at the end of filling, calculated
with the proposed solution
(Equations 29, 30, 31, and 32) by
considering H = 40 m, B = 6 m,
m = 0.2 m/h, cv = 5 m2/h, γ sat = 20
kN/m3, γ w = 10 kN/m3, and K = Ka

and the Rankine active earth pressure coefficient of Ka. One sees that the effective and total stresses decrease as the fric-
tion angle ϕ0 increases from 10 to 30 . These results indicate that the increase in the friction angle mainly has effect to
increase the arching effect and thus reduce the effective and total stresses.
Figure 9 shows the variation of the effective (Figure 9A) and total (Figure 9B) stresses along the height of the stope
at the end of filling for different stope widths B, obtained by applying the proposed solution (Equations 29, 30, 31, and
32) as considering a filling rate of m = 0.2 m/h, a consolidation coefficient of cv = 5 m2/h, an effective friction angle
ϕ0 = 20 , and an active earth pressure coefficient of K = Ka. One sees that the effective and total stresses decrease as the
stope width decreases. These results are straightforward. With a small stope width, the degree of arching effect
increases, resulting in low effective and total stresses.

5 | DISCUSSION

A new solution has been proposed to evaluate the pressures and stresses in backfilled stopes during and at the end of
stope filling operation with an impervious base. The consolidation and arching effect during the placement of backfill
have been taken into account. The proposed solution has been validated by numerical modeling with Plaxis2D. It thus
constitutes a simple and useful tool to evaluate the pressures and stresses in backfilled stopes and on barricades at the
preliminary stage of projects. However, the proposed solution was developed with several simplifications and limita-
tions. For instance, the excess PWP distribution was first calculated using the Gibson18 model. The ensuing excess PWP
solution was then taken into account in the arching solution to evaluate the effective and total stresses. The influence
of the arching effect on the dissipation of the excess PWP was not taken into account. In reality, the process of drainage
and consolidation and that of arching effect may influence each other during the placement of slurried backfill. The
drainage and consolidation lead to the generation of arching effect, which tends to prevent the settlement of backfill
particles and slow down the process of drainage and consolidation. Slower drainage and consolidation result in slower
increase in the effective stress development, which in turn favors the drainage and consolidation. A coupled process

F I G U R E 9 Variation of the
horizontal and vertical effective (A) and
total (B) stresses along the height of the
stope at the end of filling, calculated
with the proposed solution
(Equations 29, 30, 31, and 32) by
considering H = 40 m, m = 0.2 m/h,
cv = 5 m2/h, ϕ0 = 20 , γ sat = 20 kN/m3,
γ w = 10 kN/m3, and K = Ka
12 ZHENG ET AL.

