A battery management system balancing method based on real-time stream processing
By employing a real-time stream processing-based battery management system balancing method, Kalman filtering and Bayesian inference are used to analyze the state of individual battery cells, calculate cell risk scores and balancing currents, and solve the problems of low battery pack balancing efficiency and insufficient safety, thus achieving effective battery pack balancing.
Patent Information
- Application Number
- CN202611237373.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-14
- Publication Date
- 2026-11-17
AI Technical Summary
Existing battery pack balancing methods are inefficient, lack the ability to identify abnormal battery cells, make it difficult to ensure the effectiveness of balancing, and pose a risk of over-balancing and damage to low-voltage cells.
A battery management system balancing method based on real-time stream processing is adopted. By using Kalman filtering and Bayesian inference to enhance voltage rank analysis, cell risk scores and balancing currents are calculated. Combined with individual cell analysis and control methods, effective identification and protection are achieved.
It improves the efficiency and safety of battery pack balancing, ensures the effectiveness of balancing, and avoids the risk of over-balancing and damage to low-voltage cells.
Smart Images

Figure CN123121908A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of battery management system (BMS) technology, and in particular to a battery management system balancing method based on real-time stream processing. Background Technology
[0002] Energy storage systems and electric vehicle battery packs are typically composed of multiple battery cells connected in series. Due to differences in manufacturing processes, temperatures, and aging, the capacity and voltage of each battery cell inevitably vary (i.e., cell inconsistency). When cell inconsistency occurs, the usable capacity of the battery pack is determined by the weakest battery cell, which leads to reduced capacity, shorter driving range, and accelerated aging. Based on the foregoing, it is necessary to perform battery pack balancing to address "individual cell inconsistency"; however, existing methods for battery pack balancing have the following limitations: First, traditional local balancing only operates at the end of the charging process, resulting in low balancing efficiency. Second, the lack of effective identification of abnormal battery cells makes it difficult to ensure the effectiveness of the equalization process. Third, no specific analysis of the reasons for the pressure difference was conducted for the individual battery cells to be balanced, which also makes it difficult to ensure the effectiveness of the balancing process. Fourth, for the individual battery cells to be balanced, no balancing protection was provided during the balancing process, which poses a risk of over-balancing and damage to low-voltage cells.
[0003] To address the aforementioned technical problems, this application proposes a battery management system balancing method based on real-time stream processing. Summary of the Invention
[0004] In view of the above-mentioned shortcomings and deficiencies of the prior art, the present invention provides a battery management system balancing method based on real-time stream processing, which solves the current technical problem of "how to improve balancing efficiency and ensure balancing safety while ensuring the balancing effectiveness of the battery management system".
[0005] To achieve the above objectives, the main technical solutions adopted by the present invention include: This invention provides a battery management system balancing method based on real-time stream processing, used for balancing multiple battery cells in a series connection structure involved in the same battery pack; The method includes: S1, identifying all battery cells involved in the same battery pack; and calculating the cell risk score corresponding to each battery cell. Equalizing current ; S1 includes S1-1 to S1-4; S1-1 to S1-4 specifically involve: S1-1. Collect relevant data for each battery cell in real time and perform relevant data preprocessing; Based on relevant data: individual battery cells belonging to the same battery pack The corresponding charging sorting set Discharge sorting set To obtain the rank drift set ; Obtain the voltage rank set based on the rank drift set ; The relevant data includes voltage, temperature, and state of charge; Characterizes the cell number; / Characterization: For the same round of charging / discharging, the order in which different battery cells complete charging / discharging is numbered; Characterizing battery cells The voltage rank; S1-2, Based on constructing the Kalman equation to determine the voltage rank Perform Kalman filtering enhancement to obtain: corresponding to the voltage rank of State estimation set ; S1-3, Based on constructing Bayesian inference equations to determine voltage rank Perform Bayesian inference enhancement to obtain: voltage rank probability set greater than the risk rank threshold ; S1-4, based on The state estimation set and probability set are used to calculate: the cell risk score and equalization current corresponding to each battery cell; Formulas involving cell risk scoring and balancing current are as follows: (1) and (2): (1) (2) in, For each battery cell, ; for Scoring weights; Assign weights to the Bayesian score; for Normalized score; Bayesian normalized score; The number of loops. , For the preset number of times; Voltage rank of State estimation; This is the cell rank drift threshold; This is the risk rank threshold; This refers to the observed rank shift data; To balance the upper limit of current; For comparison parameters; / The rank of the current reference maximum / minimum voltage.
