Você está na página 1de 9

sensors

Article
Fatigue Damage Evaluation of Short Carbon Fiber
Reinforced Plastics Based on Phase Information of
Thermoelastic Temperature Change
Daiki Shiozawa 1, *, Takahide Sakagami 1 , Yu Nakamura 1 , Shinichi Nonaka 2 and Kenichi Hamada 2
1 Department of Mechanical Engineering, Kobe University, Kobe 657-8501, Japan;
sakagami@mech.kobe-u.ac.jp (T.S.); 169t348t@stu.kobe-u.ac.jp (Y.N.)
2 DIC Corporation, Tokyo 103-8233, Japan; shinichi-nonaka@mb.dic.co.jp (S.N.); kenichi-hamada@mb.dic.co.jp (K.H.)
* Correspondence: shiozawa@mech.kobe-u.ac.jp; Tel.: +81-78-803-6303

Received: 14 November 2017; Accepted: 1 December 2017; Published: 6 December 2017

Abstract: Carbon fiber-reinforced plastic (CFRP) is widely used for structural members of
transportation vehicles such as automobile, aircraft, or spacecraft, utilizing its excellent specific
strength and specific rigidity in contrast with the metal. Short carbon fiber composite materials are
receiving a lot of attentions because of their excellent moldability and productivity, however they
show complicated behaviors in fatigue fracture due to the random fibers orientation. In this study,
thermoelastic stress analysis (TSA) using an infrared thermography was applied to evaluate fatigue
damage in short carbon fiber composites. The distribution of the thermoelastic temperature change
was measured during the fatigue test, as well as the phase difference between the thermoelastic
temperature change and applied loading signal. Evolution of fatigue damage was detected from the
distribution of thermoelastic temperature change according to the thermoelastic damage analysis
(TDA) procedure. It was also found that fatigue damage evolution was more clearly detected than
before by the newly developed thermoelastic phase damage analysis (TPDA) in which damaged area
was emphasized in the differential phase delay images utilizing the property that carbon fiber shows
opposite phase thermoelastic temperature change.

Keywords: nondestructive evaluation; thermoelastic stress analysis; phase analysis; infrared camera;
short carbon fiber reinforced plastics

1. Introduction
Carbon fiber reinforced plastic (CFRP) has been widely employed for the primary structural
members of transportation vehicles such as automobile, aircraft, or spacecraft, utilizing its excellent
specific strength and specific rigidity in contrast with the metal. Short carbon fiber composite
materials have been receiving a lot of attentions because of their excellent moldability and productivity.
However, they show complicated behaviors in fatigue fracture due to their random fibers orientation.
Therefore, effective nondestructive evaluation techniques are required for detecting and measuring
various types of damages—such as fiber breakage, matrix cracking, and delamination—during fatigue
fracture. Nondestructive evaluation techniques using infrared thermography, i.e., thermographic
NDT has been effectively employed for the detection of delamination defects in fiber reinforced
plastics. The thermographic NDT techniques based on thermal insulation effect by the delamination
defect have been applied to the nondestructive testing of composite materials. The combination of
active pulse heating and subsequent transient temperature measurement was effectively employed
for sensitive detection of delamination damage in composites. Avdelidis et al. [1,2] reviewed NDT
techniques using transient temperature distribution for CFRP as well as data processing techniques
for pulsed thermography. Chatterjee et al. [3] compared the defect imaging performance among three

Sensors 2017, 17, 2824; doi:10.3390/s17122824 www.mdpi.com/journal/sensors


Sensors 2017, 17, 2824 2 of 9

transient thermography techniques, e.g., pulsed, lock-in, and frequency modulated thermography.
Maldague et al. [4] developed pulse phase infrared thermography for improving the resolution of
defect images and contributed to nondestructive evaluation of composite materials. As the transient
temperature data analysis scheme, principal component analysis (PCA) has been investigated by
several researchers to extract small changes in temperature trends [5–7].
In contrast with these thermal insulation techniques, damage evaluation techniques based on the
thermoelastic stress analysis (TSA) using infrared thermography were examined for CFRP composites.
TSA is a well-established experimental technique for the evaluation of the stress field in isotropic
materials, and it has come into widespread use in the industry as an effective, experimental, full-field
stress measurement technique [8–12]. Innovative research works on the TSA technique are found
in structural integrity evaluations for steel structures related with fracture mechanics evaluations
and/or fatigue damage analyses. As for steel structural members, evaluation schemes of the fracture
mechanics parameters have been investigated, taking advantage of the TSA technique [13]. The fracture
mechanics parameters such as stress intensity factor or J-integral have been directly determined from
stress distribution around crack tips [14,15]. These techniques have been successfully applied to
the fatigue life assessment for metallic materials. Tomlinson et al. [16] investigated fatigue crack
propagation under mixed mode loading. Diaz et al. [17] applied their improved TSA technique for
evaluating stress intensity factors in the fatigue test of weld specimen, demonstrating the feasibility of
TSA for crack growth analysis influenced by crack closure or residual stress field.
Thermoelasticity for orthotropic materials have been studied by many researchers and the
TSA technique has been employed as a powerful tool for evaluating impact or fatigue damage in
composite materials and structures [18–22]. Paynter and Dutton [23] applied the TSA technique
to wind turbine blade composite structure with successful results on damage evaluation using
second harmonic signal correlation. Jones and Molent [24] showed the applicability of the TSA
technique for in situ measurement during redesign, reinforcement or repair in aircraft structures.
Emery and Dulieu-Barton [25] applied the TSA technique to fatigue damage evaluation in laminated
glass fiber epoxy materials, in which TSA was employed as a powerful experimental tool for
analyzing complicated fatigue damages such as fiber breakage, matrix cracking, and delamination
damage in composites. Fruehmann and Dulieu-Barton [26] applied the TSA technique for the
assessment of fatigue damage evolution in woven composite materials. Uenoya and Fujii [27]
developed thermoelastic damage analysis (TDA) for the early damage detection in plain-woven CFRP.
Yoshida et al. [28] applied TSA technique for the characterization of impact damages in cross-plied
carbon fiber/thermoplastic composites.
The present authors [29] investigated the relationship between the fiber orientation angles and
the phase delay of the thermoelastic temperature change from the applied loading signal. In this study,
distributions of the thermoelastic temperature change and its phase delay were measured for short
carbon fiber reinforced plastics during fatigue testing. The relationship between the phase difference
and fiber orientation angle was investigated to develop a new thermoelasticity approach for fatigue
damage identification for the short carbon fiber reinforced plastics. Fatigue damages were evaluated
according to the conventional TDA procedure [27] as well as the newly developed phase-delay based
damage characterizing technique “thermoelastic phase damage analysis (TPDA)”.

