Professional Documents
Culture Documents
Analytical Solution For A Deep Tunnel Excavated in A Porous Elasto-Plastic Material Considering The Effects of Seepage Forces PDF
Analytical Solution For A Deep Tunnel Excavated in A Porous Elasto-Plastic Material Considering The Effects of Seepage Forces PDF
Grasselli)
ABSTRACT: Seepage forces generated during tunnel excavation affect the ground response and
may cause severe instability problems, including total ground collapse and flooding into the
tunnel. This paper presents a new closed-form analytical solution, for stresses and displacements
around a cylindrical tunnel excavated in a porous water-bearing elasto-plastic material, which
includes the effects of seepage forces. The analysis is carried out in terms of effective stresses
and pore water pressures. The solution shows that seepage forces may have a major effect on the
effective stresses around the tunnel, on the extent of the plastic zone, on the radial displacements, and on the tunnel support requirements. The conditions for development of a flowing
ground condition at the tunnel face have been identified. The paper presents a procedure for the
hydromechanical ground-support interaction analysis, to determine the excavation induced inflow and the pressures acting on the lining, throughout the progressive driving of the tunnel, including the effective pressures exerted by the ground and the pressures generated by the water.
1 INTRODUCTION
Tunneling in water bearing ground affects the hydraulic equilibrium of the surrounding
ground leading to seepage into the tunnel. The seepage forces generated in the ground by the
movement of water towards the tunnel can have a significant effect on the behavior of the opening. Groundwater inflow can cause severe instability problems, including total ground collapse
and flooding into the tunnel. Numerous case histories of tunnel failures during construction, at
sections with high water pressures, have been reported (e.g. Marulanda, et al, 2008, Singh et al,
2006). Failure often consists of a sudden mud inrush flowing violently into the opening, dragging equipment, facilities and sometimes even workers, and completely invading long stretches
of tunnel. Even in less striking cases, seepage forces can have a strong effect on the ground support requirements and on the structural requirements of the support system.
Nevertheless, no adequate analytical solutions that consider the effect of seepage forces are
currently available for analyzing the inelastic ground-support interaction to dimension tunnel
support elements, and for properly assessing cases of flowing ground.
A number of researchers have investigated the effect of seepage forces during excavation on
the stress and displacement fields around tunnels. The analysis techniques used by these authors
vary from advanced numerical methods to simplified elastic solutions. Alvarez (1997) used the
coupled hydro-mechanical discrete-finite-element model developed by Barbosa (1990), Shin et
al (2005) used a coupled finite element model with nonlinear permeability, Fernandez and
Moon (2006) used a coupled distinct element model, Lee and Nam (2001, 2004) used uncoupled finite element analyses, and Li (1999) and Bobet (2003) developed elastic closed-form solutions.
In this paper a proper elasto-plastic solution is derived, for a deep cylindrical tunnel excavated in a Mohr-Coulomb perfectly plastic material under the water table, considering seepage
PAPER 4293
ROCKENG09: Proceedings of the 3rd CANUS Rock Mechanics Symposium, Toronto, May 2009 (Ed: M.Diederichs and G. Grasselli)
forces. In addition to establishing the radius of the plastic zone, radial convergence, the stress
and displacement fields around the tunnel, and identifying the condition for development of
flowing ground condition at the tunnel face, a procedure for hydromechanic ground-support interaction is proposed. The basic assumptions are stated first and then the solution is developed.
2 BASIC ASSUMPTIONS
The basic configuration considered in this study is shown in Figure 1. The analysis assumes a
deep cylindrical tunnel of radius a, excavated in a naturally stressed water-bearing ground that
obeys the principle of effective stress. A supporting lining is installed after excavation. A cylindrical coordinate system with the origin at the center of the opening is used in the analysis. The
length of the tunnel is such that the problem can be treated two-dimensionally.
Figure 1. (a) Effective stresses (b) Pore water pressures (c) Transient hydraulic radius of influence.
The horizontal and vertical in situ effective stresses are assumed to be equal and to have a
magnitude P0. The installed support is assumed to exert a uniform radial effective support
pressure Pi on the walls of the tunnel.
The original groundwater pressure is Pw0. The pressure inside the supported tunnel is equal to
the atmospheric pressure. The water pressure at the lining-tunnel wall interface is Pwi. Groundwater inflow is assumed to be radial. It is assumed that the effect of the water-pressure-drop
vanishes beyond a no-flow moving boundary located at a time-dependent distance Rw(t).Beyond
such instantaneous distance, the tunnel has no effect on the original groundwater condition and
the pore water pressure becomes equal to the original ground water pressure.