[0006] Optionally, the Kalman equation involves formulas (3) and (4): (3) (4) in, This is process noise; The rank of the observed voltage; To observe noise; These are prior predictions; The prior error covariance; For the posterior error covariance; For process noise variance; Kalman gain; To observe the noise variance; in, , ; / Follows distribution / ; .
[0007] Optionally, the Bayesian inference equation involves formula (5): (5) in, Follows a prior distribution of rank ; For the specific value of the voltage rank, ; Given the voltage rank, The posterior probability; For the observed rank drift; Voltage rank The corresponding expected rank drift; The standard deviation of the observed noise; For the first Next cycle voltage rank False counting; To adjust the voltage rank after adding new observations Update pseudo-count; This is the learning rate.
[0008] Optionally, the method further includes: S2, according to the voltage rank corresponding to each battery cell. , State estimation Single voltage rank fluctuation Battery cell risk score Static voltage drop rate, internal resistance increment, and equalization current Each battery cell is analyzed, marked, and balanced through a battery cell analysis and control method. The battery cell analysis and control method includes: If the following is satisfied: single voltage rank fluctuation and State estimation If there is no monotonic trend, the cell is marked as a temporary capacity shift; and based on the equalization current... Perform standard equilibrium and wait for natural recovery; If the following condition is met: static pressure drop rate is greater than Voltage rank If the internal resistance increase is normal, then the battery cell is marked as having an abnormal self-discharge; and based on the equalization current... Perform priority charging compensation equalization to shorten the equalization cycle; If the voltage rank is satisfied... Battery cell risk score And the increase in internal resistance is greater than If the battery cell is marked as having capacity decay, charging equalization will be disabled to reduce the charging rate.
[0009] Optional, based on equalization current In addition, there are preset three-level rules and equalization scale control rules to perform battery management system equalization; The balanced scale control rules include: integrated quantity limit rules and circuit breaker protection rules.
[0010] Optionally, the three-tier rules include: The differential pressure threshold is dynamically calibrated based on the 90th quantile of the differential pressure distribution over the most recent 100 cycles; when the differential pressure convergence rate is less than 20% for five consecutive cycles, the differential pressure threshold is automatically reduced to 30%. To improve sensitivity; when the differential pressure convergence rate is greater than 80%, the differential pressure threshold automatically increases to 60. To prevent over-balancing; If the temperature of the individual battery cells is within a reasonable range If the temperature is between 20℃ and 60℃, the equalization channel will be functioning without fault; otherwise, the equalization channel will malfunction. Battery management system balancing is performed based on the priority of individual battery cells; the acquisition of the priority-ranking of individual battery cells involves: combining differential pressure, risk score, spatial proximity of balancing channels, and hot zone avoidance, and using Pareto front screening to output a non-dominated solution set.
[0011] Optionally, the integration quantity limitation rules include: The maximum number of balanced battery cells shall not exceed 50% of the total number of hardware channels; If the maximum voltage difference of the current battery pack is greater than 100 Then, the battery management system will perform equalization with the maximum number of equalized battery cells to quickly balance the voltage; If the maximum voltage difference of the current battery pack is at Then, the battery management system is used to balance the cells from high to low based on the pressure difference, and the cells with the maximum number of cells to be balanced are selected. If the maximum voltage difference of the current battery pack is less than 50 Then, the battery management system balancing is performed only on 3 to 5 extreme cells.
[0012] Optionally, the triggering condition for the fuse protection rule is: the battery cell corresponding to the lowest voltage completes a battery management system equalization. The execution action of the circuit breaker protection rule is as follows: suspend all equalization operations of the current battery pack for 6 hours; after the suspension is restored, execute a new round of battery management system equalization.
