Ferromagnetic material internal defect leakage magnetic field increasing method based on alternating current and direct current composite magnetization

Abstract

The invention discloses a ferromagnetic material internal defect leakage magnetic field increasing method based on alternating current and direct current composite magnetization, which comprises the following steps of: fixing a workpiece at the center of a composite exciting coil, and synchronously applying direct current and high-frequency alternating current; after pure direct current magnetization is saturated, current is locked, and a basic leakage magnetic field is established; the surface magnetic conductivity is dynamically reduced by overlapping alternating current, and the magnetic shielding effect is weakened; a leakage magnetic field signal is scanned and collected through a differential Hall sensor array, and background magnetic field deduction, peak value comparison and adaptive filtering processing are carried out; and determining an optimal magnetization parameter combination by adopting a gradient descent method and Bayesian optimization. The system is integrated with a composite excitation coil and a signal conditioning unit, so that synchronous excitation, signal acquisition and real-time analysis are realized. The leakage magnetic field signal at the deep defect is enhanced through alternating-current and direct-current composite magnetization, the problem of signal weakening caused by traditional direct-current saturation is solved, the defect detection sensitivity and efficiency are improved, and the method is suitable for non-destructive detection scenes of ferromagnetic materials.

Description

Technical Field

[0001] This invention relates to the field of nondestructive testing, and more particularly to a method for increasing the leakage magnetic field of internal defects in ferromagnetic materials based on AC / DC composite magnetization. Background Technology

[0002] Existing technologies face a dual bottleneck in enhancing the leakage magnetic field of internal defects in ferromagnetic materials: 1. The contradiction of saturation suppression: The DC magnetic field needs to reach magnetic saturation to excite the maximum fundamental leakage magnetic signal, but strong saturation will significantly suppress the permeability perturbation Δ at internal defects. 1. This leads to weakening of internal defect signals; 2. Magnetic shielding effect: The high μ surface layer forms a "magnetic short circuit" effect, and the leakage magnetic field of internal defects with a depth of >5mm is attenuated by more than 60%.

[0003] Limitations of existing enhancement technologies: 1. Patent CN211603029U discloses a magnetic field amplification device for improving the sensitivity of magnetic flux leakage detection. It uses a permanent magnet + electromagnetic coil superimposed magnetic field. Although the field strength can be adjusted, its DC magnetic field still operates in the saturation region, and the Δ 1. The problem of attenuation exists, and no AC desaturation mechanism is designed; 2. Patent CN201810182149.4 discloses an internal flaw detection circuit for ferromagnetic materials based on low-frequency leakage magnetic flux, which improves the penetration depth by using 20Hz low-frequency excitation, but the lowest frequency band (20Hz) still cannot avoid the magnetic shielding effect; 3. Patent CN115189494A discloses a centrifugal variable flux permanent magnet motor and a method for increasing or decreasing the leakage magnetic flux of the motor permanent magnet, which adjusts the magnetic resistance of the leakage magnetic circuit through mechanical structure, but this solution is only applicable to the permanent magnet structure of the motor and cannot be transferred to the defect detection scenario of ferromagnetic materials. 4. Patent CN102590328A discloses a method for detecting leakage magnetic field by combining permanent magnet and AC / DC. Although it also uses AC / DC composite magnetization, it is a time-division excitation mechanism with dual defects: (1) Efficiency loss: the step-by-step scanning of DC followed by AC leads to repeated detection of the same area, reducing efficiency by ≥50%; (2) Signal attenuation: in the DC stage, the permanent magnet and electromagnetic coil work together to reach the optimal operating point, but in the AC stage, the DC component of the electromagnetic coil is removed, leaving only the permanent magnet bias, which weakens the total DC field strength → shifts the operating point → attenuates the leakage magnetic field of deep defects. The AC magnetic field is limited by the skin effect and can only locate defects, but cannot maintain or enhance deep signals. Summary of the Invention

[0004] This invention overcomes the shortcomings of the prior art and provides a method for increasing the leakage magnetic field of internal defects in ferromagnetic materials based on AC / DC composite magnetization.

[0005] To achieve the above objectives, the technical solution adopted by this invention is as follows: a method for increasing the leakage magnetic field of internal defects in ferromagnetic materials based on AC / DC composite magnetization, comprising the following steps:

[0006] S1: Fix the ferromagnetic workpiece to be inspected to the center of the composite excitation coil using a bracket, connect the coil to the DC power supply and the high-frequency AC power supply, and complete the system initialization and workpiece positioning.

