Distributed BMS battery active equalization type safety management method and system

Abstract

The invention provides a distributed BMS battery active equalization type safety management method and system, and the method comprises the steps: obtaining the state information and performance change trend corresponding to each single battery through real-time monitoring, constructing a target battery pack topological structure model, combining a reinforcement learning algorithm, generating an intelligent agent model, and employing the intelligent agent model and a fuzzy logic controller to achieve the active equalization type safety management of a battery pack. Determining an energy flow management scheme; based on the feedback data and historical data of the energy flow management scheme, using a deep belief network to construct an anomaly detection model, based on the anomaly detection model, generating a risk assessment report, and constructing a security record platform based on the block chain technology to obtain security record information; optimizing each model parameter based on the risk assessment report and the security record information, and generating a security management system in combination with a transfer learning algorithm and a quantum computing method; and high-precision monitoring, intelligent electric quantity balance regulation and control, anomaly detection, risk management and continuous optimization of the distributed BMS battery system are realized.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of battery management technology, and in particular to a distributed BMS battery active balancing safety management method and system. Background Art

[0002] With the rapid development of electric vehicles, energy storage systems and other fields, distributed BMS battery management systems play a vital role in ensuring the safety of battery packs and extending their service life.

[0003] At present, the common distributed BMS battery management solutions on the market mainly include traditional rule-based control systems and centralized management architectures. Traditional solutions often rely on fixed thresholds and rules, and cannot accurately monitor the status changes of each single battery in real time, resulting in potential problems not being discovered in time, increasing safety risks. The centralized management architecture method uploads all data to the central controller for processing, which has high requirements for communication bandwidth and real-time performance, and has a high risk of single-point failure, making it difficult to provide comprehensive safety protection. Summary of the invention

[0004] The embodiment of the present invention provides a distributed BMS battery active balancing safety management method and system to solve the problems of insufficient monitoring accuracy, great safety hazards, and difficulty in providing comprehensive safety protection in the prior art.

[0005] In a first aspect, an embodiment of the present invention provides a distributed BMS battery active balancing safety management method, comprising:

[0006] Real-time monitoring of the voltage, temperature and internal resistance of each single battery to obtain battery data, and state evaluation and trend prediction of the battery data to obtain the state information and performance change trend corresponding to each single battery;

[0007] Based on the state information and the performance change trend, a target battery pack topology model is constructed, and based on the key information of the internal structure of the battery pack in the target battery pack topology model and a reinforcement learning algorithm, an intelligent agent model is generated, and a preliminary power balancing strategy is selected by using the intelligent agent model, and the preliminary power balancing strategy is optimized to obtain an optimal power balancing strategy;

[0008] Based on the optimal power balancing strategy, a fuzzy logic controller is used to optimize the regulation of energy flow between each single battery to obtain an optimized regulation result. Based on the optimized regulation result, an energy flow management plan is determined, and a sliding mode variable structure control theory and an optimized energy transfer strategy are introduced in the regulation process for optimization;

[0009] Apply the energy flow management solution to obtain feedback data, build an anomaly detection model using a deep belief network based on the feedback data and historical data, identify abnormal operation modes using the anomaly detection model, analyze the abnormal operation modes, obtain an abnormal analysis result, generate a risk assessment report based on the abnormal analysis result, build a security record platform based on blockchain technology, and obtain security record information based on the security record platform;

[0010] Based on the risk assessment report and the safety record information, the parameters of each model are optimized to obtain an optimized model. The optimized model, the fuzzy logic controller, the safety record platform, the transfer learning algorithm and the quantum computing method are combined to generate a safety management system. The models include the battery pack topology model, the intelligent agent model and the anomaly detection model.

[0011] Optionally, based on the state information and the performance change trend, a target battery pack topology model is constructed, based on key information of the internal structure of the battery pack in the target battery pack topology model and a reinforcement learning algorithm, an intelligent agent model is generated, a preliminary power balancing strategy is selected using the intelligent agent model, the preliminary power balancing strategy is optimized, and an optimal power balancing strategy is obtained, including:

[0012] Based on the state information and the performance change trend, the internal connection relationship and energy transfer path of the battery pack are modeled and processed in combination with a graph theory method to construct a target battery pack topology structure model, and the target battery pack topology structure model is used to analyze the impact of different connection modes on the performance of the distributed BMS battery system to obtain a connection analysis result, and based on the connection analysis result, key information of the internal structure of the battery pack is determined;

[0013] Based on the key information, combined with the reinforcement learning algorithm and the simulated annealing algorithm, the initial model is trained in a simulation environment to obtain an intelligent agent model, the intelligent agent model and the particle swarm optimization algorithm are used to select a preliminary power balancing strategy, and the ant colony optimization algorithm is introduced to optimize the preliminary power balancing strategy to generate a preliminary optimized power balancing strategy;

[0014] A multi-objective optimization algorithm is applied to comprehensively optimize the power balancing strategy after preliminary optimization to obtain a global optimal solution. During the comprehensive optimization process, a Bayesian optimization algorithm is used to adjust the hyperparameters of the multi-objective optimization algorithm to enhance the optimization. The global optimal solution represents the best power balancing strategy.

[0015] Optionally, based on the state information and the performance change trend, the internal connection relationship and energy transfer path of the battery pack are modeled and processed in combination with a graph theory method to construct a target battery pack topology model. The target battery pack topology model is used to analyze the impact of different connection modes on the performance of the distributed BMS battery system to obtain connection analysis results, including:

[0016] Based on the state information and the performance change trend, the internal connection relationship and energy transfer path of the battery pack are modeled and processed in combination with graph theory methods and graph neural networks to construct a preliminary battery pack topology model, and the target battery pack topology model is constructed by combining the preliminary battery pack topology model and the Bayesian network;

[0017] Analyzing conditional dependencies between variables using a Bayesian network in the target battery pack topology model, wherein the variables include components of a distributed BMS battery system, properties of the components, and interactions between the components;

[0018] Use causal reasoning technology to analyze the causal relationship between different connection modes and distributed BMS battery system performance;

[0019] Generative adversarial networks are introduced to simulate extreme working conditions to obtain the performance of different connection methods under extreme conditions and obtain the performance results of extreme working conditions;

[0020] Based on the conditional dependency, the causal relationship and the extreme operating condition performance results, the impact of different connection modes on the performance of the distributed BMS battery system is comprehensively evaluated to obtain the connection analysis results.

[0021] Optionally, based on the optimal power balancing strategy, a fuzzy logic controller is used to optimize the regulation of energy flow between each single battery to obtain an optimized regulation result. Based on the optimized regulation result, an energy flow management plan is determined, and a sliding mode variable structure control theory and an optimized energy transfer strategy are introduced in the regulation process for optimization, including:

[0022] Using the optimal power balancing strategy, the energy flow between each single battery is regulated to obtain a preliminary regulation result, and a fuzzy logic controller is used to dynamically adjust the preliminary regulation result to obtain an optimized regulation result;

[0023] According to the optimization and control results, a target energy transfer path and a target energy transfer mode are selected by using a multi-agent reinforcement learning method, the target energy transfer path and the target energy transfer mode are optimized to obtain an optimal energy transfer path and an optimal target energy transfer mode, and an energy flow management plan is determined based on the optimal energy transfer path and the optimal target energy transfer mode;

[0024] An event trigger mechanism is introduced into the regulation process, and the existing energy transfer strategy is optimized to generate an optimized energy transfer strategy so that the fuzzy logic controller can perform corresponding regulation when a specific event occurs. Based on the optimized energy transfer strategy and sliding mode variable structure control theory, the regulation of energy flow between each single cell is optimized. The specific event includes that the voltage or temperature of the single cell is higher than a preset threshold.

[0025] Optionally, according to the optimization and control result, a target energy transfer path and a target energy transfer mode are selected by using a multi-agent reinforcement learning method, the target energy transfer path and the target energy transfer mode are optimized to obtain an optimal energy transfer path and an optimal target energy transfer mode, and based on the optimal energy transfer path and the optimal target energy transfer mode, an energy flow management plan is determined, including:

[0026] Based on the optimization and control results, a multi-agent reinforcement learning method is used to simulate the energy transfer effects under different energy transfer paths, and based on the energy transfer effects, a target energy transfer path is selected, wherein the energy transfer path includes a multi-level transfer path and a centralized path based on a bus, and the energy transfer effects include transfer efficiency and thermal management performance;

[0027] Based on the optimization and control results, a multi-agent reinforcement learning method is used to simulate the performance effects under different energy transfer modes, and based on the performance effects, a target energy transfer mode is selected, where the energy transfer mode includes wireless energy transfer and transfer through an intermediate energy storage device;

[0028] A mixed integer linear programming method is introduced to perform constrained optimization on the target energy transfer path and the target energy transfer mode to obtain the optimal energy transfer path and the optimal energy transfer mode. Based on the optimal energy transfer path and the optimal target energy transfer mode, an energy flow management plan is determined.

