A defect contour recognition method based on uniform parallel induced current flow characteristics

Abstract

The present invention discloses a defect contour recognition method based on the flow characteristics of uniform parallel induced current. First, characteristic points are constructed, and the defects in the scanning direction of the metal surface are accurately recognized by analyzing the magnetic field B (z) curve group of multi-path scanning; then, the interference of the external environment is eliminated by screening the characteristic point data, so as to further improve the accuracy of defect contour recognition; the problem of mutual squeezing of currents is reduced by processing the vertical long straight contours; the problem that the recognized characteristic points deviate from the actual position due to the squeezing effect of current is solved by compensating for edge defects; the accuracy of the defect contour is improved by respectively correcting the contours in the two scanning directions, and finally the defect contour is recognized, and the defect contour information of the metal surface is quickly and accurately obtained by the flow characteristics of uniform parallel induced current, so as to realize the intuitive visualization of the defect contour and improve the application degree and characterization ability of traditional electromagnetic detection.

Description

Technical Field

[0001] The invention relates to the technical field of material surface quality and performance characterization, and in particular to a defect contour recognition method based on uniform parallel induced current flow characteristics. Background Art

[0002] With the development of modern industry, material surface defect detection technology has become an indispensable part of the industrial field. Material preparation, heat treatment and processing technology can lead to original defects in metal structures. If the product is severely damaged, unpredictable dangers will occur. Therefore, non-destructive testing can detect the existence of defects. If it can further identify the specific information of the defects, so as to analyze the impact of defects on product quality by analyzing the defect contour, it can predict the product life and optimize the product production process. At present, many industrial productions need to obtain the specific contours of workpiece surface defects to achieve comprehensive analysis of product safety, thereby optimizing production processes and equipment. This puts higher requirements on the original detection technology. Through in-depth research on the principles of non-destructive testing and technological innovation, the defect contour can be accurately identified and analyzed.

[0003] In the existing related technologies, the metal surface defect contour recognition methods based on electromagnetic methods mainly include electromagnetic imaging, magnetic flux leakage detection, and magneto-optical imaging.

[0004] A Chinese patent application numbered 2017100788954 discloses a method and device for adjusting the structure of a magnetoacoustic array for guided wave imaging of defect contours of metal plates. The method and device evenly arrange N omnidirectional magnetoacoustic transducers in the detection area of ​​the metal plate to be detected, and perform cubic smoothing spline interpolation on the scattering point position data according to the received defect guided wave scattering signal to form a defect contour curve. However, magnetoacoustic transducer guided wave imaging is mainly suitable for magnetic materials, non-magnetic materials do not produce obvious magnetoacoustic effects, and is not suitable for all types of defects. In practical applications, electromagnetic interference, environmental noise and interference from other sources may affect the clarity and accuracy of imaging. In addition, magnetoacoustic transducer guided wave imaging requires complex signal processing technology to analyze the received signals and extract signals about defects from them, which requires professional equipment and technology.

[0005] The Chinese patent application number 201510239389X discloses a method and device for reconstructing defect contours in three-dimensional magnetic flux leakage detection, which provides a defect contour inversion method based on a closed-loop iterative method of magnetic flux leakage signals. By preparing a defect magnetic flux leakage signal database in advance, the data in the defect magnetic flux leakage signal database is subjected to simple transformation operations such as indexing, interpolation, shifting, and combining to obtain the corresponding predicted defect magnetic flux leakage signal. However, the current magnetic flux leakage detection is only magnetized in a single direction, and the magnetic flux leakage distortion of complex defects is serious. The defect edge shows complex diffusion and distortion, resulting in inaccurate reconstruction of the gap contour.

[0006] Chinese patent application number 2023103304987 discloses a magneto-optical imaging surface defect contour reconstruction method based on multi-directional detection, which provides a magneto-optical imaging surface defect contour reconstruction method based on multi-directional detection, by applying the iterative inversion method to the shape reconstruction of multi-directional magnetization leakage. However, since magneto-optical imaging technology relies on the mutual conversion of magnetic field and light, the interference of external electromagnetic fields will affect the accuracy of the imaging results. In addition, magneto-optical imaging generally requires more complex equipment, including magneto-optical instruments, light sources, optical path control systems, etc. The complexity of the equipment will increase the error of the system, and professional technicians are required to operate and maintain it. Summary of the invention

[0007] In order to solve the above technical problems, the present invention provides a defect contour recognition method based on the uniform parallel induced current flow characteristics, comprising the following steps

