Application of NDT Thermographic Imaging of Aerospace Structures
Application of NDT Thermographic Imaging of Aerospace Structures
Regular article
Keywords: This work aims to address the effectiveness and challenges of Non-Destructive Testing (NDT) inspection and
Active infrared thermography improve the detection of defects without causing damage to the material or operator. It focuses on two types of
Pulsed thermography NDT methods; pulsed thermography and vibrothermography. The paper also explores the possibility of per-
Vibrothermography forming automated aerial inspection using an unmanned aerial vehicle (UAV) provided with a thermographic
UAV
imaging system. The concept of active thermography is discussed for inspecting aircraft CFRP panels along with
Composite
Kissing bond
the proposal for performing aerial inspection using the UAV for real time inspection. Static NDT results and the
further UAV research indicate that the UAV inspection approach could significantly reduce the inspection time,
cost, and workload, whilst potentially increasing the probability of detection.
⁎
Corresponding author.
E-mail address: [email protected] (S. Deane).
https://doi.org/10.1016/j.infrared.2019.02.002
Received 25 December 2018; Received in revised form 5 February 2019; Accepted 6 February 2019
Available online 13 February 2019
1350-4495/ © 2019 Elsevier B.V. All rights reserved.
S. Deane, et al. Infrared Physics and Technology 97 (2019) 456–466
1.1. Composites
457
S. Deane, et al. Infrared Physics and Technology 97 (2019) 456–466
Fig. 3. The bar chart (left) shows the use of composites in Airbus aircraft since 1985. The line graph (right) represents the demand by tons from 2005 up to 2020
[7,5].
experiment in order to control the time between the initial thermal pulse manufactured via manual woven and the other two samples where
excitation and the recording. The acquired data is stored as a 3D matrix unidirectional. The composites consist of laying up pre-preg piles to
where t is the time and x and y are the spatial coordinates. form a laminate stack, the material is then autoclave processed at
There are two main types of thermography, active and passive as 180 °C and 7 bars for a few hours suitable for the thermoset resin cure,
depicted in Fig. 4. Passive thermography is an inspection without the this is according to the supplier’s specification (Hexcel). There were
use of an external excitation source, its commonly used for inspecting three specimens, one that was undamaged, and two that had been
materials that have a temperature naturally different than its sur- impacted with a force of 15 and 20 J of energy. The infrared camera
rounding [19]. Active thermography uses external heating dissipation used in this experiment was a FLIR Phoenix, with inSb sensor material,
to create thermal contrast, once excited the thermal waves propagate 3–5 mm, 640 × 512 pixels and allows data acquisition at 50 Hz [5].
through the heated specimen, and when they reach inhomogeneities These camera needs to be cooled to 77 K (i.e. −196.15 °C) during
caused by a medium with different thermal properties, the diffusivity testing. Cooled cameras work by collecting photons of infrared energy
coefficient changes. These inhomogeneities could be defects, such as that pass through the optic. The photons are converted into electrons
cracks, surface cracks, air/water/material inclusions, delamination’s that are stored in an integration capacitor. After a certain period (in-
and debonding damages. Mathematical processing techniques are the tegration time), the charge is read out to a digital count calibrated to
key to gather the necessary data from the raw thermogram and help temperature, assigned a colour or greyscale value and then presented as
visualise the damages within the material [19,5]. a viewable infrared image [20].
The surface of the specimens where positioned parallel to the
2.3. PT equipment camera lens as depicted in Fig. 4. The data was acquired for 40 s with a
1.5 ms integration time, that includes 10 frames before the flashes, plus
The CFRP specimens where manufactured and purposely damaged 1990 frames during cooling a total of 2000 frames where recorded. The
via impact, in the Cranfield composite centre. Sample A4 was software used to acquire the data was RDac from FLIR. For signal
Fig. 4. Images showing the PT testing equipment and the experimental set up [5].
