Battery management system passive equalization control method and system based on big data

Abstract

The application discloses a battery management system passive equalization control and system based on big data, which is applied to an energy storage battery system comprising a plurality of battery groups. The method comprises the following steps: collecting historical operation data of each battery monomer, generating or updating an equalization start pressure difference threshold value and an equalization stop pressure difference threshold value which are used by each battery group; collecting real-time voltages of each battery monomer, when the pressure difference between any two battery monomers in a battery group is greater than the start threshold value, determining the battery monomer with a higher voltage as a to-be-equalized battery monomer and starting the passive equalization circuit thereof; when the pressure difference between the battery monomer in the passive equalization and each battery monomer in the same group is less than the stop threshold value, stopping the passive equalization circuit thereof. The application can dynamically adjust the equalization threshold value according to the battery consistency and aging characteristics, improve the accuracy of equalization object identification and stop control, and reduce under-equalization, over-equalization and invalid energy loss.

Description

Technical Field

[0001] This invention relates to the field of battery equalization management technology, and in particular to a passive equalization control method and system for a battery management system based on big data. Background Technology

[0002] With the development of new energy power generation and energy storage technologies, the capacity and number of battery cells in large-scale energy storage systems such as energy storage power stations are constantly increasing. Energy storage battery systems typically consist of multiple battery packs, each of which includes multiple battery cells (i.e., battery cells) connected in series or parallel.

[0003] Due to factors such as manufacturing processes, material properties, and usage environment, different battery cells vary in capacity, internal resistance, self-discharge rate, and aging rate. After long-term charge-discharge operation, the voltage difference between battery cells will gradually increase, thereby limiting the usable capacity of the battery pack, shortening its service life, and in severe cases, potentially causing safety problems such as overcharging and over-discharging.

[0004] To improve the consistency between individual battery cells, existing battery management systems typically equip each cell with a passive equalization circuit consisting of an equalization switch and a discharge resistor. When the voltage difference between cells in the battery pack exceeds a preset start threshold, the cell with the higher voltage is discharged; when the voltage difference decreases to a preset stop threshold, the discharge stops.

[0005] However, most existing passive balancing strategies rely solely on currently collected cell voltages and pre-set fixed thresholds for control, failing to fully utilize historical data such as voltage, current, temperature, health status, and balancing records accumulated by individual cells during long-term operation. As battery aging and operating conditions change, fixed balancing thresholds struggle to continuously adapt to battery characteristics, easily leading to excessively frequent balancing starts, inappropriate stopping times, insufficient balancing, or over-balancing. This not only affects the balancing effect but also increases energy loss and the thermal load on the balancing circuit.

[0006] Therefore, it is necessary to improve the passive balancing control method of existing energy storage battery systems. Summary of the Invention

[0007] The purpose of this invention is to provide a passive equalization control method and system for a battery management system based on big data, which can dynamically update the threshold conditions for passive equalization based on the historical operating data of the battery.

[0008] To achieve the above objectives, this invention provides a passive balancing control method for a battery management system based on big data, applied to an energy storage battery system. The energy storage battery system includes several battery packs, each battery pack containing several individual battery cells, and each individual battery cell equipped with a passive balancing circuit. The control method includes: Collect and store historical operating data of each of the battery cells to obtain a historical dataset; The equalization parameters of the energy storage battery system are generated or updated based on the historical dataset. The equalization parameters include at least the equalization start-up differential pressure threshold and the equalization stop-down differential pressure threshold. The real-time voltage of each battery cell is collected, and for any battery pack, when the real-time voltage difference between any two battery cells in the battery pack is greater than the equalization start voltage difference threshold, the battery cell with the relatively higher voltage among the two battery cells corresponding to the real-time voltage difference is identified as the battery cell to be equalized, and the passive equalization circuit configured for the battery cell to be equalized is activated. For any battery cell in passive balancing, when the real-time voltage difference between the battery cell and the other battery cells in the battery pack is less than the balancing stop voltage difference threshold, the passive balancing circuit configured for the battery cell is controlled to stop.

[0009] Preferably, the historical operating data includes at least one of the following: historical voltage data, historical current data, historical temperature data, historical state of charge data, historical health status data, and historical balancing records for each of the battery cells.

[0010] Preferably, the method for generating the balancing parameters based on the historical operating data includes: The historical operational data is cleaned, and the data cleaning includes at least one of noise reduction and interpolation; Based on the historical operating data after cleaning, the consistency characteristics of each battery cell are extracted. The consistency characteristics include at least one of voltage fluctuation characteristics, capacity decay characteristics, and internal resistance growth characteristics. The consistency characteristic is input into the big data analysis model, and the big data analysis model outputs the equilibrium start pressure difference threshold and the equilibrium stop pressure difference threshold.

[0011] Preferably, the energy storage battery system further includes an energy management server and a battery management controller communicatively connected to the energy management server; The energy management server collects and stores historical operating data of each battery cell, generates or updates the balancing parameters, and sends the balancing parameters to the battery management controller via a communication link. The battery management controller performs state data acquisition and passive balancing control of the individual battery cells.

[0012] Preferably, the balancing parameter further includes a priority coefficient corresponding to each of the battery cells, the priority coefficient being used to characterize the balancing priority of the corresponding battery cell; If the number of battery cells to be balanced exceeds a preset threshold, the priority order for balancing is determined according to the priority coefficient corresponding to each battery cell to be balanced.

[0013] Preferably, the method for generating the priority coefficient includes: Based on the historical operating data, the voltage fluctuation characteristics, health status degradation, and internal resistance increase of each battery cell are obtained. The priority coefficient of the battery cell is obtained by weighted summing of the voltage fluctuation characteristic, the health state decay, and the internal resistance increase.

[0014] Preferably, the balancing parameters further include a pre-balancing trigger identifier, which is determined based on the consistency change trend of the battery cells reflected by the historical operating data. The method further includes: When the real-time voltage difference is less than the equalization start voltage difference threshold, the passive equalization circuit configured in one of the two battery cells corresponding to the real-time voltage difference and bearing the pre-equalization trigger mark is activated.

[0015] Preferably, the pre-equalization trigger flag is dynamically generated in the following manner: Based on the historical operating data, predict the voltage change trend of each battery cell within a preset time period in the future; Based on the voltage change trend, the pre-equalization trigger flag is configured for battery cells whose voltage deviation meets the preset divergence condition within the predicted future preset time period.

[0016] Preferably, the equalization parameters also include the maximum equalization duration; The method further includes: For any battery cell that is discharging, the duration of continuous discharge of the passive balancing circuit configured for that battery cell is recorded. When the continuous discharge duration reaches the maximum equalization duration, the passive equalization circuit configured for the battery cell stops discharging.

[0017] Preferably, the maximum equilibrium duration is generated in the following manner: Based on the discharge duration, temperature change data during discharge, and ambient temperature corresponding to each equalization discharge in the historical operation data, a temperature rise prediction model is constructed. The temperature rise prediction model is used to predict the temperature rise of a single battery cell under continuous equalization discharge. Based on the temperature rise predicted by the temperature rise prediction model and the upper limit of the safe temperature, the longest continuous discharge time that ensures the temperature of a single battery cell does not exceed the upper limit of the safe temperature is determined as the maximum equalization time.

[0018] Preferably, the balancing parameters also include a manually balanced recommendation list; For each of the battery cells, the battery cells to be manually balanced are identified based on the voltage, internal resistance growth rate and health status indicators of the battery cells in the historical operating data, so as to generate the manual balancing recommendation list. The manually balanced recommendation list is displayed on the human-computer interaction interface; The system receives a manual balancing command input through the human-computer interaction interface, wherein the manual balancing command specifies at least one battery cell to be manually balanced from the manual balancing recommendation list. The corresponding passive equalization circuit is activated according to the manual equalization command.

