A laser particle size analyzer inverse fourier transform particle cross detection method and system

Abstract

The present application relates to the technical field of particle size analysis, solves the technical problem of signal distortion in rod-shaped abrasive particle size monitoring in a strong magnetic field environment, and particularly relates to a laser particle size analyzer inverse Fourier transform particle cross detection method and system, comprising the following steps: collecting original scattered light intensity signals and magnetic field strength data, calculating a magnetic field change gradient based on the magnetic field strength data; using the magnetic field change gradient to calculate a predicted phase offset for eliminating signal distortion caused by sudden changes in rotational speed; and identifying an oscillation risk period based on the magnetic field change gradient and a preset risk determination threshold. The present application breaks through the assumption of spherical particles by traditional Mie theory, establishes a scattering kernel function for rod-shaped particles based on T-matrix theory, and introduces a magnetic field coupling term to quantify the influence of the magnetic field on particle orientation and scattering signals. The scattering double-peak artifact caused by directional arrangement of rod-shaped particles in a strong magnetic field is avoided, single particles are not misjudged as mixed particles of multiple sizes, and the particle size inversion accuracy is significantly improved.

Description

Technical Field

[0001] This invention relates to the field of particle size analysis technology, and in particular to a laser particle size analyzer inverse Fourier transform particle cross-detection method and system. Background Technology

[0002] The inverse Fourier transform particle cross-detection method in laser particle size analyzers is a highly efficient particle size analysis technique based on light scattering theory. Its technical principle is as follows: after laser irradiation of the sample, the angular distribution of the scattered light is mapped into a spatial distribution signal through a Fourier lens group; multiple quadrant detectors simultaneously collect the scattered light intensity at different spatial locations; then, an inverse Fourier transform is performed on the scattered light intensity, and the particle size distribution is solved using the Mie scattering theory model. The core advantages of this method are its speed and wide dynamic range, and it has been widely applied in chemical, pharmaceutical, and other fields.

[0003] However, existing laser particle size analyzers rely on a fundamental assumption: that the particles are ideally spherical. This assumption stems from the mathematical foundation of Mie theory—the analytical solution of Maxwell's equations at spherical boundaries. There is no analytical solution for the scattering behavior of non-spherical particles, requiring numerical approximations (such as the T-matrix method). However, due to high computational complexity, this approach cannot be integrated into real-time particle size analyzers. Therefore, commercial devices forcibly map the scattering signals of non-spherical particles to an equivalent spherical diameter, leading to systematic distortion in certain scenarios.

[0004] Especially in the scenario of abrasive monitoring of magnetohydrodynamic (MHD) seals, it is necessary to monitor the iron oxide (Fe3O4) wear debris in the hydrocarbon-based carrier fluid in real time and to provide early warning of seal failure based on particle size growth. Wear debris is typically rod-shaped particles with an aspect ratio of 3:1-10:1 (material wear generates flaky fragments; the magnetic field orientation and the antimagnetic force of the hydrocarbon-based carrier fluid combine to stretch and lock the rod-shaped morphology). When the equipment generates a strong magnetic field (>0.5T), the wear debris particles align along the magnetic field lines (long axis parallel to the magnetic field), and their scattered signals form a bimodal distribution at the equatorial plane. However, the inversion algorithm forcibly fits non-spherical signals to a spherical model, causing a single wear debris particle to be misclassified as a mixture of particles of various sizes, for example, disassembling a 10μm wear debris particle into 5μm and 20μm spherical particles, leading to system warning failure. Furthermore, in high-speed rotating equipment (such as turbines, pumps, compressors, etc.), the magnetic moment direction of the wear particles oscillates violently when the rotational speed changes suddenly, and the intensity of scattered light shifts at high frequency. This causes the phase shift signal of the scattered light to jump at high frequency, leading to the dynamic shift being misinterpreted as a bimodal distribution, which triggers a system warning error. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a laser particle size analyzer inverse Fourier transform particle cross-detection method and system, which solves the technical problem of signal distortion in rod-shaped grinding debris particle size monitoring under strong magnetic field conditions, and achieves the purpose of real-time and accurate inversion and early warning of the true size of grinding debris particles.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a laser particle size analyzer inverse Fourier transform particle cross-detection method, the method comprising the following steps:

[0007] The system collects raw scattered light intensity signals and magnetic field strength data, calculates the magnetic field change gradient based on the magnetic field strength data, calculates the predicted phase shift to eliminate signal distortion caused by sudden changes in rotation speed using the magnetic field change gradient, and identifies oscillation risk periods based on the magnetic field change gradient and a preset risk judgment threshold. The raw scattered light intensity signals in the oscillation risk periods are marked as transient signals, and others are marked as steady-state signals.

[0008] The morphological characteristics of grinding debris particles and the magnetic field coupling effect are used to correct the bimodal pseudo-signal of the steady-state signal caused by the directional arrangement of the magnetic field; the phase jump caused by mechanical interference is extracted from the transient signal through time-frequency analysis; and the steady-state signal and the transient signal are weighted and fused into a composite signal.

[0009] During periods of oscillation risk, the magnetic field interference component is removed from the composite signal to obtain the de-oscillation phase domain signal; time-frequency analysis is performed on the de-oscillation phase domain signal to determine the bimodal characteristic parameters and calculate the true scattering characteristics; based on the results of the true scattering characteristics, the morphological characteristics of the wear particles are inverted, and the equipment wear warning is triggered based on the inversion results.

[0010] Furthermore, the method of correcting the bimodal pseudo-signal of the steady-state signal caused by the directional alignment of the magnetic field through the morphological characteristics of the wear particles and the magnetic field coupling effect includes:

[0011] Based on the image of the grinding particles, the morphological features of the grinding particles are obtained, the main axis of the grinding particle shape is identified, and the lengths of the major and minor axes are determined.

