A lithium-ion battery state-of-charge estimation method, a battery management system and a storage medium

Abstract

This invention discloses a method for estimating the state of charge (SOC) of a lithium-ion battery, a battery management system, and a storage medium, belonging to the technical field of battery management systems. The invention first conducts pulse discharge experiments on the battery to collect charge and discharge data; then, it establishes a third-order RC equivalent circuit model, fully considering ohmic internal resistance, electrochemical polarization, and concentration gradient polarization; next, it uses an improved I-LM algorithm to optimize and fit the model parameters, where the I-LM algorithm adaptively updates the damping coefficient through a multi-factor collaborative decision-making mechanism; finally, it estimates the battery SOC in real time based on the fitted parameters and the ampere-hour integration method. Compared with existing technologies, this invention not only improves the prediction accuracy of the battery model but also significantly accelerates the convergence speed of parameter extraction, enhances adaptability to complex dynamic operating conditions, and has high computational efficiency, making it suitable for application in resource-constrained embedded systems. It provides a more accurate and efficient SOC estimation solution for lithium-ion battery management systems.

Description

Technical Field

[0001] This invention relates to the field of battery management technology, and in particular to a method for estimating the state of charge (SOC) of lithium-ion batteries based on a third-order RC model using an improved Levenberg-Marquardt (I-LM) algorithm, and a battery management system. Background Technology

[0002] State of Charge (SOC) refers to the ratio of a battery's current remaining charge to its rated capacity, and is a core parameter for measuring a battery's state of use. Accurate SOC estimation is crucial for battery charge and discharge control, energy management, fault diagnosis, and extending battery life. However, due to the complexity and nonlinear characteristics of the internal electrochemical reactions in lithium-ion batteries, accurate SOC estimation faces many challenges.

[0003] In a battery management system (BMS), the equivalent circuit model is fundamental for achieving state of charge (SOC) estimation. Traditional battery equivalent circuit models mainly include the first-order RC model (1RC) and the second-order RC model (2RC). The 1RC model only considers the battery's ohmic internal resistance and one RC circuit, resulting in a simple structure but lower accuracy. The 2RC model adds an electrochemical polarization circuit to the 1RC model, improving accuracy to some extent, but it still cannot fundamentally solve the error problem, especially under dynamic operating conditions, where the model error will increase significantly.

[0004] Regarding SOC estimation methods, existing technologies mainly employ the following approaches:

[0005] Open-circuit voltage method: This method estimates the state of charge (SOC) by measuring the open-circuit voltage of the battery after it has been left to rest, based on the OCV-SOC mapping relationship. This method is simple and direct, but it requires the battery to be left to rest for a long time (usually 2-3 hours) to reach a stable state, which cannot meet the needs of real-time SOC estimation.

[0006] The ampere-hour integration method obtains the change in SOC by integrating the current in real time, and calculates the current SOC by combining it with the initial SOC value. This method is simple to implement, but it requires high accuracy of the initial SOC, and the current measurement error accumulates continuously during the integration process, leading to a decrease in long-term estimation accuracy.

[0007] Neural network method: This method utilizes the powerful nonlinear mapping capability of neural networks to establish a mapping relationship between inputs (such as voltage, current, temperature, etc.) and the SOC (System-on-Chip). This method offers high accuracy but requires a large amount of training data, resulting in high computational complexity and making it unsuitable for resource-constrained embedded systems.

[0008] Extended Kalman Filter (EKF) algorithm: This method linearizes the battery model and estimates the state of charge (SOC) recursively. It effectively suppresses noise, but requires calculating the Jacobian matrix, suffers from linearization errors, and is highly dependent on model accuracy.

[0009] Unscented Kalman Filter (UKF) algorithm: By avoiding the calculation of the Jacobian matrix through unscented transformation, it improves the estimation accuracy of nonlinear systems. However, the standard UKF does not consider the influence of unknown interference noise, which limits the estimation accuracy under complex operating conditions.

[0010] To address the aforementioned issues, recent studies have proposed SOC estimation methods based on a third-order RC equivalent circuit model. The third-order RC model, building upon the 2RC model, adds a concentration difference polarization stage, providing a more comprehensive reflection of the internal electrochemical characteristics of lithium-ion batteries. Combined with the third-order RC model, the Adaptive Unscented Kalman Filter (AUKF) algorithm improves the robustness of SOC estimation by adjusting the noise covariance matrix online.

