Battery module active equalization method and system based on clustering

Abstract

The invention relates to the field of battery equalization, in particular to a clustering-based battery pack active equalization method and system, and the method comprises the steps: carrying out the uniform-temperature and voltage-limiting charging of all battery modules in the same battery cluster under the conditions that the charging of the battery cluster is close to cut-off and the overcharge protection of a single battery is not achieved, carrying out the voltage sampling at the level of the battery modules, and carrying out the equalization of the battery pack. Obtaining a module voltage sampling data set; performing clustering matching on the battery modules according to the module voltage sampling data set, and performing active equalization among the battery modules according to a matching result; according to the invention, clustering matching is carried out on the battery modules, and active equalization between the battery modules is carried out according to the matching result, so that more effective active equalization between the battery modules is realized.

Description

[0001] This divisional application is based on the mother patent of an invention patent with an application date of December 4, 2024, an application number of 202411765286.2, and a title of "A Clustering-Based Battery Pack Balancing Method and System". Technical Field

[0002] The present invention relates to the field of new energy energy storage battery balancing, and particularly to an active balancing method and system for battery modules based on clustering. Background Art

[0003] In a battery energy storage system, due to inevitable series applications, environmental factors, as well as process and material reasons in battery manufacturing, there will be consistency differences among battery monomers. This problem is an inherent problem of energy storage carried by electrochemical batteries. There have been numerous studies on improving the consistency among battery monomers from the manufacturing process perspective, controlling environmental influencing factors at the application level, and various battery balancing methods, including active or passive balancing, balancing circuits, balancing strategies, etc. By deeply analyzing the reasons for the inconsistency of lithium-ion batteries and various balancing variables, parameters that can reflect consistency such as temperature, SOC, SOH, internal resistance, and voltage are the focus of current research in this field.

[0004] The judgment of the consistency of lithium batteries relies on accurate SOC calculation, and this parameter is related to temperature. In general SOC estimation, the influence of temperature is calibrated and corrected, increasing the workload of calibration tests and calculations; SOH is a parameter on a long time scale and does not match the control period of balancing. The consistency correction based on SOH cannot achieve ideal results in a short time; the AC internal resistance of the battery can reflect the changes in characteristics during battery material aging, and the DC internal resistance is also related to the state of charge to a certain extent, but factors such as temperature change, discharge plateau period, and production process make it impossible to distinguish SOC differences with high resolution through internal resistance; in the most commonly used balancing strategies, the consistency is judged by the monomer voltage. The open-circuit voltage OCV reflects the charge state of the material from the electrochemical principle, but for lithium-ion batteries with lithium iron phosphate as the positive electrode material, the OCV difference in the middle region of SOC is not obvious. In order to capture slight voltage changes, the battery often needs to be static for a long time, which is difficult to achieve in the case of pursuing the energy storage utilization rate of lithium batteries; the judgment based on the operating moment voltage OV avoids the voltage sampling waiting time but is easily affected by the fluctuation of the operating current.

[0005] With the increase in the number of lithium batteries connected in series, the problem of balancing efficiency gradually emerges. Active balancing has become a research hotspot due to its high balancing efficiency, leading to numerous research topics on active balancing circuits. Applying an active balancing circuit to a single cell increases the risk of circuit device failure and affects the battery. The single-cell low-voltage converter has a negative impact on both efficiency and cost, and also increases the complexity of the circuit, thereby reducing reliability. Passive balancing, with its small current and high reliability, is still widely used. However, when the SOC difference is large, especially when there is imbalance between modules, the balancing time will be very long, and control methods such as the range method and the switching method are often used, which cannot fully utilize the efficiency of cluster balancing. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a clustering-based active balancing method and system for battery modules to achieve more effective active balancing between battery modules.

[0007] To solve the above technical problem, the technical solution adopted by the present invention is as follows:

[0008] A clustering-based battery pack balancing method includes the steps of:

[0009] S1. Under the condition that the charging of the battery cluster is approaching the cut-off and the overcharge protection of the single cell has not been reached, perform temperature-equalizing and voltage-limiting charging on all battery modules within the same battery cluster, and perform voltage sampling at the battery module level and the cell level respectively to obtain a module voltage sampling data set and a single cell voltage sampling data set;

[0010] S2. According to the module voltage sampling data set, cluster and group the battery modules, and perform active balancing between the battery modules according to the grouping result;

[0011] According to the single cell voltage sampling data set of each battery module, cluster and group the single cells within the battery module, and perform passive balancing between the single cells according to the grouping result.

