Preclinical comparison of regorafenib and sorafenib efficacy for hepatocellular carcinoma using multimodality molecular imaging

materials Article

On-Site Health Monitoring of Composite Bolted Joint Using Built-In Distributed Eddy Current Sensor Network Qijian Liu † , Hu Sun † , Tao Wang and Xinlin Qing * School of Aerospace Engineering, Xiamen University, Xiamen 361005, China * Correspondence: [email protected]; Tel.: +86-187-5928-7299 † These authors contributed equally to this work and should be considered co-first author. Received: 13 August 2019; Accepted: 28 August 2019; Published: 29 August 2019







Abstract: There is an urgent need to monitor the structural state of composite bolted joints while still remaining in service; however, there are many difficulties in analyzing their strength and failure modes. In this paper, a built-in distributed eddy current (EC) sensor network based on EC array sensing film is developed to monitor the hole-edge damages of composite bolted joints. The EC array sensing film is bonded onto the bolt and consists of one exciting coil and four separate sensing coils. Experiments are conducted on unidirectional composite specimens to validate the ability of the EC array sensing film to quantitatively track the damage that occurs at the hole edge and to investigate the performances of the EC array sensing films with different configurations of the exciting coil. Experimental results show that the induced voltage of sensing coil changes only if the damage appears on the laminate structure where that particular sensing coil is located, whereas the induced voltages of the other sensing coils on other laminate plates remain unchanged. Numerical simulation based on the finite element method is also carried out to investigate and explain the phenomena observed in the experiments and to analyze the distribution of the EC around the bolt hole. Both experimental and numerical simulation results demonstrate that the developed EC array sensing film can effectively identify not only whether there is damage at the hole edge but also the damage location within the thickness and quantitative size. Keywords: composite; bolted joint; structural health monitoring; built-in; eddy current sensor network

1. Introduction Fiber-reinforced composite materials are gaining more and more importance in aerospace, automotive, shipbuilding, and other industries due to their high strength-to-weight and stiffness-to-weight ratios. By utilizing these advanced composite materials, large aircrafts are able to improve its efficiency while reducing its weight and operational costs [1]. As one of the key components of large composite structures, joints are widely used to ensure the integrity of composite structures. The mechanically fastened joint is widely used for connections within the composite structure due to its exceptionally strong bearing capacity, high reliability, and ease of disassembly and maintenance. The structural integrity of bolted joints can generate a significant impact on the stability and safety. Furthermore, the knowledge of the failure load and the response of bolted joints is critical for the design of the structures [2–4]. Nevertheless, it is very difficult to solve the strength and failure modes of composite bolted joints due to multiple factors. Due to the structural importance of joints, material degradation and other types of damages in the joint is a major concern. Consequently, timely and accurate damage detection along with active characterization and monitoring of delamination is needed. Damages in composite structures can be much more difficult to inspect and evaluate compared to metallic structures. Conventional nondestructive inspection techniques utilizing a variety of methods such as ultrasonic C-scan, eddy current, thermography, and shearography are difficult to perform Materials 2019, 12, 2785; doi:10.3390/ma12172785

www.mdpi.com/journal/materials

Materials 2019, 12, 2785

2 of 12

during operation due to the inaccessibility of the joint. Therefore, there is an urgent need to develop a method to monitor the structural state of the composite bolted joints while staying in service. By using a built-in sensor network integrated with the structure, structural health monitoring (SHM) can supply crucial information corresponding to the current condition and damage state of the structure [5–13]. SHM is showing great promise of being embraced by the industries as a capable method of monitoring the structural condition throughout its lifetime without requiring disassembly. An overview of structural health monitoring of composite joints can be found in the literature [14]. Several SHM technologies have been developed for damage inspection of bolted joints. Jalalpour et al. [15] made a combination of ultrasonic signals with fuzzy pattern recognition to monitor the integrity of 90◦ bolted joints. Rakow and Chang [16] developed a flexible eddy current (EC) sensing film bonded on the bolt to detect the existence of bolt hole-edge crack and its propagation. Sun and Qing et al [17,18] further developed the EC array sensing film bonded on the bolt to quantitatively monitor the crack growth along the radial direction and axial direction of the bolt hole. Some other methods, including acoustic emission [19], acoustic-ultrasonics [20], electromechanical impedance [21], comparative vacuum monitoring (CVM) [22], and intelligent coating monitoring (ICM) [23], can also be used to inspect the damage of bolted joints. However, these researches with flexible EC sensing film are focused on the metallic joints because of their good conductivity. The concept of a multi-field coupled sensing network based on piezoelectric sensors, EC sensors, and the Rogowski coils was developed to allow the structural state to be actively monitored while the structure remains in service [24]. However, it is difficult to construct a multi-field coupled sensing network for composite bolted joints due to some challenges and issues, including how to design the EC sensors for composite joints and how to integrate a variety of sensors together. In this study, a built-in distributed EC sensor network based on EC array sensing film is developed to quantitatively monitor the damages around the bolt hole of composite joints. The anisotropic conductivity of carbon fiber reinforced plastic (CFRP) is analyzed and employed in the numerical simulation. The principles and configuration designs of EC sensing film are first described. Then a series of experiments are conducted on unidirectional composite specimens to prove the feasibility of EC array sensing film to track the hole-edge damage and to investigate the performances of EC array sensing films in the condition of different exciting coil configurations. Numerical simulation is finally carried out to investigate and explain the phenomena observed in the experiments. 2. Principles and Configuration Design of Eddy Current Sensing Film 2.1. Anisotropic Conductivity Property of CFRP The structure of CFRP consists of carbon fibers and resin matrix. The CFRP offers excellent electrical conductivity along fiber direction. Additionally, considerable transverse electrical conductivity is also observed in carbon fiber composites since there are several contact points between fibers during the high-temperature curing process of prepregs. In other words, the composite materials are electrically conductive in all directions. Different orientations and cross-connections of carbon fibers will influence the mechanical and electrical properties of CFRP. Furthermore, the volume fraction of carbon fiber also affects the conductivity of CFRP. In general, the volume fraction of carbon fiber in CFRP is about 60%–70%. Figure 1a shows a schematic diagram of contact points between carbon fibers, and Figure 1b represents the current paths between carbon fibers in CFRP materials. The contact between the adjacent fibers constructs a closed conductive loop in CFRP and the current can flow along the path. However, the conductivity of composite material will change when damage appears. For example, fiber fracture or structure failure will cause resistance to increment due to the breakage of fibers or connection points between fibers.

