INTERNATIONAL SOCIETY FOR

SOIL MECHANICS AND

GEOTECHNICAL ENGINEERING

This paper was downloaded from the Online Library of

the International Society for Soil Mechanics and
Geotechnical Engineering (ISSMGE). The library is
available here:

https://www.issmge.org/publications/online-library

This is an open-access database that archives thousands

of papers published under the Auspices of the ISSMGE and
maintained by the Innovation and Development
Committee of ISSMGE.
 Geotechnical Aspects of Underground Construction in Soft Ground – Ng, Huang & Liu (eds)
© 2009 Taylor & Francis Group, London, ISBN 978-0-415-48475-6

Characteristics of tunneling-induced ground settlement in groundwater

drawdown environment

C. Yoo & S.B. Kim

Sungkyunkwan University, Suwon, Korea

Y.J. Lee
RIST, Yongin, Korea

ABSTRACT: This paper presents the results of an investigation on the characteristics of tunnelling-induced
ground settlement in groundwater drawdown environment. The dynamics of the effect of groundwater draw-
down on the ground settlements are first investigated using a case history concerning a conventional tunnelling
situation in which the interaction between the tunnelling and the groundwater induced excessive ground surface
settlements. A 2D stress-pore pressure coupled finite element analysis is then conducted on a tunnelling case
with groundwater drawdown, aiming at investigating ground surface settlement characteristics. The results indi-
cated among other things that significant portion of ground settlement can occur before tunnel face reaches, and
that the error function approach does not provide a good fit to the settlement troughs for tunnelling cases with
groundwater drawdown.

1 INTRODUCTION
GWT
Tunnelling beneath the groundwater table causes
changes in the state of stress and the pore water pres- Ground
sure distribution. In such tunnelling problems, the surface
tunnelling work inevitably causes water inflows into Groundwater
excavated area, thus causing the change in the pore lowering ∆U
water pressure distribution. The direct environmen-
tal consequence of water inflows during tunnelling is
Reduction in
the drawdown of groundwater level in the surround- pore pressure
ing aquifer (Yoo 2005). The related ground subsidence
occurring as a result of the reduction in water pressures Figure 1. Illustration of ground settlement associated with
in the soil layers can damage nearby structures/utilities tunnelling-induced groundwater drawdown.
(Figure 1). One of the major case histories illustrat-
ing damage due to ground settlement associated with
tunneling-induced groundwater drawn is perhaps the Xu et al. 2006). The results of their studies can-
Romeriksporten tunnel in which the highspeed rail- not be directly applied to the tunnel excavation
way tunnel construction caused more 1 m of ground problems as they focused only on the groundwater
subsidence due to groundwater drawdown, raising sig- drawdown due to groundwater pumping. As urban
nificant technical and political issues pertaining to tunnelling projects tend to involve potential prob-
the effect of tunnelling on surrounding environment lems related to groundwater drawdown ground move-
(NSREA 1995). ments during tunnelling, there is an urgent need for
Surprisingly studies concerned with the tunneling- better understanding on the mechanisms involved
induced ground movements in groundwater drawdown in tunnelling-induced ground movements associated
are scarce as indicated by Yoo (2005). Although a with groundwater drawdown.
number of studies on the ground subsidence caused This paper presents the results of investigation on
by groundwater pumping from an aquifer have been the characteristics of tunnelling-induced ground set-
conducted (Shen et al. 2006; Qiao & Liu 2006; tlement in groundwater drawdown environment. The

485
Figure 2. Typical ground profile.

Table 1. Geotechnical properties of soil/rock layers.

γ c′ φ′ E ν k Figure 3. Support pattern (typical).

Type (kN/m3 ) (kPa) (deg) (MPa) (cm/sec)

