Charging method for balancing OCV voltage consistency of energy storage battery

Abstract

The invention relates to the technical field of energy storage battery manufacturing, in particular to a charging method for balancing OCV voltage consistency of an energy storage battery, which comprises the following steps: S1, dynamic health state characterization: placing a to-be-balanced lithium iron phosphate battery pack in an environment with the temperature controlled at 20-30 DEG C, and pairing the OCV of each monomer with the calculated dynamic DC internal resistance DCR; according to the method, the dynamic direct-current internal resistance of each single battery is measured by applying the short-time high-rate discharge pulse at the initial stage of equalization processing, and the dynamic direct-current internal resistance is combined with the initial open-circuit voltage, so that a battery health state portrait can be established. On the basis of the portrait, the batteries are automatically grouped by utilizing a clustering algorithm, and customized multi-stage discharge curves are matched for different groups, so that differentiated and refined management of the batteries is realized, excessive electrochemical stress on the batteries with high internal resistance or poor states is avoided, and the safety and reliability of the whole process are improved.

Description

Technical Field

[0001] This invention relates to the field of energy storage battery manufacturing technology, and in particular to a method for replenishing energy storage batteries to balance the OCV voltage consistency. Background Technology

[0002] With the rapid development of energy storage batteries, the consistency of individual cell performance directly determines the matching of the entire battery pack when batteries are assembled into a cluster during the PACK process. In battery manufacturing, the open-circuit voltage (OCV) consistency of individual cells is a key indicator affecting the overall performance of the assembled pack. Unstable voltage consistency can lead to decreased battery pack capacity, shortened cycle life, and even safety hazards such as thermal runaway. Therefore, voltage consistency equalization of battery packs before shipment has become a core critical aspect of the industry. The charging process, as the final capacity replenishment step in the lithium-ion battery PACK supply, directly determines the voltage consistency during subsequent battery assembly. Charge and discharge tests are conducted using a charging cabinet to precisely replenish battery capacity (30% SOC) and eliminate cells with substandard capacity. Furthermore, batteries are categorized according to parameters such as capacity, OCV voltage, and internal resistance (e.g., L1 / L2 / L3 categories) to ensure consistent performance and good voltage consistency within the same battery pack.

[0003] Current technologies for equalizing energy storage batteries have limitations. They employ a standardized, open-loop, fixed procedure to treat all batteries indiscriminately. This ignores the performance inconsistencies between individual battery cells caused by microscopic differences in manufacturing processes and materials, particularly subtle differences in internal resistance and capacity. In practice, when a batch of batteries is subjected to the same constant-current charge-discharge procedure, batteries with higher internal resistance exhibit more drastic voltage responses. For example, during constant-current charging, the terminal voltage of a battery with higher internal resistance rises to the charging cutoff voltage faster than that of a battery with lower internal resistance. This causes the system to misjudge that the battery is fully charged and terminate charging prematurely, even though the actual amount of charge and state of charge (SOC) of the higher-resistance battery are lower than those of the lower-resistance battery. Therefore, improvements are needed. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and propose a method for equalizing the OCV voltage consistency of energy storage batteries.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a method for equalizing the OCV voltage consistency of an energy storage battery, comprising the following steps:

[0006] S1, Dynamic health status characterization, the lithium iron phosphate battery pack to be balanced is placed in an environment with the temperature controlled at 20-30℃, and the OCV of each cell is paired with the calculated dynamic DC internal resistance DCR to obtain the initial health status dataset.

[0007] S2, Based on SOH adaptive grouping and strategy matching, the K-means clustering algorithm is applied to the initial health state dataset to obtain battery arrays with group labels;

[0008] S3, Differentiated stepped discharge: The battery array with grouped tags is connected to the battery charging and discharging device for discharge to obtain a deeply discharged battery pack.

[0009] S4. Based on the discharge cutoff judgment of the voltage plateau slope, during the third stage current discharge of the battery pack after the deep discharge, the voltage change rate is continuously monitored to obtain the battery pack that has reached the discharge benchmark.

[0010] S5, Electrochemical equilibrium adaptive static setting: For the battery pack that has reached the discharge benchmark, disconnect all charging and discharging circuits, determine whether the battery has reached electrochemical equilibrium, and obtain an electrochemically stable battery pack.

[0011] S6. Based on the closed-loop precise charging of the OCV-SOC model, all cells of the electrochemically stable battery pack are charged to obtain a preliminary balanced battery pack.

[0012] S7, fine-tune pulse voltage calibration, and perform static and equalization charging on the battery pack after the main charging is completed to obtain a high-precision equalized battery pack.

[0013] S8, Final inspection of all parameters and data traceability: After completing all calibration steps and letting it stand for 1-2 hours, the high-precision balanced battery pack is subjected to step S1 again to measure and record the OCV and DCR values, and a finished balanced battery with a full life cycle traceability file is obtained.

[0014] Preferably, step S1 specifically involves: placing the lithium iron phosphate battery pack to be equalized in an environment with a temperature controlled at 20-30℃; applying a constant current discharge pulse of 0.8C-1.5C to each individual cell in the battery pack, with a pulse duration of 8-15 seconds; during this period, collecting voltage data at a frequency of 100-500Hz; after the pulse ends, calculating the dynamic DC internal resistance of each individual cell and recording the initial open-circuit voltage OCV before the pulse; pairing the OCV of each individual cell with the calculated dynamic DC internal resistance DCR value to obtain the initial health state dataset; the formula for calculating the dynamic DC internal resistance is: Among them, DCR j V represents the dynamic DC internal resistance of the j-th battery. pre,j V represents the average voltage of the j-th battery in the 2 seconds before the pulse is applied. pulse,j I represents the voltage of the j-th battery at the instant the pulse ends. pulse T represents the pulse discharge current. amb T represents ambient temperature.ref The reference temperature is 25°C, α represents the temperature correction factor for internal resistance, β represents the weighting factor for voltage stability, and σ represents the reference temperature. v,pre This represents the standard deviation of the voltage of the j-th battery within 2 seconds before the pulse is applied.

[0015] Preferably, step S2 specifically involves: applying the K-means clustering algorithm to the initial health state dataset, using the initial OCV and dynamic DC internal resistance as two-dimensional feature vectors, dividing all individual cells into 3-5 cluster groups, and matching a dedicated set of discharge and charge process parameters for each cluster group to obtain a battery array with group labels.