2. Thermoelastic Stress Analysis Using Infrared Thermography


Dynamic stress change cause a very small temperature change under adiabatic conditions in
a solid. This phenomenon is known as the thermoelastic effect and is described by Lord Kelvin’s
equation [30], which relates the temperature change (∆TE ) to the sum of the changes in the principal
stresses (∆σ) under cyclic variable loading as follows.
α
∆TE = − T∆σ = −kT∆σ (1)
ρC p
Sensors 2017, 17, 2824 3 of 10

Sensors 2017, 17, 2824 3 of 10



TE   T   kT  (1)
C p
Sensors 2017, 17, 2824 TE   T   kT  3(1)
of 9
 C p
Here, α is the coefficient of thermal expansion, ρ is the mass density, Cp is the specific heat at
constant
Here,pressure
α theand
is the T is the of
coefficient absolute
thermal temperature.
expansion, The ρ is coefficient
the mass k is called
mass density,
density, Cppthermoelastic
is the specificconstant.
the specific heat at
at
Here, α is coefficient of thermal expansion, ρ is the C is heat
The sum
constant of the
pressure changes
and T is in
the the principal
absolute stresses
temperature. (∆σ)
The is obtained
coefficient k by
is measuring
called the
thermoelastic temperature
constant.
constant pressure and T is the absolute temperature. The coefficient k is called thermoelastic constant.
change
The sum(∆T E) using infrared thermography.
The sum ofofthe
thechanges
changes in the
in the principal
principal stresses
stresses (∆σ) (∆σ) is obtained
is obtained by measuring
by measuring the temperature
the temperature change
change As thermoelastic
(∆T temperature changes are very small and sometimes hidden by the thermal
E) using infrared thermography.
(∆TE ) using infrared thermography.
noiseAsofthermoelastic
the infrared camera,temperature lock-in infrared
changes thermography using reference signals synchronized
As thermoelastic temperature changes arearevery very small
small andand sometimes
sometimes hiddenhiddenby the bythermal
the thermal
noise
with
noise the
of stress
the changes
infrared is commonly
camera, lock-in employed
infrared to improve
thermography the accuracy
using of
referencestress measurements.
signals synchronized The
of the infrared camera, lock-in infrared thermography using reference signals synchronized with the
TSA technique
with the uses
stress changes a lock-in algorithm
is commonly with
employed a reference-loading
to improve signal
the accuracy extracted
of stress from the
measurements. load cell
The
stress changes is commonly employed to improve the accuracy of stress measurements. The TSA
or strain
TSA gaugeuses
technique to improve
a lock-in thealgorithm
signal–noise ratio.
technique uses a lock-in algorithm with a with a reference-loading
reference-loading signal extracted
signal extracted from thefrom load the
cell load cell
or strain
or When
strain gaugea tension
to and
improve compression
the signal–noiseloading in sinusoidal waveform as shown in Figure 1 is applied
ratio.
gauge to improve the signal–noise ratio.
to materials with positive
When aa tension value of thermoelastic constantwaveform
k, the thermoelastic intemperature change
When tension and and compression
compression loadingloading in in sinusoidal
sinusoidal waveform as as shown
shown in Figure 11 is
Figure is applied
applied
shows
to opposite
materials withphase waveform
positive value of against
of the loading
thermoelastic waveform.
constant the In
k, the this study, the
thermoelastic value of the
temperature phase
change
to materials with positive value thermoelastic constant k, thermoelastic temperature change
difference
shows opposite ∆θ E was defined as the difference in phase between the thermoelastic temperature change
opposite phase
shows phase waveform
waveform againstagainst the the loading
loading waveform.
waveform. In In this
this study,
study, thethe value
value of of the
the phase
phase
and the loading
difference ∆θ ∆θE was signal as shown in the figure. In this case, phase difference
defined as the difference in phase between the thermoelastic temperature change ∆θ E is 180 deg. In this
difference E was defined as the difference in phase between the thermoelastic temperature change
study,
and the thermoelastic
loading signaltemperature
as shown
shown in measurement
in the figure.
figure. In Inwasthisconducted
case, phase
phase fordifference
the short carbon
∆θE is fiberdeg.
is 180
180 reinforced
and the loading signal as the this case, difference ∆θ E deg. In this
In this
plastics,
study, and the
thermoelastic thermoelastic
temperature temperature
measurement change
was ∆T E and the phase difference ∆θE was obtained
conducted for the short carbon fiber reinforced
study, thermoelastic temperature measurement was conducted for the short carbon fiber reinforced
from experimental
plastics, and the data.
the thermoelastic
thermoelastic temperature change change ∆T ∆TE and the phase difference ∆θE was obtained
plastics, and temperature E and the phase difference ∆θ E was obtained
from experimental
from experimental data. data.

