materials Article
Fatigue Damage Monitoring of a Composite Step Lap Joint Using Distributed Optical Fibre Sensors Leslie Wong 1, *, Nabil Chowdhury 1 , John Wang 2 , Wing Kong Chiu 1 and Jayantha Kodikara 3 1 2 3
*
Department of Mechanical & Aerospace Engineering, Monash University, Wellington Road, Clayton, VIC 3168, Australia;
[email protected] (N.C.);
[email protected] (W.K.C.) Aerospace Division, Defence Science and Technology Group, 506 Lorimer Street, Fishermans Bend, VIC 3207, Australia;
[email protected] Department of Civil Engineering, Monash University, Clayton, VIC 3168, Australia;
[email protected] Correspondence:
[email protected]; Tel.: +61-3-9905-3512
Academic Editor: Ferri M.H. Aliabadi Received: 24 March 2016; Accepted: 9 May 2016; Published: 14 May 2016
Abstract: Over the past few decades, there has been a considerable interest in the use of distributed optical fibre sensors (DOFS) for structural health monitoring of composite structures. In aerospace-related work, health monitoring of the adhesive joints of composites has become more significant, as they can suffer from cracking and delamination, which can have a significant impact on the integrity of the joint. In this paper, a swept-wavelength interferometry (SWI) based DOFS technique is used to monitor the fatigue in a flush step lap joint composite structure. The presented results will show the potential application of distributed optical fibre sensor for damage detection, as well as monitoring the fatigue crack growth along the bondline of a step lap joint composite structure. The results confirmed that a distributed optical fibre sensor is able to enhance the detection of localised damage in a structure. Keywords: fatigue test; step lap joint; structural health monitoring; distributed optical fibre sensor; adhesion; composite structure
1. Introduction Due to the requirement of high specific stiffness, strength, durability and corrosion resistance, composite materials have seen increasing use in the engineering industry, and, in particular, the aerospace sector. Adhesive bonding is one of the most efficient methods of joining lightweight composite structures. Compared to mechanically fastened joints, it reduces high stress concentration in the substrates with minimal weight penalty. Results gathered by several researchers have demonstrated that step lap joints can be an effective and efficient joint repair method to restore a structure’s strength and stiffness [1,2]. Health monitoring of the joints is becoming more significant as they often suffer from cracking and slippage, which renders them out of service prematurely. Therefore, the development of proper non-destructive evaluation (NDE) techniques and strategies have become more crucial for damage interrogation and the continuous monitoring of structural integrity to assess the state of joints. This is particularly important for fatigue loading cases, where cracks initiate in the adhesive layer of bonded and hybrid joint structures. Effective and reliable control of the bondline can provide advanced warning and enhance replacement and repair of joints before catastrophic failure, as well as extend the service life of a structure. Common conventional NDE techniques, such as visual testing (VT), Ultrasonic C-Scan, X-ray, eddy current, and thermography, are limited as they require structural components of complex geometry to be taken out of service for a substantial period of time for post-damage inspection and Materials 2016, 9, 374; doi:10.3390/ma9050374
www.mdpi.com/journal/materials
Materials 2016, 9, 374
2 of 11
assessment. Ultrasonic guided waves, using built-in thin Lead-Zirconate-Titanate piezoelectric films (PZT), are becoming more popular for in situ structural health monitoring due to their small size, high piezoelectric coupling coefficient, and ease for surface-mounting and embedding in composite structures [3,4]. The use of guided waves has been proven to be effective in locating and identifying damage [5,6]. Moreover, the usage of surface-bonded resistive foil strain gauges offers another attractive strain measurement method for continuous and in situ monitoring of structural conditions. Nevertheless, both PZT and strain gauges are susceptible to electromagnetic and electrical interference, in addition to physical damage (e.g., impact). Distributed fibre optic sensors (DOFS) have a proven strain measuring capability [7–14] with the key characteristics of being small in size, lightweight and immune to electromagnetic interference. DOFS also have the ability to monitor thousands of points along a single strand of optical fibre. They can be easily integrated into composite structures to produce a so-called ‘smart’ composite structure, which is capable of providing real-time information for inaccessible areas in a structure (which other sensing methods cannot easily probe). The basic working principle of fibre optic strain sensors is to monitor or measure the intensity, wavelength, phase and state of polarisation (SOP). It is important to monitor these properties as they are directly related to the behaviour in the host structure. With proper fibre adhesion, the distributed optical fibre sensor can measure the same strain gradient as the host structure and enhance the detection of damage initiation and progression. Moreover, they must also be located near the region where damage occurs. Based on this method, a significant level of local strain concentrations or anomalies can be used to indirectly identify damage in the host composite structure. In certain applications, it is more convenient to measure a variation of stiffness or applied load in order to assess damage. Hence, by making good use of the DOFS for structural health monitoring, one could reduce structural maintenance costs and decrease the level of structural overdesign whilst maintaining and possibly improving safety. The aim of this paper is to experimentally demonstrate the ability of a distributed optical fibre strain sensor to monitor the fatigue crack growth along the adhesive in a bonded composite step lap joint. In this paper, a swept-wavelength interrogator (SWI) based distributed optical fibre is exploited to perform a continuous and real-time distributed strain measurement on an adhesively bonded step lap joint configuration subjected to fatigue tests using a 100 kN INSTRON servo-hydraulic machine (INSTRON, Melbourne, Australia). 2. Fatigue Test Monitoring 2.1. Distributed Optical Fibre Sensor Technique A Luna Technologies’ Optical Distributed Sensor Interrogator (ODiSI-B series, LunaInc, Blacksburg, VA, USA), as shown in Figure 1, is used in the experiments presented in this paper. The sensor is based on the Rayleigh optical frequency domain reflectometry (OFDR) principles, and was coupled with swept-wave interferometry techniques (SWI) [15,16]. A brief schematic of a SWI-based system is shown in Figure 2. Light from a tuneable Continuous Wave (CW) laser source is split and propagated through the testing fibre and static reference arms of a fibre optic Mach-Zehnder interferometer [10,17] and recombined at an optical detector. The interference patterns are generated as the laser frequency is tuned. The patterns are detected and related to the optical amplitude and phase response of the fibre under test. These patterns (in spectral domain) of the fibre under test can be processed using the inverse Fourier transform to time domain. A map of the reflections as a function of distance of the fibre under test can then be constructed. Thus, OFDR can be used for both spectral and time domain reflectometry.
