Method and system for controlling uniformity of sodium ion battery slurry

Abstract

The invention discloses a uniformity control method and system for sodium-ion battery slurry, and the method comprises the steps: carrying out the data preprocessing according to the collected shear stress, shear rate and viscosity parameters of the slurry, and generating a data set containing a training group and a verification group; continuously training the training group by using a linear regression model, generating a control prediction model and deploying the control prediction model; acquiring real-time shear stress, shear rate and viscosity parameters of the slurry, inputting the real-time shear stress, shear rate and viscosity parameters into the control prediction model, and outputting a slurry rheological property prediction result and an additive prediction result at the next moment; dynamically adjusting the stirring speed and the shearing speed of the slurry according to the slurry rheological property prediction result, and dynamically adjusting the feeding amount and the feeding time of the additive according to the additive prediction result; and carrying out loop iteration until the uniformity of the slurry reaches a preset requirement. According to the uniformity control method for the sodium-ion battery slurry, the uniformity of the active material and the additive in the slurry can be improved, and material particle aggregation is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of battery slurry mixing, and in particular to a method and system for controlling the uniformity of sodium ion battery slurry. Background Art

[0002] As a low-cost and resource-rich energy storage technology, sodium-ion batteries have received widespread attention in the energy storage field in recent years. The main components of sodium-ion batteries include positive electrode materials, negative electrode materials, electrolytes and separators, and the preparation of electrode slurry plays a vital role in battery performance. The uniformity of the slurry directly affects the structure, performance and consistency of the electrode. Therefore, the uniformity control of the slurry is a key link in the production process of sodium-ion batteries.

[0003] At present, sodium-ion battery slurry is usually composed of active materials, conductive agents, binders and solvents. These components need to form a uniformly dispersed mixed system during the stirring and mixing process to ensure the conductivity, mechanical properties and electrochemical stability of the electrode; the uniformity of the slurry not only affects the thickness uniformity and coating effect of the electrode sheet, but also directly affects the rate performance, cycle life and energy density of the battery. In the existing sodium-ion battery slurry preparation process, common stirring and dispersion methods include high-speed stirring, planetary stirring, ball milling and ultrasonic dispersion, etc. These methods use mechanical force or sound waves to uniformly mix active materials, conductive agents and binders in solvents; however, due to the large differences in physical and chemical properties between different components, such as particle size, surface energy and density, the slurry is often prone to agglomeration, sedimentation or stratification during the mixing process, resulting in poor slurry uniformity.

[0004] Therefore, there is an urgent need for a uniformity control method that can improve the uniformity of active materials and additives in the slurry and reduce the agglomeration of material particles. Summary of the invention

[0005] Purpose of the invention: In order to overcome the above shortcomings, the purpose of the present invention is to provide a method and system for controlling the uniformity of sodium ion battery slurry.

[0006] In order to solve the above technical problems, the present invention provides a method for controlling the uniformity of sodium ion battery slurry, comprising: Step S1: performing data preprocessing according to the shear stress, shear rate and viscosity parameters of the collected slurry to generate a data set including a training group and a validation group; Step S2: continuously training the training group using the linear regression model, generating a control prediction model and deploying the control prediction model, wherein the training process includes: optimizing the linear regression model parameters by minimizing the mean square error using the gradient descent method, verifying the accuracy of the linear regression model through the validation group, and generating a control prediction model based on the validated linear regression model; Step S3: collecting the real-time shear stress, shear rate and viscosity parameters of the slurry and inputting them into the control prediction model, and outputting the prediction results of the slurry rheological properties and additives at the next moment; Step S4: dynamically adjusting the stirring speed and shear rate of the slurry according to the prediction results of the rheological properties of the slurry, and dynamically adjusting the amount and time of the additive according to the prediction results of the additive; Step S5: loop step S3-step S4 until the slurry uniformity reaches the preset requirement.

[0007] In one aspect, in step S1, the method comprises the following steps: Step S11: extracting the shear stress, shear rate and viscosity parameters collected at fixed time intervals during the stirring process of the historical slurry according to the historical database; Step S12: extracting a timestamp matching the shear stress, shear rate and viscosity parameters according to a historical database, and then combining the timestamp with the shear stress, shear rate and viscosity parameters to form historical slurry rheological property data; Step S13: performing outlier processing and data normalization processing on all historical slurry rheological property data to generate a data set.

[0008] In one aspect, in step S2, the method comprises the following steps: Step S21: A linear combination of shear rate, shear stress and viscosity is used as a linear regression model for predicting shear stress and viscosity: , , in, for The shear stress at time is the model coefficient of shear stress, for The viscosity of time, for The shear rate at time, for The shear stress at time for The viscosity of time, is the model coefficient of viscosity, is a random error; Step S22: Use the mean square error as the shear stress and viscosity loss function: , , in, is the sample size, and are the true shear stress and viscosity, and The shear stress and viscosity predicted by the model; Step S23: Use the gradient descent method to minimize the mean square error to optimize the coefficients of the linear regression model of shear stress and viscosity: , , in, is the learning rate, which is used to control the step size of coefficient update; is the mean square error The partial derivative of is the mean square error The partial derivative of Step S24: Repeat the coefficient optimization process for a preset number of times until a coefficient combination with the minimum mean square error is generated.

