materials

Article
Experimental Study on Dynamic Characteristics of Damaged
Post-Tensioning Concrete Sleepers Using Impact Hammer
Jung-Youl Choi 1 , Tae-Hyung Shin 2 , Sun-Hee Kim 3, * and Jee-Seung Chung 1

1 Department of Construction Engineering, Dongyang University, No. 145 Dongyangdae-ro, Punggi-eup,

Yeongju-si 36040, Gyeongsangbuk-do, Republic of Korea; [email protected] (J.-Y.C.);
[email protected] (J.-S.C.)
2 Engineering HQ, Seoul Metro, 5, Hyoryeong-ro, Seocho-gu, Seoul 06693, Republic of Korea;
[email protected]
3 Department of Architectural Engineering, Gachon University, 1342 Seongnamdaero, Sujeong-gu,
Seongnam-si 13120, Gyeonggi-do, Republic of Korea
* Correspondence: [email protected]; Tel.: +82-31-750-4718

Abstract: Concrete sleepers in operation are commonly damaged by various internal and external
factors, such as poor materials, manufacturing defects, poor construction, environmental factors, and
repeated loads and driving characteristics of trains; these factors affect the vibration response, mode
shape, and natural frequency of damaged concrete sleepers. However, current standards in South
Korea require only a subjective visual inspection of concrete sleepers to determine the damage degree
and necessity of repair or replacement. In this study, an impact hammer test was performed on
concrete sleepers installed on the operating lines of urban railroads to assess the field applicability of
the modal test method, with the results indicating that the natural frequency due to concrete sleeper
damage was lower than that of the undamaged state. Furthermore, the discrepancy between the
simulated and measured natural frequencies of the undamaged concrete sleeper was approximately
1.87%, validating the numerical analysis result. The natural frequency of the damaged concrete
sleepers was lower than that of the undamaged concrete sleeper, and cracks in both the concrete
sleeper core and the rail seat had the lowest natural frequency among all the damage categories.
Therefore, the damage degrees of concrete sleepers can be quantitatively estimated using measured
natural-frequency values.
Citation: Choi, J.-Y.; Shin, T.-H.; Kim,
S.-H.; Chung, J.-S. Experimental Study
Keywords: concrete sleepers; impact hammer test; natural frequency; cracks; numerical analysis
on Dynamic Characteristics of
Damaged Post-Tensioning Concrete
Sleepers Using Impact Hammer.
Materials 2024, 17, 1581. https://
doi.org/10.3390/ma17071581 1. Introduction
Concrete sleepers on railroad tracks can be damaged because of repeated loads and
Academic Editor: Lizhi Sun
driving characteristics of trains, resulting in the cracking and dislodging of tie bar contacts
Received: 31 January 2024 in the interior of the sleeper, tie bar breakage, breaking around fasteners, and fissures in
Revised: 11 March 2024 the concrete underneath the rail. Damage to concrete structures can affect their material
Accepted: 27 March 2024 and structural properties, such as dynamic responsiveness, natural frequency, mode forms,
Published: 29 March 2024 and damping ratio. Nonetheless, according to South Korean standards [1], the condition
of damaged sleepers and the necessity of repair or replacement are determined by a sub-
jective visual inspection of the damaged sleeper. Additionally, damaged concrete sleepers
Copyright: © 2024 by the authors.
must be removed from the railway, making the entire inspection process time-consuming
Licensee MDPI, Basel, Switzerland.
and expensive.
This article is an open access article
In South Korea, sleeper inspection is performed using two methods: visual inspec-
distributed under the terms and tions and rebound hardness tests. In visual inspections, the sleeper’s damage is judged
conditions of the Creative Commons subjectively by the inspector via a qualitative appraisal. According to Clark et al. [2], the
Attribution (CC BY) license (https:// increasing demand on track structures creates a need for an appropriate monitoring system.
creativecommons.org/licenses/by/ Visual observation methods do not record damage in real time and do not have sufficient
4.0/). efficiency to meet the significant demand for reduced track possessions. By applying

Materials 2024, 17, 1581. https://doi.org/10.3390/ma17071581 https://www.mdpi.com/journal/materials

Materials 2024, 17, 1581 2 of 15

non-destructive technologies (NDTs), the risk due to structural damage to the track can
be reduced. As a result of monitoring the structural condition of the damaged structure
through acoustic emission, it was confirmed that it was effective in detecting early cracks.
The rebound hardness test examines the material strength of the concrete but does
not assess the structural integrity of damaged concrete sleepers. Due to the limitations
of these two methods, an improved method is required to determine the condition of
concrete sleepers.
To better understand the structural behavior of damaged concrete sleepers, Choi [3]
conducted experimental and computational assessments of the changes in the dynamic
characteristics of concrete sleepers via an impact hammer test. Kim and Jung [4] performed
an impact hammer test to assess the structural integrity of old railroad plate girder bridges.
These researchers also performed a mode analysis of the full-scale field experiments to
verify their reliability. Esmaeili et al. [5] tested the dynamic resistance of concrete, wood,
and steel sleepers under lateral impact loading conditions in the laboratory, with steel
sleepers found to have insufficient dynamic resistance to impact loads compared with
concrete and wooden sleepers.
Braunfelds et al. [6] proposed a technology that can measure read deformation over a
long period of time using an optical FBG sensor. Chung et al. conducted field applicability
evaluation using the rebound hardness test method to estimate the strength of sleeper float-
ing track concrete sleepers. As a result, the strength estimation results showed differences
depending on the sleeper’s condition and support conditions [7].
Choi et al. [8] estimated the spring stiffness of the sleeper floating track sleeper re-
silience pads using finite element analysis. The vibration acceleration of the ballast in the
numerical analysis was within the range of field measurement results.
Sapidis et al. [9] proposed autonomous compression damage of fiber-reinforced con-
crete using a piezoelectric lead zirconate titanate transducer using a 1D convolutional
neural network. As a result, it showed an accuracy of 95.25% with the experimental results.
Nielsen and Palmer [10] discovered that the modal analysis of prestressed concrete
sleepers was useful for identifying dynamic behavior and reaction, with the modal analy-
sis results represented as frequency response functions (FRFs) for various sleeper states.
Through finite element analysis, Kaewunruen and Remennikov [11] analyzed the sensi-
tivity effect on the free-vibration characteristics of concrete sleepers in the field caused
by the changes in the stiffness and material properties of the rail pads using Timoshenko
beam and spring elements. The finite element analysis results confirmed that the rail pad
parameters have a nonlinear effect on the effective stiffness, influencing the field track
system and significantly affecting the frequency and mode shape of additional modes.
Kaewunruen and Remennikov [12] performed an impact hammer test, a nondestructive
test method, to evaluate the structural integrity of a gravel road track, and they compared
the time–acceleration response function from field measurements with the visual inspec-
tion results. By performing impact hammer experiments on recovered damaged concrete
sleepers, Shin [13] proved that damage to sleepers can reduce the stiffness of sleepers but
has minimal effects on their mass-and-damping ratio. Furthermore, it was proven experi-
mentally and analytically via comparison with numerical analysis results that the natural
frequency of damaged sleepers is lower than that of undamaged sleepers, depending on
the extent of the damage. Finally, the vibration of the sleeper was examined by Lam and
Wong [14] through an impact hammer test to determine whether the sleeper was damaged.
Kaewunruen and Remennikov [15] used modal analysis to evaluate the modal change in
vibration characteristics of prestressed concrete sleepers in the 0–1600 Hz frequency band.
The experiment evaluated the impacted specimen with a drop impact tester. The modal
parameters of undamaged sleepers and cracked sleepers were compared.
You et al. summarized the damage patterns of prestressed concrete sleepers to provide
better insight into the damage effects of railway-reinforced concrete sleepers so that they
could be used to improve track maintenance and inspection criteria. They analyzed the
theory of the causes of prestressed concrete sleepers based on the limit state method [16].
Materials 2024, 17, 1581 3 of 15

