Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
GEOTECHNICAL ENGINEERING
GEO-001
GEO-002
Abbas Moustafa
GEO-003
TWO DIMENSIONAL SIMULATION FOR THE EARTH RESSURE of Point Load Acting
Behind a Cantilever Wall
George Maurice Iskander
GEO-004
GEO-005
GEO-006
GEO-007
GEO-008
GEO-009
GEO-010
Abbas Moustafa
KAMAL GHAMERY METWALLY, MANAR MOHAMED HUSSEIN, and ADEL YEHIA AKL
GEO-011
GEO-012
GEO-013
GEO-014
GEO-015
GEO-016
Mostapha Boulbibane
ICSGE-13
13th ICSGE
27-29 Dec. 2009
Cairo - Egypt
ABSTRACT:
In order to investigate the spatial passive earth pressure, this boundary value problem
was modelled using the finite element method (FEM) with the program ABAQUS.
The soil was modelled as a sandy soil. The constitutive model used is based on the
theory of hypoplasticity. This special material law simulates the non-linear stress-strain
behaviour of the soil and takes implicitly the stress-dependence of the angle of internal
friction into account. The interface between the rigid walls and soil was modelled using
contact elements. The computations were developed for three different modes of the
wall movement (parallel movement, rotation around the top of the wall and rotation
around the toe of the wall).
The passive earth pressure coefficients for different wall dimensions (breadth/ height)
were calculated. The results are compared with the analytic computation models based
on the German code, in order to be able to derive statements about their quality and
range of their validity.
ICSGE-13
GEO-001-1
INTRODUCTION
ICSGE-13
GEO-001-2
earth pressure and developed an empirical formula to calculate the spatial earth
pressure. Ovesen (1964) conducted an extensive series of model tests on anchor plates
to investigate the 3-D effects. His tests showed that passive earth pressures are higher
than those predicted using conventional methods. Brinch Hansen (1966) developed a
method for correcting the results of conventional passive pressure theories behind
anchor plates to account for 3D effects. He found also that the dimension of the
supporting structure is an important factor for the analysis of 3D passive earth pressure.
Neuberg (2002) made an experimental model of a retaining wall (I-profile) at University
of Dresden in order to calculate the 3D passive earth pressure. Based on his results, a
computation approach for earth pressure at peak and a mobilization equation were
developed.
A number of investigators developed alternative theoretical approaches for determining
spatial passive earth pressure. Blum (1932) developed a failure mechanism, as in
Coulombs theory, by creating a three-dimensional flat failure surface. Based on the log
spiral method, a spreadsheet was developed by Duncan and Mokwa (2001) to calculate
the spatial passive earth pressure. They obtained a hyperbolic load-deflection
relationship to estimate the spatial passive earth pressure. Soubra and Regenass (2000)
developed a method to calculate 3-D passive earth pressure coefficients based on the
upper-bound method of limit analysis. The method consists of three phases considering
different block mechanisms. Merifield and Solan (2006) published numerical modelling
in order to calculate the capacity of anchors in frictional soils based on a certain failure
mechanism. Stefan (2007) used ABAQUS to investigate the behaviour of a solid pile
wall and derived a mobilization curve for medium-dense sand. Benmebarek et al. (2007)
performed a numerical study of 3D passive earth pressure by parallel movement of rigid
walls using FLAC3D. Their results were presented in design tables relating different
geometrical parameters and 3D passive earth pressure coefficients.
3
MATERIAL BEHAVIOUR
The numerical modelling was done for sandy soil. Regarding the well-reported
experimental data about Karlsruhe sand, the material parameters required were
determined for this sand. Karlsruhe sand consists mainly of sub round quartz grains.
The index properties of the sand are given in Table 1. The behaviour of Karlsruhe sand
in triaxial tests was investigated by Kolymbas & Wu (1990). Results with dense
samples (e0 = 0.53) for different confining pressures (3) are shown in Figure 1.
Table 1: Index properties of Karlsruhe medium sand
Unit weight of the grains, kN/m3
D10, mm
D60, mm
Uniformity coefficient, Cu
Min. void ratio, emin
Max. void ratio, emax
ICSGE-13
GEO-001-3
26.5
0.240
0.443
1.85
0.53
0.84
Stress ratio 1 / 3
-4-
Volumetric strain v
Axial strain 1
Axial strain 1
TJ = h(T, D, e )
with:
TJ:
h:
T:
D:
e:
(2)
The formulation for sands is rate independent, which means that the function h is
positive homogeneous of first order in D. The constitutive equation allows for the
application for large strain problems.
The numerical modelling is based on the version developed by von Wolffersdorf (1997)
in the following equation
ICSGE-13
GEO-001-4
[(
( ) ]
TJ = f b f e L T * , D + f d N T * D
(3)
with tensor functions L and N. D is the euclidian norm of the symmetric part of the
velocity gradient. The functions fe and fd describe the influence of density and fb the
influence of mean pressure. T* is the normalized cauchy stress tensor using the trace of
T. The constitutive equation has eight constants, which are measured in a simple way
from standard tests in soil mechanics. The eight constants used in the finite element
analysis are listed in table (2). For a detailed discussion about the mathematical
background and physical meaning of the input parameters for the constitutive model
refer to Herle (1997). For more details concerning this material law refer to Herle &
Gudehus (1999).
Table 2: Input parameters for hypoplastic material law for Karlsruhe Sand.
grain stiffness (hs) ed0
ec0
n
ei0
c
30.0
5800.0 MN/m2
0.53 0.84 0.25 1.00 0.13 1.05
4
NUMERICAL MODEL
For the numerical investigation with the finite element method (FEM), the program
ABAQUS was used. The main aspects of the modelling are listed below:
The modelling process is executed stepwise. Firstly, the primary stress state using own
weight of the soil medium is generated for different initial void ratios. The following
table (Table 3) shows the required parameters for this stage. The frictions angle was
ICSGE-13
GEO-001-5
estimated regarding the initial relative density. In the second step the required rigid
walls were moved under three different basic wall movements as mentioned before.
Table 3: Input parameters for primary stress state.
eo
void ratio
0.55
0.65
0.75
no
porosity
0.355
0.394
0.428
D
relative density
92.0%
56.8%
26.0%
()
friction angle
38.5
35.0
30.5
(/m3)
unit weight
17.1
16.1
15.1
Ko
earth press.
0.46
0.48
0.50
ICSGE-13
GEO-001-6
( B)
(4)
earth press.
earth press.
kp3Dfem/kp2DDIN D
ratio
()
0.75
30.5
3.10
5.0
1.61
1.35
0.65
35.0
3.70
6.0
1.62
1.42
0.55
38.5
4.30
8.0
1.86
1.48
From the previous table (Tab. 4), the correction factors obtained from FE-results are
higher than from the German code, which means that the German code is on the safe
side.
5.2 Effect of the wall deformation mode
In order to study the effect of wall deformation mode, another model for a mediumdense sand (e0=0.65) with the same spatiality ratio (B/H=1.0) under three different wall
movements was developed. In the case of rotation around the toe the limit value could
not be reached so an approximate value was calculated for U/H equal to 15%.
The earth pressure coefficient reaches its maximum value (6.0) by parallel movement
and the minimum value (4.0) by rotation around the toe of the wall (Fig. 5). Whereby
ICSGE-13
GEO-001-7
the rotation around the top of the wall gives a passive earth pressure coefficient equal to
5.10. This means that the coefficient by rotation around the top of the wall is almost
equal to 0.85 of the coefficient for parallel movement and the coefficient by rotation
around the toe is equal to 0.67 of the coefficient of parallel movement. Compared with
the German code DIN 4085, it is found that ratios are larger, i.e. again the German code
gives results lying on the safe side.
ICSGE-13
GEO-001-8
Fig. 6 (a): Passive earth pressure coefficient (kp) as a function of the normalized wall
displacement (U/H) for parallel movement
Fig. 6 (b): Passive earth pressure coefficient (kp) as a function of the normalized wall
displacement (U/H) for rotation around the top of the wall
Fig. 6 (c): Passive earth pressure coefficient (kp) as a function of the normalized wall
displacement (U/H) for rotation around the toe of the wall
The need for research regarding the spatial passive earth pressure problem is shown by
the numerical calculations. On one hand, the mobilization curves between the
normalized wall displacement (U/H) and the passive earth pressure should be derived
ICSGE-13
GEO-001-9
for different wall movement modes. On the other hand, earth pressure distribution
should be investigated. Also a parametric study regarding different soil-wall friction
angle and different wall heights should be performed.
REFERENCES:
[1]
[2]
[3]
Brinch Hansen. Comparison of Methods for stability Analysis Threedimensional Effect in stability analysis resistance of a rectangular anchor slabDanish Geotechnical Institute Bulletin No.21, 1966.
[4]
[5]
Duncan, M., Mokwa, R. Passive Earth Pressures -Theories and Tests, ASCE
journal of geotechnical and geoenvironment engineering, Vol. 127, No. (3),
248-257, (2001).
[6]
[7]
[8]
[9]
[10]
Ovsen, NK. Anchor slabs, Calculation methods and Model tests, Danish
Geotechnical Institute Bulletin, NO. 16, 5-39, 1964.
[11]
[12]
Stefan,J.
Nichtlinearer
Horizontaler
Bettungsmodulansatz
fr
Trgerbohlwnde in mitteldicht gelagertem Sand Verffentlichungen des
instituts fr Bodenmechanik und Grundbau, Technische Universitt
Kaiserslautern (2007).
[13]
[14]
ICSGE-13
GEO-001-10
13th ICSGE
Faculty of Engineering
Cairo - Egypt
ABSTRACT
This paper investigates the implications of the earthquake characteristics, such as, the
total energy and energy concentration of the ground motion on the earthquake input
energy, dissipated energies and damage of inelastic structures. Influences of the
structures parameters on the structural damage are also studied. The structural damage
is quantified in terms of Park and Ang damage indices. Strong ground motions recorded
in USA, Japan and Egypt are employed in the numerical investigations. The material
force-deformation relation is modelled using elastic-plastic and bilinear laws. The
numerical analyses carried out in the Matlab platform reveal the significant effect of the
earthquake parameters and the structural properties on damage of inelastic structures.
KEYWORDS
Strong ground motion, earthquake characteristics, inelastic structures, elastic-plastic,
input energy, hysteretic energy, damage indices, Egypt National Seismic Network.
1 INTRODUCTION
Damage of structures during earthquakes depends on the structural properties (natural
frequencies, mode shapes, damping mechanism, material of construction and the
structural model) and also on the earthquake characteristics (source properties, epicentre
distance, earthquake intensity, duration, Fourier amplitude and frequency content). The
12 October 1992 Dahshur earthquake that hit the Greater Cairo area is a remarkable
example on this aspect. Several buildings were severely damaged during this earthquake
while other buildings experienced minor or no damage [1]. In earthquake engineering it
is essential to model nonlinear structural behaviour, which in turn is encapsulated in the
structural model adopted [2, 3]. When dealing with inelastic structures it is of interest to
ICSGE-13
GEO-002-1
investigate the structural response in terms of modern measures using energy concepts
and damage indices rather than the structural deformations [4-6]. This is because
structures get damaged due to not only maximum displacement excursion but also due
to accumulated hysteretic energy dissipated by repeated yielding [7-9].
According to the Egyptian seismic code, the nonlinear time-history analysis of
structures constitutes the most accurate way for predicting the structural responses
under strong ground motion [10]. The establishment of the Egyptian national Seismic
Network (ENSN) during 1993-1997, following the 1992 Dahshur earthquake, provided
seismic data on recorded strong ground motions for Egypt territory. However, given
that adequate information on soil conditions beneath recording stations in Egypt is not
currently available, this paper examines implications of adopting alternative earthquake
records, in terms of their characteristics, such as, energy and energy concentration, and
also the structural properties on the structures damage. The next section demonstrates
the dynamic analysis of inelastic structures to earthquake loads. Section 3 explains the
quantification of the structural response using energy concepts. The development of
damage indices for inelastic structures under seismic loads is demonstrated in section 4.
Section 5 provides numerical illustrations and discussions.
2 RESPONSE OF INELASTIC STRUCTURES TO EARTHQUAKE LOADS
The equation of motion for an N multi-degree-of-freedom (MDOF) nonlinear structure
driven by a single component of earthquake acceleration &x&g (t ) is given by [11]
&& (t ) + CX
& (t ) + F (t ) = P(t ) = M r &x& (t )
MX
s
g
(1)
where, M and C, are the mass and the damping matrices of the structure, respectively,
Fs (t ) is the vector of hysteretic restoring forces, r is a vector of ones, X(t) is the
structure displacement vector and dot indicates differentiation with respect to time. Note
that, for nonlinear damping models, the damping matrix C is a time dependent function.
To proceed further, the incremental form of equation (1) can be written as
&& + CX
& + K X = P = M{1}[ &x& (t ) &x& (t )]
MX
(2)
s
g
k +1
g
k
&& = X
&& (t ) X
&& (t ); X
& =X
& (t ) X
& (t ); X = X(t ) X(t )
X
k +1
k
k +1
k
k +1
k
(3)
where Ks is the stiffness corresponding to the displacement from X(tk) to X(tk+1). Using
Newmark- method, the velocity and displacement responses at time tk+1 are given as
&
& + (1 ) X
&& t + X
&& t
X
=X
k +1
k +1
(4)
& t + ( 1 ) X
&& (t ) 2 + X
&& (t ) 2
X k +1 = X k + X
k
k
k +1
2
where and are the parameters of the Newmark- method and t = t k +1 t k is the
time step. Equations (4) can be recast in an incremental form as
& = (1 )X
&& t + X
&& t ; X = X
& t + ( 1 )X
&& (t )2 + X
&& (t )2
(5)
X
k
k +1
k
k
k +1
2
Substituting equations (5) into equation (2) we get
&& + C[(1 ) X
&& t + X
&& t ] + K [ X
& t + ( 1 ) X
&& ( t ) 2 + X
&& ( t ) 2 ] = P
MX
k
k +1
s
k
k
k +1
2
ICSGE-13
GEO-002-2
(6)
&& = X
&& + X
&&
Substituting X
k +1
k
&& + (1 )tCX
&& + tCX
&& + tCX
&& + tK X
& + ( 1 )(t ) 2 K X
&&
MX
k
k
s
k
s
k
2
&& + (t ) 2 K X
&& = P
+ (t ) 2 K s X
k
s
(7)
(8)
&& we get
Solving for X
&& = [M + tC + (t ) 2 K ]1[P tCX
&& tK X
& 1 ( t ) 2 K X
&& ]
X
s
k
s
k
s
k
2
(9)
1
&& ; B = M + t C + ( t ) 2 K
( t ) 2 B 1 K s ] X
k
s
2
(10)
&& is given as
From equation (10), the acceleration X
k +1
&& = X
&& + X
&& = X
&& + B 1 P [ tB 1 K ]X
& [ tB 1C + 1 ( t ) 2 B 1 K ]X
&&
X
k +1
k
k
s
k
s
k
2
(11)
1
1
1
1
2
1
&
&
&
= B P [ tB K s ]X k + [I tB C (t ) B K s ]X k
2
Substituting equations (11) into equations (5) and making use of equation (3), we get
X k +1
X k
&
&
&&
&&
(12)
X k +1 = G X
k + H P or q(t k +1 ) = G q(tk ) + H [ x g (t k +1 ) x g (t k )];
X
X
&&
&&
k +1
k
3
1
I (t )I (t ) B K s
G = 0
I (t ) 2 B 1K s
0
tB 1K s
1
1
(t ) 2 I (t ) 3 B 1C (t ) 4 B 1K s
2
2
1
2
1
tI (t ) B C (t ) 3 B 1K s ;
2
1
1
2
1
I tB C (t ) B K s
2
(t ) 2 B 1M
H = tB 1M (13)
B 1M
Note that the matrices Ks, G and H are time dependent functions. In the above
formulation a viscous damping model is considered, and, thus the damping matrix C is
constant. Hysteretic nonlinear damping can also be considered in which the damping
matrix is treated as a time dependent function (see Section 5).
The stiffness used in the solution for time step tk+1 is taken as the secant stiffness from
time step tk-1 to tk. To correct for this approximation, an iterative procedure for the
ICSGE-13
GEO-002-3
stiffness Ks is performed where the initial stiffness is replaced by Ks(1) , Ks(2), , Ks(n)
until a convergence criterion on Ks is achieved. The next section demonstrates the use
of energy concepts in characterising the structural inelastic response.
3 INPUT AND DISSIPATED ENERGIES BY INELASTIC STRUCTURES
Several studies have characterized the structure response in terms of earthquake input
energy and energy dissipated by the structure [4-6]. These energy terms are quantified
by integrating the structure equation of motion as follows
t
T
T
T
T
X& ( )MX&& ( )d + X& ( )CX& ( )d + X& ( )Fs ( )d = X& ( )M{1}&x&g ( )d
(14)
The right side of equation (14) represents the input energy to the structure since ground
starts shaking until it comes to rest. The first energy term of the left side is the relative
kinetic energy EK(t) of the masses associated with their motion relative to the ground
t
N
& T ( )MX
&& ( )d = 1 m x& 2 (t )
E K (t ) = X
(15)
i i
2
i
=
1
0
The second term in equation (14) is the energy dissipated by damping ED(t)
t
i =1 0
& T ( )CX
& ( )d = x& ( ) f ( )d
E D (t ) = X
i Di
(16)
N
For viscous damping models, the above expression reduces to cij x& i ( ) x& j ( )d .
i =1 j =1 0
The third term of equation (14) is the sum of the recoverable strain energy Es(t) and the
hysteretic cumulative plastic energy dissipated by yielding EH(t)
N
E H (t ) = x& i ( ) f si ( )d Es (t )
(17)
i =1 0
The recoverable strain and kinetic energies vanish by the end of the earthquake duration
and the earthquake input energy to the structure is absorbed by damping and hysteretic
mechanisms. The next section explains the use of maximum deformation and hysteretic
energy in developing damage indices for inelastic structures.
4 DAMAGE MEASURES FOR INELASTIC STRUCTURES
The literature on damage measures for structures under earthquake ground motions is
vast, see e.g., Refs. [12-14]. Damage indices can be based on either a single or
combination of structural response parameters. Examples of damage measures that are
based on a single response parameter are the ultimate ductility and number of yieldings
during ground shaking [12,14]. Clearly, these measures do not incorporate information
on how the earthquake input energy is imparted on the structure nor how this energy is
dissipated. Earthquake damage occurs due to not only the maximum deformation or
ductility but is associated with hysteretic energy dissipated by the structure as well. The
definition of structural damage in terms of the ductility factor is therefore inadequate.
The rate of the earthquake input energy to the structure has been used as a measure of
the structural damage [15]. Damage indices can be estimated by comparing the response
parameters demanded by the earthquake with the structural capacities. Ref. [14]
ICSGE-13
GEO-002-4
proposed a damage index in terms of the ultimate ductility (capacity) u and the
maximum ductility attained during ground shaking (demand) max
DI =
x max x y
xu xy
max 1
u 1
(18)
However DI does not include effects from hysteretic energy dissipation. References
[12,13] proposed a damage index based on the structure hysteretic energy E H
DI H =
E H /( f y x y )
(19)
u 1
where f y , x y are the yield force and displacement. A robust damage measure should
include not only the maximum response but also the effect of repeated cyclic loading.
Park and Ang developed a simple damage index given as [7,16,17]
E H /( f y x y )
x
E
(20)
DI PA = max + H = max +
u
u
xu
f y xu
Here, x max , E H are maximum absolute displacement and dissipated hysteretic energy
under the earthquake. x u is the ultimate deformation capacity under monotonic loading
and is a positive constant that weights the effect of cyclic loading on structural
damage. Note that if = 0, the contribution to DIPA from cyclic loading is omitted.
The state of the structure damage is defined as: (a) repairable damage ( DI PA < 0.40 ),
(b) damaged beyond repair ( 0.40 DI PA < 1.0 ), and (c) total collapse ( DI PA 1.0 ).
These criteria are based on calibration of DIPA against experimental results and field
observations during earthquakes [17]. Equation (20) reveals that both maximum
ductility and hysteretic energy dissipation contribute to the structure damage during
ground motions. Herein damage is expressed as a linear combination of damage caused
by excessive deformation and that contributed by repeated cyclic loading effect. The
quantities x max , E H depend on the loading history while the quantities , x u , f y are
independent of the loading history and are determined from experimental tests.
Equation (20) can be used for a structural member or a single-degree-of-freedom
structure. For MDOF structures, the global damage index is expressed as a weighted
sum of members local damage indices
N
i =1
i =1
DI G = i DI i ; i = E i / E i ; i = 1,2,..., N
(21)
Herein E i is the energy absorbed by the ith member. In this paper we adopt Park and
Ang damage indices in quantifying the structural damage. The next section provides
numerical illustrations and discussions on the formulation developed in this paper.
5 NUMERICAL ILLUSTRATIONS AND DISCUSSIONS
5.1 Elastic-plastic single-storey frame building
ICSGE-13
GEO-002-5
1966 Parkfield
Acceleration / g
Acceleration / g
1940 El Centro
0.5
-0.5
-1
0
0.5
-0.5
10
20
30
-1
0
40
10
Time [s]
20
1995 Kobe
2001 Egypt
0.5
Acceleration / g
0.5
Acceleration / g
40
Time [s]
-0.5
-1
0
30
-0.5
10
20
Time [s]
30
40
-1
0
10
20
Time [s]
30
40
Station site
M Dss (km) Component td (s)
7.0 12.99
EW
40.00
ELC#9
6.1 31.04
C02065
43.69
Chol#2
6.9 13.12
TAK000 40.96
Takatori
45.00 Ariesh quarry
6.2 30.00
N
Remarks
USA
USA
Japan
Egypt
This example examines the earthquake input energy, dissipated energy by the structure
and damage state of a single-storey inelastic frame structure to the earthquake records
shown in Fig. 1. Information on these records is provided in Table 1. These records
include recorded ground motions in USA, Japan and Egypt. Note that Parkfield and
Kobe records have narrow frequency contents compared to El Centro and Egypt records
as confirmed by the Fourier amplitude spectra. This could be attributed to source, path
and local soil site effects. All records are normalized to 0.51 g peak ground acceleration.
The Arias intensity measure (square root of area under square of ground acceleration)
for these records was computed as 28.24, 12.77, 37.74 and 17.21 m/s1.5, respectively.
The material nonlinearity is modelled using elastic-plastic force-deformation relation.
The frame supports a total mass of 9 103 kg and the initial stiffness of the columns
k 0 = 3.55 105 N/m. The initial natural frequency of the structure was computed as
n = 1.0 Hz. A viscous damping model with 0.05 damping ratio is adopted. The yield
ICSGE-13
GEO-002-6
strength of columns in tension and compression are taken as 2 104 and -2 104 N,
respectively. The dynamic analysis is performed following the formulation of Section 2.
Figure 2 depicts the input and dissipated energies by the structure from the four records.
The maximum ductility factor is computed as 3.50, 2.45, 8.89 and 0.28 for El Centro,
Parkfield, Kobe and Egypt records, respectively. Thus, Egypt record does not drive the
structure into the inelastic range although the Arias intensity is higher than that for
Parkfield record. In fact, Egypt record has its energy located at the high frequency
range. When k0 is changed such that n = 2.0 Hz, max is computed as 4.45, 13.23,
16.56 and 1.03, respectively. This points towards the effect of energy concentration of
Parkfield and Kobe accelerations compared to El Centro and Egypt records [18, 19].
The total energy alone does not significantly affect the structural maximum ductility.
10
x 10
1940 El Centro
1966 Parkfield
1995 Kobe
2001 Egypt
(a)
6
4
2
0
0
10
20
30
Time [s]
40
x 10
EH N m
EI N m
50
(b)
1940 El Centro
1966 Parkfield
1995 Kobe
2001 Egypt
0
0
10
20
30
40
Time [s]
Earthquake
1940 El Centro
1966 Parkfield
1995 Kobe
2001 Egypt
Linear behaviour, RD = repairable damage, DBR = damaged beyond repair, TC = total collapse.
The damage state of the structure is determined using Eq. (20) with = 0.15, u = 8.0
(see Table 2). The effect of the structure initial natural frequency, the frequency content
and concentration of acceleration energy on DIPA is obvious from the numerical values
of Park and Ang damage index shown in Table 2.
5.2 Bilinear inelastic two-storey frame building
The two-storey building frame shown in Fig. 3 is considered under the same set of
earthquake records of Table 1. This structure was studied by Moustafa [20] for
modelling critical design earthquake loads on inelastic structures. The behaviour of
braces 1 and 2 is taken as bilinear (k1= k0, = strain hardening ratio). The mass and
initial stiffness matrices of the structure are given as
ICSGE-13
GEO-002-7
m
M= 1
0
A E
A1E A2 E
+
2
0
L
L2
L2
; K el = cos 2 1
A2 E
m2
A2 E
L2
L2
m2
x2
E, A2
Brace
m1
9.14 m
x1
E, A1
9.14
&x&g (t )
(22)
Brace
9.14 m
&x&g (t )
1
L
L
1 1
1
(1 + )
2
AE
AE
cos 2
; K 12 = K el =
cos 2
K2 =
L
L
K el =
(23)
The first two natural frequencies of the elastic structure are computed as 1 = 6.18 rad/s
and 2 = 16.18 rad/s. A Rayleigh proportional damping C = aM + bK s with a =
0.2683, b = 0.0027 is adopted. These values are selected such that the damping ratio in
the first two modes is 0.05. Let the yield strain of braces y = 0.002 for both tension
ICSGE-13
GEO-002-8
input energy compared to the strain-hardening ratio (see Fig. 5). Table 3 summarizes the
global damage index for alternative values of which confirms these results.
25
20
20
10
1940 El Centro
1966 Parkfield
1995 Kobe
2001 Egypt
(a)
EH kN m
EI kN m
15
0
0
1940 El Centro
1966 Parkfield
1995 Kobe
2001 Egypt
15
(b)
10
5
10
20
Time [s]
30
0
0
40
10
20
30
40
Time [s]
Fig. 4: Input energy and energy dissipated by the two-storey inelastic structure (a)
Input energy (b) Hysteretic energy
20
20
15
= 0.20
= 0.10
= 0.05
=0
10
(a)
EH kN m
EI kN m
15
0
0
10
20
30
40
Time [s]
10
(b)
5
0
0
10
20
= 0.01
= 0.03
= 0.05
= 0.10
30
40
Time [s]
Fig. 5: Influence of structure properties on earthquake input energy for the two-storey
inelastic structure (a) Strain-hardening ratio (b) Damping ratio
Table 3: Global damage index for the two-storey frame building ( = 0.15, u = 8.0 )
Earthquake
1940 El Centro
1966 Parkfield
1995 Kobe
2001 Egypt
Strain-hardening ratio ( )
0
0.05
0.10
0.44 (DBR) 0.41 (DBR) 0.41 (DBR)
0.39 (RD) 0.38 (RD) 0.38 (RD)
0.75 (DBR) 0.69 (DBR) 0.68 (DBR)
-*
-*
-*
0.20
0.39 (RD)
0.36 (RD)
0.66 (DBR)
-*
Linear behaviour, RD = repairable damage, DBR = damaged beyond repair, TC = total collapse.
CONCLUDING REMARKS
This paper investigated the implications of the earthquake characteristics and the
structural properties on the damage of inelastic structures. The structural damage is
quantified in terms of Park and Ang damage indices. Earthquake accelerograms
recorded in USA, Japan and Egypt are adopted as seismic inputs to the inelastic
structures considered. The material force-deformation relation is modelled using elastic-
ICSGE-13
GEO-002-9
plastic and bilinear laws. It is shown that the concentration of acceleration energy close
to the fundamental frequency of the linear structure has a substantial effect on
producing large structural damage. Other parameters such as the acceleration energy do
not significantly affect the structural damage compared to energy concentration. The
damping ratio is also seen to substantially reduce the damage index compared to the
strain-hardening ratio which has smaller effect on the structural damage.
REFERENCES
[1] Sadek, A.W. Earthquake Engineering & Structural Dynamics 26 (1997) 529-540.
[2] Otani, S. Journal of Faculty of Engineering, Tokyo university 36(2) (1981) 407-441.
[3] Takeda, T., Sozen, M.A. and Nielsen, N. Journal of Structural Division, ASCE
96(ST12) (1970) 2557-2573.
[4] Zahrah, T.F. and Hall, W.J. Journal of Structural Engineering, ASCE 110 (1984)
1757-1772.
[5] Uang, C.M. and Bereto, V.V. Earthquake Engineering & Structural Dynamics 19
(1990) 77-90.
[6] Takewaki, I. Journal of Structural Engineering, ASCE 130 (2004) 1289-1297.
[7] Park, Y.J. and Ang, A.H-S. Journal of Structural Engineering, ASCE 111(4) (1985)
722-739.
[8] Moustafa, A. Journal of Structural Engineering, ASCE (2009) in press.
[9] Decanini, L.D. and Mollaioli, F. Soil Dynamics and Earthquake Engineering 21
(2001) 113-137.
[10] Permanent Committee for Code-201. The Egyptian code for calculation of loads
and forces on structural works and buildings, Cairo (2008).
