Soil Dynamics and Earthquake Engineering xxx (xxxx) xxx

Contents lists available at ScienceDirect

Soil Dynamics and Earthquake Engineering

journal homepage: http://www.elsevier.com/locate/soildyn

Mitigating high-speed train vibrations with EPS blocks for various

soil conditions
Alexandros Lyratzakis a, Yiannis Tsompanakis a, *, Prodromos N. Psarropoulos b
a
School of Environmental Engineering, Technical University of Crete, Chania, Greece
b
School of Rural and Surveying Engineering, National Technical University of Athens, Greece

A R T I C L E I N F O A B S T R A C T

Keywords: Ground-borne vibrations due to high-speed trains passage strongly depend, apart from the speed of the train, on
High-speed trains the geometry of the railways as well as the properties of the underlying soil layer(s). The main aim of this study is
Train-induced vibrations to investigate the effectiveness of expanded polystyrene (EPS) blocks in mitigating soil vibrations induced on
Railway embankments
railway embankments for different subsoil and railway embankment material conditions. The EPS blocks are
Mitigation measures
Expanded polystyrene (EPS)
placed in suitable locations, either as embankment’s side fill material, or trench filling material, or combination
of the above. An efficient three-dimensional numerical model has been developed -in conjunction with a user-
developed subroutine for applying the moving loads-to accurately calculate the dynamic response of the
coupled embankment-soil model. Four typical soil types - categorized as rock, dense sand with gravels, stiff and
soft clay - are investigated. In addition, the mechanical properties of the embankment material have been altered
to assess to what extend they can affect the HST vibrations.

out to investigate the optimum trench geometry and various filling

1. Introduction
In recent years, train-induced vibrations and the associated detri­
mental effects have become even more pronounced. The higher living
standards of the population worldwide have increased the need for fast,
safe, economical and comfortable traveling, a fact that has led in the
development of new and the upgrading of existing railways in order to
be used by high-speed trains. The effect of the developed vibrations on
the structural safety, as well as the distribution of residents’ activities in
adjacent buildings has become one crucial environmental issue nowa­
days [1]. Train-induced vibrations are propagated in the underlying soil
in the form of surface Rayleigh waves. This phenomenon has drawn the
attention of many researchers and engineers [2–6].
The mitigation of the detrimental effects of these vibrations to rail­
way infrastructure as well as nearby structures is a key issue. Several
authors have focused on the development of mitigation measures that
can be used to minimize railway traffic-induced vibrations [7–9]. The
most common mitigation measure is the implementation of one or more
trenches -open or filled-in the direction of wave propagation, in order to
reflect and absorb the vibrations [10]. The discontinuity between the
soil and the trench’s filling material leads to the mitigation of the ground
vibrations at the trench-soil interface. Several studies have been carried
materials have been proposed.
For instance, Beskos et al. [11] compared filled with open trenches
and concluded that trench’s depth must be greater than 0.6 m for open
trenches. Furthermore, the product of the normalized (by Rayleigh wave
length) depth and width must be more than 1.5 for a trench filled with
concrete. In another relevant study [12], it has been noted that the in­
crease of trench’s width is less effective than its depth. Nevertheless, the
implementation of very narrow trenches allows a significant part of the
vibrations to pass [13]. Feng et al. [14] investigated the efficiency of
buried open trenches covered also with an extending plate and
concluded that the source- and receiver-oriented extending plate is an
effective mitigation approach, especially when the trench is buried at a
depth equal to 0.2–0.3 Rayleigh wavelength. The most effective trenches
for the mitigation of the traffic-induced vibrations are considered to be
the open ones [15,16]. Nevertheless, several issues (e.g., maintaining
their initial geometry intact, avoiding plants and filling with water or
other materials) need to be properly addressed in order to ensure their
continuous operation as initially planned.
Undoubtedly, the efficiency of a filled trench depends on filling
material properties. Several types of filled trenches, such as concrete or
bentonite trenches [17], gas-filled cushions [18], water [19],
sand-rubber mixture [20] and expanded polystyrene (EPS) [13] were

* Corresponding author.
E-mail addresses: alyratzakis@gmail.com (A. Lyratzakis), jt@science.tuc.gr (Y. Tsompanakis), prod@central.ntua.gr (P.N. Psarropoulos).

https://doi.org/10.1016/j.soildyn.2020.106482
Received 10 March 2020; Received in revised form 29 September 2020; Accepted 21 October 2020
0267-7261/© 2020 Elsevier Ltd. All rights reserved.

