Lfp positive electrode high-density homogenate process optimization method and system and storage medium

Abstract

The application relates to the technical field of lithium battery material preparation, and discloses an LFP positive electrode high-density homogenate process optimization method, a system and a storage medium. The method comprises the following steps: collecting slurry state and process parameter data in real time, analyzing a correlation strength index, identifying an abnormal slurry state mode, extracting a quantitative feature index, obtaining historical upstream and downstream data if the quantitative feature index is not in a preset range, combining an initial slurry state to evaluate the influence degree of upstream fluctuation, generating a risk level, formulating a compensation scheme according to the risk level, verifying the effectiveness of the scheme through virtual simulation, generating an adjustment instruction and delivering the adjustment instruction to a production control system, updating process parameters and storing the process parameters in a historical database, and the application improves the self-adaptive capacity of the homogenate process to upstream fluctuation.

Description

Technical Field

[0001] This application relates to the field of lithium battery material preparation technology, and in particular to an optimized method, system and storage medium for high-density homogenization process of LFP cathode. Background Technology

[0002] Lithium iron phosphate (LFP), as a cathode material for lithium-ion batteries, is widely used in electric vehicles and energy storage due to its high safety, long cycle life, and low cost. LFP materials typically undergo a carbon coating process to form a carbon layer on the surface of the particles to improve conductivity. This is followed by a slurry preparation process, where the LFP is mixed with conductive agents, binders, and solvents to prepare a high-density slurry for electrode coating. The slurry preparation process is a crucial step in lithium battery electrode fabrication, and its quality directly affects coating consistency and the final battery performance. High-density slurry requires a high solids content while maintaining stable viscosity and good particle dispersion uniformity.

[0003] However, in actual production, the homogenization process faces numerous challenges. First, the upstream carbon coating process exhibits inherent fluctuations. Factors such as temperature changes and cooling rates can lead to batch-to-batch variations in the thickness and uniformity of the carbon layer on the LFP material surface. While these variations may be within the acceptable range for material testing, subtle changes in the carbon layer characteristics significantly impact the rheological behavior and dispersion properties of the slurry during the high-density homogenization process, resulting in viscosity fluctuations and uneven dispersion across different batches. Traditional homogenization process control often employs a post-event feedback model. When abnormal slurry quality is detected, the machine must be shut down for investigation, parameters adjusted based on experience, and production restarted for verification. This control method exhibits significant lag, leading to material waste and reduced production efficiency. Furthermore, the effectiveness of adjustments depends on operator experience, making it difficult to guarantee batch consistency. Existing technologies lack an effective response mechanism to upstream process fluctuations. The homogenization process is typically treated as an isolated step, with fixed parameter settings, failing to proactively adapt to changes in the characteristics of upstream incoming materials. When the upstream carbon coating process fluctuates, the homogenization process can only passively bear the impact, resulting in unstable slurry quality.

[0004] Therefore, how to establish a method that can proactively sense fluctuations in the upstream carbon coating process, intelligently assess its impact on the homogenization process, and dynamically adjust homogenization parameters to compensate for upstream fluctuations has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] To address the aforementioned technical issues, this application provides an optimization method, system, and storage medium for the high-density homogenization process of LFP cathodes, which improves the adaptability of the homogenization process to upstream fluctuations, enhances batch stability of slurry viscosity and dispersion uniformity, and reduces trial-and-error costs and material waste in the production line.

[0006] In a first aspect, this application provides an optimization method for the high-density homogenization process of LFP cathode, the method comprising:

[0007] Step S1: Real-time acquisition of slurry state data and process parameter data in the LFP cathode high-density homogenization process, analysis of the correlation strength index between slurry state changes and process parameter adjustments, and identification of abnormal slurry state patterns in the current batch.

[0008] Step S2: Extract the quantitative feature index corresponding to the abnormal slurry state pattern. If the quantitative feature index is not within the preset range, obtain the historical upstream carbon coating process correlation data and downstream process response data, and combine them with the initial slurry state of the current batch to determine the degree of influence of upstream fluctuations on the current homogenization process and generate the homogenization process instability risk level.

[0009] Step S3: Based on the risk level of instability in the homogenization process, formulate a homogenization process compensation scheme for the current batch, run the homogenization process simulation through the virtual simulation module, verify the effectiveness of the process compensation scheme in dealing with fluctuations in the upstream carbon coating process, and obtain simulation verification results;

[0010] Step S4: Based on the verified process compensation scheme, generate the final homogenization process parameter adjustment instruction, and send the adjustment instruction to the homogenization production control system to update the process operation parameters of the current batch and store them in the historical database.

[0011] Secondly, this application provides an optimization system for high-density homogenization process of LFP cathode, the system comprising:

[0012] The data acquisition module is used to acquire slurry state data and process parameter data in the high-density homogenization process of LFP cathode in real time, analyze the correlation strength index between slurry state changes and process parameter adjustments, and identify abnormal slurry state patterns in the current batch.

[0013] The judgment module is used to extract the quantitative feature index corresponding to the abnormal slurry state pattern. If the quantitative feature index is not within the preset range, the historical upstream carbon coating process correlation data and downstream process response data are obtained, and combined with the initial slurry state of the current batch, the influence of upstream fluctuations on the current homogenization process is judged, and the homogenization process instability risk level is generated.

[0014] The verification module is used to formulate a compensation scheme for the homogenization process for the current batch based on the risk level of instability of the homogenization process, and to run a simulation of the homogenization process through the virtual simulation module to verify the effectiveness of the compensation scheme in dealing with fluctuations in the upstream carbon coating process and obtain simulation verification results.

[0015] The execution module is used to generate the final homogenization process parameter adjustment instruction based on the verified process compensation scheme, and send the adjustment instruction to the homogenization production control system to update the process operation parameters of the current batch and store them in the historical database.

[0016] A third aspect of this application provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the aforementioned method for optimizing the high-density homogenization process of an LFP cathode.

[0017] Compared with the prior art, the beneficial effects of the present invention are at least as follows:

[0018] This application constructs a complete technical chain from anomaly identification, source analysis, scheme formulation, simulation verification to closed-loop control, achieving proactive perception and adaptive compensation of fluctuations in the upstream carbon coating process during homogenization, yielding significant beneficial effects. First, by collecting slurry state and process parameter data in real time and analyzing correlation strength indicators, it can accurately identify abnormal slurry state patterns in the current batch, providing accurate decision-making basis for subsequent process optimization. Second, by extracting quantitative characteristic indicators and comparing them with preset ranges, it matches historical batches and analyzes upstream and downstream correlations when they exceed the limits, quantifying the disturbance impact of upstream carbon coating process fluctuations on downstream slurry and generating risk levels. This achieves a quantitative assessment of the impact of upstream fluctuations, avoiding the blindness of relying on experience-based judgments in traditional methods.

[0019] Furthermore, this application extracts adjustment rules from the rule base based on risk level to formulate a compensation scheme, and conducts pre-verification through a virtual simulation module to ensure that the compensation scheme is fully validated before actual application, avoiding the risks and costs associated with trial and error on the production line. Finally, the validated adjustment instructions are sent to the production control system to update process parameters and store them in the database, forming a closed-loop mechanism for continuous optimization. Through the above technical solutions, this application significantly improves the batch stability of slurry viscosity and particle dispersion uniformity, shortens the anomaly response time from several hours to minutes, improves the batch pass rate, and significantly reduces material waste caused by trial and error adjustments, providing an intelligent, data-driven process optimization solution for the preparation of lithium battery cathode materials. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1This is a flowchart of an optimization method for high-density homogenization process of LFP cathode in an embodiment of this application;

[0022] Figure 2 This is a schematic diagram comparing the effects of the present application's embodiments with those of existing technologies;

[0023] Figure 3 This is a schematic diagram of the structure of an LFP cathode high-density homogenization process optimization system in an embodiment of this application. Detailed Implementation

[0024] This application provides an optimized method, system, and storage medium for high-density homogenization process of LFP cathodes. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0025] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 An optimization method for high-density homogenization process of LFP cathode in this application includes:

[0026] Step S1: Real-time acquisition of slurry state data and process parameter data in the LFP cathode high-density homogenization process, analysis of the correlation strength index between slurry state changes and process parameter adjustments, and identification of abnormal slurry state patterns in the current batch.

[0027] The step S1, analyzing the correlation strength index between changes in slurry state and adjustments to process parameters, includes: extracting slurry viscosity and particle size distribution values ​​from slurry state data, and extracting corresponding shear rate and stirring time values ​​from process parameter data; calculating a first correlation coefficient between slurry viscosity and shear rate values, and a second correlation coefficient between particle size distribution and stirring time values; and generating a correlation strength index to characterize the correlation between slurry state and process parameters based on the first and second correlation coefficients.