[11] Hart, G.C. and Wong, K., Structural dynamics for structural engineers, John Wiley
& Sons, NY (2000).
[12] Cosenza, C., Manfredi, G. and Ramasco, R. Earthquake Engineering & Structural
Dynamics 22 (1993) 855-868.
[13] Ghobara, A., Abou-Elfath, H. and Biddah, A. Earthquake Engineering &
Structural Dynamics 28 (1999) 79-104.
[14] Powell, G.H. and Allahabadi, R. Earthquake Engineering & Structural Dynamics
16 (1988) 719-734.
[15] Takewaki, I. Journal of Engineering Mechanics, ASCE 132 (2006) 990-1000.
[16] Park, Y.J., Ang, A.H-S. and Wen, Y.K. Journal of Structural Engineering, ASCE
111(4) (1985) 740-757.
[17] Park, Y.J., Ang, A.H-S. and Wen, Y.K. Earthquake Spectra, 3(1) (1987) 1-26.
[18] Moustafa, A. Soil Dynamics and Earthquake Engineering, 29 (2009) 1181-1183.
[19] Cao, H. and Friswell, M.I. Soil Dynamics and Earthquake Engineering, 29 (2009)
292299.
[20] Moustafa, A. Journal of Sound and Vibration, 325 (2009) 532-544.
ICSGE-13
GEO-002-10
13th ICSGE
Faculty of Engineering
Cairo - Egypt
ABSTRACT
The problem of earth pressure due to point load acting behind a cantilever wall is a three
dimensional one. The lateral earth pressure varies in the vertical and the transverse
directions. In this research, the equations of earth pressure due to point load described in
the Egyptian Code of Practice are integrated and submitted to analytical procedure
through which an equivalent two dimensional equations are developed. A new
parameter known as the design width, which is the horizontal distance retained by the
individual vertical retaining element is introduced. The used procedure and the resulted
equations are tested and verified, by adopting different cases of the design width values
covering the practical range. The resulted equations can be easily used together with the
equations of the line load described in the Egyptian Code of Practice for two
dimensional analysis of the point load.
KEYWORDS
Earth pressure, retaining wall, cantilever wall, point load, and line load.
1
INTRODUCTION
The retaining structures are used to avoid the collapse of the retained soil, and to
support the adjacent structures. The reliable prediction of the effect of external loads on
the retaining structure is very important step to achieve a safe and economic design. The
supporting wall should be designed to limit ground movement, Chandrasekaran and
Hong [1].
There are a number of methods to estimate the effect of external loads on the adjacent
retaining wall. The elasticity solution developed by Boussinesq has been commonly
adopted to estimate the earth pressure due to external loads. Many of the external loads
can be represented by point and line loads. Researchers investigated the lateral earth
pressure on a retaining wall due to point and line loads; Kim and Barker [2], Greco [3],
Wang [4], Greco [5], Wang [6], Wang [7], and El-Attar [8].
ICSGE-13
GEO-003-1
Analytical closed form solutions have the great advantage of providing easy, direct, and
powerful tool for the design process. The Egyptian Code of Practice [9] provides semiempirical solutions for the cases of point and line loads. In this research, a new attempt
has been performed and discussed to achieve an easier analytical procedure for
analyzing the problem of a point load, by adopting the Egyptian Code equations.
2
The Egyptian Code of Practice adopts the semi-empirical expressions developed from
experimental work by Terzaghi [10], to estimate the earth pressure due to external
loads.
2.1 Case of Point Load
Consider a point load on the surface of the ground at distance (x) from a cantilever
unyielding rigid wall as shown in Fig. (1). For a free height of wall (H), the lateral earth
pressure distribution on the wall due to point load (Q) in the vertical and the transverse
directions is shown in Fig. (1-a) and Fig. (1-b), respectively.
a- Vertical direction
b- Transverse direction
Fig. 1: Lateral earth pressure distribution on the wall due to point load
The induced earth pressure (
on the wall for a point at depth (z) located along the
vertical line just in front of the point load, is given by NAVFAC [11] as follows:
for m > 0.4
(1)
for m 0.4
(2)
Where;
(3)
ICSGE-13
GEO-003-2
(4)
The corresponding earth pressure (
for any other point at the same depth, located at
horizontal distance (y) apart from the load location, along the wall in the transverse
direction, is given by U.S. Army Corps of Engineers [12], as follows:
(5)
Where; (
formula:
is the horizontal angle as shown in Fig. (1-b), and can be calculated from the
(6)
Consequently:
(7)
2.2 Case of Line Load
Consider a case of line load (q) on the surface of the ground at distance (x) from a
cantilever unyielding rigid wall as shown in Fig. (2).
Fig. 2: Lateral earth pressure distribution on the wall due to line load
For a free height of wall (H), the induced earth pressure (
depth (z) is given by NAVFAC [11] as follows:
(8)
for m 0.4
(9)
Where; (m) and (n) are as previously given by equations (3) and (4), respectively.
3
The analysis of a cantilever wall in the vertical direction is a plane strain problem. For
the case of an adjacent point load, the induced earth pressure is usually approximated to
a certain plane strain distribution in order to be included in the analysis. This process is
generally inaccurate one, which may lead to uneconomic or unsafe designed sections.
ICSGE-13
GEO-003-3
Gourvenec et al. [13] highlighted the potential significance of three dimensional effects
in the case of a retaining wall. Uncertainties in the geotechnical design variables have a
significant impact on the safety of cantilever retaining walls as stated by Goh et al. [14].
Therefore, it is a great need to consider properly the three dimensional effect of the
point load when analyzing the wall in the vertical direction.
3.1 Design Width
It would be a practically oriented step to define a parameter known as the design width
(2D), which is the horizontal distance retained by the individual supporting element. In
case the supporting elements are not completely working together laterally, the design
width is the width retained by single pile for tangent or secant piles, and width retained
by single panel for case of diaphragm wall, as shown in Fig. (3). For vertical elements
completely working together laterally, this width could be chosen a considerable design
value in the lateral direction or can be evaluated based on practical experience
considering the construction circumstances.
Fig. 3: Lateral earth pressure distribution along the design width due to point load
The plane strain analysis of the wall should be carried out considering the additional
earth pressure due to the point load. The most important case to be considered is the
case of the design width is totally centered with respect to the load location as shown in
Fig. (3), which will result in the highest additional stress due to point load.
Consequently, it is important to evaluate the average earth pressure distribution in the
lateral direction throughout this design width. The average pressure value cannot be
directly obtained, but it should be evaluated through an integration procedure.
3.2 Integration Procedure
It is now required to evaluate the horizontal stress area (2A) bounded by the curve
within the design width as shown in Fig. (4-a).
An infinitesimal portion of this area (dA) of width (dy), shown in Fig. (4-b) is calculated
as follows:
(10)
ICSGE-13
GEO-003-4
a- Area (2A)
Fig. 4: Area of stress in front of the design width due to point load
Due to the symmetry of the curve, we can calculate half of this area (A) for half the
design width (D) from the following integral:
(12)
Using equation (11), the integral form becomes:
(13)
The exact closed form solution of this integral is not available. However, a very close
solution can be obtained through the following formula:
(14)
By substituting with the integral limits to evaluate its value, we finally get:
(15)
And this represents half the symmetric horizontal stress area bounded by the curve
within the design width. The value of the angle (
) should be in radians.
Before using this final result, it is important to check the validity of this formula.
4
TECHNIQUE VERIFICATION
Before applying the developed formula, it should be carefully checked, to achieve safe
design for the wall and provide sufficient stability for the ground and the neighbouring
buildings. The induced ground movement affects the building damage risk-assessment,
Aye et al. [15], and the building serviceability assessment, Hsiao et al. [16].
ICSGE-13
GEO-003-5
The method of evaluating the accuracy of the formula obtained in equation (15), is
based on calculating the area (A) by another method and comparing the results. The
other method will be the numerical integration; by dividing this area into large number
(100) of small rectangles each having area of (dA), and summing up the areas of these
rectangles to calculate the area (A). The width of each rectangle (dy) will be (D/100).
This testing procedure is been carried out through preparing an Excel sheet. Two
proposed examples are presented and each has two different cases.
4.1 First Numerical Example
The data for this example are: Q = 1000kN, x = 2m, H = 5m, and z = 1m. Therefore
using equations (3) and (4), we get m = 0.4 and n = 0.2, respectively. Since m = 0.4,
then we shall use equation (2); therefore h = 56.0kN/m2. This example is solved for
two cases.
In the first case; D = 0.5m. By substitution in equation (15), we get A = 27.44kN/m. By
evaluating this area numerically after dividing it into 100 slices as described above, we
get A = 27.31kN/m. The percentage of difference between both methods is 0.5%.
In the second case; D = 2.0m. By substitution in equation (15), we get A = 87.96kN/m.
By evaluating this area numerically after dividing it into 100 slices as described above,
we get A = 83.40kN/m. The percentage of difference between both methods is 5.5%.
4.2 Second Numerical Example
The data for this example are: Q = 2000kN, x = 3m, H = 6m, and z = 1.5m. Therefore
using equations (3) and (4), we get m = 0.5 and n = 0.25, respectively. Since m = 0.5,
then we shall use equation (1); therefore h = 50.35kN/m2. This example is solved for
two cases.
In the first case; D = 1.0m. By substitution in equation (15), we get A = 48.60kN/m. By
evaluating this area numerically after dividing it into 100 slices as described above, we
get A = 48.21kN/m. The percentage of difference between both methods is 0.8%.
In the second case; D = 3.0m. By substitution in equation (15), we get A = 118.63kN/m.
By evaluating this area numerically after dividing it into 100 slices as described above,
we get A = 112.48kN/m. The percentage of difference between both methods is 5.5%.
4.3 Technique Overview
Two different numerical examples using practical values for the different parameters,
and each has two different cases of the design width values covering the range of the
practical limits, are adopted as testing tools for the validity and applicability of the
developed formula in equation (15). The results indicate that the difference between the
developed formula and the numerical integration is less than 1.0% for values of design
width (2D) of about 1.0 to 2.0m, and is less than 6.0% for values of design width (2D)
of about 4.0 to 6.0m. All the obtained differences are in the direction of the developed
formula; the results of the developed formula are greater than those of the numerical
integration, and this is in the safe side and gives more confidence for the resulted
design.
ICSGE-13
GEO-003-6
We can conclude that the developed formula in equation (15) can be practically applied
to obtain a plane strain distribution for the earth pressure due to point load within the
distance of the design width.
5
From the Egyptian Code equations, it is clear that the earth pressure due to a point load
is a three dimensional distribution, whereas the earth pressure due to a line load is a two
dimensional one. Practically, engineers prefer simple unsophisticated tools in
calculating straining actions for design, considering both economic and safety purposes.
Therefore, the final objective of this research is to develop a two dimensional
simulation of the earth pressure due to a point load, by making use of the previously
developed and verified technique.
5.1 Average Stress
Applying equation (15), thus, the average earth pressure throughout the design width is:
(16)
By substituting with the value of (
from equations (1) and (2) into equation (16), we
obtain equations (17) and (18), respectively:
for m > 0.4
(17)
for m 0.4
(18)
ICSGE-13
GEO-003-7
certain depth. For different depths, the value of (n) varies, and consequently, the
resulted value of ( ) differs. Therefore, the earth pressure distribution along the wall
height, equivalent to a point load, will be obtained using different resulted values of
( ). Nevertheless, this obtained distribution has the privilege of being calculated once
and can be safely used as a plane strain distribution within the specified design width
(2D).
6
CONCLUSIONS
In this research, the semi-empirical expressions of earth pressure due to point load and
line load, described in the Egyptian Code of Practice are reviewed. A new parameter
known as the design width (2D), which is the horizontal distance retained by the
individual vertical retaining element is introduced. The case of the design width is
totally centered with respect to the point load location which results in the highest
additional stress due to point load, is considered and subjected to analytical procedure,
through which an integrated formula is developed to evaluate half the symmetric
horizontal stress area (A) bounded by the curve of the transverse earth pressure
distribution within the design width.
Two different numerical examples each has two different cases of the design width
values covering the range of the practical limits, are tested and compared with the
numerical integration results as a corresponding technique. The results indicate that the
difference between the developed formula and the numerical integration is less than
1.0% to 6.0%, and the difference is usually in the safe side for design. The developed
formula becomes valid and can be applied practically to obtain a plane strain
distribution for the earth pressure due to point load within the distance of design width.
The developed formula is used to calculate the equivalent corresponding line load
formula. The resulted equations can be easily used together with the equations of line
load described in the Code of Practice for providing a two dimensional analysis of the
point load.
REFERENCES
[1]
[2]
Kim, J.S., and Barker, R.M. Effect of live load surcharge on retaining walls and
Abutments, Geotechnical and Geoenvironmental Engineering, ASCE 128 (10)
(2002) 803-813.
[3]
[4]
Wang, C.D. Lateral stress caused by horizontal and vertical surcharge strip
loads on a cross-anisotropic backfill, International Journal for Numerical and
Analytical Methods in Geomechanics 29 (14) (2005) 1341-1361.
[5]
ICSGE-13
GEO-003-8
[6]
Wang, C.D. Lateral force induced by rectangular surcharge loads on a crossanisotropic backfill, Geotechnical and Geoenvironmental Engineering, ASCE
133 (10) (2007) 1259-1276.
[7]
Wang, C.D. Lateral force and centroid location caused by horizontal and
vertical surcharge strip loads on a cross-anisotropic backfill, International
Journal for Numerical and Analytical Methods in Geomechanics, 31 (13) (2007)
1443-1475.
[8]
El-Attar, A.N. The effect of external load on the excavation supporting system,
M.Sc. Thesis, Ain Shams University, Cairo, Egypt (2009).
[9]
Egyptian Code of Practice for Soil Mechanics and Foundations, Part 7 (2001).
[10]
[11]
NAVFAC. Foundations & Earth Structures, Design Manual 7.02, United States
Department of Navy, Naval Facilities Engineering Command (1986).
[12]
U.S. Army Corps of Engineers. Engineering and design retaining and flood
walls, Manual EM No. 1110-2-2502, U.S. Army, Washington, D.C. (1989).
(Quoted from Kim and Barker [2]).
[13]
[14]
Goh, A.T.C., Phoon, K.K., and Kulhawy, F.H. Reliability analysis of partial
safety factor design method for cantilever retaining walls in granular soils,
Geotechnical and Geoenvironmental Engineering, ASCE 135 (5) (2009) 616622.
[15]
Aye, Z.Z., Karki, D., and Schulz, C. Ground movement prediction and building
damage risk-assessment for the deep excavations and tunneling works in
Bangkok subsoil, International Symposium on Underground Excavation and
Tunnelling, Bangkok, Thailand (2006) 281-297.
[16]
Hsiao, E.C.L., Schuster, M., Juang, C.H., and Kung, G.T.C. Reliability analysis
and updating of excavation-induced ground settlement for building
serviceability assessment, Geotechnical and Geoenvironmental Engineering,
ASCE 134 (10) (2008) 1448-1458.
ICSGE-13
GEO-003-9
13th ICSGE
Faculty of Engineering
Cairo - Egypt
ABSTRACT
Despite the moderate seismicity of Egypt, many structures in the Greater Cairo area have been
severely damaged during the 12th October 1992 Dahshur earthquake. This is because most buildings
were not designed to withstand seismic loads before 1992. Following this earthquake, the Egyptian
Government established the Egyptian National Seismic Network (ENSN). This network consists of
the main center at Helwan and five sub-centers at Hurghada, Burg El-Arab, Mersa Alam, Aswan and
Kharga. The ENSN contains 70 recording stations, 25 of which represent strong-motion instruments.
This paper provides overviews on the seismic activity of Egypt from an engineering perspective, and
on the development of the ENSN. The paper summarizes also information on historical and recent
earthquakes hit Egypt. Ninety strong-motion records from 14 earthquakes measured by the ENSN are
analyzed. The paper emphasizes the need for upgrading existing strong-motion instruments to
measure the three acceleration components and for installing new strong-motion instruments in
seismically active regions. Given the accumulation of strong-motion earthquakes recorded by the
ENSN over more than one decade, it is essential to construct strong-motion database for the country.
This will facilitate performing seismological applications (locating hypocenters of future earthquakes,
studying source mechanisms and attenuation analyses) and engineering applications (constructing
design response spectra, providing acceleration inputs for time-history analysis of inelastic structures
and safety assessment of lifeline structures, critical facilities and historical monuments).
KEYWORDS
Egypt seismicity, ENSN, NRIAG, seismic codes, Aswan high dam, historical monuments.
1
INTRODUCTION
Egypt seismicity is known for a long period of time and some of the historical earthquakes hit
the country are documented in the annals of ancient Egyptian history [1-3]. Similarly,
destructive strong-motion earthquakes hit the country in 1981, 1987 and more recently in
1992 [2,3]. Egypt seismicity results from the interaction between the three plates of Eurasia,
African and Arabian plates, as well as the Sinai micro-plate that is partially separated from
the African plate by rifting along the Gulf of Suez.
The structural damage and loss of life caused by the 1992 Dahshur earthquake (35 km of
southwest Cairo and magnitude = 5.6) in the Greater Cairo area have been reported in [4,5].
This earthquake resulted in 561 deaths, about 10,000 injuries and structural damage of about
GEO-004-1
ICSGE-13
$35 million. Ref. [4] presents information on the earthquake characteristics (location, focal
depth, magnitude, and fault mechanism) of the main shock. Field surveys of the damaged and
collapsed structures such as bridges and buildings have been also carried out [4]. Insightful
and detailed statistical analysis of damage caused by this earthquake is contained in [5]. A
sample of 2270 buildings covering a wide geographical distribution, date of construction and
structural system properties has been considered. The specific percentages of damage
reported provide insight as to the seismic performance of different structures in terms of day
of construction, construction material and type of structural system [5]. Following this
earthquake, the Egyptian Government established the Egyptian National Seismic Network
(ENSN) during 1993-1997. This network has been under a continuous upgrading with new
seismic instruments during the last few years. Meanwhile, actual seismic provisions against
earthquake loads have not existed before 1992. Nowadays, recent versions of the Egyptian
standards contain seismic provisions against earthquake and wind loads.
This paper provides overviews on seismic activity of Egypt from an engineering perspective,
the development of the ENSN and on the evolution of national seismic provisions in Egypt.
Ninety strong-motion accelerograms recorded by the ENSN are also analyzed. The paper
sheds the lights on the need for establishing strong-motion database for the country (detailed
information on strong-motion earthquakes and knowledge on soil profile beneath recording
stations). The next section provides an overview on the seismicity of Egypt.
2
SEISMICITY OF EGYPT
Egypt has moderate seismicity and is known for its seismicity for a long time. In fact, some
of the large historical events are documented in the annals of ancient Egyptian history and in
temples [1-3]. Fig. 1 shows the location of historical earthquakes struck Egypt between 2200
BC and 1900 AD. The largest earthquake has a magnitude of 6.7 on the Richters scale.
Egypt is located close to Hellenic arc, one of the continental fracture system at the
convergence boundary of two big lithospheric plates, namely, Eurasia and Africa. The
country is also affected by the opening of the Red Sea (Mid Oceanic System) and its two
branches, the Gulf of Suez and the Gulf of Aqaba-Dead Sea transform system [2,6,7]. Thus,
Egypt seismicity results from the interaction between the three plates of Eurasia, African and
Arabian plates, as well as the Sinai micro-plate which is partially separated from the African
plate by rifting along the Gulf of Suez. In addition to the seismic activity along these plate
margins, mega-shear zones running from southern Turkey to Egypt have been defined in
some studies [8,9]. Historical and instrumental seismicity in the region is mainly associated
with six tectonic trends, namely, the Levant-Aqaba transform system, the northern Red SeaGulf of Suez-Cairo-Alexandria trend (Suez trend), the Eastern Mediterranean-Cairo-Fayoum
trend, Hellenic and Cyprian arcs, Mediterranean coastal dislocation, and the southern
Egyptian trends. Intensive historic reviews on the seismicity of Egypt, Arabia and the Red
Sea region can be found in [2,3]. The total number of historical earthquakes until 1900 is
reported to be 83 [3]. Some of these earthquakes were initiated from seismic sources outside
Egypt. The lack in number of earthquakes in some eras is attributed to the political situations
and disappearance of documents [3]. The period of observations has been grouped into four
intervals, namely, 2200 BC-1300 AD, 1300-1800 (starts with the beginning of Mamluk
period), 1800-1990 (starts with the end of Ottoman Empire), and 1900-2004. The number of
reported events was 34, 36, 17 and 54, respectively (i.e. 1, 6, 17 and 51 per century). Fig. 1
shows the map of local seismicity from 1900-1997 (before the installation of ENSN) [3]. The
number of observed events has remarkably increased after 1997.
GEO-004-2
ICSGE-13
In 7 August 1847 a large earthquake with an estimated magnitude of about 5.5-5.9 shook El
Faiyum-Cairo region. Hundreds of people were killed and injured and thousands of structures
were destroyed. The earthquake was felt across Egypt and much of North Africa with heavy
damage reported as far as Assuit, southeast Egypt.
deposits in the Qarun oil field led to understanding of the structural geology of El Faiyum
depression [8]. In general, the regional seismic activity of Egypt during the last few years
shows a cluster of the seismic activity to the southern part of Hellenic and Cyprean arcs. The
seismic activity extends toward north to the southern part of Greece and Turkey. A Few
events were located along Southern Jordan, Northern Sudan and Central Red Sea.
seismographs were installed in 1951. In 1962 a seismic station was installed at Helwan
Observatory as part of the Worldwide Standardized Seismograph Network (WSSN). In 1972
four analog stations (one Japanese short-period and three intermediate Russian seismographs)
were installed at Helwan, Aswan, Abu-Symbil and Matrouh. Following the 14 November
1981 Naser Lake earthquake, a radio-telemetry network of eight seismographs was installed
around the northern part of Aswan reservoir. The network was extended to eleven stations in
1984 and to thirteen stations in 1985. Six analog strong-motion accelerographs were installed
at Aswan dam (about 700 km south of Cairo). Four short-period vertical components
seismographs were added during 1986-1990 at Kottamia (KOT), Hurghada (HUR), Tell ElAmarna (TAS), and Mersa Alam (MRS) [7].
Table 1: List of recording stations of the ENSN connected to the main-center at Helwan [6]
Station code
HLW
KOT
FYM
HAG
FYD
ZNM
KHB
MYD
NAT
AYT
SAF
GLL
SQR
BNS
ZAF
SUZ
RSH
RDS
GRB
TR1
TR2
EDF
NKL
CAT
BST
DHB
NUB
SFG
QSR
ADB
SHL
AGS
DK2
SWA1
SWA2
Location
Lat.
Long.
29.858 31.343
29.927 31.829
29.692 31.043
29.952 32.093
30.293 32.231
29.375 32.877
29.928 30.972
29.275 30.799
29.633 30.617
29.704 31.153
29.622 31.555
29.577 31.708
29.881 31.196
28.952 31.212
29.282 32.550
29.840 32.830
30.960 33.722
28.711 33.297
28.270 32.786
28.007 33.952
28.385 33.723
25.102 33.181
29.929 33.980
28.775 33.979
29.216 34.732
28.722 34.618
29.027 33.648
26.568 33.929
25.755 34.387
25.351 34.623
23.107 35.399
24.375 34.987
24.320 28.955
29.263 25.709
29.243 25.455
Height
(m)
115
495
204
480
200
41
265
250
570
160
460
515
115
207
155
275
217
260
175
280
275
410
440
410
200
30
145
460
166
206
147
326
410
50
55
Connection mode
Radio to HLWC
Radio to HLWC
Radio to HLWC
Radio to HLWC
Telephone to HLWC
Satellite to HLWC
Radio to HLWC
Radio to HLWC
Radio to HLWC
Radio to HLWC
Radio to HLWC
Radio to HLWC
Radio to HLWC
Telephone to HLWC
Radio to ZNM
Radio to ZNM
Telephone to HLWC
Radio to ZNM
Radio to ZNM
Radio to ZET
Radio to ZET
Telephone to HLWC
Satellite to HLWC
Satellite to HLWC
Satellite to HLWC
Radio to BST
Radio to BST
Satellite to HLWC
Satellite to HLWC
Satellite to HLWC
Satellite to HLWC
Satellite to HLWC
Satellite to HLWC
Satellite to HLWC
Radio to SWA1
Until 1992, many parts of the country were not covered with earthquake recording
instruments. Except of the six instruments installed at Aswan high dam, there was no strongmotion instruments in the Greater Cairo area. Following the 12th October 1992 earthquake in
Dahshur area, the Egyptian Government was prompted to construct the Egyptian National
Seismic Network (ENSN). The network was completed in 1997 and included 63 instruments
GEO-004-5
ICSGE-13
covering the Egyptian territory. These instruments contain 47 seismic stations based on
satellite communication, 11 stations based on telemetry communication and 5 stations based
on telephone lines. The network records local and regional earthquakes as well as teleseismic events. The data communication system was also upgraded from telephone lines to
satellite to increase the efficiency of the network. By the end of 2002, the installation of the
whole seismic field stations was completed covering the main center and five sub-centers
were also constructed and equipped. The ENSN consists of the main center at Helwan and
five sub-centers at Hurghada, Burg El-Arab, Mersa Alam, Aswan and Kharga. An
Earthquake Disaster Reduction Data Center (EDRDC) was also established and supported by
GIS technology. In 2003, the number of recording stations was upgraded to 70 stations. Table
1 provides information on recording stations connected to the main center. This center
receives seismic data from near distance stations through telemetry communication and from
remote stations and sub-centers via telephone lines and satellite communications.
Although, the ENSN contains 70 recording stations that are geographically distributed to
effectively cover seismically active regions in Egypt, the network contains 25 strong-motion
instruments only. From an engineering point of view this number is small. Furthermore, some
of these instruments measure one or two acceleration components only (see Table 1). Given
the safety requirements of lifeline structures, critical facilities (e.g., Aswan dam, existing and
new proposed nuclear power plants and petroleum and natural gas installations) and historical
monuments, there is an urgent need for upgrading existing instruments and for installing new
three components instruments near new identified seismic regions. Specifically, the number
of strong-motion instruments in the Greater Cairo area needs to be increased.
4
Until the beginning of 1990, real consideration of seismic loadings were absent from national
codes of Egypt. Buildings were typically designed to resist gravity loads and lateral load
resistance was provided through consideration of wind loads, mainly, for tall buildings. In
1988 the Egyptian Society for Earthquake Engineering proposed a model code entitled
Regulations for earthquake-resistant design of buildings in Egypt [13]. This code draft was
based on the Uniform Building Code (UBC) of America. However, these guidelines were not
put into practice. Earthquake loads were specified as a percentage of the total weight of the
building. High-rise buildings of 30-40 floors were typically designed based on seismic design
regulations of the United States or England.
The reinforced concrete code published in 1989 by the Ministry of Housing, Utilities and
New Communities was the only exception that considered seismic loading. This code,
however, ignored basic seismic considerations, such as, influence of soil type and dynamic
characteristics of buildings. Seismic effects were represented by lateral forces as a percentage
of the building weight. Two design procedures, the allowable (working) stress design and the
ultimate strength design methods were specified. The allowable stress design method was
more widely adopted by most structural engineers. The loading code issued in 1994 following
the 1992 earthquake, defined the seismic loads based on the structure type. Although this
code adopted significantly simplified assumptions in terms of loading considerations and
design procedures it represented an improvement to previous regulations. Ref. [12] proposed
a revised draft for the Egyptian code for seismic loads on structures. In 2004 a new loading
code [14] was issued, which dealt with most of the shortcomings of preceding standards. The
2004 code introduced the concept of response spectra and force reduction factors for the
design of structures including safety verifications relevant to ultimate and serviceability limit
GEO-004-6
ICSGE-13
states. The ECCS code has also gradually introduced ductility concepts and detailing
procedures through its 2001 and 2004 versions [14]. With the recent 2007 203-Egyptian
standards, development of national seismic codes is converging to international standards.
However, the majority of current buildings were not designed to seismic regulations. Hence,
with exception of structures constructed in the last 10-15 years, a high level of seismic
vulnerability exists [15].
Table 2: Information on earthquakes recorded by the ENSN [6]
Earthquake date
31 August 1998
4 September 1999
9 September 1999
13 September 1999
8 March 2000
25 December 2000
5 January 2001
M
NA
NA
NA
NA
5.7
5.2
4.3
D
(km)
NA
NA
NA
NA
7.0
11.0
25.6
Component
N
E
V
N
E
V
N
E
V
N
E
N
E
V
N
E
N
E
V
N
E
V
N
E
V
N
E
V
N
E
V
N
E
N
E
V
N
E
N
E
N
E
N
E
N
E
PGA
(g)
0.05
0.06
0.92
0.002
0.01
0.02
0.07
0.08
0.96
0.33
0.09
0.01
0.01
0.02
0.35
0.09
0.91
0.78
1.00
0.02
0.03
0.02
0.14
0.14
0.36
0.02
0.02
0.02
0.04
0.09
0.40
0.20
0.26
0.15
0.09
0.16
0.80
1.00
0.11
0.28
0.10
0.14
0.16
0.03
0.08
0.11
td
(s)
41
Station No.