[0016] Preferably, step S3 specifically involves: connecting the battery array with grouped tags to the battery charging and discharging device; calling the corresponding discharge process parameter set to perform three-stage discharge according to the allocation tags of each battery in S2; and obtaining the battery pack after deep discharge.

[0017] Preferably, step S4 specifically involves: during the third stage of current discharge of the battery pack after deep discharge, calculating the voltage change rate of each individual cell in real time at intervals of 3-8 seconds, continuously monitoring the voltage change rate, and determining that the core discharge plateau period of the battery has ended when the voltage drop rate of any individual cell exceeds a preset threshold of 15-25mV / min within a continuous calculation cycle, thus obtaining a battery pack that has reached the discharge benchmark.

[0018] Preferably, step S5 specifically involves: disconnecting all charging and discharging circuits for the battery pack that has reached the discharge benchmark, entering an open-circuit resting state, continuously measuring the open-circuit voltage of each cell at intervals of 30-90 seconds, calculating the voltage rebound rate in real time, and determining that the battery has reached electrochemical equilibrium when the voltage rebound rate of a single cell remains below the stable threshold of 0.03-0.08 mV / s for a continuous time window of 90-180 seconds. This results in an electrochemically stable battery pack.

[0019] Preferably, step S6 specifically involves: charging all cells of the electrochemically stable battery pack with a current of 0.45C-0.55C; interrupting charging after every 240-360 seconds and allowing it to rest for 8-15 seconds; measuring the quasi-open-circuit voltage of each cell at the end of the resting period; and calculating and estimating the state of charge (SOC) by combining this with the real-time measured battery surface temperature. est,k When the estimated state of charge (SOC) of any battery reaches the target range of 28%-32%, charging is immediately stopped to obtain a preliminarily balanced battery pack. The formula for calculating the estimated SOC is as follows: Among them, SOC est,k Q represents the estimated state of charge of the k-th battery. ocv,ka represents the quasi-open-circuit voltage measured during the charging interval of the k-th battery. n The fitting coefficients represent the basic curve of the nth-order polynomial OCV-SOC, γ represents the temperature compensation coefficient of the voltage, and T represents the fitting coefficient of the basic curve of the basic curve of the polynomial OCV-SOC. cell,k T represents the real-time temperature of the k-th battery. ref The reference temperature is 25°C, δ represents the correction weighting factor for the individualized internal resistance effect, and DCR k represents the dynamic DC internal resistance value of the k-th battery obtained from S1, and N represents the order of the polynomial used to fit the OCV-SOC basic curve.

[0020] Preferably, step S7 specifically involves: after completing the main charging, the initially balanced battery pack is allowed to rest for 30-60 minutes, and the stable OCV of each cell is measured and compared with the standard voltage value corresponding to the target estimated state of charge. If the absolute value of the voltage difference is greater than 0.5mV, a current pulse of 0.008C-0.015C is applied to replenish the charge. Each pulse lasts for 1-5 seconds, and the pulse interval is 30-60 seconds to wait for the voltage to stabilize until the difference between the OCV of all cells and the standard voltage is less than 0.5mV, thus obtaining a high-precision balanced battery pack.

[0021] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0022] This invention establishes a battery health profile by applying a short-duration, high-rate discharge pulse at the initial stage of equalization processing to measure the dynamic DC internal resistance of each individual cell and combining it with the initial open-circuit voltage. Based on this profile, a clustering algorithm is used to automatically group the batteries, and customized multi-stage discharge curves are matched to different groups. This enables differentiated and refined management of the batteries, avoiding excessive electrochemical stress on batteries with high internal resistance or poor condition, and improving the safety and reliability of the entire process. During the discharge stage, a fixed voltage cutoff point is abandoned, and the discharge endpoint is determined by real-time monitoring of the voltage change rate. This ensures that all batteries reach their true and uniform electrochemical discharge reference point. Furthermore, the determination of electrochemical equilibrium no longer relies on a fixed period of rest, but rather on monitoring the voltage rebound rate until it stabilizes at a threshold, ensuring that each battery reaches a true internal steady state. In the core recharging stage, a method of periodically interrupting charging and measuring the quasi-open-circuit voltage is adopted, combined with an OCV-SOC model that simultaneously considers temperature and individualized internal resistance for closed-loop control, achieving guidance of the target state of charge. This improves the battery pack's power output, overall capacity utilization, and cycle life. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the steps of the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0025] Please see Figure 1 This invention provides a technical solution, a method for equalizing the OCV voltage consistency of an energy storage battery, comprising the following steps:

[0026] S1, Dynamic health status characterization, the lithium iron phosphate battery pack to be balanced is placed in an environment with the temperature controlled at 20-30℃, and the OCV of each cell is paired with the calculated dynamic DC internal resistance DCR to obtain the initial health status dataset.

[0027] S2, based on SOH adaptive grouping and policy matching, uses the K-means clustering algorithm on the initial health state dataset to obtain battery arrays with group labels;

[0028] S3, Differentiated stepped discharge, connects the battery array with grouped tags to the battery charging and discharging device for discharge, and obtains a battery pack after deep discharge;

[0029] S4, based on the discharge cutoff judgment of the voltage plateau slope, continuously monitors the voltage change rate during the third stage current discharge of the battery pack after deep discharge to obtain the battery pack that has reached the discharge benchmark.

[0030] S5, Electrochemical Equilibrium Adaptive Static Conditioning: For a battery pack that has reached the discharge benchmark, disconnect all charging and discharging circuits, determine whether the battery has reached electrochemical equilibrium, and obtain an electrochemically stable battery pack.

[0031] S6, based on the closed-loop precise charging of the OCV-SOC model, charges all cells in the electrochemically stable battery pack to obtain a battery pack with preliminary equilibrium.

[0032] S7, fine-tuning pulse voltage calibration, after the initial equalization of the battery pack, performs static and equalization charging to obtain a high-precision equalized battery pack.

[0033] S8, full parameter final inspection and data traceability, for high-precision balanced battery packs, after completing all calibration steps and letting it stand for 1-2 hours, step S1 is performed again to measure and record the OCV and DCR values, resulting in a finished balanced battery product with a full life cycle traceability file.