takes place between the arching effect and self-weight consolidation. More work is necessary to account for this coupled
process in the estimation of the distribution and evolution of pressures and stresses in backfilled stopes.
It should be noted that the proposed solution is developed by considering a constant filling rate with uncemented
backfill. In practice, the filling operation is usually divided into a plug pour and a final pour to avoid the generation of
excessive backfill pressure on barricades.26,46 The proposed solution can be used to evaluate the pressures and stresses
during the plug pour, but not those of the final pour. Nevertheless, the proposed solution can be used to evaluate if the
filling stage division is necessary or what is the maximum height of the plug pour for a given barricade.
In the proposed solution, the chemical shrinkage associated with the cementation and hydration of cemented back-
fill was not considered. When the cement content of the backfill is low (3 to 7% in many mines in Canada), the effect of
cementation and hydration during the filling operation may be insignificant as shown by the field stress and pressure
measurement of Thompson et al.26 as well as the laboratory test results of Beya.47 The drainage and consolidation along
with the arching effect are the dominant processes to be considered in pressure and stress estimation for barricade
design. The proposed solution can be useful to evaluate the pressures and stresses in backfilled stopes. When the
cement content is high, cementation and hydration may take place during the filling operation. The neglect of the
cement hydration may lead to an overly conservative barricade design. In the future work, more work is necessary to
take into account the cement hydration in the pressure and stresses estimation.
Same as the traditional consolidation theory of Terzaghi,38 the Gibson18 model is based on the linear theory, includ-
ing the linear relationship between the void ratio and effective stress, infinitesimal strain, constant hydraulic conductiv-
ity, constrained modulus, and consolidation coefficient. For mining backfills, the void ratio, the permeability, and the
constrained modulus can change with the drainage and consolidation. The consolidation coefficient cv is thus rarely
constant. More work is required to take into account the variation of the mechanical and hydraulic properties during
the drainage and consolidation.
As the drainage and consolidation take place, the top surface of the placed backfill will settle down, resulting in
accumulated ponding on top of the backfill.48–50 In this study, the hydraulic boundary condition along the top of each
newly added backfill was assumed to be permeable under zero PWP in Plaxis2D. This may not fully represent the field
conditions where ponding water can be observed. Jaouhar and Li51 applied impervious boundary condition along the
top surface of the backfill to reproduce the accumulated ponding. Unrealistic nonzero total stresses were observed along
the top surface of the backfill after the empty of the top surface. Jaouhar and Li51 also used pervious hydraulic bound-
ary condition along the top surface of the backfill and reservoir elements (finite elements proposed by Chapuis52 to sim-
ulate the full and empty conditions of a pipe or reservoir, which has a very high hydraulic conductivity and very low
suction at any degree of saturation) above the tope surface of the backfill. The problem of nonnegligible residual total
stress along the top surface of the backfill after the empty of the top surface remains. More work is required to consider
more representative hydraulic boundary conditions and to produce more realistic numerical results in the future.
When a high water content backfill is placed in the stope, the dissipation of the excess PWP can begin with the sedi-
mentation (i.e., no effective stress) and followed by the consolidation. This has been confirmed by laboratory (e.g., Li
et al.16) and in situ measurements (e.g., Thompson et al.26). The sedimentation is not considered in the Gibson18 model.
More work is needed to take this aspect into account in the future.
Another limitation is related to the one-dimensional drainage and consolidation. The stope base and side walls were
considered to be impermeable. In reality, lateral drainage can occur through the fractures and discontinuities. Two- or
three-dimensional consideration of the problem is necessary.
Finally, the proposed solution is developed for mine stopes with an impervious base. In practice, drainage can take
place downward through porous barricades at the base of stopes. A solution has been developed for estimating the
stresses in backfilled stopes with a pervious base.53

6 | C ON C L U S I ON

In this paper, a new solution has been proposed to estimate the pressures and stresses in mine backfilled stopes with an
impervious base by considering the combined effects of the process of drainage and consolidation and that of arching
effect during the placement of slurried cohesionless backfill. As a special case when the hydraulic conductivity is very
high or/and when the filling rate is very small, the proposed solution reduces to the solution of Li and Aubertin36 that
was developed for estimating the stresses in backfilled stope with hydrostatic water pressure. The proposed solution
was validated by numerical modeling with Plaxis2D. It can thus be used as a simple tool to evaluate the pressures and
ZHENG ET AL. 13

stresses in backfilled stope during and at the end of stope filling at the preliminary stage of projects. Sample applications
of the proposed solution have shown that the drainage and consolidation of backfill can be favored with high consolida-
tion coefficient cv and low filling rate m, resulting in high effective stress and low total stress in the backfilled stope.
With increased backfill effective friction angle or decreased stope width, low effective and total stresses can be
expected.

ACK NO WLE DGE MEN TS

The authors would like to acknowledge the financial support from the Natural Sciences and Engineering Research
Council of Canada (NSERC 402318), Fonds de recherche du Québec - Nature et Technologies (FRQNT 2015-MI-
191676), Mitacs Elevate Postdoctoral Fellowship (IT12573), and industrial partners of the Research Institute on Mines
and the Environment (RIME UQAT-Polytechnique; http://rime-irme.ca/).

