You are on page 1of 8


12, DECEMBER 2011 3261

Surface Crack Detection for Carbon Fiber Reinforced
Plastic (CFRP) Materials Using Pulsed Eddy
Current Thermography
Liang Cheng and Gui Yun Tian, Senior Member, IEEE

Abstract—There is currently a requirement in many industries the need for a couplant required for introducing acoustic waves,
to inspect carbon fiber reinforced plastic (CFRP) components, lack of sensitivity to shallow surface breaking defects, and large
such as those used in aircraft and for wind turbine blades to iden- attenuation of acoustic waves when propagating through the
tify issues leading to potential failures. To detect surface cracks,
multilayered structure of composites. Other NDT&E techniques
pulsed eddy current (PEC) thermography is proposed as a pow-
erful inspection technique, allowing the operator to observe the have also been applied to composite material inspection, such
heating developed from the eddy current distribution in a struc- as X-ray [3], acoustic emission [4], [5], eddy current [6], [7],
ture using infrared imaging, detecting defects over a relatively and microwave [8], [9]. Different NDT&E techniques have dif-
wide area within a short time (of the order of milliseconds). In ferent characteristics, but the majority of methods have limi-
this paper, a PEC thermography inspection system for CFRP ma- tations with regard to large-scale sensing, imaging, and com-
terials is studied and optimized. Using the system, the directional prehensive measurement or have safety issues. Consequently,
electrical conductivity of the CFRP material is observed through
the surface heating pattern. Then, the normalized temperature the integration of NDT techniques to achieve improved perfor-
rise and decay are investigated through the inspection of notches mances has been implemented.
with varied depths and widths. The position invariance of the coil The major advantage of thermography over other techniques
with respect to the notch along the fiber direction is also studied in is the potential for rapid inspection of a large area within a
the experiments. The work shows that PEC thermography can be short time, though currently it is mostly applied to samples
used for defect detection and characterization through analysis of in the lab instead of in situ structures [10]. However, there is
the surface heating pattern and the transient temperature change.
a tradeoff between detectable defect size and inspection area.
Index Terms—Carbon fiber reinforced plastic (CFRP), nonde- Thermography is also applicable to a wide range of materials,
structive testing and evaluation (NDT&E), pulsed eddy current including glass fiber, carbon fiber composites, and metallic
(PEC) thermography, surface defects.
materials, where specific excitation techniques are suitable for
different applications. To inspect defects over a large scale and
I. INTRODUCTION at large standoff distances, integration of thermography and
other NDE approaches have been investigated [11]–[15], e.g.,
N RECENT decades, there has been an increasing interest
I in the use of composite materials, particularly carbon fiber
reinforced plastic (CFRP), in the aerospace and renewable en-
flash thermography [16]–[18], vibrothermography, sonic ther-
mography [19], laser thermography [2], optical thermography
[20], and pulsed eddy current (PEC) thermography or induc-
ergy industries, because of the low weight and improved me- tion thermography [21]. Among the thermography techniques
chanical properties compared with metals. Components made mentioned previously, PEC thermography, combining PEC and
from CFRP, such as wind turbine blades and aircraft fuselage, thermography, has its own advantages. For composite materials,
have to be tested for quality evaluation after manufacturing and varied excitation direction can be used in PEC thermography
monitored during in-service operation to increase the compo- to investigate different layers, since the electrical and thermal
nent lifetime. To accomplish this, nondestructive testing and conductivity is the greatest along the fiber orientation. In addi-
evaluation (NDT&E) techniques are used. tion, the application of heat is not limited to the sample surface,
Ultrasonic testing is one of the most widely used methods such as in the flash thermography; rather, it can reach a certain
for composite material inspection [1], [2] and has the advantage depth, which is governed by the skin depth or penetration depth
that it can detect defects in the interior of material. However, formula
this method suffers from a number of disadvantages, including
Manuscript received March 01, 2011; revised April 27, 2011; accepted May
07, 2011. Date of publication May 23, 2011; date of current version November where is excitation frequency, is the electrical conductivity,
02, 2011. This work was supported by the Engineering and Physical Sciences and is the permeability of the material under inspection. The
Research Council (EPSRC), U.K., under Grant EP/F06151X/1. The PEC transient of electromagnetic (EM) distribution and heat diffu-
thermography system was developed through the RCNDE (Research Centre
of NDE) supported by EPSRC in collaboration with Rolls Royce, Alstom sion can derive a depth profile of defects [22], compared with
and University of Bath. The associate editor coordinating the review of this surface-heating techniques.