[0034] Step S1 is as follows: Place the lithium iron phosphate battery pack to be equalized in an environment with a temperature controlled at 20-30℃. Apply a constant current discharge pulse of 0.8C-1.5C to each individual cell in the battery pack. The pulse duration is 8-15 seconds. During this period, voltage data is collected at a frequency of 100-500Hz. After the pulse ends, calculate the dynamic DC internal resistance of each individual cell and record the initial open-circuit voltage OCV before the pulse. Pair the OCV of each cell with the calculated dynamic DC internal resistance DCR value to obtain the initial health state dataset. The formula for calculating the dynamic DC internal resistance is: Among them, DCR j V represents the dynamic DC internal resistance of the j-th battery. pre,j V represents the average voltage of the j-th battery in the 2 seconds before the pulse is applied. pulse,j I represents the voltage of the j-th battery at the instant the pulse ends. pulse T represents the pulse discharge current. amb T represents ambient temperature. ref The reference temperature is 25°C, α represents the temperature correction factor for internal resistance, β represents the weighting factor for voltage stability, and σ represents the reference temperature. v,pre This represents the standard deviation of the voltage of the j-th battery within 2 seconds before the pulse is applied.

[0035] Specifically, the formula: The advantage of the formula lies in its integration of two key correction terms: the temperature correction term [1+α(T)]. amb -T ref )] and voltage stability correction term β·σ v,pre Temperature correction is used to eliminate the interference of ambient temperature changes on internal resistance measurement results, normalizing the internal resistance values ​​measured at different temperatures to a unified reference temperature (T). ref Under this benchmark, the comparability and accuracy of internal resistance data across different batches and test times are ensured. The voltage stability correction term incorporates consideration of the battery's electrochemical stability state, taking into account small fluctuations in the pre-pulse voltage (σ). v,pre This can reflect the degree of instability factors such as internal side reactions and self-discharge of the battery. By assigning a weighting coefficient β, the calculated DCR is made more accurate. j It's not just the impedance of charge transport.

[0036] V pre,jThis represents the average voltage of the j-th battery in volts (V) over the 2 seconds prior to applying the discharge pulse. This parameter is acquired by a high-frequency data acquisition system during the resting period before the discharge pulse is applied. Specifically, within a 2-second time window before the pulse command is issued, the voltage acquisition module continuously records the voltage values ​​across the battery terminals at a frequency of at least 100Hz, forming a time-series voltage dataset containing at least 200 data points. Subsequently, by averaging all voltage readings in this dataset, a stable voltage value representing the initial state before the pulse is obtained. This averaging process effectively filters out transient noise caused by electromagnetic interference from the measuring equipment or the environment, providing a more reliable reference voltage than single-time-point sampling. Example: Before applying a pulse to battery number 5 (j=5), the system acquires 1000 voltage points at a frequency of 500Hz over 2 seconds, with the data sequence being V1, V2, ..., V... 1000 Its value fluctuated between 3.3048V and 3.3053V, and its average value was calculated to obtain

[0037] V pulse,j This represents the voltage of the j-th battery at the instant the pulse ends, expressed in volts (V). Accurate acquisition of this parameter relies on high-time-resolution voltage acquisition capabilities. Throughout the entire cycle of the constant-current discharge pulse, the voltage acquisition system operates continuously. When the pulse reaches a preset duration (e.g., 15 seconds) and is about to end, the system captures the voltage reading at the last or closest point to the end as V. pulse,j This value reflects the battery's polarization voltage under a specific current load rate and is the core voltage drop component for calculating internal resistance. To ensure accuracy, the sampling frequency should be high enough (e.g., 500Hz) to capture the true voltage at the moment the pulse cuts off, avoiding reading the voltage that begins to recover after the pulse ends due to data acquisition delay. Example: A 15-second discharge pulse is applied to battery number 5 (j=5). At time t=15.000 seconds, the voltage acquisition card records a voltage value of 3.1558V, which is adopted as V. pulse,5 .

[0038] I pulse This represents the pulse discharge current, measured in amperes (A). This parameter is precisely controlled and output by the battery charging and discharging device according to a preset program. Its value is based on the battery's rated capacity (C). rated It is calculated based on the rated capacity of a battery (100Ah) and the set discharge rate (C-rate). For example, for a battery with a rated capacity of 100Ah, if a 1.2C discharge pulse is applied, then I... pulseThe current is set to 1.2 × 100 Ah = 120 A. Throughout the pulse duration, the closed-loop control system of the charging / discharging device monitors and adjusts the output in real time to ensure current constancy, with fluctuations typically controlled within ±0.1% of the set value. This set value is directly provided to the calculation module by the test program and is a precise input parameter. Example: For a lithium iron phosphate battery with a rated capacity of 280 Ah, the test procedure requires applying a 0.8C discharge pulse; the constant discharge current set by the charging / discharging device is I. pulse =0.8×280=224A.

[0039] T amb Represents the ambient temperature, expressed in degrees Celsius (°C). This parameter is measured using multiple digital temperature sensors (such as PT100 or NTC thermistors) deployed within the temperature-controlled test chamber. To ensure the representativeness of the measurements, the sensors are typically placed at different locations around the battery pack (top, middle, bottom, left, right) to monitor the air temperature within the chamber in real time. The control system collects data from all sensors and calculates their average as the final ambient temperature T. amb The control system of the constant temperature chamber automatically adjusts the heating or cooling system based on the deviation of this average value from the set target value (20-30℃), keeping temperature fluctuations within ±1℃. Example: During the test, the readings of the three temperature sensors inside the chamber were 26.2℃, 26.5℃, and 26.3℃, respectively. The system calculates the current ambient temperature as T. amb = (26.2 + 26.5 + 26.3) / 3 = 26.33℃.

[0040] T ref This represents a reference temperature of 25°C, expressed in degrees Celsius (°C). It is a standardized industry reference value used for calibrating and comparing battery performance parameters. In battery research and development and production, 25°C is typically defined as the standard test temperature. Performance parameters (such as internal resistance and capacity) measured at different temperatures need to be corrected using a temperature coefficient to convert them to their equivalent values ​​at this reference temperature. This provides a unified benchmark for comparison of test results under different environmental conditions. This parameter is used as a fixed constant in the formula and does not require measurement; its value is preset to 25.

