Bidirectional DC / DC conversion split type parallel DC power supply lithium battery independent management method

Abstract

The invention belongs to the technical field of direct-current power supply system management, and provides an independent management method for a split type parallel direct-current power supply lithium battery of bidirectional DC / DC conversion, a direct-current power supply is of a movable structure, and the system architecture of the direct-current power supply comprises a central controller and a plurality of parallel power sub-modules, the method specifically comprises the following steps: a central controller identifies a newly added power sub-module and completes system architecture configuration based on acquired state parameters of existing power sub-modules and main bus operation state data; the central controller dynamically selects a control mode and generates an execution instruction of charging and discharging power based on the number and state parameters of the power sub-modules, and controls the charging and discharging of each power sub-module; on the premise of ensuring uninterrupted power supply of the direct-current power supply, the central controller controls the lithium battery pack in the power sub-module to perform online capacity checking or activation operation. According to the invention, plug-and-play and independent control of the power sub-modules are realized through modular design.

Description

Technical Field

[0001] This invention relates to the field of DC power system management technology, and more particularly to a method for independent management of lithium batteries in a split-type parallel DC power supply with bidirectional DC / DC conversion. Background Technology

[0002] In the field of portable DC power supplies, lithium battery packs often employ a centralized management architecture. The high integration of the battery pack and conversion module necessitates a complete shutdown and replacement when a single battery pack fails, severely impacting power continuity. Furthermore, the lack of efficient compatibility identification and system configuration mechanisms when adding or replacing battery packs hinders flexible capacity expansion. Regarding charge and discharge control, fixed control modes struggle to adapt to the dynamic power distribution requirements of multi-module parallel scenarios, easily leading to circulating current issues or uneven power distribution, reducing system efficiency and battery life. Moreover, when DC power supplies are in a prolonged discharging or quiescent state, lithium battery packs are prone to capacity decay and performance degradation. Traditional offline capacity assessment and activation methods require power interruption, failing to meet the application requirements of uninterruptible power supplies (UPS).

[0003] To address the aforementioned issues, while existing technologies attempt to employ parallel architectures, the lack of standardized, fast-plug interfaces and interlocking mechanisms between power submodules results in insufficient reliability of physical separation and electrical connections. Furthermore, module identification relies solely on basic parameter comparisons without compatibility verification based on system operating conditions, leading to reduced system stability after capacity expansion. Regarding dynamic control strategies, centralized control suffers from heavy communication burdens and high response delays in multi-module scenarios, while traditional droop control lacks dynamic compensation mechanisms, and current sharing accuracy is significantly affected by line impedance differences. For handling AC input fluctuations, existing methods mostly employ passive adjustment, lacking an active pre-adjustment mechanism based on predictive models, making it difficult to ensure rapid bus voltage stabilization. Simultaneously, battery balancing is largely limited to static balancing between individual cells, lacking dynamic balancing and capacity collaborative management strategies based on the health status and temperature characteristics of power submodules, thus limiting the overall performance of the battery pack. Moreover, during sudden power fluctuations on the main bus, existing technologies fail to fully utilize the transient buffering capabilities of the energy storage elements in each submodule, easily causing excessive voltage fluctuations.

[0004] Patent CN120546139A proposes a parallel control method for grid-connected energy storage inverters considering State of Health (SOH) and State of Charge (SOC). This method includes n energy storage inverter modules connected in parallel to a load to supply power, a SOC balancing module considering SOH, and controls the active power setpoint of the n inverters based on their SOCi, SOH, and power setpoints. A virtual synchronous motor is input, and the corresponding phase of the i-th energy storage inverter module is output. This invention more rationally allocates the power output of each inverter compared to traditional VSG control strategies. When the battery's remaining charge is low, the output power is reduced to avoid over-discharge and extend battery life. However, it primarily targets the parallel control of grid-connected energy storage inverters and does not address the independent charging / discharging control and fault isolation mechanism for lithium battery packs in a split-type bidirectional DC / DC converter architecture. Furthermore, it does not fully couple the dynamic degradation characteristics such as battery temperature field distribution and aging rate during power allocation, making it difficult to directly apply to the full lifecycle management scenario of multi-lithium battery pack parallel DC power supply systems. In a split-type structure, each lithium battery module is connected to the main bus via an independent DC / DC converter. Its SOC and SOH states not only affect power distribution but are also closely related to the dynamic response characteristics of the DC / DC converter. Existing technologies, relying solely on inverter-side power input adjustments, struggle to achieve precise control over the charging and discharging process of each lithium battery module. This is especially problematic when multiple modules are connected in parallel, potentially leading to uneven charging and discharging of some lithium batteries due to mismatched DC / DC converter control parameters. This can negatively impact the overall system stability and the consistent lifespan of the lithium batteries. Furthermore, this method lacks charging and discharging protection strategies for lithium batteries under different operating conditions and a collaborative control mechanism between the DC / DC converter and the lithium battery, failing to meet the high-precision and high-reliability requirements of independent lithium battery management in split-type DC power systems.

[0005] Therefore, there is an urgent need for an independent lithium battery management method that can enable rapid replacement and flexible capacity expansion of power submodules, dynamic optimization of charge and discharge control, support online capacity and activation, and possess efficient fluctuation suppression and energy collaborative management capabilities. Summary of the Invention

[0006] This invention provides a method for independent management of lithium batteries in a split-parallel DC power supply with bidirectional DC / DC conversion. It addresses the problems caused by centralized management architecture in existing technologies, such as complex fault maintenance, difficulty in flexible capacity expansion, uneven dynamic power distribution, and the need for power interruption for online capacity activation. By constructing a split-parallel architecture and a multi-level collaborative management strategy, it achieves independent charging and discharging control, fast plug-in and safety interlocking, dynamic compatibility identification, adaptive power distribution, and online maintenance and capacity activation functions for lithium battery packs, thereby improving system power supply continuity, scalability, and battery life.

[0007] This invention provides a method for independent management of lithium batteries in a split-type parallel DC power supply with bidirectional DC / DC conversion. The DC power supply has a portable structure, and its system architecture includes a central controller and multiple parallel power sub-modules. The system architecture is configured as follows: on the main bus of the DC power supply, at least two independent lithium battery packs are connected in parallel through independent bidirectional DC / DC conversion modules that support quick plug-in. Each lithium battery pack and its corresponding bidirectional DC / DC conversion module are integrated into a physically separable power sub-module to enable rapid replacement of faulty lithium battery packs and flexible expansion of the system architecture. Each power sub-module has a built-in bidirectional active balancing circuit for single-cell balancing control of the lithium battery pack. The method includes the following steps: Step 1: Based on the acquired status parameters of existing power submodules and the main bus operating status data, the central controller identifies the newly added power submodules and completes the system architecture configuration; the status parameters of existing power submodules are acquired from their corresponding lithium battery packs by the slave controller built into the bidirectional DC / DC converter module. Step 2: The central controller dynamically selects the control mode and generates execution commands for charging and discharging power based on the number and status parameters of the power sub-modules, thereby controlling the charging and discharging of each power sub-module. Step 3: Under the premise of ensuring uninterrupted DC power supply, the central controller controls the lithium battery pack in the power submodule to perform online capacity verification or activation operations to address the battery degradation caused by the DC power supply being in a state of discharge or quiescence for a long time.

[0008] Furthermore, the electrical interface between the bidirectional DC / DC converter module and the main bus adopts a magnetic or snap-fit ​​quick connector, and integrates mechanical and electrical interlocks, supporting hot-swapping operation, allowing connection or disconnection without power interruption.

[0009] Furthermore, the newly added power submodule is identified and the system architecture configuration is completed, specifically including: The central controller receives and parses the identification information packet of the newly added power submodule. Combining the status parameters of the existing power submodules and the operating status data of the main bus, it verifies whether the newly connected power submodule is compatible. If the verification is successful, it updates the total number of power submodules, available capacity information, and power allocation weight coefficient. The identification information packet is collected by the slave controller of the bidirectional DC / DC module in the new power submodule from the corresponding lithium battery pack, including the lithium battery pack model, rated capacity, initial SOC, nominal voltage, maximum charge / discharge current, and SOH. The power allocation weight coefficient is calculated based on a preset capacity expansion algorithm.

[0010] Furthermore, in step two, the central controller dynamically selects the control mode and generates execution commands for charging and discharging power based on the number and status parameters of the power submodules, controlling the charging and discharging of each power submodule, including: When the number of power submodules is less than or equal to the preset number threshold, a centralized control mode is adopted. The central controller generates the execution command of charging and discharging power and sends it to the power submodules to control the charging and discharging of each power submodule. When the number of power submodules exceeds a preset threshold, a distributed control mode is adopted. The central controller sends the total charging and discharging power demand to each power submodule. The slave controllers of the bidirectional DC / DC converter modules in the power submodules autonomously negotiate power allocation through a distributed algorithm based on improved droop control and autonomously execute charging and discharging. The distributed algorithm for improved droop control is as follows: a virtual impedance dynamic compensation mechanism and an adaptive adjustment strategy for the voltage-power droop curve are introduced to optimize the current sharing accuracy among power submodules.

[0011] Furthermore, step two also includes: When any power submodule fails, disconnect the output contactor of the bidirectional DC / DC converter module in the power submodule to isolate the power submodule from the main bus. The central controller recalculates based on the current load demand of the DC power supply and the available capacity information of the power sub-modules that have not experienced any faults. It then controls the power sub-modules that have not experienced any faults to take over the current load demand according to the execution command of the charging and discharging power, so as to ensure the stability of the main bus voltage and the continuous and uninterrupted power supply of the DC power supply.

