A method for analyzing physical properties of battery electrode materials for recycling

Abstract

The application discloses a physical property analysis method for recycling battery electrode materials, comprising: preparing a recycled battery electrode material powder into a suspension and applying an alternating current field, collecting the electric rotation signal and electrophoretic movement signal of each particle; extracting the particle size and surface potential of each particle according to the two signals respectively, and statistically obtaining the particle size distribution and surface potential distribution; constructing an inter-particle interaction network atlas with each particle as a node, the particle size and surface potential as attributes, and the inter-particle interaction force as an edge; performing graph convolution propagation on the atlas, aggregating neighbor node attributes and updating node features, and generating a physical feature vector after multiple iterations and pooling; and simultaneously outputting the specific surface area and tap density through multi-task mapping of the physical feature vector. The method realizes nondestructive and rapid detection of the physical properties of recycled battery electrode materials.

Description

Technical Field

[0001] This invention belongs to the field of retired battery recycling technology, specifically relating to a method for analyzing the physical properties of battery electrode materials for recycling. Background Technology

[0002] The recycling and processing of retired lithium-ion batteries is a crucial link in the downstream of the new energy vehicle industry chain. Physical sorting of electrode materials is a core step in achieving efficient separation of positive and negative electrode active materials from current collectors, binders, and other components. The mainstream physical sorting process includes a combination of operations such as crushing, sieving, magnetic separation, electrostatic separation, and air separation. Its basic principle lies in utilizing the differences in physical properties of different material components, such as density, conductivity, magnetic susceptibility, dielectric constant, and work function. By applying an external force field, different particles are caused to move along different trajectories in space, thereby achieving separation between components. In this process chain, analyzing the physical properties of the disassembled electrode material powder is the prerequisite and basis for determining the type of sorting equipment, arranging sorting processes, and setting operating parameters.

[0003] Currently, the physical property analysis of retired battery electrode material powders in actual industrial production typically employs a combination of elemental content detection and overall property measurement. Elemental content detection mainly uses techniques such as inductively coupled plasma atomic emission spectrometry (ICP-AES) or X-ray fluorescence spectrometry to obtain the total mass percentage of elements such as cobalt, nickel, manganese, lithium, iron, phosphorus, and aluminum in the powder sample, thereby inferring the approximate ratio of positive and negative electrode materials. Overall property measurement includes powder resistivity testing, tap density determination, and particle size distribution analysis, used to characterize the overall conductivity and packing properties of the powder. After obtaining the above overall property parameters, operators select the appropriate sorting equipment and set the operating parameters based on experience or by consulting process manuals.

[0004] However, retired battery electrode material powders exhibit significant particle-scale inhomogeneity. Electrode material powders obtained from the same batch of dismantling not only contain different types of positive and negative electrode active materials, but also auxiliary components such as conductive carbon black, binder debris, and current collector micro-flakes. More importantly, even different particles of the same nominal material show significant differences in the degree of crystal structure degradation, the amount of electrolyte decomposition products deposited, and the interfacial chemical state after hundreds to thousands of charge-discharge cycles. This particle-scale inhomogeneity means that analytical results based on overall average properties cannot accurately reflect the actual behavioral response of individual particles during the physical sorting process. Specifically, in the electrostatic sorting stage, the charged behavior of particles depends on the surface work function, which is influenced by both the particle surface chemical composition and crystal structure. The difference in work function between different particles of the same material in retired electrode material powder can be significant enough to affect the sorting effect, but overall measurement methods can only obtain statistical average results, completely masking this particle-scale difference information. In a high-voltage electrostatic field, the dielectric force on a particle depends on its dielectric constant and the distribution of the electric field gradient. The dielectric constant of a particle is closely related to its local elemental composition and lattice integrity. The dispersion of dielectric constant between particles in decommissioned electrode material powder is often much greater than that in newly prepared materials, and existing analytical methods lack effective characterization methods for this.

[0005] Furthermore, existing analytical procedures lack a systematic correlation mechanism between physical property measurement results and the determination of sorting operation parameters. The setting of operation parameters typically relies on empirical rules or offline experiments, making it difficult to establish a quantitative mapping relationship from particle property distribution to sorting kinetic response. This results in a lack of targeted basis for selecting operation parameters when dealing with decommissioned electrode materials with significant property fluctuations, ultimately affecting the separation accuracy between positive and negative electrode materials and the recovery rate of target products. Summary of the Invention

[0006] In view of this, the present invention provides a physical property analysis method for recycling battery electrode materials, which aims to systematically characterize the intrinsic electrical properties of decommissioned electrode material powder at the particle scale, and to correlate particle property differences with sorting kinetics, thereby providing a refined decision-making basis for industrial sorting operation parameters.

[0007] The technical solution adopted in this invention is as follows: A method for analyzing the physical properties of recycled battery electrode materials, comprising: Step 1: Obtain the surface and interface spectral data and composition data of individual particles in the electrode material powder obtained from the dismantling of retired batteries, perform spatial registration and standardization correction, and form a multi-source characterization dataset indexed by individual particles; Step 1 specifically includes: Step 1.1: Using a microscopic hyperspectral imaging device, acquire a cube of diffuse reflectance hyperspectral image of the electrode material powder in a preset continuous spectral band; Step 1.2: Using a scanning electron microscope coupled with energy-dispersive X-ray spectroscopy, acquire backscattered electron morphology and elemental concentration distribution maps corresponding to the cubic space of the hyperspectral image; Step 1.3: Using the backscattered electron topography map as a spatial reference, perform black-and-white correction and multivariate scattering correction on the hyperspectral image cube pixel by pixel, and perform background subtraction and characteristic peak area integration on the element concentration distribution map to obtain the corrected spectral reflectance curve and element concentration vector. Step 1.4: Bind the corrected spectral reflectance curve and elemental concentration vector of a single particle according to the particle profile to form a multi-source characterization dataset indexed by a single particle.

[0008] Step 2: Perform dielectric response analysis on the spectral data of each particle in the multi-source characterization dataset, and invert the work function and static dielectric constant of the particle by combining the particle composition data, and summarize them into a set of intrinsic electrical properties of the particle. Step 2 specifically includes: Step 2.1: Perform a Kramerskrenig transform on the corrected spectral reflectance curve of each particle in the multi-source characterization dataset to extract the real part of the complex permittivity function of the particle in the preset frequency band from the spectral reflectance. and the virtual part ,in The angular frequency of the incident electromagnetic wave is given, and the real and imaginary parts constitute the particle dielectric response spectrum. Step 2.2: Locate the characteristic peak frequency position of the imaginary part from the dielectric response spectrum of the particles. ,Will The predicted work function of the particle is output by weighting the molar ratio of transition metals to alkali metals in the elemental concentration vector of the corresponding particle, inputting a pre-calibrated work function regression equation, and outputting the predicted work function of the particle. ; Step 2.3: Extract the static dielectric constant from the low-frequency asymptotic region of the particle dielectric response spectrum. The predicted work function and the static dielectric constant are used together as the electrical intrinsic properties of the particle, and the set of particle electrical intrinsic properties of all particles is summarized.

