Ultrasonic defect signal automatic detection method based on instantaneous a-scan signal statistics

Abstract

The application provides an ultrasonic defect signal automatic detection method based on instantaneous a-scan signal statistics, which comprises the following steps: step 1, acquiring all ultrasonic a-scan signals [U]t1 of a material to be detected at time t; step 2, acquiring a signal statistics number table [D]t1 at time t; step 3, extracting all ultrasonic signal types in the signal statistics number table [D]t1 and performing analysis; step 4, taking the analysis result as reconstruction data and inputting the reconstruction data into a signal statistics table [D]t2 at time t; step 5, reconstructing all ultrasonic a-scan signals [U]t1 at time t into [U]t2 according to the mapping relationship between the signal statistics number table [D]t1 at time t and the signal statistics table [D]t2; and step 6, repeating steps 1-5 to reconstruct and superimpose all ultrasonic a-scan signals at all times, so as to obtain the reconstructed ultrasonic detection result. The application can realize automatic processing of ultrasonic data, can directly obtain defect distribution information after the original data is input, and is helpful for integration and development of multi-field technologies based on ultrasonic nondestructive testing technology.

Description

TECHNICAL FIELD

[0001] The present application relates to an ultrasonic defect signal automatic detection method based on instantaneous a-scan signal statistics, and belongs to the technical field of ultrasonic nondestructive testing. BACKGROUND

[0002] At present, nondestructive testing technology is widely used in the processing of various instruments and equipment and the maintenance and support of in-service equipment. Ultrasonic flaw detection is a nondestructive testing method that uses the differences in acoustic properties of materials and defects to inspect internal defects of materials by reflecting the waveform of ultrasonic wave propagation and the energy change of penetration time. It has the advantages of large detection thickness, high sensitivity, fast speed, low cost, no harm to human body, and can locate and quantify defects, and is widely used in the field of nondestructive testing.

[0003] At present, the ultrasonic nondestructive testing technology has the following defects: 1. Most of the current marketized ultrasonic detection equipment is a-scan type, which is not intuitive in displaying defects and is easily affected by subjective and objective factors; 2. The current more intuitive ultrasonic c-scan technology needs to set gate parameters such as gate width, gate starting point and gate threshold, which requires higher nondestructive testing experience of the operator, and is not conducive to the further promotion of ultrasonic nondestructive testing technology; 3. The results obtained by ultrasonic nondestructive testing are not easy to save due to the great change of gate parameters and other subjective and objective factors, so it is difficult to reuse the ultrasonic detection results in subsequent combination of various nondestructive testing technologies, and the ultrasonic detection results have more background noise, which requires experienced inspectors to accurately identify the defect types.

[0004] In existing research, Hao Yanpeng et al. (Patent No. CN110320274B) of South China University of Technology proposed a three-pillar insulator internal defect reconstruction method based on ultrasonic scanning principle, which discloses a three-pillar insulator internal defect reconstruction method based on ultrasonic scanning principle, including the following steps: S1, building an ultrasonic detection system; S2, the ultrasonic detection system detects the same position of the three-pillar insulator, records the ultrasonic reflection echo information of each detection position; S3, using the analogy method to judge the ultrasonic reflection echo information of each detection position, and determining the defect position; S4, the ultrasonic detection system performs ultrasonic scanning near the defect position, and records the ultrasonic reflection echo information of each detection position; S5, constructing the corresponding relationship between the ultrasonic reflection echo information and the defect depth and defect size, and obtaining the defect reconstruction diagram based on the detection results of each scanning position. The present application can efficiently, accurately and intuitively identify and locate the internal defects of the three-pillar insulator and quantify the defects. Although this method uses the echo difference between the normal area and the defect area of the material, the point-by-point calculation efficiency is low, and it is only suitable for small components such as insulators, and the calculation performance for large-area materials is poor.

[0005] Lin Jing et al. (patent number CN104764804A) of Xi'an Jiaotong University discloses a kind of ultrasonic Lamb wave local cyclic scanning probability reconstruction tomography method, this method first plans detection area, and sensor array is arranged on detection component, selects Lamb wave mode and designs excitation signal, then respectively gathers each sensor pair reference signal and damage signal, detection area is divided to obtain several sectors, and with adjacent several sectors as detection sub-region, after each time completing sub-region detection, sub-region is rotated clockwise one sector as next cycle detection sub-region, again carry out data selection preprocessing, according to the distance between sensor pair and the velocity of selected Lamb mode Effective data length is calculated, then local cyclic probability reconstruction tomography calculation is carried out, finally data fusion and image splicing are carried out, the present application has strong anti-interference ability, high precision, can effectively eliminate the influence of multiple faults etc.

