Secondary battery charging method based on soc optimization

Abstract

The application discloses a secondary battery charging method based on SOC optimization and belongs to the technical field of secondary battery charging. The steps include: collecting and preprocessing battery pack data, determining a historical charging and discharging state marker, and extracting a dynamic polarization time constant and a dynamic polarization resistance; superimposing a hysteresis correction compensation amount on a preset open-circuit voltage-state-of-charge mapping curve to determine an initial state-of-charge; outputting a real-time state-of-charge by using an extended Kalman filter algorithm model; updating a capacity reference of the real-time state-of-charge by using a recursive least square method with a forgetting factor to obtain a corrected state-of-charge; calculating a negative electrode surface potential estimation value by using a negative electrode potential estimation model based on a single particle model architecture; and judging an output maximum allowable charging current instruction and a current limit dynamic adjustment instruction. The application realizes dynamic adjustment of a charging current by precisely optimizing SOC estimation, adapting to battery aging and evaluating lithium precipitation risk at a mechanism level, and takes into account secondary battery charging efficiency and cycle life.

Description

Technical Field

[0001] This invention belongs to the field of secondary battery charging technology, specifically relating to a secondary battery charging method based on SOC optimization. Background Technology

[0002] Currently, secondary batteries, especially lithium-ion batteries, are widely used in consumer electronics, energy storage systems, industrial equipment, new energy vehicles and other fields. The control strategy in the charging process directly determines the charging efficiency, cycle life and safety of the battery. The accurate estimation and dynamic optimization of the state of charge (SOC) is the core technology foundation for the battery management system (BMS) to achieve efficient and safe charging control, and it is also a key component of the secondary battery charging method.

[0003] Current secondary battery charging primarily employs a constant current-constant voltage segmented charging mode. The charger controls power output based on preset voltage and current thresholds, while the battery management system monitors basic charging status by collecting external parameters such as battery terminal voltage, charging current, and ambient temperature. In the SOC estimation stage, an open-circuit voltage mapping method combined with a current integration method based on a fixed capacity benchmark is often used. Specifically, the initial SOC is determined by matching the terminal voltage at the start of charging with a preset open-circuit voltage-SOC mapping curve. Then, using the battery's rated capacity as a fixed benchmark, the SOC is updated in real-time during charging through current integration. In the charging safety constraint stage, battery terminal voltage, maximum allowable temperature, and maximum charging current are the core thresholds. When these parameters reach the thresholds, current reduction or charging termination is implemented.

[0004] However, existing technologies still have some shortcomings in practical applications, making it difficult to simultaneously achieve accuracy, efficiency, and safety in battery charging. For example:

[0005] The open-circuit voltage mapping method ignores the voltage hysteresis effect after battery charging and discharging. The terminal voltage at the start of charging is not the real open-circuit voltage. The basic SOC obtained by direct mapping has obvious errors, and these errors will continue to accumulate with current integration, resulting in insufficient accuracy of SOC estimation throughout the process and failing to provide accurate state basis for charging control.

[0006] Using the battery's rated capacity as a fixed capacity benchmark for SOC calculation does not take into account the aging characteristics of the battery during cycle use, such as capacity decay and increased ohmic internal resistance. This further widens the deviation in SOC calculation for aged batteries and makes it easy to have problems such as "phantom charge" and "overcharge".

[0007] Existing safety constraints only target externally collectable parameters of the battery and cannot detect the internal electrochemical reaction state of the battery. Especially for lithium-ion batteries, it is impossible to assess the core safety risk of lithium plating at the negative electrode online. It is difficult to avoid problems such as permanent capacity decay, internal short circuit or even thermal runaway caused by lithium plating from the mechanism level by controlling the terminal voltage threshold alone.

[0008] The threshold of charging current is mostly a preset fixed value, which cannot be dynamically adjusted according to the real-time SOC state and internal electrode potential state of the battery. This can easily lead to situations where high current charging in the high SOC range induces lithium plating, or low current charging in the low SOC range reduces charging efficiency, making it difficult to balance charging efficiency and battery cycle life.

[0009] Therefore, there is an urgent need for a battery charging optimization method based on accurate SOC estimation. This method would eliminate initial SOC estimation bias, adapt to battery aging characteristics, achieve mechanism-level online assessment of internal safety risks, and dynamically adjust the charging current in conjunction with the real-time battery status. This would solve the problems of low charging accuracy, poor aging adaptability, and insufficient safety protection in existing technologies. Summary of the Invention

[0010] In view of the shortcomings of the prior art, the present invention aims to provide a secondary battery charging method based on SOC optimization. By accurately optimizing SOC estimation, adapting to battery aging and assessing lithium plating risk at the mechanistic level, the charging current is dynamically adjusted, taking into account both secondary battery charging efficiency and cycle life.

[0011] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0012] The SOC-optimized rechargeable battery charging method includes the following steps:

[0013] S1. Collect and preprocess the battery pack's terminal voltage data, charging current data, ambient temperature data, and historical current data; determine the historical charge and discharge state markers; and extract the dynamic polarization time constant and dynamic polarization resistance.

[0014] S2. Based on the historical charge and discharge state markings, a hysteresis correction compensation is superimposed on the preset open-circuit voltage-state of charge mapping curve to determine the initial state of charge.

[0015] S3. Input the initial state of charge, dynamic polarization time constant, and dynamic polarization resistance into the extended Kalman filter algorithm model, and output the real-time state of charge.

[0016] S4. Using the recursive least squares method with a forgetting factor, the ohmic internal resistance of the battery pack is identified and the actual usable capacity is calculated by using the preprocessed terminal voltage data and charging current data. The capacity benchmark of the real-time state of charge is updated to obtain the corrected state of charge.