fill 18 0 27 5 0.40 3.8 × 10−4

2.2 Tunnel design
alluvial 20 15 30 10 0.40 3.8 × 10−4 Figure 3 shows a typical tunnel support pattern used
weathered 25 15 30 50 0.33 2.4 × 10−4 for a 100 m long section for the ground profile given
soil
in Figure 2. On account of the difficult ground con-
weathered 25 60 35 120 0.30 8.8 × 10−5
rock dition the ring cut excavation method was adopted to
hard rock 26 100 35 200 0.25 5.0 × 10−5 promote the tunnel face stability during excavation.
The primary support system consisted of a 0.2 m thick
Note: γ = unit weight, c′ = cohesion, φ = internal fric- steel fibre reinforced shotcrete (SFRS) layer with 4 m
tion angle, E = young’s modulus, ν = poisson’s ratio, long system rock bolts at 1.0 and 1.2 m, respectively,
K = coefficient of permeability longitudinal and transverse spacing. The pipe umbrella
technique using 800 mm diameter grout injected 12 m
long steel pipes was additionally implemented to pro-
dynamics of the effect of groundwater drawdown on mote the face stability through improving the load
the ground settlements are first investigated using a carrying capacity of the ground ahead of the face. Also
case history concerning a conventional tunnelling sit- adopted was a trumphet shaped micro cement injec-
uation in which the interaction between the tunnelling tion (MSG) pre-grouting around the tunnel periphery
and the groundwater induced excessive ground sur- to create a 5 m thick watertight shell for sections in
face settlements. A parametric study using a calibrated which the weathered soil layer extended to the tun-
2D stress-pore pressure coupled finite element model nel crown level. The pre-grouting scheme was later
is then conducted on a number of factors influenc- extended to cover the face after the settlement problem
ing the tunnelling-induced ground settlements with had become an issue.
groundwater drawdown. Based on the results, the inter-
action mechanism between the tunneling, groundwater
lowering, and ground settlement is identified. 2.3 Measured ground surface settlements
Figure 4 shows the progressive development of set-
tlements during the tunnel advancement at various
2 GROUND SURFACE SETTLEMENT
monitoring stations. In this figure the measured settle-
CHARACTERISTICS – FIELD MONITORING
ments are plotted against the relative distance between
DATA
the tunnel face and the monitoring stations normal-
ized by the tunnel diameter (D). These data, measured
2.1 Tunnelling condition
using the conventional leveling technique, thus indeed
A case history concerning the conventional tunnelling, represent the settlement history during the tunneling
i.e., NATM, was considered. The tunnel has excavation process for the monitoring stations.
width and height of approximately 10.5 m and 8.7 m, A total of five settlement curves are presented in
respectively, with a cover depth ranging approximately this figure. As can be seen in this figure the five curves
20 ∼ 30 m, and constructed in a multi-layered ground are similar both in qualitative and quantitative terms,
including a fill, alluvium, and a weathered zone as despite the vertical extent of the decomposed soil rel-
illustrated in Figure 2. The geotechnical properties of ative to the tunnel varies along the route, showing
the ground are given in Table 1. the maximum converged settlements in the range of

486
 of the final settlement (Sv,max ) was completed before
the face passed a monitoring station with the remain-
ing 30 ∼ 40% occurred after the full passage of the
tunnel face. Such a percentage of settlement ahead of
the face is considerably larger than the typical value of
40 ∼ 50%, suggesting a larger portion of the converged
settlement occurred prior to the arrival of tunnel face
in this tunneling condition than a tunneling condition
without the groundwater drawdown. Moreover, the set-
tlements tend to converge to a constant value after the
tunnel face advanced to a distance of 6 ∼ 7D beyond
the monitoring stations, suggesting slower settlement
convergence than a normal condition.
These results in fact are somewhat different from
Figure 4. Progressive development surface settlements at typical trends that can be observed in tunneling con-
various monitoring stations. ditions without significant groundwater lowering, and
led to a conclusion that factors other than the unloading
effect due to the tunnel excavation may have played a
role. Such a tendency is directly linked to the ground-
water drawdown as will be shown in a subsequent
chapter.

3 PARAMETRIC STUDY

3.1 Stress-pore pressure coupled analysis

A commercial finite element package ABAQUS
(Abaqus, Inc. 2002) was used for the parametric study.
A 2D stress-pore pressure coupled effective formula-
tion was adopted in order to realistically capture the
interaction mechanism between the tunnelling and the
Figure 5. Normalized settlement history curves. groundwater.
In ABAQUS a porous medium is approximately
1.6% ∼ 1.8%D. Of salient feature that can be observed modelled by attaching the finite element mesh to
in this figure is the tendency of settlement increase the solid phase. Equilibrium is expressed by writing
during which the tunnel advancement was halted, sug- the principle of virtual work for the volume under
gesting time dependent (tunneling activity indepen- consideration in its current configuration at time t:
dent) settlement development (to be discussed later).
Another of interest trend is the resemblance of the
settlement curves with a typical log t curve from a con-
solidation test, being characterized by three zones as
an initial compression, primary, and a secondary zone.
Such a trend strongly suggests a possible cause being where dv is a virtual velocity field, dε is the virtual
the volume change effect due to tunneling-induced rate of deformation, σ is the true (Cauchy) stress, t are
groundwater drawdown. surface tractions per unit area, and f body forces per
The data in Figure 4 are further analyzed by nor- unit volume. f includes the weight of the wetting liquid
malizing the settlement values with their respective fw defined as Eq. (2)
maximum values (Sv,max ) in Figure 5. As seen in this
figure, the normalized curves tend to collapse into one
curve. A further inspection of the normalized settle-
ment history curves shows that the settlements started in which s is the degree of saturation, n is the porosity,
to develop when the tunnel face was approximately 6D and nt is the volume of trapped wetting liquid per unit
away from the monitoring stations. The settlements of current volume. Eq. (1) can then be rewritten as
tend to accelerate when the tunnel face reached 3D
away from the monitoring stations, and decelerate after
the tunnel advanced 5 ∼ 6D beyond the monitoring
stations. Also shown are that approximately 60 ∼ 70%

