Battery control method of BMS (Battery Management System)

Abstract

The invention relates to the technical field of BMS management systems, and discloses a battery control method for a BMS management system, and the method comprises the steps: carrying out the topological connection of a plurality of battery packs, and obtaining a hierarchical control system composed of a main control unit, a plurality of sub-control units, and a plurality of battery packs; performing recursive least square parameter identification and battery charging and discharging power equalization optimization on the voltage data, the current data and the temperature data in the hierarchical control system to obtain a battery pack power distribution control instruction; performing Lyapunov stability analysis and sequence planning on the battery pack power distribution control instruction and the load fluctuation data to obtain battery pack safety constraint control parameters; the charging and discharging power of the battery pack is dynamically distributed according to the battery pack safety constraint control parameters, battery pack cooperative control data are obtained and fed back to the main control unit, and the overall control performance and operation efficiency of the BMS are improved.

Description

Technical Field

[0001] This invention relates to the field of BMS management system technology, and in particular to a battery control method for a BMS management system. Background Technology

[0002] Traditional BMS systems typically employ a centralized control architecture, which suffers from heavy communication burdens and high control complexity when managing large-scale battery packs. Furthermore, due to the dynamic characteristics of battery packs and the influence of environmental factors, the system faces challenges such as significant uncertainties in the charging and discharging process, uneven temperature distribution, and substantial differences in state of charge.

[0003] Current BMS systems often employ simple threshold control strategies for battery pack balancing, making it difficult to balance system efficiency and safety. Especially in the collaborative management of large-scale battery packs, traditional control methods cannot effectively handle random disturbances such as load fluctuations and temperature changes, easily leading to safety hazards such as localized overcharging, over-discharging, and temperature runaway. Furthermore, most existing BMS control algorithms rely on accurate system models, but in practical applications, battery parameters are time-varying, resulting in significant system modeling errors. Summary of the Invention

[0004] This invention provides a battery control method for a BMS management system, which improves the overall control performance and operating efficiency of the BMS management system.

[0005] In a first aspect, the present invention provides a battery control method for a BMS management system, the battery control method for the BMS management system comprising:

[0006] Multiple battery packs are topologically connected to obtain a hierarchical control system consisting of a main control unit, multiple sub-control units, and multiple battery packs. Each of the multiple battery packs is connected to a sub-control unit in a one-to-one correspondence, and the multiple sub-control units are connected to the main control unit via a data bus.

[0007] Recursive least squares parameter identification and battery charging and discharging power equalization optimization are performed on the voltage, current and temperature data in the hierarchical control system to obtain the battery pack power distribution control command.

[0008] Lyapunov stability analysis and sequence planning were performed on the battery pack power distribution control commands and load fluctuation data to obtain the battery pack safety constraint control parameters.

[0009] The charging and discharging power of the battery pack is dynamically allocated according to the battery pack safety constraint control parameters to obtain battery pack collaborative control data, and the battery pack collaborative control data is fed back to the main control unit.

[0010] Secondly, the present invention provides a BMS management system, the BMS management system comprising:

[0011] The topology connection module is used to connect multiple battery packs in a topology to obtain a hierarchical control system consisting of a main control unit, multiple sub-control units, and multiple battery packs. Each of the multiple battery packs is connected to a sub-control unit in a one-to-one correspondence, and the multiple sub-control units are connected to the main control unit through a data bus.

[0012] The equalization optimization module is used to perform recursive least squares parameter identification and battery charging and discharging power equalization optimization on the voltage data, current data and temperature data in the hierarchical control system to obtain battery pack power distribution control instructions.

[0013] The sequence planning module is used to perform Lyapunov stability analysis and sequence planning on the battery pack power distribution control commands and load fluctuation data to obtain battery pack safety constraint control parameters.

[0014] The dynamic allocation module is used to dynamically allocate the charging and discharging power of the battery pack according to the battery pack safety constraint control parameters, obtain battery pack collaborative control data, and feed the battery pack collaborative control data back to the main control unit.

[0015] The technical solution provided by this invention establishes a hierarchical BMS control system, enabling collaborative work between the main control unit and sub-control units, reducing the system's communication burden, and improving the real-time performance and reliability of control. The recursive least squares method is used for parameter identification, combined with extended Kalman filtering for state estimation, significantly improving the identification accuracy of the battery equivalent circuit model and the accuracy of the state of charge calculation. A dynamic programming-based equilibrium optimization strategy ensures a reasonable allocation of charging and discharging power among multiple battery packs, effectively solving the imbalance problem between battery packs. The introduction of Lyapunov stability analysis and adaptive control algorithms enhances the system's ability to suppress random disturbances such as load fluctuations, improving the robustness of the control system. By establishing a chance-constrained optimization model, probabilistic constraints are transformed into deterministic constraints, enabling the system to operate safely in uncertain environments. A risk assessment-based dynamic scheduling strategy, combined with temperature equalization control, ensures the safe operation of the battery packs and extends battery life. A closed-loop optimization mechanism is adopted, improving the overall control performance and operating efficiency of the system through real-time parameter updates and strategy optimization. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the steps of the battery control method of the BMS management system in an embodiment of the present invention;

[0018] Figure 2 This is a schematic diagram of the structure of the BMS management system in an embodiment of the present invention. Detailed Implementation

[0019] This invention provides a battery control method for a BMS management system. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0020] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 One embodiment of the battery control method of the BMS management system in this invention includes:

[0021] Step S1: Connect multiple battery packs in a topology to obtain a hierarchical control system consisting of a main control unit, multiple sub-control units, and multiple battery packs. Each battery pack is connected to a sub-control unit in a one-to-one correspondence, and the multiple sub-control units are connected to the main control unit through a data bus.

[0022] It is understood that the executing entity of this invention can be a BMS management system, a terminal, or a server; no specific limitation is made here. This embodiment of the invention will be described using a server as an example.

[0023] Specifically, data acquisition channels are allocated to the voltage acquisition ports of multiple battery packs. By identifying and marking the voltage acquisition ports of the battery packs, the specific data acquisition channel number corresponding to each battery pack is determined, and a data acquisition port mapping table for n battery packs is generated. Based on the data acquisition port mapping table, multiple sub-control units are configured with channels to establish multiple independent data acquisition channels. Each sub-control unit programs and configures its internal data acquisition module according to the correspondence in the mapping table to achieve independent monitoring of each battery pack. After completing the channel configuration, data transmission protocols are set for these independent acquisition channels. In this process, the RS485 protocol is selected as the communication standard. RS485 has the advantages of strong anti-interference capability, long communication distance, and support for multi-point communication, making it suitable for the needs of a distributed battery management system. By setting the channel protocol, RS485 communication link configuration data is generated to ensure that each sub-control unit can stably acquire key data such as voltage, current, and temperature of the corresponding battery pack. The RS485 communication link configuration data is input into the communication modules of multiple sub-control units to establish a one-to-one correspondence between the sub-control unit and the battery pack, ensuring that a unique and stable communication path is established between the sub-control unit and the battery pack it manages. While establishing the connection between the sub-control units and the battery pack, the CAN communication interface of the sub-control units is configured as a master-slave node. The CAN bus is a high-speed, real-time communication bus suitable for multi-node control systems with a master-slave structure. Through master-slave node configuration, each sub-control unit is configured as a slave node, and the master control unit is configured as the master node, establishing a communication link between the sub-control units and the master control unit. After configuring the communication link, the data receiving module of the master control unit is initialized. The initialization process includes setting the communication parameters of the master control unit (such as baud rate and communication address) to ensure that it can correctly receive data from the sub-control units. Through this process, a data transmission channel is formed between the master and slave controllers, opening up the information flow between the master control unit and the sub-control units. Based on the master-slave controller data transmission channel, a tree-shaped control structure is constructed, forming a hierarchical controller topology. In this topology, the master control unit is at the highest level, responsible for global control and data aggregation, while multiple sub-control units are at the secondary level, each responsible for the status monitoring and data acquisition of the battery packs they manage. This tree structure effectively simplifies the management complexity of the system while enhancing the efficiency and stability of data transmission. The hierarchical controller topology is combined with the one-to-one correspondence between the sub-control units and the battery pack to form the final hierarchical control system.

