sensors

Review
Recent Advances in Active Infrared Thermography for
Non-Destructive Testing of Aerospace Components
Francesco Ciampa * ID
, Pooya Mahmoodi, Fulvio Pinto and Michele Meo
Materials and Structures Centre (MAST), Department of Mechanical Engineering, University of Bath,
Bath BA2 7AY, UK; [email protected] (P.M.); [email protected] (F.P.); [email protected] (M.M.)
* Correspondence: [email protected]; Tel.: +44-(0)1125386383

Received: 18 January 2018; Accepted: 14 February 2018; Published: 16 February 2018

Abstract: Active infrared thermography is a fast and accurate non-destructive evaluation technique
that is of particular relevance to the aerospace industry for the inspection of aircraft and helicopters’
primary and secondary structures, aero-engine parts, spacecraft components and its subsystems.
This review provides an exhaustive summary of most recent active thermographic methods used
for aerospace applications according to their physical principle and thermal excitation sources.
Besides traditional optically stimulated thermography, which uses external optical radiation such
as flashes, heaters and laser systems, novel hybrid thermographic techniques are also investigated.
These include ultrasonic stimulated thermography, which uses ultrasonic waves and the local damage
resonance effect to enhance the reliability and sensitivity to micro-cracks, eddy current stimulated
thermography, which uses cost-effective eddy current excitation to generate induction heating, and
microwave thermography, which uses electromagnetic radiation at the microwave frequency bands
to provide rapid detection of cracks and delamination. All these techniques are here analysed
and numerous examples are provided for different damage scenarios and aerospace components
in order to identify the strength and limitations of each thermographic technique. Moreover,
alternative strategies to current external thermal excitation sources, here named as material-based
thermography methods, are examined in this paper. These novel thermographic techniques rely
on thermoresistive internal heating and offer a fast, low power, accurate and reliable assessment of
damage in aerospace composites.

Keywords: non-destructive evaluation; infrared thermography; material damage; composite

materials; smart materials; aerospace structures

1. Infrared Thermography for Aerospace Applications

Infrared thermography (IRT) is a rapid and accurate non-destructive evaluation (NDE) technique
that is widely used for the inspection of large aerospace components such as aircraft and helicopters’
primary and secondary structures, aero-engine parts, spacecraft components and its subsystems.
IRT is generally applied to different materials including aluminium, composites and hybrid fibre
metal laminates such a glass laminate aluminium reinforced epoxy (GLARE) and carbon reinforced
aluminium laminate (CARALL) [1]. Whilst aluminium is today the primary material used in the
aerospace industry, composite laminates, such as carbon fibre reinforced polymers (CFRP) and glass
fibre reinforced polymers (GFRP), have been increasingly used due their excellent thermal-mechanical
properties such as low weight, high specific strength and stiffness, corrosion resistant and thermal
insulation [2–4]. It is estimated that composite components constitute to about 50% by weight of the
current generation of aircraft and are used for engine casings, wing sections, tail plane, fuselage and
control surfaces [5]. Typical damage scenarios in aerospace components include corrosion cracking,
fatigue cracks, barely visible impact damage (BVID), delaminations, disbonding, matrix/fibre cracking,

Sensors 2018, 18, 609; doi:10.3390/s18020609 www.mdpi.com/journal/sensors

Sensors 2018, 18, 609 2 of 37

Sensors 2018, 18, x 2 of 37

voids and core crushing. These types of material defects may occur at any stage of the component’s
fuselage
cycle and control
life, from surfaces 5.toTypical
manufacturing damage
in-service andscenarios in aerospace
maintenance components
operations include
[6]. IRT has corrosion
proven to be an
cracking, fatigue cracks, barely visible impact damage (BVID), delaminations,
effective tool to detect and, in many cases, quantify these material flaws. The most common materialsdisbonding,
matrix/fibre cracking, voids and core crushing. These types of material defects may occur at any stage
and types of damage in aerospace components that can be monitored with IRT are reported in Table 1
of the component’s cycle life, from manufacturing to in-service and maintenance operations 6. IRT
and illustrated in Figure 1 [7–11].
has proven to be an effective tool to detect and, in many cases, quantify these material flaws. The
most common
Table 1. Mostmaterials and types
common materials oftypes
and damage in aerospace
of damage components
used in aerospace that can that
components be monitored with
are monitored
IRTwith
are IRT.
reported in Table 1 and illustrated in Figure 1 7.

Table 1. Materials
Most common materials and types of damage used inTypes
Applications aerospace components
of Damage that
Monitored areIRT
with
monitored with IRT.
Delamination
Materials Applications Types of Damage
Impact damage and Monitored
BVID with IRT
Aircraft and spacecraft primary Delamination
Interlaminar disbond
and secondary structures
Impact damage and BVID
Porosity
Aircraft and spacecraft primary
Interlaminar disbond
and secondary structures Fibre/Matrix Cracking
Porosity
Thermal stress cracking
Fibre/Matrix Cracking
Composite laminates (e.g., CFRP Delamination
Thermal between ceramic thermal
stress cracking
Composite
and GFRP)laminates composites Jet engine turbine blades
and hybrid(e.g., barrier coating and the composite
(e.g.,and
GLARE andand
CARALL) Jet engine turbine blades Delamination between ceramic thermal
CFRP GFRP) substrate
barrier coating and the composite
hybrid composites (e.g., Skin-to-core debonding
substrate
GLARE and CARALL) Facesheetdebonding
delamination
Skin-to-core
Facesheet delamination
Honeycomb cell wall damage
Honeycomb panels
Honeycomb cell wall damage
Absence of adhesive
Honeycomb panels
Absence of adhesive
Water ingress
Water ingress
Core crushing
Core crushing
Fatigue cracks parallel and
Fatigue cracks parallel and perpendicular
perpendicular to the surface
to the surface
Aircraft and spacecraft
Aircraft primary
and spacecraft primary PoresPores and voids
and voids
and secondary structures,
and secondary corecore for Corrosion
structures,
Metals
Metals Corrosion
honeycombsand
for honeycombs and electromagnetic
wave shields Material
Material inclusions
inclusions
electromagnetic wave shields
Welded sheets
Welded without
sheets metal
without diffusion
metal diffusion
(“kissing
(“kissing bonds”)
bonds”)

(a)

(b)

Figure 1. Cont.
 Sensors 2018, 18, x 3 of 37
Sensors 2018, 18, 609 3 of 37

Figure 1. Cont.

(c)

(d)
Figure 1. Illustration of typical material defects monitored by IRT for composite aircraft and spacecraft
Figure 1. Illustration of typical material defects monitored by IRT for composite aircraft and spacecraft
structures (a); jet engine turbine blades (b); honeycomb panels (c) and metallic aircraft and spacecraft
structures (a); jet engine turbine blades (b); honeycomb panels (c) and metallic aircraft and spacecraft
components (d).
components (d).
IRT is generally classified as “passive” (stationary) and “active” (non-stationary) thermography
12.IRT
Passive IRT is typically
is generally classifiedused for materials
as “passive” that areand
(stationary) not“active”
in thermal equilibrium and
(non-stationary) have a different
thermography [12,13].
temperature
Passive IRT is from the environment
typically in whichthat
used for materials theyare
operate.
not inPassive
thermal thermography
equilibrium isand suitable
have for cyclic
a different
loading asfrom
temperature the resulting hysteretic
the environment heat allows
in which for time-dependent
they operate. temperature
Passive thermography variations
is suitable for that
cycliccan be
loading
monitored with an infrared (IR) camera. As an example from Montesano
as the resulting hysteretic heat allows for time-dependent temperature variations that can be monitored et al. 14, passive
thermography
with an infrared (IR) can be used for
camera. As an theexample
detectionfrom
of water ingress et
Montesano just
al. after
[14], aircraft
passive landing due to acan
thermography
significant temperature difference between the aircraft and water.
be used for the detection of water ingress just after aircraft landing due to a significant temperature
Active
difference IRT, on
between theaircraft
the other hand, employs external excitation sources such as optical radiation (e.g.,
and water.
halogen heat lamps and laser), electromagnetic stimulation (induced eddy currents and microwaves),
Active IRT, on the other hand, employs external excitation sources such as optical radiation (e.g.,
mechanical ultrasonic waves and material enabled features in composites in order to generate heat
halogen heat lamps and laser), electromagnetic stimulation (induced eddy currents and microwaves),
in the component under inspection 15. In Active IRT, temperature gradients responses at the medium
mechanical ultrasonic waves and material enabled features in composites in order to generate heat in
surface are recorded and analysed by an IR camera to provide information about the integrity of the
the component under inspection [15,16]. In Active IRT, temperature gradients responses at the medium
structure. Indeed, as thermal waves flow inside the sample by diffusion, the heat diffusion rate over
surface are recorded
a material defect will anddiffer
analysed
from by thean IR camera to
surrounding provide
area and the information about the
resulting thermal integrity
response canof bethe
structure.
used to detect and quantify material damage 17. For this reason, active IRT has very competitive costs a
Indeed, as thermal waves flow inside the sample by diffusion, the heat diffusion rate over
material
compareddefect will differ
to more from theNDE
sophisticated surrounding
equipment area and
such as the resulting
ultrasonic thermal
phased arrayresponse
and X-ray cansystem
be used
to (CT-Scan).
detect andLiterature
quantify material damage [17]. For this reason, active IRT has very competitive
offers a vast review and classification of active IRT for different application costs
compared
fields and to types
more of sophisticated
materials/damageNDE equipment
18. However, such as ultrasonic
a deep phased array
review including and X-rayrecent
the principles, system
(CT-Scan).
developments,Literature offers
strengths anda vast reviewofand
limitations classification
active of activespecifically
IRT thermography IRT for different application
for the aerospace
fields and types
industry is stillofmissing
materials/damage
and constitutes [18,19]. However,
the main focusaofdeep
this review including
work. Table the principles,
2 summarises the mainrecent
developments,
active IRT methodsstrengths usedandin limitations of activeaccording
aerospace industry IRT thermography specifically
to their physical principle forand
the thermal
aerospace
sources,iswhich
industry will be analysed
still missing in this review.
and constitutes the main focus of this work. Table 2 summarises the main
active IRT methods used in aerospace industry according to their physical principle and thermal
sources, which will be analysed in this review.
Sensors 2018, 18, 609 4 of 37

Table 2. Summary of active IRT methods used for aerospace applications.

Physical Principle Thermal Source Active IRT Terminology Abbreviation

Pulsed Thermography [20]
PT
(or Flash Thermography)
Lock-in Thermography [21]
Optical flash, lamp and (or Amplitude Modulated LIT
electrical heaters Thermography)
Optically Stimulated
Optical Radiation Step-Heating Thermography [22]
Thermography (OST) SHT and LPT
and Long Pulse Thermography [23]
Frequency Modulated
FMT
Thermography [24]
Laser-spot Thermography [25] and
Optical laser LST and LLT
Laser-line Thermography [26]

Acoustic/Ultrasonic horn, Ultrasonic Stimulated Thermography [27] (or Thermosonics,

Acoustic/Ultrasonic UST
piezo-ceramic sensors, Vibro-thermography and Sonic IR Thermography)
Wave Propagation
air-coupled transducers Nonlinear Ultrasonic Stimulated Thermography [28] NUST
Electromagnetic Radiation for Induced eddy currents Eddy Current Stimulated Thermography [29] ECST
dielectric materials Microwaves Microwave Thermography [30] MWT
Electrical Resistance Change
Electrical current applied to Direct Material-based
Method (ERCM) coupled to
carbon fibres Thermography (DMT)
thermography [31]
Material enabled Electrical current applied to embedded
Metal-based Thermography [32] MT
thermo-resistive radiation for steel wires
composite materials Electrical current applied to embedded Carbon Nanotubes-based
Indirect Material-based
CNTT
carbon nanotubes Thermography (IMT) Thermography [33]
Electrical current applied to embedded Shape Memory Alloys-based
SMArT
shape memory alloys wires Thermography [34]
 Depending on the type of external optical heat source used to generate thermal waves, the most
common methods used in optically stimulated thermography (OST) for aerospace applications are
pulsed thermography (PT), lock-in thermography (LIT), step heating thermography (SHT) and long
pulse thermography (LPT), frequency modulated thermography (FMT), laser-spot thermography
Sensors and18,
(LST) 2018, 609
laser-line thermography (LLT) 35. Figure 2 shows the experimental setup of OST 5 offor
37

different optical excitation sources.

Data obtained
2. Optically by anThermography
Stimulated IR camera in OST are often contaminated by different noise sources such as
external reflections, variations in the optical properties of the specimen and non-uniform heating.
These Depending on in
noise effects thethermal
type of images
externalproduce
optical heat sourcethermal
abnormal used topatterns
generateand
thermal
makewaves, the most
the detection of
common
damage more complex. A number of signal processing techniques has been developed to reduce are
methods used in optically stimulated thermography (OST) for aerospace applications the
pulsed
noise inthermography
thermal images, (PT), lock-in
thus thermography
enhancing the damage (LIT), step heating
detection thermography
and quantification (SHT) and
capabilities of long
each
pulse thermography (LPT), frequency modulated thermography (FMT), laser-spot
OST method 39. An extensive review of signal processing methods for IRT data was recently thermography (LST)
and laser-line
provided thermography
by 41. (LLT)a [35–38].
Table 3 reports summary Figure 2 shows
of the the experimental
most used setup techniques
signal processing of OST for different
for OST
optical
and their excitation sources.
basic principles.

Figure 2.
Figure 2. Schematic
Schematic diagram
diagram of
of optically
optically stimulated
stimulated thermography
thermography (OST).
(OST).

Table 3. Summary of signal processing techniques used for OST.

Data obtained by an IR camera in OST are often contaminated by different noise sources such
Signal
as Processing
external Method variations
reflections, Abbreviation
in the optical properties of the Basic Principle
specimen and non-uniform heating.
It is one of the well-known processing technique that uses the difference
These noise effects in thermal images produce abnormal thermal patterns and make the detection
Differential Absolute
DAC between the temperature of a sound area and a defected area. DAC is
Contrast
of damage42 more complex. A number of signal processing techniques has been developed to reduce
based on the 1D solution of the Fourier heat equation.
the noise in thermal images, thus enhancing thepolynomial
It creates damage detection and quantification
filters from log-log pre-calculations capabilities
for the surface of
each OST method [39,40]. An extensive review of signal processing methods for IRT datapolynomial
temperature response on each pixel by fitting a low-order was recently
function to the temperature evolution cooling profiles. Basic TSR
provided by [41]. Table 3 reports a summary of the most used signal processing techniques for OST
significantly reduces the noise in thermal images and it can be used to
Thermographic
and their basicSignal
principles. TSR generate time-derivative images without additional noise contribution.
Reconstruction 44
These time-derivative images can reduce the effects of non-uniform
Table 3. Summary of heating,
signal processing techniques
the background used
reflection for OST.
artefacts and provide good sensitivity to
smaller and deeper defects. However, TSR only filters data temporally and
Signal Processing Method does not make use of the spatial
Abbreviation information.
Basic Principle
It uses the singular value decomposition (SVD) method to reduce an
It is one of the well-known processing technique that uses the
appropriately constructed matrix of observations to a set of orthogonal
Differential Absolute difference between the temperature of a sound area and a
DAC functions that produce a useful representation of both spatial information
Contrast
Principal [42,43]
Component defected area. DAC is based on the 1D solution of the Fourier
PCT and temporal features for pulse thermography images. This technique
Thermography 46 heat equation.
provides an estimate of the average flaw depth within a flawed zone but
It creates
does not providepolynomial filters
any indication from
of the log-logdistribution.
damage pre-calculations
PCT isfor
also
the surfaceexpensive.
computationally temperature response on each pixel by fitting a
low-order polynomial function to the temperature evolution
cooling profiles. Basic TSR significantly reduces the noise in
thermal images and it can be used to generate time-derivative
Sensors 2018, 18, x; doi:
Thermographic Signal www.mdpi.com/journal/sensors
TSR images without additional noise contribution. These
Reconstruction [44,45]
time-derivative images can reduce the effects of non-uniform
heating, the background reflection artefacts and provide good
sensitivity to smaller and deeper defects. However, TSR only
filters data temporally and does not make use of the
spatial information.
Sensors 2018, 18, 609 6 of 37

Table 3. Cont.