[0012] A clustering-based active balancing method for battery modules includes the steps of:

[0013] S1. Under the condition that the charging of the battery cluster is approaching the cut-off and the overcharge protection of the single cell has not been reached, perform temperature-equalizing and voltage-limiting charging on all battery modules within the same battery cluster, and perform voltage sampling at the battery module level to obtain a module voltage sampling data set;

[0014] S2. According to the module voltage sampling data set, cluster and group the battery modules, and perform active balancing between the battery modules according to the grouping result, including the steps of:

[0015] S21a. Calculate the balancing voltage judgment threshold according to the module voltage sampling data set;

[0016] The equalization voltage judgment threshold is used to characterize the voltage imbalance degree threshold between modules;

[0017] S22a. Sort the voltages of the k-th sampling in the module voltage sampling dataset, calculate the range and the mean value, and judge whether active equalization between modules is required according to the range and the equalization voltage judgment threshold. If active equalization is required, proceed to step S23a;

[0018] S23a. Perform clustering and grouping according to the mean value, and perform active equalization between the battery modules according to the grouping result.

[0019] To solve the above technical problems, another technical solution adopted by the present invention is:

[0020] A clustering-based battery pack equalization system, where each battery cluster in the system includes multiple battery modules;

[0021] Each of the battery modules includes multiple battery cells, and is provided with a bidirectional active equalization converter and a battery cell management unit. The bidirectional active equalization converter is used to realize the bidirectional flow of equalization energy between the low-voltage auxiliary power supply and the battery module, so as to realize active equalization;

[0022] The battery cell management unit includes a microprocessor and an analog front end, and the battery cell management unit is communicatively connected to the battery module management unit for receiving control instructions. The analog front end is used to realize passive equalization between single battery cells;

[0023] Based on the above structure, the battery cluster implements the steps in the above-mentioned clustering-based battery pack equalization method.

[0024] A clustering-based battery pack equalization system, where each battery cluster in the system includes multiple battery modules;

[0025] Each of the battery modules includes multiple battery cells, and is provided with a bidirectional active equalization converter and a battery cell management unit. The bidirectional active equalization converter is used to realize the bidirectional flow of equalization energy between the low-voltage auxiliary power supply and the battery module, so as to realize active equalization;

[0026] Based on the above structure, the battery cluster implements the steps in the above-mentioned clustering-based battery module active equalization method.

[0027] The beneficial effects of the present invention are as follows: The clustering-based battery module active equalization method and system of the present invention eliminate the influence of temperature, perform clustering of battery modules and battery cells according to voltage, and perform active equalization between battery modules according to the clustering result, so as to realize more effective active equalization between battery modules. Description of the Drawings

[0028] Figure 1 It is a flowchart of a clustering-based battery pack group balancing method according to an embodiment of the present invention;

[0029] Figure 2 It is a structural diagram of a clustering-based battery pack group balancing system according to an embodiment of the present invention. Detailed Embodiments

[0030] Please refer to Figure 1 , a clustering-based battery pack group balancing method, including the steps:

[0031] S1. Under the condition that the charging of the battery cluster is approaching the cut-off and the overcharge protection of the single battery is not reached, perform temperature-equalizing and voltage-limiting charging on all battery modules within the same battery cluster, and perform voltage sampling at the battery module level and the cell level respectively to obtain a module voltage sampling data set and a single-cell voltage sampling data set;

[0032] S2. Cluster and group the battery modules according to the module voltage sampling data set, and perform active balancing between the battery modules according to the grouping result;

[0033] Cluster and group the single cells within the battery module according to the single-cell voltage sampling data set of each battery module, and perform passive balancing between the single cells according to the grouping result.

[0034] As can be seen from the above description, the beneficial effect of the present invention is that: a clustering-based battery pack group balancing method of the present invention eliminates the influence of temperature, clusters the battery modules and battery single cells according to voltage, and performs active balancing between the battery modules and passive balancing between the single cells within the battery module according to the clustering result, realizing more effective battery balancing.

[0035] Further, step S1 includes the steps:

[0036] S11. Collect the voltages and charging currents of N battery modules within the same battery cluster. When the voltage of a certain module reaches the first charging preset voltage, reduce the charging current I C to a preset ratio of the previous moment, and at the same time start the single-cell temperature-equalizing management;

[0037] S12. Every time the voltage of a module reaches the first charging preset voltage, reduce the charging current I C to a preset ratio of the previous moment until the following conditions are met:

[0038]

[0039] Perform the first synchronous sampling on the voltages of the N battery modules;

[0040] where V m_max is the highest voltage of each battery module in the battery cluster, ΔT is the temperature range difference of M single battery cells, and ΔT limit is the temperature range limit value, C1 is the preset equalizing temperature and limiting voltage charging rate, and I CN is the rated charging current;

[0041] S13. Starting from the first synchronous sampling moment, maintain the charging current I C unchanged and perform continuous k synchronous samplings until the voltage of any module reaches the second preset charging voltage V δ2 :

[0042] V m_max ≥Vδ2;

[0043] Obtain the module voltage sampling data set;

[0044] S14. Record the energy charged into the module with the highest voltage during the k samplings:

[0045]

[0046] S15. During the kth sampling, collect the voltages of all single battery cells in each of the battery modules to obtain the single battery cell voltage sampling data set.

