operational deflection shape extraction from broadband ...researcharticle operational deflection...

Research ArticleOperational Deflection Shape Extraction from BroadbandEvents of an Aircraft Component Using 3D-DIC inMagnified Images

Angel J Molina-Viedma 1 Elıas Lopez-Alba1 Luis Felipe-Sese 2

and Francisco A Dıaz 1

1Departamento de Ingenierıa Mecanica y Minera Campus Las Lagunillas Universidad de Jaen 23004 Jaen Spain2Departamento de Ingenierıa Mecanica y Minera Campus Cientıfico Tecnologico de Linares Universidad de Jaen23700 Linares Spain

Correspondence should be addressed to Angel J Molina-Viedma ajmolinaujaenes

Received 30 November 2018 Revised 12 February 2019 Accepted 5 March 2019 Published 9 April 2019

Academic Editor Nuno M Maia

Copyright copy 2019 Angel J Molina-Viedma et al is is an open access article distributed under the Creative CommonsAttribution License which permits unrestricted use distribution and reproduction in anymedium provided the original work isproperly cited

Recently many works have shown the capabilities of noninterferometric optical techniques such as digital image correlation tocharacterise modal behaviour ey provide a global insight into the structure or component behaviour which implies massivespatial information unaffordable by traditional sensor instrumentationMoreover phase-basedmotionmagnification (PMM) is amethodology which based on a sequence of images magnifies a periodic motion encoded in phase time-domain signals of thecomplex steerable pyramid filters employed to decompose the images It provides a powerful tool to interpret deformationHowever the interpretation is just qualitative and should be avoided if out-plane motion is recorded as only one camera isemployed To overcome this issue 3D digital image correlation (3D-DIC) has been linked with PMM to provide measurementsfrom stereoscopic sets of images providing full-field displacement maps to magnified images In this work the combination ofPMM and 3D-DIC has been employed to evaluate the modal behaviour of an aircraft cabin under random excitation e studywas focused on the passenger window area due to its significance to the structural integrity as a discontinuity of the peelOperational deflection shapes at different resonances were characterised by magnifying a single resonance in the spectrum andthen measuring with 3D-DIC ese measurements were validated with those obtained in forced normal mode tests Motion anddisplacement videos improved the understanding of the identified resonance deformation Actually a relevant behaviour wasnoticed in the windowrsquos frame a quite narrow area where using traditional sensors would not provide such a detailed3D information

1 Introduction

Structural integrity of aircraft is a vital issueese structuresundergo strict controls under severe conditions to providesafety and excellent dynamic performance Many efforts aretaken during design and validation stages in prediction withpowerful numerical simulations and expensive challengingexperimental measurements respectively Modal parameterinspection is a typical way to detect damage or abnormalbehaviour and also to characterise the generated noise insidethe cabin Experimentally the instrumentation of big

structures is usually sparse to prevent the cost increasingHence low spatial resolution provides a rough character-isation regarding mode shapes or even missed due to spatialaliasing Especially the resolution is not enough to un-derstand the behaviour of small critical areas In this sensefull-field optical techniques offer an ultrahigh spatial reso-lution compared to pointwise methods and thanks to highspeed cameras represent an interesting alternative formodalidentification Digital image correlation (DIC) [1] is one ofthe most popular methods due to its robustness well-developed methodology and user-friendly commercial

HindawiShock and VibrationVolume 2019 Article ID 4039862 9 pageshttpsdoiorg10115520194039862

software But the main advantage over most of the tech-niques is the possibility of performing 3D measurements bysetting up a calibrated stereovision system of two or morecameras Different studies have explored 3D-DIC for modalidentification [2ndash5] In a previous work the authors eval-uated the effect of differential pressure on the modal be-haviour of aircraft using 3D-DIC [6] ese studies provedthe capabilities of the technique to determine full-field modeshapes avoiding any spatial aliasing

An interesting methodology recently developed is phase-based motion magnification (PMM) It is a new Eulerianapproach to motion processing developed by Wadhwa et al[7] It is intended to reveal subtle motion in videos byobtaining a magnified version e video signals aredecomposed using complex steerable pyramid filters [8ndash10]is is an overcomplete transform based on complex si-nusoids modulated by a Gaussian window function at dif-ferent spatial scales orientations and positions e time-domain local phase at every spatial scale and orientation of asteerable pyramid where the motion is encoded is tem-porally band-pass filtered and amplified by a magnificationfactor When images are reconstructed back the result is themagnification of the harmonic motion in that frequencyband is method is advantageous over the previous linearamplitude-based Eulerian magnification method [11] sinceit is possible to achieve larger amplitudes without distortionMoreover it provides substantially better noise performancesince there is no increment of spatial noise amplitude ismethodology has been employed by Chen et al to describeoperational deflection shapes (ODS) under forced normalmode tests using the edge detection algorithm to highlightthe magnified shapes [12] Sarrafi et al [13] also employedthe edge detection on magnified images for damage de-tection by noticing changes in natural frequencies andODSs PMM has also been employed to develop an au-tonomous methodology for experimental modal analysiswith a little user supervision [14 15] Natural frequency anddamping are determined from phase signals under randomexcitation whereas mode shapes were magnified andhighlighted using the edge detection is methodology hasbeen employed for damage detection localisation usingfractal dimension [16]

