Battery management system adaptive active equalization method, device and system

Abstract

The application provides a battery management system adaptive active balancing method, device and system, and relates to the technical field of energy storage batteries.The method comprises the following steps: obtaining electrochemical impedance spectrum test data and charge-discharge curve segments corresponding to each battery cell in a battery pack; in the case where a current voltage difference value of the battery cell is less than or equal to a preset voltage difference threshold value, inputting the electrochemical impedance spectrum test data and the charge-discharge curve segments into a prediction balancing model to obtain a battery cell capacity attenuation rate prediction value output by the prediction balancing model, and obtaining a battery balancing priority ranking result according to the battery cell capacity attenuation rate prediction value; and performing a corresponding energy transfer operation on a target battery cell according to the battery cell capacity attenuation rate prediction value and the battery balancing priority ranking result. The application realizes accurate aging evaluation and efficient balancing management of the battery cell.

Description

Technical Field

[0001] This invention relates to the field of energy storage battery technology, and in particular to an adaptive active balancing method, apparatus and system for a battery management system. Background Technology

[0002] Energy storage lithium battery packs typically consist of one or more battery groups connected in parallel, with each battery group composed of multiple batteries connected in series. During the charging and discharging of lithium batteries, an imbalance effect can easily occur, where the charging or discharging of the entire pack stops when one battery is fully charged or discharged. Traditional balancing methods are divided into active balancing and passive balancing. Active balancing uses energy transfer and often employs a DC / DC topology. By monitoring the external environment (such as temperature and humidity) and individual cell voltage, it controls the activation and deactivation of the battery pack to maintain a constant current balance, thereby improving the charging and discharging efficiency of the battery pack and saving energy.

[0003] Traditional active balancing involves collecting the overall voltage of the battery pack and the voltage of each individual cell, then setting an active balancing strategy for each individual cell based on the collected voltage, and controlling it to charge or discharge according to the strategy. In addition, the external environment during the active balancing process of each individual cell is monitored, and the balancing process is adjusted according to the results.

[0004] However, in existing lithium battery charging and discharging processes, due to the large capacity of individual battery cells, the passive balancing current is small and the balancing time is long. The active balancing strategies employed often rely on voltage difference triggering, failing to fully consider the changes in electrochemical parameters under cell aging conditions, resulting in deviations between balancing decisions and actual aging states. Furthermore, traditional active balancing technologies do not adjust the balancing current according to aging conditions; a fixed balancing current accelerates the aging rate of relatively aged cells, thus exacerbating the aging differences across the entire battery pack. Therefore, there is an urgent need for an adaptive active balancing method, device, and system for battery management systems to address these issues. Summary of the Invention

[0005] To address the problems existing in the prior art, the present invention provides an adaptive active balancing method, apparatus and system for a battery management system.

[0006] This invention provides an adaptive active balancing method for a battery management system, comprising:

[0007] Acquire electrochemical impedance spectroscopy test data and charge-discharge curve segments for each cell in the battery pack;

[0008] If the current voltage difference of the battery cell is determined to be less than or equal to a preset voltage difference threshold, the electrochemical impedance spectroscopy test data and the charge-discharge curve segment are input into the prediction equalization model to obtain the predicted value of the battery cell capacity decay rate output by the prediction equalization model, and the battery equalization priority ranking result is obtained based on the predicted value of the battery cell capacity decay rate.

[0009] Based on the predicted cell capacity decay rate and the battery balancing priority ranking result, the corresponding energy transfer operation is performed on the target cell in the battery pack.

[0010] According to the adaptive active balancing method for a battery management system provided by the present invention, the step of acquiring electrochemical impedance spectroscopy test data corresponding to each cell in the battery pack includes:

[0011] Electrochemical impedance spectroscopy was performed on the battery cell to collect data on the solid electrolyte interface membrane impedance, charge transfer resistance, and diffusion impedance.

[0012] The electrochemical impedance spectroscopy test data are obtained based on the solid electrolyte interface membrane impedance data, the charge transfer resistance data, and the diffusion impedance data.

[0013] An adaptive active balancing method for a battery management system provided by the present invention further includes:

[0014] The current voltage difference value of the battery cell is determined based on the least mean square algorithm.

[0015] If it is determined that the current voltage difference value of the battery cell is greater than the preset voltage difference threshold, an equalization current adjustment operation is performed on the battery cell based on the recursive least squares method, the voltage difference value, the solid electrolyte interface membrane impedance data, and the charge transfer resistance change, wherein the charge transfer resistance change is obtained based on the charge transfer resistance data.

[0016] According to the adaptive active balancing method for a battery management system provided by the present invention, the step of obtaining electrochemical impedance spectroscopy test data based on the solid electrolyte interface membrane impedance data, the charge transfer resistance data, and the diffusion impedance data includes:

[0017] The stability index of the solid electrolyte interface membrane is calculated based on the reciprocal of the product between the solid electrolyte interface membrane impedance data and the solid electrolyte interface membrane capacitance.

[0018] The charge transfer efficiency value is calculated based on the reciprocal of the charge transfer resistance data.

[0019] The lithium-ion diffusion coefficient is calculated based on the reciprocal square of the diffusion impedance data.

[0020] The electrochemical impedance spectroscopy test data are constructed based on the solid electrolyte interface film stability index, the charge transfer efficiency value, and the lithium ion diffusion coefficient.

[0021] According to the adaptive active balancing method for a battery management system provided by the present invention, the predictive balancing model is trained through the following steps:

[0022] Sample data were constructed based on historical electrochemical impedance spectroscopy sample data and charge-discharge curve samples;

[0023] Obtain the sample values ​​of the cell capacity decay rate corresponding to the sample data at historical moments, and construct the label data of the sample data based on the sample values ​​of the cell capacity decay rate;

[0024] Based on the sample data and the label data, the long short-term memory neural network is trained to obtain the prediction equilibrium model.

[0025] According to the adaptive active balancing method for a battery management system provided by the present invention, the step of performing corresponding energy transfer operations on target cells in the battery pack based on the predicted cell capacity decay rate and the battery balancing priority ranking result includes:

[0026] Based on the predicted cell capacity decay rate and the battery equalization priority ranking result, an energy distribution equalization instruction corresponding to the target cell is generated.

[0027] Based on the bidirectional DC / DC converter and the energy distribution balancing command, an energy transfer operation is performed on the target battery cell.

[0028] An adaptive active balancing method for a battery management system provided by the present invention further includes:

[0029] The stability index of the solid electrolyte interface film, the charge transfer efficiency value, and the lithium ion diffusion coefficient are compared with their respective preset thresholds.

[0030] Based on the comparison results, the cell health status data of the battery cell is updated.

[0031] The present invention also provides an adaptive active balancing device for a battery management system, comprising:

[0032] The test unit is used to acquire electrochemical impedance spectroscopy test data and charge-discharge curve segments corresponding to each cell in the battery pack.

[0033] The decision unit is used to input the electrochemical impedance spectroscopy test data and the charge-discharge curve segment into the prediction equalization model when the current voltage difference value of the battery cell is less than or equal to a preset voltage difference threshold, to obtain the predicted value of the battery cell capacity decay rate output by the prediction equalization model, and to obtain the battery equalization priority ranking result based on the predicted value of the battery cell capacity decay rate.

[0034] An energy transfer unit is used to perform corresponding energy transfer operations on the target cells in the battery pack based on the predicted cell capacity decay rate and the battery equalization priority ranking result.

[0035] The present invention also provides a battery management system, including the above-described adaptive active balancing device for the battery management system.

