Method for predicting charging and discharging performance of battery according to morphology of battery pole piece

By combining the micromorphic information of lithium battery pole plates and electrochemical models, a three-dimensional geometric model is established, which solves the problem of battery performance prediction distortion in the existing technology, and achieves more accurate battery performance prediction and design guidance, reducing production costs.

CN119064792APending Publication Date: 2024-12-03GUANGXI NORMAL UNIV
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411231659.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-04
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

The existing simulation technology fails to effectively consider the micromorphic morphology of lithium battery poles, resulting in distortion of battery performance prediction, making it difficult to guide design, affecting R&D efficiency and production costs.

Method used

By obtaining the micromorphic information of the calibrated battery cell sample and combining it with the electrochemical model, a three-dimensional geometric model was established. Using technologies such as the watershed algorithm, Marching Cubes algorithm and Bruggmann equation, the electrode-diaphragm interface layer was reconstructed and the effective transport coefficient and structure were calculated. Parameters, correct the electrochemical model to achieve more accurate battery charge and discharge performance prediction.

Benefits of technology

It improves the accuracy of battery performance prediction, reduces production costs, can predict the charge and discharge curve and internal physical parameter distribution before battery production, guides the electrode interface morphology and structure design, and improves R&D efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119064792A_ABST
    Figure CN119064792A_ABST
Patent Text Reader

Abstract

The invention discloses a method for predicting the charge and discharge performance of a battery according to the morphology of a battery pole piece. The method comprises the following steps: 1) acquiring historical performance data of a calibrated battery cell sample under different working conditions; the method comprises the following steps of (1) establishing an electrochemical model, (2) establishing a preset electrochemical model, (3) calibrating and correcting the electrochemical model, and (4) carrying out performance prediction.According to the method, real microtopography information and the electrochemical model can be combined, and the performance change of a battery cell product can be predicted more truly, so that the research and development efficiency is improved, and the production cost is reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to battery simulation technology, in particular to a method for predicting battery charging and discharging performance based on battery pole piece morphology. Background Art

[0002] Lithium batteries have excellent comprehensive electrochemical properties and are widely used in consumer electronics, vehicle power units, energy storage equipment and other scenarios. The final performance of lithium batteries depends on the composition and properties of the components of the battery pole piece on the one hand, and on the battery preparation process on the other. The battery preparation process mainly includes the ingredients, coating, roll cutting, lamination, assembly, injection and formation of positive and negative pole pieces. The preparation process and parameters of the pole piece determine the final microscopic morphology of the electrode, which in turn has a decisive effect on the capacity and cycle stability of the battery. The pole piece of a lithium battery has a complex porous structure. The porous structure is connected together by active substances and conductive agents through adhesives, and adheres to the metal current collector and diaphragm to form a conductive network. In addition, the pore area is filled with electrolyte, and the electrolyte penetrates the surface of all active material particles to become a carrier for conducting lithium ions.

[0003] The thickness of lithium-ion battery pole pieces is usually 40 to 300 microns, with a deviation requirement of 1 to 2 microns. Many methods have been tried for the preparation and morphology control of lithium battery pole pieces, such as chemical vapor deposition, spray deposition, laser deposition, spin coating, dry electrode method, etc. Among them, the morphology control of the pole piece at the diaphragm interface layer is the most important. For example, in fast charging, affected by the lithium ion transport characteristics, lithium precipitation occurs preferentially in the negative electrode-diaphragm interface area, which directly affects the battery's fast charging capability and safety performance. During the experiment, R&D personnel usually use trial and error: after the battery cell is tested, the disassembly and SEM morphology characterization are performed to infer the relationship between the electrode morphology and the battery cell performance. The trial and error test process is long and costly. Predicting battery performance with the help of simulation modeling is an alternative means of accelerating design.

[0004] In terms of simulation modeling, the existing related technologies usually use electrochemical models to describe the electrochemical reactions of the battery's pole pieces to complete the prediction of battery performance. However, the related methods do not consider the influence of the pole piece's microscopic morphology, or the microscopic morphology information cannot be effectively transmitted to the model, resulting in distorted prediction results of the model and difficulty in guiding the design. Overall, the existing simulation technology is limited by computational efficiency, and the microscopic morphology information cannot be effectively transmitted to the model, resulting in inaccurate and slow calculations, and cannot effectively solve actual industrial production problems. Summary of the invention

[0005] The purpose of the present invention is to provide a method for predicting battery charge and discharge performance based on battery electrode morphology in view of the shortcomings of the prior art. This method can combine real microscopic morphology information with electrochemical models, and can more realistically predict the performance changes of battery products, thereby improving R&D efficiency and reducing production costs.

[0006] The technical solution for achieving the purpose of the present invention is:

[0007] A method for predicting battery charge and discharge performance based on battery electrode morphology comprises the following steps:

[0008] 1) Obtain the performance history data of the calibration cell samples tested under different working conditions: the calibration samples are required to be in the same material system as the subsequent samples to be tested, such as lithium iron phosphate-PP diaphragm-graphite system. The measured performance data include open circuit voltage OCV curve, DCIR, ACIR, different working conditions include different magnifications C, different operating temperatures T, and different SOCs are as follows:

[0009] 1-1) According to the specific application scenarios of the battery cells, determine the highest and lowest values ​​of the parameters at each level to ensure that the operating conditions of the battery cells cover the conventional working environment conditions. The samples tested for different level combinations are from the same batch of battery cells obtained under the same stable and same manufacturing process route;

[0010] 1-2) The computer equipment receives the physical information data of the calibration sample, including geometric dimension information: thickness of the diaphragm, positive and negative electrodes; morphological structure information of the electrode section: proportion of each component, particle information, material properties, as well as porosity, three-phase interface, specific surface area, and a preferred scheme for obtaining the morphological structure information of the electrode section, specifically using argon ion polishing technology to bombard the material sample section to obtain a flat polished section, and at the same time using a scanning electron microscope SEM to complete the observation and analysis of the microscopic characteristics of the internal structure of the sample, and using ImageJ image processing software to obtain microstructural information, and then using the above electrical performance history data as a test data benchmark for calibrating the electrochemical model;