Materials 2019, 12, 2785 Materials2019, 2019,12, 12,xxFOR FORPEER PEERREVIEW REVIEW Materials

3 of 12 of 13 13 33 of

(a) (a)

(b) (b)

Figure The schematic diagram of conductivity of carbon fiber reinforced plastic (CFRP): (a) Contact Figure1.1. 1.The Theschematic schematicdiagram diagramof ofconductivity conductivityof ofcarbon carbonfiber fiberreinforced reinforcedplastic plastic(CFRP): (CFRP):(a) (a)Contact Contact Figure pointsbetween betweencarbon carbonfibers; fibers;(b) (b)Current Currentpath pathbetween betweencarbon carbonfibers. fibers. points

The electrical in in CFRP can be expressed in threein directions (longitudinal, transversal, The electricalconductivity conductivity in CFRP CFRP can be expressed expressed in three directions directions (longitudinal, The electrical conductivity can be three (longitudinal, and through-thickness). The value of conductivity in three directions are presented as σ , σ transversal, and through-thickness). The value of conductivity in three directions are presented as cp . T , and σ transversal, and through-thickness). The value of conductivity in three directions are Lpresented as 3 4 conductivity, σL , conductivity, is between 5 ×σσ 10,, isis and 5 × 10 5S/m along and σσ .. The The longitudinal conductivity, between 5 10 10 andthe 10 The S/mtransversal along the the σσThe ,, σσlongitudinal ,, and longitudinal between and 55 fiber. 10 S/m along conductivity, σ , varies from 10 to 100 S/m [25]. The plies of CFRP laminates are contacted during , varies from 10 to 100 S/m [25]. The plies of CFRP laminates fiber. The transversal conductivity, σ T fiber. The transversal conductivity, σ , varies from 10 to 100 S/m [25]. The plies of CFRP laminates thecontacted manufacturing process which provides a through-thickness betweenconductivity the adjacent are contacted during the manufacturing manufacturing process which provides providesconductivity through-thickness conductivity are during the process which aa through-thickness plies. Nevertheless, the conductivity between each ply between of CFRP,each σcp ,ply is much smaller than the between theadjacent adjacentplies. plies. Nevertheless, theconductivity conductivity between each ply ofCFRP, CFRP, ismuch much between the Nevertheless, the of σσ ,,is transversal direction. The expression of anisotropic conductivity of CFRP can be represented as smaller than than the the transversal transversal direction. direction. The The expression expression of of anisotropic anisotropic conductivity conductivity of of CFRP CFRP can can be bea smaller conductivity tensor relied on the rotated coordinate system (Figure 2). The generalized definition of represented as as aa conductivity conductivity tensor tensor relied relied on on the the rotated rotated coordinate coordinate system system (Figure (Figure 2). 2). The The represented conductivitydefinition tensor is of calculated in a matrix in Equation (1), as and θ is a in parameter relative to generalized definition ofconductivity conductivity tensoras calculated inaamatrix matrix asshown shown inEquation Equation (1),and and generalized tensor isisshown calculated in (1), the fiber orientation. parameterrelative relativeto tothe thefiber fiberorientation. orientation. θθ isisaaparameter σσ −−σ σσ −σ σ cos θ2 L sin T 2θ σ sin sinsin sin 2θ(2θ) 00 ⎤⎤ 0  cos θθ2 (θ) sin cosθ (θ)σ+ σ ⎡⎡σ σ L T  2 2 2 = ⎢⎢  σLσ −σT 2 2 (θ) ⎥⎥ 0   σ − σ − σ ( ) ( ) σ = (1) sin 2θ σ sin θ + σ cos   σ (1) σ ⎢⎢ (1) 2 sin sin 2θ 2θ sinL θθ σσ cos cosT θθ σσ sin 00 ⎥⎥   2 2 ⎢ ⎥ 0 0 σ ⎢ ⎥ cp 00 00 σσ ⎦⎦ ⎣⎣

Figure 2. 2. Rotating Rotating coordinate coordinate system. system. Figure Figure 2. Rotating coordinate system.

2.2. Principles of Monitoring 2.2.Principles PrinciplesofofMonitoring Monitoring 2.2. The basic principles of monitoring using the EC sensor are shown in Figure 3. The EC sensor Thebasic basicprinciples principlesof ofmonitoring monitoringusing using theEC EC sensorare areshown shownin inFigure Figure3.3.The TheEC ECsensor sensorisis The is manufactured on a flexible substrate with the printedsensor flexible circuit technology and bonded on the manufacturedon on aflexible flexiblesubstrate substratewith withprinted printedflexible flexiblecircuit circuittechnology technologyand andbonded bondedon on thebolt bolt manufactured bolt of compositeajoint. To achieve the quantification of monitoring a bolt hole-edge damagethe of both of composite joint. To achieve the quantification of monitoring a bolt hole-edge damage of both radial ofradial composite joint.directions, To achievethe the EC quantification of monitoring a bolt hole-edge damage of bothsensing radial and axial sensing film is formed by an exciting coil and several andaxial axialdirections, directions,the theEC ECsensing sensingfilm filmisisformed formedby byan anexciting excitingcoil coiland andseveral severalsensing sensingcoils coils[17]. [17]. and coils [17]. According to Faraday’s Law, when an alternating electric current is applied to the exciting coil, According to to Faraday’s Faraday’s Law, Law, when when an an alternating alternating electric electric current current isis applied applied to to the the exciting exciting coil, coil, aa According a primary alternating magnetic field will be produced and surround the exciting coil, and then an EC primaryalternating alternating magneticfield field willbe beproduced producedand andsurround surroundthe theexciting excitingcoil, coil,and andthen thenan anEC EC primary will be generated inmagnetic the conductivewill composite material. The EC will generate a secondary alternating willbe begenerated generatedin inthe theconductive conductivecomposite compositematerial. material.The TheEC ECwill willgenerate generateaasecondary secondaryalternating alternating will magneticfield, field,restraining restrainingthe theoriginal originalmagnetic magneticfield. field.When Whendamage damageoccurs occursat atthe thebolt bolthole-edge, hole-edge,the the magnetic

Materials 2019, 12, 2785

Materials 2019, 12, x FOR PEER REVIEW

4 of 12

4 of 13

restraining due the original damage occurs the bolt hole-edge, will generatedmagnetic EC willfield, be disturbed to the magnetic change field. of theWhen original flow pathatand consequently, the generated EC will be disturbed due to the change of the original flow path and consequently, also affect the secondary magnetic field. By analyzing the signals captured from the sensing coil, holewill also affect the secondary magnetic field. By analyzing the signals captured from the sensing coil, edge damages in the composite laminates can be can identified. hole-edge damages in the composite laminates be identified.

