M

TI

a

ARRAA

KGVEDH

1

fomd

ctmef

isarimSes

hd

h0

Sensors and Actuators A 228 (2015) 55–61

Contents lists available at ScienceDirect

Sensors and Actuators A: Physical

j ourna l h o mepage: www.elsev ier .com/ locate /sna

agnetic sensors assessment in velocity induced eddy current testing

iago J. Rocha, Helena Geirinhas Ramos ∗, A. Lopes Ribeiro, Dário J. Pasadasnstituto de Telecomunicac ões, Instituto Superior Técnico, Universidade de Lisboa Lisboa, Portugal

r t i c l e i n f o

rticle history:eceived 20 August 2014eceived in revised form 2 February 2015ccepted 3 February 2015vailable online 14 March 2015

a b s t r a c t

This paper presents an enhancement in the probes to be used on a new nondestructive testing methodwith eddy currents induced by velocity. In this method, a permanent magnet that is attached to a movingcarriage creates eddy currents in the conductive material to be inspected. By measuring the opposingmagnetic field generated by the eddy currents, it is possible to obtain information regarding the presenceof defects. Different magnetic field sensors, such as, differential pick-up coils, giant magneto resistors

eywords:MRelocity inducedddy current testingifferential pick-up coilall sensor

(GMR) and Hall sensors have been used and compared. A permanent magnet moving above a plate wasstudied using a numerical model to allow further improvements to be made in the probe. Depending oneach sensor’s geometry, sensing axis and range, its position and orientation must be strategically chosenin order to increase defect sensitivity. The best probe’s position is the one that guarantees the highestsensibility to the defects’ presence.

© 2015 Elsevier B.V. All rights reserved.

. Introduction

Nondestructive testing (NDT) is a topic of significant relevanceor several industries, such as aeronautics, railroad, petrochemicalr nuclear. There is currently a demand for NDT progress due toaintenance and safety needs in such industries. This factor has

riven a recent growth in the development of new NDT methods.This paper is a step toward the development of a NDT system

apable of being applied to cases where the velocity factor is usedo its benefit by increasing its sensitivity. Examples of industrial

anufacturers that can benefit from these inspection systems arelectrical cable manufacturers or producers of aluminum plates andoils using cold rolled lamination.

Eddy current testing (ECT) is a commonly used technique tonspect conductive materials for defects [1] caused by materialtress or corrosion. It uses a time-varying magnetic field, gener-ted by an excitation coil with a sinusoidal [2] or transient currentegime [3], to induce eddy currents in the material. By measur-ng the disturbances in the path of the induced eddy currents this

ethod is able to detect, locate and characterize the defect [4].

ome of the most used magnetic field sensors are based on Hallffect [5], anisotropic magnetoresistor (AMR) [6], giant magnetore-istor (GMR) [7] and pick-up coils [8].

∗ Corresponding author. Tel.: +351 218418473; fax: +351 218417672.E-mail addresses: tiago.rocha@lx.it.pt (T.J. Rocha), hgramos@ist.utl.pt,

gramos@lx.it.pt (H.G. Ramos), arturlr@lx.it.pt (A. Lopes Ribeiro),pasadas@lx.it.pt (D.J. Pasadas).

ttp://dx.doi.org/10.1016/j.sna.2015.02.004924-4247/© 2015 Elsevier B.V. All rights reserved.

Recent work showed that a modified ECT method using a per-manent magnetic field fixed in a moving medium and a GMR isable to detect defects on an aluminum plate [9]. The advantage ofthis method relies of the fact that its sensitivity increases with thespeed of the probe. This was proved for an aluminum sample for aspeed range from 0 to 6 m/s [10,11].

Other authors [12] also used the same physical principle todetect the presence of cracks on a conductive material but insteadof measuring the magnetic field as our approach does, the quantitybeing assessed is the Lorentz force experienced by the sample. Theycall the method “Lorentz force eddy current testing” (LET).

