A ferromagnetic material hardness evaluation method based on magnetic barkhausen noise avalanche three-dimensional visualization

Abstract

The application discloses a ferromagnetic material hardness evaluation method based on magnetic Barkhausen noise (MBN) avalanche three-dimensional visualization. The method acquires MBN original signal and tangential magnetic field signal data of a ferromagnetic material surface; sets an adaptive threshold to extract discrete MBN avalanche events, respectively calculates the avalanche size of each event and the instantaneous tangential magnetic field intensity of the corresponding center moment; divides a grid on a "magnetic field-avalanche size" characteristic plane and counts event logarithmic characteristic density, and uses a Gaussian smoothing algorithm to construct a three-dimensional statistical manifold surface; finally, extracts the double-peak interval spread characteristics of the manifold surface and the topological evolution law of bidirectional outward migration of the manifold surface to the high magnetic field direction, so as to quantitatively evaluate the material hardness.

Description

Technical Field

[0001] This invention relates to the field of nondestructive testing and evaluation technology for materials, specifically to the microstructure and hardness testing technology during the manufacturing and service of ferromagnetic materials, and particularly to a method for material hardness evaluation based on magnetic Barkhausen noise (MBN) avalanche feature extraction, three-dimensional visualization, and material hardness evaluation. Background Technology

[0002] In the field of engineering manufacturing technology, the surface hardness and residual stress of mechanical components (such as bearings, gears, and rails) are key indicators determining their fatigue life and service reliability. Magnetic Barkhausen noise is a discontinuous, jumping signal generated when ferromagnetic materials are dynamically magnetized under an alternating magnetic field. This noise occurs because the irreversible displacement of magnetic domain walls and the reversal of magnetic moments within the material are hindered by pinning points such as grain boundaries, dislocations, precipitates, and micro-stresses. This signal is highly sensitive to changes in the microstructure of ferromagnetic materials, thus having significant application potential in material hardness assessment and non-destructive testing engineering practices.

[0003] Currently, commonly used MBN analysis methods in engineering mainly focus on extracting one-dimensional macroscopic features of the signal, such as root mean square (RMS), peak amplitude, and envelope area, and establishing calibration curves between these one-dimensional features and material hardness or stress. However, this "macroscopic averaging" approach has significant limitations: First, macroscopic RMS and other amplitude features are easily affected by sensor lift-off, workpiece surface roughness, and instrument gain fluctuations, leading to large measurement errors in actual testing; Second, macroscopic features average thousands of independent microscopic magnetic domain jumping events, smoothing out the transient dynamic details of the interaction between the magnetic domain walls and pinning points of different strengths within the material, resulting in insufficient sensitivity for distinguishing materials with similar hardness levels; Third, existing analysis results lack intuitive multidimensional spatial characterization and cannot reflect the microscopic magnetization reversal physical process of the material with visualized topological morphology.

[0004] In summary, how to deeply explore the microscopic dynamic characteristics of MBN signals, overcome the problem of the high dependence of traditional one-dimensional features on absolute signal amplitude, and provide a highly sensitive, intuitive, and visual quantitative assessment method for hardness is a pressing problem that the engineering community needs to solve. Summary of the Invention

[0005] To address the problems in existing technologies, this invention provides a method for three-dimensional visualization and hardness assessment of magnetic Barkhausen noise avalanches. This method breaks away from the conventional approach of averaging continuous signals, deconstructing the MBN signal into discrete avalanche events. It constructs a three-dimensional topological manifold using the instantaneous tangential magnetic field, avalanche size, and logarithmic characteristic density, and utilizes the phase transitions and evolution of the manifold morphology to achieve accurate hardness measurement.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] A method for three-dimensional visualization and hardness assessment of magnetic Barkhausen noise avalanches includes the following steps:

[0008] Step (1): Acquire the raw signal data of the magnetic Barkhausen noise of the ferromagnetic material under test under alternating magnetic field excitation and the synchronous tangential magnetic field signal data;