Please cite this article as: Alexandros Lyratzakis, Soil Dynamics and Earthquake Engineering, https://doi.org/10.1016/j.soildyn.2020.106482
A. Lyratzakis et al.
Abbreviations
3D Three-dimensional
DIN German Institute for Standardization
EPS Expanded polystyrene
FEM Finite element model
HSR High-speed railways
HST High-speed trains
PPV Peak particle velocity
RC Reflection coefficient
proposed over the last decades. Yamohammadi et al. [21] compared
single, double and triple trenches and concluded that the double trench
increases the vibrations’ reduction in contrast with the single trench.
Moreover, double trenches require a lower depth in order to achieve
similar reduction of the vibrations as a single trench [22]. Bo et al. [13]
presented a numerical investigation of the effect of several parameters
(e.g., trenches’ depth and width) in the mitigation of vibrations induced
by a dynamic load, using a trench filled with EPS blocks. According to
this study, larger depth leads to a higher reduction of the vibrations,
while trenches with a slight inclination from the vertical direction are
more efficient than the vertical cuts. Trenches filled with Fontainebleau
sand were investigated and some guidelines regarding the design of
geofoam infilled trenches were proposed by Murillo et al. [23]. More­
over, the effectiveness of the geometrical properties of open and geo­
foam filled trenches were compared via a full-scale field experiment by
Alzawi and EI-Naggar [24].
There exist some alternative approaches for the mitigation of vi­
brations, aiming for instance to improve soil’s stiffness [25] (e.g., deep
vibro-compaction, grouting consolidation [26], lime–cement columns
[27] or deep subsoil mix), which can be quite effective, but are not
commonly applied due to their high cost. In addition, several studies
have concluded that trenches provide higher reduction of the train
induced vibrations, compared to the local soil stiffening beneath the
embankment [28,29]. The implementation of a sheet-pile wall as a stiff
wave barrier has been studied by Dijckmans et al. [30]. The effectiveness
of this mitigation approach is determined by the depth and the stiffness
contrast between the barrier and the surrounding soil. Thompson et al.
[31] examined the mitigation of the HST’s induced vibrations by stiff­
ening track’s subgrade. The implementation of a horizontal wave bar­
rier, known as wave impedance block, has been investigated by several
authors [32–34].
Recently, the authors of the current study have proposed the
implementation of a limited number of expanded polystyrene (EPS)
blocks as partial fill material on the slopes of a typical railway
embankment, which is part of Thalys line that connects Paris with
Brussels (Lyratzakis et al. [35]). It has been demonstrated that this
-relatively low-cost and easily applicable in practice-mitigation measure
can lead to a substantial reduction of the vibrations induced by Thalys
HST. In the sequence, the implementation of the EPS blocks on the
slopes of the embankment with varying geometry (in terms of height and
slope angle) has also been investigated in another study [36].
In any case, it is very important to investigate the effectiveness of
each mitigation measure under different conditions: embankment ge­
ometry, soil properties, train dimensions, and speed. Nonetheless, only a
few studies have considered the impact of subsoil conditions [37,38],
while the efficiency of any open or filled trench strongly depends on the
underlying soil material properties. Jiang et al. [37] firstly examined the
role of surface soil layer stiffness and depth in the developed HST vi­
brations and subsequently assessed the efficiency of four different
mitigation measures (open trench, soft barrier, subgrade stiffening and
wave impeding blocks) for three different real sites and discovered that
the soil conditions (layout and properties) affect the performance of the
2
Soil Dynamics and Earthquake Engineering xxx (xxxx) xxx
mitigation measures by 5–10 dB. Therefore, it was concluded that it is
crucial to determine the role of the soil in a robust and reliable manner.
On the other hand, several authors examined the influence of the
embankment fill material on waves’ propagation. For instance, Feng
et al. [39] investigated the impact of subgrade treatment on HST
induced vibrations, incorporating different subgrade materials and
different subgrade type. Furthermore, Connolly et al. [3] examined the
effect of embankment stiffness on vibration propagation at various dis­
tances from the track and concluded that depending on the stiffness of
the embankment fill material, the embankment may act as a waveguide,
thus trapping waves within its body. Accordingly, as it will be shown in
the sequence, the current study has also focused on this very important
aspect.
Since the underlying soil and the embankment fill soil strongly in­
fluence the developed vibrations, the varying soil conditions are
examined in the current study in order to examine more thoroughly the
efficiency of the aforementioned EPS-based mitigation measure [35,36].
In addition, the proposal of using EPS material as mitigation measure is
further expanded by examining its application as trench filling material,
either alone or in conjunction with the placement of EPS blocks at the
slopes of the railway embankment. For this purpose, the embankment
(with constant geometry) has been examined numerically -utilizing the
finite-element software ABAQUS [40]- for various underlying and
embankment soil conditions with and without the three different miti­
gation approaches. More specifically, four different typical soil types,
classified as rock, dense sand with gravels, stiff and soft clay, have been
examined. In addition, the material properties of the embankment have
been altered to assess to what extend they affect the propagation of HST
vibrations and the effectiveness of the application of EPS blocks for their
mitigation. It should be noted that the current numerical study examines
hypothetical scenarios without rail irregularities, which in reality exist
and contribute to soil vibrations.
2. Description of the numerical model
2.1. General details
Due to the symmetry of the examined embankment along its longi­
tudinal axis, part of the geostructure has been analyzed in the current
investigation, as shown in Fig. 1. Accordingly, Fig. 2 presents the finite/
infinite element model that has been developed, utilizing the commer­
cial software ABAQUS [40], in order to examine the vibrations due to
HST passage. A user-developed VDload subroutine has been imple­
mented in order to simulate HST’s multiple moving loads. This numer­
ical modelling approach is capable of providing reliable results when
there are no rail irregularities, as it was validated by the authors with
field measurements from Thalys HST at North-East of Braffe, Belgium
[35]. The examined Thalys HSR is one of the most popular in Europe and
its maintenance is continuous and of high quality; thus, it has been
Fig. 1. Sketch of the track and the embankment parts.
A. Lyratzakis et al. Soil Dynamics and Earthquake Engineering xxx (xxxx) xxx

Table 2
Parameters of the examined four soil types.
Soil type Density Young’s Poisson’s Damping ξ
(kg/m3) Modulus ratio (-) (%)
Es (MPa)

A Rock 2150 10000 0.15 5

B Dense sand 2100 1000 0.20 5
with gravels
C Stiff clay 2000 600 0.25 5
D Soft clay 1850 170 0.35 5

been modelled as a solid section with dimensions equal to 30 m, 50 m

and 50 m (depth, width and length, respectively). Note that the devel­
oped numerical approach can be applied also for layered soils, as in the
Fig. 2. Finite/infinite 3D numerical model. validation site [35]. Lastly, as shown in Fig. 1, infinite elements have
been added to the edges of the finite element domain in order to prevent
assumed that the track does not have any substantial geometric irreg­ wave reflections.
ularities or defects. Rayleigh damping has been used, in which the damping matrix [C] is
In the present study, four soil types, classified as rock, dense sand analogous to the mass matrix [M] and the stiffness matrix [K] [42]:
with gravels, stiff and soft clay have been examined, while the properties
[C] = a[M] + b[K] (1)
of the silty embankment were initially kept unchanged [35]. Subse­
quently, the embankment material has been replaced, using the same where parameters a and b are properly adjusted to provide a constant
soil as the examined subsoil type, in order to investigate the impact of value of damping ratio (ξ) at a target frequency range (ω1, ω2), via the
the embankment fill material on the propagation of the vibrations. following expressions:
2ξω1 ω2
2.2. Track/embankment modelling α= (2)
ω1 + ω2

Track cross section’s geometrical and mechanical properties have 2ξ

been chosen taking into account the specifications of the International
Union of Railways [41] for high-speed railways (HSR). Fig. 2 displays
the geometry of the examined embankment, which is commonly used in
Paris – Brussels line. The embankment has a height of 5.5 m and 30◦
slope inclination. Fifty meters of track and embankment were simulated
and concrete sleepers were placed at 0.6 m. The rail has a mass of 60
kg/m and has been fixed to the sleeper. The sleepers have been con­
structed on three track layers (ballast, subballast and subgrade).
As mentioned earlier, the track is symmetrical in the direction of the
HST passage about the track centerline; thus, part of the track layers and
the embankment have been modelled in order to reduce the computa­
tional cost. Three-dimensional quadratic cuboid finite elements have
been used for all the track and embankment components, except from
the rails. The element size has been set to be 0.1 m, in order to accurately
model the vibrations induced by Thalys HST. For the simulation of the
UIC60 rail, linear beam elements (with length equal to 0.1 m) along each
axis have been used. Table 1 summarizes the mechanical properties of
the rail track materials.
2.3. Soil modelling
The main aim of the current work is to examine the vibrations
induced by Thalys HST for different scenarios of underlying soil types. In
the present study, four scenarios are examined, in which the railway
embankment is founded on: (i) rock, (ii) dense sand with gravels, (iii)
stiff, and (iv) soft clay. The mechanical properties of the examined soils
are shown in Table 2. The single soil layer under the embankment has
Table 1
Parameters of the examined HSR model.
Track Part Layer Young’s Poisson’s Density (kg/
thickness (m) Modulus (GPa) ratio (-) m3)
b= (3)
ω1 + ω2
In the present investigation, the damping ratio of the examined soils
layers is set equal to 5%. In order to achieve a damping ratio 5% within
the frequency range of 5–60 Hz, parameters α and b have been set equal
to 4.901s-1 and 0.0003s, respectively.
2.4. Modelling of train moving load
The numerical model has been developed considering the passage of
Thalys HST, moving along the railway at a constant speed of 284 km/h.
Details for Thalys are available in the literature [43,44]. In order to
avoid the increased computational cost of simulating the whole train, its
moving loads have been assigned in the numerical model via a user
developed VDload subroutine. In this manner, the multiple loads are
moving on the rail along the track at the locations of bogies axles, while
a small time-step (Δt = 1.3 × 10− 6 s) has been used in order to ensure
stability and accuracy of the numerical calculations. According to the
colleagues who collected the field data that have been used to validate
the developed numerical model [35], rail irregularities were removed
one week before the measurements [45,46]. Therefore, all numerical
calculations have been performed considering uniform track geometry
conditions. Consequently, the developed numerical modelling approach
is reliable for cases without rail irregularities.
Thalys HST has a total length of 200.18 m and consists of two lo­
comotives and eight carriages: (a) two Y230A locomotives, (b) two
Y237A side carriages, and (c) six Y237B central carriages. Fig. 3 depicts
the loads acting on the rails which replicate Thalys HST passage.
Furthermore, Fig. 3 summarizes the dimensions of each bogie type, the
distances between the axles and the bogies as well as the carriage length.
2.5. Mitigation scenarios