[0017] S5. Using a negative electrode potential estimation model based on a single-particle model architecture, the estimated value of the negative electrode surface potential is calculated by utilizing the corrected state of charge, dynamic polarization resistance, preprocessed charging current data, and the Butler-Folmer equation.

[0018] S6. Based on the corrected state of charge and the estimated negative electrode surface potential, determine the maximum allowable charging current command and the current limit dynamic adjustment command, and execute the lithium plating risk warning.

[0019] As a preferred embodiment of the present invention, the process of collecting and preprocessing the battery pack's terminal voltage data, charging current data, ambient temperature data, and historical current data in S1 includes:

[0020] A voltage sensor is connected in parallel to the positive and negative total output terminals of the battery pack to obtain terminal voltage data. A current sensor is connected in series to the main current loop of the battery pack to obtain charging current data. A temperature sensor is attached to the surface of the individual cells inside the battery pack to obtain ambient temperature data. The system is connected to the memory through the controller local area network bus communication interface to read the historical current data recorded in the memory.

[0021] A sliding sampling window with a fixed width is established for the terminal voltage data, charging current data, ambient temperature data, and historical current data. The arithmetic mean of the data within the sliding sampling window is calculated and then subjected to mean filtering to obtain the preprocessed terminal voltage data, charging current data, ambient temperature data, and historical current data. The preprocessed values ​​are determined by dividing the sum of the data at the current sampling time and the previous sampling time by the total number of samples.

[0022] As a preferred embodiment of the present invention, the process of determining the historical charge and discharge state markers and extracting the dynamic polarization time constant and dynamic polarization resistance in S1 includes:

[0023] The historical current data is accumulated over time to obtain the total historical current. The historical charge and discharge status is marked according to the polarity of the total historical current. If the total historical current is positive, the historical charge and discharge status is marked as charging. If the total historical current is negative, the historical charge and discharge status is marked as discharging.

[0024] Based on the preprocessed ambient temperature data, the parameter mapping table in the memory is retrieved. Through linear interpolation logic, the corresponding dynamic polarization time constant and dynamic polarization resistance are matched from the parameter mapping table and output.

[0025] As a preferred embodiment of the present invention, the process of determining the initial state of charge in S2 includes:

[0026] The preprocessed terminal voltage data acquired at the start of charging is determined as open-circuit voltage data. By retrieving the preset open-circuit voltage-state of charge mapping curve in the memory, the value corresponding to the open-circuit voltage data is matched to determine the basic state of charge.

[0027] Retrieve the preset hysteresis correction parameter table in the memory, and extract the hysteresis correction compensation amount corresponding to the preprocessed ambient temperature data from the hysteresis correction parameter table;

[0028] Using historical charge and discharge state markers and hysteresis correction compensation, an algebraic superposition operation is performed on the basic state of charge to obtain the initial state of charge. Here, a value of 1 in the historical charge and discharge state marker corresponds to the charging state, and a value of -1 corresponds to the discharging state.

[0029] As a preferred embodiment of the present invention, the process of outputting the real-time state of charge in S3 includes:

[0030] A state vector consisting of a state of charge component and a polarization voltage component is established. The initial value of the state vector consists of the initial state of charge and the initial value of the polarization voltage, which is preset to zero volts. The initial values ​​of the error covariance matrix, the process noise covariance matrix, and the measurement noise covariance matrix are also preset. The initial value of the capacity reference is set to the rated capacity of the battery pack.

[0031] Using the dynamic polarization time constant and sampling period at the current sampling moment, the state transition coefficient is calculated. Using the state transition coefficient, the state vector at the previous sampling moment, the dynamic polarization resistance, and the preprocessed charging current data, the prior predicted values ​​of the state of charge component and the polarization voltage component are calculated respectively, thus forming the prior state estimate value at the current moment.

[0032] The prior prediction value of the state of charge component is determined by superimposing the state of charge component from the previous moment with a dimensionless change in charge. The change in charge is determined by dividing the current integral by the current capacity reference and performing a time unit conversion. The current integral is derived from the preprocessed charging current data and the sampling period. Determine by performing a product operation;

[0033] The prior prediction of the polarization voltage component is determined by performing an exponential decay operation on the polarization voltage component of the previous moment through the state transition coefficient, and superimposed with the polarization voltage response increment determined by the dynamic polarization resistance and the preprocessed charging current data of the previous moment.

[0034] The prediction error covariance matrix is ​​calculated by combining the state transition coefficients, the error covariance matrix of the previous time step, and the process noise covariance matrix. The prediction error covariance matrix is ​​determined by multiplying the state transition coefficients and the error covariance matrix and then superimposing the process noise covariance matrix.

[0035] Based on the state of charge component in the prior state estimate, the open-circuit voltage-state of charge mapping curve in the memory is retrieved to obtain the predicted open-circuit voltage.

[0036] The polarization voltage component in the open-circuit voltage prediction, the prior state estimation, and the terminal voltage deviation compensation amount generated by the dynamic polarization resistor are algebraically added to obtain the terminal voltage estimation value.

[0037] Calculate the voltage deviation between the preprocessed terminal voltage data and the estimated terminal voltage value;

[0038] The Kalman gain matrix of the extended Kalman filter algorithm model is calculated using the prediction error covariance matrix and the measurement noise covariance matrix.

[0039] The prior state estimate is corrected using the Kalman gain matrix and the voltage deviation value to obtain the posterior state vector, and the error covariance matrix is ​​updated synchronously as the initial value for calculation at the next sampling time.

[0040] The values ​​of the charged state components in the posterior state vector are extracted and determined as the real-time charged state.

[0041] As a preferred embodiment of the present invention, the process of obtaining the corrected state of charge in S4 includes:

[0042] The parameter identification process is performed using the recursive least squares method with a forgetting factor. The ohmic internal resistance of the battery pack is determined as the parameter to be identified, and the pre-processed terminal voltage data is determined as the system output and the pre-processed charging current data is determined as the system input.