at is very valuable however numerical information orquantifying the motion in terms of displacement is notavailable To achieve this the authors developed a meth-odology to measure the magnified images using 2D-DICunder sinusoidal excitation to determine displacements inthe sensor plane [17] is demonstrated to be beneficial toprovide a full-field measurement to magnified images Inaddition it improved the quality of the ODSs measurementdue to the larger motion considering the typical limitationsof DIC for low-amplitude vibrations Nevertheless thementioned phase-based methodologies and also the com-bination with 2D-DIC employ a system with a single cameraIt is only suitable to visualise in-plane motions and thusonly for 2D displacements Considering that the authorsevaluated the validity of the measurement on magnifiedimages from a stereoscopic system using 3D-DIC [18]Under analogous conditions the obtained ODSs were

compared with numerical simulations for a cantilever beamand it was demonstrated that no distortion occurred as aconsequence of separate magnification of the images fromeach camera is methodology was then performed on alarge aeronautical panel where 3D-DIC measurementsprovided out-of-plane information that could not beassessed by simple visual inspection despite magnification

In this study the capabilities of the combination of 3D-DIC and PMM are explored during random excitation toenable multiple mode evaluation in a single test In a pre-vious study on stereophotogrammetry and magnification inrandom tests Poozesh et al [19] employed 3D-DIC to revealone low response mode of a simple cantilever beam Herethe passenger window of a full-scale aircraft demonstrator isevaluated Multiple modes have been magnified to reveal theshape after a previous identification in a stabilisation dia-gram ODSs have been obtained in a full-field manner withnegligible presence of noise or other modes e results havebeen validated using the mode shapes from forced normalmode tests Videos of the naked motion and also includingthe displacement maps allowed an intuitive interpretation ofthe deformation at allowed behaviour characterisation ofthe window frame what is of significant interest for thestructure integrity

2 Experimental Tests

e methodology for ODSs evaluation using PMM and 3D-DIC was performed in Clean Sky GRA MT2 CockpitDemonstrator shown in Figure 1 e demonstrator con-sisted in a cabin fixed to a bulkhead and was fully instru-mented to characterise and evaluate the behaviour underdifferent force and environmental conditions Howeverthese sensors are not employed in this study is work isfocused on a specific part of the demonstrator the passengerwindow highlighted in Figure 1 which involves a discon-tinuity in the composite peel Hence this part plays a criticalrole in noise generation and keeping the airtightness

An electrodynamic shaker model LDS V450 wasemployed to generate noise random excitation in thespectrum 40ndash300Hz is study investigated the local be-haviour of the window and hence lower frequencies wereavoided as they involve global structure deformation or anyother rigid body behavioure vibration was transmitted tothe cabin lateral through a stinger as seen in Figure 1(b) Anaccelerometer was placed in the shakerrsquos armature to recordthe excitation

Two high-speed cameras model Photron FastCam SA4(1 megapixel full resolution) were employed to record theresponse on the inner surface of the window under exci-tation as observed in Figure 2 Both cameras were supportedon a rigid aluminium structure which was fixed to thebulkhead in order to isolate them from the vibration eevent was recorded at 2000 fps according to the Nyquistcriterion and the shutter speed was 01ms short enough toavoid blurring Two light sources were also employed togenerate clear high-contrast images e accelerometersignal was synchronised with both cameras through a DAQdevice model NI USB-6251

2 Shock and Vibration

Two additional forced normal mode tests were per-formed to validate the ODSs obtained from magnifiedimages using the same setup and recording parameter asdescribed in Figure 2

3 Image Processing

31 Phase-Based Motion Magnification In this study thephased-based motion magnification methodology de-veloped by Wadhwa et al [7] was employed is algorithmexploits the phase signals of the complex steerable pyramidsconsisting in Gabor wavelets at different positions andorientations to identify the local motions in spatial subbandsof an image e concept is analogous to Fourier shifttheorem for sine function transforms For instance a generic1D spatial-domain intensity profile I(x) that is movedaccording to a time-domain function δ(t) can be describedby a Fourier series decomposition as follows