The original ground is assumed to be a porous liner-elastic material characterized by a
Youngs modulus E, a Poissons ratio , a Darcys coefficient of permeability K, and a specific
storage Ss. The failure characteristics of this material are defined by the Coulomb-Navier equation.
1' = c + 3' .N
N =
1 + sin '
1 sin '
(1)
= c r + .N r
'
1
'
3
N r
1 + sin r'
=
1 sin r'
(2)
It is assumed that the strength reduces suddenly from the peak defined in equation (1) to the
residual defined in equation (2).
PAPER 4293
ROCKENG09: Proceedings of the 3rd CANUS Rock Mechanics Symposium, Toronto, May 2009 (Ed: M.Diederichs and G. Grasselli)
The volumetric response of the material in the plastic zone is controlled by the dilation angle
. If the dilation angle is equal to the friction angle (=r) then the plastic flow rule is said to
be associated and the material undergoes volumetric expansion while it plastifies. If the dilation
angle is less than the friction angle the plastic flow rule is said to be non-associated. If the dilation angle is zero, there is no volumetric change while the material plastifies. The dilation angle
enters the solution through the parameter N that controls the ratio between the circumferential
(or tangential) and the radial components of the plastic strain increment.
plastic
1
=
plastic
N
r
N =
1 + sin '
1 sin '
(3)
3 FUNDAMENTAL RELATIONS
3.1 Equilibrium and compatibility
Figure 2. (a) Equilibrium of an infinitesimal element (b) Water continuity on infinitesimal element.
Figure 2(a) illustrates the stresses acting on a typical ground element. The basic differential
equation of equilibrium for such element is
r, and the third term is the seepage force. The radial and tangential strains for plane strain conditions can be stated in terms of the radial displacement u, as follows
r =
u
r
u
r
(5)
Strain components are related by the strain compatibility equation, which is obtained by
eliminating the radial displacement in Eq. (5)
=0
r
r
(6)
3.2 Continuity
Figure 2(b) illustrates the amount of water that enters a differential element and the amount of
water that leaves the element, per unit time. The difference between these two quantities represents the change in the volume of water stored in the element. For a steady state condition, the
amount of water that leaves the element is equal to the amount of water that enters it. In the
transient stage, the flow rate that exits the element is larger than the flow rate that enters, as
some water stored in voids and joints of the ground is released.
PAPER 4293
ROCKENG09: Proceedings of the 3rd CANUS Rock Mechanics Symposium, Toronto, May 2009 (Ed: M.Diederichs and G. Grasselli)
Using Darcys law to write the flow rates in and out of the element and writing the change in
volume of water in terms of the volumetric specific storage of the ground Ss, the continuity condition yields the following diffusion equation
2 Pw 1 Pw S s Pw
+
=
r r
K t 2
r 2
(7)
The volumetric specific storage is the amount of water per unit volume of a saturated formation that is released, per unit decline in hydraulic head. Specific storage can be expressed in
terms of compressibility of water Cw, compressibility of skeleton Cs, and porosity n, as
S s = w (C s + nC w ) .
4 METHOD OF ANALYSIS
The method of analysis follows the approach used by Perrochet (2004) to derive a transient solution for tunnel discharge under constant drawdown. The analysis is based on the idea that the
transient solution can be computed as successive steady state snapshots using a time dependent
radius of influence, Rw. Therefore, in deriving the solution for stresses and displacements, the
equilibrium equation is integrated, by treating the radius of influence as a known constant distance beyond which the water pressures becomes equal to the original water pressure Pw0. The
transient nature of the solution is represented by the time dependency of Rw, which is determined from the solution of the diffusion equation.
The proposed solution includes coupling between effective stresses and pore water pressures
with ground displacements. However, the solution neglects any variation of the coefficient of
permeability with changes in effective stress. More precisely the analysis assumes that the ratio
Ss/K is constant in time and space. A similar approximation is implicit in the conventional solution for consolidation of soft clays, which is a similar diffusion process involving release of
pore water from saturated ground. The mechanism for pore water release in the later diffusion
process is however different, due to differences in the relative compressibility of the skeleton
and water, and to the magnitude of the water pressures. While during consolidation of clays, the
mechanism for water release is compression of the soil skeleton; the transient flow into a deep
tunnel excavated in a water-bearing rock formation is produced mainly by water expulsed from
the voids due to expansion of pore water, as the pore water pressure is reduced.