458
S. Deane, et al. Infrared Physics and Technology 97 (2019) 456–466
processing MATLAB and Ir_view from Visiooimage inc were employed. defects is usually assumed. Despite this, it is certain that generated heat
Two advanced processing techniques where used; PCT (principal depends on frequency and position of the excitation source [21]. Vi-
component thermography) and PPT (pulsed phase thermography) [5]. brothermography is becoming a popular technique to detect cracks in
metallic structures and several types of damages in composite mate-
3. Results & discussion I rials.
Sonic waves vibrate between 20 Hz and 20 kHz, which is audible for
3.1. Experiment humans. However, the range for ultrasonic waves, which is beyond
humans’ audible capabilities, is between 20 kHz and 1 MHz, although
The figures below (Figs. 5–11) are all the processed results. A re- the transducers that are usually used in VT inspections operate between
flection test means that the camera and the excitation source are on the 15 and 50 kHz. Mechanical elastic waves such as sonic and ultrasonic
same side (as seen in Fig. 4). Generally, reflection is more suited to require a medium to travel, and travel faster through solids and liquids
detect defects closer to heated surface. PPT (pulsed phase thermo- as opposed to air [1]. The VT approach consist of using a coupling
graphy) was one of the methods used, with the acquired temperature media such as a piece of moisten fabric, water-based gels or aluminium,
data it varies with time and is transferred to phase data, which varies
with frequency, the frequencies which show adequate data are below
each image in hertz. PCT (principal component thermography) was the
other, method used, the data is re-ordered by variance, this method
finds both time series and spatial patterns, most of the variability is
contained in the first component: EOF1 (empirical orthogonal func-
tion). Usually, EOF1 shows thermal variations related to surface
heating. Later EOFs are typically related to surface and subsurface da-
mage. Comparing both techniques, for this specific data set in Fig. 5 it
seems that the PCT technique displays the damage clearer [5].
4. Experimental procedure II
Fig. 6. Sample A4 (15 J), Reflection rear face PPT and PCT. The rear surface
is cracked, this is interesting because this side is significantly more damaged than
4.1. Vibrothermography (VT)
the impacted side. This is a prime example of barely visible impact damage [5].
Fig. 5. Sample A4 (15 J woven), front face (impacted side), PPT and PCT.
The damage is barely visible in the original image of the composite (Far right Fig. 8. Sample 20 J Demo, Reflection front face, PPT. The damage is visible
image), the white marks are lettering from labelling the specimen. The 0.05 Hz on the front face, how, it’s clear that the damage is much worse subsurface and
PPT image, shows clear subsurface damage [5]. spreads further than the impacted area [5].
459
S. Deane, et al. Infrared Physics and Technology 97 (2019) 456–466
Fig. 9. Sample 20 J Demo, Reflection rear face. PPT. The rear face has some substantial damage. The PPT locates the surrounding internal damage [5].
There are two main analogues that can be considered when gath-
ering data. The two techniques are lock-in and burst VT. Note the fre-
quency can be modulated in addition to the amplitude. When acquiring
data, the wobbulation procedure is often used, this procedure seeks to
cover a range of different frequencies in a single experiment, the pur-
pose is for the user to predict the right frequency for the specific ap-
plication as each application is different and requires tailored inspec-
tions [22]. The amount of damage detection is proportional to the
amount of excitation, too much power (higher frequency) will be costly
and will risk damage to the specimen. Therefore, using the minimum
amount of power required to acquire satisfactory results is becoming a
major interest for researchers [24].
