Ground-Borne Vibration Due To Railway Traffic A Review of Excitation Mechanisms, Prediction Methods and Mitigation Measures

Ground-Borne Vibration due to Railway Traffic:
A Review of Excitation Mechanisms, Prediction

Methods and Mitigation Measures
G. Lombaert1, G. Degrande1, S. François1 , and D.J. Thompson2

1
KU Leuven, Department of Civil Engineering,
Kasteelpark Arenberg 40, 3001 Leuven, Belgium
geert.lombaert@bwk.kuleuven.be
2
Institute of Sound and Vibration Research, University of Southampton, UK
Summary. The aim of this paper is to provide a comprehensive overview of

the state of the art on railway-induced ground vibration. The governing physical
mechanisms, prediction methods, and mitigation measures of ground-borne vi-
bration are discussed, with focus on low frequency feelable vibration and the case
of railway traffic at grade. In order to clarify the importance of quasi-static and
dynamic excitation, the basic problems of a moving load with constant magnitude
and harmonic magnitude are discussed first. Dynamic excitation due to wheel and
track unevenness and parametric excitation is shown to be the dominant source of
environmental vibration in most cases. Next, an overview of prediction methods
for ground-borne vibration is given. The advantages and limitations of numer-
ical methods, based on physical or mechanical models, and empirical models,
derived from measured data, are discussed. Finally, the mitigation of railway-
induced ground vibration is considered, where the focus goes to mitigation mea-
sures at source (wheel and rail unevenness, rolling stock, track) and measures
on the transmission path (trenches and barriers, wave impeding blocks, subgrade
stiffening, and heavy masses next to the track). In conclusion, a number of open
points requiring further research is given.
1 Introduction
Ground-borne noise and vibration due to railway traffic is a problem of large societal
and economic importance. This problem has gained considerable attention recently for a
number of reasons. Problems with congestion of road traffic in densely populated areas
have led to an increased demand for public transport, stimulating the construction of
light rail systems and underground railway lines. The development of high speed train
(HST) networks in Europe and Asia has led to a significant increase in the speed of
passenger trains and raised concerns on the corresponding environmental impact. The
desired shift of freight transport by road to rail will lead to higher axle loads and an
increased number of freight trains. Furthermore, railway noise and vibration also need
to be addressed for the development of land in urban areas adjacent to railway lines.
Railway-induced vibrations are generated by quasi-static and dynamic axle loads;
the latter are due to several mechanisms such as wheel and rail unevenness, impact
c Springer-Verlag Berlin Heidelberg 2015 253

J.C.O. Nielsen et al. (eds.), Noise and Vibration Mitigation for Rail Transportation Systems, Notes on
Numerical Fluid Mechanics and Multidisciplinary Design 126, DOI: 10.1007/978-3-662-44832-8_33
254 G. Lombaert et al.
Fig. 1. Railway-induced vibrations in the built environment
excitation due to rail joints and wheel flats, and parametric excitation due to sleeper pe-
riodicity [140]. These loads are transferred to the track, its supporting structure (ballast,
subgrade, slab or tunnel) and the soil, where vibrations propagate as elastic waves and
excite the foundations of nearby buildings (figure 1). In the frequency range between
1 and 80 Hz, building vibration is perceived as mechanical vibration of the human
body, whereas between 16 and 250 Hz, ground-borne vibrations can cause re-radiated
or structure-borne noise by vibrating walls and floors [40]. Low frequency feelable
ground-borne vibration is most often due to freight traffic or rolling stock with heavy un-
sprung masses travelling on tracks at grade in areas with soft soil. Problems of ground-
borne noise are typically encountered when the airborne noise is effectively shielded off,
i.e. in case of underground railway traffic or tracks at grade equipped with noise barri-
ers [140]. Norms and guidelines recognise discomfort to people [19, 30, 66, 67, 136],
malfunctioning of sensitive equipment [137] and damage to buildings [31, 135] as pos-
sible consequences of vibrations. Guidance on the prediction of ground-borne vibration
arising from rail systems is provided in a recent ISO standard [68].
Traffic-induced vibrations have been investigated since the beginning of the previous
century. In 1929, Hyde and Lintern refer to studies that were conducted in 1901 on the
vibration that was induced by the Central London Railway [65]. They state "there was
probably no subject before engineers in the country today, particularly those dealing
with road transport, that caused more trouble, or gave rise to more concern, than vi-
bration". Since then, a lot of research has been performed on transportation-induced
noise and vibration, e.g. for the development of prediction methods and mitigation
measures. The problem of railway-induced ground vibration is still of large concern
to various stakeholders and highly relevant in view of the rapid extension of high speed
[52, 97, 125, 139] and urban railway networks [25], the increased use of land for con-
struction of buildings near railways, and an increased public awareness of noise and
vibration.
The aim of this paper is to provide a comprehensive overview of the state of the art on
railway-induced ground vibration. The focus is on the governing physical mechanisms,
prediction, and mitigation of ground-borne vibration. The response of buildings [40, 44,
149], human perception of vibration and exposure-effect relation [49, 112, 115, 144],
and limit values in norms and guidelines are not discussed. Moreover, the discussion is
Ground-Borne Vibration due to Railway Traffic 255
mostly limited to ground-borne vibration for the case of railway traffic at grade. Many
of the elements discussed will also apply, however, to ground-borne vibration and noise
generated by underground railway traffic and trains running on viaducts [55, 139, 155].
For a more general overview on rolling noise and vibration due to railway traffic, the
reader is referred to the recent book of Thompson [140] whereas a discussion of ground
vibration for a wider range of excitation sources, including traffic, construction activi-
ties, and seismic events, can be found in the book of Semblat and Pecker [127].
The outline of this paper is as follows. In section 2, a brief recapitulation is made of
the relevant characteristics governing wave propagation in the soil and the wave fields
generated by loads at fixed and moving positions are discussed. These elements are
used to clarify quasi-static and dynamic excitation of railway-induced ground vibra-
tion in section 3. Next, it is discussed how track-soil interaction determines how the
vehicle loads are transferred to the soil. In section 4, an overview of methods for the
prediction of ground-borne vibration is given, making a distinction between numeri-
cal methods, based on physical or mechanical models, and empirical models derived
from measured data. The mitigation of railway-induced ground vibration is discussed
in section 5, where the focus is given to the mitigation measures at the source and
on the transmission path. Open points requiring further investigation are listed in the
conclusions.
2 Wave Propagation in the Soil

2.1 Body Waves and Surface Waves
Apart from a small zone immediately underneath the track, the strain levels in the soil
remain relatively low during the passage of a train, so that a linear elastic constitutive
behaviour can reasonably be assumed in the study of ground-borne vibration. Inelastic
behaviour of the track subgrade has been found to be important for high speed trains
running on tracks supported by very soft soil. In order to account for stiffness degra-
dation and increased energy dissipation under large amplitude cyclic loading of soil,
an equivalent linear analysis of the track response can be performed [22, 97]. Inelastic
behaviour is expected to have less impact on the vibration levels away from the track,
however.
In a homogeneous and isotropic elastic medium of infinite extent, a distinction can be
made between two types of waves: dilatational (or: longitudinal, irrotational, primary,
P-) waves where the soil particles move parallel to the wave propagation direction and
shear (or: transverse, equivoluminal, rotational, secondary, S) waves where the soil par-
ticles move perpendicular to the wave propagation direction [1, 11]. The corresponding
wave velocities are computed as:

μ λ + 2μ
Cs = ; Cp = (2.1)
ρ ρ
where ρ is the soil density and λ and μ are the Lamé constants which are related to the
Young’s modulus E and the Poisson’s ratio ν as follows:
Eν E
λ= ;μ = (2.2)
(1 + ν)(1 − 2ν) 2(1 + ν)
From this it can be seen that elastic properties of the soil can be defined in terms of the
Lamé constants λ and μ, the Young’s modulus E and Poisson’s ratio ν or in terms of
the fundamental wave speeds Cp and Cs in each case in combination with the density ρ.
The ratio s of the dilatational and transversal wave velocities depends only on Poisson’s
ratio ν:

Cs 1 − 2ν
s= = (2.3)
Cp 2 − 2ν
Typical values for wave velocities in soft soil such as peat are Cs = 50 m/s, Cp =
357 m/s, whereas for a medium stiff soil typically Cs = 150 m/s, Cp = 300 m/s, and
for a stiff soil Cs = 400 m/s, Cp = 800 m/s. Even higher values are encountered in the
case of sandstone and rock materials.
At low frequencies, energy dissipation in the soil under cyclic excitation has been
found to be rate-independent [4]. Frequency-independent hysteretic material damping
in the soil can be modelled in the frequency domain by means of the correspondence
principle [117, 121] through the use of complex Lamé coefficients:
(λ + 2μ) = (λ + 2μ)(1 + 2βp i) (2.4)