Signal Processing Method Abbreviation Basic Principle

It uses the singular value decomposition (SVD) method to
reduce an appropriately constructed matrix of observations to
a set of orthogonal functions that produce a useful
Principal Component representation of both spatial information and temporal
PCT
Thermography [46] features for pulse thermography images. This technique
provides an estimate of the average flaw depth within a
flawed zone but does not provide any indication of the
damage distribution. PCT is also computationally expensive.
It is widely used in lock-in thermography and it is based on
the principle that the harmonic thermal response measured by
the IR camera is analysed with in-phase correlation and
Correlation Extraction
CEA cross-correlation functions in order to extract information
Algorithm [47]
about the amplitude and phase of measured thermal signals.
The feasibility of this technique depends on the length of the
image sequence.
It suppresses surface clutter conditioned by uneven heating
and lateral heat diffusion for pulsed thermography. The
algorithm does not require selecting a reference point for the
3D Normalisation sound area and involves the division of the IR image sequence
3DNA
Algorithm [48] in synthetic sequences, which are calculated by solving the
corresponding three-dimensional problem of heat conduction.
This method requires the determination of material
thermal properties.
It is used to decompose the TSR-smoothed thermal image into
Multi-dimensional the high-frequency noise, low-frequency background and the
Ensemble Empirical Mode MEEMD informative part of the signal in order to remove the noise and
Decomposition [49] non-uniform background from the thermographic data, thus
improving the damage detection capabilities.
It is a two-dimensional reference-free quantitative detection
method for sub-surface defects used for pulsed thermography.
It relies on the determination of a damage index pattern that is
Gapped Smoothing function of the thermal contrast between real and estimated
GSA
Algorithm [50] temperature profiles for all pixels. GSA method can enhance
the thermal contrast and suppress the effect of non-uniform
heating. It was also shown to improve the detection of
damage far from the heating surface.
It is mainly used for pulsed thermography and it is based on
the partial least squares regression (also known as projection
to latent structures) that computes new thermal sequences
that are correlated to the predicted block Y and the predictor
matrix X. The matrix X corresponds to the surface temperature
matrix obtained during the thermography inspection, whilst Y
is defined by the observation time during which thermal
Partial Least Squares
PLST images are captured. The new set of thermal images and
Thermography [51]
observation time vector is composed of variables that consider
only the most important signal variations. Unnecessary
information present in the original thermal sequence is
neglected. Contrary to PCT, PLST keeps track of the time.
Hence data set can be decomposed, manipulated and
recomposed, for example by omitting certain loadings with as
benefits enhanced signal-to-noise ratio (SNR).
It is a reference-based quantitative detection method based on
fitting a second order polynomial model for temperature
decay curves. The coefficients of the model are much less
Coefficient Clustering sensitive to noise and more consistent for pixels from the
CSA
Analysis [52] sound area as compared to a high order model. This technique
not only provides an enhanced visual confirmation of the
damage, but it also reduces the burden on the operator in
post-processing data.
Sensors 2018, 18, 609 7 of 37

2.1. Pulsed Thermography

PT is a rapid NDE process that uses a high intensity pulse of light to heat the surface of a
test piece via the photothermal effect [41]. Very short duration flashes (ranging between 0.1 and
50 ms) are typically employed, depending on the sample thermal properties and defect depths.
The temperature gradients of the sample surface are monitored by an IR camera with a computer
system to quantify the material damage [53]. Sub-surface defects reduce the conduction of heat away
from the surface, decreasing the surface cooling rate compared to that of a non-defective region.
The transient temperature field T(z,t) in PT is obtained as solution of the following inhomogeneous
one-dimensional (1D) heat equation [54]:

∂2 T (z, t) 1 ∂T (z, t) g(z, t)

2
− =− t>0 (1)
∂z α ∂t k

where α = k/ ρc is the thermal diffusivity (m2 /s), with k the thermal conductivity (W/mK), ρ the
density (kg/m3 ) and c the specific heat (J/kg K), g(z,t) = Q0 δ(z−z)δ(t−t0 ) is the external impulsive
heat source located in z0 = 0 and excited at time t0 = 0, with Q0 the source intensity per unit length
(Jm−1 ) and δ(z−z)δ(t−t0 ) Dirac’s delta functions in space and time, respectively. According to Green’s
theorem [55], to solve the linear inhomogeneous transient heat Equation (1), it is useful to define a
Green space-time function that satisfies the homogeneous boundary conditions on the boundary of the
material. Hence, the solution of the inhomogeneous heat Equation (1) can be expressed as follows:

Q 0 − z2
T (z, t) = √ e 4αt (2)
2ζ πt

where ξ = kρc is the thermal effusivity (Ws1/2 m−2 K−1 ) of the material, which controls the reflection
p

of heat from discontinuities. Equation (2) represents the temperature rise at a depth z beneath the
surface at a time t after a uniform impulse of energy Q0 . The surface thermal response is simply
obtained from Equation (2) by posing z = 0. Assuming a region containing a delamination-like defect
as a layer of thickness d, i.e., the depth of the defect below the surface, the impulsive thermal response
Td (0,t) at the surface becomes [53]:

∞ 
( )
(nd)2 (nL)2 ( D/2)2

Q0
Td (0, t) = √ 1 + 2 ∑ e αt − e αt
− −
1 − e− 4αt (3)
2ζ πt n =1

where, L is the thickness of the material and D is the diameter of the circular defect.
Pulsed Phase Thermography (PPT) is the phase analysis in frequency domain of PT and is a
signal processing method that solves the problem of heating inhomogeneity as it is less sensitive to
the non-uniformity of heating, surface irregularities and environmental reflections [56]. It can also
image large components on airframes relatively fast [57]. Since PPT is widely used in the aerospace
industry for the inspection of aerospace metallic and composite components [58], it can be considered
as a valid alternative to time domain PT. Low frequencies, typically up to 2–3 Hz are used in PPT to
assess deeper defects. The procedure to extract the phase from the thermal data T(k) is based on the
Fourier Transform, which is computed according to the well-known formula [59]:

N  
j2πkn
T̂n = ∑ T ( k ) exp −
N
= Ren + Imn (4)
n =1

where j is an imaginary number, n design the modulation frequency increment, N is the total number
of thermograms, and Ren and Imn are the real and imaginary parts of the Fourier Transform T̂n of the
thermal emissivity data. Hence, the phase φn is simply obtained as φn = tan−1 (Imn /Ren ). For PPT
evaluation, the term “blind frequency” fb is often used as the limiting frequency at which the defect
 Sensors 2018, 18, 609 8 of 37

presents enough phase contrast to be detected on the phase spectrum. The “blind frequency” and the
defect depth are typically correlated by the following formula [60]:
r
α
d = C1 (5)
π fb

where C1 is called a “correlation constant” and ranges between 1.5 and 2 [61].
PT can be applied to several types of aerospace materials such as composite, metallic and
polymeric components [16,62]. In Cramer’s report [63], NASA used PT for on-orbit inspection
of manned flight vehicles and the International Space Station (ISS). Figure 3 shows the PT results
coupled with solar heating achieved by astronauts during extravehicular activity (EVA) inspections.
In particular, it illustrates a PT inspection of both a CFRP component of the Space Shuttle with holes
and a damaged ISS radiator. Black holes in Figure 3a correspond to pre-damaged carbon-carbon
composite
Sensors 2018, 18,samples
x imaged by the EVA IR camera. 4 of 37

Figure
Figure3.3.PT
PTresults
resultsobtained
obtainedduring
duringEVA
EVAinspections
inspectionsofofa aCFRP
CFRPcomponent ofofthe
component theSpace
SpaceShuttle
Shuttle(a)(a)
and a damaged ISS radiator (b), from 63.
and a damaged ISS radiator (b), from [63].

Li et al. 50 assessed the performance of the Gapped Smoothing Algorithm (GSA, see Table 3) for
Li et al. [50] assessed the performance of the Gapped Smoothing Algorithm (GSA, see Table 3) for
damage detection on an aluminium plate and a glass fibre reinforced composite laminate containing
damage detection on an aluminium plate and a glass fibre reinforced composite laminate containing
different defects such flat-bottom holes, delamination and material impurities. The authors used a
different defects such flat-bottom holes, delamination and material impurities. The authors used a
TIM-160 IR camera (MICRO-EPSILON, Ortenburg , Germany) and four heating lamps (240 V, 275 W)
TIM-160 IR camera (MICRO-EPSILON, Ortenburg , Germany) and four heating lamps (240 V, 275 W)
to generate heat. The numerical results on both aluminium and composite plates demonstrated that
to generate heat. The numerical results on both aluminium and composite plates demonstrated that
the proposed GSA method was robust, accurate and effective for detecting sub-surface flaws (Figure
the proposed GSA method was robust, accurate and effective for detecting sub-surface flaws (Figure 4).
4).
Manhoar and Lanza di Scalea [64] developed and validated a three-dimensional (3D) analytical
model to simulate the heat flow interaction with defects in a CFRP panel containing rectangular
flat-bottom holes of different sizes. The authors found that experimental results were reasonably
close to theoretical predictions. Montesano et al. [14] investigated the fatigue behaviour of a triaxially
braided carbon fibre fabric (T650/35-6 K), with a 0◦ /±ϑ braid orientation, embedded in a thermosetting
polyimide resin. A SC5000 IR camera (FLIR, Wilsonville OR, United States) with spatial resolution
of 320 × 240 was used and heat was generated by mechanical loading. The authors elaborated a
fatigue threshold value that was in excellent agreement with that obtained through a conventional
PT experimental test. Guo and Vavilov [65] used PT for the detection of disbonds in the insulation of
solid rocket motors made of a low-density rubber-like elastomer. The authors compared several data
 and a damaged ISS radiator (b), from 63.

Li et al. 50 assessed the performance of the Gapped Smoothing Algorithm (GSA, see Table 3) for
damage detection on an aluminium plate and a glass fibre reinforced composite laminate containing
different defects
Sensors 2018, such flat-bottom holes, delamination and material impurities. The authors used9 ofa 37
18, 609
TIM-160 IR camera (MICRO-EPSILON, Ortenburg , Germany) and four heating lamps (240 V, 275 W)
to generate heat. The numerical results on both aluminium and composite plates demonstrated that
processing approaches based on the correlation technique and showed that the highest signal-to-noise
the proposed GSA method was robust, accurate and effective for detecting sub-surface flaws (Figure
ratio (SNR) appeared when choosing a reference point in a defect-free area of the sample.
4).

Figure 4. Comparison between PT thermal results from raw data (a); Differential Absolute Contrast
Figure 4. Comparison between PT thermal results from raw data (a); Differential Absolute Contrast
(DAC) method (b) and Gapped Smoothing Algorithm (GSA) (c), with permission from 50.
(DAC) method (b) and Gapped Smoothing Algorithm (GSA) (c), with permission from [50].

Manhoar and Lanza di Scalea 64 developed and validated a three-dimensional (3D) analytical
Assimulate
model to reportedtheinheat
Table 3, interaction
flow advanced with
signal processing
defects methods
in a CFRP panel have been rectangular
containing applied to flat-
PT in
orderholes
bottom to enhance the detection
of different sizes. Theand quantification
authors of material damage
found that experimental in aerospace
results were components.
reasonably close to
Lopez et al. [51] used the Partial Least Square Thermography (PLST) method to optimise the
theoretical predictions. Montesano et al. 14 investigated the fatigue behaviour of a triaxially braided processing
of PT fibre
carbon data on a CFRP
fabric composite
(T650/35-6 K), with
with simulated defects.
a 0°/±ϑ braid The experimental
orientation, embedded results, achieved with a
in a thermosetting
FPA SBF125 IR camera (Santa Barbara, Lowell MA, USA) and two photographic flash
polyimide resin. A SC5000 IR camera (FLIR, Wilsonville OR, United States) with spatial resolution lamps (FX60,
of
320 × 240 was used and heat was generated by mechanical loading. The authors elaborated aof
Glamox, Borehamwood, UK) powered at 6.4 kJ, showed an increase in the SNR for 96% defects
fatigue
after processing the images with the PLST technique. Zhao et al. [52] used the Coefficient Clustering
Analysis (CSA) method coupled to PT, whilst Chang et al. [49] employed the Multi-dimensional
Ensemble Empirical Decomposition (MEEMD) algorithm to quantify material damage in impacted
CFRP composites. Pawar and Vavilov [48] used a 3D Normalisation Algorithm (3DNA) for PT
to compensate the background non-uniformity in both damaged glass and carbon fibre reinforced
composites. Tang et al. [66] combined the Markov and Principal Component Thermography (PCT)
algorithm with PT in order to detect debonding in ceramic thermal barrier coatings for aerospace.
The depth and diameter prediction error was about 4–10% for defects with a diameter–depth ratio of
1.2–4.0, and depth between 1.0 and 2.5 mm, which proved the effectiveness of the proposed method
for a quantitative detection of material defects. Li et al. [67] compared the PT method with ultrasonic
stimulated thermography (UST, see Section 3). The authors found that PT is more suitable for detecting
delamination, especially with large areas, whilst UST is superior to detect small cracks such as matrix
cracking and fibre breakage. The combination of the above two methods can greatly improve the
capability to detect and evaluate impact damage in aerospace composites.
Shin et al. [57] used PPT to detect the initiation and propagation of fatigue-induced damage in
single lap joints with CFRP adherends (Figure 5). The authors developed a threshold technique that
allowed the identification of fatigued samples with early stage material damage, thus preventing
premature failures. Oswald-Tranta [68] provided a quantitative comparison between the time domain
Thermographic Signal Reconstruction (TSR) method and the frequency domain PPT for an aluminium
sample with flat bottom holes. The author found that whilst both signal-processing methods provide
good SNR, the maximum of second derivative used in TSR allows an easier calculation of the defect
composites. Tang et al. 66 combined the Markov and Principal Component Thermography (PCT)
algorithm with PT in order to detect debonding in ceramic thermal barrier coatings for aerospace.
The depth and diameter prediction error was about 4–10% for defects with a diameter–depth ratio of
1.2–4.0, and depth between 1.0 and 2.5 mm, which proved the effectiveness of the proposed method
for a quantitative detection of material defects. Li et al. 67 compared the PT method with ultrasonic
Sensors 2018, 18, 609 10 of 37
stimulated thermography (UST, see Section 3). The authors found that PT is more suitable for
detecting delamination, especially with large areas, whilst UST is superior to detect small cracks such
depth
as compared
matrix crackingto Equation
and fibre(5)breakage.
in PPT. Moreover, unlike PPT,
The combination of TSR
the does
abovenottwo
require any selection
methods of a
can greatly
sound reference point.
improve the capability to detect and evaluate impact damage in aerospace composites.

Figure
Figure 5.
5. PPT
PPT phase
phase results
results on
on aa single
single lap
lap joints
joints with
with CFRP
CFRP adherends
adherends undergone
undergone toto different
different fatigue
fatigue
loading
loading cycles, with permission from [57]. Each subfigure corresponds to the thermal phase image of
cycles, with permission from 57. Each subfigure corresponds to the thermal phase image of
the composite sample
the composite sampleafter
after200
200cycles
cycles(a);
(a);400
400 cycles
cycles (b);
(b); 400400 cycles
cycles andand pretension
pretension (c) and
(c) and 600 600 cycles
cycles (d).
(d).
2.2. Lock-In Thermography
In LIT, an external periodic heat excitation is used to generate thermal waves within the sample.
This is usually performed by employing sinusoidally modulated tungsten-halogen flood lamps [69].
Temperature variations at damage location are monitored with an IR camera and the magnitude
images are computed for each heat-generating frequency by post-processing the recorded thermal
data using the Fourier transform algorithm. Thermal images are recorded throughout the heating
period that lasts for at least a full excitation cycle. LIT has the advantage of being less sensitive to
the local variations of surface emissivity and continuous IR radiation once the incandescent lamp
is switched off [17]. For a homogenous isotropic semi-infinite material whose surface is uniformly
illuminated by a light beam of periodically modulated intensity qz = Q0 [1+exp(jωt)]/2, with ω being
the angular modulation frequency of the heat source, the temperature field can be obtained by solving
the homogeneous 1D heat diffusion equation associated with Equation (1). The solution is expressed
as follows [70]:
Q0 − z j(ω t− µz )
T (z, t) = √ e µ e (6)
2ζ ω
√
where µ = 2α/ω is the thermal diffusion length of the material. As it can be seen from Equation (6),
thermal waves are very heavily damped with a constant decay equal to the thermal diffusion length.
Moreover, thermal
√ waves are characterised by a dispersive behaviour, with their phase velocity defined
as ν = ωµ = 2αω. The response at the surface monitored by an IR camera can be obtained either
from Equation (6) by posing z = 0, or as function of the thermal surface response using PT at the time
t = 1 s [70]:
Sensors 2018, 18, 609 11 of 37

Q0
T (0, t) = √ (7)
2ζs ω
with s being the duration of the LIT experiment. Duan et al. [71] provided a quantitative evaluation
of both LIT and PT for a set of aluminium foam specimens used for aerospace applications with
flat-bottom holes. The authors found that although the post-processing methods of PT (including
TSR and PCT, see Table 3) improve the thermal contrast for sub-surface and deeper defects, it is more
difficult to analyse raw PT data due to non-uniform heating. Kordatos et al. [72] combined LIT and
acoustic emissions measurements in order to determine the fatigue limit on dog-bone aluminium and
cross-ply SiC/BMAS composite samples used in aircraft components. Grammatikos et al. analysed an
aluminium wing structure with CFRP composite patches using LIT in order to assess the durability
and efficiency of the patch repair during both maintenance (off-line) and service (on-line) conditions
(Figure 6). 2018, 18, x
Sensors 7 of 37

Sensors 2018, 18, x 7 of 37

Figure 6. Illustration of the wing and its schematic representation analysed with LIT, from 73.
Figure 6. Illustration of the wing and its schematic representation analysed with LIT, from [73].
Unit measures in Figure
Zhao et al. 6d arethe
74 analysed in mm.
phase variations of the surface temperature with LIT in order to
Figure
investigate 6. Illustration
skin-to-core of the wing
disbonds in aand its schematic
titanium alloyrepresentation analysed with
honeycomb structure LIT, from
(Figure 73. authors
7). The
Zhao et al. [74]a analysed
also developed the phase
three-dimensional variations
finite of theofsurface
element model temperature
the titanium with LIT
alloy honeycomb in order to
sandwich
Zhao et al. 74 analysedbetween
investigate skin-to-core
and found good disbondsthe
agreement in phase variations
a titanium
numerical andof
alloy the surface temperature
honeycomb
experimental structure with LIT in order to
results. (Figure 7). The authors also
investigate skin-to-core disbonds in a titanium alloy honeycomb structure (Figure 7). The authors
developed a three-dimensional finite element model of the titanium alloy honeycomb sandwich and
also developed a three-dimensional finite element model of the titanium alloy honeycomb sandwich
foundand
good agreement between numerical and experimental results.
found good agreement between numerical and experimental results.

Figure 7. Lock-in thermography testing of a titanium alloy honeycomb sandwich structure with
disbond, with permission from 74.

Figure
A number7. Lock-in thermography
of authors testing
has recently usedofLIT
a titanium
for damage alloydetection
honeycomb andsandwich structure
quantification with
in aerospace
Figure 7. Ampligram (a) and phasegram (b) obtained with LIT testing of a titanium alloy honeycomb
disbond, with permission from 74.
composite components 75. Junyan et al. 79 used LIT to quantify damage in SiC coated carbon-carbon
sandwich structure with disbond, with permission from [74].
composites used for the spacecraft re-entry thermal protection system. The same authors also
A number
developed of authors
an inverse has recently
approach applied used LITin
to LIT fororder
damage detection
to estimate theand quantification
thermal in aerospace
diffusivities and sub-
composite components 75. Junyan et al. 79 used LIT to quantify damage in SiC coated
surface defect size of CFRP laminate composites 80. Such inverse approach consisted of solving the carbon-carbon
composites used for
objective function the spacecraft
constructed from LIT re-entry thermalbecause
phase images protection system.
of being The sameto authors
less sensitive also
non-uniform
developed an inverse approach applied to LIT in order to estimate the thermal diffusivities
heating. A FLIR SC 7000 camera (FLIR, Wilsonville OR, United States) with a resolution of 320 × 256 and sub-
surface
was useddefect size
for the of CFRP laminate
experimental testing.composites
De Angelis80. Such
et al. inverse
Error! approach
Reference consisted
source of solving
not found. the
compared
objective function constructed from LIT phase images because of being less sensitive to non-uniform
Sensors 2018, 18, 609 12 of 37

A number of authors has recently used LIT for damage detection and quantification in aerospace
composite components [75–79]. Junyan et al. [80] used LIT to quantify damage in SiC coated
carbon-carbon composites used for the spacecraft re-entry thermal protection system. The same
authors also developed an inverse approach applied to LIT in order to estimate the thermal diffusivities
and sub-surface defect size of CFRP laminate composites [3]. Such inverse approach consisted of
solving the objective function constructed from LIT phase images because of being less sensitive to
non-uniform heating. A FLIR SC 7000 camera (FLIR, Wilsonville OR, United States) with a resolution
of 320 × 256 was used for the experimental testing. De Angelis et al. [4] compared LIT with other
NDE systems such as shearography, thermosonics and traditional ultrasounds in order to provide
flaw detection across vulnerable surfaces of a Typhoon air-cooling inlet composite panel. The authors
found that LIT can be used for both laboratory and/or full field applications of aeronautical composite
components. Similarly, Ranjit et al. [81] provided a quantitative analysis of artificial defects in a GFRP
composite plate of dimensions 300 × 300 × 1.4 mm by comparing LIT amplitude and phase images
with shearograhy and linear ultrasounds (Figure 8). A FLIR SC465 IR camera with a resolution of
640 × 480 was used for the thermography testing.
Meola et al. [82] used LIT to monitor impacted composite samples with different types of
reinforcement (e.g., carbon and glass fibres) and matrix (e.g., thermoset and thermoplastic with
a compatibilising agent). The authors found that although LIT is able to reveal different types of
failures such as fibre breakage and delamination, it still fails to appraise the whole extension of the
delaminated zone from the amplitude and phase images. Indeed, micro-crack and delamination can
be confused with small material non-uniformities.