Station site
61624307
BYD (school)
45
65624302
Observatory
52
61624703
BYD (school)
40
62424700
48
65625201
Observatory
41
62425200
62
61625600
BYD (school)
67
65625604
Observatory
57
62506700
Nuweiba
43
62306700
NRIAG
58
65906700
Dahab
72
66806700
El-Tur
46
69035900
Sharm Elsheikh
43
62735900
Taba
82
65935900
Dahab
45
66835900
El-Tur
43
65900501
Dahab
41
66800501
El-Tur
GEO-004-7
ICSGE-13
11 November 2001
17 December 2001
22 January 2002
18 February 2002
9 April 2002
20 August 2002
M
6.2
3.9
4.5
6.6
3.8
4.0
NA
D
(km)
30.0
23.0
6.0
25.0
13.4
13.6
NA
Component
N
E
N
E
V
N
E
V
N
E
V
N
E
N
E
N
E
N
E
V
N
E
N
E
N
E
N
E
V
V
N
E
N
E
V
N
E
N
E
V
N
E
N
E
PGA
(g)
0.23
0.25
0.14
0.64
1.00
0.01
0.06
0.21
0.01
0.06
0.21
0.10
0.46
0.16
0.20
0.27
0.09
0.25
0.11
0.06
0.08
0.27
0.41
0.13
0.15
0.52
0.07
0.11
0.23
0.30
0.23
1.00
0.30
0.23
0.17
0.11
0.20
0.24
0.17
0.17
0.14
0.20
0.15
0.16
td
(s)
45
Station No.
Station site
60517400
Ariesh quarry
28
65817400
Burg El-arab
center
27
67117400
Matrouh
26
67117401
Matrouh
25
65931500
Dahab
29
62531500
El-Tur
29
65735100
Tal El-amarna
23
60535100
Ariesh quarry
28
67102200
Matrouh
24
65502200
Ariesh655
47
65802200
33
67102201
Burg Elarab
center
Matrouh
36
69004900
Sharm Elsheokh
28
65904900
El-Tur
35
69009900
Sharm Elsheikh
26
65909904
El-Tur
45
69023201
Sharm Elsheikh
31
62323203
Hurghada
M = Richters magnitude, NA = not available, D = Focal depth, td= duration, PGA = peak ground acceleration.
0.1
Acceleration (g)
0.05
-0.05
-0.1
0
10
20
30
40
x 10
2.5
2
1.5
1
0.5
0
0
50
10
20
Time (s)
30
40
50
Frequency (Hz)
Fig. 3: Time history and Fourier spectra of EW acceleration component (9.9.1999 at Observatory)
2.5
2.5
2
EW
1.5
1
0.5
0
0
10
20
30
NS
1.5
1
0.5
40
Frequency (rad/s)
0
0
10
20
30
40
Frequency (rad/s)
CONCLUSIONS
Egypt has moderate seismicity and is known for its seismicity for a long period of time.
Large historical events are documented in annals of ancient Egyptian history and
temples. The installation of the Egyptian National Seismic Network (ENSN) has been
initiated in 1993 following the 1992 Dahshur earthquake and has been completed by the
end of 1997. This network consisted initially of 63 recording instruments and was
upgraded to 70 instruments by 2003. Nevertheless, the network contains 25 strongmotion seismographs. Moreover, some of these instruments measure single- or twoacceleration components only. Given that this network covers about 1,000,000 km2, and
given the recent seismic activity in the country and safety requirements of lifeline
structures, critical facilities (e.g. petroleum and gas installations, electrical and water
networks, nuclear plants and Aswan dam) and historical monuments, there is an urgent
GEO-004-9
ICSGE-13
REFERENCES
[1]
[2]
Ambraseys, N.N., Melville, C.P. and Adams, R.D. The Seismicity of Egypt,
Arabia and the Red Sea Cambridge University press (1994), UK.
[3]
[4]
[5]
[6]
[7]
[8]
[9]
Moharram, A.M., Elghazouli, A.Y. and Bommer, J.J. Soil Dynamics and
Earthquake Engineering 28 (2008) 795-811.
[10]
Haggag, H.M., Gaber, H.H., Sayed, A.D. and Ezzat, M.E. Acta Geodaetica
Geomaterial 5 1 (2008) 19-29.
[11]
[12]
[13]
[14]
ECS-203. Research center for housing, building and urban planning, 8th edition
(2004).
[15]
ECP-203. Research center for housing, building, and urban planning, 3rd edition
(2007).
ICSGE-13
GEO-004-10
13th ICSGE
Faculty of Engineering
Cairo - Egypt
ABSTRACT
The New Austrian Tunnelling Method (NATM) is a conventional tunnelling technique
in which the tunnel section is divided into suitable parts in area and shape. Each part is
sequentially excavated and the exposed ground by excavation is temporarily supported
by a shotcrete lining thus allow a controlled ground deformation, then the permanent
lining is carried out after the completion of the whole tunnel section and reaching
equilibrium. Therefore, NATM as a construction technique requires less mechanical
equipments and low installation costs which are significant factors in underground
projects that require major financial investments. In this paper, 2-D models are
employed to study the effect of construction sequences due to NATM tunnelling on
ground deformation. Then, a 3-D model for the optimal sequence of excavation is used
to investigate the ground deformation behaviour for different excavation lengths.
Furthermore, the 3-D model is used in the calibration of the 2-D model in order to
narrow the gap between 2-D and 3-D results. A main concern is devoted to shallow
tunnelling in soft ground to meet most of the conditions in urban areas. The results
constitute a step forward to automate the safe and economic design of NATM cross
sections in urban areas.
KEYWORDS
NATM, Tunnelling, Soft Ground, Finite Element, Modelling.
ICSGE-13
GEO-005-1
INTRODUCTION
For safe and economic application of NATM, it is important to determine carefully the
design aspects of NATM such as ground support, face stability, excavation sequence,
section profile and monitoring system. Also, engineers should understand the concept of
NATM to avoid misapplication of the method and solve any unexpected problems.
Although NATM was primarily developed for rock soil, it is now one of the appropriate
methods in soft ground conditions. However in cases such as flowing ground or ground
with short stand-up time, using NATM requires solving these unstable conditions by
dewatering, grouting or any other mean of ground improvement [1] [2].
In this paper, different construction sequences of NATM for two metro lines cross
sections are presented and analysed in both two and three dimensional finite element
models. The results of ground movements in terms of surface and crown settlements for
each excavation sequences were assessed to determine the appropriate sequence of
excavation where the ground deformation are kept to a minimum to fulfil the safety
requirements for the construction. The 2-D plain strain model of the selected excavation
sequence is then more investigated by sophisticated 3-D models with different lengths
of excavation. The Finite Element Modelling techniques using FINAL package [3] are
employed to analyse the different scenarios and construction sequences.
2
Finite element models have been used in order to understand the interaction between the
surrounding soil and the tunnel supporting system for NATM tunnelling and to predict
the effect of excavation sequence on ground deformation. The assumed tunnel cross
section is suitable for double track metro lines and satisfying the electrical and
mechanical requirements of the trolleys. Based on both geometrical and stress
optimization techniques, an optimal profile for the tunnel is considered [4].
2.1 Geotechnical parameters
The tunnel is assumed to be constructed in a clayey layer at a depth of 14.0m with no
existence of ground water and no specific strengthening technique is applied. In this
paper, the mechanical parameters used in the analysis are indicated in Fig. 1. These
geotechnical conditions can be found in one of the cross sections of Cairo Metro line 3
phase 2 [5].
2.2 Finite element modelling
For the modelling process in FINAL, The soil media is modelled using six node linearly
varying Strain Triangular finite elements (L.S.T) and the shotcrete lining by six node
curved boundary beam elements (Beam6). The mesh adopted for 2-D analysis is shown
in Fig. 1. The movements at the bottom of the model are prevented while only the
horizontal movements are prevented at the side faces of the model.
2.3 Construction sequence simulation
For simulating the excavation sequences and shotcrete installation, the 2-D plane strain
finite element model with different loading cases is employed. The initial loading case
is the geostatic state of stresses and the existing surface loads. Then each part of
excavation is simulated in two steps using stiffness reduction method [6]. In the first
ICSGE-13
GEO-005-2
step, the stiffness of the soil in the excavated part is reduced by a factor accompanied by
stress redistribution at the tunnelling periphery. In the second step, the excavated part is
removed and the temporary lining is installed in its fresh state accompanied with a stress
elimination of the excavated soil. In the 2-D analysis of the presented sequences of
excavation, the stiffness reduction factor is assumed to be the same in all sequences.
Then, the 3-D analysis will be used in estimating more accurate values for the reduction
factor needed in the 2-D analysis.
One of the critical aspects in NATM tunnel design is the excavation sequences. The
cross section of NATM is divided into suitable parts in area and shape in order to allow
efficient and practical construction. Then, these parts are sequentially excavated. The
subdivisions depend on ground condition, soil stand up time, tunnel geometry,
allowable settlement and the space requirements for typical equipments used in NATM
construction. Excavation is usually performed by blasting, hand mining or any other
mechanical means [7].
In this paper, top heading and sidewall drift for partial face excavations, Fig. 2, are
presented and studied. In top heading excavation, Section (A), the crown is excavated
and immediately supported with temporary shotcrete lining then subsequently the invert
is excavated and ring closure is achieved. In single sidewall drift, Section (B), the top
side part followed by the lower side part is excavated and supported then subsequently
the full size excavation are executed in the same manner. While in double side drift,
Section (C), the side galleries are excavated and supported first, and then the central part
is subsequently excavated. The ring closure at the invert comes at the end. The
ICSGE-13
GEO-005-3
percentages of excavated area for each part to the whole excavated section, 65.80 m2,
are presented in Table 1.
P
31.19%
20.32%
28.19%
20.30%
---
Section (C)
20.52%
16.83%
20.52%
16.83%
18.32%
6.98%
0.00
2.00
Section (A)
4.00
Section (B)
Section (B')
6.00
Section (C)
8.00
10.00
(1)
(2)
(3)
(4)
(5)
(6)
Excavated Parts
ICSGE-13
GEO-005-4
The maximum values of crown settlement in section (A), (B), (B\) and (C) are 8.46,
7.55, 5.12 and 5.94 cm respectively and 6.36, 5.77, 4.36, 4.05 cm respectively for
surface settlement.
P
The 3-D analysis is needed for studying the stability at the tunnel face and investigating
the effect of excavation steps on ground deformation due to the inability of the 2-D
analysis to model the tunnel along its longitudinal axis. Furthermore, the 3-D model is
used to calibrate the stiffness reduction factor used in the 2-D model and to narrow the
gap between 2-D and 3-D results. The analysis is carried out with the same mechanical
properties for soil and shotcrete lining. The modelled sequence of excavation is the
single sidewall drift as mentioned previously in section (B). Parallel step wise analysis
was considered for each excavated part. In this paper, a serious of systematic 3-D
models with seven different excavation lengths, ranging from 4.0m to 25.0m, were
analysed
4.1 Finite element modelling
The soil is modelled by cubic Iso-Parametric Quadratic Solid (IPQS) elements with
twenty-nodes. At each node, there are three degrees of freedom describing the
movements in X, Y and Z directions. For modelling the shotcrete lining, eight-node
shell elements are used with six degrees of freedom at each node describing three
movements and three rotations. One of the meshes adopted for 3-D analysis is shown in
Fig. 4. The movements in all directions at the bottom of the mesh and the horizontal
movements at the vertical faces of the model are prevented.
ICSGE-13
GEO-005-5
Special attention was paid to the crown deformation in order to calibrate the 2-D model
by determining an appropriate value for stiffness reduction factor (). In 2-D analysis,
each part of excavation is simulated in two steps. In the first step, stress is redistributed
at the tunnel periphery based on the factor of stress reduction. In the second step, the
removing of the excavated part is accompanied by redistribution for the rest of the
stresses. In order to study this influence, the variation of stiffness reduction for each part
of excavation is studied individually while assuming an average value for the reduction
in other parts equals to 50%. Crown settlements due to different values of for parts
(1), (2), (3) and (4) are compared in Figs. 6-a, 6-b, 6-c and 6-d respectively.
ICSGE-13
GEO-005-6
Reviewing the results in Fig.6, direct correlation is found between ground deformations
and for all parts except part (3). For the two steps of excavation simulation of each
part, increasing the reduction factor causes settlement increment in the first step and
decrement in the second step. In part (3), the induced crown settlement in the first step
is relatively lower than the second step due to the existence of inner temporary shotcrete
lining of previous parts, Fig. 7.
ICSGE-13
GEO-005-7
Based on the previous study and for simplicity, two values for will be used in
calibrating the 2-D model; 1-2 for part (1) and (2), 3-4 for part (3) and (4).
B
Figs. 8 and 9 illustrate the reduction factor 1-2 and 3-4 for different excavation lengths
to reach the maximum condition in the 3-D analysis (U0). Also, reduction factor has
been predicted to obtain 2-D ground deformation higher than that of the 3-D results by
30% (U30) as a factor of safety for 2-D analyses results.
B
The relation between and the excavation length can be expressed as follow:
1 2 = A * L + B
(1)
3 4 = C * L + D
(2)
Where; 12
ICSGE-13
GEO-005-8
U0
B
U30
B
2.3656
5.0407
-0.2016
97.339
1.5249
55.177
-0.5468
92.611
As an example, these results are used to predict the value of 1-2 & 3-4 for an
excavation length of 8.0m. The induced crown settlement profile and surface settlement
profile from 3-D analysis and 2-D analyses are plotted in Figs. 10 and 11.
B
0.00
3D result
1.00
2.00
2D result
(U0)
3.00
4.00
2D result
(U30)
5.00
6.00
7.00
-40
-30
-20
-10
10
20
30
40
0.00
3D result
1.00
2.00
2D result
(U0)
3.00
4.00
2D result
(U30)
5.00
6.00
7.00
-40
-30
-20
-10
10
20
30
40
ICSGE-13
GEO-005-9
CONCLUSIONS
The analysis was carried out to study the effect of construction sequence on ground
deformation and to obtain the optimum sequence. In soft soil with difficult ground
condition, more subdivisions for the tunnel cross section minimize the ground
deformation. In this case, sidewall drift is considered the appropriate construction
sequence. It is also recommended that the inner temporary lining should be removed
later with the ring closure of the invert.
The effect of excavation length on ground deformation was investigated and the
optimum excavation length can be determined. Furthermore, more stable deformational
behaviour can be reached when applying a suitable strengthening technique.
Sophisticated 3-D analysis is a powerful tool to predict ground deformation and to study
face stability. However, it is not widely used because its complexity. Therefore, easier
2-D models can be used provided that well calibrated values of should be estimated by
using 3-D modelling in the early design stage.
7
REFERENCES
[1]
[2]
[3]
[4]
Abd Elrehim, M.Z. & Marwan, A.M. On the geometrical optimization of NATM
in urban areas. Fourth International Young Geotechnical Engineers Conference.
Alex., Egypt (2009).
[5]
Abd Elrehim, M.Z. & Abu-Krisha, A.A. Controlling ground movements due to
NATM tunnelling in urban areas, ECCOMAS Thematic conference on
computational methods in tunnelling. (2007).
[6]
[7]
Karaku, M. & Fowell, R.J. An insight into the New Austrian Tunnelling
Method (NATM), Blgesel Kaya Mekanii Sempozyumu / ROCKMEC2004VIIth Regional Rock Mechanics Symposium (2004).
ICSGE-13
GEO-005-10
13th ICSGE
Faculty of Engineering
Cairo - Egypt
ICSGE-13
GEO-006-1
INTRODUCTION
ICSGE-13
GEO-006-2
Hung [24]. Plastic diaphragm wall seems also to be an efficient solution (Comina and
Foti, [25]).
Wave barriers can be divided into two groups, namely, active and passive isolation
systems, Woods [26]. Barriers placed around the vibratory source are active isolation
systems whereas barriers located farther from the source and close to a site where the
vibratory amplitude must be reduced are defined as passive isolation. Numerical
methods have also been developed to effectively examine vibration isolation problems.
Lysmer and Waas [27] employed the lumped mass method to assess the isolation
efficiency of open and bentonite-slurry-filled trenches in a layered soil. Segol et al. [9]
applied finite element methods (FEM) along with special non-reflecting boundaries to
the same problem. Fuykui and Matsumoto [16] used the finite difference technique to
study the scattering of a Rayleigh wavelet by a rectangular open trench.
In this study, finite element method is used to simulate two-dimensional axisymmetric
cases of vibration isolation by trenches due pile driving by impact and vibratory
machines. The study consists of two parts; the first part represents the verification of the
finite element model by comparing its results with full scale model results. In the second
part, a parametric study is performed to evaluate the efficiency of using trenches in
screening the vibrations resulting from driving piles in different types of soil.
2
In nonlinear dynamic analysis the solution of the finite element equations is usually
obtained by direct integration procedures. The major technique can be employed is the
implicit integration, using the Newmark method or Wilson theta method. A very
common technique used is the trapezoidal rule, which is Newmark's method with = 0.5
and = 0.25. Using implicit time integration, we consider the equilibrium of the system
at time t+t. This requires in nonlinear analysis that iteration be performed. Using the
modified Newton-Raphson iteration, the incremental finite element equilibrium
equations used are
U (k) =
t + t
t +t
R-
U (k 1 ) + U (k)
t +t
F (k-1)
(1)
(2)
t + t
point forces
t + t
U (k 1 ) .
Using the trapezoidal rule of time integration, the following assumptions are employed:
t + t
t + t
ICSGE-13
U = U+
t
U = U+
t t
( U+
2
t t
( U+
2
t + t
t + t
U)
(3)
U)
(4)
GEO-006-3
(k )
4 t + t ( k 1) t
4 t t
(k )
(
)
U
U
U
U U
t 2
t
(5)
K U(k) =
t +t
R-
t +t
F (k-1) M 4 2 ( t + t U ( k 1) t U ) 4 t U t U
(6)
The energy and force/moment convergence criteria used in nonlinear dynamic analysis
are:
Energy convergence criterion: For all degrees of freedom
T
t + t
t + t
U ( i ) R - M t + t U (i 1 ) C t + t U (i 1 )
F (i-1 )
ETOL
T
t
t
t
(1)
(0)
(0)
t + t
t + t
U
R -M
U C
U F
(7)
Force and moment convergence criteria: For the translational degrees of freedom
t + t
R -M
t + t
(i 1 )
C t + t U (i 1 )
t + t
F (i-1 )
2
RTOL
(8)
RTOL
(9)
RNORM
For the rotational degrees of freedom
t + t
R -M
t + t
(i 1 )
C t + t U (i 1 )
t + t
F (i-1 )
RMNORM
ICSGE-13
GEO-006-4
Cu = 50.0 kN/m2
3
E = 3.50x10 kN/m
(-4.00)
= 0.5
= 0.0
= 20.0 kN/m3
2
E = 1.0x10 kN/m
= 0.5
= 0.0
(-10.00)
C = 0.0
= 19.7 kN/m
5
E = 1.0x10 kN/m
= 0.3
= 43.0
Fig. 1: Soil profile at the site and average soil properties, (after El-Shaikh, 2005)
River
River
Driven cast-in-place concrete piles were used in that site, were closed-end steel masts of
diameter 0.60 meter and thickness of 0.019 meter were driven through soil to a depth of
15.00 meters from ground surface. The levels of vibrations were measured during pile
driving at seventeen locations at distances ranges between 0.00 and 140.00 meters from
the driving machine. Figure 2 demonstrates the position of the driving machine at the
site and the distances between it and the measurement points. The measurement of
vibrations in terms of displacement (m), velocity (mm/sec) and acceleration (mm/sec2)
were reordered using two test facilities positioned on the ground surface.
0.0 m
5.0 m
10.0 m
15.0 m
20.0 m
25.0 m
30.0 m
35.0 m
40.0 m
45.0 m
50.0 m
60.0 m
70.0 m
Traffic Road
Driving Machine
River
Traffic Road
River
Water
Breaker
Traffic Road
100.0 m
120.0 m
140.0 m
Traffic Road
ICSGE-13
GEO-006-5
In order to verify the finite element model used in this study, two-dimensional
axisymmetric solid elements are used to model the soil layers. Nine-node elements are
used in the y-z plane with dimensions 0.5 x 0.5 m. The Mohr-Coulomb model is used to
simulate the soil material formulation. The soil continuum is taken 120.00 meters in
length and 30.00 meters in depth, the soil is divided into three surfaces as presented in
Fig. 3, to simulate the soil layers at the site. The soil average properties were taken
according to the site investigation for the proposed area as shown in Fig. 1.
Concentrated Impact Load
Steel Pile
1.00 m Beam Element
Soil Continuum
0.50 x 0.50 m Solid Element
Ground Surface
Medium Stiff to
Stiff Silty Clay
Spring Element
Very Stiff
Silty Clay
Dense to Very
Dense Sand
ICSGE-13
GEO-006-6
The load is modelled as impact concentrated load with a speed of 25.00 blows/minute,
the time function of the load is demonstrated in Fig. 4, using 0.01 second time step.
ICSGE-13
GEO-006-7
ICSGE-13
GEO-006-8
ICSGE-13
GEO-006-9
Fig. 11: Relation between measured vs. calculated acceleration in vertical direction with
distance. The pile is driven 5.00 m from the ground level.
Fig. 12: Relation between measured vs. calculated acceleration in horizontal direction
with distance. The pile is driven 5.00 m from the ground level.
Fig 13: Relation between measured vs. calculated velocity in vertical direction with
distance. The pile is driven 5.00 m from the ground level.
ICSGE-13
GEO-006-10
Fig. 14: Relation between measured vs. calculated velocity in horizontal direction with
distance. The pile is driven 5.00 m from the ground level.
Fig. 15: Relation between measured vs. calculated displacement in vertical direction
with distance. The pile is driven 5.00 m from the ground level.
Fig. 16: Relation between measured vs. calculated displacement in horizontal direction
with distance. The pile is driven 5.00 m from the ground level.
ICSGE-13
GEO-006-11
A parametric study using finite element model is performed to predict the level of
vibrations produced due pile driving in sand before isolation and after isolation using
trenches and then calculating the percentage of reduction in the magnitude of
displacement whether in vertical or in horizontal directions. Also, the magnitude of
displacements in both directions is compared with the level of caution and danger on the
structures according to the hazard chart shown in Fig. 17. It is a plot developed by
Richart [29] for vibrations that gives various limits of frequency and amplitude for
different purposes.
ICSGE-13
Modulus of
Poissons
Elasticity
Ratio
(E)
()
(kN/m2)
Friction
Angle
()
1.50 x 104
0.30
30.00
5.00 x 104
0.30
34.00
GEO-006-12
b) The frequency of loading (F); which is taken according to vibratory pile driving
machine with frequencies 50.00 and 100.00 Hz.
c) The length of pile (L); which is taken 10.00 and 15.00 meters.
d) The penetration ratio of the pile in the soil (PR); three levels are taken into
consideration. First, when the pile is driven on the top surface (PR = 0.00%). Second,
when the pile is driven to 50.00% of its length. And finally, when it reaches a level of
100.00% of its length.
e) The distance of the trench from the pile (D); this is varied from 5.00 to 30.00 meters.
f) The depth of the trench from the ground surface (H); which is varied from 1.00 to
20.00 meters.
g) In-filling of the trench; in the first case the trench is open, the second case it is braced
using timber lagging and finally, the trench is in-filled with bentonite slurry.
6 OUTLINE OF THE ANALYSIS RESULTS
Results of the analytical study are presented in more than one chart form; where the
results are recorded at points 1.00, 5.00, 10.00, and 20.00 meters away from the trench
and at a depth of 1.50 meters from ground level which represents the foundation depth
of the nearby structures. These results are in the form of vertical and horizontal
displacement, velocity and acceleration. The curves represent the relationship between
the magnitude of vertical and horizontal displacement before trench and after the trench
is dug with the distance from the trench. Also the relationship between the percentage of
reduction in the magnitude of both vertical and horizontal displacements with the
distance from the trench, considering different parameters such as the frequency of the
load, the pile length, the distance of the trench from the pile and the depth of the trench
from the ground.
7 ANALYSIS PROCEDURE AND SOIL MODEL
The following section represents a part of the study which discusses the effect of filling
the trench with bentonite slurry on the efficiency of the trench as a vibration barrier. The
soil continuum is taken 60.00 meters in length and 30.00 meters in depth. Twodimensional axisymmetric solid elements are used to model the soil layer, 9-nodes
elements are used in the y-z plane with dimensions 1.00 x 1.00 meters.
The same elements are used to model bentonite slurry trench, but with different
dimensions. The dimensions of bentonite slurry elements are 0.50 x 1.00 meters, as
illustrated in Figs. 18- a and b. The material of the elements is taken as Mohr-Coulomb
formulation with the following properties; modulus of elasticity is 300.00 kN/m2,
Poissons ratio is 0.35, unit weight is 10.50 kN/m3, cohesion is 0.00 and friction angle is
28.00o.
ICSGE-13
GEO-006-13
Spring Element
Ground Surface
Soil Continuum
1.00 x 1.00 m Solid Element
Spring Element
Spring Element
Fig. 18-a: Model discretization of the soil continuum with bentonite slurry trench.
Hammering Load
Steel Pile
Observation Points
L
0.50 m
Ground Surface
1.00 m
5.00 m
H
10.00 m
20.00 m
ICSGE-13
GEO-006-14
ICSGE-13
GEO-006-15
Fig. 20: Relation between the distance from trench and the vertical displacement
before and after the trench is dug. PR = 0.00 % and D = 15.00 m.
Fig. 21: Relation between the distance from trench and the reduction in vertical
displacement after the trench is dug. PR = 0.00 % and D = 15.00 m.
Fig. 22: Relation between the distance from trench and the horizontal displacement
before and after the trench is dug. PR = 0.00 % and D = 15.00 m.
ICSGE-13
GEO-006-16
Fig. 23: Relation between the distance from trench and the reduction in horizontal
displacement after the trench is dug. PR = 0.00 % and D = 15.00 m.
A group of analytical models were prepared to simulate the advancing of the pile
through the soil. The first case represent the stage when the pile reaches 5.00 meters
from ground level, which means a penetration ratio PR = 50.00%. Various types of
analyses as previously proposed are performed with the same parameters which includes
different distances; D = 5.00, 10.00, 15.00, 20.00 and 30.00 meters. Also, the depths of
the trench used here were H = 3.00, 5.00, 6.00, 8.00 and 10.00 meters. A sample of the
results is presented in Figs. 24 and 25.
The last stage of pile driving, where the pile reaches 10.00 meters below ground surface
or 100.00% of its length is finally investigated, which includes different distances; D =
5.00, 10.00, 15.00, 20.00 and 30.00 meters. Also, the depths of the trench used in the
analyses were H = 3.00, 5.00, 6.00, 8.00, 10.00, 15.00 and 20.00 meters.
Fig. 24: Relation between the distance from trench and the vertical displacement before
and after the trench is dug. PR = 50.00 % and D = 5.00 m.
ICSGE-13
GEO-006-17
Fig. 25: Relation between the distance from trench and the vertical displacement before
and after the trench is dug. PR = 50.00 % and D = 30.00 m.
Another group of figures are constructed to show the relation between the distance of the
trench from the pile (D) versus the safe distance - points of displacement lower than the
level of caution to structures once from the trench and the other from the hammering
machine as illustrated in Figs. 26 and 27 respectively. The studied heights are H = 1.00,
2.00, 3.00 and 4.00 meters from the ground level. It is clear from the figures that the safe
distance from the trench is decreased as (D) is increased for the same depth of trench.
Also, increasing the depth of trench decreases the safe distance from the trench. Figure
27 shows that whatever the distance of the trench from the pile is; the safe distance from
the hammering machine is almost the same for the same depth.
Fig. 26: Relation between the distance of trench from the pile (D) and the safe distance
from trench, for vertical displacement. F = 50.00 Hz, PR = 0.00 % and L = 10.00 m.
ICSGE-13
GEO-006-18
Fig. 27: Relation between the distance of trench from the pile (D) and safe distance from
pile hammer, for vertical displacement. F = 50.0 Hz, PR = 0.00% and L = 10.0 m.
9
The effect of changing the relative density of sand is examined and the sand is modelled
as medium dense sand, using the properties of sand (S2) as indicated in Table 1. The
trench is in-filled with bentonite slurry. Also, the three penetration ratios are taken into
consideration. Figs. 28, 29, 30 and 31 illustrate the relationships between the distances
from trench versus the magnitude of vertical displacement before and after the trench is
dug, the reduction in vertical displacement after the trench is dug, the magnitude of
horizontal displacement before and after the trench is dug and the reduction in horizontal
displacement after the trench is dug, respectively. The distance of the trench is D =
30.00 meters and the depths of the trench are H = 5.00, 10.00, 15.00 and 20.00 meters. F
= 50.00 Hz and L = 10.00 meters and PR = 100.00%. Also, D = 5.00, 10.00, 15.00 and
20.00 meters are investigated for the same penetration ratio and for PR = 0.00% and PR
= 50.00%.
Fig. 28: Relation between the distance from trench and the vertical displacement before
and after the trench is dug. PR = 100.00 % and D = 30.00 m.