5 ELASTIC ANALYSIS
Substituting the stress-strain relations for elastic behavior and the pressure gradient for radial
flow into the equilibrium equation, and integrating the resulting differential equation, yields:
For r Rw:
( Pw0 Pwi )
a2
2
R
r
2(1 ) ln aw
a2 R
R
a2
. 2 ln w + ( 12 )(1 2 ) ln w (8a)
r
a
r
r
( Pw0 Pwi )
a2
+
2
R
r
2(1 ) ln aw
a2 R
R
a2
. 2 ln w + ( 12 )(1 2 ) + ln w (8b)
r
a
r
r
2
(Pw0 Pwi )(1+ )r a2 Rw 1
Rw
(1+ ) '
a2
' a
(P0 Pi ) +
.
ln
(
)(
1
)
(
1
2
)
ln
u=
+
(8c)
2
E
r
r
r2
2E(1 ) ln Raw r 2 a
u=
PAPER 4293
(1 + ).a
( P0' Pi ' ) + ( Pw0 Pwi )
E
(9d)
ROCKENG09: Proceedings of the 3rd CANUS Rock Mechanics Symposium, Toronto, May 2009 (Ed: M.Diederichs and G. Grasselli)
For r Rw:
Rw2
w
r2
R2
' = P0' + ( P0' R' w ) 2w
r
Rw2
(1 + ) '
'
u=
( P0 Rw )
E
r
(9a)
(9b)
(9c)
The set of equations 8 show the detrimental effect of seepage forces on the stresses around
the tunnel. Seepage forces reduce the effective radial stress, which is the minor principal stress,
and increase the tangential stress, which is the major principal stress. Therefore, seepage forces
increase the deviatoric stress r-, around the tunnel, while reducing the effective confining
stress '3.
On the other hand, Eq. 8c shows that even if the effective support pressure is equal to the farfield effective stress, there is still a ground displacement produced by seepage forces. According to Eq. (8d), the displacement at the wall face turns out to be the same as that computed for
the total net pressure. However, the above result is true only for elastic conditions and only at
the wall face.
6 ELASTO-PLASTIC ANALYSIS
The elastic solution is valid for effective support pressures greater than a critical pressure Picrit
for which the effective stresses at the tunnel wall reach the failure criterion of the original
ground. Using Eqs. (8) for r=a and substituting into Eq. 1 yields
'
Picrit
2 P0' c
( Pw0 Pwi )
+
=
N + 1
(1 )( N + 1)
(10)
For effective support pressures lower than the critical, the effective stresses induced in the
ground following excavation will exceed the original ground strength, and a plastic zone of radius Rp will develop around the tunnel. The ground outside the boundary defined by Rp is assumed to remain elastic.
6.1 Stresses and displacements in the elastic zone
The elastic solution derived above is valid in the elastic zone. If the plastic zone is within the
radius of influence of the tunnel (e.g. Rp<Rw) the solution is a set of equations similar to Eqs.
(8), in which the tunnel radius a, is replaced by Rp, and the effective internal pressure Pi, is replaced by the effective radial stress at the elastic/plastic interface, 'rp. On the other hand, if the
plastic zone extends beyond Rw, the solution is a set of equations similar to Eqs. (9), in which
the radial distance Rw is replaced by Rp and the effective stress 'rw is replaced by 'rp.
6.2 Stresses in the plastic zone
Substituting the failure criterion for the material in the plastic zone, into the equilibrium equation, and integrating the resulting differential equation, yields the following stress distribution in
the plastic zone.
'
c*r r Nr 1
c*r
= Pi +
N r 1 a
N r 1
'
r
' = c r + r' .N r
PAPER 4293
(11a)
where c*r = cr
( Pw0 Pwi )
ln Raw
(11b)
ROCKENG09: Proceedings of the 3rd CANUS Rock Mechanics Symposium, Toronto, May 2009 (Ed: M.Diederichs and G. Grasselli)
Equation (11b) shows that the effect of the seepage forces can be visualized as that of reducing
the compressive strength of the ground in the plastic zone.