[0019] Preferably, for a battery cell to be manually balanced, a recommended balancing time and a recommended target voltage are calculated based on the voltage distribution characteristics and capacity decay trajectory of the battery cell, and the recommended balancing time and recommended target voltage are associated with the manual balancing recommendation list; and or, for a battery cell to be manually balanced, a recommended balancing current level is determined based on the internal resistance and capacity decay degree of the battery cell in the historical operating data, and the recommended balancing current level is associated with the manual balancing recommendation list.

[0020] Preferably, before controlling the corresponding passive equalization circuit to start discharging according to the manual equalization command, a safety verification method is also included: Determine whether the battery cell to be manually balanced is in a fault state, or whether there is an unrecoverable alarm, or whether the predicted maximum temperature of the passive balancing circuit during the execution of the manual balancing command exceeds the upper limit of the safe temperature. If so, refuse to execute the manual balancing command and maintain the current operating mode.

[0021] Preferably, before receiving the manual equalization command input through the human-computer interaction interface, the method further includes: Determine if there is an ongoing automatic load balancing task; If so, save the current state of the automatic balancing task and then pause the automatic balancing task. The on-site status includes the identification of the battery cell undergoing equalization, the accumulated equalization duration, and the current operating parameters of the equalization circuit.

[0022] Preferably, after controlling the corresponding passive equalization circuit to start discharging according to the manual equalization command, the method further includes: During the execution of the manual balancing command, the voltage of the battery cell to be manually balanced, the discharge current of the corresponding passive balancing circuit, and the temperature of the balancing resistor are monitored in real time. The corresponding passive equalization circuit is disconnected when one of the following conditions is met: The voltage of the battery cell to be manually balanced reaches the target voltage; The accumulated balancing time has reached the preset balancing time; The temperature of the battery cell to be manually balanced or the temperature of the balancing resistor exceeds a preset safety threshold. The discharge current exceeds the preset safety range; An emergency stop command has been received.

[0023] Preferably, after disconnecting the corresponding passive equalization circuit, the method further includes: Entering the recovery transition period, during which any automatic balancing tasks are prohibited from being started; The duration of the recovery transition period is configured by the balancing parameters; After the recovery transition period ends, the automatic balancing ready state is restored, and the start-up conditions for automatic balancing are reassessed based on the current state of the individual battery cells.

[0024] The present invention also provides a passive balancing control system for a battery management system, wherein the control system performs passive balancing on individual battery cells based on the passive balancing control method for a battery management system described above.

[0025] The present invention also provides a passive balancing control system for a battery management system, comprising: One or more processors; Memory; And one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the programs including instructions for performing the passive equalization control method of the battery management system as described above.

[0026] The present invention also provides a computer-readable storage medium comprising a computer program that can be executed by a processor to perform the passive equalization control method for a battery management system as described above.

[0027] Compared with existing technologies, the passive equalization control method provided by the above technical solution dynamically generates or updates the equalization start-up voltage difference threshold and equalization stop voltage difference threshold, which are used by all battery packs in the system, based on the historical operating data of each battery cell in the energy storage battery system. This allows the equalization criterion to be adaptively adjusted according to changes in system consistency and aging characteristics, and ensures that different battery packs use a unified equalization control benchmark. At the same time, it identifies battery cells with relatively high voltages by identifying the voltage differences between each pair of cells in the group and implements passive equalization. It stops the equalization when the voltage difference between the battery cell and the other cells in the same group is lower than the stop threshold, thereby improving the accuracy of identifying the cells to be equalized and the equalization stop control, reducing under-equalization, over-equalization and ineffective energy loss caused by fixed thresholds, improving battery consistency and extending the service life of the energy storage battery system. Attached Figure Description

[0028] Figure 1 This is a diagram of the battery management system architecture in an embodiment of the present invention.

[0029] Figure 2 This is a flowchart of the passive equalization control method in an embodiment of the present invention.

[0030] Figure 3 This is a flowchart illustrating how the energy management server generates balancing parameters in an embodiment of the present invention.

[0031] Figure 4 This is a flowchart illustrating the manual balancing control and safety protection logic in an embodiment of the present invention. Detailed Implementation

[0032] To illustrate the technical content, structural features, objectives, and effects of the present invention in detail, the following description is provided in conjunction with the embodiments and accompanying drawings.

[0033] This embodiment discloses a passive balancing control method for a battery management system based on big data, applied to an energy storage battery system. This energy storage battery system can be a battery storage unit in an energy storage power station, or a system containing multi-level battery packs, such as an industrial or commercial energy storage cabinet or a backup power supply for a communication base station.

[0034] The energy storage battery system comprises several battery packs, each pack consisting of several battery cells connected in series. Each battery cell is equipped with a passive balancing circuit. The passive balancing circuit includes a balancing switch and a balancing resistor, connected in series and then in parallel across the corresponding battery cell. When the balancing switch is closed, the corresponding battery cell discharges through the balancing resistor. When the balancing switch is open, the corresponding battery cell stops discharging through the passive balancing circuit.

[0035] like Figure 2 The passive equalization control method in this embodiment includes the following steps: S11 collects and stores historical operating data for each battery cell.

[0036] During the operation of the energy storage battery system, the voltage of each battery cell is collected according to the set sampling period, and historical voltage data is stored using the battery pack identifier, battery cell identifier, and sampling time as indexes.

[0037] The sampling period can be configured in seconds, minutes, or dynamically according to the operating status. For example, during charging and discharging, data can be collected in periods of 1 to 10 seconds, while in a static state, data can be collected in periods of 1 to 5 minutes. For data with a high sampling frequency, minute-level averages, maximum values, minimum values, and standard deviations can be generated while retaining the original data to reduce the space occupied by long-term storage.

[0038] Historical operating data can be stored in a local battery management controller, energy management server, or remote data server. For energy storage battery systems with multiple battery packs, data indexes can be established according to the hierarchy of battery clusters, battery packs, and individual battery cells, so that each piece of historical operating data can correspond to the corresponding individual battery cell.

[0039] S12 generates or updates the equalization parameters corresponding to each battery cell based on historical operating data.

[0040] The balancing parameters include at least the balancing start differential pressure threshold and the balancing stop differential pressure threshold.

[0041] The equalization start-up differential pressure threshold is higher than the equalization stop differential pressure threshold to create a hysteresis range between equalization start-up and equalization stop, reducing frequent start-up and stop of the passive equalization circuit near the threshold. For example, analysis of historical data from the energy storage battery system over the past month shows that when the differential pressure between individual cells within the battery pack exceeds 30mV, the consistency differences tend to widen further; when the differential pressure between cells after equalization is less than 10mV, the improvement in differential pressure brought by continued equalization is relatively small. Therefore, the equalization start-up differential pressure threshold can be set to 30mV, and the equalization stop differential pressure threshold can be set to 10mV.

[0042] In this embodiment, a set of equalization start-up differential pressure thresholds and equalization stop differential pressure thresholds are generated for the entire energy storage battery system. All battery packs and all individual battery cells in the energy storage battery system share the same set of equalization start-up differential pressure thresholds and equalization stop differential pressure thresholds.

[0043] During subsequent parameter update cycles, the equalization execution results fed back by each battery pack can be added to the historical dataset, and the equalization start differential pressure threshold and equalization stop differential pressure threshold can be updated.

[0044] S13: Collect the real-time voltage of individual battery cells and calculate the real-time voltage difference within each battery pack.

