Applied Acoustics 107 (2016) 10–18

Contents lists available at ScienceDirect

Applied Acoustics
journal homepage: www.elsevier.com/locate/apacoust

Correlation between rolling noise generation and rail roughness

of tangent tracks and curves in time and frequency domains
Javad Sadeghi a,⇑, Araz Hasheminezhad b
a
Centre of Excellence in Railway Transportation, Iran University of Science and Technology, Tehran, Iran
b
School of Railway Engineering, Iran University of Science and Technology, Tehran, Iran

a r t i c l e i n f o a b s t r a c t

Article history: Despite considerable advantages of the railway track over other means of transportation, noise pollution
Received 14 January 2015 is the main adverse consequence of railway transportation. The basic cause of railway noise is rail corru-
Received in revised form 30 October 2015 gation. Although characteristics of railway noise have been considerably studied in the literature, rail cor-
Accepted 4 January 2016
rugation effects on rolling noise generation in tangent tracks and the curves have not been sufficiently
Available online 10 February 2016
investigated. This research addresses the limitations of the current understanding of the rolling noise
generation by investigating rail corrugation effects on rolling noise in tangent tracks and curves. This
Keywords:
research was made based on the results obtained form a thorough field investigation carried out in a rail-
Tangent tracks
Curves
way line which includes tangents tracks and sharp curves. A track geometry recording car was used to
Railway noise measure rail corrugations. For this purpose, an indirect method was developed in this research to obtain
Rail corrugation rail corrugation patterns from the data recorded by the track recording car. The effectiveness of the new
Experiments method was shown. The induced noises were recorded using two particular types of microphones and
implementing the method suggested by the ISO 3095 Standard. The rolling noise signal was distinguished
from the total noise, using Butterworth Band-Pass Filtering. The role of rail corrugations in the rolling
noise was discussed. Correlations were made between various types of corrugations and the induced
noises. The results were presented and discussed in the spatial and frequency scales. Results obtained
have led to new findings in rail corrugation effects on rolling noise generation. This research paves a
way toward a better understanding of rolling noise sources and the parameters influencing the noise.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction
With the advent of the new generation of high-speed trains,
there is no doubt that railway is preferred over other means of
transportation mainly due to its speed as well as its environmental
factors such as less energy consumption and less pollution.
Nevertheless, the railway has negative impacts on the environ-
ment, which mostly concern noise. Furthermore, in terms of
physiological parameters related to sleep, contrary to expectations,
the impact of railway noise is much greater than the road and air
traffic [1]. Intermittent sounds and vibrations endanger people’s
mental health in the long term and lead to depression and fatigue.
Therefore, maintaining the railway noise within the allowable lim-
its is essential, particularly for the new generation of high-speed
trains [2]. There are various sources of railway noise. The dominant
source of noise differs, as the track conditions vary. However, the
⇑ Corresponding author.
E-mail addresses: Javad_sadeghi@iust.ac.ir (J. Sadeghi), Araz_hasheminezhad@
rail.iust.ac.ir (A. Hasheminezhad).
most important type of railway noise is rolling noise [2]. Rolling
noise is generated by the vertical vibration of the wheel and the
rail. This vibration is induced by the relative displacements
between the rail and the wheel because of the roughness on their
surfaces. Rail corrugation is an important type of surface rough-
ness, which constitutes an undesirable periodic wear pattern on
the contact surface of the rails [2]. Short-pitch corrugation is a less
understood type of corrugation. This type of corrugation has wave-
lengths in the range of 25–80 mm. An investigation of this form of
corrugation is of interest when considering acoustic roughness
development. Although a great deal of work has been published
on the subject of corrugation, there is not much literature available
on the development of acoustic roughness [1]. An accurate knowl-
edge of rail and wheel roughness and its effect on the noise is an
essential factor in noise control policies. That is, a more mainte-
nance attention should be paid on the high roughness levels. The
maintenance includes re-profiling the wheels and grinding the
rails. The process of rail grinding to reduce roughness and corruga-
tion is time-consuming and expensive. In order to optimize the
minimization method of rail roughness, one should be aware of