[0012] Furthermore, in step three, the online nuclear capacity or activation operation specifically includes: The central controller controls the bidirectional DC / DC converter module in the target power submodule to switch to constant current discharge mode or charge-discharge cycle mode according to a preset cycle or command; wherein, constant current discharge mode corresponds to capacity core, and charge-discharge cycle mode corresponds to activation. The bidirectional DC / DC converter module in the target power submodule collects the corresponding lithium battery pack status data from the controller in real time and uploads it to the central controller. Based on the state data of the lithium battery pack, the central controller performs capacity calculation and health status assessment of the lithium battery pack, and calculates the capacity retention rate of the lithium battery pack. If the capacity retention rate is lower than the preset threshold, the controller controls the target power submodule to perform an activation operation including at least one complete charge-discharge cycle. After the activation operation is completed, the controller controls the target power submodule to switch to the normal charge-discharge control mode.

[0013] Furthermore, when the AC input voltage fluctuates, the central controller triggers a pre-regulation mechanism for the output power of the power submodule based on a preset fluctuation threshold. This pre-regulation mechanism includes: The central controller collects the fluctuation amplitude and rate of change of AC input voltage in real time, and calculates the voltage fluctuation trend in the next 50-200ms through a sliding window algorithm. When the predicted fluctuation amplitude exceeds ±5% of the rated voltage or the rate of change exceeds 2V / ms, pre-regulation is initiated. Calculate the dynamic adjustment weighting coefficient based on the current SOC, SOH, and real-time temperature of each power submodule. , ,in, Represents the state of charge. Represents health status. For reference temperature, Let α be the real-time temperature of the power submodule, and let α, β, and γ be the weighting coefficients of SOC, SOH, and temperature deviation, respectively, and let α+β+γ=1. The central controller sends pre-adjustment commands to each power submodule. The power submodules then dynamically adjust the output power of the bidirectional DC / DC converter module according to the dynamic adjustment weighting coefficient. satisfy: , The total regulating power demand is calculated based on the AC fluctuation amplitude and the energy storage characteristics of the bus capacitor. Within 10ms after the pre-adjustment is executed, the voltage fluctuation of the main bus is collected. If the voltage fluctuation still exceeds ±2% of the rated voltage, the secondary correction is initiated. The power distribution accuracy is further optimized by introducing virtual impedance compensation until the bus voltage stabilizes within ±1% of the rated voltage.

[0014] Furthermore, it also includes dynamic balancing and capacity coordination management steps: When the slave controller of any power submodule detects that its corresponding lithium battery pack has any of the following preset conditions, the central controller controls the bidirectional active balancing circuit in the power submodule to start, so as to achieve balancing of the lithium battery pack. The bidirectional active balancing circuit is configured to be connected to the lithium battery pack. The preset conditions include: individual cell voltage difference > static balancing threshold; capacity difference during charging and discharging > dynamic balancing threshold; SOH difference > health state balancing threshold.

[0015] Furthermore, the dynamic balancing and capacity coordination management steps also include: the central controller establishing a capacity contribution model for each power submodule, with the contribution coefficient as follows: , in, Represents the battery's state of charge. This indicates the battery's health status. This represents the temperature influence factor, and its value is determined based on the deviation of the real-time operating temperature of the power submodule from the preset optimal operating temperature range. When the real-time temperature is within the optimal operating temperature range, The value is 1, meaning that when the real-time temperature is higher or lower than the optimal operating temperature range, The value decreases linearly as the degree of deviation increases, and the minimum value is not less than 0.5. a, b, and c are the weight coefficients of each parameter, and satisfy a+b+c=1. The central controller dynamically adjusts the priority of each power submodule during the charging and discharging process based on its contribution coefficient. Specifically, during charging, power submodules with higher contribution coefficients receive higher priority in charging current allocation and are charged first; during discharging, power submodules with higher contribution coefficients also receive higher priority in discharging current allocation and are discharged first. At the same time, upper and lower thresholds for the contribution coefficient are set. When the contribution coefficient of a power submodule is lower than the lower threshold, its charging and discharging operation is suspended and an alarm is issued. When the contribution coefficient of a power submodule is higher than the upper threshold, its charging and discharging current rate is appropriately reduced.

[0016] Furthermore, it also includes a distributed energy buffering coordination step, specifically: the slave controller of each power submodule analyzes the main bus voltage fluctuation characteristics in real time through a sliding window algorithm, and predicts the main bus power demand change ΔP_pred within a preset time period based on the LSTM network; when ΔP_pred is predicted to exceed a preset threshold, a distributed buffering strategy is activated; the distributed buffering strategy is as follows: the energy storage inductor L of each power submodule enters a transient energy storage mode, and absorbs or releases energy by changing the current change rate di / dt; each power submodule coordinates to undertake this buffering task according to its own SOC state in a preset proportion.

[0017] Compared with the prior art, the present invention has the following advantages and beneficial effects: First, the split-parallel architecture achieves fault isolation and elastic expansion. The failure of a single power submodule only affects itself and will not cause the entire system to fail, effectively reducing the complexity of fault maintenance. Secondly, by adopting a multi-level collaborative management strategy and combining the dynamic switching between centralized and distributed control modes, the control logic can be flexibly adjusted according to the system scale and operating status, thereby improving the balance of dynamic power distribution. Third, by supporting the rapid plugging and unplugging of power submodules and the safety interlocking mechanism, combined with the dynamic compatibility identification algorithm, flexible access and online maintenance of new battery packs have been achieved, solving the problem of power interruption required for capacity expansion under the traditional centralized architecture; Fourth, by integrating the bidirectional DC / DC converter module with the lithium battery pack design, and combining the specific steps of online capacity / activation and the SOH prediction model, the capacity detection and performance recovery of the battery pack can be completed without affecting the continuity of system power supply, which significantly extends the overall service life of the battery. Fifth, a fuzzy logic dynamic power allocation and capacity contribution model is introduced to dynamically adjust the charging and discharging priority based on the SOC, SOH and temperature parameters of each power sub-module, avoiding local overload or capacity waste caused by differences in battery characteristics and improving the system's energy utilization efficiency.

[0018] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0019] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0020] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 A schematic diagram illustrating the steps of a separate parallel DC power supply for bidirectional DC / DC conversion with independent lithium battery management. Figure 2 A schematic diagram illustrating the steps of controlling the charging and discharging of each power submodule in order to dynamically select the control mode and generate execution instructions for charging and discharging power. Figure 3 This is a schematic diagram of the steps involved in online nuclear capacity or activation procedures. Detailed Implementation

