Storage battery visual charging and discharging intelligent management method and system

Abstract

The invention provides a storage battery visual charging and discharging intelligent management method and system. The method comprises the following steps: constructing a battery pack state set of a storage battery pack; analyzing a trend vector and fluctuation intensity of each battery cell parameter based on the state vectors of the plurality of battery cells to calculate an initial risk score of the battery cells, and calculating a comprehensive risk score of each battery cell in combination with a space thermal coupling relationship among the battery cells; in combination with the battery pack state set and the comprehensive risk score, performing simulation prediction on a plurality of preset charging and discharging strategies, designing a target function based on temperature suppression, voltage balance and control complexity for evaluation, and selecting an optimal charging and discharging strategy; and generating a control instruction according to the optimal charging and discharging strategy, generating visual information of battery pack state evolution based on a simulation prediction result of the charging and discharging strategy, and issuing the control instruction to a battery management system for execution. The service life of the battery can be effectively prolonged, the maintenance cost is reduced, and the scientificity of operation decision is enhanced.

Description

Technical Field

[0001] This invention belongs to the field of intelligent management of battery charging and discharging visualization, and particularly relates to a method and system for intelligent management of battery charging and discharging visualization. Background Technology

[0002] With the continuous development of energy storage technology, smart grids, and new energy vehicle industries, battery packs have become key energy storage units in energy systems. Their operational safety and management efficiency directly affect the overall performance and lifespan of the system. Under complex charging and discharging conditions, there are significant individual differences and thermal and electrical coupling effects among battery cells. Parameters such as voltage, current, temperature, and internal resistance exhibit nonlinear changes in time and space. Traditional battery management systems often use fixed threshold judgments or static models based on single-point data for state estimation. While these methods can provide rough estimates of state of charge and health, they exhibit significant lag when dealing with rapid fluctuations in cell state, local imbalances, and abnormal propagation trends, failing to effectively support refined real-time control. Furthermore, existing management methods generally lack time-series modeling capabilities and dynamic correlation analysis mechanisms between multi-dimensional parameters, making it difficult for the system to identify early signs of degradation. Some existing systems attempt to improve risk identification through fault detection or prediction models, but these often remain at the level of single-dimensional parameter analysis or independent model output, failing to achieve structured fusion of multiple parameters. Furthermore, current systems primarily display results as static curves or numerical reports, lacking dynamic visualization methods for the spatial distribution and state evolution trends of battery packs. This makes it difficult for managers to perceive potential risks from a holistic perspective. In terms of control, traditional systems often employ preset charging and discharging strategies, unable to adaptively adjust charging and discharging parameters based on the current risk state, nor can they predict potential thermal and electrical effects before strategy execution. These issues result in bottlenecks in safety protection, energy utilization, and lifespan extension. Particularly in large-scale cell packs and complex load environments, lagging management strategies, rigid control execution, and insufficient interpretability have become key obstacles restricting the development of smart energy storage systems. Summary of the Invention

[0003] The purpose of this invention is to propose a visualized intelligent management method and system for charging and discharging batteries to solve the above-mentioned problems.

[0004] To achieve the above objectives, a first aspect of the present invention provides a method for intelligent management of battery charging and discharging with visualization, the method comprising the following steps: S1. Construct a battery pack state set for the battery pack; wherein, the battery pack state set includes state vectors of several cells, and the state vectors of the cells include cell voltage, current, temperature and internal resistance; S2. Based on the state vector analysis of the several battery cells, the trend vector and fluctuation intensity of each battery cell parameter are analyzed to calculate the initial risk score of the battery cell, and combined with the spatial thermal coupling relationship between the battery cells, the comprehensive risk score of each battery cell is calculated. S3. Combining the battery pack state set and comprehensive risk score, simulate and predict multiple preset charging and discharging strategies, and design an objective function based on temperature suppression, voltage equalization and control complexity for evaluation, and select the optimal charging and discharging strategy. S4. Generate control commands based on the optimal charging and discharging strategy, and generate visualization information on the state evolution of the battery pack based on the simulation prediction results of the charging and discharging strategy. At the same time, send the control commands to the battery management system for execution.

[0005] Furthermore, the voltage is acquired by a voltage sampling chip inside the BMS; the current is measured by a Hall sensor or shunt resistor connected in series in the cell branch; the temperature is acquired by an NTC thermistor in close contact with the cell surface; and the internal resistance is calculated by measuring the voltage change by applying a short-time load pulse.