I(x + δ(t)) 1113944infin

ωsminusinfinCse

iωs(x+δ(t)) (1)

where ωs is the frequency and Cs the amplitude of the si-nusoid Phase shifting of a hypothetical single sinusoidfringe image is shown in Figure 3 when motion occurs alongthe x direction As can be observed the phase is encoded as aphase shifting x + δ(t) with respect to the initial positionConsidering this analogy the phase signal of each pyramid isthen band-pass filtered and multiplied by a factor e

video reconstruction would show a magnified version ofthe motion present in that frequency band previouslyimperceptible Hence the algorithm inputs are the imagesequence the desired frequency band and the magnificationfactor

For subsequent measurements using 3D-DIC the se-quence from each camera must be magnified using the sameparameters [18] Unlike sinusoidal excitation many signifi-cant modes are now present in the images simultaneouslyerefore magnification factors were chosen so that theresponses of the nonmagnified modes were negligible incomparison with the magnified one

Considering the random nature of the recorded vibra-tion the larger is the sequence the better are the resultsHowever phase-based motion magnification is highlymemory demanding and this limits the sequence length Inthis occasion a computer with 32GB RAM was employedallowing 1000 image sequences of 1 megapixel

32 3D Digital Image Correlation Digital image correlationis an optical technique based on intensity field correlation toperform displacement and strain measurements in images[1] For 3D-DIC images from a stereoscopic system com-posed by using two cameras are employede calibration ofthe stereovision consisted in determining both extrinsicparameters which define the relative position of the camerasand establish the relation of the coordinate systems andintrinsic parameters related to the position of the sensor thelenses and the light An area of interest is identified in the

(a)

iii

iii

(b)

Figure 1 (a) Full-scale aircraft demonstrator and (b) the lateral excitation configuration using (i) a shaker with (ii) a stinger to measure on(iii) the passenger window

i

ii

iii

Figure 2 Stereoscopic setup composed by (i) two high-speedcameras and (ii) light sources recording on (iii) the passengerwindow from inside

x

Iδ(t)

x

y

(a) (b)

Figure 3 (a) Sinusoidal fringe image and (b) the effect of x-motionin a 1D intensity profile

Shock and Vibration 3

reference image from one camera and divided into squaredsubregions of pixels called facets or subsets To make everyfacet unique the surface is typically treated to generate a grey-scale speckle pattern as shown in Figure 4(a) e size of thefacets is an odd number of pixels (2M+1) where M is aninteger number so that the facet central pixel is P at theposition (x0 y0) [20] as shown in Figure 4(b) e centralpixel of every facet is initially identified in the second cameraas themost similar intensity areaWith this first identificationit is possible to determine the spatial position of the centralpixel of every facet ie every measurement point and hence adigitalisation of the specimen surface Performing the cor-relation in the subsequent deformed states the displacementof every facet is determined as shown in Figure 4(c)

In the present case the window and the surrounding peelwas coated with white paint and then sprayed with blackpaint to create the random speckle pattern e correlationanalysis was performed with Vic 3D software from Corre-lated Solutions Inc using 19-pixel facet and 5-pixel overlapstep at generated a full-field measurement equivalent to23000 3D sensors in this surface Facet size can be compared

with the image size in Figure 5(a) and the resulting surfacedigitalisation is shown in Figure 5(b)

4 Results and Discussion

As a broadband event the expected response of the windowto such excitation is a combination of the modes in thisspectrum erefore an instantaneous out-of-plane dis-placement W in Figure 6 shows that no clear mode shapeextraction could be done by visualising the time-domainresponse Actually a significant amount of noise was alsoexpected and found due to the low sensitivity of non-interferometric techniques in comparison with traditionalpointwise sensors such as accelerometers or even laservibrometry [21 22] erefore PMM was employed in thisstudy to highlight individual ODSs making both othermodes and noise negligible [7 14]

Before performing the magnification some modes hadto be identified e full-field transfer functions were ob-tained from the 3D-DIC and the accelerometer measure-ments under the same noise excitation e excitation

(a)

x

yP(x0 y0)

(b)

x

yP(x0 y0)

Pprime(x y)

(c)

Figure 4 (a) Example of a grey-scale speckle pattern on the surface of interest (b) Facet grid generation in the reference image highlightingthat whose central pixel P (c) Position of the pixel P after deformation Pprime to determine the displacement vector

(a)

ndash250ndash1375 ndash25 875 200

250

1125

ndash25

ndash1625

ndash300

ndash20ndash1001020 Z (mm)

Y (m

m)

X (mm)

(b)