The transducer is made of a stack of pie piezoelectric elements and
Fig. 10. Sample 20 J Demo, Reflection front face, Transmission. The test concentrated in a titanium horn that acts like a hammer [22]. The
has adequately located the internal damage [5]. specimen must be firmly restraint without damage, the immobilized
specimen seeks to avoid cantilever effects, clapping, and sliding of the
between the transducer and the specimen to improve contact [1]. On transducer [22]. For maximum performance the coupling of the ultra-
the contrary, there are currently ultrasonic non-contact inspection sound into the specimen needs to be improved, to do this the transducer
methods being investigated in many areas, this is more appealing due to horn should be pressed against the specimen. It is recommended to add
the reduced risk of damage. some sort of material between the transducer and the specimen to avoid
The traditional VT inspection works by placing a transducer in damage to the specimen, to improve contact and to eliminate any
contact with the specimen assisted by a coupling media as before possible misalignment [22]. In the event of bad coupling it can result in
mentioned (material, gel, etc.). The ultrasonic waves flow through the poor ultrasound transmission, the specimen will then be subject to
specimen and will travel without disruption through a homogenous unwanted heat at the injection point, which can ultimately affect the
material, however if there is an internal defect present this will disrupt inspection results. Crack detection is extremely sensitive to whether the
the usual flow of the waves producing a compilation of effects such as specimen is fixed or free, in Fig. 12 the results from the constrained
the waves being absorbed, scattered, and dispersed, this is present in specimen proved the most accurate [23].
the form of heat. An IR camera can observe the surface of the specimen The temperature of the cracks will usually increase slowly when the
because the heat that is generated will travel by conduction in all di- excitation is injected into the specimen, as a result the thermal waves
rections and a defect signature data can be captured. VT is ideal for travel by conduction to the surface, which can then be detected by the
detecting crack and delamination’s, the ultrasonic waves assist in dis- IR camera. VT is a fast inspection technique, slightly faster than PT,
covering both internal and surface defects [22]. within seconds the defect usually appears therefore the acquisition will
460
S. Deane, et al. Infrared Physics and Technology 97 (2019) 456–466
Fig. 11. Sample No Name, Reflection front, back, transmission PCT. Undamaged composite sample, which was manufactured the exact same way as the 20 J
sample, the different NDT test prove that this sample is damage-free [5].
then end. The fast inspection is necessary because longer inspections surfaces are in intimate contact but with little bonding if any. Detection
will result in an excess of heat and increase the probability of the of kissing bonds defects is challenging for all NDT techniques given
specimen getting damaged. The experiment is controllable, this is be- that, unlike other type of defects such as delamination’s there is no
cause the heat is generated selectively at the location of defects, and if significant separation between adherent and/or adhesive surfaces.
certain parameters of excitation need changing, repeat experiments can These types of defects can considerably reduce the impact strength and
be made in quick success. the fatigue life of the material causing premature failure.
The three inspected specimens are all composite material with ar-
4.3. VT equipment tificial defects as shown in Fig. 14.
The data processing is usually easier to manage when using VT, this • Specimen N-D-F4P is composed of one layer of adhesive sand-
is since the source of the heat is the damage itself [23]. The IR camera wiched by two layers of adherent (carbon fiber) and contains no
used to capture the data, is the same as used in the PT experiment defect, it was purposely used for reference only.
explained in 2.1. The transducer used is a 20 kHz ultrasound generator • Specimen W-DA-F4 is similar to N-D-F4P but with a adhesive film
(BRANSON 2000), as seen in Fig. 13. release semi-circular (“thumb nail”) shape defect, simulating a kis-
The experiment will consist of inspecting three different specimens, sing bond defect.
to search for defects but more specifically concentrating on kissing • SpecimenW-DC-F2 is similar to W-DA-F4 but with an additional
bond defects. Kissing bonds can be defined as a kind of defect in which adhesive film on top of the one with the simulated defect.
Fig. 12. The mechanical wave can be more easily converted into heat when the specimen is entirely fixed [24].
461
S. Deane, et al. Infrared Physics and Technology 97 (2019) 456–466
Fig. 13. Image displaying the VT testing equipment and the experimental set up.
5. Results & discussion II crack, which can be detected using the IR camera. Not only as this
technique discovered a clear kissing bond defect but the results also
5.1. Experiment show an apparently real delamination.
462
S. Deane, et al. Infrared Physics and Technology 97 (2019) 456–466
Fig. 15. Image showing the processed results from the vibrothermographic inspection on sample W-DA-F4.
Fig. 16. Image showing the processed results from the vibrothermographic inspection on sample W-DC-F2.