[0036] The adaptive active balancing method, device, and system for battery management systems provided by this invention acquire electrochemical impedance spectroscopy test data and charge-discharge curve segments of each cell in the battery pack. When the cell voltage difference value meets the standard, it is input into a predictive balancing model based on neural network training to obtain the predicted value of cell capacity decay rate and sort it. Finally, based on this result, energy transfer operation is performed on the target cell to achieve accurate assessment and efficient balancing management of cell aging. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in this invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 A flowchart illustrating the adaptive active balancing method for the battery management system provided by this invention;

[0039] Figure 2 This is a schematic diagram of the overall process of the present invention;

[0040] Figure 3 A schematic diagram of the adaptive active balancing device for the battery management system provided by the present invention;

[0041] Figure 4 A schematic diagram of the deployment of the adaptive active balancing device of the battery management system provided by the present invention on an energy storage architecture;

[0042] Figure 5 This is a schematic diagram of the battery management system provided by the present invention;

[0043] Figure 6 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0045] Traditional battery management system (BMS) balancing technology has certain limitations. Because the individual capacity of energy storage batteries is relatively large (hundreds of Ah or more), while the passive balancing current is generally below 5A, the balancing time is quite long (tens of hours or more). Furthermore, active balancing strategies often rely on voltage difference triggering. This long-term, single-condition active balancing method does not fully consider the electrochemical characteristics of the cells under their aging state, such as changes in cell internal resistance and solid electrolyte interface (SEI) impedance. This leads to a deviation between the balancing decision and the actual aging state. Moreover, traditional active balancing technology does not adjust the balancing current based on the aging state, consistently using a fixed balancing current, which further accelerates the aging of relatively aged cells, thus exacerbating the aging differences across the entire battery pack.

[0046] To address the problems existing in the prior art, this invention provides an adaptive active balancing method for battery management systems based on multi-dimensional electrochemical characteristics. By real-time monitoring of the dynamic impedance spectrum of the cells and combining it with multi-parameter fusion analysis to optimize energy transfer strategies, this method achieves efficient balancing management of individual cells within the battery pack. This invention introduces the acquisition and processing of high-frequency, mid-frequency, and low-frequency impedance data during the balancing process, forming a closed-loop control mechanism to ensure the accuracy and adaptability of balancing decisions, significantly improving the battery pack's lifespan and overall performance.

[0047] Figure 1 This is a flowchart illustrating the adaptive active balancing method for the battery management system provided by the present invention, as shown below. Figure 1 As shown, the present invention provides an adaptive active balancing method for a battery management system, comprising:

[0048] Step 101: Obtain electrochemical impedance spectroscopy test data and charge / discharge curve segments corresponding to each cell in the battery pack.

[0049] In this invention, a dynamic electrochemical impedance spectroscopy (EIS) test is first performed, using a speed measurement device that supports dynamic EIS testing with a DC bias current greater than 10A. At the end of charging, when the battery's state of charge (SOC) is between 80% and 90%, a pseudo-electrochemical impedance spectroscopy (PEIS) test can be performed, with a frequency range of 100kHz to 0.1Hz. During discharge, a galvanostatic electrochemical impedance spectroscopy (GEIS) test can be dynamically performed with a bias current of 10A, monitoring the changing trend of diffusion impedance in real time.

[0050] After testing, communication with the main control unit is established via the CAN bus (Controller Area Network), with a data transmission rate of up to 500kbps. This ensures the reliable transmission of high-precision impedance data, and ultimately, electrochemical impedance spectroscopy test data for each cell in the battery pack is obtained through feature extraction. Simultaneously, during normal charging and discharging of the battery, charge / discharge curve segments of each cell (e.g., the 3.8V to 4.1V range) are collected.

[0051] Step 102: If the current voltage difference value of the battery cell is less than or equal to the preset voltage difference threshold, the electrochemical impedance spectroscopy test data and the charge-discharge curve segment are input into the prediction equalization model to obtain the predicted value of the battery cell capacity decay rate output by the prediction equalization model, and the battery equalization priority ranking result is obtained based on the predicted value of the battery cell capacity decay rate.

[0052] In this invention, the current voltage difference value of the battery cell is first determined. When the voltage difference value is less than or equal to a preset voltage difference threshold (e.g., 30mV), the electrochemical impedance spectroscopy test data and charge-discharge curve fragments obtained in step 101 are input into the prediction equilibrium model.

[0053] The predictive equalization model is trained on a Long Short-Term Memory (LSTM) neural network model using a large number of training samples. The model's input layer contains EIS data from the most recent 10 cycles (covering high-frequency, mid-frequency, and low-frequency impedance) and charge / discharge curve segments (3.8V to 4.1V range). The hidden layer consists of two layers of LSTM units (each with 128 neurons). The output layer can predict the cell capacity decay rate after 72 hours (i.e., the predicted cell capacity decay rate). Based on the predicted cell capacity decay rate, it also outputs the battery equalization priority ranking result, with a prediction error of less than 2%.

[0054] Optionally, the present invention also includes a preset prediction trigger period, with the above prediction process running once daily. Furthermore, when the State of Health (SOH) of the battery cell is below 80%, a balancing plan is triggered 48 hours in advance to prepare for battery cell balancing.

[0055] Specifically, in this invention, the voltage difference value reflects the degree of voltage inconsistency between the individual cells within the battery pack. When the voltage difference between cells is too large, it may affect the overall performance and lifespan of the battery pack, and even cause safety issues. Therefore, it is necessary to first determine whether the current voltage difference value of the cells is less than or equal to a preset voltage difference threshold, which serves as the basis for whether to initiate the predictive equalization process.

[0056] The preset voltage difference threshold is determined based on factors such as battery type, specifications, and actual application scenarios. For example, in scenarios with high battery performance requirements, the threshold may be set lower, such as 30mV, to ensure that the voltage difference between cells is within a small range and to guarantee the stable operation of the battery pack; while in scenarios where cost and efficiency requirements are more balanced, the threshold may be appropriately relaxed.

[0057] In this invention, the voltage monitoring module in the battery management system collects the voltage values ​​of each cell in the battery pack in real time, and then calculates the maximum voltage difference between the cells. This maximum voltage difference is compared with a preset voltage difference threshold. If the maximum voltage difference is less than or equal to the preset threshold, it means that the voltage difference between the cells is within an acceptable range, and the next step of predictive equalization can proceed. If the maximum voltage difference is greater than the preset threshold, it may be necessary to take some emergency equalization measures (such as rapid response of real-time equalization) to reduce the voltage difference, and then perform predictive equalization after the voltage difference meets the requirements.

[0058] Electrochemical impedance spectroscopy (EIS) is a method for analyzing the internal electrochemical characteristics of a battery by measuring its impedance at different frequencies. In step 101, EIS data of each cell in the battery pack were acquired using specific testing equipment. This data covers high-frequency, mid-frequency, and low-frequency impedance information, and can reflect electrochemical parameters such as ohmic resistance, charge transfer resistance, and diffusion impedance within the cell.

[0059] The charge-discharge curve records the changes in parameters such as voltage and current of the battery over time during the charge-discharge process. This invention selects a segment of the charge-discharge curve in the range of 3.8V to 4.1V as input data because this voltage range usually contains some key characteristics of the battery charge-discharge process, such as the battery polarization characteristics and the stability of the charge-discharge platform, which can provide valuable information for predicting the cell capacity decay rate.

[0060] Furthermore, the collected electrochemical impedance spectroscopy test data and charge-discharge curve fragments were organized and preprocessed to ensure that the data format and accuracy met the requirements of the predictive equilibrium model. Then, this data was input into the predictive equilibrium model through a specific data interface. The predictive equilibrium model is obtained by training a neural network model based on training samples. The neural network model has powerful nonlinear mapping capabilities and can learn the complex relationship between input data and cell capacity decay rate.

