See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/262639770

Finite element modelling of ground vibrations due to tunnelling activities

Article · January 2011

CITATIONS READS
0 63

1 author:

Muhammad Ekhlasur Rahman

Curtin University Sarawak
51 PUBLICATIONS   323 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Investigation on the behavior and performance of wire rope isolators View project

ASR suppression using fine, medium and ultrafine POFA treated at various elevated temperatures View project

All content following this page was uploaded by Muhammad Ekhlasur Rahman on 04 September 2014.

The user has requested enhancement of the downloaded file.

 International Journal of Civil and Environmental Engineering 3:3 2011

Finite Element Modelling of Ground Vibrations

Due to Tunnelling Activities
Muhammad E. Rahman and Trevor Orr

! Tunnelling under city streets has now become a common

Abstract—This paper presents the use of three-dimensional finite activity for nearly all the types of transportation, but long
elements coupled with infinite elements to investigate the ground urban tunnels are principally for metros, water supply and
vibrations at the surface in terms of the peak particle velocity (PPV) sewers. Shorter tunnels may be required for highway
due to construction of the first bore of the Dublin Port Tunnel. This
underpasses and pedestrian subways. Longer highway tunnels
situation is analysed using a commercially available general-purpose
finite element package ABAQUS. A series of parametric studies is are sometimes suggested but problems of access and
carried out to examine the sensitivity of the predicted vibrations to ventilation add greatly to their difficulty and cost, against
variations in the various input parameters required by finite element which preservation of surface land and gains in environmental
method, including the stiffness and the damping of ground. The amenity may prove a quite inadequate offset [4, 5]. It is
results of this study show that stiffness has a more significant effect necessary to ensure that structures of historic and economic
on the PPV rather than the damping of the ground.
importance will not be affected or that any effects will be
.
within acceptable limits. Vibration from tunnelling activities
will normally be of a temporary nature and much smaller than
Keywords—Finite Elements, PPV, Tunnelling, Vibration
earthquake vibrations[6, 7, 8, 9, 10].
I. INTRODUCTION The consequences of tunnelling activities have prompted
recent research into improved methods for predicting the
T HE environmental effects of noise and vibrations have
received considerable attention in recent years due to the
damage that vibrations can cause to buildings and because of
ground vibrations induced by tunnelling in free field
conditions and in houses. This paper presents the use of three-
dimensional finite elements coupled with infinite elements to
the sensitivity of people to noise and vibration [1, 2]. predict the ground vibrations at the surface level in terms of
Vibrations can be generated either by natural or artificial the PPV due to construction of the Dublin Port Tunnel.
sources. Natural sources are earthquakes, ocean waves and
landslide. In the case of earthquakes, the intensity of ground II. DUBLIN PORT TUNNEL
vibrations may be high enough to result in severe structural
The Dublin Port Tunnel project comprises approximately
damage or even collapse of structures accompanied by loss of
5.6km of dual carriageway, of which 4.5km are underground,
life [3]. Ground vibrations from artificial sources are called
with 2.1km of twin cut and cover tunnels and 2.4km of twin-
man-made vibration. Sources of man-made vibrations include
bored tunnels, plus associated interchanges and infrastructure.
machinery, compressors, pile drivers, road traffic, tunnel It provides a link from Dublin Port to Santry in the Dublin
construction, mechanised construction activities, demolition of suburbs and the motorways to the rest of Ireland. The depth of
structures and aircraft. The magnitude of man-made vibrations the bored tunnel section, from existing ground level to the
is normally much smaller compared to earthquake vibrations roof of the tunnel, ranges from approximately 19m to 24m
and hence these vibrations do not, in most cases, cause serious under residential houses.
structural damage to buildings and their effects are limited to
the development of cosmetic cracks in the walls and floors. III. GROUND CONDITIONS
Man-made vibrations can also induce densification in sandy
soils, resulting in foundation settlement. This settlement has The ground conditions within the tunnel area consist of
the potential of inducing more serious structural damage than glacial deposits resting on limestone bedrock. In many areas
the cosmetic cracking that results from direct vibration due to along the route of the tunnel these deposits are overlaid by
machinery in the building itself [3]. made ground of varying thickness, which is site specific. Rock
head levels along the line of the tunnel are relatively close to
the surface and generally vary from 5 to 20m below the
ground level. Due to the irregular profile of the rock head, the
Muhammad E. Rahman is with the Curtin University Sarawak, Miri, ground cover can change significantly over a short distance.
Sarawak, Malaysia (corresponding author phone: 60-128451750; e-mail:
The water table is at around 1m below the ground level.
[email protected]).
Trevor Orr is with the University of Dublin, Trinity College, Dublin-2,
Ireland.

