A method and system for online decoupling identification of model parameters of a lithium-ion battery pack

Abstract

The application discloses a kind of lithium ion battery pack model parameter online decoupling identification method and system, comprising: based on the basic working principle of lithium ion battery, the multi-physical field coupling model of lithium ion battery pack is constructed;Combined with the electrical and thermal connection between different monomer lithium ion batteries, the multi-physical field coupling model extension is completed, and parameter offline identification is carried out;Study the sensitivity relationship between offline parameters and lithium ion battery external characteristics, select the characteristic parameters of the extended multi-physical field coupling model, and complete the multi-physical field coupling model parameter online identification based on least square method and dichotomy.The multi-physical field coupling model of the application can realize the accurate simulation of battery terminal voltage and three-dimensional shell temperature and the online identification of model parameters.

Description

Technical Field

[0001] This invention belongs to the field of online decoupling identification technology for lithium-ion battery packs, specifically relating to an online decoupling identification method and system for lithium-ion battery pack model parameters. Background Technology

[0002] With increasingly serious problems such as environmental pollution and energy shortages, lithium-ion batteries, as a clean energy source, have been widely used in many industrial fields, including electric vehicles. The main factor restricting their further development is battery safety and reliability. Therefore, during battery use, a management system is needed to strengthen the monitoring of battery current, voltage, and battery status.

[0003] The construction of battery models is a crucial component of battery management systems. Data-driven models are complex to implement due to the need for large amounts of data and specific algorithms for data preprocessing. In contrast, electrochemical models can accurately describe the mechanistic characteristics of batteries, providing specific mathematical descriptions of internal reactions and mass diffusion, resulting in higher simulation accuracy. However, the internal reaction processes of batteries are complex, and their internal behavior is simultaneously influenced by the combined interactions of multiple physical fields. Furthermore, battery packs are more complex than individual cells, making characterization more difficult. Therefore, establishing accurate multi-physics coupled models of battery packs is essential. Additionally, the low real-time simulation accuracy of models under complex operating conditions and across the entire operating range urgently needs to be addressed. Integrating real-time monitoring data with mechanistic mechanisms and achieving online identification of model parameters is key. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention proposes an online decoupling identification method and system for lithium-ion battery pack model parameters. Specifically, the method for constructing a multiphysics coupling model for a square battery pack and its online decoupling identification of parameters consists of the following three parts: electrochemical model construction; battery pack model expansion and offline parameter identification; and model feature parameter screening and its online decoupling identification.

[0005] To achieve the above objectives, the present invention provides the following solution:

[0006] A method for online decoupling identification of lithium-ion battery pack model parameters includes the following steps:

[0007] Based on the basic working principle of lithium-ion batteries, a multi-physics coupling model of lithium-ion battery packs is constructed.

[0008] By combining the electrothermal relationship between different individual lithium-ion batteries, the multiphysics coupling model is extended and the parameters are identified offline.

[0009] This study investigates the relationship between offline parameters and the sensitivity of lithium-ion battery external characteristics, screens characteristic parameters of the extended multiphysics coupling model, and completes online identification of multiphysics coupling model parameters based on the least squares method and the bisection method.

[0010] Preferably, the expression for the multiphysics coupling model is:

[0011]

[0012] in, U represents the open-circuit voltage at the reference temperature. i ,i=p,n are the open-circuit potential correlation functions of the positive and negative terminals, y surf x represents the average surface lithium intercalation amount of particles of various sizes. surf Characterized by the overall lithium intercalation amount on the electrode surface, T is the battery temperature, dE ocv / dT is the entropy coefficient, E ocv T is the open-circuit voltage. ref This is a reference temperature.

[0013] Preferred parameters for offline identification include: battery equivalent specific heat capacity C. p , Battery three-dimensional thermal resistance R condx R condy R condz Entropy coefficient;

[0014] Among them, the battery's equivalent specific heat capacity C p It is obtained using a weighted average calculation method;

[0015] The three-dimensional thermal resistance R of the battery condx R condy R condz The identification method is as follows: set an initial value for the entropy coefficient, calculate the internal heat generation of the battery through the current; obtain the thermal resistance across the entire SOC range based on the shell temperature data, and obtain the thermal resistance-SOC mapping relationship based on polynomial fitting.

[0016] The method for identifying the entropy coefficient is as follows: combining the excitation response relationship between current and shell temperature, the entropy coefficient curve of the entire SOC range is obtained by the least squares method, and the entropy coefficient-SOC mapping relationship is obtained based on polynomial fitting.

[0017] Preferred methods for online parameter identification include:

[0018] Remove parameters whose sensitivity is less than the preset value;

[0019] Select parameters with high SOC sensitivity greater than low SOC sensitivity and select parameters with low SOC sensitivity greater than high SOC sensitivity to complete the online identification of differential parameters.

