Method for non-destructively testing a part made of polycrystalline material

Abstract

The invention relates to a method (100) for non-destructively testing a part made of polycrystalline material, the method being executed by a test device and comprising: measuring (1) an ultrasonic reflection matrix of the part associated with a multi-element probe containing N piezoelectric transducers; projecting (2) the ultrasonic reflection matrix into a focused basis; constructing (3) a map of the speed of the longitudinal waves and / or constructing (4) a map of the speed of the transverse waves; constructing (5) a temperature map in a volume of the material using the map of the speed of the longitudinal waves and / or the map of the transverse waves; detecting (6) a defect or a particular area in the part by analysing the determined temperature map.

Description

[0001] NON-DESTRUCTIVE INSPECTION METHOD FOR A POLYCRYSTALLINE PART

[0002] TECHNICAL FIELD

[0003] The invention relates to the field of control applied to parts made of polycrystalline material and more particularly to the field of non-destructive testing of parts made of polycrystalline material.

[0004] STATE OF PRIOR ART

[0005] A number of critical components used in aircraft engines are made of polycrystalline metal alloys, such as titanium alloys or nickel-based superalloys. These components can be forged from cylinders called billets. They may also be pre-machined or machined.

[0006] To ensure the integrity of these components before installation in the reactor and to verify that they are free from defects such as cracks, porosity, or inclusions, they can be non-destructively inspected using ultrasound to detect any indications of defects. These inspections are performed on billets and on forged components using one or more piezoelectric transducers. Single-element or multi-element (ring or linear) piezoelectric probes are commonly used.

[0007] Ultrasonic waves are sensitive to many physical properties of the material through which they propagate, such as its elastic constants or temperature. More specifically, this patent presents a method for determining the local temperature within a material by measuring the speed of ultrasonic waves propagating through it.

[0008] The temperature measurements currently available in the field of industrial metal parts inspection are:

[0009] of the surface temperature of the room;

[0010] - the temperature inside the volume of a part but having drilled through the part to insert a thermocouple (destructive); - the temperature at the surface of a part to which is associated a numerical model to determine the temperature in the volume by solving the inverse problem of thermal diffusion; this requires knowing the thermal properties of the material and generally relies on many assumptions (fully known or even homogeneous material, absence of defects, absence of drift in the forging process, etc.).

[0011] Methods using ultrasonic waves have been described in the literature: they employ a velocity chart for ultrasonic waves as a function of temperature; the velocity is measured by a time-of-flight method or by ultrasonic tomography. Advanced methods have been developed specifically for the medical field. US patent 2013 / 0116560 A1 describes the use of a single-element ultrasonic probe and an algorithm based on the cross-correlation of time-domain signals (A-scans) to determine the temperature of an object placed in the inspected medium, in a medical context. The signals are acquired for two states: a reference state before heating, and a heated state. The algorithm detects a shift in the wavefronts reflected by the object and correlates them with a change in the object's temperature.This method allows us to probe a change in temperature in the environment but not to create a map of local temperature values.

[0012] US patent 2014 / 0018676 Al also uses correlations between acquired ultrasonic signals to calculate a temperature map of the medium in the medical field, based on the shift of wavefronts reflected by objects. The developed method makes it possible to obtain a temperature map from signals generated by a phased array ultrasonic probe. To do this, the energy of the echoes reflected by a tissue of interest is compared to that of reference echoes. This method is not based on measuring the velocities of ultrasonic waves.

[0013] US patent 2014 / 0121517 Al describes curves of the variation of the longitudinal ultrasound wave velocity with temperature for several components of a biological medium and combines them to determine the temperature of a biological tissue. This method is not based on a local analysis of the ultrasound signals.

[0014] Obtaining the local temperature via a method compatible with industrial constraints would be beneficial for the non-destructive characterization of parts made of polycrystalline material, for example by identifying areas with abnormal temperatures generated by the manufacturing processes or to use the measured temperature field as input data for simulation software used for forging.

[0015] In this context, it is therefore necessary to provide a non-destructive testing method for a part made of polycrystalline material, allowing the temperature of the material inside the part to be measured non-intrusively and with a single probe.

[0016] DESCRIPTION OF THE INVENTION

[0017] To this end, according to a first aspect, a non-destructive testing method for a part made of polycrystalline material is proposed, the method being carried out by a non-destructive testing device comprising electronic circuitry adapted to implement the method, the method being characterized in that it comprises at least the following steps:

[0018] A. measure an ultrasonic reflection matrix of the part associated with a multi-element probe containing N piezoelectric transducers;

[0019] B. project the ultrasonic reflection matrix into a focused basis;

[0020] C. construct a map of the speed of longitudinal waves and / or D. construct a map of the speed of transverse waves;

[0021] E. construct a temperature map in a volume of the material using (respectively) the longitudinal wave velocity map and / or the transverse wave velocity map;

[0022] F. detect a defect or a particular area in the part by analyzing the temperature map determined in step E.

