CN113552219B - A method for ultrasonic self-focusing detection of hole defects in multilayer structures - Google Patents

Abstract

The invention provides an ultrasonic self-focusing detection method for a hole defect of a multilayer structure, which comprises the following steps: (1) determining an initial value of the sound velocity of any non-traversed layer in the multilayer structure, taking the initial value of the sound velocity as a current sound velocity value, and entering the next step; (2) taking the current sound velocity value and the full matrix data as input, adopting a phase shift method to carry out focusing imaging on the current layer, evaluating the current focusing imaging result, and outputting the current focusing imaging result and the current sound velocity value if the evaluation value meets the requirement; if the evaluation value does not meet the requirement, the next step is carried out; (3) optimizing the current sound velocity value by taking the obtained evaluation value as a reference, and returning to the step (2); (4) and (4) repeating the steps (2) to (3) until all layers are traversed, and obtaining the optimal sound velocity value and the optimal image of each layer. The method realizes the automatic focusing of hole defect imaging and improves the robustness of a phase shift method to the sound velocity.

Description

A method for ultrasonic self-focusing detection of hole defects in multilayer structures

technical field

The invention belongs to the technical field of detection methods, and in particular relates to an ultrasonic self-focusing detection method for hole defects in a multilayer structure.

Background technique

In order to integrate the advantages of high strength, light weight, corrosion resistance, durability and flexibility of different materials, multi-layer (multi-layer polymer) composite structures are widely used in the fields of automobile, electronics, energy, and aerospace. High-performance specialty engineering polymers such as polyetheretherketone (PEEK) and polyphenylene sulfide (PPS), with their high specific strength and self-lubricating properties, are increasingly used in multi-layer composite structures for load-bearing components. However, due to the complex manufacturing process, internal void defects in these products are very common and are not allowed in many cases. Therefore, high-precision automatic positioning of hole defects is a hot issue. In general, the main nondestructive testing methods for internal voids in polymer composite products are infrared (IR) thermal imaging and X-ray computed tomography. Infrared thermal imaging uses an infrared thermal imaging camera to capture the local high temperature generated at the hole defect, but the depth of the hole cannot be determined, and the resolution is limited. For X-ray tomography, radiographs at various set angles are collected for high-precision reconstruction of hole defects in products. However, the tedious and complex calculations limit its application in engineering practice.

Ultrasonic scanning has been widely used to detect defects in composite sheets, where the defect region is obtained by superimposing a piece of C-scan data in the time domain. However, the scanning data acquisition is time-consuming and the defect imaging accuracy is limited, which is not suitable for detecting hole defects in multilayer structures. Ultrasonic phased array defect imaging has high detection efficiency and precision in the field of non-destructive testing, and has been widely used in the detection of concrete or metal structures and products. So far, the research on ultrasonic hole defect imaging of polymer composite products is still limited, mainly using the total focusing method (TFM). There are two difficulties in the ultrasound imaging of multi-layered composite structures. On the one hand, coatings and laminate structures make the propagation of ultrasound very complicated. On the other hand, the sound velocity of each polymer layer in a multilayer composite structure is usually unknown because it is highly dependent on the material formation process and the measurement environment. In other words, the polymer layers do not always have the same speed of sound due to inevitable process differences, which can vary over 500 m/s.

Total focus method is a typical time domain imaging method, and gradually become the standard of non-destructive testing. However, the all-focal method was initially only applicable to single-layer structures under the assumption of constant sound velocity. The introduction of ray tracing makes the all-focusing method suitable for multi-layer structure inspection, but it requires iterative calculation of ray tracing, which reduces the inspection efficiency. The Phase Shift Method (PSM) is an angular spectral method that describes the propagation of waves in the frequency-wavenumber domain. It can extrapolate the product surface wavefield to any depth and then build a focused image, and it is suitable for multilayer structures.

