Layered collaborative battery energy storage system bypass equalization method
By employing a hierarchical collaborative bypass balancing method for battery energy storage systems, and utilizing an LSTM model to predict SOC values and differentiate current regulation, the problems of battery imbalance and voltage disturbance in battery energy storage systems are solved, thereby improving battery consistency and operational efficiency.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SICHUAN UNIV
- Filing Date
- 2026-03-13
- Publication Date
- 2026-05-15
AI Technical Summary
Existing battery energy storage systems face problems such as increased imbalance between batteries, delayed balancing control, voltage disturbances, and increased costs when dealing with heterogeneity in battery states and complex dynamic operating conditions. Existing solutions rely on additional hardware, resulting in low efficiency.
A hierarchical collaborative bypass balancing method for battery energy storage systems is adopted. By calculating the balancing index of the battery modules, the target battery cells are determined and bypass balancing is performed. The current of the battery modules is compensated by a differentiated current regulation coefficient. The SOC value is predicted by an LSTM model, and the current is adjusted in real time to achieve SOC balancing among battery modules.
Without adding extra hardware, it eliminates the effects of voltage drops, achieves SOC balance between battery modules, avoids control lag, improves battery consistency and system operating efficiency, extends system maintenance intervals, and reduces switching losses.
Smart Images
Figure CN122052237A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of electronic technology, and specifically to a bypass balancing method for a hierarchical and collaborative battery energy storage system. Background Technology
[0002] Battery energy storage systems (BESS) are increasingly widely used in current power grids, but they face challenges such as heterogeneity in battery states and complex dynamic operating conditions during operation. Balancing the grid-connected dynamic output stability of BESS with battery equalization is crucial. At the level of inter-battery group coordinated control, existing power allocation strategies mostly rely solely on the average state of charge (SOC) of the battery group, ignoring the dispersion of states among batteries within the group, which may exacerbate imbalances. Regarding intra-battery group equalization control, bypass equalization strategies are currently widely used. However, conventional bypass equalization strategies generally employ threshold triggering mechanisms, resulting in lag in imbalance monitoring and equalization control, reducing equalization efficiency. Furthermore, bypass actions may cause voltage disturbances in the BESS, and existing solutions often rely on adding additional DC / DC regulator hardware, leading to a significant increase in cost and losses. Summary of the Invention
[0003] The purpose of this invention is to provide a hierarchical collaborative bypass balancing method for battery energy storage systems to solve the technical problems existing in related technologies.
[0004] To achieve the above objectives, this invention provides a hierarchical collaborative bypass balancing method for a battery energy storage system. The battery energy storage system comprises multiple parallel battery modules, each battery module including any number of battery cells connected in series. The balancing method includes:
[0005] For any given battery module:
[0006] Calculate the balance index of the battery module, wherein the balance index is obtained by subtracting the minimum SOC prediction value from the maximum SOC prediction value among the individual battery cells in the battery module.
[0007] When the balance index exceeds the preset balance threshold, the target battery cell of the battery module is determined. The target battery cell is the battery cell with the lowest SOC prediction value in the battery module when the battery energy storage system is in the discharge state, or the battery cell with the highest SOC prediction value in the battery module when the battery energy storage system is in the charging state.
[0008] Bypass balancing is performed on the target battery cells, and the current of the battery module is compensated by a differentiated current regulation coefficient.
[0009] Optionally, the battery energy storage system further includes a first switch and a second switch. For any given battery cell, the second switch is connected in series, and the first switch is connected in parallel to the battery cell and the second switch. The bypass balancing of the target battery cell includes:
[0010] Close the first switch connected in parallel with the target battery cell and disconnect the second switch connected in series with the target battery cell.
[0011] Optionally, the equilibrium index is obtained by the following method:
[0012] The real-time SOC value of each battery cell in the battery module is obtained, and the real-time SOC value is input into the target LSTM model to obtain the predicted SOC value of the battery cell. The target LSTM model is trained by the historical SOC value of the battery cell.
[0013] Among multiple SOC predictions, the equilibrium index is obtained by subtracting the minimum SOC prediction from the maximum SOC prediction.
[0014] Optionally, the current of the compensated battery module is calculated using the following formula:
[0015] ;
[0016] ;
[0017] ;
[0018] in, To compensate for the first in the battery module The current of each battery cell To compensate for the first The current of each battery module This refers to the order of the target battery cells within the battery module. For the first battery energy storage system The port voltage of each battery module after bypass balancing For the first battery energy storage system Control current of each battery module For the first battery energy storage system The output power of each battery module For the first battery energy storage system Total output power of each battery module For the first battery energy storage system The total number of individual battery cells in a battery module.