[0012] Establish the scattering kernel function for non-spherical wear debris particles:

[0013] ;

[0014] In the formula, The scattering kernel function; The long axis of the grinding particles; For the short axis of the grinding debris particles; The scattering angle; y is the scattering amplitude; y is the wavenumber of the propagating wave; The wavelength of light;

[0015] Introducing a magnetic coupling term to correct scattering intensity:

[0016] ;

[0017] In the formula, The modified scattering kernel function; This represents the scattering kernel function calculated under conditions without a magnetic field; This is the shape correction factor; is Boltzmann constant; T is absolute temperature; H is external magnetic field; Material parameters;

[0018] Decompose the steady-state signal into morphological feature contributions and noise:

[0019] ;

[0020] m represents the number of categories of wear particles with different morphological characteristics; The weight corresponding to the i-th type of wear particles; This represents the scattering intensity of the i-th type of wear particles; For noise;

[0021] Calculate the second derivative that reveals the concavity and convexity of the steady-state signal and identify the number of extreme points of the second derivative; if the second derivative has two extreme points near the equatorial plane, it indicates that the steady-state signal is a multi-particle signal; if the second derivative has only one extreme point near the equatorial plane, it is determined to be a single-particle signal.

[0022] Furthermore, the step of extracting the phase jump from the transient signal through time-frequency analysis to compensate for the mechanical interference includes:

[0023] The transient signal is divided into preset time windows, and a Fourier transform is performed on the signal within each time window, i.e.:

[0024] ;

[0025] In the formula, It is a time-frequency plane; This is an expression for the window function; For frequency, The time delay is denoted by j; j is the imaginary unit.

[0026] Find each delay The frequency with the strongest energy, that is:

[0027] ;

[0028] In the formula, The main frequency;

[0029] The extracted main frequency is matched with the frequency of equipment rotation speed change. If the two frequency changes match, the phase information is extracted, i.e.:

[0030] ;

[0031] In the formula, This indicates a phase jump caused by mechanical interference; The phase angle represents the dominant frequency.

[0032] Furthermore, the method for removing the magnetic field interference component from the composite signal is as follows:

[0033] ;

[0034] In the formula, To remove the oscillating phase domain signal; It is the fundamental wave function of the oscillation; This is a scaling factor used to adjust the amplitude of the compensation signal. It is obtained by calibrating the historical oscillation data of the device to ensure that the magnitude of the interference signal deducted is appropriate and that the amplitude of the compensation signal matches that of the actual interference. The oscillation frequency modulation factor. The formula for predicting the phase offset is as follows:

[0035] ;

[0036] Where k is the magneto-optical coupling coefficient; The gradient of the magnetic field change; For time scale.

[0037] Furthermore, the method for calculating the true scattering characteristics is as follows:

[0038] The scale range of the wavelet transform is set based on the peak width W of the oscillating phase domain signal;

[0039] Perform a discrete wavelet transform on the oscillating phase domain signal and calculate the wavelet coefficients, i.e.:

[0040] ;

[0041] In the formula, q is the scale; p is the translation parameter; These are wavelet coefficients; It is a wavelet function;

[0042] The wavelet coefficients at each scale are examined progressively within the scale range of the wavelet transform, and the corresponding peak sharpness is calculated, i.e.:

[0043] ;

[0044] In the formula, For peak resolution; This represents the maximum value of the wavelet coefficients; This represents the minimum value of the wavelet coefficients; The variance of the wavelet coefficients;

[0045] The scale that maximizes peak sharpness is selected as the final chosen scale, and the local maxima of the wavelet coefficients at this scale are calculated to determine the significant peaks; that is:

[0046] ;

[0047] In the formula, A set of significant peaks; The peak value filtering threshold;

[0048] Two strongest local maxima are selected from the set of significant peaks and defined as bimodal peaks; the signal strength of the two peaks is compared to calculate the intensity ratio, i.e.:

[0049] ;

[0050] In the formula, Strength ratio; Indicates the position angle Signal strength at the location; Indicates the position angle Signal strength at the location;

[0051] The bimodal angular distance, used to accurately identify the morphological characteristics of wear debris particles and improve inversion accuracy, is calculated by the position angles of the two peaks.

[0052] ;

[0053] In the formula, The angular distance between the two peaks, the position angle This indicates the angular position of the first peak, position angle. This indicates the angular position of the second peak.

[0054] Furthermore, the method for calculating the true scattering characteristics is as follows:

[0055] ;

[0056] In the formula, This represents the reconstructed true scattered signal intensity distribution; This is the convolution operator; It is a unit impulse function.

[0057] Furthermore, the morphological characteristics of the wear particles inverted based on the results of real scattering features include:

[0058] ;

[0059] In the formula, Indicates the scattering angle of the wear particles. The scattering intensity at point n; n is the order of the scattering mode; and It is a correlation coefficient related to the ratio of the major and minor axes of the wear particles and the scattering angle; and This is the relationship function related to the cosine value of the scattering angle;

[0060] The reconstructed true scattered signal intensity distribution and scattered intensity are fitted using the linear least squares method to invert the major and minor axes of the wear debris particles, i.e.:

[0061] ;

[0062] In the formula, C is the calibrated concentration correction factor.

[0063] Furthermore, the method for triggering equipment wear warning based on the inversion results is as follows:

[0064] Calculate the equivalent volume diameter of the wear debris particles:

[0065] ;

[0066] In the formula, The equivalent volume diameter of the grinding debris particles; This represents the inversion value of the long axis of the grinding debris particles; This represents the short-axis inversion value of the wear debris particles;

[0067] The equivalent volume diameter of the wear particles is compared with a preset volume threshold. When the equivalent volume diameter exceeds the volume threshold N times consecutively, an equipment wear warning is triggered.

[0068] Furthermore, the magneto-optical coupling coefficient is determined as follows:

[0069] Calculate the electric field intensity when the laser beam reaches the surface of the grinding debris particles, i.e.:

[0070] ;

[0071] In the formula, The electric field strength of the laser beam propagating onto the surface of the grinding particles; The initial electric field strength; denoted by , where is the absorption coefficient of the hydrocarbon-based carrier liquid; z is the propagation distance.

