Metal particle overlap detection methods, electronic devices and storage media

By performing coherent demodulation and AC coupling processing on the amplitude-modulated wave signal of the three-coil inductive metal particle detection system, the zero-crossing time and frequency domain phase features were extracted, solving the problem of particle overlap signal identification under high-concentration conditions and achieving particle counting with high accuracy and low complexity.

CN122775504BActive Publication Date: 2026-11-10GUANGZHOU MECHANICAL ENGINEERING RESEARCH INSTITUTE CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202611274440.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2026-08-21
Publication Date
2026-11-10
Estimated Expiration
2046-08-21

AI Technical Summary

Technical Problem

Existing three-coil inductive metal particle detection systems have difficulty effectively distinguishing between single particles and overlapping signals of multiple particles under high-concentration conditions, resulting in distorted particle counting results, affecting equipment wear condition assessment and increasing operation and maintenance costs.

Method used

By coherently demodulating the amplitude-modulated wave signal output by the sensor, the baseband envelope signal is extracted and AC coupling is performed to obtain the zero-mean baseband envelope waveform. The zero-crossing time feature and frequency domain phase feature are extracted, and the overlap of multiple metal particle signals is determined by the zero-crossing time difference and phase difference.

Benefits of technology

It improves the accuracy and reliability of particle counting, avoids misjudgments caused by factors such as noise and particle speed fluctuations, requires little computation and does not require additional hardware costs, making it suitable for embedded online monitoring platforms.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122775504B_ABST
    Figure CN122775504B_ABST
Patent Text Reader

Abstract

This application provides a method, electronic device, and storage medium for detecting overlapping metal particles, belonging to the field of metal particle detection technology. The method includes: coherently demodulating an amplitude-modulated wave signal to extract a baseband envelope signal; performing AC coupling processing on the baseband envelope signal to obtain a zero-mean baseband envelope waveform; extracting the zero-crossing time feature and frequency domain phase feature of the baseband envelope waveform; comparing the zero-crossing time feature with a preset zero-crossing time reference value to obtain a zero-crossing time difference, and comparing the frequency domain phase feature with a preset phase reference value to obtain a phase difference; when the zero-crossing time difference is greater than or equal to a preset time threshold and the phase difference is greater than or equal to a preset phase threshold, the baseband envelope signal is determined to be generated by the overlap of multiple metal particle signals. This application can effectively distinguish between a single particle signal and multiple overlapping particle signals, improving the accuracy of particle counting.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of metal particle detection technology, and in particular to a method for detecting overlapping metal particles, an electronic device, and a storage medium. Background Technology

[0002] Three-coil inductive metal particle detection sensors are widely used for online oil monitoring in critical equipment such as hydraulic systems and engine lubrication systems due to their simple structure, high sensitivity, and non-clogging characteristics. These sensors typically consist of two axially symmetrically placed excitation coils and a single detection coil in the middle. The two excitation coils are energized with sinusoidal alternating currents of the same frequency, amplitude, and opposite phase, generating a near-zero net magnetic field at the detection coil. When metal particles in the oil flow axially through the detection area, they disturb the alternating magnetic field generated by the excitation coils, inducing an amplitude-modulated wave signal in the detection coil. By coherently demodulating this amplitude-modulated wave, the baseband envelope signal can be extracted. The amplitude, width, and waveform of this envelope signal are closely related to the size, material, and velocity of the particles. Therefore, by analyzing the characteristic parameters of the baseband envelope waveform, the size and count of metal particles in the oil can be assessed, providing crucial data support for predictive maintenance and wear condition assessment of equipment.

[0003] In existing three-coil inductive particle detection systems, traditional particle identification and counting methods are mainly based on the time-domain amplitude characteristics or pulse width characteristics of the baseband envelope. Specifically, the system acquires the amplitude-modulated wave signal output by the detection coil in real time, obtains the baseband envelope after coherent demodulation, and determines that a particle has passed through the detection area when the envelope amplitude exceeds a preset trigger threshold. Furthermore, the equivalent size of the particle is calculated based on the peak amplitude of the envelope, and the pulse width of the envelope helps determine the particle's motion state. This type of method is simple in principle, requires little computation, and can meet the online monitoring needs under low particle concentration conditions, thus it has been widely used in existing commercial online oil monitoring sensors.

