Structural damage real-time positioning method based on acoustic emission shear wave beam forming phased array

Abstract

The invention belongs to the technical field of structural health monitoring, and discloses a structural damage real-time positioning method based on an acoustic emission shear wave beam forming phased array, which comprises the following steps of: S1, fixing m circular ring phased arrays consisting of acoustic emission sensors on the surface of a structure to be detected; s2, acquiring an acoustic emission signal when the interior of the structure is damaged, and performing Hilbert transform; s3, giving an initial wavelength, and obtaining incoming wave direction estimation by the phased array through minimum mean square distortionless response beam forming; s4, estimating and determining an internal damage position based on the incoming wave directions of the centers of the n circular ring phased arrays; and S5, iterating the wavelength and repeating S3 and S4 to obtain a space point closest to the incoming wave direction. According to the acoustic emission beam forming internal damage positioning method, only plane arrangement is needed, the installation difficulty is reduced, and the detection cost is reduced; through a wavelength iteration method, three-dimensional damage positioning in the structure and accurate estimation of the wavelength of an acoustic emission signal are realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of structural health monitoring based on acoustic emission, and relates to the real-time positioning of damage inside various homogeneous structures (such as metal structures such as steel structures). Specifically, it is a method for real-time positioning of internal damage of a structure through acoustic emission shear wave signals of a circular phased array, and estimating the wavelength of the acoustic emission shear wave signals at the damaged location. Background Art

[0002] The internal damage localization method based on acoustic emission uses a certain algorithm to analyze and locate damage on the signal measured by the acoustic emission sensor installed on the structure; beamforming has been successfully applied to satellite communications, radar, medical imaging, etc.; both have received widespread attention and made progress.

[0003] Internal structural damage location, especially damage location in hard-to-reach areas, is a very important and difficult topic for aerospace, mechanical engineering, construction engineering, etc. When minor damage inside a structure accumulates to a certain extent at a certain location, it will cause devastating damage to the entire structure, resulting in huge personal safety risks and economic losses. How to timely diagnose and locate the damage position when the structural damage is relatively minor is the focus of structural damage detection; however, existing damage location methods, such as arrival time method, time difference method, time reversal method, etc., all require more stringent layout conditions and more sensors to ensure the accuracy of damage location. At present, the positioning of internal damage of large structures has large positioning errors due to interference from external noise, and a large number of sensors need to be deployed, which brings huge time and economic costs.

[0004] In order to overcome these difficulties, a real-time structural damage location method based on acoustic emission shear wave beamforming phased array was invented. It only requires the arrangement of multiple separate circular phased arrays on the structural plane, and the calculation of the received acoustic signal based on the minimum variance distortion-free response beamforming algorithm, which can effectively suppress the sidelobe level, while narrowing the main lobe in the beamforming, and accurately estimating the incoming wave direction of the simulated damage inside the structure. By iterating the wavelength to locate the damage, it effectively avoids the large-scale deployment of sensors in advance, and does not require material property information. Summary of the invention

[0005] In order to accurately locate the internal damage position of a homogeneous structure and perform imaging, the present invention provides a real-time damage location method based on an acoustic emission shear wave beamforming phased array in a homogeneous structure. The main steps of the method are as follows: fix m circular phased arrays composed of n acoustic emission sensors on the surface of the structure to be measured; obtain the acoustic emission signal received by the acoustic emission sensor when the structure is damaged internally and perform Hilbert transform; given the initial wavelength, the phased array phase shifts the acoustic emission signal through the minimum mean square distortionless response beamforming to obtain the estimated direction of the incoming wave; determine the position of the internal acoustic emission source point based on the estimated direction of the incoming wave of the center of the m circular phased arrays and perform internal three-dimensional positioning; obtain the spatial point closest to the ray of the incoming wave direction of the phased array by iterating the wavelength.

[0006] The technical solution of the present invention:

[0007] A method for real-time location of structural damage based on acoustic emission shear wave beamforming phased array comprises the following steps:

[0008] S1. Measurement point arrangement: multiple piezoelectric ceramic sensors are formed into a phased array in the form of a circular array and evenly arranged on the surface of the measured structure to receive the acoustic emission signal generated by the damage;

[0009] S2, based on the acoustic emission signal collected by the phased array, applying Hilbert transform to the acoustic emission shear wave waveform signal to convert the real signal into a complex signal;

