Isotropic and selective acoustic electromagnetic transducer with a Lamb mode of interest and method for designing such a transducer

Abstract

A selective acoustic electromagnetic transducer (400) for a Lamb wave propagation mode of interest propagating in a structure, comprising a circularly shaped magnet (401) and a coil (402) comprising several independent concentric circular portions and a processing unit (403) connected to the ends of each circular portion, the processing unit (403) being configured to generate a composite signal equal to a linear combination of the signals measured on each circular portion, the coefficients of the linear combination being chosen so as to minimize the sensitivity of the coil (402) to a set of Lamb modes excluding the mode of interest. Figure 4a

Description

Title of the invention: Isotropic and selective acoustic electromagnetic transducer with a Lamb mode of interest and method for designing such a transducer

[0001] The invention relates to the field of acoustic electromagnetic transducers also called EM AT (ElectroMagnetic Acoustic Transducer) sensors.

[0002] This type of sensor is notably used in the field of non-destructive testing and integrated health monitoring by guided waves of structures.

[0003] The invention relates more specifically to an acoustic electromagnetic transducer that is both isotropic and selective for a single Lamb wave propagation mode. The transducer according to the invention is also reconfigurable according to changing environmental conditions or a desired mode change. The invention also relates to a method for designing such a transducer.

[0004] In the field of non-destructive testing and integrated health monitoring by guided waves, there are control algorithms which require the selection of a single mode of propagation of guided waves as well as an isotropy of the sensors.

[0005] More specifically, in thin walls, there are several types of guided wave propagation modes: antisymmetric (A) or symmetric (S) Lamb modes, as well as shear (SH) modes. The invention aims to design a sensor that is selective for a single Lamb mode while remaining insensitive to the other modes (A, S, and SH).

[0006] There is therefore a need to design an acoustic electromagnetic transducer that is isotropic and selective to a given Lamb propagation mode while being insensitive to other propagation modes.

[0007] The transducer must also be easily reconfigurable in order to compensate for possible changes in environmental conditions which may impact the sensitivity of the coil or to change the propagation mode of interest without completely redesigning the coil.

[0008] The state of the art includes solutions which only partially address the aforementioned technical problem and / or which present certain drawbacks.

[0009] Reference [1] describes a first type of transducer made from a meander coil. An example of such a transducer is shown in [Fig. 1]. It consists of a meander coil 101 and a magnet 102 and is positioned on a part to be inspected 103.

[0010] This type of coil 101 is composed of several parallel wires, spaced at half the wavelength of the propagation mode of interest, and with opposite current directions. This solution allows for the selection of only one propagation mode of guided waves (modal selectivity) but does not allow for isotropy. It has the disadvantage of being, on the contrary, highly directional.

[0011] Reference [2] describes a second type of transducer based on the use of a The so-called "pancake" coil. An example of this transducer is shown in [Fig. 2]. It consists of two spiral coils 201, 202, each wound in a different plane, the two coils being joined by a via 203. This type of coil consists of a spiral of wires whose turns are very close together to form a disk containing a maximum number of wires in an orthoradial direction. This second solution allows for isotropic transmission / reception but does not allow the selection of a single propagation mode.

[0012] Reference [3] describes a third type of transducer based on a circular meander coil, composed of a set of concentric circles, with two adjacent circles having opposite current directions and a spacing of half a wavelength of the mode of interest. An example of such a coil is shown in [Fig. 3]. In this example, two coils 301, 302 are made on two superimposed layers and connected to each other by means of a via 303.

[0013] This type of coil guarantees good isotropy; however, modal selectivity is less efficient than with the first type of meandering coil. Indeed, the half-wavelength spacing between two adjacent circles is only maintained for a wave propagation direction that is along the axis of the coil radius.

[0014] In this case, a large number of circles must be used to obtain satisfactory modal selectivity, for example up to a diameter of 6.5 wavelengths in reference [3], which leads to space constraints at low frequency (for an aluminium plate 3 mm thick, the wavelength is about 20 mm for the A0 mode at 70 kHz, which implies using a sensor of at least 13 cm in diameter).

[0015] To improve the modal selectivity of the transducer, a judicious choice of magnet geometry can be made to make the sensor more sensitive to certain directions of movement and therefore to certain modes. This is, for example, what is done in publication [3], in which the magnets used are chosen to have a magnetic field with an essentially radial direction in the plate, which makes the sensor more sensitive to mode A0 (the mode of interest in the case of the article) and less sensitive to mode S0.