[0004] However, the aforementioned traditional detection methods have significant limitations in practical applications. When the particle concentration in the oil is high, the distance between particles may be smaller than the sensor's detection area, causing multiple particles to enter the sensing range of the detection coil simultaneously or sequentially. In this case, the amplitude-modulated wave signals sensed by each particle aliased in the time domain, and the baseband envelope output by the detection coil presented as a single composite waveform, rather than multiple independent pulses. The traditional amplitude threshold method can only identify one envelope peak of this composite waveform and misclassify it as a large single particle based on the amplitude of this peak; the pulse width method is difficult to provide a reasonable judgment result due to the abnormally increased width of the composite waveform. These misjudgments will directly lead to the undercounting of small particles and the false reporting of large particles, causing serious distortion of particle counting results and size distribution curves, thereby affecting the accurate assessment of equipment wear status, and even triggering unnecessary equipment shutdown for inspection, increasing operation and maintenance costs.

[0005] To address the misjudgment problem caused by signal aliasing, researchers have attempted to improve the situation using various technical approaches. Some studies have proposed aliasing signal separation methods based on pulse waveform decomposition or matching pursuit algorithms, identifying the number of particles by decomposing the composite waveform into a superposition of multiple single pulses. Other studies have introduced deep learning frameworks, utilizing models such as convolutional neural networks to classify and identify aliased waveforms. Furthermore, some schemes attempt to reduce the probability of aliasing by changing the sensor excitation frequency or using multi-frequency excitation. However, waveform decomposition and matching pursuit methods typically require prior knowledge of the pulse template and are sensitive to changes in particle velocity and signal-to-noise ratio fluctuations. While deep learning methods achieve high accuracy under certain conditions, their model training requires a large number of labeled samples, consuming significant computational resources and making real-time deployment on embedded online monitoring platforms difficult. Variable frequency or multi-frequency excitation schemes require additional hardware overhead and place higher demands on the sensor driving circuitry and subsequent signal processing links, significantly increasing system complexity and cost.

[0006] Therefore, how to effectively identify and distinguish between single particles and overlapping signals of multiple particles without increasing hardware costs or significantly increasing computational complexity, and improve the accuracy of particle counting under high-concentration conditions, is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0007] The purpose of this application is to provide a method for detecting overlapping metal particles, an electronic device, and a storage medium to solve the above-mentioned problems.

[0008] To achieve the above objectives, firstly, this application proposes a method for detecting overlapping metal particles, the method comprising: The amplitude-modulated wave signal output by the sensor is coherently demodulated to extract the baseband envelope signal; The baseband envelope signal is subjected to AC coupling processing to obtain a zero-mean baseband envelope waveform; Extract the zero-crossing time features and frequency domain phase features of the baseband envelope waveform; The zero-crossing time feature is compared with a preset zero-crossing time reference value to obtain the zero-crossing time difference, and the frequency domain phase feature is compared with a preset phase reference value to obtain the phase difference. When the zero-crossing time difference is greater than or equal to a preset time threshold and the phase difference is greater than or equal to a preset phase threshold, the baseband envelope signal is determined to be generated by the overlap of multiple metal particle signals.

[0009] In some embodiments, extracting the zero-crossing time features and frequency domain phase features of the baseband envelope waveform includes: Based on the distribution of the zero-crossing points of the baseband envelope waveform on the time axis, the zero-crossing time characteristic quantity is determined; Perform a Fourier transform on the pulse segment of the baseband envelope waveform, and extract the phase value at the peak frequency in the transformed spectrum as the frequency domain phase feature.

[0010] In some embodiments, the spectrum includes an amplitude spectrum and a phase spectrum, and the step of performing a Fourier transform on the pulse segment of the baseband envelope waveform and extracting the phase value at the peak frequency in the transformed spectrum as a frequency domain phase feature includes: Perform a Fourier transform on the pulse segment of the baseband envelope waveform to obtain the amplitude spectrum and phase spectrum; Determine the peak frequency in the amplitude spectrum; In the phase spectrum, the phase value at the peak frequency is extracted as a frequency domain phase feature.

[0011] In some embodiments, determining the peak frequency in the amplitude spectrum includes: Search for the maximum value point of the amplitude spectrum according to the preset frequency range; The frequency corresponding to the maximum value point is taken as the peak frequency. In some embodiments, determining the zero-crossing time characteristic based on the distribution of the zero-crossing points of the baseband envelope waveform on the time axis includes: Detect the zero-crossing point of the baseband envelope waveform; Based on the distribution of each zero-crossing point on the time axis, the time interval between adjacent zero-crossing points is calculated, and the time interval is used as a zero-crossing time feature.

