Ground Movements Due to the Construction of

Cut-and-Cover Structures and Slurry Shield Tunnel

of the Cairo Metro

M. Hamza, A. Ata, and A. Roussin

A b s t r a c t - - T h i s paper pn,.sents an evaluation of the settlement prediction techniques used to estimate the
surface settlements associated with the construction of the Greater Cairo Metro Line 2. The construction of
the Cairo Metro involved I~heconstruction of cut-and-cover underground stations and bored tunneling. A
typical underground station was executed using top-down construction technique. The twenty two meters
excavation was carried inside a watertight box with 50-m-deep diaphragm walls to form the sides and a 7-
m thick groutedplug at the bottom. Tunneling was performed using a slurry shield tunnel boring machine,
TBM, having an internal diameter of 9.48 m. This analysis is the first step in uiew of enhancing the
procedures of settlement prediction and appraising potential damages to overlying structures and utilities
for the future construction of the twin road tunnels in the historical urban environment of AI Azhar area and
If7~an El I~alily market in Cairo. © 1999 Published by Elsevier Science Ltd. All rights reserved.

KETWORDS: bored tunnel; settlement prediction; field monitoring; slurry sh~ld; TBM; cut and cover

I. Introduction 2. Subsurface Conditions

nalyzing the restdts from monitoring of ground settle The subsurface conditions along the route of the Cairo

A ments restdtingt~om construction works of the Cairo

Metro serves two main goals. The first is to record
the actual values of s~wface and subsurface settlements
Metro Line 2 can be generally described as a typical river
Nile alluvial deposit, where four main layers are recognized
as follows:
during construction in order to monitor the safety and • Fill or made ground consisting of asphalt, broken red
integrity of nearby structures and underground utilities. bricks and stones in a sandy silt/clay matrix
The second is to use the records of ground movements to • Clay which may be classified as silty clay or clayey silt
improve or redevelop the methods utilized in the settlement * Fine Sand, micaceons, very silty with lenses of clayey
prediction process and :in the design of tunnels in similar
sandy silt
soi~ ground or "Nile" alluvial formations.
The first goal is concerned with the absolute values of the • Sand, slightly gravely to gravely, medium to coarse.
recorded settlements at the selected locations or reference A typical geotechnical profile is presented in Figure 1,
points. This helps in controlling the construction opera- showing the average depths of the different soil layers and
tions, if needed, and in providing immediate alarms to the ground water depth.
designer and to the contractor of any unexpected events so
that immediate actions m a y be implemented. 3. Settlement Due to Construction of Cut-and-
On the other hand, the fulfillment of the second goal Cover Structures
requires a longer time, i.e., until a sufficient body of data is
collected in order to analyze and observe a pattern of Other than settlement caused by accidental conditions,
behavior and start the back analysis, which is the scope of the ground surface settlement associated with the construc-
this paper. tion of underground stations can be generally divided into
two main components:
• Settlement due to the installation of slurry and/or
diaphragm walls
Present address: Mamdouh Hamza, Principal Associate, Hamza • Settlement due to wall movements (e.g., movement
Associates, Giza, Egypt; Alan Ata, Lecturer, Department of due to excavation inside the wall).
structural Eng., Zagazig University and Senior Geotech. Eng. These two components are functions of m a n y factors
Hamza Assoc.,Egypt; and Alain Rouesin, Design Manager, J.V. led such as depth and shape of walls, distance from the embed-
by Campenon Bernard SGE, France.