The monitoring and inspection of metal surfaces for defectsin moving media are an area where ongoing research is stillparamount. In order to understand the effect of speed in thedetection of defects using electromagnetic NDT techniques, someauthors developed numerical models [13,14] applied to actual test-ing solutions. Modeling velocity in NDE is a non trivial problem thatcan only be solved is particular situations.

This paper is a step forward essential to increase the perfor-mance of the method. It reports and analyses in detail the datameasured with different magnetic sensors when a sample withmachined cracks perpendicular to the eddy currents induced in thematerial is inspected. The sensors used are a differential pick-upcoil, a commercial magnetic sensor based on giant magneto-resistors (GMR) and a Hall sensor.

From a preview of the magnetic field pattern in the vicinity ofa conductive plate with defects one is able to decide, based on thequantity (i.e. the magnetic field component) to be measured, whichis the best position for the sensor.

5 d Actu

ricmttfi

2

ttietmtmcip

mia

wtacrm

bio

3

ddspsrSa

In Fig. 2 it is possible to observe that the eddy current densityin front of the magnet is higher than the one behind the magnet.This justifies the location of the magnetic sensors in front of themagnet.

6 T.J. Rocha et al. / Sensors an

The paper is organized as follows: in Section 2 a small theo-etical introduction to the working physical laws of the methods presented; Section 3 introduces the experimental set up and theonditions of the experiments; Section 4 previews the results beingeasured in order to position the probes for the best outcome; Sec-

ion 5 describes the probes; Section 6 presents and concludes abouthe details of the experimental data obtained by each probe andnally in Section 6 conclusions are drawn.

. Modified ECT method

The work described in this paper allows the detection and loca-ion of defects in a metallic plate using a modified eddy currentesting method. In the traditional ECT eddy currents are inducedn the sample plate by a time-varying magnetic field created by anxcitation coil placed closely to the sampled material. The flow ofhe eddy currents is disturbed when a defect exists and the resulting

agnetic field contains the information concerning the perturba-ion. In this work eddy currents are induced by moving a constant

agnetic field in the vicinity of the surface of the sample plate. Thisonstant magnetic field is created by a permanent magnet that, ast moves, creates a relative magnetic field variation in the metalliclate surface.

From the derivation of the Maxwell equations, assuming theagnetic field from the permanent magnet is constant in time, it

s possible to obtain a relation between the magnetic field densitynd velocity:

1��

∇2B + ∇(v × B) = − 1��

∂B

∂t(1)

here B represents the total magnetic field that is a superposi-ion of the primary magnetic field due to the permanent magnetnd of the secondary magnetic field produced by the induced eddyurrents, v is the velocity of the permanent magnet and � and � areespectively the magnetic permeability and the conductivity of theaterial being tested.The detection of defects using this method is therefore possi-

le by measuring the magnetic field density as it contains all thenformation of the eddy currents perturbation due to the presencef defects.

. Experimental setup and conditions of experiments

A probe installed in a carriage moves above the sample asepicted in Fig. 1. The probe contains a permanent magnet to pro-uce the eddy currents in the conductive material and a magneticensor to measure the magnetic field in the vicinity of the sam-le. A LCB060 belt drive actuator from Parker with a 3.5 m long

liding glide system was assembled to move the 150 mm long car-iage that holds the probe. The actuator is powered by a ParkerMH100 brushless servomotor which can accelerate the carriaget a rate of 20 m/s2, to achieve a maximum speed of 6 m/s. A Parker

Fig. 1. Block scheme of the experimental setup.

ators A 228 (2015) 55–61

Compax3 driver/controller controls the motor, using the encoder asfeedback, according to a programmed PLC algorithm. The encodercontains 2048 points per revolution, providing a maximum carriageposition resolution of 0.083 mm. A National Instruments USB-6251DAQ is used to measure the output voltage of the magnetic sensorin the probe with 16 bit resolution. The whole experimental setupis depicted in Fig. 1 where it is visible the drive belt actuator, theactuator driver/controller, the probe, the DAQ a personal computerand the aluminum plate with defects.

