Modular battery pack equalization circuit control method

Abstract

The present application relates to the technical field of battery management, in particular to a modular battery PACK equalization circuit control method, the method comprising: collecting operating parameters to calculate actual available capacity and normalized residual energy indicators, and smoothing and filtering to obtain stabilized normalized residual energy indicators; taking the actual available capacity as the weight to calculate the capacity weighted equalization reference value, and making a difference to obtain the equalization deviation to determine the modules participating in the equalization; calculating the target equalization current amplitude based on the equalization deviation, the actual available capacity and the equalization time window, combining the temperature correction coefficient to generate the instruction current to drive active equalization; calculating the internal resistance relative deviation and the equalization protection trigger upper limit based on the single cell internal resistance, and triggering the equalization suspension protection and recalculating the reference value when the upper limit is exceeded. The present application solves the problem that the equalization target does not match the actual available energy under the health state differentiation working condition, and prolongs the whole group operation life.

Description

Technical Field

[0001] This invention relates to the field of battery management technology, and more specifically to a modular battery pack equalization circuit control method. Background Technology

[0002] The construction and promotion of large-scale energy storage systems are a core support for the current energy transition. In highly dynamic application scenarios such as mobile energy storage and temporary power supply for engineering construction, battery systems generally adopt a multi-module parallel or series PACK architecture, which combines several battery clusters into a battery stack to meet high power output requirements. After such modular PACK systems are put into operation, they require a matching balancing circuit to adjust the charge of each module to prevent premature failure of the entire system due to parameter deviations between individual cells or modules.

[0003] Common battery balancing technologies can be divided into two categories: dissipative balancing and non-dissipative balancing. Passive balancing achieves overall balance by consuming the energy of high-capacity individual cells, which is reliable and effective but wasteful of energy. Active balancing, while achieving energy transfer, has a relatively complex circuit structure, and the balancing control struggles to balance speed and stability. Chinese patent application CN103532189A discloses a battery pack balancing control system and method based on dynamic balancing points. The system includes a battery pack, a voltage sampling circuit, a balancing main circuit, and a balancing control circuit. The balancing control circuit compares the voltage values ​​of individual cells. When the voltage difference between different cells exceeds a preset threshold, it adjusts the frequency of the power switch corresponding to each cell, controlling the power switch frequency to track the balancing frequency of the battery pack in a dynamic balancing state. Most balancing control strategies are based on PFM control modes based on battery voltage differences. However, due to the lack of a defined frequency value, magnetic component design is difficult, and balancing performance is limited by hardware implementation, restricting the adaptability of balancing strategies under complex operating conditions.

[0004] The mainstream control method for existing equalization circuits uses the state of charge (SOC) as the sole equalization target. This involves detecting the SOC value of each module and driving a passive discharge circuit or an active inductor / capacitor transfer circuit to equalize the SOC of each module. Some more advanced solutions introduce a voltage difference threshold. Equalization is initiated when the voltage difference between individual cells within a cluster triggers an alarm, and it stops when the voltage difference returns to the normal range. The entire control logic uses equal SOC or equal voltage as the convergence criterion.

[0005] However, after the modular PACK system has been in operation for a period of time, the health of each module will vary due to differences in the thermal environment of their installation location, differences in cumulative charge and discharge depth, and frequent high-power throughput. The rated capacity degradation rate of different modules will not be consistent. At this point, the control method that uses uniform SOC as the balancing target has a fundamental flaw: when a module with an SOH of 80% and a module with an SOH of 100% are balanced to the same SOC reading, their actual stored physical charge is not equal; the usable energy of the former is only 80% of that of the latter. When the system operates under high-power discharge conditions, the low-SOH module will reach the cutoff voltage first due to insufficient actual usable capacity, triggering the over-discharge protection of that cluster, and subsequently causing a protective power-off of the entire group. At this time, the high-SOH module still has a large amount of residual energy that has not been released, resulting in the actual usable energy of the entire group being far lower than the sum of the capacities of each module. As the operating cycle extends, the SOH differentiation continues to deepen, and the aforementioned premature power-off phenomenon becomes more frequent, severely compressing the effective discharge window of the system. Summary of the Invention

[0006] To address the existing technical problem of low-health modules triggering over-discharge protection first and high-health modules having their remaining energy shelved due to mismatch in balancing targets under different health status conditions, this invention provides a modular battery pack balancing circuit control method.