ICSGE-13
GEO-006-19
Fig. 29: Relation between the distance from trench and the reduction in vertical
displacement after the trench is dug. PR = 100.00 % and D = 30.00 m.
Fig. 30: Relation between the distance from trench and the horizontal displacement
before and after the trench is dug. PR = 100.00 % and D = 30.00 m.
Fig. 31: Relation between the distance from trench and the reduction in horizontal
displacement after the trench is dug. PR = 100.00 % and D = 30.00 m.
ICSGE-13
GEO-006-20
By comparing these figures with the corresponding ones in the very loose sand; it is
clear that the level of vibrations are already very high in the medium dense sand case
compared with the very loose sand ones, even before the trench is dug; this refers to
high refusal of the medium sand to the pile driving which generates high level of
vibrations. But, the figures shows that the trench is more effective in the medium dense
sand as the percentage of reduction in the magnitude of both vertical and horizontal
vibrations is greater than that in the very loose ones, even though the vibrations are still
very high and dangerous to the structures.
10 EFFECT OF PENETRATION RATIO
The effect of penetration ratio of pile (PR) is examined for D = 30.00 meters in very
loose sand, where the needed trench depth (H) to be safe under the level of caution for
vertical displacement is plotted versus the (PR) as indicated in Fig. 32. The figure shows
clearly that; as the penetration length of pile increases the depth of the trench must be
increased too where there is a need to lengthen the trench to increase the path of
vibrations and decreasing its effect on the surrounding buildings.
Fig. 32: Relation between pile penetration ratio (PR) and the needed trench depth. For
vertical displacement in very loose sand. F = 50.00 Hz, L = 10.00 m and D = 30.00 m.
11 CONCLUSIONS
In this study, finite element model is employed to evaluate the efficiency of using
trenches in screening the vibrations resulting from pile driving in different types of sand.
different types of trenches whether open, braced with timber lagging or in-filled with
bentonite, the load frequency, the pile length , the penetration ratio of the pile, the
distance of the trench from the source of vibration and the depth of the trench from the
ground level, are the main parameters in the investigation. Results of the analytical
program are presented in the form of figures and tables, and have given the possibility to
draw out some conclusions, which are summarized as;
1. Using bentonite slurry trenches is very useful in isolating vibrations produced due to
pile driving in sand deposits.
2. The reduction in the magnitude of horizontal displacement is very much greater than
that of vertical displacement.
ICSGE-13
GEO-006-21
3. Increasing the distance of the trench from the source of vibrations increases the
reduction in the magnitude of both vertical and horizontal displacements.
4. Considering the vertical displacement of very loose sand; a trench of at least D =
20.00 meters and H = 20.00 meters - or 200.00% of the pile length - is required to
reduce its hazards on the neighbouring structures, also a trench of D = 30.00 meters
and H = 15.00 meters or more may be used in this case.
5. The safe distance from the trench is decreased as (D) is increased for the same depth
of trench. Also, increasing the depth of trench decreases the safe distance from the
trench.
6. Whatever the distance of the trench from the pile is; the safe distance from the
hammering machine is almost the same for the same depth.
7. As the penetration length of pile increases; the length of the trench must be increased
too in order to screen the vibrations from the hammering machine.
8. The percentage of reduction in the magnitude of displacement must not be the main
criteria to choose either the distance or the depth of the trench as it can be very high
in some cases but the level of vibration is still very harmful to the nearby structures.
9. The level of vibrations is already very high in the medium dense sand case compared
with the very loose sand one, even before the trench is dug.
10. Considering the vertical displacement of medium dense sand; a trench with at least
D = 30.00 meters and H = 20.00 meters or 200.00% of the pile length is needed to
reduce the vibrations under the level of caution to the structures.
11. The trench is more effective in the medium dense sand as the percentage of
reduction in the magnitude of both vertical and horizontal vibrations is greater than
that in the very loose ones, even though the vibrations are still very high and
dangerous to the structures.
From the research work of this study, it can be suggested that further detailed
investigations have to be conducted, in order to investigate the following topics:
1. Study the efficiency of using trenches in screening vibrations resulting from pile
driving in different soil formation like clay or multi layer soil conditions.
2. Perform three dimensional finite element analyses, to verify the two dimensional
axisymmetric results.
3. Conducting more sophisticated analytical models, to investigate the effect of
different materials filling the trench on the efficiency of the trench.
12 REFERENCES
[1]
[2]
[3]
Athanasopoulos G.A., Pelekis P.C., Ground Vibrations from Sheet Pile Driving
in Urban Environment: Measurements, Analysis and Effects on Buildings and
Occupants. Soil Dynamics and Earthquake Engineering; (2000) 19:371387.
[4]
ICSGE-13
GEO-006-22
[5]
Holeyman A.E., Legrand C., Van Rompaey D., A Method to Predict the
Drivability of Vibratory Driven Piles. Proceedings of the 3rd International
Conference on the Application of the Stress-wave Theory to Piles, Orlando,
U.S.A.; (1996) 11011112.
[6]
Ramshaw C.L., Selby A.R., Bettess P., Computed Ground Waves due to Piling.
In Geotechnical Earthquake Engineering and Soil Dynamics III, Seattle, Dakoulas
P, Yegian M, Holtz RD (eds); (1998) 14841495.
[7]
[8]
Al-Abdeh, R., Etude des Vibrations Induites dans le Sol Par le Battage et le
Vibrofonc Age de Peiux. Ph.D. Thesis, Universite des Sciences et Technologies
de Lille, France. (2005)
[9]
Segol, G., Lee, P.C.Y., Abel, J.R., Amplitude Reduction of Surface Waves by
Trenches. Journal of the Engineering Mechanics Division, (1978) ASCE 104 (3),
621641.
[10]
[11]
Ahmad, S., Al-Hussaini, T.M., (1991) Simplified Design for Vibration Screening
by Open and In-Filled Trenches. Journal of Geotechnical Engineering, ASCE
117 (1), 6788.
[12] Luong, M.P., Efficiency of a Stress Wave Mitigation Barrier. Proceedings of the
International Conference on Centrifuge Modelling (Centrifuge 94), Balkema,
Singapore, (1994) pp. 28388.
[13] Massarsch, K.R., Ground Vibration Isolation using Gas Cushions. In:
Proceedings of the 2nd International Conference on Recent Advances in
Geotechnical Earthquake Engineering and Soil Dynamics, St. Louis, MO, 1115
March 1991. University of Missouri at Rolla, Rolla, MO, pp. 146170.
[14] Xu, Y., Huang, J., Du, Y., Sun, D., Earth Reinforcement using Soil Bags.
Geotextiles and Geomembranes 26 (3), (2008) 279289.
[15] Aboudi, J., Elastic Waves in Half-Space with Barrier. Journal of the
Engineering Mechanics Division, (1973) ASCE 99 (1), 6983.
[16]
[17]
[18] Al-Hussaini, T.M., Ahmad, S., Simple Design Methods for Vibration Isolation by
Wave Barriers. In: Proceedings of the 2nd International Conference on Recent
Advances in Geotechnical Earthquake Engineering and Soil Dynamics, St. Louis,
MO, 1115 March 1991. University of Missouri at Rolla, Rolla, MO, pp. 149399.
ICSGE-13
GEO-006-23
[19]
Al-Hussaini, T.M., Ahmad, S., Active Isolation of Machine Foundation by InFilled Trench Barriers. Journal of Geotechnical Engineering, ASCE 122 (4),
(1996) 288294.
[20] Yeh, C.S., Liao, W.I., Tsai, J.F., Teng., T.J., Train Induced Ground Motion and
its Mitigation by Trench and WIB. Report of NCREE-97-009, National Centre for
Research on Earthquake Engineering, Taipei, Taiwan, ROC. (1997)
[21]
[22]
[23]
Wang, J.G., Sun, W., Anand, S., Numerical Investigation on Active Isolation of
Ground Shock by Soft Porous Layers. Journal of Sound and Vibration 321,
(2009) 492509.
[24] Yang, Y.B., Hung, H.H., A Parametric Study of Wave Barriers for Reduction of
Train-Induced Vibration. International Journal for Numerical Methods in
Engineering 40, (1997) 37293747.
[25]
Comina, C., Foti, S., Surface Wave Tests for Vibration Mitigation Studies.
Technical note. Journal of Geotechnical and Geoenvironmental Engineering ASCE
133 (10), (2007) 13201324.
[26] Woods, R., Screening Surface of Waves Soils. Journal of Soil Mechanics and
Foundations 94, (1968) 951979.
[27] Lysmer, J., Waas, G., Shear Waves in Plane Infinite Structures. ASCE Journal
of Engineering Mechanics 98, (1972) 85105.
[28] El-Sheikh, S. A. Effect of Vibration Generated due to Execution of Construction
Works on Nearby Structures, Ph.D. Thesis, Ain Shams University, Cairo, Egypt.
(2005)
[29]
Richart, F.E., Jr., Foundation Vibrations, Trans. A.S.C.E., Vol. 127, part 1, pp.
863-898. (1962)
[30]
[31]
Egyptian Code of Practice, Part 4, Deep Foundations, Code No. 202/4. (2001)
ICSGE-13
GEO-006-24
13th ICSGE
27-29 Dec. 2009
Cairo - Egypt
ABSTRACT
One of the geotechnical problems for onshore structures is the existence of shell as
pieces or as calcareous fines with the fine sandy soil near the seashore. Shell existence
affects the physical and mechanical sand properties. Experimental study is presented on
the formation of shelly sandy soil. The effect of different compositions between shell,
lime and sand mixed with sea water on the maximum dry density, optimum moisture
content, shear strength parameters, compressibility characteristics and stress- strain
behavior are studied. The shell content for soil samples ranges between10 to 70% of
sand weight. Different lime ratios of sand weight were added to the mixtures as
stabilizing material. For each sample, maximum dry density, and optimum moisture
content were determined using standard Proctor test. For direct shear test, shear strength
parameters were measured and compressibility behavior was studied using Oedometer
test.
KEYWORDS
Shell, fine sand, lime, shear strength, maximum dry density, optimum moisture content.
1 INTRODUCTION
Egypt is rich with its sea shores on both the Mediterranean and the Red sea. Shells are
usually found on shores mixed with soil in different percentages and sizes. Soil mainly
consists of sandy soil with shells wither it is cemented or loose depends on how far from
the shoreline; also existence of shell material changes the soil properties. There are
many projects need to be constructed on the sea shores.
Few researchers discussed the calcareous shelly sand properties and shell effect on shear
strength parameters. El Shennawy [1] studied the effect of shell size on sand shear
strength and concluded that with the increase of shell percentage increases the shearing
angle till 40 %. Mohamed [2] studied the effect of shell size on compressibility of
sand using Oedometer test and concluded that shell content from 20 to 30% improved
the compressibility characteristic of the mixture. Hamdy [3] studied three different shell
ICSGE-13
GEO-007-1
samples collected from Hurgada and concluded that the presence of shell decreases the
bearing capacity of footing, and increases the footing settlement. Joer et al. [4] found
that calcareous soils have a particular character with regard to liquefaction potential.
Ohno et al. [5] concluded that, the calcareous offshore sand has an extreme soil type in
terms of compressibility. Yasufuku and Hyde [6] concluded that the crushable nature of
the shelly sand causes unfavorable strain affecting the pile end bearing capacity. ElSaadany et al. [7] studied the effect of lime on the collapsibility of compacted fine sand.
This paper studies the properties of shelly sand, different tests were carried out such as
sieve analysis, Standard, and Modified Proctor compaction tests. Also direct shear test
was carried out for all the specimens to study the effect of different shell content, and
lime content, with existence of salty water on the shear strength parameters of shelly
sand. Oedometre test was used to measure the compressibility and stress strain behavior
with different shell contents.
2
EXPERIMENTAL PROGRAM
2.1 Materials
2.1.1 Sandy soil
The used sand has been classified as poorly graded (SP) according to (USCS). Results
of sieve analysis is shown in Fig. 1, results of compaction test and direct shear test are
shown in Table 1.
Table1: Physical and Mechanical properties of sand
Sand
C
d-min d-max
Cu
Cc
Gs
100
Shell
Fine Sand
Passing Percentage %
90
80
70
60
50
40
30
20
10
0
1000
100
10
0.1
0.01
ICSGE-13
GEO-007-2
2.1.2
Shells
The used shells were collected from Elarish beach at north Sinai. The delivered shells
were washed and left to dry in air. The shells were sieved as shown in figure1. Dry shell
were chemically analyzed, it contains mainly 88.19 % Calcium oxide (Ca O), and trace
of Sulphur trioxide (SO3), Sodium chloride (Na CL) and Magnesium oxide (Mg O),
with some of Silicon dioxide (SiO2), Ferrous oxide and Aluminum oxide (FeO+Al2O3).
Shells size was measured using sieve analysis, it was less than 25 mm as shown in
figure 1, and it was more reliable to use shells with different sizes in all the tests by
weight. The shell contents ranges between10 to 70% of sand weight. During compaction
tests or during preparation of shear box and compression tests samples, the shell size
changes to finer pieces which changes sand gradation.
2.1.3
Lime
In this study, Lime (Ca (OH2)) was used as cementing material with the shell and sand.
Also, it works as stabilizing material or as a soil treatment to improve the mixture
properties. Lime powder was used with different percentages 3, 6, 10, 17 and 25 % of
sand weight.
2.1.4 Salty water
In order to simulate the shelly sand on shore, salty water was used during compaction
tests and during samples preparation for direct shear and Oedometer tests. Water
samples from Mediterranean and Red seas were chemically analyzed to measure the
amount of total salts in both samples. Table 2 shows the analysis of both samples. In
this research, salty water with concentration 41 gm/liter was used for all the samples.
Table 2: Chemical analysis of sea water
Chemical
Properties
Mediterranean
Sea
Red Sea
T.D.S
CL
SO3
CO3&HCO3
pH
34 gm/liter
17.2 gm/liter
2.4 gm/liter
0.04 gm/liter
7.5
41 gm/liter
23.4 gm/liter
2.7 gm/liter
ICSGE-13
GEO-007-3
predetermined shell percentages with or without 10% lime. Then salty water was added
according to OMC which was found in compaction test for the same shell and lime
percentages. Shell, sand, with or without lime and salty water mixture were mixed into a
homogenous mixture. Mixture was compacted in a standard Proctor mold. Direct shear
sampler of 60mm square, was placed in the mold to get a specimen and immediately
tested in shear box with no curing time for the samples contain lime. The direct shear
test was conducted to measure the change in shear strength parameters of the mixture.
The tests were conducted using conventional step loading method with deformation rate
of 0.59 mm/min under various normal stresses 50, 100, 150, and 200 kN/m2.
3
20.5
Pure Sand
10% Shell
20% Shell
30% Shell
40% Shell
50% Shell
60% Shell
70% Shell
(a)
20.0
19.5
19.0
18.5
18.0
17.5
17.0
16.5
16.0
(b)
d =1.7301+0.00451 Sc
20.5
R =0.98633
20.0
14
18
20
22
24
20
(c)
18
17
19.5
16
19.0
15
18.5
18.0
14
Pure sand
13
12
11
17.5
10
17.0
16.5
16
19
OMC %
12
WC%
21.0
10
Pure Sand
0
10
20
30
9
40
50
60
70
80
Shell content Sc %
10
20
30
40
50
60
70
80
Shell content Sc %
ICSGE-13
GEO-007-4
The relation between maximum dry density and shell contents is presented in figure 2-b
as a linear relationship. This means that, increasing of shells content increases the
maximum dry density because the exictance of shell fragment due to compaction
process which changes the sand degradation besides the salty water increases the
maximum dry density. The results between shell content and optimum moisture content
are presented in figure 2-c, it is concluded that no relation between the optimum
moisture content and shell content.
Maximum dry densities measured for the tested specimens using different shell content
with 10% lime are presented in figure 3-a. It can be seen that by adding 10 % lime for
different shell content, maximum dry density of the specimens increased from 17 kN/m3
for pure sand to 20.6 kN/m3 for shell content 70 %. Fig. 3-b presents the relation
between maximum dry density and shell content as a nonlinear relationship with the
presence of 10% lime. The results showed that the maximum dry density for shelly sand
samples increases as shell content increases as shown in figure 3-b. Optimum moisture
content decreases as shell content increase by adding 10 % lime as shown in figure 3-c.
Comparing zero and 10 % lime samples, the decrease in optimum moisture content with
is due to lime existing. Note that in both figures 2 and 3 the samples were mixed with
salty water during compaction which means that the decrease in optimum moisture
content is due to lime existence only.
21.0
(a)
20.5
Pure sand
10% Shell
20% Shell
30% Shell
40% Shell
50% Shell
70% Shell
20.0
19.5
19.0
18.5
18.0
17.5
17.0
Pure sand
16.5
16.0
10 11 12 13 14 15 16 17 18 19 20
Wc%
20.5
14.5
(b)
=1.765+0.00756 Sc-4.7619E-5 Sc
d
2
R =0.99149
14.0
13.5
Pure Sand
12.5
12.0
19.5
11.5
19.0
d =1.7301+0.00451 Sc
18.5
R =0.98633
11.0
10.5
10.0
9.5
18.0
9.0
8.5
17.5
17.0
16.5
(c)
13.0
20.0
OMC %
21.0
Pure Sand
0
10
20
30
40
50
60
70
8.0
7.5
7.0
80
Shell content Sc %
R =0.97599
2
10
20
30
40
50
60
70
80
Shell content Sc %
Fig.3: Effects of shell content and 10% lime on maximum dry density and OMC.
ICSGE-13
GEO-007-5
(a)
Zero Lime
3% Lime
6% Lime
10% Lime
17% Lime
25% Lime
9 10 11 12 13 14 15 16 17 18 19 20
Wc %
15
d =1.866+0.0106 L-2.06E-4 L
20.2
(b)
(c)
14
Zero Lime
R =0.988
20.0
13
19.8
OMC =9.514+0.095L
2
19.6
OMC %
20.4
19.4
19.2
19.0
R =0.9898
12
11
10
18.8
18.6
18.4
18.2
18.0
Zero Lime
8
0
10
12
14
16
18
20
22
24
26
28
30
Lime content L %
10
12
14
16
18
20
22
24
26
28
Lime content L %
Fig.4: Effects of lime content on maximum dry density and optimum moisture content.
ICSGE-13
GEO-007-6
3.2
60
130
(a)
o
=62504-0.0.416 Sc
50
R =0.9972
45
40
0
=54.88429-0.38961 Sc+0.00168 Sc
35
2
10
20
90
80
70
60
50
40
30
Pure Sand
10
Pure Sand
30
R =0.99824
100
20
R =0.97944
30
40
50
60
70
(b)
110
2
55
C =0.02686-0.0012 Sc+2.21429E-4 Sc
120
Zero Lime
10% Lime
Cohesion c (kN/m )
65
80
Shell content Sc %
10
20
30
40
50
60
70
80
Shell content Sc %
ICSGE-13
GEO-007-7
(with shell content 30%) decreases the friction angle more than as zero lime, as
shown in figure. 6-a. Shear parameter, cohesion, c, measured for specimens at different
lime ratios and constant shell content (30%) are shown in figure 6-b, it can be seen that
the increase in lime content causes small increases the cohesion, c, due to the
cementation. According to the non linear fitting line, the cohesion gains increase from 0
to 6 kN/m2 at 3 % lime content, more gain took place in the shearing resistance
parameter c, depending on increase of lime content to 17%.
In summary, the angle of shearing resistance, , decreases with the increasing of lime
content in existence of Sodium Chloride. While the cohesion, c, increases with the
increase of lime content to 17%. Lime and fines due to shell crushing during sample
preparation processing, with sand particles created cementing agent as semi gel with no
curing time for the samples. This semi gel created lubricates the grains and causes
decrease in shearing angle of friction while increasing the cohesion.
(b)
20
50
45
=63.65-1.746L
Zero Lime
55
25
(a)
Cohesion c (kN/m )
60
R =0.9929
40
35
30
25
20
15
10
12
14
R =0.99946
10
Zero Lime
C =-0.00191+0.02236 L-5.21481E-4 L
15
16
18
20
22
24
26
28
Lime content L %
0
0
10
12
14
16
18
20
22
24
26
28
30
Lime content L %
ICSGE-13
GEO-007-8
200
400
600
10% Shell
20% Shell
30% Shell
40% Shell
50% Shell
70% Shell
Axial Strain %
1
2
3
4
5
6
200
400
600
800
1000
1200
1600
1800
Loose mixture
standard comp.
Modified comp.
Axial Strain %
1400
4
6
8
10
12
ICSGE-13
GEO-007-9
CONCLUSIONS
From previous study, it is concluded that, increasing shell content with no lime and with
salty water, increases the maximum dry density as a linear relationship, and does not
affect on the optimum moisture content. Also increasing shell content to 10% increases
the angle of shearing resistance, but higher shell content decreases it to higher than the
pure sand and no cohesion was created even with existing salty water during sample
preparation. Using sea water in the mixture with increasing the lime percentages,
increases the maximum dry density, decreases the OMC. Increasing the lime contents,
causes slight increase the cohesion while angle of shearing resistance , increases till
3% lime content, and decreases for higher lime content to higher than the pure sand.
Increasing shell content leads to increase the compressibility of the mixture. The
increase of density or sedimentation decreases axial strain.
5
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
REFERENCES
El-Shennawy, M.T., Effect of Shell on Shear Strength of Sand, Master thesis
Cairo University, Egypt, (1981) 95.
Mohamed, A.M.., Effect of Shell on Compressibility of Sand, Master thesis
Cairo University, Egypt, (1981) 110.
Hamdy D. B., Geotechnical Properties of Shelly Sand at Hurgada, Eleventh
International Colloquium and Geotechnical Engineering. Ain Shams University,
Egypt, (2005) E05GE14-1- 14-8.
Joer, H. A., Bahattaria, B., Fahhey, M., and Randolph, M. F., Assessing the
liquefaction of Calcareous Soil, Proceeding 11th European Conference of Soil
Mech. And Found. Engineering. Copenhagen, Vol.3, (1995) 107-112.
Ohno, S., Ochjai, H., and Yasufuku, N., Estimation of Pile Settlement in
Calcareous Sands, Proceeding 2nd International Engineering for Calcareous
Sediments, Vol. 1, Bahrain (1999) 21-24.
Yasufuku, N. and Hyde, A. F., Pile End Bearing Capacity in Crushable Sands
Geotechnique 45, No. 4, (1995) 663-676.
El-Saadany, M.M., Abou-Rayan, A. M., and Easa, A. A., Effect of Lime on the
Collapsibility of Compacted Sand, Civil Engineering Research Magazine,
Azhar University, Vol. (24) No. 1, (2002) 214-232.
Senoon, A., A., Effect of Lime Admixture on The Properties of Compacted
Sandy Soil, Journal of the Egyptian Geotechnical Society, Vol. 17, part 2,
(2006)161-175.
Lamb, T.W. and Whitman, R.V., Soil Mechanics, John Wiley and Sons Inc.,
New York, (1969).
ICSGE-13
GEO-007-10
13th ICSGE
Faculty of Engineering
Cairo - Egypt
ABSTRACT
Due to highly increasing population rate and urban development, many structural
problems occurred in many parts of the world when constructing the wastewater
networks. One of the main problems is cracking of pipes as a result of many
factors like bad structures, overloading the roads, erosion and agitation of soil.
These factors lead to a series of problems that can lead to the collapse of
networks by the seepage of groundwater through the cracks into the pipes or
wastewater from the pipes to the soil. The collapse of network affects the soil and
lead to the occurrence of ground surface settlement and soil settlement around
the pipes, which cause settlement of adjacent buildings and structures, in many
cases.
KEYWORDS
Pipelines, sewer, soil, settlement, cracks.
1
INTRODUCTION
The study methodology and the derived numerical model were applied to the network
system in one of the most crowded area in downtown Cairo (the intersection between
Magles Al Shaab St. and Mansour St.) where an incident of soil settlement occurred in
August 2008, followed by the collapse of a structure next to this settlement [1]. All
ICSGE-13
GEO-008-1
concerning data of the pipes in this area were collected, including historical indicators
and incidents that happened before the main settlement in 2008. The numerical model
derived enabled stimulating the entrances of the settlement, accurately, over two years
and the accurate simulation of soil settlement in the area of the study, as well as the
value of generated cracks in the surrounding buildings and stages of the damaged
building. A numerical model was designed using all observed and monitored data for
settlements (CCTV, GPR, ) and the size of the mesh used depends on results from
previous extensive parametric study for different factors. Different scenarios was
proposed for failure mechanisms until we choose the one that gives very good
agreement with observed data.
At this location, a concrete sewer pipe, two meters in diameter, which flows wastewater
to CWO (Cairo Wastewater Organization) Tunnel, is buried at a depth of approximately
five meters according to the as-built drawings and field inspections. The selection of
this site is due to the subsidence accident of pavement surface over the last year due to a
leakage and an inflow and infiltration across the buried pipes. In the past years a
significant number of ground subsidences have occurred in this area.
2
Microtunnel
ICSGE-13
GEO-008-2
The sewer pipeline at this location passed entirely through the 4m clay stratum.
Overlying this is a layer of 4m of made ground and road surfacing [3]. The
pipeline stratum is above sand layer to the end of boring at -15.45m.
Groundwater level is found at -1.20 m from ground surface. A detailed ground
investigation was conducted after the accident. Three boreholes were carried out
at the location of the study area as shown in Fig. 1.
2.3 Pipelines
The study area presented herein contains two sewer pipelines with burial depth
5.00 m; the first pipeline is 1600 mm internal diameter precast reinforced
concrete sewer pipeline with pipe segments of 2.00 m length, and the second is
600 mm internal diameter vitrified clay sewer pipeline with pipe segments of
3.00 m length. The first pipeline was constructed in 1988 by Trenchless-Jacking
method within the Greater Cairo for Sanitary Drainage project to improve the
sanitary of Greater Cairo. The second pipeline was constructed by cut-and-cover
method in 1989.
Historically, the first ground subsidence, had taken place in September 2006, was
a consequence for structural deterioration of the 600 mm sewer pipeline. Figure
2 illustrates the CCTV photo for 600mm sewer pipeline diameter. This photo
shows the fracture in pipes. The fracture is a crack line along pipe wall in which
the pipe material is still intact with a crack line that is visibly open in the sewer
pipe [4]. The Greater Cairo Sanitary Drainage Company (GCSDC) had made the
repair for the three defected pipes by using stainless steel pipes encased in the
fractured pipes.
.
Fracture
Longitudinal
Many important structures, buildings and fences, surround the accident area.
These structures belong to four organizations; ministry of justice, ministry of
monuments, ministry of finance, and national authority for tunnels. About 15
ICSGE-13
GEO-008-3
structures were monitored after the starting of this accident, three of them
belonging to the ministry of justice were affected. Figure 3 shows the ministry of
justice damaged Notary building before the demolition.
Fig. 3: Views of the front and side of the damaged Notary building.
3 SETTLEMENT MONITORING AND ITS EFFECTS ON STRUCTURES
Three methods of monitoring have been conducted in the selected site. The first
method is the settlement monitoring for the structures by using total station tool.
The second method is the structural monitoring for the sewer pipelines by using
closed circuit television (CCTV). The third method is the subsurface monitoring
by using ground penetrating radar (GPR).
3.1 Settlement Monitoring
The project consultant has completed the daily readings report from 8 to 30
August 2008 for settlement of the ground at the selected site [5]. Survey markers
installed on the selected structures are shown in Fig. 4. Survey readings were
recorded on a daily basis. The elevation data from these readings are provided in
Fig. 5 and 6.
ICSGE-13
GEO-008-4
CCTV inspections was conducted to the sewer pipelines within the study area.
This inspection was required to assess structural conditions of the sewer pipeline.
Condition assessment based on CCTV inspection has traditionally been the
approach used for rehabilitation decisions in sewer networks. Current practice for
CCTV evaluation involves examination with a pan and tilt CCTV camera to
record the service condition of the sewer. Survey data is coded by trained
operators according to an accepted protocol such as the WRC Manual of Sewer
Classification 3rd edition and entered into a proprietary database.
ICSGE-13
GEO-008-5
The CCTV video has provided us with information that is beneficial for
determining points of extraneous flow (e.g. displaced joints, cracks, voids, and
leaking, or broken, etc) and identifies mainline and service connection problems.
Figure 7 illustrates the CCTV photo for 600mm sewer pipeline diameter. This
photo shows the breaking in pipes. The Breaking is a visible displacement of
pipe material from its original position [4].
Broken
Fig. 7: CCTV of 600mm diameter pipe (Cairo 8/2008).
Figure 8 illustrates the CCTV photo for 2000mm sewer pipeline diameter. This photo
shows the open-joint in pipes. The open-joint defects describe visible displacement at
joints in sewer pipes. The open-joint in the pipes was so severe that the soil was visible
through it [4].
Joint open
Fig. 8: CCTV of 2000mm diameter pipe (Cairo 8/2008).
3.3 Ground Penetrating Radar (GPR)
Ground penetrating radar (commonly called GPR) is a high resolution electromagnetic
technique that is designed primarily to investigate the shallow subsurface of the earth,
building materials, and roads and bridges.