[0011] However, existing technologies still have the following shortcomings:

[0012] Slow convergence speed: The AUKF algorithm takes a long time to reach a stable state in the initial stage, especially when the initial SOC error is large, the convergence process is significantly prolonged.

[0013] The accuracy of parameter estimation needs to be improved: Although the third-order RC model is theoretically more accurate, existing parameter identification methods (such as the traditional Levenberg-Marquardt algorithm) are prone to getting trapped in local optima when facing high-dimensional parameters and strongly nonlinear problems, which limits the accuracy of the model.

[0014] Limited adaptability to noise and dynamic operating conditions: Existing algorithms show a significant decrease in estimation accuracy under complex dynamic operating conditions, such as temperature changes and high-current charging and discharging scenarios.

[0015] High computational complexity: Algorithms such as AUKF require a large number of matrix operations, which place high demands on computing resources and are not suitable for application in resource-constrained embedded BMS.

[0016] The algorithm lacks robustness: the existing methods have limited adaptability to different battery types and aging levels, and require frequent parameter recalibration.

[0017] Therefore, there is an urgent need for a SOC estimation method that can improve model accuracy, accelerate convergence speed, and reduce computational complexity, so as to meet the needs of electric vehicles, energy storage systems and other fields for high-precision, real-time battery management. Summary of the Invention

[0018] This invention addresses the shortcomings of existing technologies by providing a method for estimating the state of charge of a lithium-ion battery, a battery management system, and a storage medium.

[0019] To achieve the above-mentioned objectives, the technical solution adopted by the present invention is as follows:

[0020] A method for estimating the state of charge of a lithium-ion battery includes the following steps:

[0021] 1) Data acquisition: Conduct pulse discharge experiments on the battery to collect voltage and current data under different load conditions;

[0022] 2) Model construction: A third-order RC equivalent circuit model is used as the battery model. The model includes ohmic internal resistance R0, ohmic polarization internal resistance R1, ohmic polarization capacitance C1, electrochemical polarization internal resistance R2, electrochemical polarization capacitance C2, concentration difference polarization internal resistance R3, and concentration difference polarization capacitance C3.

[0023] 3) Parameter fitting: The parameters of the third-order RC equivalent circuit model are fitted using the improved I-LM algorithm. The I-LM algorithm adaptively adjusts the damping coefficient through a multi-factor collaborative decision-making mechanism, including: calculating the combination factor based on the residual term, gradient term and model quality term, and dynamically scaling the damping coefficient according to the combination factor to optimize the parameter fitting process.

[0024] 4) SOC estimation: Using the fitted model parameters, the real-time state of charge (SOC) of the battery is calculated by the ampere-hour integration method or the state observer.

[0025] Furthermore, in the data acquisition step, the pulse discharge experiment includes a hybrid pulse power characteristic (HPPC) test, which is repeated multiple times to obtain average data.

[0026] Furthermore, in step 2), the relationship between the open-circuit voltage Uoc and the SOC of the third-order RC equivalent circuit model is defined by the following formula:

[0027]

[0028] in , , , , These are the parameters obtained by fitting using the I-LM algorithm.

[0029] Furthermore, in step 3), the damping coefficient adjustment process of the I-LM algorithm includes:

[0030] Calculate the residual term based on the ratio of the current residual vector norm to the preset reference residual value;

[0031] Calculate the gradient term based on the ratio of the current gradient vector norm to a preset reference gradient value;

[0032] The model quality term is calculated based on the ratio of the actual decrease in the objective function to the predicted decrease.

[0033] A weighted geometric mean is calculated for the residual term, gradient term, and model quality term, and combined with the iterative decay factor to generate a combined factor.

[0034] Based on the magnitude of the combined factors, a three-level strategy is adopted to adjust the scaling factor of the damping coefficient, including aggressively increasing, moderately increasing or decreasing the damping coefficient.

[0035] Furthermore, the calculation of the residual term uses the hyperbolic tangent function to compress and map the residual proportion; the calculation of the gradient term uses the exponential decay function to map the gradient proportion; and the calculation of the model quality term uses the sigmoid function to map the ratio of the actual to the predicted decrease.