www.elsevier.com/locate/tust
Tunnelling and Underground Space Technology,Vol. 14, No. 3, pp. 281-289, 1999
0886-7798/99/$ - see front matter © 1999 Published by Elsevier Science Ltd.
All rights reserved.
~ Pergamon
PII: S0886-7798(99)00044-9
 ded wall, subsoil properties, depth of excavation inside the
walls, type and stiffness of supporting system, time period
ofcoustruction, surrounding structures, and surcharge loads.
Layer 1: Fill
p 3.1. Installation of Diaphragm Walls (So-Curve)
i m m G.W.Tablc The vertical settlement at the ground surface due to the
., ~",~'',,, ,. ,,
installation of diaphragm walls (or slurry walls) is gener-
ally represented as a function of the trench depth and the
4.00m Layer 2: Clay horizontal distance from the trench (Peck 1969). It is
generally accepted that in normal conditions, based on
data from case histories, the surface movements are lim-
ited to a small percentage of the excavation depth and
decrease with increasing distance from the wall. In view of
the new construction technologies, m a n y researchers have
contributed to the database of the subject by updating it
Layer 3: Silty sand with field measurements and by specifying patterns of
3.111 m movements in the vicinity of deep excavations [e.g., Clough
and Schmidt (1981), Uriel and Sagesta (1989), and Clough
and O'Rourke (1990)].
Hamza (1993) adopted the following expression to de-
scribe the settlement component due to the installation of
diaphragm walls (S o curve),
So )=s,[1- lm (1)
Layer 4: Sand
where:
S t is the maximum settlement adjacent to the wall,
x is the distance from the wall,
x ois the distance from the wall where the settlement due
to wall installation is equal to zero, and
Figure 1. Soil profile at the Cairo Metro Line 2 (Lot 12 at m is an empirical exponent.
kP 3.358).
Because local experiences are essential for the estimate
of this settlement component, it was decided to use initial
settlement measurements at one of the
Table 1. Measured settlements in the vicinity of Rod El Farag Station. first stations t h a t was constructed for
the Cairo Metro Line 2, namely Rod E1
Distance from D. Wall (m) 5 10 20 40 60 Farag Station. The measured in-situ
settlements are presented in Table 1.
Settlement (mm) 15 11 6 1.5 0.9 In this particular case, for the set of
measurements at the West Side of the
station, the m a x i m u m value of settle-
ment is extrapolated as S t = 20 ram.
This value is approximately 0.041% of
the excavated trench depth, which is
Distance from wall (m) typically in the middle range reported
0 l0 40 6O S0 100 120 by Clough and O'Rourke (1990). Figure
2 presents a plot of the measured values
9 ,,,,,

and the values calculated by Eqn. (1) for

various m p a r a m e t e r s and for x o= twice
the trench depth = 97.2 m. Figure 2
shows t h a t the best fit for the proposed
0.COS equation is for a value of the exponent
m = 5. The values of settlements pre-
dicted by Eqn. (1) and using the above
mentioned parameters at all stations
along the Cairo Metro Line 2 are in good
o.01
| agreement with the values and the pat-
tern of the measured settlements.
Figures 3 and 4 present typical ex-
amples of the accuracy of the predictions
o.o]s compared to the field measurements at
two stations: St. Theresa (Phase 1A) and
Mohamed Naguib (Phase 1B). The best
fit curves shown in Figures 3 and 4
0.02 confirm the adequacy ofEqn. (1) to model
the pattern of settlement when using
the above mentioned parameters (S t =
0.041% D and m = 5). The measured
settlements at these two locations are
o.oz5 slightly less t h a n the values predicted,
especially at distances close to the walls.
Figure 2. Parameters of the SO curve (data from Rod El Farag Station).

~82 TUNNELLING AND UNDERGROUND SPACE TECHNOLOGY Volume 14, Number 3, 1999
It is also noted that existing buildings
adjacent to these two locations are ap- ~tal;ft' (m t
proximately 8.0 m away from the in- 0 la 20 30 -;0 50 6O "0
stalled walls. The stiffzless of the build- IIIII

ings affected both the value and the -- •O " ~I- ' ~ ' ' ~ ' ' A I ~

J
:
pattern of movements by reducing the
settlement adjacent ~ the wall and
keeping it constant for some distance
below the building. 0.01
A summary of the results from the
Cairo Metro Line 2 (Phase 1A, 1B and
2A) is presented in Figure 5, where the
data are normalized by the trench depth.
Statistics of the ratio of the maximum "~ 1t.02
settlement to the trench depth are as
follows: .... P r e d i c t e d
Minimum S~depth = 0.014 % • Measured
Maximum S]depth = 0.052 % o~o3 - ..........
Average S/depth = 0.031%
The envelope of the data (S0-curve) Figure 3. Settlement trough at St. Theresa Station.
is also plotted on Figure 5 using the
average value S~= 0.041% D. From Fig- Distance (in)
ure 5, it is clear that the adopted method 0 I0 20 30 40
provides an adequate approximation
for most measurements along the route
of the Cairo Metro Line 2.