manuscript and approving it for publication was Prof. Okyay Kaynak. PEC thermography involves the application of high-current
The authors are with the School of Electrical, Electronic and Computer Engi-
neering, University of Newcastle, NE1 7RU Newcastle upon Tyne, U.K. (e-mail:
EM pulse to the conductive material under inspection for a short; period (typically less than 1 s). When the eddy currents en-
Digital Object Identifier 10.1109/JSEN.2011.2157492 counter a discontinuity (e.g., notches or delaminations), they are
1530-437X/$26.00 © 2011 IEEE

Taking CFRP material as there is no need to model over the whole length. Ramdane et al.3262 IEEE SENSORS JOURNAL. . only one edge of the rectangular coil is tigated. simulation studies on how notch depth and width impacts on and S/m [27]. The results showed that delaminations in rectangular coil is used in experiments. where higher levels of Joule (ohmic) heating is achieved [23]. simulated. as shown in Fig. 1. A conductive ma- terial is heated by Joule heating. The thermal diffusivity of the CFRP as mea- the transient temperature change are described in Section III. Thus. [28]. 1. and cross-ply directions Experiments on directional conductivity inspection. the penetration depth is much larger than that of a notch in the sample with varied depth or width is simulated. This thermal properties. while and higher temperatures at the crack bottom in nonmagnetic ma. . After the period of eddy current heating. and is governed by (2) where is the electrical conductivity. The “notch” in this paper is defined as (4) a surface crack over the full width of the sample. The specific sition invariance are reported in Section IV. DECEMBER 2011 II. However. Thus. as metallic material. heat is set at 850 J/kg K from the COMSOL Multiphysics . eddy The heat conduction equation of a specimen caused by a Joule current diversion results in increased eddy current density at the heating source is governed by notch bottom. The density of CFRP is set as 1540 kg/m . or 2 mm) bottom of the notch will be investigated because it is higher than and depth (0.1. Section II describes the analyt. leading to regions of increased and decreased eddy current density. 1. 11. but finite in depth and width. the length of the sample is reduced for faster calculation since terials with a large penetration depth. In this paper. transversal and cross-ply directions are set at . via numerical simulations and experiments. They vary in different directions due to the paper is organized as follows. in this study the temperature at the shown in Fig. the electric field intensity vector and the eddy current density can be derived via (1) and forced to divert. the eddy CFRP could be detected using PEC thermography. According to Maxwell’s equations. the dinal. The results showed the crack can the experiments. 0. Therefore.5. Subsequently. The excitation that at the edge of the notch. the mixed phenomena ductivity. the defect can be identified from a charac. VOL. However. 12. . Oswald-Tranta and Wally [24] investigated the temperature distribution around a crack with In this study. A posite materials [26]. CFRP samples (dimensions 100 mm 38 different penetration depths using FEM modeling and experi. the thickness and width of be identified by lower temperatures at surface edges of the crack the sample are set the same as the experimental sample. the notch also affects the heat where . the impact are . currents are induced dominantly by the coil edge which is close the other types of defect in composite material were not inves. 2-D geometry setup for PEC thermography simulation. transversal. carried out 3-D nu. the magnetic vector potential can be calculated from (3) Fig. Both carbon fiber and the from surface defect detection in metallic components to CFRP matrix (normally epoxy resin) contribute the electrical and components. PEC thermography is proposed and extended CFRP is an anisotropic material. Simulation Setup in previous studies [23]–[25]. sured in the longitudinal. which is introduced in Section IV. drawn as a cylindrical wire in Fig. m s. (5) teristic heat distribution in the thermal image/video. the experimental setup. The relationship between . an analytical model is established for the eddy current (EM) and heat diffusion phenomena. THEORETICAL CONSIDERATIONS In this paper. and thermal con- diffusion in the cooling phase. . 1. and are density. The electrical conductivities in the longitu- ical model for eddy current and heat diffusion mechanisms. and notch po. at 20 C of notch depth on the transient temperature rise. or 2 mm) are simulated. The generated resistive heat is proportional to the square of the eddy current density or electric field intensity . For the notch case. of induction heating dominating the heating phase and heat dif- fusion dominating the cooling phase and their specific behavior is used for the quantitative NDE (QNDE) of defects. 1. which is caused by resistive heating from the eddy currents. Therefore. frequency and current are set as 256 kHz and 380 to match merical simulations and experiments on both metallic and com. NO. NUMERICAL SIMULATION STUDIES PEC thermography has been used to mainly inspect metals A. In the simulation. 1. III. mm 6 mm) provided by Exel Composites UK are used in ments with metallic materials. fiber orientation. respectively. Notches with width (0. heat capacity. respectively. The single an example. to the sample.5.