However, most ultrasound imaging reconstructions require the sound velocity of each slice as an input, and the deviation of the input sound velocity from the true sound velocity will lead to the deterioration of the reconstructed image. Therefore, the calculation and estimation of the speed of sound is a key factor for the success of the reconstruction. The speed of sound can be calculated using transmission techniques or ray-theoretic homogenization methods, but these methods are difficult to implement and are only suitable for single-layer structures. In addition, due to fluctuations in the manufacturing process, variations in the speed of sound in batches are unavoidable. Therefore, the algorithm with preset fixed sound velocity is not robust enough in hole defect detection of multilayer structure products. The variation of the sound velocity of each layer in the multilayer composite structure hinders the application of phase shift in on-line detection. Therefore, an ultrasonic autofocus imaging method that can automatically determine the sound velocity parameters is urgently needed.

SUMMARY OF THE INVENTION

In order to solve the above-mentioned problems in the prior art, the present invention provides a method for self-focusing imaging of internal hole defects in a multi-layer structure by using an ultrasonic phased array. Ultrasound data and focus all hole defects in the multilayer structure.

A method for ultrasonic self-focusing detection of hole defects in a multilayer structure, comprising the following steps:

(1) Collect the full matrix data of the multi-layer structure, and determine the sound speed range of each layer;

(2) for any untraversed layer in the multilayer structure, determine the initial value of the sound speed of this layer, and use the initial value of the sound speed as the current sound speed value, enter step (3);

(3) Taking the current sound velocity value and the full matrix data as input, adopting the phase shift method to focus and image the current layer to obtain the current hole defect image;

(4) Evaluate the current hole defect image, if the evaluation value meets the requirements, output the current hole defect image and the current sound speed value as the optimal image and the optimal sound speed value respectively; if the evaluation value does not meet the requirements, go to step (5) );

(5) take the obtained evaluation value as a reference, optimize the current speed of sound, and return to step (3);

(6) Repeat steps (2) to (4) until all layers are traversed, and obtain the optimal sound velocity value and optimal image of each layer.

In the above step (1), when determining the sound velocity range of each layer, it can be set according to the performance of the material. For example, non-metallic materials can be set to 0 to 4000 m/s, and metal materials can be set to 0 to 8000 m/s; of course, the sound velocity range of all layers can also be determined to be the same larger range, such as 0 to 10000 m/s.

The invention proposes an ultrasonic self-focusing detection method for hole defects in a multi-layer structure, so as to improve the robustness of the phase shift to the sound speed, and determine the sound speed value of each layer of the polymer when the hole defect is imaged. In general, if the correct sound velocity value is used in the imaging algorithm (phase shift method), its output image (hole defect image) will be focused on the hole position, and the corresponding focusing criterion value (evaluated value) will be maximized. Conversely, bad images resulting from wrong sound velocity values correspond to lower focus criterion values. Therefore, the focus evaluation criterion is used to evaluate the hole defect imaging results and estimate the sound velocity parameters.

Preferably, in step (1), the full-matrix data is obtained by measuring the multilayer structure by a phased array sensor. Among them, the detection parameters of the phased array sensor (such as the frequency of the probe, the number of vibration elements, the spacing of the array elements, the effective aperture and the length of the array elements) are determined according to the material, geometric size and detection range of the multilayer structure to be detected. When measuring, the probe of the phased array sensor can be directly placed on the multi-layer structure.

In step (2), the initial value of the sound speed of the layer is selected by random sampling according to the determined sound speed range of each layer.

Preferably, in step (2), the multi-layer structure is traversed in a top-to-bottom order.

Preferably, in step (3), when focusing and imaging the current layer, the phase offset Φ(k _x , z, ω) is obtained by the following formula:

Φ(k _x , z, ω)=Φ ₁ (k _x , z _m-1 , ω)·Φ ₂ (k _x , z, ω)

Among them, m represents the current layer, m∈[1, M], M is the number of layers of the multi-layer structure, i represents the imaginary unit, k _{z, j} represents the offset wave number of the j-th layer in the depth direction, and d _j represents the Thickness of layer j, k _{z, m} is the offset wavenumber of layer m in the depth direction, z represents the depth of any imaging point in layer m, z _m-1 represents the interface depth of layer m-1;

Φ ₁ represents the product of the phase shift amounts corresponding to the optimal sound velocity of the 1st to m-1 layers, and Φ ₂ represents the phase offset amount corresponding to the current sound velocity of the m th layer.