[0009] Step (2): Process the original magnetic Barkhausen noise signal, set an adaptive noise threshold, segment the continuous signal and extract several discrete magnetic Barkhausen noise avalanche events;

[0010] Step (3) Extract the feature parameter pairs for each avalanche event ( , ),in This refers to the instantaneous tangential magnetic field strength data corresponding to the moment at which the event occurred. This refers to the avalanche size data for this event;

[0011] Step (4) Divide the area into a grid on the two-dimensional feature plane formed by the tangential magnetic field and the avalanche size, count the number of avalanche events falling into each grid to calculate the logarithmic feature density, and use a smoothing algorithm to construct a grid with the tangential magnetic field as the reference. X-axis, avalanche size A three-dimensional statistical manifold surface with Y-axis and logarithmic characteristic density on Z-axis;

[0012] Step (5) Visualize the topological features of the three-dimensional statistical manifold surface and evaluate the hardness of the tested ferromagnetic material based on these topological features.

[0013] In one embodiment, the formula for calculating the adaptive noise threshold in step (2) is: .

[0014] In one embodiment, the avalanche size in step (3) is calculated by integrating the effective voltage over the duration of a single avalanche event, as shown in the following formula: in and These represent the start and end times of the avalanche event crossing the adaptive noise threshold, respectively. The signal is a magnetic Barkhausen noise signal, and the avalanche size is determined in step (4). Perform logarithmic scaling .

[0015] In one embodiment, the construction process of the three-dimensional statistical manifold surface in step (4) specifically includes: in Divide the plane into a uniform grid and count the number of events falling into each grid. Logarithmic transformation of the number of events ;right The matrix is ​​smoothed using a two-dimensional Gaussian filter to obtain a continuous logarithmic eigendensity matrix. And generate a three-dimensional mesh surface.

[0016] In one embodiment, the topological features based on the three-dimensional statistical manifold surface in step (5) are specifically manifested as follows: due to the symmetry of the alternating magnetization period, the three-dimensional statistical manifold surface exhibits a bimodal structure; when it exhibits a narrow-spacing bimodal shape, due to the low domain wall motion barrier, the bimodal peaks move towards the central low field region, resulting in lower material hardness; when it exhibits a wide-spacing bimodal shape, due to the enhanced micro-pinning effect leading to delayed domain wall motion, the magnetic field spacing between the bimodal peaks increases significantly, and the avalanche distribution centroid migrates bidirectionally towards the high magnetic field direction, resulting in higher material hardness.

[0017] In one embodiment, step (5) further includes the step of constructing a peak offset layered evolution diagram: the logarithmic characteristic density of the three-dimensional statistical manifold surface is summed and projected along the avalanche size axis to obtain the magnetic field-density evolution curve; the magnetic field-density evolution curves of multiple sets of standard test blocks with known hardness are translated and layered at equal intervals along the vertical axis, the peak point coordinates of each curve are extracted, and a peak offset trajectory calibration model on the tangential magnetic field axis is generated; the evolution curve of the material to be tested is substituted into the calibration model to obtain its quantified hardness value.

[0018] In one embodiment, step (5) further includes the step of constructing a two-dimensional topological contour fingerprint: extracting the feature contours of the three-dimensional statistical manifold surface, projecting the contours of different hardness states onto the same two-dimensional plane, and using the magnetic field broadening and region separation features of the closed contour to realize the hardness detection of the material state.

[0019] On the other hand, the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps of the above-mentioned method for three-dimensional visualization and hardness assessment of magnetic Barkhausen noise avalanche.