[0061] During the model training phase, a large number of training samples need to be collected. These samples include EIS data, charge-discharge curve segments, and corresponding actual capacity degradation rates of battery cells under different aging states and usage conditions. The training samples are divided into training and validation sets. The neural network model is trained using the training set, and the model parameters (such as weights and biases) are continuously adjusted to minimize the error between the model's predictions and the actual capacity degradation rates. Simultaneously, the trained model is validated using the validation set to evaluate its generalization ability and ensure that the model can provide accurate predictions even on new data.

[0062] In this invention, the predictive equalization model outputs a predicted capacity decay rate of the battery cells after 72 hours (adjustable according to actual conditions) based on the input electrochemical impedance spectroscopy test data and charge-discharge curve segments. This predicted value reflects the performance degradation trend of the battery cells over a future period. Then, based on the predicted capacity decay rate, the cells within the battery pack are sorted to determine the priority of battery equalization. Cells with higher predicted capacity decay rates are more likely to experience performance degradation in the future and require priority equalization to balance the performance of the cells within the battery pack and extend the overall lifespan of the battery pack. For example, by sorting the predicted capacity decay rates from high to low, energy transfer and other equalization operations are prioritized for the cells ranked higher.

[0063] Step 103: Based on the predicted cell capacity decay rate and the battery balancing priority ranking result, perform the corresponding energy transfer operation on the target cell in the battery pack.

[0064] In this invention, based on the predicted cell capacity decay rate and battery balancing priority ranking results obtained in step 102, a target cell (i.e., the cell to be subjected to energy transfer operation) is determined from multiple cells within the battery pack. Then, a bidirectional DC / DC converter is used to implement the energy transfer operation, which supports energy distribution between clusters or between cells. The main control unit in the battery management system sends balancing commands to relevant control units via the CAN bus. The relevant control units then activate the bidirectional DC / DC converter according to the commands, transferring energy from high-capacity cells to low-capacity cells at a 5A current. This method can shorten the balancing time to one-third of traditional methods, effectively reducing energy loss and meeting the high-efficiency balancing requirements of large-capacity battery packs.

[0065] Specifically, in this invention, the target cell is determined based on the predicted capacity decay rate and the battery balancing priority ranking. The predicted capacity decay rate reflects the degree of performance degradation of the cell over a future period, while the balancing priority ranking clarifies the order in which each cell needs to undergo balancing operations. Cells with higher predicted capacity decay rates and higher rankings in the balancing priority ranking are prioritized as target cells. For example, in a battery pack, if multiple cells have predicted capacity decay rates of 5%, 3%, 7%, and 2%, after balancing priority ranking, the cell with a predicted capacity decay rate of 7% is first identified as the target cell for energy transfer operations. After balancing is completed, the cells ranked lower in the priority ranking are then processed sequentially.

[0066] A bidirectional DC / DC converter is a power electronic device capable of enabling bidirectional flow of DC power. It achieves bidirectional energy transfer by controlling the on / off state of switching transistors to alter the magnitude and direction of input and output voltage and current. In battery balancing systems, it can transfer energy from high-capacity cells to low-capacity cells, achieving energy balance among the cells within the battery pack. Compared to traditional unidirectional energy transfer methods, bidirectional DC / DC converters offer greater flexibility and efficiency. They can flexibly control the direction and magnitude of energy flow according to actual needs, better adapting to battery balancing requirements under different operating conditions. Furthermore, their high energy conversion efficiency effectively reduces energy loss and improves the overall energy utilization efficiency of the battery pack.

[0067] In some large-scale battery energy storage systems, battery packs typically consist of multiple battery clusters. Inter-cluster energy distribution refers to the transfer of energy from one battery cluster to another. For example, when the cells in a battery cluster become unbalanced in charge due to long-term use or other reasons, a bidirectional DC / DC converter can transfer energy from other battery clusters with sufficient charge to that cluster, thus achieving energy balance between clusters.

[0068] For a battery system consisting of multiple battery boxes, inter-box energy distribution refers to transferring energy from one battery box to another. This distribution method can further expand the scope of energy balance and improve the stability and reliability of the entire battery system. For example, in a distributed battery energy storage system, battery boxes in different locations may have inconsistent charge levels due to factors such as ambient temperature and usage frequency. Inter-box energy distribution can achieve system-wide balance.

[0069] In this invention, based on the predicted cell capacity decay rate and the balancing priority ranking results, combined with the current state of the battery pack (such as the voltage, current, temperature, and other parameters of each cell), specific balancing instructions are generated. The balancing instructions clearly specify the target cell for energy transfer, the direction of energy transfer (from which cell to which cell), and the amount of energy transferred. The balancing instructions are then sent to the slave control module via the CAN bus. The CAN bus has advantages such as strong real-time performance, high reliability, and strong anti-interference capabilities, ensuring that the balancing instructions are accurately and promptly transmitted to the slave control module.

[0070] Upon receiving an equalization command, the bidirectional DC / DC converter first parses the command to extract key information, such as the target cell identifier, the direction and magnitude of energy transfer, etc. Then, based on the parsed command, it adjusts parameters such as input and output voltage and current to achieve directional energy transfer. During the energy transfer process, the voltage and current parameters of the cell are monitored in real time to ensure the safety and stability of the energy transfer process. If any abnormality is detected (such as excessively high or low cell voltage, excessive current, etc.), the bidirectional DC / DC converter's operating state is adjusted promptly or the energy transfer operation is stopped, and abnormal information is reported back.

[0071] Under the action of the bidirectional DC / DC converter, the energy of the high-capacity cells is gradually transferred to the low-capacity cells with a certain current (such as 5A). During this process, the voltage of each cell in the battery pack will gradually tend to be balanced, and the difference in capacity between the cells will gradually decrease.

[0072] In this invention, the energy transfer operation can end when one of the following conditions is met: First, the voltage difference between the cells reaches a preset equalization target value, that is, the voltage difference between each cell is less than or equal to a small threshold, indicating that the battery pack has reached a good equalization state; Second, a preset equalization time is reached. In order to avoid the energy transfer process being too long and affecting the normal use of the battery pack, a maximum equalization time is set, and the energy transfer operation will automatically stop when the time is reached; Third, if an abnormal situation occurs, such as the cell temperature is too high or the inverter fails, the energy transfer operation will be stopped immediately in order to protect the safety of the battery pack and the equipment.

[0073] The adaptive active balancing method for battery management system provided by this invention acquires electrochemical impedance spectroscopy test data and charge-discharge curve segments of each cell in the battery pack. When the cell voltage difference value meets the standard, it is input into a predictive balancing model based on neural network training to obtain the predicted value of cell capacity decay rate and sort it. Finally, based on this result, energy transfer operation is performed on the target cell to achieve accurate assessment and efficient balancing management of cell aging.

[0074] Based on the above embodiments, the step of obtaining electrochemical impedance spectroscopy test data corresponding to each cell in the battery pack includes:

[0075] Electrochemical impedance spectroscopy was performed on the battery cell to collect data on the solid electrolyte interface membrane impedance, charge transfer resistance, and diffusion impedance.

[0076] The electrochemical impedance spectroscopy test data are obtained based on the solid electrolyte interface membrane impedance data, the charge transfer resistance data, and the diffusion impedance data.