Rail – 210 0.25 7900 As it has been earlier mentioned, several mitigation techniques have
Sleepers – 30 0.4 2400 been investigated in order to reduce the developed vibrations due to
Ballast 0.3 0.1 0.35 1800 HST passage. In the present study, the focus is given on the application
Subballast 0.2 0.3 0.35 2200 of EPS blocks in three different configurations (Models 2, 3 and 4, as
Subgrade 0.5 0.127 0.35 2100
described in Fig. 4), which are compared with the initial railway

3
A. Lyratzakis et al. Soil Dynamics and Earthquake Engineering xxx (xxxx) xxx

Fig. 3. Thalys train axle loads (adopted from Refs. [43]) and bogies dimensions (adopted from Ref. [44]).

Fig. 4. HSR embankment cross-section: (a) Model 1 refers to the soil embankment, (b) Model 2 refers to the mitigated embankment with EPS blocks at the slopes, (c)
Model 3 refers to the mitigated embankment with an EPS-filled trench, (d) Model 4 refers to the embankment with both mitigation measures.

embankment (Model 1).
Over the last decades EPS has been applied in many engineering
applications, including normal speed railway embankments [47,48].
Due to its low weight, geofoam fill can be easily implemented without
the use of large mechanical equipment at zones with difficult access,
reducing the construction time and cost. Apart from its practicality and
ease of use, it should also be mentioned that EPS geofoam is a relatively
low-cost material. Nevertheless, the cost of EPS material varies,
depending on material type, dimensions of the blocks, required quan­
tities, transportation costs, etc. In addition, covering soil can be placed
directly on the geofoam blocks, eliminating the need for expensive and
time-consuming forming and casting. Lastly, if it is well-constructed and
protected EPS has durability and low maintenance cost.
A dense EPS material, i.e., EPS46, has been used, which is the stiffer
material among the seven types of EPS described in ASTM D6817 [49]),
since it was proven more efficient compared to softer EPS materials for
mitigating HST vibrations [35]. EPS46 is a low-density material (45.7
kg/m3) with relatively high Young’s Modulus (12800 kPa) and
extremely low Poisson ratio (0.05).
According to the first mitigation scenario (Model 2), a limited
number of EPS46 blocks have been placed at the sides of the railway
embankment that were covered with a very thin layer of soil. The
optimal configuration of EPS blocks has been investigated by the authors
in recent papers [35,36] Accordingly, a few EPS blocks (with dimensions
equal to 1 m) have been placed as illustrated in Fig. 4b. The efficiency of
this mitigation measure has been investigated by the authors for a real
embankment that is founded on silty soil layers [35]. Note that alter­
native EPS configurations, i.e., to use an EPS embankment (fully or
partially made by EPS blocks) had been examined [35]. Although higher
reductions on HST induced vibrations were achieved, these EPS schemes
are not suggested, since apart from their increased cost, they also
resulted in higher displacements that can cause derailments, since HST
are much more sensitive to settlements compared to normal speed
trains. This cost-efficient technique has managed to reduce substantially
the vibration levels by altering the wave propagation within the em­
bankment’s body due to the contrast between the properties of the two
materials (soil and EPS) which leads to high values of reflection coeffi­
cient (RC), which strongly affects the vibration patterns. This is analo­
gous with the discussion in Section 5 related to the wave propagation
under the embankment. The placement of the EPS blocks at the slopes

4
A. Lyratzakis et al.
actually creates a waveguide, which leads the waves to a certain path
and does not allow them to be freely transmitted along the slopes and
further. In addition, the EPS blocks are slightly buried under 0.2 m of
soil, hence, this soil cover may play a role in wave propagation at the
surface similar to the case of a buried trench [14].
Secondly, the use of a trench with EPS blocks as filling material has
also been examined (Model 3). As presented in the introductory section,
trenches are the most commonly used mitigation measure for HST
induced vibrations. According to several authors (e.g. Ref. [50]), the
most effective type of trench, is the open-air trench, i.e., without any
filling material. On the other hand, problems related to stabilization and
maintenance of open-air trenches, has led to the application of several
filling materials, as discussed in the literature review. In the present
study, the use of EPS46 blocks as trench filling material has also been
examined. After a preliminary investigation regarding the optimal di­
mensions and location, a single trench has been placed at 14 m from the
middle of the railway embankment, having 50 cm width and 3 m depth
(see Fig. 4c). Finally, it was assumed that using EPS blocks at the slopes
of the embankment and as filling material in the trench could further
decrease the HST induced vibrations. Hence, the hybrid mitigation
scenario, depicted in Fig. 4d (Model 4), has also been studied.
3. Numerical results
In the sequence, a comparison between the conventional and the
three mitigated embankments is presented for the four hypothetical soil
scenarios in terms of velocity time-histories and Fourier spectra. Note
that the scales in the plots are kept constant in order to illustrate that -as
expected-the HST vibrations are increasing as underlying soil gets softer.
The results are given for two characteristic locations, i.e., d = 15 m and
d = 35 m from the track, to illustrate the impact of the proposed miti­
gation measures both near-field and far-field from the track. These lo­
cations are often used in field measurements [2,45] and were also used
in the validation of the applied numerical modelling approach [35].
Fig. 5. Comparison of soil and EPS-retrofitted embankments on soft clay in terms of vertical velocity time-histories: (a) Model 1, (b) Model 2, (c) Model 3, (d)
Model 4.
5
Soil Dynamics and Earthquake Engineering xxx (xxxx) xxx
3.1. Soft clay
A soft soil, classified as soft clay, is the first scenario that has been
examined in the present study. The implementation of a limited number
of EPS blocks (Model 2) in such soils plays a very beneficial role in the
mitigation of the increased HST vibrations. More specifically, the
maximum vertical velocity has been decreased from 0.8 mm/s to 0.35
mm/s at the near-field location (at 15 m). The same trend is observed for
Model 3, where the maximum vertical velocity has been decreased from
0.8 mm/s to 0.5 mm/s (the reductions is less than the corresponding one
for Model 2). Lastly, as it can be noticed from Fig. 5d, the decrease of the
vibrations levels is even higher in the case of Model 4.
The most critical vibrations are the vibrations in the low-frequency
range [2]. Accordingly, Fig. 6 presents the impact of the examined
mitigation measures on the velocity spectra at the low frequency range
(0–60 Hz). In the case of Model 1, three frequencies (21.4 Hz, 25.4 Hz
and 29.3 Hz) can be spotted in Fig. 6a. The plot in Fig. 6b displays the
improvement of the soil response in the whole low-frequency range in
the case of Model 2. For example, the peak of the most dominant fre­
quency (29.3 Hz) is reduced from 0.2 mm/s to 0.0 mm/s. In contrast,
Fig. 6c illustrates that the implementation of the EPS-filled trench
(Model 3) has led to a slight reduction of the vibrations’ peaks in the
whole examined frequency range (0–40 Hz). It is obvious from Fig. 6d
that regarding near-field vibrations (i.e., at 15 m), Model 4 has led to a
remarkable improvement of the soil response. The same observation can
be drawn at 35 m, where the examined measures mitigate effectively the
HST vibrations.
3.2. Stiff clay
As shown in Fig. 7, the beneficial influence of the mitigation with
EPS46 blocks is also significant for stiff clay. In this case the Young’s
modulus of the soil is 3.5 times higher than in the previous scenario. This
results in lower vertical velocities compared to soft clay, e.g., at 15 m the
A. Lyratzakis et al. Soil Dynamics and Earthquake Engineering xxx (xxxx) xxx

Fig. 6. Comparison of soil and EPS-retrofitted embankments on soft clay in terms of vertical velocity Fourier spectra: (a) Model 1, (b) Model 2, (c) Model 3, (d)
Model 4.