[0023] Thus, the process according to the invention very cleverly uses longitudinal waves and / or transverse waves to locally determine the temperature of the material, in its volume.

[0024] In a particular embodiment, when the process includes step D, then step D comprises separating the reflection matrix into four submatrices: a first submatrix composed of emitters from a part of the probe located to the left of the focal point, a second submatrix of receivers from a part of the probe located to the left of the focal point, a third submatrix composed of emitters from a part of the probe located to the left of the focal point, and a fourth submatrix of receivers from a part of the probe located to the left of the focal point.

[0025] In a particular embodiment, when the process includes step D, then at step D, the reflection matrix is ​​expressed as follows: |

[0026]

[0027] R^ % | = l^dd I + + l^d^l + with Rdd the first submatrix, R gg the second submatrix, Rg to the third sub-matrix and Ra g the fourth sub-matrix...

[0028] In a particular embodiment, when the process includes step D, then at step D, the reflection matrix is ​​expressed using three or fewer of the four submatrices.

[0029] In a particular embodiment, when the process includes step D, then step D includes weighting the reflection matrix by a function dependent on the angle between a transducer and the focal point.

[0030] In a particular embodiment, the reflection matrix is ​​filtered using a function expressed as i(0) = |sin(20) x 0“| with h the filter function and 0 the angle between the normal to the probe and the axis along the line joining the center of the probe and the focal point; or as i(0) = D T (0) x (1 — D L(0)) with h the filter function, 0 the angle between the normal to the probe and the axis along the line joining the center of the probe and the focal point, D T (0) the directivity of a network element for transverse waves and D L (0) the directivity of a network element for longitudinal waves.

[0031] In one particular embodiment, step E comprises performing, for at least one predetermined temperature, a calibration phase including placing the part in an environment exhibiting said predetermined temperature and performing steps A to D of the process to construct a longitudinal wave velocity map and a transverse wave velocity map for the predetermined temperature. Thus, in one particular embodiment, the calibration phase can be repeated for a plurality of predetermined temperatures.

[0032] In a particular embodiment, step E includes a phase of mapping a room temperature from the mapping of longitudinal wave velocities and / or the mapping of transverse wave velocities, using the curves constructed during the calibration phase. In a particular embodiment, the ultrasonic probe used is a matrix probe.

[0033] According to a second aspect, a non-destructive testing device is proposed for a part made of polycrystalline material, characterized in that it comprises electronic circuitry for implementing a method for detecting a defect in a part made of polycrystalline material, which includes at least the following steps:

[0034] A. measure an ultrasonic reflection matrix of the part associated with a multi-element probe containing N piezoelectric transducers;

[0035] B. project the ultrasonic reflection matrix into a focused basis;

[0036] C. construct a map of the speed of longitudinal waves and / or D. construct a map of the speed of transverse waves;

[0037] E. construct a temperature map in a volume of the material using (respectively) the longitudinal wave velocity map and / or the transverse wave velocity map;

[0038] F. detect a defect or a particular area in the part by analyzing the temperature map determined in step E.

[0039] BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The features of the invention mentioned above, as well as others, will become clearer upon reading the following description of at least one exemplary embodiment, said description being made in relation to the accompanying drawings, among which:

[0041] [Fig. 1] schematically illustrates the process of a non-destructive testing procedure for parts made of polycrystalline material;

[0042] [Fig. 2] illustrates a focal spot measurement (RPSF) as a function of velocity integrated in a polycrystalline material;

[0043] [Fig. 3] illustrates a focused basis reflection matrix measurement for transverse Fonde without any additional steps performed compared to the longitudinal Fonde processing;

[0044] [Fig. 4] schematically illustrates a defined angle for a filter;

[0045] [Fig. 5] schematically illustrates a superimposed applied filter with simulated amplitudes; [Fig. 6] schematically illustrates the directivity of an element for longitudinal and transverse basis;

[0046] [Fig. 7] illustrates experimental maps of longitudinal Fonde velocity and transverse Fonde velocity in a polycrystalline material;

[0047] [Fig. 8] schematically illustrates the implementation of a calibration phase;

[0048] [Fig. 9] illustrates longitudinal Fonde velocity maps measured in a metallic sample;

[0049] [Fig. 10] illustrates temperature maps as a function of ultrasonic wave velocities during the calibration phase;

[0050] [Fig. 11] schematically illustrates a longitudinal and transverse Fonde velocity curve as a function of temperature;

[0051] [Fig. 12] schematically illustrates a computer system adapted to implement the process.