Real et al. investigated the dynamic effect of cracks on the vibration characteristics
of railway-prestressed concrete sleepers to identify structural defects. They highlighted
changes in the modal parameters of healthy and cracked sleepers in terms of natural
frequencies and modal damping [17]. Matsuoka and Watanabe applied a frequency-based
detection method to evaluate damaged concrete sleepers. As the frequency-based damage-
detection method, which has limited practical application, was used, damage to concrete
was confirmed through experiments to validate its feasibility [18].
According to the results of previous studies, we propose the impact hammer test, a
modal test technique, as a quantitative method for evaluating the structural soundness
of concrete sleepers. The impact hammer test was employed to compare field data with
analytical results. A post-tensioned sleeper with rail support points buried in the concrete
road surface at both ends of the sleeper was used, and the sleeper floating track was elasti-
cized by inserting resilience pads at the corresponding positions. The dynamic response
changes of concrete sleepers induced by the post-tensioning method were analyzed via
field measurements and finite element analysis to compare the damage to concrete sleep-
ers and the dynamic characteristic changes using the impact hammer test. Furthermore,
field measurements were performed on a public urban railroad’s sleeper floating track
to examine the effect of the damage degree of concrete sleepers on the natural frequency
change. The impact hammer test is a portable, easy-to-apply method for maintaining
concrete sleepers.

2. Materials and Methods

2.1. Measurement Sections and System
Field measurements were performed on concrete sleepers of antivibration track struc-
tures built on urban railroad lines in Korea. The antivibration track structure was used
on steeply curved concrete sleepers with a curve radius (R) of less than 400 m and is one
of the sleeper floating orbits with a resilience pad placed on the bottom. The measuring
track comprised a concrete track in a tunnel with a curve radius of 250 m (cant: 110 mm), a
vibration-reducing track structure, and a train speed of 50 km/h.
Impact hammer tests were performed on one undamaged concrete sleeper and three
damaged concrete sleepers with cracks in the sleeper center and rail seat based on visual
inspection. The impact hammer test is a method of evaluating the dynamic properties of
concrete through the impact of the hammer and the response signal measured with an
accelerometer. To check the stiffness of concrete, the influence of dynamic characteristics
must be checked. Acceleration gauges are installed inside and outside of track gauge. For
the damaged concrete sleepers, the damage was concentrated on the rail seat and sleeper
center, as shown in Figure 1. Figure 1a,b,d,f shows the damage caused to the sleeper directly
below the rail. Figure 1e shows damage to the sleeper directly below the rail; here, damage
occurred in the center of the sleeper.
The field measurement section is shown in Figure 2a, and accelerometers were placed
on the top surface of the sleeper to measure the vertical acceleration of the sleeper during
the impact hammer blow, as shown in Figure 2b,c. Table 1 shows the sensitivity of each
sensor of the impact hammer and accelerometer employed in the measuring method. The
dynamic response (acceleration) measurement data against the load created by the impact
hammer blow were used to study the frequency response function (FRF). The FRF allows
us to calculate the characteristics of dynamic responses in the frequency domain and the
relationship between mass and stiffness, which is expressed as the ratio of the dynamic
response to the impact load.
Materials 2024, 17, x FOR PEER REVIEW 4 of 15
accelerometer. To check the stiffness of concrete, the influence of dynamic characteristics
must be checked. Acceleration gauges are installed inside and outside of track gauge. For
the damaged concrete sleepers, the damage was concentrated on the rail seat and sleeper
center, as shown in Figure 1. Figure 1a,b,d,f shows the damage caused to the sleeper di-
Materials 2024, 17, 1581 4 of 15
rectly below the rail. Figure 1e shows damage to the sleeper directly below the rail; here,
damage occurred in the center of the sleeper.

(c) (d)

Materials 2024, 17, x FOR PEER REVIEW 4 of 15

(a) (b)

(e) (f)
Figure 1. Visual inspection of concrete sleepers. (a) Crack #1 (rail seat); (b) Crack #1 (center); (c)
Crack #2 (rail seat); (d) Crack #2 (rail seat); (e) Crack #3 (rail seat, center); (f) Crack #3 (rail seat).

(c)
The field measurement (d)
section is shown in Figure 2a, and accelerometers were placed
on the top surface of the sleeper to measure the vertical acceleration of the sleeper during
the impact hammer blow, as shown in Figure 2b,c. Table 1 shows the sensitivity of each
sensor of the impact hammer and accelerometer employed in the measuring method. The
dynamic response (acceleration) measurement data against the load created by the impact
hammer blow were used to study the frequency response function (FRF). The FRF allows
us to calculate the characteristics of dynamic responses in the frequency domain and the
relationship between mass and stiffness, which is expressed as the ratio of the dynamic
response to the impact load.

Table 1. Sensitivity of sensors for impact hammer test.

Sensor Type Impact Hammer

(e) Accelerometer
(f)
Channels Ch.1 Ch.2 Ch.3 Ch.4
Figure1.1.Visual
Visual inspection
of of concrete sleepers. (a) Crack #1 (rail seat); (b) Crack #1 (center); (c)
Sensitivityinspection 0.2301
Figure concrete
mV/Nsleepers. (a)1024.0
Crack #1 (rail
mV/g seat); (b) Crack
1018.0 mV/g#1 (center);
993.2(c) Crack
mV/g
Crack
#2 (rail #2 (rail(d)
seat); seat); (d)#2
Crack Crack
(rail #2 (rail(e)
seat); seat);
Crack(e)#3
Crack
(rail #3 (rail
seat, seat, center);
center); (f)#3
(f) Crack Crack
(rail #3 (rail seat).
seat).

The field measurement section is shown in Figure 2a, and accelerometers were placed
on the top surface of the sleeper to measure the vertical acceleration of the sleeper during
the impact hammer blow, as shown in Figure 2b,c. Table 1 shows the sensitivity of each
sensor of the impact hammer and accelerometer employed in the measuring method. The
dynamic response (acceleration) measurement data against the load created by the impact
hammer blow were used to study the frequency response function (FRF). The FRF allows
us to calculate the characteristics of dynamic responses in the frequency domain and the
relationship between mass and stiffness, which is expressed as the ratio of the dynamic
response to the impact load.

Table 1. Sensitivity of sensors for impact hammer test.