http://dx.doi.org/10.1016/j.apacoust.2016.01.006
0003-682X/Ó 2016 Elsevier Ltd. All rights reserved.
 J. Sadeghi, A. Hasheminezhad / Applied Acoustics 107 (2016) 10–18
the rail wear and roughness development mechanisms and hence
the costs of noise control and its maintenance. One of the main
topics in railway acoustics is the role of the infrastructure and roll-
ing stock in the total noise radiated into the environment [1]. In
this regard, the most controversial questions is related to the noise
source. The controversy primarily focused on whether the wheel or
the rail is the main source. The physics of noise generation by
means of the roughness on the running surfaces of the wheels
and rails is very complex [3–5]. Consequently, to fully understand
the main controlling factor of wheel/rail noise generation,
researchers have relied on the results obtained from analytical
models. Although the theoretical modelling approach can deter-
mine the wavelength of some corrugations types, they have limita-
tions in explanation of all corrugation formations [6]. This remains
corrugation as an area which needs further researches. Despite the
extensive studies of rail corrugations including short-pitch corru-
gation and rolling contact corrugation developments in the last
30 years, there is no clear sense on the effect of these two impor-
tant and conventional types of corrugation on the railway rolling
noise generation. This research is a response to this need.
2. Review of literature
Railway noise has been studied extensively since the late 1970s
[1–3]. Remington [3,4] in 1976 developed the first theoretical
model of rolling noise. In 1993, Thompson [1,5,6] extended Rem-
ington’s theory and developed the model as a computer program
called TWINS. During the last decade, the sources of rolling noise
and the importance of surface roughness in railway noise has been
well studied using TWINS. Results indicate that the roughness of
the running surface of the rail and the wheel is the predominant
source of railway noise. In 2006, Kitagawa and Thompson [7]
investigated the impact of wheel/rail noise radiation on Japanese
railways using the TWINS model and microphone array measure-
ments. Hardy and his colleagues [8] carried out a statistically based
study into the effects of railhead roughness on the rolling noise
prediction. Their works were further developed by Nielsen et al.
[9–11], Gullers et al. [12] and Jiang et al. [13]. Although, theoretical
models for prediction of noise such as TWINS [14–22] provide a
clear understanding of railway noise sources, they just import
the roughness levels into the model and do not take into account
the role of the track system in the development of roughness or
corrugation over the time. Moreover, although the type of rail cor-
rugation in the case of short wavelength corrugation can have a
large influence on the rolling noise, it has not been considered in
the noise prediction models. Most of the previous studies have
used the experimental results of measured noise during train
pass-by and compared them with those predicted by the TWINS
program. The experimental works reported in the available litera-
ture are very limited and just used for the validation of TWINS.
A review of the literature [23–30] indicates that surface rough-
ness and wheel–rail contact, as important factors in railway rolling
noise, have not been adequately investigated. That is, there is a
need for further investigations (in particular field investigation)
into the effect of rail irregularities such as rail corrugation on roll-
ing noise. In response to this need, these authors have made a lim-
ited investigation into the effects of corrugations on rolling noise in
the time domain [2]. Their research was developed further in this
paper in both time and frequency domains using an improved
filtering approach. That is, this research is a comprehensive
experimental study on noise and rail corrugation in railway track
system, leading to develop correlations between corrugation and
rolling noise. In this paper, an indirect method for rail corrugation
measurement was developed. Using this method, rail corrugation