Each 0.083 mm step the probe is moved, the drive belt controlleroutputs a pulse that triggers the ADC included in the DAQ board todigitize the probe’s output voltage.

To analyze the effects of velocity and defect depth on theacquired signals, a 4 mm thick aluminum 1050 plate with eightmachined defects was inspected. All defects are 50 mm long,0.5 mm wide and depth varies from 0.5 mm to 4 mm (through plate)with intervals of 0.5 mm.

Tests were performed by programming the controller to acceler-ate the carriage to the desired speed, to maintain the speed whileconducting the measurement, to decelerate the carriage to a fullstop and finally makes it to return to the initial position at a rate of0.5 m/s.

Each experiment is performed at a constant speed. The out-put voltage from the magnetic sensor that denotes the presenceof a defect is an instantaneous peak. Considering that the probe intesting situations shall be attached to a moving carriage with con-siderable inertia, the velocity derivative is always much lower thatthe one related with the amplitude change due to the defect. Thus,the defect is always detectable even if during the measurement anincident occurs that causes a variation of speed.

4. Measurement preview and probe positioning

Considering that the final target of this ECT-based technique isto detect cracks in a material when there is a relative movementbetween the probe and the material then, in order to maximizethe output signal from the magnetic sensor included in the probe,the position of the permanent magnet was chosen to create eddycurrents directed perpendicularly to the cracks’ line on a limitedarea. This situation is depicted in Fig. 2. It was obtained using thecommercial finite element numerical model COMSOL simulatingthe eddy currents distribution on the surface of a 4 mm thick alu-minum sample 1050 without defects when a permanent magnetmoves along x at a constant speed of 20 m/s above the sample.

Fig. 2. Eddy currents distribution on the Al surface when the magnet moves alongx at 20 m/s.

T.J. Rocha et al. / Sensors and Actuators A 228 (2015) 55–61 57

Fp

ai

dpeolbAc

fiFpddhtfwodfinm

t(t

ig. 3. Permanent magnet positioning and current flow perturbation due to theresence of a defect; paths covered by the magnetic sensors.

Fig. 3 depicts the magnet’s alignment with the aluminum platend shows a pictorial representation of the induced eddy currentn the very close vicinity of the magnet.

In Refs. [15,16] it is demonstrated that the eddy currents pathepends on the geometrical parameters of the crack and on thearameters related to the standard skin depth. These papers reportddy current patterns obtained for a spatially parallel distributionf the eddy currents density on a limited area that included theinear crack and oriented perpendicularly to the crack’s line inducedy a sinusoidal magnetic field created by a current in a planar coil.lthough different conditions apply an analogy with the presentonfiguration can be established.

Fig. 4 depicts the two measured components of the magneticeld obtained when the probe scans over the linear defect. Inig. 4(a) the magnetic field component measured in the directionarallel to the eddy currents (and perpendicular of the machinedefect) is presented and in Fig. 4(b) the component parallel to theefect. A larger current density occurs at the surface and a veryigh density on the crack tops. For the linear crack the perturba-ion of the magnetic field component measured along x presentsour peaks, two on each crack end. One single peak is observablehen the field component being measured is along the y-axis. It is

bvious that this makes the y-axis more suitable to be used in cracketection. On the other end one has to be careful as the magneticeld intensity along y may saturate the magnetic sensor and sig-al processing techniques have to be used to subtract the constantagnetic field to highlight the perturbation due to the defect.

Similar results are expected for the geometry proposed within

his paper. The two paths along the movement direction, (a) andb), represented in Fig. 3, will be traversed by the probe to obtainhe magnetic field values that may inform about the existence of

Fig. 4. Output voltage corresponding to the magnetic field components

Fig. 5. Bottom view of the probes. (a) Differential coil; (b) GMR; (c) Hall effect sensor.

defects. Thus, one may expect that when crossing the defect alongthe path (a) (in the middle of the crack), if the magnetic sensorsensing axis is in the y-direction, then the magnetic field has a min-imum due to the conductivity discontinuity caused by the crack.For the same reason, if the magnetic component to be detected isalong x then the path to be traversed must be (b) and two peaks(one negative and another positive) are observed.

