Battery management method and battery management system

Abstract

This application discloses a battery management method and a battery management system. The battery management method includes: acquiring multiple sampling data points of a battery cell stack at a preset frequency within each statistical cycle; the sampling data includes real-time data of different operating parameters; determining the mean and standard deviation of different operating parameters of the battery cell stack based on the multiple sampling data; configuring the mean and standard deviation of the corresponding dynamic threshold models for each operating parameter based on the mean and standard deviation of the different operating parameters; the dynamic threshold models including dynamic coefficients corresponding to the standard deviations of different operating parameters; determining the abnormal state type of the battery cell stack based on the configured dynamic threshold models for each operating parameter and each sampling data point of the battery cell stack in the next statistical cycle; and invoking the corresponding optimization strategy based on the determined abnormal state type of the battery cell stack to optimize and adjust the battery cell stack. This application can effectively reduce the misjudgment rate in battery management.

Description

Technical Field

[0001] This application relates to the field of battery management technology, and in particular to battery management methods and battery management systems. Background Technology

[0002] Liquid-cooled battery systems are widely used in critical scenarios such as energy storage power stations, electric vehicle power battery systems, and data center uninterruptible power supplies due to their excellent heat dissipation and thermal equalization capabilities. A liquid-cooled battery consists of multiple cell stacks. As usage time increases, the dynamic changes in operating parameters such as voltage, temperature, and internal resistance of the cell stacks during charging and discharging directly affect the overall performance and lifespan of the liquid-cooled battery.

[0003] To ensure the safe operation of liquid-cooled batteries, the operating parameters of the cell stack are typically monitored, and abnormal cell stacks are controlled based on the monitoring results. In practice, fixed thresholds are generally used to determine whether the operating parameters of the cell stack are abnormal. That is, fixed upper / lower limits for voltage, upper limit for temperature, upper limit for internal resistance, etc. are preset. When the operating parameters of the cell stack exceed the fixed thresholds, it is considered abnormal.

[0004] However, the operating parameters of a battery cell stack are affected by various factors such as ambient temperature, charge / discharge rate, and aging degree, and their normal fluctuation range has time-varying characteristics. Fixed thresholds are difficult to adapt to the time-varying characteristics of the battery cell stack's operating parameters, resulting in a high false alarm rate for anomaly detection.

[0005] Therefore, how to effectively reduce the misjudgment rate in battery management remains a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0006] In view of this, in order to solve the above-mentioned technical problems, this application provides a battery management method and a battery management system.

[0007] To address the aforementioned technical problems, one technical solution adopted in this application is to provide a battery management method for managing a liquid-cooled battery, which includes multiple cell stacks. The battery management method includes:

[0008] Multiple sampling data of the battery cell stack are acquired at a preset frequency within each statistical period. The sampling data includes real-time data of different operating parameters. The statistical data of the battery cell stack was determined based on multiple sampling data of the battery cell stack. The statistical data included the mean and standard deviation of different operating parameters. Based on the mean and standard deviation of different operating parameters, configure the mean and standard deviation of the dynamic threshold model corresponding to each operating parameter. The dynamic threshold model includes dynamic coefficients corresponding to the standard deviation of different operating parameters. Based on the dynamic threshold models corresponding to different configured operating parameters and the sampling data of the cell stack in each statistical period, the abnormal state type of the cell stack is determined. Based on the determined abnormal state type of the battery cell stack, the corresponding optimization strategy is invoked to optimize and adjust the battery cell stack accordingly.

[0009] To address the aforementioned technical problems, another technical solution adopted in this application is to provide a battery management system, which includes: Liquid-cooled battery, consisting of multiple cell stacks; Multiple sampling units are set up, each corresponding to a different cell stack, to monitor different operating parameters of the multiple cell stacks in order to obtain real-time data of different operating parameters of the multiple cell stacks; The optimization and adjustment unit is used to optimize and adjust multiple battery cell stacks. The system is connected to the sampling unit and the optimization adjustment unit via communication to control the sampling unit and the optimization adjustment unit to implement the battery management method described above.

[0010] Beneficial effects: Unlike existing technologies that rely on fixed thresholds, this technology does not depend on fixed thresholds. Instead, it uses a dynamic threshold model to adapt to individual differences in different cell stacks and environmental changes. This reduces the false positive rate, lowers maintenance costs caused by inaccurate anomaly identification, and improves the operational safety and lifespan of the liquid-cooled battery system. Attached Figure Description

[0011] Figure 1 This is a flowchart illustrating the battery management method of this application; Figure 2 This is a schematic diagram of a module of an embodiment of the battery management system of this application; Figure 3 This is a schematic diagram of the liquid-cooled battery module of the battery management system of this application. Detailed Implementation

[0012] To enable those skilled in the art to better understand the technical solutions of this application, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0013]

Terminology Definition

[0014] [Example 1: Basic Method Example] like Figure 1 As shown, the battery management method is used to manage a liquid-cooled battery, which includes multiple cell stacks. The battery management method includes steps S101 to S105.

