Proceedings of the Institution of

Civil Engineers
Geotechnical Engineering 161
October 2008 Issue GE5
Pages 227233
doi: 10.1680/geng.2008.161.5.227
Paper 80035
Received 25/01/2006
Accepted 08/07/2008
Keywords: tunnels & tunnelling

M. E. Rahman
Department of Civil, Structural
and Environmental
Engineering, University of
Dublin, Trinity College, Ireland

T. L. L. Orr
Department of Civil, Structural
and Environmental
Engineering, University of
Dublin, Trinity College, Ireland

Using trenches to reduce tunnelling vibrations

M. E. Rahman

MScEng, PhD

and T. L. L. Orr

MSc, PhD, EurIng, CEng, MICE, FIEI

Wave barriers are often used in engineering practice to

reduce the ground vibrations induced by man-made
sources such as traffic and machine foundations. In this
paper, the results of a numerical study, with threedimensional finite elements coupled with infinite
elements, are presented to examine the effectiveness of
open and backfilled trenches as wave barriers in reducing
the ground surface vibrations due to tunnelling activities.
The effects of the geometrical dimensions of the
trenches and material properties of the backfill on the
reduction of ground vibrations are studied. It is found
that the percentage reduction in the vibration velocities
increases with increasing depth of the trenches, and
softer backfill materials yield a better screening effect
than stiffer materials.
1. INTRODUCTION
Wave barriers are often installed in the ground to reduce the
ground vibrations induced by man-made sources such as traffic
and machine foundations. 1 Such ground vibrations and those
due to a tunnel-boring machine (TBM) can be a major problem
in densely populated urban areas and for structures that house
sensitive equipment. The environmental effects of vibrations
have received considerable attention in recent years, owing to
the damage they can cause both to buildings and to people,
and in many countries new rules and regulations have been
introduced to control vibrations. 2 Therefore screening of
vibrations induced by traffic and construction activities has
become an important issue in recent times and it has been
shown that it is possible to reduce the ground vibrations
significantly by placing a suitable wave barrier in the ground.1
The methods of vibration isolation for soilstructure
interaction systems may be classified into two families: active
isolation and passive isolation. Active isolation, also known as
source isolation, refers to the installation of barriers at a
distance so close to the wave source that energy transmitted
from it can be directly cut off, or to the installation of barriers
surrounding the vibration source. 3 Passive isolation, also
known as receiver isolation, involves placing suitable barriers
in the ground before the structure or surrounding the
structure. 4 An active isolation barrier can effectively isolate
stationary sources of vibration, whereas a passive isolation
barrier is effective for a wide variety of wave-generating
sources, including moving sources. Different kinds of wave
barrier, among them open and backfilled trenches, are the most
Geotechnical Engineering 161 Issue GE5

common in practical applications as they are the most effective

and their installation cost is low. 5,6
Much research, both experimental and numerical, has been
carried out in the past few decades to study the effectiveness of
wave barriers for reducing ground vibrations. An analytical
approach has rarely been used, because closed-form solutions
are extremely difficult to obtain, except for very simple
geometries and boundary conditions that hardly exist in
practice. While full-scale experimentation is expensive to carry
out, small-scale model test results can be difficult to
extrapolate to prototype situations. An efficient numerical
technique, on the other hand, can be an effective alternative
method for conducting a thorough investigation into the
vibration isolation phenomenon. 1,4
Barkan7 investigated the screening of waves by means of a
trench and found that the amplitude of the waves decreased
with increasing the depth of the trench. Woods 8 also carried
out a series of field tests on vibration isolation by installing
open trenches from very close to the wave source to the far
field and found that reductions in the amplitude of ground
vibrations of up to 75% could be achieved.
In the past few decades, researchers have used several
numerical techniques to study the vibration isolation problem.
Ju 9 used three-dimensional finite element analyses with
absorbing boundary conditions to isolate structures from traininduced vibrations by means of soil improvement and open
and infilled trenches. As a result of this research it was found
that soil improvements near bridges did not significantly
reduce low-frequency vibrations. Open and backfilled trenches
could reduce vertical vibrations, but their efficiency seemed
disproportionate to their cost. Yang and Hung 10 carried out a
parametric study of the use of open and backfilled trenches and
elastic foundations for the reduction of train-induced
vibrations and found that all three-wave barriers are suitable
for screening vibrations associated with waves of higher
frequencies. Segol et al. 11 used the FEM along with special
non-reflecting boundaries to study vibration isolation by
means of open and backfilled trenches in stratified soils. They
concluded that the width of the trench is insignificant. Lysmer
and Wass 12 applied the lumped mass method to investigate
stratified soil systems as a means of reducing vibrations.
Aboudi 13 carried out a combined perturbationfinite difference
method analysis of the screening effect due to a thin barrier in