[0019] Optionally, in the battery energy storage system, the first Control current of each battery module This can be expressed by the following formula:
[0020] ;
[0021] ;
[0022] in, This represents the total current demand on the grid side of the battery energy storage system. For the first battery energy storage system Differentiated current regulation coefficients for each battery module Let be the differentiated current regulation coefficient for the k-th battery module. This refers to the number of battery modules connected in parallel in a battery energy storage system. For the first battery energy storage system Boundary penalty factor for each battery module For the first battery energy storage system Internal penalty factor for each battery module For the first battery energy storage system The demand potential for individual battery modules.
[0023] Optionally, the boundary penalty factor This can be expressed by the following calculation formula:
[0024] ;
[0025] in, For the first battery energy storage system SOC of a battery module The lower limit SOC of the battery module, To ensure safe bandwidth buffering, This represents the upper limit of the SOC (State of Charge) of the battery module.
[0026] The internal penalty factor This can be expressed by the following calculation formula:
[0027] ;
[0028] in, For consistency sensitivity coefficient, As a balance indicator;
[0029] The demand potential energy This can be expressed by the following calculation formula:
[0030] ;
[0031] ;
[0032] in, Adjust the sensitivity index to adjust the SOC. For the first battery energy storage system SOC of a battery module This is the control current for the battery module. The first in the battery module The remaining capacity of each battery cell The first in the battery module SOC of a single battery cell This represents the total number of individual battery cells in the battery module.
[0033] Through the above technical solution, for any battery module in the battery energy storage system, the balance index of the battery module is calculated. When the balance index exceeds the preset balance threshold, the target battery cell of the battery module is determined, and bypass balance is performed on the target battery cell. Then, the current of the battery module is compensated by a differentiated current control coefficient. Without adding additional hardware structure, the current of the battery module is compensated according to the differentiated current control coefficient, thereby eliminating the voltage drop caused by the bypass balance topology switching and realizing SOC balance between battery modules. At the same time, the SOC of the battery cell can be calculated based on the compensated current, realizing SOC balance of the battery cells within the battery module. This avoids the lag in SOC balance control between battery cells or between battery modules.
[0034] Other features and advantages of the present invention will be described in detail in the following detailed description section. Attached Figure Description
[0035] Figure 1 This is a schematic diagram illustrating a hierarchical collaborative bypass balancing method for a battery energy storage system according to an exemplary embodiment of the present invention.
[0036] Figure 2 This is a schematic diagram illustrating the parallel connection of battery modules according to an exemplary embodiment of the present invention.
[0037] Figure 3 This is a schematic diagram of bypass balancing according to an exemplary embodiment of the present invention.
[0038] Figure 4 This is a flowchart illustrating a hierarchical collaborative bypass balancing method for a battery energy storage system according to an exemplary embodiment of the present invention.
[0039] Figure 5 This is illustrated in an exemplary embodiment of the present invention. and Schematic diagram of the curve.
[0040] Figure 6 This is an exemplary embodiment of the present invention illustrating the interaction between battery cells. Schematic diagram of the curve. Detailed Implementation
[0041] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, so as to provide a better understanding of the concept of the present invention, the technical problem solved, the technical features constituting the technical solution, and the technical effects brought about.
[0042] In recent years, with the acceleration of the global energy transition, the large-scale grid connection of intermittent renewable energy sources, such as wind power and photovoltaics, has led to dramatic dynamic fluctuations in the load demand of power systems. Battery-Effective Systems (BESS), with its rapid response and bidirectional regulation capabilities, has become a key resource for smoothing power fluctuations and tracking dynamic output commands. However, most existing BESS control strategies treat them as an ideal, uniform whole, ignoring the differences between cells. In practical applications, manufacturing tolerances such as capacity and internal resistance are common among cells, and this heterogeneity significantly intensifies over time. When the BESS operates under complex dynamic output conditions with high-frequency fluctuations, this heterogeneity can cause the weaker cells to prematurely reach their charge / discharge cutoff voltage, limiting the overall usable capacity and continuous operating time of the BESS, and in severe cases, even triggering thermal runaway or causing unplanned BESS shutdowns. Therefore, establishing a hierarchical collaborative equilibrium control system capable of adapting to both high heterogeneity among cells and dynamic output conditions is particularly urgent.