2.3. Step Heating and Long Pulse Thermography

In SHT and LPT a low intensity step-pulse (heat source) is applied for a long period (typically
from milliseconds to few seconds), thus enabling a longer heating time to locate deeper defects. The
main difference between the two techniques is that in SHT data are measured during the application
of the step-pulse, whereas in LPT thermal signals are acquired during the cooling phase [23].
The temperature response T(0,t) measured at the surface of a homogeneous and semi-infinite
medium is obtained as solution of the homogeneous 1D heat diffusion equation associated with
Equation (1) by applying a step-pulse heating source Q0 [54]. This is expressed as follows:
r
2Q0 t
T (0, t) = (8)
ζ π

In the case of a material with finite thickness L, the surface temperature T(0,t) is given by [22,83]:

√ ∞
r "  #
2Q0 t nL
T (0, t) = 1 + π ∑ 2ier f c √ (9)
ζ π n =1 αt
√
with ier f c( x ) = 1/ π exp − x2 − x ·er f c( x ) . Roche and Balageas [84] compared the thermal

imaging results obtained with both SHT and PT in order to detect material damage located at different
depth within a CFRP composite plate. The authors found that without any signal processing technique
applied to both thermographic methods (i.e., during visual examination), SHT was able to reveal a
surface defects but not in-depth damage. However, when TSR was applied to both SHT and PT, the last
technique was better contrasted and allowed a better damage identification even for deepest defects.
These results were also confirmed by Almond et al. [23], who applied SHT to both aluminium and
composite samples with flat bottom holes of different dimensions and depth. The authors provided an
analytical study of the SHT process, which revealed that the magnitude of defect image contrast is
dependent on the thermal response rate of the material.
Sensors 2018, 18, x 8 of 37

300 × 1.4 mm by comparing LIT amplitude and phase images with shearograhy and linear
ultrasounds
Sensors 2018, 18,(Figure
609 8). A FLIR SC465 IR camera with a resolution of 640 × 480 was used for the
13 of 37
thermography testing.

Figure 8. Lock-in thermography amplitude and phase images on a GFRP composite plate with
Figure 8. Lock-in thermography amplitude and phase images on a GFRP composite plate with artificial
artificial defects, with permission from Error! Reference source not found..
defects, with permission from [84].
Meola et al. 81 used LIT to monitor impacted composite samples with different types of
2.4. Frequency Modulated
reinforcement (e.g., carbonThermography
and glass fibres) and matrix (e.g., thermoset and thermoplastic with a
compatibilising
FMT uses aagent). The authors
non-stationary form found that although
of thermal excitation LITinis which
able toa reveal differentfrequency
low intensity types of
failures such as fibre breakage and delamination, it still fails to appraise
modulated input signal is used to generate thermal waves within the material [24]. Unlike the whole extension of LIT
the
delaminated zone from the amplitude and phase images. Indeed, micro-crack and
that uses a single excitation frequency for a fixed damage depth resolution, FMT can be used to detect delamination can
be confused
material with small at material non-uniformities.
hflaws located
h various
i depths
i in the test sample [85]. The temperature field of intensity
πBt2
qz = Q0 1 + exp j ωt + τ /2] , with τ the duration of the frequency modulated heat cycle of
2.3. Step Heating and Long Pulse Thermography
bandwidth B, is obtained as solution of the homogeneous 1D heat diffusion equation associated with
In SHT
Equation (1).and
ThisLPT a low intensity
is defined as follows step-pulse
[24]: (heat source) is applied for a long period (typically
from milliseconds to few seconds), thus enabling a longer heating time to locate deeper defects. The
Q0 − µ in z πBt2 − z )
j(ωt+data
main difference between the twoT (techniques
z, t) = √is thate FM SHT e τ are
µ FMmeasured during the application
(10)
of the step-pulse, whereas in LPT thermal2ζsignals ω are acquired during the cooling phase 23.
The temperature
r  response  T(0,t) measured at the surface of a homogeneous and semi-infinite
medium 2Bt
with µ FMis = obtained
2α/ asω+ solution
τ of thermal
the the homogeneous
diffusion length1D heat of diffusion equation
the frequency associated
modulated with
heating.
Equation (1) by applying a step-pulse heating source Q 0 54. This is expressed as follows:
Arora et al. [86] applied FMT on a CFRP sample with flat bottom holes and analysed the depth
scanning performance using both time domain (magnitude)
2Q0 t and frequency domain (phase) imaging
T 0phase
data. The experimental results showed that , t   retained (8)
its advantages of energy concentration,
 
which led to enhanced depth scanning with improved resolution and sensitivity for defect detection.
Sensors 2018, 18, 609 14 of 37

2.5. Laser-Spot Thermography and Laser-Line Thermography

The optical flash lamp heating process used for LIT, SHT and FMT produces a uniform area of
heating across the surface of the test specimen. This results in a heat flow from the surface into the
bulk, which limits the detected material flaw to delaminations lying in a plane parallel to the surface
(interlaminar) or at the interfaces between a surface coating and its substrate. Hence, material damage
that forms in planes perpendicular to the surface (intralaminar) is not detectable with traditional flash
lamps [25]. Cracks of this type can be detected if the heating process is localised to a spot or a line on a
surface. In both LST and LLT, damage can be identified either by a change in the shape of the heated
spot or by an apparent increase in temperature [87]. A complete inspection of a test piece is achieved
by processing thermal images collected during scanning the spot or line over the test piece surface.
In the case of LLT, either the sample is moving with a constant speed passing in front of a fixed line
heat source or vice versa [26]. The IR camera records the temperature distribution at the surface in
a specified distance after the heating (cooling effect [88]). Figure 9 illustrates the schematic diagram
for LST.
In LST, the temperature field T(r,z,t) after the laser spot is switched off is analytically obtained by
convoluting a Gaussian shape round spot source (i.e., the laser pulse) with continuous heating. This is
expressed as follows [54,89]:

r Zt z2 − r2
Imax a2 α p(t − t0 )e− 4αt0 e 4αt0 +a2 0
T (r, z, t) = √ dt (11)
k π t0 (4αt0 + a2 )
0

where (r,z)pare cylindrical coordinates with the origin on the surface at the centre of the irradiated spot,
with r = x2 + y2 , Imax and a are the maximum power density of the laser pulse and the radius of the
laser beam, respectively, and p(t) is the normalised temporal profile of the laser pulse at the time t. The
temperature
Sensors 2018, 18,response
x T(r,0,t) measured at the surface is obtained by posing z = 0 in Equation (11).
10 of 37

Figure 9. Experimental set-up for laser scanning thermography, with permission from 86.
Figure 9. Experimental set-up for laser scanning thermography, with permission from [90].

In LST, the temperature field T(r,z,t) after the laser spot is switched off is analytically obtained
In LLT, a laser
by convoluting line source
a Gaussian canround
shape be approximated
spot source to a Gaussian
(i.e., the laser shape
pulse)elliptical spot source
with continuous with
heating.
continuous heating and the temperature
This is expressed as follows 54: profile after cooling is given by [54]:

r Zt 2
z2 z −  xr
22 y2
−
Imax ab α2 p(t t p−tt0 )te−e4αt40et e 4αt
40t+
aa22 e 4αt0 +b2
T ( x, y, z,Tt)r=
, z, t  k max π
I a dt0
k   p
0
t (4αt 0

t  4+t  a )(
2
 0
a 4αt + b )
2
dt  2 (11)
(12)
0 0

where
with x, (r,z)
y and are cylindrical
z the Cartesian coordinates
coordinates with
andthe origin
a and b theonlaser
the beam
surface at the
radii. centre of
Similarly to the irradiated
Equation (11),
spot, with 𝑟 = √𝑥 2 + 𝑦 2 , Imax and a are the maximum power density of the laser pulse and the radius
the temperature response T(x,y,0,t) measured at the surface is obtained by posing z = 0 in Equation (12).
of the laser beam, respectively, and p(t) is the normalised temporal profile of the laser pulse at the
time t. The temperature response T(r,0,t) measured at the surface is obtained by posing z = 0 in
Equation (11).
In LLT, a laser line source can be approximated to a Gaussian shape elliptical spot source with
continuous heating and the temperature profile after cooling is given by 54:
Sensors 2018, 18, 609 15 of 37

Ley et al. [90,91] used laser thermography to detect sub-surface defects in different composite
materials commonly used in aerospace applications such as composite sandwich panels with artificial
flaws (i.e., embedded Teflon patches) and real impact damage. The authors found good agreement
between the laser thermography results and those obtained with ultrasonic C-Scan mesurements.
Fedala et al. [92] combined LIT with a modulated laser excitation for the assessment of surface
cracks in metallic turbine blades. Roemer et al. [93] compared LST with UST (see Section 3) to detect
fatigue cracks on an aluminium sample (Figure 10). Khodayar et al. [94] used robotized Line Scan
Thermography to inspect large composite materials, whilst Zhang et al. [95] used pulsed micro-laser
line thermography to detect submillimeter porosity in CFRP composites. Fernandes et al. [96] employed
a flying laser spot to heat a line region on a CFRP sample and assess the fiber orientation over that
region. The non-contact nature and predispositions for automatization makes LST and LLT very
promising for aerospace applications. Moreover, by simply combining IR cameras with zooming and
focusing lenses,
Sensors 2018, 18, x these techniques can be also used to identify micro-defects. 11 of 37

Figure 10. Laser-spot thermography result on fatigue damaged aluminium sample, from 92.
Figure 10. Laser-spot thermography result on fatigue damaged aluminium sample, from [93].

3. Ultrasonic Stimulated Thermography

3. Ultrasonic Stimulated Thermography
Optically Stimulated Thermography (OST) uses external optical flashes to heat the sample on its
Optically Stimulated Thermography (OST) uses external optical flashes to heat the sample on
surface and then an IR camera to record the temperature decay curve. Although good results can be
its surface and then an IR camera to record the temperature decay curve. Although good results
obtained for surface inspection, this technique may not be suitable for detecting in-depth damage,
can be obtained for surface inspection, this technique may not be suitable for detecting in-depth
closed cracks and damage failures at the early stages. Indeed, the usage of external optical heat
damage, closed cracks and damage failures at the early stages. Indeed, the usage of external optical
sources may limit the detection of defects to those located within a few millimetres from the material
heat sources may limit the detection of defects to those located within a few millimetres from the
surface. In addition, the lack of a significant air gap between the cracked surfaces may not generate
material surface. In addition, the lack of a significant air gap between the cracked surfaces may not
any significant variation of the local temperature needed for damage detection. UST (also known as
generate any significant variation of the local temperature needed for damage detection. UST (also
thermosonics, sonic IR or vibro-thermography) is an alternative means of thermography that
known as thermosonics, sonic IR or vibro-thermography) is an alternative means of thermography
involves the generation of powerful vibrations in a test piece to cause frictional heating at crack
that involves the generation of powerful vibrations in a test piece to cause frictional heating at crack
surfaces. An IR camera with temperature resolution ranging between 20 and 40 mK is usually used
surfaces. An IR camera with temperature resolution ranging between 20 and 40 mK is usually used
to measure the frictional heating on the sample’s surface 96. Figure 11 shows the experimental set-up
to measure the frictional heating on the sample’s surface [97–99]. Figure 11 shows the experimental
for thermosonics. Elastic vibrations are generally produced by an ultrasonic plastic welding horn
set-up for thermosonics. Elastic vibrations are generally produced by an ultrasonic plastic welding
being pressed against a surface on the part under inspection (Figure 12). A pulse of high acoustic
horn being pressed against a surface on the part under inspection (Figure 12). A pulse of high acoustic
power (between 1 kW and 80 kW) in the 15–50 kHz frequency range is typically applied and the
power (between 1 kW and 80 kW) in the 15–50 kHz frequency range is typically applied and the
duration can vary from 30 to 200 ms.
duration can vary from 30 to 200 ms.
Similarly to OST, depending on the shape of the ultrasonic driving signal, additional excitations
can be used in thermosonics. These include “ultrasonic lock-in thermography” with continuous
monochromatic elastic input 99 and “ultrasonic frequency modulated thermography” with chirp
excitation signals 100. Umar et al. 100 recently reported a review of earlier and later research on
thermosonics for different application areas.
 duration can vary from 30 to 200 ms.
Similarly to OST, depending on the shape of the ultrasonic driving signal, additional excitations
can be used in thermosonics. These include “ultrasonic lock-in thermography” with continuous
monochromatic elastic input 99 and “ultrasonic frequency modulated thermography” with chirp
excitation
Sensors 2018, 18,signals
609 100. Umar et al. 100 recently reported a review of earlier and later research16on
of 37
thermosonics for different application areas.

Figure 11. Experimental set-up of ultrasonic stimulated thermography, from 97.

Figure 11. Experimental set-up of ultrasonic stimulated thermography, from [98].
Sensors 2018, 18, x 12 of 37

Figure 12. Experimental set-up of ultrasonic stimulated thermography, with permission from 98.
Figure 12. Experimental set-up of ultrasonic stimulated thermography, with permission from [98].

Specifically for the aerospace industry, UST is mainly used for the inspection of micro-cracks (of
theSimilarly
order of to OST, depending
microns) in aluminium on the shape ofcomponents
aerospace the ultrasonic 101,driving
aircraftsignal,
engineadditional
fan blades excitations
102 and
cancomposite
be used primary
in thermosonics. Thesestructures.
and secondary include “ultrasonic
Early testing lock-in thermography”
on composite materialswith continuous
included the
monochromatic
evaluation of thickelastic input [100]
multilayer and “ultrasonic
carbon/carbon frequency
composites usedmodulated thermography”
in wing leading edges of thewith chirp
space
excitation signals
shuttle 104, the [101]. UmarofetF16
inspection al. airplane
[101] recently
main reported
landing gearsa review
105 of andearlier and laterofresearch
the detection internalon
delaminations
thermosonics forin graphite-fibre
different composites
application areas. and disbonds in airplane vertical stabilizers 106. Unlike
OST, the “clapping”
Specifically motion
for the and friction
aerospace industry,(“rubbing”) of closedused
UST is mainly cracks forstimulated by ultrasonic
the inspection waves
of micro-cracks
is an “ideal” condition for UST. Renshaw et al. 107 have shown that heat
(of the order of microns) in aluminium aerospace components [102], aircraft engine fan blades [103,104] generation in damage
and regions duringprimary
composite a thermosonics experiment
and secondary is mainly due
structures. Earlyto testing
three different mechanisms:
on composite (i) frictional
materials included
therubbing of contact
evaluation of thickregions of damage
multilayer interfaces; (ii)composites
carbon/carbon plastic deformations
used in wing during crack or
leading damage
edges of the
growth that generate heat in the plastic zone surrounding these areas;
space shuttle [105], the inspection of F16 airplane main landing gears [106] and the detection of and (iii) viscoelastic losses that
generate
internal significant heat
delaminations in regions ofcomposites
in graphite-fibre stress concentration
and disbonds (i.e., in
around
airplane delaminations and other
vertical stabilizers [107].
defects). Polimeno et al. 98 developed a compact thermosonics inspection system with a
Unlike OST, the “clapping” motion and friction (“rubbing”) of closed cracks stimulated by ultrasonic
microbolometer array camera to detect artificial delamination (i.e., an embedded Teflon patch) in
waves is an “ideal” condition for UST. Renshaw et al. [108] have shown that heat generation in
composite materials. The authors used a parameter, known as Heating Index 108, to predict the
damage regions during a thermosonics experiment is mainly due to three different mechanisms:
vibration level in the presence of vibrations governed by “acoustic chaos”. The calculation of the
(i) frictional rubbing of contact regions of damage interfaces; (ii) plastic deformations during crack or
Heating Index involved calibration methods and accurate measurements of the strain energy for each
damage growth that generate heat in the plastic zone surrounding these areas; and (iii) viscoelastic
test case. Gaudenzi et al. 109 compared the performance of UST with standard OST and ultrasonic
losses that
phased arraygenerate
system significant
on a CFRP heat in regions
samples of stress
undergone concentration
to low velocity impact (i.e., damage.
around delaminations
The authors
and found that UST provided a quick and reliable estimation of material micro-defects. Figuresystem
other defects). Polimeno et al. [99] developed a compact thermosonics inspection 13 shows with
a microbolometer
the experimental UST results of a composite sample with delaminations obtained at an impact energy in
array camera to detect artificial delamination (i.e., an embedded Teflon patch)
composite
of 20 J. materials. The authors used a parameter, known as Heating Index [109], to predict the
vibration level in the presence of vibrations governed by “acoustic chaos”. The calculation of the Heating
Index involved calibration methods and accurate measurements of the strain energy for each test case.
Gaudenzi et al. [110] compared the performance of UST with standard OST and ultrasonic phased array
system on a CFRP samples undergone to low velocity impact damage. The authors found that UST
provided a quick and reliable estimation of material micro-defects. Figure 13 shows the experimental UST
results of a composite sample with delaminations obtained at an impact energy of 20 J.
Heating Index involved calibration methods and accurate measurements of the strain energy for each
test case. Gaudenzi et al. 109 compared the performance of UST with standard OST and ultrasonic
phased array system on a CFRP samples undergone to low velocity impact damage. The authors
found that UST provided a quick and reliable estimation of material micro-defects. Figure 13 shows
the experimental
Sensors 2018, 18, 609 UST results of a composite sample with delaminations obtained at an impact energy
17 of 37
of 20 J.