[0020] As can be seen from the above description, this application innovatively proposes a new method for dimensionality reduction and three-dimensional visualization based on the statistical manifold of micro avalanches with magnetic Barkhausen noise. Its beneficial effects are as follows: (1) It achieves a leap in analytical dimensions and extremely high detection sensitivity. This method starts from the statistical perspective of micro avalanche events, avoiding the defect of micro-features being wiped out by averaging. For materials with similar hardness levels, it can clearly distinguish them through micro-topological evolution such as "peak splitting"; (2) It has excellent anti-interference and anti-lift-off capabilities. This method does not rely on the absolute voltage amplitude of the signal, but relies on the relative lateral offset of the feature surface in physical space (tangential magnetic field axis), effectively eliminating the errors caused by sensor lift-off and instrument gain fluctuations; (3) The results are visualized intuitively and are easy to interpret in engineering automation. The generated three-dimensional manifold diagram, peak offset layer evolution diagram, etc. provide a unique "topological fingerprint" for the material, which is extremely suitable for combining computer vision and deep learning algorithms to realize the rapid automatic sorting of the heat treatment quality of large batches of components. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments will be briefly introduced below.

[0022] Figure 1 This is a flowchart illustrating the method for three-dimensional visualization and hardness assessment of magnetic Barkhausen noise avalanche in an embodiment of the present invention.

[0023] Figure 2 This is a close-up schematic diagram of the preprocessing of the original magnetic Barkhausen noise signal and the extraction of discrete avalanche events in an embodiment of the present invention.

[0024] Figure 3 In a specific application example of the present invention, the MBN avalanche three-dimensional topographic map of a low-hardness (soft state) specimen (Example 23.0 HRC);

[0025] Figure 4 In a specific application example of the present invention, the MBN avalanche three-dimensional topographic topography map of a high-hardness (hard state) specimen (Example 62.8 HRC);

[0026] Figure 5 This is a specific application example of the present invention, showing the evolution of the MBN avalanche active peak shift and stacking.

[0027] Figure 6 In a specific application example of the present invention, a two-dimensional contour fingerprint of microscopic avalanche hotspots shifting with hardness is shown.

[0028] Figure 7 In a specific application example of the present invention, the error bar and correlation calibration model diagram of the variation of the tangential magnetic field of the main peak of avalanche group with hardness are shown. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0030] When ferromagnetic materials are alternating magnetized, irreversible domain wall motion within the material releases discrete electromagnetic pulses. The presence of stress or hardness significantly affects the pinning and depinning processes of these domain walls. Existing research has largely focused on the total energy (RMS) of macroscopic continuous signals, while the physical principle of this invention lies in treating these discontinuous jumps as independent "microscopic avalanches." By analyzing the scale (size) of each avalanche and the driving force (tangential magnetic field) required to trigger it, the joint probability distribution of the pinning strength within the material can be reconstructed.

[0031] Based on the aforementioned physical principles, the specific implementation of the method provided by this invention is as follows, see below. Figure 1 :

[0032] Step (1): Obtain the raw signal data of the magnetic Barkhausen noise of the ferromagnetic material under test under alternating magnetic field excitation and the synchronous tangential magnetic field signal data.

[0033] This specific application example uses a calibration test block of GCr15 bearing steel, a common material in engineering, as an example. The acquisition system uses a low-frequency sinusoidal signal (typically below 200Hz) as the excitation signal to drive a U-shaped magnetic yoke to AC magnetize the test piece. Simultaneously, it acquires the original high-frequency MBN signal picked up by the receiving coil on the test piece surface. and the reference tangential magnetic field signal after low-pass filtering In this embodiment, 11 sets of sample datasets covering a hardness gradient from 23.0 HRC to 62.8 HRC were selected, and each set of samples contains multi-period data.

[0034] Step (2): Discretization extraction of MBN avalanche events.

[0035] In order to start from continuous Extracting the actual physical avalanche events from the data, and then... Find the absolute envelope after removing DC. And establish an adaptive detection threshold. Adaptive threshold Set as a multiple of the root mean square (RMS) of the background signal, i.e. In this embodiment, the preferred option is... This threshold effectively filters out instrument background noise and minor fluctuations caused by reversible magnetization. Using a differential algorithm, when the signal envelope continuously crosses and eventually falls below [a certain threshold], [the threshold is determined]. The time segment was identified as an independent microdomain avalanche event. .