[0029] Optionally, the energy flow management solution is applied to obtain feedback data, and based on the feedback data and historical data, an anomaly detection model is constructed using a deep belief network, and the abnormal operation mode is identified using the anomaly detection model, and the abnormal operation mode is analyzed to obtain an abnormal analysis result, and a risk assessment report is generated based on the abnormal analysis result, and a security record platform based on blockchain technology is constructed, and security record information is obtained based on the security record platform, including:

[0030] Using the energy flow management scheme, the energy flow between the single cells is regulated to obtain feedback data, wherein the feedback data includes the voltage, temperature and internal resistance of each single cell after regulation;

[0031] Based on the feedback data and historical data, an anomaly detection model is constructed using a deep belief network, and the operating state of the distributed BMS battery system is monitored in real time using the anomaly detection model to identify abnormal operation modes, wherein the abnormal operation modes include an overcharge mode, an over-discharge mode, and an abnormal temperature mode;

[0032] Analyze the abnormal operation mode, determine the abnormal cause and the abnormal impact range, and obtain an abnormal analysis result based on the abnormal cause and the abnormal impact range;

[0033] Based on the abnormal analysis results, assess the risk level of the distributed BMS battery system and generate a risk assessment report, wherein the risk assessment report includes the risk type and the probability of risk occurrence;

[0034] A security record platform is constructed using blockchain technology. Based on the security record platform, operation logs, abnormal events and risk assessment reports are extracted to obtain security record information.

[0035] Optionally, based on the abnormal analysis result, the risk level of the distributed BMS battery system is evaluated and a risk assessment report is generated. The risk assessment report includes the risk type and the probability of risk occurrence, including:

[0036] Based on the abnormal analysis results, the risk level of the distributed BMS battery system is evaluated, and based on the risk level, the risk scenarios and the probability of occurrence of the risk scenarios are simulated using a Monte Carlo simulation method, wherein the risk scenarios include overcharge risk scenarios, over-discharge risk scenarios, and temperature abnormality risk scenarios;

[0037] Using the Markov chain method, the transition probability of the distributed BMS battery system between different risk states is defined, and the risk evolution trend of the distributed BMS battery system is predicted. Based on the transition probability and the risk evolution trend, the risk propagation analysis result is obtained. The risk state includes a normal state, an abnormal state, and a fault state.

[0038] A risk assessment report is generated based on the risk level, the risk scenario, the probability of occurrence of the risk scenario, and the risk propagation analysis result.

[0039] In a second aspect, an embodiment of the present invention provides a distributed BMS battery active balancing safety management system, including:

[0040] The monitoring module is used to monitor the voltage, temperature and internal resistance of each single battery in real time, obtain battery data, perform status evaluation and trend prediction on the battery data, and obtain the status information and performance change trend corresponding to each single battery;

[0041] A construction module is used to construct a target battery pack topology model based on the state information and the performance change trend, generate an intelligent agent model based on key information of the internal structure of the battery pack in the target battery pack topology model and a reinforcement learning algorithm, select a preliminary power balancing strategy using the intelligent agent model, optimize the preliminary power balancing strategy, and obtain an optimal power balancing strategy;

[0042] A determination module is used to optimize the regulation of energy flow between each single battery using a fuzzy logic controller based on the optimal power balancing strategy to obtain an optimized regulation result, determine an energy flow management plan based on the optimized regulation result, and introduce sliding mode variable structure control theory and optimized energy transfer strategy in the regulation process for optimization;

[0043] an identification module, used for applying the energy flow management solution to obtain feedback data, constructing an anomaly detection model based on the feedback data and historical data using a deep belief network, identifying abnormal operation modes using the anomaly detection model, analyzing the abnormal operation modes, obtaining an abnormal analysis result, generating a risk assessment report based on the abnormal analysis result, constructing a security record platform based on blockchain technology, and obtaining security record information based on the security record platform;

[0044] A generation module is used to optimize the parameters of each model based on the risk assessment report and the safety record information to obtain an optimized model, and to generate a safety management system by combining the optimized model, the fuzzy logic controller, the safety record platform, the transfer learning algorithm and the quantum computing method. The various models include the battery pack topology model, the intelligent agent model and the anomaly detection model.

[0045] In a third aspect, an embodiment of the present invention provides a computing device, comprising a processor and a memory, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute a distributed BMS battery active balancing safety management method as described in any one of the first aspects.

[0046] In a fourth aspect, an embodiment of the present invention provides a computer storage medium having computer program instructions stored thereon, wherein the computer program instructions, when executed by a processor, implement a distributed BMS battery active balancing safety management method as described in any one of the first aspects.

[0047] In the embodiment of the present invention, the voltage, temperature and internal resistance of each single battery are monitored in real time to obtain battery data, and the battery data is evaluated and trend predicted to obtain the state information and performance change trend corresponding to each single battery; based on the state information and the performance change trend, a target battery pack topology model is constructed, and based on the key information of the internal structure of the battery pack in the target battery pack topology model and the reinforcement learning algorithm, an intelligent agent model is generated, and a preliminary power balancing strategy is selected by using the intelligent agent model, and the preliminary power balancing strategy is optimized to obtain an optimal power balancing strategy; based on the optimal power balancing strategy, a fuzzy logic controller is used to optimize the regulation of energy flow between each single battery to obtain an optimized regulation result, and based on the optimized regulation result, an energy flow management plan is determined, and the sliding mode variable structure control theory and The optimized energy transfer strategy is optimized; the energy flow management scheme is applied to obtain feedback data, based on the feedback data and historical data, an anomaly detection model is constructed using a deep belief network, the anomaly detection model is used to identify abnormal operation modes, the abnormal operation modes are analyzed, and an abnormal analysis result is obtained, based on the abnormal analysis result, a risk assessment report is generated, and a security record platform based on blockchain technology is constructed, and security record information is obtained based on the security record platform; based on the risk assessment report and the security record information, the parameters of each model are optimized to obtain an optimized model, and a security management system is generated by combining the optimized model, the fuzzy logic controller, the security record platform, the transfer learning algorithm and the quantum computing method, wherein each model includes the battery pack topology model, the intelligent agent model and the anomaly detection model. The technical solution provided by the present invention realizes high-precision real-time monitoring, intelligent power balancing control, anomaly detection and risk management, as well as continuous optimization and adaptive adjustment of distributed BMS battery systems. It helps to detect potential problems in a timely manner and prevent events such as overcharging, over-discharging and abnormal temperature. It improves the accuracy of monitoring, can also warn of possible risks in advance, and enhances the system's prevention capabilities. It uses fuzzy logic controllers to further optimize energy flow regulation to ensure a smoother and more efficient regulation process, reduce energy loss, and extend battery life. The generated safety management system not only has high-precision monitoring, intelligent regulation and anomaly detection functions, but can also continuously optimize itself during operation, ensuring the stable operation of the battery pack under various working conditions, reducing maintenance costs, and providing all-round safety protection.

[0048] These and other aspects of the present invention will become more apparent from the following description of the embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0050] Figure 1 A flow chart of a distributed BMS battery active balancing safety management method provided by an embodiment of the present invention;

[0051] Figure 2 A schematic diagram of the structure of a distributed BMS battery active balancing safety management system provided by an embodiment of the present invention;

[0052] Figure 3 A schematic diagram of the structure of a computing device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0053] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0054] In some of the processes described in the specification and claims of the present invention and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0055] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0056] Figure 1 A flowchart of a distributed BMS battery active balancing safety management method is provided for an embodiment of the present invention, such as Figure 1 As shown, the method includes:

[0057] With the rapid development of applications such as electric vehicles and energy storage systems, the efficient and safe operation of battery management systems has become crucial. Traditional battery management systems have many deficiencies in monitoring accuracy, control flexibility and risk management, and are difficult to meet the increasingly complex application requirements. In order to meet the challenges faced by modern battery management, based on this, the present invention provides a distributed BMS battery active balancing safety management method, such as Figure 1 ,include:

[0058] Step 101: monitor the voltage, temperature and internal resistance of each single battery in real time to obtain battery data, perform status evaluation and trend prediction on the battery data, and obtain status information and performance change trend corresponding to each single battery;

[0059] In this step, state assessment refers to evaluating the current health status of the battery based on the collected data. Trend prediction refers to predicting the performance change trend in the future based on historical data and current data.

[0060] This step assumes that there is a battery pack consisting of 24 single lithium batteries, and the voltage, temperature and internal resistance data of each single battery collected by high-precision sensors every second are monitored in real time. For example, the voltage of a single battery at a certain moment is monitored to be 3.7V, the temperature is 25℃, and the internal resistance is 0.5Ω. These data are transmitted to the central control system in real time and stored in a database; machine learning algorithms are used to perform status assessment and trend prediction on these data. For example, by analyzing the data of recent days, it is found that the internal resistance of a single battery is gradually increasing, which may indicate that the battery is about to have problems, that is, the current state of the battery is slightly abnormal, and it is predicted that if it continues to be used, its internal resistance may further increase to 0.8Ω in the next week, thereby affecting the performance of the entire battery pack.

[0061] Step 102: constructing a target battery pack topology model based on the state information and the performance change trend, generating an intelligent agent model based on key information of the internal structure of the battery pack in the target battery pack topology model and a reinforcement learning algorithm, selecting a preliminary power balancing strategy using the intelligent agent model, optimizing the preliminary power balancing strategy, and obtaining an optimal power balancing strategy;

[0062] In this step, the energy balancing strategy refers to a method to ensure that the power levels of all cells in the battery pack are as consistent as possible.

[0063] This step constructs a topological structure model of the battery pack based on the state evaluation and performance change trend obtained in step 101. For example, assuming that the battery pack is connected in series, the connection relationship between each single cell is accurately modeled; the intelligent agent model is generated using a reinforcement learning algorithm, and the model can select a preliminary power balancing strategy based on the current battery state. For example, the intelligent agent model may choose a simple balancing strategy, that is, when the voltage of a single cell is too high, its charging current is reduced through a bypass circuit. Then, this preliminary strategy is optimized, considering multiple factors such as minimizing energy loss and maximizing balancing speed, and finally the best power balancing strategy is obtained. For example, the optimized strategy may include adjusting the charging or discharging rate of certain single cells to ensure a more uniform power distribution of the entire battery pack.