Figure 1. Definition of phase in thermoelastic temperature change.


Figure
Figure 1.
1. Definition
Definition of
of phase
phase in
in thermoelastic temperature change.
thermoelastic temperature change.
3. Experimental Setup
3. Experimental
3. Experimental Setup
Configurations
Setup of the CFRP specimen employed in this study are shown in Figure 2. The
specimens were cut from
Configurations laminated short fiber CFRP sheet withstudy
vinyl are
estershown
resin andFigure
25.4 mm 2. long
Configurations ofofthe
the CFRP
CFRP specimen
specimen employed
employed in thisinstudy
this are shown in Figurein The
2. The specimens
carbon
specimensfiber bundles.
were cut fromEach bundle
laminated was
short composed
fiber CFRP of 12,000
sheet short
with carbon
vinyl ester fibers.
resin The
and mass
25.4 mm content
long
were cut from laminated short fiber CFRP sheet with vinyl ester resin and 25.4 mm long carbon fiber
(wt %) of
carbon resin and fiber
fiber wasbundle
67 and 33, respectively.ofThe specimen has circular notches with acontent
radius
bundles. Eachbundles.
bundle was Eachcomposed was composed
of 12,000 12,000
short carbon short
fibers. Thecarbon fibers.
mass content The%)
(wt mass
of resin and
of 2
(wt %)mm.
of 67
resin
fiber was andand
33,fiber was 67 and
respectively. The33, respectively.
specimen The specimen
has circular has circular
notches with a radiusnotches
of 2 mm.with a radius
of 2 mm.

Figure 2. Configurations
Figure 2. Configurations of
of employed
employedshort
shortcarbon
carbonfiber
fiberreinforced
reinforcedplastic
plasticspecimen.
specimen.
Figure 2. Configurations of employed short carbon fiber reinforced plastic specimen.
Cyclic-axis sinusoidal waveform loading with a frequency f of 7 Hz and a stress ratio R = 0.1
was applied to the specimen by an electrohydraulic fatigue testing machine. Microscopic visible
Sensors 2017, 17, 2824 4 of 10

Sensors 2017, 17, 2824 4 of 9


Cyclic-axis sinusoidal waveform loading with a frequency f of 7 Hz and a stress ratio R = 0.1 was
applied to the specimen by an electrohydraulic fatigue testing machine. Microscopic visible images
images
on on the specimen
the specimen surface surface
and sideand side were
surface surface werebytaken
taken by microscope.
optical optical microscope. The temperature
The temperature change
change on the specimen surface was measured by infrared thermography with
on the specimen surface was measured by infrared thermography with an MCT array detector an MCT array detector
(CEDIP Inc. (FLIR
(FLIR Systems
Systems Inc.),
Inc.), Croissy
Croissy Beaubourg,
Beaubourg, Fracnce
Fracnce(Wilsonville,
(Wilsonville, USA),
OR, USA), Titanium530L).
Titanium530L). The
The specifications
specifications andand setting
setting of this
of this infrared
infrared camera
camera areare listed
listed inin Table1.1.The
Table Thethermoelastic
thermoelastictemperature
temperature
∆TEE and
change ∆T and the
the phase
phase difference
difference ∆θ ∆θEE were
were obtained
obtained fromfrom experimental
experimental data
data as
as shown
shown in the
forgoing paragraph.