[0045] For any given battery pack, the real-time voltage of each individual battery cell is collected.

[0046] Suppose a battery pack consists of n individual cells, with the real-time voltage of the i-th cell being Vi and the real-time voltage of the j-th cell being Vj. The real-time voltage difference between them can be expressed as: ΔVij=∣Vi Vj∣,i≠j.

[0047] To reduce redundant calculations, the highest and lowest real-time voltages within the battery pack can be determined first. If the difference between the highest and lowest real-time voltages is not greater than the equalization start-up voltage difference threshold, the start-up voltage difference comparison between other battery cells will not be performed.

[0048] S14, determine the battery cells to be balanced based on the equalization start-up differential pressure threshold.

[0049] For any given battery pack, when the real-time voltage difference between any two cells in the battery pack is greater than the equalization start-up voltage difference threshold, the cell with the relatively higher real-time voltage among the two cells corresponding to that real-time voltage difference is identified as the cell to be equalized.

[0050] For example, if the equalization start-up voltage difference threshold is 30mV, and the real-time voltages of battery cells B1 and B2 in a certain battery pack are 3.435V and 3.400V respectively, the real-time voltage difference between them is: ΔV12=|3.435 3.400∣=35mV.

[0051] Since 35mV is greater than 30mV, the battery cell B1 with the relatively high real-time voltage is identified as the battery cell to be balanced, and the passive balancing circuit configured for battery cell B1 is activated, so that battery cell B1 discharges through the corresponding balancing resistor.

[0052] If the same battery cell forms a real-time voltage difference greater than the equalization start-up voltage difference threshold with multiple other battery cells, then only that battery cell is identified as a single object to be equalized, and its passive equalization circuit is not repeatedly activated.

[0053] If multiple battery cells are identified as battery cells to be balanced, the corresponding passive balancing circuit can be activated within the range of the number of channels that the passive balancing circuit is allowed to conduct simultaneously.

[0054] S15, stop passive balancing according to the balancing stop pressure difference threshold.

[0055] After the passive balancing circuit is activated, it continues to collect the real-time voltage of each battery cell and calculates the real-time voltage difference between each battery cell in passive balancing and the other battery cells in its battery pack.

[0056] Let the i-th battery cell be the one undergoing passive balancing. The i-th battery cell is defined as follows: The i-th battery cell satisfies the following relationship with every other battery cell in its battery pack: |Vi Vj∣ <Vstop,j≠i Where Vstop is the equalization stop voltage threshold used by the entire energy storage battery system, then a disconnect signal is output to the equalization switch corresponding to the i-th battery cell to stop the passive equalization of that battery cell.

[0057] For example, the energy storage battery system includes three battery packs, each consisting of 16 battery cells connected in series.

[0058] A set of balancing parameters is generated based on the historical operating data of all 48 battery cells in the three battery packs. The balancing start-up differential pressure threshold is 30mV, and the balancing stop differential pressure threshold is 10mV.

[0059] If the maximum real-time voltage difference of the first battery pack is 36mV, the maximum real-time voltage difference of the second battery pack is 24mV, and the maximum real-time voltage difference of the third battery pack is 33mV, then the first and third battery packs determine the individual battery cells to be balanced based on the shared 30mV equalization start-up voltage difference threshold, and the second battery pack does not start passive equalization for the time being.

[0060] When the real-time voltage difference between a battery cell in the first battery pack that is in passive balancing and all the other battery cells in the first battery pack is less than 10mV, the passive balancing of that battery cell is stopped; the third battery pack independently determines its balancing stop condition according to the same rule.

[0061] By using the above method, the historical operating data of all battery cells in the energy storage battery system are used to dynamically generate the equalization start-up differential pressure threshold and equalization stop differential pressure threshold that are used by the entire energy storage battery system. This allows the thresholds to be updated according to the consistency changes and aging status of the energy storage battery system. At the same time, each battery pack independently performs real-time differential pressure judgment and passive equalization control under the unified threshold, thereby reducing over-equalization or under-equalization caused by using fixed thresholds.

[0062] On the other hand, historical operating data includes at least one of the following: historical voltage data, historical current data, historical temperature data, historical state of charge data, historical health status data, and historical balance records for each battery cell.

[0063] Historical current data can include the battery pack's charge and discharge currents and corresponding timestamps, and is linked to individual battery cells within the pack through battery pack topology. Historical temperature data can include the surface temperature of individual battery cells, the temperature at which the equalization resistance is measured, and the ambient temperature. Historical state of charge (SOC) data characterizes the remaining capacity of individual battery cells or the battery pack, while historical health status data characterizes the remaining capacity relative to the rated capacity, changes in internal resistance, or overall health. Historical equalization records can include equalization start time, equalization end time, target battery cell, starting voltage, ending voltage, discharge current, cumulative discharge energy, temperature changes during equalization, and reason for termination.

[0064] When analyzing historical voltage data, the data can be stratified according to historical current, state of charge, and temperature. For example, the voltage dispersion under resting, late-charge, and discharging conditions can be analyzed separately to avoid directly assuming transient voltage differences caused by load changes as consistent degradation. For the same battery cell, the changes in voltage and internal resistance over different time periods under the same state of charge and similar temperature conditions can be compared to identify characteristic drift caused by aging.

[0065] On the other hand, regarding historical operational data, such as Figure 3 In this embodiment, the data is first cleaned, then the consistency features are extracted, and the consistency features are input into the big data analysis model.

[0066] Data cleaning includes at least one of noise reduction and interpolation.

[0067] After cleaning, at least one of the following characteristics is extracted from historical operating data: voltage fluctuation, capacity decay, and internal resistance growth. Voltage fluctuation characteristics can be the voltage standard deviation, range, interquartile range, or average deviation relative to the median voltage of the battery pack within a set time window. Capacity decay characteristics can be the difference between the current available capacity and the initial available capacity, the capacity decrease rate per unit cycle, or the change in health status over a certain period. Internal resistance growth characteristics can be the difference between the current internal resistance and the reference internal resistance, or the internal resistance growth rate per unit time.

[0068] Big data analytics models can employ regression models, decision tree models, gradient boosting tree models, clustering models, neural network models, or time-series prediction models. The model's input consists of the consistency characteristics of each battery cell and optional operating condition labels; the output is the equalization start-up differential pressure threshold and the equalization stop-down differential pressure threshold.

[0069] For example, the analysis model determines the voltage fluctuation standard deviation of a certain battery cell to be 8mV, the capacity decay rate to be 0.6% per 100 cycles, and the internal resistance growth rate to be 1.2% per 100 cycles based on data from the past 30 days. It then outputs the equalization start-up voltage difference threshold as 28mV and the equalization stop voltage difference threshold as 9mV.

[0070] In another implementation, such as Figure 1 The energy storage battery system also includes an energy management server (EMS) and a battery management controller (BMS) that communicates with the energy management server.

[0071] Energy management servers can be EMS servers deployed locally at energy storage power stations, or cloud servers or edge computing servers.

[0072] The battery management controller includes a data acquisition module, a main control MCU, an equalization drive circuit, and a communication interface.

[0073] In a typical deployment, the energy storage battery system comprises multiple battery clusters, each configured with a battery management controller. All battery management controllers are connected to the energy management server via industrial Ethernet, CAN bus, or other communication links.

[0074] The energy management server collects and stores historical operating data of each battery cell, generates or updates balancing parameters, and sends the balancing parameters to the battery management controller via a communication link.