[0024] Step S2: Perform recursive least squares parameter identification and battery charging and discharging power equalization optimization on the voltage, current and temperature data in the hierarchical control system to obtain the battery pack power distribution control command.

[0025] Specifically, the error covariance matrix of the voltage and current data collected by the hierarchical control system is initialized to establish an initial matrix for parameter identification. Based on the initial matrix, the open-circuit voltage source in the battery's equivalent circuit model is recursively calculated. The recursive least squares method, as the core algorithm for dynamic parameter identification, gradually optimizes the calculation results of the open-circuit voltage based on updates to historical and real-time data, obtaining accurate open-circuit voltage parameters. The open-circuit voltage parameters and the battery's internal resistance data are input into a second-order RC network model. The second-order RC network model is a battery equivalent circuit model that realistically reflects the dynamic characteristics inside the battery by modeling and calculating the ohmic internal resistance and polarization internal resistance. The ohmic internal resistance and polarization internal resistance obtained through the model calculation are used to solve for the polarization time constant. The polarization time constant, as an important parameter describing the battery's dynamic response, can intuitively reflect the time characteristics of the battery's internal chemical processes. In this process, the ohmic internal resistance reflects the transient impedance inside the battery, while the polarization internal resistance reflects the energy loss caused by the polarization phenomenon inside the battery; together, they constitute the complete parameters of the battery's equivalent circuit model. Extended Kalman filter (EKF) state equations are constructed based on voltage, current, and temperature data in a hierarchical control system. The EKF accurately estimates the dynamic state of the battery by combining the prediction and update processes of the nonlinear system state, including the dynamic estimation of the battery's state of charge (SOC). A state observation matrix is ​​established by real-time tracking and state prediction of voltage and current changes. The observed values ​​are combined with model calculations, and time updates are performed using a dynamic iterative calculation unit to obtain accurate SOC values. Based on the obtained SOC values, the changing trend of the battery's internal resistance is calculated using temperature data to determine the battery's health status. The battery's state of health (SOH) is a crucial indicator of battery life. By analyzing the impact of temperature on internal resistance, the degree of battery aging and its impact on charge / discharge capacity are effectively predicted. Combining the SOC and SOH values ​​generates complete battery state data reflecting the battery's operating status. Based on the battery equivalent circuit model parameters and battery state data, the charge / discharge power of the battery is optimized for balance. The core of charge / discharge power optimization is to balance the power distribution of each battery pack while ensuring safe operation of the battery pack, thereby extending the overall lifespan of the system. During the optimization process, objective functions are constructed, such as minimizing battery pack power imbalance and maximizing system efficiency. Optimization algorithms are then used to find the optimal power allocation scheme for each battery pack. By decoding and transforming the optimization results, power allocation control commands for the battery packs are generated and fed back to the main control unit, achieving global coordinated control of the system.

[0026] The state of charge (SOC) values ​​in the battery state data are used to quantify state differences. By analyzing the SOC values, a SOC difference matrix is ​​constructed between adjacent battery packs, reflecting the SOC difference of each battery pack relative to others. Based on this, weighting coefficients are configured on the SOC difference matrix to generate an objective function for system balancing optimization. By introducing weighting coefficients, the optimization objective is adjusted according to the specific characteristics and operational requirements of the battery packs, such as prioritizing reducing the SOC difference of certain key battery packs, thereby achieving system balancing more efficiently. Based on the internal resistance parameters in the battery equivalent circuit model, dynamic power limits for charging and discharging are calculated to obtain the threshold range for charging and discharging power. The internal resistance characteristics of the battery are analyzed in conjunction with the SOC, as the internal resistance changes with SOC and temperature; therefore, the charging and discharging power limits also need to be dynamically adjusted. Limit calculations ensure that the battery does not exceed its safe range during power distribution and provide a physical basis for establishing constraints. Based on the charging and discharging power threshold range, constraint equations are constructed to obtain mathematical models for voltage constraints, current constraints, and power constraints. These constraints describe the operating boundaries of the battery pack, such as limiting the charging voltage to the maximum allowable voltage and the discharging current to the maximum allowable current. The construction of the constraint equations ensures that the power allocation optimization process always conforms to the safe operation specifications of the battery pack. Based on the objective function of the equalization optimization and the above mathematical model, a state transition cost matrix is ​​constructed to quantify the transition costs between various possible states in the power allocation path. For example, at different time steps, the power adjustment required to achieve state-of-charge balance and the potential losses. This cost matrix transforms the dynamic power allocation problem into an optimal path search problem. Optimal path search calculations are performed on the cost matrix using optimization algorithms (such as dynamic programming or linear programming) to obtain power allocation data at each time step, reflecting the power adjustment required by each battery pack at different time points. Based on the power allocation data, charging and discharging power adjustment coefficients between adjacent battery packs are generated, indicating how each battery pack should adjust its power output at the current time point to achieve system-wide equalization optimization. Dynamic power commands are time-series integrated to ensure that the dynamic commands of each battery pack are coordinated in time, thereby avoiding unnecessary power fluctuations or system instability caused by command timing issues. The power allocation control commands for the battery pack are then obtained. This instruction can dynamically adapt to changes in the battery pack's operating status and fluctuations in load demand, thereby achieving balanced optimization of battery charging and discharging power on a global scale.

[0027] Step S3: Perform Lyapunov stability analysis and sequence planning on the battery pack power distribution control command and load fluctuation data to obtain the battery pack safety constraint control parameters;

[0028] Specifically, amplitude features are extracted from load fluctuation data to obtain a load fluctuation amplitude sequence, reflecting the main characteristics of the system under external disturbances. Based on the load fluctuation amplitude sequence, probabilistic statistical operations are performed to generate a probabilistic model of load disturbances. The probabilistic model can describe the statistical laws and distribution characteristics of load fluctuations. This probabilistic model is combined with the power distribution control commands of the battery pack, and through state-space modeling, a stochastic disturbance state equation containing load disturbance terms is obtained. To simplify the analysis of the stochastic disturbance state equation, it is linearized, transforming the complex nonlinear stochastic model into a linear state-space model containing disturbance terms. The linearized model has higher computability while retaining an accurate description of the disturbance terms. On this basis, a positive definite quadratic Lyapunov function is constructed, which serves as the system energy functional, quantifying the energy change of the system under disturbance conditions. By taking the time derivative of the Lyapunov function, the system stability discriminant is derived. The system stability discriminant is transformed into a nonnegative term matrix by performing full square term collocation and coefficient matching. Eigenvalue decomposition is performed on the non-negative term matrix to extract parameters and boundaries that play a key role in system stability, yielding system stability constraints. Based on these constraints, adaptive control parameters are recursively calculated to obtain a feedback control gain sequence. This sequence is the core of adaptive control; these gain coefficients are used to adjust the control strategy in real time to cope with disturbance changes. A parameter adjustment law is constructed using the feedback control gain sequence, forming an adaptive parameter update rule. This rule defines how to dynamically adjust control parameters based on the current system state and disturbance conditions to ensure the system always operates within a stable and safe range. Inputting this rule into the feedback controller yields a dynamic compensation control quantity. This quantity is used to adjust the system's current operating state in real time to offset the adverse effects of disturbances. Through this dynamic compensation process, online parameter adjustment of the battery pack's operating state is achieved, generating adaptive feedback control data. Probability distribution analysis and sequence planning are performed on this data, comprehensively considering the randomness of disturbances and the dynamic characteristics of the system response, ultimately generating the battery pack's safety constraint control parameters.

