A novel dual modality sensor with sensitivities to permittivity, conductivity and permeability

Transcription

1 A nove dua modaity sensor with sensitivities to permittivity, conductivity and permeabiity Jorge R. Saas Avia, Kin Yau How, Mingyang Lu and Wuiang Yin, Senior Member, IEEE Schoo of Eectrica and Eectronic Engineering The University of Manchester, Manchester M3 9PL, UK Abstract - In this paper, an eectromagnetic (EM) sensor which can operate simutaneousy in capacitive and inductive modaities with sensitivities to permittivity, conductivity and permeabiity is deveoped, and a nove measurement strategy is proposed accordingy. The sensor is composed of two panar spira cois with a track width of 4 mm, which promotes its capacitive mode. The capacitive couping is measured in common mode whie the inductive couping is measured in differentia mode. In capacitive mode, the sensor is sensitive to changes in permittivity, i.e. the dieectric materia distribution; whie in inductive mode, it is sensitive to magneticay permeabe materia and eectricay conductive materia. Furthermore, it is demonstrated that the sensor can simutaneousy measure dieectric and conductive materias. This nove sensing eement has been designed and impemented. Experimenta resuts verified its effectiveness in dua modaity measurement. Keywords - panar sensors; EM sensor; dua modaity; eddycurrent testing; combined sensing. I. INTRODUCTION Evauation of materias by using eectric or magnetic fieds has been extensivey performed for various inspection purposes, such as faiure detection, quaity assurance and materia composition inspection [], []. The seection of the measurement method is determined by the fundamenta eectrica and magnetic properties of the materia of interest i.e. permittivity, conductivity and permeabiity [3]. Capacitance measurements are appropriate for evauating dieectric materias; for exampe, panar capacitance sensors have been used to inspect variations in dieectric properties of materias [4], [5]. Magnetic induction / eddy-current testing is suitabe for evauating and inspecting conductive/permeabe materias with many different coi configurations having been investigated, incuding panar spira cois [6]. Measuring the change in both capacitance and mutua inductance with a suitabe sensor gives the possibiity of inspecting the fundamenta eectrica and magnetic properties (permittivity, conductivity and permeabiity) with a singe sensor. Therefore, insuators, conductors and composite materias can be inspected with one sensor. Attempts of combining capacitive and inductive measurements have been reported previousy. In [7] and [8], by switching between modes of operation or mutipexing, the presence of conductive and dieectric materias is detected with a dua mode sensor, but the sensor is sti a physica combination of two sensors (separate capacitive and inductive eements). A printed sensor was reported in [9]; by identifying the predominant sensor response above and beow the resonant frequency, it was possibe to distinguish between conductive and dieectric materias. In [0], meander and mesh panar sensors were empoyed for inspection of conductive and dieectric materias; the effects of some dieectric sampes on the transfer impedance using frequencies up to hundreds of megahertz were reported. This sensor is sensitive to both conductive and dieectric materias, but was not capabe of determining both properties simutaneousy. In this paper, we present a nove sensor which inherenty is a dua inductive/capacitive sensing eement and thus is sensitive to changes in conductivity, permittivity and permeabiity; and importanty, inductive/capacitive effects can be separated by using different modes of measurement. In differentia mode, the change in mutua inductance is measured; and in common mode, the change in capacitive couping is measured. Measurements can be taken in differentia and common modes simutaneousy with an impedance anayser with a suitabe configuration, therefore, the sensor can work simutaneousy in inductive and capacitive modes and there is no need for switching between different sensing eements. This aows fast measurements to be carried out and avoids the associated disadvantage in Ref. [7], i.e. the setting time and system stabiity need to be considered associated with the switching and a programmabe deay has to be introduced to avoid interference. Moreover, the designed panar sensor has some advantages incuding, easy manufacturing, good repeatabiity, ow cost as in [], and can be buit of fexibe materias for inspection of irreguar surfaces as in [6]. The sensor was designed and buit, and experimenta resuts for measuring conductive, dieectric and permeabe materias are presented. We tested the sensor with a range of materias and combinations: such as water, air, pastic pates, copper Jorge R. Saas Avia woud ike to thank the Nationa Counci of Science and Technoogy (CONACYT) of Mexico for sponsoring his PhD studies.