[0011] 2) Establishing a preset electrochemical model: The computer equipment uses the physical information of the calibration cell to obtain the microstructural physical and chemical information of the calibration cell sample, including: the proportion of active material particles, the proportion of binder, the proportion of conductive agent, porosity, irregularity of active particles, particle size, particle size distribution, and intrinsic physical properties of the material, specifically:

[0012] 2-1) Reconstructing the interface layer of the electrode-diaphragm: After determining the cross-sectional morphology input information, the computer equipment randomly reconstructs the interface layer image set under the morphological feature, and then uses the watershed algorithm to perform grayscale processing and phase recognition on the image to obtain a grayscale image sequence, and the grayscale image sequence images are stacked to form a 3D microscopic geometric structure;

[0013] 2-2) Processing to obtain the physical properties and active interface of the interface layer: Based on the 3D microscopic geometric structure of the interface layer in step 2-1), the effective transport coefficient and structural parameters, i.e., porosity and tortuosity, are calculated, including:

[0014] 2-2-1) Identify the active surface based on the three-dimensional microstructure of the interface layer: The active surface of the electrode is the two-phase interface between the active material and the electrolyte. The image is identified and counted using the Marching Cubes algorithm to obtain an accurate approximation of the real surface configuration to obtain accurate active reaction surface area data;

[0015] 2-2-2) Calculation of local pore size distribution: In order to accurately calculate the local diffusion coefficient and local conductivity in the anisotropic porous structure, the computer equipment calculates the accurate distribution of local pore size values ​​in three-dimensional space based on the aforementioned three-dimensional reconstructed image data. The local pore size is calculated as follows:

[0016] The local aperture is calculated as follows: First, determine whether the current position is a pore pixel grid point. If so, start the calculation and expand along the 18 adjacent grid points until the solid phase grid point is encountered. Calculate the Euclidean distance along this direction. Finally, take the average of the distances in all directions as the effective aperture of the current grid point. Traverse the pixel grid points of the entire electrode structure space to obtain the aperture distribution in three-dimensional space, as shown in formula (1).

[0017]

[0018] In the formula, (x0, y0, z0) is the position information of the current coordinate point in three-dimensional space, (x n ,y n ,z n ) is the position information of adjacent grid points in three-dimensional space, r loc The distance between the current coordinate point and the adjacent coordinate point, i.e. the inscribed sphere aperture;

[0019] 2-2-3) Calculate the conductivity and local diffusion coefficient: According to the different phase regions, assign and store different physical parameters on the corresponding grid points, calculate the conductivity distribution in 3D space, and obtain the corrected local diffusion coefficient of each component i based on the effective pore size and Bruggmann equation, as shown in formula (2) and formula (3):

[0020]

[0021] in, is the molecular diffusion coefficient of component i, is the Knudsen diffusion coefficient of component i, ε is the porosity of the electrode or diaphragm, τ is the tortuosity, and M is the molecular mass of component i;

[0022] 2-3) Construct the three-dimensional geometry of the whole battery: Couple the advantages of high precision at the pore scale and fast calculation speed at the continuous medium scale, convert the aforementioned three-dimensional reconstructed geometry into a high-quality tetrahedral mesh, and use hexahedral meshes for the remaining areas of the battery, i.e., the diaphragm and the current collector, which can reduce the degrees of freedom and calculation amount of the model solution;

[0023] 3) Calibrate and correct the electrochemical model: Calibrate and correct the electrochemical model based on the historical experimental data of the calibration sample performance and the microstructure physical information. The calibration parameters include exchange current density, contact resistance, electrode conductivity, and electrode ion diffusion coefficient, so as to obtain a calibrated cell electrochemical model, specifically:

[0024] 3-1) Simulate the charging and discharging process of the battery cell: Set material properties and boundary conditions on the three-dimensional geometric calculation domain of the battery, apply electrochemical reaction source terms according to the activity, further solve the BV equation of lithium ion conversion and electrochemical reaction rate, the constitutive equation of components and ion-electron conservation, and simulate the charging and discharging process of the battery cell. The diffusion and transfer of electrons and ions follow the formulas (4), (5), (6), and (7):

[0025]

[0026]

[0027] i s =F∑ i z i N i (7)

[0028] In the formula, w(t,e) is the specific surface area of ​​the active particles of the electrode, which is related to time t and porosity, i s and i e are the surface currents of the solid phase and the liquid phase during reaction j, s s is the electrical conductivity of the solid phase, f el is the solid phase potential, f e2 is the potential of the electrolyte phase, F is the Faraday constant, z i is the charge number of component i, N i is the material surface flux, and the reaction rate on the active reaction surface is described by the BV equation, as shown in formula (8):

[0029]

[0030] Among them, s i,j is the stoichiometric coefficient of substance i in electrochemical reaction j, i j is the current density of electrochemical reaction j at the solid-liquid interface, i 0,j is the exchange current density of reaction j, C i,refis the reference concentration of substance i,

[0031] The overpotential of reaction j is determined according to formula (9):

[0032] h j =f e1 -f e2 -U j,ref (9)

[0033] U j,ref Corresponding to the reference concentration C i,ref The equilibrium potential is expressed as shown in formula (10):

[0034]