[0052] DETAILED DESCRIPTION OF IMPLEMENTATION METHODS

[0053] Control method

[0054] With reference to Fig. 1, a non-destructive testing method 100 for parts made of polycrystalline material is proposed in its first aspect. As described below, the method 100 is carried out by a non-destructive testing device comprising electronic circuitry 200 adapted to implement the method 100. The method 100 comprises at least the following steps:

[0055] (Step 1) measure an ultrasonic reflection matrix of the part associated with a multi-element probe containing N piezoelectric transducers;

[0056] - (Step 2) project the ultrasonic reflection matrix into a focused basis;

[0057] - (Step 3) construct a map of the velocity of longitudinal waves and / or (Step 4) construct a map of the velocity of transverse waves;

[0058] - (Step 5) construct a temperature map in a volume of the material using (respectively) the longitudinal wave velocity map and / or the transverse wave velocity map;

[0059] - (Step 6) Detect a specific area in the room by analyzing the temperature map determined in step 5. Step 1 - Measure an ultrasonic reflection matrix

[0060] As previously stated, process 100 includes an initial step 1 of signal generation with a multi-element probe. According to a particular arrangement, process 100 uses a multi-element probe containing N piezoelectric transducers to generate a wave in the polycrystalline material. It is specified that N is a positive integer strictly greater than zero.

[0061] The reflection matrix can be measured by acquiring NxN ultrasonic signals, where N is a positive integer. These ultrasonic signals are the impulse responses between each transducer and will be denoted R(u in ,u 0Ut , t), with t being the echo time. To do this, the position of each transducer is denoted u in when a transducer is used as a transmitter and u 0Ut When a transducer is used as a receiver, this notation allows us to define a reflection matrix for the material under study (Full Matrix Capture, abbreviated FMC). The reflection matrix is ​​thus presented in the form of an R matrix. uu (t) = [R(Ui n ,u 0Ut , t)].

[0062] Alternatively, the reflection matrix can be measured in a plane wave basis.

[0063] Step 2 - Project the reflection matrix onto a focused basis

[0064] According to a particular arrangement, a change of basis allows the reflection matrix to be projected into a basis focused on a set of points {xi n ,z} and {x ou t,z} to search for local elastic constants of the inspected material. It is specified that the matrix R uu presented in the form of an FMC matrix, to study the medium locally, it may be necessary to focus the waves into the medium and thus transform the matrix R uu (t) en a matrix R xx (z) = [Æ(x in ,x out , z)]. The focused basis reflection matrix can be seen as an extension of the classical (confocal) ultrasonic image where a focal point in emission and reception can be separated (xi n x ou t). This can, for example, be achieved by using different focusing delays at transmission and reception.

[0065] Using a frequency-domain approach to image construction, it is possible to express the focused basis reflection matrix from the reflection matrix. In particular, by defining:

[0066]

[0067] E) TFÇR(Ui n ,U out ,t')')

[0068] where TF represents the discrete Fourier transform along the time dimension of the reflection matrix and f is the frequency. It is then possible to define the reflection matrix in the focused basis as:

[0069] ^( x in< x out< ' ( x in< Uin> Z, / ^^(Uin, U ou t> / Df' ( x out< Uout< Z' D

[0070]

[0071] f u in u out

[0072] In this equation, x in represents the x-coordinate of the emitting focal points, xout represents the x coordinate of the focal points in reception, z a focal depth and G is a Green's matrix which contains Green's functions describing the propagation between the transducers and the focal points for a given ultrasonic wave velocity c0. The symbol * represents the conjugate matrix.

[0073] According to the embodiment presented here, the reflection matrix is ​​defined with a two-dimensional Green's function:

[0074] 2TT / G(x it u jt z t f) = —Hl , with k = - .

[0075]

[0076] c0

[0077] Step 3 - Construct a map of the speed of longitudinal waves

[0078] The expression of R xx (z = [Æ(x in , x out, z)] implicitly depends on the speed of the ultrasonic wave defined to achieve focusing c0. In simple scattering, the focused reflection matrix appears as a matrix whose energy is essentially concentrated on the main diagonal (it is logical to "probe" the maximum energy in reception at the point where focusing was achieved in transmission). The energy spread around the diagonal depends on the difference between the model speed of sound c0 and the actual velocity distribution in the medium.

[0079] Thus, to estimate the speed of the longitudinal wave, the energy spreading outside the diagonal of R xx (z) can perhaps be quantified by expressing the reflection matrix in a de-scanned basis, i.e., as a function of the distance between the points x in and x out : Ax = x out — x inFurthermore, the coefficients of the reflection matrix must be expressed as a function of the echo time t, rather than as a function of the depth z = 0^1 / 2 of the plane, which depends on the velocity model:

[0080] Æ({x in > , Mx, c0}) = Æ(x in , x out , z, c0)

[0081] This change of variable allows us to track the evolution of the point spread function (PSF) for each speckle grain (i.e., speckle or iridescence) defined by its spatiotemporal coordinate {x in , t}. Generally measured by imaging a point object and observing the size of its image, here it is measured by reflection and corresponds to the dependence of the backscattered energy on Ax:

[0082] R

[0083]

[0084] PSF(àx, c0) =< | / ?({x in , t}, {Ax, c0}) | 2 > {Xin , t}

[0085] where the symbol (...) denotes an average over the subscript coordinate. This function then provides an estimate of the width of the focal spot; in the absence of aberrations, this is given by diffraction theory: θx = Δz / D, with Δz the wavelength, z the distance between the emitter and the object, and D the size of the probe aperture. Experimentally, the presence of aberrations (inhomogeneous medium, imprecise knowledge of the ultrasonic wave velocity, multiple scattering, etc.) widens the RPSF. Its effective width Δ corresponds to the width of the diagonal of the focused basis reflection matrix measured around a position x. in , for a given velocity c0 and echo time t.

[0086] Thus, the measurement of A as a function of the assumption of the speed of the ultrasonic wave c0 in the calculation of the focused basis reflection matrix provides information on the actual speed of Fonde in the medium: a minimization of A is directly correlated to a decrease in aberrations, therefore a minimization of the difference between the assumption of the speed of Fonde and its actual speed in the medium.

[0087] In the study of the microstructure of metallic alloys, many configurations lack a strong reflector for directly measuring the PSF (Peak-Scatter Factor). Therefore, it is estimated using speckle, which is the superposition of all the signals backscattered by the microstructure. Since the intensity of speckle is inherently spatially fluctuating and locally zero, it is necessary to spatially average the measured PSFs to smooth out the speckle fluctuations and obtain a correct estimate of A; this spatial average can be consistent or inconsistent. A typical size for the averaging area to obtain noise-free measurements is approximately six wavelengths in both the x and z dimensions.

[0088] In the case of, for example, the nickel-based superalloy Inconel 600, focusing is more accurate when the velocity used to calculate the reflection matrix is ​​the actual longitudinal wave velocity (5850 m / s) than when the velocity used in the algorithm is overestimated. In this way, the velocity that minimizes the width and maximizes the intensity of the focal spot will be considered the average velocity between the multi-element probe and the focal point, called the integrated velocity. A RPSF measured in a sample of Inconel 600 is shown in Fig. 2.

[0089] From integrated velocity measurements taken at a point grid, it is possible to construct an integrated velocity map between the probe and all focal points. This map is obtained by estimating the optimal longitudinal velocity of Fonde around a given point. By then shifting the averaging zone, it is possible to construct an integrated velocity map within the inspected area of ​​the material.

[0090] Several methods are possible for calculating a local longitudinal Fonde velocity map from an integrated longitudinal Fonde velocity map. For example, one can consider the Fonde slowness s, defined as the inverse of the Fonde velocity (s = 1 / c). The integrated slowness s int is given by the average local slowness If ocalong the Fonde trajectory. Neglecting oblique paths and refraction effects, the integrated slowness between the probe and each point is related to the local slowness by the following relationship:

[0091] 1 f Zt

[0092] z t) I dz S[ocÇx> %)

[0093]

[0094] z Jo

[0095] with z t = t / (2s int ), the depth of the isochronous volume at the considered echo time t. The previous equation can be rewritten in matrix form by discretizing the values ​​of the integrated and local slownesses at a set of depths z t :0 0

[0096] 1 / If oc (x, zj \ / s int (x, zj \ 0

[0097] 2 Sloc(.X> ^2) Sint(.X> ^2) 1 1 Sloc(.X> ^3) Sint(.X> ^2) with A =

[0098] 3 3 \si 0C (x, Z N ) / k^int Z;v) / ii

[0099]

[0100] NN

[0101] By then inverting this equation, the local slowness is given by:

[0102] / SiocÇx.Z^XS' SS s / 1 0 0 ... 0\ Sloc(.X> ^2) 5? 5? 5? 5? ••• -1 2 0 ... 0 Sloc = A- 1 x NNNN (.X> ^3) 3 NNR)) with A 1 = 0 -2 3 ... 0

[0103]

[0104] \If 0C (,X,Z N ) / \ 0 0 0 ... IV /

[0105] This calculation is very sensitive to noise, which inevitably pollutes the integrated speed map. In practice, a Gaussian filter is therefore applied to the integrated speed map to eliminate this noise before performing the calculation shown in the equation above.

[0106] By applying this calculation to all columns (lateral x positions) of the longitudinal wave velocity map, it is thus possible to reconstruct a complete map of the local longitudinal wave velocity.