[0043] The prediction error is calculated using the system input and system output, and the gain vector is calculated by combining the preset forgetting factor, the covariance matrix of the previous time step, and the system input.

[0044] The parameter correction term is obtained by multiplying the gain vector and the prediction error, and then the parameter correction term is added to the ohmic internal resistance of the battery pack at the previous moment to update the ohmic internal resistance of the battery pack in real time.

[0045] During the charging process, an analysis window with a preset duration is selected, and the cumulative amount of preprocessed charging current data within the analysis window is calculated and determined as the current variation value.

[0046] Obtain the first real-time state of charge at the start of the analysis window and the second real-time state of charge at the end of the analysis window, calculate the algebraic difference between the second real-time state of charge and the first real-time state of charge, and determine it as the change in charge value.

[0047] The actual available capacity is calculated using the current variation and the energy variation.

[0048] Replace the rated capacity parameters used in the real-time state of charge with the actual available capacity to complete the update of the capacity reference for the real-time state of charge.

[0049] Using the updated capacity reference and sampling period, perform a recursive calculation based on current integral for the current state of charge and output a corrected state of charge.

[0050] As a preferred embodiment of the present invention, the process of calculating the estimated value of the negative electrode surface potential in S5 includes:

[0051] A negative electrode potential estimation model based on a single-particle model architecture is established. By using the corrected state of charge, the negative electrode potential estimation model retrieves the preset open-circuit potential curves of the negative electrode material in the memory. Through linear interpolation logic, the potential values ​​corresponding to the corrected state of charge are matched from the open-circuit potential curves of the negative electrode material to determine the negative electrode equilibrium potential.

[0052] The negative electrode potential estimation model retrieves the preset ideal gas constant, charge transfer coefficient, and Faraday constant from the memory, and uses the absolute temperature obtained from the preprocessed ambient temperature data to determine the kinetic constant term;

[0053] The preset negative electrode active area and exchange current density are retrieved from the memory, and the electrochemical overpotential is calculated by using the kinetic constant term and the preprocessed charging current data at the current moment through the transformation function of the Butler-Folmer equation.

[0054] The estimated value of the negative electrode surface potential is obtained by subtracting the electrochemical overpotential from the negative electrode equilibrium potential, and by subtracting the product of the preprocessed charging current data and the dynamic polarization resistance at the current moment.

[0055] As a preferred embodiment of the present invention, the process of determining the maximum allowable charging current command and the current limit dynamic adjustment command in S6 includes:

[0056] Retrieve the preset charging range, safety potential threshold, and proportional-integral-derivative adjustment coefficient from the memory, compare the corrected state of charge with the preset charging range, and compare the estimated value of the negative electrode surface potential with the safety potential threshold.

[0057] Under the condition that the state of charge is within the preset charging range and the estimated value of the negative electrode surface potential is not lower than the safe potential threshold, the maximum allowable charging current command is output.

[0058] Under the condition that the state of charge is in the preset charging range and the estimated value of the negative electrode surface potential is lower than the safe potential threshold, the algebraic difference between the safe potential threshold and the estimated value of the negative electrode surface potential is calculated and determined as the potential deviation value. The proportional, integral and derivative adjustment coefficients are used to perform proportional, integral and derivative algebraic operations on the potential deviation value to determine the current adjustment increment in amperes. The current adjustment increment is subtracted from the preprocessed charging current data of the previous moment to determine the charging current limit at the current moment, and the current limit dynamic adjustment command is output.

[0059] When the state of charge exceeds the preset charging range, a charging termination command is output.

[0060] As a preferred embodiment of the present invention, the process of performing lithium plating risk warning in S6 includes:

[0061] Subtract the estimated negative electrode surface potential from the previous sampling time from the current estimated negative electrode surface potential, and divide by the sampling period to obtain the rate of change of the estimated negative electrode surface potential over time.

[0062] The preset lithium plating change rate threshold is retrieved from the memory. The rate of change of the estimated negative electrode surface potential over time is compared with the lithium plating change rate threshold. When the rate of change of the estimated negative electrode surface potential over time exceeds the lithium plating change rate threshold, a lithium plating risk warning signal is triggered.

[0063] As a preferred embodiment of the present invention, after S6, it further includes:

[0064] S7. Present a three-dimensional charging characteristic curve composed of the corrected state of charge, actual usable capacity, and negative electrode surface potential estimate through a graphical interface: Construct a three-dimensional spatial coordinate system in the graphical interface, map the corrected state of charge to the first axis coordinate, the actual usable capacity to the second axis coordinate, and the negative electrode surface potential estimate to the third axis coordinate; perform surface fitting operation based on continuously sampled coordinate point data to generate a three-dimensional charging characteristic curve composed of the corrected state of charge, actual usable capacity, and negative electrode surface potential estimate in real time.

[0065] The beneficial effects of this invention are as follows:

[0066] This invention addresses the issue of state estimation accuracy during the initial charging stage of secondary batteries. By analyzing historical current data to determine charge and discharge state markers and superimposing hysteresis correction compensation, it effectively eliminates the interference of voltage hysteresis effect on the initial state of charge judgment, significantly improving the accuracy of the state of charge value during the initial charging stage of secondary batteries. This lays a reliable initial data foundation for the precise execution of subsequent charging control strategies.

[0067] This invention addresses the aging adaptation requirements of secondary batteries during long-term charging. It utilizes a recursive least squares method with a forgetting factor to identify the ohmic internal resistance in real time and calculate the actual usable capacity. This enables the dynamic updating of the state of charge (SOC) calculation logic as the battery pack ages, solving the problem of SOC calculation deviation caused by traditional fixed capacity benchmarks. This ensures the stability and reliability of SOC estimation throughout the entire life cycle of the secondary battery.