6.3 Radius of plastic zone
Considering equilibrium of effective radial stresses at the plastic/elastic boundary yields
Rp
Rw
( P w 0 P wi ) ln
Rp
cr*
2 P 0' c
+
+
Rw
N + 1
Nr 1
( N + 1 )( 1 ) ln
a
=
cr*
'
Pi +
Nr 1
(12)
The radius of the plastic zone is determined by solving iteratively Eq. (12). Equation (12)
presumes that the plastic zone is within the radius of influence of the tunnel (e.g. Rp<Rw). If the
plastic zone extends beyond de radius of influence, additional computations are required to determine the actual extent of the plastic zone. First Eq. (8a) is used to compute 'Rw, the effective
radial stress at a Rw. Next a simplified version of Eq. (12) is used to compute the radius of the
plastic zone, in which the effective support pressure Pi is replaced by 'Rw and the water pressure gradient term is ignored. Such equation can be solved directly without any iterative procedure.
6.4 Flowing ground condition
Equation (12) will not converge (i.e. continuity of effective radial stresses at the plastic/elastic
boundary cannot be accomplished) if the water pressure gradient, effective support pressure, and
strength of the ground in the plastic zone are such that
( Pw0 Pwi )
ln
Rw
a
Pi ' ( N r 1) + cr
(13)
For the above condition, which makes the denominator of Eq. (12) negative, the ground cannot
resist the seepage forces and collapses into the tunnel. A flowing ground condition will develop at the tunnel face, with the ground invading the tunnel.
Equation (13) includes three terms, which means there are three ways to handle flowing
ground conditions. (a) Reduce the net water pressure with drainage (or compressed air), increase
effective support pressure by adopting a stiffer support, and (c) increase ground strength with
grouting.
PAPER 4293
ROCKENG09: Proceedings of the 3rd CANUS Rock Mechanics Symposium, Toronto, May 2009 (Ed: M.Diederichs and G. Grasselli)
Rp
u=
r
(1 + ) R p
E
[(P
'
0
'
Rp
) + ( Pw0 Pwp )
(14)
where Rp and Pwp are the effective radial stress and pore water pressures at the plastic/elastic
interface. The above solution allows computing the stress and displacement fields around the
tunnel. Fig. 3 shows de distribution of pore water pressures, effective stresses and displacements
for an application example described in section 9. The effective and water pressures at the wall
face are Pi = 0.84 MPa and Pwi= 1.09 MPa respectively, while the in situ effective stress and
pore pressures are P0 = 4.9 MPa and Pw0 = 3.2 MPa respectively.
7 TRANSIENT ANALYSIS
7.1 Transient radius of influence
The radius of influence Rw, has been treated so far as a constant distance beyond which the pore
water pressure becomes equal to the original water pressure. However, in reality, such radius is
transient in nature. Its initial value is the radius of the opening when the excavation is made, and
it expands radially as time proceeds, until a maximum value is reached, corresponding to the
development of a steady state condition.
Jacob and Lohman (1952) derived the first transient solution of the diffusion equation, for a
tunnel subjected to a sudden, constant hydraulic head. The solution is a complicated one, which
involves first and second kind zero-order Bessel functions. Recently, Perrochet (2005) developed a much simpler analytical solution, which yields essentially the same results. From the
formula for the transient discharge in such solution, and accounting for the presence of a liner,
the following equation for the transient radius of influence is obtained
K ( Pw 0 Pwi )
Rw
= 1+
t
2
a
S s a Pw 0
0 .5
R wmax
a
(15)
where t is the time elapsed since flow started at the section analyzed, and Rwmax is the maximum
radius of influence corresponding to the final steady state condition.
7.2 Progressive excavation
Combining the above elasto-plastic solution and the concept of transient radius of influence, it is
possible to evaluate the variation in time of the tunnel support requirements, extent of the plastic
zone, distributions of pore water pressures, effective stresses, and displacements, during tunnel
excavation, in terms of the tunnel advance rate. The analysis follows the approach used by Perrochet (2005) to determine transient discharge into a tunnel during progressive drilling.
Consider the progressive driving of a tunnel into a permeable zone at an average advance rate v,
as illustrated in Figure 4. At time t, the tunnel face is located at a distance vt, and the time at
which a position x <vt was reached is x/v. Thus, the time elapsed since that position was reached
and during which inflow occurred at that section is t-x/v, which is the time that goes into equation 15. Therefore, the transient radius of influence is a non-uniform function of space over the
distance vt. Thus for a long tunnel, it is possible that by the time the tunnel face is near the end
of the water-bearing zone (x=L), while the transient expansion of the radius of influence is beginning at such location, near the entrance to the water-bearing zone (x=0) the radius of influence may be approaching the steady state maximum.