μ = μ(1 + 2βs i) (2.5)
where βp and βs represent the frequency- independent hysteretic material damping ratios
for dilatational and shear waves, respectively. Alternatively, loss factors ηp = 2βp , ηs =
2βs are used to characterize energy dissipation.
At the free surface of a halfspace, interaction between dilatational and shear waves
results in a surface or Rayleigh wave [116]. In the case of a homogeneous halfspace, a
single Rayleigh wave mode exists with a phase velocity CR approximately equal to [1]:
0.862 + 1.14ν
CR ≈ Cs (2.6)
1+ν
For a realistic range of Poisson’s ratios, the Rayleigh wave velocity is in the range of
90% to 95% of the shear wave velocity. Equation (2.6) shows that in a homogeneous
halfspace, the Rayleigh wave velocity does not depend on the frequency, meaning that
the wave is non-dispersive. The Rayleigh wave propagates along the free surface of the
halfspace and has a limited penetration depth of approximately one wavelength.
Natural soils often have a horizontal stratification as their formation is generally gov-
erned by phenomena affecting large areas of land, such as erosion, sediment transport,
and weathering processes [39]. In this case, the soil can be modelled as a horizontally
layered elastic halfspace, where the material properties Cs , Cp , ρ, βs , and βp vary in the
vertical direction only. In a layered halfspace, multiple Rayleigh modes occur, which
are dispersive due to the variation with depth of the dynamic soil characteristics. This is
illustrated by results for a site at Lincent, Belgium, located next to the high speed line
between Brussels and Köln, where ground vibration measurements have been made for
validation of numerical models [91, 94]. Borings performed prior to the construction
of the high speed line show that the soil consists of a shallow quaterny top layer of silt
with a thickness of 1.2 m, followed by a layer of fine sand up to a depth of 3.2 m. Be-
tween 3.2 m and 7.5 m is a sequence of stiff layers of arenite (a sediment of a sandstone
0 0
2 2
Depth [m]
Depth [m]
4 4
6 6
8 8
10 10
−1.5 −1 −0.5 0 0.5 1 1.5 −1.5 −1 −0.5 0 0.5 1 1.5
Displacement [−] Displacement [−]
(a) (b) (c)
Fig. 2. Rayleigh wave modes calculated for a layered halfspace (site at Lincent, Belgium): (a)
dispersion curves of the first three Rayleigh wave modes, (b) horizontal and (c) vertical compo-
nent of the fundamental Rayleigh wave mode as a function of depth at 10 Hz (solid line), 20 Hz
(dashed line), and 40 Hz (dotted line)
residue) embedded in clay. The small strain dynamic soil characteristics at the site have
been determined by means of two Spectral Analysis of Surface Wave tests and five Seis-
mic Cone Penetration Tests. These results show that, in the frequency range of interest
for railway-induced vibration, the soil can be represented by a single layer with a thick-
ness of 3.0 m and a shear wave velocity of about 150 m/s on top of a stiffer halfspace
with a shear wave velocity of 280 m/s. A summary of the dynamic soil characteristics
at this site as assumed in previous studies [91, 94] is shown in Table 1.
Table 1. Dynamic soil characteristics at the site at Lincent, Belgium
Layer d Cs Cp E ν ρ β
[m] [m/s] [m/s] [×106 N/m2 ] [-] [kg/m3 ] [-]
1 3 150 300 120 0.333 2000 0.03
2 ∞ 280 560 418 0.333 2000 0.03
Figure 2a shows the dispersion curve of the first three Rayleigh wave modes, com-
puted with the ElastoDynamics Toolbox (EDT) for MATLAB [123]. At low frequen-
cies, the fundamental Rayleigh wave has a phase velocity close to 261 m/s at low
frequencies, corresponding to the phase velocity of Rayleigh waves in the halfspace.
At high frequencies, the phase velocity converges asymptotically to a value of about
140 m/s, corresponding to the Rayleigh wave velocity in the top layer. The variation
with depth of the fundamental mode shape is shown in figures 2b and 2c. At low fre-
quencies, the Rayleigh wave reaches very deep, explaining why the phase velocity is
affected by the stiff deeper layers whereas at high frequencies, the motion is concen-
trated near the surface and dominated by the properties of the soft surface layer.
Higher order Rayleigh wave modes with a more complicated mode shape develop
at higher frequencies. The cut-on frequencies of the second and third Rayleigh wave
are observed at 20 Hz and 44 Hz, where the higher order waves have an initial phase
velocity of 280 m/s, corresponding to the shear wave velocity of the halfspace [145]. In
the Spectral Analysis of Surface Waves test, information on the dispersive behaviour of
the Rayleigh wave modes is extracted from an in situ experiment and used for identify-
ing the variation with depth of the shear wave velocity and the corresponding material
damping ratio [101, 120]. The grey area in the background of figure 2a is a represen-
tation in the frequency - phase velocity domain of the Green’s function or fundamental
solution for the vertical displacement due to a vertical point load, both at the surface
of the layered halfspace. These results illustrate the relative importance of the different
surface wave modes in this particular problem.
In the presence of ground water, the pores between the solid skeleton may be com-
pletely saturated with water. Biot’s theory [15, 16] can be used to describe wave prop-
agation in saturated poroelastic media for the prediction of railway-induced ground
vibration [103]. According to Biot’s theory, two P-waves and a single S-wave are found
in a saturated poroelastic medium [122]. A characteristic frequency that is inversely
proportional to the permeability of the poroelastic medium, determines the importance
of the corresponding relative motion of the solid and fluid phase. For typical soils, the
characteristic frequency is in the order of several kHz, which is much higher than the
frequency range of interest here. At frequencies that are very low compared to the char-
acteristic frequency, no relative motion between the fluid and solid phase occurs for
the first type of P-wave and the S-wave. For the second type of P-wave, the fluid and
solid phase move perfectly out-of-phase. This type of wave, however, is characterized
by very high attenuation coefficients, meaning that wave propagation essentially occurs
through a single P-wave and S-wave. In the frequency range of interest for railway in-
duced ground vibration, the soil can therefore be modelled as an equivalent dry elastic
medium, provided that the density and incompressibility of the saturated soil layers are
accounted for by equivalent Lamé coefficients μeq and λeq [122]. The resulting P-wave
velocity is high as the presence of the pore fluid results in low compressibility while
the S-wave velocity of the orginal dry solid is only weakly affected due to the change
in density. Moderate seasonal variations of the ground water level were found not to
significantly affect peak values of ground vibrations produced by road traffic [122].
2.2 Wave Field Generated by a Point Load at a Fixed Position
For a point load acting at a fixed position, the relation between the soil’s displacement
response û(r, ω) and the load amplitude fˆ(ω) can be written in the frequency domain as
follows because of the assumed linear elastic constitutive behaviour:
û(r, ω) = Ĥ(r, ω) fˆ(ω) (2.7)
where Ĥ(r, ω) is the transfer function and r is the distance between the receiver and
the source. A hat above a variable denotes its representation in the frequency domain.
Figure 3a shows the modulus of the transfer function that represents the vertical dis-
placement response at 8, 16, 32 and 64 m from a vertical point source for the site at
Lincent. At very low frequencies, the attenuation with distance is proportional to 1/r.
A steep rise as a function of the frequency is found between 10 and 20 Hz. This occurs
at the onset of wave propagation in the surface layer, when one quarter of the Rayleigh
wavelength computed with the properties of the top soil material fits within the thick-
ness of this layer. A moderate peak is found in the transfer functions near the frequency
at which half the Rayleigh wavelength fits within the surface layer [8]. From this fre-
quency on, the surface wave mainly propagates in the soft top layer (figures 2b and 2c).
The peaks and troughs observed for each of the transfer functions in figure 3a are due
to interference between different types of waves [122].
−9 −9
10 10
Displacement [m/N]
Displacement [m/N]
−10 −10
10 10
−11 −11
10 10
−12 −12
10 10
0 20 40 60 80 100 0 20 40 60 80 100
Frequency [Hz] Distance [m]
Fig. 3. Modulus of the transfer function Ĥ(r, ω) (a) as a function of the frequency f = ω/2π at
r = 8 m (solid line), r = 16 m (dashed line), r = 32 m (dotted line), and r = 64 m (dashed-dotted
line) from the source and (b) as a function of the distance r from the source at f = 10 Hz (solid
line), f = 20 Hz (dashed line), and f = 40 Hz (dotted line)
Figure 3b shows the modulus of the transfer function as a function of the distance
from the source at 10, 20, and 40 Hz. When the distance from the source is small com-
pared to the dominant wavelength in the soil, the receiver is located in the so-called near
field region where different wave components can generally not be distinguished [4]. At
larger distances, P-, S- and Rayleigh wave fronts have fully developed and their arrival
can be detected. The attenuation with distance is governed by geometric attenuation and
material damping. Geometric attenuation is caused by the expansion of the wavefronts,
resulting in the spreading of energy over an increasing area. In a homogeneous halfs-
pace, √geometric attenuation of Rayleigh waves generated by a point load is proportional
to 1/ r and frequency independent. This is no longer the case for a layered halfspace,
however, as multiple Rayleigh wave modes (figure 2a) will contribute to the response
at the soil’s surface. Material damping introduces an additional frequency dependent
attenuation with distance which is due to energy dissipation in the soil. The stronger
reduction with increasing frequency of the transfer functions at larger distances from
the source (figure 3) is mainly due to the effect of material damping. As a result, the
frequency at which the peak of the transfer function is found decreases with increasing
distance from the source. The observed attenuation with distance of the transfer func-
tions of a layered halfspace as shown in figure 3 is hard to capture in analytical form
and needs to be retrieved by numerical simulations, e.g. by means of the direct stiffness
method [78, 79] or a transfer matrix approach [54, 141].
2.3 Wave Field Generated by a Moving Point Load

Whereas a point load at a fixed position can only generate a time-varying response when
its amplitude is also time-varying, this is no longer true for the case of a moving point
load. Even when the load amplitude is constant in time, the motion of the load will lead
to a time-varying response at a fixed point. The response depends on the magnitude of
the load speed relative to the wave velocities in the soil. Figure 4 shows the calculated
wave field generated by a point load of constant unit amplitude for three load speeds:
v = 0, v = 100 m/s, and v = 200 m/s moving on the layered halfspace representing the
soil at the site of Lincent (table 1). These results have been obtained from the fundamen-
tal solution or Green’s function of the layered halfspace by applying the Betti-Rayleigh
reciprocity theorem and exploiting the invariance of the problem geometry in the di-
rection in which the load moves [92]. Comparison of figures 4a and 4b shows that the
wave field generated by a load moving at a speed of 100 m/s, which is below the lower
limit of the phase velocities of surface waves (figure 2a), is similar to the one of a load
at a fixed position. No propagating waves are emitted by the moving load in this case.
An observer at a fixed position in the free field will observe a time variant response,
however, as the deflection shape travels with the load at a speed of 100 m/s. When the
load speed exceeds the lower limit of the phase velocities of surface waves, the wave
field generated by the moving load changes drastically. In this case, propagating waves
are generated by the moving load, giving rise to a sharp discontinuity in the wave field,
and the formation of a Mach cone. The displacement amplitudes are significantly larger
in this case.
When considering railway traffic, the critical load speed at which large amplifications
of track displacements are expected, also depends on the properties of the track, but is
generally controlled by the Rayleigh wave velocities in the soil [97]. For high speed
−8 −8
x 10 x 10
1 1
Displacement [m]
Displacement [m]
0 0
−1 −1
−25 −25
0 25 0 25
0 0
25 −25 25 −25
x [m] y [m] x [m] y [m]
(a) (b)
−8
x 10
1
Displacement [m]
−1
−25
0 25
0
25 −25
(c) x [m] y [m]
Fig. 4. Wave field generated by a unit point load with constant amplitude as seen in a moving
frame: (a) at a fixed position, (b) moving at a speed of 100 m/s, and (c) moving at a speed of
200 m/s (peak amplitudes clipped). Calculated results for soil properties of Lincent, Belgium.
trains (v ≥ 180 km/h) travelling on tracks supported by very soft soil (Cs ≤ 50 m/s),
the train speed can actually be close to or even larger than this critical load speed. In
the 1990’s, the problem of trains running at trans-critical speeds was addressed in a se-
ries of papers by Krylov [85, 86, 88] focussing on ground vibration and Dieterman and
Metrikine [32, 33] investigating track-soil interaction. These studies have indicated that
a significant amplification of vibration levels and track displacements is obtained com-
pared with the subcritical speed range, leading to problems of track stability and safety.
This was confirmed by field measurements at the site of Ledsgård along the West Coast
Line in Sweden where track displacements up to 10 mm have been measured during the
passage of the X2000 train [3]. This case has been considered in numerical simulations
by many authors, among which the first were Madshus and Kaynia [97]. Problems en-
countered with high speed trains running at trans-Rayleigh speeds are primarily of large
concern for reasons of track stability, however, necessitating subgrade stiffening or piled
track foundations, and will rarely be an issue for problems of environmental ground-
borne vibration. In the following, focus will therefore be on the subcritical range of
train speeds.
−8 −8
x 10 x 10
1 1
Displacement [m]
Displacement [m]
0 0
−1 −1
−25 −25
0 25 0 25
0 0
25 −25 25 −25
x [m] y [m] x [m] y [m]
(a) (b)
−8
x 10
1
Displacement [m]
−1
−25
0 25
0
25 −25
(c) x [m] y [m]
Fig. 5. Wave field generated by a unit point load with harmonic time variation ( f = 25 Hz) as
seen in a moving frame: (a) at a fixed position, (b) moving at a speed of 100 m/s, and (c) moving
at a speed of 200 m/s (peak amplitudes clipped). Calculated results for soil properties of Lincent,
Belgium.
Whereas a moving load with constant amplitude only generates propagating waves
at trans-Rayleigh load speeds, a moving load with time-varying amplitude will gener-
ate propagating waves, irrespective of the load speed, similarly to the case of a fixed
position. This is illustrated in figure 5 which shows the wave field generated by a unit
point load with harmonic variation in time ( f = 25 Hz) for the same three load speeds.
Comparison of figures 5a and 5b clearly shows the Doppler effect in the case of a mov-
ing load. In front of the moving load (figure 5b), shorter wavelengths are observed, in
agreement with the higher frequency observed at a fixed receiver ahead of the source,
whereas the longer wavelengths behind the load correspond to the lower frequencies
observed at a point behind the source. The Doppler effect has been identified in field
measurements of railway-induced vibrations by Ditzel et al. [35]. When the load speed
exceeds the lowest Rayleigh wave velocity in the soil, the wave field displays a Mach
cone (figure 5c) as in the case of a load with constant amplitude (figure 4c), but without
the large amplification observed in the latter case.
Due to the load motion, equation (2.7) that relates the response to the load amplitude
for a load at a fixed position, is no longer valid. For train speeds that are low compared
with the wave velocities in the soil, however, the effect of the moving load and the
corresponding Doppler shift are small and equation (2.7) holds approximately at the
time when the load and receiver are at an intermediate distance r.
3 Excitation Mechanisms
3.1 Quasi-Static Excitation
The loads applied to the track by a running train can be decomposed into a static and
dynamic load component. Assuming a linear behaviour of the track and the support-
ing soil, the resulting ground vibration u(t) can be decomposed into the quasi-static
contribution us (t) and the dynamic contribution ud (t):
u(t) = us (t) + ud (t) (3.1)
When the train speed is situated in the subcritical range, the quasi-static response of
the soil resembles a sequence of bowl shaped deflections, each similar in form to the
one for an upward load shown in figure 4b, that travel with the train. The time vari-
ation of the response at a fixed point is therefore due to successive rising and falling
of the response at the passage of each axle. The repeated passage of axles leads to the
characteristic peaks and troughs in the narrow-band frequency spectrum of the response
that are determined by the axle and bogie passage frequencies [9, 10, 28]. This can be
understood by writing the quasi-static response as a superposition of the contributions
of different train axles:
na
us (t) = wk us0 (t − yk0 /v) (3.2)
k=1
where na is the number of axles of the train, wk is the weight carried by axle k, us0 (t)
is the response due to a moving load with unit magnitude and yk0 is the position of the
axle on the track at a reference time t = 0. In the frequency domain, this expression
becomes:
⎡ na y ⎤⎥
⎢⎢⎢ k0 ⎥
⎥⎥⎥
ûs (ω) = ûs0 (ω) ⎢⎢⎣ wk exp iω (3.3)
k=1
v ⎦
where the bracketed term that depends on the distribution of the weight over the axles
and the train speed v gives rise to the characteristic shape of the narrow band response
spectrum [9, 10, 28].
This is now illustrated for the computed response of the sleeper due to the passage
of an InterCity train at the site of Lincent. For more details regarding the model and
input parameters, the reader is referred to Lombaert and Degrande [91]. Figures 6a and
6b show the time history and narrow-band spectrum of the sleeper velocity due to the
passage of the first axle of an InterCity train at a speed of 156 km/h (43.3 m/s), which is
well below the lower limit of the Rayleigh wave velocities in the soil. Due to the short
duration of the quasi-static sleeper response for a single axle (figure 6a), the passages
of individual axles do not overlap in time and are still observed in the time history
response for the entire train (figure 6c). The results show that the InterCity train is in
"pull mode" where the axles of the locomotive that carry the largest weight come first.
Comparison of the narrow-band spectra for a single axle (figure 6b) and the full train
(figure 6d) shows how the characteristic peaks and troughs appear by multiplication
with the bracketed term in equation (3.3).
−3
x 10
0.05
1
Velocity [m/s/Hz]
Velocity [m/s]
0.8
0 0.6
0.4
0.2
−0.05 0
−0.1 −0.05 0 0.05 0.1 0 20 40 60 80 100
Time [s] Frequency [Hz]
(a) (b)
0.05 0.012
0.01
Velocity [m/s/Hz]
Velocity [m/s]
0.008
0 0.006
0.004
0.002
−0.05 0
−5 0 5 0 20 40 60 80 100
(c) (d)
Fig. 6. Quasi-static contribution to the sleeper response due to the passage of an InterCity train
at a speed of 156 km/h: (a) time history and (b) narrow-band spectrum of the velocity during the
passage of the first axle of the locomotive and (c) time history and (d) narrow-band spectrum of
the velocity during the passage of the entire train
In the subcritical speed range considered here, the narrow-band spectrum of the
quasi-static contribution to the free field response becomes increasingly concentrated
at low frequencies with increasing distance from the track. This is due to the charac-
teristic bowl shape of the deflection field for each axle (figure 4b). At a larger distance
from the track, the arrival of a single axle is detected earlier and its passage lasts for
a longer time. The time scale at which the response rises and falls therefore increases
with the distance from the track, implying that its representation in the frequency do-
main gets more concentrated at lower frequencies. This is fundamentally different from
the aforementioned reduction in the frequency at which the peak value of the transfer
functions (figure 3) is found, as the latter is due to energy dissipation characterized by
the material damping ratio. Furthermore, as the distance from the track increases, con-
tributions from different axles and bogies coalesce and can no longer be identified in
the time history of the vibration response.
3.2 Dynamic Excitation