[0064] Step 103: Based on the optimal power balancing strategy, a fuzzy logic controller is used to optimize the regulation of energy flow between each single battery to obtain an optimized regulation result. Based on the optimized regulation result, an energy flow management plan is determined, and a sliding mode variable structure control theory and an optimized energy transfer strategy are introduced in the regulation process for optimization;

[0065] In this step, the Energy Flow Management Plan refers to a detailed energy distribution and control plan to ensure efficient operation of the battery pack.

[0066] This step is based on the optimal power balancing strategy obtained in step 102, and uses a fuzzy logic controller to optimize the energy flow between each single cell. For example, the fuzzy logic controller can dynamically adjust the charge / discharge current according to the real-time status of the battery to ensure that the energy flow is smoother and more efficient. In order to further optimize the control effect, the sliding mode variable structure control theory is introduced. For example, when adjusting the charging current of a single cell, the sliding mode control can ensure that the system can respond quickly and remain stable even when external conditions change. In addition, the optimized energy transfer strategy is also applied to ensure efficient energy transfer between different single cells and reduce energy loss. Finally, a detailed energy flow management plan is generated, which clarifies the specific control measures for each single cell to ensure that the battery pack can operate efficiently under various working conditions.

[0067] Step 104: applying the energy flow management solution to obtain feedback data, using a deep belief network to build an anomaly detection model based on the feedback data and historical data, using the anomaly detection model to identify abnormal operation modes, analyzing the abnormal operation modes, obtaining an abnormal analysis result, generating a risk assessment report based on the abnormal analysis result, building a security record platform based on blockchain technology, and obtaining security record information based on the security record platform;

[0068] In this step, the Risk Assessment Report refers to a report summarizing the current risk level and potential threats facing the system.

[0069] In this step, the distributed BMS battery system is actually operated according to the energy flow management scheme obtained in step 103, and feedback data of the system is collected. For example, the actual voltage, temperature and internal resistance data of a single cell deviate from the expected values, indicating that there may be an abnormal situation. The deep belief network is used to build an anomaly detection model, and the model can identify abnormal operation modes by learning a large amount of historical data. For example, the model detects that the temperature of a single cell rises abnormally, which may be caused by overcharging; the abnormal operation mode is analyzed in detail to obtain the cause and scope of the abnormality, and the abnormal analysis results are generated. Based on these results, a risk assessment report is generated to indicate the current risk type and probability of occurrence. For example, the report indicates that there is a 5% probability that a single cell will have an overheating failure in the next week. At the same time, a security recording platform based on blockchain technology is built to save all operation logs, abnormal events and risk assessment reports to ensure that the data is transparent and cannot be tampered with. Finally, the security record information is extracted from the security recording platform to help managers understand the operating status and potential risks of the system and take preventive measures in time.

[0070] Step 105: Based on the risk assessment report and the safety record information, optimize the parameters of each model to obtain an optimized model, and combine the optimized model, the fuzzy logic controller, the safety record platform, the transfer learning algorithm and the quantum computing method to generate a safety management system, wherein each model includes the battery pack topology model, the intelligent agent model and the anomaly detection model;

[0071] This step optimizes the parameters of each model based on the risk assessment report and safety record information obtained in step 104. For example, the connection relationship parameters in the battery pack topology model are updated, the selection strategy of the intelligent agent model is improved, and the threshold of the anomaly detection model is adjusted. A comprehensive safety management system is generated by combining the optimized model, fuzzy logic controller, safety record platform, transfer learning algorithm and quantum computing method. For example, the transfer learning algorithm can help the safety management system adapt to the new working environment more quickly, while the quantum computing law can provide higher efficiency in complex computing tasks. In the end, this safety management system not only has high-precision monitoring, intelligent control and anomaly detection functions, but also can continuously optimize itself during operation to provide all-round safety protection. In this way, the safety and reliability of the battery pack are significantly improved, the battery life is extended, and the maintenance cost is reduced.

[0072] The embodiments of the present invention can achieve the following beneficial effects through the above steps:

[0073] Ensure that the status of each single battery can be accurately tracked and potential problems can be discovered in time;

[0074] Improve the flexibility and response speed of the safety management system through adaptive control strategies, reduce energy loss and extend battery life;

[0075] Effectively identify potential risks, provide early warnings and generate detailed risk assessment reports to enhance system security;

[0076] The safety management system can continuously improve itself, adapt to new working conditions and environmental changes, and improve overall performance;

[0077] Combining multiple advanced technologies, it provides all-round safety management and monitoring to ensure stable operation of the battery pack under various working conditions.

[0078] In a distributed BMS battery management system, achieving efficient and safe power balancing is the key to ensuring the performance and life of the battery pack. Traditional methods often rely on static models and fixed rules, and are difficult to adapt to complex dynamic conditions. In order to overcome these limitations and improve the adaptability and response speed of the system, based on this, the present invention provides a specific embodiment, step 102, based on the state information and the performance change trend, construct a target battery pack topology model, based on the key information of the internal structure of the battery pack in the target battery pack topology model and the reinforcement learning algorithm, generate an intelligent agent model, use the intelligent agent model to select a preliminary power balancing strategy, optimize the preliminary power balancing strategy, and obtain the optimal power balancing strategy, which specifically includes the following steps:

[0079] Step 210: Based on the state information and the performance change trend, the internal connection relationship and energy transfer path of the battery pack are modeled and processed in combination with a graph theory method to construct a target battery pack topology structure model, and the target battery pack topology structure model is used to analyze the impact of different connection modes on the performance of the distributed BMS battery system to obtain a connection analysis result, and based on the connection analysis result, key information of the internal structure of the battery pack is determined;

[0080] In this step, key information refers to the information extracted from the connection analysis results that has a significant impact on the performance of the distributed BMS battery system, including connection topology, energy transfer path, performance influencing factors, and fault sensitivity.

[0081] This step assumes that there is a battery pack consisting of 48 single lithium batteries. The status information and performance change trend of each single battery have been obtained through real-time monitoring and status evaluation. In order to further optimize the performance of the battery pack, the graph theory method is used to model the connection relationship and energy transfer path inside the battery pack. Specifically, each single battery is regarded as a node in the graph, and the connection between them is regarded as an edge to construct a graph model. For example, if the battery pack is connected in series, there is an edge between each single battery to represent its direct connection relationship; using this target battery pack topology model, the impact of different connection methods on the performance of the distributed BMS battery system is analyzed. For example, by comparing the effects of series connection and parallel connection, it is found that although the series connection increases the total voltage, the failure of a single battery will affect the entire battery pack; while the parallel connection can improve redundancy, there may be uneven current distribution; through this analysis, detailed connection analysis results are obtained, from which key information of the internal structure of the battery pack is determined. For example, it is found that series connection is more effective in high-voltage applications, but requires stronger monitoring and protection mechanisms; while parallel connection is suitable for low-voltage and high-capacity applications, but the current distribution strategy needs to be optimized.

[0082] Step 211: Based on the key information, the initial model is trained in a simulation environment in combination with a reinforcement learning algorithm and a simulated annealing algorithm to obtain an intelligent agent model, a preliminary power balancing strategy is selected using the intelligent agent model and a particle swarm optimization algorithm, and an ant colony optimization algorithm is introduced to optimize the preliminary power balancing strategy to generate a preliminary optimized power balancing strategy;

[0083] In this step, the simulation environment is a key training platform, which provides a learning scenario close to the real world for the training of the initial model, ensuring that the trained intelligent agent model can perform well in complex actual environments and enhancing the ability of the intelligent agent model to deal with various uncertainties and challenges.

[0084] This step is based on key information, combined with reinforcement learning algorithm and simulated annealing algorithm, to train the initial model in a simulated environment. For example, a virtual environment is created to simulate the operation of the battery pack under different working conditions, and the intelligent agent model is allowed to continuously try different power balancing strategies in this environment to find the best solution. The reinforcement learning algorithm helps the intelligent agent model learn how to choose the best action based on the current state, while the simulated annealing algorithm ensures that the model can jump out of the local optimal solution and find the global optimal solution. After multiple iterations of training, an intelligent agent model is finally obtained, which can select a preliminary power balancing strategy based on the state information of the battery pack. In order to further optimize this strategy, the particle swarm optimization algorithm is introduced. For example, the intelligent agent model selects a preliminary power balancing strategy, that is, when the voltage of a single cell is too high, its charging current is reduced through a bypass circuit. The particle swarm optimization algorithm optimizes this strategy, considering multiple factors such as minimizing energy loss and maximizing balancing speed, and generates a preliminary optimized power balancing strategy. To further improve the optimization effect, the system introduces the ant colony optimization algorithm. For example, the ant colony optimization algorithm simulates the foraging behavior of ants to find the shortest path, thereby optimizing the energy transfer path in the power balancing strategy. After optimization by the ant colony optimization algorithm, a preliminary optimized power balancing strategy was generated to ensure efficient energy transfer between different single cells and reduce energy loss.

[0085] Step 212: applying a multi-objective optimization algorithm to comprehensively optimize the power balancing strategy after the preliminary optimization to obtain a global optimal solution. In the comprehensive optimization process, a Bayesian optimization algorithm is used to adjust the hyperparameters of the multi-objective optimization algorithm to enhance the optimization. The global optimal solution represents the best power balancing strategy.