[0008] S1. A rectangular excitation coil is set parallel to the metal to be tested. When the excitation coil is energized, a uniformly distributed and parallel induced current is generated on the surface of the metal to be tested. The length direction of the metal to be tested is the x-axis, the width direction is the y-axis, and the thickness direction is the z-axis. The magnitude of the magnetic field in the z-axis direction is recorded as B(z);

[0009] S2, determining the detection interval D, threshold A and threshold B according to the parameters of the excitation coil, taking the defect on the metal to be tested as the center and taking multiple scanning paths parallel to each other and arranged along the x-axis direction at intervals of D to the upper and lower sides of the defect, and forming a scanning curve group based on the amplitude of the magnetic field B(z) on the scanning path through multi-path scanning;

[0010] S3, marking the maximum value of each scanning curve in the scanning curve group, and mapping the horizontal coordinate value of the maximum value point to the corresponding scanning path, and constructing feature points respectively;

[0011] S4, rotating the scanning path of the excitation coil by 90° on the xy plane and repeating steps S2 to S3 to obtain feature points in the same manner;

[0012] S5. Screen all feature points by relative comparison method and absolute value method, and remove singular points and invalid feature points;

[0013] S6, determining whether the defect profile on the metal to be tested is a vertical long straight profile, and if so, optimizing the vertical long straight profile;

[0014] S7, performing data correction on the feature points located at the edge of the metal to be measured;

[0015] S8, connecting the feature points in sequence according to their coordinates to form a closed defect contour, and obtaining a preliminary defect fitting contour;

[0016] S9. Set the excitation current frequency of the excitation coil, and optimize the preliminary defect fitting contour according to the actual excitation current frequency to obtain the final defect contour.

[0017] The technical solution further defined in the present invention is:

[0018] Furthermore, in step S1, the cross section of the excitation coil in the yz plane is set to a rectangle, and the excitation coil is wound along the circumference of the rectangle. A sliding roller is provided at the bottom of the excitation coil, and the excitation coil moves along the x-axis direction through the roller to scan the metal surface to be tested.

[0019] As described above, in a defect contour recognition method based on the flow characteristics of uniform parallel induced current, in step S3, the maximum value of each scanning curve includes two positive and negative maximum values, that is, the magnetic field B(z) presents two positive and negative maximum values, and the magnetic field B(z) value at the middle position of the defect is 0.

[0020] As described above, in a defect contour recognition method based on the flow characteristics of uniform parallel induced current, in step S5, a relative comparison method is performed on each feature point on the scanning curve except the first point and the last point to exclude singular points therein; the relative comparison method is specifically as follows: for a group of scanning curves in the same direction, any feature point on any scanning curve is recorded as a2, and the two closest points on both sides are recorded as a1 and a3; only when |a1-a2|>threshold A and |a2-a3|>threshold A, the feature point a2 is identified as a singular point and the singular point is excluded.

[0021] As described above, in a defect contour recognition method based on the flow characteristics of uniform parallel induced current, in step S5, invalid feature points are removed by an absolute value method, specifically: the size of the maximum value of each feature point is determined, when the absolute value of the maximum value is greater than a threshold value B, the feature point is considered to be a valid feature point and is labeled, when the absolute value of the maximum value is less than or equal to the threshold value B, the feature point is considered to be an invalid feature point and is removed; when the maximum value of a valid feature point is a minimum value, the minimum value is labeled by an odd number, and when the maximum value of a valid feature point is a maximum value, the maximum value is labeled by an even number;

[0022] Define two two-dimensional arrays H[i][j] and Z[i][j], where i represents the scanning direction, which is 1 along the x-axis and 2 along the y-axis; j represents the number of the feature point; the horizontal coordinate information of the feature point is stored in the horizontal coordinate array H[i][j], and the vertical coordinate information of the feature point is stored in the vertical coordinate array Z[i][j].

[0023] As described above, in a defect contour recognition method based on the flow characteristics of uniform parallel induced current, in step S5, feature points are firstly screened for data obtained by scanning along the x-axis direction, and then feature points are screened for data obtained by scanning along the y-axis direction.