[0047] As can be seen from the above description, through the single battery cell equalizing temperature management, the equalizing temperature and limiting voltage charging of the battery modules are realized, the temperature influence is eliminated, and voltage sampling is performed under the condition of equalizing temperature and limiting voltage charging.

[0048] Furthermore, in step S2, according to the module voltage sampling data set, the battery modules are clustered and grouped, and active equalization is performed among the battery modules according to the grouping result, including the steps:

[0049] S21a. Calculate the equalizing voltage judgment threshold according to the module voltage sampling data set;

[0050] S22a. Sort the voltages of the kth sampling in the module voltage sampling data set, calculate the range and the mean value, and judge whether active equalization among the modules is required according to the range and the equalizing voltage judgment threshold. If active equalization is required, go to step S23a;

[0051] S23a. Perform clustering and grouping according to the mean value, and perform active equalization among the battery modules according to the grouping result.

[0052] As described above, according to the range and the equilibrium voltage judgment threshold, it is judged whether active balancing is required between battery modules. In the case where active balancing is required, clustering and grouping are performed according to the average value of the voltages, so as to determine each battery module with a voltage lower than and higher than the average value, and then active balancing is performed.

[0053] Further, step S21a is specifically as follows:

[0054] Respectively take the maximum values in the first sampling data set V and the k-th sampling data set V M_1 and V M_k in the module voltage sampling data set, and the difference between the two is used as the equilibrium voltage judgment threshold ε M :

[0055] ε M = max(V M_k ) - max(V M_1 ).

[0056] As described above, according to the above steps, the equilibrium voltage judgment threshold is determined.

[0057] Further, step S22a is specifically as follows:

[0058] Sort the voltages of the k-th sampling in the module voltage sampling data set and calculate the range. If the range is greater than the equilibrium voltage judgment threshold ε M , it is considered that one-time active balancing between modules is required, and step S23a is entered.

[0059] As described above, it is determined whether the range of the voltage of the k-th sampling is greater than the equilibrium voltage judgment threshold to confirm whether active balancing between battery modules is required.

[0060] Further, step S23a is specifically as follows:

[0061] Calculate the average value V m_k_avg of the voltages of the k-th sampling in the module voltage sampling data set, and divide the data set into two subsets V M_k = V ML_ ∪V MU_ ;

[0062] Among them, all voltage values less than or equal to the average value V m_k_avg are classified into V ML_ , and vice versa into V MU_ ;

[0063] Control the two-way active balancing converter of the battery module corresponding to the maximum value V MU_ in the data set V mu__ to start, and perform energy conversion from the battery module to the auxiliary power supply direction for a duration of TMU is:

[0064]

[0065] wherein, V m_k_min is the minimum value in the dataset V ML_ , s1 is the adjustment coefficient, and P1 is the constant power of the bidirectional active equalization converter;

[0066] Calculate the dataset T formed by the turn-on times of the bidirectional active equalization converters of the battery modules corresponding to each value in the dataset V ML_ ; ML ;

[0067] wherein, for the j-th value among the total h values in V ML_ , the turn-on time of the bidirectional active equalization converter of the corresponding battery module:

[0068]

[0069] Obtain the dataset T ML , sort the data in the dataset T ML , and sequentially turn on the bidirectional active equalization converters of the corresponding battery modules in order, perform energy conversion from the auxiliary power supply to the module direction, and work for the corresponding duration until all the bidirectional active equalization converters of the battery modules corresponding to the data in the dataset T ML have completed working.

[0070] As can be seen from the above description, the battery module with the maximum voltage discharges through the bidirectional active equalization converter, and each battery module with a voltage lower than the average value receives the electric energy released by the battery module with the maximum voltage one by one, so that the input and output energy are kept balanced.

[0071] Furthermore, according to the single-cell voltage sampling dataset of each battery module, the single cells in the battery module are clustered and grouped, and passive equalization between the single cells is performed according to the grouping result, including the steps of:

[0072] S21b. For each battery module, calculate the range and mean according to the single-cell voltage sampling dataset, and sort them;

[0073] S22b. When the range is greater than the preset equalization threshold, obtain all the battery cells with a voltage greater than the mean, and perform discharge passive equalization.

[0074] As can be seen from the above description, it is judged whether equalization between battery cells is needed according to the range, and clustering and grouping are realized according to the mean to obtain all the battery cells with a voltage greater than the mean, and discharge passive equalization is performed.

[0075] Furthermore, the discharge passive equalization is specifically as follows:

[0076] Apply different PWM controls to each of the battery cells. The duty cycle d of the PWM is proportional to the degree of deviation of the voltage of the battery cell from the average value. For the j-th battery cell in the battery module that needs to be equalized, there is d j :

[0077]

[0078] wherein, V c_j represents the voltage of the j-th battery cell in the battery module that needs to be equalized, V c_avg represents the average value of the monomer voltage sampling data set, and V c_max represents the maximum value in the monomer voltage sampling data set.