(with various amplitudes and wavelengths) was precisely obtained
11
from a track recording car. Due to the speed and availability of
track recording cars, this method is cost effective compared with
the conventional ones. This research has focused on the effects of
particular rail corrugation types (i.e., short-pitch rail corrugation
and rolling contact corrugation) on only one type of railway noise
(i.e. rolling noise) as this has not been attempted in the available
literature. For this purpose, railway noise was measured from the
passage of a railway vehicle using microphones in a railway field.
Using the vertical longitudinal profile of the railhead recorded by
the EM50 track recording car, rail corrugation in the tangent track
and the curve was derived [2]. The short-pitch rail corrugation
signal and rolling contact fatigue signal were extracted from the
corrugation signals by means of an improved Band-Pass Filtering.
Utilizing a new procedure, a rolling noise signal was extracted from
the original recorded noise signal for the tangent track and the
curve. Through analyses of rail corrugation and rolling noise in
various track conditions, correlations between the rail roughness
effect and rolling noise generation were derived and compared in
the frequency and time domains.
3. Rail corrugation and noise in time and frequency domains
The conditions required to achieve reliable and repeatable mea-
surements of the noise emitted by railway system have been
defined in standard ISO 3095 [14]. This standard also provides us
with the descriptions of processing techniques for rail roughness
measurements. This research was made according to ISO 3095.
For the field tests, noise prone areas were identified, such that
the noise generated by the passage of a railway vehicle could be
recorded using measurement equipment. As a result, Tehran metro
was identified as the most appropriate location for the field tests
[2]. Field noise measurements and the corrugation measurements
in a tangent track and a sharp curve were carried out initially in
2013, and their results were presented in [2]. Improved, extended
and completed field measurements were conducted in this
research (2014–15). The lines tested were exposed to severe corru-
gation growth. The metro line under investigation including loca-
tion of the field tests (the tangent track and the curve) is
presented in Fig. 1. Since one of the aims of this research was to
find the main reason of high level of noise near residential areas,
the test site is located near a region of high residential density,
with limited but significant traffic. The track segment in this inves-
tigation includes an approximately 50 m length tangent track and a
curve with an approximately 300 m radius and 25 m in length. The
track is ballasted with CWR-UIC 60 rail, rail cant of 1:20 rested on
B70 concrete sleepers and a Vossloh fastening system [2]. Accord-
ing to ISO 3095, direct roughness measurements are taken over the
reference section, whose length is proportional to the microphone
distance r from the track and varies from 2r to +2r (i.e.; 4r)
(r = 7.5 m) relative to the center of the reference section where
the noise measurement microphone is positioned. In addition,
the track at the measuring section shall be laid without rail joints
(welded rail) and free of visible surface defects such as rail burns or
pits and spikes caused by the compression of external material
between wheel and rail: no audible impact noise due to welds or
loose sleepers should be presented [14]. According to ISO 3095,
for Indirect roughness measurement as used in this paper, this dis-
tance can be considered more depending on track conditions such
as welded rail and being free of visible surface defects such as rail
burns or pits and spikes caused by the compression of external
material between wheel and rail [14]. So in this paper the mea-
surement distance in the test was considered about 50 m for the
tangent track and 25 m for the curve length which is in accordance
with ISO 3095.
As indicated above, the track instrumentation for noise mea-
surements were performed in accordance with the ISO 3095
12 J. Sadeghi, A. Hasheminezhad / Applied Acoustics 107 (2016) 10–18