[0009] Step 3: Simulate the continuous physical process of industrial triboelectric charging and electrostatic sorting, substitute the set of intrinsic electrical properties of the particles into the contact charging equation and the electric field force equation, and calculate the charge decay curve and electric field migration deflection of each particle. Step 3 specifically includes: Step 3.1: In the coupled simulation space of discrete element method and computational fluid dynamics, a three-dimensional virtual sorting flow field is set up, which includes a triboelectric pipe section and a high-voltage electrostatic sorting section. The inner wall material of the triboelectric pipe section is set to a preset metal, and the high-voltage electrostatic sorting section is equipped with a corona electrode array and a static electrode plate. Step 3.2: Instantiate each particle in the set of intrinsic electrical properties of the particles as a discrete element particle, set the surface work function of the particle according to the predicted work function of the particle, set the dielectric property of the particle according to the static dielectric constant of the particle, and calculate the mass density according to the weighted atomic weight of each element in the element concentration vector of the particle and assign it to the particle. Step 3.3: In the triboelectrically charged pipe section, simulate the multiple contact and collision processes between the particles and the metal wall. Calculate the charge transfer amount of each collision based on the difference between the work function of the particle surface and the work function of the wall. Accumulate the initial cumulative charge of the particles before entering the high-voltage electrostatic separation section. Step 3.3 specifically includes: Step 3.31: Within the triboelectric pipe section, track the trajectory of each particle. When the distance between the particle and the metal wall is less than the preset contact detection distance, a contact collision event is determined to have occurred. Step 3.32: For each contact collision event, calculate the normal contact force based on the particle incident angle and velocity, and determine the contact area using Hertzian contact theory. ; Step 3.33: Based on the contact work function difference Calculate the charge transfer amount in a single contact: , in Predicted work function for particles With wall work function The difference, The amount of charge transferred in a single contact. The preset contact charge coefficient; And according to The sign of the charge determines the direction of charge transfer. Electrons transfer from the wall to the particle surface, causing the particle to become negatively charged. Electrons transfer from the particle surface to the wall, causing the particle to become positively charged; Step 3.34: Sum the charge transfer amounts of all contact and collision events experienced by the particles within the triboelectrically charged pipe section to obtain the initial accumulated charge of the particles before entering the high-voltage electrostatic separation section. .

[0010] Step 3.4 specifically includes: Step 3.41: Solve the Laplace equation between the corona electrode array and the static electrode plate in the high-voltage electrostatic separation section using the finite element method to obtain the electric field distribution between the electrodes. ,in and These are the spatial coordinates parallel to the static electrode plate and perpendicular to the static electrode plate, respectively. Step 3.42: Simulate the real-time charge of particles after they enter the high-voltage electrostatic separation section. The exponential decay process, according to Calculate the charge of the particle at each time step, where For particles at time Real-time charge level, The initial decay time constant, For the natural constant An exponential function with base 0; Step 3.43: Calculate the dielectric force generated by the polarization effect of particles in a non-uniform electric field:

[0011] in The dielectric force acting on the particles, The vacuum permittivity, The dielectric constant of air is The equivalent radius of the particle. The gradient is the square of the electric field strength. The Clausius Mossotti factor has a value of This reflects the effect of the dielectric difference between the static dielectric constant of particles and the dielectric constant of air on the polarization effect; Step 3.44: Calculate the Coulomb force on the particle surface charge in the electric field. Dielectrophoresis force The vector sum of the Coulomb forces is the electric force on the particle. By combining the gravitational force and fluid drag acting on the particles, the trajectory of the particles within the high-voltage electrostatic separation section is solved using Newton's equations of motion, and the deflection displacement of the particles along the electric field direction at the exit of the high-voltage electrostatic separation section is recorded. and real-time charge sequence.

[0012] Step 3.4: After the particles enter the high-voltage electrostatic separation section, the charge decay process of the particles in the corona field and the effect of the electric field force are simulated. The motion trajectory of the particles under the combined action of electric field force, gravity and fluid drag force is calculated, and the deflection displacement and real-time charge sequence of the particles at the exit of the high-voltage electrostatic separation section are recorded.

[0013] Step 4: Extract the decay time constant from the charge decay curve, perform feature cross-combination of the decay time constant and the work function in the set of intrinsic electrical properties of the particles, and perform correlation screening with the electric field migration deflection to generate a charge separation tendency vector. Step 4 specifically includes: Step 4.1: Perform exponential decay fitting on the real-time charge sequence of each particle, and extract the time constant corresponding to the time when the charge decays to a preset proportion of the initial accumulated charge. As a characteristic of charge decay; Step 4.2: Nonlinearly combine the time constant with the predicted work function of the corresponding particle to form a combined feature set, which includes various mathematical transformation combinations of the predicted work function and the time constant; Step 4.3: Calculate the Pearson correlation coefficient between each feature dimension in the combined feature set and the deflection displacement. ,reserve Feature dimensions exceeding a preset correlation threshold are concatenated with the static dielectric constant in the set of intrinsic electrical properties of the particles to form a sensitive feature subset; Step 4.4: Use the subset of sensitive features as the charge separation tendency vector of the particle.

[0014] Step 5: Project the charge separation tendency vector of all particles onto a low-dimensional decision plane, calculate the separation degree between the particle distribution areas of different material components, solve for the optimal separation line that maximizes the separation degree between adjacent material components, and extract the normal vector of the optimal separation line. Step 5 specifically includes: The charge separation tendency vectors of all particles are standardized by standard fractions according to their dimensions. The standardized vectors are then projected onto a low-dimensional decision plane composed of the first and second principal components through principal component analysis. In the low-dimensional decision plane, each scatter point is labeled according to the actual electrode material type of the particle, and the mean vector of the scatter point distribution for each material type is calculated. Covariance Matrix subscript Indicates the first Consider two classes of materials and calculate the values ​​of any two classes of materials. and Mahalanobis distance between them:

[0015] in The pooled covariance matrix is ​​the result of merging the two classes, with superscript... Indicates matrix transpose, superscript This represents finding the inverse of a matrix. Mahalanobis distance For adjacent material categories with values ​​less than a preset threshold, calculate the inter-class divergence under the Fisher linear discrimination criterion. With intra-class divergence The projection direction that maximizes the ratio ; Extract the components of each dimension of the projection direction, sort them according to the absolute value of the components, and identify the first preset number of dimensions with the largest absolute value of the components as the key feature dimensions that determine the separation effect.

[0016] Step 6: Map the normal vector of the optimal separation line back to the operating parameter space of the industrial sorting machine, determine the contribution weight of each parameter in the operating parameter space to the separation effect, and generate a parameter configuration sequence to guide physical sorting.

[0017] Step 6 specifically includes: Based on the original physical properties corresponding to the key feature dimensions, adjustable operating parameters affected by the key feature dimensions are determined in the operating parameter space of the industrial triboelectric and electrostatic separator. The adjustable operating parameters include at least the geometric dimension parameters and airflow dynamic parameters of the triboelectric section, and the electric field strength parameters of the high-voltage electrostatic separation section. Using the adjustable operating parameters as independent variables and the Mahalanobis distance between adjacent material categories in the low-dimensional decision plane as response variables, a parameter sampling scheme is generated through experimental design methods. Simulations are performed at each sampling point in the three-dimensional virtual sorting flow field to obtain the Mahalanobis distance response values ​​under each parameter combination. Response surface regression modeling is performed on the sampling results to obtain the regression mapping relationship of Mahalanobis distance with respect to the adjustable operating parameters, and the parameter combination that maximizes the regression mapping relationship is solved. The optimal parameter combination obtained from the solution is arranged into a parameter configuration sequence according to the actual sorting process sequence. The parameter configuration sequence includes the operating parameter setting values ​​of the triboelectric section and the operating parameter setting values ​​of the high-voltage electrostatic separation section, and is output to the sorting machine control system.