3.2. Wall Movement During

o,oi
Construction (S and 81 Curves) w

The procedure used for estimating

the magnitude and distribution of the
settlements that are likely to occur due
to excavation inside the diaphragm 0.02
Walls is consistent with the applied
methods of computation for the design
Predi¢led
of the diaphragm walls in the Cairo
Metro Line 2 project. The procedure is • MeRsured
(~°~ . . . . . . . . . . . . . .,,.,,, ,.,,

basically an observational method that

utilizes the available database in the Figure 4. Settlement trough at Mohamed Naguib Station.
literature and initial experiences from
the Cairo Metro Line 2 project. The
general pattern of wall :movement and
of adjacent ground settlement can be
summarized into two main curves: the Distance from D. wall/trench depth
S curve and the S 1 curve (Hamza 1993). o o.2s o.Ts , 1.5
These two curves and the parameters
involved are presented in Figures 6a • ....
0 . . . . . . . . . . . . . . . . .

and 6b.
The two curves pertain only to settle ~+ o , . 0 ~ , ~'- ....:- .......
ments associated with the excavation, A ~ ~ ~ i • ROD
design and execution stages under nor-
real construction conditions. The pre- " _°~i ~ i • oPntA
0,02
dictions do not account ibr other activi-
ties, such as dewatering, poor work-
manship or external eftbcts that may
cause instability.
The results of the computations for ~..~..................... ~ ............ -.. X LOT-I0
the design of the shafts contain the o.0
deformed line of the diaphragm wall. • NAGUIR
The volume of the soil confined be-
tween the initial and final position of X -t-MAZE
the wall, Vw, and the m ~ m u m value
of the horizontal movem~.nt ofthe wall, ¢jl 0 . 0 6 O MAZ-W
y ~ , are used as the main input data
for the settlement estimate of the A ROBE
ground surface and under the neigh-
boring structures. The volume is calcu- - - - - - Emc|ope
lated by numerical integration of the 0.08
horizontal displacement along the
depth of the wall. Figure 5. Summary of normalized settlement due to installation of the
Cairo Metro D. Walls.

Volume 14, Number 3, 1999 ~ L L n ~ G ANDUNDERGROUNDSPACETECHNOLOGY283

 3.3. First Analytical Approximation (S-Curve)
Previous attempts to match the measured settlement
curves have indicated that the shape of the settlement
trough can be approximated by the error function that is
widely used in the prediction of the settlement trough due
to tunneling. Parameters of this curve, which for easier
reference will be called S curve, are shown in Figure 6a.
The S-Curve shown in Figure 6a provides the settlement
trough adjacent to the wall that m a y represent particular
cases or stages of construction where the movements at the
top of the wall are not totallyrestrained (flexibleor canti-
lever system). A similar triangular pattern has been de-
scribed by Clough and O'Rourke (1990) where the settle-
ment trough is approximated by triangular bounds. How-
t ever, the authors consider that the first analytical ap-
/ proach (S-Curve) presented herein corresponds to previous
0.f~Sm= . ""
Vw / observed modes of wall displacements and low shearing
Vs i ~ SIn=
resistance at the soil-wall interface from instrumented
sections at the Cairo Metro Line 1 (Hamza 1993).
If Sin= is the m a x i m u m settlement at the wall contact,
¥=.. the equation of the settlement trough can be as:
S (x) = S , ~ exp [ -x~ / 2i 2 ] (4)
where the maximum settlement is given by
S ==Y,/(¢Y~i) (5)
\ Substituting from Eqns. (2) and (3) in Eqn. (4) and
\
solving for i, we get the location of the inflection point which
:\i corresponds to the settlement of 0.606 Sm~x:
i =RsR v Vw/Ym~ ~ (6)
Fis. 6=. ~ m m m ~ o f ~ m S- c u r ~
Having obtained the value ofi from Eqn. (6) and S ~ from
Eqn. (2), the settlement profile can be constructed using
Eqn. (4). The effective width of the settlement trough for
this settlement component can be taken as:
v. ~ - - . . . . . //]~ . . . . . i j --- X
x~=~/z/2i =2.507i . (7)
/ I v, o~s...z/ s..
The maximum gradient of the surface settlement profile
for this curve is:

ddlnaiam~
/ Vw L. il ~ il ---,¢ i (8)
II I' Yao cota ~ - 0.606 S m~
and the average slope is:
_ x max (9)
cota ~g S max

Hamza (1993) studied the two R v and R,, in view of data

'---- wJl available from the Cairo Metro Line 1 at Saad Zaghloul
underground station and using dimensional analysis. The
computed volume ratio of R at Saad Zaghloul Station was
0.65. A volume ratio of 0.75 has been finally adopted to be on
Fi~ 6b. l ~ m ~ oftbo Si-
the conservative side. The computed
settlement
. . .
ratio R $, was 1.40, which
is identical to the value proposed by
Figure 6. Parameters of the S and $1 curves. Clough and Schmidt (1981).

