Tunnelling and Underground Space Technology 71 (2018) 505–517

Contents lists available at ScienceDirect

Tunnelling and Underground Space Technology

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

Groundwater Seepage Rate (GSR); a new method for prediction of MARK

groundwater inflow into jointed rock tunnels
⁎
Mahdi Rasouli Maleki
Senior Advisor of Water & Power Resources Development Co. (IWPCO) and Head of Engineering Geology Department at Harazrah Consulting Engineers Group, Tehran,
Iran

A R T I C L E I N F O A B S T R A C T

Keywords: The main purpose of this research is to introduce a new method for estimating the groundwater inflow rate into
Groundwater Seepage Rate (GSR) tunnels excavated in rock environments. The Groundwater Seepage Rate (GSR) is a novel analytical method that
Groundwater inflow has the capability to assess the rock mass potential in conducting the groundwater into tunnels. Geological and
Hydraulic conductivity hydraulic parameters and radius of the tunnel are the main parameters used in the GSR method. In this method,
Fracture and joint sets
geological parameters are defined based on the characteristics of the joint sets including the strike, number,
3D dimensional
spacing and aperture of joints of each joint set. In addition, hydraulic head and hydraulic conductivity are other
Elastic behavior
principle inputs of the proposed model. In GPR method, the rate of groundwater inflow has been evaluated in 3D
dimensions and efforts were made to obtain acceptable values by creating rational correlations among input
data. The taking account of joints condition in tunnel and the direct effect of tunnel radius, the separate study of
the joint sets and the role of fracture systems in the groundwater conductivity into excavated openings are the
main advantages of this model. According to the results of this study, GSR method can provide better estimations
of the inflow volume for rock masses with elastic behavior in which fracture systems have been developed.
Finally, in the end this paper for validation of GSR values, the results of this method were compared with the
obtained results of empirical methods and observed groundwater inflow in various geological units of the Zagros
tunnel in Iran. Additionally, the calculated value of groundwater inflow based on GSR method shows good
compatibility with empirical methods.

1. Introduction groundwater inflow into tunnels.

Groundwater inflows during construction pose one of the greatest
risks to the successful completion of tunnel projects. Tunnels con-
structed below the water table are exposed to some level of risk asso-
ciated with water inflows with magnitude of highly variable risk.
Groundwater inflows may constitute a potential hazard as well as an
important factor controlling the rate of advancement in driving a
tunnel.
Indeed, some of the most disastrous experiences in tunneling have
been the result of interception of large flows of water from highly
fractured water-saturated rocks (Freeze and Cherry, 1979). As for the
discontinuous rock masses, the water flow is strongly dependent upon
the hydraulic characteristics of joint sets and rocks fracturing condi-
tions (Gattinoni et al., 2009).
Since the presence of joint and fracture systems in rock masses
around the tunnel is the main factor affecting the amount of ground-
water conducted into tunnel, the use of engineering characteristics of
joints and fractures can be a proper way of estimating the quantity of
In this research, efforts were made to represent a new analytical
method, known as Groundwater Seepage Rate (GSR) via using joint and
fracture characteristics as well as the hydraulic parameters of the rock
mass. In GSR model, the amount of water inflow is estimated by ap-
plying aperture surface area of joints of rock mass with respect to the
tunnel surface area. Finally, compared with the results of empirical
methods, the results of this method for different units of Zagros tunnel
can be more credible.
2. A review of previous methods
So far many researchers all over the world have presented different
equations for calculating the extent of groundwater discharge into un-
derground openings. Most of these equations are based on one of the
empirical, analytical or numerical methods.
The empirical method is a procedure adopted by Heuer (1995) and
subsequently developed by Raymer (2001). This method is based on the
experiences obtained from other tunnels, excavated in similar

⁎
Corresponding author. Tel.: +98 914 400 3015; fax: +98 21 88768555.
E-mail address: Mahdi.Rasouli@yahoo.com.