[0018] A physical property analysis system for recycling battery electrode materials, the system comprising: The multi-source data acquisition and registration module acquires the surface and interface spectral data and composition data of individual particles in the electrode material powder obtained from the dismantling of retired batteries, performs spatial registration and standardization correction, and forms a multi-source characterization dataset indexed by individual particles. The electrical intrinsic property inversion module performs dielectric response analysis on the spectral data of each particle in the multi-source characterization dataset, and inverts the work function and static dielectric constant of the particle by combining the particle composition data, and summarizes them into a set of particle electrical intrinsic properties. The charging behavior simulation module simulates the continuous physical process of industrial triboelectric charging and electrostatic sorting. It substitutes the set of intrinsic electrical properties of the particles into the contact charging equation and the electric field force equation to calculate the charge decay curve and electric field migration deflection of each particle. The separation tendency characterization generation module extracts the decay time constant from the charge decay curve, performs feature cross-combination of the decay time constant with the work function in the set of intrinsic electrical properties of the particles, and performs correlation screening with the electric field migration deflection to generate a charge separation tendency vector. The separation boundary decision module projects the charge separation tendency vector of all particles onto a low-dimensional decision plane, calculates the separation degree between the distribution areas of particles of different material components, solves the optimal separation line that maximizes the separation degree between adjacent material components, and extracts the normal vector of the optimal separation line. The sorting parameter generation and output module reverse maps the normal vector of the optimal separation line back to the operating parameter space of the industrial sorting machine, determines the contribution weight of each parameter in the operating parameter space to the separation effect, and generates a parameter configuration sequence to guide physical sorting.

[0019] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: 1. In this invention, by collecting surface and interface spectral data and composition data at the single-particle scale and inverting the work function and static dielectric constant of each particle, a set of intrinsic electrical properties of particles is constructed. This allows the differences in physical properties between different particles in decommissioned electrode material powder to be characterized in detail at the electrical level, rather than relying on the overall average properties, thus providing particle-level differentiated input for subsequent sorting behavior simulation.

[0020] 2. In this invention, by substituting the set of intrinsic electrical properties of particles into the contact charging equation and the electric field force equation, the charge decay curve and electric field migration deflection of each particle in the continuous physical process of triboelectric charging and electrostatic sorting are simulated in the virtual simulation space. This allows the dynamic behavior of particles in actual sorting equipment to be pre-calculated and quantified, establishing a mapping bridge between the intrinsic properties of particles and the sorting behavior response.

[0021] 3. In this invention, by extracting the decay time constant from the charge decay curve and performing feature cross-combination with the work function, and then performing correlation screening with the deflection displacement, a charge separation tendency vector is generated. This enables the key physical factors affecting particle separation effect to be automatically identified and extracted from multi-dimensional features, reducing the blindness of manually selecting features. Attached Figure Description

[0022] The present invention will be described by way of example and with reference to the accompanying drawings, wherein: Figure 1 This is a schematic diagram of the overall process structure of Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the process structure of step 1 in Embodiment 1 of the present invention; Figure 3 This is a schematic diagram of the process structure of step 2 in Embodiment 1 of the present invention; Figure 4 This is a schematic diagram of the process structure of step 3 in Embodiment 1 of the present invention; Figure 5 This is a schematic diagram of the process structure of step 4 in Embodiment 1 of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0024] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0025] It should be noted that, unless otherwise specified, the embodiments and features described in this invention can be combined with each other.

[0026] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0027] In this invention, unless otherwise explicitly specified and limited, "above" or "below" the second feature can include direct contact between the first and second features, or contact between the first and second features through another feature between them. Furthermore, "above," "over," and "on top" of the second feature includes the first feature directly above or diagonally above the second feature, or simply indicates that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature includes the first feature directly below or diagonally below the second feature, or simply indicates that the first feature is at a lower horizontal level than the second feature.

[0028] It should be noted that, unless otherwise specified, the embodiments and features described in this invention can be combined with each other.

[0029] Example 1

[0030] like Figures 1-5 As shown in the figure, an embodiment of the present invention discloses a method for analyzing the physical properties of recycled battery electrode materials, including: Step 1: Obtain the surface and interface spectral data and composition data of individual particles in the electrode material powder obtained from the dismantling of retired batteries, perform spatial registration and standardization correction, and form a multi-source characterization dataset indexed by individual particles.

[0031] Step 1 specifically includes: Step 1.1: Using a microscopic hyperspectral imaging device, acquire a cube of diffuse reflectance hyperspectral images of the electrode material powder under a preset continuous spectral band. Specifically, the electrode material powder obtained after discharging, disassembling, crushing, and preliminary sieving of a retired battery is uniformly spread on a quartz sample stage. The powder spreading thickness is controlled to a single particle layer to avoid severe particle stacking. The magnification of the objective lens of the microscopic hyperspectral imaging device is selected according to the powder particle size range. For example, for electrode material powder with a particle size distribution of 20 micrometers to 200 micrometers, a 10x to 20x objective lens is selected to ensure that a single particle occupies enough pixels in the image to guarantee the signal-to-noise ratio of spectral extraction. The preset continuous spectral band is selected to cover the visible to near-infrared range, for example, 400 nm to 2500 nm, and the spectral sampling interval is set to 5 nm to 10 nm. The hyperspectral image cube contains two-dimensional spatial pixel coordinates and one spectral dimension, with each pixel recording a continuous spectral reflectance curve. During the acquisition process, a dark current image is acquired first, followed by a standard white board reference image for subsequent correction processing.

[0032] Step 1.2: Using a scanning electron microscope coupled with energy-dispersive X-ray spectroscopy (EDXPS), acquire backscattered electron morphology and elemental concentration distribution maps corresponding to the cubic space of the hyperspectral image. After acquiring the hyperspectral image, transfer the same sample stage to the sample chamber of the scanning electron microscope and locate the same field of view as the hyperspectral image using the positioning marks on the sample stage. The accelerating voltage of the scanning electron microscope is set according to the conductivity of the electrode material powder. For mixed powders with high carbon content, an accelerating voltage of 15 kV is selected, and a moderate beam current is chosen to ensure elemental excitation efficiency and spatial resolution. The EDXPS operates in area scanning mode, with the single-point residence time optimized according to the count rate, typically set to 0.5 ms to 2 ms, and the total number of pixels scanned is no less than 512 × 512. After acquisition, the system's built-in quantitative software performs background subtraction, overlapping peak removal, and ZAF correction, outputting the mass percentage of each element at each pixel to form an elemental concentration distribution map. Commonly detected elements include cobalt, nickel, manganese, iron, phosphorus, lithium, aluminum, carbon, and oxygen.

[0033] Step 1.3: Using the backscattered electron topography map as a spatial reference, perform black-and-white correction and multivariate scattering correction pixel-by-pixel on the hyperspectral image cube, and perform background subtraction and feature peak area integration on the elemental concentration distribution map to obtain the corrected spectral reflectance curve and elemental concentration vector. The multimodal data registration process is as follows: First, extract the scale-invariant feature transformation feature points of the backscattered electron topography map and a certain band of the hyperspectral image cube. Calculate the affine transformation matrix between the two images by matching the feature points, and uniformly map all bands of the hyperspectral image cube to the coordinate system of the backscattered electron topography map. The elemental concentration distribution map itself is acquired from the same field of view, and only requires a small translation correction to align with the backscattered electron topography map. After registration, perform black-and-white correction on the hyperspectral image cube, and calculate the reflectance value pixel-by-pixel using the pixel values ​​of the dark current image and the standard white board image. Based on this, multivariate scattering correction is performed on the spectrum of each particle region. The process involves first calculating the average spectrum of all spectra within that region, then performing a univariate linear regression on each spectrum using the average spectrum as a reference, and using the regression coefficients to correct for scattering effects, resulting in the corrected spectral reflectance curve. For the elemental concentration distribution map, a background subtraction method based on iterative polynomial fitting is first used to remove continuous background signals. Then, Gaussian fitting or simplex method peak area integration is performed on the characteristic X-ray peaks of each element to obtain the net peak area count. After conversion using the standard sample calibration curve, the mass percentage of each element is output, forming the elemental concentration vector.