[0107] It should be noted that the invention is not limited to this method of measuring the speed of longitudinal waves and that it can be replaced or coupled with other approaches such as, for example, the one consisting of finding the speed model optimizing a coherence factor in place of RPSF [Reference: M. Jakovljevic, S. Hsieh, R. Ali, G. Chau Loo Kung, D. Hyun, and JJ Dahl, Local speed of sound estimation in tissue using pulseecho ultrasound: Model-based approach, Journal of the Acoustical Society of America 144, 254 (2018)], the CUTE method which allows to perform the velocity of sound tomography by analyzing the phase shift of ultrasound images constructed on different angular spectra [P. Stahli, M. Kuriakose, M. Frenz, and M. Jaeger, Improved forward model for quantitative pulse-echo speed-of-sound imaging, Ultrasonics 108, 106168 (2020)] or this latter method for only axial variations of the speed of sound [B. Heriard-Dubreuil, A.Besson, F. Wintzenrieth, C. Cohen-Bacrie, and J.-P. Thiran, Refraction-Based Speed ​​of Sound Estimation in Layered Media: an Angular Approach, IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control 70, 486 (2023)]. These different approaches have in common that they rely on the measurement of the reflection matrix and manipulate ultrasonic data from different study bases to perform a tomography of the speed of sound in reflection.

[0108] Step 4 - Construct a map of the velocity of transverse waves

[0109] Determining the velocity of transverse ultrasonic waves presents a challenge due to the presence of side lobes in the focal spot when attempting to coherently sum the signals from different transducers. These side lobes are related to the angular spectrum of the transverse waves generated by the ultrasonic probe. This spectrum exhibits a dipolar shape with field cancellation along the z-axis, while the maximum amplitude of the longitudinal waves is found directly in front of the transducer. For transverse waves, the maximum is typically located around 40° from the transducer axis.

[0110] This directivity profile manifests as a low-intensity diagonal for the reflection matrix in the focused basis obtained from the transverse wave velocity, as shown in Fig. 3. A second problem is the presence of residual signals associated with the longitudinal wave. It is necessary to reduce their contribution to preserve the majority of the transverse wave signal. These two specific properties of the matrix in the focused basis associated with the transverse wave necessitate an adaptation of the algorithm described previously for longitudinal waves. Therefore, additional processing must be performed on the matrix in the focused basis Æ(u in , u out , ) to obtain a usable transverse wave velocity map.

[0111] To this end, a filter is applied to give less weight to longitudinal waves and more weight to transverse waves. To do this, the Green's matrix is ​​weighted by a value that depends on the angle 9 between the normal to the probe and the axis of the line joining the center of the probe and the focal point. The angle 9 is shown schematically in Fig. 4.

[0112] This filter function h(0) can be the simulated amplitude of the transverse Fonde propagating in the material of interest or its approximation by an analytical function. The chosen function can be of the form:

[0113] / i(0) = |sin(20) x 0“|, where a is the parameter optimized to approximate the simulation. For example, for a simulation in the case of a TA6V titanium alloy and a transducer with a width of 0.4 mm and a center frequency of 3.5 MHz, a = 0.21. This function h can also be normalized. The curve resulting from the simulation and the deduced analytical function are shown in Fig. 5. Alternatively, the choice of the function h can be made with respect to the theoretical directivity of the elements composing the transducer array.

[0114] The calculation of directivity for longitudinal and transverse waves is performed for a transducer of infinite length (a valid assumption because its width is much greater than the wavelength), of a given width *a*, and for a known material. It is based on the calculation of three terms. The first is the radiation from a source line emitting a signal *s(t)* for a given frequency *f*:

[0115] LS L(0J) = TF(s(t)) . -L-^cosCS)

[0116] J \ (JL)

[0117] 1 1

[0118] LS T Ç6,n = TF(s(^Y— — sinW

[0119]

[0120] < J \ (JL)

[0121] with :

[0122] m = 2nf the angular frequency with f the frequency of the emitted signal;

[0123] T (s(t)) the Fourier transform of the signal emitted at frequency f;

[0124] v L / T the speed of the longitudinal and transverse waves respectively;

[0125] Z L / T = pv L / T the impedance of the medium respectively for longitudinal and transverse waves.

[0126] The second term allows us to take into account the scalar opening of width a relative to the source line:

[0127] (scal L (_0,f) = 2a sinc(k L a sin(0y)

[0128]

[0129] scal T (0,f) = 2a sinc(k T a sin(6y)

[0130] with k L / T = the wave number of the longitudinal and transverse waves respectively.

[0131]

[0132] The third term is the Miller-Pursey factor, which accounts for the fact that generation occurs on the surface of the sample: k a 2 - 2sin(0) 2 MP L (0) = 2k 2

[0133] (fc 2 — 2sin (0) 2 ) 2 + 2sin(0) x sin(20) x ^ / fc 2 — sin (0) 2 <

[0134] 71 - Zc 2 sin(0) 2 MP T (0) = 4cos(0) fc a (l — 2sin(0) 2 ) 2 + 2sin(0) x sin(20) x — fc 2 sin(0) 2 7 with k a = —

[0135]

[0136] Vj 7Finally, the directivities of a network element for longitudinal and transverse waves are given, for a frequency f, by:

[0137] r D L (&) = x scaZ L (0, / ) x W L (0)|

[0138] f

[0139] <

[0140] D T (0) = ^\LS T (e,f) x scaZ T (0, / ) x MP r (0)|

[0141]

[0142] f

[0143] These guidelines are then standardized.