[0020] The identified differential parameters are filtered based on sensitivity to obtain differential parameters that meet the sensitivity requirements;

[0021] Based on the differential parameters that meet the sensitivity requirements, other feature parameters are selected when the online identification result is negative, violates the physical meaning of the parameter, or the external characteristic correction result is not ideal.

[0022] The present invention also provides an online decoupling identification system for lithium-ion battery pack model parameters, comprising: a construction module, an offline parameter identification module, and an online parameter identification module;

[0023] The building module is used to construct a multi-physics coupling model of a lithium-ion battery pack based on the basic working principle of lithium-ion batteries.

[0024] The offline parameter identification module is used to combine the electrothermal relationship between different individual lithium-ion batteries to complete the multiphysics coupling model extension and perform offline parameter identification.

[0025] The online parameter identification module is used to study the relationship between offline parameters and the sensitivity of lithium-ion battery external characteristics, screen the feature parameters of the extended multi-physics coupling model, and complete the online identification of multi-physics coupling model parameters based on the least squares method and the bisection method.

[0026] Preferably, the expression for the multiphysics coupling model is:

[0027]

[0028] in, U represents the open-circuit voltage at the reference temperature. i ,i=p,n are the open-circuit potential correlation functions of the positive and negative terminals, y surf x represents the average surface lithium intercalation amount of particles of various sizes. surf Characterized by the overall lithium intercalation amount on the electrode surface, T is the battery temperature, dE ocv / dT is the entropy coefficient, E ocv T is the open-circuit voltage. ref This is a reference temperature.

[0029] Preferred parameters for offline identification include: battery equivalent specific heat capacity C. p , Battery three-dimensional thermal resistance R condx R condy R condz Entropy coefficient;

[0030] Among them, the battery's equivalent specific heat capacity C p Obtained using the weighted average calculation process;

[0031] The three-dimensional thermal resistance R of the battery condx R condy Rcondz The identification process is as follows: set an initial value for the entropy coefficient, calculate the internal heat generation of the battery through the current; obtain the thermal resistance across the entire SOC range based on the shell temperature data, and obtain the thermal resistance-SOC mapping relationship based on polynomial fitting.

[0032] The process of identifying the entropy coefficient is as follows: combining the excitation response relationship between current and shell temperature, the entropy coefficient curve of the entire SOC range is obtained by the least squares method, and the entropy coefficient-SOC mapping relationship is obtained based on polynomial fitting.

[0033] Preferably, the online parameter identification process includes:

[0034] Remove parameters whose sensitivity is less than the preset value;

[0035] Select parameters with high SOC sensitivity greater than low SOC sensitivity and select parameters with low SOC sensitivity greater than high SOC sensitivity to complete the online identification of differential parameters.

[0036] The identified differential parameters are filtered based on sensitivity to obtain differential parameters that meet the sensitivity requirements;

[0037] Based on the differential parameters that meet the sensitivity requirements, other feature parameters are selected when the online identification result is negative, violates the physical meaning of the parameter, or the external characteristic correction result is not ideal.

[0038] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0039] This invention targets individual cells, combining size effect and stress description, and based on the basic working principle of the battery, establishes the correlation between internal processes and external characteristics. It constructs a multi-physics coupling model by combining the battery's heat generation and dissipation descriptions. By considering the electrothermal relationship between different cells, it expands the battery pack model and designs intermittent charge-discharge identification conditions. It proposes a parameter acquisition method based on least squares and optimization algorithms. Furthermore, it studies the sensitivity relationship between parameters and battery external characteristics, filters characteristic parameters of the battery pack model, and completes online identification of model parameters based on least squares and bisection methods. Attached Figure Description

[0040] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 This is an overall framework diagram of an online intelligent education system according to an embodiment of the present invention;

[0042] Figure 2This is a schematic diagram of the offline parameter identification process in an embodiment of the present invention;

[0043] Figure 3 This is a wiring diagram of the battery pack in an embodiment of the present invention;

[0044] Figure 4 This is a structural diagram of the battery pack in an embodiment of the present invention;

[0045] Figure 5 This is a flowchart illustrating the autonomous updating of electrochemical parameters in an embodiment of the present invention.

[0046] Figure 6 This is a flowchart of the bisection method for updating thermal resistance in an embodiment of the present invention;

[0047] Figure 7 This is a schematic diagram comparing the external characteristics of the battery pack under constant current conditions at 0.75C in an embodiment of the present invention. Among them, (a) is a schematic diagram of the voltage comparison curve at the end of series unit 1, (b) is a schematic diagram of the voltage comparison curve at the end of series unit 2, (c) is a schematic diagram of the voltage comparison curve at the end of the battery pack, (d) is a schematic diagram of the temperature comparison curve of the casing of cell 1, (e) is a schematic diagram of the temperature comparison curve of the casing of cell 2, (f) is a schematic diagram of the temperature comparison curve of the casing of cell 3, and (g) is a schematic diagram of the temperature comparison curve of the casing of cell 4.