[0072] The total scattering intensity is calculated by integrating the scattering modes, i.e.:

[0073] ;

[0074] In the formula, For grinding particles at a specific scattering angle The total field intensity below; Let be the distribution function of the grinding debris particles; The scattering mode is related to the interaction intensity between the grinding particles and the incident laser beam;

[0075] The magneto-optical coupling coefficient is calculated based on the total scattering intensity and the electric field intensity, i.e.:

[0076] ;

[0077] In the formula, It is the vacuum permittivity; The incident field intensity; The scattered field intensity is calculated using the following formula:

[0078] ;

[0079] in, To monitor the volume of the container; Let V be the volume of the grinding debris particles.

[0080] A laser particle size analyzer inverse Fourier transform particle cross-detection system includes:

[0081] The signal acquisition module is used to acquire laser scattering signals and equipment operating parameters;

[0082] The signal preprocessing module is used to classify the initial scattered light intensity distribution and eliminate environmental noise;

[0083] The signal correction module is used to correct the bimodal artifact signal in response to the magnetic field orientation effect of rod-shaped particles.

[0084] The phase compensation module is used to eliminate the phase shift of magnetic moment oscillation caused by sudden changes in rotational speed.

[0085] The signal reconstruction module is used to remove magnetic field artifacts and extract the true scattering characteristics of particles;

[0086] The particle inversion and early warning module is used to invert the geometric size of wear debris particles to achieve wear early warning.

[0087] By employing the above technical solution, the present invention provides a method and system for detecting cross-particle movement using inverse Fourier transform in a laser particle size analyzer, which has at least the following beneficial effects:

[0088] 1. This invention breaks through the traditional Mie theory's assumption of spherical particles, establishes a scattering kernel function for rod-shaped particles based on T-matrix theory, and introduces a magnetic field coupling term to quantify the influence of the magnetic field on particle orientation and scattering signal. It solves the problem of bimodal scattering artifacts caused by the directional alignment of rod-shaped particles under strong magnetic fields, avoids misclassifying single particles as multi-sized mixed particles, and significantly improves the accuracy of particle size inversion.

[0089] 2. This invention extracts the magnetic moment oscillation frequency caused by sudden changes in rotational speed through time-frequency analysis, dynamically compensates for phase jumps, and fuses steady-state and transient signals to generate an anti-interference composite signal. This eliminates high-frequency jumps in the scattered signal caused by sudden changes in equipment rotational speed, prevents the misinterpretation of dynamic phase shifts as sudden changes in particle size, and improves the system's stability under transient conditions.

[0090] 3. This invention combines wavelet transform and magnetic field gradient data to remove oscillation artifacts, extract the true bimodal angular distance, and reconstruct the essential scattering distribution using a unit impulse function. It accurately separates magnetic field interference from the true particle signal, and inversely derives the long axis. and short axis The error was reduced to less than 5%, providing a reliable data basis for wear analysis. Attached Figure Description

[0091] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0092] Figure 1 This is a schematic diagram of the detection method in an embodiment of the present invention. Detailed Implementation

[0093] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. This will allow for a full understanding of how the present application uses technical means to solve technical problems and achieve technical effects, and to facilitate its implementation.

[0094] This embodiment proposes an inverse Fourier transform particle cross-detection method for laser particle size analyzers. It solves the problems of: under strong magnetic fields, the long axis of grinding particles is parallel to the magnetic field lines, resulting in an incorrect bimodal distribution of scattered signals at the equatorial plane; the inversion algorithm, due to the forced assumption of spherical particles, mistakenly interprets these signals as a mixture of particles of various sizes; and during high-speed operation, sudden changes in rotation speed cause violent oscillations in the magnetic moment direction of the particles, resulting in high-frequency phase jumps in the scattered light, leading to signal instability and interfering with the accurate measurement of particle size. By correctly decoupling the magnetic field influence and the oscillating signal, the true size parameters (long and short axes) of rod-shaped particles can be accurately identified and inverted during the analysis process, greatly reducing measurement errors caused by improper model assumptions. Figure 1 As shown, the method includes the following steps:

[0095] A correlation was established between the raw scattered light intensity signal and the equipment operating parameters to filter out the effective signal range. In the context of magnetohydrodynamic (MHD) seals, sudden changes in equipment rotation speed lead to drastic changes in magnetic field strength. The magnetic moment of the abrasive particles precesses under the Lorentz force, causing high-frequency phase shifts in the scattered light. These phase shifts overlap with the actual particle scattering signal in the time domain, making them impossible to separate using traditional inversion methods. By collecting raw scattered light intensity signals and magnetic field strength data from the abrasive particle monitoring scenario of MHD seals and aligning them according to the acquisition time, a direct mapping relationship between magnetic field strength changes and scattered signals was established, resolving the phase shift problem caused by sudden changes in rotation speed. Specifically:

[0096] Based on the time-series data of magnetic field strength, the gradient of magnetic field change at each discrete time point is calculated, i.e.:

[0097] ;

[0098] In the formula, The gradient of the magnetic field reflects the instantaneous rate of change of the magnetic field. In order to be in The magnetic field strength collected at any given time, i.e., the magnetic field strength at the current moment; In order to be in The magnetic field strength collected at any given time is the magnetic field strength at the previous moment. t represents the sampling time interval; t represents the time scale.

[0099] The predicted phase offset used to eliminate signal distortion caused by sudden changes in rotational speed is calculated using the gradient of magnetic field changes and material parameters, i.e.:

[0100] ;

[0101] In the formula, To predict the phase shift, specifically the phase shift of scattered light caused by magnetic field interference, quantification is used; k is the magneto-optical coupling coefficient, with dimensions... This reflects the phase shift generated per unit change in magnetic field gradient under specific material susceptibility conditions, and is determined based on experimental data fitting. The magnetic susceptibility of the wear particles is a material parameter that reflects the intensity of the magnetization response of the wear particles under the action of a magnetic field.

[0102] Simultaneously, based on the magnetic field change gradient and a preset risk assessment threshold, the oscillation risk period is used to trigger the compensation processing window, i.e.:

[0103] ;

[0104] In the formula, Risk markers for periods of oscillation risk; The risk assessment threshold is determined using historical fault data.