[0034] Step 1.4: Bind the corrected spectral reflectance curve and elemental concentration vector of a single particle according to its contour to form a multi-source representation dataset indexed by a single particle. Particle contour recognition employs adaptive thresholding combined with a watershed algorithm: First, Gaussian smoothing is applied to the backscattered electron topography image for noise reduction. Then, Otsu's method is used to automatically determine the grayscale threshold, binarizing the image to distinguish particle regions from background regions. For adherent particles, a distance-transform-based watershed algorithm is used for segmentation to obtain the contour boundary of each independent particle. Using the contour of each particle as a mask region, the average value of the corrected spectral reflectance curve of all pixels within that region is taken across all bands to obtain the overall spectral reflectance curve representing the particle. Simultaneously, the average value of the elemental concentration vector of all pixels within that region is taken to obtain the average elemental concentration vector of the particle. The particle number, the discrete point sequence of the corrected spectral reflectance curve, and the elemental concentration vector are bound and stored to form a multi-source representation dataset indexed by a single particle.

[0035] Step 2: Perform dielectric response analysis on the spectral data of each particle in the multi-source characterization dataset, and invert the work function and static dielectric constant of the particle by combining the particle composition data, and summarize them into a set of intrinsic electrical properties of the particle.

[0036] Step 2 specifically includes: Step 2.1: Perform a Kramerskrenig transform on the corrected spectral reflectance curve of each particle in the multi-source characterization dataset to extract the real part of the complex permittivity function of the particle in the preset frequency band from the spectral reflectance. and the virtual part ,in Let be the angular frequency of the incident electromagnetic wave, and let the real and imaginary parts constitute the particle dielectric response spectrum. The specific execution process of the Kramerskrenig transform is as follows: first, the corrected spectral reflectance curve is... From wavelength Convert to angular frequency ,in For the speed of light, we obtain Then calculate the reflection amplitude. Reflection phase angle The expression is obtained by the Kramersk-Lenich integral:

[0037] in This represents the Cauchy principal value integral, which is actually solved numerically using the truncated Fourier transform method or the double fast Fourier transform method. After obtaining the phase, the complex reflection coefficient is... Complex permittivity From the relation The inverse kinematics is used to obtain the dielectric function real and imaginary part curves of each particle within a preset frequency band, which is the particle dielectric response spectrum.

[0038] Step 2.2: Locate the imaginary part from the dielectric response spectrum of the particles. Characteristic peak frequency position ,in This is the angular frequency corresponding to the maximum value of the imaginary part. The predicted work function of the particle is output by weighting the molar ratio of transition metals to alkali metals in the elemental concentration vector of the corresponding particle, inputting a pre-calibrated work function regression equation, and outputting the predicted work function of the particle. The process of establishing the pre-calibrated work function regression equation is as follows: Several standard electrode material samples with known work functions are selected, including lithium cobalt oxide, lithium nickel cobalt manganese oxide, lithium iron phosphate, natural graphite, and artificial graphite with different lithium contents. The actual work function values ​​of each standard sample are measured using Kelvin probe force microscopy or ultraviolet photoelectron spectroscopy as reference true values. Simultaneously, the characteristic peak frequency positions of each standard sample are obtained according to the method described above in this step. The molar ratio of elements. Using these data as training samples, a multiple linear regression equation is established:

[0039] in This represents the molar ratio of transition metals to alkali metals. The coefficient represents the mass percentage of carbon. to The work function is determined by least squares fitting. For example, for lithium cobalt oxide particles, where the transition metal element is cobalt, if the cobalt-lithium molar ratio is 1.05 and the characteristic peak frequency is at a specific value, the regression equation outputs a predicted work function of approximately 4.7 eV; for graphite particles, where the carbon content is higher than 90%, the predicted work function is approximately 4.3 eV.

[0040] Step 2.3: Extract the static dielectric constant from the low-frequency asymptotic region of the particle dielectric response spectrum. ,in This represents the limiting value of the real part of the dielectric function as the angular frequency approaches zero. In practice, several data points on the real part curve of the dielectric function are selected at the low-frequency end, typically within the first 10% of the interval corresponding to the lowest frequency in the measurement band. These data points are then linearly extrapolated to zero frequency to obtain the static dielectric constant. The predicted work function and static dielectric constant These collectively constitute the intrinsic electrical properties of the particle. A set of intrinsic electrical properties for all particles is compiled, with each particle in the dataset corresponding to a set of these properties. Numerical value.

[0041] Step 3: Simulate the continuous physical process of industrial triboelectric charging and electrostatic sorting. Substitute the set of intrinsic electrical properties of the particles into the contact charging equation and the electric field force equation to calculate the charge decay curve and electric field migration deflection of each particle.

[0042] Step 3 specifically includes: Step 3.1: In the coupled simulation space of Discrete Element Method (DEM) and Computational Fluid Dynamics (CFD), a three-dimensional virtual sorting flow field is set up, including a triboelectric pipe section and a high-voltage electrostatic sorting section. The inner wall material of the triboelectric pipe section is set to a preset metal, and the high-voltage electrostatic sorting section is equipped with a corona electrode array and a static electrode plate. In specific implementation, the coupled simulation platform of DEM and CFDEM uses the EDEM and Fluent coupling framework or the open-source CFDEM coupling framework. The geometric model of the triboelectric pipe section is established according to the actual size of the industrial triboelectric electrostatic sorting machine. The pipe cross-section is rectangular, with a width of 0.15 meters, a height of 0.08 meters, and a length of 1.5 meters. The inner wall material of the pipe is set to copper, and the work function is set to 4.65 electron volts. The high-voltage electrostatic sorting section is located downstream of the pipe outlet and includes a row of corona electrode wires and a static electrode plate. The corona electrode wires have a diameter of 0.3 mm and are connected to a negative high-voltage power supply. The static electrode plate is a flat plate electrode, grounded, and the distance between the corona electrode and the static electrode plate is set to 0.04 meters. The airflow velocity inside the pipeline is set to 5 to 15 meters per second, depending on the actual working conditions.

[0043] Step 3.2: Instantiate each particle in the set of intrinsic electrical properties of the particles as a discrete element particle, based on the predicted work function of the particle. The work function of the particle surface is set based on the static dielectric constant of the particle. The dielectric properties of the particles are set, and the mass density is calculated and assigned to the particles based on the weighted average atomic weights of each element in the particle's elemental concentration vector. The particle size is set according to the equivalent circle diameter obtained from the scanning electron microscope image analysis in step 1, allowing a certain distribution of particle size within the range of 10 micrometers to 200 micrometers. Taking lithium cobalt oxide particles as an example, the theoretical density of lithium cobalt oxide crystals is calculated to be approximately 5.1 g / cm³ based on the mass percentages of lithium, cobalt, and oxygen in the elemental concentration vector and their respective atomic weights. This is then corrected by multiplying by the packing factor of 0.8, resulting in an apparent density of 4.1 g / cm³. The particle's Young's modulus, Poisson's ratio, and coefficient of restitution are obtained from material handbooks and literature and used for subsequent discrete element contact force calculations. After completing the property settings, the particles are released at the inlet of the triboelectric conduit with random positions and initial velocities.

[0044] Step 3.3: In the triboelectrically charged pipe section, simulate the multiple contact and collision processes between the particles and the metal wall. Calculate the charge transfer amount of each collision based on the difference between the work function of the particle surface and the work function of the wall. Accumulate these values ​​to obtain the initial cumulative charge of the particles before they enter the high-voltage electrostatic separation section.

[0045] Step 3.3 specifically includes: Step 3.31: Within the triboelectrically charged pipe section, track the trajectory of each particle. When the distance between a particle and the metal wall is less than the preset contact threshold, a contact collision event is determined. The preset contact threshold is one-thousandth of the particle radius to ensure that the discrete element contact model is activated when the particle is about to contact the wall. During the simulation, at each computation time step, all particles are checked for their distance from the pipe wall. Once the distance is less than the threshold, the particle's number, position, velocity, and collision time are recorded.