[0006] Yuxiaonan et al. (patent number CN109087246B) of Jiangnan University discloses an ultrasonic image reconstruction method based on sparse reconstruction, relating to the field of image processing, the method comprises: obtaining the ultrasonic echo time domain signal of the sample to be measured using an ultrasonic probe, constructing an overcomplete dictionary using a reference echo signal, performing sparse decomposition on the ultrasonic echo time domain signal of the sample to be measured using the overcomplete dictionary to obtain sparse decomposition coefficients, reconstructing the ultrasonic echo time domain signal according to the sparse decomposition coefficients and the overcomplete dictionary, and combining the reconstructed ultrasonic echo time domain signals according to the spatial positions to form a reconstructed ultrasonic image of the sample to be measured; the method can improve the horizontal and vertical resolution of the image, and can quickly and effectively observe the position, size and distribution of micro-defects in the sample. Although this method can effectively improve the horizontal and vertical resolution of the ultrasonic image, the construction dictionary based on Gabor function needs a large amount of data for fitting, and its applicability is poor in the environment of ultrasonic nondestructive testing with limited data volume. SUMMARY

[0007] The present application is to solve the technical problem of manual setting of gate parameters in ultrasonic testing, and further proposes an ultrasonic defect signal automatic detection method based on instantaneous a-scan signal statistics.

[0008] The technical solution adopted by the present application to solve the above problems is as follows: the present application proposes an ultrasonic defect signal automatic detection method based on instantaneous a-scan signal statistics, which comprises:

[0009] Step 1: obtain all ultrasonic a-scan signals [U]t1 of the material to be measured at time t;

[0010] Step 2: Statistics of all ultrasonic signals in the ultrasonic a-scan signal [U]t1 is obtained to obtain a signal statistics table [D]t1 at t time;

[0011] Step 3: All ultrasonic signal types in the signal statistics table [D]t1 are extracted and analyzed;

[0012] Step 4: The analysis result is counted as reconstruction data into the reconstruction signal statistics table [D]t2 at t time;

[0013] Step 5: According to the mapping relationship of the physical position of the signal statistics table [D]t1 and the reconstruction signal statistics table [D]t2, all ultrasonic a-scan signals [U]t1 at t time are reconstructed into [U]t2, and the extraction of the defect information in the ultrasonic a-scan signal is completed;

[0014] Step 6: Steps 1-5 are repeated to reconstruct and superimpose all time ultrasonic a-scan signals to obtain the reconstructed ultrasonic detection result.

[0015] Optionally, the step of obtaining all ultrasonic a-scan signals [U]t1 of the measured material at t time in step 1 comprises:

[0016] Step 1.1: The material is excited by using a phased array ultrasonic array element;

[0017] Step 1.2: Obtain the change of pressure at the excited position of the material with time, convert the sound pressure into a voltage signal through a piezoelectric sensor, and form a-scan data of a point on the surface of the material;

[0018] Step 1.3: Repeat steps 1.1-1.2 to obtain all a-scan data of the material surface, assemble all a-scan data of the material surface according to the physical position, and obtain a three-dimensional matrix [U]t of the sound pressure signal of the material surface changing with time;

[0019] Step 1.4: Obtain all ultrasonic a-scan signals [U]t1 of the measured material at t time based on the three-dimensional matrix [U]t of the sound pressure signal of the material surface changing with time.

[0020] The application can effectively solve the problem that the detection personnel requires high experience and needs a long time to accumulate detection experience due to the gate parameters, and realizes full-automatic processing of ultrasonic a-scan data, and effectively reduces the popularization and use threshold of ultrasonic nondestructive testing technology.

