Low-expansion and high-compaction-density artificial graphite material and preparation method thereof

Abstract

The invention provides a low-expansion and high-compaction-density artificial graphite material and a preparation method thereof. Compared with the prior art, the preparation method comprises the following steps: selecting a carbon-containing raw material, calcining, crushing, grinding, screening and grading to obtain carbonaceous precursor powder; mixing the precursor powder with a binder, sequentially performing granulation, spheroidization shaping and drying to obtain precursor particles, and acquiring particle characteristic parameters such as D10, D50, D90, sphericity, roundness, length-diameter ratio and surface adhesion fine powder proportion through an image acquisition and graphic processing unit; when the particle morphology and the particle size distribution deviate from a preset range, the correction unit adjusts granulation, shaping and / or drying parameters to realize closed-loop control; placing the precursor particles in a carbonization furnace, and performing heating and heat preservation in an inert atmosphere according to a matching temperature curve output by the temperature matching model to form a carbonized precursor; and graphitizing at 2500-3000 DEG C in an inert atmosphere, cooling, crushing, screening, and carrying out acid pickling and surface coating modification to obtain the artificial graphite material.

Description

Technical Field

[0001] This invention relates to the field of artificial graphite preparation technology, and in particular to an artificial graphite material with low expansion and high compaction density and its preparation method. Background Technology

[0002] In the field of lithium-ion battery anode materials, artificial graphite is widely used due to its scalable preparation, adjustable performance consistency, and controllable cost. With the increasing demands for energy density and fast-charging performance in power and storage batteries, anode materials are generally developing towards higher compaction density, lower expansion, and higher rate capability. Among these, process routes that improve packing density while maintaining electrochemical performance through spheroidization and secondary granulation are common. However, the microstructure evolution and macroscopic particle morphology / size distribution of artificial graphite have a significant impact on compaction density, expansion rate, and cycle stability. This is especially true when raw material fluctuations, equipment conditions change, and carbonization / graphitization heat treatment conditions are unstable, which can easily cause batch-to-batch differences, affecting the consistency and yield of the final product.

[0003] Our research team has long been reviewing and studying a large amount of relevant data on artificial graphite materials. Relying on relevant resources and conducting numerous experiments, we discovered existing technologies such as CN115415003B, CN115893401B, and CN117566735B. One such technology discloses a secondary particle artificial graphite anode material and its preparation method. This preparation method includes the following steps: mixing the anode raw material and a binder uniformly, then adding water and stirring to obtain a wet material; pelletizing the wet material to obtain a spherical mixture; wherein the binder is a composite of starch-based binder and asphalt-based binder or resin-based binder; heat-treating the spherical mixture to obtain dried and hardened pellets, which are then graphitized; and finally, pulverizing and sieving the graphitized pellets to obtain the secondary particle artificial graphite anode material. The binder and anode raw material powder are mixed and then subjected to pelletizing and graphitization. After graphitization, some material remains bound together in the pellets. After crushing, dispersing, and sieving, secondary granulated artificial graphite anode products are obtained. Compared with the traditional production process, this reduces the granulation step, significantly lowers production costs, and improves the rate capability and cycle performance of the material.

[0004] To address the common problems in this field, such as the reliance on experience-based settings for key processes like granulation, shaping, and drying, and the lack of online quantitative characterization and closed-loop correction mechanisms for precursor particle morphology and size distribution, leading to issues like particle tailing, insufficient sphericity, and excessive fine powder adhering to the surface, resulting in increased compaction density fluctuations and expansion risks; the heating and holding regimes during the carbonization stage typically employ fixed temperature curves or minor empirical adjustments, failing to fully consider the coupled influence of precursor particle characteristics and inert atmosphere conditions on the escape of pyrolysis volatiles and internal pressure defects, easily causing defects such as cracks, blistering, and pores, or incomplete carbonization, thus affecting the structural stability after subsequent graphitization; and the insufficient adaptability of existing processes to different operating conditions under fluctuating raw materials, varying charge amounts, and changes in inert atmosphere purity / dryness, making it difficult to simultaneously achieve low expansion and high compaction density targets, and leaving room for improvement in batch consistency and yield, this invention was developed. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings existing in the field by proposing a low-expansion, high-density artificial graphite material and its preparation method.

[0006] To overcome the shortcomings of the prior art, the present invention adopts the following technical solution:

[0007] A method for preparing a low-expansion, high-density artificial graphite material, the method comprising:

[0008] S1: Select carbon-containing raw materials such as petroleum coke, needle coke, or pitch coke, and calcine, crush, grind, and screen the raw materials to obtain carbonaceous precursor powder that meets the target particle size range;

[0009] S2: The carbonaceous precursor powder is mixed with a binder, and then granulated, spheroidized, and dried sequentially through a granulation device, a shaping device, and a drying device to obtain precursor particles with controlled morphology and particle size distribution. Furthermore, image information of the precursor particles is acquired through an image acquisition device, and the image information is input to a graphics processing unit to obtain particle characteristic parameters of the precursor particles. When the morphology and particle size distribution of the precursor particles do not meet the preset range, the granulation parameters, shaping parameters, and / or drying parameters are adjusted by a correction unit to bring the particle size distribution and morphology of the precursor particles back to the preset range.

[0010] The particle characteristic parameters include 10% cumulative quantile particle size D10, 50% cumulative quantile particle size D50, 90% cumulative quantile particle size D90, sphericity, roundness, aspect ratio and the proportion of fine powder adhering to the surface. The binder includes, but is not limited to, asphalt, resin or a binder system formed by compounding asphalt and resin in a certain proportion.

[0011] S3: The precursor particles are placed in a carbonization furnace and heated and kept at a temperature according to a matching temperature curve under an inert atmosphere to pyrolyze and carbonize the binder to form a carbonized precursor; the matching temperature curve is obtained based on particle characteristic parameters and output through a temperature matching model; the inert atmosphere includes nitrogen, argon, helium or a combination of nitrogen, argon and helium.

[0012] S4: The carbonization precursor is further heated to about 2500-3000 ℃ in an inert atmosphere and held at that temperature to cause the carbon structure to undergo an ordered transformation and increase the degree of graphitization, thereby obtaining a graphitized product.

[0013] S5: The graphitized products are cooled, crushed, and screened for classification. After acid washing and surface coating modification, artificial graphite materials are finally obtained.

[0014] Furthermore, the image processing unit processes image information through the following steps:

[0015] S201: The image information includes a two-dimensional image of the precursor particles. The two-dimensional image is scaled so that the pixel size in the two-dimensional image corresponds to the actual length unit.

[0016] S202: Perform image preprocessing on the two-dimensional image. The image preprocessing includes at least one of grayscale conversion, noise reduction filtering, background correction, and contrast enhancement to reduce the impact of noise and uneven illumination on precursor particle recognition.