The dynamic load component is determined by train-track interaction resulting from
several excitation mechanisms, such as wheel and track unevenness, impact excitation
due to rail joints and wheel flats, and parametric excitation due to the spatial variation
of support stiffness, e.g. due to the discrete nature of sleeper support [56, 82, 105, 140].
In the following, wheel and track unevenness, including impact excitation due to local-
ized defects, are discussed first, as these excitation mechanisms are at present the best
understood and quantified. Parametric excitation, which has received less attention in
the context of ground-borne vibration, is considered next.
Wheel and Track Unevenness. In order to understand how parameters of the vehicle,
track, and supporting soil affect the dynamic load component, it is instructive to con-
sider how the dynamic load component can be computed from the combined wheel and
track unevenness [9, 94, 131]. First, it is assumed that a perfect contact exists between
the train and the track, implying that the following expression should hold:
ûa (ω) = ût (ω) + ûw/r (ω) (3.4)
where ua (ω) and ut (ω) collect the displacements of the axles and the track at the na mov-
ing contact points. The vector uw/r (ω) collects the combined wheel and rail unevenness
perceived by the axles.
Next, the displacements ua (ω) and ut (ω) are expressed in terms of the dynamic train
loads by means of the track and vehicle compliance matrices Ĉt (ω) and Ĉv (ω) that
relate the displacements at the multiple moving contact points to the vector ĝd (ω) of
dynamic train loads [94]:

Ĉt (ω) + Ĉv (ω) ĝd (ω) = −ûw/r (ω) (3.5)
An additional compliance matrix representing the contact spring is in some cases con-
sidered in equation (3.5) to account for Hertzian contact between the wheel and the rail
[130]. In the frequency range of interest for ground-borne vibration, the contact stiffness
is relatively high compared to the track and vehicle stiffness, so that the corresponding
compliance can generally be disregarded. Note that equation (3.5) does not allow ac-
counting for large deflections or loss of contact, e.g. in the presence of wheel flats [154];
this would require a non-linear model. Furthermore, it has been implicity assumed that
the track dynamic properties are translationally invariant in the longitudinal direction
of the track. This assumption is also acceptable for discretely supported ballasted tracks
as the rail receptance at a sleeper and in between two sleepers are similar in the low fre-
quency range of interest for ground-borne vibration [82, 154]. Equation (3.5) therefore
provides a reasonable basis for investigating how vehicle-track interaction affects the
dynamic train loads.
Simplifying equation (3.5) for a single axle allows effects from vehicle-track inter-
action to be highlighted. If the train speed is relatively low, the corresponding track
compliance in a moving frame of reference is approximately equal to the track recep-
tance, and can roughly be represented as the inverse of the track stiffness 1/kt when
its imaginary part and frequency dependence are disregarded. At frequencies of more
than a few Hertz [82], the vehicle’s primary and secondary suspension isolate the bogie
and the body from the wheelset. At sufficiently high frequencies (above approximately
10 Hz), the vehicle’s unsprung mass Mu is therefore the only component that affects
the vertical dynamic loads and can be represented as a rigid body [82]. In this case,
the compliance of a single axle becomes −1/(Mu ω2 ). Introducing these expressions in
equation (3.5) leads to the following expression for the dynamic load at axle k:

kt Mu ω2
ĝdk (ω) = − ûk (ω) (3.6)
kt − Mu ω2 w/r
The denominator
√ on the right hand side of equation (3.6) becomes zero at the frequency
ω = kt /Mu where "resonance" of the unsprung mass on the track stiffness (sometimes
known as the P2 resonance) occurs. At this frequency, the narrow-band spectrum of the
dynamic vehicle load ĝdk (ω) displays a resonance peak. The peak is strongly damped
in reality as the track stiffness has a significant imaginary part which is mainly due to
radiation damping.
Disregarding wheel unevenness for the moment, the frequency domain representa-
tion of the unevenness ûkw/r (ω) experienced by axle k in equation (3.6) can be computed
from the wavenumber domain representation ũw/r (ky ) of the unevenness uw/r (y) along
the track:
1 ω y
k0
ûkw/r (ω) = ũkw/r − exp iω (3.7)
v v v
where yk0 is the position of the axle on the track at a reference time t = 0. The ge-
ometrical irregularities considered here include broad-band unevenness of a random
nature, typically modelled as a stationary random process characterized by its power
spectral density (PSD) [108], but can also represent localized irregularities such as rail
joints giving rise to impact excitation. Equation (3.7) indicates that, in the absence of
wheel unevenness, all axles experience the same excitation apart from a shift in time
accounted for by the phase shift exp (iωyk0 /v) in equation (3.7). When all train carriages
have similar characteristics and track unevenness is the dominant source of excitation,
the resulting dynamic loads will also be similar and the narrow-band spectrum of the
dynamic response contribution will have a similar shape as the one of the quasi-static
response contribution.
Equation (3.7) relates the spectral content of the unevenness in the wavenumber do-
main or, equivalently, in terms of its reciprocal, the wavelength λy = 2π/ky, to the
Table 2. Relevant range of wavelengths λy for ground-borne vibration and noise as a function of
the train speed v
Ground-borne vibration Ground-borne noise

Train speed/frequency 1 Hz 80 Hz 16 Hz 250 Hz
v = 72 km/h 20 m 0.25 m 1.25 m 0.08 m
v = 360 km/h 100 m 1.25 m 6.25 m 0.40 m
excitation experienced by the axle in the frequency domain. Excitation at a frequency

f = ω/(2π) is due to unevenness characterized by a wavenumber ky = 2π f /v or wave-
length λy = v/ f .
Table 2 shows that the relevant range of wavelengths for ground borne vibration and
noise extends from several centimetres up to a hundred metres, depending on the train
speed and the specific problem at hand. The measurement bandwidth of track recording
cars is usually restricted, however, to wavelengths between a few metres and 20 to
30 m [38] and these data therefore need to be supplemented by data obtained from other
measurement devices such as trolleys. This is not trivial, as the unevenness measured by
a track recording car is the one experienced by the car that includes a load dependent
contribution from parametric excitation, and is therefore not a pure representation of
imperfect track geometry.
For wheel unevenness, a similar range of wavelengths as presented in Table 2 for
track unevenness is important. Recent measurements along the conventional line Paris-
Bordeaux have shown that the time history of the ground vibration velocity during the
passage of a freight train has a significantly more irregular character than for high speed
trains and passenger trains [93]. This suggests that wheel unevenness is more important
in the case of freight traffic, which in Europe is characterised by cast-iron brake blocks.
In order to illustrate the importance of the dynamic excitation due to track uneven-
ness, the total response in the free field is now considered for the same passage of an
InterCity train at the site in Lincent. The dynamic response has been calculated consid-
ering the track unevenness measured by a track recording car [91]. Figure 7a shows the
−3
x 10
2
200
Velocity [dBref 10−8 m/s]
Velocity [m/s]
1 150
0 100
50
−1
0
−2 0 1 2
−5 0 5 10 10 10
(a) (b)
Fig. 7. Free field response at 16 m from the track due to the passage of an InterCity train at a
speed of 156 km/h: (a) time history and (b) one-third octave band spectra of the total response
(solid line), quasi-static (dashed line) and dynamic (dotted line) response contribution
time history of the free field velocity at 16 m from the track. In contrast to what was
observed for the quasi-static contribution to the sleeper response in figure 6c, the pas-
sage of individual axles can no longer be identified. Even when the dynamic response
contribution was not accounted for, this would be the case as at a larger distance from
the track, the quasi-static response contributions from different axles and bogies coa-
lesce. Figure 7b shows the one-third octave band spectra of the free field velocity at
16 m from the track. The one-third octave band spectra are computed according to the
German standard DIN 45672-2 [29] for a reference period T 2 which is here considered
as the duration of the stationary part of the response. Comparing the one-third octave
band spectra of the total response, and the quasi-static and dynamic response contri-
butions in figure 7b shows that the latter dominates the total response. A significant
contribution of the quasi-static excitation to the total response is in this case only found
in the frequency range below 3 Hz.
The quasi-static contribution to the response generally remains important, however,
in the immediate vicinity of the track. No attempt can therefore be made to use equation
(2.7) for estimating the dynamic loads applied to the track from the response measured
close to the track and transfer functions Ĥ(r, ω) between the track and the free field
[143]. Furthermore, this also implies that the estimation of insertion loss values from
field tests requires measurements at a sufficiently large distance from the track. The
relative importance of quasi-static and dynamic excitation depends on the train speed,
the ratio of the static and dynamic axle loads, and the dynamic characteristics of the
track and the soil as indicated by Sheng et al. [130, 131], Auersch [10], Lombaert and
Degrande [91], and Triepaischajonsak et al. [142, 143].
Parametric Excitation. Parametric excitation is another source of excitation that re-

lates to the spatial variation of the support stiffness. A first obvious source is the spatial
variation of the dynamic track stiffness of a conventional ballasted track within one
sleeper bay. A wheel running over the rail experiences the relatively small spatial vari-
ation in stiffness and is excited at the sleeper-passing frequency v/d, with d the sleeper
spacing [9, 142, 154]. For conventional railway traffic, the sleeper-passing frequency
is generally situated within the relevant range for ground-borne vibration, whereas for
high speed trains it is in the range for ground-borne noise. For example at 80 km/h and
a sleeper spacing of 0.6 m the sleeper-passing frequency occurs at 37 Hz whereas at
300 km/h it occurs at 139 Hz.
A second source of parametric excitation occurs at transition zones where the change
in stiffness, e.g. resulting from a transition from one type of track to another, will also
excite the vehicle [47, 146]. The length and suddenness of the transition zone will de-
termine the frequency range of excitation.
Parametric excitation may also be caused by other less obvious variations in the
stiffness of the track and the subsoil. Experimental investigations have shown significant
scatter in the sleeper support stiffness [109] and the Young’s modulus and thickness of
the ballast and subgrade layers [118]. Variations in support stiffness from one sleeper
to another and support stiffness variations on a larger spatial scale can also lead to
excitation of the vehicle and ground borne vibration.
At present, the importance of this source of excitation is not fully understood, as

track stiffness measurements are not routinely performed by infrastructure managers
and require specially designed measurement cars such as the Rolling Stiffness Mea-
surement Vehicle (RSMV) [12, 13, 152]. One could argue that the distinction between
geometric imperfections and support stiffness variations is unimportant to some extent
as both sources contribute to the unevenness measured by track recording cars. Using
these data for a similar type of carriage should therefore allow the dynamic vehicle loads
to be predicted and the resulting ground-borne vibration to be assessed. The distinction
between geometric unevenness and parametric excitation is very important, however,
for determining the effectiveness of mitigation measures at the track, e.g. resilient rail
support systems [61], or track maintenance measures.
4 Prediction of Ground-Borne Vibration due to Railway Traffic