Table 1. Specifications and setting of employed infrared camera


Table

Infrared Detector MCT


Infrared Detector MCT
Detectable wavelength 7.7–9.3 μm
Detectable wavelength 7.7–9.3 µm
Number of detectors 320 × 256
Number of detectors 320 × 256
Temperature resolution (NETD) 25 mK
Temperature resolution (NETD) 25 mK
Framing rate
Framing rate 373 Hz 373 Hz
Time of data acquisition
Time of data acquisition 10 s 10 s

4. Experimental Results
4. Experimental Results
4.1.
4.1. Effect
Effect of
of Fiber
Fiber Orientation
Orientation on
on Thermoelastic
Thermoelastic Temperature Change
Temperature Change
Before
Before conducting
conducting fatigue
fatigue test,
test, the
the effect
effect of the fiber
of the fiber orientation
orientation angle
angle onon the
the thermoelastic
thermoelastic
temperature
temperature change
change was was experimentally
experimentally investigated
investigated using
using aa CFRP
CFRP specimen
specimen without
without notches.
notches.
Measurement
Measurement of the thermoelastic temperature change was conducted under the applied maximum
of the thermoelastic temperature change was conducted under the applied maximum
stress
stress σσmax
max = 100 MPa. The fiber orientation angle φf is defined as shown in Figure 3; thus φf is equal
= 100 MPa. The fiber orientation angle φf is defined as shown in Figure 3; thus φf is equal to
to 0 degrees
0 degrees when when a fiber
a fiber bundle
bundle is oriented
is oriented in parallel
in parallel withwith the loading
the loading axis.axis. The distribution
The distribution of theoffiber
the
fiber orientation angle φ f was measured after the thermoelastic stress measurement as described
orientation angle φf was measured after the thermoelastic stress measurement as described below:
below:
(1) (1) thin
surface surface thin
resin resin
layer waslayer was removed
removed by polishing
by polishing for exposingfor carbon
exposing carbon
fiber fiber(2)
bundles; bundles;
optical
(2) optical image of the surface carbon fiber bundles was taken by digital camera; (3) fiber
image of the surface carbon fiber bundles was taken by digital camera; (3) fiber orientation angle φf orientation
angledetermined
was φf was determined
by imageby image processing
processing program program
developed developed
by Enomae by [31].
Enomae [31].

Figure 3.
Figure Definition of
3. Definition of fiber
fiber orientation
orientation angle.
angle.

optical image
Obtained optical image ofofthe thesample
sample surface,
surface, distribution
distribution of fiber
of fiber orientation
orientation angles,
angles, and
and thermoelastic temperature change
thermoelastic temperature change ∆TE obtained ∆T E obtained for CFRP specimen without notches
for CFRP specimen without notches are shown in are shown
in Figure
Figure 4. 4. The
The relationship
relationship betweenthe
between thethermoelastic
thermoelastictemperature
temperaturechange
changeandandthethe fiber
fiber orientation
angle φ is shown
φff is shown in in Figure
Figure 5.
5.ItItisisfound
foundfrom
fromthese
thesefigures
figuresthat
that∆T∆T measured
E measured
E inin
thethe area
area where
where φf φisf
is around
around ◦
90°90takes
takes higher
higher values
values comparedwith
compared withthose
thosemeasured
measuredininthe thearea
areawhere
whereφφff isis around 0◦ .
around 0°.
Furthermore, the relationship between the fiber orientation angle φ φff and the phase difference ∆θ ∆θEE is
shown in Figure 6. It can be seen from the figure that the phase difference difference ∆θ 180°◦
∆θEE takes the value of 180
where the fiber orientation angles are around 90◦ . On the other hand, ∆θ E takes 0◦ where φf is around 0◦ .
Sensors 2017, 17, 2824 5 of 10

where the fiber orientation angles are around 90°. On the other hand, ∆θE takes 0° where φf is
Sensors2017,
Sensors 2017, 17,2824
2824 55of
of10
10
around 0°. 17,17,
Sensors 2017, 2824 5 of 9
It was
where thereported by Sugimoto
fiber orientation
orientation anglesand areIshikawa
around 90°. [32]On
90°. that
thethe coefficient
other hand, ∆θ ∆θofE takes
thermal expansion
0° where φf isof
where the fiber angles are around On the other hand, E takes 0° where φf is
carbon
around fiber0°. in longitudinal direction has negative value, hence the thermoelastic constant k of
around
It was 0°.
reported
unidirectional
It was CFRPby
reported
It was reported
Sugimoto
takes negative
by Sugimoto
by Sugimoto and Ishikawa
value.
and
and Ishikawa
Ishikawa
[32] [32]
It is supposed that that
[32] that
the
thatcoefficient
carbon
the
the
of thermal
fibers
coefficient
coefficient mainly expansion
shared
of thermal
of thermal theofapplied
expansion
expansion
carbon
of
of
fiber
axial in
carbon longitudinal
load in the
fiber region
in direction
where
longitudinal has
the negative
orientation
direction hasvalue, hence
angles
negative of the
the
value, thermoelastic
carbon
hence fiber
the constant
bundles k
are
thermoelastic of
0° unidirectional
(parallel
constant k with
of
carbon fiber in longitudinal direction has negative value, hence the thermoelastic constant k of
CFRP takes
unidirectional
theunidirectional negative
loading axis). CFRP value. It is
takesnegative
Therefore,negative supposed that carbon
value.ItItisissupposed
the thermoelastic supposed fibers
temperature mainly
thatcarbon
carbon
change shared
fibers the applied
mainlyshared
showed shared axial
applied in
load
theapplied
a coordinate-phase
CFRP takes value. that fibers mainly
◦ (parallel the
the region
axialload
waveform load where
inthe
with the the
that orientation
region whereloading.
of applied angles of
theorientation the
orientationanglescarbon
anglesof fiber
ofthe bundles
thecarbon
carbonfiber are 0
fiberbundles
bundlesare are0° with
(parallelloading
the
0°(parallel with
axial in region where the with
axis).
the Therefore,
loading the
axis). thermoelastic
Therefore, the temperature
thermoelastic change showed
temperature a coordinate-phase
change showed a waveform
coordinate-phase with
the loading axis). Therefore, the thermoelastic temperature change showed a coordinate-phase
that of applied
waveform
waveform withloading.
with thatof
that ofapplied
appliedloading.
loading.