143
 International Journal of Civil and Environmental Engineering 3:3 2011

The glacial deposits consist of brown boulder clay TABLE I

underlain by black boulder clay. The brown boulder clay VALUES OF SOIL PARAMETERS USED IN REFERENCE MODELS
consists of stiff sandy silty clay and sandy clayey silt with Top layer DBC1 Limestone Concrete Steel
some gravel. The black boulder clay can generally be
described as very stiff to hard slightly sandy clay with much Density, 20 23 27 24 78
fine to coarse, subrounded to subangular, limestone gravel and kN/m3
occasional cobbles and boulders. The underlying Dublin
Young’s 0.2 1.0 60 40 200
formation bedrock comprises carboniferous limestone and modulus, GPa
shales known as ‘Calp’ limestone. The limestone can be
described as a predominantly strong to very strong, very thin Poisson’s ratio 0.49 0.49 0.25 0.2 0.3
to medium bedded, dark grey, predominantly medium to Damping, % 2.85 1.75 1.0 1.0 0.5
coarsely crystalline and fossiliferous limestone, with
1
occasional interbedded, moderately strong, very thin to thin, DBC= Dublin Boulder Clay
dark grey to black, unweathered shaley mudstones and strong
to very strong fine grained argillaceous limestone [11]. The V.NUMERICAL MODELLING
limestone within the project area lies wholly beneath the water
table with artesian pressure to near surface observed in Three dimensional finite element models were used in this
piezometers. Water inflows during tunnelling in the limestone paper to predict the ground vibrations at the surface. Three-
were generally around 250litres/min for the full 11.8m dimensional finite element models allow the analyst to
diameter. The ground conditions were modelled as three account for the extent and geometry of each of stratum when
different layers of soil strata and the values of the material constructing a representative numerical model to predict the
properties used for the each layer were based on the results of ground vibrations due to the TBM. The problem to be
ground investigations. The material properties that were used analysed is shown in Fig. 1. In Fig. 1, the geometry is
in the finite element analyses are given in Table I. idealised by a finite region (Part I) and a semi-infinite far field
region (Part II). The ABAQUS V6.4 finite element software
IV. EMPIRICAL MODELS was used to construct the model and three-dimensional eight
As so many variables are involved in the determination of noded linear brick and reduced integration elements, C3D8R,
vibration levels in the ground due to tunnelling, no explicit were employed for the ground and the tunnel lining which is
equations exist which allow the magnitude of the ground under Part I. Reduced-integration elements use fewer
vibrations at the surface to be predicted accurately for any integration points in each direction than the fully integrated
given source and ground conditions. Approximate empirical elements. These reduce the running time, especially in three-
equations have been developed based on a limited number of dimensional finite element analyses. In Part II, the soils and
case studies, and these concentrate on the simulated resultant tunnel lining in the far field that extends to infinity, were
peak particle velocity at the ground surface. Prediction modelled by the infinite elements, CIN3D8. In the model, the
methods are required that can predict the ground vibrations top layer is brown clay, the middle layer is Dublin boulder
taking account of both the nature of the activity causing the clay and the bottom layer is limestone. The finite element
vibrations and the characteristics of site. Mechanised mesh is shown in Fig. 2. The elementary boundary conditions
tunnelling generates levels of groundborne vibration that are are applied on the four sides of the vertical section to obtain
unlikely to cause damage to structures. horizontal fixities, which allow only vertical movement, and
Godio et al., (1992 cited in Hillar & Crabb, 2000) proposed these boundaries are also applied on the bottom side to obtain
an equation for predicting an upper bound to the peak particle a full fixity [12].
velocities, v in mm/s due to ground vibrations. This is a useful The dimensions of the finite element model are 132m in the
first estimate for the vibration levels likely to be generated by transverse direction, 50m in the vertical direction and 40m in
mechanised bored tunnelling works and is given by the the direction of the tunnel. The thickness of the top layer is
Equation (1): 2.5m, the middle layer is 11m and the bottom layer is 36.5m.
v = Ar-1.3 (1) (1) machine is a
By advancing in the ground, the tunnel-boring
where A is a constant whose value depends on the ground vibrating source from which vibrations are transmitted into the
condition (for hard ground A=180) and r is the direct distance ground. Therefore, the impacts of the TBM on the ground
(m) from the vibration source to the measurement location. during boring have been modelled by a dynamic loading
Nishimatsu also produced equations for upper and lower system. The finite elements representing the TBM cutter have
bounds for v based on monitored vibration values on other been loaded by a short triangular pulse. Due to this pulse,
construction sites. Their upper and lower bound equations for waves propagate through the soil and the response at the
the peak particle velocity are: surface is determined. The analyses have been carried out in
v = 176r–1.18 (upper bound) (2) the time domain.
v = 7.4r–1.07 (lower bound) (3)