[0028] Specifically, in the LFP cathode high-density homogenization process, slurry viscosity and particle dispersion uniformity are the core indicators determining the electrode coating quality, while shear rate and stirring time are key process parameters for controlling these states. However, traditional homogenization processes often use fixed parameter modes, failing to perceive the dynamic correlation between process parameter adjustments and slurry state changes in real time, leading to delayed anomaly identification. Therefore, this application first collects slurry state and process parameter data in real time using sensors and analyzes the correlation strength between the two, thereby accurately identifying abnormal slurry state patterns in the current batch and laying the foundation for subsequent process optimization. Specifically, after the LFP cathode high-density homogenization process starts running, slurry state data is collected in real time using an online viscometer and an online particle size analyzer installed in the homogenization equipment. The online viscometer continuously monitors the viscosity changes of the slurry during stirring, collecting viscosity values ​​once per second to generate viscosity time-series data reflecting the rheological properties of the slurry. The online particle size analyzer simultaneously collects the particle size distribution in the slurry, outputting a particle size distribution curve every ten seconds, including the particle size value corresponding to the cumulative distribution percentage of particles. Meanwhile, the control system of the homogenizing equipment reads current process parameter data in real time, including shear rate, stirring time, and solvent addition flow rate. The shear rate refers to the velocity gradient between the rotor and stator during the homogenizing process, usually expressed as reciprocal of a second or revolutions per minute (RPM). It is a core process parameter characterizing the intensity of shearing action; a higher shear rate indicates a stronger shearing effect on the particles, which is beneficial for breaking up agglomerates. However, excessively high shear rates may lead to secondary agglomeration of already dispersed particles or overheating of the slurry system, therefore it needs to be controlled within a reasonable range. The stirring time refers to the duration from the start to the stop of stirring in the homogenizing equipment. The stirring time is usually measured in minutes. This parameter is a key factor in determining whether particles can obtain sufficient dispersion energy in the slurry. The longer the stirring time, the longer the cumulative time of shear force acting on the particle agglomerates, which is conducive to the gradual opening of agglomerates and uniform distribution in the slurry system. However, excessive stirring time may lead to increased slurry temperature, solvent evaporation, or secondary agglomeration of dispersed particles. Therefore, it is necessary to set a reasonable stirring time range according to process requirements. The solvent addition flow rate refers to the volume or mass flow rate of liquid solvent added to the slurry per unit time, usually expressed in liters per minute or kilograms per hour. This parameter is the main means of adjusting the solid content of the slurry and directly affects the concentration and rheological properties of the slurry.

[0029] These process parameter data and slurry state data are aligned with a unified timestamp to form a process dataset reflecting the real-time changes in the homogenization process. After obtaining the above process dataset, the slurry state data sequence and process parameter data sequence are extracted from the dataset. Specifically, the slurry viscosity value and particle size distribution value are extracted from the slurry state data, and the particle size distribution value specifically includes D10 particle size, D50 particle size, and D90 particle size; the shear rate value and stirring time value corresponding to the time of the slurry state data are extracted from the process parameter data to ensure that each set of slurry state data has matching process parameter data, forming a one-to-one data pair.

[0030] Then, correlation analysis was used to calculate the correlation strength between slurry state data and process parameter data. First, the first correlation coefficient between slurry viscosity and shear rate was calculated. In the homogenization process, shear rate is the core means of adjusting the rheological properties of the slurry, and shearing directly affects the change in slurry viscosity; therefore, there is an inherent physical correlation between the two. Calculating the first correlation coefficient is precisely to quantitatively assess the strength of this correlation, using the Pearson correlation coefficient as a quantification tool, with a value range of -1 to 1. Simultaneously, the second correlation coefficient between particle size distribution and stirring time was calculated. Particle dispersion in the slurry is a kinetic process, requiring sufficient time for shear force to act on each agglomerate, gradually opening it and distributing it evenly in the slurry system. Therefore, there is a clear causal relationship between stirring time and particle dispersion uniformity. Calculating the second correlation coefficient is precisely to quantitatively assess whether the current stirring time setting is within the effective dispersion range, again using the Pearson correlation coefficient as a quantification tool. By calculating these two correlation coefficients, the dynamic response relationship between process parameters and slurry state can be comprehensively evaluated from different dimensions, providing a reliable quantitative basis for subsequent identification of abnormal slurry state patterns.

[0031] Based on the first and second correlation coefficients, a correlation strength index is generated to characterize the relationship between slurry state and process parameters. This correlation strength index is a comprehensive quantification of the two correlation coefficients, used to evaluate the overall effect of process parameter adjustments on slurry state under current process conditions. In generating the correlation strength index, the first and second correlation coefficients are first normalized, converting them into positive indices within the range of 0 to 1. Then, according to the actual needs of process control, different weight coefficients are assigned to the normalized first and second correlation coefficients. In high-density homogenization processes, viscosity stability has a more direct impact on subsequent coating processes, therefore the normalized first correlation coefficient is given a higher weight, such as 0.6. Dispersion uniformity, as a secondary control objective, is given a lower weight, such as 0.4, to the normalized second correlation coefficient. The first weighted value is obtained by multiplying the normalized first correlation coefficient by its corresponding weight, and the second weighted value is obtained by multiplying the normalized second correlation coefficient by its corresponding weight. The sum of these two values ​​yields the correlation strength index, which represents the degree of matching between process parameter adjustments and slurry state changes under current process conditions.

[0032] The correlation strength index generated in the above manner integrates the correlation information of the two dimensions into a comprehensive evaluation value. It not only retains the individual effects of shear rate on viscosity and stirring time on dispersion, but also reflects the overall control efficiency of the current process parameter system, providing a simple and intuitive quantitative basis for subsequent identification of abnormal slurry state patterns.

[0033] The step S1, identifying the abnormal slurry state mode of the current batch, includes: comparing the correlation strength index with a preset benchmark correlation range; if it is lower than the lower limit, it is determined to be a first abnormal mode; if it is higher than the upper limit, it is determined to be a second abnormal mode; for the first abnormal mode, based on the particle agglomeration inhibition, identifying whether there is poor dispersion due to insufficient shear rate or uneven mixing due to excessively short stirring time, and assigning a corresponding specific mode; for the second abnormal mode, based on the particle agglomeration inhibition, identifying whether there is sensitive abnormality due to fluctuations in incoming material characteristics or viscosity mutation due to improper solvent addition, and assigning a corresponding specific mode; and determining the abnormal slurry state mode of the current batch based on the identified specific mode, and outputting the corresponding mode type.

[0034] Specifically, after obtaining the correlation strength index, it needs to be compared with a preset benchmark correlation range to determine whether the current batch's process operation status is within the normal range. The benchmark correlation range is preset based on the statistical distribution of correlation strength indices in historical normal production batches. Usually, the mean of the correlation strength index under normal production conditions plus or minus two standard deviations is taken as the upper and lower limits. When the correlation strength index falls within the benchmark correlation range, it indicates that the control relationship between the process parameters and the slurry state of the current batch is within the normal fluctuation range, and the process operation is stable. When the correlation strength index is lower than the lower limit of the benchmark correlation range, it means that the process parameter adjustment has insufficient control effect on the slurry state, that is, the change in shear rate has failed to effectively cause a viscosity response or the change in stirring time has failed to effectively improve particle dispersion. At this time, it is judged as the first abnormal mode. When the correlation strength index is higher than the upper limit of the benchmark correlation range, it means that the slurry state is too sensitive to the adjustment of process parameters. Small parameter adjustments cause drastic changes in the slurry state, indicating that the slurry system may be in an unstable critical state. At this time, it is judged as the second abnormal mode.

[0035] For the first abnormal mode, further analysis of particle agglomeration inhibition is needed to identify the specific cause. Particle agglomeration inhibition refers to the volume ratio and trend of particle agglomerates in the slurry, obtained through real-time monitoring with an online particle size analyzer, and is usually expressed as a percentage of agglomerate volume. For the first abnormal mode, if the agglomerate volume ratio is consistently higher than a preset ratio and shows an upward trend at the current shear rate, and the absolute value of the first correlation coefficient between shear rate and viscosity is lower than a preset threshold, the shear rate's effect on viscosity regulation is ineffective. That is, increasing the shear rate fails to effectively reduce the slurry viscosity, and the agglomerate volume ratio continues to rise, indicating that the current shear rate is insufficient to break up the particle agglomerates, exacerbating the poor dispersion problem. This situation is identified as a poor dispersion mode caused by insufficient shear rate. If, during the current stirring time, a D50 particle size (D50 particle size refers to the particle size corresponding to a cumulative distribution reaching 50%) is detected... If the particle size (D50 particle size) does not decrease significantly or even increases with prolonged stirring time, and the volume ratio of agglomerates does not improve, and the absolute value of the second correlation coefficient between stirring time and D50 particle size is lower than a preset threshold, then the regulating effect of stirring time on particle dispersion is ineffective. In other words, extending the stirring time fails to effectively reduce the D50 particle size; instead, the particle size tends to increase, and the volume ratio of agglomerates does not improve. This situation indicates that the current stirring time is insufficient to achieve sufficient particle dispersion, or that excessive stirring time causes secondary agglomeration of already dispersed particles. This is identified as a mixing unevenness mode caused by improper stirring time. By refining the first abnormal mode into two specific situations—poor dispersion due to insufficient shear rate and mixing unevenness due to improper stirring time—precise location of the cause of process control failure is achieved.