3.4. Second Analytical

Given the maximum value ofthe horizontal movement of Approximation (S1-Curve or Truncated Error Function
wall, y=,=, the m a x i m u m value of the ground surface, S=~=, Curve)
can be expressed in general by
The mode of movement of some other walls, particularly
S,= =y,~/R e (2)
deep ones which are supported at the top in the early stages
where R is a settlement ratio. by roof slabs, will be significantly restrained at the top, and
Given the volume, V , the volume of the settlement profile larger movements at the base are likely to occur at some
of the ground surface, Vo, can be expressed in general as: depth with negligible movements at the base level. In
V = R V. (3) addition, we assume that the adhesion at the soil-wall
interface for the cast-in-place concrete would be higher,
where R is a volume ratio. reducing the possible slip at the soil-concrete contact and,
In the text that follows, analytical expressions relating therefore, changing the shape of the settlement trough at
ground surface settlements resulting from wag movements the ground surface. It is therefore worth considering an
are given. alternative definition for the shape of the settlement trough
from t h a t observed in the instrumented section of the Cairo

2 8 4 TUNNELLING AND UNDERGROUND SPACE TECHNOLOGY Volume 14, Number 3, 1999

Metro Line 1. In such circumstances,
the location of the point with maximum
settlement is not necessarily at the point Distance from D. Wall (m)
just next to the wall, but at some dis- 0 10 20 30 40 50 60
tance away. Observed settlement pat-
terns of this type have been extensively 0
reported in the literature, for example, O~ ~'~IP~'"
t-g~' " J
5
by Dyslis and F o n t a n a (1982) and
Clough and O'Rourke (1990). 10
Accordingly, a new settlement curve 15
(S1-Curve) can be constructed based on
the probability equation as shown in
~" 2o ]o ........... ......... i ............. i ............. 4 ..........

Figure 5 and having the same maximum 25

settlement and the ratio of volumes as
for the case of the S curve,
35 ~
Slm~ = y,~= / R s , and (10) I
40 ~
V = R V , where y ix and t h e r=L~ fred ~20 I ............
I S (Measured &flectioa
: ...............................................................
volume ratios are as previou~sT~defined. 45 I
i 25 ~ _ ~ SI (Measured deflectioe
By integration from x = 0 to infinity, 50
. . . . . . S (Design)
we get 0 20 40
30 - "+-- St (Design)
V = 2.109 i 1 $1~~ (11) Deflection (mm)
Given the value ofym.~ and knowing (a) Deflection of D. Wall
V , the required parameter i I is simply: (b) G r o u n d Surface Settlement

R oR w V w (12) Figure 7. Predicted a n d measured settlements at R o d E l Farag station.

i 1 = 2.109y m,~
The effective width of the settlement
trough for this case is:
x lm~ =(1 + ~ : ~ i 1 = 3"507i I Distance from D. Wall (m)
(13) 0 10 20 30 40 50 60
The maximum gradient of the sur- o !
face settlement profile tbr this curve is: 0
5
i1 (14) 10
cota lma~= 0.606S lm,x
15
and occurs at two point~l, i.e., for x~ = 0
(on the soil surface at the. wall interface)
A