Figure 5 (a) Region of interest and facet size and (b) 3D digitalisation using 3D-DIC


displacement amplitudes were obtained by accelerometersignal double integration As RAM was not a limitationfor 3D-DIC processing but only for PMM 5400 frame se-quences were considered for this estimation that eventuallyyielded 1Hz frequency resolution e stabilisation diagramcorresponding to the least squares complex exponentialmethod was chosen to highlight modes is diagram isrepresented in Figure 7 and shows the averaged transferfunctionMultiple peaks were observedmany of which can beattributed to global structural modes that do not show the

local behaviour of the region ese modes were not of in-terest in this studye selected local modes for magnificationwere found at the peaks 124Hz 155Hz 215Hz and 268HzIn fact these had the highest stabilisation trend and the se-lection is also supported by the previous study results [6]en magnification was performed in a narrow band for eachone e magnification factors were 50x for 124Hz and 100xfor the rest ose values were determined according to thelevel of response to yield clear displacement maps withoutproducing blurring derived from excessive magnification

00875

0075

00625

005

00375

0025

00125

0

ndash00125

ndash0025

ndash00375

ndash005

ndash00625

ndash0075

ndash00875

ndash01

W (mm)01

Figure 6 Instantaneous out-of-plane displacements W measured by 3D-DIC under random excitation

50 100 150 200 250 300Frequency (Hz)

0

5

10

15

20

Mod

el o

rder

102

103

104

105

Tran

sfer f

unct

ion

(mm

mm

)

Stable in frequencyStable in frequency and damping

Not stable in frequencyAveraged response function

Figure 7 Stabilisation diagram obtained from the least squares complex exponential method applied to the full-field transfer functions


As a result videos were obtained showing a clear motionfor the naked eye e videos corresponding to the right-hand side camera can be found as Videos 1ndash4 in supple-mentary material for each respective mode As can be seenthese videos describe the local behaviour of the window andthe surrounding aircraftrsquos peel Comparing them differentdeformations are noticed from one mode to another and thecomplexity is increasing for higher frequency modes as ex-pected is makes the interpretation of the motion difficultfor the naked eye Under this limitation 3D-DIC measure-ments were performed in the magnified sequences Takinginto account that bending is mainly described by out-of-planedisplacements called asW focus was placed on this directionFour respective videos were obtained that included the dis-placement maps evolution (Videos 5ndash8) e instantaneousW maps at the signal crest are shown in Figure 8 emeasurement allowed the quantification of the virtual motionof these images and improved the interpretation of the de-formation It is also worth mentioning the capabilities of themethodology to extract ODSs from a random event where justa combination of noise and modes was obtained ey all

correspond to bending modes of increasing order mainlyproduced in the window itself ey can be checked againstthe ODSs obtained from forced normal mode tests A quickidentification of two resonances at 124Hz and 215Hz wasperformed with an impact hammer test Hence the ODSs forthese frequencies under sinusoidal excitation are shown inFigure 9 Magnified ODSs from random tests show the samebehaviour what proves they are actual ODSs and the influenceof the remaining modes is negligible

Overall both sorts of video provide a meaningful insightof the behaviour of the window For instance the travellingwave effect produced by different phase lag along the surfaceis evidenced For the first mode the phase lag is morehomogeneous but this effect is especially relevant for thethree higher frequency modes However there is an addi-tional behaviour that might be of major importance for theintegrity of the joint with the peel and hence for the air-tightness In all the analysed resonances the frame expe-rienced deformation in certain zones with the highestamplitudes of the whole surface By observing the dis-placement maps in Figure 8 (more notorious in Videos 5ndash8)

W (mm)0504375037503125025018750125006250ndash00625ndash0125ndash01875ndash025ndash03125ndash0375ndash04375ndash05

(a)


(b)

0

W (mm)0620542504650387503102325015500775

ndash00775ndash0155ndash02325ndash031ndash03875ndash0465ndash05425ndash062

(c)

0

W (mm)03202802402016012008004


(d)

Figure 8 ODSs of the window in the out-of-plane direction W obtained with 3D-DIC after magnifying the resonances (a) 124Hz(b) 155Hz (c) 215Hz and (d) 268Hz


this deformation can be localised e motion may beintuited in the motion Videos 1ndash4 but zoom-in videos areprovided for detailed inspection In the first mode largeamplitudes are detected on the left part of the frame and alsoclose to the right lower corner ey can be observed re-spectively in zoom-in Videos 9ndash10 Namely the maximumdeformation of the left part is compared with the nonloadedstate in Figure 10 Blurring appeared in the frame as aconsequence of the large displacements after magnificationSince the displacement in the window was smaller themagnification factor was chosen to provide good displace-ment fields for the window despite the little blurring of theframe As observed maximum displacements for the secondresonance mainly occur in the upper and lower parts shownthe former in Video 11 It is worthy to note that the thirdmode shows higher order deformation especially close tothe right upper corner as seen in Video 12 A particular case

is the last mode as no deformation was detected in theframe in the W direction However vertical displacementsV shown in Figure 11 revealed a significant deformation inthe upper region of the frame is vertical motion can beconfirmed watching the zoom Video 13