[0144] An example in the case of the TA6V titanium alloy for an element of width 0.42 mm and for a Gaussian pulse s(t) of center frequency 3.5 MHz is shown in Fig. 6.

[0145] Finally, the second filter proposed to favor the signal from transverse waves and limit that from longitudinal waves is Zi(0) = D T 0') X (1 —

[0146] Taking into account the chosen h function, the modified Green's function applied to focus the reflection matrix is ​​as follows:

[0147] G'(xpii,-, z, / ) = — HQ I — k (Xi — u ) 2 + z 2 ) h(0\ with k = - and

[0148] 4 \ \ / c0

[0149]

[0150] By applying this filter during the matrix calculation, the longitudinal wave signal is reduced relative to the transverse wave signal, thus improving the matrix's signal-to-noise ratio and, in practice, resulting in a smaller off-diagonal signal relative to the on-diagonal signal. The second step consists of eliminating interference due to the coherent summation of the signal from the side lobes. For this, the reflection matrix can be viewed as composed of four summable terms: a reflection matrix composed of emitters and receivers from a portion of the probe located to the right of the focal point (denoted Raa), and a reflection matrix composed of emitters and receivers from a portion of the probe located to the left of the focal point (R). gg) and a reflection matrix composed of emitters and receivers originating from a different part (right or left) of the probe relative to the focal point (R g a and Ra g ).

[0151] To perform the calculation of these four contributions, the matrix G' is separated into two "left" and "right" contributions, G g ' and G d ', whose coefficients are calculated as follows:

[0152] G'Çxi,Uj,z,f^ if Uj <

[0153] G^(xi,Uj,z,f) =

[0154] 0 otherwise

[0155] G'(xi,Uj,z,f^ if Uj > Xi

[0156] Gâ(^'UpZ, / ) =

[0157]

[0158] 0 otherwise

[0159] The partial reflection matrices are calculated from G g ' and G A ':

[0160] Gd' * (Xin' U in , Z, / )7?(u in , U out , / ) Rdd (Xi n X out , G d (xO ut< U O ut< X, Z 7 ) f u in u out

[0161] Rgg (Xin< x O ut' f u in u out

[0162] Rgdfain' x out' Z) = 111 Gg (Xin> Uj n , Z, / ) / ?(u in , U ou t> f^G d (x O ut' Uout' / ) f u in u out

[0163] ^ (Xin' Xout' Z) = / JZJ À G d ( X in' Uin, Z, / )Æ(u in< U O ut< DGg (X out , Uout' X, / )

[0164]

[0165] f u i n u An interference-free reflection matrix between transverse waves propagating to the right (towards x>0) and to the left (towards x<0) can then be obtained by the incoherent sum of the four sub-matrices defined above:

[0166] L

[0167]

[0168] ^xxl - \Rdd ​​I + + \Rdg I + \Rgd I

[0169] According to a particular arrangement, the reflection matrix (which could be called the total reflection matrix) can be expressed using three or fewer of the four submatrices. In particular, the information of R x ' x can mostly be contained in the matrices R dg and R gd when the interaction of the wave with the medium is primarily specular (reflectors of a size comparable to or greater than the wavelength) or is mainly contained within the R matrices dd and R gg when the scatterers in the medium are smaller than the wavelength. In an intermediate case, all four matrices contain information of interest.

[0170] Next, it is possible to construct a transverse velocity map in the same way as for longitudinal velocities, but using the R matrix. x ' xinstead of R xx Fig. 7 presents experimental maps of longitudinal wave velocity and transverse wave measured in a sample of Inconel 600.

[0171] Step 5 - Map the local temperature within the volume of the material

[0172] Once the maps of the velocities of the transverse and longitudinal waves have been made, it becomes possible to make a temperature map in a volume of the material using the map of the velocities of the longitudinal waves and / or the map of the velocities of the transverse waves.

[0173] Step 5 includes a calibration phase involving placing the part in an environment with a predetermined temperature and performing steps 1 and 2, as well as steps 3 and / or 4 of the process, to construct a longitudinal wave velocity map and / or a transverse wave velocity map for the predetermined temperature. In other words, calibration allows a temperature to be associated with a longitudinal wave velocity and / or a transverse wave velocity in the material. Depending on the specific arrangement, calibration involves placing the part in a temperature-controlled environment (e.g., an oven) for a sufficient time to ensure a homogeneous and controlled temperature.

[0174] According to a specific arrangement, the calibration phase is repeated for a plurality of predetermined temperatures to obtain maps of wave velocities as a function of temperature, which allows the construction of curves in which the transverse wave velocities and / or the longitudinal wave velocities are determined as a function of temperature. Such curves are shown as an example in Fig. H.