[0077] This invention supports dynamic EIS testing with a DC bias current greater than 10A. In actual battery usage scenarios, battery cells often operate under high current conditions. The testing process can simulate the impedance characteristics of the battery cell under real high current loads, making the test results closer to the actual operating conditions of the battery cell and improving the practicality and accuracy of the test data.

[0078] Specifically, a pseudo-potential constant EIS test is performed at the end of the charging process (80% to 90% SOC). The end of the charging process is a stage where the battery state changes are more complex. Performing the test at this time can capture the impedance characteristics of the cell when it is close to being fully charged, which is of great significance for understanding the charging performance and health status of the cell.

[0079] Furthermore, a galvanic EIS test is dynamically performed during the discharge process. The discharge process is the stage where the cell releases energy; by testing during this process, the impedance changes of the cell under different discharge states can be monitored in real time. During discharge, a bias current of 10A is used for testing, which accurately captures the changing trend of the diffusion impedance. As discharge progresses, the lithium-ion concentration and distribution inside the cell change, and the diffusion impedance changes accordingly. Dynamic testing can capture these changes, providing detailed data for evaluating the cell's discharge performance.

[0080] In the above tests, the frequency range was set from 100kHz to 0.1Hz. The high-frequency region (around 100kHz) primarily reflects the impedance characteristics of the solid electrolyte interphase (SEI) membrane. SEI impedance data (RI) can be acquired through testing in this frequency band. SEI SEI is a passivation film formed during the first charge and discharge of a battery cell, and its impedance characteristics have a significant impact on the performance and lifespan of the battery cell.

[0081] Mid-frequency region (frequency range in the middle part) corresponding charge transfer resistance data (R) ct It can collect charge transfer resistance data, which reflects the ease of charge transfer inside the cell and is closely related to the chemical reactivity of the cell.

[0082] In the low-frequency region (around 0.1Hz), the diffusion impedance data (R) is relevant. w Related to this, diffusion impedance data can be collected. Diffusion impedance reflects the diffusion ability of lithium ions inside the cell and has a significant impact on the charge and discharge performance of the cell.

[0083] In this invention, an equivalent circuit model is used to fit the collected impedance spectrum data. An equivalent circuit model is a model that represents the electrochemical process of a battery cell as a combination of circuit elements. By appropriately selecting and setting the parameters of these circuit elements, the impedance characteristics of the battery cell at different frequencies can be simulated. For example, the SEI impedance, charge transfer resistance, and diffusion impedance can be represented as resistive elements in a circuit, and combined with other components such as capacitors and inductors, an equivalent circuit model that accurately describes the impedance characteristics of the battery cell can be constructed.

[0084] For example, based on the fitted SEI impedance data, the SEI stability index (SSI) is calculated using a specific algorithm. SSI reflects the stability of the SEI, which significantly impacts the cell's performance and lifespan. For instance, an unstable SEI may lead to side reactions during charging and discharging, increasing the cell's internal resistance and reducing its capacity and cycle life. The charge transfer efficiency (CTE) is calculated based on charge transfer resistance data. CTE reflects the effectiveness of charge transfer within the cell; efficient charge transfer means faster and more complete chemical reactions, thus improving the cell's charging and discharging performance. Finally, the lithium-ion diffusion coefficient (D) is calculated using diffusion impedance data. Li Lithium-ion diffusion coefficient D Li It reflects the diffusion rate of lithium ions inside the cell. The larger the diffusion coefficient, the easier it is for lithium ions to diffuse inside the cell, and the better the charge and discharge rate and performance of the cell.

[0085] Furthermore, the calculated SEI stability index SSI, charge transfer efficiency CTE, and lithium-ion diffusion coefficient D are used to further analyze the results. Li These parameters, as important components of electrochemical impedance spectroscopy (EIS) test data, are used to update the cell health status profile. The cell health status profile records various performance indicators of the cell at different points in time. Continuous updating of the profile allows for a comprehensive and dynamic understanding of the cell's health status trends, providing a scientific basis for cell maintenance, management, and replacement. For example, if it is found that a cell's SEI stability index (SSI), charge transfer efficiency (CTE), or lithium-ion diffusion coefficient (D) is continuously decreasing, the profile will be updated. Li When the value decreases significantly, it indicates that the health of the battery cell is deteriorating, and appropriate measures need to be taken in a timely manner, such as performing equalization maintenance or replacing the battery cell.

[0086] Based on the above embodiments, the method further includes:

[0087] The current voltage difference value of the battery cell is determined based on the least mean square algorithm.

[0088] If it is determined that the current voltage difference value of the battery cell is greater than the preset voltage difference threshold, an equalization current adjustment operation is performed on the battery cell based on the recursive least squares method, the voltage difference value, the solid electrolyte interface membrane impedance data, and the charge transfer resistance change, wherein the charge transfer resistance change is obtained based on the charge transfer resistance data.

[0089] In this invention, the Least Mean Square (LMS) algorithm is an adaptive filtering algorithm that can quickly determine the current voltage difference between cells in a cell voltage balancing scenario. This algorithm continuously adjusts its parameters to minimize the mean square value of the error signal. During real-time monitoring of cell voltage, it compares the actual measured voltage values ​​of each cell with a preset reference voltage value (or the average voltage value of the cell group) to obtain the voltage difference value.

[0090] When voltage differences occur between battery cells, the LMS algorithm can react quickly. For example, if a preset voltage difference threshold of 30mV is set, the algorithm will immediately identify this anomaly once a voltage difference greater than 30mV is detected between cells. This is because the LMS algorithm has a fast convergence speed, with a convergence time of less than 50ms, enabling it to determine the voltage difference in a very short time and provide timely information for subsequent equalization operations. This rapid response capability is crucial for maintaining the consistency and stability of the battery cell assembly, preventing overcharging or over-discharging problems caused by excessive voltage differences, and extending the lifespan of the battery cells.

[0091] After determining that the current voltage difference of the battery cell exceeds a preset voltage difference threshold, the balancing current needs to be adjusted. Recursive Least Squares (RLS) dynamically adjusts model parameters as new data is continuously acquired, thus more accurately reflecting the current state of the system. In a battery cell balancing system, the RLS algorithm can dynamically adjust the balancing current based on real-time voltage difference values, solid electrolyte interface membrane impedance data, and changes in charge transfer resistance to achieve more precise balancing control. Specifically, the balancing current... The calculation formula is:

[0092] ;

[0093] in, This indicates the voltage difference between battery cells. When... When the voltage difference exceeds the preset threshold, it indicates an imbalance in the cell assembly, requiring real-time adjustment of the balancing current to eliminate this difference. SEI (Self-Interference Layer) is a passivation film formed during the first charge and discharge cycle of the cell; its impedance characteristics significantly impact cell performance. The SEI (Self-Impeding Insulation) reflects the resistance of the current to the current, and it is one of the important impedance parameters in equalization current calculations. The charge transfer resistance reflects the ease of charge transfer within the battery cell. It is obtained from charge transfer resistance data, which represents the change in charge transfer resistance at different times or under different conditions. The change in charge transfer resistance will affect the charging and discharging performance of the battery cell. It is an adjustment coefficient used to balance the influence of changes in charge transfer resistance on the equalization current. This can be achieved by setting it appropriately. The value of can make the calculation of the balancing current more in line with the actual situation and improve the balancing effect.

[0094] This invention combines the LMS and RLS algorithms to achieve dynamic adjustment of the balancing current. As the battery cells are used and their condition changes, the SEI impedance and charge transfer resistance also change. The RLS algorithm can continuously update the calculation parameters of the balancing current based on real-time data, ensuring that the balancing current remains within a suitable range. This dynamic adjustment capability ensures that the battery cell assembly can achieve effective balancing under different operating conditions, improving the consistency and overall performance of the battery cell assembly. For example, during the aging process of the battery cells, the SEI impedance may increase, and the charge transfer resistance may also change. By dynamically adjusting the balancing current, these changes can be compensated for in a timely manner, maintaining the balanced state of the battery cell assembly.