144
 International Journal of Civil and Environmental Engineering 3:3 2011

VI. NUMERICAL RESULTS

Part I
Finite element models were constructed of the ground, the
Part II
tunnel lining and the frame of the TBM at different sections
along the tunnel bore. Consistent modelling techniques were
used for each section points with material properties for the
ground of these sections based on ground investigations.
The choice of element to be used for the ground, the tunnel
lining and the frame of the TBM is limited in ABAQUS V6.4 Fig. 1 Typical model
because the three-dimensional eight noded linear brick and
reduced integration element, C3D8R, is the only element
available both in the continuum solid element and the infinite
element family. Eight-noded linear brick elements with a
linear material model have been used to define the ground, the
tunnel lining and the frame of the TBM.
In each case, the thicknesses of the different layers of
ground were selected from the ground investigation
information and extended in the tunnel direction to generate
the solid model. For simplification, the material properties in
each layer were assumed to be constant in the tunnel direction
and in the transverse direction in the three-dimensional finite
element model.
Fig. 2 Finite element mesh
VII. RESULTS AND DISCUSSION
1
The vibrations were generated in the finite element model
by applying a unit triangular pulse (Fig. 3) to the ground in 0.8
Velocity, mm/s.

front of frame of the TBM. The output results presented in 0.6

Fig. 4 and Fig. 5 are the ground response in the transverse 0.4
direction on the surface just above the TBM in the time
0.2
domain and the frequency domain respectively at measuring
0
Section 1 due to the unit triangular pulse. In Fig. 4, it can be
0 0.02 0.04 0.06 0.08 0.1
seen that the finite element predicted ground vibrations are
Time, s
free from any reflections and this indicates that the infinite
elements in ABAQUS are working effectively. In Fig. 5 it can Fig. 3 A short triangular pulse
be seen that the frequency band of greatest vibration 0.020
amplitudes is in the range of 8 to 30Hz and peak amplitudes 0.015
Velocity, mm/s.