[0041] α represents the temperature correction factor for internal resistance, expressed in degrees Celsius (°C). This factor characterizes the sensitivity of the battery's internal resistance to temperature changes, and its value needs to be obtained through experimental calibration. The calibration process is as follows: Select a batch of reference batteries of the same model, and measure their uncorrected baseline DC internal resistance (DCR) at multiple different stable ambient temperature points (e.g., 15°C, 20°C, 25°C, 30°C, 35°C). base =(V pre -V pulse) / I pulse Subsequently, the DCR was measured with temperature as the x-axis. base Plot a scatter plot with the vertical axis as the ordinate. Perform linear regression on these data points to obtain a line in the form of DCR. base (T)=DCR ref ·(1+α·(TT ref The value of α can be calculated by fitting the slope and intercept of the line. Example: The basic internal resistance of a battery was experimentally measured to be 2.15mΩ, 2.00mΩ, and 1.86mΩ at 15℃, 25℃, and 35℃, respectively. Using 25℃ as the baseline, a system of equations can be established to solve for α. For example, using the data at 35℃: 1.86=2.00·(1+α·(35-25)), solving for α gives α=-0.007℃-1.

[0042] β represents the weighting coefficient for voltage stability and is dimensionless. This coefficient quantifies the contribution of pre-pulse voltage instability to dynamic DC internal resistance, and its setting is based on the correlation analysis of a large amount of long-term battery aging data and initial characteristic parameters. The setting process is as follows: First, a database containing hundreds of battery samples is collected, which includes the σ measured for each battery in the initial stage. v,pre The values ​​and the capacity decay rate or internal resistance growth rate measured after long-term cycling (e.g., 1000 cycles) are then used to establish a multiple linear regression model with the capacity decay rate or internal resistance growth rate as the dependent variable and the base internal resistance and σ as the variables. v,pre The independent variable is c0 + c1 * DCR. The model form is: Aging index = c0 + c1 * DCR base +c2·σ v,pre The coefficients c1 and c2 obtained through regression analysis reflect the relative impact of basic internal resistance and voltage instability on aging. The weighting coefficient β is set proportional to... Its specific value is determined by scaling this ratio to a range that matches the dimensions of internal resistance. Example: Regression analysis yields c1 = 50.0, c2 = 15000.0. The relative influence ratio is 300. To make β·σ v,pre The magnitude of the value matches the internal resistance (mΩ), so β ​​is set to 0.3 because σ v,pre The unit of V is V, while the unit of internal resistance is mΩ, so dimensional adjustment is required.

[0043] σ v,pre This represents the standard deviation of the voltage of the j-th battery in the 2 seconds before the pulse is applied, in volts (V). It is a standard deviation of the voltage of the j-th battery in V. pre,j The voltage time series data (V1, V2, ..., V) used in the calculation process n The dispersion of a dataset is a statistical measure. Calculating this value requires first calculating the mean of the data set. (i.e. V) pre,jThen apply the definition of standard deviation: Where n is the total number of data points within the time window. A smaller σ v,pre A value indicates that the battery voltage is very stable when at rest, while a larger value may suggest unstable electrochemical activity inside, such as a high self-discharge rate or side reactions. Example: For battery number 5, in calculating V... pre,5 When the voltage was 3.3051V, 1000 voltage readings were used. The standard deviation was calculated using the above formula for these 1000 data points, yielding the voltage standard deviation σ. v,pre,5 =0.00012V.

[0044] Calculation process:

[0045] Taking battery number 5 as an example, based on the aforementioned parameter acquisition steps, the following specific values ​​are obtained:

[0046] V pre,5 =3.3051V;

[0047] V pulse,5 =3.1558V;

[0048] I pulse =224A;

[0049] T amb =26.33℃;

[0050] T ref =25℃;

[0051] α = -0.007℃-1

[0052] β = 0.3;

[0053] σ v,pre,5 =0.00012V;

[0054] The calculation process is as follows:

[0055] 1. Calculate the basic DC internal resistance:

[0056]

[0057] 2. Calculate the temperature correction factor:

[0058] =1+α(T) amb -T ref )=1+(-0.007)·(26.33-25)=1-0.00931=0.99069;

[0059] 3. Temperature correction for basic internal resistance:

[0060] =0.6665mΩ·0.99069=0.6603mΩ;

[0061] 4. Calculate the voltage stability penalty term:

[0062] =β·σ v,pre,5 =0.3·0.00012V=0.000036;

[0063] Note the unit mismatch issue here; the setting for β should include unit conversion. In practical applications, the unit of β should be mΩ / V to match σ. v,pre The product is multiplied in mΩ. For example, if β is set to 300 mΩ / V, then:

[0064] =300mΩ / V·0.00012V=0.036mΩ;

[0065] 5. Calculate the final dynamic DC internal resistance DCR5:

[0066] DCR5=0.6603mΩ+0.036mΩ=0.6963mΩ;

[0067] The result indicates that the dynamic DC internal resistance (DCR) of battery number 5 is 0.6963 mΩ. This value comprehensively reflects its charge transport capability at a standard temperature of 25°C and its electrochemical stability. A lower DCR value (e.g., below 0.7 mΩ) generally indicates good battery health, low internal resistance, and stable internal chemistry. A significantly higher calculated value (e.g., above 1.0 mΩ) suggests potential issues such as aging, high internal resistance, or internal instability. This DCR... j The value will be paired with the battery's initial OCV value to form a two-dimensional health state feature vector (OCV, DCR), which will serve as the input data for the K-means clustering algorithm in the subsequent S2 step, used to accurately group the batteries by health state.

[0068] Step S2 is as follows: For the initial health state dataset, the K-means clustering algorithm is used to divide all individual cells into 3-5 cluster groups with the initial OCV and dynamic DC internal resistance as two-dimensional feature vectors. A dedicated set of discharge and charge process parameters is matched for each cluster group to obtain a battery array with group labels.

