Lithium ion battery equalizing charge control method and system based on BMS

Abstract

The invention relates to the field of battery charging control, in particular to a lithium ion battery equalizing charge control method and system based on a BMS (Battery Management System). The method comprises the following steps: continuously obtaining battery cell monitoring parameters of a battery charging state based on a BMS, and carrying out high-frequency filtering and denoising processing and battery cell charging behavior analysis to obtain a battery cell behavior characteristic time sequence; identifying a charging voltage response signal of each battery cell based on a BMS, performing multi-frequency impedance change trend evolution, performing battery cell charging response characteristic perception based on a battery cell behavior characteristic time sequence, and constructing a plurality of battery cell charging state maps; and carrying out quantitative calculation on charging energy flow affinity according to the battery cell monitoring parameters, carrying out state deviation trend distinguishing by combining the plurality of battery cell charging state maps, constructing a topological connection diagram among the battery cells, and carrying out energy deviation complementary path mining and dynamic deviation balance planning on weak battery cells and strong battery cells. According to the invention, efficient and accurate energy balanced distribution and charging control are realized.

Description

Technical Field

[0001] The present invention relates to the field of battery charging control, and in particular to a BMS-based lithium-ion battery balanced charging control method and system. Background Art

[0002] With the rapid development of new energy vehicles, energy storage systems, and portable electronic devices, lithium-ion batteries (Li-ion batteries) have become the mainstream electrochemical energy storage device due to their high energy density, long cycle life, and low self-discharge rate. Battery management systems (BMS) are widely used in various power systems and energy storage systems to ensure safe, reliable, and efficient operation of batteries. They perform key functions such as real-time monitoring, state estimation, and charge and discharge control of Li-ion batteries. In large-capacity battery packs composed of multiple series-connected cells, manufacturing variations and varying degrees of aging can easily lead to voltage and charge inconsistencies between cells, further exacerbating battery imbalance and, in severe cases, compromising the performance, safety, and service life of the entire pack. During actual operation, cell voltage differences in Li-ion battery packs accumulate, causing some cells to overcharge or overdischarge during charging or discharging, reducing the system's available capacity and posing significant safety risks. To mitigate this imbalance and improve overall battery pack performance, cell balancing technology has become a core function in BMS systems. Common balancing methods currently include passive balancing and active balancing. Passive balancing consumes the excess energy of high-voltage batteries as heat, which has a simple structure but low efficiency. Active balancing redistributes energy between batteries through electrical energy transfer. Although it is more efficient, the system complexity and cost also increase accordingly.

[0003] However, existing lithium-ion battery balancing and charge control methods still face numerous challenges. For one thing, traditional balancing strategies are often based on static threshold judgments, making them difficult to adapt to complex and changing operating environments and battery state variations, resulting in suboptimal balancing results and potentially even introducing new battery safety risks. Furthermore, existing BMS systems often have a delayed response to balancing control and lack the ability to comprehensively assess factors such as battery aging and dynamic load variations, making efficient and accurate energy distribution and charging control impossible. Therefore, a more intelligent and dynamically responsive lithium-ion battery balancing and charge control method and system is urgently needed. This method and system can fully leverage the real-time data acquisition capabilities of the BMS and implement precise and dynamic balancing control strategies through in-depth analysis and modeling of battery operating states. Leveraging advanced algorithm design, state estimation models, and energy scheduling mechanisms, this method should effectively improve battery pack consistency and efficiency, extend the overall system lifespan, and significantly enhance the battery system's safety capabilities under complex operating conditions. Building a lithium-ion battery balancing and charge control system for intelligent BMS platforms is a key direction for driving the development of next-generation battery management technologies. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention proposes a lithium-ion battery balanced charging control method and system based on BMS to solve at least one of the above technical problems.

[0005] To achieve the above object, the present invention provides a lithium-ion battery balanced charging control method based on a BMS, comprising the following steps: Step S1: Based on the BMS, the battery cell monitoring parameters of the battery charging status are continuously acquired, and high-frequency filtering and noise removal processing and battery cell charging behavior analysis are performed to obtain a battery cell behavior characteristic time series sequence; Step S2: Based on the BMS, the charging voltage response signal of each battery cell is identified, the multi-frequency impedance change trend evolution is performed, and the battery cell charging response characteristics are perceived based on the battery cell behavior characteristic time series sequence, and multiple battery cell charging state maps are constructed; Step S3: performing quantitative calculation of charging energy flow affinity according to the cell monitoring parameters, and distinguishing state deviation trends based on the multiple cell charging state maps, thereby classifying weak cells and strong cells; Step S4: Construct a topological connection diagram between cells, and mine complementary paths for energy offsets of weak cells and strong cells, perform dynamic offset balancing planning, and build an intelligent offset path planning strategy; Step S5: performing asynchronous hysteresis predictive drive adjustment according to the cell monitoring parameters, and performing real-time dynamic charge balancing control based on an intelligent offset path planning strategy to execute an intelligent balanced charge control operation.

[0006] In this specification, a BMS-based lithium-ion battery balanced charging control system is provided, which is used to execute the BMS-based lithium-ion battery balanced charging control method described above, including: The charging behavior analysis module is used to continuously obtain the battery cell monitoring parameters of the battery charging status based on the BMS, perform high-frequency filtering and noise removal, and analyze the battery cell charging behavior to obtain the battery cell behavior characteristic time series sequence; The state map module is used to identify the charging voltage response signal of each battery cell based on the BMS, perform multi-frequency impedance change trend evolution, and perceive the battery cell charging response characteristics based on the battery cell behavior characteristic time series sequence to construct multiple battery cell charging state maps; A cell status classification module is used to perform quantitative calculation of charging energy flow affinity based on the cell monitoring parameters, and distinguish state deviation trends based on the multiple cell charging state maps, thereby classifying weak cells and strong cells; The complementary path mining and planning module is used to construct a topological connection diagram between cells, and to mine complementary paths for the energy offset of weak and strong cells, perform dynamic offset balancing planning, and build an intelligent offset path planning strategy. The asynchronous balancing adjustment module is used to perform asynchronous hysteresis predictive drive adjustment according to the battery cell monitoring parameters, and perform real-time dynamic charging balancing control based on the intelligent offset path planning strategy to perform intelligent balanced charging control operations.

[0007] The beneficial effects of this invention include: By performing high-frequency filtering on battery cell monitoring parameters (such as voltage, current, and temperature), environmental noise and signal distortion during the BMS acquisition process are effectively removed, improving data stability. The dynamic behavior patterns of battery cells during charging (such as voltage fluctuation trends and charging rate changes) are analyzed to establish a time-series sequence of cell behavior characteristics, providing a high-resolution dynamic feature foundation for subsequent analysis. This enhances early detection of abnormal battery cell behavior (such as overcharging and sudden changes in internal resistance), improving system safety and reliability. Multi-frequency impedance change analysis of the voltage-time curves of each battery cell's charging response reveals differences in internal chemical processes and identifies potential risks such as cell aging and imbalance. By integrating time-series behavior characteristics, the response dynamics of each battery cell are comprehensively perceived, and a cell charging state map is constructed, achieving a spatial representation of the multidimensional performance status of the battery cell. This map provides the input foundation for advanced algorithms such as graph neural networks for subsequent state assessment and path optimization, improving data mining accuracy. A quantitative calculation method for energy flow affinity is introduced to identify the "affinity" between different battery cells during energy exchange and flow, thereby accurately classifying the charge and discharge compatibility between cells. Combined with the cell state map, this method analyzes the state deviation trends between cells and systematically distinguishes between "weak cells" (degraded performance or low charging capacity) and "strong cells" (responsive and healthy performance). This provides a key foundation for energy scheduling and helps implement a "weak-strong complementarity" strategy to avoid uneven energy distribution or excessive loading of individual cells. A topological connection graph is constructed between cells, systematically representing the connection structure and interchangeable paths between each cell, improving the accuracy of overall system modeling. Graph analysis techniques are used to discover energy compensation paths between weak and strong cells, enabling dynamic energy shifting from excess to insufficient cells. By matching the actual cell demand with the path capacity, a personalized energy shift path planning strategy is developed, improving charge balancing efficiency and system adjustment flexibility. Identifying asynchronous and delayed cell responses (e.g., slow response and regulation in some cells), and proactively adjusting parameters through a predictive drive mechanism for preventive control. Based on the aforementioned intelligent offset path planning strategy, real-time charge balancing control is implemented, independently of traditional static balancing circuits, dynamically adapting to changes in cell state. This significantly improves charging efficiency, extends the overall life of the battery pack, reduces energy loss, and enhances system energy efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 This is a schematic flow chart of the steps of a BMS-based lithium-ion battery balanced charging control method of the present invention; Figure 2 Detailed implementation flow chart of step S1; Figure 3 Detailed implementation flow chart of step S2; Figure 4Schematic diagram of the detailed implementation steps of step S3. DETAILED DESCRIPTION

[0009] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0010] This application provides a BMS-based lithium-ion battery balanced charging control method and system. The execution entities of the BMS-based lithium-ion battery balanced charging control method and system include, but are not limited to, mechanical equipment, data processing platforms, cloud server nodes, network upload devices, etc. that are equipped with the system, which can be regarded as general computing nodes of this application. The data processing platform includes, but is not limited to, at least one of an audio and image management system, an information management system, and a cloud data management system.