[0021] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0022] This invention provides a method for independent management of lithium batteries in a split-type parallel DC power supply with bidirectional DC / DC conversion, such as... Figure 1 As shown, the DC power supply has a portable structure. The system architecture includes a central controller and multiple parallel power sub-modules. The system architecture is configured as follows: On the main bus of the DC power supply, at least two independent lithium battery packs are connected in parallel via independent bidirectional DC / DC converter modules that support quick plug-in / out. Each lithium battery pack and its corresponding bidirectional DC / DC converter module are integrated into a physically separable power sub-module to enable rapid replacement of faulty lithium battery packs and flexible expansion of the system architecture. Each power sub-module has a built-in bidirectional active balancing circuit for single-cell balancing control of the lithium battery packs. The central controller communicates with the slave controllers of each power sub-module via a CAN bus. The power sub-modules are electrically connected to the main bus via magnetic quick-connect connectors, with a rated current of not less than 50A. The method includes the following steps: Step 1: Based on the acquired status parameters of existing power submodules and the main bus operating status data, the central controller identifies the newly added power submodules and completes the system architecture configuration; the status parameters of existing power submodules are acquired from their corresponding lithium battery packs by the slave controller built into the bidirectional DC / DC converter module. Step 2: The central controller dynamically selects the control mode and generates execution commands for charging and discharging power based on the number and status parameters of the power sub-modules, thereby controlling the charging and discharging of each power sub-module. Step 3: Under the premise of ensuring uninterrupted DC power supply, the central controller controls the lithium battery pack in the power submodule to perform online capacity verification or activation operations to address the battery degradation problem caused by the DC power supply being in a state of discharge or quiescence for a long time. In specific applications, taking a backup power system of a communication base station as an example, in emergency scenarios, this backup power system of the communication base station originally used a traditional fixed lead-acid battery pack, which had problems such as difficulty in replacement and high system expansion costs. In order to solve the above problems, the base station adopted the bidirectional DC / DC conversion split parallel DC power lithium battery independent management scheme provided by the present invention. This DC power supply system is a portable structure. Its system architecture includes a central controller and multiple parallel power sub-modules. On the main bus of the DC power supply (nominal voltage value is DC220V), three independent lithium battery packs (model can be: LFP-120Ah / 48V, each pack consists of 16 3.2V 120Ah lithium iron phosphate batteries connected in series) are connected in parallel through independent bidirectional DC / DC converter modules that support quick plug-in. Each lithium battery pack and its corresponding bidirectional DC / DC converter module are integrated into a physically separable power sub-module through mechanical structure and electrical interface (size can be: 400mm×300mm×200mm, weight can be: approximately 25kg). Each power sub-module is equipped with an independent cooling fan and status indicator lights (power, operation, fault), and is designed with standardized DIN rail mounting slots and quick plug-in aviation plugs to realize the rapid replacement of faulty lithium battery packs and the flexible expansion of the system architecture. In step one, assuming the system is initially running with two power submodules (submodule A and submodule B) functioning normally, a new power submodule C needs to be added to increase system capacity. Technicians slide the new power submodule C into the designated position in the cabinet via a guide rail and insert its quick-plug aviation connector into the corresponding socket on the main busbar, completing the mechanical and electrical connection. Next, after the bidirectional DC / DC converter module of power submodule C is powered on and started by the built-in controller (model: MCU-DCDC-01), it immediately collects parameters from its corresponding lithium battery pack C. The collected status parameters include: individual battery voltage (16 channels), total voltage, total current, battery pack temperature (3 measurement points: battery pack surface, positive electrode, negative electrode), initial estimated SOC (State of Charge), initial estimated SOH (State of Health), and battery pack cycle time. The controller collects the status parameters of the lithium battery pack C, such as the number of cycles, the temperature of the bidirectional DC / DC converter module itself, and the input and output voltage and current. These parameters are then reported to the central controller in real time via the internal CAN bus. The central controller scans the devices on the bus in real time. When it detects the handshake signal and parameter data packet sent by the controller of the newly connected power submodule C, it initiates the new module identification process. The central controller first verifies whether the communication protocol version and hardware ID of the submodule C are compatible with the system. After verification, it reads the key parameters reported by the submodule C, such as the lithium battery pack type, rated capacity, nominal voltage, and maximum charge and discharge current. At the same time, the central controller obtains the current operating status data of the main bus (nominal voltage DC220V), including the main bus voltage, total output current, load power, and the current charging and discharging mode of the system (float charging / equalizing charging / discharging). Based on the acquired status parameters of existing power submodules A and B (such as current SOC of 85% and 83% respectively, and SOH of 92% for both) and main bus operating status data (main bus voltage 220.5V, load power 200W), the central controller performs system architecture configuration for the newly added power submodule C: assigning a unique submodule address (0x03) to submodule C, updating the system power submodule topology information, incorporating submodule C into the unified system management, setting its initial working mode to "standby", and limiting its initial charging and discharging current to 20% of the rated value (i.e., 10A) to avoid impacting the system. The central controller sends a configuration confirmation command to submodule C, and the status indicator light of submodule C changes from "power" light to "running" light, indicating that identification and configuration are complete and submodule C is ready to participate in system charging and discharging. In step two, assuming the base station is powered normally, the AC mains power supplies the main bus (nominal voltage DC220V) through the rectifier module and charges the lithium battery pack; when the mains power is interrupted, the lithium battery pack discharges through the bidirectional DC / DC converter module to ensure the main bus (nominal voltage DC220V) is powered. In step three, assuming that the mains power supply of the base station is stable, the DC power system (the nominal voltage of the main bus is DC220V) is in a floating charge state for a long time, which may lead to a decrease in battery activity and an inflated capacity. To address this issue, the central controller performs online capacity verification on each lithium battery pack periodically (e.g., every 3 months). The central controller is set to perform online capacity verification at 2:00 AM on the 1st of each month (during the off-peak period when the base station load power drops to about 150W). When submodule A to be charged is removed from normal charging and discharging, the central controller sends a command to submodule A to control its bidirectional DC / DC converter module to stop charging and discharging, enter "charge-capacity mode", and softly isolate it from the main bus (controlling its output current to 0); Submodule A discharges to the cutoff voltage: The central controller commands the bidirectional DC / DC converter module of submodule A to work in discharge mode, discharging a built-in high-power resistive load (only enabled during charge-capacity) with a constant current of 0.2C (24A). The controller monitors the total voltage and individual cell voltage of the lithium battery pack of submodule A in real time. When the total voltage drops to 40V (or any individual cell voltage drops to 2.5V), the discharge stops, and the discharge time t1 is recorded (assuming the discharge time is 4 hours and 50 minutes, i.e. 4.833 hours). The actual discharged capacity Cdischarge = 24A * 4.833h ≈ 116Ah; after discharge, the central controller instructs the bidirectional DC / DC converter module of submodule A to switch to charging mode, draw power from the main bus (nominal voltage DC220V), and charge the lithium battery pack A with a constant current of 0.2C (24A). When the total voltage reaches 54V (or the voltage of any single cell reaches 3.65V), it switches to constant voltage charging until the charging current drops below 0.02C (2.4A) and stops charging. The charging time t2 is recorded, and the actual charged capacity Ccharge is calculated; the central controller updates the actual capacity of the lithium battery pack in submodule A based on the average value of Cdischarge and Ccharge (taking Cdischarge because discharge is closer to the actual usable capacity), and recalculates its SOH. During the capacity verification and discharge process, the central controller controls submodules B and C to bear the entire load (150W) of the main bus (nominal voltage DC220V) and the power consumed by the discharge resistor of submodule A. The bidirectional DC / DC converter modules of submodules B and C increase the output current according to the instructions of the central controller to ensure that the main bus voltage is stable within the range of 220V±0.5V, ensuring uninterrupted power supply to the base station equipment. After the capacity verification of submodule A is completed, the central controller switches it back to the normal charging and discharging mode and restores its participation in system power distribution. Subsequently, the central controller performs online capacity verification operations on power submodules B and C in the same manner. If it is found that the actual capacity of a lithium battery pack is less than 80% of the rated capacity during the capacity verification process, the central controller issues an alarm signal to prompt the maintenance personnel to replace it.

[0023] The working principle of the above technical solution is as follows: The core working principle of this bidirectional DC / DC converter split parallel DC power supply lithium battery independent management method is to realize the independent and efficient management of lithium battery pack, system elastic expansion and uninterrupted power supply guarantee through modular design and intelligent collaborative control.

[0024] Firstly, at the system architecture level, the DC power supply adopts a portable structure, with a central controller at its core and multiple parallel power sub-modules. Each power sub-module consists of an independent lithium battery pack and an integrated bidirectional DC / DC converter module. This integrated design makes the power sub-module a physically separable unit. Crucially, these power sub-modules are connected in parallel to the main bus of the DC power supply via a quick-plug bidirectional DC / DC converter module. This design directly gives the system two core advantages: firstly, when a lithium battery pack fails, its corresponding power sub-module can be quickly plugged in and replaced, achieving rapid replacement of the faulty lithium battery pack and greatly shortening maintenance time; secondly, the system architecture can be flexibly expanded or reduced in capacity by simply increasing or decreasing the number of power sub-modules to adapt to different power demand scenarios.

[0025] Secondly, during the new power submodule integration and configuration phase (Step 1), the central controller plays a crucial role in system integration. When a new power submodule is added to the system, the central controller does not blindly accept it, but rather makes a comprehensive judgment and configuration based on two aspects of information: one is the status parameters obtained from the power submodules already running online. These status parameters are collected by the slave controllers built into the bidirectional DC / DC converter modules within each power submodule, which collect information about their corresponding lithium battery packs (such as voltage, current, temperature, SOC, SOH, etc.) and report them to the central controller; the other is the current operating status data of the main bus (such as bus voltage, total output current / power, etc.). The central controller uses this information to identify, verify compatibility, and configure the parameters of the newly added power submodule, ensuring that the new module can seamlessly integrate into the existing system and work collaboratively with other modules.

[0026] Secondly, in the daily charging and discharging management and control phase (step two), the central controller makes global optimization decisions based on the actual number of power submodules currently connected to the system and the real-time status parameters (such as SOC, health status, and charging / discharging capacity of each lithium battery pack) reported by each module from the controller. Its core task is to dynamically select the most suitable control mode (e.g., voltage equalization control, SOC equalization control, or power distribution control), and generate specific charging and discharging power execution commands based on the selected control mode and the overall charging and discharging requirements of the system (such as the power deficit or surplus of the main bus). These commands are sent to the slave controllers of each power submodule, thereby controlling the operating state of the bidirectional DC / DC converter module (operating in Buck mode to discharge to the bus, or Boost mode to charge from the bus), to precisely control the charging and discharging power and direction of each power submodule, achieving optimized energy allocation and efficient utilization of the entire DC power system, while ensuring the charging and discharging safety of each lithium battery pack.

[0027] Finally, in the lithium battery pack maintenance and performance assurance phase (step three), to address the battery performance degradation caused by the DC power supply potentially being in a discharged state for extended periods (leading to untimely equalization after deep discharge) or a quiescent state (leading to uneven self-discharge and passivation of active materials), the central controller is designed with an online capacity verification and activation mechanism. Crucially, these maintenance operations are performed under the premise of "ensuring uninterrupted DC power supply." This means the central controller will intelligently schedule operations. For example, when a lithium battery pack needs capacity verification (precisely measuring its actual capacity) or activation (restoring its activity through specific charge-discharge cycles), it will first ensure that other parallel power submodules can temporarily handle the power supply or load requirements when that module is disconnected. Then, it will control the target power submodule to disconnect from the main bus or enter a specific maintenance mode for operation. After the operation is completed, it will be smoothly reconnected to the system. In this way, the system can perform regular health maintenance on each lithium battery pack without interrupting external power supply, effectively extending battery life and ensuring long-term stable system operation.