Taking this into account, one can conclude that it is possible todetect a defect by measuring the magnetic field in the plate vicinityin either axis, however the component along the crack may be moreadvantageous as only one peak is due. Nevertheless, one has tostress that when doing measurements along the y-direction themagnetic sensor may saturate with the permanent magnetic field.

4.1. ECT probes

Fig. 5 presents photos of the three different probes implementedin our laboratory to be used and compared within this work.

All the probes include the same permanent magnet to induceeddy currents on the aluminum plate. It is a cylinder made from aneodymium alloy N45 grade with 6 mm of diameter and 25 mm ofheight. Its poles are on the flat surfaces.

Each probe includes a different magnetic sensor. Fig. 5(a) depictsthe one with a differential sensing coil with ferrite core. Fig. 5(b)has a GMR based magnetic sensor and Fig. 5(c) depicts a Hall sensor.

The output voltage of each sensor is amplified by instrumen-tation amplifiers INA118 from Texas Instruments. The gain of theamplifier is set accordingly to the sensor output voltage.

4.1.1. Differential coilsEach sensing coil of the differential coil probe presented in

measured for the defect displayed as a dotted line in the center.

Fig. 5(a) is a Murata 1300R Series, 33 �H radial lead inductive ele-ment with a ferrite core with 9 mm of diameter and 13 mm ofheight. They are positioned to be sensitive to the gradient alongx of the time derivative of the component of the total magnetic

5 d Actuators A 228 (2015) 55–61

fiol

4

0Ii3

Gamnbmstw

bt

4

atmpa

5

aapo

5

31

o

adoyt

ma

wad

5

d

Fig. 6. Differential coil output voltage amplified by a factor of 100 for the 3.5 mmdeep defects at six velocities: 1–6 m/s.

top with the eddy currents first deviating away from the center ofdefect and then returning back to it.

8 T.J. Rocha et al. / Sensors an

eld in the y axis. Two coils were used in differential mode insteadf just one, mainly because it reduces the noise caused by slightift-off vibrations while the probe is moving.

.1.2. GMRFig. 5(b) shows the GMR-based magnetic sensor. It is a AA002-

2 from Non Volatile Electronics powered by a ±12 V power supply.t has an omnipolar linear response to the applied magnetic fieldn a wide range of frequencies (DC to 1 MHz) with a sensitivity of.5 mV/V/Oe and can sense magnetic flux densities up to 10.5 Oe.

Considering the magnetic field generated by the magnet, theMR sensor would saturate if placed next to it with its sensingxis along the y direction. One could increase the distance betweenagnet and sensor until it is biased in its linear working zone, but

ot only the intensity of the eddy currents below the sensor woulde significantly lower, but also the velocity at which the probe isoving would affect the biasing field. Considering this, the GMR

ensor was carefully positioned next to the permanent magnet withhe sensing axis parallel to x and at the same time biased in its linearorking zone.

Despite the high sensitivity and bandwidth, this sensor can onlye practical when used to detect the edges of defects when usinghis modified ECT method.

.1.3. Hall sensorFig. 5(c) depicts a Hall sensor placed to sense the magnetic field

long the y axis. The Hall effect sensor is a linear magnetic fieldransducer CYSJ362A from ChenYang Technologies. It was chosen

ainly because of its linear range of ±1.5 T, as it can withstand theowerful magnetic field generated by the permanent magnet. It has

3.5 mV/mT sensitivity and it is powered by a ±5 V power supply.

. Analysis of the results

Measurements of the output voltage were done for each sensort six different speeds (from 1 m/s to 6 m/s with 1 m/s intervals) forll the defects (0.5–4 mm deep) of the aluminum sample. One sam-le of the magnetic sensor output was acquired for every 0.083 mmf probe displacement.