[0024] In the defect contour recognition method based on the uniform parallel induced current flow characteristics as described above, in step S6, a class abnormal situation is defined, and the abnormal situation is that the maximum difference between the horizontal coordinates of the feature points on the upper and lower sides perpendicular to the scanning direction of the feature point is within 5 times of the scanning step length, the maximum difference between the horizontal coordinates of the feature point and the feature point in the middle area perpendicular to the scanning direction is between 10-15 times of the scanning step length, and the maximum difference between the horizontal coordinates of the feature point and the feature point of the actual middle point is between 5-10 times of the scanning step length;

[0025] For the two-dimensional arrays H[i][j] and Z[i][j], if a similar abnormal situation occurs, the defect contour is considered to be a vertical long straight contour, and the value of the horizontal coordinate of the feature point in the middle area perpendicular to the scanning direction is reduced so that the maximum difference between the feature point with the minimum or maximum value and the feature point in the middle area perpendicular to the scanning direction is reduced to within 5 times of the scanning step.

[0026] As described above, in a defect contour recognition method based on the flow characteristics of uniform parallel induced current, in step S6, the optimization method of the vertical long straight contour is set to: rotate the scanning path of the excitation coil by 45° on the xy plane and then repeat steps S2 to S5.

[0027] As described above, in a defect contour recognition method based on the flow characteristics of uniform parallel induced current, in step S7, the distance between the feature point on the scanning curve and the edge of the metal being measured is recorded as d, and data compensation is performed only for feature points with d values ​​less than 2 mm, and the compensation value is (2-d)×0.2.

[0028] As described above, a defect contour recognition method based on the flow characteristics of uniform parallel induced current, in step S9, the excitation current frequency of the excitation coil is set to 200Hz~20kHz, and the excitation current frequency reference value is set to f0=1kHz. The angle θ between the tangent line at different points on the preliminary defect fitting contour and the scanning direction and the proportional coefficient K are set. For each point on the preliminary defect fitting contour, the value of K*|f-f0|*(θ-90°) is calculated. If the value is a positive number, the point is translated in the positive direction of the x-axis; if the value is a negative number, the point is translated in the negative direction of the x-axis, and the translation distance is the absolute value of the value.

[0029] The beneficial effects of the present invention are:

[0030] (1) In the present invention, the conversion between the defect contour, induced current flow, and magnetic field change is strictly correlated with each other, so it has the advantage of high detection accuracy; the uniform and mutually parallel induced current ensures that the current flows along the defect contour and makes the original "background field" zero. At this time, the measured magnetic field will accurately restore the flow state of the induced current on the surface of the object being measured;

[0031] (2) In the present invention, there is no need to magnetize the object to be tested, and the coil excitation is relatively flexible, thus avoiding the current situation where the leakage magnetic field detection is only magnetized in a single direction, resulting in serious leakage magnetic field distortion of complex defects;

[0032] (3) In the present invention, the detection equipment is simple and easy to operate. It does not require more complex imaging equipment. The complexity of the equipment will increase the error of the system and also requires professional technicians to operate and maintain it, thereby reducing the implementation cost and difficulty of the method. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 This is a principle diagram of uniform parallel induction current flow characteristics in an embodiment of the present invention;

[0034] Figure 2 It is a schematic diagram of the overall process of an embodiment of the present invention;

[0035] Figure 3 This is a schematic diagram of an implementation model in an embodiment of the present invention;

[0036] Figure 4 This is a comparison diagram between the defect contour structure and the actual defect in an embodiment of the present invention. DETAILED DESCRIPTION

[0037] This embodiment provides a defect contour recognition method based on the flow characteristics of uniform parallel induced current, which quickly and accurately obtains metal surface defect contour information through the flow characteristics of uniform parallel induced current, realizes intuitive visualization of the defect contour, and improves the application level and characterization ability of traditional electromagnetic detection.

[0038] According to the principle of electromagnetic induction, when a coil with alternating current is close to the metal being measured, an induced current will be formed on the metal surface. The direction of the induced current is closely related to the structure of the excitation coil and the form of the excitation current. Uniform parallel induced current is a new detection technology based on the principle of electromagnetic induction. Through the specific planning of the excitation probe structure, coil winding process, and excitation current technology, it ensures that the induced current is evenly distributed and parallel to each other on the surface of the workpiece being measured.

[0039] If there are no defects on the metal surface being measured, the current under the coil is evenly distributed, and the current lines are parallel to each other. If there are defects on the metal surface being measured, the change in the resistance value of the defective part will affect the distribution of the uniform parallel induced current, causing the uniform parallel induced current to "walk along the track" along the defect contour at the defect, that is, the uniformly distributed induced current will flow along the defect contour. Changes in the uniform parallel induced current will cause corresponding changes in the distribution of the induced magnetic field. The spatial magnetic field on the metal surface can be divided into three components, and this embodiment selects the vertical component B(z).