[0009] In one aspect, in step S2, the method further comprises the following steps: Step S25: After the coefficient optimization is completed, the shear stress and viscosity included in the validation group are predicted using the trained linear regression model; Step S26: Calculate the mean square error of the predicted shear stress and viscosity; Step S27: Calculate the coefficient of determination of the predicted shear stress and viscosity: , , in, is the determination coefficient of shear stress, is the mean of the true values ​​of shear stress, is the coefficient of determination of viscosity, is the mean of the true values ​​of viscosity; Step S28: The linear regression model whose mean square error and determination coefficient meet the preset conditions is set as the control prediction model.

[0010] In one aspect, in step S4, the method comprises the following steps: Step S41: adjusting the stirring speed of the slurry according to the predicted viscosity and the target viscosity: , in, for The stirring speed at the time, is the target viscosity, for The viscosity of time, Gain factor for adjusting stirring speed; Step S42: adjusting the shear rate of the slurry according to the predicted shear stress and the target shear stress: , in, for The shear rate at time, is the target shear stress, for The shear stress at time Gain factor for adjusting stirring speed; Step S43: looping step S41-step S42 until the viscosity and shear stress of the slurry are consistent with the target values.

[0011] In one aspect, in step S4, the method further comprises the following steps: Step S44: judging whether the predicted shear stress and viscosity of the slurry deviate from the target value according to the real-time prediction results of the slurry rheological properties and the additive prediction results; Step S45: If yes, calculate the correction value of the additive dosage: , in, is the additive dosage correction value, is the proportional, integral and differential control coefficient; if not, then return to obtain the prediction results of the slurry rheological characteristics and additives at the next moment; Step S46: Calculate the corrected additive dosage according to the additive dosage correction value: , in, is the corrected additive dosage, is the predicted amount of additives added; Step S47: Calculate the corrected additive delivery time according to the corrected additive delivery amount; , in, is the revised additive administration time, The predicted time for adding additives.

[0012] In one aspect, when adjusting the shear rate of the slurry in step S4, the method comprises the following steps: Step S401: classifying particles contained in the slurry according to the particle size distribution of the particles in the slurry, wherein the classification includes large particle size particles, medium particle size particles and small particle size particles; Step S402: Calculate the shear rate suitable for different particle sizes according to the particle size distribution: , , , in, is the initial high shear rate, Large particle size particles, For medium shear rate, Medium-sized particles. For low shear rates, Small particle size particles, A coefficient for adjusting the base value of shear rate; is an empirical index that determines the sensitivity of shear rate to particle size changes; Step S403: According to the particle size distribution, a matching shear rate stage is automatically selected and the mixing process is performed according to the set shear rate curve.

[0013] In one aspect, in step S403, the method comprises the following steps: Step S4031: automatically selecting a matching large, medium or small shear rate stage according to the particle size distribution detected in real time; Step S4032: setting a shear rate curve for each shear rate stage, and then selecting a matching shear rate curve to execute according to the shear rate stage; Step S4033: When performing slurry mixing according to the set shear rate curve, dynamically fine-tune the shear rate according to the slurry viscosity and shear stress.

[0014] In one aspect, in step S403, the method further comprises the following steps: Step S4034: Calculate the ultrasonic time, power, and frequency of the ultrasonic generator according to the selected shear rate stage and particle size distribution: , , , in, is the empirical coefficient, is the current target particle size, is the adjustment factor used to optimize the power and frequency parameters, is the initial ultrasonic frequency, is the ultrasound time, is the ultrasonic power, is the ultrasonic frequency; Step S4035: Based on the environmental parameters, control the ultrasonic generator to operate according to the calculated ultrasonic time, power, and frequency parameters.

[0015] The present invention also provides a sodium ion battery slurry uniformity control system for executing the uniformity control method, comprising: A data processing module is used to perform data preprocessing based on the shear stress, shear rate and viscosity parameters of the collected slurry to generate a data set including a training group and a validation group; A model training module is used to continuously train the training group using a linear regression model, generate a control prediction model and deploy the control prediction model, wherein the training process includes: optimizing the linear regression model parameters by minimizing the mean square error using a gradient descent method, verifying the accuracy of the linear regression model through a verification group, and generating a control prediction model based on the verified linear regression model; The uniformity control module is used to input the real-time shear stress, shear rate and viscosity parameters of the slurry into the control prediction model, and output the prediction results of the slurry rheological characteristics and additives at the next moment; according to the prediction results of the slurry rheological characteristics, the feedback control method is used to dynamically adjust the stirring speed and shear rate process parameters of the slurry; at the same time, according to the additive prediction results, the adaptive control method is used to dynamically adjust the amount and time of additives; and then the adjustment is repeated until the slurry uniformity meets the preset requirements.

