Eddy current defect detection method based on multi-direction scanning, electronic equipment and readable storage medium

Abstract

The invention discloses an eddy current defect detection method based on multi-direction scanning, electronic equipment and a readable storage medium. The basic idea of working is that excitation current with fixed frequency is introduced into the coil so as to generate an alternating excitation magnetic field in the space. The magnetic field sensor is used for measuring the magnetic flux density change caused by the defect at a proper position, so as to sense the defect. By means of the characteristics that the sizes of hole defects in different directions are similar and the length-width ratio of crack defects is large, a proper probe is adopted to conduct continuous motion scanning in multiple directions, and defect classification is achieved by obtaining more defect information. According to the method, after multi-direction scanning based on continuous movement is adopted, information of defects in different directions can be obtained on the premise that the detection efficiency is not seriously influenced, and classification of holes and cracks can be smoothly realized by adopting a proper signal processing method.

Description

Technical Field

[0001] The present invention mainly relates to the field of industrial detection, and in particular to an eddy current defect detection method based on multi-directional scanning, an electronic device and a readable storage medium. Background Art

[0002] Eddy current defect detection technology has many advantages. First, it is a non-contact detection technology that avoids the contact process that may cause damage. Secondly, eddy current detection is not affected by external environmental factors such as oil pollution and lighting conditions, and can be used under relatively harsh working conditions. In addition, eddy current detection responds quickly and can achieve rapid detection of large areas or large batches of workpieces. At the same time, eddy current technology is highly adaptable and can adapt to various workpiece shapes and surface conditions. Finally, it can not only detect the existence of defects, but also quantitatively evaluate defects to a certain extent with the help of the measured signals, such as the length, width, depth and other information of the crack. Therefore, it is widely used in quality control and safety detection in the industrial field. However, the existing eddy current detection technology often takes the known defect type as a prerequisite when quantitatively evaluating defects, which is inconsistent with the actual application scenario. Therefore, distinguishing the defect type helps to accurately select the appropriate defect quantitative evaluation algorithm in the future, and take corresponding solutions based on the results, which is of great significance for improving production efficiency and product quality. At the same time, it also helps to improve product design and manufacturing processes and reduce the occurrence of future defects.

[0003] Through the existing document search, it is found that there are mainly the following solutions for eddy current defect detection:

[0004] The Chinese patent application number is CN201611125184X, and its name is "A composite processing method for high-energy beam additive and subtractive online eddy current detection of metal materials". This invention provides a method for processing metal materials. First, a high-energy beam is used to melt the metal layer by layer to form a workpiece. Then, the formed material is trimmed to make the surface smoother. Next, defects are detected on the workpiece through eddy current detection. If there are defects, they are removed by trimming and re-additive until there are no defects. If there are no defects, the workpiece is directly completed. This method solves the problem of the inability to detect and repair pure additive manufacturing parts online, and improves product quality and production efficiency.

[0005] The Chinese patent application number is CN202310210215.5, and its name is a GMM-based eddy current detection probe and detection method. The invention discloses a GMM-based eddy current detection probe and detection method, which belongs to the field of eddy current non-destructive testing technology. The invention uses the principle that the magnetostrictive effect of the GMM changes with the change of the magnetic field on the GMM, and detects the defects of the test piece by detecting the magnetostrictive strain of the GMM rod. The GMM sensing unit replaces the traditional receiving coil as the detection probe. The invention can detect tiny defects and has the advantages of high detection accuracy and sensitivity.

[0006] The Chinese patent application number is CN202410393625.2, and its name is a model training method, a defect detection method, and related devices. The application provides a model training method, a defect detection method, and related devices, which are suitable for various scenarios, such as computer vision. The model training method uses a partially trained first sampling image set and a fully newly added second training image set to train the initial detection model. Compared with using all historical training data, such a training method saves time and reduces the consumption of GPU cards. By adaptively evaluating the corresponding weights of network parameters, the update of network parameters by historical training data is limited, and the problem of knowledge forgetting that may be caused by fine-tuning using only the full amount of newly added data is solved. This improves the learning ability of the detection model. The optimized detection model obtained by using the model training method can more accurately detect defects in the image to be detected, thereby improving the effect and accuracy of defect recognition.