4.1 Numerical Models
The prediction of ground-borne vibration in buildings is usually performed in a two
step procedure, where the free-field response due to a running train is calculated first
and subsequently used for computating the building response. In the following, the
focus will be on the prediction of the free-field response and the building response is
not considered. For more information regarding the latter, the reader is referred to the
literature [44, 114, 149]. Furthermore, the discussion is limited to models that take into
account dynamic excitation, as this is of main interest for problems involving ground-
borne vibration.
Numerical prediction models for railway-induced ground vibration are generally ob-
tained by coupling submodels for the train, the track, and the soil. The dynamic be-
haviour of the train is usually represented by a relatively simple multi-body vehicle
model whereas a much larger modelling and computational effort is required for cap-
turing the dynamic behaviour of the track and the soil. In order to obtain accurate nu-
merical prediction of ground-borne vibration, detailed information is needed regarding
the parameters that characterize the dynamic behaviour of these components as well as
the excitation. Of particular importance are the dynamic soil characteristics that must
be derived indirectly from in situ geophysical tests [72, 94].
The prediction of the incident wave field usually also consists of two steps. First, the
train-track-soil interaction problem is solved in order to compute the dynamic vehicle
loads. For example this can be done by means of equation (3.5) when assuming trans-
lational invariance of the problem geometry and a perfect contact between the train and
the track. Second, the dynamic vehicle loads are applied to the track and the free field
response is computed. When a transfer function Hts (x, x , t) is available that relates the
response at a point x to the load at a point x on the track, this can be formally written
as follows:
na t

u(x , t) = Hts (xk (τ), x , t − τ)gk (τ)dτ (4.1)
k=1 −∞
where the coupling between the position xk (τ) of the k-th axle (k = 1, . . . , na ) and the
time history of the load gk (τ) through the time τ gives rise to the Doppler effect. Due to
this coupling, the expression cannot be rewritten in the simple form of equation (2.7)
unless the train is assumed to be at a fixed position, i.e. xk (τ) ≈ xk0 , as for calculating
the stationary part of the vibration response [148].
The transfer function Hts (x, x , t) between the track and the free field in equation (4.1)
determines how the load applied at a position x on the track is transferred to the soil,
to provide response at a point x . The transfer is determined by track-soil interaction,
which will give rise to waves of the coupled system that propagate along the track.
The resulting load transfer could be compared with a combined filtering in the time
and space domain. Generally, a distribution of the load over a larger area, determined
by the track width and load spreading in the direction along the track, will result in
a reduction in the high frequency vibration transmitted to the free field. This is due
to destructive interference of waves transmitted by different parts of the contact area.
An accurate prediction of the stress distribution at the track-soil interface is therefore
essential for predicting ground-borne vibration due to railway traffic, in particular when
the wavelength in the soil is comparable or smaller than the characteristic dimension of
the stress distribution area [134].
General purpose 3D finite element (FE) methods offer the largest flexibility in mod-
elling, but require appropriate procedures to avoid spurious reflections at the bound-
aries of the finite volume of soil accounted for in the analysis [36, 37]. Alternatively,
3D coupled finite element - boundary element (BE) methods can be used [47, 107]. The
versatility of 3D FE and 3D FE/BE models comes at a very high computational cost,
however. Dedicated models have therefore been developed that exploit the (assumed)
regularity of the track and the underlying soil.
When the track and the soil are regarded as translationally invariant, a Fourier trans-
formation with respect to the coordinate y along the track leads to an efficient solution
in the frequency-wavenumber domain [7, 45]:
+∞
ũ(x, ky , z, ω) = û(x, y, z, ω) exp(iky y) dy
−∞
+∞
1
û(x, y, z, ω) = ũ(x, ky , z, ω) exp(−iky y) dky (4.2)
2π −∞
A tilde above a variable denotes its representation in the frequency-wavenumber do-

main. In this so-called 2.5D methodology, a problem with 2D geometry is solved for
each frequency ω and wavenumber ky to compute ũ(x, ky , z, ω) and the 3D solution
û(x, y, z, ω) is recovered by an inverse Fourier transformation with respect to the
wavenumber ky . Because of their high computational efficiency and relatively mod-
est modelling effort, 2.5D methods have been applied by a large number of researchers
to study ground-borne vibration due to railway traffic at grade, for example the re-
search groups at ISVR [128, 129, 133], NGI [97], NTU [156], BAM [9], FEUP [21, 22],
Chalmers [76], and KU Leuven [94], as well as for underground railway traffic, for ex-
ample at Cambridge University [42, 43, 64], NTU [157], TU Munich [100], and ISVR
[5, 132].
One of the drawbacks of these 2.5D models is that, due to the assumed translational
invariance, they do not allow accounting for periodic rail support as in a conventional
ballasted track. This implies that the stress distribution under the sleepers is not entirely
correctly predicted, which is important at high frequencies when wavelengths in the

track and the soil are of the same order of magnitude as the sleeper dimensions. For the
same reason, 2.5D models cannot account for parametric excitation, unless represented
by an equivalent geometric unevenness [9]. In order to resolve these issues and take
into account the periodicity of the track structure, a similar methodology may be used
which is based on the Floquet instead of the Fourier transform [20, 26]. In this case, the
3D solution is obtained based on the discretization of a single periodic cell.
Still other types of models are required for the analysis of vibrations generated by
trains crossing transition zones, switches and crossings [47, 113]. Limitations arising
from the use of a linear translationally invariant or periodic model for the track can
be partially circumvented [72] by using a more detailed model for the train and the
track with a suitably chosen soil stiffness [70, 105, 106, 142] in the first step where the
dynamic vehicle loads are considered. The use of time domain finite element analysis
allows for the consideration of nonlinear components, e.g. rail pad behaviour, as well
as loss of contact between the wheel and the rail [106].
Numerical models have undoubtedly contributed to a better understanding of physi-
cal mechanisms in the generation of ground-borne vibration, providing insight into re-
sults of field measurements and vibration problems encountered in practice. In design,
numerical models are particularly useful in new-build situations where a new building
is to be built close to an existing track or tunnel or where a new railway line is to be
constructed close to existing buildings. They can be used to assess the environmental
impact of new railway lines or to develop and design mitigation measures aimed at
reducing vibration nuisance to an acceptable level.
There are also, however, important limitations in the use of numerical models which
can be attributed to model and model parameter uncertainty. First, the simplifications
introduced for modelling may be too restrictive for the model to be useful. The geome-
try of the track and the soil may not be translationally invariant or periodic, e.g. due to
the presence of transition zones in the track, inclined soil layers or heterogeneities in the
soil. Hunt and his co-workers have recently assessed a number of simplifying assump-
tions in the prediction of ground-borne vibration due to underground railway traffic.
Deviations of up to 10 dB have been found at particular locations and frequencies, when
assuming horizontal soil stratification and disregarding the slightly inclined nature of
soil layers [74], disregarding voids at the tunnel–soil interface [75] or leaving out of
consideration a neighbouring tunnel [89]. Second, even when situations are adequately
represented by simple model geometries, the model parameters are also subject to sig-
nificant uncertainty. For new-build situations, estimations will have to be made based
on prior experience or engineering judgement. In existing situations, measurements of
track unevenness [91] or in situ tests for identification of track and soil properties will
not allow all model parameter uncertainty to be eliminated. Schevenels et al. [124] have
recently quantified the uncertainty in the prediction of ground vibration transmission
that arises from the limited resolution of a Spectral Analysis of Surface Waves test.
Hunt and Hussein [63] have shown that deviations due to model parameter uncertainty
are expected to be of the same order of magnitude as those arising from model uncer-
tainty. Although the importance of prediction uncertainty is generally recognized, its
quantification in the numerical prediction of ground-borne vibrations has seldom been
addressed. This is essential, however, in order to arrive at robust predictions as needed

in engineering practice.
4.2 Empirical Models
Notwithstanding the large recent progress in the development of numerical models for
railway-induced ground vibration, they are still mainly used for research. In contrast,
engineering practice mostly still makes use of empirical methods. The ISO 14837-1
standard that provides general guidance on ground-borne noise and vibration arising
from rail systems [68] indicates that requirements for absolute predictions change dur-
ing the various stages of development and distinguishes between scoping models (ear-
liest stage), environmental assessment models (planning process) and detailed design
models (part of construction and design). This categorization is seen in many of the
empirical models in use but, of course, also applies to numerical models.
Examples of empirical methods include the procedures developed by the Federal
Railroad Administration (FRA) and the Federal Transit Administration (FTA) of the
U.S. Department of Transportation [52, 53], the method developed by the Swiss Federal
Railways (SBB) [90], the method of Madshus et al. [96] which was based on measure-
ments in Norway and Sweden, and the method of Hood et al. [58] which was developed
within the frame of the Channel Tunnel Rail Link project in the UK. The procedures
developed by FRA and FTA distinguish between the three different levels of assess-
ment of the ISO 14837-1 standard [68]. The Detailed Vibration Assessment is based on
a prediction technique developed by Bovey [18] and Nelson and Saurenman [104] and
presents a more elaborate method for the prediction of ground-borne vibrations and re-
radiated noise in buildings. The method developed by SBB [90] distinguishes between
two prediction models, VIBRA-1 and VIBRA-2, where the latter is more detailed and
considers, for example, frequency-dependent attenuation models. The empirical meth-
ods by Madshus et al. [96] and Hood et al. [58] follow a similar structure as the one by
SBB, and additionally consider the issue of prediction uncertainty.
The aforementioned empirical methods aim at predicting the quasi-stationary re-
sponse during a train passage and can be cast in the following general form of the ISO
14837-1 standard [68]:
A( f ) = S ( f )P( f )R( f ) (4.3)
where A( f ) is the magnitude of ground vibration, typically a root mean square value in
one-third octave bands for detailed design situations, S ( f ) is the source strength, P( f )
characterizes the propagation path, and R( f ) the receiver. ISO 14837-1 [68] stipulates
that each of these terms should be further divided into relevant components, which in-
teract and can only be assumed uncoupled in some situations for simplified models.
Leaving out of consideration the receiver term R( f ), equation (4.3) has the same struc-
ture as the one in equation (2.7) for a dynamic load at a fixed position. The latter ex-
pression is strictly speaking not valid for the narrow-band spectrum of the response due
to a moving dynamic load because of the Doppler effect. A reasonable estimate for the
quasi-stationary response is still obtained, however, when the train speed is relatively
small compared to the wave velocities in the soil [148].
In the Detailed Vibration Assessment of the U.S. Department of Transportation, the

term characterizing the propagation path P( f ) is determined from field measurements
by adding contributions from incoherent point sources at different positions along the
track, leading to the so-called line source transfer mobility. The source strength S ( f ),
which is termed the force density, is determined indirectly from the measured response
and the experimental line source transfer mobility. A database with source strengths
obtained at different sites can then be used to predict ground vibration levels at sites
where the line source transfer mobility has been determined. The advantage of this
method when compared with numerical models is that it inherently takes into account
characteristics of the vibration transmission at a given site by the direct use of measured
transfer functions. In this way, simplifying assumptions, such as the horizontal nature
of the soil stratification, and identification of the dynamic soil characteristics from in
situ geophysical tests are avoided.
In the methods by SBB [90] and Madshus et al. [96], the source magnitude S ( f ) in
equation (4.3) is eliminated by taking measurements at a reference distance r0 , leading
to:
P(r, f )
A(r, f ) = A(r0 , f ) (4.4)
P(r0 , f )
and the ratio P(r, f )/P(r0 , f ) is computed assuming attenuation with distance of the
form r−n for some value of n which may depend on frequency. Madshus et al. [96]
propose a reference distance of 15 m whereas a much shorter distance of 3 m is given
in Kuppelwieser and Ziegler [90]. At such a small distance from the source, however,
quasi-static excitation will significantly contribute to the response and equation (4.4)
does not allow for a valid extrapolation of the response to larger distances. Again, vi-
bration velocities A(r0 , f ) known from past measurements can be used for predictions
at other sites using equation (4.4), correcting for train speed [96] and track quality
[90, 96]. Madshus et al. [96] also indicate that the reference vibration level A(r0 , f ) and
the exponent n in the attenuation law depend on the soil type.
Equation (4.3), that provides the general framework for empirical models, only holds
approximately in case of a moving load due to the Doppler effect. As a result, the source
magnitude S ( f ) will also depend on the distance r. Empirical prediction methods have
nevertheless largely shown their value in practice by providing reasonable estimates
of vibration velocity levels. Crucial for the prediction quality is the availability of a
suitable source characterization, either in terms of a force density or a reference vibra-
tion level. Equation (3.6) that relates the dynamic train loads to the wheel and track
unevenness shows that this requires a match in the type and level of excitation, train
characteristics, as well as characteristics of the track and the soil. Matching soil con-
ditions are particularly important in the low frequency range where they significantly
affect the track receptance [83] and, therefore, the dynamic train loads. This implies
that the method cannot be used in situations where new train or track types are imple-
mented or at sites with deviating soil conditions. A study of prediction errors arising
from a mismatch in soil conditions for the Detailed Vibration Assessment of the U.S.
Department of Transportation was recently presented by Verbraken et al. [147].
4.3 Hybrid Models
The use of empirical methods is limited to those cases where a suitable characterization
of the source strength (force density, reference vibration level) and the vibration trans-
mission (line source transfer mobility, attenuation law) is available. In order to over-
come these limitations, empirical methods could be combined with numerical methods
in a hybrid or semi-empirical prediction procedure.
Hybrid prediction methods seem particularly appealing in the case of new tracks,
new rolling stock or modifications of existing tracks. Numerical models can here be
used to assess the influence of rolling stock characteristics and track design on ground
vibration, avoiding the need for in situ tests possibly requiring the construction of test
tracks. In the case of new tracks, a numerically predicted source strength can be intro-
duced in equation (4.3) and combined with an experimentally determined transfer func-
tion or attenuation law that inherently takes into account the particularities of vibration
transmission. Since, for the case of new track infrastructure, these transfer functions
are determined in the absence of the track, an additional correction may be needed.
If modifications are made to the rolling stock or track structure, numerical modelling
can be used to evaluate the change in source strength and vibration transfer in equation
(4.3), so as to correct the existing vibration level. Accurately predicting relative levels
of vibration due to changes in track or vehicle parameters is expected to be much more
reliable than making predictions in absolute terms [72].
Hybrid prediction methods may therefore allow the accuracy of numerical models to
be improved by an adequate characterization of the vibration transfer, while at the same
time providing the flexibility of numerical models to assess a wide range of rolling stock
and track parameters. In order to introduce results from numerical models in empirical
predictions, the terms in equation (4.3) or corrections to them could be computed by
direct simulation of experiments as e.g. prescribed in the Detailed Vibration Assessment
of the U.S. Department of Transportation or an analytical expression of these terms can
be derived [148].
5 Mitigation of Railway-Induced Vibrations
5.1 Mitigation at Source