[0095] Based on the above embodiments, obtaining the electrochemical impedance spectroscopy test data based on the solid electrolyte interface membrane impedance data, the charge transfer resistance data, and the diffusion impedance data includes:

[0096] The stability index of the solid electrolyte interface membrane is calculated based on the reciprocal of the product between the solid electrolyte interface membrane impedance data and the solid electrolyte interface membrane capacitance.

[0097] The charge transfer efficiency value is calculated based on the reciprocal of the charge transfer resistance data.

[0098] The lithium-ion diffusion coefficient is calculated based on the reciprocal square of the diffusion impedance data.

[0099] The electrochemical impedance spectroscopy test data are constructed based on the solid electrolyte interface film stability index, the charge transfer efficiency value, and the lithium ion diffusion coefficient.

[0100] The solid electrolyte interphase (SEI) film is a passivation film formed during the charging and discharging process of the battery cell. Its impedance characteristics (R) SEI The impedance and capacitance characteristics of the solid electrolyte interface film have a significant impact on cell performance. This invention uses solid electrolyte interface film impedance data (R0) to... SEI ) and solid electrolyte interface film capacitance (C SEI The solid electrolyte interfacial membrane stability index (SSI) is calculated by taking the reciprocal of the product between R and R. Specifically, R SEI C reflects the degree to which the SEI impedes the flow of current. SEI This reflects the SEI's ability to store charge. The reciprocal of their product combines these two characteristics, providing an overall reflection of the SEI's stability. For example, if R...SEI A larger value indicates that the SEI (Sediment Ion) significantly impedes current flow, potentially indicating a thicker film or poorer performance; if C... SEI A smaller product indicates a weaker charge storage capacity, which may also affect the stability of the membrane. The smaller the reciprocal of the product, the more stable the SEI may be, because a stable SEI should have moderate impedance and capacitance characteristics, which can effectively conduct current and store charge reasonably.

[0101] Charge transfer resistance (R) ct The charge transfer efficiency (CTE) reflects the ease of charge transfer within the battery cell and is closely related to the cell's chemical reactivity. This invention calculates the CTE by taking the reciprocal of the charge transfer resistance data. Specifically, R... ct The smaller the value, the easier it is for charge to transfer inside the cell, resulting in higher electrochemical reactivity and higher charge transfer efficiency. Taking the reciprocal, a larger CTE value indicates higher charge transfer efficiency. For example, when R... ct A lower CTE value indicates that the charge transfer reaction on the electrode surface proceeds more smoothly, and the cell can charge and discharge more efficiently, resulting in a higher CTE value.

[0102] Diffusion resistance (R) w The diffusion coefficient (D) reflects the diffusion ability of lithium ions within the battery cell. This invention calculates the lithium-ion diffusion coefficient (D) based on the reciprocal square of the diffusion impedance data. Li ). Specifically, R w The smaller the value, the easier it is for lithium ions to diffuse inside the cell, and the stronger the diffusion ability. Taking the square of the inverse of the diffusion impedance, D... Li The larger the value, the greater the lithium-ion diffusion coefficient. For example, when R... w When the density is small, lithium ions can diffuse more quickly in the electrode material, which is beneficial to the charging and discharging process of the cell, so D Li The value will be relatively large.

[0103] In one embodiment, SSI (Sequence Instability Sequence Index) can be used to determine whether SEI (Sequence Identification Index) is stable. If it is unstable, the charge / discharge strategy may need to be adjusted to prevent further deterioration of SEI. CTE (Chemical Reactivity Test) value can reflect the chemical reactivity of the cell. If the activity decreases, the operating conditions of the cell may need to be optimized. Li It can indicate the diffusion of lithium ions; if diffusion is hindered, the temperature or charge / discharge rate may need to be adjusted.

[0104] The solid electrolyte interfacial film stability index (SSI), charge transfer efficiency (CTE), and lithium-ion diffusion coefficient (D) were used to measure the stability of the solid electrolyte interfacial film (SSI), charge transfer efficiency (CTE), and lithium-ion diffusion coefficient (D). Li The electrochemical impedance spectroscopy (EIS) data constructed can provide important information for battery management systems. Long-term monitoring and analysis of this data can assess the health status and performance degradation of the battery cells. For example, if the SSI gradually decreases, the CTE gradually decreases, or the D...Li A gradually decreasing size may indicate a decline in the cell's performance, requiring timely maintenance or replacement. At the same time, this data can also be used to optimize battery charging and discharging strategies, improving battery efficiency and lifespan.

[0105] In this invention, the solid electrolyte interfacial film stability index (SSI), charge transfer efficiency (CTE), and lithium-ion diffusion coefficient (D) are used. Li These parameters reflect the electrochemical performance of the battery cell from different perspectives. SSI focuses on the stability of SEI, CTE reflects the efficiency of charge transfer, and D... Li By reflecting the diffusion ability of lithium ions and integrating these parameters to construct electrochemical impedance spectroscopy test data, the internal electrochemical characteristics of the battery cell can be described more comprehensively and holistically.

[0106] Based on the above embodiments, the predictive equilibrium model is trained through the following steps:

[0107] Sample data were constructed based on historical electrochemical impedance spectroscopy sample data and charge-discharge curve samples;

[0108] Obtain the sample values ​​of the cell capacity decay rate corresponding to the sample data at historical moments, and construct the label data of the sample data based on the sample values ​​of the cell capacity decay rate;

[0109] Based on the sample data and the label data, the long short-term memory neural network is trained to obtain the prediction equilibrium model.

[0110] In this invention, the construction of sample data mainly relies on historical electrochemical impedance spectroscopy (EIS) sample data and charge-discharge curve samples. EIS data reflects the impedance characteristics of the battery cell at different frequencies. Impedance at different frequencies (high frequency, mid frequency, and low frequency) contains information about different electrochemical processes within the battery cell. For example, high-frequency impedance may be related to charge transfer processes on the electrode surface, while mid-frequency and low-frequency impedance may involve diffusion processes within the electrode material. The charge-discharge curve samples record the changes in parameters such as voltage and current of the battery cell over time during the charge-discharge process. The charge-discharge curve segment in the 3.8V to 4.1V range contains important electrochemical behavior information of the battery cell within this voltage range.

[0111] The input layer of the predictive equalization model contains EIS data (high-frequency, mid-frequency, and low-frequency impedance) from the most recent 10 charge-discharge cycles, as well as charge-discharge curve segments (3.8V to 4.1V range). When constructing the sample data, EIS data of the battery cell is collected over the most recent 10 charge-discharge cycles. The impedance values ​​at different frequencies in each cycle are used as one data dimension, and the charge-discharge curve segments in the 3.8V to 4.1V range within each cycle are extracted as another data component. These data from different cycles and of different types are integrated to form a complete sample dataset. For example, for a specific battery cell, high-frequency, mid-frequency, and low-frequency impedance values ​​from 10 consecutive charge-discharge cycles, along with voltage-time curve data in the 3.8V to 4.1V range for each cycle, are collected. These data are then arranged and combined according to a specific format to form a sample dataset.

