sensors Article

An Improved Design of the Spiral-Coil EMAT for Enhancing the Signal Amplitude Xiaojuan Jia, Qi Ouyang * and Xinglan Zhang Institute of Smart Engineering, School of Automation, Chongqing University, Chongqing 400044, China; [email protected] (X.J.); [email protected] (X.Z.) * Correspondence: [email protected]; Tel.: +86-139-8365-1722 Academic Editor: Xiaoning Jiang Received: 12 March 2017; Accepted: 22 April 2017; Published: 12 May 2017

Abstract: The low energy transition efficiency of electromagnetic ultrasonic transducer (EMAT) is a common problem in practical application. For the purpose of enhancing the amplitude of the received signal, an improved double-coil bulk wave EMAT is proposed for the thickness measurement of metallic block. This new configuration of magnets consists of a solid cylindrical magnet and a ring-shaped magnet encircling the outer side of the solid cylindrical one. A double-coil was applied instead of a single spiral-coil. Numerical simulations were performed to analyze and optimize the proposed configuration of the EMAT by the 2-D axisymmetric finite element model (FEM). The experiment effectively verifies the rationality of the new configuration and the feasibility of improving the signal strength. Keywords: electromagnetic ultrasonic transducer; double-coil; finite element model; signal-to-noise ratio; bulk wave

1. Introduction Nondestructive testing (NDT) is an indispensable tool for industrial development, and its method mainly includes ultrasonic testing (UT), radiation testing (RT), magnetic particle testing (MT), liquid penetration testing (PT) and AC field measurement (ACFMT). Electromagnetic acoustic transducer (EMAT) is a type of non-contact ultrasonic transducer, which is the core component of ultrasonic excitation and acceptance in UT [1–5]. Due to a plenty of outstanding advantages such as non-contact operation, requiring no coupling fluid, the flexibility to generate and detect multiple wave modes, and the capability of operating in high temperature, isolation layer and other harsh environment, EMATs are particularly useful and active in nondestructive evaluation (NDE) and internal structure evaluation (ISE) [6–12]. Besides, EMATs have be distinctively attractive in thickness measurement for their good penetration ability. However, compared with conventional piezoelectric transducers, the poor energy efficiency is the predominant drawback of EMATs, which leads to the low signal to noise ratio (SNR) of the ultrasonic signal [13–16]. This drawback severely limits the application of EMATs in various fields. So it is urgently needed to make EMATs have a higher SNR and extraordinary pure ultrasound with a reasonable design. Earlier research was made based on the experimental approach, which was costly and time-consuming. Until recently, researchers established the mathematical model of EMATs, and made remarkable improvement in theoretical analysis and numerical simulation [17]. Now, there are substantial conspicuous improvements in the performance and transduction efficiency of EMATs, such as the application of high-frequency pulse generator and low-noise receiver, the introduction of phased array and the digital signal processing [18,19]. However, these methods will make electronic devices bulky. Under the high voltage environment, there are hidden dangers. For example, the static electricity is easily to result fire in gas and oil industries [20]. Sensors 2017, 17, 1106; doi:10.3390/s17051106

www.mdpi.com/journal/sensors

Sensors 2017, 17, 1106

2 of 12

Obviously, it can be obtained from Equation (1) that the square of the magnetic flux density is proportional to the SNR of EMATs with the mechanism of Lorentz force [7–15]. The SNR can be given as √   VEMAT P0 N 2 B2 A αh p exp − = (1) Vnoise D 2ZS KTβREMAT where VEMAT is the received voltage across the terminals of the coil (V), Vnoise is the received noise level (V), P0 is the power of the EMAT (W), N is the number of turns per unit length, B is the magnetic flux density (T), A is the active area of EMAT (m2 ), ZS is the acoustic impedance of the part (Ω), either longitudinal or shear wave, K is Boltzmann’s constant times electronic charge (Meter-Kilogram-Second units), T is temperature in degrees Kelvin (K), β is the bandwidth of the amplifier (Hz), REMAT is the resistance of EMAT coil (Ω), α is a geometry dependent constant, h is the lift-off distance between EMAT coil and the surface of the part (m), and D is the diameter of the coil (m). Therefore, it is an extremely effective way to improve the transducer efficiency by enhancing the magnetic flux density. Consequently, improvement in the strength of the magnetic field based on simulation and experiment methods has been widely applied to ameliorate SNR of the EMAT. Pei et al. enhanced the efficiency of meander-line-coil (MLC) EMAT by improving the magnetic field [7]. Wang et al. increased the excitation efficiency of a torsional wave PPM EMAT array for pipe inspection by optimizing the element number of the array based on 3-D FEM [8]. Kang et al. improved the signal amplitude of surface wave EMAT based on 3-D simulation analysis and orthogonal test method [9]. Xie et al. proposed a wholly analytical modelling method to study the Rayleigh waves’ beam directivity on the surface of the material [10]. Seung et al. presented a new EMAT for generation and measurement of omnidirectional shear-horizontal (SH) guided waves in metallic plates [11]. There have been previous studies of SH wave EMAT, but they mainly focused on the meander-line-coil (MLC) EMAT for surface wave generation and detection in the metallic plate or pipe [7–10]. Until now, there is hardly research about the bulk wave EMAT, which is composed of spiral-coil and cylindrical magnet, except that Julio Isla et al. presented an in-depth analysis of the bias magnetic field strength and resulted signal amplitude of different magnet configurations for shear wave EMAT on mild steel [20]. Nevertheless, in practice, the thickness measurement has a wide range of applications by bulk waves EMAT. And a higher signal amplitude can help to improve the penetration depth and the measurement accuracy of the EMAT system. Accordingly, it is of great practical significance to improve the energy efficiency of bulk wave EMAT for thickness measuring of the metallic material. The main work of this paper is to propose an improved configuration related to cylindrical magnet to strengthen the magnetic field of the spiral-coil EMAT for generation and detection of the bulk waves, which is mainly used to measure the thickness of the metallic plate. It has been reported that the signal intensity of the bulk wave is increased by changing the configuration of a magnet with a racetrack coil, but not referenced to the spiral-coil [21]. Figure 1 shows a comparison of the diagrammatic sketch of the modified magnet arrangement with the conventional one. The detailed structure of the proposed EMAT will be described in Section 2. Then, by using the axisymmetric finite element model (FEM), the distribution of the magnetic flux density with respect to the traditional and improved double-coil EMAT is analyzed by numerical simulation. Subsequently, in Section 3, the radius ratio and height of the new EMAT were optimized to find the most suitable configuration when other dimensions are constant. In the last section of the paper (Section 4), we conclude that the proposed magnet configuration can tremendously improve the signal amplitude compared with the traditional configuration by numerical simulation and experimental verification.