[0006] Furthermore, before constructing the battery pack state set, the cell voltage, current, temperature, and internal resistance are normalized within a time window.

[0007] Furthermore, the trend vector is a slope index of temperature, internal resistance, and voltage; the fluctuation intensity is the variance of the regression residual of the corresponding cell's state vector within that window.

[0008] Furthermore, the comprehensive risk score for each cell is calculated based on the spatial thermal coupling relationship between the cells, specifically as follows: Assuming the battery pack structure is a nearest neighbor graph, and defining the second cell as a thermal neighbor of the first cell; wherein, the nearest neighbor graph includes the second cell; Based on the second cell, a neighbor propagation term based on thermal coupling weight is constructed. Combined with the initial risk score of the first cell, the comprehensive risk score of the first cell is obtained.

[0009] Furthermore, the preset multiple charging and discharging strategies correspond to multiple sets of charging and discharging control operations; The simulation prediction of multiple preset charging and discharging strategies specifically includes: The state vector of the battery cell is input into a set of independently trained lightweight one-dimensional convolutional temporal networks, which output the predicted battery cell temperature and predicted voltage trajectory for the next H steps. The lightweight one-dimensional convolutional temporal network consists of 3 convolutional layers, each with 16 channels, a kernel size of 3, an activation function of ReLU, and an output layer consisting of a multi-channel regression unit per battery cell.

[0010] Furthermore, the temperature suppression is based on the current charge / discharge strategy under the first... The predicted cell temperature for each cell in the next H steps is calculated; the voltage equalization is calculated based on the predicted voltage trajectory for the next H steps; and the control complexity is calculated based on the operands generated in the cell control channel by the current charging and discharging strategy.

[0011] Furthermore, the simulation prediction results based on the charge-discharge strategy include the predicted cell temperature, predicted voltage trajectory, and predicted current for each cell.

[0012] Furthermore, the control commands include cell number, current limit value, equalization flag, and cooling control.

[0013] A second aspect of the present invention provides a battery visualization charging and discharging intelligent management system, the system comprising: The data acquisition and processing module is used to construct a battery pack state set for the battery pack; wherein, the battery pack state set includes state vectors of several cells, and the state vectors of the cells include cell voltage, current, temperature and internal resistance; The risk analysis module is used to analyze the trend vector and fluctuation intensity of the parameters of each battery cell based on the state vector of the several battery cells, so as to calculate the initial risk score of the battery cell, and calculate the comprehensive risk score of each battery cell by combining the spatial thermal coupling relationship between the battery cells. The strategy simulation and decision-making module is used to combine the battery pack state set and comprehensive risk score to simulate and predict multiple preset charging and discharging strategies, and to design an objective function based on temperature suppression, voltage equalization and control complexity for evaluation, and select the optimal charging and discharging strategy. The visualization and control execution module is used to generate control commands based on the optimal charging and discharging strategy, generate visualization information of the battery pack state evolution based on the simulation prediction results of the charging and discharging strategy, and send the control commands to the battery management system for execution.

[0014] The beneficial technical effects of the present invention are at least as follows: This invention's system utilizes a structured multi-dimensional time-series data modeling mechanism to construct a cell-level dynamic state tensor using four core parameters: voltage, current, temperature, and internal resistance. This enables continuous characterization and computationalization of the charging and discharging process, providing a stable data foundation for subsequent analysis. Through trend identification and risk scoring models, the system comprehensively analyzes the slope of cell changes, fluctuation intensity, and spatial thermal coupling characteristics in the time series, proposing a risk regularization assessment method to achieve early detection of local anomalies and potential faults. In the strategy generation stage, based on the state sequence and risk score, the system employs a lightweight predictive simulation model to evaluate the future response effects of different charging and discharging strategies. It also integrates temperature suppression, voltage equalization, and control complexity regularization terms into the loss function, thereby achieving optimal strategy selection through multi-objective optimization. In the execution stage, the system maps the optimal strategy to a standardized control instruction set, directly applying it to execution modules such as current limiting, equalization, and cooling. A three-dimensional dynamic visualization interface displays the state evolution of the battery pack under strategy execution, allowing managers to observe the consistency between predicted trends and control results in real time. This method is the first to realize a fully intelligent closed loop from data acquisition, risk identification, strategy prediction to visualized execution. While ensuring safety, it improves the transparency and adjustability of the system, effectively extends battery life, reduces maintenance costs, and enhances the scientific nature of operational decisions. Attached Figure Description