[0021] Optionally, the step of exciting the material by using a phased array ultrasonic array element in step 1.1 comprises:

[0022] Step 1.1.1: ultrasonic a-scan is performed on the measured material to obtain the propagation route of the phased array ultrasonic wave in the wedge and the measured material, and the radiated sound wave ray of a single ultrasonic array element is obtained based on the propagation route of the phased array ultrasonic wave in the wedge and the measured material;

[0023] Step 1.1.2: the radiated sound wave ray of a single ultrasonic array element is modeled to obtain the particle vibration velocity generated in the wedge;

[0024] Step 1.1.3: the integral function in the particle vibration velocity is calculated by memory approximation to obtain a model of the ultrasonic longitudinal wave radiated by a single ultrasonic array element with a length of lx and a width of ly into the solid medium;

[0025] Step 1.1.4: the ultrasonic array element is divided into m blocks and n blocks in the x direction and the y direction respectively, and based on the high-frequency approximation and the stationary phase method, a three-dimensional sound field model of a single array element in the wedge is obtained in combination with the model of the ultrasonic longitudinal wave radiated by a single ultrasonic array element into the solid medium;

[0026] Step 1.1.5: the plane wave transmission coefficient T of the ultrasonic wave passing through the liquid coupling layer between the wedge and the measured material is introduced into the three-dimensional sound field model of a single array element in the wedge to obtain a three-dimensional sound field model of a single array element in the measured material;

[0027] Step 1.1.6: the propagation of the ultrasonic wave inside the measured material is obtained according to the three-dimensional sound field model of a single array element in the measured material, when the material has a defect, the ultrasonic wave propagation distance changes, and the related information of the first echo is calculated to complete the excitation of the material;

[0028] The expression of the particle vibration velocity generated in the wedge is:

[0029]

[0030] In formula (1), K p (θ′) is the sound field directivity function, r is the length of the path between the center of the array element and the point x, θ′ is the angle between the path of the point x to a certain point x′ on the array element and the horizontal plane, k p1 is the wave number, p0(ω) is the surface pressure of the array element, κ is the wave velocity ratio, G is the gravitational constant, c p1 is the longitudinal wave velocity, i is the number of incident points, S is the surface area of the array element, d′ p is the unit vector between the point x and the certain point x′ on the array element;

[0031] The expression of the sound field directivity function K p (θ′) is:

[0032]

[0033] In formula (2), κ is the ratio of the longitudinal wave speed to the transverse wave speed of the sound wave propagating in the solid;

[0034] The expression of the model of the single ultrasonic element radiating the ultrasonic longitudinal wave into the solid medium is:

[0035]

[0036] In formula (3), is a directivity function in the spherical coordinate system;

[0037] The directivity function in the spherical coordinate system The expression of is:

[0038]

[0039] The expression of the relevant information of the first echo is:

[0040]

[0041] Optionally, the step of obtaining the signal statistical number table [D]t1 at the time t in step 2 comprises:

[0042] All types of ultrasonic signals in the entire ultrasonic a-scan signal [U]t1 of the measured material at the time t are extracted for statistics, and a signal statistical number table [D]t1 at the time t is established. The signal statistical number table [D]t1 at the time t includes all types of ultrasonic signals and the number of ultrasonic signals of the same type.

[0043] Optionally, the step of extracting all types of ultrasonic signals in the signal statistical number table [D]t1 and performing analysis in step 3 comprises:

[0044] Step 3.1: extracting all types of ultrasonic signals in the signal statistical number table [D]t1 at the time t, and folding repeated signal values;

[0045] Step 3.2: calculating the difference between each type of ultrasonic signal after folding and other types of ultrasonic signals,

[0046] Step 3.3: taking the ratio of the difference between each type of ultrasonic signal and other types of ultrasonic signals and the average difference between the ultrasonic signals closest to 1 / 3 of the ultrasonic signals of the corresponding type and all types of ultrasonic signals as the reconstruction data, and completing the analysis of all signal types in the signal statistical number table [D]t1.

[0047] Through the reconstruction of the a-scan data, the regularity of the defect information is realized, and the background information is flattened, which is beneficial to the secondary use and development of the ultrasonic nondestructive testing data.

[0048] The expression of the folded repeated signal values is:

[0049]

[0050] In formula (6), num is the number of the corresponding type of ultrasonic signal, p1, p2, p n are different types of ultrasonic signals, p i is the difference between the ith ultrasonic signal and other types of ultrasonic signals.