This effect is illustrated to an area of increased eddy current density and a resultant hot in the simulation results shown in Fig. 3. 2. changes in coil orientation will cause increased current flow in different layers. Since con. variation of the applied EM a finite body.1 mm for steel). the skin depth is normally larger than where is the thermal diffusivity and is the notch depth. the notch at the maximum PEC-induced heating time of 200 ms. the skin depth is normally smaller than the notch depth.95 mm through the sample thickness) is much larger than that of metal (normally less than 0. Hence. as defined by the skin depth For cases where skin depth is larger than sample thickness formula (1). tion between the notch and uniform eddy currents in the sample Thus. d = 2 mm notch. where the temperature spot at the bottom of the notch. for the heating phase ductivity is greatest in the fiber direction and fibers are orien. 2 shows the temperature distribution in the area around 1 mm: (a) normalized responses and (b) nonnormalized (raw) responses. a such as a notch. These results agree with experimental results discussed University of Bath [29]. 1 at notch bottom for varied notch depth at notch width w = Fig. The schematic representation of rise at the 4-mm-deep notch is larger than that at the shallower eddy current behavior around different types of defect is also notch. for (7) A smaller skin depth causes the eddy currents to be concentrated at the surface and much denser at the surface edge of the notch. [25]. The interac. A large skin depth results in eddy currents at the distance from the evaluated point to the rear side of the sample. mainly confined to the surface layer. The geometry of sample with notch defects in the numerical simulation is the same as that in the experiment. COMSOL is used to perform finite-element method (FEM)-based simulation via the AC/DC module. More heating is observed at the bottom of the notch due to the increased eddy current density in that area caused by eddy bottom of the notch being denser than that at the surface edge current diversion around the bottom of the defect. the skin depth at 256 kHz (around 9. When a discontinuity. To solve the EM and heat diffusion problem. 1. . along the sample thickness. The investigated point in Fig. interrupts in the eddy current path. they will deeper notch causes more eddy current diversion at the bottom be forced to divert around the bottom of the notch. Derived from (2) and (5) using Green’s function solution for ferent directions. they will be ulation studies. Conversely. As a consequence. [32]: tated in different directions in different layers. as shown in the sim- resistance. the following two equations. 3. the temperature rise at the bottom of the notch is higher shows that the eddy currents will always follow the path of least than the other regions in the CFRP sample. at the excitation frequency used in simulation and experiment. Therefore. Fig. Simulation results for transient temperature against time at “pos 1” B. respectively. As the electric and thermal conductivity of CFRP vary in dif.CHENG AND TIAN: SURFACE CRACK DETECTION FOR CFRP MATERIALS USING PEC THERMOGRAPHY 3263 Fig. [30] and for the cooling phase [31]. for CFRP. in a sample without a defect. Simulation results of thermal images at the maximum heating (200 ms) for w =1 mm. For metallic ma- terials. the estimation of temperature against time for an field orientation through changes in the excitation coil direction infinite length and finite thickness plate can be expressed with can be used for the investigation of different layers. simulation package (COMSOL for short) library for granite (granite: carbon content in excess of 99%). 3 is marked as “pos1” in discussed in previous work by Newcastle University and the Fig. because the notch blocks the surface current flow. for (6) Besides. which leads of the notch than the shallower notch. Influence of Notch Depth shown in Fig. of the notch. in the Section IV. the electric conductivity of CFRP is much smaller than that of metal.