[0015] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0016] Figure 1 This is a flowchart of the intelligent management method for visualized charging and discharging of batteries according to the present invention. Detailed Implementation

[0017] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0018] like Figure 1 As shown in the embodiment of the present invention, the intelligent management method for visualized charging and discharging of batteries includes: S1. Construct a battery pack state set for the battery pack; wherein, the battery pack state set includes state vectors of several cells, and the state vectors of the cells include cell voltage, current, temperature and internal resistance.

[0019] Specifically, during the charging and discharging process, the operating state of a battery system is reflected not only in the static values ​​such as voltage and current at the current moment, but also in the dynamic evolution of these parameters over time. Therefore, to achieve management of the battery's true state, this step proposes a complete operating state data construction scheme to represent the continuous state evolution process of each cell over a time period. This scheme not only collects current values ​​but also constructs a unified data structure through a sliding time window mechanism, covering the historical states of all cells, thereby providing stable, coherent, and structured input tensors for subsequent anomaly trend identification and charging / discharging strategy prediction. The goal of this design is to achieve reproducible cell-level time-series state modeling with the fewest possible measurement channels.

[0020] Furthermore, in the actual data collection process, the key parameters of each battery cell include four categories: voltage, current, temperature, and internal resistance. Voltage Voltage is acquired by the internal voltage sampling chip of the BMS, and the sampling channel is typically a 16-bit ADC; current... Temperature is measured by a Hall sensor or shunt resistor connected in series in the cell branch. Data is collected by an NTC thermistor closely attached to the surface of the battery cell; internal resistance. The voltage change is calculated by applying a short-term load pulse. All data sampling frequencies are fixed and driven by the MCU timing. The sampling results are timestamped and then transmitted to the host computer or control module for further processing. Taking a 48-cell battery pack as an example, 48 voltage values, 48 ​​current values, 24 temperature values ​​(each pair of cells shares a thermistor), and 48 internal resistance values ​​can be obtained per second. All data is uploaded via the CAN bus.

[0021] At every moment, for the first Individual cell state vectors: ; in , , , These represent the cell voltage, current, temperature, and internal resistance, respectively, all derived from corresponding physical sensors or computing modules. The state vectors of all cells combine at the same time to form the battery pack state. ; in The total number of battery cells is determined by their physical connection sequence, ensuring a constant spatial position of different cell states within the data structure. To capture trends during charging and discharging, past data is retained in the time dimension. Construct a time window based on the state at each moment: ; In the implementation, this tensor is a three-dimensional array (time × cell number × parameter dimension). During data stream updates, the system samples only the latest frame each time. Insert at the end of the window and remove the oldest frame to implement a sliding window update mechanism.

[0022] Because the dynamic characteristics of each parameter differ significantly—voltage and current change rapidly, while temperature and internal resistance change slowly—normalization is performed on each parameter dimension within the time window to ensure the model can perceive different feature dimensions in a balanced manner. ; in Represents the original value of a parameter. and These are the moving mean and moving standard deviation of the parameter within the current window, respectively. Normalization is performed independently within each parameter dimension to ensure uniform numerical scale across different parameters; to ensure consistency, and It is fixed based on training data before system deployment and does not change dynamically during operation.

[0023] The final output is a state sequence tensor. This tensor retains the most recent The cell's operating state at any given moment is the sole input to the subsequent anomaly trend identification model. This method integrates the raw sensor data into a three-dimensional structure of time, space, and features, balancing the feasibility of data acquisition with the structural integrity of subsequent analysis, thereby achieving an accurate description of the battery's charge and discharge state.

[0024] S2. Based on the state vector analysis of the several battery cells, the trend vector and fluctuation intensity of each battery cell parameter are analyzed to calculate the initial risk score of the battery cell, and combined with the spatial thermal coupling relationship between the battery cells, the comprehensive risk score of each battery cell is calculated.

