Article

A Numerical Tool for Assessing Random Vibration-Based

Fatigue Damage Diagnosability in Thermoplastic Coupons
Niki Tsivouraki 1,2 , Spilios Fassois 2 and Konstantinos Tserpes 1, *

1 Laboratory of Technology & Strength of Materials, Department of Mechanical Engineering & Aeronautics,
University of Patras, 26504 Patras, Greece; niki.tsivouraki@ac.upatras.gr
2 Stochastic Mechanical Systems and Automation Laboratory, Department of Mechanical Engineering &
Aeronautics, University of Patras, 26504 Patras, Greece; fassois@upatras.gr
* Correspondence: kitserpes@upatras.gr; Tel.: +30-2610-969498

Abstract: A numerical tool is developed to simulate the random vibration-response-only-

based fatigue delamination diagnosability in thermoplastic coupons. That is the ability to
both detect damage and identify its current severity, aiming to establish a virtual framework
for optimizing diagnosability methods. The numerical tool employs the FE method. It
comprises two modules: a fatigue delamination module and a random vibration module.
The first module implements a fatigue crack growth model based on the cohesive zone
modeling method to predict delamination accumulation, while the second module uses an
experimentally verified FE model of the delaminated coupon to predict its random vibration
response. Delamination accumulation is evident in the ‘predicted’ FE-based power spectral
densities. The model’s capability to diagnose delamination is demonstrated using seven
different damage metrics based on simulated random vibration responses, enabling damage
detection and severity assessment (increasing trend guides to distinguishing each fatigue
state from its counterparts). Comparisons with their experimentally obtained counterparts
are also used in the assessment. The procedure clearly suggests that the proposed numerical
tool may be reliably used for virtually assessing the efficacy of random vibration-based
fatigue damage diagnosability for any given structure and also to aid the user in selecting
the method’s parameters for virtual diagnosability optimization.

Keywords: thermoplastic composites; CFRPs; vibration-based monitoring; numerical

Academic Editor: Francesco Aymerich
random vibration model; fatigue damage; damage diagnosis
Received: 23 February 2025
Revised: 16 March 2025
Accepted: 19 March 2025
Published: 23 March 2025
1. Introduction
Citation: Tsivouraki, N.; Fassois, S.;
Since the launch of the Boeing 787 Dreamliner, in which composites account for
Tserpes, K. A Numerical Tool for
Assessing Random Vibration-Based
nearly 50% of its structural weight, and the Airbus A350 XWB, where composites make up
Fatigue Damage Diagnosability in about 52% of its structural weight, Carbon Fiber Reinforced Plastics (CFRP) have become
Thermoplastic Coupons. J. Compos. a dominant material in the aeronautical industry [1]. While thermosets are primarily
Sci. 2025, 9, 153. https://doi.org/ used in aircraft structures, thermoplastics are gaining popularity due to their advantages,
10.3390/jcs9040153
particularly in terms of dynamic behavior, recyclability, and weldability.
Copyright: © 2025 by the authors. Fatigue behavior remains a critical design factor for composite aircraft structures due
Licensee MDPI, Basel, Switzerland. to its stochastic nature, which is further enhanced by the inherent anisotropy of CFRPs.
This article is an open access article
To date, CFRP structures have primarily been designed based on experimental data, as
distributed under the terms and
reliable predictive models are still lacking.
conditions of the Creative Commons
Attribution (CC BY) license
Improving safety has become a paramount concern for aircraft manufacturers and
(https://creativecommons.org/ operators, particularly in light of recent incidents and accidents that have resulted in the
licenses/by/4.0/). grounding of hundreds of aircraft for extended periods. Structural Health Monitoring

J. Compos. Sci. 2025, 9, 153 https://doi.org/10.3390/jcs9040153

J. Compos. Sci. 2025, 9, 153 2 of 22

(SHM) offers a valuable solution by providing reliable diagnostic and prognostic tools that
can help prevent critical structural failures [2,3]. Additionally, SHM has the potential to
enhance the efficiency and effectiveness of maintenance procedures, ultimately contributing
to improved operational safety and reduced downtime.
SHM of composite structures has undergone rapid advancements over the past
decade [3,4]. Traditionally, offline experimental techniques have been used to detect
the initiation of damage, track its progression, and, in some cases, predict the residual
fatigue life and strength of structures [5]. Common methods include X-rays [6,7], acoustic
emission [8–10], optical and surface observation [11,12], and post-damage analysis through
Scanning Electron Microscopy (SEM) [13,14]. In contrast, a promising family of online SHM
methods relies on vibration-data-driven methods. These offer two key advantages [15]:
Continuous, real-time information on the structure’s health state without disrupting normal
operation and operating via sensors that are not necessarily in the neighborhood of the
damage location. This renders vibration-based methods especially useful for the real-time
monitoring of complex composite structures.
Vibration-based SHM has been widely used across various engineering sectors,
though predominantly in beam-like structures. Most studies have employed exper-
imental approaches [16–27], with fewer studies utilizing a combined experimental–
numerical approach [28–36], an experimental–analytical approach [37], or a purely nu-
merical approach [38–43]. Certain studies [44–46] employed the wavelet transform, while
others [47–49] utilized artificial neural networks. The majority of the published works
focus on damage detection, with only a small subset [24,27] addressing fatigue damage
of composite materials. All experimental–numerical approaches and purely numerical
approaches have exploited experimental data to predict damage accumulation or fatigue
life of the structures, while only the work of Sartorato et al. [39] employed numerical
analyses to support the design of a piezoelectric-based SHM system for metallic materials.
Additionally, some important studies [19,50,51] examined the non-linearities, which are
inserted in the problem of vibration–fatigue damage.
The goal of this study is the development of an integrated fatigue damage–random vi-
bration numerical tool for the parametric simulation of random vibration-based progressive
fatigue damage detectability for thermoplastic coupons. The implementation of the numer-
ical tool requires no experimental data, with the exception of the fatigue model. Following
its validation against mechanical fatigue and random vibration tests, the numerical models
are used to assess the effects of specific parameters on fatigue damage diagnosability.
The study’s innovative aspects may be summarized as follows:
• Development of a numerical methodology of random vibration-response-only-based
fatigue damage diagnosability for thermoplastic coupons;
• Development of an efficient simulation process for random vibration experiments;
• Examination of progressive fatigue damage through numerical random vibration
responses in the frequency domain;
• Employment of several damage metrics for fatigue damage diagnosability in thermo-
plastic coupons using numerical random vibration response data;
• Generation of an integrated numerical model/tool, which can be used to support
random vibration-response-only-based SHM systems for composite structures.

2. Conceptual Framework
The concept of detecting damage in materials and structures by exploiting random
vibration response is based on the premise that damage alters their characteristics (stiffness
and damping), thus their dynamical behavior, which subsequently leads to changes in their
vibration response characteristics [52]. Yet, as damage characteristics can also be affected by
J. Compos. Sci. 2025, 9, 153 3 of 22

other factors, such as operational and environmental conditions, care should be exercised
to distinguish between actual damage and other external influences [53–58].
Especially, detecting fatigue damage from random vibration responses and modal
characteristics, such as frequency, damping ratio, spectral distance, PSD estimates, etc.,
requires the ability to deeply understand the initial propagation, constant evolution, and
the critical damage modes before failure, with the aim of understanding the vibration
behavior and the correlation of it with fatigue damage. The proposed fatigue degradation
model is based on the loading envelope approach proposed by [59] in which the cyclic
loading is simulated by a linear force incremental to the maximum value of applied force.
The implementation of the model was undertaken on an element basis and was based on a
modified Paris’ law. For more details, Section 3 of this present paper, as [59–61], describes
the fatigue delamination growth model.
The implementation of the random vibration model also needs a comprehensive un-
derstanding of the exact and with which tendency the modal characteristics are influenced
by progressive fatigue damage. The random vibration simulation process lies in the modal
superposition method to characterize the linear dynamic behavior of the structure [62].

3. The Numerical Tool for Fatigue Damage Diagnosability

Both the Fatigue Delamination Module (FDM) and the Random Vibration Module
(RVM) were developed in LS-Dyna. These are two distinct models functioning dependently
on each other. The FDM has been described in detail in [61], so only a brief overview of it is
presently provided for the sake of presentation completeness.

3.1. The Fatigue Delamination Module (FDM)

The FDM was based on the fatigue crack growth model developed in [60] for simu-
lating interfacial fracture in co-consolidated thermoplastic joints. The model utilizes the
cohesive zone modeling method and incorporates a modified Paris’ law, expressed as:

da
= c( Gmax )m (1)
dN
da
where dN is the fatigue crack growth rate, c and m are the modified Paris’ law parameters
for each simulated FS, and Gmax represents the maximum energy release rate for the affected
element at a given FS. The instantaneous energy release rate Gi for each cohesive element
is consistently set to Gmax , reflecting the maximum energy release rate of the actual loading
spectrum due to the use of the load envelope technique. The intermediate mixed-mode
parameters c and m are derived using the prediction model of Russel and Street [63], which
requires only pure mode I and mode II fatigue experimental data as inputs.
Numerical fatigue degradation was implemented using the method of cumulative
static and fatigue element damage parameters, described by:

dtot = ds + d f (2)

where dtot is the total damage variable (ranging from 0 for undamaged to 1 for failed),
while ds and d f are the static and fatigue contributing damage variables, respectively. The
stress state σ for each element in the fatigue-activated zone is then degraded using:

σ = (1 − dtot )σmax (3)

where σmax is the maximum traction of the affected element.

In this study, the modified Paris’ law was applied, accounting solely for Gmax . The
fatigue crack growth model was fully validated using the same experimental parameters,
 where 𝜎 is the maximum traction of the affected element.
In this study, the modified Paris’ law was applied, accounting solely for 𝐺
J. Compos. Sci. 2025, 9, 153 fatigue crack growth model was fully validated using the same experimental 4 of 22 param
such as stress ratio and maximum load percentage, for the same interfaces as deta
[60].
such as stress ratio and maximum load percentage, for the same interfaces as detailed
Fatigue loading was implemented using the loading envelope technique [59],
in [60].
the actual
Fatiguesinusoidal
loading wasspectrum was
implemented approximated
using by a constant
the loading envelope forcewhere
technique [59], correspond
the maximum fatigue load value, as illustrated in Figure 1.
the actual sinusoidal spectrum was approximated by a constant force corresponding to the
maximum fatigue load value, as illustrated in Figure 1.

Figure 1. The load envelope for the modeling of fatigue loading.