[0046] Step 3.32: For each contact collision event, calculate the normal contact force based on the particle incident angle and velocity, and determine the contact area using Hertzian contact theory. The formula for calculating the Hertzian contact area is:

[0047] in The normal contact force is calculated from the normal overlap and contact stiffness between the particle and the wall in the discrete element model. The equivalent contact radius is equal to the particle radius for the contact between a particle and a planar wall. The equivalent elastic modulus is derived from the elastic modulus of granular materials. Poisson's ratio and the elastic modulus of wall material Poisson's ratio Calculate using the formula:

[0048] Normal contact force The relationship between the amount of overlap follows the Hertzian contact model and is automatically calculated by the discrete element solver.

[0049] Step 3.33: Based on the contact work function difference Calculate the charge transfer amount in a single contact ,in Predicted work function for particles With wall work function The difference, The amount of charge transferred in a single contact. The preset contact charge coefficient, The contact area is determined in step 3.32. Contact charge coefficient. Through calibration experiments, it was determined that its physical meaning is the amount of charge transferred per unit area and per unit work function difference, with a typical value of approximately [value missing]. Coulombs per square meter per electron volt. According to The sign of the charge determines the direction of charge transfer. When electrons transfer from the wall to the particle surface, the particle gains electrons and becomes negatively charged; when When electrons transfer from the particle surface to the wall, the particle loses electrons and becomes positively charged. For example, a lithium cobalt oxide particle with a predicted work function of 4.7 eV comes into contact with a copper wall with a work function of 4.65 eV. When the value is positive, electrons flow from the copper wall to the particle, and the particle becomes negatively charged.

[0050] Step 3.34: Sum the charge transfer amounts of all contact and collision events experienced by the particles within the triboelectrically charged pipe section to obtain the initial accumulated charge of the particles before entering the high-voltage electrostatic separation section. During the simulation, each particle maintains an accumulated charge variable, initially set to zero. Each time a contact collision occurs, the charge calculated in step 3.33 is... The number of collisions, accumulated according to their positive or negative signs, varies from a few to dozens, depending on the pipe length, airflow velocity, and particle size. The accumulated charge of the particles when they reach the pipe outlet with the airflow is the initial accumulated charge. .

[0051] Step 3.4: After the particles enter the high-voltage electrostatic separation section, the charge decay process of the particles in the corona field and the effect of the electric field force are simulated. The motion trajectory of the particles under the combined action of electric field force, gravity and fluid drag force is calculated, and the deflection displacement and real-time charge sequence of the particles at the exit of the high-voltage electrostatic separation section are recorded.

[0052] Step 3.4 specifically includes: Step 3.41: Solve the Laplace equation between the corona electrode array and the static electrode plate in the high-voltage electrostatic separation section using the finite element method to obtain the electric field distribution between the electrodes. ,in and These represent the spatial coordinates parallel and perpendicular to the direction of the stationary electrode plate, respectively. Considering the distortion effect of space charge generated by corona discharge on the electric field, the current continuity equation and Poisson equation need to be coupled and iterated during the actual solution process. The boundary conditions are set as follows: the surface potential of the corona electrode is a negative high voltage value, such as -20 kV; the surface potential of the stationary electrode plate is zero; and the remaining boundaries are set as zero normal electric field or far-field boundaries. The solution is obtained using commercial finite element software or a self-written program. The mesh is refined near the electrodes to ensure calculation accuracy, and finally, the electric field vectors at each discrete node in space are obtained.

[0053] Step 3.42: Simulate the real-time charge of particles after they enter the high-voltage electrostatic separation section. The exponential decay process, according to Calculate the charge of the particle at each time step, where For particles at time Real-time charge level, The initial accumulated charge obtained in step 3.34 The initial decay time constant, For the natural constant An exponential function with base 0.05. Initial decay time constant. The value of is related to the particle conductivity, the content of adsorbed water molecules on the particle surface, and the relative humidity of the environment. It can be calibrated before simulation by measuring the discharge curves of different material particles in an electric field. For example, for graphite particles in a dry environment, Approximately 0.1 to 0.5 seconds; for metal oxide cathode particles, It could be between 0.05 seconds and 0.2 seconds. At each calculation time step, based on the current time... Update the real-time charge of the particles .

[0054] Step 3.43: Calculate the dielectric force generated by the polarization effect of particles in a non-uniform electric field:

[0055] in The dielectric force acting on the particles, Let be the vacuum permittivity, and take the value of . Farad per meter, The dielectric constant of air is approximately 1.0. The equivalent radius of the particle. The gradient is the square of the electric field strength. The Clausius-Mosotti factor has the following values:

[0056] This factor reflects the effect of the dielectric difference between the static dielectric constant of the particles and the dielectric constant of air on the polarization effect. The value of this factor is usually between -0.5 and 1.0: when the dielectric constant of the particles is greater than that of air, the factor is positive, and the dielectric force points in the direction of increasing electric field; conversely, it points in the direction of decreasing electric field. The electric field distribution obtained in step 3.41 is obtained by numerical differentiation calculation, and the Green's Gaussian gradient calculation method is used on the finite element mesh.

[0057] Step 3.44: Calculate the Coulomb force on the particle surface charge in the electric field. ,in The Coulomb force is the force exerted on the particle. The dielectric force... With Coulomb force The vector sum of the electric force on the particle is the sum of the vectors. In addition, the particles are also subject to gravity. ,in For particle mass, Let gravitational acceleration be 9.81 m / s²; and fluid drag force be calculated using Stokes' law or a more general drag force model. The vector sum of these forces is then taken as the net force acting on the particle, and applied using Newton's equations of motion. The trajectory of the particles within the high-voltage electrostatic separation section is determined. Explicit Euler or Runge-Kutta methods are used for time integration, and the particle's position coordinates and velocity are recorded at each time step. When the particle reaches the exit section of the separation section, its deflection displacement in the electric field direction is recorded. and the real-time charge throughout the process Sequence. For example, negatively charged lithium cobalt oxide particles are subjected to Coulomb repulsion away from the corona electrode in a negative corona electric field, and with the influence of dielectric force, their deflection displacement can reach several centimeters; graphite particles, due to their different work function and dielectric constant, have a significantly different deflection displacement from lithium cobalt oxide particles.

[0058] Step 4: Extract the decay time constant from the charge decay curve, perform feature cross-combination of the decay time constant with the work function in the set of intrinsic electrical properties of the particles, and perform correlation screening with the electric field migration deflection to generate a charge separation tendency vector.

[0059] Step 4 specifically includes: Step 4.1: Perform exponential decay fitting on the real-time charge sequence of each particle, and extract the time constant corresponding to the time when the charge decays to a preset proportion of the initial accumulated charge. As a characteristic of charge decay, the fitting method employs nonlinear least squares, and the fitting function has the following form: The fitting parameters are The preset ratio is acceptable. That is, approximately 36.8%, at which point the corresponding time constant is the characteristic time constant of exponential decay. In the actual fitting process, the real-time charge sequence recorded in step 3.42 was used as the observation data, and the initial guess value was taken as... The optimal fitting parameters are obtained by iterative optimization until convergence. .

[0060] Step 4.2: Set the time constant... Predicted work function of the corresponding particle Nonlinear cross-combination is performed to form a combined feature set, which includes and Various combinations of mathematical transformations. Specific combinations of mathematical transformations include: original values... and The product of the two The ratio of the two The logarithmic product of the two The sum of the squares of the two These combined features can capture the nonlinear coupling relationship that may exist between the work function and the decay time constant, providing rich candidate dimensions for subsequent correlation analysis and feature selection.