[0025] Specifically, during the charging and discharging operation of a battery system, the degradation of cell performance often begins with extremely subtle physical changes, such as a slight increase in internal resistance, a minor temperature drift, or a slight uneven discharge trend. If left unchecked, these trends will gradually amplify in subsequent cycles, ultimately leading to reduced charging and discharging efficiency, deterioration of cell consistency, and even thermal runaway. Therefore, relying solely on static rules or single-point thresholds cannot meet the requirements of highly reliable intelligent battery management. This step builds upon the structured state-time series tensor constructed in the previous step. Based on this, a trend-driven anomaly scoring mechanism is proposed, aiming to identify potential risks in advance and provide an actionable basis for identifying key targets for the strategy prediction module. This step directly uses... As input, where It is the first The state set of all cells in the battery pack at any given time, for each cell exist The state vector at time t is These variables originate from hardware acquisition devices (voltage sampling chip, current sensor, thermistor, internal resistance estimation circuit), and have been preprocessed in the previous step to form a unified and complete historical state window.

[0026] In the trend modeling process, considering the differences in structural layout and load environment of different cells, even if the physical indicators show the same numerical changes, their risk levels may be different. Therefore, this step introduces two key innovations in parameter evolution modeling: (1) joint anomaly expression based on time series trend slope and fluctuation intensity; (2) regularized risk calibration mechanism combining cell position weight and deviation mode.

[0027] The first innovation lies in the anomaly trend detection structure. We extract the slope indices of temperature, internal resistance, and voltage by fitting a first-order linear regression of each parameter within the window over the time dimension to construct a trend vector. Simultaneously calculate the residual variance This indicates the intensity of the fluctuation. To control the number of variables, we use the residuals and slope norm together as the main criteria for judgment: ; in, This is the initial risk score for the i-th battery cell. They represent battery cells respectively. The slope of the time fitting of temperature, internal resistance, and voltage within the sliding window. This represents the variance of the regression residuals for the corresponding parameters within this window. This represents the weighting of trends and fluctuations. The series of parameters were obtained by fitting using the standard least squares method. The variances from the fitted error sequence are all in the tensor It is implemented internally and has no external dependencies.

[0028] The second innovation lies in the structural regularization of the risk scoring. Considering the thermal coupling characteristics of battery cells in the physical topology, we further... A structural regularization term is introduced to reflect the weighted impact of high-risk cells on spatial propagation. The battery pack structure is assumed to be a nearest-neighbor graph. ,definition For battery cells The neighbor who is close to you, on the other hand Add neighbor propagation item: ; in, The overall risk score for the i-th cell is... The thermal coupling weights of adjacent edges in the structure graph. This is the regularization coefficient, used to suppress isolated anomalous points in the structure and amplify local consistency differences, thereby improving the system's ability to detect "local hotspot clusters." The final output... This is a comprehensive risk score for each battery cell, reflecting multiple pieces of information including short-term trend changes, volatility intensity, and structural inconsistencies. This scoring vector... It will be used as input for the next step, participating in the target weighting and path selection of the policy prediction model.

[0029] in, The overall risk score for the i-th cell is... The thermal coupling weights of adjacent edges in the structure graph. This is the regularization coefficient, used to suppress isolated anomalous points in the structure and amplify local consistency differences, thereby improving the system's ability to detect "local hotspot clusters." The final output... This is a comprehensive risk score for each battery cell, reflecting multiple pieces of information including short-term trend changes, volatility intensity, and structural inconsistencies. This scoring vector... It will be used as input for the next step, participating in the target weighting and path selection of the policy prediction model.

[0030] S3. Combining the battery pack state set and comprehensive risk score, simulate and predict multiple preset charging and discharging strategies, and design an objective function based on temperature suppression, voltage equalization and control complexity for evaluation, and select the optimal charging and discharging strategy.

[0031] Specifically, in large-scale battery systems, the selection of charging and discharging strategies is often influenced by multiple factors, including performance differences between cells, thermal coupling in spatial layout, distribution of risk states, and execution complexity. Traditional strategies are often based on uniform current limiting or single-point triggering logic, lacking forward-looking prediction of risk propagation trends and local response capabilities. This results in the inability to proactively adjust the strategy in a timely manner when high-risk cells experience early degradation. This step is based on the state sequence tensor output from step one. and the cell risk score vector output in step two This paper proposes a strategy simulation and selection method that integrates risk guidance, multi-objective trade-offs, and local group optimization, aiming to realize a structure-aware and trend-responsive optimal charging and discharging strategy generation mechanism.