[0036] For the second abnormal mode, it is also necessary to identify the specific cause by combining the particle agglomeration inhibition status. For the second abnormal mode, if the viscosity value is detected to fluctuate significantly in a short period of time under the current process parameters, and the fluctuation amplitude exceeds the preset proportion of the normal fluctuation range, and the fluctuation change is not proportional to the shear rate adjustment, and the particle agglomeration inhibition status shows that the agglomerate volume ratio is not obviously abnormal and is within the normal fluctuation range, and the historical database shows that the upstream carbon coating process of the corresponding batch has a record of the carbon layer thickness distribution standard deviation exceeding the preset threshold, then it indicates that the drastic viscosity fluctuation is not caused by particle agglomeration, but rather by the change in material surface properties caused by the fluctuation of the upstream carbon coating process, which affects the slurry system. The material exhibits abnormal sensitivity to process parameter adjustments, identified as a sensitive anomaly caused by fluctuations in incoming material characteristics. If a viscosity mutation is detected simultaneously with a record of solvent addition flow rate adjustments, and the mutation amplitude is not linearly related to the amount of solvent added, and the volume ratio of agglomerates changes drastically in a short period of time when combined with particle agglomeration suppression monitoring, it means that the adjustment of solvent addition flow rate has disrupted the stability of the slurry system. The normal correspondence between solvent addition and viscosity change has been broken. Excessive or insufficient solvent causes drastic changes in the interparticle forces, leading to the re-agglomeration of dispersed particles or local concentration imbalance in the slurry. This situation is identified as a viscosity mutation pattern caused by improper solvent addition.

[0037] Based on the specific patterns identified above, the abnormal slurry state pattern of the current batch caused by the mismatch between process parameters and slurry state is determined. If any specific pattern is identified, the corresponding pattern type identifier is output. For example, when the pattern is identified as poor dispersion due to insufficient shear rate, the output pattern type identifier is "shear failure, poor dispersion"; when the pattern is identified as uneven mixing due to insufficient stirring time, the output pattern type identifier is "insufficient stirring, uneven mixing"; when the pattern is identified as sensitive abnormality due to fluctuations in incoming material characteristics, the output pattern type identifier is "sensitive fluctuation abnormality in incoming material"; when the pattern is identified as viscosity mutation due to improper solvent addition, the output pattern type identifier is "improper solvent addition". Each pattern is assigned a unique identifier, which will be used in subsequent steps to trigger the corresponding risk assessment and process compensation process to achieve targeted abnormality handling.

[0038] By comparing the correlation strength index with the benchmark correlation range and identifying the first and second abnormal modes in a hierarchical manner, and then combining the particle agglomeration inhibition situation to further identify the specific causes, this step achieves accurate classification and tracing of abnormal slurry state modes. This provides a clear guide for the subsequent targeted development of process compensation schemes, avoids blind adjustments, and significantly improves the efficiency and accuracy of abnormal handling.

[0039] Step S2: Extract the quantitative feature index corresponding to the abnormal slurry state pattern. If the quantitative feature index is not within the preset range, obtain the historical upstream carbon coating process correlation data and downstream process response data, and combine them with the initial slurry state of the current batch to determine the degree of influence of upstream fluctuations on the current homogenization process and generate the homogenization process instability risk level.

[0040] Step S2 includes: comparing the quantitative characteristic index with a preset range; if it is not within the preset range, retrieving historical batches that match the current quantitative characteristic index from the historical database, extracting upstream carbon coating process correlation data and downstream process response data of the historical batches, analyzing the first correlation between upstream carbon layer thickness distribution and downstream slurry viscosity, and the second correlation between upstream coating uniformity and downstream particle dispersion uniformity; based on the first and second correlations, combined with the initial slurry state of the current batch, calculating the disturbance impact of upstream carbon coating process fluctuations on the current homogenization process, mapping the disturbance impact level to a preset risk level threshold, and determining the homogenization process instability risk level.

[0041] Specifically, after identifying the abnormal slurry state pattern of the current batch and obtaining the corresponding pattern type, it is necessary to further assess whether the anomaly has reached a severity level requiring tracing upstream causes, and to quantify the actual impact of fluctuations in the upstream carbon coating process on the downstream homogenization process, thereby providing a basis for decision-making in subsequent process compensation. First, extract the quantitative characteristic indicators corresponding to the identified abnormal slurry state patterns. Quantitative characteristic indicators refer to feature parameters that can quantitatively describe the severity of the abnormal pattern. Different abnormal patterns correspond to different quantitative characteristic indicators. For example, for the shear failure and poor dispersion pattern, the root cause is insufficient shear rate, leading to the ineffective opening of particle agglomerates. Therefore, the severity of this pattern directly reflects the quality of particle dispersion. A quantitative characteristic indicator can be selected as the deviation between the D50 particle size and the target D50 particle size in the current batch. The D50 particle size refers to the particle size value corresponding to a cumulative distribution reaching 50%. When the shear rate is insufficient, the agglomerates cannot be fully dispersed. In cases of dispersion, the D50 particle size will be significantly higher than the target value. The larger this deviation, the more severe the poor dispersion caused by shear failure. For the insufficient stirring and uneven mixing mode, the root cause is that insufficient stirring time prevents the particles from obtaining sufficient dispersion energy in the slurry, and the agglomerates cannot be fully opened within a limited time. Therefore, the severity of this mode is directly reflected in the evolution of the particle dispersion state over time. A quantitative characteristic index can be selected as the rate of change of D50 particle size with stirring time in the current batch, that is, the decrease in particle size per unit time. The lower the rate of change, the more severe the uneven mixing caused by insufficient stirring time. For the incoming material sensitive fluctuation abnormal mode, the root cause is... The surface properties of LFP materials change due to fluctuations in the upstream carbon coating process, causing the slurry system to exhibit abnormal sensitivity to adjustments in process parameters. Therefore, the severity of this pattern is directly reflected in the drastic response of the slurry state to fine-tuning of process parameters. A quantifiable characteristic index can be the ratio of the viscosity fluctuation amplitude in the current batch to the normal fluctuation range. Specifically, under the same process parameter adjustment range, if the characteristics of the incoming material fluctuate, the slurry viscosity will exhibit a drastic change far exceeding the normal level. This abnormal sensitivity is quantified by the ratio of the actual viscosity fluctuation amplitude to the historical average fluctuation amplitude of normal batches. The larger this ratio, the more abnormal the sensitivity to fluctuations in the incoming material. The severity of this mode is directly reflected in the degree of abnormal viscosity response to solvent addition. The root cause of the improper solvent addition mode is the mismatch between the solvent addition flow rate and the actual demand of the slurry system, which leads to abnormal viscosity changes. Therefore, the severity of this mode is directly reflected in the degree of abnormal viscosity response to solvent addition. The quantitative characteristic index can be the ratio of the viscosity change amplitude to the amount of solvent added in the current batch. Specifically, under normal process conditions, the amount of solvent added and the viscosity change should show a stable correspondence. When the solvent addition is improper, even a small change in the amount of solvent added will cause a violent fluctuation in viscosity. This ratio deviates significantly from the historical normal range. The larger the ratio, the more severe the viscosity change caused by improper solvent addition.

[0042] After extracting the quantitative characteristic indicators corresponding to each abnormal mode, the extracted quantitative characteristic indicators are compared with the preset normal range to determine whether the current abnormality should trigger the subsequent analysis process. The preset normal range is set according to the statistical distribution of the corresponding indicators in historical normal production batches. For example, for the D50 particle size deviation value corresponding to the shear failure dispersion poor mode, its normal range can be set to -0.3μm to 0.3μm. When the D50 particle size deviation value falls within this range, it indicates that the dispersion state is normal; when the D50 particle size deviation value is lower than the lower limit or higher than the upper limit, it indicates that the degree of dispersion poorness has exceeded the process's own adjustment capability. When any quantitative characteristic indicator exceeds its corresponding preset normal range, it means that the process state of the current batch has deviated from the normal fluctuation range. The abnormality has reached the severity level that requires the initiation of the subsequent analysis process, and further investigation of the cause of the abnormality and assessment of the risk level are required through historical data matching and upstream and downstream correlation analysis.