PAPER 4293
ROCKENG09: Proceedings of the 3rd CANUS Rock Mechanics Symposium, Toronto, May 2009 (Ed: M.Diederichs and G. Grasselli)
Figure 4. Flow regime, radius of influence Rw, and water inflow per unit length of tunnel, for a tunnel
excavated at a constant velocity.
As a result, the radius of influence at the tunnel face, at the point where the support is being installed, and elsewhere, depend on the tunneling rate. If the tunnel is driven fast, the radius of influence along the tunnel will be small, resulting in large hydraulic gradients and seepage forces,
which could cause stability problems at the face. For a slower advance rate, a larger hydraulic
influence zone develops, reducing hydraulic gradients and seepage forces. Therefore, in poor
permeable ground a slow rate of advance is preferred. However, on less pervious soft ground,
with a short-term strength (undrained) greater than the long-term strength (drained), rapid advance improves stability. Analysis for the later case is not considered in this study. In both
cases tunnel support, should be installed as fast as possible to provide early mechanical support.
On the other hand, as shown in Fig. 4, the larger initial heading inflow or flush flow
(Heuer, 1995, Fernandez et al 2006), and the continuous decrease with time of the inflows per
unit length that have been measured at tunnel walls as the steady state condition is approached,
are all consistent with the concept of transient radius of influence. As the radius of influence expands the hydraulic gradient decreases and the inflow reduces as well.
7.3 Steady state radius of influence
Pore pressures, seepage forces, groundwater inflow, and the risk for instabilities are in general
greatest immediately after excavation. The long-term condition would not be critical unless
there is a significant deterioration of ground properties over time (e.g. due to weathering, slaking, creep, etc). Still, the long-term condition may be significant for the tunnel operation in
terms of inflow. For long times, as the radius of influence expands away from the tunnel wall,
the flow deviates from a circular fashion as the steady state condition is approached, and adjust
to the site groundwater boundary conditions. Some possible steady state equilibrium configurations are shown in Figure 5.
The final steady state configuration of the water table depends on the available recharge. In
case (a) recharge is from above and there is an unlimited amount of water available for recharge.
The tunnel does not modify the horizontal water table. This is a case of confined flow where the
ground remains saturated. In case (b) the water table is below the ground surface and the ground
receives limited recharge water directly from the surface, mainly from precipitation. The recharge however is insufficient to maintain a horizontal water table and a significant drawdown
develops. This is a case of unconfined flow where drainage of pore water makes part of the
ground partially saturated. In case (c) the recharge is also limited, but greater than that in case
(b), resulting in a small drawdown. In case (d) recharge is from an aquifer underlying the
ground around the tunnel. Recharge is unlimited and the type of flow is confined.
Based on the flow net, one could estimate a steady state radius of influence Rwmax as a distance beyond which the pore pressure is not affected significantly by the presence of the tunnel.
Typically Rwmax can be estimated as a factor (1 to 2) of the original groundwater pressure head
Pwo/w. Alternatively, an equivalent Rwmax could be computed as the radius of influence for a
symmetrical flow regime that yields the same inflow as that of the actual flow net.
PAPER 4293
ROCKENG09: Proceedings of the 3rd CANUS Rock Mechanics Symposium, Toronto, May 2009 (Ed: M.Diederichs and G. Grasselli)
For case (a) in Fig. 5, the equivalent steady state radius of influence can be computed in
terms of the depth to the tunnel D, using the actual analytical solution for flow rate (Harr, 1962),
as
Rwmax = 2 D
(16)
PAPER 4293
ROCKENG09: Proceedings of the 3rd CANUS Rock Mechanics Symposium, Toronto, May 2009 (Ed: M.Diederichs and G. Grasselli)
4 K ( Pw 0 Pwi )
t
2
X t = X max 1 e S w X max
0.5
(17)
Thus, according to Eq. (17) the width of the drawdown cone expands exponentially with
time, at a decreasing rate. Although Eq. (17) is applicable to a well or a vertical shaft instead of
a horizontal tunnel, it suggests that the rate of drawdown is much slower than the rate of expansion of the radius of influence. Therefore, it appears that for unconfined flow cases, the expansion of the radius of influence predominates, in the early stages of flow after tunnel excavation,
whereas drawdown dominates in the late stages.