[0016] The above technical solution of the present application has the following advantages compared with the prior art: 1. During the training of the linear regression model, the model's prediction accuracy for the rheological properties of the slurry is improved through the linear combination and optimization of shear rate, shear stress and viscosity. This high-precision prediction capability can identify possible uniformity problems in the slurry in advance, thereby adjusting the production process parameters in a timely manner.

[0017] 2. The gradient descent method is used to minimize the mean square error to optimize the linear regression model parameters, ensuring the rapid convergence and accurate fitting of the model, and then the coefficients of the model can be accurately adjusted, making the control prediction model more adaptable and accurate.

[0018] 3. By controlling the prediction model to output the rheological properties and additive prediction results at the next moment, the stirring speed and shear rate are dynamically adjusted to achieve precise control of the rheological properties of the slurry; dynamically adjusting the amount and time of additives can more effectively respond to changes in slurry uniformity and ensure that high-quality slurry properties are always maintained in a complex production environment.

[0019] 4. According to the distribution of particles of different sizes in the slurry, the shear rate is calculated and dynamically adjusted to adapt to the characteristics of large, medium and small particles, thereby optimizing the slurry mixing effect; by calculating the time, power and frequency parameters of the ultrasound, the operating parameters of the ultrasonic generator are automatically adjusted according to the slurry characteristics to further improve the particle dispersion efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0021] Figure 1 It is a flow chart of a uniformity control method provided by an embodiment of the present invention.

[0022] Figure 2 It is a flow chart of a data preprocessing method provided by an embodiment of the present invention.

[0023] Figure 3 It is a flow chart of the model training method provided by an embodiment of the present invention.

[0024] Figure 4 It is a flow chart of the model verification method provided by an embodiment of the present invention.

[0025] Figure 5 It is a flow chart of a method for dynamically adjusting stirring speed and shear rate provided in an embodiment of the present invention.

[0026] Figure 6 It is a flow chart of the additive delivery method provided in an embodiment of the present invention.

[0027] Figure 7 It is a flow chart of a material particle size and shear rate adjustment method provided in an embodiment of the present invention.

[0028] Figure 8 It is a flow chart of a method for matching material force and shear rate provided in an embodiment of the present invention.

[0029] Fig. 9 4 is a flow chart of an ultrasound-assisted dispersion method provided in an embodiment of the present invention.

[0030] Fig.10 It is a schematic diagram of module connection of a uniformity control system provided in an embodiment of the present invention.

[0031] Description of the Figures in the Specification: 101. Data processing module, 102. Model training module, 103. Uniform control module, 104. Model deployment module. DETAILED DESCRIPTION

[0032] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0033] refer to Figure 1 As shown, in some embodiments, the method for controlling the uniformity of the sodium ion battery slurry includes the following steps: Step S1: Perform data preprocessing based on the shear stress, shear rate and viscosity parameters of the collected slurry to generate a data set including a training group and a validation group.

[0034] Specifically, refer to Figure 2 As shown, in step S1, the method includes the following steps: Step S11: extracting the shear stress, shear rate and viscosity parameters collected at fixed time intervals during the stirring process of the historical slurry according to the historical database, Among them, the fixed time interval is set by the operator according to actual needs, such as sampling once per second; shear stress refers to the force applied to the slurry during the flow process under the action of shear force, which is usually calculated by the shear force and the force area; shear rate refers to the relative sliding speed of adjacent liquid layers in the slurry, which is usually measured by a shear rate sensor; viscosity refers to the measure of fluid resistance, indicating the internal friction applied to the slurry during flow, which is usually obtained by the ratio of shear stress to shear rate; after collecting the parameters, the collected parameters are stored in the historical database in the form of time series for subsequent analysis.

[0035] Step S12: extracting a timestamp matching the shear stress, shear rate and viscosity parameters according to a historical database, and then combining the timestamp with the shear stress, shear rate and viscosity parameters to form historical slurry rheological property data.

[0036] Each data is stored as a sample in the historical database, where each sample is: .

[0037] Step S13: performing outlier processing and data normalization processing on all historical slurry rheological property data to generate a data set.

[0038] Among them, existing statistical methods or machine learning methods (such as box plot method, z-score analysis, etc.) are used to identify and eliminate abnormal data, and then existing data normalization methods are used to normalize all samples to the same scale (usually in the range of [0, 1]) to avoid data of different magnitudes affecting training.

[0039] Step S2: Use the linear regression model to continuously train the training group, generate a control prediction model and deploy the control prediction model, the training process includes: using the gradient descent method to minimize the mean square error to optimize the linear regression model parameters, verifying the accuracy of the linear regression model through the verification group and generating a control prediction model based on the verified linear regression model.