[0007] According to the search results, the content of eddy current detection technology in existing patents is mostly focused on improving detection accuracy and sensitivity, and there are fewer patents based on eddy current detection to distinguish different types of defects. As mentioned above, in actual production, simply detecting the existence of defects cannot meet the requirements of product quality control. Understanding the type and nature of defects is crucial for the next step in the production process. For example, for some surface defects, simple repair work may be required, while for more serious internal defects, more thorough treatment measures may be required, such as remanufacturing or replacing parts. In addition, different types of defects may originate from different production links or reasons. Timely identification and classification of defects can help find problem points in production and make improvements to reduce future defects. Therefore, classifying defects can help manufacturers manage quality more effectively, improve product consistency and reliability, and thus enhance overall competitiveness. Summary of the invention

[0008] In view of the above-mentioned defects of the prior art, the technical problems to be solved by the present invention include:

[0009] How to design an eddy current defect detection method to overcome the above technical problems.

[0010] To achieve the above object, the present invention provides an eddy current defect detection method based on multi-directional scanning, comprising the steps of:

[0011] Step 1: Use eddy current probes to measure the Scan the workpiece from multiple directions to obtain magnetic flux density scanning images with different directions, the number of which is N r ;

[0012] Step 2: N obtained in different directions r The scanned images are normalized relative to the maximum pixel value within each image;

[0013] Step 3: Use threshold S t Binarize all normalized images to obtain one or more closed figures in each image;

[0014] Step 4: Calculate the N surrounding the estimated defect center in each normalized image. p The average value of pixels

[0015]

[0016] Step 5: For a series of 2D scan images of the same defect obtained in different directions, With S f Compare and determine the defect type.

[0017] Furthermore, in step 1, the number Nr of the magnetic flux density scanning images is obtained by the following method:

[0018]

[0019] Further, in step 1, after each scanning is completed, the eddy current probe and the preset scanning path need to be Rotate relative to the workpiece at angular intervals.

[0020] Furthermore, in step 1, there is no special requirement for the scanning path, as long as it can achieve complete two-dimensional scanning and form a two-dimensional scanning image according to the position of the eddy current probe and the measured signal.

[0021] Furthermore, during the scanning process, the posture of the eddy current probe should be kept constant.

[0022] Furthermore, the step 3 also includes considering the centroid position of all closed figures in each image as the defect center position.

[0023] Furthermore, the step 3 also includes averaging the centroid positions of all closed figures in each image to obtain a true defect center.

[0024] Further, in step 5, the With S f By comparison, the defect type is determined as follows: if the calculated values ​​of all images are relatively small, that is, S f is a predetermined threshold, the defect can be regarded as a hole; however, if the calculated value of any scanned image exceeds the predetermined threshold S f ,Right now Then the defect will be considered as a crack.

[0025] An electronic device comprising

[0026] Memory,

[0027] processor,

[0028] Eddy current probes,

[0029] Mobile platforms, and

[0030] A computer program stored in the memory and executable on the processor, when the processor executes the computer program, controls the eddy current probe and the mobile platform to implement the eddy current defect detection method based on multi-directional scanning as described above.

[0031] A computer-readable storage medium stores a computer program, which, when executed by a processor, controls an eddy current probe and a mobile platform to implement the eddy current defect detection method based on multi-directional scanning as described above.

[0032] Compared with the prior art solutions, the technical effects of the present invention are:

[0033] The present invention takes advantage of the fact that hole defects have similar sizes in different directions while crack defects have a larger aspect ratio, and proposes to use a suitable probe to perform continuous motion scanning from multiple directions to achieve defect classification by obtaining more defect information.