[0105] when When the magnetic field oscillates violently, it enters the transient region, and the original scattered light intensity signal within this region is marked as the transient signal. Drastic changes in the magnetic field cause precession of the magnetic moment of the wear particles, resulting in phase oscillation of the scattered light. Simultaneously, the wear particles may also be subjected to other dynamic forces (such as the Lorentz force) causing them to rotate. Therefore, the transient signal is not only related to the scattering angle... The transient signal is related to the fact that it changes rapidly over time, exhibiting a dynamic process; therefore, the transient signal cannot be directly processed by the static inversion method, and dynamic compensation is required to eliminate phase distortion before inversion.

[0106] when When the magnetic field stabilizes and enters the steady-state region, the original scattered light intensity signal within this region is marked as the steady-state signal. Since the magnetic field is stable, the magnetic moment direction of the wear particles is stable, and the phase of the scattered light is not disturbed by the sudden change in the magnetic field. Therefore, the steady-state signal is only related to the scattering angle and does not change with time, and can be represented as a static distribution. Furthermore, the steady-state signal can be directly used in classical static light scattering inversion algorithms (such as Mie theory inversion) to obtain the particle size distribution of the wear particles.

[0107] Using known material parameters and particle shape models of the wear debris, the bimodal pseudo-signal caused by the directional alignment of the magnetic field is corrected. Specifically:

[0108] Images of abrasive particles in a magnetic fluid seal abrasion monitoring scenario are acquired, and the morphological characteristics of the abrasive particles are obtained. The main axis of the abrasive particle shape is identified, and the lengths of the major and minor axes are determined.

[0109] Based on the T-matrix theory, the wear particles are considered as rotating ellipsoids, and the relationship between the morphological characteristics of the wear particles and the scattering intensity is established, namely:

[0110] ;

[0111] In the formula, Let be the scattering kernel function, representing the scattering angle. The morphological feature is the long axis. short axis The scattering intensity of the wear debris particles; The major axis of the wear particle is the straight line at the farthest point between the two ends of the wear particle, which corresponds to the main extension direction of the wear particle. In non-spherical wear particles, the value of the major axis directly affects the angular distribution of scattered light. The short axis of the wear particles is perpendicular to the long axis and points to the shortest distance between the two ends in the cross section. It can provide information about the influence of the wear particle shape on the rheological properties. The scattering angle is the angle between the incident laser beam and the direction of the scattered light. This angle affects the intensity and distribution of the scattering of the wear particles and is usually used to define different scattering modes. Let be the scattering amplitude, which is related to the geometry and physical properties of the grinding particles. It represents the complex intensity of the scattered light at a specific angle, and its expression is:

[0112] ;

[0113] in, Let be a complex exponential function; y is the wave number of the propagating wave, calculated using the following formula:

[0114] ;

[0115] in, The wavelength of light is denoted by the wave number, which determines the propagation characteristics of light waves in space.

[0116] The scattering kernel function is used to critically calculate the scattering behavior of wear particles under different conditions (including shape and magnetic field). Subsequent magnetic field coupling corrections are all based on this fundamental model.

[0117] The scattering kernel function under no magnetic field influence is modified by introducing a magnetic coupling term. If the abrasive particles have good magnetism, their orientation will adjust under the influence of an external magnetic field, thereby changing their optical properties during scattering. That is:

[0118] ;

[0119] In the formula, The modified scattering kernel function describes the scattering intensity of grinding particles under the influence of an external magnetic field H. This represents the scattering kernel function calculated under conditions without a magnetic field; This is a magnetic field correction term, used to account for the influence of an external magnetic field on the scattering behavior of wear debris particles; where, This is the shape correction coefficient, obtained through experimental calibration, used to adjust the degree of influence of the magnetic field; is Boltzmann's constant, which connects temperature and energy and is commonly used in statistical physics; T is absolute temperature, expressed in Kelvin, and characterizes the degree of thermal motion of a material.

[0120] steady-state signal The contribution and noise of the scattered signal are decomposed into wear debris particles with different morphological characteristics, thereby clarifying the influence of each type of wear debris particle on the overall signal, i.e.:

[0121] ;

[0122] In the formula, m represents the number of different types of wear particles with different morphological characteristics involved in the calculation; The weight corresponding to the i-th type of wear particles represents the relative density of wear particles with this morphological characteristic in the whole; This represents the scattering intensity of the i-th type of wear particles; Let be the long axis of the i-th type of wear particles; Let be the short axis of the i-th type of wear particles; It is noise.

[0123] This decoupling process allows us to more accurately analyze the contributions of wear particles with different time periods and morphological characteristics in the original scattered light intensity signal, thereby improving overall accuracy and reducing errors caused by signal overlap. Furthermore, separating noise components from the total signal effectively suppresses random fluctuations caused by external interference or wear particle movement, making subsequent analysis more stable and reducing misjudgments. This provides a reasonable data foundation for subsequent inversion of wear particle morphology characteristics.

[0124] Calculating the second derivative reveals the concavity and convexity of steady-state signals (i.e., rapidly changing regions), which shows significant differences between single-particle and multi-particle signals. The nature of the signal is determined by identifying the number of extrema of the second derivative of the steady-state signal. If the second derivative of the steady-state signal has two extrema near the equatorial plane, it indicates a multi-particle signal with two scattering peaks; if the extrema are smooth and there is only one, it is identified as a single-particle signal. This distinguishes single-particle and multi-particle signals, ensuring the accuracy and denoising capabilities of subsequent analysis.

[0125] To eliminate phase shifts and high-frequency noise caused by magnetic moment oscillations of grinding debris particles due to sudden changes in rotational speed, the reconstructed true scattered signal intensity distribution is obtained. Specifically:

[0126] The transient signal is divided into preset time windows, for example, each time window is: Perform a Fourier transform on the signal within each time window to obtain its frequency domain representation, i.e.:

[0127] ;

[0128] In the formula, is the time-frequency plane, is the Fourier transform result of the signal, is a complex number that describes the characteristics of the signal at a given frequency and time delay; An expression for a window function (such as the Hanning window), a function used to smooth a signal; For frequency, is the time delay, representing the central function of the window function; j is the imaginary unit, used to construct the complex space.