[0068] This invention focuses on balancing the safety and efficiency of secondary battery charging. By using a single-particle model architecture and the Butler-Folmore equation to calculate the estimated value of the negative electrode surface potential, it achieves a mechanistic-level online assessment of the risk of lithium plating. While maintaining the negative electrode surface potential within a safe threshold range, it dynamically adjusts the current limit, effectively avoiding the risk of lithium plating during the secondary battery charging process and maximizing charging efficiency, thus balancing the dual requirements of charging speed and battery cycle life. Attached Figure Description

[0069] Figure 1 This is a flowchart of the secondary battery charging method based on SOC optimization in Embodiment 1 of the present invention;

[0070] Figure 2 This is a positive charge state curve diagram during the verification process of this invention;

[0071] Figure 3 This is a graph showing the surface potential curve of the negative electrode during the verification process of this invention;

[0072] Figure 4 This is a dynamic adjustment curve of the charging current command during the verification process of the present invention.

[0073] Figure 5 This is a flowchart of the secondary battery charging method based on SOC optimization in Embodiment 2 of the present invention. Detailed Implementation

[0074] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0075] Example 1: As Figure 1 As shown, the SOC-optimized secondary battery charging method is mainly applicable to the charging process of lithium-ion batteries and includes the following steps:

[0076] S1. Collect and preprocess the battery pack's terminal voltage data, charging current data, ambient temperature data, and historical current data; determine the historical charge and discharge state markers; and extract the dynamic polarization time constant and dynamic polarization resistance.

[0077] S2. Based on the historical charge and discharge state markings, a hysteresis correction compensation is superimposed on the preset open-circuit voltage-state of charge mapping curve to determine the initial state of charge.

[0078] S3. Input the initial state of charge, dynamic polarization time constant, and dynamic polarization resistance into the extended Kalman filter algorithm model, and output the real-time state of charge.

[0079] S4. Using the recursive least squares method with a forgetting factor, the ohmic internal resistance of the battery pack is identified and the actual usable capacity is calculated by using the preprocessed terminal voltage data and charging current data. The capacity benchmark of the real-time state of charge is updated to obtain the corrected state of charge.

[0080] S5. Using a negative electrode potential estimation model based on a single-particle model architecture, the estimated value of the negative electrode surface potential is calculated by utilizing the corrected state of charge, dynamic polarization resistance, preprocessed charging current data, and the Butler-Folmer equation.

[0081] S6. Based on the corrected state of charge and the estimated negative electrode surface potential, determine the maximum allowable charging current command and the current limit dynamic adjustment command, and execute the lithium plating risk warning.

[0082] In S1, the process of collecting and preprocessing the battery pack's terminal voltage data, charging current data, ambient temperature data, and historical current data includes:

[0083] A voltage sensor is connected in parallel to the positive and negative total output terminals of the battery pack to obtain terminal voltage data. A current sensor is connected in series to the main current loop of the battery pack to obtain charging current data. A temperature sensor is attached to the surface of the individual cells inside the battery pack to obtain ambient temperature data. The system is connected to a memory via a controller local area network bus communication interface to read the historical current data recorded in the memory.

[0084] A sliding sampling window with a fixed width is established for the terminal voltage data, charging current data, ambient temperature data, and historical current data. The arithmetic mean of the data within the sliding sampling window is calculated and then subjected to mean filtering to obtain the preprocessed terminal voltage data, charging current data, ambient temperature data, and historical current data. The preprocessed values ​​are determined by dividing the sum of the data at the current sampling time and the previous sampling time by the total number of samples.

[0085] The process of determining historical charge / discharge state markers and extracting dynamic polarization time constants and dynamic polarization resistances includes:

[0086] The historical current data is accumulated over time to obtain the historical current sum. The historical charge / discharge state is marked based on the polarity of the historical current sum. If the historical current sum is positive, the historical charge / discharge state is marked as charging state. If the historical current sum is negative, the historical charge / discharge state is marked as discharging state.

[0087] Based on the preprocessed ambient temperature data, the parameter mapping table in the memory is retrieved. Through linear interpolation logic, the corresponding dynamic polarization time constant and dynamic polarization resistance are matched from the parameter mapping table and output.

[0088] In S2, the process of determining the initial state of charge includes:

[0089] The preprocessed terminal voltage data acquired at the start of charging is determined as the open-circuit voltage data. By retrieving the preset open-circuit voltage-state-of-charge mapping curve in the memory, the value corresponding to the open-circuit voltage data is matched to determine the basic state of charge. .

[0090] The system retrieves the preset hysteresis correction parameter table from the memory and extracts the hysteresis correction compensation amount corresponding to the preprocessed ambient temperature data. .

[0091] Using historical charge / discharge state markers L and hysteresis correction compensation amount Regarding the fundamental state of charge Perform algebraic superposition to obtain the initial state of charge. Where L takes a value of 1 to correspond to the charging state and takes a value of -1 to correspond to the discharging state.

[0092] In S3, the process of outputting the real-time state of charge includes:

[0093] A state vector consisting of a state of charge component and a polarization voltage component is established. The initial value of the state vector consists of the initial state of charge and the initial value of the polarization voltage, which is preset to zero volts. The initial values ​​of the error covariance matrix, the process noise covariance matrix, and the measurement noise covariance matrix are also preset. The initial value of the capacity reference is set to the rated capacity of the battery pack.

[0094] Using the dynamic polarization time constant at the current sampling time With sampling period The state transition coefficients are calculated. Using the state transition coefficients, the state vector at the previous sampling time, the dynamic polarization resistance R, and the preprocessed charging current data, the prior predicted values ​​of the state of charge component and the polarization voltage component are calculated respectively, thus forming the prior state estimate value at the current time.