[0075] Specifically, each battery management controller periodically reports the collected battery cell status data via a communication link during operation. The energy management server writes the reported data into a database or data lake, forming a historical dataset for each battery cell since its commissioning. The energy management server periodically (e.g., when system load is low in the early morning each day) extracts data from the database, runs a parameter generation process, and packages the generated balancing parameters into data packets, which are then sent to the corresponding battery management controller via TCP protocol or CAN message. Upon receiving the data, the battery management controller stores the balancing parameters in its local non-volatile memory as the basis for subsequent balancing decisions.

[0076] The battery management controller performs state data acquisition and passive balancing control of individual battery cells. This means that real-time voltage acquisition, differential voltage comparison, and the opening and closing of the balancing switch are all completed locally by the battery management controller, without relying on real-time communication with an energy management server.

[0077] Since the generation of balancing parameters requires processing long-term, large-scale historical data, it has high requirements for storage capacity and computing power, but low requirements for real-time performance, making it suitable for the energy management server to handle. On the other hand, the start and stop judgment of balancing action depends on real-time voltage in the millisecond to second range, which has high requirements for real-time performance, but low calculation volume per operation, making it suitable for the battery management controller to execute locally.

[0078] In this embodiment, the architecture of server-side parameter analysis and generation, and controller-side local real-time execution, ensures the continuity of the balancing function. Even if the communication link between the energy management server and the battery management controller is briefly interrupted, the battery management controller can still continue to perform passive balancing based on the most recent balancing parameters stored locally. Simultaneously, the computational load of historical data analysis does not consume the battery management controller's resources, allowing the battery management controller to be implemented using a low-cost MCU, thus reducing system costs. Furthermore, the battery management controller sends the balancing execution results back to the energy management server. This returned data is incorporated into the historical dataset for the next round of parameter generation, forming a data closed loop that allows the balancing parameters to be continuously optimized as data accumulates.

[0079] On the other hand, the number of channels that can be simultaneously activated in a passive balancing circuit is usually limited in hardware: firstly, the driving capability of the balancing drive circuit is limited; secondly, the simultaneous discharge of multiple balancing resistors will cause concentrated heat generation. Due to heat dissipation constraints, the number of channels that can be balanced simultaneously is generally a portion of the total number of battery cells in the group. For example, for a 16-cell battery pack, the hardware can only allow 8 channels to be balanced simultaneously. Therefore, the balancing parameters also include a priority coefficient corresponding to each battery cell. The priority coefficient is used to characterize the balancing priority of the corresponding battery cell and is generated by the energy management server based on historical datasets and distributed along with the balancing parameters.

[0080] During operation, if the number of identified battery cells to be balanced exceeds the preset limit threshold, the priority order of balancing is determined according to the priority coefficient corresponding to each battery cell: battery cells with higher priority coefficients are given priority to obtain the balancing channel, while battery cells with lower priority coefficients enter the waiting queue and are started in sequence after the balancing task that has already started is completed and the channel is released.

[0081] The threshold is typically set based on the maximum number of simultaneous equalization channels allowed by the hardware, such as half the total number of battery cells in the group. It can also be dynamically lowered in high-temperature environments with poor heat dissipation. For example, if a battery pack identifies 10 battery cells to be equalized at a certain moment, and the threshold is 8, they are sorted from highest to lowest priority. The passive equalization circuit is activated for the top 8 battery cells, while the remaining 2 battery cells wait for the channel to be released.

[0082] In this embodiment, due to the introduction of a priority coefficient generated based on historical data, the allocation of balancing channels is no longer solely based on the magnitude of the real-time pressure difference, but rather takes into account the long-term performance of each battery cell. This allows battery cells with more severe consistency degradation and greater need for balancing intervention to obtain balancing resources first. Thus, under the condition of limited balancing channels, the contribution of unit balancing resources to improving battery pack consistency is increased, avoiding the situation where healthy cells occupy channels due to large instantaneous pressure differences, while truly lagging cells cannot obtain balancing for a long time.

[0083] Furthermore, the method for generating the priority coefficient includes: obtaining the voltage fluctuation characteristics, health status degradation, and internal resistance growth of each battery cell based on historical operating data. The priority coefficient of the battery cell is obtained by weighted summing of voltage fluctuation characteristics, health state decay, and internal resistance growth.

[0084] Expressed as a formula, the priority coefficient of the i-th battery cell It can be represented as:

[0085] in, This is the standard deviation of the voltage of the battery cell within a statistical time window (e.g., the past week), which is a characteristic quantity of voltage fluctuation. This represents the amount of health state degradation of the individual battery cell within the statistical time window. α represents the increase in internal resistance of the battery cell within the statistical time window; α, β, and γ are weighting coefficients. These three characteristics can be normalized before weighting to eliminate dimensional differences. The values ​​of the weighting coefficients can be configured according to the operational priorities of the energy storage battery system: for example, for systems where extending lifespan is the primary objective, the weight of β can be increased to give higher priority to battery cells with rapidly declining health; for systems where short-term consistency is the primary objective, the weight of α can be increased.

[0086] In this embodiment, since the priority coefficient integrates both current performance and aging trend, the balancing resources are tilted towards unhealthy battery cells. This achieves key care for lagging battery cells at the balancing execution level, slows down the widening gap between lagging battery cells and the main body of the group, and thus slows down the decay of the overall usable capacity of the battery pack.

[0087] On the other hand, in order to intervene earlier in the process of consistency deterioration, rather than waiting until the pressure difference has exceeded the equilibrium start pressure difference threshold before responding passively, the equilibrium parameters also include a pre-equilibrium trigger flag.

[0088] The pre-balancing trigger flag is determined based on the consistency trend of individual battery cells as reflected in historical operating data. Specifically, during parameter generation, the energy management server analyzes the consistency trend of each battery cell, assigns a pre-balancing trigger flag to battery cells identified as having a tendency to diverge, and sends a list of battery cells with pre-balancing trigger flags to the battery management controller along with the balancing parameters.

[0089] Based on this, at the battery management controller, when the real-time voltage difference is less than the equalization start voltage difference threshold, the passive equalization circuit configured in the one of the two battery cells with the pre-equalization trigger mark corresponding to the real-time voltage difference is activated.

[0090] In other words, for a battery cell with a pre-equalization trigger indicator, even if its real-time voltage difference with other battery cells in the group has not yet reached the equalization start voltage difference threshold, its passive equalization circuit is allowed to be activated in advance to perform a small-energy pre-equalization.

[0091] To control the energy input for pre-equilibration, stricter execution conditions can be attached to pre-equilibration, such as performing it only during system rest periods, limiting the duration of a single pre-equilibration to a portion of the regular equilibration duration, and stopping the pre-equilibration process once the pressure difference converges to the normal level within the group.

[0092] The degradation of battery cell consistency is a gradual process. When the self-discharge rate of a battery cell begins to be high or its capacity begins to decline rapidly, its voltage deviation is slight in the early stages and insufficient to trigger regular equalization. However, the deviation will continue to accumulate over time. If we wait until the deviation accumulates to exceed the equalization start-up voltage threshold before intervention, the required equalization discharge is already large, and the battery cell may have already undergone multiple charge-discharge cycles in a deviated state, accelerating aging. By pre-equalizing and pulling the voltage back with a small energy discharge in the early stages of deviation, we can suppress the accumulation of differentiation at a relatively low cost.

[0093] Specifically, the pre-balancing trigger flag is dynamically generated in the following way: S71: Based on historical operating data, predict the voltage change trend of each battery cell within a preset time period in the future. The prediction can be made using a Long Short-Term Memory (LSTM) network model or other intelligent models. The LSTM model is trained to learn the temporal evolution of voltage deviations by using the voltage deviation sequence of each battery cell under reference operating conditions (such as daily rest periods) over the past several cycles as input, and outputs the predicted voltage deviation sequence of each battery cell within a preset time period in the future (e.g., the next 24 hours or the next 72 hours).

