A BMS active advance equalization method and system based on time series prediction and intelligent planning

Abstract

The application provides a BMS active leading equalization method and system based on time series prediction and intelligent planning, belonging to the technical field of energy storage, which comprises the following steps: collecting cell voltage, SOC, SOH, temperature and charging and discharging current data, inputting the data into a time series prediction model after denoising and normalization preprocessing to obtain cell state prediction values; completing fault diagnosis based on the prediction values and the current state, dividing normal cells into high-voltage groups, normal groups and low-voltage groups according to the voltage prediction values after isolating abnormal cells; when the voltage difference between groups exceeds a preset threshold, taking the maximization of the total available capacity of the battery system as the target, combining the equalization loop distance minimization and the total equalization power maximization constraints to perform global path planning, and controlling the energy transfer from the high-voltage groups to the low-voltage groups; at the same time, marking the cells in the low-voltage groups for multiple times in succession and performing differentiated equalization strategies. The application realizes leading equalization, improves the consistency, energy utilization rate and operation safety of the battery pack.

Description

Technical Field

[0001] This application belongs to the field of energy storage technology, specifically relating to a BMS active advance balancing method and system based on time series prediction and intelligent planning. Background Technology

[0002] As lithium battery applications continue to evolve towards higher rates and larger capacities, the consistency issue of battery packs is becoming increasingly prominent. Even cells produced in the same batch may have slight differences in parameters such as capacity, voltage, and internal resistance. Over long-term use, these slight differences can be amplified, triggering a "weakest link" effect. Ultimately, this results in the actual usable capacity of the battery pack being far lower than the theoretical capacity of a single cell, leading to a significant decrease in charge and discharge efficiency and severely restricting the energy utilization rate of lithium batteries.

[0003] To address the issue of cell inconsistency, existing technologies primarily employ two approaches: passive balancing and active balancing. Passive balancing methods often dissipate excess energy from cells with high voltage as heat through resistors. While this approach is simple and inexpensive, it suffers from significant drawbacks: severe energy waste, with all excess energy converted into heat, reducing the energy efficiency of the battery system and increasing the burden on the cooling system; balancing speed is limited by resistor power, preventing rapid balancing; and it can only balance cells with high voltage at the end of charging, failing to regulate energy during discharge or compensate for degraded cells, thus failing to fundamentally solve the battery pack inconsistency problem.

[0004] Active balancing schemes utilize energy storage components such as inductors, capacitors, transformers, or switch arrays to transfer energy from high-energy cells to low-energy cells, achieving energy redistribution and theoretically resulting in higher energy utilization. However, most current active balancing systems still suffer from the following shortcomings: The balancing control strategy is lagging, relying solely on current cell state parameters such as individual cell voltage, state of charge (SOC), and state of health (SOH) to mechanically respond according to set thresholds, without considering future state changes. This is a reactive balancing approach, unable to intervene before cell differences amplify, leading to poor balancing performance. Furthermore, they lack global planning capabilities; the balancing process is often limited to local cell-to-cell transitions, requiring one or several cells to complete the energy transfer. This not only increases intermediate losses but also makes it difficult to achieve balancing across the entire battery pack or cluster, failing to maximize the overall usable capacity of the battery system. Finally, while the existing Long Short-Term Memory (LSTM) prediction model can handle time-series data, it lacks the ability to capture long-distance dependencies in long-sequence data, resulting in limited prediction accuracy.

[0005] Therefore, there is an urgent need to develop a BMS active advance balancing system and method that can predict changes in cell state in advance, has global intelligent planning capabilities, and has high balancing efficiency, in order to solve the above-mentioned problems in the existing technology. Summary of the Invention

[0006] The purpose of this invention is to provide a BMS active advance equalization method and system based on time-series prediction and intelligent planning, which solves the problems of lagging equalization control, lack of global planning, low equalization efficiency and insufficient prediction accuracy in existing active equalization technologies. It realizes advance prediction of cell status and global optimal equalization of the battery system, effectively suppresses the amplification of cell differences, and improves the overall consistency and energy utilization of the battery pack.

[0007] In a first aspect, embodiments of this application provide a BMS proactive advance equalization method based on time series prediction and intelligent planning, the method comprising: S1. Collect historical operating data of each cell in the battery system, including voltage, SOC, SOH, temperature and charge / discharge current, and perform noise reduction and normalization preprocessing on the historical operating data; S2. Input the preprocessed historical operating data into the time series prediction model to predict the operating status of each cell after a preset time, and obtain the state prediction value of each cell, including voltage prediction value, SOC prediction value and SOH prediction value. S3. Based on the predicted status value and the current operating status of each battery cell, perform fault diagnosis and anomaly warning for each battery cell, and isolate abnormal battery cells; S4. Sort all normal battery cells in descending order of their predicted voltage values, and divide the top 20% into the high-voltage group, the bottom 20% into the low-voltage group, and the middle 60% into the normal group. S5. Calculate the average voltage of the three groups respectively to obtain the average voltage of the high voltage group, the average voltage of the normal group and the average voltage of the low voltage group. Calculate the difference between the average voltage of the high voltage group and the average voltage of the normal group, and record it as the first difference. The difference between the average voltage of the normal group and the average voltage of the low voltage group is recorded as the second difference. S6. If the first difference or the second difference is greater than the preset equalization threshold, the global equalization process is started. With the goal of maximizing the total available capacity of the battery system, the global equalization path is planned in combination with the constraints of minimizing the loop distance and maximizing the total equalization power, and equalization execution instructions are generated. S7. According to the equalization execution instruction, control the equalization execution module to complete the energy transfer from the high-voltage group cell to the low-voltage group cell; S8. Mark the cells that have been assigned to the low-voltage group more than m times consecutively as abnormal, and implement a differentiated equalization strategy that increases the upper limit of the equalization voltage threshold and the actual voltage conversion.

[0008] Further, in step S2, the temporal prediction model is a BiLSTM-MultiHeadAttention model combining a bidirectional long short-term memory network and a multi-head self-attention mechanism. Its construction and training process specifically includes: S21. Construct a BiLSTM-MultiHeadAttention model, which includes an input layer, a bidirectional LSTM layer, a multi-head self-attention layer, a fully connected layer, and an output layer in sequence. S22. Divide the preprocessed historical running data into training set, validation set and test set in chronological order, and generate training samples using the sliding window method; S23. Train the BiLSTM-MultiHeadAttention model using the training set, using the mean squared error (MSE) as the loss function, and use the Adam optimizer to update the parameters; S24. Use the validation set to validate the trained model, monitor the changes in the validation set loss function value, and stop training and save the optimal model parameters when the conditions for stopping training are met. S25. Test the saved optimal model using the test set. When the average absolute error of voltage prediction in the test set is less than the preset voltage value, the average absolute error of SOC prediction is less than the first preset percentage, and the average absolute error of SOH prediction is less than the second preset percentage, the model training is deemed complete.