http://dx.doi.org/10.1016/j.tust.2017.10.006
Received 25 December 2016; Received in revised form 17 September 2017; Accepted 18 October 2017
Available online 02 November 2017
0886-7798/ © 2017 Elsevier Ltd. All rights reserved.
M. Rasouli Maleki
hydrogeological conditions and on the basis of available hydro-
geological information (in situ tests, chemical and physical measures,
etc.) (Dematteis et al., 2007).
The analytical approach is based on the application of Dupuit’s
formula, with corrections to the geometry of aquifer layers, the intake
area and permeability reduction associated with hydraulic head varia-
tions and the effective stress around deep tunnels.
Water inflow into tunnels can be modeled using numerical methods
and the seepage into tunnels can be calculated according to various site
conditions. These methods require comprehensive data regarding the
site condition. Numerical methods are quite complex and their appli-
cation is time consuming, however, the results are more precise in
comparison to analytical methods, particularly when the tunnel is ex-
cavated in fractured rock mass and the impact of geo-structural aniso-
tropy of fractured rocks on tunnel inflows needs to be addressed.
In both of the empirical and analytical methods, the equivalent
hydraulic conductivity obtained from packer test is applied for pre-
dicting the amount of water inflow into tunnel.
Polubarinova-Kochina (1962), Goodman et al. (1965), Lohman
(1972), Herth and Arndts (1973), Custodio (1983), Zhang and Franklin
(1993), Heuer (1995), Knutsson et al. (1996), El Tani (1999), Raymer
(2001), Karlsrud (2001), Ribacchi et al. (2002), Cesano et al. (2003),
Perrochet and Dematteis (2007), Hwang and Lu (2007) and Park et al.
(2008) have introduced different ways for estimating the groundwater
flow into tunnels.
Goodman’s formula is probably the most commonly applied ap-
proximation for quick estimation of water inflow rates. It also serves as
the basis of the empirical method calculations proposed by Heuer
(1995). It should be noted that relations obtained for prediction of
tunnel water discharge rates by Goodman et al. (1965), Chisyaki (1984)
and El Tani (2003) are developed according to analytical solutions for
steady-state final values, and those adopted by Perrochet (2005a,
2005b), Perrochet and Dematteis (2007) are for transient inflows.
Moreover, Jang et al. (1996) carried out a groundwater flow analysis
based on 2D fracture network.
3. Groundwater Seepage Rate (GSR)
The main prerequisite for estimating the water inflow rate into the
tunnels excavated in rock masses with many joints is adequate knowl-
edge upon joint and fracture systems of tunnel excavation perimeter.
Since the rate of amount of entrance water from a discontinuity always
depends on the aperture surface area, it can be concluded that by de-
termining the total aperture surface area of joints surrounding the
tunnel, a rather precise estimation of water inflow rate into the tunnel
can be obtained.
GSR model is a method based on the following data: geology, hy-
drology and the tunnel radius. The main purpose for introducing this
analytical model is to provide a somewhat accurate estimation of the
water inflow rate into tunnels being excavated in rocks based on joint
and fracture characteristics of the excavation area. One of the ad-
vantages of this method is taking account of each joint system with its
own geometrical features including the height, width and length
through the calculation process.
In this proposed model, it is assumed that a tunnel with circular
cross section is excavated in an isotropic media with even permeability
coefficient. The main input parameters used in this model are the
geological and hydraulic parameters as well as the radius of the tunnel
(Table 1). Among the geological parameters, aperture and spacing of
joints are of crucial importance, since they have direct relationships
with the amount of water.
Unlike other models, in GSR model, the cross section from which
water passes into the tunnel is assumed equivalent to the aperture
surface area of joint sets available in the rock mass. Accordingly, for
determining the cross section of joints in tunnel circumference, firstly
by determining the number of joints (NJ), aperture surface area (AJ) in
506
Tunnelling and Underground Space Technology 71 (2018) 505–517
Table 1
The effective parameters in assessment inflow water into tunnel based Groundwater
Seepage Rate (GSR).
1. Geological parameters 2. Hydraulic parameters 3. Tunnel properties
• Strike of joints, α • Hydraulic heads, H • Radius of tunnel, r
• Number of joints, NJ • Hydraulic conductivity, k
• Spacing of joints, SJ
•
Aperture of joints, AJ
each joint set and the total aperture area of joints (Atotal) around the
tunnel is calculated. Then the rate of water inflow into the tunnel can be
estimated in a form of analytical relationships via simple correlation
between these data and hydraulic parameters.
The groundwater head (h) and hydraulic conductivity (k) of rock
masses surrounding the tunnel play an important role for determining
the water inflow rate. Due to the variability of these parameters along
the tunnel route, it is necessary to divide the tunnel path into separate
zones with similar structural and hydraulic characteristics for more
accurate evaluations. In order to distinguish each zone from others, one
should bear in mind specific criteria such as lithological changes, per-
meability variations with respect to lugeon value in boreholes, hy-
draulic head of groundwater changes and so on.
In GSR, the water inflow rate into the tunnel can be calculated in
three different scenarios:
1. Groundwater inflow into the whole zones of the tunnel.
2. Groundwater inflow into each zone of the tunnel.
3. Groundwater inflow into drilling effective length of the tunnel.
3.1. Geological parameters
As stated earlier, the main purpose of applying geological para-
meters is to involve their vital effect through the water inflow evalua-
tion procedure. Indeed, this important factor has not been adequately
considered in other models. The analysis and assessment results using
the GSR model can subsequently be compared with the actual condi-
tions of the excavated zone.
Based on GSR model, the angle between the joint strike and tunnel
axis (α), number (NJ), spacing (SJ) and aperture (AJ) of joints in each
joint set are the main geological parameters involved in estimation of
water inflow into tunnels. To sum up, the main purpose of involving
geological parameters in GSR model is to determine the aperture sur-
face area of total joints with respect to the surrounding area of tunnel
(i.e. to determine the parameter of reduction (n) of joints and fractures
around the tunnel). Fig. 1 shows the orientation, spacing and number of
joint sets with respect to tunnel axis.
3.1.1. Angle between joint strike and tunnel axis (α)
The intersection of each joint with the tunnel path could form cir-
cular, elliptical or rectangular shapes in the cross section of a tunnel.
The angle between joint strike and tunnel axis is used for determining:
1) the number of joints of tunnel and 2) the radius of the circular or
elliptical shape or large length of rectangular shape.
3.1.2. Number of joints (NJ)
The development of joints around the tunnel results in a porous
media with high permeability. Therefore, it could be mentioned that
this geological parameter is an important factor which facilitates the
groundwater inflow into the tunnel. As a result, this parameter has been
applied in GSR model due to its significant role for accurate estimation
of water inflow.
3.1.3. Spacing of joints (SJ)
A lower joint spacing means a greater number of joints, a lower rock
mass quality and a higher permeability in the tunnel surface. The joint
M. Rasouli Maleki Tunnelling and Underground Space Technology 71 (2018) 505–517

Fig. 1. The orientation, spacing and number of joint sets with respect to tunnel axis.

spacing as another geological parameter is used to determine the Where C is the ellipsoid circumference (m); a and b are the one-half
number of joints in each joint set in the whole tunnel path which is of the ellipse’s major and minor axis (m), respectively.
defined using Eq. (1) as follows. In GSR method, to calculate a parameter, can be used Eq. (4) and b is
equal to radius tunnel (m).
L sin(α ) ⎤ L sin(α ) ⎤ L sin(α ) ⎤
NJ 1 = ⎡ , NJ 2 = ⎡ . .. NJn = ⎡ 2
⎢ SJ 1 ⎦
⎣ ⎥ ⎢ SJ 2 ⎦
⎣ ⎥ ⎢
⎣ SJn ⎦ ⎥ (1) 1 2 1 ⎛ 2r ⎞
a= (L + 2r 2)0.5 = ⎜ + 2r 2⎟