[0028] The beneficial effects of the above technical solution are as follows: By adopting the solution provided in this embodiment, through a modular architecture of separate parallel connections, intelligent identification and configuration by the central controller, dynamic power distribution control, and online maintenance scheduling, independent management of lithium battery packs, flexible system expansion, efficient charging and discharging, and high availability are jointly achieved. The separate parallel architecture enables each lithium battery pack and bidirectional DC / DC converter module to form an independent functional unit, avoiding the impact of a single battery failure on the overall system in traditional centralized management, thus improving the system's fault tolerance and reliability. The intelligent identification and configuration of newly connected modules by the central controller simplifies the system expansion process and reduces costs. It reduces the complexity of manual operation, ensuring that new modules can be quickly integrated into the existing system and work collaboratively; dynamic power distribution control can accurately adjust the charging and discharging power according to the real-time status of each battery pack, which not only maximizes energy utilization but also effectively prevents overcharging and over-discharging, ensuring battery safety; the online maintenance scheduling mechanism solves the problem of traditional system maintenance requiring downtime. By intelligently scheduling other modules to temporarily take over the load, it can complete the capacity verification and activation of the target battery pack without interrupting power supply, significantly improving the system's continuous operation capability and battery lifespan, and overall meeting the application requirements of high reliability, high scalability, and long lifespan for DC power supply systems.

[0029] In one embodiment, the electrical interface between the bidirectional DC / DC converter module and the main bus adopts a magnetic or snap-fit ​​quick connector, and integrates mechanical and electrical interlocks, supporting hot-swapping operation, allowing connection or disconnection without power interruption.

[0030] The working principle of the above technical solution is as follows: the bidirectional DC / DC converter module and the main bus are mechanically connected and electrically connected via magnetic or snap-fit ​​quick connectors. An integrated mechanical interlock ensures that the connector is correctly aligned and locked before establishing an electrical connection, preventing misoperation; the electrical interlock monitors the connection status and only allows current to flow after confirming a secure mechanical connection. During hot-swapping, the connector's structural design and interlocking mechanism work together to ensure that, under energized conditions, auxiliary signals or control circuits are reliably connected / disconnected first, followed by safe connection / disconnection of the main power circuit, avoiding arcing or equipment damage. This allows for safe and rapid connection or disconnection of the module and the main bus without power interruption.

[0031] The beneficial effects of the above technical solution are as follows: By adopting the solution provided in this embodiment, the combination of magnetic or snap-fit ​​quick connectors with mechanical and electrical interlocks improves the maintenance convenience of the bidirectional DC / DC converter module and the continuity of system operation.

[0032] In one embodiment, the newly added power submodule is identified and the system architecture configuration is completed, specifically including: The central controller receives and parses the identification information packet of the newly added power submodule. Combining this with the status parameters of existing power submodules and the operating status data of the main bus, it verifies the compatibility of the newly added power submodule. If the verification passes, it updates the total number of power submodules, available capacity information, and power allocation weighting coefficient. The identification information packet is collected by the slave controller of the bidirectional DC / DC module in the new power submodule from the corresponding lithium battery pack, including the lithium battery pack's model, rated capacity, initial SOC, nominal voltage, maximum charge / discharge current, and SOH. The power allocation weighting coefficient is calculated based on a preset capacity expansion algorithm. This algorithm is based on the ratio of the rated capacity of the new power submodule to the existing power submodules, the current SOC balance, and the SOH health status, and incorporates a dynamic adjustment factor. The factors comprehensively consider the real-time load demand fluctuations of the main bus, the differences in the temperature field distribution of the battery pack, and the historical charge-discharge cycle decay characteristics. The optimal weight coefficient matrix is ​​solved through a multi-objective optimization algorithm. During the weight coefficient calculation process, a capacity decay compensation coefficient is set for power sub-modules with SOH below the health threshold, and a temperature impact correction term is added for battery packs in extreme temperature ranges. This ensures that the power allocation meets the current load power supply requirements while maximizing the balance of cycle life loss of each lithium battery pack. After the weight coefficient is updated, the central controller issues a new power allocation command to all power sub-modules. Newly connected modules gradually increase their output to the target value according to the command, and existing modules adjust their output power according to the weight. The entire process is achieved through a smooth transition control strategy to avoid impacting the voltage and current of the main bus. The complete execution process of the capacity expansion algorithm is as follows: First, the central controller collects the operating parameters of all power sub-modules (including newly connected and existing modules) in real time, covering the rated capacity, current state of charge (SOC), state of health (SOH), real-time output power, temperature value and its temperature range (low temperature / normal temperature / high temperature) of each module. At the same time, it acquires key information such as the load current and voltage fluctuation of the main bus and the number of historical charge and discharge cycles. In addition, distributed temperature sensors are used to monitor the spatial distribution of the battery pack temperature field, identify local hot spots or cold spots, and provide a basis for subsequent thermal management. Subsequently, the collected multidimensional heterogeneous data were standardized and normalized to eliminate the dimensional differences between different physical quantities, and the SOC was uniformly mapped to the [0,1] interval. The SOH was calculated based on the initial capacity, and 80% was set as the health threshold. Modules below this value were considered to be aging. The temperature effect was divided into three levels: -20°C to 10°C was the low-temperature suppression zone, 10°C to 45°C was the high-efficiency operating zone, and above 45°C was the high-temperature derating zone. Corresponding correction weights were assigned to each level to ensure that environmental factors were reasonably reflected in the decision-making process. Based on this, a dynamic adjustment factor is constructed to characterize the comprehensive response capability of the i-th power submodule under the current operating conditions. It is composed of three core components: load response factor, temperature field equalization factor, and historical attenuation compensation factor. Next, the capacity ratio initialization weight is set, and subsequent optimization and correction will be further integrated with health status, temperature conditions, and dynamic response capability. The process then proceeds to the multi-objective optimization modeling stage, constructing a comprehensive optimization function containing three sub-objectives: the first objective is to minimize the total power deviation, ensuring that the system output accurately matches the load demand; the second objective aims to minimize the SOC dispersion among modules, improving the system consistency level; and the third objective focuses on the balanced control of lifetime loss, extending the overall service life. The comprehensive objective function is a weighted sum of these three objectives, solved using an improved NSGA-II algorithm. This algorithm searches for the Pareto front in the feasible solution set that satisfies all constraints, selecting the solution closest to the ideal point as the optimal weight coefficient matrix. After obtaining the optimized weights, the process moves to the weight fusion and instruction generation stage: the original capacity weights and the optimized weights are smoothly weighted and fused to prevent sudden power allocation changes from causing system disturbances. Subsequently, the central controller broadcasts the updated allocation weights to all power submodules. For newly connected modules, a ramp-up control strategy is adopted, with the initial output set to zero and the output power increased in stages until the target value is reached, effectively suppressing inrush current. Existing modules synchronously adjust their output power according to the new weights, with the adjustment rate limited to within ±10% of rated power / second to ensure dynamic stability of the system. After the expansion is completed, the system enters the steady-state operation phase. Every 300 seconds, a complete re-optimization process is automatically triggered to re-sensitize the state, reconstruct the model and update the weight matrix to achieve continuous dynamic optimization. If a module is significantly down-weighted in three consecutive optimization cycles and its SOH shows a continuous downward trend, it is marked as a unit to be maintained and included in the warning list to prompt the operation and maintenance personnel to arrange inspection or replacement.

[0033] The working principle of the above technical solution is as follows: After the newly added power submodule is connected to the system, the slave controller of its internal bidirectional DC / DC module will first collect parameters of its corresponding lithium battery pack. The collected parameters include the lithium battery pack model, rated capacity (referring to the standard electrical energy that the battery pack can provide), initial SOC (State of Charge, which is the initial state of charge of the battery pack, representing the percentage of the current battery capacity relative to the rated capacity), nominal voltage (the standard voltage when the battery pack is working normally), maximum charging and discharging current (the maximum charging and discharging current value allowed by the battery pack under safe conditions), and SOH (State of Health, which reflects the degree of degradation of the current performance of the battery relative to the nominal performance). These collected parameters are integrated and packaged to form an identification information package. Subsequently, the identification information packet is sent to the system's central controller. After receiving the identification information packet, the central controller parses it and extracts various key parameters of the new power submodule lithium battery pack. Then, the central controller comprehensively considers and compares these new parameters with the current status parameters of other existing power submodules in the system (such as the remaining capacity, charge and discharge capacity, health status, etc. of these submodules) and the real-time operating status data of the main bus (such as the current voltage level, total load power, current magnitude, power flow direction, etc. of the main bus). Through this multi-dimensional information integration and analysis, the core task of the central controller is to verify whether the newly connected power submodules are compatible with the existing system in terms of electrical characteristics, communication protocols, capacity matching, and safe operation. If the verification confirms that the new power submodules are compatible with the system, the central controller will execute a series of system architecture configuration update operations. First, it updates the total number of power submodules in the system; second, it updates the total available capacity information of the system based on parameters such as the rated capacity of the new submodules; finally, and crucially, the central controller will calculate and allocate an appropriate power allocation weight coefficient for the newly connected power submodules based on a preset expansion algorithm, combined with the various performance parameters of the new submodules (such as rated capacity, SOH, maximum charging and discharging current, etc.) and the status of existing submodules and the load of the main bus. This weight coefficient will determine the proportion or priority that the new submodule should bear in the total power when the system performs power allocation or load scheduling, thereby ensuring that all power submodules, including the new submodules, can operate collaboratively, efficiently, safely, and stably, jointly providing reliable power support or absorption for the system.