[0175] It is specified that the number of repetitions of the calibration phase can vary depending on the desired accuracy. In other words, the more the calibration phase is repeated for distinct predetermined temperatures, the more accurate the resulting curves will be. Conversely, a small number of repetitions of the calibration phase allows for faster curve generation, but these curves will be comparatively less accurate.

[0176] As shown schematically in Fig. 8, part 10 can be positioned on a heating element 20 at a temperature T p and an ultrasonic probe 30 can be positioned on part 10. Part 10 is heated from below from ambient temperature (~20°C), inducing a temperature gradient in the part when T p > 20°. This temperature gradient induces a velocity gradient for longitudinal and transverse ultrasonic waves, as shown in Fig. 9 (longitudinal Fonde velocity measurement). Using the previously constructed calibration curve, it is possible to measure the temperature locally within the material, as shown in Fig. 10.

[0177] It is specified that it is possible to construct only a curve of transverse velocities as a function of temperature or only a curve of longitudinal velocities as a function of temperature. Alternatively, both curves can be constructed and used to reduce the uncertainty in the estimation of the local temperature of the material.

[0178] Step 5 then includes a phase of mapping a room temperature from the mapping of longitudinal wave velocities and the mapping of tangential wave velocities, using the curves constructed during the calibration phase. In other words, once the maps of the longitudinal wave velocity and / or the transverse wave are obtained, a velocity / temperature relationship can be applied to deduce a temperature map in the volume of the room, using the curves constructed during the calibration phase.

[0179] Using velocity measurements to perform temperature mapping offers a significant advantage: employing a non-destructive testing method for parts that can serve other purposes (such as detecting defects or determining local elastic constants) to perform temperature mapping. This saves time and allows for more comprehensive control of a part's characteristics, for example, to validate a manufacturing step or use the data as input for part behavior simulation software.

[0180] The construction of the temperature map can also be used to validate a manufacturing step of the part involving a heat treatment (the part is brought to a set temperature and / or is allowed to cool).

[0181] The construction of the temperature map in step 5 also allows for the estimation of a temperature field inside the part to provide data to predict the microstructure of a material and / or its mechanical properties.

[0182] Step 6 - Detect an area of ​​different temperature in the room

[0183] According to a particular provision, process 100 may then include a step 6 of detecting a different temperature zone in the room by analyzing the temperature map determined in step 5. As an example, step 6 may allow the detection of a point or zone exhibiting a temperature deviation from the expected value or from an adjacent zone considered as a reference zone.

[0184] Thus, process 100 allows for the extraction of quantitative properties from the inspected medium, such as its temperature. Obtaining a temperature map enables the non-destructive characterization of the polycrystalline material part, identifying areas with abnormal thermal properties generated by the manufacturing processes or corresponding to a defect that locally alters the part's thermal properties. A particularly advantageous aspect is that, if a defect is detected in the part, it can then be repaired or replaced using a third-party device.

[0185] computer program product

[0186] According to another aspect, a computer program product is proposed comprising program code instructions for executing the non-destructive testing process 100 in any of its embodiments, when said instructions are executed by a processor.

[0187] storage media

[0188] According to another aspect, a non-transient storage medium is proposed on which is stored a computer program comprising program code instructions to execute the non-destructive testing process 100 in any of its embodiments, when said instructions are read from said non-transient storage medium and executed by a processor.

[0189] Non-destructive testing device

[0190] According to another aspect, a non-destructive testing device is proposed, comprising an ultrasonic probe and electronic circuitry (computer system 200) adapted to implement a process 100.

[0191] As shown schematically in Fig. 12, the computer system 200 can include, connected by a communication bus 210: a processor 201; a random access memory 202; a read-only memory 203, for example of type ROM (“Read Only Memory”) or EEPROM (“Electrically-Erasable Programmable Read Only Memory”); a storage unit 204, such as a hard disk drive (HDD) or a storage media reader, such as an SD card reader (“Secure Digital”); and an input / output interface manager 205.

[0192] The processor 201 is capable of executing instructions loaded into RAM 202 from ROM 203, external memory, a storage medium (such as an SD card), or a communication network. When the computer system 200 is powered on, the processor 201 can read instructions from RAM 202 and execute them. These instructions form a computer program that allows the processor 201 to implement process 100.

[0193] All or part of process 100 can thus be implemented in software form by executing a set of instructions by a programmable machine, for example a DSP (Digital Signal Processor) or a microcontroller, or implemented in hardware form by a dedicated machine or component, for example an FPGA (Field Programmable Gate Array) or ASIC (Application-Specific Integrated Circuit). Generally speaking, the computer system 200 includes electronic circuitry adapted and configured to implement process 100 in software and / or hardware form.

[0194] Alternative implementation methods

[0195] As mentioned previously, typically the ultrasonic probe used is a linear multi-element probe.