(a) (b)
Sensor Type Impact Hammer Accelerometer
2. Cont.
FigureChannels Ch.1 Ch.2 Ch.3 Ch.4
Sensitivity 0.2301 mV/N 1024.0 mV/g 1018.0 mV/g 993.2 mV/g
 Materials 2024, 17, x FOR PEER REVIEW 5 of 15
Materials 2024, 17, 1581 5 of 15

Materials 2024, 17, x FOR PEER REVIEW 5 of 15

(c)
Figure 2. Photographs of modal test using impact hammer. (a) Overview of test track; (b) impact
Figure 2. Photographs of modal test using impact hammer. (a) Overview of test track; (b) impact
hammer test; (c) schematic of instrumentation.
hammer test; (c) schematic of instrumentation.
2.2. Dynamic Mass-and-Stiffness Estimation Using FRF
(c)impact hammer test.
Table 1. Sensitivity of sensors for
The relationship between the loads acting on a structure and its dynamic reaction is
Figure 2. for
critical
Sensor Photographs of modal
determining
Type Impact test using
itsHammer
dynamic impactThis
features. hammer. (a)
reaction Overview
is knownofastest
Accelerometer an track;
FRF in(b)the
impact
fre-
hammer test; (c) schematic of instrumentation.
quency domain. The FRFCh.1
Channels can be used to estimate
Ch.2 values aboveCh.3the natural frequency
Ch.4 for
dynamic masses and0.2301
Sensitivity belowmV/N
the natural1024.0
frequency
mV/gfor dynamic stiffness. 993.2 mV/g
1018.0 mV/g
2.2. Dynamic Mass-and-Stiffness Estimation Using FRF
The vibration response of a structure can be used to generate several forms of FRFs.
In Thestudy,
this relationship between
the standard FRFthe
was loads
usedacting
to on a structure
determine and its dynamic reaction is
2.2. Dynamic Mass-and-Stiffness Estimation Using FRF the structure’s dynamic mass.
criticalThe forrelationship
determiningbetween
its dynamic features.load
the applied Thisand
reaction is known
dynamic response as an
in FRF in the fre-
the frequency
quency The relationship
domain. The between
FRF can bethe
usedloads
to acting onvalues
estimate a structure
above and
the its dynamic
natural reaction
frequency for
domain is summarized in Table 2 [13]. In Table 2, 𝐹 is the load and u is the displace-
is critical
dynamic for determining its dynamic features. This reaction is known as an FRF in the
ment. masses
frequency
and below the natural frequency for dynamic stiffness.
domain. The FRF can be used to estimate values above the natural frequency for
The vibration response of a structure can be used to generate several forms of FRFs.
dynamic masses and below the natural frequency for dynamic stiffness.
InTable
this 2.
study, the standard
Load-response FRF was
relationship used
in the to determine
vibration domain.the structure’s dynamic mass.
The vibration response of a structure can be used to generate several forms of FRFs. In
The relationship between the applied load and dynamic response in the frequency
this study,
Responsethe standard FRF was used to
Inverse FRFdetermine the structure’s dynamicFRF
Standard mass.
domain The isrelationship
summarizedbetween
in Tablethe 2 [13]. In Table
applied load 2, 𝐹dynamic
and is the load and u in
response is the
the frequency
displace-
Acceleration 𝐹(𝜔)/𝑢̈ (𝜔) 𝑢̈ (𝜔)/𝐹(𝜔)
ment.
domain is summarized in Table𝐹(𝜔)/𝑢̇ In Table 2, F is the load and 𝑢̇u(𝜔)/𝐹(𝜔)
2 [13]. (𝜔) is the displacement.
Velocity
Table 2. Load-response relationship𝐹(𝜔)/u(𝜔)
Displacement in the vibration domain.
Table 2. Load-response relationship in the vibration domain.
u(𝜔)/𝐹(𝜔)

Response
Through
Response Inverse
FRF, the dynamic massFRFwas estimated
Inverse FRF Standard
in the range exceedingFRF
Standard the natural
FRF
Acceleration
frequency, as shown in Figure 𝐹(𝜔)/𝑢̈
3a. (𝜔) stiffness
Dynamic .. was estimated𝑢̈ below
(𝜔)/𝐹(𝜔)
.. the natural fre-
Acceleration F (ω )/u(ω ) u(ω )/F (ω )
quency, shown in Figure 3b.𝐹(𝜔)/𝑢̇ (𝜔)
as Velocity
Velocity . .
𝑢̇ (𝜔)/𝐹(𝜔)
F (ω )/u(ω ) u(ω )/F (ω )
Displacement
Displacement 𝐹(𝜔)/u(𝜔) F (ω )/u(ω) u(𝜔)/𝐹(𝜔)
u(ω)/F (ω )

Through
ThroughFRF,
FRF, the
the dynamic
dynamic mass
mass was estimated in
was estimated in the
the range
range exceeding
exceeding the
thenatural
natural
frequency, as shown in Figure 3a. Dynamic stiffness was estimated below the natural
frequency, as shown in Figure 3a. Dynamic stiffness was estimated below the natural fre-
quency, as shown in Figure 3b.
frequency, as shown in Figure 3b.

(a) (b)
Figure 3. Dynamic mass-and-stiffness estimation of single degree of freedom. (a) Dynamic mass
(m); (b) dynamic stiffness.

Using FRF, the dynamic mass-and dynamic stiffness of a structure can be estimated,
and
(a) the relationship is summarized in Table 3 [13]. (b)
Figure 3. Dynamic mass-and-stiffness estimation of single degree of freedom. (a) Dynamic mass
Figure 3. Dynamic mass-and-stiffness estimation of single degree of freedom. (a) Dynamic mass (m);
(m); (b) dynamic stiffness.
(b) dynamic stiffness.

Using FRF, the dynamic mass-and dynamic stiffness of a structure can be estimated,
and the relationship is summarized in Table 3 [13].
Materials 2024, 17, x FOR PEER REVIEW 6 of 15

Materials 2024, 17, 1581 6 of 15

Table 3. Relationship between FRF and dynamic mass and dynamic stiffness.
Using FRF, the dynamic Dynamic
𝒍𝒐𝒈|𝑭𝑹𝑭| mass-and dynamic
Mass (m)stiffness of a structure
Dynamiccan be estimated,
Stiffness (k)
and the relationship is summarized
1 in Table 3 [13]. 2
𝜔
log|𝑢̈ (𝜔)/𝐹(𝜔)| − log(𝑚) − 2log(𝜔) − log(𝑘)
𝑚 𝑘
Table 3. Relationship between FRF and dynamic mass and dynamic stiffness.
1 𝜔
log|𝑢̇ (𝜔)/𝐹(𝜔)| −𝑖 − log(𝑚) − log(𝜔) 𝑖 − log(𝜔) − log(𝑘)
log|FRF| 𝜔𝑚 Dynamic Mass (m) 𝑘
Dynamic Stiffness (k)
.. 1 1
log|𝑢(𝜔)/𝐹(𝜔)|
log u(ω )/F (ω ) − 2 − log(𝑚) 1 − 2log(𝜔)
m − log( m )
2
−) log(𝑘)
k − 2 log − log(k)
ω
(ω
. 𝜔 𝑚 1 𝑘
log u(ω )/F (ω ) −i ωm − log(m) − log(ω ) ω
i k − log(ω ) − log(k)
log|u(ω )/F (ω )| − ω12 m − log(m) − 2 log(ω ) 1
k − log( k )
2.3. Numerical Analysis
2.3.1.
2.3. Numerical
Numerical Analysis Conditions
Analysis
2.3.1. The rail-connection
Numerical mechanism and spring element of the sleeper floating track were
Analysis Conditions
usedThe
in rail-connection
the same manner as in the
mechanism andfield.
springThe internal
element of thesteel barsfloating
sleeper of thetrack
post-tensioning
were
sleeper
used were
in the samesimulated,
manner asand tension
in the field.forces for each
The internal steel
steel barsbar
of were added, as shown in
the post-tensioning
Figure were
sleeper 4. Self-weight
simulated,and
andexternal-impact
tension forces forloads
each were considered
steel bar were added,in the numerical
as shown in anal-
Figure 4. Self-weight and external-impact loads were considered in the numerical
yses. The impact load was 700 N, as determined via the impact hammer test. Rail pads (k1, analyses.
The impact load
50 kN/mm) was 700
inserted intoN,
theasrail
determined
seat, underviasleeper
the impact
padshammer test. Railinserted
(k2, 10 kN/mm) pads (k1into
, the
50 kN/mm) inserted into the rail seat, under sleeper pads (k , 10 kN/mm)
lower section of the sleeper, and rubber boots (k3, 2000 kN/mm) were used as spring ele-
2 inserted into
the lower section of the sleeper, and rubber boots (k3 , 2000 kN/mm) were used as spring
ments. The stiffness of the rail pad, under sleeper pad, and rubber boots used the values
elements. The stiffness of the rail pad, under sleeper pad, and rubber boots used the values
presented in Seoul Metro’s internal specifications.
presented in Seoul Metro’s internal specifications.