2 Fig.. Sensor ayout. Units are in miimetres. pates and ferrite rings. In addition, the sensor was coated and resuts are presented for an immersion experiment. II. SENSOR DESIGN The sensor is composed of two panar spira cois printed on a P. Fig. depicts the ayout of the sensor. The trace width is 4 mm, and the gap between the traces is mm; the separation between the nearest traces of the coi pair is 5 mm, and the distance between the centres of the coi pair is 4 mm. The P substrate is made of.6 mm FR-4, with a reative permittivity vaue of 4.4. The sef-inductance of each coi is ~30 nh, mutua inductance 0 nh and direct couping capacitance.56 pf at 00 khz measured with an impedance anayser (SL 60). The same instrument was used in the foowing experiments; connections between the sensor and the impedance anayser are shown in Fig.. The instrument has two measurement channes which can be independenty configured as differentia and common modes; thus, differentia and common mode measurements can be taken simutaneousy with this configuration. III. SENSING MODES AND MEASUREMENTS MODES A. Inductive sensing mode Currents fowing in the tracks on the excitation side produce magnetic fied, which induces votage in the receiving side due to magnetic induction. So, the same panar structure that is used for capacitive measurements can be treated as cois for magnetic induction measurement. The conductivity and magnetic permeabiity of the sampe affect the magnetic induction due to eddy currents and magnetic poarisation, and the effects can be measured through the induced votage across the receiver coi []. As the magnetic fied depends on the coi geometry, the sensitivity of the sensor is intrinsicay reated to its geometry. The anaytica soution for the change in impedance of a panar circuar spira coi can be derived from Dodd and Deeds theory as presented by Ditchburn [6]. Circuar and rectanguar geometries for panar cois have been compared due to its simiar behaviour [3]. The Dodd and Deeds anaytica soution describes the inductance change of an air-core coi pair caused by a metaic pate for both non-magnetic and magnetic cases. The difference in the compex mutua inductance is L(ω) = L(ω) L A (ω) where the coi inductance above a pate is L(ω), and L A (ω) is the inductance in free space. In the region between and, the vector potentia can be expressed as (), where N, N denote the number of turns in the excitation and pickup coi; α is a spatia frequency variabe; μ0 denotes the permeabiity of free space; e and e denote the height of bottom and top of the excitation coi; whie p and p denote the height of bottom and top of the pickup coi; re and re denote the inner and outer radii of the excitation coi; whie rp and rp denote the inner and outer radii of the pickup coi; and c denotes the thickness of the pate. J(x) is a first-order Besse function of the first kind. I(x, x) represent the production of J(x) for radii of x and x. A, r,z IN ze ze I 3 (r e,r e ) J ( r)[ e e 0 0 re -r e e - e e e ( )( ) ( )( )e c z e e e ]d c ( )( ) ( )( )e The votage induced in the reciver with a singe turn can be expressed as s p s () j A r,z ds j A r,z r cos d () where φ = θ + tg (r p sin θ/(w r p cos θ)) is the ange between A and ds; s denotes the transect area in the pickup coi; whie r = r p sin θ + (w r p cos θ) is the distance GEN Impedance/Gain-Phase Anayser Fig.. Connections between the sensor and the impedance anayser. Fig. 3. Dodd and Deeds simuations setup.