DC-F2 and W-DA-F4, from one side and the other (front and back), after Fig. 19 is an ultrasound testing inspection. The defects are suc-
processing by PPT and PCT. cessfully located and are more visible than the PT results in Fig. 18. The
Additionally, several indications of real defects seem to be observed. size of the defect can also be measured during processing, the defect
For instance, possible delamination’s were found and indicated with present has a radius of 5 mm. The vibrothermography results are pro-
1
red circles. (See Fig. 15). mising for the detection of kissing bond defects.
Sample W-DC-F2: Vibrothermography:
There was no kissing bond defect detected in specimen N-D-F4P (as 5.3. Proposed UAV
expected) however, delamination-like defects were identified. Further
testing using alternative NDT techniques or destructive testing could be Although a cooling camera was used in the beforementioned NDT
used in the future to validate the nature of these indications. (See tests, this type of camera will not be feasible for a UAV thermographic
Fig. 16 and Fig. 17). inspection. Cooling cameras usually require cryogenic cooling, which is
Sample N-D-F4P: Vibrothermography: an excess of weight. The equipment that will be used in future UAV
For instance, additional tests using long-pulse thermography in NDT testing is a DJI M210, which carries an RGB and thermal camera.
transmission mode could possibly detect such kind of defects. However, The weight of the UAV is approximately 3.84 kg with a maximum
this test would require access to both sides of the specimen, which is not payload of 2.3 kg with 2 batteries [26]. The thermal camera weighs
always possible in industry e.g. an aircraft wing panel after assembly. 0.27 kg and is 640 × 512 pixels, which is the exact same number of
Figs. 18 and 19 show some examples of previous IR NDT inspections pixels as the FLIR Phoenix IR camera that was used [26]. It has a 13 mm
of similar specimens [26]. Fig. 18 is a pulsed thermographic inspection; lens and is radiometric, therefore it uses a set of techniques to measure
the defects are located however they are not easily visible in this case. electromagnetic radiation, including visible light, this means that the
acquired data can be extensively analysed after the survey in order to
visualise and understand the data such as the temperature variations
1
For interpretation of color in Fig. 15, the reader is referred to the web [27,5].
version of this article. UAV dynamic inspection has the potential to provide a complete
463
S. Deane, et al. Infrared Physics and Technology 97 (2019) 456–466
Fig. 17. Image showing the postprocessed results from the vibrothermographic inspection on sample N-D-F4P.
portrait of the aircraft structure much faster than static inspections. For images, which will reveal heat patterns and thermal anomalies [12].
a UAV inspection an external excitation source will be employed to Transient thermography with flash and step heating where the methods
excite the surface. When inspecting something of this significance, the used in this inspection. An inspection took place and gathered data on
best results need to be obtained. Therefore, finding the most suitable different distances between the camera and the samples (2, and 6 m).
excitation source is critical, vibrothermography may not be the most Along with the UAV equipped with the IR camera, there was a fixed IR
suitable due to the complexity of the setup. On the contrary, there are camera, therefore two sets of data could be acquired and compared. A
several alternative excitation sources more practical for UAV applica- thermally uniform area of a sample is delimited, and the IR cameras
tions. For instance, halogen lamps (heating the surface for longer), Heat captured the data [12,5].
Blankets or a Laser.
Along with the liquid penetrant NDT method as mentioned in the
introduction, a UAV thermographic inspection was tested for civil en- 5.4. Performing an NDT inspection
gineering applications (see Fig. 20B). The UAV is equipped with a FLIR
A320, this IR camera is connected to a computer via a cable, one is a The purpose of this paper is to present various results from active
power cable and the other provides real time radiometric thermal thermography on composite materials (locating different types of defects
within the materials), and integrate the proposed developed methodology
Fig. 18. Past pulsed thermographic inspection from similar specimens as used in the before mentioned VT inspection [25].