[0011] See also Figures 1 to 4 The present invention provides a lithium-ion battery balanced charging control method based on BMS, comprising the following steps: Step S1: Based on the BMS, the battery cell monitoring parameters of the battery charging status are continuously acquired, and high-frequency filtering and noise removal processing and battery cell charging behavior analysis are performed to obtain a battery cell behavior characteristic time series sequence; Step S2: Based on the BMS, the charging voltage response signal of each battery cell is identified, the multi-frequency impedance change trend evolution is performed, and the battery cell charging response characteristics are perceived based on the battery cell behavior characteristic time series sequence, and multiple battery cell charging state maps are constructed; Step S3: performing quantitative calculation of charging energy flow affinity according to the cell monitoring parameters, and distinguishing state deviation trends based on the multiple cell charging state maps, thereby classifying weak cells and strong cells; Step S4: Construct a topological connection diagram between cells, and mine complementary paths for energy offsets of weak cells and strong cells, perform dynamic offset balancing planning, and build an intelligent offset path planning strategy; Step S5: performing asynchronous hysteresis predictive drive adjustment according to the cell monitoring parameters, and performing real-time dynamic charge balancing control based on an intelligent offset path planning strategy to execute an intelligent balanced charge control operation.

[0012] In the embodiment of the present invention, see Figure 1 , is a schematic flow chart of the steps of a BMS-based lithium-ion battery balanced charging control method of the present invention. In this example, the steps of the BMS-based lithium-ion battery balanced charging control method include: Step S1: Based on the BMS, the battery cell monitoring parameters of the battery charging status are continuously acquired, and high-frequency filtering and noise removal processing and battery cell charging behavior analysis are performed to obtain a battery cell behavior characteristic time series sequence; In this embodiment, battery cell charging state parameters are continuously collected. These include key operating variables such as each cell's instantaneous voltage, current, temperature, internal resistance, and SOC (State of Charge). To ensure data continuity and high accuracy, the BMS sets a sampling frequency of every 100ms, using a 32-bit ADC for analog sampling with accuracy within ±1mV (voltage) and ±0.1°C (temperature). Data is uploaded to the central processing unit via the CAN bus protocol and enters the filtering and feature analysis process. After data collection is complete, the system first performs high-frequency filtering and noise reduction on the acquired battery cell monitoring data. Because the charging process may be affected by current disturbances, temperature fluctuations, and electromagnetic interference, the raw data may contain high-frequency jitter and abnormal spikes. To address this, the system combines a sliding window average filter (with a window size of 5 sampling points) with a weighted median filter to first smooth out abrupt data changes and then remove outliers with deviations greater than three times the standard deviation. After processing, the cell voltage curve becomes smoother and the internal resistance trend becomes more continuous, laying the foundation for subsequent behavioral modeling. After filtering, the system enters the cell charging behavior analysis phase. The core of this phase is to extract the behavioral characteristics of each cell within a specific time period by analyzing the interactive changes between multiple monitoring parameters. During the constant-current charging phase, the system observes changes in the voltage rise slope, the SOC growth rate, and the temperature rise amplitude to identify cell response sluggishness or heat dissipation abnormalities. During the constant-voltage phase, the system focuses on the duration of the voltage plateau and the internal resistance trend to determine whether there is a deviation from the charge saturation response. Specifically, the system incorporates a "perturbation voltage fluctuation analysis method." This method monitors the voltage response sensitivity under small current perturbations (for example, introducing a ±0.5A perturbation signal to the original charging current) to assess the cell's electrochemical activity and interfacial polarization. Finally, the system constructs a "cell behavior characteristic time series" by converting these behavioral characteristic parameters into a time series. This series is recorded for each cell individually, including the SOC change rate, internal resistance change amplitude, voltage slope, temperature gradient, and other information at the corresponding timestamp. Taking experimental data as an example, for a battery pack consisting of 96 battery cells, after a 1-hour constant current charging cycle, the system can record 57,600 sets of behavioral characteristic data points (sampled 10 times per second), and perform clustering and trend modeling on them, providing accurate data support for subsequent state identification and balance control.

[0013] Step S2: Based on the BMS, the charging voltage response signal of each battery cell is identified, the multi-frequency impedance change trend evolution is performed, and the battery cell charging response characteristics are perceived based on the battery cell behavior characteristic time series sequence, and multiple battery cell charging state maps are constructed; In this embodiment, the voltage response signal of each battery cell during the charging process is captured in real time by the BMS. During this process, the BMS uses a millisecond sampling frequency (for example, 10ms / time) to collect the voltage response of each battery cell under a specific perturbation current excitation. The excitation current is intermittently injected into the charging circuit with a controllable amplitude (such as ±0.5A), and the duration is controlled between 200 and 500ms to avoid interfering with the actual charging process while stimulating the equivalent impedance characteristics inside the battery cell. The voltage response signal is captured by the BMS's high-precision ADC module (16 bits or above) and enters the feature extraction process immediately after signal acquisition. The system then processes the voltage response signal and enters the multi-frequency impedance change trend evolution analysis stage. The goal of this stage is to extract the voltage-current relationship of the battery cell at different frequency responses through a multi-scale time-frequency decomposition method, thereby deriving the equivalent impedance spectrum line. Common methods include wavelet packet decomposition and short-time Fourier transform (STFT), which decompose the original response signal into multiple frequency bands, such as 0.1Hz, 0.5Hz, 1Hz, and 5Hz, from which the complex impedance variation trends at different frequencies can be derived. In the experimental environment, by superimposing the perturbation current, the system accurately observes the amplitude-frequency response of each battery cell under different frequency excitations. The analysis focuses on the evolution trend of the impedance modulus, the rate of change of the phase angle, and the evolution path of the equivalent series resistance and capacitance effects. Next, the system synchronizes and jointly analyzes the extracted impedance evolution characteristics with the time series of battery cell behavioral characteristics established in step S1. By constructing a one-to-one timestamp mapping model, the system integrates behavioral characteristics (such as SOC change rate, voltage slope, and temperature gradient) with impedance characteristics (such as impedance amplitude at specific frequencies and fitted pseudo-impedance trajectories), thereby identifying the response characteristics of each battery cell over different time periods. If the impedance modulus of a battery cell at a frequency of 0.5Hz suddenly rises, accompanied by a decrease in its SOC climbing rate, it may indicate a trend of lithium deposition, increased polarization, or increased thermal diffusion resistance. Ultimately, the system visualizes and models the above-mentioned fusion features in the form of a time series graph to construct multiple "battery cell charging state graphs." Each graph represents the multi-dimensional charging state evolution process of a battery cell, with time on the horizontal axis and the combination of multiple state parameters (such as frequency domain impedance, behavioral characteristic indicators, local thermal response, etc.) on the vertical axis. Taking 96 battery cells as an example, the system generates 96 independent graphs, which clearly present the state consistency, offset, and possible aging path between the battery cells through heat maps, trend lines, and threshold markers.

[0014] Step S3: performing quantitative calculation of charging energy flow affinity according to the cell monitoring parameters, and distinguishing state deviation trends based on the multiple cell charging state maps, thereby classifying weak cells and strong cells; In this embodiment, the system quantitatively calculates the charging energy flow affinity. The raw parameters used include each cell's real-time voltage (V_cell), internal resistance (R_cell), current (I_cell), SOC (State of Charge), and temperature during the charging process. To comprehensively reflect a cell's ability to accept charging energy flow, the system designed a comprehensive "energy affinity index" that incorporates factors such as the voltage rise rate per unit time, the cell temperature rise, the internal resistance trend, and the SOC increase. In the experiment, the unit charging cycle was 10 seconds, with a sampling period of 100ms. The system evaluated each cell's energy absorption efficiency and reverse resistance during this cycle, generating a normalized affinity score. The value range is typically set between 0 and 1, with higher values ​​indicating better charging acceptance. The system then compares these energy affinity results with the cell charging state map generated in the previous step. In the map, the changing trends of each cell across multiple dimensions (such as frequency response, SOC evolution, and internal resistance trajectory) are time-series-based. By dynamically comparing the energy affinity profile with the energy affinity score, the system can effectively identify cells that exhibit consistently low affinity and negative profile trends during the global charging process (e.g., increasing internal resistance, increasing impedance modulus, and slowly increasing SOC). These cells are considered "weak cells." Cells that exhibit positive trends across multiple dimensions, such as consistently high affinity, fast charging response, relatively stable internal resistance, and linear SOC growth, are classified as "strong cells." To enhance the reliability of these judgments, the system also incorporates a state deviation trend analysis mechanism. This mechanism performs cluster comparative analysis of the historical state profiles of all cells, calculating the degree of deviation of each cell's state trajectory from the overall state center. Cells with large deviations, particularly those that consistently deviate from the average trajectory in a negative direction, are flagged as weak cells with a high risk of performance degradation. In measured data, if 5-8 of the 96 cells have an energy affinity below 0.35 and an average internal resistance above 10mΩ, while also experiencing a temperature fluctuation exceeding 3°C, the system automatically classifies them as weak cells and prioritizes them for balancing. Ultimately, the system labels weak and strong cells separately and records them in a unified cell status classification model, providing structured input for subsequent offset balancing control, energy path planning, and dynamic charging adjustment strategies. This classification result is also continuously updated through the BMS system, supporting online dynamic adjustments and achieving truly adaptive cell status identification and classification optimization.