[0013] The beneficial effects of this application include: In this application, Kalman filtering enhancement is performed based on constructing the Kalman equation to obtain: the voltage rank corresponding to the... State estimation set; Bayesian inference enhancement is performed based on the constructed Bayesian inference equation to obtain: the probability set of voltage rank greater than the risk rank threshold; based on the above, the cell risk score and equalization current corresponding to each battery cell are calculated. Based on this, each battery cell is analyzed, marked and equalized by the battery cell analysis and control method, which can effectively identify abnormal battery cells and thus ensure the effectiveness of equalization. In this application, the aforementioned battery cell analysis and control method not only involves charging, discharging, and resting periods, but also involves specific analysis and control of battery cells. Based on this, it can not only ensure balancing efficiency, but also perform specific analysis of the causes of pressure difference to ensure the effectiveness of balancing. In this application, the equalization scale control rules involve integration quantity limits and circuit breaker protection, thereby enabling battery cell equalization protection during the equalization process. Attached Figure Description
[0014] Figure 1 This is a schematic diagram of the battery management system balancing process provided in one embodiment of the present invention. Detailed Implementation
[0015] To better explain and facilitate understanding of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.
[0016] To better understand the above technical solutions, exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that the present invention can be understood more clearly and thoroughly, and that the scope of the present invention can be fully conveyed to those skilled in the art.
[0017] Example 1 This embodiment proposes a battery management system balancing method based on real-time stream processing, which is used to balance and manage multiple battery cells in the same battery pack with a series connection structure. In this embodiment, Figure 1 This is a schematic diagram of the battery management system balancing process; for example... Figure 1 As shown, the aforementioned battery management system balancing method involves the following S1 to S2; S1. Identify all individual battery cells involved in the same battery pack; calculate the cell risk score corresponding to each individual battery cell. Equalizing current ; It should be noted that the aforementioned S1 includes S1-1 to S1-4; S1-1. Collect relevant data for each battery cell in real time and perform relevant data preprocessing; Based on relevant data: individual battery cells belonging to the same battery pack The corresponding charging sorting set Discharge sorting set To obtain the rank drift set ; Obtain the voltage rank set based on the rank drift set ; The relevant data includes voltage, temperature, and state of charge; Characterizes the cell number; / Characterization: For the same round of charging / discharging, the order in which different battery cells complete charging / discharging is numbered; Characterizing battery cells The voltage rank; S1-2, Based on constructing the Kalman equation to determine the voltage rank Perform Kalman filtering enhancement to obtain: corresponding to the voltage rank of State estimation set ; It should be noted that the aforementioned Kalman equation involves formulas (3) and (4): (3) (4) in, This is process noise; The rank of the observed voltage; To observe noise; These are prior predictions; The prior error covariance; For the posterior error covariance; For process noise variance; Kalman gain; To observe the noise variance; in, , ; / Follows distribution / ; .
[0018] S1-3, Based on constructing Bayesian inference equations to determine voltage rank Perform Bayesian inference enhancement to obtain: voltage rank probability set greater than the risk rank threshold ; It should be noted that the aforementioned Bayesian inference equation involves formula (5): (5) in, Follows a prior distribution of rank ; For the specific value of the voltage rank, ; Given the voltage rank, The posterior probability; For the observed rank drift; Voltage rank The corresponding expected rank drift; The standard deviation of the observed noise; For the first Next cycle voltage rank False counting; To adjust the voltage rank after adding new observations Update pseudo-count; This is the learning rate.
[0019] S1-4, based on The state estimation set and probability set are used to calculate: the cell risk score and equalization current corresponding to each battery cell; It should be noted that the formulas involving cell risk scoring and balancing current are as follows: (1) and (2): (1) (2) in, For each battery cell, ; for Scoring weights; Assign weights to the Bayesian score; for Normalized score; Bayesian normalized score; The number of loops. , For the preset number of times; Voltage rank of State estimation; This is the cell rank drift threshold; This is the risk rank threshold; This refers to the observed rank shift data; To balance the upper limit of current; For comparison parameters; / The rank of the current reference maximum / minimum voltage.