8 HYDROMECHANICAL GROUND-SUPPORT INTERACTION
A ground-interaction analysis is used to determine the pressures acting on the lining, including
the hydraulic pore water pressure Pwi and the mechanical effective pressure Pi. The analysis is
carried out for a time t, at which the tunnel support is installed. The support system considered
includes a concrete lining and (or) rockbolts.
The concrete or shotcrete lining is assumed to be porous and elastic-perfectly plastic. The properties of the lining are thickness tc, Youngs modulus Ec, Poissons ratio c, compressive
strength fc, and coefficient of permeability Kc. The maximum support pressure Pcmax is computed based on the specified compressive strength.
The ungrouted mechanically or chemically anchored bolts are characterized by a stiffness Kb
and a maximum support pressure Pbmax. These properties are computed based on the spacing,
length, individual capacity, and other characteristics of the individual elements (Hoeck &
Brown, 1980).
8.1 Hydraulic interaction
The flow rate seeping through the ground is
PAPER 4293
10
ROCKENG09: Proceedings of the 3rd CANUS Rock Mechanics Symposium, Toronto, May 2009 (Ed: M.Diederichs and G. Grasselli)
q ground =
2K ( Pw0 Pwi )
R
w ln w
a
(18)
qlining =
2K c Pwi
a
w ln
a tc
(19)
Figure (6a) illustrates the variation with Pwi of both, the flow rate through the lining and the
seepage through the ground. Both are linear relations. From continuity, both flow rate quantities
must be equal. Thus, as shown in Figure 6a, the inflow into the tunnel and the actual water pressure acting on the lining are represented by the intersection of the above two lines. If a relatively
impervious lining is selected, the resulting water pressure acting on the lining is high, close to
original water pressure Pw0 and the tunnel inflow is minimal On the contrary, if a permeable lining is selected, the water pressure on the lining is minimal but the tunnel inflow is large.
8.2 Hydromechanical interaction
8.2.1 Ground response curve
Figure 6 illustrates the ground response curve as computed with the solution developed in the
preceding section, by computing the wall displacement for the pore water pressure Pwi determined above, and for a variable effective support pressure Pi. The effective support provided
by the face effect is shown in Fig. 6 as Pf.
8.2.2 Support characteristic curve
The support characteristic curve is obtained by subjecting the lining to the hydraulic pressure
Pwi and to a variable mechanical pressure Pi. The characteristic curve is positioned at an initial
displacement uio representing the displacement that has occurred in the ground prior to the installation of the support. Such displacement can be estimated from a longitudinal deformation
profile, as usual in the Convergence Confinement Method.
The solution for a porous liner subjected to hydraulic load Pwi and a mechanical load Pi is
obtained using the same approach followed to obtain the elastic solution in section 5.
Pi ' a Pwi a
+
+ u i0
Kw
K'
(20)
2
2
a
1 (1+ c ) (1 2 c )a + (a t c )
1
(1+c ) (1 2c )a2 + (a tc)2 (1 2c)(1 lnat ) (21)
and
=
=
+
K'
Ec a 2 (a t c ) 2
Kw 2Ec (1c ) a2 (a tc )2
lnaat
=
'
2Pi '
1 ( aatc2 )
Pwi 2
1 2 c
+
+
2
2(1 c ) 1 ( atc2 )
ln a atc
a
(22)
The maximum support pressure generated by the concrete lining is the one that makes the
tangential stress equal to the compressive strength of the material (i.e. ' = f c' ).
Although the analysis assumes both support systems are installed at time t, in combining their
characteristic curves, the actual sequence must be considered. As Fig. 7 illustrates, the sequence
determines whether or not rockbolts help to support the external water pressure acting on the
concrete lining. If shotcreting is done first followed by rock-bolting, rockbolts help to support
PAPER 4293
11
ROCKENG09: Proceedings of the 3rd CANUS Rock Mechanics Symposium, Toronto, May 2009 (Ed: M.Diederichs and G. Grasselli)
the fluid pressure. On the contrary, if the tunnel wall is bolted before shotcreting, the concrete
liner alone has to support the fluid pressure.
Figure 7. Combined support characteristic curve for two sequences of support installation. (a) shotcreting-bolting, (b) bolting-shotcreting
9 EXAMPLE
A 6 m diameter tunnel is excavated in a jointed sandstone at a depth of 300 m below the surface
where the total in situ stress is Po=8.1 MPa and the pore water pressure is Pwo= 3.2 MPa (pressure head 20 m above ground surface, Case (a) Fig. 5). The following material data is given:
Original rock mass: E = 1.5 GPa, = 0.33, c = 1 MPa, = 30, K = 0.001 mm/sec, Ss = 0.00005 m-1.