[0040] Specifically, refer to Figure 3 As shown, in step S2, the method includes the following steps: Step S21: A linear combination of shear rate, shear stress and viscosity is used as a linear regression model for predicting shear stress and viscosity: , , in, for The shear stress at time is the model coefficient of shear stress, for The viscosity of time, for The shear rate at time, for The shear stress at time for The viscosity of time, is the model coefficient of viscosity, is a random error.

[0041] Among them, the goal of the linear regression model is to predict the rheological properties at future times through historical data of shear rate, shear stress and viscosity.

[0042] Step S22: In order to evaluate the difference between the model prediction value and the true value, the mean square error (MSE) is used as the loss function. The loss functions for shear stress and viscosity are: , , in, is the sample size, and are the true shear stress and viscosity, and Shear stress and viscosity predicted by the model.

[0043] Step S23: Use the gradient descent method to minimize the mean square error to optimize the coefficients of the linear regression model of shear stress and viscosity: , , in, is the learning rate, which is used to control the step size of coefficient update; is the mean square error The partial derivative of is the mean square error The partial derivative of .

[0044] Step S24: Repeat the coefficient optimization process for a preset number of times until a coefficient combination with the minimum mean square error is generated.

[0045] Among them, the preset number of times is set by the operator according to actual needs.

[0046] After the model training is completed, the model is verified using the validation set that did not participate in the training. The mean square error and determination coefficient on the validation set are calculated. , to measure the predictive performance of the model.

[0047] refer to Figure 4 shown.

[0048] Step S25: After the coefficient optimization is completed, the shear stress and viscosity included in the validation group are predicted using the trained linear regression model.

[0049] Among them, the shear stress is predicted as: ; Viscosity prediction: ;in, and are the shear stress and viscosity predicted by the model, respectively.

[0050] Step S26: Calculate the mean square error of the predicted shear stress and viscosity.

[0051] Among them, the smaller the mean square error, the better, which represents the average error of the model in prediction.

[0052] Step S27: Calculate the coefficient of determination of the predicted shear stress and viscosity: , , in, is the determination coefficient of shear stress, is the mean of the true values ​​of shear stress, is the coefficient of determination of viscosity, is the mean of the true value of viscosity; the numerator is the residual sum of squares of the model, which represents the error of model prediction; the denominator is the total sum of squares, which represents the variability of the true value.

[0053] The coefficient of determination is used to measure the ability of the model to explain the variables. The closer the value is to 1, the better the model is.

[0054] Step S28: The linear regression model whose mean square error and determination coefficient meet the preset conditions is set as the control prediction model.

[0055] Among them, the preset conditions of the mean square error and the determination coefficient are set by the operator according to actual needs.

[0056] Step S3: collect the real-time shear stress, shear rate and viscosity parameters of the slurry and input them into the control prediction model, and output the prediction results of the slurry rheological properties and additives at the next moment.

[0057] In step S3, the real-time shear stress, shear rate and viscosity parameters of the slurry are collected and calculated by the rheometer and viscometer integrated in the control system, and real-time data are extracted from the equipment at a preset sampling frequency (for example, once per second). All parameters are expressed in the form of time series. At the same time, statistical methods or machine learning methods are used to identify and eliminate abnormal data, and then all data are normalized.

[0058] Step S4: dynamically adjusting the stirring speed and shear rate of the slurry according to the prediction results of the rheological properties of the slurry, and dynamically adjusting the dosage and time of the additive according to the prediction results of the additive.

[0059] Specifically, refer to Figure 5 As shown, when the stirring speed and shear rate of the slurry are dynamically adjusted according to the prediction results of the slurry rheological properties, the method includes the following steps: Step S41: adjusting the stirring speed of the slurry according to the predicted viscosity and the target viscosity: , in, for The stirring speed at the time, is the target viscosity, for The viscosity of time, Gain factor for stirring speed adjustment.

[0060] Step S42: adjusting the shear rate of the slurry according to the predicted shear stress and the target shear stress: , in, for The shear rate at time, is the target shear stress, for The shear stress at time Gain factor for adjusting stirring speed; Step S43: loop step S41 to step S43 until the viscosity and shear stress of the slurry are consistent with the target values.

[0061] In the actual implementation process, the additives of the present application include but are not limited to conductive agents, adhesives, dispersants, thickeners, wetting agents and stabilizers, etc. As long as they are added during the preparation process and have a function, they can be set as additives; among them: Conductive agents are mainly used to improve the conductivity of electrode materials and ensure the electron transmission efficiency of the battery during the charging and discharging process; common conductive additives include: carbon black, conductive carbon black (such as Super P), graphene, carbon nanotubes (CNTs), carbon fibers, etc.

[0062] The function of the binder is to combine the active material and the conductive additive with the current collector to provide mechanical strength and stability of the electrode; commonly used binders include: polyvinylidene fluoride (PVDF), polyacrylic acid (PAA), styrene butadiene rubber (SBR), etc.