[0007] S2: In the pure DC magnetization mode, the DC power supply is turned on, and the current is gradually increased until the amplitude of the Hall sensor array output signal no longer increases. The DC current value is locked to achieve saturated DC magnetization lock and establish the maximum basic leakage magnetic field.

[0008] S3: The AC / DC composite magnetization mode is to apply a high-frequency AC current simultaneously on the basis of locking the DC current, so that the magnetic permeability of the workpiece surface layer decreases periodically and the magnetic shielding layer is dynamically weakened.

[0009] S4: A linear Hall sensor is used to scan the workpiece surface at a speed of 5~10mm / s and collect leakage magnetic field signals under the action of AC and DC composite fields in real time.

[0010] S5: Perform background magnetic field subtraction and peak value comparison processing on the acquired signal, eliminate background magnetic field interference by using the two-sample difference method, plot the leakage magnetic field peak value as a function of magnetization current or AC parameters, determine the optimal magnetization conditions, and verify the signal gain effect.

[0011] In a preferred embodiment of the present invention, the composite excitation coil is wound with multi-strand enameled copper wire, the coil type is a composite excitation coil assembly, the DC current value ranges from 0 to 30A, the frequency range of the high-frequency AC current is 500 to 1500Hz, and the amplitude is 1 to 6A.

[0012] In a preferred embodiment of the present invention, the Hall sensor array adopts a differential pair structure, with two Hall sensors vertically mounted 8-10 mm apart, and geomagnetic field interference is eliminated by a differential amplifier; the sensor output signal is sampled by a 24-bit ADC with a sampling rate set to 10 kHz, and power frequency interference and high-frequency noise are filtered out by a digital filter; the original signal and the background magnetic field signal are subtracted in real time by an FPGA, and the net leakage magnetic field data is sent to the host computer for storage via a wireless transmission module.

[0013] In a preferred embodiment of the present invention, the background magnetic field subtraction adopts dual Hall sensor dynamic calibration technology, which automatically corrects the sensor sensitivity difference through correlation analysis of the two sensor signals; an adaptive filtering algorithm is introduced in the signal conditioning stage, which dynamically adjusts the filtering parameters according to the real-time noise level to ensure that the effective signal retention rate is not less than 95%; the optimal magnetization current is determined by gradient descent method, which locates the current value corresponding to the maximum gradient value by iteratively calculating the second derivative of the leakage magnetic field signal amplitude and the current value.

[0014] In a preferred embodiment of the present invention, the original signal is derived in pure DC magnetization mode and the background magnetic field is subtracted to obtain the reference leakage magnetic field; in AC-DC composite magnetization mode, the original signal is derived by adding the optimal magnetization current and the background magnetic field is subtracted to obtain the enhanced leakage magnetic field; by comparing the peak value changes of the reference and enhanced leakage magnetic fields, the optimal parameter combination is determined.

[0015] In a preferred embodiment of the present invention, the determination of the optimal parameter combination is based on the curve of the leakage magnetic field peak value as a function of magnetizing current, AC amplitude and frequency.

[0016] In a preferred embodiment of the present invention, a magnetization system based on a method for increasing the leakage magnetic field of internal defects in ferromagnetic materials using AC / DC composite magnetization is proposed. The method for increasing the leakage magnetic field of internal defects in ferromagnetic materials includes: a composite excitation coil assembly and a signal conditioning unit; the coil assembly is simultaneously supplied with DC current and high-frequency AC current to achieve synchronous excitation; the signal conditioning unit performs background magnetic field subtraction and peak value comparison analysis.

[0017] In a preferred embodiment of the present invention, the signal conditioning unit integrates a Hall sensor array and a data processing module; the sensor array is linearly arranged with a lift-off value of 0.4~0.6mm and a scanning speed of 5~10mm / s; the data processing module performs raw signal acquisition, background magnetic field subtraction, peak extraction and curve fitting.

[0018] This invention addresses the shortcomings of the prior art and has the following beneficial effects:

[0019] (1) The composite excitation coil is simultaneously supplied with DC current and high-frequency AC current. Combined with the pure DC magnetization mode, the saturation DC current value is locked. The current is gradually increased until the amplitude of the Hall sensor output signal no longer increases, thus locking the maximum basic leakage magnetic field. The high-frequency AC current is applied simultaneously to make the permeability of the workpiece surface decrease periodically. While suppressing the magnetic shielding effect, the permeability disturbance Δ at the internal defect is selectively retained. Thus, under the background of maintaining the maximum basic leakage magnetic field, the leakage magnetic field signal at the internal defect is effectively enhanced, which solves the contradiction between signal weakening caused by strong saturation and coexistence of magnetic shielding in traditional DC saturation magnetization.