[0069] Specifically, for the initial health state dataset, the initial OCV and dynamic DC internal resistance data of all batteries in the dataset are first subjected to min-max normalization, mapping their values ​​to the [0,1] interval. The calculation formula is X. norm =(XX) min ) / (X max -X min ), where X is the original data, Xmin and X max These are the minimum and maximum values ​​among all data in this dimension. This avoids the clustering algorithm from over-relying on one feature due to excessive differences in the dimensions and numerical ranges of the two features. Subsequently, the K-means clustering algorithm is used to process the normalized two-dimensional feature vector. Before formal clustering, the elbow method is used to determine the optimal number of clusters K. Specifically, the sum of squares within each cluster (SSE) is calculated sequentially for K values ​​of 2, 3, 4, 5, 6, and 7, and the relationship curve between K and SSE is plotted. When K increases from 3 to 4, the rate of decrease in SSE slows significantly, forming a... The obvious "elbow" effect led to the determination of an optimal number of clusters: 4. Next, four battery data points were randomly selected from the dataset as the initial four cluster centers. The process then proceeded iteratively. In the allocation step, the Euclidean distance from each battery's data point to these four cluster centers was calculated, and the battery was assigned to the cluster containing the nearest cluster center. In the update step, the centroid of each cluster was recalculated, i.e., the average of the components of the two-dimensional feature vectors of all batteries within the cluster was taken as the new cluster center. This allocation and update step was repeated until... If the sum of the distances the cluster centers move in two consecutive iterations is less than the preset convergence threshold of 0.0001, or if the maximum number of iterations of 200 is reached, clustering is completed. Then, feature analysis is performed on the four resulting cluster groups. For example, Group 1, exhibiting low DCR and high OCV, is labeled "High Health, High Battery"; Group 2, with low DCR and low OCV, is labeled "High Health, Low Battery"; Group 3, with high DCR and medium-high OCV, is labeled "Moderate Aging"; and Group 4, with high DCR and low OCV, is labeled "Severe Aging". Finally, for each of the four groups, a unique and differentiated set of discharge and charge process parameters is matched. For example, for the "severe aging" group, the discharge process parameter set is set to a smaller discharge current (e.g., 0.5C), a higher discharge cutoff voltage (e.g., 2.8V), and a longer inter-stage rest time (e.g., 30 minutes), while the "high health" group can use more aggressive parameters (e.g., 1.0C discharge current, 2.5V cutoff voltage). These parameter sets are stored in a structured manner and associated with group tags, ultimately resulting in a battery array with group tags.

[0070] Step S3 specifically involves connecting the battery array with grouped tags to the battery charging and discharging device, and calling the corresponding discharge process parameter set to perform three-stage discharge based on the allocation tags of each battery in S2, thereby obtaining the battery pack after deep discharge.

[0071] Specifically, the battery array with group tags is connected to the battery charging and discharging device. The central control system of the device first reads the pre-assigned group tags of each individual battery in the array through the communication bus, such as "high health, high capacity" and "moderate aging". Then, based on the read tags, the system calls the corresponding discharge process parameter set for each channel from the built-in process parameter database. This parameter set defines the specific procedures for the three-stage discharge. The first stage is the constant current main discharge stage. The system discharges according to the first stage current value set in the parameter set (for example, setting a discharge current of 1.0C for the battery channel of the "high health" group and a discharge current of 0.5C for the battery channel of the "severely aged" group) and continuously monitors the real-time voltage of each battery. When the voltage of any battery drops to the first stage cutoff voltage set in the parameter set (for example, ... When the battery voltage reaches 3.0V, the channel automatically switches to the second stage, the constant current supplementary discharge stage. At this time, the system will call a lower second stage current value set in the parameter set (e.g., 0.3C for the "high health" group and 0.15C for the "severe aging" group) to continue discharging until the battery voltage reaches the cutoff voltage of the second stage (e.g., 2.8V). Then, the channel enters the third stage, the trickle discharge stage, using a very small constant current (e.g., 0.05C for all groups) to discharge. The purpose of this stage is to safely deplete the battery to a near-completely discharged state. Throughout the process, the control system independently controls and monitors each channel to ensure that the batteries in different groups strictly follow the differentiated strategies tailored to them for discharge. Finally, when all batteries have completed their respective three-stage discharge processes, a deeply discharged battery pack is obtained.

[0072] Step S4 is as follows: During the third stage of current discharge of the battery pack after deep discharge, the voltage change rate of each individual cell is calculated in real time at intervals of 3-8 seconds. The voltage change rate is continuously monitored. When the voltage drop rate of any individual cell exceeds the preset threshold of 15-25mV / min within a continuous calculation cycle, it is determined that the core discharge plateau period of the battery has ended, and the battery pack that has reached the discharge benchmark is obtained.

[0073] Specifically, during the third stage of current discharge of the battery pack after deep discharge, the system initiates real-time calculation and monitoring of the voltage change rate of each individual cell at fixed intervals of 5 seconds. The specific calculation method is to record the current voltage value V at each sampling point. current and compared with the voltage value V recorded 5 seconds ago. previous Comparisons were made using the formula: voltage change rate = (V previous -V currentThe instantaneous voltage drop rate, expressed in V / s, is obtained by multiplying the result by 60 over 5 seconds to convert it to mV / min. Simultaneously, a preset voltage drop rate threshold is set to 20 mV / min. This threshold is based on the following analysis: Statistical analysis of voltage curves from over 500 lithium iron phosphate batteries of the same model during complete discharge revealed that at the "inflection point" where the voltage drops sharply from a flat plateau at the end of discharge, the average voltage drop rate is 19.8 mV / min, with a standard deviation of 3.5 mV / min. To ensure reliable capture of the end of the plateau period with a certain safety margin, and to avoid misjudgments due to occasional voltage fluctuations, the threshold is set close to the average. If the value is slightly higher than 20mV / min, the system will maintain a queue for each individual cell containing the voltage drop rate values ​​of the most recent three calculation cycles (i.e., within 15 seconds). After each calculation cycle, the system checks the queue. When three consecutive values ​​in the queue are greater than 20mV / min, for example, the rates calculated at t seconds, t+5 seconds, and t+10 seconds are 21.5mV / min, 24.3mV / min, and 28.1mV / min, respectively, the system determines that the core discharge plateau period of that individual cell has ended and marks it as having reached the discharge benchmark. This judgment logic will be applied independently to each individual cell in the battery pack until all cells are marked, finally resulting in a battery pack that has reached the discharge benchmark.

[0074] Step S5 is as follows: For the battery pack that has reached the discharge benchmark, disconnect all charging and discharging circuits and enter the open circuit resting state. Continuously measure the open circuit voltage of each cell at intervals of 30-90 seconds, calculate the voltage rebound rate in real time, and when the voltage rebound rate of a single cell is continuously lower than the stable threshold of 0.03-0.08mV / s for a continuous time window of 90-180 seconds, it is determined that the battery has reached electrochemical equilibrium, and an electrochemically stable battery pack is obtained.