5 Conclusions

is study shows the benefits of the combination of phase-based motion magnification and 3D-DIC especially forevaluating the complex structures where no conclusive in-formation can be obtained from a single camera MultipleODSs of the passenger window of a full-scale aircraftdemonstrator were deduced visualized and quantitativelymeasured from a single random excitation test e mag-nification of an individual resonance from a broadbandresponse makes it predominant against the other and its only


(a)


(b)

Figure 9 ODSs of the window in the out-of-plane direction W obtained with 3D-DIC in forced normal mode tests for the resonances(a) 124Hz and (b) 215Hz

(a) (b)

Figure 10 Windowrsquos frame upper right corner in the nonloaded state (a) and at maximum deformation (b) for the magnified resonance124Hz


presence can be assumed Larger amplitude motion achievedby magnification allowed clearer 3D-DIC displacementmaps to be obtained which was little influenced by noise Onthe contrary 3D measurement provided the necessary in-formation to understand themotion of themagnified videosese measurements agreed with the ODSs obtained inforced normal mode tests As a conclusion visual andquantitative information proved to be complementary andprovided a deep insight into an intuitive interpretation andunderstanding of the deformation

Furthermore this is a powerful tool for such big structuressince it is able to give high-density information in critical partswhich could not be achieved by using spaced sensors beingpossible to lose localised effects is paper has demonstratedthe potential when detecting significant deformation of thewindowrsquos frame is provides relevant information andfeedback to improve the union of the window and the peelthat ensures the integrity and airtightness

Data Availability

e data used to support the findings of this study areavailable from the corresponding author upon request

Disclosure

Activities reported in this paper were developed in the frameof the European Community Seventh Framework Programwhere Airbus Defence and Space SAU was a partner of theClean Sky Green Regional Aircraft Integrated TechnologyDemonstrator e University of Jaen (Spain) participatedunder contract with Airbus Defence and Space SAU in thefield of testing technologies

Conflicts of Interest

e authors declare that there are no conflicts of interestregarding the publication of this paper

Supplementary Materials

A description of the videos supplied as supplementarymaterials is here provided ey can be found in the onlineversion of the present paperSupplementary 1Video 1 nakedmotion of the windowODSat 124Hz using 50x magnification factor from the right-hand side camera viewpoint

Supplementary 2Video 2 nakedmotion of the windowODSat 155Hz using 100x magnification factor from the right-hand side camera viewpoint



Supplementary 5 Video 5 ODSs of the window in the out-of-plane direction W obtained with 3D-DIC after mag-nifying the resonance 124Hz with 50x factor




Supplementary 9 Video 9 naked motion of the windowODS at 124Hz using 50x magnification factor zooming atthe left part of the frame from the right-hand side cameraviewpoint

Supplementary 10 Video 10 naked motion of the windowODS at 124Hz using 50x magnification factor zooming atthe right lower corner of the frame from the right-hand sidecamera viewpoint

Supplementary 11 Video 11 naked motion of the windowODS at 155Hz using 100x magnification factor zooming atthe upper part of the frame from the left-hand side cameraviewpoint



References

[1] H Schreier J-J Orteu and M A Sutton Image Correlationfor Shape Motion and Deformation Measurements SpringerBoston MA USA 2009

0350302502015010050ndash005ndash01ndash015ndash02ndash025ndash03ndash035ndash04

V (mm)04

Figure 11 ODSs of the window in the vertical direction V obtainedwith 3D-DIC in forced normal mode tests for the resonance 268Hz


[2] R Hunady and M Hagara ldquoA new procedure of modalparameter estimation for high-speed digital image correla-tionrdquo Mech Syst Signal Process vol 93 pp 66ndash79 2017

[3] D A Ehrhardt M S Allen S Yang and T J Beberniss ldquoFull-field linear and nonlinear measurements using continuous-scan laser Doppler vibrometry and high speed three-dimensional digital image correlationrdquo Mechanical Systemsand Signal Processing vol 86 pp 82ndash97 2017

[4] P L Reu D P Rohe and L D Jacobs ldquoComparison of DICand LDV for practical vibration and modal measurementsrdquoMechanical Systems and Signal Processing vol 86 pp 2ndash162017

[5] T J Beberniss and D A Ehrhardt ldquoHigh-speed 3D digitalimage correlation vibration measurement recent advance-ments and noted limitationsrdquo Mechanical Systems and SignalProcessing vol 86 pp 35ndash48 2017