The radar profile obtained along adjacent damaged building [6]. In the field, the GPR
survey along the mentioned grid line was performed in a continuous base. The graphic
representation of the reflected radar signals shows clearly several anomalies along the
surveyed area. The radar reflections describe certain uniformity along the 5-meter long
ICSGE-13
GEO-008-6
survey in the last part of the profile. This uniformity is interrupted in the area at the start
of damaged building indicating a variation in the continuity of the soil stratigraphy.
These irregularities can be the result of disturbances to the sedimentation of the soil
caused by the soil infiltration toward the sewer pipeline.
4
NUMERICAL MODEL
ICSGE-13
GEO-008-7
ICSGE-13
GEO-008-8
pipeline failure. It is performed using a finite element calculation considering the own
weight of both the soil and the structure. Displacements are reset to zero at the end of
this stage; consequently, results referred to hereafter are due to the pipeline
deterioration. The second step (pipeline deterioration state) deals with the numerical
simulation for the failure of the pipeline in presence of the structure. The pipeline
deterioration operation is modelled by reducing the stiffness of soil above the pipeline.
The failure mechanism in study area is assumed here by reducing stiffness around and
above the R.C. pipeline in three stages. The reduction in soil stiffness is due to two
reasons; the first reason is the sewage water coming from the deteriorated pipeline, and
the second reason is the infiltration of soil to the 2000 mm diameter pipeline. Figure 10
illustrates the four zones of soil which are reduced in stiffness. Table 2 contains the
different values of soil stiffness (E) for four zones of soil in the three stages. These
values are relative to the value of the existing soil stiffness; Emg (stiffness of made
ground stratum) and Ecl (stiffness of clay stratum).
Table 2: Soil stiffness of different zones in the three stages.
Stage No.
Emg1
Emg2
2
Ecl1
2
Ecl2
2
(400t/m )
(400t/m )
(3000t/m )
(3000t/m2)
Stage 1
0.50 Emg
Emg
Ecl
Ecl
Stage 2
0.25 Emg
0.50 Emg
0.50 Ecl
Ecl
Stage 3
0.25 Emg
0.25 Emg
0.50 Ecl
0.50 Ecl
ICSGE-13
GEO-008-9
ICSGE-13
GEO-008-10
In the Magles Al Shaab Street model, the distribution of settlement looks reasonable,
with the building responding almost as a rigid body with relatively light damage.
Moreover, the magnitude of the building settlement from the model is approximately
equal to what was observed. The best agreement between the model and observations
was obtained when the stage of failure mechanism was stage-3 and with the observed
settlement on the last day (30 August 2008). The initial analyses with stage 1 did not
give particularly good agreement with observations in the detail of the building
performance.
The results show the influence of soil properties on the damage category predicted for
the building, and the study implies that a high quality investigations should be carried to
complement the use of the analysis techniques in practice.
The results are presented in Table 3, for the faades of damaged building, which
illustrate damage category in different stages of failure mechanism. Good agreement
was found in the distribution and severity of damage. This confirmed the applicability
of the general approach of modeling used in this study.
Figure 12 shows the relation between the vertical settlement and the horizontal distance
of damaged building at axis 1 (Fig. 9 and 10), and the soil stiffness changing around and
above pipeline. We can find out that the building deformation increases with decreasing
the soil stiffness above the pipeline. In addition, The settlement profile along the west
and east facades show sudden changes concentrated beneath columns, the highest value
lies under the nearest row of columns to the damaged pipeline.
Table 3: Evaluation of potential damage to in-fill walls and beams in frame
structures due to soil stiffness above pipeline.
Case
Properties
Unit
Stage 1
58.9
Stage 2
143.6
Stage 3
147.2
-6.6
-19.6
-20.5
Mm
65.4
163.2
167.7
Mm
Angle of Tilt ()
Cumulative Tensile crack width (Ct)
0.250
0.624
0.640
deg.
0.0
0.9
1.0
Mm
7.5
18.5
18.5
Mm
Moderate
Severe
Severe
Mm
The settlement profile along the cars shed from the field and model are compared in Fig.
13. In this case, the settlements recorded 20 days after the accident on 30th August 2008
are given. It can be seen that along cars shed the settlements predicted by the model
agree much better with the final settlements in the field rather than with the short-term
settlements.
Figure 14 shows the settlement profile obtained along the west and east faades
footprint of electric room. The profiles for stage 2 and stage 3 are approximately
ICSGE-13
GEO-008-11
similar, as are the extent and severity of settlement to the footings. The settlement
profile along the electric room from the field and model are compared in Fig. 14. In this
case, the settlements recorded 20 days after the accident on 30th August 2008 are given.
The remarkably good agreement of the observed settlements with the model at stage 3
has to be noted. The settlement profiles along the facades of electric room show sudden
changes concentrated beneath columns of the room .
ICSGE-13
GEO-008-12
From our study we can conclude that pipeline failure affects ground surface and
adjacent buildings. The ground surface deformation and buildings differential
displacements are highly affected by the decrease of soil stiffness above pipelines.
After analyzing our case study, we recommend that:
In the event of any major settlement, a total study should be carried within
sufficient area around the incident.
Data should be collected from any previous cases with all available means such
as observations, CCTV or GPR etc. this should be analyzed for future
prevention and remedial of existing damage.
A total analysis should be run for any accident such as pipe cracks and soil
settlement to predict the main reason so that it wont recur again after
temporary remedial and affect adjacent structures.
After such incidents, adjacent buildings should be supported to prevent building
deterioration due to continuous settlement of pipes.
7
REFERENCES
[1]
[2]
[3]
[4]
ICSGE-13
GEO-008-13
[5]
A &A, "Monitoring Report for Structures for Period from 13-8-2008 to 1-92008, A&A Engineers and Consultant, Cairo, Egypt (2008).
[6]
[7]
ECP, The Egyptian Code for Soil Mechanics, Design, and Construction of
Foundations, 7th Edition, The continua Committee for Preparing the Egyptian
Code for Soil Mechanics, Design, and Construction of Foundations, Part 3, ECP,
Egypt (2001) 202-2001.
[8]
Swanson, P.G., ANSYS, Inc. Theory, Theoretical Manual, Release 11.0, U.S.A.
(2007).
[9]
[10]
[11]
ICSGE-13
GEO-008-14
13th ICSGE
Faculty of Engineering
Cairo - Egypt
ABSTRACT
The study here uses the three-dimensional (3-D) Finite Elements Method,
utilizing the "ANSYS" software, to simulate the relationship between: pipes,
structures and surrounding soil, with different effects (type and diameter of pipes,
type of soil and wastewater) to calculate the shape of potential settlement of
ground surface under such effects. The results of this numerical analysis were
mapped on a Geographic Information System (GIS) software, to evaluate the
effect of pipe settlement on the surrounding structures and the extent, nature,
shape and values of cracks expected to be generated in these structures. The
study was thus able to provide an integrated technique that combines the results
of numerical analysis (FEM) and Geographic Information Systems (GIS) to deal
with problems associated with settlement of sewer pipes and forecasting the
shape and value of expected soil settlement and the type of cracks generated in
the surrounding structures.
KEYWORDS
Pipelines, sewer, soil, GIS, settlement, cracks.
ICSGE-13
GEO-009-1
INTRODUCTION
The ability of GIS to search databases and perform geographic queries has
revolutionized many areas of science and business. It can be invaluable during a
decision-making process. The information can be presented succinctly and clearly in the
form of a map and accompanying report, allowing decision makers to focus on the real
issues rather than trying to understand the data. Because GIS products can be produced
quickly, multiple scenarios can be evaluated efficiently and effectively. For this reason,
in todays world, the ability to use GIS is increasingly important.
The design process of using FEM and GIS is presented in this study for the example of
an area in Shobra region, suburbs of Cairo, Egypt. A settlement risk assessment is
performed to determine the influence of structural deterioration of a pipeline. The two
main stages in this process are the FE-settlement predictions and its implementation in
GIS. GIS uses the FEM output (settlement predictions in the surrounding soil mass) to
analyze the risks of damage impact in form of a damage classification for each adjacent
building. Both stages are described in this paper. Besides the design stage, these
information tools are of high importance during pipeline or building rehabilitation,
especially when using the observational method. A GIS system, being a powerful
visualization tool, makes it possible to compare the predictions with the actual
monitoring readings and thereby assists in steering the damage assessment process.
In this study, the settlement prediction maps are created by interpolating the settlement
point data, which are obtained from the FEM results. The spatial analyses by using the
GIS modules and the calculation of building deformation by using spreadsheets are used
to categorize the buildings damage. The ArcGIS software is used in spatial analysis and
maps producing. The equations of buildings damage (differential settlements, tilt,
strains, crack width, and building classification) are programmed in Excel spreadsheets.
2
LITERATURE REVIEW
Haasnoot and Netzel [1] presented the design process of using FEM and GIS for the
example of a station box of the North South Line in Amsterdam. A settlement risk
assessment was performed to determine the influence of construction. It was pointed out
that, the two main stages in this process are the FE-settlement predictions and its
implementation in the GIS. GIS uses the FEM output (settlement predictions in the
surrounding soil mass) to analyze the risks of damage impact in form of a damage
classification for each adjacent building. Beside the design stage, these information
tools are of high importance during construction, especially when using the
observational method.
Yenigul [2] used three-dimensional GIS and two-dimensional (2-D) finite element
modeling to analyze the settlement of subsurface, where these applications were carried
out and the results are compared and correlated. Settlement analysis by using threedimensional GIS approach is performed for two different geometrical conditions by
using modified Terzaghi approach or the so-called Koppejan Formula for long-term
settlement. Thereafter, the settlement results are presented as settlement prediction maps
for comparison purposes and as a 3-D isovolumes to reflect the real picture of
settlement predictions. After performing the 2-D based settlement analysis on the
sections taken from the 3-D geology model, the results are compared and correlated
visually and statistically with the 3-D analysis results at the same sections. Statical
ICSGE-13
GEO-009-2
analysis results showed that there is a significant linear relationship and a very strong
correlation between 2-D FEM based and 3-D GIS based settlement analysis results.
2.1 Building Damage Prediction and Classification
Settlement of structures, whether from nearby construction or other causes, can result in
noticeable damage. Such damage can be significant and costly. Usually, the most
settlement sensitive buildings are those with masonry load bearing walls or frames with
masonry in-fill walls. By combining ground deformation patterns, well-known damage
category criteria, strain superposition and critical strain concepts, the potential effects of
building deformations can be readily assessed without undue oversimplification.
Boscardin and Cording [3] extended the framework of limiting tensile strain by linking
strain values to building damage observed from case studies subjected to excavation
induced subsidence. In their work they linked values of limiting strain with the
categories of damage shown in Table 1.
Burland [4] compared the behavior of load bearing masonry walls undergoing
settlement to the bending of a deep beam subjected to a point load at the beam centre.
One of the more important aspects of their work included a summarization of damage
severity based on observed crack width as illustrated in Table 1. This damage
categorization is widely referenced and is intuitive, practical, and related to measurable
evidence. Yet, the deformation and critical strain criteria were not directly linked to
damage category.
Table 1: Severity of cracking damage
Category of Normal degree of
Approximate Individual Crack
Damage
severity
Width (mm)
0
Negligible
< 0.1
1
Very Slight
0.1 to 1
2
Slight
1 to 5
3
Moderate
5 to 15 or a number of cracks > 3
4
Severe
15 to 25 also depends on the number of cracks
5
Very Severe
> 25 also depends on the number of cracks
2.2 Geographic Information Systems (GIS)
The literature presents a large number of GIS definitions. A single, reasonably inclusive
definition is that a GIS consists of computer-based tools that are used to capture, store,
update, manipulate, retrieve, analyze, display, print, and otherwise manage large
amounts of geographic and attribute data. Originally, these tools were developed to ease
cartography, but they are currently being used for diverse applications such as,
planning, facilities management, and hydrologic and hydraulic (H&H) modeling.
Alternatively, one can simply define a GIS as a link between a map and a database as
shown in Fig. 1.
A geographic information system (GIS) is a computer-based tool for mapping and
analyzing spatial data. GIS technology integrates common database operations such as
query and statistical analysis with the unique visualization and geographic analysis
benefits offered by maps. These abilities distinguish GIS from other information
systems and make it valuable to a wide range of public and private enterprises for
explaining events, predicting outcomes, and planning strategies. GIS is considered to be
ICSGE-13
GEO-009-3
one of the most important new technologies, with the potential to revolutionize many
aspects of society through increased ability to make decisions and solve problems.
The different scenarios of pipeline failure and its impact on the nearby buildings were
studied as well as the calculation and classification of building damage by Metwally [5].
These scenarios of pipeline failure are pipeline settlement, the influence of the soil
stiffness changing above pipeline, and the influence of the sewage water around
pipeline. The results of these calculations are automatically imported in the GIS system.
The GIS is capable of extracting deformation data along any given line. The Settlement
Risk Module of the GIS will verify the FEM results in relation to the surrounding
structures.
The damage classification system relates the differential ground movements to the
anticipated damage of a structure. The damage classification system for the case of
study is derived from numerical calculations, determining the deformation capacity of
specific reinforced concrete and masonry structures in the study area in combination
with analytical, empirical methods generally used for assessment of damage due to
excavation-induced ground movement. The damage classification describes the
deformation capacity of the buildings in terms of tolerable values for a damage
parameter (angular distortion, horizontal strain deflection ratio, tilt, slope, settlement),
which are related to the expected degree of damage (damage class).
The predicted damage class of each building is determined by projecting the predicted
settlements on the location of the buildings. The damage parameters (Cumulative
Tensile crack width and Cumulative Principal crack width) are calculated with the
(differential) ground movements for each structure and compared with the tolerable
values for the various classes in the damage classification system.
ICSGE-13
GEO-009-4
A. Model geometry
B. FEM Model
=2000 t/m2
=1.60 m
=0.35
=0.20 m
Soil cohesion, C
=2.00 t/m2
Pipe length, Lp
=2.00 m
=15 pipes
o
Angle of internal friction, =30
=1.85 t/m3
=3.5E6 t/m2
=5.0 m
=0.20
=0.32
=0.60
ICSGE-13
GEO-009-5
between the vertical settlement and the horizontal displacement of ground surface, and
the pipeline settlement. The figures exhibits that the minimum results are for the
minimum value of pipeline settlement.
ICSGE-13
GEO-009-6
The contents of the database (data layer) for the study area comprise spatial data and its
attribute data. These data layer includes basemap (streets, parcels, and buildings), soil,
and sewer networks (pipeline and manholes).
ICSGE-13
GEO-009-7
Trough
ICSGE-13
GEO-009-8
ICSGE-13
GEO-009-9
Fig. 10: Settlement prediction map of buildings due to pipeline settlement (5%
Diameter).
CONCLUSION
The major effect of pipe deterioration occurs within soil width equal six times the pipe
diameter or two times the burial depth from vertical axis of pipelines. Therefore, this
ICSGE-13
GEO-009-10
zone should be monitored and tested on a regular basis for early prevention of ground
surface failure and pipeline failures.
GIS is a very powerful tool that should be integrated with FEM for damage assessment
of sewer pipelines to determine the impact of the structural deterioration of pipelines on
ground surface and adjacent buildings. This will be time saving for important buildings
and main pipelines in heavy traffic areas.
6
REFERENCES
[1]
Netzel, H., and Haasnoot, J.K., Settlement Risk Assessment; Integrating FEM
and GIS, Proc. of the NUMOG 2002 conference in Rome, GL1822, Rome
(2002).
[2]
Yenigul, N.B., Comparative Settlement Analysis by Using 3D-GIS and 2DFEM Approaches, M. Sc. Thesis, Section Engineering Geology, International
Institute for Aerospace Survey and Earth Sciences (ITC), Delft, Netherlands,
July (2000).
[3]
[4]
[5]
[6]
[7]
[8]
[9]
Shamsi, U.M., GIS Tools for Water, Wastewater, and Stormwater Systems,
1st Edition, ASCE press, Library of congress catalog card No: 20020744, 7,
ISBN: 0-7844-0583-5, USA (2002).
ICSGE-13
GEO-009-11
13th ICSGE
Faculty of Engineering
Cairo - Egypt
ABSTRACT
The impacts of groundwater on underground construction projects are enormous.
Groundwater affects the design of the structures, the construction procedures and the
overall project cost. This study presents the results of a three-dimensional numerical
analysis of a dewatering system that was required for construction of an inlet pumping
station at the Abu Rawash waste water treatment plant. A comparison between the
numerical results and monitored response is presented to exhibit the adequacy of the
proposed numerical scheme. The numerical results compared favourably with the field
measurements. In addition, the results of a numerical sensitivity study are presented to
highlight the influence of some controlling parameters such as the aquifer hydraulic
conductivity and its anisotropy. Furthermore, the dewatering system was analyzed using
the conventional analytical equations of groundwater flow to assess the influence of the
analysis assumptions on the predicted response of the system.
KEYWORDS
Groundwater, semi-confined, aquifer, dewatering, three-dimensional, numerical
analysis.
1
INTRODUCTION
ICSGE-13
GEO-010-1
The basic assumptions of the analytical formulas are a uniform hydraulic conductivity
through the entire aquifer, full efficient well, well penetrates the entire aquifer, a state of
laminar flow through the aquifer is established, the initial water table is flat and the
drawdown and radius of influence are no longer change with continuous pumping at a
constant rate [1]. The analytical formulas were basically developed by Theim [2] and
Muskat [3]. A good review of these formulas is reported by Mansour and Kaufman [4],
Hausmann [5] and Powers [6]. Complicated flow problems can not be solved by
analytical methods. Some analytical solutions [7, 8, 9] address issues such as leaky
aquifers but they always assume a homogeneous system. Recent contributions [10, 11,
12, 13] provided analytical solutions that can deal with a complex aquifer by dividing it
into several regions of homogeneous properties. However, these types of models can
not represent many important regional hydrologic features that affect groundwater
movement (e.g., aquifer anisotropy, structural features, recharge and discharge) [1].
Numerical models are capable of solving the groundwater flow equations considering
the different site conditions. These models have the ability to overcome the limitations
associated with analytical equations. The continuous advances in computer systems,
regarding computational speed and storage capacity, facilitate handling the modeling of
such big problems. Computer packages of the finite element programs [14] and the
finite difference programs [15, 16] are widely used in numerical analysis of
groundwater flow problems. To study the impacts of a pumping well on groundwater
flow regime, Pinder and Bredehoeft [17] presented an axisymmetric numerical model
designed for a well-test interpretation. Many other efforts focused on developing
numerical models that investigated well-test interpretation, pumping impacts and the
impact of heterogeneity upon the uncertainty of the interpreted results [18, 19, 20, 21].
Numerical groundwater models were found to be the most appropriate tool that allows
simulation of complex three dimensional dewatering problems that contain gravity or
confined aquifers, soil anisotropy, partially penetrating wells, vertical boundaries,
vertical flowetc [22].
This study presents the results of a three-dimensional numerical analysis of a
dewatering system that was required to dry out a wide and deep excavation necessary
for construction of the inlet pumping station at the site of the Abu Rawash waste water
treatment plant (WWTP). A comparison between the numerical modeling results and
the monitored response is presented to evaluate the adequacy of the proposed numerical
scheme. The paper presents, also, the results of a sensitivity study that was carried out
to examine the effect of the aquifer permeability and anisotropy on the system response.
In addition, the dewatering system is analyzed using the analytical equation of semiconfined aquifer conditions together with the cumulative drawdown concept to
highlight the influence of the analytical equation assumptions on the analytically
predicted response of the dewatering system..
2
Construction of the inlet pump station foundations of the extension of the Abu Rawash
WWTP necessitated drying out the pump excavation site that was around 72m by 85m
in plan and 10.6m deep. The soil stratification at the inlet pump station site consisted of
a surface graded sand formation, 3m thick, underlain by a stiff clay layer that extended
to a depth of 8.5m below the ground surface (bgs). The clay formation was underlain by
a medium to fine sand with some silt that extended to the end of the carried out
ICSGE-13
GEO-010-2
borehole, 25m bgs. The piezometer reading of the groundwater of the bottom confined
aquifer was 5.6m bgs, no groundwater was noticed in the top sand layer.
The dewatering system constructed at the inlet pump station site consisted of 18
working deep wells in addition to 4 standby deep wells. Figure 1 shows the arrangement
of the deep wells around the excavation perimeter in addition to the piezometer location
within the site. The operating steady state discharge of the wells was 95 m3/hr. The
wells were 24m in length. The entire pipe of each well was slotted except the 5.0m at
the well top and the bottom 2.0m of the well pipe.
Fig. 1 Construction dewatering system of the inlet pump station: layout of the deep
wells.
3
ICSGE-13
GEO-010-3
the modeled area. Table 1 exhibits the hydraulic parameters adopted in the current study
for different formations. The hydraulic conductivity of the confined aquifer, bottom
sand layer, was assessed based on the results of an in-situ pumping test.
Table 1: Hydraulic parameters of the different soil layers
Parameters
Clay layer
Sand aquifer
110-9
-0.55
0.65
2.510-4
0.001
0.35
0.45
ICSGE-13
GEO-010-4
obtained at the piezometer location changed from 6.2m to 4.5m due to change of the
hydraulic conductivity from 2.5x10-4 m/s to 4x10-4 m/s, respectively. This reflects the
importance of proper assessment of the aquifer hydraulic conductivity. In other words,
for appropriate design of groundwater lowering systems, it is crucial to asses the
aquifers hydraulic conductivity from in-situ pumping tests.
Fig. 2 Contour lines of the numerically predicted drawdown values of the bottom
sand aquifer, all dimensions are in meter.
ICSGE-13
GEO-010-5
0.0003
0.0004
0.0005
Drawdown (m)
10
Fig. 3 Variation of the groundwater drawdown within the bottom sand aquifer with
change of its hydraulic conductivity
Hydraulic Conductivity Ratio (Kh/Kv)
0
Drawdown (m)
10
Dewatering systems are conventionally designed using the analytical equations of the
groundwater flow because of the simplicity of such equations and the ease of their
implementation in design spread sheets. For a multiple wells system, the analytical
equation is used together with the superposition/cumulative drawdown concept. In this
ICSGE-13
GEO-010-6
concept it is assumed that the total drawdown at any point will be the sum of the
drawdown that would have been caused by each well if it were operating alone.
The dewatering system under consideration was analyzed using the analytical approach.
Since the excavation and dewatering activities resulted in making the bottom aquifer
semi/partially confined, the drawdown due to each well was estimated by Eq. (1) that is
used for partially confined aquifer [6].
(1)
Where;(H-h)
H
D
Q
K
R
The influence radius was assessed from the analysis of the pumping test results, 290m.
Equation (1) delineates that the drawdown is dependent on the considered aquifer
thickness. Figure (5) shows the variation of the analytically estimated drawdown at the
piezometer location with the aquifer thickness. The figure shows that the drawdown is
increased with decrease of the aquifer thickness that is referred to the point that Eq. (1)
was developed based on the assumption that the pumping well is fully penetrating the
aquifer. Considering the difficulty of assessing the value of the aquifer thickness
especially for semi confined conditions, it is recommended to depend on the numerical
methods in analyzing/designing dewatering systems especially for partially confined
aquifers and partially penetrating wells.
Aquifer Thickness(m)
40
50
60
70
80
Drawdown (m)
2
4
6
8
10
12
14
Fig. 5 Effect of the aquifer thickness on the analytically predicted drawdown of the
semi-confined aquifer at the piezometer location.
ICSGE-13
GEO-010-7
REFERENCES
[1]
[2]
[3]
Muskat, M. The flow of Homogenous Fluids through Porous Media, (Mc GrawHill, New York 1937).
[4]
[5]
[6]
[7]
ICSGE-13
GEO-010-8
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
Harbaugh, A. W. and McDonald, M. G. User's Documentation for MODFLOW96, an Update to the U.S. Geological Survey Modular Finite Difference
Groundwater Flow Model, U.S. Geological Survey Open-File Report, (1996)
96-486.
[17]
[18]
[19]
[20]
[21]
[22]
ICSGE-13
GEO-010-9
13th ICSGE
Faculty of Engineering
Cairo - Egypt
S. A. Mazek 2
M. K. El Ghamrawy 3
ABSTRACT
Development of tunneling is an effective tool to overcome high-density population
challenges such as transportation and utilities activities. Tunnels are used for several
purposes such as construction of sewers, railway, and roadway tunnels. The Greater
Cairo Metro tunnel, El-Azhar road tunnels, and the sewage tunnel in the Greater Cairo
city (CWO) are considered as major projects to solve transportation and environmental
problems. There are technologies to assist in excavation such as the tunnel boring
machine (TBM) and the cut and cover method. To better understand the performance of
the tunnel system, it is necessary to investigate the geotechnical problems related to
tunneling.
In the present study, the finite element model (FEM) is proposed to predict the
performance of the tunnel system under the shadow of three case histories. The three
case studies presented and discussed in this study give good and rare opportunity to
understand the performance of the tunnel system. The constitutive model for this
analysis utilizes elasto-plastic materials. A yielding function of the Mohr-Coulomb
type and a plastic potential function of the Drucker-Prager type are employed. A linear
constitutive model is employed to represent the tunnel liner.
The response of the tunnel system is described by the ground movement caused by
tunneling. The ground movement is presented and calculated using the proposed 3-D
FEM. The recorded measurements are also presented for the three case histories. The
results obtained by the 3-D numerical model are compared with those obtained by the
field measurement to assess the accuracy of the proposed 3-D numerical model. There
is a good agreement between the results obtained by the 3-D FEM and those obtained
by the field recording.
KEYWORDS: tunnels, numerical modelling and analysis, finite element analysis,
deformations.
1
2
3
Ph.D. Student, Lecturer, Civil Engineering Department, Military Technical College, Cairo, Egypt.
Ph.D. Lecturer, Civil Engineering Department, Military Technical College, Cairo, Egypt
Professor, Civil Engineering Department, El-Azhar University, Cairo, Egypt
ICSGE-13
GEO-011-1
1. INTRODUCTION
Tunneling leads to ground movement due to the associated stress change around an
advancing tunnel. The numerical techniques are widely used to predict the ground
movements [1]. Finite element method is considered the most appropriate analytical
technique to solve geotechnical problems [2], [3], [4], [5], [6], [7], [8]. Modeling of
geotechnical properties and tunneling procedure is the sophisticated problem [9], [10].
In this study, the Greater Cairo Metro tunnel, El- Azhar road tunnels, and the Greater
Cairo Wastewater tunnel (CWO) in central Cairo city are considered as three case
studies. The tunnel system performance is studied. The tunnel system model is proposed
using 3-D nonlinear finite element analysis (FEA) under the shadow of the three case
histories. The aim of this study is to understand the performance of the tunnel systems.
In the 3-D nonlinear finite element modeling, the tunneling process and the interaction
effects between the tunnel and the soil around the tunnel are investigated. The 3-D
nonlinear finite element analysis is used to estimate the vertical displacement at ground
surface due to tunneling.
In this Study, the Greater Cairo Metro tunnel, El Azhar road tunnels, and the Greater
Cairo wastewater tunnel are discussed through a comparison between the results
calculated by the 3-D nonlinear finite element analysis and the results recorded by field
measurements to assess the accuracy of the proposed 3-D finite element analysis (FEA).
The typical geotechnical and soil properties used in this study are presented in Table 1.
However, the results calculated by the finite element analysis agree well with those
obtained by the field measurements.
Table 2: Geotechnical properties
Soil parameter
Fill
Silty clay
Sand
b (t/m 3 )
1.8
0.58
0.4
1.9
0.8
0.35
2.0
0.37
0.30
25
26
37
1.0
300
0.74
0.0 to 4.0
350
0.6
4.0 to 10.0
400-600
0.5-0.6
10.0 to end
ko
(Degree)
C (t/m2)
m
n
Depth (m)
ICSGE-13
GEO-011-2
modeling the soil media and the thick shell elements for modeling the tunnel liner. The
thick shell element models both membrane (in plane) and bending (out plane) behavior
of the tunnel structure. The solid element is chosen since it posses in-plane and out-ofplane stiffness. The solid element allows for both in-plane and out-of-plane loads. The
solid element is prismatic in shape. The prismatic solid element and the triangular shell
element interface are used between the soil media and the tunnel liner to ensure the
compatibility conditions at the interface between them as well as the associated stress
and strains along the interface surface.
The vertical boundaries of the 3-D finite elements model are restrained by roller
supports to prevent a movement normal to the boundaries. The horizontal plane at the
bottom of the mesh represented a rigid bedrock layer and the movement at this plane is
restrained in all three directions. The movement at the upper horizontal plane is free to
simulate a free ground surface, as shown in Fig. 1. In the finite element analysis, the
loading attributed to the construction process is considered.
ICSGE-13
GEO-011-3
pa
(1)
In which, the modulus number (m) and the exponent number (n) are both pure numbers
and ( a ) is the value of the atmospheric pressure expressed in appropriate units. The
nonlinear soil parameters used for elasto-plastic finite element analysis for different
types of the soil are presented in Table 1 [3], [14].
The ground water table varies between two meters to four meters from the ground
surface. Figure 2 shows the soil profile along the Greater Cairo City. The ground
surface displacement due to the construction of the tunnel is calculated in this study.