[0075] Specifically, for battery packs that have reached the discharge benchmark, the control system immediately executes a command to disconnect all connected charging and discharging circuits, putting the entire battery pack into a completely open-circuit resting state. Starting from the moment the circuits are disconnected, the system continuously measures and records the open-circuit voltage of each cell at fixed 60-second intervals. After each measurement, the system calculates the voltage rebound rate since the last measurement in real time. The calculation method is to subtract the open-circuit voltage value from 60 seconds ago from the current open-circuit voltage value, and then divide by 60 seconds to obtain the rate value in mV / s. The key to this process is determining when the battery reaches internal electrochemical equilibrium. For this purpose, the system sets a stability threshold of 0.05 mV / s. This threshold is determined by experimental analysis of the battery relaxation effect: by fitting the voltage rebound curves of a large number of batteries after removing the load, it was found that the rebound process conforms to a double exponential decay model, representing electrochemical polarization and concentration. The polarization dissipation process, according to calculations, shows that when the voltage rebound rate is below 0.05 mV / s, the ion concentration gradient inside the battery is basically balanced, and further changes in the electrochemical state are extremely slow, which can be considered as reaching stability. To ensure the robustness of the judgment and avoid misjudgment caused by instantaneous disturbances, the system uses a continuous time window of 120 seconds for evaluation. That is, the system checks whether the calculated voltage rebound rate is below 0.05 mV / s in two consecutive 60-second measurement intervals. For example, if the rate of a single cell is 0.048 mV / s in the interval from 300 seconds to 360 seconds and 0.045 mV / s in the interval from 360 seconds to 420 seconds, the system determines that the single cell has reached electrochemical equilibrium and marks it as a stable state. This judgment process is executed in parallel for all cells until all cells obtain the stable state mark, thus obtaining an electrochemically stable battery pack.

[0076] Step S6 specifically involves charging all cells of the electrochemically stable battery pack at a current of 0.45C-0.55C. After each 240-360 seconds of charging, the charging is interrupted and the cells are allowed to rest for 8-15 seconds. At the end of the resting period, the quasi-open-circuit voltage of each cell is measured, and the state of charge (SOC) is estimated by combining this with the real-time measured battery surface temperature. est,k When the estimated state of charge (SOC) of any battery reaches the target range of 28%-32%, charging is immediately stopped to obtain a preliminarily balanced battery pack. The formula for calculating the estimated SOC is as follows: Among them, SOC est,k Q represents the estimated state of charge of the k-th battery. ocv,k a represents the quasi-open-circuit voltage measured during the charging interval of the k-th battery. n The fitting coefficients represent the basic curve of the nth-order polynomial OCV-SOC, γ represents the temperature compensation coefficient of the voltage, and T represents the fitting coefficient of the basic curve of the basic curve of the polynomial OCV-SOC. cell,k T represents the real-time temperature of the k-th battery.ref The reference temperature is 25°C, δ represents the correction weighting factor for the individualized internal resistance effect, and DCR k represents the dynamic DC internal resistance value of the k-th battery obtained from S1, and N represents the order of the polynomial used to fit the OCV-SOC basic curve.

[0077] Specifically, the formula: The advantage of the formula lies in the basic model. Polynomial fitting was employed to accurately characterize the highly nonlinear relationship between OCV and SOC in lithium iron phosphate batteries within the target SOC range (28%-32%). Secondly, a temperature compensation term γ(T) was introduced. cell,k -T ref )·Q ocv,k The model dynamically corrects the OCV drift caused by the actual battery operating temperature deviating from the standard temperature, improving its adaptability and accuracy under different ambient temperatures. Most importantly, it incorporates a correction weight term δ·ln(1+DCR) to account for the individualized internal resistance effect. k Using the dynamic DC internal resistance (DCR) measured for each battery in step S1, k Using this as a quantitative indicator of its state of health (SOH), the SOC estimation is individually calibrated, which allows the model to take into account the differences in internal resistance caused by battery aging and manufacturing differences, thereby making a more accurate SOC judgment for each battery.

[0078] Q ocv,k This represents the quasi-open-circuit voltage of the k-th battery measured during the charging interval, in volts (V). After charging at a 0.5C current for 300 seconds, the charging circuit is interrupted for 15 seconds. This parameter is obtained by averaging the voltage readings obtained during the last second of this 15-second rest period through high-frequency sampling (e.g., 100Hz). Because the rest period is short, this voltage value has not yet fully reached the true open-circuit voltage, but it has significantly eliminated most of the influence of polarization voltage; therefore, it is called the quasi-open-circuit voltage and is an effective input for quickly estimating SOC. Example: For battery number 10 (k=10), during a charging interval, 100 voltage points were collected between the 14.0 and 15.0 seconds of rest. The average value was 3.3095V. Then Q... ocv,10 =3.3095V.

[0079] a nThe coefficients represent the dimensionless fitting coefficients of the nth-order polynomial OCV-SOC baseline curve. These coefficients were obtained through OCV-SOC calibration experiments on a batch of standard lithium iron phosphate batteries. The experimental procedure was as follows: the standard batteries were fully charged to 100% SOC in a 25°C constant temperature chamber, and then discharged with a very small current (e.g., 0.02C). After each 1% of the rated capacity was discharged, the batteries were allowed to stand for 4 hours until the voltage was completely stable. The OCV and corresponding SOC at this point were recorded. All data points from 0% to 100% SOC were collected, and then the least squares method was used to fit a polynomial to these data points, with OCV as the independent variable and SOC as the dependent variable. To achieve high accuracy within the target range, an order N = 7 was chosen. Example: By fitting the calibration data with a 7th-order polynomial, a set of fitting coefficients may be obtained as follows: a0 = -15.65, a1 = 20.80, a2 = -11.54, a3 = 3.42, a4 = -0.58, a5 = 0.052, a6 = -0.0024, a7 = 0.000045.

[0080] γ represents the temperature compensation coefficient for voltage, with units of V⁻¹℃⁻¹. This coefficient is used to correct the effect of temperature on OCV. It is obtained by repeating the OCV-SOC calibration experiment at multiple different constant temperatures (e.g., 15℃, 25℃, 35℃) to obtain multiple sets of OCV-SOC curves. For the same SOC point, comparing the OCV values ​​at different temperatures reveals that OCV changes approximately linearly with temperature. By calculating ΔOCV / ΔT and correlating it with OCV itself, a temperature compensation model is obtained. The linear compensation coefficient γ can be calculated using the following formula: Where i represents different SOC points, and M is the number of selected SOC points. Example: Select 10 SOC points within the 20%-40% SOC range, compare the OCV values ​​at 15℃ and 35℃, and calculate the average temperature compensation coefficient γ = -0.00015V-1℃-1.