[0036] Refer to the instruction manual. Figure 2 Provide a detailed explanation: such as Figure 2 The diagram shown is a close-up illustration of the MBN raw signal preprocessing and discrete micro avalanche event extraction in this invention. The upper part of the diagram shows the synchronously acquired low-frequency tangential magnetic field signal, and the lower part shows the MBN envelope voltage signal obtained after full-wave rectification. When the signal envelope continuously crosses and eventually falls below the adaptive threshold... The time segments (thick black dashed lines in the diagram) are strictly identified as independent micro-avalanche events. For example... Figure 2 As shown in the magnified picture-in-picture example, the avalanche size of a single micro-avalanche event is defined as the area of ​​voltage integration (gray shaded region) of the portion of the envelope signal above the threshold within that segment. Simultaneously, the instantaneous value of the synchronous tangential magnetic field at the center moment of that segment is extracted as the driving force feature. This method accurately deconstructs continuous macroscopic noise into discrete microscopic dynamic events.

[0037] Step (3): Calculation and mapping of feature parameter pairs.

[0038] For extracting independent avalanche events Calculate its core physical characteristics:

[0039] (1) Avalanche size The energy released during this domain jump is represented by integrating the effective voltage above a threshold within the time segment. Since the distribution of avalanche size typically follows a power-law distribution with an extremely wide dynamic range, the obtained... Perform logarithmic scaling .

[0040] (2) Instantaneous tangential magnetic field : Represents the external magnetic field driving force required to trigger this jump. Extract the center index position of the event start and end times. Mapping to obtain synchronously acquired tangential magnetic field data .

[0041] Step (4): Construction of three-dimensional topological statistical manifold surface.

[0042] Obtain discrete After scattering the data, transform it into a continuous statistical space. and Divide on the two-dimensional plane The grid is sized (preferably 40×40 in this embodiment). The frequency of events falling into each grid (m,n) is calculated using a two-dimensional histogram. To highlight local features and enhance contrast, a logarithmic feature density transformation is performed. Finally, in order to eliminate the sawtooth noise caused by discrete statistics, [the following is done]: The matrix is ​​processed by a two-dimensional Gaussian filter to generate a smooth and topologically meaningful three-dimensional manifold surface height matrix.

[0043] Step (5): Topological feature evolution analysis and hardness interpretation.

[0044] This step overcomes the limitations of traditional methods that rely on finding signal amplitude for calibration, and provides three visualization analysis and quantitative evaluation methods based on "topological geometric deformation and offset":

[0045] Method 1: 3D Topological Terrain Feature Determination. (Reference) Figure 3 and Figure 4 The outputs are manifold diagrams for hardnesses of 23.0 HRC (soft state) and 62.8 HRC (hard state), respectively. Figure 3 In soft materials, where there are no strong pinning points, avalanches are prone to concentrated eruption under low electric field strength, forming a "compact bimodal structure with a close center and small spacing," as shown in the figure; while in... Figure 4 In hard materials, due to the presence of numerous dislocations and carbide pinnings, the domain wall motion barrier increases sharply, leading to a topological "phase transition" that manifests as a "bimodal crater structure" with dramatic expansion along the tangential magnetic field axis, and a reduction in the number of large-scale avalanches (the Y-axis centroid sinks). The heat treatment state of the material can be qualitatively or semi-quantitatively determined by extracting the "bimodal splitting degree" of the surface.