[0015] Step S101: Acquire multiple sampling data of the battery cell stack at a preset frequency within each statistical period. The sampling data includes real-time data of different operating parameters.

[0016] Optionally, in this embodiment, the system monitors different operating parameters of the battery cell stack in real time at a preset frequency (such as per minute or per hour) to obtain real-time data of different operating parameters, so as to obtain multiple sampling data of the battery cell stack in each statistical cycle (such as daily).

[0017] It should be noted that in other alternative examples, the preset frequency can be every 30 seconds, every 3 hours, or every day, and the statistical period can be every 6 hours, a daily cycle (a cycle of one day), or a cycle of 7 days, but is not limited to these.

[0018] Step S102: Determine the statistical data of the battery cell stack based on multiple sampling data of the battery cell stack. The statistical data includes the mean and standard deviation of different operating parameters.

[0019] Optionally, if different operating parameters include voltage, temperature, and internal resistance, then the statistical data for the cell stack includes: Voltage-related: The mean voltage μV and the standard deviation voltage σV are calculated based on real-time data from multiple voltage samplings of the battery cell stack within each statistical period; Temperature-related parameters: The mean temperature μT and the standard deviation of temperature σT are calculated based on real-time data from multiple temperature samplings of the battery cell stack within each statistical period. Internal resistance: The mean value μR and the standard deviation σR of the internal resistance are calculated based on real-time data of multiple samplings of the internal resistance of the battery cell stack within each statistical period.

[0020] Step S103: Configure the mean and standard deviation of the dynamic threshold model corresponding to each of the different operating parameters based on the mean and standard deviation of the different operating parameters. The dynamic threshold model includes dynamic coefficients corresponding to the standard deviation of the different operating parameters.

[0021] For example, and not as a limitation, a dynamic threshold model can be described by a threshold interval [a, b], where a is the dynamic lower threshold and b is the dynamic upper threshold. Where μ is defined as the mean of the operating parameters, k as the dynamic coefficient corresponding to the standard deviation of the operating parameters, and σ as the standard deviation of the operating parameters, then the dynamic lower threshold satisfies the dynamic lower threshold acquisition formula, which is: μ-k·σ; the dynamic upper threshold satisfies the dynamic upper threshold acquisition formula, which is: μ+k·σ.

[0022] Specifically, to distinguish the mean μ, standard deviation σ, and corresponding dynamic coefficient k of different operating parameters, the mean μ, standard deviation σ, and corresponding dynamic coefficient k of voltage, temperature, and resistance can be as follows: The mean voltage is denoted as μV, the standard deviation of voltage is denoted as σV, and the dynamic coefficient corresponding to the standard deviation of voltage is denoted as kV; The mean temperature is denoted as μT, the standard deviation of temperature is denoted as σT, and the dynamic coefficient corresponding to the standard deviation of temperature is denoted as kT. The mean value of the internal resistance is denoted as μR, the standard deviation of the temperature is denoted as σR, and the dynamic coefficient corresponding to the standard deviation of the resistance is denoted as kR.

[0023] Therefore, the threshold ranges for voltage, temperature, and resistance can be respectively: The threshold range corresponding to the voltage can be: [μV-kV·σV, μV+kV·σV]; The threshold range corresponding to the temperature can be: [μT-kT·σT, μT+kT·σT]; The threshold range corresponding to the internal resistance can be: [μR-kR·σR, μR+kR·σR].

[0024] Where k is a dynamic coefficient, and its initial value can be preset to 1.2~1.7, but is not limited to this. For example, but not as a limitation, the initial value of k for voltage, temperature and internal resistance is set to 1.5, and can be adjusted later according to the error rate.

[0025] Step S104: Based on the dynamic threshold models corresponding to the different configured operating parameters and the sampling data of the cell stack in each statistical period, determine the abnormal state type of the cell stack.

[0026] Optionally, step S104 may include steps S11 and S12.

[0027] Step S11: Based on the threshold intervals [a, b] corresponding to the different configured operating parameters and the sampling data of the cell stack in the next statistical period, obtain the judgment result of whether the different operating parameters of the cell stack are abnormal.

[0028] Optionally, step S11 includes steps S11-1 and S11-2.

[0029] Step S11-1: Determine whether the result of the judgment of whether the operating parameters that deviate from the corresponding threshold range in each sampling data of the cell stack in the next statistical period are abnormal.