[0036] Furthermore, the weighting parameters of the weighted geometric mean include , and ,in The residual weighting coefficient is... These are the gradient weight coefficients. These are the model quality weight coefficients, and .

[0037] Furthermore, in step 3), the initial parameters of the I-LM algorithm include the initial value of the damping coefficient. Weight parameters , , Gradient accuracy threshold Reference scale and The damping coefficient is updated iteratively until the convergence condition is met.

[0038] Furthermore, in step 4), the SOC is calculated using the discretized ampere-hour integral formula:

[0039]

[0040] in For Coulomb efficiency, For the battery's rated capacity, The initial time, The sampling interval is denoted as .

[0041] This invention also discloses a battery management system configured to execute the above-described lithium-ion battery state-of-charge estimation method for real-time monitoring and control of the lithium-ion battery's state of charge, specifically including:

[0042] The data acquisition module is configured to perform pulse discharge experiments on the battery and collect voltage and current data of the battery under different load conditions.

[0043] The model building module is configured to use a third-order RC equivalent circuit model to describe battery characteristics. The model includes ohmic internal resistance, ohmic polarization internal resistance, electrochemical polarization internal resistance, concentration difference polarization internal resistance, ohmic polarization capacitance, electrochemical polarization capacitance, and concentration difference polarization capacitance.

[0044] The parameter fitting module is configured to fit the parameters of the third-order RC equivalent circuit model using an improved I-LM algorithm, wherein the I-LM algorithm adaptively adjusts the damping coefficient through a multi-factor collaborative decision-making mechanism to optimize parameter estimation.

[0045] The SOC calculation module is configured to calculate the battery's state of charge (SOC) value based on the fitted model parameters, using either the ampere-hour integration method or the model output.

[0046] The monitoring and control module is configured to achieve real-time SOC estimation, monitor battery status, and output control signals to manage the battery charging and discharging process.

[0047] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method for estimating the state of charge of a lithium-ion battery.

[0048] Compared with the prior art, the advantages of the present invention are as follows:

[0049] 1. By employing a third-order RC model based on the improved Levenberg-Marquardt (I-LM) algorithm, the internal electrochemical characteristics of lithium-ion batteries can be more accurately reflected, especially the concentration gradient polarization phenomenon. This enables the battery model to provide more accurate SOC estimates under various operating conditions.

[0050] 2. Using the I-LM algorithm for parameter fitting not only improves the accuracy of parameter estimation but also significantly accelerates the convergence speed. Compared with traditional algorithms and other optimization algorithms, such as AUKF, quantum particle swarm optimization hybrid neural networks, and genetic algorithms, the I-LM algorithm exhibits higher acceptance rate, faster convergence speed, and higher parameter accuracy in continuous optimization problems.

[0051] 3. This invention solves the problem of poor adaptability of SOC estimation to complex dynamic operating conditions in existing technologies. By adaptively adjusting the noise covariance, it effectively addresses the influence of unknown interference noise and improves the estimation accuracy in complex environments.

[0052] 4. This invention achieves higher accuracy and faster convergence speed, and can be applied in resource-constrained embedded systems, providing more effective technical support for battery management systems (BMS). Attached Figure Description

[0053] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be 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.

[0054] Figure 1 This is a schematic diagram of the third-order RC equivalent circuit model structure used in the embodiments of the present invention;

[0055] Figure 2 This is a graph showing the fitting result of the curve between the battery open-circuit voltage (OCV) and the state of charge (SOC) in an embodiment of the present invention;

[0056] Figure 3 This is a schematic diagram of the voltage change curve in the pulse discharge experiment of this invention. Detailed Implementation

[0057] 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, not all, of the embodiments of the present invention. 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.

[0058] I. Overall Implementation Framework

[0059] The core of this invention lies in achieving high-precision SOC estimation through four main stages: experimental data acquisition, battery model construction, intelligent parameter fitting, and SOC calculation verification. The implementation process follows the principles of systematization and repeatability to ensure the applicability of the method in actual battery management systems (BMS). The overall scheme is based on a third-order RC equivalent circuit model and utilizes the I-LM algorithm to optimize parameter fitting, thereby addressing the problems of slow convergence and low accuracy associated with traditional methods.