[0009] Furthermore, the specific steps of step S3 include: S31. Real-time acquisition of the predicted voltage, SOC, and SOH values ​​of each individual cell, as well as the current voltage, current SOC, current SOH, and current charge / discharge status. S32. If the predicted voltage value of a certain cell is higher than the preset upper limit of charging voltage Vh when charging, or lower than the preset lower limit of discharging voltage Vl when discharging, the cell is determined to be a seriously abnormal cell, and an instruction to disconnect the battery cluster where the cell is located is immediately generated, and the location information and abnormality type of the cell are recorded in the system log. S33. Calculate the deviation rate between the current voltage of a certain battery cell and the predicted voltage value at the previous moment. The formula for calculating the deviation rate is: Deviation rate = |Current voltage - Previous predicted voltage value| / Previous predicted voltage value × 100% If the deviation rate exceeds 20% of the preset deviation threshold, the cell is determined to be a voltage abnormal cell, a first-level alarm command is immediately generated, and its location information and deviation rate value are recorded. S34. If the predicted SOC value of a certain cell exceeds the preset normal SOC range, or the predicted SOH value is lower than the preset lower limit of SOH, or the deviation between the current SOC and the predicted SOC value of the previous moment exceeds 5%, or the deviation between the current SOH and the predicted SOH value of the previous moment exceeds 2%, then the cell is determined to be an abnormal cell, a level 2 alarm command is immediately generated, and its location information and abnormal state parameters are recorded. S35. Based on the generated abnormality handling instructions, control the main contactor of the battery management system to disconnect the battery cluster containing the corresponding abnormal cell, and at the same time send the alarm information and abnormal cell location information to the host computer module for display.

[0010] Furthermore, in step S6, the constraints of the global equilibrium path planning include: C1. During the balancing process, the total distance of the balancing current loop between all high-voltage group cell units and low-voltage group cell units participating in the balancing is minimized. The total distance of the balancing current loop is the sum of the physical distances between each pair of balancing cell units. C2. During the balancing process, the total balancing power of all high-voltage group cells and low-voltage group cells participating in the balancing process is maximized within the next minute. The total balancing power is the sum of the balancing power of each balancing cell pair. C3. During the balancing process, the real-time voltage of any single cell in the high-voltage group must not be lower than the real-time average voltage of the low-voltage group; C4. During the balancing process, the real-time voltage of any single cell in the low-voltage group must not be higher than the real-time average voltage of the high-voltage group.

[0011] Furthermore, in step S6, the convergence conditions for the global equilibrium path planning include: When the first difference is greater than the second difference, continue to perform the balancing operation until one of the following conditions is met: T1. The difference between the real-time average voltage of the high-voltage group and the real-time average voltage of the low-voltage group is less than the preset convergence threshold D; T2. The real-time average voltage of the high-voltage group is lower than the real-time average voltage of the normal group; T3. The real-time average voltage of the low-voltage group reaches the sum of the real-time average voltage of the normal group and 50% of the initial first difference; When the first difference is less than or equal to the second difference, gradually increase the number of individual cells in the high-voltage group, increasing the total number of normal cells by 1% each time, and recalculate the new average voltage of the high-voltage group and the first difference, until the new first difference satisfies the following: The new first difference × (1 + N × 5%) > the second difference Where N represents the number of times the high-voltage group is reselected, and then the equilibration operation is performed according to the constraints and convergence conditions.

[0012] Furthermore, in step S6, when the number of individual battery cells in the battery system exceeds 1000, a step-by-step balancing method is used to perform the balancing process, specifically including: S61. Divide the individual cells of the entire battery system into several cell groups according to their physical location. Each cell group contains a fixed number of individual cells. Each battery cluster has a total of M cell groups and each cell group has a total of K individual cells. S62. Perform a balancing process once within each cell group. Based on the predicted state value of each cell, perform balancing planning within the group. The goal is to make the voltage deviation of the K cells in the group within the first acceptable range V0±D0. The default is to minimize balancing loss as the secondary goal. The host computer can switch to make the fastest balancing rate as the secondary goal. S63. During the group balancing process, the constraints of minimizing the total distance of the balancing current loop and maximizing the total balancing power are followed. The 20% of cells with the lowest voltage in the group are marked as low-voltage cells in the group, and their actual voltage is converted according to a preset ratio. If a cell is marked as a low-voltage cell in the group more than 3 times in a row, the cell is recorded as a potential abnormal cell. S64. If the proportion of potentially abnormal cells in a certain cell group exceeds 10%, then the entire cell group is marked as a potentially abnormal cell group, and a level 3 alarm command is generated and sent to the host computer module. S65. After completing one equalization of all cell groups, calculate the available power C of each cell group and the variance F of the available power of all cell groups based on the state prediction value of each cell group, and sort the M cell groups in descending order of available power. S66. If the variance F of the available power is greater than the preset inter-group balancing threshold Fset, then the secondary balancing process is started, and inter-group balancing planning is carried out on a cell-by-cell basis. The goal is to make the average voltage deviation of the M cell groups within the second acceptable range Vg±Dg. The default is to minimize the balancing loss as the secondary goal, and the host computer switches to the fastest balancing rate as the secondary goal. S67. During the inter-group balancing process, the constraints of minimizing the total distance of the balancing current loop and maximizing the total balancing power are also followed to complete the inter-group energy transfer; S68. Regularly generate abnormal cell reports, record the location information, operating status parameters and balancing history data of all potential abnormal cells and groups of cells, and prompt maintenance personnel to check and replace them.

[0013] Secondly, embodiments of this application also provide a BMS proactive advance equalization method and system based on time series prediction and intelligent planning, the system comprising: The data acquisition module is used to collect historical operating data of each cell in the battery system, including voltage, SOC, SOH, temperature and charge / discharge current, and to perform noise reduction and normalization preprocessing on the historical operating data. The time-series prediction module is used to input preprocessed historical operating data into the time-series prediction model to predict the operating status of each cell after a preset time, and obtain the state prediction value of each cell, including voltage prediction value, SOC prediction value and SOH prediction value. The fault diagnosis module is used to diagnose and provide early warning of abnormalities for individual cells based on the predicted status values ​​and the current operating status of each cell, and to isolate abnormal cells. The cell grouping module is used to sort all normal cells in descending order of their predicted voltage values, dividing the top 20% into the high-voltage group, the bottom 20% into the low-voltage group, and the middle 60% into the normal group. The voltage calculation module is used to calculate the average voltage of the three groups respectively, to obtain the average voltage of the high voltage group, the average voltage of the normal group, and the average voltage of the low voltage group. It calculates the difference between the average voltage of the high voltage group and the average voltage of the normal group, which is recorded as the first difference, and the difference between the average voltage of the normal group and the average voltage of the low voltage group, which is recorded as the second difference. The equalization decision module is used to initiate a global equalization process if the first difference or the second difference is greater than the preset equalization threshold. With the goal of maximizing the total available capacity of the battery system, it performs global equalization path planning by combining the constraints of minimizing loop distance and maximizing total equalization power, and generates equalization execution instructions. The equalization execution module is used to control the equalization execution module to complete the energy transfer from high-voltage group cell to low-voltage group cell according to the equalization execution instruction. The abnormal cell management module is used to mark cells that have been assigned to the low-voltage group more than m times consecutively as abnormal, and to implement a differentiated equalization strategy that increases the upper limit of the equalization voltage threshold and the actual voltage conversion.