[0012] In some embodiments, detecting the zero-crossing point of the baseband envelope waveform includes: Calculate the amplitude product of adjacent sampling points in the baseband envelope waveform; When the amplitude product is less than zero, the zero-crossing point between the adjacent sampling points is determined by a linear interpolation algorithm. In some embodiments, the baseband envelope signal includes a baseband in-phase component and a baseband quadrature component, and the coherent demodulation of the amplitude-modulated wave signal output by the sensor to extract the baseband envelope signal includes: Generate in-phase and quadrature reference signals that are in the same frequency as the amplitude-modulated wave signal output by the sensor; The amplitude-modulated wave signal is multiplied by the in-phase reference signal and the quadrature reference signal respectively, and then low-pass filtered to obtain the baseband in-phase component and the baseband quadrature component. In some embodiments, the method further includes: When the zero-crossing time difference is greater than the first extreme threshold or the phase difference is greater than the second extreme threshold, the baseband envelope signal is determined to be generated by the overlap of multiple metal particle signals, wherein the first extreme threshold is greater than the preset time threshold and the second extreme threshold is greater than the preset phase threshold.

[0013] Secondly, to achieve the above objectives, this application also proposes an electronic device, comprising: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors perform the metal particle overlap detection method as described above.

[0014] Thirdly, to achieve the above objectives, this application also proposes a computer storage medium storing executable instructions, which, when executed by a processor, cause the processor to perform the metal particle overlap detection method described above.

[0015] Compared with the prior art, the beneficial effects of this application include: Firstly, this application coherently demodulates the amplitude-modulated wave signal output by the sensor to extract the baseband envelope signal, and further obtains a zero-mean baseband envelope waveform through AC coupling, making the baseband envelope signal symmetrical around the zero axis. Based on this, zero-crossing time features are extracted from the time domain, effectively characterizing the symmetry of the pulse waveform and the uniformity of the zero-crossing point distribution. Since the zero-crossing time features are insensitive to changes in particle velocity and signal amplitude fluctuations, they exhibit strong robustness. Compared to traditional methods that rely solely on amplitude thresholds for particle detection, this application provides richer discrimination information from the waveform morphology dimension.

[0016] Secondly, since the baseband envelope of a single particle signal has a definite waveform shape, its dominant frequency phase has a stable reference value; however, the spectral dominant frequency phase of multiple overlapping particle signals will shift significantly with different particle spacing, and the shift amount is correlated with the degree of overlap. This application extracts frequency domain phase features from the frequency domain to characterize the overall waveform shape from the frequency domain dimension, which is independent of and complementary to the zero-crossing features in the time domain.

[0017] Thirdly, this application only determines a multi-particle overlapping signal when the zero-crossing time difference reaches a preset time threshold and the phase difference reaches a preset phase threshold. This effectively avoids accidental misjudgments caused by noise, particle velocity fluctuations, or material differences in a single dimension, significantly improving the accuracy and reliability of the determination. Compared to existing solutions that require complex calculations (such as waveform decomposition, matching pursuit, and deep learning inference), this application involves minimal computation, only zero-crossing detection and FFT operations, which can be implemented in real time on an embedded online monitoring platform without increasing hardware costs. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation on the scope of this application.

[0019] Figure 1 This is a flowchart illustrating a method for detecting overlapping metal particles in one embodiment; Figure 2 This is a schematic diagram illustrating the process of coherently demodulating the amplitude-modulated wave signal output by the sensor and extracting the baseband envelope signal in one embodiment. Figure 3 This is a detailed flowchart illustrating the extraction process of zero-crossing time features and frequency domain phase features of the baseband envelope waveform in one embodiment. Figure 4 This is a comparison diagram of the zero-crossing time distribution of the baseband envelope waveform of a single metal particle and two overlapping metal particles in one embodiment; Figure 5 This is a comparison diagram of the baseband envelope main frequency peak phase of single metal particles of different sizes and overlapping metal particles of different spacing in one embodiment; Figure 6 This is a schematic diagram of the electronic device involved in the metal particle overlap detection method in the embodiments of this application. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0021] All terms used in this application (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0022] For example, the terms "first" and "second" in this application are used only to distinguish and describe similar objects, not to describe a specific order or sequence, nor should they be construed as indicating or implying relative importance. For instance, without departing from the scope of this application, a first extreme threshold can be referred to as a second extreme threshold, and similarly, a second extreme threshold can be referred to as a first extreme threshold. Both the first extreme threshold and the second extreme threshold are thresholds, but they are not the same threshold.

[0023] For example, the terms "comprising" or "including" used in this application indicate the presence of features, steps, operations and / or components, but do not exclude the presence or addition of one or more other features, steps, operations or components.

[0024] As mentioned earlier, when the particle concentration in the oil is high, multiple particles enter the sensor's detection area simultaneously or sequentially, causing the amplitude-modulated wave signals sensed by each particle to alias in the time domain. Traditional amplitude thresholding methods easily misclassify multiple overlapping small particles as a single large particle, resulting in counting errors and size distribution distortion. Although there are improved schemes based on waveform decomposition, matching pursuit, or deep learning in the existing technology, they generally suffer from problems such as computational complexity, reliance on prior templates, and difficulty in real-time deployment on embedded platforms. To address this, this application proposes a metal particle overlap detection method, electronic device, and storage medium that can effectively distinguish between single particle signals and multiple overlapping particle signals without increasing hardware costs or significantly increasing computational complexity, thereby improving the accuracy of particle counting.