[0015] Step S4: Construct a topological connection diagram between cells, and mine complementary paths for energy offsets of weak cells and strong cells, perform dynamic offset balancing planning, and build an intelligent offset path planning strategy; In this embodiment, the system first obtains internal battery module connection data based on the battery pack structure information stored in the BMS. This data is typically derived from the battery pack's manufacturing configuration file or identified through real-time detection methods (such as an embedded BMS diagnostic module with multi-channel sampling capabilities). By comprehensively scanning the cell number, physical connection method (series or parallel), and wiring topology, the system constructs a physical topological connection diagram between the cells. In practical applications, for example, in a battery pack consisting of 96 18650 cells, the system uses 12 cells as a module unit. Each module is connected in series, while modules may be connected in parallel or a mixed series-parallel configuration. The BMS constructs a graph structure using each cell as a node and each connection as an edge, thereby forming a complete topological network between cells. Next, the system marks the "weak cells" and "strong cells" identified in the previous step in this topological diagram and explores potential energy-shifting and complementary paths based on factors such as physical path distance, accessibility, thermal resistance correlation coefficient, and current path switching capability between cells. The system uses a weighted graph-based path search algorithm (such as the A* shortest path algorithm with heuristic factors or the weighted Dijkstra algorithm) to discover paths. The weighting factors take into account not only distance but also path switching costs, channel temperature rise coefficients, and bypass switch delays. In one experiment, the system limited the maximum energy transfer distance to no more than two module spans and set a path weight threshold of 0.8 (after normalization). The system successfully discovered 12 pairs of weak and strong cell combinations with good energy complementarity potential. After identifying feasible energy complementarity paths, the system further performs "dynamic offset balancing planning." This planning is not a static execution, but rather generates a dynamic scheduling sequence based on the dynamic changes in the SOC of each cell, real-time voltage differences, current path switching states, and thermal distribution. The system maps the scheduling sequence into an adjustable current control matrix and defines permissible energy transfer windows (no more than 30 seconds per cycle, with a maximum transfer current of no more than 1A, and prioritizing paths with temperatures below 45°C). This ensures efficient energy transfer from strong cells to weak cells, thereby achieving rapid SOC convergence. Ultimately, the system integrates all offset strategies and path information into an "intelligent offset path planning strategy" and injects it into the balancing control module of the BMS main control unit in real time. This strategy supports dynamic interruption, path reconstruction, and priority switching, and has the ability to adapt to environmental changes, avoid faults, and avoid high temperatures. In actual road condition simulation tests, the system's offset path planning strategy effectively shortened the battery cell SOC convergence time by an average of 18%, while reducing the local battery cell peak temperature rise by 3.2°C, significantly improving balancing charging efficiency and system safety.

[0016] Step S5: performing asynchronous hysteresis predictive drive adjustment according to the cell monitoring parameters, and performing real-time dynamic charge balancing control based on an intelligent offset path planning strategy to execute an intelligent balanced charge control operation.

[0017] In this example, the system first constructs a time-series trajectory of each cell's response during the charging process based on cell monitoring parameters collected by the BMS, including each cell's charging voltage, charging current, temperature, SOC, internal resistance, and response delay. Specifically for the charging pulse response sequence, the system analyzes perturbation signals during the charging process (e.g., a step current stimulus within 5ms) and uses a sliding window method (with a window width of 100ms and a step size of 10ms) to extract the delayed response interval of the voltage change. Based on this, the system estimates the response hysteresis parameters for each cell. These hysteresis parameters reflect the differences in the response speed of different cells to the same current stimulus and are an important indicator for measuring asynchronous behavior. The system then normalizes the extracted hysteresis parameters and uses them as a basis for predicting asynchronous behavior. The system uses time series-based nonlinear regression modeling (such as LSTM regression or local linear embedding regression) to predict the cell response state within the next few tens of seconds. Taking actual test data as an example, in a 96-cell 18650 battery pack, 16 cells exhibited response lags exceeding 50ms and significant SOC ramp-up hysteresis during charging. Using the aforementioned modeling, the system predicted that these cells would exhibit further response delays within the next 20 seconds. After this prediction, the system dynamically schedules the charging regulation mechanism, combining it with the "intelligent offset path planning strategy" generated in the previous step. Specifically, the system implements "pre-load reduction" and "delayed charging" strategies for cells predicted to exhibit hysteresis. It also phases in energy input to cells with strong response rates and, through the path selection module, prioritizes energy offset paths to achieve peak-shifting energy compensation. To ensure accurate and real-time regulation, the system integrates a high-resolution PWM current control unit at the hardware level, combined with the cell-side MOS control matrix, to dynamically adjust current on / off and amplitude based on the regulation strategy. During charge balancing, each strategy execution cycle is controlled within 500ms, ensuring timely response to each predicted asynchronous state. Ultimately, the entire charge balancing process is no longer a static "charge until full" mode. Instead, it uses real-time adjustments and path reconstruction to intelligently distribute loads and adjust peak loads based on the dynamic state of each cell. Experimental data shows that applying this strategy in simulated road conditions can shorten the entire battery pack charging time by approximately 7.6%, while reducing the SOC variation between cells from a maximum of 8.5% to within 1.2%, effectively achieving the goal of efficient, safe, and adaptive dynamic balancing control.

[0018] In this embodiment, refer to Figure 2, is a flowchart of the detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include: Based on BMS, millisecond-level micro-time sampling is performed to continuously obtain battery cell monitoring parameters of the battery charging status; Perform high-frequency filtering and denoising on the cell monitoring parameters to obtain filtered and denoised monitoring parameters; Calculate the real-time SOC parameters, internal resistance state, thermal parameters, charge and discharge efficiency, and voltage of each battery cell based on the filtered and denoised monitoring parameters to obtain the multi-dimensional state characteristics of the battery cell; Divide the multi-dimensional state characteristics of the battery cell into time series windows to obtain multi-dimensional state characteristics in different time periods; The charging behavior of the battery cell is analyzed based on the multi-dimensional state characteristics of different time periods. The internal resistance perturbation changes, micro-voltage fluctuations and transient thermal responses are analyzed to obtain the time series of the battery cell behavior characteristics.

[0019] In this embodiment, to achieve high-precision control of the lithium-ion battery's state of charge, a high-frequency, millisecond-level micro-time series data acquisition mechanism must first be implemented at the battery management system (BMS) level. This sampling system primarily collects data including cell voltage, current, temperature, and their rate of change. These parameters reflect the cell's current state of charge, temperature rise, and electrochemical stability. In the experimental design, data from each cell is continuously sampled with a 1ms sampling period. TI's BQ76PL455A-Q1 battery monitoring chip is used. This chip supports cascaded acquisition of up to 16 cells, with a sampling accuracy of ±1mV and a current sampling error within ±0.2%. The sampling system must integrate a real-time data transmission mechanism, implementing high-speed ADC conversion and CAN bus communication on an STM32 microcontroller platform to ensure stable upload of sampled data. A data acquisition period of 1ms means that 1000 sets of cell data can be acquired per second, capturing critical transient responses such as rapid voltage drops, current ripples, and transient temperature rises. During high-power charging (such as a 2C rate) or fast current switching stages (such as increasing from 0.5C to 1.5C), this micro-timing sampling is crucial for capturing the dynamic response of the battery cell at the millisecond level, providing basic data support for subsequent behavioral analysis and control.