[0020] (2) The magnetic field generated by the high frequency alternating current periodically changes the permeability of the workpiece surface. Combined with the differential pair structure of the linear Hall sensor array, the alternating magnetic field dynamically reduces the permeability of the high μ surface layer, reducing the "magnetic short circuit" effect. The differential pair structure eliminates common-mode interference such as the geomagnetic field through differential processing of the signals from two Hall sensors, thus realizing effective detection of deep defects.

[0021] (3) The dynamic parameter optimization algorithm adopts Bayesian optimization and combines gradient descent method to determine the optimal magnetization conditions. By iteratively calculating the second derivative of the leakage magnetic field signal amplitude and the current value, the current value corresponding to the maximum gradient point is located. Parameters such as DC current, AC frequency, and scanning speed are dynamically adjusted. The optimization objective function is a weighted sum of maximizing signal gain and minimizing detection time. The detection efficiency is significantly improved, and the problems of repeated detection and signal attenuation caused by time-division excitation are avoided. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a flowchart of a preferred embodiment of the present invention for detecting the increase of leakage magnetic field in internal defects of ferromagnetic materials;

[0024] Figure 2 This is a schematic diagram of a magnetization system according to a preferred embodiment of the present invention;

[0025] Figure 3 This is a data graph of DC magnetization minus the background magnetic field according to a preferred embodiment of the present invention;

[0026] Figure 4 This is a peak leakage magnetic field diagram under pure DC excitation and different magnetization currents in a preferred embodiment of the present invention.

[0027] Figure 5 These are different amplitude diagrams of the original AC / DC composite magnetization data after subtracting the background magnetic field, according to a preferred embodiment of the present invention.

[0028] Figure 6 These are different frequency diagrams of the original AC / DC composite magnetization data after subtracting the background magnetic field, according to a preferred embodiment of the present invention.

[0029] Figure 7 This is a graph showing the variation of the peak value of the original AC / DC composite magnetization data with AC amplitude and frequency according to a preferred embodiment of the present invention.

[0030] In the figure: 101, composite excitation coil assembly; 104, workpiece to be inspected; 105, defect; 106, Hall sensor array; 108, DC magnetization region; 109, AC / DC magnetization region. Detailed Implementation

[0031] 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. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein. Therefore, the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0033] In the description of this application, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as limiting the scope of protection of this application. Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, features defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0034] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art will understand the specific meaning of the above terms in this application based on the specific circumstances.

[0035] like Figure 1 As shown, the method for increasing the leakage magnetic field of internal defects in ferromagnetic materials based on AC / DC composite magnetization includes the following steps:

[0036] S1: Fix the ferromagnetic workpiece to be inspected to the center of the composite excitation coil using a bracket, connect the coil to a DC power supply and a high-frequency AC power supply, and complete system initialization and workpiece positioning.

[0037] Furthermore, the composite excitation coil is wound with multi-strand enameled copper wire, and the coil type is a composite excitation coil assembly. The DC current value ranges from 0 to 30A, the frequency range of the high-frequency AC current is 500 to 1500Hz, and the amplitude is 1 to 6A.

[0038] In another embodiment, a PT100 platinum resistance temperature sensor is embedded in the coil surface. When the temperature exceeds 85°C, a PID control algorithm is triggered to reduce the DC current amplitude, ensuring the operating temperature remains stable between 60 and 85°C. A Keyence LK-H050 laser displacement sensor is used to calibrate the core-workpiece distance. The calibration process includes a reference distance acquisition stage and a dynamic deviation compensation verification stage. The compensation formula for the dynamic deviation compensation verification stage is: .

[0039] The dual-power redundancy design uses Schneider NSX400N circuit breakers, with a main and backup power voltage difference of ≤2% and a switching time of ≤20ms. During the switching period, power is maintained by a supercapacitor bank. The material identification algorithm collects the reflection spectrum through an Ocean Optics USB4000 spectrometer, filters it with Savitzky-Golay, and inputs it into the SVM model. It matches 12 material libraries to form a complete closed-loop evidence chain of temperature-permeability mapping, spacing calibration log, power switching record, and material identification report.

[0040] S2: In the pure DC magnetization mode, the DC power supply is turned on, and the current is gradually increased until the amplitude of the Hall sensor array output signal no longer increases. The DC current value is locked to achieve saturated DC magnetization lock and establish the maximum basic leakage magnetic field.

[0041] Furthermore, the linear Hall sensor array adopts a differential pair structure, with two Hall sensors vertically mounted 8-10mm apart. The differential amplifier eliminates geomagnetic field interference. The sensor output signal is sampled by a 24-bit ADC with a sampling rate set to 10kHz. A digital filter removes power frequency interference and high-frequency noise. The original signal and the background magnetic field signal are subtracted in real time by an FPGA. The net leakage magnetic field data is sent to the host computer for storage via a wireless transmission module.