[0007] This invention provides a modular battery pack equalization circuit control method, employing the following technical solution: S1. Collect operating parameters at three levels: battery stack, battery cluster, and individual cells. Each battery cluster is considered a module. Operating parameters include cluster-level state of charge (SOC), cluster-level health status (SHS), average cluster temperature, and individual cell internal resistance. Calculate the actual usable capacity based on the cluster-level SHS and rated capacity. Combine this with the cluster-level SOC to calculate the normalized remaining energy index. Smooth the normalized remaining energy index to obtain a stable normalized remaining energy index. S2. Using the actual usable capacity as the weight, perform a weighted calculation on the stable normalized remaining energy index of each effective module to obtain a capacity-weighted equilibrium benchmark value. Subtract the stable normalized remaining energy index from the capacity-weighted equilibrium benchmark value. S3. Obtain the balance deviation and determine the effective modules participating in the balance based on the balance deviation; S4. For the effective modules participating in the balance, calculate the target balance current amplitude based on the absolute value of the balance deviation, the actual available capacity, and the balance time window, and determine the temperature correction coefficient based on the average cluster temperature. Combine the target balance current amplitude with the temperature correction coefficient to generate the command current to drive the active balance output; S5. During the balance output period, calculate the relative deviation of the internal resistance and the upper limit of the balance protection trigger based on the internal resistance of the individual unit. In response to the relative deviation of the internal resistance exceeding the upper limit of the balance protection trigger, trigger the balance suspension protection of the corresponding module, remove the module from the effective modules, and recalculate the capacity-weighted balance reference value.

[0008] By replacing the traditional state of charge with normalized residual energy as the balancing target, the balancing benchmark truly reflects the physical energy storage level of each module under its current aging state, eliminating the contradiction between the balancing target and the actual available energy under different health conditions. Using the capacity-weighted average as the overall balancing benchmark, modules with higher health and larger capacity contribute more to the benchmark value, avoiding the problem of aging modules lowering the benchmark and healthy modules being over-discharged, effectively extending the overall operational life of the group.

[0009] Preferably, the actual available capacity is calculated based on the cluster-level health status value and the rated capacity, and the normalized remaining energy index is calculated in combination with the cluster-level state of charge value, including: multiplying the cluster-level health status value by the rated capacity to obtain the actual available capacity; and multiplying the cluster-level state of charge value by the cluster-level health status value to finally obtain the normalized remaining energy index.

[0010] Preferably, the normalized residual energy index is smoothed and filtered to obtain a stable normalized residual energy index, including: obtaining the normalized residual energy index within multiple consecutive sampling periods; taking the moving average of the normalized residual energy index within a set smoothing window to finally obtain the stable normalized residual energy index.

[0011] Preferably, the stabilized and normalized residual energy index of each effective module is weighted and calculated to obtain the capacity weighted equilibrium benchmark value, including: calculating the sum of the products of the stabilized and normalized residual energy index and the actual available capacity; calculating the sum of the actual available capacity of all effective modules; dividing the sum of the products by the sum of the actual available capacity to finally obtain the capacity weighted equilibrium benchmark value.

[0012] By embedding the actual health differences of each module into the formation process of the benchmark value, the equilibrium objective shifts from statistical consistency to optimal allocation of available energy in a physical sense. This ensures that the dynamic correspondence between the overall equilibrium benchmark and the actual schedulable energy can be maintained even during the operation phase when the differentiation of health status continues to deepen.

[0013] Preferably, the equilibrium deviation is obtained by subtracting the stabilized normalized residual energy index from the capacity-weighted equilibrium benchmark value, and the modules participating in the equilibrium are determined based on the equilibrium deviation. This includes: subtracting the capacity-weighted equilibrium benchmark value from the stabilized normalized residual energy index to obtain the equilibrium deviation; sorting all modules in descending order of the absolute value of the equilibrium deviation; and selecting the module with the largest absolute value as the final module to participate in the equilibrium.

[0014] Preferably, the target equalization current amplitude is calculated based on the absolute value of the equalization deviation, the actual available capacity, and the equalization time window, including: obtaining the remaining duration of the current charge / discharge cycle and the rated equalization response duration; selecting the smaller value between the remaining duration of the current charge / discharge cycle and the rated equalization response duration as the equalization time window; multiplying the absolute value of the equalization deviation by the actual available capacity and then dividing by the equalization time window to finally obtain the target equalization current amplitude.