[0030] It should be noted that if the real-time voltage data is denoted as Vt, the real-time temperature data as Tt, and the real-time internal resistance data as Rt, then the operating parameters that satisfy the deviation of the real-time data from the corresponding threshold range can be: the voltage that satisfies the preset voltage condition, the temperature that satisfies the preset temperature condition, and the internal resistance that satisfies the preset internal resistance condition.

[0031] Preset voltage conditions: Vt>μV+kV·σV or Vt<μV-kV·σV.

[0032] Preset temperature conditions: Tt>μT+kT·σT or Tt<μT-kT·σT.

[0033] Preset internal resistance condition: Rt>μR+kR·σR or Rt<μR-kR·σR.

[0034] Prioritize that when the operating parameters deviate from the corresponding threshold range in real-time data, a supervised learning model can be further used to determine whether the operating parameters exhibit a continuous abnormal trend. If so, the result of determining whether the operating parameters are abnormal is considered abnormal. If not, the result of determining whether the operating parameters are abnormal is considered normal. For example, and not as a limitation, this supervised learning model is, for instance, a random forest model.

[0035] Step S11-2: Determine whether the result of the judgment of whether the operating parameters that meet the real-time data falling within the corresponding threshold range in each sampling data of the cell stack in the next statistical period are abnormal or normal.

[0036] Step S12: Based on the judgment results of whether different operating parameters of the battery cell stack are abnormal, determine the abnormal state type of the battery cell stack.

[0037] Optionally, step S12 may include steps S12-1 and S12-2.

[0038] Step S12-1: When only one of the different operating parameters of the battery cell stack is abnormal, the abnormal state type is determined to be a single-dimensional abnormality.

[0039] For example, and not as a limitation, when the judgment result of whether the voltage of the battery cell stack is abnormal is abnormal, the judgment result of whether the temperature of the battery cell stack is abnormal is normal, and the judgment result of whether the internal resistance of the battery cell stack is abnormal is normal, the abnormal state type of the battery cell stack is determined to be voltage drift type.

[0040] For example, and not as a limitation, when the judgment result of whether the voltage of the battery cell stack is abnormal is normal, the judgment result of whether the temperature of the battery cell stack is abnormal, and the judgment result of whether the internal resistance of the battery cell stack is abnormal is normal, the abnormal state type of the battery cell stack is determined to be thermal imbalance type.

[0041] For example, and not as a limitation, when the judgment result of whether the voltage of the battery cell stack is abnormal is normal, the judgment result of whether the temperature of the battery cell stack is abnormal is normal, and the judgment result of whether the internal resistance of the battery cell stack is abnormal, the abnormal state type of the battery cell stack is determined to be internal resistance degradation type.

[0042] Step S12-2: When more than one of the different operating parameters of the battery cell stack is abnormal, the abnormal state type is determined to be a composite abnormality.

[0043] For example, and not limited to, when the judgment result of at least two of the voltage, temperature and internal resistance is abnormal, the abnormal state type of the cell stack is determined to be a composite abnormality type.

[0044] For example, and not limited to, clustering algorithms (such as K-Means) can be used to automatically classify the abnormal state types of the battery cell stack based on the judgment results of whether different operating parameters of the battery cell stack are abnormal, so as to determine the abnormal state type of the battery cell stack.

[0045] Step S105: Invoke the corresponding optimization strategy based on the determined abnormal state type of the battery cell stack, and optimize and adjust the battery cell stack based on the corresponding optimization strategy.

[0046] Optionally, step S105 may include steps S21 and S22.

[0047] Step S21: If the abnormal state type of the battery cell stack is a single-dimensional abnormality, then the corresponding single-dimensional optimization strategy is invoked. The single-dimensional optimization strategy includes adjusting the operating parameters related to the abnormal operating parameters of the battery cell stack.

[0048] Optionally, step S21 may include steps S21-1, S21-2, and S21-3.

[0049] Step S21-1: If the abnormal state type of the cell stack is voltage drift, then the voltage optimization strategy is invoked. The voltage optimization strategy includes reducing the power output ratio of the cell stack in the liquid-cooled battery.

[0050] In step S21-1, by reducing the power output ratio of the cell stack in the liquid-cooled battery, load fluctuations caused by excessively high or low voltage can be avoided.

[0051] Step S21-2: If the abnormal state type of the battery cell stack is thermal imbalance, then the temperature optimization strategy is invoked. The temperature optimization strategy includes adjusting the flow rate of the liquid cooling channel corresponding to the battery cell stack.

[0052] Optionally, in step S21-2, the flow rate of the liquid cooling channel corresponding to the battery cell stack is adjusted to match the required cooling efficiency of the battery cell stack. Optionally, the temperature of the battery cell stack can be balanced with the temperature of other battery cell stacks.