occurs at 18Hz. Fig. 6 shows the measured TBM excitation in 0.010

the transverse direction at measuring Section 1 in the time 0.005
domain and Fig. 7 is the Fast Fourier Transform (FFT) of Fig. 0.000
6 and shows that most of the energy is concentrated in the -0.005
frequency range of 3 to 30Hz; these are also called the source -0.010
function. Fig. 8 shows the vibration velocity v(!) in the 0 0.2 0.4 0.6 0.8 1
transverse direction plotted against frequency, which derived Time, s
from the following equation:
v(!) = F(!)*{output(!)/input(!)} (4) Fig. 4 Predicted time history in surface level at measuring Section 1;
where F(!) is the source function, i.e. the TBM excitation. transverse direction
Fig. 9, Fig. 10 and Fig. 11 show the vibration velocity at the
surface level just above TBM at measuring Section 1 in the
time domain in the three orthogonal directions (x, y and z) due Resultant PPV = TPPV 2 "VPPV 2 " LPPV 2 (6)
to the transverse input load. The peak particle velocity due to where Tx = Vibration velocity in x direction (transverse) due
a transverse load and the resultant peak particle velocity are to a transverse input load
calculated using Equations 5 and 6 respectively [14]. Ty = Vibration velocity in y direction (vertical) due to a
Peak Particle Velocity (PPV) due to a transverse load transverse input load
(TPPV) = Tx 2 "T y 2 "Tz 2 (5) Tz = Vibration velocity in z direction (longitudinal) due to a
transverse input load

145
 International Journal of Civil and Environmental Engineering 3:3 2011

150
TPPV = Peak Particle Velocity (PPV) due to a transverse input

Amplitude, (mm/s)/Hz
load
VPPV = Peak Particle Velocity (PPV) due to a vertical input 100
load
LPPV = Peak Particle Velocity (PPV) due to a longitudinal 50
input load.
0
0.50 0 1 10 100 1000
Amplitude, (mm/s)/Hz.

0.40 Frequency, Hz
0.30
Fig. 8 Transverse vibration velocity (Frequency domain) at the
0.20
surface at measuring Section 1; transverse direction
0.10
0.00 1.00

Velocity, mm/s.
1 10 100 1000 0.50
Frequency, Hz 0.00
Fig. 5 FFT of time history in surface level at measuring Section 1; -0.50
transverse direction
-1.00
15
-1.50
Velocity, mm/s.

10
0.00 0.20 0.40 0.60 0.80 1.00
5
Time, s
0
-5 Fig. 9 Transverse vibration velocity at the surface level measuring
-10 Section 1 due to transverse vibration.
-15
0 0.2 0.4 0.6 0.8 1 0.02
Velocity, mm/s.

Time, s 0.01
Fig. 6 Measured TBM excitation at measuring Section 1; transverse
direction 0.00

2000 -0.01
Amplitude, (mm/s)/Hz.

1500 -0.02
0.00 0.20 0.40 0.60 0.80 1.00
1000
Time, s
500 Fig. 10 Vertical vibration velocity at the surface level at measuring
Section 1 due to transverse vibration
0
0 1 10 100 1000 0.02
Velocity, mm/s.

Frequency, Hz 0.01
Fig. 7 FFT of TBM excitation at measuring Section 1; transverse 0.00
direction
-0.01
The 3-D FE predicted resultant PPV values together with
the Godio and Nishimatsu upper and lower bound values are -0.02
plotted in Fig. 12, 13, 14 and 15. The PPV values plotted in 0.00 0.20 0.40 0.60 0.80 1.00
Fig. 12, 13, 14 and 15 show that there is good agreement Time, s
between the 3D-FE predicted values and the Godio and
Nishimatsu upper and lower bound values. It can be seen Fig. 11 Longitudinal vibration velocity at the surface level at
measuring Section 1 due to transverse vibration
from the results plotted for each section that the 3D-FE lie
between the Godio and Nishimatsu upper and lower bound
the source and the point of interest, while the groundborne
values. It can also be seen from Fig. 12, 13, 14 and 15 that all
vibrations depend also on the soil properties particularly, on
the models have similar attenuation characteristics. Empirical
the stiffness and the resonance frequency of the soil layer, and
models are a function of only the direct distance between
on the tunnel boring machine excitation.