[0081] T cell,k This represents the real-time temperature of the k-th battery, in degrees Celsius (°C). This parameter is measured in real-time by an NTC thermistor attached to the surface of each individual battery cell. These sensors are connected to the battery management system's acquisition module via data cables, updating the temperature data at a frequency of at least 1 Hz. This is used in calculating the State of Charge (SOC). est,k At that time, the system will read and measure Q. ocv,k Temperature values ​​at the same point in time are used to ensure the immediacy and accuracy of temperature compensation. Example: Measuring the Q value of battery number 10. ocv,10 At that time, the surface temperature sensor reading was 27.5℃, then T cell,10 =27.5℃.

[0082] T refThis represents a reference temperature of 25°C, expressed in degrees Celsius (°C). This is a fixed reference standard, and all temperature-related compensations are calculated based on deviations from this temperature. This value is set to a constant of 25 in the algorithm and does not require measurement.

[0083] δ represents a dimensionless, dimensionless corrected weighting coefficient for the individualized internal resistance effect. This coefficient is used to quantify the impact of battery aging (characterized by increased internal resistance) on the OCV-SOC relationship. Its setup process is as follows: Select a dataset containing different health states (i.e., different DCRs). k A sample library of batteries (with values) was precisely adjusted to the same true SOC (e.g., 30%) at 25°C, and then their quasi-open-circuit voltage Q was measured. ocv,k Calculate the voltage deviation ΔV for each battery. k =Q ocv,k -Q ocv,ref Q ocv,ref This is the standard open-circuit voltage of a healthy battery at this SOC. Simultaneously, the DCR of each battery is recorded. k Value. For the data pair (ln(1+DCR) k ),ΔV k Performing linear regression analysis yields a slope that is an estimate of δ. Since this term in the formula is directly added to SOC, the unit of δ should be 1, representing a direct correction to SOC. Example: By conducting the above experiment on 100 batteries with different aging levels, performing linear regression analysis on the data, the resulting regression equation is: SOC deviation = δ·ln(1+DCR) k The fitting yielded δ = -0.85.

[0084] DCR k This represents the dynamic DC internal resistance value of the k-th battery obtained from step S1, in ohms (Ω). This value is a unique health status fingerprint for each battery, calculated and stored in step S1. In step S6, the system directly calls this value for personalized SOC estimation. Example: Based on the calculation in step S1, the dynamic DC internal resistance value of battery number 10 is DCR. 10 =0.82mΩ =0.00082Ω.

[0085] N represents the order of the polynomial used to fit the OCV-SOC baseline curve. This is a preset model parameter; for lithium iron phosphate batteries, a 7th-order polynomial (N=7) typically provides sufficient fitting accuracy within the SOC range of interest without introducing excessive oscillations. This value is set to a constant of 7 in the algorithm.

[0086] Substituting the parameters yielded a value of 0.2983, indicating that the current estimated state of charge (SOC) of battery #10 is 29.83%. This value is a comprehensive assessment combining its quasi-open-circuit voltage, real-time temperature, and individual health status (DCR). This result is correlated with the step objective: the system compares this value to the target range of 28%-32%. Because 29.83% falls within this range, the system immediately stops charging battery #10 and marks it as "preliminary equalization complete." This calculation and judgment will be performed independently for all batteries until the SOC of all batteries reaches its target value. est All cells fall within the target range, thus obtaining a preliminary balanced battery pack with highly consistent SOC across all cells.

[0087] Step S7 is as follows: After the initial equalization of the battery pack is completed, allow it to rest for 30-60 minutes. Measure the stable OCV of each cell and compare it with the standard voltage value corresponding to the target estimated state of charge. If the absolute value of the voltage difference is greater than 0.5mV, apply a current pulse of 0.008C-0.015C to replenish the charge. Each pulse lasts for 1-5 seconds, with a pulse interval of 30-60 seconds to allow the voltage to stabilize. Continue this process until the difference between the OCV of all cells and the standard voltage is less than 0.5mV, thus obtaining a high-precision equalized battery pack.

[0088] Specifically, for the initially balanced battery pack, after completing the main charge, the system controls all channels to enter a resting mode. This resting process is set to last 45 minutes. This duration is determined experimentally and is sufficient to allow lithium ions enriched on the electrode surface during charging to fully diffuse into the electrode material, enabling the battery voltage to reach a relatively stable state. After the 45-minute resting period, a high-precision voltmeter measures and records the stable open-circuit voltage (OCV) of each cell. Subsequently, the system compares these measured OCV values ​​with a standard voltage value. This standard voltage value is calculated based on the OCV-SOC fundamental polynomial curve established in S6, inputting the center value of the target SOC (e.g., 30%), and represents the ideal OCV at a reference temperature of 25°C, for example, 3.3080V. The judgment threshold of an absolute value of the voltage difference greater than 0.5mV is based on strict requirements for battery consistency, around 30% SOC. A 0.5mV OCV deviation corresponds to approximately a 0.2% to 0.3% SOC deviation. Setting the threshold to 0.5mV aims to control the SOC difference of all cells in the final battery pack to within 0.3%. For any cell with an absolute voltage difference exceeding 0.5mV, the system will initiate a fine-tuning pulse voltage calibration procedure. Specifically, if the measured OCV is lower than the standard voltage value, the system will apply a 0.01C constant current charging pulse through an independent micro-current source. The pulse duration is fixed at 3 seconds. After that, charging is paused and a 30-second rest observation period is entered to allow the charge from this small replenishment to diffuse internally. After the voltage stabilizes, the OCV is measured again and compared. This cycle of "replenishment-rest-measurement-comparison" will be repeated until the difference between the OCV of the cell and the standard voltage is less than 0.5mV. This fine calibration process will be applied to all cells that do not meet the accuracy requirements, ultimately resulting in a high-precision balanced battery pack.