Figure 3. Operating principle of eddy current sensing film mounted on the bolt of the composite joint.

Figure 3. Operating principle of eddy current sensing film mounted on the bolt of the 3. Experimental Investigation

composite joint.

A series of experiments were designed to prove the capability of the EC array sensing film to quantitatively track the damage around hole edge, and the performances of the EC array sensing films 3. Experimental Investigation were investigated with different configurations of the exciting coil. EC array sensing used in the experiments has one exciting the entirefilm to A series Each of experiments werefilm designed to prove the capability of thecoil ECcovering array sensing region track of the inner wall and four separate sensing coils laying the axial direction the bolt hole. quantitatively the damage around hole edge, and thealong performances of theof EC array sensing As shown in Figure 4, each sensing coil is constructed with two sections of traces, which are separated films were investigated with different configurations of the exciting coil. by the dielectric film and connected through welding at the weld hole near the center of the traces. EachThe ECexciting array coil sensing film in the experiments has one exciting coil covering the entire has the sameused structure as sensing coils. Two different configurations of the exciting region of coil thewere inner wall andinfour separate sensing laying along the direction of the bolt investigated the experiments. Both ofcoils the two configurations areaxial similar to the sensing coil, whereas all of the exciting traces are series connection one sections coil. The main difference hole. As shown in Figure 4, each sensing coilinistheconstructed withastwo of traces, which are between the two configurations of exciting coil lies in the direction of current path on the boundary of separated by the dielectric film and connected through welding at the weld hole near the center of two adjacent exciting sub-coils. In one configuration, the direction of current at the boundary of two the traces. The exciting coil has the same structure as sensing coils. Two different configurations of adjacent sub-coils is the same, while the opposite current direction is set in the other configuration. the exciting coil were investigated in the experiments. Both of the two configurations are similar to Figure 4 shows the plot of exciting coil with the exciting current in the boundary trace of one exciting the sensing coil, being whereas of direction the exciting are in the seriestrace connection one coil. The main sub-coil in theall same as thattraces in the adjacent boundary of anotheras exciting sub-coil, can bethe achieved by applying the current from external differencewhich between two configurations of exciting coil liescircuits. in the direction of current path on the

boundary of two adjacent exciting sub-coils. In one configuration, the direction of current at the boundary of two adjacent sub-coils is the same, while the opposite current direction is set in the other configuration. Figure 4 shows the plot of exciting coil with the exciting current in the boundary trace of one exciting sub-coil being in the same direction as that in the adjacent boundary trace of another exciting sub-coil, which can be achieved by applying the current from external circuits.

Materials 2019, 12, 2785 x FOR PEER REVIEW Materials 2019, 12, x FOR PEER REVIEW

55of of 13 12 5 of 13

Figure 4. Schematic plot of exciting coil. Figure 4. Schematic plot of exciting coil.

3.1. Experimental Experimental Set-Up Set-Up 3.1. 3.1. Experimental Set-Up The EC above were were The EC array array sensing sensing films films with with different different configurations configurations of of exciting exciting coil coil described described above The EC array sensing films withBoth different configurations of and exciting coil described above were mounted on the bolts, respectively. the width of each trace the distance between all of the mounted on the bolts, respectively. Both the width of each trace and the distance between all of the mounted on the bolts, respectively. Both the width of each trace and the distance between all of the traces are are 0.25 The whole whole structure structure of of the the EC EC sensing film was of sensing traces 0.25 mm. mm. The sensing film was comprised comprised of sensing coils coils with with traces are and 0.25an mm. The whole structure of theThe ECparameter sensing film was comprised of sensing coils with 20 traces exciting coil with 80 traces. for the height and the length of the EC 20 traces and an exciting coil with 80 traces. The parameter for the height and the length of the EC 20 traces and an exciting coiland with 80 traces. The parameter for the height and thehole length ofmm the and EC array sensing film is 20 mm 43.96 mm, respectively. The diameter of the bolt is 14 array sensing film is 20 mm and 43.96 mm, respectively. The diameter of the bolt hole is 14 mm and array sensing of film isCFRP 20 mm and 43.96ismm, respectively. The diameter of the bolt hole is 14 mm and the thickness thickness the of the the CFRP specimens specimens is 20 20 mm. mm. the thickness of the CFRP specimens is 20 mm. Figure 55 shows Figure shows the the detailed detailed information information regarding regarding the the experimental experimental setup. setup. A A sinusoidal sinusoidal signal signal Figure 5 shows the detailed information regarding thegenerated experimental setup. A sinusoidal signal with a frequency of 8 MHz and output voltage of 1.2V was by a signal generator, Rigol DG with a frequency of 8 MHz and output voltage of 1.2V was generated by a signal generator, Rigol DG with a frequency of 8 MHz and output voltage of 1.2V was generated by a signal generator, Rigol DG 3061A (RIGOL (RIGOL Technologies Technologies USA, OR, USA), USA), which which was was amplified by an 3061A USA, INC., INC., Beaverton, Beaverton, OR, amplified by an AG1020 AG1020 3061A (RIGOL Technologies USA, INC.,Inc., Beaverton, OR,MN, USA), which was amplified by an AG1020 RF amplifier (T&C Power Conversion, Rochester, USA) and then inputted to the RF amplifier (T&C Power Conversion, Inc., Rochester, MN, USA) and then inputted to the exciting exciting RF (T&Cvoltage Power Conversion, Inc.,was Rochester, MN, and then inputted to the exciting coil.amplifier The induced induced of sensing sensing coils coils measured byUSA) aa lock-in coil. The voltage of was measured by lock-in amplifier amplifier (SYSU (SYSU SCIENTIFIC SCIENTIFIC coil. The induced Guangzhou, voltage of sensing was measured lock-in voltage amplifier (SYSU SCIENTIFIC INSTRUMENTS, China),coils OE2041. The values valuesby ofainduced induced from different sensing INSTRUMENTS, Guangzhou, China), OE2041. The of voltage from different sensing INSTRUMENTS, Guangzhou, China), OE2041. The values of induced voltage from different sensing coils were obtained by the switch used to connect each coil. coils were obtained by the switch used to connect each coil. coils were obtained by the switch used to connect each coil.