[0086] In order to further improve the effect of the power balancing strategy, this step applies a multi-objective optimization algorithm to comprehensively optimize the power balancing strategy after preliminary optimization. For example, two objective functions are set: one is to minimize energy loss, and the other is to maximize the balancing speed. While considering these two objectives, the multi-objective optimization algorithm seeks the Pareto optimal solution, that is, to achieve the best balance point between the two. In the comprehensive optimization process, the Bayesian optimization algorithm is used to adjust the hyperparameters of the multi-objective optimization algorithm. For example, the Bayesian optimization algorithm predicts the optimization effect under different hyperparameter settings by constructing a probability model, and gradually converges to the optimal solution. This not only speeds up the optimization speed, but also improves the optimization quality. After the joint action of the multi-objective optimization algorithm and the Bayesian optimization algorithm, the global optimal solution is finally obtained to characterize the optimal power balancing strategy. For example, a power balancing strategy that can minimize energy loss while maximizing the balancing speed is found, ensuring that the power distribution of the entire battery pack is more uniform, extending the battery life and improving the overall performance of the system.

[0087] The embodiments of the present invention can achieve the following beneficial effects through the above steps:

[0088] The battery pack topology model is constructed by graph theory, and the impact of different connection methods on system performance is deeply analyzed to provide a scientific basis for optimal design;

[0089] Combine multiple algorithms such as reinforcement learning, simulated annealing, particle swarm optimization and ant colony optimization to generate intelligent agent models, select and optimize power balancing strategies, and significantly improve system flexibility and response speed;

[0090] Apply multi-objective optimization algorithms and Bayesian optimization algorithms to ensure that the power balancing strategy achieves the best balance between multiple objectives, find the global optimal solution, and further improve system performance;

[0091] By optimizing the power balancing strategy, we ensure that the battery pack can operate efficiently under various working conditions, extend battery life, reduce maintenance costs, and provide all-round safety protection.

[0092] In a distributed BMS battery management system, the internal connection relationship and energy transfer path of the battery pack have an important impact on the overall performance and stability of the system. Traditional methods often find it difficult to accurately describe and optimize these complex relationships. In order to achieve more efficient system design and performance optimization, a scientific basis is provided for optimizing system configuration. Based on this, the present invention provides a specific embodiment, step 210, based on the state information and the performance change trend, combined with the graph theory method, the internal connection relationship and energy transfer path of the battery pack are modeled to construct a target battery pack topology model, and the target battery pack topology model is used to analyze the impact of different connection methods on the performance of the distributed BMS battery system to obtain a connection analysis result, which specifically includes the following steps:

[0093] Step 221: Based on the state information and the performance change trend, the internal connection relationship and energy transfer path of the battery pack are modeled and processed in combination with graph theory methods and graph neural networks to construct a preliminary battery pack topology model, and the target battery pack topology model is constructed by combining the preliminary battery pack topology model and the Bayesian network;

[0094] This step assumes that there is a battery pack consisting of 48 single lithium batteries. The system has obtained the status information and performance change trend of each single battery through real-time monitoring and status evaluation. For example, the gradual increase in the internal resistance of some single batteries may indicate that these batteries are about to have problems. In order to further optimize the performance of the battery pack, the system combines graph theory methods and graph neural networks to model the connection relationship and energy transfer path within the battery pack. Specifically: each single battery is regarded as a node in the graph, and the connection between them is regarded as an edge to build a graph model. For example, if the battery pack is connected in series, there is an edge between each single battery to represent its direct connection relationship. The graph neural network is used to capture the complex relationship between nodes (single batteries), including but not limited to voltage, temperature, internal resistance and other attributes. This helps to more accurately model the energy transfer path within the battery pack. After the above processing, the system constructs a preliminary battery pack topology model. Next, the system introduces the Bayesian network to further improve the model. The Bayesian network can represent the conditional dependency between variables and support causal reasoning, thereby building a more accurate target battery pack topology model. For example, a Bayesian network can represent the relationship between the temperature anomaly of a single cell and the status of other single cells, and how this relationship affects the performance of the entire battery pack.

[0095] Step 222: Analyze the conditional dependencies between variables using the Bayesian network in the target battery pack topology model, where the variables include components of the distributed BMS battery system, properties of the components, and interactions between the components;

[0096] In this step, in the target battery pack topology model constructed, the Bayesian network is used to analyze the conditional dependencies between variables. Specifically: components, including single cells, sensors, controllers, etc. Component properties, such as the voltage, temperature, and internal resistance of single cells; sensor accuracy and response time; controller algorithm type, etc. Interactions, such as the connection method between single cells, energy transfer path, bypass circuit, etc. Through the Bayesian network, the system can identify the conditional dependencies between different components and their properties. For example: the temperature abnormality of a single cell may affect the temperature distribution of adjacent single cells. The algorithm type of the controller determines the choice of power balancing strategy, which in turn affects the energy transfer efficiency between single cells. These conditional dependencies provide the basis for subsequent causal reasoning, helping the system better understand the interactions between components and their impact on overall performance.

[0097] Step 223: Analyze the causal relationship between different connection modes and the performance of the distributed BMS battery system using causal reasoning technology;

[0098] This step uses causal reasoning technology to analyze the causal relationship between different connection methods and the performance of the distributed BMS battery system. Specifically, for series connection, through causal reasoning, it is found that the series connection increases the total voltage, but any single cell failure will affect the normal operation of the entire battery pack, so a stronger monitoring and protection mechanism is needed. For parallel connection, the parallel connection increases redundancy, but there may be uneven current distribution, causing some single cells to overheat or overcharge. Causal reasoning not only reveals the direct impact of different connection methods on system performance, but also discovers potential indirect impacts. For example, although a certain connection method increases the total voltage, it may cause the temperature of some single cells to rise abnormally, thereby affecting the battery life. These causal relationships provide an important basis for optimizing battery pack design.

[0099] Step 224: introducing a generative adversarial network to simulate extreme working conditions, so as to obtain the performance of different connection modes under extreme conditions and obtain the performance results of extreme working conditions;

[0100] In order to evaluate the performance of different connection methods under extreme conditions, this step introduces a generative adversarial network (GAN). Specifically, the generator generates simulated extreme operating condition data, such as high temperature environment, high load operation, etc. The discriminator determines whether the generated data is true and ensures that the generated extreme operating condition data is consistent with the actual situation. By simulating extreme conditions through GAN, the system can test the performance of different connection methods under these extreme conditions. For example, in a high temperature environment, a certain connection method may cause the temperature of some single cells to be too high, affecting the battery life; while under high load operation, another connection method may show better energy transfer efficiency and stability. Finally, the performance results of extreme conditions are obtained, which provide an important reference for the comprehensive evaluation of different connection methods.

[0101] Step 225: Based on the condition dependency, the causal relationship, and the extreme operating condition performance results, comprehensively evaluate the impact of different connection modes on the performance of the distributed BMS battery system to obtain a connection analysis result;

[0102] This step conducts a comprehensive evaluation based on the conditional dependencies, causal relationships, and extreme operating performance results obtained in the previous steps. Specifically, the conditional dependencies reveal the interactions between different components and their properties. The causal relationships clarify the direct and indirect effects of different connection methods on system performance. The extreme operating performance results show the actual performance of different connection methods under extreme conditions. Through comprehensive evaluation, the system obtains detailed connection analysis results. For example, in high-voltage applications, although the series connection increases the total voltage, a stronger monitoring and protection mechanism is required to prevent single cell failures from affecting the entire battery pack. In low-voltage and high-capacity applications, hybrid connection methods (partial series, partial parallel) can provide higher redundancy and better current distribution while ensuring a certain voltage, reducing local hot spots and energy losses. These connection analysis results provide a scientific basis for optimizing battery pack design, helping to select the connection method that best suits specific application scenarios and ensure efficient operation and safety of the system.

[0103] The embodiments of the present invention can achieve the following beneficial effects through the above steps:

[0104] Combining graph theory and graph neural network, an accurate battery pack topology model is constructed to deeply analyze the impact of different connection methods on system performance, providing a scientific basis for optimal design.

[0105] Using Bayesian networks and causal reasoning techniques, the causal relationship between different connection modes and system performance is revealed, helping to understand the interaction between components and their impact on overall performance;

[0106] Generative adversarial networks are introduced to simulate extreme working conditions, evaluate the performance of different connection methods under extreme conditions, and ensure the stability and reliability of the system in various complex environments;

[0107] Based on conditional dependencies, causal relationships, and extreme operating conditions, a comprehensive evaluation is conducted to provide a scientific basis for selecting the optimal connection method and improve the overall performance and safety of the system.

[0108] In a distributed BMS battery management system, precise regulation of the energy flow between individual cells is crucial to maintaining system stability and improving overall performance. Traditional regulation methods often have difficulty coping with complex dynamic changes and nonlinear characteristics. In order to achieve more intelligent and efficient energy management and ensure efficient operation and excellent performance of the system. Based on this, the present invention provides a specific embodiment, step 103, based on the optimal power balancing strategy, a fuzzy logic controller is used to optimize the regulation of energy flow between individual cells to obtain an optimized regulation result, based on the optimized regulation result, an energy flow management plan is determined, and the sliding mode variable structure control theory and the optimized energy transfer strategy are introduced in the regulation process for optimization, which specifically includes the following steps:

[0109] Step 311: using the optimal power balancing strategy to regulate the energy flow between the individual cells to obtain a preliminary regulation result, and using a fuzzy logic controller to dynamically adjust the preliminary regulation result to obtain an optimized regulation result;

[0110] This step assumes that there is a battery pack consisting of 48 lithium-ion cells. The system has obtained the optimal power balancing strategy through a multi-objective optimization algorithm. The strategy specifies the charge / discharge rate of each cell and the use of the bypass circuit to ensure uniform power distribution throughout the battery pack. Next, the system performs initial regulation of the energy flow between the cells based on this optimal power balancing strategy. For example, when the voltage of a cell is too high, the system reduces its charging current through the bypass circuit; and when the voltage of another cell is too low, it increases its charging current. After the initial regulation, the system obtains preliminary regulation results. In order to further optimize the regulation effect, the system introduces a fuzzy logic controller (FLC). The FLC can dynamically adjust the regulation strategy based on the real-time monitored cell status (such as voltage, temperature, etc.). For example: if the voltage of a cell is close to the upper threshold, the FLC will reduce its charging current more significantly. If the temperature of a cell rises abnormally, the FLC will temporarily stop charging it until the temperature returns to normal. Through the dynamic adjustment of the FLC, the system finally obtains the optimized regulation result, ensuring a smoother and more efficient energy flow.