[0014] Furthermore, the equalization execution module includes multiple equalization sub-modules and a bidirectional equalization connection line. Each equalization sub-module is responsible for the equalization of 5 individual battery cells, and adjacent equalization sub-modules are connected by a bidirectional equalization connection line. Each equalization submodule specifically includes: The equalization circuit, including a wide-input-range bidirectional DC-DC converter circuit and its control circuit, is used to realize energy transfer between individual cells during the equalization process. The first circuit selection switch includes a 5-channel cell selection switch array. One end can individually select the positive and negative terminals of any one of the 5 series-connected cells. The other end is connected to one end of the equalization circuit to connect the high-voltage cell that needs to output energy to the low-voltage cell that needs to input energy. The second-path selection switch circuit includes a 5-channel cell selection switch array. One end can individually select the positive and negative terminals of any single cell among the 5 cells connected in series; the other end is connected to the other end of the equalization circuit, and together with the first-path selection switch circuit, they form a complete equalization loop. A bidirectional balancing connection line, which is a bidirectional switch, is provided. One end of the bidirectional switch is connected to the other end of the balancing circuit, and the other end is connected to the bidirectional switch of the adjacent balancing submodule, so as to realize energy interaction between modules. The drive control unit, connected to the equalization decision module, is used to receive equalization execution instructions sent by the equalization decision module, and control the on / off state of the first path selection switch circuit, the second path selection switch circuit and the bidirectional switch according to the instructions, so as to realize energy transfer between any two battery cells or between a battery cell and an adjacent equalization submodule.

[0015] Furthermore, the abnormal cell management module specifically includes: The statistics unit is used to count the number of times each cell is assigned to the low-voltage group during multiple consecutive equalization processes, and the number of times it is marked as a low-voltage cell within the group during the equalization process within the group. An anomaly marking unit, connected to the statistics unit, is used to mark a cell as a potential abnormal cell when a cell is assigned to a low-voltage group more than 3 times or is marked as a low-voltage cell in a group more than m times; and to mark a cell group as a potential abnormal cell group when the proportion of potential abnormal cells in a cell group exceeds 10%. The differentiated strategy generation unit, connected to the anomaly marking unit, is used to generate a differentiated balancing strategy for battery cells marked as potentially abnormal. Specifically, it includes: increasing the upper limit of the balancing voltage threshold of the battery cell by 10mV, and calculating its actual voltage according to a preset ratio during the subsequent balancing process, so that the balancing decision module can give priority to energy compensation when planning. The report generation unit, connected to the anomaly marking unit, is used to periodically generate abnormal cell reports. The report content includes the location information, operating status parameters, historical balancing data, and anomaly level of potential abnormal cells and groups of cells. The reporting cycle can be set to daily, weekly, or monthly by the host computer.

[0016] Furthermore, the system also includes a host computer module; The host computer module is connected to the equalization decision module, fault diagnosis module, and abnormal cell management module via a CAN bus, and is used for: Send the equalization task instruction to the equalization decision module and set the equalization-related parameters, including the preset equalization threshold G, the preset convergence threshold D, the preset deviation threshold, the preset upper limit of charging voltage Vh, the preset lower limit of discharging voltage Vl, the preset normal range of SOC, the preset lower limit of SOH, the first acceptable range of voltage deviation within the group V0±D0, the second acceptable range of voltage deviation between groups Vg±Dg, and the equalization threshold Fset between groups. It receives alarm information at all levels and abnormal cell location information sent by the fault diagnosis module, displays them visually on the screen, and alerts maintenance personnel through an audible and visual alarm device; Receive abnormal cell reports sent by the abnormal cell management module, and manage the storage, query, and printing of the reports; The system monitors the operating status and balancing process of the entire battery system in real time, displaying the real-time voltage, SOC, SOH, and temperature parameters of each cell, as well as the balancing current, balancing power, and energy transfer during the balancing process. It supports historical data query function, allowing users to query cell operation data, balance history records and abnormal alarm records for any time period, and generate corresponding trend analysis charts.

[0017] As can be seen from the above technical solutions, the present invention has the following advantages: This invention employs a BiLSTM-MultiHeadAttention time-series prediction model, which can simultaneously extract bidirectional time-series features and capture long-distance dependencies, improving the prediction accuracy of cell voltage, SOC, and SOH. Based on the prediction results, it achieves advanced equalization, fundamentally suppressing the amplification of cell parameter differences. It adopts a global intelligent planning strategy, with the core objective of maximizing the total usable capacity of the battery system. It matches the optimal high- and low-voltage cell pairings across the entire battery pack / cluster, avoiding multi-step energy over-loss during local equalization. Simultaneously, it considers minimizing the equalization loop distance and maximizing total power constraints, improving equalization efficiency and overall battery pack consistency. Through a step-by-step, layered equalization architecture, it solves the problem of global planning computational overload in large-scale battery systems, significantly reducing hardware requirements while ensuring equalization effectiveness. It possesses strong scalability and can adapt to lithium battery systems of different capacity levels. By statistically analyzing historical equalization data, it automatically identifies potentially degraded cells, executes differentiated energy compensation strategies to delay performance degradation, and promptly alerts battery packs with abnormal cell ratios exceeding the standard, improving the safety and reliability of system operation. Attached Figure Description

[0018] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the BMS proactive advance equilibrium method based on time series prediction and intelligent planning proposed in this application. Figure 2 This is a schematic diagram of the BMS active advance equilibrium system based on time series prediction and intelligent planning in this application. Figure 3 This is a schematic diagram of the balancing execution module of the BMS proactive leading balancing system based on time series prediction and intelligent planning in this application. Detailed Implementation

[0020] The various embodiments of the present invention will be described more fully in the detailed steps of the BMS proactive advance equalization method and system based on time series prediction and intelligent planning described below. The present invention may have various embodiments, and adjustments and changes may be made therein. However, it should be understood that there is no intention to limit the various embodiments of the present invention to the specific embodiments disclosed herein, but rather the present invention should be understood to cover all adjustments, equivalents, and / or alternatives falling within the spirit and scope of the various embodiments of the present invention.