[0040] The original uniform and parallel current has its flow path changed due to the existence of defects, resulting in the redistribution of the current on the metal surface, such as Figure 1 Therefore, by measuring and analyzing the changes in the magnetic field around the metal surface, the flow characteristics of the uniform parallel current on the metal surface can be analyzed, thereby further realizing the accurate identification of the defect profile of the metal surface.

[0041] The solution for realizing metal surface defect contour recognition based on the uniform parallel induced current flow characteristics has the following advantages: on the one hand, the uniform and mutually parallel induced current ensures that the current flows along the defect contour, that is, the induced current follows the "runway" law; on the other hand, the uniform and mutually parallel induced current ensures that the original induced current magnetic field is zero, that is, the "background field" is zero. At this time, the measured magnetic field will truly restore the flow state of the induced current on the surface of the measured object. The conversion between the three physical quantities of defect contour, induced current flow, and magnetic field change has a strict corresponding relationship with each other.

[0042] This embodiment provides a defect contour recognition method based on the uniform parallel induced current flow characteristics, such as Figure 2 As shown, the following steps are included

[0043] S1, such as Figure 3 As shown in FIG. 1 , it is a schematic diagram of the implementation model of this embodiment. A metal aluminum plate is selected as the metal to be tested. A circular defect is selected on the metal aluminum plate for defect contour recognition. The diameter of the circular defect is 6 mm and the depth is 1 mm. A rectangular excitation coil is set parallel to the metal to be tested. After the excitation coil is energized, a uniformly distributed and parallel induced current is generated on the surface of the metal to be tested. The size of the excitation coil is 25×25×40 mm, the number of clockwise winding turns is 1000 N, the current size is 10 A, and the conductivity of the excitation coil wire is 6×10 7 S / m, a sinusoidal current with a frequency of 2000Hz is passed through the exciting coil.

[0044] The length direction of the metal to be tested is taken as the x-axis, the width direction is taken as the y-axis, and the thickness direction is taken as the z-axis. The z-axis is perpendicular to the metal to be tested and points into the air. The magnitude of the magnetic field in the z-axis direction is recorded as B(z). The rectangular excitation coil is placed along the x-axis direction with a lifting height of 2mm. A sliding roller is provided at the bottom of the excitation coil to facilitate scanning on the surface of the metal to be tested.

[0045] Since the effective excitation area of ​​the excitation coil is limited, the position of the excitation coil is continuously moved to make it act on the metal to be tested at different positions; the excitation coil is moved in two mutually perpendicular directions of 0° (x-axis) and 90° (y-axis), and the magnetic field size in the z-axis direction is obtained 1.5 mm directly below the center of the excitation coil; during the actual operation, a magnetic sensor is installed under the rectangular excitation coil to obtain the magnetic field values ​​at different positions.

[0046] S2. Determine the detection spacing D, threshold A and threshold B according to the parameters of the excitation coil. In this embodiment, the detection spacing D is set to 1 mm. In different cases, the values ​​of threshold A and threshold B are different. When the number of turns of the excitation coil is 1000N, the coil current is 10A, and the conductivity of the coil wire is 6×10 7 S / m, when a sinusoidal current with a frequency of 2000Hz is passed, the threshold A is 0.0045T and the threshold B is 0.0066789T; after experimental data analysis, when the number of turns of the excitation coil is 1000N and the frequency range is between 500Hz and 11000Hz, the threshold B satisfies B=3×10 —16 x 3 -6×10 -11 x 2 +1×10 —6 The relationship is x+0.0043, where x represents the current frequency in Hertz.

[0047] Taking the defect as the center, multiple parallel scanning paths are taken at intervals of 1 mm to the upper and lower sides of the defect. The B(z) magnetic field corresponding to each scanning path is collected to obtain a 0° multi-path scanning curve group of the circular defect. For this curve group, each scanning curve will have a peak and a trough. The positions of the peaks or troughs of all scanning curves correspond exactly to the edge positions of the defect contour.

[0048] S3. Mark the maximum value of each scanning curve in the scanning curve group, and map the horizontal coordinate value of the maximum value point to the corresponding scanning path to construct feature points respectively.