[0034] After the present invention adopts multi-directional scanning based on continuous motion, the information of defects in different directions can be obtained without seriously affecting the detection efficiency, and the classification of holes and cracks can be smoothly realized by adopting a suitable signal processing method.

[0035] The concept, specific structure and technical effects of the present invention will be further described below in conjunction with the accompanying drawings to fully understand the purpose, characteristics and effects of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1It is a schematic diagram of an eddy current detection system for multi-directional scanning to distinguish holes and crack defects of the present invention.

[0037] Figure 2 It is a simulation result diagram of the present invention for determining a characteristic parameter threshold to distinguish between hole and crack defects.

[0038] Figure 3 The present invention discloses an experimental device and a test piece photo for multi-directional scanning to distinguish holes and crack defects.

[0039] Figure 4 It is a schematic diagram of a detection probe used for multi-directional scanning to distinguish holes and crack defects of the present invention.

[0040] Figure 5 The invention provides an example of a normalized scanning image of a hole and crack detection result obtained by using a probe for multi-directional scanning to distinguish hole and crack defects.

[0041] Figure 6 The present invention is a binary scanning image example of a hole and crack detection result of a probe used for multi-directional scanning to distinguish hole and crack defects.

[0042] Figure 7 It is an actual characteristic parameter distribution result used for multi-directional scanning to distinguish holes and crack defects in the present invention.

[0043] Reference numerals

[0044] 1. Eddy current probe; 2. Test piece; 3. Defect; 4. Simulation solution domain; 5. Single direction scanning path; 6. Rectangular excitation coil; 7. Magnetic field sensor; DETAILED DESCRIPTION

[0045] The following describes several preferred embodiments of the present invention with reference to the drawings in the specification, so that the technical content is clearer and easier to understand. The present invention can be embodied in many different forms of embodiments, and the protection scope of the present invention is not limited to the embodiments mentioned in the text.

[0046] The detection target of the present invention is to detect and classify hole and crack defects. Figure 1 The schematic diagram of the multi-directional eddy current detection system is shown. First, three coordinate systems are established: XYZ, xyz and Coordinate system. The XYZ coordinate system is defined on the eddy current probe, and the xyz coordinate system is defined on the workpiece, where the X and Y axes are parallel to the x and y axes respectively. It is a defect coordinate system that is additionally defined to describe defects. Its origin coincides with the defect center on the workpiece surface. The axes are perpendicular to the upper surface of the workpiece and point outward. The orientation of the defect is based on the X-axis and Axis angle To characterize.

[0047] In order to successfully identify and distinguish crack defects and hole defects in metal materials, the present invention is based on the following principles:

[0048] (1) For the non-conductive area, the eddy current field in the defect area is always zero, which indicates that the disturbed eddy current field in the local area is equal in magnitude to the original eddy current field induced by the probe, but opposite in direction.

[0049] (2) For holes and cracks, the eddy current field j at the defect boundary satisfies the constraint: j(r)·n(r)=0, where n(r) represents the unit external normal vector at the defect edge position r.

[0050] (3) The disturbed eddy current field also satisfies the law of conservation of charge and appears in the workpiece in the form of a closed loop without any starting point or end point.

[0051] Based on the above principles, it can be deduced that at different relative positions of the defect and the probe Comparing the disturbed eddy current fields of conductors with hole defects and cracks, it is found that the eddy current field near the crack defect will change with the The eddy current field rotates relative to the X-axis due to the change of the eddy current field, but the rotation of the eddy current field near the hole defect is not obvious. Therefore, holes and cracks can be detected and classified by multi-directional scanning.