Rock mass in plastic zone: cr = 0.5 MPa, r = 25, r = 20.
Shotcrete lining: tc = 200 mm, Ec = 16 GPa, c = 0.25, fc = 50 MPa, Kc = 3x10-5 mm/sec.
Rockbolts: 25 mm rockbolt, spacing pattern: 1.00 m x 1.00 m.
The support is installed at a distance of 2 m from the advancing tunnel face, about 4 hours after tunnel face advance. Based on the above distance and a longitudinal deformation profile, the
initial ground displacement prior to the installation of the support is estimated to be 0.477 the
maximum displacement away from face (displacement at which an unstable flowing ground
condition develops). The equivalent steady state radius of influence is Rwmax= 2 x 300 = 600 m.
This example was solved for: (1) the original parameters, (2) a liner with numerous shrinkage
cracks and gaps represented by increasing its permeability an order of magnitude (Kc = 3x10-4
mm/sec), (3) no pore water pressure or conventional solution in terms of total stresses without
considering pore water pressures and seepage forces. The results of the analysis are presented in
Fig. 8 and summarized in Tables 1 and 2. Partial results for a case intermediate between 1 and 2
(Kc = 6x10-5 mm/sec) were presented in Fig. 3.
Comparing to case 3 (no pore pressures/ conventional analysis), case 1 demonstrates the effect of pore pressures and case 2 demonstrates the effect of seepage forces. In case 1, the liner is
relatively impermeable as compared to the ground, thus seepage is limited. As Fig. 8a shows,
the external water pressure in the liner is relatively high in this case. Hence the pore pressures in
the ground near the tunnel remain high.
On the other hand, in case 2, the cracked lining allows more inflow and the release of the external pressure. As Fig 8a shows, the external water pressure in the lining is relatively low.
Thus, there is a more significant hydraulic gradient in the ground in this case, leading to larger
seepage forces. Fig. 8b shows the distribution of pore water pressure in the ground. The radius
of influence is larger in the case of the cracked liner because the net hydraulic head is larger
leading to a faster rate of expansion of Rw. Table 2 shows the expansion with time of Rw. Thus,
even in permeable ground, it takes a few months to develop a steady state condition.
The ground reaction curves are shown in Fig 8c. The water pressure for each case is shown as
a dotted line. The point of equilibrium is reached for much larger displacements in case 3, due
to pore pressures and seepage forces, next in case 2, due to pore pressures and least in case 3,
which includes neither pore pressures nor seepage forces.
Effective stresses and the extent of the plastic zone are shown in Fig. 8d. The conventional
analysis gives the most optimistic results. The radius of the plastic zone increases from 2.12 to
2.64 because of the water pressures, and to 2.86 due to seepage forces.
PAPER 4293
12
ROCKENG09: Proceedings of the 3rd CANUS Rock Mechanics Symposium, Toronto, May 2009 (Ed: M.Diederichs and G. Grasselli)
Finally, the radial displacements are shown on Figure 8e. Pore pressures increase the radial
displacement at the tunnel wall from 114 mm to 164 mm, and seepage forces to 243 mm.
Figure 8. Hydromechanical interaction, and effective stresses, pore pressures, and displacement around
tunnel for (1) original parameters (2) cracked liner (3) conventional analysis with no seepage forces
Table
1. Summary of Results.
____________________________________________________________________
No
P0
Pw0
Pi
Pwi
Ua
Rp/a
Rw/a
MPa
MPa
MPa
MPa
mm
____________________________________________________________________
1
4.9
3.2
0.73
1.69
164
2.64
7.90
2
4.9
3.2
1.00
0.28
243
2.86
10.57
3
8.1
0
1.03
0.00
114
2.12
____________________________________________________________________
PAPER 4293
13
ROCKENG09: Proceedings of the 3rd CANUS Rock Mechanics Symposium, Toronto, May 2009 (Ed: M.Diederichs and G. Grasselli)