[0095] The prior prediction value of the state of charge component is determined by superimposing the state of charge component from the previous moment with a dimensionless change in charge. The change in charge is determined by dividing the current integral by the current capacity reference and performing a time unit conversion. The current integral is derived from the preprocessed charging current data and the sampling period. Perform the product operation to determine.

[0096] Prior prediction of polarization voltage component The polarization voltage component from the previous moment The state transition coefficients are subjected to exponential decay calculations, and the data is superimposed with the dynamic polarization resistor R and the preprocessed charging current data from the previous time step. The polarization voltage response increment is determined by common determination, where e is the natural constant.

[0097] The prediction error covariance matrix is ​​calculated by combining the state transition coefficients, the error covariance matrix of the previous time step, and the process noise covariance matrix. The prediction error covariance matrix is ​​determined by multiplying the state transition coefficients and the error covariance matrix, and then superimposing the process noise covariance matrix.

[0098] Based on the state of charge component in the prior state estimate, the open-circuit voltage-state of charge mapping curve in the memory is retrieved to obtain the predicted open-circuit voltage value.

[0099] The polarization voltage component in the open-circuit voltage prediction, the prior state estimation, and the terminal voltage deviation compensation amount generated by the dynamic polarization resistor are algebraically added together to obtain the terminal voltage estimation value.

[0100] Calculate the voltage deviation between the preprocessed terminal voltage data and the estimated terminal voltage value.

[0101] The Kalman gain matrix of the extended Kalman filter algorithm model is calculated using the prediction error covariance matrix and the measurement noise covariance matrix.

[0102] The prior state estimate is corrected using the Kalman gain matrix and the voltage deviation value to obtain the posterior state vector, and the error covariance matrix is ​​updated synchronously as the initial value for the calculation at the next sampling time.

[0103] The values ​​of the charged state components in the posterior state vector are extracted and determined as the real-time charged state.

[0104] In S4, the process of obtaining the corrected state of charge includes:

[0105] The parameter identification process is performed using the recursive least squares method with a forgetting factor. The ohmic internal resistance of the battery pack is determined as the parameter to be identified, and the preprocessed terminal voltage data is determined as the system output and the preprocessed charging current data is determined as the system input.

[0106] The prediction error is calculated using the system input and system output, and the gain vector is calculated by combining the preset forgetting factor, the covariance matrix of the previous time step, and the system input.

[0107] The parameter correction term is obtained by multiplying the gain vector and the prediction error, and then the parameter correction term is added to the ohmic internal resistance of the battery pack at the previous moment to update the ohmic internal resistance of the battery pack in real time.

[0108] During the charging process, select an analysis window with a preset duration (e.g., 5 minutes, 10 minutes, 20 minutes), calculate the cumulative amount of preprocessed charging current data within the analysis window, and determine it as the current variation value.

[0109] Obtain the first real-time state of charge at the start of the analysis window. And the second real-time state of charge at the end of the analysis window. Calculate the second real-time state of charge. With the first real-time state of charge The algebraic difference is determined as the change in electricity consumption.

[0110] Calculate the actual available capacity using current and charge fluctuations. ,in, This is the preprocessed charging current data corresponding to sampling point i. The sampling period is in seconds, k is the current sampling time, and m is the total number of sampling points corresponding to the analysis window.

[0111] The rated capacity parameters used in the real-time state of charge are replaced with the actual available capacity to complete the update of the capacity reference for the real-time state of charge.

[0112] Using the updated capacity reference and sampling period, perform a recursive calculation based on current integral for the current state of charge and output a corrected state of charge.

[0113] In S5, the process of calculating the estimated value of the negative electrode surface potential includes:

[0114] A negative electrode potential estimation model based on a single-particle model architecture is established. Using a corrected state of charge, the model retrieves preset open-circuit potential curves of the negative electrode material from memory. Through linear interpolation logic, the potential values ​​corresponding to the corrected state of charge are matched from the open-circuit potential curves to determine the negative electrode equilibrium potential. .

[0115] The negative electrode potential estimation model retrieves the preset ideal gas constant from memory. Charge transfer coefficient The Faraday constant F is used, and the absolute temperature T obtained from the preprocessed ambient temperature data is used to determine the kinetic constant term.

[0116] Retrieve the preset negative electrode active area from the memory. With exchange current density And using the dynamic constant term and the preprocessed charging current data at the current time. Electrochemical overpotentials were calculated using a modified function of the Butler-Folmer equation. :

[0117] .

[0118] Negative electrode equilibrium potential Subtract electrochemical overpotential And subtract the preprocessed charging current data at the current moment. The product of the potential and the dynamic polarization resistance R yields the estimated value of the negative electrode surface potential. .

[0119] In S6, the process of determining the maximum allowable charging current command and the current limit dynamic adjustment command includes:

[0120] The preset charging range, safety potential threshold, and proportional-integral-derivative adjustment coefficients are retrieved from the memory. The corrected state of charge is compared with the preset charging range, and the estimated value of the negative electrode surface potential is compared with the safety potential threshold.

[0121] Under the condition that the state of charge is within the preset charging range and the estimated value of the negative electrode surface potential is not lower than the safe potential threshold, the maximum allowable charging current command is output.

[0122] Under the condition that the state of charge is within the preset charging range and the estimated value of the negative electrode surface potential is lower than the safe potential threshold, the algebraic difference between the safe potential threshold and the estimated value of the negative electrode surface potential is calculated and determined as the potential deviation value. The proportional, integral and derivative adjustment coefficients are used to perform proportional, integral and derivative algebraic operations on the potential deviation value to determine the current adjustment increment in amperes. The current adjustment increment is subtracted from the preprocessed charging current data of the previous moment to determine the charging current limit at the current moment, and the dynamic adjustment command of the current limit is output.