[0052] Based on the above principles, the basic ideas of the present invention are as follows:

[0053] (1) Use eddy current probes at the same angle intervals From multiple directions (quantity Scan the workpiece to obtain magnetic flux density scanning images with different defect directions, the number of which is also N r To achieve the above multi-directional scanning, after each scan, the eddy current probe and the preset scanning path need to be The eddy current probe rotates relative to the workpiece at an angle interval, and during a single scan, the eddy current probe posture should remain constant. The present invention has no special requirements for the scanning path, and can achieve complete two-dimensional scanning and form a two-dimensional scanning image based on the position of the probe and the measured signal. When performing multi-directional scanning, the present invention does not make any requirements for the initial scanning angle, that is, the initial scanning angle will not affect the subsequent processing process and the final classification result of the present invention.

[0054] (2) For each defect, the N obtained in different directions rThe scanned images are normalized relative to the maximum pixel value in each image. Then, the threshold S is used t All normalized images are binarized to obtain one or more closed figures in each image. Then, the centroid position of all closed figures in each image is found and regarded as the defect center position. For the same defect, the defect center position estimated in different scanning directions may be slightly different, but it does not affect the subsequent classification process. The true defect center can be further obtained by averaging the defect center positions estimated in all images to further reduce the error.

[0055] (3) For each defect, calculate the N surrounding the estimated defect center in each normalized image. p The average value of pixels For a series of 2D scanning images of the same defect obtained in different directions, if the calculated values ​​of all images are relatively small, that is, S f is a predetermined threshold, the defect can be regarded as a hole; however, if the calculated value of any scanned image exceeds the predetermined threshold S f ,Right now In this step, the purpose of calculating the average value of multiple pixels is to eliminate the influence of signal noise on the classification results. p The specific value of is related to the defect size, noise size and the probe space distance between adjacent pixels on the scan pattern. Generally speaking, N p Not too big.

[0056] In order to cooperate with the above defect detection strategy and determine specific parameters, especially the classification threshold sf, the present invention studies the system symmetry to more efficiently predict the scanning image. Since the eddy current probe has system symmetry and the workpiece is assumed to be an infinitely large flat plate, the eddy current induced by it and the magnetic flux density field it generates show a certain symmetry. This symmetry can be used to avoid some redundant calculations when predicting 2D scanning images. According to the Biot-Savart law, the magnetic flux density generated by the eddy current at the sensing point can be calculated by formula (1):

[0057]

[0058] Among them, r s and r represent the coordinates of the sensing point and the source point in the xyz coordinate system, r O represents the coordinate of point O in the xyz coordinate system. Now consider three cases, namely: case 1, case 2, and case 3, where there is a defect in the workpiece in each case. Assuming that the geometric shape of the defect and the position of the probe in case 1 and case 2 are symmetrical about the xz plane, the eddy current field satisfies the following in the xyz coordinate system:

[0059]

[0060] Among them, the superscripts (1) and (2) are used to represent the physical quantities in case 1 and case 2, respectively. In addition, since the sensing point in the eddy current probe is located on the Z axis, its coordinates in the xyz coordinate system can be expressed as follows in both cases:

[0061]

[0062] From formula (2) and (3), we can get:

[0063]

[0064] Therefore, the relationship between the magnetic flux density generated by the eddy current field in the two cases can be expressed as equation (5). The scanning image of case 1 is inverted and is symmetric to the scanning image of case 2 about the x-axis.

[0065]

[0066] Assuming that the geometric shapes of the defects and the positions of the eddy current probes in Case 1 and Case 3 are symmetrical about the yz plane, similar derivation can be used to obtain Equation (6) for the case where the symmetry is about the yz plane. The scanned image of Case 1 is inverted and is symmetrical to the scanned image of Case 3 about the y axis.

[0067]

[0068] Then the geometric shapes of the defects and the positions of the eddy current probes in Case 2 and Case 3 are symmetrical about the origin o of the xyz coordinate system, and equation (7) can be derived.