[0129] By analyzing the spectrum, we can find each time delay. The frequency with the strongest energy, that is:

[0130] ;

[0131] In the formula, The primary frequency indicates the time delay. At that time, frequency The main frequency.

[0132] The extracted primary frequency is matched with the frequency of equipment speed change acquired by the speed sensor. If the two frequency changes match, it indicates that the frequency component is caused by mechanical interference from the equipment.

[0133] Extract phase information from the main frequencies caused by mechanical interference from the equipment, and identify phase jumps and nonlinear interference components caused by mechanical interference. That is:

[0134] ;

[0135] In the formula, This indicates a phase jump caused by mechanical interference; The phase angle represents the dominant frequency.

[0136] By compensating for the phase jump of transient signals, unnecessary interference caused by mechanical oscillations can be corrected, that is:

[0137] ;

[0138] In the formula, The compensated transient signal removes the phase shift caused by mechanical interference or other factors, thus making the signal closer to its true characteristics.

[0139] Steady-state signals represent the particle size distribution of abrasive particles under normal operating conditions, while transient signals capture instantaneous characteristics caused by changes in rotational speed, etc. Combining the two provides more comprehensive information, reflecting the true characteristics of abrasive particles under actual operating conditions. In subsequent particle size inversion, without merging the corrected signals, each signal only provides partial information. That is, steady-state signals may fail to reflect instantaneous changes and unpredictable events, while transient signals may have their true characteristics obscured by various noise effects. Relying solely on a single signal for inversion can lead to biased results, especially when transient signals fluctuate significantly or have poor quality, making it impossible to accurately capture particle characteristics. By weighted merging of the two signals to obtain a composite signal, the stability and reliability of the signal can be further ensured. Specifically, the merging method for the compensated transient and steady-state signals is as follows:

[0140] The compensated transient and steady-state signals are aligned according to the acquisition timestamps. For signals that cannot be directly aligned, interpolation algorithms (such as linear interpolation or spline interpolation) are used to adjust them to the same time axis, ensuring that the corresponding signal value is obtained at each time point. The compensated transient and steady-state signals are then weighted and fused according to preset weights, i.e.:

[0141] ;

[0142] In the formula, The resulting composite signal; and The preset weights are satisfied. Calibration based on experimental data.

[0143] During periods of oscillation risk marked by risk, spurious signal components caused by the magnetic field are removed from the composite signal; this avoids misinterpreting sudden changes in rotational speed as a change in particle size due to high-frequency peak shifts in the composite signal caused by magnetic moment oscillations.

[0144] ;

[0145] In the formula, The signal in the deoscillation phase domain, i.e. the essential scattered light intensity distribution after stripping away the transient interference of the magnetic field, is a three-dimensional tensor, a two-dimensional matrix composed of scattering angle and time, superimposed with the real part of the complex signal (describing the intensity distribution of scattering from the wear particles, which is directly related to the particle size and shape) and the imaginary part (characterizing the phase information of the light wave that has suppressed magneto-induced phase jumps). This is the fundamental oscillation function, used to restore the time-varying periodicity of the interference signal; This is a scaling factor used to adjust the amplitude of the compensation signal. It is obtained by calibrating the historical oscillation data of the device to ensure that the magnitude of the interference signal deducted is appropriate and that the amplitude of the compensation signal matches that of the actual interference. The oscillation frequency modulation factor, measured in rad / s, determines the oscillation frequency of the compensation signal.

[0146] Extract the peak width W of the deoscillation-free phase domain signal to determine the wavelet transform scale range, and set the wavelet transform scale range based on the peak width W, i.e.:

[0147] ;

[0148] in, This represents the lower limit of the scale range for wavelet transform. This represents the upper limit of the scale range for wavelet transform.

[0149] The deoscillating phase domain signal is transformed to the time-frequency domain using discrete wavelet transform, and the fine signal structure is obtained through the Daubechies wavelet function, i.e.:

[0150] ;

[0151] In the formula, q is the scale, which controls the degree of expansion of the wavelet function; p is the translation parameter, which controls the position of the wavelet function and can locate a specific time point of the signal. These are wavelet coefficients, representing the characteristic intensity of the deoscillation-free phase domain signal at a specific location under a specified scale and translation parameters. They reflect the details of the deoscillation-free phase domain signal at that scale and location, and are the representation of the signal in the time-frequency domain. It is a wavelet function.

[0152] The wavelet coefficients at each scale are examined progressively within the scale range of the wavelet transform, and the corresponding peak sharpness is calculated, i.e.:

[0153] ;

[0154] In the formula, Peak resolution represents the signal sharpness metric for a specific scale; Given a scale and all translation parameters, the maximum value of the wavelet coefficients; Given a scale and all translation parameters, the minimum value of the wavelet coefficients; The variance of the wavelet coefficients is used to quantify the distribution range of the wavelet coefficients and reflect the randomness and fluctuation degree in the signal.

[0155] The scale that maximizes peak sharpness is selected as the final chosen scale, and the local maxima of the wavelet coefficients at this scale are calculated to determine possible significant peaks; that is:

[0156] ;

[0157] In the formula, It is a set of significant peaks, including possible locations of local maxima and their corresponding times; A peak selection threshold is set; only when the absolute value of the wavelet coefficient exceeds this threshold can the time point be considered a potential peak. This threshold is determined based on fitting experimental data.

[0158] From the set of significant peaks, select the two strongest local maxima, define them as a double peak on the equatorial plane, and determine the position angles of the two peaks; compare the signal strengths of the two peaks and calculate the intensity ratio, i.e.:

[0159] ;

[0160] In the formula, Strength ratio; Indicates the position angle Signal strength at the location; Indicates the position angle Signal strength at that location.

[0161] The bimodal angular distance, used to accurately identify the morphological characteristics of wear debris particles and improve inversion accuracy, is calculated by the position angles of the two peaks.