∆λ ∆λ λ λ
∆υ ∆υ υ υ
ε
ε
(1) (1)
where λ, υ, KT and Kε are the mean optical wavelength, frequency, temperature and strain calibration where λ, υ, KT and Kε are the mean optical wavelength, frequency, temperature and strain calibration constants, respectively. This technique results in distributed strain measurements with a 1 με constants, respectively. This technique results in distributed strain measurements with a 1 με3 of 11 Materials 2016,or 9, temperature 374 resolution measurements with a 0.1 °C resolution [17]. resolution or temperature measurements with a 0.1 °C resolution [17].
Figure ODiSI-Bseries series from from Luna Inc. Figure 1. 1.ODiSI-B LunaInnovations Innovations Inc. Figure 1. ODiSI-B series from Luna Innovations Inc.
Figure 2. 2.Schematic drawingof of swept-wavelength interrogator (SWI)-based system Figure Schematic drawing swept-wavelength interrogator (SWI)-based system with Mach- with Zehnder interferometer. Mach-Zehnder interferometer.
Figure Schematic drawing ofthe swept-wavelength interrogator systemresolution with MachThe2.ODiSI-B sensor used in experiments presented in this (SWI)-based paper has a spatial of
The Rayleigh scatter amplitude reflected from a single strand of fibre as a function of distance is a interferometer. 5Zehnder mm, a maximum measuring length of 10 m and a temporal resolution of 0.01 s. Acrylate coated random pattern. Although random, it is static and unique for every fibre. When fibreasisthe stretched single mode fibre (SMF 28e from AFW Technologies, Hallam, VIC, Australia) wasaused fibre or strained, the spatial frequency of this pattern is also stretched. This elongation leads to a change The ODiSI-B sensor used in the experiments presented in this paper has a spatial resolution of sensor of choice in the study. The diameter of the fibre sensors (core with cladding) was in the frequency spectrum reflected from this10section fibre. These changes can s. be measured and 5 mm, a maximum measuring length of m tight-buffered andofa the temporal resolution of sensor 0.01 Acrylate coated approximately 250 μm. The lack of protective jacket on the fibre meant that it had totobedetermine handled carefully during installation to avoid Bare fibres preferred calibrated the strain and temperature inbreakage. the fibre [18]. This were shift inused the spectrum in single mode fibre (SMF 28elocal from AFW Technologies, Hallam, VIC, Australia) was asover the fibre tight-buffered fibres due to slippage can introduce error to the results. response strain, ε,inorthe temperature, T, isdiameter analogous a change the spectral sensor ofto choice study. The oftothe fibre in sensors (core shift, with ∆υ: cladding) was approximately 250 μm. The lack of protective tight-buffered jacket on the fibre sensor meant that it 2.2. Specimen Design ∆λ ∆υ to avoid breakage. Bare fibres were preferred over had to be handled carefully during installation “´ “ K T T ` Kε ε (1) λ υ are The adherends in the step lap joint manufactured from 24 plies thick HexPly tight-buffered fibres dueused to slippage can introduce error to the results. carbon fibre (HEXCEL., Stamford, frequency, CT, USA) [19], Table 1; with ply stacking whereM18/1/G939 λ, υ, KT and Kε are theprepreg mean optical wavelength, temperature andastrain calibration sequence of [(0/90)/(45/-45)/(45/-45)/(0/90)] 6. Figure 3 shows a schematic of the composite structure 2.2. Specimen Design constants, respectively. This technique results in distributed strain measurements with a 1 µε resolution with a width of 25.4 mm with each ply˝having a nominal uncured thickness of 0.41 mm and every or temperature measurements 0.1 C resolution The having adherends used in with theofastep are [17]. manufactured from 24 plies thick HexPly step a uniform thickness 0.908lap mm.joint The specimen contains a total bondline area of 90 mm The ODiSI-B sensor used in the experiments presented in this paper has a1;spatial of M18/1/G939 carbon prepreg (HEXCEL., CT,by USA) [19], Table with aresolution plyVIC, stacking × 25.4 mm usingfibre FM300-2K film adhesive Stamford, manufactured Cytec Engineering (Carlton, 5sequence mm, a maximum measuring lengthofofthe 106structure and ato resolution ofwere 0.01bonded s. Acrylate coated Australia) [20], Table 2. The location which the fibre sensorsof was structure first of [(0/90)/(45/-45)/(45/-45)/(0/90)] .m Figure 3temporal shows a schematic the composite single mode fibre (SMF 28e from AFW Technologies, Hallam, VIC, Australia) was used as the fibre marked out. The outer surface layer on these surfaces was cleaned with acetone. The acrylate coated with a width of 25.4 mm with each ply having a nominal uncured thickness of 0.41 mm and every Single Mode (SMF) fibre optic sensors were then heldwith in place with strain tape sensor of choice in theFibre study. The 28e diameter of the fibre sensors (core wasgauge approximately step bare having a uniform thickness of 0.908 mm. The specimen contains a cladding) total bondline area of 90 mm 250 µm. The lack of protective tight-buffered jacket on the fibre sensor meant that it had to be handled × 25.4 mm using FM300-2K film adhesive manufactured by Cytec Engineering (Carlton, VIC, carefully installation to avoidof breakage. Bareto fibres were overwere tight-buffered fibres Australia)during [20], Table 2. The location the structure which thepreferred fibre sensors bonded was first due to slippage can introduce error to the results. marked out. The outer surface layer on these surfaces was cleaned with acetone. The acrylate coated
bare Single Mode Fibre (SMF) 28e fibre optic sensors were then held in place with strain gauge tape 2.2. Specimen Design The adherends used in the step lap joint are manufactured from 24 plies thick HexPly M18/1/G939 carbon fibre prepreg (HEXCEL., Stamford, CT, USA) [19], Table 1; with a ply stacking sequence of [(0/90)/(45/-45)/(45/-45)/(0/90)]6 . Figure 3 shows a schematic of the composite structure with a width of 25.4 mm with each ply having a nominal uncured thickness of 0.41 mm and every step having
Materials 2016, 9, 374
4 of 11
a uniform thickness of 0.908 mm. The specimen contains a total bondline area of 90 mm ˆ 25.4 mm using FM300-2K film adhesive manufactured by Cytec Engineering (Carlton, VIC, Australia) [20], Table 2. The location of the structure to which the fibre sensors were bonded was first marked out. Materials 2016, 9, 374 layer on these surfaces was cleaned with acetone. The acrylate coated bare Single 4 of 11 The outer surface 9, 374 4 of 11 ModeMaterials Fibre 2016, (SMF) 28e fibre optic sensors were then held in place with strain gauge tape and bonded and bonded tosurface the prepped surface Loctite 406. Where possible, thecontinuously fibre sensors were to theand prepped Loctite 406. with Where possible, fibre possible, sensors were bonded bonded to thewith prepped surface with Loctite 406.the Where the fibre sensors were continuously bonded to the structure’s surface, to the structure’s surface, as shown in Figure 4. as shown in Figure 4. continuously bonded to the structure’s surface, as shown in Figure 4.