Fig. 7. Comparison of soil and EPS-retrofitted embankments on stiff clay in terms of vertical velocity time-histories: (a) Model 1, (b) Model 2, (c) Model 3, (d)
Model 4.

maximum vertical velocity is equal to 0.4 mm/s (see Fig. 7a), almost half
than previously (see Fig. 5a). Fig. 7b depicts the significant vibrations
mitigation in the case of Model 2. More specifically, the maximum
vertical velocity is close to 0.19 mm/s, 50% lower than for Model 1.
Similarly, the vertical velocities have also been reduced in Model 3, in
which the maximum vertical velocity is 0.18 mm/s at 15 m. Τhe same
observations can be made at 35 m, in this case all the examined miti­
gation measures contribute to about 50% reduction of the vertical ve­
locity. For example, the maximum vertical velocity is decreased from
0.11 mm/s to 0.06 mm/s after the implementation of the EPS-filled
trench (Model 2).
Fig. 8 displays the comparison of the Fourier spectra for the four

6
A. Lyratzakis et al. Soil Dynamics and Earthquake Engineering xxx (xxxx) xxx

Fig. 8. Comparison of soil and EPS-retrofitted embankments on stiff clay in terms of vertical velocity Fourier spectra: (a) Model 1, (b) Model 2, (c) Model 3, (d)
Model 4.

examined models on stiff clay. As it is clearly illustrated, the most critical
frequencies remain the same as in the case of soft clay, while the peak
values are decreased. For instance, the peak at the frequency of 25.4 Hz
is equal to 0.12 mm/s at 15 m, about 50% the peak at the same fre­
quency for soft clay. The proposed measures are efficiently mitigating
the HST vibrations within the low frequency range. It can be noticed that
Model 3 is more effective in the frequency range from 40 Hz to 60 Hz,
while Model 2 reduces significantly the vibrations in the lower fre­
quency range (0–40 Hz). Fig. 8d illustrates that Model 4 leads to the
maximum mitigation of the vibrations. Similarly, at 35 m, the proposed
mitigation measures contribute to the spectral velocity reduction in the
whole low frequency range.

Fig. 9. Comparison of soil and EPS-retrofitted embankments on dense sand with gravels in terms of vertical velocity time-histories: (a) Model 1, (b) Model 2, (c)
Model 3, (d) Model 4.

7
A. Lyratzakis et al.
3.3. Dense sand with gravels
Fig. 9 illustrates the vertical velocity time-histories when the
embankment is located on dense sand with gravels, where there is a
significant decrease on the vibrations compared to stiff clay, while the
beneficial impact of EPS is not so pronounced as in the previous cases of
softer soils. More specifically, the maximum vertical velocity of the soil
embankment model at 15 m is equal to 0.2 mm/s. Due to the higher
weight of the first and the last locomotive bogies, higher near-field soil
response has been observed when these carriages are passing. As it is
clearly depicted in Fig. 9b, Model 2 has slightly reduced the vibrations
induced by Thalys HST at both examined locations. Furthermore, Model
3 results in a higher decrease of the developed vibrations. Lastly, Model
4 leads to the optimum mitigation of the vibrations, since the maximum
vertical velocity is less than 0.16 mm/s (see Fig. 9d).
Fig. 10 presents the comparison of the Fourier spectra for the four
embankments on dense sand with gravels. The main vibrations peaks at
25.4 Hz and 29.3 Hz in the case of the soil embankment (Model 1) at 15
m are equal to 0.06 mm/s and 0.08 mm/s, respectively. Furthermore,
two lower peaks have been observed at 21.4 Hz and 33.8 Hz. Model 2
has contributed to the reduction of the most dominant peak at 29.3 Hz
from 0.08 mm/s to 0.03 mm/s. In addition, the vibrations peak at 25.4
Hz has been reduced to 0.04 mm/s. On the other hand, the peak at 21.4
Hz has been slightly increased. Furthermore, in Model 3, the vibrations
peaks at 15 m are slightly reduced. Model 4 exhibits the higher reduction
of the vibrations peak at 29.3 Hz, although the reduction of the vibra­
tions level at the other vibrations peaks is marginal. Similar observations
are made for the examined models at the far-field location (35 m).
3.4. Rock
In the last soil scenario, the passage of Thalys HST over a very stiff
soil layer has been examined. The examined soil has been classified as
Fig. 10. Comparison of soil and EPS-retrofitted embankments on dense sand with gravels in terms of vertical velocity Fourier spectra: (a) Model 1, (b) Model 2, (c)
Model 3, (d) Model 4.
8
Soil Dynamics and Earthquake Engineering xxx (xxxx) xxx
rock and its properties are shown in Table 2. In this case, the vibrations
induced by Thalys HST are limited to very low levels (i.e., less than 0.06
mm/s) and they are not causing any problems to the train and the
embankment. Therefore, the application of any mitigation measure is
not required/recommended and the focus of the parametric study is
concentrated on the other three cases.
4. Mitigation measures efficiency
The efficiency of the examined mitigation measures, i.e., Models 2, 3
and 4, which are compared with the reference case of the initial railway
embankment (Model 1), is summarized in Fig. 11. More specifically, the
attenuation in terms of peak particle velocity (PPV) for dense sand with
gravels, stiff and soft clay at 15 m is illustrated in this chart. In all soil
types the proposed mitigation measures have led to the reduction of the
vibrations induced by Thalys HST passage. It is clearly depicted that the
most effective mitigation measure is Model 4, since the attenuation is
close to 65% for the soft soil case, where it is noteworthy to observe that
the side-fill (Model 2) is much more efficient compared to EPS-filled
trench (Model 3).
Fig. 12 illustrates the effectiveness of the examined mitigation ap­
proaches in terms of PPV, which is calculated as follows:
⃒√⃒̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
⃒ ⃒
PPV = max⃒ v2x (t) + v2y (t) + v2z (t)⃒ (4)
where vx(t), vy(t) and vz(t) are the longitudinal, the transversal and the
vertical velocity time-histories, while the vertical velocity is the domi­
nant component of PPV. More specifically, the PPV in the modelling
scenario for dense sand with gravels without any mitigation measure
(Model 1) has been compared with the Model 4 of the mitigated
embankment laid on soft clay. As it is clearly shown in the plot, the PPV
at varying distances (from 15 m to 35 m) in the case of soft clay after the
combined EPS mitigation (Model 4) is comparable with the
A. Lyratzakis et al. Soil Dynamics and Earthquake Engineering xxx (xxxx) xxx

Fig. 11. Attenuation of PPV for soil type B, C and D at d = 15 m.