[0020] Because high-speed data sampling is often accompanied by strong electromagnetic interference (EMI) and switching noise (such as pulsation from DC-DC converters and chargers), the raw sampled data inevitably contains high-frequency noise components. Therefore, after data acquisition, high-frequency filtering and noise suppression are required for the cell monitoring parameters. In this experiment, a multi-resolution analysis method based on the wavelet transform was used for filtering. This method has excellent localization capabilities in both time and frequency dimensions, effectively identifying short-term non-stationary noise. The wavelet filter specifically uses the db4 mother wavelet, reconstructing the signal through a three-layer decomposition to preserve the primary low-frequency characteristics of the cell response while removing high-frequency interference components above 50Hz. Compared with traditional moving average methods, the wavelet method more accurately preserves edge information during voltage jumps and current pulses, avoiding the loss of detail caused by oversmoothing. Furthermore, for hardware filtering, a second-order Butterworth low-pass filter with a cutoff frequency of 500Hz was introduced in the sampling front-end circuit to effectively reduce interference from externally coupled signals on the ADC input. Through collaborative filtering using hardware and software, the peak-to-peak noise amplitude in the raw voltage data can be reduced from 30mV to approximately 5mV, significantly improving the availability and accuracy of cell state parameters and laying the foundation for subsequent state calculations. After obtaining high-quality monitoring data, key state indicators for each cell are calculated based on physical and data-driven models to form a multi-dimensional state signature. First, SOC (State of Charge) calculation is performed using a model based on the Extended Kalman Filter (EKF). This method combines historical voltage and current data with the cell's open-circuit voltage (OCV) curve to dynamically update the SOC estimate. The initial SOC is estimated based on the pre-power-on charging history. Under 1C charging conditions, the EKF model maintains an error within ±2%. Internal resistance state calculation utilizes a modified version of the HPPC (Hybrid Pulse Power Characterization) method. 1C pulse current is injected at SOC levels of 60%, 40%, and 20%. Voltage sag and steady-state voltage drop are analyzed to separate ohmic internal resistance from polarization resistance. The internal resistance of each cell was measured at 2-4 mΩ at room temperature (25°C), rising to approximately 6 mΩ at 10°C, reflecting thermal correlation. Thermal parameters (such as thermal resistance and heat capacity) were modeled using equivalent thermal networks, derived from continuous temperature change data combined with heat flux calculations. The thermal resistance range is generally between 0.2°C / W and 0.5°C / W, depending on the cell packaging and heat dissipation structure. Charge and discharge efficiency, calculated based on coulombic efficiency and energy efficiency, was measured to be over 98%. Under standard 1C conditions, the efficiency of some cells decreased significantly at high temperatures (45°C). Ultimately, the multi-dimensional characteristics of each cell include more than eight indicators, including SOC, internal resistance, temperature, thermal conductivity parameters, voltage, current history, and charging efficiency. These are timestamped to form a cell state vector, providing the data foundation for temporal behavior modeling.After obtaining the multi-dimensional cell state parameters, they need to be structured in the time dimension to capture the stage-by-stage behavior during the charging process. A sliding window approach is used for sequence segmentation. Each window is set to 5 seconds in length, with a sliding step of 1 second, resulting in overlapping windows. This window design maintains continuity while enabling fine-grained segmentation of cell state change trends.

[0021] Within each window, statistics such as the mean, maximum, minimum, standard deviation, and gradient rate of change are calculated for each state feature within that time period. The rate of change of SOC reflects the stability of the charging current, while fluctuations in internal resistance indicate material structural changes or local thermal disturbances. Experiments observed that during the initial stages of CC (constant current) charging, voltage fluctuations were large, with standard deviations reaching 15mV. This fluctuation gradually stabilized at the end of CV (constant voltage) charging, decreasing to approximately 3mV. Furthermore, to enhance the characterization of dynamic characteristics, a Fourier transform was introduced to extract the distribution of each state feature across frequency bands. The spectral energy distribution can help identify abnormal periodic oscillations, such as high-frequency interference or load instability. Ultimately, each window generates a set of time series feature vectors containing indicators such as state trends, fluctuation amplitudes, and frequency components. These vectors are converted into a high-dimensional sequence data format, providing continuous input for subsequent cell behavior modeling and balancing strategy development. After obtaining structured time window features, in-depth analysis of cell behavior is required from multiple dimensions to uncover the underlying electrochemical dynamics. Using multivariate time series analysis, combined with principal component analysis (PCA) and dynamic time warping (DTW), micro-behavioral changes such as internal resistance perturbations, voltage micro-fluctuations, and transient thermal responses are identified. Internal resistance perturbations are primarily analyzed by analyzing small upward trends in internal resistance over short time windows and their coupling with temperature and current changes. When the SOC remains constant at 80%, if the internal resistance slowly rises from 2.5mΩ to 2.8mΩ over 5 minutes, accompanied by a slight temperature increase (<1°C), it can be preliminarily determined that the cell has entered a region of enhanced thermal coupling. Micro-voltage fluctuations are analyzed by extracting energy variations in the 1-10Hz range of the voltage signal using FFT transformation. A sudden increase in spectral energy in this range may indicate electrochemical imbalances caused by controller oscillation or material particle expansion during charging. Transient thermal response features are primarily extracted based on a temperature-current hysteresis response model. By analyzing the temperature change delay and maximum temperature rise rate before and after a sudden current change (e.g., 1C to 2C), heat diffusion paths and the effectiveness of the thermal management system can be estimated. Experimental data shows that shorter response times indicate stronger thermal coupling in the battery cell, which is beneficial for maintaining thermal stability. Through the above methods, a time-ordered sequence of battery cell behavior characteristics is ultimately constructed. This sequence not only includes numerical trends but also incorporates physically relevant "behavioral segments," such as "impedance surge," "stable charging," and "thermal-sensitive transition." This time sequence will serve as the basis for subsequent balanced charging strategies, guiding when to bypass a specific battery cell, adjust the charging current, or trigger the temperature control mechanism.

[0022] In this embodiment, refer to Figure 3 , is a flowchart of the detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include: Based on the BMS, a perturbation excitation signal is introduced into the charging current, and the charging voltage response signal of each battery cell is identified; Perform multi-time point response state fitting on the charging voltage response signal to construct a steady-state voltage response waveform; Perform multi-scale time-frequency decomposition on the steady-state voltage response waveform and extract the voltage response impedance spectra at different frequencies; Performing multi-frequency impedance change trend evolution on the voltage response impedance spectrum to obtain impedance change trend characteristics at different frequencies; Based on the impedance change trend characteristics and the cell behavior characteristic time series, the cell charging response characteristics are perceived and multiple cell charging state maps are constructed.

[0023] In this embodiment, during the balanced charging control process for lithium-ion batteries, conventional charging strategies typically employ a constant current constant voltage (CC-CV) mode, without actively perturbing system parameters. However, to determine the impedance characteristics and response sensitivity of the battery cells during charging, this step innovatively introduces a perturbation excitation signal into the charging current through a BMS control strategy, inducing a response in the battery cells and thereby identifying their dynamic voltage variations. This perturbation signal typically utilizes a low-amplitude sinusoidal or step signal with an amplitude controlled within ±0.2C and a frequency set between 0.1Hz and 5Hz, covering the low- and mid-frequency ranges of the electrochemical impedance spectroscopy. In the experiment, a sinusoidal signal with a frequency of 1Hz and an amplitude of ±200mA was superimposed on a standard 1C charging current. The application duration was controlled within 5 seconds to ensure that the perturbation did not affect the overall SOC by more than 0.2%. Using a BMS embedded controller (such as the STM32F4 series) to sample the battery cell voltage at a 1ms frequency during charging, the transient response changes in the voltage before and after the excitation signal injection were clearly captured.

[0024] The voltage response signal is uploaded to the main control platform via the CAN bus. The system automatically identifies key characteristics of each cell, including instantaneous voltage fluctuations, response delays, and amplitude variations. By superimposing multiple perturbations and statistically averaging the responses, the signal-to-noise ratio can be effectively improved, reducing random errors caused by ambient temperature and current fluctuations. This technique provides a key experimental foundation for subsequent impedance modeling and state determination. After collecting the charging voltage response signals of the cells under perturbation current, these signals need to be analyzed at multiple time points to extract their stable response characteristics. To achieve this, this step uses time-domain curve fitting and signal steady-state response modeling to process the voltage response curve of each cell and reconstruct the steady-state voltage response waveform of the cell under perturbation. Specifically, the corresponding voltage sequence is collected within the duration of each perturbation signal (e.g., 5 seconds) and divided into three time periods: rising edge, steady-state segment, and falling edge. In the experiment, the voltage variation range is approximately ±8mV, and the typical response time is around 200ms. During the fitting process, the voltage response curve was fitted using the least squares method. A combined model of a Gaussian function and a first-order exponentially decreasing function was selected to effectively simulate the inertial delay and steady-state stability of the voltage response. Furthermore, a sliding window median filter and a Savitzky-Golay smoothing filter were used to address noise interference, preserving the key inflection points of the response waveform. Through repeated measurements and averaging, a stable and reproducible voltage response waveform was obtained for each cell. This waveform was then normalized based on frequency and amplitude normalization. This steady-state voltage waveform served as input for subsequent time-frequency analysis, laying the foundation for a deeper understanding of the cell's response mechanism to different perturbation frequencies. To reveal the cell's responsiveness to perturbation current signals of varying frequencies, the steady-state voltage waveform was mapped from the time domain to the time-frequency domain to obtain the frequency response characteristics. The continuous wavelet transform (CWT) method was used to perform a multi-scale time-frequency decomposition of the voltage response waveform to extract the impedance characteristics of the cell at different frequency scales.