Figure 5. Experimental setup. Figure 5. Experimental setup.

The experiments on unidirectional CFRP specimens made of carbon fiber T300 prepreg were The experiments on unidirectional CFRP specimens made of carbon fiber T300 prepreg were conducted. The length, width, and thickness of unidirectional CFRP specimens are 100 mm, 100 mm, conducted. The length, width, and thickness of unidirectional CFRP specimens are 100 mm, 100 mm, and 20 mm, respectively. In order to make artificial defects accurately, the joint was designed to and 20 mm, respectively. In order to make artificial defects accurately, the joint was designed to consist of four laminate plates, each having a 5 mm thickness. The stacking sequence of each laminate consist of four laminate plates, each having a 5 mm thickness. The stacking sequence of each laminate plate is [0/45/90/−45] . plate is [0/45/90/−45] .

Materials 2019, 12, 2785

6 of 12

The experiments on unidirectional CFRP specimens made of carbon fiber T300 prepreg were conducted. The length, width, and thickness of unidirectional CFRP specimens are 100 mm, 100 mm, and 20 mm, respectively. In order to make artificial defects accurately, the joint was designed to consist Materials 12, x FOR PEEReach REVIEW of 13 of four 2019, laminate plates, having a 5 mm thickness. The stacking sequence of each laminate6 plate is [0/45/90/−45]3s . During During the the experiment, experiment, artificial artificial cracks cracks (notches) (notches) with with different different lengths lengths were were introduced introduced at at the the hole edges of unidirectional CFRP specimens, as shown in Figure 6. The cracks were increased in hole edges of unidirectional CFRP specimens, as shown in Figure 6. The cracks were increasedthe in radial direction of the bolt hole by 1 mm each step, or in the axial direction by 5 mm each step. The the radial direction of the bolt hole by 1 mm each step, or in the axial direction by 5 mm each step. variation of induced voltages related to the cracks propagating along The variation of induced voltages related to the cracks propagating alongthe theaxial axialdirection directionand and radial radial direction were investigated. direction were investigated.

Figure 6. Schematic plot of the artificial crack propagating around radial and axial directions. Figure 6. Schematic plot of the artificial crack propagating around radial and axial directions.

3.2. Experimental Results 3.2. Experimental Results Figure 7 presents the changes of induced voltages of the sensing coils, caused by the damage at Figure 7 presents changesplate of induced voltages from of the0sensing byradial the damage at the hole edge in the topthe laminate when increased mm to 4coils, mm caused along the direction the hole edge in the top laminate plate when increased from 0 mm to 4 mm along the radial direction of the bolt hole of the joint. From the results, the induced voltage of sensing coil 1 at the top laminate of the raised bolt hole of thewhen joint.the From theoccurred results, the induced voltage plate. of sensing coil 1 at the top plate rapidly crack at the top laminate In the meantime, thelaminate induced plate raised rapidly when thesensing crack occurred at the laminate plate.plates In thewithout meantime, theremained induced voltages from the other three coils located at top the other laminate cracks voltages from the otherindicate three sensing located at the other laminate plateswhen without cracks unchanged. The results that the coils values of the induced voltages are higher the exciting remained The results indicate that the values the induced are higher when current in unchanged. the boundary trace of one sub-coil is in the same of direction as thatvoltages in the adjacent boundary the exciting current in the boundary trace of one sub-coil is in the same direction as that in the adjacent trace of another sub-coil compared to when the adjacent boundary trace currents move in opposite boundary another the sub-coil compared to when the adjacent boundary move in directions.trace This of is because intensity of current is increased when the currenttrace in thecurrents boundary traces opposite directions. is because the intensity of current is increased when the current in the of adjacent coils is inThis the same direction. boundary traces of adjacent coils is in same The changes of induced voltages the from the direction. sensing coils caused by the artificial crack at the hole changes inducedplate voltages the sensing caused the artificial at theofhole edgeThe in the secondoflaminate whenfrom increased from 0coils mm to 4 mmby along the radialcrack direction bolt edge in the second when increased from 0inmm to 47,mm along the radialfrom direction of hole are depicted in laminate Figure 8. plate Similar to the results shown Figure the induced voltage sensing bolt hole are depicted in Figure 8. Similar to the results shown in Figure 7, the induced voltage from coil 2 at the second laminate plate raised rapidly when the crack occurred in the second laminate sensing coil the 2 atinduced the second laminate plate raised the crack occurred in the second plate, while voltages from the other threerapidly sensingwhen coils located at the other laminate plates laminate plate, while the induced voltages from the other three sensing coils located at the other remained unchanged. laminate plates remained unchanged. Furthermore, Figures 7 and 8 illustrate that the change in the induced voltage of the sensing coil at Furthermore, Figures 7 andincreases 8 illustrate that the the crack change in the induced voltage of the sensing the damaged area continuously when propagates to about 3 mm. Passed that,coil the at the damaged area continuously increases when the crack propagates to about 3 mm. Passed that, induced voltage of the sensing coil remains unchanged, even as the crack propagates further. This is the induced of the sensing coil remains unchanged, even as the crack propagates further. because the voltage EC mainly distributes near the surface of the bolt hole, and the penetration depth ofThis the is the EC mainly distributes near the surface of the bolt hole, and the penetration depth of ECbecause is limited. the EC is limited.