Where NJ is the number of joints; SJ is the spacing of joints (m); α is the
angle between strike of joints and tunnel axis (°); L is the length of each
zone (m) and J1, J2… Jn are the joint sets.
3.1.4. Aperture of joints (AJ)
According to the proposed GSR method, the distance between the
walls of joints known as the aperture is directly related to the rate of
water inflow into the tunnel. In this model, it is presumed that the total
aperture surface area in tunnel circumference equals to the cross-sec-
tion from which water flows into the tunnel. Accordingly, considering
the field aperture of joints (αF) in each joint set and the circumference
of the shape resulted from the intersection of joints with the tunnel, the
2 2 ⎝ sin(α ) ⎠ (4)
Where a is the one-half of the ellipse’s major axis (m); L is the length
of each zone (m); r is the tunnel radius (m) and α is the angle be-
tween the strike of joint and tunnel axis (°).
c. In case the joint strike is parallel to the tunnel axis (α ∼ 0), i.e. if the
angle between joint strike and tunnel axis is smaller than the critical
angle (α < θ), (Eq. (5)), the formed shape from their intersection
will not be elliptical any more. Instead, a rectangular shape is cre-
ated in which one side has been excavated. Therefore, its cir-
cumference can be calculated via Eq. (6).
2r
θ = Arc tan ⎛ ⎞ ⎜ ⎟

aperture surface area of joints (αJ) in tunnel circumference can be es- ⎝L⎠ (5)
timated for each joint set. Where θ is the critical angle (°); r is the radius of circle (m); L is the
According to different possible angles between the joint strike and length of each zone (m).
the tunnel axis (α), different shapes will be resulted as a consequence of
C = 2a + 2b or C = 2L + 2b (6)
the intersection of joints strike and the tunnel axis (Fig. 2).
Where C is the rectangular circumference (m); a and b are one-half
a. Provided that the joint strike is perpendicular to the tunnel axis of the rectangular major and minor dimensions (m), respectively.
(α = 90), the created shape from their intersection will be circular
and the circumference can be calculated via Eq. (2). By considering the circumference of the resulted shape and the
aperture of joints in the field (aF ), the aperture surface area (aJ ) of each
C = 2πr (2)
joint set can be calculated via Eq. (7).
Where C is the circle circumference (m) and r is the radius of circle (m).
aF (J 1) × C(J 1) ⎞ aF (J 2) × C(J 2) ⎞ aF (Jn) × C(Jn)
b. If the joint strike is oblique to the tunnel axis (0 < α < 90), the aJ 1 = ⎛
⎜ ⎟ , aJ 2 = ⎛ ⎜ ⎟ . .. aJn = ⎛
⎜ ⎟
⎞
⎝ 1000 ⎠ ⎝ 1000 ⎠ ⎝ 1000 ⎠
resulted shape from their intersection will be elliptical. Since the
calculation of elliptical circumference involves using complicated (7)
integral equations, it is recommended to apply Eq. (3) for calcu- Where αJ is the aperture surface area of joints (m2); aF (J ) is the aperture
lating the elliptical circumference in a more simplified way of joints in the field (mm) and CJ is the obtained circumference of shape
(Almkvist and Berndt, 1988). from intersection of joint strike with tunnel axis (m).
In order to calculate the aperture surface area of all joints in a specific
2
⎛
C = π (a + b) ⎜1 +
a−b
3 a+b ( ) ⎞
⎟
zone with length of (L), a new parameter called total aperture surface
area (AJ) is defined. This parameter can be calculated via Eq. (8).
a−b 2
⎜
⎝
10 + 4−3 ( ) a+b
⎟
⎠ (3) AJ 1 = NJ1 × aJ 1, AJ 2 = NJ2 × aJ 2 . .. AJn = NJn × aJn (8)

507
M. Rasouli Maleki
Fig. 2. Various orientations of joint strike to tunnel axis and the possible shapes created from their intersection.
Fig. 3. Changes of hydraulic head via water seepage into tunnel over time (from a to f).
Where AJ is the total aperture surface area (m2); NJ is the number of
joints; aJ is the aperture surface area of joints (m2).
Lastly, the sum of total aperture surface area (AJ) of several joint
sets can be considered equal to the total aperture surface area of all
joint sets (AJ (total)) around the tunnel according to Eq. (9).
AJ (total) = ∑ (AJ 1 + AJ 2 + ...+Ajn ) (9)
Where AJ (total) is the total aperture surface area of all joint sets (m2).
3.2. Hydraulic parameters
Based on conducted studies by numerous researchers (Jacob and
Lohman, 1952; Goodman, 1965; El Tani, 2003; Perrochet (2005a,
2005b) and Park et al., 2008), it can be perceived that the hydraulic
head (H) and permeability coefficient (k) of rock mass are other ef-
fective factors for estimating the rate of water inflow into the tunnel.
These characteristics of rock mass are highly affected by groundwater
level, characteristics of joint and fracture systems, crushed intensity of
rock mass etc.
Piezometering and carrying out pumping test (lugeon test) in
508
Tunnelling and Underground Space Technology 71 (2018) 505–517
explorative boreholes are among the most efficient ways for de-
termining the hydraulic heads and permeability coefficient of the
tunnel, respectively. But, in the absence of lugeon test results, the
proposed equation by Barton (2008) can be employed to determine the
permeability coefficient (k) of rock mass. This equation is confirmed by
the Dekchu and Bhasmey hydroelectric projects in Sikkim and in the
Himalayas.
0.002 ⎞ 100 1
k≈⎛ ⎜ ⎛⎟ ⎞⎛
5/3
⎞
⎝ Q ⎠ ⎝ JCS ⎠ ⎝ H ⎠ (10)
Where k is the permeability of the rock mass (m s−1), Q is the in situ
rock mass quality (Q = 0.1 to 100) that can be calculated by (RQD/Jn)
(Jr/Ja) (Jw/SRF), JCS is the joint wall compressive strength (MPa), and
H is the depth of a specific point under consideration below the ground
surface (m).
3.3. Tunnel properties
Another effective parameter for estimating the quantity of water in-
flow is the radius of underground structure. The effect of this parameter
M. Rasouli Maleki Tunnelling and Underground Space Technology 71 (2018) 505–517

Fig. 4. Excavation of tunnel with fixed (a) and cone (b) hydraulic head.

Fig. 5. The effective discharge length (Le) with regards to the tunnel face advance.

has been considered directly in the geological parameters to estimate the Q = AV = Aki (11)
resulted shape from intersection of joint sets with tunnel section. 3 −1
Where Q is the groundwater inflow (m s ); V is the discharge
velocity (m s−1); A is the surface area (m2); k is the hydraulic con-
4. Calculation of groundwater inflow rate (Q) ductivity (m s−1); i is hydraulic gradient (without unit).