[0043] Furthermore, historical batches matching the current quantitative characteristic indicators are retrieved from the historical database. The historical database pre-stores upstream carbon coating process data and corresponding downstream homogenization process data for each batch of LFP material. A similarity matching algorithm, such as Euclidean distance, is used to calculate the distance between the current quantitative characteristic indicator and the corresponding historical batch. The historical batch with the smallest distance is selected as the matching historical batch. Upstream carbon coating process correlation data is extracted from these batches, including records of carbon layer thickness distribution, coating uniformity indicators, and temperature fluctuation range records in the carbon coating process. Simultaneously, using these matching batches as indexes, the downstream process response data for the corresponding batches is queried from the historical database, including slurry viscosity control records. The data includes solid content ratio records and particle dispersion uniformity data. The slurry viscosity control record refers to the data sequence of slurry viscosity changes over time, collected and stored in real-time using online viscometers and other equipment during historical batch homogenization processes. The solid content ratio record refers to the actual mixing ratio data of LFP material with solvents, binders, conductive agents, and other components during historical batch homogenization processes. The particle dispersion uniformity data refers to the particle size distribution data in the slurry collected and stored in real-time using an online particle size analyzer during historical batches. This data is typically expressed as D10, D50, and D90 particle sizes, where D50 particle size refers to the particle size value corresponding to a cumulative distribution reaching 50%, and is the core indicator characterizing particle dispersion uniformity.

[0044] Based on the aforementioned upstream and downstream data, we analyzed the first correlation between upstream carbon layer thickness distribution and downstream slurry viscosity, and the second correlation between upstream coating uniformity and downstream particle dispersion uniformity. The first correlation was calculated by extracting the average carbon layer thickness from historical batches and its corresponding initial viscosity value from the downstream homogenization process. The average carbon layer thickness refers to the statistical average carbon layer thickness measured by a laser scanning microscope for each batch of LFP material after carbon coating, reflecting the average thickness level of the carbon coating layer on the particle surface of that batch. The initial viscosity value from the downstream homogenization process refers to the baseline viscosity value of the slurry measured by an online viscometer at the start of the homogenization process or in the early stages of operation for the corresponding historical batch, reflecting the original rheological state of the slurry before being affected by process parameter adjustments. After obtaining the above two sets of data sequences, we used the average carbon layer thickness as the independent variable sequence and the initial viscosity value from the downstream homogenization process as the dependent variable sequence, and used the Pearson correlation coefficient method to calculate the correlation value between the two, which is the first correlation. The Pearson correlation coefficient is obtained by calculating the product of the covariance of two sets of data and their respective standard deviations. The value ranges from -1 to 1. The first correlation reflects the degree and direction of the influence of the upstream carbon layer thickness distribution on the downstream slurry viscosity. The larger the value, the stronger the transmission effect of carbon layer thickness fluctuation on slurry viscosity, that is, the thicker or thinner the carbon layer, the more significant the viscosity change.

[0045] The second correlation coefficient is calculated by extracting the coating uniformity index and its corresponding downstream homogenization process D50 particle size value from historical batches. The coating uniformity index refers to the percentage of carbon layer coverage obtained by analyzing the scanned image of LFP material after carbon coating in the matched historical batch using image processing software, reflecting the completeness of carbon layer coverage on the particle surface. The downstream homogenization process D50 particle size value reflects the original dispersion state of particles in the batch of slurry before being subjected to sufficient shearing. After obtaining the above two sets of data sequences, the coating uniformity index is used as the independent variable sequence, and the downstream homogenization process D50 particle size value is used as the dependent variable sequence. The correlation value between the two is obtained by using the Pearson correlation coefficient method. This value is the second correlation coefficient, which reflects the degree and direction of the influence of upstream coating uniformity changes on downstream particle dispersion. The larger the value, the stronger the transmission effect of coating uniformity fluctuations on particle dispersion effect, that is, the more significant the increase in D50 particle size caused by more uneven coating.

[0046] Based on the first and second correlation coefficients, and combined with the initial slurry state of the current batch, the degree of disturbance impact of upstream carbon coating process fluctuations on the current homogenization process is calculated. The initial slurry state of the current batch refers to the basic characteristic parameters of the slurry measured before the start of the homogenization process or in the early stage of operation, specifically including the initial viscosity reference value and the initial dispersibility reference value. The initial viscosity reference value refers to the viscosity measurement value of the slurry at the start of homogenization, and the initial dispersibility reference value refers to the dispersion state of particles in the slurry at the start of homogenization, usually expressed as the D50 particle size value, reflecting the original degree of agglomeration of particles before being subjected to sufficient shearing. The first and second correlation coefficients are used to calculate the quantitative transmission relationship between upstream and downstream process parameters based on the historical batch data that best matches the current batch, reflecting the transmission of upstream carbon coating process fluctuations to the downstream homogenization process. Following the general principle of process flow, and combining the initial slurry state of the current batch, i.e., the initial viscosity benchmark value and the initial dispersion benchmark value, the viscosity disturbance component is obtained by multiplying the first deviation between the measured value of the carbon layer thickness distribution of the upstream carbon coating process and the historical benchmark value by the first correlation degree. The dispersion disturbance component is obtained by multiplying the second deviation between the measured value of the coating uniformity and the historical benchmark value by the second correlation degree. Then, the viscosity disturbance component and the dispersion disturbance component are respectively superimposed on the initial viscosity benchmark value and the initial dispersion benchmark value to obtain the slurry state assessment value after being affected by upstream fluctuations. This quantifies the degree of disturbance impact of upstream carbon coating process fluctuations on the current homogenization process. This degree of disturbance impact is a quantitative indicator used to characterize the extent to which the expected fluctuations of the upstream carbon coating process will cause the downstream slurry state to deviate from the target value. The detailed calculation process will be explained later.

[0047] The calculated disturbance impact level is mapped to a preset risk level threshold to determine the instability risk level of the homogenization process. The preset risk level thresholds include low risk, medium risk, and high risk. For example, a disturbance impact level below 10% is mapped to low risk, 10% to 30% to medium risk, and above 30% to high risk. The determined risk level will serve as the basis for subsequent steps to formulate process compensation plans. The higher the risk level, the greater the required compensation adjustment, thereby enabling differentiated response strategies for anomalies of different severity.

[0048] By quantifying abnormal patterns into characteristic indicators and comparing them with preset ranges, and matching historical batches and analyzing upstream and downstream correlations when they exceed the limits, a precise quantitative assessment of the impact of upstream fluctuations on downstream processes is achieved, thereby generating risk levels. This provides a scientific basis for subsequent differentiated process compensation, avoids blind adjustments, and significantly improves the pertinence and effectiveness of anomaly handling.

[0049] The calculation of the impact of upstream carbon coating process fluctuations on the current homogenization process includes: obtaining the first deviation between the measured value of carbon layer thickness distribution of the current batch's upstream carbon coating process and the historical benchmark value, and the second deviation between the measured value of coating uniformity and the historical benchmark value; obtaining the viscosity disturbance component based on the first deviation and the first correlation, and obtaining the dispersion disturbance component based on the second deviation and the second correlation; superimposing the viscosity disturbance component and the dispersion disturbance component onto the viscosity benchmark value and dispersion benchmark value of the initial slurry state of the current batch, respectively, to obtain the slurry state evaluation value after being affected by upstream fluctuations; and calculating the comprehensive deviation between the slurry state evaluation value and the target slurry state as the impact of upstream carbon coating process fluctuations on the current homogenization process.

[0050] Specifically, after obtaining the first and second correlation degrees, it is necessary to convert the fluctuations in the upstream carbon coating process into specific impact values ​​on the downstream slurry state based on these two correlation degrees. This quantifies the actual disturbance degree of upstream fluctuations on the current homogenization process, providing accurate numerical basis for subsequent risk level determination. Specifically, when calculating the disturbance impact of upstream carbon coating process fluctuations on the current homogenization process, the first deviation between the measured carbon layer thickness distribution of the current batch's upstream carbon coating process and the historical benchmark value, and the second deviation between the measured coating uniformity value and the historical benchmark value, are obtained. The measured carbon layer thickness distribution value refers to the statistical value of the carbon layer thickness actually measured by a laser scanning microscope after carbon coating of the current batch of LFP material, typically including the average thickness and the standard deviation of the thickness distribution. The measured coating uniformity value refers to the percentage of carbon layer coverage obtained by analyzing the scanned image using image processing software. The historical baseline value is the statistical average of carbon layer thickness and coating uniformity in normal production batches throughout history. It represents the standard level under stable process conditions. The first deviation is the measured value of carbon layer thickness distribution minus the historical baseline value, and the second deviation is the measured value of coating uniformity minus the historical baseline value. The deviation value can be positive or negative. A positive value indicates that the measured value is higher than the historical baseline value, and a negative value indicates that the measured value is lower than the historical baseline value.

[0051] When calculating the impact of upstream carbon coating process fluctuations on the current homogenization process based on the first and second correlations, it is necessary to multiply the upstream fluctuation amplitude by the upstream-downstream transmission coefficient to quantify the expected change in downstream slurry state caused by upstream fluctuations. The rationale for this calculation logic is that the first and second correlations are transmission coefficients obtained based on historical batch data, reflecting how many units the downstream slurry viscosity or dispersion will change on average when the upstream carbon layer thickness or coating uniformity changes by one unit. The first and second deviations are the actual amplitudes of upstream fluctuations in the current batch. Multiplying the deviation value by the correlation can yield the specific impact that upstream fluctuations are expected to transmit to the downstream, namely the viscosity disturbance component and the dispersion disturbance component. Specifically, the viscosity disturbance component is calculated by multiplying the first deviation between the measured value of the carbon layer thickness distribution of the current batch and the historical benchmark value by the first correlation degree. This product quantifies the degree of influence of upstream carbon layer thickness fluctuations on downstream slurry viscosity. The dispersion disturbance component is calculated by multiplying the second deviation between the measured value of the coating uniformity of the current batch and the historical benchmark value by the second correlation degree. This product quantifies the degree of influence of upstream coating uniformity fluctuations on downstream particle dispersion.