Materials 2019, 12, x FOR PEER REVIEW

7 of 13

Materials Materials 2019, 2019, 12, 12, 2785 x FOR PEER REVIEW

of 12 77 of 13

(a)

(b)

Figure 7. Change of induced voltage caused by the crack in the top laminate plate along the

(a) (b) radial direction of bolt hole: (a) Exciting coils with the same current direction in boundary traces; Exciting withvoltage the opposite current direction in boundary traces. Figure(b) ofcoils induced voltage caused by crack in top the top laminate plate the Figure 7. Change Change of induced caused by thethe crack in the laminate plate along along the radial radial direction of bolt hole: (a) Exciting coils with the same current direction in boundary direction of bolt hole: (a) Exciting coils with the same current direction in boundary traces; (b) Exciting coils with oppositecoils current direction in boundary traces. traces; (b)the Exciting with the opposite current direction in boundary traces. 0.12

0.12

Coil 1 Coil 2 Coil 3 Coil4 1 Coil Coil 2 Coil 3 Coil 4

Change of induced voltage/V Change of induced voltage/V

0.10 0.12 0.09 0.11 0.08 0.10 0.07 0.09 0.06 0.08 0.05 0.07 0.04 0.06 0.03 0.05 0.02 0.04 0.01 0.03 0.00 0.02

0

1

2

3

Crack length/mm

0.01 0.00

0.10 0.12 0.09 0.11 0.08 0.10 0.07 0.09 0.06 0.08 0.05 0.07 0.04 0.06 0.03 0.05 0.02 0.04 0.01 0.03 0.00 0.02

4

5

Coil 2 Coil 3 Coil4 1 Coil Coil 2 Coil 3 Coil 4

1

(a)

2

3

Crack length/mm

0

Unidirectional CFRP

1

2

4

5

0

3

Crack length/mm

0.01 0.00

0

Unidirectional CFRP

Coil 1

0.11

Change of induced voltage/V Change of induced voltage/V

0.11

1

2

3

Crack length/mm

4

5

4

5

(b)

Figure of induced inducedvoltage voltage caused by crack the crack the second laminate plate Figure8.8.Change Change of caused by the in theinsecond laminate plate along thealong radial (a) (b) the radialof direction ofExciting bolt hole: coils with the in same current direction in direction bolt hole: (a) coils (a) withExciting the same current direction boundary traces; (b) Exciting coils with the opposite current direction in boundary traces. boundary traces; (b) Exciting coils with the opposite current direction boundary traces. Figure 8. Change of induced voltage caused by the crack in the secondinlaminate plate along

the radial direction of bolt hole: (a) Exciting coils with the same current direction in Figure presents voltages corresponding to the crack of 2 mm in theinradial Figure 99 presents thechanges changesof ofinduced induced voltages corresponding to the crack of 2 traces. mm the boundary traces; the (b) Exciting coils with the opposite current direction in boundary direction propagating in the axial direction along the bolt hole from 0 mm to 20 mm. The crack was radial direction propagating in the axial direction along the bolt hole from 0 mm to 20 mm. The crack increased from 0 mm tothe 20 mm with mm per step. obvious of induced induced voltage was increased 0 mm to changes 20 mm with 5 mm pervoltages step.ItItisis obviousthat that the the value of voltage Figure 9 from presents of5 induced corresponding to value the crack of 2 mm in the increases onlyin inpropagating thesensing sensingin coil ataxial thelocation locationwhere where thedamage damage appears, while the value of the increases only the coil at the while the value the radial direction the direction alongthe the bolt holeappears, from 0 mm to 20 mm. Theof crack remaining sensing coils only changes slightly or are unvaried. As mentioned previously, the response remaining sensing coils only slightly orper are step. unvaried. As mentioned the response was increased from 0 mm tochanges 20 mm with 5 mm It is obvious that thepreviously, value of induced voltage of the induced induced voltage more sensitive forthe the configuration configuration ofwhen when theexciting exciting current in the of the isismore sensitive for of the current increases only voltage in the sensing coil at the location where the damage appears, while the valuein ofthe the boundary trace of one exciting sub-coil is in the same direction as that in the adjacent boundary trace of boundary trace of one exciting sub-coil is in the same direction as that in the adjacent boundary trace remaining sensing coils only changes slightly or are unvaried. As mentioned previously, the response another exciting sub-coil than when the adjacent boundary trace currents point in opposite directions. ofof another exciting sub-coil thansensitive when the boundaryoftrace pointcurrent in opposite the induced voltage is more for adjacent the configuration whencurrents the exciting in the The damage destroys path sub-coil ofthe current flow CFRP, and theasintensity of the secondary magnetic directions. The damage destroys pathisof flow in CFRP, and the intensity of boundary the secondary boundary trace of onethe exciting incurrent theinsame direction that in the adjacent trace field generated by EC decreases. Consequently, there is an increment of total magnetic field based on magnetic field generated by ECthan decreases. there is an trace increment of total magnetic field of another exciting sub-coil when Consequently, the adjacent boundary currents point in opposite directions. The damage destroys the path of current flow in CFRP, and the intensity of the secondary magnetic field generated by EC decreases. Consequently, there is an increment of total magnetic field

Materials 2019, 12, x FOR PEER REVIEW Materials 2019, 12, 2785

8 of 13 8 of 12

based on subtraction of the original magnetic field and the secondary magnetic field. This is the reason why voltage increases the crack magnetic occurs in field. both axial and directions. subtraction ofthe the induced original magnetic field andwhen the secondary This is theradial reason why the induced voltage increases when the crack occurs in both axial and radial directions. 0.16

0.16

Coil 2 Coil 3

0.12

Coil 4

0.10 0.08 0.06 0.04 0.02 0.00

Coil 2

0.14

Change of induced voltage/V

Change of induced voltage/V

0.14

Unidirectional CFRP

Coil 1

Coil 1

Coil 3 Coil 4

0.12 0.10 0.08 0.06 0.04 0.02

0

5

10

Crack depth/mm

(a)

15

20

0.00

0

5

10

Crack depth/mm

15

20

(b)