[0032] Regarding input data, Provided the past The battery pack's full-state information within each time step includes the voltage, current, temperature, and internal resistance of each cell, possessing complete temporal variability; while the risk vector This integrates information on cell slope trends, residual fluctuations, and structural regularization, indicating key areas requiring intervention in the current system. Together, these form the input basis for the strategy prediction model.

[0033] This step first constructs a set of policy candidates. Each strategy This corresponds to a set of charge and discharge control operations, such as setting different current limits. Options include whether to enable active balancing and whether to activate local cooling fans. These strategies are limited combinations of actions defined in the system settings before deployment, and have a structure that can be directly converted into BMS control commands.

[0034] To achieve the simulation and prediction of future states, a set of independently trained lightweight one-dimensional convolutional temporal networks is used. tensor For input, different strategies Output the future Step cell temperature and voltage trajectory The model structure consists of three convolutional layers, each with 16 channels, a kernel size of 3, and the ReLU activation function. The output layer is a cell-by-cell multi-channel regression unit. During the training phase, the model uses actual running data under the policy for end-to-end learning. During deployment, it is frozen and only performs forward inference.

[0035] In the strategy evaluation phase, design an objective function that integrates trend response, energy balance, and control complexity. This function is used to measure the strategy. The future operating effect. The formula is as follows: ; In the formula, Representation Strategy Under the action of the first The first battery cell will be the first in the future Predicted temperature step; This indicates its predicted voltage sequence; The variance represents the voltage fluctuation and reflects the load balance of the system. It is the current risk score of the battery cell, which serves as an anomaly weighting factor to guide the system to adopt stronger constraint strategies for high-risk battery cells. It is the control strategy complexity term, specifically the number of operands generated by the strategy in the cell control channel, reflecting the execution cost of the control system. and To adjust the parameters.

[0036] Furthermore, the objective function exhibits three innovative design features: First, the difference between the temperature response term and the current temperature emphasizes the ability to "suppress trends" rather than controlling absolute temperature; second, the voltage fluctuation term, calculated through variance, makes the system more inclined to choose a balanced strategy that stabilizes voltage, especially when multiple medium-risk cells are present; and third, the control complexity term avoids high-frequency control commands or activating too many channels, reducing system power consumption and controller load. For example, in a typical application, if a strategy needs to simultaneously control the balanced current channels of 16 cells, then... This value will increase significantly. The final score was used to exclude the candidate if the risk control benefits were not obvious.

[0037] The strategy selection mechanism ultimately outputs the optimal strategy. : ; This strategy will be used in the next step to generate control commands that can be issued to drive the battery pack into a new operating mode.

[0038] This step introduces a risk-driven trend suppression objective function, local voltage stability assessment, and control execution cost regularization term to construct a strategy simulation and evaluation method that conforms to the actual operation of battery packs. Compared with traditional average loss-based evaluation methods, this scheme emphasizes the balance between precise intervention in key cells and the physical cost of the controller, making it particularly suitable for energy storage, electric vehicles, and UPS systems that need to extend system life under cell aging conditions and avoid excessive maintenance triggers. This strategy mechanism is the core decision-making unit of the entire invention, and its bidirectional adaptability, combining risk prediction and strategy control, forms the central hub of a complete intelligent closed loop.

[0039] S4. Generate control commands based on the optimal charging and discharging strategy, and generate visualization information on the state evolution of the battery pack based on the simulation prediction results of the charging and discharging strategy. At the same time, send the control commands to the battery management system for execution.

[0040] Specifically, in the intelligent management system for battery pack operation, strategy generation is only one part of decision-making. More crucial is how to translate the strategy into physically executable control commands, enabling system operators or engineers to clearly and intuitively understand the predictive logic and potential impact behind the strategy. To this end, this step designs a mechanism that integrates state evolution visualization and automatic control command generation, relying on the optimal strategy obtained in the previous step. and its corresponding future state prediction sequence This completes the engineering closed loop "from prediction to execution".