Figure 3. Schematic drawing of step lap joint composite specimen (red dotted lines indicate the
Figure 3. Schematic drawing of step lap joint composite specimen (red dotted lines indicate the location Figure 3. Schematic step lap composite specimen (red dotted lines indicate the location of bondeddrawing fibre) andof showing the joint clamped side. of bonded fibre) and showing the clamped side. location of bonded fibre) and showing the clamped side.
Figure 4. Demonstration of the bonded fibre (dotted line) along the composite specimen. 1. Material properties for HexPly M18/1/G939 prepreg [19].specimen. Figure 4. Table Demonstration of the bonded fibre (dotted line)carbon alongfibre the composite Figure 4. Demonstration of the bonded fibre (dotted line) along the composite specimen. Tensile Strength Compressive Strength E11 E22 G12 G13 Material Table 1. Material properties for HexPly ν12 M18/1/G939 σ11 σcarbon 22 σ12 prepreg [19]. fibre (GPa) (GPa) properties (MPa) (MPa) σ11 (MPa) Table 1. Material for HexPly M18/1/G939 carbon fibre prepreg [19]. σ22 (MPa) (MPa) (MPa) (MPa) Compressive M18/G939 E11 65 E22 67 0.04 800Tensile 800Strength 100 800 800Strength G124.0 G4.0 13
Tensile Strength ν12 σ 11 σ22 σ12 Compressive Strength E11 E22 G12 G(MPa) (GPa) (GPa) (MPa) σ11 (MPa) σ22 (MPa) 13 ν12 (MPa) (MPa) Table(MPa) 2. Material properties for FM300-2K film adhesive [20]. σ(MPa) σ22 σ12 (GPa) (GPa) (MPa) 11 σ11 (MPa) σ22 (MPa) M18/G939 65 67 4.0 4.0 0.04 (MPa) 800 (MPa) 800 (MPa) 100 800 800 Adhesive GIC GIIC E (MPa) G (MPa) ν X t (MPa) σ 12 (MPa) γ e γ p M18/G939 67 4.0 4.0 0.04 800 800 100 800 2) (kJ/m 800 2) Material65 (kJ/m Table 2. Material properties for FM300-2K film adhesive [20]. FM300-2K 2400 840 0.4 94.2 54.4 0.055 0.580 1.3 5 Table 2. Material properties for FM300-2K film adhesive [20]. Adhesive GIC GIIC E (MPa) G (MPa)Testing ν Xt (MPa) σ12 (MPa) γe γp 2.3. Experimental Setup—Fatigue 2 Material (kJ/m ) (kJ/m2)
Material Material
Adhesive GIC G The prepared INSTRONσ12 100(MPa) kN testing shown inIIC FM300-2K 2400 specimen 840was set0.4 54.4 0.055 0.580 1.3 5 2 E (MPa) G (MPa) νup onXan (MPa) γemachine γp [21], as t94.2 2
Material (kJ/m ) (kJ/m ) Figure 5. In Figure 5, the fibre bonded along the region of interest was labelled as Fibre 1 (left) and FM300-2K 2400 840 Testing 94.2 0.055 0.580 and1.3 Fibre 2 (right). Fibre 1 measured the 0.4 strain from the bottom 54.4 to the top of the structure vice versa.5 2.3. Experimental Setup—Fatigue In general, aircraft components are usually subjected to a vicinity of 3000 με loading region. In order The prepared specimen was region, set up on an INSTRON 100 kN testing machine [21], as shown in to cover the operating loading 2.3. Experimental Setup—Fatigue Testing the specimen are subjected to tension-tension block loading Figure 5. In as Figure 5, the fibre 3. bonded along the region ofainterest was labelled as Fibre 1 (left) regime identified in Table The experiment started with loading regime with peak strain of 1000 and The prepared specimen was set up on an INSTRON 100 kN testing machine [21], as shown in με and R-ratio of 0.1. On completion of 100,000 cycles, the peak strain was increased with the R-ratio Fibre 2 (right). Fibre 1 measured the strain from the bottom to the top of the structure and vice versa. maintained at 0.1. The entire fatigue test was conducted at a frequency of 5 Hz with a sinusoidal Figure 5. Inaircraft Figurecomponents 5, the fibre bonded along the region interestofwas as Fibre 1 (left) and In general, are usually subjected to aofvicinity 3000labelled με loading region. In order waveform and an1extensometer wasstrain used correlate strain to to load. until final Fibre 2 (right). Fibre measured the the bottom theThe top of was the run structure andfailure vice versa. to cover the operating loading region, thetofrom specimen are subjected totest tension-tension block loading of the test specimen. In general, aircraft components are usually subjected to a vicinity of 3000 µε loading region. In order to
regime as identified in Table 3. The experiment started with a loading regime with peak strain of 1000 με and R-ratio of 0.1. On completion of 100,000 cycles, the peak strain was increased with the R-ratio maintained at 0.1. The entire fatigue test was conducted at a frequency of 5 Hz with a sinusoidal waveform and an extensometer was used to correlate strain to load. The test was run until final failure of the test specimen.