Using trenches to reduce tunnelling vibrations

Rahman Orr

227

a half space, and suggested that an open trench barrier will

yield the best results for screening purposes. Ju 14 applied the
3-D finite element method to simulate soil vibrations due to a
high-speed train moving across a bridge, and investigated the
screening of vibrations by means of open and backfilled
trenches. Ju found that these trenches could not provide a
barrier against low-frequency waves. Yang and Hung3
proposed a 2.5-D finite/infinite element procedure to model
ground vibrations induced by moving loads. Besides the two
in-plane degrees of freedom per node conventionally used for
plane-strain elements, an extra degree of freedom was
introduced to account for out-of-plane wave transmission. The
profile of the half-space was divided into a near field and a
semi-infinite far field. The near field, containing loads and
irregular structures, was modelled by the finite elements, while
the far field, covering the soils extending to infinity, was
modelled by the infinite elements, with due account taken of
the radiation effects produced by the moving loads.
In the past few decades, the boundary element method (BEM)
has also emerged as a very efficient numerical technique for
solving a wide class of engineering problems. This method is
especially well suited for wave propagation problems in soils
involving a semi-infinite domain. A number of researchers
have carried out 2-D and 3-D numerical analyses using the
BEM to investigate the screening of vibrations by means of
open and backfilled trenches, and the use of different types of
foundation to screen the vibrations due to machines 1,4,6,15,16
The BEM is not well suited to the modelling of irregular
geometries. To overcome this problem, several procedures for
coupling finite elements with boundary elements or finite
elements with infinite elements have been proposed by
researchers. 5
It can be observed from the above literature review that
researchers have carried out much work on systems to provide
isolation from machine foundation or high-speed train-induced
vibrations. To the best knowledge of the authors, no work has
been reported in the published geotechnical literature on the
screening by wave barriers of the ground vibrations due to a
TBM. Therefore it is important to study the performance of
wave barriers in reducing the ground vibrations due to a TBM.
In this paper numerical studies, using 3-D finite element (FE)
model coupling with infinite
elements, are presented that
investigate the effectiveness
Finite element
of wave barriers (both open
trenches and backfilled
trenches) in reducing the
ground surface vibration due
to tunnelling activities.

compared with the measured vibrations. 17 The effectiveness of

wave barriers for isolating TBM-induced ground vibrations was
then investigated, and is reported in this paper. Threedimensional finite element models enable account to be taken
of the extent and geometry of each stratum when constructing
a representative numerical model to predict the ground
vibrations due to a TBM. The problem analysed is shown in
Fig. 1. The geometry was idealised by a finite region (Zone I)
and by a semi-infinite far-field region (Zone II). The Abaqus
v. 6.4 finite element software was used to construct the model,
and three-dimensional eight-noded linear brick and reduced
integration elements, C3D8R, were employed for the ground
and the tunnel lining in Zone I. Reduced-integration elements
use fewer integration points in each direction than the fully
integrated elements. These reduce the running time, especially
in three-dimensional finite element model analyses. In Zone II
the soil and rock strata and the tunnel lining in the far field
that extends to infinity, were modelled by the infinite elements,
CIN3D8. In the model the top stratum is Dublin Brown Boulder
Clay, the middle stratum is Dublin Black Boulder Clay, and the
bottom stratum is limestone. The finite element mesh is shown
in Fig. 2. The dimensions of the finite element mesh are 132 m
in the transverse direction, 50 m in the vertical direction, and
40 m in the direction of the tunnel. The thickness of the top
stratum is 2.5 m, the middle stratum is 11 m thick, and the
bottom stratum is 36.5 m thick. The diameter of the tunnel is
12.0 m. The material properties that were used for the different
soil strata in the finite element analyses are given in Table 1.
The geometric parameters of the trench barrier and the
properties of the backfill material are given in Tables 2 and 3
respectively. The values given in these tables show that a range