Figure 1. The load envelope for the modeling of fatigue loading.
During each numerical iteration, the rate of fatigue crack growth was calculated using
Equation (1), which operates in a singular direction. This resulted in the establishment
During each numerical iteration, the rate of fatigue crack growth was calcula
of an equivalent rate of damage development uniformly across all axes, disregarding the
ing Equation (1), which operates in a singular direction. This resulted in the establis
orthotropic nature of thermoplastics and the complexity of three-dimensional interfaces.
of
Toan equivalent
address rate of adamage
this limitation, development
specific length measurement uniformly across
was assigned all axes,
to each disregard
element,
orthotropic nature of thermoplastics
guiding the progression of damage by consideringand theinputcomplexity
from theof three-dimensional
cohesive integration inte
points. This measurement is particularly significant in cases where
To address this limitation, a specific length measurement was assigned to each el there is considerable
variation in element aspect ratios and irregularities within the interfaces. The calculation
guiding the progression of damage by considering input from the cohesive integ
of the fatigue damage rate, d f , requires the characteristic element length, le , as described
points. This measurement is particularly significant in cases where there is consid
in detail in [29]. le was computed for each iteration through an internal loop, which
variation
determined inthe
element
minimum aspect ratios
distance fromandtheirregularities
midpoint of awithin the element
non-failed interfaces. The calc
to the
ofmidpoint
the fatigue damage rate,
of neighboring 𝑑𝑓, Following
elements. requires the this,characteristic element
the elements’ data, length,
including G I and𝑙 , as des
G , were
in Idetail
I in [29]. 𝑙 was computed for each iteration through an internal loop, wh
calculated, and a check for fatigue module activation was performed. If either
G I or G I I exceeded the threshold value Gth , the fatigue module was activated; otherwise,
termined the minimum distance from the midpoint of a non-failed element to th
static degradation was applied. The value of Gth was based on literature. The stress
point
σ of the of cohesive
neighboring elementelements.
is governed Following this, the elements’
by the mixed-mode stress, whichdata, the 𝐺 a
including
combines
were calculated,
contributions from and a checkand
both normal fortangential
fatigue stress
module activation
components, was performed.
according to LS-Dyna’sIf either
𝐺MAT_138
exceeded the threshold
formulation [35]. Therefore, 𝐺 , theelements
valuecohesive fatiguesubjected
moduletowas activated;
fatigue damage are otherwise
degradation was applied. The value of 𝐺 was based on literature. The stress 𝜎
fully governed by mixed-mode stress ( σ ) for their response, degradation, and failure. The
iterative numerical procedure for the fatigue crack growth model is illustrated in Figure 2.
cohesive element is governed by the mixed-mode stress, which combines the co
The pure mode I and mode II modified Paris’ laws used in the study are:
tions from both normal and tangential stress components, according to LS-
 
dα
MAT_138 formulation [35]. Therefore, cohesive
4.23 elements subjected to fatigue dam
= 0.0027G I,max (4)
dN t
fully governed by mixed-mode stress 𝜎 for their response, degradation, and
The iterative numerical procedure for the fatigue crack growth model is illustrated
 
dα
= 0.00040G5.31
I I,max (5)
ure 2. dN II
   
ThedN
where dapure mode
da
, dN I and
and mode
G I,max II modified
, G I I,max Paris’
are the fatigue laws
crack used
growth in the
rates study
and the are:
maxi-
I II
mum energy release rates per cycle for mode I and mode II, respectively. More details on
𝑑𝛼 .
0.0027𝐺
the derivation of Equations (1) and (2) can be found in [60].
,
𝑑𝑁
 𝑑𝑁

where , and 𝐺 , , 𝐺 , are the fatigue crack growth rates and the maxi-
J. Compos. Sci. 2025, 9, 153
mum energy release rates per cycle for mode I and mode II, respectively. More details5on
of 22

the derivation of Equations (1) and (2) can be found in [60].

Figure 2. 2.
Figure Flowchart of of
Flowchart thethe
fatigue crack
fatigue growth
crack model.
growth model.

3.2.
3.2. The
The Random
Random Vibration
Vibration Module
Module (RVM)
(RVM)
The
The RVMwas
RVM wasdeveloped
developed using
using LS-Dyna’s
LS-Dyna’sgraphical interface
graphical pre-processor
interface LS-PrePost
pre-processor LS-Pre-4.8.
The4.8.
Post damping ratio was
The damping determined
ratio using: using:
was determined

Δ𝑓 ∆f
𝜉 ξ= (6) (6)
2𝑓 2 f n
where𝑓 f nisisthe
where thefirst
first natural
natural frequency
frequency andand ∆𝛥𝑓f isisthe
thebandwidth
bandwidthofofthe half-power
the points
half-power [40].
points
To To
[40]. prevent
prevent hourglass-induced
hourglass-inducedinstabilities,
instabilities,aa conversion of the
conversion of theexcitation
excitationamplitude
amplitude
was necessary.
was necessary.
The
The applied
applied excitation
excitation was
was calculated
calculated using
using [64]:
[64]:
 
dB = 20 × log 𝑃 P
dB [ ]
20 ∗ log 10 5 (7) (7)
2 ∗ 102 × 10

where𝑃 Pisisthe and22∗ × 5 is reference sound pressure

where thepressure
pressureofofwhite
whitenoise
noiseininPascal
Pascaland 10 10is reference sound pressure
in Pascal.
in Pascal.
Respectively,
Respectively, thenumerical
the numericalresults
resultswere
wereconverted
convertedfrom
from[mm/s]
[mm/s] toto [dB]
[dB] using
using [65]:
[65]:

𝑣 20 ∗ log 𝑣 ∗ 10 (8) (8)

 
vdB = 20 × log10/ vmm/s × 106
where 𝑣 is the resulting velocity in dB and 𝑣 / is the resulting velocity in mm/s.
where vdB is the resulting velocity in dB and vmm/s is the resulting velocity in mm/s.
With the aim of calibrating the model parameters, and checking how basic parameters
(such as boundary conditions, measurement points, and mesh density—more details
in Section 4.5) influence the coupon’s behavior, Frequency Response Functions (FRF)
analyses were implemented. The geometry (see below—Section 4.4 and the material model
(orthotropic elastic—see below Table 1) were the same as described in the RVM model.
J. Compos. Sci. 2025, 9, 153 6 of 22

Table 1. Orthotropic elastic properties of the thermoplastic lamina [61].

Property Value
Density (ton/mm3 ) 1.75 × 10−9
Elastic modulus—a direction (MPa) 95,000
J. Compos. Sci. 2025, 9, x FOR PEER REVIEWElastic modulus—b direction (MPa) 8500 6 of 23
Elastic modulus—c direction (MPa) 8500
Shear modulus—ab direction (MPa) 4300
Shear modulus—bc direction (MPa) 3571
With the
Shear aim of calibrating
modulus—ac directionthe model parameters, and checking
(MPa) 4300 how basic parame-
ters (such asPoisson’s
boundaryratio—νba
conditions, measurement points, and mesh 0.027
density—more details
Poisson’s ratio—νca
in Section 4.5) influence the coupon’s behavior, Frequency Response 0.024
Functions (FRF) anal-
Poisson’s ratio—νcb
yses were implemented. The geometry (see below—Section 4.4 and the material model 0.172
(orthotropic elastic—see below Table 1) were the same as described in the RVM model.
The FE modeling
modeling procedure
procedurefor forthe
thedamaged
damagedcoupon
couponisisoutlined
outlined
ininthethe flowchart
flowchart of
of Figure
Figure 3. This
3. This process
process combined
combined the the
RVM RVM
withwith
the the
FDM. FDM.For For
eacheach FS, FDM
FS, the the FDM was was
exe-
executed to predict
cuted to predict thethe delamination
delamination state,which
state, whichcorresponds
correspondsto todeleted
deleted cohesive
cohesive elements
representing the the interfaces
interfacesamong amonglayers.
layers.All
Allcohesive
cohesive elements
elements were exported
were exported to atody-
a
nain.ASCII file, from which the deleted cohesive elements were isolated
dynain.ASCII file, from which the deleted cohesive elements were isolated into a new into a new keyfile
(set.k). The set.k
keyfile (set.k). Thefile was
set.k filethen
wasimported into the
then imported intoinitial FE model,
the initial and and
FE model, by reversing
by reversingthe
deleted elements, duplicated cohesive layers (both healthy and
the deleted elements, duplicated cohesive layers (both healthy and damaged) were cre-damaged) were created.
Finally, the deleted
ated. Finally, cohesive
the deleted elements
cohesive saved in
elements the set.k
saved fileset.k
in the were removed,
file and the and
were removed, remain-
the
ing duplicate nodes were merged. This procedure effectively “inserts”
remaining duplicate nodes were merged. This procedure effectively “inserts” the delam- the delaminated
(deleted) cohesive
inated (deleted) elements
cohesive into theinto
elements cohesive layers. layers.
the cohesive

Figure 3. The
Figure 3. The RVM
RVM development
development process
process for
for the
the damaged
damaged coupon.
coupon.

An important clarification should be given at this point: the coupons that were
An important clarification should be given at this point: the coupons that were sub-
submitted to T-T fatigue had transverse and off-axis cracks, except for delamination, which
mitted to T-T fatigue had transverse and off-axis cracks, except for delamination, which
surely contributed to the final failure. The initiation and propagation of fatigue damage,
surely contributed to the final failure. The initiation and propagation of fatigue damage,
which was validated from C-Scan images, showed that delamination was the main cause
which was validated from C-Scan images, showed that delamination was the main cause
of failure [61]—so, as with the FDM, the RVM also considers only the delamination as
damage mode.