[0048] Figure 8 This is a schematic diagram of the DST operating current in an embodiment of the present invention;

[0049] Figure 9 This is a schematic diagram comparing the external characteristics of the battery pack under DST operating conditions in an embodiment of the present invention, wherein (a) is a schematic diagram of the voltage comparison curve at the end of series unit 1, (b) is a schematic diagram of the voltage comparison curve at the end of series unit 2, (c) is a schematic diagram of the voltage comparison curve at the end of the battery pack, (d) is a schematic diagram of the temperature comparison curve of the casing of cell 1, (e) is a schematic diagram of the temperature comparison curve of the casing of cell 2, (f) is a schematic diagram of the temperature comparison curve of the casing of cell 3, and (g) is a schematic diagram of the temperature comparison curve of the casing of cell 4. Detailed Implementation

[0050] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0052] Example 1

[0053] like Figure 1 As shown, an online decoupling identification method for lithium-ion battery pack model parameters includes the following steps:

[0054] Based on the basic working principle of lithium-ion batteries, a multi-physics coupling model of lithium-ion battery packs is constructed.

[0055] By combining the electrothermal relationship between different individual lithium-ion batteries, the multiphysics coupling model is extended and the parameters are identified offline.

[0056] This study investigates the relationship between offline parameters and the sensitivity of lithium-ion battery external characteristics, screens characteristic parameters of the extended multiphysics coupling model, and completes online identification of multiphysics coupling model parameters based on the least squares method and the bisection method.

[0057] In this embodiment, the lithium-ion battery consists of a positive electrode, a negative electrode, and a separator, with the positive and negative electrodes separated by the separator. During the charging and discharging process, lithium ions are inserted into and extracted from the positive and negative electrodes. The relevant mathematical description of the battery's multiphysics coupling model behavior is as follows:

[0058] (1) Basic Working Process. The basic working process of a lithium-ion battery is the insertion and extraction of lithium ions within the spherical particles of the positive and negative electrodes. Based on the assumption of multiple particle sizes in the battery electrodes, it is assumed that the electrodes consist of ten particles of different sizes, and that they conform to a normal distribution. Referring to the single-particle model particle size r, the model particle size distribution is set as follows: r1 = 0.2r, r2 = 0.6r, r3 = 0.6r, r4 = r, r5 = r, r6 = r, r7 = r, r8 = 1.4r, r9 = 1.4r, r 10 =1.8r. For a single-particle-size model, its specific surface area a can be directly calculated using the following formula:

[0059]

[0060] Where, ε s,i ε i ε f,i These represent the particle volume fraction, porosity, and effective porosity of the negative and positive electrode active materials, respectively.

[0061] Since the dimension of the particle specific surface area is m -1 The relationship between the specific surface area of ​​particles with different sizes can be obtained:

[0062]

[0063] Assuming that the surface reactive ion current density is equal for particles of different sizes, and relating the working current to the reactive ion current density based on charge conservation, the average specific surface area of ​​the multi-particle-size particles is used in the calculation to obtain the surface reactive ion current density j of the positive and negative electrodes.

[0064]

[0065] Where I is the current; F is the Faraday constant; l i , i = p, n are the thicknesses of the positive and negative electrode plates; A is the effective area of ​​the electrode.

[0066] Assuming lithium ions are uniformly distributed throughout the electrode particles, and the lithium ion concentration at each location is the average value, the average lithium intercalation amount x of the positive and negative electrodes during charging and discharging is... avg,k y avg,k The calculation is as follows:

[0067]

[0068] Among them, y 0，k x represents the initial lithium intercalation amount at the cathode for particles of various sizes; 0，k c represents the initial lithium intercalation amount at the negative electrode for particles of various sizes; max,s,i The maximum concentration of lithium ions in the positive and negative electrode materials is assumed to be equal, assuming that the concentration of lithium ions in particles of different radii is equal.

[0069] Define Q p Q n The positive and negative capacitances are expressed as follows:

[0070] Q i =c max,s,i ε s,i Fl i A,i=n,p (5)

[0071] By rearranging equations (3), (4), and (5), we can obtain the equations characterizing the basic working process of the battery, as shown below:

[0072]

[0073] Assume the initial lithium intercalation amount of particles of different sizes in the positive and negative electrodes is related as follows:

[0074]

[0075] (2) Solid-phase diffusion process. The solid-phase diffusion process in lithium-ion batteries is the movement of lithium ions from the center to the surface of the spherical particles in the positive and negative electrodes, driven by a concentration gradient. During this process, uneven concentration distribution at different locations leads to uneven deformation of the active material, resulting in diffusion-induced stress at those locations. This diffusion-induced stress also affects the insertion and extraction of lithium ions. Therefore, the solid-phase diffusion process is simultaneously affected by both the concentration gradient and stress. The solid-phase lithium ion concentrations c of the positive and negative electrodes are... s,i (t,r), i=n, p distribution at particle radius r conforms to Fick's second law considering diffusion-induced stress.