[0021] It should be understood that, when used in this specification, the term "comprising" indicates the presence of the described feature, integral, step, operation, element, and / or component, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or collections thereof. The terms "comprising," "including," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0022] The terms "one embodiment" or "some embodiments" used in this application mean that one or more embodiments of this application include the specific features, structures, or characteristics described in that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this application do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized.

[0023] To facilitate a clear description of the technical solutions of this application, the terms "first" and "second" are used to distinguish identical or similar items with essentially the same function and effect. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and that the terms "first" and "second" do not necessarily imply that they are different.

[0024] To make the objectives, features, and advantages of this invention more apparent and understandable, specific embodiments and accompanying drawings will be used to clearly and completely describe the technical solutions protected by this invention. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0025] Please see Figure 1 The diagram shows a flowchart of a BMS proactive advance balancing method based on time series forecasting and intelligent planning. The method includes: S1. Collect historical operating data of each cell in the battery system, including voltage, SOC, SOH, temperature and charge / discharge current, and perform noise reduction and normalization preprocessing on the historical operating data; S2. Input the preprocessed historical operating data into the time series prediction model to predict the operating status of each cell after a preset time, and obtain the state prediction value of each cell, including voltage prediction value, SOC prediction value and SOH prediction value. S3. Based on the predicted status value and the current operating status of each battery cell, perform fault diagnosis and anomaly warning for each battery cell, and isolate abnormal battery cells; S4. Sort all normal battery cells in descending order of their predicted voltage values, and divide the top 20% into the high-voltage group, the bottom 20% into the low-voltage group, and the middle 60% into the normal group. S5. Calculate the average voltage of the three groups respectively to obtain the average voltage of the high voltage group, the average voltage of the normal group and the average voltage of the low voltage group. Calculate the difference between the average voltage of the high voltage group and the average voltage of the normal group, and record it as the first difference. The difference between the average voltage of the normal group and the average voltage of the low voltage group is recorded as the second difference. S6. If the first difference or the second difference is greater than the preset equalization threshold, the global equalization process is started. With the goal of maximizing the total available capacity of the battery system, the global equalization path is planned in combination with the constraints of minimizing the loop distance and maximizing the total equalization power, and equalization execution instructions are generated. S7. According to the equalization execution instruction, control the equalization execution module to complete the energy transfer from the high-voltage group cell to the low-voltage group cell; S8. Mark the cells that have been assigned to the low-voltage group more than m times consecutively as abnormal, and implement a differentiated equalization strategy that increases the upper limit of the equalization voltage threshold and the actual voltage conversion.

[0026] As a refinement and extension of the specific implementation of the above embodiments, in order to fully illustrate the specific implementation process of this embodiment, another BMS proactive advance equalization method based on time series prediction and intelligent planning is provided. This method includes the following steps: S1. Collect historical operating data of each cell in the battery system, including voltage, SOC, SOH, temperature and charge / discharge current, and perform noise reduction and normalization preprocessing on the historical operating data; S2. Input the preprocessed historical operating data into the time series prediction model to predict the operating status of each cell after a preset time (e.g., 1 minute) and obtain the state prediction value of each cell, including voltage prediction value, SOC prediction value and SOH prediction value. In step S2, the temporal prediction model is a BiLSTM-MultiHeadAttention model that combines a bidirectional long short-term memory network with a multi-head self-attention mechanism. Its construction and training process specifically includes: S21. Construct a BiLSTM-MultiHeadAttention model, which sequentially includes an input layer, a bidirectional LSTM layer, a multi-head self-attention layer, a fully connected layer, and an output layer. The input layer has an input dimension of 5, corresponding to 5 features: voltage, SOC, SOH, temperature, and charging / discharging current. The bidirectional LSTM layer has 2 layers, each with 128 hidden units. The multi-head self-attention layer has 8 attention heads, with an attention dimension of 64. The fully connected layer has 2 layers, with 64 and 32 neurons respectively. The output layer has an output dimension of 3, corresponding to the predicted voltage, predicted SOC, and predicted SOH. S22. Divide the preprocessed historical running data into training set, validation set and test set in chronological order, with the training set accounting for 70%, the validation set accounting for 20% and the test set accounting for 10%. Use the sliding window method to generate training samples with a window size of 60 and a step size of 1. S23. The BiLSTM-MultiHeadAttention model is trained using the training set, with mean squared error (MSE) as the loss function and the Adam optimizer used for parameter updates. The initial learning rate is set to 0.001, and the learning rate is reduced to 0.9 after every 10 epochs of training. S24. Use the validation set to validate the trained model, monitor the change in the validation set loss function value, and stop training when the conditions for stopping training are met, such as when the validation set loss function value no longer decreases for 5 consecutive epochs. Save the optimal model parameters. S25. Test the saved optimal model using the test set. When the average absolute error of voltage prediction in the test set is less than 5mV of the preset voltage value, the average absolute error of SOC prediction is less than 1% of the first preset percentage, and the average absolute error of SOH prediction is less than 0.5% of the second preset percentage, the model training is deemed complete.

[0027] S3. Based on the predicted status value and the current operating status of each battery cell, perform fault diagnosis and anomaly warning for each battery cell, and isolate abnormal battery cells; The specific steps of step S3 include: S31. Real-time acquisition of the predicted voltage, SOC, and SOH values ​​of each individual cell, as well as the current voltage, current SOC, current SOH, and current charge / discharge status. S32. If the predicted voltage value of a certain cell is higher than the preset upper limit of charging voltage Vh when charging, or lower than the preset lower limit of discharging voltage Vl when discharging, the cell is determined to be a seriously abnormal cell, and an instruction to disconnect the battery cluster where the cell is located is immediately generated, and the location information and abnormality type of the cell are recorded in the system log. S33. Calculate the deviation rate between the current voltage of a certain battery cell and the predicted voltage value at the previous moment. The formula for calculating the deviation rate is: Deviation rate = |Current voltage - Previous predicted voltage value| / Previous predicted voltage value × 100% If the deviation rate exceeds 20% of the preset deviation threshold, the cell is determined to be a voltage abnormal cell, a first-level alarm command is immediately generated, and its location information and deviation rate value are recorded. S34. If the predicted SOC value of a certain cell exceeds the preset normal SOC range (0%-100%), or the predicted SOH value is lower than the preset lower limit of SOH by 80%, or the deviation between the current SOC and the predicted SOC value of the previous moment exceeds 5%, or the deviation between the current SOH and the predicted SOH value of the previous moment exceeds 2%, then the cell is determined to be an abnormal cell, a level 2 alarm command is immediately generated, and its location information and abnormal status parameters are recorded. S35. Based on the generated abnormality handling instructions, control the main contactor of the battery management system to disconnect the battery cluster containing the corresponding abnormal cell, and at the same time send the alarm information and abnormal cell location information to the host computer module for display.