Zone II, width 10

132
40

Zone I

25

11
365

12 diameter

Limestone

Dublin Boulder Clay

Fig. 1. Typical model of problem (dimensions in m)

Trench
Infinite element

Study point

30 m

2. NUMERICAL
MODELLING
Initially 3-D FE analyses were
carried out to model the
ground vibrations at the
surface due to a TBM. The
accuracy of the model has
been verified by good
agreement obtained when the
predicted vibrations were
228

Fig. 2. Three-dimensional FE model with open trench

Geotechnical Engineering 161 Issue GE5

Using trenches to reduce tunnelling vibrations

Rahman Orr

the actual TBM excitation is

because it ensures
parsimonious handling of the
FE model. The use of an
input pulse saved FE
Density: kg/m3
2000
2300
2710
2400
7800
modelling time of a couple of
Youngs modulus: GPa
0.2
1.0
60
40
200
days for one model run, and
.
.
.
.
.
Poissons ratio,
0 495
0 495
0 25
02
03
for the entire project it saved
Damping: %
2.85
1.75
1.0
1.0
0.5
a couple of months, which is
very significant. A short
Table 1. Values of soil parameters used in models
triangular pulse has a wide
power spectrum (Fig. 4), and
is able to analyse waves ranging from low to high frequency.
of trench dimensions together with a range of backfill material
In the FEM analyses the pulse is applied at the beginning of the
properties were assumed in order to investigate the effects of
calculation, and the vibration energy is damped out during the
different trench geometries and different backfill material
rest of the calculation. As a result, there is no possibility of
properties in reducing the TBM-induced ground vibrations. To
reflection of waves in the model due to this short triangular
examine the sensitivity of the predicted vibrations to variations
pulse. Because of the short input pulse, it is possible to
in the values of the input parameters for the trench dimensions
calculate the transfer function. The transfer function is the
and the properties of the backfill material, a series of
ratio between the output wave and the input wave and hence
parametric studies were carried out. It should be noted that
the magnitude of the input wave is not an important factor.
when the sensitivity of the vibrations to variations in a
The input wave has a larger high-frequency component than
particular parameter was being investigated, the other
parameters were kept constant.
When advancing in the ground, a TBM is a vibrating source,
from which vibrations are transmitted into the ground. The
source of the ground vibration is located at the front of the
TBM face as the TBM cutters impact on the ground during
boring. This was modelled by a dynamic loading system. The
finite elements representing the TBM cutters were loaded by a
short triangular load pulse, which was applied directly to the
elements representing the limestone just in front of the TBM.
Because of this pulse, waves propagated through the soil and
the response at the surface was determined. The analyses were
carried out in the time domain and the characteristics of the
input vibration, shown as vibration velocity plotted against
time and vibration amplitude plotted against frequency, are
shown in Figs 3 and 4 respectively. The output response is
presented in Fig. 5. In Fig. 5 it can be seen that the finiteelement-predicted ground vibrations are free from any
reflections and this indicates that the infinite elements in
Abaqus are working effectively. The main reason for using an
input pulse in the FE model rather than a direct application of

Limestone Concrete

Steel

12

Velocity: mm/s

Dublin black
boulder clay

Top soil
stratum,
(Dublin brown
boulder clay)