[0076]

[0077]

[0078] Where t is time; D s,i σ is the solid-phase diffusion coefficient of the positive and negative electrodes; R is the ideal gas constant; Ω is the partial molar volume; σ h For hydrostatic stress, the calculation is as follows:

[0079]

[0080] Among them, E i The Young's modulus of the positive and negative electrode active materials; υ i Poisson's ratio represents the positive and negative poles.

[0081] Using a simplified method based on the two-parameter parabolic equivalence, the lithium intercalation amount y on the positive and negative electrode surfaces of the particles can be obtained. surf,k x surf,k Calculation equation:

[0082]

[0083]

[0084] To avoid model errors caused by difficulties in radius identification, the radius is combined with the solid-phase diffusion coefficients of the positive and negative electrodes to obtain the solid-phase diffusion time constant τ of the positive and negative electrodes, which is proportional to the square of the radius. i ,i=p,n.

[0085]

[0086] By rearranging equations (3), (11), and (13), we can obtain

[0087]

[0088] The average value of surface lithium intercalation of particles of various sizes, y surf x surf The overall lithium intercalation amount on the electrode surface is characterized, providing input for subsequent calculations of open-circuit potential and reaction polarization overpotential. The open-circuit voltage at the reference temperature is calculated as follows, where U i ,i=p,n are the open-circuit potential correlation functions of the positive and negative terminals.

[0089]

[0090] (3) Concentration polarization. Along the l-direction of the positive and negative electrodes of the battery, there is a gradient in the concentration distribution of liquid lithium ions, leading to a concentration polarization overpotential η. con The equation for the generation of is shown below:

[0091]

[0092] Where t+ is the cation transport number; T is the battery temperature; Δc1 and Δc2 are the changes in the liquid phase lithium-ion concentration at the positive and negative electrode current collectors, and their calculation equations are shown below:

[0093]

[0094] Among them, P conp P conn The positive and negative electrode liquid-phase diffusion ratio coefficients reflect the degree of concentration polarization; τ e This is the diffusion time constant, which reflects the time it takes for the concentration gradient to establish.

[0095] (4) Reaction polarization. To maintain a certain electrochemical reaction rate, a reaction polarization overpotential η is required. act The calculation formula is as follows:

[0096]

[0097]

[0098]

[0099] Where c0 is the initial lithium-ion concentration in the liquid phase; k i , i = n, p is the positive and negative electrode reaction rate constant, which is an inherent parameter for the electrochemical reaction of active particles, and the value is the same for particles of different sizes.

[0100] By rearranging equations (3), (5), (19), and (20), we can obtain:

[0101]

[0102]

[0103]

[0104] To avoid model errors caused by difficulties in radius identification, the radius is combined with the positive and negative electrode reaction rate constants to obtain the reaction polarization coefficient P. act The calculation formula is as follows:

[0105]

[0106] (5) Ohmic polarization. The polarization overpotential caused by the solid electrolyte interphase (SEI) film on the surface of the active particles of the positive and negative electrodes, the positive and negative electrode materials, and the electrolyte impedance is the ohmic polarization overpotential η. ohm The calculation formula is:

[0107] η ohm =R ohm I (25)

[0108] Among them, R ohm It is the internal resistance of the Ohm.

[0109] Battery terminal voltage U app The following formula can be used to calculate:

[0110]

[0111] (6) Battery heat generation. During battery charging and discharging, the main sources of heat generation are: heat generated by internal reactions within the battery and heat generated due to internal resistance. Ignoring other forms of heat generation, the heat generation rate can be obtained based on Bernardi's heat generation formula. That is, the heat generated per unit time:

[0112]

[0113] Among them, Q r It is a reversible heat; Q ir Irreversible heat; dE ocv / dT is the entropy coefficient.

[0114] (7) Battery temperature conduction. Battery temperature conduction conforms to the following heat balance equation:

[0115]

[0116] Where ρ is the material density; C is the battery heat capacity; and λ is the thermal conductivity. This is the net heat production rate.

[0117] To simplify the process of identifying thermal model parameters, the battery thermal conduction phenomenon is ignored. The average temperature of the electrode winding is used to replace the temperature at different locations of the battery. Therefore, equation (28) can be simplified to:

[0118]

[0119] Based on this, thermal resistance is defined as the ratio of temperature difference to heat, to characterize the amount of heat transferred. A thermal resistance model is constructed to describe the heat exchange process between the battery winding and the casing, and the heat exchange process between the battery casing and the environment.