[0016] Apart from the size, the second drawback is that this type of sensor is very specific to a guided mode of interest, which also implies a range of conditions environmental conditions in which the sensor has the expected properties, which leads to limitations in an application context for integrated health monitoring.

[0017] Reference [4] describes yet another type of solution that reduces the size problem of the previous solution while maintaining good modal selectivity. It is based on a circular meander coil of the type presented previously, with the addition of a magnetic concentrator below the magnet. The geometric properties of this concentrator are numerically optimized to eliminate the sensor's sensitivity to all propagation modes except the one to be isolated. A magnetic concentrator is a ferromagnetic component that concentrates and directs the magnetic field.

[0018] As with the previous solution, this sensor is designed for a specific propagation mode and for specific parameters of the instrumented structure, which can vary with environmental conditions and therefore reduces its range of use in an integrated health control context.

[0019] Furthermore, this solution has the drawback that the geometry of the magnetic concentrator cannot be modified once manufactured. Thus, any differences between the numerical optimization results and the experimental realization of the sensor (in particular the appearance of modes intended to be eliminated) cannot be corrected by modifying the concentrator geometry accordingly. Consequently, the numerical simulations must be based on an extremely comprehensive and accurate model, and the sensor's performance cannot be improved if it proves not to be sufficiently representative. Moreover, the sensor's modal selectivity is calculated and therefore applicable only for given structural parameters, which limits its use to a certain range of environmental conditions outside of which these parameters can change and impact the sensor's performance.

[0020] The proposed invention overcomes the drawbacks of the aforementioned prior art solutions by using a sensor comprising a circular coil with several independent circular segments and as many acquisition channels as there are circular segments. A processing unit coupled to the acquisition channels performs a linear combination of the signals captured on each segment of the coil so as to produce a signal characteristic of a single Lamb mode, in which the other modes are filtered out. In this way, the transducer is made selective for this Lamb mode and insensitive to the other modes.

[0021] The invention also proposes a method for designing such a coil which aims to determine the coefficients of the linear combination implemented in the processing unit.

[0022] The invention has the particular advantage of offering an isotropic sensor with modal selectivity that is compact and reconfigurable if the environmental conditions of the sensor change or if the user wishes to modify the Lamb mode of interest for a different application.

[0023] The invention thus relates to an acoustic electromagnetic transducer selective to a Lamb wave propagation mode of interest propagating in a structure, comprising a magnet with circular geometry and a coil comprising several independent concentric circular portions and a processing unit connected to the ends of each circular portion, the processing unit being configured to generate a composite signal equal to a linear combination of the signals measured on each circular portion, the coefficients of the linear combination being chosen so as to minimize the sensitivity of the coil to a set of Lamb modes excluding the mode of interest.

[0024] According to a particular aspect of the invention, each circular portion is made up of a winding of several turns.

[0025] According to a particular aspect of the invention, the coils of a circular portion are superimposed on each other.

[0026] According to a particular aspect of the invention, the circular portions have diameters substantially equal to (k+l)X / 2 where X is the wavelength of the mode of interest and k is a positive or zero integer.

[0027] According to a particular aspect of the invention, the number of circular portions is at least equal to a specified number of Lamb wave propagation modes.

[0028] In one embodiment, the acoustic electromagnetic transducer according to the invention comprises a magnetostrictive patch.

[0029] The invention also relates to a method for designing the acoustic electromagnetic transducer according to the invention, the method comprising the steps of: - Determining the sensitivity of each circular portion of the coil to several respective Lamb wave propagation modes, - Define a cost function dependent on a linear combination of the sensitivities of each circular portion to a set of Lamb wave propagation modes excluding the Lamb mode of interest, - Find the values ​​of the coefficients of the linear combination that minimize the cost function, - Implement (503) in the processing unit said linear combination of the signals measured from the coefficients obtained

[0030] According to a particular aspect of the invention, the cost function depends on the sum, over the set of Lamb wave propagation modes excluding the Lamb mode of interest, the modulus or the squared modulus of the sum of the sensitivities of each circular portion, weighted by the coefficients of the linear combination.