[0010] The acoustic emission signal received by the piezoelectric ceramic sensor in the phased array is divided into data. First, the shear wave real signal z(t) is fast Fourier transformed to obtain the signal spectrum; then, the positive half of the spectrum is multiplied by -j, the negative half is multiplied by j, the DC component is set to zero, and the signal is converted back to the time domain; by superimposing the original shear wave signal and the processed signal in the complex domain, the conversion from the real signal to the acoustic emission shear wave complex signal is realized:

[0011]

[0012] Where: x(t) is the complex signal after Hilbert transform, is the imaginary part of the complex signal after Fourier processing, a(t) is the amplitude of the exponential expression of the complex signal, is the phase angle of the complex signal in exponential form;

[0013] S3. Using the known phased array arrangement and the given initial wavelength, the material and the received signal characteristics are estimated; the minimum variance distortionless response beamformer is used to process the acoustic emission shear wave complex signal to estimate the direction of the damage wave; when estimating the actual damage wave direction, the highest point in the power spectrum density is selected as the azimuth and elevation angles for damage location; the details are as follows:

[0014] S3.1, Assume that the signal S(t) in any direction in space is in the form of a plane wave from the direction relative to the center of the phased array The signal matrix X(t) with a length of L received by the n piezoelectric ceramic sensors in the phased array can be assumed to be:

[0015]

[0016] Among them, A(θ,φ) is the phase compensation matrix; N(t) is the noise matrix received by the sensor;

[0017] S3.2, using formula (3) to calculate the projection distance l of each piezoelectric ceramic sensor in the phased array to the center of the phased array at different azimuth angles θ and elevation angles φ, where the azimuth angle θ is 0° to 360° and the elevation angle φ is 0° to 180°;

[0018] l=Rsinθ i (cosφ j cosφ n +sinφ j sinφ n )=Rsinθ i [cos(φ j -φ n )] (3)

[0019] Where R is the radius of the phased array, θ i is the azimuth direction of beam focusing, φ j is the elevation angle direction of beam focusing, φ n is the azimuth direction of the piezoelectric ceramic sensor element relative to the center of the phased array;

[0020] The phase shift matrix of the beam is obtained by using equation (3) through the damaged wavelength, and the phase compensation matrix with azimuth angle θ of 0° to 360° and elevation angle φ of 0° to 180° is obtained by combination. The phase compensation matrix A(θ, φ) is obtained by calculating the geometric relationship of the phased array and the wavelength:

[0021]

[0022] Among them, k i,j represents the wave number vector of the spatial signal in the circular coordinate system of the phased array, P n represents the position vector of the nth piezoelectric ceramic sensor in the spatial coordinate system of the phased array, λ0 is the initial wavelength, and during the iteration process it is the iterative wavelength;

[0023] S3.3. According to the constraint of minimum variance distortion-free response, a set of weighted filters is set to require the signal in the target direction to be completely output without distortion, as shown in formula (6):

[0024] diag(W H a(θ,φ))=I (6)

[0025] Among them, W H is the conjugate transposed matrix of the weighted filter, a(θ,φ) is the phase compensation matrix A(θ i ,φ j ) is the phase compensation vector of the correct incoming wave direction, and I is the unit matrix;

[0026] At the same time, the weighted filter needs to minimize the interference of signals in other directions, as shown in formula (7):

[0027]

[0028] Where X(t) is the complex signal after Hilbert transform;

[0029] The optimized beamformer W is obtained by Lagrange multiplier method optimization. H , and calculate the maximum power P(θ,φ) at the azimuth and elevation angles:

[0030]

[0031] Where R represents the autocorrelation matrix of the acoustic emission signal;

[0032] The autocorrelation matrix of the acoustic emission signal is enhanced by adding the eigenvalue matrix to highlight its characteristics, and the regularization factor f is set to 10 according to the acoustic emission signal intensity. -2 ;

[0033] R -1 =(R+fλ1I) -1 (10)

[0034] Where λ1 represents the maximum eigenvalue of the autocorrelation matrix of the acoustic emission signal, and I represents the unit matrix;

[0035] S4, according to the estimation of the wave direction of the azimuth angle θ and the pitch angle φ of the acoustic emission source in step S3, determine the wave direction of each circular array; solve the focus of multiple rays in space by matrix least squares method, and then determine the damage point; the matrix least squares method needs to consider the wave direction of multiple circular arrays, select the closest point of the space straight line as the damage position, and realize the accurate positioning of the acoustic emission signal; the details are as follows:

[0036] S4.1. Any point in space passing through (x1, y1, z1) and having a direction vector (u x1 ,u y1 ,u z1 ) is defined by the following formula:

[0037]