Materials 2016, 9, 374
5 of 11
Table 3. Block loading regime—strain amplitude increased every 105 cycles starting at 1000 με. Materials 2016, 9, 374
5 of 11
Cycles (N) Strain (µε) R-Ratio Upper Limit (kN) Lower Limit (kN) 100,000 1000 0.1 6.30 0.63 200,000 loading 2000 0.1 12.60to tension-tension block 1.26 loading regime cover the operating region, the specimen are subjected 300,000 3000 18.90 regime with peak 1.89strain of 1000 µε as identified in Table 3. The experiment 0.1 started with a loading 400,000 4000 0.1 25.20 2.52 with the R-ratio and R-ratio of 0.1. On completion of 100,000 cycles, the peak strain was increased 4500fatigue test0.1 28.35 2.84with a sinusoidal maintained500,000 at 0.1. The entire was conducted at a frequency of 5 Hz 600,000 5000was used to 0.1correlate strain 31.50 waveform and an extensometer to load. The test was 3.15 run until final failure 700,000 5500 0.1 34.65 3.47 of the test specimen.
Figure 5. 5. Experimental setup of the composite specimen on testing machine with the red dotted line indicating the the bonded bonded fibre. fibre. The bonded fibre along the region of interest was labelled as Fibre 1 (left) indicating Fibre 22 (right). (right). and Fibre
3. Results and Table 3. Discussion Block loading regime—strain amplitude increased every 105 cycles starting at 1000 µε. 3.1. Dynamic Strain Cycles (N)Measurement Strain (µε)
R-Ratio
Upper Limit (kN)
Lower Limit (kN)
100,000 1000 In the experiment, the distributed strain0.1was sampled 6.30 at 17 Hz whilst the 0.63 fatigue loading was 200,000 2000 0.1 12.60 1.26 applied at 5 Hz. To show the ability to monitor dynamic strain, the measured strain at the centre 300,000 3000 0.1 18.90 1.89 point along 400,000 the Fibre 1 is displayed over a period of four seconds, Figure 6a. The 4000 0.1 25.20 2.52 measured strain shows a sinusoidal with maximum 500,000 waveform 4500 0.1 and minimum 28.35 strain recording 2.84at approximately 5000A Fast Fourier 0.1 Transform (FFT) 31.50 is performed on3.15 2000 με and 600,000 200 με, respectively. the time series and 0.1 34.65 the result is 700,000 shown in Figure 5500 6b. The cyclic frequency is calculated to be 5.11 Hz3.47 with an amplitude of 870 με (where amplitude = maximum extent of an oscillation from the equilibrium position). The 3. Results and Discussion results gathered clearly show that the strain measured by distributed optical fibre sensor corresponded to the fatigue loading profile whilst also providing real-time structural health 3.1. Dynamic Strain Measurement monitoring capabilities. The potential of using this high-speed distributed optical fibre sensor to monitor theexperiment, dynamic strain response along thewas length of theat test evident. In the the distributed strain sampled 17specimen Hz whilstisthe fatigue loading was applied at 5 Hz. To show the ability to monitor dynamic strain, the measured strain at the centre point along the Fibre 1 is displayed over a period of four seconds, Figure 6a. The measured strain shows a sinusoidal waveform with maximum and minimum strain recording at approximately 2000 µε and 200 µε, respectively. A Fast Fourier Transform (FFT) is performed on the time series and the result is shown in Figure 6b. The cyclic frequency is calculated to be 5.11 Hz with an amplitude of 870 µε (where amplitude = maximum extent of an oscillation from the equilibrium position). The results gathered clearly show that the strain measured by distributed optical fibre sensor corresponded to the fatigue loading profile whilst also providing real-time structural health monitoring capabilities. The potential of using this high-speed distributed optical fibre sensor to monitor the dynamic strain response along the length of the test specimen is evident.
Materials 2016, 9, 374
6 of 11
Materials 2016, 9, 374
6 of 11
Materials 2016, 9, 374
6 of 11
FigureFigure 6. Time series (4 strain measurement 2000 με loading regime; Fourier 6. (a) (a)6. Time series (4 s) s) of the strain measurement at2000 2000με µεloading loading regime; and (b) Fast Fast Fourier (a) Time series (4of s) the of the strain measurement at at regime; andand (b) (b) Fast Fourier Transform (FFT)(FFT) of the series. Transform of time the time series.