Figure4.4.Concrete
Figure Concrete sleeper
sleeper spring
spring model.
model.

2.3.2.
2.3.2.Modeling
Modeling
Figure
Figure5 5shows
shows a numerical
a numerical model corresponding
model correspondingto thetofield-measured concrete
the field-measured concrete
sleeper in Section 2.1. A numerical analysis was performed using Ansys Workbench Ver.
sleeper in Section 2.1. A numerical analysis was performed using Ansys Workbench Ver.
2021 R2 [19] to analyze the natural frequency of the undamaged concrete sleepers. For
2021 R2 [19] to analyze the natural frequency of the undamaged concrete sleepers. For the
the concrete sleepers, modal analysis was performed on a total of six modes supported by
concrete
the Ansys sleepers,
program. modal
Modal analysis was
analysis is performed
a method on a total
of analyzing theofdynamic
six modes supported by the
characteristics
of a structure or the characteristics of a structure subjected to dynamic external forces. of a
Ansys program. Modal analysis is a method of analyzing the dynamic characteristics
structurecharacteristics
Dynamic or the characteristics of a structure
include resonance frequencysubjected
and modeto dynamic external Each
shape attenuation. forces. Dy-
namicrepresents
mode characteristics includephenomenon
a free vibration resonance frequency
at the modeand mode which
frequency, shapeisattenuation.
influenced Each
by the weight
mode and rigidity
represents of the structure.
a free vibration phenomenon at the mode frequency, which is influenced
by the weight and rigidity of the structure.
In general, simple structures can be said to have the first mode shape in the most
flexible direction. In the case of a cantilever with one side fixed, the first mode shape can
be predicted because the largest deformation is at the opposite end.
Materials 2024, 17, x FOR PEER REVIEW 7 of 15
Materials 2024, 17, 1581 7 of 15

Figure5.5.Post-tensioning
Figure Post-tensioning concrete
concrete sleeper
sleeper modeling.
modeling.

In general,
The nodessimple structures
and elements cancalculated
were be said to automatically
have the first mode
during shape
theiringeneration
the most in the
flexible
ANSYSdirection. In the program.
Workbench case of a cantilever
The mesh withof
onethe
sidesleeper
fixed, the first mode248,777
contained shape cannodes
be and
predicted because the largest deformation is at the opposite end.
148,707 elements, with the element size of the sleeper assumed to be 20 mm. The size of
The nodes and elements were calculated automatically during their generation in
the steel bar element was 10 mm and that of the rail element was 15 mm. Table 4 presents
the ANSYS Workbench program. The mesh of the sleeper contained 248,777 nodes and
the specifications
148,707 of the
elements, with the element
analysissize
model.
of the sleeper assumed to be 20 mm. The size of the
steel bar element was 10 mm and that of the rail element was 15 mm. Table 4 presents the
Table 4. Properties
specifications of theof FE model.
analysis model.
Items Properties
Table 4. Properties of FE model.
Rail weight (kg/m) 60
Items type
Sleeper Properties
Post-tensioning Sleeper
Sleeper
Rail weightsize (mm)
(kg/m) 212
60 × 240 × 2300
Sleeper type
Compressive strength (MPa) Post-tensioning Sleeper
59
Sleeper size (mm) 212 × 240 × 2300
Compressive strength (MPa) Tensile strength
59 Diameter Length
Post-tensioning steel bar (MPa) Diameter(mm)
Tensile strength Length (mm)
Post-tensioning steel bar (MPa)1230 (mm) Ø9.2 (mm) 2240
1230 Ø9.2 2240
Post-tensioning force (kN/EA)
Post-tensioning force (kN/EA) 66.1 66.1
Spring stiffness
Spring stiffness of of
railrail
padpad (kN/mm)
(kN/mm) 50 50
Spring stiffness of under sleeper pad (kN/mm)
Spring stiffness of under sleeper pad (kN/mm) 10 10
Spring stiffness of rubber boot (kN/mm) 2000
Spring stiffness of rubber boot (kN/mm) 2000
3. Results and Discussion
3. Results and Discussion
3.1. FRF Results
3.1. FRF
TimeResults
acceleration was measured using an impact hammer test on damaged concrete
sleepers, as shown
Time in Figure
acceleration 6. The
was y-coordinate
measured usinginanFigure 7 is hammer
impact the magnitude
test onof damaged
acceleration.
concrete
Figure 7 shows
sleepers, the FRF
as shown of each 6.
in Figure concrete sleeper computed
The y-coordinate from7the
in Figure impact
is the hammerof
magnitude test.
accelera-
The
tion.FRF was used
Figure to compute
7 shows the FRFthe
of measured natural
each concrete frequency
sleeper for eachfrom
computed sleeper.
the impact hammer
The FRF was used to measure the natural frequencies of
test. The FRF was used to compute the measured natural frequency the concrete sleepers,
for eachwith the
sleeper.
natural frequency of the undamaged concrete sleeper found to be approximately 122.96 Hz,
as shown in Figure 7a. The natural frequencies of Crack #1 (91.09 Hz), Crack #2 (85.47 Hz),
and Crack #3 (53.59 Hz) in the three damaged sleepers, which had similar levels of damage,
were measured as shown in Figure 7b–d, experimentally proving that the damage degree of
concrete sleepers that are difficult to assess visually can be quantitatively examined using
the natural frequency.

(a) (b)
 3. Results and Discussion
3.1. FRF Results
Time acceleration was measured using an impact hammer test on damaged concrete
sleepers, as shown in Figure 6. The y-coordinate in Figure 7 is the magnitude of accelera-
Materials 2024, 17, 1581 8 of 15
tion. Figure 7 shows the FRF of each concrete sleeper computed from the impact hammer
test. The FRF was used to compute the measured natural frequency for each sleeper.