[0042] In another embodiment, a three-dimensional magnetoresistive sensor reconstruction algorithm and a deep learning defect recognition system are extended. An HMC1053L magnetoresistive sensor with a sensitivity of 0.7 mV / Gs and a range of ±6 Gs is used. An 8x8 sensor array with a spacing of 5 mm is constructed. The three-dimensional leakage magnetic field distribution is reconstructed iteratively using gradient descent, with the iteration termination condition being RMSE < 0.001 Gs. The deep learning model adopts an improved U-Net architecture, taking a 64x64 pixel grayscale image of the leakage magnetic field as input, with 3x3 convolutional kernels, a stride of 1, and a ReLU activation function. The training dataset contains 10,000 simulated defect images and 5,000 measured images, labeled with bounding boxes and category labels. The model is trained using the Adam optimizer, with a test set accuracy ≥ 95%; otherwise, the network depth or the number of convolutional kernels is adjusted. System components include an NI USB-6361 data acquisition card, an NVIDIA Jetson TX2 edge computing module, and LabVIEW 2021 visualization software. The operation process covers data preprocessing, model training, verification deployment, and real-time defect recognition output, forming a complete detection closed loop.

[0043] S3: The AC / DC co-magnetization mode is to apply a high-frequency AC current synchronously on the basis of a locked DC current, so that the magnetic permeability of the workpiece surface layer decreases periodically, and the magnetic shielding layer is dynamically weakened.

[0044] S4: A linear Hall sensor is used to scan the workpiece surface at a speed of 5~10mm / s to collect leakage magnetic field signals under the action of AC and DC combined fields in real time.

[0045] Furthermore, the AC / DC composite magnetization mode achieves synergistic enhancement of surface and deep defect signals in ferromagnetic materials by simultaneously applying DC and high-frequency AC magnetic fields. Its physical essence lies in overcoming the suppression effect of traditional DC saturation magnetization and the skin effect limitations of high-frequency AC magnetization. The DC magnetic field establishes the basic leakage magnetic field through Ampere's circuital law. When the current increases to 25A, the Hall sensor output signal saturates, indicating that the material has reached magnetic saturation. At this point, the permeability perturbation Δ at the internal defects... Because the strong DC field is suppressed, the high-frequency alternating magnetic field (500-1500Hz) induces periodic changes in permeability on the workpiece surface through Faraday's law of electromagnetic induction, causing the surface permeability μ to decrease periodically. This dynamic modulation weakens the surface "magnetic short circuit" effect and reduces the attenuation of the deep leakage magnetic field during conduction. Mathematical derivation shows that the μ change caused by the alternating magnetic field satisfies... ,in Let μ be the modulation depth and f be the AC frequency. As μ decreases, the magnetoresistance of the leakage magnetic field generated by deep defects increases during conduction at the surface, which in turn enhances the leakage magnetic field signal acquired by the sensor. Experimental data show that when detecting a 20mm thick steel plate with a sensor on one side and a defect on the other, the defect signal gain under composite magnetization reaches 26.47%; under the same arrangement conditions on a 10mm thick steel plate, the defect signal gain reaches 47.2%, verifying the physical effectiveness of this mode. This mode enhances the deep defect signal by dynamically adjusting the surface permeability, solving the problem of difficult deep defect detection in traditional methods.

[0046] The Hall sensor array employs a differential pair structure, consisting of two vertically mounted linear Hall elements spaced 8-10 mm apart. Its operating principle is based on the Hall effect: when a current I passes through a semiconductor placed in a magnetic field B, a potential difference is generated perpendicular to both the current and the magnetic field. The differential pair eliminates common-mode interference such as that from the Earth's magnetic field by subtracting the signals from the two sensors. For example, when the Earth's magnetic field... When present, both sensors output After difference Effectively inhibited The anti-interference design also includes 24-bit ADC sampling (sampling rate 10kHz) and digital filters. The former reduces quantization error through high-resolution quantization, while the latter eliminates 50Hz power frequency interference and high-frequency noise through bandpass filtering (0-2kHz). An FPGA-implemented real-time background magnetic field subtraction algorithm subtracts the original signal from the background magnetic field signal calibrated by the dual Hall sensors, further improving signal purity. Experiments show that this design achieves an effective signal retention rate of no less than 95%. The differential pair structure improves signal detection accuracy by eliminating common-mode interference, while the high-resolution ADC and digital filters ensure the quality of signal acquisition.