Ground-borne vibrations can be controlled at different levels at the source (train-track-
soil interaction), on the transmission path, or at the receiver. In the following, the focus
is placed on mitigation measures at source and in the transmission path as these are
considered to be more effective and economical [57] than measures at the receiver such
as base isolation of the building or box in box arrangement of rooms [41]. Many of
the mitigation measures at source discussed in the following, however, will mainly lead
to a reduction in the frequency range of relevance to ground-borne noise. In general,
the mitigation of low frequency feelable vibration which involves surface waves with
longer wavelengths and larger penetration depths is far more difficult [140] than the
mitigation of high frequency ground vibration leading to ground-borne noise.
Wheel and Track Unevenness. At the source, train-track-soil interaction generated

by combined wheel and track unevenness as well as parametric excitation determine
the dynamic vehicle loads transferred to the track. In order to reduce dynamic vehicle
loads, one could therefore try to reduce the excitation, i.e. geometric unevenness and
parametric excitation, or modify the dynamic characteristics of the rolling stock and the
track. Reduction of combined wheel and track unevenness will only lead to vibration
mitigation when it is in the relevant range of wavelengths (table 2). Measures to reduce
excitation arising from wheels include wheel reprofiling or truing [102, 151]. Remov-
ing wheel flats will clearly avoid large impact forces and lead to a substantial reduction
of dynamic loads. Generally, however, excitation due to wheel out-of-roundness occurs
in the frequency range of interest for ground-borne noise. Wheel reprofiling will there-
fore mainly reduce excitation in this frequency range. For minimizing track unevenness,
techniques as rail grinding or reprofiling [50] can be used. Rail reprofiling has shown to
be effective for reducing irregularities with wavelengths between 30 mm and 500 mm
and will therefore mainly allow reducing ground-borne noise (table 2) and rolling noise.
Ballast tamping may allow unevenness to be reduced at longer wavelengths [140]. For
properly maintained vehicles and tracks, the possibilities of reducing the source of ex-
citation seem limited, however.
Rolling Stock. A second possibility of vibration mitigation by reducing dynamic ve-

hicle loads is through modification of the rolling stock characteristics [151]. Mirza et
al. [99] performed an elaborate study of the effect of vehicle parameters on railway-
induced vibration for a two car EMU train and found that the parameters of most in-
fluence are the stiffness of the primary suspension and the unsprung mass with higher
vibration levels occurring for stiffer primary suspensions and heavier unsprung masses.
Changes in the geometrical parameters (e.g. bogie and axle distance) mainly lead to
shifts in the one-third octave band spectrum [99], as these parameters affect the axle
and bogie passage frequencies determining the shape of the narrow-band response spec-
trum. For the range of vehicle parameters investigated by Mirza et al. [99], the effect
on the levels of ground-borne vibration is relatively small. A much larger change of the
primary suspension stiffness was considered in a test with a prototype of the Metropoli-
tan Atlanta Rapid Transit Authority (MARTA) C-car [102, 151] where a reduction of
vibration of up to 6 dB was found at 30 Hz. In another case of urban traffic, resilient
wheels were found to effectively reduce ground vibrations arising from rail defects [84].
Opportunities for reduction of ground-borne vibration by modification of rolling stock
characteristics may be limited for conventional trains, however. Considerable room for
improvement may exist for freight trains which generally lack a primary or secondary
suspension.
Track Structure. A third way to intervene with the aim of reducing ground-borne noise
and vibration is by modification of the track structure. In order to isolate dynamically
part of the track superstructure, resilient elements can be included at different levels
in the track structure. Examples of resilient track elements include resilient fasteners
[102], sleeper pads [17, 51, 70, 126], ballast mats [21, 53, 102, 150], and slab mats in
floating slab tracks [102, 151]. As a general rule, measures providing resilience at a
lower level in the track will dynamically isolate a larger part of the track mass and work
in a lower frequency range. Ballast mats and floating slab tracks therefore seem best
suited for treatment of ground-borne vibration whereas other measures are, at least at
first sight, mostly of benefit for reduction of ground-borne noise.
The working principle of resilient elements such as ballast mats [150] or floating
slab tracks [80, 151] is usually demonstrated by considering the force transmissibility,
i.e. the ratio of force applied to the system and transmitted force, of a single degree of
freedom (SDOF) system [140]. When considering a SDOF system consisting of a mass
including the vehicle’s unsprung mass and the track mass and a spring representing
the track stiffness, the force transmissibility is controlled by the coupled wheel/track
resonance frequency [140]. The resonance frequency and, therefore, the track stiffness
should be as low as possible for an effective vibration reduction. Resonance frequen-
cies between 8 and 16 Hz are reported for floating slab tracks in the USA and Canada
[102], the Singapore Mass Rapid Transit system [25] and the North-South high speed
train connection through the city of Antwerp [125]. In the following, the effect of in-
cluding additional resilience in the track is assessed through equations (3.5) and (4.1)
that determine the dynamic vehicle loads and the transfer of vibrations into the free
field, respectively. Geometric track unevenness is assumed to be the primary source of
excitation for the moment. Parametric excitation will be discussed at the end.
First, equation (3.5) shows that a reduction of the track stiffness will affect train-track
interaction and, therefore, the dynamic vehicle loads.√ Reduction of the track stiffness kt
will, for example, result in a lower frequency ω = kt /Mu at which resonance occurs of
the unsprung mass on the track stiffness. This leads to a shift of the frequency spectrum
of the dynamic vehicle load to lower frequencies and, in the case of unmodified wheel
and track unevenness, amplification of the dynamic vehicle load around the new reso-
nance frequency and a reduction at higher frequencies, in particular around the original
resonance frequency.
Second, the introduction of resilient elements will also modify the transfer of vibra-
tion from the track to the soil as characterized by the transfer function Hts (xk (τ), x , t−τ)
in equation (4.1). The introduction of a resilient layer will lead to a (reduced) cut-on
frequency of waves which predominantly travel in the upper part of the track on top
of the resilient layer. When the soil flexibility is disregarded, the cut-on frequencies of
propagating track waves are the natural frequencies of the 2D model of the track sec-
tion [94]. At sufficiently high frequencies, the vibration energy is effectively confined
in the upper track part and less energy is radiated into the soil. The actual reduction in
the free field is determined by the relation between the phase velocities of the waves
propagating in the track and Rayleigh waves in the soil [94]. In case of unmodified dy-
namic vehicle loads, an amplification of ground-borne vibration or negative insertion
loss values are found around the newly introduced cut-on frequency and a reduction
or positive insertion loss values at higher frequencies. For an effective vibration reduc-
tion, the cut-on frequency and, therefore, the stiffness of the resilient layer should be as
low as possible. The latter is limited, however, by the maximum allowable static track
deflection.
The combined effect of the modified dynamic load and the modification of the trans-
fer function determines the overall effect of the resilient element on the ground vibration
during a train passage. For an accurate and detailed quantification of the vibration reduc-
tion, three-dimensional numerical models are needed that allow for interaction between
waves propagating in the track and in the soil [62, 95].
In the previous discussion, geometric track unevenness was assumed to be the main
source of excitation. Adding resilience in the track system may also lead to a reduction
of parametric excitation due to differences in support stiffness [60, 70]. The additional
resilience in the track support will to some extent smooth variations in support stiffness,
e.g. due to variation in ballast or subgrade properties. In this way, resilient rail fastening
systems, rail pads or sleeper pads will lead to a reduction of vibration at relatively low
frequencies [60], which is unexpected from the discussion above.
Track Subgrade. Stiffening of subgrade under the track is essentially performed for
improving the bearing capacity of soft soils and avoiding excessive track settlements,
but has also been shown to be effective in reducing ground vibration [2], in particular
when quasi-static excitation is important [111]. Various techniques for soil stiffening
have been developed, some of which also allow for the treatment of soil under existing
tracks. Subgrade stiffening is not considered as a practical option for vibration mitiga-
tion by railway operators because the works generally require interruption of railway
operation.
5.2 Mitigation in the Transmission Path
Mitigation measures in the transmission path aim at impeding the propagation of elastic
waves travelling in the soil from the railway track to nearby buildings. Examples include
open trenches, soft and stiff wave barriers, wave impeding barriers, and placement of
heavy masses next to the track.
An open trench (figure 8a) is effective in cutting off the propagation of surface or
Rayleigh waves if the depth of the trench is at least half the Rayleigh wavelength
[119, 153]. The benefit is lost beyond a certain distance, however, due to the fact that
diffraction can occur around the bottom of the barrier. For stability reasons, an open
trench is limited to shallow depths, and the presence of ground water compromises its
efficiency. This necessitates the use of soft or stiff in-fill materials (figure 8b). When
including soft in-fill materials, the stiffness of the material should be as low as possi-
ble to limit the transmission of vibration and approach the ideal case of an open trench
[46, 69]. Gas cushions [98], rubber chips [81], and polystyrene [27] have been used for
creating soft wave barriers. Design rules that have been determined using calculations
and experiments [14, 59] indicate that impractical depths of trench are mostly required
to achieve high attenuations of vibration in a homogeneous ground at low frequencies.
Often, however, a relatively thin upper layer of soil is important in determining the
nature of surface wave propagation, with increased propagation for frequencies where
surface waves can propagate predominantly in this upper layer. A trench may be ex-
pected to have a greater effect in such a layered ground than in the homogeneous case
[48, 69, 73].
A stiff wave barrier [23] is realized by a concrete slab, a row of steel or concrete
piles [77], a sheet pile wall [34] or a jet grouting wall [24]. In this case, the lateral
stability of the screen is not an issue and the installation of the screen may be more
straightforward. A stiff wave barrier essentially behaves as a stiff beam embedded in a
softer material and the transmission of waves propagating in the soil is impeded when
the trace wavelength of the incident waves is smaller than the bending wavelength in
the stiffened block [24]. The mitigation measure will be more effective at sites with a
soft soil.
(a) (b) (c)
(d)
Fig. 8. Vibration mitigation measures on the transmission path: (a) open trech, (b) soft or stiff
wave barrier, (c) wave impeding block, and (d) heavy masses next to the track
Wave impeding blocks (figure 8c) are stiff inclusions placed under or next to the
railway track in order to modify the wave propagation in the soil [110, 138]. The idea
is based on the existence of a cut-off frequency in a soil. Below the cut-off frequency
(figure 3a), no propagating modes exist in the soft top layer. By installing a sufficiently
stiff wave impeding block in the soil [6, 133], an attempt is made to reduce the thickess
of the soft top layer so as to increase the cut-off frequency of the soil. In some cases
it has been found, however, that the vibration level is amplified at frequencies higher
than the cut-off frequency [110], so that appropriate care must be taken in the design.
Ideally, a rigid substratum with infinite lateral dimensions is created, in order to obtain
the desired cut-off of surface waves. In practice, however, only wave impeding blocks
with finite dimensions can be created. According to Takemiya and Fujiwara [138], the
width of the WIB should be sufficiently large compared with the dominant wavelength
in the soil.
A simple idea that has been proposed to reduce ground-borne vibration due to rail-
way traffic at grade is to place heavy masses such as gabion walls composed of stone
baskets or concrete blocks on the ground surface next to the track [71, 87]. These ir-
regularities are expected to cause a scattering of the incident surface waves, resulting
in a reduction of the transmitted wave field. When effective, these measures could be
designed in conjunction with noise barriers such as gabion walls in order to obtain an
integrated solution for ground-borne vibration and airborne noise. Up to now, few stud-
ies have been carried out, however about the practical use of masses for mitigation of
railway-induced vibration.
Initial numerical analysis and preliminary experiments show that mitigation mea-
sures on the transmission path offer the prospect of obtaining vibration reduction at low
frequencies [140]. More research is currently needed, however, to confirm these initial
findings.
6 Conclusion
Ground-borne vibration due to railway traffic is an important societal issue, which is of