[0015] By combining the normalized energy deviation with the available capacity of the module to derive the equalization current amplitude, while ensuring that the equalization current is accurately matched with the actual transfer requirements of each module, the equalization current magnitude protection is inherently applied to modules with low health, thus avoiding overload of absolute charge transfer.

[0016] Preferably, the temperature correction coefficient is determined based on the average cluster temperature, including: when the average cluster temperature is within the normal operating range, setting the temperature correction coefficient to a fixed value that does not decay; when the average cluster temperature is higher than the upper limit of the normal operating range, controlling the temperature correction coefficient to decrease linearly to zero; when the average cluster temperature is lower than the lower limit of the normal operating range, controlling the temperature correction coefficient to decrease linearly based on the difference between the average cluster temperature and the low-temperature cutoff charging temperature, and finally obtaining the temperature correction coefficient.

[0017] Preferably, after combining the target equalization current amplitude with the temperature correction coefficient to generate the command current to drive the active equalization output, the method further includes: calculating the ratio of the intra-group range to the group mean of the stabilized and normalized residual energy index of the entire group to obtain the relative range; recording the relative range sequence for multiple consecutive equalization cycles; when the decrease in the relative range sequence within a consecutive cycle is less than a set proportion of the historical mean of the relative range sequence, it is determined that the equalization has reached the convergence plateau and exits the current round of equalization.

[0018] By calculating the relative range of the stable normalized residual energy index of the entire group and monitoring whether its downward trend tends to stagnate during the continuous equilibrium period, an adaptive convergence determination mechanism that does not rely on absolute value thresholds is provided. It can accurately identify the equilibrium entering the plateau period and exit in time, avoiding invalid equilibrium and energy oscillation. At the same time, it is applicable to systems of different sizes and different degrees of differentiation, without the need to manually set a fixed convergence threshold.

[0019] Preferably, the calculation of the relative deviation of internal resistance and the upper limit of equalization protection triggering based on the internal resistance of individual cells includes: within the static acquisition window before equalization starts, calculating the arithmetic mean of the internal resistance of all individual cells in the module to obtain the baseline mean of the internal resistance of the module, and calculating the static fluctuation standard deviation of internal resistance based on the module internal resistance mean sequence at each sampling time within the static window; after equalization starts, calculating the arithmetic mean of the current internal resistance of all individual cells in the module to obtain the module internal resistance mean at the current time; dividing the difference between the internal resistance mean and the baseline mean by the baseline mean to finally obtain the relative deviation of internal resistance.

[0020] By establishing individualized baseline mean and standard deviation of internal resistance based on the static window before equalization startup, and calculating the change in internal resistance by combining the rated self-heating temperature rise caused by equalization current, the upper limit of equalization protection trigger can adapt to modules with different aging levels and impedance characteristics. When the internal resistance rises abnormally beyond the range of normal thermal effects and measurement noise, protection is triggered in a timely manner, effectively identifying fault risks such as loose connections and local lithium plating, and preventing abnormal modules from continuously participating in equalization and causing safety hazards.

[0021] Preferably, the calculation of the relative deviation of internal resistance and the upper limit of equalization protection based on the internal resistance of a single unit further includes: calculating the change in internal resistance corresponding to the rated self-heating temperature rise based on the command current, the estimated value of thermal resistance and the internal resistance temperature coefficient; summing the change in internal resistance with the standard deviation of the static fluctuation of internal resistance by a preset multiple; and dividing the summation result by the baseline mean to finally obtain the upper limit of equalization protection.

[0022] By dynamically binding the equalization protection trigger threshold with the standard deviation of static internal resistance fluctuation and rated self-heating capacity of each module, the internal resistance anomaly monitoring has adaptive sensitivity for modules with different aging levels and different internal resistance bases. When connection abnormalities or local impedance changes occur, the abnormal module can be switched off in time while maintaining the continuity of the equalization process of the remaining modules.