4. Numerical Tool Assessment and Experimental Verification

 cess for structural parts. A key difference to be kept in mind is that while in the experi-
mental procedure, all damage modes may be present, the numerical tool presently in-
cluded only delamination damage. Thus, the assessment was based on the premise that
J. Compos. Sci. 2025, 9, 153 7 of 22
delamination is the predominant damage mode in coupons subjected to tension-tension
fatigue—which was validated by C-Scan images [61]. The numerical tool assessment was
the application of the above two modules into a thermoplastic coupon, which was exper-
of failure [61]—so, as with the FDM, the RVM also considers only the delamination as
imentally validated at various stages. The FDM was validated by C-Scan images [61], and
damage mode.
the RVM was validated by the experimental PSD Welch estimates [27,66]. Any deviation
between the numerical and experimental results is justified below.
4. Numerical Tool Assessment and Experimental Verification
In [27],and
4.1. Material theCoupons
concept of using random vibration response to detect fatigue damage in
thermoplastic coupons was experimentally demonstrated. The objective of this present
The coupons employed in the study were fabricated from T700 carbon fiber rein-
work was the numerical development of [27] that comprehensively examines the influence
forced low-melt semi-crystalline PAEK (polyaryletherketone) matrix prepregs (Toray
of testing parameters on the method’s efficiency and virtually optimizes the process for
Cetex® TC1225) [67–69]. The fiber volume fraction was 66%, and each ply was 0.14 mm
structural parts. A key difference to be kept in mind is that while in the experimental
thick. The lay-up was quasi-isotropic, with the sequence [−45°/0°/45°/90°]2s. The dimen-
procedure, all damage modes may be present, the numerical tool presently included only
sions of the coupons were 250 mm 25 mm 2.24 mm. Tabs made of the same mate-
delamination damage. Thus, the assessment was based on the premise that delamination is
rial, measuring 50 mm 25 mm 2.24 mm, were bonded to the coupons. In total, 23
the predominant damage mode in coupons subjected to tension-tension fatigue—which
coupons were subjected to mechanical loads and excitation. The properties of the material
was validated by C-Scan images [61]. The numerical tool assessment was the application of
are provided in Table 1 [61] and the dimensions od the coupon are illustrated in Figure 4.
the above two modules into a thermoplastic coupon, which was experimentally validated at
various
Table stages. Theelastic
1. Orthotropic FDMproperties
was validated
of the by C-Scan images
thermoplastic [61],
lamina [61].and the RVM was validated
by the experimental PSD Welch estimates [27,66]. Any deviation between the numerical
and experimental results is Property
justified below. Value
Density (ton/mm3) 1.75 × 10−9
4.1. Material andElastic modulus—a direction (MPa)
Coupons 95,000
The coupons employed in the study were fabricated from T700 carbon8500
Elastic modulus—b direction (MPa) fiber reinforced
Elastic modulus—c direction (MPa)
low-melt semi-crystalline PAEK (polyaryletherketone) matrix prepregs (Toray Cetex® 8500
Shear modulus—ab direction (MPa) 4300
TC1225) [67–69]. The fiber volume fraction was 66%, and each ply was 0.14 mm thick. The
Shear modulus—bc direction (MPa) ◦ ◦ 3571
lay-up was quasi-isotropic, with the sequence [−45 /0 /45◦ /90◦ ] 2s. The dimensions of
Shear modulus—ac direction (MPa) 4300
the coupons were 250 mm × 25 mm × 2.24 mm. Tabs made of the same material, measur-
Poisson’s ratio—νba 0.027
ing 50 mm × 25 mm × 2.24 mm, were bonded to the coupons. In total, 23 coupons were
Poisson’s ratio—νca 0.024
subjected to mechanical loads and excitation. The properties of the material are provided
Poisson’s ratio—νcb 0.172
in Table 1 [61] and the dimensions od the coupon are illustrated in Figure 4.

Figure 4. Dimensions of coupons.

Figure 4. Dimensions of coupons.
4.2. Brief Overview of the Experimental Procedure
4.2. Brief Overview of the Experimental Procedure
The random vibration test setup is illustrated in Figure 5, while Figure 6 provides a
sketch ofrandom
The vibration
the testing test setup
and data is illustrated
acquisition in Figure
processes. 5, while
The test setupFigure 6 provides
consisted a
of a Laser
sketch
Dopplerof Vibrometer
the testing and data
(LDV) acquisition
OFV-505 processes.
(Polytec GmbH,The test setup
Karlsruhe, consistedits
Germany), ofOFV-5000
a Laser
Doppler Vibrometer (LDV) OFV-505 (Polytec GmbH, Karlsruhe, Germany),
controller (Polytec GmbH, Karlsruhe, Germany), an NI-9230 acquisition card (National its OFV-5000
controller
Instruments,(Polytec GmbH,
Austin, Karlsruhe,
TX, USA), and a PC.Germany), an NI-9230
The applied acquisition
excitation card (National
was a MATLAB (R2024a)-
Instruments, Austin, TX, USA), and a PC. The applied excitation
generated band-limited (within the 0–6.4 kHz range) white noise. The coupons were was a MATLAB
(R2024a)-generated
supported via elastic band-limited
cords to simulate the 0 boundary
(withinfree-free 6.4 kHz range) white Vibration
conditions. noise. Thevelocity
cou-
pons were supported
measurements via using
were taken elasticthe
cords
laser to simulate which
vibrometer, free-free
wasboundary
focused onconditions.
a reflective
surface positioned at the center of the upper face of each coupon. The random vibration
test was implemented every 10,000 fatigue cycles for all coupons.
 cused on a reflective surface positioned at the center of the upper face of each coupon. The
random vibration test was implemented
Vibration every 10,000
velocity measurements fatigue
were taken cycles
using the for allvibrometer,
laser coupons. which was fo-
J. Compos. Sci. 2025, 9, 153 cused on a reflective surface positioned at the center of the upper face of each coupon. 8The
of 22
random vibration test was implemented every 10,000 fatigue cycles for all coupons.

Figure 5. Random vibration test setup (adapted from [27]).

Figure5.5.Random
Figure Randomvibration
vibrationtest
testsetup
setup (adapted
(adapted from [27]).

Figure 6. Sketch of the testing and acquisition processes for the random vibration test (adapted from
Figure 6. Sketch of the testing and acquisition processes for the random vibration test (adapted
[27]).
Figure 6. Sketch of the testing and acquisition processes for the random vibration test (adapted from
from [27]).
[27]).
Non-parametric signal analysis was performed, based on which Welch-based Power
Non-parametric signal analysis was performed, based on which Welch-based Power
Spectral Density (PSD) estimates were obtained for the Healthy State (HS) and various
Non-parametric signal
Spectral analysis
Density was
(PSD) performed,
estimates were based
obtained on for
which Welch-based
the Healthy Power
State (HS) and various
Fatigue States (FSs). The signal characteristics and the Welch estimation parameters are
Spectral Density Fatigue
(PSD) States (FSs).
estimates The obtained
were signal characteristics
for the Healthyand the Welch
State estimation
(HS) parameters are
and various
summarized in Table 2. Progressive fatigue damage detection was conducted on a popu-
summarized
Fatigue States (FSs). Theofsignalin Table 2. Progressive
characteristics fatigue
and the Welch damage detection
estimation was conducted
parameters areon a popu-
lation 13 coupons, with one representative sample selected for model comparison and
lation of 13 coupons, with one representative sample selected for model comparison and
summarized in Table 2. Progressive
validation. fatigue
The results acrossdamage detection
the coupons was conducted
were grossly consistent,on a popu-
with random devia-
validation. The results across the coupons were grossly consistent, with random deviations
lation of 13 coupons,tionswith one representative
attributed to the stochasticsample
nature ofselected
fatigue for model
damage comparison
accumulation. and
This variability
attributed to the stochastic nature of fatigue damage accumulation. This variability is
is discussed by the authors in references [66,70]. The PSD
validation. The results across the coupons were grossly consistent, with random devia- Welch estimates for Coupon #3
discussed by the authors in references [66,70]. The PSD Welch estimates for Coupon #3
under
tions attributed tounder HS and various
the stochastic FSs
natureFSs (70 PSD damage
of (70
fatigue estimates under the HS and This35 under each FS) are
HS and various PSD estimatesaccumulation.
under the HS and 35variability
under each FS) are
presented in Figure 7. A clear and consistent effect of fatigue damage was observed on the
is discussed by the authorsininFigure
presented references
7. A [66,70].
clear and The PSD Welch
consistent effect estimates
of fatigue for Coupon
damage #3
was observed on
PSD Welch estimates in that certain resonant frequencies (in particular, those in the 3
under HS and various
the FSs (70 PSD estimates under the HS and 35 under each FS) are
3.6PSD
kHz Welch
range)estimates
shifted fromin that certain
higher resonant
to lower frequencies
frequencies (in particular,
as the damage those in the
progressed.
presented in Figure3–3.67. A
kHzclear andshifted
range) consistent
from effect
higheroftofatigue damage was
lower frequencies observed
as the damageon the
progressed.
PSD Welch estimates
Tablein2.that certain
Signal resonant
characteristics frequencies
and Welch (in particular,
PSD estimation those in
details for random the 3 testing [27].
vibration
Table 2. Signal characteristics and Welch PSD estimation details for random vibration testing [27].
3.6 kHz range) shifted Random
from higher to lower frequencies as the damage progressed.
Vibration Test with Acoustic Excitation
Signal Characteristics Random Vibration Test with Acoustic Excitation
Welch Parameters
Table 2. Signal
Sampling frequency 𝑓 characteristics and Welch
12,800 Hz
Signal PSD estimation
Characteristics details for random
Segment length vibration testingsamples
25,600
Welch Parameters [27].
Signal duration Sampling 60 s
frequency fs 12,800 Hz OverlapSegment length 90% samples
25,600
Random Vibration Test with Acoustic Excitation
Signal bandwidth 0–6400
Signal duration Hz Frequency
60 s resolution 𝛿𝑓
Overlap 1 Hz 90%
Signal Characteristics Welch Parameters
Signal bandwidth 0–6400 Hz Frequency resolution δ f 1 Hz
g frequency 𝑓 12,800 Hz Segment length 25,600 samples
al duration 60 s Overlap 90%
bandwidth 0–6400 Hz Frequency resolution 𝛿𝑓 1 Hz
2025, 9, x FOR PEER REVIEW
J. Compos. Sci. 2025, 9, 153
9 of 23
9 of 22

Figure 7. Experimental results: Fatigue effects on random vibration Welch-based PSD estimates for
Coupon #3 (Fatigue Stages, FS1–FS13).