In order to improve the stability of the imaging method, the sound velocity is estimated layer by layer according to the phase offset of the current layer. For the mth layer, Φ ₁ (k _x , z _m-1 , ω) is fixed, Φ(k _x , z, ω) is adjusted by Φ ₂ (k _x , z, ω) (Φ(k _x , z , ω) , ω)=Φ ₁ (k _x , z _m-1 , ω)·Φ ₂ (k _x , z, ω)), obtain the hole defect image I _m with z in the range of [z _m-1 , z _m ] (x, z), namely the focused imaging result of the mth layer. Then, the quality of _Im (x, z) is evaluated with the focus evaluation criterion, and this evaluation value is used as a reference to optimize the parameter _cm . After obtaining the optimal sound velocity for the mth layer, Φ ₁ (k _x , z _m-1 , ω) is updated and m is set to (m+1) in preparation for hole defect imaging in the next layer. When the number of calculated layers reaches M+1 (m>M), the optimal sound velocity of each layer and the optimal focus image of hole defects are output.

Preferably, in step (3), before the full matrix data is input to the phase shift method, two-dimensional Fourier transform preprocessing is performed on the full matrix data.

The focused imaging results can clearly describe the void defects of material interfaces and components. The grayscale value of the image near the hole is required to be large enough for it to form a good contrast with the background and have clear boundaries, which means that the image should be sharp. Therefore, the degree of focus can be quantified by image sharpness.

Preferably, in step (4), the maximum focus evaluation criterion is used to evaluate the hole defect image of the current layer.

The better the focusing effect in the image ROI, the clearer the current hole defect image obtained, and the higher the output value (evaluation value) of the focusing evaluation criterion. When the evaluation value is the largest (satisfying the requirements), the current hole defect image is the clearest and the imaging result is the most accurate, which is the optimal image, and the corresponding current sound speed is the optimal sound speed.

There are many focusing evaluation criteria, each of which has its own characteristics and is suitable for different conditions. As a further preference, the focusing evaluation criteria is the Brenner gradient criterion, the Tenenbaum gradient criterion or the normalized variance criterion.

Preferably, when the initial value of the sound velocity of the current layer is determined in step (2) and the current sound velocity value is optimized in step (5), a global optimization algorithm not based on gradient is used for processing.

The evaluation value of the current hole defect image is fed back to the global optimization algorithm, and the global optimization algorithm uses the evaluation value as a reference to optimize the current sound velocity value.

As a further preference, the global optimization method is particle swarm optimization, genetic algorithm or differential evolution algorithm.

More preferably, the differential evolution algorithm is used to determine the initial value of the sound velocity of the current layer and to optimize the current sound velocity. Through the present invention, the phase shift algorithm can be used to image the defects of each layer structure, the imaging quality of each layer can be evaluated by the maximum focusing evaluation criterion, and the imaging sound velocity value can be optimized by using the evaluation value as a reference by the global optimization algorithm. The optimal sound speed is obtained, and the sound speed value is also a key parameter of material properties, and then focused imaging of hole defects can be achieved, and accurate defect locations in the multilayer structure can be obtained.

The automatic focusing imaging method of the present invention realizes automatic focusing by performing layer-by-layer imaging evaluation and sound speed optimization on the hole defects and interlayer interfaces of the multi-layer structure in the order from top to bottom, and optimizes the optimal value determined by the previous layer. The sound velocity is applied to the next-layer phase migration method imaging, which improves the robustness of the phase migration method to the sound speed and the accuracy of the hole defect position measurement.