146
 International Journal of Civil and Environmental Engineering 3:3 2011

TABLE II
RANGES OF PARAMETERS USED IN SENSITIVITY ANALYSES
Top layer DBC1 Limestone Concrete Steel

Stiffness, 0.01, 0.1, 0.2, 1.0 60 40 200

GPa 0.5, 1.0

Frequencies 8 to 35 8 to 35 8 to 35 8 to 35 8 to 35
Ranges, Hz 1 to 60 1 to 60 1 to 60 1 to 60 1 to 60

Damping, % 2.85 1.75 1.0 1.0 0.5

1
DBC= Dublin Boulder Clay
Fig. 12 FE, empirical and measured PPV values at Section 1

TABLE III
NORMALISED PPV WITH RESPECT TO REFERENCE MODEL
Frequency -10m from TBM At TBM face +10m from TBM
Range face face
8 to 35 Hz 1.0 1.0 1.0
(Reference)

1 to 60 Hz 1.1 1.18 1.05

sensitivity of the vibration predictions to variations in the

various input parameters, which include the stiffness and the
damping properties of the soil, are discussed.
Fig. 13 3D-FE, empirical and measured PPV values at Section 2 The Nishimatsu supplied the values of the material
parameters used for the finite element modelling of the strata.
The samples of soil were collected at different locations and
different depths along the tunnel line for testing in the
laboratory. Field tests were also carried out to determine the
values of the soil parameters required for the design of the
tunnel and for use in the finite element analyses.

A. The Effects of Varying the Stiffness of the Soil

A reference model of the ground was selected in which the
soil parameter values were based on the results of the soil
tests. The parameters required for the dynamic analyses were
Fig. 14 3D-FE, empirical and measured PPV values at Section 3
the weight density, dynamic Young’s modulus, Poisson’s ratio
and the damping ratio and the values of these parameters are
given in Table I.
In order to investigate the effect of variations in the material
parameters on the produced vibrations, the stiffness of the
upper soil layer of the reference model was changed. The
range of values used in this study for the soil parameters in the
upper soil layer are summarised in Table II.The results of the
parametric study on the effects of the upper layer’s modulus
of elasticity on the amplitude of the vibrations are presented in
Fig. 16, 17 and 18. As can be seen in Fig. 16, the resonance
frequency increases with increasing the stiffness of the upper
Fig. 15 3D-FE, empirical and measured PPV values at Section 4 soil layer. The resonance frequency is dependent on the
stiffness of the soil layer. When the resonance frequency of
the soil layer and the TBM excitation coincide or are very
VIII.PARAMETRIC STUDY close, this results in a maximum vibration velocity. In Fig.16,
it can be seen that the resonance frequency of the soil layer at
The level of vibration observed at any point depends on the
upper layer’s stiffnesses of 0.01GPa, 0.1GPa, 0.2GPa, 0.5GPa
energy transmitted into the ground at the source and its
and 1.0GPa are 4Hz, 15Hz, 18Hz, 20Hz and 22Hz
subsequent attenuation within the ground. In this section, the
respectively and in Fig. 7 the resonance frequency of the TBM
results of a series of parametric studies to examine the

147
 International Journal of Civil and Environmental Engineering 3:3 2011

excitation is 14Hz. Since the resonance frequency of 15Hz of

the soil layers at stiffness 0.1GPa of upper soil layer is very 0.60

Amplitude, (mm/s)/Hz.
0.01 GPa
close to the resonance frequency of 14Hz of the TBM 0.50
0.1 GPa
excitation, the maximum vibration velocity is obtained at this 0.40 0.2 GPa
stiffness which is shown in Fig. 17. The minimum vibration 0.5 GPa
0.30
velocity is obtained at stiffnesses of upper soil layer of 0.01 1.0 GPa
0.20
GPa and 1.0 GPA, since at these stiffnesses there is a
significant difference between the resonance frequency of the 0.10
soil layer and the resonance frequency of the TBM excitation. 0.00
Fig. 18 shows the variations of velocity reduction with the 1 10 100 1000
distance. The vibrations at points on the ground surface
Frequency, Hz
directly above the tunnel axis and up to 50m an either side are
most sensitive to variations in the stiffness of the upper soil Fig. 16 Transverse output response (frequency domain) in surface
layer. After 50m from the tunnel axis the influence of the level
material parameters on the surface vibrations is very small.
350
B. The Effects of Varying the Damping of the Soil 0.01 GPa