464
S. Deane, et al. Infrared Physics and Technology 97 (2019) 456–466
Fig. 19. Past ultrasonic thermographic inspection from similar specimens as used in the before mentioned VT inspection [25].
into a UAV thermographic inspection system. The autonomous system is separate regions of interest, and separated images will be taken and then
to be tested in real life application on a full-size passenger aircraft wing stitched together as a mosaic image, this segregation will ensure the
before its assembly to the fuselage. In theory, it is the same method as the subject is excited uniformly. Once the data is acquired, it will then need to
beforementioned experiments, the only difference is the camera will be be processed to improve results, when post processing, many properties
moving. A UAV carrying an IR camera can provide much more mobility need to be considered [5];
and allow for inspections that would usually be difficult and time con-
suming when using the traditional inspection techniques. There are • Thermal Properties: conductivity, diffusivity, effusivity, specific
however, several challenges that need to be taken into consideration for heat.
using a UAV for NDT. The future plan is to inspect full size composite • Spectral Properties: emissivity, absorption, reflection, transmission.
aircraft and also composite panels straight from the manufacturing line, • Material’s properties-characteristics: density, porosity, thickness,
to verify their worthiness and quality assurance. There are some limita- geometry [19,5].
tions for this proposed idea. Due to the fact the inspection is inside it
could provide problems for the UAVs GPS system, which will result in an With these properties understood, data process can be performed,
unstable flight and affect the data as before-seen. There are a few ways to for instance; when a material presents voids or pores in its structure,
combat this issue, one of them is to set up a localisation scheme by dis- then its thermal conductivity and density decreases, its thermal diffu-
tributing some ultrasonic sensors around a specific geographic area, the sivity is altered and so the conduction of heat transfer within the ma-
UAV will communicate with these sensors to understand its location, terial is affected. For instance, materials with low-effusivity values will
which will ultimately result in stabilisation. Once the flight is ready – due present higher temperatures. With all this information taken into con-
to the size of aircraft parts, the subject will need to be segregated into sideration, a better damage assessment can be done [19,5].
Fig. 20. A: DJI M210 equipped with thermal camera [27]. Fig. 20 B: UAV equipped with thermal camera [12]. Fig. 20 C: Comparison of front face temperature
responses to flash heating in the case of a large uniform irradiation with a fixed camera and a stabilized camera with UAV at two distances (2 m and 6 m) [12,5].
465
S. Deane, et al. Infrared Physics and Technology 97 (2019) 456–466
6. Conclusion [4] T. Ullmann, T. Schmidt, S. Hofmann, R. Jemmali, In-line quality assurance for the
manufacturing of carbon fiber reinforced aircraft structures, 2nd International
Symposium on NDT in Aerospace, (2010).
This study contributes to reducing the time and cost of NDT com- [5] S. Deane, N. Avdelidis, C. Ibarra Castanedo, H. Zhang, H. Yazdani Nezhad,
posite inspections. The research shows the effectiveness of two different A.A. Williamson, X.P.V. Maldague, A. Tsourdos, Autonomous systems imaging of
excitation sources used for NDT inspection. The raw data that is ac- aerospace structures, 10th International Symposium on NDT in Aerospace. 2018
Dresden: DGZfp, (2018) https://www.researchgate.net/publication/329558466_
quired often contains noise caused by the environment such as, the Autonomous_Systems_Imaging_of_Aerospace_Structures.
reflections, the emissivity variations of the specimen or if the excitation [6] J. Pora, Airbus, Composite materials in the Airbus A380 - from history to future,
was non-uniform, it is important to take this into consideration and to paper-1695, 13th international conference on composite materials Beijing:
International Committee On Composite Materials, (2001) http://www.iccm-
employ specific algorithms for image processing in order to produce central.org/Proceedings/ICCM13proceedings/SITE/PAPERS/paper-1695.pdf.
accurate results. Signal processing techniques, such as PPT and PCT, [7] G. Hellard, Composites in airbus, a long story of innovations and experiences, in: Dr
allow to improve defect contrast and hence, visualise the extent of the Roland Thévenin (Ed.), Global Investor Forum. Airbus, 2016.