Sensors 2017, 17, 1106 Sensors 2017, 17, 1106

3 of 12 3 of 12

Figure 1. Schematics Schematicsofofthe theconventional conventionalspiral-coil spiral-coil EMAT EMAT and and the the improved improved double-coil double-coil EMAT. EMAT. (a,d) 3-D model; (b,e) cross-sectional view; (c,f) top view of the coil of the two configurations, (a,d) 3-D model; (b,e) cross-sectional view; (c,f) top view of the coil of the two configurations, respectively. respectively.

2. Configuration of the Double-Coil EMAT and Numerical Simulation 2. Configuration of the Double-Coil EMAT and Numerical Simulation In this section, the modified configuration of the bulk-wave EMAT configuration will be described In this modified the bulk-wave EMAT configuration will be in detail, and section, it will bethe compared withconfiguration the traditionalofconfiguration. In order to verify the rationality of described in detail, and it will be compared with the traditional configuration. In order to verify the the new configuration, a two dimensional axisymmetric finite element model (2-D FEM) is established rationality of thesymmetry. new configuration, a two dimensional finite element model FEM) due to its axial The numerical simulation axisymmetric and calculation indicate that the (2-D improved is established due to its axial symmetry. The numerical simulation and calculation indicate that the configuration dramatically heighten the magnetic flux density. improved configuration dramatically heighten the magnetic flux density. 2.1. Improved Double-Coil EMAT Design 2.1. Improved Double-Coil EMAT Design It is well known that the combination of the cylindrical magnet providing static bias magnetic is well knowntothat combination of thesample cylindrical magnet statichigh-frequency bias magnetic field It perpendicular the the surfaces of the tested and the spiralproviding coil exerting field perpendicular to the ultrasonic surfaces ofguided the tested sample and spiral coil high-frequency pulse excitation produces waves within thethe specimen [22].exerting As shown in Figure 1a, pulse excitation produces ultrasonicbulk guided waves within the As shown and in Figure 1a, the spiral-coil EMAT can generate waves in samples forspecimen thickness[22]. measurement internal the spiral-coil EMAT can generate waves in samples for thickness measurement internal defect detection of a metallic plate. bulk But the conversion efficiency is not ideal in practice.and In order to defect detection of a metallic But the conversion efficiency is not ideal in practice. to overcome this conundrum, anplate. improved magnet arrangement is proposed, which consistsInoforder a solid overcome this conundrum, an improved magnet arrangement is proposed, which consists of a solid cylindrical magnet (Magnet 1) and a hollow annular magnet (Magnet 2) with axial polarization. cylindrical magnet (Magnet 1) and hollow annular magnet (Magnet 2) with polarization. And the polarization direction of athe two magnets is placed oppositely, asaxial shown in Figure And 1b,e. theis polarization ofthe thesolid two cylinder magnets isistightly placedsurrounded oppositely, by as the shown in Figurepermanent 1b,e. It is It important todirection notice that ring-shaped important notice the1d. solid cylinder is tightly surrounded ring-shaped magnet, as to shown in that Figure In other words, the outer diameter ofbythethe solid cylindricalpermanent magnet is magnet, as shown in Figureof1d. Inannular other words, outer diameter of the solid cylindrical magnet is equal to the inner diameter the hollowthe magnet. equal to the inner diameter of the annular hollow magnet.

Sensors 2017, 17, 1106

4 of 12

To make the direction of the Lorentz force all outward along the radial direction, a double-coil is used instead of a single spiral-coil, as shown in Figure 1f. The double-coil includes two coils that are interconnected to each other. The width of the inner coil is W21 which is equal to the radii of the Magnet 1. And the width of the outer coil is W22 , which is equal to the difference between the radii of the Magnet 1 (R21 ) and Magnet 2 (R22 ). The current flow direction in the coil located below the cylindrical magnet (Magnet 1) is opposite to that of the coil under the hollow circular magnet (Magnet 2). Note that all the magnets (Neodymium Iron Boron) have the same remanence. An aluminum plate with thickness of 50 mm is taken as a specimen. Generally speaking, when two magnets are close to each other, like magnetic poles repel each other, whereas opposite poles attract each other. Therefore, the magnetic flux density of the conventional EMAT is scattered around the magnet due to reversal placement of the magnetic pole and far distance between two poles. Differently, the magnetic flux density will focus on the surface of the magnet in the case of the improved EMAT configuration, as shown in Figure 1e. This analogical configuration combined with a wound coil has been used to generate omnidirectional SH waves, but not yet combined with the spiral-coil to generate bulk waves [11]. 2.2. Numerical Simulation of the Static Magnetic Field In this section, the distribution of the magnetic flux density with respect to the conventional and modified configuration will be compared by numerical calculation. The 2-D FEM was established with COMSOL Multiphysics [23]. A 2-D axisymmetric space dimension chosen for the coil and magnet are axisymmetric about z axis. The AD/DC, magnet field module were used to generate Lorentz force and the structural mechanics module was applied to produce and propagate the bulk wave. Because of a coupling of varieties of physical field, the coupling variables need to be set up. The vector product of the magnetic flux intensity (B0 ) of the bias magnetic and eddy current density (Je ) set to be the volume force ( f L ) in the specimen for the transmitting process. The f L can be given by f L = B0 × Je

(2)

And the vector product of the magnetic flux density (B0 ), electrical conductivity of the specimen (σ) and the velocity of the charged particles (v) set to be the eddy current density (JL ) of the pulsed eddy current field for the receiving process. The JL under the surface of a conductive specimen with density JL = σ (v × B0 )

(3)