[0120] Discretizing the internal temperature balance equation of the battery yields the iterative equation for the change of internal temperature over time, as shown in the following equation:

[0121]

[0122] Where k is the time iteration coefficient; m rollΔt represents the mass of the wound body; Δt represents the time step.

[0123] Discretizing the battery casing temperature balance equation yields the following iterative equation for the battery casing temperature over time:

[0124]

[0125] Where, m can For the quality of the outer casing.

[0126] (8) Multiphysics coupling. The internal temperature of the battery affects the reaction process. Several related parameters of the electrochemical model change with the internal temperature, and it is assumed that the changes in these parameters follow the Arrhenius equation. In addition, temperature also directly affects η. con η act η con The stress description, which is itself a function of temperature, is finally corrected by the Nernst equation for E. ocv As shown in equation (36).

[0127]

[0128] In this embodiment, the stress behavior in the solid-phase diffusion of each individual cell is considered to be independent, and the stress between different individual cells caused by cell deformation is ignored.

[0129] Assuming that the internal electrochemical behavior of each individual cell in the battery pack is independent, and constructing the electrochemical relationships between different cells based on the external characteristics of each cell. Ignoring the influence of wire resistance, and based on Kirchhoff's laws, it is considered that the voltage and current of the battery pack's main circuit are linear superpositions of the voltages and currents of each individual cell.

[0130] As the principle of single-cell model construction shows, the external characteristics of a battery are determined by parameters, and the parameter sets of different single cells differ. Different model parameter sets are used for each battery to characterize the differences between them. Since the terminal voltages of all single cells in a parallel battery pack are consistent, to reduce the number of parameter sets and simplify the parameter identification process, the differences between single cells in the parallel battery pack can be ignored. It can be assumed that the parameter sets of each single cell are identical, and that the main circuit current is evenly distributed among the branch currents.

[0131] The battery pack thermal model is also constructed using a thermal resistance network. It assumes that the heat generation and conduction behaviors of each cell in the battery pack are independent, and establishes the relationship between the thermal behaviors of different cells through heat exchange between the gaps. Since the gap distance and gap area between different cells are different, the thermal behavior of the battery pack is closely related to the packing method. Square batteries are mostly packed in the direction of the large end face.

[0132] It is assumed that the thermal behavior of each cell in the battery pack in the two directions outside the pack direction is consistent with the thermal behavior of the individual cells, and only the heat exchange of the cells in the pack direction is considered.

[0133] During heat exchange between the casing and the environment, with a constant heat exchange area, the heat dissipation mainly depends on the heat transfer coefficient. The heat transfer coefficient within the gap decreases as the gap decreases. If the individual cells are closely arranged, the heat transfer coefficient between adjacent cells is considered to be small. Due to differences between individual cells, the casing temperatures obtained from different individual cell thermal models also differ. Therefore, after grouping, during the iterative calculation of the individual cell casing temperature, the average casing temperature of two adjacent cells is taken as the casing temperature at both ends of the corresponding gap.

[0134] The multiphysics coupling relationship of each cell in the battery pack is consistent with the single cell model representation method. Combined with the above inter-cell heat exchange model, the multiphysics coupling model of the battery pack is completed.

[0135] Battery geometric parameters, electrochemical parameters, and stress-related parameters during solid-phase diffusion can be obtained by consulting relevant literature and battery manufacturers; no identification is required. To simplify the parameter identification process, the parameter identification procedure for a single-particle-size model with radius r is adopted, and the relevant parameters for the multi-particle-size model are obtained by utilizing the functional relationship between parameters and particle size in the model. The parameters to be identified and their definitions are shown in Table 1, the model parameter table.

[0136]

[0137]

[0138] Under the same current excitation, the response times of different processes in the battery reaction vary, thus enabling the design of a parameter identification method. From the established mathematical descriptions of each process, it is clear that reactive polarization and ohmic polarization generate polarization overpotentials upon current excitation, while solid-state diffusion, being a state change, requires time to establish its state variables, exhibiting a significant time difference compared to polarization. Furthermore, facing the same current excitation, the intensity of the response varies among different processes. Based on this difference, the various components in the terminal voltage calculation can be separated and extracted, and further, the parameters can be gradually obtained through the relationship between open-circuit voltage, overpotentials of each component, and corresponding parameters.

[0139] To identify model parameters at the reference temperature, based on the fundamental principles of excitation-response analysis and the requirements for parameter decoupling identification, a parameter identification process was designed: each constant current discharge segment was set to 20 min, the constant current charging time to 10 min, and the rest time to 15 min. The charge / discharge current rate range was 0.16–0.5 C, to reduce the impact of internal battery temperature changes on the electrochemical parameter identification results. The specific parameter identification process at the reference temperature is as follows: Figure 2As shown.