[8] M. Mrazova, Advanced composite materials of the future in aerospace industry,
damage and the surrounding areas. When acquiring data, the value of
Incas Bullet. 5 (3) (2013) 139–150.
each pixel will represent the temperature at a specific point. The tem- [9] K.T.N. Materials, R. Young, K. Newton, T. Dunhill, C. Huggins, A landscape for the
perature profile of the point corresponding to the healthy area is dif- future of NDT in the UK economy, Brochure, Materials KTN, Knowledge Transfer
ferent from that of the point corresponding to the defective area. Using Networks, Materials KTN, NPL Product Verification Programme, Materials
Knowledge Transfer Network (2014) Accessed December 13, 2018 http://www.
further digital image processing it is possible to enhance the images to bindt.org/downloads/Materials-KTN-Future-of-NDT-in-UK-economy.pdf.
reveal the location and geometry of the defects [3]. These methods can [10] N.V. Nayak, Composite materials in aerospace applications, Accessed December 12,
potentially become beneficial to the aerospace industry and possibly 2018. International Journal of Scientific and Research Publications Volume 4 (Issue
9): ISSN 2250-3153. (2014) https://pdfs.semanticscholar.org/3a73/
integrated onto a UAV. The results from the UAV inspection indicate an 9ca726a88cf833ba89708a4944c46457e42d.pdf.
excess of noise from the UAV, since this flight was performed without [11] K. Malandrakis, A. Savvaris, J.A.G. Domingo, N. Avdelidis, P. Tsilivis, F. Plumacker,
the use of GPS, this affected the results and the noise was due to the L.Z. Fragonara, A. Tsourdos, Inspection of aircraft wing panels using unmanned
aerial vehicles, 2018 IEEE MetroAeroSpace. Rome, (2018).
stabilisation problem. It is necessary to perform pixel by pixel treatment [12] L.E. Mavromatidis, J.L. Dauvergne, R. Saleri, J.C. Batsale, First experiments for the
in addition to the standard average signal study [12]. To do this, the IR diagnosis and thermophysical sampling using impulse IR thermography from
camera needs to be as static as possible, GPS or other localisation Unmanned Aerial Vehicle (UAV), Proceedings of QIRT Conf., Bordeaux, France,
2014, https://doi.org/10.21611/qirt.2014.213.
techniques can be used to compensate for any stability problems. The [13] W.J. Cantwell, J. Morton, The impact resistance of composite materials — a review,
data acquired has shown that different techniques are better suited for Composites (1991) 347–362, https://doi.org/10.1016/0010-4361(91)90549-V.
different defects. Pulsed thermography can accurately locate delami- [14] Y. Zhao, J. Mehnen, W. Xu, M. Alrashed, S. Abineri, R. Roy, Degradation assessment
of industrial composites using thermography, Proc. CIRP (2015) 147–152, https://
nation’s. Delamination is where the material fails due to the layers se-
doi.org/10.1016/j.procir.2015.07.045.
parating which result in significant loss of mechanical toughness. In the [15] N. Razali, M.T.H. Sultan, F. Mustapha, N. Yidris, M.R. Ishak, Impact damage on
future, a UAV inspection will be performed, and the fidelity and per- composite structures – a review, Int. J. Eng. Sci. (2014) 08–20.
formance of the image processing algorithms will be further evaluated. [16] J.E. King, Failure in composite materials, Metals Mater. Bury St Edmunds 5 (12)
(1989) 720–726 https://www.scopus.com/record/display.uri?eid=2-s2.0-
An excitation source that is small, light and produces enough energy to 0024923342&origin=inward&txGid=8ae9d8890817115ae04936fb5108292a.
penetrate the material, can possibly be equipped on the UAV soon, this [17] D.S. Cairns, Composite Materials for Aircraft Structures, Accessed September 09,
will offer an all in one inspection product and minimise time as there 2018 Montana State University, 2009, http://www.montana.edu/dcairns/
documents/composites/MSUComposites2009.pdf.
will be no need to set up an external excitation source. To quantify [18] M. Genest, M. Martinez, N. Mrad, G. Renaud, A. Fahr, Pulsed thermography for non-
these parameters, the excitation source that is planned to be equipped is destructive evaluation and damage growth monitoring of bonded repairs, Compos.
approximately 200 * 254 * 254 mm, the mass is 1.5 kg and consist of Struct 88 (1) (2009) 112–120, https://doi.org/10.1016/j.compstruct.2008.02.010.