[0111] Step 312: According to the optimization and control result, a target energy transfer path and a target energy transfer mode are selected by using a multi-agent reinforcement learning method, and the target energy transfer path and the target energy transfer mode are optimized to obtain an optimal energy transfer path and an optimal target energy transfer mode, and an energy flow management plan is determined based on the optimal energy transfer path and the optimal target energy transfer mode;

[0112] This step is based on the optimization and control results in the first step. The system uses a multi-agent reinforcement learning method to select the target energy transfer path and target energy transfer method. Specifically, the system creates multiple agents, each representing a single cell. These agents work together to optimize the energy transfer path and method through trial and error learning. For example, the agents try different energy transfer paths and methods and evaluate the effect of each solution. If a certain path causes some single cells to have too high temperatures or too much energy loss, the agent will record it and adjust the strategy. After multiple iterations of training, the system finally found the optimal energy transfer path and the optimal target energy transfer method. Based on these optimal results, the system determines the energy flow management plan. For example, the plan may stipulate that under certain operating conditions, a bypass circuit should be used for energy transfer to reduce the load on the main path; or in a high temperature environment, energy transfer between certain single cells that are prone to heat should be avoided as much as possible to reduce risks.

[0113] Step 313: Introducing an event trigger mechanism in the control process, optimizing the existing energy transfer strategy, generating an optimized energy transfer strategy, so that the fuzzy logic controller performs corresponding control when a specific event occurs, and optimizing the control of energy flow between each single battery based on the optimized energy transfer strategy and sliding mode variable structure control theory, wherein the specific event includes that the voltage or temperature of the single battery is higher than a preset threshold;

[0114] In order to further improve the response speed and adaptability of the system, the system introduces an event trigger mechanism in the regulation process. Specifically, when the voltage or temperature of a single cell is higher than the preset threshold, the system will trigger the corresponding regulation measures. For example, if the temperature of a single cell exceeds the safety threshold (such as 45°C), the event trigger mechanism will start immediately and notify the fuzzy logic controller (FLC) to take emergency measures. According to the optimized energy transfer strategy, the FLC quickly adjusts the charging and discharging behavior of the single cell to prevent the temperature from continuing to rise. For example, the FLC may temporarily stop charging the single cell or transfer excess energy to other single cells through a bypass circuit. In order to ensure the stability and robustness of the regulation process, the system also introduces the sliding mode variable structure control theory. Sliding mode control can respond quickly and maintain the stability of the system when external conditions change. For example, when adjusting the charging current of a single cell, sliding mode control can ensure that the system can quickly return to the ideal state even when the ambient temperature fluctuates. Finally, the system optimizes the regulation of energy flow between each single cell by combining the optimized energy transfer strategy and sliding mode variable structure control theory, ensuring the efficient operation and safety of the system.

[0115] The embodiments of the present invention can achieve the following beneficial effects through the above steps:

[0116] Through the optimal power balancing strategy and dynamic adjustment of the fuzzy logic controller, accurate regulation and optimization of energy flow are ensured, and the flexibility and response speed of the system are improved;

[0117] The multi-agent reinforcement learning method is used to select the optimal energy transfer path and method, which significantly improves the efficiency and reliability of energy transfer;

[0118] The introduction of event trigger mechanism enables the system to respond quickly when a specific event occurs, ensuring the safe operation of single batteries and reducing potential risks;

[0119] Combined with sliding mode variable structure control theory, the stability and robustness of the regulation process are ensured and the overall performance of the system is enhanced.

[0120] In a distributed BMS battery management system, selecting and optimizing energy transfer paths and methods is the key to achieving efficient energy management. Traditional control methods often have difficulty coping with complex dynamic changes and changing working environments. In order to improve the flexibility and response speed of the system, ensure that the system can achieve efficient and stable energy management under various working conditions. Based on this, the present invention provides a specific embodiment, step 312, according to the optimization control results, using a multi-agent reinforcement learning method to select a target energy transfer path and a target energy transfer method, optimize the target energy transfer path and the target energy transfer method, obtain the optimal energy transfer path and the optimal target energy transfer method, and determine the energy flow management plan based on the optimal energy transfer path and the optimal target energy transfer method, which specifically includes the following steps:

[0121] Step 321: Based on the optimization control result, a multi-agent reinforcement learning method is used to simulate energy transfer effects under different energy transfer paths, and based on the energy transfer effects, a target energy transfer path is selected, wherein the energy transfer path includes a multi-level transfer path and a bus-based centralized path, and the energy transfer effects include transfer efficiency and thermal management performance;

[0122] This step assumes that there is a battery pack consisting of 48 single lithium batteries, and the system has obtained the optimization control results through the fuzzy logic controller. Next, in order to further optimize the energy transfer path, the system uses the multi-agent reinforcement learning method (MARL) to simulate the energy transfer effect under different energy transfer paths. For example, the system creates multiple agents, each representing a single battery. These agents work together to try different energy transfer paths and evaluate the effect of each path. Specifically, multi-level transfer path: energy is gradually transferred through multiple intermediate nodes. For example, energy is transferred from battery A to battery B, and then from battery B to battery C. The advantage of this path is high flexibility, but it may increase energy loss and heat accumulation. Bus-based centralized path: energy is transferred through a centralized bus. For example, all single batteries that need to be charged are connected to the same bus and are managed uniformly by a central controller. The advantages of this path are simple structure and easy control, but it may cause some nodes to overload or heat. The system simulates the energy transfer effect under these two paths through MARL, mainly evaluating two aspects: Transfer efficiency: how much energy is lost during the transfer process, the higher the transfer efficiency, the better. Thermal management performance: Whether the heat generated during energy transfer can be effectively managed to avoid overheating problems. For example, after multiple simulations, the system found that although the multi-level transfer path is flexible, it can cause local hot spots in some cases and affect battery life; and the centralized path based on the bus is simple in structure, but it is prone to overload problems under high load. In the end, the system selected the centralized path based on the bus as the target energy transfer path because it provides better overall performance in most cases.

[0123] Step 322: Based on the optimization control result, using a multi-agent reinforcement learning method, simulating the performance effects under different energy transfer modes, and selecting a target energy transfer mode based on the performance effect, wherein the energy transfer mode includes wireless energy transfer and transfer through an intermediate energy storage device;

[0124] Based on the optimization and control results in the first step, the system continues to use the multi-agent reinforcement learning method (MARL) to simulate the performance of different energy transfer methods to select the best energy transfer method. For example, the system simulates the following two energy transfer methods: Wireless energy transfer: does not rely on physical contact, and transfers energy through electromagnetic waves or other wireless methods. The advantage of this method is that it reduces the risk of wear and failure caused by physical connections, but it may be affected by environmental factors such as electromagnetic interference or distance limitations. Transfer through an intermediate energy storage device: Energy is transferred through an intermediate energy storage device (such as a supercapacitor). The advantage of this method is that it can buffer energy fluctuations and improve the stability and response speed of the system, but it increases the complexity and cost of the system. The system simulates the performance of these two methods through MARL, mainly evaluating the following aspects: Transfer efficiency: how much energy is lost during the transfer process, the higher the transfer efficiency, the better. Response speed: the speed of energy transfer, the faster the response speed, the better. Stability: stability during energy transfer, avoiding voltage or current fluctuations. For example, after many simulations, the system found that wireless energy transfer performs well over short distances, but may be unstable over long distances or in complex environments; while transfer through an intermediate energy storage device can maintain high transfer efficiency and stability under various conditions. Ultimately, the system chose transfer through an intermediate energy storage device as the target energy transfer method because it provides better overall performance in most cases.

[0125] Step 323: Introduce a mixed integer linear programming method to perform constrained optimization on the target energy transfer path and the target energy transfer mode to obtain an optimal energy transfer path and an optimal energy transfer mode, and determine an energy flow management plan based on the optimal energy transfer path and the optimal target energy transfer mode;

[0126] In order to ensure that the selected target energy transfer path and target energy transfer mode achieve the best effect in practical applications, this step introduces the mixed integer linear programming method (MILP) for constrained optimization. Specifically, the system defines a series of constraints, such as maximum allowable current, upper temperature limit, energy loss limit, etc., to ensure the safety and efficiency of energy transfer. For example, the system sets the following constraints: Maximum allowable current: The maximum charge / discharge current of each single battery cannot exceed its rated value. Upper temperature limit: The temperature of any single battery cannot exceed the safety threshold (such as 45°C). Energy loss limit: The total energy loss must be lower than a preset value to ensure the overall efficiency of the system. Through MILP, the system constrains and optimizes the target energy transfer path and target energy transfer mode, and finally obtains the optimal energy transfer path and optimal energy transfer mode. For example, the system finds that under certain working conditions, the use of a centralized path based on the bus and the transmission through an intermediate energy storage device can provide the highest transmission efficiency and the best thermal management performance while meeting all constraints. Based on these optimal results, the system determines the energy flow management solution. For example, the plan stipulates that under normal operating conditions, energy transfer should be carried out through a centralized bus-based path, and buffered through an intermediate energy storage device; under extreme conditions (such as high temperature or high load), the system will automatically switch to an alternative path or method to ensure safety and stability.