[0023] The present invention has the following technical effects: 1. The equalization current amplitude of this invention is objectively derived from the deviation, module available capacity, and temperature ledger parameters. Low-temperature cutoff and high-temperature protection conditions are both read from the manufacturer's ledger parameters, eliminating adaptation errors introduced by global empirical settings. The internal resistance protection trigger threshold is dynamically calculated based on the static fluctuation standard deviation of each module and its rated self-heating capacity, enabling adaptation for modules with different aging levels and different internal resistance substrates, avoiding continuous injection of equalization current into cells with connection failures or lithium plating risks.

[0024] 2. The battery management system uses the ledger parameters and real-time collected data as the sole input driver. All thresholds and limits are read from the physical parameters written into the ledger by the manufacturer. No global empirical constants are introduced. It can be directly applied to modular PACK systems with different battery models and different aging levels without the need to recalibrate control parameters for specific deployment scenarios. Attached Figure Description

[0025] Figure 1 This is a flowchart of a modular battery PACK equalization circuit control method according to the present invention.

[0026] Figure 2 This is a comparison of the convergence process of the relative range of normalized residual energy within a group as a function of the number of equilibrium cycles under two equilibrium methods. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] This invention discloses a modular battery pack equalization circuit control method, referring to... Figure 1 This includes the following steps: S1. Collect parameters and calculate the normalized residual energy index.

[0029] In an optional embodiment, the BMS reads operating parameters at three levels—the battery stack, battery cluster, and individual cells—with a 1-second acquisition cycle. The acquired data includes the average cluster temperature. Cluster-level state of charge Cluster-level health status value and the internal resistance of each individual unit ,in, This is the battery cluster number. This refers to the individual cell number. The three-level parallel reporting acquisition architecture enables the BMS to simultaneously sense the stack-level power status, cluster-level energy deviation, and individual cell-level impedance anomalies, providing complete parameters for subsequent calculations. Each battery cluster is recorded as a module.

[0030] Based on the collected With rated capacity Calculate the first The actual available capacity of each module:

[0031] in, For the first The actual available capacity of each module; For the first Cluster-level health status values ​​for each module; For the first The rated capacity of each module.

[0032] in The cluster-level health status value field reported by BMS is used in the calculation as a percentage. The rated capacity of the module at the time of manufacture, in Ah, is permanently stored in the BMS ledger and is written by the manufacturer along with the factory inspection report. It cannot be overwritten during system operation.

[0033] Furthermore, based on the calculation of the actual available capacity, the cluster-level state of charge is combined to calculate the first... Normalized residual energy index of each module :

[0034] in, For the first Normalized residual energy index of each module This represents the current cluster-level state of charge value for this module. For the corresponding cluster-level health status values, both are multiplied as percentages. This reflects the module's true state of charge after deducting aging losses; that is, the ratio of the module's current actual remaining energy to its rated capacity. Even if two modules in different health states show the same state of charge reading, their... The value will also be affected by The differences in energy storage status between modules can be directly reflected in the control layer, providing a physically meaningful input for the calculation of the capacity weighted equilibrium benchmark.

[0035] To avoid misjudgment of equalization due to transient noise from a single sampling, the normalized residual energy index is calculated for a continuous preset number of sampling periods. The moving average is used for smoothing filtering to obtain the stable normalized residual energy index. This serves as the input for subsequent equilibrium calculations. For example, the preset number can be 5 sampling periods. It should be noted that there are many other methods in the art that can implement short-term smoothing filtering. Those skilled in the art can choose different smoothing schemes such as exponentially weighted moving average or Kalman filtering according to actual needs. The above-described moving average window is merely an example.

[0036] In this way, by collecting data at three levels and normalizing the definition of the remaining energy index, the actual energy state of each module under different aging conditions is uniformly mapped to a comparable quantity of the same dimension. This eliminates the false equality caused by differences in health status when directly comparing the state of charge, and provides a physically consistent data foundation for subsequent capacity-weighted benchmark calculations.

[0037] S2, Calculate the capacity weighted benchmark value to determine the equilibrium module.

[0038] In an optional embodiment, all online and normal clusters in the system are selected for the load balancing calculation, excluding modules in maintenance, charging-discharge-prohibited, discharging-prohibited, or fault states. Let there be a total of... One effective module. Based on the actual usable capacity of each module. Using the weights, the stable normalized residual energy index of the effective modules in the system is weighted and calculated to obtain the capacity-weighted equilibrium benchmark value for the entire group. :

[0039] in, For the first The stabilized and normalized residual energy index of each module. For the first The actual available capacity of each module, with the denominator being the sum of the available capacities of all valid modules. This represents the number of valid modules.