[0017] S203: Perform image segmentation on the two-dimensional image after image preprocessing to separate the precursor particles from the background and obtain candidate regions of the precursor particles; the image segmentation adopts existing threshold segmentation methods or edge detection segmentation methods, the threshold segmentation methods include Otsu threshold segmentation or adaptive threshold segmentation, and the edge detection segmentation methods include Canny edge detection;

[0018] S204: Morphological processing is performed on the candidate regions of the precursor particles to remove noise regions and fill pores; when there are adhering precursor particles, adhering separation processing is performed on the adhering regions to obtain single-particle regions; the morphological processing includes at least one of opening operation, closing operation, erosion and expansion, and the adhering separation processing includes a separation method of distance transformation combined with watershed segmentation or a label-based separation method.

[0019] S205: Perform connected component labeling and contour extraction on the precursor particles to obtain the contour information of each precursor particle, and extract the basic geometric feature parameters of the precursor particles based on the contour information. The basic geometric feature parameters include projected area, perimeter, and major axis and minor axis.

[0020] S206: Determine the equivalent particle size of each precursor particle based on the basic geometric feature parameters, and form an equivalent particle size sample set; convert the equivalent particle size sample set into actual particle size units according to the scale calibration.

[0021] Wherein, the equivalent particle size represents the diameter of the equivalent circle that has the same projected area as the precursor particle in the two-dimensional image;

[0022] S207: Construct a cumulative particle size distribution based on the equivalent particle size sample set. The cumulative particle size distribution is the cumulative distribution obtained by sorting the equivalent particle sizes of all precursor particles in the same production batch image information in ascending order. And determine the 10% cumulative quantile particle size D10, the 50% cumulative quantile particle size D50 and the 90% cumulative quantile particle size D90 according to the cumulative particle size distribution.

[0023] D10 refers to the particle size value corresponding to a cumulative proportion of 10% in the cumulative particle size distribution; D50 refers to the particle size value corresponding to a cumulative proportion of 50% in the cumulative particle size distribution; and D90 refers to the particle size value corresponding to a cumulative proportion of 90% in the cumulative particle size distribution.

[0024] S208: Calculate the sphericity, roundness, and aspect ratio of the precursor particles based on the contour information and basic geometric feature parameters, wherein roundness is used to characterize the degree to which the projected contour of the precursor particles is close to a circle, aspect ratio is used to characterize the ratio of the major axis to the minor axis of the precursor particles, and sphericity is used to characterize the degree to which the precursor particles are close to a sphere.

[0025] S209: Based on the contour information, determine the boundary neighborhood of each precursor particle. The boundary neighborhood is an annular region surrounding the contour of the precursor particle. The annular region can be obtained by expanding and / or shrinking the contour by a preset width. Within the boundary neighborhood, identify connected regions and determine small connected regions that meet a preset area threshold as surface-attached fine powder. Calculate the total projected area of ​​the surface-attached fine powder within the boundary neighborhood as the attached fine powder area. Simultaneously, calculate the projected area of ​​the boundary neighborhood as the neighborhood area. Then, define the ratio of the attached fine powder area to the neighborhood area as the surface-attached fine powder ratio.

[0026] S210: Statistically summarize and output the D10, D50, D90, sphericity, roundness, aspect ratio, and proportion of fine powder adhering to the surface of all precursor particles in the image information.

[0027] Furthermore, the morphology and grain size parameters are referred to as the morphology-grain size index M, which is calculated as follows:

[0028] ,

[0029] This represents the average percentage of fine powder adhering to the surface of all precursor particles within the same production batch's image information. This represents the average sphericity of all precursor particles within the same production batch's image information.

[0030] Furthermore, the granulation device, shaping device, and drying device can be implemented using equipment existing in the art, and their specific structures will not be described in detail here; however, the granulation device, shaping device, and drying device each have at least adjustable granulation parameters, shaping parameters, and drying parameters, and the parameters can be adjusted under the control of the correction unit;

[0031] The granulation parameters include at least the binder addition rate; the shaping parameters include at least the classifying wheel speed, classifying air volume, and shaping rotor speed; the drying parameters include at least the drying temperature, drying air volume, and / or residence time.

[0032] Furthermore, the specific operating steps of the correction unit are as follows:

[0033] S301: Obtain the particle characteristic parameters and calculate the particle size index M; when the particle size index M meets the preset range, the correction unit keeps the operating parameters of the granulation device, shaping device and drying device unchanged or operates according to the predetermined process; when the particle size index M does not meet the preset range, the correction unit further determines whether the parameters related to particle size distribution in the particle characteristic parameters exceed the preset range.

[0034] S302: When the average D50 of all precursor particles in the same batch of image information is higher than the preset upper limit of D50 and the average D90 is higher than the preset upper limit of D90, the correction unit outputs the first correction command to the shaping device to adjust the grading wheel speed or grading air volume in the shaping parameters of the shaping device, so as to reduce the proportion of coarse particle size of precursor particles and bring the particle size distribution back to the preset range.

[0035] S303: When the average D10 of all precursor particles in the same batch of image information is lower than the preset lower limit of D10, the correction unit outputs a second correction command to the granulation device to adjust the binder addition rate in the granulation parameters of the granulation device, so as to reduce the proportion of fine particle size of the precursor particles and bring the particle size distribution back to the preset range.

[0036] S304: When the average sphericity of all precursor particles in the same batch of image information is lower than the preset lower limit of sphericity or the average roundness is lower than the preset lower limit of roundness or the average aspect ratio is higher than the preset upper limit of aspect ratio, the correction unit outputs a third correction command to the shaping device to adjust the shaping rotor speed in the shaping parameters of the shaping device in order to improve the degree of sphericity and reduce the aspect ratio.

[0037] S305: When the surface is covered with fine powder proportion When the ratio of surface-attached fine powder exceeds the preset upper limit, the correction unit outputs a fourth correction command to the drying device to adjust the drying temperature, drying air volume, or residence time in the drying parameters of the drying device, so as to reduce the ratio of surface-attached fine powder and suppress fine powder adhesion caused by particle surface adhesion.

[0038] S306: After executing the first correction instruction, the second correction instruction, the third correction instruction, or the fourth correction instruction, the image acquisition device is triggered to acquire image information of the updated precursor particles again, and the image processing unit updates and outputs particle feature parameters. The correction unit recalculates the particle appearance index M based on the updated particle feature parameters and repeats steps S301–S305 until the particle appearance index M meets the preset range.

[0039] Furthermore, the training steps for the temperature matching model are as follows:

[0040] S401: Using precursor particles from the same production batch as a training sample, collect and record the particle characteristic parameters and inert atmosphere parameters of that batch. The particle characteristic parameters include D10, D50, D90, and sphericity. Roundness C, aspect ratio ar, and the proportion of fine powder adhering to the surface rf;

[0041] Inert atmosphere parameters include the total inert gas flow rate F and oxygen content in the carbonization furnace. The furnace pressure P and the inert atmosphere of the carbonization furnace measured by the dew point meter are cooled to the condensate temperature Ltem at the current pressure to which condensate begins to appear; at the same time, the performance indicators of the carbonization precursor obtained after carbonization of this batch are recorded as reference indicators for training targets. The performance indicators include the compaction density of the carbonization precursor, the carbonization expansion rate and the defect index. When the performance indicators of the carbonization precursor meet the production requirements, the batch of samples is marked as a valid training sample; otherwise, it is discarded.