[0060] II. Data Collection

[0061] Data is fundamental to the method of this invention, and its accuracy and completeness directly affect the accuracy of subsequent models. The implementation process is as follows:

[0062] 1. Experimental Equipment Preparation: Select standard lithium-ion batteries (such as ternary lithium or lithium iron phosphate batteries), high-precision charge-discharge testing equipment (programmable load), data acquisition card (sampling frequency not less than 10Hz), and temperature control box. The test environment temperature is controlled at 25°C±2°C to reduce the impact of temperature fluctuations.

[0063] 2. Pulse Discharge Experiment: The battery's hybrid pulse power characteristic (HPPC) was tested. The specific steps are as follows:

[0064] Charge the battery to full charge (SOC=100%) and let it sit for 1 hour to allow the voltage to stabilize.

[0065] Apply a short, high-current discharge pulse (e.g., 10 seconds), followed by a resting recovery period (e.g., 40 seconds). This process is repeated at multiple SOC points (e.g., 100%, 90%, 80%, ..., 10%) to cover the full operating range of the battery.

[0066] Repeat the test 10 times to ensure data reliability.

[0067] 3. Data recording and preprocessing: Acquire the terminal voltage within each pulse cycle. and current The time series data is processed by filtering and denoising the original data (e.g., using a moving average method), and the average of 10 test data points is calculated as the final dataset for subsequent parameter fitting.

[0068] III. Battery Model Construction

[0069] This invention employs a third-order RC equivalent circuit model to accurately describe the dynamic characteristics of a battery. The structure of this model is as follows: Figure 1 As shown. The implementation process includes:

[0070] 1. Model Equation Establishment: Based on Kirchhoff's laws, the continuous-time state-space equations of the model are established, as shown in the following formula:

[0071]

[0072] in, Open circuit voltage, For ohmic internal resistance, , , These are the internal resistances for ohmic polarization, electrochemical polarization, and concentration gradient polarization, respectively. , , For the corresponding capacitor, Represents current. Indicates ohmic polarization capacitor The voltage at both ends, Electrochemically polarized capacitor The voltage at both ends, Concentration difference polarization capacitance The voltage at both ends.

[0073] 2. Model Discretization: To facilitate microprocessor implementation, the continuous equations are discretized to obtain discrete state equations:

[0074]

[0075] in, Sampling time, For noise terms, This represents the battery's rated capacity.

[0076] 3. OCV-SOC Relationship Modeling: The relationship between open-circuit voltage and SOC is fitted using a composite model, with the following formula:

[0077]

[0078] parameter to Extracted from experimental data using the I-LM algorithm. The fitted OCV-SOC curve is shown below. Figure 2 As shown, this curve forms the basis for parameter fitting.

[0079] IV. Parameter Fitting of I-LM Algorithm

[0080] The specific method for parameter fitting in the I-LM algorithm is as follows: The open circuit voltage (OCV) of the battery is the terminal voltage of the battery under no-load conditions, and it has a complex nonlinear relationship with the battery's state of charge (SOC). This relationship is determined by the battery's chemical characteristics and is usually expressed as:

[0081]

[0082] in It is the vector of physical parameters to be identified.

[0083] Based on the analysis of battery electrochemical characteristics, the following hybrid model is used to describe the OCV-SOC relationship:

[0084]

[0085] Index Term : Describes the nonlinear voltage behavior of a battery in extreme SOC regions (close to 0% or 100%).

[0086] constant term The fundamental potential of the battery;

[0087] Linear terms Linear voltage response caused by changes in SOC;

[0088] Quadratic terms Consider the curvature variation of the voltage curve;

[0089] cubic term : To capture higher-order nonlinear characteristics of voltage curves.