[0038] S4.2, by introducing two straight lines passing through the points (x1, y1, z1) and (x2, y2, z2), and their direction vectors are (u x1 ,u y1 ,u z1 ) and (u x2 ,u y2 ,u z2 ), the Euclidean lengths are a1 and a2 respectively, and the following expression is used to represent the combination of any two straight lines:

[0039] G×m=d (12)

[0040] in:

[0041]

[0042] S4.3. When multiple lines intersect, use matrix least squares to find the intersection point:

[0043] m L2 =(G T G) -1 G T d (14)

[0044] S4.4. When multiple straight lines do not intersect, the rank of the G matrix is ​​defective, so it is necessary to first perform singular value decomposition on the matrix and then perform the least squares operation:

[0045]

[0046] Where: V and is the orthogonal matrix after singular value decomposition, S P Diagonal matrix after singular value decomposition;

[0047] The spatial position closest to these two lines is MID L2 or MID SVD , respectively, by m L2 or m svd The first three elements of are determined as shown in the following formula:

[0048] MID L2 =m L2 [1:3] or MID svd =m svd [1:3] (16)

[0049] S5, iterate the wavelength of the acoustic emission signal and calculate the sum of the distances from the damage point to each straight line in the Cartesian coordinate system, as shown in formula (17); when the distance meets the convergence criterion: d i-1≤d i ≤d i+1 , terminate the iteration, output the positioning result and the damaged wavelength; otherwise, iterate and calculate i=i+1 and return to step S3;

[0050]

[0051] Where: d i represents the sum of the distances from the damage point to each straight line, and represents the vector from the center of the phased array to the positioning point, and Represents the direction vector of the central ray of the phased array.

[0052] Beneficial effects of the present invention: The method proposed in the present invention suppresses interference and noise signals by minimizing the total output energy of beamforming using variance-free beamforming constraints. Compared with the traditional damage location method using arrival time, this method can achieve adaptive solution of damage wavelength through iterative solution of damage wavelength, which helps to timely identify damage type and damage stage, and achieve high-precision three-dimensional positioning of damage. At the same time, it has the advantages of small sensor layout area, labor saving, simple operation, etc., and has good application value in practical engineering. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a flow chart of wavelength iterative acoustic emission positioning in a specific example of the present invention;

[0054] Figure 2 It is a schematic diagram of the arrangement of phased array sensors in an example of the present invention;

[0055] Figure 3 Schematic diagram of the estimation of the direction of incoming waves in step 3 of the example of the present invention;

[0056] Figure 4 This is a schematic diagram of multi-array damage location in step 4 of an example of the present invention;

[0057] Figure 5 It is a schematic diagram of the acoustic emission source positioning result in the Cartesian coordinate system in an example of the present invention. DETAILED DESCRIPTION

[0058] The specific implementation of the method is further described below in conjunction with the accompanying drawings and technical solutions.

[0059] like Figure 1 As shown, the present invention provides a real-time damage location method based on acoustic emission shear wave beamforming phased array in a homogeneous structure, which estimates the wave direction based on minimum variance distortionless response beamforming, and includes five steps S1-S5:

[0060] In step S1, multiple piezoelectric ceramic sensors are made into sensor clusters in the form of a circular array and evenly distributed on the surface of the structure to be measured. The sensors are numbered in a counterclockwise direction perpendicular to the outer side of the surface of the object to be measured. Taking the damage location of an iron alloy specimen of size 20cm×20cm×20cm as an example, experiments and numerical simulations are carried out for verification. A finite element model of a cubic specimen is established, and the excitation points and receiving points are arranged to simulate the acoustic emission signal and the sensor receiving signal. There are 8 sensors arranged in the phased array, and the center positions of the array are (10, 0, 10), (5, 0, 10), (10, 0, 5) and (5, 0, 5), respectively. The radii of the phased array are 1, 1.5 and 2, respectively, and the units are all cm.

[0061] In step S2, each signal received by the phased array is subjected to Hilbert transformation to convert the real signal into a complex signal. The signals obtained by the experiment and simulation are subjected to Hilbert transformation.

[0062] In step S3, Figure 2 As shown in Figure 3, according to the composition of any signal in the assumed space in S3.1, the projection distance in each direction is calculated using the geometric information of the phased array. The phase difference is calculated by the initial wavelength, and then the phase compensation is obtained for the azimuth angle θ ranging from 0° to 360° and the elevation angle φ ranging from 0° to 180°. Finally, the array phase compensation matrix A in the experiment and simulation is determined respectively. n×ij (θ,φ).

[0063]

[0064] Furthermore, it can be seen from the arrangement of sensors in step S1 that the radius R of the circular array is 1, 1.5 and 2 respectively, all in cm, and the number n of sensors is 8.