[0028] S4. Sort all normal battery cells in descending order of their predicted voltage values, and divide the top 20% into the high-voltage group, the bottom 20% into the low-voltage group, and the middle 60% into the normal group. S5. Calculate the average voltage of the three groups respectively to obtain the average voltage of the high voltage group, the average voltage of the normal group and the average voltage of the low voltage group. Calculate the difference between the average voltage of the high voltage group and the average voltage of the normal group, and record it as the first difference. The difference between the average voltage of the normal group and the average voltage of the low voltage group is recorded as the second difference. S6. If the first difference or the second difference is greater than the preset equalization threshold, the global equalization process is started. With the goal of maximizing the total available capacity of the battery system, the global equalization path is planned in combination with the constraints of minimizing the loop distance and maximizing the total equalization power, and equalization execution instructions are generated. In step S6, the constraints of the global equilibrium path planning include: C1. During the balancing process, the total distance of the balancing current loop between all high-voltage group cell units and low-voltage group cell units participating in the balancing is minimized. The total distance of the balancing current loop is the sum of the physical distances between each pair of balancing cell units. C2. During the balancing process, the total balancing power of all high-voltage group cells and low-voltage group cells participating in the balancing process is maximized within the next minute. The total balancing power is the sum of the balancing power of each balancing cell pair. C3. During the balancing process, the real-time voltage of any single cell in the high-voltage group must not be lower than the real-time average voltage of the low-voltage group; C4. During the balancing process, the real-time voltage of any single cell in the low-voltage group must not be higher than the real-time average voltage of the high-voltage group.

[0029] In step S6, the convergence conditions for the global equilibrium path planning include: When the first difference is greater than the second difference, continue to perform the balancing operation until one of the following conditions is met: T1. The difference between the real-time average voltage of the high-voltage group and the real-time average voltage of the low-voltage group is less than the preset convergence threshold D; T2. The real-time average voltage of the high-voltage group is lower than the real-time average voltage of the normal group; T3. The real-time average voltage of the low-voltage group reaches the sum of the real-time average voltage of the normal group and 50% of the initial first difference; When the first difference is less than or equal to the second difference, gradually increase the number of individual cells in the high-voltage group, increasing the total number of normal cells by 1% each time, and recalculate the new average voltage of the high-voltage group and the first difference, until the new first difference satisfies the following: The new first difference × (1 + N × 5%) > the second difference Where N represents the number of times the high-voltage group is reselected, and then the equilibration operation is performed according to the constraints and convergence conditions.

[0030] In step S6, when the number of individual battery cells in the battery system exceeds 1000, a step-by-step balancing method is used to perform the balancing process, which specifically includes: S61. Divide the individual cells of the entire battery system into several cell groups according to their physical location. Each cell group contains a fixed number of individual cells. Each battery cluster has a total of M cell groups and each cell group has a total of K individual cells. S62. Perform a balancing process once within each cell group. Based on the predicted state value of each cell, perform balancing planning within the group. The goal is to make the voltage deviation of the K cells in the group within the first acceptable range V0±D0. The default is to minimize balancing loss as the secondary goal. The host computer can switch to make the fastest balancing rate as the secondary goal. S63. During the group balancing process, the constraints of minimizing the total distance of the balancing current loop and maximizing the total balancing power are followed. The 20% of cells with the lowest voltage in the group are marked as low-voltage cells in the group, and their actual voltage is converted according to a preset ratio. If a cell is marked as a low-voltage cell in the group more than 3 times in a row, the cell is recorded as a potential abnormal cell. S64. If the proportion of potentially abnormal cells in a certain cell group exceeds 10%, then the entire cell group is marked as a potentially abnormal cell group, and a level 3 alarm command is generated and sent to the host computer module. S65. After completing one equalization of all cell groups, calculate the available power C of each cell group and the variance F of the available power of all cell groups based on the state prediction value of each cell group, and sort the M cell groups in descending order of available power. S66. If the variance F of the available power is greater than the preset inter-group balancing threshold Fset, then the secondary balancing process is started, and inter-group balancing planning is carried out on a cell-by-cell basis. The goal is to make the average voltage deviation of the M cell groups within the second acceptable range Vg±Dg. The default is to minimize the balancing loss as the secondary goal, and the host computer switches to the fastest balancing rate as the secondary goal. S67. During the inter-group balancing process, the constraints of minimizing the total distance of the balancing current loop and maximizing the total balancing power are also followed to complete the inter-group energy transfer; S68. Regularly generate abnormal cell reports, record the location information, operating status parameters and balancing history data of all potential abnormal cells and groups of cells, and prompt maintenance personnel to check and replace them.

[0031] S7. According to the equalization execution instruction, control the equalization execution module to complete the energy transfer from the high-voltage group cell to the low-voltage group cell; S8. Mark the cells that have been assigned to the low-voltage group more than m times consecutively as abnormal, and implement a differentiated equalization strategy that increases the upper limit of the equalization voltage threshold and the actual voltage conversion.

[0032] For example, this specifically includes: increasing the upper limit of the equalization voltage threshold of the single cell by 10mV, and calculating its actual voltage according to a preset ratio during the subsequent equalization process, such as taking 99% of the actual value, that is, calculating 3V voltage as 2.97V.

[0033] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0034] Please see Figure 2 The diagram shows a BMS proactive advance balancing system based on time series prediction and intelligent planning. The system includes: The data acquisition module is used to collect historical operating data of each cell in the battery system, including voltage, SOC, SOH, temperature and charge / discharge current, and to perform noise reduction and normalization preprocessing on the historical operating data. The time-series prediction module is used to input preprocessed historical operating data into the time-series prediction model to predict the operating status of each cell after a preset time, and obtain the state prediction value of each cell, including voltage prediction value, SOC prediction value and SOH prediction value. The fault diagnosis module is used to diagnose and provide early warning of abnormalities for individual cells based on the predicted status values ​​and the current operating status of each cell, and to isolate abnormal cells. The cell grouping module is used to sort all normal cells in descending order of their predicted voltage values, dividing the top 20% into the high-voltage group, the bottom 20% into the low-voltage group, and the middle 60% into the normal group. The voltage calculation module is used to calculate the average voltage of the three groups respectively, to obtain the average voltage of the high voltage group, the average voltage of the normal group, and the average voltage of the low voltage group. It calculates the difference between the average voltage of the high voltage group and the average voltage of the normal group, which is recorded as the first difference, and the difference between the average voltage of the normal group and the average voltage of the low voltage group, which is recorded as the second difference. The equalization decision module is used to initiate a global equalization process if the first difference or the second difference is greater than the preset equalization threshold. With the goal of maximizing the total available capacity of the battery system, it performs global equalization path planning by combining the constraints of minimizing loop distance and maximizing total equalization power, and generates equalization execution instructions. The equalization execution module is used to control the equalization execution module to complete the energy transfer from high-voltage group cell to low-voltage group cell according to the equalization execution instruction. The abnormal cell management module is used to mark cells that have been assigned to the low-voltage group more than m times consecutively as abnormal, and to implement a differentiated equalization strategy that increases the upper limit of the equalization voltage threshold and the actual voltage conversion.