[0089] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A method for replenishing energy storage batteries to balance the OCV voltage consistency, characterized in that, Includes the following steps: S1, Dynamic health status characterization, the lithium iron phosphate battery pack to be balanced is placed in an environment with the temperature controlled at 20-30℃, and the OCV of each cell is paired with the calculated dynamic DC internal resistance DCR to obtain the initial health status dataset. S2, Based on SOH adaptive grouping and strategy matching, the K-means clustering algorithm is applied to the initial health state dataset to obtain battery arrays with group labels; S3, Differentiated stepped discharge: The battery array with grouped tags is connected to the battery charging and discharging device for discharge to obtain a deeply discharged battery pack. S4. Based on the discharge cutoff judgment of the voltage plateau slope, during the third stage current discharge of the battery pack after the deep discharge, the voltage change rate is continuously monitored to obtain the battery pack that has reached the discharge benchmark. S5, Electrochemical equilibrium adaptive static setting: For the battery pack that has reached the discharge benchmark, disconnect all charging and discharging circuits, determine whether the battery has reached electrochemical equilibrium, and obtain an electrochemically stable battery pack. S6. Based on the closed-loop precise charging of the OCV-SOC model, all cells of the electrochemically stable battery pack are charged to obtain a preliminary balanced battery pack. S7, fine-tune pulse voltage calibration, and perform static and equalization charging on the battery pack after the main charging is completed to obtain a high-precision equalized battery pack. S8, Final inspection of all parameters and data traceability: After completing all calibration steps and letting it stand for 1-2 hours, the high-precision balanced battery pack is subjected to step S1 again to measure and record the OCV and DCR values, and a finished balanced battery with a full life cycle traceability file is obtained.

2. The method for equalizing OCV voltage consistency in energy storage batteries according to claim 1, characterized in that, Step S1 specifically involves: placing the lithium iron phosphate battery pack to be equalized in an environment with a temperature controlled at 20-30℃; applying a constant current discharge pulse of 0.8C-1.5C to each individual cell in the battery pack, with a pulse duration of 8-15 seconds; during this period, collecting voltage data at a frequency of 100-500Hz; after the pulse ends, calculating the dynamic DC internal resistance of each individual cell and recording the initial open-circuit voltage OCV before the pulse; pairing the OCV of each cell with the calculated dynamic DC internal resistance DCR value to obtain the initial health state dataset; the formula for calculating the dynamic DC internal resistance is: Among them, DCR j V represents the dynamic DC internal resistance of the j-th battery. pre,j V represents the average voltage of the j-th battery in the 2 seconds before the pulse is applied. pulse,j I represents the voltage of the j-th battery at the instant the pulse ends. pulse T represents the pulse discharge current. amb Represents ambient temperature, T ref The reference temperature is 25°C, α represents the temperature correction factor for internal resistance, β represents the weighting factor for voltage stability, and σ represents the reference temperature. v,pre This represents the standard deviation of the voltage of the j-th battery within 2 seconds before the pulse is applied.

3. The method for equalizing OCV voltage consistency in energy storage batteries according to claim 1, characterized in that, The S2 step specifically involves: using the K-means clustering algorithm on the initial health state dataset, with the initial OCV and dynamic DC internal resistance as two-dimensional feature vectors, dividing all individual cells into 3-5 cluster groups, and matching a dedicated set of discharge and charge process parameters for each cluster group to obtain a battery array with group labels.

4. The method for equalizing OCV voltage consistency in energy storage batteries according to claim 1, characterized in that, The S3 step specifically involves connecting the battery array with group tags to the battery charging and discharging device, and calling the corresponding discharge process parameter set to perform three-stage discharge according to the allocation tags of each battery in S2, so as to obtain the battery pack after deep discharge.

5. The method for equalizing OCV voltage consistency in energy storage batteries according to claim 1, characterized in that, The S4 step is as follows: during the third stage of current discharge of the battery pack after deep discharge, the voltage change rate of each individual cell is calculated in real time at intervals of 3-8 seconds, and the voltage change rate is continuously monitored. When the voltage drop rate of any individual cell exceeds the preset threshold of 15-25mV / min within a continuous calculation cycle, it is determined that the core discharge plateau period of the battery has ended, and the battery pack that has reached the discharge benchmark is obtained.

6. The method for equalizing OCV voltage consistency in energy storage batteries according to claim 1, characterized in that, The S5 step is as follows: For the battery pack that has reached the discharge benchmark, disconnect all charging and discharging circuits and enter the open circuit resting state. Continuously measure the open circuit voltage of each cell at intervals of 30-90 seconds, calculate the voltage rebound rate in real time, and when the voltage rebound rate of a single cell is continuously lower than the stable threshold of 0.03-0.08mV / s for a continuous time window of 90-180 seconds, it is determined that the battery has reached electrochemical equilibrium, and an electrochemically stable battery pack is obtained.

7. The method for equalizing OCV voltage consistency in energy storage batteries according to claim 1, characterized in that, The S6 step specifically involves: charging all cells of the electrochemically stable battery pack with a current of 0.45C-0.55C. After each 240-360 seconds of charging, the charging is interrupted and the cells are allowed to rest for 8-15 seconds. At the end of the resting period, the quasi-open-circuit voltage of each cell is measured, and the state of charge (SOC) is estimated by combining this with the real-time measured battery surface temperature. est,k When the estimated state of charge (SOC) of any battery reaches the target range of 28%-32%, charging is immediately stopped to obtain a preliminarily balanced battery pack. The formula for calculating the estimated SOC is as follows: Among them, SOC est,k Q represents the estimated state of charge of the k-th battery. ocv,k a represents the quasi-open-circuit voltage measured during the charging interval of the k-th battery. n The fitting coefficients represent the basic curve of the nth-order polynomial OCV-SOC, γ represents the temperature compensation coefficient of the voltage, and T represents the fitting coefficient of the basic curve of the basic curve of the polynomial OCV-SOC. cell,k T represents the real-time temperature of the k-th battery. ref The reference temperature is 25°C, δ represents the correction weighting factor for the individualized internal resistance effect, and DCR k represents the dynamic DC internal resistance value of the k-th battery obtained from S1, and N represents the order of the polynomial used to fit the OCV-SOC basic curve.

8. The method for equalizing OCV voltage consistency in energy storage batteries according to claim 1, characterized in that, The S7 step specifically involves: after completing the main charging, the initially balanced battery pack is allowed to rest for 30-60 minutes, and the stable OCV of each cell is measured and compared with the standard voltage value corresponding to the target estimated state of charge. If the absolute value of the voltage difference is greater than 0.5mV, a current pulse of 0.008C-0.015C is applied to replenish the charge. Each pulse lasts for 1-5 seconds, and the pulse interval is 30-60 seconds to wait for the voltage to stabilize until the difference between the OCV of all cells and the standard voltage is less than 0.5mV, thus obtaining a high-precision balanced battery pack.

Images