ORCID
Jian Zheng https://orcid.org/0000-0002-6730-6464

R EF E RE N C E S
1. Hassani F, Archibald J. Mine Backfill. (CD-ROM). Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, 1998.
2. Kump D. Backfill—whatever it takes. Miner Eng. 2001;53(1):50-52.
3. Jung SJ, Biswas K. Review of current high density paste fill and its technology. Miner Resour Eng. 2002;11(2):165-182.
4. Cui L, Fall M. Modeling of self-desiccation in a cemented backfill structure. Int J Numer Anal Methods Geomech. 2018;42(3):558-583.
5. Aubertin M, Bussiere B, Bernier L. Environnement et gestion des rejets miniers. (CD-ROM). Presses Internationales Polytechnique,
Montreal (in French), 2002.
6. Bussière B. Colloquium 2004: hydrogeotechnical properties of hard rock tailings from metal mines and emerging geoenvironmental
disposal approaches. Can Geotech J. 2007;44(9):1019-1052.
7. Simms P, Grabinsky M, Zhan GS. Modelling evaporation of paste tailings from the Bulyanhulu mine. Can Geotech J. 2007;44(12):
1417-1432.
8. Grice T. Stability of hydraulic backfill barricades. In: Bloss M, ed. Proc., the 6th International Symposium On Mining with Backfill.
Carlton, Victoria, Australia: AusIMM; 1998:117-120.
9. Sivakugan N, Rankine K, Rankine R. Permeability of hydraulic fills and barricade bricks. Geotech Geol Eng. 2006a;24(3):661-673.
10. Sivakugan N, Rankine RM, Rankine KJ, Rankine KS. Geotechnical considerations in mine backfilling in Australia. J Clean Prod. 2006b;
14(12):1168-1175.
11. Sivakugan N, Rankine K, Lovisa J, Hall W. Flow rate computations in hydraulic fill mine stopes. Indian Geotech J. 2013;43(3):195-202.
12. Revell MB, Sainsbury DP. Paste bulkhead failures. Proc. the 9th International Symposium on Mining with Backfill (Minefill'07),
Montréal, Quebec, Canada, 2007.
13. Yumlu M, Guresci M. Paste backfill bulkhead monitoring—a case study from Inmet's Cayeli mine. Proc., the 9th International
Symposium in Mining with Backfill (CD-ROM), Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, 2007.
14. Pedroni L. Étude expérimentale et numérique de la sédimentation et de la consolidation des boues de traitement des eaux acides. Ph.D.
thesis, École Polytechnique de Montréal, Montreal, Canada, 2011.
15. Li L. Beneficial experience from teaching and education to research and development. Creat Educ. 2013;3(7):148-153.
16. Li L, Alvarez IC, Aubertin JD. Self-weight consolidation of a slurried deposition: tests and interpretation. Int J Geotech Eng. 2013;7(2):
205-213.
17. El Mkadmi N, Aubertin M, Li L. Effect of drainage and sequential filling on the behavior of backfill in mine stopes. Can Geotech J. 2014;
51(1):1-15.
18. Gibson RE. The progress of consolidation in a clay layer increasing in thickness with time. Geotechnique. 1958;8(4):171-182.
19. Fahey M, Helinski M, Fourie A. Consolidation in accreting sediments: Gibson's solution applied to backfilling of mine stopes.
Géotechnique. 2010;60(11):877-882.
20. Doherty JP. A numerical study into factors affecting stress and pore pressure in free draining mine stopes. Comput Geotech. 2015;63:
331-341.
21. Shahsavari M, Grabinsky M. Cemented paste backfill consolidation with deposition-dependent boundary conditions. Proc., the 67th
Canadian Geotechnical Conference. CGS, Regina, Saskatchewan, Canada, 2014.
22. Shahsavari M, Grabinsky M. Mine backfill porewater pressure dissipation: numerical predictions and field measurements. Proc. the 68th
Canadian Geotechnical Conference. CGS, Quebec City, Canada, 2015.
23. Wood MD, Doherty JP, Walske ML. Deposition and self-weight consolidation of a shrinking fill. Géotech L. 2016;6(1):72-76.
24. Zheng J, Li L, Mbonimpa M, Pabst T. An analytical solution of Gibson's model for estimating the pore water pressures in accreting depo-
sition of Slurried material under one-dimensional self-weight consolidation. Part I: pervious base. Ind Geotech J. 2018a;48(1):72-83.
25. Zheng J, Li L, Mbonimpa M, Pabst T. An analytical solution of Gibson's model for estimating pore water pressures in accreting deposi-
tion of slurried material under one-dimensional self-weight consolidation. Part II: impervious base. Ind Geotech J. 2018b;48(1):188-195.
14 ZHENG ET AL.