[0049] When an excitation coil with alternating current is close to the metal being measured, if there are no defects on the surface of the metal being measured, the current is evenly distributed and the current lines are parallel to each other; if there are defects on the surface of the metal being measured, the change in resistivity of the defect will affect the distribution of the induced current, causing the current lines to deflect at the edge of the defect. The change in the induced current will cause a corresponding change in the distribution of the induced magnetic field. The deflection of the current direction determines the magnetic field B(z), and the maximum value of the current deflection appears on both sides of the defect. Since the current deflection directions on both sides of the defect are opposite, the magnetic field B(z) presents two maximum values, positive and negative, and the magnetic field B(z) value in the middle position is 0.

[0050] S4. Rotate the scanning path of the excitation coil by 90° on the xy plane and repeat steps S2 to S3 to obtain feature points in the same manner.

[0051] The single scanning curve corresponding to the magnetic field B(z) amplitude measured on each scanning path is constructed. The scanning curve usually consists of two positive and negative extreme values, and the distance between the peaks on the horizontal axis corresponds to the length of the cross section of the corresponding defect on the scanning path. Through multiple scans in parallel directions, a scanning curve group is further formed. According to the scanning curve group of the magnetic field B(z), the extreme value of each scanning curve is marked.

[0052] On this basis, according to the scanning path position corresponding to the scanning curve group, the horizontal coordinate value of the maximum point is mapped to the corresponding scanning path, and the characteristic points are constructed respectively. Finally, the adjacent characteristic points are connected to realize the quantitative identification of all or part of the specific contour of the defect. Then, in the same way, the excitation coil is rotated along the 90° direction to construct the magnetic field B (z) scanning curve group on different scanning paths with a 90° orientation. Repeating the above operation can realize the accurate identification of the entire defect contour of the metal surface. The scanning curve group formed based on the above method and the constructed defect contour (black bold solid line) and the actual defect contour are shown in Figure 2. Figure 4 shown.

[0053] In some cases, multiple scans in one direction can get the outline of the defect, but it is still recommended to scan in different directions. Because by rotating the scanning direction by an angle, the defect outlines obtained twice are comprehensively analyzed, and the feature points are selected to make the defect judgment more accurate. Therefore, the defects in the scanning direction of the metal surface can be accurately identified by analyzing the magnetic field B (z) curve group of multi-path scanning.

[0054] S5. In the actual detection process, feature points are affected by many factors. For example, due to interference from the external environment, some data unrelated to the defect may be mistaken for useful feature point data. Therefore, effective screening of feature point data is very important for accurately identifying defect contours. All feature points are screened by relative comparison method and absolute value method, and singular points and invalid feature points are removed to improve the accuracy of defect contour recognition.

[0055] A relative comparison method is performed on each feature point except the first and last points on the scanning curve to exclude singular points. Specifically, for a group of scanning curves in the same direction, any feature point on any scanning curve is recorded as a2, and the two closest points on both sides are recorded as a1 and a3. Only when |a1-a2|>threshold A and |a2-a3|>threshold A, the feature point a2 is identified as a singular point and excluded. In this embodiment, the threshold A is 0.0045T.

[0056] The absolute value method is as follows: the size of the maximum value of each feature point is determined. When the absolute value of the maximum value is greater than the threshold B, the feature point is considered to be a valid feature point and is labeled. When the absolute value of the maximum value is less than or equal to the threshold B, the feature point is considered to be an invalid feature point and is removed. When the maximum value of a valid feature point is the minimum value, the minimum value is labeled with an odd number (1, 3, 5, ...). When the maximum value of a valid feature point is the maximum value, the maximum value is labeled with an even number (2, 4, 6, ...).

[0057] Define two two-dimensional arrays H[i][j] and Z[i][j], where i represents the scanning direction, which is 1 along the x-axis (i.e., 0° direction) and 2 along the y-axis (i.e., 90° direction); j represents the number of the feature point; the horizontal coordinate information of the feature point is stored in the horizontal coordinate array H[i][j], and the vertical coordinate information of the feature point is stored in the vertical coordinate array Z[i][j]. Then, the minimum feature points and the maximum feature points of the two scanning directions are connected in sequence, and the obtained feature points are effectively connected, which reflects various information of the defect contour, and the visualization of the defect contour of the metal surface can be realized.

[0058] S6. Determine whether the defect profile on the metal to be tested is a vertical long straight profile. If so, optimize the vertical long straight profile. For defects with different profiles, the construction accuracy of the general method is different. For example, the construction error of the square defect is large. The essence of the existence of such a phenomenon is still related to the flow characteristics of the induced current, that is, there are differences in the distribution state of defects with different profiles. For a square defect, taking the leftmost side profile as an example, its shape is in a vertical state. When parallel induced currents pass through, all currents are blocked by obstacles in the same direction and cannot pass through. At this time, all currents accumulate near it. Due to the particularity of the defect profile, there is a phenomenon of mutual squeezing between these blocked currents, which increases the change in the direction of the induced current, and then brings some errors. Since the changes in the direction of these current flows are caused by the backlog between each other, rather than the defect profile, it is necessary to optimize the defect data of this vertical long straight line.