[0047] S5: Perform background magnetic field subtraction and peak value comparison processing on the acquired signal, eliminate background magnetic field interference by using the two-sample difference method, plot the leakage magnetic field peak value as a function of magnetization current or AC parameters, determine the optimal magnetization conditions, and verify the signal gain effect.

[0048] The two-sample difference method eliminates background magnetic field interference by comparing leakage magnetic field signals in pure DC magnetization mode and AC / DC composite magnetization mode. The specific procedure is as follows: In pure DC mode, the leakage magnetic field signal on the workpiece surface is acquired. The data is collected and stored as a background magnetic field reference; in the composite magnetization mode, the signal is acquired. Real-time subtraction achieved through FPGA The net leakage magnetic field signal is obtained. Mathematical derivation shows that this method is equivalent to frequency domain filtering, i.e. Where f is the frequency component. Since the background magnetic field mainly contains low-frequency components (such as the Earth's magnetic field), while the defect signal contains high-frequency components (such as harmonics generated by AC modulation), low-frequency interference is suppressed after differential processing, while the high-frequency defect signal is preserved. Experimental data show that this method improves the signal-to-noise ratio by 12 dB and increases the defect detection rate to 95%. The two-sample differential method improves the accuracy of defect detection by eliminating background magnetic field interference, solving the problem of background interference affecting detection results in traditional methods.

[0049] Furthermore, the background magnetic field subtraction adopts dual Hall sensor dynamic calibration technology, which automatically corrects the sensor sensitivity difference through correlation analysis of the two sensor signals; an adaptive filtering algorithm is introduced in the signal conditioning stage, which dynamically adjusts the filtering parameters according to the real-time noise level to ensure that the effective signal retention rate is not less than 95%; the optimal magnetization current is determined by the gradient descent method, which iteratively calculates the second derivative of the leakage magnetic field signal amplitude and the current value to locate the current value corresponding to the maximum gradient point.

[0050] Adaptive filtering algorithms dynamically adjust filter parameters by monitoring noise levels in real time. Their core is the Least Mean Square Error (LMS) algorithm, and the filter coefficient update rule is as follows: ,in For the current coefficient, Step size factor This is the error signal (the difference between the desired signal and the filtered output). Input signal. Step size factor. The noise power is dynamically adjusted; when the noise power increases... Reduce the noise power to decrease the convergence speed but improve stability; when the noise power is reduced... Increase the value to accelerate convergence. Experiments show that this mechanism enables the filter to maintain an effective signal retention rate of no less than 95% when the noise power changes by ±20dB. The adaptive filtering algorithm improves the robustness of signal processing by dynamically adjusting the filtering parameters to adapt to different noise environments, and solves the problem that fixed filtering parameters in traditional methods cannot adapt to environmental changes.

[0051] In pure DC magnetization mode, the original signal is extracted and the background magnetic field is subtracted to obtain the reference leakage magnetic field; in AC / DC composite magnetization mode, the original signal is extracted and the background magnetic field is subtracted to obtain the enhanced leakage magnetic field by adding the optimal magnetization current and AC current; by comparing the peak changes of the reference and enhanced leakage magnetic fields, the optimal parameter combination is determined.

[0052] The optimal parameter combination is determined by the curve of the leakage magnetic field peak value as a function of magnetizing current, AC amplitude and frequency.

[0053] In another embodiment, a dynamic parameter optimization algorithm is introduced. The objective function is set as a weighted sum of maximizing signal gain and minimizing detection time, with adjustable weight coefficients. Bayesian optimization uses a Gaussian process regression model, and the search space includes DC current (0~30A), AC frequency (500~1500Hz), and scanning speed (5~10mm / s). The optimization process includes an initialization phase (randomly sampling 20 sets of parameters), an iterative optimization phase (selecting the optimal parameters each time), and a convergence judgment (gain <1% for 3 consecutive iterations). A real-time monitoring module collects parameters such as leakage magnetic field signal-to-noise ratio, detection time, and temperature as optimization feedback. A user-defined module supports setting priority targets (accuracy priority or speed priority) and adjusting optimization weights. Optimization results are stored in a parameter database, supporting historical optimal parameter retrieval. The system supports online optimization, automatically updating the optimization model after detecting 10 workpieces. Algorithm parameters include a population size of 50, number of iterations of 100, crossover probability of 0.8, mutation probability of 0.1 (genetic algorithm part), inertia weight of 0.7, individual learning factor of 1.5, and social learning factor of 1.5 (particle swarm optimization part). Objective function: The constraints are DC current ≤30A, AC frequency 500~1500Hz, and AC amplitude 1~6A. The edge computing module uses the NVIDIA Jetson AGX Xavier algorithm, and the host computer software is developed and verified using MATLAB R2021b, forming a complete dynamic parameter optimization closed loop.