[0052] After obtaining the viscosity and dispersion perturbation components, these components need to be combined with the initial slurry state of the current batch to assess the actual changes that fluctuations in the upstream carbon coating process may cause to the downstream slurry state. To address the dimensionless issues that may arise from inconsistent units between different physical quantities, this application first normalizes the deviation values ​​of the upstream carbon coating process before combining the perturbation components with the initial slurry state, converting them into dimensionless deviation indices. Specifically, the dimensionless carbon layer thickness deviation index is obtained by dividing the first deviation between the measured value of the current batch's carbon layer thickness distribution and the historical baseline value by its historical fluctuation range; the dimensionless coating uniformity deviation index is obtained by dividing the second deviation between the measured value of the current batch's coating uniformity and the historical baseline value by its historical fluctuation range. Meanwhile, the first and second correlation coefficients in this application are dimensionless correlation coefficients obtained based on statistical analysis of historical batch data, representing the correlation strength between upstream carbon layer thickness and downstream slurry viscosity, and between upstream coating uniformity and downstream particle dispersion, respectively.

[0053] Based on this, the dimensionless carbon layer thickness deviation index is multiplied by the first correlation coefficient to obtain the dimensionless viscosity perturbation coefficient; the dimensionless coating uniformity deviation index is multiplied by the second correlation coefficient to obtain the dimensionless dispersion perturbation coefficient. Subsequently, these dimensionless perturbation coefficients are multiplied by the initial viscosity reference value and the initial dispersion reference value of the current batch, respectively, to obtain slurry viscosity assessment values ​​and slurry dispersion assessment values ​​with correct units. Specifically, the initial viscosity reference value is multiplied by the viscosity perturbation coefficient to obtain the expected viscosity change caused by upstream carbon layer thickness fluctuations. This change is then superimposed on the initial viscosity reference value to obtain the slurry viscosity assessment value affected by upstream fluctuations. Similarly, the initial dispersion reference value is multiplied by the dispersion perturbation coefficient to obtain the expected dispersion change caused by upstream coating uniformity fluctuations. This change is then superimposed on the initial dispersion reference value to obtain the slurry dispersion assessment value affected by upstream fluctuations. These two assessment values ​​together constitute the slurry state assessment value after being affected by upstream fluctuations, reflecting the slurry state that the homogenization process is expected to achieve under the current fluctuations in the upstream carbon coating process without any process compensation.

[0054] After obtaining the slurry state assessment value, it is necessary to further calculate its comprehensive deviation from the target slurry state to quantify the impact of upstream carbon coating process fluctuations on the current homogenization process. The target slurry state is the ideal slurry parameters preset according to the requirements of subsequent coating processes, including the target viscosity value and the target dispersion value. First, calculate the percentage deviation between the slurry viscosity assessment value and the target viscosity value, and the percentage deviation between the slurry dispersion assessment value and the target dispersion value. The formula for calculating the percentage deviation is the absolute value of the assessment value minus the target value, divided by the target value, and then multiplied by 100%, reflecting the relative degree of deviation of each indicator from the target value. Since the viscosity deviation percentage and the dispersion deviation percentage have the same dimensions, they can be directly weighted and summed. The weighting coefficient is determined according to the process control requirements. For example, viscosity stability has a more direct impact on subsequent coating processes, so a higher weight can be assigned to viscosity deviation, such as 0.6, and a lower weight can be assigned to dispersion deviation, such as 0.4. Multiply the viscosity deviation percentage by 0.6 and the dispersion deviation percentage by 0.4, and add them together to obtain the comprehensive deviation, which is expressed as a percentage. The greater the overall deviation, the more severe the expected impact of upstream carbon coating process fluctuations on the current homogenization process; the smaller the overall deviation, the less severe the impact of upstream fluctuations. This overall deviation is the degree of disturbance of upstream carbon coating process fluctuations on the current homogenization process, and will be used as a quantitative basis for subsequent risk level determination.

[0055] Through the above calculations, the fluctuations in the upstream carbon coating process are transformed into specific impact values ​​on the downstream slurry state, realizing a quantitative assessment of the impact of upstream fluctuations. This provides a precise numerical basis for mapping the degree of disturbance impact to a risk level, ensuring the objectivity and accuracy of risk assessment.

[0056] Step S3: Based on the risk level of instability in the homogenization process, formulate a homogenization process compensation plan for the current batch. Run the homogenization process simulation through the virtual simulation module to verify the effectiveness of the process compensation plan in dealing with fluctuations in the upstream carbon coating process and obtain simulation verification results.

[0057] Step S3 involves developing a homogenization process compensation scheme for the current batch, including: extracting corresponding adjustment rules from a pre-defined adjustment mechanism rule library based on the homogenization process instability risk level; calculating shear rate adjustment values, stirring time adjustment values, and solvent addition ratio adjustment values ​​based on the extracted adjustment rules and the initial process parameters of the current batch, generating a compensation scheme that includes the adjusted process parameters; retrieving historical excellent adjustment records that match the current risk level from the historical database, comparing the similarity between the compensation scheme and the historical excellent adjustment records, verifying the feasibility of the compensation scheme, and outputting a homogenization process compensation scheme that has passed feasibility verification.

[0058] Specifically, after determining the instability risk level of the current batch's homogenization process, a targeted process compensation plan needs to be developed based on this risk level to address the impact of fluctuations in the upstream carbon coating process on homogenization quality. First, based on the determined instability risk level of the homogenization process, corresponding adjustment rules are extracted from a pre-built adjustment mechanism rule library. This rule library is a pre-constructed knowledge base that stores process parameter adjustment rules for various abnormal modes under different risk levels. Each rule includes the applicable abnormal mode type, risk level range, and corresponding parameter adjustment direction and magnitude template. For example, for the medium-risk level of shear failure and poor dispersion mode, the rule library stores the adjustment rule of increasing the shear rate by 10% to 15% and extending the stirring time by 5% to 8%. For the high-risk level of incoming material sensitive fluctuation abnormal mode, the adjustment rule is to reduce the shear rate by 5% to 10% to weaken the system's sensitivity to disturbances, while fine-tuning the solvent addition ratio to stabilize viscosity. When extracting adjustment rules, a matching query is performed in the rule library based on the current batch's risk level and the output abnormal mode type to obtain the adjustment rule corresponding to that abnormal mode and risk level.

[0059] Based on the extracted adjustment rules and the initial process parameters of the current batch, specific parameter adjustment values ​​are calculated. These initial parameters include the current shear rate, stirring time, and solvent addition ratio. According to the adjustment direction and range specified in the adjustment rules, and combined with the initial parameter values ​​of the current batch, the shear rate adjustment value, stirring time adjustment value, and solvent addition ratio adjustment value are calculated. For example, if the current shear rate is 1000 rpm and the adjustment rules require an increase of 10% to 15%, then the midpoint of 12.5% ​​can be taken as the adjustment range, resulting in a calculated shear rate adjustment value of 1125 rpm. If the current stirring time is 30 minutes and the adjustment rules require an extension of 5% to 8%, then the calculated stirring time adjustment value is 31.5 minutes to 32.4 minutes. The specific values ​​can be determined based on historical experience or process requirements. The calculated adjustment values ​​are combined to generate a compensation scheme containing the adjusted process parameters. This scheme clearly specifies the target values ​​that the shear rate, stirring time, and solvent addition ratio should be adjusted to in subsequent runs of the current batch.

[0060] After the compensation plan is generated, its feasibility needs to be verified to ensure that it can be effectively implemented in actual production without introducing new problems. To this end, historical best-performing adjustment records matching the current risk level are retrieved from the historical database. These records refer to successful process adjustment cases implemented under the same or similar risk levels in past production batches, achieving good results. Each case includes the anomaly mode at the time, the risk level, the initial parameters before adjustment, the target parameters after adjustment, and the slurry quality results after adjustment. A similarity comparison algorithm is used to compare the currently generated compensation plan with the historical best-performing adjustment records, calculating the similarity between each adjustment value in the plan and the corresponding adjustment value in the record. Similarity calculation can use Euclidean distance or cosine similarity methods, comprehensively considering the matching degree across three dimensions: shear rate, stirring time, and solvent addition ratio.

[0061] If the similarity between the compensation scheme and a historical excellent adjustment record exceeds a preset threshold, such as 85%, it indicates that the scheme has reliable historical practice support and its feasibility verification is passed. If the similarity is below the threshold, the compensation scheme needs to be fine-tuned, or alternative rules need to be extracted from the rule base to regenerate the scheme, until a scheme that has passed feasibility verification is obtained. The output of the slurry process compensation scheme that has passed feasibility verification will serve as the input for subsequent virtual simulation verification, ensuring that the scheme entering the simulation stage has preliminary feasibility and reliability.