[0069]

[0070] Taking advantage of the symmetry of the system, for general defects, the scanned image only needs to be calculated when the direction of the defect is in the range [0,π]. The scanned images of defects in other ranges can be obtained by scanning the scanned images of defects that are symmetrical to the defect in the range [0,π] without additional calculation. In addition, for axisymmetric defects, the direction range that needs to be calculated is further reduced to This significantly reduces the number of simulations required.

[0071] With the help of the distributed current source method and the system symmetry, a large number of scanned images of different defects in different directions were simulated, and the obtained results were processed according to the processing flow proposed in the present invention. All calculations were completed using MATLAB on a PC equipped with a Xeon 6248R, 3GHz CPU, 383GB RAM and 64-bit operating system. The characteristic parameters obtained from the simulated detection data of all different types of defects The occurrence ratio in different intervals is Figure 2 The results show that for hole defects, although the characteristic parameters have a large distribution range, they are always less than 0.82. On the contrary, for crack defects, the characteristic parameters are concentrated and always in the range of 0.98 to 1. For all defects, the characteristic parameters do not appear in the range of 0.82 to 0.94, so they are regarded as transition areas. Figure 2 The shaded part in the middle represents the area. Selecting an appropriate threshold in this range can effectively distinguish between holes and cracks. Considering that the crack width is ignored in the numerical simulation, the distribution range of the characteristic parameters of such defects in actual applications may be slightly larger than the simulation results. Therefore, the threshold is selected as sf = 0.85, making it closer to the distribution range corresponding to the hole. It is worth noting that the threshold selection for defects in different size ranges may be different, and should be determined after simulation according to the specific situation according to the method proposed in the present invention.

[0072] A specific embodiment of the present invention is further described in detail below with reference to the accompanying drawings.

[0073] in accordance with Figure 1 The test platform is built as shown in the schematic diagram of the detection system structure. Figure 3 As shown. In this embodiment, a two-dimensional linear moving platform is selected as the motion device. The moving platform is driven by a linear motor, and the two moving axes are perpendicular to each other. The probe and the test piece are respectively fixed to axis 1 and axis 2 by a fixture made of POM material. In this embodiment, the test piece is a square conductive plate (WP#1, WP#2, WP#3, WP#4) made of aluminum, and the detection target is the hole defect and crack defect located in the test piece. Among them, WP#2 and WP#4 are tested from both sides to verify the ability of the present invention to deal with sub-surface defects. During the detection, the eddy current probe forms a two-dimensional scanning image through multiple one-dimensional scans. During the scanning process, the bottom surface of the coil of the probe should be kept parallel to the upper surface of the test piece. In this embodiment, the scanning direction interval Therefore, for each defect, N can be obtained at each initial scanning angle. r = 3 images. In addition, the binarization threshold S in the present invention t Choose 0.5, the number of pixels N to calculate the average value p Select 9.

[0074] Figure 4 Schematic diagram of the probe structure used in this embodiment. According to the defect size range to be detected in this embodiment, the specific parameters are selected as follows: (w X , w Y ,h c , w i , d c , l c)=(30, 14, 5, 8, 16, 0.75) mm, and the number of coil turns is 175. The magnetic field sensor used in this example is TMR2103 (Multidimensional Technology) linear magnetic field sensor. The numerical simulation is completed by programming in MATLAB based on the distributed current source (DCS) method. The X-axis (long side of the rectangular coil) direction is selected as the probe signal acquisition direction. This example uses NI-9264 as a signal generator and NI-9252 as a signal acquisition device. The two are connected to the NI CompactRIO 9040 controller and controlled by the host computer. The sinusoidal excitation signal generated by NI-9264 is input into the power amplifier module designed using OPA549 to achieve current amplification and finally used to drive the coil. The excitation current frequency is selected as 1kHz based on experience. The response of the magnetic field sensor and the coil current are sampled by NI-9252 at a sampling frequency of 50kHz. During the scanning process, axis 2 drives the specimen to reciprocate at a speed of 100 mm / s. When the specimen reaches the limit position, axis 1 drives the probe to move 0.5 mm toward the outside of the edge.