[0031] According to a particular aspect of the invention, the sensitivities are determined for several frequency values ​​in a given frequency range and the cost function is equal to the double sum over the set of Lamb wave propagation modes excluding the Lamb mode of interest and over the set of frequencies, of the magnitude or the squared magnitude of the sum of the sensitivities of each circular portion, weighted by the coefficients of the linear combination.

[0032] According to a particular aspect of the invention, the cost function is divided by the modulus or the squared modulus of the sum of the sensitivities to the Lamb mode of interest, of each circular portion, weighted by the coefficients of the linear combination.

[0033] According to a particular aspect of the invention, the sensitivities are determined for several frequency values ​​in a given frequency range and said modulus or squared modulus of the sum of the sensitivities is integrated over said frequency range.

[0034] According to a particular aspect of the invention, the sensitivity of each of the circular portions is determined analytically via a formula depending at least on the radius of the circular portion, the components of the propagation mode, its wavelength and the components of the magnetic field to which the circular portion is subjected or by simulation.

[0035] According to a particular aspect of the invention, the sensitivity of each of the circular portions is determined experimentally by means of the following steps: - Perform a guided wave measurement for each circular portion of the coil, - For each of the said Lamb wave propagation modes, isolate the propagation mode on the measurement and determine the sensitivity of the circular portion of said propagation mode from the isolated portion of the measurement.

[0036] According to a particular aspect of the invention, guided wave measurement is carried out by: - ​​positioning said transducer and a second transducer on a sample, the second transducer being positioned at a predetermined distance from said transducer as a function of the group velocity of at least one of said propagation modes, - Measuring the guided wave emitted by the second transducer and propagated in the sample.

[0037] According to a particular aspect of the invention, guided wave measurement is carried out by: - positioning said transducer and a second transducer on a sample, the second transducer being configured to selectively emit one of said propagation modes, - Measuring the guided wave emitted by the second transducer and propagated in the sample.

[0038] Other features and advantages of the present invention will become more apparent from the following description in relation to the following accompanying drawings.

[0039] [Fig. 1] represents a diagram of a transducer comprising a meander coil according to the prior art,

[0040] [Fig.2] represents a diagram of a spiral coil according to the prior art,

[0041] [Fig.3] represents a diagram of a circular meander coil according to the prior art,

[0042] [Fig.4a] represents a diagram of an acoustic electromagnetic transducer according to an embodiment of the invention,

[0043] [Fig.4b] represents a diagram of a coil according to an embodiment of the invention,

[0044] [Fig. 4c] represents a diagram of an alternative embodiment of the coil of [Fig. 4b],

[0045] [Fig.5] represents a flowchart of a method for designing a transducer according to an embodiment of the invention,

[0046] [Fig.6a] represents a temporal measurement of an ultrasonic wave by means of a sensor comprising a coil including a circular portion of a first diameter,

[0047] [Fig.6b] represents a temporal measurement of an ultrasonic wave by means of a sensor comprising a coil including a circular portion of a second diameter,

[0048] [Fig.7] represents a diagram showing a signal resulting from an optimized combination of the signals in Figures 6a and 6b to obtain a selective coil for a given propagation mode,

[0049] Figure 4a shows a schematic of an acoustic electromagnetic transducer 400 according to an embodiment of the invention. The transducer comprises a circular magnet 401 coupled to a circular coil 402 positioned opposite the magnet. The coil is connected to a processing unit 403.

[0050] Figure 4b shows a top view of a first embodiment 412 of the coil, which consists of several concentric circular portions Ci, C2, C3 (three in the example of Figure 4b), each of which is independently connected to the processing unit 403. In other words, the different circular portions are not connected to each other and operate independently. In the example of Figure 4b, each circular portion consists of a single turn.

[0051] Figure 4c represents an alternative embodiment of the coil in Figure 4b, in which each circular portion consists of a winding of several turns. Preferably, the number of turns is the same for each circular portion, but it may also be different. An advantage of the embodiment in Figure 4c is improved sensor sensitivity.

[0052] Advantageously, the turns of the same circular portion are formed in the same plane with the smallest possible inter-turn spacing. Alternatively, the turns are formed in several superimposed planes, for example they are printed on several layers of dielectric material such as Kapton and are connected to each other by vias passing through the layers.