The final diameter (D) for the metro tunnel is 8.98 m, and the excavation diameter of
the metro tunnel is 9.48 m. The final diameter (D) for the road tunnel is 9.06 m and the
excavation diameter of the metro tunnel is 9.56 m. The final diameter (D) for the
sewage tunnel is 4.65 m and the excavation diameter of the sewage tunnel is 5.15 m.
The tunnel lining is assumed to behave in a linear manner in the 3-D nonlinear finite
element analysis. The characteristic of the tunnel lining is tabulated in Table 2.
Table 2: Characteristics of the metro, sewage, and road tunnel
Tunnel
Eb (t/m 2 )
Sewage
2.1 10 6
Metro
Road
f c (t/m 2 )
(t) cm
0.18
4000
25
2.1 10 6
0.18
4000
25
2.1 10 6
0.18
4000
25
The 3-D finite element mesh used in the analysis models a soil block with width, height, and
depth in x, y, and z directions, respectively, as shown in Fig. 1. Studies have been adopted for
the metro tunnel located at 18 meter depth from the ground surface, as shown in Fig. 2. Drain
analysis is adopted in the numerical model, where the tunnels pass through the sandy soil
layer. The ground loss of 4 % is adopted in this study. The ground loss will be discussed in
detail in the following section. The suitable geometric boundaries (model width and
model height) are studied to reflect the accurate performance of the tunnel system. By
studying the stresses at the geometric boundaries, the study is conducted to determine
the optimum model width beyond which no changes in stresses are occurred.
ICSGE-13
GEO-011-4
The 3-D finite element analysis is used to choose the suitable width so as to realistically
reflect the behaviour of the tunnel system. The height of the numerical model is some
where the soil depth beneath the invert the tunnel is 2.5 times diameter of the tunnel
[16]. The studied model width varies from 40 meters to 120 meters. The calculated
settlements at surface of the metro tunnel with different model width are shown in Fig.
3. As the model width exceeds 100 meters there is no change in the estimated surface
settlement. The calculated settlement at crown of metro tunnel with different model
width tunnel also is presented in Fig. 4. The results also show that the model width
exceeds 100 meters there is hardly any change in the estimated crown settlement of the
tunnel.
10.5
10
Optimum
model
width
8.5
8
40
50
60
70
80
90
100
110
120
130
ICSGE-13
GEO-011-5
13.75
23
-123-
21
Optimum
model
width
19
17
15
40
50
60
70
80
90
100
110
120
130
13.5
Optimum model
width
12.75
30
40
50
60
70
80
The calculated invert heaves of the metro tunnel with different model width are also
presented in Fig 4. The results also show when the model width exceeds 100 meters
there is no change in the estimated invert heave. Hence, the one-hundred-meter model
width is chosen to realistically reflect the performance of the tunnel system.
The one-hundred-meter model width and the fifty-meter model height are used in the
analysis to compare the results obtained by the 3-D finite element analysis and the field
measurement.
For the El-Azhar road tunnel (second case study), similar analysis is also carried out to
help choose the appropriate geometric boundaries. The results also show that the model
width exceeds 100 meters there is hardly any change in the estimated surface settlement
of the road tunnel. The one-hundred-meter model width and the fifty-meter model
height are used in the analysis to compare the results obtained by the 3-D finite element
analysis and the field measurement.
For the sewage tunnel (third case study), similar analysis is also carried out to help
choose the appropriate geometric boundaries. The model width varies from 40 meters to
100 meters. As the model width exceeds 80 meters there is no change in the estimated
surface settlement. The eighty-meter model width is adopted in the 3-D finite element
analysis to compare the results obtained by finite element analysis and the field
measurement.
4.2 Mesh size
The 3-D finite element mesh used in the analysis models a soil block with width, length,
and depth in x, y, and z directions, respectively, as shown in Fig. 1. The suitable mesh
size will be selected to reflect the accurate performance of the tunnel system. The
element size is varied (2 m, 3 m, 4 m, and 5 m) along the outer boundary of the soil
block. The element size is also varied (1 m, 1.5 m, 2 m, 2.5 m, and 3 m) along tunnel
liner. The calculated settlement of ground surface considering different element sizes is
presented in Table 3. The results show that for the soil element size being smaller than 3
m, there is hardly any change in the calculated surface settlement of the tunnel. The
one-meter element size of the tunnel liner is chosen for the rest of the 3-D finite element
analyses.
ICSGE-13
GEO-011-6
For the mesh size, similar analysis is also conducted for the El-Azhar road tunnel and
the sewage tunnel to choose suitable element mesh size. A three-meter element size of
the soil block and the one-meter element size of the tunnel liner are chosen for the rest
of the 3-D finite element analyses.
Table 3: Estimated settlement of ground surface considering different elements sizes
(Greater Cairo Metro)
The construction of a tunnel leads to a subsurface movement due to the moved soil towards
the excavated tunnel. The volume loss VL is the ratio of the difference between volume of
excavated soil and tunnel volume over the excavate soil volume. The ground loss ranged from
1.5 % to 4.5 % and reached to 6 % at some location along the Greater Cairo metro project
[17]. The ground loss impact on the behavior of the tunnel system is investigated to assess the
accuracy of the 3-D nonlinear finite element. Surface settlement due to the construction of the
metro tunnel is calculated by the 3-D finite element model. The suitable ground loss will be
studied to reflect the accurate performance of the tunnel system.
In the parametric study, the ground loss various from 2% to 6% to choose the suitable
ground loss adopted in finite element analysis due to tunnelling. The ground loss of 4%
is chosen to realistically reflect the performance of the tunnel system. The calculated
surface settlements against different ground loss for the metro tunnel are analyzed and
presented in Fig. 5. The calculated surface settlement due to tunneling is affected by the
ground loss impact. The results show that the increase of the ground loss due to
tunneling leads to increase the maximum surface settlements.
For the El-Azhar road tunnel, similar analysis is also carried out to choose the suitable
ground loss. The ground loss of 4% is also chosen to realistically reflect the
performance of the tunnel system.
For the third case study (sewage tunnel), the ground loss of 6% is chosen to realistically
reflect the performance of the tunnel system.
ICSGE-13
GEO-011-7
13
12
11
10
9.4
9
8
7
6
5
4
1%
2%
3%
4%
5%
6%
7%
5. SOIL BEHAVIOR
The stress changes in surrounding soils due to tunneling are investigated to study the
detailed soil behavior around the tunnel systems. The stress changes in soil around the
tunnel system due to tunneling are investigated to study the detailed soil behavior. The
stresses in the subsoil have undergone three phases of change. At these phases, the
loading steps of the tunnel construction are simulated using the 3-D finite element
analysis. First, the initial principal stresses are computed with the absence of the tunnel.
Second, the excavation of the tunnel is modeled by means of the finite element method.
The excavation is simulated by the removal of those elements inside the boundary of the
metro tunnel surface to be exposed by the excavation. The excavated tunnel boundary is
free to move until the soil comes into contact with the tunnel liner resulting in ground
loss. The ground loss is considered in this study. The ground loss of 4 % is adopted in
this study. Third, the calculated changes in stresses are then added to the initial
principal stresses computed from the first phase to determine the final principal stresses
resulting from the tunnel construction. The initial in-situ stresses of the excavated tunnel
boundary before tunneling are calculated and plotted in Fig. 6. The final vertical stress
change after tunneling is calculated and compared with the initial in-situ vertical stress
before tunneling, as shown in Fig 6. The results show that the soil above the crown of
the metro tunnel settles downward and the soil under the invert of the metro tunnel
excavation heaves.
The performance of the three case studies is modeled using the 3-D nonlinear finite
element elements analysis. In this study, the Greater Cairo Metro tunnel, El-Azhar road
tunnel, and the Greater Cairo wastewater tunnel are studied through a comparison
ICSGE-13
GEO-011-8
between the results calculated by the 3-D nonlinear model and the results recorded by
field measurements.
This case study is located along the Greater Cairo Metro Line 2, as shown in Figure 2.
The nonlinear 3-D finite element model is used to predict the performance of the metro
tunnel. The computed surface settlement is compared with those obtained by the field
measurements so as to understand the behavior of the metro tunnel, as shown in Fig. 7.
This comparison is used to assess the accuracy of the proposed numerical model,. The
comparison shows that there is good agreement between the computed and measured
readings.
6.2 El-Azhar road tunnel (Second case history)
For the El-Azhar road tunnel project, similar analysis is also carried out to predict the
performance of the road tunnels. The 3-D nonlinear numerical model is also used to predict
soil-tunnel interaction of the El-Azhar road tunnel to assess the accuracy for the numerical
model. The comparison between the calculated results using the numerical model and the
measured values are presented in Fig. 8. The study shows that the results calculated by the
numerical model have a good agreement with the field data.
ICSGE-13
GEO-011-9
-40
-30
-20
-10
0
-2 0
-4
-6
-8
-10
-12
10
20
30
40
50
Metro
Tunnel
Fig. 7: Comparison between measured and calculated surface settlement due to the
construction of the Greater Cairo metro tunnel
-40
-30
-20
-10
0
-2 0
-4
-6
-8
-10
-12
10
30
40
50
13.5 mm
Road
Tunnel
20
Road
Tunnel
Measured settlement
Fig. 8: Comparison between measured and calculated surface settlement due to the
construction of El-Azhar road tunnels
6.3 Greater Cairo wastewater project (third case history)
For the Greater Cairo wastewater tunnel project, the nonlinear 3-D finite element model
is also used to predict the performance of the sewage tunnel. The computed surface
settlement obtained by the finite element analysis is compared with those obtained by
the field measurements so as to understand the behavior of the sewage tunnel, as shown
in Fig.9. The comparison shows that there is good agreement between the computed and
measured readings.
Generally, the calculated surface settlement due to the tunnel construction
underestimates by up to 10% with respect to the field measurement for the three case
studies. This discrepancy between calculated and measured readings may be caused by
the accuracy of soil strength parameters, soil stress parameters, soil modelling, or
instrumentation.
ICSGE-13
GEO-011-10
-30
-20
-10
0
-1 0
-2
-3
-4
-5
-6
-7
10
20
30
40
Sweage
Tunnel
7. CONCLUSIONS
A 3-D nonlinear finite element analysis is used to understand the performance of the
tunnel system under shadow of three case studies. The analysis takes into account the
changes in stress, the non-linear behavior of the soil, and the construction progress, etc.
The following conclusions can be drawn regarding the performance of the tunnel under
the effects of different factors.
1- A 3-D nonlinear numerical model is applicable to analyze and predict the detailed
performance of the tunnel systems under the shadow of the three case histories.
2- The results calculated by the proposed 3-D nonlinear numerical model have a good
agreement with the field data. The predicted surface settlements underestimate by up 10
% for all case histories with respect to the field measurement.
3- The minimum width of the 3-D nonlinear model is set to be ten times the tunnel
diameter in the 3-D numerical model.
4- Ground loss is an important parameter effect on the performance of the tunnel
system.
8. REFERENCES
[1]
Berant, S., and cambou, B. 1998. Soil-Structure Interaction in Shield tunneling in soft
ground. Computers and Geotechnics. Vol. 22. No. 34. pp. 221-242.
[2]
Karakus, M., and Fowell, R.J. 2005. Back analysis for tunneling induced ground
movement and stress redistribution. Tunneling and Underground Space
Technology 20, pp. 514 524.
ICSGE-13
GEO-011-11
[3]
Mazek, S.A., Belal, A.M., Kamal, I.M., and Amer, N.H. 2006. Numerical
Modeling of Soil-Tunnel Interaction under new surface load. Procp of the 6th
ICCAE conf16-18, May, pp. 248 257.
[4]
[5]
[6]
[7]
El-Nahas, F., El-Kadi, F., and Shalaby, A. 1990. Field measurements during
construction of compressed air tunnel in Cairo. Proc. International and Future,
Chengdu, China.
[8]
Addenbrooke, T. I., and Potts, D. M. 2001. Twin tunnel interaction surface and
subsurface effects. The International Journal of Geomechanics. Vol. 1, No2, pp. 249
271.
[9]
[10]
[11]
[12]
[13]
Mazek, S.A., Law, K.T., and Lau, D.T. 2001b. 3-D Analysis on the Performance
of a Grouted Tunnel. Proceeding of Canadian Geotechnical Conference. An
Earth odyssey 2001. Calgary, Alberta, Canada, pp. 111-118.
[14]
National Authority for Tunnels (NAT), 1993 and 1999, Project Documents.
[15]
[16]
Mazek, S.A, and Tehawy, E.M. 2008. Impact of Tunneling Running Side-bySide to An Existing Tunnel on Tunnel Performance using Non-linear Analysis.
Proc of the 7th ICCAE. 27-29 May, 2008. Cairo, Egypt
[17]
ICSGE-13
GEO-011-12
13th ICSGE
Faculty of Engineering
Cairo - Egypt
ABSTRACT
In order to achieve accurate pore water pressure measurements and to gain some understanding
of the fundamental behaviour of pore of the saturated clay under slow monotonic loading two
series of tests were performed. The work reported in this paper is about the behaviour of clay
under monotonic triaxial loading. 102mm high x102mm diameter samples of remoulded
undrained Cowden clay.
Depending upon the drainage conditions, there will usually be either a volume decrease
or a generation of excess pore pressure if the volume changes are delayed or prevented,
due to long drainage paths or if the loading imposes high shear stresses or high strain
rate.
Compared to the monotonic loading, under earthquake loading the drainage paths for
excess pore pressure dissipation in clay of a deep excavation are comparatively long.
Considerable excess pore water pressure may develop as a result. Accordingly, it is
reasonable and conservative to assume completely undrained conditions
Although accurate pore water pressure measurements are of paramount importance for a
better understanding of the behaviour of clay under slow monotonic, few studies of clay
ICSGE-13
GEO-012-1
soils have included accurate measurement of pore water pressure, because of the low
permeability of clays and the corresponding long response time for pore water pressure
measurement.
KEYWORDS
Pore pressure, rate effect, clay, monotonic loading, stress, compression.
INTRODUCTION
The aim of the present work is to assess the behaviour of pore water pressure under
triaxial compression loading.
Compared to the monotonic loading, under earthquake loading the drainage paths for
excess pore pressure dissipation in clay of a deep excavation are comparatively long.
Considerable excess pore water pressure may develop as a result. Accordingly, it is
reasonable and conservative to assume completely undrained conditions.
Anderson et al A series of undrained strain-controlled monotonic triaxial tests carried
out on an undisturbed plastic marine clay (Drammen clay) [1] at a strain rate of 0.05
%/min. They suggested that the difference in undrained strengths may be the result of
the anisotropic preconsolidation and the different stress paths.
Craig [2], conducted a series of undrained triaxial compression tests on saturated
remoulded clay.
He observed a linear relation between strength and the logarithm of time to failure
(between 0.002 and 270 minutes) and that a ten-fold decrease in time to failure caused a
% increase in shear strength. He also reported that, for strain-controlled tests, the
increase in shear strength could be as much as 10%, or higher.
While it is evident that the most obvious undrained shearing resistance may be obtain
from tests on identically prepared samples at different controlled strain rates, the time
requirement associated with slow rates of testing is of a major inconvenience. To
overcome this problem, two other methods of testing can be introduced.
A step-changing procedure: Richardson and Whitman [3]introduced this technique in
which the strain rate applied to a sample is step-changed during a test. Each strain rate
would be applied only long enough to establish the stress-strain relationship for that
stage of the test. remoulded samples do not agree precisely with results from constant
rate tests.
B Kenny ,as reported by Graham [4] developed the step-changing and relaxation
method suitable for the determination of the strain rate effect from a single sample. The
difference between the two methods is that the sample tested with the second method
would go through a period a relaxation after each strain rate application.
Graham et al [4] used the step-changing and relaxation method to perform triaxial
compression and extension tests as well as simple shear tests on slightly
overconsolidated clays.
They reported a linear relation between strength and logarithm of rate. They also
reported that undrained shear strengths decrease with decreasing logarithm of strain
rate. Crawford [5] reported that there was a significant influence of the strain rate on
pore water pressure at failure, and that the high pore pressure in a slow test may be due
to a secondary consolidation effect because more time was allowed for structural
breakdown . He added that a number of tests in which the deviator stress was held
constant for several hours in which the strain was held constant were carried out. He
ICSGE-13
GEO-012-2
reported that under constant deviator stress the pore pressure continued to rise, whereas
under a constant strain condition, the pore pressure decreased. He concluded that pore
pressures are dependent on strain and may be considerably influenced by sample
disturbance and isotropic conditions.
Discussing the data reported by M.I.T and Harvard University, Whitman [4] suggested
that as rate increases, gradients of pore water pressure will be set up within the
undrained triaxial samples.
Thus, the pore water pressure within the central zone may be time dependent.
He also suggested that the mineral may, under rapidly applied loads, have a resistance to
compression approaching that of water. Such behaviour, he concluded, would be the
result of structural viscosity. If so, he continued, then the excess pore water pressures
set up during undrained shear in soft saturated soils would decrease.
2 EXPERIMENTAL APPARATUS AND MATERIAL
The main laboratory equipments used in this research consisted of a conventional
102mm diameter sample triaxial cell, a conventional triaxial loading machine.
The mid-height pore water pressure measurements were made by means of a miniature
pressure transducer [6]. Cowden clay was used in this investigation.
The index properties, as reported in reference [7], are as follows: Liquid limit=44%;
Plastic= 19%; Plasticity index=25%; Clay fraction(D<0.002mm)=32%; Activity.78; CV
of remoulded soil (V = 480kN/m2); mv=1.5m2/yr.
One final consolidation pressure of 600kN/m2 was used for preparing remoulded
samples of Cowden. The preparation procedure was as described below.
The vertical consolidation pressure was applied by means of hydraulic pressure
The cell allowed top and bottom drainage.
3
ICSGE-13
GEO-012-3
Figures.1 shows combined plots of UM, UB and D (deviator stress) versus axial strain
(A) for the tests Comp.1, whichn is similar to tests Comp.2, Comp.3, Comp.4.
I t can be deduced from plots of tests Comp 2 to Comp.4 that U, the magnitude of the
difference between mid-height and base pore pressure generation is larger at small
strains (A < 3%) than at large strains.
This is believed to be due to the rapid rate of change of the deviator stress at the start of
the test. However, in Figure1, for test Comp.1, large values of U are observed
throughout the test. This is considered to be erroneous and may be due to the low B
values associated with this test. It can be observed that, while tests Cornp.2, Comp.3
and Comp.4 achieved a fully saturated condition (BM = BB = 1), test Comp.1 achieved
BM = 0.93 and BB = 0.9. Bearing in mind the high compressibility of the soil skeleton,
these values probably indicate a significantly large loss of saturation.
A combined plot of excess mid-pressure(IJM) versus axial strain (A) for the whole set
is performed. Curves for tests Comp.1 and Comp.2 peak at about. 1% axial strain,
whereas curves for whereas curves for2% and 2.8% axial strain respectively.
It can also be seen that tests Comp.1 and Comp. 2 had higher peak IJM than tests
Comp. 3 and Comp. Also indicated by an arrow for each test is the excess pore pressure
at maximum deviator stress. It. is seen that tests Comp. I, Comp. 2 developed higher
excess pore pressure at maximum deviator stress than test Comp. 3.
However, test Comp.4, although run at the fastest strain rate of 0.3%/min shows higher
excess pore pressure at maximum deviator stress than test Comp.3.
All compression tests presented in Figure 2 show a dilatant behaviour which is a feature
associated with an overconsolidation state.
The author believes that, due to the difference in the initial effective stress (o) from
one sample to another, resulting from the sample preparation process, a better
comparison of pore water pressure behaviour could be obtained by plotting UM/0
versus axial strain (A), as shown in Figure 3 (test Comp.1 has not been included in
Figure 3 because of its low B values.) It is seen that all the curves show similar peak
values of UM/o.However, when the final pore water pressure is expressed in terms of
ICSGE-13
GEO-012-4
a ratio (UMC/O'), it is clear that, both in compression and tension increasing strain rate
results in an increasing dilation.
The pore water pressure behaviour discussed clearly suggest that samples used for
monotonic testing were lightly over-consolidated. This is due to the sampling process
(no further reconsolidation in the triaxial cell was carried out).
The over-consolidation ratios (OCRs) for the monotonic tests were computed by
dividing the maximum effective stress in the dometer Pmax, assuming k0= 1/2 for
normally consolidated clays, by the initial effective stress in the triaxial cell.
The author believes that at strain rates of 0.03 and 0.3%/min correct failure pore
pressures were still measured. If these are higher than the true values due to pore water
pressure gradients, then the effective stress paths would be closer to the origin than they
should be.
4
CONCLUSIONS
Under monotonic (compression and tension) loading the normalised shear strength
(Dmax/0) at failure increased with increasing strain rate.
The effective failure envelope inferred from the slowest monotonic compression tests
had C=0; = 23.4 in compression.
The limited amount of data indicate that increasing strain rate may affect C and .
Accurate mid-height pore water pressures relevant to failure conditions were
consistently measured.
Expressed in terms of the initial effective stress, the excess pore pressure at failure
decreased with increasing strain rate. In terms of effective stress, the increase in shear
strength with increasing strain rate may be partly explained by the reduction of pore
pressure at failure and partly by strain rate effects on C and parameters.
5
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
ICSGE-13
GEO-012-5
13th ICSGE
Faculty of Engineering
Cairo - Egypt
ABSTRACT
In this paper, a three-dimensional numerical approach, based on the combination of
Boundary Element Method and Thin Layer Method, is used to study the the dynamic
interaction effects between two surface ajacent foundations. The foundations are placed
on the surface of a homogeneous soil layer limited by a substratum. Two modes of
vibrations are considered in the study: horizontal and rocking modes. The approach
consists of an appropriate discretisation of the foundation-soil-foundation domain, and
has the advantage to transform the force displacement relationship from an integral
equation formulation to a linear system of equations, for which the displacements are
expressed as unknowns. The foundation-soil-foundation interaction effects are
formulated in the study through the coupling compliance matrices of foundations. This
approach has been applied to characterize the dynamic interaction effects. Obtained
results show the importance of these effects on the response of each foundation
especially when the two foundations are very close .
KEYWORDS
Discretisation, foundation, interaction, load, soil, response, vibration.
ICSGE-13
GEO-013-1
INTRODUCTION
ICSGE-13
GEO-013-2
Foundation 1
Foundation 2
My
Px
l
x
z
H
Soil layer
,,,G
u = U .t .dS
Where;U
u
t
(1)
As long as the model is continuous, equation (1) remains difficult to evaluate. However,
if the model is appropriately discretised, this equation may be transformed to an
algebraic system of linear equations with displacements as unknowns.
2.2 Discrete model
ICSGE-13
GEO-013-3
2b
2l
Thin layer
In the first discretisation, the soil layer is subdivided in the vertical direction into a
number of thin layers. The thickness is taken very small, in such manner that the
displacement could be assumed to vary linearly within each thin layer, with sufficient
accuracy. This discretisation provide the Greens functions of the discrete soil
medium according to the work of Kausel et al [14].
For the second discretisation, only the foundation interfaces are dicretised in horizontal
direction into a number of square elements with small sides compared to the
dimensions of the foundation. Under this condition, the distribution of forces on the
elements could be assumed uniform, without significant errors. The average
displacement of the element is approximate to the displacement of the centre.Within the
discrete model, equation (1) is transformed to a linear system of equations where
unknowns are the displacements:
Nt
u =
=1
Where; Nt
U .t .dS
(2)
The flexibility matrix for the discrete medium, could be assembled from the influence
coefficients, calculated from equation (2) by sucessive application of unit distributed
load on all elements. This matrix is written in a form that can separate the mutual
effetcts between foundations 1 and 2.
F
F
[Ft ] = 11 12
(3)
F21 F22
Where; Fii (2Nix2Ni) flexibility matrix of the soil foundation interface, i (i =1,2)
F12 (2N1 x 2N2) , F21 (2N2 x 2N1) coupling matrices for medium 1 and 2
number of discrete elements of soil foundation interface 1 and 2
N1 , N2
Clearly, the coupling matrices F12 and F21. disappear if the interaction between the
foundation is not taken into account.
The same separate form is used in the expression of the force displacement
relationship:
ICSGE-13
GEO-013-4
(4-a)
(4-b)
Where;{t1} (2N1 x 2), {t2} (2N2 x 2) applied load vectors, for medium 1 and 2
{d1} (2N1 x 2), {d2} (2N2 x 2) corespoding displacement vectors
In order to account for the interaction effects, compatibility condition between the soil
and foundation (rigid body motion) must be imposed to the elements at the interface.
This condition is simply formulated as:
{d1} = [R1 ].{D1}
(5-a)
{d 2 } = [R2 ].{D2 }
(5-b)
(6-b)
Where; {Pi} = {Pxi , Myi}t external load vectors for the foundations i
Pxi horizontal force
Myi rocking moment
2.3 Numerical Response of Foundations
From minor transformations, relations (4), (5), and (6), give the relation between the
external loads and the rigid body displacements:
{P1} = [K11 ].{D1} + [K12 ].{D2 }
(7-a)
(7-b)
ICSGE-13
GEO-013-5
(9-a)
(9-b)
(10-b)
RESULTS
The present results are compared against those obtained by Karabalis et al.[14], by
using the 3-D frequency domain BEM formulation.
The comparison concerns the case of a square foundations of unit side (b =1/2), loaded
by a unit horizontal force Px = 1 and placed at the distance ratio d/b = 2 from a free
foundation having the same size. The foundation are resting on the surface of a semiinfinite soil. The dimensionless compliance obtained by the present numerical study are
represented on Figure 3 as functions of the dimensionless frequency a0 = b/2Cs . The
curves show a good agreement between present results and those reported in previous
reference .
3.2 Application
This application analyses the dynamic interaction effects for various ditances between
the two surface foundations. The first foundation is loaded by unit horizontal force Px =
1 and unit moment My = 1, while the second is free . The soil layer is characterized by:
= 1, G = 1 , = 0.333 and = 0.05. To minimize the reflecting effects from the
substratum, the depth ratio was fixed to H/b = 8 which corresponds to a relatively
deep substratum. The soil layer properties are = 0.333 and = 0.05, with unit values
for and G.
ICSGE-13
GEO-013-6
c 11; c 12
0,25
0,21
0,17
0,13
0,09
0,05
0,01
-0,03
ao
0,14
c11
d/b=2
d/b=4
d/b=6
d/b=8
0,12
0,1
0,08
0,06
0,04
0,02
0
0
Fig.4
ICSGE-13
ao
Variation of the dynamic response of loaded foundation for various distance ratio
GEO-013-7
The numerical responses of the two foundations are calculated from relations (9), using
the dimensionless compliances functions c11 and c12 defined in the relation (10).The
dynamic responses are reported on Figure 4 (loaded foundation), and Fig. (5)
(unloaded foundation). The results are presented as function of dimensionless
frequency a0 = b/2Cs, for various values of the distance ratio (d/b = 2, 4, 6, 8).
d/b=2
0,08
c12
d/b=4
0,06
0,04
0,02
0
-0,02
-0,04
0
Fig.5
ao
Variation of the dynamic response of free foundation for various distance ratio
CONCLUSION
In this paper, the Boundary Element Method was combined with Thin Layer Method
to study the effects of foundation-soil-foundation interaction on the dynamic response
of ajacent foundations. In this approach, only the soil-foundations interfaces and a
limited domain of the soil were apropriately discretised. The numerical responses
derived from the discrete model were presented in a form of impedance or compliance
matrix, where the coupling terms are expressing the dynamic interaction effects
between the two foundations. This study illustrates the significant effect of the
foundation-soil-foundation interaction on the response of foundations. To this end, it is
recommended to take into account this phenomenon for the study of special
structure.The analysis of numerical responses of the two foundations, calculated at
different distance from the excited foundation showed that the unloaded foundation is
much more sensitive to the effects of dynamic interaction. These effects are amplified
when the two foundations are close, and become insignificant for large distances. In
contrast, for loaded foundation, these effects have less influence on the dynamic
response.
ICSGE-13
GEO-013-8
ACKNOWLEDGMENT
The authors are grateful to the Ministry of higher education and scientific research for
supporting this work under grant NJ0400920070143.
7
NOTATIONS
C
F
H
G
K
My
Px
compliance matrix
flexibility matrix
depth [m]
shear modulus [kN/cm]
impedance matrix
rocking moment [kN-m]
horizontal[kN]
a0
b
d
dimensionless frequency
half width of foundation [m]
distance between foundation [m]
damping coefficient
angular frequency [sec-1]
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
ICSGE-13
GEO-013-9
13th ICSGE
Faculty of Engineering
Cairo - Egypt
ABSTRACT
Constructing a soft ground tunnel changes the state of stresses in the soil (or soft rock)
mass resulting in associated strain and deformation fields. The main geotechnical
challenge during tunnelling is to keep the deformations of existing nearby structures
within acceptable limits. The use of bentonite slurry Tunnel Boring Machines (TBMs)
has contributed to achieving this task during advancing urban tunnels through waterbearing sand deposits. Several designs of these machines have been utilized to advance
several kilometres of relatively large size urban tunnels under the Greater Cairo, Egypt
for the Cairo Wastewater Spine Tunnel, Lines 2 and 3 of Cairo Metro Tunnels as well
as Al-Azhar Road Tunnels. Numerical modelling using different 2D and 3D finite
element techniques has been successfully used to represent almost all the construction
details and to predict the overall deformation field around such tunnels. In particular,
the maximum settlement directly above the tunnel centreline can be accurately
calculated during the design stages using the available numerical techniques. However,
prediction of the associated differential settlement, along the settlement trough at/or
near the ground surface, has been less accurate and tends to be on the unsafe side.