Figure 7. Experimental results:

Figure
4.3. FDM Fatigue effects
Experimental
7. Results results:on random
Fatigue vibration
effects Welch-based
on random PSD estimates
vibration Welch-based PSDfor
estimates for
Coupon #3 (FatigueCoupon
Stages,#3FS1–FS13).
(Fatigue Stages, FS1–FS13).
Figure 8 illustrates a typical fatigue delamination propagation pattern predicted by
4.3.
the FDM(Section
FDM Results 3.1) for a thermoplastic coupon. Delamination was initiated at the free
4.3. FDM Results
edges ofFigure 8 illustrates
the coupon, a typical
driven fatigue
by high delamination propagation
through-the-thickness pattern
normal andpredicted by the shear
intralaminar
Figure 8 illustrates a
FDM (Sectiontypical fatigue delamination
3.1) for a thermoplastic propagation
coupon. pattern predicted by
stresses, and propagated towards the centerDelamination
of the coupon. wasThe
initiated at the freepattern
propagation edges was
the FDM (Sectionof3.1)the for a thermoplastic
coupon, driven by high coupon. Delamination
through-the-thickness was initiated
normal and at the free
intralaminar shear stresses,
not fully symmetric, which is attributed to the nature of fatigue—repeated loading in spe-
edges of the coupon, driven by high
and propagated through-the-thickness
towards the center of the coupon.normalTheand intralaminar
propagation shear
pattern was not fully
cific directions can create progressive damage in certain ply orientations before others, as
symmetric, which is attributed to the nature of fatigue—repeated
stresses, and propagated towards the center of the coupon. The propagation pattern was loading in specific direc-
the material model used here [60,61] is sensitive to element orientation, which can create
tionswhich
not fully symmetric, can create progressive
is attributed todamage
the naturein certain ply orientations before
of fatigue—repeated others,
loading in as the material
spe-
directional
model used differences
here [60,61]inis damage
sensitive evolution. The consecutive
to element orientation, Fatigue
which can createStates (FS) corre-
directional
cific directions can create progressive damage in certain ply orientations before others, as
sponded to the
differences successive
in damage fatigueThe
evolution. cycles—each
consecutive FS corresponded
Fatigue to 20,000 cycles.
States (FS) corresponded The sig-
to the
the material model used here [60,61] is sensitive to element orientation, which can create
nificance
successiveof fatigue
the fatigue delamination
cycles—each pattern lies
FS corresponded to in its ability
20,000 cycles.to accurately
The significancemodel
of thedelam-
directional differences in damage evolution. The consecutive Fatigue States (FS) corre-
ination
fatiguegrowth and capture
delamination patternthe
liescharacteristics
in its ability toof thermoplastic
accurately modelcoupon’s
delaminationfatigue behavior.
growth
sponded to the successive
and capture fatigue cycles—each
the characteristics FS corresponded
of thermoplastic to 20,000
coupon’s cycles.
fatigue The sig-
behavior.
nificance of the fatigue delamination pattern lies in its ability to accurately model delam-
ination growth and capture FS4
FS2 the characteristics
FS6 ofFS8
thermoplastic
FS10 coupon’s
FS12 fatigue behavior.

FS2 FS4 FS6 FS8 FS10 FS12

Figure 8. A typical fatigue delamination pattern predicted by the FDM for a thermoplastic coupon
Figure 8. A typical fatigue delamination pattern predicted by the FDM for a thermoplastic coupon
(FS: Fatigue States correspond to increasing fatigue cycles).
(FS: Fatigue States correspond to increasing fatigue cycles).

Figure 8. A typical
4.4.fatigue
RVM delamination
Modeling pattern predicted by the FDM for a thermoplastic coupon
(FS: Fatigue States correspond to increasing fatigue cycles).
A typical FE mesh of the coupon is shown in Figure 9. The coupon’s damping ratio
was calculated as 0.01 from Equation (6). The thermoplastic plies were modeled using
4.4. RVM Modeling
A typical FE mesh of the coupon is shown in Figure 9. The coupon’s damping ratio
ci. 2025, 9,J.xCompos. Sci. 2025,
FOR PEER 9, 153
REVIEW 10 of 23 10 of 22

J. Compos. Sci. 2025, 9, x FOR PEER REVIEW 10 of 23

4.4. RVM
J. Compos. Sci. 2025, 9, x FOR PEER REVIEW Modeling 10 of 23
solid 8-noded elements with a constant stress formulation (ELFORM 1). The orthotropic
A typical FE mesh of the coupon is shown in Figure 9. The coupon’s damping ratio
elastic materialwas
properties used for the elements are listed in Table 1. As boundary condi-
calculated as 0.01 from Equation (6). The thermoplastic plies were modeled using solid
solid
tions, only the8-noded 8-noded
nodes that elements
were with a with
in contact constant stress formulation
theformulation
elastic cords were (ELFORM 1). to
restricted The orthotropic
the y- elastic
solid elements with awith
constant stress (ELFORM 1). The1).
orthotropic
elastic8-noded
materialelements
properties a constant
used stress formulation
for the elements are listed in(ELFORM
Table 1. As The orthotropic
boundary condi-
axis. Velocity was measured
material
elastic material
at the
properties usedmiddle
propertiesforused and upper
the elements are face
listednodes.
for the elements in Table
are listed1.inAs boundary
Table conditions,
1. As boundary only
condi-
tions, only the nodes that were in contact with the elastic cords were restricted to the y-
the nodes
tions, onlythat were
thewas
nodes inthat
contact
werewith the elastic cords
in contact were restricted to the y-axis. to
Velocity
axis. Velocity measured at the middle with the elastic
and upper cords were
face nodes. restricted the y-
was
axis.measured at the
Velocity was middle and
measured upper
at the faceand
middle nodes.
upper face nodes.

Figure 9. A typical FE mesh of the coupon.

Figure 9. A typical FE mesh of the coupon.
Figure 9. A typical FE mesh
Figure 9. mesh of
of the
the coupon.
coupon.
Figure 10 shows the 10
Figure applied
shows white noisewhite
the applied excitation in the time
noise excitation in and frequency
the time (spec- (spec-
and frequency
tral) domains. tral) Figure
Figure 10
11 10 shows
shows the
the applied white noise excitation in the time and frequency (spectral)
Figure
domains. shows
Figure 11applied
the shows boundary
appliedthe
white conditions
noiseboundary
applied inand
the the
excitationconditions measurement
timeand
andthe
frequency (spec-
measurement
domains.
location in the tral) Figure
domains.
coupon. 11 shows the applied boundary conditions and the measurement
Figure 11 shows the applied boundary conditions and the measurement location
location in the coupon.
in the coupon.
location in the coupon.

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

(a) (b)
Figure
Figure 10. WhiteFigure 10.
noise10. White
White noise
noise
excitation excitation
excitation
in the in the
the time
inand
time (a) time (a)
(a) and
frequency frequency
frequency (spectral)
and(spectral) (spectral)domains
domains domains(b).
(b). (b).
Figure 10. White noise excitation in the time (a) and frequency (spectral) domains (b).

Figure 11. Boundary conditions, excitation region, and measurement point of the FE model.
Figure 11. Boundary conditions, excitation region, and measurement point of the FE model.
Figure 11. Boundary conditions, excitation region, and measurement point of the FE model.
Figure 11. Boundary conditions, excitation region, and measurement point of the FE model.
4.5. RVM Results
4.5. RVM Results
4.5. RVM Results4.5.1. Calibration of the RVM
4.5.1. Calibration of the RVM
4.5.1. Calibration ofAthe
stepwise
RVM validation and calibration process was employed for the RVM. The mod-
A stepwise
eling parameters validation and calibration
that were calibrated process
includedwas employed
the for the RVM.
mesh density, The mod-
the boundary
A stepwise validation
eling and calibration
parameters that were process was
calibrated employed
included the for the RVM.
mesh Thethe
density, mod-
boundary
eling parameters that were calibrated included the mesh density, the boundary condi-
pos. Sci. 2025, 9, x FOR PEER REVIEW 11 of 23

J. Compos. Sci. 2025, 9, 153 11 of 22

conditions, the measurement point, and the excitation area. To achieve this, in-house ex-
perimental4.5.
results
RVM from
Results[27] and laboratory-scale experimental results of a simplified test
were utilized.
4.5.1. Calibration of the RVM
A stepwise validation and calibration process was employed for the RVM. The model-
4.5.2. Parametric Frequency
ing parameters that Response Function
were calibrated (FRF)
included theAnalysis
mesh density, the boundary conditions,
the measurement
As a first step, parametricpoint, andanalysis
FRF the excitation
of the area.
healthyTo achieve
couponthis,
wasin-house
conducted experimental
and the
results from [27] and laboratory-scale experimental results of
FRF amplitude and phase considering the above parameters were compared. The param-a simplified test were utilized.
eters used4.5.2.
are listed in Table 3.
Parametric Frequency Response Function (FRF) Analysis
The FRF amplitude and phase for varying mesh densities, measurement locations,
As a first step, parametric FRF analysis of the healthy coupon was conducted and
boundary theconditions, and excitation intensities are compared in Figure 12.
FRF amplitude and phase considering the above parameters were compared. The
As the mesh density
parameters used are increased, the computed
listed in Table 3. results progressively converged, alt-
hough the differences between the FE meshes remained minor. Final convergence was
Table 3.
achieved with Parameter
the mesh of values used elements,
432,000 in the calibration
which of the
wasRVM.
then applied consistently in all
subsequent analyses. The comparison between free–free and elastic
Parameter Values cord support bound-
ary conditions revealed only minor differences in FRF characteristics, and the elastic cord
Measurement location Center, Upper right corner, Lower left corner
support was ultimately
Mesh density selected to align235,000
with the experimental
elements, setup, while
432,000 elements, fixedelements
2,080,000 bound-
Boundary
ary conditions conditions
affected the dynamic behaviorCord-support, Free, Clamped
of the coupon. Meanwhile, the location of
Excitation
the measurement intensity
point × 10−15 MPa,
and excitation 2intensity 79 MPa, 200
(expectedly) MPa
influenced the magnitude
intensity; yet the resonant frequencies (spectral peaks) remained unaffected. The dynam-
The FRF amplitude
ics remained approximately and
linear phasedifferent
under for varying mesh densities,
excitation measurement
intensities, as the shape locations,
did
boundary conditions, and excitation intensities are compared in
not change, and the measurement point and the boundary conditions could be replicated Figure 12.
As the mesh density increased, the computed results progressively converged, al-
in the FEM successfully.
though the differences between the FE meshes remained minor. Final convergence was
To ensure consistency with the experimental procedure, the measurement point was
achieved with the mesh of 432,000 elements, which was then applied consistently in all
positionedsubsequent
at the center of theThe
analyses. coupon, corresponding
comparison to the setup
between free–free usedcord
and elastic in the experiment.
support boundary
conditions revealed only minor differences in FRF characteristics, and the elastic cord
Table 3. Parameter
support values used in the
was ultimately calibration
selected of the
to align withRVM.
the experimental setup, while fixed bound-
ary conditions affected the dynamic behavior ofValues
Parameter the coupon. Meanwhile, the location of
the measurement point and excitation intensity (expectedly) influenced the magnitude
Measurement location Center, Upper right corner, Lower left corner
intensity; yet the resonant frequencies (spectral peaks) remained unaffected. The dynamics
Mesh density 235,000 elements, 432,000 elements, 2,080,000 elements
remained approximately linear under different excitation intensities, as the shape did not
Boundary conditions Cord-support, Free, Clamped
change, and the measurement point and the boundary conditions could be replicated in
Excitationthe
intensity 2
FEM successfully.
× 10−15 MPa, 79 MPa, 200 MPa