.1. Differential coils

Fig. 6 depicts the output voltage of the differential coils for a.5 mm deep defect at six different speeds (from 1 m/s to 6 m/s with

m/s intervals).Fig. 7 depicts the same output voltage for all the defect depths

f 4 m/s.The depicted waveforms are similar for all the measurements

nd their amplitude increases with the speed and depth of theefect. Measurements were taken with the probe moving through-ut path (a) (see Fig. 3) and detecting the single peak that the-component of the magnetic field exhibits. The waveform showshe time derivative of the magnetic field component along y.

As expected, no voltage disturbance is present due to the verticalovement of the probe, which is an advantage for defect detection

nd waveform processing.Fig. 8 is a two dimensional representation of all the acquired

aveforms where each row represents a measurement made overll the defects at a given velocity. It is visible the effect of defectepth and velocity in the output signal.

.2. GMR

Fig. 9 depicts the output voltage of the GMR for a 3.5 mmeep defect at six different speeds (from 1 m/s to 6 m/s with 1 m/s

Fig. 7. Differential coil output voltage amplified by a factor of 100 for all the defects(0.5 mm deep to 4 mm deep) at 4 m/s.

intervals). Fig. 10 depicts the same output voltage for all the defectdepths at 4 m/s.

Waveforms are obtained moving the probe along path (b) (seeFig. 3) with the GMR positioned with its sensing axis along x. Theoutput voltage is expected to present one maximum and one min-imum, since it senses the peaks of the magnetic field at the crack’s

Fig. 8. 2D representation of the differential coil output voltage for all the defects at1–6 m/s.

T.J. Rocha et al. / Sensors and Actuators A 228 (2015) 55–61 59

Fig. 9. GMR output voltage amplified by a factor of 12 for the 3.5 mm deep defectat 1–6 m/s.

F

d

vmtp

5

di

F

Fig. 12. Hall sensor output voltage amplified by a factor of 100 for the 3.5 mm deepdefect at six velocities: 1–6 m/s.

slow measurements and less deep defects.

ig. 10. GMR output voltage amplified by a factor of 12 for all the defects at 4 m/s.

Fig. 11 is a two dimensional representation the same effect ofefect depth and velocity in the output signal is visible.

In Figs. 9 and 11 a constant magnetic field proportional to theelocity can be observed due to small GMR misalignment. Thisakes the GMR to sense a small portion of the magnetic field in

he y axis. Also, since this is not a differential sensor, some noise ifresent due to vertical vibrations of the probe while it is moving.

.3. Hall effect sensor

Fig. 12 depicts the output voltage of the Hall sensor for a 3.5 mmeep defect at six different speeds (from 1 m/s to 6 m/s with 1 m/s

ntervals) and Fig. 13 presents the output voltage for all the defect

ig. 11. 2D representation of the GMR output voltage for all the defects at 1–6 m/s.

Fig. 13. Hall sensor output voltage amplified by a factor of 100 for all the defects at4 m/s.

depths at 4 m/s using the same sensor. Fig. 14 depicts the same 2Drepresentation as Figs. 8 and 11 for this sensor.

Measurements were taken moving the probe throughout path(a) (see Fig. 3) and detecting the single peak that the y-componentof the magnetic field exhibits. The Hall sensor contains a noise com-ponent present due to the vibrations of the probe while it is beingmoved whose amplitude masks the defect’s perturbations for the

Basically the probe with the differential coils and the probe withthe Hall sensor measure the same component of the magnetic field.However, the output voltage of the differential coil is proportional

Fig. 14. 2D representation of the Hall sensor output voltage for all the defects at1–6 m/s.

60 T.J. Rocha et al. / Sensors and Actu

Fd

ticHafwamti

5

aotuaoe

0l

Fa

ig. 15. Comparison between the derivative of the Hall effect sensor output and theifferential coil output when crossing the 4 mm deep defect at 6 m/s.

o the time derivative of the magnetic flux through its section whilen the Hall sensor it is proportional to the magnetic field value. Toompare the results between the differential coil sensor and theall sensor, the waveform from the latter was differentiated andmplified by a factor of 400 to match the differential coil wave-orm. The result is depicted in Fig. 15. Both waveforms are similar,hich shows that both sensors can be used to detect and eventu-

lly characterize defects when measuring the y component of theagnetic field. The difference in the sensitivity is mainly due to

he ferrite core present in the differential coils, which significantlyncrease the sensitivity.