3.2. Fatigue Monitoring 3.2. Monitoring 3.2. Fatigue Fatigue Monitoring The distributed strain measured alongFibre Fibre11 and and Fibre structure waswas subjected to to The strain measured along Fibre when the structure subjected The distributed distributed strain measured along Fibre 1 and Fibre222when whenthe the structure was subjected to different loading regime is reported in Figure 7. It can be seen that the measurement obtained with different seen that that the the measurement measurement obtained different loading loading regime regime is is reported reported in in Figure Figure 7. 7. It It can can be be seen obtained with with Fibre 2 is essentially a mirror image of the strain distribution measured by Fibre 1. The strain profiles Fibre 22 is aa mirror image the measured by Fibre strain profiles Fibre measured is essentially essentially mirror image of of the strain strain distribution distribution measured by distributed Fibre 1. 1. The The strainfibre profiles by both fibres indicate a consistent strain measurement by the optical measured by both fibres indicate a consistent strain measurement by the distributed optical measured byThe both fibres indicate aalso consistent strain deformation measurement byboth the edges distributed optical fibre fibre sensors. strain measurement shows uniform along of the specimen. sensors. The strain measurement also shows uniform deformation along both edges of the specimen. sensors. The strain measurement also shows uniform deformation along both edges of the specimen. The consistency of the measured strain along Fibre 1 and 2 confirms that an adequate bond between The the measured strain Fibre and that bond between The consistency consistency of thestructure measured strain along along Fibreis11achieved. and 22 confirms confirms that an an adequate adequate bondfrom between the fibre andof host (composite specimen) Furthermore, the strain measured the fibre and host structure (composite specimen) is achieved. Furthermore, the strain measured bothand fibres show that the(composite top region of the specimen is experiencing a higher the strain thanmeasured the bottomfrom the fibre host structure specimen) is achieved. Furthermore, strain from region.show This isthat attributed the geometry the test specimen (Figure 3). both the of the of specimen is experiencing aa higher both fibres fibres show that the top toptoregion region of the specimen is experiencing higher strain strain than than the the bottom bottom region. (Figure 3). 3). region. This This is is attributed attributed to to the the geometry geometry of of the the test test specimen specimen (Figure
Figure 7. Distributed strain measurement along the bonded loop of fibre.
Ideally, the specimen should experience uniform strain along the entire specimen under tensile Figure Distributed strain7), measurement along thethe bonded loop has of fibre. testing, but from the 7. observation (Figure it clearly shows that top region greater strain than Figure 7. Distributed strain measurement along the bonded loop of fibre. the bottom of the specimen partly due to the localised change in adherend thickness provided by the stepped the adherend geometry. Additionally,uniform the result shown in Figure 7 shows a small strain Ideally, strain along the entire entire specimen under tensile Ideally, the specimen specimen should should experience experience uniform strain along the specimen under tensile measurement being detected at the “looped” region which is in between Fibre 1 and 2. This is because testing, but from the observation (Figure 7), it clearly shows that the top region has greater testing, but from the observation (Figure 7), it clearly shows that the top region has greater strainstrain than the fibre section between Fibre 1 and 2 was bonded to the surface of the host structure with aim of thanbottom the bottom of the specimen partly due to the localised change in adherend thickness provided the of the specimen partly due to the localised change in adherend thickness provided by avoiding high strain gradient. As the fibre can only measure the mechanical strain in axial direction, the
by the stepped adherend geometry. Additionally, the result shown in Figure showsaasmall small strain strain stepped adherend geometry. Additionally, the result shown in Figure 7 7shows measurement being detected at the “looped” region which is in between Fibre 1 and 2. This is because measurement being detected at the “looped” region which is in between Fibre 1 and 2. This is because the fibre fibre section section between between Fibre Fibre 11 and and 22 was was bonded bonded to to the the surface surface of of the the host host structure structure with with aim aim of of the avoiding high high strain strain gradient. gradient. As As the the fibre fibre can can only only measure measure the the mechanical mechanical strain strain in in axial axial direction, direction, avoiding
The entire fatigue test was conducted for approximately 14 h. The strain measured at the centre point of Fibre 1 over the last 100,000 cycles (before failure occurs) is presented in Figure 8. The strain measurement shows that around 200,000 cycles, the strain measurement detected a sudden increase in the tension-tension block loading regime (2000 με to the next tension-tension block loading regime with a strain amplitude of 3000 με). The INSTRON machine was programmed to stop for one minute Materials 2016, 9, 374 7 of 11 before it moves on to the next level of tension–tension block loading cycle. This is clearly detected by the distributed optical fibre sensors. This result demonstrates the capability of distributed optical fibrebonded sensor to in will situ monitoring of a structure. the fibre inprovide the loop still experience the strain in the “loop” direction, however this is not The failure of the stepped lap joint composite specimen occurs after 254,362 cycles in this fatigue an area of interest. The results presented showed that a reliable strain distribution can be measured. Materials 2016, 9, 374 7 of 11 loading test. As highlighted by the dotted region in Figures 8 and 9, the strain gradually changes in The entire fatigue test was conducted for approximately 14 h. The strain measured at the centre the last 3500 cycles before failure occurs. It is important to note that the distributed optical fibre fibre bonded thelast loop will still experience thefailure strain in the “loop” direction, in however is notstrain pointthe of Fibre 1 over in the 100,000 cycles (before occurs) is presented Figurethis 8. The sensors was able to record the change in strain measured before failure occurs. These features can be an area of interest. The results presented showed that a reliable strain distribution can be measured. measurement thatof around 200,000 cycles, theThese strain measurement detected a sudden increase in used as anshows indication damage or crack growth. results demonstrate the ability of distributed The entire fatigue test was conducted for approximately 14 h. The strain measured at the centre the tension-tension block loadingfor regime (2000 µε to theofnext tension-tension block loading regime strain measurement technique the in-situ monitoring damage along the bonded step lap joints. point of Fibre 1 over the last 100,000 cycles (before failure occurs) is presented in Figure 8. The strain with aThe strain amplitude of 3000the µε). Theduration INSTRON machine to stop minute optical fibre survived entire of this fatiguewas test programmed as the failure strain of for the one optical measurement shows that around 200,000 cycles, the strain measurement detected a sudden increase fibremoves exceeds of the adhesive joint. When the fibreblock is broken, the breakage point detectable by by before onthat to the next level of tension–tension loading cycle. This is is clearly initthe tension-tension block loading regime (2000 με to the next tension-tension block loading detected regime using an optical backscattered reflectometry (OBR) [22,23]. The location of the fracture point along fibre the distributed optical fibreofsensors. demonstrates theprogrammed capability oftodistributed with a strain amplitude 3000 με).This Theresult INSTRON machine was stop for one optical minute the fibres situ can be used as anof indication of damage and cracks along the structure. sensor tooptical provide monitoring structure. before it movesinon to the next level of atension–tension block loading cycle. This is clearly detected by the distributed optical fibre sensors. This result demonstrates the capability of distributed optical fibre sensor to provide in situ monitoring of a structure. The failure of the stepped lap joint composite specimen occurs after 254,362 cycles in this fatigue loading test. As highlighted by the dotted region in Figures 8 and 9, the strain gradually changes in the last 3500 cycles before failure occurs. It is important to note that the distributed optical fibre sensors was able to record the change in strain measured before failure occurs. These features can be used as an indication of damage or crack growth. These results demonstrate the ability of distributed strain measurement technique for the in-situ monitoring of damage along the bonded step lap joints. The optical fibre survived the entire duration of this fatigue test as the failure strain of the optical fibre exceeds that of the adhesive joint. When the fibre is broken, the breakage point is detectable by using an optical backscattered reflectometry (OBR) [22,23]. The location of the fracture point along the optical fibres can be used as an indication of damage and cracks along the structure.