Then, the eddy current is inductively detected by the coil of the EMAT. Because the square of the magnetic flux density is proportional to the SNR of EMAT, it contributes twice in the process of transmitting and receiving for an EMAT in pulse-echo configuration [20]. The 2-D FEM consists of the magnet, coil, aluminum plate, air and infinite element domains. The side and bottom of the aluminum plate be defined as absorbing boundary. In order to ensure the consistency, the magnet of the two configurations is loaded with the same residual magnetic flux density of 1.2 T (grade: 35 N) and axial polarization, differing only in the configuration of the magnets and the structure of the coil. To give a fair comparison, the volume of the conventional magnet is equal to that of the improved one. In the conventional configuration, the radii of the magnet is 15 mm. While, in the improved configuration, the solid magnet (Magnet 1) with radii of 10 mm and the ring magnet (Magnet 2) with inner and outer radii of 10 mm and 15 mm were applied in the 2-D FEM. And all the height of the magnet is 20 mm. The coil size used in both configurations is same with a diameter of 0.51 mm and a number of turns of 30, and the material of the coil is copper wire. All parameters of the 2-D FEM are listed in Table 1. Under the two configurations, the distribution of the magnetic flux density in the surface of the specimen was illustrated after numerical calculation, as shown in Figure 2. It should be noted that the graph is a cross section of a 2-D axisymmetric graph rotated around the z axis.

Sensors 2017, 17, 1106

5 of 12

Table 1. Parameters of 2-D FEM used in this paper. Sensors 2017, 17, 1106

Object

Parameters

Symbol

Value

5 of 12

Spiral-coil Table 1. Parameters Diameterof 2-D FEM used Din 0.51 mm 1 this paper.

Object Double-coil Spiral-coil

Width of the inner coil Parameters Width of the outer coil Diameter Diameter

W21 Symbol W22 D2

10 mm 5 mm 0.51 mm

Value 0.51 mm 10 mm Width of the inner coil Radii R1 15 mm Double-coil Magnet Width of the outer coil 5 mm Height h1 20 mm Diameter 0.51 mm Radii R21 10 mm Radii 15 mm Magnet 1 Magnet Height h2 20 mm Height ℎ 20 mm Inner radii R21 10 mm 10 mm Radii Magnet 1 Magnet 2 Outer radii R22 ℎ 15 mm Height 20 mm Height h2 20 mm Inner radii 10 mm Massradii density ρ 2700 kg/m3 15 mm Magnet 2 Outer 7 Electrical conductivity σ Height ℎ 3.77 × 10 9 S/m 20 mm Young’s modulus E Al 70 × 10 Pa2700 kg/m Mass density Passion’s ratio µ 0.33 Electrical conductivity 3.77 × 107 S/m Thickness H 50 mm Al Young’s modulus 70 × 109 Pa Passion’s ratio 0.33 Thickness 50 Figure 2 shows the distribution of the magnetic flux density in two cases. Bluemm arrows represent

the direction of the magnetic field. It can be seen from Figure 2a, the magnetic flux density of the Figure 2 shows the distribution of the magnetic flux density in two cases. Blue arrows represent traditional configuration mainly scatters around thefrom magnet, perpendicular the direction of the magnetic field. It can be seen Figurewhich 2a, thedirection magnetic is flux density of the to the surfacetraditional of the sample. As the distance from the magnet is farther, the magnetic flux densitytodecreases configuration mainly scatters around the magnet, which direction is perpendicular the gradually within a sample. certain As range aroundfrom the magnet. themagnetic magnetic flux density of the new surface of the the distance the magnetHowever, is farther, the flux density decreases gradually certain range around magnet. thethe magnetic flux as density of the new configuration iswithin more aconcentrated in the the fringe andHowever, internal of magnet, shown in Figure 2b. configuration is more concentrated in the fringe and internal of the magnet, as shown in Figure 2b. flux In a certain range around the magnet, with the distance from magnet increasing, the magnetic In a certain range around the magnet, with the distance from magnet increasing, the magnetic flux density decreases sharply. Significantly, it can be seen from the cutline on the right side of the each density decreases sharply. Significantly, it can be seen from the cutline on the right side of the each photograph that the magnetic flux density of the modified configuration has a larger peak at the photograph that the magnetic flux density of the modified configuration has a larger peak at the intersection of twoofmagnets, which is helpful theconversion conversion efficiency. intersection two magnets, which is helpfultotoimprove improve the efficiency.

(a)

(b)

Figure 2. Magnetic flux density distributions in two configurations for (a) conventional and

Figure 2. Magnetic flux density distributions in two configurations for (a) conventional and (b) improved (b) improved double-coil EMAT. double-coil EMAT. In order to accurately analyze the magnitude of the magnetic flux density, the magnetic flux density vector in the surface the magnitude aluminum plate was also calculated. Figure the 3 shows the flux In order to accurately analyzeofthe of the magnetic flux density, magnetic distribution of magnetic flux density with respect to the conventional and modified EMAT. It is density vector in the surface of the aluminum plate was also calculated. Figure 3 shows the distribution important to see that the magnitude scales of Figure 3a,b are different. It can be seen from Figure 3a, of magnetic flux density with respect to the conventional and modified EMAT. It is important to the radial component of magnetic flux density ( _ ) is equal to 0 in the center of the magnet, and a see thatspike the peak magnitude Figure are different. It can beofseen fromflux Figure 3a,( the) radial appears scales near theofedge of the3a,b magnet. The axial component magnetic density _ component of magnetic flux density (B ) is equal to 0 in the center of the magnet, andthe a spike has an arched peak on the surface of thes_rmagnet and sharply decreases to 0 or even negative near edge of the magnet. The peak of the magnetic flux density ( ) occurs at the edge of the magnet, with peak appears near the edge of the magnet. The axial component of magnetic flux density (Bs_z ) has a maximum of 0.56 distribution of the magnetic density that respect an arched peak onmagnitude the surface of T. theHowever, magnetthe and sharply decreases to 0flux or even negative near the to the modified EMAT is different from that of the conventional one, as shown in Figure 3b. _ and edge of the magnet. The peak of the magnetic flux density (B ) occurs at the edge of the magnet, s

with a maximum magnitude of 0.56 T. However, the distribution of the magnetic flux density that respect to the modified EMAT is different from that of the conventional one, as shown in Figure 3b.

Sensors 2017, 17, 1106

6 of 12

Sensors 2017, 17, 1106

6 of 12

Bs_r and Bs_z have two spike peaks and arched peaks, respectively. Bs emerges a large peak at the emerges a large peak at the junction of _ have two spike peaks and arched peaks, respectively. junction two magnets, with a maximum twoof magnets, with a maximum amplitudeamplitude of 0.78 T. of 0.78 T.