3 between O and ds. Considering () and (), the induced votage on the receiver can be yieded in (3). rp NNj 0I rp sin 3 e e e cos tg I(r, r ) r -r w r e e - e r p -r p p - p 0 0 rp p cos J( r sin w r cos ){ - [e e e ( )( ) ( )( )e e e e e - p p e e c ]}dr pd d c ( )( ) ( )( )e Consequenty, the mutua inductance between the aircored coi pair can be presented by dividing the induced votage by the current fowing through the excitation coi, as shown in (4). rp NN rp sin 0 L 3 e e cos tg I(r, r ) re -r w r e e - e r p -r p p - p 0 0 rp p cos J( r sin w r cos ){ - [e e e ( )( ) ( )( )e e e e e - p p e e c ]}dr pd d c ( )( ) ( )( )e B. Capacitive sensing mode With arge surface track width, each of the panar cois acts as a capacitive pate, where one is the transmitter and other the receiver. Therefore the capacitive sensing mechanism is simiar to that of a two copanar pate configuration [4]. When a sinusoida votage is appied to the transmitter, a potentia difference is estabished and thus a capacitive couping deveoped. Introducing permittivity change in the sensing area wi perturb the estabished potentia distribution and hence the capacitive couping, the change of which can then be measured. Depending on the nature of the sampe, different effects are expected as discussed in [5]. A grounded object reduces the eectric fux reaching the receiver due to a eakage through the newy formed ground path, and therefore wi reduce the capacitive couping; a foating sampe with a higher permittivity generay increases the capacitive couping. These effects are referred as shunt mode and transmission mode (3) (4) respectivey [4]. Both effects were observed in our sensor, but the interest of this work is for the atter case where the sampe is eectricay foating. Goss et a. [6] identified six couping mechanisms for an excitation/detection coi pair with a sampe in-between. It was stated that the potentia difference between the cois, the surface area of the target, and the direct capacitive couping between the cois strongy infuence the capacitive excitation capacitive detection mode. Whie in magnetic inductive measurements, the capacitive couping effect needs to be minimised, the sensor deveoped here intentionay expoits this effect. By using a arge track width, a significant direct capacitive couping between the tracks deveops. The track width is imited by the overa sensor size and therefore a track width of 4 mm was seected. -D finite-eement simuations were carried out to expore the sensitivity distribution of the sensor over a sampe. Both cois, excitation and detection, were segmented in 6 traces T A to T F and D A to D F as shown in Fig. 4. Treated as a coi, a sinusoida potentia is appied to the excitation eement. For simuation, different potentias were assigned to the excitation coi traces due to the fact that it is the ony oad for the signa generator and the votage drop must occur aong the excitation tracks. The excitation coi potentias going from to 0. Assuming no inductive couping, as in conventiona capacitive measurements [7], a the traces of the detection coi are at the same potentia. For the sensor working simutaneousy in inductive and capacitive modes, a potentia difference exists over the detection coi between its different segments due to inductive couping. Different potentias were assigned to the detection coi; considering the sensor inductive couping coefficient, the receiver coi potentias were set to /6 of the excitation coi potentias. The eectric fied distribution for an inductivey couped receiver is shown in Fig. 4. The reative permittivity of the sampe was set to one. TABLE I PACITANCE BETWEEN A TRANSMITTER TRACK AND THE RECEIER Couping Reative permittivity of the sampe 3 80 T A -Receiver 446 ff 6 ff 0 ff T B -Receiver 60 ff 5 ff 97 ff T C -Receiver 59 ff 78 ff 955 ff TABLE II PACITANCE BETWEEN INDIIDUAL TRACKS Fig. 4. Eectric fied distribution. Units are in vots. Couping Reative permittivity of the sampe 3 80 T A -D A 66 ff 36 ff 9 ff T A -D B 33 ff 49 ff 85 ff T A -D C 66 ff 09 ff 05 ff