(a) (b) (c)


(a)
(a) (b)
(b) (c)
(c)
Figure 4. Fiber orientation angles and thermoelastic temperature change (σmax = 100 MPa, f = 7 Hz):
(a) optical
Figure
Figure
Figure image;
4.
4. 4. Fiber
Fiber (b) orientation
Fiber orientation
orientation
orientation angle;
angles
angles
angles and(c)
and
and thermoelastic
thermoelastic
thermoelastic
thermoelastic temperature
temperature
temperature
temperature change.
change
change
change (σmax
(σmax
(σ == 100
=100
100MPa, ff == 77 Hz):
MPa, f = Hz):
MPa, 7 Hz):
max
(a)optical
opticalimage;
image;(b)
(b)orientation
orientationangle;
angle;(c)
(c)thermoelastic
thermoelastic temperaturechange.
change.
(a)(a)
optical image; (b) orientation angle; (c) thermoelastic temperature
temperature change.

Figure 5. Relationship
Figure between
5.Relationship
Relationship betweenthermoelastic
thermoelastictemperature change
temperaturechange ∆TEEand
change ∆T andfiber
fiberorientation
orientationangle
angle φf.
Figure
Figure 5. 5.
Relationship between
between thermoelastic
thermoelastic temperature
temperature change∆T∆TE Eand
andfiber orientation
fiber angle
orientation φφf.f.φf .
angle

Figure 6. Relationship between phase difference ∆θ E and fiber orientation angle φf .: the phase
difference ∆θ E was measured for specimen in Figures 4 and 5 under σmax = 100 MPa and f =7 Hz.
Sensors 2017, 17, 2824 6 of 10

Figure 6. Relationship between phase difference ∆θE and fiber orientation angle φf.: the phase
difference ∆θE was measured for specimen in Figure 4 and 5 under σmax = 100 MPa and f=7 Hz.
Sensors 2017, 17, 2824 6 of 9

4.2. Fatigue Damage Identification Based on Thermoelastic Temperature Change and Its Phase Delay
Information
4.2. Fatigue Damage Identification Based on Thermoelastic Temperature Change and Its Phase Delay Information
Fatigue
Fatigue test
test was
was conducted
conducted for for the
the short
short carbon
carbon fiber reinforced plastic
fiber reinforced plastic specimen
specimen withwith circular
circular
notches
notches under
underthe theapplied
appliedmaximum
maximum stress σmaxσmax
stress = 180=MPa. As described
180 MPa. in the introduction,
As described Uenoya
in the introduction,
et al. [25]
Uenoya et developed the thermoelastic
al. [25] developed damage
the thermoelastic analysis
damage (TDA)
analysis for the
(TDA) for early damage
the early damage detection
detectionin
plain-woven long carbon fiber reinforced plastics. In this technique, a
in plain-woven long carbon fiber reinforced plastics. In this technique, a differential thermoelastic differential thermoelastic
temperature distribution image
temperature distribution image waswas generated
generated by by subtracting
subtracting an an image
image of of thermoelastic
thermoelastic temperature
temperature
change obtained at certain loading cycle from a reference initial image of
change obtained at certain loading cycle from a reference initial image of thermoelastic temperaturethermoelastic temperature
change. Fatiguedamage
change. Fatigue damage evolution
evolution causes
causes locallocal
stressstress
change, change,
and thisand this is emphasized
is emphasized in the
in the differential
differential
thermoelastic thermoelastic
temperaturetemperature
distributiondistribution
image. Therefore,image. the Therefore, the fatigue
fatigue damage damage
in CFRP in CFRP
sample can
sample can be detected in the differential thermoelastic temperature distribution
be detected in the differential thermoelastic temperature distribution image at the early stage in the image at the early
stage
fatigueinlife.
the fatigue life. In the
In this study, thisreference
study, theinitial
reference
image initial image of thermoelastic
of thermoelastic temperature temperature
change was change
set at
was set at 200 cycles. The obtained TDA images at 10,000; 30,000; 40,000; and
200 cycles. The obtained TDA images at 10,000; 30,000; 40,000; and 44,000 cycles are shown in Figure 7, 44,000 cycles are shown
in Figure
with 7, with the
the infrared imageinfrared image
obtained obtained
after failure after
of thefailure
CFRP of the CFRP
specimen at specimen at 44,221
44,221 cycles. cycles.from
It is found It is
foundimages
TDA from TDA images
that the that the temperature
thermoelastic thermoelasticchange temperature
∆TE onchange ∆TE on
the fracture partthe
wasfracture part with
decreasing was
decreasing with the increasing loading cycles. And the decreasing area of
the increasing loading cycles. And the decreasing area of ∆TE was expanded in the transverse direction∆T E was expanded in the

transverse direction
of the specimen. This of the specimen.
indicates Thisdamage
that fatigue indicates canthat fatigue damage
be detected from the can localbe detected
stress change from the
caused
local
by thestress
changechange
in loadcaused
sharingby the change in
conditions due load sharing
to the damage conditions
evolution.due to the damage evolution.