[0112] Simultaneously, this invention requires obtaining sample values ​​of the cell capacity decay rate at historical moments. The cell capacity decay rate is a crucial indicator for measuring the degree of performance degradation of a cell; it reflects the reduction in cell capacity during use. It can be obtained through experimental testing or long-term monitoring of the cell's charge and discharge capacity, calculating the decay ratio relative to the initial capacity at different times. For example, after a period of use, the current charge and discharge capacity of the cell is accurately measured and compared with the initial capacity to calculate the capacity decay rate.

[0113] This invention constructs labeled data for sample data based on sample values ​​of battery cell capacity decay rate. Label data is an annotation of the result corresponding to the sample data, i.e., the capacity decay rate of the battery cell at a specific historical moment. Label data provides target values ​​for training the Long Short-Term Memory Neural Network (LSTM), enabling the network to learn the relationship between the input sample data (EIS data and charge / discharge curve segments) and the output battery cell capacity decay rate. For example, for a sample data point, its corresponding label data is the capacity decay rate of the corresponding battery cell at a certain historical moment, such as 5%.

[0114] In this invention, a Long Short-Term Memory (LSTM) neural network is selected for training. LSTM can process sequential data and has the ability to remember long-term information. In battery cell performance prediction, the historical EIS data and charge / discharge curve data of the battery cell are both sequential data, and the performance change of the battery cell is a long-term process. LSTM can better capture the long-term dependencies in these data, thereby improving the accuracy of the prediction.

[0115] In this invention, the hidden layer of the predictive equalization model consists of two-layer LSTM units (128 neurons per layer). This two-layer LSTM structure further enhances the model's ability to learn complex sequence data. The 128 neurons per layer represent 128 processing units within each LSTM layer, capable of performing complex nonlinear transformations and feature extraction on the input data. Through the combination of multiple layers and neurons, the model can learn deeper features and patterns from cell EIS data and charge / discharge curve data.

[0116] The output layer of the predictive equalization model can predict the cell capacity decay rate and equalization priority ranking after 72 hours. In this invention, the trained LSTM model can not only predict the capacity decay of a cell in the next 72 hours, but also rank the cells' equalization priorities based on the prediction results. For example, the predictive equalization model may output that a certain cell's capacity decay rate is 8% after 72 hours, and based on this decay rate and the prediction results of other cells, rank this cell in a higher equalization priority position.

[0117] In this invention, the training objective is to reduce the model's prediction error to less than 2%. By continuously adjusting the parameters of the LSTM model (such as weights and biases), the error between the model's predicted value (cell capacity decay rate after 72 hours) and the actual labeled data (the true cell capacity decay rate corresponding to historical moments) is minimized, given the sample data. When the prediction error is less than 2%, it indicates that the model has high prediction accuracy and can provide a reliable basis for balanced cell management.

[0118] Optionally, in this invention, a predictive balancing model is set to run once a day, and when the state of health (SOH) of a cell is below 80%, a balancing plan is triggered 48 hours in advance. If the prediction result shows that the SOH of a certain cell is below 80%, it indicates that the performance of the cell has deteriorated to a certain extent, and a balancing plan needs to be formulated and triggered 48 hours in advance to prevent further deterioration of the cell's performance and ensure the overall performance and safety of the battery pack.

[0119] Based on the above embodiments, the step of performing corresponding energy transfer operations on the target cells in the battery pack according to the predicted cell capacity decay rate and the battery equalization priority ranking result includes:

[0120] Based on the predicted cell capacity decay rate and the battery equalization priority ranking result, an energy distribution equalization instruction corresponding to the target cell is generated.

[0121] Based on the bidirectional DC / DC converter and the energy distribution balancing command, an energy transfer operation is performed on the target battery cell.

[0122] In this invention, the cell capacity decay rate is a core indicator for measuring the degree of performance degradation of a cell, reflecting the change in the amount of charge that a cell can store and release over time during use. For example, a cell with an initial capacity of 100Ah may have its capacity decay to 80Ah after a period of use; its capacity decay rate can then be calculated. A higher capacity decay rate indicates a more severe decline in cell performance, necessitating equalization operations to restore or maintain the overall performance of the battery pack.

[0123] The battery balancing priority ranking is determined by comprehensively evaluating various factors such as the health status and performance parameters of each cell in the battery pack. For example, cells with lower SEI stability, abnormal charge transfer efficiency, or unsatisfactory lithium-ion diffusion coefficients may be given higher priority. This ranking helps to prioritize the cells that most need balancing when resources are limited, thereby improving balancing efficiency and effectiveness.

[0124] Furthermore, by combining the predicted cell capacity degradation rate and the balancing priority ranking results, it is determined which cells require energy transfer operations; these cells are the target cells. For example, cells ranked high in the ranking and with high capacity degradation rates are prioritized as target cells. Based on the specific situation of the target cells, an energy allocation balancing command is generated, including parameters such as balancing current magnitude and balancing time. The adjustable balancing current range is supported from 0.1A to 5A, and in actual operation, it will be adjusted according to the state of the target cells. For example, if a single cell has a low SEI stability index and a high charge transfer efficiency, the balancing current needs to be reduced (e.g., selecting a smaller value within the 0.1A to 5A range, such as 0.5A) to protect the SEI. Considering the high charge transfer efficiency, the balancing time may be appropriately shortened or a suitable balancing duration may be maintained to achieve the overall balancing effect. For cells with severe capacity degradation and poor health, a larger balancing current (e.g., 3A to 5A) may be selected to accelerate the energy transfer speed, while other factors are also considered to ensure operational safety.

[0125] Bidirectional DC / DC converters are key devices for energy distribution between battery clusters or battery boxes. They can change the magnitude and direction of DC voltage, enabling bidirectional energy transfer between different cells, battery clusters, or battery boxes. For example, in a battery pack, when one battery cluster has a higher charge and another has a lower charge, a bidirectional DC / DC converter can transfer energy from the higher-charge cluster to the lower-charge cluster, achieving a more efficient energy distribution.

[0126] This invention uses a bidirectional DC / DC converter to support multiple energy distribution methods, meeting the needs of different scenarios. It can precisely control the direction and amount of energy transfer according to the energy distribution balancing command, and can flexibly realize energy distribution between clusters or between boxes.

[0127] Based on the above embodiments, the method further includes:

[0128] The stability index of the solid electrolyte interface film, the charge transfer efficiency value, and the lithium ion diffusion coefficient are compared with their respective preset thresholds.

[0129] Based on the comparison results, the cell health status data of the battery cell is updated.

[0130] In this invention, the preset thresholds include a first preset threshold, a second preset threshold, and a third preset threshold, which are the preset thresholds corresponding to the solid electrolyte interface film stability index, charge transfer efficiency value, and lithium ion diffusion coefficient, respectively.

[0131] When comparing the stability index of the solid electrolyte interphase (SEI) membrane, the first preset threshold can be set to 800. The SEI stability index is a quantitative indicator used to measure the stability of the SEI; its value reflects the structural integrity and chemical stability of the SEI during battery charging and discharging. The SEI stability index obtained through real-time monitoring or calculation is compared with the first preset threshold. If the SEI stability index is below 800, it indicates poor SEI stability. For example, during long-term battery use, due to side reactions, the SEI may gradually thicken and develop cracks, leading to a decrease in its stability. In this case, the monitored stability index will be below 800. When this occurs, it indicates that the SEI may further deteriorate during subsequent charging and discharging processes, thus affecting the battery's performance and lifespan.