[0035] Based on this, by solving equations (4) to (10), iteratively solving the variables of each node until convergence, the SOC-OCV curve of the battery cell charge and discharge process and the change of material distribution are obtained. The finite element method, lattice Boltzmann method, smooth particle fluid dynamics method, and physical knowledge neural network method are used for the solver to discretize the control equation. The convergence termination condition is defined as: the difference between the electrochemical reaction current and the inlet current density Iin is less than 0.1%, and the potential difference between the current collector interfaces on both sides of the battery cell is less than the cut-off voltage of 2.0V under discharge conditions, or greater than the cut-off voltage of 3.6V under charging conditions;

[0036] 3-2) Calibrate the model based on experimental data: Solving physical equations requires many parameters, which mainly include two categories: battery design parameters and material kinetic performance parameters. Among them, battery design parameters such as thickness, capacity, component geometry, electrolyte properties, etc. can be determined by historical experimental data. For material kinetic performance parameters such as exchange current density, diffusion coefficient, conductivity, and charge transfer coefficient, they cannot be directly obtained, and their value ranges can be determined from historical data in existing literature. In summary, before calibrating the model, it is necessary to combine historical data from material experimental tests and existing literature data to determine the reasonable range of parameter values. In this process, the rationality of the values ​​should meet the physical meaning, that is, the parameter values ​​should be physically explainable, such as the diffusion coefficient should not be negative. Then, all parameters to be corrected are inverted and corrected within the value range, and the least squares error method is used to construct the objective function. The objective function is expressed as the simulated voltage V at different times i under constant current charge and discharge conditions. sim With the experimental test voltage V exp The relative deviation between them is the smallest, as shown in formula (11), where i = 1...N, and N is a natural number:

[0037]

[0038] In the formula, N is the number of calibration points selected on the discharge curve. By solving the optimization formula (11), the optimal parameter combination is found by using optimization algorithms such as genetic algorithm, particle swarm optimization (PSO), simulated annealing, etc. The parameters are modified according to the sensitivity level to minimize the residual of the OBJECT function to achieve the purpose of electrochemical model correction. The specific correction is:

[0039] 3-2-1) The parameter set to be corrected is set as a vector, and a plurality of correction parameters and value ranges are stored in the computer device, and the value range is determined by combining existing literature data and historical test data;

[0040] 3-2-2) Based on the Monte Carlo method, the parameter vector value space is sampled and sensitivity analysis is performed to obtain key correction parameters with high sensitivity;

[0041] 3-2-3) Perform hierarchical correction based on the sensitivity results. First, correct the parameters with high sensitivity. Take all the corrected parameters as variables, pass the variables to the calculation model for analysis, obtain the discharge curve calculation results, and feed the results back to the correction module; further correct the local key parameters with high sensitivity, take the corrected parameters in the previous step as the initial values, and use optimization algorithms such as genetic algorithms, particle swarm optimization (PSO), simulated annealing, Powell method, etc. to find the best parameter combination. In this way, the sensitive parameters are iteratively corrected in turn to finally obtain the calibrated model;

[0042] 4) Performance prediction: Obtain the cross-sectional morphology information of the battery product to be tested, reconstruct the cross-sectional morphology into a heterogeneous morphology layer, determine whether the analyzed cross-sectional morphology information is located on the positive electrode or negative electrode side, and update the corresponding electrode-diaphragm calculation area grid to obtain a complete three-dimensional calculation domain of the battery to be tested, and predict the battery cell to be tested based on the calibrated battery cell electrochemical model to obtain the product's charge and discharge performance curve and the distribution of current, lithium ions, and intermediate products, including:

[0043] 4-1) Obtain the initial cross-sectional morphology physical information of the electrode of the battery product to be tested, including the proportion of active material particles, the proportion of binder, the proportion of conductive agent, porosity, irregularity of active particles, particle size, particle size distribution, and intrinsic physical properties of the material;

[0044] 4-2) Determine whether the electrode under study is the positive electrode side or the negative electrode side, perform three-dimensional random reconstruction using the QSGS method based on the cross-sectional morphology information, obtain a three-dimensional structure consistent with the actual microscopic morphology, generate a grid and update the electrode calculation domain of the sample to be tested accordingly;

[0045] 4-3) Run the model to predict the SOC-OCV curve and internal physical quantity distribution of the battery cell to be tested, and obtain the performance prediction of the product.

[0046] This technical solution obtains the historical test data of the electrical performance of the calibration battery sample under different working conditions. Different working conditions refer to different operating temperatures, different current densities, and different numbers of hot and cold cycles. The electrical performance includes open circuit voltage OCV (opencircuit potential, referred to as OCV)-battery state of charge SOC (State of Charge, referred to as SOC) curve, direct current impedance DCIR, (Direct Current Internal Resistance, referred to as DCIR), alternating current internal resistance ACIR (Alternating Current Internal Resistance, referred to as ACIR), electrochemical impedance spectroscopy EIS (Electrochemical Impedance Spectrum In addition, the microstructural physical information of the electrode-diaphragm cross section of the calibration cell sample is obtained, including pore distribution, particle size and distribution, and the ratio of binder and conductive agent. Specifically, the argon ion polishing technology is used to bombard the cross section of the material sample to obtain a flat polished cross section. At the same time, the scanning electron microscope SEM is used to observe and analyze the microscopic characteristics of the internal structure of the sample. The ImageJ image processing software is used to obtain the microstructural information. Then, the above electrical performance historical data is used as the test data benchmark for calibrating the electrochemical model.