[0029] Based on adaptive feedback control data, historical sample statistics of load demand data are performed to obtain a load demand probability distribution matrix, reflecting the distribution of load at different time points. A load probability density model is constructed by fitting a density function to the load demand probability distribution matrix, which describes the stochastic characteristics and variation patterns of load demand. Simultaneously, considering the impact of environmental conditions on battery operation, distribution features of temperature change data are extracted to generate a temperature fluctuation probability matrix. A temperature change probability model is obtained by calculating the cumulative distribution of the temperature fluctuation probability matrix. Combining the load probability density model and the temperature change probability model, an uncertainty probability space is constructed. This probability space represents the combined impact of load and temperature on battery pack operation in the form of a joint distribution. Confidence level analysis is performed on the uncertainty probability space to determine confidence level boundary values, defining the probabilistic requirements that the battery pack must meet under safe operating conditions. Based on the confidence level boundary values, the voltage and temperature constraints in battery operation are transformed from the probabilistic domain into deterministic constraints. The transformation process is completed using the inverse mapping method in probability theory, mapping the region in the uncertainty probability space that satisfies a given confidence level to specific constraints, ensuring that the generated constraints meet both safety requirements and operational feasibility. After obtaining the deterministic constraints, a risk optimization objective function is constructed. This objective function aims to minimize risk and maximize system benefits, comprehensively considering the performance and safety of the battery pack, and generating a complete sequence programming mathematical model. Based on the sequence programming mathematical model, KKT conditions (first-order necessity conditions) are constructed to ensure that the optimization problem satisfies the constraints while reaching the optimal solution. By constructing the first-order optimality condition matrix, the optimization problem is transformed into a solvable mathematical problem. The optimization problem is numerically solved using the gradient iteration calculation method to obtain the optimization result. The introduction of gradient iteration calculation can improve the efficiency and accuracy of the solution process, ensuring that the generated numerical solution conforms to the complex constraints of the system. The numerical solution of the optimization problem is input into the feedback correction unit for closed-loop compensation. The role of the feedback correction unit is to dynamically adjust the optimization result according to the deviation between the current state and the expected state of the system, generating compensation control parameters. The compensation control parameters enhance the system's anti-disturbance capability through closed-loop compensation, ensuring the operational stability of the battery pack under complex operating conditions. To improve safety, a safety margin is set for the compensation control parameters, that is, a certain safety redundancy is introduced into the control parameters to ensure that the operating requirements can still be met under extreme conditions. The safety constraint control parameters of the battery pack are obtained.

[0030] Step S4: Dynamically allocate the charging and discharging power of the battery pack according to the battery pack safety constraint control parameters to obtain battery pack collaborative control data, and feed the battery pack collaborative control data back to the main control unit.

[0031] Specifically, risk assessment calculations are performed on the current system operating state based on the battery pack safety constraint control parameters to obtain real-time operating risk indicators. The risk assessment calculation utilizes the dynamic correlation between the safety constraint control parameters and real-time system data. By constructing a risk assessment matrix, the safety risk level of the current battery pack under different operating conditions is quantified, and system risk level data is generated, reflecting potential problems the battery pack may face under complex conditions, such as overcharging, over-discharging, current overload, or temperature exceeding limits. Based on the system risk level data, power limit adjustments are made to dynamically adjust the battery pack's power output range. Through detailed analysis of different risk levels and optimization of adjustment strategies, dynamic power limit parameters are generated. To improve computational efficiency and enhance the practicality of control, the dynamic power limit parameters are piecewise linearized, transforming complex nonlinear power constraints into simple linear power control interval data. Differential weight allocation is applied to the state of charge (SOC) of each battery pack based on the power control interval data. Weighting factors balance the SOC differences between battery packs to reduce accelerated battery aging or power output imbalance caused by uneven SOC. The weighting result determines the generation of charge / discharge power adjustment coefficients. These coefficients are used to dynamically correct the power allocation strategy of each battery pack, resulting in an optimized dynamic power allocation strategy. Temperature equalization control is applied to the dynamic power allocation strategy to address temperature unevenness issues during battery pack operation. A temperature control target value is calculated through dynamic matching of temperature data and the power allocation strategy. Based on the temperature control target value, closed-loop control calculations generate temperature adjustment commands to adjust the temperature distribution of each battery pack in real time, ensuring the battery pack operates within a safe temperature range. According to the generated temperature adjustment commands, temperature compensation is applied to the charge / discharge power of each battery pack, resulting in a temperature compensation coefficient. This temperature compensation coefficient is used to further optimize the power adjustment strategy, balancing both temperature equalization and power output requirements. After combining the temperature compensation coefficients, real-time power control commands are generated to ensure precise control of the system under complex dynamic conditions. Timing coordination of the real-time power control commands ensures optimal collaborative operation between battery packs. Timing coordination effectively avoids control delays or power fluctuations caused by command synchronization issues, generating battery pack collaborative control data. By feeding back the battery pack's collaborative control data to the main control unit, the system's closed-loop management is completed, enabling the BMS to adjust its strategies in real time during operation, ensuring the safe, efficient, and long-life operation of the battery pack.

[0032] In this embodiment of the invention, a hierarchical BMS control system is established, enabling collaborative work between the main control unit and sub-control units, reducing the system's communication burden, and improving the real-time performance and reliability of the control. Parameter identification is performed using recursive least squares, combined with extended Kalman filtering for state estimation, significantly improving the identification accuracy of the battery equivalent circuit model and the accuracy of the state of charge calculation. A dynamic programming-based equilibrium optimization strategy ensures a reasonable allocation of charging and discharging power among multiple battery packs, effectively solving the imbalance problem between battery packs. The introduction of Lyapunov stability analysis and adaptive control algorithms enhances the system's ability to suppress random disturbances such as load fluctuations, improving the robustness of the control system. By establishing a chance-constrained optimization model, probabilistic constraints are transformed into deterministic constraints, enabling the system to operate safely in uncertain environments. A risk assessment-based dynamic scheduling strategy, combined with temperature equalization control, ensures the safe operation of the battery packs and extends battery life. A closed-loop optimization mechanism is adopted, improving the overall control performance and operating efficiency of the system through real-time parameter updates and strategy optimization.

[0033] In one specific embodiment, the process of performing step S1 may specifically include the following steps:

[0034] Data acquisition channels are allocated to the voltage acquisition ports of multiple battery packs to obtain a data acquisition port mapping table for n battery packs.

[0035] Based on the data acquisition port mapping table, multiple sub-control units are configured with channels to obtain n independent data acquisition channels. Then, the data transmission protocol is set for the n independent data acquisition channels to obtain RS485 communication link configuration data.

[0036] Input the RS485 communication link configuration data into the communication module of multiple sub-control units to obtain the one-to-one correspondence between the sub-control units and the battery pack;

[0037] Configure the CAN communication interfaces of multiple sub-control units as master and slave nodes to obtain the communication link between the sub-control units and the master control unit. Then, initialize the data receiving module of the master control unit according to the communication link to obtain the master-slave controller data transmission channel.

[0038] A tree-shaped control structure is constructed based on the master-slave controller data transmission channel to obtain a hierarchical controller topology. The hierarchical controller topology is then combined with the one-to-one correspondence connection relationship to construct a layered control system.