In the Groundwater Seepage Rate (GSR) model, the flow of
groundwater is presumed to be governed by Darcy’s law (1856), which
states that the velocity of the flow is directly proportional to the per-
meability coefficient and hydraulic gradient. By defining Q as the total
volume of flow per unit time through a cross-sectional area A, Darcy’s
law would be as follows in Eq. (11):
4.1. The maximum volume of water inflow into tunnel in the each zone
(QMax)
Based on GSR model, for estimating the maximum volume of water
inflow, the area of cross section from which water flows into the tunnel
is considered equal to the total aperture surface area (AJ (total)) in the

509
M. Rasouli Maleki
Fig. 6. Correlation between the steady state inflow and equivalent permeability (Heuer, 2005).
excavation length of tunnel. Therefore, the maximum amount of water
inflow into the tunnel, in a condition that the whole tunnel path is
excavated under a fixed hydraulic head, is considered equal to the
amount of water flown into the tunnel through all joints of tunnel path .
It can be defined as Eq. (12).
QMax = AT nkH or QMax = AJ (total) kH
Where QMax is the maximum value of water inflow into tunnel in
each zone (m3 s−1 in L zone); AT is the surface area of tunnel (m2); n is
the fracture ratio (without unit); AJ (total) is the total aperture surface
area of all joint sets (m2); k is the hydraulic conductivity (m s−1) and H
is the distance from the groundwater level in the tunnel (m).
4.2. The volume of water inflow into unit length of tunnel (Qu)
The amount of water inflow into unit length of a tunnel can be
obtained by dividing the maximum volume of water inflow (QMax) by
the tunnel length in each zone (L). (Eq. (13)).
QMax H
Qu = or Qu = AJ (total) K
L L
Where Qu is the groundwater inflow for unit length of tunnel (m3
s−1) and L is the tunnel length (m).
4.3. Actual groundwater inflow into tunnel (QActual)
In addition to the geological and hydraulic parameters (ground-
water head and permeability) of rock mass in the tunnel path, effective
discharge length is also among the main factors affecting the calcula-
tion process of the proposed GSR model.
4.3.1. Fixed and variable hydraulic heads
One of the most important challenges presented by different re-
searchers in the water inflow calculation process is the lack of sufficient
Tunnelling and Underground Space Technology 71 (2018) 505–517
attention towards fluctuation of water level while excavating tunnels.
In other words, the hydraulic head is not fixed and varies as water is
drained during tunnel excavation process (Fig. 3).
Therefore, it should be noted that the amount of water inflow ob-
tained from Eq. (12) cannot be actual unless the whole tunnel path is
excavated under a fixed hydraulic head (Fig. 4 – groundwater surface (a
(12)
and Le-a)) . However, this condition is not plausible. It is due to the fact
that in the first stages of excavation, the depression cone is formed
around the tunnel and brings about variability in the hydraulic head
and the effective discharge length (Le) (Fig. 4 –groundwater surface (b
and Le-b)). As a result, it can be inferred that the results obtained from
Eq. (14) indicate the maximum capability of the surrounding rock mass
to discharge water into the tunnel. Accordingly, the main advantage of
Eq. (14) is to estimate the maximum amount of water inflow into the
tunnel.
4.3.2. The effective discharge length (Le)
All in all, the quantity of water inflow into tunnels is highly affected
by different factors. One of the effective factors is the part of tunnel
(13)
length which is situated below the groundwater table, which depends
on the ratio of excavation speed to drainage rate into the tunnel.
In GSR method for determining the logical volume of water inflow
into tunnel the effective discharge length (Le) is defined. It is worth
mentioning that effective discharge length (Le) is highly dependent
upon the excavation speed and permeability coefficient (k) of the sur-
rounding rock mass that can be a constant value in isotropic environ-
ments. As the excavation operation progresses, the specific effective
discharge length (Le) also advances forward (See Fig. 5). It means that
in the high excavation rate the and low permeability coefficient (k) of
rock mass the effective discharge length (Le) will have a higher value.
Therefore, it should be noted that the calculated quantity of water
based on the Le is the maximum expected inflow of water into the
tunnel.
510
 Table 2a
Basic input parameters for calculation of groundwater inflow into Zagros tunnel in accordance with GSR method.

Zone ROCK MASS 1- GEOLOGICAL PARAMETERS

M. Rasouli Maleki

Name Tunnel length Length of zone Overburden, H Rock mass quality, Q Joint wall compressive Joint Sets Azimuths or Spacing, SJ Aperture, αF Number of joints
(m) (m) (m) strength, JCS (MPa) Direction to (min–max (min–max in zone, NJ
North (°) (mean)), (m) (mean)), (mm)

P1 0–753 753 80–100(90) 3.1–5.4(4.25) 5–25(15) J1 155 0.17–0.37(0.27) 1.4–1.6(1.5) 2682

(SH-ML1) J2 270 0.12–0.22(0.17) 0.7–0.8(0.75) 2910
J3 300 0.2–0.5(0.35) 0.35–0.65(0.5) 2035
J4 40 1.2–1.8(1.5) 0.2–0.8(0.5) 78
Bedding 41 1.15–1.55(1.35) 2.8–3.2(3) 78
G1 2265–3715 1450 100–140(1 2 0) 0.5–0.9(0.7) 25–50(37.5) J1 127.5 1.3–1.7(1.5) 0.35–0.75(0.5- 947
5)
(SH-LS1,4) J2 230 0.45–0.75(0.6) 1.35–1.65(1.5) 46
J3 135 0.8–1.1(0.95) 0.4–0.7(0.55) 1523
J4 250 0.2–0.4(0.3) 1.4–1.6(1.5) 1738
Bedding 24 0.6–0.9(0.75) 3.85–4.15(4) 817
Ki2 4000–4715 715 200–250(2 2 5) 0.01–0.05(0.03) 60–100(80) J1 60 0.65–0.85(0.75) 3.4–3.6(3.5) 182
(LI2) J2 170 0.4–0.6(0.5) 1.4–1.6(1.5) 1227
J3 140 0.55–0.95(0.75) 3.3–3.7(3.5) 953
J4 135 1.4–2.1(1.75) 3.65–4.35(4) 408
Bedding 125 0.25–0.45(0.35) 0.4–0.6(0.5) 1982
P5 5855–6110 225 165–180(172.5) 0.1–0.18(0.14) 40–60(50) J1 310 0.6–0.9(0.75) 3.35–3.65(3.5) 296
(ML-SH5) J2 140 0.35–0.65(0.5) 1.35–1.65(1.5) 450
J3 305 0.65–0.85(0.75) 3.4–3.6(3.5) 291
J4 210 1.45–2.05(1.75) 3.7–4.3(4) 42
Bedding 50 0.3–0.4(0.35) 0.45–0.55(0.5) 11