[0062] Through the above steps, the transformation from risk level to specific compensation plan was realized, and the practical basis of the plan was guaranteed by historical data verification, providing a reliable guarantee for subsequent simulation verification and practical application.

[0063] Step S3, obtaining simulation verification results, includes: inputting the shear rate adjustment value, stirring time adjustment value, and solvent addition ratio adjustment value from the process compensation scheme into the virtual simulation module, which has a built-in simulation model of the homogenization process based on finite element analysis; loading slurry viscosity control records and particle agglomeration suppression data that match the fluctuation characteristics of the upstream carbon coating process of the current batch from the historical database as the initial boundary conditions of the simulation model; running simulation iterations to obtain the predicted slurry viscosity change curve, particle size distribution evolution data, and batch stability index after the implementation of the compensation scheme; performing a matching degree analysis between the predicted slurry viscosity change curve and the preset target viscosity window, and performing a conformity judgment between the particle size distribution evolution data and the preset dispersion uniformity threshold; and generating simulation verification results including applicability scores and effect prediction indicators based on the results of the matching degree analysis and conformity judgment.

[0064] Specifically, after developing a process compensation scheme that has passed feasibility verification, it is necessary to further verify the effectiveness of the scheme through a virtual simulation module to ensure that the compensation scheme can achieve the expected results in practical applications. Specifically, the adjustment values ​​of various parameters included in the generated process compensation scheme are first input into the virtual simulation module. These parameters include shear rate adjustment values, stirring time adjustment values, and solvent addition ratio adjustment values. The virtual simulation module has a built-in simulation model of the homogenization process based on finite element analysis. This model is constructed by mathematically modeling the hydrodynamic behavior, particle motion trajectory, and shear force distribution in the homogenization process. It can simulate the change law of slurry viscosity over time, the opening process of particle agglomerates, and the evolution trend of dispersion uniformity under different process parameter conditions. In order to make the simulation results more consistent with the actual situation of the current batch, it is necessary to load the slurry viscosity control records and particle agglomeration suppression data that match the fluctuation characteristics of the upstream carbon coating process of the current batch into the historical database as the initial boundary conditions of the simulation model. The fluctuation characteristics of the upstream carbon coating process refer to the specific deviations in the carbon layer thickness distribution and coating uniformity of the current batch. Batches with similar fluctuation characteristics are retrieved from the historical database, and the initial viscosity records and particle agglomerate volume ratio monitoring data of these batches before the start of the homogenization process are extracted. These data are used as input to the simulation model to ensure that the simulation process can truly reflect the initial state of the slurry under the influence of the current upstream fluctuations.

[0065] After completing the parameter input and boundary condition settings, the simulation iterative calculation is run. Based on the input process parameters and initial conditions, the simulation model simulates the changes in the slurry state throughout the homogenization process by solving the fluid dynamics equations and particle motion equations. Specifically, the simulation model divides the flow field within the homogenizing equipment into hundreds of thousands of tiny computational grid cells. On each grid cell, the Navier-Stokes equations are solved to calculate the spatiotemporal evolution of the shear rate distribution, velocity field, and pressure field. Simultaneously, the model uses the Euler-Lagrange method to track the motion trajectories of tens of thousands of representative LFP particles, calculates the shear force and collision probability experienced by the particles in the flow field, and simulates the breakup process of agglomerates under shear action and the tendency of dispersed particles to re-agglomerate. The model also couples the slurry's rheological constitutive equations, updating the slurry viscosity value in real time based on the current shear rate and particle concentration. The simulation process iterates in seconds-level time steps. Within each time step, the flow field distribution is first calculated based on the current process parameters, and then the particle state and slurry viscosity are updated based on the flow field results. This process is repeated until the set total stirring time is reached. After the simulation iteration, the predicted slurry viscosity change curve, particle size distribution evolution data, and batch stability index are extracted from the simulation results after the implementation of the compensation scheme. The slurry viscosity change curve refers to the continuous record of the viscosity value changing with time from the start to the end of homogenization. Each time point corresponds to a specific viscosity value, reflecting the dynamic response of the slurry rheological properties under the action of process parameters. The particle size distribution evolution data refers to the statistical values ​​of particle size at different time points, including the change trajectory of D10, D50, and D90 particle sizes over time. These data intuitively show the process of particle agglomerates gradually opening under shear and the final dispersed state. The batch stability index includes parameters such as the time required for viscosity to reach steady state, the viscosity fluctuation amplitude after stabilization, and the concentration of the final particle size distribution, which are used to evaluate the intra-batch consistency of slurry quality under the compensation scheme.

[0066] After obtaining the simulation output data, the effectiveness of the compensation scheme is quantitatively evaluated. First, the predicted slurry viscosity change curve is compared with the preset target viscosity window. The target viscosity window is an ideal viscosity range set according to the requirements of the subsequent coating process, including an upper and lower limit. During the matching analysis, the proportion of time the predicted viscosity curve falls within the target viscosity window throughout the entire stirring cycle is calculated, along with the average deviation of the viscosity curve from the center value of the window. The proportion of time and the degree of deviation are weighted and combined to obtain a viscosity matching score. A higher score indicates a better viscosity control effect. Simultaneously, the conformity of the particle size distribution evolution data with the preset dispersion uniformity threshold is judged. The dispersion uniformity threshold includes the upper limit of the D50 particle size and the upper limit of the D90 particle size. It is determined whether the D50 particle size and the D90 particle size are lower than the upper limit at the end of the simulation, and whether the particle size shows a continuous decreasing trend and remains stable in the later stage throughout the stirring process. Based on the judgment results, a dispersion conformity index is generated. If all conditions are met, the conformity judgment is passed.

[0067] Based on the results of the matching degree analysis and compliance judgment, simulation verification results are generated, including an applicability score and effect prediction indicators. The applicability score is a comprehensive evaluation value of the overall feasibility of the compensation scheme. It is obtained by weighted summation of the viscosity matching degree score and the dispersion compliance indicator, resulting in a score from 0 to 100. For example, the viscosity matching degree score accounts for 60% of the weight, and the dispersion compliance indicator accounts for 40%. The effect prediction indicators include specific values ​​such as the predicted final viscosity value, the final D50 particle size value, the final D90 particle size value, and the batch stability level. If the applicability score is higher than a preset threshold, such as 80 points, and all effect prediction indicators meet the process requirements, it indicates that the compensation scheme is effective and feasible and can be put into practical application. If the score is lower than the threshold, it is necessary to return to step S3 to revise the compensation scheme or optimize and adjust the existing scheme. The above simulation verification process ensures that the process compensation scheme is fully verified before practical application.

[0068] The compensation scheme is pre-validated by using a virtual simulation module. By combining the matching degree analysis of the slurry viscosity change curve with the target window and the conformity judgment of particle size distribution with dispersion threshold, an applicability score and effect prediction index are generated. This ensures that the compensation scheme is fully validated before actual application, avoids the risks and costs of trial and error on the production line, and significantly improves the reliability and success rate of process adjustment.

[0069] The process involves performing a matching degree analysis between the predicted slurry viscosity change curve and a preset target viscosity window. This includes: extracting the viscosity characteristic values ​​of the predicted slurry viscosity change curve throughout the entire stirring cycle; comparing the viscosity characteristic values ​​with the preset target viscosity window and generating corresponding characteristic value matching degrees based on the comparison results; and weighting and summing the characteristic value matching degrees of multiple viscosity characteristic values ​​to obtain a matching degree analysis result that characterizes the overall degree of agreement between the predicted slurry viscosity and the target viscosity window.

[0070] Specifically, after obtaining the predicted viscosity change curve of the slurry, a matching degree analysis needs to be performed on the curve and the preset target viscosity window to quantitatively evaluate whether the slurry viscosity under the compensation scheme meets the process requirements. The matching degree analysis first extracts the viscosity characteristic values ​​throughout the entire stirring cycle from the predicted viscosity change curve. These characteristic values ​​are mathematical descriptions of the key shape of the curve and are used to characterize the overall characteristics of viscosity change over time from different dimensions. The extracted viscosity characteristic values ​​specifically include the mean viscosity, the viscosity fluctuation amplitude, and the time required for the viscosity to reach a steady state. The mean viscosity refers to the arithmetic mean of the viscosity values ​​at all time points during the entire stirring cycle, reflecting the average viscosity of the slurry as a whole. The viscosity fluctuation amplitude refers to the range of fluctuation of the viscosity curve around the center value after reaching a steady state, usually expressed as the standard deviation or maximum deviation, reflecting the stability of the slurry state. The time required for the viscosity to reach a steady state refers to the time elapsed from the start of stirring until the viscosity value first enters and remains within the target viscosity window, reflecting the response speed of the process parameters.