Figure 13. Ultrasonic lock-in thermography results on a composite panel with BVID, with permission
Figure 13. Ultrasonic lock-in thermography results on a composite panel with BVID, with permission
from. Figures (a) and (c) show the lock-in results with delamination at◦ -45° and +-45°, and -45° and
from. Figure (a,c) shows the lock-in results with delamination at −45 and +/−45◦ , and −45◦ and
90°◦ respectively, whilst figure (b) and (d) illustrate the apparent temperature results at 0.04 Hz and
90 respectively, whilst figure (b,d) illustrates the apparent temperature results at 0.04 Hz and 0.1 Hz,
0.1 Hz, respectively. 109.
Sensorsrespectively.
2018, 18, x [110]. 13 of 37

Vavilov et al. al. [111] developedan

110 developed ananalytical
analyticalmodel
model for for ultrasonic
ultrasonic stimulated
stimulated thermography
thermography and
compared the analytical
compared the analytical resultsresults with experimental
experimental data data on a carbon-carbon composite sample with
BVID. Following
Followingthese experimental
these experimental investigations, numerical
investigations, models
numerical were also
models were developed to simulate
also developed to
UST and study qualitatively the crack heating phenomenon at damage location
simulate UST and study qualitatively the crack heating phenomenon at damage location caused by caused by internal and
external and
internal energy lossesenergy
external [112,113]. Derusova
losses et al. [114]
111. Derusova numerically
et al. modelled
113 numerically with finite
modelled element
with (FE) the
finite element
impact
(FE) thedamage
impactindamage
a carbon in
epoxy composite
a carbon as a composite
epoxy pyramid-like as defect with a set ofdefect
a pyramid-like delaminations
with a setacting
of
as air-filled heat sources. As illustrated in Figure 14, the authors found a good agreement
delaminations acting as air-filled heat sources. As illustrated in Figure 14, the authors found a good between the
experimental
agreement temperature
between variations intemperature
the experimental a defective area and thein
variations numerical results.
a defective area They alsonumerical
and the were able
to estimate
results. They the energy
also weregenerated by each
able to estimate thedefect
energy (i.e., the “defect
generated power”).
by each defect (i.e., the “defect power”).

(a) (b)
Figure 14. Modelling
Figure 14. Modellingofofimpact
impact damage
damage as aas a pyramid-like
pyramid-like defectdefect
shape shape
(a) and(a)
theand the comparison
comparison between
between experimental and numerical temperature results (b),
experimental and numerical temperature results (b), from [114].from 113.

Parvasi et al. 114 used a coupled thermo-electro-mechanical FE model to simulate the interaction
Parvasi et al. [115] used a coupled thermo-electro-mechanical FE model to simulate the interaction
of ultrasonic wave transmitted by low-power piezoceramic-based transducer with the material
of ultrasonic wave transmitted by low-power piezoceramic-based transducer with the material damage
damage and the associated temperature response. The authors used a static-kinetic exponential decay
and the associated temperature response. The authors used a static-kinetic exponential decay model to
model to simulate the frictional heating at crack interfaces. However, the exposure to high-power
simulate the frictional heating at crack interfaces. However, the exposure to high-power excitation
excitation generated by the acoustic/ultrasonic horn used in UST may even further degrade the
generated by the acoustic/ultrasonic horn used in UST may even further degrade the structural
structural integrity of the components 115. Zalameda et al. 116 developed a contactless ultrasonic
integrity of the components [116]. Zalameda et al. [117] developed a contactless ultrasonic stimulated
stimulated thermography with air-coupled transducers for damage detection of an helicopter blade
sandwich. The acoustic source consisted of an array of four amplified loudspeakers emitting pulses
at different frequencies ranging between 700 Hz and 1.3 kHz. The authors compared their results
with standard flash thermography and were able to reveal the presence of disbonds between the skin
and core.
Sensors 2018, 18, 609 18 of 37

thermography with air-coupled transducers for damage detection of an helicopter blade sandwich.
The acoustic source consisted of an array of four amplified loudspeakers emitting pulses at different
frequencies ranging between 700 Hz and 1.3 kHz. The authors compared their results with standard
flash thermography and were able to reveal the presence of disbonds between the skin and core.

3.1. Ultrasonic Stimulated Thermography Using the Local Damage Resonance Effect
The ultrasonic horn used in thermosonics is a bulky and crude means of exciting high-power
vibrations. The coupling between the test specimen and the horn typically results in an uncontrolled
generation of frequency components known as “acoustic chaos” [99]. Such a condition makes UST
“non-reproducible”, thus leading to cracks being undetected if sufficient vibrational energy is not
applied at the crack location. To overcome this issue, a new material elastic effect, known as local
resonance defect (LDR), has recently gained considerable attention as it allows selective ultrasonic
activation and higher sensitivity to the presence of structural flaws [118–120]. LDR shall be referred
as the interaction of acoustic/ultrasonic waves with the damaged area at a frequency matching the
defect resonance, which results in a substantial enhancement of the vibration amplitude only in the
localised damaged region [121]. Assuming that the internal structural flaw such as a delamination in
a composite laminate of thickness h is represented by a flat bottom hole, i.e., a thin circular defect of
radius r, the18,expression
Sensors 2018, x of the LDR frequency fd becomes [119]: 14 of 37

s
1.6d E
fd ∼= 1.6d E (13)
f d  r2 2 12ρ(1 − υ22)
12 1    (13)
r
where E and υ are the effective elastic modulus and Poisson’s ratio of the composite laminate,
where E and υ are the effective elastic modulus and Poisson’s ratio of the composite laminate,
respectively, ρ is the density and d is the depth of the portion of volume below the defect.
respectively,  is the density and d is the depth of the portion of volume below the defect. Equation
Equation (13) corresponds to the first bending mode of a circular plate with clamped boundaries. For a
(13) corresponds to the first bending mode of a circular plate with clamped boundaries. For a
quadratic-shaped
quadratic-shapeddefect
defectwith
withside length
side l, Equation
length (13) (13)
l, Equation becomes [121]:Error! Reference source not
becomes
found.: s
4πd E
fd ∼
= 2 (14)
4d 6ρ(1 E
3l − υ2 )
fd 
6 1   2 
(14)
3l 2
In the case of a notch of width w, the LDR frequency is [119]:
In the case of a notch of width w, the LDR frequency
s is 118:
2π Bs
fd ∼= 2 B (15)
f d  w2 3ρds (15)
w 2
3d
Ed3 /12
1 − ν22 . Figure 15 shows an example of LDR


where Bs is the bending stiffness given by Bs = 3
where Bs is the bending stiffness given by 𝐵𝑠 = 𝐸𝑑 /12(1 − 𝜈 ). Figure 15 shows an example of LDR
frequency (3.4 kHz) on a glass fibre reinforced plastic (GFRP) sample.
frequency (3.4 kHz) on a glass fibre reinforced plastic (GFRP) sample.

Figure 15. Vibration patterns at the GFRP specimen at 3.4 kHz (a) and at the LDR frequency, 20.9 kHz,
Figure 15. Vibration patterns at the GFRP specimen at 3.4 kHz (a) and at the LDR frequency, 20.9 kHz,
on a GFRP sample (b), with permission from 118.
on a GFRP sample (b), with permission from [119].

3.2. Nonlinear Ultrasonic Stimulated Thermography

LDR also exhibits transitions from linear to nonlinear regime with the effect of generating higher
and sub-harmonics of the input frequency 120. Such an effect is typically known as “nonlinear LDR”
Error! Reference source not found., and the combination of thermosonics with this effect is known
as “Nonlinear Ultrasonic Stimulated Thermography” (NUST). Solodov et al. 122 used NUST to
Sensors 2018, 18, 609 19 of 37

3.2. Nonlinear Ultrasonic Stimulated Thermography

LDR also exhibits transitions from linear to nonlinear regime with the effect of generating
higher and sub-harmonics of the input frequency [119,120]. Such an effect is typically known as
“nonlinear LDR” [121], and the combination of thermosonics with this effect is known as “Nonlinear
Ultrasonic Stimulated Thermography” (NUST). Solodov et al. [122] used NUST to improve the
driving frequency-selection process for the generation of frictional heat on a composite sample using
commercial loudspeakers. They matched the NUST thermal results with those obtained using a
shearography system. Fierro et al. developed a nonlinear narrow sweep excitation method for NUST
in order to detect delamination fatigue damage on aluminium sample [123] and debonds on an
aerospace composite stiffener panel (see Figure 16) [28]. In the last paper, the authors developed a
coupled thermal-structural finite element model to analyse the generation of frictional heat and the
associated nonlinear elastic effects on the composite sample. Good agreement between the numerical
and experimental results was found. Recently, Dionysopoulos et al. [124] used NUST with frequency
modulated signals generated by the crack in order to image barely visible impact damage on an
aerospace composite
Sensors 2018, 18, x panel. 15 of 37

.
Figure 16. Nonlinear ultrasonic stimulated results thermography on a composite stiffened panel while
16. Nonlinear
Figure conducting ultrasonic stimulated results thermography on a composite stiffened panel while
a sweep between 20 kHz and 30 kHz, with permission from 28.
conducting a sweep between 20 kHz and 30 kHz, with permission from [28].
4. Eddy Current Stimulated Thermography
4. Eddy Current Stimulated
Eddy Current Thermography
Stimulated Thermography (ECST) induces eddy currents in a conductive material
via aCurrent
Eddy coil andStimulated
the associated induction heating
Thermography is measured
(ECST) induceswith
eddy an currents
IR camerain125. When eddymaterial
a conductive
currents encounter a material damage (discontinuity), they are forced to divert and change the
via a coil and the associated induction heating is measured with an IR camera [125]. When eddy
direction of the current flow. This leads to areas of increased and decreased eddy current density
currentsresulting
encounter a material
in relatively hot damage (discontinuity),
and cool areas they areTypically,
due to Joule heating. forced toa divert and change
high-current the direction
eddy current
of the current flow. This leads to areas of increased and decreased eddy current
pulse with typical central frequency ranging between 150 and 450 kHz is used for excitation (also density resulting
in relatively hot and cool areas due to Joule heating. Typically, a high-current eddy current pulse
known as “pulsed ECST”). The excitation period can change from few milliseconds for high-
conductivity
with typical central materials (e.g., ranging
frequency metals) tobetween
few seconds150forand
low-conductivity specimens
450 kHz is used (e.g., plastics(also
for excitation and known
carbon fibre reinforced plastic laminates). Figure 17 shows the experimental set-up for pulsed ECST.
as “pulsed ECST”). The excitation period can change from few milliseconds for high-conductivity
materials (e.g., metals) to few seconds for low-conductivity specimens (e.g., plastics and carbon fibre
reinforced plastic laminates). Figure 17 shows the experimental set-up for pulsed ECST.
Because of its working principle, in pulsed ECST, the application of heat is not limited to
the sample surface, like in optical pulsed thermography, and it can reach a certain depth. For a
homogeneous field excitation parallel to the surface, the penetration depth of a magnetic field in a
material is governed by the skin effect. The skin depth or penetration depth, δ is given by [126]:

1
δ= p (16)
f πm p σ

where f is the frequency of the pulse excitation and mp and σ are the magnetic permeability (H/m)
and electrical conductivity of the material (S/m), respectively. Typical electrical conductivity of

Figure 17. Experimental set-up of pulsed eddy current stimulated thermography, with permission
from 125.
Sensors 2018, 18, 609 20 of 37

unidirectional single layered CFRP is approximately 5 MS/m in the longitudinal direction and 1 kS/m
in the transversal one [127]. Moreover, inductors used for ECST of composites operate at high excitation
frequencies from 100 kHz to 100 MHz [128]. Hence, the skin depth for CFRP with small conductivity
(~1 kS/m) and no magnetic permeability is significantly great (~50 mm under 100 kHz excitation) [129].
Sensors 2018, 18, xto [54] the Green’s function formulation, the estimation of the temperature profile
According 15 of 37 in

ECST for an infinite length and finite thickness plate can be expressed via the following two equations,
for the heating phase ([29,130]):
√ s
Q2 t t
T (0, t) = √ ∝ ar for t < t p (17)
ρc πα tp

and the cooling phase:

∞
( )
n2 π 2

Q 1
1 + 2 ∑ exp − 2 α t − t p ∝ q


T (0, t) =
for t < t p (18)
ρcL n =1 L an t − t p

where Q = | Js |2 t/σ is the generated resistive heat with Js the eddy current density, tp is the duration
of the heating pulse and L. is the thickness of the sample. The coefficient ar is the amplitude of
temperature
Figurerise determined
16. Nonlinear by local
ultrasonic electricresults
stimulated conductivity variations
thermography at the damage
on a composite stiffenedlocation, whilst an
panel while
is the normalized
conducting atemperature
sweep between decay rate,
20 kHz anddetermined
30 kHz, withby local thermal
permission property changes and material
from 28.
defect dimensions. According to the theory of electromagnetic induction, some power is lost in the coil
and 4. cannot
Eddy Current
be usedStimulated Thermography
for excitation. Hence, the heating efficiency η c can be estimated as follows [131]:
Eddy Current Stimulated Thermography (ECST) sinduces !−eddy
1
currents in a conductive material
via a coil and the associated induction heating2hisc measured σµ pc with an IR camera 125. When eddy
ηc ≈ 1 + (19)
currents encounter a material damage (discontinuity), rc µpσthey
c are forced to divert and change the
direction of the current flow. This leads to areas of increased and decreased eddy current density
resulting
where in relatively
hc is the distancehot andcoil
of the coolform
areasthe
duematerial
to Joule surface
heating. (also
Typically,
known a high-current eddy current
as “lift-off distance”), rc is
thepulse with
radius typical
of the coil,central
µpc andfrequency
σc are theranging
magneticbetween 150 and and
permeability 450 electrical
kHz is used for excitation
conductivity (also
of the coil,
known as “pulsed ECST”). The excitation period can change from few milliseconds
respectively. Equation (19) shows that to increase the heat efficiency, the permeability of the coil shouldfor high-
be conductivity materialsto(e.g.,
as close as possible 1 andmetals) to few seconds
the electrical for low-conductivity
conductivity specimens
of the coil should be as(e.g.,
highplastics and
as possible.
carbon fibre reinforced plastic laminates). Figure 17 shows
Typical values of hc and rc are 2–10 mm and 2–2.5 mm, respectively. the experimental set-up for pulsed ECST.

Figure
Figure 17.17. Experimental
System set-up of(a)
set-up diagram pulsed eddy current stimulated
and experimental thermography,
set-up (b) of pulsed ECST,with
withpermission
permission
from 125.
from [125].

Because of its working principle, in pulsed ECST, the application of heat is not limited to the
sample surface, like in optical pulsed thermography, and it can reach a certain depth. For a
homogeneous field excitation parallel to the surface, the penetration depth of a magnetic field in a
material is governed by the skin effect. The skin depth or penetration depth, δ is given by 126:
Sensors 2018, 18, 609 21 of 37

Equation (19) is generally applicable to a wide range of materials and material failure scenarios for
aerospace applications, including corrosion on aluminium samples [132,133] and crack/delamination
in carbon fibre composites [134]. This last material has complex electromagnetic properties in which
electrically conductive carbon fibres are embedded into a dielectric resin matrix. Since the electrical
and thermal conductivity are the greatest along the fibre direction, recent advances of ECST involve
the determination of buried defects in the dielectric phase of composite materials (resin), such as
moisture ingress, aging of the material due to service or environmental/thermal exposure, voids and
delamination [134].
He et al. [135] and Liang et al. [136] used the ECST combined with the principal components
thermography (PCT) to detect material damage on CFRP samples impacted with different energies
between 4 J and 12 J (Figure 18).
Sensors 2018, 18, x 17 of 37

Figure 18. ECST results combined with PCT on a CFRP composite sample impacted at 6 J (a); 8 J (b);
Figure 18. ECST results combined with PCT on a CFRP composite sample impacted at 6 J (a); 8 J (b);
10 J (c) and 12 J (d), from 135.
10 J (c) and 12 J (d), with permission from [135].

Ishikawa et al. 137 and He and Yang 138 converted thermal ECST data from time domain to
Ishikawa
frequency et al.and
domain [137]used
andphase
He and Yang in
images [138] converted
order thermal
to increase ECST data from
the detectability time domain
of delamination in
to frequency domain and used phase images in order to increase the detectability
CFRP sample. Since the influence of non-uniform heating is periodically suppressed in phase images,of delamination
in
theCFRP
authorssample.
foundSince thedamage
that the influence of non-uniform
imaging resolution heating is periodically
(accuracy) was enhanced. suppressed in phase
Renil Thomas and
images, the authors found that the damage imaging resolution (accuracy) was
Balasubramaniam 139 developed a scanning induction method for ECST on impact damaged CFRP enhanced. Renil Thomas
and Balasubramaniam
samples, in which the[139] developed
induction a scanning
coil moved overinduction
the sample method
at thefor ECSTofon~40
speed impact
mm/sdamaged
and the
CFRP samples, in which the induction coil moved over the sample at the speed
temperature profiles caused by Joule heating were captured with an IR camera. The final ECST of ~40 mm/simage
and
the
wastemperature
compared withprofiles caused by
a standard Joule heating
ultrasonic C-Scanwere captured
showing with
good an IR camera.
agreement The final
in terms ECST
of damage
image was compared with a
location and severity (Figure 19). standard ultrasonic C-Scan showing good agreement in terms of damage
location and severity (Figure 19).
Yang and He [129] combined a selective heating thermography through electromagnetic induction
with cross correlation match filtering for the inspection of damaged CFRP composites. This technique
reduced the non-uniform heating and lateral blurring effects that are typical of ECST.
ECST can be easily automated/robotised and has shown high sensitivity for damage quantification
on dry, wet and consolidated CFRP. Moreover, since eddy currents flow parallel to the surface, it is
very complex to detect delamination in CFRP unless there are interlaminar fibre contacts. To overcome
this issue, an alternative technique recently developed for aerospace components is the Capacitive
Imaging (CI) method [140], which is based on the variation of dielectric properties of the resin matrix

Figure 19. Comparison between eddy current stimulated thermography (a) and ultrasonic C-Scan (b)
on impact damaged CFRP sample, with permission from 139.
 Ishikawa et al. 137 and He and Yang 138 converted thermal ECST data from time domain to
frequency domain and used phase images in order to increase the detectability of delamination in
CFRP sample. Since the influence of non-uniform heating is periodically suppressed in phase images,
the authors found that the damage imaging resolution (accuracy) was enhanced. Renil Thomas and
Balasubramaniam 139 developed a scanning induction method for ECST on impact damaged22CFRP
Sensors 2018, 18, 609 of 37
samples, in which the induction coil moved over the sample at the speed of ~40 mm/s and the
temperature profiles caused by Joule heating were captured with an IR camera. The final ECST image
in order
was to enablewith
compared detection of delamination
a standard ultrasoniccaused
C-Scanbyshowing
impact damage. Further studies
good agreement of CIof
in terms combined
damage
to thermography are under development
location and severity (Figure 19). to verify its effectiveness for aerospace applications.