26. Thompson BD, Bawden WF, Grabinsky MW. In situ measurements of cemented paste backfill at the Cayeli mine. Can Geotech J. 2012;
49(7):755-772.
27. Doherty JP, Hasan A, Suazo GH, Fourie A. Investigation of some controllable factors that impact the stress state in cemented paste back-
fill. Can Geotech J. 2015;52(12):1901-1912.
28. Askew JE, McCarthy PL, Fitzerald DJ. Backfill research for pillar extraction at ZC/NBHC. Proc., Mining with Backfill: 12th Canadian
Rock Mechanics Symposium, CIM, Montreal, 1978, 12, 100–110.
29. Knutsson S. Stresses in the hydraulic backfill from analytical calculations and in-situ measurements. In: Stephansson O, Jones MJ, eds.
Proc., on the Application of Rock Mechanics to Cut and Fill Mining. London: Institution of Mining and Metallurgy; 1981:261-268.
30. Aubertin M. Application de la Mécanique des Sols pour l'Analyse du Comportement des Remblais Miniers Souterrains. Short Course
(unpublished notes), Proc., the 14th Colloque en Contrôle de Terrain, Val-d'Or, 23-24 mars 1999. Association Minière du Québec
31. Aubertin M, Li L, Arnoldi S, Belem T, Bussière B, Benzaazoua M, Simon R. Interaction between backfill and rock mass in narrow stopes.
Proc., Soil and Rock America 2003, Vol. 1, Edited by Culligan PJ, Einstein HH, and Whittle AJ, eds., Verlag Glückauf Essen (VGE),
Essen, Germany, 2003, 1157–1164.
32. Li L, Aubertin M, Simon R, Bussière B, Belem T. Modeling arching effects in narrow backfilled stopes with FLAC. Proc., the 3rd Interna-
tional FLAC Symposium, A. A. Balkema, Rotterdam, Netherlands, 2003, 211–219.
33. Li L, Aubertin M, Belem T. Formulation of a three dimensional analytical solution to evaluate stress in backfilled vertical narrow open-
ings. Can Geotech J. 2005;42(6):1705-1717. (with Erratum 43(3):338–339)
34. Pirapakaran K, Sivakugan N. Arching within hydraulic fill stopes. Geotech Geol Eng. 2007a;25(1):25-35.
35. Pirapakaran K, Sivakugan N. A laboratory model to study arching within a hydraulic fill stope. Geotech Test J. 2007b;30(6):496-503.
36. Li L, Aubertin M. Influence of water pressure on the stress state in stopes with cohesionless backfill. Geotech Geol Eng. 2009a;27(1):1-11.
37. Li L, Aubertin M. A three-dimensional analysis of the total and effective stresses in submerged backfilled stopes. Geotech Geol Eng.
2009b;27(4):559-569.
38. Terzaghi K. Theory of Consolidation. New York: Wiley; 1943.
39. Holtz RD, Kovacs WD. An Introduction to Geotechnical Engineering. Civil engineering and engineering mechanics series. Englewood
Cliffs, New Jersey: Prentice-Hall Inc; 1981.
Ð +∞
40. Goodwin ET. The evaluation of integrals of the form −∞ f ðx Þe −x dx. In: Proc., Cambridge Philosophical Society. Cambridge University;
2