(a)

(b)

Figure 3. Magnetic flux density distributions of 2-D axisymmetric model for (a) conventional and

Figure 3. Magnetic flux density distributions of 2-D axisymmetric model for (a) conventional and (b) improved double-coil EMAT on the upper surface of the specimen. (b) improved double-coil EMAT on the upper surface of the specimen.

It is worth noting that, in both configurations, the circumferential magnetic flux density (

_

) in

that perpendicular the cross section 0, perpetually. Thisflux is due to the(Bs_θ ) Itthe is direction worth noting that, in both to configurations, themaintains circumferential magnetic density direction of the magnetic in section axis. Since only the arrangement This of magnets in thepolarization direction that perpendicular to thefield cross maintains 0, perpetually. is due is to the changed, the magnetic flux density of the double-coil EMAT increases by about 40%. According to polarization direction of the magnetic field in z axis. Since only the arrangement of magnets is changed, Equationflux (1), the magnetic flux density is proportional to the by SNR. Therefore, this new configuration the magnetic density of the double-coil EMAT increases about 40%. According to Equation (1), is advantageous for improving the conversion efficiency of the transducer. the magnetic flux density is proportional to the SNR. Therefore, this new configuration is advantageous for improving the conversion efficiency of the transducer. 2.3. Simulation of the Dynamic Characteristics The ultrasonic signalsCharacteristics under two configurations are compared and analyzed in this section. 2.3. Simulation of the Dynamic Generally, there are three mechanisms influencing on the ultrasonic transmission of EMAT: Lorentz

The under two configurations are compared and in this section. forceultrasonic that appliessignals to non-ferromagnetic materials, magnetostriction force andanalyzed magnetization force, Generally, there are three mechanisms influencing on theinultrasonic of EMAT: which are suitable for ferromagnetic materials [24]. Therefore, aluminum, transmission Lorentz force plays a Lorentz force that to non-ferromagnetic materials, magnetostriction and magnetization dominant role applies in the numerical simulation of the transmission and reception force of bulk waves. Based Lorentz mechanism, the working materials process of [24]. EMATs is that the wavesforce force,on which areforce suitable for ferromagnetic Therefore, in ultrasound aluminum,bulk Lorentz propagating perpendicular to the surface of the sample is excited when the specimen is subjected to plays a dominant role in the numerical simulation of the transmission and reception of bulk waves. Lorentz force acting along the radial direction of the coil under the action of bias magnetic field. Based on Lorentz force mechanism, the working process of EMATs is that the ultrasound bulk waves In this paper, the 2-D axisymmetric FEM was established based on Lorenz force mechanism and propagating perpendicular to the surface of the sample is excited when the specimen is subjected to ignored other mechanisms for an aluminum plate as a sample to be tested, which includes both the Lorentz force acting along the radial direction of the coil under the action of bias magnetic field. coil and the specimen transmitting ultrasound guided waves. In this thethe 2-Dnumerical axisymmetric FEM established based Lorenz mechanism Topaper, facilitate analysis, thewas transmitting EMAT is on assumed toforce be installed on a and ignored other mechanisms an The aluminum plate a sample to be tested, whichand includes 200-mm-thick aluminumfor plate. dimension andas material properties of the magnet coil areboth the the coil and the specimen transmitting ultrasound guided waves. same as those of the static simulation, respectively. The lift-off distance between the magnet and the upper surface the of the aluminum analysis, plate is set the to 0.5transmitting mm. The coil is drivenisbyassumed a three-cycle-long current on To facilitate numerical EMAT to be installed signal with the peak value of 200 A, dimension which centered 2 MHz and modulated sine squared a 200-mm-thick aluminum plate. The andatmaterial properties of by thea magnet and coil window. Considering, thestatic first step is to calculate the static The bias magnetic field in the finite element are the same as those of the simulation, respectively. lift-off distance between the magnet simulation, which of requires a certain plate amount of to time establish a stable magnetic field, so the and the upper surface the aluminum is set 0.5tomm. The coil is driven by a three-cycle-long excitation signal was loaded into the model at 6 μs. In order to guarantee the smoothness and stability current signal with the peak value of 200 A, which centered at 2 MHz and modulated by a sine of the results, the step length was set to 0.1 μs. squared window. Considering, the first step is to calculate the static bias magnetic field in the finite Figure 4 shows the radial component of Lorenz force ( ) produced by conventional spiral-coil element simulation, whichone, requires a certain amount of time to establish a stable magnetic field, so the EMAT and improved respectively. And the graphs are cross section of 2-D axisymmetric graph excitation signal was loaded into the model at 6 µs. In order to guarantee the smoothness and stability rotating around z axis. Obviously, the improved double-coil EMAT can generate a consistent of thecircumferential results, the step length was set conventional to 0.1 µs. one in Figure Lorenz force as the  4b.  Figure 4 shows the radial component of Lorenz force f r L produced by conventional spiral-coil EMAT and improved one, respectively. And the graphs are cross section of 2-D axisymmetric graph rotating around z axis. Obviously, the improved double-coil EMAT can generate a consistent circumferential Lorenz force as the conventional one in Figure 4b.

Sensors 2017, 17, 1106

7 of 12

Sensors 2017, 17, 1106

7 of 12

Sensors 2017, 17, 1106

7 of 12

Figure 4. The radial Lorenz force L produced by (a) the conventional spiral-coil EMAT and Figure 4. The radial Lorenz force f r produced by (a) the conventional spiral-coil EMAT and (b) improved one. Pistol graphs of the circumferential eddy density ( ) generated by (c) conventional (b) improved one. Pistol graphs of the circumferential eddy density (Jθ ) generated by (c) conventional spiral-coil and (d) improved 50 μs. by (a) the conventional spiral-coil EMAT and Figure 4. EMAT The radial Lorenz force one atproduced spiral-coil EMAT and (d) improved one at 50 µs. (b) improved one. Pistol graphs of the circumferential eddy density ( ) generated by (c) conventional