[0063] Dispersants are used to improve the dispersibility of conductive additives in the slurry, prevent particle aggregation, and improve the uniformity of the slurry; commonly used dispersants are: polyethylene glycol (PEG), sodium polyphosphate (Na-Polyphosphate), polyvinyl pyrrolidone (PVP), etc.

[0064] The function of the thickener is to control the rheological properties of the slurry so that it maintains appropriate viscosity and fluidity during the coating process; common thickeners include: carboxymethyl cellulose (CMC), hydroxypropyl methylcellulose (HPMC), carbomer, xanthan gum, etc.

[0065] Wetting agents are used to improve the adhesion between the slurry and the substrate and to improve the uniformity during coating. Commonly used wetting agents include polyvinyl alcohol (PVA), surfactants (such as SDS), etc.

[0066] Stabilizers play a role in stabilizing the dispersion state in the slurry, prolonging the shelf life of the slurry, and preventing particles from settling or agglomerating; some organic acids or their salts, such as citric acid and sodium oxalate, can be used as stabilizers.

[0067] refer to Figure 6 As shown, when dynamically adjusting the amount and time of the additive according to the additive prediction result, the method further includes the following steps: Step S44: judging whether the predicted shear stress and viscosity of the slurry deviate from the target value according to the real-time prediction results of the slurry rheological properties and the additive prediction results.

[0068] Among them, the shear stress and viscosity of the slurry are monitored in real time. If it is detected that they deviate from the target value set by the operator, the amount of additive added will be dynamically adjusted based on the feedback.

[0069] Step S45: If yes, calculate the correction value of the additive dosage: , in, is the additive dosage correction value, are the proportional, integral and differential control coefficients; if not, it returns to obtain the prediction results of the slurry rheological characteristics and additives at the next moment.

[0070] Step S46: Calculate the corrected additive dosage according to the additive dosage correction value: , in, is the corrected additive dosage, is the predicted amount of additive added.

[0071] Step S47: Calculate the corrected additive delivery time according to the corrected additive delivery amount; , in, is the revised additive administration time, The predicted time for adding additives.

[0072] In the actual implementation process, the rheological properties of the slurry will be continuously monitored, and the dosing strategy will be continuously modified according to the actual situation, for example: Initial dosing: At high shear rates, larger dosing amounts may be recommended.

[0073] Mid-term adjustment: When the shear rate decreases, appropriately reduce the amount of additives added or extend the addition time based on the feedback of shear stress and viscosity.

[0074] End of feeding: When the slurry is mixed evenly and the shear stress is stable, the feeding amount will be gradually reduced and the feeding will be ended.

[0075] Step S5: loop step S3-step S4 until the slurry uniformity reaches the preset requirement.

[0076] Among them, the preset requirements are set by the operators according to actual needs. refer to Figure 7 As shown, in some embodiments, when adjusting the shear rate of the slurry in step S4, the method includes the following steps: Step S401: Classifying particles contained in the slurry according to the particle size distribution of the particles in the slurry, wherein the classification includes large-size particles, medium-size particles and small-size particles.

[0077] Among them, particles are classified according to their size range, for example: Large particle size: , Medium size particles: , Small particle size: .

[0078] Step S402: Calculate the shear rate suitable for different particle sizes according to the particle size distribution: , , , in, is the initial high shear rate, Large particle size particles, For medium shear rate, Medium-sized particles. For low shear rates, Small particle size particles, A coefficient for adjusting the base value of shear rate; is an empirical exponent that determines the sensitivity of shear rate to changes in particle size.

[0079] Among them, according to the particle size distribution, a shear rate curve is set to adapt to different particle sizes; a higher shear rate is set for large particles in the initial stage to break and disperse the particles; the shear rate is reduced in the middle stage to adapt to the dispersion of small and medium particles; and a low shear rate is set in the final stage to avoid damage to ultrafine particles.

[0080] Step S403: According to the particle size distribution, a matching shear rate stage is automatically selected and the mixing process is performed according to the set shear rate curve.

[0081] Specifically, refer to Figure 8 As shown, in step S403, the method includes the following steps: Step S4031: Automatically select a matching large, medium or small shear rate stage according to the particle size distribution detected in real time.

[0082] Step S4032: setting a shear rate curve for each shear rate stage, and then selecting a matching shear rate curve to execute according to the shear rate stage; Among them, for large particle size materials (such as conductive additives, active material agglomerates), a higher shear rate is used to quickly break up the particle agglomerates and disperse the solid phase materials; when the particle size distribution tends to be stable, the shear rate is automatically reduced to , to adapt to the dispersion of medium-sized particles and avoid excessive shear; in the later stage of the mixing process, for ultrafine particle materials and binders, maintain a low shear rate to prevent particle breakage while maintaining the uniformity of the slurry.

[0083] Step S4033: When performing slurry mixing according to the set shear rate curve, dynamically fine-tune the shear rate according to the slurry viscosity and shear stress.