Table
2. Expansion of Radius of Influence with Time.
________________________________________________________________________________________________
Time
0
2hrs 4 hrs 12 hrs 1 day 1 week 1 month 3 months 6 months
________________________________________________________________________________________________
Case 1, Rw (m): 3
16.9 23.7 41.0
58.3
158
333
586
600
(m):
3
23.1
31.7
53.1
74.1
192
397
600
600
Case
2,
R
w
________________________________________________________________________________________________
10 CONCLUSIONS
Seepage forces generated during tunnel excavation may cause serious instability problems,
which cannot be assessed with current ground interaction analyses. An elasto-plastic solution
has been derived that includes the effect of pore water pressures and seepage forces, on the
stress and displacement field around the opening, and on the extent of the plastic zone. The
conditions for development of flowing ground at the tunnel face have been identified. The
transient nature of the hydraulic radius of influenced has been described and a procedure to predict its transient expansion has been proposed. The effects of the progressive excavation have
been discussed. Procedures for hydromechanic ground interaction analysis have been developed. An application example demonstrated the significant effect of pore water pressures and
seepage forces on the behavior of tunnels. The solution is applicable to tunneling in weak rocks,
highly fractured rocks, crushed rocks (e.g. fault zones), and general soil-like materials, subjected to high water pressures.
REFERENCES
Alvarez, T.A. 1997. A Study of the Coupled Hydromechanical Behavior of Jointed Rock Masses Around
Pressure Tunnels, Ph.D. Thesis, University of Illinois at Urbana-Champaign, Urbana, IL.
Barbosa, R.E. 1990. Discrete Element Models for Granular Materials and Rock Masses, Ph.D. Thesis,
University of Illinois at Urbana-Champaign, Urbana, IL.
Barbosa, R.E. 2009. EngSolutions TUNNEL, Users Manual. EngSolutions, Inc., Ft. Lauderdale, FL.
Brown, E.T., Bray, J.W., Ladanyi, B., Hoek, E. 1983. Ground response curves for rock tunnels. ASCE J.
Geotech. Eng. Div. 109 (1), 15-30
Bobet, A. 2003. Effect of pore water pressure on tunnel support during static and seismic loading. Tunneling and Underground Space Technology 18, 377-393.
Chu, S.T. 1994. Transient Radius of Influence Model. ASCE J. Irrig. and Drainage Eng. 120 (5), 964-969
Fernandez, G. and J.S. Moon 2006. Evaluation and assessment of inflow rates in tunnels excavated in
jointed rock mass. VI South American Rock Mechanics Congress, Cartagena, Colombia.
Harr, M.E. 1962. Groundwater and seepage. McGraw-Hill Book Co, Inc.., New York, N.Y.
Hendron A.J., and A.K. Aiyer 1972. Stresses and Strains Around a Cylindrical Tunnel in an ElastoPlastic Material with Dilatancy. Omaha District, Corps of Engineers, Technical Report No. 10,
Omaha, NE..
Heuer, R.E. 1995. Estimating rock tunnel water inflow. Rapid Excavation and Tunneling Conference,
Chap. 3, pp. 41-60.
Jacob CE, Lohman S.W. 1952. Nonsteady flow to a well of constant drawdown in an extensive aquifer.
Trans AGU 33(4):559-569
Li, X. 1999. Stress and displacement fields around a deep circular tunnel with partial sealing. Computers
and Geotechnics 24, 125-140.
Lee, I. and Nam, S. 2001. The study of seepage forces acting on the tunnel lining and tunnel face in shallow tunnels. Tunneling and Underground Space Technology 16, 31-40.
Lee, I. and Nam, S. 2004. Effect of tunnel advance rate on seepage forces acting on the underwater tunnel
faces. Tunneling and Underground Space Technology 19, 273-281.
Marulanda, A., C. Marulanda and R. Gutierrez 2008. Experiences in tunnel excavation by the conventional method and by TBMs in the Andes Mountain Range. Case histories. I South American Symposium on Rock Excavations, Bogota, Colombia.
Perrochet, P. 2005. A simple solution to tunnel or well discharge under constant drawdown. Hydrology
Journal 13, 886-888.
Perrochet, P. 2005. Confined flow into a tunnel during progressive drilling: an analytical solution.
Ground Water 43, No. 6, 943-946.
Shin, J.H., Potts D.M., and Zdravkovic L. 2005. The effect of pore-water pressure on NATM tunnel linings in decomposed granite soil. Canadian Geotechnical Journal 42, 1585-1599.
Singh, B. and R.K. Goel (2006) Tunneling in Weak Rocks, Volume 5 Geo-Engineering Book Series,
Elsevier.
PAPER 4293
14