[0079] As can be seen from the above description, by applying different PWM controls to the battery cells and adjusting the duty cycle, passive equalization is achieved.

[0080] Furthermore, it further includes the step:

[0081] S3. Repeatedly execute step S1 and step S2 until the equalization ends.

[0082] As can be seen from the above description, by repeatedly executing until the equalization ends, the effectiveness of the equalization is ensured.

[0083] Please refer to Figure 2 , a battery pack group equalization system based on clustering. Each battery cluster in the system includes multiple battery modules;

[0084] Each of the battery modules includes multiple battery cells, and is provided with a bidirectional active equalization converter and a battery cell management unit. The bidirectional active equalization converter is used to realize the bidirectional flow of equalization energy between the low-voltage auxiliary power supply and the battery module, thereby realizing active equalization;

[0085] The battery cell management unit includes a microprocessor and an analog front end, and the battery cell management unit is communicatively connected to the battery module management unit for receiving control instructions. The analog front end is used to realize the passive equalization between the single battery cells;

[0086] The battery cluster realizes the steps in the above-mentioned battery pack group equalization method based on clustering based on the above structure.

[0087] The battery pack group equalization method and system of the present invention are applicable to the battery equalization of the battery energy storage system.

[0088] Please refer to Figure 1, Embodiment 1 of the present invention is as follows:

[0089] A clustering-based battery pack equalization method, comprising the steps:

[0090] S1. Under the condition that the charging of the battery cluster is approaching the cut-off and the overcharge protection of the single battery is not reached, perform temperature-equalizing and voltage-limiting charging on all battery modules in the same battery cluster, and perform voltage sampling at the battery module level and the cell level respectively to obtain a module voltage sampling data set and a single cell voltage sampling data set;

[0091] Step S1 includes the steps:

[0092] S11. Collect the voltages and charging currents of N battery modules in the same battery cluster. When the voltage of a certain module reaches the first charging preset voltage, reduce the charging current I C to a preset ratio of the previous moment, and at the same time start the single cell temperature equalizing management.

[0093] In this embodiment, the preset ratio is 0.5 times, and in other equivalent embodiments, it can be adjusted according to the actual situation.

[0094] In this embodiment, collect the voltages and charging currents of N modules in the same battery cluster. When the voltage of a certain module reaches the first charging preset voltage, that is, V m_ ≥V δ1 when, reduce the charging current I C to 0.5 times of the previous moment, and at the same time start the single cell temperature equalizing management to make the temperatures of all M battery cells of all N modules in the same battery cluster tend to be consistent.

[0095] S12. Every time a module voltage reaches the first charging preset voltage, reduce the charging current I C to a preset ratio of the previous moment until the following conditions are met:

[0096]

[0097] Perform the first synchronous sampling of the voltages of the N battery modules;

[0098] wherein, V m_max is the highest voltage of each battery module in the battery cluster, ΔT is the temperature difference of M battery cells, ΔT limit is the temperature difference limit value, C1 is the preset ratio of temperature-equalizing and voltage-limiting charging, and I CN is the rated charging current.

[0099] In this embodiment, every time a module voltage reaches the first charging preset voltage subsequently, reduce IC to 0.5 times of the previous moment again until the above conditions are met, and then perform the first synchronous sampling of the voltages of N modules. Keep the temperature equalizing management on during the process.

[0100] Specifically, for the charged battery cluster, N battery module voltages are synchronously collected to form a data set V M_1 ={V m1 (1), …, V mN (1)};

[0101] When the first two conditions in the conditions are satisfied and ΔT > ΔT limit , then reduce I C to zero and keep the uniform temperature management on until the temperature condition ΔT < ΔT limit is satisfied, then resume charging the battery at rate C1, and collect the data set V m_i ≥ V δ1 at the moment of V M .

[0102] S13. Starting from the first synchronous sampling moment, maintain the charging current I C unchanged and perform continuous k synchronous samplings until the voltage of any module reaches the second charging preset voltage V δ2 :

[0103] V m_max ≥ Vδ2;

[0104] Obtain the module voltage sampling data set;

[0105] S14. Record the energy charged into the highest voltage module during k samplings:

[0106]

[0107] S15. During the kth sampling, collect the voltages of all battery cells in each of the battery modules to obtain the cell voltage sampling data set.

[0108] In this embodiment, starting from the first synchronous sampling moment, maintain I C unchanged and continue charging until the voltage of any module reaches the second charging preset voltage V δ2 , that is, V m_max ≥ V δ2 , and during this period, perform continuous k synchronous samplings and record the energy charged into the highest voltage module during k samplings.