[0043] Currently, the following technical challenges exist for BESS equalization control under dynamic output conditions:
[0044] 1. Coordinated control between battery packs: Existing power distribution strategies between battery packs in BESS usually only consider the SOC of each battery pack and give little consideration to the dispersion between batteries. This can easily lead to battery packs with poor internal consistency bearing excessive loads, accelerating the aging and failure of weaker batteries.
[0045] 2. Balancing Strategy under Complex Dynamic Conditions: Unlike balancing during static or constant current charging, the current under dynamic output conditions is random and fluctuates rapidly. Traditional bypass balancing strategies typically only activate after the imbalance index exceeds the limit, resulting in monitoring blind spots and control lag. This has limited effectiveness in suppressing BESS aging and may increase device losses due to repeated switching actions.
[0046] 3. Topology reconfiguration and stable support for the external power grid: Bypass balancing is essentially a topology change operation. The balancing action may cause voltage disturbances in the BESS. Existing solutions mostly involve adding additional DC / DC voltage regulation hardware, which significantly increases costs and losses. How to improve the control algorithm and use the current gain adjustment to offset the impact of voltage drops in real time and maintain constant BESS output power is a control problem that must be overcome in practical engineering applications.
[0047] In the first related technology, the battery energy storage system is balanced and controlled by a converter-based equalization control strategy; in the second related technology, the battery energy storage system is controlled by a hierarchical equalization and ripple suppression strategy based on a modular DC reconfigurable battery pack; in the third related technology, the system is controlled by a two-layer active equalization topology based on a model prediction algorithm; and in the fourth related technology, the battery energy storage system is controlled by a bypass equalization strategy based on optimized equalization indices.
[0048] However, the inventors discovered that when controlling the battery energy storage system using the first and second correlation technologies, although balanced control of each battery is achieved by adjusting the converter duty cycle, it essentially relies on a cascaded power electronic converter architecture, requiring a full-power converter for each battery or battery pack. This dependence on additional power hardware not only significantly increases the size and manufacturing cost of the BESS, but also introduces additional power losses due to the multi-stage power electronic conversion. When controlling using the second correlation technology, this scheme limits the execution frequency of the balancing sorting algorithm to a low level in order to reduce the communication and computational burden on the controller. This makes it difficult to guarantee the constantness and high-precision response of the BESS's output power when facing dynamic output conditions with high-frequency changes. In addition, the hierarchical architecture of this strategy determines that the system-level controller can only obtain the aggregation state at the module level, making it difficult to perceive the differences between batteries within the group. Therefore, when allocating control between battery packs, only the average SOC of the battery pack is considered. Long-term operation will accelerate the aging of the internal weakest battery and fail to achieve full life-cycle health management. When controlled by the third correlation technique, although a complex algorithm is used to select the optimal balancing path, it relies on additional bidirectional flybacks and Cuk circuits for energy transfer. Unlike this invention, which only uses existing PCS for current regulation, this scheme introduces a large number of magnetic components and power switches, increasing losses due to frequent power conversions under high-frequency dynamic conditions. Furthermore, the core objective function of this strategy is optimal internal consistency, lacking a real-time compensation mechanism for grid-connected power fluctuations and ignoring the interference of voltage ripple generated by balancing actions on the external output characteristics of the BESS. When controlled by the fourth correlation technique, the balancing index is optimized based on the traditional fixed balancing index, allowing for the selection of a more efficient balancing index according to the real-time operating status of the BESS. However, this scheme is essentially still a threshold-based triggering mechanism; the bypass action is only initiated after the balancing index actually exceeds the set threshold. This lag makes it impossible to effectively clamp the differences between batteries within a safe threshold and easily leads to repeated ineffective switching actions, increasing device losses.
[0049] In view of this, the present invention provides a hierarchical collaborative bypass balancing method for battery energy storage systems to solve the technical problems existing in the above-mentioned related technologies.
[0050] like Figure 1 As shown, Figure 1 This is a schematic diagram illustrating a hierarchical collaborative bypass balancing method for a battery energy storage system according to an exemplary embodiment of the present invention, with reference to... Figure 1 The battery energy storage system comprises multiple parallel battery modules, each battery module including any number of battery cells connected in series, and the balancing method includes;
[0051] For any given battery module:
[0052] S101: Calculate the balance index of the battery module, wherein the balance index is obtained by subtracting the minimum SOC prediction value from the maximum SOC prediction value among the individual battery cells in the battery module.