Fig. 12. Comparison of PPV for the soil embankment (Model 1) on dense sand with gravels with the retrofitted embankment (Model 4) on soft clay.

corresponding ones of the soil embankment (Model 1) that is located on associated unit is m/s. This indicator provides an alternative way of
dense sand with gravels. In other words, by adopting this cost-effective assessing the effect of HST induced vibrations. Vehicle vibratory
intervention, the geotechnical conditions are notably upgraded and the discomfort is then assessed by comparing the maximum level of KBF,max
HST can operate in a better and safer manner, even when it has to cross with a threshold limit for residential areas, which is set equal to 15
areas with soft soil layers. mm/s, while KBF,max is derived as a function of PPV [52]:
Since the vibration is often non-stationary, the German Institute for √̅̅̅
Standardization (DIN) proposed the implementation of a running root 2 PPV
KBF,max = √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
( )2̅ (7)
mean square to the velocity signal. In particular, according to DIN 4150- 2
5.6
1+ f
2 [51], a weighted time-averaged signal (KBF) is computed using the
following expression [2,50]:
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ Consequently, Fig. 13 depicts the maximum level of KBF at six
√ ∫τ
√1 equally spaced observation points (15 m–35 m) from the track. It is
√
(5) evident that KBF,max has been decreased with distance from the track. In
t− ξ
KBF (t) = √ KB2 (ξ)e− τ dξ
τ
O addition, the KBF,max levels have been reduced after the implementation
of all the examined mitigation measures for all soil scenarios. More
where the weighted velocity signal KB(t) is derived by flowing the initial specifically, as it is clearly illustrated in Fig. 13a, the hybrid mitigation
velocity signal via the high-pass filter [2,50]: technique is the most efficient one for all soil types.
As displayed in Fig. 13c, in the case of dense sand with gravels, the
1
HKB (f ) = √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
( )̅ (6) KBF,max of Model 1 is lower than the threshold value of 15 mm/s for
distances between 17 m and 35 m. Hence, it can be deduced that the
2
5.6
1+ f
implementation of any mitigation is not necessary for such soil condi­
tions, although the implementation of all the examined mitigation
while the integration time, τ, is set equal to 0.125s [2,50]. It should be measures contributes to a substantial reduction of KBF,max. The benefi­
mentioned that while no specific unit is defined in DIN standards, the cial effect of the proposed mitigation approaches is more evident in the

9
A. Lyratzakis et al.
Fig. 13. Comparison of soil and EPS-retrofitted embankments at increasing distance from the track in terms of KBF,max: (a) soft clay, (b) stiff clay, (c) dense sand
with gravels.
case of soil stiff clay, where initially KBF,max values are higher than the
DIN limit for Model 1 for distances between 15 m and 25 m. In contrast,
the implementation of all EPS-based mitigation types has substantially
decreased KBF,max values below the limit of 15 mm/s for almost all the
examined distances.
Finally, as it was expected, the highest values of KBF,max have been
observed for soft soil soft clay, where KBF,max values are much higher (up
to 4 times at near-field distances) than the DIN limit for all locations.
Model 3 has reduced the values of KBF,max, but they still remain above
15 mm/s for most of the examined points (<29 m from the track). On the
other hand, Model 2 (and subsequently Model 4) has managed to reduce
effectively the KBF,max levels in all distances to acceptable levels, apart
from near-field (15 m), where it remains slightly higher than the limit.
Furthermore, the insertion loss is illustrated for the three examined
10
Soil Dynamics and Earthquake Engineering xxx (xxxx) xxx
mitigation approaches in Fig. 14, The calculations have been carried out
at the centre frequency of each one-third octave from 1.25 to 40 Hz and
the results were averaged to one-third octave bands. The insertion loss
(IL) is computed using the following expression [53]:
Vrms,soil
IL = 20log 10 (8)
Vrms,0,mit
where:
vrms, soil: root mean square of the spectral velocity of each 1/3 octave
band of Model 1,
vrms,mit: root mean square of the spectral velocity of each 1/3 octave
band for each of the examined mitigation approaches shown in Fig. 4.
As it is illustrated in the plots of Fig. 14, the insertion loss is higher
A. Lyratzakis et al.
Fig. 14. Comparison of insertion loss at 15 m from the track for the three mitigation approaches for: (a) soft clay, (b) stiff clay, (c) dense sand with gravels.
for the softer soils for all the examined mitigation scenarios in the fre­
quency range of interest (20–35 Hz). It should be noted that according to
Fourier spectral analysis, the dominant frequencies for all the examined
soil layers are located at the 1/3 octave bands with centre frequencies at
25 Hz and 31.5 Hz, in which the insertion loss curve is always positive.
In particular, when the embankment is constructed on soft clay
(Fig. 14a), Model 4 reduces the vibration levels up to 5 db for the whole
examined frequency range, while Model 2 and Model 3 are effective only
for frequencies higher than 10 Hz. When the embankment is located on
stiff clay (Fig. 14b), Model 2 exhibits a better performance in most fre­
quencies. On the other hand, in the case of the stiffer soil (dense sand
with gravels) the role of all mitigation measures is not so beneficial. It is
evident from Fig. 14c that the addition of the trench (i.e., Model 3),
increases the vibration levels at the low octave bands from 1.25 Hz to 8
Hz. In contrast, Model 4, reduces the vibration levels for all the exam­
ined centre frequencies, except from 20 Hz, where the insertion loss is
negative both for Model 2 and Model 4.
It should be mentioned that in existing HSR lines, residential build­
ings are commonly located at distances greater than 15 m from the track,
thus, it is reasonable to examine the IL curves at distances where
buildings are usually cited. Nonetheless, in the future upgrading of
existing normal speed railways, which pass through urban environ­
ments, the disturbance of adjacent buildings in much lower distances
will become a common problem. Fig. 15 depicts the IL curves at 35 m
from the track, where the combined approach (Model 4) is clearly the
most effective mitigation scheme. More specifically, this scheme led to
the reduction of the vibrations level from 3 to 16 dB for all subsoil
conditions within the whole examined frequency range. Furthermore, in
the case of soft clay (Fig. 15a), the higher IL is observed at the octave
bands with centre frequencies higher than 20 Hz, which are the most
dominant ones according to the Fourier spectra (see Fig. 6). Similarly,
the IL at the octave band with centre frequency 20 Hz is higher than 15
dB when the embankment is laid on stiff clay (Fig. 15b). Lastly, in the
Fig. 15. Comparison of insertion loss at 35 m from the track for the three mitigation approaches for: (a) soft clay, (b) stiff clay, (c) dense sand with gravels.
11
Soil Dynamics and Earthquake Engineering xxx (xxxx) xxx
case of dense sand with gravels (Fig. 15c), the higher IL is again observed
at the octave band with centre frequency 20 Hz.
5. Impact of embankment fill material
Initially the embankment has been constructed by the same silty soil
-as described in Section 2.2- for all the examined soil scenarios in order
to investigate the subsoil response for the same conditions. However, in
engineering practice, the surface soil is commonly used as railway
embankment fill material. Hence, the vibrations during the passage of
Thalys have been calculated when the subsoil (dense sand with gravels,
stiff or soft clay) has been used as embankment fill material and the
corresponding models are presented in Table 3.
Fig. 16a illustrates the vertical velocity Fourier spectra of Model 1B.
In this case, the implementation of the hybrid mitigation (Model 4B) has
led to the reduction of the vibrations’ peaks within the whole frequency
range (0–60 Hz). More specifically, the two dominant peaks at 29.3 Hz
and 33 Hz have been reduced from 0.13 mm/s and 0.1 mm/s to 0.06
mm/s and 0.04 mm/s, respectively.
The same observation can be made when the embankment has been
constructed with soil obtained from the underlying stiff clay (see
Fig. 16b). The most dominant peak is at 29.3 Hz and it is decreased from
0.18 mm/s (Model 1C) to 0.06 mm/s (Model 4C). Furthermore, all the
Table 3
Models with different embankment soil material.
Embankment fill Regular Embankment with hybrid EPS
material embankment mitigation
Dense sand with Model 1B Model 4B
gravels
Stiff clay Model 1C Model 4C
Soft clay Model 1D Model 4D
A. Lyratzakis et al. Soil Dynamics and Earthquake Engineering xxx (xxxx) xxx