20 ~ 10
and at the distance x = 2i measured from 25
the soil surface to wall interface. 15 ....... ; : 4- • • ................
30
3.5. Comparison between Predicted and 35
20 - . ~ 4 ~. . . . . . .
Measured Settlements 40
This section presents actual field 45 s (Measured deflection)
measurements of the ground surface i 25
5O SI (Measured deflection)
settlements and the D. wall deflection at
selected Stations at the Cairo Metro line 0 10 20 30 . . . . . . S (Design)
30 - + - - S! (Design)
2. The actual D. wall deflections were Deflection (mm)
measured by installing inclinometers
inside the D. wall. Elewation reference (a) Deflection of D. Wall (b) G r o u n d Surface Settlement
points were installed at the ground sur-
face on the same line with the inclinom- Figure 8. Predicted a n d measured settlements at E1Behoos Station.
eter and perpendicular to the D. wall
(wherever possible). The results of these
measurements are presented in Figures
7 and 8 and are indicated as "measured,"
The design values of the wall deflection are also shown The measured settlements at the selected stations show
on these figures for comparison. Each figure presents the that they follow, to a reasonable extent, the general pat-
actual measured settlements (solid points) in mm versus tern predicted by either S or S1 curves. However, the
the distance from the D. wall. Two sets of the theoretical S measured settlements at Rod E1 Farag do not exactly
and S1 curves are also shown on these figures where the follow the expected (predicted pattern). It is logical to
designated settlement and volume ratios are R = 1.4 and
• • B
understand that the field measurements can involve sig-
R v = 0.75, respectively. The curves in sohd hnes are the nificant settlements caused by other activities at the site
predictions using the measured deflection of D. walls as such as dewatering, frequent and heavy traffic loads, other
input, while the curves in dashed lines represent the excavation or construction in the vicinity of the reference
predictions using the theoretical design values for the points. However, the main focus in viewing these measure-
deflection of D. walls. For the selected stations the maxi- ments is to justify the general pattern of movements.
mum deflection of the D. wall was similar, if not close, to Furthermore, the calculated wall movements are usually
the design value. higher than the expected ones because the designer some-

Volume 14, Number 3, 1999 TUNNELLINGANDUNDERGROUNDSPACETECHNOLOGY285

 function) curve to represent the shape of
the settlement trough due to tunneling in
C.L
soft ground (e.g., Peck 1969, Schmidt 1969,
and Attewell and Selby 1989) that may be
x_ given by
S (x) = S ( x ) ~ exp (x2 ! 2i 2 ), (15)
-t X
/ [.=y=r (1) c~y where
C Kc = 1.0 nc = 1.0 S is the vertical settlement at a distance x
from the centerline,
S is the maximum surface settlement (at
~ (2) Sand t l ~ t u n n e l centerline), and
I~ = 0.63 ~ = O.~C/ i is the distance to the point of inflection of
the error function curve (Fig. 9).
The definition of the error function curve
is shown in Figure 9 after Peck (1969) and
Schmidt (1969). The use of Equation (15) to
predict the settlement S(x) requires the
knowledge of two parameters; the maxi-
m u m settlement S and the positionofthe
• • • m~x