Materials
Materials2024,
2024,17,
17,x xFOR
FORPEER
PEERREVIEW
REVIEW 8 8ofof1515

(a) (b)

(c)
(c) (d)
(d)
Figure
Figure6.6.
Figure 6.Measurement
Measurement
Measurement results ofofdynamic
results
results ofdynamic response
dynamic response
responsefunction
function(FRF)
(FRF)
function for
foreach
(FRF) for sleeper.
each sleeper.
each (a)
(a)Time
sleeper. (a)accel-
Time accel-
Time
eration (Good);
eration (b)
(Good); time
(b) timeacceleration
acceleration (Crack #1);
(Crack (c)
#1); time
(c) acceleration
time acceleration(Crack
(Crack#2); (d)
#2); time
(d) timeacceleration
acceleration
acceleration (Good); (b) time acceleration (Crack #1); (c) time acceleration (Crack #2); (d) time
(Crack
(Crack#3).
#3). (Crack #3).
acceleration

(a)
(a) (b)
(b)

(c)
(c) (d)
(d)
Figure
Figure7.7.Measurement
Measurementresults
resultsofoffrequency
frequencyresponse
responsefunction
function(FRF)
(FRF)for
foreach
eachsleeper.
sleeper.(a)
(a)FRF
FRF(good);
(good);
Figure
(b) FRF 7. Measurement
(Crack #1); (c) results
FRF (Crackof frequency
#2); (d) response
FRF (Crack function
#3). (FRF) for each sleeper. (a) FRF (good);
(b) FRF (Crack #1); (c) FRF (Crack #2); (d) FRF (Crack #3).
(b) FRF (Crack #1); (c) FRF (Crack #2); (d) FRF (Crack #3).
The
TheFRF
FRFwas wasused
usedtotomeasure
measurethethenatural
naturalfrequencies
frequenciesofofthe
theconcrete
concretesleepers,
sleepers,with
with
the natural frequency of the undamaged concrete sleeper found to be
the natural frequency of the undamaged concrete sleeper found to be approximately approximately
122.96
122.96Hz,Hz,asasshown
shownininFigure
Figure7a.
7a.The
Thenatural
naturalfrequencies
frequenciesofofCrack
Crack#1#1(91.09
(91.09Hz),
Hz),Crack
Crack#2#2
(85.47
(85.47 Hz), and Crack #3 (53.59 Hz) in the three damaged sleepers, which had similarlev-
Hz), and Crack #3 (53.59 Hz) in the three damaged sleepers, which had similar lev-
els
elsofofdamage,
damage,wereweremeasured
measuredasasshown
shownininFigure
Figure7b–d,
7b–d,experimentally
experimentallyproving
provingthat
thatthe
the
Materials 2024, 17, 1581 9 of 15

Among the damage types, Crack #2 (a crack in the rail seat only) and Crack #1 (a
crack in the sleeper center that did not penetrate the ground) exhibited similar natural
Materials 2024, 17, x FOR PEER REVIEW
frequencies. Crack #3 had the lowest natural frequency because it penetrated the bottom 9 of 15
of the concrete sleeper core and because there was a partial loss of concrete in the crack
in the rail seat, suggesting that the damage degree of concrete sleepers can be estimated
according
aloneto was
the natural frequency
insufficient measured
to assess in the field.
the structural Furthermore,
integrity visual sleepers.
of the concrete inspection In addi-
alone was
tion,insufficient to assess
the conditions the structural
of nighttime integrity
inspections onofoperating
the concrete
linessleepers.
made it In addition,
difficult to secure
the conditions
reliable of nighttime results
assessment inspections onsleeper
for the operating lines made
conditions it difficult
through to inspection.
visual secure reliable
assessment results for the sleeper conditions through visual inspection.
3.2. Dynamic Mass and Stiffness Prediction Using FRF
3.2. Dynamic Mass and Stiffness Prediction Using FRF
3.2.1. Dynamic Mass-Prediction Results
3.2.1. Dynamic Mass-Prediction Results
According to Esmaeill et al. [5], the impact load intensity has a significant impact on
According to Esmaeill et al. [5], the impact load intensity has a significant impact on
the dynamic resistance. Therefore, the FRF was analyzed based on the impact hammer
the dynamic resistance. Therefore, the FRF was analyzed based on the impact hammer test
measuredtest in
measured
the field in
to the field to understand
understand the effect ofthe effect
state of state
changes of achanges
concreteof a concrete
sleeper on itssleeper
on its dynamic mass, and the dynamic mass of the
dynamic mass, and the dynamic mass of the structure was predicted. structure was predicted.
Figure Figure 8a–dthe
8a–d show show the dynamic
dynamic mass ofmass of the concrete
the concrete sleeperssleepers as a function
as a function of state.of state.

(a)

(b)

(c)

Figure 8. Cont.
 Materials
Materials 2024,
2024, 17, 17, x FOR PEER REVIEW
1581 10 of 15
10 of 15

(d)
FigureFigure 8. Dynamic
8. Dynamic mass mass prediction
prediction results.
results. (a) Undamaged
(a) Undamaged concrete
concrete sleeper;
sleeper; (b) Crack
(b) Crack #1; (c)#1; (c) Crack
Crack
#2; (d) Crack
#2; (d) Crack #3. #3.

As shown
As shown in Figurein Figure
8a, when8a, when the concrete
the concrete sleepersleeper
is good,is good, the dynamic
the dynamic massmassap- ap-
proaches 1. It shows a tendency to converge to 1 × 10 N/(m/s ).
proaches 1. It shows a tendency to converge to 1 3× 10 3 N/(m/s 2 2).
In theIncase
the ofcaseCrackof Crack
#1, the #1,side
the section
side section
of theofsleeper
the sleeperwhere where the accelerometer
the accelerometer was was
installed was completely
installed was completely split split
in theinvertical direction,
the vertical so thesoacceleration
direction, the acceleration response was was
response
measured
measuredseparately
separately fromfromthe parent bodybody
the parent in terms of mass.
in terms Figure
of mass. 8b shows
Figure 8b showsthat in thein the
that
case of
caseCrack #1, where
of Crack a crack
#1, where existsexists
a crack in theincenter of theofsleeper,
the center the sleeper,the initial dynamic
the initial massmass
dynamic
measured
measuredby the bynatural frequency
the natural frequencyis upistoup 1× 1 ×4 10
to 10 N/(m/s
4 N/(m/s 2 ).2 It rapidly decreases up to
). It rapidly decreases up to the
the measured
measured natural frequency of 91.09 Hz and shows a tendencytotoconverge
natural frequency of 91.09 Hz and shows a tendency convergebetweenbetween 1 ×
1 × 10 2 2 3
and11××1010N/(m/s 2
10 N/(m/s
2 N/(m/s ) )and
2 3 N/(m/s 2 ). ).
In theIncase
theofcase
Crack of #3,
Crackthe #3,
condition was thatwas
the condition the crack
that the couldcrackbe repaired
could beand attached,
repaired and at-
but ittached,
would but not itcontribute
would not tocontribute
rigidity. Figure 8c shows
to rigidity. Figurethat8cwhenshows a crack occursa directly
that when crack occurs
below the railbelow
directly (Crack the#2), the
rail initial#2),
(Crack dynamic massdynamic
the initial measured mass at the naturalat
measured frequency
the natural is fre-
1 × 10 5 N/(m/s2 ). It5 decreases up to the measured natural frequency of 85.47 Hz and
quency is 1 × 10 N/(m/s ). It decreases up to the measured natural frequency of 85.47 Hz
2

shows anda tendency to converge

shows a tendency between 1between
to converge × 103 N/(m/s 2 ) and 1 × 104 N/(m/s2 ).
1 × 103 N/(m/s 2) and 1 × 104 N/(m/s2).