[0040] Calculated Then, the difference between the stabilized and normalized residual energy index and the capacity-weighted equilibrium benchmark value is calculated, and the equilibrium deviation is obtained for each module:

[0041] in, For the first The balance deviation of each module For the first The stabilized and normalized residual energy index of each module. This serves as the capacity-weighted equilibrium benchmark value for the entire group.

[0042] This indicates that the normalized remaining energy of this module is higher than the baseline, indicating that it is on the balanced release side; This indicates that the module is below the baseline, and is on the equalization charging side. Further, the modules participating in the equalization are determined based on the equalization deviation; specifically, all modules are charged according to the absolute value of the equalization deviation. The modules with the largest absolute values ​​are selected from the largest to the smallest and prioritized for participation in this round of equilibration. This is to narrow the energy distribution range within the group as quickly as possible and reduce the risk of local over-discharge or over-charge caused by uneven energy distribution over a long period of time.

[0043] In this way, by replacing the traditional arithmetic mean with a capacity-weighted mean, the actual health differences of each module are embedded into the formation process of the benchmark value, so that the equilibrium goal shifts from statistical consistency to optimal allocation of available energy in a physical sense. Thus, even in the operation phase where the differentiation of health status continues to deepen, the dynamic correspondence between the overall equilibrium benchmark and the actual schedulable energy can still be maintained.

[0044] S3. Calculate the equalization current amplitude and generate the command current.

[0045] In an optional embodiment, for the modules participating in the equalization process, the target equalization current amplitude is calculated based on the absolute value of the equalization deviation, the actual available capacity, and the equalization time window. This amplitude is determined by the deviation. With respect to the actual available capacity of the module Joint decision:

[0046] in, For the first The target equalization current amplitude of each module, in amperes (A); For the first The absolute value of the balance deviation of each module; For the first The actual available capacity of each module, in Ah; To balance the time window, the unit is hours (h). Multiply The normalized energy deviation is then restored to the absolute amount of charge that this module actually needs to transfer under the current health state, in Ah; and then divided by the equalization time window. This yields the average current amplitude required to complete the transfer, expressed in amperes (A). The value is taken as the smaller of the remaining duration of the current charge / discharge cycle and the rated equalization response time issued by the system scheduling layer, to ensure that the equalization process can be completed before the next charge / discharge state switch and does not cross the operating condition boundary.

[0047] Furthermore, calculations were performed. Then, the temperature correction factor is determined based on the average cluster temperature. The target equalization current amplitude is combined with the temperature correction coefficient to generate the command current. To drive the active equalization output:

[0048] in, For the first The command current of each module, in amperes (A); For the first The target equalization current amplitude of each module, in amperes (A); This is the temperature correction factor.

[0049] The average cluster temperature of the current module The decision will be made to process the issue in three stages. When... When operating within the normal temperature range of 10℃ to 45℃, the temperature correction factor is set to a fixed value that does not decay. The equalization current does not attenuate. When When the temperature exceeds the upper limit of the normal operating range of 45℃, control the temperature correction factor. The energy transfer to the module is paused when the temperature linearly decreases from 45℃ to 55℃ and then to zero, to prevent high-temperature charging from accelerating damage to the positive electrode structure. 55℃ is the upper limit of the high-temperature cutoff charging temperature for this type of battery recorded in the battery management system ledger, which can be read directly from the ledger parameters without manual setting.

[0050] when When the normal operating temperature is below 10°C, the charging component in the equalization current poses a risk of lithium plating, based on the average cluster temperature and the low-temperature cutoff charging temperature. The difference, control the temperature correction factor From 10°C to the low-temperature cutoff charging temperature The temperature correction factor decreases linearly within the interval, ultimately yielding the temperature correction coefficient:

[0051] in, This is a temperature correction factor; For the first Average cluster temperature of each module, in °C; The low-temperature cutoff charging temperature of this battery model, recorded in the battery management system ledger, is expressed in °C.

[0052] when hour, The equalization charging action of this module is completely suspended. The corresponding low-temperature rated charging rate limit is read from the ledger and written into the ledger by the battery manufacturer along with the factory parameters, ensuring that the low-temperature equalization current limit strictly corresponds to the physical characteristics of the specific battery model, without introducing any global empirical constant.