[0051] The expression of the difference between each type of ultrasonic signal and other types of ultrasonic signals is:

[0052] D = [d1, d n ], i = 1, 2, K, n.(7).

[0053] d i = [p i -p j |, num i ], j = 1, 2, K, n.(8).

[0054] In formula (7) and (8), D is a difference matrix composed of the difference between each type of ultrasonic signal and other types of ultrasonic signals, d1, d n are the difference sequences between the first and nth ultrasonic signals and other types of ultrasonic signals, d i is the difference between the ith ultrasonic signal and other types of ultrasonic signals, and the calculation method is the difference between the ith ultrasonic signal p i and the jth ultrasonic signal p j ;

[0055] The calculation formula of the reconstructed data is:

[0056]

[0057]

[0058] In formula (9) and (10), RD(p i ) is the difference between the ultrasonic signal closest to the corresponding type of ultrasonic signal and all other types of ultrasonic signals, F(p i ) is the ratio of the difference between each type of ultrasonic signal and other types of ultrasonic signals to the difference between the ultrasonic signal closest to the corresponding type of ultrasonic signal and all other types of ultrasonic signals, N 1 / 3 (p i ) is the difference between the ultrasonic signal closest to the corresponding type of ultrasonic signal and all other types of ultrasonic signals.

[0059] Optionally, the step of obtaining the reconstructed signal statistical table [D]t2at time t in step 4 includes:

[0060] Constructing a reconstructed signal statistics table [D]t2 at time t based on the reconstructed data;

[0061] The expression of the reconstructed signal statistics table [D]t2 at time t is:

[0062]

[0063] In formula (11), F(p1), F(p2), M and F(p n ) are the reconstructed data.

[0064] The beneficial effects of the present application are:

[0065] 1. The present application can effectively solve the problem of high experience requirement of detection personnel and long time accumulation of detection experience caused by the gate parameter, realize full-automatic processing of ultrasonic a-scan data, and effectively reduce the popularization and use threshold of ultrasonic nondestructive testing technology.

[0066] 2. The present application solves the problem of misjudgment caused by background noise interference of detection results when the ultrasonic detection data is used twice, and the problem of manual setting of gate parameters when ultrasonic detection is performed.

[0067] 3. The present application can realize automatic processing of ultrasonic data, directly obtain defect distribution information after importing original data, and is helpful for integration and development of multi-field technology based on ultrasonic nondestructive testing technology.

[0068] 4. The present application can effectively solve the problem of disordered data structure and excessive background noise of ultrasonic nondestructive testing data, realize regularity of defect information and flattening of background information by reconstructing a-scan data, and is beneficial to secondary use and development of ultrasonic nondestructive testing data. BRIEF DESCRIPTION OF DRAWINGS

[0069] Figure 1 A flow chart of an ultrasonic defect signal automatic detection method based on instantaneous a-scan signal statistics provided by the present application;

[0070] Figure 2 A radiation sound wave sound ray schematic diagram of an ultrasonic array element provided by the present application;

[0071] Figure 3 An ultrasonic phased array detection equipment diagram provided by the present application;

[0072] Figure 4 A detection sample ES1 front face (a) and back face (b) appearance and scanning area schematic diagram provided by the present application;

[0073] Figure 5 A same group of a-scan data different processing effect comparison diagram provided by the present application;

[0074] Figure 6 The image provided by this invention shows the carbon fiber reinforced composite material being tested.

[0075] Figure 7 The image shows the C-scan result of the Dopler phased array, which shows the gate parameters adjusted to the optimal value provided by this invention.

[0076] Figure 8 The present invention provides an automatic detection method for ultrasonic defect signals based on instantaneous a-scan signal statistics. Figure 6 Image of scan results for the a-scan region in the middle scan area. Detailed Implementation

[0077] Combination Figure 1 and Figure 2 This embodiment will be described as follows: Figure 1 As shown, the steps of the automatic detection method for ultrasonic defect signals based on instantaneous a-scan signal statistics described in this embodiment include:

[0078] S1: Acquire all ultrasonic a-scan signals [U]t1 of the material under test at time t;

[0079] S101: Using Rayleigh-Sommerfeld integral form to... Figure 2 The propagation paths of phased array ultrasonic waves in the wedge and specimen, as well as the radiated acoustic ray models of a single array element, are shown to obtain the particle vibration velocities generated in the wedge:

[0080] The expression for the vibration velocity of the particles generated in the wedge is:

[0081]

[0082] In formula (1), K p (θ′) is the sound field directivity function, r is the length of the path from the center of the array element to point x, θ′ is the angle between the path from point x to a point x′ on the array element and the horizontal plane, and k p1 Let p0(ω) be the wave number, p0(ω) be the surface pressure of the array element, κ be the wave velocity ratio, G be the gravitational constant, and c be the wave number. p1 Let d' be the longitudinal wave velocity, i be the number of incident points, S be the surface area of ​​the array element, and d' be the longitudinal wave velocity. p Let x be the unit vector between point x and some point x′ on the matrix element;

[0083] Sound field directivity function, K p The expression for (θ′) is:

[0084]

[0085] In formula (2), κ is the ratio of the longitudinal wave velocity to the transverse wave velocity when the sound wave propagates in a solid.

[0086] S102: Substitute the wave number kp1, the longitudinal wave velocity cp1, the density of the wedge ρ and the wave velocity ratio κ to obtain the vibration velocity of the ultrasonic wave at a certain point in the wedge. Since the actual array element size is small, when the sound wave transmission distance is much larger than the array element size, Figure 2 The distance r' from the calculation point x to a certain point x' on the array element can be approximated as r. The approximated calculation of the integrated function is obtained, and the model of the single array element radiating ultrasonic longitudinal waves into the solid medium with a length lx and a width ly is obtained:

[0087] The expression of the model of the single ultrasonic array element radiating ultrasonic longitudinal waves into the solid medium is:

[0088]

[0089] In formula (3), D l (θ,φ) is the directivity function in the spherical coordinate system;

[0090] The expression of the directivity function in the spherical coordinate system is:

[0091]

[0092] S103: Considering the liquid coupling agent between the wedge and the test piece, assuming that the wedge and the test sample interface are in smooth contact, the ultrasonic array element is divided into m and n blocks in the x direction and the y direction respectively, and based on the high-frequency approximation and the stationary phase method, the three-dimensional acoustic field model of the single array element in the wedge is obtained. Then, the plane wave transmission coefficient T of the ultrasonic wave passing through the liquid coupling layer between the wedge and the test sample is introduced to obtain the three-dimensional acoustic field model of the single array element in the test piece. According to the acoustic field model, the propagation of the ultrasonic wave in the material can be obtained. When there is a defect in the material, the propagation distance of the ultrasonic wave changes, and thus the related information of the first echo can be calculated;

[0093] The expression of the related information of the first echo is:

[0094]

[0095] S104: After the material is excited by the phased array ultrasonic array element in S101, the pressure variation with time at the excited position is obtained, the acoustic pressure is converted into an electrical signal by the piezoelectric sensor, and finally the a-scan data of a point on the surface of the material is formed. By assembling all the a-scan signals according to the physical position, a three-dimensional matrix [U]t of the acoustic pressure signal of the material surface varying with time is finally obtained;

[0096] S2: Statistics are performed on all the ultrasonic signals in the ultrasonic a-scan signal [U]t1 to obtain a signal statistical number table [D]t1 at time t;

[0097] The types of all the ultrasonic signals in the whole ultrasonic a-scan signal [U]t1 of the measured material at time t are counted to establish a signal statistical number table [D]t1 at time t, which includes all types of ultrasonic signals and the number of ultrasonic signals of the same type;

[0098] S3: All the ultrasonic signal types in the signal statistical number table [D]t1 are extracted and analyzed, and the analysis results are taken as reconstruction data into a reconstruction signal statistical table [D]t2 at time t;

[0099] S301: Taking the signal matrix [U]t1 at a certain time in a three-dimensional matrix as an example, first, [U]t1 is counted to obtain the number of each signal appearing in the matrix, and the repeated signal values are folded to reduce the number of subsequent calculation steps:

[0100] The expression for folding the repeated signal values is:

[0101]

[0102] In formula (6), num is the number of the corresponding type of ultrasonic signal, p1, p2, M, p n are different types of ultrasonic signals, p i is the difference between the i-th ultrasonic signal and other types of ultrasonic signals;