[0053] According to one embodiment, the circular portions Ci, C2, C3 have an average diameter equal to (λ + 1 / 2)2, where 2 is the wavelength of the wave for the propagation mode of interest and the operating frequency, and k is a positive or zero integer. In the case of a frequency range of interest, the wavelength is calculated at the center frequency of the range.

[0054] Alternatively, the diameters of the circular portions can take other values ​​that are not multiples of the wavelength.

[0055] The transducer 400 thus produced is isotropic due to its circular geometry. It is insensitive to the shear propagation modes SH0, SH1,... due to the circular symmetry of the coil and the magnet.

[0056] In the case where the structure to be inspected is not in a conductive material, a magnetostrictive patch is positioned between the transducer coil and the structure.

[0057] The processing unit 403 is programmed to perform a linear combination of the signals respectively measured on each circular portion of the coil via a set of weighting coefficients whose values ​​are optimized so as to make the transducer 400 sensitive to a single mode of propagation of Lamb waves, for example an antisymmetric mode A or a symmetric mode S.

[0058] Figure 5 describes, in a flowchart, the steps for implementing a design method for the transducer 400 according to the invention. The method aims to determine the weighting coefficients that are programmed into the processing unit 403 to generate an overall measurement signal at the coil that is sensitive to a single mode of interest.

[0059] The method begins at step 500 with the selection of the number of circular sections of the coil. The number of circular sections N is chosen to be at least equal to the total number of Lamb wave propagation modes in the structure to be inspected, and according to the target frequency range. For example, for low-frequency operation, only the guided modes A0, S0, and SH0 exist; therefore, the minimum number of circular sections is 2 in order to render The coil is sensitive to one of the two modes AO, SO, and insensitive to the other. Due to its circular symmetry, the coil is naturally insensitive to the SHO shear mode.

[0060] In step 501, for each of the circular portions, its sensitivity to each mode of propagation of Lamb waves likely to be present is then determined as a function of the frequency range and the structure to be inspected.

[0061] If the circular portion comprises a single turn (embodyment of [Fig. 4b]), the sensitivity is that of the turn. If the circular portion comprises several turns (embodyment of [Fig. 4c]), the sensitivity is calculated for all the turns of the circular portion.

[0062] Several methods are possible to estimate the sensitivity of a circular portion to a Lamb mode.

[0063] A first variant of the embodiment of step 501 consists of determining the sensitivity analytically, for example by calculating beforehand the Lamb modes existing in the wall of the structure that one wishes to inspect and in the range of frequencies of interest.

[0064] The following formula can be used based on a model where the eddy currents and ferromagnetic effects (in the case where the material of the structure is ferromagnetic) created by the passage of the current in the coil are concentrated linearly below it, on the surface of the wall of the structure at z=0 (skin effect).

[0065] E^f) = «[fl 0) f)Bz(Rh O)cos0 ] RM ] (1) = ^4^(0, / )6,(^0) i ¢7^(0, / )6.(^,0) J^)]

[0066] E^f) is the sensitivity of a circular portion indexed by £ to a Lamb mode of index met at a frequency f, for one turn. This sensitivity is to be multiplied by the number of turns of the circular portion if this is not 1. [°°671 and f) SOnt 'CS coPosantes respectively longitudinal and normal mode of propagation m,

[0068] is the wavelength at the given frequency, j^ / cle radius of the portion circular (for example Rk= (^+l / 2)2,„( / ) ),

[0069] S denotes an integration variable which describes the portion of the coil which is integrated into equation (1),

[0070] Br(r, z) and Bz(r, z) are respectively the radial and normal components of the magnetic field of the magnet, Jn are the Bessel functions of the first kind, and a is a predefined constant depending on the height of the coil.

[0071] Without departing from the scope of the invention, any other model for calculating the sensitivity of a circular portion of a coil to a Lamb wave propagation mode may be considered.

[0072] A second embodiment of step 501 consists of determining the sensitivity by numerical simulation, for example using multi-physics simulation software that takes into account all electromagnetic (eddy currents, ferromagnetism) and mechanical (wave propagation) phenomena. The sensitivity of each circular portion for each mode at a given frequency is then defined as the complex ratio (amplitude and phase shift) between the signal in the circular portion of the coil and the amplitude of the mode.

[0073] A third alternative embodiment of step 501 consists of determining the sensitivity by experimental measurements.