the coefficients and can also be used for width in- around a narrower area at the notch bottom. it is clearer to investigate after the normalization of the tempera- ture curve. when the notch width is fixed. shown in Fig.3264 IEEE SENSORS JOURNAL. while is the normalized temperature decay rate. after the nor- malization of the transient temperature change with time for the same width and varied depth notches. which matches (7). is the duration of the heating pulse. Therefore. a deeper notch leads to the higher temperature rise in the heating phase at the bottom of the notch (coefficient ) and a faster tem. determined by local thermal property changes and notch dimensions etc. 3(b). From (6) and (7). 6. 3(a). NO. it is As the simulation results presented in the previous section clearer to investigate after the normalization of the temperature closely agree with the experimental results. It implies = d 2 mm: (a) normalized responses and (b) nonnomalized (raw) responses. VOL. For a defect-free sample. as C. DECEMBER 2011 Fig. varied notch widths are shown in Fig. it can be seen that deeper notch has a greater Fig. we can use coefficients and to describe the thermal response. In conclusion. a convergence of the tran- sient temperature change is obtained for the heating phase. as shown in Fig. Induction Heating System amplitude of eddy current density (as notch depth influence in The experimental setup is illustrated in Fig. 1 at notch bottom for varied notch width at notch depth efficient and notch depth is illustrated in Fig. The results indicate the deeper notch has a faster temperature decay in the cooling phase. Simulation results for transient temperature against time at “pos temperature rise (coefficient ). a deeper notch leads to larger temperature rise in the heating phase. Amplitude of temperature rise (a) versus notch depth d. for a sample with a notch. The transient temperature changes against time for as larger results in smaller value. shown in Fig. is the thickness of the sample. Although depth information can be derived from transient bottom of the notch due to the notch geometry change. perature decay in the cooling phase (coefficient ). Unlike the notch depth in- fluence. a narrower vestigation. From Fig. As for coefficient . where the figure indicates that the maximum amplitude of temperature change IV. 12. An Easyheat the Section III-B). 5(b)]. EXPERIMENTS increases as becomes smaller. 3(a). As for coefficient . Influence of Notch Width well as a greater rate of change in temperature in the early stages of the heating phase [seen in Fig. as larger (smaller . This im- temperature change with time in both the heating and cooling plies that the narrower notch will force eddy currents to divert phases. 4. 1 for the estimation in this study. it is feasible to use curve. 5(a). 5(b). However. but also the eddy current distribution at the 224 from Cheltenham Induction Heating is used for coil excita- . Thus. Experimental setup. It shows that notch depth does not affect the normalized transient heating behavior. 6. 5. that a deeper notch will interact with and change the course of more eddy currents. 5(a)]. the eddy current density at the bottom of the notch). 4. where ) results in larger value. The results also indicate that the nar- simulations to predict the impact of notch width on thermal rower notch has faster temperature decay in the cooling phase. notch leads to a greater temperature rise [seen in Fig. as shown in Fig. Fig. responses. 11. Coefficient is the amplitude of temper- ature rise determined by local electric conductivity change (in this case. The relationship between co. the variation of notch width not only changes the A. 1” shown in Fig.

high excitation frequencies will lead zontal [Fig. fixed in the vertical direction in the following experiments. horizontal direction. 7(a) and (b) shows the thermal image in terms of digital The SC7500 is a Stirling cooled camera with a 320 256 level (DL) after 2 s of heating using horizontal and vertical coil array of 1. at maximum heating. example the thermal image at the maximum heating time for a pact on experimental results. From the compar- 383 Hz. are investigated. The Easyheat has a maximum excitation power of 2. hence. it can be ascertained the fiber orien- The rectangular coil is constructed from a 6. shown in Fig. However. tion. before inspecting the sample 2-mm-deep notch is shown in Fig.6 ms. As a reference. the temperature changes has a sensitivity of 20 mK and a maximum full frame rate of at the coil in Fig. 6. normalization of transient temperature change is applied. which was verified direction. the conductivity of Exel sample in formation. In this study. The notches are manufactured in the workshop at New- using division of the transient temperature by the temperature castle University. Water cooling of coil is implemented to coun. horizontal direction. in real applications. the directional (horizontal and vertical) conductivity The transient temperature change at the same point at the is first ascertained. shown in Fig. Images were acquired for a total of 500 ms (200 ms heating followed by 300-ms cooling) at the maximum frame B. three notches with a width of 1 mm. This allows optimization of the direction of bottom of the notch and close to the coil. coil orientation has a large im. A 200-ms heating duration is selected for inspection. 7(a) and (b) are similar. For vertical coil the heating period to full power) of 5 ms. Directional Conductivity Experiment data acquisition rate of 383 fps. for defects.5–5 m InSb detectors. varying in depth from 0.35-mm high-con. 8. 