511
P6 7235–7350 115 165–205(1 8 5) 0.1–0.05(0.075) 25–100(62.5) J1 310 0.5–0.6(0.55) 0.5–0.6(0.55) 207
(ML-SH5) J2 14 0.42–0.52(0.47) 0.5–0.6(0.55) 140
J3 305 0.06–0.08(0.07) 1.24–1.26(1.2- 1594
5)
Bedding 210 0.25–0.35(0.3) 1.45–1.55(1.5) 125
G5 9415–9910 495 210–230(2 2 0) 0.1–0.5(0.3) 25–50(37.5) J1 50 0.2–0.3(0.25) 0.65–0.75(0.7) 35
(SH-LS4) J2 240 0.08–0.18(0.13) 0.65–0.75(0.7) 732
J3 205 0.39–0.41(0.4) 1.99–2.01(2) 505
Bedding 50 0.3–0.4(0.35) 0.65–0.75(0.7) 25

Zone 1- GEOLOGICAL PARAMETERS 2- HYDRAULIC PARAMETERS 3- TUNNEL PROPERTISE Groundwater inflow into tunnel, Qt (L s-1)

Name Aperture area Total Fracture ratio, n Hydraulic head, Hydraulic Radius, r (m) Axis Orientation - For each joint sets For unit length For total zone
surface of one aperture area (%) H (m) conductivity, Trend (°)
joint, AJ (mm2) surface of one k (m s-1)
joint, AJ(total)
(m2)

P1 24460 65.6 0.85% 65 3.90E−06 3.36 49 0.022 0.042 31.5

(SH-ML1) 13923 40.5 0.014
7974 16.2 0.005
22362 1.8 0.001
147686 11.5 0.000
G1 8667 8.2 0.62% 110 6.60E−06 0.004 0.054 78.3
(SH-LS1,4) 432696 19.9 0.010
8612 13.1 0.007
38309 66.6 0.033
98557 80.5 0.000
Ki2 136922 24.9 1.08% 205 1.65E−04 1.178 6.941 4963.0
Tunnelling and Underground Space Technology 71 (2018) 505–517

(LI2) 25435 31.2 1.476 (continued on next page)

M. Rasouli Maleki Tunnelling and Underground Space Technology 71 (2018) 505–517

4.3.3. The actual and expected water inflow into tunnel (QActual)
In the GSR model, actual water inflow into tunnel (QActual) can be
For total zone
calculated via multiplying the amount of water inflow for the unit
length of the tunnel (Qu) by the effective discharge length (Le) ac-

50.6

37.4

86.7
cording to Eq. (14).
H
Q Actual = Le QU = Le ⎛AJ (total) k ⎞
⎝ L⎠ (14)
Groundwater inflow into tunnel, Qt (L s-1)

For unit length

Where QActual is the actual groundwater inflow into tunnel in the

effective discharge length (m3 s−1 m); Qu is the groundwater inflow for
unit length of tunnel (m3 s−1); AJ (total) is the total aperture surface area
0.225

0.325

0.175
of all joint set (m2) and Le is the effective discharge length under
groundwater table (m).
The obtained equations for estimating the water inflow into tunnel
For each joint sets

can be summarized as Eq. (15).

SH-ML: Shale, Marly Limestone (Pd formation), SH-LS: Shale, Shaly Limestone, Sandstone (Gu formation), LI: Limestone (LI formation), ML-SH: Marly Limestone, Shale (Pd formation).
a ×C × sin (α )J1
⎛ F (J 1) (J 1) +⎞
1000SJ 1
⎜a ⎟
2.848
1.439
0.001
0.080
0.046
0.079
0.020
0.000
0.014
0.012
0.258
0.041
0.023
0.061
0.074
0.017
F (J 2) × C(J 2) × sin (αJ 2)
+⎟
Q Actual =⎜ 1000SJ 2 kHLe
⎜ ...+ ⎟
⎜ aF (Jn) × C(Jn) × sin (αJn) ⎟
⎜ ⎟
Axis Orientation -

⎝ 1000SJn ⎠
(15)
Trend (°)

Where QActual is the actual groundwater inflow into tunnel in the ef-
3- TUNNEL PROPERTISE

fective discharge length (m3 s−1 m); aF (J ) is the aperture of joints in the
field (mm); aJ the is the aperture surface area of joints (m2); CJ is the
obtained circumference of shape from intersection of joint strike with
Radius, r (m)

tunnel axis (m); SJ is the spacing of joints (m); k is the hydraulic con-
ductivity (m s−1); L is the tunnel length (m) and H is the distance from
the groundwater level in the tunnel (m) and J1, J2… Jn are the joint
sets.

5. Comparison of the GSR results with the empirical method

representing the actual condition (Case Study: Zagros Tunnel)
conductivity,

6.60E−06

6.60E−06

9.24E−06
Hydraulic

k (m s-1)
2- HYDRAULIC PARAMETERS

5.1. Heuer’s empirical method

Heuer (1995, 2005) introduced an empirical method to estimate the

water inflow based on borehole water pressure test results using
Hydraulic head,

Goodman’s model with a reduction factor to calculate for many of the

idealizations associated with the use of this equation. The method
presents an estimation of the proportion of tunnel length that will ex-
H (m)

145

140

165

perience different intensities of long term steady-state inflow. It also

provides an estimation of possible initial heading inflow by steady-state
inflow per unit tunnel length times a heading inflow factor.
Fracture ratio, n

Heuer's approach incorporates for this disagreement by grouping

the rock permeability data in “bins” corresponding to various propor-
tions of tunnel length and choosing an inflow intensity curve, flow over
1.15%

1.67%

0.54%

headthe ratio of flow rate to hydraulic head q/Hs (Fig. 6), relationship
(%)

based on the ratio of hydraulic head to tunnel radius. The variance in

the measured packer hydraulic conductivity values is considered to
surface of one
aperture area

show the distribution of hydraulic conductivity along the tunnel path.

joint, AJ(total)
1- GEOLOGICAL PARAMETERS

5.2. Validation of the proposed method

Total

(m2)

60.2
30.4
15.7
18.8
10.9
18.5

32.1

19.7
24.2
4.7
1.8
1.8
1.5

5.1
7.6

5.4

In this section, GSR method results have been compared with the
inflow rate obtained by the empirical method (Heuer, 1995) and actual
joint, AJ (mm2)
surface of one
Aperture area

condition in the Zagros tunnel. This tunnel is located in the western part
of Iran and is under construction to convey water from Sirvan river
112237
156879