Figure 7. TDA
Figure7. TDA images
imagesand
andinfrared
infraredimage
imageobtained
obtainedafter
afterspecimen
specimenfailure
failure(σ(σmax
max =
= 180 MPa, f ==77Hz).
Hz).

Distribution
Distribution of of the
the phase
phase difference
difference ∆θ ∆θE obtained at 200; 10,000; 30,000; 40,000; and 44,000 cycles
E obtained at 200; 10,000; 30,000; 40,000; and 44,000 cycles
are
are shown in Figure 8. Change in phase values were found in
shown in Figure 8. Change in phase values were found in the
the region
region where
where the the structural
structural fracture
fracture
of
of CFRP sample was detected. The phase values were changed from 0 to 180 , and the
CFRP sample was detected. The phase values were changed from 0 to 180°,◦ and the phase
phase change
change
region was expanded with the increasing loading cycles. As described in the
region was expanded with the increasing loading cycles. As described in the foregoing paragraph, foregoing paragraph,
∆θ
∆θ EE ==0°
0◦means
meansthat
thatthe
thethermoelastic
thermoelasticeffect
effectof
ofcarbon
carbonfiber
fiberbundles
bundles isis predominant,
predominant, on on the
the other
other hand
hand
∆θ E = 180°◦ means that the thermoelastic effect of matrix resin is predominant. The change in phase
∆θ E = 180 means that the thermoelastic effect of matrix resin is predominant. The change in phase
values indicates
values indicates the
the change
change of load sharing
of load sharing conditions
conditions between
between resin
resin and
and carbon
carbon fiber bundles due
fiber bundles due
to the fatigue damage evolution. We thought that high-sensitivity early fatigue
to the fatigue damage evolution. We thought that high-sensitivity early fatigue damage detection damage detection
technique
technique can canbebe developed
developed by utilizing
by utilizing the change
the change in phasein phase difference
difference ∆θE; we
∆θ E ; we propose propose a
a thermoelastic
thermoelastic phase damage analysis (TPDA). In TPDA a differential phase
phase damage analysis (TPDA). In TPDA a differential phase delay distribution image is generated delay distribution image
by
is generated by subtracting a phase delay image obtained at certain loading cycle
subtracting a phase delay image obtained at certain loading cycle from a reference initial phase image. from a reference
initial
In this phase image.
study the In thisinitial
reference studyphase
the reference initial
delay image wasphase
set atdelay imageThe
200 cycles. wasobtained
set at 200 cycles.
TPDA The
images
obtained TPDA images at 10,000; 30,000; 40,000; and 44,000 cycles are shown in Figure
at 10,000; 30,000; 40,000; and 44,000 cycles are shown in Figure 9. It was found from TPDA images that 9. It was found
from TPDA images
the significant changethat
of the
phasesignificant change
values from 0 toof phase
180 values from
◦ (indicated green0color
to 180° (indicated
in the figures)green color
was found
in
in the
the figures) was found
region where in the region
the structural where
fracture wasthe structural fracture was detected.
detected.
Sensors 2017, 17, 2824 7 of 9
Sensors 2017, 17, 2824 7 of 10

Figure 8. Distribution
Figure of phase
8. Distribution delay
of phase measured
delay measuredatatincreasing loadingcycles
increasing loading cycles (σ=max
(σmax 180=MPa,
180 fMPa, f = 7 Hz).
= 7 Hz).

Figure 9. Results
Figure of thermoelastic
9. Results of thermoelastic phase
phasedamage analysis(TPDA)
damage analysis (TPDA) indicating
indicating early early
fatiguefatigue
damagedamage
evolution
evolution (σmax (σ
= max
180 = 180
MPa,MPa,
f =f =7 7Hz).
Hz).

In this experiment, visible images on the side surfaces of the CFRP specimen were taken by
In this experiment,
digital microscope.visible images
Microscopic on the
visible sidetaken
images surfaces of the
at 30,000; CFRP
35,000; specimen
and werecycles
40,000 loading takenare
by digital
microscope. Microscopic visible images taken at 30,000; 35,000; and 40,000 loading
shown in Figure 10 compared with TPDA results. The microscopic imaging area was indicated by cycles are shown in
Figure 10 compared with TPDA results. The microscopic imaging area was indicated by red broken lines
red broken lines in the figures. It was found that delamination damage was initiated at 30,000 cycles
and it grew
in the figures. and found
It was reachedthat
the front surface of the
delamination specimen
damage was(infrared observation
initiated at 30,000 side). TPDA
cycles andimages
it grew and
show the change in phase values was clearly detected in the region where the delamination fracture
reached the front surface of the specimen (infrared observation side). TPDA images show the change in
was initiated. The phase values were changed from 0 to 180°, and the phase change region was
phase values was clearly detected in the region where the delamination fracture was initiated. The phase
expanded with the increasing loading cycles. As described in the foregoing paragraph, ∆θE = 0° means
values were changed from effect
0 to 180 ◦ , and the phase change region was expanded with the increasing
that the thermoelastic of carbon fiber bundles is predominant, on the other hand ∆θE = 180°