[0065] In step S3.3, if Figure 3 As shown in Figure 2, the optimized beamformer W can be estimated using the minimum variance distortion-free response beamforming algorithm. H , and calculate the maximum power P(θ,φ) at the azimuth and elevation angles:

[0066]

[0067] Where: R is the autocorrelation matrix of the acoustic emission signal.

[0068] Furthermore, the autocorrelation matrix of the acoustic emission signal is enhanced by adding the eigenvalue matrix to highlight its characteristics, and the regularization factor f is set to 10 according to the acoustic emission signal intensity. -2 .

[0069] R -1 =(R+fλ1I) -1 (5)

[0070] Where: λ1 represents the maximum eigenvalue of the autocorrelation matrix of the acoustic emission signal, and I represents the unit matrix.

[0071] In step S4, Figure 4 As shown, according to the azimuth and elevation angles obtained in S3 and the position of the center of the phased array in S1, the least square method is implemented using a matrix method to solve the point in space that is closest to the straight line determined by the phased array.

[0072] In step S4.2, the centers of the phased array are (10, 0, 10), (5, 0, 10), (10, 0, 5), and (5, 0, 5), all in cm. According to the calculation in S3, the incoming wave directions (θ1, φ1), (θ2, φ2), (θ3, φ3), and (θ4, φ4) can be obtained respectively, and the matrices G and d can be constructed.

[0073] Furthermore, the four straight lines in space often have no intersection, so it is necessary to first perform singular value decomposition (SVD) on the matrix G and then perform the least squares operation:

[0074]

[0075] The spatial position closest to these two straight lines is m svd 5×1 The first three elements of are determined as shown in the following formula:

[0076] MID svd =m svd [1:3] (7)

[0077] In step S5, Figure 5 As shown, the sum of the distances from the damage point to each ray is iterated, and the iterative process is repeated. When the sum of the distances is the minimum, the iteration is terminated, and the positioning result and the damage wavelength are output. The iterative formula is as follows:

[0078] d i-1 ≤d i ≤d i+1 (8)

[0079] In order to verify the accuracy of damage location in the present invention, the layout position of the finite element model excitation is changed, and the above steps are repeated for positioning. The results of simulating the positioning of different acoustic emission sources are shown in Tables 1 and 2. It can be seen that the present invention can accurately identify the position of the acoustic emission source in the structure based only on the shear wave signal received in the phased array, which is more accurate than the traditional beamforming positioning algorithm and does not rely on wave velocity information. When the damage frequency and position change, the present invention can accurately identify the damage position and wavelength, which effectively verifies the accuracy of the damage location of the present invention.

[0080] Table 1. Numerical simulation results of damage location of the specimen under 150kHz excitation

[0081]

[0082] Table 2. Damage location test results of the specimen under 150kHz excitation

[0083]

Claims

1. A real-time structural damage location method based on acoustic emission shear wave beamforming phased array, characterized in that: The steps include: S1. Measurement point arrangement: multiple piezoelectric ceramic sensors are formed into a phased array in the form of a circular array and evenly arranged on the surface of the measured structure to receive the acoustic emission signal generated by the damage; S2, based on the acoustic emission signal collected by the phased array, applying Hilbert transform to the acoustic emission shear wave waveform signal to convert the real signal into a complex signal; The acoustic emission signal received by the piezoelectric ceramic sensor in the phased array is divided into data. First, the shear wave real signal z(t) is fast Fourier transformed to obtain the signal spectrum; then, the positive half of the spectrum is multiplied by -j, the negative half is multiplied by j, the DC component is set to zero, and the signal is converted back to the time domain; by superimposing the original shear wave signal and the processed signal in the complex domain, the conversion from the real signal to the acoustic emission shear wave complex signal is realized: Where: x(t) is the complex signal after Hilbert transform, is the imaginary part of the complex signal after Fourier processing, a(t) is the amplitude of the exponential expression of the complex signal, is the phase angle of the complex signal in exponential form; S3. Using the known phased array arrangement and the given initial wavelength, the material and the received signal characteristics are estimated; the minimum variance distortionless response beamformer is used to process the acoustic emission shear wave complex signal to estimate the direction of the damage wave; when estimating the actual damage wave direction, the highest point in the power spectrum density is selected as the azimuth and elevation angles for damage location; the details are as follows: S3.1, Assume that the signal S(t) in any direction in space is in the form of a plane wave from the direction relative to the center of the phased array The signal matrix X(t) with a length of L received by the n piezoelectric ceramic sensors in the phased array is assumed to be: Among them, A(θ,φ) is the phase compensation matrix; N(t) is the noise matrix received by the sensor; S3.2, using formula (3), calculate the projection distance l of each piezoelectric ceramic sensor in the phased array to the center of the phased array at different azimuth angles θ and elevation angles φ, where the azimuth angle θ is 0°-360° and the elevation angle φ is 0°-180°; l=Rsinθ i (cosφ j cosφ n +sinφ j sinφ n )=Rsinθ i [cos(φ j -f n )] (3) Where R is the radius of the phased array, θ i is the azimuth direction of beam focusing, φ j is the elevation angle direction of beam focusing, φ n is the azimuth direction of the piezoelectric ceramic sensor element relative to the center of the phased array; The phase shift matrix of the beam is obtained by using equation (3) through the damage wavelength, and the phase compensation matrix with an azimuth angle θ of 0°-360° and an elevation angle φ of 0°-180° is obtained by combination. The phase compensation matrix A(θ, φ) is obtained by calculating the geometric relationship of the phased array and the wavelength: Among them, k i,j represents the wave number vector of the spatial signal in the circular coordinate system of the phased array, P n represents the position vector of the nth piezoelectric ceramic sensor in the spatial coordinate system of the phased array, λ0 is the initial wavelength, and during the iteration process it is the iterative wavelength; S3.