Amplitude, (mm/s)/Hz.
300
The damping ratio of the soil is a parameter that is difficult 0.1 GPa
250
to determine. Often the damping ratio is chosen as 1-6% of the 0.2 GPa
200 0.5 GPa
critical damping ratio. In the finite element code, the damping
150 1.0 GPa
matrix, C is defined as a having two components, one
100
proportional to the mass and the other proportional to the
stiffness [15]: 50
0
[C] = " [M] + # [K] (7)
1 10 100 1000
where [K] is the stiffness matrix, [M] is the mass matrix, [C]
Frequency, Hz
is the damping ratio matrix, " is the coefficient of mass
damping and # is the coefficient of stiffness damping. Fig. 17 Transverse vibration velocity (Frequency domain) in surface
In this formulation of the damping, two frequencies can be level
chosen at which a given proportional damping applies. At
1.4
other frequencies, the damping ratio will be different. Figure
1.2 0.01 GPa 0.1 GPA
19 shows two cases. In the reference model, the frequency is
Velocity, mm/s.

1 0.2 GPa 0.5 GPa

chosen so that it is exactly 1.0% at 8 and 35 Hz. In that case, 1.0 GPa
0.8
the damping ratio is close to 1.0% between 8 and 35 Hz.
0.6
However, at lower and larger frequencies, the damping ratio is
0.4
considerably larger. The two frequencies values at which the
0.2
damping ratio matches the given value should not be far away
0
from each other. In Fig. 19 can be seen that the damping ratio
0 10 20 30 40 50
will drop to 0.28% if the lower and upper bounds are 1 and 60
Hz; i.e. if the chosen frequency range is wide, the damping Distance from tunnel centre line, m
will be small between the lower and upper bound frequencies.
Fig. 18–The effects of stiffness on ground surface vibration
In order to investigate the effect of the material parameters
on the results, the frequency range to calculate the " and # 7
values for the reference model have been changed. The data 6
used in this study are summarised in Table II. 8Hz and 35Hz
Damping ratio, %

5 1Hz and 60Hz

Over the whole frequency range the response function will
4 Damping ratio 1%
increase if the chosen frequency range is wide and
3
consequently the amplitude of the vibrations will increase. It
2
can be seen from Fig. 20 and Fig. 21 that the response
function increased with increasing the chosen frequency 1

range. The normalised results of the parametric study on the 0

effects of the damping are presented in Table III. In Table III, 0 20 40 60 80 100
Frequency, Hz
it can be seen that the vibration velocity increases with
increasing the frequency ranges i.e. with decreasing damping Fig. 19 Damping ratio of 0.01 (1%) dependent on frequency
value. However, the increase in the vibration velocity is the
highest on the surface directly above the TBM and decreases
with distance from the tunnel face.

148
 International Journal of Civil and Environmental Engineering 3:3 2011

insignificant compared to the effect of vibrations in the

0.80
8Hz and 35Hz
stiffness.
Amplitude, (mm/s)/Hz.

0.60 1Hz and 60Hz

ACKNOWLEDGMENT
0.40 The work was carried out at the Department of Civil
Engineering, University of Dublin, Trinity College. The
0.20
project was supported by the Nishimatsu Construction Co.,
0.00 Ltd and the Geotechnical Trust Fund of the Institution of
0 1 10 100 1000 Engineers of Ireland. This support is gratefully acknowledged.
Frequency, Hz
REFERENCES
Fig. 20 Response function on surface directly above the TBM face
[1] Hillar, D.M., Hope, V.S., “Groundborne vibration generated by
mechanised construction activities,” Proceedings Institution of Civil
0.80 Engineers: Geotechnical Engineering, vol. 131, no. 4, pp 223-232, 1998.
8Hz and 35Hz [2] Hall, L., “Simulations and analyses of train-induced ground vibrations in
Amplitude, (mm/s)/Hz.