[0047] This technical solution calibrates the pre-constructed electrochemical model parameters according to the historical calibration data obtained. The calibration parameters include exchange current density, contact resistance, conductivity, and diffusion coefficient, so as to obtain a calibrated electrochemical model. Specifically, the physical domain of the pole piece-diaphragm-current collector is grid discretized to obtain a numerical calculation domain, the input material properties are assigned to the corresponding calculation domain, and the input working conditions are converted into the corresponding boundary conditions of the calculation domain; the numerical model is constructed by combining the BV equation of lithium ion conversion and electrochemical reaction rate, the component conservation equation, and the ion-electron conservation equation; the electrochemical reaction source term is applied in the heterogeneous region of the pole piece-diaphragm according to the distribution of the identified active reaction surface; the above equation group is numerically solved, and the iterative calculation is performed until convergence;

[0048] This technical solution inputs the microstructure information of the cross-sectional morphology of the pole piece-diaphragm of the battery product to be tested, and then predicts the charge and discharge performance of the product to be tested, specifically: the cross-sectional morphology image of the pole piece-diaphragm is obtained as input, and a set of pole piece-diaphragm intermediate layer images under the morphological feature is obtained by random three-dimensional reconstruction, and the reconstruction thickness is 5-10μm. The images are stacked to obtain a reconstructed 3D geometric structure. According to the 3D structure, the Marching cube algorithm is used to identify the active area of ​​the two-phase interface, and the interface activity is judged according to whether it constitutes an electron\ion path. The local pore size distribution is calculated and converted into the local conductivity and local diffusion coefficient distribution; the three-dimensional geometry of the whole battery is constructed, in which the heterogeneous morphology layer is divided and encrypted by tetrahedral mesh, and the remaining areas are modeled by homogenized continuous media and connected to the interface layer through unstructured meshes. Then, the input operating temperature and current information are converted into boundary conditions, updated to the input into the calibrated electrochemical model, and the model calculation is performed to output the charge and discharge curve of the product to be tested and the prediction of the distribution of key parameters.

[0049] This technical solution combines real microscopic morphology information with electrochemical models, which can more realistically predict the performance changes of battery products. With this technical solution, engineers can predict the charge and discharge curves of battery products based on the microscopic morphology images of the electrode-diaphragm cross section prepared by obtaining specific incoming materials, different process technologies and equipment before making batteries, and obtain the law between its capacity and voltage, as well as the distribution law of the key internal physical parameters of the battery when it is in service. Based on this, the morphology structure of the electrode interface and the process raw materials can be positively designed, thereby improving R&D efficiency and reducing production costs.

[0050] The advantages of this technical solution are as follows:

[0051] (1) Although there are models in the prior art to predict battery performance, the focus is not on prediction through battery electrode morphology. The present technical solution is to combine the electrode-diaphragm cross-sectional morphology information with the electrochemical model, which can more realistically predict the performance changes of battery cell products;

[0052] (3) This technical solution can be used to positively design the morphology structure and process raw materials of the electrode interface, thereby improving R&D efficiency and reducing production costs. For example, before battery production, engineers can obtain specific incoming materials, electrode interface morphology information prepared by different process technologies and equipment, predict the charge and discharge SOC-OCV curve of the battery product, and obtain the law between its capacity and voltage. Based on this, they can positively design the electrode interface morphology structure, greatly reducing production costs.

[0053] (3) This technical solution can be used to monitor key internal physical parameters of battery cells during service. Since the battery is a closed system, certain key internal parameters cannot be directly obtained through sensors, which makes it impossible to monitor potential risks during battery operation. This technical solution can be used to establish a digital twin of the battery to truly reflect the multi-field distribution inside the battery.

[0054] This method can combine real microscopic morphology information with electrochemical models, and can more realistically predict the performance changes of battery cell products, thereby improving R&D efficiency and reducing production costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 A schematic diagram of a method flow chart of an embodiment;

[0056] Figure 2 Schematic diagram of the positive electrode-diaphragm interface layer morphology of the battery cell in Example 1;

[0057] Figure 3 This is a schematic diagram of the calibration of the 0.2C discharge curve of the battery cells with different positive electrode-separator interface layer morphologies in Example 1;

[0058] Figure 4 This is a schematic diagram of the prediction of current and material distribution during the 0.2C discharge process of the battery cell in Example 1;

[0059] Figure 5 Schematic diagram of the positive electrode-diaphragm interface layer morphology of the battery cell in Example 2;

[0060] Figure 6 Schematic diagram of the prediction of 1C charging SOC-OCV curves of battery cells with different negative electrode-diaphragm interface morphologies in Example 2. DETAILED DESCRIPTION

[0061] The content of the present invention is further described below in conjunction with the drawings and embodiments, but the present invention is not limited thereto.

[0062] Embodiment 1:

[0063] Reference Figure 1 A method for predicting battery charge and discharge performance based on battery electrode morphology comprises the following steps:

[0064] 1) Obtain the performance history data of the calibration cell samples tested under different working conditions: the calibration samples are required to be in the same material system as the subsequent samples to be tested, such as lithium iron phosphate-PP diaphragm-graphite system. The measured performance data include open circuit voltage OCV curve, DCIR, ACIR, different working conditions include different magnifications C, different operating temperatures T, and different SOCs are as follows:

[0065] 1-1) According to the specific application scenarios of the battery cells, determine the highest and lowest values ​​of the parameters at each level to ensure that the operating conditions of the battery cells cover the conventional working environment conditions. The samples tested for different level combinations are from the same batch of battery cells obtained under the same stable and same manufacturing process route;

[0066] 1-2) The computer equipment receives the physical information data of the calibration sample, including geometric dimension information: thickness of the diaphragm, positive and negative electrodes; morphological structure information of the electrode section: proportion of each component, particle information, material properties, as well as porosity, three-phase interface, specific surface area, and a preferred scheme for obtaining the morphological structure information of the electrode section, specifically using argon ion polishing technology to bombard the material sample section to obtain a flat polished section, and at the same time using a scanning electron microscope SEM to complete the observation and analysis of the microscopic characteristics of the internal structure of the sample, and using ImageJ image processing software to obtain microstructural information, and then using the above electrical performance history data as a test data benchmark for calibrating the electrochemical model;

[0067] 2) The computer equipment establishes a preset electrochemical model through the physical information of the calibration cell: obtains the microstructural physical and chemical information of the calibration cell sample, including: the proportion of active material particles, the proportion of binder, the proportion of conductive agent, porosity, irregularity of active particles, particle size, particle size distribution, and intrinsic physical properties of the material, specifically:

[0068] 2-1) Reconstructing the interface layer of the electrode-diaphragm: After determining the cross-sectional morphology input information, the computer equipment randomly reconstructs the interface layer image set under the morphological feature, and then uses the watershed algorithm to perform grayscale processing and phase recognition on the image to obtain a grayscale image sequence, and the grayscale image sequence images are stacked to form a 3D microscopic geometric structure;

[0069] 2-2) Processing to obtain the physical properties and active interface of the interface layer: Based on the 3D microscopic geometric structure of the interface layer in step 2-1), the effective transport coefficient and structural parameters, i.e., porosity and tortuosity, are calculated, including:

[0070] 2-2-1) Identify the active surface based on the three-dimensional microstructure of the interface layer: The active surface of the electrode is the two-phase interface between the active material and the electrolyte. The image is identified and counted using the Marching Cubes algorithm to obtain an accurate approximation of the real surface configuration to obtain accurate active reaction surface area data;

[0071] 2-2-2) Calculate the local pore size distribution: The local pore size is calculated as follows: First, determine whether the current position is a pore pixel grid point. If so, start the calculation and expand along the 18 adjacent grid points until the solid phase grid point is encountered. Calculate the Euclidean distance along this direction. Finally, take the average of the distances in all directions as the effective pore size of the current grid point. Traverse the pixel grid points of the entire electrode structure space to obtain the pore size distribution in three-dimensional space, as shown in formula (1).

[0072]

[0073] In the formula, (x0, y0, z0) is the position information of the current coordinate point in three-dimensional space, (x n ,y n ,z n ) is the position information of adjacent grid points in three-dimensional space, r loc The distance between the current coordinate point and the adjacent coordinate point, i.e. the inscribed sphere aperture;

[0074] 2-2-3) Calculate the conductivity and local diffusion coefficient: According to the different phase regions, assign and store different physical parameters on the corresponding grid points, calculate the conductivity distribution in 3D space, and obtain the corrected local diffusion coefficient of each component i based on the effective pore size and Bruggmann equation, as shown in formula (2) and formula (3):

[0075]

[0076] in, is the molecular diffusion coefficient of component i, is the Knudsen diffusion coefficient of component i, ε is the porosity of the electrode or diaphragm, τ is the tortuosity, and M is the molecular mass of component i;

[0077] 2-3) Construct the three-dimensional geometry of the whole battery: By coupling the advantages of high precision at the pore scale and fast calculation speed at the continuous medium scale, the three-dimensional reconstructed geometry is converted into a high-quality tetrahedral mesh, and the remaining areas of the battery, i.e., the diaphragm and the current collector, are meshed with hexahedral meshes; this can reduce the degrees of freedom and computational complexity of the model.

[0078] In this example, refer to Figure 2 Taking lithium-sulfur battery as an example, after obtaining the cross-sectional morphology of sulfur-carbon cathode, the random reconstruction algorithm, namely the random growth four-parameter method QSGS (Quartet Structure Generation Set, referred to as QSGS), was used to calculate and generate Figure 3 The displayed cathode-separator microstructure morphology layers correspond to the long carbon fiber-like, short carbon fiber-like, and powder-like morphologies from left to right;

[0079] 3) Calibrate and correct the electrochemical model: Calibrate and correct the electrochemical model based on the historical experimental data of the calibration sample performance and the microstructure physical information. The calibration parameters include exchange current density, contact resistance, electrode conductivity, and electrode ion diffusion coefficient, so as to obtain a calibrated cell electrochemical model, specifically:

[0080] 3-1) Simulate the charging and discharging process of the battery cell: Set material properties and boundary conditions on the three-dimensional geometric calculation domain of the battery, apply electrochemical reaction source terms according to the activity, further solve the BV equation of lithium ion conversion and electrochemical reaction rate, the constitutive equation of components and ion-electron conservation, and simulate the charging and discharging process of the battery cell. The diffusion and transfer of electrons and ions follow the formulas (4), (5), (6), and (7):

[0081]

[0082] i s =F∑ i z i N i (7)

[0083] In the formula, w(t,e) is the specific surface area of ​​the active particles of the electrode, which is related to time t and porosity, i s and i e are the surface currents of the solid phase and the liquid phase during reaction j, s s is the electrical conductivity of the solid phase, f el is the solid phase potential, f e2 is the potential of the electrolyte phase, F is the Faraday constant, z i is the charge number of component i, N i is the material surface flux, and the reaction rate on the active reaction surface is described by the BV equation, as shown in formula (8):

[0084]

[0085] Among them, s i,j is the stoichiometric coefficient of substance i in electrochemical reaction j, i j is the current density of electrochemical reaction j at the solid-liquid interface, i 0,j is the exchange current density of reaction j, C i,ref is the reference concentration of substance i,

[0086] The overpotential of reaction j is determined according to formula (9):

[0087] h j =f e1 -f e2 -U j,ref (9)

[0088] U j,ref Corresponding to the reference concentration C i,ref The equilibrium potential is expressed as shown in formula (10):

[0089]

[0090] Based on this, by solving equations (4) to (10), iteratively solving the variables of each node until convergence, the SOC-OCV curve of the battery cell charge and discharge process and the change of material distribution are obtained. The finite element method, lattice Boltzmann method, smooth particle fluid dynamics method, and physical knowledge neural network method are used for the solver to discretize the control equation. The convergence termination condition is defined as: the difference between the electrochemical reaction current and the inlet current density Iin is less than 0.1%; and the potential difference between the current collector interfaces on both sides of the battery cell, which is less than the cut-off voltage of 2.0V under discharge conditions, or greater than the cut-off voltage of 3.6V under charging conditions;