[0127] The embodiments of the present invention can achieve the following beneficial effects through the above steps:

[0128] By simulating the effects of different energy transfer paths through multi-agent reinforcement learning methods, the optimal energy transfer path was selected, significantly improving the efficiency of energy transfer and thermal management performance;

[0129] By simulating the performance of different energy transfer methods through multi-agent reinforcement learning methods, the optimal energy transfer method was selected to ensure the high efficiency and stability of energy transfer;

[0130] The mixed integer linear programming method is introduced to perform constrained optimization on the energy transfer path and mode, ensuring the safety and efficiency of energy transfer and avoiding potential risks.

[0131] Based on the optimal energy transfer path and method, a detailed energy flow management plan was determined, providing a scientific basis for the efficient operation of the system.

[0132] In a distributed BMS battery management system, ensuring the safety and stability of the system not only depends on efficient energy management, but also requires a powerful anomaly detection and risk management mechanism. In order to achieve comprehensive security protection, based on this, the present invention provides a specific embodiment, step 104, applying the energy flow management solution, obtaining feedback data, based on the feedback data and historical data, using a deep belief network to build an anomaly detection model, using the anomaly detection model to identify abnormal operation modes, analyzing the abnormal operation modes, obtaining abnormal analysis results, generating a risk assessment report based on the abnormal analysis results, building a security record platform based on blockchain technology, and obtaining security record information based on the security record platform, specifically including the following steps:

[0133] Step 401: using the energy flow management scheme to regulate the energy flow between the single cells, and obtaining feedback data, wherein the feedback data includes the voltage, temperature and internal resistance of each single cell after regulation;

[0134] This step assumes that there is a battery pack consisting of 48 lithium-ion cells. The system has determined the energy flow management scheme by optimizing the control results and multi-agent reinforcement learning methods. The scheme specifies the charge / discharge rate of each cell and the use of the bypass circuit to ensure that the power of the entire battery pack is evenly distributed. Next, the system regulates the energy flow between the cells according to this energy flow management scheme. For example, when the voltage of a cell is too high, the system reduces its charging current through the bypass circuit; when the voltage of another cell is too low, it increases its charging current. After regulation, the system collects the voltage, temperature and internal resistance of each cell in real time as feedback data. For example, the voltage of a cell after regulation is 3.7V, the temperature is 25℃, and the internal resistance is 0.5Ω. These feedback data are transmitted to the central control system and stored in a database for subsequent analysis and adjustment.

[0135] Step 402: Based on the feedback data and historical data, an anomaly detection model is constructed using a deep belief network, and the operating state of the distributed BMS battery system is monitored in real time using the anomaly detection model to identify abnormal operation modes, wherein the abnormal operation modes include an overcharge mode, an over-discharge mode, and an abnormal temperature mode;

[0136] This step builds an anomaly detection model using a deep belief network (DBN) based on the feedback data and historical data in the first step. DBN can automatically identify normal and abnormal operating modes by learning from a large amount of data. For example, the system collects a large amount of historical data, including voltage, temperature, and internal resistance changes during normal operation, as well as abnormal events that occurred in the past. By training DBN, the system learns to distinguish between normal and abnormal operating modes. Specifically: Overcharge mode: When the voltage of a single cell is continuously higher than the safety threshold (such as 4.2V), it may indicate overcharge. Overdischarge mode: When the voltage of a single cell is continuously lower than the safety threshold (such as 2.5V), it may indicate overdischarge. Temperature abnormality mode: When the temperature of a single cell is continuously higher than the safety threshold (such as 45℃), it may indicate temperature abnormality. Use the constructed anomaly detection model to monitor the operating status of the distributed BMS battery system in real time, and identify and warn abnormal operating modes in a timely manner. For example, if the voltage of a single cell suddenly rises to 4.3V, the anomaly detection model will immediately issue an alarm, indicating that there may be a risk of overcharging.

[0137] Step 403: Analyze the abnormal operation mode, determine the abnormal cause and the abnormal impact range, and obtain an abnormal analysis result based on the abnormal cause and the abnormal impact range;

[0138] Once abnormal operation modes are identified in this step, the system will conduct in-depth analysis of these anomalies to determine the cause of the anomaly and the scope of the abnormal impact. For example: Overcharge mode: If the voltage of a single cell is continuously higher than the safety threshold, the system will check whether the charging circuit is faulty or whether there are external factors (such as excessive ambient temperature) that cause the charging current to be too large. Overdischarge mode: If the voltage of a single cell is continuously lower than the safety threshold, the system will check whether the discharge load is too large or whether there is uneven energy transfer from other single cells. Temperature abnormality mode: If the temperature of a single cell is continuously higher than the safety threshold, the system will check whether the cooling system is working properly or whether there are local hot spots. Through this analysis, the system can determine the specific cause of the abnormality and the scope of impact. For example, the system found that the overcharge of a single cell was caused by a sensor failure in the charging circuit, and the scope of impact was limited to the single cell and several adjacent cells. Finally, the system obtained detailed abnormality analysis results, indicating the specific cause and scope of impact of the abnormality, providing a basis for subsequent processing.

[0139] Step 404: Based on the abnormal analysis result, the risk level of the distributed BMS battery system is evaluated, and a risk assessment report is generated, wherein the risk assessment report includes the risk type and the probability of risk occurrence;

[0140] This step is based on the abnormal analysis results in the third step. The system evaluates the risk level of the distributed BMS battery system and generates a risk assessment report. Specifically: Risk type: The report lists all identified risk types, such as overcharge risk, over-discharge risk, temperature abnormality risk, etc. Probability of risk occurrence: For each risk type, the system calculates its probability of occurrence. For example, the probability of overcharge risk of a certain single cell is 5%, while the probability of temperature abnormality risk is 3%. The system generates a detailed risk assessment report by comprehensively evaluating the severity and probability of occurrence of various risk types. For example, the report points out that the current system has a high risk of overcharging, and it is recommended to take immediate measures to repair the sensor failure in the charging circuit; at the same time, the report also reminds you to pay attention to the risk of temperature abnormality and recommends strengthening the maintenance of the cooling system.

[0141] Step 405: construct a security record platform using blockchain technology, and extract operation logs, abnormal events, and risk assessment reports based on the security record platform to obtain security record information;

[0142] In this step, in order to ensure the security and transparency of all operation logs, abnormal events and risk assessment reports, the system uses blockchain technology to build a secure recording platform. Blockchain technology is characterized by decentralization, immutability and high transparency, which is very suitable for recording and managing key data. For example, the system records the logs of each control operation, the identified abnormal events and the generated risk assessment reports on the blockchain. Each record has a unique hash value to ensure the authenticity and integrity of the data. Users can view all historical records through the secure recording platform to ensure that all operations are traceable and immutable. In this way, the system obtains detailed security record information, which helps managers fully understand the operating status and potential risks of the system and take preventive measures in time. For example, managers can regularly review operation logs to ensure that there are no improper operations; check abnormal event records to respond quickly to potential problems; review risk assessment reports and formulate corresponding risk management strategies.

[0143] The embodiments of the present invention can achieve the following beneficial effects through the above steps:

[0144] Through precise regulation of energy flow management solutions, real-time feedback data collection ensures accurate monitoring of system operation status;

[0145] Using deep belief networks to build anomaly detection models, real-time monitoring of system operation status, timely identification of abnormal operation modes, and enhanced system security;

[0146] Through in-depth analysis of abnormal operation modes, the causes and impact range of abnormalities are determined, providing a scientific basis for subsequent processing;

[0147] Based on the abnormal analysis results, the system risk level is evaluated and a risk assessment report is generated to help decision makers understand the current risk status and formulate effective preventive measures;

[0148] The use of blockchain technology to build a secure recording platform ensures the security and transparency of all operation logs, abnormal events and risk assessment reports, enhancing the credibility and traceability of the system.

[0149] In a distributed BMS battery management system, accurately assessing the system risk level is a key step to ensure safe operation and prevent potential failures. In order to provide comprehensive risk management, based on this, the present invention provides a specific embodiment, step 404, based on the abnormal analysis results, assessing the risk level of the distributed BMS battery system, generating a risk assessment report, the risk assessment report includes the risk type and the probability of risk occurrence, specifically including the following steps:

[0150] Step 411: Based on the abnormal analysis result, the risk level of the distributed BMS battery system is evaluated. Based on the risk level, a Monte Carlo simulation method is used to simulate risk scenarios and the probability of occurrence of the risk scenarios. The risk scenarios include overcharge risk scenarios, over-discharge risk scenarios, and temperature abnormality risk scenarios.

[0151] This step assumes that there is a battery pack consisting of 48 lithium-ion cells. The system has identified and analyzed multiple abnormal operation modes through the anomaly detection model and obtained detailed anomaly analysis results. For example, it is found that a certain cell has an overcharge risk, which may be caused by a sensor failure in the charging circuit; another cell has a temperature anomaly risk, which may be caused by a problem with the cooling system. Based on these anomaly analysis results, the system evaluates the risk level of the distributed BMS battery system. For example, the overcharge risk is evaluated as a medium risk (risk level 3) and the temperature anomaly risk is evaluated as a high risk (risk level 5). The higher the risk level, the greater the potential threat to the system. In order to further understand the actual impact of these risks, the system uses the Monte Carlo simulation method to simulate different risk scenarios and their occurrence probabilities. Specifically: Overcharge risk scenario: simulates the possibility of the cell voltage exceeding the safety threshold under different circumstances. For example, when the charging current is too large, the probability of overcharge risk is 5%. Over-discharge risk scenario: simulates the possibility of the cell voltage being lower than the safety threshold under different circumstances. For example, when the discharge load is too large, the probability of over-discharge risk is 3%. Temperature anomaly risk scenario: simulates the possibility of the cell temperature exceeding the normal range under different circumstances. For example, when the ambient temperature is too high, the probability of abnormal temperature risk is 7%. Through multiple simulations, the system obtains the probability of occurrence of each risk scenario and establishes a probability distribution model for the risk scenario. This helps to more accurately assess the impact of different risks on the system.