particular concern in the extension of urban and high speed railway networks and for the
development of land next to railways. Studies in the past have led to improved insight
in the governing physical phenomena, numerical and empirical prediction methods, as
well as possible mitigation methods. In all of these areas, some open points are still
present, however, that require further investigation.
Governing physical phenomena. Besides geometric unevenness, parametric excita-
tion, which is due to spatial variation of the track stiffness, is also a source of railway-
induced ground vibration. The importance of this term is difficult to asses, however, due
to the fact that data of track stiffness measurements are scarce. More research is needed
to quantify the importance of this type of excitation which is essential for the develop-
ment of vibration mitigation measures at source. This also holds for some other types
of dynamic excitation sources, e.g. singular points such as switches and crossings.
Prediction methods. Numerical prediction methods have allowed for a better un-
derstanding of the governing physical phenomena but need to be applied with care in
practical design situations because of the simplifying assumptions involved (type of
excitation, horizontal soil stratification, free field conditions). More research is needed
to quantify prediction uncertainties as these are essential for robust predictions in en-
gineering practice and design. Empirical methods allow the particularities of vibration
transmission at a given site to be taken into account, but for new-build or modified situa-
tions their application is limited to cases for which suitable prior experience is available.
The limits of empirical prediction methods may be better understood by numerical sim-
ulations of the proposed procedures. Hybrid models, combining results from numerical
and empirical predictions, may be developed to overcome limitations of both methods.
Mitigation measures. Many of the mitigation measures at source discussed will lead
to a reduction of ground vibration in the frequency range of structure-borne noise, but
are less well suited for mitigation of low frequency feelable vibration. Effective reduc-
tion of low frequency vibration is hard due to the long wavelengths and large penetra-
tion depths of the corresponding surface waves. An improved insight in the governing
physical phenomena may allow identifying novel solutions for effective low frequency
mitigation measures.
Acknowledgements. The third author is a postdoctoral fellow of the Research Foun-

dation - Flanders. The financial support is gratefully acknowledged.
References
[1] Achenbach, J.: Wave propagation in elastic solids. North-Holland Series in Applied Math-
ematics and Mechanics, vol. 16. North-Holland, Amsterdam (1973)
[2] Adam, D., Vogel, A., Zimmermann, A.: Ground improvement techniques beneath existing
rail tracks. Ground Improvement 11(4), 229–235 (2007)
[3] Adolfsson, K., Andréasson, B., Bengtson, P.-E., Bodare, A., Madshus, C., Massarch, R.,
Wallmark, G., Zackrisson, P.: High speed lines on soft ground. Evaluation and analyses of
measurements from the West Coast Line. Technical report, Banverket, Sweden (1999)
[4] Aki, K., Richards, P.: Quantitative seismology, 2nd edn. University Science Books, Sausal-
ito (2002)
[5] Andersen, L., Jones, C.: Coupled boundary and finite element analysis of vibration from
railway tunnels – a comparison of two- and three-dimensional models. Journal of Sound
and Vibration 293(3-5), 611–625 (2006)
[6] Andersen, L., Nielsen, S.: Reduction of ground vibration by means of barriers or soil
improvement along a railway track. Soil Dynamics and Earthquake Engineering 25, 701–
716 (2005)
[7] Aubry, D., Clouteau, D., Bonnet, G.: Modelling of wave propagation due to fixed or mo-
bile dynamic sources. In: Chouw, N., Schmid, G. (eds.) Wave Propagation and Reduction
of Vibrations Workshop Wave 1994, pp. 109–121. Ruhr Universität Bochum, Germany
(1994)
[8] Auersch, L.: Wave propagation in layered soils: theoretical solution in wavenumber do-
main and experimental results of hammer and railway traffic excitation. Journal of Sound
and Vibration 173(2), 233–264 (1994)
[9] Auersch, L.: The excitation of ground vibration by rail traffic: theory of vehicle-track-soil
interaction and measurements on high-speed lines. Journal of Sound and Vibration 284(1-
2), 103–132 (2005)
[10] Auersch, L.: Ground vibration due to railway traffic – The calculation of the effects of mov-
ing static loads and their experimental verification. Journal of Sound and Vibration 293,
599–610 (2006)
[11] Bedford, A., Drumheller, D.: Introduction to elastic wave propagation. John Wiley and
Sons (1993)
[12] Berggren, E.: Railway Track Stiffness. Dynamic Measurements and Evaluation for Effi-
cient Maintenance. Ph.D. thesis, Royal Institute of Technology (KTH) (2009)
[13] Berggren, E., Kaynia, A., Dehlbom, B.: Identification of substructure properties of railway
tracks by dynamic stiffness measurements and simulations. Journal of Sound and Vibra-
tion 329(19), 3999–4016 (2010)
[14] Beskos, D., Dasgupta, B., Vardoulakis, I.: Vibration isolation using open or filled trenches.
Part I: 2-D homogeneous soil. Computational Mechanics 1, 43–63 (1986)
[15] Biot, M.: Theory of propagation of elastic waves in a fluid-saturated porous solid. I. low-
frequency range. Journal of the Acoustical Society of America 28(2), 168–178 (1956)
[16] Biot, M.: Theory of propagation of elastic waves in a fluid-saturated porous solid. II. high-
frequency range. Journal of the Acoustical Society of America 28(2), 179–191 (1956)
[17] Bongini, E., Lombaert, G., François, S., Degrande, G.: A parametric study of the impact
of mitigation measures on ground borne vibration due to railway traffic. In: De Roeck,
G., Degrande, G., Lombaert, G., Müller, G. (eds.) Proceedings of the 8th International
Conference on Structural Dynamics EURODYN 2011, pp. 663–670. CD-ROM, Leuven
(2011)
[18] Bovey, E.: Development of an impact method to determine the vibration transfer charac-
teristics of railway installations. Journal of Sound and Vibration 87(2), 357–370 (1983)
[19] British Standards Institution. BS 6472:1992: Evaluation of human exposure to vibration