[0132] When comparing charge transfer efficiency values, the second preset threshold can be set to 3. Charge transfer efficiency reflects the battery's ability to transfer charge between the electrode and electrolyte interface, and is an important parameter for evaluating the battery's electrochemical reaction kinetics. In this invention, the obtained charge transfer efficiency value is compared with the second preset threshold. If the charge transfer efficiency value is lower than 3, it indicates insufficient charge transfer capability. For example, when electrode materials age, the electrode surface becomes contaminated, or the electrolyte composition changes, charge transfer at the interface is hindered, resulting in reduced charge transfer efficiency. In this case, during charging or discharging, charge cannot be transferred quickly and effectively between the electrode and electrolyte, affecting the battery's charging and discharging efficiency and performance.

[0133] When comparing lithium-ion diffusion coefficients, the third preset threshold can be set to 0.5. The lithium-ion diffusion coefficient is a physical quantity describing the ease with which lithium ions diffuse within the electrode material, directly affecting the battery's charging and discharging speed and performance. In this invention, the measured lithium-ion diffusion coefficient is compared with the third preset threshold. When the lithium-ion diffusion coefficient is below 0.5, it indicates that the diffusion ability of lithium ions in the electrode material is limited. For example, as the number of battery uses increases, the structure of the electrode material may change, causing the lithium-ion diffusion channels to narrow or become blocked, thus reducing the lithium-ion diffusion coefficient. In this case, during battery charging and discharging, lithium ions cannot diffuse in the electrode material in a timely and uniform manner, easily leading to localized overcharging or over-discharging.

[0134] In this invention, when the SEI stability index is below 800, due to poor SEI stability, it may be necessary to reduce the charge / discharge rate to mitigate further SEI degradation. When updating cell health status data, states with poor SEI stability are marked in the relevant data records, and recommended measures (reducing the charge / discharge rate) are recorded. Simultaneously, based on the degree of SEI stability decline, the overall cell health score is adjusted accordingly, for example, by lowering the score to reflect the impact of SEI degradation on cell health.

[0135] If the charge transfer efficiency value is below 3, it indicates insufficient charge transfer capability, and the battery should be prioritized for charge equalization. When updating the cell health status data, the cell will be marked as requiring priority charge equalization, and the specific details of the low charge transfer efficiency will be recorded. Furthermore, based on the magnitude of the decrease in charge transfer efficiency, the cell's health status will be quantitatively assessed, such as lowering the cell's health level or health score, to reflect the negative impact of charge transfer issues on cell performance.

[0136] When the lithium-ion diffusion coefficient is below 0.5, it indicates that the lithium-ion diffusion capacity is limited, and the balancing current needs to be adjusted to avoid the risk of overcharging or over-discharging. When updating cell health data, situations with low lithium-ion diffusion coefficients are recorded, and a recommendation to adjust the balancing current is noted. Simultaneously, based on the degree of limited lithium-ion diffusion, the cell's health status is updated; for example, the cell's safety factor or health indicators may be lowered to remind users of potential risks during charging and discharging.

[0137] This invention compares the stability index of the solid electrolyte interface film, the charge transfer efficiency value, and the lithium-ion diffusion coefficient with corresponding preset thresholds, and updates the cell health status data based on the comparison results. This allows for a comprehensive and accurate understanding of the cell's health status, providing a scientific basis for battery management and maintenance, thereby extending battery life and improving the safety and reliability of the battery system.

[0138] Figure 2 This is a schematic diagram of the overall process of the present invention, which can be referred to. Figure 2 As shown, firstly, the voltage difference value Vdiff is collected, and then it is determined whether Vdiff is greater than 30mV. If Vdiff is greater than 30mV, it indicates that the voltage difference between the cells is large, and real-time balancing is required immediately, entering the real-time balancing process on the left. If Vdiff is not greater than 30mV, it enters the predictive balancing stage on the right.

[0139] In the real-time balancing process, the LMS algorithm is activated for fast response: when Vdiff is greater than 30mV, the Least Mean Square (LMS) algorithm is activated for fast response, and then combined with the RLS algorithm to adjust the balancing current Ieq. In this invention, based on the fast response of the LMS algorithm, the Recursive Least Squares (RLS) algorithm is combined to further precisely adjust the balancing current Ieq. The RLS algorithm can dynamically adjust the balancing current according to the real-time voltage difference and other parameters to achieve a better balancing effect. Furthermore, balancing priority is planned. In this process, based on the voltage difference of the cells and other relevant parameters, the balancing priority of each cell is planned to ensure that the battery management system can prioritize the balancing operation of cells with large voltage differences. Finally, the battery management system outputs a specific balancing plan, including information such as the cells to be balanced, the magnitude of the balancing current, and the balancing time, in order to execute the balancing operation.

[0140] After entering the predictive equalization process, the system first inputs the most recent 10 cycles of EIS data and charge / discharge curve segments. This data contains performance information of the cells under different conditions, providing input for the predictive model. Then, the predictive equalization model is run. This model, based on historical data and machine learning algorithms, can predict the future capacity degradation of the cells. Based on the input data, the model predicts the capacity degradation rate after 72 hours, thus understanding the health status of the cells in advance and providing a basis for subsequent equalization operations. Further, based on the predicted capacity degradation rate, the system generates an equalization priority ranking for each cell. Cells with higher capacity degradation rates are prioritized for equalization to extend the overall lifespan of the battery pack. Finally, the system outputs the prediction results, including the capacity degradation rate and equalization priority ranking of each cell, for subsequent equalization operations and management decisions. The entire process ends after completing real-time or predictive equalization operations. The system continues to cyclically monitor the voltage differences between the cells to ensure the balanced state of the battery pack.

[0141] The adaptive active balancing device for the battery management system provided by the present invention is described below. The adaptive active balancing device for the battery management system described below can be referred to in correspondence with the adaptive active balancing method for the battery management system described above.

[0142] Figure 3This is a schematic diagram of the adaptive active balancing device for the battery management system provided by the present invention, as shown below. Figure 3 As shown, this invention provides an adaptive active balancing device for a battery management system, including a testing unit 301, a decision-making unit 302, and an energy transfer unit 303. The testing unit 301 acquires electrochemical impedance spectroscopy (EIS) test data and charge-discharge curve segments corresponding to each cell in the battery pack. The decision-making unit 302, when determining that the current voltage difference of the cell is less than or equal to a preset voltage difference threshold, inputs the EIS test data and the charge-discharge curve segments into a predictive balancing model to obtain a predicted cell capacity decay rate output by the predictive balancing model, and obtains a battery balancing priority ranking result based on the predicted cell capacity decay rate. The energy transfer unit 303 performs corresponding energy transfer operations on the target cells in the battery pack based on the predicted cell capacity decay rate and the battery balancing priority ranking result.

[0143] The adaptive active balancing device for battery management system provided by this invention acquires electrochemical impedance spectroscopy test data and charge-discharge curve segments of each cell in the battery pack. When the cell voltage difference value meets the standard, it is input into a predictive balancing model based on neural network training to obtain the predicted value of cell capacity decay rate and sort it. Finally, based on this result, energy transfer operation is performed on the target cell to achieve accurate assessment and efficient balancing management of cell aging.

[0144] Based on the above embodiments, the device further includes a feature extraction unit, used to perform the following steps: calculating the solid electrolyte interface membrane stability index based on the reciprocal of the product between the solid electrolyte interface membrane impedance data and the solid electrolyte interface membrane capacitance; calculating the charge transfer efficiency value based on the reciprocal of the charge transfer resistance data; calculating the lithium ion diffusion coefficient based on the square of the reciprocal of the diffusion impedance data; and constructing the electrochemical impedance spectroscopy test data based on the solid electrolyte interface membrane stability index, the charge transfer efficiency value, and the lithium ion diffusion coefficient.