[0039] Specifically, for multiple battery packs in the system, assuming the total number of battery packs is n, the voltage acquisition port of each battery pack is represented as P. i,jWhere i represents the i-th sub-control unit (from 1 to m, m≤n), and j represents the j-th battery pack acquisition port within the sub-control unit (from 1 to k≥1, i×k=n). The mapping relationship between the acquisition ports is defined as M(i,j)=P i,j This yields a complete data acquisition port mapping table, which shows the correspondence between each sub-control unit and the acquisition ports of its subordinate battery packs. For example, assuming there are n = 12 battery packs and m = 3 sub-control units, with each sub-control unit managing k = 4 battery packs, the data acquisition port mapping table would be:

[0040]

[0041] Based on the mapping table above, data acquisition channels are configured for each sub-control unit. The acquisition channel configuration for each sub-control unit is represented as C. i ={P i,1 ,P i,2 ,…,P i,k}, and set T by configuring the data transmission protocol. i The communication link is typically established using the RS485 protocol. The RS485 protocol, with its advantages of supporting multi-point communication and strong anti-interference capabilities, is suitable for the needs of battery pack management. For each sub-control unit i, the communication link configuration data is represented as follows:

[0042] T i ={B,A,R,D};

[0043] Where B is the baud rate, used to define the data transmission rate (e.g., 9600bps); A is the address, used to uniquely identify the communication address of the sub-control unit; R is the data format, such as 8-N-1 (8 data bits, no parity bit, 1 stop bit); D is the data frame structure, used to define the length and format of the data frame. This is achieved by configuring n independent data acquisition channels C1, C2, ..., C... m RS485 communication link settings are configured to obtain a complete communication configuration table. This configuration data is then input into the communication modules of each sub-control unit, establishing a one-to-one connection between the sub-control unit and the battery pack. After connecting the sub-control units to the battery pack, the CAN communication interface of the sub-control units is configured as a master-slave node to establish a communication link between the sub-control units and the master control unit. Assume the master control unit is node N0, and the sub-control units are N1, N2, ..., N... m The communication relationship between master and slave nodes is represented as follows:

[0044] L={(N0,N1),(N0,N2),…,(N0,N m )};

[0045] Where L represents the set of communication links. Each communication link requires configuration of a node ID and communication rate, typically set to the CAN standard rate of 500kbps. The data receiving module of the main control unit also needs to be initialized according to the structure of the communication links to ensure correct reception of data transmitted from the slave nodes. After completing the master-slave communication link configuration, a tree-structured control structure is constructed based on the data transmission channels of the master and slave controllers. The main control unit is the root, each sub-control unit is a first-level node, and the battery packs managed by each sub-control unit are second-level nodes. The tree structure is represented as follows:

[0046] T = (V, E);

[0047] Where V is the set of nodes, including the main control unit, sub-control units, and battery packs, and E is the set of edges, representing the communication links from the main control unit to the sub-control units, and from the sub-control units to the battery packs. In the above example, the tree topology is represented by a graph structure as follows:

[0048] V = {N0, N1, N2, N3, P} 1,1 ,P 1,2 ,…,P 3,4};

[0049] E={(N0,N1),(N0,N2),(N0,N3),(N1,P 1,1 ),…,(N3,P 3,4 )};

[0050] By combining the tree-like control structure with the one-to-one correspondence M(i,j) between the sub-control units and battery packs, a complete hierarchical control system is obtained. This system is highly modular and scalable. The main control unit is responsible for global management and coordination, the sub-control units are responsible for local data acquisition and control, and each battery pack achieves precise status monitoring and power management through the sub-control units.

[0051] In one specific embodiment, the process of performing step S2 may specifically include the following steps:

[0052] The error covariance matrix of the voltage and current data collected by the hierarchical control system is initialized to obtain the parameter identification initial matrix. Based on the parameter identification initial matrix, the open-circuit voltage source of the battery equivalent circuit is recursively calculated to obtain the open-circuit voltage parameters.

[0053] The open-circuit voltage parameters and internal resistance data are input into the second-order RC network model to obtain the equivalent circuit ohmic internal resistance and polarization internal resistance. Based on the equivalent circuit ohmic internal resistance and polarization internal resistance, the polarization time constant is calculated to obtain the battery equivalent circuit model parameters.

[0054] Extended Kalman filter state equations are constructed from voltage, current and temperature data in the hierarchical control system to obtain the state observation matrix. The state observation matrix is ​​then input into the dynamic iterative calculation unit for time update to obtain the battery state of charge value.

[0055] The internal resistance change trend is calculated based on the battery equivalent circuit model parameters and temperature data to obtain the battery health status value. The battery state of charge value and battery health status value are then combined to obtain battery status data.

[0056] Based on the battery equivalent circuit model parameters and battery state data, the battery charging and discharging power is balanced and optimized to obtain the battery pack power distribution control command.

[0057] Specifically, the hierarchical control system collects real-time voltage data V(t) and current data I(t) of the battery pack from multiple sub-control units. By initializing these data using an error covariance matrix, a parameter identification initial matrix P0 describing the statistical characteristics of the data is obtained. Let the acquisition errors of voltage and current be... and The initial covariance matrix is ​​then expressed as:

[0058]

[0059] in, The variance represents the voltage measurement error. This represents the variance of the current measurement error. Matrix P0 is used to initialize the weight update formula for the recursive least squares algorithm. The initial matrix P0 is identified based on the parameters, and the open-circuit voltage source E of the battery equivalent circuit is calculated using the recursive least squares method. oc Perform recursive calculations. The formula for identifying the open-circuit voltage parameter is expressed as:

[0060] E oc (t)=V(t)+R int ·I(t);

[0061] Among them, E oc V(t) is the open-circuit voltage at time t, V(t) is the battery terminal voltage, I(t) is the battery current, and R is the open-circuit voltage at time t. int It is the internal resistance of the battery. By measuring E... oc The recursive calculation of (t) dynamically updates the open-circuit voltage value of the battery. The open-circuit voltage parameter E... oc With internal resistance data R int The input is fed into a second-order RC network model to extract the dynamic characteristics of the battery's equivalent circuit. The second-order RC network model includes an ohmic internal resistance R0 and a polarization internal resistance R0. p and a polarization capacitor C p Its equivalent circuit description is as follows:

[0062] V(t)=E oc (t)-R0·I(t)-R p ·I p (t);

[0063] Among them I p (t) is the polarization current, defined as:

[0064]

[0065] Based on this model, R0 and R are solved using a fitting method. p The polarization time constant τ is calculated based on the ohmic resistance and polarization resistance. p Its formula is:

[0066] τ p =R p ·C p ;

[0067] Among them, C p These are polarization capacitors, determined experimentally or through identification algorithms. The final values ​​of R0 and R are obtained. p and τ p The equivalent circuit model parameters of the battery are determined. An extended Kalman filter state equation is constructed using the voltage data V(t), current data I(t), and temperature data T(t) in the hierarchical control system to estimate the battery's state of charge (SOC). The extended Kalman filter state equation is:

[0068]

[0069] Where Δt is the sampling time interval, and Q is the nominal capacity of the battery. The observation equation is:

[0070] V(t)=E oc (SOC)-R0·I(t)-R p ·I p (t);

[0071] By combining the state equation and the observation equation using an extended Kalman filter, a state-observation matrix H is generated. This matrix is ​​then input into a dynamic iterative calculation unit for time updates, ultimately obtaining the real-time SOC value of the battery. Based on the obtained SOC, the internal resistance R is calculated according to the battery equivalent circuit model parameters and temperature data T(t). int The trend of change is expressed by the formula:

[0072] R int (T)=R int,ref ·(1+α·(TT ref ));

[0073] Among them, R int,refIt is the internal resistance at the reference temperature, T ref This is the reference temperature, and α is the temperature coefficient. Through analysis of R... int The changing trend of (T) is used to assess the battery's state of health (SOH), for example, by measuring the rate of increase in internal resistance to indicate the degree of battery aging. Combining SOC and SOH forms complete battery state data for charge / discharge power optimization. The goal of power optimization is to balance the SOC of each battery pack, avoiding overcharging and over-discharging, while simultaneously delaying battery aging. The optimization model is defined as follows:

[0074]

[0075] Among them, SOC i It is the SOC of the i-th battery pack; SOC avg It is the average SOC of all battery packs; w i These are weighting factors. The optimization constraints include voltage constraints, current constraints, and power constraints, which are as follows:

[0076] V min ≤V i ≤V max ,I min ≤I i ≤I max ,P min ≤P i ≤P max ;

[0077] By solving the above optimization problem, the power allocation strategy of the battery pack is obtained, and the final power allocation control command is generated to realize the collaborative optimization control of the battery pack.