Fig. 1. Metro line under investigation.

standard [14]. Based on this standard, the microphones were posi-

tioned at a distance of 7.5 m from the track center and at a height
of 1.2 m above the top of the rail. For unification of the tests, noise
was recorded in these sections using only one type of train
(TM3-51). In this research, noise has been measured with two
types of microphones, AT8015 CT and E-818 SII. Both are capable
of recording noise within the frequency range of railway rolling
noise and human hearing (20 Hz–20 kHz) [2]. The noise recorded
by microphones was captured by a laptop and the Spectra-LAB
Software. Characteristics of the microphones used in the field tests
for the noise measurements are provided in Table 1. Figs. 2 and 3
indicate the test set up, the microphone position and the test con-
ditions for the tangent track and the curve.
In this paper, the noise measurement signals were filtered using
an improved 4th order Butterworth Band-Pass Filter. It was made
to extract the frequency range of rolling noise, i.e., 100–5000 Hz
Fig. 2. Test set up and position of the microphone in the curve.
[2]. Figs. 4 and 5 illustrate the original recorded noise signal and
the extracted rolling noise signal using an improved Band-Pass
Filtering for the tangent track and the curve. tangent track and the curve obtained from the EM50 recording
The measured noise in Figs. 4 and 5 indicates that the total car. Since rail corrugation appears as periodic unevenness on rail
noise levels in the curved sections of the track were higher than spatial axis, the results are presented in spatial axis rather than fre-
the ones in the tangent track. This proves that the track properties quency scale. That is, rail corrugation components might be
have influences on acoustic behavior of the rail. In addition, the removed if the results are presented in frequency scale (it depends
rolling noise role in the overall noise level is significant in the tan- on the speed of track recording car). In the frequency scale, it is
gent track when compared with the curve. The measured noise in also likely that other components rather than rail corrugation are
Figs. 6 and 7 indicate that the noise in the curve appeared in the appeared in the recorded signal. Roughness signals often contain
frequency domain of 250–4000 Hz with one peak at 4000 Hz, while sharp spikes relating to pits or spikes which are on the surface.
the noise at the tangent track appeared in the frequency domain of These defects can influence the roughness spectrum level, though
250–2000 Hz with two peaks at 500 Hz and 1200 Hz. they are not completely sensed by the wheel/rail contact. The
The line tested has had severe rail corrugations. It is indicated in advantage of the spatial scale is the possibility of recognizing these
Figs. 8 and 9 which indicate the rolling contact fatigue corrugation defects. This was made in this paper using Band-Pass filtering tech-
and the short-pitch rail corrugation in the curve and the tangent nique on spatial axis.
track, respectively. Rail corrugations were measured using a track To produce a vertical profile of the railhead from the analysis of
geometry car called EM50 as shown in Fig. 10. Fig. 11 illustrates the data obtained from the EM50 track recording car, an algorithm
samples of the vertical profile of the railhead measured in the was developed using Simpson’s area rule. This algorithm was

Table 1
Characteristics of two microphones used in field tests for noise measurements.

Type AT8015 CARACTERISTIQUES TECHNIQUES E-818 SII

Picture

Frequency Response 40–20000 Hz 80–20000 Hz

Weight 195 g 330 g
Dimensions 460 mm 180 mm
 J. Sadeghi, A. Hasheminezhad / Applied Acoustics 107 (2016) 10–18 13

Fig. 3. Test set up and position of the microphone in the tangent track. Fig. 6. Frequency spectra of recorded noise for the curve.

developed by taking the slope of the rail at each point equal to the
difference between the EM50 recorded data at the point [2]. Figs. 12
and 13 indicate the vertical profile of the railhead for the curve and
the tangent track using the developed algorithm. A comparison
between the vertical profiles using the EM50 data and that
obtained from the algorithm for the curve and tangent track are
presented in Figs. 14 and 15 [2]. As indicated in these figures, there
is a good agreement between the vertical profiles using the EM50
data and that obtained from the algorithm. Fig. 16 shows the resul-
tant vertical profile for the curve and the tangent track.
Rail corrugation is originally an oscillatory wear of rail surface
due to the interaction between the rail and the wheel. Since the
main purpose of this paper is to use the wavelength of irregulari-
ties which are relevant to rolling noise, wavelengths in the range
of about 25–80 mm and 150–450 mm were used for the short-
pitch rail corrugation and the rolling contact fatigue corrugation Fig. 7. Frequency spectra of recorded noise for the tangent track.
respectively [2]. In this paper, an improved 4th order Butterworth

15
Recorded Noise Signal
Rolling Noise Signal
Sound Pressure Level (dB)

10

-5

-10
0 2000 4000 6000 8000 10000 12000
Time (milliseconds)

Fig. 4. Extracted rolling noise from original noise signal at the curve.