[0053] S102: When the balance index exceeds the preset balance threshold, the target battery cell of the battery module is determined, wherein the target battery cell is the battery cell with the lowest SOC prediction value in the battery module when the battery energy storage system is in the discharge state, or the battery cell with the highest SOC prediction value in the battery module when the battery energy storage system is in the charging state.
[0054] S103: Perform bypass balancing on the target battery cell and compensate the current of the battery module through a differentiated current regulation coefficient.
[0055] Through the above technical solution, for any battery module in the battery energy storage system, the balance index of the battery module is calculated. When the balance index exceeds the preset balance threshold, the target battery cell of the battery module is determined, and bypass balance is performed on the target battery cell. Then, the current of the battery module is compensated by a differentiated current control coefficient. Without adding additional hardware structure, the current of the battery module is compensated according to the differentiated current control coefficient, thereby eliminating the voltage drop caused by the bypass balance topology switching and realizing SOC balance between battery modules. At the same time, the SOC of the battery cell can be calculated based on the compensated current, realizing SOC balance of the battery cells within the battery module. This avoids the lag in SOC balance control between battery cells or between battery modules.
[0056] To enable those skilled in the art to better understand the spatiotemporal data prediction method provided by this invention, the above steps are illustrated in detail below.
[0057] For example, a battery energy storage system can be a power system that achieves bidirectional conversion of electrical energy and chemical energy through electrochemical cells, and allows for on-demand charging and discharging. Its structure is as follows: Figure 2As shown, the system includes multiple battery modules connected in parallel, and each battery module includes any number of battery cells connected in series. For each battery cell, a second switch is connected in series, and a first switch is connected in parallel to the battery cell and the second switch. The first and second switches can be used to perform bypass balancing of the battery cells. Bypass balancing can be a method of balancing cells that are charged too fast or have too high voltage by using resistors to discharge excess power, so that the voltage of the entire battery string remains consistent.
[0058] For example, the equilibrium index can be the maximum difference between the predicted SOC values. In this embodiment of the invention, for each battery module in the battery energy storage system, the equilibrium index of the battery module is calculated. The equilibrium index is obtained by selecting the maximum and minimum predicted SOC values among the predicted SOC values of multiple battery cells, and subtracting the minimum predicted SOC value from the maximum predicted SOC value.
[0059] In one possible manner, the equilibrium index is obtained by the following method:
[0060] The real-time SOC value of each battery cell in the battery module is obtained, and the real-time SOC value is input into the target LSTM model to obtain the predicted SOC value of the battery cell. The target LSTM model is trained by the historical SOC value of the battery cell.
[0061] Among multiple SOC predictions, the equilibrium index is obtained by subtracting the minimum SOC prediction from the maximum SOC prediction.
[0062] It should be understood that when calculating the predicted SOC value of a single battery cell, prediction can be made using a target LSTM (Long Short-Term Memory) model. This embodiment of the invention does not specifically limit this approach. Specifically, as... Figure 3 As shown, when obtaining the target LSTM model, the historical SOC values of individual battery cells are first acquired. These historical SOC values are then used to train the LSTM model offline, thus obtaining the target LSTM model. Next, the real-time SOC values of each battery cell in the battery module are acquired and input into the target LSTM model to obtain the predicted SOC value for that battery cell. Specifically, in... Figure 3 In the middle, Tanh and These represent activation functions: Input gate, Output gate, and Forget gate. The input at time t, This is the output at time t-1. For time The cell state. After calculating the SOC prediction value of all battery cells in the battery module, multiple SOC prediction values can be obtained. Among the multiple SOC prediction values, the maximum SOC prediction value and the minimum SOC prediction value can be determined. Subtracting the minimum SOC prediction value from the maximum SOC prediction value yields the equilibrium index.
[0063] For example, this balancing index can then be compared with a preset balancing index, and the individual battery cells in the battery module can be controlled based on the comparison result. Specifically, when the balancing index does not exceed the preset balancing threshold, it indicates that the battery module is operating stably and no additional control is required. When the balancing index exceeds the preset balancing threshold, it indicates that the battery module is operating abnormally, and bypass balancing is required. Before performing bypass balancing, the target battery cell in the battery module can be determined. This target battery cell needs to be determined based on the state of the battery energy storage system. When the battery energy storage system is charging, the target battery cell is the one with the highest predicted SOC value in the battery module; when the battery energy storage system is discharging, the target battery cell is the one with the lowest predicted SOC value in the battery module.