This paper focuses on the potential of using a constitutive law utilizing the small strain
stiffness to model the unique deformation field around such tunnels in order to predict a
more reliable settlement trough that can be used to assess the structural safety of
existing adjacent buildings and other structures.
KEYWORDS
Tunnelling, Water-Bearing Sand, Bentonite Slurry TBM, Settlement Trough, Finite
Element, Constitutive Law, Small Strain Stiffness
ICSGE-13
GEO-014-1
INTRODUCTION
ICSGE-13
GEO-014-2
(1)
In which:
E50ref : deformation modulus for primary loading (Reference value)
m:
c:
:
ICSGE-13
GEO-014-3
It was found that soil stiffness detected by shear wave velocity (Vs) measurements can
be applied to the initial static monotonic loading, as well as the dynamic loading of soil
(Burland, 1989). Therefore, the dynamic shear modulus (Gdyn) could be seen as the
maximum shear modulus, designated Gmax or Go, that provides upper limit stiffness.
This is a fundamental stiffness and can be measured in all soil types and fractured rocks.
The corresponding equivalent elastic modulus is determined from: Eo = 2Go (1+)
where is the corresponding value of Poissons ratio of soil at small strains.
Figure 2 shows a typical relationship between the soil stiffness (G/Go) and the soil strain
in a semi-log scale according to Atkinson and Sallfors (1991). A characteristic S-shaped
stiffness reduction curve can be detected for most soil types. It can also be recognized
that soil stiffness corresponding to strain levels which can be reliably measured in
classical laboratory tests, i.e. triaxial tests and oedometer tests is often decreased to less
than half its initial value at small strain level.
ICSGE-13
GEO-014-4
ICSGE-13
GEO-014-5
ICSGE-13
GEO-014-6
/
cu
c
ko
[-]
[kN/m3]
[kN/m2]
[]
[-]
0.3
18/8
1
30
0.5
0.4
20/10
80
0.8
0.3
19/9
1
33
0.46
10
15
30 + 2 z
Very dense
sand
50
0.3
19/9
1
35
0.43
Very dense
sand
50 + 4 z
E50ref
[MN/m2]
2
Clay-Silt
Medium
to dense sand
Very dense
sand
10
15
30
50
[MN/m ]
30
45
90
150
m
Rf
/
cu
c
Ko
[-]
[-]
[kN/m3]
0.6
0.9
18/8
1
30
0.5
0.8
0.9
20/10
80
0.8
0.5
0.9
19/9
1
33
0.46
0.5
0.9
19/9
1
35
0.43
ref
ur
[kN/m2]
[]
[-]
Medium
Very dense sand
to dense sand
0.7
[-]
2 x 10-4
2 x 10-4
2 x 10-4
2 x 10-4
Go
[MN/m2]
1.25 x 105
1.875 x 105
3.75 x 105
6.25 x 105
Where:
E
Elastic modulus,
Poissons ratio
ref
E50 Primary loading stiffness
ICSGE-13
GEO-014-7
ur
Ko
m
Rf
c
/
Go
0.7
1/1
-20
-10
Distance (m)
10
20
30
Settlement (cm)
-4
0
1/2
12
16
1/1
1/2
2
3
Measurements
Mohr Coulomb Model (MC) with constant E with depth for each layers
Mohr Coulomb Model (MC) with increased E with depth for sand layers
Double Hardening Model (HSM)
Small Strain Double Hardening Model (SHSM)
ICSGE-13
GEO-014-8
independent and the soil behaves as a linear elastic material till reaching the failure
conditions according to Mohr-Coulomb. These restrictions in the model make it
unsuitable to determine the complex behavior of soil due to tunnel activity. In case of
increasing E-Modulus with depth (Model. 1/2), the settlement trough seems to be more
realistic. Nevertheless, the results are very sensitive depending on the initial value of the
deformation modulus at the layer surface and its increase value with depth.
On the contrary, the double hardening model gives realistic results concerning the
settlement trough of soil surface. Nevertheless, the calculated settlement trough is flatter
than the measured values. Therefore, the differential settlement and hence the distortion
angle between two points on the settlement trough is underestimated applying the
double hardening model. The double hardening model with small strain leads to more
accurate estimation of the full shape of settlement trough. It should be noted that the
shape of settlement trough is a decisive design criterion for tunnelling in urban areas.
Figure 7 shows the trajectories of the principal stresses using the double hardening
model with small strain in comparison with only double hardening model. The arching
action above the tunnel crown is more pronounced in case of double hardening with
small strain than in case of conventional double hardening model.
b- Double hardening
Fig. 7: Principal stress trajectories in case of double hardening model with and
without small strain
In case of Mohr-Coulomb with increasing E-Modulus with depth (MC 1/2), double
hardening (HSM) and double hardening with small strain (SHSM) models, the straining
actions have the same shape with different values. Table 2 summarizes the values of
maximum straining actions for the three models.
Table 2: Tunnel lining straining actions
Straining action
MC Model 1/2
ICSGE-13
GEO-014-9
HSM
SHSM
199.1
1160
83.3
110
941
46.3
It can be recognized that there is a big difference in the lining straining actions
calculated by different constitutive laws. As the Mohr-Coulomb model gives small
deformations it is expected that the earth pressure acting on the tunnel lining applying
the MC Model is too high and hence the straining actions are overestimated. The
difference between the HSM and SHSM is mainly due to the strong arching in the
SHSM (figure 7). It is apparent that this point needs still more research to identify the
most realistic model regarding the lining straining actions. Nevertheless, it is
recommended to apply the traditional HSM to determine the straining actions.
3.3 Parametric study
The Mohr-Coulomb Model requires simple and conventional soil parameters (EModulus, c and ). These parameters can be determined for most soil types and stated
in conventional geotechnical reports. In the case of double hardening model, additional
parameters are required (e.g. E50ref , Eurref and m). Nevertheless, there is a sound existing
experience with such parameters that can also be determined applying conventional
laboratory tests such as triaxial test. In contrary, the parameters applying the double
hardening model with small strain (e.g. Go and 0.7) require special laboratory tests and
need certain experience with the model. Therefore, a parametric study showing the
sensitivity of these parameters on the results is given in the following paragraphs.
Effect of initial shear modulus at small strain
Several researches (Burland, 1989; Tatsuoka & Shibuya, 1992; LoPresti et al., 1993)
showed that the small-strain stiffness from shear wave velocity (Vs) measurements
applies to the initial static monotonic loading, as well as the dynamic loading of
geomaterials. Thus, the dynamic shear modulus (Gdyn) has been renamed the maximum
shear modulus (Go). The corresponding equivalent elastic modulus is found from:
Eo = 2 Go (1+) where = 0.2 is the approximate value of Poissons ratio at small
strains. There is different published correlation between soil type and the ratio of Eo/Estat
(e.g. Alpan 1970 and German recommendations for soil dynamic). Figure 8 shows the
effect of the chosen Eo related to Eur on the results of the calculated settlement trough. A
reasonable range of stiffness ratio Eo /Estat has a minor effect on the settlement trough.
ICSGE-13
GEO-014-10
ICSGE-13
GEO-014-11
Fig. 11: Effect of coefficient of earth pressure at rest of the Clay-Silt layer
ICSGE-13
GEO-014-12
4. CONCLUSIONS
The shape of the settlement trough is a decisive design criterion for tunnelling in urban
areas. Bentonite slurry TBMs have been used to control the ground deformations during
advancing urban tunnels through water-bearing sand deposits. Estimating of the
maximum settlement as well as the expected differential settlement within near-by
structures are the most important factors for choosing the required additional measures
for structural safety of these structures.
Advanced numerical models with enhanced constitutive laws describing the stress-strain
behavior of soil mass are becoming a main part of each tunnel design. Choosing the
appropriate constitutive law is a basic important design step. It is illustrated in the
presented case history that the double hardening model can be considered as a realistic
model dealing with both straining action of tunnel lining as well as prediction of
settlement trough. Nevertheless, this model underestimates the predicted differential
settlements (distortion angle) at or near the ground surface. On the other hand, the
double hardening model with small strain can give more realistic results concerning the
settlement trough. Nevertheless, there is still a need for more experience with this model
especially regarding the straining action of the lining. This point needs more research.
The elastic perfect plastic (Mohr-coulomb) model with constant E-Modulus should not
be applied dealing with problems where the soil deformation parameters are highly
dependent on the stress path and stress level such as due to stress release (e.g. tunnelling
and deep pit excavation). In such cases, this simple model leads to completely
unrealistic deformations. An increasing E-Modulus with depth can enhance the
performance of this model. Nevertheless, it should be emphasized that the simple
elastic-perfectly plastic model can work relatively good in other cases such as shallow
foundations, single pile analyses or estimating bearing capacity (ultimate conditions) or
stability analyses of slopes. In most these cases, a realistic estimation of the governed
elastic modulus and shear strength parameters is essential to get realistic results.
The conducted analyses show that the most effective way to control settlement trough
due to tunnel activity is to reduce the induced ground loss to minimum values, for
instance; using bentonite slurry TBMs. Special additional measures, such as grouting
through the TBM tail, can be also very effective in reducing the undesired detrimental
effect of settlement on neighboring structures.
5. REFERENCES
Abdel Salam, M. E. (1984) "Grouting of alluvial soil in Cairo underground metro".
Proc. of Conf. on Tunnelling in Soft and Water Bearing Ground, A. A. Balkema,
pp. 135-142.
Ahmed, A. A., Ali, H., El-Sayed, S., and Nour, S. (2003) "Neuronal prediction of
tunnelling induced settlements. " Proc. of the Tenth Int. Colloquium on
Structural and Geotechnical Eng., Ain Shams Univ., Cairo, Paper No. GE37.
Atkinson, J.H. and Sallfors, G. (1991) "Experimental determination of soil properties."
Proc. 10th ECSMFE, Florence, Vol. 3, pp. 915-956
Alpan, I. (1970) "The geotechnical properties of soils." Earth-Science Reviews 6, 5-49.
Brinkgrave, R.B.J. (2002). "Plaxis manual." Balkema, Rotterdam
Burland, J.B. (1989). "Small is beautiful: The stiffness of soils at small strains."
Canadian Geotechnical Journal 26 (4), 499-516.
ICSGE-13
GEO-014-13
Chanz, T., Vermeer, P.A. and Bonnier P.G. (1999) "Formulation and verification of the
hardening soil model." Proceeding of the International Symposium Beyond 2000
in Computational Geotechnics: Beyond 2000 in Computational Geotechnics 10 Years of PLAXIS International, Balkema, Rotterdam.
Duncan, J.M. and Chang, Y. (1970). "Nonlinear analysis of stress and strain in soils."
ASCE, Vol. 96, SM, pp. 1629-1654.
Eisenstein, Z. (1986). "The contribution of numerical analysis to the design of shallow
tunnels. " Proc. of Conf. on Geotechnical Modeling in Engineering Practice, pp.
135-165.
El-Mossallamy, Y.; Stahlmann, J. (1999) "Monitoring and numerical analysis of tunnels
in complex geological conditions." Proceeding of the International Symposium
Beyond 2000 in Computational Geotechnics: Beyond 2000 in Computational
Geotechnics - 10 Years of PLAXIS International, Balkema, Rotterdam.
El-Nahhas, F. (1993) "Tunnelling under Cairo using bentonite slurry machines. " Proc.
of Int. Congress on Options for Tunnelling. Amsterdam, ITA, Elsevier Science
Publishers B. V., pp. 351-361.
El-Nahhas, F.; Ahmed, A.; and Esmail, K. (1997) "Prediction of ground subsidence
above tunnels in Cairo. " Proc. of Fourteenth International Conf. on Soil
Mechanics and Foundation Eng., Hamburg, Germany, A. A. Balkema, Vol. 3,
pp. 1453- 1456.
El-Nahhas, F. (1999) "Soft ground tunnelling in Egypt: geotechnical challenges and
expectations. " Journal of Tunnelling and Underground Space Technology,
Pergamon Press, Vol. 14, No. 3, 245-256.
El-Nahhas, F. (2003) "Geotechnical aspects of controlling groundwater levels in urban
areas. " A Keynote Paper, Proc. of the Tenth Int. Colloquium on Structural and
Geotechnical Eng., Ain Shams Univ., Cairo, Vol. 6, KL4.
El-Nahhas, F. M. (2006) "Tunnelling and supported deep excavations in the Greater
Cairo. " Int. Symposium on Utilization of Underground Space in Urban Areas,
ETS & ITA, Sharm El-Sheikh, pp. 27-56.
El-Sayed, S. (2001) "Elasto-plastic three-dimensional analysis of shielded tunnels. "
Ph.D. Thesis, Geotechnical and Foundation Engineering, Ain Shams University,
Cairo, Egypt.
Esmail, K. (1997) "Numerical modeling of deformations around closed face tunnelling
boring machines." Ph. D. Thesis, Geotechnical and Foundation Engineering, Ain
Shams University, Cairo, Egypt.
Ezzeldine, O. (1995) "Design of tunnels constructed using pressurized shield methods. "
Ph. D. Thesis, Geotechnical Engineering, University of Alberta, Canada.
Fahey, M. and Carter, J. (1993). "A finite element study of the pressuremeter using a
nonlinear elastic plastic model". Canadian Geotechnical Journal 30 (2), 348-362.
German recommendations for soil dynamic (2001) "Empfehlung des Arbeitskreises 1.4
Baugrunddynamik der Deutschen Gesellschaft fr Geotechnik e.V. (DGGT)"
Ghaboussi, J., Ranken, R. E., and Karshenas, M. (1978) "Analysis of subsidence over
soft-ground tunnels. " Proc. of International Conf. on Evaluation and Prediction
of Subsidence, ASCE, Florida, pp. 182-196.
Hardin, B.O. and Drnevich, V.P. (1972). "Shear modulus and damping in soils. "
Journal of the Soil Mechanics & Foundations Div. (ASCE) 98 (7), 667-692.
ICSGE-13
GEO-014-14
LoPresti, D.C.F., Pallara, O., Lancellotta, R., Armandi, M., and Maniscalco, R. (1993).
"Monotonic and cyclic loading behavior of two sands at small strains." ASTM
Geotechnical Testing Journal 16 (4), 409-424.
Mansour, M. (1996) "Three dimensional numerical modeling of hydroshield tunnelling.
" Ph. D. Thesis, University of Innsbruck, Austria.
Santos, J.A. and. Correia, A.G (2001) "Reference threshold shear strain of soils.
Application to obtain a unique strain-dependent shear modulus curve for soils".
Proc. 15th SMGE, Istanbul, Vol.1, pp. 267-270.
Stokoe, K.H. and Santamarina, J.C. (2000) "Seismic-wave-based testing in geotechnical
engineering" GeoEng 2000: Vol.1, pp. 1490-1536.
Tatsuoka, F. and Shibuya, S. (1992). "Deformation characteristics of soils and rocks
from field and laboratory tests. " Report of the Institute of Industrial Science,
Vol. 37, No.1, Serial No. 235, The University of Tokyo.
Vucetic, M., Hsu, C.C. and Doroudian, M. (1998) "Results of cyclic and dynamic
simple shear tests on soils from La Cienega Site conducted for ROSRINE
project and other research purposes". UCLA, Research Report No. ENG-98-200,
Civil and Environmental Engineering Department, University of California, Los
Angeles.
Ward, W. H. and Pender, M. J. (1981) "Tunnelling in soft ground." General Report.
Proc. of Tenth International Conf. on Soil Mechanics and Foundation Eng.,
Stockholm, pp. 261-275.
ICSGE-13
GEO-014-15
13th ICSGE
Faculty of Engineering
Cairo - Egypt
ABSTRACT
The performance of a pavement material beyond the elastic limit has substantial
influence on the bearing capacity of a pavement and its serviceability. It is therefore,
important to know whether a pavement structure will experience progressive
accumulation of permanent deformation leading to a state of incremental collapse, or if
this accumulation will cease resulting in a stable and elastic response. Although
experimental data and practical experience have been introduced into the design method
through empirical parameters, there is not a simple relationship between the elastic
strain and the long-term plastic behaviour of pavement materials. This paper describes a
procedure for constructing a mathematical model to predict the long-term behaviour of
pavement structures under repeated loads imposed by vehicle loads. It also enables
characterisation of asphalt pavement deterioration using Mohr-Coulomb based
shakedown. The model differs from extant models, in that the cyclic nature of the
loading on a pavement is recognised from the outset, and the current method, which is
based upon foundation analysis is replaced by a procedure employing shakedown
theory. The basic concepts are described, and results of studies of failure by rut
formation are presented. Implications for design procedures are also discussed.
KEYWORDS
Shakedown theory, cyclic loading, pavements design, granular bases, shear strength, pavement
rutting.
INTRODUCTION
Pavements, i.e. roads and runways, are amongst the most difficult geotechnical
structures to analyse and design. Unlike foundations and slopes, pavements have to be
designed against failure resulting from a large number of load applications, occurring
over long time periods, instead of just a single monotonic load application. Although
current design procedures exhibit some elements of rational analysis, they are virtually
all based on empirical relationships between elastic performance and plastic failure. A
ICSGE-13
GEO-015-1
strain criterion limiting the vertical elastic subgrade strain is used to limit the risk of
rutting within pavement layers. Base-course/subbase performance is controlled with
material specifications such as grading limits. Moreover, these methods do not consider
the plastic deformation performance of the base-course and subgrade layers. Hence, the
predicted life in terms of ESAs (Equivalent Standard Axles) using these procedures can
sometimes indicate a long pavement life. However, the observed performance (rutting,
surface integrity and roughness etc) can indicate much shorter lives. The complex
nature of the problem clearly requires the development of a mechanistic approach to
flexible pavement design, which incorporates the plastic as well as elastic properties of
the pavement layers, provides prediction whether a given pavement structure will
experience progressive accumulation of permanent deformation and consequently
gradual damage, or the increase of permanent deformation will cease, thereby leading to
a stable and resilient response. This behaviour is referred to as shakedown behaviour.
Field observations indicate that many pavements do in fact shakedown rather than
deform continuously.
Over the last two decades, there has been growing interest in the application of the
fundamental shakedown theorems of Bleich [1], Melan [2], and Koiter [3] to describe
the behaviour of materials/structures under repeated cyclic loads. Initial applications
were to frame structures and pressure vessels. Then later have been extended to cover a
broad category of applications that account for the effects of high temperature, strain or
work hardening, nonlinear geometry e.g. Johnson [4]; Knig [5], dynamic behaviour,
and non-associated plastic flow. More recently applied to model wear resistance of
metal surfaces to repeated rolling and sliding loads, Ponter et al. [6]. Using finite
element analysis made the shakedown theory a useful tool in the design of a wide range
of practical applications such as railway foundations, pavement engineering, roller
bearings, and the machining industry. Thus, it enables the long-term behaviour of
complex structures to be assessed without resorting to computationally expensive stepby-step analyses.
The application of shakedown theory to the geomaterials that usually make up the
pavement layers requires consideration of a general cohesive-frictional yield criterion.
This feature significantly complicates the determination of the shakedown limit, both in
numerical and analytical computations. In the past, a number of attempts of both types
of calculations have been made. Firstly, in their pioneering paper, Sharp and Booker [7]
introduced the so-called method of conics, which can be seen as an application of
Melans lower bound theorem, and solved a number of two-dimensional problems
where the load was approximated in terms of a trapezoidal pressure distribution. This
method was later given a kinematic interpretation by Collins and Cliffe [8] and the
resulting generalized upper bound method was applied to a number of more realistic
three-dimensional problems by Collins and Boulbibane [9] and Boulbibane et al. [10].
In this paper we discuss a simplified approach that has the potential of predicting the
load-induced deformation, e.g. rutting and surface cracking, of flexible pavements. The
method differs from extant models, in that the cyclic nature of the loading on a
pavement is recognised from the outset, and the current method, which is based upon
foundation analysis is replaced by a procedure employing shakedown theory. In essence
this method seeks the mechanism, from within a class of mechanisms that minimises the
shakedown limit load. A simplified illustrative example is given to show how the
procedure might be utilised. Extensive computer analyses based on the elastic layer
ICSGE-13
GEO-015-2
program BISAR developed by De Jong et al. [11] and global optimisation, the
Simulated Annealing procedure devised by Goffe et al. [12], are conducted to simulate
rutting in the asphalt surfaces, base-courses, and subgrades. Also investigations into the
relative structural contributions of wearing course and base course to the load carrying
capacity of the pavement structure, when used at different layer thicknesses/or different
materials, are studied and implications for design procedures are discussed.
2
10 6
N f = 1.077 10
v
18
4.4843
(1)
Where N f is the number of cycles to failure. The above equation defines failure as
12.5mm (0.5-inch) depressions in the wheel-paths of the pavement. For a non standard
axle P, the number of the repetitions of the equivalent design axle can be calculated
from the fourth power law which was deduced from the AASHTO road test:
P
N f =
Ps
(2)
In which N is the number of the repetitions of load P, and Ps is the standard axle load. In
this section, equation (1) will be used as the design criterion for flexible pavement. By
ICSGE-13
GEO-015-3
substituting equation (2) into (1), the following relationship between the load P and the
critical strain for a given pavement thickness profile is obtained:
v = 0.0105 N 0.223
(3)
P 0.892 v = 0.0129
(4)
X
Original
surface
11
Direction
of travel
21
31
Asphalt
Base-course
Asphalt
Layer
Subgrade
deformation
Subgrade
In order to utilise the results from the shakedown analysis in the practical pavement
engineering, a rutting development method based on the shakedown analysis is
discussed here. This method takes into account the plastic behaviour of the pavement
layers instead of just using a purely elastic model. It is mainly based on assuming a
certain class of failure mechanisms. Apart from the material deterioration in the
pavement, the structure hardening due to residual stress in the pavement is considered
through the shakedown limit loads. Early research in this area [9], concentrated on
failure in which subsurface or surface slip occurred in the travel direction, paralleling
the metal surfaces wear study of Ponter et al. [6]. More recently calculations have been
made for the complementary mode of failure in unbound pavement structures, where the
subgrade motion is perpendicular to the travel direction and the loaded tyre penetrates
into the pavement so forming a rut [10].
According to the classical theorem given by Koiter [3], the kinematic method provides
an upper bound of the true limit load inducing collapse. This upper bound is obtained by
assuming a failure mechanism and calculating the load factor, which is defined to be the
ratio of the rate of plastic dissipation to the virtual rate of dissipation of the elastic stress
field in the proposed failure mechanism. Illustration in Figure 2 depicts the crosssection of the road as well as the assumed model for rut failure mechanism:
p
ij
e ijp dV
e
ij
e ijp dV
(5)
ICSGE-13
GEO-015-4
where e ijp and ijp are the time-independent strain rate and plastic stress fields associated
with the assumed kinematically admissible velocity field and ije is the elastic stress
field produced by the applied load P. The shakedown load is then definitely less than or
equal to P . It is similar to the classic technique of limit analysis, except that the
calculation involves knowledge of the elastic stress distribution. In the calculations
described here we choose mechanisms in which the plastic dissipation is confined to
discrete slip lines. Thus, following Chen [16], for a Coulomb material, the two
dissipation functions in equation (5) can be evaluated from:
c [ v ti ] l i
i
[v ti ] c ie d l i
(6)
where c is the cohesion, and [ v ti ] and l i are, respectively, the jump in tangential
velocity and length of the ith discontinuity, whilst c ie is termed the elastic cohesion that
can be evaluated on this discontinuity by:
c e = et en tan
(7)
It is not appropriate here to describe in detail the procedures used for calculating
shakedown load factors. These can be found in the references cited above. In essence
the estimates are obtained by postulating a kinematically admissible velocity field -i.e.
failure mechanism- and equating the elastic and plastic rates of working in this
mechanism. In this investigation, we are mostly concerned with rut failure in three
layered pavement structures as illustrated in Figure 2. It is to be emphasised that this
model only predicts the onset of the rutting and does not attempt to model the complete
development of the rut. This failure mode has one or more defining angles i , j (i, j
=1,2,3 where i and j represent variable and layer number, respectively) and depth D.
The discontinuity lines must have a change in slope as they cross the asphalt/basecourse
and base-course/subgrade boundaries whenever the value of the internal friction angle
j is different in the two layers:
i , j j = i , j+1 j+1
(8)
These parameters ( i , j , D) can be varied, together with the co-ordinate x 0 defining the
position of the (y, z) plane of deformation and the ratio in equation (6) optimised to find
the minimum value of associated with given family of mechanisms. The resulting
value of the load factor is then an upper bound to the actual shakedown load, so that if a
number of possible mechanisms are studied, that producing the smallest load estimate is
the most likely to occur. The minimum value for each class of mechanism is found by
numerical optimization. The selected mechanism shown above is translated into a
nonlinear optimization problem and the simulated annealing has been successfully
applied to determine the optimum mechanism, which is described by the defining
parameters. In this method the ratio of plastic rate work to elastic rate work is the
function to be minimized and the control parameter in the algorithm, which is initially
set to high value decreased in steps until the optimal solution is reached.
ICSGE-13
GEO-015-5
APPLICATIONS
Layer
Asphalt
Concrete
Granular Base
type I
Granular Base
type II
Soft Subgrade
(fine grained)
Stiff Subgrade
(fine grained)
Friction
angle,
[]
17
Cohesion,
c
[kPa]
4,500
45
20
45
60
0.45
Layer
thickness, h
[mm]
50, 100, 150,
200
150, 250, 400,
500
150, 250, 400,
500
Infinite
20
0.45
Infinite
80
Elastic
modulus
[MPa]
4,000
Poissons
ratio
0.35
150
0.35
250
0.35
20
80
Initial results suggest that shakedown limits are influenced by the geometry of the
assumed failure mechanism. A comparison of the critical shakedown limit pressures
predicted by different rut mechanisms is given in Table 2.
Table 2: Shakedown limits for three-layered pavement structures [MPa]
Number of parameters
1, j = 2, j = 3, j and x 0
Type of
Subgrade
Soft
1, j 2, j 3, j , D and x 0
Soft
2.06
2.28
2.66
2.98
3.13
1, j = 2, j = 3, j and x 0
Stiff
3.51
3.69
3.98
4.26
4.38
1, j 2, j 3, j , D and x 0
Stiff
2.91
3.07
3.37
3.65
3.79
ICSGE-13
GEO-015-6
0.05
0.1
0.15
0.2
Asphalt
0.10
0.15
Base course
0.20
0.25
0.30
0.35
Subgrade
To get an appropriate picture of the size of influence of layer thickness and material
properties, some calculations of response are made and presented in graphs below. For a
given layered pavement (see Fig. 2) with known material parameters, the shakedown
limits are determined from shakedown analysis and compared against the design traffic
loads. The design traffic load obviously has to be less than the shakedown load to
ensure that accumulation of permanent strains will not occur. In the case the design
traffic load is much smaller than the shakedown limit load, the design of the pavement
is conservative and the adjustment of the base-course thickness should be made.
The variations of v with h, the depth of the base-course according to equation (1), can
be found from the program BISAR with dual load 40 KN giving a non linear loaddepth relation as is shown in Figure 4 for both soft and stiff subgrades. It can be seen
that pavement loading associated with rutting of the subgrade, is lower or higher than
the shakedown limit, depending on applied load repetitions and pavement component
characteristics. It should be emphasised though that for larger load repetitions used in
conventional mechanistic pavement design (generally > 106), it seems that the
shakedown load is essentially higher than the applied design load. From these figures, it
is also seen that as the depth increases, the shakedown load approaches a finite limiting
value while the design load increases indefinitely. This means that there is a critical
thickness of the base-course layer for which the shakedown load would become more
critical than the design load as calculated by the existing design method.
3.2 Parametric study on shakedown of layered pavements
A part from the geometry of the layered pavement, the mechanical parameters of the
materials in pavements, like stiffness and strength parameters (see Table 1), also
influence the shakedown behaviour of pavements.
ICSGE-13
GEO-015-7
6
5
N=102
103
Shakedown Limit
103
Shakedown Limit
104
105
104
105
10
106
0
0
100
200
300
400
500
100
300
400
500
10
5
4
Shakedown Limit
104
105
106
N=102
10
12
N=102
200
N=102
103
6
Shakedown Limit
104
105
106
0
0
100
200
300
400
500
100
200
300
400
500
Fig. 4: Dependence of the axle load on base-course thickness for given load repetitions
For a granular material, these parameters depend on a variety of state variables like
moisture content and temperature in a non-linear manner. Moreover, in the history of
the pavement lifetime, the material parameters could change with time due to
environment changes. The dependency of the shakedown load on these material
parameters is very complicated and it will be very difficult to give a complete analysis.
However, some aspects of this dependency can be predicted from the present theory
based on some basic assumptions.