[0034] The beneficial effects of the above technical solution are as follows: by adopting the solution provided in this embodiment, it is possible to achieve automated identification and compatibility verification during the access process of new power submodules, avoid the risk of parameter mismatch that may be caused by manual configuration, and significantly improve the convenience and security of system expansion.

[0035] In one embodiment, such as Figure 2 As shown, in step two, the central controller dynamically selects the control mode and generates execution commands for charging and discharging power based on the number and status parameters of the power submodules, controlling the charging and discharging of each power submodule, including: When the number of power submodules is less than or equal to a preset threshold, a centralized control mode is adopted. The central controller generates an execution command for charging and discharging power and sends it to the power submodules to control the charging and discharging of each power submodule. The preset threshold is 3-5. When the number of power submodules exceeds a preset threshold, a distributed control mode is adopted. The central controller sends the total charging and discharging power demand to each power submodule. The slave controllers of the bidirectional DC / DC converter modules in the power submodules autonomously negotiate power allocation through a distributed algorithm based on improved droop control and autonomously execute charging and discharging. The distributed algorithm for improved droop control is as follows: a virtual impedance dynamic compensation mechanism and an adaptive adjustment strategy for the voltage-power droop curve are introduced to optimize the current sharing accuracy among power submodules.

[0036] The working principle of the above technical solution is as follows: based on the number of power sub-modules in the system and their status parameters, an appropriate control mode is intelligently and dynamically selected, and corresponding charging and discharging power execution commands are generated accordingly, thereby achieving precise control over the charging and discharging process of each power sub-module. Specifically, this dynamic control logic is reflected in two aspects: when the number of power sub-modules in the system is less than or equal to a preset threshold, the central controller will start a centralized control mode. In this mode, the central controller assumes all decision-making and command generation responsibilities. It directly calculates and generates specific charging and discharging power execution commands, and then sends these commands one by one to each power sub-module, thereby directly controlling the charging and discharging behavior of each power sub-module and ensuring that the entire system operates in a unified and coordinated manner according to the central command. When the number of power submodules exceeds a preset threshold, the system automatically switches to distributed control mode. In this mode, the central controller no longer issues individual execution commands but instead broadcasts a total charging and discharging power demand to all power submodules. Upon receiving this total demand, the slave controllers within each power submodule's internal bidirectional DC / DC converter module play an autonomous coordination role. These slave controllers negotiate power allocation autonomously by running a distributed algorithm based on improved droop control. The core improvement of this distributed algorithm lies in the introduction of a virtual impedance dynamic compensation mechanism and an adaptive adjustment strategy for the voltage-power droop curve. Traditional droop control simulates the primary frequency regulation characteristics of a synchronous generator by setting a negative feedback relationship between voltage and power (VP) or current and power (IP), allowing each module to independently adjust its output power based on local information, thereby achieving power sharing without communication. However, in practical applications, due to differences in line impedance, the discreteness of device parameters, and dynamic changes in load, traditional droop control is prone to causing deviations in active / reactive power distribution, affecting current sharing accuracy, especially in multi-module parallel systems. To improve current sharing performance, this embodiment introduces two key improvement mechanisms: a virtual impedance dynamic compensation mechanism and an adaptive adjustment strategy for the voltage-power droop curve; First, the virtual impedance dynamic compensation mechanism injects adjustable virtual impedance components into the control loops of each power submodule, effectively adjusting the equivalent output impedance at the converter output to reduce circulating current and power distribution imbalance caused by inconsistent actual line impedance. This virtual impedance is not a physical component, but is estimated in real time by acquiring local voltage and current signals from the controller and combining them with the status information of neighboring modules obtained through inter-module communication (or quasi-communication). In specific implementation, the resistive and inductive components of the virtual impedance are dynamically updated according to the power deviation and voltage difference of adjacent modules: when the output power of a module is significantly higher than the average value, its virtual inductive reactance is increased to reduce its power output; conversely, it is decreased. This process achieves limited information interaction through low-bandwidth communication or broadcasting, ensuring a balance between dynamic response speed and system stability. Secondly, to address the problem of balancing "static error" and "dynamic response" caused by the traditional fixed droop coefficient, an adaptive adjustment strategy for the voltage-power droop curve is proposed. This strategy employs an online adjustment method based on real-time operational status feedback. Each power submodule's slave controller continuously monitors the deviation between its own output power and the system's average power, and, combined with the bus voltage fluctuation amplitude, constructs an evaluation function. ,in, Represents the evaluation function. This represents the deviation between its own output power and the system's average power. This represents the amplitude of bus voltage fluctuation. , These are weighting coefficients used to balance the priority of current sharing accuracy and voltage stability; the controller bases them on... The droop coefficient is dynamically adjusted according to the changing trend. As shown below, Where k represents time, This represents the droop coefficient at the next moment. ) represents the droop coefficient at the current moment. The learning step size is used to control the magnitude of coefficient adjustments. To evaluate the gradient estimate of the evaluation function, when the evaluation function increases, it indicates that the current sharing accuracy or voltage stability of the system deteriorates under the current droop coefficient, and the controller will adjust along the negative gradient direction. In other words, if the gradient estimate of the evaluation function is positive, the droop coefficient at the current moment is reduced to decrease the sensitivity of power output to voltage changes, which may improve current sharing or improve voltage fluctuations; if the gradient estimate of the evaluation function is negative, the droop coefficient at the current moment is increased to enhance the response of power output to voltage changes and promote faster power balance. Through this adaptive mechanism, the system can dynamically optimize droop characteristics under different operating conditions (such as load changes, module input / output), and significantly improve the static accuracy and dynamic response speed of power distribution while ensuring that the bus voltage fluctuation is within the allowable range. The virtual impedance dynamic compensation mechanism can effectively suppress the uneven power distribution caused by factors such as line impedance differences. By dynamically adjusting the virtual impedance, each module can exhibit relatively consistent external characteristics under different operating conditions. The adaptive adjustment strategy of the voltage-power droop curve allows key parameters such as the slope and intercept of the droop curve to be automatically adjusted according to the real-time operating status of the system (such as total power demand, SOC (state of charge) of each module, temperature and other status parameters). This further optimizes the current sharing accuracy among power sub-modules while ensuring system stability. Ultimately, each power sub-module can autonomously perform charging and discharging operations according to the negotiation results, share and complete the total charging and discharging power demand issued by the central controller, and achieve efficient, balanced and stable power distribution and execution under large-scale module configuration.

[0037] The beneficial effects of the above technical solution are as follows: by adopting the solution provided in this embodiment, the centralized and efficient management of small-scale systems and the distributed and flexible collaboration of large-scale systems can be taken into account by dynamically switching control modes, which significantly improves the adaptability and scalability of lithium battery pack management.

[0038] In one embodiment, step two further includes: When any power submodule fails, disconnect the output contactor of the bidirectional DC / DC converter module in the power submodule to isolate the power submodule from the main bus. The central controller recalculates based on the current load demand of the DC power supply and the available capacity information of the power sub-modules that have not experienced any faults. It then controls the power sub-modules that have not experienced any faults to take over the current load demand according to the execution command of the charging and discharging power, so as to ensure the stability of the main bus voltage and the continuous and uninterrupted power supply of the DC power supply.

[0039] The working principle of the above technical solution is as follows: When any power submodule in the system fails, the power submodule will immediately disconnect the contactor at the output end of its internal bidirectional DC / DC converter module. The direct result of this action is to electrically isolate the faulty power submodule from the main bus of the system, prevent the faulty submodule from having an adverse effect on the voltage stability of the main bus, and also avoid further expansion of the fault range, ensuring that the main bus is physically separated from the faulty part. After that, the central controller first obtains the current load demand of the DC power supply in real time, which is the amount of power that the system actually needs to consume. At the same time, the central controller will also collect the available capacity information of all power sub-modules that have not failed. This information reflects the maximum charging and discharging capacity that each normal power sub-module can provide. Based on the current load demand and the available capacity data of the power sub-modules that have not failed, the central controller will recalculate and formulate new charging and discharging power execution instructions. Subsequently, the central controller, based on the new execution instructions, coordinates and controls all power sub-modules that have not experienced any faults, directing them to jointly undertake the current load demand.

[0040] The beneficial effects of the above technical solution are as follows: By adopting the solution provided in this embodiment, through this dynamic adjustment and allocation, the non-faulty power sub-modules can work together to make up for the power gap caused by the exit of the faulty sub-module, thereby ensuring that the voltage of the main bus can be maintained within a stable normal operating range, and ultimately realizing the continuous and uninterrupted reliable power supply of the DC power supply to the load.