[0061] Step 4.3: Calculate the relationship between each feature dimension in the combined feature set and the deflection displacement. Pearson correlation coefficient between ,in The statistic used to measure the degree of linear correlation between two variables is calculated using the following formula: in For the first The value of each particle in this feature dimension For the first The deflection displacement of each particle and For the corresponding mean, This represents the total number of particles. (Retain) Feature dimensions exceeding a preset correlation threshold, typically between 0.7 and 0.9, indicate a strong linear correlation between the feature dimension and the deflection displacement. The retained feature dimensions are then compared with the static dielectric constant in the set of intrinsic electrical properties of the particles. The data is then spliced ​​together to form a subset of sensitive features.

[0062] Step 4.4: Use the subset of sensitive features as the charge separation tendency vector of the particle. This vector centrally reflects the intrinsic electrical properties and their combination features that are closely related to the deflection behavior of the particle during electrostatic separation. It realizes the automatic selection of the feature dimensions that have the greatest influence on the separation effect from the original multidimensional data, avoiding the subjectivity of manual feature selection.

[0063] Step 5 specifically includes: Step 5.1: Perform standard score standardization on the charge separation tendency vectors of all particles according to their dimensions. Project the standardized vectors onto a low-dimensional decision plane composed of the first and second principal components using principal component analysis. The standard score standardization process is as follows: For each feature dimension, calculate the mean of the values ​​taken by all particles in that dimension. and standard deviation Then, the original value of each particle in that dimension. Transform into This process ensures that the mean of each dimension after transformation is 0 and the standard deviation is 1, eliminating the influence of differences in dimensions and numerical ranges on subsequent projection. After standardization, the covariance matrix is ​​calculated for the standardized vectors of all particles, and the eigenvalues ​​and eigenvectors are solved. The eigenvectors corresponding to the two largest eigenvalues ​​are taken as the first principal component direction and the second principal component direction. The standardized vector of each particle is projected onto these two directions to obtain the coordinates in the low-dimensional decision plane.

[0064] Step 5.2: In the low-dimensional decision plane, label each scatter point according to the actual electrode material type of the particles. The actual electrode material type can be predetermined by combining the elemental concentration vector from Step 1 with chemometric methods or X-ray diffraction phase identification results. For example, lithium cobalt oxide particles are labeled red, graphite particles are labeled black, and lithium nickel cobalt manganese oxide particles are labeled green. Calculate the mean vector of the scatter point distribution for each type of material. Covariance Matrix subscript Indicates the first similar materials, For the first The vector formed by the mean values ​​of the coordinates of all particles in the low-dimensional decision plane of the material-like material. For the first Calculate the covariance matrix of all particle coordinates for all material classes. Also, calculate the covariance matrix for any two material classes. and Mahalanobis distance between them:

[0065] in To take into account the inter-class distance metric based on the intra-class distribution covariance structure, The pooled covariance matrix is ​​the result of merging the two classes, with superscript... Indicates matrix transpose, superscript This represents the inversion of a matrix. The advantage of Mahalanobis distance over Euclidean distance is that it considers the dispersion and correlation within each class. When two classes of particles are widely distributed in a certain direction, the difference weight in that direction will decrease accordingly, making the distance metric more consistent with the actual difficulty of separation.

[0066] Step 5.3: For Mahalanobis distance For adjacent material categories with values ​​less than a preset threshold, calculate the inter-class divergence under the Fisher linear discrimination criterion. With intra-class divergence The projection direction that maximizes the ratio:

[0067] in The inter-class scatter matrix is ​​a measure of the degree of separation between categories. The intra-class scatter matrix is ​​used to measure the degree of dispersion within a category. The optimal projection direction is determined by maximizing the ratio of inter-class divergence to intra-class divergence. A preset threshold for the Mahalanobis distance is set based on the actual separation accuracy requirements. When the Mahalanobis distance between two types of material particles is less than this threshold, their distribution regions in the low-dimensional decision plane are considered to overlap significantly, making efficient separation difficult through simple linear boundaries. Therefore, it is necessary to solve for the optimal separation direction. Fisher linear discriminant analysis ensures optimal separation of the two types of particles along this projection direction by maximizing the ratio of inter-class distance to intra-class distance after projection. The corresponding optimal separating line in the low-dimensional decision plane, the direction of this line is... Vertical alignment allows for maximum separation of the two types of particles in a projective sense.

[0068] Step 5.4: Extract the projection direction The components of the particle are sorted by their absolute values, and the dimensions with the largest absolute values ​​are identified as the key feature dimensions that determine the separation effect. The preset number is typically 3 to 5, and can be adjusted according to the required precision and computational complexity. The larger the absolute value of a component, the greater its contribution to the optimal separation direction; that is, the feature plays a dominant role in distinguishing the separation behavior of two types of material particles. For example, if the component analysis results show that the dimension corresponding to the product of the work function and the decay time constant has the largest absolute value component, it indicates that the coupling effect between the work function and charge decay characteristics is the key physical mechanism determining whether the two types of particles can be effectively separated.

[0069] Step 6: Map the normal vector of the optimal separation line back to the operating parameter space of the industrial sorting machine, determine the contribution weight of each parameter in the operating parameter space to the separation effect, and generate a parameter configuration sequence to guide physical sorting.

[0070] Step 6 specifically includes: Step 6.1: Based on the original physical properties corresponding to the key feature dimensions, determine the adjustable operating parameters affected by the key feature dimensions in the operating parameter space of the industrial triboelectric and electrostatic separator. These adjustable operating parameters include at least the geometric dimensions and aerodynamic parameters of the triboelectric section, and the electric field strength parameters of the high-voltage electrostatic separation section. Specifically, the mapping relationship is as follows: if the key feature dimension mainly involves the work function, the corresponding adjustable operating parameters include the inner wall material type and pipe length of the triboelectric pipe section, because different materials have different wall work functions, and the pipe length determines the number of collisions between the particles and the wall. If the key feature dimension mainly involves the static dielectric constant or charge decay time constant, the corresponding adjustable operating parameters include the corona electrode voltage and static electrode voltage of the high-voltage electrostatic separation section, because these parameters determine the magnitude and gradient distribution of the electric field strength, directly affecting the relative magnitudes of the dielectric force and Coulomb force. Furthermore, the airflow velocity simultaneously affects the particle transport time of the triboelectric section and the residence time of the electrostatic separation section, and is also a key adjustable parameter.

[0071] Step 6.2: Using the adjustable operating parameters as independent variables, and the Mahalanobis distance between adjacent material categories in the low-dimensional decision plane... As the response variable, a parameter sampling scheme is generated using experimental design methods. Simulations are performed at each sampling point in the three-dimensional virtual sorting flow field to obtain the Mahalanobis distance response values ​​under each parameter combination. Central composite design or Box-Behnken design can be used as the experimental design method; these two methods can effectively capture the nonlinear relationship between parameters and response with fewer simulations. Each sampling point corresponds to a specific set of operating parameter values. These parameter values ​​are substituted into the simulation model established in step 3, and the complete simulation process of triboelectric charging and electrostatic sorting is re-executed. The particle characteristics and Mahalanobis distance in the low-dimensional decision plane are recalculated according to steps 4 and 5. Simulations are performed one by one for all sampling points in the design matrix, and the Mahalanobis distance values ​​corresponding to each parameter combination are recorded to form an experimental data table.