point of reflection, 1.
Cording and Hansmire (1975) conducted
a statistical analysis for the case of tunnel-
ing in medium to dense sand and concluded
that it is satisfactory to assume that the
Figure 9. Definition of the settlement curve and variables used in the volume of the settlement trough is equal to
computations. the volume of the soil lost into the tunnel.
This has been also satisfactory in case of
cohesive soils (Hamza 1995). Accordingly
times considers a factor of safety or even makes assump- the volume loss is calculated by
tions regarding the material parameters.
V,=V=WS(x)m= , (16)
4. Settlement D u e to B o r e d T u n n e l l i n g where Vs is the volume of the settlement trough and is equal
The methods currently adopted to predict settlements to the half the width of the settlement trough, W, multiplied
upon bored tunneling are generally categorized as sophisti- by the maximum settlement. Therefore,
cated/numerical and empirical correlation. Many research- V~ (%)= 2.507 i S(X)m / V x 100 (17)
ers [e.g,. Ghaboussi et al. (1983), Finno and Clough (1985),
Row and Lee (1992)] have adopted numerical methods Equations (15) and (17) have three unknowns (i, Vl and
based on simplifying hypotheses with respect to the geom- S(X)m:). Numerous empirical studies have been previously
etry and the soil constitutive behavior. Fang, Lin and Su attempted by different authors to estimate the distance to
(1994) pointed out t h a t although these methods might the point ofirdlection, i. One approach, presented by O'Reilly
provide a rational analysis of the problem, the cost and time and New (1982), is to assume that the distance i is a linear
consumed to perform such analyses are quite substantial. function of the depth of the tunnel. O'Reilly and New gave
The simulation of the variably different tunneling shields the following relationships
used in practice and the level of workmanship provide a i = 0.43z +1.1 for cohesive soils, and (18)
major difficulty, if not a challenge, for implementation in
i = 0.28z - 0.1 for cohesionless soils (19)
sophisticated numerical analyses. Accordingly, the use of
empirical correlation, based on actual field measurements, where i and z are in meters.
is mostly favored by m a n y designers as a more practical Equations (18) and (19) have been used successfully by
prediction method. other researchers to predict the ground surface settlements
Using the empirical prediction method requires the avail- in the case of homogenous soils, where the ground is uni-
ability of a reasonable database of measurements from case form from the surface to the tunnel. Unfortunately, this is
histories of tunneling in similar ground and under similar not the case for the Cairo Metro Tunnel, where the soil
conditions. profile is characterized by different layers.
It is now well established t h a t the shape of the settle- Other empirical studies for estimating the distance i
ment trough at the surface due to tunneling in soft ground have been summarized by H a m z a (1995) in the comprehen-
can be reasonably represented by the error function curve. sive report on Tunnel Monitoring. Most of these studies
Many authors have shown t h a t this approach adequately have shown that the value of i can be estimated with very
models the shape of the settlement trough (e.g. Peck 1969; close approximation from the geometric relationship,
Schmidt 1969; O'Roilly and New 1982; Ata 1996). i_ (20)
4.1. Procedure for Estimating the Settlement Trough where:
a = tunnel radius, and
Underground tunneling results in a state of stress relax-
ation in the surrounding soil. The volume of the soil that is K and n are empirical parameters that depend on the
displaced or relaxed across the perimeter of the tunnel is type of soil above the tunnel axis.
termed as the ground loss. The percentage ground loss, V~ The use of the following values of K and n has been
(%), is conventionally expressed as a percentage of the ratio previously suggested:
of the volume of lost ground to the theoretical volume of the (1) K = 1.00 and n = 0.80 for clays (Schmidt 1969;
tunnel, V. As a result of this volume loss the ground surface Clough and Schmidt 1981)
depresses, forming what is known as the settlement trough. (2) K = 1.00 and n = 1.00 for clays (Attewell 1981)
Many authors have used the normal distribution (or error

286 TUNNELLING ANDUNDERGROUND SPACETECHNOLOGY Volume 14, Number 3, 1999

 (3) K = 0.74 and n = 0.90 for granular soilsbelow water ratio of the depth of d a y to the tunnel depth. From the
table (Attewell 1981) figure we m a y conclude t h a t the volume loss decreases as
(4) K = 0.63 and n = 0..97 for granular soilsirrespective the depth of clay above the tunnel crown increases. For the
of water table (AtteweU 1981). ratio clay/tunnel depth ranging between 0.35-0.75, which
is the typical range for the future A1 Azhar Road Tunnels,
Since the profilefor t~e Cairo tunnel consistsgenerally the expected volume loss (using the same TBM) ranges
of a top made ground and cohesive layers underlain by a between 0.15% and 0.45%.
layer ofgranular soil,the positionofthe inflectionpoint was
obtained by combining the relationshipsgiven by (1)and (4)
as follows: 5. C o n c l u s i o n s
Top Layer Clay: Kc = 1 and nc = 1 The prediction techniques adopted for the Cairo Metro
(Attewell 1981) Line 2 to estimate the settlements due to cut-and-cover and
bored tunneling construction works are presented. C-ener-
Bottom Layer CohesLonless: Ks = 0.83 and ns =0.97 ally, the predicted settlements are in good agreement with
(Attewell 1981) the monitored records.
The equivalent consl~ant parameters will be taken by A new expression (So-Curve) is successfully used to
interpolation for the depth range starting from the ground estimate the settlement component due to installation of
surface to the point between the spring line and the crown diaphragm walls. Two patterns are proposed to model the
as adopted by H a m z a (1995) by ground surface settlements induced by the deflection of the
diaphragm wall (S and $1 Curves). The measured ground
K CK c+SK,,and n C n ~ + S n, surface settlements due to the deflection of the diaphragm
= C +S = C +S (21) wall (induced by excavation inside the walls) are in good
Values used in the above expression are defined in agreement and within the ranges predicted by the two
Figure 9. proposed patterns.
A simple procedure, based on commonly used methods,
is developed to predict the surface settlements due to bored
4.2. Analysis of the Measured Settlements tunneling. A new set of p a r a m e t e r s is presented to account
The data from Cairo Metro Line 2 for the monitored for the stratified soil formation of the Nile alluvial depos-
settlement profiles, porl~.ndicular to the
tunnel axis, are presented in Figure 10.
The data from perpendicular sections at Offset from center line of tunnel (m)
the different Lots are fitted to an error
function curve using the least square -20 -15 -I0 -S 0 5 I0 15 20
error technique. The back-calculated dis- 0
tance i varied at different Lots and dif-
ferent locations between 5.8 m and 8.3
m. Some of the observed relationships 0.005
for i/a versus z/2a are presented in Fig-
ure 11. The ratio i/a varied between 1.0
and 2.0 for Phase I and between 1.3 and
1.85 for LOts 40, 42 and ,46 (Phase 2A). - 0.01
The back-calculated volume losses at
the Cairo Metro Line 2 are presented in
Figures 12 and 13 versus the distance 0.01¢
along the tunnel route (kilopoint). Fig-
ure 12 presents the volume loss at the
beginning of tunneling :for Phase 1 at
Lots 12, 14 and 16, where the volume 0.02
[ ~