(a) (b)
Figure 12. Cont.
pos. Sci. 2025, 9, x FOR PEER REVIEW 12 of 23
J. Compos. Sci. 2025, 9, 153 12 of 22

(c) (d)

(e) (f)

(g)

Figure 12. Figure

FRF analysis
12. FRFusing different
analysis mesh densities:
using different (a) Magnitude
mesh densities: (mm/s
(a) Magnitude 2 ) and
2) and
(mm/s (b) (b)
Phase (°),(◦ ),
Phase
different boundary conditions: (c) Magnitude 2) and 2(d) ◦ ), at different measurement
different boundary conditions: (c) Magnitude (mm/s(mm/s ) and (d) Phase
Phase (°), at( different measurement
points: (e) Magnitude (mm/s2 ), (f) Phase (◦ ), and (g) RVM analysis.
points: (e) Magnitude (mm/s2), (f) Phase (°), and (g) RVM analysis.
To ensure consistency with the experimental procedure, the measurement point was
4.5.3. Validation of the
positioned RVM
at the (Healthy
center State) corresponding to the setup used in the experiment.
of the coupon,
Given the complexity of the random vibration response that the RVM aims to simu-
4.5.3. Validation of the RVM (Healthy State)
late, validating its accuracy is crucial. To this end, different comparison cases against ex-
Given were
perimental results the complexity of the reflecting
considered, random vibration response
different that thescenarios
vibration RVM aims forto simulate,
both
validating its accuracy is crucial. To this end, different comparison cases against experi-
healthy and damaged coupons. The validation had two primary objectives: (a) To confirm
mental results were considered, reflecting different vibration scenarios for both healthy
the model’s capacity to accurately represent the random vibration behavior of the ther-
and damaged coupons. The validation had two primary objectives: (a) To confirm the
moplastic model’s
coupons and (b)toto
capacity assess the
accurately model’s
represent theability
random to vibration
capture the effectofofthe
behavior damage on
thermoplas-
the coupon’s vibration behavior.
For the healthy state, two validation cases involving different random vibration-
based excitation and support conditions were considered.
In the first case, the random vibration test performed in [27] was considered. Figure
12 compares the numerical-based and experimental PSD estimates. A satisfactory agree-
 J. Compos. Sci. 2025, 9, 153 13 of 22

tic coupons and (b) to assess the model’s ability to capture the effect of damage on the
coupon’s vibration behavior.
J. Compos. Sci. 2025, 9, x FOR PEER REVIEW For the healthy state, two validation cases involving different random vibration-based 13 of 23
excitation and support conditions were considered.
In the first case, the random vibration test performed in [27] was considered. Figure 12
compares
captured three the numerical-based and experimental
of the resonant frequencies indicated PSD byestimates.
the peaksAinsatisfactory
the PSD Welch agreement
esti-
of theThe
mate. intensity
additionalof the response
peaks in theamplitude
experimentalwas PSDevident.
Welch Furthermore,
estimate may thebemodel captured
attributed to
three of the resonant frequencies indicated by the peaks
various factors and uncertainty, including unmodelled phenomena, complicated excita-in the PSD Welch estimate. The
additional
tion peaks in the
due to reflections of theexperimental
white noisePSD Welch
acoustic estimateinmay
excitation be attributed
the room during the to various
exper-
iments, inherent instrumentation effects, and noise in the measurement devices, and due
factors and uncertainty, including unmodelled phenomena, complicated excitation so
to reflections of the white noise acoustic excitation in the room during the experiments,
on.
inherent instrumentation
To ensure effects, and
that the deviation noise in
observed in the measurement
Figure 13 was indeed devices, dueandto so
theon.
experi-
mentalToparameters
ensure thatand the deviation
conditionsobserved in Figure
of the random 13 was indeed
vibration tests, andduenotto the experimental
because the FE
model contained misleading information, a second validation case using a simplifiedmodel
parameters and conditions of the random vibration tests, and not because the FE ran-
contained
dom vibration misleading
test was information,
considered. This a second validation
simplified case conducted
test was using a simplified random
on a cantilever
vibration
coupon. Intest
thiswas considered.
case, white noise Thiswassimplified
generated testby
was anconducted on a cantilever
electromechanical excitercoupon.
(LDS
In thisV406
Model case,with
white itsnoise was generated
controller, LDS COMET by anUSB
electromechanical
COM-200) and exciter applied(LDS to theModel
fixed V406
end
with its controller, LDS COMET USB COM-200) and applied
(tab) of the coupon. A laser vibrometer was used to measure the velocity at 3 points (the to the fixed end (tab) of the
coupon. A
midpoint andlaser vibrometer
2 points near thewasright
usedtab)to measure the velocity
on the upper surface at of
3 points (the midpoint
the coupon. A key dif-and
2 points near the right tab) on the upper surface of the coupon.
ference from the previous experiment is that non-acoustic excitation was used, and the A key difference from the
previous
tests were experiment
conducted in is athat non-acoustic
significantly excitation
larger room. A was used, and
schematic the tests were
illustration conducted
of the simpli-
in a significantly larger room. A schematic illustration of the
fied random vibration test and a photograph of the test rig are provided in Figure 14. Thesimplified random vibration
test and asignal
excitation photograph of the test
was recorded andrig are provided
applied with theinsameFigure 14. Theasexcitation
intensity signalThe
the FE model. was
recorded and applied with the same intensity as the FE model.
electromechanical exciter stimulated frequencies ranging from 0 to 2 kHz, with direct con- The electromechanical
exciter
tact stimulated
between frequencies
the exciter and the ranging
coupon. from 0 to 2 kHz,
Regarding the FEwithmodel,directthecontact
modelingbetween
process the
exciter and the coupon. Regarding the FE model, the modeling
remained consistent with the previous setup, using the same mesh (432,000 elements), the process remained consistent
withthermoplastic
same the previous setup, coupon, using
andthe same mesh
geometry (432,000
(Figure 4). Theelements), the samewere
only differences thermoplastic
the loca-
coupon, and geometry (Figure 4). The only differences were the location
tion of the excitation point (on the tab), the intensity of the excitation (−100 dB), the bound- of the excitation
point
ary (on the (clamped
conditions tab), the intensity
on the left of tab),
the excitation (−100 dB), the
and the measurement boundary
points. conditions
This simplified
(clamped
setup helped on isolate
the left and
tab),eliminate
and the measurement
factors, suchpoints. This simplified
as acoustic reflections,setup helped
ensuring isolate
a more
and eliminate factors, such as acoustic reflections,
controlled and accurate validation of the FE model’s performance. ensuring a more controlled and accurate
validation of the FE model’s performance.

Figure 13. Comparison of PSD Welch estimate of the random vibration test and the FE model.
Figure 13. Comparison of PSD Welch estimate of the random vibration test and the FE model.

The signal characteristics and the Welch PSD estimation details for the simplified
random vibration tests are summarized in Table 4. The signal bandwidth was shorter than
in the random vibration tests excited by white noise, as the goal was the examination of
lower-frequency bandwidth.
J. Compos. Sci. 2025, 9, x FOR PEER REVIEW 14 of 23
J. Compos. Sci. 2025, 9, 153 14 of 22

J. Compos. Sci. 2025, 9, x FOR PEER REVIEW 14 of 23

(a) (b)

Figure 14. (a) Sketch of the simplified random vibration test, (b) Photo of the test rig of the simplified
random
Figure 14.vibration test.
(a) Sketch of the simplified random vibration test, (b) Photo of the test rig of the simplified
random vibration test.
The signal characteristics and the Welch PSD estimation details for the simplified
random
Table vibration
4. Signal tests are summarized
characteristics and Welch PSD in Table 4. The
estimation signal
details forbandwidth was shorter
simplified random than
vibration
in the
testing. random vibration tests excited by white noise, as the goal was the examination of
(a) (b)
lower-frequency bandwidth.
Simplified Random Vibration Test
Signal Characteristics
Figure 14. (a) Sketch of the simplified random vibration test, (b) Photo ofWelch
Table 4. Signal characteristics and Welch PSD estimation details forParameters
the test rig of the simplified
simplified random vibration testing.
Sampling frequency 𝑓 random vibration test.
2000 Hz Segment length 4000 samples
Signal duration 10 s Simplified RandomOverlapVibration Test 90%
Table 4. Signal characteristics and Welch PSD estimation details for simplified random vibration
Signal bandwidth testing. 0–2000
Signal Hz Characteristics Frequency resolution 𝛿𝑓 Welch 1
Parameters Hz
Sampling frequency
Simplified f s Vibration
Random 2000Test
Hz Segment length 4000 samples
Figureduration
Signal
Signal Characteristics 15 presents the comparison
10 s Welchbetween theOverlap
ParametersFEM-based analytical 90%PSD and its
Sampling frequency 𝑓 Welch-based2000 Hzestimate
Signal bandwidth obtained from
Segment
0–2000 Hz the
lengthexperimental procedure.
4000 samples Evidently,
Frequency resolution δ f 1 Hzresonant
the
Signal duration 10 s Overlap 90%
frequencies, as well as the overall PSD, were quite accurately captured by the FE model.
Signal bandwidth 0–2000 Hz Frequency resolution 𝛿𝑓 1 Hz
This validated the FE model’s capability to properly represent the coupon dynamics.
Figure 15 presents the comparison between the FEM-based analytical PSD and its
Figure 15 presents the comparison between the FEM-based analytical PSD and its
Welch-based estimate obtained from the experimental procedure. Evidently, the resonant
Welch-based estimate obtained from the experimental procedure. Evidently, the resonant
frequencies, as well as the overall PSD, were quite accurately captured by the FE model.
frequencies, as well as the overall PSD, were quite accurately captured by the FE model.
This
This validated
validated the FEthe FE model’s
model’s capabilitycapability
to properly to properly
represent represent
the coupon the coupon dynamics.
dynamics.