[0071] Each extracted viscosity feature value is compared with a preset target viscosity window, which includes a center value, an upper limit value, and a lower limit value. The center value represents the ideal target viscosity value, while the upper and lower limits constitute the allowable fluctuation range. During the comparison, the mean viscosity is compared with the center value of the target viscosity window, and the absolute or relative deviation between the two is calculated. The smaller the deviation, the closer the overall viscosity is to the ideal state. The viscosity fluctuation amplitude is compared with the allowable fluctuation range of the target viscosity window; if the fluctuation amplitude is less than half the window width, it indicates good stability. The time required for the viscosity to reach a steady state is compared with a preset steady-state time threshold; the shorter the time, the faster the process response. Based on the comparison results, corresponding feature value matching degrees are generated. For the mean viscosity, the matching degree can be calculated based on the degree of deviation. First, the absolute deviation between the mean viscosity and the center value of the target viscosity window is calculated, and then this absolute deviation is compared with the allowable deviation range of the target viscosity window. The matching degree is inversely proportional to the degree of deviation. Specifically, when the mean viscosity is exactly equal to the center value, the matching degree is 100%; when the mean viscosity deviates from the center value but is still within the upper and lower limits of the window, the matching degree is calculated linearly. For example, for every 10% increase in the window width due to the deviation, the matching degree decreases by 10 percentage points accordingly; when the mean viscosity exceeds the upper and lower limits of the window, the matching degree can be set to 0% or further reduced according to the degree of deviation. This calculation method ensures that the matching degree can objectively reflect the degree of closeness between the mean viscosity and the ideal target.

[0072] After obtaining the matching degree of each viscosity characteristic value, these matching degrees need to be integrated to obtain a matching degree analysis result that can comprehensively reflect the overall degree of agreement between the predicted slurry viscosity and the target viscosity window. Since different viscosity characteristic values ​​describe the characteristics of slurry viscosity from different dimensions, their contribution to the overall matching degree also varies. Therefore, it is necessary to assign corresponding weights to each characteristic value according to its importance in process control. Specifically, the weight coefficients of each viscosity characteristic value are first determined. The mean viscosity reflects the overall viscosity level of the slurry and is the core factor determining the coating quality, so it can be given the highest weight. The viscosity fluctuation range reflects the stability of the slurry state and has an important impact on coating consistency, so it can be given the second highest weight. The time required for the viscosity to reach steady state reflects the response speed of the process parameters and has a certain impact on production efficiency, so it can be given a relatively low weight.

[0073] Then, the matching degree of each viscosity characteristic value is multiplied by its corresponding weighting coefficient to obtain the weighted matching degree of each characteristic value. For example, the matching degree of the mean viscosity is multiplied by the weighting coefficient of the mean viscosity to obtain the weighted matching degree of the mean viscosity; the matching degree of the viscosity fluctuation amplitude is multiplied by the weighting coefficient of the viscosity fluctuation amplitude to obtain the weighted matching degree of the viscosity fluctuation amplitude; and the matching degree of the time required for viscosity to reach steady state is multiplied by the weighting coefficient of the time required for viscosity to reach steady state to obtain the weighted matching degree of the time required for viscosity to reach steady state. Finally, the weighted matching degrees of each characteristic value are summed to obtain the matching degree analysis result, which characterizes the overall degree of agreement between the predicted slurry viscosity and the target viscosity window. This result is a comprehensive score; the higher the score, the better the overall agreement between the predicted slurry viscosity change curve and the target viscosity window, and the more ideal the effect of the compensation scheme in viscosity control. This matching degree analysis result will serve as an important part of the simulation verification results and will be used to subsequently determine the feasibility and effectiveness of the compensation scheme.

[0074] By extracting multi-dimensional feature values ​​such as mean viscosity, fluctuation amplitude, and steady-state time, and comparing them with the target window to generate a matching degree, and then combining them with weighted coefficients for weighted summation, a comprehensive quantitative evaluation of the degree of fit between the slurry viscosity and the target window is achieved. This matching degree analysis result can objectively reflect the comprehensive effect of the compensation scheme in viscosity control, providing an accurate and reliable quantitative basis for subsequent assessment of the scheme's feasibility.

[0075] Step S4: Based on the verified process compensation scheme, generate the final homogenization process parameter adjustment instruction, and send the adjustment instruction to the homogenization production control system to update the process operation parameters of the current batch and store them in the historical database.

[0076] Specifically, after obtaining the process compensation scheme that has passed simulation verification, the scheme needs to be converted into executable process parameter adjustment instructions and sent to the production control system to achieve actual adjustments to the homogenization process of the current batch. Specifically, firstly, the verified process compensation scheme is obtained from the output simulation verification results. This scheme clearly specifies the various process parameters that need to be adjusted in the subsequent operation of the current batch, including shear rate adjustment values, stirring time adjustment values, and solvent addition ratio adjustment values. These adjustment values ​​are compared with the currently running process parameters in the production control system to determine the specific adjustment amount for each parameter. For example, if the current shear rate is 1000 rpm and the compensation scheme requires adjustment to 1125 rpm, the adjustment amount is an increase of 125 rpm; if the current stirring time is 30 minutes and the compensation scheme requires adjustment to 32 minutes, the adjustment amount is an increase of 2 minutes; if the current solvent addition ratio is 5.0% and the compensation scheme requires adjustment to 5.3%, the adjustment amount is an increase of 0.3 percentage points. These adjustment amounts are then encapsulated according to an instruction format recognizable by the production control system to generate adjustment instructions containing the updated process parameters.

[0077] The generated adjustment instructions are sent to the homogenization production control system. The production control system automatically updates the process operating parameters for the current batch based on the instructions. During the issuance process, the accuracy and real-time nature of the instruction transmission must be ensured. Instructions can be sent to the execution unit of the control system via industrial Ethernet or fieldbus communication methods. Upon receiving the instruction, the execution unit immediately adjusts parameters such as shear rate, stirring time, and solvent addition ratio to the target values ​​specified in the instruction and maintains these parameters until the end of the batch or the next adjustment. Simultaneously, the production control system monitors the actual operating status after parameter adjustments in real time to ensure accurate execution of the adjustment instructions. After the adjustment instructions are executed, the entire process data for this process optimization is stored in a historical database. The stored data includes the abnormal slurry state mode type for the current batch, quantitative characteristic indicators, upstream carbon coating process correlation data, calculated risk level, generated compensation scheme, simulation verification results, and the finally issued adjustment instructions and updated process parameters. This data is indexed by batch number and timestamp to form a complete process optimization record. The stored data will serve as a reference for subsequent batch process optimization, used to optimize rules in the adjustment mechanism rule base, update historical excellent adjustment records, and improve the accuracy of the similarity matching algorithm, thereby achieving continuous learning and iterative upgrades of the process optimization system.

[0078] To further verify the actual effectiveness of the technical solution of this application, multiple batches of comparative experiments were conducted under the same production conditions using both the traditional homogenization process control method and the method of this application. Figure 2 The diagram shown is a comparison of the process effects of this application and existing technologies. Figure 2 As can be seen, after adopting the method of this application, the batch stability of the slurry increased from 72.0% in the prior art to 94.0%, indicating that the fluctuations in slurry viscosity and dispersion uniformity between batches were significantly reduced, and batch consistency was greatly improved; production efficiency increased from 65.0% to 88.0%, mainly due to the shortened anomaly response time and the rapid implementation of the process compensation scheme, which reduced the downtime of the production line; and material waste rate decreased from 8.5% to 2.3%. The above comparative data fully demonstrate the significant effects of the technical solution of this application in improving the stability of slurry quality, increasing production efficiency, and reducing material costs.

[0079] The above describes an optimization method for a high-density homogenization process of LFP cathode in an embodiment of this application. The following describes an optimization system for a high-density homogenization process of LFP cathode in an embodiment of this application. Please refer to [link / reference]. Figure 3 An LFP cathode high-density homogenization process optimization system according to an embodiment of this application includes:

[0080] The data acquisition module is used to acquire slurry state data and process parameter data in the high-density homogenization process of LFP cathode in real time, analyze the correlation strength index between slurry state changes and process parameter adjustments, and identify abnormal slurry state patterns in the current batch.

[0081] The judgment module is used to extract the quantitative feature indicators corresponding to the abnormal slurry state mode. If the quantitative feature indicators are not within the preset range, the historical upstream carbon coating process correlation data and downstream process response data are obtained. Combined with the initial slurry state of the current batch, the influence of upstream fluctuations on the current homogenization process is judged, and the homogenization process instability risk level is generated.

[0082] The verification module is used to formulate a compensation plan for the homogenization process for the current batch based on the risk level of instability in the homogenization process. The virtual simulation module runs the homogenization process simulation to verify the effectiveness of the process compensation plan in dealing with fluctuations in the upstream carbon coating process and obtain simulation verification results.

[0083] The execution module is used to generate the final homogenization process parameter adjustment instructions based on the verified process compensation scheme, and send the adjustment instructions to the homogenization production control system to update the process operation parameters of the current batch and store them in the historical database.