[0041] The two input data used in this step are derived from step three. Optimal strategy Based on risk scoring vector and state sequence The derived strategy combination for minimizing overall losses includes specific control decisions such as current limiting schemes for each cell, equalization control flags, and cooling unit switching logic. Predicted state sequence. Generated by a structured time series prediction model, containing the predicted voltage of each cell over several future time steps. ,temperature and current The visualization module uses the spatial topology of the battery pack as a basis to construct a 3D interactive interface. The system reads the physical arrangement number of each cell (e.g., in some 48-cell lithium battery systems, the cell numbers are arranged in a 12×4 matrix from 1 to 48), and displays the current state. The state of each cell in the model is mapped to a cubic entity in a 3D model. Color channels are used to represent the current temperature. The level of risk is indicated by varying shades (e.g., using a blue-green-red gradient), while transparency is used to indicate the current risk level. The size of the mapping remains consistent throughout the prediction process and transitions gradually over time frames to simulate the cell state under the strategy. Evolutionary trends under execution.

[0042] Predicted state Taking the temperature component as an example, the system calculates the predicted trend offset index for each cell: ; This metric is used to assess the magnitude of cell temperature changes during strategy execution. In the visualization interface, it is mapped as an upward arrow, heat diffusion border, or abnormal symbol in the animation, helping maintenance personnel to identify potential hot spots, poor local heat dissipation areas, or control blind spots.

[0043] Furthermore, the system also provides a "before-policy" comparison perspective, displayed through overlay frame comparison. and The system differentiates between different groups of cells. For example, during charging, if the predicted temperature change of a certain group of cells exceeds 5°C, the system will highlight that group of cells and display its risk score. The curves showing changes over time are presented to the operations team in a combination of graphs and numerical values, improving the interpretability of operations.

[0044] In terms of control command generation, the system analyzes the optimal strategy. The structure is defined in the code and organized into a standardized instruction set for the actual communication protocol of the controller (such as CAN or RS485). Each control command For battery cells It contains four fields: Cell number Corresponding system wiring identifier; Rate limiting value Extracted from the strategy, in units of A, with numerical precision accurate to 0.1A; Balance sign : 0 indicates off, 1 indicates active balancing control is enabled; Cooling control If the predicted cell temperature is higher than 45℃, the corresponding fan module will be forcibly turned on.

[0045] The instruction encoding logic needs to be matched with the BMS hardware platform during actual deployment. If a domestic main control MCU solution (such as the GD32 series) is used, batch writing can be achieved through DMA + interrupt mode, with an instruction refresh frequency of 1~2Hz per second to ensure response speed.

[0046] To ensure the reproducibility and verifiability of the control strategy, the system will... and The data is stored in the local policy cache and the execution time is recorded for later backtracking and policy effect analysis. After receiving the control command, the controller directly writes the current limiting value into the PWM control register, writes the equalization flag into the voltage equalization switch control module (e.g., through a MOSFET switching matrix), and the cooling control is linked to the fan drive circuit to adjust the speed or switch it on and off.

[0047] A second aspect of the present invention provides a battery visualization charging and discharging intelligent management system, the system comprising: The data acquisition and processing module is used to construct a battery pack state set for the battery pack; wherein, the battery pack state set includes state vectors of several cells, and the state vectors of the cells include cell voltage, current, temperature and internal resistance; The risk analysis module is used to analyze the trend vector and fluctuation intensity of the parameters of each battery cell based on the state vector of the several battery cells, so as to calculate the initial risk score of the battery cell, and calculate the comprehensive risk score of each battery cell by combining the spatial thermal coupling relationship between the battery cells. The strategy simulation and decision-making module is used to combine the battery pack state set and comprehensive risk score to simulate and predict multiple preset charging and discharging strategies, and to design an objective function based on temperature suppression, voltage equalization and control complexity for evaluation, and select the optimal charging and discharging strategy. The visualization and control execution module is used to generate control commands based on the optimal charging and discharging strategy, generate visualization information of the battery pack state evolution based on the simulation prediction results of the charging and discharging strategy, and send the control commands to the battery management system for execution.

[0048] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0049] In the embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of apparatuses or units may be electrical, mechanical, or other forms.

[0050] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0051] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for visualized intelligent charging and discharging management of batteries, characterized in that: The method includes: S1. Construct a battery pack state set for the battery pack; wherein, the battery pack state set includes state vectors of several cells, and the state vectors of the cells include cell voltage, current, temperature and internal resistance; S2. Based on the state vector analysis of the several battery cells, the trend vector and fluctuation intensity of each battery cell parameter are analyzed to calculate the initial risk score of the battery cell, and combined with the spatial thermal coupling relationship between the battery cells, the comprehensive risk score of each battery cell is calculated. S3. Combining the battery pack state set and comprehensive risk score, simulate and predict multiple preset charging and discharging strategies, and design an objective function based on temperature suppression, voltage equalization and control complexity for evaluation, and select the optimal charging and discharging strategy. S4. Generate control commands based on the optimal charging and discharging strategy, and generate visualization information on the state evolution of the battery pack based on the simulation prediction results of the charging and discharging strategy. At the same time, send the control commands to the battery management system for execution.