Figure
Figure 15.
15. Random
RandomVibration
VibrationModule
Modulevalidation:
validation:FE-based
FE-basedanalytical
analyticalPSD
PSDofofthe
therandom
randomvibration
vibration
Figure 15. Random Vibration Module validation: FE-based analytical PSD of the random vibration
response
response
response versus
versus
versus its experimental
its experimental
its experimental Welch-based
Welch-based
Welch-based estimated counterpart.
estimated
estimated counterpart. counterpart.
J. Compos. Sci. 2025, 9, 153 15 of 22
J. Compos. Sci. 2025, 9, x FOR PEER REVIEW 15 of 23

4.6.
4.6. Numerical
Numerical Fatigue
Fatigue Damage
Damage Identifiability
Identifiability Assessment
Assessment
AsAs already
already mentioned,
mentioned, the damage
the damage identifiability
identifiability assessment
assessment of the numerical
of the numerical tool tool
was
was also
also based
based on the
on the experimental
experimental procedure
procedure and theand the database
database of our
of our recent recent
work [27] work [27]
which
which employed
employed various
various Fatigue
Fatigue States
States (FSs) and(FSs)
a moreand a more
complex complex
acoustic acoustic
excitation. The excitation.
complexity of the acoustic
The complexity excitation,excitation,
of the acoustic which was which
not only spatially
was not onlydependent
spatiallybutdependent
also but
included reflections
also included from the from
reflections surrounding lab walls, implies
the surrounding that implies
lab walls, some discrepancies be-
that some discrepancies
tween the experimental
between and simulated
the experimental results were
and simulated expected.
results Yet, viewed
were expected. Yet,inviewed
a positive
in a positive
light, this was intended to provide a sense of the robustness of the assessment results
light, this was intended to provide a sense of the robustness of the assessment results when
when such discrepancies—that are difficult to model precisely—are present. The assess-
such discrepancies—that are difficult to model precisely—are present. The assessment
ment results are then presented in Section “Assessment Results with Experimental Vali-
results are then presented in Section “Assessment Results with Experimental Validation”.
dation”.

Assessment Results with Experimental Validation

Assessment Results with Experimental Validation
In In Figure
Figure 16,16,
thethe FEM-based
FEM-based analytical
analytical PSDs
PSDs areare compared
compared to their
to their experimental Welch-
experimental
based counterparts
Welch-based counterparts forforthe
thethermoplastic coupon(Coupon
thermoplastic coupon (Coupon#3) #3) under
under different
different Fa- Fatigue
States
tigue (FSs).
States Even
(FSs). Even from
from thetheearly
earlybeginning
beginning (FS2), thedynamic
(FS2), the dynamicbehavior
behavior of of
thethe
cou-coupon was
pon was captured from the FEM-based spectrum, which confirmed the
captured from the FEM-based spectrum, which confirmed the usefulness the proposed usefulness of the
proposed
numericalnumerical
tool. As tool. As expected,
expected, in the
in the FatigueState
Fatigue State1010 (FS10),
(FS10), the
theeffects of of
effects damage
damage on the
onPSDs
the PSDs were intense.
were intense.

(a)

(b)

Figure 16. Random Vibration Module validation (a) between Healthy and Fatigue State 2 (FS2), and
(b) between Healthy and Fatigue State 10 (FS10): FE-based analytical PSDs of the random vibration
response versus their experimental Welch-based estimated counterparts.
 sive fatigue damage was shown through its effects on the PSD and, in part
the resonant frequencies. These are more evident in Figure 17, where the FE
J. Compos. Sci. 2025, 9, 153 16 of 22
trums for six FSs are depicted, clearly highlighting the model’s ability to m
of fatigue damage accumulation on the dynamics.
As demonstrated in the present section, the RVM’s ability to properly model progres-
The model successfully predicted the decrease in resonant frequencie
sive fatigue damage was shown through its effects on the PSD and, in particular, shifts
tigue
in thedamage accumulation.
resonant frequencies. These areConsistent
more evidentwith the17,experimental
in Figure results, th
where the FEM-based
spectrums for six FSs are depicted,
more pronounced at higher frequencies. clearly highlighting the model’s ability to model the
effects of fatigue damage accumulation on the dynamics.

Figure 17. The Random Vibration Module was used: ‘Predicted’ FEM-based PSDs of the random
Figure 17.
vibration The Random
response Vibration
for the Healthy Module
and various was used: ‘Predicted’ FEM-based PSD
Fatigue States.

vibration response
The model for the Healthy
successfully predicted and variousinFatigue
the decrease resonantStates.
frequencies caused by
fatigue damage accumulation. Consistent with the experimental results, the decrease was
moreDamage
pronounceddiagnosability
at higher frequencies. is based on dynamics-based Damage Metr
Damage diagnosability is based on dynamics-based Damage Metrics (DMs) that
characterize
characterize thethe ‘distance’
‘distance’ of the State
of the Healthy Healthy from anyState from
Fatigue any
State (FS)Fatigue State (FS) w
within a proper
dynamical feature-based
dynamical feature-based space. space.
Specifically,Specifically,
in this presentin thisseven
work, present work,
Damage seven D
Metrics
were employed based on the numerical vibration PSD spectrum: Three (Metrics 1 to 3)
were employed based on the numerical vibration PSD spectrum: Three (
previously used in [39], and four (Metrics 4 to 7) presently introduced. The DMs were
previously
dimensionless,used
exceptinfor[39],
DM5. andThesefour metrics(Metrics
are defined4asto 7) presently introduced.
follows:
dimensionless,
DM 1: except for DM5. These
metrics are  defined as follows:
log S FSi ( f i ) − S H ( f i ) dB
DM 1: × 100% (9)
log(S H ( f i )) dB

DM 2:
log
∑in=1 𝑆S FSi ( f𝑓i ) − S H 𝑆
( f i ) 𝑓 dB
 
𝑑𝐵
∑in=1 S FSi ( f i ) dB 100% (10)
log 𝑆 𝑓 𝑑𝐵
DM 3:
i
1 ∑ j=1 S FSi f j − S H f j


 
n
dB
DM 2: ∑ fc j dB
(11)
∑ i =1 S H f j


i =1

∑ 𝑆 𝑓 𝑆 𝑓 𝑑𝐵
∑ 𝑆 𝑓 𝑑𝐵

DM 3:

1∑ 𝑆 𝑓 𝑆 𝑓 𝑑𝐵
J. Compos. Sci. 2025, 9, 153 17 of 22

DM 4:
n
∑


S FSi − |S H | [dB] (12)
i =1

DM 5:
S H − S FSi
 
dB
× 100% (13)
S FSi dB

DM 6:
∑n S H ·S FSi
 
dB
= i =1


cos S H , S FSi (14)
∥S H ∥ SFSi dB

DM 7:
q
T
S H − S FSi C −1 S H − S FSi

 

dB
d S H , S FSi = q  (15)

T dB
S H − S FSi C −1 S H − S FSi


max

where
S FSi : FEM-based analytical PSD for the i − th FS;
S H : FEM-based analytical PSD for the Healthy State;
S FSi : absolute of the FEM-based analytical PSD for the i − th FS;
|S H |: absolute value of the FEM-based analytical PSD for the Healthy State,
f i : a specific frequency;
f c : the upper (max) frequency within the range of interest;


cos S H , S FSi : cosine of the angle between the two FEM-based analytical PSDs;
S FSi : l2 norm of the FEM-based analytical PSD;
d: normalized Mahalanobis distance between two health states;
S H − S FSi : difference between two FEM-based analytical PSDs;
C: estimated covariance matrix of the two FEM-based analytical PSDs (Healthy and corre-
sponding Fatigue State); and
C −1 : inverse of the estimated covariance matrix of the two FEM-based analytical PSDs
(Healthy and corresponding Fatigue State).
The values of the Damage Metrics (DM) computed from the FEM-based analytical
PSDs for six Fatigue States (FSs) are compared to those of the Healthy State (HS), as well as
to their experimentally obtained counterparts, in Figure 18. The DMs for all FSs exhibited
considerable differences from the Healthy State, with noticeable deviations starting from
FS2, clearly indicating the onset of fatigue damage. DM1, DM2, DM3, and DM5 showed a
consistent increase across the FSs, suggesting their suitability not only for detecting damage
but also for classifying its severity (extent). In contrast, DM4 displayed random variation,
making it unsuitable for damage classification. Numerical-based damage metrics tended to
have better behavior, due to a clear mode of damage and the separation of each Fatigue
State from one another.
 J. Compos. Sci. 2025, 9, x FOR PEER REVIEW 18 of 23
J. Compos. Sci. 2025, 9, 153 18 of 22

(a) (b)

(c) (d)

(e) (f)

(g)
Figure 18. The
Figure18. The Random
Random Vibration
VibrationModule
Modulewaswasused:
used: Fatigue
Fatiguedamage
damagediagnosability
diagnosability for
forvarious
various
Fatigue
Fatigue States via the Damage Metrics (a) DM1, (b) DM2, (c) DM3, (d) DM4, (e) DM5, (f) DM6,and
States via the Damage Metrics (a) DM1, (b) DM2, (c) DM3, (d) DM4, (e) DM5, (f) DM6, and
(g)
(g)DM7.
DM7.

5. Concluding Remarks
5. Concluding Remarks
In this work, we developed an integrated numerical model to simulate random-
In this work, we developed an integrated numerical model to simulate random-vi-
vibration fatigue damage detectability in thermoplastics. The model comprised two key
bration fatigue damage detectability in thermoplastics. The model comprised two key
modules: a fatigue damage simulation module and a random vibration simulation module.
modules: a fatigue damage simulation module and a random vibration simulation mod-
This numerical approach digitally replicates a previously established experimental process.
ule. This numerical approach digitally replicates a previously established experimental
Based on the numerical results, the following conclusions can be drawn:
process. Based on the numerical results, the following conclusions can be drawn:
• The Fatigue Delamination Module (FDM) accurately simulates the fatigue delamina-
 The Fatigue Delamination Module (FDM) accurately simulates the fatigue delamina-
tion initiation and propagation. Furthermore, FDM captures the fatigue behavior of
tion initiation and propagation. Furthermore, FDM captures the fatigue behavior of
thermoplastic coupons, proving how severe delamination is for the coupon;
thermoplastic coupons, proving how severe delamination is for the coupon;
J. Compos. Sci. 2025, 9, 153 19 of 22

• The Random Vibration Model (RVM) accurately simulates the vibration response
of both healthy and damaged composite coupons under various excitation condi-
tions. This was validated in two different cases using in-house experimental data.
Additionally, the RVM demonstrated similar sensitivity to damage as observed in the
experiments, showing a shift in the natural frequencies to the left;
• A parametric study using the FRF model revealed that the Finite Element (FE) mesh
density had a minimal effect, while the measurement location, boundary conditions,
and excitation intensity significantly influenced the FRF amplitude’s intensity;
• All damage metrics, using the FEM-based Power Spectral Density (PSD) spec-
trums, achieved highly effective damage detectability, reaching 100% performance.
Frequency-related damage metrics also showed a monotonic increase, making them
suitable for damage classification;
• The proposed numerical tool supports the design and optimization of a random
vibration-based Structural Health Monitoring (SHM) system for composite structures
through the following steps:
1. Predicting fatigue damage progression and residual strength as functions of the num-
ber of cycles.
2. Simulating the random vibration response of the healthy structure.
3. Introducing fatigue damage into the RVM and simulating the random vibration
response of the damaged structure.
4. Maximizing detection efficiency by comparing the FEM-based PSD spectrums of the
healthy and damaged states.