.4. Hall effect sensor – defect orientation

In order to get better insight on how the defect orientationffects the probe’s sensitivity, four tests were made using differentrientations of the defects. Tests were accomplished by rotatinghe sample plate around the 3 mm deep defect, and scanning itsing the Hall sensor probe moving at 4 m/s. The four tested anglesre 90, 60, 30 and 0◦, where the angle is defined between the linef the defect and the motion axis (x). Fig. 16 depicts the obtained

xperimental results.

In Fig. 16 it is possible to observe that as the angle approaches◦, the amplitude of the peak is reduced. This effect occurs because

ess x oriented currents are perturbed by the presence of a defect

ig. 16. Experimental results for different defect orientations with Hall sensor probet 4 m/s.

[

[

[

[

[

[

ators A 228 (2015) 55–61

at lower angles when compared to the defect perpendicular to themotion axis.

6. Conclusions

This paper reports the measurements obtained with three dif-ferent magnetic sensors using a modified eddy current testingmethod. One of the sensors is composed by two differential sensingcoils, the other is GMR-based and another is a Hall effect sensor. Allsensors successfully detect defects with a depth greater than 1 mm,however the GMR sensor can only detect them if the measurementis made crossing the edge of the defect. The differential coil andHall effect sensor both detect when crossing the whole length of thedefect. This is because each probe measures a different componentof the magnetic field density. Experiments allow us to conclude thatthe defect orientation also influences the sensitivity of the probedetection.

In all three cases the sensitivity increases with the depth ofthe defects and with the probe speed, which demonstrates thatthis modified eddy current method is especially well adapted tosituations where the speed is a factor to be considered.

Acknowledgments

This work was developed under the Instituto deTelecomunicac ões project EvalTubes and supported in part by thePortuguese Science and Technology Foundation (FCT) project: PEst-OE/EEI/LA0008/2013, SFRH/BD/81856/2011 and SFRH/BD/81857/2011. This support is gratefully acknowledged.

References

[1] B. Venkatrama, B. Rao, Practical Eddy Current Testing, Alpha Science, B.P.C. Rao,2007.

[2] J.R. Bowler, N. Harfield, Evaluation of probe impedance due to thin-skineddy current interaction with surface cracks, IEEE Trans. Magn. 34 (2) (1998)515–523.

[3] R. Smith, G. Hugo, Deep corrosion and crack detection in aging aircrafts usingtransient eddy currents, in: Proc. 5th Joint FAA/DoD/NASA Conference on AgingAircraft, San Francisco, USA, 2001.

[4] O. Postolache, H.G. Ramos, A.L. Ribeiro, Detection and characterization ofdefects using gmr probes and artificial neural networks, Comput. Stand. Inter-faces 33 (2011) 191–200.

[5] Y. He, F. Luo, M. Pan, F. Weng, X. Hu, J. Gao, B. Liu, Pulsed eddy current techniquefor defect detection in aircraft riveted structures, NDT & E Int. 43 (2) (2010)176–181.

[6] P. Ripka, M. Vopálensky, A. Platil, M. Döscher, K. Lenssen, H. Hauser, AMR mag-netometer, J. Magn. Magn. Mater. 254–255 (January 2003) (2003) 639–641.

[7] T. Dogaru, S.T. Smith, Giant magnetoresistance-based eddy current sensor, IEEETrans. Magn. 37 (5) (2001) 3831–3838.

[8] A.L. Ribeiro, H.G. Ramos, Inductive probe for flaw detection in non-magneticmetallic plates using eddy currents, in: IEEE Instrum. & Meas. Tech. Conf.,I2MTC, Victoria, Canada, 2008, May, pp. 1447–1451.

[9] H.M.G. Ramos, T. Rocha, D. Pasadas, A.L. Ribeiro, Velocity induced eddy currentstechnique to inspect cracks in moving conducting media, in: Instrumentationand Measurement Technology Conference (I2MTC), Minneapolis, MN, 2013,May, pp. 931–934.