[0053] After generating the command current to drive the active equalization output, it is necessary to continuously monitor the equalization convergence state. The stable normalized residual energy index of the entire group needs to be calculated. Within-group range with group mean The ratio of the two values ​​yields the relative range. ,in , The system records the relative range sequence over multiple consecutive equilibrium periods. When the decrease in the relative range sequence within a consecutive period is less than a preset convergence threshold of the historical mean of the relative range sequence, the system determines that equilibrium has reached a convergence plateau and exits the current equilibrium cycle. For example, this preset convergence threshold can be 5%. This criterion reflects the stagnation of the range narrowing trend, rather than the magnitude of the absolute value of the range. Therefore, it is adaptable to systems of different sizes and varying degrees of health differentiation, without requiring manual setting of the absolute value of the convergence threshold.

[0054] Energy management system based on A PWM duty cycle modulation signal is sent to the active equalization circuit corresponding to each cluster. The equalization circuit adopts an inductive transfer topology, and realizes inter-cluster charge transfer through bidirectional DC-DC switches. The switching frequency is fixed, and the amplitude of the transfer current is linearly controlled by adjusting the duty cycle to achieve equalization. Precise tracking.

[0055] In this way, by combining the normalized energy deviation with the available capacity of the module to derive the equalization current amplitude, and superimposing the temperature segmentation correction coefficient directly read from the manufacturer's ledger, the equalization current is accurately matched with the actual transfer requirements of each module, while incorporating the risk of low-temperature lithium plating and the risk of high-temperature electrode damage into the automatic constraint of the current limit, without the need for manual calibration of the temperature protection setting value for each model.

[0056] S4. Calculate the deviation based on the internal resistance of the individual unit to trigger protection.

[0057] In an optional embodiment, during equalization output, the battery management system continuously reports the internal resistance of each cell at 1-second intervals. This value is calculated using the voltage response pulse at the instant of charging / discharging current switching. The relative deviation of the internal resistance and the upper limit of the equalization protection trigger are calculated based on the individual unit's internal resistance. Specifically, this includes: within the static acquisition window before equalization activation, the value is calculated for the first... The internal resistance of each individual cell within a module Calculate the arithmetic mean to obtain the baseline mean of the internal resistance of the module. The standard deviation of the internal resistance fluctuation during rest is calculated based on the mean sequence of the module's internal resistance at each sampling time within the resting window. This serves as an individualized reference base for subsequent internal resistance anomaly judgment. The static acquisition window is a preset number of sampling periods before equalization is started; for example, this preset number can be 10 sampling periods.

[0058] After balanced startup, for the first The current internal resistance of all individual cells within each module. Calculate the arithmetic mean to obtain the average internal resistance of the module at the current time. The difference between the mean internal resistance and the baseline mean is divided by the baseline mean to obtain the relative deviation of the internal resistance. :

[0059] in, For the first The relative deviation of the internal resistance of each module; For the first The average internal resistance of each module at the current moment, in mΩ; For the first The baseline average internal resistance of each module, in mΩ.

[0060] Furthermore, based on command current The estimated thermal resistance of this battery model is pre-stored in the battery management system ledger. and internal resistance temperature coefficient Calculate the change in internal resistance corresponding to the rated self-heating temperature rise. The unit is mΩ; the change in internal resistance is based on the formula. Calculation, where , The preset temperature rise integral time; where, For the first The baseline mean of the internal resistance of each module, The internal resistance temperature coefficient of this battery model is pre-stored in the battery management system ledger. This is the estimated thermal resistance value for this battery model that is pre-stored in the ledger. For the first The command current of each module; the standard deviation of the internal resistance static fluctuation relative to the internal resistance change by a preset multiple. Summate; divide the sum by the baseline mean. Ultimately, the upper limit for triggering the balance protection is obtained. :

[0061] in, For the first The upper limit of balanced protection triggering for each module; For the first The change in internal resistance of each module under the rated self-heating temperature rise, in mΩ; This is a preset multiple; For the first Standard deviation of static internal resistance fluctuation of each module, in mΩ; For the first The baseline mean internal resistance of each module, in mΩ. For example, a preset multiple. 3 is an acceptable value. This represents the reasonable increase in internal resistance caused by normal, balanced self-heating. This represents the statistical tolerance based on the static fluctuation characteristics of this module.