[0025] like Figure 1 As shown in the figure, this application provides a method for detecting overlapping metal particles, the method comprising the following steps: Step S10: Coherently demodulate the amplitude-modulated wave signal output by the sensor to extract the baseband envelope signal.

[0026] In this embodiment, the sensor is a three-coil inductive metal particle detection sensor. The sensor includes two excitation coils symmetrically placed along the axial direction and a single detection coil located in the middle. The two excitation coils are energized with sinusoidal alternating currents of the same frequency, equal amplitude, and opposite phase (frequency denoted as ). (e.g., 150kHz) generates a net magnetic field of approximately zero at the detection coil. When a metal particle passes axially through the sensor's detection area, it disturbs the magnetic field distribution, inducing an amplitude-modulated wave signal in the detection coil. The carrier frequency and excitation frequency of this amplitude-modulated wave signal The shape of its envelope is consistent with the size, material, velocity, and position of the particles in the sensitive area.

[0027] The baseband envelope signal refers to the low-frequency modulation component extracted after coherent demodulation of the high-frequency amplitude-modulated wave signal output by the sensor, including the baseband in-phase component and the baseband quadrature component.

[0028] In some implementations, such as Figure 2 As shown, step S10 includes: Step S11: Generate an in-phase reference signal and a quadrature reference signal that are in the same frequency as the amplitude-modulated wave signal output by the sensor.

[0029] Specifically, a local reference signal is generated using a frequency synthesizer or phase-locked loop circuit. This local reference signal includes both an in-phase reference signal and a quadrature reference signal. The in-phase reference signal is: The orthogonal reference signal is: .

[0030] The frequencies of the in-phase and quadrature reference signals are related to the excitation frequency of the sensor. Strict consistency is required to ensure the accuracy of coherent demodulation.

[0031] Step S12: Multiply the amplitude-modulated wave signal with the in-phase reference signal and the quadrature reference signal respectively, and then perform low-pass filtering to obtain the baseband in-phase component and the baseband quadrature component.

[0032] Specifically, the amplitude-modulated wave signal respectively with and Multiplication, i.e., mixing:

[0033]

[0034] Obtain in-phase mixing signal and quadrature mixing signals Since the mixed signal contains high-frequency components and low-frequency baseband components, the high-frequency components can be filtered out and the low-frequency baseband components can be retained by a low-pass filter (such as a 4th or 6th order Butterworth low-pass filter, the cutoff frequency of which needs to be higher than the reciprocal of the speed of the metal particles, such as 1kHz).

[0035] In some implementations, zero-phase filtering techniques (such as the `filtfilt` function in MATLAB) can be used for low-pass filtering to eliminate the phase delay introduced by the filter and ensure that the zero-crossing position of the baseband envelope signal does not shift. It should be noted that zero-phase filtering is implemented through forward-backward filtering. Although it introduces a small delay (approximately half the filter order), this delay is fixed and known, and can be compensated for through time alignment, without affecting the accurate calculation of the zero-crossing time difference.

[0036] After low-pass filtering, the baseband in-phase component is obtained. Orthogonal components of baseband .

[0037] Step S10 enables the extraction of baseband envelope signal from high-frequency amplitude-modulated wave signal, which can convert the magnetic field disturbance caused by particles passing through the sensor into a low-frequency waveform that is easy to process digitally, providing input for subsequent extraction of zero-crossing time features and frequency domain phase features.

[0038] In some implementations, the baseband envelope signal (baseband in-phase component and baseband quadrature component) can also be downsampled to reduce computation while retaining sufficient time-domain resolution for zero-crossing detection.

[0039] Step S20: Perform AC coupling processing on the baseband envelope signal to obtain a zero-mean baseband envelope waveform.

[0040] In this embodiment, the zero-mean baseband envelope waveform refers to the AC-coupled waveform that is symmetrical around the zero axis after removing the DC bias component from the baseband envelope signal. Since the baseband envelope signal output in step S10 typically contains a DC bias, it causes the overall waveform of the baseband envelope signal to deviate from the zero axis, leading to zero-crossing detection failure. This DC bias may originate from residual coupling between the sensor excitation coil and the detection coil, circuit asymmetry, or zero drift of the preamplifier. By performing AC coupling processing on the baseband envelope signal, the DC component can be removed, making the baseband envelope waveform symmetrical around the zero axis, thereby enabling effective zero-crossing detection.