[0094] S72: Based on the voltage change trend, configure a pre-equalization trigger flag for battery cells whose voltage deviation meets the preset divergence conditions within a preset time period in the future.

[0095] The preset divergence condition can be set as follows: the predicted voltage deviation exceeds the equalization start-up voltage difference threshold by a preset percentage (e.g., 70%) at the end of a preset time period in the future, or the predicted voltage deviation growth rate exceeds the median growth rate within the group by a preset multiple.

[0096] For example, if the equalization start-up voltage difference threshold is 30mV, and the current voltage deviation of a certain battery cell is 12mV, which has not yet met the standard, but the LSTM model predicts that its deviation will increase to 24mV in the next 24 hours (more than 70% of 30mV, i.e. 21mV), then it is determined that the battery cell meets the preset divergence condition and a pre-equalization trigger flag is configured for it.

[0097] The pre-balancing trigger flag is dynamically refreshed with each parameter update cycle. For battery cells whose divergence trend is eliminated after pre-balancing, if the preset divergence conditions are no longer met in the prediction of the next cycle, the pre-balancing trigger flag will be canceled to avoid unnecessary energy loss caused by continuous pre-balancing of battery cells that have recovered to normal.

[0098] On the other hand, considering that passive balancing is a physical process of heating up through resistive discharge, long-term continuous balancing will bring the risk of temperature rise in the balancing resistance and individual battery cells. The balancing parameters also include the maximum balancing time.

[0099] Based on this, the equalization control method also includes: for any battery cell that is discharging, counting the continuous discharge duration of the passive equalization circuit configured for that battery cell. When the continuous discharge duration reaches the maximum equalization duration, the passive equalization circuit configured for that battery cell stops discharging.

[0100] Specifically, the battery management controller maintains a timer for each passive balancing circuit in the discharge state. The timer starts when the balancing switch is closed and resets when the balancing switch is open. When the continuous discharge duration of a certain circuit reaches the maximum balancing duration, the balancing switch is forcibly disconnected even if the battery cell has not yet met the balancing stop condition determined by the voltage difference. After forced stop, a cooling interval can be set for the battery cell. When the cooling interval ends and the balancing start condition is still met, balancing is restarted, so that balancing is performed in an intermittent manner of discharge-cooling-discharge.

[0101] Furthermore, the maximum equilibrium duration is generated as follows: First, based on the discharge duration, temperature change data during discharge, and ambient temperature corresponding to each equalization discharge in historical operating data, a temperature rise prediction model is constructed. The temperature rise prediction model is used to predict the temperature rise of individual battery cells under continuous equalization discharge.

[0102] The temperature rise prediction model can employ a thermal network model, which equates the equalization resistance, individual battery cells, and their installation environment to a thermal network composed of thermal resistance and thermal capacity. The equalization discharge power serves as the heat source input, and the model outputs curves showing the temperature change of each node over the discharge duration. The thermal resistance and thermal capacity parameters in the thermal network model are identified from measured temperature rise data in historical equalization records. Specifically, the model uses the discharge duration, initial temperature, and ambient temperature of each equalization cycle as input, and the measured temperature change as the fitting target. Parameter identification ensures the model output matches the historical measured temperature rise. In some optional implementations, the temperature rise prediction model can also employ empirical formulas or machine learning regression models based on historical data regression, as long as it can predict the temperature rise based on the discharge duration and environmental conditions.

[0103] Then, based on the temperature rise predicted by the temperature rise prediction model and the upper limit of the safe temperature, the longest continuous discharge time that keeps the temperature of the battery cell below the upper limit of the safe temperature is determined as the maximum equalization time.

[0104] For example, if the upper limit of the safe temperature is 65℃ and the current ambient temperature is 25℃, and the temperature rise prediction model predicts that the temperature of the equalization resistor will rise according to the curve under continuous discharge conditions, reaching 65℃ after 42 minutes of continuous discharge, then the maximum equalization time will be determined to be 42 minutes, or a preset margin (e.g., 40 minutes) will be deducted from it. It can be seen that the maximum equalization time dynamically changes with the ambient temperature. In high-temperature environments in summer (e.g., ambient temperature 40℃), the time predicted by the same model to reach the upper limit of the safe temperature is shortened, and the maximum equalization time is automatically shortened accordingly. In low-temperature environments in winter, the time is appropriately relaxed to avoid unnecessarily interrupting the equalization process.

[0105] On the other hand, considering the need for manual balancing in scenarios such as system debugging, maintenance, specific operating condition testing, or emergency response, and the fact that maintenance personnel select balancing objects based on experience lacks data support, is prone to misjudgment, and poses security risks, this embodiment further provides a data-driven manual balancing mechanism. That is, the balancing parameters also include a recommended list of manual balancing parameters.

[0106] Based on this, such as Figure 4 On the energy management server, for each battery cell, the system identifies the cells that need to be manually balanced based on the voltage, internal resistance growth rate, and health status indicators of the battery cells in historical operating data, and generates a recommended list for manual balancing.

[0107] Specifically, during parameter generation, the energy management server filters out lagging battery cells with abnormally high voltage, excessively rapid internal resistance growth, or health status indicators significantly lower than the group's level, and adds them to the manual balancing recommendation list. The filtering can employ a threshold method (e.g., the internal resistance growth rate exceeding a preset multiple of the group's median value) or a clustering method (identifying battery cells deviating from the main cluster as lagging cells). The identification criteria can be any one of the three listed indicators or a combination thereof.

[0108] The recommended list for manual balancing is displayed on the human-machine interface. This interface can be the monitoring interface of the energy management server or a display interface on the maintenance terminal. The recommended list can be presented in tabular form, listing the identifier, current voltage, and voltage difference deviation of each battery cell to be manually balanced, for maintenance personnel to view and make judgments.

[0109] Receive manual balancing commands input through the human-computer interaction interface. The manual balancing command specifies at least one battery cell in the manual balancing recommendation list to be manually balanced.

[0110] After selecting one or more battery cells from the recommended list, maintenance personnel issue a command, which is transmitted to the corresponding battery management controller via the communication link. In some optional implementations, the manual balancing command can also be generated by a local button or local debugging interface of the battery management controller to cover on-site maintenance scenarios when the communication link is unavailable.

[0111] The manual balancing command activates the corresponding passive balancing circuit. The battery management controller parses the cell identifier specified in the command, closes the corresponding balancing switch, and performs manual balancing discharge.

[0112] Since the targets for manual balancing are provided by a recommended list based on historical operating data, the intervention of maintenance personnel changes from selection based on experience to selection based on data. This allows the manual intervention to accurately target the truly lagging battery cells, thereby reducing unnecessary losses caused by mistakenly discharging healthy cells, shortening the time required for maintenance personnel to identify lagging cells, and improving the efficiency and accuracy of maintenance operations. At the same time, the manual mechanism coexists with the aforementioned automatic mechanism, taking into account both the automated operation of the system and the human flexibility in maintenance scenarios.

[0113] Furthermore, to eliminate the need for maintenance personnel to set balancing execution parameters based on experience after selecting the balancing targets, this embodiment further associates recommended execution parameters with the manual balancing recommendation list. That is, in addition to the battery cell identifier, the recommendation list also associates recommended balancing duration, recommended target voltage, and recommended balancing current level.

[0114] For a single battery cell to be manually balanced, a recommended balancing time and a recommended target voltage are calculated based on the cell's voltage distribution characteristics and capacity decay trajectory. The recommended balancing time and the recommended target voltage are then associated with the recommended list for manual balancing.