Figure 8. Strain measurement at at the Fibre11100,000 100,000 cycles before failure occurs. Figure 8. Strain measurement thecentre centrepoint point along along Fibre cycles before failure occurs.
The failure of the stepped lap joint composite specimen occurs after 254,362 cycles in this fatigue loading test. As highlighted by the dotted region in Figures 8 and 9 the strain gradually changes in the last 3500 cycles before failure occurs. It is important to note that the distributed optical fibre sensors was able to record the change in strain measured before failure occurs. These features can be used as an indication of damage or crack growth. These results demonstrate the ability of distributed strain measurement technique for the in-situ monitoring of damage along the bonded step lap joints. The optical fibre survived the entire duration of this fatigue test as the failure strain of the optical fibre exceeds that of the adhesive joint. When the fibre is broken, the breakage point is detectable by using an optical backscattered reflectometry (OBR) [22,23]. The location of the fracture point along the optical fibres can8.be used as an indication of damage and cracks along thebefore structure. Figure Strain measurement at the centre point along Fibre 1 100,000 cycles failure occurs.
Figure 9. Strain measurement at the centre point along Fibre 1 35,000 cycles before failure occurs.
Figure 9. Strain measurement thecentre centrepoint point along cycles before failure occurs. Figure 9. Strain measurement atatthe alongFibre Fibre1 135,000 35,000 cycles before failure occurs.
Materials 2016, 9, 374 Materials 2016, 9, 374
8 of 11 8 of 11
A 3D plot presented in Figure 10 shows the distributed strain measured along Fibre 1 during the 3000 με block loading regime prior to specimen failure. This is one of the characteristics of the Materials 2016, 9, 374 8 of 11 A 3D plot presented in Figurewhich 10 shows theused distributed strainstructural measuredhealth along in Fibre 1 during distributed optical fibre sensors can be to monitor term of both the 3000 µε block loading regime prior to specimen failure. This is one of the characteristics of distributed and temporal information. showsstrain the capability of DOFS perform A 3Dspatial plot presented in Figure 10 shows theThis distributed measured along Fibreto 1 during thea realthe distributed optical fibre sensors can be used toThis monitor structural health in term of botha time andμεcontinuous monitoring ofwhich the measured structure. Inone Figure 10,characteristics the results also show 3000 block loading regime prior to specimen failure. is of the of the distributed spatial and temporal information. This shows the capability of1DOFS perform aboth real-time distributed optical fibre sensors which can be the used to monitor structural healthtoin term significant change in strain measurement along entire bonded Fibre during the last of 4000 loading distributed spatial and temporal information. This shows the με capability DOFSalso toFigure perform realand continuous monitoring of measured Figure 10, the of results show11a aasignificant cycles. The structure failed at the 254,362 cyclesstructure. during theIn3000 loading regime. shows the time and continuous monitoring of the measured structure. In Figure 10, the results also show acycles. change in strain measurement bonded 1 during the last 4000showed loadingfirst composite specimen after final along failurethe hasentire occurred. TheFibre bonded joint configuration ply significant change in strain measurement along the entire bonded Fibre 1 during the last 4000 loading The structure failed at 254,362 cycles during the 3000 µε loading regime. Figure 11a shows the failure followed by net-tension failure as well as the adhesive failure, Figure 11b,c. cycles. The structure failed at 254,362 cycles during the 3000 με loading regime. Figure 11a shows the composite specimen after final failure has occurred. The bonded joint configuration showed first ply composite specimen after final failure has occurred. The bonded joint configuration showed first ply failure followed by net-tension failure as well as the adhesive failure, Figure 11b,c. failure followed by net-tension failure as well as the adhesive failure, Figure 11b,c.
Figure Figure10. 10.Distributed Distributedstrain strainmeasurement measurementalong alongFibre Fibre11for forthe the3000 3000µε μεloading loadingregime. regime. Figure 10. Distributed strain measurement along Fibre 1 for the 3000 με loading regime.