However, there is an(d) obvious disparity about spiral-coil EMAT and improved one at 50 μs. the amplitude between radial and axial component of the Lorenz forces the change of magnet The radial component the component Lorenz However, there is anfor obvious disparity aboutconfiguration. the amplitude between radial andofaxial ( )forces produced byan the improved EMAT is dramatically enhanced about times, ascomponent shown However, there is obvious about the amplitude between radial4and axial of theforce Lorenz for the change ofdisparity magnet configuration. The radial component of the Lorenzinforce Figure 5a. by And, the axial component the Lorenz force ( )The generated surface isin greatly of the Lorenz forces for the change configuration. radial component of the Lorenz ( f r L ) produced the improved EMAT of isofmagnet dramatically enhanced about 4 times, aswaves shown Figure 5a. weakened 17 times inby Figure 5b. This result willisbe more conducive to theabout excitation and propagation force ( ) produced the improved EMAT dramatically enhanced 4 times, as shown in L And, the axial component of the Lorenz force ( f z ) generated surface waves is greatly weakened of bulk 5a. waves metal plate. The areaofmarked by the yellow lines in Figurewaves 4c,d shows the Figure And,inthe axial component the Lorenz force ( )solid generated surface is greatly 17 times in Figure 5b. This result will be more conducive to the excitation and propagation of bulk snapshots of times the circumferential eddy current ( ) generated by spiral-coil EMAT and weakened 17 in Figure 5b. This result will bedensity more conducive to the excitation and propagation wavesimproved in metalone plate. marked by the theeddy yellow soliddensity linessolid in 4c,d shows the snapshots of atin50The μs.area Itplate. is obvious that current ofFigure the improved double-coil EMAT of bulk waves metal The area marked by the yellow lines in Figure 4c,d shows the the circumferential eddy current density (J ) generated by spiral-coil EMAT and improved one at 50 µs. θcurrent is more concentrated when the guided propagate aluminum by plate. snapshots of the circumferential eddywaves densityin( the ) generated spiral-coil EMAT and It is obvious that eddy density ofeddy the improved double-coil EMAT double-coil is more concentrated improved onethe at 50 μs. Itcurrent is obvious that the current density of the improved EMAT whenisthe guided waves propagate in thewaves aluminum plate. more concentrated when the guided propagate in the aluminum plate.

(a)

(b)

Figure 5. Comparisons between (a) radial and (b) axial component of Lorenz forces under two (a) (b) configurations, respectively. Figure 5. Comparisons between (a) radial and (b) axial component of Lorenz forces under two

Figure 5. Comparisons and (b) component axial component of Lorenz forces under Moreover, in the timebetween domain, (a) the radial circumferential of the eddy current density ( ) configurations, respectively. two configurations, respectively. measured near the receiving coil is higher than that of the conventional EMAT, as shown in Figure 6.

Moreover, in the time domain, the circumferential component of the eddy current density ( ) measured near the time receiving coil is the higher than that of thecomponent conventionalof EMAT, as shown in Figure 6. (Jθ ) Moreover, in the domain, circumferential the eddy current density measured near the receiving coil is higher than that of the conventional EMAT, as shown in Figure 6.

Sensors 2017, 17, 1106 Sensors 2017, 17, 1106

8 of 12

8 of 12

Sensors 2017, 17, 1106

8 of 12

Therefore, the numerical simulation shows that the new configuration is more effective for generating Therefore, the numerical simulation shows configurationis is more effective for generating Therefore, the numerical simulation showsthat thatthe the new configuration more effective for generating bulk waves. bulk waves. bulk waves.

Figure 6. Comparisons of the circumferentialeddy eddycurrent current density density (( ))generated by and Figure Comparisons the circumferential byconventional conventional and Figure 6. 6.Comparisons ofofthe circumferential eddy current density (Jθ generated ) generated by conventional and improved configurations, respectively. improved configurations, respectively. improved configurations, respectively.

3. Optimization of the Improved Double-Coil EMAT and Experiment Verification 3. Optimization of the Improved Double-Coil EMAT and Experiment Verification 3. Optimization of the Improved Double-Coil EMAT and Experiment Verification 3.1. Optimization

Optimization 3.1.3.1. Optimization Based on the results of the numerical calculation and analysis of the previous chapter, it is Based on the results of the numerical calculation and analysis of the previous chapter, it is Based thearesults the numerical and density analysiscan of the previousby chapter, it isthe proved provedonthat higherofamplitude of thecalculation magnetic field be obtained changing proved that a higher amplitude of the magnetic fieldradius density be=obtained by changing of the of magnet. However, thedensity optimum ratiocan ( by /( − configuration )) of solidthe of that aconfiguration higher amplitude the magnetic field can be obtained changing the configuration of the1)magnet. However, the optimum radius ratio ( = /(EMAT − is unknown )) of solid magnet However, (Magnet andoptimum toroidal magnet 2) of the magnet. the radius(Magnet ratio (RR =the R21modified /(R22 −double-coil R21 )) of solid magnet (Magnet 1) magnet (Magnet 1) and toroidal magnet (Magnet 2) of the modified double-coil EMAT is unknown when themagnet maximum magnitude themodified magnetic double-coil flux density isEMAT obtained. To search the optimal result and toroidal (Magnet 2) ofofthe is unknown when the maximum when the the maximum magnitude of thethe magnetic flux density is obtained. To optimized. search the optimal under improved configuration, and radius of two magnets are That the result magnitude of the magnetic flux density is obtained. To search the optimal result underThat the improved under the improved the as and radius two2.magnets are optimized. the is divided into fiveconfiguration, cases to be considered, shown in of Table Here, the overall dimension of the configuration, the RR and radius of two magnets are optimized. That the RR is divided into five cases is divided into five cases to be (considered, as ℎshown in Table Here, the overall ofofthe improved magnet is constant = 15 mm, = 20 mm). It is 2. very interesting that dimension the influence to improved be the considered, as shown in Table 2. Here, the overall dimension of the improved magnet is constant magnet 15 mm,( ℎ ) =for 20the mm). It is very interesting that is thedifferent, influence different RR is onconstant magnetic( flux=density improved double-coil EMAT as of (R22 15 mm, h2 =on 20 mm). It is very that influence of the different RRdifferent, on magnetic in Figure 7a. Obviously, it density caninteresting be seen Figure 7b that, the amplitude of the magnetic flux as the=shown different RR magnetic flux ( from ) for thethe improved double-coil EMAT is Bs ) for flux density improved double-coil EMAT is different, asamplitude shown in of Figure 7a. after Obviously, (increases density sharply when the increases to the point ,the then decreases gradually it shown in Figure 7a.the Obviously, it can be seen from Figure 7b that, the magnetic flux reaches the maximum value. it can be seen from Figure 7b that, the amplitude of the magnetic flux density increases sharply density increases sharply when the increases to the point , then decreases gradually afterwhen it When RR is roughly equal to 2 ( = 10, gradually = 15) forafter theseitcases, therethe is amaximum maximum magnetic thereaches RR increases to the point reaches value. the maximum value.Pr , then decreases flux density in the improved EMAT. It = is wellRknown amplitude of the ultrasonic When RR roughly equal forthe these cases,there there a maximum magnetic When RRisis roughly equalto to 22 (R ( 21 = 10, = 15)that for these cases, isisagenerated maximum magnetic 22 = wave is proportional to the magnetic flux density [15]. Therefore, the SNR of the new double-coil flux density in the improved EMAT. It is well known that the amplitude of the generated ultrasonic flux density in the improved EMAT. It is well known that the amplitude of the generated ultrasonic EMAT can be significantly improved flux by selecting optimal RR. the SNR of the new double-coil wave proportional themagnetic magnetic densitythe [15]. wave is is proportional totothe flux density [15]. Therefore, Therefore, the SNR of the new double-coil EMAT can significantlyimproved improvedby byselecting selecting the the optimal optimal RR. EMAT can bebe significantly RR.