Fig. 16. Comparison of soil and EPS-retrofitted embankments in terms of vertical velocity Fourier spectra for: (a) dense sand with gravels, (b) stiff clay, (b) soft clay.

other peaks between 0 and 60 Hz have been substantially reduced. As it

is clearly illustrated in Fig. 16c, the vibrations peaks are much higher for
soil soft clay compared to the other two soil types. In this scenario, the
most critical frequency range is between 0 and 40 Hz. The most domi­
nant peaks at 25.4 Hz and 29.3 Hz have been significantly reduced by
applying the proposed hybrid mitigation technique, similarly to the
behavior observed for the silty embankment for soil soft clay (refer to
Fig. 8 and Section 3.1).
Fig. 17 presents the IL curves of the three examined backfill em­
bankments. The proposed mitigation approach contributes to the
reduction of the vibrations level from 1 to 17 dB for all the examined
octave bands. In the case of soft clay, the IL is higher than 10 dB, for all
octave bands with centre frequencies higher than 20 Hz, where the vi­
brations peaks are observed in the Fourier spectra. The IL is lower in the
cases of backfill embankment constructed from stiff clay or dense sand
with gravels, although the vibrations level is again considerably
reduced.
Fig. 18 illustrates how the KBF,max indicator evolves with increasing
distance from the rail track for the examined cases. Note that the KBF,max
levels are shown both for the initial silty soil and the embankment with
dense sand with gravels, stiff or soft clay filling material. As it can be
seen in Fig. 18a, the embankment material can significantly influence
the HST induced vibrations. The KBF,max for the embankment con­
structed from dense sand with gravels (Model 1B) is increased compared
to the initial silty embankment (Model 1) and it exceeds the DIN limit for
distances between 15 m and 25 m. On the other hand, this indicator is Fig. 17. Comparison of insertion loss curves of the backfill embankment of the
less than the threshold value at all distances for the silty embankment. three examined subsoils.
Nonetheless, the hybrid mitigation measure has a beneficial effect on the
vibrations at all points, regardless of the embankment’s fill material. In the last scenario, the properties of the soft clay embankment (Model
Along the same lines, KBF,max values for stiff clay are displayed in 1D) and the initial silty embankment (Model 1) are quite similar. Hence,
Fig. 18b. In this case, the KBF,max values for Model 1 are lower than the change of the embankment’s fill material has a minor impact on the
Model 1C and exceed the limit from low to medium distances from the KBF,max indicator values, which are remarkably reduced in both cases
track. Furthermore, the proposed mitigation measure successfully when applying the hybrid mitigation measure.
reduced the KBF,max level, regardless of the embankment’s fill material. According to the previous discussion, it can be deduced that the

12
A. Lyratzakis et al. Soil Dynamics and Earthquake Engineering xxx (xxxx) xxx

Fig. 18. Comparison of KBF,max for different embankent filling material at increasing distance from the track for: (a) dense sand with gravels, (b) stiff clay, (b)
soft clay.

contrast between the subsoil and embankment material properties

strongly influences the developed vibrations levels due to HST passage. Table 4
Reflection coefficient of the initial soil embankment (Model 1) for different
The transmissions and reflections of the waves along the interface be­
subsoil conditions.
tween the embankment and the underlying soil material can be deter­
mined via the so-called reflection coefficient (RC), which is calculated as Soil type Density (kg/m3) vs,30 (m/sec) RC (%)

follows [3]: B Dense sand with gravels 2100 580 62

C Stiff clay 2000 240 26
ρS ⋅vS,S − ρE ⋅vS,E D Soft clay 1850 180 8
RC = (9)
ρS ⋅vS,S + ρE ⋅vS,E

where ρS is the density of the subsoil, vS,S is the shear wave velocity of
the subsoil, ρE is the density of the embankment fill material, and vS,E is
the shear wave velocity of the embankment.
As presented in Table 4, when the silty embankment (i.e., Model 1,
with ρE = 200 kg/m3 and vS,E = 142 m/s) is founded on stiff soil, i.e.,
dense sand with gravels, then the interface has a high RC value equal to
62%. This causes wave energy to be reflected from the subsoil surface
back into the embankment. In other words, the wave energy is trapped
within the geostructure body, i.e., the embankment acts as a waveguide
[3]. On the other hand, the value of RC is significantly lower for stiff
clay, as it is only 26%. Lastly, the RC in the case of soft clay is very low
(8%) due to the close shear wave velocity of the silty embankment and
the underlying soil materials, thus, the HST vibrations are more easily
propagated from the embankment to the surrounding soil.