[0109] S2. According to the module voltage sampling data set, perform clustering and grouping on the battery modules, and perform active balancing between the battery modules according to the grouping result; including the steps:

[0110] S21a. Calculate the equalization voltage judgment threshold according to the module voltage sampling data set;

[0111] Step S21a is specifically:

[0112] Respectively take the maximum values in the first and k-th sampling data sets V in the module voltage sampling data set M_1 and V M_k , and use the difference between the two as the balancing voltage judgment threshold ε M :

[0113] ε M = max(V M_k ) - max(V M_1 ).

[0114] In this embodiment, respectively take the maximum values in the first and k-th sampling data sets V M_1 and V M_k , and use the difference between the two as the module balancing voltage judgment threshold.

[0115] S22a. Sort the voltages of the k-th sampling in the module voltage sampling data set, calculate the range and the mean value, and judge whether active balancing between modules is required according to the range and the balancing voltage judgment threshold. If active balancing is required, go to step S23a.

[0116] Specifically: Sort the voltages of the k-th sampling in the module voltage sampling data set and calculate the range. If the range is greater than the balancing voltage judgment threshold ε M , it is considered that one active balancing between modules is required, and go to step S23a.

[0117] In this embodiment, sort the voltages in V M_k , calculate the range and the mean value, cluster and group the modules that need to be balanced according to the set threshold, and implement active balancing at the module level.

[0118] The reasons for the imbalance between battery modules include, for example, working for a long time under large temperature differences, abnormal power consumption of the internal circuit of the module, different initial SOC or SOH caused by module maintenance or replacement, etc. In these cases, the consistency of the single cells within the same module may be better than that between the single cells of different modules. Passive balancing for single cells will make the balancing time of all single cells very long. ε M is used to characterize the voltage imbalance degree threshold between modules.

[0119] In this embodiment, the module management unit (BMU) of the battery management system sorts the N voltage data in the data set V M_k to obtain {V m_k_min , …, V m_k_max}. When V m_k_max - V m_k_min > ε M, indicating that the module voltage difference exceeds the consistency threshold and module-level balancing is required, that is, starting a module balancing operation once.

[0120] S23a. Perform clustering and grouping according to the mean value, and perform active balancing between the battery modules according to the grouping result;

[0121] Step S23a is specifically as follows:

[0122] Calculate the voltage calculation mean value V of the k-th sampling in the module voltage sampling data set m_k_avg , and divide the data set into two subsets V M_k = V ML_ ∪ V MU_ ;

[0123] Among them, all voltage values less than or equal to the mean value V m_k_avg are classified into V ML_ , and vice versa into V MU_ ;

[0124] In this embodiment, according to the consistency requirement threshold of the module, the energy to be balanced, the deviation degree between the highest voltage module and the average value, and the balancing power, the required balancing duration is calculated:

[0125] The BMU controls the start of the bidirectional active balancing converter of the battery module corresponding to the maximum value V MU_ in the data set V mu__ , and the energy conversion is carried out from the battery module to the auxiliary power supply direction for a duration T Mu is:

[0126]

[0127] Among them, V m_k_min is the minimum value in the data set V ML_ , s1 is an adjustment coefficient, and P1 is the constant power of the bidirectional active balancing converter.

[0128] In this embodiment, the adjustment coefficient s1 is used to compensate for the energy loss caused by the converter efficiency.

[0129] Calculate the data set T formed by the on-time of the bidirectional active balancing converter of each battery module corresponding to each value in the data set V ML_ ; ML ;

[0130] Among them, for the on-time of the bidirectional active balancing converter of the battery module corresponding to the j-th value among the total h values in V ML_ :

[0131]

[0132] Obtain the data set T ML , sort the data in the data set T ML , and sequentially turn on the bidirectional active equalization converters of the corresponding battery modules in order, perform energy conversion from the auxiliary power supply to the module direction, and work for the corresponding duration until all the data in the data set T ML The bidirectional active equalization converters of all the battery modules corresponding to the data have completed their operations.

[0133] In this embodiment, the BMU calculates the converter turn-on time data set T of each module corresponding to each value in the data set V ML_ , and controls these converters to be sequentially turned on simultaneously with the start of the bidirectional active equalization converters of the corresponding battery modules in V ML mentioned above, and perform energy conversion from the auxiliary power supply to the module direction. Sort the data in T mu__ , sequentially turn on the converters in a corresponding module in order and work for the corresponding duration, until all k converters complete their operations and then all are turned off. ML The energy transmitted by the converter in the maximum voltage module is received by k modules lower than the voltage average value, and one converter is sequentially turned on in descending order of duration, so that the energy entering and leaving the auxiliary power supply line remains balanced.

[0134] When a single equalization cannot be completed within one equalization working condition period in this embodiment, no new voltage acquisition and equalization time calculation are performed in the next period until this round of equalization is completed.

[0135] S3. Repeat steps S1 and S2 until the equalization ends.

[0136] S3. Repeat steps S1 and S2 until the equalization ends.