3. According to the constraint of minimum variance distortion-free response, a set of weighted filters is set to require the signal in the target direction to be completely output without distortion, as shown in formula (6): diag(W H a(θ,φ))=I (6) Among them, W H is the conjugate transposed matrix of the weighted filter, a(θ,φ) is the phase compensation matrix A(θ i ,φ j ) is the phase compensation vector of the correct incoming wave direction, and I is the unit matrix; At the same time, the weighted filter needs to minimize the interference of signals in other directions, as shown in formula (7): Where X(t) is the complex signal after Hilbert transform; The optimized beamformer W is obtained by Lagrange multiplier method optimization. H , and calculate the maximum power P(θ,φ) at the azimuth and elevation angles: Where R represents the autocorrelation matrix of the acoustic emission signal; The autocorrelation matrix of the acoustic emission signal is enhanced by adding the eigenvalue matrix to highlight its characteristics, and the regularization factor f is set to 10 according to the acoustic emission signal intensity. -2 ; R -1 =(R+fλ1I) -1 (10) Among them, λ1 represents the maximum eigenvalue of the autocorrelation matrix of the acoustic emission signal; S4, according to the estimation of the wave direction of the azimuth angle θ and the pitch angle φ of the acoustic emission source in step S3, determine the wave direction of each circular array; solve the focus of multiple rays in space by matrix least squares method, and then determine the damage point; the matrix least squares method needs to consider the wave direction of multiple circular arrays, select the closest point of the space straight line as the damage position, and realize the accurate positioning of the acoustic emission signal; the details are as follows: S4.

1. Any point in space passing through (x1, y1, z1) and having a direction vector (u x1 ,u y1 ,u z1 ) is defined by the following formula: S4.2, by introducing two straight lines passing through the points (x1, y1, z1) and (x2, y2, z2), and their direction vectors are (u x1 ,u y1 ,u z1 ) and (u x2 ,u y2 ,u z2 ), the Euclidean lengths are a1 and a2 respectively, and the following expression is used to represent the combination of any two straight lines: G×m=d (12) in: S4.

3. When multiple lines intersect, use matrix least squares to find the intersection point: m L2 =(G T G) -1 G T d (14) S4.

4. When multiple straight lines do not intersect, the rank of the G matrix is ​​defective, so it is necessary to first perform singular value decomposition on the matrix and then perform the least squares operation: Where: V and is the orthogonal matrix after singular value decomposition, S P Diagonal matrix after singular value decomposition; The spatial position closest to these two lines is MID L2 or MID SVD , respectively, by m L2 or m svd The first three elements of are determined as shown in the following formula: MID L2 =m L2 [1:3] or MID svd =m svd [1:3] (16) S5, iterate the wavelength of the acoustic emission signal and calculate the sum of the distances from the damage point to each straight line in the Cartesian coordinate system, as shown in formula (17); when the distance meets the convergence criterion: d i-1 ≤d i ≤d i+1 , terminate the iteration, output the positioning result and the damaged wavelength; otherwise, iterate and calculate i=i+1 and return to step S3; Where: d i represents the sum of the distances from the damage point to each straight line, and represents the vector from the center of the phased array to the positioning point, and Represents the direction vector of the central ray of the phased array.

Images