219632

219632
63149
74653

63313
24154
63558

10857
20130
40542

26858
47894

basin to Zahab plain. The approximate length of this tunnel is

7901

8642

50 km that has been divided into two lots of (I) and (Π) with length of
Table 2a (continued)

24 and 26 km, respectively. An approximate length of 23 km from lot

(Π) of this tunnel has been excavated by a Tunnel Boring Machine
(ML-SH5)

(ML-SH5)

(SH-LS4)

(TBM). From the geological aspect, the region which is being excavated
Name

is located in Zagros Folded zone based on the structural divisions of

Zone

G5
P5

P6

Iran. The main lithology of this region consists of a sequence of Jurassic

512
 Table 2b
The remaining contents of Table 2a.

Zone ROCK MASS 1- GEOLOGICAL PARAMETERS

M. Rasouli Maleki

Name Tunnel length Length of zone Overburden, H Rock mass quality, Q Joint wall compressive Joint Sets Azimuths or Spacing, SJ Aperture, αF Number of joints
(m) (m) (m) strength, JCS (MPa) Direction to (min–max (min–max in zone, NJ
North (°) (mean)), (m) (mean)), (mm)

G8 13290–13425 135 170–230(2 0 0) 0.40 25–50(37.5) J1 30 0.35–0.55(0.45) 1.9–2.1(2) 98

(SH-LS4) J2 355 0.2–0.3(0.25) 0.95–1.05(1) 437
J3 145 0.55–0.75(0.65) 0.9–1.1(1) 207
Bedding 130 0.5–0.6(0.55) 1.95–2.05(2) 242
Ki7 14250–15065 815 170–190(1 8 0) 0.45 75–125(1 0 0) J1 65 0.2–0.3(0.25) 1.95–2.05(2) 898
(LI2) J2 0 0.3–0.6(0.45) 1.85–2.15(2) 1366
J3 0 1.8–2.2(2) 1.8–2.2(2) 307
J4 70 0.55–0.65(0.6) 0.95–1.05(1) 487
Bedding 340 0.74–0.76(0.75) 0.99–1.01(1) 1014
Ki10 17400–18030 630 180–190(1 8 5) 0.26 75–125(1 0 0) J1 50 0.45–0.55(0.5) 6.45–6.55(6.5) 22
(LI2) J2 30 2.2–2.8(2.5) 6.2–6.8(6.5) 82
J3 330 1.55–1.95(1.75) 6.3–6.7(6.5) 353
Bedding 170 1.3–1.7(1.5) 0.3–0.7(0.5) 360
Ki12 18045–18640 595 170–190(1 8 0) 0.06 100–150(1 2 5) J1 50 0.65–0.85(0.75) 3.4–3.6(3.5) 14
(LI2) J2 210 0.45–0.55(0.5) 1.45–1.55(1.5) 389
J3 150 0.7–0.8(0.75) 3.45–3.55(3.5) 779
Bedding 170 1.55–1.95(1.75) 3.8–4.2(4) 292
Ki21 24680–25683 1003 180–220(2 0 0) 1.25 75–100(87.5) J1 50 1.1–1.4(1.25) 1.85–2.15(2) 14
(LI2) J2 60 0.55–0.75(0.65) 1.9–2.1(2) 294
J3 115 2.2–2.8(2.5) 6.2–6.8(6.5) 366
Bedding 140 2.2–2.8(2.5) 1.7–2.3(2) 401

513
Zone 1- GEOLOGICAL PARAMETERS 2- HYDRAULIC PARAMETERS 3- TUNNEL PROPERTISE Groundwater inflow into tunnel, Qt (L s-1)

Name Aperture area Total aperture area Fracture ratio, n (%) Hydraulic head, Hydraulic Radius, r (m) Axis Orientation - For each joint sets For unit length of For total zone
surface of one surface of one joint, H (m) conductivity, k (m Trend (°) tunnel
joint, AJ (mm2) AJ(total) (m2) s-1)

G8 40656 4.0 0.46% 185 5.20E−05 3.36 49 0.283 0.9 125.6

(SH-LS4) 12751 5.6 0.397
7830 1.6 0.115
7850 1.9 0.136
Ki7 91534 82.2 0.79% 165 1.45E−04 2.413 3.24 2640.2
(LI2) 9593 13.1 0.385
26687 8.2 0.241
13989 6.8 0.200
24702 25.1 0.001
Ki10 1255291 27.6 0.51% 145 4.62E−04 2.934 7.20 4538.6
(LI2) 97498 8.0 0.850
24297 8.6 0.913
65458 23.6 2.507
Ki12 1255291 17.4 0.63% 145 4.75E−04 2.011 9.19 5468.4
(LI2) 13436 5.2 0.605
63397 49.4 5.716
25435 7.4 0.859
Ki21 1098326 15.4 0.32% 175 4.20E−04 1.127 4.99 5000.2
(LI2) 157143 46.2 3.389
8078 3.0 0.217
8604 3.5 0.253
Tunnelling and Underground Space Technology 71 (2018) 505–517

SH-LS: Shale, Shaly Limestone, Sandstone (Gu formation), LI: Limestone (LI formation).
M. Rasouli Maleki Tunnelling and Underground Space Technology 71 (2018) 505–517

Actual groundwater inflow into tunnel

and Cretaceous formations. These units contain limestone and shale
belonging to Ilam, Gourpi and Pabdeh formations (Rasouli Maleki and
Yousefi Mahmoud, 2015). As for the hydrogeological point of view,
only the data of groundwater inflow pertaining to a length of 13 km at
the end of the tunnel is recorded with acceptable precision.

for Le ∼ 50 m (L/s)

157.5–232.5(1 9 5)
In this research, investigations were conducted in four steps: In the

295–440(367.5)
10.5–14.5(12.5)
first step, 13 zones were selected randomly among excavated zones of

260–350(3 0 5)

215–355(2 8 5)

165–255(2 1 0)
37.5–42.5(40)
3.5–9.5(6.5)
5.5–10.5(8)
the Zagros tunnel. In the second step, geological and hydraulic para-

1–2(1.5)
1–2(1.5)
meters in the GSR method were determined based on joint mapping
from rock outcrops in the field and water pressure tests (lugeon test) in
the boreholes, respectively. The magnitude of water inflow was calcu-
lated based on GSR and Heuer methods for unit length and 50 m length

Actual Groundwater inflow for

unit length of tunnel (L/s/m)
of tunnel in the third step. The details and results of these steps are
presented in Tables 2a, 2b and 3, respectively
Finally, at the last step, water inflow rate into the tunnel was measured

Observed (Actual)
during excavation of every 50 m length in each zone of the tunnel path.