10
08
06
04
02
0
0

002

004

008

010

25
20
15
10
05
0
10

Parameter

006
Time: s

Fig. 3. Short triangular pulse (time domain)

Amplitude: (mm/s)/Hz

Parameter

100

1000
Frequency: Hz

10000

Assumed values: m
Fig. 4. Short triangular pulse (frequency domain)
2.5, 5.0, 10, 13.5, 18, 24,
30
0.0, 3.0, 6.0, 9.0, 12, 15
0010

1.5, 3.0, 4.5

Table 2. Geometric parameters of the trench barrier

Density: kg/m3

Shear wave velocity:

m/s

Poissons ratio

30, 60, 90

0.25, 0.33, 0.45

Velocity: mm/s

Depth of trench below ground

surface, D
Location of trench from study
zone, L
Width of backfilled trench, W

0005

0005

1200, 1600, 2000

Table 3. Material properties of backfill material

Geotechnical Engineering 161 Issue GE5

02

04
06
Time: s

08

10

Fig. 5. Predicted time history of vibration at surface level

Using trenches to reduce tunnelling vibrations

Rahman Orr

229

The isolation effect of a trench acting as a barrier has been

calculated using the percentage velocity reduction, defined as

Velocity reduction, %
0
1
vibration velocity of ground surface
B
C
B
C
with barrier
C 3 100
B1

@
vibration velocity of ground surface A
without barrier

3. PARAMETRIC STUDIES OF OPEN TRENCH

Analyses have been carried out to predict the vibrations at a
study point on the ground surface 41 m from the tunnel
centreline when there is an open trench parallel to the tunnel,
as shown in Fig. 2. Parametric studies have been carried out to
investigate the effects of trenches of width W 1.5 m and
with different depths D and distances L from the study point on
the percentage reduction in the velocity of the vibrations at the
surface due to tunnelling, calculated using equation (1). The
effect of different stiffnesses of the top soil stratum has also
been examined.

3.1. Effect of normalised trench depth

In this study it has been assumed that the depth of the open
trench varies between 2.5 and 30.0 m, which is normalised
with respect to the tunnel invert depth of 30.0 m. It should be
noted, however, that the open trench can be considered as a
limit situation, since in reality, depending on the soil condition,
its vertical sides will need special supporting measures when
exceeding a certain depth. 5
The percentage reduction in the vibration velocity at the study
point on the ground surface 41 m from the tunnel centreline
for different values of the normalised trench depth for a trench
at the study point is shown in Fig. 6. It can be seen from Fig. 6
that the percentage reduction of vibration velocity at the study
point increases with increase in the depth of the trench. Based
on these results, it is concluded that the open trench intercepts
a proportion of the waves, and that the velocity of the waves at
the study point is reduced by over 60% for D/30 . 0.8.

3.2. Effect of trench location

The reduction in vibration velocity at the study point,
computed when varying the horizontal distance L between the
study point on the surface and the open trench, has been
plotted in Fig. 7 for a trench with a depth of 10 m. It can be
seen from the graph in Fig. 7 that the predicted percentage
reduction in the vibration velocity decreases with increasing
distance of the trench from the study point, having a
maximum value of 40% when the trench is beside the study
point.
Based on the above results, it is concluded that deep open
trenches are most effective and provide the greatest percentage
reduction of velocity when they are close to the study point.
The reason for this is that a deep trench that is close to the
study point intercepts more waves, making the screening
effects significant, whereas when a trench is far from the study
point a significant number of waves can pass below the trench,
making its screening effect less significant.
3.3. Effect of stiffness of top stratum
The results computed by varying the trench depth for different
stiffness of the top stratum are presented in Fig. 8. As can be
seen in the graphs in Fig. 8, when the depth of the trench is
shallow, the percentage reduction in the vibration velocity is
greater for top stratum stiffnesses that are either low or high
(e.g. 10, 500 and 1000 MPa) than for intermediate top stratum
stiffnesses of 100 and 200 MPa. However, when the depth of
the trench is equal to the depth of the tunnel invert, the
percentage reduction is almost the same (between 60 and 65%)
for all top stratum stiffnesses. The rate of change of velocity
reduction with respect to the trench depth is greatest when the
50

Velocity reduction: %

the measured output signal. High frequencies are filtered out in

the soil of the model.