[0020] S2. Based on the voltage rank of each battery cell , State estimation Single voltage rank fluctuation Battery cell risk score Static voltage drop rate, internal resistance increment, and equalization current Each battery cell is analyzed, marked, and balanced through a battery cell analysis and control method. It should be noted that the aforementioned battery cell analysis and control methods include: If the following is satisfied: single voltage rank fluctuation and State estimation If there is no monotonic trend, the cell is marked as a temporary capacity shift; and based on the equalization current... Perform standard equilibrium and wait for natural recovery; If the following condition is met: static pressure drop rate is greater than Voltage rank If the internal resistance increase is normal, then the battery cell is marked as having an abnormal self-discharge; and based on the equalization current... Perform priority charging compensation equalization to shorten the equalization cycle; If the voltage rank is satisfied... Battery cell risk score And the increase in internal resistance is greater than If the battery cell is marked as having capacity decay, charging equalization will be disabled to reduce the charging rate.
[0021] It should be noted that: based on equalization current In addition, there are preset three-level rules and equalization scale control rules to perform battery management system equalization; among them, the equalization scale control rules include: integration quantity limit rules and fuse protection rules.
[0022] It should be noted that the aforementioned three-tiered rules include: The differential pressure threshold is dynamically calibrated based on the 90th quantile of the differential pressure distribution over the most recent 100 cycles; when the differential pressure convergence rate is less than 20% for five consecutive cycles, the differential pressure threshold is automatically reduced to 30%. To improve sensitivity; when the differential pressure convergence rate is greater than 80%, the differential pressure threshold automatically increases to 60. To prevent over-balancing; If the temperature of the individual battery cells is within a reasonable range If the temperature is between 20℃ and 60℃, the equalization channel will be functioning without fault; otherwise, the equalization channel will malfunction. Battery management system balancing is performed based on the priority of individual battery cells; the acquisition of the priority-ranking of individual battery cells involves: combining differential pressure, risk score, spatial proximity of balancing channels, and hot zone avoidance, and using Pareto front screening to output a non-dominated solution set.
[0023] It should be noted that the aforementioned integration quantity limit rules include: The maximum number of balanced battery cells shall not exceed 50% of the total number of hardware channels; If the maximum voltage difference of the current battery pack is greater than 100 Then, the battery management system will perform equalization with the maximum number of equalized battery cells to quickly balance the voltage; If the maximum voltage difference of the current battery pack is at Then, the battery management system is used to balance the cells from high to low based on the pressure difference, and the cells with the maximum number of cells to be balanced are selected. If the maximum voltage difference of the current battery pack is less than 50 Then, the battery management system balancing is performed only on 3 to 5 extreme cells.
[0024] It should be noted that the triggering condition for the aforementioned fuse protection rule is: the battery cell corresponding to the lowest voltage completes a battery management system equalization. The aforementioned circuit breaker protection rule will be executed as follows: all balancing operations of the current battery pack will be suspended for 6 hours; after the suspension is restored, a new round of battery management system balancing will be performed.
[0025] Based on the battery management system balancing method based on real-time stream processing described in Embodiment 1 above, Kalman filtering enhancement is performed by constructing a Kalman equation to obtain: the voltage rank... State estimation set; Bayesian inference enhancement is performed based on the constructed Bayesian inference equation to obtain: the probability set of voltage rank greater than the risk rank threshold; based on the above, the cell risk score and equalization current corresponding to each battery cell are calculated. Based on this, each battery cell is analyzed, marked and equalized by the battery cell analysis and control method, which can effectively identify abnormal battery cells and thus ensure the effectiveness of equalization. Based on the battery management system balancing method based on real-time stream processing described in the first embodiment, the aforementioned battery cell analysis and control method not only involves charging, discharging, and resting periods, but also involves specific analysis and control of battery cells. Based on this, not only can balancing efficiency be ensured, but also specific pressure difference cause analysis can be performed to ensure the effectiveness of balancing. Based on the battery management system balancing method based on real-time stream processing described in the aforementioned Embodiment 1, the balancing scale control rules involve integration quantity limits and circuit breaker protection, thereby enabling battery cell balancing protection during the balancing process.
