Frequency domain full-focusing imaging method for bolt thread crack detection

Abstract

The invention relates to a frequency domain full-focus imaging method, in particular to a frequency domain full-focus imaging method for bolt thread detection. The invention aims to solve the problems of huge data volume and long imaging time of a time domain full-focus algorithm image generation technology. According to the method, an ultrasonic phased array is used for detecting bolt crack defects, and a three-dimensional echo signal matrix of all transmitting / receiving combinations is obtained; then three-dimensional Fourier transform is carried out, and a time domain and a space domain are converted into a frequency domain and a wavenumber domain; the emission wave number is kept unchanged, a new wave number value is calculated through Stolt mapping interpolation, two-dimensional slices are extracted and superposed, and scatterer distribution of the target area after two-dimensional Fourier transform is obtained; and finally, carrying out two-dimensional inverse Fourier transform to obtain an ultrasonic image. The invention belongs to the technical field of frequency domain full-focusing imaging.

Description

Technical Field

[0001] The present invention relates to a frequency-domain full-focus imaging method, belonging to the technical field of frequency-domain full-focus imaging. Background Art

[0002] Bolts are important standard fasteners, widely used in multiple fields such as aerospace and wind power generation. The thread root region of the bolt is a key position where structural stress concentration occurs. Under the action of alternating stress for a long time, cracks are extremely likely to form, eventually leading to fatigue fracture of the bolt.

[0003] Traditional bolt defect detection methods, such as magnetic particle inspection and visual inspection, are mainly applicable to the detection of exposed surface defects. However, in the actual application process, the cracks generated by bolts are usually located inside the equipment, making it difficult for the above methods to effectively detect in-service bolts. As an emerging non-destructive testing technology, ultrasonic imaging testing method has been widely used in the field of industrial non-destructive testing in recent years due to its advantages such as high sensitivity, environmental protection and portability. This testing method can detect internal damage of workpieces and is applicable to in-service detection of bolt thread defects.

[0004] The commonly used ultrasonic testing method for bolts is the single-probe testing method. This method uses a single-element piezoelectric wafer to emit and receive ultrasonic waves, and obtains information about the size and location of defects by analyzing the voltage signal that changes with time. However, the single-probe testing method can only evaluate the defects in the area directly below the probe each time. In order to comprehensively inspect the entire bolt, the probe must be moved along the bolt end face to collect all data. This not only takes time, but also cannot meet the real-time requirements in the on-line detection process.

[0005] In recent years, phased array detection technology has emerged and demonstrated its unique advantages. The core of this method lies in controlling the time delay of each element in the array transducer to emit and receive pulses, thereby changing the phase relationship of the sound waves emitted or received by the wafer to reach a certain point in the object, realizing the synthesis, focusing, deflection, etc. of the sound beam, and forming a scanned image. The image generated by conventional phased array detection is a fan-shaped scanned image. Such images often have large deformations and lack a simulated display of the bolt structure. In contrast, the image generated by the time-domain full-focus algorithm can directly present the cross-sectional view of the bolt, thus clearly characterizing the location and size of the defects. However, its huge data volume and long imaging time have become bottlenecks restricting its application efficiency. Summary of the Invention

[0006] In order to solve the problems of the huge data volume and long imaging time of the image generation technology of the time-domain full-focus algorithm, the present invention further proposes a frequency-domain full-focus imaging method for bolt thread detection.

[0007] The technical solution adopted by the present invention to solve the above problems is as follows: The steps of the present invention include:

[0008] Step 1: Use an ultrasonic array to detect the object to be detected, obtain the echo signal data of all transmit / receive combinations, and record these data as a three-dimensional matrix e(t, u, v). The dimensions of the matrix correspond to the transmit position u, the receive position v, and the time axis t;

[0009] Step 2: Perform a three-dimensional Fourier transform on the collected three-dimensional echo signal matrix to convert it from the time domain and spatial domain to the frequency domain and wavenumber domain, obtaining the data E(ω, k u , k v ) after the three-dimensional Fourier transform;

[0010] Step 3: In the frequency domain and wavenumber domain, process E(ω, k u , k v );

[0011] Step 4: Perform an inverse two-dimensional Fourier transform on to obtain the final ultrasonic image

[0012] Furthermore, in Step 2, the frequency response of each transmit-receive pair is denoted as E(ω, u, v), and the expression is:

[0013] E(ω, u, v) = ∫∫f(x, z)G(ω, x - u, z)G(ω, x - v, z)dxdz (1),

[0014] In formula (1), ω is the angular frequency, and f(x, z) is the scattering signal of each imaging grid point (x, z);

[0015] G(ω, x, z) is the propagation Green's function, and the expression is:

[0016]

[0017] Substitute formula (2) into formula (1), then the frequency response E(ω, u, v) is expressed as:

[0018]

[0019] Identify the two-dimensional spatial integral as the Fourier transform of the scatterer distribution F(k x , k z ), and perform a Fourier transform on u and v to obtain:

[0020]

[0021] In formula (4), F is the Fourier transform of the scattering signal f(x, z); k uand k v corresponds to the wavenumber domain of the array transmit / receive signal;

[0022] k u ≈ksinθ1, k v ≈ksinθ2; k is the wavenumber, and θ1 and θ2 are the angles between the transmit / receive element and the scatterer, respectively.