Figure9. 9.Change Change of induced voltage ofgrowth crack in growth indirection the axial direction theWith bolt Figure of induced voltage of crack the axial along the boltalong hole: (a) hole: (a) With exciting current in the same direction in the boundary between exciting subexciting current in the same direction in the boundary between exciting sub-coils; (b) With exciting coils; in (b)the With exciting current in boundary the opposite direction insub-coils. the boundary between exciting current opposite direction in the between exciting

sub-coils. 4. Numerical Simulation 4. Numerical Simulation Preliminary numerical simulation was conducted to verify and explain the phenomena observed in the Preliminary experiments.numerical Two types of artificial cracks around hole, propagating along the axial simulation was conducted to bolt verify andone explain the phenomena observed direction and the otherTwo propagating along thecracks radial around direction arehole, introduced in the CFRPalong specimen to in the experiments. types of artificial bolt one propagating the axial validate the change of the induced voltage from the EC array sensing coils. direction and the other propagating along the radial direction are introduced in the CFRP specimen

to validate the change of the induced voltage from the EC array sensing coils. 4.1. Numerical Model three-dimensional simulation model based on the finite element method (FEM) has been 4.1.A Numerical Model developed to determine the changes of induced voltages of EC sensing coils resulting from the cracks A three-dimensional simulation model based on the finite element method (FEM) has been propagating around the hole-edge in the composite bolted joints. The configuration of the FEM model developed to determine the changes of induced voltages of EC sensing coils resulting from the cracks is shown in Figure 10. The dimensions of CFRP composite specimen are 30 mm × 30 mm × 20 mm propagating around the hole-edge in the composite bolted joints. The configuration of the FEM model with one-quarter of a hole having a diameter of 7 mm. In order to facilitate calculation and save time, is shown in Figure 10. The dimensions of CFRP composite specimen are 30 mm × 30 mm × 20 mm the model used only contains one-quarter of a circular hole because the distribution of EC around the with one-quarter of a hole having a diameter of 7 mm. In order to facilitate calculation and save time, hole is symmetrical. The layup of CFRP composite is [0/45/90/−45]12s All adjacent plies in the model the model used only contains one-quarter of a circular hole because the distribution of EC around the are assumed to be in contact with each other. hole is symmetrical. The layup of CFRP composite is [0/45/90/−45] All adjacent plies in the As shown in Figure 10, the inner traces are four sensing coils and the outer traces are the exciting model are assumed to be in contact with each other. coil with four exciting sub-coils. The flow path of the current for the nearby traces of the exciting As shown in Figure 10, the inner traces are four sensing coils and the outer traces are the exciting coil at the boundary of two adjacent coils is set in the same direction in the simulation. Alternating coil with four exciting sub-coils. The flow path of the current for the nearby traces of the exciting coil current with a frequency of 8 MHz and 2 V is applied to the exciting coil. The mesh of the FEM at the boundary of two adjacent coils is set in the same direction in the simulation. Alternating current model is generated with tetrahedral elements. The conductivity in the fiber direction is 29300 S/m, with a frequency of 8 MHz and 2 V is applied to the exciting coil. The mesh of the FEM model is the conductivity perpendicular to fiber is 77.8 S/m, and the conductivity in the thickness direction is generated with tetrahedral elements. The conductivity in the fiber direction is 29300 S/m, the 0.000794 S/m. The crack was set to increase 2.5 mm each step in the axial direction, and 0.5 mm each conductivity perpendicular to fiber is 77.8 S/m, and the conductivity in the thickness direction is step in the radial direction. 0.000794 S/m. The crack was set to increase 2.5 mm each step in the axial direction, and 0.5 mm each step in the radial direction.

Materials Materials2019, 2019,12, 12,2785 x FOR PEER REVIEW Materials 2019, 12, x FOR PEER REVIEW

99 of of12 13 9 of 13

Figure 10. Finite element method (FEM) simulation model. 10. Finite element method (FEM) simulationmodel. model. FigureFigure 10. Finite element method (FEM) simulation

4.2.Simulation SimulationResults Results 4.2.4.2. Simulation Results TheEC ECdistributions distributions on first the first plies four [0/45/90/−45] plies [0/45/90/-45] from the top composite of CFRP composite The on the from the top CFRP laminates The EC distributions on the firstfour four plies [0/45/90/-45] from theoftop of CFRP composite laminates with and without a crack is shown in Figures 11 and 12, respectively. Under the simulation with and without a crack is shown Figures 11 and 12, respectively. Under the simulation condition, laminates with and without a crack is in shown in Figures 11 and 12, respectively. Under the simulation condition, the distribution of each the EC for eachthe ply is not the same because of the effect of the fiber the distribution of the EC for ply is not same because of the effect of the fiber orientation. condition, the distribution of the EC for each ply is not the same because of the effect of the fiber orientation. For the plies with fiber of 0° and 90°, the of EC is in For the plies fiber direction 0◦direction and 90of◦ ,0° the distribution of distribution EC is concentrated inconcentrated the position orientation. Forwith the plies with fiberofdirection and 90°, the distribution of EC is concentrated in of the position of the opposite sides of the hole edge. The EC flows in the opposite direction when opposite of the hole edge.ofThe flows in the direction located on the when ply with thethe position ofsides the opposite sides theEC hole edge. Theopposite EC flows in the when opposite direction ◦ . The ◦ is fiber located on the ply with fiberEC orientation inof 45°. The ECfiber repartition ofof ply with direction of −45° fiber orientation in 45 repartition ply with direction −45 mainly concentrated located on the ply with fiber orientation in 45°. The EC repartition of ply with fiber direction of −45° mainly concentrated athole the edge. centralWhen area of hole edge. Whenatthe crack set at hole the central atisthe central area of at the thethe artificial theartificial central area edge, is mainly concentrated the central area of the hole edge.crack Whenset the artificial crack setofatthe the central ◦ area of the hole edge, the EC distribution of ply with fiber direction of −45° changed dramatically due the EC distribution of ply with fiber direction of −45 changed dramatically due to the current density area of the hole edge, the EC distribution of ply with fiber direction of −45° changed dramatically due to the current density decreasing. Thisofindicates that the effect ofby magnetic fieldlower generated by EC indicates that theThis effect magnetic generated EC becomes leading to decreasing. the current This density decreasing. indicates that field the effect of magnetic field generated by ECthe becomes lower leading the induced voltage of the sensing coil to increase. induced voltage of the sensing coilvoltage to increase. becomes lower leading the induced of the sensing coil to increase.