Compared with the prior art, the beneficial effects of the present invention are:

The invention adopts the phase shift method to image the defects of each layer structure of the multi-layer structure, and evaluates the imaging results, and uses the evaluation value as a reference to not optimize the sound speed value, so as to obtain the optimal sound speed and optimal sound speed of each layer. image, realizes the automatic focusing of hole defect imaging, and improves the robustness of the phase shift method to the speed of sound. Finally, by identifying and locating the defects in the optimal image, the position of each defect in the multi-layer structure can be determined with high accuracy.

Description of drawings

Fig. 1 is a schematic diagram of the measurement process of a multi-layer structure by a phased array sensor;

2 is a flowchart of an embodiment of the present invention;

In Figure 3: (a) is a schematic diagram of the measurement of a single-layer polymer by a phased array sensor and a schematic diagram of the real structure of the single-layer polymer; (b) is a schematic diagram of the measurement of a double-layer polymer by a phased array sensor and the double-layer polymerization. The schematic diagram of the real structure of the polymer; (c) is the schematic diagram of the measurement of the three-layer polymer by the phased array sensor and the real structure of the three-layer polymer;

In Fig. 4: (a) is the optimal hole defect image of the monolayer polymer obtained by the Brenner gradient criterion; (b) is the optimal hole defect image of the monolayer polymer obtained by the Tenenbaum gradient criterion; (c) is The optimal hole defect image of the monolayer polymer obtained using the normalized equation criterion;

In Fig. 5, (a) is the optimal hole defect image of the bilayer polymer obtained by the Brenner gradient criterion; (b) is the optimal hole defect image of the bilayer polymer obtained by the Tenenbaum gradient criterion; (c) is The optimal hole defect image of the bilayer polymer obtained using the normalized equation criterion;

In Fig. 6, (a) is the optimal hole defect image of the three-layer polymer obtained by using the Brenner gradient criterion; (b) is the optimal hole defect image of the three-layer polymer obtained by using the Tenenbaum gradient criterion; (c) is Optimal void defect image of the three-layer polymer obtained using the normalized equation criterion.

Detailed ways

The technical solutions of the present invention will be further described below with reference to the accompanying drawings.

As shown in Figure 2, a method for ultrasonic self-focusing detection of hole defects in a multilayer structure includes the following steps:

(1) Use the phased array sensor to collect the full matrix data of the multi-layer structure, and determine the sound velocity value of each layer in the range of 0-10000m/s according to the material of each layer;

(2) for any untraversed layer in this multilayer structure, utilize differential evolution algorithm to determine the initial value of sound speed of this layer, and take this initial value of sound speed as current sound speed value, enter step (3);

(3) Perform two-dimensional Fourier transform preprocessing on the full matrix data, take the current sound velocity value and the preprocessed full matrix data as input, and use the phase shift method to focus and image the current layer to obtain the current hole defect image;

(4) The current hole defect image is evaluated using the maximum focus evaluation criterion. If the evaluation value meets the requirements (when it is the maximum value), the current hole defect image and the current sound speed value are output as the optimal image and the optimal sound speed value respectively; If the evaluation value does not meet the requirements, then enter step (5);

(5) feedback the obtained evaluation value to the differential evolution algorithm as a reference, optimize the current speed of sound, and return to step (3);

(6) Repeat steps (2) to (4) until all layers are traversed, and obtain the optimal sound velocity value and optimal image of each layer.

Phase shift method:

The phase-shift method started out for the imaging of hole defects in uniform media. The basic principle is to extrapolate downward from the surface wave field to obtain the wave field of the entire measurement area to achieve the focusing of hole defects. The sound field propagation model in a two-dimensional isotropic medium is shown in formula (1):

Set the array arrangement direction of the phased array sensor as the positive x-axis, and the depth direction of the multilayer structure as the positive z-axis; where c is the speed of sound, p is the sound pressure, which is the space (x, z) and time (t) function.