Figure 8d shows the case of Crack #3, where

Figure 8d shows the case of Crack #3, where a crack penetrates a crack penetrates totothe
thebottom
bottominin the
the center of the concrete sleeper and a portion of the concrete
center of the concrete sleeper and a portion of the concrete cross-section is lost along cross-section is lost along with
with the
the crack
crackdirectly
directlybelow belowthe therail.
rail.TheThe initial
initial dynamic
dynamic mass mass measured
measured at theatnatural
the natural
frequency
frequency is up to 1 × 10 5 N/(m/s2 ). It decreases rapidly up to the measured natural
is up to 1 × 10 N/(m/s ). It decreases rapidly up to the measured natural 2frequency2of 53.59 Hz
5 2
frequency of 53.59
and shows Hz and to
a tendency shows
convergea tendency
between to1converge
× 102 N/(m/s between
2) and 1 1× ×10103 N/(m/s
N/(m/s 2). ) and
1 × 103 N/(m/s2 ).
3.2.2. Dynamic Stiffness Prediction Results for Concrete Sleepers
3.2.2. Dynamic Stiffness Prediction Results for Concrete Sleepers
The dynamic stiffness of the concrete sleepers was determined by evaluating the FRF
The dynamic stiffness of the concrete sleepers was determined by evaluating the FRF
from the impact hammer test measured in the field to understand the effect of state
from the impact hammer test measured in the field to understand the effect of state changes
changes of a concrete sleeper on its dynamic stiffness. Figure 9a–d shows the dynamic
of a concrete sleeper on its dynamic stiffness. Figure 9a–d shows the dynamic stiffness of
stiffnesssleepers.
the concrete of the concrete sleepers.
A sleeper with a crack in the center that did not penetrate the ground (Crack #1),
a sleeper with a crack in only the rail sheet (Crack #2), and a sleeper with a crack that
penetrated the bottom of the concrete sleeper core and lost part of the concrete from the rail
sheet crack (Crack #3) were confirmed to have an effect on the dynamic stiffness. When
concrete is damaged, its dynamic stiffness is measured to be low. As shown in Figure 9b,d,
the initial dynamic stiffness was measured to be low in Cracks #1 and #3, where critical
sleeper cracks occurred. It was possible to predict using FRF that this phenomenon affected
the stiffness of concrete sleepers depending on the degree of damage caused to the sleepers.
In addition, as shown in Figure 9a,c, it was confirmed that in the case of undamaged
concrete sleeper and Crack #2, no direct damage to the concrete sleeper occurred, and the
rigidity of the concrete sleeper was not significantly affected.
(a)
 3.2.2. Dynamic Stiffness Prediction Results for Concrete Sleepers
The dynamic stiffness of the concrete sleepers was determined by evaluating the FR
from the impact hammer test measured in the field to understand the effect of sta
Materials 2024, 17, 1581 changes of a concrete sleeper on its dynamic stiffness. Figure 9a–d shows the dynam
11 of 15
stiffness of the concrete sleepers.

Materials 2024, 17, x FOR PEER REVIEW 11 o

(a)

(b)

(c)

(d)
Figure 9. Dynamic stiffness prediction results. (a) Undamaged concrete sleeper; (b) Crack #1
Figure 9. Dynamic stiffness prediction results. (a) Undamaged concrete sleeper; (b) Crack #1;
Crack #2; (d) Crack #3.
(c) Crack #2; (d) Crack #3.

A sleeper with a crack in the center that did not penetrate the ground (Crack #1
sleeper with a crack in only the rail sheet (Crack #2), and a sleeper with a crack that pe
trated the bottom of the concrete sleeper core and lost part of the concrete from the
sheet crack (Crack #3) were confirmed to have an effect on the dynamic stiffness. W
concrete is damaged, its dynamic stiffness is measured to be low. As shown in Figure 9
the initial dynamic stiffness was measured to be low in Cracks #1 and #3, where crit
sleeper cracks occurred. It was possible to predict using FRF that this phenomenon
 Materials 2024, 17, 1581 12 of 15

Materials 2024, 17, x FOR PEER REVIEW 12 of 15

3.3. Numerical Analysis Results
Materials 2024, 17, x FOR PEER REVIEW A harmonic analysis was performed on the mode with the highest frequency among 12 of 15
the six modes obtained through modal analysis. A harmonic analysis was used to perform
Figure 10a had a natural frequency of 125 Hz and had a bending deformation character-
the numerical analysis, and the mode shape of the concrete sleeper in bending mode was
istic. The second
investigated, asmode ininFigure
Figure10b, third mode analysis
in Figureshowed
10c, fourth
that mode in mode
Figurein10d,
Figure
fifth mode 10a
inhad ashown
natural
Figure 10e, frequency
and sixth
10.of
mode
The
125in
mode
Hz and had
Figure 10f a bending
had natural
the first
deformation
frequencies character-
of 198, 355,
Figure 10asecond
istic. The had a natural
mode in frequency of 125
Figure 10b, Hzmode
third and had a bending
in Figure deformation
10c, fourth mode characteristic.
in Figure 10d,
455,
The 514, andmode
749 Hz, respectively.
fifthsecond in Figure
mode in Figure 10e, and10b, third
sixth mode
mode in in Figure
Figure 10f10c,
hadfourth
naturalmode in Figureof10d,
frequencies 198,fifth
355,
mode in Figure 10e, and sixth mode
455, 514, and 749 Hz, respectively. in Figure 10f had natural frequencies of 198, 355, 455,
514, and 749 Hz, respectively.

(a) (b)
(a) (b)

(c) (d)
(c) (d)

(e)(e) (f)
(f)
Figure 10.10.
Figure Analysis
Analysisresults (mode
results(mode
(mode shape).
shape). (a) 1st
1st mode
mode (125.30
(125.30Hz);
Hz);(b)
(b) 2ndmodemode (197.50 Hz);
(c) (c)
Figure 10. Analysis results shape). (a)(a)
1st mode (125.30 Hz); (b) 2nd2nd
mode (197.50(197.50
Hz);Hz);
(c) 3rd
3rd3rd
mode
mode (355.22
(355.22Hz);
Hz);(d)
(d)4th
4thmode
mode(454.98
(454.98 Hz); (e)
(e) 5th
5thmode
mode(514.41
(514.41Hz);
Hz);(f)(f)
6th6th mode
mode (748.66
(748.66 Hz).Hz).
mode (355.22 Hz); (d) 4th mode (454.98 Hz); (e) 5th mode (514.41 Hz); (f) 6th mode (748.66 Hz).

Figure
Figure
Figure1111shows
11 showsthat
shows thatdeformation
that deformation occurs
deformationoccurs inthe
occursin
in theconcrete
the concrete
concrete sleepers
sleepers
sleepers owing
owing
owing toato
to a bending
abending
bending
mode
mode
mode with
with
withaaanatural
naturalfrequency
natural frequencyof
frequency of125.30
of 125.30Hz
125.30 Hzin
in the
inthe first
thefirst mode.
firstmode.
mode.