[0115] Specifically, the recommended target voltage can be the median or mean of the voltage distribution within the group. The goal of balancing is to bring the voltage of the lagging battery cell back to the level of the main group.

[0116] The recommended equalization time is estimated based on the difference between the current voltage and the recommended target voltage of the battery cell, the equalization discharge current, and the actual capacity of the battery cell corrected by the capacity decay trajectory. For the same voltage drop, a battery cell with greater capacity decay requires less discharge, and the recommended equalization time is correspondingly shorter. For example, for a battery cell to be manually equalized, if it is calculated that discharging at the current equalization current for 5 minutes will bring its voltage drop back to the median value within the group, then the recommended equalization time is 5 minutes, and the recommended target voltage is the median voltage within the group.

[0117] In addition, for battery cells requiring manual equalization, a recommended equalization current level can be determined based on the cell's internal resistance and capacity degradation from historical operating data, and this recommended equalization current level can be added to the manual equalization recommendation list. The principle is that aged battery cells with significantly increased internal resistance and substantial capacity degradation experience more severe internal heat generation and polarization when subjected to high-current discharge. High-current equalization would cause further damage to them. Therefore, a lower current level is recommended for battery cells with deeper aging, while a higher current level can be recommended for battery cells in better condition to shorten equalization time.

[0118] After maintenance personnel select the battery cells to be manually balanced on the human-machine interface, the recommended balancing time, recommended target voltage, and recommended balancing current level are automatically filled in. These can be directly applied with a single click or manually modified based on the recommended values. Since the balancing targets, balancing time, target voltage, and current level are all calculated from historical data, all key parameters for manual balancing are data-driven. This avoids over-discharge and energy waste caused by excessively long experience-based balancing times, as well as accelerated damage caused by applying high currents to aging battery cells. Manual intervention is both more efficient and safer.

[0119] Furthermore, to prevent the manual balancing command from being executed when the system or individual battery cells are in an unsuitable state, this embodiment adds a safety verification step before the command is executed. That is, before controlling the corresponding passive balancing circuit to start discharging according to the manual balancing command, a safety verification method is also included.

[0120] Security checks include: See again Figure 4The system determines whether the battery cell to be manually balanced is in a fault state, whether there is an unrecoverable alarm, or whether the predicted maximum temperature of the passive balancing circuit during the execution of the manual balancing command exceeds the safe temperature limit; if so, it refuses to execute the manual balancing command and maintains the current operating mode.

[0121] Among them, the fault states of individual battery cells include overvoltage, undervoltage, overtemperature, sampling failure, etc. The voltage and temperature information of battery cells in a fault state are unreliable or their physical state is not suitable for discharge. Performing equalization on them may exacerbate the fault.

[0122] Unrecoverable alarms refer to critical system-level alarms (such as insulation faults, relay sticking, etc.). When such alarms exist, the system should remain in a protective state and should not introduce additional discharge operations.

[0123] The prediction of the highest temperature during the balancing period reuses the temperature rise prediction model of the above embodiment. The set duration and current level in the instruction, as well as the current ambient temperature and the current heat accumulation state of the balancing resistor, are substituted into the temperature rise prediction model to calculate the predicted highest temperature of the balancing resistor during the execution of the instruction. If the predicted highest temperature exceeds the upper limit of the safe temperature, it is determined that there is a thermal risk.

[0124] If any of the above three checks fails, the battery management controller will refuse to execute the manual balancing command, maintain the current operating mode (if automatic balancing was being performed before the rejection, it will continue automatic balancing), and can send a rejection reason code to the energy management server through the communication interface, such as "system abnormality" or "thermal risk", so that the operation and maintenance personnel can know the reason for the rejection and take appropriate action.

[0125] It is worth noting that the temperature rise verification occurs before the equalization switch is closed, which is a predictive verification. This, together with the real-time temperature monitoring during execution (see the embodiments described later), constitutes dual thermal protection of pre-emptive prediction and in-process monitoring. Because the battery cell status, system alarm status, and predicted temperature rise are verified before manual equalization, unsuitable commands are intercepted before discharge occurs. This eliminates risk scenarios such as discharging faulty cells, superimposed operations under severe system alarms, and foreseeable overheating of the equalization resistor, thereby improving system safety in scenarios requiring manual intervention.

[0126] Furthermore, to further resolve the conflict between manual balancing and the ongoing automatic balancing task, this embodiment introduces a context-saving mechanism during mode switching. That is, before receiving the manual balancing command (i.e., before manual balancing actually takes over), a processing step for the automatic balancing task is added: like Figure 4 Determine if there is an automatic balancing task currently in progress; if so, save the current state of the automatic balancing task and then pause the task.

[0127] The on-site status includes the identification of the battery cells undergoing equalization, the accumulated equalization duration, and the current operating parameters of the equalization circuit.

[0128] Specifically, when the battery management controller passes the safety check and prepares to switch to manual balancing mode, it first checks the status of each balancing channel. If there are any automatically balancing channels that are discharging, it writes the corresponding battery cell identifiers, the accumulated balancing time of each channel since this startup, and the current operating parameters of the balancing circuit (such as the current current level or PWM duty cycle configuration) into the field status record. Then, it disconnects the balancing switches of these channels, pauses the automatic balancing task, sets the manual mode flag, and enters manual balancing mode. The field status record can be stored in the RAM of the battery management controller or written to non-volatile memory to prevent loss during power failure during switching.

[0129] After manual balancing ends and the system returns to automatic operation, the saved field status allows the system to monitor the progress of the interrupted automatic balancing task. The accumulated balancing time can be included in the continuous discharge time statistics for subsequent balancing of that battery cell, and the interrupted battery cell identifier can be prioritized for review during recovery assessment to check whether its differential pressure still meets the balancing conditions.

[0130] Without a save mechanism, the insertion of manual commands will cause the execution progress information of the automatic balancing task to be lost, which may result in the same battery cell being repeatedly discharged for a long time before and after the interruption, or the interrupted balancing task being forgotten.

[0131] In this embodiment, the switching method of first saving the current situation and then pausing the automatic task ensures that the preemption of automatic balancing by manual balancing is orderly and traceable, thereby ensuring that no control conflict occurs when switching between manual and automatic modes, and also ensuring the continuity of thermal safety-related duration statistics before and after the mode switch.

[0132] Furthermore, after the corresponding passive equalization circuit is started to discharge according to the manual equalization command, real-time monitoring and multi-condition stop logic are added during the execution period, as follows.

[0133] During the execution of the manual balancing command, the voltage of the battery cell to be manually balanced, the discharge current of the corresponding passive balancing circuit, and the temperature of the balancing resistor are monitored in real time. Monitoring is performed cyclically, and each monitoring cycle checks the following stop conditions sequentially. When one of the following conditions is met, the corresponding passive balancing circuit is disconnected: First, the voltage of the battery cells to be manually balanced must reach the target voltage. The target voltage is the target voltage specified in the instruction (which can be a recommended target voltage or a value modified by the maintenance personnel). Reaching the target voltage indicates that the balancing objective has been achieved, and continuing to discharge will result in over-balancing.

[0134] Secondly, the accumulated balancing time has reached the preset balancing time. The preset balancing time is the smaller value between the time set in the instruction and the maximum balancing time. That is, the manually set time is also subject to the thermal safety time constraint, and maintenance personnel cannot bypass thermal protection by setting an excessively long time.

[0135] Third, the temperature of the battery cell to be manually balanced or the temperature of the balancing resistor exceeds the preset safety threshold.