The measurement of the multi-layer structure by the ultrasonic phased array is shown in Figure 1. In the figure, M represents the number of layers of the multi-layer structure, d _m represents the thickness of the m-th layer, and _cm represents the current sound velocity value of the m-th layer, where m ∈ [1, M].

The probe of the phased array sensor is placed at the position of z=0 (the upper surface of the multilayer structure), and the hole defect is located in the interior of the multilayer structure with z>0.

If the material is isotropic, the Helmholtz equation is obtained by Fourier transforming the x and t directions:

where ω and k _x are the frequency and the wavenumber in the x-direction, respectively.

Equation (2) can be obtained analytically as shown in equation (3):

Among them, U(k _x , ω) and D(k _x , ω) are up-going waves and down-going waves, respectively, which can be obtained through boundary constraints.

k _z is the wave number in the z direction, which is defined by equation (4):

sgn is a sign function, which ensures that k _z and ω are in the same direction; k _x is the wave number in the x direction.

Upgoing waves are generated by reflections from hole defects, so defects can be regarded as sound sources of upgoing waves. The wave field of the detection surface is p _U (x, 0, t), which can be collected by the probe of the phased array sensor and used as the constraint condition of the upgoing wave. Therefore, U(k _x , ω) can be obtained by transforming p _U (x, 0, t) into the ω-k _x space using formula (5):

得到，如式(6)：i is an imaginary unit, the wave field in the measurement area can be determined by the phase shift factor

get, as formula (6):

外推得到，如式(7)：At the same time, the sound source signal s(x, 0, t) is also excited on the surface by the phased array sensor, which can be used as the boundary condition of the downgoing wave, so the downgoing wave field can pass the phase shift factor

It can be extrapolated, such as formula (7):

Fourier transform of P _U (x, z, ω) and P _D (x, z, ω) by P _U (k _x , z, ω) and P _D (k _x , z, ω) in the k _x direction converted. That is, down-going (incident) waves and up-going (reflected) waves at a specific frequency and location are obtained simultaneously.

Therefore, the reflectivity in the measurement area can be obtained by formula (8):

Since the upward wave will focus at the hole defect, the reflectivity at the hole defect is much larger than other positions, so it can reflect the position of the defect.

In order to enhance the stability of imaging, the imaging conditions of formula (9) are used:

是P_D(x，z，ω_j)的共轭，它使得分母为实数，N是采用的离散的频率数目，ε是一个较小的常数，它避免分母过小，随着下行波外推深度的增加而增加，如式(10)所示in,

is the conjugate of P _D (x, z, ω _j ), which makes the denominator a real number, N is the number of discrete frequencies used, and ε is a small constant that prevents the denominator from being too small and extrapolated with the downgoing wave increases as the depth increases, as shown in Eq. (10)

ε=ε(ω, z)=λ[max(|P _D (x, z, ω)| ² )] (10)

where λ is a coefficient between 0 and 1.

In practical applications, multi-layer detection is common, and it can be simplified to the model shown in Figure 1. Here, ultrasonic waves pass through the M-layer material to the hole defect, and are then reflected and received by the phased array sensor.

The change in amplitude of the wave passing through the interface can be modeled by the transmission coefficient, which is a function of the impedance of the material and the angle of incidence of the wave. Under the narrow beam assumption, the variation of the transmission coefficient with the angle of incidence is negligible.

Therefore, as shown in equation (11), there is a proportional relationship between the wavefield amplitudes above and below the interface.

和

分别表示第n层界面的上侧和下侧。因此在z处的相位偏移；量修正为式(12)：in,

and

represent the upper and lower sides of the n-th layer interface, respectively. Hence the phase shift at z; the amount is corrected to equation (12):

where Φ(k _x , z, ω) is the corrected phase offset, and the current wavenumber k _{z, m} of the mth layer in the z direction is given by equation (13):

Among them, cm is the current sound velocity value of the _mth layer.