[0140] The specific methods for identifying thermal model parameters are as follows:

[0141] (1) Battery equivalent specific heat capacity C p The weighted average method is usually used to obtain it.

[0142]

[0143] Where C p,i m is the specific heat capacity of the corresponding component material in the battery. i This refers to the mass of the materials used in the corresponding battery components.

[0144] (2) The three-dimensional thermal resistance R of the battery condx R condy R condz It is a physical quantity that characterizes how much heat is transferred from the inside of the battery to the outer casing. The specific identification process is as follows: set an initial value for the entropy coefficient, calculate the heat generated inside the battery through the current; based on the outer casing temperature data, use the following formula to iterate and obtain the thermal resistance over the entire SOC range, and obtain the thermal resistance-SOC mapping relationship based on polynomial fitting.

[0145]

[0146] (3) Based on the thermal resistance identification results, a data-driven method is used to identify the entropy coefficient and establish the mathematical relationship between the entropy coefficient and SOC. The specific identification method is as follows: Combining the excitation response relationship between current and shell temperature, the entropy coefficient curve of the entire SOC range is obtained by the least squares method, and the entropy coefficient-SOC mapping relationship is obtained based on polynomial fitting.

[0147] In this embodiment, to simplify the online identification process, a sensitivity analysis of the model parameters needs to be performed before designing the online identification method for battery pack model parameters to determine the degree of influence of the parameters on the battery terminal voltage, thereby determining the specific parameters to be identified.

[0148] As the model construction principle states, model parameters characterize different internal behaviors of the battery and have varying impacts on the battery's external characteristics across different SOC ranges. Therefore, by conducting parameter sensitivity analysis across different SOC ranges, this study investigates the influence of model parameters on the battery's external characteristics and explores their relationship with battery SOC. The sensitivity value is characterized by the maximum relative average error, and its calculation equation is as follows:

[0149]

[0150] Among them, U app The model simulation terminal voltage curve is based on specific parameters; U app,i, i = 50%, 60%, ..., 150% are the model simulation terminal voltage curves when specific parameters change relative to the reference value.

[0151] The steps for screening feature parameters using sensitivity analysis are as follows:

[0152] (1) Remove parameters with low sensitivity;

[0153] (2) In order to prevent the selected feature parameters from affecting each other's external characteristics due to different identification order, parameters with high SOC sensitivity greater than low SOC sensitivity and parameters with low SOC sensitivity greater than high SOC sensitivity are selected to achieve online identification of differential parameters.

[0154] (3) The parameters selected in (2) are screened based on sensitivity: if the sensitivity is too small, it is easy to fail to achieve external characteristic correction; if the sensitivity is too large, it is easy to cause the external characteristic correction result to deviate from the external characteristic experimental result by a large margin repeatedly.

[0155] (4) For the parameters in (3), filter out other feature parameters whose online identification results are negative, violate the physical meaning of the parameters, and whose external characteristic correction results are not ideal.

[0156] Thermal parameters include entropy coefficient and thermal resistance coefficient. According to the thermal model, the entropy coefficient affects the casing temperature by influencing the heat generation inside the battery, while the thermal resistance coefficient can independently affect the casing temperature in any dimension. To make full use of the measured casing temperature data of the battery entity and simplify the identification process, only the thermal resistance coefficient is used as a characteristic parameter of the thermal model for online identification.

[0157] During the online identification of battery pack model parameters, the actual total discharge capacity of the battery under different operating conditions differs from the model's calibrated total capacity. This leads to a decrease in SOC estimation accuracy under complex operating conditions. Therefore, using SOC-segmented data for parameter differentiation updates is unreliable; instead, terminal voltage-segmented data is chosen. A 100-second online parameter identification time interval is used, employing the least squares method and the bisection method to achieve online identification of the battery pack's multiphysics model parameters.

[0158] Example 2

[0159] The present invention also provides an online decoupling identification system for lithium-ion battery pack model parameters, comprising: a construction module, an offline parameter identification module, and an online parameter identification module;

[0160] The building module is used to construct a multi-physics coupling model of a lithium-ion battery pack based on the basic working principle of lithium-ion batteries.

[0161] The offline parameter identification module is used to combine the electrothermal relationship between different individual lithium-ion batteries to complete the multiphysics coupling model extension and perform offline parameter identification.

[0162] The online parameter identification module is used to study the relationship between offline parameters and the sensitivity of lithium-ion battery external characteristics, screen the feature parameters of the extended multi-physics coupling model, and complete the online identification of multi-physics coupling model parameters based on the least squares method and the bisection method.