[0041] In one embodiment, in step three, such as Figure 3 As shown, the online nuclear capacity or activation operation specifically includes: The central controller controls the bidirectional DC / DC converter module in the target power submodule to switch to constant current discharge mode or charge-discharge cycle mode according to a preset cycle or command; wherein, constant current discharge mode corresponds to capacity core, and charge-discharge cycle mode corresponds to activation. The bidirectional DC / DC converter module in the target power submodule collects the corresponding lithium battery pack status data from the controller in real time and uploads it to the central controller. Based on the state data of the lithium battery pack, the central controller performs capacity calculation and health status assessment of the lithium battery pack, calculates the capacity retention rate of the lithium battery pack, and if the capacity retention rate is lower than the preset threshold, it controls the target power submodule to perform an activation operation including at least one complete charge and discharge cycle. After the activation operation is completed, it controls the target power submodule to switch to the normal charge and discharge control mode. The working principle of the above technical solution is as follows: the central controller actively triggers online capacity balancing or activation operations on the target power submodule according to a preset time period or received external instructions; specifically, the central controller sends a control signal to the bidirectional DC / DC converter module in the target power submodule to switch its working mode. When capacity balancing is required, the bidirectional DC / DC converter module is switched to constant current discharge mode; and when activation is required, it is switched to charge-discharge cycle mode. During the process of the target power submodule performing capacity or activation operations, the slave controller inside the submodule continuously and in real time collects various status data of the lithium battery pack connected to it. These status data typically include, but are not limited to, key parameters such as battery voltage, discharge current, battery temperature, and discharge time. The slave controller uploads these real-time status data to the central controller through the communication link to provide a basis for subsequent analysis and decision-making. After receiving the status data of the lithium battery pack, the central controller will use specific algorithms and models to comprehensively analyze and process the data. First, based on the collected data such as discharge current, discharge time and voltage change, it will accurately calculate the current actual capacity of the lithium battery pack. Then, combined with the initial information such as the nominal capacity of the lithium battery pack, it will further evaluate its health status. One of the core indicators is to calculate the capacity retention rate (i.e., the percentage of the current actual capacity to the nominal capacity). The central controller compares the calculated capacity retention rate with a preset threshold. If the capacity retention rate is higher than or equal to the preset threshold (80% of the rated capacity), it indicates that the lithium battery pack is in good condition and may not require activation. It will return to normal after the capacity recovery operation is completed. If the capacity retention rate is lower than the preset threshold, it indicates that the performance of the lithium battery pack has degraded and activation is required to try to restore its capacity. At this time, the central controller will immediately control the target power submodule to perform the activation operation. This operation specifically includes at least one complete charge-discharge cycle, that is, first charging the lithium battery pack to full charge with constant current and constant voltage, and then discharging it to the set cutoff voltage with constant current to activate the active materials inside the battery. After the activation operation is completed, the central controller will check the capacity of the lithium battery pack again (possibly through a new round of capacity verification operation) to confirm the activation effect. Regardless of the activation effect, after the activation operation is completed, the central controller will control the bidirectional DC / DC converter module in the target power submodule to switch from the activation mode (charge and discharge cycle mode) back to the normal charge and discharge control mode, so that the lithium battery pack is reintegrated into the system and continues to participate in the energy storage and release process according to the conventional charge and discharge strategy, thereby completing the closed-loop control of the entire online capacity verification and activation.

[0042] The beneficial effects of the above technical solution are as follows: by adopting the solution provided in this embodiment, the lithium battery pack can be subjected to online capacity testing and activation without interrupting the normal operation of the system, which effectively solves the system downtime problem caused by traditional offline maintenance.

[0043] In one embodiment, when the AC input voltage fluctuates, the central controller triggers a pre-regulation mechanism for the output power of the power submodule based on a preset fluctuation threshold. The pre-regulation mechanism includes: The central controller collects the fluctuation amplitude and rate of change of AC input voltage in real time, and calculates the voltage fluctuation trend in the next 50-200ms through a sliding window algorithm. When the predicted fluctuation amplitude exceeds ±5% of the rated voltage or the rate of change exceeds 2V / ms, pre-regulation is initiated. Calculate the dynamic adjustment weighting coefficient based on the current SOC, SOH, and real-time temperature of each power submodule. , ,in, Represents the state of charge. Represents health status. For reference temperature, Let α be the real-time temperature of the power submodule, and let α, β, and γ be the weighting coefficients of SOC, SOH, and temperature deviation, respectively, and let α+β+γ=1. The central controller sends pre-adjustment commands to each power submodule. The power submodules then dynamically adjust the output power of the bidirectional DC / DC converter module according to the dynamic adjustment weighting coefficient. satisfy: , The total regulating power demand is calculated based on the AC fluctuation amplitude and the energy storage characteristics of the bus capacitor. Within 10ms after the pre-adjustment is executed, the voltage fluctuation of the main bus is collected. If the voltage fluctuation still exceeds ±2% of the rated voltage, the secondary correction is initiated. The power distribution accuracy is further optimized by introducing virtual impedance compensation until the bus voltage stabilizes within ±1% of the rated voltage.

[0044] The working principle of the above technical solution is as follows: When the AC input voltage fluctuates, the system uses the following mechanism to pre-adjust the output power of the power sub-module to maintain system stability. First, the central controller continuously monitors the fluctuation amplitude and rate of change of the AC input voltage, and uses a sliding window algorithm to predict the voltage fluctuation trend in the next 50 to 200 milliseconds. Once it is predicted that the fluctuation amplitude will exceed ±5% of the rated voltage, or the rate of change will exceed 2V / ms, the pre-adjustment mechanism will be activated immediately. After the pre-conditioning is initiated, the central controller calculates the dynamic adjustment weight coefficient K for each power submodule based on its current state of charge (SOC), state of health (SOH), and real-time temperature (Ti). The formula for calculating this coefficient K means that the higher the SOC, the better the SOH, and the greater the difference between the real-time temperature and the reference temperature (i.e., the lower the temperature), the larger its dynamic adjustment weight coefficient K will be, and the more adjustment it will undertake in power regulation. Subsequently, the central controller sends pre-adjustment commands to each power submodule. Each power submodule determines its own power adjustment amount ΔPi based on its own dynamic adjustment weight coefficient K and the total adjustment power required by the system ΔPtotal (the total adjustment power ΔPtotal is calculated by the fluctuation amplitude of AC voltage and the energy storage characteristics of the main bus capacitor). Specifically, ΔPi = K × ΔPtotal. Based on this adjustment amount, the power submodule changes its output power by adjusting its internal bidirectional DC / DC converter module, thereby achieving the initial adjustment of the system power. Within 10 milliseconds after the pre-adjustment is executed, the system will collect the fluctuation of the main bus voltage again. If the fluctuation of the bus voltage still exceeds ±2% of the rated voltage at this time, the system will start the secondary correction mechanism. The secondary correction further optimizes the power distribution accuracy between each power sub-module by introducing virtual impedance compensation until the fluctuation of the main bus voltage is controlled within ±1% of the rated voltage, ensuring stable operation of the system.

[0045] The beneficial effects of the above technical solution are as follows: By adopting the solution provided in this embodiment, the problem of large fluctuations in bus voltage under the traditional passive adjustment method can be avoided by predicting the AC input voltage fluctuation trend in advance and starting the pre-adjustment mechanism; the introduction of the dynamic adjustment weight coefficient K realizes differentiated power allocation based on the state of the power submodule itself; the application of virtual impedance compensation in the secondary correction mechanism further improves the accuracy of power allocation and system response speed, ensuring the stability of the main bus voltage under complex fluctuation scenarios.

[0046] In one embodiment, dynamic balancing and capacity coordination management steps are also included: When the slave controller of any power submodule detects that its corresponding lithium battery pack has any of the following preset conditions, the central controller controls the bidirectional active balancing circuit in the power submodule to start, so as to achieve balancing of the lithium battery pack. The bidirectional active balancing circuit is configured to be connected to the lithium battery pack. The preset conditions include: single cell voltage difference > static balancing threshold; capacity difference during charging and discharging > dynamic balancing threshold; SOH difference > health state balancing threshold. The static balancing threshold is 20-50mV.

[0047] The working principle of the above technical solution is as follows: by introducing a dynamic balancing and capacity collaborative management mechanism, it achieves refined management and protection of lithium battery packs; its working principle is as follows: When the system is running, the slave controller of each power submodule continuously monitors various key parameters of the lithium battery pack it manages. These parameters include the voltage of each individual cell, capacity changes during charging and discharging, and the state of health (SOH) of the battery. Once the slave controller detects that the lithium battery pack has any of the following preset conditions, it will feed this status information back to the central controller: First, when the voltage difference between any two individual cells in the lithium battery pack exceeds the system's preset "static equalization threshold," this threshold is set to prevent the battery pack from being affected by excessive individual cell voltage differences when it is in a static state without charging or discharging. Second, during the dynamic process of charging or discharging the lithium battery pack, if the difference between the actual capacity and the expected capacity of the battery pack is detected, or the capacity difference between different cells / battery modules exceeds the preset "dynamic balance threshold", this threshold is for charging and discharging dynamic scenarios to ensure that the battery pack maintains relative consistency of capacity during energy conversion and avoids the risk of low charging and discharging efficiency or overcharging and over-discharging due to capacity imbalance. Third, when the State of Health (SOH) assessment results of individual cells or the whole lithium battery pack show that the difference exceeds the "State of Health Balance Threshold", SOH reflects the degree of degradation of the current performance of the battery relative to the nominal performance. The setting of this threshold is used to intervene in the system imbalance caused by the different aging of batteries in a timely manner. Upon receiving any of the aforementioned trigger signals from the controller, the central controller immediately issues a control command to activate the bidirectional active balancing circuit within the power submodule. This circuit is specifically configured to connect to the lithium battery pack, and its core function is to enable bidirectional energy transfer between individual cells within the pack. Through the activation of the bidirectional active balancing circuit, the system balances and regulates the lithium battery pack, transferring energy from cells with higher voltage, fuller capacity, or better health to cells with lower voltage, insufficient capacity, or poorer health, and vice versa (bidirectional energy flow based on actual needs), until the voltage and capacity of each cell in the pack reach equilibrium, or the differences in health status are controlled within a set threshold range. This restores the overall balance and optimal performance of the lithium battery pack, extends its lifespan, and ensures the safety and stability of the entire system. Among them, the bidirectional active balancing circuit is a battery balancing technology that can actively transfer energy from one or a group of batteries (cells / modules) to another or another group of batteries (cells / modules). Energy can flow bidirectionally, rather than just unidirectional discharge or replenishment, thereby achieving battery pack balancing more efficiently and improving energy utilization and balancing speed. The static balancing threshold refers to the limit value of the maximum voltage difference between individual cells when the battery pack is in a static, non-charging / discharging state. When the voltage difference exceeds this value, the static balancing mechanism is triggered. The dynamic balancing threshold refers to the limit value of the maximum capacity difference between individual cells or between the battery pack as a whole and the expected capacity during the dynamic operation of the battery pack during charging or discharging. The health status balancing threshold refers to the maximum limit value of the allowable difference in the state of health (SOH) between individual cells in the battery pack or the overall state of health (SOH) of the battery pack. When the SOH difference exceeds this value, the health status-based balancing mechanism is triggered.