0.60 1Hz and 60Hz finite element models,” Int. J. of Soil Dynamics and Earthquake Eng.,
vol. 23, pp 403-413, 2003.
[3] Athanasopoulos, G.A., Pelekis, P.C., “Ground vibrations from sheet pile
0.40 driving in urban environment: measurements, analysis and effects on
buildings and occupants,” Int. J. of Soil Dynamics and Earthquake Eng.,
0.20 vol. 19, pp. 371-387, 2000.
[4] Megaw, T. M., Bartlett, J.V., “Tunnels-Planning, Design, Construction,”
Ellis Horwood Ltd Chichester, 1983.
0.00 [5] Burd, H.J., Houlsby, G.T., Augarde, C.E., Liu, G., “Modelling the
0 1 10 100 1000 effects on masonry buildings of tunnelling-induced settlement,”
Proceedings Institution of Civil Engineers: Geotechnical Engineering,
Frequency, Hz
vol. 143, pp. 17-29, 2000.
Fig. 21 Response function on surface at 10m behind of the TBM face [6] Whyley, P.J., Sarsby, R.W., “Ground borne vibration from piling”
Ground Engineering, vol. 26, pp. 32-37, 1992.
[7] New, B.M., “Ground vibration caused by construction works,”
IX. CONCLUSION Tunnelling and Underground Space Technology, vol. 5, pp. 179-190,
1990.
3-D finite element analyses have been carried out to [8] New, B. M., “Vibration caused by underground construction. Tunnelling
predict the TBM induced ground vibrations at the surface ’82,” The Institution of Mining and Metallurgy, pp. 217-229, 1982.
level. Although 3-dimensional finite element analyses are very [9] Flanagan, R.F., “Ground vibration from TBMs and shields,” Tunnels &
Tunnelling, vol. 25, no. 10, pp 30-33, 1993.
time consuming, good agreement has been found, in terms of [10] Kramer, S.L., “Geotechnical Earthquake Engineering,” Prentice Hall,
PPV, between the empirical model PPV values at the surface New Jersey, 1996.
and the values calculated using 3-D finite element analyses. [11] Farrell, E.R., O’Brien, S, Lehane, B., Orr, T., “Stiffness of Dublin Black
Based on the predictions, it is concluded that the surface Boulder Clay,” Proceedings XI ECSMFE, Copenhagen, pp 1-6, 1995.
[12] Potts, D. M., Axelsson, K., Grande, L., Schweiger, H., Long. M.,
vibrations are low enough not to cause any damage to “Guidelines for the use of advanced numerical analysis,” Thomas
structures on the surface. Empirical models are crude at best, Telford London, 2002.
as these are a function of only the direct distance between the [13] Hillar, D. M., Crabb, G. I., “Groundborne vibration caused by
mechanised construction works,” Report TRL, No-429, pp1-79, 2000.
source and the point of interest, while the groundborne [14] Head, J.M., Jardine, F.M., “Ground-borne vibrations arising from
vibrations also depend on the soil properties and especially on piling,” CIRIA Technical Note 142, 1992.
the stiffness and the resonance frequency of the soil layer, and [15] ABAQUS, “Analysis User’s Manual”, Hibbit, Karlsson & Sorensen,
ABAQUS Europe BV, Maastricht, The Netherlands, 2004.
on the tunnel boring machine excitation.
The stiffness of the ground is a most important parameter,
which affects the resonance frequency of the ground. If the
resonance frequency of the ground coincide with the
resonance frequency of the TBM excitation or is very close to
it, this will lead to larger vibrations level and vice versa. In the
Dublin Port Tunnel, these two-resonance frequencies are quite
close and yielded higher surface vibrations. Changes in the
stiffness of the ground close to the tunnel have significant
effects on the surface vibrations, while changes in the stiffness
of the ground further away from the tunnel have less effect on
the surface vibrations.
In the finite element, formulation the damping is defined as
a combination of mass matrices and stiffness metrics. The
differences between the results are small indicating that the
effects of the damping ratio on the surface vibrations are

149

View publication stats