[0046] Method 2: Precise calibration based on the trajectory of the peak shift stacking evolution map. (Reference) Figure 5 The three-dimensional logarithmic characteristic density matrix of the standard test blocks with different hardness in step (4) is summed and integrated along the logarithmic dimension axis (Y-axis) to obtain the total avalanche activity curve that varies purely with the tangential magnetic field. Using the polygon occlusion algorithm in computer graphics, the curves of multiple sets of samples are stacked with equal spacing along the vertical axis from low hardness to high hardness. Figure 5 A clear "offset trajectory line" formed by connecting the peaks is visible in the image (shown by the dashed arrow). As hardness increases, the material's coercivity increases, causing the peaks to drift laterally with extremely high linearity. In actual testing, the coordinates of the peaks of the workpiece under test are captured and substituted into this calibration trajectory to accurately calculate its actual HRC value. Because it uses magnetic field coordinates rather than voltage amplitude coordinates, this method is highly immune to sensor lift-off changes.

[0047] Method 3: Classification and recognition based on 2D contour fingerprint maps. (Reference) Figure 6Two-dimensional contour lines are extracted from specific logarithmic feature density levels (e.g., 65% of the maximum density) of the three-dimensional manifold. The hotspot distributions of low hardness (solid lines), medium hardness (dashed lines), and high hardness (dotted lines) exhibit non-overlapping closed contour features. Specifically, as material hardness increases, the span of this closed contour significantly widens along the tangential magnetic field axis, the centroid of the contour shifts outward towards the high magnetic field regions on both sides, and the overall distribution centroid shifts downward along the avalanche size axis. In engineering applications, this can be used as a "digital fingerprint" of the material state, input into modules supporting linear regression or image pattern recognition, enabling rapid and high-tolerance automatic sorting of the hardness of large batches of bearings and other parts without complex modeling.

[0048] Refer to the instruction manual. Figure 7 Detailed verification and explanation: To improve the robustness of the evaluation and verify the accuracy of the peak shift calibration model, this embodiment uses a wide-window moving average algorithm to smooth the envelope energy of the single-cycle MBN signal, extracting the "macro-scale avalanche active main peak" representing the concentrated outbreak of the mass avalanche and its corresponding tangential magnetic field. .like Figure 7 The error bars and correlation calibration diagrams shown indicate that, through large-scale multi-period statistical analysis, as material hardness increases, the density of internal pinning points increases, leading to a higher domain wall motion barrier, and the macroscopic avalanche eruption center strictly shifts towards the direction of high magnetic field. Statistical analysis verifies that there is a very strong, strictly monotonically positive correlation between the main peak tangential magnetic field and Rockwell hardness, with a Spearman rank correlation coefficient as high as 0.9545 and a Pearson correlation coefficient reaching 0.8333. The extremely high rank correlation coefficients demonstrate the excellent monotonicity of this evolution trajectory, and the calibration model constructed based on this characteristic ( Figure 7 (Medium-thick dashed line) can achieve extremely stable and accurate quantitative assessment of the hardness state of ferromagnetic materials.

[0049] On the other hand, specific embodiments are used to illustrate the principles and implementation methods of this invention. Those skilled in the art should understand that the embodiments of this invention can be provided as a set of computer program-based software products, electronic devices, or testing systems. The steps in this invention, such as MBN avalanche feature extraction, logarithmic mapping, Gaussian smoothing, and visualization, can be entirely implemented by computer program instructions and stored in a computer-readable storage medium. The above description of the embodiments is only for helping to understand the method and core ideas of this invention; furthermore, there may be changes in specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. A method for evaluating the hardness of ferromagnetic materials based on three-dimensional visualization of magnetic Barkhausen noise avalanche, characterized in that... Includes the following steps: (1) Acquire the raw signal data of the magnetic Barkhausen noise of the ferromagnetic material under test under alternating magnetic field excitation and the synchronous tangential magnetic field signal data; (2) The original magnetic Barkhausen noise signal is processed, an adaptive noise threshold is set, and the continuous signal is segmented and several discrete magnetic Barkhausen noise avalanche events are extracted. (3) Extract the feature parameter pairs for each avalanche event. , ),in This refers to the instantaneous tangential magnetic field strength data corresponding to the moment at which the event occurred. This refers to the avalanche size data for this event; (4) Divide the area into a grid on the two-dimensional feature plane formed by the tangential magnetic field and the avalanche size, count the number of avalanche events falling into each grid to calculate the logarithmic feature density, and use a Gaussian smoothing algorithm to construct a grid with the tangential magnetic field as the reference. X-axis, avalanche size A three-dimensional statistical manifold surface with Y-axis and logarithmic characteristic density on Z-axis; (5) Based on the topological morphological features of the three-dimensional statistical manifold surface, a visualization graphic output is generated, the peak features or spatial distribution features of the three-dimensional statistical manifold surface along the tangential magnetic field axis are extracted, and the results are substituted into the pre-established calibration model to calculate and evaluate the quantitative hardness state of the tested ferromagnetic material.