[0062] Response to relative deviation of internal resistance Exceeding the balance protection trigger limit within a single balance cycle The system determines that the increase in the module's internal resistance has exceeded the superposition range of normal self-heating and measurement noise, triggering the equalization pause protection of the corresponding module: immediately disconnecting the equalization circuit of the module from the transfer topology, and simultaneously reporting a custom alarm to the energy management system. The alarm content is written into the anomaly field of the message, and the field value is the measured value. After the capacity balance protection is triggered, the abnormal module is removed and the capacity weighted balance benchmark value is recalculated. The remaining modules continue to be balanced until the entire group is stabilized and the relative range of the normalized residual energy index distribution is determined. Once the convergence platform condition is met, the current round of balancing ends, and the controller enters the next data acquisition and judgment cycle.

[0063] In this way, by dynamically binding the equalization protection trigger threshold with the standard deviation of static internal resistance fluctuation and rated self-heating capacity of each module, the internal resistance anomaly monitoring has adaptive sensitivity for modules with different aging levels and different internal resistance bases. When connection abnormalities or local impedance changes occur, the abnormal module can be switched off in time and the equalization process of the remaining modules can be maintained, avoiding cascading failures caused by the anomaly of a single module leading to the interruption of the equalization of the entire group.

[0064] The technical effects of this invention can also be illustrated in conjunction with the accompanying drawings. Figure 2 This is a comparison of the convergence process of the relative range of normalized residual energy within a group as a function of the number of equilibrium cycles under the two equilibrium methods of this invention. The horizontal axis represents the number of equilibrium cycles (dimensionless integer index), and the vertical axis represents the relative range of normalized residual energy within a group (%). The figure uses two curves to represent the convergence evolution trajectory of the relative range of normalized residual energy of each module within a group as the equilibrium process progresses under the same health state differentiation scenario, using the existing method and the method of this invention, respectively. The existing method has a slower convergence speed and a larger final residual relative range, indicating that the equilibrium method targeting the state of charge is difficult to achieve effective uniformity of energy distribution under health state differentiation conditions. The method of this invention has a significantly larger convergence slope and a significantly lower final relative range, indicating that the capacity-weighted normalized energy equilibrium method proposed in this invention can make the energy distribution of the entire group tend to be uniform in fewer equilibrium cycles, effectively extending the overall system lifespan.

[0065] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A modular battery pack equalization circuit control method, characterized in that, include: S1. Collect operating parameters at three levels: battery stack, battery cluster, and individual cell. Each battery cluster is considered a module. Operating parameters include cluster-level state of charge (SOC), cluster-level health status (SHS), average cluster temperature, and individual cell internal resistance. Calculate the actual usable capacity based on the cluster-level SHS and rated capacity. Combine this with the cluster-level SOC to calculate the normalized residual energy index. Smooth the normalized residual energy index to obtain a stable normalized residual energy index. S2. Using the actual usable capacity as the weight, perform a weighted calculation on the stable normalized residual energy index of each effective module to obtain a capacity-weighted equilibrium benchmark value. Subtract the stable normalized residual energy index from the capacity-weighted equilibrium benchmark value to obtain the equilibrium deviation. Determine the effective modules participating in the equilibrium based on the equilibrium deviation. S3. For the effective modules participating in the equilibrium, calculate the target equilibrium current amplitude based on the absolute value of the equilibrium deviation, the actual usable capacity, and the equilibrium time window. Determine the temperature correction coefficient based on the average cluster temperature. Combine the target equilibrium current amplitude with the temperature correction coefficient to generate a command current to drive the active equilibrium output. S4. During the equalization output period, the relative deviation of the internal resistance and the upper limit of the equalization protection are calculated based on the internal resistance of the individual unit. In response to the relative deviation of the internal resistance exceeding the upper limit of the equalization protection, the equalization suspension protection of the corresponding module is triggered, and the module is removed from the effective modules. The capacity weighted equalization benchmark value is then recalculated.

2. The modular battery pack equalization circuit control method according to claim 1, characterized in that, The actual available capacity is calculated based on cluster-level health status values ​​and rated capacity, and the normalized remaining energy index is calculated in conjunction with cluster-level state of charge values, including: Multiply the cluster-level health status value by the rated capacity to obtain the actual available capacity; Multiplying the cluster-level state of charge value by the cluster-level health value yields the normalized residual energy index.