13
A. Lyratzakis et al.
6. Conclusions
In the present study a reliable three-dimensional numerical model
has been used in order to predict the developed vibrations due to high-
speed train passage. The numerical model has been used to investigate
the impact of underlying soil conditions, as well as the embankment’s
fill material on the induced vibrations. Furthermore, the efficiency of
three vibrations mitigation measures by applying EPS blocks in three
different configurations has been investigated. The presented applica­
tion of EPS as a mitigation measure in HSR embankments has several
advantages compared to other approaches (e.g., subgrade stiffening)
that can lead to overall project cost and time savings. It has been shown
that this relatively low-cost approach is more effective in the case of soft
subsoil conditions, where the placement of EPS blocks in critical sites is
deemed necessary. On the other hand, the application of EPS is not
suggested in the case of stiffer soil conditions.
Based on the presented results, the main conclusions that can be
drawn are the following:
− The vibrations induced by HST are decreased when the soil layer is
stiffer, especially in the case of rock are marginal. Hence, mitigation
measures are not required for firm soil conditions.
− The impact of the implementation of -a limited number-of EPS blocks
at the slopes of the embankment led to the successful mitigation of
the vibrations when the embankment is founded on dense sand with
gravels, stiff or soft clay. More specifically, this mitigation measure
was very effective in the frequency range of 0–40 Hz.
− The use of filled trench with EPS blocks can also lead to a decrease of
the vibrations for dense sand with gravels, stiff or soft clay, mainly
within the frequency range between 30 Hz and 50 Hz.
− The optimal mitigation measure was the implementation of both the
EPS-filled trench and EPS blocks at the embankment’s slopes. In this
case, the vibrations have been decreased within the whole low fre­
quency range.
− The embankment material plays also an important role in the
ground-borne vibrations, especially for stiffer soils. It has been found
that depending on the difference between the subsoil and the
embankment soil properties, the waves may be trapped within the
geostructure. Nonetheless, the proposed hybrid mitigation measure
reduces the vibrations, regardless of the embankment fill material.
Certainly, this complex dynamic phenomenon should be further
investigated and the proposed mitigation measures should be applied in
different conditions, such as train speed, railway geometries (e.g., cut­
tings), etc. In addition, track irregularities, which also play a crucial role
in soil vibrations, should also be included in the numerical models.
CRediT authorship contribution statement
Alexandros Lyratzakis: Conceptualization, Investigation, Method­
ology, Software, Validation, Data curation, Writing - original draft,
preparation. Yiannis Tsompanakis: Project administration, Supervi­
sion, Conceptualization, Investigation, Methodology, Writing -
Reviewing and Editing. Prodromos N. Psarropoulos: Conceptualiza­
tion, Investigation, Methodology, Writing - Reviewing and Editing.
Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
Acknowledgments
This research is co-financed by Greece and the European Union
(European Social Fund - ESF) through the Operational Programme
14
Soil Dynamics and Earthquake Engineering xxx (xxxx) xxx
“Human Resources Development, Education and Lifelong Learning
2014–2020” in the context of the project “Strengthening Human Re­
sources Research Potential via Doctorate Research – 2nd Cycle” (MIS
5000432).
References
[1] Yao J, Zhao R, Zhang N, Yang D. Vibration isolation effect study of in-filled trench
barriers to train-induced environmental vibrations. Soil Dynam Earthq Eng 2019;
125:105741. https://doi.org/10.1016/j.soildyn.2019.105741.
[2] Kouroussis G, Verlinden O, Conti C. Free field vibrations caused by high-speed
lines: measurement and time domain simulation. Soil Dynam Earthq Eng 2011;31:
692–707. https://doi.org/10.1016/j.soildyn.2010.11.012.
[3] Connolly D, Giannopoulos A, Forde MC. Numerical modelling of ground borne
vibrations from high speed rail lines on embankments. Soil Dynam Earthq Eng
2013;46:13–9. https://doi.org/10.1016/j.soildyn.2012.12.003.
[4] Balendra T, Chua KH, Lo KW, Lee SL. Steady-state vibration of subway-soil-
building system. J Eng Mech 1989;115:145–62. https://doi.org/10.1061/(ASCE)
0733-9399(1989)115:1(145).
[5] Mezeh R, Mroueh H, Hosseingholian M, Sadek M. New approach for the assessment
of train/track/foundation dynamics using in-situ measurements of high-speed train
induced vibrations. Soil Dynam Earthq Eng 2019;116:50–9. https://doi.org/
10.1016/j.soildyn.2018.10.024.
[6] Ren X, Wu J, Tang Y, Yang J. Propagation and attenuation characteristics of the
vibration in soft soil foundations induced by high-speed trains. Soil Dynam Earthq
Eng 2019;117:374–83. https://doi.org/10.1016/j.soildyn.2018.11.004.
[7] Connolly DP, Kouroussis G, Laghrouche O, Ho CL, Forde MC. Benchmarking
railway vibrations - track, vehicle, ground and building effects. Construct Build
Mater 2015;92:64–81. https://doi.org/10.1016/j.conbuildmat.2014.07.042.
[8] Karlström A, Boström A. Efficiency of trenches along railways for trains moving at
sub- or supersonic speeds. Soil Dynam Earthq Eng 2007;27:625–41. https://doi.
org/10.1016/j.soildyn.2006.12.005.
[9] Dijckmans A, Coulier P, Jiang J, Toward MGR, Thompson DJ, Degrande G, et al.
Mitigation of railway induced ground vibration by heavy masses next to the track.
Soil Dynam Earthq Eng 2015;75:158–70. https://doi.org/10.1016/j.
soildyn.2015.04.003.
[10] Sitharam TG, Sebastian R, Fazil F. Vibration isolation of buildings housed with
sensitive equipment using open trenches – case study and numerical simulations.
Soil Dynam Earthq Eng 2018;115:344–51. https://doi.org/10.1016/j.
soildyn.2018.08.033.
[11] Beskos DE, Dasgupta B, Vardoulakis IG. Engineering M. Vibration isolation using
open or filled trenches. Comput Mech 1986;1:43. https://doi.org/10.1007/
BF00298637. 1975–6.
[12] Adam M, von Estorff O. Reduction of train-induced building vibrations by using
open and filled trenches. Comput Struct 2005;83:11–24. https://doi.org/10.1016/
j.compstruc.2004.08.010.
[13] Bo Q, Ali L, Irini D-M. Numerical study of wave barrier and its optimization design.
Finite Elem Anal Des 2014;84:1–13. https://doi.org/10.1016/j.finel.2014.02.002.
[14] Feng S, Li J, Zhang X, Chen Z, Zheng Q, Zhang D. Numerical analysis of buried
trench in screening surface vibration. Soil Dynam Earthq Eng 2019;126:105822.
https://doi.org/10.1016/j.soildyn.2019.105822.
[15] Hung HH, Yang YB, Chang DW. Wave barriers for reduction of train-induced
vibrations in soils. J Geotech Geoenviron Eng 2004;130:1283–91. https://doi.org/
10.1061/(ASCE)1090-0241(2004)130:12(1283).
[16] Yang YB, Ge P, Li Q, Liang X, Wu Y. 2.5D vibration of railway-side buildings
mitigated by open or infilled trenches considering rail irregularity. Soil Dynam
Earthq Eng 2018;106:204–14. https://doi.org/10.1016/j.soildyn.2017.12.027.
[17] Al-Hussaini TM, Ahmad S. Active isolation of machine foundations by in-filled
trench barriers. J Geotech Eng 1996;122:288–94. https://doi.org/10.1061/(ASCE)
0733-9410(1996)122:4(288).
[18] Massarsch KR. Vibration isolation using gas-filled cushions. 2005. https://doi.org/
10.1061/40780(159)7.
[19] Ekanayake SD, Liyanapathirana DS, Leo CJ. Attenuation of ground vibrations using