[0025] CWT uses the Morlet mother wavelet, which has excellent local time-frequency resolution. In the experiment, the scale was divided into 20 levels, corresponding to a frequency distribution ranging from 0.1 Hz to 5 Hz. For each scale, the phase delay and amplitude ratio between the voltage signal and the input perturbation current signal were analyzed to construct a complex impedance spectrum (Z(f) = V(f) / I(f)). After synchronous sampling and Fourier transform of the voltage and current signals, amplitude-frequency response and phase-frequency response curves were obtained. Experimental data showed that in the low-frequency region (<1 Hz), the impedance is dominated by polarization resistance, exhibiting a large phase lag (approximately 60° or more). In the high-frequency region (3-5 Hz), the impedance amplitude tends to be stable, mainly influenced by ohmic resistance, with a smaller phase difference. By comparing the impedance spectra of multiple cells at different frequencies, the frequency sensitivity characteristics, internal diffusion dynamics, and response delay characteristics of the active materials in individual cells can be identified. Ultimately, a multi-frequency voltage-response impedance spectrum encompassing 20 frequency points was constructed, providing detailed data support for subsequent multi-frequency trend analysis and state map construction. To further understand the dynamic variations in cell impedance across different operating stages, the multi-frequency impedance spectrum obtained in the previous step required time series modeling and trend evolution analysis. Specifically, the multi-frequency impedance spectra collected at different time periods were stacked into a three-dimensional time-frequency-impedance matrix, and the time-varying impedance trend at each frequency point was modeled. A combination of weighted moving average (WMA) and locally weighted regression (LOWESS) was used to extract a smoothed impedance trend at each frequency point. In the experimental setup, the impedance spectrum was sampled every 10 seconds for a total of 300 seconds, resulting in 30 time series segments for each frequency point. In the low-frequency region (0.2 Hz), the impedance slowly rises as the SOC increases, reflecting an increase in polarization impedance. In the high-frequency region (>3 Hz), the impedance remains largely stable, indicating a relatively constant internal ohmic resistance.

[0026] By quantitatively analyzing the slope and amplitude of trend curves, the "Frequency Sensitivity Index" (FSI) and "Impedance Variation Index" (IZI) were defined to characterize the behavioral activity and health status of a cell at a specific frequency. Experimental results show that when the FSI exceeds 0.5 at a frequency of 0.2Hz, the cell is more likely to have internal issues such as severe polarization and structural diffusion barriers. This trend characteristic can serve as a core indicator for cell anomaly warning, status screening, and triggering balancing strategies, effectively identifying cells with gradually deteriorating or mismatched responsiveness during charging. After obtaining the cell impedance trend characteristics and the cell behavior characteristics time series, these two types of information are fused to construct a cell state map reflecting the entire charging process, providing a comprehensive understanding of the entire battery pack status. This map uses cells as nodes and integrates their time series behavioral characteristics (such as SOC change rate, internal resistance dynamics, and temperature rise response) with frequency domain impedance characteristics (such as impedance curves at different frequencies, change trends, and frequency response delay) to construct a multidimensional state space. In the specific implementation, an autoencoder is used to reduce the dimensionality of high-dimensional state features, compressing them into three-dimensional feature vectors (representing charging dynamics, impedance stability, and thermal response activity). The distribution of multiple battery cells in the state space is then visualized using t-SNE visualization technology. In the graph, nodes with close positions indicate similar battery cell characteristics, while nodes with prominent distances indicate state deviation or degradation trends. In addition, by superimposing the time dimension evolution trajectory on the state graph, a "battery cell state trajectory diagram" can be formed, which intuitively shows the state migration process of a battery cell at different charging stages. Experiments have found that when the SOC of certain battery cells reaches above 80%, their state trajectory shifts rapidly, indicating an imbalance trend, requiring current bypass or compensation control through active balancing strategies.

[0027] In this embodiment, reference Figure 4 The above is a flowchart of the detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include: Extracting the global maximum voltage of the battery, the real-time charging voltage and the reverse value of the internal resistance of each battery cell according to the battery cell monitoring parameters; Calculating the voltage difference of each battery cell based on the global maximum voltage of the battery and the real-time charging voltage of each battery cell, and extracting the remaining voltage of each battery cell; Performing quantitative calculation of charging energy flow affinity based on the voltage residual and the internal resistance reverse value to generate the charging energy flow affinity of each battery cell; Based on the charging energy flow affinity and multiple battery cell charging state maps, an average battery cell state evaluation is performed, and a state deviation trend is distinguished, thereby dividing weak battery cells and strong battery cells.

[0028] In this embodiment, key parameters, including voltage, current, temperature, and internal resistance, are extracted from the cell monitoring data uploaded in real time by the BMS. Every 100ms, the system obtains the current voltage value of each cell in the entire battery pack from the cell sampling module and automatically identifies the cell with the highest global voltage in each acquisition round. This parameter is used to calibrate the optimal reference point for the battery pack's current state of charge. Furthermore, the reverse internal resistance value, an indicator for assessing the charging energy absorption potential of a cell, is typically derived by inverting the slope of the cell's voltage-current curve. A larger reverse internal resistance value indicates that the cell can withstand a greater charging power per unit voltage. In a specific experiment, a battery pack containing 96 18650 lithium-ion cells was charged and tested with a 100ms acquisition cycle. During a certain sampling period, the highest cell voltage was measured to be 4.187V, the lowest cell voltage was 4.108V, and the reverse internal resistance value ranged from 0.35 to 0.62. The data from this stage provides the foundation for subsequent voltage difference calculations and energy affinity quantification. To quantify the "voltage headroom" of each cell in its current charging state, the system compares each cell's real-time voltage with the global maximum voltage during the current detection cycle and calculates the voltage difference. This difference is defined as the "voltage reserve," which reflects the cell's ability to continue accepting energy input under the current charging conditions. The greater the voltage reserve, the greater the cell's charging potential.

[0029] In the above experimental environment, if the current highest voltage cell is cell 5 (4.187V) and cell 37 has a voltage of 4.135V, then their remaining voltage is 0.052V. This calculation is performed individually throughout the entire battery pack to form a set of cell voltage remaining vectors. This step is critical for assessing the degree of charging imbalance between cells and predicting potential deviations, and serves as essential foundational data for balancing control strategies. After obtaining the remaining voltage and reverse internal resistance values, the system performs a composite quantitative analysis to construct a cell "charging energy flow affinity" metric. This metric combines the remaining voltage with the cell's ability to respond to current changes (i.e., reverse internal resistance) to reflect each cell's actual affinity for absorbing current and completing charge transfer under current conditions. To enhance recognition accuracy, the system fuses these two sets of parameters using a normalized vector superposition method to ensure that the metric is not biased by numerical differences. Experimental operations revealed that cells with high residual voltage but high internal resistance exhibit moderate affinity, while cells with low residual voltage but extremely low internal resistance may also exhibit high affinity. Therefore, this metric more comprehensively reflects actual charge absorption characteristics, providing a scientific basis for further cell state classification. Finally, the system integrates the cell's charging energy affinity metric with the previously constructed cell charge state map to assess the average state across cells. The average affinity of all cells in the current cycle is calculated as a group benchmark, and each cell's affinity is then compared to this average to identify cells that deviate from the state. If a cell's affinity remains consistently below the average by more than 10% for multiple consecutive cycles and exhibits behavioral characteristics such as low residual voltage and high internal resistance, it is labeled a "weak cell." Conversely, if a cell's affinity consistently exceeds the average by more than 15% and demonstrates a stable thermal response and positive SOC changes as shown in the state map, it is classified as a "strong cell." The results of this type of division will serve as the strategic input point for BMS balancing control: preferentially suppressing the charging rate of strong cells and extending the charging time of weak cells, thereby achieving energy balance of the entire battery pack.

[0030] In this embodiment, step S4 includes the following steps: Obtain a diagram of the internal structure of the battery; identify the cell connections on the diagram and extract the cell connection information; Perform physical topology association mining based on cell connection information to identify topological connection relationships between cells; Fit the spatial distribution of topological connection paths of the topological connection relationships between battery cells and construct a topological connection graph between battery cells; Conduct energy-complementary cell mining on weak and strong cells and extract the energy-complementary cell matrix; Based on the topological connection diagram between cells, the energy offset complementary path of the energy complementary cell matrix is ​​mined to extract the nearest energy offset priority complementary path; Dynamic offset balancing planning is performed on the energy offset priority complementary path to construct an intelligent offset path planning strategy.