Figure 11. (a) Failed specimen; (b) side view of the failed specimen; and (c) the features of failed specimen thespecimen; bondline. Figure 11. (a) (a)along Failed specimen; (b) (b) side side view view of ofthe thefailed failedspecimen; specimen; and and(c) (c)the thefeatures featuresofoffailed failed Figure 11. Failed
specimenalong alongthe thebondline. bondline. specimen The distributed strain measurement technique is implemented to monitor bondline cracking during the fatigue test. High stress concentrations are found at the ends of the bondline overlap and The distributed strain measurement technique is implemented to monitor bondline cracking The distributed measurement is implemented to more monitor bondline hence it is expectedstrain bondline cracking willtechnique originate from these regions. A detailed analysiscracking of during the fatigue test. High stress concentrations arethe found at the cycles ends of the bondline overlap and the the strain measurement alongstress Fibre 1concentrations was looked at inare last 50,000 before final failure. The during fatigue test. High found at the ends of the bondline overlap hence it is expected bondline cracking will originate from these regions. Acalculated more detailed analysis of in expected strain measured along Fibre 1 at theoriginate bondeline overlap wasregions. = and difference hence it is bondline cracking will from these Ausing: moreΔεdetailed the strain measurement along Fibre 1 was looked at inmeasurement the last 50,000 before final failure. The corresponding strain measurement—reference (at cycles 200,000 cycles). This was final analysis of the strain measurement along Fibre 1strain was looked at in the last 50,000 cycles before difference in strain measured along Fibre 1 atin the bondeline overlap was calculated using: Δε = plotted over the last 50,000 cycles and is shown Figure 12. failure. The difference in strain measured along Fibre 1 at the bondeline overlap was calculated using:
corresponding strain measurement—reference strain measurement (at 200,000 cycles). This was ∆ε = corresponding strain measurement—reference strain measurement (at 200,000 cycles). This was plotted over the last 50,000 cycles and is shown in Figure 12. plotted over the last 50,000 cycles and is shown in Figure 12.
with prior understanding that damage initaition occurs at the ends of the overlap due to the high stress concentration in these outer overlap regions. Overall the repeated loading and unloading of the composite joint structure promotes damage progression. Eventually in the last 4000 cycles of the fatigue loading regime, cracks are seen from either side of the bondline overlap. This promotes greater out of plane bending resulting in a significant increase in the strain measured by the optical Materials 2016, 9, 374 9 of 11 fibre (252,000 cycles to 254,000 cycles).
Difference in strain strain measurement measurement along along the the bonded bonded fibre fibre (Fibre (Fibre 1) at 50,000 cycles before Figure 12. Difference failure occurs.
The findingsarea overall that the distributed optical sensors seemed to offerthe a The bonded has conclude a total overlap length of 90 mm. Fromfibre Figure 12, 0 mm represents considerable promise in damage assessment and monitoring of fatigue crack growth along bondlines. starting point of the data gathered by the fibre which is located closest to the bottom edge of the This distributed also an alternative method for be detecting and overlap whilst 90strain mm ismeasurement the location oftechnique the fibre at theoffers top edge of the specimen. It can clearly seen monitoring a structure whilst in flight as well as on the ground. The distributed optical fibre strain that the strain at the bottom edge of the specimen remains relatively constant between 220,000 cycles measurement approach has the the effects of damage interactive and and 250,000 cycles. However, atpotential the otherfor endevaluating of the overlap (90 mm) there is awith gradual increase in predictive capabilities. Thus, effective and efficient repair techniques can be applied to the damage strain. Due to the localised change in adherend thickness from the stepped geometry, the top section structures before the likelihood of catastrophic of the overlap expereineces a greater localisedfailure. strain measurement due to only being 4 plies thick whilst the bottom end is 22 plies thick. Furthermore, the gradual increase in strain as the number of 4. Conclusions cycles increase suggest the likelihood of damage progression i.e., disbond in this case. This agrees with Distributed prior understanding that damage initaition at the ends for of the overlapstructural due to thehealth high optical fibre sensing system is an occurs attractive scheme composite stress concentration in these outer overlap regions. Overall the repeated loading and unloading monitoring from the point of manufacture through to operation. The distributed strain technique of is the compositeexperimentally joint structure promotes damage progression. in the 4000 cycles implemented to monitor the onset of fatigueEventually damage and the last eventual failureofofthe a fatigue loading regime, cracks are seen from either side of the bondline overlap. This promotes greater step lap joint. The results gathered clearly show that the onset and propagation of the fatigue damage out be of monitored plane bending in a significant increase in the strain measured the optical fibre can usingresulting this distributed strain measurement technique prior to its by catastrophic failure (252,000 cyclesThe to 254,000 cycles). under fatigue. distributed strain measured along the fibre also shows the propagation of a crack conclude distributed sensors to offer alongThe the findings adhesionoverall of the stepped lapthat jointthe due to fatigue optical loading.fibre Through thisseemed distributed straina considerable promise in damage assessment and monitoring of fatigue crack growth along bondlines. measurement technique, in-situ and continuous the condition of the bonded step lap This distributed strain measurement technique also offers an alternative method for detecting and monitoring a structure whilst in flight as well as on the ground. The distributed optical fibre strain measurement approach has the potential for evaluating the effects of damage with interactive and predictive capabilities. Thus, effective and efficient repair techniques can be applied to the damage structures before the likelihood of catastrophic failure. 4. Conclusions Distributed optical fibre sensing system is an attractive scheme for composite structural health monitoring from the point of manufacture through to operation. The distributed strain technique is implemented experimentally to monitor the onset of fatigue damage and the eventual failure of a step lap joint. The results gathered clearly show that the onset and propagation of the fatigue damage can be monitored using this distributed strain measurement technique prior to its catastrophic failure under fatigue. The distributed strain measured along the fibre also shows the propagation of a crack along the adhesion of the stepped lap joint due to fatigue loading. Through this distributed strain measurement technique, in-situ and continuous monitoring of the condition of the bonded step lap
Materials 2016, 9, 374
10 of 11
joints without taking the structure out of service for inspection and assessment. This system offers an alternative method for detecting and monitoring a structure whilst in flight as well as on the ground. Acknowledgments: This work was supported by “Advanced Condition Assessment & Pipeline Failure Prediction Project” funded by Monash University. The research partners are Monash University (lead), the University of Technology Sydney, and the University of Newcastle. This work is also partly supported by Institute of Railway Technology (IRT) from Monash University. Author Contributions: Leslie Wong and Nabil Chowdhury assist in designing the study, developing the methodology, collecting the data, performing the analysis and wrote the manuscript. While, the equipment used in the presented work are provided by Wing Kong Chiu and Jayantha Kodikara. Conflicts of Interest: The authors declare no conflict of interests regarding the publication of this paper.