[0023] Furthermore, in step 3, the Stolt interpolation method is applied to the data in F to obtain a uniform data distribution. The Stolt mapping formula is:

[0024] k x = k u + k v (5),

[0025]

[0026] In formulas (5) and (6), k x and k z are the lateral and longitudinal wavenumbers of the image, respectively; by performing a non-linear coordinate transformation, k u , k v and k are mapped to k x and k z which helps to complete the conversion from the data domain to the image domain;

[0027] Keeping the incident wavenumber k u constant, the scattered signal F(k x , k z |k u ) is expressed as:

[0028]

[0029] In formula (7), E(ω, k v |k u ) is the two-dimensional slice data, and S -1 {·} represents the inverse Stolt mapping of the constant k u , and the formula is:

[0030] k v = k x - k u (8),

[0031]

[0032] By recalculating the 2-D Fourier transform for each k u value and then taking the average, the influence of noise and sidelobes is reduced, i.e.:

[0033]

[0034] Further, in step 3, keep k u unchanged and interpolate to calculate the new k v and k values; for each transmitted wavenumber sample k u , extract the two-dimensional slice E(ω, k v |k u ); interpolate the two-dimensional slice onto a two-dimensional grid in the image wavenumber domain through Stolt mapping to obtain F(k x , k z |k u ); superimpose the interpolation results for all k u to obtain the scatterer distribution after two-dimensional Fourier transform of the target area

[0035] Further, by calculating the two-dimensional inverse Fourier transform, an image of the scatterer distribution is obtained:

[0036]

[0037] The beneficial effects of the present invention are as follows: The present invention uses ultrasonic phased arrays to detect bolt crack defects and obtains a three-dimensional echo signal matrix for all transmit / receive combinations; then performs a three-dimensional Fourier transform to convert it from the time domain and spatial domain to the frequency domain and wavenumber domain; keeps the transmitted wavenumber unchanged, interpolates and calculates new wavenumber values through Stolt mapping, extracts two-dimensional slices and superimposes them to obtain the scatterer distribution after two-dimensional Fourier transform of the target area; finally, performs a two-dimensional inverse Fourier transform on it to obtain an ultrasonic image. The present invention significantly shortens the imaging time, improves the detection efficiency, can accurately characterize the position and size of defects, and is applicable to the in-service detection of bolt thread cracks. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is a flowchart of the present invention;

[0039] Figure 2 is a schematic diagram of a phased array transceiver model. DETAILED DESCRIPTION OF THE INVENTION

[0040] DETAILED DESCRIPTION OF THE INVENTION I: As shown in Figure 1 and Figure 2 , a frequency-domain full-focus imaging method for bolt thread detection, step 1, use an ultrasonic array to detect the object to be detected, obtain the echo signal data for all transmit / receive combinations, and record these data as a three-dimensional matrix e(t, u, v), where the dimensions of the matrix correspond to the transmit position u, the receive position v, and the time axis t;

[0041] Step 2: Perform three-dimensional Fourier transform on the collected three-dimensional echo signal matrix, convert it from the time domain and spatial domain to the frequency domain and wavenumber domain, and obtain the data E(ω,k u ,k v ) after three-dimensional Fourier transform;

[0042] Record the frequency response of each transmit-receive pair as E(ω,u,v), and the expression is:

[0043] E(ω,u,v) = ∫∫f(x,z)G(ω,x - u,z)G(ω,x - v,z)dxdz (1),

[0044] In formula (1), ω is the angular frequency, and f(x,z) is the scattering signal of each imaging grid point (x,z);

[0045] G(ω,x,z) is the propagation Green's function, and the expression is:

[0046]

[0047] Substitute formula (2) into formula (1), then the frequency response E(ω,u,v) is expressed as:

[0048]

[0049] Identify the two-dimensional spatial integral as the Fourier transform of the scatterer distribution F(k x ,k z ), and perform Fourier transform on u and v to obtain:

[0050]

[0051] In formula (4), F is the Fourier transform of the scattering signal f(x,z); k u and k v correspond to the wavenumber domain of the array transmit / receive signal;

[0052] k u ≈ ksinθ1, k v ≈ ksinθ2; k is the wavenumber, and θ1 and θ2 are the angles between the transmit / receive element and the scatterer respectively;

[0053] Step 3: In the frequency domain and wavenumber domain, process E(ω,k u ,k v ); Keep k u unchanged, interpolate to calculate new k v and k values; For each transmit wavenumber sample k u , extract the two-dimensional slice E(ω,k v |k u); Interpolate the two-dimensional slice onto the two-dimensional grid of the image wavenumber domain through Stolt mapping to obtain F(k x , k z |k u ); Superimpose the interpolation results for all k u to obtain the scatterer distribution after two-dimensional Fourier transform of the target area