[0031] In this embodiment, the lithium-ion battery pack is a multi-cell integrated system. The connection method between its internal cells (series, parallel, or hybrid) directly affects the energy transmission path and balancing control strategy. To achieve refined control based on the physical connection structure, the battery's internal structure diagram must first be obtained and analyzed. This step uses battery module structural drawings or point cloud data obtained from production data or 3D CT imaging, and extracts cell connection information through image recognition and structural analysis methods. In practice, a 3D structural scanner is used to capture the internal structure of the battery module, and the spatial layout and tab connection method of each cell in the diagram are analyzed. Edge detection algorithms and connected component analysis are used to extract the spatial coordinates of each cell, and their electrical connection status is determined by combining color identification and polarity characteristics. In a 96S battery pack, the positions and polarity identification of all 96 cells are identified using automatic image segmentation methods. After obtaining the cell connection information matrix, topological mining and structural association identification of the physical connection relationships between the cells are further performed. This step uses graph theory modeling to transform the cells and their connections into a topological graph, where nodes represent cells and edges represent electrical connections. Specifically, an undirected graph G = (V, E) is constructed, where V is the set of cells and E is the set of edges connecting them. Edge weights can be assigned based on physical distance or resistance. In the experiment, Euclidean distance was used to measure the relative position of cells in physical space. Edge weights were then assigned based on measured connection resistance values ​​(ranging from approximately 0.5 to 2 mΩ) to reflect the actual conductivity and energy transfer cost. Subsequently, graph traversal and adjacency matrix analysis were used to identify key topological relationships, such as series path chains, parallel distribution clusters, and branch connection areas. Cluster analysis was then used to delineate the topological subgraph regions. In a PHEV battery pack, 12 groups of series cells were identified, each containing eight parallel cells connected in parallel via a busbar. This topological relationship not only reveals the regularity of the physical connections between cells but also provides a graph structure for subsequent power scheduling and path planning. The mining results are standardized and stored as a topological vector graph, forming the intermediate layer for modeling cell network relationships and enabling cross-layer data conversion from structural identification to control decision-making. After identifying topological connectivity relationships, the connection paths between cells are spatially fitted and structurally mapped to construct a cell topological connection graph in three-dimensional space. This graph intuitively represents the physical proximity and conduction paths between cells, providing a basic spatial model for energy flow simulation and balanced path optimization. The fitting method utilizes multidimensional scaling (MDS) and a force-directed graph drawing algorithm.MDS is used to project high-dimensional cell connection relationships into three-dimensional space, preserving the relative distances between nodes; while the force-directed algorithm simulates the "spring-repulsion" behavior between nodes, optimizing the graph layout and avoiding node overlap. In modeling, the connection strength (edge ​​weight) is used as the attraction parameter, and the spatial distance is used as the repulsion parameter to achieve a natural distribution fit of the graph. In a certain battery pack structure, after generating a topological diagram, it was found that the cells were connected in a "hierarchical mesh" manner, with parallel connections on the periphery and series connections in the middle, forming a typical fan-shaped conduction structure. The diagram is standardized into a set of connection maps in a three-dimensional coordinate system, which can be visualized and displayed in conjunction with real-time BMS status data.

[0032] The final output is a structured cell space topology diagram, including node coordinates, edge lengths, connection weights and other information. This diagram not only supports subsequent energy flow path modeling, but can also be used for engineering applications such as thermal management simulation and electrical fault path prediction. It is the core data foundation for realizing space-aware balancing control strategies. In the previous steps, the identification and classification of weak cells and strong cells have been completed, and a cell topology connection diagram has been constructed. Based on this, the goal of this step is to discover cell pairs with potential energy complementarity in the topological structure, and construct an "energy complementary cell matrix" as the input for offset balancing path planning. The specific method is to perform energy complementation on each weak cell. , and all powerful batteries The complementarity index is defined by the topological path length (the shortest path hop number or the minimum weight path) between the two, combined with the affinity difference between the two (the degree of matching between the energy overflow potential of the strong cell and the energy gap of the weak cell): A_j and A_i represent the cell affinity, D_ij represents the topological path weight distance, is the adjustment coefficient, and ϵ is a small constant to prevent division by zero. Based on the C value, cell pairs with strong complementarity are selected to construct an energy complementarity matrix, C∈R^(m×n), where m is the number of weak cells and n is the number of strong cells. In experiments, 5 weak cells and 8 strong cells were identified in a 32-cell system, ultimately resulting in 40 valid complementary pairs. The average path length was 2.5 hops, and the average voltage difference exceeded 40mV, demonstrating significant complementary adjustment potential. After completing the energy complementarity matrix, this step focuses on extracting "energy offset priority paths" from the cell topological connection diagram. These paths are the shortest paths required to transfer energy from strong cells to weak cells under actual physical connection conditions. The goal is to identify candidate paths with the lowest transmission loss and minimal interference, thereby optimizing energy flow at the path level. Path extraction is based on a graph traversal algorithm, using Dijkstra's shortest path method or A* heuristic search. Starting from each strong cell node, the path with the minimum edge weight is found to the weak cell. The weight is a weighted combination of factors such as physical distance, connection impedance, and control response delay. In the experiment, the edge weight is set as: ; Where R_ij is the resistance, L_ij is the path length, τ_ij is the control delay, and the coefficient is set to Through algorithmic execution, an optimal energy offset path is generated for each pair of cells, and these paths are aggregated into a "priority complementary path set." In one configuration, a path from cell #2 (strong) to cell #27 (weak) requires only two hops, transmitted via relay cell #15, with a total path impedance of only 1.2 mΩ, demonstrating significant transmission potential. These paths will serve as candidate control routes for the final balancing plan, ensuring optimal energy compensation in physical space using the shortest possible links. This matrix will be used in subsequent energy offset path analysis and serves as a fundamental data structure for planning balanced energy transfer directions and path strengths. After all priority complementary paths are determined, a dynamic offset control strategy must be developed to automatically adjust the energy allocation for each path based on the system's real-time status, achieving intelligent, low-loss offset balancing. This step constructs the "offset path planning strategy," the core of which is to dynamically allocate energy transfer tasks based on path load capacity, cell state evolution trends, and scheduling frequency limits. This strategy utilizes a multi-objective optimization approach, with objective functions including maximizing energy compensation rate, minimizing path load, and accelerating cell aging. During the scheduling process, the BMS monitors parameters such as cell voltage, internal resistance, and temperature in real time. It re-evaluates the dispatchability of all current paths every sampling period (e.g., 10 seconds) and dynamically activates the optimal sub-path. For the cell pairs with the largest affinity difference and the smallest path weight, the bypass discharge (for strong cells) and weak charge (for weak cells) control actions of the current cycle are prioritized, and the scheduling ratio is limited by the path transmission capacity. In the experiment, after the strategy was implemented, the weak cells with an original voltage difference of 50mV converged to the average level within the group to ±5mV within 20 minutes, and the balancing efficiency was improved by nearly 40% compared to the traditional fixed balancing method. The planning strategy is highly adaptable and can automatically adjust the control strategy under different discharge rates, ambient temperatures, or cell aging conditions. It is a key link in achieving intelligent balancing control and improving the consistency and life of battery packs.

[0033] In this embodiment, the specific steps of step S5 are: Calculate the charging pulse frequency according to the battery cell monitoring parameters and extract the pulse frequency timestamp; Performing a sliding window regression analysis on the steady-state voltage response waveform according to the pulse frequency timestamp to extract the charging response hysteresis parameter of each battery cell; Identify the differentiated hysteresis characteristics of each battery cell for charging response hysteresis parameters, perform peak-shifting response timing analysis, and generate a global battery cell peak-shifting response sequence; Perform asynchronous hysteresis prediction and drive regulation on the global battery cell peak-shifting response sequence to build an asynchronous charging balancing regulation strategy; Real-time dynamic charging balancing control is performed based on the asynchronous charging balancing adjustment strategy and the intelligent offset path planning strategy to execute intelligent balancing charging control operations.

[0034] In this embodiment, during actual BMS operation, to more accurately analyze the response characteristics of each battery cell during charging, the system embeds slightly perturbed charging pulse signals during the standard constant-current and constant-voltage charging phase. These pulses are superimposed on the base charging current in the form of small amplitude steps, and the voltage response is acquired at a high sampling rate. The system detects the instantaneous trigger points of the current change rate (ΔI) and the voltage change rate (ΔV) to identify the location of each valid pulse and accurately extract the timestamp information for each pulse. In experimental testing, the system applied a low-amplitude (approximately 100mA) charging pulse at a frequency of 0.5Hz every 10 seconds. 96 battery cells were continuously monitored for 15 minutes, extracting 900 pulse timestamps and marking them on the charging voltage response curve of each battery cell. These timestamps provide a precise reference point for subsequent identification of voltage hysteresis. After obtaining the pulse frequency timestamps, the system sets a sliding time window of fixed width (e.g., ±0.5 seconds) around each pulse time point. It then performs a local regression fit on the voltage curve within this time period, aiming to extract the delay time of the voltage response onset and the lag time of the maximum response amplitude. This method uses linear or nonlinear regression (such as Loess or Savitzky-Golay regression) to fit the curve trend and identify the actual response start and peak delay based on residual analysis.

[0035] The system established a database of charging response hysteresis parameters for 96 battery cells. The average hysteresis time for each cell under different pulses ranged from 40ms to 110ms. This inconsistency in hysteresis time reveals subtle differences in charge transfer, interface reactions, and internal impedance among different cells, and is an important indicator of cell consistency and health. After extracting the hysteresis parameters for each cell, the system then conducts a horizontal differentiation analysis. Specifically, the hysteresis time series of all cells are clustered (for example, using DBSCAN or K-Means), grouping cells with similar response hysteresis characteristics. This allows identification of groups of cells with higher hysteresis and slower response times, as well as clusters of cells with faster response times. Based on these classifications, the system ranks all cells by their time response capabilities, generating a set of "off-peak response time series" from fastest to slowest. In the experimental model, hysteresis characteristics are categorized into five levels, from level 1 (fastest response) to level 5 (slowest response). By staggering the sequence timeline, the system generates a global time schedule for cell charging responses. This ensures that the response rhythms of different cells during subsequent charging control processes can be logically staggered, reducing the risk of concentrated current fluctuations and improving the refinement of balanced scheduling. Based on the staggered response timing sequence, the system further designs an "asynchronous charging balancing adjustment strategy" to achieve compensatory delay control for lagging cells and incentive control for fast-responding cells by nonlinearly regulating the charging drive start times of different cells. This strategy primarily relies on a hysteresis trend prediction model (such as an LSTM or sliding regression predictor) to determine the hysteresis change trend of the cell at the next moment and dynamically allocates the adjustment time window accordingly.