4 Fig. 5. Normaised sensitivity distribution of the sensor for capacitive sensing mode. It is worth noting that the capacitive couping between segments is independent of the potentia set up. From the previous simuation, it can be observed that individua capacitive couping between each segment of the transmitter and the receiver deveops; i.e. between segment T A and the receiver segments D A to D F, T B and the receiver segments D A to D F, and so on; Tabe I shows the simuation vaues for the segments T A, T B and T C. Simuation resuts of the capacitance between the individua transmitter tracks T A, T B and T C with the receiver indicate that the nearest track T A has the strongest couping with the receiver. Tabe II shows the deveoped capacitance between the nearest track of the excitation coi T A and the segments D A, D B and D C. Therefore, the strongest capacitive couping is between the adjacent tracks of the excitation and detection cois i.e. between the segments T A and D A. The sensor sensitivity distribution to permittivity cacuated according to the E dot E formuation [8], [9] is shown in Fig. 5. The centre of the sensor is positioned at the coordinate x = 0 mm, y is the distance between the sensor and the sampe. As expected the sensitivity is concentrated in the centre region of the sensor. Overa, the average couping effect from a the segments is measured. Simuation resuts give an overa capacitance between the transmitter and the receiver of.34 pf which is in accordance with measurement resuts. C. Measurements modes: differentia mode, common mode, and simutaneous mode An equivaent circuit of the sensor and a sampe in-between is shown in Fig. 6 [6]. Couping is both capacitive and inductive between the cois (direct couping: Cd and Md) and through the sampe (indirect couping: Cs, Ms, Cs and Ms). The track resistance and parasitic capacitance of the Ms Cs C3 R3 L3 Ms Cs cois are not shown. Each of the panar sensors is treated as a coi, represented as L for the transmitter and for the receiver. The target is modeed as an equivaent RLC parae circuit, where R3 represents the osses due to eddy currents for a conductive sampe; L3 is the inductive eement reated to the eddy currents; and C3 is the capacitive couping eement reated to the dispacement currents. The change in mutua inductance between the transmitter and the receiver can be detected by measuring the differentia votage change at the two terminas of the receiver coi []. As stated by Equation (5), due to magnetic induction, the change in votage Δ is proportiona to the changes in mutua inductance ΔM, the current in the transmitter coi I and the anguar frequency ω; j is the imaginary unit. = jω MI (5) The measured capacitance Cm is Cd in parae with the series equivaent of Cs, C3 and Cs. In order to separate the inductive and capacitive couping effects, different measurement modes were used, i.e. common mode and differentia mode. With common mode, the measurement is sensitive to the potentia difference between the transmitter and receiver and therefore it is reated to capacitive couping. With differentia mode, it is sensitive to the votage difference between the receiver coi terminas and therefore reated to inductive couping. In differentia mode, the circuit can be treated as two mutuay couped cois. In common mode, the circuit can be simpified as shown in Fig. 7. A is the common mode votage that is determined by the unknown capacitance Cm and the input impedance of the impedance anayser. can be treated as a constant RC parae circuit once the measurement setup is fixed. Simutaneous mode (simutaneous capacitive and inductive measurement) was deveoped in order for the proposed sensor to be abe to sense conductivity, permittivity and permeabiity simutaneousy. Common mode votage sees the receiver coi as one conductive surface at a reference potentia eve at the point of connection. Therefore, paths for the movement of charge due to the potentia difference between the transmitter and the receiver are created. However, common mode measurements do not force a uniform potentia eve on the receiver; hence a differentia votage across the receiver terminas due to inductive couping can be measured simutaneousy. An equivaent circuit for simutaneous mode is shown in Fig. 8. L and represent the excitation and receiver Cm A L Md Cd Fig. 6. Equivaent circuit of the sensor and a sampe. Fig.7. Simpified common mode setup.