Figure 11. Eddy current distributions on the first four plies [0/45/90/−45] from the top of CFRP Figure 11.without Eddy current distributions on the first four plies [0/45/90/-45] from the top of CFRP composite damage. Figure 11. Eddy current distributions on the first four plies [0/45/90/-45] from the top of CFRP composite without damage. composite without damage.

Materials 2019, 12, 2785 Materials 2019, 12, x FOR PEER REVIEW

10 of 12

10 of 13

Figure 12. Eddy current distributions on the first four plies [0/45/90/−45] from the top of CFRP composite Figure 12. Eddy current distributions on the first four plies [0/45/90/-45] from the top of CFRP with an artificial crack. composite with an artificial crack.

Figure 13 shows the simulation results of the induced voltage changes of the sensing coils, which Figure 13 shows the simulation results of the induced voltage changes of the sensing coils, which is caused by the cracks at the hole edge in the first and second laminate plates when it was propagated is caused by the cracks at the hole edge in the first and second laminate plates when it was propagated from 0 mm to 6 mm along the radial direction of the bolt hole of the unidirectional CFRP. Agreeing from 0 mm to 6 mm along the radial direction of the bolt hole of the unidirectional CFRP. Agreeing with experimental results, the variation in the induced voltage of sensing coil 1 at the top laminate with experimental results, the variation in the induced voltage of sensing coil 1 at the top laminate plate raised rapidly when the crack plate, while whilethe theinduced inducedvoltages voltages plate raised rapidly when the crackoccurred occurredatatthe thetop top laminate laminate plate, ofof thethe other three sensing coils located at the other laminate plates without cracks remained unchanged. other three sensing coils located at the other laminate plates without cracks remained unchanged. When thethe crack occurred at at the second of induced inducedvoltage voltageininsensing sensing coil When crack occurred the secondlaminate laminateplate, plate,the the variation variation of coil 2 raised rapidly. Figure 14 shows the simulation results of the induced voltage changes corresponding 2 raised rapidly. Figure 14 shows the simulation results of the induced voltage changes to the crack with 2to mm the radial in the axial direction along bolt hole from corresponding thein crack with 2direction mm in thepropagating radial direction propagating in the axialthe direction along 0 mm 20hole mmfrom of the0 unidirectional CFRP. Matching well withMatching the experimental the changes the to bolt mm to 20 mm of the unidirectional CFRP. well withresults, the experimental Materials 2019, 12, x FOR PEER REVIEW 11 of 13 in induced voltage increases onlyvoltage when increases the damage in damage the location of the corresponding results, the changes in induced onlyappears when the appears in the location of the corresponding coil, whereas thethe induced voltage of theonly otherchanges sensingslightly coils only sensing coil, whereas sensing the induced voltage of other sensing coils or changes unvaried. slightly or unvaried. 4.5

3.5

Coil 1 Coil 2

3.0

Coil 3 Coil 4

Coil 1 4.0

Change of induced voltage/mV

Change of induced voltage/mV

4.0

2.5 2.0 1.5 1.0 0.5 0.0

Coil 2 Coil 3

3.5

Coil 4

3.0 2.5 2.0 1.5 1.0 0.5

0

1

2

3

4

Crack length/mm

(a)

5

6

0.0

0

1

2

3

4

5

6

Crack length/mm

(b)

Figure 13. Simulation of induced voltage changecaused caused by in two different laminate plates: plates: Figure 13. Simulation of induced voltage change bythe thecrack crack in two different laminate (a) The top laminate plate; (b) The second laminate plate. (a) The top laminate plate; (b) The second laminate plate.

3.5

Coil 1 3.0

Coil 2 Coil 3

(a)

(b)

Figure 13. Simulation of induced voltage change caused by the crack in two different laminate plates: (a) The top laminate plate; (b) The second laminate plate.

Materials 2019, 12, 2785

11 of 12

3.5

Coil 1 Coil 2 Coil 3 Coil 4

Change of induced voltage/mV

3.0 2.5 2.0 1.5 1.0 0.5 0.0

0

5

10

15

20

Crack depth/mm

Figure 14. Simulation of the induced voltage change caused by the crack propagating in the axial direction along the bolt hole of the unidirectional CFRP. Figure 14. Simulation of the induced voltage change caused by the crack propagating in the axial

5. Conclusions direction along the bolt hole of the unidirectional CFRP. A novel built-in distributed EC sensor network has been developed to monitor the hole-edge 5. Conclusions damages of composite bolted joints. The built-in distributed EC sensor network contains one exciting A novel built-in distributed EC sensor network has been developed to monitor the hole-edge coil covering the entire area of the inner wall and four separate sensing coils laying along the axial damages of composite bolted joints. The built-in distributed EC sensor network contains one exciting directioncoil of the bolt hole. Botharea experiments simulations conducted to investigate covering the entire of the inner and wall numerical and four separate sensing were coils laying along the axial the performance ofthe the EChole. array filmsand tonumerical quantitatively track theconducted hole-edge damage. Based direction of bolt Bothsensing experiments simulations were to investigate the performance the EC arrayresults, sensing it films to quantitatively track the hole-edge damage. Based of the on the experimental andofsimulation is obviously that the changes in induced voltage on the experimental and simulation results, it is obviously that the changes in induced voltage the 0 mm sensing coil rises quickly when the hole edge crack propagates in the radial direction of bolt of from sensing coil rises quickly when the hole edge crack propagates in the radial direction of bolt from 0 to 3 mm at the laminate plate where the sensing coil is located, whereas the induced voltages of the mm to 3 mm at the laminate plate where the sensing coil is located, whereas the induced voltages of other sensing coils located at the other laminate plates without cracks only change slightly or remain unchanged. The configuration in which the exciting current in the boundary trace of one sub-coil is in the same direction as that in the adjacent boundary trace of another sub-coil is more beneficial because it can improve the intensity of the induced voltage better than when the adjacent boundary trace currents are in opposite directions. The results demonstrated that the developed EC array sensing film can effectively identify not only whether there is damage at the hole edge, but also the damage location in the thickness and quantitative size. The proposed EC array sensing film shows a good promise for real-time monitoring the crack in the composite bolted joint with several benefits, such as lightweight and low effect on original structural integrity. However, the ability of EC array sensing film to quantitatively track the damage in the radial direction needs to be further improved. Additionally, other improvements include how to apply this SHM technology into real engineering applications. Author Contributions: Conceptualization, Q.L. and X.Q.; methodology, Q.L. and H.S.; software, T.W.; formal analysis, Q.L. and H.S.; writing—original draft preparation, Q.L.; writing—review and editing, Q.L. and X.Q.; supervision, X.Q. Funding: This research was funded by the National Natural Science Foundation of China (Grant No. 11472308, 11902280) and the Natural Science Foundation of Fujian Province (Grants No. 2018J05094). Acknowledgments: The authors would like to thank Yishou Wang, Jianbo Shao, and other colleagues at Xiamen University for their technical assistance. Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2.