[0054] The dynamic parameter optimization algorithm aims to maximize signal gain and minimize detection time, and constructs a multi-objective optimization function. ,in , These are the weighting coefficients. and The peak value of the leakage magnetic field before and after composite magnetization is given, and T is the detection time. A Bayesian optimization framework is used, and a Gaussian process regression model is employed to establish the mapping relationship between the parameter space (DC current 0~30A, AC frequency 500~1500Hz, scanning speed 5~10mm / s) and the objective function. The optimization process is divided into an initialization phase (randomly sampling 20 sets of parameters) and an iterative optimization phase (each time selecting the parameter with the highest posterior probability for evaluation). Gradient descent is used to accurately locate the optimal parameters by calculating the second derivative of the objective function with respect to the DC current I, thus locating the value corresponding to the maximum gradient point. For example, when... The iteration terminates when the current I reaches the optimal magnetization current. Experiments show that this algorithm improves detection efficiency by 60% and achieves a signal gain of 47.2%. The dynamic parameter optimization algorithm achieves a balance between detection efficiency and accuracy by automatically adjusting parameters, solving the problem of parameter setting relying on experience in traditional methods.

[0055] like Figure 2As shown, the detection system consists of a composite excitation coil assembly 101, a workpiece to be inspected 104, a defect 105, and a Hall sensor array 106. The composite excitation coil has an inner diameter of 200mm, is wound with multi-strand enameled copper wire, has 1000 turns, and achieves stepless adjustment (adjustment range 0~50mm) relative to the workpiece surface via a hydraulic support. The Hall sensor array adopts an 8*8 layout with a spacing of 5mm and a sensitivity of 13mV / T, synchronously scanning the workpiece surface along with the coil.

[0056] By using high-frequency alternating current to generate a magnetic field that tends to the DC magnetization region 108 on the workpiece surface under DC excitation, the μ of the AC / DC magnetization region 109 on the workpiece surface is periodically reduced. This has almost no effect on the leakage magnetic field generated by the defect. Although it reduces the leakage magnetic field caused by permeability disturbance, the leakage magnetic field caused by magnetic shielding is reduced even more, thus obtaining a stronger leakage magnetic field for sensor acquisition.

[0057] Key component parameters

[0058] DC power supply module: 0~30A adjustable, 0.5A step, accuracy ±0.1A.

[0059] High-frequency AC power supply module: adjustable from 500 to 1500 Hz, preferably 800 Hz; amplitude adjustable from 1 to 6 A, preferably 3 A.

[0060] Scanning system: driven by servo motor, speed adjustable from 5 to 10 mm / s, positioning accuracy ±0.1 mm.

[0061] Signal acquisition module: 24-bit ADC, sampling rate 10kHz, bandwidth 0~2kHz.

[0062] Data processing unit: NVIDIA Jetson TX2 edge computing module, running signal conditioning algorithms.

[0063] The detailed steps of the detection method include:

[0064] System initialization and workpiece positioning

[0065] The 10mm thick steel plate (300×200×10mm) to be inspected is fixed to the center of the composite excitation coil using a robotic arm. The coil is connected to a DC power supply and a high-frequency AC power supply, and the system parameters are initialized as follows: scanning speed 8mm / s, lift-off value 0.5mm, DC current limit 30A, AC frequency 800Hz, amplitude 3A.

[0066] Saturated DC magnetization lock

[0067] Turn on the DC power supply and gradually increase the current from 0A in 0.5A increments. Monitor the leakage magnetic field signal in real time using a Hall sensor array. When the current increases to 25A, the signal amplitude no longer increases with the current, indicating that the workpiece has reached magnetic saturation. Lock the DC current value at 25A and establish the maximum basic leakage magnetic field.

[0068] High-frequency AC modulation

[0069] A high-frequency alternating current of 800Hz and 3A is synchronously applied to the coil, generating a high-frequency magnetic field that periodically reduces the permeability of the workpiece surface. The distance between the coil and the workpiece is monitored in real time by a laser displacement sensor. When the deviation exceeds ±0.3mm, the height of the coil support is automatically adjusted by a stepper motor, with a compensation amount Δh = 0.5 × the measured deviation value.

[0070] Signal dynamic acquisition

[0071] A Hall sensor array is used to scan the workpiece surface at a speed of 8 mm / s to acquire leakage magnetic field signals under the combined AC and DC fields in real time. After the signals are digitized by a 24-bit ADC, the background magnetic field is subtracted in real time through an FPGA, and the net leakage magnetic field data is transmitted to a cloud database via a 5G module.