Similarly, the wave field extrapolation of _PD in the case of multiple layers can be directly obtained by taking the conjugate of Φ(k _x , z, ω) in Eq. (12).

In practical applications, since the amplitude scaling effect introduced by the interface has little influence on the image structure and can be ignored, it is the relative amplitudes in each layer that reflect the hole defects. Therefore, the phase shift amount of the multilayer structure conforms to equation (12).

There are many focusing criteria, each of which has its own characteristics and is suitable for different conditions. The focusing evaluation criteria for evaluating the current hole defect image in the above technical solution can use the Brenner gradient criterion, the Tenenbaum gradient criterion or the normalized variance criterion.

The Brenner gradient criterion computes the finite difference between each pixel and its surrounding two-pixel interval, written as (14):

where I _{x, z} =I(x, z) is the pixel intensity of the image at (x, z);

The Tenenbaum gradient is convolved with the image I(x, z) by the Sobel operator to extract boundary information, and the sum of the squares of the gradient components is added to obtain, as shown in formula (15):

The Sobel operator is defined as formula (16):

Normalized variance computes the variance between image pixels and normalizes with the mean intensity, written as (17):

where μ represents the expected value of I(x, z).

In this embodiment, a 64-element phased array sensor (Olympus. Ltd, USA), with a center frequency of 5 MHz and an array element spacing of 0.6 mm, is used as an ultrasonic transceiver. The full matrix data was acquired by 64/64 OEM-PA (AOS. Ltd, America).

The single-layer, double-layer, and three-layer polymers were measured by the above autofocus imaging method. During the measurement, the probe of the phased array sensor was directly placed on the polymers of three different structures, and the polymers of each structure were measured. The full matrix data of , as shown in Figure 3.

Test Example 1. Measurement of Monolayer Polymer

The single-layer polymer is an acrylonitrile-butadiene-styrene (ABS) layer with three hole defects (layer thickness is 40mm), and the three hole defects are artificially fabricated, and its real structure is shown in Figure 3(a) .

The Brenner gradient criterion, the Tenenbaum criterion and the normalized variance criterion were used for auto-focusing, respectively, and the focused sound velocities (optimal sound velocity values) of the ABS layer were 2194m/s, 2194m/s and 2219m/s, respectively. The output hole defect image (Optimal image) is shown in Figure 4.

It can be seen from Figure 4 that the depth of hole defects and the relative positions of the hole defects in the optimal images output by the three-criteria self-focusing methods are relatively close, and are consistent with the real structure of the ABS layer in Figure 3(a). Because there is no specific installation position between the test component and the sensor during the experiment, the actual lateral position of the hole defect is unknown, but the depth of the hole defect and the relative position between the hole defects can reflect the localization accuracy of the method.

Test example 2. Double layer polymer

The double-layer polymer was formed by connecting a wedge (SB26-N0L, Doppler. Ltd, China) (layer thickness 20mm) to the above-mentioned inverted ABS layer, and its actual structure is shown in Fig. 3(b).

The autofocus method adopts Brenner gradient criterion, Tenenbaum gradient criterion and normalized variance criterion respectively.

Table 1 The optimal sound velocity values of each layer obtained by three focusing evaluation criteria

The depths of hole defects in the ABS layer and the relative positions between the hole defects in the three optimal images in Fig. 5 are similar, and are consistent with the real structure of the bilayer polymer in Fig. 3(b). In addition, the optimal sound velocity value of the ABS layer calculated in this experiment is basically consistent with the optimal sound velocity value obtained in Experiment 1. It shows that the autofocus imaging method of the embodiment of the present invention has high robustness.

Test Example 3. Three-layer polymer

A thin (layer thickness of 5.7mm) polyetheretherketone (PEEK) layer is covered on the ABS layer, and the wedge is covered on the polyetheretherketone layer. The real structure is shown in Figure 3(c). Show.