0.02–0.04(0.03)
0.02–0.04(0.03)

0.07–0.19(0.13)
0.11–0.21(0.16)
0.21–0.29(0.25)
0.75–0.85(0.8)
3.15–4.65(3.9)
Subsequently, the average measured values considered as actual volume

5.9–8.8(7.35)
4.3–7.1(5.7)

3.3–5.1(4.2)
5.2–7(6.1)
inflow into the zone (In the P1 zone with 735 m length, for example, the
first zone was divided into 15 parts (L = 50 m) and then the amount of
water inflow was measured and recorded for each part and the average of
these parts were finally determined as actual inflow volume into the P1
zone). Fig. 7 shows the water inflow volume measured for each of the zones.

tunnel for Le ∼ 50 m (L/s)

Groundwater inflow into
Summarily, the comparisons were drawn between the conducted
calculations according to Table 3. The results indicate that:

A. The values of calculated groundwater via both GSR and empirical

methods for most geological units of tunnel path are compatible

119.58

137.50
217.50
229.58
240.63
33.92
with each other. But for some cases, differences were observed

0.98
2.57

3.38
3.27
5.50
which are due to the effect of geological parameters in the GSR
Empirical method (Heuer, 1995)

model contrary to other methods.

Groundwater inflow for unit

B. In the limestone units (Ki2, Ki7, Ki10 and Ki12), the magnitude of
length of tunnel (L/s/m)

water inflow volume calculated using GSR method is higher than the
empirical method. it can be argued that in rock masses with elastic
behavior (such as limestone), more actual results can be obtained by
taking the role of joint and fracture systems into consideration. This
fact is verified according to a drawn comparison between the water
0.02
0.05

2.39
0.07
0.07
0.11
0.68
2.75
4.35
4.59
4.81
inflow values calculated by the GSR method and actual observed
ones for Ki units (Fig. 8).
tunnel for Le ∼ 50 m (L/s)
Groundwater inflow into

Groundwater inflow rate for the unit length of Zagros tunnel based
Groundwater inflow Calculated into the Zagros tunnel by dint of GSR and empirical methods.

on GSR and empirical methods are presented in Fig.8. It can be easily

perceived from this figure that the values of calculated groundwater for
P1, G1, P5, G5 and G8 units by dint of both methods are compatible
with each other. Moreover, the obtained results using GSR method for
347.1

162.0
360.2
459.5
249.3

the Ki2 (LI2), Ki7 (LI2), Ki10 (LI2), Ki12 (LI2) and Ki21 (LI2) units
11.2
16.3

46.5
2.1
2.7

8.8

prove that the amount of water inflow into the tunnel is at a critically
high level and can be dangerous for the personnel and also hinder the
Groundwater inflow for unit

excavation process.
length of tunnel (L/s/m)

Fig.9 presents the results of actual observed water inflow rate and
those calculated using GSR and empirical methods. The drawn com-
parison between these three methods demonstrates the capability of the
proposed GSR method for estimating the groundwater inflow rate into
Groundwater Seepage Rate (GSR)

the tunnel. Indeed, the values obtained via GSR method are close to
0.04
0.05

6.94
0.22
0.33
0.18
0.93
3.24
7.20
9.19
4.99

actual conditions.

6. Limitation of GSR method

water inflow into
Total volume of

Similar to other methods, the GSR model also has some limitations
tunnel (L/s)

when being applied to the calculation process of groundwater inflow

4963.0

2640.2
4538.6
5468.4
5000.2

into underground spaces. Summarily, these considerations are men-

125.6
31.5
78.3

50.6
37.4
86.7

tioned as follows:
P1 (SH-ML1)

P5 (ML-SH5)
P6 (ML-SH5)
G5 (SH-LS4)
G8 (SH-LS4)

a) The necessity for conducting joint studies with acceptable precision

Ki10 (LI2)
Ki12 (LI2)
Ki21 (LI2)
Ki2 (LI2)

Ki7 (LI2)
G1 (SH-

on rock mass outcrops.

LS1,4)
Table 3

Zone

b) The inability to carry out joint studies on the intense crushed and
faulted regions.

514
M. Rasouli Maleki Tunnelling and Underground Space Technology 71 (2018) 505–517

Fig. 7. The water inflow volume measured for every 50m length of zones.

Fig. 8. The groundwater inflow rate for unit length of the Zagros tunnel (Qu) using GSR method in comparison to the values obtained by empirical methods in different geological zones.

c) The inability to predict the volume of groundwater inflow into the
tunnels containing large and buried fractures.
d) The difficulties for accurate determination of the effective discharge
length (Le) in the tunnel path which requires doing statistical ana-
lyses and creating logical relationship on the obtained data from
different tunnels around the world. This can be fulfilled in the
coming years.
7. Conclusion
Groundwater Seepage Rate (GSR) is a new method introduced to
estimate the amount of groundwater flow into tunnels excavated in
rock. The main inputs in this method are geological and hydraulic
parameters of the tunnel path. The GSR method is generally based on
the effect of rock mass joints and fractures in conducting the