487
Table 2. Geotechnical properties.

γ E c φ k
(kN/m3 ) (kPa) (kPa) (◦ ) (cm/sec)

soil 25 50,000 30 30 5.8 × 10−3

Weathered 25 100,000 50 30 1.3 × 10−4
rock
Figure 6. Finite-element model used in the analysis.

The ground and the shotcrete lining were discretized

where f are all body forces except the weight of the using 8-node displacement and pore pressure elements
wetting liquid. with reduced integration (CPE8RP). The rock bolts
The continuity equation is satisfied approximately were modeled using the 2-node truss elements. With
in the finite element model by using excess wetting regard to the material modelling, the soil and rock
liquid pressure as the nodal variable (degree of free- layers were assumed to be an elasto-plastic material
dom 8), interpolated over the elements. The backward conforming to the Mohr-Coulomb failure criterion
Euler approximation is used to integrate the equa- together with the nonassociated flow rule proposed by
tion over time and the Newton iterations are used to Davis (1968), while the shotcrete lining and the rock
solve the nonlinear, coupled, equilibrium and continu- bolts were assumed to behave in a linear elastic man-
ity equations. Fundamentals of the stress-pore pressure ner. The geotechnical properties for the ground given
coupled formulation adopted in ABAQUS can be in Table 1 were used for analysis. The young’s moduli
found in the ABAQUS user’s manual (Abaqus, Inc. of the shotcrete and rock bolts were chosen as 15 GPa
2005). and 21 GPa, respectively.
The actual tunnelling process consisting of a series
3.2 Condition analyzed of excavation and support installation stages was
closely simulated in the analysis by adding and remov-
A tunnelling condition frequently encountered in ing corresponding elements at designated steps. After
urban situations was considered in the analysis. The establishing the initial stress and pore pressure condi-
tunnel considered is a 10 m diameter horseshoe shaped tions with appropriate boundary conditions, the step-
tunnel with a cover depth of 3.0D, excavated by the by-step tunnelling process pertinent to the bench cut
bench cut method. The primary support system con- excavation method, was then simulated. The 3D effects
sists of a 20 cm thick shotcrete lining with system of advancing a tunnel heading was taken into consider-
rock bolts installed at 1.5 m center-to-center spacing. ation using the stress relaxation method in which the
A 1.5D thick soil layer was assumed to exist above a boundary stresses arising from the removal of exca-
weathered rock layer through which the tunnel is exca- vated elements are progressively applied to simulate
vated. Tables 2 summarizes geotechnical properties of the progressive release of the excavation forces as the
the ground. tunnel heading advances.

3.3 Finite element model

3.4 Ground settlement characteristics
Figure 6 shows the finite element model adopted in
this study. The finite-element mesh extends to a depth Figure 7 presents the relationship between the max-
of two times the tunnel diameter (D) below the tun- imum surface settlement (Sv,max ), directly above the
nel spring line and laterally to a distance of 15D from tunnel crown, obtained during various stages of tun-
the tunnel center depending on the cover depth H . The neling. As seen the maximum surface settlement
lateral location was selected based on a series of pre- Sv,max tends to linearly increase with the increase in
liminary analysis as it has a significant influence on the groundwater drawdown level HD . The settlement
the results of a stress-pore pressure coupled analysis. occurred after the completion of tunnel is in fact
At the lateral boundary displacements perpendicular to twice that during the tunnel excavation. It should be
the boundaries were restrained whereas pin supports noted that the plot given in this figure represent those
were applied to the bottom boundary. caused by the groundwater inflow into the tunnel after
With regard to the hydraulic boundary conditions the completion of tunnel excavation, until a steady
and with reference to Figure 6, a no-flow condition state condition is achieved. This suggests a direct link
was assigned to the vertical boundaries perpendicular between the ground settlement and the groundwater
to the tunnel drive. At the lateral vertical boundary the drawdown, thus demonstrating the importance of cre-
groundwater table was assumed to be at the ground ating a watertight shell for tunnelling cases where the
surface and constant throughout the analysis. controlling ground surface settlement is of concern.