[0064] Specifically, when the battery energy storage system is in a charging state, the method for determining the target battery cell is as follows: Determine the predicted SOC value of each battery cell in the battery module, and calculate the average predicted SOC value of the battery cells in the battery module based on multiple predicted SOC values; calculate the compensated SOC value of the battery module based on the average predicted SOC value; sort the multiple predicted SOC values in descending order, and sum the predicted SOC values sequentially from high to low; when the sum of the summed predicted SOC values is greater than the compensated SOC value, the battery cell corresponding to the summed predicted SOC value is determined as the target battery cell. This can be expressed by the following formula:
[0065] ;
[0066] in, The first in the battery module Predicted SOC value for each battery cell To compensate for the SOC value,
[0067] For example, after the target battery cell is determined, bypass balancing can be performed on the target battery cell.
[0068] In some possible configurations, the battery energy storage system further includes a first switch and a second switch. For any given battery cell, the second switch is connected in series, and the first switch is connected in parallel to both the battery cell and the second switch. The bypass balancing of the target battery cell includes:
[0069] Close the first switch connected in parallel with the target battery cell and disconnect the second switch connected in series with the target battery cell.
[0070] It should be understood that in each battery cell of the battery energy storage system, a second switch is connected in series and a first switch is connected in parallel with the second switch. Therefore, during bypass balancing, the first switch connected in parallel with the target battery cell can be turned off, and the second switch connected in series with the target battery cell can be disconnected, thus preventing current from flowing through the target battery cell. Simultaneously, after bypass balancing of the target battery cell, the current and output power of its battery module will change, leading to differences in the State of Charge (SOC) of the battery cells within the battery module. Therefore, in this embodiment of the invention, while bypass balancing of the target battery cell, the current of the battery module is compensated, and the SOC of the battery cell is calculated based on the compensated current of the battery module and the remaining capacity of the battery cell, to balance the SOC of the battery cell and the SOC between battery modules, while also keeping the output power of the battery module constant.
[0071] In the most probable way, the current of the compensated battery module is calculated using the following formula:
[0072] ;
[0073] ;
[0074] ;
[0075] in, To compensate for the first in the battery module The current of each battery cell To compensate for the first The current of each battery module is greater than the original set value, which can achieve active current gain. This refers to the order of the target battery cells within the battery module. For the first battery energy storage system The port voltage of a battery module after bypass balancing, i.e., the voltage drop after bypass balancing of the target battery cell. For the first battery energy storage system Control current of each battery module For the first battery energy storage system The output power of each battery module For the first battery energy storage system Total output power of each battery module For the first battery energy storage system The total number of battery cells in a battery module. Voltage drop calculation: When the prediction algorithm triggers a bypass action in a target battery cell within the battery module, the series branch topology of that battery module changes instantaneously. Assume the total voltage of the battery module before the bypass action is... for:
[0076] ;
[0077] in, For the first battery energy storage system The first battery module Real-time voltage of each individual battery cell.
[0078] After the bypass balancing action is performed, the target battery cell in the battery module is bypassed, and the battery module... It can be calculated using the following formula.
[0079] ;
[0080] in, The number of target battery cells.
[0081] In one possible manner, the first in the battery energy storage system Control current of each battery module This can be expressed by the following formula:
[0082] ;
[0083] ;
[0084] in, This represents the total current demand on the grid side of the battery energy storage system. For the first battery energy storage system Differentiated current regulation coefficients for each battery module Let be the differentiated current control coefficient for the k-th battery module. The sum of the differentiated current control coefficients for all modules forms the denominator for differentiated control. This refers to the number of battery modules connected in parallel in a battery energy storage system. For the first battery energy storage system Boundary penalty factor for each battery module For the first battery energy storage system Internal penalty factor for each battery module For the first battery energy storage system The demand potential for individual battery modules.
[0085] It should be understood that an internal penalty factor can be introduced into the battery module. This is to mitigate the charging and discharging intensity of battery packs with poor internal consistency, thereby avoiding the exacerbation of the bottleneck effect. Furthermore, to prevent overcharging or over-discharging of the battery pack, the state of charge (SOC) of the battery pack is considered simultaneously. This is because the SOC of each individual battery cell within the battery module... The difference allows for the calculation of the required potential energy of the battery module. Secondly, to address the issue of significant SOC fluctuations under complex dynamic output conditions, a boundary penalty factor is set. When the SOC approaches the safe upper and lower limits, a forced decay is performed to build a safe buffer.
[0086] Specifically, the boundary penalty factor This can be expressed by the following calculation formula:
[0087] ;
[0088] in, For the first battery energy storage system SOC of a battery module The lower limit SOC of the battery module, To ensure safe bandwidth buffering, This represents the upper limit of the SOC (State of Charge) of the battery module.