[0163] In this embodiment, the expression for the multiphysics coupling model is:

[0164]

[0165] in, U represents the open-circuit voltage at the reference temperature. i ,i=p,n are the open-circuit potential correlation functions of the positive and negative terminals, y surf x represents the average surface lithium intercalation amount of particles of various sizes. surf Characterized by the overall lithium intercalation amount on the electrode surface, T is the battery temperature, dE ocv / dT is the entropy coefficient, E ocv T is the open-circuit voltage. ref This is a reference temperature.

[0166] In this embodiment, the parameters identified offline include: the battery's equivalent specific heat capacity C. p , Battery three-dimensional thermal resistance R condx R condy R condz Entropy coefficient;

[0167] Among them, the battery's equivalent specific heat capacity C p Obtained using the weighted average calculation process;

[0168] The three-dimensional thermal resistance R of the battery condx R condy R condz The identification process is as follows: set an initial value for the entropy coefficient, calculate the internal heat generation of the battery through the current; obtain the thermal resistance across the entire SOC range based on the shell temperature data, and obtain the thermal resistance-SOC mapping relationship based on polynomial fitting.

[0169] The process of identifying the entropy coefficient is as follows: combining the excitation response relationship between current and shell temperature, the entropy coefficient curve of the entire SOC range is obtained by the least squares method, and the entropy coefficient-SOC mapping relationship is obtained based on polynomial fitting.

[0170] In this embodiment, the online parameter identification process includes:

[0171] Remove parameters whose sensitivity is less than the preset value;

[0172] Select parameters with high SOC sensitivity greater than low SOC sensitivity and select parameters with low SOC sensitivity greater than high SOC sensitivity to complete the online identification of differential parameters.

[0173] The identified differential parameters are filtered based on sensitivity to obtain differential parameters that meet the sensitivity requirements;

[0174] Based on the differential parameters that meet the sensitivity requirements, other feature parameters are selected when the online identification result is negative, violates the physical meaning of the parameter, or the external characteristic correction result is not ideal.

[0175] Example 3

[0176] To illustrate the method used in this invention, the construction of a multiphysics coupling model of a two-parallel, two-series (2P2S) 8Ah lithium iron phosphate square battery pack and its online parameter identification process are used as an implementation example. In this embodiment, the battery pack connection method and battery pack structure are as follows: Figure 3 , 4 As shown.

[0177] Table 2 shows the sensitivity values ​​for the model parameters. Parameter selection is as follows:

[0178] (1) With other parameters remaining constant, τ e , τ p It has little correlation with the external characteristics of the battery, so it is not involved in the online identification process.

[0179] (2) x0 and y0 are only related to the external characteristics of the battery under full charge, and are not related to the high and low SOC regions. act Sensitivity decreases as SOC decreases; Q p Q n τ n Sensitivity increases as SOC decreases.

[0180] (3)τ n Q p Q n The affected regions overlap, both falling within the low SOC range, but Q... n and Q p The sensitivity is too high. Therefore, choose P. act and τ n These are characteristic parameters.

[0181] (4) Updating parameter P using the least squares fitting method act τ n At that time, the update result may contain negative values, violating the physical meaning of the parameter. Therefore, when obtaining P... act τ n Next, we need to determine whether it is positive or negative. If the parameter is less than 0, then while maintaining the result of the previous update, we update x0, y0, and Q.n According to the model construction principle, τ n and Q n Coupling affects the external characteristic curve, therefore when τ n When the online identification results and external characteristic correction results are not ideal, online Q identification is required. n However, due to the high sensitivity of this parameter, in order to prevent the simulation curve from deviating from the experimental curve by a large margin due to the high frequency of correction of this parameter, it is assumed that it will only be corrected once during the entire simulation process.

[0182] Table 2

[0183]

[0184]

[0185] The specific online identification process for electrical parameters in this implementation is as follows: Figure 5 , 6 As shown.

[0186] The simulation curves of the battery pack at an ambient temperature of 25℃, a discharge rate of 0.75C, and a casing temperature are shown in the following figures. Figure 7 As shown in (a), (b), (c), (d), (e), (f), and (g).

[0187] To verify the adaptability of the online model parameter identification method in this implementation to dynamic operating conditions, the simulation and measured results of the battery pack model under DST conditions were compared and verified. Figure 8 The current excitation for this operating condition is given. Figure 9 (a), (b), (c), (d), (e), (f), and (g) are the simulation curves comparing the terminal voltage and shell temperature of the model.

[0188] Simulation results show that, for square lithium iron phosphate battery packs, the multiphysics coupling model implemented in this paper can achieve accurate simulation of battery terminal voltage and three-dimensional casing temperature, as well as online identification of model parameters.