[0041] Specifically, the mean values ​​of the baseband in-phase component and the baseband quadrature component in the baseband envelope signal are calculated within a time window, and then their respective mean values ​​are subtracted from the baseband in-phase component and the baseband quadrature component to obtain the baseband envelope waveform with zero mean.

[0042]

[0043] in, The baseband in-phase component of the baseband envelope waveform with zero mean. The baseband quadrature component of the baseband envelope waveform with zero mean. mean(·) represents the arithmetic mean of the signal over a preset time window. The length of the time window should be sufficient to cover the duration of the entire pulse segment, and can be set to 1.5 to 2 times the length of the sensor detection area to ensure the accuracy of the mean calculation.

[0044] In some implementations, to prevent inaccurate mean estimation due to the slow drift of the DC bias over time, a sliding window method can be used to calculate the mean. That is, for each sampling point, a local window (N can be 500~2000) is formed to calculate the mean, thereby achieving adaptive AC coupling. This dynamic mean removal method can effectively track the slow changes in DC bias and improve the robustness of zero-crossing detection.

[0045] Step S20 converts the baseband envelope signal into a zero-mean bipolar waveform. When a complete pulse segment of the baseband envelope waveform transitions from the positive half-cycle to the negative half-cycle or from the negative half-cycle to the positive half-cycle, it will inevitably cross the zero axis.

[0046] Step S30, Extract the zero-crossing time features and frequency domain phase features of the baseband envelope waveform.

[0047] In this embodiment, the zero-crossing time feature refers to the time-domain characteristic parameter extracted from the zero-mean baseband envelope waveform, used to quantify the distribution of zero-crossing points on the time axis. Specifically, the zero-crossing time feature can be the time interval between adjacent zero-crossing points within the same pulse segment, the mean, variance, or coefficient of variation of all adjacent zero-crossing time intervals, etc. The frequency-domain phase feature refers to the phase value located at the peak frequency extracted from the spectrum of the pulse segment of the zero-mean baseband envelope waveform.

[0048] In some implementations, such as Figure 3 As shown, step S30 includes: Step S31: Determine the zero-crossing time characteristic quantity based on the distribution of the zero-crossing points of the baseband envelope waveform on the time axis.

[0049] Specifically, the zero-crossing points of the baseband envelope waveform are detected; the time interval between adjacent zero-crossing points is calculated based on the distribution of each zero-crossing point on the time axis, and the time interval is used as a zero-crossing time feature.

[0050] A zero-crossing point refers to the moment when the zero-mean baseband envelope waveform crosses the time axis (i.e., its amplitude is zero). Since the actual sampled signal consists of discrete data points, the probability that the amplitude of the baseband envelope waveform is exactly zero is extremely low. The existence of a zero-crossing point can be determined by observing the change in the sign of the amplitudes of adjacent sample points. Specifically, the amplitude product of adjacent sample points in the baseband envelope waveform can be calculated; when the amplitude product is less than zero, a linear interpolation algorithm is used to determine the zero-crossing point between the adjacent sample points.

[0051] For example, for adjacent sampling points in the baseband envelope waveform ( , )and( , ),in Represents the time point of any pulse. This represents the amplitude at the corresponding time point.

[0052] when When the time passes through zero:

[0053] If the zero-crossing points detected within the same baseband envelope waveform are sequentially... , Then the time interval between adjacent zero-crossing points is: .

[0054] It should be noted that for the baseband envelope waveform generated by a single particle passing through the sensor, due to the good waveform symmetry, the time between the zero-crossing point of the rising edge and the peak point is approximately equal to the time between the peak point and the zero-crossing point of the falling edge, and the interval between adjacent zero-crossing points is uniform. For the baseband envelope waveform generated by multiple overlapping metal particles passing through the sensor, the waveform is distorted, and the zero-crossing points are unevenly distributed. The zero-crossing time characteristic can be used to quantify this uneven distribution.

[0055] Step S32: Perform a Fourier transform on the pulse segment of the baseband envelope waveform, and extract the phase value at the peak frequency in the transformed spectrum as the frequency domain phase feature.

[0056] In this embodiment, a pulse segment is a continuous time interval within the baseband envelope waveform where the amplitude exceeds a preset trigger threshold. For example, when the amplitude in the baseband envelope waveform exceeds the preset trigger threshold (e.g., 3 times the noise floor), this moment is recorded as the start point of the pulse segment; when the amplitude drops back and remains below the preset trigger threshold for a preset duration (e.g., remaining below the preset trigger threshold for more than 0.1 ms), this is recorded as the end point. The signal segment from the start point to the end point is extracted as the pulse segment.

[0057] In some implementations, a Fourier transform can be performed on the pulse segment of the baseband envelope waveform to obtain an amplitude spectrum and a phase spectrum; a peak frequency can be determined in the amplitude spectrum; and the phase value at the peak frequency can be extracted from the phase spectrum as a frequency domain phase feature.