[0053] Step S21-3: If the abnormal state type of the cell stack is internal resistance degradation, then the internal resistance optimization strategy is invoked. The internal resistance optimization strategy includes reducing the proportion of the cell stack in the charge and discharge current of the liquid-cooled battery.

[0054] In step S21-3, reducing the proportion of the charge / discharge current of the cell stack in the liquid-cooled battery can prevent accelerated aging. Optionally, the internal resistance growth trend can be accumulated, and if the internal resistance growth trend exceeds a preset growth threshold, a maintenance prompt is triggered. For example, and not limited to, the increase in internal resistance relative to the initial value can be accumulated, and if the increase exceeds the corresponding preset growth threshold, a maintenance prompt is triggered.

[0055] Step S22: If the abnormal state type of the battery cell stack is a composite abnormality, then the corresponding comprehensive optimization strategy is invoked. The comprehensive optimization strategy includes the coordinated adjustment of the working parameters related to different operating parameters of the battery cell stack.

[0056] Optionally, in step S22, a comprehensive optimization strategy can be implemented, which can simultaneously consider voltage, temperature, and internal resistance factors to avoid side effects caused by single regulation (such as a sudden increase in internal resistance caused by excessively rapid cooling).

[0057] Optionally, step S22 includes step S22-1.

[0058] Step S22-1: If the abnormal state type of the cell stack is a composite abnormality, then the comprehensive optimization strategy is invoked. The comprehensive optimization strategy includes reducing the power output ratio of the cell stack in the liquid-cooled battery, adjusting the flow rate of the liquid-cooled channel corresponding to the cell stack, and reducing the charge and discharge current ratio of the cell stack in the liquid-cooled battery.

[0059] Optionally, the comprehensive optimization strategy may also include achieving a globally optimal balance among multiple cell stacks. Achieving a globally optimal balance among multiple cell stacks includes: Based on the temperature difference between the battery cell stack and adjacent battery cell stacks, the coolant flow distribution is balanced to reduce the temperature difference between the battery cell stack and adjacent battery cell stacks; The total output current is redistributed based on the internal resistance of the battery cell stack, so that the low internal resistance battery cell stack can handle more power output.

[0060] In this embodiment, unlike the limitations of fixed thresholds in the prior art, there is no need to rely on fixed thresholds. A dynamic threshold model can adapt to individual differences in different cell stacks and environmental changes. This reduces the misjudgment rate, lowers maintenance costs caused by inaccurate anomaly identification, and improves the operational safety and service life of the liquid-cooled battery system.

[0061] [Example 2: Key Monitoring List Example] After acquiring the sampling data of the cell stack at a preset frequency in the next statistical period, the battery management method further includes steps S201, S202, S203, S204 and S205.

[0062] Step S201: Obtain the dimensionless deviation of different operating parameters relative to the corresponding threshold interval in each sampling data of the battery cell stack.

[0063] Optionally, the dimensionless deviation can be obtained according to the dimensionless deviation acquisition formula. For any operating parameter of the battery cell stack, the dimensionless deviation acquisition steps include steps S31 to S35.

[0064] Step S31: Obtain the interval center value of the corresponding threshold interval according to the interval center value acquisition formula. The interval center value acquisition formula is C=(a+b) / 2, where C is the interval center value.

[0065] Step S32: Obtain the interval half-width value of the corresponding threshold interval according to the interval half-width value acquisition formula. The interval half-width value acquisition formula is Rw=(ba) / 2, where Rw is the interval half-width value.

[0066] Step S33: Obtain the absolute value of the difference between the real-time data of the operating parameter and the center value of the interval in each sampling data of the cell stack.

[0067] Step S34: In response to the fact that the absolute value of the difference between the real-time data of the operating parameter and the center value of the interval is not greater than the half width of the interval, the dimensionless deviation of the operating parameter relative to the corresponding threshold interval is determined to be 0.

[0068] Step S35: In response to the absolute value of the difference between the real-time data of the operating parameter and the center value of the interval being greater than the half-width of the interval, the dimensionless deviation of the operating parameter relative to the corresponding threshold interval is determined to be the absolute value of the difference between the real-time data of the operating parameter and the center value of the interval.

[0069] Step S202: Based on the dimensionless deviation of different operating parameters relative to the corresponding threshold interval and the preset weight coefficients corresponding to different operating parameters in each sampling data, obtain the sum score characterizing the ease with which anomalies occur in each cell stack.