loadingmeans
cycles.that
As thedescribed in the effect
thermoelastic foregoing
of matrix resin is ∆θ
paragraph, E = 0 means
predominant. Thethat
changethe in
thermoelastic
phase valueseffect of
indicates
carbon fiber the change
bundles of load sharing
is predominant, conditions
on the other hand ∆θ Eresin
between = 180 ◦ means
and carbon that
fiberthe
bundles due to theeffect of
thermoelastic
evolution
matrix resin of delamination
is predominant. Thedamage.
change in phase values indicates the change of load sharing conditions
Sensors resin
between 2017, 17,
and2824carbon fiber bundles due to the evolution of delamination damage. 8 of 10

Figure 10. Microscopic visible images of specimen side surface indicating delamination damage
Figure 10. Microscopic visible images of specimen side surface indicating delamination damage
evolution (σmax = 180 MPa, f = 7 Hz).
evolution (σmax = 180 MPa, f = 7 Hz).

5. Conclusions
In this study, TSA using an infrared thermography was applied to the evaluation of fatigue
damage in short carbon fiber composites. In addition to the conventional thermoelastic damage
analysis, the phase difference between the thermoelastic temperature change and the applied loading
Sensors 2017, 17, 2824 8 of 9

5. Conclusions
In this study, TSA using an infrared thermography was applied to the evaluation of fatigue
damage in short carbon fiber composites. In addition to the conventional thermoelastic damage
analysis, the phase difference between the thermoelastic temperature change and the applied loading
signal was measured for short carbon fiber reinforced plastics. Obtained results are summarized as
follows:

(1) It was found from the relationship between the thermoelastic temperature change ∆TE and the
fiber orientation angle φf that ∆TE in the area where φf = 90◦ takes higher value compared with
that in the area where φf = 0◦ .
(2) It was also found from the relationship between the fiber orientation angle φf and the phase
difference ∆θ E that ∆θ E takes 180◦ in the area where φf = 90◦ . On the other hand, ∆θ E takes 0◦
where φf = 0◦ . This is due to the negative thermoelastic constant of carbon fibers and the load
sharing condition between resin and carbon fibers.
(3) Fatigue damage was evaluated according to the conventional TDA procedure as well as the
newly developed phase-delay based damage characterizing technique “thermoelastic phase
damage analysis (TPDA)”. It was found from TPDA images that the significant change of phase
values from 0 to 180◦ (indicating the change in load sharing condition between resin and carbon
fibers due to fatigue damage evolution) was found in the region where the structural fracture
was detected.

Acknowledgments: The authors would like to acknowledge that this research was partly supported by
a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (B: 26289009).
Author Contributions: The work presented in this paper was a collaboration of all authors. S. Nonaka and
K. Hamada made the specimen. Y. Nakamura acquired and processed the experimental data. D. Shiozawa and
T. Sakagami analysis and contributed writing the paper. The manuscript was discussed by all authors.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Avdelidis, N.P.; Almond, D.P.; Dobbinson, A.; Hawtin, B.C.; Ibarra-Castanedo, C.; Maldague, X. Aircraft
composites assessment by means of transient thermal NDT. Prog. Aerosp. Sci. 2004, 40, 143–162. [CrossRef]
2. Avdelidis, N.P.; Gan, T.-H. Non-destructive evaluation (NDE) of Composites: Infrared (IR) thermography of
wind turbine blades. In Non-Destructive Evaluation (NDE) of Polymer Matrix Composites; Karbhari, V.M., Ed.;
Woodhead Publishing Limited: Cambridge, UK, 2013; pp. 634–648, ISBN 978-0-85709-344-8.
3. 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]
4. Maldague, X.; Marinetti, S. Pulse phase infrared thermography. J. Appl. Phys. 1996, 79, 2694–2698. [CrossRef]
5. Rajic, N. Principal component thermography for flaw contrast enhancement and flaw depth characterization
in composite structures. Compos. Struct. 2002, 58, 521–528. [CrossRef]
6. Griefahn, D.; Wollnack, J.; Hintze, W. Principal component analysis for fast and automated thermographic
inspection of internal structures in sandwich parts. J. Sens. Sens. Syst. 2014, 3, 105–111. [CrossRef]
7. Marinetti, S.; Grinzato, E.; Bison, P.G.; Bozzi, E.; Chimenti, M.; Pieri, G.; Salvetti, O. Statistical analysis of IR
thermographic sequences by PCA. Infrared Phys. Technol. 2004, 46, 85–91. [CrossRef]
8. Greene, R.J.; Patterson, E.A.; Rowlands, R.E. Thermoelastic Stress Analysis. In Springer Handbook of
Experimental Solid Mechanics; Sharpe, W.N., Jr., Ed.; Springer Science + Business Media, LLC: New York,
NY, USA, 2008; pp. 743–767, ISBN 978-0-387-26883-5.
9. Dulieu-Barton, J.M.; Stanley, P. Development and applications of thermoelastic stress analysis. J. Strain Anal.
Eng. Des. 1998, 33, 93–104. [CrossRef]
10. Dulieu-Barton, J.M. Introduction to thermoelastic stress analysis. Strain 1999, 35, 35–39. [CrossRef]
11. Pitarresi, G.; Patterson, E.A. A review of the general theory of thermoelastic stress analysis. J. Strain Anal.
Eng. Des. 2003, 38, 405–417. [CrossRef]
Sensors 2017, 17, 2824 9 of 9