7(b)]. 1. horizontal coil orientation. 7. the eddy currents are following in a main information will allow the derivation of defect profile in. marked as “pos 1” . 7(a)] and vertical [Fig. with the option to increase the frame rate with increase in temperature at the sample surface with the coil ori- windowing of the image. Thus. range of 150–400 kHz (380 and 256 kHz are used in the Two coil directions with respect to the sample surface. As an fiber orientation in the composite.CHENG AND TIAN: SURFACE CRACK DETECTION FOR CFRP MATERIALS USING PEC THERMOGRAPHY 3265 Fig. it may be unnecessary ductivity in the vertical direction is much larger than that in the to use such a high-end camera. 6. parallel eddy currents in the direction of maximum conductivity With awareness of the fiber orientation. Influence of Notch Depth notch. In our studies. The system has a quoted rise time (from the start of horizontal direction is investigated in this case. teract direct heating of the coil. it can be seen that the every 2. tion is investigated. The maximum 383-Hz frame rate provides one frame ison of these two excitation directions. dependent on the the same positional relationship between notch and coil. it can be concluded that the con- for QNDE. as determined in Section IV-B. The 350 mm 38 mm 6 mm CFRP sample contains the same position with respect to the coil for each test. For to high thermal contrast (or high temperature rise). According to (2). The time-do. the coil orientation is in the sample. In general. Thermal image of sample at maximum heating for (a) horizontal and (b) vertical current excitation.4 kW. 2 mm. The rectangular coil is fixed and the sample can be PEC thermography was used to inspect the sample shown in moved. Fig. Thus. hori- experiments). Therefore. Notches with different depths were inspected while retaining As CFRP exhibits directional conductivity. The camera directions. a high-speed thermal entated vertically is much larger than when the coil is orientated camera is used for investigation including feature optimization horizontally.5 to tion. The coil is selected to introduce along the fiber orientation. which is long enough to elicit an observable heat pattern around the C. notches with different depths are placed at Fig. the conductivity of Exel sample in the vertical direc- experimentally. as the conductivity is greater ductivity hollow copper tube. the liftoff influence can be The sample was heated for 200 ms using the rectangular coil eliminated. In addi. tation is in the vertical direction. respectively. the applied field and notch direction to achieve the best temper- a maximum current of 400 and an excitation frequency ature contrast between the regions with and without defects.

VOL. To compare the influence of notch location. A 1-mm-wide and 2-mm-deep experimental results agrees with that from simulation results notch is tested in this experiment. 8. 10(c). it can be ascertained that the deeper investigated in this section. is investigated. The thermal responses for varied notch depths are shown in Fig. 9(a). 8. notch is the largest. and 8-cm coil-notch distances are tested. 3(a) and (b). 10(a) and (b). The relationship Fig. D. 9. 9. Experimental results for transient temperature change against time at “pos 1” shown in Fig. 12. Fig. we can see that the temperature decay rate for the 2 mm deep 0-. unit: digital level. 10. When the distance between the coil and the 2-mm notch plitude of the temperature rise (coefficient ) for a 2-mm-deep is increased to 8 cm. Experimental results of thermal image at the maximum heating (200 ms) for a 2-mm-deep and 1-mm-wide notch. As the fiber orientation is identified the notch is. DECEMBER 2011 Fig. because in Section IV-A. the greater the increase in temperature. Fig. The normalized . illustrated in Fig. 7 at notch bottom for w = 1 mm and varied depth notches: the smallest ( ). From the comparison of by the the 2-cm distance between the coil and the notch. 11. notch is the largest due to highest eddy current density at but the temperature variation is less than one third of that for the bottom of the notch. From the comparison The variation in thermal response with a varied distance be- of the thermal responses at the investigated point for three tween the notch and the coil along the direction of the fiber is notches shown in Fig. Notch Position Invariance Along the Fiber Direction shown in Fig. 9. From Fig. Therefore. NO. heating of the notch can still be seen. we can find the am. the mutual position between the eddy currents are concentrated at the bottom of the notch the notch and the coil is changed by moving the coil. the value for the 2-mm-deep notch (a) normalized responses and (b) nonnormalized (raw) responses. (c) Diagram of the position between the coil and the notch. in the experiment. Varied distance between the coil and 2-mm-deep notch: excitation cur- rent in length direction: (a) 2-cm coil-notch distance and (b) 8-cm coil-notch distance. shown in normalized transient temperature behavior shown in Fig. 2-. Thermal videos are captured at different notch po- between notch depth and transient temperature change from sitions with respect to the coil. shown in when notch depth is smaller than skin depth.3266 IEEE SENSORS JOURNAL. Because in (7) for a 2-mm-deep notch is Fig. 9. the notch depth can be discriminated by coefficients and . is the smallest among the three notches.