u
Lot 42
Lot 40
J
m
*
Lot 42
Lot 42
I i
--
: ...... Lot 46
Lot 46
I

loss averaged 0.5% and exceeded 0.8%

at some locations. Figure 10. Settlement trough at Lots 40, 42 and 46.
In contrast, the back-calculated vol-
ume losses for Phase 2A, (average 0.3%
and a maximum 0f0.5%) were much less 4
than those of the earlier Phase. Accord-
ingly, for Phase 2A, most of the predic-
tions for the m a x i m u m settlements --A- . . . . . . . . . . . . ..... -,;7 ....

(above the tunnel centerline) were 3

greater than the actual measured settle-
ments where the ratio of predicted to

i'i
measured averaged 1.7.2%e predictions
overestimated the measured values by ~/ 2 a 2
approximately 74%. The comparison of
volumes of settlements for Phase 2A
versus those at the begimning of tunnel-
ing (Phase 1) show a clear improvement n
with regard to the application of the
slurry overpressure. o Phase i
The results from the back-analysis
of the settlement data proved t h a t the o 8 I i I | J •

sensitivity of the vohtme loss to the 0 I 2 3 4

depth of clay above the tunnel crown. i/a
Figure 14 presents the back-calculated
volume loss for Phase 2A versus the Figure 11. Relationship (i / a) vs. (z l 2a).

Volume 14, Number 3, 1999 TUNNELL~GANDUNDERGROUNDSPACETECHNOLOGY287

 its. The procedure can be programmed eas-
ily on a PC and can be adopted for settle-
d ment prediction due to tunneling in similar
0.8 [ stratified formations. Successful implemen-
Lot 12 - ' ~ ' L o t 14 ~ L o t 16 ......
tation of such a procedure to other projects
requires the calibration of the parameters
involved based on actual field measurements
~ 0.6 that should be acquired during the first
phase of the project.
The settlement predictions due to tunnel-
= 0.4 ing for Line 2, Phase 2A, generally overesti-
mated the actual (measured) settlements by
approximately 74%, and the ratio of the
0.2 predicted to measured averaged 1.7. The
volume loss due to bored tunneling decreases
as the depth of clay above the tunnel crown
increases. For the future A1 Azhar Road
0 Tunnels, the expected volume loss (using the
3000 4000 5000 6000 same TBM) ranges between 0.15% and 0.45%.
Distance along Tunnel Axis (Kilopoints)
6. Acknowledgment
Figure 12. Volume loss from back analysis (Phase 1A). The authors wish to acknowledge the
National Authority for Tunnels for their
sincere cooperation during the execution of
the monitoring program of the Cairo Metro
Line 2.