[0072] Step 6.3: Perform response surface regression modeling on the sampling results to obtain the Mahalanobis distance. Regarding the regression mapping relationship of the adjustable operating parameters, the solution is to find the parameter combination that maximizes the regression mapping relationship. The response surface model adopts a second-order polynomial form, namely:

[0073] in For the first One adjustable operating parameter For the total number of parameters, For the intercept term, The linear effect coefficient, This is the secondary effect coefficient. This is the interaction effect coefficient. This represents the random error term. The regression coefficients are estimated using the least squares method. After obtaining the regression equation, the Lagrange multiplier method or the direct search method is used to solve for the parameter values ​​within the range of possible values. The optimal parameter combination that achieves the maximum value. If multiple local extrema exist, a genetic algorithm or particle swarm optimization algorithm can be used for global optimization.

[0074] Step 6.4: Arrange the optimal parameter combination obtained from the solution into a parameter configuration sequence according to the actual sorting process sequence. This parameter configuration sequence includes the operating parameter settings for the triboelectric section and the high-voltage electrostatic separation section, and is output to the sorting machine control system. Specifically, according to the material flow direction, first set parameters such as the airflow velocity and pipe length in the triboelectric section pipe, then sequentially set parameters such as the corona electrode voltage and electrostatic electrode voltage in the high-voltage electrostatic separation section, forming a list of settings arranged in time or station order. Transmit this parameter configuration sequence to the sorting machine's programmable logic controller (PLC) via a communication interface to guide the sorting machine to operate according to the optimal parameters, achieving efficient physical sorting of retired battery electrode material powder.

[0075] Example 2

[0076] This embodiment proposes a physical property analysis system for recycling battery electrode materials, the system comprising: The multi-source data acquisition and registration module acquires the surface and interface spectral data and composition data of individual particles in the electrode material powder obtained from the dismantling of retired batteries, performs spatial registration and standardization correction, and forms a multi-source characterization dataset indexed by individual particles. The electrical intrinsic property inversion module performs dielectric response analysis on the spectral data of each particle in the multi-source characterization dataset, and inverts the work function and static dielectric constant of the particle by combining the particle composition data, and summarizes them into a set of particle electrical intrinsic properties. The charging behavior simulation module simulates the continuous physical process of industrial triboelectric charging and electrostatic sorting. It substitutes the set of intrinsic electrical properties of the particles into the contact charging equation and the electric field force equation to calculate the charge decay curve and electric field migration deflection of each particle. The separation tendency characterization generation module extracts the decay time constant from the charge decay curve, performs feature cross-combination of the decay time constant with the work function in the set of intrinsic electrical properties of the particles, and performs correlation screening with the electric field migration deflection to generate a charge separation tendency vector. The separation boundary decision module projects the charge separation tendency vector of all particles onto a low-dimensional decision plane, calculates the separation degree between the distribution areas of particles of different material components, solves the optimal separation line that maximizes the separation degree between adjacent material components, and extracts the normal vector of the optimal separation line. The sorting parameter generation and output module reverse maps the normal vector of the optimal separation line back to the operating parameter space of the industrial sorting machine, determines the contribution weight of each parameter in the operating parameter space to the separation effect, and generates a parameter configuration sequence to guide physical sorting.

[0077] The circuits, electronic components, and modules involved are all existing technologies, which can be fully implemented by those skilled in the art, and need not be elaborated upon. The scope of protection of this invention does not involve any improvement to the software and methods.

[0078] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0079] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for analyzing the physical properties of recycled battery electrode materials, characterized in that, include: Step 1: Obtain the surface and interface spectral data and composition data of individual particles in the electrode material powder obtained from the dismantling of retired batteries, perform spatial registration and standardization correction, and form a multi-source characterization dataset indexed by individual particles; Step 2: Perform dielectric response analysis on the spectral data of each particle in the multi-source characterization dataset, and invert the work function and static dielectric constant of the particle by combining the particle composition data, and summarize them into a set of intrinsic electrical properties of the particle. Step 3: Simulate the continuous physical process of industrial triboelectric charging and electrostatic sorting, substitute the set of intrinsic electrical properties of the particles into the contact charging equation and the electric field force equation, and calculate the charge decay curve and electric field migration deflection of each particle. Step 4: Extract the decay time constant from the charge decay curve, perform feature cross-combination of the decay time constant and the work function in the set of intrinsic electrical properties of the particles, and perform correlation screening with the electric field migration deflection to generate a charge separation tendency vector. Step 5: Project the charge separation tendency vector of all particles onto a low-dimensional decision plane, calculate the separation degree between the particle distribution areas of different material components, solve for the optimal separation line that maximizes the separation degree between adjacent material components, and extract the normal vector of the optimal separation line. Step 6: Map the normal vector of the optimal separation line back to the operating parameter space of the industrial sorting machine, determine the contribution weight of each parameter in the operating parameter space to the separation effect, and generate a parameter configuration sequence to guide physical sorting.

2. The method for analyzing the physical properties of recycled battery electrode materials according to claim 1, characterized in that, Step 1 specifically includes: Using a microscopic hyperspectral imaging device, a cube of diffuse reflectance hyperspectral image of the electrode material powder in a preset continuous spectral band is acquired; Using a scanning electron microscope coupled with energy-dispersive X-ray spectroscopy, backscattered electron morphology and elemental concentration distribution maps corresponding to the cubic space of the hyperspectral image were acquired. Using the backscattered electron topography as a spatial reference, black-and-white correction and multivariate scattering correction are performed pixel by pixel on the hyperspectral image cube, and background subtraction and characteristic peak area integration are performed on the element concentration distribution map to obtain the corrected spectral reflectance curve and element concentration vector. The corrected spectral reflectance curves and elemental concentration vectors of individual particles are bound together according to their particle profiles to form a multi-source characterization dataset indexed by individual particles.

3. The method for analyzing the physical properties of recycled battery electrode materials according to claim 1, characterized in that, Step 2 specifically includes: Perform a Kramerskrenig transform on the corrected spectral reflectance curve of each particle in the multi-source characterization dataset to extract the real part of the complex permittivity of the particle in a preset frequency band from the spectral reflectance. and the virtual part ,in The angular frequency of the incident electromagnetic wave is given, and the real and imaginary parts constitute the particle dielectric response spectrum. Locate the characteristic peak frequency position of the imaginary part from the dielectric response spectrum of the particles. ,Will The predicted work function of the particle is output by weighting the molar ratio of transition metals to alkali metals in the elemental concentration vector of the corresponding particle, inputting a pre-calibrated work function regression equation, and outputting the predicted work function of the particle. ; The static dielectric constant is extracted from the low-frequency asymptotic region of the dielectric response spectrum of the particles. The predicted work function and the static dielectric constant are used together as the electrical intrinsic properties of the particle, and the set of particle electrical intrinsic properties of all particles is summarized.

4. The method for analyzing the physical properties of recycled battery electrode materials according to claim 3, characterized in that, Step 3 specifically includes: Step 3.1: In the coupled simulation space of discrete element method and computational fluid dynamics, a three-dimensional virtual sorting flow field is set up, which includes a triboelectric pipe section and a high-voltage electrostatic sorting section. The inner wall material of the triboelectric pipe section is set to a preset metal, and the high-voltage electrostatic sorting section is equipped with a corona electrode array and a static electrode plate. Step 3.2: Instantiate each particle in the set of intrinsic electrical properties of the particles as a discrete element particle, set the surface work function of the particle according to the predicted work function of the particle, set the dielectric property of the particle according to the static dielectric constant of the particle, and calculate the mass density according to the weighted atomic weight of each element in the element concentration vector of the particle and assign it to the particle. Step 3.3: In the triboelectrically charged pipe section, simulate the multiple contact and collision processes between the particles and the metal wall. Calculate the charge transfer amount of each collision based on the difference between the work function of the particle surface and the work function of the wall. Accumulate the initial cumulative charge of the particles before entering the high-voltage electrostatic separation section. Step 3.4: After the particles enter the high-voltage electrostatic separation section, the charge decay process of the particles in the corona field and the effect of the electric field force are simulated. The motion trajectory of the particles under the combined action of electric field force, gravity and fluid drag force is calculated, and the deflection displacement and real-time charge sequence of the particles at the exit of the high-voltage electrostatic separation section are recorded.