Dixon. [1] R. the time delay of thermal or eddy current prop. Through the simulation and experimental results shown in Figs. Fig. respectively. because the notch tance will also be investigated further. 11. notch at different locations with respect to the coil only influences the amplitude of the temperature change. 11(b). Therefore. location and transient temperature change will [Fig.D. Univ. the attenuation delaminations and impact damages rather than surface defects of the eddy current leads to a reduction in the amplitude of the (notches) will be carried out and the relationship between de- temperature change against the distance between coil and notch lamination size. . the location of the heating and cooling thermal response. Sichuan Uni- Unfortunately. • Notch width: The results show that narrower notches lead to not only a greater temperature rise. the transient temperature change with time at varied notch positions in both ACKNOWLEDGMENT the heating and cooling phase is not changed. D. Faculty of Eng. Yang. A.CHENG AND TIAN: SURFACE CRACK DETECTION FOR CFRP MATERIALS USING PEC THERMOGRAPHY 3267 samples. shape and dimensions are not changed. work on subsurface defects like coil-notch distance increases to 8 cm. P. Production Eng. [2] S. respectively. S. but faster temper- ature rise and decay rate at the beginning of the heating phase and in the cooling phase. Burrows. The authors would like to thank Prof. for the first time. re- spectively. and. CONCLUSION AND FUTURE WORK Stuttgart. the maximum standoff dis- ization is not affected. implemented for notch detection in CFRP 2–3. Rashed. 10(b)].. pp. for their useful discussions. Y. The influences of notch depth. 2004. The time delay of either thermal wave or eddy current is much shorter than the minimum detectable REFERENCES time interval from the thermal camera (2. Direc- tional conductivity in CFRP makes the eddy current distribution different from metallic materials. 22. but the normalized transient temperature behavior in the heating phase is not affected. Xi’an Jiaotong Univer- agation from the region beneath the coil to the notch cannot be sity. 10(a). Li. “Combined laser spot imaging thermography and ultrasonic measurements for In this paper. such as in flash thermography. Thus. respectively. width. China.” Ph. China.6 ms). the notch width and depth notch only influences the amplitude of the temperature change can be determined for feature extraction and QNDE. a PEC thermography system has been proposed crack detection. 11. The interaction be. the It can be ascertained from the results that. The notch as a surface crack was detected and observed using PEC thermography through both simulation and experiment. as shown in Fig. 2007. Finally. work. dissertation. be investigated for QNDE. • Notch depth: the relationship between the notch depth and transient temperature change shows that a deeper notch leads to a greater temperature rise at the bottom of the notch. as seen in Fig. the authors observed in thermal videos. In addition. shown in Fig.. Thermal response at notch bottom versus distance between coil and • Notch position invariance with fiber direction: the same notch: (a) normalized responses and (b) nonnormalized (raw) responses. and Automotive Eng. The temperature rise and decay rate after normal. In addition. “Air-coupled ultrasound inspection as a new non-destruc- tive testing tool for quality assurance. and S. vol. 217–227. versity.” Nondestructive Testing and Evaluation. The normalized heating and cooling transient temperature be- and nonnormalized thermal responses at the notch bottom are haviors are not changed. and Associate Prof. The method allows the user to observe the eddy cur- rent distribution in a structure using infrared imaging and detect defects over a relatively wide area. A deeper notch results in a faster normalized tem- perature decay rate in the cooling phase. E. Jun. Germany. Stoessel. because the propagation velocities also would like to thank Exel Composites UK for providing the of the thermal wave and eddy currents are in the order of 10 m/s samples used in the experiments. based on the amplitude of the temperature rise. 2 and 8. and 10 m/s. it can be seen that the heat is mainly generated at the notch. Almond. However. and position are investi- gated in terms of the amplitude of the temperature rise and tran- sient temperature behaviors in the heating and cooling phases. V. Both numerical simulations and experimental investigations have been performed. It has been proven that the PEC thermography technique is feasible for surface defect detection in low conductivity composite materials and is not limited to the sample surface. no. of Stuttgart. further investigations on natural cracks will be under- tween eddy currents and the notch can still be seen when the taken in the future. The conclusions can be drawn as follows. In future in the heating phase. Design.