[0070] Optionally, a comprehensive score representing the likelihood of an anomaly occurring in each cell stack can be obtained using a comprehensive score acquisition formula. The comprehensive score acquisition formula is as follows: Sraw = wV × dV + wT × dT + wR × dR Wherein, Sraw is the sum score characterizing the ease with which anomalies occur in each cell stack, wV is the weight corresponding to the dimensionless deviation of voltage, dV is the dimensionless deviation of voltage relative to the corresponding threshold interval in each sampled data in the next statistical period, wT is the weight corresponding to the dimensionless deviation of temperature, dT is the dimensionless deviation of temperature relative to the corresponding threshold interval in each sampled data in the next statistical period, wR is the weight corresponding to the dimensionless deviation of internal resistance, and dR is the dimensionless deviation of internal resistance relative to the corresponding threshold interval in each sampled data in the next statistical period.

[0071] Optionally, wT can be greater than wV and greater than wR, so that the effect of temperature on the safety of the cell stack is greater than that of voltage and internal resistance.

[0072] For example, and not as a limitation, the preset weighting coefficients can be set to wV=0.3, wT=0.4, wR=0.3, but are not limited to these. In other alternative examples, other preset weighting coefficients can be set as needed.

[0073] Optionally, to avoid misjudgment caused by fluctuations in a single measurement, a sliding weighted average or exponential smoothing algorithm is used to smooth Sraw, and the current smoothed score St is used as the comprehensive score in subsequent steps.

[0074] Step S203: Based on the sum score characterizing the likelihood of an anomaly occurring in each cell stack, the preset comprehensive score admission threshold, and the preset comprehensive score exit threshold, determine whether the operating parameters of the currently sampled cell stack meet the conditions for adding the cell stack to the list of key monitoring options.

[0075] Optionally, the preset comprehensive score admission threshold is greater than the preset comprehensive score exit threshold. Further, the preset comprehensive score admission threshold can be 0.5 to 0.7, but is not limited to this. The preset comprehensive score exit threshold can be 0.2 to 0.4, but is not limited to this. For example, and not as a limitation, the preset comprehensive score admission threshold can be 0.6, and the preset comprehensive score exit threshold can be 0.3.

[0076] If the sum of the scores representing the likelihood of an anomaly occurring in each cell stack is not less than the preset comprehensive score threshold, it is determined that the operating parameters of the cell stack sampled in the current sampling meet the conditions for adding the cell stack to the list of key monitoring options.

[0077] If the sum of the scores representing the likelihood of an anomaly occurring in each battery cell stack is not greater than the preset comprehensive score exit threshold, it is determined that the operating parameters of the battery cell stack sampled in the current sampling do not meet the conditions for adding the battery cell stack to the list of key monitoring options.

[0078] Step S204: Add the battery cell stack to the key monitoring list if the number of times the operating parameters of the battery cell stack meet the conditions for adding the battery cell stack to the key monitoring list reaches the preset access threshold.

[0079] Step S205: If the number of times the operating parameters of the battery cell stack do not meet the conditions for adding the battery cell stack to the key monitoring list reaches a preset exit threshold, the battery cell stack is removed from the key monitoring list.

[0080] Optionally, the preset entry threshold can be lower than the preset exit threshold to reflect the principle of prevention-oriented and cautious exit in battery safety management. On the one hand, this lowers the threshold for abnormal cell stacks to enter the key monitoring unit, and on the other hand, it raises the exit threshold to prevent misjudgments caused by a single or a few drops in sampling data. It ensures that the key monitoring is lifted only after the cell stack's operating status has truly returned to normal and stable, thereby improving the effective utilization of monitoring resources and reducing safety risks.

[0081] For example, but not as a limitation, the preset threshold for the number of times an entry is allowed can be 3, and the preset threshold for the number of times an exit is allowed can be 5, but it is not limited to these.

[0082] Step S206: In response to the fact that the battery cell stack is on the key monitoring list, implement a differentiated control strategy for the battery cell stack.

[0083] The differentiated management and control strategy is as follows.

[0084] Sampling side: Increase the data sampling frequency of the battery cell stack to a second sampling frequency, which is greater than the first sampling frequency of the battery cell stack. The first sampling frequency is the preset frequency before the sampling frequency was increased.

[0085] For example, and not as a limitation, the data sampling frequency can be increased to twice the first sampling frequency to achieve high-frequency monitoring.

[0086] Temperature control side: Prioritize the allocation of coolant flow to this cell stack.

[0087] For example, rather than limiting it, the coolant flow rate is preferentially allocated to the cell stack to improve local heat exchange efficiency.

[0088] On the power control side: current limiting or peak shaving adjustments are performed on the battery cell stack during power distribution.

[0089] For example, rather than limiting it, current limiting or peak clipping adjustments can be introduced in power distribution to slow down the heat buildup in the battery cell stack.

[0090] This mechanism allows for targeted monitoring of battery cell stacks that are more prone to anomalies, enabling timely identification and targeted handling.

[0091] [Example 3: Dynamic Coefficient Feedback Optimization Example] Optionally, the battery management method further includes steps S301 to S303.