[0072] Signal Conditioning and Analysis

[0073] Multi-dimensional processing of the acquired raw signals:

[0074] Background magnetic field subtraction: A dual Hall sensor differential method is used, and the sensor sensitivity difference is automatically corrected through correlation coefficient analysis.

[0075] Spectrum analysis: The FFT transform is applied to analyze the frequency domain characteristics of the signal, filtering out 50Hz power frequency interference and high-frequency noise above 2kHz.

[0076] Peak detection: The peak value of the leakage magnetic field is extracted by the sliding window method, and the curve of the peak value as a function of magnetizing current / AC parameters is plotted.

[0077] Determining the optimal magnetization conditions: The second derivative of the leakage magnetic field signal amplitude and the current value is calculated using the gradient descent method, and the current value corresponding to the maximum gradient value is located.

[0078] Experimental Data and Results Analysis

[0079] like Figures 3-4 As shown, peak leakage magnetic field analysis under DC magnetization

[0080] Under DC magnetization, the peak value of the leakage magnetic field increases logarithmically with the increase of the magnetizing current. In the range of 0~25A, the peak value increases from 0.016T to 0.0546T, with an increase rate of 241.25%. When the current exceeds 25A, the signal increase slows down, verifying the accuracy of the determination of magnetic saturation state.

[0081] AC / DC composite magnetization effect verification

[0082] The signal enhancement effect of the defect signal is verified by comparing the signal amplitudes of DC magnetization and AC / DC composite magnetization of a single coil.

[0083] 20mm deep defect: The amplitude of a single DC magnetization signal is 0.016T, which increases to 0.0204T after composite magnetization, with a gain of 26.47%.

[0084] 10mm deep defect: The amplitude of a single DC magnetization signal is 0.0371T, which increases to 0.0546T after composite magnetization, with a gain of 47.2%.

[0085] AC parameter optimization experiment

[0086] like Figures 5-7 As shown, the influence of AC frequency and amplitude on the peak value of the leakage magnetic field is analyzed through orthogonal experiments:

[0087] Frequency effect: In the range of 500~1500Hz, the signal gain is optimal at 800Hz, which is 18.7% higher than that at 500Hz.

[0088] Amplitude effect: Within the range of 1~6A, the signal gain is optimal at 3A, which is 12.3% higher than that at 1A.

[0089] Parameter coupling effect: A peak prediction model was established using the response surface methodology, with a determination coefficient R² = 0.963, to verify the reliability of the model.

[0090] System self-test and fault diagnosis

[0091] Constructing a closed-loop self-inspection system includes:

[0092] Power system self-test: Real-time monitoring of DC / AC power supply output voltage fluctuations (threshold ±3%).

[0093] Sensor calibration and verification: Standard Helmholtz coils are used for periodic calibration, with an error requirement of <±1%.

[0094] Communication link detection: Monitor 5G / Wi-Fi 6 communication latency (threshold <100ms) through heartbeat packet mechanism.

[0095] Fault Diagnosis Expert System: Built-in 100 diagnostic rules, supporting remote expert consultation function.

[0096] This embodiment is applicable to the detection of internal defects in ferromagnetic components such as pressure vessels, pipelines, and bearings. Compared with traditional technologies, it has the following advantages:

[0097] Improved detection accuracy: Deep defect detection capability exceeds 20mm, signal gain reaches 47.2%.

[0098] Improved inspection efficiency: The inspection time for a single item is reduced to 5 minutes, which is 60% faster than traditional methods.

[0099] Non-contact inspection: Avoids damage to workpieces caused by mechanical contact and is suitable for inspection of in-service equipment.

[0100] Intelligent analysis: Integrates a deep learning defect identification module to achieve automatic classification of defect types.

[0101] This embodiment fully describes the structural composition, detection method, signal processing flow, and experimental verification results of a ferromagnetic material internal defect detection system based on AC / DC composite magnetization. Through a "saturated DC lock-in + high-frequency AC shield breaking" mechanism, it effectively solves the dual bottleneck problem in traditional detection technologies, achieving a significant enhancement of the leakage magnetic field signal and providing an innovative technical solution for the field of industrial non-destructive testing.

[0102] Based on the preferred embodiments of the present invention described above, those skilled in the art can make various changes and modifications without departing from the inventive concept. The technical scope of this invention is not limited to the contents of the specification, but must be determined according to the scope of the claims.