[0084] This application also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the method for optimizing the high-density homogenization process of LFP cathode.

[0085] In summary, this application provides a method, system, and storage medium for optimizing the high-density homogenization process of LFP cathode materials. This method collects slurry state and process parameter data in real time, analyzes the correlation strength between the two, and accurately identifies abnormal slurry state patterns in the current batch. It then extracts quantitative characteristic indicators and compares them with preset ranges. When these ranges are exceeded, it matches historical batches and analyzes the upstream and downstream correlation, quantifying the impact of upstream carbon coating process fluctuations on downstream slurry disturbances and generating a risk level. Based on the risk level, it extracts adjustment rules from a rule base, formulates compensation schemes, and verifies their effectiveness through a virtual simulation module. Finally, it generates adjustment instructions, sends them to the production control system, updates process parameters, and stores them in the database. This application constructs a complete technology chain from anomaly identification, source analysis, scheme formulation, simulation verification to closed-loop control, realizing proactive perception and adaptive compensation of upstream fluctuations in the homogenization process. This significantly improves the batch stability and consistency of slurry quality, avoids the lag and blindness of traditional post-adjustment, reduces the trial-and-error costs of the production line, and provides an intelligent, data-driven process optimization solution for lithium battery cathode material preparation.

[0086] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0087] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0088] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. An optimization method for high-density homogenization process of LFP cathode, characterized in that, The method includes: Step S1: Real-time acquisition of slurry state data and process parameter data in the LFP cathode high-density homogenization process, analysis of the correlation strength index between slurry state changes and process parameter adjustments, and identification of abnormal slurry state patterns in the current batch. Step S2: Extract the quantitative feature index corresponding to the abnormal slurry state pattern. If the quantitative feature index is not within a preset range, obtain historical upstream carbon coating process correlation data and downstream process response data, and combine them with the initial slurry state of the current batch to determine the degree of influence of upstream fluctuations on the current homogenization process, and generate a homogenization process instability risk level. Step S2 includes: comparing the quantitative feature index with a preset range; if it is not within the preset range, retrieving historical batches matching the current quantitative feature index from the historical database, and extracting the upstream carbon coating process correlation data and downstream process response data of the historical batches; analyzing the first correlation degree between the upstream carbon layer thickness distribution and the downstream slurry viscosity, and the second correlation degree between the upstream coating uniformity and the downstream particle dispersion uniformity; based on the first and second correlation degrees, and combined with the initial slurry state of the current batch, calculating the disturbance impact degree of upstream carbon coating process fluctuations on the current homogenization process, mapping the disturbance impact degree to a preset risk level threshold, and determining the homogenization process instability risk level. The calculation of the impact of upstream carbon coating process fluctuations on the current homogenization process includes: obtaining the first deviation between the measured value of carbon layer thickness distribution of the upstream carbon coating process in the current batch and the historical benchmark value, and the second deviation between the measured value of coating uniformity and the historical benchmark value; obtaining the viscosity disturbance component based on the first deviation and the first correlation, and obtaining the dispersion disturbance component based on the second deviation and the second correlation; superimposing the viscosity disturbance component and the dispersion disturbance component onto the viscosity benchmark value and dispersion benchmark value of the initial slurry state in the current batch, respectively, to obtain the slurry state evaluation value after being affected by upstream fluctuations; calculating the comprehensive deviation between the slurry state evaluation value and the target slurry state as the impact of upstream carbon coating process fluctuations on the current homogenization process; Step S3: Based on the risk level of instability in the homogenization process, formulate a homogenization process compensation scheme for the current batch, run the homogenization process simulation through the virtual simulation module, verify the effectiveness of the process compensation scheme in dealing with fluctuations in the upstream carbon coating process, and obtain simulation verification results; Step S4: Based on the verified process compensation scheme, generate the final homogenization process parameter adjustment instruction, and send the adjustment instruction to the homogenization production control system to update the process operation parameters of the current batch and store them in the historical database.

2. The method for optimizing the high-density homogenization process of LFP cathode according to claim 1, characterized in that, The correlation strength indicators analyzed in step S1 between changes in slurry state and adjustments to process parameters include: Extract the slurry viscosity and particle size distribution values ​​from the slurry state data, and extract the corresponding shear rate and stirring time values ​​from the process parameter data; Calculate the first correlation coefficient between the slurry viscosity value and the shear rate value, and the second correlation coefficient between the particle size distribution value and the stirring time value; Based on the first and second correlation coefficients, a correlation strength index is generated to characterize the correlation between slurry state and process parameters.

3. The method for optimizing the high-density homogenization process of LFP cathode according to claim 1, characterized in that, Step S1 identifies the abnormal slurry state pattern of the current batch, including: The correlation strength index is compared with a preset benchmark correlation range. If it is lower than the lower limit, it is determined to be the first abnormal mode; if it is higher than the upper limit, it is determined to be the second abnormal mode. For the first abnormal mode, based on the particle agglomeration inhibition, identify whether there is poor dispersion due to insufficient shear rate or uneven mixing due to insufficient stirring time. For the second abnormal mode, in combination with the particle agglomeration inhibition, identify whether there is a sensitive abnormality caused by fluctuations in the characteristics of the incoming material or a sudden change in viscosity caused by improper addition of solvent; Based on the specific pattern identified, determine the abnormal slurry state pattern of the current batch and output the corresponding pattern type.

4. The method for optimizing the high-density homogenization process of LFP cathode according to claim 1, characterized in that, Step S3 involves developing a homogenization process compensation plan for the current batch, including: Based on the instability risk level of the homogenization process, the corresponding adjustment rules are extracted from the preset adjustment mechanism rule base; Based on the extracted adjustment rules and combined with the initial process parameters of the current batch, the adjustment values ​​for shear rate, stirring time, and solvent addition ratio are calculated to generate a compensation scheme that includes the adjusted process parameters. Retrieve historical excellent adjustment records that match the current risk level from the historical database, compare the compensation scheme with the historical excellent adjustment records to verify the feasibility of the compensation scheme, and output the homogenization process compensation scheme that has passed the feasibility verification.

5. The method for optimizing the high-density homogenization process of LFP cathode according to claim 1, characterized in that, Step S3 obtains the simulation verification results, including: The shear rate adjustment value, stirring time adjustment value, and solvent addition ratio adjustment value in the process compensation scheme are input into the virtual simulation module, which has a built-in simulation model of the homogenization process based on finite element analysis. Load slurry viscosity control records and particle agglomeration suppression data from the historical database that match the fluctuation characteristics of the upstream carbon coating process of the current batch, as the initial boundary conditions for the simulation model; Run simulation iterations to obtain the predicted slurry viscosity change curve, particle size distribution evolution data, and batch stability index after the implementation of the compensation scheme; The predicted slurry viscosity change curve is matched with the preset target viscosity window, and the particle size distribution evolution data is judged to be consistent with the preset dispersion uniformity threshold. Based on the results of matching analysis and compliance judgment, simulation verification results including applicability scores and effect prediction indicators are generated.

6. The method for optimizing the high-density homogenization process of LFP cathode according to claim 5, characterized in that, The predicted slurry viscosity change curve is compared with the preset target viscosity window to perform a matching degree analysis, including: Extract the viscosity characteristic values ​​of the predicted slurry viscosity change curve over the entire stirring cycle; The viscosity feature value is compared with a preset target viscosity window, and the corresponding feature value matching degree is generated according to the comparison result. By weighted summing of the eigenvalue matching degrees of multiple viscosity characteristic values, the matching degree analysis results, which characterize the overall degree of agreement between the predicted slurry viscosity and the target viscosity window, are obtained.

7. A high-density homogenization process optimization system for LFP cathodes, used to implement the high-density homogenization process optimization method for LFP cathodes as described in any one of claims 1-6, characterized in that, The system includes: The data acquisition module is used to acquire slurry state data and process parameter data in the high-density homogenization process of LFP cathode in real time, analyze the correlation strength index between slurry state changes and process parameter adjustments, and identify abnormal slurry state patterns in the current batch. The judgment module is used to extract the quantitative feature index corresponding to the abnormal slurry state pattern. If the quantitative feature index is not within the preset range, the historical upstream carbon coating process correlation data and downstream process response data are obtained, and combined with the initial slurry state of the current batch, the influence of upstream fluctuations on the current homogenization process is judged, and the homogenization process instability risk level is generated. The verification module is used to formulate a compensation scheme for the homogenization process for the current batch based on the risk level of instability of the homogenization process, and to run a simulation of the homogenization process through the virtual simulation module to verify the effectiveness of the compensation scheme in dealing with fluctuations in the upstream carbon coating process and obtain simulation verification results. The execution module is used to generate the final homogenization process parameter adjustment instruction based on the verified process compensation scheme, and send the adjustment instruction to the homogenization production control system to update the process operation parameters of the current batch and store them in the historical database.

8. A computer-readable storage medium storing instructions thereon, characterized in that, When the instruction is executed by the processor, it implements the LFP cathode high-density homogenization process optimization method as described in any one of claims 1-6.

Images