12. Sakagami, T. Remote nondestructive evaluation technique using infrared thermography for fatigue cracks in
steel bridges. Fatigue Fract. Engng. Mater. Struct. 2015, 38, 755–779. [CrossRef]
13. Tomlinson, R.A.; Olden, E.A. Thermoelasticity for the analysis of crack tip stress fields—A review. Strain
1999, 35, 49–55. [CrossRef]
14. Ju, S.H.; Lesniak, J.R.; Sandor, B.I. Numerical simulation of stress intensity factors via the thermoelastic
technique. Exp. Mech. 1997, 37, 278–284. [CrossRef]
15. Diaz, F.A.; Patterson, E.A.; Tomlinson, R.A.; Yates, J.R. Measuring stress intensity factors during fatigue crack
growth using thermoelasticity. Fatigue Fract. Engng. Mater. Struct. 2004, 27, 571–583. [CrossRef]
16. Tomlinson, R.A.; Marsavina, L. Thermoelastic investigations for fatigue life assessment. Exp. Mech. 2004, 44,
487–494. [CrossRef]
17. Diaz, F.A.; Yates, J.R.; Patterson, E.A. Some improvements in the analysis of fatigue cracks using
thermoelasticity. Int. J. Fatigue 2004, 26, 365–376. [CrossRef]
18. Lin, S.T.; Rowlands, R.E. Thermoelastic stress analysis of orthotropic composites. Exp. Mech. 1995, 35,
257–265. [CrossRef]
19. Dulieu-Barton, J.M.; Stanley, P. Applications of thermoelastic stress analysis to composite materials. Strain
1999, 35, 41–48. [CrossRef]
20. Emery, T.R.; Dulieu-Barton, J.M.; Earl, J.S.; Cunningham, P.R. A generalised approach to the calibration of
orthotropic materials for thermoelastic stress analysis. Compos. Sci. Technol. 2008, 68, 743–752. [CrossRef]
21. Palumbo, D.; De Finis, R.; Demelio, G.P.; Galietti, U. Study of damage evolution in composite materials
based on the Theroelastic Phase Analysis (TPA) method. Compos. Part B Eng. 2017, 117, 49–60. [CrossRef]
22. Sfarra, S.; Perilli, S.; Ambrosini, D.; Paoletti, D.; Nardi, I.; de Rubeis, T.; Santulli, C. A proposal of
a new material for greenhouses on the basis of numerical, optical, thermal and mechanical approaches.
Constr. Build. Mater. 2017, 155, 332–347. [CrossRef]
23. Paynter, R.J.H.; Dutton, A.G. The use of a second harmonic correlation to detect damage in composite
structures using thermoelastic stress measurements. Strain 2003, 39, 73–78. [CrossRef]
24. Jones, R.; Molent, L. Application of constitutive modeling and advanced repair technology to F111C aircraft.
Compos. Struct. 2004, 66, 145–157. [CrossRef]
25. Emery, T.R.; Dulieu-Barton, J.M. Thermoelastic Stress Analysis of damage mechanisms in composite materials.
Compos. Part A 2010, 41, 1729–1742. [CrossRef]
26. Fruehmann, R.K.; Dulieu-Barton, J.M.; Quinn, S. Assessment of fatigue damage evolution in woven
composite materials using infra-red techniques. Compos. Sci. Technol. 2010, 70, 937–946. [CrossRef]
27. Uenoya, T.; Fujii, T. Damage characterization of woven fabric composite materials by thermoelastic analysis.
J. Soc. Mater. Sci. Jpn. 2000, 49, 941–947. [CrossRef]
28. Yoshida, T.; Uenoya, T.; Miyamoto, H. Impact damage characterization in cross-plied carbon fiber/
thermoplastic composites using thermoelastic stress analysis. Proc. SPIE 2012, 8347. [CrossRef]
29. Sakagami, T.; Shiozawa, D.; Nakamura, Y.; Nonaka, S.; Hamada, K. Fatigue damage evaluation of short fiber
CFRP based on phase information of thermoelastic temperature change. Proc. SPIE 2017, 10214. [CrossRef]
30. Thomson, W. On the dynamical theory of heat. Trans. R. Soc. 1853, 20, 261–283.
31. Enomae, T. FiberOri8single03.exe. Available online: http://www.enomae.com/FiberOri/index.htm
(accessed on 4 December 2017).
32. Sugimoto, S.; Ishikawa, T. Examination of quantitative infrared stress measurement of CFRP laminates and
its application to non-destructive evaluation. Tech. Rep. Natl. Aerosp. Lab. 1999, 1396, 1–23.

© 2017 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/).