“Pulsed eddy with leading industrial companies such as Airbus. nPower and current testing with variable duty cycle on rivet joints. 26. Eval. M. 205–218. and H. and M. vol. “A thermographic comparison study for the as. 2006. E. 64. Derby. Esposito. Newcastle upon Tyne. Instrum.. S. Z. 44. Y. Dept. vol. pp. and three Chinese organizations. and R.D. 1.” Insight. M. Dr. Kharkovsky. 43. “Millimeter wave detection of localized anomalies in the destr. assessment of wind turbine systems at Newcastle sessment of composite patches. May 1997.: Oxford Univ. Johnson. Z. degree in electronics from pp. M. Dynam. Belkadi. “Transient thermog. 2010. 1–3. 2009. pp.K. Graduate Conf. G.. 2009. he is Group Head of Communications . Ganchev. ed. G. Apr. China. vol. 1998. Test. 2010. elling and evaluation of eddy current stimulated thermography. method of determining thermal diffusivity. 33. Chemie. de- W. M. 1679–1684. vol. Bendada.. thesis. H. D’Amore.-M. Stanford Univ. Ranu.. J.D. Aug.. DECEMBER 2011 [3] E. no. Busse and P. Yang. pp. Y. He is currently working toward [13] N. Rep. Chengdu. B. A. 19. 1959. Chen. Morbidini and P..” Insight Non-Destructive Testing and [28] J. Phys. Appl. Reader. Professor. 2011.” IEEE Trans. T... H. and F. May 2001. China. E. in 1985 and 1988. in 2008.3. and the Ph. Avdelidis and D. Y. pp. polymer-composite materials.” NDT&E Int.” J. P. Ng. Sophian.” Sichuan. Technol. Technol. “NDT of wind turbine blades using adapted tection in metallic materials. degree in communications and signal of aircraft structures using pulsed thermography.” IEEE Trans. Electronic and Computer En- [21] I. Nico. pp.K. degree in failure models and life cycle and D. J. M. Royal raphy. Vladišauskas. “Thermal wave remote and nondestructive in. pp. Berlin. T. Butler. 3037–3040. pp. C. 133. ments. CA. no.–J. Syst. no. C. 40. and J. He has coordinated several research projects titative evaluation of angular defects by pulsed eddy current thermog. Calabro. Beukers. IR. pp. J. U. Ibarra-Castenedo. R. and In- [23] S. 093530-1–093530-9. “Theory of frequency modulated thermal wave imaging for non-destructive sub-surface defect detection. Eyerer. Tobinson. G.Sc. and Warwick University). J. Press.Sc. 58. Almond.K. Nino. U. D. 37. pp. Abidin. and J. 2258–2264. 2006. C. sities (Durham University. Sci. 105. J. Phys. N. no. Since 2007. He was the recipient of the John Grimwade Award from the British cracks using pulsed eddy current stimulated thermography. Šliteris. paper We. pp. Noethen. and as well as being on several editorial boards of international journals. F. search Fellow with the University of Derby and the [19] M. 1. D. Schulz. “3-D numer- [5] M. Tian. 2001. Condition Monitoring. 9. 2009.K. Ger- ultrasonic and radiographic techniques. Space Shuttle external fuel tank insulating foam. Senior Lecturer. 2004. Avdelidis. R. pp. L. Giordano. 477–483. [27] H. C. plastics. vol. Carlslaw and J. in 1985. 291–299. “3-D eddy current computation in carbon- 12. Menana and M.. V.” in Proc. Qaddoumi. 433–439. Case. NO.” lais. Maldague. pp. [16] R. [10] N.” Appl. [32] H. M.3. U. industry. [11] C.” J. Tian is a Fellow of the Institution of Electrical Engineers. P. VOL. no. 89.K. Appl. no.” Composites Sci.. “Electromagnetic 1008–1011. [29] J. Wolter. Mech. pp. pp. From 2000 to 2006. Wally. I. N. Bowler and M. “RF pulsed heating. Jasiūnienė.. 2010. He received the B.Sc. Abidin.. Apr. Raišutis. Y.. Conduction of Heat in Solids. vol. respec- Mater. Edinburgh University. Appl..K. U. 2009. 353–359. Morozov. 9. degree from the University of [18] G. R. L.. vol. BP. J. 55. N. Liu.und Werkstoffwissenschaften. vol. on top of this he also has good collaboration [22] I. Wilson. 36. “The detectability of cracks using sonic University of East Anglia. [30] D. Cawley.” IEEE Trans. “Thermal Derby. der Université Paul-Verlaine/Metz. gree in metrology and instrumentation and M. Carruthers.. vol. P. 9th Eur. [26] B. Tian. 191913–191913-3. Wu. five Chinese uni- for thermal non destructive characterization of carbon fiber reinforced versities. Almond. China. which is in collaboration with five U. “Comparative study of active thermography techniques for the nondestructive evaluation of Liang Cheng was born in Beijing. Y. ical modelling of the thermo-inductive technique using shell elements. Chung. U. “Surface crack detec- [4] A.” NDT & E Int. no. re- by means of optical lockin thermography. 