[6] A Molina-Viedma E Lopez-Alba L Felipe-Sese F DıazJ Rodrıguez-Ahlquist and M Iglesias-Vallejo ldquoModal pa-rameters evaluation in a full-scale Aircraft demonstratorunder different environmental conditions using HS 3D-DICrdquoMaterials vol 11 no 2 p 230 2018

[7] N Wadhwa M Rubinstein F Durand and W T FreemanldquoPhase-based video motion processingrdquo ACM Transactionson Graphics vol 32 no 4 p 1 2013

[8] E P Simoncelli W T Freeman E H Adelson andD J Heeger ldquoShiftable multiscale transformsrdquo IEEE Trans-actions on Information 8eory vol 38 no 2 pp 587ndash6071992

[9] E P Simoncelli andW T Freeman ldquoe steerable pyramid aflexible architecture for multi-scale derivative computationrdquoin Proceedings of the International Conference on ImageProcessing vol 3 pp 444ndash447 Washington DC USAOctober 1995

[10] J Portilla and E P Simoncelli ldquoA parametric texture modelbased on joint statistics of complex wavelet coefficientsrdquoInternational Journal of Computer Vision vol 40 no 1pp 49ndash70 2000

[11] H-Y Wu M Rubinstein E Shih J Guttag F Durand andW Freeman ldquoEulerian video magnification for revealingsubtle changes in the worldrdquo ACM Transactions on Graphicsvol 31 no 4 pp 1ndash8 2012

[12] J G Chen N Wadhwa Y-J Cha F DurandW T Freeman and O Buyukozturk ldquoModal identificationof simple structures with high-speed video using motionmagnificationrdquo Journal of Sound and Vibration vol 345pp 58ndash71 2015

[13] A Sarrafi Z Mao C Niezrecki and P Poozesh ldquoVibration-based damage detection in wind turbine blades using Phase-based Motion Estimation and motion magnificationrdquo Journalof Sound and Vibration vol 421 pp 300ndash318 2018

[14] Y Yang C Dorn T Mancini et al ldquoBlind identification offull-field vibration modes from video measurements withphase-based video motion magnificationrdquo Mechanical Sys-tems and Signal Processing vol 85 pp 567ndash590 2017

[15] Y Yang C Dorn T Mancini et al ldquoBlind identification of full-field vibration modes of output-only structures fromuniformly-sampled possibly temporally-aliased (sub-Nyquist)videomeasurementsrdquo Journal of Sound and Vibration vol 390pp 232ndash256 2017

[16] Y Yang C Dorn T Mancini et al ldquoReference-free detectionof minute non-visible damage using full-field high-resolution mode shapes output-only identified from digitalvideos of structuresrdquo Structural Health Monitoring vol 17no 3 pp 514ndash531 2017

[17] A J Molina-Viedma L Felipe-Sese E Lopez-Alba andF Dıaz ldquoHigh frequency mode shapes characterisation usingDigital Image Correlation and phase-based motion magni-ficationrdquo Mechanical Systems and Signal Processing vol 102pp 245ndash261 2018

[18] A J Molina-Viedma L Felipe-Sese E Lopez-Alba andF A Dıaz ldquo3D mode shapes characterisation using phase-based motion magnification in large structures using ste-reoscopic DICrdquo Mechanical Systems and Signal Processingvol 108 pp 140ndash155 2018

[19] P Poozesh A Sarrafi Z Mao P Avitabile and C NiezreckildquoFeasibility of extracting operating shapes using phase-basedmotion magnification technique and stereo-photogramme-tryrdquo Journal of Sound and Vibration vol 407 pp 350ndash3662017

[20] B Pan K Qian H Xie and A Asundi ldquoTwo-dimensionaldigital image correlation for in-plane displacement and strainmeasurement a reviewrdquo Measurement Science and Technol-ogy vol 20 no 6 p 062001 2009

[21] M N Helfrick C Niezrecki P Avitabile and T Schmidt ldquo3Ddigital image correlation methods for full-field vibrationmeasurementrdquo Mechanical Systems and Signal Processingvol 25 no 3 pp 917ndash927 2011

[22] C Warren C Niezrecki P Avitabile and P Pingle ldquoCom-parison of FRF measurements and mode shapes determinedusing optically image based laser and accelerometer mea-surementsrdquoMechanical Systems and Signal Processing vol 25no 6 pp 2191ndash2202 2011