[0089] The internal penalty factor This can be expressed by the following calculation formula:
[0090] ;
[0091] in, For consistency sensitivity coefficient, As a balance indicator;
[0092] The demand potential energy This can be expressed by the following calculation formula:
[0093] ;
[0094] ;
[0095] in, Adjust the sensitivity index to adjust the SOC. For the first battery energy storage system SOC of a battery module This is the control current for the battery module. The first in the battery module The remaining capacity of each battery cell The first in the battery module SOC of a single battery cell This represents the total number of individual battery cells in the battery module.
[0096] Through the above calculations, it is possible to achieve the following during the discharge process: When the battery module is positive, the battery module with a higher SOC and higher internal balance is allocated a larger amount of power. ;vice versa.
[0097] In one possible manner, the SOC of a battery cell is calculated based on the current of the compensated battery module and the remaining capacity of each battery cell using the following formula:
[0098] ;
[0099] in, The first in the battery module Each battery cell is in SOC at any moment The first in the battery module Each battery cell is in SOC at any moment It is the integral variable.
[0100] It should be understood that heterogeneity exists between battery modules and individual battery cells during the operation of a battery energy storage system. Therefore, a mathematical model can be established to quantify this state heterogeneity. The State of Charge (SOC) of a battery cell represents the ratio of its current stored energy to its maximum stored energy. For batteries with initial aging capacity differences, the SOC of a compensated battery cell can be calculated in real-time using the coulomb counting method during charging and discharging, based on the formula described above.
[0101] Different initial aging levels of batteries lead to Differences exist, even if all batteries are initially in the same state, the batteries will exhibit variations during synchronous charging and discharging. The heterogeneity still exhibits a divergent trend, and this heterogeneity accumulates over time. In this embodiment of the invention, the difference between the maximum predicted SOC values within the battery module can be selected as the equilibrium index. This equilibrium index directly reflects the degree of the bottleneck effect in a battery pack composed of N batteries connected in series, and its expression is: ,in, As a balance indicator, The first in the battery module SOC of each battery cell.
[0102] In actual processing, such as Figure 4 As shown, the processing flow is as follows:
[0103] Obtain the historical SOC data sequence of each battery cell in the battery module and train the LSTM model offline.
[0104] In each control cycle, the real-time collected SOC data is added to the end of the input sequence, and the trained LSTM model is used to predict the SOC value of each battery in the next time step.
[0105] Based on the predicted data, calculate the balance index of the battery modules within the group at the predicted time. and compare it with the preset equalization threshold. Compare;
[0106] like Exceeding the preset equalization threshold If the target cell has the lowest predicted SOC value (during discharge) or the highest predicted SOC value (during charging), then the corresponding parallel switch S_2i-1 is closed and the series switch S_2i is opened, physically isolating the cell. Otherwise, the current topology remains unchanged.
[0107] The SOC of the individual battery cells is then updated and the data acquisition process is returned to begin the next prediction cycle.
[0108] The above technical solution first involves system initialization and data acquisition, obtaining real-time grid commands and the State of Charge (SOC) of individual battery cells within the battery module. Subsequently, it acquires the heterogeneous states of the batteries and implements differentiated current control strategies based on the internal characteristics of the battery pack. Building upon this, the real-time SOC value is used to predict the battery pack's equilibrium index at the next time step using a target LSTM model. and compare it with the preset equalization threshold. Comparisons are made to serve as the basis for decision-making regarding hierarchical control strategies: Not exceeding the preset equalization threshold Without changing the topology of individual battery cells within the battery module, the battery adjusts its current accordingly. Charge and discharge; if Exceeding the preset equalization threshold The battery energy storage system immediately activates the bypass balancing action within the battery pack, causing a change in the battery topology within the pack. Simultaneously, it executes a current regulation and compensation algorithm to ensure constant output, reconstructing the current of the compensated battery module in real time. Differentiated regulation is then implemented. Ultimately, the battery charges and discharges based on the current generated by the above decisions, updates the battery state closed loop, returns to the data acquisition stage, and enters the next control cycle, thereby achieving hierarchical collaborative balancing within the BESS.