[0058] Specifically, a Fast Fourier Transform (FFT) is performed on the extracted pulse segment signal to obtain a complex spectrum. The complex spectrum takes complex values ​​at each frequency point. The magnitudes constitute the amplitude spectrum, reflecting the energy distribution of the signal at different frequencies, and the arguments constitute the phase spectrum, reflecting the phase shift of the signal at different frequencies. The number of points in the FFT is denoted as... , The value of can be an integer power of 2, and is not less than the length of the pulse segment. Frequency resolution: .in The sampling rate.

[0059] Since the DC component (frequency near 0) may retain a small amount of energy due to incomplete AC coupling, to avoid misjudging the DC component as the peak frequency, a preset frequency range is used […]. , Search for the maximum value point of the amplitude spectrum; and use the frequency corresponding to the maximum value point. As the peak frequency. It can be set to a value slightly greater than 0 (such as 5Hz). The frequency can be determined based on the highest possible frequency of the baseband envelope signal generated by the metal particle passing through the sensor, which is related to the particle pulse width. For example, if the minimum particle pulse width is 1ms, the corresponding main frequency is approximately 1kHz. It can be set to 2kHz.

[0060] Find the peak frequency in the phase spectrum. Corresponding phase value The phase value is expressed in radians (rad) and ranges from [ ]. π,π].

[0061] It should be noted that for the baseband envelope waveform generated by a single particle passing through the sensor, due to its definite shape, the phase at its peak frequency has a relatively stable value (with very small fluctuations under different particle sizes). For the baseband envelope waveform generated by multiple overlapping metal particles passing through the sensor, due to the time-domain misalignment of the multiple particle sub-signals, the phase at its peak frequency will be significantly shifted, and the amount of shift is correlated with the particle spacing.

[0062] Step S40: The zero-crossing time feature is compared with a preset zero-crossing time reference value to obtain the zero-crossing time difference, and the frequency domain phase feature is compared with a preset phase reference value to obtain the phase difference. When the zero-crossing time difference is greater than or equal to a preset time threshold and the phase difference is greater than or equal to a preset phase threshold, it is determined that the baseband envelope signal is generated by the overlap of multiple metal particle signals.

[0063] In this embodiment, the zero-crossing time reference value This refers to the statistical average (or median) of the zero-crossing time characteristic of the baseband envelope waveform when a single metal particle passes through the sensor. Phase reference value. This refers to the statistical average (or median) of the phase value at the peak frequency of the baseband envelope waveform when a single metal particle passes through the sensor. For example, multiple baseband envelope signal samples of single metal particles (covering different sizes and materials within the expected detection range) passing through the sensor can be selected, and the zero-crossing time feature and frequency domain phase feature of each sample can be extracted. Their statistical average can then be calculated to obtain a reference value.

[0064] Zero crossing time difference .in This refers to the currently detected zero-crossing time feature. The preset time threshold is a judgment boundary determined through single-particle sample statistics. For example, it can be set to 3 times the standard deviation of the zero-crossing time of a single-particle sample (i.e., the 3σ criterion), or set to a fixed value (such as 0.0001s) based on experimental verification.

[0065] Phase difference The preset phase threshold is determined by statistical analysis of single-particle samples. For example, it can be set to 3 times the standard deviation of the phase of a single-particle sample, or set to a fixed value (such as 0.02 rad).

[0066] when A preset time threshold (e.g., 0.0001s) is set, and A preset phase threshold (e.g., 0.02 rad) is used to determine that the current baseband envelope signal is generated by the overlap of multiple metal particle signals. The AND-AND-AND decision logic improves the reliability of the determination and reduces the false positive rate caused by random fluctuations of a single feature. Only when the waveform exhibits significant distortion in both the time and frequency domains is it identified as a multi-particle overlapping signal.

[0067] In some embodiments, the method further includes: determining that the baseband envelope signal is generated by the overlap of multiple metal particle signals when the zero-crossing time difference is greater than a first extreme threshold or the phase difference is greater than a second extreme threshold, wherein the first extreme threshold is greater than the preset time threshold and the second extreme threshold is greater than the preset phase threshold. The first extreme threshold and the second extreme threshold can be set to 5 to 10 times the preset time threshold and the preset phase threshold, respectively. When one of the features, the zero-crossing time difference or the phase difference, far exceeds the normal range, even if the other feature has not yet reached the preset threshold, it is sufficient to indicate that the waveform has been severely distorted. In this case, the OR judgment logic is used to directly determine that it is a multi-particle overlap. This dual judgment logic ensures high accuracy under normal conditions and also takes into account high recall under extreme distortion conditions.