[0054] Apply the Stolt interpolation method to the data in F to obtain a uniform data distribution. The Stolt mapping formula is:

[0055] k x = k u + k v (5),

[0056]

[0057] In formulas (5) and (6), k x and k z are the transverse and longitudinal wavenumbers of the image respectively; by performing a non-linear coordinate transformation, mapping k u , k v and k to k x and k z helps to complete the conversion from the data domain to the image domain;

[0058] Keeping the incident wavenumber k u constant, the scattered signal F(k x , k z |k u ) is expressed as:

[0059]

[0060] In formula (7), E(ω, k v |k u ) is the two-dimensional slice data, and S -1 {·} represents the inverse Stolt mapping of the constant k u , and the formula is:

[0061] k v = k x - k u (8),

[0062]

[0063] By recalculating the 2-D Fourier transform for each k u value and then taking the average, the influence of noise and sidelobes is reduced, that is:

[0064]

[0065] Step 4, perform an inverse two-dimensional Fourier transform on to obtain the final ultrasound image

[0066] By calculating the inverse two-dimensional Fourier transform, an image of the scatterer distribution is obtained:

[0067]

[0068] As described above, it is only a preferred embodiment of the present invention and does not impose any form of limitation on the present invention. Although the present invention has been disclosed above with a preferred embodiment, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to the above-disclosed technical content to obtain equivalent embodiments with equivalent changes within the scope of the technical solution of the present invention. However, as long as it does not depart from the content of the technical solution of the present invention and is based on the technical essence of the present invention, any simple modification, equivalent replacement, and improvement made to the above embodiments still fall within the protection scope of the technical solution of the present invention.

Claims

1. A frequency-domain full-focus imaging method for bolt thread detection, characterized in that, The specific steps include: Step 1: Use an ultrasonic array to detect the object to be detected, obtain the echo signal data of all transmit / receive combinations, and record these data as a three-dimensional matrix e(t, u, v), where the dimensions of the matrix correspond to the transmit position u, the receive position v, and the time axis t; Step 2: Perform three-dimensional Fourier transform on the collected three-dimensional echo signal matrix, convert it from the time domain and spatial domain to the frequency domain and wavenumber domain, and obtain the data E(ω,k u ,k v ) after the three-dimensional Fourier transform; Step 3: In the frequency domain and the wavenumber domain, process E(ω, k u , k v ); Step 4. Perform an inverse two-dimensional Fourier transform on to obtain the final ultrasound image 2. The frequency-domain full-focus imaging method for bolt thread detection according to claim 1, characterized in that In Step 2, the frequency response of each transmit-receive pair is denoted as E(ω, u, v), and the expression is: E(ω, u, v) = ∫∫f(x, z)G(ω, x - u, z)G(ω, x - v, z)dxdz (1), In formula (1), ω is the angular frequency, and f(x, z) is the scattering signal of each imaging grid point (x, z); G(ω, x, z) is the propagation Green's function, and the expression is: Substitute formula (2) into formula (1), then the frequency response E(ω, u, v) is expressed as: Identify the two-dimensional spatial integral as the Fourier transform of the scatterer distribution F(k x ,k z ), and perform Fourier transforms on u and v to obtain: In formula (4), F is the Fourier transform of the scattered signal f(x,z); k u and k v correspond to the wavenumber domain of the array transmit / receive signal; k u ≈ k sin θ1, k v ≈ k sin θ2; where k is the wave number, and θ1 and θ2 are the angles between the transmitting / receiving element and the scatterer, respectively.

3. A frequency-domain full-focus imaging method for bolt thread detection according to claim 1, characterized in that In Step 3, apply the Stolt interpolation method to the data in F to obtain a uniform data distribution. The Stolt mapping formula is: k x = k u + k v (5), In formulas (5) and (6), k x and k z are the horizontal and vertical wave numbers of the image respectively; by performing a non - linear coordinate transformation, k u , k v and k are mapped to k x and k z , which helps to complete the conversion from the data domain to the image domain; Keep the incident wave number k u constant, then the scattered signal F(k x , k z |k u ) is expressed as: In formula (7), E(ω,k v |k u ) is two-dimensional slice data, and S -1 {·} represents the inverse Stolt mapping of the constant k u , and the formula is: k v = k x -k u (8), By recalculating the 2-D Fourier transform for each k u value and then averaging, the influence of noise and sidelobes is reduced, i.e.:

4. A frequency-domain full-focus imaging method for bolt thread detection according to claim 1, characterized in that Keep k unchanged in step 3 u and interpolate to calculate the new k v and k values; for each emitted wavenumber sample k u , extract the two-dimensional slice E(ω, k v |k u ); interpolate the two-dimensional slice onto a two-dimensional grid in the image wavenumber domain through Stolt mapping to obtain F(k x , k z |k u ); superimpose the interpolation results for all k u to obtain the scatterer distribution after two-dimensional Fourier transform of the target area 5. A frequency-domain full-focus imaging method for bolt thread detection according to claim 1, characterized in that By calculating the two-dimensional inverse Fourier transform, the image of the scatterer distribution is obtained:

Images