488
Figure 7. Variation of Sv,max with HD after completion of Figure 9. Normalized surface horizontal displacement pro-
tunnel excavation. files.

tend to collapse into one curve despite some discrep-

ancies in the region father away, i.e., ≥ 4D, from the
tunnel center.
Normalized horizontal displacement (Sh ) profiles
are shown in Figure 9 for different levels of ground-
water drawdown. As seen, the maximum horizontal
displacements (Sh,max ) tend to develop at locations
3D away from the tunnel center with decreases in
magnitudes thereafter. Again the Sh profiles tend to
collapse into one curve although some discrepancies
are observed in the region away from the tunnel center.
The results obtained in this study suggest that the set-
tlement trough as well as the horizontal displacement
profile may be constructed using normalized curves
Figure 8. Normalized surface settlement troughs. when relationships between the surface movements
and other factors can be established.

In urban tunnelling situations, characteristics of

a ground surface settlement trough, such as slope
and width of inflection point, are important as they, 4 CONCLUSIONS
together with lateral displacements, determine poten-
tial for damage of adjacent structures. Figure 8 shows This paper presents the results of an investigation on
normalized ground surface settlement troughs for dif- the characteristics of tunnelling-induced ground set-
ferent groundwater drawdown levels together with tlement in groundwater drawdown environment using
error functions constructed using different values of the measured surface settlement for a site where
inflection point. Of interest trends are two fold. First, the tunnelling-induced groundwater drawdown caused
the extent of ground settlement trough is significantly significant surface settlement. A stress-pore pressure
greater than typical tunnelling conditions without coupled finite element model was additionally con-
groundwater drawdown. In fact, for the particular tun- ducted aiming at identifying the ground movement
nelling condition considered, the ground settlement characteristics when tunnelling induces a significant
zone extends more than 10 times the tunnel diameter level of groundwater drawdown. Based on the results
from the tunnel centerline. Second, the error function the following conclusions can be drawn.
approach (Attewell et al. 1986; Peck 1969) known to 1. For tunnelling cases in which tunnel excavation
well describe the surface settlement trough for tun- causes significant groundwater drawdown, the per-
nelling cases without ground water drawdown does centage of settlement that develop prior to the
not provide a good fit to the settlement troughs for tunnel face arrival to the final settlement is signif-
cases with groundwater drawdown. Another important icantly larger than for cases without groundwater
observation is that the computed settlement troughs drawdown.

489
2. Continued groundwater drawdown after the com- tunneling and their effects on pipeline and structures,
pletion of tunnel excavation may can cause settle- Blackie, Glasgow: 53–66.
ment larger than that occur during excavation. Norwegian Soil and Rock EngineeringAssociation (NSREA).
3. The error function does not provide a good fit to 1995. Norwegian urban tunnelling. Publication No. 10.,
Norway.
the settlement troughs for cases with groundwater Peck, R.B. 1969. Deep excavations and tunneling in soft
drawdown. ground Proc., 7th Int. Conf. on Soil Mech. And Found.
4. Normalization can hold for the surface settlement Engrg. : 225–290.
and horizontal displacement profiles for tunnelling Qiao, S. & Liu, B. 2006. Prediction of ground displacement
cases with groundwater drawdown. and deformation induced by dewatering of groundwa-
ter. Underground Construction and Ground Movement:
73–79.
ACKNOWLEDGEMENT Shen, S.L., Tang C.P., Bai, Y. & Xu, Y.S. 2006. Analysis of
settlement due to withdrawal of groundwater around an
unexcavated foundation pit. Underground Construction
This research is supported by Korea Ministry of Con- and Ground Movement: 377–384.
struction and Transportation under Grant No. C4-01. Xu, Y.S., Shen, S.L. & Bai, Y. 2006. State-of-art of land
The financial support is gratefully acknowledged. subsidence prediction due to Groundwater Withdrawal
in China. Underground Construction and Ground Move-
ment: 58–65.
REFERENCES Yoo, C. 2005. Interaction between Tunnelling and
Groundwater-Numerical Investigation Using Three
Abaqus users manual, Version 6.5. 2005. Hibbitt, Karlsson, Dimensional Stress-Pore Pressure Coupled Analysis.
and Sorensen, Inc., Pawtucket, Providence, R.I. Journal of Geotechnical and Geoenvironmental
Attewell, P.B., Yeates, J. & Selby, A.R. 1986. Ground defor- Engineering, ASCE 131(2): 240–250.
mation and strain equations. Soil movements induced by

490