Marsh, G. Airbus A350 XWB update. Reinf. Plast. 2010, 54, 20–24. [CrossRef] Thoppul, S.D.; Finegan, J.; Gibson, R.F. Mechanics of mechanically fastened joints in polymer-matrix composite structures—A review. Compos. Sci. Technol. 2009, 69, 301–329. [CrossRef]

Materials 2019, 12, 2785

3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

25.

12 of 12

Sun, H.T.; Chang, F.K.; Qing, X. The response of composite joints with bolt-clamping loads, Part I: Model development. J. Compos. Mater. 2002, 36, 47–67. [CrossRef] Chowdhury, N.; Chiu, W.K.; Wang, J. Review on the fatigue of composite hybrid joints used in aircraft structures. Adv. Mater. Res. 2014, 891, 1591–1596. [CrossRef] Qing, X.; Li, W.; Wang, Y.; Sun, H. Piezoelectric transducer-based structural health monitoring for aircraft applications. Sensors 2019, 19, 545. [CrossRef] [PubMed] Staszewski, W.J.; Mahzan, S.; Traynor, R. Health monitoring of aerospace composite structures—Active and passive approach. Compos. Sci. Technol. 2009, 69, 1678–1685. [CrossRef] Qing, X.; Beard, S.; Kumar, A.; Chan, H.; Ikegami, R. Advances in the development of built-in diagnostic system for filament wound composite structures. Compos. Sci. Technol. 2006, 66, 1694–1702. [CrossRef] Farrar, C.; Park, G.; Sohn, H.; Inman, D.J. Overview of piezoelectric impedance-based health monitoring and path forward. Shock Vib. Dig. 2003, 35, 451–464. Su, Z.; Ye, L.; Lu, Y. Guided Lamb waves for identification of damage in composite structures: A review. J. Sound Vib. 2006, 295, 753–780. [CrossRef] Giurgiutiu, V.; Zagrai, A.; Bao, J. Piezoelectric wafer embedded active sensors for aging aircraft structural health monitoring. Struct. Health Monit. 2002, 1, 41–61. [CrossRef] Qing, X.P.; Beard, S.J.; Kumar, A.; Ooi, T.K.; Chang, F.K. Built-in sensor network for structural health monitoring of composite structure. J. Intell. Mater. Syst. Struct. 2006, 18, 39–49. [CrossRef] Qiu, L.; Yuan, S.; Boller, C. An adaptive guided wave-Gaussian mixture model for damage monitoring under time-varying conditions: Validation in a full-scale aircraft fatigue test. Struct. Health Monit. 2017, 16, 501–517. [CrossRef] Qiu, L.; Liu, M.; Qing, X.; Yuan, S. A quantitative multidamage monitoring method for large-scale complex composite. Struct. Health Monit. 2013, 12, 183–196. [CrossRef] Wang, Y.; Qing, X. Progress on study of structural health monitoring of composite joints. Acta Mater. Compos. Sin. 2016, 33, 1–16. Jalalpour, M.; El-Osery, A.I.; Austin, E.M.; Reda Taha, M.M. Health monitoring of 90◦ bolted joints using fuzzy pattern recognition of ultrasonic signals. Smart Mater. Struct. 2014, 23, 015017. [CrossRef] Rakow, A.; Chang, F.K. A structural health monitoring fastener for tracking fatigue crack growth in bolted metallic joints. Struct. Health Monit. 2011, 11, 253–267. [CrossRef] Sun, H.; Wang, T.; Liu, Q.; Qing, X. A novel eddy current array sensing film for quantitatively monitoring hole-edge crack growth in bolted joints. Smart Mater. Struct. 2019, 28, 015018. [CrossRef] Sun, H.; Wang, T.; Liu, Q.; Wang, Y.; Qing, X. A two-dimensional eddy current array-based sensing film for estimating failure modes and tracking damage growth of bolted joints. Struct. Health Monit. 2019. [CrossRef] Dang Hoang, T.; Herbelot, C.; Imad, A. Rupture and damage mechanism analysis of a bolted assembly using coupling techniques between A.E. and D.I.C. Eng. Struct. 2010, 32, 2793–2803. [CrossRef] Zhang, Z.; Xu, H.; Liao, Y.; Su, Z.; Xiao, Y. Vibro-acoustic modulation (VAM)-inspired structural integrity monitoring and its applications to bolted composite joints. Compos. Struct. 2017, 176, 505–515. [CrossRef] Fan, X.; Li, J.; Hao, H. Piezoelectric impedance based damage detection in truss bridges based on time frequency ARMA model. Smart Struct. Syst. 2016, 18, 501–523. [CrossRef] Wheatley, G.; Kollgaard, J.; Register, J.; Zaidi, M. Comparative vacuum monitoring (CVM™) as an alternate means of compliance (AMOC). Insight-Non-Destr. Test. Cond. Monit. 2005, 47, 153–156. [CrossRef] Sun, Y.; Wang, W.; Li, B.; Liu, M. Investigation on behavior of crack penetration/deflection at interfaces in intelligent coating system. Appl. Math. Mech. 2015, 36, 465–474. [CrossRef] Wang, Y.; Qing, X.; Dong, L.; Banerjee, S. Multi-field coupled sensing network for health monitoring of composite bolted joint. In Proceedings of the Health Monitoring of Structural and Biological Systems, Las Vegas, NV, USA, 1 April 2016; Volume 9805. Grimberg, R.; Savin, A.; Steigmann, R.; Bruma, A. Eddy current examination of carbon fibres in carbon-epoxy composites and Kevlar. Int. J. Mater. Prod. Technol. 2006, 27, 221–228. [CrossRef] © 2019 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/).