12
Recorded Noise Signal
10
Sound Pressure Level (dB)

Rolling Noise Signal

8
6
4
2
0
-2
-4
-6
-8
0 2000 4000 6000 8000 10000 12000 14000 16000 18000
Time (milliseconds)

Fig. 5. Extracted rolling noise from original noise signal at the tangent track.
14 J. Sadeghi, A. Hasheminezhad / Applied Acoustics 107 (2016) 10–18
Fig. 8. Rolling contact fatigue corrugation at curve.
Fig. 9. Short-pitch rail corrugation in tangent track.
Band-Pass Filter was implemented to extract frequency ranges of
135–450 Hz of the short-pitch rail corrugation and 25–75 Hz of
the rolling contact fatigue corrugation. Fig. 17 depicts the extracted
corrugation signal against the developed rail corrugation signal for
Fig. 10. EM50 track recording car used in Tehran Railway Network.
the short-pitch rail corrugation and the rolling contact fatigue cor-
rugation using the 4th order Butterworth Band-Pass Filter.
This paper focuses only on track components and their
parameters. That is, rail corrugation is examined rather than wheel
corrugation. According to the previous studies [20–22], the wheel
roughness levels on disc-braked vehicles with composite brake
blocks are generally low. As disc-braked vehicles with composite
brake blocks are used in Tehran Railway Network and wheels were
new (in good conditions), rolling noise emissions are governed by
rail corrugation in the measurement site. That is, the wheel corru-
gation and its role in rolling noise are negligible in this research.
Figs. 18 and 19 present the rail corrugation wavelength for the
curve and the tangent track (roughness spectra) obtained from the
analysis of EM50 track recording car data using the new algorithm
developed in this paper. Based on these Figures, there are the sev-
ere short-pitch rail corrugation in the range of 25–100 mm for the
tangent track and the severe rolling contact fatigue corrugation in
the range of 150–450 mm for the curve.
4. Discussions
In order to evaluate the rail corrugation effects on the rolling
noise in the tangent track and the curve, correlations between
the rolling noise measured by the field tests and the vertical pro-
files of the track (obtained from the processing of the EM50 data)
were investigated. To obtain the same numbers of data points for
the rolling noise and the track profile, data points from the rolling
noise were extracted from the rail at 2 mm intervals. Therefore, the
rolling noise data have been conformed to the rail corrugation pro-
file data. MATLAB software was used for the analysis where the rail
corrugation data was defined as an input and the corresponding
response (rolling noise data) as an output.
To compare rail corrugation and rolling noise, rail corrugation
was expressed in terms of decibels using root mean square of the
rail roughness. Fig. 20 presents the coherence plot for correlations
between the measured corrugations (on the running surface of the
tangent track and the curve) and the rolling noise of the same
locations. To conduct a more detailed analysis on the correlation
graph, Fig. 20 was drawn for a 0.5 m length of the tangent track
and the curve. The difference between the growth rates of the rail
corrugation and the rolling noise can be observed in this Figure.
The results obtained indicate that the rolling noise level increases
with short-pitch rail corrugation in tangent track. As illustrated in
Fig. 20, there is a meaningful correlation between short-pitch rail
corrugation and corresponding rolling noise. On the other hand,
the rolling noise on the tangent track is very sensitive to the
short-pitch rail corrugation. However, there is no meaningful
 J. Sadeghi, A. Hasheminezhad / Applied Acoustics 107 (2016) 10–18 15

Fig. 11. Vertical profile of the railhead for the curve and the tangent track using an EM50 track recording car.

Fig. 14. Comparison of original profile and calculated vertical profile for the curve
Fig. 12. Vertical profile of the railhead for the curve using the algorithm. [2].

Fig. 15. Comparison of the original profile and calculated vertical profile for tangent
track [2].