[19] N.P. Avdelidis, D.P. Almond, A. Dobbinson, B.C. Hawtin, C. Ibarra-Castanedo,
four 100w LEDs. Several active thermography tests on a sample com- X.P.V. Maldague, Aircraft composites assessment by means of transient thermal
posite wing will determine how effective this excitation source is before NDT, Prog. Aerosp. Sci. 40 (3) (2004) 143–162.
it is equipped onto the UAV. There are multiple excitation sources on [20] FLIR Systems Inc, What's the Difference Between Cooled & Uncooled Thermal
Detectors? Edited by FLIR Systems Inc. Accessed 12 13, 2018 FLIR Systems Inc. 02
the market that offer a sufficient weight to energy ratio [5].
08, 2016, https://www.youtube.com/watch?v=PYHjAdf8bU0.
[21] M. Szwedo, L. Pieczonka, T. Uhl, Application of vibrothermography in non-
Acknowledgements destructive testing of structures, 6th European Workshop on Structural Health
Monitoring. Krakow: DGZfp, (2013).
[22] C. Ibarra-Castanedo, J.R. Tarpani, X.P.V. Maldague, Non-destructive testing with
This research was part supported by the British Engineering and thermography, Eur. J. Phys. (2013) s91–s109.
Physics Sciences Research Council. Data used in this paper can be found [23] L. Ambrozinski, B. Piwakowski, T. Stepinski, L. Pieczonka, T. Uhl, Pitch-catch air-
at https://doi.org/10.17862/cranfield.rd.7718477. coupled ultrasonic technique for detection of barely visible impact damagesin
composite laminates, EWSHM- 7th European Workshop on Structural Health
Monitoring, Vincent and Mevel, Laurent and Schoefs, Franck, Nantes: Le Cam, 2014
References hal-01022043.
[24] Smart materials and structures lab. 2015. Youtube. 19th October. Accessed
November 27th, 2018. Available at: https://www.youtube.com/watch?v=f-
[1] C. Ibarra-Castanedo, J.M. Piau, S. Guilbert, N.P. Avdelidis, M. Genest, A. Bendada, fYmrQtNsI.
X.P.V. Maldague, Comparative study of active thermography techniques for the [25] H. Yazdani Nezhad, D. Stratakis, D. Ayre, S. Addepalli, Y. Zhao, Mechanical per-
nondestructive, Res. Nondestr. Eval. (2009) 1–31, https://doi.org/10.1080/ formance of composite bonded joints in the presence of localised process-induced
09349840802366617. zero-thickness defects, Proc. Manuf. 16 (2018) 91–98, https://doi.org/10.1016/j.
[2] Bin Industrial Training, What is Thermography (thermal imaging), 25 November. promfg.2018.10.175.
Accessed 12 13, 2018 Bin Industrial Training, 2014, https://www.youtube.com/ [26] DJI Official, 2019. MATRICE 200 SERIES SPECS. [online] Accessed February 04
watch?v=4xulzkBOFKI&t=3s. 2019 Available at: https://www.dji.com/uk/matrice-200-series/info.
[3] A.P. Chrysafi, N. Athanasopoulos, N.J. Siakavellas, Damage detection on composite [27] A.A. Williamson, 2018. Mapair, Frequently Asked Questions. Accessed September
materials with active thermography and digital image processing, Int. J. Therm. Sci. 09, 2018. Available at: https://www.mapair.co.uk/faqs.
(2017) 242–253.
466