[0042] S402: To uniformly characterize the effects of total inert gas flow rate, oxygen content, condensate temperature, and pressure on volatile matter removal capacity and atmosphere protection capability, an inert atmosphere effectiveness index is constructed. And use it as one of the derived features of the model input or as a training constraint variable:

[0043] ,

[0044] This is a reference value for traffic volume; This is a reference value for oxygen content; This is a reference value for the condensate temperature under standard charging and stable operating conditions in the carbonization furnace; This is the scaling factor for condensate temperature deviation, used to characterize the sensitivity of condensate temperature deviation to the decay of the inert atmosphere effectiveness index. This is a pressure reference value; The pressure effect weighting coefficient is positive;

[0045] The physical meaning of the inert atmosphere effectiveness index is: the higher the inert flow rate, the lower the oxygen content, the smaller the deviation of the condensate temperature, and the more favorable the pressure, the stronger the comprehensive ability of the inert atmosphere to discharge and protect volatiles.

[0046] S403: Based on the particle size distribution, morphology, and surface-attached fine powder of precursor particles from the same production batch, an exhaust risk index is constructed. And used to generate a piecewise platform strategy for matching temperature curves or as a constraint feature of the temperature matching model, the satisfy:

[0047] ,

[0048] in, The surface-attached fine powder weighting coefficient is used to characterize the influence of the proportion of surface-attached fine powder on the blockage of the gas release path and surface adhesion. The morphology deviation weighting coefficient is used to characterize the influence of sphericity deviation on particle packing pores and gas escape paths;

[0049] S404: Constructing the input feature vector The input feature vector is in the following form:

[0050] ,

[0051] Among them, These represent the corresponding parameters. , , The normalized value, and the normalization or standardization can be achieved by min-max normalization;

[0052] S405: Parameterize the matching temperature curve of the carbonization process into an executable segmented curve, wherein the matching temperature curve includes a first heating segment, a first plateau segment, a second heating segment, a second plateau segment, a third heating segment, and a third plateau segment set sequentially; wherein the first heating segment is characterized by a first heating rate. Characterization, the first plateau segment is characterized by the first plateau temperature. With the first heat preservation time Characterization; the second heating stage is characterized by the second heating rate. Characterization, the second plateau segment is characterized by the second plateau temperature. With the second heat preservation time Characterization; the third heating stage is characterized by the third heating rate. Characterization, the third plateau segment is characterized by the third plateau temperature. With the third insulation time Characterization;

[0053] Based on this, the curve parameter vector is defined. : ;

[0054] S406: Construct a training sample set based on historical production data and process verification data, and input feature vectors of precursor particles from the same production batch. As model input, the parameter vector of the target matching temperature curve corresponding to this batch is used. As supervisory labels; neural network regression model, random forest regression model, and gradient boosting tree regression model were selected for training; during training, the parameter vector of the curve output by the model was used as the supervisory label. With the supervision label The prediction error between the two is used as the optimization objective. Training is completed by iteratively updating the model parameters to obtain a model that can predict based on the input feature vector. The output temperature matching model is a temperature matching curve parameter vector.

[0055] A low-expansion, high-density artificial graphite material is obtained according to steps S1-S5 in the method for preparing artificial graphite materials.

[0056] The beneficial effects achieved by this invention are:

[0057] 1. By introducing image acquisition equipment and graphics processing units into the granulation, shaping and drying stages, online quantitative characterization of precursor particles' D10, D50, D90, sphericity, roundness, aspect ratio and the proportion of fine powder adhering to the surface is performed. When the preset range is not met, the correction unit performs closed-loop adjustment of granulation parameters, shaping parameters and / or drying parameters to stabilize the particle size distribution and morphology of precursor particles back to the target range, thereby reducing particle size tailing and morphology fluctuations, and improving batch consistency and preparation yield.

[0058] 2. By constructing a particle morphology index M, the sphericity, particle size distribution width, and the degree of fine powder adhesion on the surface are comprehensively characterized, enabling unified evaluation and rapid determination of precursor particle morphology and particle size. This facilitates the establishment of a correspondence between correction commands and specific process parameters, thereby achieving directional control over the coarse-grained end, fine-grained end, and degree of sphericity, improving particle packing density, and reducing the risk of defects caused by fine powder adhesion, thus providing a prerequisite for obtaining high compaction density.

[0059] 3. By coupling the precursor particle characteristic parameters with the inert atmosphere parameters of the carbonization furnace through a temperature matching model, and introducing the inert atmosphere effectiveness index and gas release risk index as derived features, the matching temperature curve of the carbonization process can be adaptively adjusted according to the fluctuation of raw materials, changes in charge amount and inert atmosphere conditions. This reduces the internal pressure defects and abnormal expansion caused by thermal degassing, improves the structural stability and densification basis of the carbonization precursor, and ultimately helps to obtain low-expansion, high-pressure dense artificial graphite materials. Attached Figure Description

[0060] The invention will be further understood from the following description taken in conjunction with the accompanying drawings. The components in the drawings are not necessarily drawn to scale, but rather the emphasis is on illustrating the principles of the embodiments. In different views, the same reference numerals designate corresponding parts.

[0061] Figure 1 This is a schematic flowchart of the method for preparing low-expansion, high-density artificial graphite material according to the present invention.

[0062] Figure 2 This is a flowchart illustrating the graphics processing unit of the present invention.

[0063] Figure 3 This is a flowchart illustrating the correction unit of the present invention.

[0064] Figure 4 This is a schematic diagram of the correlation between the inert atmosphere effectiveness index and the defect index of the present invention.

[0065] Figure 5 This is a schematic diagram of the correlation between the gas release risk index and the defect index of the present invention. Detailed Implementation

[0066] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to its embodiments. It should be noted that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of this invention. Other systems, methods, and / or features of this embodiment will become apparent to those skilled in the art after reviewing the following detailed description. Furthermore, the terminology used to describe positional relationships in the accompanying drawings is for illustrative purposes only and should not be construed as limiting this patent. Those skilled in the art can understand the specific meaning of the above terms according to the specific circumstances.

[0067] Example 1: Combined with Appendix Figure 1 Appendix Figure 2 Appendix Figure 3 Appendix Figure 4 and attached Figure 5This embodiment constructs a low-expansion, high-density artificial graphite material and its preparation method. The method for preparing the low-expansion, high-density artificial graphite material includes:

[0068] S1: Select carbon-containing raw materials such as petroleum coke, needle coke, or pitch coke, and calcine, crush, grind, and screen the raw materials to obtain carbonaceous precursor powder that meets the target particle size range;

[0069] S2: The carbonaceous precursor powder is mixed with a binder, and then granulated, spheroidized, and dried sequentially through a granulation device, a shaping device, and a drying device to obtain precursor particles with controlled morphology and particle size distribution. Furthermore, image information of the precursor particles is acquired through an image acquisition device, and the image information is input to a graphics processing unit to obtain particle characteristic parameters of the precursor particles. When the morphology and particle size distribution of the precursor particles do not meet the preset range, the granulation parameters, shaping parameters, and / or drying parameters are adjusted by a correction unit to bring the particle size distribution and morphology of the precursor particles back to the preset range.