International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018


Active and Passive Electronic Components

VLSI Design



Shock and Vibration


Civil EngineeringAdvances in

Acoustics and VibrationAdvances in



Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of



Journal ofEngineeringVolume 2018

SensorsJournal of



RotatingMachinery


Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018


Chemical EngineeringInternational Journal of Antennas and

Propagation




Navigation and Observation


Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

software But the main advantage over most of the tech-niques is the possibility of performing 3D measurements bysetting up a calibrated stereovision system of two or morecameras Different studies have explored 3D-DIC for modalidentification [2ndash5] In a previous work the authors eval-uated the effect of differential pressure on the modal be-haviour of aircraft using 3D-DIC [6] ese studies provedthe capabilities of the technique to determine full-field modeshapes avoiding any spatial aliasing

An interesting methodology recently developed is phase-based motion magnification (PMM) It is a new Eulerianapproach to motion processing developed by Wadhwa et al[7] It is intended to reveal subtle motion in videos byobtaining a magnified version e video signals aredecomposed using complex steerable pyramid filters [8ndash10]is is an overcomplete transform based on complex si-nusoids modulated by a Gaussian window function at dif-ferent spatial scales orientations and positions e time-domain local phase at every spatial scale and orientation of asteerable pyramid where the motion is encoded is tem-porally band-pass filtered and amplified by a magnificationfactor When images are reconstructed back the result is themagnification of the harmonic motion in that frequencyband is method is advantageous over the previous linearamplitude-based Eulerian magnification method [11] sinceit is possible to achieve larger amplitudes without distortionMoreover it provides substantially better noise performancesince there is no increment of spatial noise amplitude ismethodology has been employed by Chen et al to describeoperational deflection shapes (ODS) under forced normalmode tests using the edge detection algorithm to highlightthe magnified shapes [12] Sarrafi et al [13] also employedthe edge detection on magnified images for damage de-tection by noticing changes in natural frequencies andODSs PMM has also been employed to develop an au-tonomous methodology for experimental modal analysiswith a little user supervision [14 15] Natural frequency anddamping are determined from phase signals under randomexcitation whereas mode shapes were magnified andhighlighted using the edge detection is methodology hasbeen employed for damage detection localisation usingfractal dimension [16]

at is very valuable however numerical information orquantifying the motion in terms of displacement is notavailable To achieve this the authors developed a meth-odology to measure the magnified images using 2D-DICunder sinusoidal excitation to determine displacements inthe sensor plane [17] is demonstrated to be beneficial toprovide a full-field measurement to magnified images Inaddition it improved the quality of the ODSs measurementdue to the larger motion considering the typical limitationsof DIC for low-amplitude vibrations Nevertheless thementioned phase-based methodologies and also the com-bination with 2D-DIC employ a system with a single cameraIt is only suitable to visualise in-plane motions and thusonly for 2D displacements Considering that the authorsevaluated the validity of the measurement on magnifiedimages from a stereoscopic system using 3D-DIC [18]Under analogous conditions the obtained ODSs were

compared with numerical simulations for a cantilever beamand it was demonstrated that no distortion occurred as aconsequence of separate magnification of the images fromeach camera is methodology was then performed on alarge aeronautical panel where 3D-DIC measurementsprovided out-of-plane information that could not beassessed by simple visual inspection despite magnification

In this study the capabilities of the combination of 3D-DIC and PMM are explored during random excitation toenable multiple mode evaluation in a single test In a pre-vious study on stereophotogrammetry and magnification inrandom tests Poozesh et al [19] employed 3D-DIC to revealone low response mode of a simple cantilever beam Herethe passenger window of a full-scale aircraft demonstrator isevaluated Multiple modes have been magnified to reveal theshape after a previous identification in a stabilisation dia-gram ODSs have been obtained in a full-field manner withnegligible presence of noise or other modes e results havebeen validated using the mode shapes from forced normalmode tests Videos of the naked motion and also includingthe displacement maps allowed an intuitive interpretation ofthe deformation at allowed behaviour characterisation ofthe window frame what is of significant interest for thestructure integrity

2 Experimental Tests

e methodology for ODSs evaluation using PMM and 3D-DIC was performed in Clean Sky GRA MT2 CockpitDemonstrator shown in Figure 1 e demonstrator con-sisted in a cabin fixed to a bulkhead and was fully instru-mented to characterise and evaluate the behaviour underdifferent force and environmental conditions Howeverthese sensors are not employed in this study is work isfocused on a specific part of the demonstrator the passengerwindow highlighted in Figure 1 which involves a discon-tinuity in the composite peel Hence this part plays a criticalrole in noise generation and keeping the airtightness

An electrodynamic shaker model LDS V450 wasemployed to generate noise random excitation in thespectrum 40ndash300Hz is study investigated the local be-haviour of the window and hence lower frequencies wereavoided as they involve global structure deformation or anyother rigid body behavioure vibration was transmitted tothe cabin lateral through a stinger as seen in Figure 1(b) Anaccelerometer was placed in the shakerrsquos armature to recordthe excitation