[0162] ;

[0163] In the formula, The bimodal angular distance represents the angular difference between two peaks identified in the deoscillating phase domain signal at the equatorial plane. It is an important parameter used to assess particle morphology, especially in rod-shaped particles that exhibit a bimodal distribution; position angle. The position angle represents the angular location of the first peak, i.e., the angle of the first local extremum point of the oscillating phase domain signal. It typically corresponds to a small scattering angle, indicating the scattering characteristics of the wear particles in a certain direction; position angle. The angle of the second peak is the angle of the second local extremum of the oscillating phase domain signal. It usually corresponds to a large scattering angle and is generally related to the geometry and motion characteristics of the wear particles.

[0164] The bimodal angular offset used to correct measurement errors caused by the magnetic field is calculated using the gradient of the changing magnetic field, material parameters, and the ratio of the major and minor axes of the wear debris particles.

[0165] ;

[0166] In the formula, The bimodal angular offset is the angular difference caused by the magnetic field, which is used to quantify the effect of changes in the shape of wear debris particles on the deoscillation phase domain signal.

[0167] Based on the bimodal angular distance and its offset, and given the magnetic field, the true bimodal angular distance is obtained by subtracting the artifact offset caused by the magnetic field from the measured bimodal angular distance.

[0168] ;

[0169] In the formula, This represents the true bimodal angular distance.

[0170] The true scattered signal intensity distribution is extracted by multiplying the unit impulse function with the deoscillated phase domain signal, i.e.:

[0171] ;

[0172] In the formula, The reconstructed true scattering signal intensity distribution provides the true scattering characteristics of the corrected and cleaned wear debris particles, providing a reliable data foundation for subsequent analysis and particle performance evaluation; This is the convolution operator; The unit impulse function represents the scattering angle at a specific scattering angle. The response at a given point is 1, while at other points it is 0. By setting a specific scattering angle, the unit impulse function ensures that only signal information related to the true bimodal angular distance is retained, helping to extract the true scattering signal information and effectively isolating the influence of other angles.

[0173] The scattering intensity for inverting the geometry and size of wear debris particles is calculated using the improved T-matrix method based on scattering theory, i.e.:

[0174] ;

[0175] In the formula, Indicates the scattering angle of the wear particles. The scattering intensity at a certain angle depends on the geometry (major and minor axes) of the wear particles, reflecting the scattering characteristics of the wear particles at a specific angle, and is used for subsequent inversion of the particle geometry and size; n is the order of the scattering mode, which increases from 1 upwards, covering all possible scattering modes, and thus can take into account all complex scattering behaviors in the calculation of scattering intensity. This is a weighting factor used to adjust the contribution of different scattering modes to the final scattering intensity; and It is a correlation coefficient related to the ratio of the major and minor axes of the wear particles and the scattering angle, used to describe the scattering intensity response under different modes, and determined by fitting experimental data; and The relationship function is related to the cosine value of the scattering angle, and is usually related to the phase of the light wave and the scattering direction. It is determined by fitting experimental data.

[0176] The reconstructed true scattered signal intensity distribution and scattered intensity are fitted using linear least squares method to optimize the major and minor axes of the inverted wear debris particles, i.e.:

[0177] ;

[0178] In the formula, C is the calibrated concentration correction factor; This represents the inversion value of the long axis of the grinding debris particles; This represents the inversion value of the minor axis of the wear particles. By minimizing the error between the actual scattered signal and the theoretical model, the geometric characteristics of the wear particles, especially the specific values ​​of the major and minor axes, are accurately inverted.

[0179] Based on the geometric characteristics of the wear particles obtained through inversion, the equivalent volume diameter of the wear particles is calculated, i.e.:

[0180] ;

[0181] In the formula, The equivalent volume diameter of the wear particles is used to assess the impact of wear particles on the wear process.

[0182] By comparing the equivalent volume diameter of the wear particles with a preset volume threshold, a wear warning is triggered when the equivalent volume diameter exceeds the volume threshold N times consecutively. This enables real-time monitoring of the wear process, timely feedback on equipment operating status, reduces the risk of equipment damage, and ensures safe operation.

[0183] The long axis inversion values ​​of the wear particles were obtained by inversion. and minor axis inversion value This feedback is incorporated into the calculation formula for the bimodal angular offset, forming a closed-loop correction that effectively reduces the impact of the morphological feature assumption error of the wear particles on the intensity distribution of the reconstructed true scattered signal.

[0184] It is important to note that accurate calculation of the magneto-optical coupling coefficient is crucial in the scenario of abrasive monitoring of magnetohydrodynamic (MHD) seals. Traditional methods rely on experimental data to determine this coefficient, but this can introduce errors and is unsuitable for rapidly changing operating conditions and variable material properties. Therefore, based on a combination of electromagnetic field theory and particle scattering models, the magneto-optical coupling coefficient is dynamically calculated, thereby improving the prediction accuracy of scattering signals in abrasive monitoring of MHD seals and enhancing early warning capabilities. Specifically:

[0185] A coupled model of the laser field and magnetic field was established using the finite element method. Boundary conditions were set according to the instrument's geometry, including the incident angle of the laser beam, the position and orientation of the wear particles in the liquid, and the externally applied magnetic field. Based on the known initial electric field strength, absorption coefficient, and propagation distance of the laser beam in the hydrocarbon-based carrier liquid, the electric field strength when the laser beam reaches the surface of the wear particles was calculated, i.e.:

[0186] ;

[0187] In the formula, The electric field intensity of the laser beam propagating to the surface of the grinding particles is the intensity of the incident light after passing through the hydrocarbon-based carrier liquid. It is used to describe the attenuation of the laser beam as it propagates in the hydrocarbon-based carrier liquid. It can help analyze the changes in electric field intensity at different depths and provide a basis for subsequent analysis of the scattering behavior of grinding particles. The initial electric field strength is the electric field strength measured at the laser source. is the absorption coefficient of the hydrocarbon-based carrier fluid; z is the propagation distance, representing the distance the laser beam travels from the laser source to the grinding particles.