[0092] Step S301: The judgment results of whether the running parameters are abnormal in the historical statistical period and the corresponding sampling data used for judgment are associated and stored to form a historical dataset.

[0093] Step S302: Receive modifications from staff based on their judgments of whether the operating parameters in the historical dataset are abnormal, in order to determine the actual misjudgment rate of the corresponding operating parameters.

[0094] Optionally, the actual misclassification rate can be the actual false positive rate. For example, and not restrictively, if a staff member confirms that an anomaly in the historical dataset is a genuine anomaly (e.g., through subsequent detection or analysis), it is counted as a correct judgment; if a staff member confirms that an anomaly in the historical dataset is a false positive (there is actually no anomaly or the anomaly is negligible, but it is incorrectly judged as an anomaly), it is counted as an incorrect judgment. The actual misclassification rate P_actual is the ratio of the number of incorrect judgments to the total number of judgments.

[0095] Step S303: Adjust the corresponding dynamic coefficients based on the misclassification rate of the corresponding operating parameters and the preset target misclassification rate.

[0096] Optionally, the actual false positive rate includes the actual false alarm rate, and the preset target false positive rate may include the preset target false alarm rate. Then, the corresponding dynamic coefficient can be adjusted according to the adaptive adjustment formula of the first dynamic coefficient, which is as follows: k_new = k_old + β·(target false positive rate - actual false positive rate) Where β is the sensitivity adjustment coefficient (e.g., 0.05~0.2).

[0097] Optionally, the actual false positive rate may include the actual false negative rate, and the preset target false positive rate may include the preset target false negative rate. Then, the corresponding dynamic coefficient can be adjusted according to the adaptive adjustment formula of the second dynamic coefficient, which is as follows: k_new = k_old - β·(target false negative rate - actual false negative rate) Using the above method, if the false alarm rate is high, the value of k is appropriately increased to relax the judgment threshold; if the false negative rate is high, the value of k is decreased to enhance the anomaly sensitivity. Dynamic coefficient adaptive optimization is achieved through steps S301 to S303.

[0098] [Experimental Verification] In the experiment, using method A (traditional fixed threshold method) and method B (battery management method of this application), the experimental results show that: Compared with Method A, Method B reduces the false alarm rate by approximately 40% and the false negative rate by approximately 30% in identifying abnormal cell stacks. In Method B, the average temperature difference between the battery cells decreased by 2.5°C.

[0099] The above results verify the advantages of the proposed method in terms of anomaly identification accuracy and system security improvement.

[0100] [Battery Management System Example] This application also provides a battery management system; please refer to [link / reference]. Figures 2-3 The battery management system 10 includes a liquid-cooled battery 100, multiple sampling units 200, an optimization and adjustment unit 300, and a control system 400.

[0101] The liquid-cooled battery 100 includes multiple cell stacks 110, which can be connected in series, parallel, or mixed to form a power supply circuit for the liquid-cooled battery 100. Each cell stack 110 may include one or more individual cells. When the cell stack 110 includes multiple individual cells, the individual cells can be connected in series, parallel, or mixed.

[0102] Multiple sampling units 200 are respectively configured for multiple cell stacks 110, and are used to sample real-time data of different operating parameters of the multiple cell stacks 110. For example, but not limited to, the sampling unit 200 may include a voltage sensor, a temperature sensor, and an internal resistance detection module, used to monitor real-time data of the voltage, temperature, and internal resistance of the cell stack 110, respectively. Optionally, the temperature sensor may include a thermistor or a thermocouple, and the internal resistance detection module may detect the internal resistance of the cell stack 110 by AC injection or DC discharge.

[0103] The optimization and adjustment unit 300 is used to optimize and adjust multiple cell stacks 110. For example, and not limited to, the optimization and adjustment unit 300 may include a power distribution module, a thermal management module, and a charge / discharge control module. The power distribution module is used to adjust the proportion of the power output of each cell stack 110 in the total power output of the liquid-cooled battery 100; the thermal management module is used to adjust the cooling efficiency of each cell stack 110, for example, by adjusting the coolant flow rate or fan speed; the charge / discharge control module is used to adjust the proportion of the charge / discharge current of each cell stack 110 in the total charge / discharge current of the liquid-cooled battery 100.

[0104] The control system 400 is communicatively connected to the sampling unit 200 and the optimization adjustment unit 300 to control the sampling unit 200 and the optimization adjustment unit 300 to implement the battery management method of this application. The control system 400 may include a microprocessor, a memory, and a communication interface. The memory is used to store computer programs, collected data, and statistical data. The microprocessor is used to execute the computer programs to implement the battery management method of this application. The communication interface is used to interact with the sampling unit 200 and the optimization adjustment unit 300.

[0105] [Media Example] This application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method as described in any of the above embodiments.