[0103] S302: Calculate the difference between each folded ultrasonic signal and other types of ultrasonic signals;

[0104] The expression for the difference between each ultrasonic signal and other types of ultrasonic signals is:

[0105] D=[d1,K,d n ](7);

[0106] d i =[p i -p j |,num i ],j=1,2,K,n.(8);

[0107] In formulas (7) and (8), D is a difference matrix composed of the difference between each ultrasonic signal and other types of ultrasonic signals, d1, d n are the difference sequences between the first and the n-th ultrasonic signals and other types of ultrasonic signals, d i is the difference between the i-th ultrasonic signal and other types of ultrasonic signals, and the calculation method is the difference between the i-th ultrasonic signal p i and the j-th ultrasonic signal p j ;

[0108] S303: taking the ratio of the difference between each type of ultrasonic signal and other types of ultrasonic signals and the difference between the 1 / 3 of ultrasonic signals closest to the corresponding type of ultrasonic signal and all other types of ultrasonic signals as reconstruction data, and constructing a reconstruction signal statistical table [D]t2 at time t;

[0109] The calculation formula of the reconstruction data is:

[0110]

[0111]

[0112] In formulas (9) and (10), RD(p i ) is the difference between the 1 / 3 of ultrasonic signals closest to the corresponding type of ultrasonic signal and all other types of ultrasonic signals, F(p i ) is the ratio of the difference between each type of ultrasonic signal and other types of ultrasonic signals and the difference between the 1 / 3 of ultrasonic signals closest to the corresponding type of ultrasonic signal and all other types of ultrasonic signals, N 1 / 3 (p i ) is the difference between the 1 / 3 of ultrasonic signals closest to the corresponding type of ultrasonic signal and all other types of ultrasonic signals.

[0113] The expression of the reconstruction signal statistical table [D]t2 at time t is:

[0114]

[0115] In formula (11), F(p1), F(p2), M and F(p n ) are reconstruction data.

[0116] S4: reconstructing all ultrasonic a-scan signals [Ut1] at time t to [Ut2] according to the mapping relationship between the physical positions of the signal statistical table [Dt1] at time t and the reconstruction signal statistical table [Dt2], and completing the extraction of the defect information in the ultrasonic a-scan signals;

[0117] S5: repeating S1-S4 to reconstruct and superimpose all ultrasonic a-scan signals at all times to obtain the reconstructed ultrasonic detection result.

[0118] Embodiment

[0119] Embodiment 1

[0120] In combination Figures 3-5 This embodiment can first reduce the intervention of personnel operation in the detection process, realize the automation of the detection process, and reduce the requirement for the nondestructive testing experience of the operator. This embodiment uses, for example, the ultrasonic a-scan signal as the input signal, and the ultrasonic a-scan signal at time t is taken as an example for description. Figure 3The shown multi-pulse ROBUST 32-128(64)_TOFD ultrasonic phased array board card system scans the LLA482 phased array probe;

[0121] The shown phenolic resin-based glass fiber reinforced composite sample ES1 is scanned, and the scanning area is Figure 4 Figure 4 (a) The internal yellow area. The scanning results are respectively set to the best effect, no gate is set, and the a-scan data is processed by the algorithm proposed in the patent to obtain the results as shown in Figure 5

[0122] As can be seen from Figure 5 , the algorithm proposed in this embodiment effectively avoids the intervention of manual gate parameter setting in the detection process, and obtains better defect distribution results than the gate c-scan reconstruction.

[0123] Example 2

[0124] This embodiment is described in combination with Figure 3 and Figures 6-8 , this embodiment uses the multi-pulse ROBUST 32-128(64)_TOFD ultrasonic phased array board card system and the LLA482 phased array probe as shown in Figure 3 to scan the carbon fiber reinforced composite ES2 marked area with a prefabricated flat bottom hole as shown in Figure 6 To ensure the detection effect, the sample ES2 is immersed in water for detection, and the gate parameter is adjusted to the best effect to obtain the two-dimensional and three-dimensional effect diagrams of the detection results as shown in Figure 7 The two-dimensional and three-dimensional effect diagrams of the a-scan data obtained using the method proposed in the patent are as shown in Figure 8 .