[0074] Two sub-variants are also conceivable for determining sensitivity by experimental measurements.

[0075] A first measurement sub-variant consists of using a first transducer of the type described in [Fig. 4a] operating in reception and a second conventional transducer operating in transmission. The second transducer is used to emit a guided Lamb wave in a chosen structure, and the first transducer is used to measure the wave, after its propagation in the structure, independently through each of the circular portions.

[0076] It is then necessary to separate the different Lamb modes, which are indistinguishable to the same extent. One solution for this is to take advantage of the difference in group velocity or energy velocity between the Lamb modes, which makes it possible to separate the wave packets between the different propagation modes in the time domain. This solution requires a sufficiently large propagation structure to ensure a sufficient distance between the two transducers relative to the group velocity (or energy velocity) of the modes, so as to prevent the first reflections of the faster modes from mixing with the direct paths of the slower modes.

[0077] A second sub-variant consists of using as a second transducer, a sensor capable of selectively emitting each of the modes alone.

[0078] In the case where the instrumented wall is flat and homogeneous (or with planar symmetry) and the frequency range of interest is small enough that only the Lamb modes A0 and S0 exist, for example, the use in phase or in opposite phase of two piezoelectric pellets glued on either side of the wall of the structure to be inspected is sufficient to separate these two modes.

[0079] Otherwise, selective actuators such as a state-of-the-art EMAT meander coil sensor can be used.

[0080]

[0081]

[0082]

[0083]

[0084]

[0085]

[0086]

[0087]

[0088]

[0089]

[0090]

[0091] Sensitivity is, for example, estimated directly by selecting a wave packet over a time measurement and calculating the signal energy within the selected time window. The integral of the square of the signal over the corresponding time window is then calculated. In step 502, a cost function is then determined from a linear combination of the sensitivities calculated in step 501. The cost function is characteristic of the overall sensitivity of the coil to propagation modes other than the mode of interest. It depends on linear combinations of the sensitivities calculated for each circular portion VN, r-, where Ek(i) is the sensitivity, calculated in step 501, of the vr 1 A circular portion of index k, for the propagation mode of index m and as a function of the frequency f. m varies over the set of so-called "parasitic" modes A / par, that is, all modes except the mode of interest m0 for which we want the coil to be sensitive. A is the number of circular portions. ("j, ..., ¾) are the coefficients of the linear combination, which are positive or negative real numbers. For example, the cost function is equal to Lj . I2 if the sensitivities are calculated for a single frequency or 9 if the sensitivities are 0(^1' • • •, (¾) — J km( f | J min calculated for a frequency range [ / , ; / max] or Alternatively, the cost function is taken as equal to Q(a„ ..^aN) =........--------- JK=1akEtaJf)j df Hk) is the mode of interest Another possible cost function is, for example: Q ( ®1' • ■ • ' — | l^k^km (^)) Or: f J max Q ( ^1' ' JI km ( f ) | J min

[0092] Or: N, L:xu J |Ek=1CkBUqdt Q (c 15 ..., c N ) — ™ IE,ckEk4fJdf - ntiji

[0093] Step 502 then consists of minimizing the cost function Q to determine the coefficients (cq, which allow us to obtain an average sensitivity to all the parasitic modes that are closest to 0.

[0094] To ensure a good numerical resolution of this optimization, advantageously, the value of the first coefficient is fixed, for example, equal to "j = 1.

[0095] Without departing from the scope of the invention, the above cost function examples can be replaced by any other cost function that allows the sensitivity of the coil to parasitic propagation modes to be characterized.

[0096] In step 503, the transducer 400 is designed by implementing in the processing unit 403 the function _ yA , where sk is the signal measured on the portion circular of index k and sf is the final signal calculated by the processing unit, this signal comprising the wave packets of the propagation mode of interest, the wave packets of the other modes being filtered.