[0109] In the specific implementation process, to verify the effectiveness of the present invention, this embodiment built a simulation model based on the parameters of a commercially available 5Ah soft-pack lithium-ion battery manufactured by DOW KOKAM. In the model settings, to simulate the capacity dispersion characteristics caused by battery aging, a battery pack consisting of six batteries with different aging levels connected in series was constructed. The initial capacity of each battery was configured differently within the range of 5.00 Ah to 4.73 Ah, and the initial SOC of all batteries was set to 100%. Based on this, a BESS consisting of three of the above battery packs connected in parallel was constructed, with battery pack 1 having the worst internal consistency and battery pack 3 having the best internal consistency. The equalization start-up threshold was set to 1.5%. The dynamic operating condition data was selected from historical frequency data of the UK power grid in 2024.
[0110] The standard deviation of SOC is introduced to evaluate the consistency between cells and between battery packs. The calculation formula is as follows:
[0111] ;
[0112] In the formula: Let be the standard deviation of the state of charge (SOC) within the battery pack at time t. for Time of the first SOC of a single battery cell for The state of charge (SOC) of the battery module at any time.
[0113] When the imbalance between batteries reaches a certain level, the machine must be shut down for maintenance to avoid safety risks. A standard deviation coefficient is used. As an indicator for maintenance early warning, select The moment when the threshold is first exceeded is designated as the maintenance time. The calculation formula is as follows:
[0114] ;
[0115] The effectiveness of this current distribution strategy was evaluated in three parallel battery packs. Figure 5 As shown, when the battery pack is in its healthy operating range, the inter-cell consistency of battery pack 1 is the worst. Therefore, the differentiated current regulation strategy between battery packs of this invention controls the current of battery pack 1. To minimize further deterioration of the balance index, battery pack 3, which exhibits better consistency, takes the opposite approach. Furthermore, in conjunction with a prediction-based bypass balance strategy within the group, each battery module... All values are strictly controlled below a preset threshold. Quantitative analysis shows that, compared to a scheme without an equalization strategy, the current initial allocation strategy proposed in this invention improves the consistency between batteries by 2.6%, and the consistency among the three battery packs by 26.0%. This result confirms that the strategy improves the consistency between and within battery packs, effectively mitigating the battery bottleneck effect, thereby maximizing the continuous operating time and available capacity of the BESS.
[0116] Furthermore, in a BESS consisting of three battery packs, the overall performance of four different strategies over a long time scale was compared. Table 1 shows the performance comparison data of no balancing strategy, only inter-pack current regulation strategy, only intra-pack prediction-based bypass balancing, and the hierarchical collaborative strategy of the present invention; Figure 6 Demonstrates battery inter-cell communication under different strategies The trajectory of change.
[0117] Table 1 Comparison of performance metrics under different equilibrium strategies
[0118] .
[0119] Table 1 and Figure 6 The results show that the hierarchical collaborative equilibrium strategy proposed in this invention significantly suppresses the imbalance interval. The divergent trend shows that the effect is superior to single-level control. Quantitative analysis shows that, compared with relying solely on intra-group balancing strategies, the hierarchical strategy proposed in this invention enhances the inter-cell control. The root mean square (RMS) value is reduced by 3.2%. While maintaining a superior balancing effect, this invention utilizes battery pack-level current regulation margin to reduce the frequent switching caused by single bypass balancing. Therefore, the total number of bypass battery cycles is reduced by 5.9%, lowering switching losses and mechanical wear of power devices and significantly improving system operating efficiency. Compared to relying solely on inter-pack balancing strategies, this invention reduces inter-cell balancing... The RMS is reduced by 4.3%, which, compared to the unbalanced scheme, makes the present invention... The RMS was reduced by 7.3%, indicating that hierarchical collaborative control can more effectively suppress imbalance divergence. The most significant advantage lies in the system's continuous operation capability. As shown in Table 1, traditional strategies or single-level strategies have limited effectiveness in mitigating the bottleneck effect, causing the system to trigger the inconsistency maintenance threshold after approximately 1.40 days of operation, requiring downtime for maintenance. This invention, through hierarchical collaboration, achieves refined and balanced control of battery state, extending the system's maintenance interval to 6.36 days, 4.5 times that of traditional solutions. In summary, this invention not only achieves optimal control of battery consistency in terms of technical indicators, but also directly demonstrates the significant engineering value of this strategy in improving BESS operating efficiency and reducing maintenance costs by extending uninterrupted operation time and reducing switching losses.