[0059] To deal with the above problems, this embodiment proposes two different solutions. The first method is to define a class of abnormal situations based on the analysis and processing of detection data features. The abnormal situation is that the maximum difference between the horizontal coordinates of the feature points on the upper and lower sides perpendicular to the scanning direction of the feature point is within 5 times of the scanning step, the maximum difference between the horizontal coordinates of the feature point and the feature points in the middle area perpendicular to the scanning direction is between 10-15 times of the scanning step, and the maximum difference between the horizontal coordinates of the feature point and the actual middle point is between 5-10 times of the scanning step.

[0060] For the two-dimensional arrays H[i][j] and Z[i][j], if a similar abnormal situation occurs, the defect contour is considered to be a vertical long straight contour, and the value of the horizontal coordinate of the feature point in the middle area perpendicular to the scanning direction is reduced so that the maximum difference between the minimum or maximum feature point and the horizontal coordinate of the feature point in the middle area perpendicular to the scanning direction is reduced to within 5 times of the scanning step length. This correction method can reduce the error of special contours to a certain extent.

[0061] The second method is based on the improvement and optimization of the probe scanning method. In addition to directly analyzing and processing the test data, the accuracy can be improved by rotating the direction of the excitation coil scanning by a certain angle, such as rotating 45° on the xy plane and then rescanning. The change in the probe scanning direction makes the angle between the induced current and the defect contour no longer a vertical relationship. Of course, the same effect can be achieved by keeping the probe scanning direction unchanged and adjusting the position of the metal to be tested. At this time, the flow characteristics of the induced current will change differently for contours at different angles, thereby reducing the problem of mutual extrusion. This improves the accuracy of the contour construction to a certain extent.

[0062] S7. Correct the data of the feature points at the edge of the metal to be tested; the edge effect of the metal to be tested is a difficult point for defect detection, so the influence of this factor must be considered and effective measures must be taken to solve it. If the defect is located far away from the edge of the metal to be tested, the scanning curve will not only have a peak at the edge of the defect, but also similar peaks at the left and right edges of the metal to be tested. However, since the defect is far away from the edge of the metal to be tested, the magnetic field change caused by the edge of the metal to be tested has little effect on the magnetic field change caused by the defect, and the recognition of the defect contour can be ignored.

[0063] In addition, experiments have confirmed that there is a certain difference between the peak value of the magnetic field caused by the edge of the conductor and the change caused by the defect. The approximate straight line fitted by the edge of the metal to be tested has a certain deviation in position from the actual edge of the metal to be tested, and is offset to the outside. It can be inferred that this is related to the particularity of the distribution of the induced current at the edge. Therefore, such characteristics can be used to effectively identify the edge of the metal to be tested.

[0064] If the defect is located near the right edge of the metal to be tested, the scanning curve will not only have a peak at the defect edge, but also similar peaks at the left and right edges of the metal to be tested. Since the defect is close to the right edge of the metal to be tested, the magnetic field change caused by the right edge of the defect is affected to a certain extent, which has a non-negligible impact on the recognition of the right contour of the defect, thereby reducing the defect recognition accuracy.

[0065] Therefore, the edge of the metal to be measured not only affects the magnetic field distribution of the scanning curve, but also has an adverse effect on the identification of defects near the edge of the metal to be measured. Close to the edge of the metal to be measured, the identified feature points will deviate from the actual position due to the squeezing effect of the current. This embodiment proposes to correct the data by the distance d between the feature point on the scanning curve and the edge of the metal to be measured, while no compensation is performed for defects that are far away from the edge of the metal to be measured. Specifically, the following scheme is used to correct the data: data compensation is performed for feature points with d values ​​less than 2 mm, and the compensation value is (2-d)×0.2; feature points with d values ​​greater than or equal to 2 mm are not processed.

[0066] S8. Connect each feature point in sequence according to its coordinates to form a closed defect contour, and obtain a preliminary defect fitting contour.

[0067] S9. Set the excitation current frequency of the excitation coil, and optimize the preliminary defect fitting contour according to the actual excitation current frequency to obtain the final defect contour, and realize the visualization of the specific defect contour and the display of defect contour information by writing a host computer application.