[0097] A particular embodiment of the invention applied to the inspection of a 3 mm thick aluminum plate structure measuring 120 by 120 cm is now described. The frequency range of interest is below 100 kHz, a range for which only three guided modes exist in the plate: A0, S0, and SH0. The mode of interest that is to be isolated in this example is A0, with an X wavelength of 20 mm. The EM AT sensors developed consist of a coil comprising two circles with radii of 30 mm (3ÀZ2) and 50 mm (5ÀZ2), respectively, printed in six layers on a Kapton sheet, and a cylindrical neodymium magnet with a diameter of 60 mm and a height of 20 mm. The Kapton sheet connectors are designed to allow each of the six layers corresponding to each turn of a circle to be connected. The sensor is used in the receiving mode. In transmission, a piezoelectric transducer is used for wave generation.This is excited by a 3-cycle signal at 50 kHz with a Hann windowing.

[0098] To perform step 501, the six layers of each circle are connected in series to increase the sensitivity of each circle. Acquisitions are made for the inner and outer circles, shown in Figures 6a and 6b. At the beginning of the measurement, between approximately 0.05 ms and 0.3 ms, three wave packets are clearly visible. The first, 601,611, corresponds to the first packet of mode S0 (direct path from the piezoelectric transducer to the EMAT), the second, 602,612, to the first packet of mode A0 (also direct path), and the third, 603,613, to the first reflection of mode S0 (the group velocity of mode S0 is approximately three times greater than that (AO mode in this example). This temporal separation, achieved through a judicious choice of the two sensor positions, allows for the separation of the two modes, AO and SO. As a reminder, the EMAT sensor is completely insensitive to SHO mode due to its circular symmetry.

[0099] For the implementation of step 502, the energy of the signal of the parasitic modes, which are here the SO modes, divided by the energy of the signal of the mode of interest AO is used as the cost function. The calculation is performed on the corresponding time windows.

[0100] The cost function is therefore A

[0101] Numerical optimization gives the coefficients «1= 1 and «2 = -0.39

[0102] Fig. 7 represents the signal obtained by linear combination of the signals from Figures 6a and 6b with the coefficients determined above.

[0103] It is observed that the wave packets of the selected mode A0 (zone 702) are well preserved while those of the parasitic modes (mode S0 zone 701 and zone 703) are strongly attenuated.

[0104] The signal in [Fig.7] corresponds to the final signal calculated by the processing unit.

[0105] Thus, it is possible to modify the propagation mode of interest by recalculating the weighting coefficients implemented by the processing unit. It is also possible to update them if the sensor's environmental conditions change, which may lead to modifications in the sensitivity values ​​calculated in step 501.

[0106] In one embodiment of the invention, the magnet associated with the coil to form the transducer is chosen to promote the modal selectivity of certain guided modes. For example, choosing a cylindrical magnet with a diameter larger than that of the coil results in a magnetic field at the surface of the direction plate that is primarily normal, and therefore a sensitivity to displacements in the plane, and thus more to the S0 mode than to the A0 mode.

[0107] To favor sensitivity to A modes rather than S modes, a radial magnetic field (i.e. in the plane) should be favored, for example by using a small cylindrical magnet and a larger toroidal magnet as is done in article [3].

[0108] Another variant involves using a magnetic concentrator to maximize the magnetic field strength in the vicinity of the coil wires. This increases the overall sensitivity of the sensor.

[0109] The processing unit 403 of the transducer can be implemented using hardware and / or software components. It can be implemented in software and / or hardware form, in particular by using one or more processors and one or more memory units. The processor can be a generic processor, a processor specific, an application-specific integrated circuit (also known as an ASIC for "Application-Specific Integrated Circuit") or an in-situ programmable gate array (also known as an FPGA for "Field-Programmable Gate Array"). References

[0110] [1] “Directivity analysis of meander-line coil EMATs with a purely analytical method”, Yuedong Xie et al, Ultrasonics, vol 73, Jan 2017, p 262-270

[0111] [2] “The excitation and detection of Lamb Waves with planar coil electromagnetic acoustic transducers », Paul D. Wilcox et al, IEEE Transactions on ultrasonics, ferroelectrics, and frequency control, vol 52, n°12, dec 2005

[0112] [3] « Development of omnidirectional A0 mode EMAT employing a concentric permanent magnet pairs with opposite polarity for plate inspection »,Zenghua Liu et al, NDT& E International, vol 94, Mardi 2018, p 13-21

[0113] [4] « Development of an omni-directional magnetic-concentrator-type electromagnetic acoustic transducer », Zenghua Liu et al, NDT& E International, vol 109, Jan 2020,

Claims

Demands

1. A selective acoustic electromagnetic transducer (400) for a Lamb wave propagation mode of interest propagating in a structure, comprising a magnet (401) with circular geometry and a coil (402) comprising several independent concentric circular portions (Ci,C2,C3) and a processing unit (403) connected to the ends of each circular portion (Ci,C2,C3), the processing unit (403) being configured to generate a composite signal equal to a linear combination of the signals measured on each circular portion, the coefficients of the linear combination being chosen so as to minimize the sensitivity of the coil (402) to a set of Lamb modes excluding the mode of interest.