[0120] The above technical solution employs a hierarchical strategy of current regulation for inter-pack balance and SOC prediction bypass balance for intra-pack balance. A constant output closed-loop compensation mechanism is designed to address voltage drops caused by bypass actions, enabling real-time reconstruction of current commands. First, this invention constructs a hierarchical current regulation architecture and utilizes an algorithm-level output current regulation compensation mechanism to support and respond to external dynamic commands without increasing additional hardware costs or losses. Second, this invention introduces an LSTM prediction mechanism to predict and actively intervene in battery inconsistency trends, effectively overcoming the inherent lag and switching oscillation problems of traditional threshold triggering strategies. Through these mechanisms, this invention can effectively reduce switching losses while actively suppressing internal battery inconsistency divergence, thereby achieving a comprehensive improvement in battery consistency and operating efficiency within the BESS. In summary, this invention, through a dynamic hierarchical collaborative strategy of differentiated inter-pack current regulation and predictive bypass and compensation within the pack, significantly reduces switching losses while achieving a significant improvement in internal BESS consistency and precise smoothing of external output, demonstrating high engineering practical value.
[0121] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.
Claims
1. A bypass balancing method for a hierarchical and coordinated battery energy storage system, characterized in that, The battery energy storage system comprises multiple parallel battery modules, each battery module including any number of battery cells connected in series, and the balancing method includes: For any given battery module: Calculate the balance index of the battery module, wherein the balance index is obtained by subtracting the minimum SOC prediction value from the maximum SOC prediction value among the individual battery cells in the battery module. When the balance index exceeds the preset balance threshold, the target battery cell of the battery module is determined. The target battery cell is the battery cell with the lowest SOC prediction value in the battery module when the battery energy storage system is in the discharge state, or the battery cell with the highest SOC prediction value in the battery module when the battery energy storage system is in the charging state. Bypass balancing is performed on the target battery cells, and the current of the battery module is compensated by a differentiated current regulation coefficient.
2. The hierarchical collaborative bypass balancing method for battery energy storage systems according to claim 1, characterized in that, The battery energy storage system further includes a first switch and a second switch. For any given battery cell, the second switch is connected in series, and the first switch is connected in parallel to both the battery cell and the second switch. The bypass balancing of the target battery cell includes: Close the first switch connected in parallel with the target battery cell and disconnect the second switch connected in series with the target battery cell.
3. The hierarchical collaborative bypass balancing method for battery energy storage systems according to claim 1, characterized in that, The equilibrium index is obtained through the following method: The real-time SOC value of each battery cell in the battery module is obtained, and the real-time SOC value is input into the target LSTM model to obtain the predicted SOC value of the battery cell. The target LSTM model is trained by the historical SOC value of the battery cell. Among multiple SOC predictions, the equilibrium index is obtained by subtracting the minimum SOC prediction from the maximum SOC prediction.
4. The hierarchical collaborative bypass balancing method for battery energy storage systems according to any one of claims 1-3, characterized in that, The current of the compensated battery module is calculated using the following formula: ； ； ； in, To compensate for the first in the battery module The current of each battery cell To compensate for the first The current of each battery module This refers to the order of the target battery cells within the battery module. For the first battery energy storage system The port voltage of each battery module after bypass balancing For the first battery energy storage system Control current of each battery module For the first battery energy storage system The output power of each battery module For the first battery energy storage system Total output power of each battery module For the first battery energy storage system The total number of individual battery cells in a battery module.
5. The hierarchical collaborative bypass balancing method for battery energy storage systems according to claim 4, characterized in that, The first battery energy storage system Control current of each battery module This can be expressed by the following formula: ； ； in, This represents the total current demand on the grid side of the battery energy storage system. For the first battery energy storage system The current regulation coefficient of each battery module Let be the differentiated current regulation coefficient for the k-th battery module. This refers to the number of battery modules connected in parallel in a battery energy storage system. For the first battery energy storage system Boundary penalty factor for each battery module For the first battery energy storage system Internal penalty factor for each battery module For the first battery energy storage system The demand potential for individual battery modules.
6. The hierarchical collaborative bypass balancing method for battery energy storage systems according to claim 5, characterized in that, The boundary penalty factor This can be expressed by the following calculation formula: ； in, For the first battery energy storage system SOC of a battery module The lower limit SOC of the battery module, To ensure safe bandwidth buffering, This represents the upper limit of the SOC (State of Charge) of the battery module. The internal penalty factor This can be expressed by the following calculation formula: ； in, For consistency sensitivity coefficient, As a balance indicator; The demand potential energy This can be expressed by the following calculation formula: ； ； in, Adjust the sensitivity index to adjust the SOC. For the first battery energy storage system SOC of a battery module This is the control current for the battery module. The first in the battery module The remaining capacity of each battery cell The first in the battery module SOC of a single battery cell This represents the total number of individual battery cells in the battery module.