[0036] In actual testing, for cells with a response lag exceeding 100ms, the system delayed the charge trigger window by 50ms to minimize impact. For cells with a response faster than 50ms, a low-amplitude pre-charge was performed in advance. The regulation strategy sends control commands to the BMS main control chip via the serial port, adjusting the PWM duty cycle and current channel switching logic in real time. This enables asynchronous regulation of each cell and establishes a responsive balancing mechanism that dynamically adapts to changes in cell status. This asynchronous hysteresis regulation strategy is integrated with the previously established energy shift path planning strategy to form a complete real-time balancing control system. In this system architecture, the asynchronous strategy regulates the response timing and rhythm of each cell, while the shift path strategy schedules the spatial energy transfer path between cells, achieving coordinated control in both temporal and spatial dimensions. During testing, the system calculated the current charge state, hysteresis trend, and energy demand of each cell in real time, using a minimum time regulation unit of 1 second. Based on the regulation strategy, it automatically selected asynchronous delay, advance, or skip processing, while also controlling the direction and rate of energy shift. During continuous charging cycles, the control system effectively suppressed the SOC deviation rate between battery cells, reducing the maximum deviation from the initial 8.7% to 2.1%, and significantly reduced the thermal load difference between battery cells, improving the stability and safety of the overall system.

[0037] In this embodiment, the specific steps of performing real-time dynamic charging balancing control based on the asynchronous charging balancing adjustment strategy and the intelligent offset path planning strategy to execute the intelligent balanced charging control operation are as follows: Real-time dynamic charge balancing control is performed based on asynchronous charge balancing adjustment strategy and intelligent offset path planning strategy, and the temperature detection value of each battery cell is identified based on thermistors; Performing a cell temperature difference analysis on the cell temperature detection values ​​and performing distribution fitting to obtain a cell temperature difference gradient; Identifying the internal resistance of the battery cells during the real-time dynamic charge balancing process; Performing a nonlinear correlation analysis on the internal resistance of the battery cell and the temperature difference gradient of the battery cell to obtain a nonlinear correlation law between temperature and internal resistance; Based on the temperature-internal resistance correlation law, local thermal excitation decisions are made and current bias path driving is performed, thereby constructing a dynamic local thermal excitation collaborative engine; Intelligent balanced charging control operations are performed based on a dynamic local thermal excitation collaborative engine.

[0038] In this embodiment, while implementing asynchronous regulation and intelligent path offset strategies, real-time acquisition of cell temperature status is a key component of the closed-loop control strategy. This step utilizes temperature sensing via NTC thermistors (NTCs) placed externally on each cell or near heat sources, establishing a real-time temperature sampling channel for each cell and ensuring the control strategy's timely response to thermal state changes. Specifically, each cell in the cell module is equipped with at least one 10kΩ negative temperature coefficient thermistor (NTC), with a sampling frequency set between 1Hz and 10Hz. The temperature is collected in parallel based on the number of ADC channels in the BMS main control chip. The sampled temperature values ​​are converted to actual temperature values ​​using a thermistor voltage divider calculation formula, with an error within ±0.5°C. Temperature data is collected synchronously with charging current, voltage, SOC, internal resistance, and other parameters, and aggregated in real time in the BMS main control logic module. This temperature acquisition system provides a "thermal feedback" channel for balanced regulation, particularly when executing high-frequency offset paths, where some paths may pose a risk of localized overheating. Temperature monitoring enables regulation rhythm adjustment and current amplitude limiting. When the temperature of cell #12 is detected to be rising at a rate exceeding 0.2°C / min, the BMS automatically delays its scheduling priority for the offset receiving path. This step closes the loop of the thermally aware balancing strategy, laying the foundation for accurate monitoring for subsequent temperature difference analysis and thermal excitation regulation. After obtaining the temperature measurement values ​​of all cells, the system first performs a unified standardization preprocessing on each set of sampled data to remove occasional data jitter and perform a sliding median filter (with a window size of 3) to eliminate temperature measurement errors caused by environmental disturbances. The temperature difference between each cell and the average temperature within its battery module is then calculated, constructing a 96-dimensional cell temperature difference vector. To comprehensively evaluate the spatial distribution structure of the cell temperatures, the system uses a combination of polynomial regression and kernel density estimation to fit the spatial distribution of the cell temperature difference data and output a cell temperature difference gradient map. In experiments, the map revealed that some cells (such as those located near the heat sink) consistently exceeded the average temperature by approximately 3-5°C, while others remained below the average by more than 2°C. This formed a clear high-low temperature gradient, providing a basis for temperature-driven energy regulation. While processing temperature data, the system leveraged embedded pulse perturbation technology to sample and calculate the response ratio of the voltage difference before and after the pulse current was applied, enabling online identification of the cell's dynamic internal resistance. To improve accuracy, the system selected multiple perturbation windows (one every 10 seconds) during each charging cycle to sample the voltage-current response values. The sliding average of these differences was used to determine the cell's real-time internal resistance. During the internal resistance identification process, the system paid particular attention to the internal resistance's variation with temperature, SOC, and current. In the test model, the internal resistance of different cells varied by 0.7mΩ to 2.1mΩ, clearly indicating that cells with high internal resistance exhibited weaker charging capabilities.At the same time, all internal resistance values ​​will be dynamically bound with temperature fluctuations, and will be updated synchronously with the temperature difference gradient map in the data structure, providing accurate support for the subsequent establishment of the temperature-internal resistance relationship model. In order to deeply understand the coupling relationship between the thermal characteristics of the battery cell and the electrochemical reaction efficiency, the system performs bivariate nonlinear modeling on the internal resistance of all battery cells and their corresponding temperature differences. In terms of specific implementation, the system uses methods such as random forest regression and support vector regression (SVR) to identify the changing patterns between temperature and internal resistance at different SOC stages and different charging rates. The model training data comes from actual measurement results under 5 sets of environmental temperature gradient conditions (10°C, 20°C, 30°C, 40°C, 50°C), with a sample size of more than 20,000 sets.

[0039] Modeling results show that battery cells exhibit their lowest internal resistance response at around 30°C. Below 20°C or above 40°C, internal resistance increases significantly, demonstrating a typical "U"-shaped nonlinear effect of temperature on internal resistance. Extracting this correlation allows the system to determine the charging potential and load capacity of a battery cell based on its current temperature during charging control, thereby facilitating more precise current routing and energy matching. Based on this identified nonlinear correlation between temperature and internal resistance, the system further develops a "dynamic local thermal excitation synergy engine." Its core mechanism is to provide targeted charging stimulation to cells with lower temperatures and higher internal resistance by intelligently adjusting the current path and charging intensity, while ensuring thermal safety. This helps rapidly raise their temperatures to the optimal internal resistance range (approximately 30°C), thereby improving overall charging efficiency. Based on a path planning strategy, the system appropriately reduces current flow to cells with high heat capacity (or higher temperatures) while increasing current input to cooler cells, achieving "thermal energy guidance" in the spatial dimension. In testing, this strategy successfully raised the temperature of five groups of cooler cells by approximately 2°C, correspondingly reducing their internal resistance by 0.6mΩ, enabling them to release more charging power and demonstrating excellent local thermal control capabilities. Ultimately, the system integrated a "dynamic local thermal excitation coordination engine" into the overall BMS balancing control strategy, forming a three-dimensional intelligent balancing control model with spatial energy regulation, timed asynchronous response, and thermal excitation guidance compensation. This model outputs the optimal current bias path in real time based on each cell's current SOC, temperature, internal resistance, and other parameters, achieving coordinated execution of multiple objectives such as local thermal regulation, response timing staggering, and energy flow balancing. During long-term operation (three hours of continuous charging), the system controlled the maximum inter-cell temperature deviation from 6.3°C to 2.8°C and the maximum internal resistance difference from 1.4mΩ to 0.6mΩ. It also improved the SOC convergence rate between cells by over 30%, effectively extending the effective period of the system's balancing charging window, providing a solid technical foundation for the safe and efficient operation of large-scale lithium battery packs in new energy vehicles.