[0136] Fourth, the discharge current exceeds the preset safety range. An abnormally large discharge current may indicate a short circuit in the equalization resistor or a malfunction in the current regulation circuit. An abnormally small discharge current may indicate an open circuit in the equalization branch or that the equalization is not actually effective. In both cases, the process should be stopped and reported.

[0137] Fifth, receiving an emergency stop command. Emergency stop commands can be issued by maintenance personnel through the human-machine interface or generated by the local emergency stop button, giving manual intervention to terminate the balancing process at any time.

[0138] In a preferred embodiment, in addition to the aforementioned monitoring and stopping logic, a hardware watchdog timer independent of the main control program is also provided. The hardware watchdog is activated simultaneously with the balancing switch closing. The main control program feeds the watchdog every monitoring cycle (e.g., 100ms), and the watchdog overflow time is set to a preset balancing duration plus a fixed margin (e.g., plus 2 seconds). If the main control program stops running due to a software malfunction and cannot execute the aforementioned stopping logic, the hardware directly cuts off the balancing switch after the watchdog overflows. The hardware watchdog and the software stopping logic are independent of each other, so even if the main control MCU program malfunctions, the balancing discharge will not continue uncontrollably.

[0139] After the balancing process stops, the complete event of this manual balancing is recorded, including the starting voltage, ending voltage, cumulative balancing duration, cumulative discharge energy, and termination reason. This information is then reported to the energy management server, and the reported data is incorporated into the historical dataset for iterative optimization of subsequent recommendation models and temperature rise prediction models. Because multi-dimensional real-time monitoring of voltage, current, and temperature is implemented during execution, and a multi-layered stop defense system is constructed using five conditions—target achievement, duration limit, temperature exceeding limit, current anomaly, and manual emergency stop—overlaid with a hardware watchdog, manual balancing can be promptly interrupted under any abnormal path, thus ensuring the controllability and safety of manual intervention under various abnormal operating conditions.

[0140] Based on the above embodiments, in order to further prevent mode switching oscillations caused by the system immediately resuming automatic equalization after manual equalization, after disconnecting the corresponding passive equalization circuit, the following is also included: Entering the recovery transition period, during which no automatic balancing tasks may be initiated. The duration of the recovery transition period is configured by the balancing parameters. After the recovery transition period ends, the automatic balancing ready state is restored, and the start-up conditions for automatic balancing are reassessed based on the current state of each battery cell.

[0141] When manual equalization ends, the equalized battery cells have undergone continuous discharge, and their terminal voltages contain polarization components that have not yet dissipated. A period of rest is required for them to relax back to their steady-state open-circuit voltage. If automatic equalization is immediately resumed during this period, the voltage difference calculated based on the transient voltage containing polarization components will be distorted, potentially triggering incorrect equalization actions. This could even lead to repeated start-stop cycles—a mode-switching oscillation—where manual equalization lowers the voltage, automatic equalization determines other cells are too high and initiates, followed by voltage relaxation and a reversal. The recovery transition period can be configured to 3 minutes, or adjusted within a range of 1 to 10 minutes depending on the relaxation characteristics of the battery type. This timeframe is optimized by the energy management server based on historical data (e.g., statistical time required for voltage relaxation to stabilize after equalization in historical equalization records).

[0142] During the recovery transition period, any automatic balancing task is prohibited from being started, but the system can still receive new manual balancing instructions. If a new manual balancing instruction is received during the transition period and passes the security check, the transition timer will be reset and the system will re-enter the manual balancing mode to meet the needs of maintenance personnel to continuously process multiple battery cells.

[0143] After the recovery transition period, the battery management controller returns to the automatic balancing ready state, reloads the latest balancing parameters issued by the energy management server, and reassesses whether the automatic balancing start conditions are met based on the currently collected real-time voltage of each battery cell. The assessment is based on the current actual state, rather than simply continuing the interrupted automatic task, because manual balancing has changed the voltage distribution within the group, and the balancing decisions made before the interruption may no longer be applicable.

[0144] In this embodiment, since a transition period of disabling automatic equalization is inserted between the end of manual equalization and the resumption of automatic equalization, the differential voltage judgment is based on the voltage after the polarization has basically subsided. This eliminates misjudgment and equalization oscillation after mode switching, realizes a smooth connection between manual mode and automatic mode, and improves the system robustness of the mode switching process.

[0145] In another preferred embodiment of the present invention, a passive balancing control system for a battery management system is also disclosed. The control system performs passive balancing on individual battery cells based on the passive balancing control method for a battery management system of any of the above embodiments.

[0146] In summary, this invention discloses a passive equalization control method and system for a battery management system based on big data. Based on historical operating data throughout the entire lifecycle of each battery cell, equalization parameters, including equalization start-up voltage difference threshold and equalization stop voltage difference threshold, are dynamically generated or updated. This allows the passive equalization strategy to adaptively adjust with the battery aging process, overcoming the shortcomings of fixed-threshold strategies that cannot adapt to battery characteristic drift. Furthermore, priority coefficients are used to achieve resource scheduling that tilts towards lagging battery cells when equalization channels are limited. Pre-equalization trigger flags enable predictive intervention against consistency divergence trends. A maximum equalization duration based on a temperature rise prediction model achieves thermal safety constraints that match actual heat dissipation conditions. Further, a manual equalization recommendation list and associated recommended equalization duration, recommended target voltage, and recommended equalization current level make manual intervention entirely data-driven. Mechanisms such as pre-emptive safety verification, on-site status saving, multiple monitoring stops during execution, and recovery transition periods ensure safe, smooth, and oscillating switching between manual and automatic modes. The battery management controller transmits the equalization execution results back to the energy management server, forming a data closed loop, allowing various equalization parameters to be continuously optimized with data accumulation.

[0147] This invention also discloses another passive equalization control system, which includes one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors. The programs include instructions for performing the passive equalization control method as described above. The processors may be general-purpose central processing units (CPUs), microprocessors, application-specific integrated circuits (ASICs), or one or more integrated circuits, used to execute relevant programs to achieve the functions required by the modules in the passive equalization control system of this application embodiment, or to execute the passive equalization control method of this application method embodiment.

[0148] This invention also discloses a computer-readable storage medium comprising a computer program executable by a processor to perform the passive equalization control method described above. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center integrating one or more available media. The available medium can be read-only memory (ROM), random access memory (RAM), or magnetic media, such as floppy disks, hard disks, magnetic tapes, magnetic disks, or optical media, such as digital versatile discs (DVDs), or semiconductor media, such as solid-state disks (SSDs).

[0149] This application also discloses a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. The processor of an electronic device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the electronic device to perform the aforementioned passive equalization control method.

[0150] The above-disclosed embodiments are merely preferred embodiments of the present invention and should not be construed as limiting the scope of the present invention. Therefore, any equivalent variations made in accordance with the claims of the present invention are still within the scope of the present invention.

Claims

1. A passive balancing control method for a battery management system based on big data, applied to an energy storage battery system, wherein the energy storage battery system includes several battery packs, each battery pack has several individual battery cells, and each individual battery cell is equipped with a passive balancing circuit; characterized in that, The control method includes: Collect and store historical operating data of each of the battery cells to obtain a historical dataset; The equalization parameters of the energy storage battery system are generated or updated based on the historical dataset. The equalization parameters include at least the equalization start-up differential pressure threshold and the equalization stop-down differential pressure threshold. The real-time voltage of each battery cell is collected, and for any battery pack, when the real-time voltage difference between any two battery cells in the battery pack is greater than the equalization start voltage difference threshold, the battery cell with the relatively higher voltage among the two battery cells corresponding to the real-time voltage difference is identified as the battery cell to be equalized, and the passive equalization circuit configured for the battery cell to be equalized is activated. For any battery cell in passive balancing, when the real-time voltage difference between the battery cell and the other battery cells in the battery pack is less than the balancing stop voltage difference threshold, the passive balancing circuit configured for the battery cell is controlled to stop.