[0152] Step 412: using the Markov chain method, defining the transition probability of the distributed BMS battery system between different risk states, and predicting the risk evolution trend of the distributed BMS battery system, and obtaining the risk propagation analysis result based on the transition probability and the risk evolution trend, wherein the risk state includes a normal state, an abnormal state, and a fault state;

[0153] In order to better understand and predict the risk evolution trend of the distributed BMS battery system, the system uses the Markov chain method to define the transition probability between different risk states. Specifically, the system divides the risk state into three categories: Normal State: The system operates well and there is no obvious sign of risk. Abnormal State: The system has some abnormal conditions, but has not yet reached the fault state. Failure State: The system fails and cannot work properly. Next, the system calculates the transition probability between different risk states based on historical data and current risk levels. For example: The transition probability from normal state to abnormal state is 0.05. The transition probability from abnormal state to fault state is 0.03. The transition probability of returning from abnormal state to normal state is 0.92. The transition probability of recovering from fault state to normal state is 0.05. By establishing a Markov chain model, the system can predict the risk evolution trend of the distributed BMS battery system. For example, if the current system is in an abnormal state, according to the transition probability, the probability of entering the fault state in the future is 0.03, and the probability of returning to the normal state is 0.92. This prediction helps to take preventive measures in advance to avoid the system from entering a fault state. Finally, the system obtains detailed risk propagation analysis results, points out the transition paths and probabilities between different risk states, and helps managers better understand the risk evolution trend of the system.

[0154] Step 413: Generate a risk assessment report based on the risk level, the risk scenario, the probability of occurrence of the risk scenario, and the risk propagation analysis result;

[0155] This step generates a comprehensive risk assessment report based on the risk levels, risk scenarios and their occurrence probabilities in the first two steps, as well as the results of risk propagation analysis. Specifically, the report contains the following parts: Risk Type and Risk Level: List all identified risk types and their corresponding risk levels. For example, overcharge risk (risk level 3), over-discharge risk (risk level 2), and temperature abnormality risk (risk level 5). Risk Scenario and Occurrence Probability: Describe each risk scenario and its probability of occurrence in detail. For example, the probability of occurrence of the overcharge risk scenario is 5%, the probability of occurrence of the over-discharge risk scenario is 3%, and the probability of occurrence of the temperature abnormality risk scenario is 7%. Risk Propagation Analysis Results: Summarize the transition probabilities and risk evolution trends between different risk states. For example, the current system is in an abnormal state, and the probability of entering a fault state in the future is 0.03, while the probability of returning to a normal state is 0.92. Recommended Measures: Based on the above analysis results, specific prevention and response measures are proposed. For example, it is recommended to immediately repair the sensor fault in the charging circuit, strengthen the maintenance of the cooling system, and regularly review the system operation log. This risk assessment report not only helps managers fully understand the risk status of the system, but also provides a scientific basis to guide them in formulating effective risk management strategies to ensure the safe and stable operation of the system.

[0156] The embodiments of the present invention can achieve the following beneficial effects through the above steps:

[0157] Based on the abnormal analysis results, the Monte Carlo simulation method is used to simulate risk scenarios and their probability of occurrence to ensure the accuracy of risk assessment;

[0158] The Markov chain method is used to define the transition probability between different risk states, predict the evolution trend of system risk, and enhance the predictability and controllability of the system;

[0159] Generate a detailed risk assessment report based on the risk level, risk scenario and its probability of occurrence, and risk propagation analysis results to help managers fully understand the risk status of the system;

[0160] Provide scientific basis to guide managers to formulate effective risk management strategies and ensure the safe and stable operation of the system.

[0161] Figure 2 A structural diagram of a distributed BMS battery active balancing safety management system is provided for an embodiment of the present invention, such as Figure 2 As shown, the system includes:

[0162] The monitoring module 21 is used to monitor the voltage, temperature and internal resistance of each single cell in real time, obtain battery data, perform status evaluation and trend prediction on the battery data, and obtain status information and performance change trend corresponding to each single cell;

[0163] A construction module 22 is used to construct a target battery pack topology model based on the state information and the performance change trend, generate an intelligent agent model based on key information of the internal structure of the battery pack in the target battery pack topology model and a reinforcement learning algorithm, select a preliminary power balancing strategy using the intelligent agent model, optimize the preliminary power balancing strategy, and obtain an optimal power balancing strategy;

[0164] The determination module 23 is used to optimize the regulation of energy flow between each single battery using a fuzzy logic controller based on the optimal power balancing strategy to obtain an optimized regulation result, determine an energy flow management plan based on the optimized regulation result, and introduce a sliding mode variable structure control theory and an optimized energy transfer strategy in the regulation process for optimization;

[0165] an identification module 24, for applying the energy flow management solution to obtain feedback data, constructing an anomaly detection model using a deep belief network based on the feedback data and historical data, identifying abnormal operation modes using the anomaly detection model, analyzing the abnormal operation modes, obtaining an abnormal analysis result, generating a risk assessment report based on the abnormal analysis result, constructing a security record platform based on blockchain technology, and obtaining security record information based on the security record platform;

[0166] The generation module 25 is used to optimize the parameters of each model based on the risk assessment report and the safety record information to obtain an optimized model, and to generate a safety management system by combining the optimized model, the fuzzy logic controller, the safety record platform, the transfer learning algorithm and the quantum computing method. The various models include the battery pack topology model, the intelligent agent model and the anomaly detection model.

[0167] Figure 2 The distributed BMS battery active balancing safety management system can execute Figure 1 The implementation principle and technical effect of the distributed BMS battery active balancing safety management method described in the embodiment are not repeated here. The specific way in which each module and unit performs operations in the distributed BMS battery active balancing safety management system in the above embodiment has been described in detail in the embodiment of the method, and will not be elaborated here.

[0168] In one possible design, Figure 2 A distributed BMS battery active balancing safety management system of the embodiment shown can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0169] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0170] The processing component 32 is used to: monitor the voltage, temperature and internal resistance of each single battery in real time to obtain battery data, perform state evaluation and trend prediction on the battery data, and obtain the state information and performance change trend corresponding to each single battery; construct a target battery pack topology structure model based on the state information and the performance change trend, generate an intelligent agent model based on the key information of the internal structure of the battery pack in the target battery pack topology structure model and the reinforcement learning algorithm, use the intelligent agent model to select a preliminary power balancing strategy, optimize the preliminary power balancing strategy, and obtain an optimal power balancing strategy; based on the optimal power balancing strategy, use a fuzzy logic controller to optimize the regulation of energy flow between each single battery to obtain an optimized regulation result, determine an energy flow management plan based on the optimized regulation result, and introduce sliding mode variable structure control theory in the regulation process and the optimized energy transfer strategy; applying the energy flow management scheme to obtain feedback data, based on the feedback data and historical data, using a deep belief network to build an anomaly detection model, using the anomaly detection model to identify abnormal operation modes, analyzing the abnormal operation modes, obtaining an abnormal analysis result, generating a risk assessment report based on the abnormal analysis result, building a security recording platform based on blockchain technology, and obtaining security recording information based on the security recording platform; based on the risk assessment report and the security recording information, optimizing the parameters of each model to obtain an optimized model, combining the optimized model, the fuzzy logic controller, the security recording platform, the transfer learning algorithm and the quantum computing method to generate a security management system, wherein each model includes the battery pack topology model, the intelligent agent model and the anomaly detection model.

[0171] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0172] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0173] Of course, the computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0174] The input / output interface provides an interface between the processing component and the peripheral interface module, which may be an output device, an input device, etc.

[0175] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0176] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0177] The embodiment of the present invention further provides a computer storage medium storing a computer program, which can achieve the above-mentioned Figure 1 A distributed BMS battery active balancing safety management method according to the illustrated embodiment.

[0178] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0179] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0180] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0181] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A distributed BMS battery active balancing safety management method, characterized in that: include: Real-time monitoring of the voltage, temperature and internal resistance of each single battery to obtain battery data, and state evaluation and trend prediction of the battery data to obtain the state information and performance change trend corresponding to each single battery; Based on the state information and the performance change trend, a target battery pack topology model is constructed, and based on the key information of the internal structure of the battery pack in the target battery pack topology model and a reinforcement learning algorithm, an intelligent agent model is generated, and a preliminary power balancing strategy is selected by using the intelligent agent model, and the preliminary power balancing strategy is optimized to obtain an optimal power balancing strategy; Based on the optimal power balancing strategy, a fuzzy logic controller is used to optimize the regulation of energy flow between each single battery to obtain an optimized regulation result. Based on the optimized regulation result, an energy flow management plan is determined, and a sliding mode variable structure control theory and an optimized energy transfer strategy are introduced in the regulation process for optimization; Apply the energy flow management solution to obtain feedback data, build an anomaly detection model using a deep belief network based on the feedback data and historical data, identify abnormal operation modes using the anomaly detection model, analyze the abnormal operation modes, obtain an abnormal analysis result, generate a risk assessment report based on the abnormal analysis result, build a security record platform based on blockchain technology, and obtain security record information based on the security record platform; Based on the risk assessment report and the safety record information, the parameters of each model are optimized to obtain an optimized model. The optimized model, the fuzzy logic controller, the safety record platform, the transfer learning algorithm and the quantum computing method are combined to generate a safety management system. The models include the battery pack topology model, the intelligent agent model and the anomaly detection model.