[0106] The above are merely embodiments of this application and do not limit the scope of this patent application. Any equivalent structural or procedural changes made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of this application.

Claims

1. A battery management method, characterized in that, The battery management method is used to manage a liquid-cooled battery, the liquid-cooled battery comprising multiple cell stacks, and the battery management method includes: Multiple sampling data of the battery cell stack are acquired at a preset frequency within each statistical period, and the sampling data includes real-time data of different operating parameters; The statistical data of the battery cell stack is determined based on multiple sampling data of the battery cell stack, and the statistical data includes the mean and standard deviation of different operating parameters; Based on the mean and standard deviation of different operating parameters, the dynamic threshold model corresponding to each operating parameter is configured with the mean and standard deviation of each operating parameter. The dynamic threshold model includes dynamic coefficients corresponding to the standard deviation of different operating parameters. Based on the dynamic threshold model corresponding to each of the different configured operating parameters and the sampling data of the battery cell stack in the next statistical period, the abnormal state type of the battery cell stack is determined. Based on the determined abnormal state type of the battery cell stack, the corresponding optimization strategy is invoked to optimize and adjust the battery cell stack accordingly.

2. The battery management method according to claim 1, characterized in that, The dynamic threshold model is described by a threshold interval [a, b], where a is the dynamic lower threshold and b is the dynamic upper threshold. Wherein, if μ is defined as the mean value corresponding to the operating parameter, k is the dynamic coefficient corresponding to the standard deviation of the operating parameter, and σ is the standard deviation of the operating parameter, then the dynamic lower limit threshold satisfies the dynamic lower limit threshold acquisition formula, which is: μ-k·σ; the dynamic upper limit threshold satisfies the dynamic upper limit threshold acquisition formula, which is: μ+k·σ.

3. The battery management method according to claim 2, characterized in that, Based on the dynamic threshold models corresponding to different configured operating parameters and the sampling data of the cell stack in each statistical period, the abnormal state type of the cell stack is determined, including: Based on the threshold intervals [a, b] corresponding to the different configured operating parameters and the sampling data of the battery cell stack in the next statistical period, the judgment result of whether the different operating parameters of the battery cell stack are abnormal is obtained; Based on the judgment results of whether different operating parameters of the battery cell stack are abnormal, the abnormal state type of the battery cell stack is determined.

4. The battery management method according to claim 3, characterized in that, The method of obtaining the judgment result of whether the different operating parameters of the battery cell stack are abnormal based on the threshold interval [a, b] corresponding to the different configured operating parameters and the sampling data of the battery cell stack in the next statistical period includes: The result of determining whether the operating parameter that deviates from the corresponding threshold range in each sampled data of the battery cell stack in the next statistical period is abnormal is defined as abnormal. The judgment result of whether the operating parameter that meets the requirement of real-time data falling within the corresponding threshold range in each sample data of the battery cell stack in the next statistical period is abnormal is determined as normal.

5. The battery management method according to claim 4, characterized in that, The abnormal state type of the battery cell stack is determined based on the judgment results of whether different operating parameters of the battery cell stack are abnormal, including: When only one of the different operating parameters of the battery cell stack is abnormal, the abnormal state type is determined to be a single-dimensional abnormal type. When more than one of the different operating parameters of the battery cell stack is abnormal, the abnormal state type is determined to be a composite abnormality type. The step of invoking the corresponding optimization strategy based on the determined abnormal state type of the battery cell stack, and optimizing and adjusting the battery cell stack based on the corresponding optimization strategy, includes: If the abnormal state type of the battery cell stack is the single-dimensional abnormal type, then the corresponding single-dimensional optimization strategy is invoked. The single-dimensional optimization strategy includes adjusting the working parameters related to the operating parameters of the battery cell stack that are abnormal. If the abnormal state type of the battery cell stack is a composite abnormality, then the corresponding comprehensive optimization strategy is invoked. The comprehensive optimization strategy includes the coordinated adjustment of the working parameters related to different operating parameters of the battery cell stack.