5 eements respectivey. The potentia difference between points and at the receiver due to inductive couping is represented with resistors to and coi. Capacitances to represent the capacitive couping at different points on the receiver track. is the input impedance of the impedance anayser. Differentia mode votage is the votage difference between the points and. As shown above, the strongest capacitive couping effect occurs between the nearest tracks of the transmitter and the receiver. Therefore, when taking common mode measurement, the point at which the votage is measured is at. Considering the equivaent circuit of Fig. 8, differentia mode measurement Δ = contains inductive couping. Common mode votage can be defined as A =I C, where I C is the current from the transmitter to the receiver due to overa capacitive couping i.e. the sum of a the currents I A to I F in the form of (6). I A = jω(c TA TA + C TB TB + C TC TC + ) I. EXPERIMENTAL SETUP AND SAMPLES An impedance/gain-phase anayser SL 60 was used to carry out the measurements. The instrument has a signa generator output and two input votage measurement channes that can be configured as either common mode or differentia mode. As shown in Fig., the excitation coi was connected to the signa generator and the receiver coi terminas were connected to both votage channes in parae; channe one was configured as differentia mode and channe two as common mode. Fig. 9 shows the schematic connections between the equivaent circuit and the instrument for each channe separatey. Fig. 9(a) represents common mode and Fig. 9(b) differentia mode. R and R in both figures represent the input impedance of the impedance anayser with the vaues taken from the instrument manua. As can be seen from Fig. 9(a), when common mode is seected, ony one of the terminas is internay connected for measuring purposes. Thus, the receiver coi is represented as ony one pate. Cm represents the couping capacitance between the sensor pair. From Fig. 9(b) it can be seen that the measured votage in differentia mode corresponds to the L L Excitation L L L Detection (6) L L A A Cm ΔM (a) Common mode (b) Differentia mode Fig. 9. Instrument connections for each mode separatey. R M Δ R M differentia votage between the two terminas of the receiver coi. Once a connections are made, simutaneous data of both channes (one in differentia mode and the other in common mode) can be obtained with the instrument. Sampes with different eectromagnetic properties were prepared to test the behaviour of the sensor. In order to test the inductive couping, a set of conductive sampes with different thickness was created by stacking -5 copper foi ayers. Each ayer has a thickness of 60 μm. A set of high permeabe sampes (nicke-zinc 0 mm x 8 mm ferrite rings) was used to introducing permeabiity changes. Ferrite rings were centred between the transmitter and the receiver. Lift-off is.6 mm. For testing the capacitive effect, pastic sheets were measured and a water immersion experiment was carried out. Lift-off for pastic sheets is.6 mm; each pastic pate is.5 mm thickness and has dimensions of 30 mm by 65 mm. For the immersion experiment, the sensor was coated; the surface area of the water voume was constant, and the height is ineary reated to the voume, approximatey mm for each 0 of water once the sensor is fuy submerged. For the foowing discussion, reative permittivity vaues of for air, 3 for pastic sampes and 80 for water sampes are assumed; which are typica vaues for these materias. A representation of the sampes positioning is shown in Fig. 0. Excitation frequency was set to MHz for common mode and simutaneous experiments; and to 00 khz for the differentia mode experiments. A R M Sensor substrate Sampe ayer Sampe ayer Sampe ayer 3 Fig. 8. Equivaent circuit of the sensor in simutaneous mode. Fig. 0. Setup for measuring sampe ayers.