in buildings (1 Hz to 80 Hz) (1992)
[20] Clouteau, D., Arnst, M., Al-Hussaini, T., Degrande, G.: Freefield vibrations due to dy-
namic loading on a tunnel embedded in a stratified medium. Journal of Sound and Vibra-
tion 283(1-2), 173–199 (2005)
[21] Costa, P., Calçada, R., Cardoso, A.: Ballast mats for the reduction of railway traffic vibra-
tions. numerical study. Soil Dynamics and Earthquake Engineering 42, 137–150 (2012)
[22] Costa, P., Calçada, R., Cardoso, A., Bodare, A.: Influence of soil non-linearity on the
dynamic response of high-speed railway tracks. Soil Dynamics and Earthquake Engineer-
ing 30(4), 221–235 (2010)
[23] Coulier, P., Dijckmans, A., Jiang, J., Thompson, D.J., Degrande, G., Lombaert, G.: Stiff
wave barriers for the mitigation of railway induced vibrations. In: Nielsen, J.C.O., Ander-
son, D., Gautier, P.-E., Iida, M., Nelson, J.T., Thompson, D., Tielkes, T., Towers, D.A.,
de Vos, P. (eds.) Noise and Vibration Mitigation for Rail Transportation Systems. NNFM,
vol. 126, pp. 539–546. Springer, Heidelberg (2015)
[24] Coulier, P., François, S., Degrande, G., Lombaert, G.: Subgrade stiffening next to the track
as a wave impeding barrier for railway induced vibrations. Soil Dynamics and Earthquake
Engineering 48, 119–131 (2013)
[25] Cui, F., Chew, C.: The effectiveness of floating slab track system - Part I. receptance meth-
ods. Applied Acoustics 61, 441–453 (2000)
[26] Degrande, G., Clouteau, D., Othman, R., Arnst, M., Chebli, H., Klein, R., Chatterjee, P.,
Janssens, B.: A numerical model for ground-borne vibrations from underground railway
traffic based on a periodic finite element - boundary element formulation. Journal of Sound
and Vibration 293(3-5), 645–666 (2006)
[27] Degrande, G., De Roeck, G.: Effectiviteit van een trillingsisolerend scherm in het station
van het Leopoldkwartier te Brussel. Report to Dynamic Engineering, Department of Civil
Engineering, KU Leuven (December 1992)
[28] Degrande, G., Lombaert, G.: An efficient formulation of Krylov’s prediction model for
train induced vibrations based on the dynamic reciprocity theorem. Journal of the Acous-
tical Society of America 110(3), 1379–1390 (2001)
[29] Deutsches Institut für Normung. DIN 45672 Teil 2: Schwingungsmessungen in der Umge-
bung von Schienenverkehrswegen: Auswerteverfahren (1995)
[30] Deutsches Institut für Normung. DIN 4150 Teil 2: Erschütterungen im Bauwesen, Ein-
wirkungen auf Menschen in Gebäuden (1999)
[31] Deutsches Institut für Normung. DIN 4150 Teil 3: Erschütterungen im Bauwesen, Ein-
wirkungen auf bauliche Anlagen (1999)
[32] Dieterman, H., Metrikine, A.: The equivalent stiffness of a halfspace interacting with a
beam. Critical velocities of a moving load along the beam. European Journal of Mechanics,
A/Solids 15(1), 67–90 (1996)
[33] Dieterman, H., Metrikine, A.: Critical velocities of a harmonic load moving uniformly
along an elastic layer. Journal of Applied Mechanics, Transactions of the ASME 64(3),
596–600 (1997)
[34] Dijckmans, A., Ekblad, A., Smekal, A., Degrande, G., Lombaert, G.: A sheet piling wall
as a wave barrier for train induced vibrations. In: Papadrakakis, M., Papadopoulos, V.,
Plevris, V. (eds.) Proceedings of the 4th International Conference on Computational Meth-
ods in Structural Dynamics and Earthquake Engineering, COMPDYN 2013, Kos Island,
Greece. CD-ROM (June 2013)
[35] Ditzel, A., Herman, G., Drijkoningen, G.: Seismograms of moving trains: a comparison of
theory and measurements. Journal of Sound and Vibration 248(4), 635–652 (2001)
[36] Ekevid, T., Lane, H., Wiberg, N.-E.: Adaptive solid wave propagation – influences of
boundary conditions in high-speed train applications. Computer Methods in Applied Me-
chanics and Engineering 195, 236–250 (2006)
[37] Ekevid, T., Wiberg, N.-E.: Wave propagation related to high-speed train. A scaled bound-
ary FE-approach for unbounded domains. Computer Methods in Applied Mechanics and
Engineering 191, 3947–3964 (2002)
[38] Esveld, C.: Modern railway track, 2nd edn. MRT-Productions, Zaltbommel (2001)
[39] Fenton, G.: Random field modeling of CPT data. Journal of Geotechnical and Geoenvi-
ronmental Engineering, Proceedings of the ASCE 125(6), 486–535 (1999)
[40] Fiala, P., Degrande, G., Augusztinovicz, F.: Numerical modelling of ground borne noise
and vibration in buildings due to surface rail traffic. Journal of Sound and Vibration 301(3-
5), 718–738 (2007)
[41] Fiala, P., Gupta, S., Degrande, G., Augusztinovicz, F.: A parametric study on counter-
measures to mitigate subway traffic induced vibration and noise in buildings. In: Sas, P.,
Bergen, B. (eds.) Proceedings of ISMA 2008 International Conference on Noise and Vi-
bration Engineering, Leuven, pp. 2751–2764 (September 2008)
[42] Forrest, J., Hunt, H.: Ground vibration generated by trains in underground tunnels. Journal
of Sound and Vibration 294, 706–736 (2006)
[43] Forrest, J., Hunt, H.: A three-dimensional tunnel model for calculation of train-induced
ground vibration. Journal of Sound and Vibration 294, 678–705 (2006)
[44] François, S., Pyl, L., Masoumi, H., Degrande, G.: The influence of dynamic soil-structure
interaction on traffic induced vibrations in buildings. Soil Dynamics and Earthquake En-
gineering 27(7), 655–674 (2007)
[45] François, S., Schevenels, M., Lombaert, G., Galvín, P., Degrande, G.: A 2.5D coupled
FE-BE methodology for the dynamic interaction between longitudinally invariant struc-
tures and a layered halfspace. Computer Methods in Applied Mechanics and Engineer-
ing 199(23-24), 1536–1548 (2010)
[46] François, S., Schevenels, M., Thyssen, B., Borgions, J., Degrande, G.: Design and effi-
ciency of a vibration isolating screen in the soil. Soil Dynamics and Earthquake Engineer-
ing 39, 113–127 (2012)
[47] Galvín, P., Romero, A., Domínguez, J.: Fully three-dimensional analysis of high-speed
train–track–soil–structure dynamic interaction. Journal of Sound and Vibration 329,
5147–5163 (2010)
[48] Garcia-Bennett, A., Jones, C.J.C., Thompson, D.J.: A numerical investigation of railway
ground vibration mitigation using a trench in a layered soil. In: Maeda, T., Gautier, P.-E.,
Hanson, C.E., Hemsworth, B., Nelson, J.T., Schulte-Werning, B., Thompson, D., de Vos,
P. (eds.) Noise and Vibration Mitigation for Rail Transportation Systems. NNFM, vol. 118,
pp. 315–322. Springer, Heidelberg (2012)
[49] Gidlöf-Gunnarsson, A., Ögren, M., Terson, J., Öhrström, E.: Railway noise annoyance
and the importance of number of trains, ground vibration, and building situational factors.
Noise Health 14, 190–201 (2012)
[50] Grassie, S.: Rail irregularities, corrugation and acoustic roughness: characteristics, signifi-
cance and effects of reprofiling. Journal of Rail and Rapid Transit 226(5), 542–557 (2012)
[51] Guigou-Carter, C., Villot, M., Guillerme, B., Petit, C.: Analytical and experimental study
of sleeper SAT S312 in slab track SATEBA system. Journal of Sound and Vibration 293(3-
5), 878–887 (2006); Proceedings of the 8th International Workshop on Railway Noise
[52] Hanson, C., Towers, D., Meister, L.: High-speed ground transportation noise and vibration
impact assessment. HMMH Report 293630-4, U.S. Department of Transportation, Federal
Railroad Administration, Office of Railroad Development (October 2005)
[53] Hanson, C., Towers, D., Meister, L.: Transit noise and vibration impact assessment. Report
FTA-VA-90-1003-06, U.S. Department of Transportation, Federal Transit Admininistra-
tion, Office of Planning and Environment (May 2006)
[54] Haskell, N.: The dispersion of surface waves on multilayered media. Bulletin of the Seis-
mological Society of America 73, 17–43 (1953)
[55] He, X., Kawatani, M., Nishiyama, S.: An analytical approach to train-induced site vibra-
tion around shinkansen viaducts. Structure and Infrastructure Engineering: Maintenance,
Management, Life-Cycle Design and Performance 6(6), 689–701 (2010)
[56] Heckl, M., Hauck, G., Wettschureck, R.: Structure-borne sound and vibration from rail
traffic. Journal of Sound and Vibration 193(1), 175–184 (1996)
[57] Hemsworth, B.: Reducing groundborne vibrations: state of the art study. Journal of Sound
and Vibration 231(3), 703–709 (2000)
[58] Hood, R., Greer, R., Breslin, M., Williams, P.: The calculation and assessment of ground-
borne noise and perceptible vibration from trains in tunnels. Journal of Sound and Vibra-
tion 193, 215–225 (1996)
[59] Hung, H., Yang, Y., Chang, D.: Wave barriers for reduction of train-induced vibrations in
soils. Journal of Geotechnical Engineering, Proceedings of the ASCE 130(12), 1283–1291
(2004)
[60] Hunt, H.: Modelling of rail roughness for the evaluation of vibration-isolation measures.
In: 12th International Congress on Sound and Vibration, Lisbon, Portugal (July 2005)
[61] Hunt, H.E.M.: Types of rail roughness and the selection of vibration isolation measures. In:
Schulte-Werning, B., Thompson, D., Gautier, P.-E., Hanson, C., Hemsworth, B., Nelson,
J., Maeda, T., de Vos, P. (eds.) Noise and Vibration Mitigation for Rail Transportation
Systems. NNFM, vol. 99, pp. 341–347. Springer, Heidelberg (2008)
[62] Hunt, H., Forrest, J.: Floating slab track for vibration reduction: why simple models don’t
work. In: 7th International Congress on Sound and Vibration, Garmisch-Partenkirchen,
Germany (July 2000)
[63] Hunt, H., Hussein, M.: Vibration from railways: can we achieve better than +/-10 dB pre-
diction accuracy? In: 14th International Congress on Sound and Vibration, Cairns, Aus-
tralia (July 2007)
[64] Hussein, M., Hunt, H.: A numerical model for calculating vibration from a railway tunnel
embedded in a full-space. Journal of Sound and Vibration 305, 401–431 (2007)
[65] Hyde, J., Lintern, H.: The vibrations of roads and structures. In: Minutes of Proceedings
of the Institution of Civil Engineers, pp. 187–243 (1929)
[66] International Organization for Standardization. ISO 2631-2:1997: Mechanical vibration
and shock - Evaluation of human exposure to whole-body vibration - Part 1: General re-
quirements, 2nd edn. (1997)
[67] International Organization for Standardization. ISO 2631-2:1999: Mechanical vibration
and shock - Evaluation of human exposure to whole-body vibration - Part 2: Vibration in
buildings (1 to 80 Hz) (1999)
[68] International Organization for Standardization. ISO 14837-1:2005 Mechanical vibration
- Ground-borne noise and vibration arising from rail systems - Part 1: General guidance
(2005)
[69] Jiang, J., Toward, M.G.R., Dijckmans, A., Thompson, D.J., Degrande, G., Lombaert, G.,
Ryue, J.: Reducing railway induced ground-borne vibration by using trenches and buried
soft barriers. In: Nielsen, J.C.O., Anderson, D., Gautier, P.-E., Iida, M., Nelson, J.T.,
Thompson, D., Tielkes, T., Towers, D.A., de Vos, P. (eds.) Noise and Vibration Mitiga-
tion for Rail Transportation Systems. NNFM, vol. 126, pp. 555–562. Springer, Heidelberg
(2015)
[70] Johansson, A., Nielsen, J., Bolmsvik, R., Karlstrom, A., Lunden, R.: Under sleeper pads -
Influence on dynamic train-track interaction. Wear 265, 1479–1487 (2008)
[71] Jones, C.: Use of numerical-models to determine the effectiveness of anti-vibration sys-
tems for railways. Proceedings of the Institution of Civil Engineers-Transport 105(1),
43–51 (1994)
[72] Jones, C., Block, J.: Prediction of ground vibration from freight trains. Journal of Sound
and Vibration 193(1), 205–213 (1996)
[73] Jones, C., Thompson, D., Andreu-Medina, J.: Initial theoretical study of reducing surface-
propagating vibration from trains using earthworks close to the track. In: Proceedings
of the 8th International Conference on Structural Dynamics, EURODYN 2011, Leuven,
Belgium, pp. 684–691 (July 2011)
[74] Jones, S., Hunt, H.: The effect of inclined soil layers on surface vibration from under-
ground railways using the thin layer method. ASCE Journal of Engineering Mechan-
ics 137(12), 887–900 (2011)
[75] Jones, S., Hunt, H.: Voids at the tunnel–soil interface for calculation of ground vibration
from underground railways. Journal of Sound and Vibration 330(2), 245–270 (2011)
[76] Karlström, A., Boström, A.: An analytical model for train induced ground vibrations from
railways. Journal of Sound and Vibration 292, 221–241 (2006)
[77] Kattis, S., Polyzos, D., Beskos, D.: Vibration isolation by a row of piles using a 3-D fre-
quency domain BEM. International Journal for Numerical Methods in Engineering 46,
713–728 (1999)
[78] Kausel, E.: Fundamental solutions in elastodynamics: a compendium. Cambridge Univer-
sity Press, New York (2006)
[79] Kausel, E., Roësset, J.: Stiffness matrices for layered soils. Bulletin of the Seismological
Society of America 71(6), 1743–1761 (1981)
[80] Kawaharazuka, T., Hiramatsu, T., Ohkawa, H., Koyasu, M.: Experimental study on vi-
bration reduction by isolated railway. In: Proceedings of InterNoise 1996, pp. 1549–1552
(1996)
[81] Kim, M., Lee, P., Kim, D., Kwon, H.: Vibration isolation using flexible rubber chip bar-
riers. In: Schmid, G., Chouw, N. (eds.) Proceedings of the International Workshop Wave
2000, Wave Propagation, Moving Load, Vibration Reduction, pp. 289–298. A.A Balkema,
Amsterdam (2000)
[82] Knothe, K., Grassie, S.: Modelling of railway track and vehicle/track interaction at high
frequencies. Vehicle Systems Dynamics 22, 209–262 (1993)
[83] Knothe, K., Wu, Y.: Receptance behaviour of railway track and subgrade. Archive of Ap-
plied Mechanics 68, 457–470 (1998)
[84] Kouroussis, G., Verlinden, O., Conti, C.: On the interest of integrating vehicle dynamics
for the ground propagation of vibrations: the case of urban railway traffic. Vehicle Systems
Dynamics 48(12), 1553–1571 (2010)
[85] Krylov, V.: On the theory of railway-induced ground vibrations. Journal de Physique
IV 4(C5), 769–772 (1994)
[86] Krylov, V.: Generation of ground vibrations by superfast trains. Applied Acoustics 44,
149–164 (1995)
[87] Krylov, V.: Scattering of rayleigh waves by heavy masses as method of protection against
traffic-induced ground vibrations. In: Takemiya, H. (ed.) Environmental Vibrations. Pre-
diction, Monitoring, Mitigation and Evaluation, pp. 393–398. Taylor and Francis Group,
London (2005)
[88] Krylov, V., Ferguson, C.: Recent progress in the theory of railway-generated ground vibra-
tions. Proceedings of the Institute of Acoustics 17(4), 55–68 (1995)
[89] Kuo, K., Hunt, H., Hussein, M.: The effect of a twin tunnel on the propagation of ground-
borne vibration from an underground railway. Journal of Sound and Vibration 330(25),
6203–6222 (2011)
[90] Kuppelwieser, H., Ziegler, A.: A tool for predicting vibration and structure-borne noise
immissions caused by railways. Journal of Sound and Vibration 193, 261–267 (1996)
[91] Lombaert, G., Degrande, G.: Ground-borne vibration due to static and dynamic axle loads
of InterCity and high speed trains. Journal of Sound and Vibration 319(3-5), 1036–1066
(2009)
[92] Lombaert, G., Degrande, G., Clouteau, D.: Numerical modelling of free field traffic in-
duced vibrations. Soil Dynamics and Earthquake Engineering 19(7), 473–488 (2000)
[93] Lombaert, G., Degrande, G., Galvín, P., Bongini, E., Poisson, F.: A comparison of pre-
dicted and measured ground vibrations due to high speed, passenger, and freight trains. In:
Maeda, T., Gautier, P.-E., Hanson, C.E., Hemsworth, B., Nelson, J.T., Schulte-Werning,
B., Thompson, D., de Vos, P. (eds.) Noise and Vibration Mitigation for Rail Transporta-
tion Systems. NNFM, vol. 118, pp. 231–238. Springer, Heidelberg (2012)
[94] Lombaert, G., Degrande, G., Kogut, J., François, S.: The experimental validation of a
numerical model for the prediction of railway induced vibrations. Journal of Sound and
Vibration 297(3-5), 512–535 (2006)
[95] Lombaert, G., Degrande, G., Vanhauwere, B., Vandeborght, B., François, S.: The control
of ground borne vibrations from railway traffic by means of continuous floating slabs.
Journal of Sound and Vibration 297(3-5), 946–961 (2006)
[96] Madshus, C., Bessason, B., Hårvik, L.: Prediction model for low frequency vibration from
high speed railways on soft ground. Journal of Sound and Vibration 193(1), 195–203
(1996)
[97] Madshus, C., Kaynia, A.: High-speed railway lines on soft ground: dynamic behaviour at
critical train speed. Journal of Sound and Vibration 231(3), 689–701 (2000)
[98] Massarsch, K.: Vibration isolation using gas-filled cushions. In: Proceedings of the Geo-
Frontiers 2005 Congress. American Society of Civil Engineers, Austin (2005)
[99] Mirza, A.A., Frid, A., Nielsen, J.C.O., Jones, C.J.C.: Ground vibrations induced by railway
traffic - the influence of vehicle parameters. In: Maeda, T., Gautier, P.-E., Hanson, C.E.,
Hemsworth, B., Nelson, J.T., Schulte-Werning, B., Thompson, D., de Vos, P. (eds.) Noise
and Vibration Mitigation for Rail Transportation Systems. NNFM, vol. 118, pp. 259–266.
Springer, Heidelberg (2012)
[100] Müller, K., Grundmann, H., Lenz, S.: Nonlinear interaction between a moving vehicle
and a plate elastically mounted on a tunnel. Journal of Sound and Vibration 310, 558–586
(2008)
[101] Nazarian, S., Desai, M.: Automated surface wave method: field testing. Journal of
Geotechnical Engineering, Proceedings of the ASCE 119(7), 1094–1111 (1993)
[102] Nelson, J.: Recent developments in ground-borne noise and vibration control. Journal of
Sound and Vibration 193(1), 367–376 (1996)
[103] Nelson, J.: Prediction of ground vibrations from train using seismic reflectivity methods
for a porous soil. Journal of Sound and Vibration 231(3), 727–737 (2000)
[104] Nelson, J., Saurenman, H.: A prediction procedure for rail transportation groundborne
noise and vibration. Transportation Research Record 1143, 26–35 (1987)
[105] Nielsen, J., Igeland, A.: Vertical dynamic interaction between train and track-influence of
wheel and rail imperfections. Journal of Sound and Vibration 187(5), 825–839 (1995)
[106] Nielsen, J., Oscarsson, J.: Simulation of dynamic train-track interaction with state-
dependent track properties. Journal of Sound and Vibration 275, 515–532 (2004)
[107] O’Brien, J., Rizos, D.: A 3D FEM-BEM methodology for simulation of high speed train
induced vibrations. Soil Dynamics and Earthquake Engineering 25, 289–301 (2005)
[108] ORE. Question C116: Wechselwirkung zwischen Fahrzeugen und gleis, Bericht Nr. 1:
Spektrale Dichte der Unregelmässigkeiten in der Gleislage. Technical report, Office for
Research and Experiments of the International Union of Railways, Utrecht, NL (1971)
[109] Oscarsson, J.: Simulation of train-track interaction with stochastic track properties. Vehicle
Systems Dynamics 37(6), 449–469 (2002)
[110] Peplow, A., Jones, C., Petyt, M.: Surface vibration propagation over a layered halfspace
with an inclusion. Applied Acoustics 56, 283–296 (1999)
[111] Peplow, A., Kaynia, A.: Prediction and validation of traffic vibration reduction due to
cement column stabilization. Soil Dynamics and Earthquake Engineering 27, 793–802
(2007)
[112] Peris, E., Woodcock, J., Sica, G., Moorhouse, A., Waddington, D.: Annoyance due to
railway vibration at different times of the day. Journal of the Acoustical Society of Amer-
ica 131(2), 191–196 (2012)
[113] Phillips, J.E., Nelson, J.T.: Analysis and design of new floating slab track for special track-
work using finite element analysis (FEA). In: Maeda, T., Gautier, P.-E., Hanson, C.E.,
Hemsworth, B., Nelson, J.T., Schulte-Werning, B., Thompson, D., de Vos, P. (eds.) Noise
and Vibration Mitigation for Rail Transportation Systems. NNFM, vol. 118, pp. 275–282.
Springer, Heidelberg (2012)
[114] Pyl, L., Degrande, G., Clouteau, D.: Validation of a source-receiver model for road traffic
induced vibrations in buildings. II: Receiver model. ASCE Journal of Engineering Me-
chanics 130(12), 1394–1406 (2004)
[115] Klæboe, R., Hårvik, L., Turunen-Rise, I.H., Madshus, C.: Vibration in dwellings from
road and rail traffic-part ii: exposure-effect relationships based on ordinal logit and logistic
regression models. Applied Acoustics 64, 89–109 (2003)
[116] Rayleigh, J.: On waves propagated along the plane surface of an elastic solid. Proceedings
of the London Mathematical Society 17, 4–11 (1887)
[117] Read, W.: Stress analysis for compressible viscoelastic materials. Journal of Applied
Physics 21, 671–674 (1950)
[118] Rhayma, N., Bressolette, P., Breul, P., Fogli, M., Saussine, G.: A probabilistic approach for
estimating the behavior of railway tracks. Engineering Structures 33, 2120–2133 (2011)
[119] Richart, F., Hall, J., Woods, R.: Vibrations of soils and foundations. Prentice-Hall, Engle-
wood Cliffs (1970)
[120] Rix, G., Lai, C., Spang Jr., A.: In situ measurement of damping ratio using surface
waves. Journal of Geotechnical and Geoenvironmental Engineering, Proceedings of the
ASCE 126(5), 472–480 (2000)
[121] Rizzo, F., Shippy, D.: An application of the correspondence principle of linear viscoelas-
ticity theory. SIAM Journal on Applied Mathematics 21(2), 321–330 (1971)
[122] Schevenels, M., Degrande, G., Lombaert, G.: The influence of the depth of the ground
water table on free field road traffic induced vibrations. International Journal for Numerical
and Analytical Methods in Geomechanics 28(5), 395–419 (2004)
[123] Schevenels, M., François, S., Degrande, G.E.: An ElastoDynamics Toolbox for MATLAB.
Computers & Geosciences 35(8), 1752–1754 (2009)
[124] Schevenels, M., Lombaert, G., Degrande, G., François, S.: A probabilistic assessment of
resolution in the SASW test and its impact on the prediction of ground vibrations. Geo-
physical Journal International 172(1), 262–275 (2008)
[125] Schillemans, L.: Impact of sound and vibration of the north-south high-speed railway
connection through the city of Antwerp, Belgium. Journal of Sound and Vibration 267,
637–649 (2003)
[126] Schulte-Werning, B., Asmussen, B., Behr, W., Degen, K.G., Garburg, R.: Advancements
in noise and vibration abatement to support the noise reduction policy strategy of deutsche
bahn. In: Maeda, T., Gautier, P.-E., Hanson, C.E., Hemsworth, B., Nelson, J.T., Schulte-
Werning, B., Thompson, D., de Vos, P. (eds.) Noise and Vibration Mitigation for Rail
Transportation Systems. NNFM, vol. 118, pp. 9–16. Springer, Heidelberg (2012)
[127] Semblat, J.-F., Pecker, A.: Waves and vibrations in soils: earthquakes, traffic, shocks, con-
struction works. IUSS Press, Pavia (2009)
[128] Sheng, X., Jones, C., Petyt, M.: Ground vibration generated by a harmonic load acting on
a railway track. Journal of Sound and Vibration 225(1), 3–28 (1999)
[129] Sheng, X., Jones, C., Petyt, M.: Ground vibration generated by a load moving along a
railway track. Journal of Sound and Vibration 228(1), 129–156 (1999)
[130] Sheng, X., Jones, C., Thompson, D.: A comparison of a theoretical model for quasi-
statically and dynamically induced environmental vibration from trains with measure-
ments. Journal of Sound and Vibration 267(3), 621–635 (2003)
[131] Sheng, X., Jones, C., Thompson, D.: A theoretical model for ground vibration from trains
generated by vertical track irregularities. Journal of Sound and Vibration 272(3-5), 937–
965 (2004)
[132] Sheng, X., Jones, C., Thompson, D.: Modelling ground vibrations from railways using
wavenumber finite- and boundary-element methods. Proceedings of the Royal Society A -
Mathematical, Physical and Engineering Sciences 461, 2043–2070 (2005)
[133] Sheng, X., Jones, C., Thompson, D.: Prediction of ground vibration from trains using the
wavenumber finite and boundary element methods. Journal of Sound and Vibration 293,
575–586 (2006)
[134] Steenbergen, M., Metrikine, A.: The effect of the interface conditions on the dynamic
response of a beam on a half-space to a moving load. European Journal of Mechanics,
A/Solids 26, 33–54 (2007)
[135] Stichting Bouwresearch. SBR deel A: Schade aan gebouwen door trillingen: meet- en
beoordelingsrichtlijn (2002)
[136] Stichting Bouwresearch. SBR deel B: Hinder voor personen in gebouwen door trillingen:
meet- en beoordelingsrichtlijn (2002)
[137] Stichting Bouwresearch. SBR deel C: Storing aan apparatuur door trillingen: meet- en
beoordelingsrichtlijn (2002)
[138] Takemiya, H., Fujiwara, A.: Wave propagation/impediment in a stratum and wave imped-
ing block (WIB) measured for SSI response reduction. Soil Dynamics and Earthquake
Engineering 13, 49–61 (1994)
[139] Takemiya, H., Bian, X.C.: Shinkansen high-speed train induced ground vibrations in view
of viaduct-ground interaction. Soil Dynamics and Earthquake Engineering 27, 506–520
(2007)
[140] Thompson, D.: Railway noise and vibration: mechanisms, modelling, and means of con-
trol. Elsevier, Oxford (2009)
[141] Thomson, W.: Transmission of elastic waves through a stratified solid medium. Journal of
Applied Physics 21, 89–93 (1950)
[142] Triepaischajonsak, N.: The influence of various excitation mechanisms on ground vibra-
tion from trains. Ph.D. thesis, University of Southampton (2011)
[143] Triepaischajonsak, N., Thompson, D.J., Jones, C.J.C., Ryue, J.: Track-based control mea-
sures for ground vibration. The influence of Quasi-static Loads and Dynamic Excitation.
In: Maeda, T., Gautier, P.-E., Hanson, C.E., Hemsworth, B., Nelson, J.T., Schulte-Werning,
B., Thompson, D., de Vos, P. (eds.) Noise and Vibration Mitigation for Rail Transportation
Systems. NNFM, vol. 118, pp. 249–257. Springer, Heidelberg (2012)
[144] Turunen-Rise, I., Brekke, A., Hårvik, L., Madshus, C., Klæboe, R.: Vibration in dwellings
from road and rail traffic-part i: a new norwegian measurement standard and classification
system. Applied Acoustics 64, 71–87 (2003)
[145] Udías, A.: Principles of seismology. Cambridge University Press, Cambridge (1999)
[146] Varandas, J., Hölscher, P., Silva, M.: Dynamic behaviour of railway tracks on transitions
zones. Computers and Structures 89, 1468–1479 (2011)
[147] Verbraken, H., Eysermans, H., Dechief, E., François, S., Lombaert, G., Degrande, G.: Ver-
ification of an empirical prediction method for railway induced vibration. In: Maeda, T.,
Gautier, P.-E., Hanson, C.E., Hemsworth, B., Nelson, J.T., Schulte-Werning, B., Thomp-
son, D., de Vos, P. (eds.) Noise and Vibration Mitigation for Rail Transportation Systems.
NNFM, vol. 118, pp. 239–247. Springer, Heidelberg (2012)
[148] Verbraken, H., Lombaert, G., Degrande, G.: Verification of an empirical prediction method
for railway induced vibrations by means of numerical simulations. Journal of Sound and
Vibration 330(8), 1692–1703 (2011)
[149] Villot, M., Ropars, P., Jean, P., Bongini, E., Poisson, F.: Modeling the influence of struc-
tural modifications on the response of a building to railway vibration. Noise Control Engi-
neering Journal 59(6), 641–651 (2011)
[150] Wettschureck, R., Kurze, U.: Einfügungsdämmass von Unterschottermatten. Acustica 58,
177–182 (1985)
[151] Wilson, G., Saurenman, H., Nelson, J.: Control of ground-borne noise and vibration. Jour-
nal of Sound and Vibration 87(2), 339–350 (1983)
[152] With, C., Metrikine, A., Bodare, A.: Identification of effective properties of the railway
substructure in the low-frequency range using a heavy oscillating unit on the track. Archive
of Applied Mechanics 80, 959–968 (2010)
[153] Woods, R.: Screening of surface waves in soils. Journal of the Soil Mechanics and Foun-
dation Division, Proceedings of the ASCE 94(SM4), 951–979 (1968)
[154] Wu, T., Thompson, D.: On the parametric excitation of the wheel/track system. Journal of
Sound and Vibration 278(4-5), 725–747 (2004)
[155] Xia, H., Zhang, N., Cao, Y.: Experimental study of train-induced vibrations of environ-
ments and buildings. Journal of Sound and Vibration 280, 1017–1029 (2005)
[156] Yang, Y., Hung, H.: A 2.5D finite-infinite element approach for modelling visco-elastic
bodies subjected to moving loads. International Journal for Numerical Methods in Engi-
neering 51, 1317–1336 (2001)
[157] Yang, Y., Hung, H.: Soil vibrations caused by underground moving trains. Journal of
Geotechnical and Geoenvironmental Engineering, Proceedings of the ASCE 134(11),
1633–1644 (2008)