Figure19.
Figure 19. Comparison
Comparison between
between eddy
eddy current
current stimulated
stimulated thermography
thermography(a)
(a)and
and ultrasonic
ultrasonic C-Scan
C-Scan(b)
(b)
onimpact
on impactdamaged
damagedCFRP
CFRPsample,
sample,with
withpermission
permissionfrom
from[139].
139.

Yang andThermography
4.1. Microwave He 129 combined a selective heating thermography through electromagnetic
induction with cross correlation match filtering for the inspection of damaged CFRP composites. This
Microwave
technique reducedThermography
the non-uniform(MWT) uses
heating electromagnetic
and lateral blurringradiation with
effects that are frequencies ranging
typical of ECST.
between 300 MHz and 30 GHz (microwaves) in order to generate heat due
ECST can be easily automated/robotised and has shown high sensitivity for damage to the dielectric loss
of materials [141].
quantification These
on dry, wetfrequencies are chosen
and consolidated CFRP.by international
Moreover, since agreement
eddy currentsin order to minimise
flow parallel to the
interference
surface, it is very complex to detect delamination in CFRP unless there are interlaminar fibrethere
with aircarft communication services. The idea behind MWT is that where is a
contacts.
damage such as
To overcome delaminations,
this cracks and
issue, an alternative voids,recently
technique the water trapped in
developed forthe open defect
aerospace is released
components is for
the
vapour tension and the associated heat is measured by the IR camera. In MWT, the dissipated
Capacitive Imaging (CI) method 140, which is based on the variation of dielectric properties of the power
per unit volume P in dielectric media is provided by the following equation [19,142]:

P = 2π f ε 0 ε00 E2f (20)

where f is the frequency of the microwave excitation, Ef is the electric field, ε0 is the permittivity of the
air and ε00 is the loss factor that quantifies the power dissipation. The temperature change at heating
time t by power dissipation of continuous microwave is [142]:

P
T (t) = (21)
ρc

From Equation (21), it can be seen that the temperature rise is approximately linear with time.
The main advantage of the MWT is its volumetric character, meaning that a certain volume of the
specimen can be heated at once, thus leading to a very fast inspection of large parts. However,
according to Equations (20) and (21), the heating ratio is dependent not only on the thermal properties
of the material, but also on its electrical properties. MWT has been used only recently for the inspection
of CFRP composites for aerospace applications [143]. Palumbo et al. [144] compared MWT with LIT
for the detection of impact damage on sandwich composites. The authors provided a quantitative
analysis of the damaged area by using a CT-Scan (X-ray tomography) and showed that both LIT and
MWT were able to provide similar results, whit the additional advantage of the latter technique being
much faster (only two seconds of heating and simple signal processing).

5. Material-Based Thermography
One of the key parameters in IRT is the selection of the thermal stimulation source. This selection
determines the physical constraints of the thermographic system such as: (i) the requirements of
clearance and accessibility of the component under examination; (ii) the power consumption in order
Sensors 2018, 18, 609 23 of 37

to comply with other aircraft/spacecraft systems; (iii) the inspection costs and (iv) the limitations
in terms of resolution of the analysis and its effectiveness in identifying in-depth defects. During
the last two decades, beside optical thermal stimulation, ultrasonic excitation and electromagnetic
radiation, several alternative stimulation sources have been investigated to enhance the resolution
of the thermal analysis while reducing the power requirement [19]. In particular, by analysing the
results obtained with all these different excitation techniques, it is possible to identify three major
drawbacks linked with the use of external thermal stimulation. Firstly, the need of uniform thermal
heating involves accurate position of the external heat source, which, inevitably, limits in-situ analysis.
Secondly, there is a need for cost-effective thermal excitation systems that do not require high input
power. Thirdly, thermal responses in fibre composite laminates, due to their anisotropic structure
behaviour, are generally affected by a high level of noise that needs to be filtered with advanced signal
processing techniques in order to avoid misdetection of damaged areas or false positives.
Based on these considerations, a number of authors has investigated alternative strategies to
current external thermal excitation sources in composites. In this section, we identify two different
approaches, the “direct material-based thermography” (DMT) and the “indirect material-based
thermography” (IMT). The DMT method consists of exploiting the specific properties of the material
under analysis in order to assess its internal health status. This can be done in conductive metal parts
and composite laminates by using one of the components (e.g., the resin matrix or the reinforcement
fibre) to generate heat via Joule effect. The IMT approach, instead, involves the embodiment of
additional thermoresistive components within the composite laminate during the manufacturing
process in order to generate internal thermal stimulation. In both approaches, the material itself is
able to autonomously assess its internal health state without the use of external excitation heaters or
flash lamps and, similarly to traditional IRT, the heat flow interacting with the material damage can be
detected with an IR camera.

5.1. Direct Material-Based Thermography

Sagami et al. [31] developed two different DMT techniques based on the application of electrical
current on electrically conductive samples and the associated generation of heat via Joule effect.
The first technique, called “singular method” (see Figure 20a), is based on the increase of the current
density near the tip of a intralaminar crack (oriented at 90◦ ) in a stainless steel plate as result of the
interaction with the stream line of an applied DC current. This increase, in turn, results in a raise of
the temperature field with a significant heat concentration at the crack tip that can be easily detected
from the top surface with an IR camera. Further experiments with cracks at 0◦ , 30◦ , 45◦ and 60◦ were
carried out by observing a less significant heat concentration with the reduction of the inclination
angle due to the decreasing behaviour of the current intensity factor. The “singular method” depends
on the intensity of the applied current field and on the thermal diffusion rate. Indeed, high values of
thermal conductivity may lead to poor detection of flaws when the analysis is not conducted in a short
inspection time.
The second approach, named “insulation method” (see Figure 20b), was designed for the detection
of interlaminar flaws in CFRP composites. In this case, the crack tip does not generate any singularities
in the temperature field but the defect acts as an insulator to the thermal conduction, thus resulting in a
low-temperature region detectable from the top surface. This technique was experimentally validated
by applying a DC current to a twelve layers CFRP sample (500 × 100 mm) with an internal patch
of polyimide film. The electrical connections were obtained by removing the resin matrix from two
100 × 100 mm areas on the two edges. Experimental results showed that both DMT methods are able
to detect the presence of internal flaws. However, high input currents (i.e., ~20 A for the “singular
method” and ~40 A for the “insulation method”) are needed.
 singularities in the temperature field but the defect acts as an insulator to the thermal conduction,
thus resulting in a low-temperature region detectable from the top surface. This technique was
experimentally validated by applying a DC current to a twelve layers CFRP sample (500 × 100 mm)
with an internal patch of polyimide film. The electrical connections were obtained by removing the
resin matrix from two 100 × 100 mm areas on the two edges. Experimental results showed that both
SensorsDMT 18, 609
2018, methods are able to detect the presence of internal flaws. However, high input currents (i.e., 24 of 37
~20 A for the “singular method” and ~40 A for the “insulation method”) are needed.

Figure 20. Illustration of the “singular method” (a) and the “insulation method” (b), from 31.
Figure 20. Illustration of the “singular method” (a) and the “insulation method” (b), with permission
from [31].
Sensors 2018, 18, x 20 of 37

5.1.1.5.1.1.
Electrical Resistance
Electrical Change
Resistance ChangeMethod
Methodand
and Thermography
Thermography
The The
ideaidea
of exploiting
of exploiting the
theconductive
conductive properties
properties ofofcarbon
carbonfibres
fibresin in
CFRPCFRP composites
composites for the
for the
autonomous
autonomous assessment
assessment of their
of their internal
internal healthhealth
statusstatus has been
has been investigated
investigated in thein“electrical
the “electrical
resistance
change resistance
method” change
(ERCM) method” (ERCM)
[145–147]. In 145. In this technique,
this technique, carboncarbon fibres
fibres are areas
used used as “sensors”
“sensors” exploiting
exploiting their electrical conductivity and variations in the electrical resistance
their electrical conductivity and variations in the electrical resistance are monitored to detect cracks are monitored to or
detect cracks or indentation damage. Although ERCM was proved to be efficient in detecting early
indentation damage. Although ERCM was proved to be efficient in detecting early stage material flaws,
stage material flaws, its reliability is not high enough for practical applications. This is mainly caused
its reliability is not high enough for practical applications. This is mainly caused by: (i) fluctuations of
by: (i) fluctuations of the electrical conductivity of laminated composites; (ii) the nature of electrical
the electrical
connectionsconductivity
that require of laminated composites;
a large amount (ii) the
of wiring fornature of electrical
complex structuresconnections
and have that require
strong
a largedependence with temperature and (iii) the limitation of ERCM for damage detection in thick and
amount of wiring for complex structures and have strong dependence with temperature
(iii) the limitation of ERCM for damage detection in thick laminates.
laminates.
Suzuki et al.
Suzuki et [148]
al. 148recently
recentlycombined ERCMwith
combined ERCM withthermography
thermography in order
in order to detect
to detect damage
damage in in
aerospace components. An electrical voltage was applied on the two outermost
aerospace components. An electrical voltage was applied on the two outermost layers of a composite layers of a composite
laminate
laminate andvariations
and the the variations
of theoftemperature
the temperature field observed
field were were observed
with an with an IR camera.
IR camera. Since
Since indentation
indentation damage increases the contact between the fibres of two separate layers and reduces the
damage increases the contact between the fibres of two separate layers and reduces the matrix interphase,
matrix interphase, the intralaminar electrical resistivity in the damage area decreases, thus leading to
the intralaminar electrical resistivity in the damage area decreases, thus leading to a local increase of the
a local increase of the current density that can be identified on the material’s surface as a hot spot
current density that can be identified on the material’s surface as a hot spot (Figure 21).
(Figure 21).

Figure 21. Schematic of the technique developed by Suzuki et al.: difference in the path of the electrical
Figure 21. Schematic of the technique developed by Suzuki et al.: difference in the path of the electrical
current in the case of undamaged and indented areas, with permission from 148.
current in the case of undamaged and indented areas, with permission from [148].
The authors evaluated the electrical properties of an IM600/133 CFRP composite laminate both
experimentally and numerically to test the effectiveness of the proposed ERCM combined to
thermography. Particularly, undamaged, indented and delaminated CFRP samples were
investigated using an input voltage ranging between 5 and 10 V applied on two electrodes. Results
showed that delamination of 0.15 mm could be easily detected by applying 5 V, thus generating a
Sensors 2018, 18, 609 25 of 37

The authors evaluated the electrical properties of an IM600/133 CFRP composite laminate
both experimentally and numerically to test the effectiveness of the proposed ERCM combined to
thermography. Particularly, undamaged, indented and delaminated CFRP samples were investigated
using an input voltage ranging between 5 and 10 V applied on two electrodes. Results showed that
delamination of 0.15 mm could be easily detected by applying 5 V, thus generating a temperature
variation of more than 30 ◦ C between damaged and undamaged areas. As for internal defects, the
technique was not able to detect the presence of an internal delamination, showing only a slight
decrease of the top surface temperature in correspondence with the damaged area. An advantage of
this method is that traditional aircraft lightening protection strips can be used as electrical contacts and
wiring, thus simplifying the manufacturing procedure and reducing production costs, as demonstrated
in a further work by the same authors [149].
Grammatikos et al. [150] also combined ERCM with thermography and compared this method
with different optical thermographic techniques including PT, PPT and LIT in order to monitor the
structural integrity of aircraft composite parts. The internal heat stimulus was provided by a square
electric pulse at low frequency applied to the bulk laminate in order generate a diffuse thermal field
via Joule heating. The authors assumed that if an undamaged carbon layer is characterised by a
resistance R1 , a damaged area will be characterised by a higher resistance R1 = nR2 (with n > 1) due to
the presence of the internal flaw (e.g., crack, delamination or BVID). By considering the 1st law of
thermodynamics and both the Ohm’s and Joule’s law, the electrical power P1 and P2 for both an
undamaged and damaged composite can be expressed as follows:

P1 = I 2 R1 and P2 = I 2 R2 = I 2 R1 n (22)

with I the input current. Assuming that I is a constant, the increase in the electrical resistance in the
damaged area will lead to an increase of the electric power since P1 = nP2 , which translates in a
temperature increase at damage location. To validate this finding, experimental data were collected by
testing CFRP square coupons 60 × 60 × 1 mm and electrical connections were obtained by removing
resin from both the edges and applying silver paint and silver loaded adhesive tape, showing good
results in terms of accuracy.
In addition, since thermal waves propagate through the thickness of the sample, the
thermal-electrical properties along the through-thickness direction constitute a key parameter for
the effectiveness of ERCM and thermography. In particular, the electrical conductivity along the
sample’s thickness is given by the random interlaminar fibre-fibre connections that lead to a pattern
of locations with higher conductivity. This, in turn, determines the anisotropy of the heat flow
propagation through the laminate. As a consequence, in order to increase the number of random
contacts and enhance the heating front propagation, 0.5% w/w of carbon nanotubes (CNTs) was
included within the epoxy matrix [151]. The nanomodified samples were subjected to low velocity
impacts at different energies (3 and 4 J) and the material damage was evaluated in “live mode” by
first recording the surface temperature during the application of electrical current at 10 A for 60 s
and then during the cooling down process for other 150 s. Results from these tests showed that the
difference in the temperature field between damaged and undamaged areas was higher during the
heating phase rather than the cooling one. The authors explained these results with the different
mechanisms of the two processes: whilst during heating the heat flow is directed through the injection
path, during the cooling phase part of the thermal energy is dissipated through the edges of the
monitored surface. However, these observations seem to be in contrast with what is usually seen in
traditional thermography with external heat sources as cooling ramps are more efficient in detecting
internal flaws [152]. In order to enhance the resolution of the results achieved in “live mode” and,
simultaneously, keep the power consumption to lower levels, PPT was employed by connecting the
camera with a pulse generator and applying 200 mA (0.5 V) to the sample for 50 s. The comparison
between the PPT phase images and traditional C-scans showed that in case of CNT-modified samples,
Sensors 2018, 18, 609 26 of 37

ERCM combined to thermography was able to reveal the presence of low velocity impacts through the
identification of cold-spots on the surface.

5.2. Indirect Material-Based Thermography

Although DMT has demonstrated its effectiveness as material enabled NDE technique, electrical
connections applied directly to fibres may lead to thermomechanical stresses especially on the edges of
the composite sample if the power required for the inspection or the time of the analysis are very high.
In addition, DMT only works if the material is characterised by good electro-thermal properties, thus
ruling out completely any glass fibre composite application. IMT is, conversely, based on a different
approach than DMT, in which a thermoresistive phase is added within the material structure and acts
as an embedded heat source that can be used for both structural and non-structural purposes [34].

5.2.1. Metal-Based Thermography

Ahmed et al. [32] developed a technique called “Heat Emitting Layer” (HEL) for Metal-based
Thermography (MT), which was based on the development of an engineered additional layer fully
integrated into the laminate structure in order to guarantee a fast inspection of large areas. The HEL
was manufactured using AISI316 type stainless steel wires with a diameter of 70 µm woven into a
satin 8H glass fabric at specific intervals and was then used within the lamination sequence of several
GFRP laminates in which polyamide (PI) foils where embedded to simulate the presence of internal
Sensors 2018, 18, x 22 of 37
defects (Figure 21).
As itit is
As is possible
possible to to see
see from
from Figure
Figure 22,
22, the
the experimental
experimental campaign
campaign was was aimed
aimed to to investigate
investigate
several aspects of MT such as the inter-distance between stainless steel
several aspects of MT such as the inter-distance between stainless steel wires within the HEL wires within the HEL layer, its
layer,
position along the laminate’s thickness and the resolution of the technique with
its position along the laminate’s thickness and the resolution of the technique with increasing distance increasing distance
between the
between the thermal
thermalemitting
emittinglayer
layerand thethe
and embedded
embedded defect. UponUpon
defect. an accurate analysis
an accurate of MT thermal
analysis of MT
results, the authors conclude that the fibre spacing played an important role
thermal results, the authors conclude that the fibre spacing played an important role in the resolution in the resolution of the
thermogram. Indeed, by decreasing the inter-wire distance from 10 to 2.5
of the thermogram. Indeed, by decreasing the inter-wire distance from 10 to 2.5 mm, the inspectionmm, the inspection produced
a more defined
produced a more view of theview
defined PI insert duePItoinsert
of the the higher
due to contrast between
the higher wiresbetween
contrast and the wires
damaged and area.
the
In addition, MT experimental results and numerical data collected from finite
damaged area. In addition, MT experimental results and numerical data collected from finite element element analysis showed
that by placing
analysis showedthe HEL
that byon the bottom
placing the HELof the
onlaminate,
the bottom the propagation
of the laminate, of the thermal
the front through
propagation of the
the thickness was more uniform, thus guaranteeing the highest resolution.
thermal front through the thickness was more uniform, thus guaranteeing the highest resolution. However, because of
the distance from the top surface, the difference in temperature between
However, because of the distance from the top surface, the difference in temperature between damaged and undamaged
areas wasand
damaged 0.5 ◦ C, which
onlyundamaged means
areas wasthat
onlyhigh
0.5 injection
°C, which currents
meanswerethat required. In particular,
high injection currents whenwere
low power was used on samples with different thickness (~0.02 W/cm 2 for the 6 plies laminate and
required. In particular, when low power was used on samples with different thickness (~0.02 W/cm2
~0.13 2 for the 24 plies one) not all2 the inserts could be detected even when the excitation current
for theW/cm
6 plies laminate and ~0.13 W/cm for the 24 plies one) not all the inserts could be detected even
was applied for 10 s.current
By increasing the input power 2 , most of defects became visible
when the excitation was applied for 10 s. Bytoincreasing
~0.19 W/cm the input power to ~0.19 W/cm2, most for
2
of defects became visible for the 6 plies laminates and two out of three patches were detectable for).
the 6 plies laminates and two out of three patches were detectable for the 24 layers one (~0.42 W/cm
It is24
the important
layers one to (~0.42
highlight
W/cm that
2). when MT is applied
It is important on structures
to highlight that when moreMT complex thanon
is applied plates, other
structures
parameters
more complex suchthanas the waviness
plates, degree of the
other parameters stainless
such as thesteel wiresdegree
waviness or the reflective surfacesteel
of the stainless of curved
wires
parts
or thebecome
reflective fundamental for the correct
surface of curved detection
parts become of flaws and
fundamental forneed to be taken
the correct in account
detection in order
of flaws and
to enhance
need the effectiveness
to be taken in account in oforder
this technique.
to enhance the effectiveness of this technique.