[0123] When the state of charge exceeds the preset charging range, a charging termination command is output.

[0124] The process of implementing lithium plating risk warning includes:

[0125] Subtract the estimated negative electrode surface potential from the previous sampling time from the current estimated negative electrode surface potential, and divide by the sampling period to obtain the rate of change of the estimated negative electrode surface potential over time.

[0126] The preset lithium plating change rate threshold is retrieved from the memory. The rate of change of the estimated negative electrode surface potential over time is compared with the lithium plating change rate threshold. When the rate of change of the estimated negative electrode surface potential over time exceeds the lithium plating change rate threshold, a lithium plating risk warning signal is triggered.

[0127] The verification process in this embodiment is as follows:

[0128] Simulation experiments were conducted to verify the ability of the method in this embodiment to accurately estimate the real-time state of charge and achieve mechanistic-level safety boundary constraints under the conditions of battery pack cyclic aging and voltage hysteresis interference.

[0129] Using an aged battery pack with a rated capacity of 100Ah and an actual usable capacity reduced to 60Ah as the controlled object, a simulation experiment was performed with a sampling period of 1 second. The simulation experiment diagram is shown below. Figure 2-4 As shown, these correspond to the dynamic monitoring process of the method in this embodiment for the dynamic adjustment command of the corrected state of charge, the estimated value of the negative electrode surface potential, and the current limit.

[0130] Figure 2 The diagram shows the stage of obtaining the corrected state of charge (SOC) in this embodiment. The SOC starts at approximately 13% at the beginning of charging, rather than the 15% baseline SOC obtained through open-circuit voltage-SOC mapping curve matching. This is because a -2% hysteresis correction compensation is superimposed based on historical charge and discharge state markers, effectively eliminating the interference of voltage hysteresis effect on the determination of the initial SOC. Furthermore, at the 20-minute mark of the simulation, the slope of the red curve becomes significantly steeper. This reflects the use of recursive least squares method with forgetting factor to identify the ohmic internal resistance and calculate the actual usable capacity, completing the update of the capacity benchmark for the real-time SOC. The calculation denominator is switched from rated capacity to actual usable capacity (60Ah), ensuring the robustness of obtaining the corrected SOC under aging conditions.

[0131] Figure 3 The diagram shows the calculation and monitoring stage of the estimated negative electrode surface potential. The blue curve reflects the evolution trend of the estimated negative electrode surface potential in real time. As charging proceeds, the potential decreases smoothly. However, when it approaches the safe potential threshold of 0.01V (dashed line), the potential curve stops falling and precisely and smoothly stays above the safe potential threshold. It does not fall below 0V throughout the process. This verifies that the method in this embodiment achieves a mechanism-level online assessment of lithium plating risk through the Butler-Folmore equation, and constructs a deeper level of safety defense than traditional terminal voltage control.

[0132] Figure 4 The diagram shows the output stage of the current limit dynamic adjustment command. In the initial stage of charging, a smooth soft-start process is performed, and the current quickly rises to the high level set by the maximum allowable charging current command. When the estimated value of the negative electrode surface potential approaches the safe potential threshold, the proportional-integral-derivative adjustment logic is triggered, and the current limit dynamic adjustment command is output. The current (green curve) begins to adaptively and smoothly decrease. This phenomenon verifies the ability of the method in this embodiment to dynamically adjust the charging current while maintaining the safety of the negative electrode surface potential. When the corrected state of charge exceeds the preset charging range (reaching 100%), a charging termination command is output, ensuring closed-loop safety throughout the charging process.

[0133] This embodiment of the method effectively solves the problems that traditional charging logic cannot cope with battery pack aging and cannot detect the risk of internal lithium plating in real time by combining historical charge and discharge state marking correction, online parameter identification and capacity benchmark update, and negative electrode potential estimation model based on single particle model architecture. By combining these methods, the method ensures the safety and efficiency of the charging process through dynamic adjustment commands for output current limits.

[0134] In summary, the SOC-optimized recharge method for secondary batteries proposed in this embodiment demonstrates extremely high accuracy in estimating the state of charge and adaptive adjustment capability under complex aging environments. At the same time, through potential constraints at the mechanistic level, a complete lithium plating risk warning and prevention system is constructed.

[0135] Example 2: Figure 5 As shown, based on Example 1, after S6, it further includes:

[0136] S7. A three-dimensional charging characteristic curve, composed of the corrected state of charge, actual usable capacity, and estimated negative electrode surface potential, is presented through a graphical interface:

[0137] A three-dimensional spatial coordinate system is constructed in the graphical interface, the corrected state of charge is mapped to the first axis coordinate, the actual available capacity is mapped to the second axis coordinate, and the estimated value of the negative electrode surface potential is mapped to the third axis coordinate.

[0138] Surface fitting calculations are performed based on continuously sampled coordinate point data to generate a three-dimensional charging characteristic curve in real time, consisting of the corrected state of charge, actual usable capacity, and estimated negative electrode surface potential.

[0139] By constructing a three-dimensional spatial coordinate system, the corrected state of charge, actual usable capacity, and estimated negative electrode surface potential are mapped to three-dimensional coordinates and fitted to generate real-time three-dimensional charging characteristic curves. The dynamic changes of the core key parameters of battery charging are presented intuitively in a graphical interface. This not only allows for visual and comprehensive monitoring of the core state of the battery during the charging process and precise understanding of the correlation and evolution of various parameters, but also provides a more intuitive assessment of the execution effect of the charging control strategy and the real-time state of the battery. This provides clear and intuitive visualization support for the monitoring, debugging, and optimization of the charging process, and improves the convenience and accuracy of charging status monitoring.