[0040] In this embodiment, a BMS-based lithium-ion battery balanced charging control system is provided, which is used to execute the BMS-based lithium-ion battery balanced charging control method described above, including: The charging behavior analysis module is used to continuously obtain the battery cell monitoring parameters of the battery charging status based on the BMS, perform high-frequency filtering and noise removal, and analyze the battery cell charging behavior to obtain the battery cell behavior characteristic time series sequence; The state map module is used to identify the charging voltage response signal of each battery cell based on the BMS, perform multi-frequency impedance change trend evolution, and perceive the battery cell charging response characteristics based on the battery cell behavior characteristic time series sequence to construct multiple battery cell charging state maps; A cell status classification module is used to perform quantitative calculation of charging energy flow affinity based on the cell monitoring parameters, and distinguish state deviation trends based on the multiple cell charging state maps, thereby classifying weak cells and strong cells; The complementary path mining and planning module is used to construct a topological connection diagram between cells, and to mine complementary paths for the energy offset of weak and strong cells, perform dynamic offset balancing planning, and build an intelligent offset path planning strategy. The asynchronous balancing adjustment module is used to perform asynchronous hysteresis predictive drive adjustment according to the battery cell monitoring parameters, and perform real-time dynamic charging balancing control based on the intelligent offset path planning strategy to perform intelligent balanced charging control operations.

[0041] The present invention is therefore intended to be illustrative and non-restrictive in all respects, with the scope of the invention being defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the application documents are intended to be embraced therein.

[0042] The foregoing description is intended only to provide specific embodiments of the present invention, which are intended to enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest manner consistent with the principles and novel features disclosed herein.

Claims

1. A lithium-ion battery balanced charging control method based on BMS, characterized in that: The following steps are involved: Step S1: Based on the BMS, the battery cell monitoring parameters of the battery charging status are continuously acquired, and high-frequency filtering and noise removal processing and battery cell charging behavior analysis are performed to obtain a battery cell behavior characteristic time series sequence; Step S2: Based on the BMS, the charging voltage response signal of each battery cell is identified, the multi-frequency impedance change trend evolution is performed, and the battery cell charging response characteristics are perceived based on the battery cell behavior characteristic time series sequence, and multiple battery cell charging state maps are constructed; Step S3: performing quantitative calculation of charging energy flow affinity according to the cell monitoring parameters, and distinguishing state deviation trends based on the multiple cell charging state maps, thereby classifying weak cells and strong cells; Step S4: Construct a topological connection diagram between cells, and mine complementary paths for energy offsets of weak cells and strong cells, perform dynamic offset balancing planning, and build an intelligent offset path planning strategy; Step S5: performing asynchronous hysteresis predictive drive adjustment according to the cell monitoring parameters, and performing real-time dynamic charge balancing control based on an intelligent offset path planning strategy to execute an intelligent balanced charge control operation.

2. The BMS-based lithium-ion battery balanced charging control method according to claim 1, characterized in that: The specific steps of step S1 are: Based on BMS, millisecond-level micro-time sampling is performed to continuously obtain battery cell monitoring parameters of the battery charging status; Perform high-frequency filtering and denoising on the cell monitoring parameters to obtain filtered and denoised monitoring parameters; Calculate the real-time SOC parameters, internal resistance state, thermal parameters, charge and discharge efficiency, and voltage of each battery cell based on the filtered and denoised monitoring parameters to obtain the multi-dimensional state characteristics of the battery cell; Divide the multi-dimensional state characteristics of the battery cell into time series windows to obtain multi-dimensional state characteristics in different time periods; The charging behavior of the battery cell is analyzed based on the multi-dimensional state characteristics of different time periods. The internal resistance perturbation changes, micro-voltage fluctuations and transient thermal responses are analyzed to obtain the time series of the battery cell behavior characteristics.

3. The BMS-based lithium-ion battery balanced charging control method according to claim 1, characterized in that: The specific steps of step S2 are: Based on the BMS, a perturbation excitation signal is introduced into the charging current, and the charging voltage response signal of each battery cell is identified; Perform multi-time point response state fitting on the charging voltage response signal to construct a steady-state voltage response waveform; Perform multi-scale time-frequency decomposition on the steady-state voltage response waveform and extract the voltage response impedance spectra at different frequencies; Performing multi-frequency impedance change trend evolution on the voltage response impedance spectrum to obtain impedance change trend characteristics at different frequencies; Based on the impedance change trend characteristics and the cell behavior characteristic time series, the cell charging response characteristics are perceived and multiple cell charging state maps are constructed.

4. The BMS-based lithium-ion battery balanced charging control method according to claim 1, characterized in that: The specific steps of step S3 are: Extracting the global maximum voltage of the battery, the real-time charging voltage and the reverse value of the internal resistance of each battery cell according to the battery cell monitoring parameters; Calculating the voltage difference of each battery cell based on the global maximum voltage of the battery and the real-time charging voltage of each battery cell, and extracting the remaining voltage of each battery cell; Performing quantitative calculation of charging energy flow affinity based on the voltage residual and the internal resistance reverse value to generate the charging energy flow affinity of each battery cell; Based on the charging energy flow affinity and multiple battery cell charging state maps, an average battery cell state evaluation is performed, and a state deviation trend is distinguished, thereby dividing weak battery cells and strong battery cells.

5. The BMS-based lithium-ion battery balanced charging control method according to claim 1, characterized in that: The specific steps of step S4 are: Obtain a diagram of the internal structure of the battery; identify the cell connections on the diagram and extract the cell connection information; Perform physical topology association mining based on cell connection information to identify topological connection relationships between cells; Fit the spatial distribution of topological connection paths of the topological connection relationships between battery cells and construct a topological connection graph between battery cells; Conduct energy-complementary cell mining on weak and strong cells and extract the energy-complementary cell matrix; Based on the topological connection diagram between cells, the energy offset complementary path of the energy complementary cell matrix is ​​mined to extract the nearest energy offset priority complementary path; Dynamic offset balancing planning is performed on the energy offset priority complementary path to construct an intelligent offset path planning strategy.

6. The BMS-based lithium-ion battery balanced charging control method according to claim 1, characterized in that: The specific steps of step S5 are: Calculate the charging pulse frequency according to the battery cell monitoring parameters and extract the pulse frequency timestamp; Performing a sliding window regression analysis on the steady-state voltage response waveform according to the pulse frequency timestamp to extract the charging response hysteresis parameter of each battery cell; Identify the differentiated hysteresis characteristics of each battery cell for charging response hysteresis parameters, perform peak-shifting response timing analysis, and generate a global battery cell peak-shifting response sequence; Perform asynchronous hysteresis prediction and drive regulation on the global battery cell peak-shifting response sequence to build an asynchronous charging balancing regulation strategy; Real-time dynamic charging balancing control is performed based on the asynchronous charging balancing adjustment strategy and the intelligent offset path planning strategy to execute intelligent balancing charging control operations.

7. The BMS-based lithium-ion battery balanced charging control method according to claim 6, characterized in that: The specific steps of performing real-time dynamic charging balancing control based on the asynchronous charging balancing adjustment strategy and the intelligent offset path planning strategy to execute the intelligent balancing charging control operation are as follows: Real-time dynamic charge balancing control is performed based on asynchronous charge balancing adjustment strategy and intelligent offset path planning strategy, and the temperature detection value of each battery cell is identified based on thermistors; Performing a cell temperature difference analysis on the cell temperature detection values ​​and performing distribution fitting to obtain a cell temperature difference gradient; Identifying the internal resistance of the battery cells during the real-time dynamic charge balancing process; Performing a nonlinear correlation analysis on the internal resistance of the battery cell and the temperature difference gradient of the battery cell to obtain a nonlinear correlation law between temperature and internal resistance; Based on the temperature-internal resistance correlation law, local thermal excitation decisions are made and current bias path driving is performed, thereby constructing a dynamic local thermal excitation collaborative engine; Intelligent balanced charging control operations are performed based on a dynamic local thermal excitation collaborative engine.

8. A lithium-ion battery balanced charging control system based on BMS, characterized in that: The method for controlling the balanced charging of a lithium-ion battery based on a BMS according to claim 1 comprises: The charging behavior analysis module is used to continuously obtain the battery cell monitoring parameters of the battery charging status based on the BMS, perform high-frequency filtering and noise removal, and analyze the battery cell charging behavior to obtain the battery cell behavior characteristic time series sequence; The state map module is used to identify the charging voltage response signal of each battery cell based on the BMS, perform multi-frequency impedance change trend evolution, and perceive the battery cell charging response characteristics based on the battery cell behavior characteristic time series sequence to construct multiple battery cell charging state maps; A cell status classification module is used to perform quantitative calculation of charging energy flow affinity based on the cell monitoring parameters, and distinguish state deviation trends based on the multiple cell charging state maps, thereby classifying weak cells and strong cells; The complementary path mining and planning module is used to construct a topological connection diagram between cells, and to mine complementary paths for the energy offset of weak and strong cells, perform dynamic offset balancing planning, and build an intelligent offset path planning strategy. The asynchronous balancing adjustment module is used to perform asynchronous hysteresis predictive drive adjustment according to the battery cell monitoring parameters, and perform real-time dynamic charging balancing control based on the intelligent offset path planning strategy to perform intelligent balanced charging control operations.

Images