[0188] Then, by calculating the total scattering intensity at a specific scattering angle using the particle scattering intensity and scattering mode of wear debris particles with different equivalent volume diameters, the necessary scattering characteristic data is provided for solving the magneto-optical coupling coefficient, thereby improving the accuracy and reliability of wear particle monitoring in magnetohydrodynamic seals. That is:

[0189] ;

[0190] In the formula, For grinding particles at a specific scattering angle The total field intensity below; The distribution function of the wear particles describes the relative number of wear particles with different equivalent volume diameters. The scattering modes are related to the interaction intensity between the grinding particles and the incident laser beam. The scattering modes at the scattering angle are described after the incident laser beam interacts with the grinding particles of equivalent volume diameter.

[0191] The magneto-optical coupling coefficient is calculated based on the total scattering intensity and the electric field intensity, i.e.:

[0192] ;

[0193] In the formula, It is the vacuum permittivity; The incident field intensity; The scattered field intensity is calculated using the following formula:

[0194] ;

[0195] in, To monitor the volume of the container; The volume of the grinding debris particles is calculated using the following formula:

[0196] ;

[0197] The magneto-optical coupling coefficient is a crucial parameter describing the strength of the interaction between electromagnetic waves and magnetic particles. By accurately calculating the magneto-optical coupling coefficient, the coupling effect between the laser beam and the magnetic field can be quantified, thereby determining the response of wear particles to the laser. Furthermore, an accurate magneto-optical coupling coefficient can improve the inversion of wear particle size and morphology, ensuring that the monitored wear particle characteristics are more realistic and reliable under changing environmental conditions; this is essential for improving the accuracy of wear prediction and fault early warning.

[0198] This embodiment also proposes a laser particle size analyzer inverse Fourier transform particle cross-detection system, including a signal acquisition module, a signal preprocessing module, a signal correction module, a phase compensation module, a signal reconstruction module, and a particle inversion and early warning module.

[0199] The signal acquisition module is used to collect laser scattering signals and equipment operating parameters, providing raw data for subsequent analysis. It emits a monochromatic laser beam to illuminate the wear debris particles under test, collects light intensity signals at different scattering angles, and outputs the raw scattered light intensity distribution. A magnetic field sensor monitors the magnetic field strength during equipment operation to correlate magnetic field changes with scattering signal distortion. A speed sensor also collects equipment rotation speed data to identify transient interference caused by sudden changes in rotation speed.

[0200] The signal preprocessing module is used to classify the initial scattered light intensity distribution and eliminate environmental noise. By calculating the magnetic field gradient and marking periods of oscillation risk, the original scattered light intensity distribution is classified into steady-state and transient signals.

[0201] The signal correction module is used to correct the bimodal artifact signal caused by the magnetic field orientation effect of rod-shaped particles. Based on T-matrix theory, a scattering kernel function for non-spherical particles is established, and magnetic susceptibility and magnetic field strength are introduced to calculate the corrected scattering intensity. Then, the steady-state signal is decomposed into morphological contribution and noise.

[0202] The phase compensation module is used to eliminate the phase shift of magnetic moment oscillations caused by sudden changes in rotational speed. It extracts the dominant frequency by performing a short-time Fourier transform on the transient signal and calculates the predicted phase shift to compensate for the transient signal. The steady-state and compensated transient signals are then weighted and combined to generate a composite signal.

[0203] The signal reconstruction module is used to remove magnetic field artifacts and extract the true scattering characteristics of particles. It dynamically removes oscillation interference based on the magnetic field gradient, then extracts the bimodal angular distance at the equatorial plane using wavelet transform to distinguish between single-particle and multi-particle signals. Finally, it uses a unit impulse function to extract the true scattering distribution and outputs the reconstructed true scattering signal.

[0204] The particle inversion and early warning module is used to invert the geometric dimensions of wear particles and achieve wear early warning. It fits the major and minor axes based on an improved T-matrix method and calculates the equivalent volume diameter. If the equivalent volume diameter exceeds a threshold N times consecutively, an equipment wear alarm is triggered.