6. The battery management method according to claim 5, characterized in that, The different operating parameters include voltage, temperature, and internal resistance; the single-dimensional anomaly types are classified into voltage drift type, thermal imbalance type, and internal resistance degradation type; when only one of the different operating parameters of the battery cell stack is abnormal, the anomaly state type is determined to be a single-dimensional anomaly type, including: When the result of the judgment on whether the voltage of the battery cell stack is abnormal is abnormal, the result of the judgment on whether the temperature of the battery cell stack is abnormal is normal, and the result of the judgment on whether the internal resistance of the battery cell stack is abnormal is normal, the abnormal state type of the battery cell stack is determined to be voltage drift type. When the result of judging whether the voltage of the battery cell stack is abnormal is normal, the result of judging whether the temperature of the battery cell stack is abnormal, and the result of judging whether the internal resistance of the battery cell stack is abnormal is normal, the abnormal state type of the battery cell stack is determined to be thermal imbalance type. When the result of judging whether the voltage of the battery cell stack is abnormal is normal, the result of judging whether the temperature of the battery cell stack is abnormal is normal, and the result of judging whether the internal resistance of the battery cell stack is abnormal, the abnormal state type of the battery cell stack is determined to be internal resistance deterioration type. When more than one of the different operating parameters of the battery cell stack is abnormal, the abnormal state type is determined to be a composite abnormality type. If the judgment result of at least two of the voltage, temperature and internal resistance is abnormal, the abnormal state type of the cell stack is determined to be a composite abnormality type.

7. The battery management method according to claim 6, characterized in that, If the abnormal state type of the battery cell stack is the single-dimensional abnormal type, then the corresponding single-dimensional optimization strategy is invoked. The single-dimensional optimization strategy includes adjusting the operating parameters related to the operating parameters of the battery cell stack that are abnormal, including: If the abnormal state type of the cell stack is voltage drift type, then the voltage optimization strategy is invoked, which includes reducing the power output ratio of the cell stack in the liquid-cooled battery. If the abnormal state type of the battery cell stack is thermal imbalance, then the temperature optimization strategy is invoked, which includes adjusting the flow rate of the liquid cooling channel corresponding to the battery cell stack. If the abnormal state type of the cell stack is the internal resistance degradation type, then the internal resistance optimization strategy is invoked. The internal resistance optimization strategy includes reducing the proportion of the charge and discharge current of the cell stack in the liquid-cooled battery. If the abnormal state type of the battery cell stack is a composite abnormality, then a comprehensive optimization strategy is invoked. This comprehensive optimization strategy includes the coordinated adjustment of the operating parameters related to different operating parameters of the battery cell stack, including: If the abnormal state type of the cell stack is the composite abnormal type, then a comprehensive optimization strategy is invoked. The comprehensive optimization strategy includes reducing the power output ratio of the cell stack in the liquid-cooled battery, adjusting the flow rate of the liquid-cooled channel corresponding to the cell stack, and reducing the charge and discharge current ratio of the cell stack in the liquid-cooled battery.

8. The battery management method according to claim 1, characterized in that, After acquiring sampling data of the battery cell stack at a preset frequency each time in the next statistical period, the battery management method further includes: Obtain the dimensionless deviation of different operating parameters relative to the corresponding threshold interval in each sampled data of the battery cell stack; Based on the dimensionless deviation of the different operating parameters relative to the corresponding threshold interval of each sampling data and the preset weight coefficients corresponding to the different operating parameters, a total score representing the ease with which the cell stack will experience an anomaly each time is obtained; Based on the sum score characterizing the likelihood of an anomaly occurring in each battery cell stack, a preset comprehensive score admission threshold, and a preset comprehensive score exit threshold, it is determined whether the operating parameters of the battery cell stack in the current sampling meet the conditions for adding the battery cell stack to the list of key monitoring options. The battery cell stack is added to the key monitoring list if the number of times the operating parameters of the battery cell stack meet the conditions for adding the battery cell stack to the key monitoring list reaches a preset access threshold. If the number of times the operating parameters of the battery cell stack fail to meet the conditions for adding the battery cell stack to the key monitoring list reaches a preset exit threshold, the battery cell stack will be removed from the key monitoring list; wherein, the preset entry threshold is less than the preset exit threshold. In response to the battery cell stack being included in the key monitoring list, a differentiated control strategy is implemented for the battery cell stack.

9. The battery management method according to claim 1, characterized in that, The battery management method includes: The results of judging whether the operating parameters are abnormal or not in the historical statistical period and the corresponding sampling data used for judgment are associated and stored to form a historical dataset; The system receives modifications from staff based on their judgments of whether the operating parameters in the historical dataset are abnormal, in order to determine the actual misjudgment rate of the corresponding operating parameters. The corresponding dynamic coefficients are adjusted based on the actual misjudgment rate and the preset target misjudgment rate of the corresponding operating parameters.

10. A battery management system, characterized in that, The battery management system includes: Liquid-cooled battery, consisting of multiple cell stacks; Multiple sampling units are set up, each corresponding to a multiple battery cell stack, to monitor different operating parameters of the multiple battery cell stacks in order to obtain real-time data of the different operating parameters of the multiple battery cell stacks; An optimization and adjustment unit is used to optimize and adjust the multiple cell stacks. The system is also connected to the sampling unit and the optimization adjustment unit, respectively, to control the sampling unit and the optimization adjustment unit to implement the battery management method according to any one of claims 1-9.

Images