[0068] For electromagnetic detection technology, the selection of excitation frequency is crucial. Theoretically, the greater the excitation current frequency, the more concentrated the current is on the metal surface to be tested. Therefore, the excitation current frequency will directly affect the distribution of the induced current in the depth direction, and will not have much effect on the identification of the defect surface contour. However, for this embodiment, the defect contour is identified based on the change law of the flow characteristics of the uniform parallel induced current. In some special cases, the parallel induced current is hindered by the defect contour, and its flow direction will also extend along the depth direction.

[0069] By setting the coil frequency to 200Hz, 2000Hz and 20kHz, the influence of the excitation frequency on the identification of circular defects and the fitting results is studied and analyzed. The results show that the three excitation current frequencies can complete the identification of the defect contour with high accuracy. Of course, the amplitude and degree of influence of the detection magnetic field are also different due to the influence of the excitation current frequency. As the excitation current frequency increases, the magnetic induction intensity B(z) is also larger, and the degree of influence by the defect edge is also aggravated. The lower the excitation current frequency, the magnetic induction intensity changes slightly in advance when gradually approaching the defect. This is not conducive to truly reflecting the actual contour size of the defect. The occurrence of these phenomena is related to the change in the distribution of the induced current on the metal surface to be tested caused by the change in the excitation current frequency. The selection of the excitation current frequency is more important and should be selected according to different situations.

[0070] Therefore, the excitation current frequency range of this detection method is set between 200Hz and 20kHz. When the excitation current frequency changes within this range, the error in defect contour recognition is small. Of course, this error still exists, and further measures are still needed when higher accuracy requirements are required.

[0071] According to the above change rules, in order to improve the measurement accuracy of the defect contour, an excitation current frequency reference value f0=lkHz is set, and the fitting defect contour is optimized according to the actual excitation current frequency f. Assume the angle θ between the tangent line of different points on the preliminary defect fitting contour and the scanning direction and the proportional coefficient K. For each point of the defect contour on the left side of the scanning direction (i.e., the negative side of the x-axis), calculate the value of K*|f-f0|*(θ-90°). If the value is positive, translate the point to the right (i.e., the positive side of the x-axis); if the value is negative, translate the point to the left; for each point of the defect contour on the right side, calculate the value of K*|f-f0|*(θ-90°), and translate the point according to the same rule according to the value. The translation distance is the absolute value of the value. By calculating the proportional coefficient K, it is taken as 0.4350649×10 -4 The accuracy of the defect contour is improved by correcting the contours in the two scanning directions respectively.

[0072] In addition to the above embodiments, the present invention may also have other implementation modes. Any technical solution formed by equivalent replacement or equivalent transformation falls within the protection scope required by the present invention.

Claims

1. A defect contour recognition method based on uniform parallel induced current flow characteristics, characterized in that: The following steps are included S1. A rectangular excitation coil is arranged parallel to the metal to be tested, and a magnetic sensor is installed below the rectangular excitation coil to obtain the magnetic field values ​​at different positions. After the excitation coil is energized, a uniformly distributed and parallel induced current is generated on the surface of the metal to be tested. The length direction of the metal to be tested is the x-axis, the width direction is the y-axis, and the thickness direction is the z-axis. The magnitude of the magnetic field in the z-axis direction is recorded as B(z); S2, determining the detection interval D, threshold A and threshold B according to the parameters of the excitation coil, taking the defect on the metal to be tested as the center and taking multiple scanning paths parallel to each other and arranged along the x-axis direction at intervals of D to the upper and lower sides of the defect, and forming a scanning curve group based on the amplitude of the magnetic field B(z) on the scanning path through multi-path scanning; S3, marking the maximum value of each scanning curve in the scanning curve group, and mapping the horizontal coordinate value of the maximum value point to the corresponding scanning path, and constructing characteristic points respectively; the maximum value of each scanning curve includes positive and negative maximum values, that is, the magnetic field B(z) presents positive and negative maximum values, and the magnetic field B(z) value at the middle position of the defect is 0; S4, rotating the scanning path of the excitation coil by 90° on the xy plane and repeating steps S2 to S3 to obtain feature points in the same manner; S5. Screen all feature points by relative comparison method and absolute value method, and remove singular points and invalid feature points; perform relative comparison method on each feature point except the first point and the last point on the scanning curve to remove singular points; The relative comparison method is as follows: for a group of scanning curves in the same direction, any feature point on any scanning curve is recorded as a2, and the two closest points on both sides are recorded as a1 and a3; only when |a1-a2|>threshold A and |a2-a3|>threshold A, the feature point a2 is identified as a singular point and excluded; S6, determining whether the defect profile on the metal to be tested is a vertical long straight profile, and if so, optimizing the vertical long straight profile; S7, performing data correction on the feature points located at the edge of the metal to be measured; S8, connecting the feature points in sequence according to their coordinates to form a closed defect contour, and obtaining a preliminary defect fitting contour; S9. Set the excitation current frequency of the excitation coil, and optimize the preliminary defect fitting contour according to the actual excitation current frequency to obtain the final defect contour.