[0140] Example 3: A secondary battery charging device based on SOC optimization, comprising:

[0141] One or more processors;

[0142] Memory, used to store one or more computer programs;

[0143] When one or more programs are executed by one or more processors, the one or more processors perform the method in Embodiment 1 or Embodiment 2.

[0144] Example 4: A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, implement the method in Example 1 or Example 2.

[0145] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A secondary battery charging method based on SOC optimization, characterized in that, Includes the following steps: S1. Collect and preprocess battery pack terminal voltage data, charging current data, ambient temperature data, and historical current data; determine historical charge / discharge state markers; and extract dynamic polarization time constant and dynamic polarization resistance. The process includes: The historical current data is accumulated over time to obtain the total historical current. The historical charge and discharge status is marked according to the polarity of the total historical current. If the total historical current is positive, the historical charge and discharge status is marked as charging. If the total historical current is negative, the historical charge and discharge status is marked as discharging. Based on the preprocessed ambient temperature data, the parameter mapping table in the memory is retrieved, and the corresponding dynamic polarization time constant and dynamic polarization resistance are matched and output from the parameter mapping table through linear interpolation logic. S2. Based on the historical charge and discharge state markings, a hysteresis correction compensation is superimposed on the preset open-circuit voltage-state of charge mapping curve to determine the initial state of charge. S3. Input the initial state of charge, dynamic polarization time constant, and dynamic polarization resistance into the extended Kalman filter algorithm model, and output the real-time state of charge. S4. Using the recursive least squares method with a forgetting factor, the ohmic internal resistance of the battery pack is identified and the actual usable capacity is calculated by using the preprocessed terminal voltage data and charging current data. The capacity benchmark of the real-time state of charge is updated to obtain the corrected state of charge. S5. Using a negative electrode potential estimation model based on a single-particle model architecture, the estimated value of the negative electrode surface potential is calculated by utilizing the corrected state of charge, dynamic polarization resistance, preprocessed charging current data, and the Butler-Folmer equation. S6. Based on the corrected state of charge and the estimated negative electrode surface potential, determine the maximum allowable charging current command and the current limit dynamic adjustment command, and execute the lithium plating risk warning.

2. The secondary battery charging method based on SOC optimization according to claim 1, characterized in that, In step S1, the process of collecting and preprocessing the battery pack's terminal voltage data, charging current data, ambient temperature data, and historical current data includes: A voltage sensor is connected in parallel to the positive and negative total output terminals of the battery pack to obtain terminal voltage data. A current sensor is connected in series to the main current loop of the battery pack to obtain charging current data. A temperature sensor is attached to the surface of the individual cells inside the battery pack to obtain ambient temperature data. The system is connected to the memory through the controller local area network bus communication interface to read the historical current data recorded in the memory. A sliding sampling window with a fixed width is established for the terminal voltage data, charging current data, ambient temperature data, and historical current data. The arithmetic mean of the data within the sliding sampling window is calculated and then subjected to mean filtering to obtain the preprocessed terminal voltage data, charging current data, ambient temperature data, and historical current data. The preprocessed values ​​are determined by dividing the sum of the data at the current sampling time and the previous sampling time by the total number of samples.

3. The secondary battery charging method based on SOC optimization according to claim 1, characterized in that, In S2, the process of determining the initial state of charge includes: The preprocessed terminal voltage data acquired at the start of charging is determined as open-circuit voltage data. By retrieving the preset open-circuit voltage-state of charge mapping curve in the memory, the value corresponding to the open-circuit voltage data is matched to determine the basic state of charge. Retrieve the preset hysteresis correction parameter table in the memory, and extract the hysteresis correction compensation amount corresponding to the preprocessed ambient temperature data from the hysteresis correction parameter table; Using historical charge and discharge state markers and hysteresis correction compensation, an algebraic superposition operation is performed on the basic state of charge to obtain the initial state of charge. Here, a value of 1 in the historical charge and discharge state marker corresponds to the charging state, and a value of -1 corresponds to the discharging state.

4. The secondary battery charging method based on SOC optimization according to claim 1, characterized in that, In S3, the process of outputting the real-time state of charge includes: A state vector consisting of a state of charge component and a polarization voltage component is established. The initial value of the state vector consists of the initial state of charge and the initial value of the polarization voltage, which is preset to zero volts. The initial values ​​of the error covariance matrix, the process noise covariance matrix, and the measurement noise covariance matrix are also preset. The initial value of the capacity reference is set to the rated capacity of the battery pack. Using the dynamic polarization time constant and sampling period at the current sampling moment, the state transition coefficient is calculated. Using the state transition coefficient, the state vector at the previous sampling moment, the dynamic polarization resistance, and the preprocessed charging current data, the prior predicted values ​​of the state of charge component and the polarization voltage component are calculated respectively, thus forming the prior state estimate value at the current moment. The prior prediction value of the state of charge component is determined by superimposing the state of charge component of the previous moment with the dimensionless change in charge. The change in charge is determined by dividing the current integral by the capacity reference at the current moment and performing time unit conversion. The current integral is determined by multiplying the preprocessed charging current data with the sampling period. The prior prediction of the polarization voltage component is determined by performing an exponential decay operation on the polarization voltage component of the previous moment through the state transition coefficient, and superimposed with the polarization voltage response increment determined by the dynamic polarization resistance and the preprocessed charging current data of the previous moment. The prediction error covariance matrix is ​​calculated by combining the state transition coefficients, the error covariance matrix of the previous time step, and the process noise covariance matrix. The prediction error covariance matrix is ​​determined by multiplying the state transition coefficients and the error covariance matrix and then superimposing the process noise covariance matrix. Based on the state of charge component in the prior state estimate, the open-circuit voltage-state of charge mapping curve in the memory is retrieved to obtain the predicted open-circuit voltage. The polarization voltage component in the open-circuit voltage prediction, the prior state estimation, and the terminal voltage deviation compensation amount generated by the dynamic polarization resistor are algebraically added to obtain the terminal voltage estimation value. Calculate the voltage deviation between the preprocessed terminal voltage data and the estimated terminal voltage value; The Kalman gain matrix of the extended Kalman filter algorithm model is calculated using the prediction error covariance matrix and the measurement noise covariance matrix. The prior state estimate is corrected using the Kalman gain matrix and the voltage deviation value to obtain the posterior state vector, and the error covariance matrix is ​​updated synchronously as the initial value for calculation at the next sampling time. The values ​​of the charged state components in the posterior state vector are extracted and determined as the real-time charged state.