3. The modular battery pack equalization circuit control method according to claim 1, characterized in that, The normalized residual energy index is smoothed and filtered to obtain a stable normalized residual energy index, including: Obtain the normalized residual energy index over multiple consecutive sampling periods; Within a set smoothing window, the normalized residual energy index is averaged to obtain the final stable normalized residual energy index.

4. The modular battery pack equalization circuit control method according to claim 1, characterized in that, The stabilized and normalized residual energy index of each effective module is weighted and calculated to obtain the capacity-weighted equilibrium benchmark value, including: Calculate the sum of the products of the stabilized normalized residual energy index and the actual available capacity; Calculate the sum of the actual available capacity of all valid modules; Dividing the sum of the products by the sum of the actual available capacities yields the capacity-weighted equilibrium benchmark value.

5. The modular battery pack equalization circuit control method according to claim 1, characterized in that, The equilibrium deviation is obtained by subtracting the stabilized and normalized residual energy index from the capacity-weighted equilibrium benchmark value. Based on the equilibrium deviation, the modules participating in the equilibrium are determined, including: Subtracting the capacity-weighted equilibrium benchmark value from the stabilized normalized residual energy index yields the equilibrium deviation. All modules are sorted from largest to smallest based on the absolute value of the balance deviation; The module with the largest absolute value is selected as the module to participate in the equilibrium process.

6. The modular battery pack equalization circuit control method according to claim 1, characterized in that, The target equalization current amplitude is calculated based on the absolute value of the equalization deviation, the actual available capacity, and the equalization time window, including: Obtain the remaining duration of the current charge / discharge cycle and the rated equalization response time; The smaller value between the remaining duration of the current charge / discharge cycle and the rated equalization response duration is selected as the equalization time window; Multiply the absolute value of the equalization deviation by the actual available capacity, and then divide by the equalization time window to obtain the target equalization current amplitude.

7. The modular battery pack equalization circuit control method according to claim 1, characterized in that, The temperature correction factor is determined based on the average cluster temperature, including: When the average cluster temperature is within the normal operating range, the temperature correction factor is set to a fixed value that does not decay. When the average cluster temperature is higher than the upper limit of the normal operating range, the temperature correction coefficient is linearly reduced to zero. When the average cluster temperature is lower than the lower limit of the normal operating range, the temperature correction coefficient is controlled to decrease linearly based on the difference between the average cluster temperature and the low-temperature cutoff charging temperature, and finally the temperature correction coefficient is obtained.

8. The modular battery pack equalization circuit control method according to claim 1, characterized in that, After combining the target equalization current amplitude with the temperature correction coefficient to generate the command current to drive the active equalization output, the following steps are also included: The relative range is obtained by calculating the ratio of the within-group range to the group mean of the stabilized normalized residual energy index of the whole group. Record the relative range sequence of multiple consecutive equilibrium periods; When the decrease in the relative range sequence within a continuous period is less than a set proportion of the historical mean of the relative range sequence, it is determined that the equilibrium has reached the convergence plateau and exits the current equilibrium.

9. The modular battery pack equalization circuit control method according to claim 1, characterized in that, The relative deviation of internal resistance and the upper limit of equalization protection triggering are calculated based on the internal resistance of a single unit, including: Within the static acquisition window before equalization startup, the arithmetic mean of the internal resistance of all individual cells in the module is calculated to obtain the baseline mean of the internal resistance of the module. The static fluctuation standard deviation of the internal resistance is calculated based on the sequence of the mean internal resistance of the module at each sampling time within the static window. After equalization startup, the arithmetic mean of the current internal resistance of all individual cells in the module is calculated to obtain the mean internal resistance of the module at the current moment. The difference between the mean internal resistance and the mean baseline is divided by the mean baseline to obtain the relative deviation of the internal resistance.

10. The modular battery pack equalization circuit control method according to claim 1, characterized in that, The calculation of the relative deviation of internal resistance and the upper limit of equalization protection triggering based on the internal resistance of a single unit also includes: Calculate the change in internal resistance corresponding to the rated self-heating temperature rise based on the command current, thermal resistance estimate, and internal resistance temperature coefficient. The change in internal resistance is summed with the standard deviation of the static fluctuation of internal resistance by a preset multiple; Dividing the summation result by the baseline mean yields the final upper limit for equalization protection triggering.

Images