Fig. 13. Vertical profile of the railhead for the tangent track using the algorithm.
correlation between rolling contact fatigue corrugation and corre-
sponding rolling noise. In fact, the rolling noise on the curve is not
influenced by the rolling contact fatigue corrugation. This can be
due to the increased slippage and wear in the wheel–rail interface.
According to Fig. 20, both rail corrugation and rolling noise
graphs have hills and valleys. The hills of the rail corrugation graph
indicate shiny crests and darker troughs which are the causes of
the short-pitch rail corrugation. In these data points, the detected
rail corrugation causes rolling noise, which makes hills on the same
data point and proves that there is a meaningful correlation
between short-pitch rail corrugation and rolling noise.
Furthermore, in the valleys of the rolling noise graph in which roll-
ing noise has lower values, the short-pitch rail corrugation graph
corresponds to periodic hills. This phenomenon is due to the higher
wear resistance of these points. It can be seen that in the valleys of
the short-pitch rail corrugation graph, the rolling noise is not
excited anymore. In fact, the formation of valleys and hills in the
rail corrugation graph is due to the structural change of the rail
16 J. Sadeghi, A. Hasheminezhad / Applied Acoustics 107 (2016) 10–18

Fig. 16. Developed vertical profile for the curve and the tangent track.

Fig. 17. Extracted rolling contact fatigue corrugation and short-pitch rail corrugation signals from the original rail corrugation signal for the curve and the tangent track. (For
interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

30
Rolling Contact Rail Corrugation
20
roughness, dB

10

0
0 100 200 300 400 500 600
-10

-20
corrugation wavelength, mm

Fig. 18. Rolling contact fatigue corrugation wavelength in the curve.

20
Short-pitch Rail Corrugation
10
roughness,dB

0
0 50 100 150 200
-10

-20
corrugation wavelength, mm

Fig. 19. Short-pitch rail corrugation wavelength in the tangent track.

 J. Sadeghi, A. Hasheminezhad / Applied Acoustics 107 (2016) 10–18
Fig. 20. Coherence of rolling noise and rail corrugation for 0.5 m length of the curve and tangent track.
along its axis, knowing that the structural change is usually not
homogeneous along the rail axis. This appears in terms of hills
and valleys in the rail corrugation graph. As indicated in Fig. 20,
consecutive hills for the curve are much more than those of the
tangent track. In fact, the rolling contact fatigue corrugation in
the curve has led to the rolling contact fatigue defect. This is due
to the heavy rail corrugation in the curve. Furthermore, due to
the periodic nature of track loading and heavy rail corrugation in
the inner rail, the rolling contact fatigue corrugation and the rolling
contact fatigue damage have been combined to produce high wear
resistance parts. This has led consecutive peaks on the correspond-
ing graph. It can be concluded that there is no meaningful correla-
tion between the rolling contact fatigue corrugation and the rolling
noise in the curves as well as in the tangent tracks. Short-pitch rail
corrugation is the source of excitations which manifest itself in the
form of noise generated by the system components. As a result,
short wavelength on the surface plays a critical role in the rolling
noise generation. On the other hand, long wavelengths tend to
reduce the excitation and thus the overall noise level. The obtained
results have led to predict the rolling noise and find its counter-
measure. Since the best method of reducing noise is to manage it
at its source, the noise reduction was found to be dependent on
the conditions of rail corrugation. With regard to grinding as a
noise reduction measure, roughness monitoring is required to
maintain the required acoustical qualifications.
5. Conclusions
Nowadays, railway tracks often pass through densely populated
areas, especially in central parts of metropolises. In order to reduce
railway noise effectively, a good knowledge of its source mecha-