[0026] Example 2 This embodiment proposes an implementation case based on the "Battery Management System Balancing Method described in Embodiment 1", specifically involving: Hardware specifications: 48-cell lithium iron phosphate battery pack, rated capacity 50Ah, supporting up to 24 simultaneous passive equalization channels; Data acquisition specifications: 1-second interval, single-cell voltage accuracy 1mV, temperature accuracy 0.1℃; Relevant parameters: cell rank drift threshold. 10 (5~48 recommended for 48-string systems); process noise variance The observation noise variance is 0.5. The learning rate is 8. It is 0.3; Rating weight The Bayesian scoring weight is 0.65. The value is 0.65; the voltage difference threshold is 50mV; the maximum number of equalized battery cells is 12; the test conditions are: the first 12 strings are manually discharged to about 60% capacity to simulate the difference, full charge-full discharge condition, the equalization time is set, and 17 days of data are recorded; the test results are shown in Table 1. Table 1. 17-day test results
[0027] Based on the foregoing, the simulation conditions involved in the algorithm verification are: 20 lithium iron phosphate cells in series, rated at 50Ah, 3 weak cells (39~45Ah), 30 charge-discharge cycles, and an observation noise of 3.0. The algorithm verification results are shown in Table 2.
[0028] Table 2. Algorithm Validation Results
[0029] It should be noted that any reference numerals placed between parentheses in the claims should not be construed as limiting the claims. The word "comprising" does not exclude the presence of components or steps not listed in the claims. The words "a" or "an" preceding a component do not exclude the presence of a plurality of such components. The use of the terms first, second, third, etc., is for convenience only and does not indicate any order. These terms can be understood as part of the component names.
[0030] Furthermore, it should be noted that in the description of this specification, the terms "one embodiment," "some embodiments," "embodiment," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Furthermore, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.
[0031] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the claims should be interpreted to include both the preferred embodiments and all changes and modifications falling within the scope of the invention.
[0032] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, then this invention should also include these modifications and variations.
Claims
1. A battery management system balancing method based on real-time stream processing, used for balancing multiple battery cells in a series connection structure within the same battery pack, characterized in that: The method includes: S1, identifying all battery cells involved in the same battery pack; and calculating the cell risk score corresponding to each battery cell. Equalizing current ; S1 includes S1-1 to S1-4; S1-1 to S1-4 specifically involve: S1-1. Collect relevant data for each battery cell in real time and perform relevant data preprocessing; Based on relevant data: individual battery cells belonging to the same battery pack The corresponding charging sorting set Discharge sorting set To obtain the rank drift set ; Obtain the voltage rank set based on the rank drift set ; The relevant data includes voltage, temperature, and state of charge; Characterizes the cell number; / Characterization: For the same round of charging / discharging, the order in which different battery cells complete charging / discharging is numbered; Characterizing battery cells The voltage rank; S1-2, Based on constructing the Kalman equation to determine the voltage rank Perform Kalman filtering enhancement to obtain: corresponding to the voltage rank of State estimation set ; S1-3, Based on constructing Bayesian inference equations to determine voltage rank Perform Bayesian inference enhancement to obtain: voltage rank probability set greater than the risk rank threshold ; S1-4, based on The state estimation set and probability set are used to calculate: the cell risk score and equalization current corresponding to each battery cell; Formulas involving cell risk scoring and balancing current are as follows: (1) and (2): (1) (2) in, For each battery cell, ; for Scoring weights; Assign weights to the Bayesian score; for Normalized score; Bayesian normalized score; The number of loops. , For the preset number of times; Voltage rank of State estimation; This is the cell rank drift threshold; This is the risk rank threshold; This refers to the observed rank shift data; To balance the upper limit of current; For comparison parameters; / The rank of the current reference maximum / minimum voltage.