1949:241-245.
41. CGS. Canadian Foundation Engineering Manual. 4th ed. Canadian Geotechnical Society; 2006.
42. Das BM. Principles of Geotechnical Engineering. 7th ed. Cengage Learning; 2010.
43. Jaky J. Pressure in silos. Proc., the 2nd International Conference on Soil Mechanics and Foundation Engineering, Rotterdam, The
Netherlands, 21–30 June 1948.
44. Brinkgreve RBJ, Engin E, Swolfs WM. PLAXIS 2014. Delft, Netherlands: PLAXIS bv; 2014.
45. Herget G. Stresses in rock. A.A. Balkema, Rotterdam, Netherlands: Rotterdam; 1988.
46. Li L. Analytical solution for determining the required strength of a side-exposed mine backfill containing a plug. Can Geotech J. 2014;
51(5):508-519.
47. Beya FK. Étude du transfert de chaleur dans les remblais en pâte cimentés curant sous les conditions aux frontières des chantiers miniers
dans le pergélisol. Canada: Master's thesis, École Polytechnique de Montréal; 2016.
48. Dalcé JB, Li L, Yang PY. Experimental study of uniaxial compressive strength (UCS) distribution of hydraulic backfill associated with
segregation. Minerals. 2019;9(3):147.
49. Potvin Y, Thomas EG, Fourie AB. Handbook on Mine Fill, Australian Centre for Geomechanics (ACG). Australia: Perth; 2015.
50. Yang PY, Li L. Evolution of water table and pore-water pressure in stopes with submerged hydraulic fill. Int J Geomech. 2017;17(9):
04017052.
51. Jaouhar EM, Li L. Effect of drainage and consolidation on the pore water pressures and total stresses within backfilled stopes and on
barricades. Adv Civ Eng. 2019;2019:1802130.
52. Chapuis RP. Numerical modeling of reservoirs or pipes in groundwater seepage. Comput Geotech. 2009;36(5):895-901.
53. Zheng J, Li L, Li YC. Total and effective stresses in backfilled stopes during the fill placement on a pervious base for barricade design.
Minerals. 2019;9(1):38.

How to cite this article: Zheng J, Li L, Li Y-C. A solution to estimate the total and effective stresses in
backfilled stopes with an impervious base during the filling operation of cohesionless backfill. Int J Numer Anal
Methods Geomech. 2020;1–17. https://doi.org/10.1002/nag.3079
ZHENG ET AL. 15

A P P END I X A .: Process for the deduction of Equation 27

Equation 26 can be rewritten as follows:

0
ð 
0 −2Ktanϕ
σv = e B l
f ðlÞdl + A , ðA1Þ

where


0 du 2Ktanϕ0 l
f ðl Þ = γ − e B : ðA2Þ
dl

Assuming,
ð
0
F ðlÞ = f ðlÞ or f ðlÞdl = F ðlÞ + C: ðA3Þ

Substituting Equation A3 into Equation A1 leads to

0
0 − 2Ktanϕ
σv = e B l
½F ðlÞ + C + A: ðA4Þ

Considering the boundary condition, l = 0, σ 0 v = 0, one can obtain

0 = F ð0Þ + C + A ) F ð0Þ = −C −A: ðA5Þ

Substituting Equation A5 into Equation A4 leads to

0
0 −2Ktanϕ
σv = e B l
½F ðlÞ− F ð0Þ: ðA6Þ

With Equation A3, one can obtain

ðl
f ðlÞdl = F ðlÞ −F ð0Þ: ðA7Þ
0

Substituting Equation A7 into Equation A6 leads to

0 ðl 0 ðl

0 −2Ktanϕ
− 2Ktanϕ 0 du 2Ktanϕ0 l
σv = e B l
f ðlÞdl = e B l
γ− e B dl: ðA8Þ
0 0 dl
16 ZHENG ET AL.

A P P END I X B.: MATLAB program for solving Equation 32

A MATLAB program is given below to solve Equation 32. Sample calculation is conducted with the proposed program
by considering a cohesionless backfill having a saturated unit weight of γ sat = 20 kN/m3 and filled to a final height of
H = 40 m. The other parameters are considered as B = 6 m, m = 0.2 m/h, cv = 5 m2/h, γ w = 10 kN/m3, ϕ0 = 20 , and
K = Ka.