20
10

3
6
9
12
Distance of trench from study point, L: m

15

Fig. 7. Effect of distance from study point, L, on screening the

effectiveness of open trench

70

Velocity reduction: %

Velocity reduction: %

30

80
60
40
20

60
50
10 MPa
100 MPa
200 MPa (ref.)
500 MPa
1000 MPa

40
30
20
10
0

0
0

025
050
075
Normalised depth of trench, D/30

Fig. 6. Influence of trench depth D on screening the

effectiveness of open trench

230

40

Geotechnical Engineering 161 Issue GE5

100

025
050
075
Normalised depth of the trench: D/30

100

Fig. 8. Influence of open trench depth D with different top

stratum stiffnesses on screening the effectiveness of open
trench

Using trenches to reduce tunnelling vibrations

Rahman Orr

stiffness of the top stratum is 100 or 200 MPa. The range of top
stratum stiffnesses examined corresponds to those ranging
from loose sand or silt (10 MPa) to weak rock (1000 MPa).
Though in practice the stiffness of the surface stratum would
not vary, the graphs indicate that the stiffness of the top
stratum could have a significant influence on the reduction in
the velocity of the TBM-induced vibrations at the study point.
The resonance frequency depends on the stiffness of the soil
and increases with increasing soil stiffness. 18 When the overall
resonance frequency of the soil system and the TBM excitation
frequency coincide, or are very close, there is a maximum
vibration velocity and a minimum reduction by the trench. The
FFT (fast Fourier transform) of the TBM excitation presented in
Fig. 9 shows that the TBM resonance frequency is 14 Hz. The
resonance frequency was obtained from the TBM excitation,
which is presented in Fig. 9. For the situation analysed, the
overall resonance frequency of the soil system is closest to the
resonance frequency of the TBM excitation when the stiffness
of the top stratum is 100 and 200 MPa. The least reduction of
vibration velocity was obtained at stiffnesses of 100 and
200 MPa and when the depth of the trench is shallow, owing to
more waves passing below the trench and because the
resonance frequencies are close, resulting in less vibration
reduction. For shallow trench depths, the maximum reduction
of vibration velocity was obtained at stiffnesses of 10, 500 and
1000 MPa, because at these stiffnesses there is a significant
difference between the resonance frequency of the soil strata
and the input TBM excitation frequency, resulting in a higher
percentage of velocity reduction. In the case of deeper trenches,
however, more waves are intercepted and there is a greater
vibration reduction for all top stratum stiffnesses.

softer than the natural soil. The advantage of a backfilled

trench is that one can achieve larger trench depths with no
need for permanent lateral supports to the vertical sides. 5 The
analyses were carried out for trenches with different depths D
and widths W, for different distances L from the study point,
and for different damping ratios, densities, Poissons ratios and
impedance ratios (IR) of the backfill material. The geometric
parameters of the trenches and the material properties of the
backfill used in the parametric studies are shown in Tables 2
and 3 respectively.
The impedance ratio is the ratio between the mass density
times the shear wave velocity of the backfill material and the
original soil and is expressed as:
IR

rb V b
rs V s

where rb and rs denote the mass densities of the backfill

material and the soil respectively, and Vb and Vs are the shear
wave velocities of the two materials. An IR value less than 1.0
means that the trench barrier is softer than the surrounding soil
and an IR value of 0.1 indicates a very soft backfill material.

Based on the above observations, it is concluded that the

stiffness of the soil is one of the most important factors
controlling the reduction in the vibration velocities. The
amount of velocity reduction increases with increasing depth
of the trench when the resonance frequency of the soil stratum
and the TBM excitation are close. However, the reduction in
the vibration velocity is greatest when the overall resonance
frequency of the soil strata and the TBM excitation frequency
differ greatly, and for these conditions a shallow depth of
trench is better, reducing the vibration velocity by up to
about 50%.