The autofocus method adopts the Brenner gradient criterion, the Tenenbaum gradient criterion and the normalized variance criterion, respectively. The optimal sound velocity values of each layer are shown in Table 2, and the imaging results of hole defects in each layer are shown in Figure 6.

Table 2 The optimal sound velocity values of each layer obtained by three focusing evaluation criteria

It can be seen from Table 2 that the sound velocity parameters estimated by the Brenner gradient criterion and the normalized variance criterion are very close, and the output image is better focused on the hole defect. For the Tenenbaum gradient criterion, the focusing effect at the gap in the middle of the third layer is not as good as that of the other layers. The estimated optimal sound velocity values for the wedge and ABS layer are close to the results of experiments 1 and 2, respectively. As for the localization accuracy, the hole defect depths and the relative positions between the hole defects in Fig. 6 agree with the real structure in Fig. 3(c). As the depth of the hole increases, the imaging amplitude decreases significantly, and the third interface completely disappears. This is because the sound attenuation coefficient of the polymer is large, and the interface between the different layers also has a significant attenuation effect on the sound wave. In order to further improve the imaging quality, an attenuation compensation method can be introduced.

Claims (2)

1. a multi-layer structure hole defect ultrasonic self-focusing detection method, is characterized in that, comprises the following steps: (1) Collect the full matrix data of the multi-layer structure, and determine the sound speed range of each layer; (2) for any untraversed layer in this multi-layer structure, determine the initial value of sound speed of this layer, and take this initial value of sound speed as the current speed of sound value, enter step (3); (3) The current sound velocity value and the full matrix data are used as input, and the phase shift method is used to focus and image the current layer to obtain the current hole defect image; (4) Evaluate the current hole defect image. If the evaluation value meets the requirements, output the current hole defect image and the current sound speed as the optimal image and the optimal sound speed value respectively; if the evaluation value does not meet the requirements, go to step (5) ; (5) take the gained evaluation value as a reference, optimize the current speed of sound value, and return to step (3); (6) Repeat steps (2) to (5) until all layers are traversed, and obtain the optimal sound velocity value and optimal image of each layer; In step (2), the multi-layer structure is traversed according to the order from top to bottom; In step (3), before the full matrix data is input to the phase shift method, the full matrix data is subjected to two-dimensional Fourier transform preprocessing; When focusing and imaging the current layer, the phase offset Φ(k _x , z, ω) is obtained by the following formula: Φ(k _x , z, ω)=Φ ₁ (k _x , z _m-1 , ω)·Φ ₂ (k _x , z, ω)

Among them, m represents the current layer, m∈[1, M], M is the number of layers of the multi-layer structure, i represents the imaginary unit, k _{z, j} represents the offset wave number of the j-th layer in the depth direction, and d _j represents the Thickness of layer j, k _{z, m} is the offset wavenumber of layer m in the depth direction, z represents the depth of any imaging point in layer m, z _m-1 represents the interface depth of layer m-1; Φ ₁ represents the product of the phase offsets corresponding to the optimal sound speed of the 1st to m-1 layers, and Φ ₂ represents the phase offset corresponding to the current sound speed of the mth layer; In step (4), the current hole defect image is evaluated using the maximum focus evaluation criterion; when the evaluation value is the largest, it is considered that the evaluation value meets the requirements; the focus evaluation criterion is the Brenner gradient criterion, the Tenenbaum gradient criterion or the normalized variance criterion. ; When determining the initial value of the sound velocity of the current layer in the step (2) and optimizing the current sound velocity value in the step (5), the global optimization algorithm is used for processing; The evaluation value of the current hole defect image is fed back to the global optimization algorithm, and the global optimization algorithm uses the evaluation value as a reference to optimize the current sound velocity value; The global optimization method is particle swarm algorithm, genetic algorithm or differential evolution algorithm. 2 . The ultrasonic self-focusing detection method for hole defects in a multilayer structure according to claim 1 , wherein in step (1), the full-matrix data is obtained by measuring the multilayer structure by a phased array sensor. 3 .

Images