2. The battery management system balancing method according to claim 1, characterized in that: The Kalman equations involve formulas (3) and (4): (3) (4) in, This is process noise; The rank of the observed voltage; To observe noise; These are prior predictions; The prior error covariance; For the posterior error covariance; For process noise variance; Kalman gain; To observe the noise variance; in, , ; / Follows distribution / ; .
3. The battery management system balancing method according to claim 2, characterized in that: The Bayesian inference equation involves formula (5): (5) in, Follows a prior distribution of rank ; For the specific value of the voltage rank, ; Given the voltage rank, The posterior probability; For the observed rank drift; Voltage rank The corresponding expected rank drift; The standard deviation of the observed noise; For the first Next cycle voltage rank False counting; To adjust the voltage rank after adding new observations Update pseudo-count; This is the learning rate.
4. The battery management system balancing method according to claim 3, characterized in that: The method further includes: S2, according to the voltage rank of each battery cell. , State estimation Single voltage rank fluctuation Battery cell risk score Static voltage drop rate, internal resistance increment, and equalization current Each battery cell is analyzed, marked, and balanced through a battery cell analysis and control method. The battery cell analysis and control method includes: If the following is satisfied: single voltage rank fluctuation and State estimation If there is no monotonic trend, the cell is marked as a temporary capacity shift; and based on the equalization current... Perform standard equilibrium and wait for natural recovery; If the following condition is met: static pressure drop rate is greater than Voltage rank If the internal resistance increase is normal, then the battery cell is marked as having an abnormal self-discharge; and based on the equalization current... Perform priority charging compensation equalization to shorten the equalization cycle; If the voltage rank is satisfied... Battery cell risk score And the increase in internal resistance is greater than If the battery cell is marked as having capacity decay, charging equalization will be disabled to reduce the charging rate.
5. The battery management system balancing method according to claim 4, characterized in that: Based on balanced current In addition, there are preset three-level rules and equalization scale control rules to perform battery management system equalization; The balanced scale control rules include: integrated quantity limit rules and circuit breaker protection rules.
6. The battery management system balancing method according to claim 5, characterized in that: The three-tier rules include: The differential pressure threshold is dynamically calibrated based on the 90th quantile of the differential pressure distribution over the most recent 100 cycles; when the differential pressure convergence rate is less than 20% for five consecutive cycles, the differential pressure threshold is automatically reduced to 30%. To improve sensitivity; when the differential pressure convergence rate is greater than 80%, the differential pressure threshold automatically increases to 60. To prevent over-balancing; If the temperature of the individual battery cells is within a reasonable range If the temperature is between 20℃ and 60℃, the equalization channel will be functioning without fault; otherwise, the equalization channel will malfunction. Battery management system balancing is performed based on the priority of individual battery cells; the acquisition of the priority-ranking of individual battery cells involves: combining differential pressure, risk score, spatial proximity of balancing channels, and hot zone avoidance, and using Pareto front screening to output a non-dominated solution set.
7. The battery management system balancing method according to claim 5, characterized in that: The integration quantity restriction rules include: The maximum number of balanced battery cells shall not exceed 50% of the total number of hardware channels; If the maximum voltage difference of the current battery pack is greater than 100 Then, the battery management system will perform equalization with the maximum number of equalized battery cells to quickly balance the voltage; If the maximum voltage difference of the current battery pack is at Then, the battery management system is used to balance the cells from high to low based on the pressure difference, and the cells with the maximum number of cells to be balanced are selected. If the maximum voltage difference of the current battery pack is less than 50 Then, the battery management system balancing is performed only on 3 to 5 extreme cells.
8. The battery management system balancing method according to claim 5, characterized in that: The trigger condition for the circuit breaker protection rule is: the battery cell corresponding to the lowest voltage completes a battery management system equalization. The execution action of the circuit breaker protection rule is as follows: suspend all equalization operations of the current battery pack for 6 hours; After the recovery is suspended, a new round of battery management system balancing will be performed.