[0048] The beneficial effects of the above technical solution are as follows: By adopting the solution provided in this embodiment, precise equalization control of lithium battery packs under different operating conditions can be achieved through dynamic equalization and capacity collaborative management mechanisms. On the one hand, by setting multi-dimensional preset conditions (voltage difference, capacity difference, SOH difference), the battery pack status can be comprehensively monitored to ensure timely triggering of equalization operations under different scenarios such as static, dynamic, and health degradation. On the other hand, the application of bidirectional active equalization circuit realizes efficient bidirectional energy transfer, which improves energy utilization and equalization speed compared to traditional unidirectional equalization methods, and avoids energy waste. At the same time, through the collaborative work of the central controller and the slave controller, independent and coordinated equalization management of lithium battery packs in each power submodule can be carried out, effectively solving the problem of poor battery pack consistency in split parallel systems, extending the overall service life of the battery pack, and improving the stability and reliability of system operation.

[0049] In one embodiment, the dynamic balancing and capacity coordination management steps further include: the central controller establishing a capacity contribution model for each power submodule, with the contribution coefficient being: , in, Represents the battery's state of charge. This indicates the battery's health status. This represents the temperature influence factor, and its value is determined based on the deviation of the real-time operating temperature of the power submodule from the preset optimal operating temperature range. When the real-time temperature is within the optimal operating temperature range, The value is 1, meaning that when the real-time temperature is higher or lower than the optimal operating temperature range, The value decreases linearly with increasing deviation, and the minimum value is not lower than 0.5. a, b, and c are the weighting coefficients of each parameter, and a + b + c = 1. Specifically, when the real-time temperature is higher than the upper limit of the optimal range... hour, ,in, The attenuation coefficient is... The real-time temperature; when the real-time temperature is below the lower limit of the optimal range. hour, ; The central controller dynamically adjusts the priority of each power submodule during the charging and discharging process based on its contribution coefficient. Specifically, during charging, power submodules with higher contribution coefficients receive higher priority in charging current allocation and are charged first; during discharging, power submodules with higher contribution coefficients also receive higher priority in discharging current allocation and are discharged first. At the same time, upper and lower thresholds for the contribution coefficient are set. When the contribution coefficient of a power submodule is lower than the lower threshold, its charging and discharging operation is suspended and an alarm is issued. When the contribution coefficient of a power submodule is higher than the upper threshold, its charging and discharging current rate is appropriately reduced.

[0050] The working principle of the above technical solution is as follows: By constructing a capacity contribution model for each power submodule, the charging and discharging priority and current rate are dynamically adjusted to achieve dynamic balance and capacity collaborative management of the system. The central controller calculates a contribution coefficient for each power submodule. This coefficient comprehensively considers the submodule's state of charge, health status, and temperature influence factor. These three factors are weighted and summed using weighting coefficients a, b, and c (a+b+c=1). The temperature influence factor is determined based on the deviation of the submodule's real-time operating temperature from the optimal operating temperature range: when the temperature is within the optimal range, the temperature influence factor is 1; when the temperature is higher or lower than the optimal range, the temperature influence factor is 1. Within the specified range, the temperature influence factor decreases linearly with increasing deviation, and the minimum value is not lower than 0.5. Based on the contribution coefficient, the central controller dynamically adjusts the charging and discharging priority of each power submodule: during charging and discharging, submodules with higher contribution coefficients receive higher current allocation priority and are charged and discharged first. At the same time, the system sets upper and lower thresholds for the contribution coefficient: when the contribution coefficient of a submodule is lower than the lower threshold, its charging and discharging operation is suspended and an alarm is issued; when the contribution coefficient of a submodule is higher than the upper threshold, its charging and discharging current rate is appropriately reduced, thereby achieving protection of the submodules and optimized management of the system.

[0051] The beneficial effects of the above technical solution are as follows: By adopting the solution provided in this embodiment, the actual performance of the power submodule can be comprehensively evaluated through multi-dimensional parameters, avoiding the problem of local overload or capacity waste caused by a single indicator (such as SOC) dominating the allocation logic; the introduction of temperature influence factor can effectively avoid the lifespan degradation of the submodule caused by forced charging and discharging under high temperature or low temperature environment, and the dynamic adjustable characteristics of the weight coefficient allow the system to flexibly adjust the optimization target according to different operating scenarios (such as emergency power supply, daily maintenance, and capacity expansion and debugging); in addition, the upper and lower limit threshold mechanism of contribution coefficient forms a dual protection, and the lithium battery pack capacity is maximized and the overall lifespan is extended through a refined collaborative management strategy.

[0052] In one embodiment, a distributed energy buffering coordination step is further included, specifically: the slave controller of each power submodule analyzes the main bus voltage fluctuation characteristics in real time through a sliding window algorithm, and predicts the main bus power demand mutation ΔP_pred within a preset time period based on the LSTM network; when ΔP_pred is predicted to exceed a preset threshold, a distributed buffering strategy is initiated; the distributed buffering strategy is as follows: the energy storage inductor L of each power submodule enters a transient energy storage mode, and absorbs or releases energy by changing the current change rate di / dt; each power submodule coordinates to undertake the buffering task according to its own SOC state in a preset proportion, the window length of the sliding window algorithm is 10-20ms; the prediction time range of the LSTM network is 50-200ms; The training and data processing of the LSTM network are as follows: First, historical voltage, current, and load change data of the main bus are collected to construct a time series dataset. The raw data is preprocessed, including denoising, normalization, and sliding window slicing, to generate sequence samples suitable for LSTM input. The sliding window width is set according to the dynamic response characteristics of the system to ensure that it contains sufficient historical fluctuation feature information. Then, the dataset is divided into training, validation, and test sets. A multi-layer LSTM neural network structure is used for model training. The input sequence is the power change sequence of the main bus over a past period, and the output is the predicted power demand value within a preset time (e.g., 50ms or 100ms). During training, mean squared error (MSE) is used as the loss function, combined with the Adam optimizer to adjust the network parameters, and early stopping is used to prevent overfitting. After training, the model is embedded in the slave controller of each power sub-module to predict the changing trend of the main bus power demand in real time. Each slave controller, based on the trained LSTM model and combined with the current real-time sampled main bus voltage data, continuously extracts local time-series features using the sliding window algorithm, and predicts the power fluctuation amplitude within the future short time window. When the prediction result exceeds the preset power change threshold, the distributed buffer strategy is triggered. At this time, all power sub-modules participating in the coordination enter the transient energy regulation state, and their energy storage inductors switch to transient energy storage mode. In this mode, each module adjusts the rate of change of inductor current (di / dt) to quickly absorb or release instantaneous excess or deficiency energy, thereby suppressing bus voltage fluctuations. The allocation of energy buffer tasks is dynamically allocated according to the current state of charge (SOC) of each sub-module according to a preset ratio. High SOC modules appropriately reduce absorption capacity and enhance release capacity, while low SOC modules do the opposite, in order to achieve balanced utilization and coordinated response of system-level energy storage resources.

[0053] The working principle of the above technical solution is as follows: A distributed energy buffering and coordination mechanism is used to mitigate power surges in the main bus. Specifically, the slave controllers of each power submodule continuously use a sliding window algorithm to monitor and analyze the main bus voltage signal in real time, extracting the dynamic characteristics of voltage fluctuations. Simultaneously, based on a Long Short-Term Memory (LSTM) neural network model, historical and real-time voltage and power data are used to predict the potential power demand surge value ΔP_pred for the main bus within a preset time period. When the system predicts that the absolute value of ΔP_pred exceeds the set power surge threshold, a distributed buffering strategy is immediately triggered. Under this strategy, the energy storage inductors L in all participating power submodules quickly switch to transient energy storage mode, actively adjusting the rate of change of the inductor current. The di / dt ratio enables rapid energy absorption (increasing di / dt when ΔP_pred is positive to allow the inductor to absorb excess energy) or release (decreasing di / dt when ΔP_pred is negative to allow the inductor to release stored energy), thereby dynamically compensating for instantaneous power imbalances in the main bus. To ensure the coordination and stability of the buffering process, each power submodule shares the total buffer power demand ΔP_pred according to its current state of charge (SOC) and a preset power allocation ratio (e.g., modules with higher SOC undertake a larger proportion of buffering tasks, while modules with lower SOC undertake a smaller proportion to avoid overcharging or over-discharging of a single module). Through distributed collaborative control among submodules, rapid and smooth suppression of power surges in the main bus is achieved, ensuring the overall stability of the system operation.