Author Contributions: Conceptualization, N.T., K.T. and S.F.; methodology, N.T., K.T. and S.F.;
software, N.T.; validation, N.T., K.T. and S.F.; formal analysis, N.T.; investigation, N.T., K.T. and S.F.;
resources, N.T.; data curation, N.T.; writing—original draft preparation, N.T.; writing—review and
editing, N.T., K.T. and S.F.; visualization, N.T.; supervision, K.T. and S.F.; project administration, K.T.
and S.F.; funding acquisition, K.T. and S.F. All authors have read and agreed to the published version
of the manuscript.

Funding: This research received no external funding.

Data Availability Statement: The data presented in this study is available on request from the
corresponding author (the data are not publicly available due to privacy).

Acknowledgments: The authors want to thank KVE Composites for manufacturing the test coupons.

Conflicts of Interest: The authors declare no conflicts of interest.

Abbreviations
The following abbreviations are used in this manuscript:

CFRP Carbon Fiber Reinforced Plastics

DM Damage Metric
LDV Laser Doppler Vibrometer
FDM Fatigue Delamination Module
FE Finite Element
FEM Finite Element Model
FRF Frequency Response Function
FS Fatigue State
PSD Power Spectral Density
RVM Random Vibration Module
SEM Scanning Electron Microscopy
SHM Structural Health Monitoring
J. Compos. Sci. 2025, 9, 153 20 of 22

References
1. Pantelakis, S.; Tserpes, K. (Eds.) Revolutionizing Aircraft Materials and Processes; Springer International Publishing: Cham,
Switzerland, 2020; ISBN 978-3-030-35345-2.
2. Structural Health Monitoring (SHM) in Aerospace Structures; Woodhead Publishing: Cambridge, UK; Elsevier Ltd.: Amsterdam,
The Netherlands, 2016; ISBN 978-0-08-100148-6.
3. Güemes, A.; Fernandez-Lopez, A.; Pozo, A.R.; Sierra-Pérez, J. Structural Health Monitoring for Advanced Composite Structures:
A Review. J. Compos. Sci. 2020, 4, 13. [CrossRef]
4. Zorman, A.; Slavič, J.; Boltežar, M. Vibration Fatigue by Spectral Methods—A Review with Open-Source Support. Mech. Syst.
Signal Process. 2023, 190, 110149. [CrossRef]
5. Bjørheim, F.; Siriwardane, S.C.; Pavlou, D. A Review of Fatigue Damage Detection and Measurement Techniques. Int. J. Fatigue
2022, 154, 106556. [CrossRef]
6. Withers, P.J.; Preuss, M. Fatigue and Damage in Structural Materials Studied by X-Ray Tomography. Annu. Rev. Mater. Res. 2012,
42, 81–103. [CrossRef]
7. Shah, C.; Bosse, S.; Zinn, C.; von Hehl, A. Optimization of Non-Destructive Damage Detection of Hidden Damages in Fiber
Metal Laminates Using X-Ray Tomography and Machine Learning Algorithms. In Proceedings of the Advances in System-Integrated
Intelligence; Valle, M., Lehmhus, D., Gianoglio, C., Ragusa, E., Seminara, L., Bosse, S., Ibrahim, A., Thoben, K.-D., Eds.; Springer
International Publishing: Cham, Switzerland, 2023; pp. 387–402.
8. Noorsuhada, M.N. An Overview on Fatigue Damage Assessment of Reinforced Concrete Structures with the Aid of Acoustic
Emission Technique. Constr. Build. Mater. 2016, 112, 424–439. [CrossRef]
9. Dia, A.; Dieng, L.; Gaillet, L.; Gning, P.B. Damage Detection of a Hybrid Composite Laminate Aluminum/Glass under Quasi-Static
and Fatigue Loadings by Acoustic Emission Technique. Heliyon 2019, 5, e01414. [CrossRef] [PubMed]
10. Guo, Y.-E.; Shang, D.-G.; Zuo, L.-X.; Qu, L.-F.; Hou, G.; Cai, D.; Jin, T.; Yin, X. Identification of Fatigue Damage Modes for
Carbon Fiber/Epoxy Composites Using Acoustic Emission Monitoring under Fully Reversed Loading. Polym. Compos. 2022, 43,
3371–3385. [CrossRef]
11. Zuluaga-Ramírez, P.; Frövel, M.; Belenguer, T.; Salazar, F. Non Contact Inspection of the Fatigue Damage State of Carbon Fiber
Reinforced Polymer by Optical Surface Roughness Measurements. NDT E Int. 2015, 70, 22–28. [CrossRef]
12. Wong, L.; Chowdhury, N.; Wang, J.; Chiu, W.K.; Kodikara, J. Fatigue Damage Monitoring of a Composite Step Lap Joint Using
Distributed Optical Fibre Sensors. Materials 2016, 9, 374. [CrossRef]
13. Khan, S.U.; Munir, A.; Hussain, R.; Kim, J.-K. Fatigue Damage Behaviors of Carbon Fiber-Reinforced Epoxy Composites
Containing Nanoclay. Compos. Sci. Technol. 2010, 70, 2077–2085. [CrossRef]
14. Ansari, M.T.A.; Singh, K.K.; Azam, M.S. Fatigue Damage Analysis of Fiber-Reinforced Polymer Composites—A Review. J. Reinf.
Plast. Compos. 2018, 37, 636–654. [CrossRef]
15. Limongelli, M.P.; Manoach, E.; Quqa, S.; Giordano, P.F.; Bhowmik, B.; Pakrashi, V.; Cigada, A. Vibration Response-Based Damage
Detection. In Structural Health Monitoring Damage Detection Systems for Aerospace; Sause, M.G.R., Jasiūnienė, E., Eds.; Springer
Aerospace Technology; Springer International Publishing: Cham, Switzerland, 2021; pp. 133–173, ISBN 978-3-030-72191-6.
16. Bull, M.K.; Norton, M.P. On the Hydrodynamic and Acoustic Wall Pressure Fluctuations in Turbulent Pipe Flow Due to a 90◦
Mitred Bend. J. Sound Vib. 1981, 76, 561–586. [CrossRef]
17. Fan, Z.; Jiang, Y.; Zhang, S.; Chen, X. Experimental Research on Vibration Fatigue of CFRP and Its Influence Factors Based on
Vibration Testing. Shock. Vib. 2017, 2017, e1241623. [CrossRef]
18. Bachoo, R.; Elishakoff, I. Random Vibrations of Laminated Planar Frames. J. Sound Vib. 2023, 562, 117860. [CrossRef]
19. Proso, U.; Slavič, J.; Boltežar, M. Vibration-Fatigue Damage Accumulation for Structural Dynamics with Non-Linearities. Int. J.
Mech. Sci. 2016, 106, 72–77. [CrossRef]
20. Andreaus, U.; Baragatti, P. Experimental Damage Detection of Cracked Beams by Using Nonlinear Characteristics of Forced
Response. Mech. Syst. Signal Process. 2012, 31, 382–404. [CrossRef]
21. Kim, H.-Y. Vibration-based damage identification using reconstructed frfs in composite structures. J. Sound Vib. 2003, 259,
1131–1146. [CrossRef]
22. Kessler, S.S.; Spearing, S.M.; Atalla, M.J.; Cesnik, C.E.S.; Soutis, C. Damage Detection in Composite Materials Using Frequency
Response Methods. Compos. Part B Eng. 2002, 33, 87–95. [CrossRef]
23. Mei, H.; Migot, A.; Haider, M.F.; Joseph, R.; Bhuiyan, M.Y.; Giurgiutiu, V. Vibration-Based In-Situ Detection and Quantification of
Delamination in Composite Plates. Sensors 2019, 19, 1734. [CrossRef]
24. Samareh-Mousavi, S.S.; Chen, X.; McGugan, M.; Semenov, S.; Berring, P.; Branner, K.; Ludwig, N. Monitoring Fatigue Delami-
nation Growth in a Wind Turbine Blade Using Passive Thermography and Acoustic Emission. Struct. Health Monit. 2024, 23,
2906–2921. [CrossRef]
25. Caminero, M.A.; Pavlopoulou, S.; Lopez-Pedrosa, M.; Nicolaisson, B.G.; Pinna, C.; Soutis, C. Analysis of Adhesively Bonded
Repairs in Composites: Damage Detection and Prognosis. Compos. Struct. 2013, 95, 500–517. [CrossRef]
J. Compos. Sci. 2025, 9, 153 21 of 22