[0196] According to a particular arrangement, the ultrasonic probe used is a matrix (two-dimensional) probe, thus allowing Fonde to be focused in three dimensions and obtaining a three-dimensional map of the transverse ultrasonic wave velocity, the longitudinal ultrasonic Fonde velocity and the temperature in a single measurement.

[0197] According to a specific design, the ultrasonic probe used is designed to withstand high temperatures. In particular, it may have a high-temperature resistant coating or be mounted on a shoe machined from a high-temperature resistant material.

[0198] In a specific configuration, the reflection matrix is ​​acquired using an ultrasound-laser system. A pulsed laser source generates ultrasound through thermoelastic conversion; the laser impact is successively focused onto a set of N points on the surface of the structure. The elastic waves are detected at the surface using an interferometer or vibrometer, at a set of M points, which may differ from the N excitation points. A matrix of MxN time-domain responses is thus acquired.

[0199] Method 100 is applicable to all media for which it is possible to obtain ultrasonic wave velocity mapping. Temperature calibration is always possible. This implies that the medium is diffusive to acoustic waves (concrete, composites, medical applications, etc.). However, excessive attenuation or anisotropy in the medium will hinder the mapping process.

Claims

DEMANDS 1. A method (100) for non-destructive testing of a part made of polycrystalline material, the method being carried out by a non-destructive testing device comprising electronic circuitry adapted to implement the method, the method being characterized in that it comprises at least the following steps: - A. measure (1) an ultrasonic reflection matrix of the part associated with a multi-element probe containing N piezoelectric transducers; - B. project (2) the ultrasonic reflection matrix into a focused basis; - C. construct (3) a map of the speed of longitudinal waves and / or D. construct (4) a map of the speed of transverse waves; - E. construct (5) a temperature map in a volume of the material using the longitudinal wave velocity map and / or the transverse wave velocity map; - F. detect (6) a defect or a particular area in the part by analyzing the temperature map determined in step E.

2. A method (100) according to the preceding claim, wherein, when the method includes step D, then step D comprises separating the reflection matrix into four sub-matrices: a first sub-matrix composed of emitters from a portion of the probe located to the left of the focal point, a second sub-matrix of receivers from a portion of the probe located to the left of the focal point, a third sub-matrix composed of emitters from a portion of the probe located to the left of the focal point, and a fourth sub-matrix of receivers from a portion of the probe located to the left of the focal point.

3. Method (100) according to claim 2, wherein, when the method includes step D, then at step D, the reflection matrix is ​​expressed as: | R^ % | = |R dd | + |R pp | + | R dp | + |Rgd|> avec R <ld L a first submatrix, R gg the second submatrix, R g to the third sub-matrix and Ra g the fourth submatrix.

4. Method (100) according to claim 2 or 3, wherein, when the method includes step D, then at step D, the reflection matrix is ​​expressed using three or fewer of the four submatrices.

5. Method (100) according to any one of the preceding claims, wherein, where the method includes step D, then step D comprises weighting the reflection matrix by a function depending on the angle between a transducer and the focal point.

6. Method (100) according to claim 5, wherein the reflection matrix is ​​filtered using a function expressed as h(0) = |sin(20) x 0 a | with h the filter function and 0 the angle between the normal to the probe and the axis along the line joining the center of the probe and the focal point; or according to h(0) = D T (0) x (1 — D L (0)) with h the filter function, 0 the angle between the normal to the probe and the axis along the line joining the center of the probe and the focal point, ε>7(0) the directivity of a grating element for transverse waves and D L (0) the directivity of a network element for longitudinal waves.

7. A method (100) according to any one of the preceding claims, wherein step E comprises performing, for at least one predetermined temperature, a calibration phase comprising placing the part in an environment presenting said predetermined temperature and carrying out steps A to D of the method to construct a map of longitudinal wave velocities and a map of transverse wave velocities, for the predetermined temperature.

8. Method according to claim 7, wherein step E comprises a phase of mapping a room temperature from the mapping of longitudinal wave velocities and / or the mapping of transverse wave velocities, using the curves constructed during the calibration phase.

9. Method (100) according to any one of the preceding claims, wherein the ultrasonic probe used is a matrix probe.

10. Non-destructive testing device for a part made of polycrystalline material characterized in that it comprises electronic circuitry (200) for implementing a method (100) for detecting a defect in a part made of polycrystalline material which comprises at least the following steps: - A. measure (1) an ultrasonic reflection matrix of the part associated with a multi-element probe containing N piezoelectric transducers; - B. project (2) the ultrasonic reflection matrix into a focused basis; - C. construct (3) a map of the speed of longitudinal waves and / or D. construct (4) a map of the speed of transverse waves; - E. construct (5) a temperature map in a volume of the material using the longitudinal wave velocity map and / or the transverse wave velocity map; - F. detect (6) a defect or a particular area in the part by analyzing the temperature map determined in step E.

Images