[0035] In this embodiment of the invention, the balancing execution module includes multiple balancing sub-modules and bidirectional balancing connection lines, such as... Figure 3 As shown, each balancing submodule is responsible for balancing five individual battery cells, and adjacent balancing submodules are connected by a bidirectional balancing connection line. Each equalization submodule specifically includes: The equalization circuit, including a wide-input-range bidirectional DC-DC converter circuit and its control circuit, is used to realize energy transfer between individual cells during the equalization process. The first circuit selection switch includes a 5-channel cell selection switch array. One end can individually select the positive and negative terminals of any one of the 5 series-connected cells. The other end is connected to one end of the equalization circuit to connect the high-voltage cell that needs to output energy to the low-voltage cell that needs to input energy. The second-path selection switch circuit includes a 5-channel cell selection switch array. One end can individually select the positive and negative terminals of any single cell among the 5 cells connected in series; the other end is connected to the other end of the equalization circuit, and together with the first-path selection switch circuit, they form a complete equalization loop. A bidirectional balancing connection line, which is a bidirectional switch, is provided. One end of the bidirectional switch is connected to the other end of the balancing circuit, and the other end is connected to the bidirectional switch of the adjacent balancing submodule, so as to realize energy interaction between modules. The drive control unit, connected to the equalization decision module, is used to receive equalization execution instructions sent by the equalization decision module, and control the on / off state of the first path selection switch circuit, the second path selection switch circuit and the bidirectional switch according to the instructions, so as to realize energy transfer between any two battery cells or between a battery cell and an adjacent equalization submodule.

[0036] In this embodiment of the invention, the abnormal cell management module specifically includes: The statistics unit is used to count the number of times each cell is assigned to the low-voltage group during multiple consecutive equalization processes, and the number of times it is marked as a low-voltage cell within the group during the equalization process within the group. An anomaly marking unit, connected to the statistics unit, is used to mark a cell as a potential abnormal cell when a cell is assigned to a low-voltage group more than 3 times or is marked as a low-voltage cell in a group more than m times; and to mark a cell group as a potential abnormal cell group when the proportion of potential abnormal cells in a cell group exceeds 10%. The differentiated strategy generation unit, connected to the anomaly marking unit, is used to generate a differentiated balancing strategy for battery cells marked as potentially abnormal. Specifically, it includes: increasing the upper limit of the balancing voltage threshold of the battery cell by 10mV, and calculating its actual voltage according to a preset ratio during the subsequent balancing process, so that the balancing decision module can give priority to energy compensation when planning. The report generation unit, connected to the anomaly marking unit, is used to periodically generate abnormal cell reports. The report content includes the location information, operating status parameters, historical balancing data, and anomaly level of potential abnormal cells and groups of cells. The reporting cycle can be set to daily, weekly, or monthly by the host computer.

[0037] In this embodiment of the invention, the system further includes a host computer module; The host computer module is connected to the equalization decision module, fault diagnosis module, and abnormal cell management module via a CAN bus, and is used for: Send the equalization task instruction to the equalization decision module and set the equalization-related parameters, including the preset equalization threshold G, the preset convergence threshold D, the preset deviation threshold, the preset upper limit of charging voltage Vh, the preset lower limit of discharging voltage Vl, the preset normal range of SOC, the preset lower limit of SOH, the first acceptable range of voltage deviation within the group V0±D0, the second acceptable range of voltage deviation between groups Vg±Dg, and the equalization threshold Fset between groups. It receives alarm information at all levels and abnormal cell location information sent by the fault diagnosis module, displays them visually on the screen, and alerts maintenance personnel through an audible and visual alarm device; Receive abnormal cell reports sent by the abnormal cell management module, and manage the storage, query, and printing of the reports; The system monitors the operating status and balancing process of the entire battery system in real time, displaying the real-time voltage, SOC, SOH, and temperature parameters of each cell, as well as the balancing current, balancing power, and energy transfer during the balancing process. It supports historical data query function, allowing users to query cell operation data, balance history records and abnormal alarm records for any time period, and generate corresponding trend analysis charts.

[0038] The above description of the disclosed embodiments enables those skilled in the art to make or use the 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 invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A proactive advance equilibrium method for BMS based on time-series forecasting and intelligent planning, characterized in that, The method includes the following steps: S1. Collect historical operating data of each cell in the battery system, including voltage, SOC, SOH, temperature and charge / discharge current, and perform noise reduction and normalization preprocessing on the historical operating data; S2. Input the preprocessed historical operating data into the time series prediction model to predict the operating status of each cell after a preset time, and obtain the state prediction value of each cell, including voltage prediction value, SOC prediction value and SOH prediction value. S3. Based on the predicted status value and the current operating status of each battery cell, perform fault diagnosis and anomaly warning for each battery cell, and isolate abnormal battery cells; S4. Sort all normal battery cells in descending order of their predicted voltage values, and divide the top 20% into the high-voltage group, the bottom 20% into the low-voltage group, and the middle 60% into the normal group. S5. Calculate the average voltage of the three groups respectively to obtain the average voltage of the high voltage group, the average voltage of the normal group and the average voltage of the low voltage group. Calculate the difference between the average voltage of the high voltage group and the average voltage of the normal group, and record it as the first difference. The difference between the average voltage of the normal group and the average voltage of the low voltage group is recorded as the second difference. S6. If the first difference or the second difference is greater than the preset equalization threshold, the global equalization process is started. With the goal of maximizing the total available capacity of the battery system, the global equalization path is planned in combination with the constraints of minimizing the loop distance and maximizing the total equalization power, and equalization execution instructions are generated. S7. According to the equalization execution instruction, control the equalization execution module to complete the energy transfer from the high-voltage group cell to the low-voltage group cell; S8. Mark the cells that have been assigned to the low-voltage group more than m times consecutively as abnormal, and implement a differentiated equalization strategy that increases the upper limit of the equalization voltage threshold and the actual voltage conversion.