[0068] Finally, to verify the effectiveness of the proposed method for detecting overlapping metal particles, the applicant conducted an experimental verification using a three-coil inductive metal particle detection sensor.

[0069] The experiment used copper particles as the detection object, with particle diameters of 400 μm, 600 μm, and 800 μm, and a particle velocity of 2.5 m / s. The spacing between the overlapping metal particles was 5 mm. The sensor sampling rate... =1MHz, excitation voltage 30V, excitation frequency =150kHz. The three-coil sensor consists of two excitation coils placed symmetrically along the axis and a single detection coil located in the middle.

[0070] like Figure 4 As shown in the figure, the zero-crossing time distribution of the baseband envelope waveform is compared when a single metal particle and two metal particles spaced 5 mm apart pass through the sensor. Experimental results show that the zero-crossing time of the baseband envelope waveform generated by the two 5 mm-spaced metal particles differs significantly from that of a single particle passing through the sensor, with a difference of approximately 1 ms, far exceeding the preset time threshold (0.0001 s) set in this application. This indicates that when multiple particles overlap in time, the temporal symmetry of the baseband envelope waveform is disrupted, and the zero-crossing distribution deviates from the reference value for a single particle. The zero-crossing time characteristic can serve as an effective indicator for distinguishing between single-particle and multi-particle overlap.

[0071] like Figure 5As shown in the figure, the baseband envelope peak phase of a single copper particle of different sizes is compared with that of two copper particles with different spacing and a diameter of 0.3 mm. Experimental results show that variations in the size of a single particle (from 400 μm to 800 μm) do not significantly affect the peak phase of the baseband envelope; the peak phase of a single particle remains near a stable reference value. However, a significant phase difference exists between the peak phases of multiple particles and those of a single particle, and the particle spacing is the main factor affecting the magnitude of this phase difference, exhibiting a certain linear relationship—the smaller the particle spacing, the higher the degree of overlap, and the greater the phase difference.

[0072] comprehensive Figure 4 and Figure 5 The experimental results show that the zero-crossing time feature and the peak phase feature of the dominant frequency in the frequency domain characterize the morphology of the baseband envelope waveform from two independent dimensions. A single size change does not cause a significant shift in the dominant frequency phase, while multiple overlapping particles simultaneously produce detectable changes in both the time domain (zero-crossing point distribution distortion) and the frequency domain (dominant frequency phase shift). Therefore, jointly determining the zero-crossing time difference and the peak phase difference of the dominant frequency can effectively distinguish between single-particle and multi-particle overlapping signals and has good adaptability to particles of different sizes. The experiments also show that when the particle spacing further decreases, both the zero-crossing time difference and the phase difference increase, and the detection sensitivity improves accordingly. When the particle spacing increases to exceed the sensor's sensitive area size, the two pulses completely separate in the time domain, and both feature values ​​fall back to near the single-particle reference value, preventing misjudgment.

[0073] In the metal particle overlap detection method proposed in this application, firstly, the application coherently demodulates the amplitude-modulated wave signal output by the sensor to extract the baseband envelope signal, and further obtains a zero-mean baseband envelope waveform through AC coupling, making the baseband envelope signal symmetrical around the zero axis. Based on this, zero-crossing time features are extracted from the time domain, which can effectively characterize the symmetry of the pulse waveform and the uniformity of the zero-crossing point distribution. Since the zero-crossing time features are insensitive to particle velocity changes and signal amplitude fluctuations, they have strong robustness. Compared to traditional methods that rely solely on amplitude thresholds for particle detection, this application provides richer discrimination information from the waveform morphology dimension.

[0074] Secondly, since the baseband envelope of a single particle signal has a definite waveform shape, its dominant frequency phase has a stable reference value; however, the spectral dominant frequency phase of multiple overlapping particle signals will shift significantly with different particle spacing, and the shift amount is correlated with the degree of overlap. This application extracts frequency domain phase features from the frequency domain to characterize the overall waveform shape from the frequency domain dimension, which is independent of and complementary to the zero-crossing features in the time domain.

[0075] Thirdly, this application only determines a multi-particle overlapping signal when the zero-crossing time difference reaches a preset time threshold and the phase difference reaches a preset phase threshold. This effectively avoids accidental misjudgments caused by noise, particle velocity fluctuations, or material differences in a single dimension, significantly improving the accuracy and reliability of the determination. Compared to existing solutions that require complex calculations (such as waveform decomposition, matching pursuit, and deep learning inference), this application involves minimal computation, only zero-crossing detection and FFT operations, which can be implemented in real time on an embedded online monitoring platform without increasing hardware costs.

[0076] Fourthly, this application also sets up a judgment logic under extreme threshold conditions, which further improves the detection recall rate under extreme distortion conditions.