4.1. Effect of normalised depth and location of trench

The effects on the amplitude of the vibration velocity of the
backfilled trench depth D and the location of the backfilled
trench with respect to the horizontal distance L of the study
point from the trench have been investigated. Figs 10 and 11
show the influence of D and L on the vibration velocity
reduction in the case of a trench backfilled with a bentonite
soil mixture having an assumed density of 1200 kg/m3 and
shear wave velocity of 30 m/s in soil with an assumed
Poissons ratio of 0.25, a density of 2000 kg/m3 and a shear
wave velocity of 180 m/s. Using these values in equation (2)
results in a calculated IR value of 0.1, which, as noted above,
indicates a very soft backfill material. From Figs 10 and 11 it
can be observed that the achieved reduction in vibration
velocity increases with increasing depth of the trench D,
whereas it decreases with increasing distance L of the trench
from the study point. The graphs in Figs 10 and 11 show that
the parameters D and L are both important with regard to the
reduction of vibration velocity.

4. PARAMETRIC STUDIES OF BACKFILLED TRENCH

Parametric studies were carried out to investigate the effect of
backfilling the trench with a bentonitesoil mixture, which is

4.2. Effect of trench width

The reduction in the vibration velocity was investigated using
the finite element analyses for different widths of the backfilled
80

Velocity reduction: %

Amplitude: (mm/s)/Hz

1000
800
600
400
200

60
40
20
0
0

0
1

10

100
Frequency: Hz

Fig. 9. Fast Fourier transform of TBM excitation

Geotechnical Engineering 161 Issue GE5

1000

025
050
075
Normalised depth of trench, D/30

100

Fig. 10. Influence of trench depth D on screening effectiveness

of backfilled trench

Using trenches to reduce tunnelling vibrations

Rahman Orr

231

100

40

80

Velocity reduction: %

Velocity reduction: %

50

30
20
10
0
0

3
6
9
12
Distance of trench from study point, L: m

60
40
20

15
0
05

Fig. 11. Effect of distance from study point, L, on screening

effectiveness of backfilled trench

trench varying between 1.5 and 4.5 m. It was found that the
variations in the width of the trench have little effect on the
vibration velocity. The calculated reductions in the vibration
velocity were 40, 41 and 44.2% for trench widths of 1.5, 3.0
and 4.5 m respectively.
4.3. Effect of Poissons ratio of backfill material
The effect of variations in the Poissons ratio of the backfill
material in the trench on the vibration velocity has also been
investigated using the same model. The depth of the trench was
10.0 m. The reductions in the vibration velocity at the study
point obtained from numerical analyses when varying the
Poissons ratio were found to be 40, 40 and 39% for Poissons
ratio values of 0.25, 0.33 and 0.45 respectively. It is concluded,
therefore, that changes in the Poissons ratio effectively have
no significant influence on the vibration velocities.
4.4. Effect of damping ratio of backfill material
Damping is the property of a material to absorb vibration
energy. The ratio between the actual damping of the material
and the critical damping is called the damping ratio. Critical
damping is the smallest amount of damping for which no
oscillation occurs in the vibration system. In order to
investigate the effect of the damping ratio of the backfill
material on the effectiveness of a backfilled trench for
screening vibrations, a trench with a depth of 5.0 m has been
examined, and it has been assumed that the damping ratio
varies between 2 and 8%. The computed reduction in the
vibration velocity was found to be 13.68% for damping ratios
of 2, 5 and 8%. Hence it is concluded that the vibration is
unaffected by the damping ratio, and therefore the effect of
damping can be neglected. This result was obtained because
the analyses were carried out assuming the material was linear
elastic and assuming Rayleigh damping.
4.5. Effect of density of backfill material
The effect of variations in the density of the backfill material
on the effectiveness of the backfilled trench in vibration
screening has also been investigated. The effects of changing
the density of the backfill material as a ratio of the top soil
stratum density are shown in Fig. 12. While the density of the
backfill can be adjusted by controlling the consistency of the
bentonitesoil mixture, the margin for varying the density is
rather narrow in practice: therefore the value of the ratio
ranges only from 0.6 to 1.0 in the present analysis. The depth
of the trench was 10.0 m. As can be seen from the results
plotted in Fig. 12, variations in the density of the backfill
232