Figure
Figure 22.
22. HEL
HEL laminates:
laminates: Schematic
Schematiclay-up
lay-upofofthe
theHEL
HELlaminates
laminatesand
andspecifications
specificationsof
ofthe
the different
different
laminates
laminates prepared
prepared during
during the
the experimental
experimental campaign,
campaign, with
with permission
permission from
from 32.
[32].

Following a similar approach, Orlowska et al. 153 proposed a further development of MT by

embedding a specially designed 3D electrical grid composed of through-the-layer (vertical) and
surface-layer (horizontal) metal elements. The grid was designed so that its mechanical properties
where the same as the composite material and, in case of loss of adhesion between two layers in a
localised portion of the structure, the embedded grid could fracture simultaneously and break the
Sensors 2018, 18, 609 27 of 37

Following a similar approach, Orlowska et al. [153] proposed a further development of MT

by embedding a specially designed 3D electrical grid composed of through-the-layer (vertical) and
surface-layer (horizontal) metal elements. The grid was designed so that its mechanical properties
where the same as the composite material and, in case of loss of adhesion between two layers in a
localised portion of the structure, the embedded grid could fracture simultaneously and break the
conductive mesh. This local disruption of the electrical field generates a decrease in the thermal
density at damage location that can be experimentally observed by a long-wave IR camera. Numerical
finite element analysis was also conducted by considering different thicknesses of the sample and
various dimensions of the internal delamination (represented by an embedded Teflon patch). Results
indicated that the temperature difference between a damaged and undamaged area was strongly
affected by the position along the thickness of the delamination, thus showing a temperature difference
of only 0.2 ◦ C in the case of a patch inserted within a 24 layers sample. The extension of the patch also
represented a key parameter for MT and could potentially constitute a limitation by showing only
0.1 ◦ C of temperature difference in case of a 6 × 6 mm embedded patch. By using a non-stationary
thermal excitation with a rectangular pulse of 20 s, temperature differences were enhanced up to
3.63 ◦ C. This method was validated experimentally by manufacturing a 8 layers GFRP laminate with
a grid of Aluchrome wires with 120 µm of diameter. In order to embed the conductive mesh within
the laminate, tiny holes were drilled within the thickness of the sample and the resistive elements
were placed in three parallel lines. One of the connection was left intentionally detached in order to
simulate the disruption of the conductive grid generated by an internal delamination and a current
of 1.1 A was applied to the entire structure. Results from the thermal image demonstrated a good
accuracy of the method in terms of damage detection. However, as the authors acknowledged in the
paper [153], in order to avoid the weakening of the material due to the drilling operation, further work
is still required in order to design a different manufacturing procedure in which the conductor wires
are knitted on the different fibre layers before the impregnation with the resin.

5.2.2. Carbon Nanotubes-Based Thermography

The simultaneity of damage in both the embedded sensing device/mesh/grid and the sample
itself may limit the sensitive of MT, as false positives could be identified in case of damaged sensor
and no structural damage. To overcome this limitation, one of the most promising approaches is to
intend the material-enabled thermography feature as one of the multiple functions of a more complex
multi-physics material system. In other words, the additional phase within the laminate’s structure
needs to be fully integrated with the existent phases in order to work not only as a sensing device but
also as a combined structural reinforcement, thus leading to a multifunctional material system whose
structural and non-structural properties results enhanced in comparison with a traditional laminate.
As seen in Section 5.1.1, the inclusion of CNTs within a resin matrix enhances the thermomechanical
properties and, in turn, increases the resolution of carbon nanotubes-based thermography (CNTT)
when a current is applied at the edges of a CFRP laminate. By following a similar approach, de
Villoria et al. [33] developed a CNTT technique termed by the authors “Nano-Engineered Thermal
NDE” (NET-NDE) by manufacturing hybrid alumina fabrics with aligned CNTs forest. The 30 µm
multi-walled nanotubes were grown through chemical vapour deposition on the surface of the satin
cloth following a “Mohawk” morphology and were characterised by a diameter of 17 nm (Figure 23).
Composite specimens were manufactured via hand lay-up impregnating three layers of fabric with
epoxy matrix. The final laminates were characterised by 50 vol % of alumina fibres and 2 vol % of
CNTs. The presence of the nano-scale phase within the composite was able to bridge the ply interfaces,
thus enhancing both intralaminar and interlaminar strength, as demonstrated by previous works from
the same research group [154,155]. In order to test the efficiency of CNTT, an in-situ thermal analysis
was performed by applying a 6.8 V (2 W) on the edges of an hybrid sample with a 6.5 mm hole subject
to a failure tensile test. Results showed that superficial cracks opening and stress concentration areas
Composite specimens were manufactured via hand lay-up impregnating three layers of fabric with
epoxy matrix. The final laminates were characterised by 50 vol % of alumina fibres and 2 vol % of
CNTs. The presence of the nano-scale phase within the composite was able to bridge the ply
interfaces, thus enhancing both intralaminar and interlaminar strength, as demonstrated by previous
works from the same research group 154. In order to test the efficiency of CNTT, an in-situ thermal
Sensors 2018, 18, 609 28 of 37
analysis was performed by applying a 6.8 V (2 W) on the edges of an hybrid sample with a 6.5 mm
hole subject to a failure tensile test. Results showed that superficial cracks opening and stress
could be easilyareas
concentration followed viabe
could thermographic inspection
easily followed by monitoring
via thermographic the variation
inspection of the thermal
by monitoring the
gradient on the damaged areas.
variation of the thermal gradient on the damaged areas.

Figure 23.
Figure 23. CNT/Alumina
CNT/Alumina hybrid
hybrid composite:
composite: internal
internal nanoscale
nanoscale interaction
interaction between alumina fibres
between alumina fibres
and carbon nanotubes, with permission from 154.
and carbon nanotubes, with permission from [154].

Good results were also observed from tests conducted on a riveted panel with internal cracks
Good results were also observed from tests conducted on a riveted panel with internal cracks not
not visible from the top surface using input voltages of 0.16 V and 12.V, and on damaged composite
visible from the top surface using input voltages of 0.16 V and 12.V, and on damaged composite plate
plate undergo to low velocity impact by applying 0.34 V on both ends.
undergo to low velocity impact by applying 0.34 V on both ends.

5.2.3. Shape Memory Alloys-Based Rapid Thermography

In order to develop a fully multifunctional material based on carbon composites that can be used
for aerospace applications, another approach is based on the development of a new hybrid material that
includes Shape Memory Alloy (SMA) wires within the lamination sequence of a traditional CFRP parts.
CFRP hybridisation with SMAs has been investigated by several authors in order to exploit the unique
properties of superelasticity and shape memory effect of SMAs to increase impact resistance and reduce
the extent of internal delamination [156,157]. These properties rise from the transitions between two
different crystalline structures (martensite and austenite) that can be activated by applying temperature
gradients or by loading the material with an external force. By combining the enhancement of the
mechanical properties with strain sensing, damage localisation and de-icing features, Pinto et al. [34]
developed a new multifunctional material embedding SMA wires within traditional CFRP composites.
This technique, named Shape Memory Alloys-based rapid Thermography (SMArT), was based on the
generation of an in situ low-power resisting heating that allowed the detection and the assessment of
structural damage in aerospace composite structures. Experiments to validate SMArT were undertaken
on several samples in order to analyse different variables such as dimensions of the embedded damage
(20 × 20 mm, 10 × 10 mm and 5 × 5 mm), resolution of the analysis, effect of the feeding current
intensity (0.1–0.9 A) and relative position of the wires within the sample’s thickness. Results from the
experimental campaign showed that the presence of the embedded hybrid phase was able to guarantee
uniform heating flow through the sample’s thickness. When thermal waves propagating from the
SMA interact with the internal defect, they are slowed down due to the different thermal diffusion
rate and generate a localised variation in the apparent temperature on the sample’s surface that can be
detected with an IR camera.
From a mathematical point of view, after a uniform impulse of thermal energy is released
instantaneously by the SMA wire at time t = 0 and propagates along the through-the-thickness
direction, the 1D analytical model of the temperature rise Tnd (h,t) on the surface of an undamaged
material is given by [158]:

∞ ∞
" #
Q0 (2mL+h)2 [2( m +1) L − h ]2
Tnd (h, t) =
4πkt ∑ e− 4αt + ∑ e− 4αt (23)
m =0 m =0
Sensors 2018, 18, 609 29 of 37

where h is depth location of the wire and L is the thickness of the composite plate. Assuming a
delamination-like defect layer of thickness d lying beneath the material surface in a plane parallel to
the surface, the temperature rise Td (h,t) on the surface of the defective material becomes:

∞ ∞ ∞ ∞
" #
Q0 (2mL+h+2nd)2 [2(m+1) L−h+2nd]2
Td (h, t) =
4πkt ∑ (1 − Γ ) 2m+1
∑Γ n −
e 4αt + ∑ (1 − Γ ) 2m+1
∑Γ n −
e 4αt (24)
m =0 n =0 m =0 n =0

where Γ is the effective thermal reflectivity of the defect. For a circular defect of diameter D, the
diffusion distance is D/2 and the temperature rise on the surface of the defective material can be
rewritten as [158]:
D2
T (h, t) = Td (h, t) + [ Tnd (h, t) − Td (h, t)]e− 16αAt (25)

where A is the thermal diffusivity anisotropy of the composite material, which is equal to 1 for
thermally isotropic materials. The exponential term in Equation (25) accounts for the physics of the
diffusion of heat from the edge of the circular defect to the centre at a distance D/2 away. Particular
attention was given by Pinto et al. [34,159] to determine the depth of embedded Teflon patches in
multi-damage samples. This was accomplished by studying both magnitude (Figure 24a) and phase
of thermal waves (Figure 24b). Low modulation frequencies (between 0.01 and 2 Hz) were used to
observe deep lying defects and to provide information about the depth variation between multiple
damage sites.
Sensors 2018, 18, x 25 of 37

(a)

(b)
Figure 24. SMArt multifunctional laminate: time history of the temperature difference variation
Figure 24. SMArt multifunctional laminate: time history of the temperature difference variation
between damaged and undamaged part (4 plies CFRP in cross-ply stacking sequence, single damage)
between damaged and undamaged part (4 plies CFRP in cross-ply stacking sequence, single
(a); phase profile of two different damage sites (16 plies, cross-ply stacking sequence, multi-damage)
damage) (a); phase profile of two different damage sites (16 plies, cross-ply stacking sequence,
(b), with permission from 34.
multi-damage) (b), with permission from [34].

Furthermore, in order to exploit the multifunctionality of the material in real applications,

Furthermore, in order to exploit the multifunctionality of the material in real applications,
particular attention was given to the design of a connection system able to be integrated into pre-
particular attention was given to the design of a connection system able to be integrated into
existing structures. This system needed to be characterised by high level of flexibility and a good
pre-existing structures. This system needed to be characterised by high level of flexibility and a
trade-off between resolution of the analysis and power consumption. Hence, each wire of the SMA
good trade-off between resolution of the analysis and power consumption. Hence, each wire of the
grid was connected to a circuit board in order to regulate the resolution of the thermal images
SMA grid was connected to a circuit board in order to regulate the resolution of the thermal images
acquired with the IR camera. By powering directly the circuit board with a fixed amount of power, it
acquired with the IR camera. By powering directly the circuit board with a fixed amount of power, it
was possible to divert a larger amount of current through a specific portion of the sample by reducing
was possible to divert a larger amount of current through a specific portion of the sample by reducing
the number of SMAs required for the analysis: by focusing on a small number of wires in a localised
the number of SMAs required for the analysis: by focusing on a small number of wires in a localised
area the heat generated by Joule effect in that specific area increased, thus enhancing the resolution
area the heat generated by Joule effect in that specific area increased, thus enhancing the resolution of
of the analysis. For this reason, smaller defects could be detected without the need to increase the
total feeding current, thus keeping the total energy consumption of the system to low levels.

6. Conclusions
This paper presented an overview of active infrared thermography techniques recently used for
Sensors 2018, 18, 609 30 of 37

the analysis. For this reason, smaller defects could be detected without the need to increase the total
feeding current, thus keeping the total energy consumption of the system to low levels.

6. Conclusions
This paper presented an overview of active infrared thermography techniques recently used for
the non-destructive evaluation of aerospace components. Optically stimulated thermographic methods
are today the most widely used thermal imaging systems. These techniques are rapid, contactless,
relatively low-cost compared to X-ray tomography and ultrasonic phased array and, when combined
with advanced signal processing tools, they can successfully assess various types of material defects
including flaws at the interfaces between surface coatings and their substrate, corrosion and fatigue
damage in metals and debonding and delamination in composites. However, optically stimulated
thermography is still not very sensitive to in-depth damage and micro-cracks with dimensions ranging
from ten microns to few millimetres. Among optically stimulated thermography techniques, laser
thermography is able to reveal cracks perpendicular to the sample’s surface (intralaminar), but
this method is still limited to near field heating. Ultrasonic stimulated thermography provides
a fast, full-filed and accurate quantification of micro-cracks due to the frictional heating caused
by the interaction of ultrasonic waves with the material damage. Although this thermographic
method is less affected by non-uniform heating, the ultrasonic excitation requires high-power, bulky
contact transducers and may generate chaotic ultrasonic wave propagation, thus making the thermal
acquisition process non-reproducible. The combination of ultrasonic stimulated thermography with
the local damage resonance effect and nonlinear ultrasounds has mitigated this issue by increasing
the selectivity of damage. Further quantitative studies are currently under development to enhance
the capabilities of this thermal imaging method. Eddy current stimulated thermography is sensitive
to a wide range of surface and sub-surface defects such as surface cracks in metals and delamination
and voids in composites. Moreover, it is also able to reveal the fibre pattern in carbon fibre laminates.
However, eddy current stimulated thermography is limited to near field planar defects in conductive
or semi-conductive materials and is severely affected by non-uniform heating. Similarly to eddy
current stimulated thermography, also microwave thermography is sensitive to superficial cracks
when water is trapped in the open defect. However, this thermal method is still in its infancy for
the non-destructive evaluation of aerospace structures. Indeed, the complexity of the microwave
operational heating system may be a risk for the integrity of the component if high power radiation is
applied. Moreover, alternative strategies to current external thermal excitation sources for composite
materials, here named as material-based thermography methods, were analysed in this review. These
thermographic techniques rely on the measurement of internal heating via Joule effect originated by
applying electrical current either directly to main fibrous reinforcement (i.e., carbon fibres) or through
the embodiment of materials such as metals inserts, carbon nanotubes and shape memory alloys.
Material-based thermography methods have proved to offer a fast, low power, accurate and reliable
assessment of delamination and cracks in aerospace composite components. Further work is currently
under development to enhance the thermal characteristics of material enabled thermography and
investigate the effects of embedded thermoresistive components on the laminate mechanical properties.

Acknowledgments: The authors would like to pay tribute to Darryl Almond who passed away before submission
of this article. He was a good scientist, a kind man, and more.
Author Contributions: All authors have equally contributed to this review paper.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Boccardi, S.; Boffa, N.D.; Carlomagno, G.M.; Maio, L.; Meola, C.; Ricci, F. Infrared thermography and ultrasonics
to evaluate composite materials for aeronautical applications. J. Phys. 2015, 658, 012007. [CrossRef]
Sensors 2018, 18, 609 31 of 37