[0090] 1. Definition of objective function:

[0091] Given a set of experimental measurement data points Where N is the number of measurement points. Define the residual function:

[0092]

[0093] 2. Least squares objective function:

[0094]

[0095] 3. Jacobian matrix calculation:

[0096] Jacobian matrix Includes the partial derivatives of the residuals with respect to each parameter:

[0097]

[0098] The specific partial derivatives are:

[0099]

[0100] 4. The gradient of the objective function is:

[0101]

[0102]

[0103] 5. Algorithm Iteration Formula:

[0104] The I-LM algorithm obtains the parameter update step size by solving the following system of linear equations. :

[0105]

[0106] in:

[0107] It is the damping coefficient;

[0108] yes identity matrix;

[0109] It is the parameter update vector;

[0110] 6. Parameter update rules:

[0111]

[0112] This embodiment proposes an innovative adaptive adjustment strategy for the damping coefficient:

[0113] A. Residual term:

[0114]

[0115]

[0116] B. Gradient term:

[0117]

[0118]

[0119] C. Model quality items:

[0120] Calculate the ratio of the actual decline to the predicted decline:

[0121]

[0122]

[0123]

[0124]

[0125]

[0126]

[0127] D. Iterative decay term:

[0128] When the iteration count iter < 20:

[0129]

[0130] otherwise:

[0131]

[0132]

[0133] E. Combination and attenuation:

[0134]

[0135] in: Let j be the j-th computational term, and let it form a vector. , for The corresponding normalized weights, which form the weight vector ; and The calculation method is as follows:

[0136]

[0137]

[0138]

[0139] F. Damping coefficient adjustment rules:

[0140]

[0141]

[0142] G. Convergence Criterion:

[0143] Gradient convergence: ;

[0144] Parameter changes converge: ;

[0145] The function value changes converge: ;

[0146] Maximum number of iterations: ;

[0147] Initial parameters: , , ; , , ; , , , ; .

[0148] This embodiment is based on the algorithm described in [the document], and the specific steps are as follows:

[0149] 1. Parameter initialization:

[0150] Set the initial guess vector of the parameters to be fitted ,correspond to The initial value (can be set based on experience).

[0151] Setting I-LM algorithm parameters: initial damping coefficient Weighting coefficient (satisfy ), gradient precision threshold Reference scale = 20、 = 150, damping coefficient boundary , .

[0152] 2. Iterative optimization process:

[0153] Step 1: Calculate the current state. Convert the current parameter vector... Substitute into the model and calculate the residual vector. (Difference between measured and predicted values), Jacobian matrix and gradient vector .

[0154] Step 2: Convergence check. If the gradient norm... If the iteration fails, the iteration terminates; otherwise, it continues.

[0155] Step 3: Solve for the LM step size. Solve the equation. Get the step size Update parameters .

[0156] Step 4: Adaptive Damping Adjustment:

[0157] Step 4.1: Calculate the actual decrease and predicted decline ;

[0158] It is the objective function. It is the current point. It's an LM step. It is a candidate point;

[0159] Step 4.2: If If step size is invalid, adjust the damping coefficient according to the number of iterations. If the number of iterations is less than 10, multiply by 3.0; if it is greater than or equal to 10, multiply by 1.8 and return to step 1.

[0160] Step 4.3: If If the step size is valid, then accept the step size and update based on multi-factor collaborative decision-making. :

[0161] Calculate the residual term (using the hyperbolic tangent function to handle the residual scaling), the gradient term (using the exponential decay function to handle the gradient scaling), and the model quality term (based on...). Factors, processed through the Sigmoid function.

[0162] The weighted geometric mean of the three terms is combined with the iterative decay factor (0.02 when the number of iterations is <20, and 0.05 when the number of iterations is ≥20) to obtain the combined factor.

[0163] A three-level scaling strategy is adopted based on the size of the combination factor: when the combination factor is >0.1, it is aggressively increased (scaling factor 1 + 1.5 × combination factor); when it is 0.01–0.1, it is moderately increased (scaling factor 1 + combination factor); and when it is ≤0.01, it is conservatively decreased (scaling factor 0.9).

[0164] Will Limited to Within the range.

[0165] Step 4.4: Repeat steps 1–4 until convergence.

[0166] 3. Parameter extraction application:

[0167] Using voltage relaxation data from HPPC experiments (such as...) Figure 3 (BD segment), using the formula Fit the parameters, and then calculate using formulas (9) and (10). .

[0168] The average value of the fitting results from 10 experiments was taken as the final model parameters.

[0169] V. SOC Calculation and Verification

[0170] 1. SOC Estimation: Based on the fitted parameters, the SOC is calculated using the ampere-hour integration method.