Geotechnical Engineering 161 Issue GE5

06
07
08
09
Density of backfill/Density of top stratum

10

Fig. 12. Effect of density ratio on screening effectiveness of

backfilled trench

material have very little effect on the vibration reduction, as

the percentage velocity reduction is approximately constant
with density ratio.
4.6. Effect of shear modulus ratio (Gb /Gs ) of backfill
material
Gb and Gs are the shear moduli of the backfill material in the
trench and the top soil stratum respectively. In order to relate
the shear modulus ratio Gb /Gs to the velocity reduction the
more practical parameter, the impedance ratio IR, presented
above in equation (2) and used by geotechnical engineers for
determining whether a barrier is soft or hard, has been adopted.
When investigating the effect of the shear modulus ratio on the
velocity reduction, the Youngs modulus was used instead of
the shear modulus and was calculated using the equation
3

E 2rV 2 1

The Youngs modulus of the backfill material was considered to

be variable, whereas the Youngs moduli of the different soil
strata were kept constant at the values shown in Table 1. The
density and shear wave velocity values given in Table 3 were
used to calculate the Youngs modulus using equation (3) and
the IR values using equation (2). The IR values were found to
be between 0.1 and 0.3. The depth of the trench was 10 m. The
percentage velocity reduction is plotted against the impedance
ratio IR in Fig. 13. It can be observed from Fig. 13 that the
percentage velocity reduction decreased with increasing IR,
indicating that the softer backfill material, which has a lower
IR value, is more effective in reducing the TBM-induced
ground vibrations than a stiffer backfill material.
5. CONCLUSIONS
The screening of TBM-induced ground vibrations by means of
a wave barrier consisting of open and backfilled trenches in
stratified soil has been studied using 3-D finite elements
coupled with infinite elements. The influences of various
parameters, such as the trench depth, width and location and
the Poissons ratio, density, damping ratio and impedance ratio
of the backfill material in the trench, have been investigated.
The important observations may be summarised as follows.
(a) TBM-induced ground vibrations can be reduced
significantly by means of open and backfilled trenches.
(b) An open trench is much better for reducing ground

Using trenches to reduce tunnelling vibrations

Rahman Orr

Velocity reduction: %

50
40

4.

30
20
10

5.

0
0

01
02
Impedance ratio, IR

03

Fig. 13. Effect of impedance ratio on screening effectiveness of

backfilled trench

6.

7.

(c)

(d )

(e)

(f)

vibrations than a backfilled trench, but it is not possible in

practice to construct stable deep, open trenches. Deep
backfilled trenches may be constructed without the need to
provide any support to the vertical walls of the trenches.
Increasing the distance between a trench and the study
point decreases its reduction effect, owing to the
propagation of waves below the trench, making the
screening effect less effective.
For open trenches, the screening of vibrations is greater
when the resonance frequency of the top soil stratum and
the TBM excitation frequency are not close.
The width of the backfilled trench and the damping ratio,
density ratio and the Poissons ratio of the backfill material
compared with the top soil stratum all have no significant
effects in reducing the ground vibrations.
The use of softer backfill material with respect to the
surrounding soil increases the effectiveness of trenches at
screening vibrations. This is shown by the finding that
reducing the impedance ratio IR of the backfill by using a
softer backfill increases a trenchs screening effect
compared with the use of a stiffer backfill material.

8.

9.

10.

11.

12.

13.

14.
ACKNOWLEDGEMENTS
The work is carried out at the department of civil engineering,
University of Dublin, Trinity College. The project is supported
by the Nishimatsu Construction Co. Ltd and the Geotechnical
Trust Fund of the Institution of Engineers of Ireland. This
support is gratefully acknowledged.