2009. Magn. A. Humphrey. Lett. A. Phys. “Simulation of edge stNDT. 46. Oswald-Tranta and G. A. Gibson. Newcastle upon Tyne. Aug.8. Oct. heat capacity. no. A. 1250–1257. 2010. Avdelidis. Phys. U. fiber reinforced composites. 2010. Rudlin.. Beijing. He is also a Chinese Changjiang scholar Syst. univer- 1983. N. “Cross correlation based approach tingham University. 302–306. S.” in Proc. “Mod- [8] S. Stanford. Parker. no. London. 43. P. Instrum. S. L.. 537–546. C.Sc. and thermal 1995. Control. the Ph. Y. 12. tion optimization of disbond in layered dielectrics with varying thick. and X. 1–31. Newcastle University. pp. Abbott. 2009. G. Mandache. Ghali. no. Jenkins. TWI among others. Y. Ahmed. Lett. and D. 2009. Martin. P. in 1998. Kim. Currently. [9] S. de- [17] Y. Hung.” Infrared Phys. and D. H. U. pp. no. He is currently working on future reliable renew- [14] G. D.” IEEE Trans. Meas. May 2009. Zoughi. J. Abou-Khousa. Ibarra-Castanedo. Voleišis. J. Tian. Magn. 1923–1928. he has been wissenschaftlischen Technischen Fakultät III. X. Zoughi. Yang.K. Trichet. “Carbon fibre reinforced plastic (CFRP) thermal diffusivity measure- [7] J. no.. L. and P. 2011. J. P. K. 355–357. with the University of Huddersfield. “Quantitative analysis of defects in composite material and Head of the group of Systems Engineering. conductivity.” (ISSE).” Insight.” Composites B. R. one research institute. Chattopadhyay. Tuli. University. Nondestructive Eval. vol. Avdelidis. A. [25] M. continuous sensors and wavelet analysis. G. Spring Seminar Electron. he was a Lecturer. Meyendorf. K. S. Wilson. Ghoshal.D. no. 2004. Huang. 4. P. upon Tyne. gineering. I. Peking University. vol. [6] A. and N. 2007. pp. 237–248. Pritzkau. “Review and comparison of gree in precision engineering from the University of shearography and active thermography for nondestructive evaluation. Maldague.Sc. 11. Féliachi. Mar. 32. 45. vol. from the Engineering and Physical Sciences Research Council (EPSRC). and G. Almond. 51.” Non- Hepburn. L. dissertation. and eddy current NDT: A review. Eng. where he has been Bio. no.” Dr. P. “Quan. Bersee. no. N.3268 IEEE SENSORS JOURNAL. and Signal Processing in the School of Electrical. and A. Sch. Abidin. “Thermo-inductive surface crack de- Mitchard. der Natur. 2nd raphy in the assessment of defects of aircraft composites. 20. Wilson. C... S. 3. 3. Stanford Linear Meas. S. W.” NDT & E Int. Yang. Ing. Conf. vol. laborative UK-China project. Z. spectively. He then became a Research Fellow and Senior Re- vol. Jaeger.” Ph.K. Accelerator Center.. vol. Reinforced Plastics Composites. pp. Newcastle University. A. pp. [20] C. Luk. Zöcke. in 2007. Fouladgar. and M. Universitat des Saarlandes und der Chair Professor in Sensor Technologies and M. vol.. A. Phys. vol. 33rd Int. Piau. tively. P.” in Proc. S. C. Almond.” Appl. pp. Technol. honeycomb structures. and [12] N. 8.K. 326–328. S. 1. no. Eng. 599–605. “Microwave detec. R. Hawtin. Newcastle ducting half-space. Ramdane. [24] B. Pharmazie. 5–6. “Flash nesses. U. 3. J. Eng. Magn. Marioli-Riga. Saarbrucker vanced Sensor Technology. processing from Imperial College London. Tian.. 42. 1961. Guilbert. 43. Rolls Royce.. Genest. P. NDT. 249–254..K. Almond. 4. Institute of Non-Destructive Testing. vol. Tian. 7. Academy of Engineering and FP7. Z. 45. 2003. 73–112. acoustic wave propagation. W. 7. R. Not- [15] R. based at Newcastle University. able energy conversion systems and networks: a col- spection of polymers.. and M. many. P. G.. 95–112. vol. Meas. vol. 5. .” J. and J. NDI of resistance welded composite structures. vol. vol.” Res. tion in ferritic and austenitic steel components using inductive heated Prosser. U. and D. Gui Yun Tian (M’01–SM’03) received the B. pp. “Pulsed eddy-current response to a con. Y.K. Ip.. “Health monitoring of composite plates using thermography. 53. no. “Through skin sensing assessment the M. and L. no. Oxford. 2. pp.” NDT & E Int. [31] W. Program Director of Ad- Ecole Doctorale MIM. B. Abidin. 3. Taylor. pp. Mulaveesala and S. Amos. 25. Mulaveesala and V. vol. vol. “An acoustic-emission characterization of the failure modes in IEEE Trans. 2006. P. S. U. no.