[0054] The beneficial effects of the above technical solution are as follows: by adopting the solution provided in this embodiment, the distributed energy buffer coordination mechanism can be used to proactively suppress power fluctuations in the main bus, thus avoiding the problem of excessive voltage fluctuations caused by the lag in the response of traditional centralized buffers.

[0055] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for managing independent lithium batteries of a split parallel DC power supply with bidirectional DC / DC conversion, the DC power supply being a movable structure, and the system architecture of the DC power supply comprising a central controller and a plurality of parallel power sub-modules, characterized in that: the system architecture is configured such that at least two independent lithium battery packs are connected in parallel to the main bus of the DC power supply through independent bidirectional DC / DC conversion modules supporting quick plugging, each lithium battery pack and its corresponding bidirectional DC / DC conversion module are integrated into a physically separable power sub-module to achieve quick replacement of failed lithium battery packs and flexible expansion of the system architecture; each power sub-module is equipped with a bidirectional active balancing circuit for single cell balancing control of the lithium battery pack; the method comprises the following steps: Step 1: The central controller identifies and completes the system architecture configuration for the newly added power sub-module based on the obtained state parameters of the existing power sub-modules and the main bus operating state data; wherein the state parameters of the existing power sub-modules are obtained from the corresponding lithium battery packs by the slave controller built-in the bidirectional DC / DC conversion module; Step 2: The central controller dynamically selects the control mode and generates the execution instruction of the charging and discharging power to control the charging and discharging of each power sub-module based on the number and state parameters of the power sub-modules; Step 3: Under the premise of ensuring uninterrupted power supply of the DC power supply, the central controller controls the lithium battery packs in the power sub-modules to perform online capacity verification or activation operation to address the battery degradation problem caused by long-term discharge or static state of the DC power supply. The electrical interface between the bidirectional DC / DC conversion module and the main bus adopts a magnetic or buckle type quick connector integrated with mechanical and electrical interlocking, supporting hot plugging operation with power on, and connection or disconnection without power off.

2. The method according to claim 1, wherein The newly added power sub-module is identified and the system architecture configuration is completed, which specifically includes:

3. The method according to claim 1, wherein The central controller receives and analyzes the identification information package of the newly added power sub-module, verifies the compatibility of the newly connected power sub-module in combination with the state parameters of the existing power sub-modules and the operating state data of the main bus, and if the verification is passed, updates the total number of power sub-modules, available capacity information and power distribution weight coefficient; wherein the identification information package is obtained from the corresponding lithium battery pack by the slave controller of the bidirectional DC / DC module in the new power sub-module, including the model, rated capacity, initial SOC, nominal voltage, maximum charging and discharging current and SOH of the lithium battery pack; the power distribution weight coefficient is calculated based on the preset expansion algorithm. In Step 2, the central controller dynamically selects the control mode and generates the execution instruction of the charging and discharging power to control the charging and discharging of each power sub-module based on the number and state parameters of the power sub-modules, including:

4. The method according to claim 1, wherein when the number of power sub-modules is less than or equal to a preset number threshold, a centralized control mode is adopted, the central controller generates the execution instruction of the charging and discharging power and sends it to the power sub-modules to control the charging and discharging of each power sub-module; ​ When the number of power sub-modules is greater than the preset number threshold, a distributed control mode is adopted, the central controller sends the total power demand for charging and discharging to each power sub-module, and the slave controller of the bidirectional DC / DC conversion module in the power sub-module autonomously negotiates power distribution through a distributed algorithm based on improved droop control and autonomously executes charging and discharging; wherein the distributed algorithm based on improved droop control is: a virtual impedance dynamic compensation mechanism and an adaptive adjustment strategy of voltage-power droop curve are introduced to realize the optimization of current sharing accuracy between power sub-modules.

5. The method of claim 1, wherein the method is characterized by: Step two also includes: When any power sub-module fails, the output end contactor of the bidirectional DC / DC conversion module in the power sub-module is disconnected, and the power sub-module is isolated from the main bus; The central controller re-calculates and controls the power sub-modules that have not failed to take over the current load demand according to the current load demand of the DC power supply and the available capacity information of the power sub-modules that have not failed, and ensures the stability of the main bus voltage and the continuous uninterrupted power supply of the DC power supply.

6. The method of claim 1, wherein the method is characterized by: In step three, the online capacity calculation or activation operation specifically includes: The central controller controls the bidirectional DC / DC conversion module in the target power sub-module to switch to the constant-current discharge mode or the charging and discharging cycle mode according to the preset period or instruction; wherein the constant-current discharge mode corresponds to capacity calculation, and the charging and discharging cycle mode corresponds to activation; The slave controller of the bidirectional DC / DC conversion module in the target power sub-module collects the state data of the corresponding lithium battery pack in real time and uploads it to the central controller; The central controller calculates the capacity of the lithium battery pack and evaluates the state of health based on the state data of the lithium battery pack, calculates the capacity retention rate of the lithium battery pack, and if the capacity retention rate is lower than the preset threshold, controls the target power sub-module to perform an activation operation including at least one complete charging and discharging cycle, and after the activation operation is completed, controls the target power sub-module to switch to the normal charging and discharging control mode.

7. The method of claim 1, wherein the method further comprises: When the AC input voltage fluctuates, the central controller triggers the pre-adjustment mechanism of the power sub-module output power based on the preset fluctuation threshold, and the pre-adjustment mechanism includes: The central controller collects the fluctuation amplitude and rate of the AC input voltage in real time, calculates the voltage fluctuation trend in the future 50-200ms through a sliding window algorithm, and when the predicted fluctuation amplitude exceeds ±5% of the rated voltage or the rate exceeds 2V / ms, the pre-adjustment is started; According to the current SOC, SOH and real-time temperature of each power submodule, a dynamic adjustment weight coefficient is calculated , , represents a state of charge, represents a state of health, is a reference temperature, is a real-time temperature of the power submodule, and α, β and γ are weight coefficients of the SOC, SOH and temperature deviation term, respectively, and α+β+γ=1. The central controller sends pre-adjustment instructions to each power submodule, and the power submodule dynamically adjusts the output power of the bidirectional DC / DC conversion module according to the dynamic adjustment weight coefficient, and the adjustment amount Satisfies: , The total adjustment power demand is calculated according to the AC fluctuation amplitude and the bus capacitor energy storage characteristics; Within 10ms after the pre-adjustment is executed, the main bus voltage fluctuation is collected, and if the voltage fluctuation still exceeds ±2% of the rated voltage, the secondary correction is started, and the power distribution accuracy is further optimized by introducing a virtual impedance compensation, until the bus voltage is stabilized within ±1% of the rated voltage.

8. The method of claim 1, wherein the method is characterized by: It also includes a dynamic balancing and capacity collaborative management step: When the slave controller of any power sub-module detects that the corresponding lithium battery pack meets any of the following preset conditions, the central controller controls the bidirectional active balancing circuit in the power sub-module to start and balance the lithium battery pack, and the bidirectional active balancing circuit is configured to connect the lithium battery pack; The preset conditions include: cell voltage difference > static balancing threshold; capacity difference > dynamic balancing threshold during charging and discharging; SOH difference > health state balancing threshold.

9. The method according to claim 8, wherein The dynamic balancing and capacity collaborative management step further includes: the central controller establishes a capacity contribution degree model for each power sub-module, and the contribution degree coefficient is: ， wherein, represents a battery state of charge, represents a battery state of health, represents a temperature influence factor, a value of which is determined according to a deviation degree of a real-time working temperature of the power sub-module from a preset optimal working temperature interval, when the real-time temperature is within the optimal working temperature interval, is 1, when the real-time temperature is higher or lower than the optimal working temperature interval, decreases linearly with the increase of the deviation degree, and a minimum value is not less than 0.5, a, b, and c are weight coefficients of respective parameters, and satisfy a+b+c=1; The central controller dynamically adjusts the priority of each power sub-module during charging and discharging according to the contribution degree coefficient, specifically: during charging, the power sub-module with a higher contribution degree coefficient obtains a higher charging current distribution priority and is preferentially charged; during discharging, the power sub-module with a higher contribution degree coefficient also obtains a higher discharging current distribution priority and is preferentially discharged; at the same time, upper and lower threshold values of the contribution degree coefficient are set, when the contribution degree coefficient of a certain power sub-module is lower than the lower threshold value, the charging and discharging operation of the power sub-module is suspended and an alarm prompt is issued, and when the contribution degree coefficient of a certain power sub-module is higher than the upper threshold value, the charging and discharging current rate of the power sub-module is appropriately reduced.

10. The method of claim 1, wherein the method is characterized by: The step further includes a distributed energy buffer coordination step, specifically: the slave controller of each power sub-module analyzes the main bus voltage fluctuation characteristics in real time through a sliding window algorithm, and predicts the main bus power demand mutation in a future preset time based on an LSTM network; when it is predicted that the threshold value is exceeded, a distributed buffer strategy is started; the distributed buffer strategy is: the energy storage inductance of each power sub-module enters a transient energy storage mode, and absorbs or releases energy by changing the current change rate; each power sub-module collaboratively undertakes the energy buffer task according to the preset proportion according to the SOC state of the power sub-module.

Images