[0137] In this embodiment, during the operation of the battery cluster, every time the conditions described in S1 are met, the above-mentioned sampling - clustering - control process is repeated until V m_k_max -V m_k_min <s2ε M when this module active equalization working condition ends.

[0138] s2 is the set active equalization judgment dead zone coefficient, generally taking values from 0.5 to 0.8.

[0139] According to the single - cell voltage sampling data set of each battery module, perform clustering and grouping of the single - cells in the battery module, and perform passive equalization between the single - cells according to the grouping result, including the steps:

[0140] S21b. For each battery module, calculate the range and mean according to the single - cell voltage sampling data set, and perform sorting.

[0141] In this embodiment, the cell management unit (CSC) of the battery management system sorts all the data in the cell voltage data set V within this module and calculates the mean value to obtain (V C , …, V c_min , …, V c_max ) and V c_avg . When V c_max - V c_min > ε C , the discharge passive equalization of all cells greater than the average value is started.

[0142] Among them, according to different cell types, the selection of ε C is generally from 0 to several hundred mV.

[0143] S22b. When the range is greater than the preset equalization threshold, obtain all battery cells with voltages greater than the mean value and perform discharge passive equalization.

[0144] The specific discharge passive equalization is as follows:

[0145] Apply different PWM controls to each of the battery cells. The duty cycle d of the PWM is proportional to the degree of deviation of the voltage of the battery cell from the average value. For the j-th battery cell to be equalized in the battery module, there is d j :

[0146]

[0147] Among them, V c_j represents the voltage of the j-th battery cell to be equalized in the battery module, V c_avg represents the mean value of the cell voltage sampling data set, and V c_max represents the maximum value in the cell voltage sampling data set.

[0148] In this embodiment, the passive equalization is implemented by the analog front end (AFE) chip on the CSC by applying different PWM controls to the equalization circuits corresponding to each cell. The duty cycle d of the PWM is proportional to the degree of deviation of the voltage of the cell from the average value.

[0149] S3. Repeat steps S1 and S2 until the equalization ends.

[0150] In this embodiment, the steps of S21b and S22b are repeated at each voltage acquisition moment. When V c_max - V c_min < s3ε C , the equalization ends.

[0151] s3 is the set passive equalization judgment hysteresis coefficient, generally taking 0.5 to 0.8.

[0152] Please refer to Figure 2 , Embodiment 2 of the present invention is as follows:

[0153] A battery pack group balancing system based on clustering, where each battery cluster in the system includes multiple battery modules;

[0154] Each of the battery modules includes multiple battery cells, and is provided with a bidirectional active balancing converter and a battery cell management unit. The bidirectional active balancing converter is used to realize the bidirectional flow of balanced energy between the low-voltage auxiliary power supply and the battery module, thereby realizing active balancing;

[0155] In this embodiment, the N battery modules in the battery cluster each contain an isolated bidirectional active balancing converter, which can realize the bidirectional flow of balanced energy between the low-voltage auxiliary power supply and the battery module, and realize the transfer of energy between one battery module and another through the auxiliary power supply line.

[0156] The battery cell management unit includes a microprocessor and an analog front end, and the battery cell management unit is communicatively connected to the battery module management unit for receiving control instructions. The analog front end is used to realize the passive balancing between single battery cells.

[0157] In this embodiment, the start-stop and power regulation of the bidirectional active balancing converter in each battery module are controlled by the battery cell management unit (CSC) in the battery module. There is communication between the CSC and the battery module management unit (BMU) for receiving control instructions. The CSC includes an MCU and an AFE (analog front end), where the MCU implements start-stop and power control of the balancing converter.

[0158] Explanation: The converter of a certain battery module steps down the module voltage to the low-voltage auxiliary power supply of 12V or 24V, and at the same time, the converter of another module steps up 12V or 24V to the module voltage, realizing the transfer of energy between the two modules.

[0159] The battery cluster implements the steps in the above-mentioned battery pack group balancing method based on clustering according to the above structure.

[0160] In summary, the battery pack group balancing method and system of the present invention eliminate the influence of temperature, cluster battery modules and battery cells according to voltage, and perform active balancing between battery modules and passive balancing between battery cells within the battery module according to the clustering results, realizing more effective battery balancing.

[0161] The present invention adopts active balancing between modules and passive balancing within modules; in terms of control, it is based on the OV method, and at the same time eliminates the influence of temperature, forming a hierarchical balancing strategy based on the clustering idea and an active balancing circuit.

[0162] For an energy storage system with a large number of battery cells, general passive equalization methods discharge and equalize high-voltage battery cells within the entire battery cluster. When high- and low-voltage battery cells are relatively concentrated in a certain battery module, the equalization efficiency is low, resulting in energy waste. Module-level active equalization is achieved without adding too many circuits and converters, which is beneficial to the automatic maintenance and operation of energy storage batteries and the correction of consistency deviations caused by working conditions or environmental differences. The energy balance between the outgoing battery module and the incoming battery module ensures that the auxiliary power line does not cause excessive fluctuations.