[0171]

[0172] in For Coulomb efficiency, This is the rated capacity. Simultaneously, a third-order RC model is used for real-time correction, and a state observer is used to reduce the accumulated error of the ampere-hour integral.

[0173] 2. Validation and Optimization: Compare the SOC estimation results with the high-precision reference values, and calculate the mean absolute error (MAE) and root mean square error (RMSE). If the error is large, adjust the I-LM algorithm parameters (such as weighting coefficients or damping boundaries) and refit until the accuracy requirements are met.

[0174] Through the above embodiments, the present invention achieves:

[0175] High-precision prediction: The combination of the I-LM algorithm and the third-order RC model improves the accuracy of parameter fitting and significantly reduces the SOC estimation error.

[0176] Fast convergence: The adaptive damping mechanism enables the algorithm to converge within 10–20 iterations, which is about 30% faster than the traditional AUKF algorithm.

[0177] High adaptability: It remains robust under dynamic operating conditions and is suitable for embedded BMS.

[0178] This embodiment fully discloses the technical details and complies with the requirements of patent law. Those skilled in the art can make adjustments without departing from the core of the invention, and such adjustments should be considered to fall within the scope of protection.

[0179] The above embodiments demonstrate that the present invention, through the multi-factor adaptive damping adjustment mechanism of the I-LM algorithm, significantly improves parameter fitting efficiency and SOC estimation accuracy, providing a reliable technical solution for battery management systems. Those skilled in the art can make various modifications and alterations to the present invention based on the above teachings, and these equivalent forms also fall within the scope defined by the appended claims.

[0180] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0181] In another embodiment, a battery management system is provided, which corresponds one-to-one with the lithium-ion battery state-of-charge estimation method described in the above embodiments. Specifically, it includes:

[0182] The data acquisition module is configured to perform pulse discharge experiments on the battery and collect voltage and current data of the battery under different load conditions.

[0183] The model building module is configured to use a third-order RC equivalent circuit model to describe battery characteristics. The model includes ohmic internal resistance, ohmic polarization internal resistance, electrochemical polarization internal resistance, concentration difference polarization internal resistance, ohmic polarization capacitance, electrochemical polarization capacitance, and concentration difference polarization capacitance.

[0184] The parameter fitting module is configured to fit the parameters of the third-order RC equivalent circuit model using an improved I-LM algorithm, wherein the I-LM algorithm adaptively adjusts the damping coefficient through a multi-factor collaborative decision-making mechanism to optimize parameter estimation.

[0185] The SOC calculation module is configured to calculate the battery's state of charge (SOC) value based on the fitted model parameters, using either the ampere-hour integration method or the model output.

[0186] The monitoring and control module is configured to achieve real-time SOC estimation, monitor battery status, and output control signals to manage the battery charging and discharging process.

[0187] For specific limitations regarding the battery management system, please refer to the limitations on the lithium-ion battery state-of-charge estimation method mentioned above, which will not be repeated here. Each module in the above system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0188] In another embodiment of the present invention, a terminal device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions to achieve a corresponding method flow or corresponding function. The processor described in this embodiment of the present invention can be used for the operation of a lithium-ion battery state-of-charge estimation method.

[0189] In another embodiment of the present invention, a storage medium is provided, specifically a computer-readable storage medium (Memory). This computer-readable storage medium is a memory device in a terminal device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the terminal device and extended storage media supported by the terminal device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, this storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device.

[0190] One or more instructions stored in a computer-readable storage medium can be loaded and executed by a processor to implement the corresponding steps of the lithium-ion battery state-of-charge estimation method in the above embodiments; one or more instructions in the computer-readable storage medium are loaded and executed by a processor.