6 pf Common mode [pf] Differentia mode (normaised) pf Common mode [pf] Differentia mode (normaised) normaised. RESULTS Resuts were obtained first for capacitive sensing/common mode measurement, then for inductive sensing/differentia mode, and asty for the simutaneous mode. A pot for capacitive-sensing/common-mode-measurement is shown in Fig.. The first datum, abeed Air, is the interpates capacitance vaue of the sensor in air; in this case, the capacitive couping is through the P substrate and air. As the thickness of the pastic pate increases, the measured capacitance increases as expected. Resuts of a D finite eement simuation are aso shown in Fig.. The capacitive couping increases as the pastic sampe thickness increases, same trend as the in the experiments. The capacitance range is.34 pf to.7 pf for simuation and.57 pf to.76 pf for experiments. The error can be attributed to the D nature of the simuation. Measurements for different voumes of water were aso carried out in common mode. As shown in Fig., the capacitive couping increases with increasing voume of water as expected. A arger change in the measured capacitance due to the presence of water than pastic pates can aso be seen, which is attributed to a much higher permittivity of water than that of the pastic sampe. To test inductive couping in differentia mode, copper sampes with different thickness positioned at 5 mm away were measured at 00 khz (Fig. 3). A reduction in the measured votage is observed when the thickness of the sampe increases. This reduction is in accordance with simuations using the method in Section III and the magnetic induction effect for highy conductive, nonmagnetic sampes []. To test the simutaneous mode, two experiments were carried out:. A 300 μm copper pate sampe was positioned 5 mm away from the sensor, and then, pastic sampes of different thicknesses were paced in-between. As expected, the differentia measurement remained the ayer ayers 3 ayers 4 ayers 5 ayers Copper sampes Differentia mode Simuation Fig. 3. otage change in differentia mode: copper sampes 5 mm away (00 khz). normaised = sampe air / air Simuation.35 Measurements.5.5 Air ayer ayers 3 ayers 4 ayers 5 ayers Pastic ayers Fig.. Capacitance in common mode: pastic sampes Air Copper at 5 mm pastic ayer inbetween Common mode Differentia mode pastic ayers inbetween 3 pastic ayers inbetween Fig. 4. Common mode and differentia mode: pastic sampes in-between a five ayers copper sampe positioned 5 mm away. normaised = + ( sampe air )/ air Fuy subm. Water Pastic container Water 5 Ferrite ring Ferrite rings Common mode Differentia mode 3 Ferrite rings 4 Ferrite 5 Ferrite rings rings Fig.. Capacitance in common mode: water immersion. Fuy submerged corresponds to 60. Fig. 5. Common mode and differentia mode: water and ferrite rings. normaised = + ( sampe air )/ air.

7 same but the common mode measurements increase with the thickness of the pastic pates (Fig. 4). Therefore, it was verified that the sensor can simutaneousy operate in both modes.. Ferrite rings were introduced in a container with 5 of water (Fig. 5). The experiment demonstrates that differentia mode measurement is sensitive to permeabe materias (ferrite rings) but not to materias with permittivity (water); common mode measurement is mainy sensitive to materias with permittivity (water) with a change of capacitance of.3 pf, but ony weaky sensitive to permeabiity (ferrite rings) with a sma change of 0.05 pf. I. CONCLUSIONS AND FUTURE WORK In this work, a nove dua modaity sensor and the corresponding measuring strategy are presented. The sensor acts as a spira coi pair, and as a panar capacitive sensor pair depending on the measurement mode. Equivaent circuits of the sensor were utiised to anayse the sensor response and deveop the measurement strategy for each mode. Resuts from measurements indicate that in differentia mode, the change in mutua inductance is measured; in common mode, the change in capacitance couping is measured; and that simutaneous measurements for inductive and capacitive couping can be performed. Tests aso suggest that the sensor is sensitive to conductivity and permeabiity in differentia mode and permittivity in common mode. Therefore, this sensor and the measurement strategy have the potentia to inspect insuators, conductors and composite materias. The versatiity of the sensor is aso demonstrated with an immersion experiment. Future research wi focus on optimisation of sensor geometries