15.

16.
REFERENCES
1. AL-HUSSAINI T. M. and AHMAD S. Design of wave barriers for
reduction of horizontal ground vibration. Journal of
Geotechnical Engineering, ASCE, 1991, 117, No. 4, 616636.
2. HALL L. Simulations and analyses of train-induced ground
vibrations in finite element models. International Journal
of Soil Dynamics and Earthquake Engineering, 2003, 23,
No. 5, 403413.
3. YANG Y. B. and HUNG H. H. A 2.5-D finite/infinite element

17.

18.

approach for modelling viscoelastic bodies subjected to

moving loads. International Journal for Numerical Methods
in Engineering, 2001, 51, 13171336.
AHMAD S. and AL-HUSSAINI T. M. Simplified design for
vibration screening by open and in-filled trenches, Journal
of Geotechnical Engineering, ASCE, 1991, 117, No. 1,
6788.
ADAM M. and ESTORFF O. V. Reduction of train-induced
building vibrations by using open and filled trenches.
Computers and Structures, 2005, 83, No. 1, 1124.
KLEIN R., ANTES H. and HOUEDEC D. L. Efficient 3D
modelling of vibration isolation by open trenches.
Computers and Structures, 1997, 64, No. 14, 809817.
BARKAN D. D. Dynamics of bases and foundations. McGrawHill, New York, 1962.
WOODS R. D. Screening of surface waves in soils. Journal of
Soil Mechanics and Foundation Engineering Division,
ASCE, 1968, 94, No. 4, 951979.
JU S. H. Three-dimensional analyses of wave barriers for
reduction of train-induced vibrations. Journal of
Geotechnical and Geoenvironmental Engineering, ASCE,
2004, 130, No. 7, 740748.
YANG Y. B. and HUNG H. H. A parametric study of wave
barriers for reduction of train-induced vibrations.
International Journal for Numerical Methods in
Engineering, 1997, 40, No. 20, 37293747.
SEGOL G., LEE P. C. and ABEL J. F. Amplitude reduction of
surface waves by trenches. ASCE Journal of Engineering
Mechanics, 1978, 104, No. 3, 621638.
LYSMER J. and WASS G. Shear waves in plane infinite
structures. Journal of Engineering Mechanics, ASCE, 1972,
98, No. 1, 85105.
ABOUDI J. Elastic waves in half-space with thin barrier.
Journal of Engineering Mechanics, ASCE, 1973, 99, No. 1,
6983.
JU S. H. Finite element analyses of wave propagations due
to high-speed train across bridges. International Journal
for Numerical Methods in Engineering, 2002, 54, No. 9,
13911408.
AHMAD S., AL-HUSSAINI T. M. and FISHERMAN K. L. An
Investigation on active isolation of machine foundation by
open trenches. Journal of Geotechnical Engineering, ASCE,
1996, 122, No. 6, 454464.
KATTIS S. E., POLYZOS D. and BESKOS D. E. Modelling of pile
wave barriers by effective trenches and their screening
effectiveness. Soil Dynamics and Earthquake Engineering,
1999, 18, No. 1, 110.
RAHMAN M. E. Prediction of ground vibrations due to a
tunnel boring machine. PhD thesis, Trinity College, Dublin,
2007.
GARDIEN W. and STUIT H. G. Modelling of soil vibrations
from railway tunnels. Journal of Sound and Vibration,
2003, 267, No. 3, 605619.

What do you think?

To comment on this paper, please email up to 500 words to the editor at journals@ice.org.uk
Proceedings journals rely entirely on contributions sent in by civil engineers and related professionals, academics and students. Papers
should be 20005000 words long, with adequate illustrations and references. Please visit www.thomastelford.com/journals for author
guidelines and further details.

Geotechnical Engineering 161 Issue GE5

Using trenches to reduce tunnelling vibrations

Rahman Orr

233