515
M. Rasouli Maleki
Fig. 9. The observed water inflow rate into Zagros tunnel in comparison with the obtained results using GSR and empirical methods for excavated units.
groundwater flow into tunnels. Attempts were made to correlate geo-
logical and hydraulic parameters of rock mass experimentally and lo-
gically in order to estimate the amount of water flow into tunnels. By
taking advantage of this method, new parameters including the total
aperture surface area (AJ), the effective discharge length (Le) and rock
mass hydraulic parameters are defined by which the magnitude of
water inflow can be calculated with acceptable precision. An analysis
on the magnitude of groundwater flow in different geological units of
Zagros tunnel verifies the compatibility of estimations made by GSR
method with empirical methods and the actual conditions of tunnel. To
conclude, the main advantage of GSR method is that it is capable of
considering the effect of joints condition in the calculations leading to
more actual results. Additionally, the potential risks associated with
water ingress into the tunnels can be predicted and dealt with accord-
ingly.
Author's Note and Reminders
In the end, I would like to thank the enthusiastic researchers in the
field of groundwater in underground spaces and the related topics. It
should be noted that GSR model is a recently adopted method for es-
timating the groundwater inflow in tunnels which is introduced for the
first time in this paper. Therefore, it can be improved by similar in-
vestigations by interested researchers throughout the world in the
coming years.
References
Almkvist, G., Berndt, B., 1988. Gauss, Landen, Ramanujan, the Arithmetic-Geometric
Mean, Ellipses, π, and the Ladies' Diary. Am. Math Monthly 95, 585–608.
Barton, N., 2008. Training course on rock engineering. Organized by ISRMTT & CSMRS,
Course Coordinator Rajbal Singh, December 10–12, New Delhi, India, p. 502.
Cesano, D., Bagtzoglou, A.C., Olofsson, B., 2003. Quantifying fractured rock hydraulic
heterogeneity and groundwater inflow prediction in underground excavations: the
heterogeneity index. Tunn. Undergr. Space Technol. 18, 19–34.
Chisyaki, T., 1984. A study of confined flow of ground water through a tunnel. Ground
Water 22 (2), 162–167.
Custodio, E., 1983. Hidraulica de captaciones de agua subterránea. In: Custodio, E.,
Llamas, M.R. (Eds.), Hidrologı́a SubterráNea. Ediciones Omega, Barcelona, pp.
516
Tunnelling and Underground Space Technology 71 (2018) 505–517
614–695 (In Spanish).
Darcy, H., 1856. Les Fontaines Publiques De La Ville De Dijon. Dalmont, Paris, France.
Dematteis, A., Torrl, R., and Looser, M., 2007. Water Resources Management in
Tunneling: insights in the decisionmaking process to improve tunnels environmental
sustainability. SEA Consulting S.r.l., Via Cernaia 27, 10121 Torino, Italy.
El Tani, M., 1999. Water inflow into tunnels. World Tunnel Congress – Challenges for the
21st Century.
El Tani, M., 2003. Circular tunnel in a semi-infinite aquifer. Tunn. Groundwater Space
Technol. 18, 49–55.
Freeze, R.A., Cherry, J.A., 1979. Groundwater. Prentice-Hall, New Jersey, pp. 604.
Gattinoni, P., Scesi, L., Terrana, S., 2009. Water flow in fractured rock masses: numerical
modeling for tunnel inflow assessment, Geophysical Research Abstracts, vol. 11,
EGU2009-468, 2009, EGU General Assembly 2009.
Goodman, R.E., 1965. Ground water inflow into a tunnel drive. Bull. Assoc. Eng. Geol. 2
(1), 41–56.
Goodman, R.E., Moye, D.G., Van Schalkwyk, A., Javandel, I., 1965. Ground water inflow
during tunnel driving. Eng. Geol. 2, 39–56.
Herth, W., Arndts, E., 1973. Theorie Und Praxis Der Grundwasserabsenkung.
Ernst & Sohn, Berlin.
Heuer, R. E., 1995. Estimating Rock Tunnel Water Inflow. RETC Proceedings, p. 41–60
(Chapter 3).
Heuer, R.E., 2005. Estimating rock tunnel water inflow, Rapid Excavation and Tunneling
Conference, pp. 394–407 (Chapter 30).
Hwang, J.-H., Lu, C.-C., 2007. A semi-analytical method for analyzing the tunnel water
inflow. Tunn. Undergr. Space Technol. 22, 39–46.
Jacob, C.E., Lohman, S.W., 1952. Nonsteady flow to a well of constant drawdown in an
extensive aquifer. Trans. Am. Geophys. Union 33 (4), 559–569.
Jang, H.I., Chang, K.M., Lee, C.I., 1996. Groundwater flow analysis of discontinuous rock
mass with probabilistic approach. In: Proc. Korea-Japan Joint Symp. Rock Eng.,
Seoul, Korea, pp. 519–523.
Karlsrud, K., 2001. Water control when tunneling under urban areas in the Olso region.
NFF Publication 12 (4), 27–33.
Knutsson, G., Olofsson, B., Cesano, D., 1996. Prognosis of groundwater inflows and
drawdown due to the construction of rock tunnels in heterogeneous media. Res. Proj.
Rep. Kungl Tekniska, Stokholm.
Lohman, S.W., 1972. Ground-Water Hydraulics. U.S. Geological Survey Prof. Paper 708,
US Government. Printing Office, Washington, DC.
Park, K.-H., Owatsiriwong, A., Lee, G.-G., 2008. Analytical solution for steady-state
groundwater inflow into a drained circular tunnel in a semi-infinite aquifer: a revisit.
Tunn. Undergr. Space Technol. 23, 206–209.
Perrochet, P., 2005a. Confined flow into a tunnel during progressive drilling: an analy-
tical solution. Ground Water 43 (6), 943–946.
Perrochet, P., 2005b. A simple solution to tunnel or well discharge under constant
drawdown. Hydrogeol. J. 13, 886–888.
Perrochet, P., Dematteis, A., 2007. Modelling transient discharge into a tunnel drilled in
heterogeneous formation. Ground Water 45 (6), 786–790.
Polubarinova-Kochina, P.Ya., 1962. Theory of ground water movement. Princeton
University, Princeton.
M. Rasouli Maleki
Rasouli Maleki, M., Yousefi Mahmoud, N., 2015. Detection of hydrogen cyanide (HCN)
gas in tunnels and fundamental strategies to deal with it. Environ. Earth Sci. 74 (9),
6919–6927.
Raymer, J.H., 2001. Predicting groundwater inflow into hard-rock tunnels: estimating the
high-end of the permeability distribution. RETC Proc. 1027–1038 Chapter 83.
517
Tunnelling and Underground Space Technology 71 (2018) 505–517
Ribacchi, R., Graziani, A., Boldini, D., 2002. Previsione degli afflussi d’acqua in galleria e
influenza sull’ambiente”. Meccanica e Ingegneria Delle Rocce 143–199.
Zhang, L., Franklin, J.A., 1993. Prediction of water flow into rock tunnels: An analytical
solution assuming a hydraulic conductivity gradient. Int. J. Rock Mech. Min.
Sci. & Geomech. Abstr. 30 (1), 37–46.