[0070] The particle characteristic parameters include 10% cumulative quantile particle size D10, 50% cumulative quantile particle size D50, 90% cumulative quantile particle size D90, sphericity, roundness, aspect ratio and the proportion of fine powder adhering to the surface. The binder includes, but is not limited to, asphalt, resin or a binder system formed by compounding asphalt and resin in a certain proportion.

[0071] The image acquisition equipment includes, but is not limited to, industrial camera and telecentric lens combination, microscopic imaging system or dynamic image particle size analyzer.

[0072] S3: The precursor particles are placed in a carbonization furnace and heated and kept at a temperature according to a matching temperature curve under an inert atmosphere to pyrolyze and carbonize the binder to form a carbonized precursor; the matching temperature curve is obtained based on particle characteristic parameters and output through a temperature matching model; the inert atmosphere includes nitrogen, argon, helium or a combination of nitrogen, argon and helium.

[0073] S4: The carbonization precursor is further heated to about 2500-3000 ℃ in an inert atmosphere and held at that temperature to cause the carbon structure to undergo an ordered transformation and increase the degree of graphitization, thereby obtaining a graphitized product.

[0074] S5: The graphitized products are cooled, crushed, and screened for classification. After acid washing and surface coating modification, artificial graphite materials are finally obtained.

[0075] The image processing unit processes image information through the following steps:

[0076] S201: The image information includes a two-dimensional image of the precursor particles. The two-dimensional image is scaled so that the pixel size in the two-dimensional image corresponds to the actual length unit.

[0077] S202: Perform image preprocessing on the two-dimensional image. The image preprocessing includes at least one of grayscale conversion, noise reduction filtering, background correction, and contrast enhancement to reduce the impact of noise and uneven illumination on precursor particle recognition.

[0078] S203: Perform image segmentation on the two-dimensional image after image preprocessing to separate the precursor particles from the background and obtain candidate regions of the precursor particles; the image segmentation adopts existing threshold segmentation methods or edge detection segmentation methods, the threshold segmentation methods include Otsu threshold segmentation or adaptive threshold segmentation, and the edge detection segmentation methods include Canny edge detection;

[0079] S204: Morphological processing is performed on the candidate regions of the precursor particles to remove noise regions and fill pores; when there are adhering precursor particles, adhering separation processing is performed on the adhering regions to obtain single-particle regions; the morphological processing includes at least one of opening operation, closing operation, erosion and expansion, and the adhering separation processing includes a separation method of distance transformation combined with watershed segmentation or a label-based separation method.

[0080] S205: Perform connected component labeling and contour extraction on the precursor particles to obtain the contour information of each precursor particle, and extract the basic geometric feature parameters of the precursor particles based on the contour information. The basic geometric feature parameters include projected area, perimeter, and major axis and minor axis.

[0081] S206: Determine the equivalent particle size of each precursor particle based on the basic geometric feature parameters, and form an equivalent particle size sample set; convert the equivalent particle size sample set into actual particle size units according to the scale calibration.

[0082] Wherein, the equivalent particle size represents the diameter of the equivalent circle that has the same projected area as the precursor particle in the two-dimensional image;

[0083] S207: Construct a cumulative particle size distribution based on the equivalent particle size sample set. The cumulative particle size distribution is the cumulative distribution obtained by sorting the equivalent particle sizes of all precursor particles in the same production batch image information in ascending order. And determine the 10% cumulative quantile particle size D10, the 50% cumulative quantile particle size D50 and the 90% cumulative quantile particle size D90 according to the cumulative particle size distribution.

[0084] D10 refers to the particle size value corresponding to a cumulative proportion of 10% in the cumulative particle size distribution; D50 refers to the particle size value corresponding to a cumulative proportion of 50% in the cumulative particle size distribution; and D90 refers to the particle size value corresponding to a cumulative proportion of 90% in the cumulative particle size distribution.

[0085] S208: Calculate the sphericity, roundness, and aspect ratio of the precursor particles based on the contour information and basic geometric feature parameters, wherein roundness is used to characterize the degree to which the projected contour of the precursor particles is close to a circle, aspect ratio is used to characterize the ratio of the major axis to the minor axis of the precursor particles, and sphericity is used to characterize the degree to which the precursor particles are close to a sphere.

[0086] S209: Based on the contour information, determine the boundary neighborhood of each precursor particle. The boundary neighborhood is an annular region surrounding the contour of the precursor particle. The annular region can be obtained by expanding and / or shrinking the contour by a preset width. Within the boundary neighborhood, identify connected regions and determine small connected regions that meet a preset area threshold as surface-attached fine powder. Calculate the total projected area of ​​the surface-attached fine powder within the boundary neighborhood as the attached fine powder area. Simultaneously, calculate the projected area of ​​the boundary neighborhood as the neighborhood area. Then, define the ratio of the attached fine powder area to the neighborhood area as the surface-attached fine powder ratio.

[0087] S210: Statistically summarize and output the D10, D50, D90, sphericity, roundness, aspect ratio, and proportion of fine powder adhering to the surface of all precursor particles in the image information.

[0088] The morphology and grain size parameters are referred to as the morphology-grain size index M, which is calculated as follows:

[0089] ,

[0090] This represents the average percentage of fine powder adhering to the surface of all precursor particles within the same production batch's image information. This represents the average sphericity of all precursor particles within the same production batch's image information.

[0091] The granulation device, shaping device, and drying device can be implemented using existing equipment in the art, and their specific structures will not be described in detail here; however, the granulation device, shaping device, and drying device each have at least adjustable granulation parameters, shaping parameters, and drying parameters, and the parameters can be adjusted under the control of the correction unit.

[0092] The granulation parameters include at least the binder addition rate; the shaping parameters include at least the classifying wheel speed, classifying air volume, and shaping rotor speed; the drying parameters include at least the drying temperature, drying air volume, and / or residence time.

[0093] The classifying wheel speed and classifying air volume are used to perform gas-solid classification on the shaped precursor particles, thereby controlling the particle size distribution. The classifying wheel speed is used to adjust the classifying cutting particle size to change the separation boundary between coarse and fine particles, and the classifying air volume is used to adjust the carrying and separation capacity of the classification field to change the ratio of fine particles being carried out and coarse particles being retained, thereby controlling the proportion of coarse particles and D50 and D90. The shaping rotor speed and reflux ratio are used to adjust the impact, friction and shear intensity and the number of repeated shapings experienced by the precursor particles in the shaping device. The shaping rotor speed is used to adjust the shaping intensity to improve sphericity and roundness and reduce the aspect ratio, and the reflux ratio is used to adjust the proportion of particles repeatedly entering the shaping zone to improve morphological consistency and control the particle size distribution in conjunction with the classification parameters.