Two high-speed cameras model Photron FastCam SA4(1 megapixel full resolution) were employed to record theresponse on the inner surface of the window under exci-tation as observed in Figure 2 Both cameras were supportedon a rigid aluminium structure which was fixed to thebulkhead in order to isolate them from the vibration eevent was recorded at 2000 fps according to the Nyquistcriterion and the shutter speed was 01ms short enough toavoid blurring Two light sources were also employed togenerate clear high-contrast images e accelerometersignal was synchronised with both cameras through a DAQdevice model NI USB-6251



3 Image Processing


I(x + δ(t)) 1113944infin

ωsminusinfinCse

iωs(x+δ(t)) (1)






(a)

iii

iii

(b)


i

ii

iii


x

Iδ(t)

x

y

(a) (b)









(a)

x

yP(x0 y0)

(b)

x

yP(x0 y0)

Pprime(x y)

(c)


(a)


250

1125

ndash25

ndash1625

ndash300


Y (m

m)

X (mm)

(b)





00875

0075

00625

005

00375

0025

00125

0

ndash00125

ndash0025

ndash00375

ndash005

ndash00625

ndash0075

ndash00875

ndash01

W (mm)01


50 100 150 200 250 300Frequency (Hz)

0

5

10

15

20

Mod

el o

rder

102

103

104

105

Tran

sfer f

unct

ion

(mm

mm

)









(a)


(b)

0

W (mm)0620542504650387503102325015500775


(c)

0

W (mm)03202802402016012008004


(d)





5 Conclusions



(a)


(b)


(a) (b)





Data Availability


Disclosure


















References



V (mm)04



























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia



3 Image Processing


I(x + δ(t)) 1113944infin

ωsminusinfinCse

iωs(x+δ(t)) (1)






(a)

iii

iii

(b)


i

ii

iii


x

Iδ(t)

x

y

(a) (b)









(a)

x

yP(x0 y0)

(b)

x

yP(x0 y0)

Pprime(x y)

(c)


(a)


250

1125

ndash25

ndash1625

ndash300


Y (m

m)

X (mm)

(b)





00875

0075

00625

005

00375

0025

00125

0

ndash00125

ndash0025

ndash00375

ndash005

ndash00625

ndash0075

ndash00875

ndash01

W (mm)01


50 100 150 200 250 300Frequency (Hz)

0

5

10

15

20

Mod

el o

rder

102

103

104

105

Tran

sfer f

unct

ion

(mm

mm

)









(a)


(b)

0

W (mm)0620542504650387503102325015500775


(c)

0

W (mm)03202802402016012008004


(d)





5 Conclusions



(a)


(b)


(a) (b)





Data Availability


Disclosure


















References



V (mm)04



























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia








(a)

x

yP(x0 y0)

(b)

x

yP(x0 y0)

Pprime(x y)

(c)


(a)


250

1125

ndash25

ndash1625

ndash300


Y (m

m)

X (mm)

(b)





00875

0075

00625

005

00375

0025

00125

0

ndash00125

ndash0025

ndash00375

ndash005

ndash00625

ndash0075

ndash00875

ndash01

W (mm)01


50 100 150 200 250 300Frequency (Hz)

0

5

10

15

20

Mod

el o

rder

102

103

104

105

Tran

sfer f

unct

ion

(mm

mm

)









(a)


(b)

0

W (mm)0620542504650387503102325015500775


(c)

0

W (mm)03202802402016012008004


(d)





5 Conclusions



(a)


(b)


(a) (b)





Data Availability


Disclosure


















References



V (mm)04



























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia




00875

0075

00625

005

00375

0025

00125

0

ndash00125

ndash0025

ndash00375

ndash005

ndash00625

ndash0075

ndash00875

ndash01

W (mm)01


50 100 150 200 250 300Frequency (Hz)

0

5

10

15

20

Mod

el o

rder

102

103

104

105

Tran

sfer f

unct

ion

(mm

mm

)









(a)


(b)

0

W (mm)0620542504650387503102325015500775


(c)

0

W (mm)03202802402016012008004


(d)





5 Conclusions



(a)


(b)


(a) (b)





Data Availability


Disclosure


















References



V (mm)04



























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia






(a)


(b)

0

W (mm)0620542504650387503102325015500775


(c)

0

W (mm)03202802402016012008004


(d)





5 Conclusions



(a)


(b)


(a) (b)





Data Availability


Disclosure


















References



V (mm)04



























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia




5 Conclusions



(a)


(b)


(a) (b)





Data Availability


Disclosure


















References



V (mm)04



























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia




Data Availability


Disclosure


















References



V (mm)04



























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia


























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia


operational deflection shape extraction from broadband ...researcharticle operational deflection...

Documents