[0205] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0206] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. Since the above embodiments are substantially similar to the method embodiments, their descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0207] The above embodiments provide a detailed description of the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for detecting cross-particle movement using inverse Fourier transform in a laser particle size analyzer, characterized in that, The method includes the following steps: Raw scattered light intensity signals and magnetic field strength data are acquired. The magnetic field gradient is calculated based on the magnetic field strength data. This gradient is then used to calculate the predicted phase shift used to eliminate signal distortion caused by sudden changes in rotational speed. ； In the formula, To predict the phase offset; The gradient of the magnetic field change; k is the magneto-optical coupling coefficient; This refers to material parameters, specifically the magnetic susceptibility of the wear debris particles. Based on the gradient of magnetic field change and the preset risk judgment threshold, the oscillation risk period is identified, and the original scattered light intensity signal in the oscillation risk period is marked as transient signal, and others are marked as steady-state signal; The bimodal signal of the steady-state signal caused by the directional alignment of the magnetic field is corrected by utilizing the morphological characteristics of the wear particles and the magnetic field coupling effect, including: Based on the image of the wear debris particles, the morphological features of the wear debris particles are obtained, the principal axis of the wear debris particle shape is identified, and the lengths of the major and minor axes are determined; and the scattering kernel function of non-spherical wear debris particles is established, namely: ； In the formula, Let be the scattering kernel function, representing the scattering angle. The morphological feature is the long axis. short axis The scattering intensity of the wear debris particles; The scattering angle is the angle between the incident laser beam and the direction of the scattered light. The scattering amplitude represents the complex intensity of the scattered light at a specific angle; The wave number of the propagating wave, , The wavelength of light; Then, a magnetic coupling term is introduced to correct the scattering intensity in the scattering kernel function, that is: ； In the formula, The modified scattering kernel function; This represents the scattering kernel function calculated under conditions without a magnetic field; This is a magnetic field correction term, used to account for the influence of an external magnetic field on the scattering behavior of wear debris particles; where, This is the shape correction coefficient, obtained through experimental calibration, used to adjust the degree of influence of the magnetic field; is Boltzmann constant; T is absolute temperature in Kelvin; H is external magnetic field strength. The steady-state signal is decomposed into morphological feature contributions and noise, i.e.: ； In the formula, This represents the steady-state signal; m represents the number of different types of wear particles with varying morphological characteristics. The weight corresponding to the i-th type of wear particles; This represents the scattering intensity of the i-th type of wear particles; For noise; Calculate the second derivative that reveals the concavity and convexity of the steady-state signal and identify the number of extreme points of the second derivative; if the second derivative has two extreme points near the equatorial plane, it indicates that the steady-state signal is a multi-particle signal; if the second derivative has only one extreme point near the equatorial plane, it is determined to be a single-particle signal. Time-frequency analysis is used to extract phase jumps in transient signals used to compensate for mechanical interference, including: The transient signal is divided into preset time windows, and a Fourier transform is performed on the signal within each time window. The result of the transform is found at each time delay. The main frequency with the strongest energy ; The main frequency is matched with the frequency of equipment rotation speed variation. If the two frequency changes match, the phase information is extracted, i.e.: ； In the formula, This indicates a phase jump caused by mechanical interference; The phase angle represents the dominant frequency; by compensating for the phase jump of transient signals, unnecessary interference caused by mechanical oscillations is corrected, i.e.: ； In the formula, The compensated transient signal; It is a transient signal; The imaginary unit is used to construct the complex number space; The steady-state signal and the transient signal are weighted and fused into a composite signal; During periods of high oscillation risk, the magnetic field interference component is removed from the composite signal to obtain the de-oscillating phase domain signal, i.e.: ； In the formula, To remove the oscillating phase domain signal; It is a composite signal; It is the fundamental wave function of the oscillation; This is a scaling factor used to adjust the amplitude of the compensation signal; The oscillation frequency modulation factor determines the oscillation frequency of the compensation signal; The risk assessment threshold is determined using historical fault data. Time-frequency analysis is performed on the deoscillation phase domain signal to determine the bimodal characteristic parameters, and the true scattering characteristics are calculated, including: The scale range of the wavelet transform is set based on the peak width W of the oscillating phase domain signal; Perform discrete wavelet transform on the oscillating phase domain signal and calculate the wavelet coefficients; The wavelet coefficients at each scale are examined step by step within the scale range of the wavelet transform, and the corresponding peak sharpness is calculated. The scale that maximizes the peak sharpness is selected as the final selected scale, and the local maxima of the wavelet coefficients at that scale are calculated to determine the significant peaks. The two strongest local maxima selected from the set of significant peaks are defined as bimodal. The bi-peak angular distance, used to accurately identify the morphological characteristics of wear debris particles and improve inversion accuracy, is calculated by using the position angles of the two peaks. The bimodal angular offset used to correct measurement errors caused by the magnetic field is calculated using the gradient of the changing magnetic field, material parameters, and the ratio of the major and minor axes of the wear debris particles. ； In the formula, This represents the offset of the bi-peak angular distance. In order to be in The magnetic field strength collected at any given time, i.e., the magnetic field strength at the current moment; Based on the bimodal distance and bimodal distance offset, under the condition of a known magnetic field, the true bimodal distance is obtained by subtracting the artifact offset caused by the magnetic field from the measured bimodal distance. The true scattered signal intensity distribution is extracted by multiplying the unit impulse function with the deoscillated phase domain signal, i.e.: ； In the formula, This represents the reconstructed true scattered signal intensity distribution; This is the convolution operator; It is a unit impulse function; This represents the true bimodal angular distance. The scattering angle; The morphological characteristics of wear particles are inverted based on the results of the actual scattering features, and the equipment wear warning is triggered based on the inversion results, including: ； In the formula, Indicates the scattering angle of the wear particles. The scattering intensity at point n; n is the order of the scattering mode; and It is a correlation coefficient related to the ratio of the major and minor axes of the wear particles and the scattering angle; and This is the relationship function related to the cosine value of the scattering angle; The reconstructed true scattered signal intensity distribution and scattered intensity are fitted using the linear least squares method to invert the major and minor axes of the wear debris particles, i.e.: ； In the formula, C is the calibrated concentration correction factor; This represents the inversion value of the long axis of the grinding debris particles; This represents the short-axis inversion value of the wear debris particles.

2. The method according to claim 1, characterized in that, The method for triggering equipment wear warning based on inversion results is as follows: Calculate the equivalent volume diameter of the wear particles, compare the equivalent volume diameter of the wear particles with a preset volume threshold, and trigger an equipment wear warning when the equivalent volume diameter exceeds the volume threshold N times consecutively.

3. The method according to claim 1, characterized in that, The magneto-optical coupling coefficient is determined as follows: Calculate the electric field intensity when the laser beam reaches the surface of the grinding debris particles, i.e.: ； In the formula, The electric field strength of the laser beam propagating onto the surface of the grinding particles; The initial electric field strength; denoted by , where is the absorption coefficient of the hydrocarbon-based carrier liquid; z is the propagation distance. The total scattering intensity is calculated by integrating the scattering modes, i.e.: ； In the formula, For grinding particles at a specific scattering angle The total field intensity below; Let be the distribution function of the grinding debris particles; The scattering mode is related to the interaction intensity between the grinding particles and the incident laser beam; The magneto-optical coupling coefficient is calculated based on the total scattering intensity and the electric field intensity, i.e.: ； In the formula, It is the vacuum permittivity; The incident field intensity; The scattered field intensity is calculated using the following formula: ； in, To monitor the volume of the container; Let V be the volume of the grinding debris particles.

4. A system for implementing the laser particle size analyzer inverse Fourier transform particle cross-detection method according to any one of claims 1-3, characterized in that, include: The signal acquisition module is used to acquire laser scattering signals and equipment operating parameters; The signal preprocessing module is used to classify the initial scattered light intensity distribution and eliminate environmental noise; The signal correction module is used to correct the bimodal artifact signal in response to the magnetic field orientation effect of rod-shaped particles. The phase compensation module is used to eliminate the phase shift of magnetic moment oscillation caused by sudden changes in rotational speed. The signal reconstruction module is used to remove magnetic field artifacts and extract the true scattering characteristics of particles; The particle inversion and early warning module is used to invert the geometric size of wear debris particles to achieve wear early warning.

Images