%===========================================================================%
% A solution to estimate the total and effective stresses in backfilled stopes with an impervious base during the
filling operation of cohesionless backfill %
%===========================================================================%

nmax= 101; % number of division of H (here nmax= 101)

time = 200; % time of filling (here t = 200 h);
m = 0.2; % filling rate (here m = 0.2 m/h);
H = m*time; % final thickness of the backfill (here H = 40 m);
c_v = 5; % consolidation coefficient of the backfill (here cv = 5 m2/h);
h_0 = 0.3; % step interval of y (see Fig. 1);
gama_sat = 20; % saturated unit weight of the backfill (here γ sat = 20 kN/m3);
gama_w = 10; % unit weight of water (γ w = 10 kN/m3);
gama_sub = gama_sat-gama_w; % submerged unit weight of the backfill;
B = 6; % width of the mine stope (here B = 6 m)
Phi = 20; % effective internal friction angle of the backfill (here ϕ’ = 20 );
K_a = (1-sind (Phi))/(1+sind (Phi)); % Rankine's active earth pressure coefficient of the backfill (used in this
sample calculation);
K_0 = 1-sind (phi); %Jaky's at-rest earth pressure coefficient (not used in this sample calculation);
A = 2*K_a*tand (Phi)/B;

% calculate the excess PWP, u, in Eq.(10)

for i = 1:nmax
l(i) = H/(nmax-1)*(i-1);
y(i) = H-l(i);
u(i) = 0.0;
for j = -91:91;
temp1 = m*j*h_0*sqrt (time)/sqrt(c_v);
temp2 = y(i)*j*h_0/sqrt(c_v*time);
u(i) = u(i)+h_0/2.*4.*c_v*time*j*h_0*tanh (temp1)*cosh (temp2)*exp(-(j*h_0)^2);
end
pre_multi = gama_sub/sqrt (pi*c_v*time)*exp(-(y(i))^2/4./c_v/time);
u(i) = (gama_sub*H-pre_multi*u(i));
end

% calculate the du/dy in Eq.(14)

for i = 1:nmax
l(i) = H/(nmax-1)*(i-1);
y(i) = H-l(i);
derive_u(i) = 0.0;
for j = -91:91;
temp1 = m*j*h_0*sqrt (time)/sqrt(c_v);
temp2 = (y(i))*j*h_0/sqrt(c_v*time);
temp3 = j*h_0/sqrt(c_v*time);
temp4 = -(y(i))/2./c_v/time;
a = tanh (temp1)*cosh (temp2)*exp(-(j*h_0)^2);
ZHENG ET AL. 17

b = tanh (temp1)*sinh (temp2)*(j*h_0/sqrt(c_v*time))*exp(-(j*h_0)^2);

derive_u(i) = derive_u(i)+temp4*h_0/2.*4.*c_v*time*j*h_0*a+h_0/2.*4*c_v*time*j*h_0*b;
end
pre_multi = -gama_sub/sqrt (pi*c_v*time)*exp(-(y(i))^2/4./c_v/time);
derive_u(i) = pre_multi*derive_u(i);
end

% calculate the horizontal total stress in Eq.(32)

for i = 1:nmax
l(i) = H/(nmax-1)*(i-1);
n_z = i;
sigma_sub(i)=0.0;
for k = 1:n_z
delta_x = l(i)/(n_z);
sigma_sub(i) = sigma_sub(i)+(gama_sub+derive_u(k))*exp(A*(k*delta_x))*delta_x;
end
sigma_sub(i) = exp(-A*l(i))*sigma_sub(i);
sigma_H(i)=K_a*sigma_sub(i)+gama_w*l(i)+u(i);
end
% results output and saved in the file “Output filename.txt”
fid = fopen('Output filename.txt','w')
for i=1:nmax
fprintf (fid,'%g %g\r\n',l(i),sigma_H(i))
end
fclose (fid)

plot (sigma_H,l,'-.');
set (gca,'YDir','reverse');