26. Patil, S.; Mallikarjuna Reddy, D. Impact Damage Assessment in Carbon Fiber Reinforced Composite Using Vibration-Based New
Damage Index and Ultrasonic C-Scanning Method. Structures 2020, 28, 638–650. [CrossRef]
27. Tsivouraki, N.; Fassois, S.; Tserpes, K. On the Random Vibration-Based Progressive Fatigue Damage Detection and Classification
for Thermoplastic Coupons under Population and Operational Uncertainty. Compos. Struct. 2024, 349, 118524. [CrossRef]
28. Giagopoulos, D.; Arailopoulos, A.; Dertimanis, V.; Papadimitriou, C.; Chatzi, E.; Grompanopoulos, K. Structural Health
Monitoring and Fatigue Damage Estimation Using Vibration Measurements and Finite Element Model Updating. Struct. Health
Monit. 2019, 18, 1189–1206. [CrossRef]
29. Sahu, S.K.; Das, P. Experimental and Numerical Studies on Vibration of Laminated Composite Beam with Transverse Multiple
Cracks. Mech. Syst. Signal Process. 2020, 135, 106398. [CrossRef]
30. Moon, T.-C.; Kim, H.-Y.; Hwang, W. Natural-Frequency Reduction Model for Matrix-Dominated Fatigue Damage of Composite
Laminates. Compos. Struct. 2003, 62, 19–26. [CrossRef]
31. Valašková, V.; Melcer, J. Bridge Natural Frequencies, Numerical Solution versus Experiment. Appl. Sci. 2022, 12, 1765. [CrossRef]
32. Medeiros, R.; Sartorato, M.; Ribeiro, M.; Vandepitte, D.; Tita, V. Numerical and Experimental Analyses about SHM Metrics Using
Piezoelectric Materials. In Proceedings of the International Conference on Noise and Vibration Engineering—ISMA, Leuven,
Belgium, 17–19 September 2012; pp. 3285–3300.
33. Gomes, G.F.; Mendéz, Y.A.D.; da Cunha, S.S.; Ancelotti, A.C. A Numerical–Experimental Study for Structural Damage Detection
in CFRP Plates Using Remote Vibration Measurements. J. Civ. Struct. Health Monit. 2018, 8, 33–47. [CrossRef]
34. de Menezes, V.G.S.; Souza, G.S.C.; Vandepitte, D.; Tita, V.; de Medeiros, R. Defect and Damage Detection in Filament Wound
Carbon Composite Cylinders: A New Numerical-Experimental Methodology Based on Vibrational Analyses. Compos. Struct.
2021, 276, 114548. [CrossRef]
35. Dubey, A.; Denis, V.; Serra, R. Localization and Quantification of Damage by Frequency Based Methods: Numerical and
Experimental Applications on Bending Vibration Beams. In Proceedings of the Surveillance, Vishno and AVE Conferences,
INSA-Lyon, Université de Lyon, Lyon, France, 8–10 July 2019.
36. Li, X.; Qu, P.; Kong, H.; Zhu, Y.; Hua, C.; Guo, A.; Wang, S. Multi-Scale Numerical Analysis of Damage Modes in 3D Stitched
Composites. Int. J. Mech. Sci. 2024, 266, 108983. [CrossRef]
37. Sharma, J. Theoretical and Experimental Modal Analysis of Beam. In Engineering Vibration, Communication and Information
Processing; Lecture Notes in Electrical Engineering; Springer International Publishing: Cham, Switzerland, 2019; pp. 177–186,
ISBN 9789811316418.
38. Karczub, D.G.; Fagerlund, A.C. Dynamic Stress Predictions of Acoustic-Induced Pipe Vibration Failures. In Proceedings of the
ASME 2005 24th International Conference on Offshore Mechanics and Arctic Engineering, Halkidiki, Greece, 12–17 June 2005;
American Society of Mechanical Engineers Digital Collection; pp. 637–640.
39. Sartorato, M.; de Medeiros, R.; Vandepitte, D.; Tita, V. Computational Model for Supporting SHM Systems Design: Damage
Identification via Numerical Analyses. Mech. Syst. Signal Process. 2017, 84, 445–461. [CrossRef]
40. Xia, J.; Li, G.; Li, B.; Cheng, L.; Zhou, B. Fatigue Life Prediction of Package-on-Package Stacking Assembly under Random
Vibration Loading. Microelectron. Reliab. 2017, 71, 111–118. [CrossRef]
41. Barraza-Contreras, J.M.; Piña-Monarrez, M.R.; Molina, A.; Torres-Villaseñor, R.C. Random Vibration Fatigue Analysis Using a
Nonlinear Cumulative Damage Model. Appl. Sci. 2022, 12, 4310. [CrossRef]
42. Liu, F.; Lu, Y.; Wang, Z.; Zhang, Z. Numerical Simulation and Fatigue Life Estimation of BGA Packages under Random Vibration
Loading. Microelectron. Reliab. 2015, 55, 2777–2785. [CrossRef]
43. Gao, D.; Yao, W.; Wen, W.; Huang, J. A Multiaxial Fatigue Life Prediction Method for Metallic Material Under Combined Random
Vibration Loading and Mean Stress Loading in the Frequency Domain. Int. J. Fatigue 2021, 148, 106235. [CrossRef]
44. Yam, L.H.; Yan, Y.J.; Jiang, J.S. Vibration-Based Damage Detection for Composite Structures Using Wavelet Transform and Neural
Network Identification. Compos. Struct. 2003, 60, 403–412. [CrossRef]
45. Bayissa, W.L.; Haritos, N.; Thelandersson, S. Vibration-Based Structural Damage Identification Using Wavelet Transform. Mech.
Syst. Signal Process. 2008, 22, 1194–1215. [CrossRef]
46. Silik, A.; Noori, M.; Altabey, W.A.; Ghiasi, R.; Wu, Z. Comparative Analysis of Wavelet Transform for Time-Frequency Analysis
and Transient Localization in Structural Health Monitoring. Struct. Durab. Health Monit. 2021, 15, 1–22. [CrossRef]
47. Khan, A.; Ko, D.-K.; Lim, S.C.; Kim, H.S. Structural Vibration-Based Classification and Prediction of Delamination in Smart
Composite Laminates Using Deep Learning Neural Network. Compos. Part B Eng. 2019, 161, 586–594. [CrossRef]
48. Jayasundara, N.; Thambiratnam, D.P.; Chan, T.H.T.; Nguyen, A. Damage Detection and Quantification in Deck Type Arch Bridges
Using Vibration Based Methods and Artificial Neural Networks. Eng. Fail. Anal. 2020, 109, 104265. [CrossRef]
49. Liang, Z.; Ramakrishnan, K.R.; Ng, C.-T.; Zhang, Z.; Fu, J. Vibration-Based Prediction of Residual Fatigue Life for Composite
Laminates through Frequency Measurements. Compos. Struct. 2024, 329, 117771. [CrossRef]
50. Capponi, L.; Česnik, M.; Slavič, J.; Cianetti, F.; Boltežar, M. Non-Stationarity Index in Vibration Fatigue: Theoretical and
Experimental Research. Int. J. Fatigue 2017, 104, 221–230. [CrossRef]
J. Compos. Sci. 2025, 9, 153 22 of 22

51. Palmieri, M.; Česnik, M.; Slavič, J.; Cianetti, F.; Boltežar, M. Non-Gaussianity and Non-Stationarity in Vibration Fatigue. Int. J.
Fatigue 2017, 97, 9–19. [CrossRef]
52. Fassois, S.D.; Sakellariou, J.S. Statistical Time Series Methods for SHM. In Encyclopedia of Structural Health Monitoring; John Wiley
& Sons, Ltd.: Hoboken, NJ, USA, 2009; ISBN 978-0-470-06162-6.
53. Hios, J.D.; Fassois, S.D. Statistical Damage Detection in a Smart Structure under Different Temperatures via Vibration Testing: A
Global Model Based Approach. Key Eng. Mater. 2009, 413–414, 261–268. [CrossRef]
54. Kopsaftopoulos, F.P.; Fassois, S.D. Vibration Based Health Monitoring for a Lightweight Truss Structure: Experimental Assessment
of Several Statistical Time Series Methods. Mech. Syst. Signal Process. 2010, 24, 1977–1997. [CrossRef]
55. Vamvoudakis-Stefanou, K.J.; Sakellariou, J.S.; Fassois, S.D. Vibration-Based Damage Detection for a Population of Nominally
Identical Structures: Unsupervised Multiple Model (MM) Statistical Time Series Type Methods. Mech. Syst. Signal Process. 2018,
111, 149–171. [CrossRef]
56. Panagiotopoulos, A.; Dmitri, T.; Spilios, F.D. Damage Detection on the Blade of an Operating Wind Turbine via a Single Vibration
Sensor and Statistical Time Series Methods: Exploring the Performance Limits of Robust Methods. Struct. Health Monit. 2022, 22,
433–448. [CrossRef]
57. Kakolyris, G.K.; Fassois, S.D.; Kefal, A.; Belur, M.Y. Random vibration based robust structural health monitoring for a composite
wing-shaped structure under temperature uncertainty: An exploratory study. In Proceedings of the International Conference on
Noise and Vibration Engineering (ISMA 2024), Leuven, Belgium, 9–11 September 2024.
58. Saramantas, I.E.; Konis, P.E.; Kriatsiotis, I.M.; Ofir, Y.; Kressel, I.; Spiliotopoulos, P.E.; Fera, F.T.; Sakellariou, J.S.; Fassois, S.D.;
Giannopoulos, F.; et al. An Exploratory Study on Data-Driven Vibration Based Damage Detection and Characterization for a
Population of Composite Aerostructures. J. Phys. Conf. Ser. 2024, 2647, 182029. [CrossRef]
59. Robinson, P.; Galvanetto, U.; Tumino, D.; Bellucci, G.; Violeau, D. Numerical Simulation of Fatigue-Driven Delamination Using
Interface Elements. Int. J. Numer. Meth. Engng. 2005, 63, 1824–1848. [CrossRef]
60. Sioutis, I.; Tserpes, K. A Mixed-Mode Fatigue Crack Growth Model for Co-Consolidated Thermoplastic Joints. Int. J. Fatigue 2023,
173, 107682. [CrossRef]
61. Tsivouraki, N.; Tserpes, K.; Sioutis, I. Modelling of Fatigue Delamination Growth and Prediction of Residual Tensile Strength of
Thermoplastic Coupons. Materials 2024, 17, 362. [CrossRef] [PubMed]
62. Wijker, J. Random Vibrations in Spacecraft Structures Design; Solid Mechanics and Its Applications; Springer: Dordrecht, The
Netherlands, 2009; Volume 165, ISBN 978-90-481-2727-6.
63. Russell, A.; Street, K. Predicting Interlaminar Fatigue Crack Growth Rates in Compressively Loaded Laminates. In Compos-
ite Materials: Fatigue and Fracture, Second Volume; ASTM International: West Conshohocken, PA, USA, 1989; pp. 162–178,
ISBN 978-0-8031-1190-5.
64. Sound Pressure. Available online: https://www.engineeringtoolbox.com/sound-pressure-d_711.html (accessed on 15 July 2024).
65. Vibration Converter. Available online: https://strutt.arup.com/help/Vibration/VibrationConvert.htm (accessed on 15 July 2024).
66. Tsivouraki, N.; Tserpes, K.; Fassois, S. Random Vibration-Based Progressive Fatigue Damage Monitoring in Thermoplastic
Coupons: A Preliminary Investigation. J. Phys. Conf. Ser. 2024, 2692, 012033. [CrossRef]
67. Toray Cetex. Datasheet of TC 1225 LM PAEK; Toray Cetex: Morgan Hill, CA, USA, 2023; p. 5.
68. Luinge, H.; Howell, D.D.; Straetker, J. Low Melt Paek (LM PAEK) Prepreg for Improved Processing of Thermoplastic Composites.
In Proceedings of the SAMPE 2019, Charlotte, NC, USA, 11 April 2019; SAMPE.
69. Miller, S.G.; Heimann, P.J.; Miller, A.J.; Scheiman, D.A.; McCorkle, L.S. Manufacturing and Mechanical Testing of TC1225/LM-PAEK
and TC1200/PEEK Thermoplastic Composite Panels; National Aeronautics and Space Administration, Glenn Research Center:
Cleveland, OH, USA, 2023.
70. Binetruy, C.; Jacquemin, F. Vibration-based progressive fatigue damage detection and classification for thermoplastic coupons
under population and operational uncertainty. In Proceedings of the 21st European Conference on Composite Materials, Nantes,
France, 2–5 July 2024; Nantes Université: Nantes, France, 2024; Volume 3, pp. 604–611.

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.