[0094] The specific operating steps of the correction unit are as follows:

[0095] S301: Obtain the particle characteristic parameters and calculate the particle size index M; when the particle size index M meets the preset range, the correction unit keeps the operating parameters of the granulation device, shaping device and drying device unchanged or operates according to the predetermined process; when the particle size index M does not meet the preset range, the correction unit further determines whether the parameters related to particle size distribution in the particle characteristic parameters exceed the preset range.

[0096] S302: When the average D50 of all precursor particles in the same batch of image information is higher than the preset upper limit of D50 and the average D90 is higher than the preset upper limit of D90, the correction unit outputs the first correction command to the shaping device to adjust the grading wheel speed or grading air volume in the shaping parameters of the shaping device, so as to reduce the proportion of coarse particle size of precursor particles and bring the particle size distribution back to the preset range.

[0097] S303: When the average D10 of all precursor particles in the same batch of image information is lower than the preset lower limit of D10, the correction unit outputs a second correction command to the granulation device to adjust the binder addition rate in the granulation parameters of the granulation device, so as to reduce the proportion of fine particle size of the precursor particles and bring the particle size distribution back to the preset range.

[0098] S304: When the average sphericity of all precursor particles in the same batch of image information is lower than the preset lower limit of sphericity or the average roundness is lower than the preset lower limit of roundness or the average aspect ratio is higher than the preset upper limit of aspect ratio, the correction unit outputs a third correction command to the shaping device to adjust the shaping rotor speed in the shaping parameters of the shaping device in order to improve the degree of sphericity and reduce the aspect ratio.

[0099] S305: When the surface is covered with fine powder proportion When the ratio of surface-attached fine powder exceeds the preset upper limit, the correction unit outputs a fourth correction command to the drying device to adjust the drying temperature, drying air volume, or residence time in the drying parameters of the drying device, so as to reduce the ratio of surface-attached fine powder and suppress fine powder adhesion caused by particle surface adhesion.

[0100] S306: After executing the first correction instruction, the second correction instruction, the third correction instruction, or the fourth correction instruction, the image acquisition device is triggered to acquire image information of the updated precursor particles again, and the image processing unit updates and outputs particle feature parameters. The correction unit recalculates the particle appearance index M based on the updated particle feature parameters and repeats steps S301–S305 until the particle appearance index M meets the preset range.

[0101] The aforementioned preset upper or lower limits can be determined based on the target product specifications and statistical analysis of valid training samples. Preferably, for qualified batches that meet the requirements of low expansion and high compaction density, the value ranges of D10, D50, D90, sphericity, roundness, aspect ratio, and the proportion of fine powder adhering to the surface are statistically analyzed. The lower boundary of the D10 value range is set as the preset lower limit of D10, the upper boundary of the D50 value range is set as the preset upper limit of D50, the upper boundary of the D90 value range is set as the preset upper limit of D90, the lower boundary of the sphericity value range is set as the preset lower limit of sphericity, the lower boundary of the roundness value range is set as the preset lower limit of roundness, the upper boundary of the aspect ratio value range is set as the preset upper limit of aspect ratio, and the upper boundary of the surface fine powder adhering to the surface fine powder proportion is set as the preset upper limit of surface fine powder proportion. Furthermore, the threshold values ​​can be periodically updated according to raw material fluctuations and equipment status.

[0102] In summary, the technical solution of this embodiment takes "visualization of precursor particle quality and closed-loop process control" as its core idea: In S2, online image analysis of precursor particles is performed through image acquisition equipment and graphics processing unit, outputting particle characteristic parameters including D10, D50, D90, sphericity, roundness, aspect ratio, and the proportion of fine powder adhering to the surface, and further constructing a particle size index M as a comprehensive characterization index of morphology and particle size; when the particle characteristic parameters and particle size index M deviate from the preset range, the correction unit performs directional correction and adjustment of the operating parameters of the granulation device, shaping device, and drying device to ensure the quality of precursor particles. The particle size distribution and morphology stabilize back to the target range, reducing the impact of particle tailing, fine powder adhesion, and morphology fluctuations on subsequent processes from the source. In S3, a temperature matching model is further introduced, which uses particle characteristic parameters and inert atmosphere parameters in the carbonization furnace as inputs and outputs matching temperature curve parameters to achieve adaptive matching of carbonization heating rate, plateau temperature, and holding time. This effectively suppresses defects and abnormal expansion caused by thermal decomposition gas under conditions of raw material fluctuations and changes in inert atmosphere, and improves the structural stability and densification basis of carbonization precursors. Ultimately, this is conducive to obtaining low-expansion, high-density artificial graphite materials.

[0103] Example 2: Combined with Appendix Figure 1 Appendix Figure 2 Appendix Figure 3 Appendix Figure 4 and attached Figure 5 In addition to the content of the above embodiments, the training steps of the temperature matching model are as follows:

[0104] S401: Using precursor particles from the same production batch as a training sample, collect and record the particle characteristic parameters and inert atmosphere parameters of that batch. The particle characteristic parameters include D10, D50, D90, and sphericity. Roundness C, aspect ratio ar, and the proportion of fine powder adhering to the surface rf;

[0105] Inert atmosphere parameters include the total inert gas flow rate F and oxygen content in the carbonization furnace. The furnace pressure P and the inert atmosphere of the carbonization furnace, as measured by the dew point meter, are cooled to the condensate temperature Ltem at the current pressure, at which condensate begins to appear.

[0106] Simultaneously, the performance indicators of the carbonized precursor obtained after carbonization are recorded to determine the validity of the batch of samples. The performance indicators include the compaction density of the carbonized precursor, the carbonization expansion rate, and the defect index. When the performance indicators of the carbonized precursor meet the production requirements (the compaction density of the carbonized precursor is not lower than the production density threshold, the carbonization expansion rate is not higher than the production expansion threshold, and the defect index is not higher than the production defect threshold), the batch of samples is marked as valid training samples; otherwise, they are discarded to reduce the interference of abnormal batches on the temperature matching model training.

[0107] The defect index qc is used to characterize the degree of defects such as cracks, blistering and abnormal pores in the carbonized precursor. The defect index qc is determined by the ratio of the defect area in the image of the carbonized precursor to the projected area of ​​the sample.

[0108] S402: To uniformly characterize the effects of total inert gas flow rate, oxygen content, condensate temperature, and pressure on volatile matter removal capacity and atmosphere protection capability, an inert atmosphere effectiveness index is constructed. And use it as one of the derived features of the model input or as a training constraint variable:

[0109] ,

[0110] This is a reference value for traffic volume; This is a reference value for oxygen content; This is a reference value for the condensate temperature under standard charging and stable operating conditions in the carbonization furnace; This is the scaling factor for condensate temperature deviation, used to characterize the sensitivity of condensate temperature deviation to the decay of the inert atmosphere effectiveness index. This is a pressure reference value; The pressure effect weighting coefficient is positive;

[0111] The total flow rate of inert gas in the carbonization furnace is measured in real time by a mass flow meter, vortex flow meter or rotor flow meter in the carbonization furnace inlet or circulating gas line, and the average value of the stable period is taken. The unit is L / min. The oxygen content of the inert atmosphere inside the carbonization furnace is measured online by an oxygen content analyzer and the average value over a stable period is taken. The unit is volume fraction. The dew point temperature is the inert atmosphere inside the carbonization furnace. Its value is measured online by a dew point meter. The dew point temperature is the temperature at which the inert atmosphere begins to condense when cooled to the current pressure. The unit is degrees Celsius. The pressure inside the carbonization furnace is measured by the furnace body pressure sensor or pressure transmitter and the average value of the stable period is taken. The unit is kilopascal.