[0084] Exemplarily, when the slurry viscosity or shear stress deviates from the set range, the shear rate is automatically adjusted. For example, if the viscosity decreases, it means that the particles have been dispersed to a more uniform state, and the shear rate will be reduced to prevent the particles from breaking; if the viscosity increases, it means that the particles may re-aggregate, and the shear rate will be increased to disperse the particles again.

[0085] In the actual implementation process, the operator regularly calibrates the rheological properties of the slurry so that the control system of the present application can dynamically adjust the shear rate setting value to ensure that the best dispersion effect can be achieved in different production batches.

[0086] The conductive additives and active materials in the sodium-ion battery slurry tend to agglomerate during the mixing process, resulting in uneven dispersion of the slurry, affecting the performance of the battery. Ultrasonic technology can effectively break up particle agglomeration and improve the uniformity of the slurry through the acoustic cavitation effect. That is, under the action of ultrasound, high-frequency vibrations and local high and low pressure alternations will occur in the liquid. This phenomenon is called the "acoustic cavitation effect". At low pressure, tiny bubbles are formed in the liquid; at high pressure, these bubbles collapse instantly, releasing strong local energy, which can effectively destroy the van der Waals forces between particles and promote the dispersion of particles.

[0087] Therefore, reference Fig. 9 As shown, in some embodiments, in step S403, the method further includes the following steps: Step S4034: Calculate the ultrasonic time, power, and frequency of the ultrasonic generator according to the selected shear rate stage and particle size distribution: , , , in, is the empirical coefficient, is the current target particle size, is the adjustment factor used to optimize the power and frequency parameters, is the initial ultrasonic frequency, is the ultrasound time, is the ultrasonic power, is the ultrasonic frequency.

[0088] Among them, ultrasonic time refers to the length of time of ultrasonic action determined according to shear rate and particle size; ultrasonic power refers to the power of ultrasonic wave. By setting the power of ultrasonic wave to determine the output of ultrasonic energy, the dispersion of particles is promoted; ultrasonic frequency refers to the frequency of ultrasonic wave. The frequency of ultrasonic wave is adjusted according to the particle size and required energy to improve the dispersion efficiency.

[0089] Step S4035: According to environmental parameters (such as temperature), the ultrasonic generator is controlled to operate according to the calculated ultrasonic time, power, and frequency parameters to maintain the uniformity of the slurry.

[0090] Among them, an ultrasonic generator with appropriate power is selected (for example, a laboratory-scale ultrasonic device with a frequency of 20kHz and a power of 500W, or a large-scale industrial device with a power of 2-3kW, etc.). Usually, the ultrasonic power increases in direct proportion to the slurry viscosity, and the frequency affects the dispersion effect and particle size control. In order to avoid local overheating, the slurry is circulated through the ultrasonic generating area to ensure that the slurry is evenly treated as a whole, while ensuring that the slurry temperature is maintained at a low level (usually not exceeding 50°C) to prevent overheating from causing agglomeration or decomposition of particles.

[0091] refer to Fig.10 As shown, in some embodiments, it also relates to a uniformity control system for sodium ion battery slurry that executes the uniformity control method, including: The data processing module 101 is used to perform data preprocessing based on the shear stress, shear rate and viscosity parameters of the collected slurry to generate a data set including a training group and a validation group. It is also used to extract the shear stress, shear rate and viscosity parameters collected at fixed time intervals during the stirring process of the historical slurry according to the historical database; extract the timestamp matching the shear stress, shear rate and viscosity parameters according to the historical database, and then combine the timestamp with the shear stress, shear rate and viscosity parameters to form the historical slurry rheological property data; perform outlier processing and data normalization processing on all the historical slurry rheological property data to generate a data set.

[0092] The model training module 102 is used to continuously train the training group using the linear regression model, generate a control prediction model and deploy the control prediction model. The training process includes: optimizing the linear regression model parameters by minimizing the mean square error using the gradient descent method, verifying the accuracy of the linear regression model through the verification group, and generating a control prediction model based on the verified linear regression model. It is also used to use a linear combination of shear rate, shear stress and viscosity as a linear regression model for predicting shear stress and viscosity; using the mean square error as the shear stress and viscosity loss function; minimizing the mean square error using the gradient descent method to optimize the coefficients of the linear regression model of shear stress and viscosity; repeating the optimization of the coefficients for a preset number of times until a coefficient combination with the minimum mean square error is generated. It is also used to predict the shear stress and viscosity contained in the verification group using the trained linear regression model after the coefficient optimization is completed; calculating the mean square error of the predicted shear stress and viscosity; calculating the determination coefficient of the predicted shear stress and viscosity: setting the linear regression model whose mean square error and determination coefficient meet the preset conditions as the control prediction model.