in- filled wave barriers. Soil Dynam Earthq Eng 2014;67:290–300. https://doi.org/
10.1016/j.soildyn.2014.10.004.
[20] Chew JH, Leong EC. Field and numerical modelling of sand-rubber mixtures
vibration barrier. Soil Dynam Earthq Eng 2019;125:105740. https://doi.org/
10.1016/j.soildyn.2019.105740.
[21] Yarmohammadi F, Rafiee-Dehkharghani R, Behnia C, Aref AJ. Design of wave
barriers for mitigation of train–induced vibrations using a coupled genetic-
algorithm/finite-element methodology. Soil Dynam Earthq Eng 2019;121:262–75.
https://doi.org/10.1016/j.soildyn.2019.03.007.
[22] Jayawardana P, Thambiratnam DP, Perera N, Chan T. Dual in-filled trenches for
vibration mitigation and their predictions using artificial neural network. Soil
Dynam Earthq Eng 2019;122:107–15. https://doi.org/10.1016/j.
soildyn.2019.04.006.
[23] Murillo C, Thorel L, Caicedo B. Geotextiles and geomembranes ground vibration
isolation with geofoam barriers : centrifuge modeling. Geotext Geomembranes
2009;27:423–34. https://doi.org/10.1016/j.geotexmem.2009.03.006.
[24] Alzawi A, Hesham El Naggar M. Full scale experimental study on vibration
scattering using open and in-filled (GeoFoam) wave barriers. Soil Dynam Earthq
Eng 2011;31:306–17. https://doi.org/10.1016/j.soildyn.2010.08.010.
A. Lyratzakis et al.
[25] Coulier P, François S, Degrande G, Lombaert G. Subgrade stiffening next to the
track as a wave impeding barrier for railway induced vibrations. Soil Dynam
Earthq Eng 2013;48:119–31. https://doi.org/10.1016/j.soildyn.2012.12.009.
[26] Coulier P, Cuéllar V, Degrande G, Lombaert G. Experimental and numerical
evaluation of the effectiveness of a stiff wave barrier in the soil. Soil Dynam Earthq
Eng 2015;77:238–53. https://doi.org/10.1016/j.soildyn.2015.04.007.
[27] With C, Bahrekazemi M, Bodare A. Wave barrier of lime–cement columns against
train-induced ground-borne vibrations. Soil Dynam Earthq Eng 2009;29:1027–33.
https://doi.org/10.1016/j.soildyn.2008.12.005.
[28] Andersen L, Nielsen SRK. Reduction of ground vibration by means of barriers or
soil improvement along a railway track. Soil Dynam Earthq Eng 2005;25:701–16.
https://doi.org/10.1016/j.soildyn.2005.04.007.
[29] Pflanz G, Hashimoto K, Chouw N. Reduction of structural vibrations induced by a
moving load. J Appl Mech 2002;5:555–63. https://doi.org/10.2208/
journalam.5.555.
[30] Dijckmans A, Ekblad A, Smekal A, Degrande G, Lombaert G. Efficacy of a sheet pile
wall as a wave barrier for railway induced ground vibration. Soil Dynam Earthq
Eng 2016;84:55–69. https://doi.org/10.1016/j.soildyn.2016.02.001.
[31] Thompson DJ, Jiang J, Toward MGR, Hussein MFM, Dijckmans A, Coulier P, et al.
Mitigation of railway-induced vibration by using subgrade stiffening. Soil Dynam
Earthq Eng 2015;79:89–103. https://doi.org/10.1016/j.soildyn.2015.09.005.
[32] Gao G, Li N, Gu X. Field experiment and numerical study on active vibration
isolation by horizontal blocks in layered ground under vertical loading. Soil Dynam
Earthq Eng 2015;69:251–61. https://doi.org/10.1016/j.soildyn.2014.11.006.
[33] Antes H, von Estorff O. Dynamic response of 2D and 3D block foundations on a
halfspace with inclusions. Soil Dynam Earthq Eng 1994;13:305–11. https://doi.
org/10.1016/0267-7261(94)90022-1.
[34] Takemiya H. Field vibration mitigation by honeycomb WIB for pile foundations of
a high-speed train viaduct. Soil Dynam Earthq Eng 2004;24:69–87. https://doi.
org/10.1016/j.soildyn.2003.07.005.
[35] Lyratzakis A, Tsompanakis Y, Psarropoulos PN. Efficient mitigation of high-speed
trains induced vibrations of railway embankments using expanded polystyrene
blocks. Transp Geotech 2020;22:100312. https://doi.org/10.1016/j.
trgeo.2019.100312.
[36] Lyratzakis A, Tsompanakis Y, Psarropoulos PN. Mitigation of HSΤ-induced
vibrations by EPS blocks applied in railway embankments. Invited paper for special
issue on: noise and vibration from transportation. J Zhejiang Univ Sci A 2020.
https://doi.org/10.1631/jzus.A1900680.
[37] Jiang J, Toward MGR, Dijckmans A, Thompson DJ, Degrande G, Lombaert G,
Hussein MFM. The influence of soil conditions on railway induced ground-borne
vibration and relevant mitigation measures. In: The 21st International Congress on
Sound and Vibration, Beijing, China, 13-17 July; 2014.
[38] Thompson DJ, Jiang J, Toward MGR, Hussein MFM, Ntotsios E, Dijckmans A, et al.
Reducing railway-induced ground-borne vibration by using open trenches and soft-
filled barriers. Soil Dynam Earthq Eng 2016;88:45–59. https://doi.org/10.1016/j.
soildyn.2016.05.009.
15
Soil Dynamics and Earthquake Engineering xxx (xxxx) xxx
[39] Feng SJ, Zhang XL, Zheng QT, Wang L. Simulation and mitigation analysis of
ground vibrations induced by high-speed train with three dimensional FEM. Soil
Dynam Earthq Eng 2017;94:204–14. https://doi.org/10.1016/j.
soildyn.2017.01.022.
[40] ABAQUS. Analysis user’s manual Ver, 6.14. Providence, RI, USA: Simulia Corp.;
2014.
[41] International union of Railways. UIC code 719R: earthworks and trackbed layers
for railway lines. France: UIC. Paris; 1994.
[42] Nakamura N. Time history response analysis using extended Rayleigh damping
model. Procedia Eng 2017;199:1472–7. https://doi.org/10.1016/j.
proeng.2017.09.408.
[43] Kouroussis G, Verlinden O. Prediction of railway induced ground vibration through
multibody and finite element modelling. Mech Sci 2013;4:167–83. https://doi.org/
10.5194/ms-4-167-2013.
[44] Degrande G, Schillemans L. Free field vibrations during the passage of a thalys
high-speed train at variable speed. J Sound Vib 2001;247:131–44. https://doi.org/
10.1006/jsvi.2001.3718.
[45] Connolly DP, Kouroussis G, Woodward PK, Alves Costa P, Verlinden O, Forde MC.
Field testing and analysis of high speed rail vibrations. Soil Dynam Earthq Eng
2014;67:102–18. https://doi.org/10.1016/j.soildyn.2014.08.013.
[46] Kouroussis G, Connolly DP, Olivier B, Laghrouche O, Costa PA. Railway cuttings
and embankments: experimental and numerical studies of ground vibration. Sci
Total Environ 2016;557–558:110–22. https://doi.org/10.1016/j.
scitotenv.2016.03.016.
[47] Neupane R. Expanded polystyrene geofoam embankment for support of railways
and bridges. PhD Thesis. Utah: The University of Utah; 2015.
[48] Li S. Evaluation and numerical modeling of deflections and vertical displacement
of rail systems supported by EPS geofoam embankments. Utah: MSc Thesis, The
University of Utah; 2014.
[49] ASTM D6817/D6817M-17. Standard specification for rigid cellular polystyrene
geofoam. West Conshohocken, PA: ASTM International; 2017. www.astm.org.
[50] Yang Y-B, Hung H-H. A parametric study of wave barriers for reduction of train-
induced vibrations. Int J Numer Methods Eng 1997;40:3729–47. https://doi.org/
10.1002/(SICI)1097-0207(19971030)40:20<3729::AID-NME236>3.0.CO;2-8.
[51] Deutsches Institut fur Normung. Din 4150-2: structural vibrations – Part 2: Human
exposure to vibration in buildings, Germany. 1999.
[52] Kouroussis G, Conti C, Verlinden O. Building vibrations induced by human
activities: a benchmark of existing standards. Mec Ind 2014;15:345–53. https://
doi.org/10.1051/meca/2014041.
[53] Coulier P, Dijckmans A, Jiang J, Thompson DJ, Degrande G, Lombaert G. Stiff wave
barriers for the mitigation of railway induced vibrations. Noise and Vibration
Mitigation for Rail Transportation Systems. In: Nielsen J, et al., editors. Stiff wave
barriers for the mitigation of railway induced vibrations, 126. Berlin, Heidelberg:
Springer; 2013. p. 539546.