nisms is required. Therefore, a thorough investigation of the track
factors which influence the railway vibration and noise is a need.
One of the main factors influencing the railway noise is the surface
roughness of the wheel and rail, and in particular rail corrugations.
The effects of general rail corrugation on the total noise have been
investigated in the available literature. However, the works in the
available literature have two main limitations: they have been
made through only theoretical modeling; and the effects of two
particular corrugations (i.e., short-pitch rail corrugation in the
tangent track and rolling contact corrugation in the curve) on the
rolling noise (as the main source of noise) have not been investi-
gated. Addressing these limitations, the effects of short-pitch rail
corrugation of the tangent track and rolling contact fatigue corru-
gation of the curve on rolling noise were experimentally investi-
gated in this research. For this purpose, railway noise was
measured during the passage of a railway vehicle in a comprehen-
sive field test carried out based on the ISO 3095 standard. An indi-
17
rect method for rail corrugation measurement was developed in
this research. Using this method, rail corrugations with various
wavelengths were derived from processing the data obtained from
a track recording car. It was shown that this method is an effective
method of obtaining rail corrugations and cost effective compared
with the conventional methods. The Butterworth Band-Pass
Filtering was used to distinguish the rolling noise signal from the
total noise and to extract the short-pitch rail corrugation and the
rolling contact corrugation from the original (measured) rail corru-
gation signals.
It was shown that the formation and occurrence of rolling con-
tact corrugation at the curves is much more than short-pitch rail
corrugation in tangent tracks. So, rolling contact corrugation in
curves can led to severe corrugation and damage rather than
short-pitch rail corrugation in tangent tracks. Through analyses
of the rail corrugation in the tangent track and the curve and their
corresponding rolling noise, correlations between rail corrugation
and rolling noise were developed for the first time. The influences
of rail corrugation on the rolling noise were illustrated. According
to the obtained results, the growth rates of the roughness and the
noise are different. It was shown that the high levels of rolling
noise on the curve are not caused by the track rolling contact cor-
rugation. This is due to the increased slippage and wears in the
wheel–rail interface that can occur on curved tracks. The contribu-
tion of short-pitch rail corrugation levels in rolling noise genera-
tion has been shown experimentally. Results indicate that rolling
noise is very sensitive to the short-pitch rail corrugation in the tan-
gent track. However, the effect of rolling contact fatigue corruga-
tion on rolling noise was shown to be negligible for the curve. It
was shown that there is a meaningful correlation between reduc-
tions in the rolling noise level and reductions in the short-pitch rail
corrugation level.
The results obtained in this research provide a better under-
standing of the correlations between rail corrugation and railway
noise. They can be used to validate the existing rail rolling noise
prediction models. They also provide us with guides for the design
of track maintenance strategies with the aim of increasing safety
and reducing railway noise pollution.
Acknowledgments
The study was performed in the Tehran Metro Network. The
technical support of the Tehran Urban and Suburban Railway
Operation Co is most appreciated. The authors acknowledge with
deep appreciation their untiring support, so vital to the completion
of this research. This appreciation is extended to all the enthusias-
tic and hardworking site engineers and technicians who so care-
fully carried out all the site activities.
18 J. Sadeghi, A. Hasheminezhad / Applied Acoustics 107 (2016) 10–18