[0078] In one specific embodiment, the process of performing equalization optimization of battery charging and discharging power based on battery equivalent circuit model parameters and battery state data to obtain battery pack power distribution control commands can specifically include the following steps:

[0079] The state of charge (SCC) values ​​in the battery state data are quantified to obtain the SCC difference matrix between adjacent battery packs. The weighting coefficients of the SCC difference matrix between adjacent battery packs are then configured to obtain the objective function for balanced optimization.

[0080] Dynamic power limits are calculated based on the internal resistance parameters in the battery equivalent circuit model to obtain the charging and discharging power threshold range. Constraint equations are then constructed based on the charging and discharging power threshold range to obtain mathematical models for voltage constraints, current constraints, and power constraints.

[0081] Based on the objective function and mathematical model of equilibrium optimization, a state transition cost matrix is ​​constructed, and the optimal path search is performed on the state transition cost matrix to obtain power allocation data for each time step.

[0082] Based on the power allocation data, the charging and discharging power of adjacent battery packs is adjusted to generate adjustment coefficients, thereby obtaining the dynamic power command for each battery pack. The dynamic power commands of each battery pack are then integrated in a timing sequence to obtain the battery pack power allocation control command.

[0083] Specifically, by collecting the state of charge (SOC) values ​​of the battery packs, the SOC of each battery pack is defined as SOC. i where i = 1, 2, ..., n represents the state value of the i-th battery pack in the system. To quantify the state differences between adjacent battery packs, a state-of-charge difference matrix ΔSOC is constructed, whose elements ΔSOC... i,j The state difference between the i-th and j-th battery packs is represented by the following formula:

[0084] ΔSOC i,j =SOC i -SOC j ;

[0085] This matrix represents the SOC difference among all battery packs. Weighting coefficients are configured on the difference matrix, and the weighting coefficients are defined as w. i,j , representing the weight of the state difference between the i-th and j-th battery packs on the optimization objective. Combining the weighting coefficients, the objective function for equilibrium optimization is expressed as:

[0086]

[0087] Here, J represents the total optimization cost of the objective function. Minimizing J achieves a balanced distribution of State of Charge (SOC) among the battery packs. The weighting coefficients are adjusted based on the importance of the battery packs, temperature distribution, or aging state. Based on the constructed balanced optimization objective function, the internal resistance R in the battery equivalent circuit model parameters is also considered. int Perform dynamic power limit calculation. Assume the maximum charging power of the i-th battery pack is P. ch,i The maximum discharge power is P dis,i Its threshold range is expressed as:

[0088] P min,i ≤P i ≤P max,i ;

[0089] Among them, P min,i =-P dis,i P represents the maximum discharge power. max,i =P ch,i P represents the maximum charging power. i This is the actual power of the i-th battery pack. Combined with the internal resistance parameter R... int The relationship between battery power and voltage and current is expressed as follows:

[0090] P i =V i ·I i =(E oc,i -I i ·R int,i )·I i ;

[0091] Using this formula, the power limit is converted into voltage and current constraints, resulting in the following mathematical model:

[0092] V min,i ≤E oc,i -I i ·R int,i ≤V max,i ,I min,i ≤I i ≤I max,i ;

[0093] These constraints ensure that the battery pack operates within safe charging and discharging ranges. A state transition cost matrix C(t) is constructed based on the objective function and mathematical model, where C... i,j (t) represents the cost required to transition from state i to state j at time t. Assume the state transition and power difference P are constants. i,j If they are proportional, then the cost matrix can be expressed as:

[0094] C i,j (t)=α·(ΔP i,j ) 2

[0095] Wherein, ΔP i,j =P i (t)-P j (t) represents the power difference between adjacent battery packs, and α is the cost scaling factor used to adjust the sensitivity of the optimization. By performing optimal path search calculations on the state transition cost matrix, for example using a dynamic programming algorithm, the power allocation data P for each time step is obtained. i (t). Based on the power allocation data, the charge / discharge power adjustment coefficient β is generated. i Its definition is:

[0096]

[0097] By adjusting the coefficient β i The power allocation strategy is modified to generate dynamic power commands for each battery pack.

[0098]

[0099] The dynamic power commands of each battery pack are integrated in timing to generate the battery pack power distribution control command P.ctrl (t), which is defined as:

[0100]

[0101] This control command will be fed back to the main control unit to adjust the power distribution of each battery pack in real time, ensuring that the system achieves balanced optimization of SOC while meeting the constraints.

[0102] In one specific embodiment, the process of performing step S3 may specifically include the following steps:

[0103] Amplitude features are extracted from the load fluctuation data to obtain the load fluctuation amplitude sequence. Probability statistics are then performed on the load fluctuation amplitude sequence to obtain the load disturbance probability model.

[0104] The load disturbance probability model and the battery pack power distribution control command are modeled in state space to obtain the random disturbance state equation. The random disturbance state equation is then linearized to obtain a linear state space model containing the disturbance term.

[0105] Based on the linear state-space model containing the disturbance term, a positive definite quadratic Lyapunov function is constructed to obtain the system energy functional. The time derivative of the system energy functional is then obtained to obtain the system stability criterion.

[0106] The system stability criterion is subjected to full square term configuration and coefficient matching to obtain a non-negative term matrix. Then, the non-negative term matrix is ​​subjected to eigenvalue decomposition to obtain the system stability constraints.

[0107] The adaptive control parameters are recursively calculated based on the system stability constraints to obtain the feedback control gain sequence. Then, the parameter adjustment law is constructed based on the feedback control gain sequence to obtain the adaptive parameter update rule.

[0108] The adaptive parameter update rule is input into the feedback controller to obtain the dynamic compensation control quantity, and the parameters are adjusted online according to the dynamic compensation control quantity to obtain the adaptive feedback control data.

[0109] Probability distribution analysis and sequence planning are performed on the adaptive feedback control data to obtain the battery pack safety constraint control parameters.

[0110] Specifically, amplitude features are extracted from load fluctuation data. Assuming the time series of load fluctuation is L(t), its amplitude is defined as:

[0111]

[0112] Among them, A L (t) is the fluctuation amplitude at time t. It is the average value of the load. The amplitude characteristic sequence {A} is obtained through calculation. L (t1),A L (t2),…,A L (t n The dynamic characteristics of load fluctuations are quantified. Based on the fluctuation amplitude sequence A L (t), perform probability and statistical calculations, assuming that the fluctuation amplitude follows a certain distribution p(A) L The probability density function is constructed using statistical frequencies. For example, the histogram estimation method is used to fit the probability distribution model P of load fluctuations. L (A). If the load fluctuation approximately follows a normal distribution, its probability density function is expressed as:

[0113]

[0114] Where μ is the mean of the fluctuation amplitude, σ 2 It is the variance of the fluctuation amplitude. By fitting a probability density model P... L (A) provides a complete description of the statistical characteristics of load fluctuations. The load disturbance probability model P... L (A) Power distribution control command P of the battery pack i Combining (t), we generate stochastic perturbation state equations through state-space modeling. Assuming the dynamic state of the battery pack is x(t), the input is u(t), and the perturbation is w(t), the state-space equations are expressed as:

[0115]

[0116] Where x(t) is the system state vector, such as state of charge (SOC), temperature, etc., u(t) is the power distribution control command, and w(t) is the random disturbance term, following the probability distribution P. L (A), where A is the state matrix, B is the control input matrix, and G is the disturbance input matrix. To simplify the calculation, the above random disturbance state equation is linearized. Assuming the system operates near a certain equilibrium point, δx(t) = x(t) - x0 and δu(t) = u(t) - u0 are defined, resulting in the linearized state-space model:

[0117]

[0118] After obtaining the linearized model, based on the dynamic characteristics of the system, a positive definite quadratic Lyapunov function is constructed, defined as:

[0119] V(x)=x T Px;

[0120] Where V(x) is the Lyapunov function, representing the energy functional of the system, and P is a symmetric positive definite matrix used to measure the energy of the system state. By taking the time derivative of V(x), the stability criterion of the system is obtained:

[0121]

[0122] If proof If the condition holds true for all x ≠ 0, then the system is asymptotically stable. To verify the system's stability, the quadratic form in the discriminant is collocated with perfect square terms and coefficients matched to obtain the nonnegative term matrix Q, which is defined as:

[0123] Q = A T P+PA+PGG T P;

[0124] If Q is positive definite, then the system satisfies the Lyapunov stability condition. The positive definiteness of Q is verified by eigenvalue decomposition, and the system stability constraints are extracted. After obtaining the stability constraints, the control parameters are optimized, and the feedback control gain K is defined as:

[0125] K = -R -1 B T P;

[0126] Where R is the control weight matrix. The feedback control gain sequence K(t) is obtained by recursively calculating the feedback control gain, and a parameter adjustment law is constructed based on the gain sequence, for example:

[0127] u(t) = K(t)x(t);

[0128] This regulation law adjusts the control input in real time to ensure system stability. Based on this, an adaptive parameter update rule is constructed using the feedback control gain, enabling the control parameters to be dynamically updated as the system state changes. After inputting the update rule into the feedback controller, a dynamic compensation control quantity Δu(t) is generated to compensate for deviations caused by disturbances. The control quantity after online adjustment is defined as:

[0129] u′(t)=u(t)+Δu(t);

[0130] By combining dynamic compensation control variables, adaptive feedback control data is obtained through online parameter adjustment, which is used to further optimize system performance. Probability distribution analysis is performed on the feedback control data, and the final battery pack safety constraint control parameters are generated through sequence programming.

[0131] In one specific embodiment, the process of performing probability distribution analysis and sequence planning on the adaptive feedback control data to obtain the battery pack safety constraint control parameters may specifically include the following steps:

[0132] Based on the adaptive feedback control data, historical sample statistics of load demand data are performed to obtain the load demand probability distribution matrix. Then, the density function is fitted to the load demand probability distribution matrix to obtain the load probability density model.

[0133] Based on the load probability density model, the distribution characteristics of temperature change data are extracted to obtain the temperature fluctuation probability matrix. The cumulative distribution of the temperature fluctuation probability matrix is ​​then calculated to obtain the temperature change probability model.

[0134] The probability space of the load probability density model and the temperature change probability model is constructed to obtain the uncertainty probability space. Then, the confidence level boundary value is obtained by performing confidence analysis on the uncertainty probability space.

[0135] Based on the confidence level boundary values, the voltage and temperature constraints are probabilistically transformed to obtain deterministic constraints. Based on the deterministic constraints, a risk optimization objective function is constructed to obtain a sequence programming mathematical model.

[0136] The KKT conditions are constructed on the sequential programming mathematical model to obtain the first-order optimality condition matrix. Then, gradient iteration is performed based on the first-order optimality condition matrix to obtain the numerical solution of the optimization problem.

[0137] The numerical solution of the optimization problem is input into the feedback correction unit for closed-loop compensation to obtain the compensation control parameters. The safety margin of the compensation control parameters is then set to obtain the battery pack safety constraint control parameters.

[0138] Specifically, based on adaptive feedback control data, historical sample statistics are performed on load demand data to obtain the load demand probability distribution matrix. Let the load demand time series be L(t), and the sample set {L1, L2, ..., L...} is obtained through sampling. N These samples are used to describe the variation of load over time. For ease of analysis, the probability distribution matrix P of the load demand is... L Defined as:

[0139] P L (i,j)=P(L i ∈[a i ,b i ],L j ∈[a j ,b j ]);

[0140] Where P L (i,j) represents the load value L. i and L j They fall into the interval [a] respectively i ,b i ] and [a j ,b jThe probability of P. L (i,j), construct a two-dimensional probability distribution matrix of load demand. For P L To fit the probability density function, assuming the load demand approximately follows a certain distribution, such as a normal distribution, its probability density function is expressed as:

[0141]

[0142] Where, μ L It is the average load demand. This is the variance of the load demand. By fitting f... L (x), thus obtaining the probability density model f of the load. L (x) is used to quantify the statistical characteristics of load demand. After obtaining the load probability density model, the temperature change data is analyzed based on the impact of load fluctuations on temperature changes. Let the temperature change sequence be T(t), its distribution characteristics are extracted, and the temperature fluctuation probability matrix P is constructed. T Its definition is similar to the load demand probability distribution matrix P. L By calculating the cumulative distribution of the temperature fluctuation probability matrix, a probabilistic model F for temperature change is generated. T (x), whose expression is:

[0143]

[0144] Among them, F T (x) is the cumulative distribution function of temperature change, f T (t) is the probability density function of temperature change. The temperature change probability model F... T (x) describes the probability distribution of temperature within a specific range. The load probability density model f... L (x) and the temperature change probability model F T (x0 combined, an uncertainty probability space is constructed. This space S is defined as the joint distribution of load and temperature, with probability density function f) L,T (x,y) can be represented as:

[0145] f L,T (x,y)=f L (x)·f T (y);

[0146] The coupling relationship between load and temperature is analyzed using the uncertainty probability space S. Based on the probability space S, a confidence analysis is performed on the system to calculate the boundary values ​​that satisfy a specific probability level α. For example, the confidence level boundary values ​​are determined by solving the following equation:

[0147]

[0148] Where Sα This is a probability subset that satisfies the confidence level, where α is typically set to a high confidence level value such as 95% or 99%. Based on the confidence level boundary values, the voltage-temperature constraint is transformed from a probabilistic domain constraint to a deterministic constraint. Let the battery voltage range be [V]. min V max ], temperature range is [T min ,T max Then the deterministic constraint condition is expressed as:

[0149] V min ≤V≤V max ,T min ≤T≤T max ;

[0150] Combining deterministic constraints, we construct a risk optimization objective function, assuming the objective function is to minimize the risk cost J, which has the following form:

[0151]

[0152] Among them, w i It is the weighting factor for the state of charge, SOC i It is the state of charge (SOC) of the i-th battery pack. avg λ is the average value of the state of charge, and λ is the weighting factor of the risk constraint. To optimize the objective function J, KKT conditions (Karush-Kuhn-Tucker conditions) need to be constructed to ensure the constraint satisfaction of the optimal solution. The Lagrangian function is defined. for:

[0153]

[0154] Among them, g i (x) is an inequality constraint, h j (x) is an equality constraint, μ i ,ν j They are Lagrange multipliers. Through the analysis of... Taking the partial derivatives and setting them to zero yields the first-order optimality condition matrix, which, when solved, provides the numerical solution to the optimization problem. This numerical solution is then input into the feedback correction unit to generate the compensated control parameters K. The control quantity after online compensation is defined as:

[0155] u′(t)=u(t)+Kx(t);

[0156] Set a safety margin Δ for the compensation control parameters:

[0157] K′=K+Δ;

[0158] Through the above steps, the safety constraint control parameters of the battery pack are finally generated.

[0159] In one specific embodiment, the process of performing step S4 may specifically include the following steps:

[0160] Based on the battery pack safety constraint control parameters, a risk assessment calculation is performed on the current system operating status to obtain real-time operating risk indicators. A risk assessment matrix is ​​then constructed based on these real-time operating risk indicators to obtain system risk level data.