[0093] The uniform control module 103 is used to input the real-time shear stress, shear rate and viscosity parameters of the slurry into the control prediction model, and output the prediction results of the slurry rheological characteristics and additives at the next moment; according to the prediction results of the slurry rheological characteristics, the feedback control method is used to dynamically adjust the stirring speed and shear rate process parameters of the slurry, and according to the additive prediction results, the adaptive control method is used to dynamically adjust the amount and time of the additive; and then the adjustment is repeated until the uniformity of the slurry meets the preset requirements. It is also used to adjust the stirring speed of the slurry according to the predicted viscosity and the target viscosity, and adjust the shear rate of the slurry according to the predicted shear stress and the target shear stress, and repeat the cycle until the viscosity and shear stress of the slurry are consistent with the target value. It is also used to determine whether the predicted shear stress and viscosity of the slurry deviate from the target value based on the real-time prediction results of the rheological properties of the slurry and the prediction results of the additives; if so, the correction value of the additive dosage is calculated; if not, the prediction results of the rheological properties of the slurry and the prediction results of the additives at the next moment are returned; the corrected additive dosage is calculated based on the correction value of the additive dosage, and the corrected additive dosage time is calculated based on the corrected additive dosage. It is also used to classify the particles contained in the slurry according to the particle size distribution in the slurry, calculate the shear rate adapted to different particle sizes based on the particle size distribution, and automatically select the matching shear rate stage according to the particle size distribution and perform the mixing process according to the set shear rate curve. It is also used to automatically select the matching large, medium or small shear rate stage according to the particle size distribution detected in real time; set the shear rate curve for each shear rate stage, and then select the matching shear rate curve for execution according to the shear rate stage; when performing slurry mixing according to the set shear rate curve, dynamically fine-tune the shear rate according to the slurry viscosity and shear stress. It is also used to calculate the ultrasonic time, power and frequency of the ultrasonic generator according to the selected shear rate stage and particle size distribution, and control the ultrasonic generator to operate according to the calculated ultrasonic time, power and frequency parameters according to environmental parameters.

[0094] The model deployment module is used to deploy the control prediction model generated by training in the specified object.

[0095] In some embodiments, a computer medium is also provided, on which a computer program is stored, and the computer program is executed by a processor to implement the above-mentioned method for controlling the uniformity of a sodium ion battery slurry.

[0096] In some embodiments, a computer is also provided, comprising the above-mentioned computer medium.

[0097] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0098] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. A method for controlling the uniformity of sodium ion battery slurry, characterized in that: The following steps are involved: Step S1: performing data preprocessing according to the shear stress, shear rate and viscosity parameters of the collected slurry to generate a data set including a training group and a validation group; Step S2: continuously training the training group using the linear regression model to generate a control prediction model and deploy the control prediction model, wherein the training process includes: optimizing the linear regression model parameters by minimizing the mean square error using the gradient descent method, verifying the accuracy of the linear regression model through the validation group, and generating a control prediction model based on the validated linear regression model; Step S3: collecting the real-time shear stress, shear rate and viscosity parameters of the slurry and inputting them into the control prediction model, and outputting the prediction results of the slurry rheological properties and additives at the next moment; Step S4: dynamically adjusting the stirring speed and shear rate of the slurry according to the prediction results of the rheological properties of the slurry, and dynamically adjusting the amount and time of the additive according to the prediction results of the additive; Step S5: loop step S3-step S4 until the slurry uniformity reaches the preset requirement.

2. The method for controlling uniformity of sodium ion battery slurry according to claim 1, characterized in that: In step S1, the method comprises the following steps: Step S11: extracting the shear stress, shear rate and viscosity parameters collected at fixed time intervals during the stirring process of the historical slurry according to the historical database; Step S12: extracting a timestamp matching the shear stress, shear rate and viscosity parameters according to a historical database, and then combining the timestamp with the shear stress, shear rate and viscosity parameters to form historical slurry rheological property data; Step S13: performing outlier processing and data normalization processing on all historical slurry rheological property data to generate a data set.

3. A method for controlling uniformity of sodium ion battery slurry according to claim 1 or 2, characterized in that: In step S2, the method comprises the following steps: Step S21: A linear combination of shear rate, shear stress and viscosity is used as a linear regression model for predicting shear stress and viscosity: ， ， in, for The shear stress at time is the model coefficient of shear stress, for The viscosity of time, for The shear rate at time, for The shear stress at time for The viscosity of time, is the model coefficient of viscosity, is a random error; Step S22: Use the mean square error as the shear stress and viscosity loss function: ， ， in, is the sample size, and are the true shear stress and viscosity, and The shear stress and viscosity predicted by the model; Step S23: Use the gradient descent method to minimize the mean square error to optimize the coefficients of the linear regression model of shear stress and viscosity: ， ， in, is the learning rate, which is used to control the step size of coefficient update; is the mean square error The partial derivative of is the mean square error The partial derivative of Step S24: Repeat the coefficient optimization process for a preset number of times until a coefficient combination with the minimum mean square error is generated.