[0145] Figure 4 This is a schematic diagram of the deployment of the adaptive active balancing device of the battery management system provided by the present invention on an energy storage architecture, which can be referred to. Figure 4 As shown, the test unit can be set in the Battery Management Unit (BMU) corresponding to each battery module, the feature extraction unit and the decision unit can be set in the Battery Control Unit (BCU), and the energy transfer unit can be set in each active equalization board.

[0146] Figure 5This is a schematic diagram of the battery management system provided by the present invention, as shown below. Figure 5 As shown, the present invention provides a battery management system, including the battery management system adaptive active balancing device 501 described in the above embodiments.

[0147] The battery management system provided by this invention acquires electrochemical impedance spectroscopy test data and charge-discharge curve segments of each cell in the battery pack. When the cell voltage difference value meets the standard, it inputs it into a prediction and equalization model based on neural network training to obtain the predicted value of cell capacity decay rate and sort it. Finally, based on this result, it performs energy transfer operation on the target cell to achieve accurate assessment and efficient equalization management of cell aging.

[0148] The apparatus provided in this embodiment of the invention is used to execute the above-described method embodiments. For specific processes and details, please refer to the above embodiments, which will not be repeated here.

[0149] Figure 6 This is a schematic diagram of the structure of the electronic device provided by the present invention, such as... Figure 6 As shown, the electronic device may include: a processor 601, a communication interface 602, a memory 603, and a communication bus 604, wherein the processor 601, the communication interface 602, and the memory 603 communicate with each other through the communication bus 604. The processor 601 can call logical instructions in the memory 603 to execute an adaptive active balancing method for the battery management system. This method includes: acquiring electrochemical impedance spectroscopy test data and charge-discharge curve segments corresponding to each cell in the battery pack; when it is determined that the current voltage difference value of the cell is less than or equal to a preset voltage difference threshold, inputting the electrochemical impedance spectroscopy test data and the charge-discharge curve segments into a predictive balancing model to obtain a predicted cell capacity decay rate output by the predictive balancing model, and obtaining a battery balancing priority ranking result based on the predicted cell capacity decay rate; and performing corresponding energy transfer operations on the target cells in the battery pack based on the predicted cell capacity decay rate and the battery balancing priority ranking result.

[0150] Furthermore, the logical instructions in the aforementioned memory 603 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0151] On the other hand, the present invention also provides a computer program product, the computer program product including a computer program stored on a non-transitory computer-readable storage medium, the computer program including program instructions, when the program instructions are executed by the computer, the computer is able to execute the adaptive active balancing method of the battery management system provided by the above methods, the method including: acquiring electrochemical impedance spectroscopy test data and charge-discharge curve segments corresponding to each cell in the battery pack; when it is determined that the current voltage difference value of the cell is less than or equal to a preset voltage difference threshold, inputting the electrochemical impedance spectroscopy test data and the charge-discharge curve segments into a predictive balancing model to obtain a predicted value of cell capacity decay rate output by the predictive balancing model, and obtaining a battery balancing priority ranking result based on the predicted value of cell capacity decay rate; and performing corresponding energy transfer operations on the target cells in the battery pack according to the predicted value of cell capacity decay rate and the battery balancing priority ranking result.

[0152] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the adaptive active balancing method for the battery management system provided in the above embodiments. The method includes: acquiring electrochemical impedance spectroscopy test data and charge-discharge curve segments corresponding to each cell in the battery pack; when it is determined that the current voltage difference value of the cell is less than or equal to a preset voltage difference threshold, inputting the electrochemical impedance spectroscopy test data and the charge-discharge curve segments into a predictive balancing model to obtain a predicted cell capacity decay rate output by the predictive balancing model, and obtaining a battery balancing priority ranking result based on the predicted cell capacity decay rate; and performing a corresponding energy transfer operation on the target cell in the battery pack based on the predicted cell capacity decay rate and the battery balancing priority ranking result.

[0153] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0154] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0155] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A battery management system adaptive active balancing method, characterized in that, include: Acquire electrochemical impedance spectroscopy test data and charge-discharge curve segments for each cell in the battery pack; If the current voltage difference of the battery cell is determined to be less than or equal to a preset voltage difference threshold, the electrochemical impedance spectroscopy test data and the charge-discharge curve segment are input into the prediction equalization model to obtain the predicted value of the battery cell capacity decay rate output by the prediction equalization model, and the battery equalization priority ranking result is obtained based on the predicted value of the battery cell capacity decay rate. Based on the predicted cell capacity decay rate and the battery equalization priority ranking result, perform the corresponding energy transfer operation on the target cell in the battery pack. The acquisition of electrochemical impedance spectroscopy test data for each cell in the battery pack includes: Electrochemical impedance spectroscopy was performed on the battery cell to collect data on the solid electrolyte interface membrane impedance, charge transfer resistance, and diffusion impedance. Based on the solid electrolyte interface membrane impedance data, the charge transfer resistance data, and the diffusion impedance data, the electrochemical impedance spectroscopy test data is obtained; The method further includes: The current voltage difference value of the battery cell is determined based on the least mean square algorithm. If it is determined that the current voltage difference value of the battery cell is greater than the preset voltage difference threshold, an equalization current adjustment operation is performed on the battery cell based on the recursive least squares method, the voltage difference value, the solid electrolyte interface membrane impedance data, and the charge transfer resistance change, wherein the charge transfer resistance change is obtained based on the charge transfer resistance data.

2. The battery management system adaptive active balancing method of claim 1, wherein, The process of obtaining the electrochemical impedance spectroscopy test data based on the solid electrolyte interface membrane impedance data, the charge transfer resistance data, and the diffusion impedance data includes: The stability index of the solid electrolyte interface membrane is calculated based on the reciprocal of the product between the solid electrolyte interface membrane impedance data and the solid electrolyte interface membrane capacitance. The charge transfer efficiency value is calculated based on the reciprocal of the charge transfer resistance data. The lithium-ion diffusion coefficient is calculated based on the reciprocal square of the diffusion impedance data. The electrochemical impedance spectroscopy test data are constructed based on the solid electrolyte interface film stability index, the charge transfer efficiency value, and the lithium ion diffusion coefficient.

3. The battery management system adaptive active balancing method of claim 1, wherein, The predictive equilibrium model is trained through the following steps: Sample data were constructed based on historical electrochemical impedance spectroscopy sample data and charge-discharge curve samples; Obtain the sample values ​​of the cell capacity decay rate corresponding to the sample data at historical moments, and construct the label data of the sample data based on the sample values ​​of the cell capacity decay rate; Based on the sample data and the label data, the long short-term memory neural network is trained to obtain the prediction equilibrium model.

4. The battery management system adaptive active equalization method of claim 1, wherein, The step of performing corresponding energy transfer operations on the target cells within the battery pack based on the predicted cell capacity decay rate and the battery balancing priority ranking result includes: Based on the predicted cell capacity decay rate and the battery equalization priority ranking result, an energy distribution equalization instruction corresponding to the target cell is generated. Based on the bidirectional DC / DC converter and the energy distribution balancing command, an energy transfer operation is performed on the target battery cell.

5. The battery management system adaptive active balancing method of claim 2, wherein, The method further includes: The stability index of the solid electrolyte interface film, the charge transfer efficiency value, and the lithium ion diffusion coefficient are compared with their respective preset thresholds. Based on the comparison results, the cell health status data of the battery cell is updated.

6. A battery management system adaptive active equalization device, characterized in that, The apparatus is used to implement the adaptive active balancing method of the battery management system as described in any one of claims 1 to 5.

7. A battery management system, characterized by, Includes the adaptive active balancing device for the battery management system as described in claim 6.

Images