[0112] The inert gas flow rate is a reference value, which is the rated total inert gas flow rate under standard charging and stable operating conditions in the carbonization furnace. It is preferably the median or mean of the total inert gas flow rate in the effective training sample. The oxygen content reference value is taken as the representative value of the oxygen content in the effective training sample, preferably the median or upper quartile value of the oxygen content of qualified batches to characterize the safety margin. It is a reference value for the condensate temperature under standard charging and stable operating conditions in the carbonization furnace. The value is taken as a representative value of the dew point temperature in the effective training sample, preferably the median or mean of the dew point temperature of qualified batches. This is the condensate temperature deviation scaling factor, used to characterize the sensitivity of condensate temperature deviation to the decay of the effectiveness index. Its value is obtained by calibration from historical data, preferably taken as the standard deviation of the dew point temperature of qualified batches or a preset allowable deviation (e.g., half of the allowable dew point temperature range). ; This is a pressure reference value, which is the median or mean of the rated furnace pressure or the effective training sample furnace pressure under stable operating conditions. The pressure influence weighting coefficient is set to a positive value, and its value is obtained by fitting or cross-validating historical production data and process verification data. Preferably, The range of values ​​is .

[0113] The positive values ​​of the pressure influence weighting coefficient are obtained by calibration using historical production data and process verification data. Specifically, based on historical batch data, furnace pressure, reference pressure (as a normalization benchmark), and the carbonization expansion rate and defect index of the corresponding batch of carbonization precursor are extracted. The deviation between the model prediction results and the actual measured carbonization expansion rate and defect index under different furnace pressure conditions is fitted. The pressure influence weighting coefficient is adjusted and calibrated by minimizing the prediction residual and combining it with the defect judgment threshold constraint fitting method. The final value of the pressure influence weighting coefficient is obtained so that the sensitivity of the inert atmosphere effectiveness index to changes in furnace pressure matches the actual defect risk of the carbonization process, thereby improving the stability and generalization performance of the temperature matching model under different charge amounts and different inert atmosphere conditions.

[0114] The physical meaning of the inert atmosphere effectiveness index is: the higher the inert flow rate, the lower the oxygen content, the smaller the deviation of the condensate temperature, and the more favorable the pressure, the stronger the comprehensive ability of the inert atmosphere to discharge and protect volatiles.

[0115] S403: Based on the particle size distribution, morphology, and surface-attached fine powder of precursor particles from the same production batch, an exhaust risk index is constructed. And used to generate a piecewise platform strategy for matching temperature curves or as a constraint feature of the temperature matching model, the satisfy:

[0116] ,

[0117] in, , is the weighting coefficient for surface-attached fine powder, taken as a positive value, used to characterize the influence of the proportion of surface-attached fine powder on the blockage of the gas release path and surface adhesion; This is the morphology deviation weighting coefficient, which is positive and used to characterize the influence of sphericity deviation on particle packing channels and gas escape paths.

[0118] , The values ​​are obtained by calibration using historical production data and process verification data. Specifically, on the training sample set, with carbonization expansion rate and defect index as constraints, grid search or cross-validation methods are used to select the values ​​within a preset range that minimize the model's prediction error. , Preferably, , The value range is 0.1 to 10.

[0119] S404: Constructing the input feature vector The input feature vector satisfies the following form:

[0120] ,

[0121] Among them, These represent the corresponding parameters. , , The normalized value, whereby the normalization or standardization can be achieved using minimum-maximum normalization.

[0122] S405: Parameterize the matching temperature curve of the carbonization process into an executable segmented curve, wherein the matching temperature curve includes a first heating segment, a first plateau segment, a second heating segment, a second plateau segment, a third heating segment, and a third plateau segment set sequentially; wherein the first heating segment is characterized by a first heating rate. Characterization, the first plateau segment is characterized by the first plateau temperature. With the first heat preservation time Characterization; the second heating stage is characterized by the second heating rate. Characterization, the second plateau segment is characterized by the second plateau temperature. With the second heat preservation time Characterization; the third heating stage is characterized by the third heating rate. Characterization, the third plateau segment is characterized by the third plateau temperature. With the third insulation time Characterization;

[0123] Based on this, the curve parameter vector is defined. : ;

[0124] S406: Construct a training sample set based on historical production data and process verification data, and input feature vectors of precursor particles from the same production batch. As model input, the parameter vector of the target matching temperature curve corresponding to this batch is used. As supervisory labels; neural network regression model, random forest regression model, and gradient boosting tree regression model were selected for training; during training, the parameter vector of the curve output by the model was used as the supervisory label. With the supervision label The prediction error between the two is used as the optimization objective. Training is completed by iteratively updating the model parameters to obtain a model that can predict based on the input feature vector. A temperature matching model is output, which is a parameter vector of the temperature curve. The predictive performance of the temperature matching model is quantitatively evaluated and the parameters are tuned using at least one of the root mean square error (RMSE) and mean absolute error (MAE). The model is validated using a cross-validation strategy to reduce the evaluation bias caused by different batch sample divisions and to improve the model's generalization ability to raw material fluctuations and changes in inert atmosphere conditions.

[0125] As attached Figure 4 and attached Figure 5 As shown in the sample verification experiment of this invention, the inert atmosphere effectiveness index and the defect index show a negative correlation trend. That is, under relatively stable particle characteristic parameters and other process conditions, a higher inert atmosphere effectiveness index indicates a stronger comprehensive ability of the inert atmosphere to carry away volatiles and protect the atmosphere, and the corresponding defect index is lower. At the same time, the gas release risk index and the defect index show a positive correlation trend. That is, a higher gas release risk index indicates a higher risk of obstruction of the volatile escape path and internal pressure, and the corresponding defect index is larger. Based on the above correlation, the inert atmosphere effectiveness index and the gas release risk index can be used as input features and / or constraint features of the temperature matching model to guide the segmented heating and plateau setting of the matching temperature curve, thereby reducing the probability of defects such as cracks, bubbling and abnormal pores during carbonization and improving the quality stability of the carbonized precursor.