2. The passive balancing control method for a battery management system according to claim 1, characterized in that, The historical operating data includes at least one of the following: historical voltage data, historical current data, historical temperature data, historical state of charge data, historical health status data, and historical balancing records for each of the battery cells.

3. The passive balancing control method for a battery management system according to claim 1, characterized in that, The method for generating the equilibrium parameters based on the historical operating data includes: The historical operational data is cleaned, and the data cleaning includes at least one of noise reduction and interpolation; Based on the historical operating data after cleaning, the consistency characteristics of each battery cell are extracted. The consistency characteristics include at least one of voltage fluctuation characteristics, capacity decay characteristics, and internal resistance growth characteristics. The consistency characteristic is input into the big data analysis model, and the big data analysis model outputs the equilibrium start pressure difference threshold and the equilibrium stop pressure difference threshold.

4. The passive balancing control method for a battery management system according to claim 1, characterized in that, The energy storage battery system also includes an energy management server and a battery management controller that is communicatively connected to the energy management server. The energy management server collects and stores historical operating data of each battery cell, generates or updates the balancing parameters, and sends the balancing parameters to the battery management controller via a communication link. The battery management controller performs state data acquisition and passive balancing control of the individual battery cells.

5. The passive balancing control method for a battery management system according to claim 1, characterized in that, The balancing parameters also include a priority coefficient corresponding to each of the battery cells, the priority coefficient being used to characterize the balancing priority of the corresponding battery cell; If the number of battery cells to be balanced exceeds a preset threshold, the priority order for balancing is determined according to the priority coefficient corresponding to each battery cell to be balanced.

6. The passive balancing control method for a battery management system according to claim 5, characterized in that, The method for generating the priority coefficient includes: Based on the historical operating data, the voltage fluctuation characteristics, health status degradation, and internal resistance increase of each battery cell are obtained. The priority coefficient of the battery cell is obtained by weighted summing of the voltage fluctuation characteristic, the health state decay, and the internal resistance increase.

7. The passive balancing control method for a battery management system according to claim 1, characterized in that, The equalization parameters also include a pre-equalization trigger identifier, which is determined based on the consistency change trend of the battery cells reflected by the historical operating data. The method further includes: When the real-time voltage difference is less than the equalization start voltage difference threshold, the passive equalization circuit configured in one of the two battery cells corresponding to the real-time voltage difference and bearing the pre-equalization trigger mark is activated.

8. The passive balancing control method for a battery management system according to claim 7, characterized in that, The pre-equalization trigger identifier is dynamically generated in the following manner: Based on the historical operating data, predict the voltage change trend of each battery cell within a preset time period in the future; Based on the voltage change trend, the pre-equalization trigger flag is configured for battery cells whose voltage deviation meets the preset divergence condition within the predicted future preset time period.

9. The passive balancing control method for a battery management system according to claim 1, characterized in that, The equalization parameters also include the maximum equalization duration; The method further includes: For any battery cell that is discharging, the duration of continuous discharge of the passive balancing circuit configured for that battery cell is recorded. When the continuous discharge duration reaches the maximum equalization duration, the passive equalization circuit configured for the battery cell stops discharging.

10. The passive equalization control method for a battery management system according to claim 9, characterized in that, The maximum equalization duration is generated in the following manner: Based on the discharge duration, temperature change data during discharge, and ambient temperature corresponding to each equalization discharge in the historical operation data, a temperature rise prediction model is constructed. The temperature rise prediction model is used to predict the temperature rise of a single battery cell under continuous equalization discharge. Based on the temperature rise predicted by the temperature rise prediction model and the upper limit of the safe temperature, the longest continuous discharge time that ensures the temperature of a single battery cell does not exceed the upper limit of the safe temperature is determined as the maximum equalization time.

11. The passive balancing control method for a battery management system according to claim 1, characterized in that, The balancing parameters also include a manually balanced recommendation list; For each of the battery cells, the battery cells to be manually balanced are identified based on the voltage, internal resistance growth rate and health status indicators of the battery cells in the historical operating data, so as to generate the manual balancing recommendation list. The manually balanced recommendation list is displayed on the human-computer interaction interface; The system receives a manual balancing command input through the human-computer interaction interface, wherein the manual balancing command specifies at least one battery cell to be manually balanced from the manual balancing recommendation list. The corresponding passive equalization circuit is activated according to the manual equalization command.

12. The passive balancing control method for a battery management system according to claim 11, characterized in that, For a battery cell to be manually balanced, a recommended balancing time and a recommended target voltage are calculated based on the voltage distribution characteristics and capacity decay trajectory of the battery cell, and the recommended balancing time and recommended target voltage are associated with the manual balancing recommendation list. Alternatively, for a battery cell to be manually balanced, a recommended balancing current level is determined based on the internal resistance and capacity decay of the battery cell in the historical operating data, and the recommended balancing current level is associated with the manual balancing recommendation list.

13. The passive balancing control method for a battery management system according to claim 11, characterized in that, Before the corresponding passive equalization circuit is started to discharge according to the manual equalization command, a safety verification method is also included: Determine whether the battery cell to be manually balanced is in a fault state, or whether there is an unrecoverable alarm, or whether the predicted maximum temperature of the passive balancing circuit during the execution of the manual balancing command exceeds the upper limit of the safe temperature. If so, refuse to execute the manual balancing command and maintain the current operating mode.

14. The passive balancing control method for a battery management system according to claim 11, characterized in that, Before receiving the manual equalization command input through the human-computer interaction interface, the method further includes: Determine if there is an ongoing automatic load balancing task; If so, save the current state of the automatic balancing task and then pause the automatic balancing task. The on-site status includes the identification of the battery cell undergoing equalization, the accumulated equalization duration, and the current operating parameters of the equalization circuit.

15. The passive equalization control method for a battery management system according to claim 14, characterized in that, After controlling the corresponding passive equalization circuit to start discharging according to the manual equalization command, the process also includes: During the execution of the manual balancing command, the voltage of the battery cell to be manually balanced, the discharge current of the corresponding passive balancing circuit, and the temperature of the balancing resistor are monitored in real time. The corresponding passive equalization circuit is disconnected when one of the following conditions is met: The voltage of the battery cell to be manually balanced reaches the target voltage; The accumulated balancing time has reached the preset balancing time; The temperature of the battery cell to be manually balanced or the temperature of the balancing resistor exceeds a preset safety threshold. The discharge current exceeds the preset safety range; An emergency stop command has been received.

16. The passive equalization control method for a battery management system according to claim 15, characterized in that, After disconnecting the corresponding passive equalization circuit, the process further includes: Entering the recovery transition period, during which any automatic balancing tasks are prohibited from being started; The duration of the recovery transition period is configured by the balancing parameters; After the recovery transition period ends, the automatic balancing ready state is restored, and the start-up conditions for automatic balancing are reassessed based on the current state of the individual battery cells.

17. A passive balancing control system for a battery management system, characterized in that, The control system performs passive balancing on individual battery cells based on the passive balancing control method of the battery management system according to any one of claims 1 to 16.

18. A passive balancing control system for a battery management system, characterized in that, include: One or more processors; Memory; And one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the programs including instructions for performing the passive equalization control method of the battery management system as described in any one of claims 1 to 16.

19. A computer-readable storage medium, characterized in that, Includes a computer program, which can be executed by a processor to perform the passive equalization control method for a battery management system as described in any one of claims 1 to 16.

Images