References 1.
2. 3. 4. 5.
6. 7. 8.
9. 10.
11. 12. 13. 14.
15. 16. 17. 18.
Mollenhauer, D.; Fredrickson, B.; Schoeppner, G.; Iarve, E.; Palazotto, A. Analysis and measurement of scarf-lap and step-lap joint repair in composite laminates. In Proceedings of the 16th International Conference on Composite Materials, Kyoto, Japan, 2 July 2007. Kim, J.; Park, B.; Han, Y. Evaluation of fatigue characteristic for adhesively-bonded composite stepped lap joints. Compos. Struct. 2004, 66, 69–75. [CrossRef] Viktorov, I. Rayleigh and Lamb Waves: Physical Theory and Applications; Springer: Berlin, Germany, 1967. Rose, J. Ultrasonic Waves in Solid Media; Cambridge University Press: Cambridge, UK, 2004. Oleksii, K.; Anton, K.; Ermias, K.; Mahmoodul, H.; Lalita, U. Monitoring of Fatigue Damage in Composite Lap-Joints using Guided Waves and FBG Sensors. In Proceedings of the AIP Conference Proceedings, 1706, Minneapolis, MN, USA, 10 February 2016. Gandhi, N.; Michaels, J.; Lee, S. Acoustoelastic Lamb Wave Propagation in Biaxially Stressed Plates. J. Acoust. Soc. Am. 2012, 132, 1284–1293. [CrossRef] [PubMed] Sorin, W.; Baney, D. Measurement of rayleigh backscatter at 1.55 µm with 32 µm spatial resolution. IEEE Photonics Technol. Lett. 1992, 4, 374–376. [CrossRef] Zou, L.; Ferrier, G.; Afshar, S.; Yu, Q.; Chen, L.; Bao, X. Distributed Brillouin scattering sensor for discrimination of wall-thinning defects in steel pipe under internal pressure. Appl. Opt. 2004, 43, 1583–1588. [CrossRef] [PubMed] Zou, L.; Bao, X.; Ravet, F.; Chen, L. Distributed Brillouin fiber sensor for detecting pipeline buckling in an energy pipe under internal pressure. Appl. Opt. 2006, 45, 3372–3377. [CrossRef] [PubMed] Gifford, D.K.; Kreger, S.T.; Sang, A.K.; Froggatt, M.E.; Duncan, R.G.; Wolfe, M.S.; Soller, B.J. Swept-wavelength Interferometric Interrogation of Fiber Rayleigh Scatter for Distributed Sensing Application. Fiber Opt. Sens. Appl. V 2007, 6770, F1–F9. Inaudi, B.; Glisic, D. Integration of distributed strain and temperature sensors in composite coiled tubing. Smart Struct. Mater. Int. Soc. Opt. Photonics 2006, 6167, 319–328. Lim, K.; Wong, L.; Chiu, W.; Kodikara, J. Distributed fiber optic sensors for monitoring pressure and stiffness changes in out-of-round pipes. Struct. Control Health Monit. 2016, 23, 303–314. [CrossRef] Graham, D.; Guillaume, R.; Raymond, M. The mechanical load transfer into distributed optical fiber sensor due to a linear strain gradient: Embedded and surface bonded cases. Smart Mater. Struct. 1999, 8, 175–181. Bernasconi, A.; Carboni, M.; Comolli, L.; Galeazzi, R.; Gianneo, A.; Kharshiduzzaman, M. Fatigue crack growth monitoring in composite bonded lap joints by a distributed fibre optic sensing system and comparison with ultrasonic testing. J. Adhes. 2015. [CrossRef] Eickhoff, W.; Ulrich, R. Optical frequency domain reflectometry in single-mode fiber. Appl. Phys. 1981, 39, 693–695. [CrossRef] Glombiza, U.; Brinkmeyer, E. Coherent frequency domain reflectometry for characterization of single-mode integrated optical waveguides. J. Lightwave Technol. 1993, 11, 1377–1384. [CrossRef] Soller, B.J.; Wolfe, M.; Froggatt, M.E. Polarization Resolved Measurement of Rayleigh Backscatter in Fiber-Optic Components; OFC Technical Digest: Los Angeles, CA, USA, 2005. Rahim, N.A.A.; Thoreson, M.A.; Gorney, T.; Garg, N.; Gifford, D.K.; Froggatt, M.E.; Sang, A.K. Superior Fatigue Characteristics of Fiber Optic Strain Sensors; LunaInc Engineering Note: Roanoke, VA, USA, 2013.
Materials 2016, 9, 374
19. 20. 21. 22.
23.
11 of 11
HEXCEL. HexPly M18/1 180 Degree Celcius Curing Epoxy Matrix Product Data; HEXCEL: Stamford, CT, USA, 2011. FM 300-2 Film Adhesive Technical Data Sheet. 2011. Available online: https://www.cytec.com/sites/ default/files/datasheets/FM_300-2_092711.pdf (accessed on 23 February 2016). Chowdhury, N.; Chiu, W.; Wang, J.; Chang, P. Static and fatigue testing thin riveted, bonded & hybrid carbon fiber double lap joints used in aircraft structures. Compos. Struct. 2015, 121, 315–323. Alfredo, G.; Julian, S.; Frank, G.; Toni, K.; Pavlos, M.; Dimitrios, H.; Moshe, T.; Nahum, G.; Christos, K.; Nikolaos, K. Methodologies for the Damage Detection Based on Fibre-Optic Sensors. Applications to the Fuselage Panel and Lower Wing Panel. In Smart Intelligent Aircraft Structures (SARISTU); Springer International Publishing: Berlin, Germany, 2016; pp. 407–431. Stancalie, A.; Gelu, I. Experimental Characterization of the Optical Fibre Sensors. UPB Sci. Bull. Ser. A 2016, 78, 299–308. © 2016 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/).