2. The method according to claim 1, characterized in that, The adaptive noise threshold mentioned in step (2) The calculation formula is: ,in The root mean square value of the magnetic Barkhausen noise signal after DC removal is given. The threshold coefficient is preferably in the range of 1.5 to 3.

0.

3. The method according to claim 1, characterized in that, The avalanche size mentioned in step (3) The calculation method is to integrate the effective voltage over the duration of a single avalanche event, as shown in the following formula: in and These represent the start and end times of the avalanche event crossing the adaptive noise threshold, respectively. The signal is a magnetic Barkhausen noise signal, and the avalanche size is determined in step (4). Perform logarithmic scaling .

4. The method according to claim 1, characterized in that, The construction process of the three-dimensional statistical manifold surface described in step (4) specifically includes: exist Divide the plane into a uniform grid and count the number of cells falling into each grid. Number of events within Logarithmic transformation of the number of events ;right The matrix is ​​smoothed by two-dimensional Gaussian filtering to obtain a continuous logarithmic characteristic density matrix Z, and a three-dimensional mesh surface is generated.

5. The method according to claim 1, characterized in that, The evaluation of the hardness of the tested ferromagnetic material based on the topological morphology characteristics of the three-dimensional statistical manifold surface in step (5) specifically includes: if the three-dimensional statistical manifold surface presents a narrow-spacing double-peak shape that approaches the central low magnetic field region, then the hardness of the tested ferromagnetic material is determined to be low; if the magnetic field spacing between the double peaks of the three-dimensional statistical manifold surface increases and the distribution centroid shifts towards the high magnetic field direction, presenting a wide-spacing double-peak shape, then the hardness of the tested ferromagnetic material is determined to be high.

6. The method according to claim 1 or 5, characterized in that, Step (5) also includes the step of constructing a peak shift stack evolution diagram: The logarithmic characteristic density of the three-dimensional statistical manifold surface is integrally projected along the avalanche size axis to obtain the magnetic field-density evolution curve. The magnetic field-density evolution curves of multiple standard test blocks with known hardness are plotted by equidistant translation along the vertical axis, and the coordinates of the peak points of each curve are extracted to generate a calibration model of the offset trajectory of the peak on the tangential magnetic field axis. The evolution curve of the material to be tested is substituted into the calibration model to obtain its quantitative hardness value.

7. The method according to claim 6, characterized in that, The step of extracting the peak coordinates of each curve specifically includes: using a moving average algorithm to smooth and filter the magnetic field-density evolution curve to filter out local extreme value interference caused by microscopic random jumps, extracting the global maximum point on the curve that represents the concentrated outbreak of the avalanche, and recording the tangential magnetic field corresponding to the global maximum point as the peak coordinates.

8. The method according to claim 1, characterized in that, Step (5) also includes the step of constructing a two-dimensional topological contour fingerprint: extracting the feature contours of the three-dimensional statistical manifold surface, and projecting the contours of different hardness states onto the same two-dimensional plane. By utilizing the evolution characteristics of the tangential magnetic field span widening, the center of gravity expanding outward to the high magnetic field region, and the avalanche size center of gravity shifting downward as the closed contour increases with hardness, the material state classification and evaluation can be achieved.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the method according to any one of claims 1 to 8.

Images