References [16] ISO, EN. 13231-3: 2006. Railway applications – track – acceptance of works –
Part 3: acceptance of rail grinding, milling and planning work in track; 2006.
[17] Nielsen Jens CO, Ekberg Anders, Lundén Roger. Influence of short-pitch wheel/
[1] Thompson DJ. Railway noise and vibration: mechanisms, modeling and means
rail corrugation on rolling contact fatigue of railway wheels. Proc Inst Mech
of control. Elsevier; 2008.
Engineers, Part F: J Rail Rapid Transit 2005;219(3):177–87.
[2] Sadeghi J, Hasheminezhad A. Field investigation of rail corrugation effect on
[18] Grassie SL, Saxon MJ, Smith JD. Measurement of longitudinal rail irregularities
rolling noise. J Transport Res 2015;12(2):145–55 [in Persian].
and criteria for acceptable grinding. J Sound Vib 1999;227(5):949–64.
[3] Remington Paul J. Wheel/rail noise—Part IV: rolling noise. J Sound Vib 1976;46
[19] Nielsen Jens CO, Lundén Roger, Johansson Anders, Vernersson Tore. Train–
(3):419–36.
track interaction and mechanisms of irregular wear on wheel and rail surfaces.
[4] Remington Paul J. Wheel/rail rolling noise: what do we know? What don’t we
Vehicle Syst Dyn 2003;40(1–3):3–54.
know? Where do we go from here? J Sound Vib 1988;120(2):203–26.
[20] Grassie Stuart L. Rail corrugation: advances in measurement, understanding
[5] Thompson DJ, Hemsworth B, Vincent N. Experimental validation of the TWINS
and treatment. Wear 2005;258(7):1224–34.
prediction program for rolling noise, Part 1: description of the model and
[21] Tassilly E. Rail corrugations: analytical model and field tests. Mechanics and
method. J Sound Vib 1996;193(1):123–35.
fatigue in wheel/rail contact; 2012. p. 163.
[6] Thompson DJ, Janssens MHA. TWINS–track-wheel interaction noise software.
[22] Martínez-Casas José, Di Gialleonardo Egidio, Bruni Stefano, Baeza Luis. A
Theoretical manual, version 2.4. TNO Institute of Applied Physics report TPD-
comprehensive model of the railway wheelset–track interaction in curves. J
HAG-RPT-93-0214, Delft; 1997.
Sound Vib 2014;333(18):4152–69.
[7] Kitagawa T, Thompson DJ. Comparison of wheel/rail noise radiation on
[23] Kaewunruen Sakdirat, Marich Steve. Severity and growth evaluation of rail
Japanese railways using the TWINS model and microphone array
corrugations on sharp curves using wheel/rail interaction. In: Proceedings of
measurements. J Sound Vib 2006;293(3):496–509.
the 20th national convention on civil engineering, vol. 1 (1); 2015. p. 1–10.
[8] Hardy AEJ, Jones RRK, Turner S. The influence of real-world rail head roughness
[24] Bellette PA, Meehan PA, Daniel WJT. Validation of a tangent track corrugation
on railway noise prediction. J Sound Vib 2006;293(3):965–74.
model with a two disk test rig. Wear 2011;271(1):268–77.
[9] Nielsen Jens CO, Ekberg Anders. Acceptance criterion for rail roughness level
[25] Jiang S, Meehan PA, Thompson DJ, Jones CJC. Railway rolling noise prediction:
spectrum based on assessment of rolling contact fatigue and rolling noise.
field validation and sensitivity analysis. Int J Rail Transport 2013;1(1–
Wear 2011;271(1):319–27.
2):109–27.
[10] Nielsen Jens CO. High-frequency vertical wheel–rail contact forces—validation
[26] Jiang Shijie, Meehan Paul A. Sensitivity analysis of simulation model on
of a prediction model by field testing. Wear 2008;265(9):1465–71.
railway rolling noise prediction. Int J Vehicle Noise Vib 2014;10(3):241–56.
[11] Ekberg Anders, Kabo Elena, Nielsen Jens CO, Lundén Roger. Subsurface
[27] Meehan PA, Bellette PA, Batten RD, Daniel WJT, Horwood RJ. A case study of
initiated rolling contact fatigue of railway wheels as generated by rail
wear-type rail corrugation prediction and control using speed variation. J
corrugation. Int J Solids Struct 2007;44(24):7975–87.
Sound Vib 2009;325(1):85–105.
[12] Gullers Per, Andersson Lars, Lundén Roger. High-frequency vertical wheel–rail
[28] Jiang S, Meehan PA, Thompson DJ, Jones CJC. Railway rolling noise prediction
contact forces—field measurements and influence of track irregularities. Wear
under European conditions. Proc Acoustics 2011;2011:2–4.
2008;265(9):1472–8.
[29] Baeza Luis, Vila Paloma, Roda Alejandro, Fayos Juan. Prediction of corrugation
[13] Jiang S, Meehan PA, Bellette PA, Thompson DJ, Jones CJC. Validation of a
in rails using a non-stationary wheel–rail contact model. Wear 2008;265
prediction model for tangent rail roughness and noise growth. Wear 2014;314
(9):1156–62.
(1):261–72.
[30] Li SG, Li ZL, Dollevoet RPBJ. Wear study of short pitch corrugation using an
[14] ISO, EN. 3095: 2005. Railway applications-acoustics-measurement of noise
integrated 3D FE train–track interaction model. In: The 9th international
emitted by rail bound vehicles; 2005.
conference on contact mechanics and wear of rail/wheel systems, Chengdu,
[15] ISO, EN. 15610: 2009. Railway applications – noise emission – rail roughness
China; 2012. p. 216–22.
measurement related to rolling noise generation; 2009.