[0125] As shown in Figure 7 , the c-scan effect obtained after improving the coupling environment is greatly improved compared with Figure 5 in Example 1. From the two-dimensional graph, the background noise is greatly relieved, but from the three-dimensional effect diagram, it is found that the flatness of the background signal in the detection result is poor, and the defect area edge and the center signal are distributed on both sides of the background signal. This signal format is only suitable for manual identification. Since the defect information is distributed in disorder, the neural network, machine vision and other automatic identification technologies cannot accurately extract the defect information from the background signal, which greatly challenges the secondary use of ultrasonic detection results.

[0126] As shown in Figure 8 , the ultrasonic defect automatic detection method proposed in this embodiment can achieve a higher smoothness of the background signal without manual intervention, and the defect information is on one side of the background signal, which is very convenient for subsequent data fusion or artificial intelligence reuse.​​

[0127] In summary, after the verification of example 1 and example 2, the present application can obtain the following technical effects:

[0128] 1. The present method can effectively solve the problem that the detection personnel requires high experience and needs a long time to accumulate detection experience caused by the gate parameters, and effectively reduces the popularization and use threshold of ultrasonic nondestructive testing technology;

[0129] 2. The present method can realize the automatic processing of ultrasonic data, and directly obtain the defect distribution information after importing the original data, which is helpful for the integration and development of multi-field technology based on ultrasonic nondestructive testing technology;

[0130] 3. The present method can effectively solve the problem of disordered data structure and excessive background noise of ultrasonic nondestructive testing data, and realize the regularization of defect information and the flattening of background information by reconstructing the a-scan data, which is conducive to the secondary use and development of ultrasonic nondestructive testing data.

[0131] The above is only a preferred embodiment of the present application, and does not limit the present application in any form. Although the present application has been disclosed as above with a preferred embodiment, it is not intended to limit the present application. Any skilled person in the art can make some changes or modifications to the above disclosed technical content without departing from the scope of the technical solution of the present application, and make equivalent embodiments with equivalent changes. Any simple modification, equivalent replacement and improvement of the above embodiments within the scope of the technical solution of the present application, according to the technical essence of the present application, within the spirit and principles of the present application, are still within the protection scope of the present application.

Claims

1. A method for automatic detection of ultrasonic defect signals based on instantaneous a-scan signal statistics, characterized in that, The steps of the ultrasonic defect signal automatic detection method based on instantaneous a-scan signal statistics include: Step 1: obtaining all ultrasonic a-scan signals [U]t1 of the measured material at t time; Step 2: performing statistics on all ultrasonic signals in the ultrasonic a-scan signal [U]t1 to obtain a signal statistics table [D]t1 at t time; Step 3: extracting all ultrasonic signal types in the signal statistics table [D]t1 and performing analysis; Step 4: taking the analysis result as reconstruction data into a reconstructed signal statistics table [D]t2 at t time; Step 5: reconstructing all ultrasonic a-scan signals [U]t1 at t time into [U]t2 according to the mapping relationship between the physical positions of the signal statistics table [D]t1 and the reconstructed signal statistics table [D]t2, and completing extraction of defect information in the ultrasonic a-scan signal; Step 6: repeating steps 1-5 to reconstruct and superimpose all-time ultrasonic a-scan signals to obtain a reconstructed ultrasonic detection result.

2. The method of claim 1, wherein, The step of obtaining all ultrasonic a-scan signals [U]t1 of the measured material at t time in step 1 includes: Step 1.1: exciting the material using a phased array ultrasonic element; Step 1.2: obtaining the change of pressure intensity of the material at the excited position with time, converting the sound pressure into a voltage signal through a piezoelectric sensor to form a-scan data of a point on the surface of the material; Step 1.3: repeating steps 1.1-1.2 to obtain all a-scan data of the material surface, assembling all a-scan data of the material surface according to the physical position, and obtaining a three-dimensional matrix [U]t of the material surface sound pressure signal changing with time; Step 1.4: obtaining all ultrasonic a-scan signals [U]t1 of the measured material at t time based on the three-dimensional matrix [U]t of the material surface sound pressure signal changing with time.