2. The BMS proactive advance equilibrium method based on time series prediction and intelligent planning according to claim 1, characterized in that, In step S2, the temporal prediction model is a BiLSTM-MultiHeadAttention model that combines a bidirectional long short-term memory network with a multi-head self-attention mechanism. Its construction and training process specifically includes: S21. Construct a BiLSTM-MultiHeadAttention model, which includes an input layer, a bidirectional LSTM layer, a multi-head self-attention layer, a fully connected layer, and an output layer in sequence. S22. Divide the preprocessed historical running data into training set, validation set and test set in chronological order, and generate training samples using the sliding window method; S23. Train the BiLSTM-MultiHeadAttention model using the training set, using the mean squared error (MSE) as the loss function, and use the Adam optimizer to update the parameters; S24. Use the validation set to validate the trained model, monitor the changes in the validation set loss function value, and stop training and save the optimal model parameters when the conditions for stopping training are met. S25. Test the saved optimal model using the test set. When the average absolute error of voltage prediction in the test set is less than the preset voltage value, the average absolute error of SOC prediction is less than the first preset percentage, and the average absolute error of SOH prediction is less than the second preset percentage, the model training is deemed complete.

3. The BMS proactive advance equilibrium method based on time series prediction and intelligent planning according to claim 1, characterized in that, The specific steps of step S3 include: S31. Real-time acquisition of the predicted voltage, SOC, and SOH values ​​of each individual cell, as well as the current voltage, current SOC, current SOH, and current charge / discharge status. S32. If the predicted voltage value of a certain cell is higher than the preset upper limit of charging voltage Vh when charging, or lower than the preset lower limit of discharging voltage Vl when discharging, the cell is determined to be a seriously abnormal cell, and an instruction to disconnect the battery cluster where the cell is located is immediately generated, and the location information and abnormality type of the cell are recorded in the system log. S33. Calculate the deviation rate between the current voltage of a certain battery cell and the predicted voltage value at the previous moment. The formula for calculating the deviation rate is: Deviation rate = |Current voltage - Previous predicted voltage value| / Previous predicted voltage value × 100% If the deviation rate exceeds 20% of the preset deviation threshold, the cell is determined to be a voltage abnormal cell, a first-level alarm command is immediately generated, and its location information and deviation rate value are recorded. S34. If the predicted SOC value of a certain cell exceeds the preset normal SOC range, or the predicted SOH value is lower than the preset lower limit of SOH, or the deviation between the current SOC and the predicted SOC value of the previous moment exceeds 5%, or the deviation between the current SOH and the predicted SOH value of the previous moment exceeds 2%, then the cell is determined to be an abnormal cell, a level 2 alarm command is immediately generated, and its location information and abnormal state parameters are recorded. S35. Based on the generated abnormality handling instructions, control the main contactor of the battery management system to disconnect the battery cluster containing the corresponding abnormal cell, and at the same time send the alarm information and abnormal cell location information to the host computer module for display.

4. The BMS proactive advance equilibrium method based on time-series prediction and intelligent planning according to claim 1, characterized in that, In step S6, the constraints of the global equilibrium path planning include: C1. During the balancing process, the total distance of the balancing current loop between all high-voltage group cell units and low-voltage group cell units participating in the balancing is minimized. The total distance of the balancing current loop is the sum of the physical distances between each pair of balancing cell units. C2. During the balancing process, the total balancing power of all high-voltage group cells and low-voltage group cells participating in the balancing process is maximized within the next minute. The total balancing power is the sum of the balancing power of each balancing cell pair. C3. During the balancing process, the real-time voltage of any single cell in the high-voltage group must not be lower than the real-time average voltage of the low-voltage group; C4. During the balancing process, the real-time voltage of any single cell in the low-voltage group must not be higher than the real-time average voltage of the high-voltage group.

5. The BMS proactive advance equilibrium method based on time series prediction and intelligent planning according to claim 4, characterized in that, In step S6, the convergence conditions for the global equilibrium path planning include: When the first difference is greater than the second difference, continue to perform the balancing operation until one of the following conditions is met: T1. The difference between the real-time average voltage of the high-voltage group and the real-time average voltage of the low-voltage group is less than the preset convergence threshold D; T2. The real-time average voltage of the high-voltage group is lower than the real-time average voltage of the normal group; T3. The real-time average voltage of the low-voltage group reaches the sum of the real-time average voltage of the normal group and 50% of the initial first difference; When the first difference is less than or equal to the second difference, gradually increase the number of individual cells in the high-voltage group, increasing the total number of normal cells by 1% each time, and recalculate the new average voltage of the high-voltage group and the first difference, until the new first difference satisfies the following: The new first difference × (1 + N × 5%) > the second difference Where N represents the number of times the high-voltage group is reselected, and then the equilibration operation is performed according to the constraints and convergence conditions.

6. The BMS proactive advance equilibrium method based on time-series prediction and intelligent planning according to claim 5, characterized in that, In step S6, when the number of individual battery cells in the battery system exceeds 1000, a step-by-step balancing method is used to perform the balancing process, which specifically includes: S61. Divide the individual cells of the entire battery system into several cell groups according to their physical location. Each cell group contains a fixed number of individual cells. Each battery cluster has a total of M cell groups and each cell group has a total of K individual cells. S62. Perform a balancing process once within each cell group. Based on the predicted state value of each cell, perform balancing planning within the group. The goal is to make the voltage deviation of the K cells in the group within the first acceptable range V0±D0. The default is to minimize balancing loss as the secondary goal. The host computer can switch to make the fastest balancing rate as the secondary goal. S63. During the group balancing process, the constraints of minimizing the total distance of the balancing current loop and maximizing the total balancing power are followed. The 20% of cells with the lowest voltage in the group are marked as low-voltage cells in the group, and their actual voltage is converted according to a preset ratio. If a cell is marked as a low-voltage cell in the group more than 3 times in a row, the cell is recorded as a potential abnormal cell. S64. If the proportion of potentially abnormal cells in a certain cell group exceeds 10%, then the entire cell group is marked as a potentially abnormal cell group, and a level 3 alarm command is generated and sent to the host computer module. S65. After completing one equalization of all cell groups, calculate the available power C of each cell group and the variance F of the available power of all cell groups based on the state prediction value of each cell group, and sort the M cell groups in descending order of available power. S66. If the variance F of the available power is greater than the preset inter-group balancing threshold Fset, then the secondary balancing process is started, and inter-group balancing planning is carried out on a cell-by-cell basis. The goal is to make the average voltage deviation of the M cell groups within the second acceptable range Vg±Dg. The default is to minimize the balancing loss as the secondary goal, and the host computer switches to the fastest balancing rate as the secondary goal. S67. During the inter-group balancing process, the constraints of minimizing the total distance of the balancing current loop and maximizing the total balancing power are also followed to complete the inter-group energy transfer; S68. Regularly generate abnormal cell reports, record the location information, operating status parameters and balancing history data of all potential abnormal cells and groups of cells, and prompt maintenance personnel to check and replace them.