for specific measurement appications such as inspection of composite materias in NDT [0], [] and mutiphase fow measurements; where components with conductivity, permittivity and permeabiity are present. In addition, a custom instrument based on FPGA [] [4] wi be buit to repace the commercia instrument (SL 60) to impement the measurement strategy. ERENCES [] U. Kaatze, Measuring the dieectric properties of materias. Ninety-year deveopment from ow-frequency techniques to broadband spectroscopy and high-frequency imaging, Meas. Sci. Techno., vo. 4, no., p. 005, 03. [] J. García-Martín, J. Gómez-Gi, and E. ázquez-sánchez, Nondestructive techniques based on eddy current testing, Sensors, vo., no. 3, pp , 0. [3] J. Bitz, Eectrica and magnetic methods of non-destructive Testing, nd ed. Springer, 997. [4] A. a Nassr, W. H. Ahmed, and W. W. E-Dakhakhni, Copanar capacitance sensors for detecting water intrusion in composite structures, Meas. Sci. Techno., vo. 9, no. 7, p. 7570, 008. [5] X. Yin and D. A. Hutchins, Non-destructive evauation of composite materias using a capacitive imaging technique, Compos. Part B Eng., vo. 43, no. 3, pp. 8 9, 0. [6] R. J. Ditchburn, S. K. Burke, and M. Posada, Eddy-current nondestructive inspection with thin spira cois: Long cracks in stee, J. Nondestruct. Eva., vo., no., pp , 003. [7] B. George, H. Zang, T. Bretterkieber, and G. Brasseur, A combined inductive-capacitive proximity sensor for seat occupancy detection, IEEE Trans. Instrum. Meas., vo. 59, no. 5, pp , 00. [8] X. Hu, Imaging of hidden objects by panar capacitive and magnetic sensor arrays for security appications, University of Manchester, 009. [9] J. Long and B. Wang, A metamateria-inspired combined inductivecapacitive sensor, SPIE Mutisensor, Mutisource Inf. Fusion Archit. Agorithms, App., 04. [0] S. C. Mukhopadhyay, Nove panar eectromagnetic sensors: modeing and performance evauation, Sensors, vo. 5, no., pp , 005. [] L. Abdia, L. Mangion, and C. Sammut, Eddy current sensing using panar cois, IEEE Instrum. Meas. Mag., vo. 5, no. 6, pp. 4 8, 0. [] A. J. Peyton, Eectromagnetic induction tomography, in Industria Tomography: Systems and Appications, Woodhead Pubishing, 05, pp [3] R. J. Ditchburn and S. K. Burke, Panar rectanguar spira cois in eddycurrent non-destructive inspection, NDT E Int., vo. 38, no. 8, pp , 005. [4] X. Hu and W. Yang, Panar capacitive sensors designs and appications, Sens. Rev., vo. 30, no., pp. 4 39, 00. [5] J. R. Smith, Eectric Fied Imaging, Massachusetts Institute of Technoogy, 999. [6] D. Goss, R. O. Mackin, E. Crescenco, H. S. Tapp, and A. J. Peyton, Understanding the couping mechanisms in high frequency EMT, 3rd Word Congr. Ind. Process Tomogr., pp , 003. [7] W. Q. Yang, Hardware design of eectrica capacitance tomography Systems, Meas. Sci. Techno., vo. 7, no. 3, pp. 5 3, 996. [8] D. N. Dyck, D. A. Lowther, and E. M. Freeman, A method of computing the sensitivity of eectromagnetic quantities to changes in materias and sources, IEEE Trans. Magn., vo. 30, no. 5, pp , 994. [9] W. Yin and A. J. Peyton, Sensitivity formuation incuding veocity effects for eectromagnetic induction systems, IEEE Trans. Magn., vo. 46, no. 5, pp. 7 76, 00. [0] W. Yin, P. J. Withers, U. Sharma, and A. J. Peyton, Noncontact characterization of carbon-fiber-reinforced pastics using mutifrequency eddy current sensors, IEEE Trans. Instrum. Meas., vo. 58, no. 3, pp , 009. [] J. Cheng, J. Qiu, H. Ji, E. Wang, T. Takagi, and T. Uchimoto, Appication of ow frequency ECT method in noncontact detection and visuaization of RP materia, Compos. Part B Eng., vo. 0, pp. 4 5, 07. [] T. Yang, Q. Zhao, K. Y. How, K. Xu, M. Lu, Y. Xie, and W. Yin, Leve measurement for saine with a sma surface area using high frequency eectromagnetic sensing technique, Measurement, vo. 0, pp. 8 5, 07. [3] W. Yin, G. Chen, L. Chen, and B. Wang, The design of a digita magnetic induction tomography (MIT) system for metaic object imaging based on haf cyce demoduation, IEEE Sens. J., vo., no. 0, pp , 0. [4] W. Yin, S. J. Dickinson, and A. J. Peyton, A muti-frequency impedance anaysing instrument for eddy current testing, Meas. Sci. Techno., vo. 7, no., pp , 006.