"--'--Lot 40 ---"--Lot 42 - - L o I 46 7. References

Ata, A. 1996. Ground Settlements Induced by
0.8
Slurry Shield Tunneling in Stratified Soils.
North American Tunneling, Vol. 1 (L. Ozdemir,
g ed.), 43-50.
0.6 .................................................................................. Attewell, P.B. 1981. Engineering contract. Site
Investigation and Surface Movements. In
Tunneling Works, Soft-Ground Tunneling
-- 0.4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Failures and Displacements (D. Resendis and
M. P. Romo, eds.), 5-12. Rotterdam: Balkema
Attewell, P. and Selby, A. 1989. Tunneling in
compressible soils: large ground movements
0.2 and structural implications. Tunneling and
Underground Space Technology 4(4), 481--487.
Clough, G. W. and Schmidt, B. 1981. Design and
0 Performance of Excavations and Tunnels in
10000 ! 1000 ! 2000 ! 3000 14000 sott clay. InSoft Clay Engineering (E. W. Brand
and R.P. Brenner, eds.), 569-634. Oxford:
Distance along Tunnel Axis (Kilopoints) Elsevier.
Clough, G. W. and O'Rourke, T. D., 1990.
Figure 13. Volume loss from back analysis (Phase 2A). Construction induced movements of in situ
walls. ASCE Specialty Conference on Design
and PerformanceofEarth RetainingStructures,
GeotechnicalSpecialty Publication No. 25,439-
1.4 470. New York: ASCE.
Cording, E. J. and Hansmire, W. H. 1975.
Displacements around soft ground tunnels.
i.2 Proc. Fifth Panama Conf. on Soil Mechanics
and Foundation Engineering, Buenos Aires,
,.., ! Argentina, 572-633.
I. Road Tunnels_i Dyslis, M, and Fontana, A. 1982. Deformations
around excavations in clayey soil. Proc. Int.
0.S Syrup. on Numerical Models in Geomeehanics,
Zurich, Switzerland, 634-642. Rotterdam:
0.6 Balkema.
Fang, Y. S., Lin, J. S. and Su, C. S. 1994. An
estimation of ground settlement due to shield
0.4 tunneling bythe Peck-Fujita method. Canadian
Geotechnique 31, 431-443.
0.2 Finno, R. J. and Clough, G. W. 1985. Evaluation of
soil response to EPB shield tunneling. ASCE,
I i I
Journal of Geoteehnical Engineering 1112),
0 I
155-173.
0 0.2 0.4 0.6 0.8 1 1.2 1.4 Ghaboussi, J.; Hansmire, W. H.; Parker H.W.; and
Kim, K.J. 1983. Finite element simulation of
Depth of Clay /Tunnel Depth tunneling over subways. ASCE Journal of
Geotechnical Engineering 109(3), 318-344.
Figure 14. Effect of depth of clay on the percentage volume loss.

288 TUNNELLINGAND UNDERGROUNDSPACE TECHNOLOGY V o l u m e 14, N u m b e r 3, 1999

Hamza Associates. 1993. "Procedure for Settlement Prediction: Peck R. B. 1969. Deep excavations and tunneling in soft ground.
Diaphragm Walls, Cairo Metro Line No. 2. ~Report Preparecl for Stateofthe art report.Proc. 7th Int. Conference on Soil Mechanics
Campenon Bernard SGE. and Foundation Engineering, Mexico City, Mexico, 225--290.
Hamza Associates. 1993. ~Procedure for Settlement Prediction due Rowe, R. If, and Lee, K. M. 1992. An evaluation of simplified
to Bored Tunneling: Cairo Metro Line No. 2." Report submitted techniques for estimating three dimensional undrained Ground
to Caponon Bernard SGE, HamT.a Associates, Egypt. movements due to tunneling in soft soils. Canadian Geotechnical
Hamza Associate. 1995. ~Pu~nel Monitoring Comprehensive Report, Journal 29, 39-52.
Greater Cairo Metre Line 2 Phase 1.~ Report submitted to Sclunidt, B. 1969. Settlements and Ground Movements Associated
Campenon Bernard SGE, Hamza Associates, Egypt. with Tunneling in Soil. Ph.D. Thesis, University of Illinois,
O~Reilly, M. P. and New, B. M. 1992. Settlements above tunnels in 1969.
the U.K-- their magnitude and prediction. Tunneling "82, 173- Uriel, A. O, and Sagaseta, C. 1989. General Report, Discussion
181. Session 9. Proc. of the 12th Int. Conf. on Soil Mechanics and
Foundation Engineering, Rio de Janeiro, Vol. 4, 2521-2550.

Volume 14, Number 3, 1999 Tubr~hLINO AND UNDERGROUND SPACE TECHNOLOGY 289