7. A BMS proactive lead balancing system based on time-series prediction and intelligent planning, characterized in that, The system is used to implement the method as described in any one of claims 1-6, the system comprising: The data acquisition module is used to collect historical operating data of each cell in the battery system, including voltage, SOC, SOH, temperature and charge / discharge current, and to perform noise reduction and normalization preprocessing on the historical operating data. The time-series prediction module is used to input preprocessed historical operating data into the time-series prediction model to predict the operating status of each cell after a preset time, and obtain the state prediction value of each cell, including voltage prediction value, SOC prediction value and SOH prediction value. The fault diagnosis module is used to diagnose and provide early warning of abnormalities for individual cells based on the predicted status values ​​and the current operating status of each cell, and to isolate abnormal cells. The cell grouping module is used to sort all normal cells in descending order of their predicted voltage values, dividing the top 20% into the high-voltage group, the bottom 20% into the low-voltage group, and the middle 60% into the normal group. The voltage calculation module is used to calculate the average voltage of the three groups respectively, to obtain the average voltage of the high voltage group, the average voltage of the normal group, and the average voltage of the low voltage group. It calculates the difference between the average voltage of the high voltage group and the average voltage of the normal group, which is recorded as the first difference, and the difference between the average voltage of the normal group and the average voltage of the low voltage group, which is recorded as the second difference. The equalization decision module is used to initiate a global equalization process if the first difference or the second difference is greater than the preset equalization threshold. With the goal of maximizing the total available capacity of the battery system, it performs global equalization path planning by combining the constraints of minimizing loop distance and maximizing total equalization power, and generates equalization execution instructions. The equalization execution module is used to control the equalization execution module to complete the energy transfer from high-voltage group cell to low-voltage group cell according to the equalization execution instruction. The abnormal cell management module is used to mark cells that have been assigned to the low-voltage group more than m times consecutively as abnormal, and to implement a differentiated equalization strategy that increases the upper limit of the equalization voltage threshold and the actual voltage conversion.

8. The BMS active advance equalization system based on time-series prediction and intelligent planning according to claim 7, characterized in that, The equalization execution module includes multiple equalization sub-modules and bidirectional equalization connection lines. Each equalization sub-module is responsible for the equalization of 5 individual cells, and adjacent equalization sub-modules are connected by a bidirectional equalization connection line. Each equalization submodule specifically includes: The equalization circuit, including a wide-input-range bidirectional DC-DC converter circuit and its control circuit, is used to realize energy transfer between individual cells during the equalization process. The first circuit selection switch includes a 5-channel cell selection switch array. One end can individually select the positive and negative terminals of any one of the 5 series-connected cells. The other end is connected to one end of the equalization circuit to connect the high-voltage cell that needs to output energy to the low-voltage cell that needs to input energy. The second-path selection switch circuit includes a 5-channel cell selection switch array. One end can individually select the positive and negative terminals of any single cell among the 5 cells connected in series; the other end is connected to the other end of the equalization circuit, and together with the first-path selection switch circuit, they form a complete equalization loop. A bidirectional balancing connection line, which is a bidirectional switch, is provided. One end of the bidirectional switch is connected to the other end of the balancing circuit, and the other end is connected to the bidirectional switch of the adjacent balancing submodule, so as to realize energy interaction between modules. The drive control unit, connected to the equalization decision module, is used to receive equalization execution instructions sent by the equalization decision module, and control the on / off state of the first path selection switch circuit, the second path selection switch circuit and the bidirectional switch according to the instructions, so as to realize energy transfer between any two battery cells or between a battery cell and an adjacent equalization submodule.

9. The BMS active advance equalization system based on time series prediction and intelligent planning according to claim 7, characterized in that, The abnormal battery cell management module specifically includes: The statistics unit is used to count the number of times each cell is assigned to the low-voltage group during multiple consecutive equalization processes, and the number of times it is marked as a low-voltage cell within the group during the equalization process within the group. An anomaly marking unit, connected to the statistics unit, is used to mark a cell as a potential abnormal cell when a cell is assigned to a low-voltage group more than 3 times or is marked as a low-voltage cell in a group more than m times; and to mark a cell group as a potential abnormal cell group when the proportion of potential abnormal cells in a cell group exceeds 10%. The differentiated strategy generation unit, connected to the anomaly marking unit, is used to generate a differentiated balancing strategy for battery cells marked as potentially abnormal. Specifically, it includes: increasing the upper limit of the balancing voltage threshold of the battery cell by 10mV, and calculating its actual voltage according to a preset ratio during the subsequent balancing process, so that the balancing decision module can give priority to energy compensation when planning. The report generation unit, connected to the anomaly marking unit, is used to periodically generate abnormal cell reports. The report content includes the location information, operating status parameters, historical balancing data, and anomaly level of potential abnormal cells and groups of cells. The reporting cycle can be set to daily, weekly, or monthly by the host computer.

10. The BMS active advance equalization system based on time-series prediction and intelligent planning according to claim 7, characterized in that, The system also includes a host computer module; The host computer module is connected to the equalization decision module, fault diagnosis module, and abnormal cell management module via a CAN bus, and is used for: Send the equalization task instruction to the equalization decision module and set the equalization-related parameters, including the preset equalization threshold G, the preset convergence threshold D, the preset deviation threshold, the preset upper limit of charging voltage Vh, the preset lower limit of discharging voltage Vl, the preset normal range of SOC, the preset lower limit of SOH, the first acceptable range of voltage deviation within the group V0±D0, the second acceptable range of voltage deviation between groups Vg±Dg, and the equalization threshold Fset between groups. It receives alarm information at all levels and abnormal cell location information sent by the fault diagnosis module, displays them visually on the screen, and alerts maintenance personnel through an audible and visual alarm device; Receive abnormal cell reports sent by the abnormal cell management module, and manage the storage, query, and printing of the reports; The system monitors the operating status and balancing process of the entire battery system in real time, displaying the real-time voltage, SOC, SOH, and temperature parameters of each cell, as well as the balancing current, balancing power, and energy transfer during the balancing process. It supports historical data query function, allowing users to query cell operation data, balance history records and abnormal alarm records for any time period, and generate corresponding trend analysis charts.

Images