2. Montesano, J.; Bougherara, H.; Fawaz, Z. Application of infrared thermography for the characterization
of damage in braided carbon fiber reinforced polymer matrix composites. Compos. Part B Eng. 2014, 60,
137–143. [CrossRef]
3. De Angelis, G.; Dati, E.; Bernabei, M.; Betti, L.; Menner, P. Non Destructive Investigation of disbonding damage
in air cooling inlet composite panel using active thermography and shearography. MAYFEB J. Mater. Sci. 2016,
1, 10–20.
4. Ranjit, S.; Choi, M.; Kim, W. Quantification of defects depth in glass fiber reinforced plastic plate by infrared
lock-in thermography. J. Mech. Sci. Technol. 2016, 30, 1111–1118. [CrossRef]
5. Rana, S.; Fangueiro, R. (Eds.) Advanced Composite Materials for Aerospace Engineering: Processing, Properties and
Applications; Woodhead Publishing: Cambridge, UK, 2016.
6. Abrate, S. Impact on Composite Structures; Cambridge University Press: Cambridge, UK, 2005.
7. Hull, D.; Clyne, T. An Introduction to Composite Materials; Cambridge University Press: Cambridge, UK, 1996.
8. Soutis, C. Fibre reinforced composites in aircraft construction. Prog. Aerosp. Sci. 2005, 41, 143–151. [CrossRef]
9. Brown, W.F.; Ho, C.Y.; Mindlin, H. Aerospace Structural Metals Handbook; CINDAS-USAF CRDA Handbooks
Operation; Purdue University: West Lafayette, IN, USA, 1979.
10. Lopez, F.; Sfarra, S.; Ibarra-Castanedo, C.; Paoletti, D.; Maldague, X. Integration of infrared and optical
imaging techniques for the nondestructive inspection of aeronautic parts. In SPIE Thermosense Conference:
Thermal Infrared Applications XXXVII; International Society for Optics and Photonics: Baltimore, MD, USA,
2015; Volume 9485, p. 948506.
11. Sfarra, S.; Ibarra-Castanedo, C.; Avdelidis, N.P.; Genest, M.; Bouchagier, L.; Kourousis, D.; Tsimogiannis, A.;
Anastassopoulous, A.; Bendada, A.; Maldague, X.; et al. A comparative investigation for the nondestructive
testing of honeycomb structures by holographic interferometry and infrared thermography. J. Phys. 2010,
214, 012071. [CrossRef]
12. Ibarra-Castanedo, C.; Galmiche, F.; Darabi, A.; Pilla, M.; Klein, M.; Ziadi, A.; Maldague, X.P. Thermographic
nondestructive evaluation: Overview of recent progress. Proc. SPIE 2003, 5073, 450–459.
13. Balageas, D.; Maldague, X.; Burleigh, D.; Vavilov, V.P.; Oswald-Tranta, B.; Roche, J.M.; Carlomagno, G.M.
Thermal (IR) and Other NDT Techniques for Improved Material Inspection. J. Nondestruct. Eval. 2016, 35,
1–17. [CrossRef]
14. Montesano, J.; Fawaz, Z.; Bougherara, H. Use of infrared thermography to investigate the fatigue behavior
of a carbon fiber reinforced polymer composite. Compos. Struct. 2013, 97, 76–83. [CrossRef]
15. Bates, D.; Smith, G.; Lu, D.; Hewitt, J. Rapid thermal non-destructive testing of aircraft components.
Compos. Part B Eng. 2000, 31, 175–185. [CrossRef]
16. Vavilov, V.P.; Pawar, S.S. A novel approach for one-sided thermal nondestructive testing of composites by
using infrared thermography. Polym. Test. 2015, 44, 224–233. [CrossRef]
17. Pickering, S.; Almond, D.-P. Matched excitation energy comparison of the pulse and lock-in thermography.
NDT E Int. 2008, 41, 501–509. [CrossRef]
18. Ibarra-Castanedo, C.; Maldague, X.P.V. Infrared Thermography. In Handbook of Technical Diagnostics;
Czichos, H., Ed.; Springer: Berlin, Heidelberg, Germany, 2013; pp. 175–220.
19. Yang, R.; He, Y. Optically and non-optically excited thermography for composites: A review. Infrared Phys. Technol.
2016, 75, 26–50. [CrossRef]
20. Maldague, X.P. Theory and Practice of Infrared Technology for Nondestructive Testing; John Wiley Interscience:
New York, NY, USA, 2001.
21. Bai, W.; Wong, B.S. Nondestructive evaluation of aircraft structure using lockin thermography. In Proceedings
of the SPIE’s 5th Annual International Symposium on Nondestructive Evaluation and Health Monitoring
of Aging Infrastructure, Newport Beach, CA, USA, 6–8 March 2000; International Society for Optics and
Photonics: Bellingham, WA, USA, 2000; pp. 37–46.
22. Badghaish, A.A.; Fleming, D.C. Non-destructive inspection of composites using step heating thermography.
J. Compos. Mater. 2008, 42, 1337–1357. [CrossRef]
23. Almond, D.P.; Angioni, S.L.; Pickering, S.G. Long pulse excitation thermographic non-destructive evaluation.
NDT E Int. 2017, 87, 7–14.
24. Mulaveesala, R.; Tuli, S. Theory of frequency modulated thermal wave imaging for nondestructive subsurface
defect detection. Appl. Phys. Lett. 2006, 89, 191913. [CrossRef]
Sensors 2018, 18, 609 32 of 37

25. Li, T.; Almond, D.P.; Rees, D.A.S. Crack imaging by scanning pulsed laser spot thermography. NDT E Int.
2011, 44, 216–225. [CrossRef]
26. Woolard, Deonna F.; Cramer, K.E. Line scan versus flash thermography: Comparative study on reinforced
carbon-carbon. In Thermosense XXVII; International Society for Optics and Photonics: Bellingham, WA, USA,
2005; Volume 5782, pp. 315–324.
27. Zweschper, T.; Riegert, G.; Dillenz, A.; Busse, G. Ultrasound excited thermography-advances due to
frequency modulated elastic waves. Quant. InfraRed Thermogr. J. 2005, 2, 65–76. [CrossRef]
28. Fierro, G.P.M.; Ginzburg, D.; Ciampa, F.; Meo, M. Imaging of barely visible impact damage on a complex
composite stiffened panel using a nonlinear ultrasound stimulated thermography approach. J. Nondestruct. Eval.
2017, 36, 69. [CrossRef]
29. Wilson, J.; Tian, G.Y.; Abidin, I.Z.; Yang, S.; Almond, D. Modelling and evaluation of eddy current stimulated
thermography. Nondestruct. Test. Eval. 2010, 25, 205–218. [CrossRef]
30. Levesque, P.; Deom, A.; Balageas, D. Non destructive evaluation of absorbing materials using microwave
stimulated infrared thermography. In Review of Progress in Quantitative Nondestructive Evaluation; Springer:
Berlin, Germany, 1993; pp. 649–654.
31. Sakagami, T.; Ogura, K. A New flaw inspection technique based on infrared thermal images under Joule
effect heating. JSME Int. J. Ser. A 1994, 37, 380–388. [CrossRef]
32. Ahmed, T.; Nino, G.; Bersee, H.; Beukers, A. Heat emitting layers for enhancing NDE of composite structures.
Compos. Part A Appl. Sci. Manuf. 2008, 39, 1025–1036. [CrossRef]
33. De Villoria, R.G.; Yamamoto, N.; Miravete, A.; Wardle, B.L. Multi-physics damage sensing in nano-engineered
structural composites. Nanotechnology 2011, 22, 185502. [CrossRef] [PubMed]
34. Pinto, F.; Ciampa, F.; Meo, M.; Polimeno, U. Multifunctional SMArt composite material for in situ NDT/SHM
and de-icing. Smart Mater. Struct. 2012, 21, 105010. [CrossRef]
35. Czichos, H. Handbook of Technical Diagnostics: Fundamentals and Application to Structures and Systems; Springer
Science & Business Media: Berlin, Germany, 2013.
36. Fazeli, H.; Mirzaei, M. Shape identification problems on detecting of defects in a solid body using inverse
heat conduction approach. J. Mech. Sci. Technol. 2012, 26, 3681–3690. [CrossRef]
37. Junyan, L.; Liqiang, L.; Yang, W. Experimental study on active infrared thermography as a NDI tool for
carbon–carbon composites. Compos. Part B Eng. 2013, 45, 138–147. [CrossRef]
38. Li, Y.; Yang, Z.W.; Zhu, J.T.; Ming, A.B.; Zhang, W.; Zhang, J.Y. Investigation on the damage evolution in the
impacted composite material based on active infrared thermography. NDT E Int. 2016, 83, 114–122. [CrossRef]
39. Zalameda, J.N.; Rajic, N.; Winfree, W.P. A comparison of image processing algorithms for thermal
nondestructive evaluation. In AeroSense 2003; International Society for Optics and Photonics: Bellingham,
WA, USA, 2003; pp. 374–385.
40. Madruga, F.J.; Ibarra-Castanedo, C.; Conde, O.M.; López-Higuera, J.M.; Maldague, X. Infrared thermography
processing based on higher-order statistics. NDT E Int. 2010, 43, 661–666. [CrossRef]
41. Vavilov, V.P.; Burleigh, D.D. Review of pulsed thermal NDT: Physical principles, theory and data processing.
NDT E Int. 2015, 73, 28–52. [CrossRef]
42. Pilla, M.; Klein, M.; Maldague, X.; Salerno, A. New absolute contrast for pulsed thermography. In Proceeding
QIRT; 2002; Volume 5. [CrossRef]
43. Meola, C.; Carlomagno, G.M. Impact damage in GFRP: New insights with infrared thermography.
Compos. Part A Appl. Sci. Manuf. 2010, 41, 1839–1847. [CrossRef]
44. Shepard, S.M.; Lhota, J.R.; Rubadeux, B.A.; Wang, D.; Ahmed, T. Reconstruction and enhancement of active
thermographic image sequences. Opt. Eng. 2003, 42, 1337–1342. [CrossRef]
45. Shepard, S.M. Advances in pulsed thermography. In Thermosense XXIII; International Society for Optics and
Photonics: Bellingham, WA, USA, 2001; Volume 4360, pp. 511–516.
46. Rajic, N. Principal component thermography for flaw contrast enhancement and flaw depth characterisation
in composite structures. Compos. Struct. 2002, 58, 521–528. [CrossRef]
47. Tang, Q.; Dai, J.; Bu, C.; Qi, L.; Li, D. Experimental study on debonding defects detection in thermal barrier coating
structure using infrared lock-in thermographic technique. Appl. Therm. Eng. 2016, 107, 463–468. [CrossRef]
48. Pawar, S.S.; Vavilov, V.P. Applying the heat conduction-based 3D normalization and thermal tomography to
pulsed infrared thermography for defect characterization in composite materials. Int. J. Heat Mass Transf.
2016, 94, 56–65. [CrossRef]
Sensors 2018, 18, 609 33 of 37

49. Chang, Y.-S.; Yan, Z.; Wang, K.-H.; Yao, Y. Non-destructive testing of CFRP using pulsed thermography
and multi-dimensional ensemble empirical mode decomposition. J. Taiwan Inst. Chem. Eng. 2016, 61, 54–63.
[CrossRef]
50. Li, B.; Ye, L.; Li, E.; Shou, D.; Li, Z.; Chang, L. Gapped smoothing algorithm applied to defect identification
using pulsed thermography. Nondestruct. Test. Eval. 2015, 30, 171–195. [CrossRef]
51. Lopez, F.; Ibarra-Castanedo, C.; de Paulo Nicolau, V.; Maldague, X. Optimization of pulsed thermography
inspection by partial least-squares regression. NDT E Int. 2014, 66, 128–138. [CrossRef]
52. Zhao, Y.; Tinsley, L.; Addepalli, S.; Mehnen, J.; Roy, R. A coefficient clustering analysis for damage assessment
of composites based on pulsed thermographic inspection. NDT E Int. 2016, 83, 59–67. [CrossRef]
53. Almond, D.P.; Pickering, S.G. An analytical study of the pulsed thermography defect detection limit.
J. Appl. Phys. 2012, 111, 093510. [CrossRef]
54. Carslaw, H.S.; Jaeger, J.C. Conduction of Heat in Solids, 2nd ed.; Clarendon Press: Oxford, UK, 1959; Volume 1959.
55. Duffy, D.G. Green’s Functions with Applications; CRC Press: Boca Raton, FL, USA, 2015.
56. Ishikawa, M.; Hatta, H.; Habuka, Y.; Fukui, R.; Utsunomiya, S. Detecting deeper defects using pulse phase
thermography. Infrared Phys. Technol. 2013, 57, 42–49. [CrossRef]
57. Shin, P.H.; Webb, S.C.; Peters, K.J. Pulsed phase thermography imaging of fatigue-loaded composite
adhesively bonded joints. NDT E Int. 2016, 79, 7–16. [CrossRef]
58. Ibarra-Castanedo, C.; Maldague, X.P. Review of pulse phase thermography. In SPIE Sensing Technology+
Applications; International Society for Optics and Photonics: Bellingham, WA, USA, 2015; p. 94850T.
59. Maldague, X.; Marinetti, S. Pulse phase infrared thermography. J. Appl. Phys. 1996, 79, 2694–2698. [CrossRef]
60. Ibarra-Castanedo, C. Quantitative Subsurface Defect Evaluation by Pulsed Phase Thermography Depth
Retrieval with the Phase. Ph.D. Thesis, Faculte des Etudes Superieurs de Universite Laval, Québec City, QC,
Canada, 2005.
61. Meola, C.; Carlomagno, G.M.; Giorleo, L. The use of infrared thermography for materials characterization.
J. Mater. Process. Technol. 2004, 155, 1132–1137. [CrossRef]
62. Yang, B.; Huang, Y.; Cheng, L. Defect detection and evaluation of ultrasonic infrared thermography for
aerospace CFRP composites. Infrared Phys. Technol. 2013, 60, 166–173. [CrossRef]
63. Cramer, K.E. Research Developments in Nondestructive Evaluation and Structural Health Monitoring
for the Sustainment of Composite Aerospace Structures at NASA. In Proceedings of the Congress of the
International Council of the Aeronautical Sciences, Daejeon, Korea, 25–30 September 2016.
64. Manohar, A.; di Scalea, F.L. Modeling 3D heat flow interaction with defects in composite materials for
infrared thermography. NDT E Int. 2014, 66, 1–7. [CrossRef]
65. Guo, X.; Vavilov, V. Pulsed thermographic evaluation of disbonds in the insulation of solid rocket motors
made of elastomers. Polym. Test. 2015, 45, 31–40. [CrossRef]
66. Tang, Q.; Dai, J.; Liu, J.; Liu, C.; Liu, Y.; Ren, C. Quantitative detection of defects based on Markov–PCA–BP
algorithm using pulsed infrared thermography technology. Infrared Phys. Technol. 2016, 77, 144–148. [CrossRef]
67. Li, Y.; Zhang, W.; Yang, Z.W.; Zhang, J.Y.; Tao, S.J. Low-velocity impact damage characterization of carbon fiber
reinforced polymer (CFRP) using infrared thermography. Infrared Phys. Technol. 2016, 76, 91–102. [CrossRef]
68. Oswald-Tranta, B. Time and frequency behaviour in TSR and PPT evaluation for flash thermography.
Quant. InfraRed Thermogr. J. 2017, 14, 164–184. [CrossRef]
69. Meola, C.; Carlomagno, G.M.; Squillace, A.; Vitiello, A. Non-destructive evaluation of aerospace materials
with lock-in thermography. Eng. Fail. Anal. 2006, 13, 380–388. [CrossRef]
70. Chatterjee, K.; Tuli, S.; Pickering, S.G.; Almond, D.P. A comparison of the pulsed, lock-in and frequency
modulated thermography nondestructive evaluation techniques. NDT E Int. 2011, 44, 655–667. [CrossRef]
71. Duan, Y.; Huebner, S.; Hassler, U.; Osman, A.; Ibarra-Castanedo, C.; Maldague, X.P. Quantitative evaluation
of optical lock-in and pulsed thermography for aluminum foam material. Infrared Phys. Technol. 2013, 60,
275–280. [CrossRef]
72. Kordatos, E.Z.; Dassios, K.G.; Aggelis, D.G.; Matikas, T.E. Rapid evaluation of the fatigue limit in composites
using infrared lock-in thermography and acoustic emission. Mech. Res. Commun. 2013, 54, 14–20. [CrossRef]
73. Grammatikos, S.A.; Kordatos, E.Z.; Matikas, T.E.; Paipetis, A.S. Service and maintenance damage assessment
of composite structures using various modes of infrared thermography. In IOP Conference Series: Materials
Science and Engineering; IoP Publishing: Bristol, UK, 2015; Volume 74, p. 012006.
Sensors 2018, 18, 609 34 of 37

74. Zhao, H.; Zhou, Z.; Fan, J.; Li, G.; Sun, G. Application of lock-in thermography for the inspection of disbonds
in titanium alloy honeycomb sandwich structure. Infrared Phys. Technol. 2016, 81, 69–78. [CrossRef]
75. Montanini, R.; Freni, F. Non-destructive evaluation of thick glass fiber-reinforced composites by means of
optically excited lock-in thermography. Compos. Part A Appl. Sci. Manuf. 2012, 43, 2075–2082. [CrossRef]
76. Harizi, W.; Chaki, S.; Bourse, G.; Ourak, M. Mechanical damage assessment of Polymer–Matrix Composites
using active infrared thermography. Compos. Part B Eng. 2014, 66, 204–209. [CrossRef]
77. Junyan, L.; Yang, L.; Fei, W.; Yang, W. Study on probability of detection (POD) determination using lock-in
thermography for nondestructive inspection (NDI) of CFRP composite materials. Infrared Phys. Technol. 2015,
71, 448–456. [CrossRef]
78. Meola, C.; Carlomagno, G.M. Lock-in thermography to detect delamination in carbon fibres reinforced
polymers. In Damage Growth in Aerospace Composites; Springer International Publishing: Geneva, Switzerland,
2015; pp. 263–279.
79. Junyan, L.; Qingju, T.; Yang, W. The study of inspection on SiC coated carbon–carbon composite with
subsurface defects by lock-in thermography. Compos. Sci. Technol. 2012, 72, 1240–1250. [CrossRef]
80. Junyan, L.; Fei, W.; Yang, L.; Yang, W. Inverse methodology for identification the thermal diffusivity
and subsurface defect of CFRP composite by lock-in thermographic phase (LITP) profile reconstruction.
Compos. Struct. 2016, 138, 214–226. [CrossRef]
81. Meola, C.; Boccardi, S.; Carlomagno, G.M.; Boffa, N.D.; Ricci, F.; Simeoli, G.; Russo, P. Impact damaging
of composites through online monitoring and non-destructive evaluation with infrared thermography.
NDT E Int. 2017, 85, 34–42. [CrossRef]
82. Balageas, D.L.; Roche, J.M. Common tools for quantitative time-resolved pulse and step-heating
thermography—Part I: Theoretical basis. Quant. InfraRed Thermogr. J. 2014, 11, 43–56. [CrossRef]
83. Roche, J.M.; Balageas, D.L. Common tools for quantitative pulse and step-heating thermography–part II:
Experimental investigation. Quant. InfraRed Thermogr. J. 2015, 12, 1–23. [CrossRef]
84. Ghali, V.S.; Mulaveesala, R.; Takei, M. Frequency-modulated thermal wave imaging for non-destructive
testing of carbon fiber-reinforced plastic materials. Meas. Sci. Technol. 2011, 22, 104018. [CrossRef]
85. Arora, V.; Siddiqui, J.A.; Mulaveesala, R.; Muniyappa, A. Pulse compression approach to nonstationary
infrared thermal wave imaging for nondestructive testing of carbon fiber reinforced polymers. IEEE Sens. J.
2015, 15, 663–664. [CrossRef]
86. Burrows, S.E.; Dixon, S.; Pickering, S.G.; Li, T.; Almond, D.P. Thermographic detection of surface breaking
defects using a scanning laser source. NDT E Int. 2011, 44, 589–596. [CrossRef]
87. Oswald-Tranta, B.; Shepard, S.M. Comparison of pulse phase and thermographic signal reconstruction
processing methods. In Thermosense: Thermal Infrared Applications XXXV; International Society for Optics
and Photonics: Bellingham, WA, USA, 2013; Volume 8705, p. 87050S.
88. Li, T.; Almond, D.P.; Rees, D.A.S. Crack imaging by scanning laser-line thermography and laser-spot
thermography. Meas. Sci. Technol. 2011, 22, 035701. [CrossRef]
89. Ley, O.; Butera, M.; Godinez, V. Inspection of composite structures using line scanning thermography.
In SPIE Defense, Security, and Sensing; International Society for Optics and Photonics: Bellingham, WA, USA,
2012; p. 835406.
90. Ley, O.; Godinez-Azcuaga, V. Line Scanning Thermography and its Application Inspecting Aerospace
Composites. In Proceedings of the 5th International Symposium on NDT in Aerospace, Singapore,
13–15 November 2013.
91. Fedala, Y.; Streza, M.; Sepulveda, F.; Roger, J.P.; Tessier, G.; Boué, C. Infrared lock-in thermography crack
localization on metallic surfaces for industrial diagnosis. J. Nondestruct. Eval. 2014, 33, 335–341. [CrossRef]
92. Roemer, J.; Uhl, T.; Pieczonka, L. Laser spot thermography for crack detection in aluminum structures.
In Proceedings of the 7th International Symposium on NDT in Aerospace, Bremen, Germany,
16–18 November 2015; pp. 1–7.
93. Khodayar, F.; Lopez, F.; Ibarra-Castanedo, C.; Maldague, X. Optimization of the inspection of large composite
materials using robotized line scan thermography. J. Nondestruct. Eval. 2017, 36, 32. [CrossRef]
94. Zhang, H.; Fernandes, H.; Dizeu, F.B.D.; Hassler, U.; Fleuret, J.; Genest, M.; Ibarra-Castanedo, C.; Robitaille, F.;
Joncas, S.; Maldague, X. Pulsed micro-laser line thermography on submillimeter porosity in carbon fiber
reinforced polymer composites: Experimental and numerical analyses for the capability of detection.
Appl. Opt. 2016, 55, D1–D10. [CrossRef] [PubMed]
Sensors 2018, 18, 609 35 of 37