10] H.G. Ramos, T. Rocha, D. Pasadas, A.L. Ribeiro, Faraday induction effect appliedto the detection of defects in a moving plate, in: 39th Annual Review of Progressin Quantitative Nondestructive Evaluation, AIP Conf. Proc., vol. 1511, 2012, pp.1490–1497.

11] H.M.G. Ramos, T. Rocha, D. Pasadas, A.L. Ribeiro, Velocity induced eddy currentstechnique to inspect cracks in moving conducting media, in: Instrumenta-tion and Measurement Technology Conference (I2MTC), Montevideo, Uruguay,2014, May, pp. 931–934.

12] R.P. Uhling, M. Zec, H. Brauer, A. Tess, Lorentz force eddy currents testing: aprototype model, J. Nondestruct. Eval. 31 (2012) 357–372.

13] H. Song, N. Ida, Modeling of velocity terms in 3D eddy current problems, IEEETrans. Magn. 28 (2) (1992) 1178–1181.

14] Y.K. Shin, W. Lord, Numerical modeling of moving probe effects for electromag-netic nondestructive evaluation, IEEE Trans. Magn. 29 (2) (1993) 1865–1868.

15] H.G. Ramos, T. Rocha, D. Pasadas, A.L. Ribeiro, Determination of linear defectdepths from eddy currents disturbances, in: 40th Annual Review of Progress inQuantitative Nondestructive Evaluation, AIP Conf. Proc., 2013, pp. 1448–1455.

d Actu

[

B

University of Lisbon (UTL) in 2010. He is currently aPh.D. student in electrical and computer engineering atIST/UTL researching new eddy current probes, digitalsignal processing and inverse problem algorithms for

T.J. Rocha et al. / Sensors an

16] D. Pasadas, A.L. Ribeiro, T. Rocha, H.G. Ramos, Open crack depth evaluation usingeddy current methods and GMR detection, in: IEEE Workshop on Metrology forAeroSpace, Benevento, Italy, 2014, May.

iographies

Tiago Jorge Rocha was born in Lisbon, Portugal, in 1987.Received the M.Sc. degree in electronic engineering fromInstituto Superior Técnico, Technical University of Lisbon(IST/UTL), in 2011. He is currently a Ph.D. student in elec-trical and computer engineering at IST/UTL researchingvelocity induced eddy current testing.

Helena Geirinhas Ramos was born in Lisbon, Portugal,on October 1957. She received the M.Sc. and Ph.D.degrees in electrical and computer engineering fromthe Instituto Superior Técnico of the Technical Univer-sity of Lisbon (IST/UTL), in 1987 and 1995 respectively.In 1981, she joined the Department of Electrical andComputer Engineering at IST/UTL, where she has beena member of the faculty. She joined the Instituto de

Telecomunicac ões in 1995. Her main research interestsare in the area of instrumentation, transducers and mea-surement techniques, automated measurement systems,applied electromagnetism, nondestructive evaluation andeddy current testing.

ators A 228 (2015) 55–61 61

Artur Lopes Ribeiro was born in Lisbon, on April 8, 1950.He received the degree in electrical engineering fromthe Instituto Superior Técnico (IST), Technical Univer-sity of Lisbon (UTL), Lisbon, Portugal, in 1973, and thePh.D. degree in electrical and computer engineering fromIST/UTL in 1990. He joined the Department of Electricaland Computer Engineering, IST/UTL, in 1977, where he hasbeen a member of the faculty. He joined the Instituto deTelecomunicac ões in 1991. His current research interestsinclude the instrumentation and electric measurement,non destructive evaluation, numerical modeling of elec-tric and optoelectronic components.

Dário Jerónimo Pasadas was born in Geneva,Switzerland, on April 28,1986 and he came to Portugalin 1997. He received his M.Sc. degree in electronics engi-neering from Instituto Superior Técnico (IST), Technical

application on non-destructive evaluation.