(a)

(b)

Figure 7. The relationship between the RR of the modified magnet and (a) the magnetic flux density; (b) (a) and (b) the maximum amplitude of the magnetic flux density.

Figure Therelationship relationshipbetween between the the RR RR of of the magnetic flux density; Figure 7. 7. The the modified modifiedmagnet magnetand and(a)(a)the the magnetic flux density; and (b) the maximum amplitude of the magnetic flux density. and (b) the maximum amplitude of the magnetic flux density.

Sensors 2017, 17, 1106

9 of 12

Sensors 2017, 17, 1106

9 of 12

Table 2. Cases of the RR and the variation of the magnetic flux density. Sensors 2017, 17, 1106 Table 2. Cases of the RR and the variation of the magnetic flux density.

Case Case 11 Case 22 1 33 2 44 53 5 4 5

9 of 12

((T)) Increment Increment(%) (%) − R21of(mm) (mm) (mm) (mm) R21(mm) (mm) R22Cases − R22 (mm) h2 (mm) RR B Table 2. the RR and the variation of the magnetic flux _ density. s_max 33 (mm) 55 3 7.5 7.5 5 1010 12 7.5 12 10 12

−

12 12 (mm) 10 10 12 7.5 7.5 5510 37.5 3 5 3

(mm)

(mm)

1515

20 20

15

20

1/4 1/4 1/2 1/2 1/4 11 1/2 22 41 4 2 4

0.568 0.568 ( ) 0.576 0.576 0.568 0.712 0.712 0.576 0.78 0.78 0.716 0.712 _

0.716 0.78 0.716

0.14 0.14 Increment (%) 2.86 2.86 0.14 27.14 27.14 2.86 39.29 39.29 27.86 27.14 27.86 39.29 27.86

According Accordingtotothe theprevious previousanalysis, analysis,the themagnetic magneticflux fluxdensity densitygenerated generatedunder underdifferent differentradius radiusof ofimproved the According improved double-coil EMAT were compared. Here RR is set to 2, since the magnitude of the the double-coil EMAT were compared. Here RR is set to 2, since the magnitude of the flux to the previous analysis, the magnetic flux density generated under different radius flux density is maximum at this point. As shown in Figure 8a, the magnetic flux density of the density is maximum at this point. As shown in Figure 8a, the magnetic flux density of the modified of the improved double-coil EMAT were compared. Here RR is set to 2, since the magnitude of the modified EMAT tends to grow rapidly with of the radius of the(R magnet (density ) when theof EMAT to is grow rapidly the increase ofincrease theinradius of8a, the magnet when the radius 22 )flux flux tends density maximum atwith this point. As the shown Figure the magnetic of the radius of the magnet (Magnet 2) changes from 6 mm to 27 mm. It is necessary to be reminded that the magnet (Magnet 2) changes from 6 mm to 27 mm. It is necessary to be reminded that RR = 2 is modified EMAT tends to grow rapidly with the increase of the radius of the magnet ( ) when the = 2 is maintained continuously. maintained radius of continuously. the magnet (Magnet 2) changes from 6 mm to 27 mm. It is necessary to be reminded that = 2 is maintained continuously.

(a) (a)

(b) (b)

Figure The relationship between radius of modified the modified magnet and the magnetic flux Figure 8. 8. The relationship between the the radius of the magnet and (a) the(a) magnetic flux density; Figure and 8. The relationship between the radius of the modified magnet and (a) the magnetic flux density; (b) the maximum amplitude of the magnetic flux density. and (b) the maximum amplitude of the magnetic flux density. density; and (b) the maximum amplitude of the magnetic flux density.

The comparison between the maximum magnetic flux density and the height of the magnet can TheThe comparison the maximum magnetic flux and theheight height themagnet magnet can comparison the maximum magneticincreases flux density density and the ofofheight the be seen from Figure between 9bbetween that the maximum amplitude with increasing the fromcan the berange seen from Figure 9b that the maximum amplitude increases with increasing the height from the be seen 9b that the amplitude increasesof with the density height from the of 3from to 27Figure mm. However, themaximum growth rate of the amplitude the increasing magnetic flux becomes range ofafter 3ofto3 to 2727 mm. However, the of amplitude of the theresult magnetic flux density becomes range mm. However, the growth rate ofathe the of magnetic flux density small the height increases togrowth point rate . As result, the optimal with respect to becomes height is small after the height increases to point P . As a result, the optimal result with respect to height small after the height increases to point . As a result, the optimal result with respect to height is is h approximately 20 mm. approximately 20 20 mm. approximately mm. v

v

(a) (a)

(b) (b)

Figure 9. The relationship between the height of the modified magnet and (a) the magnetic flux Figure 9. The relationship between the height of the modified magnet and (a) the magnetic flux Figure 9. The between the height of the modified magnet and (a) the magnetic flux density; density; andrelationship (b) the maximum amplitude of the magnetic flux density. density; and (b) the maximum amplitude of the magnetic flux density. and (b) the maximum amplitude of the magnetic flux density.