[0189] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for online decoupling and identification of parameters in a lithium-ion battery pack model, characterized in that, Includes the following steps: Based on the basic working principle of lithium-ion batteries, a multi-physics coupling model of lithium-ion battery packs is constructed. By combining the electrothermal relationship between different individual lithium-ion batteries, the multiphysics coupling model is extended and the parameters are identified offline. The study investigates the relationship between offline parameters and the sensitivity of external characteristics of lithium-ion batteries, screens the characteristic parameters of the extended multi-physics coupling model, and completes the online identification of multi-physics coupling model parameters based on the least squares method and the bisection method. The expression for the multiphysics coupling model is: ， in, Represents the open-circuit voltage at the reference temperature. This is a function relating the open-circuit potential of the positive and negative terminals. This represents the average surface lithium intercalation amount for particles of various sizes. Characterizes the total amount of lithium intercalated on the electrode surface. T For battery temperature, The entropy coefficient, Open circuit voltage, For reference temperature; Online parameter identification methods include: Remove parameters whose sensitivity is less than the preset value; Select parameters with high SOC sensitivity greater than low SOC sensitivity and select parameters with low SOC sensitivity greater than high SOC sensitivity to complete the online identification of differential parameters. The identified differential parameters are filtered based on sensitivity to obtain differential parameters that meet the sensitivity requirements; Based on the differential parameters that meet the sensitivity requirements, other feature parameters are selected when the online identification result is negative, violates the physical meaning of the parameter, or the external characteristic correction result is not ideal.

2. The online decoupling identification method for lithium-ion battery pack model parameters according to claim 1, characterized in that, The parameters identified offline include: battery equivalent specific heat capacity. , battery three-dimensional thermal resistance , , Entropy coefficient; Among them, the battery's equivalent specific heat capacity It is obtained using a weighted average calculation method; Three-dimensional thermal resistance of battery , , The identification method is as follows: set an initial value for the entropy coefficient, calculate the internal heat generation of the battery through the current; obtain the thermal resistance across the entire SOC range based on the shell temperature data, and obtain the thermal resistance-SOC mapping relationship based on polynomial fitting. The method for identifying the entropy coefficient is as follows: combining the excitation response relationship between current and shell temperature, the entropy coefficient curve of the entire SOC range is obtained by the least squares method, and the entropy coefficient-SOC mapping relationship is obtained based on polynomial fitting.

3. A lithium-ion battery pack model parameter online decoupling identification system, characterized in that, include: The module consists of a construction module, an offline parameter identification module, and an online parameter identification module; The building module is used to construct a multi-physics coupling model of a lithium-ion battery pack based on the basic working principle of lithium-ion batteries. The offline parameter identification module is used to combine the electrothermal relationship between different individual lithium-ion batteries to complete the multiphysics coupling model extension and perform offline parameter identification. The online parameter identification module is used to study the relationship between offline parameters and the sensitivity of external characteristics of lithium-ion batteries, screen the feature parameters of the extended multi-physics coupling model, and complete the online identification of multi-physics coupling model parameters based on the least squares method and the bisection method. The expression for the multiphysics coupling model is: ， in, Represents the open-circuit voltage at the reference temperature. This is a function relating the open-circuit potential of the positive and negative terminals. This represents the average surface lithium intercalation amount for particles of various sizes. Characterizes the total amount of lithium intercalated on the electrode surface. T For battery temperature, The entropy coefficient, Open circuit voltage, For reference temperature; The online parameter identification process includes: Remove parameters whose sensitivity is less than the preset value; Select parameters with high SOC sensitivity greater than low SOC sensitivity and select parameters with low SOC sensitivity greater than high SOC sensitivity to complete the online identification of differential parameters. The identified differential parameters are filtered based on sensitivity to obtain differential parameters that meet the sensitivity requirements; Based on the differential parameters that meet the sensitivity requirements, other feature parameters are selected when the online identification result is negative, violates the physical meaning of the parameter, or the external characteristic correction result is not ideal.

4. The online decoupling identification system for lithium-ion battery pack model parameters according to claim 3, characterized in that, The parameters identified offline include: battery equivalent specific heat capacity. , battery three-dimensional thermal resistance , , Entropy coefficient; Among them, the battery's equivalent specific heat capacity Obtained using the weighted average calculation process; Three-dimensional thermal resistance of battery , , The identification process is as follows: set an initial value for the entropy coefficient, calculate the internal heat generation of the battery through the current; obtain the thermal resistance across the entire SOC range based on the shell temperature data, and obtain the thermal resistance-SOC mapping relationship based on polynomial fitting. The process of identifying the entropy coefficient is as follows: combining the excitation response relationship between current and shell temperature, the entropy coefficient curve of the entire SOC range is obtained by the least squares method, and the entropy coefficient-SOC mapping relationship is obtained based on polynomial fitting.

Images