[0161] The power limit is adjusted on the system risk level data to obtain dynamic power limit parameters, and the dynamic power limit parameters are then processed by piecewise linearization to obtain power control range data.

[0162] Based on the power control range data, the state of charge of each battery pack is weighted differently to obtain the charge and discharge power adjustment coefficient. Then, the power allocation is corrected based on the charge and discharge power adjustment coefficient to obtain the dynamic power allocation strategy.

[0163] Temperature equalization control is performed on the dynamic power distribution strategy to obtain the temperature control target value, and closed-loop control calculation is performed on the temperature control target value to obtain the temperature adjustment command.

[0164] Temperature compensation is performed on the charging and discharging power based on the temperature adjustment command to obtain the temperature compensation coefficient, and the power is adjusted according to the temperature compensation coefficient to obtain the real-time power control command.

[0165] The real-time power control commands are time-coordinated to obtain battery pack collaborative control data, which is then fed back to the main control unit.

[0166] Specifically, a risk assessment is performed on the current system operating state based on the battery pack safety constraint control parameters. Let the system operating state include voltage V. i Current I i Temperature T i and State of charge (SOC) i (where i = 1, 2, ..., n represents the i-th battery pack), the risk assessment calculates the real-time operational risk index R using the following formula. i :

[0167]

[0168] Where w1, w2, w3, and w4 are the weighting factors for each parameter, and V safe ,I safe ,T safe For reference values ​​of safe voltage, current, and temperature, SOC avg Let R be the average state of charge of all battery packs. Using the above formula, the operational risk of each battery pack is quantified, and a risk assessment matrix R is constructed:

[0169] R = [R1 R2…R]n ] T ;

[0170] Based on the risk assessment matrix R, the overall risk of the system is classified, generating system risk level data L. i :

[0171]

[0172] Where θ1 and θ2 are the thresholds for risk level classification. Based on the system risk level data L... i Power limits are adjusted for each battery pack. Assume the dynamic power limit parameter for each battery pack is P. i,max and P i,min The calculation formula is as follows:

[0173]

[0174] Where P safe This represents the maximum safe power output of the battery pack. The dynamic power limit parameters are piecewise linearized to obtain the power control range data Q. i :

[0175]

[0176] Through the above processing, suitable power control range data is generated for each battery pack. Based on the power control range data, differentiated weighting is applied to the state of charge (SOC) of each battery pack. A weighting factor β is defined. i for:

[0177]

[0178] Using weighting factor β i The charging and discharging power is corrected to generate a dynamic power allocation strategy.

[0179]

[0180] After correcting the power distribution strategy, temperature equalization control is applied, with a target temperature of T. target Temperature control target value Represented as:

[0181]

[0182] Where κ is the temperature adjustment coefficient. The temperature adjustment command ΔT is generated through closed-loop control calculations. i :

[0183]

[0184] Based on the temperature regulation command, temperature compensation is performed on the charging and discharging power to obtain the temperature compensation coefficient γ. i :

[0185]

[0186] By incorporating a temperature compensation coefficient, the dynamic power allocation strategy is modified to generate real-time power control commands.

[0187]

[0188] Real-time power control commands are time-coordinated to avoid system conflicts caused by multiple battery packs simultaneously adjusting their power. A timing integration algorithm generates battery pack collaborative control data C(t), which is expressed as:

[0189]

[0190] After the collaborative control data is fed back to the main control unit, global optimization control of the system is achieved.

[0191] The battery control method of the BMS management system in the embodiments of the present invention has been described above. The BMS management system in the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 2 One embodiment of the BMS management system in this invention includes:

[0192] The topology connection module is used to connect multiple battery packs in a topology to obtain a hierarchical control system consisting of a main control unit, multiple sub-control units, and multiple battery packs. Each battery pack is connected to a sub-control unit in a one-to-one correspondence, and the multiple sub-control units are connected to the main control unit through a data bus.

[0193] The equalization optimization module is used to perform recursive least squares parameter identification and battery charging and discharging power equalization optimization on voltage, current and temperature data in the hierarchical control system to obtain battery pack power distribution control commands.

[0194] The sequence planning module is used to perform Lyapunov stability analysis and sequence planning on battery pack power distribution control commands and load fluctuation data to obtain battery pack safety constraint control parameters.

[0195] The dynamic allocation module is used to dynamically allocate the charging and discharging power of the battery pack according to the battery pack safety constraint control parameters, obtain battery pack collaborative control data, and feed the battery pack collaborative control data back to the main control unit.

[0196] Through the collaborative efforts of the aforementioned components, a hierarchical BMS control system was established, enabling coordinated operation between the main control unit and sub-control units. This reduced the system's communication burden and improved the real-time performance and reliability of the control. The recursive least squares method was used for parameter identification, combined with extended Kalman filtering for state estimation, significantly improving the identification accuracy of the battery equivalent circuit model and the accuracy of state-of-charge calculation. A dynamic programming-based equilibrium optimization strategy ensured a reasonable allocation of charging and discharging power among multiple battery packs, effectively solving the imbalance problem between battery packs. The introduction of Lyapunov stability analysis and adaptive control algorithms enhanced the system's ability to suppress random disturbances such as load fluctuations, improving the robustness of the control system. By establishing a chance-constrained optimization model, probabilistic constraints were transformed into deterministic constraints, enabling safe operation of the system under uncertain environments. A risk assessment-based dynamic scheduling strategy, combined with temperature equalization control, ensured the safe operation of the battery packs and extended battery life. A closed-loop optimization mechanism was adopted, improving the overall control performance and operating efficiency of the system through real-time parameter updates and strategy optimization.

[0197] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0198] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0199] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A battery control method for a BMS management system, characterized in that, The method includes: Multiple battery packs are topologically connected to obtain a hierarchical control system consisting of a main control unit, multiple sub-control units, and multiple battery packs. Each of the multiple battery packs is connected to a sub-control unit in a one-to-one correspondence, and the multiple sub-control units are connected to the main control unit via a data bus. Recursive least squares parameter identification and battery charging and discharging power equalization optimization are performed on the voltage, current and temperature data in the hierarchical control system to obtain the battery pack power distribution control command. Lyapunov stability analysis and sequence planning were performed on the battery pack power distribution control commands and load fluctuation data to obtain the battery pack safety constraint control parameters. The charging and discharging power of the battery pack is dynamically allocated according to the battery pack safety constraint control parameters to obtain battery pack collaborative control data, and the battery pack collaborative control data is fed back to the main control unit.

2. The battery control method of the BMS management system according to claim 1, characterized in that, The process involves topologically connecting multiple battery packs to obtain a hierarchical control system consisting of a main control unit, multiple sub-control units, and multiple battery packs. Each battery pack is connected to a sub-control unit in a one-to-one correspondence. The sub-control units are connected to the main control unit via a data bus. Data acquisition channels are allocated to the voltage acquisition ports of the multiple battery packs to obtain a data acquisition port mapping table for n battery packs; Based on the data acquisition port mapping table, multiple sub-control units are configured with channels to obtain n independent data acquisition channels, and data transmission protocols are set for the n independent data acquisition channels to obtain RS485 communication link configuration data. The RS485 communication link configuration data is input into the communication module of multiple sub-control units to obtain a one-to-one correspondence between the sub-control units and the battery pack. The CAN communication interfaces of the multiple sub-control units are configured as master-slave nodes to obtain the communication link between the sub-control units and the master control unit. The data receiving module of the master control unit is initialized according to the communication link to obtain the master-slave controller data transmission channel. A tree-shaped control structure is constructed based on the master-slave controller data transmission channel to obtain a hierarchical controller topology. The hierarchical controller topology is then combined with the one-to-one correspondence connection relationship to construct a layered control system.