Figure 11. Analyzed natural frequency of concrete sleeper with normal condition.
Figure
Figure11.
11.Analyzed
Analyzednatural
natural frequency ofconcrete
frequency of concretesleeper
sleeperwith
with normal
normal condition.
condition.
3.4. Correlation between Concrete Sleeper Damage and Natural Frequency
3.4. Correlation between
To examine Concrete Sleeper
the structural Damage
soundness andnumerical
of the Natural Frequency
analysis and the field-meas-
ured
Tocondition,
examine the
the findings
structuralof the FRF analysis
soundness and
of the the calculation
numerical of dynamic
analysis and themass-and-
field-meas-
stiffness values were compared with the field-measured concrete sleepers.
ured condition, the findings of the FRF analysis and the calculation of dynamic mass-and-Damage oc-
curred mostly in the rail seat and sleeper core of the concrete sleepers
stiffness values were compared with the field-measured concrete sleepers. Damage oc-measured in this
investigation.
curred mostly inFigure 12 shows
the rail seat andthesleeper
damaged concrete
core of the sleeper.
concreteFigure 12a measured
sleepers shows damage
in this
investigation. Figure 12 shows the damaged concrete sleeper. Figure 12a shows damage
 Materials 2024, 17, 1581 13 of 15

3.4. Correlation between Concrete Sleeper Damage and Natural Frequency

Materials 2024, 17, x FOR PEER REVIEW 13 of 15
To examine the structural soundness of the numerical analysis and the field-measured
condition,
Materials 2024, 17, x FOR PEER REVIEW the findings of the FRF analysis and the calculation of dynamic mass-and- 13 of 15
stiffness values were compared with the field-measured concrete sleepers. Damage oc-
occurring at the bottom of the rail, and Figure 12b shows damage occurring at the center
curred mostly in the rail seat and sleeper core of the concrete sleepers measured in this
ofinvestigation.
the concrete sleeper.
Figure 12 shows the damaged concrete sleeper. Figure 12a shows damage
occurring at the
occurring at the bottom
bottom of
of the
therail,
rail,and
andFigure
Figure12b
12bshows
showsdamage
damageoccurring
occurringat at
thethe center
center of
of
thethe concrete
concrete sleeper.
sleeper.

(a) (b)
Figure
(a) 12. Damage of concrete sleepers. (a) Bottom of the rail; (b)
(b)center of concrete sleeper.
Figure 12. Damage
Damage of
of concrete
concrete sleepers.
sleepers. (a) Bottom of
of the rail; (b)
(b) center of
of concrete sleeper.
Figure
The12.field impact hammer tests (a) Bottom
reveal that the
therail;
change center concretemass
in dynamic sleeper.
of concrete
sleepersThe causedimpact by damage wastests minimal. However, there wasdynamic a difference inofthe dynamic
The field
field impact hammerhammer tests reveal reveal that
that the
the change
change in in dynamic mass mass of concrete
concrete
stiffness.
sleepers Differences
sleepers caused
caused by in
by damage the
damage was natural
was minimal. frequency
minimal. However, developed
However, therethere was depending
was aa difference on
difference in the
in the damage
dynamicde-
the dynamic
gree of theDifferences
stiffness. concrete sleeper,
Differences inthe
in whichfrequency
thenatural
natural isfrequency
consistent with the
developed
developed findings on
depending
depending ofon
Lam
the the and Wong
damage
damage [10].
de-
degree
greeThe
of natural
the concrete frequency
sleeper, in the
which bending
is mode
consistent of
with the
the
of the concrete sleeper, which is consistent with the findings of Lam and Wong [10]. undamaged
findings of Lamsleeper
and calculated
Wong [10]. by
numerical
The analysis
The natural was
naturalfrequency 125.30
frequencyininthe Hz, whereas
thebending
bendingmode the
mode natural
of of
thethe frequency
undamaged
undamaged of the
sleeper undamaged
sleepercalculated
calculated con-
by
crete sleeperanalysis
numerical
by numerical measured
analysiswaswasin the
125.30 field
Hz, Hz,
125.30 waswhereas
whereas 122.96 Hz.
the naturalThefrequency
the natural error between
frequency the
thetwo
of theofundamaged natural
undamaged con-fre-
crete
quencies sleeper
concrete measured
issleeper
approximately
measured in the
in field
1.87%, was 122.96
confirming
the field was Hz.
that
122.96 The
Hz. the error
The between
numerical
error the two
analysis
between natural
andnatural
the two fre-
measure-
quencies
frequencies is approximately
is approximately 1.87%,
1.87%,confirming
confirming that
thatthethe numerical
ment results are comparable. Since the numerical simulation and field measurement re- numerical analysis
analysis and
and measure-
measure-
mentfor
sults results
results arecomparable.
are
the undamaged comparable. Sincethe
Since
concrete thenumerical
sleepers numerical simulation
weresimulation
similar, isand
itand field
field measurement
measurement
inferred re-
results
that the numerical
for the
sults forundamaged
simulation the undamaged
model concrete
accurately sleepers
concrete were
sleepers
captures similar,
were
the behavior itofisconcrete
similar, inferred that the
it is inferred
sleepersthatnumerical
inthe simu-
thenumerical
field. There-
lation model
simulation accurately
model accuratelycaptures the
captures behavior
the of
behavior concrete
of concretesleepers
fore, the applicability of the numerical model and results were validated. Numerical anal- in
sleepers the
in field.
the Therefore,
field. There-
the applicability
fore, the applicability of theofnumerical
thefrequency
numericalmodel andand
model results were
results validated.
were validated.Numerical analysis
ysis confirmed the natural by modeling a concrete sleeper Numerical anal-
in good condition.
confirmed
ysis confirmed the natural
the naturalfrequency
frequencyby modeling
by modeling a concrete
a concretesleeper in good
sleeper condition.
in good The
condition.
The natural frequencies of intact concrete sleepers and the natural frequencies obtained
natural
The frequencies
natural frequenciesof intact concrete
of intact sleepers
concrete and theand
sleepers natural frequencies
the natural obtainedobtained
frequencies through
through the numerical
the numerical analysis analysis were confirmed
were confirmed to be similar.to be similar.
through the numerical analysis were confirmed to be similar.
The natural
The natural frequencies
natural frequencies observed
frequencies observed
observed in in the
in the damaged
the damaged
damaged concrete concretesleepers
concrete sleepers were
were approxi-
approxi-
The sleepers were approxi-
mately
mately27–57%
mately 27–57% lower
27–57% lower
lowerthanthan those
thanthose
those of
ofof
thethethe numerical
numerical
numerical analysis
analysis
analysis results results
(125.30
results (125.30 as Hz),
Hz),Hz),
(125.30 as illus-
illustrated
as illus-
trated in
in Figure
trated Figure 13:
13: Crack
in Figure Crack
#1 (91.09
13: Crack #1 (91.09 Hz),
Hz), Crack
#1 (91.09 Crack
#2 (85.47
Hz), Crack #2 (85.47
Hz), and
#2 (85.47 Hz),
Hz),Crack and Crack
#3 (53.59
and Crack #3 (53.59
Hz). Hz).
#3 (53.59 Hz).

Figure 13. Comparison of modal test and numerical analysis results.

Figure
Figure13.
13.Comparison
Comparisonof
of modal
modal test and numerical
test and numericalanalysis
analysisresults.
results.
From the impact hammer tests, it can be seen that differences in the natural frequency
From the
developed impact hammer
depending tests, it can
on the damage be seen
degree of thethat differences
concrete in Thus,
sleeper. the natural frequency
even though
developed depending on the damage degree of the concrete sleeper. Thus,
the visible damage of the concrete sleepers is comparable, they may have varying struc- even though
the visible
tural damage
soundness of the
values concrete
due sleepers
to unseen is comparable,
damage. theycomputational
Field studies and may have varying struc-
analyses
tural
show soundness values due
that a quantitative to unseen
condition damage.based
assessment Fieldonstudies and natural
measured computational analyses
frequencies is
show that ausing
attainable quantitative
an impact condition
hammer assessment
test, which isbased
simpleontomeasured natural frequencies is
use in the field.
Materials 2024, 17, 1581 14 of 15

From the impact hammer tests, it can be seen that differences in the natural frequency
developed depending on the damage degree of the concrete sleeper. Thus, even though
the visible damage of the concrete sleepers is comparable, they may have varying struc-
tural soundness values due to unseen damage. Field studies and computational analyses
show that a quantitative condition assessment based on measured natural frequencies is
attainable using an impact hammer test, which is simple to use in the field.