[0126] In summary, this second embodiment, based on the first embodiment, further introduces a training and construction mechanism for a temperature matching model. Using production batches as units, it jointly models the particle size distribution, morphology, and surface-attached fine powder characteristics of the precursor particles with the inert atmosphere parameters within the carbonization furnace. An inert atmosphere effectiveness index and a gas release risk index are constructed as derived features, and the performance indicators of the carbonized precursor after carbonization are used to effectively screen the training samples. This allows the temperature matching model to output matching segmented heating and platform heat preservation strategies under different raw material fluctuations, different charge amounts, and changes in inert atmosphere conditions. This solves the problems of existing fixed carbonization temperature regimes, which struggle to balance gas release pressure control and carbonization sufficiency, easily leading to defects such as bubbling and cracking, as well as excessive expansion and large fluctuations in compaction density. By establishing a data-driven mapping relationship between "particle state—atmosphere state—temperature curve," the adaptability and stability of the carbonization process are improved, the risk of defects and abnormal expansion is reduced, and the consistency of the carbonized precursor is enhanced, providing support for ultimately obtaining low-expansion, high-compaction-density artificial graphite materials.

[0127] While the invention has been described above with reference to various embodiments, it should be understood that many changes and modifications can be made without departing from the scope of the invention. That is, the methods, systems, and devices discussed above are examples. Various configurations can be appropriately omitted, substituted, or added to various processes or components. For example, in alternative configurations, methods can be performed in a different order than described, and / or various components can be added, omitted, and / or combined. Moreover, features described with respect to certain configurations can be combined in various other configurations, such as different aspects and elements of the configuration can be combined in a similar manner. Furthermore, the elements therein can be updated as the technology develops; many elements are examples and do not limit the scope of this disclosure or the claims. It should also be understood that after reading the description of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent changes and modifications also fall within the scope defined by the claims of this invention.

Claims

1. A method for preparing a low-expansion, high-compact-density artificial graphite material, characterized in that, The preparation method of the low-expansion, high-compact-density artificial graphite material includes: S1: Select carbonaceous raw materials such as petroleum coke, needle coke, or pitch coke, and calcine, crush, grind, and screen the raw materials to obtain carbonaceous precursor powder that meets the target particle size range; S2: The carbonaceous precursor powder is mixed with a binder, and then granulated, spheroidized, and dried sequentially through a granulation device, a shaping device, and a drying device to obtain precursor particles with controlled morphology and particle size distribution. Furthermore, image information of the precursor particles is acquired through an image acquisition device, and the image information is input to a graphics processing unit to obtain particle characteristic parameters of the precursor particles. When the morphology and particle size distribution of the precursor particles do not meet the preset range, the granulation parameters, shaping parameters, and / or drying parameters are adjusted by a correction unit to bring the particle size distribution and morphology of the precursor particles back to the preset range. The particle characteristic parameters include 10% cumulative quantile particle size D10, 50% cumulative quantile particle size D50 and 90% cumulative quantile particle size D90, sphericity, roundness, aspect ratio and the proportion of fine powder adhering to the surface, and the binder includes, but is not limited to, asphalt, resin or a binder system formed by compounding asphalt and resin in a certain proportion. S3: The precursor particles are placed in a carbonization furnace and heated and kept at a temperature according to a matching temperature curve under an inert atmosphere to pyrolyze and carbonize the binder to form a carbonized precursor; the matching temperature curve is obtained based on particle characteristic parameters and output through a temperature matching model; the inert atmosphere includes nitrogen, argon, helium or a combination of nitrogen, argon and helium. S4: The carbonization precursor is further heated to about 2500-3000 ℃ in an inert atmosphere and held at that temperature to cause the carbon structure to undergo an ordered transformation and increase the degree of graphitization, thereby obtaining a graphitized product. S5: The graphitized products are cooled, crushed, and screened for classification. After acid washing and surface coating modification, artificial graphite materials are finally obtained.

2. The method for preparing artificial graphite material as described in claim 1, characterized in that, The image processing unit processes image information through the following steps: S201: The image information includes a two-dimensional image of the precursor particles. The two-dimensional image is scaled so that the pixel size in the two-dimensional image corresponds to the actual length unit. S202: Perform image preprocessing on the two-dimensional image. The image preprocessing includes at least one of grayscale conversion, noise reduction filtering, background correction, and contrast enhancement to reduce the impact of noise and uneven illumination on precursor particle recognition. S203: Perform image segmentation on the two-dimensional image after image preprocessing to separate the precursor particles from the background and obtain candidate regions of the precursor particles; the image segmentation adopts existing threshold segmentation methods or edge detection segmentation methods, the threshold segmentation methods include Otsu threshold segmentation or adaptive threshold segmentation, and the edge detection segmentation methods include Canny edge detection; S204: Morphological processing is performed on the candidate regions of the precursor particles to remove noise regions and fill pores; when there are adhering precursor particles, adhering separation processing is performed on the adhering regions to obtain single-particle regions; the morphological processing includes at least one of opening operation, closing operation, erosion and expansion, and the adhering separation processing includes a separation method of distance transformation combined with watershed segmentation or a label-based separation method. S205: Perform connected component labeling and contour extraction on the precursor particles to obtain the contour information of each precursor particle, and extract the basic geometric feature parameters of the precursor particles based on the contour information. The basic geometric feature parameters include projected area, perimeter, and major axis and minor axis. S206: Determine the equivalent particle size of each precursor particle based on the basic geometric feature parameters, and form an equivalent particle size sample set; convert the equivalent particle size sample set into actual particle size units according to the scale calibration. Wherein, the equivalent particle size represents the diameter of the equivalent circle that has the same projected area as the precursor particle in the two-dimensional image; S207: Construct a cumulative particle size distribution based on the equivalent particle size sample set. The cumulative particle size distribution is the cumulative distribution obtained by sorting the equivalent particle sizes of all precursor particles in the same production batch image information in ascending order. And determine the 10% cumulative quantile particle size D10, the 50% cumulative quantile particle size D50 and the 90% cumulative quantile particle size D90 according to the cumulative particle size distribution. D10 refers to the particle size value corresponding to a cumulative proportion of 10% in the cumulative particle size distribution; D50 refers to the particle size value corresponding to a cumulative proportion of 50% in the cumulative particle size distribution; and D90 refers to the particle size value corresponding to a cumulative proportion of 90% in the cumulative particle size distribution. S208: Calculate the sphericity, roundness, and aspect ratio of the precursor particles based on the contour information and basic geometric feature parameters, wherein roundness is used to characterize the degree to which the projected contour of the precursor particles is close to a circle, aspect ratio is used to characterize the ratio of the major axis to the minor axis of the precursor particles, and sphericity is used to characterize the degree to which the precursor particles are close to a sphere. S209: Based on the contour information, determine the boundary neighborhood of each precursor particle. The boundary neighborhood is an annular region surrounding the contour of the precursor particle. The annular region can be obtained by expanding and / or shrinking the contour by a preset width. Within the boundary neighborhood, identify connected regions and determine small connected regions that meet a preset area threshold as surface-attached fine powder. Calculate the total projected area of ​​the surface-attached fine powder within the boundary neighborhood as the attached fine powder area. Simultaneously, calculate the projected area of ​​the boundary neighborhood as the neighborhood area. Then, define the ratio of the attached fine powder area to the neighborhood area as the surface-attached fine powder ratio. S210: Statistically summarize and output the D10, D50, D90, sphericity, roundness, aspect ratio, and proportion of fine powder adhering to the surface of all precursor particles in the image information.