5. The secondary battery charging method based on SOC optimization according to claim 1, characterized in that, In step S4, the process of obtaining the corrected state of charge includes: The parameter identification process is performed using the recursive least squares method with a forgetting factor. The ohmic internal resistance of the battery pack is determined as the parameter to be identified, and the pre-processed terminal voltage data is determined as the system output and the pre-processed charging current data is determined as the system input. The prediction error is calculated using the system input and system output, and the gain vector is calculated by combining the preset forgetting factor, the covariance matrix of the previous time step, and the system input. The parameter correction term is obtained by multiplying the gain vector and the prediction error, and then the parameter correction term is added to the ohmic internal resistance of the battery pack at the previous moment to update the ohmic internal resistance of the battery pack in real time. During the charging process, an analysis window with a preset duration is selected, and the cumulative amount of preprocessed charging current data within the analysis window is calculated and determined as the current variation value. Obtain the first real-time state of charge at the start of the analysis window and the second real-time state of charge at the end of the analysis window, calculate the algebraic difference between the second real-time state of charge and the first real-time state of charge, and determine it as the change in charge value. The actual available capacity is calculated using the current variation and the energy variation. Replace the rated capacity parameters used in the real-time state of charge with the actual available capacity to complete the update of the capacity reference for the real-time state of charge. Using the updated capacity reference and sampling period, perform a recursive calculation based on current integral for the current state of charge and output a corrected state of charge.

6. The secondary battery charging method based on SOC optimization according to claim 1, characterized in that, In step S5, the process of calculating the estimated value of the negative electrode surface potential includes: A negative electrode potential estimation model based on a single-particle model architecture is established. By using the corrected state of charge, the negative electrode potential estimation model retrieves the preset open-circuit potential curves of the negative electrode material in the memory. Through linear interpolation logic, the potential values ​​corresponding to the corrected state of charge are matched from the open-circuit potential curves of the negative electrode material to determine the negative electrode equilibrium potential. The negative electrode potential estimation model retrieves the preset ideal gas constant, charge transfer coefficient, and Faraday constant from the memory, and uses the absolute temperature obtained from the preprocessed ambient temperature data to determine the kinetic constant term; The preset negative electrode active area and exchange current density are retrieved from the memory, and the electrochemical overpotential is calculated by using the kinetic constant term and the preprocessed charging current data at the current moment through the transformation function of the Butler-Folmer equation. The estimated value of the negative electrode surface potential is obtained by subtracting the electrochemical overpotential from the negative electrode equilibrium potential, and by subtracting the product of the preprocessed charging current data and the dynamic polarization resistance at the current moment.

7. The secondary battery charging method based on SOC optimization according to claim 1, characterized in that, In step S6, the process of determining the maximum allowable charging current command and the current limit dynamic adjustment command includes: Retrieve the preset charging range, safety potential threshold, and proportional-integral-derivative adjustment coefficient from the memory, compare the corrected state of charge with the preset charging range, and compare the estimated value of the negative electrode surface potential with the safety potential threshold. Under the condition that the state of charge is within the preset charging range and the estimated value of the negative electrode surface potential is not lower than the safe potential threshold, the maximum allowable charging current command is output. Under the condition that the state of charge is in the preset charging range and the estimated value of the negative electrode surface potential is lower than the safe potential threshold, the algebraic difference between the safe potential threshold and the estimated value of the negative electrode surface potential is calculated and determined as the potential deviation value. The proportional, integral and derivative adjustment coefficients are used to perform proportional, integral and derivative algebraic operations on the potential deviation value to determine the current adjustment increment in amperes. The current adjustment increment is subtracted from the preprocessed charging current data of the previous moment to determine the charging current limit at the current moment, and the current limit dynamic adjustment command is output. When the state of charge exceeds the preset charging range, a charging termination command is output.

8. The secondary battery charging method based on SOC optimization according to claim 1, characterized in that, In S6, the process of performing lithium plating risk warning includes: Subtract the estimated negative electrode surface potential from the previous sampling time from the current estimated negative electrode surface potential, and divide by the sampling period to obtain the rate of change of the estimated negative electrode surface potential over time. The preset lithium plating change rate threshold is retrieved from the memory. The rate of change of the estimated negative electrode surface potential over time is compared with the lithium plating change rate threshold. When the rate of change of the estimated negative electrode surface potential over time exceeds the lithium plating change rate threshold, a lithium plating risk warning signal is triggered.

9. The secondary battery charging method based on SOC optimization according to claim 1, characterized in that, Following S6, it also includes: S7. Present a three-dimensional charging characteristic curve consisting of the corrected state of charge, actual usable capacity, and estimated negative electrode surface potential through a graphical interface: Construct a three-dimensional spatial coordinate system in the graphical interface, map the corrected state of charge to the first axis coordinate, the actual usable capacity to the second axis coordinate, and the estimated negative electrode surface potential to the third axis coordinate; perform surface fitting calculations based on continuously sampled coordinate point data to generate a three-dimensional charging characteristic curve consisting of the corrected state of charge, actual usable capacity, and estimated negative electrode surface potential in real time.

Images