95. Fernandes, H.; Zhang, H.; Figueiredo, A.; Malheiros, F.; Ignacio, L.H.; Sfarra, S.; Ibarra-Castanedo, C.;
Guimaraes, G.; Maldague, X. Machine Learning and Infrared Thermography for Fiber Orientation
Assessment on Randomly-Oriented Strands Parts. Sensors 2018, 18, 288. [CrossRef] [PubMed]
96. Barden, T.J.; Almond, D.P.; Pickering, S.G.; Morbidini, M.; Cawley, P. Detection of impact damage in CFRP
composites by thermosonics. Nondestruct. Test. Eval. 2007, 22, 71–82. [CrossRef]
97. Sakagami, T.; Kuroki, K.; Kubo, S.; Katsumata, R.; Matsumoto, Y.; Harada, Y. Detection of Stress Corrosion
Cracking by Sonic-IR Technique. In Proceedings of the 7th International Conference on NDE in Relation to
Structural Integrity for Nuclear and Pressurized Components, Yokohama, Japan, 12–14 May 2009; pp. 12–15.
98. Polimeno, U.; Almond, D.P.; Weekes, B.; Chen, E.W.J. A compact thermosonic inspection system for the
inspection of composites. Compos. Part B Eng. 2014, 59, 67–73. [CrossRef]
99. Castelo, A.; Mendioroz, A.; Celorrio, R.; Salazar, A. Optimizing the Inversion Protocol to Determine the
Geometry of Vertical Cracks from Lock-in Vibrothermography. J. Nondestruct. Eval. 2017, 36, 3. [CrossRef]
100. Umar, M.Z.; Vavilov, V.; Abdullah, H.; Ariffin, A.K. Ultrasonic infrared thermography in non-destructive
testing: A review. Rus. J. Nondestruct. Test. 2016, 52, 212–219. [CrossRef]
101. Guo, X.; Vavilov, V. Crack detection in aluminum parts by using ultrasound-excited infrared thermography.
Infrared Phys. Technol. 2013, 61, 149–156. [CrossRef]
102. Gao, C.; Meeker, W.Q.; Mayton, D. Detecting cracks in aircraft engine fan blades using vibrothermography
nondestructive evaluation. Reliabil. Eng. Syst. Saf. 2014, 131, 229–235. [CrossRef]
103. Zhang, D.; Han, X.; Newaz, G. Sonic IR crack detection of aircraft turbine engine blades with multi-frequency
ultrasound excitations. In AIP Conference Proceedings; Chimenti, D.E., Bond, L.J., Thompson, D.O., Eds.; AIP:
College Park, MD, USA, 2014; Volume 1581, pp. 1644–1651.
104. Burke, M.W.; Miller, W.O. Status of VibroIR at Lawrence Livermore national laboratory. In Defense and
Security; International Society for Optics and Photonics: Bellingham, WA, USA, 2004; pp. 313–321.
105. Dillenz, A.; Zweschper, T.; Busse, G. Burst phase-angle thermography with elastic waves. In AeroSense 2002;
International Society for Optics and Photonics: Bellingham, WA, USA, 2002; pp. 572–577.
106. Han, X.; Favro, L.D.; Thomas, R.L. Sonic IR imaging of delaminations and disbonds in composites. J. Phys. D
Appl. Phys. 2010, 44, 034013. [CrossRef]
107. Renshaw, J.; Chen, J.C.; Holland, S.D.; Thompson, R.B. The sources of heat generation in vibrothermography.
NDT E Int. 2011, 44, 736–739. [CrossRef]
108. Morbidini, M.; Cawley, P. A calibration procedure for sonic infrared nondestructive evaluation. J. Appl. Phys.
2009, 106, 023504. [CrossRef]
109. Gaudenzi, P.; Bernabei, M.; Dati, E.; De Angelis, G.; Marrone, M.; Lampani, L. On the evaluation of impact
damage on composite materials by comparing different NDI techniques. Compos. Struct. 2014, 118, 257–266.
[CrossRef]
110. Vavilov, V.; Świderski, W.; Derusova, D. Ultrasonic and optical stimulation in IR thermographic NDT of
impact damage in carbon composites. Quant. InfraRed Thermogr. J. 2015, 12, 162–172. [CrossRef]
111. Pieczonka, L.; Aymerich, F.; Brozek, G.; Szwedo, M.; Staszewski, W.J.; Uhl, T. Modelling and numerical
simulations of vibrothermography for impact damage detection in composites structures. Struct. Control
Health Monit. 2013, 20, 626–638. [CrossRef]
112. Rizi, A.S.; Hedayatrasa, S.; Maldague, X.; Vukhanh, T. FEM modeling of ultrasonic vibrothermography of a
damaged plate and qualitative study of heating mechanisms. Infrared Phys. Technol. 2013, 61, 101–110. [CrossRef]
113. Derusova, D.A.; Vavilov, V.P.; Pawar, S.S. Evaluation of equivalent defect heat generation in carbon epoxy
composite under powerful ultrasonic stimulation by using infrared thermography. In IOP Conference Series:
Materials Science and Engineering; IoP Publishing: Bristol, UK, 2015; Volume 81, p. 012084.
114. Parvasi, S.M.; Xu, C.; Kong, Q.; Song, G. Detection of multiple thin surface cracks using vibrothermography
with low-power piezoceramic-based ultrasonic actuator—A numerical study with experimental verification.
Smart Mater. Struct. 2016, 25, 055042. [CrossRef]
115. Tao, N.; Zeng, Z.; Feng, L.; Chen, D.; Li, Y.; Zhang, C. Noncontact ultrasonic excitation thermography
inspection technique. In Photonics Asia 2010; International Society for Optics and Photonics: Bellingham,
WA, USA, 2010; p. 78543S78543S.
116. Zalameda, J.N.; Winfree, W.P.; Yost, W.T. Air coupled acoustic thermography (ACAT) inspection technique.
In AIP Conference Proceedings; Thompson, D.O., Chimenti, D.E., Eds.; AIP: College Park, MD, USA, 2008;
Volume 975, pp. 467–474.
Sensors 2018, 18, 609 36 of 37

117. de Angelis, G.; Meo, M.; Almond, D.P.; Pickering, S.G.; Angioni, S.L. A new technique to detect defect size
and depth in composite structures using digital shearography and unconstrained optimization. NDT E Int.
2012, 45, 91–96. [CrossRef]
118. Solodov, I.; Bai, J.; Busse, G. Resonant ultrasound spectroscopy of defects: Case study of flat-bottomed holes.
J. Appl. Phys. 2013, 113, 223512. [CrossRef]
119. Ciampa, F.; Scarselli, G.; Meo, M. On the generation of nonlinear damage resonance intermodulation for
elastic wave spectroscopy. J. Acoust. Soc. Am. 2017, 141, 2364–2374. [CrossRef] [PubMed]
120. Rahammer, M.; Solodov, I.; Bisle, W.; Scherling, D.; Kreutzbruck, M. Thermosonic Testing with Phase
Matched Guided Wave Excitation. J. Nondestruct. Eval. 2016, 35, 47. [CrossRef]
121. Rahammer, M.; Kreutzbruck, M. Fourier-transform vibrothermography with frequency sweep excitation
utilizing local defect resonances. NDT E Int. 2017, 86, 83–88. [CrossRef]
122. Solodov, I.; Rahammer, M.; Gulnizkij, N.; Kreutzbruck, M. Noncontact Sonic NDE and Defect Imaging via
Local Defect Resonance. J. Nondestruct. Eval. 2016, 35, 48. [CrossRef]
123. Fierro, G.P.M.; Ginzburg, D.; Ciampa, F.; Meo, M. Nonlinear ultrasonic stimulated thermography for damage
assessment in isotropic fatigued structures. J. Sound Vib. 2017, 404, 102–115. [CrossRef]
124. Dionysopoulos, D.; Fierro, G.P.M.; Meo, M.; Ciampa, F. Imaging of barely visible impact damage on a
composite panel using nonlinear wave modulation thermography. NDT E Int. 2018, 95, 9–16. [CrossRef]
125. Abidin, I.Z.; Tian, G.Y.; Wilson, J.; Yang, S.; Almond, D. Quantitative evaluation of angular defects by pulsed
eddy current thermography. NDT E Int. 2010, 43, 537–546. [CrossRef]
126. Yin, A.; Gao, B.; Yun Tian, G.; Woo, W.L.; Li, K. Physical interpretation and separation of eddy current pulsed
thermography. J. Appl. Phys. 2013, 113, 064101. [CrossRef]
127. Yang, R.; He, Y. Eddy current pulsed phase thermography considering volumetric induction heating for
delamination evaluation in carbon fiber reinforced polymers. Appl. Phys. Lett. 2015, 106, 234103. [CrossRef]
128. Cheng, L.; Tian, G.Y. Transient thermal behavior of eddy-current pulsed thermography for nondestructive
evaluation of composites. IEEE Trans. Instrum. Meas. 2013, 62, 1215–1222. [CrossRef]
129. Yang, R.; He, Y. Polymer-matrix composites carbon fibre characterisation and damage inspection using
selectively heating thermography (SeHT) through electromagnetic induction. Compos. Struct. 2016, 140,
590–601. [CrossRef]
130. Cheng, L.; Tian, G.Y. Surface crack detection for carbon fiber reinforced plastic (CFRP) materials using pulsed
eddy current thermography. IEEE Sens. J. 2011, 11, 3261–3268. [CrossRef]
131. Vrana, J.; Goldammer, M.; Bailey, K.; Rothenfusser, M.; Arnold, W. Induction and conduction
thermography: Optimizing the electromagnetic excitation towards application. In AIP Conference Proceedings;
Thompson, D.O., Chimenti, D.E., Eds.; AIP: College Park, MD, USA, 2009; Volume 1096, pp. 518–525.
132. Libin, M.N.; Balasubramaniam, K.; Maxfield, B.W.; Krishnamurthy, C.V. Modeling and Simulation for crack
detection in aluminium components using tone burst eddy current thermography. In Proceedings of the
National Seminar & Exhibition on Non Destructive Evaluation, Pune, India, 4–6 December 2014.
133. Frackowiak,
˛ W.; Bruchwald, O.; Reimche, W.; Bach, F.W.; Maier, H.J. High Frequency Eddy-Current and
Induction Thermography Inspection Techniques for Turbine Components. In Electromagnetic Nondestructive
Evaluation XVII; Capova, K., Udpa, L., Janousek und BPC Rao, L., Eds.; IOS Press: Amsterdam,
The Netherlands, 2014; pp. 226–233.
134. Ren, W.; Liu, J.; Tian, G.Y.; Gao, B.; Cheng, L.; Yang, H. Quantitative non-destructive evaluation method for
impact damage using eddy current pulsed thermography. Compos. Part B Eng. 2013, 54, 169–179. [CrossRef]
135. He, Y.; Tian, G.; Pan, M.; Chen, D. Impact evaluation in carbon fiber reinforced plastic (CFRP) laminates
using eddy current pulsed thermography. Compos. Struct. 2014, 109, 1–7. [CrossRef]
136. Liang, T.; Ren, W.; Tian, G.Y.; Elradi, M.; Gao, Y. Low energy impact damage detection in CFRP using eddy
current pulsed thermography. Compos. Struct. 2016, 143, 352–361. [CrossRef]
137. Ishikawa, M.; Shiiya, Y.; Kogo, Y.; Hatta, H.; Habuka, Y.; Fukui, R.; UTSUNOMIYA, S. Inspection of CFRP
laminates using phase-transformed induction heating thermography. In Proceedings of the 11th European
Conference of Non-Destructive Testing, Prague, Czech Republic, 6–10 October 2014.
138. He, Y.; Yang, R. Eddy current volume heating thermography and phase analysis for imaging characterization
of interface delamination in CFRP. IEEE Trans. Ind. Inform. 2015, 11, 1287–1297. [CrossRef]
Sensors 2018, 18, 609 37 of 37

139. Thomas, K.R.; Balasubramaniam, K. Scanning induction thermography (SIT) for imaging damages in
carbon-fibre reinforced plastics (CFRP) components. In AIP Conference Proceedings; Chimenti, D.E., Bond, L.J.,
Eds.; AIP: College Park, MD, USA, 2015; Volume 1650, pp. 306–313.
140. Morozov, M.; Jackson, W.; Pierce, S.G. Capacitive imaging of impact damage in composite material.
Compos. Part B Eng. 2017, 113, 65–71. [CrossRef]
141. Foudazi, A.; Donnell, K.M.; Ghasr, M.T. Application of active microwave thermography to delamination
detection. In Proceedings of the 2014 IEEE International Instrumentation and Measurement Technology
Conference (I2MTC) Proceedings, Montevideo, Uruguay, 12–15 May 2014; pp. 1567–1571.
142. Yang, R.; Zhang, H.; Li, T.; He, Y. An investigation and review into microwave thermography for NDT
and SHM. In Proceedings of the 2015 IEEE Far East NDT New Technology & Application Forum (FENDT),
Zhuhai, China, 28–31 May 2015; pp. 133–137.
143. Sam-Ang, K.; Didier, D.; Florin, B.; Franck, B. Comparison between microwave infrared thermography and
CO2 laser infrared thermography in defect detection in applications with CFRP. Mater. Sci. Appl. 2013, 4, 600.
144. Palumbo, D.; Ancona, F.; Galietti, U. Quantitative damage evaluation of composite materials with microwave
thermographic technique: Feasibility and new data analysis. Meccanica 2015, 50, 443–459. [CrossRef]
145. Schulte, K.; Baron, C. Load and failure analyses of CFRP laminates by means of electrical resistivity
measurements. Compos. Sci. Technol. 1989, 36, 63–76. [CrossRef]
146. Abry, J.; Bochard, S.; Chateauminois, A.; Salvia, M.; Giraud, G. In situ detection of damage in CFRP laminates
by electrical resistance measurements. Compos. Sci. Technol. 1999, 59, 925–935. [CrossRef]
147. Suzuki, Y.; Todoroki, A.; Mizutani, Y.; Matsuzaki, R. Impact damage detection in CFRP using statistical
analysis of resistance-temperature characteristics. J. Solid Mech. Mater. Eng. 2011, 5, 33–43. [CrossRef]
148. Suzuki, Y.; Todoroki, A.; Matsuzaki, R.; Mizutani, Y. Impact-damage visualization in CFRP by resistive
heating: Development of a new detection method for indentations caused by impact loads. Compos. Part A
Appl. Sci. Manuf. 2012, 43, 53–64. [CrossRef]
149. Suzuki, Y.; Todoroki, A.; Matsuzaki, R.; Mizutani, Y. Indentation-damage visualization in CFRP by resistive
heating: Analytical verification of the inspection of aircraft using its lightning protection systems. J. Solid
Mech. Mater. Eng. 2012, 6, 213–226. [CrossRef]
150. Grammatikos, S.; Kordatos, E.; Barkoula, N.; Matikas, T. Paipetis, A. Innovative non-destructive evaluation
and damage characterisation of composite aerostructures using thermography. Plast. Rubber Compos. 2011,
40, 6–7. [CrossRef]
151. Grammatikos, S.; Kordatos, E.; Matikas, T.; David, C.; Paipetis, A. Current injection phase thermography for
low-velocity impact damage identification in composite laminates. Mater. Des. 2014, 55, 429–441. [CrossRef]
152. Avdelidis, N.P.; Hawtin, B.C.; Almond, D.P. Transient thermography in the assessment of defects of aircraft
composites. NDT E Int. 2003, 36, 433–439. [CrossRef]
153. Orłowska, A.; Kołakowski, P.; Holnicki-Szulc, J. Detecting delamination zones in composites by embedded
electrical grid and thermographic methods. Smart Mater. Struct. 2011, 20, 105009. [CrossRef]
154. Garcia, E.J.; Wardle, B.L.; Hart, A.J.; Yamamoto, N. Fabrication and multifunctional properties of a hybrid
laminate with aligned carbon nanotubes grown in situ. Compos. Sci. Technol. 2008, 68, 2034–2041. [CrossRef]
155. Wicks, S.S.; de Villoria, R.G.; Wardle, B.L. Interlaminar and intralaminar reinforcement of composite laminates
with aligned carbon nanotubes. Compos. Sci. Technol. 2010, 70, 20–28. [CrossRef]
156. Paine, J.S.N.; Rogers, C.A. The Response of SMA Hybrid Composite Materials to Low Velocity Impact.
J. Intell. Mater. Syst. Struct. 1994, 5, 530–535. [CrossRef]
157. Angioni, S.L.; Meo, M.; Foreman, A. Impact damage resistance and damage suppression properties of shape
memory alloys in hybrid composites—A review. Smart Mater. Struct. 2011, 20, 013001. [CrossRef]
158. Angioni, S.L.; Ciampa, F.; Pinto, F.; Scarselli, G.; Almond, D.P.; Meo, M. An Analytical Model for Defect
Depth Estimation Using Pulsed Thermography. Exp. Mech. 2016, 56, 1111–1122. [CrossRef]
159. Pinto, F.; Maroun, F.; Meo, M. Material enabled thermography. NDT E Int. 2014, 67, 1–9. [CrossRef]

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).