3. The method for preparing artificial graphite material as described in claim 2, characterized in that, The morphology and grain size parameters are referred to as the morphology-grain size index M, which is calculated as follows: ， This represents the average percentage of fine powder adhering to the surface of all precursor particles within the same production batch's image information. This represents the average sphericity of all precursor particles within the same production batch's image information.

4. The method for preparing artificial graphite material as described in claim 3, characterized in that, The granulation device, shaping device, and drying device each have at least adjustable granulation parameters, shaping parameters, and drying parameters, and the parameters can be adjusted under the control of the correction unit. The granulation parameters include at least the binder addition rate; the shaping parameters include at least the classifying wheel speed, classifying air volume, and shaping rotor speed; the drying parameters include at least the drying temperature, drying air volume, and / or residence time.

5. The method for preparing artificial graphite material as described in claim 4, wherein the specific operating steps of the correction unit are as follows: S301: Obtain the particle characteristic parameters and calculate the particle size index M; when the particle size index M meets the preset range, the correction unit keeps the operating parameters of the granulation device, shaping device and drying device unchanged or operates according to the predetermined process; when the particle size index M does not meet the preset range, the correction unit further determines whether the parameters related to particle size distribution in the particle characteristic parameters exceed the preset range. S302: When the average D50 of all precursor particles in the same batch of image information is higher than the preset upper limit of D50 and the average D90 is higher than the preset upper limit of D90, the correction unit outputs the first correction command to the shaping device to adjust the grading wheel speed or grading air volume in the shaping parameters of the shaping device, so as to reduce the proportion of coarse particle size of precursor particles and bring the particle size distribution back to the preset range. S303: When the average D10 of all precursor particles in the same batch of image information is lower than the preset lower limit of D10, the correction unit outputs a second correction command to the granulation device to adjust the binder addition rate in the granulation parameters of the granulation device, so as to reduce the proportion of fine particle size of the precursor particles and bring the particle size distribution back to the preset range. S304: When the average sphericity of all precursor particles in the same batch of image information is lower than the preset lower limit of sphericity or the average roundness is lower than the preset lower limit of roundness or the average aspect ratio is higher than the preset upper limit of aspect ratio, the correction unit outputs a third correction command to the shaping device to adjust the shaping rotor speed in the shaping parameters of the shaping device in order to improve the degree of sphericity and reduce the aspect ratio. S305: When the surface is covered with fine powder proportion When the ratio of surface-attached fine powder exceeds the preset upper limit, the correction unit outputs a fourth correction command to the drying device to adjust the drying temperature, drying air volume, or residence time in the drying parameters of the drying device, so as to reduce the ratio of surface-attached fine powder and suppress fine powder adhesion caused by particle surface adhesion. S306: After executing the first correction instruction, the second correction instruction, the third correction instruction, or the fourth correction instruction, the image acquisition device is triggered to acquire image information of the updated precursor particles again, and the image processing unit updates and outputs particle feature parameters. The correction unit recalculates the particle appearance index M based on the updated particle feature parameters and repeats steps S301–S305 until the particle appearance index M meets the preset range.

6. The method for preparing artificial graphite material as described in claim 5, characterized in that, The training steps for the temperature matching model are as follows: S401: Using precursor particles from the same production batch as a training sample, collect and record the particle characteristic parameters and inert atmosphere parameters of that batch. The particle characteristic parameters include D10, D50, D90, and sphericity. Roundness C, aspect ratio ar, and the proportion of fine powder adhering to the surface rf; Inert atmosphere parameters include the total inert gas flow rate F and oxygen content in the carbonization furnace. The furnace pressure P and the inert atmosphere of the carbonization furnace measured by the dew point meter are cooled to the condensate temperature Ltem at the current pressure to which condensate begins to appear; at the same time, the performance indicators of the carbonization precursor obtained after carbonization of this batch are recorded as reference indicators for training targets. The performance indicators include the compaction density of the carbonization precursor, the carbonization expansion rate and the defect index. When the performance indicators of the carbonization precursor meet the production requirements, the batch of samples is marked as a valid training sample; otherwise, it is discarded. S402: To uniformly characterize the effects of total inert gas flow rate, oxygen content, condensate temperature, and pressure on volatile matter removal capacity and atmosphere protection capability, an inert atmosphere effectiveness index is constructed. And use it as one of the derived features of the model input or as a training constraint variable: ， This is a reference value for traffic volume; This is a reference value for oxygen content; This is a reference value for the condensate temperature under standard charging and stable operating conditions in the carbonization furnace; This is the scaling factor for condensate temperature deviation, used to characterize the sensitivity of condensate temperature deviation to the decay of the inert atmosphere effectiveness index. This is a pressure reference value; The pressure effect weighting coefficient is positive; The physical meaning of the inert atmosphere effectiveness index is: the higher the inert flow rate, the lower the oxygen content, the smaller the deviation of the condensate temperature, and the more favorable the pressure, the stronger the comprehensive ability of the inert atmosphere to discharge and protect volatiles. S403: Based on the particle size distribution, morphology, and surface-attached fine powder of precursor particles from the same production batch, an exhaust risk index is constructed. And used to generate a piecewise platform strategy for matching temperature curves or as a constraint feature of the temperature matching model, the satisfy: ， in, The surface-attached fine powder weighting coefficient is used to characterize the influence of the proportion of surface-attached fine powder on the blockage of the gas release path and surface adhesion. The morphology deviation weighting coefficient is used to characterize the influence of sphericity deviation on particle packing pores and gas escape paths; S404: Constructing the input feature vector The input feature vector is in the following form: , Among them, These represent the corresponding parameters. , , The normalized value, and the normalization or standardization can be achieved by min-max normalization; S405: Parameterize the matching temperature curve of the carbonization process into an executable segmented curve, wherein the matching temperature curve includes a first heating segment, a first plateau segment, a second heating segment, a second plateau segment, a third heating segment, and a third plateau segment set sequentially; wherein the first heating segment is characterized by a first heating rate. Characterization, the first plateau segment is characterized by the first plateau temperature. With the first heat preservation time Characterization; the second heating stage is characterized by the second heating rate. Characterization, the second plateau segment is characterized by the second plateau temperature. With the second heat preservation time Characterization; the third heating stage is characterized by the third heating rate. Characterization, the third plateau segment is characterized by the third plateau temperature. With the third insulation time Characterization; Based on this, the curve parameter vector is defined. : ; S406: Construct a training sample set based on historical production data and process verification data, and input feature vectors of precursor particles from the same production batch. As model input, the parameter vector of the target matching temperature curve corresponding to this batch is used. As supervisory labels; neural network regression model, random forest regression model, and gradient boosting tree regression model were selected for training; during training, the parameter vector of the curve output by the model was used as the supervisory label. With the supervision label The prediction error between the two is used as the optimization objective. Training is completed by iteratively updating the model parameters to obtain a model that can predict based on the input feature vector. The output temperature matching model is a temperature matching curve parameter vector.

7. A low-expansion, high-density artificial graphite material, prepared according to any one of claims 1 to 6.

Images