3. The battery control method of the BMS management system according to claim 1, characterized in that, The process of performing recursive least-squares parameter identification and battery charging / discharging power balancing optimization on voltage, current, and temperature data in the hierarchical control system to obtain battery pack power distribution control commands includes: The voltage and current data collected by the hierarchical control system are initialized with an error covariance matrix to obtain an initial parameter identification matrix. Based on the initial parameter identification matrix, the open-circuit voltage source of the battery equivalent circuit is recursively calculated to obtain the open-circuit voltage parameters. The open-circuit voltage parameters and internal resistance data are input into a second-order RC network model to obtain the equivalent circuit ohmic internal resistance and polarization internal resistance. Based on the equivalent circuit ohmic internal resistance and polarization internal resistance, the polarization time constant is calculated to obtain the battery equivalent circuit model parameters. The voltage, current and temperature data in the hierarchical control system are used to construct the state equation of the extended Kalman filter to obtain the state observation matrix. The state observation matrix is ​​then input into the dynamic iterative calculation unit for time update to obtain the battery state of charge value. The internal resistance change trend is calculated based on the battery equivalent circuit model parameters and temperature data to obtain the battery health status value. The battery state of charge value and the battery health status value are then combined to obtain battery status data. Based on the battery equivalent circuit model parameters and the battery state data, the battery charging and discharging power is balanced and optimized to obtain the battery pack power distribution control command.

4. The battery control method of the BMS management system according to claim 3, characterized in that, The step of optimizing the battery charging and discharging power based on the battery equivalent circuit model parameters and the battery state data to obtain battery pack power distribution control commands includes: The state of charge (SCC) values ​​in the battery state data are used to quantify the state differences to obtain the SCC difference matrix between adjacent battery packs. The weighting coefficients of the SCC difference matrix between adjacent battery packs are then configured to obtain the objective function for balanced optimization. Dynamic power limits are calculated based on the internal resistance parameters in the battery equivalent circuit model parameters to obtain the charge and discharge power threshold range. Constraint equations are then constructed based on the charge and discharge power threshold range to obtain mathematical models of voltage constraints, current constraints, and power constraints. Based on the objective function of the equilibrium optimization and the mathematical model, a state transition cost matrix is ​​constructed, and the optimal path search is performed on the state transition cost matrix to obtain power allocation data for each time step. Based on the power allocation data, adjustment coefficients are generated for the charging and discharging power of adjacent battery packs to obtain dynamic power commands for each battery pack. The dynamic power commands of each battery pack are then integrated in a timing sequence to obtain battery pack power allocation control commands.

5. The battery control method of the BMS management system according to claim 1, characterized in that, The Lyapunov stability analysis and sequence planning are performed on the battery pack power distribution control commands and load fluctuation data to obtain battery pack safety constraint control parameters, including: The load fluctuation data is subjected to amplitude feature extraction to obtain a load fluctuation amplitude sequence, and a probability statistical operation is performed on the load fluctuation amplitude sequence to obtain a load disturbance probability model. The load disturbance probability model and the battery pack power distribution control command are modeled in state space to obtain a random disturbance state equation. The random disturbance state equation is then linearized to obtain a linear state space model containing disturbance terms. Based on the linear state-space model containing the disturbance term, a positive definite quadratic Lyapunov function is constructed to obtain the system energy functional. The time derivative of the system energy functional is then calculated to obtain the system stability criterion. The system stability discriminant is configured with perfect square terms and coefficients matched to obtain a non-negative term matrix. Then, the non-negative term matrix is ​​subjected to eigenvalue decomposition to obtain the system stability constraints. The adaptive control parameters are recursively calculated based on the system stability constraints to obtain the feedback control gain sequence, and a parameter adjustment law is constructed based on the feedback control gain sequence to obtain the adaptive parameter update rule. The adaptive parameter update rule is input into the feedback controller to obtain the dynamic compensation control quantity, and the online parameters are adjusted according to the dynamic compensation control quantity to obtain adaptive feedback control data. Probability distribution analysis and sequence planning are performed on the adaptive feedback control data to obtain the battery pack safety constraint control parameters.

6. The battery control method of the BMS management system according to claim 5, characterized in that, The process of performing probability distribution analysis and sequence planning on the adaptive feedback control data to obtain battery pack safety constraint control parameters includes: Based on the adaptive feedback control data, historical sample statistics are performed on the load demand data to obtain the load demand probability distribution matrix, and the density function is fitted to the load demand probability distribution matrix to obtain the load probability density model. Based on the load probability density model, the distribution characteristics of the temperature change data are extracted to obtain the temperature fluctuation probability matrix, and the cumulative distribution of the temperature fluctuation probability matrix is ​​calculated to obtain the temperature change probability model. The load probability density model and the temperature change probability model are used to construct a probability space to obtain an uncertainty probability space. Confidence analysis is then performed on the uncertainty probability space to obtain confidence level boundary values. Based on the confidence level boundary value, the voltage and temperature constraints are probabilistically transformed to obtain deterministic constraints. Based on the deterministic constraints, a risk optimization objective function is constructed to obtain a sequence programming mathematical model. The KKT conditions are constructed on the sequence programming mathematical model to obtain the first-order optimality condition matrix, and gradient iteration calculation is performed based on the first-order optimality condition matrix to obtain the numerical solution of the optimization problem. The numerical solution of the optimization problem is input into the feedback correction unit for closed-loop compensation to obtain the compensation control parameters. The safety margin of the compensation control parameters is then set to obtain the battery pack safety constraint control parameters.

7. The battery control method of the BMS management system according to claim 1, characterized in that, The step of dynamically allocating the charging and discharging power of the battery pack according to the battery pack safety constraint control parameters to obtain battery pack collaborative control data, and feeding back the battery pack collaborative control data to the main control unit, includes: Based on the battery pack safety constraint control parameters, a risk assessment calculation is performed on the current system operating status to obtain real-time operating risk indicators. A risk assessment matrix is ​​then constructed based on the real-time operating risk indicators to obtain system risk level data. The system risk level data is adjusted by power limit to obtain dynamic power limit parameters, and the dynamic power limit parameters are then processed by piecewise linearization to obtain power control interval data. Based on the power control range data, the state of charge of each battery pack is weighted differently to obtain the charge and discharge power adjustment coefficient. Based on the charge and discharge power adjustment coefficient, the power allocation is corrected to obtain a dynamic power allocation strategy. Temperature equalization control is applied to the dynamic power distribution strategy to obtain a temperature control target value, and closed-loop control calculation is performed on the temperature control target value to obtain a temperature adjustment command. Based on the temperature adjustment command, the charging and discharging power is temperature compensated to obtain a temperature compensation coefficient, and the power is adjusted according to the temperature compensation coefficient to obtain a real-time power control command. The real-time power control command is time-coordinated to obtain battery pack collaborative control data, and the battery pack collaborative control data is fed back to the main control unit.

8. A BMS management system, characterized in that, A battery control method for performing a BMS management system as described in any one of claims 1-7, the BMS management system comprising: The topology connection module is used to connect multiple battery packs in a topology to obtain a hierarchical control system consisting of a main control unit, multiple sub-control units, and multiple battery packs. Each of the multiple battery packs is connected to a sub-control unit in a one-to-one correspondence, and the multiple sub-control units are connected to the main control unit through a data bus. The equalization optimization module is used to perform recursive least squares parameter identification and battery charging and discharging power equalization optimization on the voltage data, current data and temperature data in the hierarchical control system to obtain battery pack power distribution control instructions. The sequence planning module is used to perform Lyapunov stability analysis and sequence planning on the battery pack power distribution control commands and load fluctuation data to obtain battery pack safety constraint control parameters. The dynamic allocation module is used to dynamically allocate the charging and discharging power of the battery pack according to the battery pack safety constraint control parameters, obtain battery pack collaborative control data, and feed the battery pack collaborative control data back to the main control unit.

Images