[0075] Figure 5 is Disturbed magnetic flux density images of two defects, holes and cracks, under three conditions. When the defects are scanned from multiple directions, for hole defects, the signal amplitude is always relatively small in the central area of ​​all scanned images, while for cracks, the signal amplitude reaches the maximum in the central area of ​​some scanned images, which proves the rationality of the defect detection strategy proposed by the analysis of the present invention for distinguishing hole and crack defects.

[0076] Figure 6 yes Figure 5 The binarized scanned images of the two examples of holes and cracks are shown. According to the description of step (2), a threshold of 0.5 is selected to binarize the image, thereby estimating the position of the defect center (the cross position in the figure). The results show that the error between the estimated defect center position and the actual position is small. Based on this result, the characteristic parameters corresponding to different defects can be calculated.

[0077] Figure 7 The characteristic parameter distribution of different types of defects obtained in the experiment, including two sub-surface defects, can be seen from the results that the hole and crack defects can be correctly distinguished according to the characteristic parameters with the help of the selected threshold, which proves the ability of the eddy current defect detection and classification method based on multi-directional scanning proposed in the present invention.

[0078] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0079] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0080] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0081] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An eddy current defect detection method based on multi-directional scanning, characterized in that: Includes steps: Step 1: Use eddy current probes to measure the Scan the workpiece from multiple directions to obtain magnetic flux density scanning images with different directions, the number of which is N r ; Step 2: N obtained in different directions r The scanned images are normalized relative to the maximum pixel value within each image; Step 3: Use threshold S t Binarize all normalized images to obtain one or more closed figures in each image; Step 4: Calculate the N surrounding the estimated defect center in each normalized image. p The average value of pixels Step 5: For a series of 2D scan images of the same defect obtained in different directions, With S f Compare and determine the defect type.

2. The eddy current defect detection method based on multi-directional scanning according to claim 1 is characterized in that: In step 1, the number Nr of the magnetic flux density scanning images is obtained by the following method:

3. The eddy current defect detection method based on multi-directional scanning according to claim 2 is characterized in that: In step 1, after each scan is completed, the eddy current probe and the preset scanning path need to be Rotate relative to the workpiece at angular intervals.

4. The eddy current defect detection method based on multi-directional scanning according to claim 1 is characterized in that: In step 1, there is no special requirement for the scanning path, as long as it can achieve complete two-dimensional scanning and form a two-dimensional scanning image according to the position of the eddy current probe and the measured signal.

5. The eddy current defect detection method based on multi-directional scanning according to claim 3 is characterized in that: During the scanning process, the posture of the eddy current probe should be kept constant.

6. The eddy current defect detection method based on multi-directional scanning according to claim 1 is characterized in that: The step 3 also includes considering the centroid position of all closed figures in each image as the defect center position.

7. The eddy current defect detection method based on multi-directional scanning according to claim 6 is characterized in that: The step 3 also includes averaging the centroid positions of all closed figures in each image to obtain a true defect center.

8. The eddy current defect detection method based on multi-directional scanning according to claim 1, characterized in that: In step 5, the With S f By comparison, the defect type is determined as follows: if the calculated values ​​of all images are relatively small, that is, S f is a predetermined threshold, the defect can be regarded as a hole; however, if the calculated value of any scanned image exceeds the predetermined threshold S f ,Right now Then the defect will be considered as a crack.

9. An electronic device, characterized in that: include Memory, processor, Eddy current probes, Mobile platforms, and A computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the eddy current probe and the mobile platform are controlled to implement the eddy current defect detection method based on multi-directional scanning as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by the processor, the eddy current probe and the mobile platform are controlled to implement the eddy current defect detection method based on multi-directional scanning as described in any one of claims 1 to 8.

Images