2. The method according to claim 1, characterized in that Based on the state information and the performance change trend, a target battery pack topology model is constructed; based on key information of the internal structure of the battery pack in the target battery pack topology model and a reinforcement learning algorithm, an intelligent agent model is generated; a preliminary power balancing strategy is selected using the intelligent agent model; the preliminary power balancing strategy is optimized to obtain an optimal power balancing strategy, including: Based on the state information and the performance change trend, the internal connection relationship and energy transfer path of the battery pack are modeled and processed in combination with a graph theory method to construct a target battery pack topology structure model, and the target battery pack topology structure model is used to analyze the impact of different connection modes on the performance of the distributed BMS battery system to obtain a connection analysis result, and based on the connection analysis result, key information of the internal structure of the battery pack is determined; Based on the key information, combined with the reinforcement learning algorithm and the simulated annealing algorithm, the initial model is trained in a simulation environment to obtain an intelligent agent model, the intelligent agent model and the particle swarm optimization algorithm are used to select a preliminary power balancing strategy, and the ant colony optimization algorithm is introduced to optimize the preliminary power balancing strategy to generate a preliminary optimized power balancing strategy; A multi-objective optimization algorithm is applied to comprehensively optimize the power balancing strategy after preliminary optimization to obtain a global optimal solution. During the comprehensive optimization process, a Bayesian optimization algorithm is used to adjust the hyperparameters of the multi-objective optimization algorithm to enhance the optimization. The global optimal solution represents the best power balancing strategy.

3. The method according to claim 2, characterized in that Based on the state information and the performance change trend, the internal connection relationship and energy transfer path of the battery pack are modeled and processed in combination with the graph theory method to construct a target battery pack topology structure model. The target battery pack topology structure model is used to analyze the impact of different connection modes on the performance of the distributed BMS battery system to obtain connection analysis results, including: Based on the state information and the performance change trend, the internal connection relationship and energy transfer path of the battery pack are modeled and processed in combination with graph theory methods and graph neural networks to construct a preliminary battery pack topology model, and the target battery pack topology model is constructed by combining the preliminary battery pack topology model and the Bayesian network; Analyzing conditional dependencies between variables using a Bayesian network in the target battery pack topology model, wherein the variables include components of a distributed BMS battery system, properties of the components, and interactions between the components; Use causal reasoning technology to analyze the causal relationship between different connection modes and distributed BMS battery system performance; Generative adversarial networks are introduced to simulate extreme working conditions to obtain the performance of different connection methods under extreme conditions and obtain the performance results of extreme working conditions; Based on the conditional dependency, the causal relationship and the extreme operating condition performance results, the impact of different connection modes on the performance of the distributed BMS battery system is comprehensively evaluated to obtain the connection analysis results.

4. The method according to claim 1, characterized in that: Based on the optimal power balancing strategy, a fuzzy logic controller is used to optimize the regulation of energy flow between each single battery to obtain an optimized regulation result. Based on the optimized regulation result, an energy flow management plan is determined. In the regulation process, a sliding mode variable structure control theory and an optimized energy transfer strategy are introduced for optimization, including: Using the optimal power balancing strategy, the energy flow between each single battery is regulated to obtain a preliminary regulation result, and a fuzzy logic controller is used to dynamically adjust the preliminary regulation result to obtain an optimized regulation result; According to the optimization and control results, a target energy transfer path and a target energy transfer mode are selected by using a multi-agent reinforcement learning method, the target energy transfer path and the target energy transfer mode are optimized to obtain an optimal energy transfer path and an optimal target energy transfer mode, and an energy flow management plan is determined based on the optimal energy transfer path and the optimal target energy transfer mode; An event trigger mechanism is introduced into the regulation process, and the existing energy transfer strategy is optimized to generate an optimized energy transfer strategy so that the fuzzy logic controller can perform corresponding regulation when a specific event occurs. Based on the optimized energy transfer strategy and sliding mode variable structure control theory, the regulation of energy flow between each single cell is optimized. The specific event includes that the voltage or temperature of the single cell is higher than a preset threshold.

5. The method according to claim 4, characterized in that According to the optimization and control results, a target energy transfer path and a target energy transfer mode are selected by using a multi-agent reinforcement learning method, and the target energy transfer path and the target energy transfer mode are optimized to obtain an optimal energy transfer path and an optimal target energy transfer mode. Based on the optimal energy transfer path and the optimal target energy transfer mode, an energy flow management plan is determined, including: Based on the optimization and control results, a multi-agent reinforcement learning method is used to simulate the energy transfer effects under different energy transfer paths, and based on the energy transfer effects, a target energy transfer path is selected, wherein the energy transfer path includes a multi-level transfer path and a centralized path based on a bus, and the energy transfer effects include transfer efficiency and thermal management performance; Based on the optimization and control results, a multi-agent reinforcement learning method is used to simulate the performance effects under different energy transfer modes, and based on the performance effects, a target energy transfer mode is selected, where the energy transfer mode includes wireless energy transfer and transfer through an intermediate energy storage device; A mixed integer linear programming method is introduced to perform constrained optimization on the target energy transfer path and the target energy transfer mode to obtain the optimal energy transfer path and the optimal energy transfer mode. Based on the optimal energy transfer path and the optimal target energy transfer mode, an energy flow management plan is determined.

6. The method according to claim 1, characterized in that Apply the energy flow management solution to obtain feedback data, build an anomaly detection model based on the feedback data and historical data using a deep belief network, identify abnormal operation modes using the anomaly detection model, analyze the abnormal operation modes, obtain an abnormal analysis result, generate a risk assessment report based on the abnormal analysis result, build a security record platform based on blockchain technology, and obtain security record information based on the security record platform, including: Using the energy flow management scheme, the energy flow between the single cells is regulated to obtain feedback data, wherein the feedback data includes the voltage, temperature and internal resistance of each single cell after regulation; Based on the feedback data and historical data, an anomaly detection model is constructed using a deep belief network, and the operating state of the distributed BMS battery system is monitored in real time using the anomaly detection model to identify abnormal operation modes, wherein the abnormal operation modes include an overcharge mode, an over-discharge mode, and an abnormal temperature mode; Analyze the abnormal operation mode, determine the abnormal cause and the abnormal impact range, and obtain an abnormal analysis result based on the abnormal cause and the abnormal impact range; Based on the abnormal analysis results, assess the risk level of the distributed BMS battery system and generate a risk assessment report, wherein the risk assessment report includes the risk type and the probability of risk occurrence; A security record platform is constructed using blockchain technology. Based on the security record platform, operation logs, abnormal events and risk assessment reports are extracted to obtain security record information.

7. The method according to claim 6, characterized in that Based on the abnormal analysis results, the risk level of the distributed BMS battery system is evaluated and a risk assessment report is generated. The risk assessment report includes the risk type and the probability of risk occurrence, including: Based on the abnormal analysis results, the risk level of the distributed BMS battery system is evaluated, and based on the risk level, the risk scenarios and the probability of occurrence of the risk scenarios are simulated using a Monte Carlo simulation method, wherein the risk scenarios include overcharge risk scenarios, over-discharge risk scenarios, and temperature abnormality risk scenarios; Using the Markov chain method, the transition probability of the distributed BMS battery system between different risk states is defined, and the risk evolution trend of the distributed BMS battery system is predicted. Based on the transition probability and the risk evolution trend, the risk propagation analysis result is obtained. The risk state includes a normal state, an abnormal state, and a fault state. A risk assessment report is generated based on the risk level, the risk scenario, the probability of occurrence of the risk scenario, and the risk propagation analysis result.

8. A distributed BMS battery active balancing safety management system, characterized in that: include: The monitoring module is used to monitor the voltage, temperature and internal resistance of each single battery in real time, obtain battery data, perform status evaluation and trend prediction on the battery data, and obtain the status information and performance change trend corresponding to each single battery; A construction module is used to construct a target battery pack topology model based on the state information and the performance change trend, generate an intelligent agent model based on key information of the internal structure of the battery pack in the target battery pack topology model and a reinforcement learning algorithm, select a preliminary power balancing strategy using the intelligent agent model, optimize the preliminary power balancing strategy, and obtain an optimal power balancing strategy; A determination module is used to optimize the regulation of energy flow between each single battery using a fuzzy logic controller based on the optimal power balancing strategy to obtain an optimized regulation result, determine an energy flow management plan based on the optimized regulation result, and introduce sliding mode variable structure control theory and optimized energy transfer strategy in the regulation process for optimization; an identification module, used for applying the energy flow management solution to obtain feedback data, constructing an anomaly detection model based on the feedback data and historical data using a deep belief network, identifying abnormal operation modes using the anomaly detection model, analyzing the abnormal operation modes, obtaining an abnormal analysis result, generating a risk assessment report based on the abnormal analysis result, constructing a security record platform based on blockchain technology, and obtaining security record information based on the security record platform; A generation module is used to optimize the parameters of each model based on the risk assessment report and the safety record information to obtain an optimized model, and to generate a safety management system by combining the optimized model, the fuzzy logic controller, the safety record platform, the transfer learning algorithm and the quantum computing method. The various models include the battery pack topology model, the intelligent agent model and the anomaly detection model.

9. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a distributed BMS battery active balancing safety management method as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a computer, a distributed BMS battery active balancing safety management method as described in any one of claims 1 to 7 is implemented.

Images