The magnet with radii of 15 mm in the conventional configuration, the solid magnet (Magnet 1) with radii of 10 mm and the ring magnet (Magnet 2) with inner and outer radii of 10 mm and 15 mm in the improved configuration were applied in the experiment. And all the height of the magnet is 20 mm. A double-coil is applied, as shown in Figure 1f, since it is easy to produce a consistent radial Lorenz force is Sensors 2017, 17, with 1106 an improved magnet configuration. And the coil dimension of the experiment 10 of 12 the same as that of the previous simulation. The EMAT system is driven by a one-cycle burst signal at 2 MHz to generate and receive bulk waves for measuring the thickness of the aluminum block. All 3.2. Experiment Analysis experiments were performed in a pulse-echo mode for thickness measurement. Figure 10 shows received signals the aluminum plate withdouble-coil same thickness (50 mm) In order to verify the the characteristic andfrom performance of the improved EMAT, several under the conventional configuration (Figure 10a), the improved magnet configuration without using contrastive experiments between the traditional configuration and the improved one are done. the double-coil (Figure thetheimproved magnet configuration with the double-coil The magnet with radii10b), of 15and mm in conventional configuration, the solid magnet (Magnet 1) (Figure 10c), respectively. Apparently, it is nearly 38 μs that the first reflected wave themm back with radii of 10 mm and the ring magnet (Magnet 2) with inner and outer radii of 10 mmfrom and 15 in wall of the aluminum platewere in allapplied experimental arrangements. is the worth noting that the amplitude the improved configuration in the experiment. AndItall height of the magnet is 20 mm. scale of Figureis10a–c is different. A double-coil applied, as shown in Figure 1f, since it is easy to produce a consistent radial Lorenz Compared with Figure 10a,configuration. not only the amplitude ofdimension the received signal increasedis by force with an improved magnet And the coil of the experiment theabout same 25%, also the noise distortion is reduced Figure burst 10b. In theatimproved as thatbut of the previous simulation. The EMAT systemsignificantly is driven by ainone-cycle signal 2 MHz to configuration, the noise level is similar regardless whether the coilaluminum is changedblock. or not, the amplitude generate and receive bulk waves for measuring theofthickness of the All experiments significantly increases when using the double-coil, as shown in Figure 10b,c. Under the traditional were performed in a pulse-echo mode for thickness measurement. configuration, the noise is ±0.5 V asfrom shown in Figure 10a, noise amplitude to Figure 10 shows the level received signals the aluminum platethe with same thicknesscorresponds (50 mm) under approximately 1/5 of the signal generated by the improved configuration, as shown in Figure 10c. the conventional configuration (Figure 10a), the improved magnet configuration without using the Furthermore, under10b), the configuration of Figure 10c,configuration the amplitude peak the largest (Figure in the three double-coil (Figure and the improved magnet with theisdouble-coil 10c), cases. According to Equation (1), the logarithm of the SNR (20log( / )) under three cases respectively. Apparently, it is nearly 38 µs that the first reflected wave from the back wall of the were calculated listed in Table 3. It can be seen that theworth SNR of the improved configuration with aluminum plateand in all experimental arrangements. It is noting that the amplitude scale of using double-coil is 20.58 dB, which is 5 times bigger than that of the conventional one. Figurethe 10a–c is different.

Figure Figure 10. 10. Comparisons Comparisons of of pulse pulse echo echo signals signals from from the the back back of of the the 50-mm-thick 50-mm-thick aluminum aluminum block. block. (a) The conventional configuration; (b) The improved configuration without using the double-coil; (a) The conventional configuration; (b) The improved configuration without using the double-coil; (c) configuration with with using using the the double-coil. double-coil. (c) The The improved improved configuration

Compared with Figure 10a, not only the amplitude of the received signal increased by about 25%, but also the noise distortion is reduced significantly in Figure 10b. In the improved configuration, the noise level is similar regardless of whether the coil is changed or not, the amplitude significantly increases when using the double-coil, as shown in Figure 10b,c. Under the traditional configuration, the noise level is ±0.5 V as shown in Figure 10a, the noise amplitude corresponds to approximately 1/5 of the signal generated by the improved configuration, as shown in Figure 10c. Furthermore, under the configuration of Figure 10c, the amplitude peak is the largest in the three cases. According to

Sensors 2017, 17, 1106

11 of 12

Equation (1), the logarithm of the SNR (20 log(VEMAT /Vnoise )) under three cases were calculated and listed in Table 3. It can be seen that the SNR of the improved configuration with using the double-coil is 20.58 dB, which is 5 times bigger than that of the conventional one. Table 3. Comparisons of SNR in three experiments. Cases

SNR (dB)

Conventional configuration Improved configuration without using double-coil Improved configuration with using double-coil

4.08 13.96 20.58

4. Conclusions This paper presents an improved spiral-coil EMAT for the generation and detection of bulk waves to address the shortcoming of conventional spiral-coil EMAT which is with lower transmission efficiency and higher noise level. Instead of a single cylinder magnet, the improved EMAT configuration consists of an annular magnet (Magnet 2) and a cylinder magnet (Magnet 1) surrounded by the annular one, which are placed oppositely each other in the direction of polarization. Because the transmission medium of the bulk wave is an aluminum block, 2-D axisymmetric FEM of the two configurations with Lorenz force mechanism were established, respectively. The numerical simulation results show that the maximum amplitude of the magnetic flux density in the surface of the improved magnet is 40% higher than that of the conventional one. And the amplitude of the radial component of Lorenz force generating the bulk wave is increased, but that of the axial component of Lorenz force producing the surface wave is substantially reduced. Based on the simulation for optimizing the dimension of the improved magnet, it was found that there exists an optimal RR which makes the magnetic flux density reach the maximum. When the height and the lift-off distance of the new magnet arrangement is 20 and 0.5 mm, respectively, the RR changes from 1/4 to 4, the maximum amplitude of the magnetic flux density increases firstly and then decreases slowly after reaching the peak. Therefore, the optimum RR is found to be approximately 2 as the radius of the cylindrical magnet (Magnet 1) is between 3 and 12 mm. Furthermore, it can be concluded that improving the arrangement of the magnet and applying the double-coil can effectively improve the low energy conversion efficiency and SNR. According to the experimental results, the improved double-coil EMAT configuration with the combination of Magnet 1 and Magnet 2 yields twice larger amplitude of the pulse-echo signal and less noise disturbance than the conventional one with a single cylinder magnet. It is worth emphasizing that the improved magnet configuration has the function of noise reduction regardless of whether the double-coil is used or not. But it is helpful to improve the amplitude of the pulse-echo signal and conversion efficiency when the double-coil is used. Acknowledgments: This work is supported by the National Natural Science Foundation of China (No. 51374264) and the basic and frontier research project of Chongqing City (No. CSTC2014jcjA50003). Author Contributions: Xiaojuan Jia contributed to the simulation research, designing the experiments, writing and revising the manuscript. Qi Ouyang contributed to the theoretical research and helped revising the manuscript. Xinglan Zhang gave help in the result analysis. Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2.