[0077] In one embodiment, a computer storage medium is provided that stores executable instructions that, when executed by a processor, cause the processor to perform the steps in the above method embodiments.

[0078] In one embodiment, an electronic device is also provided, including one or more processors; and a memory storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors perform the steps in the above method embodiments.

[0079] In one embodiment, such as Figure 6 The diagram illustrates the structure of an electronic device used to implement an embodiment of this application. The electronic device includes a central processing unit (CPU) 101, which can perform various appropriate actions and processes based on a program stored in a read-only memory (ROM) 102 or a program loaded from a storage portion 108 into a random access memory (RAM) 103. The RAM 103 also stores various programs and data required for the operation of the electronic device. The CPU 101, ROM 102, and RAM 103 are interconnected via a bus 104. An input / output (I / O) interface 105 is also connected to the bus 104.

[0080] The following components are connected to I / O interface 105: an input section 106 including a keyboard, mouse, etc.; an output section 107 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 108 including a hard disk, etc.; and a communication section 109 including a network interface card such as a LAN card, modem, etc. The communication section 109 performs communication processing via a network such as the Internet. A drive 110 is also connected to I / O interface 105 as needed. A removable medium 111, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 110 as needed so that computer programs read from it can be installed into storage section 108 as needed.

[0081] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer-readable medium carrying instructions that, in such embodiments, can be downloaded and installed from a network via communication section 109, and / or installed from removable medium 111. When the instructions are executed by central processing unit (CPU) 101, the various method steps described in this application are performed.

[0082] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0083] Furthermore, those skilled in the art will understand that although some embodiments herein include certain features included in other embodiments but not others, combinations of features from different embodiments are intended to be within the scope of this application and form different embodiments. For example, any of the embodiments or implementations claimed above can be used in any combination. The information disclosed in this background section is intended only to enhance the understanding of the general background of this application and should not be construed as an admission or in any way implying that such information constitutes prior art known to those skilled in the art.

Claims

1. A method for detecting overlapping metal particles, characterized in that, The method includes: The amplitude-modulated wave signal output by the sensor is coherently demodulated to extract the baseband envelope signal; The baseband envelope signal is subjected to AC coupling processing to obtain a zero-mean baseband envelope waveform; Calculate the amplitude product of adjacent sampling points in the baseband envelope waveform. When the amplitude product is less than zero, determine the zero-crossing point between the adjacent sampling points using a linear interpolation algorithm. Calculate the time interval between adjacent zero-crossing points based on the distribution of each zero-crossing point on the time axis, and use the time interval as a zero-crossing time feature. Fourier transform is performed on the pulse segment of the baseband envelope waveform to obtain the amplitude spectrum and phase spectrum. According to the preset frequency range, the maximum value point of the amplitude spectrum is searched, and the frequency corresponding to the maximum value point is taken as the peak frequency. In the phase spectrum, the phase value at the peak frequency is extracted as the frequency domain phase feature quantity. The zero-crossing time feature is compared with a preset zero-crossing time reference value to obtain the zero-crossing time difference, and the frequency domain phase feature is compared with a preset phase reference value to obtain the phase difference. When the zero-crossing time difference is greater than or equal to a preset time threshold and the phase difference is greater than or equal to a preset phase threshold, the baseband envelope signal is determined to be generated by the overlap of multiple metal particle signals.

2. The method for detecting overlapping metal particles according to claim 1, characterized in that, The baseband envelope signal includes a baseband in-phase component and a baseband quadrature component. The coherent demodulation of the amplitude-modulated wave signal output from the sensor to extract the baseband envelope signal includes: Generate in-phase and quadrature reference signals that are in the same frequency as the amplitude-modulated wave signal output by the sensor; The amplitude-modulated wave signal is multiplied by the in-phase reference signal and the quadrature reference signal respectively, and then low-pass filtered to obtain the baseband in-phase component and the baseband quadrature component.

3. The method for detecting overlapping metal particles according to claim 1, characterized in that, The method further includes: When the zero-crossing time difference is greater than the first extreme threshold and the phase difference is greater than the second extreme threshold, the baseband envelope signal is determined to be generated by the overlap of multiple metal particle signals, wherein the first extreme threshold is greater than the preset time threshold and the second extreme threshold is greater than the preset phase threshold.

4. An electronic device, characterized in that, include: One or more processors; Memory, used to store one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors perform the metal particle overlap detection method as described in any one of claims 1 to 3.

5. A computer storage medium, characterized in that, The storage medium stores executable instructions, which, when executed by a processor, cause the processor to perform the metal particle overlap detection method as described in any one of claims 1 to 3.

Citation Information

Patent Citations

  • Laser sensor for particle size detection

    CN108027312A

  • Signal demodulation identification circuit and method for micron-sized metal abrasive particles in oil

    CN122130561A