2. The defect contour recognition method based on uniform parallel induced current flow characteristics according to claim 1, characterized in that: In step S1, the cross section of the excitation coil on the yz plane is set to a rectangle, and the excitation coil is wound along the circumference of the rectangle. A sliding roller is provided at the bottom of the excitation coil, and the excitation coil moves along the x-axis direction through the roller to scan the metal surface to be measured.

3. The defect contour recognition method based on uniform parallel induced current flow characteristics according to claim 1, characterized in that: In step S5, invalid feature points are removed by using the absolute value method, specifically: the size of the maximum value of each feature point is determined, when the absolute value of the maximum value is greater than the threshold value B, the feature point is considered to be a valid feature point and is labeled, when the absolute value of the maximum value is less than or equal to the threshold value B, the feature point is considered to be an invalid feature point and is removed; when the maximum value of the valid feature point is the minimum value, the minimum value is labeled with an odd number, and when the maximum value of the valid feature point is the maximum value, the maximum value is labeled with an even number; Define two two-dimensional arrays H[i][j] and Z[i][j], where i represents the scanning direction, which is 1 along the x-axis and 2 along the y-axis; j represents the number of the feature point; the horizontal coordinate information of the feature point is stored in the horizontal coordinate array H[i][j], and the vertical coordinate information of the feature point is stored in the vertical coordinate array Z[i][j].

4. The defect contour recognition method based on uniform parallel induced current flow characteristics according to claim 1, characterized in that: In step S5, feature points are firstly screened for data obtained by scanning along the x-axis direction, and then feature points are screened for data obtained by scanning along the y-axis direction.

5. The defect contour recognition method based on uniform parallel induced current flow characteristics according to claim 1, characterized in that: In the step S6, a class of abnormal conditions is defined, where the maximum difference between the horizontal coordinates of the feature points on the upper and lower sides of the feature point perpendicular to the scanning direction is within 5 times of the scanning step length, the maximum difference between the horizontal coordinates of the feature point and the feature point in the middle area perpendicular to the scanning direction is between 10-15 times of the scanning step length, and the maximum difference between the horizontal coordinates of the feature point and the feature point of the actual middle point is between 5-10 times of the scanning step length; For the two-dimensional arrays H[i][j] and Z[i][j], if a similar abnormal situation occurs, the defect contour is considered to be a vertical long straight contour, and the value of the horizontal coordinate of the feature point in the middle area perpendicular to the scanning direction is reduced so that the maximum difference between the feature point with the minimum or maximum value and the feature point in the middle area perpendicular to the scanning direction is reduced to within 5 times of the scanning step.

6. The defect contour recognition method based on uniform parallel induced current flow characteristics according to claim 1, characterized in that: In step S6, the optimization method of the vertical long straight profile is set as follows: the scanning path of the excitation coil is rotated 45° on the xy plane and then steps S2 to S5 are repeated.

7. The defect contour recognition method based on uniform parallel induced current flow characteristics according to claim 1, characterized in that: In step S7, the distance between the characteristic point on the scanning curve and the edge of the metal being measured is recorded as d, and data compensation is performed only for the characteristic points with d values ​​less than 2 mm, and the compensation value is (2-d)×0.

2.

8. The defect contour recognition method based on uniform parallel induced current flow characteristics according to claim 1, characterized in that: In step S9, the excitation current frequency of the excitation coil is set to 200 Hz to 20 kHz, and the excitation current frequency reference value is set to f0 = 1 kHz. The angle θ between the tangent line at different points on the preliminary defect fitting contour and the scanning direction and the proportional coefficient K are set. For each point on the preliminary defect fitting contour, the value of K*|f-f0|*(θ-90°) is calculated. If the value is a positive number, the point is translated in the positive direction of the x-axis; if the value is a negative number, the point is translated in the negative direction of the x-axis, and the translation distance is the absolute value of the value.

Images