Xie, Y.; Yin, W.; Liu, Z.; Peyton, A. Simulation of ultrasonic and EMAT arrays FEM and FDTD. Ultrasonics 2015, 66, 154–165. [CrossRef] [PubMed] Thring, C.B.; Fan, Y.; Edwards, R.S. Focused Rayleigh wave EMAT for characterisation of surface-breaking defects. NDT E Int. 2016, 81, 20–27. [CrossRef]

Sensors 2017, 17, 1106

3. 4.

5. 6.

7. 8.

9. 10. 11. 12. 13.

14. 15. 16.

17. 18.

19. 20. 21. 22. 23. 24.

12 of 12

Hernandez-Valle, F.; Dixon, S. Initial tests for designing a high temperature EMAT with pulsed electromagnet. NDT E Int. 2010, 43, 171–175. [CrossRef] Kogia, M.; Gan, T.H.; Balachandran, W.; Livadas, M.; Kappatos, V.; Szabo, I.; Mohimi, A.; Round, A. High Temperature Shear Horizontal Electromagnetic Acoustic Transducer for Guided Wave Inspection. Sensors 2016, 16, 582. [CrossRef] [PubMed] Mirkhani, K.; Chaggares, C.; Masterson, C.; Jastrzebski, M.; Dusatko, T.; Sinclair, A.; Shapoorabadi, R.J.; Konrad, A.; Papini, M. Optimal design of EMAT transmitters. NDT E Int. 2004, 37, 181–193. [CrossRef] Dutton, B.; Boonsang, S.; Dewhurst, R.J. A new magnetic configuration for a small in-plane electromagnetic acoustic transducer applied to laser-ultrasound measurements: Modelling and validation. Sens. Actuators A Phys. 2006, 125, 249–259. [CrossRef] Pei, C.; Zhao, S.; Xiao, P.; Chen, Z. A modified meander-line-coil EMAT design for signal amplitude enhancement. Sens. Actuators A Phys. 2016, 247, 539–546. [CrossRef] Wang, Y.; Wu, X.; Sun, P.; Li, J. Enhancement of the excitation efficiency of a torsional wave PPM EMAT array for pipe inspection by optimizing the element number of the array based on 3-D FEM. Sensors 2015, 15, 3471–3490. [CrossRef] [PubMed] Kang, L.; Dixon, S.; Wang, K.; Dai, J. Enhancement of signal amplitude of surface wave EMATs based on 3-D simulation analysis and orthogonal test method. NDT E Int. 2013, 59, 11–17. [CrossRef] Xie, Y.; Liu, Z.; Deng, P.; Yin, W. Directivity analysis of Meander-Line-Coil EMATs with a wholly analytical method. Ultrasonics 2016, 73, 262–270. [CrossRef] [PubMed] Hong, M.S.; Park, C.I.; Kim, Y.Y. An omnidirectional shear-horizontal guided wave EMAT for a metallic plate. Ultrasonics 2016, 69, 58–66. Huang, S.; Zhao, W.; Zhang, Y.; Wang, S. Study on the lift-off effect of EMAT. Sens. Actuators A Phys. 2009, 153, 218–221. [CrossRef] Murayama, R.; Makiyama, S.; Kodama, M.; Taniguchi, Y. Development of an ultrasonic inspection robot using an electromagnetic acoustic transducer for a Lamb wave and an SH-plate wave. Ultrasonics 2004, 42, 825–829. [CrossRef] [PubMed] Kang, L.; Wang, S.; Jiang, T.; Zhai, G. Optimal Design of Lamb Wave Electromagnetic Acoustic Transducers for Improving Their Excitation Performance. Jpn. J. Appl. Phys. 2011, 50. [CrossRef] Maclauchlan, D.; Clark, S.; Cox, B.; Doyle, T.; Grimmett, B.; Hancock, J.; Hour, K.; Rutherford, C. Recent Advancements in the Application of EMATs to NDE; BWXT Services: Lynchburg, VA, USA, 2004. Wang, S.; Huang, S.; Zhang, Y.; Zhao, W. Multiphysics Modeling of a Lorentz Force-Based Meander Coil Electromagnetic Acoustic Transducer via Steady-State and Transient Analyses. IEEE Sens. J. 2016, 16, 6641–6651. [CrossRef] Muscia, R. Equivalent magnetic charge in helicoidal magnets. J. Appl. Phys. 2008, 104, 103916. [CrossRef] Wang, S.; Su, R.; Chen, X.; Kang, L.; Zhai, G. Numerical and experimental analysis of unidirectional meander-line coil electromagnetic acoustic transducers. IEEE Trans. Ultrason. Ferroelectr. Freq. Control 2013, 60, 2657–2664. [CrossRef] [PubMed] Wang, S.; Kang, L.; Li, Z.; Zhai, G.; Zhang, L. 3-D modeling and analysis of meander-line-coil surface wave EMATs. Mechatronics 2012, 22, 653–660. [CrossRef] Isla, J.; Cegla, F. Optimization of the Bias Magnetic Field of Shear Wave EMATs. IEEE Trans. Ultrason. Ferroelectr. Freq. Control 2016, 63, 1148–1160. [CrossRef] Ohtaki, K.; Uchimoto, T.; Takagi, T. Material characterization of cast irons with an EMAT/EC dual probe. Trans. Jpn. Soc. Mech. Eng. A 2010, 33, 1135–1141. Hirao, M.; Ogi, H. EMATs for Science and Industry: Noncontacting Ultrasonic Measurements; Kluwer: Norwell, MA, USA, 2003. Introduction to the AC/DC Module; COMSOL: Burlington, MA, USA, 2013. Ogi, H. Field dependence of coupling efficiency between electromagnetic field and ultrasonic bulk waves. J. Appl. Phys. 1997, 82, 3940–3949. [CrossRef] © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).