5. The method for analyzing the physical properties of recycled battery electrode materials according to claim 4, characterized in that, Step 3.3 specifically includes: Step 3.31: Within the triboelectric pipe section, track the trajectory of each particle. When the distance between the particle and the metal wall is less than the preset contact detection distance, determine that a contact collision event has occurred. Step 3.32: For each contact collision event, calculate the normal contact force based on the particle incident angle and velocity, and determine the contact area using Hertzian contact theory. ; Step 3.33: Based on the contact work function difference Calculate the charge transfer amount in a single contact: ， in Predicted work function for particles With wall work function The difference, The amount of charge transferred in a single contact. The preset contact charge coefficient; And according to The sign of the charge determines the direction of charge transfer. Electrons transfer from the wall to the particle surface, causing the particle to become negatively charged. Electrons transfer from the particle surface to the wall, causing the particle to become positively charged; Step 3.34: Sum the charge transfer amounts of all contact and collision events experienced by the particles within the triboelectrically charged pipe section to obtain the initial accumulated charge of the particles before entering the high-voltage electrostatic separation section. .

6. The method for analyzing the physical properties of recycled battery electrode materials according to claim 4, characterized in that, Step 3.4 specifically includes: Step 3.41: Solve the Laplace equation between the corona electrode array and the static electrode plate in the high-voltage electrostatic separation section using the finite element method to obtain the electric field distribution between the electrodes. ,in and These are the spatial coordinates parallel to the static electrode plate and perpendicular to the static electrode plate, respectively. Step 3.42: Simulate the real-time charge of particles after they enter the high-voltage electrostatic separation section. The exponential decay process, according to Calculate the charge of the particle at each time step, where For particles at time Real-time charge level, The initial decay time constant, For the natural constant An exponential function with base 0; Step 3.43: Calculate the dielectric force generated by the polarization effect of particles in a non-uniform electric field: in The dielectric force acting on the particles, The vacuum permittivity, The dielectric constant of air is The equivalent radius of the particle. The gradient is the square of the electric field strength. The Clausius Mossotti factor has a value of This reflects the effect of the dielectric difference between the static dielectric constant of particles and the dielectric constant of air on the polarization effect; Step 3.44: Calculate the Coulomb force on the particle surface charge in the electric field. Dielectrophoresis force The vector sum of the Coulomb forces is the electric force on the particle. By combining the gravitational force and fluid drag acting on the particles, the trajectory of the particles within the high-voltage electrostatic separation section is solved using Newton's equations of motion, and the deflection displacement of the particles along the electric field direction at the exit of the high-voltage electrostatic separation section is recorded. and real-time charge sequence.

7. The method for analyzing the physical properties of recycled battery electrode materials according to claim 4, characterized in that, Step 4 specifically includes: Step 4.1: Perform exponential decay fitting on the real-time charge sequence of each particle, and extract the time constant corresponding to the time when the charge decays to a preset proportion of the initial accumulated charge. As a characteristic of charge decay; Step 4.2: Nonlinearly combine the time constant with the predicted work function of the corresponding particle to form a combined feature set, which includes various mathematical transformation combinations of the predicted work function and the time constant; Step 4.3: Calculate the Pearson correlation coefficient between each feature dimension in the combined feature set and the deflection displacement. ,reserve Feature dimensions exceeding a preset correlation threshold are concatenated with the static dielectric constant in the set of intrinsic electrical properties of the particles to form a sensitive feature subset; Step 4.4: Use the subset of sensitive features as the charge separation tendency vector of the particle.

8. The method for analyzing the physical properties of recycled battery electrode materials according to claim 7, characterized in that, Step 5 specifically includes: The charge separation tendency vectors of all particles are standardized by standard fractions according to their dimensions. The standardized vectors are then projected onto a low-dimensional decision plane composed of the first and second principal components through principal component analysis. In the low-dimensional decision plane, each scatter point is labeled according to the actual electrode material type of the particle, and the mean vector of the scatter point distribution for each material type is calculated. Covariance Matrix subscript Indicates the first Consider two classes of materials and calculate the values ​​of any two classes of materials. and Mahalanobis distance between them: in The pooled covariance matrix is ​​the result of merging the two classes, with superscript... Indicates matrix transpose, superscript This represents finding the inverse of a matrix. Mahalanobis distance For adjacent material categories with values ​​less than a preset threshold, calculate the inter-class divergence under the Fisher linear discrimination criterion. With intra-class divergence The projection direction that maximizes the ratio ; Extract the components of each dimension of the projection direction, sort them according to the absolute value of the components, and identify the first preset number of dimensions with the largest absolute value of the components as the key feature dimensions that determine the separation effect.

9. The method for analyzing the physical properties of recycled battery electrode materials according to claim 8, characterized in that, Step 6 specifically includes: Based on the original physical properties corresponding to the key feature dimensions, adjustable operating parameters affected by the key feature dimensions are determined in the operating parameter space of the industrial triboelectric and electrostatic separator. The adjustable operating parameters include at least the geometric dimension parameters and airflow dynamic parameters of the triboelectric section, and the electric field strength parameters of the high-voltage electrostatic separation section. Using the adjustable operating parameters as independent variables and the Mahalanobis distance between adjacent material categories in the low-dimensional decision plane as response variables, a parameter sampling scheme is generated through experimental design methods. Simulations are performed at each sampling point in the three-dimensional virtual sorting flow field to obtain the Mahalanobis distance response values ​​under each parameter combination. Response surface regression modeling is performed on the sampling results to obtain the regression mapping relationship of Mahalanobis distance with respect to the adjustable operating parameters, and the parameter combination that maximizes the regression mapping relationship is solved. The optimal parameter combination obtained from the solution is arranged into a parameter configuration sequence according to the actual sorting process sequence. The parameter configuration sequence includes the operating parameter setting values ​​of the triboelectric section and the operating parameter setting values ​​of the high-voltage electrostatic separation section, and is output to the sorting machine control system.

10. A physical property analysis system for recycling battery electrode materials, characterized in that, The system for performing the method according to any one of claims 1 to 9 comprises: The multi-source data acquisition and registration module acquires the surface and interface spectral data and composition data of individual particles in the electrode material powder obtained from the dismantling of retired batteries, performs spatial registration and standardization correction, and forms a multi-source characterization dataset indexed by individual particles. The electrical intrinsic property inversion module performs dielectric response analysis on the spectral data of each particle in the multi-source characterization dataset, and inverts the work function and static dielectric constant of the particle by combining the particle composition data, and summarizes them into a set of particle electrical intrinsic properties. The charging behavior simulation module simulates the continuous physical process of industrial triboelectric charging and electrostatic sorting. It substitutes the set of intrinsic electrical properties of the particles into the contact charging equation and the electric field force equation to calculate the charge decay curve and electric field migration deflection of each particle. The separation tendency characterization generation module extracts the decay time constant from the charge decay curve, performs feature cross-combination of the decay time constant with the work function in the set of intrinsic electrical properties of the particles, and performs correlation screening with the electric field migration deflection to generate a charge separation tendency vector. The separation boundary decision module projects the charge separation tendency vector of all particles onto a low-dimensional decision plane, calculates the separation degree between the distribution areas of particles of different material components, solves the optimal separation line that maximizes the separation degree between adjacent material components, and extracts the normal vector of the optimal separation line. The sorting parameter generation and output module reverse maps the normal vector of the optimal separation line back to the operating parameter space of the industrial sorting machine, determines the contribution weight of each parameter in the operating parameter space to the separation effect, and generates a parameter configuration sequence to guide physical sorting.

Images