3. The method of claim 2, wherein, The step of exciting the material using a phased array ultrasonic element in step 1.1 includes: Step 1.1.1: performing ultrasonic a-scan on the measured material to obtain the propagation route of the phased array ultrasonic wave in the wedge and the measured material, and obtaining the radiation sound wave ray of a single ultrasonic element based on the propagation route of the phased array ultrasonic wave in the wedge and the measured material; Step 1.1.2: modeling the radiation sound wave ray of the single ultrasonic element to obtain the particle vibration velocity generated in the wedge; Step 1.1.3: approximately calculating the integral function in the particle vibration velocity to obtain a model of the ultrasonic longitudinal wave radiated by the single ultrasonic element into the solid medium with a length of lx and a width of ly; Step 1.1.4: dividing the ultrasonic element into m blocks and n blocks in the x direction and the y direction respectively, obtaining a three-dimensional sound field model of the single element in the wedge based on the high-frequency approximation and the stationary phase method, and combining the model of the ultrasonic longitudinal wave radiated by the single ultrasonic element into the solid medium; Step 1.1.5: introducing a plane wave transmission coefficient T of the liquid coupling layer between the ultrasonic wave passing through the wedge and the measured material into the three-dimensional sound field model of the single element in the wedge to obtain a three-dimensional sound field model of the single element in the measured material; Step 1.1.6: Obtain the propagation of ultrasonic wave in the material under test according to the three-dimensional sound field model of individual array element in the material under test. When the material has defects, the propagation distance of ultrasonic wave changes, and the related information of the first echo is calculated to complete the excitation of the material.

4. The method of claim 1, wherein, The step of obtaining the signal statistics table [D]t1 in step 2 includes: Extract all types of ultrasonic signals in the all ultrasonic A-scan signals [Ut1] of the material under test at time t, and count the number of ultrasonic signals of the same type to establish the signal statistics table [Dt1] at time t, which includes all types of ultrasonic signals and the number of ultrasonic signals of the same type.

5. The method of claim 1, wherein, The step of extracting and analyzing all types of ultrasonic signals in the signal statistics table [D]t1 in step 3 includes: Step 3.1: Extract all types of ultrasonic signals in the signal statistics table [Dt1] at time t, and fold the repeated signal values; Step 3.2: Calculate the difference between each type of ultrasonic signal and other types of ultrasonic signals after folding; Step 3.3: Take the ratio of the difference between each type of ultrasonic signal and other types of ultrasonic signals and the average difference between the nearest 1 / 3 ultrasonic signals of the corresponding type and all types of ultrasonic signals as the reconstruction data, and complete the analysis of all types of signals in the signal statistics table [Dt1]; The expression for folding repeated signal values is: In formula (1), num is the number of the corresponding type of ultrasonic signals, p1, p2, p n are the first, second, and nth ultrasonic signals, pi is the i-th ultrasonic signal. i are the first, second, and nth ultrasonic signals, pi is the i-th ultrasonic signal. The expression for the difference between each type of ultrasonic signal and other types of ultrasonic signals is: D = [d1, K, d n ](2); d i = [p i -p j |, num i ], j = 1, 2, K, n. (3); In the formulas (2) and (3), D is a difference matrix composed of the differences of each type of ultrasound signal and other types of ultrasound signals, di, d n are the difference sequences of the first and nth ultrasound signals and other types of ultrasound signals, di i is the difference of the ith ultrasound signal and other types of ultrasound signals, and the calculation method is the difference between the ith ultrasound signal p i and the jth ultrasound signal p j . The calculation formula of the reconstruction data is: In the formulas (4) and (5), RD(p i ) is the difference mean value of the ultrasound signals closest to 1 / 3 of the corresponding type ultrasound signals and all other types of ultrasound signals, F(p i ) is the ratio of the difference value of each type of ultrasound signal and other types of ultrasound signals to the difference mean value of the ultrasound signals closest to 1 / 3 of the corresponding type ultrasound signals and all other types of ultrasound signals, N 1 / 3 (p i ) is the difference value of the ultrasound signals closest to 1 / 3 of the corresponding type ultrasound signals and all other types of ultrasound signals.

6. The method of claim 1, wherein, The step of obtaining the reconstructed signal statistics table [Dt2] at time t in step 4 includes: Construct the reconstructed signal statistics table [Dt2] at time t based on the reconstruction data; The expression for the reconstructed signal statistics table [Dt2] at time t is: In Equation (6), F(p1), F(p2), and F(p n ) are the reconstructed data.

Images