Claims

1. A method for increasing the leakage magnetic field of internal defects in ferromagnetic materials based on AC / DC composite magnetization, characterized in that, Includes the following steps: S1: Fix the ferromagnetic workpiece to be inspected to the center of the composite excitation coil using a bracket, connect the coil to the DC power supply and the high-frequency AC power supply, and complete the system initialization and workpiece positioning. S2: In the pure DC magnetization mode, the DC power supply is turned on, and the current is gradually increased until the amplitude of the Hall sensor array output signal no longer increases. The DC current value is locked to achieve saturated DC magnetization lock and establish the maximum basic leakage magnetic field. S3: The AC / DC composite magnetization mode is to apply a high-frequency AC current simultaneously on the basis of locking the DC current, so that the magnetic permeability of the workpiece surface layer decreases periodically and the magnetic shielding layer is dynamically weakened. S4: A linear Hall sensor is used to scan the workpiece surface at a speed of 5~10mm / s and collect leakage magnetic field signals under the action of AC and DC composite fields in real time. S5: Perform background magnetic field subtraction and peak value comparison processing on the acquired signal, eliminate background magnetic field interference by using the two-sample difference method, plot the leakage magnetic field peak value as a function of magnetization current or AC parameters, determine the optimal magnetization conditions, and verify the signal gain effect.

2. The method for increasing the leakage magnetic field of internal defects in ferromagnetic materials based on AC / DC composite magnetization according to claim 1, characterized in that: The composite excitation coil is wound with multi-strand enameled copper wire. The coil type is a composite excitation coil assembly. The DC current value ranges from 0 to 30A, the frequency range of the high-frequency AC current is 500 to 1500Hz, and the amplitude is 1 to 6A.

3. The method for increasing the leakage magnetic field of internal defects in ferromagnetic materials based on AC / DC composite magnetization according to claim 1, characterized in that: The Hall sensor array adopts a differential pair structure, with two Hall sensors installed vertically 8-10mm apart. The differential amplifier eliminates geomagnetic field interference. The sensor output signal is sampled by a 24-bit ADC with a sampling rate set to 10kHz. A digital filter removes power frequency interference and high-frequency noise. The original signal and the background magnetic field signal are subtracted in real time by an FPGA. The net leakage magnetic field data is sent to the host computer for storage via a wireless transmission module.

4. The method for increasing the leakage magnetic field of internal defects in ferromagnetic materials based on AC / DC composite magnetization according to claim 1, characterized in that: The background magnetic field subtraction adopts dual Hall sensor dynamic calibration technology, which automatically corrects the sensor sensitivity difference through correlation analysis of the two sensor signals; an adaptive filtering algorithm is introduced in the signal conditioning stage, which dynamically adjusts the filtering parameters according to the real-time noise level to ensure that the effective signal retention rate is not less than 95%; the optimal magnetization current is determined by gradient descent method, which iteratively calculates the second derivative of the leakage magnetic field signal amplitude and the current value to locate the current value corresponding to the maximum gradient point.

5. The method for increasing the leakage magnetic field of internal defects in ferromagnetic materials based on AC / DC composite magnetization according to claim 1, characterized in that: In pure DC magnetization mode, the original signal is extracted and the background magnetic field is subtracted to obtain the reference leakage magnetic field; in AC / DC composite magnetization mode, the original signal is extracted and the background magnetic field is subtracted to obtain the enhanced leakage magnetic field by adding the optimal magnetization current and AC current; by comparing the peak changes of the reference and enhanced leakage magnetic fields, the optimal parameter combination is determined.

6. The method for increasing the leakage magnetic field of internal defects in ferromagnetic materials based on AC / DC composite magnetization according to claim 5, characterized in that: The criteria for determining the optimal parameter combination are the curves showing the change of the peak leakage magnetic field with magnetizing current, AC amplitude, and frequency.

7. A magnetization system for increasing the leakage magnetic field of internal defects in ferromagnetic materials based on AC / DC composite magnetization, wherein the method for increasing the leakage magnetic field of internal defects in ferromagnetic materials according to any one of claims 1 to 6 is characterized in that, include: A composite excitation coil assembly and a signal conditioning unit are provided; the coil assembly is simultaneously supplied with DC current and high-frequency AC current to achieve synchronous excitation; the signal conditioning unit performs background magnetic field subtraction and peak value comparison analysis.

8. The magnetization system according to claim 7, characterized in that: The signal conditioning unit integrates a Hall sensor array and a data processing module; the sensor array is linearly arranged with a lift-off value of 0.4~0.6mm and a scanning speed of 5~10mm / s; the data processing module performs raw signal acquisition, background magnetic field subtraction, peak extraction, and curve fitting.

Images