[0191] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0192] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0193] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for estimating the state of charge of a lithium-ion battery, characterized in that, Includes the following steps: 1) Data acquisition: Conduct pulse discharge experiments on the battery to collect voltage and current data under different load conditions; 2) Model construction: A third-order RC equivalent circuit model is adopted as the battery model. The third-order RC equivalent circuit model includes ohmic internal resistance R0, ohmic polarization internal resistance R1, ohmic polarization capacitance C1, electrochemical polarization internal resistance R2, electrochemical polarization capacitance C2, concentration difference polarization internal resistance R3, and concentration difference polarization capacitance C3. 3) Parameter fitting: The parameters of the third-order RC equivalent circuit model are fitted using the improved I-LM algorithm. The I-LM algorithm adaptively adjusts the damping coefficient through a multi-factor collaborative decision-making mechanism, including: calculating the combination factor based on the residual term, gradient term and model quality term, and dynamically scaling the damping coefficient according to the combination factor to optimize the parameter fitting process. The damping coefficient adjustment process of the I-LM algorithm includes: The residual term is calculated based on the ratio of the current residual vector norm to the preset reference residual value; The gradient term is calculated based on the ratio of the current gradient vector norm to the preset reference gradient value; The model quality term is calculated based on the ratio of the actual decrease in the objective function to the predicted decrease. A weighted geometric mean is calculated for the residual term, gradient term, and model quality term, and combined with the iterative decay factor to generate a combined factor. Based on the magnitude of the combined factors, a three-level strategy is adopted to adjust the scaling factor of the damping coefficient, including aggressively increasing, moderately increasing or decreasing the damping coefficient; 4) SOC estimation: Using the fitted model parameters, the real-time state of charge (SOC) of the battery is calculated by the ampere-hour integration method or the state observer.

2. The method for estimating the state of charge of a lithium-ion battery according to claim 1, characterized in that, In the data acquisition step, the pulse discharge experiment includes a mixed pulse power characteristic test, which is repeated multiple times to obtain average data.

3. The method for estimating the state of charge of a lithium-ion battery according to claim 1, characterized in that, In step 2), the open-circuit voltage of the third-order RC equivalent circuit model The relationship with SOC is defined by the following formula: ， in , , , , These are the parameters obtained by fitting using the I-LM algorithm.

4. The lithium-ion battery state-of-charge estimation method according to claim 3, characterized in that, The residual term is calculated using a hyperbolic tangent function to compress and map the residual proportion; the gradient term is calculated using an exponential decay function to map the gradient proportion; and the model quality term is calculated using a sigmoid function to map the ratio of the actual to the predicted decrease.

5. The method for estimating the state of charge of a lithium-ion battery according to claim 3, characterized in that, The weighting parameters of the weighted geometric mean include , and ,in These are the residual weighting coefficients. These are the gradient weight coefficients. These are the model quality weight coefficients, and .

6. The method for estimating the state of charge of a lithium-ion battery according to claim 1, characterized in that, In step 3), the initial parameters of the I-LM algorithm include the initial value of the damping coefficient. Weight parameters , , Gradient accuracy threshold Reference scale and The damping coefficient is updated iteratively until the convergence condition is met.

7. The method for estimating the state of charge of a lithium-ion battery according to claim 1, characterized in that, In step 4), the SOC is calculated using the discretized ampere-hour integral formula: ， in For Coulomb efficiency, For the battery's rated capacity, The initial time, The sampling interval is... This represents the real-time current.

8. A battery management system, characterized in that, Configured to execute a lithium-ion battery state-of-charge estimation method according to any one of claims 1 to 7, for real-time monitoring and control of the state of charge of lithium-ion batteries, specifically including: The data acquisition module is configured to perform pulse discharge experiments on the battery and collect voltage and current data of the battery under different load conditions. The model building module is configured to use a third-order RC equivalent circuit model to describe battery characteristics. The third-order RC equivalent circuit model includes ohmic internal resistance, ohmic polarization internal resistance, electrochemical polarization internal resistance, concentration difference polarization internal resistance, ohmic polarization capacitance, electrochemical polarization capacitance, and concentration difference polarization capacitance. The parameter fitting module is configured to fit the parameters of the third-order RC equivalent circuit model using an improved I-LM algorithm, wherein the I-LM algorithm adaptively adjusts the damping coefficient through a multi-factor collaborative decision-making mechanism to optimize parameter estimation. The SOC calculation module is configured to calculate the battery's state of charge (SOC) value based on the fitted model parameters, using either the ampere-hour integration method or the model output. The monitoring and control module is configured to achieve real-time SOC estimation, monitor battery status, and output control signals to manage the battery charging and discharging process.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements a method for estimating the state of charge of a lithium-ion battery according to any one of claims 1 to 7.

Images