[0163] While performing module-level equalization, the single-cell level passive equalization adopts a clustering group method, which is also beneficial to improving the equalization efficiency.

[0164] The equalization sampling point is selected in the small-current stage before the end of charging, which reduces the influence of current in the operating voltage sampling and also avoids static waiting. At the same time, the temperature consistency and the temperature equalization acquisition strategy are also considered, realizing the dimensionality reduction of SOC data, reducing the calculation load and increasing the accuracy.

[0165] The above are only embodiments of the present invention, and do not limit the patent scope of the present invention accordingly. Any equivalent transformation made by using the content of the specification and drawings of the present invention, or directly or indirectly applied in the relevant technical fields, shall be equally included in the patent protection scope of the present invention.

Claims

1. An active balancing method for battery modules based on clustering, characterized in that, Including the steps: S1. Under the condition that the battery cluster charging is approaching the cut-off and the overcharge protection of the single battery is not reached, perform temperature-equalizing and voltage-limiting charging on all battery modules within the same battery cluster, and perform voltage sampling at the battery module level to obtain a module voltage sampling data set; S2. According to the module voltage sampling data set, perform clustering and grouping on the battery modules, and perform active balancing between the battery modules according to the grouping result, including the steps: S21a. Calculate an equalizing voltage judgment threshold according to the module voltage sampling data set; The equalizing voltage judgment threshold is used to represent the voltage imbalance degree threshold between modules; S22a. Sort the voltages of the k-th sampling in the module voltage sampling data set, calculate the range and the mean value, and judge whether active balancing between modules is required according to the range and the equalizing voltage judgment threshold. If active balancing is required, go to step S23a; S23a. Perform clustering and grouping according to the mean value, and perform active balancing between the battery modules according to the grouping result.

2. The active balancing method for battery modules based on clustering according to claim 1, wherein Step S21a is specifically: Respectively take the maximum values in the first sampling data set V and the k-th sampling data set V M_1 and V M_k in the data set, and use the difference between the two as the equalization voltage judgment threshold ε M : ε M = max(V M_k ) - max(V M_1 )。 3. The active balancing method for battery modules based on clustering according to claim 1, characterized in that Step S22a is specifically: Sort the voltages of the k-th sampling in the module voltage sampling data set and calculate the range. If the range is greater than the equalization voltage judgment threshold ε M , it is considered that an active equalization between modules is required, and step S23a is entered.

4. A clustering-based active equalization method for battery modules according to claim 1, characterized in that Step S23a is specifically: Calculate the mean voltage V of the k-th sampled voltage in the module voltage sampling dataset m_k_avg , and divide the dataset into two subsets V M_k = V ML_k ∪ V MU_k ; Among them, all voltage values less than or equal to the mean value V m_k_avg are classified into V ML_k , and vice versa into V MU_k ; Control data set V MU_k The maximum value V in mu_k_max The bidirectional active equalization converter of the battery module corresponding thereto is started, and energy conversion is performed from the battery module in the direction of the auxiliary power supply for a duration T MU is as follows: Among them, V m_k_min is the minimum value in the data set V ML_k , and V m_k_max is the maximum value in the data set V ML_k . s1 is the adjustment coefficient, and P1 is the constant power of the bidirectional active equalization converter; Calculate the dataset V ML_k The turn-on time of the bidirectional active equalization converter of the battery module corresponding to each value in the dataset T is formed ML ; Among them, for V ML_k the j-th value V of the total h values ml_k_j corresponding to the turn-on time of the bidirectional active equalization converter of the battery module: Obtain the data set T ML , sort the data in the data set T ML , and sequentially activate the bidirectional active equalization converters of the corresponding battery modules in order, perform energy conversion from the auxiliary power supply to the module direction, and work for the corresponding duration until the bidirectional active equalization converters of all the battery modules corresponding to the data in the data set T ML have all completed their operations.

5. A clustering-based active equalization method for battery modules according to claim 1, characterized in that Step S1 includes the steps: S11. Collect the voltages and charging currents of N battery modules in the same battery cluster. When the voltage of a certain module reaches the first charging preset voltage, reduce the charging current I C to a preset ratio of the previous moment, and at the same time start the single-cell equal-temperature thermal management; S12. Each time the module voltage reaches the first charging preset voltage V δ1 , reduce the charging current I C to a preset ratio of the previous moment until the following conditions are met: Perform the first synchronous sampling on the voltages of N battery modules; Among them, V m_max is the highest voltage of each battery module in the battery cluster, ΔT is the temperature difference of M single cells of the battery, and ΔT limit is the temperature difference limit value, C1 is the preset magnification of equal-temperature and voltage-limited charging, and I CN is the rated charging current; S13. From the first synchronous sampling moment, maintain the charging current I C constant and perform continuous synchronous sampling for k times until the voltage of any module reaches the second charging preset voltage V δ2 : V m_max ≥V δ2 ; Obtain the module voltage sampling data set; S14. Record the energy charged into the highest voltage module in the k-th sampling:

6. A clustering-based active battery module equalization system, characterized in that, Each battery cluster in the system includes multiple battery modules; Each battery module includes multiple battery cells, and is provided with a bidirectional active equalizing converter and a battery cell management unit. The bidirectional active equalizing converter is used to realize the bidirectional flow of equalizing energy between the low-voltage auxiliary power supply and the battery module, and further realize active balancing, including the steps: S1. Under the condition that the battery cluster charging is approaching the cut-off and the overcharge protection of the single battery is not reached, perform temperature-equalizing and voltage-limiting charging on all battery modules within the same battery cluster, and perform voltage sampling at the battery module level to obtain a module voltage sampling data set; S2. According to the module voltage sampling data set, perform clustering and grouping on the battery modules, and perform active balancing between the battery modules according to the grouping result, including the steps: S21a. Calculate an equalizing voltage judgment threshold according to the module voltage sampling data set; The equalizing voltage judgment threshold is used to represent the voltage imbalance degree threshold between modules; S22a. Sort the voltages of the k-th sampling in the module voltage sampling data set, calculate the range and the mean value, and judge whether active balancing between modules is required according to the range and the equalizing voltage judgment threshold. If active balancing is required, go to step S23a; S23a. Perform clustering and grouping according to the mean value, and perform active balancing between the battery modules according to the grouping result.

7. The active balancing system for battery modules based on clustering according to claim 6, wherein Step S21a is specifically: Respectively take the maximum values in the first sampling data set V and the k-th sampling data set V M_1 and V M_k in the module voltage sampling data set, and use the difference between the two as the balancing voltage judgment threshold ε M : ε M = max(V M_k ) - max(V M_1 )。 8. The active equalization system for battery modules based on clustering according to claim 6, wherein Step S22a is specifically: Sort the voltages of the k-th sampling in the module voltage sampling data set and calculate the range. If the range is greater than the equalizing voltage judgment threshold ε M , it is considered that an active equalization between modules is required, and step S23a is entered.

9. The active equalization system for battery modules based on clustering according to claim 6, characterized in that, Step S23a is specifically: Calculate the mean voltage V of the k-th sampled voltage in the module voltage sampling dataset m_k_avg , and divide the dataset into two subsets V M_k = V ML_k ∪ V MU_k ; Among them, all voltage values less than or equal to the mean value V m_k_avg are classified into V ML_k , and vice versa into V MU_k ; Control data set V MU_k The maximum value V in mu_k_max The bidirectional active equalization converter of the battery module corresponding thereto is started, and energy conversion is performed from the battery module to the auxiliary power supply direction for a duration T MU Is: Among them, V m_k_min is the minimum value in the data set V ML_k , and V m_k_max is the maximum value in the data set V ML_k . s1 is the adjustment coefficient, and P1 is the constant power of the bidirectional active equalization converter; Calculate the dataset V ML_k The turn-on time of the bidirectional active equalization converter of the battery module corresponding to each value in it forms the dataset T ML ; Among them, for V ML_k the j-th value V of a total of h values in the Communist Party ml_k_j The turn-on time of the bidirectional active equalization converter of the battery module corresponding thereto is: Obtain the data set T ML , sort the data in the data set T ML , and sequentially turn on the bidirectional active equalization converters of the corresponding battery modules in order, perform energy conversion from the auxiliary power supply to the module direction, and work for the corresponding duration until the bidirectional active equalization converters of all the battery modules corresponding to the data in the data set T ML have all completed their operations.

10. A clustering-based active equalization system for battery modules according to claim 6, characterized in that Step S1 includes the steps: S11. Collect the voltages and charging currents of N battery modules in the same battery cluster. When the voltage of a certain module reaches the first charging preset voltage, reduce the charging current I C to a preset ratio of the previous moment, and at the same time start the single-cell equal-temperature thermal management; S12. Every time the module voltage reaches the first charging preset voltage V δ1 reduce the charging current I C to a preset ratio of the previous moment until the following conditions are met: Perform the first synchronous sampling on the voltages of N battery modules; Among them, V m_max is the highest voltage of each battery module in the battery cluster, ΔT is the temperature difference of M single cells, and ΔT limit is the temperature difference limit value, C1 is the preset rate for equalizing temperature and limiting voltage during charging, and I CN is the rated charging current; S13. From the first synchronous sampling moment, maintain the charging current I C constant, and perform continuous synchronous sampling for k times until the voltage of any module reaches the second preset charging voltage V δ2 : V m_max ≥V δ2 ; Obtain the module voltage sampling data set; S14. Record the energy charged into the highest voltage module in the k-th sampling:

Images