4. Conclusions
This study proposed the impact hammer test as a quantitative assessment method to
evaluate the structural soundness of concrete sleepers. It is a portable test method that is
easy to apply in practice, and the results are as follows.
The error between the simulated and measured natural frequencies of the undamaged
concrete sleeper was approximately 1.87%, confirming the appropriateness of the numerical
analysis model. The measured natural frequencies of the damaged concrete sleepers were
approximately 27–57% lower than the studied natural frequencies of the undamaged
concrete sleeper, demonstrating that the damage to the concrete sleepers had a direct effect
on their natural frequency.
The dynamic masses of the natural frequency prediction of the undamaged and
damaged concrete sleepers were compared using the impact hammer test in the field,
and it was found that the dynamic masses of the damaged concrete sleepers were at least
10 times smaller than that of the undamaged concrete sleeper. The undamaged concrete
sleeper exhibited a higher dynamic stiffness than the damaged concrete sleepers, and the
natural frequencies estimated for the comparably damaged sleepers (53.59–91.09 Hz) were
1.34–2.29 times lower than the undamaged concrete sleeper (122.96 Hz). These findings
show that the dynamic mass and stiffness of damaged concrete sleepers have a lower
natural frequency than undamaged concrete sleepers, demonstrating the relationship
between concrete sleeper damage and natural frequency values.
The structural integrity of a large number of sleepers can be quickly and easily tested
quantitatively by determining their natural frequency in the field using the impact hammer
test. In the future, the relationship between the degree of concrete sleeper damage (e.g.,
crack length, crack width, and crack pattern) and natural frequency values should be
further investigated. Classifying concrete sleepers based on their state using natural
frequency values recorded in the field will allow these structures to be more effectively
maintained. In addition, it is believed that the structural integrity of all concrete sleepers
can be quantitatively evaluated through analysis of natural frequencies measured using
impact hammer testing.

Author Contributions: Conceptualization, J.-Y.C., J.-S.C. and T.-H.S.; methodology, J.-Y.C.; software,
T.-H.S. and S.-H.K.; formal analysis, T.-H.S. and S.-H.K.; investigation, T.-H.S. and J.-S.C.; data curation,
T.-H.S. and J.-S.C.; writing—original draft preparation, J.-Y.C., T.-H.S. and S.-H.K.; writing—review and
editing, S.-H.K. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Data are contained within the article.
Conflicts of Interest: Author Tae-Hyung Shin was employed by the company Seoul Metro. The
remaining authors declare that the research was conducted in the absence of any commercial or
financial relationships that could be construed as a potential conflict of interest.
Materials 2024, 17, 1581 15 of 15

References
1. Korea National Railway. Guidelines for Track Facility Performance Evaluation; Korea Ministry of Land Infrastructure and Transport
(MOLIT): Sejong, Republic of Korea, 2022. (In Korean)
2. Clark, A.; Kaewunruen, S.; Janeliukstis, R.; Papaelias, M. Damage detection in railway prestressed concrete sleepers using
acoustic emission. In Proceedings of the IOP Conference Series: Materials Science and Engineering, IMST 2017, Riga, Latvia,
27–29 September 2017; Volume 251, p. 012068.
3. Choi, J.Y. Qualitative Analysis for Dynamic Behavior of Railway Ballasted Track. Ph.D. Thesis, Technical University of Berlin,
Berlin, Germany, February 2014.
4. Kim, J.W.; Jung, H.Y. Vibrational characteristics of the deteriorated railway plate girder bridge by full-scale experimental modal
analysis. J. Korean Soc. Steel Constr. 2012, 24, 119–128. [CrossRef]
5. Esmaeili, M.; Majidi-Parast, S.; Hosseini, A. Comparison of dynamic lateral resistance of railway concrete, wooden and steel
sleepers subjected to impact loading. Road Mater. Pavement Des. 2019, 20, 1779–1806. [CrossRef]
6. Braunfelds, J.; Senkans, U.; Skels, P.; Porins, J.; Haritonovs, V.; Spolitis, S.; Bobrovs, V. Development of the strain measurement
calibration technique for road pavement structural health monitoring applications using optical FBG sensors. In Proceedings of
the 2023 PhotonIcs & Electromagnetics Research Symposium (PIERS), Prague, Czech Republic, 3–7 July 2023; pp. 1060–1065.
7. Chung, J.S.; Lee, J.S.; Choi, J.Y. A study on estimation method of concrete sleeper strength for sleeper floating track using rebound
hardness test method. J. Converg. Cult. Technol. (JCCT) 2022, 8, 277–282.
8. Choi, J.Y.; Park, S.W.; Chung, J.S. Estimation method of resilience pads spring stiffness for sleeper floating tracks based on track
vibration. J. Converg. Cult. Technol. (JCCT) 2023, 9, 1057–1063.
9. Sapidis, G.M.; Kansizoglou, I.; Naoum, M.C.; Papadopoulos, N.A.; Chalioris, C.E. A deep learning approach for autonomous
compression damage identification in fiber-reinforced concrete using piezoelectric lead zirconate titanate transducers. Sensors
2023, 24, 386. [CrossRef]
10. Nielsen, D.; Palmer, E. Modal analysis experiments on prestressed concrete sleepers. In Proceedings of the 16th East Asian-Pacific
Conference on Structural Engineering and Construction, Brisban, Australia, 3–6 December 2019.
11. Kaewunruen, S.; Remennikov, A.M. Sensitivity analysis of free vibration characteristics of an in situ railway concrete sleeper to
variations of rail pad parameters. J. Sound Vib. 2006, 298, 453–461. [CrossRef]
12. Kaewunruen, S.; Remennikov, A. Non-destructive evaluation for dynamic integrity of railway track structure. In Proceedings of
the International Conference on Structural Integrity and Failure—SIF2006, Sydney, Australia, 27–29 September 2006; pp. 294–299.
13. Shin, T.H. Integrity Assessment of Concrete Sleeper Using Modal Test Technique. Ph.D. Thesis, Dongyang University, Yeongju-si,
Republic of Korea, February 2022.
14. Lam, H.F.; Wong, M.T. Railway ballast diagnose through impact hammer test. Procedia Eng. 2011, 14, 185–194. [CrossRef]
15. Kaewunruen, S.; Remennikov, A.M. Dynamic effect on vibration signatures of cracks in railway prestressed concrete sleepers.
Adv. Mater. Res. 2008, 41, 233–239. [CrossRef]
16. You, R.; Wang, J.; Ning, N.; Wang, M.; Zhang, J. The typical damage form and mechanism of a railway prestressed concrete
sleeper. Materials 2022, 15, 8074. [CrossRef] [PubMed]
17. Real, J.J.; Sanchez, M.E.; Real, T.; Sanchez, F.J. Experimental modal analysis of railway concrete sleepers with cracks. Struct. Eng.
Mech. 2012, 44, 51–60. [CrossRef]
18. Hassan, M.H.M. Advances in Structural Health Monitoring; IntechOpen: London, UK, 2019; Chapter 6.
19. ANSYS. ANSYS®2007 ANSYS Workbench, version 2021 R2; ANSYS Inc.: Cannonsberg, PA, USA, 2021.

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.