Ground-Borne Vibration Due To Railway Traffic A Review of Excitation Mechanisms, Prediction Methods and Mitigation Measures

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ground-Borne Vibration Due To Railway Traffic A Review of Excitation Mechanisms, Prediction Methods and Mitigation Measures

Uploaded by

Copyright:

Available Formats

Ground-Borne Vibration due to Railway Traﬃc:

A Review of Excitation Mechanisms, Prediction

G. Lombaert1, G. Degrande1, S. François1 , and D.J. Thompson2

Summary. The aim of this paper is to provide a comprehensive overview of

c Springer-Verlag Berlin Heidelberg 2015 253

Fig. 1. Railway-induced vibrations in the built environment

2 Wave Propagation in the Soil

(λ + 2μ) = (λ + 2μ)(1 + 2βp i) (2.4)

(a) (b) (c)

Table 1. Dynamic soil characteristics at the site at Lincent, Belgium

2.2 Wave Field Generated by a Point Load at a Fixed Position

û(r, ω) = Ĥ(r, ω) fˆ(ω) (2.7)

2.3 Wave Field Generated by a Moving Point Load

3.2 Dynamic Excitation

ûa (ω) = ût (ω) + ûw/r (ω) (3.4)

Ground-borne vibration Ground-borne noise

excitation experienced by the axle in the frequency domain. Excitation at a frequency

Parametric Excitation. Parametric excitation is another source of excitation that re-

At present, the importance of this source of excitation is not fully understood, as

4 Prediction of Ground-Borne Vibration due to Railway Traﬃc

A tilde above a variable denotes its representation in the frequency-wavenumber do-

correctly predicted, which is important at high frequencies when wavelengths in the

addressed. This is essential, however, in order to arrive at robust predictions as needed

4.2 Empirical Models

A( f ) = S ( f )P( f )R( f ) (4.3)

In the Detailed Vibration Assessment of the U.S. Department of Transportation, the

4.3 Hybrid Models

5 Mitigation of Railway-Induced Vibrations

5.1 Mitigation at Source

Wheel and Track Unevenness. At the source, train-track-soil interaction generated

Rolling Stock. A second possibility of vibration mitigation by reducing dynamic ve-

5.2 Mitigation in the Transmission Path

(a) (b) (c)

Ground-borne vibration due to railway traﬃc is an important societal issue, which is of

Acknowledgements. The third author is a postdoctoral fellow of the Research Foun-

[19] British Standards Institution. BS 6472:1992: Evaluation of human exposure to vibration

You might also like

(λ + 2μ) = (λ + 2μ)(1 + 2βp i) (2.4)