2. Acoustic electromagnetic transducer according to claim 1 in which each circular portion (Ci,C2,C3) is made up of a winding of several turns.

3. Acoustic electromagnetic transducer according to claim 2 in which the turns of a circular portion (Ci,C2,C3) are superimposed on each other.

4. Acoustic electromagnetic transducer according to any one of the preceding claims wherein the circular portions (Ci,C2,C3) have diameters substantially equal to (k+l)X / 2 where X is the wavelength of the mode of interest and k is a positive or zero integer.

5. Acoustic electromagnetic transducer according to any one of the preceding claims wherein the number of circular portions (Ci,C2,C3) is at least equal to a specified number of Lamb wave propagation modes.

6. Acoustic electromagnetic transducer according to any one of the preceding claims comprising a magnetostrictive patch.

7. A method for designing the acoustic electromagnetic transducer according to any one of the preceding claims, the method comprising the steps of: - Determining (501) the sensitivity of each circular portion of the coil to several respective Lamb wave propagation modes, - Defining a cost function dependent on a linear combination of the sensitivities of each circular portion to a set of Lamb wave propagation modes excluding the Lamb mode of interest, - Find (502) the values ​​of the coefficients of the linear combination that minimize the cost function, - Implement (503) in the processing unit said linear combination of the signals measured from the coefficients obtained

8. Method of designing an acoustic electromagnetic transducer according to claim 7 wherein the cost function depends on the sum, over the set of Lamb wave propagation modes excluding the Lamb mode of interest, of the modulus or the squared modulus of the sum of the sensitivities of each circular portion, weighted by the coefficients of the linear combination.

9. Method of designing an acoustic electromagnetic transducer according to claim 8 wherein the sensitivities are determined for several frequency values ​​in a given frequency range and the cost function is equal to the double sum over all Lamb wave propagation modes excluding the Lamb mode of interest and over all frequencies, of the magnitude or the squared magnitude of the sum of the sensitivities of each circular portion, weighted by the coefficients of the linear combination.

10. Method of designing an acoustic electromagnetic transducer according to claim 9 wherein the cost function is divided by the modulus or the squared modulus of the sum of the sensitivities to the Lamb mode of interest, of each circular portion, weighted by the coefficients of the linear combination.

11. Method of designing an acoustic electromagnetic transducer according to claim 10 wherein the sensitivities are determined for several frequency values ​​in a given frequency range and said modulus or squared modulus of the sum of the sensitivities is integrated over said frequency range.

12. A method for designing an acoustic electromagnetic transducer according to any one of claims 7 to 11, wherein the sensitivity of each of the circular portions is determined analytically via a formula depending at least on the radius of the circular portion, the components of the mode propagation, its wavelength and the components of the magnetic field to which the circular portion is subjected or by simulation.

13. Method of designing an acoustic electromagnetic transducer according to any one of claims 7 to 11 wherein the sensitivity of each of the circular portions is determined experimentally by means of the steps of: - Performing a guided wave measurement for each circular portion of the coil, - For each of said Lamb wave propagation modes, isolating the propagation mode on the measurement and determining the sensitivity of the circular portion to said propagation mode from the isolated portion of the measurement.

14. Method of designing an acoustic electromagnetic transducer according to claim 13 wherein the guided wave measurement is carried out by: - ​​positioning said transducer and a second transducer on a sample, the second transducer being positioned at a predetermined distance from said transducer as a function of the group velocity of at least one of said propagation modes, - Measuring the guided wave emitted by the second transducer and propagated in the sample.

15. Method of designing an acoustic electromagnetic transducer according to claim 14 wherein the guided wave measurement is carried out by: - ​​positioning said transducer and a second transducer on a sample, the second transducer being configured to selectively emit one of said propagation modes, - measuring the guided wave emitted by the second transducer and propagated in the sample.