[0091] 3-2) Calibrate the model based on experimental data: Solving physical equations requires many parameters, which mainly include two categories: battery design parameters and material kinetic performance parameters. Among them, battery design parameters such as thickness, capacity, component geometry, electrolyte properties, etc. can be determined by historical experimental data. The value range of material kinetic performance parameters such as exchange current density, diffusion coefficient, conductivity, and charge transfer coefficient cannot be directly obtained, and the value range can be determined from historical data in existing literature. In summary, before calibrating the model, it is necessary to combine historical data of material experimental tests and existing literature data to determine the reasonable range of parameter values. In this process, the rationality of the value should meet the physical meaning, that is, the parameter value should be physically explainable, such as the diffusion coefficient should not be negative. In this case, taking the lithium iron phosphate-PP diaphragm-graphite system as an example, when studying the influence of graphite negative electrode morphology, the parameters to be corrected, the reasonable value range and the basis for the value are summarized as shown in Table 1:

[0092] Table 1 Battery parameters to be revised, value ranges and value basis

[0093]

[0094] Then, all parameters to be corrected are inversely corrected within the range of values, and the objective function is constructed using the least squares error method. The objective function is expressed as the simulated voltage V at different times i under constant current charge and discharge conditions. sim With the experimental test voltage V exp The relative deviation between them is the smallest, as shown in formula (11), where i = 1...N, and N is a natural number:

[0095]

[0096] In the formula, N is the number of calibration points selected on the discharge curve. By solving the optimization formula (11), the optimal parameter combination is found by using optimization algorithms such as genetic algorithm, particle swarm optimization (PSO), simulated annealing, etc. The parameters are modified according to the sensitivity level to minimize the residual of the OBJECT function to achieve the purpose of electrochemical model correction. The specific correction is:

[0097] 3-2-1) The parameter set to be corrected is set as a vector, and a plurality of correction parameters and value ranges are stored in the computer device, and the value ranges are determined by combining existing literature data and historical test data, as shown in Table 1;

[0098] 3-2-2) Based on the Monte Carlo method, the parameter vector value space is sampled and sensitivity analysis is performed to obtain key correction parameters with high sensitivity;

[0099] 3-2-3) Perform hierarchical correction based on the sensitivity results. First, correct the parameters with high sensitivity. Take all the corrected parameters as variables, pass the variables to the calculation model for analysis, obtain the discharge curve calculation results, and feed the results back to the correction module; further correct the local key parameters with high sensitivity, take the corrected parameters in the previous step as the initial values, and use optimization algorithms such as genetic algorithms, particle swarm optimization (PSO), simulated annealing, Powell method, etc. to find the best parameter combination. In this way, the sensitive parameters are iteratively corrected in turn to finally obtain the calibrated model;

[0100] In this example, if Figure 3 As shown in the figure, taking lithium-sulfur battery as an example, the discharge curve calibration results of a battery cell with different cathode-diaphragm interface morphology at 0.2C and 25℃ are shown. It can be seen that the model prediction value is in good agreement with the experimental value. In particular, since the model fully considers the heterogeneous structure of the diaphragm-cathode interface layer and the conversion process of polysulfides, it can completely predict the dual-platform discharge voltage distribution curve of lithium-sulfur battery.

[0101] 4) Performance prediction: After obtaining the cross-sectional morphology information of the battery product to be tested, the cross-sectional morphology is reconstructed into a heterogeneous morphology layer, and the analyzed cross-sectional morphology information is determined to be located on the positive electrode or negative electrode side, and the corresponding electrode-diaphragm calculation area grid is updated to obtain a complete three-dimensional calculation domain of the battery to be tested. The battery cell to be tested is predicted based on the calibrated electrochemical model of the battery cell to obtain the product's charge and discharge performance curve and the distribution of current, lithium ions, and intermediate products, such as Figure 4 As shown, including:

[0102] 4-1) Obtain the initial cross-sectional morphology physical information of the electrode of the battery product to be tested, including the proportion of active material particles, the proportion of binder, the proportion of conductive agent, porosity, irregularity of active particles, particle size, particle size distribution, and intrinsic physical properties of the material;

[0103] 4-2) Determine whether the electrode under study is the positive electrode side or the negative electrode side, perform three-dimensional random reconstruction using the QSGS method based on the cross-sectional morphology information, obtain a three-dimensional structure consistent with the actual microscopic morphology, generate a grid and update the electrode calculation domain of the sample to be tested accordingly;

[0104] 4-3) Run the model to predict the SOC-OCV curve and internal physical quantity distribution of the battery cell to be tested, and obtain the performance prediction of the product.

[0105] In this case, taking lithium-sulfur batteries as an example, the current density distribution in sulfur electrodes with different morphologies and the predicted distribution results of the intermediate product Li2S6 are shown. It can be seen that at low discharge depth (20% DoD), the main reactions are the dissolution of S8 and the generation of Li2S8. At this time, the reaction zone is concentrated in the electrode layer. At 80% DoD, with the subsequent dissolution of polysulfides and entering the interface layer, the reaction zone is extended to the entire electrode. In particular, since the long carbon fiber interface layer has the highest active surface area, the current density of the sulfur electrode containing the long carbon fiber interface layer is the highest at the same DoD. From the distribution of the intermediate product Li2S6, it can be seen that as the discharge depth increases from 20% to 80%, Li2S6 is gradually consumed and the concentration decreases. Due to the low porosity of the powder morphology interface layer, its Li2S6 concentration is the lowest at the same DoD.