4. The method for controlling uniformity of sodium ion battery slurry according to claim 3, characterized in that: In step S2, the method further comprises the following steps: Step S25: After the coefficient optimization is completed, the shear stress and viscosity included in the validation group are predicted using the trained linear regression model; Step S26: Calculate the mean square error of the predicted shear stress and viscosity; Step S27: Calculate the coefficient of determination of the predicted shear stress and viscosity: ， ， in, is the determination coefficient of shear stress, is the mean of the true values ​​of shear stress, is the coefficient of determination of viscosity, is the mean of the true values ​​of viscosity; Step S28: The linear regression model whose mean square error and determination coefficient meet the preset conditions is set as the control prediction model.

5. The method for controlling uniformity of sodium ion battery slurry according to claim 1, characterized in that: In step S4, the method comprises the following steps: Step S41: adjusting the stirring speed of the slurry according to the predicted viscosity and the target viscosity: ， in, for The stirring speed at the time, is the target viscosity, for The viscosity of time, Gain factor for adjusting stirring speed; Step S42: adjusting the shear rate of the slurry according to the predicted shear stress and the target shear stress: ， in, for The shear rate at time, is the target shear stress, for The shear stress at time Gain factor for adjusting stirring speed; Step S43: loop step S41 to step S43 until the viscosity and shear stress of the slurry are consistent with the target values.

6. The method for controlling uniformity of sodium ion battery slurry according to claim 5, characterized in that: In step S4, the method further comprises the following steps: Step S44: judging whether the predicted shear stress and viscosity of the slurry deviate from the target value according to the real-time prediction results of the slurry rheological properties and the additive prediction results; Step S45: If yes, calculate the correction value of the additive dosage: ， in, is the additive dosage correction value, is the proportional, integral and differential control coefficient; if not, then return to obtain the prediction results of the slurry rheological characteristics and additives at the next moment; Step S46: Calculate the corrected additive dosage according to the additive dosage correction value: ， in, is the corrected additive dosage, is the predicted amount of additives added; Step S47: Calculate the corrected additive delivery time according to the corrected additive delivery amount; ， in, is the revised additive administration time, The predicted time for adding additives.

7. The method for controlling uniformity of sodium ion battery slurry according to claim 1, characterized in that: When adjusting the shear rate of the slurry in step S4, the method comprises the following steps: Step S401: classifying particles contained in the slurry according to the particle size distribution of the particles in the slurry, wherein the classification includes large particle size particles, medium particle size particles and small particle size particles; Step S402: Calculate the shear rate suitable for different particle sizes according to the particle size distribution: ， ， ， in, is the initial high shear rate, Large particle size particles, For medium shear rate, Medium-sized particles. For low shear rates, Small particle size particles, A coefficient for adjusting the base value of shear rate; is an empirical index that determines the sensitivity of shear rate to particle size changes; Step S403: According to the particle size distribution, a matching shear rate stage is automatically selected and the mixing process is performed according to the set shear rate curve.

8. The method for controlling uniformity of sodium ion battery slurry according to claim 1, characterized in that: In step S403, the method includes the following steps: Step S4031: automatically selecting a matching large, medium or small shear rate stage according to the particle size distribution detected in real time; Step S4032: setting a shear rate curve for each shear rate stage, and then selecting a matching shear rate curve to execute according to the shear rate stage; Step S4033: When performing slurry mixing according to the set shear rate curve, dynamically fine-tune the shear rate according to the slurry viscosity and shear stress.

9. The method for controlling uniformity of sodium ion battery slurry according to claim 7, characterized in that: In step S403, the method further includes the following steps: Step S4034: Calculate the ultrasonic time, power, and frequency of the ultrasonic generator according to the selected shear rate stage and particle size distribution: ， ， ， in, is the empirical coefficient, is the current target particle size, is the adjustment factor used to optimize the power and frequency parameters, is the initial ultrasonic frequency, is the ultrasound time, is the ultrasonic power, is the ultrasonic frequency; Step S4035: Based on the environmental parameters, control the ultrasonic generator to operate according to the calculated ultrasonic time, power, and frequency parameters.

10. A sodium ion battery slurry uniformity control system that implements the uniformity control method according to any one of claims 1 to 9, characterized in that: include: A data processing module is used to perform data preprocessing based on the shear stress, shear rate and viscosity parameters of the collected slurry to generate a data set including a training group and a validation group; A model training module is used to continuously train the training group using a linear regression model, generate a control prediction model and deploy the control prediction model, wherein the training process includes: optimizing the linear regression model parameters by minimizing the mean square error using a gradient descent method, verifying the accuracy of the linear regression model through a verification group, and generating a control prediction model based on the verified linear regression model; The uniformity control module is used to input the real-time shear stress, shear rate and viscosity parameters of the slurry into the control prediction model, and output the prediction results of the slurry rheological characteristics and additives at the next moment; according to the prediction results of the slurry rheological characteristics, the feedback control method is used to dynamically adjust the stirring speed and shear rate process parameters of the slurry; at the same time, according to the additive prediction results, the adaptive control method is used to dynamically adjust the amount and time of additives; and then the adjustment is repeated until the slurry uniformity meets the preset requirements.

Images