Results shown in Figure 5 indicate that increasing the thickness of the asphalt concrete
surface could have a significant effect on the increase in shakedown resistance of the
pavement. This increase is more for pavements with stiff subgrade compared to soft
subgrade. The shakedown resistance also increases with increased thickness of the basecourse as illustrated in Figure 6. However, this increase is less significant than that
observed for the asphalt concrete surface. Also improving the base properties in terms
of modulus and shear strength i.e. cohesion and/or friction angle, will result in improved
shakedown resistance as shown in Figure 4. A similar tendency in term of relationships
between shakedown and the thickness of the asphalt and/or the granular base-course can
be observed from Raad and Minassian [19] research for similar pavement structures,
although the shakedown limits are slightly different.
ICSGE-13
GEO-015-8
16
14
12
10
Stiff subgrade
8
6
4
3.5
Stiff
subgrade
This study
Soft subgrade
Soft
subgrade
3
2.5
2
Raad & Minassian [19 ]
This study
1.5
0
0
50
100
150
200
250
100
150
200
250
300
350
In the above calculations the shakedown load is calculated assuming a uniform load
distribution as an approximation to the tire load on the pavement. In reality, this is not
exactly true as the applied load on the pavement is distributed in a very complicated
way, which depends on a number of factors such as the speed of the vehicle, the
pressure of the tires and the flexibility of the pavement. Here, an effort is made to
simulate more realistic load distributions using a superposition of 10 uniform load
distributions. Figure 7 shows geometrically the superposition of uniform loads with
different radii to approximate both Hertzian and trapezoidal load distributions. In the
latter, the pressure distribution takes the form of a cone frustum with upper radius b and
lower radius a=150mm. The dependence of the shakedown limit load on the assumed
shape of the loaded distribution is illustrated in Figure 8. It is seen that the shakedown
limit decreases when the uniform pressure is replaced by graded pressure.
Hertzian load
3.5
Trapezoidal load
Pb
3
2.5
2
Uniform
1.5
Trapezoidal b/a=0.2
Hertzian
0.5
0
100
200
300
400
500
ICSGE-13
GEO-015-9
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
Ponter, A.R., Hearle, A.D. and Johnson, K.L. Journal Mechanics Physics Solids,
33 (1985) 339-362.
[7]
Sharp, R.W. and Booker, J.R. Journal of Transport Eng. 110, (1984) 1-14.
[8]
Collins, I.F. and Cliffe, P.F. J. Num. and Anal. M. Geomech. 11 (1987) 409-420.
[9]
Collins, I.F. and Boulbibane, M. J. of Geotech. and Geo. Eng. 126 (2000) 50-59.
[10]
Boulbibane, M., Collins, I.F., Ponter, A.R.S. and Weichert, D. Road Materials
and Pavement Design 6 (1) (2005) 81-96.
[11]
De Jong, D.L., Peutz, M.G.F. and Korswagen, A.R. Computer Program Bisar.
Shell, Amsterdam (1979).
[12]
Goffe, B., Ferri, G.D. and Rogers, J. Journal of Econometrics, 60 (1994) 65-99.
[13]
HMSO, Design Manual for Roads and Bridges, Vol. 7, HD 25/94, Part 2 (1994).
[14]
[15]
[16]
Chen, W.F. Limit analysis and soil plasticity. Elsevier, Amsterdam (1975).
[17]
[18]
Bolton, M.D. and Lau, C.K. Canadian Geotech. Journal 30 (1993) 1024-1033.
[19]
Raad, L. and Minassian, G. Road Materials & Pavement Design 6 (2005) 53-79.
ICSGE-13
GEO-015-10
13th ICSGE
Faculty of Engineering
Cairo - Egypt
ABSTRACT
The pile working load depends on the imperfections which may be taken place in pilesoil system, during pile construction, among many other factors. This subject attracted
the researcher's attention world wide in the last decades. Types of imperfections either
geotechnical or structural are documented in literature and well explained. Nevertheless,
the influence of these imperfections in pile load calculations still ambiguous. The work
presented herein is devoted to study soil disturbance during construction of piles using
continuous flight auger, CFA. The study of soil disturbance due to drilling needs some
evidence. The source of this evidence is field observations collected from four different
construction sites, which are documented in this paper. The study concluded that the
disturbed zone of soil by CFA has a conical shape and extending laterally to a distance
equivalent to ten times of the pile diameter around the auger at the cutting bits and has
an inclined surface of 4:1 (vertical : horizontal). Furthermore excess pore water pressure
was induced in soil in the vicinity of pile drilling. Due to this excess pore water
pressure, 3.5% to 6.5 % of piles constructed by CFA showed percolation of water from
the top of the piles through fresh concrete. Also, subsidence of fresh concrete in pile
hole was recorded in few of the constructed piles. Pile loading tests showed that the
percolation of water and/or subsidence of fresh concrete have not appreciable influence
on the load-displacement characteristics of the piles. Moreover, percolation of water at
pile heads.
KEYWORDS
Pore water pressure, pile drilling, imperfections, CFA, and water percolation.
ICSGE-13
GEO-016-1
INTRODUCTION
The process of drilling vertical holes to form cast in-situ bored piles affects the
properties of soil around the piles. These effects are difficult to quantify their influence
on the preformed piles. During boring of a hole, horizontal geostatic stresses at the
exposed surface of the drilled shaft decreases to almost zero, and soil relax towards the
drilling hole. Soil relaxation is followed by water migration towards the surface of the
hole, so softening zone starts to occur. A similar phenomenon occurs at the base of the
shaft, where the vertical geostatic stresses are reduced to about zero. The extent of the
relaxation and softening zone depends on the properties of soil and the length of time
that hole is open before concrete is poured. On the other side, during boring a hole using
continuous flight auger, the auger retained part of the excavated material and the other
part transport to the ground surface. At the same time, narrow gap is formed between
the outer periphery of the auger and the sides of the hole. This gap is performed due to
the relatively flexibility of the auger. The auger flexibility depends upon the auger
length, auger diameter, and thickness of the auger flights. Hence the auger rotates
around a vertical axis passing through the centre of the rotary head and at the same time
around an imaginary axis away form the centre of the rotary head. Keeping in mind that
there is slackness between the auger and the guide fixed in drilling rig. Furthermore,
during drilling the pile, the drilling rig is not stationary in place and may exhibits side
sway. The value of side sway depends upon the auger stiffness and the height of drilling
rig. The gap between the auger and the hole is filled with excavated geomaterial.
Consequently the horizontal stresses acting on sides of the hole is dropped from a
geostatic value which equivalent to (K0. -v) to relatively small value. While the base of
the hole formed by drilling is not stress free, contrary to the bored piles. But once the
auger with retained soil is raised up before pouring the concrete to open the bottom tap
of the stem, unbalanced stresses are formed at the hole base. Therefore the drilling
process relieve the geostatic stresses acting on the hole sides and on the hole base.
Abdrabbo [1 & 2] and Polous [3] summarized the sources of pile imperfections during
construction of bored piles. Kenny et al [4] reported soil disturbance during CFA piling
in sand. Carrier [5] illustrated that during construction of continuous flight auger pile, a
single stream of clear water was frequently observed percolating from the top of the pile
through the fresh cement grout as well as one to three streams through the surface of the
surrounding ground. Carrier [5] attributed the percolation of water due to liquefaction
induced by the auger in a zone of loose to very loose sand. Carrier [5] showed that the
reported observation takes place during drilling in sand with average SPT below count
of 5 with about one third of the values equal to 1 or 2.
The work presented herein sheds some light on the phenomenon of water percolation
from the top of the piles through fresh concrete. The paper attempts to discuss the
reasons of the phenomenon and to investigate the frequency of occurring this
phenomenon during construction of CFA piles. Moreover the paper aims to study the
effect of this phenomenon on the pile working load. The subject is considered one of the
most interesting topics in geotechnical engineering, especially in piled foundations.
2
It was observed that at many construction sites, on which continuous flight auger is
implemented in pile construction and after filling the pile holes with fresh concrete, a
stream of water percolate from the top of the piles through fresh concrete. The water
ICSGE-13
GEO-016-2
percolation is continued up to six to eight hours. Another observation was also noted at
some construction sites that a big amount of water suddenly flood to the ground surface
with the excavated material. Furthermore, the transported sand to the ground surface
appears in semi-liquid state. Subsidence of fresh concrete takes place through a pile has
been recently constructed, while auguring a pile hole nearby. Sometimes subsidence of
fresh concrete takes place when topping the pile hole with concrete and without any
drilling nearby. Finally subsidence of ground surface takes place around the drilling
hole. It is worth noting that, the above observations are not noted for all piles at a site.
But some of these observations are noted in some piles at a site and the other piles at the
same site do not exhibit any of theses observations. However all these field observations
confirm that, during drilling the pile hole in sand, pore water pressure increases. The
increase of pore water pressure causes hydraulic gradient towards the pile hole followed
by flow of water, with or without sand particles, up to ground surface through fresh
concrete along the centreline of the pile. The observation was noted in case of ground
water table up to 1.5 m below the ground surface. It was noticed that subsidence of
concrete takes place immediately after filling the pile hole with fresh concrete. Of
course, topping of pile is necessary with more fresh concrete. This subsidence occurred
whether water percolation was noticed in the fresh concrete or not. One meter of
concrete subsidence was recorded, which is corresponding to 7% of the theoretical pile
volume. Also it was noticed that during drilling a pile hole, at a distance equal to ten
times the pile diameter, away from fresh concrete pile which has been formed a
subsidence of 1.2 m of concrete takes place in that pile which corresponding to 8.6% of
the theoretical volume of the pile. It was noticed that percolation of water from top of
piles through fresh concrete never observed in all piles constructed at a site, but this
phenomenon was noticed in 4.3 % to 6.5 % of the constructed piles. From these
observations, it can be concluded that the effects of drilling process on soil properties
extend to more than 10 times of the pile diameter. The subsidence of concrete can be
attributed to volumetric strain takes place in soil as a result of dissipation of excess pore
water pressure. Carrier [5] illustrated that an average grout subsidence of 1.1 m which
corresponds to 8.5% of the theoretical pile volume was noticed. The above observations
are resulted from the effect of drilling process on the properties of soil nearby. The
following sections may explain this mechanism.
During auguring in a soil mass to form a pile hole, cyclic shear strains are imposed on
soil around the auger. Nearly free stress surface is expected to be formed at auger-soil
interface. The amount of excess pore water pressure induced in soil mass around the
drilling hole depends upon many factors such as penetration ratio (L/NP), in-situ density
of sand, stratification of soil which have to be drilled, borehole diameter, depth of
borehole, and normalized shear stress ration (/ -0)/(/ -0)L, fig. (1-a). Where L = depth
of penetration, N = number of turns and P = pitch between flights. Fig. (1-a) illustrates
that the constrained modulus of soil increases from a reduced value at the hole side to
the unaffected value at the outer boundary of the disturbed zone. The excess pore water
pressure developed in the disturbed zone has the profile of bell shape with a small
values at the hole sides and at the outer boundary of the disturbed soil zone. So it is
anticipated that dissipation of excess pore water pressure takes place, horizontally, in
two directions towards the hole and away from the hole. The figure demonstrates
schematic drawing for tangential cyclic strain. The tangential cyclic strain has a
maximum value at pile hole and decreases with distance away from the hole. The
ICSGE-13
GEO-016-3
resulted cyclic radial stress has an opposite trend, it is increased with distance away
from the hole. The build-up of permanent volume strain is related to relative density of
sand, number of cycles, and the cyclic shear strain; Silver & Seed [6]. Ishihara et al [7]
and Martin et al [8] produced fundamental models for pore water calculations due to
cyclic loads. While Silver & Seed [6] and Dobry et al [9] confirmed that 0.01% cyclic
shear strain produce inappreciable volume change and hence no significant increase in
pore water pressure. Even though liquefaction may not have occurred during drilling
process, the dissipation of pore water pressure results volumetric change in soil. In this
conditions, pore water pressure generated may be expressed in terms of the normalized
shear stress ratio, (/ -0)/ (/ -0)L. As this ratio increases the pore pressure ratio (U/ -0)
increased reaching 1.0 when the normalized shear stress ratio (/ -0)/ (/ -0)L becomes
1.0. Tokimatsu & Yoshimi [10] showed that excess pore pressure is generated when
(/ -0)/(/ -0)L is 0.50. Tokimatsu & Seed [11] point out that volumetric strain
expressed as percentage attained 0.2 just before liquefaction. The excess pore water
pressure at hole-soil interface has a residual small value, due to the expulsion of water
into the hole, fig. (1-a).
For non liquefaction conditions, the amount of reconsolidation volumetric strain
increases with increasing grain size of soil, decreasing relative density, and increasing
excess pore pressure generated during the undrained cyclic loading. But for drilling
numerous pile holes adjacent to each others, the distribution of excess pore water
pressures in soil mass becomes complicated due to the interaction effects between holes.
Furthermore, it is worth noting that the movements of soil grains in soil mass around a
drilling hole will be reversed in the vicinity between a hole which has been recently
drilled and that hole. Fig. (1-b) presents a proposed soil disturbance zone adjacent to a
drilling hole, in which the disturbed zone has a conical shape with side inclination of 1
horizontal to 4 vertical.
Dissipation
direction
Ground surface
Drilled hole
Slippage/sliding
of soil grains
Tangential strain
Radial stress
10 D
ICSGE-13
GEO-016-4
The above explanations of soil disturbed due to drilling need some evidence. The source
of this evidence is field observations collected from four different construction sites,
which are documented in the subsequent sections.
2.1 Site No. 1
Twelve boreholes were conducted at the site, the recovered soil samples revealed the
following soil stratification; a top fill layer of grey loamy sand which was explored up
to a depth varies between 2.0 m to 6.0 m. The fill overlies very soft clay which is
extending to depth varies from 7.0 m to 9.0 m. A relatively thick layer of medium/fine
sand was encountered at a depth varies from 7.0 m to 9.0 m and extending to a depth
varies between 16 m and 20 m, where a layer of stiff silty clay underlies. The stiff clay
is extending to a depth of 26 m, where a bed of medium to fine sand is extending to the
end of exploration depth which is 40.0 m. The corrected N70% of the top sand layer
varies from 30 to 45 which indicated that the sand is very dense; fig. (2) presents typical
borehole log. The very soft clay has the following properties: undrained shear strength
7.0 kPa, undrained angle of shearing resistance of 3.0 degree, natural unit weight 13.0
kN/m3, natural water content 90%. compression index 0.84, coefficient of consolidation
0.014 cm2/day, in-situ void ratio 2.2, liquid limit 100%, and plastic limit 35%. The
percentage of fines in sand is 19% and the relative density of sand is about 80%. The
ground water table lies at depth varies from 0.20 m to 0.40 m below the ground surface.
Continuous flight auger was used for the construction of bored piles to support the
structure loads. The piles are extending to depth varies between 12 m to 14 m according
to the soil stratification under each building. For 0.50 m pile diameter, the pile working
load is 550 kN and the pile test load is 900 kN whereas for 0.60 m pile diameter, the
pile working load is 700 kN and the pile test load is 1100 kN
Crushed limestone layer of thickness 0.30 m was placed on ground surface to facilitate
machinery manoeuvring. Therefore, the ground water table becomes at a depth varies
from 0.50 m to 0.70 m below working platform. Careful field observations revealed that
there is no stream of water percolated from any pile heads through fresh concrete of the
constructed piles at north side courts of the site. While streams of water were
percolating from some pile heads through fresh concrete of piles constructed in the
other three courts. The number of piles showed percolation of water through fresh
concrete represents 4.3 % of the total number of piles, and decreased towards the north
direction. The process of pile construction was started at north side of the site and
proceeds towards south. The penetration depth of piles in sand varies between 5.0 m at
north side and 7.0 m at south side. The pile bearing stratum is very dense sand. So it is
impossible to correlate the occurrence of phenomenon of soil disturbance to the initial
condition of sand, because Carrier [5] reported that the phenomenon occurs in loose
sand and in this study occurs in very dense sand. This may be true because pile drilling
process initially changes the initial state of sand, following by a process of liquefaction
at a certain extent. The state of liquefaction depends upon the pre-liquefaction geostatic
state of soil and not on the initial state of soil. This may explain also why some piles
showed percolation of water from pile heads through fresh concrete and other did not.
The Pre-liquefaction geostatic state of sand is very complicated and depends upon order
of pile drilling in a group, distance between piles, and time lag between drilling the
closest pile and the pile under consideration.
ICSGE-13
GEO-016-5
Pile loading tests were conducted on seven piles at the construction site where stream of
water percolate from pile heads through fresh concrete. The tested piles of 0.50 m
diameter have code as 51, 217, 378, and 526. For piles of 0.60 m diameter, the tested
piles are 19, 391, and 730. Test pile No. 378 has a grout ratio of about unity. The pile
load tests were conducted in accordance with maintained load test procedure. Fig. (3)
presents the achieved load-settlement of pile loading tests. The results indicated that the
pile displacement after 24 hours after the application of load equal to 1.5 times the pile
working load varies between 1.01% and 2.16% times the pile diameter. The
corresponding non-recoverable displacements are some between 0.8 % and 1.66 times
the pile diameter. The figure revealed that, there is no obvious appreciable effect of
concrete subsidence, and/or grout ratio on the response of the tested piles. All tested
piles carry safely the desired working load. Also there is no appreciable effect of soil
liquefaction due to pile drilling on the behaviour of piles, keeping in mind that the tests
were carried out after 30 days from pile construction.
0
400
Load (kN)
600
800
1000
(3.00)
w = 90 %
c
Cu = 7.00 kN m2
u = 3 o
4
Displacement (mm)
(9.00)
200
6
8
Pile No. 51
Pile No. 217
Pile No. 378
Pile No. 526
10
12
14
N 70%= 35 - 40
16
18
(19.00)
Light grey pieces of crushed cemented sand
(21.00)
Brown grey stiff/hard silty clay
(26.00)
Yell. Grey med./fine sand
0
0
Displacement (mm)
200
400
Load (kN)
600
800
1000
1200
Pile No. 19
Pile No. 391
Pile No. 730
4
6
8
10
12
14
16
ICSGE-13
GEO-016-6
which is extending to depth 19.5 m. A layer of sand was explored up to a depth varies
from 24.0 m to 25.5 m, underlies by layers of medium to stiff clay and fine /medium
sand up to depth of 40.0 m. The very soft clay has the following properties; natural
water content 75 %, natural unit weight 12.5 kN/m3, undrained shear strength 7.0 kPa,
undrained angle of shearing resistance of 4.0 degrees, compression index 2.15, preconsolidation pressure 100 kN/m2, and in-situ void ratio 2.0. The SPT values N70
recorded in top sand layer vary from 70 to 89 which mean that the sand is very dense.
The percentage of fines in sand varies from 1.2% to 6.5 %.
Continuous flight auger was used to construct piles of 0.60 m diameter to support the
structure loads. The piles are designed to transfer structural loads to top sand layer, the
pile length is 21.0 m. Careful field observations of piles during construction revealed
that about 3.5% to 6.5 % of the number of piles showed percolation of water at pile
heads through fresh concrete. The drilling time of piles which is showing percolation of
water varies from 7 to 55 minutes. Also the slump of the poured concrete in these piles
varies from 175 to 220 mm. These observations indicated that percolation of water at
pile heads through fresh concrete takes place irrespective of pile drilling time, or
concrete slump.
2.3 Site No. 3
At the third construction site, twenty four boreholes were carried out to explore the
subsoil conditions. Fig. (5) illustrates one boring log showing the soil stratification at
the site. The retrieved soil samples revealed that the top fill layer of dark grey loamy
sand with little crushed stones is extending to depth of 5.0 m below ground surface. The
fill overlies formation of very soft clay, silty sand, marine shells, silty sand, sandy silt,
Gypsum, and sandy clayey silt up to depth 22.0 m. A bed of very dense sand was
explored up to a depth of 35.0 m, fig. (5). The very soft clay has the following
properties; natural water content 85 %, natural unit weight 13.5 kN/m3, undrained shear
strength 6.0 kPa, undrained angle of shearing resistance of 4.0 degrees, compression
index 2.10, pre-consolidation pressure 90 kN/m2, and in-situ void ratio 2.1.
Continuous flight auger of 0.50 m diameter was used for construction of bored piles to
support the structure loads. The pile length is 24.0 m. The pile working load is 800 kN
and the pile test load is 1200 kN. During construction of a pile, subsidence in ground
surface was noticed. The subsidence of ground surface has a conical shape with a
maximum diameter of 11.0 m at ground surface. This means that, the noticeable
influenced zone of conical shape has a maximum diameter equal to ten times of the pile
diameter at pile tip and extended to ground surface with an inclination of 4 : 1 (vertical
: horizontal). The work was suspended in this pile, and the drilling hole was backfilled
by rotating the auger in opposite direction to drilling operation. After 40 days, the pile
was re-drilled in the same location and poured by fresh concrete. When the pile hole
was filled by concrete, subsidence in fresh concrete was noted as 1.0 m. Therefore, it
can be pointed out that, the excess pore water pressure during pile drilling last more
than 40 days without complete dissipation. During construction of another pile at the
same site, subsidence of ground surface takes place for 0.90 m depth while the height of
fresh concrete in the hole is 18 m, measured from the bottom of the hole. The fresh
concrete was excavated and the work in this pile was suspended. After one month, the
drilling process and casting concrete for this pile was resumed in the same location. The
ICSGE-13
GEO-016-7
Ground surface
Ground water table
(0.50)
Fill of grey loamy sand
(3.00)
Dark grey clayey silty sand
(4.50)
Dark grey very soft clay, trace silt
(5.00)
(6.00) Shells, some grey silty sand
Dark grey very soft/soft clay, trace silt,
trace crushed shells.
N 70%= 70 - 89
(12.00)
(13.00) Crushed gypsum, little sand
Grey soft clay, trace silt
(15.00)
Grey lime, some fine sand,
little Gypsum
(24.00)
Br. grey stiff silty clay
(18.00)
(19.00) Grey Med./fine sand
(20.00) Crushed gypsum, little sand
Grey sandy clayey silt
(31.50)
Yell. med./fine sand
(33.00)
Brown stiff silty clay
(34.50)
(22.00)
(23.00) Yell. weak crushed sand stone
Yell. Grey med./fine sand
pile was constructed successfully without any problem. Furthermore, at the same site,
streams of water from top of some piles through fresh concrete were noted. The amount
of percolating water is sufficient enough to wash out the cement from the top three
meters of fresh concrete in piles. As a result, the crew of drilling pile machine was
unable to insert the steel cage through fresh concrete. This phenomenon was observed in
about 15% of the constructed piles at this site. The final observation in this site is that,
some constructed piles consume more than 2.76 times the theoretical volume of
concrete to form the pile. The stability of fresh concrete in the drilling hole was checked
as; cu (c.kc - s.ko).H/Nc, where c=12 kN/m3, kc=0.70, s=3.5 kN/m3, ko=1.0, H=12 m,
and Nc= 5.7. The above equation gives that; undrained shear strength of very soft clay
should be more than 10.3 kPa, which less than the measured value. Therefore collapse
of some drilling holes in very soft clay is evidence. It is important to note that, the
implementation of the above stability equation for pile drilling holes does not take into
consideration the effects of nearby piles which had been previously constructed.
The pile construction process at this site caused many imperfections in piles, if
compared by other sites. It is worth noting that, the construction site is adjacent to 8story building constructed on raft foundation since 25 years ago and surrounding by 15story building which recently constructed on piled foundation. Two roads of width 8.0
ICSGE-13
GEO-016-8
m and 6.0 m are bounded to the construction site. The building of raft foundation
exhibits tilting away from the pile construction site due to consolidation settlement. It is
anticipated that the excess pore water pressure induced in clay layer due to super
structure loads dissipates vertically and horizontally to the pile construction site.
Especially the construction site is free superimposed stress on subsoil layers. As a result
of this, the pore water pressure built up at the pile construction site is bigger than the
hydrostatic value, which adds another adverse effect and complicated the problem of
pile imperfection during drilling process.
2.4 Site No. 4
At the fourth construction site, it is required to construct two basement floors below the
ground surface. The subsoil investigation revealed that a top fill layer of 4.0 m overlies
soft clay layer. The ground water table lies at a depth of 0.70 m below the ground
surface. A retaining wall of secant CFA piles of 800 mm diameter was constructed. The
piles are extended to depth of 20.0 m below the ground surface. During construction of
piled-wall, cracks were observed in the pavement layers of the adjacent road. These
cracks are extending to a distance of 13.0 m from centreline of the constructed piles.
These cracks conform well that the soil disturbance zone adjacent to a drilling hole
initiated to ten times of the pile diameter at the pile tip and extended with a side
inclination of 1 horizontal to 4 vertical to the ground surface.
3
To explain the phenomenon of water percolation from top of some CFA piles through
fresh concrete, it is essential to realize the effect of drilling process on the properties of
soil around the pile. During drilling, the cutting bits at the auger head cut the
geomaterial and shear a cylinder of soil around the auger head from a soil mass. It is
known that in many soils, the shear strength differs with depth. As a result of that, the
control of penetration rate of the auger and rate of shear strain are extremely difficult in
drilling process. Therefore, the pile drilling process imposes cyclic shear strain on
surrounding soil. Cyclic shear strain causes degradation of normal stresses in soil, and
results degradation in shear strength of soil. At the same time, cyclic shear decreases
volume of loose granular soil followed by an increase in pore water pressure. The
consequence of cyclic loading under undrained conditions is generally an increase in the
pore water pressure, which causes the effective stress to decrease and the cyclic
deformations of the soil to increase. If the soil is saturated, changes in pore pressure will
occur during fast cycling. That is where the rate of cyclic is such that changed in pore
pressure are not fully dissipated. Furthermore, the grain structures may be mobilized
when the stress state changes. The stress change may be due to coupling effects between
volumetric and shear strains created by drilling process. The grain structure
mobilization resulted soil particle reorientation including soil particles rotation and
sliding relative to each other. This rotation and sliding of soil particles may cause the
soil grain structure to compress or dilate and the soil strength to increase or decrease. In
undrained conditions, large pore water pressure increase may suddenly occur which is
accompanied by a rapid decrease in strength. The phenomenon is described as grain
structure collapse under undrained conditions; the soil grain structure collapse may be
accompanied by a sudden large volumetric compression, Gu & Kraha [12]. The excess
pore water pressure build up in soil mass due to cyclic loading was reported by many
ICSGE-13
GEO-016-9
researchers; Silver & Seed [6], Seed & Idriss [13], and El-gamal et al [14], while excess
pore water pressure build up during drilling process is not presented in literature. The
excess pore water pressure generated in a soil during cyclic loading can be separated
into two components coded as transient and residual. In saturated soils, the transient
excess pore water pressures are equal to the changes in the applied mean normal stresses
resulting from the dynamic loading. The transient excess pore water pressures have little
influence on the effective stresses acting on the soil. Conversely, the residual excess
pore water pressures result from the progressive collapse of the soil skeleton alter the
effective stresses acting on the soil. Consequently, the residual excess pore water
pressures are directly influence the strength and stiffness of the soil, Polito et al [15].
The shearing motion of the soil resulted from the rotation of the auger, which causes a
slippage or sliding of soil grains, causes the constrained modulus (Mc) to be reduced.
The authors believe that the effect of shearing motion extends in the surrounding soil
around the auger to a distance of radius equal to about ten times of the hole diameter at
the cutting head. This zone has a conical shape of 4:1 (vertical: horizontal) inclined
surface. With further cyclic shearing motion, the soil particles do not slide any more and
the soil skeleton recover much of it is original strength but in a slightly consolidated
form. The excess hydrostatic pressure in the disturbed zone will partly dissipate by
horizontal flow towards the hole. In loose soil, cyclic shear strain causes a decrease in
volume and increase in pore water pressure. In dense soil, cyclic shear strain causes
grain structure collapse under undrained conditions accompanied by big volume
compression and an increase in pore water pressure. Therefore the authors believed that,
the phenomenon of pore water pressure increase is not related only to dense soil as
reported by Carrier [5].
4
CONCLUSIONS
Drilling hole in soil mass using continuous flight auger induce cyclic shear strain on
soil, which cause; degradation of stresses in soil , grain structure collapse, decreased in
constrained modulus of soil and increase in pore water pressure. The excess pore water
pressure induced in soil during drilling of pile hole depends upon, normalized shear
stress ratio, pre-liquefaction geostatic state of soil, and number of cyclic shear strain.
Percolation of water from pile heads through fresh concrete of piles constructed using
continuous flight auger takes place irrespective of the initial condition of sand, whatever
loose, dense or very dense. Moreover, drilling time and concrete slump haven't
appreciable effects on excess pore water pressure induced in soil adjacent to drilling
piles. Piles exhibit percolation of water from top of piles through fresh concrete
represents 3.5% to 6.5 % of the number of piles. There are no appreciable effects of
fresh concrete subsidence in a drilling pile hole, percolation of water from pile head
through fresh concrete, and/or grout ratio on load displacement characteristics of the
tested piles. The zone of soil disturbed by continuous flight auger piling has a conical
shape extending laterally to a distance equal to ten times the pile diameter at the cutting
head and has an inclined surface of 4 (vertical) : 1 (horizontal).
ICSGE-13
GEO-016-10
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
ICSGE-13
GEO-016-11