[0106] Embodiment 2:

[0107] Taking lithium iron phosphate-graphite system battery cells as an example, the negative electrode-diaphragm morphology affects the performance prediction. Graphite is an intercalated layered material. During rapid charge and discharge, the Li ion flux near the diaphragm side is high, and the corresponding SOC and diffusion stress are higher, resulting in early lithium precipitation, particle crack formation and degradation. Therefore, the morphology of graphite particles in the negative electrode-diaphragm interface layer has a significant effect on the charge and discharge curve. In step 4), based on the microstructure information of the negative electrode-diaphragm cross-section of the battery cell to be tested, the negative electrode-diaphragm interface layer morphology layer is reconstructed, such as Figure 5 As shown, after that, the morphology update information is input into the calibrated model of step 3), and the charging performance and material distribution prediction results of the battery cell are calculated. After executing step 4), as shown in FIG. Figure 6 As shown in the figure, the predicted results of 1C charging SOC-OCV curves of cells with different morphologies of the negative electrode-diaphragm interface layer in this example, it can be seen that the morphology difference of the graphite active material in the interface layer affects the size of the active reaction area and the Li ion concentration polarization, which in turn leads to differences in the SOC-OCV curves. The OCV of cells with different morphologies undergoes a step of about 50mV near 50% SOC, which is mainly due to the lithium intercalation phase transformation of the negative electrode graphite.

[0108] The rest of this example is the same as in Example 1.

[0109] In summary, it can be seen that in this embodiment, a three-dimensional electrochemical model integrating the real microstructural morphology is established through the microstructural information of the actual battery cell electrode cross section. The electrochemical model created in this way can more realistically and accurately predict the influence of the electrode-diaphragm cross-section microstructural morphology, and the electrochemical performance information obtained is comprehensive, which is conducive to the optimization design of subsequent processes.