2. The intelligent management method for visualized charging and discharging of batteries according to claim 1, characterized in that, The voltage is obtained by a voltage sampling chip inside the BMS; the current is measured by a Hall sensor or shunt resistor connected in series in the cell branch; the temperature is collected by an NTC thermistor in close contact with the cell surface; and the internal resistance is calculated by measuring the voltage change by applying a short-time load pulse.

3. The intelligent management method for visualized charging and discharging of batteries according to claim 1, characterized in that, Before constructing the battery pack state set, the cell voltage, current, temperature, and internal resistance are normalized within a time window.

4. The intelligent management method for visualized charging and discharging of batteries according to claim 1, characterized in that, The trend vector is a slope index of temperature, internal resistance, and voltage; the fluctuation intensity is the variance of the regression residual of the corresponding cell's state vector in the corresponding time window.

5. The intelligent management method for visualized charging and discharging of batteries according to claim 4, characterized in that, The comprehensive risk score for each battery cell is calculated based on the spatial thermal coupling relationship between the cells, specifically as follows: Assuming the battery pack structure is a nearest neighbor graph, and defining the second cell as a thermal neighbor of the first cell; wherein, the nearest neighbor graph includes the second cell; Based on the second cell, a neighbor propagation term based on thermal coupling weight is constructed. Combined with the initial risk score of the first cell, the comprehensive risk score of the first cell is obtained.

6. The intelligent management method for visualized charging and discharging of batteries according to claim 1, characterized in that, The preset multiple charging and discharging strategies correspond to multiple sets of charging and discharging control operations. The simulation prediction of multiple preset charging and discharging strategies specifically includes: The state vector of the battery cell is input into a set of independently trained lightweight one-dimensional convolutional temporal networks, which output the predicted battery cell temperature and predicted voltage trajectory for the next H steps. The lightweight one-dimensional convolutional temporal network consists of 3 convolutional layers, each with 16 channels, a kernel size of 3, an activation function of ReLU, and an output layer consisting of a multi-channel regression unit per battery cell.

7. The intelligent management method for visualized charging and discharging of batteries according to claim 6, characterized in that, The temperature suppression is based on the current charge / discharge strategy under the first... The predicted cell temperature for each cell in the next H steps is calculated; the voltage equalization is calculated based on the predicted voltage trajectory for the next H steps; and the control complexity is calculated based on the operands generated in the cell control channel by the current charging and discharging strategy.

8. The intelligent management method for visualized charging and discharging of batteries according to claim 1, characterized in that, The simulation prediction results based on the charge-discharge strategy include the predicted cell temperature, predicted voltage trajectory, and predicted current for each cell.

9. The intelligent management method for visualized charging and discharging of a battery according to claim 1, characterized in that, The control commands include cell number, current limit value, equalization flag, and cooling control.

10. A visualized intelligent charging and discharging management system for batteries, characterized in that: The system includes: The data acquisition and processing module is used to construct a battery pack state set for the battery pack; wherein, the battery pack state set includes state vectors of several cells, and the state vectors of the cells include cell voltage, current, temperature and internal resistance; The risk analysis module is used to analyze the trend vector and fluctuation intensity of the parameters of each battery cell based on the state vector of the several battery cells, so as to calculate the initial risk score of the battery cell, and calculate the comprehensive risk score of each battery cell by combining the spatial thermal coupling relationship between the battery cells. The strategy simulation and decision-making module is used to combine the battery pack state set and comprehensive risk score to simulate and predict multiple preset charging and discharging strategies, and to design an objective function based on temperature suppression, voltage equalization and control complexity for evaluation, and select the optimal charging and discharging strategy. The visualization and control execution module is used to generate control commands based on the optimal charging and discharging strategy, generate visualization information of the battery pack state evolution based on the simulation prediction results of the charging and discharging strategy, and send the control commands to the battery management system for execution.

Images