Claims

1. A method for predicting battery charge and discharge performance based on battery electrode morphology, It is characterized in that The steps include: 1) Obtain the performance history data of the calibration cell samples tested under different working conditions: The calibration samples are required to be in the same material system as the subsequent samples to be tested. The measured performance data include open circuit voltage OCV curve, DCIR, ACIR, and different working conditions include different magnifications C, different operating temperatures T, and different SOCs, specifically: 1-1) According to the specific application scenarios of the battery cells, determine the highest and lowest values ​​of the parameters at each level to ensure that the operating conditions of the battery cells cover the conventional working environment conditions. The samples tested for different level combinations are from the same batch of battery cells obtained under the same stable and same manufacturing process route; 1-2) The computer equipment receives the physical information data of the calibration sample, including geometric dimension information: thickness of the diaphragm, positive and negative electrodes; morphological structure information of the electrode section: proportion of each component, particle information, material properties, as well as porosity, three-phase interface, specific surface area, and a preferred scheme for obtaining the morphological structure information of the electrode section, specifically using argon ion polishing technology to bombard the material sample section to obtain a flat polished section, and at the same time using a scanning electron microscope SEM to complete the observation and analysis of the microscopic characteristics of the internal structure of the sample, and using ImageJ image processing software to obtain microstructural information, and then using the above electrical performance history data as a test data benchmark for calibrating the electrochemical model; 2) Establish a preset electrochemical model: The computer equipment uses the physical information of the calibration cell to obtain the microstructure physical and chemical information of the calibration cell sample, including: the proportion of active material particles, the proportion of binder, the proportion of conductive agent, porosity, irregularity of active particles, particle size, particle size distribution, and intrinsic physical properties of the material, specifically: 2-1) Reconstructing the interface layer of the electrode-diaphragm: After determining the cross-sectional morphology input information, the computer equipment randomly reconstructs the interface layer image set under the morphological feature, and then uses the watershed algorithm to perform grayscale processing and phase recognition on the image to obtain a grayscale image sequence, and the grayscale image sequence images are stacked to form a 3D microscopic geometric structure; 2-2) Processing to obtain the physical properties and active interface of the interface layer: Based on the 3D microscopic geometric structure of the interface layer in step 2-1), the effective transport coefficient and structural parameters, i.e., porosity and tortuosity, are calculated, including: 2-2-1) Identify the active surface based on the three-dimensional microstructure of the interface layer: The active surface of the electrode is the two-phase interface between the active material and the electrolyte. The image is identified and counted using the Marching Cubes algorithm to obtain an accurate approximation of the real surface configuration to obtain accurate active reaction surface area data; 2-2-2) Calculate the local pore size distribution: The local pore size is calculated as follows: First, determine whether the current position is a pore pixel grid point. If so, start the calculation and expand along the 18 adjacent grid points until the solid phase grid point is encountered. Calculate the Euclidean distance along this direction. Finally, take the average of the distances in all directions as the effective pore size of the current grid point. Traverse the pixel grid points of the entire electrode structure space to obtain the pore size distribution in three-dimensional space, as shown in formula (1). In the formula, (x 0 ,y 0 ,z 0 ) is the position information of the current coordinate point in three-dimensional space, (x n ,y n ,z n ) is the position information of adjacent grid points in three-dimensional space, r loc The distance between the current coordinate point and the adjacent coordinate point, i.e. the inscribed sphere aperture; 2-2-3) Calculate the conductivity and local diffusion coefficient: According to the different phase regions, assign and store different physical parameters on the corresponding grid points, calculate the conductivity distribution in 3D space, and obtain the corrected local diffusion coefficient of each component i based on the effective pore size and Bruggmann equation, as shown in formula (2) and formula (3): in, is the molecular diffusion coefficient of component i, is the Knudsen diffusion coefficient of component i, ε is the porosity of the electrode or diaphragm, τ is the tortuosity, and M is the molecular mass of component i; 2-3) Constructing the three-dimensional geometry of the whole battery: The three-dimensional reconstructed geometry is converted into a high-quality tetrahedral mesh, and the remaining areas of the battery, i.e., the separator and the current collector, are meshed with hexahedral meshes; 3) Calibrate and correct the electrochemical model: Calibrate and correct the electrochemical model based on the historical experimental data of the calibration sample performance and the microstructure physical information. The calibration parameters include exchange current density, contact resistance, electrode conductivity, and electrode ion diffusion coefficient, so as to obtain a calibrated cell electrochemical model, specifically: 3-1) Simulate the charging and discharging process of the battery cell: Set material properties and boundary conditions on the three-dimensional geometric calculation domain of the battery, apply electrochemical reaction source terms according to the activity, further solve the BV equation of lithium ion conversion and electrochemical reaction rate, the constitutive equation of components and ion-electron conservation, and simulate the charging and discharging process of the battery cell. The diffusion and transfer of electrons and ions follow the formulas (4), (5), (6), and (7): i s =F∑ i z i N i (7), In the formula, ω(t,e) is the specific surface area of ​​the active particles of the electrode, which is related to time t and porosity, i s and i e are the surface currents of the solid phase and the liquid phase during reaction j, s s is the electrical conductivity of the solid phase, φ el is the solid phase potential, φ e2 is the potential of the electrolyte phase, F is the Faraday constant, z i is the charge number of component i, N i is the material surface flux, and the reaction rate on the active reaction surface is described by the BV equation, as shown in formula (8): Among them, s i,j is the stoichiometric coefficient of substance i in electrochemical reaction j, i j is the current density of electrochemical reaction j at the solid-liquid interface, i 0,j is the exchange current density of reaction j, C i,ref is the reference concentration of substance i, The overpotential of reaction j is determined according to formula (9): or j =φ e1 -f e2 -U j,ref (9), U j,ref Corresponding to the reference concentration C i,ref The equilibrium potential is expressed as shown in formula (10): Based on this, by solving equations (4) to (10), iteratively solving the variables of each node until convergence, the SOC-OCV curve of the battery cell charge and discharge process and the change of material distribution are obtained. The finite element method, lattice Boltzmann method, smooth particle fluid dynamics method, and physical knowledge neural network method are used for the solver to discretize the control equation. The convergence termination condition is defined as: the difference between the electrochemical reaction current and the inlet current density Iin is less than 0.1%, and the potential difference between the current collector interfaces on both sides of the battery cell is less than the cut-off voltage of 2.0V under discharge conditions, or greater than the cut-off voltage of 3.6V under charging conditions; 3-2) Calibrate the model based on experimental data: The parameters required to solve the physical equations are divided into two categories: battery design parameters and material kinetic performance parameters. The battery design parameters, namely thickness, capacity, component geometry, and electrolyte properties, have their value ranges determined by historical experimental data. The material kinetic performance parameters, namely exchange current density, diffusion coefficient, conductivity, and charge transfer coefficient, cannot be directly obtained, and their value ranges are determined from historical data in existing literature. Before calibrating the model, the parameters are determined by combining historical data from material experimental tests and existing literature data. The rationality of the values ​​should meet the physical meaning, that is, the parameter values ​​should be physically explainable. Then, all parameters to be corrected are inverted and corrected within the value range, and the least squares error method is used to construct the objective function. The objective function is expressed as the simulated voltage V at different times i under constant current charge and discharge conditions. sim With the experimental test voltage V exp The relative deviation between them is the smallest, as shown in formula (11), where i = 1...N, and N is a natural number: In the formula, N is the number of calibration points selected on the discharge curve. By solving the optimization formula (11), the optimization algorithm, namely genetic algorithm, particle swarm optimization PSO, and simulated annealing, is used to find the best parameter combination. The parameters are modified according to the sensitivity level so that the residual of the OBJECT function is minimized to achieve the purpose of electrochemical model correction. The specific correction is: 3-2-1) The parameter set to be corrected is set as a vector, and a plurality of correction parameters and value ranges are stored in the computer device, and the value range is determined by combining existing literature data and historical test data; 3-2-2) Based on the Monte Carlo method, the parameter vector value space is sampled and sensitivity analysis is performed to obtain key correction parameters with high sensitivity; 3-2-3) Perform hierarchical correction based on the sensitivity results. First, correct the parameters with high sensitivity. Take all the corrected parameters as variables, pass the variables to the calculation model for analysis, obtain the discharge curve calculation results, and feed the results back to the correction module; further correct the local key parameters with high sensitivity, take the corrected parameters in the previous step as the initial values, use optimization algorithms including genetic algorithms, particle swarm optimization PSO, simulated annealing method, and Powell method to find the best parameter combination, iteratively correct the sensitive parameters in turn, and finally obtain the calibrated model; 4) Performance prediction: Obtain the cross-sectional morphology information of the battery product to be tested, reconstruct the cross-sectional morphology into a heterogeneous morphology layer, determine whether the analyzed cross-sectional morphology information is located on the positive electrode or negative electrode side, and update the corresponding electrode-diaphragm calculation area grid to obtain a complete three-dimensional calculation domain of the battery to be tested, and predict the battery cell to be tested based on the calibrated battery cell electrochemical model to obtain the product's charge and discharge performance curve and the distribution of current, lithium ions, and intermediate products, including: 4-1) Obtain the initial cross-sectional morphology physical information of the electrode of the battery product to be tested, including the proportion of active material particles, the proportion of binder, the proportion of conductive agent, porosity, irregularity of active particles, particle size, particle size distribution, and intrinsic physical properties of the material; 4-2) Determine whether the electrode under study is the positive electrode side or the negative electrode side, perform three-dimensional random reconstruction using the QSGS method based on the cross-sectional morphology information, obtain a three-dimensional structure consistent with the actual microscopic morphology, generate a grid and update the electrode calculation domain of the sample to be tested accordingly; 4-3) Run the model to predict the SOC-OCV curve and internal physical quantity distribution of the battery cell to be tested, and obtain the performance prediction of the product.

Citation Information

Cited By

  • Ternary lithium battery pole piece performance test method and system

    CN120820115A

  • Method for evaluating integrity of silane electrophoresis film of automobile sheet based on conjoint analysis of surface topography characteristics and conductivity

    CN121114112A